Dendritic cells (DCs) are critical for pathogen recognition and Ag processing/presentation. Human monocyte-derived DCs (moDCs) have been extensively used in experimental studies and DC-based immunotherapy approaches. However, the extent of human moDC and peripheral DCs heterogeneity and their interrelationship remain elusive. In this study, we performed single-cell RNA sequencing of human moDCs and blood DCs. We identified seven subtypes within moDCs: five corresponded to type 2 conventional DCs (cDC2s), and the other two were CLEC10A+CD127+ cells with no resemblance to any peripheral DC subpopulations characterized to date. Moreover, we defined five similar subtypes in human cDC2s, revealed the potential differentiation trajectory among them, and unveiled the transcriptomic differences between moDCs and cDC2s. We further studied the transcriptomic changes of each moDC subtype during maturation, demonstrating SLAMF7 and IL15RA as maturation markers and CLEC10A and SIGLEC10 as markers for immature DCs. These findings will enable more accurate functional/developmental analyses of human cDC2s and moDCs.

Visual Abstract

Dendritic cells (DCs), found in the blood, lymphoid organs, and all of the other body tissues, “ingest” endogenous Ags and pathogens and present processed epitopes to T cells, B cells, and NK cells (1). DCs serve as the bridge between the innate and adaptive immune systems and are critical in the initiation of primary immune responses. Human DCs comprise multiple subtypes with unique transcriptomic features and biological functions, such as plasmacytoid DCs, conventional DCs (cDCs), and Langerhans cells (2). Additionally, cDCs, defined by the expression of ITGAX (CD11c) and MHC class II (MHCII) molecules consist of two subsets, CD1C cDC1s and CD1C+ cDC2s. Our conceptualization of DC is recurrently updated, mostly driven by the development of new technologies such as lineage-tracing and single-cell transcriptomics.

The classification of human DCs was extended to a more accurate taxonomy by Villani et al. (3). This study revealed six DC subtypes (DC1–DC6); of note, two of them, DC2 and DC3, were linked to cDC2s based on defined marker genes (3). Although DC1s and DC6s correspond to cDC1s and plasmacytoid DCs, respectively, some researchers have indicated that DC5s comprise multiple types of cells, including precursors of DCs, and some DC2s (4). In addition, the DC4 subtype was reclassified as CD16+ nonconventional monocytes. More recently, researchers have proposed a different cDC2 classification strategy in both mice and humans, based on the transcription regulators T-bet and RORγt (5). Some researchers also indicated that all cDC2s should express CLEC10A, a previously defined marker gene for the cDC2 subtype DC2 (6, 7). A study based on high-dimensional protein and RNA single-cell analyses revealed the presence of a CD1clow CD14+ cDC2 subtype and identified another set of surface markers for DC2s and DC3s (4). The heterogeneity of DCs is still unclear, especially considering cDC2s. Therefore, it is essential to explore in detail DC heterogeneity and to provide clear markers for each identified subpopulation, which could benefit further functional and developmental analyses.

Human monocyte-derived DCs (moDCs) are DCs generated from monocytes via incubation with GM-CSF and IL-4 (8). Functionally and phenotypically, moDCs are believed to be typical immature DCs, characterized by their low expression of MHCII and costimulatory molecules. Immature moDCs can be subsequently matured after treatment with compounds known to induce DC maturation, such as LPS, TNF-α, INF-γ, or CD40L (9). Importantly, because of the scarcity of DC populations in the peripheral blood, moDCs have been the most widely used models to investigate human DC biology and function since their discovery in the 1990s (8, 10, 11). Moreover, moDCs are also currently used in most DC-based immunotherapeutic treatments for cancers and autoimmune diseases. However, little is known about the heterogeneity of moDCs, as well as with respect to the differences between them and peripheral DCs.

In this study, we studied the transcriptomes of moDCs and human peripheral DCs using single-cell RNA sequencing (scRNA-seq) to unveil their heterogeneity. Importantly we identified seven subtypes and further characterized the transcriptomic changes during their maturation. Altogether, our data support the revision of the current classification of cDC2s, show transcriptomic differences between moDCs and cDC2s, and study the transcriptomic changes of moDCs during maturation.

