diff --git a/paper-releases/pbisc-diffusion/updates/20261015T141849Z-rare-cell-7k-8k-panels/diffusion/dataset/server_1/model8/plans/rare_gene_context_expansion_20261015/research/ilc_dc/sources/PMC7445430.txt b/paper-releases/pbisc-diffusion/updates/20261015T141849Z-rare-cell-7k-8k-panels/diffusion/dataset/server_1/model8/plans/rare_gene_context_expansion_20261015/research/ilc_dc/sources/PMC7445430.txt
new file mode 100644
index 0000000000000000000000000000000000000000..e1d8970a080f665a639186f6d9ab8b8f1f492536
--- /dev/null
+++ b/paper-releases/pbisc-diffusion/updates/20261015T141849Z-rare-cell-7k-8k-panels/diffusion/dataset/server_1/model8/plans/rare_gene_context_expansion_20261015/research/ilc_dc/sources/PMC7445430.txt
@@ -0,0 +1,290 @@
+Human dendritic cells (DCs) are sentinel cells of the immune system specialized in controlling T cell function ( Banchereau and Steinman, 1998 ;  Palucka and Banchereau, 2013 ;  Steinman et al., 2003 ). The mouse model has brought important concepts to our understanding of DCs and suggests that multiple DC subsets arising from specialized ontogenetic pathways are endowed with specific immune functions ( Briseño et al., 2016 ;  Guermonprez et al., 2019 ;  Merad et al., 2013 ;  Murphy et al., 2016 ).
+
+Definition of human DC subsets is a prerequisite to understanding the division of labor underpinning induction of various types of immune responses. At homeostasis, conventional DCs (cDCs) include cDC1s (CD141 + XCR1 + CLEC9A + IRF8 + ) and cDC2s (CD1c + CD11c + CD172a + IRF4 + ) ( Bachem et al., 2010 ;  Crozat et al., 2010 ;  Heidkamp et al., 2016 ;  Jongbloed et al., 2010 ;  Schlitzer et al., 2013 ). cDC1s and cDC2s arise through a specialized ontogenetic pathway from a common DC precursor (CDP) ( Lee et al., 2015b ) or from early IRF8 +  multipotent lympho-myeloid progenitors (MLPs) ( Helft et al., 2017 ;  Lee et al., 2017 ). Bone marrow progenitors for cDCs generate a common circulating precursor that progressively diverges into pre-cDC1 and pre-cDC2 ( Breton et al., 2015 ,  2016 ;  See et al., 2017 ). This is further complicated by inclusion of AXL + SIGLEC6 + CD11c + CD1c +  cells (AS-DCs also called type 5 DCs), which have been proposed to act as precursors for cDCs (pre-cDCs) or a lineage on its own ( See et al., 2017 ;  Villani et al., 2017 ).
+
+An additional layer of complexity in the DC network lies in its responsiveness to perturbations. For instance, inflammation affects hematopoiesis and phagocyte trafficking, resulting in leukocyte mobilization and tissue infiltration. Specifically, inflammation affects DC diversity and triggers mobilization of CD14 + CD1c +  DCs, called inflammatory DCs (iDCs) ( Binnewies et al., 2019 ;  Granot et al., 2017 ;  Segura et al., 2012 ,  2013 ;  Wollenberg et al., 1996 ;  Zaba et al., 2009 ). The expression of CD1c lectin is shared between iDCs and cDCs. However, CD1c +  iDCs also express multiple monocytic markers, such as CD14, CCR2, and FcγRI/CD64. iDCs have been reported in inflamed skin, synovial fluid, ovarian cancer ascites, solid tumor infiltrates, and lymph nodes ( Bakdash et al., 2016 ;  Binnewies et al., 2019 ;  Granot et al., 2017 ;  Lavin et al., 2017 ;  Segura et al., 2012 ,  2013 ;  Wollenberg et al., 1996 ;  Zaba et al., 2009 ). The developmental pathway of human CD1c + CD14 +  iDCs is poorly understood.  In vitro  studies suggest that iDCs obtained after differentiation of CD14 +  monocytes in granulocyte macrophage colony-stimulating factor (GM-CSF) and interleukin-4 (IL-4) ( Sallusto and Lanzavecchia, 1994 ) might correspond to  in vivo  iDCs ( Granot et al., 2017 ;  Segura et al., 2012 ,  2013 ). In this context, IL-4 acts through induction of the transcriptional regulator NCOR2 ( Sander et al., 2017 ). In addition, triggering the aryl hydrocarbon receptor in monocytes supports activation of IRF4-dependent differentiation of iDCs ( Goudot et al., 2017 ). Together, these studies support the prevailing notion that CD14 +  monocytes act as immediate precursors for iDCs.
+
+Re-evaluation of circulating mononuclear phagocyte diversity has been enabled by single-cell RNA sequencing (scRNA-seq). Recent studies have revealed that a subset of DC-like cells, called DC3s, express mRNA for the CD14 and CD1c genes ( Villani et al., 2017 ). However, this analysis was performed after excluding cells expressing the highest amount of CD14 ( Villani et al., 2017 ). As a consequence, this approach renders a problematic distinction between DC3s and  bona fide  CD14 +  monocytes ( Villani et al., 2017 ). This discrimination is further complicated by previous reports of CD14 + CD1c +  “inflammatory” DCs recruited at inflammatory sites ( Binnewies et al., 2019 ;  Granot et al., 2017 ;  Segura et al., 2012 ,  2013 ;  Wollenberg et al., 1996 ;  Zaba et al., 2009 ).
+
+Here we intended to re-evaluate the definition of DC3s using unbiased scRNA-seq and high-dimensional flow cytometry by exploring the full spectrum of CD14 and CD1c expression. In addition, we identify DC3 growth factor requirements and developmental pathways. Finally, we show that DC3s activate CD103 +  T cells and that DC3 infiltration in human breast tumors correlates with the abundance of CD8 + CD103 + CD69 +  tissue-resident memory (T RM ) T cells.
+
+To probe the diversity of CD16 − CD141 − CD123 −  blood mononuclear phagocytes, we developed a sorting strategy including all phenotypic intermediates between CD14 hi CD1c lo  and CD14 lo CD1c hi  cells. The proportions between cell populations were compensated to enrich in less abundant CD14 lo CD1c hi  cells ( Figure S1 A). Flow cytometry-sorted cells isolated from blood were analyzed using a droplet-based scRNA-seq approach ( Figure 1 A;  Figure S1 A). We found that cells expressing CD14 and/or CD1c could be separated into four CD33 +  clusters (A, B, C, and D) ( Figure 1 A;  Figure S1 B). Contaminating clusters containing B and T lymphocytes and neutrophils were excluded from the analysis ( Figure S1 B). Hierarchical clustering performed on averaged single cell expression data within clusters showed that A and B were closer to each other than any of the other subsets ( Figures 1 B–1D). Cluster D fell between the group of clusters A and B and cluster C ( Figure 1 B). Classical cDC2 markers, such as C LEC10A ,  FCER1A , and major histocompatibility complex (MHC) class II genes ( HLA-DQA ,  HLA-DPA ,  HLA-DRA , and the MHC class II-associated invariant chain  CD74 ) were expressed prominently in clusters A, B, and D ( Figures 1 D and 1E;  Figure S1 C;  Heidkamp et al., 2016 ;  Lavin et al., 2017 ;  Segura et al., 2013 ). In contrast, monocytic markers such as  CD14 ,  S100A8 ,  S100A9 ,  S100A12 , and  VCAN  were more expressed in clusters C and D, with higher expression in C compared with D ( Figures 1 D and 1E). Finally, expression of the C5 receptor  C5AR1  (CD88) was found to be restricted to cluster C together with  SOD2  and  RBP7  ( Figures 1 D and 1E). Figure 1 DC3s Are a Discrete Subset of CD88 − CD1c + CD163 +  Cells in Human Peripheral Blood (A) Gating strategy used to define mononuclear phagocytes expressing CD14 and/or CD1c. Cells expressing CD14 and/or CD1c were sorted by flow cytometry from 3 healthy donors and pooled before scRNA-seq analysis. To improve the resolution of CD1c +  subsets, the cellular input was enriched in CD1 high  cells ( Figure S1 A). Single cells were isolated using a droplet-based approach and sequenced. Dimensionality reduction of scRNA-seq data was performed using dimensionality reduction (t-distributed stochastic neighbor embedding [tSNE]). Clusters A, B, C, and D were identified using the shared nearest neighbor (SNN) clustering algorithm. Each dot represents an individual cell (n = 1,622). (B) Hierarchal clustering of groups A, B, C, and D based on average gene expression (14,933 genes). (C) Absolute number of differentially expressed genes (DEGs) for pairwise comparisons between groups A, B, and D. (D) Heatmaps displaying relative expression of up to 20 DEGs defining each cluster. (E) Violin plots illustrating expression probability distributions across clusters of representative DEGs (226 total DEGs). Feature plots display the average expression of groups of genes (identified in violin plots) in each cell of the tSNE plot defined in (A). (F) Expression distribution across clusters A, B, C, and D of gene signatures identified by  Villani et al. (2017)  and  Yin et al. (2017) . ( ∗∗ p < 0.01,  ∗∗∗∗ p < 0.0001, one-way ANOVA test) (G) Identification of 4 subsets within CD14 lo to hi  CD1c lo to hi  cells by unsupervised clustering of flow cytometry data using the FlowSOM algorithm. tSNE and unsupervised clustering were performed using the following markers: CD88, CD1c, FcεRI, CD14, CD163, BTLA, CD123, and CD5. tSNE plots (right) display the relative expression of each marker among the subsets. Dot plots (below) show the expression of specific markers in clusters 1, 2, and 3 when combined in 2-dimensional analysis. (H) Improved gating strategy for identification of cDC2s, DC3s, and CD14 +  monocytes in circulating PBMCs and histograms showing expression of S100A8/9, FcεRI, CD5, CD14, CD116, and CD206. (I) Principal-component analysis (PCA) for bulk-sequenced mononuclear phagocyte populations as defined in (H). (J) Cluster dendrogram of the different cell types using the 2,000 most variable genes. (K) Heatmaps comparing the relative expression of markers discriminating clusters in scRNA-seq analysis (A, B, C, and D, left) and in bulk RNA-seq analysis on sorted subsets based on the gating strategy defined in (H) (right). See also  Figure S1  and  Table S1 .
+
+DC3s Are a Discrete Subset of CD88 − CD1c + CD163 +  Cells in Human Peripheral Blood
+
+(A) Gating strategy used to define mononuclear phagocytes expressing CD14 and/or CD1c. Cells expressing CD14 and/or CD1c were sorted by flow cytometry from 3 healthy donors and pooled before scRNA-seq analysis. To improve the resolution of CD1c +  subsets, the cellular input was enriched in CD1 high  cells ( Figure S1 A). Single cells were isolated using a droplet-based approach and sequenced. Dimensionality reduction of scRNA-seq data was performed using dimensionality reduction (t-distributed stochastic neighbor embedding [tSNE]). Clusters A, B, C, and D were identified using the shared nearest neighbor (SNN) clustering algorithm. Each dot represents an individual cell (n = 1,622).
+
+(B) Hierarchal clustering of groups A, B, C, and D based on average gene expression (14,933 genes).
+
+(C) Absolute number of differentially expressed genes (DEGs) for pairwise comparisons between groups A, B, and D.
+
+(D) Heatmaps displaying relative expression of up to 20 DEGs defining each cluster.
+
+(E) Violin plots illustrating expression probability distributions across clusters of representative DEGs (226 total DEGs). Feature plots display the average expression of groups of genes (identified in violin plots) in each cell of the tSNE plot defined in (A).
+
+(F) Expression distribution across clusters A, B, C, and D of gene signatures identified by  Villani et al. (2017)  and  Yin et al. (2017) . ( ∗∗ p < 0.01,  ∗∗∗∗ p < 0.0001, one-way ANOVA test)
+
+(G) Identification of 4 subsets within CD14 lo to hi  CD1c lo to hi  cells by unsupervised clustering of flow cytometry data using the FlowSOM algorithm. tSNE and unsupervised clustering were performed using the following markers: CD88, CD1c, FcεRI, CD14, CD163, BTLA, CD123, and CD5. tSNE plots (right) display the relative expression of each marker among the subsets. Dot plots (below) show the expression of specific markers in clusters 1, 2, and 3 when combined in 2-dimensional analysis.
+
+(H) Improved gating strategy for identification of cDC2s, DC3s, and CD14 +  monocytes in circulating PBMCs and histograms showing expression of S100A8/9, FcεRI, CD5, CD14, CD116, and CD206.
+
+(I) Principal-component analysis (PCA) for bulk-sequenced mononuclear phagocyte populations as defined in (H).
+
+(J) Cluster dendrogram of the different cell types using the 2,000 most variable genes.
+
+(K) Heatmaps comparing the relative expression of markers discriminating clusters in scRNA-seq analysis (A, B, C, and D, left) and in bulk RNA-seq analysis on sorted subsets based on the gating strategy defined in (H) (right).
+
+See also  Figure S1  and  Table S1 .
+
+To determine whether our clustering reflects previously published data, we decided to evaluate the expression of gene signatures obtained from  Villani et al. (2017) . We found that signature genes discriminating cDC2 within CD14 lo  cells were mostly represented in clusters A and B, confirming their identity ( Figure 1 F). Likewise, signature genes defining DC3s within CD14 lo  cells ( Villani et al., 2017 ) and CD14 +  monocytes were significantly enriched in cluster D and cluster C, respectively ( Figure 1 F). Of note, genes enriched in DC3s compared with cDC2s ( S100A8 ,  S100A9 , and  CD14 ) were highly expressed in clusters C and D ( Figure 1 E;  Figure S1 E).This underlines the need to integrate a monocyte reference in any comparison aiming to define DC3s. We conclude that scRNA-seq analysis identifies DC3s (cluster D) as a specific subset sharing transcriptional features of cDC2s (clusters A and B) and monocytes (cluster C).
+
+Next we sought to define a flow cytometry-based strategy enabling analysis and prospective isolation of DC3s in blood. To this end, we performed an unsupervised flow cytometry data analysis based on genes identified by scRNA-seq (e.g., CD88, CD14, FcεRI, and CD1c) as well as markers previously associated with cDC2s (e.g., BTLA and CD5) ( Yin et al., 2017 ) and DC3s (e.g., CD163) ( Villani et al., 2017 ), even though they were not detected in the scRNA-seq analysis. We identified three main clusters (1, 2, and 3) together with a rarer cluster (4) of CD5 hi  cells ( Figure 1 G;  Figure S1 D). Cluster 4 appeared to be CD123 hi  contamination of AS-DCs ( Figure S1 D). Cluster 1 highly expressed CD88, aligning with monocyte cluster C identified by scRNA-seq. Cluster 2 did not express the monocyte-associated markers CD14 and CD88 but was characterized by expression of CD1c, FcεRI, and B and T lymphocyte attenuator (BTLA) ( Figure 1 G). In addition, cluster 2 showed heterogeneous expression of CD5 ( Figure 1 G). Similar to cluster 2, cluster 3 did not express the monocytic marker CD88 and displayed higher amounts of CD1c and FcεRI. However, cluster 3 could be distinguished from cluster 2 by higher expression of CD163 and heterogeneous expression of CD14 ( Figure 1 G). The unsupervised flow cytometry analysis allowed us to define a simple gating strategy enabling prospective isolation of monocytes (CD88 + CD14 + ), cDC2s (CD5 +  and CD5 − CD88 − CD1c + BTLA + CD163 − ), and DC3s (CD88 − CD1c + BTLA − CD163 + ) ( Figure 1 H).
+
+To further validate the flow cytometry-based identification of cell subsets, we performed bulk RNA-seq analysis of sorted monocytes, cDC2s (CD5 +  and CD5 − ), DC3s, plasmacytoid DCs (pDCs), and AS-DCs ( Figure S1 E). Principal-component analysis (PCA) indicated that, overall, CD5 +  and CD5 −  cDC2s, DC3s, and monocytes separated from pDCs and AS-DCs along the PC1 axis, accounting for 61% variance ( Figure 1 I). CD5 +  and CD5 −  cDC2 clustered closely together, and DC3s separated clearly from cDC2s and monocytes ( Figure 1 I;  Figure S1 F). Hierarchical clustering (HC) and differentially expressed gene analysis led to the same conclusion, with DC3s sitting between monocytes and cDC2s (CD5 +  and CD5 − ) ( Figure 1 J;  Figure S1 G). Overall, DC3s are closer to cDC2s than monocytes ( Figures 1 I and 1J).
+
+Together, this validates that cellular clusters isolated by a flow cytometry-based approach align to clusters identified by unbiased scRNA-seq ( Figure 1 K).
+
+We conclude that DC3s are a separable entity within CD1c +  cells, defined by a distinct gene expression profile, and that they can be prospectively isolated using CD88, CD1c, CD163, and BTLA ( Figure S1 H). scRNA-seq and bulk gene expression profiling identify markers shared between DC3s and cDC2s (e.g.,  CLEC10A  and  FCER1A ) and markers shared between DC3s and monocytes (such as  S100A8 ,  S100A9 ,  CD14 , and  CD163 ).
+
+Tumor-infiltrating CD14 + CD1c +  DCs have been reported in multiple instances, including ovarian cancer ascites ( Segura et al., 2013 ), breast cancer ( Michea et al., 2018 ), and melanoma ( Bakdash et al., 2016 ;  Binnewies et al., 2019 ). Therefore, we asked whether CD14 + CD1c +  cells would align with DC3s. To this end, we analyzed mononuclear phagocytes infiltrating luminal breast cancer primary tumors. Using the gating strategy described in  Figure 1 , we found that, after exclusion of CD88 +  monocytes and macrophages, the remaining CD45 + HLA-DR + CD123 − CD88 −  fraction contained cDC1s, CD14 − CD1c + CD163 − CD5 +  cDC2s (CD5 +  cDC2s), and CD14 + CD1c + CD163 + CD5 −  DC3s (CD14 +  DC3s) ( Figure 2 A;  Figure S2 A). At this stage, we observed that CD163 and BTLA were particularly susceptible to enzymatic digestion of solid tissue, preventing consistent and reliable quantification of cells throughout the cohort of samples. Hence, we adopted a more restrictive definition of cDC2s and DC3s as CD1c + CD14 − CD5 +  and CD1c + CD14 + CD5 −  cells, respectively ( Figure 2 A). This strategy enabled isolation of CD5 +  cDC2s and CD14 +  DC3s in peripheral tissues even though it might result in underestimation of their absolute numbers. Nevertheless, both subsets aligned phenotypically with their blood counterparts ( Figures S2 B and S2C) and expressed markers reported previously for CD1c + CD14 +  iDCs, such as CD11c and FcεRI ( Figure 2 A;  Segura et al., 2013 ). Figure 2 DC3s Infiltrate Human Breast Tumors (A) Representative gating strategy used to define macrophages, CD5 +  cDC2s, and CD14 +  DC3s and histograms showing the expression of CD163, FcεRI, BTLA, and CD11c in human breast cancer primary tumors. (B) Violin plot quantifying cDC1, CD5 + cDC2, CD14 +  DC3, and CD14 + CD88 +  macrophage subsets identified in (A) in human breast cancer primary tumors (n = 25;  ∗ p < 0.05,  ∗∗ p < 0.01,  ∗∗∗ p < 0.001,  ∗∗∗∗ p < 0.0001, one-way ANOVA test). (C) Pearson correlations of the frequencies of macrophages, cDC1s, CD5 + cDC2s, and CD14 +  DC3s within HLA-DR +  cells in human breast cancer primary tumors (red, significantly correlated p < 0.05; black, not correlated). (D) HC showing the relative expression of markers used for subset identification in  Figure 1  in CD1c + , CD1c + CD14 + , and CD14 +  cells from invaded lymph nodes draining human breast cancer primary tumors. (E) GSEA of pairwise comparisons of CD1c + CD14 +  cells with CD1c +  or CD14 +  from invaded lymph nodes draining human breast cancer primary tumors. Gene signatures of blood DC3s compared with cDC2s (DC3 > cDC2) or CD14 +  monocytes (DC3 > Mono) and, vice versa, of blood cDC2s (cDC2 > DC3) or CD14 +  monocytes (Mono > DC3) compared with DC3s were used ( Villani et al., 2017 ). See also  Figure S2  and  Table S2 .
+
+DC3s Infiltrate Human Breast Tumors
+
+(A) Representative gating strategy used to define macrophages, CD5 +  cDC2s, and CD14 +  DC3s and histograms showing the expression of CD163, FcεRI, BTLA, and CD11c in human breast cancer primary tumors.
+
+(B) Violin plot quantifying cDC1, CD5 + cDC2, CD14 +  DC3, and CD14 + CD88 +  macrophage subsets identified in (A) in human breast cancer primary tumors (n = 25;  ∗ p < 0.05,  ∗∗ p < 0.01,  ∗∗∗ p < 0.001,  ∗∗∗∗ p < 0.0001, one-way ANOVA test).
+
+(C) Pearson correlations of the frequencies of macrophages, cDC1s, CD5 + cDC2s, and CD14 +  DC3s within HLA-DR +  cells in human breast cancer primary tumors (red, significantly correlated p < 0.05; black, not correlated).
+
+(D) HC showing the relative expression of markers used for subset identification in  Figure 1  in CD1c + , CD1c + CD14 + , and CD14 +  cells from invaded lymph nodes draining human breast cancer primary tumors.
+
+(E) GSEA of pairwise comparisons of CD1c + CD14 +  cells with CD1c +  or CD14 +  from invaded lymph nodes draining human breast cancer primary tumors. Gene signatures of blood DC3s compared with cDC2s (DC3 > cDC2) or CD14 +  monocytes (DC3 > Mono) and, vice versa, of blood cDC2s (cDC2 > DC3) or CD14 +  monocytes (Mono > DC3) compared with DC3s were used ( Villani et al., 2017 ).
+
+See also  Figure S2  and  Table S2 .
+
+DC3s were consistently identified in 25 samples of primary tumors of clinical stages I, II, and III ( Figure S2 D). Macrophages represented by far the most abundant population. DC3s outnumbered cDC1s but were on par with  bona fide  CD5 +  cDC2s ( Figure 2 B). The relative abundance of DC3s did not correlate with disease progression ( Figure S2 D) or with macrophages or cDCs ( Figure 2 C). In contrast, cDC1 infiltration correlated with cDC2s ( Figure 2 C).
+
+Bulk RNA-seq analysis of CD1c + CD14 +  cells sorted from tumor-invaded lymph nodes indicated that they displayed a similar expression profile as blood DC3s (high expression of cDC2 markers such as  CLEC10A  and  FCER1A  combined with low expression of monocyte-associated markers such as  C5AR1  and  SOD2 ) ( Figure 2 D). Gene set enrichment analysis (GSEA) revealed that CD14 + CD1c +  were more enriched for the DC3 > cDC2 signature compared with CD1c +  cells (normalized enrichment score [NES], 2.06; p = 0.0;  Subramanian et al., 2005 ;  Villani et al., 2017 ). In addition, CD14 + CD1c +  were more enriched for the DC3 > monocyte (Mono) signature compared with CD14 +  cells (NES, 1.42; p = 0.02;  Villani et al., 2017 ;  Figure 2 E). Conversely, compared with CD1c + CD14 + , the cDC2 > DC3 and Mono > DC3 gene signatures were enriched in CD1c +  (NES, −1.57; p = 0.004) and CD14 +  (NES, −2.07; p = 0.0) cells, respectively ( Figure 2 E).
+
+We conclude that CD11c + FcεRI  + CD14 + CD1c +  iDCs infiltrating breast cancer align with DC3s.
+
+Inflammatory cues promoting mobilization of CD14 + CD1c +  cells at the site of inflammation are not fully defined. Mouse studies define GM-CSF as a likely candidate ( Mach et al., 2000 ;  Menezes et al., 2016 ). For this reason, we decided to test whether GM-CSF was sufficient to mobilize human CD14 + CD1c +  DCs in a humanized mouse metastatic lung model. We generated B16 mouse melanoma engineered to overexpress human GM-CSF (B16_huGM) or FLT3L (B16_huFLT3L) ( Figure S3 A). Immunodeficient NOD.Cg- Prkdc scid 
+ Il2rg tm1Wjl /SzJ (NSG) mice were injected intravenously with B16 control (CTRL), B16_huGM, or B16_huFLT3L ( Figure 3 A). Lung metastasis-bearing mice were engrafted with human peripheral blood mononuclear cells (PBMCs). Two days later, human CD45 +  leukocytes were found in lung tumor foci and juxta-tumor areas ( Figure 3 B;  Figures S3 B and S3C). Flow cytometry analysis of metastatic lungs showed that FLT3L, but not GM-CSF, promotes expansion of CD1c + CD5 +  cells aligning with blood cDC2s ( Figure 3 C, blue). In contrast, GM-CSF, but not FLT3L, led to accumulation of CD1c + CD14 +  cells aligning with blood DC3s ( Figure 3 C, orange;  Figure S3 D, orange). As shown for circulating peripheral blood subsets in  Figure 1 , FLT3L-dependent cDC2s and GM-CSF-dependent DC3s shared expression of Clec10A, CD11c, and FcεRI ( Figure 3 C;  Figure S3 E). All tumor-bearing lungs contained some monocytes and/or macrophages (CD14 + CD88 + ) ( Figure 3 C;  Figure S3 B). To further establish alignment of GM-CSF-mobilized DC3s, we performed an unbiased scRNA-seq on human CD45 + HLA-DR +  cells expressing CD14 and/or CD1c ( Figures S3 F and S3G). Two major clusters could be identified. Cluster 0 was characterized by expression of genes associated with DC3s and cDC2s, such as  CLEC10A ,  FCER1A , and  CD74  ( Figures S3 H and S3I), and was enriched in DC3 differentially expressed transcripts ( Villani et al., 2017 ;  Figure S3 J). Conversely, cluster 1 expressed markers defining monocytes ( SOD2 ,  C5AR1 , and  G0S2 ) and  CXCL2 ,  CXCL3 , and  CXCL8  chemokines ( Figures S3 H and S3I). We conclude that GM-CSF drives mobilization of CD14 + CD1c +  cells aligning phenotypically with circulating DC3s. Figure 3 DC3s Give Rise to CD14 + CD1c +  DCs at Inflammatory Sites (A) Experimental model. NSG mice were injected intravenously (i.v.) with B16_CTRL, B16_huFLT3L, or B16_huGM on day 0. On days 7 and 8, 10 8  human PBMCs were injected i.v. Metastatic lungs were collected on day 9. (B) Pseudocolor images of B16_huGM (green) metastatic lung on day 9 post-injection, stained for human CD45 (red). Nuclei were stained with Hoechst (blue). Scale bar, 100 μm. (C) Gating strategy for cDC2 and DC3 identification in B16_huGM and B16_huFLT3L metastatic mouse lung and histograms showing the expression of CD163, CD206, and Clec10A. The bar graph summarizes the frequency of cDC2s and DC3s among total HLA-DR +  cells in metastatic B16_CTRL, B16_huFLT3L, or B16_huGM mouse lungs (n = 3 independent mice;  ∗∗ p < 0.01,  ∗∗∗ p < 0.001,  ∗∗∗∗ p < 0.0001, one-way ANOVA test). (D) Flow cytometry analysis of flow cytometry-sorted cDC2s, DC3s, and CD14 +  monocytes after 2 days of culture with MS5 stromal cells expressing human GM-CSF (MS5_GM). Bar graphs show the frequency of output cells among total huCD45 +  cells (n = 4–5 healthy donors). (E) Histograms showing CD14 expression on cDC2s, DC3s, and CD14 +  monocytes before and after 2 days of coculture with MS5_GM and bar graphs summarizing the frequency of CD14 expression within each cell type (n = 5 healthy donors,  ∗∗ p < 0.01, Mann-Whitney two-tailed t test). See also  Figures S3  and  S4  and  Table S1 .
+
+DC3s Give Rise to CD14 + CD1c +  DCs at Inflammatory Sites
+
+(A) Experimental model. NSG mice were injected intravenously (i.v.) with B16_CTRL, B16_huFLT3L, or B16_huGM on day 0. On days 7 and 8, 10 8  human PBMCs were injected i.v. Metastatic lungs were collected on day 9.
+
+(B) Pseudocolor images of B16_huGM (green) metastatic lung on day 9 post-injection, stained for human CD45 (red). Nuclei were stained with Hoechst (blue). Scale bar, 100 μm.
+
+(C) Gating strategy for cDC2 and DC3 identification in B16_huGM and B16_huFLT3L metastatic mouse lung and histograms showing the expression of CD163, CD206, and Clec10A. The bar graph summarizes the frequency of cDC2s and DC3s among total HLA-DR +  cells in metastatic B16_CTRL, B16_huFLT3L, or B16_huGM mouse lungs (n = 3 independent mice;  ∗∗ p < 0.01,  ∗∗∗ p < 0.001,  ∗∗∗∗ p < 0.0001, one-way ANOVA test).
+
+(D) Flow cytometry analysis of flow cytometry-sorted cDC2s, DC3s, and CD14 +  monocytes after 2 days of culture with MS5 stromal cells expressing human GM-CSF (MS5_GM). Bar graphs show the frequency of output cells among total huCD45 +  cells (n = 4–5 healthy donors).
+
+(E) Histograms showing CD14 expression on cDC2s, DC3s, and CD14 +  monocytes before and after 2 days of coculture with MS5_GM and bar graphs summarizing the frequency of CD14 expression within each cell type (n = 5 healthy donors,  ∗∗ p < 0.01, Mann-Whitney two-tailed t test).
+
+See also  Figures S3  and  S4  and  Table S1 .
+
+We next wondered whether GM-CSF could induce  trans -differentiation of circulating cDC2s or monocytes into CD14 + CD1c +  DC3s. To test this hypothesis, cDC2s, DC3s, and monocytes were sorted by flow cytometry from blood and cocultured  in vitro  in the presence of GM-CSF-expressing stromal cells (MS5_GM). After 2 days, cDC2s did not acquire CD163 or CD14 expression ( Figure 3 D). In contrast, DC3s did upregulate CD14 ( Figures 3 D and 3E). We found that monocytes differentiated into CD16 +  and CD16 −  macrophages ( Figure 3 D). Importantly, CD88 + CD1c − CD14 +  monocytes did not give rise to CD1c hi CD163 hi  DC3s ( Figure 3 D). Addition of IL-4 to GM-CSF culture did not affect the outcome ( Figure S4 A). The lack of  trans -differentiation of cDC2s or monocytes into DC3s was also confirmed  in vivo  upon adoptive transfer into immunodeficient NSG mice carrying GM-CSF-expressing tumors ( Figures S4 B and S4C). Overall, we conclude that CD88 − CD14 + CD1c +  cells can differentiate from DC3s independent of cDC or monocytic lineages.
+
+We next wondered how DC3s would differentiate from bone marrow progenitors. To this end, we sought to define a culture system capable of generating DC3s together with cDC2s and macrophages. We cocultured human cord blood-derived CD34 +  hematopoietic stem and progenitor cells (HSPCs) with stromal cells engineered to overexpress cDC-promoting factors (membrane-bound FLT3L together with stem cell factor (SCF) and CXCL12 [MS5_FS12];  Anselmi et al., 2020 ) in the presence or absence of GM-CSF ( Figure 4 A;  Figure S5 A). We found that GM-CSF increased differentiation of CD1c + CD14 + CD163 +  cells (orange) phenotypically aligning with blood DC3s ( Figures 4 A and 4B). CD1c + CD14 −  cells did not express CD163, suggesting that they mostly align with cDC2s. In addition, CD163 expression was restricted to CD1c + CD14 +  cells ( Figure 4 B). GM-CSF alone (MS5_GM), but not FLT3L, was sufficient to induce differentiation of DC3s  in vitro  ( Figure 4 C). In contrast, FLT3L (MS5_FL) was sufficient to induce differentiation of cDC2s ( Figure 4 C). We conclude that cDC2s and DC3s have distinct growth factor requirements. Figure 4 DC3s Differentiate from Hematopoietic Progenitors upon GM-CSF Exposure Independent of Mono-Committed Progenitors (cMoPs) or cDC-Committed Progenitors (CDPs) (A) Flow cytometry analysis of cord blood-derived CD34 +  HSPCs cultured on stromal cells expressing human FLT3L, SCF, and CXCL12 (MS5_FS12) with or without human recombinant GM-CSF (MS5_FS12+recGM-CSF) for 14 days. (B) Flow cytometry plots of BTLA and CD163 expression within CD1c + CD14 −  and CD1c + CD14 +  cells identified in (A). Bar graphs summarize the absolute numbers of differentiated CD1c + CD14 −  cDC2s and CD1c + CD14 +  DC3s (a line represents the median; n = 6 independent cord blood donors,  ∗ p < 0.05, Wilcoxon test). (C) Bar graphs summarizing the absolute numbers of CD1c + CD14 − CD163 −  cDC2s and CD1c + CD14 + CD163 +  DC3s differentiated from cord blood-derived CD34 +  HSPCs cocultured for 14 days with stromal cells expressing human FLT3L (MS5_FL), GM-CSF (MS5_GM), or neither (MS5_CTRL) (a line represents the median; n = 5–7 independent cord blood donors,  ∗ p < 0.05,  ∗∗ p < 0.01,  ∗∗∗ p < 0.001, Wilcoxon test). (D) HC based on 19,791 protein-coding genes of  in-vitro- generated subsets differentiated from CD34 +  HSPCs cultured with MS5_FS12 supplemented with human recombinant GM-CSF (MS5_FS12+recGM-CSF). Each dot represents an average of three donors. (E) Volcano plots showing the DEGs between  in-vitro- generated DC3s cells (orange) compared with cDC2s (blue and turquoise, left plot) or macrophages (gray and brown, right plot). Genes with Log 2 (fold change, FC) > ±2 and a false discovery rate (FDR)-adjusted p value of less than 0.05 were considered significant. (F) GSEA of pairwise comparisons of DC3s with cDC2s or macrophages generated i n vitro . Gene signatures (gene set) defining genes upregulated in blood DC3s compared with cDC2s (DC3 > cDC2) or blood DC3s compared with CD14 +  monocytes (DC3 > CD14 Mono) were used ( Villani et al., 2017 ) (NES, normalized enrichment score). (G) BubbleMap summarizing the enrichment of defined gene sets in pairwise comparisons of  in-vitro- differentiated DC3s versus  in vitro  cDC2s or  in vitro  macrophages. Gene signatures (gene sets) of blood DC3s compared with cDC2s (DC3 > cDC2) or CD14 +  monocytes (DC3 > Mono) and, vice versa, of blood cDC2s (cDC2 > DC3) or CD14 +  monocytes (Mono > DC3) compared with DC3s were used ( Villani et al., 2017 ). (H) Flow cytometry analysis of cord blood-derived CDPs, cMoPs, and GMDPs cultured for 7 days with MS5_FS12 or MS5_GM. Bar graphs summarize the absolute number of differentiated cells from each progenitor (a line represents the median, n = 4–7 independent cord blood donors). See also  Figure S5  and  Table S1 .
+
+DC3s Differentiate from Hematopoietic Progenitors upon GM-CSF Exposure Independent of Mono-Committed Progenitors (cMoPs) or cDC-Committed Progenitors (CDPs)
+
+(A) Flow cytometry analysis of cord blood-derived CD34 +  HSPCs cultured on stromal cells expressing human FLT3L, SCF, and CXCL12 (MS5_FS12) with or without human recombinant GM-CSF (MS5_FS12+recGM-CSF) for 14 days.
+
+(B) Flow cytometry plots of BTLA and CD163 expression within CD1c + CD14 −  and CD1c + CD14 +  cells identified in (A). Bar graphs summarize the absolute numbers of differentiated CD1c + CD14 −  cDC2s and CD1c + CD14 +  DC3s (a line represents the median; n = 6 independent cord blood donors,  ∗ p < 0.05, Wilcoxon test).
+
+(C) Bar graphs summarizing the absolute numbers of CD1c + CD14 − CD163 −  cDC2s and CD1c + CD14 + CD163 +  DC3s differentiated from cord blood-derived CD34 +  HSPCs cocultured for 14 days with stromal cells expressing human FLT3L (MS5_FL), GM-CSF (MS5_GM), or neither (MS5_CTRL) (a line represents the median; n = 5–7 independent cord blood donors,  ∗ p < 0.05,  ∗∗ p < 0.01,  ∗∗∗ p < 0.001, Wilcoxon test).
+
+(D) HC based on 19,791 protein-coding genes of  in-vitro- generated subsets differentiated from CD34 +  HSPCs cultured with MS5_FS12 supplemented with human recombinant GM-CSF (MS5_FS12+recGM-CSF). Each dot represents an average of three donors.
+
+(E) Volcano plots showing the DEGs between  in-vitro- generated DC3s cells (orange) compared with cDC2s (blue and turquoise, left plot) or macrophages (gray and brown, right plot). Genes with Log 2 (fold change, FC) > ±2 and a false discovery rate (FDR)-adjusted p value of less than 0.05 were considered significant.
+
+(F) GSEA of pairwise comparisons of DC3s with cDC2s or macrophages generated i n vitro . Gene signatures (gene set) defining genes upregulated in blood DC3s compared with cDC2s (DC3 > cDC2) or blood DC3s compared with CD14 +  monocytes (DC3 > CD14 Mono) were used ( Villani et al., 2017 ) (NES, normalized enrichment score).
+
+(G) BubbleMap summarizing the enrichment of defined gene sets in pairwise comparisons of  in-vitro- differentiated DC3s versus  in vitro  cDC2s or  in vitro  macrophages. Gene signatures (gene sets) of blood DC3s compared with cDC2s (DC3 > cDC2) or CD14 +  monocytes (DC3 > Mono) and, vice versa, of blood cDC2s (cDC2 > DC3) or CD14 +  monocytes (Mono > DC3) compared with DC3s were used ( Villani et al., 2017 ).
+
+(H) Flow cytometry analysis of cord blood-derived CDPs, cMoPs, and GMDPs cultured for 7 days with MS5_FS12 or MS5_GM. Bar graphs summarize the absolute number of differentiated cells from each progenitor (a line represents the median, n = 4–7 independent cord blood donors).
+
+See also  Figure S5  and  Table S1 .
+
+We next intended to determine whether the transcriptional landscape of  in-vitro -generated DC3s from stromal cell cocultures aligned with their  in vivo  counterparts. Bulk RNA-seq analysis of  in-vitro -generated cells showed that CD14 + CD1c +  DC3s (orange) sat between cDC2s (blue and turquoise) and macrophages (gray and brown) ( Figure 4 D). Of note, CD1c + CD14 −  cells generated  in vitro  were heterogenous for CD206 expression and were therefore analyzed as two independent subsets (CD1c + CD206 −  [blue] and CD1c + CD206 +  [turquoise];  Figure S5 B). However, both subsets displayed very similar transcriptomes regardless of CD206 expression, and the CD1c + CD14 − CD206 +  and CD1c + CD14 − CD206 −  fractions strongly resembled circulating cDC2s ( Anselmi et al., 2020 ).
+
+Further analysis of differentially expressed genes (DEGs) showed that  in vitro  DC3s differed from  in vitro  cDC2s by expression of monocyte-associated markers such as  S100A8 ,  S100A9 ,  S100A12 , and  CD14  ( Figure 4 E;  Figure S5 C). Conversely,  in vitro  cDC2s displayed higher expression of  CD1C ,  LAMP3 ,  CD52 , and  LTB , as reported for lung cDC2s ( Lavin et al., 2017 ;  Figure 4 E;  Figure S5 C). Using the GSEA methodology ( Subramanian et al., 2005 ), we found that the set of genes upregulated in primary DC3s compared with cDC2s (DC3 > cDC2 from  Villani et al., 2017 ) was enriched in GM-CSF-dependent  in vitro  DC3s compared with  in vitro  cDC2s ( Figure 4 F). Of note, GM-CSF-exposed cDC2s did not convert to DC3s ( Figures S5 D and S5E). We found that markers common for primary cDC2s and DC3s ( CLEC10A ,  Figure 1 ) were higher in  in vitro  DC3s compared with  in vitro  macrophages ( Figure 4 E;  Heidkamp et al., 2016 ). Conversely,  in vitro  macrophages expressed more  FCGR1A ,  C5AR1  (CD88),  CXCL8 ,  CXCL1 ,  CXCL2 ,  CXCL3 ,  CCL2 ,  CCL3 , and  CCL7  compared with  in vitro  DC3s ( Ruffell et al., 2009 ;  Figure 4 E;  Figures S5 C and S5F). Genes upregulated in primary DC3s compared with CD14 +  monocytes (DC3 > CD14 +  Mono from  Villani et al., 2017 ) were significantly enriched in  in vitro  DC3s compared with macrophages ( Figures 4 F and 4G). We conclude that GM-CSF drives,  in vitro , the differentiation of DC3s aligning to primary blood DC3s.
+
+The developmental relationship between DC3s and cDCs or monocyte lineages is not known. A classical view defines the development of phagocytes as a stepwise and ordered loss of developmental potential concomitant with lineage commitment. Historically, this process has been identified using prospective isolation of progenitor populations of decreasing potential. Specifically, early granulocyte-monocyte and DC progenitors (GMDPs) carries a tri-lineage potential ( Lee et al., 2015b ). Loss of neutrophil potential defines monocyte and DC progenitors (MDPs) ( Lee et al., 2015b ). Loss of monocyte potential defines CDPs, which generate cDCs via a pre-cDC intermediate ( Breton et al., 2015 ,  2016 ;  Lee et al., 2015b ;  Naik et al., 2007 ;  Onai et al., 2007 ;  See et al., 2017 ). Finally, loss of DC and neutrophil potential defines monocyte-committed progenitors (cMoPs) ( Kawamura et al., 2017 ). Having established the growth factor requirements for DC3 development, we decided to test the contribution of CDPs and cMoPs to generation of DC3s. Flow cytometry-sorted CDPs and cMoPs and the remaining GMDP-containing fraction (CD34 + CD38 + CD45RA + CD123 − CD64 − ) isolated from cord blood-derived HSPCs (see the gating strategy in  Figure S5 G) were cocultured with stromal cells supporting cDC (MS5_FS12) or DC3 (MS5_GM) differentiation.
+
+Flow cytometry analysis after 7 days of culture showed that CDP gave rise exclusively to cDC1s and cDC2s in MS5_FS12 cocultures, as described previously ( Figure 4 H;  Figure S5 H;  Lee et al., 2015b ). cMoP cultures gave rise solely to CD14 + CD1c −  cells, as described previously ( Figure 4 H;  Figure S5 H;  Kawamura et al., 2017 ). As expected, the GMDP-containing fraction gave rise to granulocytes, macrophages, and cDC2s. Importantly, the GMDP-containing fraction also gave rise to DC3s in MS5_GM coculture ( Figure 4 H;  Figure S5 H). Therefore, we asked whether DC3s would arise directly from a multipotent progenitor or via formation of an intermediate DC3-committed progenitor devoid of any other lineage potential.
+
+To address this question in unbiased settings, we developed a single cell culture of CD34 + CD38 + CD123 − CD64 −  progenitors distinct from CDPs or cMoPs ( Figure 5 A). We chose to combine MS5_FS12 with soluble GM-CSF for two reasons: (1) GM-CSF alone did not support growth of individual progenitors ( Figure S5 I), and (2) MS5_FS12 coculture was found to more efficiently support cDC and monocyte differentiation ( Figure 4 G). Figure 5 Single-Cell Analysis of DC3 Commitment (A) Flow cytometry analysis of bulk (500 cells) or single CD34 + CD38 + CD123 − CD64 −  progenitor cells cocultured for 14 days with MS5_FS12 supplemented with recombinant human GM-CSF (MS5_FS12+recGM-CSF). Flow cytometry plots resulting from single CD34 + CD38 + CD123 − CD64 −  progenitor cells with different potentials are shown as representative examples (n = 355 cells from 2 independent experiments). (B) HC of lineage potential from single CD34 + CD38 + CD123 − CD64 −  progenitor cells (n = 355). (C) Bar graph and Venn diagram summarizing the frequency of the potential of mono-, bi-, tri-, or multipotent individual CD34 + CD38 + CD123 − CD64 −  cells within the total wells analyzed (n = 355). (D) Bar graphs summarizing the frequency of mono-, bi-, tri-, or multipotent individual CD34 + CD38 + CD123 − CD64 −  cells among DC3-generating progenitors only. An orange bar represents the frequency of DC3-restricted progenitors. (E) Cell surface phenotype of DC3-restricted progenitors before differentiation cultures inferred by index flow cytometry sorting. tSNE plots display an overlay of total CD45 +  cells (gray) and DC3-restricted progenitor cells (orange) (top left). Shown is relative expression of the markers CD45RA, CD38, CD34, CD10, Clec12A, CD64, CD123, CD163, and SIRPα. (F) Validation experiment for identification of Clec12A as a marker for DC3-committed progenitors. Shown is flow cytometry analysis of bulk-sorted CD34 + CD38 + CD45RA + CD123 − CD64 − Clec12A −  and Clec12A +  cells. 500 cells were cocultured with MS5_GM for 7 days. The bar graph summarizes the number of differentiated DC3s from each bulk population (n = 4 healthy donors,  ∗ p < 0.05, Mann-Whitney two-tailed t test). See also  Figure S5 .
+
+Single-Cell Analysis of DC3 Commitment
+
+(A) Flow cytometry analysis of bulk (500 cells) or single CD34 + CD38 + CD123 − CD64 −  progenitor cells cocultured for 14 days with MS5_FS12 supplemented with recombinant human GM-CSF (MS5_FS12+recGM-CSF). Flow cytometry plots resulting from single CD34 + CD38 + CD123 − CD64 −  progenitor cells with different potentials are shown as representative examples (n = 355 cells from 2 independent experiments).
+
+(B) HC of lineage potential from single CD34 + CD38 + CD123 − CD64 −  progenitor cells (n = 355).
+
+(C) Bar graph and Venn diagram summarizing the frequency of the potential of mono-, bi-, tri-, or multipotent individual CD34 + CD38 + CD123 − CD64 −  cells within the total wells analyzed (n = 355).
+
+(D) Bar graphs summarizing the frequency of mono-, bi-, tri-, or multipotent individual CD34 + CD38 + CD123 − CD64 −  cells among DC3-generating progenitors only. An orange bar represents the frequency of DC3-restricted progenitors.
+
+(E) Cell surface phenotype of DC3-restricted progenitors before differentiation cultures inferred by index flow cytometry sorting. tSNE plots display an overlay of total CD45 +  cells (gray) and DC3-restricted progenitor cells (orange) (top left). Shown is relative expression of the markers CD45RA, CD38, CD34, CD10, Clec12A, CD64, CD123, CD163, and SIRPα.
+
+(F) Validation experiment for identification of Clec12A as a marker for DC3-committed progenitors. Shown is flow cytometry analysis of bulk-sorted CD34 + CD38 + CD45RA + CD123 − CD64 − Clec12A −  and Clec12A +  cells. 500 cells were cocultured with MS5_GM for 7 days. The bar graph summarizes the number of differentiated DC3s from each bulk population (n = 4 healthy donors,  ∗ p < 0.05, Mann-Whitney two-tailed t test).
+
+See also  Figure S5 .
+
+Flow cytometry analysis of 14-day progeny of individual progenitor cultures revealed multiple patterns of developmental potential ( Figure 5 A). Overall, the granulocyte potential was segregated from the potential for mononuclear phagocytes ( Figure 5 B). We found that only 0.3% of progenitors could differentiate into all four lineages corresponding to the GMDP functional definition ( Figure 5 D). Most individual progenitors gave rise to single lineage progeny ( Figures 5 C and 5D). Unipotent wells containing only neutrophils represented the most abundant outcome (39.4%), followed by macrophage-only (26.8%) and cDC-only (5.63%) wells ( Figure 5 C). Most individual progenitors endowed with DC3 potential had multi-lineage potential. DC3 potential was more associated with mononuclear phagocytes rather than granulocyte potential (61.4% and 19.3% of DC3-containing wells, respectively) ( Figure 5 D). Importantly, we also identified a minor fraction of individual progenitors giving rise exclusively to DC3s ( Figures 5 B–5D).
+
+We next aimed to further define the cell surface phenotype of progenitors endowed with DC3 potential. To this end, we attempted to establish correlations between the cell surface phenotype of single sorted cells (inferred from index flow cytometry sorting) and their developmental potential.  A posteriori  identification of the cell surface phenotype of DC3-committed progenitors revealed that they had a CD34 + CD38 + CD45RA int CD123 − CD64 − SIRPα lo CD10 − Clec12A +  phenotype ( Figure 5 E). As validation, we showed that the potential for DC3s lay in the Clec12A +  population of the GMDP-containing fraction ( Figure 5 F). We conclude that DC3s can develop via a DC3-restricted intermediate distinct from cDC-restricted CDPs or monocyte-restricted cMoPs. Even though the existence of a DC3-commited unipotent progenitor is not formally proven, our data are compatible with the notion that DC3 specification arises downstream of MDP.
+
+We next aimed to understand the immunological function of DC3s and to compare it with cDC2s and monocytes. First, we decided to test the responsiveness of DC3s, cDC2s, and monocytes to a cocktail of Toll-like receptor (TLR) agonists. cDC2s, DC3s, and monocytes were sorted by flow cytometry from blood and stimulated overnight  ex vivo . PCA analysis of the total transcriptome of unstimulated and stimulated populations evidenced that all subsets underwent a certain degree of convergence in their transcriptome ( Figure 6 A). In support of this, we found an important overlap in the set of activation-induced genes defined for each subset (1,344 genes;  Figure 6 B). Despite the relative convergence of activated cells, we found that overnight activation did not compromise cell surface discrimination of DC3s from cDC2s and monocytes ( Figure S6 A). Indeed, TLR-activated DC3s could still be discriminated from TLR-activated cDC2s by 437 DEGs or from TLR-activated monocytes by 1,293 genes ( Figure 6 C). The same was true for the pairwise comparison of activated cDC2s and circulating DC3s ( Figure S6 B). In sum, we conclude that innate activation does not trigger conversion of cDC2s or monocytes into DC3s despite induction of a common transcriptional response to TLR stimulation. Figure 6 DC3s Respond to TLR Stimulation We performed bulk RNA-seq analysis of BTLA + CD5 +  and BTLA + CD5 −  cDC2s, DC3s, and monocytes sorted as shown in  Figure 1 H and stimulated overnight (16 h, 3 donors) or not (4 donors) with a TLR agonist cocktail (25 μg/mL poly(I:C), 1 μg/mL R848, and 10 ng/mL LPS). For activation of cDC2s, BTLA + CD5 +  and BTLA + CD5 −  were pooled. (A) PCA analysis for all genes. (B) Venn diagram summarizing the number of activation-induced DEGs upregulated in stimulated compared with unstimulated cells within each cell population. (C) Volcano plots showing DEGs between TLR agonist-stimulated DC3s compared with TLR agonist-stimulated cDC2s or TLR agonist-stimulated monocytes. Genes with Log 2 (FC) > ±2 and a FDR-adjusted p value of less than 0.05 were considered significant. (D) Bar graph summarizing relative CCR7 gene expression within TLR-agonist stimulated or unstimulated mononuclear phagocyte populations (n = 3–4; a line represents the median;  ∗ p < 0.05,  ∗∗ p < 0.01,  ∗∗∗ p < 0.001,  ∗∗∗∗ p < 0.0001, one-way ANOVA test). (E) Histograms showing the frequency and median of fluorescence intensity (MFI) of CD86 on TLR agonist-stimulated or unstimulated mononuclear phagocyte populations. (F) Heatmap showing the relative gene expression of selected costimulatory molecules on TLR agonist-stimulated or unstimulated mononuclear phagocyte populations. (G) Quantification of cytokines and chemokines secreted by cDC2s, DC3s, and CD14 +  monocytes in response to overnight stimulation with a cocktail of TLR agonists (n = 9 healthy donors; a line represents the median;  ∗ p < 0.05,  ∗∗ p < 0,01,  ∗∗∗∗ p < 0,0001, one-way ANOVA test). (H) Heatmap showing the relative gene expression of cytokines and chemokines analyzed in (G) within TLR-agonist stimulated or unstimulated mononuclear phagocyte populations. See also  Figure S6  and  Table S1 .
+
+DC3s Respond to TLR Stimulation
+
+We performed bulk RNA-seq analysis of BTLA + CD5 +  and BTLA + CD5 −  cDC2s, DC3s, and monocytes sorted as shown in  Figure 1 H and stimulated overnight (16 h, 3 donors) or not (4 donors) with a TLR agonist cocktail (25 μg/mL poly(I:C), 1 μg/mL R848, and 10 ng/mL LPS). For activation of cDC2s, BTLA + CD5 +  and BTLA + CD5 −  were pooled.
+
+(A) PCA analysis for all genes.
+
+(B) Venn diagram summarizing the number of activation-induced DEGs upregulated in stimulated compared with unstimulated cells within each cell population.
+
+(C) Volcano plots showing DEGs between TLR agonist-stimulated DC3s compared with TLR agonist-stimulated cDC2s or TLR agonist-stimulated monocytes. Genes with Log 2 (FC) > ±2 and a FDR-adjusted p value of less than 0.05 were considered significant.
+
+(D) Bar graph summarizing relative CCR7 gene expression within TLR-agonist stimulated or unstimulated mononuclear phagocyte populations (n = 3–4; a line represents the median;  ∗ p < 0.05,  ∗∗ p < 0.01,  ∗∗∗ p < 0.001,  ∗∗∗∗ p < 0.0001, one-way ANOVA test).
+
+(E) Histograms showing the frequency and median of fluorescence intensity (MFI) of CD86 on TLR agonist-stimulated or unstimulated mononuclear phagocyte populations.
+
+(F) Heatmap showing the relative gene expression of selected costimulatory molecules on TLR agonist-stimulated or unstimulated mononuclear phagocyte populations.
+
+(G) Quantification of cytokines and chemokines secreted by cDC2s, DC3s, and CD14 +  monocytes in response to overnight stimulation with a cocktail of TLR agonists (n = 9 healthy donors; a line represents the median;  ∗ p < 0.05,  ∗∗ p < 0,01,  ∗∗∗∗ p < 0,0001, one-way ANOVA test).
+
+(H) Heatmap showing the relative gene expression of cytokines and chemokines analyzed in (G) within TLR-agonist stimulated or unstimulated mononuclear phagocyte populations.
+
+See also  Figure S6  and  Table S1 .
+
+From the perspective of adaptive immunity, activated DC3s shared a lot of common features with activated cDC2s but less with activated monocytes: (1) stimulated DC3s and cDC2s upregulated  CCR7  upon activation ( Figure 6 D), potentially enabling their ability to migrate toward T cell zones; (2) activated DC3s and cDC2s upregulated cell surface co-stimulatory molecules ( CD80 ,  CD86 ,  CD70 , and  CD40 ;  Figure 6 E and 6F); (3) activated DC3s and cDC2s efficiently increased the expression of T cell-attracting chemokines such as CCL5 ( Figures 6 G and 6H),  CCL19 ,  CCL17 ,  CCL22 ,  CXCL9 ,  CXCL10 ,  CXCL11 , and  CXCL13  ( Figure S6 C); and (4) activated DC3s and cDC2s produced higher amounts of IL-12p70, IL-23,  IL27 , and IL-10 ( Figures 6 G and 6H;  Figure S6 E).
+
+In addition, activated DC3s shared some common features with activated monocytes that are less pronounced in activated cDC2s: (i) activated DC3s and monocytes secreted more inflammatory cytokines, such as tumor necrosis factor alpha (TNF-α) and IL-1β ( Figures 6 G and 6H); and (2) activated DC3s and activated monocytes upregulated inflammatory chemokines, such as CCL2 ( Figures 6 G and 6H),  CCL1 , and  CCL3  ( Figure S6 C) or granulocytes attracting CXCL1,  CXCL3 , and  CXCL5  ( Figures S6 C and S6D).
+
+Collectively, our data suggest that the transcriptome and secretome of activated DC3s, unlike the ones of activated monocytes, are consistent with a function in priming of naive T cells.
+
+To directly test the T cell priming capabilities of DC3s, we performed 5-day cocultures of flow cytometry-sorted, activated DC3s, cDC2s, and monocytes (see the cell sorting strategy in  Figure S7 A) with allogenic CD45RA +  naive T cells in the presence of a synthetic superantigen. We found that activated DC3s and cDC2s, unlike monocytes, triggered proliferative expansion and effector differentiation in CD4 +  and CD8 +  T cells probed by CD45RO acquisition ( Figure S7 B). The same results were also obtained using  in-vitro- generated DC3s ( Figure S7 C). In contrast to monocytes, activated DC3s and cDC2s induced interferon γ (IFN-γ)- and TNF-α-producing CD4 +  and CD8 +  T cells but not IL- 17A ( Figure 7 A;  Figure S7 D). We found that DC3s had a specific ability to efficiently trigger CD103 expression in CD8 +  T cells ( Figure 7 B), even without the presence of superantigen ( Figure S7 E). CD103 expression is a hallmark of tissue-resident memory T (T RM ) cells because of its interaction with E-cadherin ( Mueller and Mackay, 2016 ). Mechanistically, multiple factors, including transforming growth factor β (TGF-β), have been proposed to induce CD103 expression on T cells ( Mueller and Mackay, 2016 ;  Rihs et al., 1996 ;  Yu et al., 2013 ). Here we showed that anti-TGF-β neutralizing antibodies, but not others, blocked the expression of CD103 on CD8 +  T cells ( Figure 7 C;  Figure S7 F). Figure 7 DC3s Prime Naive T Cells and Drive Acquisition of the CD103 +  T RM  Phenotype (A and B) Representative flow cytometry plots and quantification of CD4 +  and CD8 +  naive T cells cultured for 5 days with flow cytometry-sorted blood cDC2s, DC3s, or CD14 +  monocytes after overnight activation with TLR agonists (25 μg/mL poly(I:C), 1 μg/mL R848, and 10 ng/mL LPS) in the presence of a synthetic superantigen (Cytostim). Absolute numbers and frequencies of cytokine-producing and other activated T cells (A) and CD103 +  T cells (B) are shown (n = 5 healthy donors in 5 independent experiments; a line represents the median;  ∗ p < 0.05,  ∗∗ p < 0.01,  ∗∗∗ p < 0.001,  ∗∗∗∗ p < 0.0001, one-way ANOVA test). (C) Representative flow cytometry plots and quantification showing CD103 expression on CD8 +  naive T cells cocultured with blood DC3s sorted by fluorescence-activated cell sorting (FACS) in the presence of 10 μg/mL of neutralizing antibodies against TNF-α or TGF-β or an isotype CTRL (n = 4 healthy donors in 3 independent experiments; a line represents the median;  ∗ p < 0.05,  ∗∗ p < 0.01,  ∗∗∗ p < 0.001,  ∗∗∗∗ p < 0.0001, one-way paired ANOVA test). (D and E) Bulk RNA-seq analysis of CD8 + CD103 −  T cells (n = 3) and CD8 + CD103 +  T cells (n = 3) sorted by flow cytometry after 5 days  in vitro  coculture of naive blood CD8 +  T cells with blood DC3s activated overnight by TLR agonists. (D) GSEA of pairwise comparisons of CD8 + CD103 +  T cells with CD8 + CD103 −  T cells. Gene signatures (gene set) defining genes upregulated in breast or lung CD103 + CD69 + CD8 +  T RM  cells were used ( Hombrink et al., 2016 ;  Kumar et al., 2017 ;  Savas et al., 2018 ). (E) Heatmap displaying 56 representative genes significantly upregulated in CD8 + CD103 +  cells compared with CD8 + CD103 −  induced by blood DC3s  in vitro  (of 205 DEGs). Selected genes are shared with at least one of the previously reported gene signatures defining human T RM  cells ( Hombrink et al., 2016 ;  Kumar et al., 2017 ;  Savas et al., 2018 ). (F–H) Correlative analysis of T RM  cell content in luminal breast cancer primary tumors. (F) Representative flow cytometry plots showing the gating strategy for CD103 + CD69 + CD8 +  T cells in 21 human luminal breast cancer primary tumors. (G) Quantification of CD103 + CD69 + CD8 +  T cells in different stages of human breast tumors (stage I, n = 3; stage II, n = 13; stage III, n = 5). (H) Pearson correlation of the frequencies of the macrophages and cDC1, CD1c +  CD14 − , and CD1c +  CD14 −  cells and the frequencies of CD103 + CD69 + CD8 +  T cells in human breast cancer primary tumors (red, significantly correlated p < 0.05; black, not correlated). See also  Figure S7  and  Tables S1  and  S2 .
+
+DC3s Prime Naive T Cells and Drive Acquisition of the CD103 +  T RM  Phenotype
+
+(A and B) Representative flow cytometry plots and quantification of CD4 +  and CD8 +  naive T cells cultured for 5 days with flow cytometry-sorted blood cDC2s, DC3s, or CD14 +  monocytes after overnight activation with TLR agonists (25 μg/mL poly(I:C), 1 μg/mL R848, and 10 ng/mL LPS) in the presence of a synthetic superantigen (Cytostim). Absolute numbers and frequencies of cytokine-producing and other activated T cells (A) and CD103 +  T cells (B) are shown (n = 5 healthy donors in 5 independent experiments; a line represents the median;  ∗ p < 0.05,  ∗∗ p < 0.01,  ∗∗∗ p < 0.001,  ∗∗∗∗ p < 0.0001, one-way ANOVA test).
+
+(C) Representative flow cytometry plots and quantification showing CD103 expression on CD8 +  naive T cells cocultured with blood DC3s sorted by fluorescence-activated cell sorting (FACS) in the presence of 10 μg/mL of neutralizing antibodies against TNF-α or TGF-β or an isotype CTRL (n = 4 healthy donors in 3 independent experiments; a line represents the median;  ∗ p < 0.05,  ∗∗ p < 0.01,  ∗∗∗ p < 0.001,  ∗∗∗∗ p < 0.0001, one-way paired ANOVA test).
+
+(D and E) Bulk RNA-seq analysis of CD8 + CD103 −  T cells (n = 3) and CD8 + CD103 +  T cells (n = 3) sorted by flow cytometry after 5 days  in vitro  coculture of naive blood CD8 +  T cells with blood DC3s activated overnight by TLR agonists.
+
+(D) GSEA of pairwise comparisons of CD8 + CD103 +  T cells with CD8 + CD103 −  T cells. Gene signatures (gene set) defining genes upregulated in breast or lung CD103 + CD69 + CD8 +  T RM  cells were used ( Hombrink et al., 2016 ;  Kumar et al., 2017 ;  Savas et al., 2018 ).
+
+(E) Heatmap displaying 56 representative genes significantly upregulated in CD8 + CD103 +  cells compared with CD8 + CD103 −  induced by blood DC3s  in vitro  (of 205 DEGs). Selected genes are shared with at least one of the previously reported gene signatures defining human T RM  cells ( Hombrink et al., 2016 ;  Kumar et al., 2017 ;  Savas et al., 2018 ).
+
+(F–H) Correlative analysis of T RM  cell content in luminal breast cancer primary tumors.
+
+(F) Representative flow cytometry plots showing the gating strategy for CD103 + CD69 + CD8 +  T cells in 21 human luminal breast cancer primary tumors.
+
+(G) Quantification of CD103 + CD69 + CD8 +  T cells in different stages of human breast tumors (stage I, n = 3; stage II, n = 13; stage III, n = 5).
+
+(H) Pearson correlation of the frequencies of the macrophages and cDC1, CD1c +  CD14 − , and CD1c +  CD14 −  cells and the frequencies of CD103 + CD69 + CD8 +  T cells in human breast cancer primary tumors (red, significantly correlated p < 0.05; black, not correlated).
+
+See also  Figure S7  and  Tables S1  and  S2 .
+
+We next wondered whether CD103 +  T cells aligned with  bona fide  T RM  cells isolated from human tissue. To this end, we evaluated the gene expression profile of CD103 −  and CD103 +  CD8 +  T cells obtained after coculture of naive T cells with activated DC3s ( Figure S7 G). Using the GSEA methodology ( Subramanian et al., 2005 ), we found that the signatures obtained for breast cancer or lung tissue CD103 +  T RM  cells ( Hombrink et al., 2016 ;  Kumar et al., 2017 ;  Savas et al., 2018 ) were enriched in CD103 +  T cells compared with CD103 −  T cells induced by DC3s ( Figure 7 D). DEG analysis revealed that human T RM  cell markers such as  NUSAP1 ,  DUSP4 ,  CXCR6 , and  FASLG  ( Figure 7 E;  Hombrink et al., 2016 ;  Kumar et al., 2017 ;  Savas et al., 2018 ) were upregulated in CD103 +  compared with CD103 −  CD8 +  T cells. In addition, DC3-activated CD103 +  CD8 +  T cells expressed core components of the cytotoxic machinery ( PRF1  and  GZMA ), as reported earlier for breast cancer-invading T RM  cells ( Figure 7 E;  Savas et al., 2018 ).
+
+CD8 + CD103 +  T RM  cell infiltration has a protective prognosis value in breast cancer ( Savas et al., 2018 ;  Wang et al., 2016 ). To test the physiological relevance of DC3-dependent induction of CD8 + CD103 +  T cells, we analyzed CD103 and CD69 expression in CD3 + CD8 +  T cells from 18 samples of primary luminal breast cancer ( Figure 7 F). We found that CD8 + CD103 + CD69 +  T cells were present across different cancer stages ( Figure 7 G). The frequency of CD8 + CD103 + CD69 +  T cells was positively correlated with DC3 infiltration but not with other mononuclear phagocytes ( Figure 7 H). Despite the statistical significance of this correlation, it is important to underline that this was mainly driven by a subset of the samples. This suggests that further stratification of patients could help improve our understanding of the relationship between T RM  cells and DC3s in breast cancer.
+
+Together, our results anchor DC3 function within the DC lineage; DC3s, just like cDC2s but unlike monocytes, are competent for priming and polarization of CD45RA +  naive T cells. In addition, we define induction of  bona fide  T RM  cell-like CD103 + CD8 +  T cells ( Yu et al., 2013 ) as a specific privilege of DC3s but not cDC2s.
+
+Using scRNA-seq and high-dimensional flow cytometry, we provided evidence that DC3s represent a DC subset that can be separated and isolated from other DC subtypes. Regarding gene expression, our results are largely congruent with findings reported in previous studies ( Dutertre et al., 2019 ;  Villani et al., 2017 ). In addition, we provide a robust flow cytometry strategy to identify and purify DC3s by taking in account CD14 +  monocytes and cDC2s. In agreement with previous studies, our results challenge the widely accepted notion that CD14 is a specific marker for monocytes ( Dutertre et al., 2019 ;  Villani et al., 2017 ). Indeed, we clearly demonstrate that a large fraction of DC3s expresses cell surface CD14. We identified CD88 (encoded by the  C5AR1  gene) as a proper monocyte marker, enabling prospective purification of monocytes devoid of CD14 + CD1c +  DC3s. We found that  ex vivo  GM-CSF cultures of pure CD88 + CD14 +  monocytes did not give rise to CD88 - CD14 + CD1c +  iDCs. Therefore, our findings provide an incentive to carefully revisit the prevailing notion that CD14 + CD1c +  iDCs arise exclusively from monocytes  in vivo .
+
+From the DC perspective, we refined the strategy to analyze the functional heterogeneity of CD1c +  DCs, including cDC2s and DC3s. Indeed, previous studies have reported heterogeneous expression of CD5 in CD1c +  DCs ( Dutertre et al., 2019 ;  Yin et al., 2017 ). However, our findings highlight that CD5 − CD1c +  DCs contain CD163 −  cDC2s and CD163 +  DC3s, which are transcriptionally distinct. Overall, our scRNA-seq and bulk RNA-seq analyses cluster together CD5 +  and CD5 −  cDC2s. However, the developmental relationship between CD5 −  cDC2s, CD5 +  cDC2s, and AS-DCs remains to be clarified ( Dutertre et al., 2019 ;  See et al., 2017 ;  Villani et al., 2017 ).
+
+Because of CD1c expression and their relative similarity to cDC2s, DC3s have been embedded in the group of cDCs. Here we have provided substantial evidence arguing against this notion. Indeed, a conservative definition of cDCs includes (1) dependence on the FLT3L growth factor ( Breton et al., 2015 ;  Guermonprez et al., 2013 ,  2019 ;  McKenna et al., 2000 ;  Pulendran et al., 1998 ;  Waskow et al., 2008 ) and (2) reliance on the CDP and pre-cDC developmental pathway ( Breton et al., 2015 ,  2016 ;  Guermonprez et al., 2019 ;  Lee et al., 2015b ;  Naik et al., 2007 ;  Onai et al., 2007 ). Here we have shown that DC3s do not meet any of these criteria. First, FLT3L alone was poorly active in stimulating the production of DC3s from CD34 +  HSPCs in a controlled setting  in vitro . Conversely, GM-CSF drives the commitment of DC3s under the same conditions. Moreover, we have shown that CDP did not give rise to DC3s although they are competent to generate cDC1s and cDC2s ( Lee et al., 2015b ). Taking in account the heterogeneity of defined progenitor populations ( Paul et al., 2015 ), we developed single-cell cultures enabling analysis of multiple lineage populations. Our single-cell cultures demonstrated that DC3s develop from CLEC12A +  DC3-restricted progenitors. In addition, the most frequent progenies differentiating along with DC3s from a bi-potent progenitor were cDCs and macrophages. This strongly suggest that DC3 progenitors diverge downstream of the MDP stage ( Fogg et al., 2006 ;  Lee et al., 2015a ). Further studies will define more precisely the cell surface and molecular phenotype of the DC3-restricted progenitor. In support of a distinct regulation of cDC2s and DC3s, we have shown that cDC2 infiltration in breast cancer is correlated with cDC1s but not DC3s. Also,  Dutertre et al. (2019)  have shown recently that DC3s, but not cDC2s, expand in the blood of systemic lupus erythematosus patients ( Dutertre et al., 2019 ). Further delineation of the inflammatory cues and transcription factors underpinning the development of the DC3 lineage is needed.
+
+Identification of DC3 as a cellular entity arising from a specific lineage brings forward the question of their specific immune function. We have shown that activated DC3s, just as cDC2s but unlike monocytes, secrete high amounts of T cell-polarizing cytokines (IL-12p70 and IL-23) and T cell-attracting chemokines (CXCL9, CXCL10, CXCL11, and CCL5). In addition, DC3s also secrete other cytokines (IL-10, IL-6, and TNF-α), some of which were poorly secreted by cDC2s (e.g., TNF-α). DC3s are polyvalent phagocytes with a cytokine pattern encompassing T cell and inflammation cues.
+
+A quintessential defining feature of DCs is their capacity to activate naive T cells. For instance, infiltration of CD1c +  DCs is associated with priming of T cell effectors when regulatory T (Treg) cell infiltration is low or Treg cell-mediated suppression is alleviated by checkpoint blockade ( Binnewies et al., 2019 ). However, even when the heterogeneity of CD1c +  cells is appreciated, including recruitment of CD14 + CD1c +  DCs in tumor-draining lymph nodes ( Binnewies et al., 2019 ), little is known about the function of CD1c +  subtypes. Therefore, we asked whether the functions attributed to CD1c +  DCs would be carried out by cDC2s and/or DC3s. We found that DC3s, like cDC2s but unlike monocytes, primed and drove robust activation of naive T cells into IFN-γ- and TNF-α-secreting polyfunctional effectors ( Acosta-Rodriguez et al., 2007 ;  Leal Rojas et al., 2017 ;  Napolitani et al., 2005 ;  Nizzoli et al., 2013 ;  Schlitzer et al., 2013 ;  Yin et al., 2017 ).
+
+αE integrin (CD103) is a key marker of T RM  cells through its interaction with E-cadherin, participating in retention of T cells at epithelial and mucosal sites ( Mueller and Mackay, 2016 ).  Yu et al. (2013)  have identified a feature of total CD1c +  DCs in their ability to drive acquisition of CD103 in CD8 +  T cells. Here we have shown that DC3s, but not cDC2s, induced expression of CD103 on CD8 +  and CD4 +  T cells. In addition, we have shown that CD103 +  T cells primed by DC3s  ex vivo  align with  bona fide  T RM  cells isolated from lung or breast cancer ( Hombrink et al., 2016 ;  Kumar et al., 2017 ;  Savas et al., 2018 ). This finding is consistent with the existence of early imprinting of the T RM  cell program at the level of T cell priming by DC3s. This notion is supported by a recent  in vivo  study in a mouse model, evidencing early imprinting of the T RM  cell program during T cell priming in the lymph nodes by αV integrin-expressing migratory DCs ( Mani et al., 2019 ). Further supporting this view, we found that (1) DC3s upregulated CCR7 when activated by TLR agonists, potentially acting as migratory DCs, and (2) blocking studies indicated that TGF-β was required to prime CD103 +  T cells. This role of priming in lymph nodes does not exclude that tumor-infiltrating DC3s might also provide TGF-β and other signals important for maintenance of T RM  cells, as demonstrated in a mouse model ( Mani et al., 2019 ). In support of this view, we found that infiltration of DC3s was selectively associated with the abundance of CD8 + CD103 +  T cells in luminal breast cancer primary tumors. Further studies are needed to delineate the full molecular mechanisms of CD103 induction in T cells by DC3s and their physiological  in vivo  relevance during immune responses. Indeed, CD103 + CD8 +  T cells are a protective biomarker in triple-negative breast cancer ( Savas et al., 2018 ;  Wang et al., 2016 ), and lung CD103 + CD8 +  T cells are a hallmark of protective immunity afforded by influenza vaccination ( Yu et al., 2013 ). This underlines the potential of DC3s to regulate tissue immunity and defines them as targets for vaccines and immunotherapeutic interventions.
+
+We identify DC3s as CD88 - CD1c + CD163 + CD14 +/- . Further studies will be needed to determine if this phenotype corresponds to all CD1c + CD14 +  cells that had been reported in various inflammatory settings.
+
+We showed that DC3 differentiation is driven by GM-CSF in humanized mouse metastatic lung model. However, human circulating PBMCs were used in these experiments, preventing the assessment of GM-CSF impact on bone marrow DC3 progenitors  in vivo . Hence, we cannot exclude a role of GM-CSF on DC3s survival instead of differentiation. In addition, we showed that GM-CSF drives DC3 differentiation from CD34 +  umbilical cord HSPCs  in vitro . This finding does not exclude that other growth factors control DC3 development  in vivo .
+
+We aligned CD1c + CD14 +  cells infiltrating breast tumor-draining lymph node with blood DC3s compared to cDC2s and monocytes. Due to limited availability of healthy human lymph nodes, we did not investigate if CD1c + CD14 +  cells are infiltrating secondary lymphoid organs in homeostatic conditions. Blood DC3s upregulate CCR7 upon TLR stimulation suggesting they have a migratory potential. However, we did not investigate whether lymph nodes DC3s originate from non-lymphoid peripheral tissues via lymphatic vessels or directly from blood.
+
+A defining feature of cDCs lies in their ability to prime naïve T cells. We showed that DC3s can induce the proliferation of naïve T cell in a non-autologous priming context,  i.e.  the mixed leukocyte reaction. The capability of DC3s to uptake, process and present antigens by MHCI and MHCII molecules remains to be addressed.
+
+Our  in vitro  experiments showed that  i ) DC3s induce efficiently T RM  differentiation from naïve CD8 + T cells and,  ii ) DC3 infiltration correlates with T RM  abundance  in vivo . However, this does not provide a direct evidence that DC3s control T RM  specification within tumor draining lymph nodes where they are likely to interact with naïve T cells. Indeed, we cannot rule out that,  in vivo , DC3s act selectively in the tissue to maintain T RM  populations primed in lymph nodes by a different DC subtype. Further studies will be needed to address how DC3s control CD8 +  T RM  populations  in vivo .
+
+REAGENT or RESOURCE SOURCE IDENTIFIER Antibodies Biotin anti-human CD335 (NKp46) antibody (clone 9E2) BioLegend Cat#331906; RRID:  AB_1027671 Biotin anti-human CD3 antibody (clone OKT3) BioLegend Cat# 317320; RRID:  AB_10916519 Biotin anti-human CD19 antibody (clone HIB19) BioLegend Cat# 302204; RRID:  AB_314234 Biotin anti-human CD20 antibody (clone 2H7) BioLegend Cat# 302350; RRID:  AB_2565524 Biotin anti-human CD56 (NCAM) antibody (clone HCD56) BioLegend Cat# 318320; RRID:  AB_893390 Biotin anti-human CD203c (E-NPP3) antibody (clone NP4D6) BioLegend Cat# 324604; RRID:  AB_756042 Biotin anti-human CD66b antibody (clone G10F5) BioLegend Cat# 305120; RRID:  AB_2566608 Brilliant Violet 421 anti-human CD16 antibody (clone 3G8) BioLegend Cat# 302038; RRID:  AB_2561578 Brilliant Violet 510 anti-human CD45RA antibody (clone HI100) BioLegend Cat# 304142; RRID:  AB_2561947 Brilliant Violet 510 anti-human HLA-DR antibody (clone L243) BioLegend Cat# 307646; RRID:  AB_2561948 Brilliant Violet 510 anti-human CD15 (SSEA-1) antibody (clone W6D3) BioLegend Cat# 323028; RRID:  AB_2563400 FITC anti-human CD163 antibody (clone GHI/61) BioLegend Cat# 333618; RRID:  AB_2563094 Brilliant Violet 711 anti-human CD163 antibody (clone GHI/61) BioLegend Cat# 333630; RRID:  AB_2650972 Brilliant Violet 421 anti-human CD163 antibody (clone GHI/61) BioLegend Cat# 333612; RRID: AB_2562463 Brilliant Violet 785 anti-human CD14 antibody (clone M5E2) BioLegend Cat# 301840; RRID:  AB_2563425 PE/Cy7 anti-human CD14 antibody (clone HCD14) BioLegend Cat# 325618; RRID:  AB_830691 FITC anti-human CD14 antibody (clone M5E2) BioLegend Cat# 301804; RRID:  AB_314186 ) Brilliant Violet 785 anti-human CD3 antibody (clone UCHT1) BioLegend Cat# 300472; RRID:  AB_2687178 APC anti-human CD34 antibody (clone 561) BioLegend Cat# 343608; RRID:  AB_2228972 FITC anti-human CD34 antibody (clone 561) BioLegend Cat# 343604; RRID:  AB_1732005 PerCP/Cyanine5.5 anti-human CD303 (BDCA-2) antibody (clone 201A) BioLegend Cat# 354210; RRID:  AB_11219604 PerCP/Cyanine5.5 anti-human CD141 (Thrombomodulin) antibody (clone M80) BioLegend Cat# 344112; RRID:  AB_2561625 PerCP/Cyanine5.5 anti-human CD371 (CLEC12A) antibody (clone 50C1) BioLegend Cat# 353612; RRID:  AB_2565544 PE anti-human CD370 (CLEC9A/DNGR1) antibody (clone 8F9) BioLegend Cat# 353804; RRID:  AB_10965546 PE anti-human CD5 antibody (clone L17F12) BioLegend Cat# 364014; RRID:  AB_2565284 PE anti-human CD116 antibody (clone 4H1) BioLegend Cat# 305908; RRID:  AB_2085686 PE anti-human CD206 (MMR) antibody (clone 15-2) BioLegend Cat# 321106; RRID:  AB_571911 PE anti-human CD88 (C5aR) antibody (clone S5/1) BioLegend Cat# 344304; RRID:  AB_2067175 PE/Dazzle 594 anti-human CD123 antibody (clone 6H6) BioLegend Cat# 306034; RRID:  AB_2566450 PerCP/Cyanine5.5 anti-human CD123 antibody (clone 6H6) BioLegend Cat# 306016; RRID:  AB_2264693 FITC anti-human CD1c antibody (clone L161) BioLegend Cat# 331518; RRID:  AB_2073403 PE/Cy7 anti-human CD1c antibody (clone L161) BioLegend Cat# 331516; RRID:  AB_2275574 APC anti-human CD272 (BTLA) antibody (clone MIH26) BioLegend Cat# 344510; RRID:  AB_10613101 APC anti-human CD64 antibody (clone 10.1) BioLegend Cat# 305014; RRID:  AB_1595428 APC/Cyanine7 anti-human CD45 antibody (clone HI30) BioLegend Cat# 304014; RRID:  AB_314402 APC/Cy7 Streptavidin BioLegend Cat# 405208 Brilliant Violet 785 anti-human CD95 (Fas) antibody (clone DX2) BioLegend Cat# 305646; RRID:  AB_2629742 Brilliant Violet 711 anti-human CD8 antibody (clone SK1) BioLegend Cat# 344734; RRID:  AB_2565243 Brilliant Violet 605 anti-human CD45 antibody (clone HI30) BioLegend Cat# 304042; RRID:  AB_2562106 Brilliant Violet 650 anti-human CD45RO antibody (clone UCHL1) BioLegend Cat# 304232; RRID:  AB_2563462 Brilliant Violet 421 anti-human CD103 (Integrin E) antibody (clone Ber-ACT8) BioLegend Cat# 350214; RRID:  AB_2563514 Brilliant Violet 785 anti-human CD103 (Integrin E) antibody (clone Ber-ACT8) BioLegend Cat# 350230; RRID:  AB_2734364 FITC anti-human CD223 (LAG-3) antibody (clone 7H2C65) BioLegend Cat# 369210; RRID:  AB_2716129 FITC anti-human CD3 antibody (clone UCHT1) BioLegend Cat# 300406; RRID:  AB_314060 PE/Cy7 anti-human CD183 (CXCR3) antibody (clone G025H7) BioLegend Cat# 353720; RRID:  AB_11219383 PE/Dazzle 594 anti-human CD28 antibody (clone CD28.2) BioLegend Cat# 302942; RRID:  AB_2564235 PE anti-human CD197 (CCR7) antibody (clone G043H7) BioLegend Cat# 353204; RRID:  AB_10913813 APC/Cyanine7 anti-human CD4 antibody (clone SK3) BioLegend Cat# 344616; RRID:  AB_2028483 Alexa Fluor® 700 anti-human CD127 (IL-7Ra) antibody (clone A019D5) BioLegend Cat# 351344; RRID:  AB_2566200 APC anti-human CD279 (PD-1) antibody (clone EH12.2H7) BioLegend Cat# 329908; RRID:  AB_940475 PE anti-human CD115 antibody (clone 9-4D2-1D4) BioLegend Cat# 347310; RRID:  AB_2565491 FITC anti-human CD116 antibody (clone 4H1) BioLegend Cat# 305906; RRID:  AB_2085687 FITC anti-human CD86 antibody (clone BU63) BioLegend Cat# 374204; RRID:  AB_2721574 Human Axl PE antibody (clone 108724) R&D Cat# FAB154P; no RRID PerCP/Cyanine5.5 anti-mouse CD45 antibody (clone 30-F11) BioLegend Cat# 103132; RRID:  AB_893340 APC-eFluor 780 anti-human HLA-DR antibody (clone LN3) eBioscience Cat# 47-9956; RRID: AB_1963603 PE/Cy7 anti-human CD11c antibody (clone Bu15) Biolegend Cat# 337216; RRID: AB_2129790 FITC anti-human CD14 antibody (clone 61D3) eBioscience Cat# 11-0149; RRID: AB_464951 PerCP-eFluor 710 anti-human CD1c antibody (clone L161) eBioscience Cat# 46-0015; RRID: AB_10548936 PE anti-human CD304 antibody (clone AD5-17F6) Miltenyi Cat# 130-113-517; RRID: AB_2751124 PE-Cy5 Mouse anti-Human CD1a antibody (clone HI149) BD Cat# 555808; RRID: AB_396142 Alexa Fluor® 647 anti-human CD206 (MMR) antibody (clone 15-2) Biolegend Cat# 321116; RRID: AB_571881 VioBlue anti-human CD141 antibody (clone AD5-14H12) Miltenyi Cat# 130-113-882; RRID: AB_2726374 Purified anti-human CD45 antibody (clone HI30) BioLegend Cat# 304002; RRID:  AB_314390 Purified anti-human TGF 1-2-3 antibody (clone 1d11) R&D Cat# MAB1835; RRID: AB_357931 Purified anti-human TNFα antibody (clone 1825) R&D Cat# MAB210; RRID: AB_2240620 Purified anti-human IL-15 antibody (clone 34593) R&D Cat# MAB247; RRID: AB_2124578 Purified anti-human Lymphotoxin α antibody (clone 5802) R&D Cat# MAB211; RRID: AB_2138622 Purified polyclonal anti-human OX40L antibody R&D Cat# AF1236; RRID: AB_354686 Purified polyclonal anti-human IL-33 antibody R&D Cat# AF3625; RRID: AB_1151900 Purified anti-human TNFSF8 antibody (clone 116614) R&D Cat# MAB1028; RRID: AB_2303710 Purified Mouse IgG1 (clone 11711) R&D Cat# MAB002; RRID: AB_357344 Purified polyclonal Goat IgG R&D Cat# AB-108-C; RRID: AB_354267 Purified Mouse IgG2b (clone 20116) R&D Cat# MAB004; RRID: AB_357346 Bacterial and Virus Strains pMX-IRES-GFP Cell biolabs Cat# RTV-0133 pMX-huFLT3L-IRES-GFP This paper N/A pMX-huGMCSF-IRES-GFP This paper N/A Biological Samples Leukocyte cones for bulk RNaseq NHS N/A Leukocytes cones for functionnal experiments NHS N/A umbilical cord-blood for  in vitro  diffentiation cells Anthony Nolan Cell Therapy Centre N/A PBMC from whole blood for blood scRNaseq Institut Curie Hospital (Paris, France) N/A Chemicals, Peptides, and Recombinant Proteins Liberase TL Sigma Cat# 5401020001 Collagenase D Sigma Cat# 11088866001 DNase I Sigma Cat# 10104159001 Dispase Sigma Cat# D4693 CD34 MicroBead Kit UltraPure, human Miltenyi Cat# 130-100-453 Accucheck counting beads Thermo Fischer Cat# PCB100 Matrigel® BD Cat# 354230 Poly I:C InvivoGen Cat# 31852-29-6 R848 InvivoGen Cat# 144875-48-9 lipopolysaccharide (LPS) Sigma-Aldrich Cat# L2630 Cell Trace Violet (CTV) Thermo Fischer Cat#  C34557 Naive Pan T cells isolation Kit Miltenyi Cat# 130-097-095 Cytostim Miltenyi Cat# 130-092-172 Cytofix/ cytoperm Kit BD Bioscience Cat# 554714 Critical Commercial Assays Human Flt-3 Ligand/FLT3L Quantikine ELISA Kit R&D Cat# DFK00 ELISA MAX Deluxe Set Human GM-CSF Biolegend Cat# 432004 Human Proinflammatory Chemokine Panel (13-plex) Biolegend Cat# 740003 Human Macrophage/Microglia Panel (13-plex) Biolegend Cat# 740503 Deposited Data RNaseq data raw reads and processed data This paper GEO: SuperSeries  GSE151095 Experimental Models: Cell Lines B16_Control This paper N/A B16_human FLT3L This paper N/A B16_human GM-CSF This paper N/A MS5_FLT3L Anselmi et al., 2020 N/A MS5_GM-CSF Anselmi et al., 2020 N/A MS5_FS12 Anselmi et al., 2020 N/A Experimental Models: Organisms/Strains mouse: NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ The Jakson Laboratory Stock#005557 Software and Algorithms Genomics Suite Partek N/A FlowJo V10 BD https://www.flowjo.com GraphPad Prism 8 GraphPad Software https://www.graphpad.com R4.4 The R Foundation https://www.r-project.org Fiji open source https://imagej.net/Fiji
+
+Additional information and request for resources and reagents should be directed to and will be made available by the Lead Contact, Pierre Guermonprez ( pierre.guermonprez@kcl.ac.uk ).
+
+The reagents generated in this study will be made available on request, but we may require a payment and/or a completed Materials Transfer Agreement if there is potential for commercial application.
+
+The scRNA-seq and the bulk RNA-seq datasets are deposited in the Genome Expression Omnibus under the SuperSeries accession numbers  GSE151095 .
+
+Human umbilical cord blood units were obtained from Anthony Nolan Cell Therapy Centre (ANCTC). Leukophoretic adult blood (buffy coats or leukocyte cones) were obtained from healthy volunteers through NHS.
+
+Tumor-invaded lymph nodes (tdLN) and primary tumors were collected from luminal breast cancer submitted to surgical resection at the Institut Curie Hospital (Paris, France), in accordance with institutional ethical guidelines. Patients’ clinical and pathologic characteristics are summarized in  Table S2 .
+
+NSG (NOD.Cg- Prkdc scid 
+ Il2rg tm1Wjl /SzJ) mice were bred and maintained in specific pathogen-free animal facility in accordance with institutional KCL guidelines. All procedures involving animals were conducted according to requirements of UK Animals (Scientific Procedures) Act 1986.
+
+Peripheral blood mononuclear cells (PMBCs) were obtained by gradient centrifugation using Ficoll-Paque (GE Healthcare). Progenitor cells were enriched using CD34 +  microbead isolation kit (Miltenyi).
+
+B16_CTRL, B16_huGM-CSF and B16_huFLT3L were generated by retroviral transduction of B16-F10 (C57BL/6 melanoma cell line) with an empty pMX-IRES-GFP vector or coding for human GM-CSF and human Flt3L respectively. Tumor cell lines were cultured in RPMI 1640 medium (Thermo Fisher) supplemented with 10% fetal bovine serum (FBS) (Thermo Fisher), penicillin-streptomycin (Thermo Fisher) and b-mercaptoethanol (Thermo Fisher) (complete RPMI) and maintained at 37°C and 5% CO 2 .
+
+Engineered B16-F10 cells were counted and resuspended in RPMI 1640 medium. NSG mice (8-12 weeks, males and females) were injected intravenously with 10 6  B16-F10 at day 0. 10 8  human blood PBMCs were injected intravenously at day 7 and 8. Mice were culled at day 9 and lungs were harvested.
+
+Mouse lungs were fixed with 1% PFA (Alfa Aesar) for 1hr at 4°C, washed and incubated in 34% sucrose solution (Sigma-Aldrich) overnight at 4°C. Lungs were embedded in Cryomatrix (Thermo Fischer) and frozen for cryostat sectioning (9 μm-thick). Sections were permeabilized using 0.5% saponin (Sigma-Aldrich), 2% BSA (Sigma-Aldrich), 1% FBS (Life Technologies) for 30 minutes at room temperature. Sections were labeled overnight at 4°C with mouse anti-human purified CD45 (HI30, Biolegend) followed by incubation for 1hr at room temperature with goat anti-mouse Cy3 (Jackson laboratory). All sections were labeled with Hoechst (Molecular Probes, Thermo Fisher) for nuclei staining 5 minutes at room temperature and mounted with Prolong diamond (Thermo scientific). Slides were imaged using a SP5 (Leica) and analyzed with Fiji software.
+
+Mouse lungs were harvested and transferred to 3ml digestion buffer (Hank’s Balanced Salt Solution (HBSS) with calcium and magnesium (Thermo Fisher) and with 0.1 mg/ml of Liberase TL (Roche) and 0.02μg/ml DNase I (Thermo Fisher). Lungs were dissociated using gentleMACS Octo Dissociator (Miltenyi) and incubated at 37°C for 45 minutes. The cell suspension was passed through a cell strainer (70μm, Corning) and red blood cells were lysed using ACK lysing buffer (Thermo Fischer). The absolute number of cells in the resulting cell suspension was assessed using AccuCheck Counting Beads (Thermo Fisher) on BD FACSCanto II (BD Biosciences).
+
+Mouse bone marrow-derived MS5 stromal cells engineered to express human membrane bound FLT3L alone (MS5_FL) or in combination with SCF and soluble CXCL12 (MS5_FS12) or human GM-CSF alone (MS5_GM) generated as previously described ( Anselmi et al., 2020 ). Stomal cell lines were cultured in IMDM medium supplemented with 10% heat-inactivated FBS, penicillin/streptomycin, 50 μM β-mercaptoethanol (complete IMDM) and maintained at 37°C and 5% CO 2 .
+
+For  in vitro  co-culture experiments with stromal cells, MS5_FL, MS5_FS12 and MS5_GM feeders were seeded at 10 4  cells/well density in a 96-well plate (Thermo Fischer) and maintained overnight at 37°C and 5% CO 2 .
+
+Total enriched-CD34 +  progenitors and flow cytrometry-sorted progenitor subsets (GMDP, CDP and cMoP) were plated on top of stromal cells, supplemented or not with 100ng/ml of recombinant human GM-CSF (Peprotech), for 14 days and 7 days, respectively. Blood phagocytes subsets were flow cytometry-sorted and plated on MS5_GM for 2 days.
+
+For bulk RNA sequencing analysis, total enriched-CD34 +  progenitors were plated on top of MS5_FS12 feeders and with or without 100ng/ml of recombinant human GM-CSF (Peprotech). On day 5 and 10 of differentiation, half the medium was replaced with fresh complete IMDM or complete IMDM containing 100ng/ml of recombinant human GM-CSF. Cells were collected at day 15.
+
+All cells were collected with a solution of phosphate buffered saline (PBS) (GIBCO) 5mM EDTA (Thermo Fisher) at 4°C for 10min.
+
+PBMCs from healthy donors were stained and sorted as described previously. Blood cDC2 (10 5 ) and monocytes (3 × 10 5 ) were flow cytometry-sorted and injected subcutaneously along with B16_huGM (10 5 ) in 200 μL of ice-cold Matrigel ®  (BD Biosciences). Mice were sacrificed at day 4 by cervical dislocation and Matrigel ®  plugs were collected. Subcutaneous Matrigel ®  plugs were recovered, cut in pieces and incubated in HBSS (Life Technologies) 1% FBS, 0.37 U/ml Collagenase D (Roche), 10 μg/ml DNaseI (Roche) and 1 mg/ml Dispase (Sigma-Aldrich) for 30 minutes at 37°C. After digestion, plugs were smashed on a 70 μm strainer (Corning) and cells were collected and resuspended in flow cytometry buffer for flow cytometry analysis.
+
+Cells were stained in flow cytometry buffer (PBS 5mM EDTA 1% BSA) according to antibody panels ( Key Resources Table ) for 30 min. Dead cells were stained using DAPI or Live/Dead Blue staining (Thermo Fischer). Lineage (Lin) included CD3, CD19, CD20, NKp46, CD56, CD203c and CD66b, all conjugated with biotin. Multiparameter analysis was performed on LSRFortessa and Symphony (BD Biosciences) flow cytometers and analyzed using FlowJo software (Tree Star). The flow cytometry-sort was performed on BD FACS Aria II or BD FACS Aria Fusion at the Biomedical Research Centre (BRC) Flow Core Facility (Guy’s and St Thomas’ NHS Foundation Trust and King’s College London).
+
+Unsupervised clustering of flow cytometry data ( Figure 1 F) was performed using Self-Organizing Map clustering algorithm FlowSOM (Cytofkit R package). Data was then extracted as FCS files and further analyzed in FlowJo software.
+
+For bulk sequencing of  in vitro  differentiated subsets, up to 100 flow cytometry-sorted cells from three individual donors were collected directly in Lysis buffer (Takara Clontech, Cat# 635015) containing RNase inhibitors. RNaseq libraries were prepared on the contactless liquid handling system Labcyte Echo 525 (Labcyte Inc). In brief, ERCC was added to each sample and first strand full length cDNA was generated with a modified protocol of the SMARTseq v4 Ultra Low Input RNA Kit (Takara Clontech, Cat# 634891) using poly dT primers and a template switching oligo. Full length cDNA was amplified using SeqAmp DNA Polymerase (Takara Clontech, Cat# 638509). 12 ng of amplified cDNA from each sample was used to generate non-stranded RNA libraries using a modified protocol of the Ovation Ultralow System V2 1-96 kit (NuGEN, Cat# 0347-A01). In brief, amplified cDNA was fragmented through sonication on Covaris E220 (Covaris Inc), repaired and polished followed by ligation of indexed adapters. Adaptor ligated cDNA were pooled before final amplification to add flow cell primers. Libraries were sequenced on HiSeq2500 (Illumina Cambridge) for 100 cycles PE in Rapid mode.
+
+The raw sequencing data was initially processed using open source, web-based platform Galaxy (version 18.05.rc1) ( https://usegalaxy.org ). Reads were filtered for quality with more than 80% of the sequence having quality score > 33 using FastQC tool. Mapping against reference genome was performed with Hisat2 to the hg38 human genome. Adaptor sequences were detected automatically with TrimGalore!. Reads under 20bp were discarded. All processed sequencing files were imported in Partek® Genomics Suite software®, version 7.0©; 2017 (PGS), where they were processed further.
+
+mRNA was quantified using PGS built in RNA-seq workflow. Normalization method used was Reads Per Kilobase per Million mapped reads (RPKM) and mRNA was quantified against RefSeq Transcripts 2018-11-20 database. Hierarchal clustering on average expression within the group was performed on all identified protein coding genes (19791 genes). Based on the CD14 - CD1c + CD206 +/−  two blue cluster similarity, they were both considered as cDC2-like cells and their datasets were merged for further analysis. The same was done for the two macrophage-like cells (gray and brown). Differentially expressed genes (Fold-Change ≤ −2 or ≥ 2 and p value < 0.05) were determined using one-way ANOVA in all pairwise comparisons with three donors grouped and visualized as Volcano plots in PGS. Individual samples were visualized via principal component analysis (PCA) using 500 most variable genes, which were determined based on median absolute deviation (MAD). The expression patterns of selected gene lists were displayed in the form of heatmaps, where rows and/or columns were ordered based on hierarchical clustering using Euclidean distance and average linkage in Morpheus (Broad Institute;  https://software.broadinstitute.org/morpheus/ ). A few gene names are depicted next to the heatmap.
+
+PBMCs from three healthy donors or human CD45 +  cells isolated from metastatic mouse lungs were isolated as previously described. Sorted cells from each donor were pooled together and CD1c +  and CD14 +  cells were mixed at 80:20 ratio. 3 × 10 3  cells from the resulting cell suspension were partitioned into an emulsion of nanoliter-sized droplets using a 10X Genomics Chromium Single Cell Controller and RNA sequencing libraries were constructed using the Chromium Single Cell 3′ Library & Gel Bead Kit v2 (10X Genomics, Cat# PN-120237). Briefly, droplets containing individual cells, reverse transcription reagents and a gel bead were loaded with poly(dT) primers that include a 16 base cell barcode and a 10 base unique molecular index (UMI). Reverse transcription reactions were engaged to generate barcoded full-length cDNA followed by the disruption of emulsions using the recovery agent and cDNA clean up with DynaBeads MyOne Silane Beads (Thermo Fisher Scientific, Cat# 37002D). Bulk cDNA was amplified, and indexed sequencing libraries were constructed using the reagents from the Chromium Single Cell 3′ v2 Reagent Kit. Libraries were sequenced on NovaSeq 6000 Sequencing System (Illumina Cambridge).
+
+Cell Ranger (version 2.1.1) (from 10x genomics) was used to process Chromium single cell 3′ v2 RNA-seq output files. First, we generated fastq files for the Read1 for cell barcode and UMI and Read2 for transcript applying cellranger mkfastq (with default parameters). Second, we aligned the Read2 to the human reference genome GRCh38 using STAR (version 2.5.1) with cell ranger count (with default parameters) ( https://support.10xgenomics.com/single-cell-gene-expression/software/pipelines/latest/using/mkfastq ). Further analysis was performed using Seurat package (version 2.3.4) in R (version 3.4.0) ( Butler et al., 2018 ). Before performed analysis, we applied the following filtering step: only genes expressed in 3 or more cells have been preserved and cells with less 200 unique genes and more than 4000 unique expressed genes were discarded (as they are potentially cells doublets). After filtering step, we used an expression matrix resulting in 14933 genes across 1622 cells (among 1625 cells) for the rest of the analysis. The matrix was normalized using genes expression values for each cell were divided by the total number of transcripts and multiplied by median of UMI counts. Then, these values were natural log-transformed before downstream analysis. For dimensionality reduction analysis, we first identified 3288 genes as highly variable genes across the single cells (cutoff value for dispersion = 0.5; cutoff value for average expression = 0). PCA performed using the variable genes as input and determined 10 PCs as significant PCs. These principal components were used as input for t-Distributed Stochastic Neighbor Embedding (tSNE) ( van der Maaten, 2008 ). We used the shared nearest neighbor (SNN) modularity optimization-based clustering algorithm from the Seurat package (FindClusters function with default parameters) to identify the clusters of cells, following by Clustree analysis (clustree R package, version 0.2.2) by changing the resolution parameters from 0 to 2. Finally, we kept a resolution parameter at 0.8 and defined distinct 9 clusters. After controlling expression of some quality control genes, we excluded clusters E, F, G, H and I as contamination. Finally, we identified 4 relevant clusters. We identified cell specific marker by comparing cells in a specific cluster with cells in all other clusters using FindAllMarkers from Seurat package (MAST; logFC threshold = 0.5; only positive markers). Heatmap, feature plots and violin plots were performed using Seurat package.
+
+Single-cell RNA sequencing matrix for clusters A, B, C and D was created with gene signatures from  Villani et al. (2017) . Signatures were defined as mean expression of discriminative markers for cDC2s and DC3s among lin - CD14 -  cells (cDC2 enriched and DC3 enriched, respectively) and of discriminative markers for CD14 +  monocytes within lin - CD14 +/−  and/or CD16 +/−  monocytes (CD14 +  mono enriched). The average expression of signature genes for each cell was calculated and plotted as a violin plot using R package ggplot2 (version 3.1.0).
+
+To statistically evaluate the enrichment of previously reported gene signatures (Gene Sets) ( Hombrink et al., 2016 ;  Kumar et al., 2017 ;  Savas et al., 2018 ;  Villani et al., 2017 ) in our dataset, we used pairwise comparisons using the gene set enrichment analysis (GSEA) ( Subramanian et al., 2005 ) method from the Massachussets Institute of Technology ( https://www.broadinstitute.org/gsea ). GSEA tests the relative position of a collection of genes (Gene Set) within an independent, ranked dataset (GeneList). Statistical analysis was performed by evaluation of nominal p value and false discovery rate (q value) based on 1,000 random permutations. Results were considered significant when the p value was below 0.05 and when the q value was below 0.25 (false discovery rate below 25%) accordingly to the recommendation from the software developers. For each pairwise comparison, the GSEA output can be represented as a bar code where each bar corresponds to the projection of one Probe Sets of the Gene Sets on the list of all the Probe Sets of the gene chips ranked from those having a high signal in one cell population (represented in red) to those having a high signal in the other cell population (represented in blue). Each bar code can be characterized by two parameters. The normalized enrichment score (NES) represents the number and differential expression intensity of the genes enriched in the corresponding cell subset. The NES is positive if the Gene Set is enriched in the first cell population and negative if it is enriched in the second cell population. The false discovery rate (FDR) statistical value (q) represents the likelihood that the enrichment of the Gene Set represents a false positive finding (for example if q = 0.25, 25% of the Gene Sets found enriched can be false positives). An absolute value of the NES below or around 1 means no enrichment as confirmed with associated q-values above 0.25.
+
+To simultaneously visualize pairwise comparisons of transcriptomes from cord blood-derived DC2s (light and dark blue) DC3s (orange) and macrophages (gray and brown), the BubbleMap module of BubbleGum ( Spinelli et al., 2015 ) was used. Results were considered significant when the p value was below 0.05 and the FDR (false discovery rate, q) value was below 0.25. The BubbleGum was performed using previously published gene signatures of pairwise comparison between DC2s and DC3s and DC3s and CD14 +  monocytes ( Villani et al., 2017 ).
+
+Tumor-draining lymph nodes (tdLN) and primary tumors were collected in CO 2  independent medium (GIBCO; Cat# 18045-054) within few hours after the primary surgery. Tissue were cut into small fragments and submitted to enzymatic digestion using 0.1 mg/ml of Liberase TL (Roche) and 0.1 mg/ml of DNase (Roche) for 30 min. Cells were filtered on 40-μm cell strainer (BD), washed using CO 2  independent medium (GIBCO; Cat# 18045-054) containing 0.4g/ml of human albumin and resuspended for cell counting.
+
+After tissue processing, cells obtained from tdLN were submitted to myeloid cells enrichment accordingly to ( Durand and Segura, 2016 ) prior flow cytometry-sorting. In brief, T and B lymphocytes, NK cells, erythrocytes and myelomonocytic cells were depleted using monoclonal antibodies against: CD3, CD19, CD56, CD235a and CD15, respectively. Subsequently, cell suspensions were stained for 30 min with antibody-conjugated as the following: HLA-DR, CD11c, CD14, CD1c, CD304, CD1a, CD206, CD141. Around 1,000 cells of each DC subset were sorted by flow cytometry using BD FACS ARIA II cell sorter, (purity > 98%). Cells were centrifuged and lysed with TCL buffer (QIAGEN) containing 1% of beta-mercaptoethanol before storage at −80°C. RNA were extracted and isolated using the Single Cell RNA purification kit (Norgen, Cat#51800) according to the manufacturer’s instructions and the RNA integrity number was evaluated with an Agilent RNA 6000 pico kit.
+
+PBMCs and  in vitro  generated cells were stained and sorted as described previously. In total 3 × 10 5  of blood cell subsets (cDC2, DC3 and Mono) or 10 5 
+ in vitro  generated cells (cDC2-, DC3 and Macro-like) were flow cytometry-sorted and cultured with TLR agonists cocktail containing 25 μg/ml Poly I:C (InvivoGen, Cat# 31852-29-6), 1 μg/ml R848 (InvivoGen, Cat# 144875-48-9) and 10ng/ml LPS (Sigma-Aldrich, Cat# L2630) for 16 h. Culturing supernatants were collected and stored at −20°C until the LEGENDplex™ assay execution day. LEGENDplex Human Macrophage/Microglia Panel (13-plex) with V-bottom Plate (Biolegend Cat# 740503) and Human Proinflammatory Chemokine Panel (13-plex) with V-bottom Plate (Biolegend Cat# 740003) was used according to manufacturer’s instructions. In short, samples and standards were thawed and plated with capture beads and incubated for 2 h. Plate was then washed, and Detection Antibodies were added. After 1 h incubation SA-PE was added and incubated for 30 min. Samples were acquired on BD FACSCanto II. Samples were analyzed using LEGENDplex™ Data Analysis Software.
+
+PBMCs and  in vitro  generated cells were stained and flow cytometry-sorted as described previously. 10 4  mononuclear phagocytes were cultured with TLR agonists cocktail containing 25 μg/ml PolyI:C, 1 μg/ml R848 and 10 ng/ml LPS for 16 h. T cells were isolated from fresh or frozen PBMCs using Naive Pan T Cell Isolation Kit (Miltenyi, Cat# 130-097-095) and 10 5  cells were plated on top of mononuclear phagocytes in presence of CytoStim (Miltenyi, Cat# 130-092-172), according to manufacturer’s instructions. For mixed leukocyte reaction (MLR) experiments, isolated Naive Pan T cells were labeled with Cell Tracer Violet (CTV) (Thermo Fischer, Cat#  C34557 ) as per manufacturer’s instructions and cultured with flow cytometry-sorted DCs without CytoStim. At day 5, cells were collected and stained for extracellular and intracellular marker expression and analyzed using BD LSRFortessa. For intracellular staining, fixation and permeabilization were performed using BD Cytofix/Cytoperm solution (BD Biosciences, Cat# 554714) according to manufacturer’s instructions.
+
+Statistical analysis was performed using Prism 8.3 (GraphPad Software Inc., USA). When two experimental groups were compared, non-parametric Mann-Whitney test was used. When three or more groups were compared, statistically significant differences between means were determined using the one-way or two-way analysis of variance (ANOVA) method. A p value of less than 0.05 was considered as significant.
\ No newline at end of file
diff --git a/paper-releases/pbisc-diffusion/updates/20261015T141849Z-rare-cell-7k-8k-panels/diffusion/dataset/server_1/model8/plans/rare_gene_context_expansion_20261015/research/ilc_dc/sources/PMC7445430.xml b/paper-releases/pbisc-diffusion/updates/20261015T141849Z-rare-cell-7k-8k-panels/diffusion/dataset/server_1/model8/plans/rare_gene_context_expansion_20261015/research/ilc_dc/sources/PMC7445430.xml
new file mode 100644
index 0000000000000000000000000000000000000000..67e89f6d2b09aa76bce5d2300c50f691206cc0e7
--- /dev/null
+++ b/paper-releases/pbisc-diffusion/updates/20261015T141849Z-rare-cell-7k-8k-panels/diffusion/dataset/server_1/model8/plans/rare_gene_context_expansion_20261015/research/ilc_dc/sources/PMC7445430.xml
@@ -0,0 +1,8 @@
+
+<!DOCTYPE article
+  PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD with MathML3 v1.4 20241031//EN" "JATS-archivearticle1-4-mathml3.dtd">
+<article article-type="research-article" xml:lang="en" dtd-version="1.4"><?da-xref-anchor-style autodetect?><front><journal-meta><journal-id journal-id-type="nlm-ta">Immunity</journal-id><journal-id journal-id-type="iso-abbrev">Immunity</journal-id><journal-id journal-id-type="pmc-domain-id">445</journal-id><journal-id journal-id-type="pmc-domain">elsevierwt</journal-id><journal-id journal-id-type="nlm-id">9432918</journal-id><journal-title-group><journal-title>Immunity</journal-title></journal-title-group><issn pub-type="ppub">1074-7613</issn><issn pub-type="epub">1097-4180</issn><?publisher_abbrev elsevier?><custom-meta-group><custom-meta><meta-name>pmc-is-collection-domain</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-collection-title</meta-name><meta-value>Elsevier Sponsored Documents</meta-value></custom-meta></custom-meta-group></journal-meta><article-meta><article-id pub-id-type="pmcid">PMC7445430</article-id><article-id pub-id-type="pmcid-ver">PMC7445430.1</article-id><article-id pub-id-type="pmcaid">7445430</article-id><article-id pub-id-type="pmcaiid">7445430</article-id><article-id pub-id-type="pmid">32610077</article-id><article-id pub-id-type="doi">10.1016/j.immuni.2020.06.002</article-id><article-id pub-id-type="publisher-id">S1074-7613(20)30232-6</article-id><article-version article-version-type="pmc-version">1</article-version><article-categories><subj-group subj-group-type="heading"><subject>Article</subject></subj-group></article-categories><title-group><article-title>Transcriptional and Functional Analysis of CD1c<sup>+</sup> Human Dendritic Cells Identifies a CD163<sup>+</sup> Subset Priming CD8<sup>+</sup>CD103<sup>+</sup> T Cells</article-title></title-group><contrib-group><contrib contrib-type="author" id="au1"><name name-style="western"><surname>Bourdely</surname><given-names initials="P">Pierre</given-names></name><xref rid="aff1" ref-type="aff">1</xref><xref rid="aff2" ref-type="aff">2</xref><xref rid="fn1" ref-type="fn">9</xref></contrib><contrib contrib-type="author" id="au3"><name name-style="western"><surname>Anselmi</surname><given-names initials="G">Giorgio</given-names></name><xref rid="aff1" ref-type="aff">1</xref><xref rid="aff2" ref-type="aff">2</xref><xref rid="fn1" ref-type="fn">9</xref></contrib><contrib contrib-type="author" id="au2"><name name-style="western"><surname>Vaivode</surname><given-names initials="K">Kristine</given-names></name><xref rid="aff1" ref-type="aff">1</xref><xref rid="aff2" ref-type="aff">2</xref><xref rid="fn1" ref-type="fn">9</xref></contrib><contrib contrib-type="author" id="au4"><name name-style="western"><surname>Ramos</surname><given-names initials="RN">Rodrigo Nalio</given-names></name><xref rid="aff3" ref-type="aff">3</xref></contrib><contrib contrib-type="author" id="au5"><name name-style="western"><surname>Missolo-Koussou</surname><given-names initials="Y">Yoann</given-names></name><xref rid="aff3" ref-type="aff">3</xref></contrib><contrib contrib-type="author" id="au6"><name name-style="western"><surname>Hidalgo</surname><given-names initials="S">Sofia</given-names></name><xref rid="aff3" ref-type="aff">3</xref><xref rid="aff4" ref-type="aff">4</xref></contrib><contrib contrib-type="author" id="au7"><name name-style="western"><surname>Tosselo</surname><given-names initials="J">Jimena</given-names></name><xref rid="aff3" ref-type="aff">3</xref></contrib><contrib contrib-type="author" id="au8"><name name-style="western"><surname>Nuñez</surname><given-names initials="N">Nicolas</given-names></name><xref rid="aff3" ref-type="aff">3</xref></contrib><contrib contrib-type="author" id="au9"><name name-style="western"><surname>Richer</surname><given-names initials="W">Wilfrid</given-names></name><xref rid="aff3" ref-type="aff">3</xref></contrib><contrib contrib-type="author" id="au10"><name name-style="western"><surname>Vincent-Salomon</surname><given-names initials="A">Anne</given-names></name><xref rid="aff5" ref-type="aff">5</xref></contrib><contrib contrib-type="author" id="au11"><name name-style="western"><surname>Saxena</surname><given-names initials="A">Alka</given-names></name><xref rid="aff6" ref-type="aff">6</xref></contrib><contrib contrib-type="author" id="au12"><name name-style="western"><surname>Wood</surname><given-names initials="K">Kristie</given-names></name><xref rid="aff6" ref-type="aff">6</xref></contrib><contrib contrib-type="author" id="au13"><name name-style="western"><surname>Lladser</surname><given-names initials="A">Alvaro</given-names></name><xref rid="aff4" ref-type="aff">4</xref><xref rid="aff7" ref-type="aff">7</xref></contrib><contrib contrib-type="author" id="au14"><name name-style="western"><surname>Piaggio</surname><given-names initials="E">Eliane</given-names></name><xref rid="aff3" ref-type="aff">3</xref></contrib><contrib contrib-type="author" id="au15"><name name-style="western"><surname>Helft</surname><given-names initials="J">Julie</given-names></name><xref rid="aff3" ref-type="aff">3</xref><xref rid="fn2" ref-type="fn">10</xref></contrib><contrib contrib-type="author" id="au16"><name name-style="western"><surname>Guermonprez</surname><given-names initials="P">Pierre</given-names></name><email>pierre.guermonprez@kcl.ac.uk</email><xref rid="aff1" ref-type="aff">1</xref><xref rid="aff2" ref-type="aff">2</xref><xref rid="aff8" ref-type="aff">8</xref><xref rid="fn2" ref-type="fn">10</xref><xref rid="fn3" ref-type="fn">11</xref><xref rid="cor1" ref-type="corresp">∗</xref></contrib><aff id="aff1"><label>1</label>Centre for Inflammation Biology and Cancer Immunology, The Peter Gorer Department of Immunobiology, School of Immunology &amp; Microbial Sciences, King’s College London, London, UK</aff><aff id="aff2"><label>2</label>Cancer Research UK King’s Health Partner Cancer Centre, King’s College London, London, UK</aff><aff id="aff3"><label>3</label>PSL Research University, Institut Curie Research Center, Translational Immunotherapy Team, INSERM U932, Paris, France</aff><aff id="aff4"><label>4</label>Laboratory of Immuno-oncology, Fundación Ciencia &amp; Vida, Santiago, Chile</aff><aff id="aff5"><label>5</label>PSL Research University, Institut Curie, Department of Biopathology, Paris, France</aff><aff id="aff6"><label>6</label>National Institute of Health Research Biomedical Research Centre at Guy’s and St Thomas’ Hospital and King’s College London, London, UK</aff><aff id="aff7"><label>7</label>Facultad de Medicina y Ciencia, Universidad San Sebastián, Santiago, Chile</aff><aff id="aff8"><label>8</label>Université de Paris, Centre for Inflammation Research, CNRS ERL8252, INSERM1149 Paris, France</aff></contrib-group><author-notes><corresp id="cor1"><label>∗</label>Corresponding author <email>pierre.guermonprez@kcl.ac.uk</email></corresp><fn id="fn1"><label>9</label><p id="ntpara0010">These authors contributed equally</p></fn><fn id="fn2"><label>10</label><p id="ntpara0015">These authors contributed equally</p></fn><fn id="fn3"><label>11</label><p id="ntpara0020">Lead Contact</p></fn></author-notes><pub-date pub-type="ppub"><day>18</day><month>8</month><year>2020</year></pub-date><volume>53</volume><issue>2</issue><issue-id pub-id-type="pmc-issue-id">363872</issue-id><fpage>335</fpage><lpage>352.e8</lpage><history><date date-type="received"><day>18</day><month>5</month><year>2019</year></date><date date-type="rev-recd"><day>15</day><month>4</month><year>2020</year></date><date date-type="accepted"><day>29</day><month>5</month><year>2020</year></date></history><pub-history><event event-type="pmc-release"><date><day>18</day><month>08</month><year>2020</year></date></event><event event-type="pmc-live"><date><day>28</day><month>08</month><year>2020</year></date></event><event event-type="pmc-last-change"><date iso-8601-date="2026-03-28 22:25:13.527"><day>28</day><month>03</month><year>2026</year></date></event></pub-history><permissions><copyright-statement>© 2020 The Author(s)</copyright-statement><copyright-year>2020</copyright-year><license xmlns:xlink="http://www.w3.org/1999/xlink" license-type="CC BY" xlink:href="http://creativecommons.org/licenses/by/4.0/"><ali:license_ref xmlns:ali="http://www.niso.org/schemas/ali/1.0/" specific-use="textmining" content-type="ccbylicense">https://creativecommons.org/licenses/by/4.0/</ali:license_ref><license-p>This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).</license-p></license></permissions><abstract id="abs0010"><title>Summary</title><p>Dendritic cells (DCs) are antigen-presenting cells controlling T cell activation. In humans, the diversity, ontogeny, and functional capabilities of DC subsets are not fully understood. Here, we identified circulating CD88<sup>−</sup>CD1c<sup>+</sup>CD163<sup>+</sup> DCs (called DC3s) as immediate precursors of inflammatory CD88<sup>−</sup>CD14<sup>+</sup>CD1c<sup>+</sup>CD163<sup>+</sup>FcεRI<sup>+</sup> DCs. DC3s develop via a specific pathway activated by GM-CSF, independent of cDC-restricted (CDP) and monocyte-restricted (cMoP) progenitors. Like classical DCs but unlike monocytes, DC3s drove activation of naive T cells. <italic toggle="yes">In vitro</italic>, DC3s displayed a distinctive ability to prime CD8<sup>+</sup> T cells expressing a tissue homing signature and the epithelial homing alpha-E integrin (CD103) through transforming growth factor β (TGF-β) signaling. <italic toggle="yes">In vivo</italic>, DC3s infiltrated luminal breast cancer primary tumors, and DC3 infiltration correlated positively with CD8<sup>+</sup>CD103<sup>+</sup>CD69<sup>+</sup> tissue-resident memory T cells. Together, these findings define DC3s as a lineage of inflammatory DCs endowed with a strong potential to regulate tumor immunity.</p></abstract><abstract abstract-type="graphical" id="abs0015"><title>Graphical Abstract</title><fig id="undfig1" position="anchor" orientation="portrait"><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="fx1.jpg"><?image-name fx1.jpg?><?image-size 194278?><?image-md5 133afeb16df625bc8e127d374624d970?><?image-image-server-status LOAD_COMPLETED?><?image-original-height 996?><?image-original-width 996?><?image-scaled-height 664?><?image-scaled-width 664?><?image-cloudpmc-urn urn:cdn:blobs/922d/7445430/133afeb16df6/fx1.jpg?><?thumb-name fx1.gif?><?thumb-size 12903?><?thumb-md5 25e6e5a2bb4eb3f9231907c38f78ea02?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 100?><?thumb-scaled-width 100?><?thumb-cloudpmc-urn urn:cdn:blobs/922d/7445430/25e6e5a2bb4e/fx1.gif?></graphic></fig></abstract><abstract abstract-type="author-highlights" id="abs0020"><title>Highlights</title><p><list list-type="simple" id="ulist0010"><list-item id="u0010"><label>•</label><p id="p0010">DC3s are phenotypic and functional intermediates between cDC2s and monocytes</p></list-item><list-item id="u0015"><label>•</label><p id="p0015">GM-CSF alone, but not FLT3L, supports efficient differentiation of DC3s</p></list-item><list-item id="u0020"><label>•</label><p id="p0020">DC3s do not differentiate via cDC (CDP)- or monocyte-restricted (cMoP) progenitors</p></list-item><list-item id="u0025"><label>•</label><p id="p0025">DC3s prime T<sub>RM</sub> cells <italic toggle="yes">in vitro</italic> and correlate with T<sub>RM</sub> expansion in primary breast cancer</p></list-item></list></p></abstract><abstract abstract-type="teaser" id="abs0025"><p>Bourdely et al. identify human CD88<sup>−</sup>CD1c<sup>+</sup>CD163<sup>+</sup> DC3s as a pro-inflammatory phagocyte lineage sharing features with monocytes and conventional DCs. DC3s efficiently induce differentiation of CD103<sup>+</sup>CD8<sup>+</sup> T cells <italic toggle="yes">in vitro</italic>, and their infiltration correlates with CD8<sup>+</sup>CD69<sup>+</sup>CD103<sup>+</sup> T<sub>RM</sub> accumulation in breast tumors.</p></abstract><kwd-group id="kwrds0010"><title>Keywords</title><kwd>conventional DCs</kwd><kwd>DC3s</kwd><kwd>cDC2s</kwd><kwd>mononuclear phagocytes</kwd><kwd>inflammatory DCs</kwd><kwd>monocytes</kwd><kwd>DC progenitors</kwd><kwd>T<sub>RM</sub></kwd></kwd-group><custom-meta-group><custom-meta><meta-name>pmc-status-qastatus</meta-name><meta-value>0</meta-value></custom-meta><custom-meta><meta-name>pmc-status-live</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-status-embargo</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-status-released</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-open-access</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-olf</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-manuscript</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-legally-suppressed</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-has-pdf</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-has-supplement</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-pdf-only</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-suppress-copyright</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-is-real-version</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-is-scanned-article</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-preprint</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-in-epmc</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-license-ref</meta-name><meta-value>CC BY</meta-value></custom-meta></custom-meta-group></article-meta><notes><p id="misc0010">Published: June 30, 2020</p></notes></front><body><sec id="sec1"><title>Introduction</title><p id="p0030">Human dendritic cells (DCs) are sentinel cells of the immune system specialized in controlling T cell function (<xref rid="bib5" ref-type="bibr">Banchereau and Steinman, 1998</xref>; <xref rid="bib42" ref-type="bibr">Palucka and Banchereau, 2013</xref>; <xref rid="bib55" ref-type="bibr">Steinman et al., 2003</xref>). The mouse model has brought important concepts to our understanding of DCs and suggests that multiple DC subsets arising from specialized ontogenetic pathways are endowed with specific immune functions (<xref rid="bib9" ref-type="bibr">Briseño et al., 2016</xref>; <xref rid="bib18" ref-type="bibr">Guermonprez et al., 2019</xref>; <xref rid="bib34" ref-type="bibr">Merad et al., 2013</xref>; <xref rid="bib37" ref-type="bibr">Murphy et al., 2016</xref>).</p><p id="p0035">Definition of human DC subsets is a prerequisite to understanding the division of labor underpinning induction of various types of immune responses. At homeostasis, conventional DCs (cDCs) include cDC1s (CD141<sup>+</sup>XCR1<sup>+</sup>CLEC9A<sup>+</sup>IRF8<sup>+</sup>) and cDC2s (CD1c<sup>+</sup>CD11c<sup>+</sup>CD172a<sup>+</sup>IRF4<sup>+</sup>) (<xref rid="bib3" ref-type="bibr">Bachem et al., 2010</xref>; <xref rid="bib11" ref-type="bibr">Crozat et al., 2010</xref>; <xref rid="bib19" ref-type="bibr">Heidkamp et al., 2016</xref>; <xref rid="bib22" ref-type="bibr">Jongbloed et al., 2010</xref>; <xref rid="bib50" ref-type="bibr">Schlitzer et al., 2013</xref>). cDC1s and cDC2s arise through a specialized ontogenetic pathway from a common DC precursor (CDP) (<xref rid="bib28" ref-type="bibr">Lee et al., 2015b</xref>) or from early IRF8<sup>+</sup> multipotent lympho-myeloid progenitors (MLPs) (<xref rid="bib20" ref-type="bibr">Helft et al., 2017</xref>; <xref rid="bib29" ref-type="bibr">Lee et al., 2017</xref>). Bone marrow progenitors for cDCs generate a common circulating precursor that progressively diverges into pre-cDC1 and pre-cDC2 (<xref rid="bib7" ref-type="bibr">Breton et al., 2015</xref>, <xref rid="bib8" ref-type="bibr">2016</xref>; <xref rid="bib51" ref-type="bibr">See et al., 2017</xref>). This is further complicated by inclusion of AXL<sup>+</sup>SIGLEC6<sup>+</sup>CD11c<sup>+</sup>CD1c<sup>+</sup> cells (AS-DCs also called type 5 DCs), which have been proposed to act as precursors for cDCs (pre-cDCs) or a lineage on its own (<xref rid="bib51" ref-type="bibr">See et al., 2017</xref>; <xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>).</p><p id="p0040">An additional layer of complexity in the DC network lies in its responsiveness to perturbations. For instance, inflammation affects hematopoiesis and phagocyte trafficking, resulting in leukocyte mobilization and tissue infiltration. Specifically, inflammation affects DC diversity and triggers mobilization of CD14<sup>+</sup>CD1c<sup>+</sup> DCs, called inflammatory DCs (iDCs) (<xref rid="bib6" ref-type="bibr">Binnewies et al., 2019</xref>; <xref rid="bib16" ref-type="bibr">Granot et al., 2017</xref>; <xref rid="bib52" ref-type="bibr">Segura et al., 2012</xref>, <xref rid="bib53" ref-type="bibr">2013</xref>; <xref rid="bib61" ref-type="bibr">Wollenberg et al., 1996</xref>; <xref rid="bib64" ref-type="bibr">Zaba et al., 2009</xref>). The expression of CD1c lectin is shared between iDCs and cDCs. However, CD1c<sup>+</sup> iDCs also express multiple monocytic markers, such as CD14, CCR2, and FcγRI/CD64. iDCs have been reported in inflamed skin, synovial fluid, ovarian cancer ascites, solid tumor infiltrates, and lymph nodes (<xref rid="bib4" ref-type="bibr">Bakdash et al., 2016</xref>; <xref rid="bib6" ref-type="bibr">Binnewies et al., 2019</xref>; <xref rid="bib16" ref-type="bibr">Granot et al., 2017</xref>; <xref rid="bib25" ref-type="bibr">Lavin et al., 2017</xref>; <xref rid="bib52" ref-type="bibr">Segura et al., 2012</xref>, <xref rid="bib53" ref-type="bibr">2013</xref>; <xref rid="bib61" ref-type="bibr">Wollenberg et al., 1996</xref>; <xref rid="bib64" ref-type="bibr">Zaba et al., 2009</xref>). The developmental pathway of human CD1c<sup>+</sup>CD14<sup>+</sup> iDCs is poorly understood. <italic toggle="yes">In vitro</italic> studies suggest that iDCs obtained after differentiation of CD14<sup>+</sup> monocytes in granulocyte macrophage colony-stimulating factor (GM-CSF) and interleukin-4 (IL-4) (<xref rid="bib47" ref-type="bibr">Sallusto and Lanzavecchia, 1994</xref>) might correspond to <italic toggle="yes">in vivo</italic> iDCs (<xref rid="bib16" ref-type="bibr">Granot et al., 2017</xref>; <xref rid="bib52" ref-type="bibr">Segura et al., 2012</xref>, <xref rid="bib53" ref-type="bibr">2013</xref>). In this context, IL-4 acts through induction of the transcriptional regulator NCOR2 (<xref rid="bib48" ref-type="bibr">Sander et al., 2017</xref>). In addition, triggering the aryl hydrocarbon receptor in monocytes supports activation of IRF4-dependent differentiation of iDCs (<xref rid="bib15" ref-type="bibr">Goudot et al., 2017</xref>). Together, these studies support the prevailing notion that CD14<sup>+</sup> monocytes act as immediate precursors for iDCs.</p><p id="p0045">Re-evaluation of circulating mononuclear phagocyte diversity has been enabled by single-cell RNA sequencing (scRNA-seq). Recent studies have revealed that a subset of DC-like cells, called DC3s, express mRNA for the CD14 and CD1c genes (<xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>). However, this analysis was performed after excluding cells expressing the highest amount of CD14 (<xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>). As a consequence, this approach renders a problematic distinction between DC3s and <italic toggle="yes">bona fide</italic> CD14<sup>+</sup> monocytes (<xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>). This discrimination is further complicated by previous reports of CD14<sup>+</sup>CD1c<sup>+</sup> “inflammatory” DCs recruited at inflammatory sites (<xref rid="bib6" ref-type="bibr">Binnewies et al., 2019</xref>; <xref rid="bib16" ref-type="bibr">Granot et al., 2017</xref>; <xref rid="bib52" ref-type="bibr">Segura et al., 2012</xref>, <xref rid="bib53" ref-type="bibr">2013</xref>; <xref rid="bib61" ref-type="bibr">Wollenberg et al., 1996</xref>; <xref rid="bib64" ref-type="bibr">Zaba et al., 2009</xref>).</p><p id="p0050">Here we intended to re-evaluate the definition of DC3s using unbiased scRNA-seq and high-dimensional flow cytometry by exploring the full spectrum of CD14 and CD1c expression. In addition, we identify DC3 growth factor requirements and developmental pathways. Finally, we show that DC3s activate CD103<sup>+</sup> T cells and that DC3 infiltration in human breast tumors correlates with the abundance of CD8<sup>+</sup>CD103<sup>+</sup>CD69<sup>+</sup> tissue-resident memory (T<sub>RM</sub>) T cells.</p></sec><sec id="sec2"><title>Results</title><sec id="sec2.1"><title>DC3s Represent a Discrete Subset of CD88<sup>−</sup>CD1c<sup>+</sup>CD163<sup>+</sup> Cells in Human Peripheral Blood</title><p id="p0055">To probe the diversity of CD16<sup>−</sup>CD141<sup>−</sup>CD123<sup>−</sup> blood mononuclear phagocytes, we developed a sorting strategy including all phenotypic intermediates between CD14<sup>hi</sup>CD1c<sup>lo</sup> and CD14<sup>lo</sup>CD1c<sup>hi</sup> cells. The proportions between cell populations were compensated to enrich in less abundant CD14<sup>lo</sup>CD1c<sup>hi</sup> cells (<xref rid="mmc1" ref-type="supplementary-material">Figure S1</xref>A). Flow cytometry-sorted cells isolated from blood were analyzed using a droplet-based scRNA-seq approach (<xref rid="fig1" ref-type="fig">Figure 1</xref>A; <xref rid="mmc1" ref-type="supplementary-material">Figure S1</xref>A). We found that cells expressing CD14 and/or CD1c could be separated into four CD33<sup>+</sup> clusters (A, B, C, and D) (<xref rid="fig1" ref-type="fig">Figure 1</xref>A; <xref rid="mmc1" ref-type="supplementary-material">Figure S1</xref>B). Contaminating clusters containing B and T lymphocytes and neutrophils were excluded from the analysis (<xref rid="mmc1" ref-type="supplementary-material">Figure S1</xref>B). Hierarchical clustering performed on averaged single cell expression data within clusters showed that A and B were closer to each other than any of the other subsets (<xref rid="fig1" ref-type="fig">Figures 1</xref>B–1D). Cluster D fell between the group of clusters A and B and cluster C (<xref rid="fig1" ref-type="fig">Figure 1</xref>B). Classical cDC2 markers, such as C<italic toggle="yes">LEC10A</italic>, <italic toggle="yes">FCER1A</italic>, and major histocompatibility complex (MHC) class II genes (<italic toggle="yes">HLA-DQA</italic>, <italic toggle="yes">HLA-DPA</italic>, <italic toggle="yes">HLA-DRA</italic>, and the MHC class II-associated invariant chain <italic toggle="yes">CD74</italic>) were expressed prominently in clusters A, B, and D (<xref rid="fig1" ref-type="fig">Figures 1</xref>D and 1E; <xref rid="mmc1" ref-type="supplementary-material">Figure S1</xref>C; <xref rid="bib19" ref-type="bibr">Heidkamp et al., 2016</xref>; <xref rid="bib25" ref-type="bibr">Lavin et al., 2017</xref>; <xref rid="bib53" ref-type="bibr">Segura et al., 2013</xref>). In contrast, monocytic markers such as <italic toggle="yes">CD14</italic>, <italic toggle="yes">S100A8</italic>, <italic toggle="yes">S100A9</italic>, <italic toggle="yes">S100A12</italic>, and <italic toggle="yes">VCAN</italic> were more expressed in clusters C and D, with higher expression in C compared with D (<xref rid="fig1" ref-type="fig">Figures 1</xref>D and 1E). Finally, expression of the C5 receptor <italic toggle="yes">C5AR1</italic> (CD88) was found to be restricted to cluster C together with <italic toggle="yes">SOD2</italic> and <italic toggle="yes">RBP7</italic> (<xref rid="fig1" ref-type="fig">Figures 1</xref>D and 1E).<fig id="fig1" position="float" orientation="portrait"><label>Figure 1</label><caption><p>DC3s Are a Discrete Subset of CD88<sup>−</sup>CD1c<sup>+</sup>CD163<sup>+</sup> Cells in Human Peripheral Blood</p><p>(A) Gating strategy used to define mononuclear phagocytes expressing CD14 and/or CD1c. Cells expressing CD14 and/or CD1c were sorted by flow cytometry from 3 healthy donors and pooled before scRNA-seq analysis. To improve the resolution of CD1c<sup>+</sup> subsets, the cellular input was enriched in CD1<sup>high</sup> cells (<xref rid="mmc1" ref-type="supplementary-material">Figure S1</xref>A). Single cells were isolated using a droplet-based approach and sequenced. Dimensionality reduction of scRNA-seq data was performed using dimensionality reduction (t-distributed stochastic neighbor embedding [tSNE]). Clusters A, B, C, and D were identified using the shared nearest neighbor (SNN) clustering algorithm. Each dot represents an individual cell (n = 1,622).</p><p>(B) Hierarchal clustering of groups A, B, C, and D based on average gene expression (14,933 genes).</p><p>(C) Absolute number of differentially expressed genes (DEGs) for pairwise comparisons between groups A, B, and D.</p><p>(D) Heatmaps displaying relative expression of up to 20 DEGs defining each cluster.</p><p>(E) Violin plots illustrating expression probability distributions across clusters of representative DEGs (226 total DEGs). Feature plots display the average expression of groups of genes (identified in violin plots) in each cell of the tSNE plot defined in (A).</p><p>(F) Expression distribution across clusters A, B, C, and D of gene signatures identified by <xref rid="bib58" ref-type="bibr">Villani et al. (2017)</xref> and <xref rid="bib62" ref-type="bibr">Yin et al. (2017)</xref>. (<sup>∗∗</sup>p &lt; 0.01, <sup>∗∗∗∗</sup>p &lt; 0.0001, one-way ANOVA test)</p><p>(G) Identification of 4 subsets within CD14<sup>lo to hi</sup> CD1c<sup>lo to hi</sup> cells by unsupervised clustering of flow cytometry data using the FlowSOM algorithm. tSNE and unsupervised clustering were performed using the following markers: CD88, CD1c, FcεRI, CD14, CD163, BTLA, CD123, and CD5. tSNE plots (right) display the relative expression of each marker among the subsets. Dot plots (below) show the expression of specific markers in clusters 1, 2, and 3 when combined in 2-dimensional analysis.</p><p>(H) Improved gating strategy for identification of cDC2s, DC3s, and CD14<sup>+</sup> monocytes in circulating PBMCs and histograms showing expression of S100A8/9, FcεRI, CD5, CD14, CD116, and CD206.</p><p>(I) Principal-component analysis (PCA) for bulk-sequenced mononuclear phagocyte populations as defined in (H).</p><p>(J) Cluster dendrogram of the different cell types using the 2,000 most variable genes.</p><p>(K) Heatmaps comparing the relative expression of markers discriminating clusters in scRNA-seq analysis (A, B, C, and D, left) and in bulk RNA-seq analysis on sorted subsets based on the gating strategy defined in (H) (right).</p><p>See also <xref rid="mmc1" ref-type="supplementary-material">Figure S1</xref> and <xref rid="mmc2" ref-type="supplementary-material">Table S1</xref>.</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="gr1.jpg"><?image-name gr1.jpg?><?image-size 532689?><?image-md5 f2b9bb05974b01a8520dab288e4e9a52?><?image-image-server-status LOAD_COMPLETED?><?image-original-height 4226?><?image-original-width 2965?><?image-scaled-height 1056?><?image-scaled-width 741?><?image-cloudpmc-urn urn:cdn:blobs/922d/7445430/f2b9bb05974b/gr1.jpg?><?thumb-name gr1.gif?><?thumb-size 17668?><?thumb-md5 57ec990f1d07c084ea2b25d0ee09152a?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 143?><?thumb-scaled-width 100?><?thumb-cloudpmc-urn urn:cdn:blobs/922d/7445430/57ec990f1d07/gr1.gif?></graphic></fig></p><p id="p0060">To determine whether our clustering reflects previously published data, we decided to evaluate the expression of gene signatures obtained from <xref rid="bib58" ref-type="bibr">Villani et al. (2017)</xref>. We found that signature genes discriminating cDC2 within CD14<sup>lo</sup> cells were mostly represented in clusters A and B, confirming their identity (<xref rid="fig1" ref-type="fig">Figure 1</xref>F). Likewise, signature genes defining DC3s within CD14<sup>lo</sup> cells (<xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>) and CD14<sup>+</sup> monocytes were significantly enriched in cluster D and cluster C, respectively (<xref rid="fig1" ref-type="fig">Figure 1</xref>F). Of note, genes enriched in DC3s compared with cDC2s (<italic toggle="yes">S100A8</italic>, <italic toggle="yes">S100A9</italic>, and <italic toggle="yes">CD14</italic>) were highly expressed in clusters C and D (<xref rid="fig1" ref-type="fig">Figure 1</xref>E; <xref rid="mmc1" ref-type="supplementary-material">Figure S1</xref>E).This underlines the need to integrate a monocyte reference in any comparison aiming to define DC3s. We conclude that scRNA-seq analysis identifies DC3s (cluster D) as a specific subset sharing transcriptional features of cDC2s (clusters A and B) and monocytes (cluster C).</p><p id="p0065">Next we sought to define a flow cytometry-based strategy enabling analysis and prospective isolation of DC3s in blood. To this end, we performed an unsupervised flow cytometry data analysis based on genes identified by scRNA-seq (e.g., CD88, CD14, FcεRI, and CD1c) as well as markers previously associated with cDC2s (e.g., BTLA and CD5) (<xref rid="bib62" ref-type="bibr">Yin et al., 2017</xref>) and DC3s (e.g., CD163) (<xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>), even though they were not detected in the scRNA-seq analysis. We identified three main clusters (1, 2, and 3) together with a rarer cluster (4) of CD5<sup>hi</sup> cells (<xref rid="fig1" ref-type="fig">Figure 1</xref>G; <xref rid="mmc1" ref-type="supplementary-material">Figure S1</xref>D). Cluster 4 appeared to be CD123<sup>hi</sup> contamination of AS-DCs (<xref rid="mmc1" ref-type="supplementary-material">Figure S1</xref>D). Cluster 1 highly expressed CD88, aligning with monocyte cluster C identified by scRNA-seq. Cluster 2 did not express the monocyte-associated markers CD14 and CD88 but was characterized by expression of CD1c, FcεRI, and B and T lymphocyte attenuator (BTLA) (<xref rid="fig1" ref-type="fig">Figure 1</xref>G). In addition, cluster 2 showed heterogeneous expression of CD5 (<xref rid="fig1" ref-type="fig">Figure 1</xref>G). Similar to cluster 2, cluster 3 did not express the monocytic marker CD88 and displayed higher amounts of CD1c and FcεRI. However, cluster 3 could be distinguished from cluster 2 by higher expression of CD163 and heterogeneous expression of CD14 (<xref rid="fig1" ref-type="fig">Figure 1</xref>G). The unsupervised flow cytometry analysis allowed us to define a simple gating strategy enabling prospective isolation of monocytes (CD88<sup>+</sup>CD14<sup>+</sup>), cDC2s (CD5<sup>+</sup> and CD5<sup>−</sup>CD88<sup>−</sup>CD1c<sup>+</sup>BTLA<sup>+</sup>CD163<sup>−</sup>), and DC3s (CD88<sup>−</sup>CD1c<sup>+</sup>BTLA<sup>−</sup>CD163<sup>+</sup>) (<xref rid="fig1" ref-type="fig">Figure 1</xref>H).</p><p id="p0070">To further validate the flow cytometry-based identification of cell subsets, we performed bulk RNA-seq analysis of sorted monocytes, cDC2s (CD5<sup>+</sup> and CD5<sup>−</sup>), DC3s, plasmacytoid DCs (pDCs), and AS-DCs (<xref rid="mmc1" ref-type="supplementary-material">Figure S1</xref>E). Principal-component analysis (PCA) indicated that, overall, CD5<sup>+</sup> and CD5<sup>−</sup> cDC2s, DC3s, and monocytes separated from pDCs and AS-DCs along the PC1 axis, accounting for 61% variance (<xref rid="fig1" ref-type="fig">Figure 1</xref>I). CD5<sup>+</sup> and CD5<sup>−</sup> cDC2 clustered closely together, and DC3s separated clearly from cDC2s and monocytes (<xref rid="fig1" ref-type="fig">Figure 1</xref>I; <xref rid="mmc1" ref-type="supplementary-material">Figure S1</xref>F). Hierarchical clustering (HC) and differentially expressed gene analysis led to the same conclusion, with DC3s sitting between monocytes and cDC2s (CD5<sup>+</sup> and CD5<sup>−</sup>) (<xref rid="fig1" ref-type="fig">Figure 1</xref>J; <xref rid="mmc1" ref-type="supplementary-material">Figure S1</xref>G). Overall, DC3s are closer to cDC2s than monocytes (<xref rid="fig1" ref-type="fig">Figures 1</xref>I and 1J).</p><p id="p0075">Together, this validates that cellular clusters isolated by a flow cytometry-based approach align to clusters identified by unbiased scRNA-seq (<xref rid="fig1" ref-type="fig">Figure 1</xref>K).</p><p id="p0080">We conclude that DC3s are a separable entity within CD1c<sup>+</sup> cells, defined by a distinct gene expression profile, and that they can be prospectively isolated using CD88, CD1c, CD163, and BTLA (<xref rid="mmc1" ref-type="supplementary-material">Figure S1</xref>H). scRNA-seq and bulk gene expression profiling identify markers shared between DC3s and cDC2s (e.g., <italic toggle="yes">CLEC10A</italic> and <italic toggle="yes">FCER1A</italic>) and markers shared between DC3s and monocytes (such as <italic toggle="yes">S100A8</italic>, <italic toggle="yes">S100A9</italic>, <italic toggle="yes">CD14</italic>, and <italic toggle="yes">CD163</italic>).</p></sec><sec id="sec2.2"><title>DC3s Give Rise to CD14<sup>+</sup>CD1c<sup>+</sup> DCs Infiltrating Tumors</title><p id="p0085">Tumor-infiltrating CD14<sup>+</sup>CD1c<sup>+</sup> DCs have been reported in multiple instances, including ovarian cancer ascites (<xref rid="bib53" ref-type="bibr">Segura et al., 2013</xref>), breast cancer (<xref rid="bib35" ref-type="bibr">Michea et al., 2018</xref>), and melanoma (<xref rid="bib4" ref-type="bibr">Bakdash et al., 2016</xref>; <xref rid="bib6" ref-type="bibr">Binnewies et al., 2019</xref>). Therefore, we asked whether CD14<sup>+</sup>CD1c<sup>+</sup> cells would align with DC3s. To this end, we analyzed mononuclear phagocytes infiltrating luminal breast cancer primary tumors. Using the gating strategy described in <xref rid="fig1" ref-type="fig">Figure 1</xref>, we found that, after exclusion of CD88<sup>+</sup> monocytes and macrophages, the remaining CD45<sup>+</sup>HLA-DR<sup>+</sup>CD123<sup>−</sup>CD88<sup>−</sup> fraction contained cDC1s, CD14<sup>−</sup>CD1c<sup>+</sup>CD163<sup>−</sup>CD5<sup>+</sup> cDC2s (CD5<sup>+</sup> cDC2s), and CD14<sup>+</sup>CD1c<sup>+</sup>CD163<sup>+</sup>CD5<sup>−</sup> DC3s (CD14<sup>+</sup> DC3s) (<xref rid="fig2" ref-type="fig">Figure 2</xref>A; <xref rid="mmc1" ref-type="supplementary-material">Figure S2</xref>A). At this stage, we observed that CD163 and BTLA were particularly susceptible to enzymatic digestion of solid tissue, preventing consistent and reliable quantification of cells throughout the cohort of samples. Hence, we adopted a more restrictive definition of cDC2s and DC3s as CD1c<sup>+</sup>CD14<sup>−</sup>CD5<sup>+</sup> and CD1c<sup>+</sup>CD14<sup>+</sup>CD5<sup>−</sup> cells, respectively (<xref rid="fig2" ref-type="fig">Figure 2</xref>A). This strategy enabled isolation of CD5<sup>+</sup> cDC2s and CD14<sup>+</sup> DC3s in peripheral tissues even though it might result in underestimation of their absolute numbers. Nevertheless, both subsets aligned phenotypically with their blood counterparts (<xref rid="mmc1" ref-type="supplementary-material">Figures S2</xref>B and S2C) and expressed markers reported previously for CD1c<sup>+</sup>CD14<sup>+</sup> iDCs, such as CD11c and FcεRI (<xref rid="fig2" ref-type="fig">Figure 2</xref>A; <xref rid="bib53" ref-type="bibr">Segura et al., 2013</xref>).<fig id="fig2" position="float" orientation="portrait"><label>Figure 2</label><caption><p>DC3s Infiltrate Human Breast Tumors</p><p>(A) Representative gating strategy used to define macrophages, CD5<sup>+</sup> cDC2s, and CD14<sup>+</sup> DC3s and histograms showing the expression of CD163, FcεRI, BTLA, and CD11c in human breast cancer primary tumors.</p><p>(B) Violin plot quantifying cDC1, CD5<sup>+</sup>cDC2, CD14<sup>+</sup> DC3, and CD14<sup>+</sup>CD88<sup>+</sup> macrophage subsets identified in (A) in human breast cancer primary tumors (n = 25; <sup>∗</sup>p &lt; 0.05, <sup>∗∗</sup>p &lt; 0.01, <sup>∗∗∗</sup>p &lt; 0.001, <sup>∗∗∗∗</sup>p &lt; 0.0001, one-way ANOVA test).</p><p>(C) Pearson correlations of the frequencies of macrophages, cDC1s, CD5<sup>+</sup>cDC2s, and CD14<sup>+</sup> DC3s within HLA-DR<sup>+</sup> cells in human breast cancer primary tumors (red, significantly correlated p &lt; 0.05; black, not correlated).</p><p>(D) HC showing the relative expression of markers used for subset identification in <xref rid="fig1" ref-type="fig">Figure 1</xref> in CD1c<sup>+</sup>, CD1c<sup>+</sup>CD14<sup>+</sup>, and CD14<sup>+</sup> cells from invaded lymph nodes draining human breast cancer primary tumors.</p><p>(E) GSEA of pairwise comparisons of CD1c<sup>+</sup>CD14<sup>+</sup> cells with CD1c<sup>+</sup> or CD14<sup>+</sup> from invaded lymph nodes draining human breast cancer primary tumors. Gene signatures of blood DC3s compared with cDC2s (DC3 &gt; cDC2) or CD14<sup>+</sup> monocytes (DC3 &gt; Mono) and, vice versa, of blood cDC2s (cDC2 &gt; DC3) or CD14<sup>+</sup> monocytes (Mono &gt; DC3) compared with DC3s were used (<xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>).</p><p>See also <xref rid="mmc1" ref-type="supplementary-material">Figure S2</xref> and <xref rid="mmc1" ref-type="supplementary-material">Table S2</xref>.</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="gr2.jpg"><?image-name gr2.jpg?><?image-size 303105?><?image-md5 9c233af0377064f7c060f65a71d0fb5e?><?image-image-server-status LOAD_COMPLETED?><?image-original-height 2783?><?image-original-width 2906?><?image-scaled-height 695?><?image-scaled-width 726?><?image-cloudpmc-urn urn:cdn:blobs/922d/7445430/9c233af03770/gr2.jpg?><?thumb-name gr2.gif?><?thumb-size 16401?><?thumb-md5 a3857d32563629f00a0b3758d1121d51?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 96?><?thumb-scaled-width 100?><?thumb-cloudpmc-urn urn:cdn:blobs/922d/7445430/a3857d325636/gr2.gif?></graphic></fig></p><p id="p0090">DC3s were consistently identified in 25 samples of primary tumors of clinical stages I, II, and III (<xref rid="mmc1" ref-type="supplementary-material">Figure S2</xref>D). Macrophages represented by far the most abundant population. DC3s outnumbered cDC1s but were on par with <italic toggle="yes">bona fide</italic> CD5<sup>+</sup> cDC2s (<xref rid="fig2" ref-type="fig">Figure 2</xref>B). The relative abundance of DC3s did not correlate with disease progression (<xref rid="mmc1" ref-type="supplementary-material">Figure S2</xref>D) or with macrophages or cDCs (<xref rid="fig2" ref-type="fig">Figure 2</xref>C). In contrast, cDC1 infiltration correlated with cDC2s (<xref rid="fig2" ref-type="fig">Figure 2</xref>C).</p><p id="p0095">Bulk RNA-seq analysis of CD1c<sup>+</sup>CD14<sup>+</sup> cells sorted from tumor-invaded lymph nodes indicated that they displayed a similar expression profile as blood DC3s (high expression of cDC2 markers such as <italic toggle="yes">CLEC10A</italic> and <italic toggle="yes">FCER1A</italic> combined with low expression of monocyte-associated markers such as <italic toggle="yes">C5AR1</italic> and <italic toggle="yes">SOD2</italic>) (<xref rid="fig2" ref-type="fig">Figure 2</xref>D). Gene set enrichment analysis (GSEA) revealed that CD14<sup>+</sup>CD1c<sup>+</sup> were more enriched for the DC3 &gt; cDC2 signature compared with CD1c<sup>+</sup> cells (normalized enrichment score [NES], 2.06; p = 0.0; <xref rid="bib56" ref-type="bibr">Subramanian et al., 2005</xref>; <xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>). In addition, CD14<sup>+</sup>CD1c<sup>+</sup> were more enriched for the DC3 &gt; monocyte (Mono) signature compared with CD14<sup>+</sup> cells (NES, 1.42; p = 0.02; <xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>; <xref rid="fig2" ref-type="fig">Figure 2</xref>E). Conversely, compared with CD1c<sup>+</sup>CD14<sup>+</sup>, the cDC2 &gt; DC3 and Mono &gt; DC3 gene signatures were enriched in CD1c<sup>+</sup> (NES, −1.57; p = 0.004) and CD14<sup>+</sup> (NES, −2.07; p = 0.0) cells, respectively (<xref rid="fig2" ref-type="fig">Figure 2</xref>E).</p><p id="p0100">We conclude that CD11c<sup>+</sup>FcεRI <sup>+</sup>CD14<sup>+</sup>CD1c<sup>+</sup> iDCs infiltrating breast cancer align with DC3s.</p><p id="p0105">Inflammatory cues promoting mobilization of CD14<sup>+</sup>CD1c<sup>+</sup> cells at the site of inflammation are not fully defined. Mouse studies define GM-CSF as a likely candidate (<xref rid="bib30" ref-type="bibr">Mach et al., 2000</xref>; <xref rid="bib33" ref-type="bibr">Menezes et al., 2016</xref>). For this reason, we decided to test whether GM-CSF was sufficient to mobilize human CD14<sup>+</sup>CD1c<sup>+</sup> DCs in a humanized mouse metastatic lung model. We generated B16 mouse melanoma engineered to overexpress human GM-CSF (B16_huGM) or FLT3L (B16_huFLT3L) (<xref rid="mmc1" ref-type="supplementary-material">Figure S3</xref>A). Immunodeficient NOD.Cg-<italic toggle="yes">Prkdc</italic><sup><italic toggle="yes">scid</italic></sup>