For moDC generation, we first used buffy coats of healthy blood donors at the Zhongshan Ophthalmic Center. Healthy donors were above 23 y of age. Ficoll-Paque density gradient centrifugation was used to isolate PBMCs. A Dynabeads Untouched Human Monocytes Kit was used to separate untouched monocytes from PBMCs. To generate moDCs, we plated monocytes at 1 × 106 cells/ml in RPMI 1640 (Life Technologies) supplemented with 10% FBS (Life Technologies), GlutaMAX (Life Technologies), penicillin–streptomycin (Life Technologies), human GM-CSF (100 ng/ml; PeproTech), and human IL-4 (20 ng/ml; GlutaMAX). Cells were then incubated for 6 d at 37°C and 5% CO2; IL-4 and GM-CSF were replenished on day 3. Maturation was induced by a 24-h incubation with 1 µg/ml LPS (Sigma-Aldrich). We sorted single cells chilled to 4°C and preprepared lysis buffer consisting of 10 ml of Buffer TCL (QIAGEN) supplemented with 1% 2-ME. Single-cell lysates were sealed, vortexed, spun down at 300 × g at 4°C for 1 min, immediately placed on dry ice, and transferred for storage at −80°C.

We isolated PBMCs from fresh blood using Ficoll-Paque density gradient centrifugation, as previously described. We performed flow cytometry and FACS sorting of both PBMCs and generated moDCs on a BD LSRFortessa or BD FACSAria Fusion instrument, and data acquired were analyzed with FlowJo v10.1, as shown in the Results section. The single-cell suspension was prepared the same as moDCs.

Single-cell suspensions were stained after removal of the supernatant, the cell pellet was resuspended in 100 μl of 1× PBS, and the suspension was passed through a 35-μm cell strainer (352235; Falcon) to remove cellular aggregates, followed by cell counting using a disposable Neubauer chamber (DHC-N01; NanoEnTek). Single-cell RNA-seq libraries were prepared using the 10× Genomics Single Cell Immune Profiling Solution Kit, according to the manufacturer’s instructions. For gene expression library construction, 50 ng of amplified cDNA was fragmented and end repaired, double-size selected with SPRIselect beads (average size, 450 bp), and sequenced on Illumina NovaSeq 6000 platform using 150 paired-end reads.

Cell Ranger version 4.1.0 (10× Genomics) was used to process the raw scRNA-seq data. The transcripts were aligned to human reference genome hg38 (version 4.1.0; 10× Genomics). The following steps were done with Seurat version 3.1 (https://github.com/satijalab/seurat) on R version 4.0.0. Quality control steps included the percentage of mitochondrial transcripts smaller than 15%, and the number of identified features ranging between 200 and 4000 were filtered. After quality control filtering, the samples were then normalized and scaled. We performed dimension reduction clustering and differential expression analysis following the Seurat guided tutorial.

The integration process basically followed the Seurat “sctransform” tutorial with default parameters. We performed principal component analysis and uniform manifold approximation and projection (UMAP) dimension reduction with 50 principal components. Fifty dimensions of the principal component analysis reduction were used in the “FindNeighbors()” calculation, followed by clustering using “FindClusters()” with a resolution of 0.7.

We used Monocle 3 (https://github.com/cole-trapnell-lab/monocle3) to identify the pseudotime differentiation trajectory between different moDCs subsets. UMAP embeddings and cell clusters generated from Seurat were used as input. We performed graph learning and pseudotime measurement through reversed graph embedding following the guided tutorial.

Sequencing data are available in the National Genomic Data Center (primary accession number HRA000563; https://bigd.big.ac.cn/gsa-human/browse/HRA000563).

For MLR assay, we performed FACS sorting of moDCs to acquire undefined monocyte-derived cells (UMDCs) and moDCs without UMDCs. DC maturation was induced by a 24-h incubation with 1 µg/ml LPS (Sigma-Aldrich). Allogeneic CD4+ T cells from FACS sorting labeled with CFSE (BD Biosciences) were cultured with DCs at 10:1. T cell proliferation after 5 d was measured by flow cytometry.

The study was performed in accordance with protocols approved by the Ethics Committee of Zhongshan Ophthalmic Center. Blood was collected from six healthy individuals without history of inflammatory disease, cancer, allergies, or immune treatment. All donors were nonsmokers with a normal body mass index and normal blood pressure. All donors were asked for consent for genetic research.