+<italic toggle="yes">Il2rg</italic><sup><italic toggle="yes">tm1Wjl</italic></sup>/SzJ (NSG) mice were injected intravenously with B16 control (CTRL), B16_huGM, or B16_huFLT3L (<xref rid="fig3" ref-type="fig">Figure 3</xref>A). Lung metastasis-bearing mice were engrafted with human peripheral blood mononuclear cells (PBMCs). Two days later, human CD45<sup>+</sup> leukocytes were found in lung tumor foci and juxta-tumor areas (<xref rid="fig3" ref-type="fig">Figure 3</xref>B; <xref rid="mmc1" ref-type="supplementary-material">Figures S3</xref>B and S3C). Flow cytometry analysis of metastatic lungs showed that FLT3L, but not GM-CSF, promotes expansion of CD1c<sup>+</sup>CD5<sup>+</sup> cells aligning with blood cDC2s (<xref rid="fig3" ref-type="fig">Figure 3</xref>C, blue). In contrast, GM-CSF, but not FLT3L, led to accumulation of CD1c<sup>+</sup>CD14<sup>+</sup> cells aligning with blood DC3s (<xref rid="fig3" ref-type="fig">Figure 3</xref>C, orange; <xref rid="mmc1" ref-type="supplementary-material">Figure S3</xref>D, orange). As shown for circulating peripheral blood subsets in <xref rid="fig1" ref-type="fig">Figure 1</xref>, FLT3L-dependent cDC2s and GM-CSF-dependent DC3s shared expression of Clec10A, CD11c, and FcεRI (<xref rid="fig3" ref-type="fig">Figure 3</xref>C; <xref rid="mmc1" ref-type="supplementary-material">Figure S3</xref>E). All tumor-bearing lungs contained some monocytes and/or macrophages (CD14<sup>+</sup>CD88<sup>+</sup>) (<xref rid="fig3" ref-type="fig">Figure 3</xref>C; <xref rid="mmc1" ref-type="supplementary-material">Figure S3</xref>B). To further establish alignment of GM-CSF-mobilized DC3s, we performed an unbiased scRNA-seq on human CD45<sup>+</sup>HLA-DR<sup>+</sup> cells expressing CD14 and/or CD1c (<xref rid="mmc1" ref-type="supplementary-material">Figures S3</xref>F and S3G). Two major clusters could be identified. Cluster 0 was characterized by expression of genes associated with DC3s and cDC2s, such as <italic toggle="yes">CLEC10A</italic>, <italic toggle="yes">FCER1A</italic>, and <italic toggle="yes">CD74</italic> (<xref rid="mmc1" ref-type="supplementary-material">Figures S3</xref>H and S3I), and was enriched in DC3 differentially expressed transcripts (<xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>; <xref rid="mmc1" ref-type="supplementary-material">Figure S3</xref>J). Conversely, cluster 1 expressed markers defining monocytes (<italic toggle="yes">SOD2</italic>, <italic toggle="yes">C5AR1</italic>, and <italic toggle="yes">G0S2</italic>) and <italic toggle="yes">CXCL2</italic>, <italic toggle="yes">CXCL3</italic>, and <italic toggle="yes">CXCL8</italic> chemokines (<xref rid="mmc1" ref-type="supplementary-material">Figures S3</xref>H and S3I). We conclude that GM-CSF drives mobilization of CD14<sup>+</sup>CD1c<sup>+</sup> cells aligning phenotypically with circulating DC3s.<fig id="fig3" position="float" orientation="portrait"><label>Figure 3</label><caption><p>DC3s Give Rise to CD14<sup>+</sup>CD1c<sup>+</sup> DCs at Inflammatory Sites</p><p>(A) Experimental model. NSG mice were injected intravenously (i.v.) with B16_CTRL, B16_huFLT3L, or B16_huGM on day 0. On days 7 and 8, 10<sup>8</sup> human PBMCs were injected i.v. Metastatic lungs were collected on day 9.</p><p>(B) Pseudocolor images of B16_huGM (green) metastatic lung on day 9 post-injection, stained for human CD45 (red). Nuclei were stained with Hoechst (blue). Scale bar, 100 μm.</p><p>(C) Gating strategy for cDC2 and DC3 identification in B16_huGM and B16_huFLT3L metastatic mouse lung and histograms showing the expression of CD163, CD206, and Clec10A. The bar graph summarizes the frequency of cDC2s and DC3s among total HLA-DR<sup>+</sup> cells in metastatic B16_CTRL, B16_huFLT3L, or B16_huGM mouse lungs (n = 3 independent mice; <sup>∗∗</sup>p &lt; 0.01, <sup>∗∗∗</sup>p &lt; 0.001, <sup>∗∗∗∗</sup>p &lt; 0.0001, one-way ANOVA test).</p><p>(D) Flow cytometry analysis of flow cytometry-sorted cDC2s, DC3s, and CD14<sup>+</sup> monocytes after 2 days of culture with MS5 stromal cells expressing human GM-CSF (MS5_GM). Bar graphs show the frequency of output cells among total huCD45<sup>+</sup> cells (n = 4–5 healthy donors).</p><p>(E) Histograms showing CD14 expression on cDC2s, DC3s, and CD14<sup>+</sup> monocytes before and after 2 days of coculture with MS5_GM and bar graphs summarizing the frequency of CD14 expression within each cell type (n = 5 healthy donors, <sup>∗∗</sup>p &lt; 0.01, Mann-Whitney two-tailed t test).</p><p>See also <xref rid="mmc1" ref-type="supplementary-material">Figures S3</xref> and <xref rid="mmc1" ref-type="supplementary-material">S4</xref> and <xref rid="mmc2" ref-type="supplementary-material">Table S1</xref>.</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="gr3.jpg"><?image-name gr3.jpg?><?image-size 326012?><?image-md5 1bec257516ee884d4ec3d7839de96f0a?><?image-image-server-status LOAD_COMPLETED?><?image-original-height 3448?><?image-original-width 2897?><?image-scaled-height 862?><?image-scaled-width 724?><?image-cloudpmc-urn urn:cdn:blobs/922d/7445430/1bec257516ee/gr3.jpg?><?thumb-name gr3.gif?><?thumb-size 15186?><?thumb-md5 5c2557532e7de844c37b0629fa58cef7?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 118?><?thumb-scaled-width 99?><?thumb-cloudpmc-urn urn:cdn:blobs/922d/7445430/5c2557532e7d/gr3.gif?></graphic></fig></p><p id="p0110">We next wondered whether GM-CSF could induce <italic toggle="yes">trans</italic>-differentiation of circulating cDC2s or monocytes into CD14<sup>+</sup>CD1c<sup>+</sup> DC3s. To test this hypothesis, cDC2s, DC3s, and monocytes were sorted by flow cytometry from blood and cocultured <italic toggle="yes">in vitro</italic> in the presence of GM-CSF-expressing stromal cells (MS5_GM). After 2 days, cDC2s did not acquire CD163 or CD14 expression (<xref rid="fig3" ref-type="fig">Figure 3</xref>D). In contrast, DC3s did upregulate CD14 (<xref rid="fig3" ref-type="fig">Figures 3</xref>D and 3E). We found that monocytes differentiated into CD16<sup>+</sup> and CD16<sup>−</sup> macrophages (<xref rid="fig3" ref-type="fig">Figure 3</xref>D). Importantly, CD88<sup>+</sup>CD1c<sup>−</sup>CD14<sup>+</sup> monocytes did not give rise to CD1c<sup>hi</sup>CD163<sup>hi</sup> DC3s (<xref rid="fig3" ref-type="fig">Figure 3</xref>D). Addition of IL-4 to GM-CSF culture did not affect the outcome (<xref rid="mmc1" ref-type="supplementary-material">Figure S4</xref>A). The lack of <italic toggle="yes">trans</italic>-differentiation of cDC2s or monocytes into DC3s was also confirmed <italic toggle="yes">in vivo</italic> upon adoptive transfer into immunodeficient NSG mice carrying GM-CSF-expressing tumors (<xref rid="mmc1" ref-type="supplementary-material">Figures S4</xref>B and S4C). Overall, we conclude that CD88<sup>−</sup>CD14<sup>+</sup>CD1c<sup>+</sup> cells can differentiate from DC3s independent of cDC or monocytic lineages.</p></sec><sec id="sec2.3"><title>GM-CSF Stimulates Differentiation of CD34<sup>+</sup> Hematopoietic Progenitors into DC3s <italic toggle="yes">In Vitro</italic></title><p id="p0115">We next wondered how DC3s would differentiate from bone marrow progenitors. To this end, we sought to define a culture system capable of generating DC3s together with cDC2s and macrophages. We cocultured human cord blood-derived CD34<sup>+</sup> hematopoietic stem and progenitor cells (HSPCs) with stromal cells engineered to overexpress cDC-promoting factors (membrane-bound FLT3L together with stem cell factor (SCF) and CXCL12 [MS5_FS12]; <xref rid="bib2" ref-type="bibr">Anselmi et al., 2020</xref>) in the presence or absence of GM-CSF (<xref rid="fig4" ref-type="fig">Figure 4</xref>A; <xref rid="mmc1" ref-type="supplementary-material">Figure S5</xref>A). We found that GM-CSF increased differentiation of CD1c<sup>+</sup>CD14<sup>+</sup>CD163<sup>+</sup> cells (orange) phenotypically aligning with blood DC3s (<xref rid="fig4" ref-type="fig">Figures 4</xref>A and 4B). CD1c<sup>+</sup>CD14<sup>−</sup> cells did not express CD163, suggesting that they mostly align with cDC2s. In addition, CD163 expression was restricted to CD1c<sup>+</sup>CD14<sup>+</sup> cells (<xref rid="fig4" ref-type="fig">Figure 4</xref>B). GM-CSF alone (MS5_GM), but not FLT3L, was sufficient to induce differentiation of DC3s <italic toggle="yes">in vitro</italic> (<xref rid="fig4" ref-type="fig">Figure 4</xref>C). In contrast, FLT3L (MS5_FL) was sufficient to induce differentiation of cDC2s (<xref rid="fig4" ref-type="fig">Figure 4</xref>C). We conclude that cDC2s and DC3s have distinct growth factor requirements.<fig id="fig4" position="float" orientation="portrait"><label>Figure 4</label><caption><p>DC3s Differentiate from Hematopoietic Progenitors upon GM-CSF Exposure Independent of Mono-Committed Progenitors (cMoPs) or cDC-Committed Progenitors (CDPs)</p><p>(A) Flow cytometry analysis of cord blood-derived CD34<sup>+</sup> HSPCs cultured on stromal cells expressing human FLT3L, SCF, and CXCL12 (MS5_FS12) with or without human recombinant GM-CSF (MS5_FS12+recGM-CSF) for 14 days.</p><p>(B) Flow cytometry plots of BTLA and CD163 expression within CD1c<sup>+</sup>CD14<sup>−</sup> and CD1c<sup>+</sup>CD14<sup>+</sup> cells identified in (A). Bar graphs summarize the absolute numbers of differentiated CD1c<sup>+</sup>CD14<sup>−</sup> cDC2s and CD1c<sup>+</sup>CD14<sup>+</sup> DC3s (a line represents the median; n = 6 independent cord blood donors, <sup>∗</sup>p &lt; 0.05, Wilcoxon test).</p><p>(C) Bar graphs summarizing the absolute numbers of CD1c<sup>+</sup>CD14<sup>−</sup>CD163<sup>−</sup> cDC2s and CD1c<sup>+</sup>CD14<sup>+</sup>CD163<sup>+</sup> DC3s differentiated from cord blood-derived CD34<sup>+</sup> HSPCs cocultured for 14 days with stromal cells expressing human FLT3L (MS5_FL), GM-CSF (MS5_GM), or neither (MS5_CTRL) (a line represents the median; n = 5–7 independent cord blood donors, <sup>∗</sup>p &lt; 0.05, <sup>∗∗</sup>p &lt; 0.01, <sup>∗∗∗</sup>p &lt; 0.001, Wilcoxon test).</p><p>(D) HC based on 19,791 protein-coding genes of <italic toggle="yes">in-vitro-</italic>generated subsets differentiated from CD34<sup>+</sup> HSPCs cultured with MS5_FS12 supplemented with human recombinant GM-CSF (MS5_FS12+recGM-CSF). Each dot represents an average of three donors.</p><p>(E) Volcano plots showing the DEGs between <italic toggle="yes">in-vitro-</italic>generated DC3s cells (orange) compared with cDC2s (blue and turquoise, left plot) or macrophages (gray and brown, right plot). Genes with Log<sub>2</sub>(fold change, FC) &gt; ±2 and a false discovery rate (FDR)-adjusted p value of less than 0.05 were considered significant.</p><p>(F) GSEA of pairwise comparisons of DC3s with cDC2s or macrophages generated i<italic toggle="yes">n vitro</italic>. Gene signatures (gene set) defining genes upregulated in blood DC3s compared with cDC2s (DC3 &gt; cDC2) or blood DC3s compared with CD14<sup>+</sup> monocytes (DC3 &gt; CD14 Mono) were used (<xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>) (NES, normalized enrichment score).</p><p>(G) BubbleMap summarizing the enrichment of defined gene sets in pairwise comparisons of <italic toggle="yes">in-vitro-</italic>differentiated DC3s versus <italic toggle="yes">in vitro</italic> cDC2s or <italic toggle="yes">in vitro</italic> macrophages. Gene signatures (gene sets) of blood DC3s compared with cDC2s (DC3 &gt; cDC2) or CD14<sup>+</sup> monocytes (DC3 &gt; Mono) and, vice versa, of blood cDC2s (cDC2 &gt; DC3) or CD14<sup>+</sup> monocytes (Mono &gt; DC3) compared with DC3s were used (<xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>).</p><p>(H) Flow cytometry analysis of cord blood-derived CDPs, cMoPs, and GMDPs cultured for 7 days with MS5_FS12 or MS5_GM. Bar graphs summarize the absolute number of differentiated cells from each progenitor (a line represents the median, n = 4–7 independent cord blood donors).</p><p>See also <xref rid="mmc1" ref-type="supplementary-material">Figure S5</xref> and <xref rid="mmc2" ref-type="supplementary-material">Table S1</xref>.</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="gr4.jpg"><?image-name gr4.jpg?><?image-size 496040?><?image-md5 29944b8bdf39ada7aeb94f336c9aaaa5?><?image-image-server-status LOAD_COMPLETED?><?image-original-height 4159?><?image-original-width 2762?><?image-scaled-height 1188?><?image-scaled-width 789?><?image-cloudpmc-urn urn:cdn:blobs/922d/7445430/29944b8bdf39/gr4.jpg?><?thumb-name gr4.gif?><?thumb-size 16055?><?thumb-md5 217d257773c18b9e0261a813a9ed3856?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 151?><?thumb-scaled-width 100?><?thumb-cloudpmc-urn urn:cdn:blobs/922d/7445430/217d257773c1/gr4.gif?></graphic></fig></p><p id="p0120">We next intended to determine whether the transcriptional landscape of <italic toggle="yes">in-vitro</italic>-generated DC3s from stromal cell cocultures aligned with their <italic toggle="yes">in vivo</italic> counterparts. Bulk RNA-seq analysis of <italic toggle="yes">in-vitro</italic>-generated cells showed that CD14<sup>+</sup>CD1c<sup>+</sup> DC3s (orange) sat between cDC2s (blue and turquoise) and macrophages (gray and brown) (<xref rid="fig4" ref-type="fig">Figure 4</xref>D). Of note, CD1c<sup>+</sup>CD14<sup>−</sup> cells generated <italic toggle="yes">in vitro</italic> were heterogenous for CD206 expression and were therefore analyzed as two independent subsets (CD1c<sup>+</sup>CD206<sup>−</sup> [blue] and CD1c<sup>+</sup>CD206<sup>+</sup> [turquoise]; <xref rid="mmc1" ref-type="supplementary-material">Figure S5</xref>B). However, both subsets displayed very similar transcriptomes regardless of CD206 expression, and the CD1c<sup>+</sup>CD14<sup>−</sup>CD206<sup>+</sup> and CD1c<sup>+</sup>CD14<sup>−</sup>CD206<sup>−</sup> fractions strongly resembled circulating cDC2s (<xref rid="bib2" ref-type="bibr">Anselmi et al., 2020</xref>).</p><p id="p0125">Further analysis of differentially expressed genes (DEGs) showed that <italic toggle="yes">in vitro</italic> DC3s differed from <italic toggle="yes">in vitro</italic> cDC2s by expression of monocyte-associated markers such as <italic toggle="yes">S100A8</italic>, <italic toggle="yes">S100A9</italic>, <italic toggle="yes">S100A12</italic>, and <italic toggle="yes">CD14</italic> (<xref rid="fig4" ref-type="fig">Figure 4</xref>E; <xref rid="mmc1" ref-type="supplementary-material">Figure S5</xref>C). Conversely, <italic toggle="yes">in vitro</italic> cDC2s displayed higher expression of <italic toggle="yes">CD1C</italic>, <italic toggle="yes">LAMP3</italic>, <italic toggle="yes">CD52</italic>, and <italic toggle="yes">LTB</italic>, as reported for lung cDC2s (<xref rid="bib25" ref-type="bibr">Lavin et al., 2017</xref>; <xref rid="fig4" ref-type="fig">Figure 4</xref>E; <xref rid="mmc1" ref-type="supplementary-material">Figure S5</xref>C). Using the GSEA methodology (<xref rid="bib56" ref-type="bibr">Subramanian et al., 2005</xref>), we found that the set of genes upregulated in primary DC3s compared with cDC2s (DC3 &gt; cDC2 from <xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>) was enriched in GM-CSF-dependent <italic toggle="yes">in vitro</italic> DC3s compared with <italic toggle="yes">in vitro</italic> cDC2s (<xref rid="fig4" ref-type="fig">Figure 4</xref>F). Of note, GM-CSF-exposed cDC2s did not convert to DC3s (<xref rid="mmc1" ref-type="supplementary-material">Figures S5</xref>D and S5E). We found that markers common for primary cDC2s and DC3s (<italic toggle="yes">CLEC10A</italic>, <xref rid="fig1" ref-type="fig">Figure 1</xref>) were higher in <italic toggle="yes">in vitro</italic> DC3s compared with <italic toggle="yes">in vitro</italic> macrophages (<xref rid="fig4" ref-type="fig">Figure 4</xref>E; <xref rid="bib19" ref-type="bibr">Heidkamp et al., 2016</xref>). Conversely, <italic toggle="yes">in vitro</italic> macrophages expressed more <italic toggle="yes">FCGR1A</italic>, <italic toggle="yes">C5AR1</italic> (CD88), <italic toggle="yes">CXCL8</italic>, <italic toggle="yes">CXCL1</italic>, <italic toggle="yes">CXCL2</italic>, <italic toggle="yes">CXCL3</italic>, <italic toggle="yes">CCL2</italic>, <italic toggle="yes">CCL3</italic>, and <italic toggle="yes">CCL7</italic> compared with <italic toggle="yes">in vitro</italic> DC3s (<xref rid="bib46" ref-type="bibr">Ruffell et al., 2009</xref>; <xref rid="fig4" ref-type="fig">Figure 4</xref>E; <xref rid="mmc1" ref-type="supplementary-material">Figures S5</xref>C and S5F). Genes upregulated in primary DC3s compared with CD14<sup>+</sup> monocytes (DC3 &gt; CD14<sup>+</sup> Mono from <xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>) were significantly enriched in <italic toggle="yes">in vitro</italic> DC3s compared with macrophages (<xref rid="fig4" ref-type="fig">Figures 4</xref>F and 4G). We conclude that GM-CSF drives, <italic toggle="yes">in vitro</italic>, the differentiation of DC3s aligning to primary blood DC3s.</p></sec><sec id="sec2.4"><title>DC3s Develop via a Differentiation Pathway Independent of CDPs and cMoPs</title><p id="p0130">The developmental relationship between DC3s and cDCs or monocyte lineages is not known. A classical view defines the development of phagocytes as a stepwise and ordered loss of developmental potential concomitant with lineage commitment. Historically, this process has been identified using prospective isolation of progenitor populations of decreasing potential. Specifically, early granulocyte-monocyte and DC progenitors (GMDPs) carries a tri-lineage potential (<xref rid="bib28" ref-type="bibr">Lee et al., 2015b</xref>). Loss of neutrophil potential defines monocyte and DC progenitors (MDPs) (<xref rid="bib28" ref-type="bibr">Lee et al., 2015b</xref>). Loss of monocyte potential defines CDPs, which generate cDCs via a pre-cDC intermediate (<xref rid="bib7" ref-type="bibr">Breton et al., 2015</xref>, <xref rid="bib8" ref-type="bibr">2016</xref>; <xref rid="bib28" ref-type="bibr">Lee et al., 2015b</xref>; <xref rid="bib38" ref-type="bibr">Naik et al., 2007</xref>; <xref rid="bib41" ref-type="bibr">Onai et al., 2007</xref>; <xref rid="bib51" ref-type="bibr">See et al., 2017</xref>). Finally, loss of DC and neutrophil potential defines monocyte-committed progenitors (cMoPs) (<xref rid="bib23" ref-type="bibr">Kawamura et al., 2017</xref>). Having established the growth factor requirements for DC3 development, we decided to test the contribution of CDPs and cMoPs to generation of DC3s. Flow cytometry-sorted CDPs and cMoPs and the remaining GMDP-containing fraction (CD34<sup>+</sup>CD38<sup>+</sup>CD45RA<sup>+</sup>CD123<sup>−</sup>CD64<sup>−</sup>) isolated from cord blood-derived HSPCs (see the gating strategy in <xref rid="mmc1" ref-type="supplementary-material">Figure S5</xref>G) were cocultured with stromal cells supporting cDC (MS5_FS12) or DC3 (MS5_GM) differentiation.</p><p id="p0135">Flow cytometry analysis after 7 days of culture showed that CDP gave rise exclusively to cDC1s and cDC2s in MS5_FS12 cocultures, as described previously (<xref rid="fig4" ref-type="fig">Figure 4</xref>H; <xref rid="mmc1" ref-type="supplementary-material">Figure S5</xref>H; <xref rid="bib28" ref-type="bibr">Lee et al., 2015b</xref>). cMoP cultures gave rise solely to CD14<sup>+</sup>CD1c<sup>−</sup> cells, as described previously (<xref rid="fig4" ref-type="fig">Figure 4</xref>H; <xref rid="mmc1" ref-type="supplementary-material">Figure S5</xref>H; <xref rid="bib23" ref-type="bibr">Kawamura et al., 2017</xref>). As expected, the GMDP-containing fraction gave rise to granulocytes, macrophages, and cDC2s. Importantly, the GMDP-containing fraction also gave rise to DC3s in MS5_GM coculture (<xref rid="fig4" ref-type="fig">Figure 4</xref>H; <xref rid="mmc1" ref-type="supplementary-material">Figure S5</xref>H). Therefore, we asked whether DC3s would arise directly from a multipotent progenitor or via formation of an intermediate DC3-committed progenitor devoid of any other lineage potential.</p><p id="p0140">To address this question in unbiased settings, we developed a single cell culture of CD34<sup>+</sup>CD38<sup>+</sup>CD123<sup>−</sup>CD64<sup>−</sup> progenitors distinct from CDPs or cMoPs (<xref rid="fig5" ref-type="fig">Figure 5</xref>A). We chose to combine MS5_FS12 with soluble GM-CSF for two reasons: (1) GM-CSF alone did not support growth of individual progenitors (<xref rid="mmc1" ref-type="supplementary-material">Figure S5</xref>I), and (2) MS5_FS12 coculture was found to more efficiently support cDC and monocyte differentiation (<xref rid="fig4" ref-type="fig">Figure 4</xref>G).<fig id="fig5" position="float" orientation="portrait"><label>Figure 5</label><caption><p>Single-Cell Analysis of DC3 Commitment</p><p>(A) Flow cytometry analysis of bulk (500 cells) or single CD34<sup>+</sup>CD38<sup>+</sup>CD123<sup>−</sup>CD64<sup>−</sup> progenitor cells cocultured for 14 days with MS5_FS12 supplemented with recombinant human GM-CSF (MS5_FS12+recGM-CSF). Flow cytometry plots resulting from single CD34<sup>+</sup>CD38<sup>+</sup>CD123<sup>−</sup>CD64<sup>−</sup> progenitor cells with different potentials are shown as representative examples (n = 355 cells from 2 independent experiments).</p><p>(B) HC of lineage potential from single CD34<sup>+</sup>CD38<sup>+</sup>CD123<sup>−</sup>CD64<sup>−</sup> progenitor cells (n = 355).</p><p>(C) Bar graph and Venn diagram summarizing the frequency of the potential of mono-, bi-, tri-, or multipotent individual CD34<sup>+</sup>CD38<sup>+</sup>CD123<sup>−</sup>CD64<sup>−</sup> cells within the total wells analyzed (n = 355).</p><p>(D) Bar graphs summarizing the frequency of mono-, bi-, tri-, or multipotent individual CD34<sup>+</sup>CD38<sup>+</sup>CD123<sup>−</sup>CD64<sup>−</sup> cells among DC3-generating progenitors only. An orange bar represents the frequency of DC3-restricted progenitors.</p><p>(E) Cell surface phenotype of DC3-restricted progenitors before differentiation cultures inferred by index flow cytometry sorting. tSNE plots display an overlay of total CD45<sup>+</sup> cells (gray) and DC3-restricted progenitor cells (orange) (top left). Shown is relative expression of the markers CD45RA, CD38, CD34, CD10, Clec12A, CD64, CD123, CD163, and SIRPα.</p><p>(F) Validation experiment for identification of Clec12A as a marker for DC3-committed progenitors. Shown is flow cytometry analysis of bulk-sorted CD34<sup>+</sup>CD38<sup>+</sup>CD45RA<sup>+</sup>CD123<sup>−</sup>CD64<sup>−</sup>Clec12A<sup>−</sup> and Clec12A<sup>+</sup> cells. 500 cells were cocultured with MS5_GM for 7 days. The bar graph summarizes the number of differentiated DC3s from each bulk population (n = 4 healthy donors, <sup>∗</sup>p &lt; 0.05, Mann-Whitney two-tailed t test).</p><p>See also <xref rid="mmc1" ref-type="supplementary-material">Figure S5</xref>.</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="gr5.jpg"><?image-name gr5.jpg?><?image-size 409829?><?image-md5 d1c4eaf19ffb26ee677b368e50d6efb2?><?image-image-server-status LOAD_COMPLETED?><?image-original-height 3413?><?image-original-width 2906?><?image-scaled-height 853?><?image-scaled-width 726?><?image-cloudpmc-urn urn:cdn:blobs/922d/7445430/d1c4eaf19ffb/gr5.jpg?><?thumb-name gr5.gif?><?thumb-size 17790?><?thumb-md5 2889190bfa83e623c1e29df80d3e08a5?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 117?><?thumb-scaled-width 100?><?thumb-cloudpmc-urn urn:cdn:blobs/922d/7445430/2889190bfa83/gr5.gif?></graphic></fig></p><p id="p0145">Flow cytometry analysis of 14-day progeny of individual progenitor cultures revealed multiple patterns of developmental potential (<xref rid="fig5" ref-type="fig">Figure 5</xref>A). Overall, the granulocyte potential was segregated from the potential for mononuclear phagocytes (<xref rid="fig5" ref-type="fig">Figure 5</xref>B). We found that only 0.3% of progenitors could differentiate into all four lineages corresponding to the GMDP functional definition (<xref rid="fig5" ref-type="fig">Figure 5</xref>D). Most individual progenitors gave rise to single lineage progeny (<xref rid="fig5" ref-type="fig">Figures 5</xref>C and 5D). Unipotent wells containing only neutrophils represented the most abundant outcome (39.4%), followed by macrophage-only (26.8%) and cDC-only (5.63%) wells (<xref rid="fig5" ref-type="fig">Figure 5</xref>C). Most individual progenitors endowed with DC3 potential had multi-lineage potential. DC3 potential was more associated with mononuclear phagocytes rather than granulocyte potential (61.4% and 19.3% of DC3-containing wells, respectively) (<xref rid="fig5" ref-type="fig">Figure 5</xref>D). Importantly, we also identified a minor fraction of individual progenitors giving rise exclusively to DC3s (<xref rid="fig5" ref-type="fig">Figures 5</xref>B–5D).</p><p id="p0150">We next aimed to further define the cell surface phenotype of progenitors endowed with DC3 potential. To this end, we attempted to establish correlations between the cell surface phenotype of single sorted cells (inferred from index flow cytometry sorting) and their developmental potential. <italic toggle="yes">A posteriori</italic> identification of the cell surface phenotype of DC3-committed progenitors revealed that they had a CD34<sup>+</sup>CD38<sup>+</sup>CD45RA<sup>int</sup>CD123<sup>−</sup>CD64<sup>−</sup>SIRPα<sup>lo</sup>CD10<sup>−</sup>Clec12A<sup>+</sup> phenotype (<xref rid="fig5" ref-type="fig">Figure 5</xref>E). As validation, we showed that the potential for DC3s lay in the Clec12A<sup>+</sup> population of the GMDP-containing fraction (<xref rid="fig5" ref-type="fig">Figure 5</xref>F). We conclude that DC3s can develop via a DC3-restricted intermediate distinct from cDC-restricted CDPs or monocyte-restricted cMoPs. Even though the existence of a DC3-commited unipotent progenitor is not formally proven, our data are compatible with the notion that DC3 specification arises downstream of MDP.</p></sec><sec id="sec2.5"><title>Activated DC3s Induce Priming of Naive T Cells and Differentiation of CD103<sup>+</sup> T Cells</title><p id="p0155">We next aimed to understand the immunological function of DC3s and to compare it with cDC2s and monocytes. First, we decided to test the responsiveness of DC3s, cDC2s, and monocytes to a cocktail of Toll-like receptor (TLR) agonists. cDC2s, DC3s, and monocytes were sorted by flow cytometry from blood and stimulated overnight <italic toggle="yes">ex vivo</italic>. PCA analysis of the total transcriptome of unstimulated and stimulated populations evidenced that all subsets underwent a certain degree of convergence in their transcriptome (<xref rid="fig6" ref-type="fig">Figure 6</xref>A). In support of this, we found an important overlap in the set of activation-induced genes defined for each subset (1,344 genes; <xref rid="fig6" ref-type="fig">Figure 6</xref>B). Despite the relative convergence of activated cells, we found that overnight activation did not compromise cell surface discrimination of DC3s from cDC2s and monocytes (<xref rid="mmc1" ref-type="supplementary-material">Figure S6</xref>A). Indeed, TLR-activated DC3s could still be discriminated from TLR-activated cDC2s by 437 DEGs or from TLR-activated monocytes by 1,293 genes (<xref rid="fig6" ref-type="fig">Figure 6</xref>C). The same was true for the pairwise comparison of activated cDC2s and circulating DC3s (<xref rid="mmc1" ref-type="supplementary-material">Figure S6</xref>B). In sum, we conclude that innate activation does not trigger conversion of cDC2s or monocytes into DC3s despite induction of a common transcriptional response to TLR stimulation.<fig id="fig6" position="float" orientation="portrait"><label>Figure 6</label><caption><p>DC3s Respond to TLR Stimulation</p><p>We performed bulk RNA-seq analysis of BTLA<sup>+</sup>CD5<sup>+</sup> and BTLA<sup>+</sup>CD5<sup>−</sup> cDC2s, DC3s, and monocytes sorted as shown in <xref rid="fig1" ref-type="fig">Figure 1</xref>H and stimulated overnight (16 h, 3 donors) or not (4 donors) with a TLR agonist cocktail (25 μg/mL poly(I:C), 1 μg/mL R848, and 10 ng/mL LPS). For activation of cDC2s, BTLA<sup>+</sup>CD5<sup>+</sup> and BTLA<sup>+</sup>CD5<sup>−</sup> were pooled.</p><p>(A) PCA analysis for all genes.</p><p>(B) Venn diagram summarizing the number of activation-induced DEGs upregulated in stimulated compared with unstimulated cells within each cell population.</p><p>(C) Volcano plots showing DEGs between TLR agonist-stimulated DC3s compared with TLR agonist-stimulated cDC2s or TLR agonist-stimulated monocytes. Genes with Log<sub>2</sub>(FC) &gt; ±2 and a FDR-adjusted p value of less than 0.05 were considered significant.</p><p>(D) Bar graph summarizing relative CCR7 gene expression within TLR-agonist stimulated or unstimulated mononuclear phagocyte populations (n = 3–4; a line represents the median; <sup>∗</sup>p &lt; 0.05, <sup>∗∗</sup>p &lt; 0.01, <sup>∗∗∗</sup>p &lt; 0.001, <sup>∗∗∗∗</sup>p &lt; 0.0001, one-way ANOVA test).</p><p>(E) Histograms showing the frequency and median of fluorescence intensity (MFI) of CD86 on TLR agonist-stimulated or unstimulated mononuclear phagocyte populations.</p><p>(F) Heatmap showing the relative gene expression of selected costimulatory molecules on TLR agonist-stimulated or unstimulated mononuclear phagocyte populations.</p><p>(G) Quantification of cytokines and chemokines secreted by cDC2s, DC3s, and CD14<sup>+</sup> monocytes in response to overnight stimulation with a cocktail of TLR agonists (n = 9 healthy donors; a line represents the median; <sup>∗</sup>p &lt; 0.05, <sup>∗∗</sup>p &lt; 0,01, <sup>∗∗∗∗</sup>p &lt; 0,0001, one-way ANOVA test).</p><p>(H) Heatmap showing the relative gene expression of cytokines and chemokines analyzed in (G) within TLR-agonist stimulated or unstimulated mononuclear phagocyte populations.</p><p>See also <xref rid="mmc1" ref-type="supplementary-material">Figure S6</xref> and <xref rid="mmc2" ref-type="supplementary-material">Table S1</xref>.</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="gr6.jpg"><?image-name gr6.jpg?><?image-size 366190?><?image-md5 713e3e52a83133bcbc6c7fade839f041?><?image-image-server-status LOAD_COMPLETED?><?image-original-height 3768?><?image-original-width 3321?><?image-scaled-height 837?><?image-scaled-width 738?><?image-cloudpmc-urn urn:cdn:blobs/922d/7445430/713e3e52a831/gr6.jpg?><?thumb-name gr6.gif?><?thumb-size 16994?><?thumb-md5 a28708d44e76cd4b000fc12db406c583?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 113?><?thumb-scaled-width 100?><?thumb-cloudpmc-urn urn:cdn:blobs/922d/7445430/a28708d44e76/gr6.gif?></graphic></fig></p><p id="p0160">From the perspective of adaptive immunity, activated DC3s shared a lot of common features with activated cDC2s but less with activated monocytes: (1) stimulated DC3s and cDC2s upregulated <italic toggle="yes">CCR7</italic> upon activation (<xref rid="fig6" ref-type="fig">Figure 6</xref>D), potentially enabling their ability to migrate toward T cell zones; (2) activated DC3s and cDC2s upregulated cell surface co-stimulatory molecules (<italic toggle="yes">CD80</italic>, <italic toggle="yes">CD86</italic>, <italic toggle="yes">CD70</italic>, and <italic toggle="yes">CD40</italic>; <xref rid="fig6" ref-type="fig">Figure 6</xref>E and 6F); (3) activated DC3s and cDC2s efficiently increased the expression of T cell-attracting chemokines such as CCL5 (<xref rid="fig6" ref-type="fig">Figures 6</xref>G and 6H), <italic toggle="yes">CCL19</italic>, <italic toggle="yes">CCL17</italic>, <italic toggle="yes">CCL22</italic>, <italic toggle="yes">CXCL9</italic>, <italic toggle="yes">CXCL10</italic>, <italic toggle="yes">CXCL11</italic>, and <italic toggle="yes">CXCL13</italic> (<xref rid="mmc1" ref-type="supplementary-material">Figure S6</xref>C); and (4) activated DC3s and cDC2s produced higher amounts of IL-12p70, IL-23, <italic toggle="yes">IL27</italic>, and IL-10 (<xref rid="fig6" ref-type="fig">Figures 6</xref>G and 6H; <xref rid="mmc1" ref-type="supplementary-material">Figure S6</xref>E).</p><p id="p0165">In addition, activated DC3s shared some common features with activated monocytes that are less pronounced in activated cDC2s: (i) activated DC3s and monocytes secreted more inflammatory cytokines, such as tumor necrosis factor alpha (TNF-α) and IL-1β (<xref rid="fig6" ref-type="fig">Figures 6</xref>G and 6H); and (2) activated DC3s and activated monocytes upregulated inflammatory chemokines, such as CCL2 (<xref rid="fig6" ref-type="fig">Figures 6</xref>G and 6H), <italic toggle="yes">CCL1</italic>, and <italic toggle="yes">CCL3</italic> (<xref rid="mmc1" ref-type="supplementary-material">Figure S6</xref>C) or granulocytes attracting CXCL1, <italic toggle="yes">CXCL3</italic>, and <italic toggle="yes">CXCL5</italic> (<xref rid="mmc1" ref-type="supplementary-material">Figures S6</xref>C and S6D).</p><p id="p0170">Collectively, our data suggest that the transcriptome and secretome of activated DC3s, unlike the ones of activated monocytes, are consistent with a function in priming of naive T cells.</p><p id="p0175">To directly test the T cell priming capabilities of DC3s, we performed 5-day cocultures of flow cytometry-sorted, activated DC3s, cDC2s, and monocytes (see the cell sorting strategy in <xref rid="mmc1" ref-type="supplementary-material">Figure S7</xref>A) with allogenic CD45RA<sup>+</sup> naive T cells in the presence of a synthetic superantigen. We found that activated DC3s and cDC2s, unlike monocytes, triggered proliferative expansion and effector differentiation in CD4<sup>+</sup> and CD8<sup>+</sup> T cells probed by CD45RO acquisition (<xref rid="mmc1" ref-type="supplementary-material">Figure S7</xref>B). The same results were also obtained using <italic toggle="yes">in-vitro-</italic>generated DC3s (<xref rid="mmc1" ref-type="supplementary-material">Figure S7</xref>C). In contrast to monocytes, activated DC3s and cDC2s induced interferon γ (IFN-γ)- and TNF-α-producing CD4<sup>+</sup> and CD8<sup>+</sup> T cells but not IL- 17A (<xref rid="fig7" ref-type="fig">Figure 7</xref>A; <xref rid="mmc1" ref-type="supplementary-material">Figure S7</xref>D). We found that DC3s had a specific ability to efficiently trigger CD103 expression in CD8<sup>+</sup> T cells (<xref rid="fig7" ref-type="fig">Figure 7</xref>B), even without the presence of superantigen (<xref rid="mmc1" ref-type="supplementary-material">Figure S7</xref>E). CD103 expression is a hallmark of tissue-resident memory T (T<sub>RM</sub>) cells because of its interaction with E-cadherin (<xref rid="bib36" ref-type="bibr">Mueller and Mackay, 2016</xref>). Mechanistically, multiple factors, including transforming growth factor β (TGF-β), have been proposed to induce CD103 expression on T cells (<xref rid="bib36" ref-type="bibr">Mueller and Mackay, 2016</xref>; <xref rid="bib45" ref-type="bibr">Rihs et al., 1996</xref>; <xref rid="bib63" ref-type="bibr">Yu et al., 2013</xref>). Here we showed that anti-TGF-β neutralizing antibodies, but not others, blocked the expression of CD103 on CD8<sup>+</sup> T cells (<xref rid="fig7" ref-type="fig">Figure 7</xref>C; <xref rid="mmc1" ref-type="supplementary-material">Figure S7</xref>F).<fig id="fig7" position="float" orientation="portrait"><label>Figure 7</label><caption><p>DC3s Prime Naive T Cells and Drive Acquisition of the CD103<sup>+</sup> T<sub>RM</sub> Phenotype</p><p>(A and B) Representative flow cytometry plots and quantification of CD4<sup>+</sup> and CD8<sup>+</sup> naive T cells cultured for 5 days with flow cytometry-sorted blood cDC2s, DC3s, or CD14<sup>+</sup> monocytes after overnight activation with TLR agonists (25 μg/mL poly(I:C), 1 μg/mL R848, and 10 ng/mL LPS) in the presence of a synthetic superantigen (Cytostim). Absolute numbers and frequencies of cytokine-producing and other activated T cells (A) and CD103<sup>+</sup> T cells (B) are shown (n = 5 healthy donors in 5 independent experiments; a line represents the median; <sup>∗</sup>p &lt; 0.05, <sup>∗∗</sup>p &lt; 0.01, <sup>∗∗∗</sup>p &lt; 0.001, <sup>∗∗∗∗</sup>p &lt; 0.0001, one-way ANOVA test).</p><p>(C) Representative flow cytometry plots and quantification showing CD103 expression on CD8<sup>+</sup> naive T cells cocultured with blood DC3s sorted by fluorescence-activated cell sorting (FACS) in the presence of 10 μg/mL of neutralizing antibodies against TNF-α or TGF-β or an isotype CTRL (n = 4 healthy donors in 3 independent experiments; a line represents the median; <sup>∗</sup>p &lt; 0.05, <sup>∗∗</sup>p &lt; 0.01, <sup>∗∗∗</sup>p &lt; 0.001, <sup>∗∗∗∗</sup>p &lt; 0.0001, one-way paired ANOVA test).</p><p>(D and E) Bulk RNA-seq analysis of CD8<sup>+</sup>CD103<sup>−</sup> T cells (n = 3) and CD8<sup>+</sup>CD103<sup>+</sup> T cells (n = 3) sorted by flow cytometry after 5 days <italic toggle="yes">in vitro</italic> coculture of naive blood CD8<sup>+</sup> T cells with blood DC3s activated overnight by TLR agonists.</p><p>(D) GSEA of pairwise comparisons of CD8<sup>+</sup>CD103<sup>+</sup> T cells with CD8<sup>+</sup>CD103<sup>−</sup> T cells. Gene signatures (gene set) defining genes upregulated in breast or lung CD103<sup>+</sup>CD69<sup>+</sup>CD8<sup>+</sup> T<sub>RM</sub> cells were used (<xref rid="bib21" ref-type="bibr">Hombrink et al., 2016</xref>; <xref rid="bib24" ref-type="bibr">Kumar et al., 2017</xref>; <xref rid="bib49" ref-type="bibr">Savas et al., 2018</xref>).</p><p>(E) Heatmap displaying 56 representative genes significantly upregulated in CD8<sup>+</sup>CD103<sup>+</sup> cells compared with CD8<sup>+</sup>CD103<sup>−</sup> induced by blood DC3s <italic toggle="yes">in vitro</italic> (of 205 DEGs). Selected genes are shared with at least one of the previously reported gene signatures defining human T<sub>RM</sub> cells (<xref rid="bib21" ref-type="bibr">Hombrink et al., 2016</xref>; <xref rid="bib24" ref-type="bibr">Kumar et al., 2017</xref>; <xref rid="bib49" ref-type="bibr">Savas et al., 2018</xref>).</p><p>(F–H) Correlative analysis of T<sub>RM</sub> cell content in luminal breast cancer primary tumors.</p><p>(F) Representative flow cytometry plots showing the gating strategy for CD103<sup>+</sup>CD69<sup>+</sup>CD8<sup>+</sup> T cells in 21 human luminal breast cancer primary tumors.</p><p>(G) Quantification of CD103<sup>+</sup>CD69<sup>+</sup>CD8<sup>+</sup> T cells in different stages of human breast tumors (stage I, n = 3; stage II, n = 13; stage III, n = 5).</p><p>(H) Pearson correlation of the frequencies of the macrophages and cDC1, CD1c<sup>+</sup> CD14<sup>−</sup>, and CD1c<sup>+</sup> CD14<sup>−</sup> cells and the frequencies of CD103<sup>+</sup>CD69<sup>+</sup>CD8<sup>+</sup> T cells in human breast cancer primary tumors (red, significantly correlated p &lt; 0.05; black, not correlated).</p><p>See also <xref rid="mmc1" ref-type="supplementary-material">Figure S7</xref> and <xref rid="mmc2" ref-type="supplementary-material">Tables S1</xref> and <xref rid="mmc1" ref-type="supplementary-material">S2</xref>.</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="gr7.jpg"><?image-name gr7.jpg?><?image-size 456320?><?image-md5 2a37155aa9b79a3a73b5cae020fb7f8d?><?image-image-server-status LOAD_COMPLETED?><?image-original-height 4159?><?image-original-width 2655?><?image-scaled-height 1187?><?image-scaled-width 758?><?image-cloudpmc-urn urn:cdn:blobs/922d/7445430/2a37155aa9b7/gr7.jpg?><?thumb-name gr7.gif?><?thumb-size 17038?><?thumb-md5 e5f245e636521187c641ee4687b86777?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 157?><?thumb-scaled-width 100?><?thumb-cloudpmc-urn urn:cdn:blobs/922d/7445430/e5f245e63652/gr7.gif?></graphic></fig></p><p id="p0180">We next wondered whether CD103<sup>+</sup> T cells aligned with <italic toggle="yes">bona fide</italic> T<sub>RM</sub> cells isolated from human tissue. To this end, we evaluated the gene expression profile of CD103<sup>−</sup> and CD103<sup>+</sup> CD8<sup>+</sup> T cells obtained after coculture of naive T cells with activated DC3s (<xref rid="mmc1" ref-type="supplementary-material">Figure S7</xref>G). Using the GSEA methodology (<xref rid="bib56" ref-type="bibr">Subramanian et al., 2005</xref>), we found that the signatures obtained for breast cancer or lung tissue CD103<sup>+</sup> T<sub>RM</sub> cells (<xref rid="bib21" ref-type="bibr">Hombrink et al., 2016</xref>; <xref rid="bib24" ref-type="bibr">Kumar et al., 2017</xref>; <xref rid="bib49" ref-type="bibr">Savas et al., 2018</xref>) were enriched in CD103<sup>+</sup> T cells compared with CD103<sup>−</sup> T cells induced by DC3s (<xref rid="fig7" ref-type="fig">Figure 7</xref>D). DEG analysis revealed that human T<sub>RM</sub> cell markers such as <italic toggle="yes">NUSAP1</italic>, <italic toggle="yes">DUSP4</italic>, <italic toggle="yes">CXCR6</italic>, and <italic toggle="yes">FASLG</italic> (<xref rid="fig7" ref-type="fig">Figure 7</xref>E; <xref rid="bib21" ref-type="bibr">Hombrink et al., 2016</xref>; <xref rid="bib24" ref-type="bibr">Kumar et al., 2017</xref>; <xref rid="bib49" ref-type="bibr">Savas et al., 2018</xref>) were upregulated in CD103<sup>+</sup> compared with CD103<sup>−</sup> CD8<sup>+</sup> T cells. In addition, DC3-activated CD103<sup>+</sup> CD8<sup>+</sup> T cells expressed core components of the cytotoxic machinery (<italic toggle="yes">PRF1</italic> and <italic toggle="yes">GZMA</italic>), as reported earlier for breast cancer-invading T<sub>RM</sub> cells (<xref rid="fig7" ref-type="fig">Figure 7</xref>E; <xref rid="bib49" ref-type="bibr">Savas et al., 2018</xref>).</p><p id="p0185">CD8<sup>+</sup>CD103<sup>+</sup> T<sub>RM</sub> cell infiltration has a protective prognosis value in breast cancer (<xref rid="bib49" ref-type="bibr">Savas et al., 2018</xref>; <xref rid="bib59" ref-type="bibr">Wang et al., 2016</xref>). To test the physiological relevance of DC3-dependent induction of CD8<sup>+</sup>CD103<sup>+</sup> T cells, we analyzed CD103 and CD69 expression in CD3<sup>+</sup>CD8<sup>+</sup> T cells from 18 samples of primary luminal breast cancer (<xref rid="fig7" ref-type="fig">Figure 7</xref>F). We found that CD8<sup>+</sup>CD103<sup>+</sup>CD69<sup>+</sup> T cells were present across different cancer stages (<xref rid="fig7" ref-type="fig">Figure 7</xref>G). The frequency of CD8<sup>+</sup>CD103<sup>+</sup>CD69<sup>+</sup> T cells was positively correlated with DC3 infiltration but not with other mononuclear phagocytes (<xref rid="fig7" ref-type="fig">Figure 7</xref>H). Despite the statistical significance of this correlation, it is important to underline that this was mainly driven by a subset of the samples. This suggests that further stratification of patients could help improve our understanding of the relationship between T<sub>RM</sub> cells and DC3s in breast cancer.</p><p id="p0190">Together, our results anchor DC3 function within the DC lineage; DC3s, just like cDC2s but unlike monocytes, are competent for priming and polarization of CD45RA<sup>+</sup> naive T cells. In addition, we define induction of <italic toggle="yes">bona fide</italic> T<sub>RM</sub> cell-like CD103<sup>+</sup>CD8<sup>+</sup> T cells (<xref rid="bib63" ref-type="bibr">Yu et al., 2013</xref>) as a specific privilege of DC3s but not cDC2s.</p></sec></sec><sec id="sec3"><title>Discussion</title><p id="p0195">Using scRNA-seq and high-dimensional flow cytometry, we provided evidence that DC3s represent a DC subset that can be separated and isolated from other DC subtypes. Regarding gene expression, our results are largely congruent with findings reported in previous studies (<xref rid="bib13" ref-type="bibr">Dutertre et al., 2019</xref>; <xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>). In addition, we provide a robust flow cytometry strategy to identify and purify DC3s by taking in account CD14<sup>+</sup> monocytes and cDC2s. In agreement with previous studies, our results challenge the widely accepted notion that CD14 is a specific marker for monocytes (<xref rid="bib13" ref-type="bibr">Dutertre et al., 2019</xref>; <xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>). Indeed, we clearly demonstrate that a large fraction of DC3s expresses cell surface CD14. We identified CD88 (encoded by the <italic toggle="yes">C5AR1</italic> gene) as a proper monocyte marker, enabling prospective purification of monocytes devoid of CD14<sup>+</sup>CD1c<sup>+</sup> DC3s. We found that <italic toggle="yes">ex vivo</italic> GM-CSF cultures of pure CD88<sup>+</sup>CD14<sup>+</sup> monocytes did not give rise to CD88<sup>-</sup>CD14<sup>+</sup>CD1c<sup>+</sup> iDCs. Therefore, our findings provide an incentive to carefully revisit the prevailing notion that CD14<sup>+</sup>CD1c<sup>+</sup> iDCs arise exclusively from monocytes <italic toggle="yes">in vivo</italic>.</p><p id="p0200">From the DC perspective, we refined the strategy to analyze the functional heterogeneity of CD1c<sup>+</sup> DCs, including cDC2s and DC3s. Indeed, previous studies have reported heterogeneous expression of CD5 in CD1c<sup>+</sup> DCs (<xref rid="bib13" ref-type="bibr">Dutertre et al., 2019</xref>; <xref rid="bib62" ref-type="bibr">Yin et al., 2017</xref>). However, our findings highlight that CD5<sup>−</sup>CD1c<sup>+</sup> DCs contain CD163<sup>−</sup> cDC2s and CD163<sup>+</sup> DC3s, which are transcriptionally distinct. Overall, our scRNA-seq and bulk RNA-seq analyses cluster together CD5<sup>+</sup> and CD5<sup>−</sup> cDC2s. However, the developmental relationship between CD5<sup>−</sup> cDC2s, CD5<sup>+</sup> cDC2s, and AS-DCs remains to be clarified (<xref rid="bib13" ref-type="bibr">Dutertre et al., 2019</xref>; <xref rid="bib51" ref-type="bibr">See et al., 2017</xref>; <xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>).</p><p id="p0205">Because of CD1c expression and their relative similarity to cDC2s, DC3s have been embedded in the group of cDCs. Here we have provided substantial evidence arguing against this notion. Indeed, a conservative definition of cDCs includes (1) dependence on the FLT3L growth factor (<xref rid="bib7" ref-type="bibr">Breton et al., 2015</xref>; <xref rid="bib17" ref-type="bibr">Guermonprez et al., 2013</xref>, <xref rid="bib18" ref-type="bibr">2019</xref>; <xref rid="bib32" ref-type="bibr">McKenna et al., 2000</xref>; <xref rid="bib44" ref-type="bibr">Pulendran et al., 1998</xref>; <xref rid="bib60" ref-type="bibr">Waskow et al., 2008</xref>) and (2) reliance on the CDP and pre-cDC developmental pathway (<xref rid="bib7" ref-type="bibr">Breton et al., 2015</xref>, <xref rid="bib8" ref-type="bibr">2016</xref>; <xref rid="bib18" ref-type="bibr">Guermonprez et al., 2019</xref>; <xref rid="bib28" ref-type="bibr">Lee et al., 2015b</xref>; <xref rid="bib38" ref-type="bibr">Naik et al., 2007</xref>; <xref rid="bib41" ref-type="bibr">Onai et al., 2007</xref>). Here we have shown that DC3s do not meet any of these criteria. First, FLT3L alone was poorly active in stimulating the production of DC3s from CD34<sup>+</sup> HSPCs in a controlled setting <italic toggle="yes">in vitro</italic>. Conversely, GM-CSF drives the commitment of DC3s under the same conditions. Moreover, we have shown that CDP did not give rise to DC3s although they are competent to generate cDC1s and cDC2s (<xref rid="bib28" ref-type="bibr">Lee et al., 2015b</xref>). Taking in account the heterogeneity of defined progenitor populations (<xref rid="bib43" ref-type="bibr">Paul et al., 2015</xref>), we developed single-cell cultures enabling analysis of multiple lineage populations. Our single-cell cultures demonstrated that DC3s develop from CLEC12A<sup>+</sup> DC3-restricted progenitors. In addition, the most frequent progenies differentiating along with DC3s from a bi-potent progenitor were cDCs and macrophages. This strongly suggest that DC3 progenitors diverge downstream of the MDP stage (<xref rid="bib14" ref-type="bibr">Fogg et al., 2006</xref>; <xref rid="bib27" ref-type="bibr">Lee et al., 2015a</xref>). Further studies will define more precisely the cell surface and molecular phenotype of the DC3-restricted progenitor. In support of a distinct regulation of cDC2s and DC3s, we have shown that cDC2 infiltration in breast cancer is correlated with cDC1s but not DC3s. Also, <xref rid="bib13" ref-type="bibr">Dutertre et al. (2019)</xref> have shown recently that DC3s, but not cDC2s, expand in the blood of systemic lupus erythematosus patients (<xref rid="bib13" ref-type="bibr">Dutertre et al., 2019</xref>). Further delineation of the inflammatory cues and transcription factors underpinning the development of the DC3 lineage is needed.</p><p id="p0210">Identification of DC3 as a cellular entity arising from a specific lineage brings forward the question of their specific immune function. We have shown that activated DC3s, just as cDC2s but unlike monocytes, secrete high amounts of T cell-polarizing cytokines (IL-12p70 and IL-23) and T cell-attracting chemokines (CXCL9, CXCL10, CXCL11, and CCL5). In addition, DC3s also secrete other cytokines (IL-10, IL-6, and TNF-α), some of which were poorly secreted by cDC2s (e.g., TNF-α). DC3s are polyvalent phagocytes with a cytokine pattern encompassing T cell and inflammation cues.</p><p id="p0215">A quintessential defining feature of DCs is their capacity to activate naive T cells. For instance, infiltration of CD1c<sup>+</sup> DCs is associated with priming of T cell effectors when regulatory T (Treg) cell infiltration is low or Treg cell-mediated suppression is alleviated by checkpoint blockade (<xref rid="bib6" ref-type="bibr">Binnewies et al., 2019</xref>). However, even when the heterogeneity of CD1c<sup>+</sup> cells is appreciated, including recruitment of CD14<sup>+</sup>CD1c<sup>+</sup> DCs in tumor-draining lymph nodes (<xref rid="bib6" ref-type="bibr">Binnewies et al., 2019</xref>), little is known about the function of CD1c<sup>+</sup> subtypes. Therefore, we asked whether the functions attributed to CD1c<sup>+</sup> DCs would be carried out by cDC2s and/or DC3s. We found that DC3s, like cDC2s but unlike monocytes, primed and drove robust activation of naive T cells into IFN-γ- and TNF-α-secreting polyfunctional effectors (<xref rid="bib1" ref-type="bibr">Acosta-Rodriguez et al., 2007</xref>; <xref rid="bib26" ref-type="bibr">Leal Rojas et al., 2017</xref>; <xref rid="bib39" ref-type="bibr">Napolitani et al., 2005</xref>; <xref rid="bib40" ref-type="bibr">Nizzoli et al., 2013</xref>; <xref rid="bib50" ref-type="bibr">Schlitzer et al., 2013</xref>; <xref rid="bib62" ref-type="bibr">Yin et al., 2017</xref>).</p><p id="p0220">αE integrin (CD103) is a key marker of T<sub>RM</sub> cells through its interaction with E-cadherin, participating in retention of T cells at epithelial and mucosal sites (<xref rid="bib36" ref-type="bibr">Mueller and Mackay, 2016</xref>). <xref rid="bib63" ref-type="bibr">Yu et al. (2013)</xref> have identified a feature of total CD1c<sup>+</sup> DCs in their ability to drive acquisition of CD103 in CD8<sup>+</sup> T cells. Here we have shown that DC3s, but not cDC2s, induced expression of CD103 on CD8<sup>+</sup> and CD4<sup>+</sup> T cells. In addition, we have shown that CD103<sup>+</sup> T cells primed by DC3s <italic toggle="yes">ex vivo</italic> align with <italic toggle="yes">bona fide</italic> T<sub>RM</sub> cells isolated from lung or breast cancer (<xref rid="bib21" ref-type="bibr">Hombrink et al., 2016</xref>; <xref rid="bib24" ref-type="bibr">Kumar et al., 2017</xref>; <xref rid="bib49" ref-type="bibr">Savas et al., 2018</xref>). This finding is consistent with the existence of early imprinting of the T<sub>RM</sub> cell program at the level of T cell priming by DC3s. This notion is supported by a recent <italic toggle="yes">in vivo</italic> study in a mouse model, evidencing early imprinting of the T<sub>RM</sub> cell program during T cell priming in the lymph nodes by αV integrin-expressing migratory DCs (<xref rid="bib31" ref-type="bibr">Mani et al., 2019</xref>). Further supporting this view, we found that (1) DC3s upregulated CCR7 when activated by TLR agonists, potentially acting as migratory DCs, and (2) blocking studies indicated that TGF-β was required to prime CD103<sup>+</sup> T cells. This role of priming in lymph nodes does not exclude that tumor-infiltrating DC3s might also provide TGF-β and other signals important for maintenance of T<sub>RM</sub> cells, as demonstrated in a mouse model (<xref rid="bib31" ref-type="bibr">Mani et al., 2019</xref>). In support of this view, we found that infiltration of DC3s was selectively associated with the abundance of CD8<sup>+</sup>CD103<sup>+</sup> T cells in luminal breast cancer primary tumors. Further studies are needed to delineate the full molecular mechanisms of CD103 induction in T cells by DC3s and their physiological <italic toggle="yes">in vivo</italic> relevance during immune responses. Indeed, CD103<sup>+</sup>CD8<sup>+</sup> T cells are a protective biomarker in triple-negative breast cancer (<xref rid="bib49" ref-type="bibr">Savas et al., 2018</xref>; <xref rid="bib59" ref-type="bibr">Wang et al., 2016</xref>), and lung CD103<sup>+</sup>CD8<sup>+</sup> T cells are a hallmark of protective immunity afforded by influenza vaccination (<xref rid="bib63" ref-type="bibr">Yu et al., 2013</xref>). This underlines the potential of DC3s to regulate tissue immunity and defines them as targets for vaccines and immunotherapeutic interventions.</p></sec><sec id="sec7"><title>Limitations of Study</title><p id="p0415">We identify DC3s as CD88<sup>-</sup>CD1c<sup>+</sup>CD163<sup>+</sup>CD14<sup>+/-</sup>. Further studies will be needed to determine if this phenotype corresponds to all CD1c<sup>+</sup>CD14<sup>+</sup> cells that had been reported in various inflammatory settings.</p><p id="p0420">We showed that DC3 differentiation is driven by GM-CSF in humanized mouse metastatic lung model. However, human circulating PBMCs were used in these experiments, preventing the assessment of GM-CSF impact on bone marrow DC3 progenitors <italic toggle="yes">in vivo</italic>. Hence, we cannot exclude a role of GM-CSF on DC3s survival instead of differentiation. In addition, we showed that GM-CSF drives DC3 differentiation from CD34<sup>+</sup> umbilical cord HSPCs <italic toggle="yes">in vitro</italic>. This finding does not exclude that other growth factors control DC3 development <italic toggle="yes">in vivo</italic>.</p><p id="p0425">We aligned CD1c<sup>+</sup>CD14<sup>+</sup> cells infiltrating breast tumor-draining lymph node with blood DC3s compared to cDC2s and monocytes. Due to limited availability of healthy human lymph nodes, we did not investigate if CD1c<sup>+</sup>CD14<sup>+</sup> cells are infiltrating secondary lymphoid organs in homeostatic conditions. Blood DC3s upregulate CCR7 upon TLR stimulation suggesting they have a migratory potential. However, we did not investigate whether lymph nodes DC3s originate from non-lymphoid peripheral tissues via lymphatic vessels or directly from blood.</p><p id="p0430">A defining feature of cDCs lies in their ability to prime naïve T cells. We showed that DC3s can induce the proliferation of naïve T cell in a non-autologous priming context, <italic toggle="yes">i.e.</italic> the mixed leukocyte reaction. The capability of DC3s to uptake, process and present antigens by MHCI and MHCII molecules remains to be addressed.</p><p id="p0435">Our <italic toggle="yes">in vitro</italic> experiments showed that <bold><italic toggle="yes">i</italic></bold>) DC3s induce efficiently T<sub>RM</sub> differentiation from naïve CD8<sup>+</sup>T cells and, <bold><italic toggle="yes">ii</italic></bold>) DC3 infiltration correlates with T<sub>RM</sub> abundance <italic toggle="yes">in vivo</italic>. However, this does not provide a direct evidence that DC3s control T<sub>RM</sub> specification within tumor draining lymph nodes where they are likely to interact with naïve T cells. Indeed, we cannot rule out that, <italic toggle="yes">in vivo</italic>, DC3s act selectively in the tissue to maintain T<sub>RM</sub> populations primed in lymph nodes by a different DC subtype. Further studies will be needed to address how DC3s control CD8<sup>+</sup> T<sub>RM</sub> populations <italic toggle="yes">in vivo</italic>.</p></sec><sec id="sec4"><title>STAR★Methods</title><sec id="sec4.1"><title>Key Resources Table</title><p id="p0225"><table-wrap position="float" id="undtbl1" orientation="portrait"><table frame="hsides" rules="groups"><thead><tr><th colspan="1" rowspan="1">REAGENT or RESOURCE</th><th colspan="1" rowspan="1">SOURCE</th><th colspan="1" rowspan="1">IDENTIFIER</th></tr></thead><tbody><tr><td colspan="3" rowspan="1"><bold>Antibodies</bold></td></tr><tr><td colspan="3" rowspan="1"><hr/></td></tr><tr><td colspan="1" rowspan="1">Biotin anti-human CD335 (NKp46) antibody (clone 9E2)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat#331906; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_1027671" id="intref0010">AB_1027671</ext-link></td></tr><tr><td colspan="1" rowspan="1">Biotin anti-human CD3 antibody (clone OKT3)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 317320; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_10916519" id="intref0015">AB_10916519</ext-link></td></tr><tr><td colspan="1" rowspan="1">Biotin anti-human CD19 antibody (clone HIB19)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 302204; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_314234" id="intref0020">AB_314234</ext-link></td></tr><tr><td colspan="1" rowspan="1">Biotin anti-human CD20 antibody (clone 2H7)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 302350; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2565524" id="intref0025">AB_2565524</ext-link></td></tr><tr><td colspan="1" rowspan="1">Biotin anti-human CD56 (NCAM) antibody (clone HCD56)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 318320; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_893390" id="intref0030">AB_893390</ext-link></td></tr><tr><td colspan="1" rowspan="1">Biotin anti-human CD203c (E-NPP3) antibody (clone NP4D6)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 324604; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_756042" id="intref0035">AB_756042</ext-link></td></tr><tr><td colspan="1" rowspan="1">Biotin anti-human CD66b antibody (clone G10F5)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 305120; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2566608" id="intref0040">AB_2566608</ext-link></td></tr><tr><td colspan="1" rowspan="1">Brilliant Violet 421 anti-human CD16 antibody (clone 3G8)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 302038; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2561578" id="intref0045">AB_2561578</ext-link></td></tr><tr><td colspan="1" rowspan="1">Brilliant Violet 510 anti-human CD45RA antibody (clone HI100)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 304142; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2561947" id="intref0050">AB_2561947</ext-link></td></tr><tr><td colspan="1" rowspan="1">Brilliant Violet 510 anti-human HLA-DR antibody (clone L243)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 307646; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2561948" id="intref0055">AB_2561948</ext-link></td></tr><tr><td colspan="1" rowspan="1">Brilliant Violet 510 anti-human CD15 (SSEA-1) antibody (clone W6D3)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 323028; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2563400" id="intref0060">AB_2563400</ext-link></td></tr><tr><td colspan="1" rowspan="1">FITC anti-human CD163 antibody (clone GHI/61)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 333618; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2563094" id="intref0065">AB_2563094</ext-link></td></tr><tr><td colspan="1" rowspan="1">Brilliant Violet 711 anti-human CD163 antibody (clone GHI/61)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 333630; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2650972" id="intref0070">AB_2650972</ext-link></td></tr><tr><td colspan="1" rowspan="1">Brilliant Violet 421 anti-human CD163 antibody (clone GHI/61)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 333612; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2562463" id="intref0075">AB_2562463</ext-link></td></tr><tr><td colspan="1" rowspan="1">Brilliant Violet 785 anti-human CD14 antibody (clone M5E2)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 301840; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2563425" id="intref0080">AB_2563425</ext-link></td></tr><tr><td colspan="1" rowspan="1">PE/Cy7 anti-human CD14 antibody (clone HCD14)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 325618; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_830691" id="intref0085">AB_830691</ext-link></td></tr><tr><td colspan="1" rowspan="1">FITC anti-human CD14 antibody (clone M5E2)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 301804; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_314186" id="intref0090">AB_314186</ext-link>)</td></tr><tr><td colspan="1" rowspan="1">Brilliant Violet 785 anti-human CD3 antibody (clone UCHT1)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 300472; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2687178" id="intref0095">AB_2687178</ext-link></td></tr><tr><td colspan="1" rowspan="1">APC anti-human CD34 antibody (clone 561)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 343608; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2228972" id="intref0100">AB_2228972</ext-link></td></tr><tr><td colspan="1" rowspan="1">FITC anti-human CD34 antibody (clone 561)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 343604; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_1732005" id="intref0105">AB_1732005</ext-link></td></tr><tr><td colspan="1" rowspan="1">PerCP/Cyanine5.5 anti-human CD303 (BDCA-2) antibody (clone 201A)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 354210; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_11219604" id="intref0110">AB_11219604</ext-link></td></tr><tr><td colspan="1" rowspan="1">PerCP/Cyanine5.5 anti-human CD141 (Thrombomodulin) antibody (clone M80)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 344112; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2561625" id="intref0115">AB_2561625</ext-link></td></tr><tr><td colspan="1" rowspan="1">PerCP/Cyanine5.5 anti-human CD371 (CLEC12A) antibody (clone 50C1)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 353612; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2565544" id="intref0120">AB_2565544</ext-link></td></tr><tr><td colspan="1" rowspan="1">PE anti-human CD370 (CLEC9A/DNGR1) antibody (clone 8F9)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 353804; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_10965546" id="intref0125">AB_10965546</ext-link></td></tr><tr><td colspan="1" rowspan="1">PE anti-human CD5 antibody (clone L17F12)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 364014; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2565284" id="intref0130">AB_2565284</ext-link></td></tr><tr><td colspan="1" rowspan="1">PE anti-human CD116 antibody (clone 4H1)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 305908; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2085686" id="intref0135">AB_2085686</ext-link></td></tr><tr><td colspan="1" rowspan="1">PE anti-human CD206 (MMR) antibody (clone 15-2)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 321106; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_571911" id="intref0140">AB_571911</ext-link></td></tr><tr><td colspan="1" rowspan="1">PE anti-human CD88 (C5aR) antibody (clone S5/1)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 344304; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2067175" id="intref0145">AB_2067175</ext-link></td></tr><tr><td colspan="1" rowspan="1">PE/Dazzle 594 anti-human CD123 antibody (clone 6H6)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 306034; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2566450" id="intref0150">AB_2566450</ext-link></td></tr><tr><td colspan="1" rowspan="1">PerCP/Cyanine5.5 anti-human CD123 antibody (clone 6H6)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 306016; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2264693" id="intref0155">AB_2264693</ext-link></td></tr><tr><td colspan="1" rowspan="1">FITC anti-human CD1c antibody (clone L161)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 331518; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2073403" id="intref0160">AB_2073403</ext-link></td></tr><tr><td colspan="1" rowspan="1">PE/Cy7 anti-human CD1c antibody (clone L161)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 331516; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2275574" id="intref0165">AB_2275574</ext-link></td></tr><tr><td colspan="1" rowspan="1">APC anti-human CD272 (BTLA) antibody (clone MIH26)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 344510; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_10613101" id="intref0170">AB_10613101</ext-link></td></tr><tr><td colspan="1" rowspan="1">APC anti-human CD64 antibody (clone 10.1)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 305014; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_1595428" id="intref0175">AB_1595428</ext-link></td></tr><tr><td colspan="1" rowspan="1">APC/Cyanine7 anti-human CD45 antibody (clone HI30)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 304014; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_314402" id="intref0180">AB_314402</ext-link></td></tr><tr><td colspan="1" rowspan="1">APC/Cy7 Streptavidin</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 405208</td></tr><tr><td colspan="1" rowspan="1">Brilliant Violet 785 anti-human CD95 (Fas) antibody (clone DX2)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 305646; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2629742" id="intref0185">AB_2629742</ext-link></td></tr><tr><td colspan="1" rowspan="1">Brilliant Violet 711 anti-human CD8 antibody (clone SK1)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 344734; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2565243" id="intref0190">AB_2565243</ext-link></td></tr><tr><td colspan="1" rowspan="1">Brilliant Violet 605 anti-human CD45 antibody (clone HI30)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 304042; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2562106" id="intref0195">AB_2562106</ext-link></td></tr><tr><td colspan="1" rowspan="1">Brilliant Violet 650 anti-human CD45RO antibody (clone UCHL1)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 304232; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2563462" id="intref0200">AB_2563462</ext-link></td></tr><tr><td colspan="1" rowspan="1">Brilliant Violet 421 anti-human CD103 (Integrin E) antibody (clone Ber-ACT8)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 350214; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2563514" id="intref0205">AB_2563514</ext-link></td></tr><tr><td colspan="1" rowspan="1">Brilliant Violet 785 anti-human CD103 (Integrin E) antibody (clone Ber-ACT8)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 350230; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2734364" id="intref0210">AB_2734364</ext-link></td></tr><tr><td colspan="1" rowspan="1">FITC anti-human CD223 (LAG-3) antibody (clone 7H2C65)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 369210; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2716129" id="intref0215">AB_2716129</ext-link></td></tr><tr><td colspan="1" rowspan="1">FITC anti-human CD3 antibody (clone UCHT1)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 300406; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_314060" id="intref0220">AB_314060</ext-link></td></tr><tr><td colspan="1" rowspan="1">PE/Cy7 anti-human CD183 (CXCR3) antibody (clone G025H7)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 353720; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_11219383" id="intref0225">AB_11219383</ext-link></td></tr><tr><td colspan="1" rowspan="1">PE/Dazzle 594 anti-human CD28 antibody (clone CD28.2)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 302942; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2564235" id="intref0230">AB_2564235</ext-link></td></tr><tr><td colspan="1" rowspan="1">PE anti-human CD197 (CCR7) antibody (clone G043H7)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 353204; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_10913813" id="intref0235">AB_10913813</ext-link></td></tr><tr><td colspan="1" rowspan="1">APC/Cyanine7 anti-human CD4 antibody (clone SK3)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 344616; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2028483" id="intref0240">AB_2028483</ext-link></td></tr><tr><td colspan="1" rowspan="1">Alexa Fluor® 700 anti-human CD127 (IL-7Ra) antibody (clone A019D5)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 351344; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2566200" id="intref0245">AB_2566200</ext-link></td></tr><tr><td colspan="1" rowspan="1">APC anti-human CD279 (PD-1) antibody (clone EH12.2H7)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 329908; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_940475" id="intref0250">AB_940475</ext-link></td></tr><tr><td colspan="1" rowspan="1">PE anti-human CD115 antibody (clone 9-4D2-1D4)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 347310; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2565491" id="intref0255">AB_2565491</ext-link></td></tr><tr><td colspan="1" rowspan="1">FITC anti-human CD116 antibody (clone 4H1)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 305906; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2085687" id="intref0260">AB_2085687</ext-link></td></tr><tr><td colspan="1" rowspan="1">FITC anti-human CD86 antibody (clone BU63)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 374204; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2721574" id="intref0265">AB_2721574</ext-link></td></tr><tr><td colspan="1" rowspan="1">Human Axl PE antibody (clone 108724)</td><td colspan="1" rowspan="1">R&amp;D</td><td colspan="1" rowspan="1">Cat# FAB154P; no RRID</td></tr><tr><td colspan="1" rowspan="1">PerCP/Cyanine5.5 anti-mouse CD45 antibody (clone 30-F11)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 103132; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_893340" id="intref0270">AB_893340</ext-link></td></tr><tr><td colspan="1" rowspan="1">APC-eFluor 780 anti-human HLA-DR antibody (clone LN3)</td><td colspan="1" rowspan="1">eBioscience</td><td colspan="1" rowspan="1">Cat# 47-9956; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_1963603" id="intref0275">AB_1963603</ext-link></td></tr><tr><td colspan="1" rowspan="1">PE/Cy7 anti-human CD11c antibody (clone Bu15)</td><td colspan="1" rowspan="1">Biolegend</td><td colspan="1" rowspan="1">Cat# 337216; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2129790" id="intref0280">AB_2129790</ext-link></td></tr><tr><td colspan="1" rowspan="1">FITC anti-human CD14 antibody (clone 61D3)</td><td colspan="1" rowspan="1">eBioscience</td><td colspan="1" rowspan="1">Cat# 11-0149; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_464951" id="intref0285">AB_464951</ext-link></td></tr><tr><td colspan="1" rowspan="1">PerCP-eFluor 710 anti-human CD1c antibody (clone L161)</td><td colspan="1" rowspan="1">eBioscience</td><td colspan="1" rowspan="1">Cat# 46-0015; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_10548936" id="intref0290">AB_10548936</ext-link></td></tr><tr><td colspan="1" rowspan="1">PE anti-human CD304 antibody (clone AD5-17F6)</td><td colspan="1" rowspan="1">Miltenyi</td><td colspan="1" rowspan="1">Cat# 130-113-517; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2751124" id="intref0295">AB_2751124</ext-link></td></tr><tr><td colspan="1" rowspan="1">PE-Cy5 Mouse anti-Human CD1a antibody (clone HI149)</td><td colspan="1" rowspan="1">BD</td><td colspan="1" rowspan="1">Cat# 555808; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_396142" id="intref0300">AB_396142</ext-link></td></tr><tr><td colspan="1" rowspan="1">Alexa Fluor® 647 anti-human CD206 (MMR) antibody (clone 15-2)</td><td colspan="1" rowspan="1">Biolegend</td><td colspan="1" rowspan="1">Cat# 321116; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_571881" id="intref0305">AB_571881</ext-link></td></tr><tr><td colspan="1" rowspan="1">VioBlue anti-human CD141 antibody (clone AD5-14H12)</td><td colspan="1" rowspan="1">Miltenyi</td><td colspan="1" rowspan="1">Cat# 130-113-882; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2726374" id="intref0310">AB_2726374</ext-link></td></tr><tr><td colspan="1" rowspan="1">Purified anti-human CD45 antibody (clone HI30)</td><td colspan="1" rowspan="1">BioLegend</td><td colspan="1" rowspan="1">Cat# 304002; RRID: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_314390" id="intref0315">AB_314390</ext-link></td></tr><tr><td colspan="1" rowspan="1">Purified anti-human TGF 1-2-3 antibody (clone 1d11)</td><td colspan="1" rowspan="1">R&amp;D</td><td colspan="1" rowspan="1">Cat# MAB1835; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_357931" id="intref0320">AB_357931</ext-link></td></tr><tr><td colspan="1" rowspan="1">Purified anti-human TNFα antibody (clone 1825)</td><td colspan="1" rowspan="1">R&amp;D</td><td colspan="1" rowspan="1">Cat# MAB210; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2240620" id="intref0325">AB_2240620</ext-link></td></tr><tr><td colspan="1" rowspan="1">Purified anti-human IL-15 antibody (clone 34593)</td><td colspan="1" rowspan="1">R&amp;D</td><td colspan="1" rowspan="1">Cat# MAB247; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2124578" id="intref0330">AB_2124578</ext-link></td></tr><tr><td colspan="1" rowspan="1">Purified anti-human Lymphotoxin α antibody (clone 5802)</td><td colspan="1" rowspan="1">R&amp;D</td><td colspan="1" rowspan="1">Cat# MAB211; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2138622" id="intref0335">AB_2138622</ext-link></td></tr><tr><td colspan="1" rowspan="1">Purified polyclonal anti-human OX40L antibody</td><td colspan="1" rowspan="1">R&amp;D</td><td colspan="1" rowspan="1">Cat# AF1236; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_354686" id="intref0340">AB_354686</ext-link></td></tr><tr><td colspan="1" rowspan="1">Purified polyclonal anti-human IL-33 antibody</td><td colspan="1" rowspan="1">R&amp;D</td><td colspan="1" rowspan="1">Cat# AF3625; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_1151900" id="intref0345">AB_1151900</ext-link></td></tr><tr><td colspan="1" rowspan="1">Purified anti-human TNFSF8 antibody (clone 116614)</td><td colspan="1" rowspan="1">R&amp;D</td><td colspan="1" rowspan="1">Cat# MAB1028; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_2303710" id="intref0350">AB_2303710</ext-link></td></tr><tr><td colspan="1" rowspan="1">Purified Mouse IgG1 (clone 11711)</td><td colspan="1" rowspan="1">R&amp;D</td><td colspan="1" rowspan="1">Cat# MAB002; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_357344" id="intref0355">AB_357344</ext-link></td></tr><tr><td colspan="1" rowspan="1">Purified polyclonal Goat IgG</td><td colspan="1" rowspan="1">R&amp;D</td><td colspan="1" rowspan="1">Cat# AB-108-C; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_354267" id="intref0360">AB_354267</ext-link></td></tr><tr><td colspan="1" rowspan="1">Purified Mouse IgG2b (clone 20116)</td><td colspan="1" rowspan="1">R&amp;D</td><td colspan="1" rowspan="1">Cat# MAB004; RRID:<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="nif-antibody:AB_357346" id="intref0365">AB_357346</ext-link></td></tr><tr><td colspan="3" rowspan="1"><hr/></td></tr><tr><td colspan="3" rowspan="1"><bold>Bacterial and Virus Strains</bold></td></tr><tr><td colspan="3" rowspan="1"><hr/></td></tr><tr><td colspan="1" rowspan="1">pMX-IRES-GFP</td><td colspan="1" rowspan="1">Cell biolabs</td><td colspan="1" rowspan="1">Cat# RTV-0133</td></tr><tr><td colspan="1" rowspan="1">pMX-huFLT3L-IRES-GFP</td><td colspan="1" rowspan="1">This paper</td><td colspan="1" rowspan="1">N/A</td></tr><tr><td colspan="1" rowspan="1">pMX-huGMCSF-IRES-GFP</td><td colspan="1" rowspan="1">This paper</td><td colspan="1" rowspan="1">N/A</td></tr><tr><td colspan="3" rowspan="1"><hr/></td></tr><tr><td colspan="3" rowspan="1"><bold>Biological Samples</bold></td></tr><tr><td colspan="3" rowspan="1"><hr/></td></tr><tr><td colspan="1" rowspan="1">Leukocyte cones for bulk RNaseq</td><td colspan="1" rowspan="1">NHS</td><td colspan="1" rowspan="1">N/A</td></tr><tr><td colspan="1" rowspan="1">Leukocytes cones for functionnal experiments</td><td colspan="1" rowspan="1">NHS</td><td colspan="1" rowspan="1">N/A</td></tr><tr><td colspan="1" rowspan="1">umbilical cord-blood for <italic toggle="yes">in vitro</italic> diffentiation cells</td><td colspan="1" rowspan="1">Anthony Nolan Cell Therapy Centre</td><td colspan="1" rowspan="1">N/A</td></tr><tr><td colspan="1" rowspan="1">PBMC from whole blood for blood scRNaseq</td><td colspan="1" rowspan="1">Institut Curie Hospital (Paris, France)</td><td colspan="1" rowspan="1">N/A</td></tr><tr><td colspan="3" rowspan="1"><hr/></td></tr><tr><td colspan="3" rowspan="1"><bold>Chemicals, Peptides, and Recombinant Proteins</bold></td></tr><tr><td colspan="3" rowspan="1"><hr/></td></tr><tr><td colspan="1" rowspan="1">Liberase TL</td><td colspan="1" rowspan="1">Sigma</td><td colspan="1" rowspan="1">Cat# 5401020001</td></tr><tr><td colspan="1" rowspan="1">Collagenase D</td><td colspan="1" rowspan="1">Sigma</td><td colspan="1" rowspan="1">Cat# 11088866001</td></tr><tr><td colspan="1" rowspan="1">DNase I</td><td colspan="1" rowspan="1">Sigma</td><td colspan="1" rowspan="1">Cat# 10104159001</td></tr><tr><td colspan="1" rowspan="1">Dispase</td><td colspan="1" rowspan="1">Sigma</td><td colspan="1" rowspan="1">Cat# D4693</td></tr><tr><td colspan="1" rowspan="1">CD34 MicroBead Kit UltraPure, human</td><td colspan="1" rowspan="1">Miltenyi</td><td colspan="1" rowspan="1">Cat# 130-100-453</td></tr><tr><td colspan="1" rowspan="1">Accucheck counting beads</td><td colspan="1" rowspan="1">Thermo Fischer</td><td colspan="1" rowspan="1">Cat# PCB100</td></tr><tr><td colspan="1" rowspan="1">Matrigel®</td><td colspan="1" rowspan="1">BD</td><td colspan="1" rowspan="1">Cat# 354230</td></tr><tr><td colspan="1" rowspan="1">Poly I:C</td><td colspan="1" rowspan="1">InvivoGen</td><td colspan="1" rowspan="1">Cat# 31852-29-6</td></tr><tr><td colspan="1" rowspan="1">R848</td><td colspan="1" rowspan="1">InvivoGen</td><td colspan="1" rowspan="1">Cat# 144875-48-9</td></tr><tr><td colspan="1" rowspan="1">lipopolysaccharide (LPS)</td><td colspan="1" rowspan="1">Sigma-Aldrich</td><td colspan="1" rowspan="1">Cat# L2630</td></tr><tr><td colspan="1" rowspan="1">Cell Trace Violet (CTV)</td><td colspan="1" rowspan="1">Thermo Fischer</td><td colspan="1" rowspan="1">Cat# <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="pmc:entrez-nucleotide" xlink:href="C34557">C34557</ext-link></td></tr><tr><td colspan="1" rowspan="1">Naive Pan T cells isolation Kit</td><td colspan="1" rowspan="1">Miltenyi</td><td colspan="1" rowspan="1">Cat# 130-097-095</td></tr><tr><td colspan="1" rowspan="1">Cytostim</td><td colspan="1" rowspan="1">Miltenyi</td><td colspan="1" rowspan="1">Cat# 130-092-172</td></tr><tr><td colspan="1" rowspan="1">Cytofix/ cytoperm Kit</td><td colspan="1" rowspan="1">BD Bioscience</td><td colspan="1" rowspan="1">Cat# 554714</td></tr><tr><td colspan="3" rowspan="1"><hr/></td></tr><tr><td colspan="3" rowspan="1"><bold>Critical Commercial Assays</bold></td></tr><tr><td colspan="3" rowspan="1"><hr/></td></tr><tr><td colspan="1" rowspan="1">Human Flt-3 Ligand/FLT3L Quantikine ELISA Kit</td><td colspan="1" rowspan="1">R&amp;D</td><td colspan="1" rowspan="1">Cat# DFK00</td></tr><tr><td colspan="1" rowspan="1">ELISA MAX Deluxe Set Human GM-CSF</td><td colspan="1" rowspan="1">Biolegend</td><td colspan="1" rowspan="1">Cat# 432004</td></tr><tr><td colspan="1" rowspan="1">Human Proinflammatory Chemokine Panel (13-plex)</td><td colspan="1" rowspan="1">Biolegend</td><td colspan="1" rowspan="1">Cat# 740003</td></tr><tr><td colspan="1" rowspan="1">Human Macrophage/Microglia Panel (13-plex)</td><td colspan="1" rowspan="1">Biolegend</td><td colspan="1" rowspan="1">Cat# 740503</td></tr><tr><td colspan="3" rowspan="1"><hr/></td></tr><tr><td colspan="3" rowspan="1"><bold>Deposited Data</bold></td></tr><tr><td colspan="3" rowspan="1"><hr/></td></tr><tr><td colspan="1" rowspan="1">RNaseq data raw reads and processed data</td><td colspan="1" rowspan="1">This paper</td><td colspan="1" rowspan="1">GEO: SuperSeries <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="pmc:entrez-geo" xlink:href="GSE151095">GSE151095</ext-link></td></tr><tr><td colspan="3" rowspan="1"><hr/></td></tr><tr><td colspan="3" rowspan="1"><bold>Experimental Models: Cell Lines</bold></td></tr><tr><td colspan="3" rowspan="1"><hr/></td></tr><tr><td colspan="1" rowspan="1">B16_Control</td><td colspan="1" rowspan="1">This paper</td><td colspan="1" rowspan="1">N/A</td></tr><tr><td colspan="1" rowspan="1">B16_human FLT3L</td><td colspan="1" rowspan="1">This paper</td><td colspan="1" rowspan="1">N/A</td></tr><tr><td colspan="1" rowspan="1">B16_human GM-CSF</td><td colspan="1" rowspan="1">This paper</td><td colspan="1" rowspan="1">N/A</td></tr><tr><td colspan="1" rowspan="1">MS5_FLT3L</td><td colspan="1" rowspan="1"><xref rid="bib2" ref-type="bibr">Anselmi et al., 2020</xref></td><td colspan="1" rowspan="1">N/A</td></tr><tr><td colspan="1" rowspan="1">MS5_GM-CSF</td><td colspan="1" rowspan="1"><xref rid="bib2" ref-type="bibr">Anselmi et al., 2020</xref></td><td colspan="1" rowspan="1">N/A</td></tr><tr><td colspan="1" rowspan="1">MS5_FS12</td><td colspan="1" rowspan="1"><xref rid="bib2" ref-type="bibr">Anselmi et al., 2020</xref></td><td colspan="1" rowspan="1">N/A</td></tr><tr><td colspan="3" rowspan="1"><hr/></td></tr><tr><td colspan="3" rowspan="1"><bold>Experimental Models: Organisms/Strains</bold></td></tr><tr><td colspan="3" rowspan="1"><hr/></td></tr><tr><td colspan="1" rowspan="1">mouse: NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ</td><td colspan="1" rowspan="1">The Jakson Laboratory</td><td colspan="1" rowspan="1">Stock#005557</td></tr><tr><td colspan="3" rowspan="1"><hr/></td></tr><tr><td colspan="3" rowspan="1"><bold>Software and Algorithms</bold></td></tr><tr><td colspan="3" rowspan="1"><hr/></td></tr><tr><td colspan="1" rowspan="1">Genomics Suite</td><td colspan="1" rowspan="1">Partek</td><td colspan="1" rowspan="1">N/A</td></tr><tr><td colspan="1" rowspan="1">FlowJo V10</td><td colspan="1" rowspan="1">BD</td><td colspan="1" rowspan="1"><ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="https://www.flowjo.com" id="intref0370">https://www.flowjo.com</ext-link></td></tr><tr><td colspan="1" rowspan="1">GraphPad Prism 8</td><td colspan="1" rowspan="1">GraphPad Software</td><td colspan="1" rowspan="1"><ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="https://www.graphpad.com" id="intref0375">https://www.graphpad.com</ext-link></td></tr><tr><td colspan="1" rowspan="1">R4.4</td><td colspan="1" rowspan="1">The R Foundation</td><td colspan="1" rowspan="1"><ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="https://www.r-project.org" id="intref0380">https://www.r-project.org</ext-link></td></tr><tr><td colspan="1" rowspan="1">Fiji</td><td colspan="1" rowspan="1">open source</td><td colspan="1" rowspan="1"><ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="https://imagej.net/Fiji" id="intref0385">https://imagej.net/Fiji</ext-link></td></tr></tbody></table></table-wrap></p></sec><sec id="sec4.2"><title>Resource Availability</title><sec id="sec4.2.1"><title>Lead Contact</title><p id="p0230">Additional information and request for resources and reagents should be directed to and will be made available by the Lead Contact, Pierre Guermonprez (<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="mailto:pierre.guermonprez@kcl.ac.uk" id="intref0390">pierre.guermonprez@kcl.ac.uk</ext-link>).</p></sec><sec id="sec4.2.2"><title>Materials Availability</title><p id="p0235">The reagents generated in this study will be made available on request, but we may require a payment and/or a completed Materials Transfer Agreement if there is potential for commercial application.</p></sec><sec sec-type="data-availability" id="sec4.2.3"><title>Data and Code Availability</title><p id="p0240">The scRNA-seq and the bulk RNA-seq datasets are deposited in the Genome Expression Omnibus under the SuperSeries accession numbers <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="ncbi-geo:GSE151095" id="intref0395">GSE151095</ext-link>.</p></sec></sec><sec id="sec4.3"><title>Experimental Model and Subject Details</title><sec id="sec4.3.1"><title>Human umbilical cord and adult blood</title><p id="p0245">Human umbilical cord blood units were obtained from Anthony Nolan Cell Therapy Centre (ANCTC). Leukophoretic adult blood (buffy coats or leukocyte cones) were obtained from healthy volunteers through NHS.</p></sec><sec id="sec4.3.2"><title>Clinical samples</title><p id="p0250">Tumor-invaded lymph nodes (tdLN) and primary tumors were collected from luminal breast cancer submitted to surgical resection at the Institut Curie Hospital (Paris, France), in accordance with institutional ethical guidelines. Patients’ clinical and pathologic characteristics are summarized in <xref rid="mmc1" ref-type="supplementary-material">Table S2</xref>.</p></sec><sec id="sec4.3.3"><title>Mice</title><p id="p0255">NSG (NOD.Cg-<italic toggle="yes">Prkdc</italic><sup><italic toggle="yes">scid</italic></sup>
+<italic toggle="yes">Il2rg</italic><sup><italic toggle="yes">tm1Wjl</italic></sup>/SzJ) mice were bred and maintained in specific pathogen-free animal facility in accordance with institutional KCL guidelines. All procedures involving animals were conducted according to requirements of UK Animals (Scientific Procedures) Act 1986.</p></sec></sec><sec id="sec4.4"><title>Method Details</title><sec id="sec4.4.1"><title>Human blood and CD34<sup>+</sup> progenitors</title><p id="p0260">Peripheral blood mononuclear cells (PMBCs) were obtained by gradient centrifugation using Ficoll-Paque (GE Healthcare). Progenitor cells were enriched using CD34<sup>+</sup> microbead isolation kit (Miltenyi).</p></sec><sec id="sec4.4.2"><title>Tumor cell lines</title><p id="p0265">B16_CTRL, B16_huGM-CSF and B16_huFLT3L were generated by retroviral transduction of B16-F10 (C57BL/6 melanoma cell line) with an empty pMX-IRES-GFP vector or coding for human GM-CSF and human Flt3L respectively. Tumor cell lines were cultured in RPMI 1640 medium (Thermo Fisher) supplemented with 10% fetal bovine serum (FBS) (Thermo Fisher), penicillin-streptomycin (Thermo Fisher) and b-mercaptoethanol (Thermo Fisher) (complete RPMI) and maintained at 37°C and 5% CO<sub>2</sub>.</p></sec><sec id="sec4.4.3"><title>Metastasis model</title><p id="p0270">Engineered B16-F10 cells were counted and resuspended in RPMI 1640 medium. NSG mice (8-12 weeks, males and females) were injected intravenously with 10<sup>6</sup> B16-F10 at day 0. 10<sup>8</sup> human blood PBMCs were injected intravenously at day 7 and 8. Mice were culled at day 9 and lungs were harvested.</p></sec><sec id="sec4.4.4"><title>Histology</title><p id="p0275">Mouse lungs were fixed with 1% PFA (Alfa Aesar) for 1hr at 4°C, washed and incubated in 34% sucrose solution (Sigma-Aldrich) overnight at 4°C. Lungs were embedded in Cryomatrix (Thermo Fischer) and frozen for cryostat sectioning (9 μm-thick). Sections were permeabilized using 0.5% saponin (Sigma-Aldrich), 2% BSA (Sigma-Aldrich), 1% FBS (Life Technologies) for 30 minutes at room temperature. Sections were labeled overnight at 4°C with mouse anti-human purified CD45 (HI30, Biolegend) followed by incubation for 1hr at room temperature with goat anti-mouse Cy3 (Jackson laboratory). All sections were labeled with Hoechst (Molecular Probes, Thermo Fisher) for nuclei staining 5 minutes at room temperature and mounted with Prolong diamond (Thermo scientific). Slides were imaged using a SP5 (Leica) and analyzed with Fiji software.</p></sec><sec id="sec4.4.5"><title>Preparation of cell suspensions from lung mouse</title><p id="p0280">Mouse lungs were harvested and transferred to 3ml digestion buffer (Hank’s Balanced Salt Solution (HBSS) with calcium and magnesium (Thermo Fisher) and with 0.1 mg/ml of Liberase TL (Roche) and 0.02μg/ml DNase I (Thermo Fisher). Lungs were dissociated using gentleMACS Octo Dissociator (Miltenyi) and incubated at 37°C for 45 minutes. The cell suspension was passed through a cell strainer (70μm, Corning) and red blood cells were lysed using ACK lysing buffer (Thermo Fischer). The absolute number of cells in the resulting cell suspension was assessed using AccuCheck Counting Beads (Thermo Fisher) on BD FACSCanto II (BD Biosciences).</p></sec><sec id="sec4.4.6"><title>Stromal cell line maintenance</title><p id="p0285">Mouse bone marrow-derived MS5 stromal cells engineered to express human membrane bound FLT3L alone (MS5_FL) or in combination with SCF and soluble CXCL12 (MS5_FS12) or human GM-CSF alone (MS5_GM) generated as previously described (<xref rid="bib2" ref-type="bibr">Anselmi et al., 2020</xref>). Stomal cell lines were cultured in IMDM medium supplemented with 10% heat-inactivated FBS, penicillin/streptomycin, 50 μM β-mercaptoethanol (complete IMDM) and maintained at 37°C and 5% CO<sub>2</sub>.</p></sec><sec id="sec4.4.7"><title><italic toggle="yes">In vitro</italic> differentiation from CD34<sup>+</sup> progenitors or blood phagocytes</title><p id="p0290">For <italic toggle="yes">in vitro</italic> co-culture experiments with stromal cells, MS5_FL, MS5_FS12 and MS5_GM feeders were seeded at 10<sup>4</sup> cells/well density in a 96-well plate (Thermo Fischer) and maintained overnight at 37°C and 5% CO<sub>2</sub>.</p><p id="p0295">Total enriched-CD34<sup>+</sup> progenitors and flow cytrometry-sorted progenitor subsets (GMDP, CDP and cMoP) were plated on top of stromal cells, supplemented or not with 100ng/ml of recombinant human GM-CSF (Peprotech), for 14 days and 7 days, respectively. Blood phagocytes subsets were flow cytometry-sorted and plated on MS5_GM for 2 days.</p><p id="p0300">For bulk RNA sequencing analysis, total enriched-CD34<sup>+</sup> progenitors were plated on top of MS5_FS12 feeders and with or without 100ng/ml of recombinant human GM-CSF (Peprotech). On day 5 and 10 of differentiation, half the medium was replaced with fresh complete IMDM or complete IMDM containing 100ng/ml of recombinant human GM-CSF. Cells were collected at day 15.</p><p id="p0305">All cells were collected with a solution of phosphate buffered saline (PBS) (GIBCO) 5mM EDTA (Thermo Fisher) at 4°C for 10min.</p></sec><sec id="sec4.4.8"><title><italic toggle="yes">In vivo</italic> conversion assay</title><p id="p0310">PBMCs from healthy donors were stained and sorted as described previously. Blood cDC2 (10<sup>5</sup>) and monocytes (3 × 10<sup>5</sup>) were flow cytometry-sorted and injected subcutaneously along with B16_huGM (10<sup>5</sup>) in 200 μL of ice-cold Matrigel<bold>®</bold> (BD Biosciences). Mice were sacrificed at day 4 by cervical dislocation and Matrigel<bold>®</bold> plugs were collected. Subcutaneous Matrigel<bold>®</bold> plugs were recovered, cut in pieces and incubated in HBSS (Life Technologies) 1% FBS, 0.37 U/ml Collagenase D (Roche), 10 μg/ml DNaseI (Roche) and 1 mg/ml Dispase (Sigma-Aldrich) for 30 minutes at 37°C. After digestion, plugs were smashed on a 70 μm strainer (Corning) and cells were collected and resuspended in flow cytometry buffer for flow cytometry analysis.</p></sec><sec id="sec4.4.9"><title>Flow cytometry analysis</title><p id="p0315">Cells were stained in flow cytometry buffer (PBS 5mM EDTA 1% BSA) according to antibody panels (<xref rid="sec4.1" ref-type="sec">Key Resources Table</xref>) for 30 min. Dead cells were stained using DAPI or Live/Dead Blue staining (Thermo Fischer). Lineage (Lin) included CD3, CD19, CD20, NKp46, CD56, CD203c and CD66b, all conjugated with biotin. Multiparameter analysis was performed on LSRFortessa and Symphony (BD Biosciences) flow cytometers and analyzed using FlowJo software (Tree Star). The flow cytometry-sort was performed on BD FACS Aria II or BD FACS Aria Fusion at the Biomedical Research Centre (BRC) Flow Core Facility (Guy’s and St Thomas’ NHS Foundation Trust and King’s College London).</p><p id="p0320">Unsupervised clustering of flow cytometry data (<xref rid="fig1" ref-type="fig">Figure 1</xref>F) was performed using Self-Organizing Map clustering algorithm FlowSOM (Cytofkit R package). Data was then extracted as FCS files and further analyzed in FlowJo software.</p></sec><sec id="sec4.4.10"><title>Bulk RNA sequencing</title><p id="p0325">For bulk sequencing of <italic toggle="yes">in vitro</italic> differentiated subsets, up to 100 flow cytometry-sorted cells from three individual donors were collected directly in Lysis buffer (Takara Clontech, Cat# 635015) containing RNase inhibitors. RNaseq libraries were prepared on the contactless liquid handling system Labcyte Echo 525 (Labcyte Inc). In brief, ERCC was added to each sample and first strand full length cDNA was generated with a modified protocol of the SMARTseq v4 Ultra Low Input RNA Kit (Takara Clontech, Cat# 634891) using poly dT primers and a template switching oligo. Full length cDNA was amplified using SeqAmp DNA Polymerase (Takara Clontech, Cat# 638509). 12 ng of amplified cDNA from each sample was used to generate non-stranded RNA libraries using a modified protocol of the Ovation Ultralow System V2 1-96 kit (NuGEN, Cat# 0347-A01). In brief, amplified cDNA was fragmented through sonication on Covaris E220 (Covaris Inc), repaired and polished followed by ligation of indexed adapters. Adaptor ligated cDNA were pooled before final amplification to add flow cell primers. Libraries were sequenced on HiSeq2500 (Illumina Cambridge) for 100 cycles PE in Rapid mode.</p></sec><sec id="sec4.4.11"><title>Bulk RNA sequencing data processing</title><p id="p0330">The raw sequencing data was initially processed using open source, web-based platform Galaxy (version 18.05.rc1) (<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="https://usegalaxy.org" id="intref0400">https://usegalaxy.org</ext-link>). Reads were filtered for quality with more than 80% of the sequence having quality score &gt; 33 using FastQC tool. Mapping against reference genome was performed with Hisat2 to the hg38 human genome. Adaptor sequences were detected automatically with TrimGalore!. Reads under 20bp were discarded. All processed sequencing files were imported in Partek® Genomics Suite software®, version 7.0©; 2017 (PGS), where they were processed further.</p></sec><sec id="sec4.4.12"><title>Primary data analysis and visualization</title><p id="p0335">mRNA was quantified using PGS built in RNA-seq workflow. Normalization method used was Reads Per Kilobase per Million mapped reads (RPKM) and mRNA was quantified against RefSeq Transcripts 2018-11-20 database. Hierarchal clustering on average expression within the group was performed on all identified protein coding genes (19791 genes). Based on the CD14<sup>-</sup>CD1c<sup>+</sup>CD206<sup>+/−</sup> two blue cluster similarity, they were both considered as cDC2-like cells and their datasets were merged for further analysis. The same was done for the two macrophage-like cells (gray and brown). Differentially expressed genes (Fold-Change ≤ −2 or ≥ 2 and p value &lt; 0.05) were determined using one-way ANOVA in all pairwise comparisons with three donors grouped and visualized as Volcano plots in PGS. Individual samples were visualized via principal component analysis (PCA) using 500 most variable genes, which were determined based on median absolute deviation (MAD). The expression patterns of selected gene lists were displayed in the form of heatmaps, where rows and/or columns were ordered based on hierarchical clustering using Euclidean distance and average linkage in Morpheus (Broad Institute; <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="https://software.broadinstitute.org/morpheus/" id="intref0405">https://software.broadinstitute.org/morpheus/</ext-link>). A few gene names are depicted next to the heatmap.</p></sec><sec id="sec4.4.13"><title>Single-cell RNA sequencing</title><p id="p0340">PBMCs from three healthy donors or human CD45<sup>+</sup> cells isolated from metastatic mouse lungs were isolated as previously described. Sorted cells from each donor were pooled together and CD1c<sup>+</sup> and CD14<sup>+</sup> cells were mixed at 80:20 ratio. 3 × 10<sup>3</sup> cells from the resulting cell suspension were partitioned into an emulsion of nanoliter-sized droplets using a 10X Genomics Chromium Single Cell Controller and RNA sequencing libraries were constructed using the Chromium Single Cell 3′ Library &amp; Gel Bead Kit v2 (10X Genomics, Cat# PN-120237). Briefly, droplets containing individual cells, reverse transcription reagents and a gel bead were loaded with poly(dT) primers that include a 16 base cell barcode and a 10 base unique molecular index (UMI). Reverse transcription reactions were engaged to generate barcoded full-length cDNA followed by the disruption of emulsions using the recovery agent and cDNA clean up with DynaBeads MyOne Silane Beads (Thermo Fisher Scientific, Cat# 37002D). Bulk cDNA was amplified, and indexed sequencing libraries were constructed using the reagents from the Chromium Single Cell 3′ v2 Reagent Kit. Libraries were sequenced on NovaSeq 6000 Sequencing System (Illumina Cambridge).</p></sec><sec id="sec4.4.14"><title>Single-cell RNaseq data processing and analysis</title><p id="p0345">Cell Ranger (version 2.1.1) (from 10x genomics) was used to process Chromium single cell 3′ v2 RNA-seq output files. First, we generated fastq files for the Read1 for cell barcode and UMI and Read2 for transcript applying cellranger mkfastq (with default parameters). Second, we aligned the Read2 to the human reference genome GRCh38 using STAR (version 2.5.1) with cell ranger count (with default parameters) (<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="https://support.10xgenomics.com/single-cell-gene-expression/software/pipelines/latest/using/mkfastq" id="intref0410">https://support.10xgenomics.com/single-cell-gene-expression/software/pipelines/latest/using/mkfastq</ext-link>). Further analysis was performed using Seurat package (version 2.3.4) in R (version 3.4.0) (<xref rid="bib10" ref-type="bibr">Butler et al., 2018</xref>). Before performed analysis, we applied the following filtering step: only genes expressed in 3 or more cells have been preserved and cells with less 200 unique genes and more than 4000 unique expressed genes were discarded (as they are potentially cells doublets). After filtering step, we used an expression matrix resulting in 14933 genes across 1622 cells (among 1625 cells) for the rest of the analysis. The matrix was normalized using genes expression values for each cell were divided by the total number of transcripts and multiplied by median of UMI counts. Then, these values were natural log-transformed before downstream analysis. For dimensionality reduction analysis, we first identified 3288 genes as highly variable genes across the single cells (cutoff value for dispersion = 0.5; cutoff value for average expression = 0). PCA performed using the variable genes as input and determined 10 PCs as significant PCs. These principal components were used as input for t-Distributed Stochastic Neighbor Embedding (tSNE) (<xref rid="bib57" ref-type="bibr">van der Maaten, 2008</xref>). We used the shared nearest neighbor (SNN) modularity optimization-based clustering algorithm from the Seurat package (FindClusters function with default parameters) to identify the clusters of cells, following by Clustree analysis (clustree R package, version 0.2.2) by changing the resolution parameters from 0 to 2. Finally, we kept a resolution parameter at 0.8 and defined distinct 9 clusters. After controlling expression of some quality control genes, we excluded clusters E, F, G, H and I as contamination. Finally, we identified 4 relevant clusters. We identified cell specific marker by comparing cells in a specific cluster with cells in all other clusters using FindAllMarkers from Seurat package (MAST; logFC threshold = 0.5; only positive markers). Heatmap, feature plots and violin plots were performed using Seurat package.</p></sec><sec id="sec4.4.15"><title>Signature expression analysis</title><p id="p0350">Single-cell RNA sequencing matrix for clusters A, B, C and D was created with gene signatures from <xref rid="bib58" ref-type="bibr">Villani et al. (2017)</xref>. Signatures were defined as mean expression of discriminative markers for cDC2s and DC3s among lin<sup>-</sup>CD14<sup>-</sup> cells (cDC2 enriched and DC3 enriched, respectively) and of discriminative markers for CD14<sup>+</sup> monocytes within lin<sup>-</sup>CD14<sup>+/−</sup> and/or CD16<sup>+/−</sup> monocytes (CD14<sup>+</sup> mono enriched). The average expression of signature genes for each cell was calculated and plotted as a violin plot using R package ggplot2 (version 3.1.0).</p></sec><sec id="sec4.4.16"><title>Gene set enrichment analysis</title><p id="p0355">To statistically evaluate the enrichment of previously reported gene signatures (Gene Sets) (<xref rid="bib21" ref-type="bibr">Hombrink et al., 2016</xref>; <xref rid="bib24" ref-type="bibr">Kumar et al., 2017</xref>; <xref rid="bib49" ref-type="bibr">Savas et al., 2018</xref>; <xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>) in our dataset, we used pairwise comparisons using the gene set enrichment analysis (GSEA) (<xref rid="bib56" ref-type="bibr">Subramanian et al., 2005</xref>) method from the Massachussets Institute of Technology (<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="https://www.broadinstitute.org/gsea" id="intref0415">https://www.broadinstitute.org/gsea</ext-link>). GSEA tests the relative position of a collection of genes (Gene Set) within an independent, ranked dataset (GeneList). Statistical analysis was performed by evaluation of nominal p value and false discovery rate (q value) based on 1,000 random permutations. Results were considered significant when the p value was below 0.05 and when the q value was below 0.25 (false discovery rate below 25%) accordingly to the recommendation from the software developers. For each pairwise comparison, the GSEA output can be represented as a bar code where each bar corresponds to the projection of one Probe Sets of the Gene Sets on the list of all the Probe Sets of the gene chips ranked from those having a high signal in one cell population (represented in red) to those having a high signal in the other cell population (represented in blue). Each bar code can be characterized by two parameters. The normalized enrichment score (NES) represents the number and differential expression intensity of the genes enriched in the corresponding cell subset. The NES is positive if the Gene Set is enriched in the first cell population and negative if it is enriched in the second cell population. The false discovery rate (FDR) statistical value (q) represents the likelihood that the enrichment of the Gene Set represents a false positive finding (for example if q = 0.25, 25% of the Gene Sets found enriched can be false positives). An absolute value of the NES below or around 1 means no enrichment as confirmed with associated q-values above 0.25.</p><p id="p0360">To simultaneously visualize pairwise comparisons of transcriptomes from cord blood-derived DC2s (light and dark blue) DC3s (orange) and macrophages (gray and brown), the BubbleMap module of BubbleGum (<xref rid="bib54" ref-type="bibr">Spinelli et al., 2015</xref>) was used. Results were considered significant when the p value was below 0.05 and the FDR (false discovery rate, q) value was below 0.25. The BubbleGum was performed using previously published gene signatures of pairwise comparison between DC2s and DC3s and DC3s and CD14<sup>+</sup> monocytes (<xref rid="bib58" ref-type="bibr">Villani et al., 2017</xref>).</p></sec><sec id="sec4.4.17"><title>Human tissue processing and cell suspension</title><p id="p0365">Tumor-draining lymph nodes (tdLN) and primary tumors were collected in CO<sub>2</sub> independent medium (GIBCO; Cat# 18045-054) within few hours after the primary surgery. Tissue were cut into small fragments and submitted to enzymatic digestion using 0.1 mg/ml of Liberase TL (Roche) and 0.1 mg/ml of DNase (Roche) for 30 min. Cells were filtered on 40-μm cell strainer (BD), washed using CO<sub>2</sub> independent medium (GIBCO; Cat# 18045-054) containing 0.4g/ml of human albumin and resuspended for cell counting.</p></sec><sec id="sec4.4.18"><title>Cell sorting of myeloid subsets from patients’ tdLN for RNA sequencing</title><p id="p0370">After tissue processing, cells obtained from tdLN were submitted to myeloid cells enrichment accordingly to (<xref rid="bib12" ref-type="bibr">Durand and Segura, 2016</xref>) prior flow cytometry-sorting. In brief, T and B lymphocytes, NK cells, erythrocytes and myelomonocytic cells were depleted using monoclonal antibodies against: CD3, CD19, CD56, CD235a and CD15, respectively. Subsequently, cell suspensions were stained for 30 min with antibody-conjugated as the following: HLA-DR, CD11c, CD14, CD1c, CD304, CD1a, CD206, CD141. Around 1,000 cells of each DC subset were sorted by flow cytometry using BD FACS ARIA II cell sorter, (purity &gt; 98%). Cells were centrifuged and lysed with TCL buffer (QIAGEN) containing 1% of beta-mercaptoethanol before storage at −80°C. RNA were extracted and isolated using the Single Cell RNA purification kit (Norgen, Cat#51800) according to the manufacturer’s instructions and the RNA integrity number was evaluated with an Agilent RNA 6000 pico kit.</p></sec><sec id="sec4.4.19"><title>LEGENDplex™ assay</title><p id="p0375">PBMCs and <italic toggle="yes">in vitro</italic> generated cells were stained and sorted as described previously. In total 3 × 10<sup>5</sup> of blood cell subsets (cDC2, DC3 and Mono) or 10<sup>5</sup>
+<italic toggle="yes">in vitro</italic> generated cells (cDC2-, DC3 and Macro-like) were flow cytometry-sorted and cultured with TLR agonists cocktail containing 25 μg/ml Poly I:C (InvivoGen, Cat# 31852-29-6), 1 μg/ml R848 (InvivoGen, Cat# 144875-48-9) and 10ng/ml LPS (Sigma-Aldrich, Cat# L2630) for 16 h. Culturing supernatants were collected and stored at −20°C until the LEGENDplex™ assay execution day. LEGENDplex Human Macrophage/Microglia Panel (13-plex) with V-bottom Plate (Biolegend Cat# 740503) and Human Proinflammatory Chemokine Panel (13-plex) with V-bottom Plate (Biolegend Cat# 740003) was used according to manufacturer’s instructions. In short, samples and standards were thawed and plated with capture beads and incubated for 2 h. Plate was then washed, and Detection Antibodies were added. After 1 h incubation SA-PE was added and incubated for 30 min. Samples were acquired on BD FACSCanto II. Samples were analyzed using LEGENDplex™ Data Analysis Software.</p></sec><sec id="sec4.4.20"><title><italic toggle="yes">In vitro</italic> T cell assay</title><p id="p0380">PBMCs and <italic toggle="yes">in vitro</italic> generated cells were stained and flow cytometry-sorted as described previously. 10<sup>4</sup> mononuclear phagocytes were cultured with TLR agonists cocktail containing 25 μg/ml PolyI:C, 1 μg/ml R848 and 10 ng/ml LPS for 16 h. T cells were isolated from fresh or frozen PBMCs using Naive Pan T Cell Isolation Kit (Miltenyi, Cat# 130-097-095) and 10<sup>5</sup> cells were plated on top of mononuclear phagocytes in presence of CytoStim (Miltenyi, Cat# 130-092-172), according to manufacturer’s instructions. For mixed leukocyte reaction (MLR) experiments, isolated Naive Pan T cells were labeled with Cell Tracer Violet (CTV) (Thermo Fischer, Cat# <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="pmc:entrez-nucleotide" xlink:href="C34557">C34557</ext-link>) as per manufacturer’s instructions and cultured with flow cytometry-sorted DCs without CytoStim. At day 5, cells were collected and stained for extracellular and intracellular marker expression and analyzed using BD LSRFortessa. For intracellular staining, fixation and permeabilization were performed using BD Cytofix/Cytoperm solution (BD Biosciences, Cat# 554714) according to manufacturer’s instructions.</p></sec></sec><sec id="sec4.5"><title>Quantification and Statistical Analysis</title><p id="p0385">Statistical analysis was performed using Prism 8.3 (GraphPad Software Inc., USA). When two experimental groups were compared, non-parametric Mann-Whitney test was used. When three or more groups were compared, statistically significant differences between means were determined using the one-way or two-way analysis of variance (ANOVA) method. A p value of less than 0.05 was considered as significant.</p></sec></sec></body><back><ref-list id="cebib0010"><title>References</title><ref id="bib1"><element-citation publication-type="journal" id="sref1"><person-group person-group-type="author"><name name-style="western"><surname>Acosta-Rodriguez</surname><given-names>E.V.</given-names></name><name name-style="western"><surname>Napolitani</surname><given-names>G.</given-names></name><name name-style="western"><surname>Lanzavecchia</surname><given-names>A.</given-names></name><name name-style="western"><surname>Sallusto</surname><given-names>F.</given-names></name></person-group><article-title>Interleukins 1beta and 6 but not transforming growth factor-beta are essential for the differentiation of interleukin 17-producing human T helper cells</article-title><source>Nat. Immunol.</source><volume>8</volume><year>2007</year><fpage>942</fpage><lpage>949</lpage><pub-id pub-id-type="pmid">17676045</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/ni1496</pub-id></element-citation></ref><ref id="bib2"><element-citation publication-type="journal" id="sref2"><person-group person-group-type="author"><name name-style="western"><surname>Anselmi</surname><given-names>G.</given-names></name><name name-style="western"><surname>Vaivode</surname><given-names>K.</given-names></name><name name-style="western"><surname>Dutertre</surname><given-names>C.A.</given-names></name><name name-style="western"><surname>Bourdely</surname><given-names>P.</given-names></name><name name-style="western"><surname>Missolo-Koussou</surname><given-names>Y.</given-names></name><name name-style="western"><surname>Newell</surname><given-names>E.</given-names></name><name name-style="western"><surname>Hickman</surname><given-names>O.</given-names></name><name name-style="western"><surname>Wood</surname><given-names>K.</given-names></name><name name-style="western"><surname>Saxena</surname><given-names>A.</given-names></name><name name-style="western"><surname>Helft</surname><given-names>J.</given-names></name></person-group><article-title>Engineered niches support the development of human dendritic cells in humanized mice</article-title><source>Nat. Commun.</source><volume>11</volume><year>2020</year><fpage>2054</fpage><pub-id pub-id-type="pmid">32345968</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41467-020-15937-y</pub-id><pub-id pub-id-type="pmcid">PMC7189247</pub-id></element-citation></ref><ref id="bib3"><element-citation publication-type="journal" id="sref3"><person-group person-group-type="author"><name name-style="western"><surname>Bachem</surname><given-names>A.</given-names></name><name name-style="western"><surname>Güttler</surname><given-names>S.</given-names></name><name name-style="western"><surname>Hartung</surname><given-names>E.</given-names></name><name name-style="western"><surname>Ebstein</surname><given-names>F.</given-names></name><name name-style="western"><surname>Schaefer</surname><given-names>M.</given-names></name><name name-style="western"><surname>Tannert</surname><given-names>A.</given-names></name><name name-style="western"><surname>Salama</surname><given-names>A.</given-names></name><name name-style="western"><surname>Movassaghi</surname><given-names>K.</given-names></name><name name-style="western"><surname>Opitz</surname><given-names>C.</given-names></name><name name-style="western"><surname>Mages</surname><given-names>H.W.</given-names></name></person-group><article-title>Superior antigen cross-presentation and XCR1 expression define human CD11c+CD141+ cells as homologues of mouse CD8+ dendritic cells</article-title><source>J. Exp. Med.</source><volume>207</volume><year>2010</year><fpage>1273</fpage><lpage>1281</lpage><pub-id pub-id-type="pmid">20479115</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1084/jem.20100348</pub-id><pub-id pub-id-type="pmcid">PMC2882837</pub-id></element-citation></ref><ref id="bib4"><element-citation publication-type="journal" id="sref4"><person-group person-group-type="author"><name name-style="western"><surname>Bakdash</surname><given-names>G.</given-names></name><name name-style="western"><surname>Buschow</surname><given-names>S.I.</given-names></name><name name-style="western"><surname>Gorris</surname><given-names>M.A.</given-names></name><name name-style="western"><surname>Halilovic</surname><given-names>A.</given-names></name><name name-style="western"><surname>Hato</surname><given-names>S.V.</given-names></name><name name-style="western"><surname>Sköld</surname><given-names>A.E.</given-names></name><name name-style="western"><surname>Schreibelt</surname><given-names>G.</given-names></name><name name-style="western"><surname>Sittig</surname><given-names>S.P.</given-names></name><name name-style="western"><surname>Torensma</surname><given-names>R.</given-names></name><name name-style="western"><surname>Duiveman-de Boer</surname><given-names>T.</given-names></name></person-group><article-title>Expansion of a BDCA1+CD14+ Myeloid Cell Population in Melanoma Patients May Attenuate the Efficacy of Dendritic Cell Vaccines</article-title><source>Cancer Res.</source><volume>76</volume><year>2016</year><fpage>4332</fpage><lpage>4346</lpage><pub-id pub-id-type="pmid">27325645</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1158/0008-5472.CAN-15-1695</pub-id></element-citation></ref><ref id="bib5"><element-citation publication-type="journal" id="sref5"><person-group person-group-type="author"><name name-style="western"><surname>Banchereau</surname><given-names>J.</given-names></name><name name-style="western"><surname>Steinman</surname><given-names>R.M.</given-names></name></person-group><article-title>Dendritic cells and the control of immunity</article-title><source>Nature</source><volume>392</volume><year>1998</year><fpage>245</fpage><lpage>252</lpage><pub-id pub-id-type="pmid">9521319</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/32588</pub-id></element-citation></ref><ref id="bib6"><element-citation publication-type="journal" id="sref6"><person-group person-group-type="author"><name name-style="western"><surname>Binnewies</surname><given-names>M.</given-names></name><name name-style="western"><surname>Mujal</surname><given-names>A.M.</given-names></name><name name-style="western"><surname>Pollack</surname><given-names>J.L.</given-names></name><name name-style="western"><surname>Combes</surname><given-names>A.J.</given-names></name><name name-style="western"><surname>Hardison</surname><given-names>E.A.</given-names></name><name name-style="western"><surname>Barry</surname><given-names>K.C.</given-names></name><name name-style="western"><surname>Tsui</surname><given-names>J.</given-names></name><name name-style="western"><surname>Ruhland</surname><given-names>M.K.</given-names></name><name name-style="western"><surname>Kersten</surname><given-names>K.</given-names></name><name name-style="western"><surname>Abushawish</surname><given-names>M.A.</given-names></name></person-group><article-title>Unleashing Type-2 Dendritic Cells to Drive Protective Antitumor CD4(+) T Cell Immunity</article-title><source>Cell</source><volume>177</volume><year>2019</year><fpage>556</fpage><lpage>571.e16</lpage><pub-id pub-id-type="pmid">30955881</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.cell.2019.02.005</pub-id><pub-id pub-id-type="pmcid">PMC6954108</pub-id></element-citation></ref><ref id="bib7"><element-citation publication-type="journal" id="sref7"><person-group person-group-type="author"><name name-style="western"><surname>Breton</surname><given-names>G.</given-names></name><name name-style="western"><surname>Lee</surname><given-names>J.</given-names></name><name name-style="western"><surname>Zhou</surname><given-names>Y.J.</given-names></name><name name-style="western"><surname>Schreiber</surname><given-names>J.J.</given-names></name><name name-style="western"><surname>Keler</surname><given-names>T.</given-names></name><name name-style="western"><surname>Puhr</surname><given-names>S.</given-names></name><name name-style="western"><surname>Anandasabapathy</surname><given-names>N.</given-names></name><name name-style="western"><surname>Schlesinger</surname><given-names>S.</given-names></name><name name-style="western"><surname>Caskey</surname><given-names>M.</given-names></name><name name-style="western"><surname>Liu</surname><given-names>K.</given-names></name><name name-style="western"><surname>Nussenzweig</surname><given-names>M.C.</given-names></name></person-group><article-title>Circulating precursors of human CD1c+ and CD141+ dendritic cells</article-title><source>J. Exp. Med.</source><volume>212</volume><year>2015</year><fpage>401</fpage><lpage>413</lpage><pub-id pub-id-type="pmid">25687281</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1084/jem.20141441</pub-id><pub-id pub-id-type="pmcid">PMC4354370</pub-id></element-citation></ref><ref id="bib8"><element-citation publication-type="journal" id="sref8"><person-group person-group-type="author"><name name-style="western"><surname>Breton</surname><given-names>G.</given-names></name><name name-style="western"><surname>Zheng</surname><given-names>S.</given-names></name><name name-style="western"><surname>Valieris</surname><given-names>R.</given-names></name><name name-style="western"><surname>Tojal da Silva</surname><given-names>I.</given-names></name><name name-style="western"><surname>Satija</surname><given-names>R.</given-names></name><name name-style="western"><surname>Nussenzweig</surname><given-names>M.C.</given-names></name></person-group><article-title>Human dendritic cells (DCs) are derived from distinct circulating precursors that are precommitted to become CD1c+ or CD141+ DCs</article-title><source>J. Exp. Med.</source><volume>213</volume><year>2016</year><fpage>2861</fpage><lpage>2870</lpage><pub-id pub-id-type="pmid">27864467</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1084/jem.20161135</pub-id><pub-id pub-id-type="pmcid">PMC5154947</pub-id></element-citation></ref><ref id="bib9"><element-citation publication-type="journal" id="sref9"><person-group person-group-type="author"><name name-style="western"><surname>Briseño</surname><given-names>C.G.</given-names></name><name name-style="western"><surname>Haldar</surname><given-names>M.</given-names></name><name name-style="western"><surname>Kretzer</surname><given-names>N.M.</given-names></name><name name-style="western"><surname>Wu</surname><given-names>X.</given-names></name><name name-style="western"><surname>Theisen</surname><given-names>D.J.</given-names></name><name name-style="western"><surname>Kc</surname><given-names>W.</given-names></name><name name-style="western"><surname>Durai</surname><given-names>V.</given-names></name><name name-style="western"><surname>Grajales-Reyes</surname><given-names>G.E.</given-names></name><name name-style="western"><surname>Iwata</surname><given-names>A.</given-names></name><name name-style="western"><surname>Bagadia</surname><given-names>P.</given-names></name></person-group><article-title>Distinct Transcriptional Programs Control Cross-Priming in Classical and Monocyte-Derived Dendritic Cells</article-title><source>Cell Rep.</source><volume>15</volume><year>2016</year><fpage>2462</fpage><lpage>2474</lpage><pub-id pub-id-type="pmid">27264183</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.celrep.2016.05.025</pub-id><pub-id pub-id-type="pmcid">PMC4941620</pub-id></element-citation></ref><ref id="bib10"><element-citation publication-type="journal" id="sref10"><person-group person-group-type="author"><name name-style="western"><surname>Butler</surname><given-names>A.</given-names></name><name name-style="western"><surname>Hoffman</surname><given-names>P.</given-names></name><name name-style="western"><surname>Smibert</surname><given-names>P.</given-names></name><name name-style="western"><surname>Papalexi</surname><given-names>E.</given-names></name><name name-style="western"><surname>Satija</surname><given-names>R.</given-names></name></person-group><article-title>Integrating single-cell transcriptomic data across different conditions, technologies, and species</article-title><source>Nat. Biotechnol.</source><volume>36</volume><year>2018</year><fpage>411</fpage><lpage>420</lpage><pub-id pub-id-type="pmid">29608179</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nbt.4096</pub-id><pub-id pub-id-type="pmcid">PMC6700744</pub-id></element-citation></ref><ref id="bib11"><element-citation publication-type="journal" id="sref11"><person-group person-group-type="author"><name name-style="western"><surname>Crozat</surname><given-names>K.</given-names></name><name name-style="western"><surname>Guiton</surname><given-names>R.</given-names></name><name name-style="western"><surname>Contreras</surname><given-names>V.</given-names></name><name name-style="western"><surname>Feuillet</surname><given-names>V.</given-names></name><name name-style="western"><surname>Dutertre</surname><given-names>C.A.</given-names></name><name name-style="western"><surname>Ventre</surname><given-names>E.</given-names></name><name name-style="western"><surname>Vu Manh</surname><given-names>T.P.</given-names></name><name name-style="western"><surname>Baranek</surname><given-names>T.</given-names></name><name name-style="western"><surname>Storset</surname><given-names>A.K.</given-names></name><name name-style="western"><surname>Marvel</surname><given-names>J.</given-names></name></person-group><article-title>The XC chemokine receptor 1 is a conserved selective marker of mammalian cells homologous to mouse CD8alpha+ dendritic cells</article-title><source>J. Exp. Med.</source><volume>207</volume><year>2010</year><fpage>1283</fpage><lpage>1292</lpage><pub-id pub-id-type="pmid">20479118</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1084/jem.20100223</pub-id><pub-id pub-id-type="pmcid">PMC2882835</pub-id></element-citation></ref><ref id="bib12"><element-citation publication-type="journal" id="sref12"><person-group person-group-type="author"><name name-style="western"><surname>Durand</surname><given-names>M.</given-names></name><name name-style="western"><surname>Segura</surname><given-names>E.</given-names></name></person-group><article-title>Dendritic Cell Subset Purification from Human Tonsils and Lymph Nodes</article-title><source>Methods Mol. Biol.</source><volume>1423</volume><year>2016</year><fpage>89</fpage><lpage>99</lpage><pub-id pub-id-type="pmid">27142010</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1007/978-1-4939-3606-9_6</pub-id></element-citation></ref><ref id="bib13"><element-citation publication-type="journal" id="sref13"><person-group person-group-type="author"><name name-style="western"><surname>Dutertre</surname><given-names>C.A.</given-names></name><name name-style="western"><surname>Becht</surname><given-names>E.</given-names></name><name name-style="western"><surname>Irac</surname><given-names>S.E.</given-names></name><name name-style="western"><surname>Khalilnezhad</surname><given-names>A.</given-names></name><name name-style="western"><surname>Narang</surname><given-names>V.</given-names></name><name name-style="western"><surname>Khalilnezhad</surname><given-names>S.</given-names></name><name name-style="western"><surname>Ng</surname><given-names>P.Y.</given-names></name><name name-style="western"><surname>van den Hoogen</surname><given-names>L.L.</given-names></name><name name-style="western"><surname>Leong</surname><given-names>J.Y.</given-names></name><name name-style="western"><surname>Lee</surname><given-names>B.</given-names></name></person-group><article-title>Single-Cell Analysis of Human Mononuclear Phagocytes Reveals Subset-Defining Markers and Identifies Circulating Inflammatory Dendritic Cells</article-title><source>Immunity</source><volume>51</volume><year>2019</year><fpage>573</fpage><lpage>589.e8</lpage><pub-id pub-id-type="pmid">31474513</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2019.08.008</pub-id></element-citation></ref><ref id="bib14"><element-citation publication-type="journal" id="sref14"><person-group person-group-type="author"><name name-style="western"><surname>Fogg</surname><given-names>D.K.</given-names></name><name name-style="western"><surname>Sibon</surname><given-names>C.</given-names></name><name name-style="western"><surname>Miled</surname><given-names>C.</given-names></name><name name-style="western"><surname>Jung</surname><given-names>S.</given-names></name><name name-style="western"><surname>Aucouturier</surname><given-names>P.</given-names></name><name name-style="western"><surname>Littman</surname><given-names>D.R.</given-names></name><name name-style="western"><surname>Cumano</surname><given-names>A.</given-names></name><name name-style="western"><surname>Geissmann</surname><given-names>F.</given-names></name></person-group><article-title>A clonogenic bone marrow progenitor specific for macrophages and dendritic cells</article-title><source>Science</source><volume>311</volume><year>2006</year><fpage>83</fpage><lpage>87</lpage><pub-id pub-id-type="pmid">16322423</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1126/science.1117729</pub-id></element-citation></ref><ref id="bib15"><element-citation publication-type="journal" id="sref15"><person-group person-group-type="author"><name name-style="western"><surname>Goudot</surname><given-names>C.</given-names></name><name name-style="western"><surname>Coillard</surname><given-names>A.</given-names></name><name name-style="western"><surname>Villani</surname><given-names>A.C.</given-names></name><name name-style="western"><surname>Gueguen</surname><given-names>P.</given-names></name><name name-style="western"><surname>Cros</surname><given-names>A.</given-names></name><name name-style="western"><surname>Sarkizova</surname><given-names>S.</given-names></name><name name-style="western"><surname>Tang-Huau</surname><given-names>T.L.</given-names></name><name name-style="western"><surname>Bohec</surname><given-names>M.</given-names></name><name name-style="western"><surname>Baulande</surname><given-names>S.</given-names></name><name name-style="western"><surname>Hacohen</surname><given-names>N.</given-names></name></person-group><article-title>Aryl Hydrocarbon Receptor Controls Monocyte Differentiation into Dendritic Cells versus Macrophages</article-title><source>Immunity</source><volume>47</volume><year>2017</year><fpage>582</fpage><lpage>596.e6</lpage><pub-id pub-id-type="pmid">28930664</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2017.08.016</pub-id></element-citation></ref><ref id="bib16"><element-citation publication-type="journal" id="sref16"><person-group person-group-type="author"><name name-style="western"><surname>Granot</surname><given-names>T.</given-names></name><name name-style="western"><surname>Senda</surname><given-names>T.</given-names></name><name name-style="western"><surname>Carpenter</surname><given-names>D.J.</given-names></name><name name-style="western"><surname>Matsuoka</surname><given-names>N.</given-names></name><name name-style="western"><surname>Weiner</surname><given-names>J.</given-names></name><name name-style="western"><surname>Gordon</surname><given-names>C.L.</given-names></name><name name-style="western"><surname>Miron</surname><given-names>M.</given-names></name><name name-style="western"><surname>Kumar</surname><given-names>B.V.</given-names></name><name name-style="western"><surname>Griesemer</surname><given-names>A.</given-names></name><name name-style="western"><surname>Ho</surname><given-names>S.H.</given-names></name></person-group><article-title>Dendritic Cells Display Subset and Tissue-Specific Maturation Dynamics over Human Life</article-title><source>Immunity</source><volume>46</volume><year>2017</year><fpage>504</fpage><lpage>515</lpage><pub-id pub-id-type="pmid">28329707</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2017.02.019</pub-id><pub-id pub-id-type="pmcid">PMC5415308</pub-id></element-citation></ref><ref id="bib17"><element-citation publication-type="journal" id="sref17"><person-group person-group-type="author"><name name-style="western"><surname>Guermonprez</surname><given-names>P.</given-names></name><name name-style="western"><surname>Helft</surname><given-names>J.</given-names></name><name name-style="western"><surname>Claser</surname><given-names>C.</given-names></name><name name-style="western"><surname>Deroubaix</surname><given-names>S.</given-names></name><name name-style="western"><surname>Karanje</surname><given-names>H.</given-names></name><name name-style="western"><surname>Gazumyan</surname><given-names>A.</given-names></name><name name-style="western"><surname>Darasse-Jèze</surname><given-names>G.</given-names></name><name name-style="western"><surname>Telerman</surname><given-names>S.B.</given-names></name><name name-style="western"><surname>Breton</surname><given-names>G.</given-names></name><name name-style="western"><surname>Schreiber</surname><given-names>H.A.</given-names></name></person-group><article-title>Inflammatory Flt3l is essential to mobilize dendritic cells and for T cell responses during Plasmodium infection</article-title><source>Nat. Med.</source><volume>19</volume><year>2013</year><fpage>730</fpage><lpage>738</lpage><pub-id pub-id-type="pmid">23685841</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nm.3197</pub-id><pub-id pub-id-type="pmcid">PMC3914008</pub-id></element-citation></ref><ref id="bib18"><element-citation publication-type="journal" id="sref18"><person-group person-group-type="author"><name name-style="western"><surname>Guermonprez</surname><given-names>P.</given-names></name><name name-style="western"><surname>Gerber-Ferder</surname><given-names>Y.</given-names></name><name name-style="western"><surname>Vaivode</surname><given-names>K.</given-names></name><name name-style="western"><surname>Bourdely</surname><given-names>P.</given-names></name><name name-style="western"><surname>Helft</surname><given-names>J.</given-names></name></person-group><article-title>Origin and development of classical dendritic cells</article-title><source>Int. Rev. Cell Mol. Biol.</source><volume>349</volume><year>2019</year><fpage>1</fpage><lpage>54</lpage><pub-id pub-id-type="pmid">31759429</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/bs.ircmb.2019.08.002</pub-id></element-citation></ref><ref id="bib19"><element-citation publication-type="journal" id="sref19"><person-group person-group-type="author"><name name-style="western"><surname>Heidkamp</surname><given-names>G.F.</given-names></name><name name-style="western"><surname>Sander</surname><given-names>J.</given-names></name><name name-style="western"><surname>Lehmann</surname><given-names>C.H.K.</given-names></name><name name-style="western"><surname>Heger</surname><given-names>L.</given-names></name><name name-style="western"><surname>Eissing</surname><given-names>N.</given-names></name><name name-style="western"><surname>Baranska</surname><given-names>A.</given-names></name><name name-style="western"><surname>Lühr</surname><given-names>J.J.</given-names></name><name name-style="western"><surname>Hoffmann</surname><given-names>A.</given-names></name><name name-style="western"><surname>Reimer</surname><given-names>K.C.</given-names></name><name name-style="western"><surname>Lux</surname><given-names>A.</given-names></name></person-group><article-title>Human lymphoid organ dendritic cell identity is predominantly dictated by ontogeny, not tissue microenvironment</article-title><source>Sci. Immunol.</source><volume>1</volume><year>2016</year><comment>eaai7677</comment><pub-id pub-id-type="doi" assigning-authority="pmc">10.1126/sciimmunol.aai7677</pub-id><pub-id pub-id-type="pmid">28783692</pub-id></element-citation></ref><ref id="bib20"><element-citation publication-type="journal" id="sref20"><person-group person-group-type="author"><name name-style="western"><surname>Helft</surname><given-names>J.</given-names></name><name name-style="western"><surname>Anjos-Afonso</surname><given-names>F.</given-names></name><name name-style="western"><surname>van der Veen</surname><given-names>A.G.</given-names></name><name name-style="western"><surname>Chakravarty</surname><given-names>P.</given-names></name><name name-style="western"><surname>Bonnet</surname><given-names>D.</given-names></name><name name-style="western"><surname>Reis e Sousa</surname><given-names>C.