For comparisons between expression values, the Seurat function “FindMarkers()” was used. Cell type markers were obtained using the “FindAllMarkers()” function with a negative binomial test. The p values shown in box plot figures are exact two-sided by Student t test using Graph Prism 8.

We sorted peripheral CD14+ monocytes, induced their differentiation into DCs with GM-CSF and IL-4, and collected moDCs for our study. Next, we performed scRNA-seq on the obtained DCs (8640 moDCs) and, finally, focused on 7617 moDCs with high-quality reads after the quality control process (see the Materials and Methods section) (Fig. 1A).

FIGURE 1.

Unbiased classification of moDCs subsets. (A) Research experimental design. Three types of unstimulated DCs were employed: moDCs induced from CD14+ monocytes, mature moDCs after incubation with LPS, and human DCs isolated from peripheral blood. We performed scRNA-seq on three types of DCs and studied their transcriptome profiling. (B) UMAP embedding of unstimulated moDCs colored by cell type annotation, including three subtypes within DC2 (DC2-1, DC2-2, and DC2-3), two subtypes within DC3 (DC3-1, DC3-2), and two new, to our knowledge, clusters of undefined moDC-1 (UMDC-1) and undefined moDC-2 (UMDC-2). (C) UMAP embedding shows the expression of conventional moDC and cDC2 markers. (D) Heatmap of DEGs for each cluster. Margin color bars highlight specific genes of respective unstimulated moDC subpopulations.

FIGURE 1.

Unbiased classification of moDCs subsets. (A) Research experimental design. Three types of unstimulated DCs were employed: moDCs induced from CD14+ monocytes, mature moDCs after incubation with LPS, and human DCs isolated from peripheral blood. We performed scRNA-seq on three types of DCs and studied their transcriptome profiling. (B) UMAP embedding of unstimulated moDCs colored by cell type annotation, including three subtypes within DC2 (DC2-1, DC2-2, and DC2-3), two subtypes within DC3 (DC3-1, DC3-2), and two new, to our knowledge, clusters of undefined moDC-1 (UMDC-1) and undefined moDC-2 (UMDC-2). (C) UMAP embedding shows the expression of conventional moDC and cDC2 markers. (D) Heatmap of DEGs for each cluster. Margin color bars highlight specific genes of respective unstimulated moDC subpopulations.

Close modal

We initially identified seven distinct clusters, as visualized by UMAP (Fig. 1B). All cells exhibited high levels of the moDC markers CD1A, CD1C, and MHCII molecules, as well as low levels of CD14 and FCGR3A, confirming our success in DC induction (Fig. 1C, Supplemental Fig. 1A). Of note, the upregulation of CD1C and costimulatory factors allowed the classification of different human cDC2s with similar CD1C and MHCII molecule expression profiles (Fig. 1C, Supplemental Fig. 1A). Interestingly, we distinguished clusters similar to DC2 and DC3 within moDCs based on the marker genes CLEC10A and S100A9, in line with previous studies (Fig. 1C, Supplemental Fig. 1B and 1C) (3). Remarkably, unsupervised clustering analysis allowed the extension of the current cDC2 classification; five subtypes were identified: three DC2-like moDC subtypes (DC2-1, DC2-2, and DC2-3) and two DC3-like moDC subtypes (DC3-1, DC3-2) (Fig. 1C, 1D). In addition, we found two new, to our knowledge, clusters, named undefined moDC-1 (UMDC-1) and undefined moDC-2 (UMDC-2) (Fig. 1C, 1D).

DC2-like moDCs, making up to ∼40% of all profiled moDCs, were subdivided into three subpopulations: DC2-1, DC2-2, and DC2-3 (Supplemental Fig. 1D). Based on unbiased clustering, we identified differential gene expression signatures for each DC2 cluster and perform functional analysis (Fig. 2A, 2B). The DC2-1 cluster exhibited higher expression of classical DC2 marker genes CLEC10A, SIGLEC10, and CD1A, whereas the DC2-2 cluster expressed more SPP1 and IL1R2 (Fig. 2A). Interestingly, except from higher levels of TNFRSF11A and RAB33A, the DC2-3 cluster expressed moderate levels of both DC2 and DC3 markers, suggesting that these cells might form a continuum between DC2s and DC3s (Fig. 2A, 2C).