</given-names></name></person-group><article-title>Dendritic Cell Lineage Potential in Human Early Hematopoietic Progenitors</article-title><source>Cell Rep.</source><volume>20</volume><year>2017</year><fpage>529</fpage><lpage>537</lpage><pub-id pub-id-type="pmid">28723558</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.celrep.2017.06.075</pub-id><pub-id pub-id-type="pmcid">PMC5529209</pub-id></element-citation></ref><ref id="bib21"><element-citation publication-type="journal" id="sref21"><person-group person-group-type="author"><name name-style="western"><surname>Hombrink</surname><given-names>P.</given-names></name><name name-style="western"><surname>Helbig</surname><given-names>C.</given-names></name><name name-style="western"><surname>Backer</surname><given-names>R.A.</given-names></name><name name-style="western"><surname>Piet</surname><given-names>B.</given-names></name><name name-style="western"><surname>Oja</surname><given-names>A.E.</given-names></name><name name-style="western"><surname>Stark</surname><given-names>R.</given-names></name><name name-style="western"><surname>Brasser</surname><given-names>G.</given-names></name><name name-style="western"><surname>Jongejan</surname><given-names>A.</given-names></name><name name-style="western"><surname>Jonkers</surname><given-names>R.E.</given-names></name><name name-style="western"><surname>Nota</surname><given-names>B.</given-names></name></person-group><article-title>Programs for the persistence, vigilance and control of human CD8<sup>+</sup> lung-resident memory T cells</article-title><source>Nat. Immunol.</source><volume>17</volume><year>2016</year><fpage>1467</fpage><lpage>1478</lpage><pub-id pub-id-type="pmid">27776108</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/ni.3589</pub-id></element-citation></ref><ref id="bib22"><element-citation publication-type="journal" id="sref22"><person-group person-group-type="author"><name name-style="western"><surname>Jongbloed</surname><given-names>S.L.</given-names></name><name name-style="western"><surname>Kassianos</surname><given-names>A.J.</given-names></name><name name-style="western"><surname>McDonald</surname><given-names>K.J.</given-names></name><name name-style="western"><surname>Clark</surname><given-names>G.J.</given-names></name><name name-style="western"><surname>Ju</surname><given-names>X.</given-names></name><name name-style="western"><surname>Angel</surname><given-names>C.E.</given-names></name><name name-style="western"><surname>Chen</surname><given-names>C.J.</given-names></name><name name-style="western"><surname>Dunbar</surname><given-names>P.R.</given-names></name><name name-style="western"><surname>Wadley</surname><given-names>R.B.</given-names></name><name name-style="western"><surname>Jeet</surname><given-names>V.</given-names></name></person-group><article-title>Human CD141+ (BDCA-3)+ dendritic cells (DCs) represent a unique myeloid DC subset that cross-presents necrotic cell antigens</article-title><source>J. Exp. Med.</source><volume>207</volume><year>2010</year><fpage>1247</fpage><lpage>1260</lpage><pub-id pub-id-type="pmid">20479116</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1084/jem.20092140</pub-id><pub-id pub-id-type="pmcid">PMC2882828</pub-id></element-citation></ref><ref id="bib23"><element-citation publication-type="journal" id="sref23"><person-group person-group-type="author"><name name-style="western"><surname>Kawamura</surname><given-names>S.</given-names></name><name name-style="western"><surname>Onai</surname><given-names>N.</given-names></name><name name-style="western"><surname>Miya</surname><given-names>F.</given-names></name><name name-style="western"><surname>Sato</surname><given-names>T.</given-names></name><name name-style="western"><surname>Tsunoda</surname><given-names>T.</given-names></name><name name-style="western"><surname>Kurabayashi</surname><given-names>K.</given-names></name><name name-style="western"><surname>Yotsumoto</surname><given-names>S.</given-names></name><name name-style="western"><surname>Kuroda</surname><given-names>S.</given-names></name><name name-style="western"><surname>Takenaka</surname><given-names>K.</given-names></name><name name-style="western"><surname>Akashi</surname><given-names>K.</given-names></name><name name-style="western"><surname>Ohteki</surname><given-names>T.</given-names></name></person-group><article-title>Identification of a Human Clonogenic Progenitor with Strict Monocyte Differentiation Potential: A Counterpart of Mouse cMoPs</article-title><source>Immunity</source><volume>46</volume><year>2017</year><fpage>835</fpage><lpage>848 e4</lpage><pub-id pub-id-type="pmid">28514689</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2017.04.019</pub-id></element-citation></ref><ref id="bib24"><element-citation publication-type="journal" id="sref24"><person-group person-group-type="author"><name name-style="western"><surname>Kumar</surname><given-names>B.V.</given-names></name><name name-style="western"><surname>Ma</surname><given-names>W.</given-names></name><name name-style="western"><surname>Miron</surname><given-names>M.</given-names></name><name name-style="western"><surname>Granot</surname><given-names>T.</given-names></name><name name-style="western"><surname>Guyer</surname><given-names>R.S.</given-names></name><name name-style="western"><surname>Carpenter</surname><given-names>D.J.</given-names></name><name name-style="western"><surname>Senda</surname><given-names>T.</given-names></name><name name-style="western"><surname>Sun</surname><given-names>X.</given-names></name><name name-style="western"><surname>Ho</surname><given-names>S.H.</given-names></name><name name-style="western"><surname>Lerner</surname><given-names>H.</given-names></name></person-group><article-title>Human Tissue-Resident Memory T Cells Are Defined by Core Transcriptional and Functional Signatures in Lymphoid and Mucosal Sites</article-title><source>Cell Rep.</source><volume>20</volume><year>2017</year><fpage>2921</fpage><lpage>2934</lpage><pub-id pub-id-type="pmid">28930685</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.celrep.2017.08.078</pub-id><pub-id pub-id-type="pmcid">PMC5646692</pub-id></element-citation></ref><ref id="bib25"><element-citation publication-type="journal" id="sref25"><person-group person-group-type="author"><name name-style="western"><surname>Lavin</surname><given-names>Y.</given-names></name><name name-style="western"><surname>Kobayashi</surname><given-names>S.</given-names></name><name name-style="western"><surname>Leader</surname><given-names>A.</given-names></name><name name-style="western"><surname>Amir</surname><given-names>E.D.</given-names></name><name name-style="western"><surname>Elefant</surname><given-names>N.</given-names></name><name name-style="western"><surname>Bigenwald</surname><given-names>C.</given-names></name><name name-style="western"><surname>Remark</surname><given-names>R.</given-names></name><name name-style="western"><surname>Sweeney</surname><given-names>R.</given-names></name><name name-style="western"><surname>Becker</surname><given-names>C.D.</given-names></name><name name-style="western"><surname>Levine</surname><given-names>J.H.</given-names></name></person-group><article-title>Innate Immune Landscape in Early Lung Adenocarcinoma by Paired Single-Cell Analyses</article-title><source>Cell</source><volume>169</volume><year>2017</year><fpage>750</fpage><lpage>765.e17</lpage><pub-id pub-id-type="pmid">28475900</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.cell.2017.04.014</pub-id><pub-id pub-id-type="pmcid">PMC5737939</pub-id></element-citation></ref><ref id="bib26"><element-citation publication-type="journal" id="sref26"><person-group person-group-type="author"><name name-style="western"><surname>Leal Rojas</surname><given-names>I.M.</given-names></name><name name-style="western"><surname>Mok</surname><given-names>W.H.</given-names></name><name name-style="western"><surname>Pearson</surname><given-names>F.E.</given-names></name><name name-style="western"><surname>Minoda</surname><given-names>Y.</given-names></name><name name-style="western"><surname>Kenna</surname><given-names>T.J.</given-names></name><name name-style="western"><surname>Barnard</surname><given-names>R.T.</given-names></name><name name-style="western"><surname>Radford</surname><given-names>K.J.</given-names></name></person-group><article-title>Human Blood CD1c<sup>+</sup> Dendritic Cells Promote Th1 and Th17 Effector Function in Memory CD4<sup>+</sup> T Cells</article-title><source>Front. Immunol.</source><volume>8</volume><year>2017</year><fpage>971</fpage><pub-id pub-id-type="pmid">28878767</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.3389/fimmu.2017.00971</pub-id><pub-id pub-id-type="pmcid">PMC5572390</pub-id></element-citation></ref><ref id="bib27"><element-citation publication-type="journal" id="sref27"><person-group person-group-type="author"><name name-style="western"><surname>Lee</surname><given-names>J.</given-names></name><name name-style="western"><surname>Breton</surname><given-names>G.</given-names></name><name name-style="western"><surname>Aljoufi</surname><given-names>A.</given-names></name><name name-style="western"><surname>Zhou</surname><given-names>Y.J.</given-names></name><name name-style="western"><surname>Puhr</surname><given-names>S.</given-names></name><name name-style="western"><surname>Nussenzweig</surname><given-names>M.C.</given-names></name><name name-style="western"><surname>Liu</surname><given-names>K.</given-names></name></person-group><article-title>Clonal analysis of human dendritic cell progenitor using a stromal cell culture</article-title><source>J. Immunol. Methods</source><volume>425</volume><year>2015</year><fpage>21</fpage><lpage>26</lpage><pub-id pub-id-type="pmid">26056939</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.jim.2015.06.004</pub-id><pub-id pub-id-type="pmcid">PMC4604063</pub-id></element-citation></ref><ref id="bib28"><element-citation publication-type="journal" id="sref28"><person-group person-group-type="author"><name name-style="western"><surname>Lee</surname><given-names>J.</given-names></name><name name-style="western"><surname>Breton</surname><given-names>G.</given-names></name><name name-style="western"><surname>Oliveira</surname><given-names>T.Y.</given-names></name><name name-style="western"><surname>Zhou</surname><given-names>Y.J.</given-names></name><name name-style="western"><surname>Aljoufi</surname><given-names>A.</given-names></name><name name-style="western"><surname>Puhr</surname><given-names>S.</given-names></name><name name-style="western"><surname>Cameron</surname><given-names>M.J.</given-names></name><name name-style="western"><surname>Sékaly</surname><given-names>R.P.</given-names></name><name name-style="western"><surname>Nussenzweig</surname><given-names>M.C.</given-names></name><name name-style="western"><surname>Liu</surname><given-names>K.</given-names></name></person-group><article-title>Restricted dendritic cell and monocyte progenitors in human cord blood and bone marrow</article-title><source>J. Exp. Med.</source><volume>212</volume><year>2015</year><fpage>385</fpage><lpage>399</lpage><pub-id pub-id-type="pmid">25687283</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1084/jem.20141442</pub-id><pub-id pub-id-type="pmcid">PMC4354373</pub-id></element-citation></ref><ref id="bib29"><element-citation publication-type="journal" id="sref29"><person-group person-group-type="author"><name name-style="western"><surname>Lee</surname><given-names>J.</given-names></name><name name-style="western"><surname>Zhou</surname><given-names>Y.J.</given-names></name><name name-style="western"><surname>Ma</surname><given-names>W.</given-names></name><name name-style="western"><surname>Zhang</surname><given-names>W.</given-names></name><name name-style="western"><surname>Aljoufi</surname><given-names>A.</given-names></name><name name-style="western"><surname>Luh</surname><given-names>T.</given-names></name><name name-style="western"><surname>Lucero</surname><given-names>K.</given-names></name><name name-style="western"><surname>Liang</surname><given-names>D.</given-names></name><name name-style="western"><surname>Thomsen</surname><given-names>M.</given-names></name><name name-style="western"><surname>Bhagat</surname><given-names>G.</given-names></name></person-group><article-title>Lineage specification of human dendritic cells is marked by IRF8 expression in hematopoietic stem cells and multipotent progenitors</article-title><source>Nat. Immunol.</source><volume>18</volume><year>2017</year><fpage>877</fpage><lpage>888</lpage><pub-id pub-id-type="pmid">28650480</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/ni.3789</pub-id><pub-id pub-id-type="pmcid">PMC5743223</pub-id></element-citation></ref><ref id="bib30"><element-citation publication-type="journal" id="sref30"><person-group person-group-type="author"><name name-style="western"><surname>Mach</surname><given-names>N.</given-names></name><name name-style="western"><surname>Gillessen</surname><given-names>S.</given-names></name><name name-style="western"><surname>Wilson</surname><given-names>S.B.</given-names></name><name name-style="western"><surname>Sheehan</surname><given-names>C.</given-names></name><name name-style="western"><surname>Mihm</surname><given-names>M.</given-names></name><name name-style="western"><surname>Dranoff</surname><given-names>G.</given-names></name></person-group><article-title>Differences in dendritic cells stimulated in vivo by tumors engineered to secrete granulocyte-macrophage colony-stimulating factor or Flt3-ligand</article-title><source>Cancer Res.</source><volume>60</volume><year>2000</year><fpage>3239</fpage><lpage>3246</lpage><pub-id pub-id-type="pmid">10866317</pub-id></element-citation></ref><ref id="bib31"><element-citation publication-type="journal" id="sref31"><person-group person-group-type="author"><name name-style="western"><surname>Mani</surname><given-names>V.</given-names></name><name name-style="western"><surname>Bromley</surname><given-names>S.K.</given-names></name><name name-style="western"><surname>Äijö</surname><given-names>T.</given-names></name><name name-style="western"><surname>Mora-Buch</surname><given-names>R.</given-names></name><name name-style="western"><surname>Carrizosa</surname><given-names>E.</given-names></name><name name-style="western"><surname>Warner</surname><given-names>R.D.</given-names></name><name name-style="western"><surname>Hamze</surname><given-names>M.</given-names></name><name name-style="western"><surname>Sen</surname><given-names>D.R.</given-names></name><name name-style="western"><surname>Chasse</surname><given-names>A.Y.</given-names></name><name name-style="western"><surname>Lorant</surname><given-names>A.</given-names></name></person-group><article-title>Migratory DCs activate TGF-β to precondition naïve CD8<sup>+</sup> T cells for tissue-resident memory fate</article-title><source>Science</source><volume>366</volume><year>2019</year><fpage>eaav5728</fpage><pub-id pub-id-type="pmid">31601741</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1126/science.aav5728</pub-id><pub-id pub-id-type="pmcid">PMC6939608</pub-id></element-citation></ref><ref id="bib32"><element-citation publication-type="journal" id="sref32"><person-group person-group-type="author"><name name-style="western"><surname>McKenna</surname><given-names>H.J.</given-names></name><name name-style="western"><surname>Stocking</surname><given-names>K.L.</given-names></name><name name-style="western"><surname>Miller</surname><given-names>R.E.</given-names></name><name name-style="western"><surname>Brasel</surname><given-names>K.</given-names></name><name name-style="western"><surname>De Smedt</surname><given-names>T.</given-names></name><name name-style="western"><surname>Maraskovsky</surname><given-names>E.</given-names></name><name name-style="western"><surname>Maliszewski</surname><given-names>C.R.</given-names></name><name name-style="western"><surname>Lynch</surname><given-names>D.H.</given-names></name><name name-style="western"><surname>Smith</surname><given-names>J.</given-names></name><name name-style="western"><surname>Pulendran</surname><given-names>B.</given-names></name></person-group><article-title>Mice lacking flt3 ligand have deficient hematopoiesis affecting hematopoietic progenitor cells, dendritic cells, and natural killer cells</article-title><source>Blood</source><volume>95</volume><year>2000</year><fpage>3489</fpage><lpage>3497</lpage><pub-id pub-id-type="pmid">10828034</pub-id></element-citation></ref><ref id="bib33"><element-citation publication-type="journal" id="sref33"><person-group person-group-type="author"><name name-style="western"><surname>Menezes</surname><given-names>S.</given-names></name><name name-style="western"><surname>Melandri</surname><given-names>D.</given-names></name><name name-style="western"><surname>Anselmi</surname><given-names>G.</given-names></name><name name-style="western"><surname>Perchet</surname><given-names>T.</given-names></name><name name-style="western"><surname>Loschko</surname><given-names>J.</given-names></name><name name-style="western"><surname>Dubrot</surname><given-names>J.</given-names></name><name name-style="western"><surname>Patel</surname><given-names>R.</given-names></name><name name-style="western"><surname>Gautier</surname><given-names>E.L.</given-names></name><name name-style="western"><surname>Hugues</surname><given-names>S.</given-names></name><name name-style="western"><surname>Longhi</surname><given-names>M.P.</given-names></name></person-group><article-title>The Heterogeneity of Ly6C<sup>hi</sup> Monocytes Controls Their Differentiation into iNOS<sup>+</sup> Macrophages or Monocyte-Derived Dendritic Cells</article-title><source>Immunity</source><volume>45</volume><year>2016</year><fpage>1205</fpage><lpage>1218</lpage><pub-id pub-id-type="pmid">28002729</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2016.12.001</pub-id><pub-id pub-id-type="pmcid">PMC5196026</pub-id></element-citation></ref><ref id="bib34"><element-citation publication-type="journal" id="sref34"><person-group person-group-type="author"><name name-style="western"><surname>Merad</surname><given-names>M.</given-names></name><name name-style="western"><surname>Sathe</surname><given-names>P.</given-names></name><name name-style="western"><surname>Helft</surname><given-names>J.</given-names></name><name name-style="western"><surname>Miller</surname><given-names>J.</given-names></name><name name-style="western"><surname>Mortha</surname><given-names>A.</given-names></name></person-group><article-title>The dendritic cell lineage: ontogeny and function of dendritic cells and their subsets in the steady state and the inflamed setting</article-title><source>Annu. Rev. Immunol.</source><volume>31</volume><year>2013</year><fpage>563</fpage><lpage>604</lpage><pub-id pub-id-type="pmid">23516985</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1146/annurev-immunol-020711-074950</pub-id><pub-id pub-id-type="pmcid">PMC3853342</pub-id></element-citation></ref><ref id="bib35"><element-citation publication-type="journal" id="sref35"><person-group person-group-type="author"><name name-style="western"><surname>Michea</surname><given-names>P.</given-names></name><name name-style="western"><surname>Noël</surname><given-names>F.</given-names></name><name name-style="western"><surname>Zakine</surname><given-names>E.</given-names></name><name name-style="western"><surname>Czerwinska</surname><given-names>U.</given-names></name><name name-style="western"><surname>Sirven</surname><given-names>P.</given-names></name><name name-style="western"><surname>Abouzid</surname><given-names>O.</given-names></name><name name-style="western"><surname>Goudot</surname><given-names>C.</given-names></name><name name-style="western"><surname>Scholer-Dahirel</surname><given-names>A.</given-names></name><name name-style="western"><surname>Vincent-Salomon</surname><given-names>A.</given-names></name><name name-style="western"><surname>Reyal</surname><given-names>F.</given-names></name></person-group><article-title>Adjustment of dendritic cells to the breast-cancer microenvironment is subset specific</article-title><source>Nat. Immunol.</source><volume>19</volume><year>2018</year><fpage>885</fpage><lpage>897</lpage><pub-id pub-id-type="pmid">30013147</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41590-018-0145-8</pub-id></element-citation></ref><ref id="bib36"><element-citation publication-type="journal" id="sref36"><person-group person-group-type="author"><name name-style="western"><surname>Mueller</surname><given-names>S.N.</given-names></name><name name-style="western"><surname>Mackay</surname><given-names>L.K.</given-names></name></person-group><article-title>Tissue-resident memory T cells: local specialists in immune defence</article-title><source>Nat. Rev. Immunol.</source><volume>16</volume><year>2016</year><fpage>79</fpage><lpage>89</lpage><pub-id pub-id-type="pmid">26688350</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nri.2015.3</pub-id></element-citation></ref><ref id="bib37"><element-citation publication-type="journal" id="sref37"><person-group person-group-type="author"><name name-style="western"><surname>Murphy</surname><given-names>T.L.</given-names></name><name name-style="western"><surname>Grajales-Reyes</surname><given-names>G.E.</given-names></name><name name-style="western"><surname>Wu</surname><given-names>X.</given-names></name><name name-style="western"><surname>Tussiwand</surname><given-names>R.</given-names></name><name name-style="western"><surname>Briseño</surname><given-names>C.G.</given-names></name><name name-style="western"><surname>Iwata</surname><given-names>A.</given-names></name><name name-style="western"><surname>Kretzer</surname><given-names>N.M.</given-names></name><name name-style="western"><surname>Durai</surname><given-names>V.</given-names></name><name name-style="western"><surname>Murphy</surname><given-names>K.M.</given-names></name></person-group><article-title>Transcriptional Control of Dendritic Cell Development</article-title><source>Annu. Rev. Immunol.</source><volume>34</volume><year>2016</year><fpage>93</fpage><lpage>119</lpage><pub-id pub-id-type="pmid">26735697</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1146/annurev-immunol-032713-120204</pub-id><pub-id pub-id-type="pmcid">PMC5135011</pub-id></element-citation></ref><ref id="bib38"><element-citation publication-type="journal" id="sref38"><person-group person-group-type="author"><name name-style="western"><surname>Naik</surname><given-names>S.H.</given-names></name><name name-style="western"><surname>Sathe</surname><given-names>P.</given-names></name><name name-style="western"><surname>Park</surname><given-names>H.Y.</given-names></name><name name-style="western"><surname>Metcalf</surname><given-names>D.</given-names></name><name name-style="western"><surname>Proietto</surname><given-names>A.I.</given-names></name><name name-style="western"><surname>Dakic</surname><given-names>A.</given-names></name><name name-style="western"><surname>Carotta</surname><given-names>S.</given-names></name><name name-style="western"><surname>O’Keeffe</surname><given-names>M.</given-names></name><name name-style="western"><surname>Bahlo</surname><given-names>M.</given-names></name><name name-style="western"><surname>Papenfuss</surname><given-names>A.</given-names></name></person-group><article-title>Development of plasmacytoid and conventional dendritic cell subtypes from single precursor cells derived in vitro and in vivo</article-title><source>Nat. Immunol.</source><volume>8</volume><year>2007</year><fpage>1217</fpage><lpage>1226</lpage><pub-id pub-id-type="pmid">17922015</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/ni1522</pub-id></element-citation></ref><ref id="bib39"><element-citation publication-type="journal" id="sref39"><person-group person-group-type="author"><name name-style="western"><surname>Napolitani</surname><given-names>G.</given-names></name><name name-style="western"><surname>Rinaldi</surname><given-names>A.</given-names></name><name name-style="western"><surname>Bertoni</surname><given-names>F.</given-names></name><name name-style="western"><surname>Sallusto</surname><given-names>F.</given-names></name><name name-style="western"><surname>Lanzavecchia</surname><given-names>A.</given-names></name></person-group><article-title>Selected Toll-like receptor agonist combinations synergistically trigger a T helper type 1-polarizing program in dendritic cells</article-title><source>Nat. Immunol.</source><volume>6</volume><year>2005</year><fpage>769</fpage><lpage>776</lpage><pub-id pub-id-type="pmid">15995707</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/ni1223</pub-id><pub-id pub-id-type="pmcid">PMC3760217</pub-id></element-citation></ref><ref id="bib40"><element-citation publication-type="journal" id="sref40"><person-group person-group-type="author"><name name-style="western"><surname>Nizzoli</surname><given-names>G.</given-names></name><name name-style="western"><surname>Krietsch</surname><given-names>J.</given-names></name><name name-style="western"><surname>Weick</surname><given-names>A.</given-names></name><name name-style="western"><surname>Steinfelder</surname><given-names>S.</given-names></name><name name-style="western"><surname>Facciotti</surname><given-names>F.</given-names></name><name name-style="western"><surname>Gruarin</surname><given-names>P.</given-names></name><name name-style="western"><surname>Bianco</surname><given-names>A.</given-names></name><name name-style="western"><surname>Steckel</surname><given-names>B.</given-names></name><name name-style="western"><surname>Moro</surname><given-names>M.</given-names></name><name name-style="western"><surname>Crosti</surname><given-names>M.</given-names></name></person-group><article-title>Human CD1c+ dendritic cells secrete high levels of IL-12 and potently prime cytotoxic T-cell responses</article-title><source>Blood</source><volume>122</volume><year>2013</year><fpage>932</fpage><lpage>942</lpage><pub-id pub-id-type="pmid">23794066</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1182/blood-2013-04-495424</pub-id></element-citation></ref><ref id="bib41"><element-citation publication-type="journal" id="sref41"><person-group person-group-type="author"><name name-style="western"><surname>Onai</surname><given-names>N.</given-names></name><name name-style="western"><surname>Obata-Onai</surname><given-names>A.</given-names></name><name name-style="western"><surname>Schmid</surname><given-names>M.A.</given-names></name><name name-style="western"><surname>Manz</surname><given-names>M.G.</given-names></name></person-group><article-title>Flt3 in regulation of type I interferon-producing cell and dendritic cell development</article-title><source>Ann. N Y Acad. Sci.</source><volume>1106</volume><year>2007</year><fpage>253</fpage><lpage>261</lpage><pub-id pub-id-type="pmid">17360795</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1196/annals.1392.015</pub-id></element-citation></ref><ref id="bib42"><element-citation publication-type="journal" id="sref42"><person-group person-group-type="author"><name name-style="western"><surname>Palucka</surname><given-names>K.</given-names></name><name name-style="western"><surname>Banchereau</surname><given-names>J.</given-names></name></person-group><article-title>Dendritic-cell-based therapeutic cancer vaccines</article-title><source>Immunity</source><volume>39</volume><year>2013</year><fpage>38</fpage><lpage>48</lpage><pub-id pub-id-type="pmid">23890062</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2013.07.004</pub-id><pub-id pub-id-type="pmcid">PMC3788678</pub-id></element-citation></ref><ref id="bib43"><element-citation publication-type="journal" id="sref43"><person-group person-group-type="author"><name name-style="western"><surname>Paul</surname><given-names>F.</given-names></name><name name-style="western"><surname>Arkin</surname><given-names>Y.</given-names></name><name name-style="western"><surname>Giladi</surname><given-names>A.</given-names></name><name name-style="western"><surname>Jaitin</surname><given-names>D.A.</given-names></name><name name-style="western"><surname>Kenigsberg</surname><given-names>E.</given-names></name><name name-style="western"><surname>Keren-Shaul</surname><given-names>H.</given-names></name><name name-style="western"><surname>Winter</surname><given-names>D.</given-names></name><name name-style="western"><surname>Lara-Astiaso</surname><given-names>D.</given-names></name><name name-style="western"><surname>Gury</surname><given-names>M.</given-names></name><name name-style="western"><surname>Weiner</surname><given-names>A.</given-names></name></person-group><article-title>Transcriptional Heterogeneity and Lineage Commitment in Myeloid Progenitors</article-title><source>Cell</source><volume>163</volume><year>2015</year><fpage>1663</fpage><lpage>1677</lpage><pub-id pub-id-type="pmid">26627738</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.cell.2015.11.013</pub-id></element-citation></ref><ref id="bib44"><element-citation publication-type="journal" id="sref44"><person-group person-group-type="author"><name name-style="western"><surname>Pulendran</surname><given-names>B.</given-names></name><name name-style="western"><surname>Smith</surname><given-names>J.L.</given-names></name><name name-style="western"><surname>Jenkins</surname><given-names>M.</given-names></name><name name-style="western"><surname>Schoenborn</surname><given-names>M.</given-names></name><name name-style="western"><surname>Maraskovsky</surname><given-names>E.</given-names></name><name name-style="western"><surname>Maliszewski</surname><given-names>C.R.</given-names></name></person-group><article-title>Prevention of peripheral tolerance by a dendritic cell growth factor: flt3 ligand as an adjuvant</article-title><source>J. Exp. Med.</source><volume>188</volume><year>1998</year><fpage>2075</fpage><lpage>2082</lpage><pub-id pub-id-type="pmid">9841921</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1084/jem.188.11.2075</pub-id><pub-id pub-id-type="pmcid">PMC2212387</pub-id></element-citation></ref><ref id="bib45"><element-citation publication-type="journal" id="sref45"><person-group person-group-type="author"><name name-style="western"><surname>Rihs</surname><given-names>S.</given-names></name><name name-style="western"><surname>Walker</surname><given-names>C.</given-names></name><name name-style="western"><surname>Virchow</surname><given-names>J.C.</given-names><suffix>Jr.</suffix></name><name name-style="western"><surname>Boer</surname><given-names>C.</given-names></name><name name-style="western"><surname>Kroegel</surname><given-names>C.</given-names></name><name name-style="western"><surname>Giri</surname><given-names>S.N.</given-names></name><name name-style="western"><surname>Braun</surname><given-names>R.K.</given-names></name></person-group><article-title>Differential expression of alpha E beta 7 integrins on bronchoalveolar lavage T lymphocyte subsets: regulation by alpha 4 beta 1-integrin crosslinking and TGF-beta</article-title><source>Am. J. Respir. Cell Mol. Biol.</source><volume>15</volume><year>1996</year><fpage>600</fpage><lpage>610</lpage><pub-id pub-id-type="pmid">8918367</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1165/ajrcmb.15.5.8918367</pub-id></element-citation></ref><ref id="bib46"><element-citation publication-type="journal" id="sref46"><person-group person-group-type="author"><name name-style="western"><surname>Ruffell</surname><given-names>D.</given-names></name><name name-style="western"><surname>Mourkioti</surname><given-names>F.</given-names></name><name name-style="western"><surname>Gambardella</surname><given-names>A.</given-names></name><name name-style="western"><surname>Kirstetter</surname><given-names>P.</given-names></name><name name-style="western"><surname>Lopez</surname><given-names>R.G.</given-names></name><name name-style="western"><surname>Rosenthal</surname><given-names>N.</given-names></name><name name-style="western"><surname>Nerlov</surname><given-names>C.</given-names></name></person-group><article-title>A CREB-C/EBPbeta cascade induces M2 macrophage-specific gene expression and promotes muscle injury repair</article-title><source>Proc. Natl. Acad. Sci. USA</source><volume>106</volume><year>2009</year><fpage>17475</fpage><lpage>17480</lpage><pub-id pub-id-type="pmid">19805133</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1073/pnas.0908641106</pub-id><pub-id pub-id-type="pmcid">PMC2762675</pub-id></element-citation></ref><ref id="bib47"><element-citation publication-type="journal" id="sref47"><person-group person-group-type="author"><name name-style="western"><surname>Sallusto</surname><given-names>F.</given-names></name><name name-style="western"><surname>Lanzavecchia</surname><given-names>A.</given-names></name></person-group><article-title>Efficient presentation of soluble antigen by cultured human dendritic cells is maintained by granulocyte/macrophage colony-stimulating factor plus interleukin 4 and downregulated by tumor necrosis factor alpha</article-title><source>J. Exp. Med.</source><volume>179</volume><year>1994</year><fpage>1109</fpage><lpage>1118</lpage><pub-id pub-id-type="pmid">8145033</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1084/jem.179.4.1109</pub-id><pub-id pub-id-type="pmcid">PMC2191432</pub-id></element-citation></ref><ref id="bib48"><element-citation publication-type="journal" id="sref48"><person-group person-group-type="author"><name name-style="western"><surname>Sander</surname><given-names>J.</given-names></name><name name-style="western"><surname>Schmidt</surname><given-names>S.V.</given-names></name><name name-style="western"><surname>Cirovic</surname><given-names>B.</given-names></name><name name-style="western"><surname>McGovern</surname><given-names>N.</given-names></name><name name-style="western"><surname>Papantonopoulou</surname><given-names>O.</given-names></name><name name-style="western"><surname>Hardt</surname><given-names>A.L.</given-names></name><name name-style="western"><surname>Aschenbrenner</surname><given-names>A.C.</given-names></name><name name-style="western"><surname>Kreer</surname><given-names>C.</given-names></name><name name-style="western"><surname>Quast</surname><given-names>T.</given-names></name><name name-style="western"><surname>Xu</surname><given-names>A.M.</given-names></name></person-group><article-title>Cellular Differentiation of Human Monocytes Is Regulated by Time-Dependent Interleukin-4 Signaling and the Transcriptional Regulator NCOR2</article-title><source>Immunity</source><volume>47</volume><year>2017</year><fpage>1051</fpage><lpage>1066.e12</lpage><pub-id pub-id-type="pmid">29262348</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2017.11.024</pub-id><pub-id pub-id-type="pmcid">PMC5772172</pub-id></element-citation></ref><ref id="bib49"><element-citation publication-type="journal" id="sref49"><person-group person-group-type="author"><name name-style="western"><surname>Savas</surname><given-names>P.</given-names></name><name name-style="western"><surname>Virassamy</surname><given-names>B.</given-names></name><name name-style="western"><surname>Ye</surname><given-names>C.</given-names></name><name name-style="western"><surname>Salim</surname><given-names>A.</given-names></name><name name-style="western"><surname>Mintoff</surname><given-names>C.P.</given-names></name><name name-style="western"><surname>Caramia</surname><given-names>F.</given-names></name><name name-style="western"><surname>Salgado</surname><given-names>R.</given-names></name><name name-style="western"><surname>Byrne</surname><given-names>D.J.</given-names></name><name name-style="western"><surname>Teo</surname><given-names>Z.L.</given-names></name><name name-style="western"><surname>Dushyanthen</surname><given-names>S.</given-names></name><collab>Kathleen Cuningham Foundation Consortium for Research into Familial Breast Cancer (kConFab)</collab></person-group><article-title>Single-cell profiling of breast cancer T cells reveals a tissue-resident memory subset associated with improved prognosis</article-title><source>Nat. Med.</source><volume>24</volume><year>2018</year><fpage>986</fpage><lpage>993</lpage><pub-id pub-id-type="pmid">29942092</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41591-018-0078-7</pub-id></element-citation></ref><ref id="bib50"><element-citation publication-type="journal" id="sref50"><person-group person-group-type="author"><name name-style="western"><surname>Schlitzer</surname><given-names>A.</given-names></name><name name-style="western"><surname>McGovern</surname><given-names>N.</given-names></name><name name-style="western"><surname>Teo</surname><given-names>P.</given-names></name><name name-style="western"><surname>Zelante</surname><given-names>T.</given-names></name><name name-style="western"><surname>Atarashi</surname><given-names>K.</given-names></name><name name-style="western"><surname>Low</surname><given-names>D.</given-names></name><name name-style="western"><surname>Ho</surname><given-names>A.W.</given-names></name><name name-style="western"><surname>See</surname><given-names>P.</given-names></name><name name-style="western"><surname>Shin</surname><given-names>A.</given-names></name><name name-style="western"><surname>Wasan</surname><given-names>P.S.</given-names></name></person-group><article-title>IRF4 transcription factor-dependent CD11b+ dendritic cells in human and mouse control mucosal IL-17 cytokine responses</article-title><source>Immunity</source><volume>38</volume><year>2013</year><fpage>970</fpage><lpage>983</lpage><pub-id pub-id-type="pmid">23706669</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2013.04.011</pub-id><pub-id pub-id-type="pmcid">PMC3666057</pub-id></element-citation></ref><ref id="bib51"><element-citation publication-type="journal" id="sref51"><person-group person-group-type="author"><name name-style="western"><surname>See</surname><given-names>P.</given-names></name><name name-style="western"><surname>Dutertre</surname><given-names>C.A.</given-names></name><name name-style="western"><surname>Chen</surname><given-names>J.</given-names></name><name name-style="western"><surname>Günther</surname><given-names>P.</given-names></name><name name-style="western"><surname>McGovern</surname><given-names>N.</given-names></name><name name-style="western"><surname>Irac</surname><given-names>S.E.</given-names></name><name name-style="western"><surname>Gunawan</surname><given-names>M.</given-names></name><name name-style="western"><surname>Beyer</surname><given-names>M.</given-names></name><name name-style="western"><surname>Händler</surname><given-names>K.</given-names></name><name name-style="western"><surname>Duan</surname><given-names>K.</given-names></name></person-group><article-title>Mapping the human DC lineage through the integration of high-dimensional techniques</article-title><source>Science</source><volume>356</volume><year>2017</year><comment>eaag3009</comment><pub-id pub-id-type="doi" assigning-authority="pmc">10.1126/science.aag3009</pub-id><pub-id pub-id-type="pmcid">PMC7611082</pub-id><pub-id pub-id-type="pmid">28473638</pub-id></element-citation></ref><ref id="bib52"><element-citation publication-type="journal" id="sref52"><person-group person-group-type="author"><name name-style="western"><surname>Segura</surname><given-names>E.</given-names></name><name name-style="western"><surname>Valladeau-Guilemond</surname><given-names>J.</given-names></name><name name-style="western"><surname>Donnadieu</surname><given-names>M.H.</given-names></name><name name-style="western"><surname>Sastre-Garau</surname><given-names>X.</given-names></name><name name-style="western"><surname>Soumelis</surname><given-names>V.</given-names></name><name name-style="western"><surname>Amigorena</surname><given-names>S.</given-names></name></person-group><article-title>Characterization of resident and migratory dendritic cells in human lymph nodes</article-title><source>J. Exp. Med.</source><volume>209</volume><year>2012</year><fpage>653</fpage><lpage>660</lpage><pub-id pub-id-type="pmid">22430490</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1084/jem.20111457</pub-id><pub-id pub-id-type="pmcid">PMC3328358</pub-id></element-citation></ref><ref id="bib53"><element-citation publication-type="journal" id="sref53"><person-group person-group-type="author"><name name-style="western"><surname>Segura</surname><given-names>E.</given-names></name><name name-style="western"><surname>Touzot</surname><given-names>M.</given-names></name><name name-style="western"><surname>Bohineust</surname><given-names>A.</given-names></name><name name-style="western"><surname>Cappuccio</surname><given-names>A.</given-names></name><name name-style="western"><surname>Chiocchia</surname><given-names>G.</given-names></name><name name-style="western"><surname>Hosmalin</surname><given-names>A.</given-names></name><name name-style="western"><surname>Dalod</surname><given-names>M.</given-names></name><name name-style="western"><surname>Soumelis</surname><given-names>V.</given-names></name><name name-style="western"><surname>Amigorena</surname><given-names>S.</given-names></name></person-group><article-title>Human inflammatory dendritic cells induce Th17 cell differentiation</article-title><source>Immunity</source><volume>38</volume><year>2013</year><fpage>336</fpage><lpage>348</lpage><pub-id pub-id-type="pmid">23352235</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2012.10.018</pub-id></element-citation></ref><ref id="bib54"><element-citation publication-type="journal" id="sref54"><person-group person-group-type="author"><name name-style="western"><surname>Spinelli</surname><given-names>L.</given-names></name><name name-style="western"><surname>Carpentier</surname><given-names>S.</given-names></name><name name-style="western"><surname>Montañana Sanchis</surname><given-names>F.</given-names></name><name name-style="western"><surname>Dalod</surname><given-names>M.</given-names></name><name name-style="western"><surname>Vu Manh</surname><given-names>T.P.</given-names></name></person-group><article-title>BubbleGUM: automatic extraction of phenotype molecular signatures and comprehensive visualization of multiple Gene Set Enrichment Analyses</article-title><source>BMC Genomics</source><volume>16</volume><year>2015</year><fpage>814</fpage><pub-id pub-id-type="pmid">26481321</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1186/s12864-015-2012-4</pub-id><pub-id pub-id-type="pmcid">PMC4617899</pub-id></element-citation></ref><ref id="bib55"><element-citation publication-type="journal" id="sref55"><person-group person-group-type="author"><name name-style="western"><surname>Steinman</surname><given-names>R.M.</given-names></name><name name-style="western"><surname>Hawiger</surname><given-names>D.</given-names></name><name name-style="western"><surname>Nussenzweig</surname><given-names>M.C.</given-names></name></person-group><article-title>Tolerogenic dendritic cells</article-title><source>Annu. Rev. Immunol.</source><volume>21</volume><year>2003</year><fpage>685</fpage><lpage>711</lpage><pub-id pub-id-type="pmid">12615891</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1146/annurev.immunol.21.120601.141040</pub-id></element-citation></ref><ref id="bib56"><element-citation publication-type="journal" id="sref56"><person-group person-group-type="author"><name name-style="western"><surname>Subramanian</surname><given-names>A.</given-names></name><name name-style="western"><surname>Tamayo</surname><given-names>P.</given-names></name><name name-style="western"><surname>Mootha</surname><given-names>V.K.</given-names></name><name name-style="western"><surname>Mukherjee</surname><given-names>S.</given-names></name><name name-style="western"><surname>Ebert</surname><given-names>B.L.</given-names></name><name name-style="western"><surname>Gillette</surname><given-names>M.A.</given-names></name><name name-style="western"><surname>Paulovich</surname><given-names>A.</given-names></name><name name-style="western"><surname>Pomeroy</surname><given-names>S.L.</given-names></name><name name-style="western"><surname>Golub</surname><given-names>T.R.</given-names></name><name name-style="western"><surname>Lander</surname><given-names>E.S.</given-names></name><name name-style="western"><surname>Mesirov</surname><given-names>J.P.</given-names></name></person-group><article-title>Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles</article-title><source>Proc. Natl. Acad. Sci. USA</source><volume>102</volume><year>2005</year><fpage>15545</fpage><lpage>15550</lpage><pub-id pub-id-type="pmid">16199517</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1073/pnas.0506580102</pub-id><pub-id pub-id-type="pmcid">PMC1239896</pub-id></element-citation></ref><ref id="bib57"><element-citation publication-type="journal" id="sref57"><person-group person-group-type="author"><name name-style="western"><surname>van der Maaten</surname><given-names>J.</given-names></name></person-group><article-title>Visualizing Data Using t-SNE</article-title><source>J. Mach. Learn. Res.</source><volume>9</volume><year>2008</year><fpage>2579</fpage><lpage>2605</lpage></element-citation></ref><ref id="bib58"><element-citation publication-type="journal" id="sref58"><person-group person-group-type="author"><name name-style="western"><surname>Villani</surname><given-names>A.C.</given-names></name><name name-style="western"><surname>Satija</surname><given-names>R.</given-names></name><name name-style="western"><surname>Reynolds</surname><given-names>G.</given-names></name><name name-style="western"><surname>Sarkizova</surname><given-names>S.</given-names></name><name name-style="western"><surname>Shekhar</surname><given-names>K.</given-names></name><name name-style="western"><surname>Fletcher</surname><given-names>J.</given-names></name><name name-style="western"><surname>Griesbeck</surname><given-names>M.</given-names></name><name name-style="western"><surname>Butler</surname><given-names>A.</given-names></name><name name-style="western"><surname>Zheng</surname><given-names>S.</given-names></name><name name-style="western"><surname>Lazo</surname><given-names>S.</given-names></name></person-group><article-title>Single-cell RNA-seq reveals new types of human blood dendritic cells, monocytes, and progenitors</article-title><source>Science</source><volume>356</volume><year>2017</year><comment>eaah4573</comment><pub-id pub-id-type="doi" assigning-authority="pmc">10.1126/science.aah4573</pub-id><pub-id pub-id-type="pmcid">PMC5775029</pub-id><pub-id pub-id-type="pmid">28428369</pub-id></element-citation></ref><ref id="bib59"><element-citation publication-type="journal" id="sref59"><person-group person-group-type="author"><name name-style="western"><surname>Wang</surname><given-names>Z.Q.</given-names></name><name name-style="western"><surname>Milne</surname><given-names>K.</given-names></name><name name-style="western"><surname>Derocher</surname><given-names>H.</given-names></name><name name-style="western"><surname>Webb</surname><given-names>J.R.</given-names></name><name name-style="western"><surname>Nelson</surname><given-names>B.H.</given-names></name><name name-style="western"><surname>Watson</surname><given-names>P.H.</given-names></name></person-group><article-title>CD103 and Intratumoral Immune Response in Breast Cancer</article-title><source>Clin. Cancer Res.</source><volume>22</volume><year>2016</year><fpage>6290</fpage><lpage>6297</lpage><pub-id pub-id-type="pmid">27267849</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1158/1078-0432.CCR-16-0732</pub-id></element-citation></ref><ref id="bib60"><element-citation publication-type="journal" id="sref60"><person-group person-group-type="author"><name name-style="western"><surname>Waskow</surname><given-names>C.</given-names></name><name name-style="western"><surname>Liu</surname><given-names>K.</given-names></name><name name-style="western"><surname>Darrasse-Jèze</surname><given-names>G.</given-names></name><name name-style="western"><surname>Guermonprez</surname><given-names>P.</given-names></name><name name-style="western"><surname>Ginhoux</surname><given-names>F.</given-names></name><name name-style="western"><surname>Merad</surname><given-names>M.</given-names></name><name name-style="western"><surname>Shengelia</surname><given-names>T.</given-names></name><name name-style="western"><surname>Yao</surname><given-names>K.</given-names></name><name name-style="western"><surname>Nussenzweig</surname><given-names>M.</given-names></name></person-group><article-title>The receptor tyrosine kinase Flt3 is required for dendritic cell development in peripheral lymphoid tissues</article-title><source>Nat. Immunol.</source><volume>9</volume><year>2008</year><fpage>676</fpage><lpage>683</lpage><pub-id pub-id-type="pmid">18469816</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/ni.1615</pub-id><pub-id pub-id-type="pmcid">PMC2746085</pub-id></element-citation></ref><ref id="bib61"><element-citation publication-type="journal" id="sref61"><person-group person-group-type="author"><name name-style="western"><surname>Wollenberg</surname><given-names>A.</given-names></name><name name-style="western"><surname>Kraft</surname><given-names>S.</given-names></name><name name-style="western"><surname>Hanau</surname><given-names>D.</given-names></name><name name-style="western"><surname>Bieber</surname><given-names>T.</given-names></name></person-group><article-title>Immunomorphological and ultrastructural characterization of Langerhans cells and a novel, inflammatory dendritic epidermal cell (IDEC) population in lesional skin of atopic eczema</article-title><source>J. Invest. Dermatol.</source><volume>106</volume><year>1996</year><fpage>446</fpage><lpage>453</lpage><pub-id pub-id-type="pmid">8648175</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1111/1523-1747.ep12343596</pub-id></element-citation></ref><ref id="bib62"><element-citation publication-type="journal" id="sref62"><person-group person-group-type="author"><name name-style="western"><surname>Yin</surname><given-names>X.</given-names></name><name name-style="western"><surname>Yu</surname><given-names>H.</given-names></name><name name-style="western"><surname>Jin</surname><given-names>X.</given-names></name><name name-style="western"><surname>Li</surname><given-names>J.</given-names></name><name name-style="western"><surname>Guo</surname><given-names>H.</given-names></name><name name-style="western"><surname>Shi</surname><given-names>Q.</given-names></name><name name-style="western"><surname>Yin</surname><given-names>Z.</given-names></name><name name-style="western"><surname>Xu</surname><given-names>Y.</given-names></name><name name-style="western"><surname>Wang</surname><given-names>X.</given-names></name><name name-style="western"><surname>Liu</surname><given-names>R.</given-names></name></person-group><article-title>Human Blood CD1c+ Dendritic Cells Encompass CD5high and CD5low Subsets That Differ Significantly in Phenotype, Gene Expression, and Functions</article-title><source>J. Immunol.</source><volume>198</volume><year>2017</year><fpage>1553</fpage><lpage>1564</lpage><pub-id pub-id-type="pmid">28087664</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.4049/jimmunol.1600193</pub-id></element-citation></ref><ref id="bib63"><element-citation publication-type="journal" id="sref63"><person-group person-group-type="author"><name name-style="western"><surname>Yu</surname><given-names>C.I.</given-names></name><name name-style="western"><surname>Becker</surname><given-names>C.</given-names></name><name name-style="western"><surname>Wang</surname><given-names>Y.</given-names></name><name name-style="western"><surname>Marches</surname><given-names>F.</given-names></name><name name-style="western"><surname>Helft</surname><given-names>J.</given-names></name><name name-style="western"><surname>Leboeuf</surname><given-names>M.</given-names></name><name name-style="western"><surname>Anguiano</surname><given-names>E.</given-names></name><name name-style="western"><surname>Pourpe</surname><given-names>S.</given-names></name><name name-style="western"><surname>Goller</surname><given-names>K.</given-names></name><name name-style="western"><surname>Pascual</surname><given-names>V.</given-names></name></person-group><article-title>Human CD1c+ dendritic cells drive the differentiation of CD103+ CD8+ mucosal effector T cells via the cytokine TGF-β</article-title><source>Immunity</source><volume>38</volume><year>2013</year><fpage>818</fpage><lpage>830</lpage><pub-id pub-id-type="pmid">23562160</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2013.03.004</pub-id><pub-id pub-id-type="pmcid">PMC3639491</pub-id></element-citation></ref><ref id="bib64"><element-citation publication-type="journal" id="sref64"><person-group person-group-type="author"><name name-style="western"><surname>Zaba</surname><given-names>L.C.</given-names></name><name name-style="western"><surname>Krueger</surname><given-names>J.G.</given-names></name><name name-style="western"><surname>Lowes</surname><given-names>M.A.</given-names></name></person-group><article-title>Resident and “inflammatory” dendritic cells in human skin</article-title><source>J. Invest. Dermatol.</source><volume>129</volume><year>2009</year><fpage>302</fpage><lpage>308</lpage><pub-id pub-id-type="pmid">18685620</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/jid.2008.225</pub-id><pub-id pub-id-type="pmcid">PMC2746703</pub-id></element-citation></ref></ref-list><sec id="app2" sec-type="supplementary-material"><title>Supplemental Information</title><p id="p0410"><supplementary-material content-type="local-data" id="mmc1" position="float" orientation="portrait"><caption><title>Document S1. Figures S1–S7 and Table S2</title></caption><media xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="mmc1.pdf" position="float" orientation="portrait"><?suppdata-name mmc1.pdf?><?suppdata-size 4704203?><?suppdata-md5 639fd45d648b5464dbc4583a2a7b327a?><?suppdata-image-server-status NEVER_LOAD?><?suppdata-mime-type application?><?suppdata-mime-sub-type pdf?><?suppdata-cloudpmc-urn urn:app:922d/7445430/639fd45d648b/mmc1.pdf?></media></supplementary-material><supplementary-material content-type="local-data" id="mmc2" position="float" orientation="portrait"><caption><title>Table S1. Significant DEGs Obtained from the Pairwise Comparisons Used to Generate the Volcano Plots, Related to Figures 4 and 6 and Figures S1, S3, S5, and S6</title></caption><media xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="mmc2.xlsx" position="float" orientation="portrait"><?suppdata-name mmc2.xlsx?><?suppdata-size 408823?><?suppdata-md5 268805f38a8f98c782e972509831f942?><?suppdata-image-server-status NEVER_LOAD?><?suppdata-mime-type application?><?suppdata-mime-sub-type vnd.openxmlformats-officedocument.spreadsheetml.sheet?><?suppdata-cloudpmc-urn urn:app:922d/7445430/268805f38a8f/mmc2.xlsx?></media></supplementary-material><supplementary-material content-type="local-data" id="mmc3" position="float" orientation="portrait"><caption><title>Document S2. Article plus Supplemental Information</title></caption><media xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="mmc3.pdf" position="float" orientation="portrait"><?suppdata-name mmc3.pdf?><?suppdata-size 12120679?><?suppdata-md5 d46219d91af29a5b0a98d53ceda6404f?><?suppdata-image-server-status NEVER_LOAD?><?suppdata-mime-type application?><?suppdata-mime-sub-type pdf?><?suppdata-cloudpmc-urn urn:app:922d/7445430/d46219d91af2/mmc3.pdf?></media></supplementary-material></p></sec><ack id="ack0010"><title>Acknowledgments</title><p>P.G. is a CNRS investigator. J.H. is an INSERM investigator. We thank the Anthony Nolan Cord Blood Bank and Cell Therapy Centre for providing the cord blood units used in this study, the flow core and NGS platform from Institut Curie, and Daniel Chung for technical help. This research was supported by the <funding-source id="gs1"><institution-wrap><institution-id institution-id-type="doi">10.13039/501100000272</institution-id><institution>National Institute for Health Research</institution></institution-wrap></funding-source> (NIHR) <funding-source id="gs2"><institution-wrap><institution-id institution-id-type="doi">10.13039/100014461</institution-id><institution>Biomedical Research Centre</institution></institution-wrap></funding-source> based at <funding-source id="gs3"><institution-wrap><institution-id institution-id-type="doi">10.13039/501100004941</institution-id><institution>Guy's and St Thomas' NHS Foundation Trust</institution></institution-wrap></funding-source> and <funding-source id="gs4"><institution-wrap><institution-id institution-id-type="doi">10.13039/100009360</institution-id><institution>King’s College London</institution></institution-wrap></funding-source> and/or the <funding-source id="gs5">NIHR Clinical Research Facility</funding-source>. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR, or the Department of Health. The research was supported by the <funding-source id="gs6"><institution-wrap><institution-id institution-id-type="doi">10.13039/501100000268</institution-id><institution>BBSRC</institution></institution-wrap></funding-source>, <funding-source id="gs7">NC3RS</funding-source>, <funding-source id="gs9"><institution-wrap><institution-id institution-id-type="doi">10.13039/501100000289</institution-id><institution>CRUK</institution></institution-wrap></funding-source>, <funding-source id="gs10"><institution-wrap><institution-id institution-id-type="doi">10.13039/501100000833</institution-id><institution>Rosetrees Trust</institution></institution-wrap></funding-source>, <funding-source id="gs11">CRUK King’s Health Partners Cancer Centre in United Kingdom. The research was supported by</funding-source>
+<funding-source id="gs12">IDEX</funding-source> (ANR-18-IDEX-0001) and <funding-source id="gs13"><institution-wrap><institution-id institution-id-type="doi">10.13039/100013501</institution-id><institution>INCA</institution></institution-wrap></funding-source> (PL-BIO-ICR1), <funding-source id="gs14"><institution-wrap><institution-id institution-id-type="doi">10.13039/501100010463</institution-id><institution>Institut Curie</institution></institution-wrap></funding-source>, <funding-source id="gs15"><institution-wrap><institution-id institution-id-type="doi">10.13039/501100001677</institution-id><institution>Institut National de la Santé et de la Recherche Médicale</institution></institution-wrap></funding-source>, <funding-source id="gs16">Labex DCBIOL</funding-source> (ANR-10-IDEX-0001-02 PSL), ANR-11-LABX0043, <funding-source id="gs17">SIRIC</funding-source> INCa-DGOS-INSERM_12554, in France and the <funding-source id="gs18">Programa de Apoyo a Centros con Financiamiento Basal</funding-source> (AFB 170004), and <funding-source id="gs19"><institution-wrap><institution-id institution-id-type="doi">10.13039/501100002850</institution-id><institution>FONDECYT</institution></institution-wrap></funding-source> 1171703 in Chile.</p><sec id="sec5"><title>Author Contributions</title><p id="p0395">Conceptualization, P.G., P.B., G.A., K.V., and J.H.; Software, K.V., W.R., and Y.M.K.; Methodology, G.A., P.B., K.V., E.P., S.H., and A.L.; Formal Analysis, G.A., P.B., K.V., W.R., J.H., and P.G.; Investigation, G.A., K.V., P.B., S.H., R.N.R., K.W., and A.S.; Resources, E.P., J.H., and P.G.; Writing, P.G., G.A., K.V., P.B., and J.H.; Visualization, G.A., K.V., and P.B.; Supervision, A.L., E.P., J.H., and P.G.; Funding acquisition, A.L., E.P., J.H., and P.G.</p></sec><sec sec-type="COI-statement" id="sec6"><title>Declaration of Interests</title><p id="p0400">The authors declare no competing interests.</p></sec></ack><fn-group><fn id="app1" fn-type="supplementary-material"><p id="p0405">Supplemental Information can be found online at <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="https://doi.org/10.1016/j.immuni.2020.06.002" id="intref0420">https://doi.org/10.1016/j.immuni.2020.06.002</ext-link>.</p></fn></fn-group></back></article>
\ No newline at end of file
diff --git a/paper-releases/pbisc-diffusion/updates/20261015T141849Z-rare-cell-7k-8k-panels/diffusion/dataset/server_1/model8/plans/rare_gene_context_expansion_20261015/research/ilc_dc/sources/PMC8089104.txt b/paper-releases/pbisc-diffusion/updates/20261015T141849Z-rare-cell-7k-8k-panels/diffusion/dataset/server_1/model8/plans/rare_gene_context_expansion_20261015/research/ilc_dc/sources/PMC8089104.txt
new file mode 100644
index 0000000000000000000000000000000000000000..e6de01fec3e4bd4b30d682425bcfad0d222cef30
--- /dev/null
+++ b/paper-releases/pbisc-diffusion/updates/20261015T141849Z-rare-cell-7k-8k-panels/diffusion/dataset/server_1/model8/plans/rare_gene_context_expansion_20261015/research/ilc_dc/sources/PMC8089104.txt
@@ -0,0 +1,190 @@
+Innate lymphoid cells (ILCs) are a recent addition to innate immunity, playing an important role in clearing intra- and extracellular pathogens, regulating adaptive immunity, tissue repair and contributing to inflammation. 1  Their defining feature is a lack of rearranged antigen-specific receptors, which distinguishes them from T cells. However, in other ways these cells closely mirror the function of their adaptive counterparts, being an essential source of cytokines related to type 1, 2 and 17 immune responses. 1  Currently, ILCs are classified into five subsets: ILC1, ILC2, ILC3, Natural killer (NK) cells and lymphoid tissue inducer (LTi) cells. ILC1 produce interferon-gamma (IFN-γ) and are dependent on the T-box transcription factor (TF) TBX21 (Tbet) for their development and function. NK cells share these features with ILC1, but they additionally express the TF Eomesodermin homolog (Eomes) and have the ability to produce cytolytic and cytotoxic molecules. 1  ILC2s produce IL-13 and IL-5, and are dependent on GATA-3. 1  Both LTi cells and ILC3 can produce IL-17 and IL-22 cytokines, and are dependent on the nuclear receptor ROR-γt. 1  Transcriptional analysis of ILCs obtained from different human and mouse tissues has revealed tissue-specific transcriptional imprinting. 2 – 4  However, these studies relied on biased definition of ILC populations sorted in bulk, masking potential heterogeneity and misdefinition of ILC populations across tissues. Recent developments in single-cell technologies on the transcriptional and protein level have rapidly advanced our understanding of the heterogeneity of the human immune system. 5  While protein analyses still suffer from limitations, e.g., in terms of cell number requirements, single-cell RNA-sequencing (scRNA-seq) enables an unbiased, transcriptome-wide analysis of even small numbers of cells. So far, scRNA-seq of specific subsets of ILCs (NK cells, 6  ILC2 7 ) or in specific tissues 8 , 9  has been performed in mice. Similarly, in humans, scRNA-seq has been restricted to one subset (NK cells 6 ) or one tissue (ileum 10 , 11  and tonsil 12 ) at a time, yielding limited understanding of tissue-specific imprinting of ILCs.
+
+In this study, we analyzed human ILCs through full-length scRNA-seq, combining data from four different tissues: blood, tonsil, lung and colon. Comparison between tissues revealed tissue-specific transcriptional imprinting, and gave us the possibility to, for the first time in humans, compare scRNA-seq profiles of ILCs in the circulation, secondary lymphoid organs (SLOs) and mucosal tissues. Our data, which is provided as a publically available resource, reveal previously unknown functions of human ILCs and extend our current understanding of human ILCs biology.
+
+Transcriptional profiling of ILCs across tissues on the bulk 2 , 3  or single-cell level 6 , 7 , 9  in humans and mice have revealed unique features of ILCs that are dependent on the tissue microenvironment. Here we performed flow cytometric sorting of human CD127 +  ILCs from three tissues representing two distinct immune cell compartments: circulation (blood), and mucosal tissue (lung and colon) (3–4 donors/tissue) (Fig.  1a ;  Supplementary Information  Fig.  S1 ). As reference populations we included selected NK populations along the spectrum of previously described tissue NK cell phenotypes. Predominantly recirculating CD56 dim CD16 +  NK cells were sorted from the lung 13  and predominantly tissue-resident NK cells were sorted as bulk CD56 +  from the gut 14  or as CD56 bright NKG2A +  NK cells from the tonsil. 12  Cells were next processed for full-length Smart-seq2 scRNA-seq to unravel in-depth tissue-specific transcriptional profiles and heterogeneity. Our own previously published Smart-seq2 scRNA-seq dataset of CD127 +  ILC and NK cells from tonsils, 12  representing a SLO, was included in the analysis. All datasets were combined and analyzed together in order to identify common and distinct patterns across tissues and cell types (Fig.  1a ). Low quality cells and a minor fraction of contaminating cells lacking expression of  PTPRC  (encoding CD45) and expressing transcripts of antigen-presenting cells (e.g., high expression of  HLA-DQA1 ,  HLA-DRA  and HLA- DPA1  at levels exceeding ILC) were identified and removed ( Supplementary Information , Fig.  S2 ), resulting in the analysis of a total of 2956 cells distributed across four tissues (Fig.  1b ;  Supplementary Information , Fig.  S2 ). Fig. 1 ScRNA-seq reveals subset and tissue-specific transcriptional imprinting. a  Summary of the experimental workflow. Arrows in data integration and analysis section visualize the concept of immune cell migration between the tissues analyzed. Tonsil data was obtained from 12  using the same scRNA-seq platform.  b  UMAP visualization of all cells ( n  = 2956) color-coded on the basis of tissue origin.  c  UMAP visualization of all cells, color-coded on the basis of cell surface phenotype inferred from indexed FACS sorting.  d  UMAP showing unbiased graph-based clustering analysis of all cells.  e  UMAP of annotated clusters. C_, colon; L_, lung; B_, blood; T_, tonsil; nILC, naïve ILC.  f  Dotplot displaying expression of selected and previously described human ILC subset-specific TFs and cell surface markers used to annotate clusters in  e .  g  Barplot showing the representation of each donor in each annotated cluster in  e . Data are from ten independent experiments with one tissue donor each (blood = 3, lung = 4 and colon = 3) integrated with data from. 12
+
+a  Summary of the experimental workflow. Arrows in data integration and analysis section visualize the concept of immune cell migration between the tissues analyzed. Tonsil data was obtained from 12  using the same scRNA-seq platform.  b  UMAP visualization of all cells ( n  = 2956) color-coded on the basis of tissue origin.  c  UMAP visualization of all cells, color-coded on the basis of cell surface phenotype inferred from indexed FACS sorting.  d  UMAP showing unbiased graph-based clustering analysis of all cells.  e  UMAP of annotated clusters. C_, colon; L_, lung; B_, blood; T_, tonsil; nILC, naïve ILC.  f  Dotplot displaying expression of selected and previously described human ILC subset-specific TFs and cell surface markers used to annotate clusters in  e .  g  Barplot showing the representation of each donor in each annotated cluster in  e . Data are from ten independent experiments with one tissue donor each (blood = 3, lung = 4 and colon = 3) integrated with data from. 12
+
+Cell embedding using uniform manifold approximation and projection (UMAP) 15  revealed that all ILC subsets as well as NK cells separated according to tissue of origin and FACS cell subset annotation as inferred from index sorting (Fig.  1b, c ). Unbiased graph-based clustering subdivided cells into 20 clusters that segregated cells both by tissue of origin and FACS phenotype (Fig.  1d ). On average, we observed 90% agreement between protein FACS phenotype and the clusters identified in scRNA-seq ( Supplementary Information , Fig.  S3a ). We next used previously described human ILC subset-specific TFs and cell surface markers 12  to annotate the clusters (Fig.  1e, f ). Clusters 5 (lung), 15 (tonsil) and 19 (colon), expressing  NKG7 ,  EOMES  and  GZMA/B  but lacking  IL7R  were annotated as NK cells. Clusters 8 (blood), 13 (tonsil), 18 (lung) expressed  IKZF3  (encoding the TF and zinc finger protein Aiolos) but lacked  EOMES , and  KIR2DL1/3 , and were conditionally annotated as ILC1. Cluster 17 consisted of both tonsil and blood CD127 +  ILCs, which expressed low levels of  EOMES , conditionally annotating them as  EOMES +  ILC1. Clusters 6 (blood), 11 (tonsil) and 12 (lung) were annotated as ILC2 on the basis of uniquely high  GATA3 ,  MAF ,  PTGDR2  and  HPGDS  expression (Fig.  1e, f ). As predicted by the low frequency of human colon ILC1 and ILC2 in non-inflamed colon tissue, 16  colon cells with a FACS phenotype of ILC1 and ILC2 (total 10 cells) did not generate distinct clusters. ILC3s were annotated on the basis of  RORC ,  IL23R ,  IL1R1  and  KIT  expression (cluster 4 and 9 in tonsil, 2 and 10 in lung and 0, 1, 3 and 14 in colon) (Fig.  1e, f ). We also observed  IL7R  and  KIT -expressing ILCs in the blood (cluster 7) as well as in colon and lung (cluster 16), which lacked transcripts of all master TFs (Fig.  1e, f ), reminiscent of the previously described ILC progenitors in blood and tissues. 12 , 17  These cells were annotated as naïve ILCs (nILCs), consistent with the nomenclature for naïve T cells. 18  Importantly, cells from all donors of each tissue were present in all annotated cell types (Fig.  1g ) and clusters ( Supplementary Information , Fig.  S3b, c ) for each tissue.