FIGURE 2.

Subpopulations within unstimulated DC2-like and DC3-like moDCs. (A) Dot plot shows the DEGs in three respective subpopulations within unstimulated DC2-like moDCs. Violin plot shows the expression levels of candidate genes in each subpopulation. The upper/lower part shows the expression of marker/function–related genes for each subset. (B) Lollipop chart depicts gene ontology–term functional analysis by upregulated and downregulated biological progress in DC2-2. (C) Violin plots depict distinct expression of CLEC10A, SIGLEC10, CCL22, and FABP5 in all five subtypes within unstimulated DC2-like and DC3-like moDCs. (D) The same as (A) for two subpopulations within unstimulated DC3-like moDCs. (E) The same as (B) for upregulated biological progress in DC3-1 and DC3-2.

FIGURE 2.

Subpopulations within unstimulated DC2-like and DC3-like moDCs. (A) Dot plot shows the DEGs in three respective subpopulations within unstimulated DC2-like moDCs. Violin plot shows the expression levels of candidate genes in each subpopulation. The upper/lower part shows the expression of marker/function–related genes for each subset. (B) Lollipop chart depicts gene ontology–term functional analysis by upregulated and downregulated biological progress in DC2-2. (C) Violin plots depict distinct expression of CLEC10A, SIGLEC10, CCL22, and FABP5 in all five subtypes within unstimulated DC2-like and DC3-like moDCs. (D) The same as (A) for two subpopulations within unstimulated DC3-like moDCs. (E) The same as (B) for upregulated biological progress in DC3-1 and DC3-2.

Close modal

Comparing the potential functional properties of three clusters, the DC2-1 cluster was more functionally “conventional,” distinguished by the higher expression of MHCII genes and costimulatory factors (Fig. 2A, Supplemental Fig. 1E). We identified corresponding upregulation of endocytosis-associated categories in functional analysis, possibly related to Ag uptake (Fig. 2A). Additionally, DC2-2 cells were characterized by the upregulation of proinflammatory cytokines and chemokines (CCL2, CCL3, CCL4, and CSF1) (Fig. 2A, Supplemental Fig. 1F). Of note, functional analysis revealed that the upregulated genes were associated with cell migration, recruitment, and protein secretion categories, suggesting a unique regulatory and secretory role of the DC2-2 subtype (Fig. 2B).

As mentioned above, DC3-like moDCs, comprising 35% of moDCs, consisted of two subpopulations (Supplemental Fig. 1D). Interestingly, these two subpopulations exhibited a very distinct expression pattern of ANXA1, a previously defined DC3 marker gene related to immune regulation (Supplemental Fig. 1G) (12). With lower ANXA1 expression, the DC3-1 cluster exhibited higher levels of markers, including S100A8, S100A9, CD81, and CD1C (Fig. 2D). Functionally, the DC3-1 cluster was characterized by the higher expression of function-related genes (CD40, CD74, and MRC1), chemokines (CCL17, and CCL22), and proinflammatory cytokines (IL18) (Fig. 2D). On the contrary, the DC3-2 cluster was marked by a higher level of marker ANXA1. Of note, based on the expression of ANXA1 and IL18 alone, we were able to clearly distinguish the two DC3 clusters (Supplemental Fig. 1H). The DC3-2 cluster expressed high levels of cell cycle-related genes, including JUN, SOD2, EMP3, and KLF6 (Fig. 2D). Interestingly, functional analysis revealed that the DC3-2 cluster was highly enriched in apoptosis-related categories, suggesting its terminal state, whereas the DC3-1 cluster was more associated with categories involving DC function, including Ag processing and presentation, and leukocyte migration (Fig. 2E).

Interestingly, CLEC10A, CCL22, SIGLEC10, and FABP5 exhibited gradual and progressive upregulation or downregulation along five cDC2-like subtypes (DC2-1, DC2-2, DC2-3, DC3-1, and DC3-2; in this order), suggesting an underlying differentiation or transformation pathway in these cDC2-like cells (Fig. 2C).Furthermore, trajectory analysis supported this hypothesis, identifying DC2-1 as the initiation point, DC2-3 as the transitional state, and DC3-2 as the terminal state of the pseudotime differentiation trajectory (Supplemental Fig. 1I).

Next, we turned to the two newly discovered clusters constituting 25% of all sampled moDCs: smaller UMDC-1 and larger UMDC-2 (Supplemental Fig. 1D). Compared with DC2s and DC3s, these cells were characterized by the upregulation of CLEC10A and of IL7R (CD127) (13) and exhibited no resemblance to any known DC subtypes (Fig. 3A, Supplemental Fig. 2A). For validation, we performed flow cytometry on moDCs, identified three clusters based on CD127 (encoded by IL7R) and CLEC10A, and revealed the presence of CLEC10A+CD127+ UMDCs (Fig. 3B). As for potential function, we observed the upregulation of TGFBI, MRC1, and CCL22 as well as of the regulatory genes FGL2, and MMP12, suggesting these cells have potential regulatory functions (Fig. 3A) (14, 15). Functional enrichment analysis highlighted upregulated biological processes involved in endocytosis, including phagosome, vesicle-mediated transport, positive regulation of endocytosis, and receptor-mediated endocytosis (Supplemental Fig. 2B).

FIGURE 3.

Characterization of unstimulated UMDC subpopulations. (A) Heatmap shows the DEGs of DC2-like and DC3-like moDCs and UMDC. Violin plot shows the expression levels of representative genes in three subpopulations. (B) Flow cytometry gating strategy to identify moDCs based on CD127 and CLEC10A. Violin plots are same as (A) for CD127 and CLEC10A. (C) Venn diagram shows upregulated DEGs in two UMDC subpopulations compared with five cDC2-like clusters. The counts show the number of DEGs. (D) Heatmap of upregulated DEGs in two UMDC subpopulations. Margin color bars indicate three types of genes: commonly upregulated in UMDC1 and UMDC2, specifically upregulated in UMDC1, and specifically upregulated in UMDC2. (E) Proliferation of allogeneic T cells after coculture with UMDCs (CD127+CLEC10A+) purified within moDCs and moDCs without UMDCs. Bar plots on the right show the composite data. n = 3; mean ± SD; p < 0.05 by Student t test.

FIGURE 3.

Characterization of unstimulated UMDC subpopulations. (A) Heatmap shows the DEGs of DC2-like and DC3-like moDCs and UMDC. Violin plot shows the expression levels of representative genes in three subpopulations. (B) Flow cytometry gating strategy to identify moDCs based on CD127 and CLEC10A. Violin plots are same as (A) for CD127 and CLEC10A. (C) Venn diagram shows upregulated DEGs in two UMDC subpopulations compared with five cDC2-like clusters. The counts show the number of DEGs. (D) Heatmap of upregulated DEGs in two UMDC subpopulations. Margin color bars indicate three types of genes: commonly upregulated in UMDC1 and UMDC2, specifically upregulated in UMDC1, and specifically upregulated in UMDC2. (E) Proliferation of allogeneic T cells after coculture with UMDCs (CD127+CLEC10A+) purified within moDCs and moDCs without UMDCs. Bar plots on the right show the composite data. n = 3; mean ± SD; p < 0.05 by Student t test.

Close modal

In addition, the smaller UMDC-1 cluster exhibited the clear upregulation of genes encoding chemokines associated with immune tolerance, including CCL8, CCL18, as well as the suppressive marker CD274 (Fig. 3C, 3D; Supplemental Fig. 2C) (1618). We identified the largest proportion of spliced mRNAs in the UMDC-1 cluster, indicating high transcriptional activity (Supplemental Fig. 2D). In the context of the larger UMDC-2 cluster, we found the upregulation of FGL2 and CCL22 and no change in the expression of chemokines associated with the UMDC-1 cluster (Figure 3D, Supplemental Fig. 2C). Moreover, we validated the potential function of UMDCs by culturing UMDCs, as well as moDCs without UMDCs with allogenic T cells, and found that UMDCs exhibited weaker T cell activation properties (Fig. 3E).