+
+To identify conserved gene expression patterns that are shared across tissues and cell types, we used a two-step approach rather than simple differential expression (DE), which reveals differences but not similarities between cell types and tissues. DE gene lists for each annotated cell type and the overlaps between cell types and tissues are provided in  Supplementary Information , Data  S1 . For the two-step approach, we selected the top DE genes in any of the 20 clusters (4410 genes) and subsequently performed weighted gene co-expression network analysis (WGCNA) for identification of 100 gene modules (see Materials and Methods for details) (Fig.  2a, b ;  Supplementary Information , Data  S1 ). A high number of modules were chosen in order to avoid forcing genes with different expression patterns into the same group. Modules with highly correlated expression are indicated with a connecting line in the graph (Fig.  2c ). Next, modules were scored individually for their contribution to discriminating tissues, cell types and/or cell clusters (Fig.  2c ;  Supplementary Information , Fig.  S3d ). Notably, gene modules did not contribute equally to explaining metadata and thus only modules that contributed to at least one cell type or tissue were selected for further in-depth analysis herein (Fig.  2c ). Fig. 2 Migration and tissue-resident transcriptional profile of ILCs. a  Summary of analysis approach used to identify gene modules and perform trajectory analysis.  b  UMAP visualization of all cells, color-coded on the basis of tissue origin.  c  Circular dendrogram illustrating the number and composition of each module and the contribution of each module to discriminating cell clusters, tissues and cell types (as inferred from FACS). Modules with highly correlated expressions are indicated with a 5-NN graph (connecting lines represent KNN; k nearest neighbor). Modules described in text are marked with an asterisk.  d  Modules highly correlated with tissues and/or cell types/clusters containing genes involved in migration or tissue residency. The figure shows a selection of genes per module, and their levels of expression within annotated clusters (Fig.  1e ). Data in  a – d  is from ten independent experiments with one tissue donor each (blood = 3, lung = 4 and colon = 3) integrated with data from. 12 
+ e  Co-expression of CD69 and CD18/CD62L (top row) and NKp44 and CD18/CD62L (bottom row) on total ILCs from blood, tonsil and colon biopsies.  f  Histogram showing the expression of CD18, CD62L and CD69 on mechanically or enzymatically isolated total tonsil ILCs. FMO, fluorescence minus one.  g  Bar plot quantification of the results shown in  e . MFI, mean fluorescence intensity. Data in  e – g  are from six independent experiments with at least one tissue donor each (blood = 6, tonsil = 8 and colon = 6).
+
+a  Summary of analysis approach used to identify gene modules and perform trajectory analysis.  b  UMAP visualization of all cells, color-coded on the basis of tissue origin.  c  Circular dendrogram illustrating the number and composition of each module and the contribution of each module to discriminating cell clusters, tissues and cell types (as inferred from FACS). Modules with highly correlated expressions are indicated with a 5-NN graph (connecting lines represent KNN; k nearest neighbor). Modules described in text are marked with an asterisk.  d  Modules highly correlated with tissues and/or cell types/clusters containing genes involved in migration or tissue residency. The figure shows a selection of genes per module, and their levels of expression within annotated clusters (Fig.  1e ). Data in  a – d  is from ten independent experiments with one tissue donor each (blood = 3, lung = 4 and colon = 3) integrated with data from. 12 
+ e  Co-expression of CD69 and CD18/CD62L (top row) and NKp44 and CD18/CD62L (bottom row) on total ILCs from blood, tonsil and colon biopsies.  f  Histogram showing the expression of CD18, CD62L and CD69 on mechanically or enzymatically isolated total tonsil ILCs. FMO, fluorescence minus one.  g  Bar plot quantification of the results shown in  e . MFI, mean fluorescence intensity. Data in  e – g  are from six independent experiments with at least one tissue donor each (blood = 6, tonsil = 8 and colon = 6).
+
+The analysis of four different tissues that represent unique immune niches (circulation, lymphoid- and mucosal tissue) in combination with the unbiased gene module analysis allowed us to identify conserved patterns across tissues and cell types (Fig.  2d ). We identified two gene modules (41 and 11) containing transcripts highly correlated to the expression of the prototypical T cell migratory molecules  S1PR1  and  SELL  (encoding L-selectin; CD62L), respectively (Fig.  2d ). Genes in these two modules were highly expressed in tonsil and blood ILCs, including nILCs. The modules included the glycolysis-associated transcript  LDHB , the adhesion molecule transcript  ICAM3 , as well as transcripts for several S100 proteins ( S100A4 ,  S100A6  and  S100A11 ) (Fig.  2d ), shown to be involved in migration. 19  Module 23 was also over-represented in blood and tonsil and contained transcripts for several integrins including  ITGB2  (CD18)  ITGAL  (CD11a) and  ITGAM  (CD11b) (Fig.  2d ). CD18 complexes with CD11 molecules to form the adhesion molecules lymphocyte function-associated antigen-1 (LFA-1), which together with CD62L are involved in endothelial tethering and firm adhesion. In sharp contrast to the migratory phenotype of blood and tonsil ILCs, we also identified gene modules indicative of tissue residency of lung and colon ILCs, including tissue nILCs and to some extent also of tonsil ILCs. Module 54 revealed a set of genes highly correlated to the expression of the prototypical tissue-residency-associated marker  CD69  (Fig.  2d ). We confirmed the diametrically opposed migratory vs tissue-residency profiles of blood as compared to colon ILCs by demonstrating the absence of CD69 and high expression of CD18 and CD62L on blood ILCs and low expression of CD18 and CD62L and high expression of CD69 on colon ILCs (Fig.  2e ). Tonsil showed a mixed phenotype, regardless of mechanic or enzymatic isolation of the cells (Fig.  2f ), with low expression of CD62L, intermediate expression of CD18 and some cells expressing CD69, aligning with the presence of both tissue resident and migratory ILCs in secondary lymphoid tissues in mice. 20  CD69 +  tonsil ILCs expressed lower levels of CD18 as compared to CD69 –  tonsil ILCs (Fig.  2g ). Similarly, expression of NKp44 was associated with reduced expression of CD18 (Fig.  2g ). These data not only validate the transcriptional findings but also show that in contrast to colon, tonsil CD69 +  express the endothelial adhesion integrin CD18, as described for intravascular liver-resident T cells. 21
+
+Showing a similar expression pattern, we identified module 7 as containing genes previously reported in human tissue resident T cells 22  involved in Fos/Jun signaling ( FOS ,  FOSB ,  FOSL2  and  JUNB ), NFκB signaling ( NFKBIZ ,  NFKBID ,  NFKBIA ,  NFKB1  and  NFKB2 ), adhesion ( ICAM1 ) and  CXCR4  (Fig.  2d ). In this module were also transcripts of effector molecules such as  CXCL2 ,  CXCL1  and  AREG , the latter previously reported as an ILC2 effector molecule. 23  The expression of  MAFF ,  KLF4 ,  NFIL3  and  NR4A1  in tissue ILCs suggests that these TFs are involved in transcriptional programs of activation and/or tissue-residency in colon, lung and tonsil ILCs. Module 6, which was particularly expressed in mature but not nILCs in colon and lung, contained transcripts of additional nuclear receptor subfamily four group members,  NR4A2  and  NR4A3 , as well as the hyaluronic acid receptor  CD44 , previously shown to be expressed by tissue-resident T and NK cells 13 , 22  (Fig.  2d ). This module also contained transcripts involved in lipid metabolism ( ELOVL5  and  PTGS2 ). The expression of  STAT3 ,  NOTCH2 ,  DLL1 ,  RORA  and  RARA  suggests a potential role for STAT3-, Notch-, retinoic acid- and cholesterol-derivative (e.g., oxysterol) signaling in controlling activation and/or tissue-residency of colon and lung ILCs (Fig.  2d ). Colon and lung ILC3, but to a lesser extent tonsil ILC3 expressed transcripts encoding the receptors for IL-4 ( IL4R ), IFN-γ ( IFNGR2 ) and CXCL16 ( CXCR6 ), raising the possibility that these factors are involved in the tissue-specific regulation of colon and lung ILC3. Furthermore, we identified module 34, containing lung-restricted transcripts indicative of tissue-residency 22  including  LMNA  and the TFs  KLF10  and  BCL3  (Fig.  2d ), the latter involved in NFκB signaling.
+
+In summary, we identified three gene modules allowing for the characterization of the lymphoid- and peripheral tissue homing properties of tonsil and blood ILCs, including blood nILCs. Four additional modules contained transcripts for known and potential activation and/or tissue-residency associated proteins that showed tissue-restricted expression across colon, lung and tonsils. Most of these genes were highly expressed by colon and lung ILCs, including tissue nILCs, but to some extent also by tonsil ILCs. Importantly, these genes were not expressed by blood ILCs, including blood nILCs. Furthermore, lung NK dim  cells, which have been reported as circulating 13  also showed low expression of transcripts known to be associated with tissue-residency.
+
+Module 6 specified shared characteristics of putative tissue-resident ILCs in lung and colon (Fig.  2d ). Notably, some of the genes in this module were particularly confined to ILC3 in these two tissues (Fig.  3a, b ). These included transcripts encoding HLA-DR, -DP and -DQ (Fig.  3b ), which we previously described in the tonsil. 12  Additional transcripts in this module included  IRF4  and  BCL2A1 , indicating a potential role for these transcripts in establishing tissue-unique features of human colon and lung ILC3. Also the transcript for the effector cytokine GM-CSF,  CSF2  and  BHLHE40 , the latter a metabolism-related gene reported to regulate GM-CSF production in mouse ILC3, 24  as well as  CXCL8  and  VEGFA  were expressed in lung and colon ILC3 (Fig.  3b ), further indicating colon and lung-associated transcriptional imprinting of ILC3 as compared to tonsil ILC3. However, colon and tonsil ILC3 also shared striking similarities with respect to high expression of the prototypical ILC3 transcripts  NCR2 ,  IL23R ,  NRP1 ,  IKZF2 ,  TNFSF13B ,  CCL20 ,  JAML ,  PECAM1 ,  TOX ,  RBPJ  and  CD300LF  as well as the previously described Th17 cell transcript  IL4I1  (Fig.  3b ). Of interest, colon ILC3 expressed both  LTA4H , encoding the leukotriene B4 (LTB 4 ) forming enzyme LTA4-hydrolase, and the transcript coding for the receptor for LTB 4 ,  LTB4R , indicating a role for LTB 4  in colon-specific regulation of ILC3 function. Further strengthening this hypothesis, a subset of colonic ILC3 expressed the LTB4R surface protein (Fig.  3c ). A comparison of DE gene expression between lung, colon and tonsil ILC3 is provided as part of  Supplementary Information , Data  S1 . Fig. 3 Tissue- and subset-specific profile of ILC3. a  UMAP visualization of annotated ILC3 and nILC clusters (C_, colon; L_, lung; B_, blood; T_, tonsil; nILC, naïve ILC). Clusters are color coded based on annotated clusters (Fig.  1e ).  b  Modules highly correlated with tissues and ILC3. The figure shows a selection of genes per module, and their levels of expression within annotated clusters (Fig.  1e ).  c  Expression of LTB4R on colon ILC3 (top plot) and summary bar plot showing expression of LTB4R on colon ILC3, tonsil ILC3 and blood nILCs (bottom plot). Data in  c  (top) is from one representative blood, tonsil and gut sample. Data are from six independent experiments with one tissue donor each (blood = 6, tonsil = 8 and colon = 6).  d  UMAP visualization of unbiased graph-based clustering analysis of ILC3. Clusters are color coded based on unbiased graph-based clustering analysis of all cells (Fig.  1d ).  e  Trajectory analysis of nILCs and colon ILC3 displaying a selection of DE genes for each colon ILC3 cluster (C_ILC3a–d).  f  Trajectory analysis of nILCs and lung ILC1, ILC2 and ILC3 displaying a selection of DE genes for each lung ILC3 cluster (L_ILC3a–b).  g  UMAP visualization of  NCR2  expression. Colored cells (shades of blue) indicate expression of  NCR2  transcripts.  h  UMAP visualization of  IL22  expression. Colored cells (shades of blue) indicate expression of  IL22  transcripts. Data in  a ,  b ,  d – h  is from ten independent experiments with one tissue donor each (blood = 3, lung = 4 and colon = 3) integrated with data from. 12
+
+a  UMAP visualization of annotated ILC3 and nILC clusters (C_, colon; L_, lung; B_, blood; T_, tonsil; nILC, naïve ILC). Clusters are color coded based on annotated clusters (Fig.  1e ).  b  Modules highly correlated with tissues and ILC3. The figure shows a selection of genes per module, and their levels of expression within annotated clusters (Fig.  1e ).  c  Expression of LTB4R on colon ILC3 (top plot) and summary bar plot showing expression of LTB4R on colon ILC3, tonsil ILC3 and blood nILCs (bottom plot). Data in  c  (top) is from one representative blood, tonsil and gut sample. Data are from six independent experiments with one tissue donor each (blood = 6, tonsil = 8 and colon = 6).  d  UMAP visualization of unbiased graph-based clustering analysis of ILC3. Clusters are color coded based on unbiased graph-based clustering analysis of all cells (Fig.  1d ).  e  Trajectory analysis of nILCs and colon ILC3 displaying a selection of DE genes for each colon ILC3 cluster (C_ILC3a–d).  f  Trajectory analysis of nILCs and lung ILC1, ILC2 and ILC3 displaying a selection of DE genes for each lung ILC3 cluster (L_ILC3a–b).  g  UMAP visualization of  NCR2  expression. Colored cells (shades of blue) indicate expression of  NCR2  transcripts.  h  UMAP visualization of  IL22  expression. Colored cells (shades of blue) indicate expression of  IL22  transcripts. Data in  a ,  b ,  d – h  is from ten independent experiments with one tissue donor each (blood = 3, lung = 4 and colon = 3) integrated with data from. 12
+
+Module 3 contained ILC3-related genes that were expressed by ILC3 in all tissues but at particularly high level and frequency in colon ILC3 (Fig.  3b ). These included surface molecule transcripts  KIT ,  IL1R1 ,  ICOS ,  CCR6 ,  TNFSF11  (encoding RANK ligand),  TNFSF4  (encoding OX40 ligand), the ITIM-containing  CD96  and the tetraspanin  CD81 . We also identified transcripts for the TFs  RORC  and  AHR , the signaling molecule  PTPN22  and the cytokine  LIF  as preferentially highly expressed by colon ILC3 (Fig.  3b ).
+
+Finally, we evaluated the transcriptional profile of the clusters of cells that were termed nILCs in blood and tissues (colon and lung) on the basis of lack of markers associated with mature ILCs (Fig.  1f ). From the evaluation of ILC3 (Fig.  3a, b ) it became obvious that these cells show few transcriptional parallels to ILC3 despite their phenotypical similarities as assessed by flow cytometry. Expression of genes associated with migration and tissue-residency revealed a clear migratory behavior of blood nILCs whereas tissue nILCs showed features of tissue residency (Fig.  2d ). A list of DE genes unique for blood and tissue nILCs is provided as part of  Supplementary Information , Data  S1 . This list includes TFs of potential importance for the differentiation of nILCs in blood (e.g.,  TCF7 ) and lung/colon (e.g.,  XBP1  and  BATF ) ( Supplementary Information , Data  S1 ).
+
+In summary, we unveil the transcriptional features of nILCs in blood and tissues and reveal tissue-specific differences among ILC3 and potential novel functions of these cells by description of transcripts previously not known to be expressed by human ILC3.
+
+Next we performed trajectory analysis via Slingshot 25  to understand the potential relationships between the putative nILCs identified in blood and tissues, and the more mature ILC subsets and the potential spectrum of activation states of ILC3 in colon and lung. Overlaying these trajectories with the clusters identified in Fig.  1d  (Fig.  3d ) not only illustrated the heterogeneity among ILC3 in colon, lung and tonsil but also inferred the potential transcriptional relationships between these transcriptionally distinct ILC3 subclusters (Fig.  3e, f ;  Supplementary Information , Fig.  S4a, b ). ILC3s were particularly heterogeneous in the colon where four subclusters could be identified through Louvain clustering (colon ILC3a–d, lung ILC3a–b, tonsil ILC3a–b) (Figs.  1 d,  3e, f ;  Supplementary Information , Fig.  S4a ). Some of these clusters were dominated by, but not unique to, individual donors as cells from all donors of each tissue were present in all clusters ( Supplementary Information , Fig.  S3b, c ) for each tissue. Of importance, the TF  IKZF2  and the master ILC3 TFs  RORC  and  AHR  were expressed across all ILC3 subclusters in all tissues, assigning them to the ILC3 lineage ( Supplementary Information , Fig.  S4c ). Analysis comparing clusters within each tissue and focusing on the top 20 DE genes ( Supplementary Information , Data  S2 ), revealed that colon ILC3a (cluster 1) contained cells with an activated and potentially tissue-resident phenotype, expressing transcripts for multiple effector molecules involved in chemotaxis and recruitment of immune cells including  CSF2 ,  XCL1 ,  XCL2 ,  CXCL8 ,  S100A4  and  S100A6  (Fig.  3e ).  EGR1 ,  EGR2 and KLF6 , encoding TFs involved in immune cell activation, 26 , 27  and  BCL2A1 , coding for an anti-apoptotic protein, were also enriched in this cluster of cells. The prototypical ILC3 cytokine  IL22  and the activation marker  NCR2  were expressed in colon ILC3a, as well as in colon ILC3b (cluster 0) (Fig.  3g, h ); the latter also showed an activated phenotype with expression of genes involved in glutamate response and lipid metabolism, respectively ( GRM7  and  LTA4H ). Colon ILC3a was most similar to lung ILC3a (cluster 2) in terms of  XCL1 ,  XCL2, S100A4, EGR2 and IFITM2 , but also shared similarities with lung ILC3b (cluster 10) in terms of  CXCL8, CD83  and  PTGS2  (encoding COX2) (Fig.  3e, f ). ILC3s in the lung also branched into two subclusters, indicating distinct transcriptional trajectories which both contained features of activation and tissue residency (Fig.  3f ). The activated phenotype of colon ILC3a stood in sharp contrast to colon ILC3d (cluster 14), which was most similar to lung ILC3b ( Supplementary Information , Data  S2 ). Colon ILC3d additionally expressed  SELL  (encoding CD62L), also expressed by tonsil ILC3b (cluster 4) ( Supplementary Information , Fig.  S4a ), 12 
+ S1PR1 ,  ITGAX  and  GPR183  (Fig.  3e ), indicative of a migratory behavior reminiscent of LTi-like cells contributing to colonic lymphoid structure formation in a GPR183-dependent manner in mice. 28  Importantly, colon ILC3d formed a separate branch in trajectory analysis and were transcriptionally distinct from the tissue nILCs (Fig.  3e ).
+
+In summary, ILC3s show a high degree of transcriptional heterogeneity, in particular in the colon where we identified a spectrum of ILC3 states with variable activation status which separated into different branches in trajectory analysis, ranging from cells with a migratory phenotype, to ILC3-expressing effector molecules important for metabolic adaptations, tissue repair and immune cell recruitment.
+
+Our gene module approach identified a set of genes that were specifically expressed in ILC2 across the three tissues (Fig.  4a, b ), making up an ILC2 core transcriptome. These genes included, as expected,  GATA3  and  MAF , as well as  HPGD ,  HPGDS  and  MBOAT2 , the latter three encoding proteins involved in lipid metabolism. Furthermore, ILC2 in all tissues expressed transcripts for the IL-9 receptor ( IL9R ). Other genes in the module were particularly expressed in blood ILC2 or blood and tonsil ILC2, including transcripts for receptors involved in cell trafficking and migration such as the prototypical ILC2 marker CRTH2 ( PTGDR2 ),  S1PR2  and  CCR2 , the latter shown to be involved in the regulation of S1P receptors. 29  The TNF family-related transcripts  TNFSF10  (encoding TRAIL) and  TNFRSF19  (encoding TROY) were also highly expressed by blood and tonsil ILC2, as well as  CD200R1 , possibly acting as an inhibitory receptor for human blood and tonsil ILC2. Another ILC2-related module was less restricted to ILC2 but still contained transcripts relevant for ILC2 biology, particularly for the potential negative regulation of these cells;  IL10RA , 30 
+ KLRG1 , 31 
+ PTGER2 , 32 
+ ANXA1 33  as well as  LGALS1 , encodes the immunosuppressive molecule galectin-1. Fig. 4 Tissue- and subset-specific profile of ILC2. a  UMAP visualization of annotated ILC2 clusters (L_, lung; B_, blood; T_, tonsil). Clusters are color coded based on annotated clusters (Fig.  1e ).  b  Modules highly correlated with ILC2 or tissue. The figure shows a selection of genes per module, and their level of expression within annotated clusters (Fig.  1e ). Data are from ten independent experiments with one tissue donor each (blood = 3, lung = 4 and colon = 3) integrated with data from. 12
+
+a  UMAP visualization of annotated ILC2 clusters (L_, lung; B_, blood; T_, tonsil). Clusters are color coded based on annotated clusters (Fig.  1e ).  b  Modules highly correlated with ILC2 or tissue. The figure shows a selection of genes per module, and their level of expression within annotated clusters (Fig.  1e ). Data are from ten independent experiments with one tissue donor each (blood = 3, lung = 4 and colon = 3) integrated with data from. 12
+
+Finally, we identified a module of genes (number 65) specifically expressed in lung, but not tonsil or blood ILC2 (Fig.  4b ). Expression of the main effector molecule  IL13  indicated that lung ILC2 are highly activated. Indeed, lung ILC2 showed high expression of  IL1RL1  and  IL17RB , encoding the receptors for the ILC2-activating alarmins IL-33 and IL-25, respectively. Additional transcripts for potential activating receptors included the SLAM-family receptor  SLAMF1, TNFRSF9 , the fatty acid receptors  FFAR3  and  PPARG , the latter shown to activate ILC2 in mice. 34
+
+The fact that blood ILC2 showed a phenotype associated with re-circulation and migration on both the transcriptional (Figs.  2 d,  4b ) and protein ( Supplementary Information , Fig.  S4d, e ) level support the possibility that blood ILC2 might be recruited to the lung, as previously suggested in the mouse 35 , 36  and in a human lung allergen-provocation study. 37  We therefore hypothesized that exposure of blood ILC2 to an inflammatory type 2-deviating environment could induce the lung ILC2 phenotype. To address this possibility, we sorted and exposed blood ILC2 to the prototypical ILC2-activating cocktail of alarmins (IL-2, IL-25, IL-33 and TSLP) and assessed phenotypical changes ( Supplementary Information , Fig.  S5 ). Indeed, alarmin exposure increased the expression of IL-1RL1 and IL-17RB surface proteins on blood-derived ILC2 (Fig.  5a ), phenocopying lung ILC2. In parallel, the expression of CRTH2 was downregulated, which we previously showed can be promoted by alarmin-mediated endogenous production of the CRTH2 ligand PGD 2 . 38  This finding raised the possibility that in the lung, alarmin-activated ILC2 might not express CRTH2. Indeed, utilizing our indexed flow cytometry data, evaluation of the composition of the cluster annotated as lung ILC2 (Fig.  1d ;  Supplementary Information , Fig.  S3a ) revealed a relatively high presence (35%) of CRTH2 −  ILCs (ILC1 and ILC3 by FACS phenotype) among the lung ILC2 which was not seen for blood or tonsil ILC2 where the majority of cells expressed CRTH2 protein (Fig.  1d ;  Supplementary Information , Fig.  S3a ). Evaluation of the transcriptional profile of CRTH2 –  vs CRTH2 +  cells showed that ILC2-related genes were present in CRTH2 –  ILC2, and at higher levels and frequencies as compared to lung ILC1 and ILC3 (Fig.  5b, c ). Fig. 5 Exposure of blood ILC2 to alarmins recapitulates the lung ILC2 phenotype. a  Expression of IL-1RL1, IL-17RB and CRTH2 protein on the surface of blood ILC2 treated with an alarmin cocktail (IL-2, IL-25, IL-33 and TSLP) for 48 h. FMO, fluorescence minus one.  b  UMAP visualization of annotated ILC clusters in the lung. Clusters are color coded based on annotated clusters (Fig.  1e ).  c  Violin plot visualization of differentially expressed genes in cells annotated as ILC2 with a FACS phenotype of ILC1 (L_ILC2_ILC1) or ILC3 (L_ILC2_ILC3) vs cells ILC1 or ILC3, respectively. Data are from seven independent experiments with a total of 13 blood donors ( a ) or from ten independent experiments with one tissue donor each (blood = 3, lung = 4 and colon = 3) integrated with data from 12  ( b ,  c ). Bars in  b  represent means ± SD. Mann Whitney test was used in  a ; ** P  < 0.01; *** P  < 0.001.
+
+a  Expression of IL-1RL1, IL-17RB and CRTH2 protein on the surface of blood ILC2 treated with an alarmin cocktail (IL-2, IL-25, IL-33 and TSLP) for 48 h. FMO, fluorescence minus one.  b  UMAP visualization of annotated ILC clusters in the lung. Clusters are color coded based on annotated clusters (Fig.  1e ).  c  Violin plot visualization of differentially expressed genes in cells annotated as ILC2 with a FACS phenotype of ILC1 (L_ILC2_ILC1) or ILC3 (L_ILC2_ILC3) vs cells ILC1 or ILC3, respectively. Data are from seven independent experiments with a total of 13 blood donors ( a ) or from ten independent experiments with one tissue donor each (blood = 3, lung = 4 and colon = 3) integrated with data from 12  ( b ,  c ). Bars in  b  represent means ± SD. Mann Whitney test was used in  a ; ** P  < 0.01; *** P  < 0.001.
+
+In summary, we reveal tissue-specific transcriptional differences between ILC2 in blood, tonsil and lung particularly related to migration, tissue-residency and sensing of the tissue microenvironment. Exposure of blood ILC2 to an inflammatory type-2 environment lead to development of a lung-like phenotype including downregulation of CRTH2. Thus, our transcriptional data highlight heterogeneity among lung ILC2, specifically identifying CRTH2 −  ILC2 with transcriptional similarities to bona fide CRTH2 +  ILC2, including expression of the ILC2 effector molecule  IL13 .
+
+Module 32 contained transcripts which were uniquely expressed by putative ILC1 and  EOMES +  ILC1 in blood and tonsil (Fig.  6a, b ). In line with previous reports, 12 , 39  the module consisted of T cell-related genes including  CD3D, CD3E, CD3G, CD4, CD5 ,  CD6 and CD27 .  LEF1 , a TF transcript described in tonsil ILC1, 12  was also expressed by blood ILC1. Notably, the transcripts in this module were expressed by few lung ILC1 suggesting that blood and tonsil ILC1 have more pronounced T cell features than lung ILC1 (Fig.  6b ). Furthermore, indicating that blood and tonsil ILC1, and in particular  EOMES +  ILC1, are more related to NK cells than lung ILC1, two additional transcriptional modules (Fig.  6b ) revealed transcripts expressed in NK cells as well as in ILC1 and  EOMES +  ILC1 in blood and tonsil, and to a lesser extent by ILC1 in lung (Fig.  6b ). Fig. 6 Tissue and subset specific profile of ILC1. a  UMAP visualization of annotated ILC1 and NK cell clusters. Clusters are color coded based on annotated clusters (Fig.  1e ).  b  Modules highly correlated with ILC1 and NK cells. The figure shows a selection of genes per module, and their level of expression within annotated clusters (Fig.  1e ).
+
+a  UMAP visualization of annotated ILC1 and NK cell clusters. Clusters are color coded based on annotated clusters (Fig.  1e ).  b  Modules highly correlated with ILC1 and NK cells. The figure shows a selection of genes per module, and their level of expression within annotated clusters (Fig.  1e ).
+
+Since putative ILC1 and  EOMES +  ILC1 in blood and tonsil showed characteristics of T cells, we performed a direct comparison between blood ILCs and CD4 +  T cells sorted in parallel from the same peripheral blood donors ( Supplementary Information , Figs.  S1  and  S6 ). To understand the T cell identity of ILC1, we analyzed TCR V(D)J-rearrangement using MiXCR. 40  In contrast to the T cell population, which as expected predominantly consisted of cells with rearranged TCR-α and -β chains, ILC2 and nILCs showed minimal V(D)J rearrangement across any of the TCR chains (Fig.  7a ;  Supplementary Information , Fig.  S7a ). This stood in sharp contrast to putative blood ILC1 and  EOMES +  ILC1, where most of the cells had undergone TCR rearrangement (Fig.  7a ;  Supplementary Information , Fig.  S7a ), despite lacking surface expression of TCR α/β, γ/δ or CD3 ( Supplementary Information , Fig.  S1a ). While  EOMES +  ILC1 mostly had predicted rearrangement in α- and/or β-TCR chains, blood ILC1 showed predicted rearrangement in all four TCR chains (Fig.  7a ;  Supplementary Information , Fig.  S7a ). UMAP clustering revealed that blood ILC1 with α- and/or β-TCR rearrangement chains clustered with CD4 +  T cells (AB_T_ILC1) (Fig.  7b–d ;  Supplementary Information , Fig.  S7b, c ). DE analysis showed that while the majority of ILC1 expressed transcripts of α- and/or β-TCR constant regions ( TRA-C  and  TRB-C ), expression of  CD4  as well as variable fragments of α- and/or β-TCR genes ( TRA-V  and  TRB-V ) was more pronounced in ILC1 with α- and/or β-TCR rearrangement (Fig.  7e ;  Supplementary Information , Fig.  S7d, e ). ILC1 with rearrangement in predominantly δ-, but also γ-TCR chains formed a distinct cluster (GD_ILC1) (Fig.  7c, d ;  Supplementary Information , Fig.  S7b ) and concordantly expressed fragments of γ- and/or δ-TCR variable chains ( TRG-V  and  TRD-V ) as well as  IKZF2  (Fig.  7e ;  Supplementary Information , Fig.  S7d, e ). In contrast,  EOMES +  ILC1, of which most had predicted rearrangements in α/β chains, clustered with T cells in close proximity to CXCR3 +  Th1 cells 41  (Fig.  7b–d ;  Supplementary Information , Fig.  S7c ). DE analysis unveiled that  EOMES +  ILC1 with predicted rearrangement in α/β chains showed high expression of α- and/or β-TCR constant and variable regions, as well as  CD8A  (Fig.  7e ;  Supplementary Information , Fig.  S7d, e ). Hence,  EOMES +  ILC1 are transcriptionally closely related to Th1 and cytotoxic T cells. Finally, we evaluated the clonality of ILCs vs T cells. As expected from healthy donor blood, most cells represented unique clones (Fig.  7f ;  Supplementary Information , Fig.  S7f ). Among ILC1 and  EOMES +  ILC1 with α/β rearrangement, there were six clones represented by two or more cells (Fig.  7f ;  Supplementary Information , Fig.  S7f ). Two of these clones were shared with T cells ( Supplementary Information , Fig.  S7g ). Fig. 7 Rearrangement of V(D)J T cell receptor genes in blood ILCs and T cells. a  Predicted V(D)J-rearrangement of  TRA ,  TRB ,  TRD  and/or  TRG  genes in annotated blood ILC clusters (B_ILC1, EOMES +  ILC1, B_ILC2 and B_nILC) and in donor matched CD4 +  T cells (T). TRA_B, rearrangement of  TRA  and  TRB  genes; TRG_D, rearrangement of  TRG  and  TRD  genes; TRA_B + G_D, rearrangement of  TRA  and/or  TCRB  in combination with  TRD  and/or  TRG  genes.  b  UMAP of blood ILCs and CD4 +  T cells color-coded on the basis of annotation in Fig.  1e  (B_ILC1,  EOMES +  ILC1, B_ILC2 and B_nILC) and CD4 +  T cells (T).  c  UMAP displaying each subset in  b  separately and color-coding the cells on the basis of predicted V(D)J-rearrangement of  TRA  and/or  TRB  (TRA_B),  TRG  and/or  TRD  (TRG_D) or  TRA  and/or  TCRB  in combination with  TRD  and/or  TRG  genes (TRA_B + G_D).  d  Annotation of clusters on the basis of cell composition and predicted TCR V(D)J-rearrangement pattern.  e  Violin plot representation of selected DE genes in AB_T_ILC1 cluster (top), GD_ILC1 (middle) and  EOMES +  ILC1 (bottom). Within each cluster, data is displayed separately for ILC1 and T cells, and also on the basis of TCR V(D)J-rearrangement pattern.  f  Number of cells with different clone frequencies (based on third complementarity region 3 (CDR3)) per rearranged TCR chain for all ILC1 (top) and CD4 +  T cells (bottom) in the different clusters in  d . Data are from three independent scRNA-seq experiments with one blood donor each. Total number of cells is 1036 (536 CD4 +  T cells and 473 ILCs).
+
+a  Predicted V(D)J-rearrangement of  TRA ,  TRB ,  TRD  and/or  TRG  genes in annotated blood ILC clusters (B_ILC1, EOMES +  ILC1, B_ILC2 and B_nILC) and in donor matched CD4 +  T cells (T). TRA_B, rearrangement of  TRA  and  TRB  genes; TRG_D, rearrangement of  TRG  and  TRD  genes; TRA_B + G_D, rearrangement of  TRA  and/or  TCRB  in combination with  TRD  and/or  TRG  genes.  b  UMAP of blood ILCs and CD4 +  T cells color-coded on the basis of annotation in Fig.  1e  (B_ILC1,  EOMES +  ILC1, B_ILC2 and B_nILC) and CD4 +  T cells (T).  c  UMAP displaying each subset in  b  separately and color-coding the cells on the basis of predicted V(D)J-rearrangement of  TRA  and/or  TRB  (TRA_B),  TRG  and/or  TRD  (TRG_D) or  TRA  and/or  TCRB  in combination with  TRD  and/or  TRG  genes (TRA_B + G_D).  d  Annotation of clusters on the basis of cell composition and predicted TCR V(D)J-rearrangement pattern.  e  Violin plot representation of selected DE genes in AB_T_ILC1 cluster (top), GD_ILC1 (middle) and  EOMES +  ILC1 (bottom). Within each cluster, data is displayed separately for ILC1 and T cells, and also on the basis of TCR V(D)J-rearrangement pattern.  f  Number of cells with different clone frequencies (based on third complementarity region 3 (CDR3)) per rearranged TCR chain for all ILC1 (top) and CD4 +  T cells (bottom) in the different clusters in  d . Data are from three independent scRNA-seq experiments with one blood donor each. Total number of cells is 1036 (536 CD4 +  T cells and 473 ILCs).
+
+In summary, in contrast to ILC2 and nILCs, the majority of blood ILC1 and  EOMES +  ILC1 are strikingly similar to CD4 +  or CD8 +  TCR α/β +  T cells, respectively, in terms of expression profile, TCR rearrangement and clonality. In contrast, blood ILC1 with preferential δ-TCR rearrangement constitute a subset distinct from T cells. Our conclusions were further strengthened when comparing the transcriptional profiles of ILCs to blood CD3 +  T cells and CD56 +  NK cells in a publically available scRNA-seq data set ( Supplementary Information , Fig.  S8 ), showing minimal similarity between blood ILC1 with preferential δ-TCR rearrangement and any of the T cell subsets ( Supplementary Information , Fig.  S8e, f ).
+
+Previous reports have described tissue-specific transcriptional imprinting of ILCs across multiple organs. However, these findings originated from pre-defined ILC subsets sorted in bulk from mice 2  or humans, 3  neglecting heterogeneity in the sorted populations. Here we used full-length scRNA-seq on individually sorted ILCs allowing for integrated analysis of FACS phenotype and transcriptome on the single-cell level. Additionally, we took a holistic and unbiased approach to analyzing the scRNA-seq data by complementing DE analysis with correlation analysis to identify gene modules, thereby going beyond only describing markers specific to each cell type. Through our approach, we have been able to make a number of novel observations, accompanied by a set of experimental functional studies, which advance our understanding of human ILC biology and serve as a resource to the immunology community.
+
+We identified two distinct ILC clusters that lacked markers of mature ILCs. In the blood, these cells expressed CD117, identifying them as the previously described circulating progenitors of all lineages of mature ILCs, including NK cells. 4 , 17  The other cluster consisted of predominantly CD117 +  ILC from colon and lung. In sharp contrast to circulating nILCs, these cells showed a transcriptome compatible with tissue-residency, possibly representing a resident precursor to mature ILCs in these tissues. nILCs are identified as ILC3 by flow cytometry, 3  calling for a better cell surface definition of these cells. Future studies should be aimed at targeted sorting and enrichment of these cells based on e.g., expression of CD117 and lack of markers associated with mature ILC subsets and activation such as CRTH2, HLA-DR and NKp44. This might reveal nILC subsets polarized to specific fates in certain tissues. Our in-depth unbiased transcriptional profiling of circulating and resident nILCs provide a road-map for such differentiation studies of these cells and their role in disease conditions.
+
+Parabiosis experiments in mice have suggested that ILCs in mucosal and adipose tissues are largely tissue resident 42  while other reports have unveiled a recirculating behavior of colon ILC2 during  N. brasiliensis  infection. 36  We identified gene modules particularly enriched in naïve and mature ILCs in the colon and lung including  CD69  and transcripts encoding TFs ( NR4A1-3 ,  STAT3 ,  RORA  and  RARA ) as well as transcripts involved in Notch-, Fos/Jun- and NFκB signaling, many of which have previously been detected in CD69 +  T cells and NK cells in humans and mice. 13 , 22  In sharp contrast to colon and lung, naïve and mature blood ILCs showed features of CD62L- and S1PR1-mediated recirculation, similar to that described in mice. 36  In the tonsil, we detected signs of both recirculation and tissue-residency. These data align well with a recent report on human ILC3 sorted in bulk from tonsils, spleen and blood. 4  Additionally, in mice, lymph nodes harbor both migratory and tissue-resident ILCs, the former more common for ILC1. 20  Although the tissue-residency and migratory profiles of mouse ILCs seem largely conserved in humans, we did detect some unique features of human ILCs in this regard. Most strikingly, human ILC2 uniquely expressed  S1PR2 , opening up for the possibility of selectively targeting this receptor to influence ILC2 migration during type 2 inflammation.
+
+The colonic mucosa is continuously exposed to environmental and intrinsic factors derived from the microbiota, diet and secretions in the digestive system. Indeed, we detected the highest expression of ILC3-related genes in this tissue, and also the highest degree of heterogeneity among ILC3. These ILC3 represented a spectrum of cells ranging from those with a more resting and recirculating phenotype, possibly constituting cells recently recruited from the circulation, to cells expressing genes encoding effector molecules, activation markers and molecules involved in metabolism. These data align well with a recent scRNA-seq study of ILCs in the mouse colon, 8  although the precise transcripts defining ILC3 subsets in the mouse colon was different to those we identified here. Notably, we did not identify ILC1-biased ILC3, as those recently reported in the human inflamed colon, 11  which is likely due to the non-inflamed nature of the intestinal tissue in our analysis. Furthermore, our data confirm previous reports that in stark contrast to the mouse, the human healthy colon lacks ILC2 as defined by surface protein expression. 16 , 43  Our data now further conclusively demonstrate that the human colon is also largely devoid of ILCs with a transcriptome reminiscent of ILC2. These results provide a blueprint of the homeostatic situation of ILCs in the human colon, aiding studies of the role for ILCs in colon immunopathology.
+
+In the lung, we identified ILC2 with a transcriptome indicative of activation, including expression of receptors of both IL-25 and IL-33, reminiscent of natural ILC2 described in mice. 44  The lung-specific transcriptome of ILC2 could be a result of the constant exposure of the lung to both extrinsic and intrinsic immune stimuli. Indeed, we could recapitulate the lung phenotype of ILC2, with high expression of receptors for IL-33 and IL-25, and low expression of CRTH2, by exposing circulating ILC2 to tissue-derived alarmins. This indicates that activated ILC2 might not necessarily express CRTH2, which has previously been suggested based on both mouse 35  and human ex vivo 37 , 45  and in vitro data. 38  Utilizing the power of our flow cytometry indexed scRNA-seq data we were indeed able to identify such CRTH2 –  ILC2 in the lung, expressing receptors for IL-33 and IL-25. In addition to downregulation of CRTH2 on ILC2, CRTH2 –  ILC2 might also be derived from ILC1 and/or ILC3 transdifferentiation. However, such ILC2 plasticity from bona fide ILC1 or ILC3 has, to the best of our knowledge, never been demonstrated in either mice or humans. Nevertheless, our unbiased identification of CRTH2 –  ILC2 challenges our understanding of human ILC2 biology, and has consequences for the identification of activated human ILC2 and strategies to target these cells in type 2 inflammation in e.g., allergy and asthma. The failure of the CRTH2 antagonist Fevipiprant in moderate-to-severe asthma 46 , 47  may, at least in part, be explained by our findings. The expression of the cell surface receptor SLAMF1 could be helpful to distinguish CRTH2 –  ILC2 from ILC1 and ILC3 in the lung.
+
+The existence of ILC1 in different human tissues in homeostasis and inflammation has been a topic of much debate. 48 – 51  While Tbet +  non-NK/T/B lymphocytes, fulfilling the definition of ILC1, clearly accumulate in inflamed human intestinal tissues, 16 , 52 , 53  Simoni et al. have suggested that in non-affected human tissues, ILC1s are artifacts derived from contamination, by e.g., T cells. In partial support of this, Roan et al. showed that a fraction (7%) of blood ILC1 upregulate surface expression of TCRα/β following a 7-day culture. 39  While this shows that indeed blood ILC1 might contain a small fraction of T cells that have downregulated their surface TCRs, the majority of blood ILC1 are likely not T cells in disguise despite expressing intracellular CD3 epsilon. 39  In fact, CD3 epsilon overexpression in the absence of TCR expression in mice has been suggested to block T cell development. 54  Furthermore, cytoplasmic CD3 epsilon expression might indicate that blood ILC1 are immature, as it is observed in fetal but not adult NK cells. 55  In further support of that, CD5 +  ILCs, expressing intracellular CD3 epsilon, have been suggested as precursors of mature IFN-γ-producing ILC1. 56  We identified a particularly pronounced T-cell signature, including  CD3E , encoding CD3 epsilon, in putative ILC1 and  EOMES +  ILC1 in blood and tonsil. We therefore analyzed TCR rearrangement to better understand ILC1-T cell relationships.  EOMES +  ILC1 and a subset of ILC1 showed pronounced rearrangement in TCRα/β chains as well as expression of  CD4  and  CD8A , respectively. Despite lacking surface expression of TCRs and CD3, these ILCs might very well be T cell contaminants, which was also supported by partial clonal overlap between TCRα/β-rearranged T cells and ILC1. Clonal overlap was however uncommon, and therefore it is also possible that these ILCs are “failed T cells” with non-productive TCRα/β- rearrangement, leading to a lack of surface TCR expression. Notably, in ILCs, the expression of TCRα/β variable chains was associated with TCRα/β rearrangement and could potentially be used as a proxy for rearrangement in these cells.
+
+We also identified a cluster of ILC1, which was transcriptionally distinct from T cells, with rearrangement in predominantly TCR-δ chains. Indeed, mouse NK cell precursors show V(D)J-recombination activity 57  and up to a third of human blood NK cells display rearrangement of δ-TCR chains. 58  Hence, similar to intracellular CD3 epsilon expression, 55  δ-chain rearrangement in blood ILCs might be a sign of immaturity and a history of recombination enzyme activity.
+
+In summary, our comparisons of ILCs and CD4 +  T cells in human blood unveiled clusters of putative ILC1 that show TCRα/β-rearrangement and clonal overlap with T cells. The latter are likely T cell contaminants with down-regulated surface TCRs while the remainder could be “failed T cells”, with non-productive TCR rearrangement. We also identified ILCs with TCR-δ rearrangement reminiscent of immature ILCs, rather than differentiated Tbet +  ILC1. Our findings provide important first steps towards more extensive, elaborate studies on rearrangement in subsets of T and ILCs/NK cells across the innate and adaptive landscape and the mechanisms, including development, underlying this unexpected TCR rearrangement.
+
+In conclusion, we used scRNA-seq to identify a plethora of ILC subsets ranging from circulating and tissue-resident nILCs to subsets of mature, activated and tissue-resident ILCs with altered metabolism across four tissues. Deepened analysis unveiled the presence of lung-specific CRTH2 −  ILC2, challenging our understanding of ILC2-mediated diseases. Furthermore, TCR-V(D)J rearrangement studies identified widespread predicted TCRα/β-rearrangement and a distinct subset of putative immature ILCs among the ILCs commonly referred to as ILC1 in peripheral blood. Our findings constitute a roadmap for future studies aimed at understanding tissue-specific ILC-mediated immunity in immunopathology.
+
+The Swedish ethical review authority approved all studies of human tissues and all studies were performed according to the Declaration of Helsinki. All regulations for work with human tissues were complied with. All study subjects gave their informed written consent prior to tissue collection.
+
+Lung tissue was obtained from four patients undergoing lobectomy for lung cancer (age 57–76 year old, three females, one male) at Karolinska University Hospital, Sweden. Non-affected tissue was obtained distally from the tumor. None of the patients had a diagnosis of asthma or chronic obstructive pulmonary disease (COPD). All patients were ex-smokers. None of the patients had received preoperative chemo- or radiotherapy. Lung cells were isolated from human lung tissue by a sequential combination of washing, enzymatic digestion, mechanical disruption and density centrifugation using Percoll gradient medium (WEMP protocol), as recently published by us 59  where all details of the protocol can be found. Briefly, freshly resected lung tissue was washed in PBS and cut into smaller pieces, then enzymatically digested while stirring. The resulting digested product was cut and mechanically disrupted by syringe plunging and filtering. The filtered cell suspension was washed in PBS, resuspended in 30% Percoll (one phase) and centrifuged. The cell pellet was resuspended and red blood cells were lysed and washed with PBS 2% fetal calf serum (FCS) and cells were counted. The cell suspension was then stained with fluorescence-labeled antibodies for FACS.
+
+Peripheral blood mononuclear cells (PBMCs) were isolated from whole venous blood from three healthy donors (age 29–32 year old, two males, one female). 27 mL of fresh blood was collected in EDTA tubes and directly diluted 1:1 in PBS. PBMC were isolated via lymphoprep (Axis Shield) and gradient centrifugation for 20 min at 2000× rpm. Lymphocytes were then collected, washed and counted. The cell suspension was then stained with fluorescence-labeled antibodies for FACS.
+
+Biopsies from the ascending colon were obtained from three tumor-screening patients (age 57–71 year old, two males, one female) with known genetic predisposition for colon cancer (hereditary nonpolyposis colorectal cancer (HPNCC) or Lynch Syndrome) but without any polyps or tumors at the time of sampling. Colonoscopies were performed at Karolinska University Hospital or GHP Stockholm Gastro Center, Sweden. Biopsy retrieval and cell isolation was performed as previously described, 16  where all the details of the protocol can be found. Briefly, eight punch-biopsies per patients were collected and transported in wash buffer, consisting of Hank’s Balanced Salt Solution (HBSS) supplemented with antibiotics, fungizone and HEPES. Samples were processed within 1 h of biopsy retrieval. Wash buffer was removed and biopsies were enzymatically digested while stirring. The filtered cell suspension was washed by centrifugation and the total cell pellet was used for staining with fluorescence-labeled antibodies for FACS.
+
+The previously published tonsil raw data used in this study can be found at GEO:  GSE70580 ; Sequence Read Archive: SRP060416.
+
+Isolated cells were incubated with live/dead cell marker (DCM; Invitrogen) and surface protein antibodies ( Supplementary Information , Table  S1 ) at 4 °C for 30 min, washed with PBS and resuspended in Iscove’s modified Dulbecco’s medium IMDM medium supplemented with 1% normal human serum (NHS; Invitrogen) and Yssels supplement (made in house). Cells were sorted into 384-well plates (Eppendorf) containing 2.3 μL of lysis buffer per well. Lysis buffer (per well) consisted of 1.150 μL of 0.4% Triton X-100 (Sigma), 0.4 μL dNTPs mix (25 mM) (Thermo Life Tech), 0.05 μL Smart-dT 30 VN (100uM) (IDT), 0.05 μL RNAse inhibitor (40 U/μL) (Takara), 0.0025 μL ERCC (1:40,000 dilution) (Ambion) and 0.6475 μL of water (Thermo Life Tech). Cells were sorted on an Aria FUSION sorter (BD Biosciences), with FACS Diva version 8 software and indexed data analysis enabled. Sorting of blood ILC2 were performed on a SONY MA900 Multi-Application cell sorter.
+
+Sort-purified peripheral blood ILC2 ( Supplementary Information , Fig.  S5 ) were cultured in IMDM (Fisher Scientific) supplemented with 2% NHS and Yssels supplement in a 96 well plate. Cells were left in culture for 48 h at 37 °C with IL-2 (10 U/mL) or IL-2 (10 U/mL), IL-25 (R&D), IL-33 (PeproTech) plus TSLP (PeproTech) (at 50 ng/mL each). Following culture, the cells were washed and stained with fluorescence-labeled antibodies for flow cytometry.
+
+For flow cytometry analysis, cells were stained with DCM and antibodies ( Supplementary Information , Table  S1 ) for 30 min at 4 °C. Cells were washed in PBS supplemented with 2 mM EDTA and 2% FCS and fixed with 2% paraformaldehyde for 10 min at room temperature. Cells were acquired on an LSR Fortessa flow cytometer (BD Biosciences) with FACSDiva software version 8. Data were analyzed with FlowJo software version 9.9.6 (Tree Star).
+
+Sorted cells were processed using the Smart-seq2 protocol with minor changes. 60  Sequencing was done to the average depth of 0.7 million reads per cell with 43 bp reads (single end) after demultiplexing. Blood samples were additionally sequenced for V(D)J analysis to and average depth of 0.84 million reads per cell with 150 bp reads (paired end) after demultiplexing.
+
+The reads were aligned to the human genome (hg19–GRCh38) and spike-in sequences (from the External RNA Controls Consortium) using STAR v2.5.3a, 61  and filtered for uniquely mapping reads. Counts per gene were calculated for each transcript in Ensembl release 92 using rpkmforgenes. 62
+
+Bioinformatics analyses were performed using R programming language, using Seurat v3 package 63  with modifications. Reads were normalized to transcripts per million reads (TPMs) in order to compensate for gene length in transcript count estimation. Values were then multiplied by the read length (43 bp) and rounded to the nearest integer. Cell libraries were considered of low quality and filtered out if: (1) the percentages of mitochondrial and ribosomal gene families were above 40% or 50%, respectively; (2) the percentage of protein-coding genes was below 50%; (3) the number of unique genes in a cell was below 200. Genes detected in less than five cells were excluded from the analysis. Counts were next log-normalized with scale factor of 1000 and Z-score transformed for regression of confounding factors such as effects related to mitochondrial, ribosomal, Y chromosome, cell cycle, sex and proportion of non-coding RNA.
+
+Normalized data were further processed for identification of highly variable genes using the Scran method 64  and linear data compression using principal component analysis (PCA) from which the top 50 principal components (PCs) explaining about 70% of the variance in the dataset were chosen for downstream analysis. Non-linear dimensionality reduction was run on the top 50 PCs using UMAP 15  and embedded into two final dimensions. Graph construction was also run on the top 50 PCs using approximate nearest neighbor (KNN) 65  search with k = 10, reflecting the minimum number of cells expected to be considered a potentially rare cell cluster. Spurious connections were filtered using shared nearest-neighbors (SNN) search using Jacquard’s index of 1/15. Unsupervised graph clustering was run using the Louvain method, in which the modularity resolution parameter (res = 2.4) was optimized to match both known sample metadata information as well as a coherent cluster distribution across the UMAP embedding.
+
+Identification of marker genes associated to specific cell clusters was done using differential expression via MAST. 66  For standard differential expression log 2 FC  > 0.25 and false discovery rate (FDR) < 0.01 was used.
+
+For identification of the most relevant genes across cell clusters to be later used for weighted gene correlation network analysis (WGCNA), 67  a permissive threshold with absolute log 2 FC > 0.1 and  P  value < 0.1 was used. Briefly, the 4410 genes identified in this manner were used for calculation of gene-to-gene Pearson correlations. The adjacency matrix was directly used for creation of the topological overlap matrix (TOM) using again Pearson as correlation measure. Power transformation of adjacency values were ignored since it would result in flattening all values to near 0, as even strong single-cell correlations rarely reach close to 1 due to the nature of data sparsity. No significant differences in the results were found when using Pearson and Spearman correlations, and therefore Pearson correlations were kept due to its faster overall performance. Gene module identification was done with agglomerative hierarchical clustering on the TOM using Ward’s squared linkage (Ward.D2) with a fixed number of 100 modules. Gene counts belonging to the same module were normalized between 0–1 individually and then averaged per cell for calculation of each gene module expression pattern.
+
+Modules were prioritized by fitting a general linear model (GLM) using each gene module expression pattern as the effect and the sample metadata as explaining variable. Each module and metadata parameter was tested independently. The output of the GLM fit results in the residual metric of “deviance” ( \documentclass[12pt]{minimal}
+				\usepackage{amsmath}
+				\usepackage{wasysym} 
+				\usepackage{amsfonts} 
+				\usepackage{amssymb} 
+				\usepackage{amsbsy}
+				\usepackage{mathrsfs}
+				\usepackage{upgreek}
+				\setlength{\oddsidemargin}{-69pt}
+				\begin{document}$$dev_{model}$$\end{document} d e v m o d e l ) after fitting the model and “null deviance” ( \documentclass[12pt]{minimal}
+				\usepackage{amsmath}
+				\usepackage{wasysym} 
+				\usepackage{amsfonts} 
+				\usepackage{amssymb} 
+				\usepackage{amsbsy}
+				\usepackage{mathrsfs}
+				\usepackage{upgreek}
+				\setlength{\oddsidemargin}{-69pt}
+				\begin{document}$$dev_{null}$$\end{document} d e v n u l l ) when considering a NULL model. The percent difference in these two deviances ( \documentclass[12pt]{minimal}
+				\usepackage{amsmath}
+				\usepackage{wasysym} 
+				\usepackage{amsfonts} 
+				\usepackage{amssymb} 
+				\usepackage{amsbsy}
+				\usepackage{mathrsfs}
+				\usepackage{upgreek}
+				\setlength{\oddsidemargin}{-69pt}
+				\begin{document}$${\Delta}dev$$\end{document} Δ d e v ) was used as a measure of how much a particular metadata factor explains the expression pattern of a gene module: \documentclass[12pt]{minimal}
+				\usepackage{amsmath}
+				\usepackage{wasysym} 
+				\usepackage{amsfonts} 
+				\usepackage{amssymb} 
+				\usepackage{amsbsy}
+				\usepackage{mathrsfs}
+				\usepackage{upgreek}
+				\setlength{\oddsidemargin}{-69pt}
+				\begin{document}$${\Delta}dev = \frac{{dev_{null} - dev_{model}}}{{dev_{null}}}$$\end{document} Δ d e v = d e v n u l l − d e v m o d e l d e v n u l l
+
+This way, modules with higher  \documentclass[12pt]{minimal}
+				\usepackage{amsmath}
+				\usepackage{wasysym} 
+				\usepackage{amsfonts} 
+				\usepackage{amssymb} 
+				\usepackage{amsbsy}
+				\usepackage{mathrsfs}
+				\usepackage{upgreek}
+				\setlength{\oddsidemargin}{-69pt}
+				\begin{document}$${\Delta}dev$$\end{document} Δ d e v  were prioritized for further exploration throughout the manuscript. Modules enriched for ribosomal, mitochondrial and other known housekeeping families were mentioned but not further analyzed.
+
+Trajectory inference analysis was performed using Slingshot v.1.2.0 25  on the Diffusion Map embedding (using five components), since the UMAP embedding creates a distorted representation of the original space. Lineages were defined by setting the naïve clusters (nILC) as the root, and curves were projected onto the Diffusion Map embedding for visualization.
+
+For blood data with 150 bp PE sequencing, MiXCR 40  was run on all cells using the “analyze shotgun” command and default settings. To avoid false positives MiXCR output was filtered for clones with count > 5. If multiple predicted clones were found, only dominant predicted clone from each chain (TRA, TRB, TRD or TRG) was included in downstream analysis.
+
+Human PBMC data with 10× Genomics scRNA-seq, cell surface protein detection with antibody-derived tags (ADT) and V(D)J assay was downloaded from the 10× genomics website ( https://support.10xgenomics.com/single-cell-vdj/datasets/3.0.0/vdj_v1_hs_pbmc2_5gex_protein ). Dimensionality reduction and clustering was performed with Seurat v3.1.3 with the whole dataset. T-cell and NK-cell clusters were selected based on the protein expression, e.g. CD3 +  and CD56 + CD3 –  clusters were selected for further analysis. Clustering and dimensionality reduction was performed again with this subset, as shown in Supplementary Fig.  8a, b  and cell types annotated based on top DEGs using MAST ( Supplementary Information , Fig.  S8c ). At this stage, a contaminating cluster 8 that is low in CD3 and CD56, but high in CD16 and CD19, was removed as it likely represented doublets or cells that were wrongly clustered in the first round.
+
+To compare the 10× PBMC clusters to our blood SS2 data, we looked at intersections of DEGs. All significant upregulated DEGs ( P  < 0.01) for each cluster within the 10× data and SS2 data were calculated separately. In the case of the AB_T_ILC1 cluster, cells were separated into T-cells (AB_T) and ILCs (AB_ILC) and DEG analysis was performed separately for those two groups. For each comparison, phyper test  P  values were calculated for overlap between the gene lists with all genes expressed in five cells in both datasets as background.
+
+For correlation between 10× PBMC and blood SS2 data, top 100 DEGs (or fewer if not 100 significant genes) from both datasets were used to calculate mean expression for each cluster. Spearman correlations between the clusters in each dataset were then calculated for the mean expression.