We further performed the analysis of transcription factors (TFs) using DoRothEA and reclustered our moDCs based on TF activity. Dimensionality reduction clearly distinguished five cDC2-like clusters and two newly discovered, to our knowledge, clusters, indicating distinct TF activity in line with the above-reported transcriptome differences (Supplemental Fig. 2E, 2F).

Next, we sorted DCs from the human peripheral blood, performed scRNA-seq, and separated CD1C+ cDC2s (Fig. 4A). Importantly, unbiased clustering also revealed five similar subpopulations, although with slightly weaker expression changes of marker genes compared with those observed for moDCs (Fig. 4B; Supplemental Fig. 2G, 2H). After integration, peripheral DCs were distributed relatively evenly across all five subpopulations, suggesting more complex heterogeneity than that previously addressed in peripheral cDC2s (Fig. 4C, 4D). Also, the functional analysis results were similar. Peripheral DCs include conventional DC2-1 and DC3-1, migration-associated DC2-2, transitional DC2-3, and DC3-2 in the terminal state (Supplemental Figs. 2I, 3A). Importantly, the integration process not only indicated five different subpopulations of cDC2s but also confirmed the two newly identified UMDC clusters not previously characterized in the human peripheral blood (Fig. 4E, 4F).

FIGURE 4.

Integration of unstimulated moDCs and human peripheral cDC2s. (A) Flow cytometry gating strategy to identify DCs within human blood. (B) UMAP embedding of human peripheral cDC2s colored by cell types annotation. (C) UMAP embedding colored by origins shows the integration with unstimulated moDCs and human cDC2s. (D) The same as (C) colored by cell types annotation. (E) Percentage of cells from each cluster derived from unstimulated moDCs (red) or cDC2s (blue). (F) UMAP embedding depicts expression level of CD1C, CD86, HLA-DRB1, HLA-DRB5, CLEC10A, S100A9, ANXA1, and CCL3 after integration, comparing moDCs with human peripheral DCs (cDC2).

FIGURE 4.

Integration of unstimulated moDCs and human peripheral cDC2s. (A) Flow cytometry gating strategy to identify DCs within human blood. (B) UMAP embedding of human peripheral cDC2s colored by cell types annotation. (C) UMAP embedding colored by origins shows the integration with unstimulated moDCs and human cDC2s. (D) The same as (C) colored by cell types annotation. (E) Percentage of cells from each cluster derived from unstimulated moDCs (red) or cDC2s (blue). (F) UMAP embedding depicts expression level of CD1C, CD86, HLA-DRB1, HLA-DRB5, CLEC10A, S100A9, ANXA1, and CCL3 after integration, comparing moDCs with human peripheral DCs (cDC2).

Close modal

Next, we studied the complex transcriptomic differences between moDCs and peripheral cDC2s. Of note, the numbers of differentially expressed genes (DEGs) indicated that all five cDC2-like subpopulations exhibited relatively average and evident differences (Supplemental Fig. 3B). To better characterize the transcriptomic difference between moDCs and peripheral cDC2s, we identified 584/476 genes commonly upregulated or downregulated across the five subtypes and performed functional enrichment analysis (Supplemental Fig. 3C, 3D). Importantly, we not only found the enrichment of oxidative phosphorylation and cellular response to external stimuli categories, but also of other unexpected ones involving immune regulation, including myeloid leukocyte activation, lysosome, and adaptive immune system categories (Supplemental Fig. 3D).

Among the moDCs, we identified the downregulation of MHCII molecules, C-type lectin superfamily members, and TLRs, as well as the upregulation of C1Q production in all subpopulations (Fig. 5A). Of note, protein–protein interaction analysis based on gene expression also revealed upregulated endocytosis and complement production, as well as downregulated Ag presentation and migration capacity. The upregulation of the phagosome, clathrin-mediated endocytosis, and classical Ab-mediated complement activation categories, as well as the downregulation of the Ag processing and presentation, proteasome, and CXCR4 pathway categories were detected in moDCs versus peripheral DCs (Supplemental Fig. 3E, 3F).

FIGURE 5.

Transcriptomic differences between unstimulated moDC and peripheral cDC2. (A) Dot plots show the expression levels of DEGs from each cluster in unstimulated moDC and peripheral cDC2. Significant changes for each gene of moDC versus cDC2 in five same-cell subpopulations is marked by a red circle. Margin color bars indicate six types of genes. (B) The expression of CD127 (encoded by IL7R) and CLEC10A in peripheral cDC2s and unstimulated moDCs. Bar plots show the percentage of CD127+ or CLEC10A+ cells. n = 3; mean ± SD; p < 0.05 by Student t test. (C) Table and violin plot depict distinct expression of leukocyte-differentiation Ag in five cell subpopulations of unstimulated moDC peripheral and cDC2, respectively. Genes are assigned into five function-associated groups.

FIGURE 5.

Transcriptomic differences between unstimulated moDC and peripheral cDC2. (A) Dot plots show the expression levels of DEGs from each cluster in unstimulated moDC and peripheral cDC2. Significant changes for each gene of moDC versus cDC2 in five same-cell subpopulations is marked by a red circle. Margin color bars indicate six types of genes. (B) The expression of CD127 (encoded by IL7R) and CLEC10A in peripheral cDC2s and unstimulated moDCs. Bar plots show the percentage of CD127+ or CLEC10A+ cells. n = 3; mean ± SD; p < 0.05 by Student t test. (C) Table and violin plot depict distinct expression of leukocyte-differentiation Ag in five cell subpopulations of unstimulated moDC peripheral and cDC2, respectively. Genes are assigned into five function-associated groups.

Close modal

Regarding other functional features, the difference between these two types of cells became even more complex. Although moDCs exhibited slightly but firmly higher expression of the chemokines CCL17, CCL22, and CCL13 among all five subpopulations, the human DCs expressed more CCL3 and CXCL8 in DC2-2 and DC3-2 subtypes, as well as CXCL16 in all subpopulations (Fig. 5A, Supplemental Fig. 3G). We also identified upregulated cytokines IL1B, IL18, IL16, and TGFB1 in human DCs (Supplemental Fig. 3H). Additionally, regarding cytokine receptors, we found slightly higher expression of IL7R and CCR1 and lower expression of IL6R, CXCR4, TGFBR2, and IFNRs in moDCs (Supplemental Fig. 3I). Furthermore, we also validated the expression of IL7R (CD127) and CLEC10A in cDC2s and moDCs by flow cytometry, revealing higher levels of CD127 and lower levels of CLEC10A in moDCs (Fig. 5B).

We also observed distinct preferential expression of surface markers and other genes associated with DC function (Fig. 5C). Cells from the peripheral blood tended to express higher levels of CD1C (cDC2s marker), CD68 and CD33 (myeloid markers), and CD86, CD83, and CD58 (costimulatory factors) (19, 20). Conversely, moDCs expressed higher levels of CD40 (costimulatory factor), CD84 (autophagy associated), and CD1A and CD1B (marker genes) (Fig. 5C) (21). Interestingly, we also found the upregulated expression of CD274 (PD-L1) in moDCs and inhibitory CD47 in human cDC2s, further indicating a distinct gene expression pattern (Fig. 5C) (22, 23). In addition, regarding migration-related genes, moDCs preferentially expressed CD81, whereas human DCs expressed CD44, CD37, CD99, and CD302 (Fig. 5C) (2427). We also identified upregulation of complement regulation-associated genes CD46, CD52, and CD55 in human cDC2s (Fig. 5C) (28, 29).

Next, we investigated the transcriptomic changes in moDCs during maturation. We cultured immature moDCs with LPS for 24 h and performed scRNA-seq. Following the Seurat pipeline, we integrated immature and mature moDCs and separated them into three large clusters: DC2-like, DC3-like, and moDC-specific subtypes based on marker gene expression patterns (Fig. 6A–C, Supplemental Fig. 3J, 3K). The upregulation of genes coding for MHCII molecules, costimulatory factors, cytokines, and maturation markers, such as CCR7 and LAMP3, confirmed the successful maturation (Fig. 6D) (1, 9). Additionally, we also observed the robust upregulation of SLAMF7, IL15RA, MS4A7, and CD70 (Supplemental Fig. 4A). Notably, we also observed a clear downregulation of CLEC10A, the DC2 marker gene, and of SIGLEC10 in all moDCs during the maturation process, suggesting both as markers of DCs in the immature state (Fig. 6D). These two marker genes were further validated on the protein level with the flow cytometry results, indicating significant downregulation of CLEC10A and SIGLEC10 in activated moDCs, compared with immature moDCs (Fig. 6E).