+
+Supplementary Text
+
+Supplementary Figure S1
+
+Supplementary Figure S2
+
+Supplementary Figure S3
+
+Supplementary Figure S4
+
+Supplementary Figure S5
+
+Supplementary Figure S6
+
+Supplementary Figure S7
+
+Supplementary Figure S8
+
+Supplementary Data S1
+
+Supplementary Data S2
+
+Supplementary Table S1
\ No newline at end of file
diff --git a/paper-releases/pbisc-diffusion/updates/20261015T141849Z-rare-cell-7k-8k-panels/diffusion/dataset/server_1/model8/plans/rare_gene_context_expansion_20261015/research/ilc_dc/sources/PMC8089104.xml b/paper-releases/pbisc-diffusion/updates/20261015T141849Z-rare-cell-7k-8k-panels/diffusion/dataset/server_1/model8/plans/rare_gene_context_expansion_20261015/research/ilc_dc/sources/PMC8089104.xml
new file mode 100644
index 0000000000000000000000000000000000000000..63777d0ef5271b580238e533e26cbc17238eea6b
--- /dev/null
+++ b/paper-releases/pbisc-diffusion/updates/20261015T141849Z-rare-cell-7k-8k-panels/diffusion/dataset/server_1/model8/plans/rare_gene_context_expansion_20261015/research/ilc_dc/sources/PMC8089104.xml
@@ -0,0 +1,54 @@
+
+<!DOCTYPE article
+  PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD with MathML3 v1.4 20241031//EN" "JATS-archivearticle1-4-mathml3.dtd">
+<article article-type="research-article" xml:lang="en" dtd-version="1.4"><front><journal-meta><journal-id journal-id-type="nlm-ta">Cell Res</journal-id><journal-id journal-id-type="iso-abbrev">Cell Res</journal-id><journal-id journal-id-type="pmc-domain-id">1475</journal-id><journal-id journal-id-type="pmc-domain">cellres</journal-id><journal-id journal-id-type="nlm-id">9425763</journal-id><journal-title-group><journal-title>Cell Research</journal-title></journal-title-group><issn pub-type="ppub">1001-0602</issn><issn pub-type="epub">1748-7838</issn><?publisher_abbrev naturepg?><publisher><publisher-name>Nature Publishing Group</publisher-name></publisher></journal-meta><article-meta><article-id pub-id-type="pmcid">PMC8089104</article-id><article-id pub-id-type="pmcid-ver">PMC8089104.1</article-id><article-id pub-id-type="pmcaid">8089104</article-id><article-id pub-id-type="pmcaiid">8089104</article-id><article-id pub-id-type="pmid">33420427</article-id><article-id pub-id-type="doi">10.1038/s41422-020-00445-x</article-id><article-id pub-id-type="publisher-id">445</article-id><article-version article-version-type="pmc-version">1</article-version><article-categories><subj-group subj-group-type="heading"><subject>Article</subject></subj-group></article-categories><title-group><article-title>Tissue-specific transcriptional imprinting and heterogeneity in human innate lymphoid cells revealed by full-length single-cell RNA-sequencing</article-title></title-group><contrib-group><contrib contrib-type="author"><name name-style="western"><surname>Mazzurana</surname><given-names initials="L">Luca</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid" authenticated="false">http://orcid.org/0000-0001-8150-4021</contrib-id><name name-style="western"><surname>Czarnewski</surname><given-names initials="P">Paulo</given-names></name><xref ref-type="aff" rid="Aff2">2</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Jonsson</surname><given-names initials="V">Viktor</given-names></name><xref ref-type="aff" rid="Aff3">3</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Wigge</surname><given-names initials="L">Leif</given-names></name><xref ref-type="aff" rid="Aff3">3</xref></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid" authenticated="false">http://orcid.org/0000-0001-5469-8940</contrib-id><name name-style="western"><surname>Ringnér</surname><given-names initials="M">Markus</given-names></name><xref ref-type="aff" rid="Aff4">4</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Williams</surname><given-names initials="TC">Teresa C.</given-names></name><xref ref-type="aff" rid="Aff1">1</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Ravindran</surname><given-names initials="A">Avinash</given-names></name><xref ref-type="aff" rid="Aff5">5</xref></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid" authenticated="false">http://orcid.org/0000-0003-2224-7090</contrib-id><name name-style="western"><surname>Björklund</surname><given-names initials="ÅK">Åsa K.</given-names></name><xref ref-type="aff" rid="Aff6">6</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Säfholm</surname><given-names initials="J">Jesper</given-names></name><xref ref-type="aff" rid="Aff7">7</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Nilsson</surname><given-names initials="G">Gunnar</given-names></name><xref ref-type="aff" rid="Aff5">5</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Dahlén</surname><given-names initials="SE">Sven-Erik</given-names></name><xref ref-type="aff" rid="Aff7">7</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Orre</surname><given-names initials="AC">Ann-Charlotte</given-names></name><xref ref-type="aff" rid="Aff8">8</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Al-Ameri</surname><given-names initials="M">Mamdoh</given-names></name><xref ref-type="aff" rid="Aff8">8</xref><xref ref-type="aff" rid="Aff9">9</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Höög</surname><given-names initials="C">Charlotte</given-names></name><xref ref-type="aff" rid="Aff10">10</xref></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid" authenticated="false">http://orcid.org/0000-0002-4921-8516</contrib-id><name name-style="western"><surname>Hedin</surname><given-names initials="C">Charlotte</given-names></name><xref ref-type="aff" rid="Aff10">10</xref><xref ref-type="aff" rid="Aff11">11</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Szczegielniak</surname><given-names initials="S">Sylwester</given-names></name><xref ref-type="aff" rid="Aff11">11</xref></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid" authenticated="false">http://orcid.org/0000-0001-9334-1821</contrib-id><name name-style="western"><surname>Almer</surname><given-names initials="S">Sven</given-names></name><xref ref-type="aff" rid="Aff10">10</xref><xref ref-type="aff" rid="Aff11">11</xref></contrib><contrib contrib-type="author" corresp="yes"><name name-style="western"><surname>Mjösberg</surname><given-names initials="J">Jenny</given-names></name><address><email>jenny.mjosberg@ki.se</email></address><xref ref-type="aff" rid="Aff1">1</xref></contrib><aff id="Aff1"><label>1</label>Department of Medicine, Center for Infectious Medicine, Karolinska Institutet, Stockholm, Sweden </aff><aff id="Aff2"><label>2</label>Science for Life Laboratory, Department of Biochemistry and Biophysics, National Bioinformatics Infrastructure Sweden, Stockholm University, Solna, Sweden </aff><aff id="Aff3"><label>3</label>Science for Life Laboratory, Department of Biology and Biological Engineering, National Bioinformatics Infrastructure Sweden, Chalmers University of Technology, Gothenburg, Sweden </aff><aff id="Aff4"><label>4</label>Science for Life Laboratory, Department of Biology, National Bioinformatics Infrastructure Sweden, Lund University, Lund, Sweden </aff><aff id="Aff5"><label>5</label>Immunology and Allergy Unit, Department of Medicine Solna, Department of Clinical Immunology and Transfusion Medicine, Karolinska Institutet, Karolinska University Hospital, Stockholm, Sweden </aff><aff id="Aff6"><label>6</label>Science for Life Laboratory, Department of Cell and Molecular Biology, National Bioinformatics Infrastructure Sweden, Uppsala University, Uppsala, Sweden </aff><aff id="Aff7"><label>7</label>Institutet of Environmental Medicine, Center for Allergy Research, Karolinska Institutet, Stockholm, Sweden </aff><aff id="Aff8"><label>8</label>Division of Thoracic Surgery, Karolinska University Hospital, Stockholm, Sweden </aff><aff id="Aff9"><label>9</label>Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden </aff><aff id="Aff10"><label>10</label>Department of Medicine Solna, Karolinska Institutet, Stockholm, Sweden </aff><aff id="Aff11"><label>11</label>Division of Gastroenterology, Medical Unit Gastroenterology, Dermatovenereology and Rheumatology, Karolinska University Hospital, Stockholm, Sweden </aff></contrib-group><pub-date pub-type="epub"><day>8</day><month>1</month><year>2021</year></pub-date><pub-date pub-type="ppub"><month>5</month><year>2021</year></pub-date><volume>31</volume><issue>5</issue><issue-id pub-id-type="pmc-issue-id">380802</issue-id><fpage>554</fpage><lpage>568</lpage><history><date date-type="received"><day>4</day><month>6</month><year>2020</year></date><date date-type="accepted"><day>4</day><month>11</month><year>2020</year></date></history><pub-history><event event-type="pmc-release"><date><day>08</day><month>01</month><year>2021</year></date></event><event event-type="pmc-live"><date><day>05</day><month>05</month><year>2021</year></date></event><event event-type="pmc-last-change"><date iso-8601-date="2026-04-05 17:25:13.390"><day>05</day><month>04</month><year>2026</year></date></event></pub-history><permissions><copyright-statement>© The Author(s) 2021</copyright-statement><license><ali:license_ref xmlns:ali="http://www.niso.org/schemas/ali/1.0/" specific-use="textmining" content-type="ccbylicense">https://creativecommons.org/licenses/by/4.0/</ali:license_ref><license-p><bold>Open Access</bold> This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="https://creativecommons.org/licenses/by/4.0/">http://creativecommons.org/licenses/by/4.0/</ext-link>.</license-p></license></permissions><self-uri xmlns:xlink="http://www.w3.org/1999/xlink" content-type="pmc-pdf" xlink:href="41422_2020_Article_445.pdf"><?pdf-name 41422_2020_Article_445.pdf?><?pdf-size 6328824?><?pdf-md5 84eb563570c501bfbdbc0510e2dc5ec5?><?pdf-image-server-status NEVER_LOAD?><?pdf-cloudpmc-urn urn:app:f93f/8089104/84eb563570c5/41422_2020_Article_445.pdf?></self-uri><abstract id="Abs1"><p id="Par1">The impact of the microenvironment on innate lymphoid cell (ILC)-mediated immunity in humans remains largely unknown. Here we used full-length Smart-seq2 single-cell RNA-sequencing to unravel tissue-specific transcriptional profiles and heterogeneity of CD127<sup>+</sup> ILCs across four human tissues. Correlation analysis identified gene modules characterizing the migratory properties of tonsil and blood ILCs, and signatures of tissue-residency, activation and modified metabolism in colon and lung ILCs. Trajectory analysis revealed potential differentiation pathways from circulating and tissue-resident naïve ILCs to a spectrum of mature ILC subsets. In the lung we identified both CRTH2<sup>+</sup> and CRTH2<sup>−</sup> ILC2 with lung-specific signatures, which could be recapitulated by alarmin-exposure of circulating ILC2. Finally, we describe unique TCR-V(D)J-rearrangement patterns of blood ILC1-like cells, revealing a subset of potentially immature ILCs with TCR-δ rearrangement. Our study provides a useful resource for in-depth understanding of ILC-mediated immunity in humans, with implications for disease.</p></abstract><kwd-group kwd-group-type="npg-subject"><title>Subject terms</title><kwd>Innate immunity</kwd><kwd>Bioinformatics</kwd></kwd-group><funding-group><award-group><funding-source><institution-wrap><institution-id institution-id-type="FundRef">https://doi.org/10.13039/501100004063</institution-id><institution>Knut och Alice Wallenbergs Stiftelse (Knut and Alice Wallenberg Foundation)</institution></institution-wrap></funding-source><award-id>2014.0244</award-id><award-id>2014.0244</award-id><principal-award-recipient><name name-style="western"><surname>Mazzurana</surname><given-names>Luca</given-names></name><name name-style="western"><surname>Mjösberg</surname><given-names>Jenny</given-names></name></principal-award-recipient></award-group></funding-group><funding-group><award-group><funding-source><institution-wrap><institution-id institution-id-type="FundRef">https://doi.org/10.13039/501100002794</institution-id><institution>Cancerfonden (Swedish Cancer Society)</institution></institution-wrap></funding-source><award-id>16-0664, 2016/1134</award-id><principal-award-recipient><name name-style="western"><surname>Mjösberg</surname><given-names>Jenny</given-names></name></principal-award-recipient></award-group></funding-group><funding-group><award-group><funding-source><institution-wrap><institution-id institution-id-type="FundRef">https://doi.org/10.13039/501100004359</institution-id><institution>Vetenskapsrådet (Swedish Research Council)</institution></institution-wrap></funding-source><award-id>5521-2013-2791</award-id><principal-award-recipient><name name-style="western"><surname>Mazzurana</surname><given-names>Luca</given-names></name></principal-award-recipient></award-group></funding-group><custom-meta-group><custom-meta><meta-name>pmc-status-qastatus</meta-name><meta-value>0</meta-value></custom-meta><custom-meta><meta-name>pmc-status-live</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-status-embargo</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-status-released</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-open-access</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-olf</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-manuscript</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-legally-suppressed</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-has-pdf</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-has-supplement</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-pdf-only</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-suppress-copyright</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-is-real-version</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-is-scanned-article</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-preprint</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-in-epmc</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-license-ref</meta-name><meta-value>CC BY</meta-value></custom-meta><custom-meta><meta-name>issue-copyright-statement</meta-name><meta-value>© Center for Excellence in Molecular Cell Science, CAS 2021</meta-value></custom-meta></custom-meta-group></article-meta></front><body><sec id="Sec1" sec-type="introduction"><title>Introduction</title><p id="Par2">Innate lymphoid cells (ILCs) are a recent addition to innate immunity, playing an important role in clearing intra- and extracellular pathogens, regulating adaptive immunity, tissue repair and contributing to inflammation.<sup><xref ref-type="bibr" rid="CR1">1</xref></sup> Their defining feature is a lack of rearranged antigen-specific receptors, which distinguishes them from T cells. However, in other ways these cells closely mirror the function of their adaptive counterparts, being an essential source of cytokines related to type 1, 2 and 17 immune responses.<sup><xref ref-type="bibr" rid="CR1">1</xref></sup> Currently, ILCs are classified into five subsets: ILC1, ILC2, ILC3, Natural killer (NK) cells and lymphoid tissue inducer (LTi) cells. ILC1 produce interferon-gamma (IFN-γ) and are dependent on the T-box transcription factor (TF) TBX21 (Tbet) for their development and function. NK cells share these features with ILC1, but they additionally express the TF Eomesodermin homolog (Eomes) and have the ability to produce cytolytic and cytotoxic molecules.<sup><xref ref-type="bibr" rid="CR1">1</xref></sup> ILC2s produce IL-13 and IL-5, and are dependent on GATA-3.<sup><xref ref-type="bibr" rid="CR1">1</xref></sup> Both LTi cells and ILC3 can produce IL-17 and IL-22 cytokines, and are dependent on the nuclear receptor ROR-γt.<sup><xref ref-type="bibr" rid="CR1">1</xref></sup> Transcriptional analysis of ILCs obtained from different human and mouse tissues has revealed tissue-specific transcriptional imprinting.<sup><xref ref-type="bibr" rid="CR2">2</xref>–<xref ref-type="bibr" rid="CR4">4</xref></sup> However, these studies relied on biased definition of ILC populations sorted in bulk, masking potential heterogeneity and misdefinition of ILC populations across tissues. Recent developments in single-cell technologies on the transcriptional and protein level have rapidly advanced our understanding of the heterogeneity of the human immune system.<sup><xref ref-type="bibr" rid="CR5">5</xref></sup> While protein analyses still suffer from limitations, e.g., in terms of cell number requirements, single-cell RNA-sequencing (scRNA-seq) enables an unbiased, transcriptome-wide analysis of even small numbers of cells. So far, scRNA-seq of specific subsets of ILCs (NK cells,<sup><xref ref-type="bibr" rid="CR6">6</xref></sup> ILC2<sup><xref ref-type="bibr" rid="CR7">7</xref></sup>) or in specific tissues<sup><xref ref-type="bibr" rid="CR8">8</xref>,<xref ref-type="bibr" rid="CR9">9</xref></sup> has been performed in mice. Similarly, in humans, scRNA-seq has been restricted to one subset (NK cells<sup><xref ref-type="bibr" rid="CR6">6</xref></sup>) or one tissue (ileum<sup><xref ref-type="bibr" rid="CR10">10</xref>,<xref ref-type="bibr" rid="CR11">11</xref></sup> and tonsil<sup><xref ref-type="bibr" rid="CR12">12</xref></sup>) at a time, yielding limited understanding of tissue-specific imprinting of ILCs.</p><p id="Par3">In this study, we analyzed human ILCs through full-length scRNA-seq, combining data from four different tissues: blood, tonsil, lung and colon. Comparison between tissues revealed tissue-specific transcriptional imprinting, and gave us the possibility to, for the first time in humans, compare scRNA-seq profiles of ILCs in the circulation, secondary lymphoid organs (SLOs) and mucosal tissues. Our data, which is provided as a publically available resource, reveal previously unknown functions of human ILCs and extend our current understanding of human ILCs biology.</p></sec><sec id="Sec2" sec-type="results"><title>Results</title><sec id="Sec3"><title>ScRNA-seq reveals subset- and tissue-specific transcriptional imprinting</title><p id="Par4">Transcriptional profiling of ILCs across tissues on the bulk<sup><xref ref-type="bibr" rid="CR2">2</xref>,<xref ref-type="bibr" rid="CR3">3</xref></sup> or single-cell level<sup><xref ref-type="bibr" rid="CR6">6</xref>,<xref ref-type="bibr" rid="CR7">7</xref>,<xref ref-type="bibr" rid="CR9">9</xref></sup> in humans and mice have revealed unique features of ILCs that are dependent on the tissue microenvironment. Here we performed flow cytometric sorting of human CD127<sup>+</sup> ILCs from three tissues representing two distinct immune cell compartments: circulation (blood), and mucosal tissue (lung and colon) (3–4 donors/tissue) (Fig. <xref rid="Fig1" ref-type="fig">1a</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref> Fig. <xref rid="MOESM1" ref-type="media">S1</xref>). As reference populations we included selected NK populations along the spectrum of previously described tissue NK cell phenotypes. Predominantly recirculating CD56<sup>dim</sup>CD16<sup>+</sup> NK cells were sorted from the lung<sup><xref ref-type="bibr" rid="CR13">13</xref></sup> and predominantly tissue-resident NK cells were sorted as bulk CD56<sup>+</sup> from the gut<sup><xref ref-type="bibr" rid="CR14">14</xref></sup> or as CD56<sup>bright</sup>NKG2A<sup>+</sup> NK cells from the tonsil.<sup><xref ref-type="bibr" rid="CR12">12</xref></sup> Cells were next processed for full-length Smart-seq2 scRNA-seq to unravel in-depth tissue-specific transcriptional profiles and heterogeneity. Our own previously published Smart-seq2 scRNA-seq dataset of CD127<sup>+</sup> ILC and NK cells from tonsils,<sup><xref ref-type="bibr" rid="CR12">12</xref></sup> representing a SLO, was included in the analysis. All datasets were combined and analyzed together in order to identify common and distinct patterns across tissues and cell types (Fig. <xref rid="Fig1" ref-type="fig">1a</xref>). Low quality cells and a minor fraction of contaminating cells lacking expression of <italic toggle="yes">PTPRC</italic> (encoding CD45) and expressing transcripts of antigen-presenting cells (e.g., high expression of <italic toggle="yes">HLA-DQA1</italic>, <italic toggle="yes">HLA-DRA</italic> and HLA-<italic toggle="yes">DPA1</italic> at levels exceeding ILC) were identified and removed (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S2</xref>), resulting in the analysis of a total of 2956 cells distributed across four tissues (Fig. <xref rid="Fig1" ref-type="fig">1b</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S2</xref>).<fig id="Fig1" position="float" orientation="portrait"><label>Fig. 1</label><caption><title>ScRNA-seq reveals subset and tissue-specific transcriptional imprinting.</title><p><bold>a</bold> Summary of the experimental workflow. Arrows in data integration and analysis section visualize the concept of immune cell migration between the tissues analyzed. Tonsil data was obtained from<sup><xref ref-type="bibr" rid="CR12">12</xref></sup> using the same scRNA-seq platform. <bold>b</bold> UMAP visualization of all cells (<italic toggle="yes">n</italic> = 2956) color-coded on the basis of tissue origin. <bold>c</bold> UMAP visualization of all cells, color-coded on the basis of cell surface phenotype inferred from indexed FACS sorting. <bold>d</bold> UMAP showing unbiased graph-based clustering analysis of all cells. <bold>e</bold> UMAP of annotated clusters. C_, colon; L_, lung; B_, blood; T_, tonsil; nILC, naïve ILC. <bold>f</bold> Dotplot displaying expression of selected and previously described human ILC subset-specific TFs and cell surface markers used to annotate clusters in <bold>e</bold>. <bold>g</bold> Barplot showing the representation of each donor in each annotated cluster in <bold>e</bold>. Data are from ten independent experiments with one tissue donor each (blood = 3, lung = 4 and colon = 3) integrated with data from.<sup><xref ref-type="bibr" rid="CR12">12</xref></sup></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" id="d32e649" position="float" orientation="portrait" xlink:href="41422_2020_445_Fig1_HTML.jpg"><?image-name 41422_2020_445_Fig1_HTML.jpg?><?image-size 174414?><?image-md5 80e4033aee5b22845a8989102508fbe7?><?image-image-server-status LOAD_COMPLETED?><?image-original-height 1998?><?image-original-width 1525?><?image-scaled-height 998?><?image-scaled-width 762?><?image-cloudpmc-urn urn:cdn:blobs/f93f/8089104/80e4033aee5b/41422_2020_445_Fig1_HTML.jpg?><?thumb-name 41422_2020_445_Fig1_HTML.gif?><?thumb-size 6334?><?thumb-md5 88723c03405759e4dd36e8930fbf281d?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 131?><?thumb-scaled-width 100?><?thumb-cloudpmc-urn urn:cdn:blobs/f93f/8089104/88723c034057/41422_2020_445_Fig1_HTML.gif?></graphic></fig></p><p id="Par5">Cell embedding using uniform manifold approximation and projection (UMAP)<sup><xref ref-type="bibr" rid="CR15">15</xref></sup> revealed that all ILC subsets as well as NK cells separated according to tissue of origin and FACS cell subset annotation as inferred from index sorting (Fig. <xref rid="Fig1" ref-type="fig">1b, c</xref>). Unbiased graph-based clustering subdivided cells into 20 clusters that segregated cells both by tissue of origin and FACS phenotype (Fig. <xref rid="Fig1" ref-type="fig">1d</xref>). On average, we observed 90% agreement between protein FACS phenotype and the clusters identified in scRNA-seq (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S3a</xref>). We next used previously described human ILC subset-specific TFs and cell surface markers<sup><xref ref-type="bibr" rid="CR12">12</xref></sup> to annotate the clusters (Fig. <xref rid="Fig1" ref-type="fig">1e, f</xref>). Clusters 5 (lung), 15 (tonsil) and 19 (colon), expressing <italic toggle="yes">NKG7</italic>, <italic toggle="yes">EOMES</italic> and <italic toggle="yes">GZMA/B</italic> but lacking <italic toggle="yes">IL7R</italic> were annotated as NK cells. Clusters 8 (blood), 13 (tonsil), 18 (lung) expressed <italic toggle="yes">IKZF3</italic> (encoding the TF and zinc finger protein Aiolos) but lacked <italic toggle="yes">EOMES</italic>, and <italic toggle="yes">KIR2DL1/3</italic>, and were conditionally annotated as ILC1. Cluster 17 consisted of both tonsil and blood CD127<sup>+</sup> ILCs, which expressed low levels of <italic toggle="yes">EOMES</italic>, conditionally annotating them as <italic toggle="yes">EOMES</italic><sup>+</sup> ILC1. Clusters 6 (blood), 11 (tonsil) and 12 (lung) were annotated as ILC2 on the basis of uniquely high <italic toggle="yes">GATA3</italic>, <italic toggle="yes">MAF</italic>, <italic toggle="yes">PTGDR2</italic> and <italic toggle="yes">HPGDS</italic> expression (Fig. <xref rid="Fig1" ref-type="fig">1e, f</xref>). As predicted by the low frequency of human colon ILC1 and ILC2 in non-inflamed colon tissue,<sup><xref ref-type="bibr" rid="CR16">16</xref></sup> colon cells with a FACS phenotype of ILC1 and ILC2 (total 10 cells) did not generate distinct clusters. ILC3s were annotated on the basis of <italic toggle="yes">RORC</italic>, <italic toggle="yes">IL23R</italic>, <italic toggle="yes">IL1R1</italic> and <italic toggle="yes">KIT</italic> expression (cluster 4 and 9 in tonsil, 2 and 10 in lung and 0, 1, 3 and 14 in colon) (Fig. <xref rid="Fig1" ref-type="fig">1e, f</xref>). We also observed <italic toggle="yes">IL7R</italic> and <italic toggle="yes">KIT</italic>-expressing ILCs in the blood (cluster 7) as well as in colon and lung (cluster 16), which lacked transcripts of all master TFs (Fig. <xref rid="Fig1" ref-type="fig">1e, f</xref>), reminiscent of the previously described ILC progenitors in blood and tissues.<sup><xref ref-type="bibr" rid="CR12">12</xref>,<xref ref-type="bibr" rid="CR17">17</xref></sup> These cells were annotated as naïve ILCs (nILCs), consistent with the nomenclature for naïve T cells.<sup><xref ref-type="bibr" rid="CR18">18</xref></sup> Importantly, cells from all donors of each tissue were present in all annotated cell types (Fig. <xref rid="Fig1" ref-type="fig">1g</xref>) and clusters (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S3b, c</xref>) for each tissue.</p></sec><sec id="Sec4"><title>Migration- and tissue-resident transcriptional profiles of ILCs</title><p id="Par6">To identify conserved gene expression patterns that are shared across tissues and cell types, we used a two-step approach rather than simple differential expression (DE), which reveals differences but not similarities between cell types and tissues. DE gene lists for each annotated cell type and the overlaps between cell types and tissues are provided in <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Data <xref rid="MOESM1" ref-type="media">S1</xref>. For the two-step approach, we selected the top DE genes in any of the 20 clusters (4410 genes) and subsequently performed weighted gene co-expression network analysis (WGCNA) for identification of 100 gene modules (see Materials and Methods for details) (Fig. <xref rid="Fig2" ref-type="fig">2a, b</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Data <xref rid="MOESM1" ref-type="media">S1</xref>). A high number of modules were chosen in order to avoid forcing genes with different expression patterns into the same group. Modules with highly correlated expression are indicated with a connecting line in the graph (Fig. <xref rid="Fig2" ref-type="fig">2c</xref>). Next, modules were scored individually for their contribution to discriminating tissues, cell types and/or cell clusters (Fig. <xref rid="Fig2" ref-type="fig">2c</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S3d</xref>). Notably, gene modules did not contribute equally to explaining metadata and thus only modules that contributed to at least one cell type or tissue were selected for further in-depth analysis herein (Fig. <xref rid="Fig2" ref-type="fig">2c</xref>).<fig id="Fig2" position="float" orientation="portrait"><label>Fig. 2</label><caption><title>Migration and tissue-resident transcriptional profile of ILCs.</title><p><bold>a</bold> Summary of analysis approach used to identify gene modules and perform trajectory analysis. <bold>b</bold> UMAP visualization of all cells, color-coded on the basis of tissue origin. <bold>c</bold> Circular dendrogram illustrating the number and composition of each module and the contribution of each module to discriminating cell clusters, tissues and cell types (as inferred from FACS). Modules with highly correlated expressions are indicated with a 5-NN graph (connecting lines represent KNN; k nearest neighbor). Modules described in text are marked with an asterisk. <bold>d</bold> Modules highly correlated with tissues and/or cell types/clusters containing genes involved in migration or tissue residency. The figure shows a selection of genes per module, and their levels of expression within annotated clusters (Fig. <xref rid="Fig1" ref-type="fig">1e</xref>). Data in <bold>a</bold>–<bold>d</bold> is from ten independent experiments with one tissue donor each (blood = 3, lung = 4 and colon = 3) integrated with data from.<sup><xref ref-type="bibr" rid="CR12">12</xref></sup>
+<bold>e</bold> Co-expression of CD69 and CD18/CD62L (top row) and NKp44 and CD18/CD62L (bottom row) on total ILCs from blood, tonsil and colon biopsies. <bold>f</bold> Histogram showing the expression of CD18, CD62L and CD69 on mechanically or enzymatically isolated total tonsil ILCs. FMO, fluorescence minus one. <bold>g</bold> Bar plot quantification of the results shown in <bold>e</bold>. MFI, mean fluorescence intensity. Data in <bold>e</bold>–<bold>g</bold> are from six independent experiments with at least one tissue donor each (blood = 6, tonsil = 8 and colon = 6).</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" id="d32e866" position="float" orientation="portrait" xlink:href="41422_2020_445_Fig2_HTML.jpg"><?image-name 41422_2020_445_Fig2_HTML.jpg?><?image-size 267331?><?image-md5 b38ed384f249cc4886b3e4c24bb83129?><?image-image-server-status LOAD_COMPLETED?><?image-original-height 2369?><?image-original-width 1775?><?image-scaled-height 948?><?image-scaled-width 710?><?image-cloudpmc-urn urn:cdn:blobs/f93f/8089104/b38ed384f249/41422_2020_445_Fig2_HTML.jpg?><?thumb-name 41422_2020_445_Fig2_HTML.gif?><?thumb-size 8811?><?thumb-md5 a9514634d86777d7325409fba63f1c0e?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 133?><?thumb-scaled-width 100?><?thumb-cloudpmc-urn urn:cdn:blobs/f93f/8089104/a9514634d867/41422_2020_445_Fig2_HTML.gif?></graphic></fig></p><p id="Par7">The analysis of four different tissues that represent unique immune niches (circulation, lymphoid- and mucosal tissue) in combination with the unbiased gene module analysis allowed us to identify conserved patterns across tissues and cell types (Fig. <xref rid="Fig2" ref-type="fig">2d</xref>). We identified two gene modules (41 and 11) containing transcripts highly correlated to the expression of the prototypical T cell migratory molecules <italic toggle="yes">S1PR1</italic> and <italic toggle="yes">SELL</italic> (encoding L-selectin; CD62L), respectively (Fig. <xref rid="Fig2" ref-type="fig">2d</xref>). Genes in these two modules were highly expressed in tonsil and blood ILCs, including nILCs. The modules included the glycolysis-associated transcript <italic toggle="yes">LDHB</italic>, the adhesion molecule transcript <italic toggle="yes">ICAM3</italic>, as well as transcripts for several S100 proteins (<italic toggle="yes">S100A4</italic>, <italic toggle="yes">S100A6</italic> and <italic toggle="yes">S100A11</italic>) (Fig. <xref rid="Fig2" ref-type="fig">2d</xref>), shown to be involved in migration.<sup><xref ref-type="bibr" rid="CR19">19</xref></sup> Module 23 was also over-represented in blood and tonsil and contained transcripts for several integrins including <italic toggle="yes">ITGB2</italic> (CD18) <italic toggle="yes">ITGAL</italic> (CD11a) and <italic toggle="yes">ITGAM</italic> (CD11b) (Fig. <xref rid="Fig2" ref-type="fig">2d</xref>). CD18 complexes with CD11 molecules to form the adhesion molecules lymphocyte function-associated antigen-1 (LFA-1), which together with CD62L are involved in endothelial tethering and firm adhesion. In sharp contrast to the migratory phenotype of blood and tonsil ILCs, we also identified gene modules indicative of tissue residency of lung and colon ILCs, including tissue nILCs and to some extent also of tonsil ILCs. Module 54 revealed a set of genes highly correlated to the expression of the prototypical tissue-residency-associated marker <italic toggle="yes">CD69</italic> (Fig. <xref rid="Fig2" ref-type="fig">2d</xref>). We confirmed the diametrically opposed migratory vs tissue-residency profiles of blood as compared to colon ILCs by demonstrating the absence of CD69 and high expression of CD18 and CD62L on blood ILCs and low expression of CD18 and CD62L and high expression of CD69 on colon ILCs (Fig. <xref rid="Fig2" ref-type="fig">2e</xref>). Tonsil showed a mixed phenotype, regardless of mechanic or enzymatic isolation of the cells (Fig. <xref rid="Fig2" ref-type="fig">2f</xref>), with low expression of CD62L, intermediate expression of CD18 and some cells expressing CD69, aligning with the presence of both tissue resident and migratory ILCs in secondary lymphoid tissues in mice.<sup><xref ref-type="bibr" rid="CR20">20</xref></sup> CD69<sup>+</sup> tonsil ILCs expressed lower levels of CD18 as compared to CD69<sup>–</sup> tonsil ILCs (Fig. <xref rid="Fig2" ref-type="fig">2g</xref>). Similarly, expression of NKp44 was associated with reduced expression of CD18 (Fig. <xref rid="Fig2" ref-type="fig">2g</xref>). These data not only validate the transcriptional findings but also show that in contrast to colon, tonsil CD69<sup>+</sup> express the endothelial adhesion integrin CD18, as described for intravascular liver-resident T cells.<sup><xref ref-type="bibr" rid="CR21">21</xref></sup></p><p id="Par8">Showing a similar expression pattern, we identified module 7 as containing genes previously reported in human tissue resident T cells<sup><xref ref-type="bibr" rid="CR22">22</xref></sup> involved in Fos/Jun signaling (<italic toggle="yes">FOS</italic>, <italic toggle="yes">FOSB</italic>, <italic toggle="yes">FOSL2</italic> and <italic toggle="yes">JUNB</italic>), NFκB signaling (<italic toggle="yes">NFKBIZ</italic>, <italic toggle="yes">NFKBID</italic>, <italic toggle="yes">NFKBIA</italic>, <italic toggle="yes">NFKB1</italic> and <italic toggle="yes">NFKB2</italic>), adhesion (<italic toggle="yes">ICAM1</italic>) and <italic toggle="yes">CXCR4</italic> (Fig. <xref rid="Fig2" ref-type="fig">2d</xref>). In this module were also transcripts of effector molecules such as <italic toggle="yes">CXCL2</italic>, <italic toggle="yes">CXCL1</italic> and <italic toggle="yes">AREG</italic>, the latter previously reported as an ILC2 effector molecule.<sup><xref ref-type="bibr" rid="CR23">23</xref></sup> The expression of <italic toggle="yes">MAFF</italic>, <italic toggle="yes">KLF4</italic>, <italic toggle="yes">NFIL3</italic> and <italic toggle="yes">NR4A1</italic> in tissue ILCs suggests that these TFs are involved in transcriptional programs of activation and/or tissue-residency in colon, lung and tonsil ILCs. Module 6, which was particularly expressed in mature but not nILCs in colon and lung, contained transcripts of additional nuclear receptor subfamily four group members, <italic toggle="yes">NR4A2</italic> and <italic toggle="yes">NR4A3</italic>, as well as the hyaluronic acid receptor <italic toggle="yes">CD44</italic>, previously shown to be expressed by tissue-resident T and NK cells<sup><xref ref-type="bibr" rid="CR13">13</xref>,<xref ref-type="bibr" rid="CR22">22</xref></sup> (Fig. <xref rid="Fig2" ref-type="fig">2d</xref>). This module also contained transcripts involved in lipid metabolism (<italic toggle="yes">ELOVL5</italic> and <italic toggle="yes">PTGS2</italic>). The expression of <italic toggle="yes">STAT3</italic>, <italic toggle="yes">NOTCH2</italic>, <italic toggle="yes">DLL1</italic>, <italic toggle="yes">RORA</italic> and <italic toggle="yes">RARA</italic> suggests a potential role for STAT3-, Notch-, retinoic acid- and cholesterol-derivative (e.g., oxysterol) signaling in controlling activation and/or tissue-residency of colon and lung ILCs (Fig. <xref rid="Fig2" ref-type="fig">2d</xref>). Colon and lung ILC3, but to a lesser extent tonsil ILC3 expressed transcripts encoding the receptors for IL-4 (<italic toggle="yes">IL4R</italic>), IFN-γ (<italic toggle="yes">IFNGR2</italic>) and CXCL16 (<italic toggle="yes">CXCR6</italic>), raising the possibility that these factors are involved in the tissue-specific regulation of colon and lung ILC3. Furthermore, we identified module 34, containing lung-restricted transcripts indicative of tissue-residency<sup><xref ref-type="bibr" rid="CR22">22</xref></sup> including <italic toggle="yes">LMNA</italic> and the TFs <italic toggle="yes">KLF10</italic> and <italic toggle="yes">BCL3</italic> (Fig. <xref rid="Fig2" ref-type="fig">2d</xref>), the latter involved in NFκB signaling.</p><p id="Par9">In summary, we identified three gene modules allowing for the characterization of the lymphoid- and peripheral tissue homing properties of tonsil and blood ILCs, including blood nILCs. Four additional modules contained transcripts for known and potential activation and/or tissue-residency associated proteins that showed tissue-restricted expression across colon, lung and tonsils. Most of these genes were highly expressed by colon and lung ILCs, including tissue nILCs, but to some extent also by tonsil ILCs. Importantly, these genes were not expressed by blood ILCs, including blood nILCs. Furthermore, lung NK<sup>dim</sup> cells, which have been reported as circulating<sup><xref ref-type="bibr" rid="CR13">13</xref></sup> also showed low expression of transcripts known to be associated with tissue-residency.</p></sec><sec id="Sec5"><title>Gene modules reveal subset- and tissue-specific profiles of ILC3</title><p id="Par10">Module 6 specified shared characteristics of putative tissue-resident ILCs in lung and colon (Fig. <xref rid="Fig2" ref-type="fig">2d</xref>). Notably, some of the genes in this module were particularly confined to ILC3 in these two tissues (Fig. <xref rid="Fig3" ref-type="fig">3a, b</xref>). These included transcripts encoding HLA-DR, -DP and -DQ (Fig. <xref rid="Fig3" ref-type="fig">3b</xref>), which we previously described in the tonsil.<sup><xref ref-type="bibr" rid="CR12">12</xref></sup> Additional transcripts in this module included <italic toggle="yes">IRF4</italic> and <italic toggle="yes">BCL2A1</italic>, indicating a potential role for these transcripts in establishing tissue-unique features of human colon and lung ILC3. Also the transcript for the effector cytokine GM-CSF, <italic toggle="yes">CSF2</italic> and <italic toggle="yes">BHLHE40</italic>, the latter a metabolism-related gene reported to regulate GM-CSF production in mouse ILC3,<sup><xref ref-type="bibr" rid="CR24">24</xref></sup> as well as <italic toggle="yes">CXCL8</italic> and <italic toggle="yes">VEGFA</italic> were expressed in lung and colon ILC3 (Fig. <xref rid="Fig3" ref-type="fig">3b</xref>), further indicating colon and lung-associated transcriptional imprinting of ILC3 as compared to tonsil ILC3. However, colon and tonsil ILC3 also shared striking similarities with respect to high expression of the prototypical ILC3 transcripts <italic toggle="yes">NCR2</italic>, <italic toggle="yes">IL23R</italic>, <italic toggle="yes">NRP1</italic>, <italic toggle="yes">IKZF2</italic>, <italic toggle="yes">TNFSF13B</italic>, <italic toggle="yes">CCL20</italic>, <italic toggle="yes">JAML</italic>, <italic toggle="yes">PECAM1</italic>, <italic toggle="yes">TOX</italic>, <italic toggle="yes">RBPJ</italic> and <italic toggle="yes">CD300LF</italic> as well as the previously described Th17 cell transcript <italic toggle="yes">IL4I1</italic> (Fig. <xref rid="Fig3" ref-type="fig">3b</xref>). Of interest, colon ILC3 expressed both <italic toggle="yes">LTA4H</italic>, encoding the leukotriene B4 (LTB<sub>4</sub>) forming enzyme LTA4-hydrolase, and the transcript coding for the receptor for LTB<sub>4</sub>, <italic toggle="yes">LTB4R</italic>, indicating a role for LTB<sub>4</sub> in colon-specific regulation of ILC3 function. Further strengthening this hypothesis, a subset of colonic ILC3 expressed the LTB4R surface protein (Fig. <xref rid="Fig3" ref-type="fig">3c</xref>). A comparison of DE gene expression between lung, colon and tonsil ILC3 is provided as part of <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Data <xref rid="MOESM1" ref-type="media">S1</xref>.<fig id="Fig3" position="float" orientation="portrait"><label>Fig. 3</label><caption><title>Tissue- and subset-specific profile of ILC3.</title><p><bold>a</bold> UMAP visualization of annotated ILC3 and nILC clusters (C_, colon; L_, lung; B_, blood; T_, tonsil; nILC, naïve ILC). Clusters are color coded based on annotated clusters (Fig. <xref rid="Fig1" ref-type="fig">1e</xref>). <bold>b</bold> Modules highly correlated with tissues and ILC3. The figure shows a selection of genes per module, and their levels of expression within annotated clusters (Fig. <xref rid="Fig1" ref-type="fig">1e</xref>). <bold>c</bold> Expression of LTB4R on colon ILC3 (top plot) and summary bar plot showing expression of LTB4R on colon ILC3, tonsil ILC3 and blood nILCs (bottom plot). Data in <bold>c</bold> (top) is from one representative blood, tonsil and gut sample. Data are from six independent experiments with one tissue donor each (blood = 6, tonsil = 8 and colon = 6). <bold>d</bold> UMAP visualization of unbiased graph-based clustering analysis of ILC3. Clusters are color coded based on unbiased graph-based clustering analysis of all cells (Fig. <xref rid="Fig1" ref-type="fig">1d</xref>). <bold>e</bold> Trajectory analysis of nILCs and colon ILC3 displaying a selection of DE genes for each colon ILC3 cluster (C_ILC3a–d). <bold>f</bold> Trajectory analysis of nILCs and lung ILC1, ILC2 and ILC3 displaying a selection of DE genes for each lung ILC3 cluster (L_ILC3a–b). <bold>g</bold> UMAP visualization of <italic toggle="yes">NCR2</italic> expression. Colored cells (shades of blue) indicate expression of <italic toggle="yes">NCR2</italic> transcripts. <bold>h</bold> UMAP visualization of <italic toggle="yes">IL22</italic> expression. Colored cells (shades of blue) indicate expression of <italic toggle="yes">IL22</italic> transcripts. Data in <bold>a</bold>, <bold>b</bold>, <bold>d</bold>–<bold>h</bold> is from ten independent experiments with one tissue donor each (blood = 3, lung = 4 and colon = 3) integrated with data from.<sup><xref ref-type="bibr" rid="CR12">12</xref></sup></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" id="d32e1290" position="float" orientation="portrait" xlink:href="41422_2020_445_Fig3_HTML.jpg"><?image-name 41422_2020_445_Fig3_HTML.jpg?><?image-size 198848?><?image-md5 b04b5834f554a43a07956cd83e644970?><?image-image-server-status LOAD_COMPLETED?><?image-original-height 2475?><?image-original-width 1460?><?image-scaled-height 1238?><?image-scaled-width 730?><?image-cloudpmc-urn urn:cdn:blobs/f93f/8089104/b04b5834f554/41422_2020_445_Fig3_HTML.jpg?><?thumb-name 41422_2020_445_Fig3_HTML.gif?><?thumb-size 6744?><?thumb-md5 abec9943d52986071837b839f60d13ef?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 170?><?thumb-scaled-width 100?><?thumb-cloudpmc-urn urn:cdn:blobs/f93f/8089104/abec9943d529/41422_2020_445_Fig3_HTML.gif?></graphic></fig></p><p id="Par11">Module 3 contained ILC3-related genes that were expressed by ILC3 in all tissues but at particularly high level and frequency in colon ILC3 (Fig. <xref rid="Fig3" ref-type="fig">3b</xref>). These included surface molecule transcripts <italic toggle="yes">KIT</italic>, <italic toggle="yes">IL1R1</italic>, <italic toggle="yes">ICOS</italic>, <italic toggle="yes">CCR6</italic>, <italic toggle="yes">TNFSF11</italic> (encoding RANK ligand), <italic toggle="yes">TNFSF4</italic> (encoding OX40 ligand), the ITIM-containing <italic toggle="yes">CD96</italic> and the tetraspanin <italic toggle="yes">CD81</italic>. We also identified transcripts for the TFs <italic toggle="yes">RORC</italic> and <italic toggle="yes">AHR</italic>, the signaling molecule <italic toggle="yes">PTPN22</italic> and the cytokine <italic toggle="yes">LIF</italic> as preferentially highly expressed by colon ILC3 (Fig. <xref rid="Fig3" ref-type="fig">3b</xref>).</p><p id="Par12">Finally, we evaluated the transcriptional profile of the clusters of cells that were termed nILCs in blood and tissues (colon and lung) on the basis of lack of markers associated with mature ILCs (Fig. <xref rid="Fig1" ref-type="fig">1f</xref>). From the evaluation of ILC3 (Fig. <xref rid="Fig3" ref-type="fig">3a, b</xref>) it became obvious that these cells show few transcriptional parallels to ILC3 despite their phenotypical similarities as assessed by flow cytometry. Expression of genes associated with migration and tissue-residency revealed a clear migratory behavior of blood nILCs whereas tissue nILCs showed features of tissue residency (Fig. <xref rid="Fig2" ref-type="fig">2d</xref>). A list of DE genes unique for blood and tissue nILCs is provided as part of <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Data <xref rid="MOESM1" ref-type="media">S1</xref>. This list includes TFs of potential importance for the differentiation of nILCs in blood (e.g., <italic toggle="yes">TCF7</italic>) and lung/colon (e.g., <italic toggle="yes">XBP1</italic> and <italic toggle="yes">BATF</italic>) (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Data <xref rid="MOESM1" ref-type="media">S1</xref>).</p><p id="Par13">In summary, we unveil the transcriptional features of nILCs in blood and tissues and reveal tissue-specific differences among ILC3 and potential novel functions of these cells by description of transcripts previously not known to be expressed by human ILC3.</p></sec><sec id="Sec6"><title>A spectrum of mucosal ILC3 transcriptional states branching in trajectory analysis</title><p id="Par14">Next we performed trajectory analysis via Slingshot<sup><xref ref-type="bibr" rid="CR25">25</xref></sup> to understand the potential relationships between the putative nILCs identified in blood and tissues, and the more mature ILC subsets and the potential spectrum of activation states of ILC3 in colon and lung. Overlaying these trajectories with the clusters identified in Fig. <xref rid="Fig1" ref-type="fig">1d</xref> (Fig. <xref rid="Fig3" ref-type="fig">3d</xref>) not only illustrated the heterogeneity among ILC3 in colon, lung and tonsil but also inferred the potential transcriptional relationships between these transcriptionally distinct ILC3 subclusters (Fig. <xref rid="Fig3" ref-type="fig">3e, f</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S4a, b</xref>). ILC3s were particularly heterogeneous in the colon where four subclusters could be identified through Louvain clustering (colon ILC3a–d, lung ILC3a–b, tonsil ILC3a–b) (Figs. <xref rid="Fig1" ref-type="fig">1</xref>d, <xref rid="Fig3" ref-type="fig">3e, f</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S4a</xref>). Some of these clusters were dominated by, but not unique to, individual donors as cells from all donors of each tissue were present in all clusters (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S3b, c</xref>) for each tissue. Of importance, the TF <italic toggle="yes">IKZF2</italic> and the master ILC3 TFs <italic toggle="yes">RORC</italic> and <italic toggle="yes">AHR</italic> were expressed across all ILC3 subclusters in all tissues, assigning them to the ILC3 lineage (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S4c</xref>). Analysis comparing clusters within each tissue and focusing on the top 20 DE genes (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Data <xref rid="MOESM1" ref-type="media">S2</xref>), revealed that colon ILC3a (cluster 1) contained cells with an activated and potentially tissue-resident phenotype, expressing transcripts for multiple effector molecules involved in chemotaxis and recruitment of immune cells including <italic toggle="yes">CSF2</italic>, <italic toggle="yes">XCL1</italic>, <italic toggle="yes">XCL2</italic>, <italic toggle="yes">CXCL8</italic>, <italic toggle="yes">S100A4</italic> and <italic toggle="yes">S100A6</italic> (Fig. <xref rid="Fig3" ref-type="fig">3e</xref>). <italic toggle="yes">EGR1</italic>, <italic toggle="yes">EGR2 and KLF6</italic>, encoding TFs involved in immune cell activation,<sup><xref ref-type="bibr" rid="CR26">26</xref>,<xref ref-type="bibr" rid="CR27">27</xref></sup> and <italic toggle="yes">BCL2A1</italic>, coding for an anti-apoptotic protein, were also enriched in this cluster of cells. The prototypical ILC3 cytokine <italic toggle="yes">IL22</italic> and the activation marker <italic toggle="yes">NCR2</italic> were expressed in colon ILC3a, as well as in colon ILC3b (cluster 0) (Fig. <xref rid="Fig3" ref-type="fig">3g, h</xref>); the latter also showed an activated phenotype with expression of genes involved in glutamate response and lipid metabolism, respectively (<italic toggle="yes">GRM7</italic> and <italic toggle="yes">LTA4H</italic>). Colon ILC3a was most similar to lung ILC3a (cluster 2) in terms of <italic toggle="yes">XCL1</italic>, <italic toggle="yes">XCL2, S100A4, EGR2 and IFITM2</italic>, but also shared similarities with lung ILC3b (cluster 10) in terms of <italic toggle="yes">CXCL8, CD83</italic> and <italic toggle="yes">PTGS2</italic> (encoding COX2) (Fig. <xref rid="Fig3" ref-type="fig">3e, f</xref>). ILC3s in the lung also branched into two subclusters, indicating distinct transcriptional trajectories which both contained features of activation and tissue residency (Fig. <xref rid="Fig3" ref-type="fig">3f</xref>). The activated phenotype of colon ILC3a stood in sharp contrast to colon ILC3d (cluster 14), which was most similar to lung ILC3b (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Data <xref rid="MOESM1" ref-type="media">S2</xref>). Colon ILC3d additionally expressed <italic toggle="yes">SELL</italic> (encoding CD62L), also expressed by tonsil ILC3b (cluster 4) (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S4a</xref>),<sup><xref ref-type="bibr" rid="CR12">12</xref></sup>
+<italic toggle="yes">S1PR1</italic>, <italic toggle="yes">ITGAX</italic> and <italic toggle="yes">GPR183</italic> (Fig. <xref rid="Fig3" ref-type="fig">3e</xref>), indicative of a migratory behavior reminiscent of LTi-like cells contributing to colonic lymphoid structure formation in a GPR183-dependent manner in mice.<sup><xref ref-type="bibr" rid="CR28">28</xref></sup> Importantly, colon ILC3d formed a separate branch in trajectory analysis and were transcriptionally distinct from the tissue nILCs (Fig. <xref rid="Fig3" ref-type="fig">3e</xref>).</p><p id="Par15">In summary, ILC3s show a high degree of transcriptional heterogeneity, in particular in the colon where we identified a spectrum of ILC3 states with variable activation status which separated into different branches in trajectory analysis, ranging from cells with a migratory phenotype, to ILC3-expressing effector molecules important for metabolic adaptations, tissue repair and immune cell recruitment.</p></sec><sec id="Sec7"><title>Gene modules reveal subset- and tissue-specific profiles of ILC2</title><p id="Par16">Our gene module approach identified a set of genes that were specifically expressed in ILC2 across the three tissues (Fig. <xref rid="Fig4" ref-type="fig">4a, b</xref>), making up an ILC2 core transcriptome. These genes included, as expected, <italic toggle="yes">GATA3</italic> and <italic toggle="yes">MAF</italic>, as well as <italic toggle="yes">HPGD</italic>, <italic toggle="yes">HPGDS</italic> and <italic toggle="yes">MBOAT2</italic>, the latter three encoding proteins involved in lipid metabolism. Furthermore, ILC2 in all tissues expressed transcripts for the IL-9 receptor (<italic toggle="yes">IL9R</italic>). Other genes in the module were particularly expressed in blood ILC2 or blood and tonsil ILC2, including transcripts for receptors involved in cell trafficking and migration such as the prototypical ILC2 marker CRTH2 (<italic toggle="yes">PTGDR2</italic>), <italic toggle="yes">S1PR2</italic> and <italic toggle="yes">CCR2</italic>, the latter shown to be involved in the regulation of S1P receptors.<sup><xref ref-type="bibr" rid="CR29">29</xref></sup> The TNF family-related transcripts <italic toggle="yes">TNFSF10</italic> (encoding TRAIL) and <italic toggle="yes">TNFRSF19</italic> (encoding TROY) were also highly expressed by blood and tonsil ILC2, as well as <italic toggle="yes">CD200R1</italic>, possibly acting as an inhibitory receptor for human blood and tonsil ILC2. Another ILC2-related module was less restricted to ILC2 but still contained transcripts relevant for ILC2 biology, particularly for the potential negative regulation of these cells; <italic toggle="yes">IL10RA</italic>,<sup><xref ref-type="bibr" rid="CR30">30</xref></sup>
+<italic toggle="yes">KLRG1</italic>,<sup><xref ref-type="bibr" rid="CR31">31</xref></sup>
+<italic toggle="yes">PTGER2</italic>,<sup><xref ref-type="bibr" rid="CR32">32</xref></sup>
+<italic toggle="yes">ANXA1</italic><sup><xref ref-type="bibr" rid="CR33">33</xref></sup> as well as <italic toggle="yes">LGALS1</italic>, encodes the immunosuppressive molecule galectin-1.<fig id="Fig4" position="float" orientation="portrait"><label>Fig. 4</label><caption><title>Tissue- and subset-specific profile of ILC2.</title><p><bold>a</bold> UMAP visualization of annotated ILC2 clusters (L_, lung; B_, blood; T_, tonsil). Clusters are color coded based on annotated clusters (Fig. <xref rid="Fig1" ref-type="fig">1e</xref>). <bold>b</bold> Modules highly correlated with ILC2 or tissue. The figure shows a selection of genes per module, and their level of expression within annotated clusters (Fig. <xref rid="Fig1" ref-type="fig">1e</xref>). Data are from ten independent experiments with one tissue donor each (blood = 3, lung = 4 and colon = 3) integrated with data from.<sup><xref ref-type="bibr" rid="CR12">12</xref></sup></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" id="d32e1659" position="float" orientation="portrait" xlink:href="41422_2020_445_Fig4_HTML.jpg"><?image-name 41422_2020_445_Fig4_HTML.jpg?><?image-size 81575?><?image-md5 b6277c61ae387fad6863047fdae65aa7?><?image-image-server-status LOAD_COMPLETED?><?image-original-height 1141?><?image-original-width 1525?><?image-scaled-height 570?><?image-scaled-width 762?><?image-cloudpmc-urn urn:cdn:blobs/f93f/8089104/b6277c61ae38/41422_2020_445_Fig4_HTML.jpg?><?thumb-name 41422_2020_445_Fig4_HTML.gif?><?thumb-size 3087?><?thumb-md5 2e4be2314a7dd3627092eedd3fe93d88?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 79?><?thumb-scaled-width 106?><?thumb-cloudpmc-urn urn:cdn:blobs/f93f/8089104/2e4be2314a7d/41422_2020_445_Fig4_HTML.gif?></graphic></fig></p><p id="Par17">Finally, we identified a module of genes (number 65) specifically expressed in lung, but not tonsil or blood ILC2 (Fig. <xref rid="Fig4" ref-type="fig">4b</xref>). Expression of the main effector molecule <italic toggle="yes">IL13</italic> indicated that lung ILC2 are highly activated. Indeed, lung ILC2 showed high expression of <italic toggle="yes">IL1RL1</italic> and <italic toggle="yes">IL17RB</italic>, encoding the receptors for the ILC2-activating alarmins IL-33 and IL-25, respectively. Additional transcripts for potential activating receptors included the SLAM-family receptor <italic toggle="yes">SLAMF1, TNFRSF9</italic>, the fatty acid receptors <italic toggle="yes">FFAR3</italic> and <italic toggle="yes">PPARG</italic>, the latter shown to activate ILC2 in mice.<sup><xref ref-type="bibr" rid="CR34">34</xref></sup></p><p id="Par18">The fact that blood ILC2 showed a phenotype associated with re-circulation and migration on both the transcriptional (Figs. <xref rid="Fig2" ref-type="fig">2</xref>d, <xref rid="Fig4" ref-type="fig">4b</xref>) and protein (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S4d, e</xref>) level support the possibility that blood ILC2 might be recruited to the lung, as previously suggested in the mouse<sup><xref ref-type="bibr" rid="CR35">35</xref>,<xref ref-type="bibr" rid="CR36">36</xref></sup> and in a human lung allergen-provocation study.<sup><xref ref-type="bibr" rid="CR37">37</xref></sup> We therefore hypothesized that exposure of blood ILC2 to an inflammatory type 2-deviating environment could induce the lung ILC2 phenotype. To address this possibility, we sorted and exposed blood ILC2 to the prototypical ILC2-activating cocktail of alarmins (IL-2, IL-25, IL-33 and TSLP) and assessed phenotypical changes (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S5</xref>). Indeed, alarmin exposure increased the expression of IL-1RL1 and IL-17RB surface proteins on blood-derived ILC2 (Fig. <xref rid="Fig5" ref-type="fig">5a</xref>), phenocopying lung ILC2. In parallel, the expression of CRTH2 was downregulated, which we previously showed can be promoted by alarmin-mediated endogenous production of the CRTH2 ligand PGD<sub>2</sub>.<sup><xref ref-type="bibr" rid="CR38">38</xref></sup> This finding raised the possibility that in the lung, alarmin-activated ILC2 might not express CRTH2. Indeed, utilizing our indexed flow cytometry data, evaluation of the composition of the cluster annotated as lung ILC2 (Fig. <xref rid="Fig1" ref-type="fig">1d</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S3a</xref>) revealed a relatively high presence (35%) of CRTH2<sup>−</sup> ILCs (ILC1 and ILC3 by FACS phenotype) among the lung ILC2 which was not seen for blood or tonsil ILC2 where the majority of cells expressed CRTH2 protein (Fig. <xref rid="Fig1" ref-type="fig">1d</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S3a</xref>). Evaluation of the transcriptional profile of CRTH2<sup>–</sup> vs CRTH2<sup>+</sup> cells showed that ILC2-related genes were present in CRTH2<sup>– </sup>ILC2, and at higher levels and frequencies as compared to lung ILC1 and ILC3 (Fig. <xref rid="Fig5" ref-type="fig">5b, c</xref>).<fig id="Fig5" position="float" orientation="portrait"><label>Fig. 5</label><caption><title>Exposure of blood ILC2 to alarmins recapitulates the lung ILC2 phenotype.</title><p><bold>a</bold> Expression of IL-1RL1, IL-17RB and CRTH2 protein on the surface of blood ILC2 treated with an alarmin cocktail (IL-2, IL-25, IL-33 and TSLP) for 48 h. FMO, fluorescence minus one. <bold>b</bold> UMAP visualization of annotated ILC clusters in the lung. Clusters are color coded based on annotated clusters (Fig. <xref rid="Fig1" ref-type="fig">1e</xref>). <bold>c</bold> Violin plot visualization of differentially expressed genes in cells annotated as ILC2 with a FACS phenotype of ILC1 (L_ILC2_ILC1) or ILC3 (L_ILC2_ILC3) vs cells ILC1 or ILC3, respectively. Data are from seven independent experiments with a total of 13 blood donors (<bold>a</bold>) or from ten independent experiments with one tissue donor each (blood = 3, lung = 4 and colon = 3) integrated with data from<sup><xref ref-type="bibr" rid="CR12">12</xref></sup> (<bold>b</bold>, <bold>c</bold>). Bars in <bold>b</bold> represent means ± SD. Mann Whitney test was used in <bold>a</bold>; **<italic toggle="yes">P</italic> &lt; 0.01; ***<italic toggle="yes">P</italic> &lt; 0.001.</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" id="d32e1811" position="float" orientation="portrait" xlink:href="41422_2020_445_Fig5_HTML.jpg"><?image-name 41422_2020_445_Fig5_HTML.jpg?><?image-size 127230?><?image-md5 4f884b0e369e274f0cd6db56bd6bd0d1?><?image-image-server-status LOAD_COMPLETED?><?image-original-height 2342?><?image-original-width 1775?><?image-scaled-height 937?><?image-scaled-width 710?><?image-cloudpmc-urn urn:cdn:blobs/f93f/8089104/4f884b0e369e/41422_2020_445_Fig5_HTML.jpg?><?thumb-name 41422_2020_445_Fig5_HTML.gif?><?thumb-size 5000?><?thumb-md5 4bcf67ce95fef1048a3a1da08019e9b9?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 132?><?thumb-scaled-width 100?><?thumb-cloudpmc-urn urn:cdn:blobs/f93f/8089104/4bcf67ce95fe/41422_2020_445_Fig5_HTML.gif?></graphic></fig></p><p id="Par19">In summary, we reveal tissue-specific transcriptional differences between ILC2 in blood, tonsil and lung particularly related to migration, tissue-residency and sensing of the tissue microenvironment. Exposure of blood ILC2 to an inflammatory type-2 environment lead to development of a lung-like phenotype including downregulation of CRTH2. Thus, our transcriptional data highlight heterogeneity among lung ILC2, specifically identifying CRTH2<sup>−</sup> ILC2 with transcriptional similarities to bona fide CRTH2<sup>+</sup> ILC2, including expression of the ILC2 effector molecule <italic toggle="yes">IL13</italic>.</p></sec><sec id="Sec8"><title>Gene modules reveal subset- and tissue-specific profiles of ILC1-like cells</title><p id="Par20">Module 32 contained transcripts which were uniquely expressed by putative ILC1 and <italic toggle="yes">EOMES</italic><sup>+</sup> ILC1 in blood and tonsil (Fig. <xref rid="Fig6" ref-type="fig">6a, b</xref>). In line with previous reports,<sup><xref ref-type="bibr" rid="CR12">12</xref>,<xref ref-type="bibr" rid="CR39">39</xref></sup> the module consisted of T cell-related genes including <italic toggle="yes">CD3D, CD3E, CD3G, CD4, CD5</italic>, <italic toggle="yes">CD6 and CD27</italic>. <italic toggle="yes">LEF1</italic>, a TF transcript described in tonsil ILC1,<sup><xref ref-type="bibr" rid="CR12">12</xref></sup> was also expressed by blood ILC1. Notably, the transcripts in this module were expressed by few lung ILC1 suggesting that blood and tonsil ILC1 have more pronounced T cell features than lung ILC1 (Fig. <xref rid="Fig6" ref-type="fig">6b</xref>). Furthermore, indicating that blood and tonsil ILC1, and in particular <italic toggle="yes">EOMES</italic><sup>+</sup> ILC1, are more related to NK cells than lung ILC1, two additional transcriptional modules (Fig. <xref rid="Fig6" ref-type="fig">6b</xref>) revealed transcripts expressed in NK cells as well as in ILC1 and <italic toggle="yes">EOMES</italic><sup>+</sup> ILC1 in blood and tonsil, and to a lesser extent by ILC1 in lung (Fig. <xref rid="Fig6" ref-type="fig">6b</xref>).<fig id="Fig6" position="float" orientation="portrait"><label>Fig. 6</label><caption><title>Tissue and subset specific profile of ILC1.</title><p><bold>a</bold> UMAP visualization of annotated ILC1 and NK cell clusters. Clusters are color coded based on annotated clusters (Fig. <xref rid="Fig1" ref-type="fig">1e</xref>). <bold>b</bold> Modules highly correlated with ILC1 and NK cells. The figure shows a selection of genes per module, and their level of expression within annotated clusters (Fig. <xref rid="Fig1" ref-type="fig">1e</xref>).</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" id="d32e1898" position="float" orientation="portrait" xlink:href="41422_2020_445_Fig6_HTML.jpg"><?image-name 41422_2020_445_Fig6_HTML.jpg?><?image-size 87225?><?image-md5 4953240719292be35e0c05d7aad51401?><?image-image-server-status LOAD_COMPLETED?><?image-original-height 1175?><?image-original-width 1525?><?image-scaled-height 587?><?image-scaled-width 762?><?image-cloudpmc-urn urn:cdn:blobs/f93f/8089104/495324071929/41422_2020_445_Fig6_HTML.jpg?><?thumb-name 41422_2020_445_Fig6_HTML.gif?><?thumb-size 3339?><?thumb-md5 b49101b5a4c94a0c30f546e5cc34cec6?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 79?><?thumb-scaled-width 103?><?thumb-cloudpmc-urn urn:cdn:blobs/f93f/8089104/b49101b5a4c9/41422_2020_445_Fig6_HTML.gif?></graphic></fig></p></sec><sec id="Sec9"><title>Widespread predicted TCR V(D)J-rearrangement in blood ILC1-like cells</title><p id="Par21">Since putative ILC1 and <italic toggle="yes">EOMES</italic><sup>+</sup> ILC1 in blood and tonsil showed characteristics of T cells, we performed a direct comparison between blood ILCs and CD4<sup>+</sup> T cells sorted in parallel from the same peripheral blood donors (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Figs. <xref rid="MOESM1" ref-type="media">S1</xref> and <xref rid="MOESM1" ref-type="media">S6</xref>). To understand the T cell identity of ILC1, we analyzed TCR V(D)J-rearrangement using MiXCR.<sup><xref ref-type="bibr" rid="CR40">40</xref></sup> In contrast to the T cell population, which as expected predominantly consisted of cells with rearranged TCR-α and -β chains, ILC2 and nILCs showed minimal V(D)J rearrangement across any of the TCR chains (Fig. <xref rid="Fig7" ref-type="fig">7a</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S7a</xref>). This stood in sharp contrast to putative blood ILC1 and <italic toggle="yes">EOMES</italic><sup>+</sup> ILC1, where most of the cells had undergone TCR rearrangement (Fig. <xref rid="Fig7" ref-type="fig">7a</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM8" ref-type="media">S7a</xref>), despite lacking surface expression of TCR α/β, γ/δ or CD3 (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S1a</xref>). While <italic toggle="yes">EOMES</italic><sup>+</sup> ILC1 mostly had predicted rearrangement in α- and/or β-TCR chains, blood ILC1 showed predicted rearrangement in all four TCR chains (Fig. <xref rid="Fig7" ref-type="fig">7a</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S7a</xref>). UMAP clustering revealed that blood ILC1 with α- and/or β-TCR rearrangement chains clustered with CD4<sup>+</sup> T cells (AB_T_ILC1) (Fig. <xref rid="Fig7" ref-type="fig">7b–d</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S7b, c</xref>). DE analysis showed that while the majority of ILC1 expressed transcripts of α- and/or β-TCR constant regions (<italic toggle="yes">TRA-C</italic> and <italic toggle="yes">TRB-C</italic>), expression of <italic toggle="yes">CD4</italic> as well as variable fragments of α- and/or β-TCR genes (<italic toggle="yes">TRA-V</italic> and <italic toggle="yes">TRB-V</italic>) was more pronounced in ILC1 with α- and/or β-TCR rearrangement (Fig. <xref rid="Fig7" ref-type="fig">7e</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S7d, e</xref>). ILC1 with rearrangement in predominantly δ-, but also γ-TCR chains formed a distinct cluster (GD_ILC1) (Fig. <xref rid="Fig7" ref-type="fig">7c, d</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S7b</xref>) and concordantly expressed fragments of γ- and/or δ-TCR variable chains (<italic toggle="yes">TRG-V</italic> and <italic toggle="yes">TRD-V</italic>) as well as <italic toggle="yes">IKZF2</italic> (Fig. <xref rid="Fig7" ref-type="fig">7e</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S7d, e</xref>). In contrast, <italic toggle="yes">EOMES</italic><sup>+</sup> ILC1, of which most had predicted rearrangements in α/β chains, clustered with T cells in close proximity to CXCR3<sup>+</sup> Th1 cells<sup><xref ref-type="bibr" rid="CR41">41</xref></sup> (Fig. <xref rid="Fig7" ref-type="fig">7b–d</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S7c</xref>). DE analysis unveiled that <italic toggle="yes">EOMES</italic><sup>+</sup> ILC1 with predicted rearrangement in α/β chains showed high expression of α- and/or β-TCR constant and variable regions, as well as <italic toggle="yes">CD8A</italic> (Fig. <xref rid="Fig7" ref-type="fig">7e</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S7d, e</xref>). Hence, <italic toggle="yes">EOMES</italic><sup>+</sup> ILC1 are transcriptionally closely related to Th1 and cytotoxic T cells. Finally, we evaluated the clonality of ILCs vs T cells. As expected from healthy donor blood, most cells represented unique clones (Fig. <xref rid="Fig7" ref-type="fig">7f</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S7f</xref>). Among ILC1 and <italic toggle="yes">EOMES</italic><sup>+</sup> ILC1 with α/β rearrangement, there were six clones represented by two or more cells (Fig. <xref rid="Fig7" ref-type="fig">7f</xref>; <xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S7f</xref>). Two of these clones were shared with T cells (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S7g</xref>).<fig id="Fig7" position="float" orientation="portrait"><label>Fig. 7</label><caption><title>Rearrangement of V(D)J T cell receptor genes in blood ILCs and T cells.</title><p><bold>a</bold> Predicted V(D)J-rearrangement of <italic toggle="yes">TRA</italic>, <italic toggle="yes">TRB</italic>, <italic toggle="yes">TRD</italic> and/or <italic toggle="yes">TRG</italic> genes in annotated blood ILC clusters (B_ILC1, EOMES<sup>+ </sup>ILC1, B_ILC2 and B_nILC) and in donor matched CD4<sup>+</sup> T cells (T). TRA_B, rearrangement of <italic toggle="yes">TRA</italic> and <italic toggle="yes">TRB</italic> genes; TRG_D, rearrangement of <italic toggle="yes">TRG</italic> and <italic toggle="yes">TRD</italic> genes; TRA_B + G_D, rearrangement of <italic toggle="yes">TRA</italic> and/or <italic toggle="yes">TCRB</italic> in combination with <italic toggle="yes">TRD</italic> and/or <italic toggle="yes">TRG</italic> genes. <bold>b</bold> UMAP of blood ILCs and CD4<sup>+</sup> T cells color-coded on the basis of annotation in Fig. <xref rid="Fig1" ref-type="fig">1e</xref> (B_ILC1, <italic toggle="yes">EOMES</italic><sup>+</sup> ILC1, B_ILC2 and B_nILC) and CD4<sup>+</sup> T cells (T). <bold>c</bold> UMAP displaying each subset in <bold>b</bold> separately and color-coding the cells on the basis of predicted V(D)J-rearrangement of <italic toggle="yes">TRA</italic> and/or <italic toggle="yes">TRB</italic> (TRA_B), <italic toggle="yes">TRG</italic> and/or <italic toggle="yes">TRD</italic> (TRG_D) or <italic toggle="yes">TRA</italic> and/or <italic toggle="yes">TCRB</italic> in combination with <italic toggle="yes">TRD</italic> and/or <italic toggle="yes">TRG</italic> genes (TRA_B + G_D). <bold>d</bold> Annotation of clusters on the basis of cell composition and predicted TCR V(D)J-rearrangement pattern. <bold>e</bold> Violin plot representation of selected DE genes in AB_T_ILC1 cluster (top), GD_ILC1 (middle) and <italic toggle="yes">EOMES</italic><sup>+</sup> ILC1 (bottom). Within each cluster, data is displayed separately for ILC1 and T cells, and also on the basis of TCR V(D)J-rearrangement pattern. <bold>f</bold> Number of cells with different clone frequencies (based on third complementarity region 3 (CDR3)) per rearranged TCR chain for all ILC1 (top) and CD4<sup>+</sup> T cells (bottom) in the different clusters in <bold>d</bold>. Data are from three independent scRNA-seq experiments with one blood donor each. Total number of cells is 1036 (536 CD4<sup>+</sup> T cells and 473 ILCs).</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" id="d32e2244" position="float" orientation="portrait" xlink:href="41422_2020_445_Fig7_HTML.jpg"><?image-name 41422_2020_445_Fig7_HTML.jpg?><?image-size 217368?><?image-md5 a6cd1212b6f9730de3ca9f892e02cea3?><?image-image-server-status LOAD_COMPLETED?><?image-original-height 2306?><?image-original-width 1775?><?image-scaled-height 922?><?image-scaled-width 710?><?image-cloudpmc-urn urn:cdn:blobs/f93f/8089104/a6cd1212b6f9/41422_2020_445_Fig7_HTML.jpg?><?thumb-name 41422_2020_445_Fig7_HTML.gif?><?thumb-size 7368?><?thumb-md5 b0faabc9b5066b5726bb693ddab29985?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 130?><?thumb-scaled-width 100?><?thumb-cloudpmc-urn urn:cdn:blobs/f93f/8089104/b0faabc9b506/41422_2020_445_Fig7_HTML.gif?></graphic></fig></p><p id="Par22">In summary, in contrast to ILC2 and nILCs, the majority of blood ILC1 and <italic toggle="yes">EOMES</italic><sup>+</sup> ILC1 are strikingly similar to CD4<sup>+</sup> or CD8<sup>+</sup> TCR α/β<sup>+</sup> T cells, respectively, in terms of expression profile, TCR rearrangement and clonality. In contrast, blood ILC1 with preferential δ-TCR rearrangement constitute a subset distinct from T cells. Our conclusions were further strengthened when comparing the transcriptional profiles of ILCs to blood CD3<sup>+</sup> T cells and CD56<sup>+</sup> NK cells in a publically available scRNA-seq data set (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S8</xref>), showing minimal similarity between blood ILC1 with preferential δ-TCR rearrangement and any of the T cell subsets (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S8e, f</xref>).</p></sec></sec><sec id="Sec10" sec-type="discussion"><title>Discussion</title><p id="Par23">Previous reports have described tissue-specific transcriptional imprinting of ILCs across multiple organs. However, these findings originated from pre-defined ILC subsets sorted in bulk from mice<sup><xref ref-type="bibr" rid="CR2">2</xref></sup> or humans,<sup><xref ref-type="bibr" rid="CR3">3</xref></sup> neglecting heterogeneity in the sorted populations. Here we used full-length scRNA-seq on individually sorted ILCs allowing for integrated analysis of FACS phenotype and transcriptome on the single-cell level. Additionally, we took a holistic and unbiased approach to analyzing the scRNA-seq data by complementing DE analysis with correlation analysis to identify gene modules, thereby going beyond only describing markers specific to each cell type. Through our approach, we have been able to make a number of novel observations, accompanied by a set of experimental functional studies, which advance our understanding of human ILC biology and serve as a resource to the immunology community.</p><p id="Par24">We identified two distinct ILC clusters that lacked markers of mature ILCs. In the blood, these cells expressed CD117, identifying them as the previously described circulating progenitors of all lineages of mature ILCs, including NK cells.<sup><xref ref-type="bibr" rid="CR4">4</xref>,<xref ref-type="bibr" rid="CR17">17</xref></sup> The other cluster consisted of predominantly CD117<sup>+</sup> ILC from colon and lung. In sharp contrast to circulating nILCs, these cells showed a transcriptome compatible with tissue-residency, possibly representing a resident precursor to mature ILCs in these tissues. nILCs are identified as ILC3 by flow cytometry,<sup><xref ref-type="bibr" rid="CR3">3</xref></sup> calling for a better cell surface definition of these cells. Future studies should be aimed at targeted sorting and enrichment of these cells based on e.g., expression of CD117 and lack of markers associated with mature ILC subsets and activation such as CRTH2, HLA-DR and NKp44. This might reveal nILC subsets polarized to specific fates in certain tissues. Our in-depth unbiased transcriptional profiling of circulating and resident nILCs provide a road-map for such differentiation studies of these cells and their role in disease conditions.</p><p id="Par25">Parabiosis experiments in mice have suggested that ILCs in mucosal and adipose tissues are largely tissue resident<sup><xref ref-type="bibr" rid="CR42">42</xref></sup> while other reports have unveiled a recirculating behavior of colon ILC2 during <italic toggle="yes">N. brasiliensis</italic> infection.<sup><xref ref-type="bibr" rid="CR36">36</xref></sup> We identified gene modules particularly enriched in naïve and mature ILCs in the colon and lung including <italic toggle="yes">CD69</italic> and transcripts encoding TFs (<italic toggle="yes">NR4A1-3</italic>, <italic toggle="yes">STAT3</italic>, <italic toggle="yes">RORA</italic> and <italic toggle="yes">RARA</italic>) as well as transcripts involved in Notch-, Fos/Jun- and NFκB signaling, many of which have previously been detected in CD69<sup>+</sup> T cells and NK cells in humans and mice.<sup><xref ref-type="bibr" rid="CR13">13</xref>,<xref ref-type="bibr" rid="CR22">22</xref></sup> In sharp contrast to colon and lung, naïve and mature blood ILCs showed features of CD62L- and S1PR1-mediated recirculation, similar to that described in mice.<sup><xref ref-type="bibr" rid="CR36">36</xref></sup> In the tonsil, we detected signs of both recirculation and tissue-residency. These data align well with a recent report on human ILC3 sorted in bulk from tonsils, spleen and blood.<sup><xref ref-type="bibr" rid="CR4">4</xref></sup> Additionally, in mice, lymph nodes harbor both migratory and tissue-resident ILCs, the former more common for ILC1.<sup><xref ref-type="bibr" rid="CR20">20</xref></sup> Although the tissue-residency and migratory profiles of mouse ILCs seem largely conserved in humans, we did detect some unique features of human ILCs in this regard. Most strikingly, human ILC2 uniquely expressed <italic toggle="yes">S1PR2</italic>, opening up for the possibility of selectively targeting this receptor to influence ILC2 migration during type 2 inflammation.</p><p id="Par26">The colonic mucosa is continuously exposed to environmental and intrinsic factors derived from the microbiota, diet and secretions in the digestive system. Indeed, we detected the highest expression of ILC3-related genes in this tissue, and also the highest degree of heterogeneity among ILC3. These ILC3 represented a spectrum of cells ranging from those with a more resting and recirculating phenotype, possibly constituting cells recently recruited from the circulation, to cells expressing genes encoding effector molecules, activation markers and molecules involved in metabolism. These data align well with a recent scRNA-seq study of ILCs in the mouse colon,<sup><xref ref-type="bibr" rid="CR8">8</xref></sup> although the precise transcripts defining ILC3 subsets in the mouse colon was different to those we identified here. Notably, we did not identify ILC1-biased ILC3, as those recently reported in the human inflamed colon,<sup><xref ref-type="bibr" rid="CR11">11</xref></sup> which is likely due to the non-inflamed nature of the intestinal tissue in our analysis. Furthermore, our data confirm previous reports that in stark contrast to the mouse, the human healthy colon lacks ILC2 as defined by surface protein expression.<sup><xref ref-type="bibr" rid="CR16">16</xref>,<xref ref-type="bibr" rid="CR43">43</xref></sup> Our data now further conclusively demonstrate that the human colon is also largely devoid of ILCs with a transcriptome reminiscent of ILC2. These results provide a blueprint of the homeostatic situation of ILCs in the human colon, aiding studies of the role for ILCs in colon immunopathology.</p><p id="Par27">In the lung, we identified ILC2 with a transcriptome indicative of activation, including expression of receptors of both IL-25 and IL-33, reminiscent of natural ILC2 described in mice.<sup><xref ref-type="bibr" rid="CR44">44</xref></sup> The lung-specific transcriptome of ILC2 could be a result of the constant exposure of the lung to both extrinsic and intrinsic immune stimuli. Indeed, we could recapitulate the lung phenotype of ILC2, with high expression of receptors for IL-33 and IL-25, and low expression of CRTH2, by exposing circulating ILC2 to tissue-derived alarmins. This indicates that activated ILC2 might not necessarily express CRTH2, which has previously been suggested based on both mouse<sup><xref ref-type="bibr" rid="CR35">35</xref></sup> and human ex vivo<sup><xref ref-type="bibr" rid="CR37">37</xref>,<xref ref-type="bibr" rid="CR45">45</xref></sup> and in vitro data.<sup><xref ref-type="bibr" rid="CR38">38</xref></sup> Utilizing the power of our flow cytometry indexed scRNA-seq data we were indeed able to identify such CRTH2<sup>–</sup> ILC2 in the lung, expressing receptors for IL-33 and IL-25. In addition to downregulation of CRTH2 on ILC2, CRTH2<sup>–</sup> ILC2 might also be derived from ILC1 and/or ILC3 transdifferentiation. However, such ILC2 plasticity from bona fide ILC1 or ILC3 has, to the best of our knowledge, never been demonstrated in either mice or humans. Nevertheless, our unbiased identification of CRTH2<sup>–</sup> ILC2 challenges our understanding of human ILC2 biology, and has consequences for the identification of activated human ILC2 and strategies to target these cells in type 2 inflammation in e.g., allergy and asthma. The failure of the CRTH2 antagonist Fevipiprant in moderate-to-severe asthma<sup><xref ref-type="bibr" rid="CR46">46</xref>,<xref ref-type="bibr" rid="CR47">47</xref></sup> may, at least in part, be explained by our findings. The expression of the cell surface receptor SLAMF1 could be helpful to distinguish CRTH2<sup>–</sup> ILC2 from ILC1 and ILC3 in the lung.</p><p id="Par28">The existence of ILC1 in different human tissues in homeostasis and inflammation has been a topic of much debate.<sup><xref ref-type="bibr" rid="CR48">48</xref>–<xref ref-type="bibr" rid="CR51">51</xref></sup> While Tbet<sup>+</sup> non-NK/T/B lymphocytes, fulfilling the definition of ILC1, clearly accumulate in inflamed human intestinal tissues,<sup><xref ref-type="bibr" rid="CR16">16</xref>,<xref ref-type="bibr" rid="CR52">52</xref>,<xref ref-type="bibr" rid="CR53">53</xref></sup> Simoni et al. have suggested that in non-affected human tissues, ILC1s are artifacts derived from contamination, by e.g., T cells. In partial support of this, Roan et al. showed that a fraction (7%) of blood ILC1 upregulate surface expression of TCRα/β following a 7-day culture.<sup><xref ref-type="bibr" rid="CR39">39</xref></sup> While this shows that indeed blood ILC1 might contain a small fraction of T cells that have downregulated their surface TCRs, the majority of blood ILC1 are likely not T cells in disguise despite expressing intracellular CD3 epsilon.<sup><xref ref-type="bibr" rid="CR39">39</xref></sup> In fact, CD3 epsilon overexpression in the absence of TCR expression in mice has been suggested to block T cell development.<sup><xref ref-type="bibr" rid="CR54">54</xref></sup> Furthermore, cytoplasmic CD3 epsilon expression might indicate that blood ILC1 are immature, as it is observed in fetal but not adult NK cells.<sup><xref ref-type="bibr" rid="CR55">55</xref></sup> In further support of that, CD5<sup>+</sup> ILCs, expressing intracellular CD3 epsilon, have been suggested as precursors of mature IFN-γ-producing ILC1.<sup><xref ref-type="bibr" rid="CR56">56</xref></sup> We identified a particularly pronounced T-cell signature, including <italic toggle="yes">CD3E</italic>, encoding CD3 epsilon, in putative ILC1 and <italic toggle="yes">EOMES</italic><sup>+</sup> ILC1 in blood and tonsil. We therefore analyzed TCR rearrangement to better understand ILC1-T cell relationships. <italic toggle="yes">EOMES</italic><sup>+</sup> ILC1 and a subset of ILC1 showed pronounced rearrangement in TCRα/β chains as well as expression of <italic toggle="yes">CD4</italic> and <italic toggle="yes">CD8A</italic>, respectively. Despite lacking surface expression of TCRs and CD3, these ILCs might very well be T cell contaminants, which was also supported by partial clonal overlap between TCRα/β-rearranged T cells and ILC1. Clonal overlap was however uncommon, and therefore it is also possible that these ILCs are “failed T cells” with non-productive TCRα/β- rearrangement, leading to a lack of surface TCR expression. Notably, in ILCs, the expression of TCRα/β variable chains was associated with TCRα/β rearrangement and could potentially be used as a proxy for rearrangement in these cells.</p><p id="Par29">We also identified a cluster of ILC1, which was transcriptionally distinct from T cells, with rearrangement in predominantly TCR-δ chains. Indeed, mouse NK cell precursors show V(D)J-recombination activity<sup><xref ref-type="bibr" rid="CR57">57</xref></sup> and up to a third of human blood NK cells display rearrangement of δ-TCR chains.<sup><xref ref-type="bibr" rid="CR58">58</xref></sup> Hence, similar to intracellular CD3 epsilon expression,<sup><xref ref-type="bibr" rid="CR55">55</xref></sup> δ-chain rearrangement in blood ILCs might be a sign of immaturity and a history of recombination enzyme activity.</p><p id="Par30">In summary, our comparisons of ILCs and CD4<sup>+</sup> T cells in human blood unveiled clusters of putative ILC1 that show TCRα/β-rearrangement and clonal overlap with T cells. The latter are likely T cell contaminants with down-regulated surface TCRs while the remainder could be “failed T cells”, with non-productive TCR rearrangement. We also identified ILCs with TCR-δ rearrangement reminiscent of immature ILCs, rather than differentiated Tbet<sup>+</sup> ILC1. Our findings provide important first steps towards more extensive, elaborate studies on rearrangement in subsets of T and ILCs/NK cells across the innate and adaptive landscape and the mechanisms, including development, underlying this unexpected TCR rearrangement.</p><p id="Par31">In conclusion, we used scRNA-seq to identify a plethora of ILC subsets ranging from circulating and tissue-resident nILCs to subsets of mature, activated and tissue-resident ILCs with altered metabolism across four tissues. Deepened analysis unveiled the presence of lung-specific CRTH2<sup>−</sup> ILC2, challenging our understanding of ILC2-mediated diseases. Furthermore, TCR-V(D)J rearrangement studies identified widespread predicted TCRα/β-rearrangement and a distinct subset of putative immature ILCs among the ILCs commonly referred to as ILC1 in peripheral blood. Our findings constitute a roadmap for future studies aimed at understanding tissue-specific ILC-mediated immunity in immunopathology.</p></sec><sec id="Sec11" sec-type="materials|methods"><title>Materials and methods</title><sec id="Sec12"><title>Human tissue samples</title><p id="Par32">The Swedish ethical review authority approved all studies of human tissues and all studies were performed according to the Declaration of Helsinki. All regulations for work with human tissues were complied with. All study subjects gave their informed written consent prior to tissue collection.</p></sec><sec id="Sec13"><title>Cell isolation from lung tissue</title><p id="Par33">Lung tissue was obtained from four patients undergoing lobectomy for lung cancer (age 57–76 year old, three females, one male) at Karolinska University Hospital, Sweden. Non-affected tissue was obtained distally from the tumor. None of the patients had a diagnosis of asthma or chronic obstructive pulmonary disease (COPD). All patients were ex-smokers. None of the patients had received preoperative chemo- or radiotherapy. Lung cells were isolated from human lung tissue by a sequential combination of washing, enzymatic digestion, mechanical disruption and density centrifugation using Percoll gradient medium (WEMP protocol), as recently published by us<sup><xref ref-type="bibr" rid="CR59">59</xref></sup> where all details of the protocol can be found. Briefly, freshly resected lung tissue was washed in PBS and cut into smaller pieces, then enzymatically digested while stirring. The resulting digested product was cut and mechanically disrupted by syringe plunging and filtering. The filtered cell suspension was washed in PBS, resuspended in 30% Percoll (one phase) and centrifuged. The cell pellet was resuspended and red blood cells were lysed and washed with PBS 2% fetal calf serum (FCS) and cells were counted. The cell suspension was then stained with fluorescence-labeled antibodies for FACS.</p></sec><sec id="Sec14"><title>Cell isolation from whole blood</title><p id="Par34">Peripheral blood mononuclear cells (PBMCs) were isolated from whole venous blood from three healthy donors (age 29–32 year old, two males, one female). 27 mL of fresh blood was collected in EDTA tubes and directly diluted 1:1 in PBS. PBMC were isolated via lymphoprep (Axis Shield) and gradient centrifugation for 20 min at 2000× rpm. Lymphocytes were then collected, washed and counted. The cell suspension was then stained with fluorescence-labeled antibodies for FACS.</p></sec><sec id="Sec15"><title>Cell isolation from colon biopsies</title><p id="Par35">Biopsies from the ascending colon were obtained from three tumor-screening patients (age 57–71 year old, two males, one female) with known genetic predisposition for colon cancer (hereditary nonpolyposis colorectal cancer (HPNCC) or Lynch Syndrome) but without any polyps or tumors at the time of sampling. Colonoscopies were performed at Karolinska University Hospital or GHP Stockholm Gastro Center, Sweden. Biopsy retrieval and cell isolation was performed as previously described,<sup><xref ref-type="bibr" rid="CR16">16</xref></sup> where all the details of the protocol can be found. Briefly, eight punch-biopsies per patients were collected and transported in wash buffer, consisting of Hank’s Balanced Salt Solution (HBSS) supplemented with antibiotics, fungizone and HEPES. Samples were processed within 1 h of biopsy retrieval. Wash buffer was removed and biopsies were enzymatically digested while stirring. The filtered cell suspension was washed by centrifugation and the total cell pellet was used for staining with fluorescence-labeled antibodies for FACS.</p></sec><sec id="Sec16"><title>Tonsil data</title><p id="Par36">The previously published tonsil raw data used in this study can be found at GEO: <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="pmc:entrez-geo" xlink:href="GSE70580">GSE70580</ext-link>; Sequence Read Archive: SRP060416.</p></sec><sec id="Sec17"><title>Cell sorting</title><p id="Par37">Isolated cells were incubated with live/dead cell marker (DCM; Invitrogen) and surface protein antibodies (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Table <xref rid="MOESM1" ref-type="media">S1</xref>) at 4 °C for 30 min, washed with PBS and resuspended in Iscove’s modified Dulbecco’s medium IMDM medium supplemented with 1% normal human serum (NHS; Invitrogen) and Yssels supplement (made in house). Cells were sorted into 384-well plates (Eppendorf) containing 2.3 μL of lysis buffer per well. Lysis buffer (per well) consisted of 1.150 μL of 0.4% Triton X-100 (Sigma), 0.4 μL dNTPs mix (25 mM) (Thermo Life Tech), 0.05 μL Smart-dT<sub>30</sub>VN (100uM) (IDT), 0.05 μL RNAse inhibitor (40 U/μL) (Takara), 0.0025 μL ERCC (1:40,000 dilution) (Ambion) and 0.6475 μL of water (Thermo Life Tech). Cells were sorted on an Aria FUSION sorter (BD Biosciences), with FACS Diva version 8 software and indexed data analysis enabled. Sorting of blood ILC2 were performed on a SONY MA900 Multi-Application cell sorter.</p></sec><sec id="Sec18"><title>In vitro cultures</title><p id="Par38">Sort-purified peripheral blood ILC2 (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S5</xref>) were cultured in IMDM (Fisher Scientific) supplemented with 2% NHS and Yssels supplement in a 96 well plate. Cells were left in culture for 48 h at 37 °C with IL-2 (10 U/mL) or IL-2 (10 U/mL), IL-25 (R&amp;D), IL-33 (PeproTech) plus TSLP (PeproTech) (at 50 ng/mL each). Following culture, the cells were washed and stained with fluorescence-labeled antibodies for flow cytometry.</p></sec><sec id="Sec19"><title>Flow cytometry</title><p id="Par39">For flow cytometry analysis, cells were stained with DCM and antibodies (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Table <xref rid="MOESM1" ref-type="media">S1</xref>) for 30 min at 4 °C. Cells were washed in PBS supplemented with 2 mM EDTA and 2% FCS and fixed with 2% paraformaldehyde for 10 min at room temperature. Cells were acquired on an LSR Fortessa flow cytometer (BD Biosciences) with FACSDiva software version 8. Data were analyzed with FlowJo software version 9.9.6 (Tree Star).</p></sec><sec id="Sec20"><title>Preparation of cDNA libraries and sequencing</title><p id="Par40">Sorted cells were processed using the Smart-seq2 protocol with minor changes.<sup><xref ref-type="bibr" rid="CR60">60</xref></sup> Sequencing was done to the average depth of 0.7 million reads per cell with 43 bp reads (single end) after demultiplexing. Blood samples were additionally sequenced for V(D)J analysis to and average depth of 0.84 million reads per cell with 150 bp reads (paired end) after demultiplexing.</p></sec><sec id="Sec21"><title>Read alignment and gene-expression estimation</title><p id="Par41">The reads were aligned to the human genome (hg19–GRCh38) and spike-in sequences (from the External RNA Controls Consortium) using STAR v2.5.3a,<sup><xref ref-type="bibr" rid="CR61">61</xref></sup> and filtered for uniquely mapping reads. Counts per gene were calculated for each transcript in Ensembl release 92 using rpkmforgenes.<sup><xref ref-type="bibr" rid="CR62">62</xref></sup></p></sec><sec id="Sec22"><title>Data filtering and normalization</title><p id="Par42">Bioinformatics analyses were performed using R programming language, using Seurat v3 package<sup><xref ref-type="bibr" rid="CR63">63</xref></sup> with modifications. Reads were normalized to transcripts per million reads (TPMs) in order to compensate for gene length in transcript count estimation. Values were then multiplied by the read length (43 bp) and rounded to the nearest integer. Cell libraries were considered of low quality and filtered out if: (1) the percentages of mitochondrial and ribosomal gene families were above 40% or 50%, respectively; (2) the percentage of protein-coding genes was below 50%; (3) the number of unique genes in a cell was below 200. Genes detected in less than five cells were excluded from the analysis. Counts were next log-normalized with scale factor of 1000 and Z-score transformed for regression of confounding factors such as effects related to mitochondrial, ribosomal, Y chromosome, cell cycle, sex and proportion of non-coding RNA.</p></sec><sec id="Sec23"><title>Dimensionality reduction and clustering</title><p id="Par43">Normalized data were further processed for identification of highly variable genes using the Scran method<sup><xref ref-type="bibr" rid="CR64">64</xref></sup> and linear data compression using principal component analysis (PCA) from which the top 50 principal components (PCs) explaining about 70% of the variance in the dataset were chosen for downstream analysis. Non-linear dimensionality reduction was run on the top 50 PCs using UMAP<sup><xref ref-type="bibr" rid="CR15">15</xref></sup> and embedded into two final dimensions. Graph construction was also run on the top 50 PCs using approximate nearest neighbor (KNN)<sup><xref ref-type="bibr" rid="CR65">65</xref></sup> search with k = 10, reflecting the minimum number of cells expected to be considered a potentially rare cell cluster. Spurious connections were filtered using shared nearest-neighbors (SNN) search using Jacquard’s index of 1/15. Unsupervised graph clustering was run using the Louvain method, in which the modularity resolution parameter (res = 2.4) was optimized to match both known sample metadata information as well as a coherent cluster distribution across the UMAP embedding.</p></sec><sec id="Sec24"><title>Differential expression and gene module identification using WGCNA</title><p id="Par44">Identification of marker genes associated to specific cell clusters was done using differential expression via MAST.<sup><xref ref-type="bibr" rid="CR66">66</xref></sup> For standard differential expression log<sub>2</sub>FC  &gt; 0.25 and false discovery rate (FDR) &lt; 0.01 was used.</p><p id="Par45">For identification of the most relevant genes across cell clusters to be later used for weighted gene correlation network analysis (WGCNA),<sup><xref ref-type="bibr" rid="CR67">67</xref></sup> a permissive threshold with absolute log<sub>2</sub>FC &gt; 0.1 and <italic toggle="yes">P</italic> value &lt; 0.1 was used. Briefly, the 4410 genes identified in this manner were used for calculation of gene-to-gene Pearson correlations. The adjacency matrix was directly used for creation of the topological overlap matrix (TOM) using again Pearson as correlation measure. Power transformation of adjacency values were ignored since it would result in flattening all values to near 0, as even strong single-cell correlations rarely reach close to 1 due to the nature of data sparsity. No significant differences in the results were found when using Pearson and Spearman correlations, and therefore Pearson correlations were kept due to its faster overall performance. Gene module identification was done with agglomerative hierarchical clustering on the TOM using Ward’s squared linkage (Ward.D2) with a fixed number of 100 modules. Gene counts belonging to the same module were normalized between 0–1 individually and then averaged per cell for calculation of each gene module expression pattern.</p><p id="Par46">Modules were prioritized by fitting a general linear model (GLM) using each gene module expression pattern as the effect and the sample metadata as explaining variable. Each module and metadata parameter was tested independently. The output of the GLM fit results in the residual metric of “deviance” (<inline-formula id="IEq1"><alternatives><tex-math id="M1"><?equation-image-name M1.gif?><?equation-image-status READY?><?equation-image-md5 c4607af0a99f868dfd0559aaebd9649d?><?equation-image-cloudpmc-urn urn:cdn:blobs/f93f/8089104/c4607af0a99f/M1.gif?>\documentclass[12pt]{minimal}
+				\usepackage{amsmath}
+				\usepackage{wasysym} 
+				\usepackage{amsfonts} 
+				\usepackage{amssymb} 
+				\usepackage{amsbsy}
+				\usepackage{mathrsfs}
+				\usepackage{upgreek}
+				\setlength{\oddsidemargin}{-69pt}
+				\begin{document}$$dev_{model}$$\end{document}</tex-math><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M2"><mml:mi>d</mml:mi><mml:mi>e</mml:mi><mml:msub><mml:mrow><mml:mi>v</mml:mi></mml:mrow><mml:mrow><mml:mi>m</mml:mi><mml:mi>o</mml:mi><mml:mi>d</mml:mi><mml:mi>e</mml:mi><mml:mi>l</mml:mi></mml:mrow></mml:msub></mml:math><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="41422_2020_445_Article_IEq1.gif"><?image-name 41422_2020_445_Article_IEq1.gif?><?image-size 381?><?image-md5 6af2342f02c78c6362724c54d459dc94?><?image-image-server-status NEVER_LOAD?><?image-scaled-height 29?><?image-scaled-width 58?><?image-cloudpmc-urn urn:cdn:blobs/f93f/8089104/6af2342f02c7/41422_2020_445_Article_IEq1.gif?><?thumb-name 41422_2020_445_Article_IEq1.gif?><?thumb-size 381?><?thumb-md5 6af2342f02c78c6362724c54d459dc94?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 29?><?thumb-scaled-width 58?><?thumb-cloudpmc-urn urn:cdn:blobs/f93f/8089104/6af2342f02c7/41422_2020_445_Article_IEq1.gif?></inline-graphic></alternatives></inline-formula>) after fitting the model and “null deviance” (<inline-formula id="IEq2"><alternatives><tex-math id="M3"><?equation-image-name M3.gif?><?equation-image-status READY?><?equation-image-md5 32f3dbfa923fec35dae15e3c4eeb1add?><?equation-image-cloudpmc-urn urn:cdn:blobs/f93f/8089104/32f3dbfa923f/M3.gif?>\documentclass[12pt]{minimal}
+				\usepackage{amsmath}
+				\usepackage{wasysym} 
+				\usepackage{amsfonts} 
+				\usepackage{amssymb} 
+				\usepackage{amsbsy}
+				\usepackage{mathrsfs}
+				\usepackage{upgreek}
+				\setlength{\oddsidemargin}{-69pt}
+				\begin{document}$$dev_{null}$$\end{document}</tex-math><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M4"><mml:mi>d</mml:mi><mml:mi>e</mml:mi><mml:msub><mml:mrow><mml:mi>v</mml:mi></mml:mrow><mml:mrow><mml:mi>n</mml:mi><mml:mi>u</mml:mi><mml:mi>l</mml:mi><mml:mi>l</mml:mi></mml:mrow></mml:msub></mml:math><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="41422_2020_445_Article_IEq2.gif"><?image-name 41422_2020_445_Article_IEq2.gif?><?image-size 335?><?image-md5 c55ccce912d6e8616367e3399548c8ac?><?image-image-server-status NEVER_LOAD?><?image-scaled-height 29?><?image-scaled-width 48?><?image-cloudpmc-urn urn:cdn:blobs/f93f/8089104/c55ccce912d6/41422_2020_445_Article_IEq2.gif?><?thumb-name 41422_2020_445_Article_IEq2.gif?><?thumb-size 335?><?thumb-md5 c55ccce912d6e8616367e3399548c8ac?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 29?><?thumb-scaled-width 48?><?thumb-cloudpmc-urn urn:cdn:blobs/f93f/8089104/c55ccce912d6/41422_2020_445_Article_IEq2.gif?></inline-graphic></alternatives></inline-formula>) when considering a NULL model. The percent difference in these two deviances (<inline-formula id="IEq3"><alternatives><tex-math id="M5"><?equation-image-name M5.gif?><?equation-image-status READY?><?equation-image-md5 06d13fd256a6223ae7ce229698947da0?><?equation-image-cloudpmc-urn urn:cdn:blobs/f93f/8089104/06d13fd256a6/M5.gif?>\documentclass[12pt]{minimal}
+				\usepackage{amsmath}
+				\usepackage{wasysym} 
+				\usepackage{amsfonts} 
+				\usepackage{amssymb} 
+				\usepackage{amsbsy}
+				\usepackage{mathrsfs}
+				\usepackage{upgreek}
+				\setlength{\oddsidemargin}{-69pt}
+				\begin{document}$${\Delta}dev$$\end{document}</tex-math><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M6"><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi>d</mml:mi><mml:mi>e</mml:mi><mml:mi>v</mml:mi></mml:math><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="41422_2020_445_Article_IEq3.gif"><?image-name 41422_2020_445_Article_IEq3.gif?><?image-size 301?><?image-md5 56b3901db1edf2387a268baeb08ff27a?><?image-image-server-status NEVER_LOAD?><?image-scaled-height 30?><?image-scaled-width 40?><?image-cloudpmc-urn urn:cdn:blobs/f93f/8089104/56b3901db1ed/41422_2020_445_Article_IEq3.gif?><?thumb-name 41422_2020_445_Article_IEq3.gif?><?thumb-size 301?><?thumb-md5 56b3901db1edf2387a268baeb08ff27a?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 30?><?thumb-scaled-width 40?><?thumb-cloudpmc-urn urn:cdn:blobs/f93f/8089104/56b3901db1ed/41422_2020_445_Article_IEq3.gif?></inline-graphic></alternatives></inline-formula>) was used as a measure of how much a particular metadata factor explains the expression pattern of a gene module:<disp-formula id="Equa"><alternatives><tex-math id="M7"><?equation-image-name M7.gif?><?equation-image-status READY?><?equation-image-md5 a2d7f747bf23cb444da4c1c47c300b56?><?equation-image-cloudpmc-urn urn:cdn:blobs/f93f/8089104/a2d7f747bf23/M7.gif?>\documentclass[12pt]{minimal}
+				\usepackage{amsmath}
+				\usepackage{wasysym} 
+				\usepackage{amsfonts} 
+				\usepackage{amssymb} 
+				\usepackage{amsbsy}
+				\usepackage{mathrsfs}
+				\usepackage{upgreek}
+				\setlength{\oddsidemargin}{-69pt}
+				\begin{document}$${\Delta}dev = \frac{{dev_{null} - dev_{model}}}{{dev_{null}}}$$\end{document}</tex-math><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M8"><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi>d</mml:mi><mml:mi>e</mml:mi><mml:mi>v</mml:mi><mml:mo>=</mml:mo><mml:mfrac><mml:mrow><mml:mi>d</mml:mi><mml:mi>e</mml:mi><mml:msub><mml:mrow><mml:mi>v</mml:mi></mml:mrow><mml:mrow><mml:mi>n</mml:mi><mml:mi>u</mml:mi><mml:mi>l</mml:mi><mml:mi>l</mml:mi></mml:mrow></mml:msub><mml:mo>−</mml:mo><mml:mi>d</mml:mi><mml:mi>e</mml:mi><mml:msub><mml:mrow><mml:mi>v</mml:mi></mml:mrow><mml:mrow><mml:mi>m</mml:mi><mml:mi>o</mml:mi><mml:mi>d</mml:mi><mml:mi>e</mml:mi><mml:mi>l</mml:mi></mml:mrow></mml:msub></mml:mrow><mml:mrow><mml:mi>d</mml:mi><mml:mi>e</mml:mi><mml:msub><mml:mrow><mml:mi>v</mml:mi></mml:mrow><mml:mrow><mml:mi>n</mml:mi><mml:mi>u</mml:mi><mml:mi>l</mml:mi><mml:mi>l</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:mfrac></mml:math><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="anchor" orientation="portrait" xlink:href="41422_2020_445_Article_Equa.gif"><?image-name 41422_2020_445_Article_Equa.gif?><?image-size 1016?><?image-md5 f5fc49db37c86c8794e02379a8347a7d?><?image-image-server-status NEVER_LOAD?><?image-scaled-height 53?><?image-scaled-width 179?><?image-cloudpmc-urn urn:cdn:blobs/f93f/8089104/f5fc49db37c8/41422_2020_445_Article_Equa.gif?><?thumb-name 41422_2020_445_Article_Equa.gif?><?thumb-size 1016?><?thumb-md5 f5fc49db37c86c8794e02379a8347a7d?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 53?><?thumb-scaled-width 179?><?thumb-cloudpmc-urn urn:cdn:blobs/f93f/8089104/f5fc49db37c8/41422_2020_445_Article_Equa.gif?></graphic></alternatives></disp-formula></p><p id="Par47">This way, modules with higher <inline-formula id="IEq4"><alternatives><tex-math id="M9"><?equation-image-name M9.gif?><?equation-image-status READY?><?equation-image-md5 06d13fd256a6223ae7ce229698947da0?><?equation-image-cloudpmc-urn urn:cdn:blobs/f93f/8089104/06d13fd256a6/M9.gif?>\documentclass[12pt]{minimal}
+				\usepackage{amsmath}
+				\usepackage{wasysym} 
+				\usepackage{amsfonts} 
+				\usepackage{amssymb} 
+				\usepackage{amsbsy}
+				\usepackage{mathrsfs}
+				\usepackage{upgreek}
+				\setlength{\oddsidemargin}{-69pt}
+				\begin{document}$${\Delta}dev$$\end{document}</tex-math><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M10"><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi>d</mml:mi><mml:mi>e</mml:mi><mml:mi>v</mml:mi></mml:math><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="41422_2020_445_Article_IEq4.gif"><?image-name 41422_2020_445_Article_IEq4.gif?><?image-size 301?><?image-md5 56b3901db1edf2387a268baeb08ff27a?><?image-image-server-status NEVER_LOAD?><?image-scaled-height 30?><?image-scaled-width 40?><?image-cloudpmc-urn urn:cdn:blobs/f93f/8089104/56b3901db1ed/41422_2020_445_Article_IEq4.gif?><?thumb-name 41422_2020_445_Article_IEq4.gif?><?thumb-size 301?><?thumb-md5 56b3901db1edf2387a268baeb08ff27a?><?thumb-image-server-status NEVER_LOAD?><?thumb-scaled-height 30?><?thumb-scaled-width 40?><?thumb-cloudpmc-urn urn:cdn:blobs/f93f/8089104/56b3901db1ed/41422_2020_445_Article_IEq4.gif?></inline-graphic></alternatives></inline-formula> were prioritized for further exploration throughout the manuscript. Modules enriched for ribosomal, mitochondrial and other known housekeeping families were mentioned but not further analyzed.</p></sec><sec id="Sec25"><title>Trajectory inference analysis</title><p id="Par48">Trajectory inference analysis was performed using Slingshot v.1.2.0<sup><xref ref-type="bibr" rid="CR25">25</xref></sup> on the Diffusion Map embedding (using five components), since the UMAP embedding creates a distorted representation of the original space. Lineages were defined by setting the naïve clusters (nILC) as the root, and curves were projected onto the Diffusion Map embedding for visualization.</p></sec><sec id="Sec26"><title>TCR rearrangement analysis</title><p id="Par49">For blood data with 150 bp PE sequencing, MiXCR<sup><xref ref-type="bibr" rid="CR40">40</xref></sup> was run on all cells using the “analyze shotgun” command and default settings. To avoid false positives MiXCR output was filtered for clones with count &gt; 5. If multiple predicted clones were found, only dominant predicted clone from each chain (TRA, TRB, TRD or TRG) was included in downstream analysis.</p></sec><sec id="Sec27"><title>Analysis of human 10× genomics scRNA-seq PBMC data</title><p id="Par50">Human PBMC data with 10× Genomics scRNA-seq, cell surface protein detection with antibody-derived tags (ADT) and V(D)J assay was downloaded from the 10× genomics website (<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="https://support.10xgenomics.com/single-cell-vdj/datasets/3.0.0/vdj_v1_hs_pbmc2_5gex_protein">https://support.10xgenomics.com/single-cell-vdj/datasets/3.0.0/vdj_v1_hs_pbmc2_5gex_protein</ext-link>). Dimensionality reduction and clustering was performed with Seurat v3.1.3 with the whole dataset. T-cell and NK-cell clusters were selected based on the protein expression, e.g. CD3<sup>+</sup> and CD56<sup>+</sup>CD3<sup>–</sup> clusters were selected for further analysis. Clustering and dimensionality reduction was performed again with this subset, as shown in Supplementary Fig. <xref rid="MOESM1" ref-type="media">8a, b</xref> and cell types annotated based on top DEGs using MAST (<xref rid="MOESM1" ref-type="media">Supplementary Information</xref>, Fig. <xref rid="MOESM1" ref-type="media">S8c</xref>). At this stage, a contaminating cluster 8 that is low in CD3 and CD56, but high in CD16 and CD19, was removed as it likely represented doublets or cells that were wrongly clustered in the first round.</p><p id="Par51">To compare the 10× PBMC clusters to our blood SS2 data, we looked at intersections of DEGs. All significant upregulated DEGs (<italic toggle="yes">P</italic> &lt; 0.01) for each cluster within the 10× data and SS2 data were calculated separately. In the case of the AB_T_ILC1 cluster, cells were separated into T-cells (AB_T) and ILCs (AB_ILC) and DEG analysis was performed separately for those two groups. For each comparison, phyper test <italic toggle="yes">P</italic> values were calculated for overlap between the gene lists with all genes expressed in five cells in both datasets as background.</p><p id="Par52">For correlation between 10× PBMC and blood SS2 data, top 100 DEGs (or fewer if not 100 significant genes) from both datasets were used to calculate mean expression for each cluster. Spearman correlations between the clusters in each dataset were then calculated for the mean expression.</p></sec></sec><sec sec-type="supplementary-material"><title>Supplementary information</title><sec id="Sec28"><supplementary-material content-type="local-data" id="MOESM1" position="float" orientation="portrait"><media xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="41422_2020_445_MOESM1_ESM.docx" position="float" orientation="portrait"><?suppdata-name 41422_2020_445_MOESM1_ESM.docx?><?suppdata-size 133857?><?suppdata-md5 589de6e0ee63c04e59f580f040687ddc?><?suppdata-image-server-status NEVER_LOAD?><?suppdata-mime-type application?><?suppdata-mime-sub-type vnd.openxmlformats-officedocument.wordprocessingml.document?><?suppdata-cloudpmc-urn urn:app:f93f/8089104/589de6e0ee63/41422_2020_445_MOESM1_ESM.docx?><caption><p>Supplementary Text</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="MOESM2" position="float" orientation="portrait"><media xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="41422_2020_445_MOESM2_ESM.pdf" position="float" orientation="portrait"><?suppdata-name 41422_2020_445_MOESM2_ESM.pdf?><?suppdata-size 2270864?><?suppdata-md5 bae935f556ac6c758fcc23692931c395?><?suppdata-image-server-status NEVER_LOAD?><?suppdata-mime-type application?><?suppdata-mime-sub-type pdf?><?suppdata-cloudpmc-urn urn:app:f93f/8089104/bae935f556ac/41422_2020_445_MOESM2_ESM.pdf?><caption><p>Supplementary Figure S1</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="MOESM3" position="float" orientation="portrait"><media xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="41422_2020_445_MOESM3_ESM.pdf" position="float" orientation="portrait"><?suppdata-name 41422_2020_445_MOESM3_ESM.pdf?><?suppdata-size 615258?><?suppdata-md5 337b385e05d468b86be0976013589818?><?suppdata-image-server-status NEVER_LOAD?><?suppdata-mime-type application?><?suppdata-mime-sub-type pdf?><?suppdata-cloudpmc-urn urn:app:f93f/8089104/337b385e05d4/41422_2020_445_MOESM3_ESM.pdf?><caption><p>Supplementary Figure S2</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="MOESM4" position="float" orientation="portrait"><media xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="41422_2020_445_MOESM4_ESM.pdf" position="float" orientation="portrait"><?suppdata-name 41422_2020_445_MOESM4_ESM.pdf?><?suppdata-size 686978?><?suppdata-md5 407c39553475711d7f74ce44c6ab4837?><?suppdata-image-server-status NEVER_LOAD?><?suppdata-mime-type application?><?suppdata-mime-sub-type pdf?><?suppdata-cloudpmc-urn urn:app:f93f/8089104/407c39553475/41422_2020_445_MOESM4_ESM.pdf?><caption><p>Supplementary Figure S3</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="MOESM5" position="float" orientation="portrait"><media xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="41422_2020_445_MOESM5_ESM.pdf" position="float" orientation="portrait"><?suppdata-name 41422_2020_445_MOESM5_ESM.pdf?><?suppdata-size 790036?><?suppdata-md5 0befd4db4f9b67e250b90761a2c8b979?><?suppdata-image-server-status NEVER_LOAD?><?suppdata-mime-type application?><?suppdata-mime-sub-type pdf?><?suppdata-cloudpmc-urn urn:app:f93f/8089104/0befd4db4f9b/41422_2020_445_MOESM5_ESM.pdf?><caption><p>Supplementary Figure S4</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="MOESM6" position="float" orientation="portrait"><media xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="41422_2020_445_MOESM6_ESM.pdf" position="float" orientation="portrait"><?suppdata-name 41422_2020_445_MOESM6_ESM.pdf?><?suppdata-size 1157316?><?suppdata-md5 0f8da854fb707a0b01f9fcbdc8e68e10?><?suppdata-image-server-status NEVER_LOAD?><?suppdata-mime-type application?><?suppdata-mime-sub-type pdf?><?suppdata-cloudpmc-urn urn:app:f93f/8089104/0f8da854fb70/41422_2020_445_MOESM6_ESM.pdf?><caption><p>Supplementary Figure S5</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="MOESM7" position="float" orientation="portrait"><media xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="41422_2020_445_MOESM7_ESM.pdf" position="float" orientation="portrait"><?suppdata-name 41422_2020_445_MOESM7_ESM.pdf?><?suppdata-size 1553514?><?suppdata-md5 e7da264570dc29704595395a52ed7777?><?suppdata-image-server-status NEVER_LOAD?><?suppdata-mime-type application?><?suppdata-mime-sub-type pdf?><?suppdata-cloudpmc-urn urn:app:f93f/8089104/e7da264570dc/41422_2020_445_MOESM7_ESM.pdf?><caption><p>Supplementary Figure S6</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="MOESM8" position="float" orientation="portrait"><media xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="41422_2020_445_MOESM8_ESM.pdf" position="float" orientation="portrait"><?suppdata-name 41422_2020_445_MOESM8_ESM.pdf?><?suppdata-size 1395353?><?suppdata-md5 dcb8dd33e2dcff4c316cfe9f107be152?><?suppdata-image-server-status NEVER_LOAD?><?suppdata-mime-type application?><?suppdata-mime-sub-type pdf?><?suppdata-cloudpmc-urn urn:app:f93f/8089104/dcb8dd33e2dc/41422_2020_445_MOESM8_ESM.pdf?><caption><p>Supplementary Figure S7</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="MOESM9" position="float" orientation="portrait"><media xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="41422_2020_445_MOESM9_ESM.pdf" position="float" orientation="portrait"><?suppdata-name 41422_2020_445_MOESM9_ESM.pdf?><?suppdata-size 5516404?><?suppdata-md5 0bd558af6f7a316ed93e018c4607a45b?><?suppdata-image-server-status NEVER_LOAD?><?suppdata-mime-type application?><?suppdata-mime-sub-type pdf?><?suppdata-cloudpmc-urn urn:app:f93f/8089104/0bd558af6f7a/41422_2020_445_MOESM9_ESM.pdf?><caption><p>Supplementary Figure S8</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="MOESM10" position="float" orientation="portrait"><media xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="41422_2020_445_MOESM10_ESM.xlsx" position="float" orientation="portrait"><?suppdata-name 41422_2020_445_MOESM10_ESM.xlsx?><?suppdata-size 4934137?><?suppdata-md5 b7b2c5e407bd88572a935b16aacfcb56?><?suppdata-image-server-status NEVER_LOAD?><?suppdata-mime-type application?><?suppdata-mime-sub-type vnd.openxmlformats-officedocument.spreadsheetml.sheet?><?suppdata-cloudpmc-urn urn:app:f93f/8089104/b7b2c5e407bd/41422_2020_445_MOESM10_ESM.xlsx?><caption><p>Supplementary Data S1</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="MOESM11" position="float" orientation="portrait"><media xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="41422_2020_445_MOESM11_ESM.xlsx" position="float" orientation="portrait"><?suppdata-name 41422_2020_445_MOESM11_ESM.xlsx?><?suppdata-size 416063?><?suppdata-md5 bf428c22fd7805d32603dc2ee342a7ac?><?suppdata-image-server-status NEVER_LOAD?><?suppdata-mime-type application?><?suppdata-mime-sub-type vnd.openxmlformats-officedocument.spreadsheetml.sheet?><?suppdata-cloudpmc-urn urn:app:f93f/8089104/bf428c22fd78/41422_2020_445_MOESM11_ESM.xlsx?><caption><p>Supplementary Data S2</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="MOESM12" position="float" orientation="portrait"><media xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="41422_2020_445_MOESM12_ESM.xlsx" position="float" orientation="portrait"><?suppdata-name 41422_2020_445_MOESM12_ESM.xlsx?><?suppdata-size 39558?><?suppdata-md5 a5391bbad604bd8a784bf574f6691f76?><?suppdata-image-server-status NEVER_LOAD?><?suppdata-mime-type application?><?suppdata-mime-sub-type vnd.openxmlformats-officedocument.spreadsheetml.sheet?><?suppdata-cloudpmc-urn urn:app:f93f/8089104/a5391bbad604/41422_2020_445_MOESM12_ESM.xlsx?><caption><p>Supplementary Table S1</p></caption></media></supplementary-material></sec></sec></body><back><sec><title>Supplementary information</title><p><bold>Supplementary information</bold> accompanies this paper at 10.1038/s41422-020-00445-x.</p></sec><ack><title>Acknowledgements</title><p>The authors thank the doctors and the research nurses at Karolinska University Hospital for collection of colon samples: Lisa Sundin Zarouki, Karin Thourot Nouchi, Peter Elbe, Peter Thelin Schmidt, Aiva Lundberg Båve, Miroslav Vujasinovic, Rima Eldandachi, Per Bergenzaun, Masami Omae, Alexander Waldthaler, Reza Rezaie, Fransisico Silva, Apostolos Tsolakis and Peter Borch Johnsen. The authors thank Iyadh Douagi for support with FACS sorting, Anna Rao for technical assistance and Nicole Marquardt for aiding in lung sample collection. The authors thank the Uppsala Multidisciplinary Center for Advanced Computational Science (UPPMAX) for performing computational analyses on resources provided by the Swedish National Infrastructure for Computing (SNIC), under project SNIC 2017/7-123. P.C., V.J., L.W., M.R. and Å.K.B. are financially supported by the Knut and Alice Wallenberg Foundation as part of the National Bioinformatics Infrastructure Sweden at SciLifeLab. The research was supported by the Swedish Cancer Society, the Swedish Foundation for Strategic Research, the Swedish Foundation for Medical Research, the Swedish Research Council, the Knut and Alice Wallenberg foundation and Karolinska Institutet.</p></ack><notes notes-type="author-contribution"><title>Author contributions</title><p>J.M. conceived the study, interpreted data and co-wrote the manuscript with L.M. L.M., A.R. and T.C.W. designed and performed the experiments. J.M. and G.N. supervised the experiments. L.M., P.C., V.J., L.W., M.R. and Å.K.B. performed the bioinformatics analysis. J.S., A.-C.O., S.-E.D., M.A.-A., C.Höög, C.Hedin, S.S. and S.A. provided samples and clinical data. All authors helped in manuscript writing and/or editing.</p></notes><notes notes-type="data-availability"><title>Data availability</title><p>Data produced in this manuscript is available on Gene Expression Omnibus (GEO) with accession number <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="pmc:entrez-geo" xlink:href="GSE150050">GSE150050</ext-link>.</p></notes><notes notes-type="data-availability"><title>Code availability</title><p>All scripts used for the analysis can be found at <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="https://github.com/czarnewski/ilc_tissue_imprinting">https://github.com/czarnewski/ilc_tissue_imprinting</ext-link>.</p></notes><notes id="FPar1" notes-type="COI-statement"><title>Competing interests</title><p id="Par53">L.M. is a part-time employee of Janssen Pharmaceutical. The research was not performed within this employment. L.W. is currently an employee of AstraZeneca. The research was not performed within this employment. All other authors declare no competing interests.</p></notes><ref-list id="Bib1"><title>References</title><ref id="CR1"><label>1.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Spits</surname><given-names>H</given-names></name><etal/></person-group><article-title>Innate lymphoid cells–a proposal for uniform nomenclature</article-title><source>Nat. Rev. Immunol.</source><year>2013</year><volume>13</volume><fpage>145</fpage><lpage>149</lpage><pub-id pub-id-type="pmid">23348417</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nri3365</pub-id></element-citation></ref><ref id="CR2"><label>2.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Robinette</surname><given-names>ML</given-names></name><etal/></person-group><article-title>Transcriptional programs define molecular characteristics of innate lymphoid cell classes and subsets</article-title><source>Nat. Immunol.</source><year>2015</year><volume>16</volume><fpage>306</fpage><lpage>317</lpage><pub-id pub-id-type="pmid">25621825</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/ni.3094</pub-id><pub-id pub-id-type="pmcid">PMC4372143</pub-id></element-citation></ref><ref id="CR3"><label>3.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Yudanin</surname><given-names>NA</given-names></name><etal/></person-group><article-title>Spatial and temporal mapping of human innate lymphoid cells reveals elements of tissue specificity</article-title><source>Immunity</source><year>2019</year><volume>50</volume><fpage>505</fpage><lpage>519</lpage><pub-id pub-id-type="pmid">30770247</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2019.01.012</pub-id><pub-id pub-id-type="pmcid">PMC6594374</pub-id></element-citation></ref><ref id="CR4"><label>4.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Bar-Ephraim</surname><given-names>YE</given-names></name><etal/></person-group><article-title>Cross-tissue transcriptomic analysis of human secondary lymphoid organ-residing ILC3s reveals a quiescent state in the absence of inflammation</article-title><source>Cell Rep.</source><year>2017</year><volume>21</volume><fpage>823</fpage><lpage>833</lpage><pub-id pub-id-type="pmid">29045847</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.celrep.2017.09.070</pub-id></element-citation></ref><ref id="CR5"><label>5.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Xu</surname><given-names>G</given-names></name><etal/></person-group><article-title>Dissecting the human immune system with single cell RNA sequencing technology</article-title><source>J. Leukoc. Biol.</source><year>2019</year><volume>107</volume><fpage>613</fpage><lpage>623</lpage><pub-id pub-id-type="pmid">31803960</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1002/JLB.5MR1019-179R</pub-id></element-citation></ref><ref id="CR6"><label>6.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Crinier</surname><given-names>A</given-names></name><etal/></person-group><article-title>High-dimensional single-cell analysis identifies organ-specific signatures and conserved NK cell subsets in humans and mice</article-title><source>Immunity</source><year>2018</year><volume>49</volume><fpage>971</fpage><lpage>986</lpage><pub-id pub-id-type="pmid">30413361</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2018.09.009</pub-id><pub-id pub-id-type="pmcid">PMC6269138</pub-id></element-citation></ref><ref id="CR7"><label>7.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Ricardo-Gonzalez</surname><given-names>RR</given-names></name><etal/></person-group><article-title>Tissue signals imprint ILC2 identity with anticipatory function</article-title><source>Nat. Immunol.</source><year>2018</year><volume>19</volume><fpage>1093</fpage><lpage>1099</lpage><pub-id pub-id-type="pmid">30201992</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41590-018-0201-4</pub-id><pub-id pub-id-type="pmcid">PMC6202223</pub-id></element-citation></ref><ref id="CR8"><label>8.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Gury-BenAri</surname><given-names>M</given-names></name><etal/></person-group><article-title>The spectrum and regulatory landscape of intestinal innate lymphoid cells are shaped by the microbiome</article-title><source>Cell</source><year>2016</year><volume>166</volume><fpage>1231</fpage><lpage>1246</lpage><pub-id pub-id-type="pmid">27545347</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.cell.2016.07.043</pub-id></element-citation></ref><ref id="CR9"><label>9.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Xu</surname><given-names>H</given-names></name><etal/></person-group><article-title>Transcriptional atlas of intestinal immune cells reveals that neuropeptide α-CGRP modulates group 2 innate lymphoid cell responses</article-title><source>Immunity</source><year>2019</year><volume>51</volume><fpage>696</fpage><lpage>708</lpage><pub-id pub-id-type="pmid">31618654</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2019.09.004</pub-id><pub-id pub-id-type="pmcid">PMC6991097</pub-id></element-citation></ref><ref id="CR10"><label>10.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Martin</surname><given-names>JC</given-names></name><etal/></person-group><article-title>Single-cell analysis of crohn’s disease lesions identifies a pathogenic cellular module associated with resistance to anti-TNF therapy</article-title><source>Cell</source><year>2019</year><volume>178</volume><fpage>1493</fpage><lpage>1508</lpage><pub-id pub-id-type="pmid">31474370</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.cell.2019.08.008</pub-id><pub-id pub-id-type="pmcid">PMC7060942</pub-id></element-citation></ref><ref id="CR11"><label>11.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Cella</surname><given-names>M</given-names></name><etal/></person-group><article-title>Subsets of ILC3−ILC1-like cells generate a diversity spectrum of innate lymphoid cells in human mucosal tissues</article-title><source>Nat. Immunol.</source><year>2019</year><volume>20</volume><fpage>980</fpage><lpage>991</lpage><pub-id pub-id-type="pmid">31209406</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41590-019-0425-y</pub-id><pub-id pub-id-type="pmcid">PMC6685551</pub-id></element-citation></ref><ref id="CR12"><label>12.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Björklund</surname><given-names>ÅK</given-names></name><etal/></person-group><article-title>The heterogeneity of human CD127+ innate lymphoid cells revealed by single-cell RNA sequencing</article-title><source>Nat. Immunol.</source><year>2016</year><volume>17</volume><fpage>451</fpage><lpage>460</lpage><pub-id pub-id-type="pmid">26878113</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/ni.3368</pub-id></element-citation></ref><ref id="CR13"><label>13.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Marquardt</surname><given-names>N</given-names></name><etal/></person-group><article-title>Unique transcriptional and protein-expression signature in human lung tissue-resident NK cells</article-title><source>Nat. Commun.</source><year>2019</year><volume>10</volume><fpage>3841</fpage><pub-id pub-id-type="pmid">31451696</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41467-019-11632-9</pub-id><pub-id pub-id-type="pmcid">PMC6710242</pub-id></element-citation></ref><ref id="CR14"><label>14.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Sagebiel</surname><given-names>AF</given-names></name><etal/></person-group><article-title>Tissue-resident Eomes+ NK cells are the major innate lymphoid cell population in human infant intestine</article-title><source>Nat. Commun.</source><year>2019</year><volume>10</volume><fpage>975</fpage><pub-id pub-id-type="pmid">30816112</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41467-018-08267-7</pub-id><pub-id pub-id-type="pmcid">PMC6395753</pub-id></element-citation></ref><ref id="CR15"><label>15.</label><mixed-citation publication-type="other">McInnes, L., Healy, J. &amp; Melville, J. UMAP: uniform manifold approximation and projection for dimension reduction. arXiv: 1802.03426v3 (2018).</mixed-citation></ref><ref id="CR16"><label>16.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Forkel</surname><given-names>M</given-names></name><etal/></person-group><article-title>Distinct alterations in the composition of mucosal innate lymphoid cells in newly diagnosed and established Crohn’s disease and ulcerative colitis</article-title><source>J. Crohns Colitis</source><year>2019</year><volume>13</volume><fpage>67</fpage><lpage>78</lpage><pub-id pub-id-type="pmid">30496425</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1093/ecco-jcc/jjy119</pub-id></element-citation></ref><ref id="CR17"><label>17.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Lim</surname><given-names>AI</given-names></name><etal/></person-group><article-title>Systemic human ILC precursors provide a substrate for tissue ILC differentiation</article-title><source>Cell</source><year>2017</year><volume>168</volume><fpage>1086</fpage><lpage>1100</lpage><pub-id pub-id-type="pmid">28283063</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.cell.2017.02.021</pub-id></element-citation></ref><ref id="CR18"><label>18.</label><mixed-citation publication-type="other">Bal, S. M., Golebski, K. &amp; Spits, H. Plasticity of innate lymphoid cell subsets. <italic toggle="yes">Nat. Rev. Immunol</italic>. 1–14 10.1038/s41577-020-0282-9 (2020).<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41577-020-0282-9</pub-id><pub-id pub-id-type="pmid">32107466</pub-id></mixed-citation></ref><ref id="CR19"><label>19.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Gonzalez</surname><given-names>LL</given-names></name><name name-style="western"><surname>Garrie</surname><given-names>K</given-names></name><name name-style="western"><surname>Turner</surname><given-names>MD</given-names></name></person-group><article-title>Role of S100 proteins in health and disease</article-title><source>Biochim. Biophys. Acta Mol. Cell Res.</source><year>2020</year><volume>1867</volume><fpage>118677</fpage><pub-id pub-id-type="pmid">32057918</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.bbamcr.2020.118677</pub-id></element-citation></ref><ref id="CR20"><label>20.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Dutton</surname><given-names>EE</given-names></name><etal/></person-group><article-title>Peripheral lymph nodes contain migratory and resident innate lymphoid cell populations</article-title><source>Sci. Immunol</source><year>2019</year><volume>35</volume><fpage>eaau8082</fpage><pub-id pub-id-type="doi" assigning-authority="pmc">10.1126/sciimmunol.aau8082</pub-id><pub-id pub-id-type="pmcid">PMC7018521</pub-id><pub-id pub-id-type="pmid">31152090</pub-id></element-citation></ref><ref id="CR21"><label>21.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>McNamara</surname><given-names>HA</given-names></name><etal/></person-group><article-title>Up-regulation of LFA-1 allows liver-resident memory T cells to patrol and remain in the hepatic sinusoids</article-title><source>Sci. Immunol</source><year>2017</year><volume>9</volume><fpage>eaaj1996</fpage><pub-id pub-id-type="doi" assigning-authority="pmc">10.1126/sciimmunol.aaj1996</pub-id><pub-id pub-id-type="pmcid">PMC5505664</pub-id><pub-id pub-id-type="pmid">28707003</pub-id></element-citation></ref><ref id="CR22"><label>22.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Li</surname><given-names>J</given-names></name><name name-style="western"><surname>Olshansky</surname><given-names>M</given-names></name><name name-style="western"><surname>Carbone</surname><given-names>FR</given-names></name><name name-style="western"><surname>Ma</surname><given-names>JZ</given-names></name></person-group><article-title>Transcriptional analysis of T cells resident in human skin</article-title><source>PLoS One</source><year>2016</year><volume>11</volume><fpage>e0148351</fpage><pub-id pub-id-type="pmid">26824609</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1371/journal.pone.0148351</pub-id><pub-id pub-id-type="pmcid">PMC4732610</pub-id></element-citation></ref><ref id="CR23"><label>23.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Monticelli</surname><given-names>LA</given-names></name><etal/></person-group><article-title>IL-33 promotes an innate immune pathway of intestinal tissue protection dependent on amphiregulin-EGFR interactions</article-title><source>Proc. Natl. Acad. Sci. USA</source><year>2015</year><volume>112</volume><fpage>10762</fpage><lpage>10767</lpage><pub-id pub-id-type="pmid">26243875</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1073/pnas.1509070112</pub-id><pub-id pub-id-type="pmcid">PMC4553775</pub-id></element-citation></ref><ref id="CR24"><label>24.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Hirota</surname><given-names>K</given-names></name><etal/></person-group><article-title>Autoimmune Th17 cells induced synovial stromal and innate lymphoid cell secretion of the cytokine GM-CSF to initiate and augment autoimmune arthritis</article-title><source>Immunity</source><year>2018</year><volume>48</volume><fpage>1220</fpage><lpage>1232</lpage><pub-id pub-id-type="pmid">29802020</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2018.04.009</pub-id><pub-id pub-id-type="pmcid">PMC6024031</pub-id></element-citation></ref><ref id="CR25"><label>25.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Street</surname><given-names>K</given-names></name><etal/></person-group><article-title>Slingshot: cell lineage and pseudotime inference for single-cell transcriptomics</article-title><source>BMC Genomics</source><year>2018</year><volume>19</volume><fpage>477</fpage><pub-id pub-id-type="pmid">29914354</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1186/s12864-018-4772-0</pub-id><pub-id pub-id-type="pmcid">PMC6007078</pub-id></element-citation></ref><ref id="CR26"><label>26.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Go</surname><given-names>CK</given-names></name><name name-style="western"><surname>Gross</surname><given-names>S</given-names></name><name name-style="western"><surname>Hooper</surname><given-names>R</given-names></name><name name-style="western"><surname>Soboloff</surname><given-names>J</given-names></name></person-group><article-title>EGR-mediated control of STIM expression and function</article-title><source>Cell Calcium</source><year>2019</year><volume>77</volume><fpage>58</fpage><lpage>67</lpage><pub-id pub-id-type="pmid">30553973</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.ceca.2018.12.003</pub-id><pub-id pub-id-type="pmcid">PMC6467086</pub-id></element-citation></ref><ref id="CR27"><label>27.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Date</surname><given-names>D</given-names></name><etal/></person-group><article-title>Kruppel-like transcription factor 6 regulates inflammatory macrophage polarization</article-title><source>J. Biol. Chem.</source><year>2014</year><volume>289</volume><fpage>10318</fpage><lpage>10329</lpage><pub-id pub-id-type="pmid">24385430</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1074/jbc.M113.526749</pub-id><pub-id pub-id-type="pmcid">PMC4036156</pub-id></element-citation></ref><ref id="CR28"><label>28.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Emgård</surname><given-names>J</given-names></name><etal/></person-group><article-title>Oxysterol sensing through the receptor GPR183 promotes the lymphoid-tissue-inducing function of innate lymphoid cells and colonic inflammation</article-title><source>Immunity</source><year>2018</year><volume>48</volume><fpage>120</fpage><lpage>132</lpage><pub-id pub-id-type="pmid">29343433</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2017.11.020</pub-id><pub-id pub-id-type="pmcid">PMC5772175</pub-id></element-citation></ref><ref id="CR29"><label>29.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Aili</surname><given-names>A</given-names></name><etal/></person-group><article-title>CCR2 signal facilitates thymic egress by priming thymocyte responses to sphingosine-1-phosphate</article-title><source>Front. Immunol.</source><year>2018</year><volume>9</volume><fpage>1263</fpage><pub-id pub-id-type="pmid">29930553</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.3389/fimmu.2018.01263</pub-id><pub-id pub-id-type="pmcid">PMC6001116</pub-id></element-citation></ref><ref id="CR30"><label>30.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Morita</surname><given-names>H</given-names></name><etal/></person-group><article-title>Induction of human regulatory innate lymphoid cells from group 2 innate lymphoid cells by retinoic acid</article-title><source>J. Allergy Clin. Immunol.</source><year>2019</year><volume>143</volume><fpage>2190</fpage><lpage>2201</lpage><pub-id pub-id-type="pmid">30682454</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.jaci.2018.12.1018</pub-id></element-citation></ref><ref id="CR31"><label>31.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Salimi</surname><given-names>M</given-names></name><etal/></person-group><article-title>A role for IL-25 and IL-33–driven type-2 innate lymphoid cells in atopic dermatitis</article-title><source>J. Exp. Med.</source><year>2013</year><volume>210</volume><fpage>2939</fpage><pub-id pub-id-type="pmid">24323357</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1084/jem.20130351</pub-id><pub-id pub-id-type="pmcid">PMC3865470</pub-id></element-citation></ref><ref id="CR32"><label>32.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Maric</surname><given-names>J</given-names></name><etal/></person-group><article-title>Prostaglandin E 2 suppresses human group 2 innate lymphoid cell function</article-title><source>J. Allergy Clin. Immunol</source><year>2017</year><volume>141</volume><fpage>1761</fpage><lpage>1773</lpage><pub-id pub-id-type="pmid">29217133</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.jaci.2017.09.050</pub-id><pub-id pub-id-type="pmcid">PMC5929462</pub-id></element-citation></ref><ref id="CR33"><label>33.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Yang</surname><given-names>YH</given-names></name><etal/></person-group><article-title>Deficiency of annexin A1 in CD4+ T cells exacerbates T cell-dependent inflammation</article-title><source>J. Immunol.</source><year>2013</year><volume>190</volume><fpage>997</fpage><lpage>1007</lpage><pub-id pub-id-type="pmid">23267026</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.4049/jimmunol.1202236</pub-id></element-citation></ref><ref id="CR34"><label>34.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Batyrova</surname><given-names>B</given-names></name><etal/></person-group><article-title>PD-1 expression affects cytokine production by ILC2 and is influenced by peroxisome proliferator-activated receptor-γ</article-title><source>Immun. Inflamm. Dis.</source><year>2020</year><volume>8</volume><fpage>8</fpage><lpage>23</lpage><pub-id pub-id-type="pmid">31742928</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1002/iid3.279</pub-id><pub-id pub-id-type="pmcid">PMC7016838</pub-id></element-citation></ref><ref id="CR35"><label>35.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Wojno</surname><given-names>EDT</given-names></name><etal/></person-group><article-title>The prostaglandin D<sub>2</sub> receptor CRTH2 regulates accumulation of group 2 innate lymphoid cells in the inflamed lung</article-title><source>Mucosal Immunol.</source><year>2015</year><volume>8</volume><fpage>1313</fpage><lpage>1323</lpage><pub-id pub-id-type="pmid">25850654</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/mi.2015.21</pub-id><pub-id pub-id-type="pmcid">PMC4598246</pub-id></element-citation></ref><ref id="CR36"><label>36.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Huang</surname><given-names>Y</given-names></name><etal/></person-group><article-title>S1P-dependent interorgan trafficking of group 2 innate lymphoid cells supports host defense</article-title><source>Science</source><year>2018</year><volume>359</volume><fpage>114</fpage><lpage>119</lpage><pub-id pub-id-type="pmid">29302015</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1126/science.aam5809</pub-id><pub-id pub-id-type="pmcid">PMC6956613</pub-id></element-citation></ref><ref id="CR37"><label>37.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Winkler</surname><given-names>C</given-names></name><etal/></person-group><article-title>2 innate lymphoid cells after allergen challenge in asthmatic patients</article-title><source>J. Allergy Clin. Immunol.</source><year>2019</year><volume>144</volume><fpage>61</fpage><lpage>69</lpage><pub-id pub-id-type="pmid">30731124</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.jaci.2019.01.027</pub-id></element-citation></ref><ref id="CR38"><label>38.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Maric</surname><given-names>J</given-names></name><etal/></person-group><article-title>Cytokine-induced endogenous production of prostaglandin D2 is essential for human group 2 innate lymphoid cell activation</article-title><source>J. Allergy Clin. Immunol.</source><year>2019</year><volume>143</volume><fpage>2202</fpage><lpage>2214</lpage><pub-id pub-id-type="pmid">30578872</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.jaci.2018.10.069</pub-id></element-citation></ref><ref id="CR39"><label>39.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Roan</surname><given-names>F</given-names></name><etal/></person-group><article-title>CD4<sup>+</sup> group 1 innate lymphoid cells (ILC) form a functionally distinct ILC subset that is increased in systemic sclerosis</article-title><source>J. Immunol.</source><year>2016</year><volume>196</volume><fpage>2051</fpage><lpage>2062</lpage><pub-id pub-id-type="pmid">26826243</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.4049/jimmunol.1501491</pub-id><pub-id pub-id-type="pmcid">PMC4761490</pub-id></element-citation></ref><ref id="CR40"><label>40.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Bolotin</surname><given-names>DA</given-names></name><etal/></person-group><article-title>MiXCR: software for comprehensive adaptive immunity profiling</article-title><source>Nat. Methods</source><year>2015</year><volume>12</volume><fpage>380</fpage><lpage>381</lpage><pub-id pub-id-type="pmid">25924071</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nmeth.3364</pub-id></element-citation></ref><ref id="CR41"><label>41.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Acosta-Rodriguez</surname><given-names>EV</given-names></name><etal/></person-group><article-title>Surface phenotype and antigenic specificity of human interleukin 17-producing T helper memory cells</article-title><source>Nat. Immunol.</source><year>2007</year><volume>8</volume><fpage>639</fpage><lpage>646</lpage><pub-id pub-id-type="pmid">17486092</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/ni1467</pub-id></element-citation></ref><ref id="CR42"><label>42.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Gasteiger</surname><given-names>G</given-names></name><name name-style="western"><surname>Fan</surname><given-names>X</given-names></name><name name-style="western"><surname>Dikiy</surname><given-names>S</given-names></name><name name-style="western"><surname>Lee</surname><given-names>SY</given-names></name><name name-style="western"><surname>Rudensky</surname><given-names>AY</given-names></name></person-group><article-title>Tissue residency of innate lymphoid cells in lymphoid and nonlymphoid organs</article-title><source>Science</source><year>2015</year><volume>350</volume><fpage>981</fpage><lpage>985</lpage><pub-id pub-id-type="pmid">26472762</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1126/science.aac9593</pub-id><pub-id pub-id-type="pmcid">PMC4720139</pub-id></element-citation></ref><ref id="CR43"><label>43.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Krämer</surname><given-names>B</given-names></name><etal/></person-group><article-title>Compartment-specific distribution of human intestinal innate lymphoid cells is altered in HIV patients under effective therapy</article-title><source>PLoS Pathog.</source><year>2017</year><volume>13</volume><fpage>e1006373</fpage><pub-id pub-id-type="pmid">28505204</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1371/journal.ppat.1006373</pub-id><pub-id pub-id-type="pmcid">PMC5444854</pub-id></element-citation></ref><ref id="CR44"><label>44.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Huang</surname><given-names>Y</given-names></name><etal/></person-group><article-title>IL-25-responsive, lineage-negative KLRG1hi cells are multipotential ‘inflammatory’ type 2 innate lymphoid cells</article-title><source>Nat. Immunol.</source><year>2014</year><volume>16</volume><fpage>161</fpage><lpage>169</lpage><pub-id pub-id-type="pmid">25531830</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/ni.3078</pub-id><pub-id pub-id-type="pmcid">PMC4297567</pub-id></element-citation></ref><ref id="CR45"><label>45.</label><mixed-citation publication-type="other">Liu, S. et al. Optimal identification of human conventional and nonconventional (CRTH2–IL7Rα–) ILC2s using additional surface markers. <italic toggle="yes">J. Allergy Clin. Immunol</italic>. 10.1016/J.JACI.2020.01.038 (2020).<pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.jaci.2020.01.038</pub-id><pub-id pub-id-type="pmcid">PMC7398840</pub-id><pub-id pub-id-type="pmid">32032632</pub-id></mixed-citation></ref><ref id="CR46"><label>46.</label><mixed-citation publication-type="other">Novartis Pharmaceuticals. Study of efficacy and safety of QAW039 in patients with severe asthma inadequately controlled with standard of care asthma treatment. Full text view—ClinicalTrials.gov. <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="https://clinicaltrials.gov/ct2/show/NCT02563067">https://clinicaltrials.gov/ct2/show/NCT02563067</ext-link> (2019).</mixed-citation></ref><ref id="CR47"><label>47.</label><mixed-citation publication-type="other">Novartis Pharmaceuticals. Study of efficacy and safety of QAW039 in patients with severe asthma inadequately controlled with standard of care asthma treatment. Full text view—ClinicalTrials.gov. <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="https://clinicaltrials.gov/ct2/show/NCT02555683">https://clinicaltrials.gov/ct2/show/NCT02555683</ext-link> (2019).</mixed-citation></ref><ref id="CR48"><label>48.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Simoni</surname><given-names>Y</given-names></name><etal/></person-group><article-title>Human innate lymphoid cell subsets possess tissue-type based heterogeneity in phenotype and frequency</article-title><source>Immunity</source><year>2017</year><volume>46</volume><fpage>148</fpage><lpage>161</lpage><pub-id pub-id-type="pmid">27986455</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2016.11.005</pub-id><pub-id pub-id-type="pmcid">PMC7612935</pub-id></element-citation></ref><ref id="CR49"><label>49.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Simoni</surname><given-names>Y</given-names></name><name name-style="western"><surname>Newell</surname><given-names>EW</given-names></name></person-group><article-title>Toward meaningful definitions of innate-lymphoid-cell subsets</article-title><source>Immunity</source><year>2017</year><volume>46</volume><fpage>760</fpage><lpage>761</lpage><pub-id pub-id-type="pmid">28514678</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2017.04.026</pub-id></element-citation></ref><ref id="CR50"><label>50.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Bernink</surname><given-names>JH</given-names></name><name name-style="western"><surname>Mjösberg</surname><given-names>J</given-names></name><name name-style="western"><surname>Spits</surname><given-names>H</given-names></name></person-group><article-title>Human ILC1: to be or not to be</article-title><source>Immunity</source><year>2017</year><volume>46</volume><fpage>756</fpage><lpage>757</lpage><pub-id pub-id-type="pmid">28514676</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2017.05.001</pub-id><pub-id pub-id-type="pmcid">PMC5441993</pub-id></element-citation></ref><ref id="CR51"><label>51.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Roan</surname><given-names>F</given-names></name><name name-style="western"><surname>Ziegler</surname><given-names>SF</given-names></name></person-group><article-title>Human group 1 innate lymphocytes are negative for surface CD3ε but express CD5</article-title><source>Immunity</source><year>2017</year><volume>46</volume><fpage>758</fpage><lpage>759</lpage><pub-id pub-id-type="pmid">28514677</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2017.04.024</pub-id><pub-id pub-id-type="pmcid">PMC5503209</pub-id></element-citation></ref><ref id="CR52"><label>52.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Bernink</surname><given-names>JH</given-names></name><etal/></person-group><article-title>Human type 1 innate lymphoid cells accumulate in inflamed mucosal tissues</article-title><source>Nat. Immunol.</source><year>2013</year><volume>14</volume><fpage>221</fpage><lpage>229</lpage><pub-id pub-id-type="pmid">23334791</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/ni.2534</pub-id></element-citation></ref><ref id="CR53"><label>53.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Bernink</surname><given-names>JH</given-names></name><etal/></person-group><article-title>Interleukin-12 and -23 control plasticity of CD127+ group 1 and group 3 innate lymphoid cells in the intestinal lamina propria</article-title><source>Immunity</source><year>2015</year><volume>43</volume><fpage>146</fpage><lpage>160</lpage><pub-id pub-id-type="pmid">26187413</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.immuni.2015.06.019</pub-id></element-citation></ref><ref id="CR54"><label>54.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Wang</surname><given-names>B</given-names></name><etal/></person-group><article-title>Over-expression of CD3ε transgenes blocks T lymphocyte development</article-title><source>Int. Immunol.</source><year>1995</year><volume>7</volume><fpage>435</fpage><lpage>448</lpage><pub-id pub-id-type="pmid">7794823</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1093/intimm/7.3.435</pub-id></element-citation></ref><ref id="CR55"><label>55.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Phillips</surname><given-names>JH</given-names></name><etal/></person-group><article-title>Ontogeny of human natural killer (NK) cells: fetal NK cells mediate cytolytic function and express cytoplasmic CD3 epsilon,delta proteins</article-title><source>J. Exp. Med.</source><year>1992</year><volume>175</volume><fpage>1055</fpage><lpage>1066</lpage><pub-id pub-id-type="pmid">1372642</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1084/jem.175.4.1055</pub-id><pub-id pub-id-type="pmcid">PMC2119193</pub-id></element-citation></ref><ref id="CR56"><label>56.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Nagasawa</surname><given-names>M</given-names></name><name name-style="western"><surname>Germar</surname><given-names>K</given-names></name><name name-style="western"><surname>Blom</surname><given-names>B</given-names></name><name name-style="western"><surname>Spits</surname><given-names>H</given-names></name></person-group><article-title>Human CD5+ innate lymphoid cells are functionally immature and their development from CD34+ progenitor cells is regulated by Id2</article-title><source>Front. Immunol.</source><year>2017</year><volume>8</volume><fpage>1047</fpage><pub-id pub-id-type="pmid">28912776</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.3389/fimmu.2017.01047</pub-id><pub-id pub-id-type="pmcid">PMC5583608</pub-id></element-citation></ref><ref id="CR57"><label>57.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Borghesi</surname><given-names>L</given-names></name><etal/></person-group><article-title>B lineage-specific regulation of V(D)J recombinase activity is established in common lymphoid progenitors</article-title><source>J. Exp. Med.</source><year>2004</year><volume>199</volume><fpage>491</fpage><lpage>502</lpage><pub-id pub-id-type="pmid">14769852</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1084/jem.20031800</pub-id><pub-id pub-id-type="pmcid">PMC2211824</pub-id></element-citation></ref><ref id="CR58"><label>58.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Fronková</surname><given-names>E</given-names></name><etal/></person-group><article-title>Lymphoid differentiation pathways can be traced by TCR delta rearrangements</article-title><source>J. Immunol.</source><year>2005</year><volume>175</volume><fpage>2495</fpage><lpage>2500</lpage><pub-id pub-id-type="pmid">16081821</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.4049/jimmunol.175.4.2495</pub-id></element-citation></ref><ref id="CR59"><label>59.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Ravindran</surname><given-names>A</given-names></name><etal/></person-group><article-title>An optimized protocol for the isolation and functional analysis of human lung mast cells</article-title><source>Front. Immunol.</source><year>2018</year><volume>9</volume><fpage>2193</fpage><pub-id pub-id-type="pmid">30344519</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.3389/fimmu.2018.02193</pub-id><pub-id pub-id-type="pmcid">PMC6183502</pub-id></element-citation></ref><ref id="CR60"><label>60.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Picelli</surname><given-names>S</given-names></name><etal/></person-group><article-title>Smart-seq2 for sensitive full-length transcriptome profiling in single cells</article-title><source>Nat. Methods</source><year>2013</year><volume>10</volume><fpage>1096</fpage><lpage>1098</lpage><pub-id pub-id-type="pmid">24056875</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nmeth.2639</pub-id></element-citation></ref><ref id="CR61"><label>61.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Dobin</surname><given-names>A</given-names></name><etal/></person-group><article-title>STAR: ultrafast universal RNA-seq aligner</article-title><source>Bioinformatics</source><year>2013</year><volume>29</volume><fpage>15</fpage><lpage>21</lpage><pub-id pub-id-type="pmid">23104886</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1093/bioinformatics/bts635</pub-id><pub-id pub-id-type="pmcid">PMC3530905</pub-id></element-citation></ref><ref id="CR62"><label>62.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Ramsköld</surname><given-names>D</given-names></name><name name-style="western"><surname>Wang</surname><given-names>ET</given-names></name><name name-style="western"><surname>Burge</surname><given-names>CB</given-names></name><name name-style="western"><surname>Sandberg</surname><given-names>R</given-names></name></person-group><article-title>An abundance of ubiquitously expressed genes revealed by tissue transcriptome sequence data</article-title><source>PLoS Comput. Biol.</source><year>2009</year><volume>5</volume><fpage>e1000598</fpage><pub-id pub-id-type="pmid">20011106</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1371/journal.pcbi.1000598</pub-id><pub-id pub-id-type="pmcid">PMC2781110</pub-id></element-citation></ref><ref id="CR63"><label>63.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Stuart</surname><given-names>T</given-names></name><etal/></person-group><article-title>Comprehensive integration of single-<italic toggle="yes">cell data</italic></article-title><source>Cell</source><year>2019</year><volume>177</volume><fpage>1888</fpage><lpage>1902</lpage><pub-id pub-id-type="pmid">31178118</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1016/j.cell.2019.05.031</pub-id><pub-id pub-id-type="pmcid">PMC6687398</pub-id></element-citation></ref><ref id="CR64"><label>64.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Lun</surname><given-names>ATL</given-names></name><name name-style="western"><surname>McCarthy</surname><given-names>DJ</given-names></name><name name-style="western"><surname>Marioni</surname><given-names>JC</given-names></name></person-group><article-title>A step-by-step workflow for low-level analysis of single-cell RNA-seq data with Bioconductor</article-title><source>F1000Research</source><year>2016</year><volume>5</volume><fpage>2122</fpage><pub-id pub-id-type="pmid">27909575</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.12688/f1000research.9501.1</pub-id><pub-id pub-id-type="pmcid">PMC5112579</pub-id></element-citation></ref><ref id="CR65"><label>65.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Arya</surname><given-names>S</given-names></name><name name-style="western"><surname>Mount</surname><given-names>DM</given-names></name><name name-style="western"><surname>Netanyahu</surname><given-names>NS</given-names></name><name name-style="western"><surname>Silverman</surname><given-names>R</given-names></name><name name-style="western"><surname>Wu</surname><given-names>AY</given-names></name></person-group><article-title>An optimal algorithm for approximate nearest neighbor searching fixed dimensions</article-title><source>J. ACM</source><year>1998</year><volume>45</volume><fpage>891</fpage><lpage>923</lpage></element-citation></ref><ref id="CR66"><label>66.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Finak</surname><given-names>G</given-names></name><etal/></person-group><article-title>MAST: a flexible statistical framework for assessing transcriptional changes and characterizing heterogeneity in single-cell RNA sequencing data</article-title><source>Genome Biol.</source><year>2015</year><volume>16</volume><fpage>278</fpage><pub-id pub-id-type="pmid">26653891</pub-id><pub-id pub-id-type="doi" assigning-authority="pmc">10.1186/s13059-015-0844-5</pub-id><pub-id pub-id-type="pmcid">PMC4676162</pub-id></element-citation></ref><ref id="CR67"><label>67.</label><element-citation publication-type="journal"><person-group person-group-type="author"><name name-style="western"><surname>Langfelder</surname><given-names>P</given-names></name><name name-style="western"><surname>Horvath</surname><given-names>S</given-names></name></person-group><article-title>WGCNA: an R package for weighted correlation network analysis</article-title><source>BMC Bioinform.</source><year>2008</year><volume>9</volume><fpage>559</fpage><pub-id pub-id-type="doi" assigning-authority="pmc">10.1186/1471-2105-9-559</pub-id><pub-id pub-id-type="pmcid">PMC2631488</pub-id><pub-id pub-id-type="pmid">19114008</pub-id></element-citation></ref></ref-list></back></article>
\ No newline at end of file