FIGURE 6.

Distinct transcriptome during maturation in DCs. (A) UMAP embedding of immature and mature moDCs integration colored by three large clusters: DC2, DC3, and UMDC. (B) Dot plots show the expression levels of marker genes for DC2, DC3, and UMDC. (C) Violin plots depict distinct expression of marker genes S100A9, CLEC10A, CD1C, and TGFBI in DC2, DC3, and UMDC. (D) UMAP embedding depicts expression level of CCR7, LAMP3, CLEC10A, and SIGLEC10 comparing mature moDCs (MA) with immature moDCs (IM). (E) Flow cytometry gating strategy to identify MA and IM based on CLEC10A and SIGLEC10. (F and G) Heatmap of commonly upregulated (F) or downregulated (G) genes in maturation of DC2 and DC3. Genes were assigned into nine (F) and six (G) groups according to their functions and indicated by margin color bars. (H) Violin plots depict DEGs only upregulated or downregulated in DC2 during maturation. The p values and log n (fold change) (LnFC) only mentioned in text are showed. The p values are exact two-sided p values by Wilcoxon rank-sum test. (I) The same as (H) for upregulated genes in DC3.

FIGURE 6.

Distinct transcriptome during maturation in DCs. (A) UMAP embedding of immature and mature moDCs integration colored by three large clusters: DC2, DC3, and UMDC. (B) Dot plots show the expression levels of marker genes for DC2, DC3, and UMDC. (C) Violin plots depict distinct expression of marker genes S100A9, CLEC10A, CD1C, and TGFBI in DC2, DC3, and UMDC. (D) UMAP embedding depicts expression level of CCR7, LAMP3, CLEC10A, and SIGLEC10 comparing mature moDCs (MA) with immature moDCs (IM). (E) Flow cytometry gating strategy to identify MA and IM based on CLEC10A and SIGLEC10. (F and G) Heatmap of commonly upregulated (F) or downregulated (G) genes in maturation of DC2 and DC3. Genes were assigned into nine (F) and six (G) groups according to their functions and indicated by margin color bars. (H) Violin plots depict DEGs only upregulated or downregulated in DC2 during maturation. The p values and log n (fold change) (LnFC) only mentioned in text are showed. The p values are exact two-sided p values by Wilcoxon rank-sum test. (I) The same as (H) for upregulated genes in DC3.

Close modal

Furthermore, among the commonly upregulated genes, we found MHC class I genes, CD80 and CD86 (costimulatory factors); CXCL8, CCL2, and CCL1 (chemokines); TNF, IL12B, IL1B (cytokines); IFIT1, IFITM1, and ISG20 (IFN-induced genes); and CCR7 and IL12RA (cytokine receptors), as well as other functional genes associated with maturation (e.g., NFKB1, NFKB2) (Fig. 6F). Notably, the maturation process seemed stronger in DC3s, showing higher average fold changes of most of the genes mentioned above (Fig. 6F). In contrast, the Ag uptake, proteolytic, and transport functions seemed to decline during maturation, supported by the downregulation of genes coding for C-type lectins (CLEC10A, CLEC4A, and CLEC4G), integrins (ITGAM, and ITGB1), IgG receptors (FCGR2B), proteolytic cathepsins (CTSD and CTSZ), and Ras-associated proteins (RAB32 and RAB7A) (Fig. 6G). Interestingly, most of these genes were strongly downregulated in DC2s, suggesting a unique mature phenotype (Fig. 6G).

In fact, two subpopulations showed specific and distinct maturation features (Supplemental Fig. 4B). Regarding the upregulated genes, DC2s seemed less inflammatory than DC3s, exhibiting unique upregulation of chemokine CXCL9, tolerance-associated TGFBR1, and the surface marker CD1C (Fig. 6H) (30). We also identified 2133 specifically downregulated genes in DC2s, including those coding for multiple integrins, for members of the annexin protein family and for TNFRSFs (Fig. 6H). In contrast, within DC3s, the specifically upregulated genes encoding proinflammatory chemokines (CXCL1, CXCL3, and CCL7), migration-associated chemokine receptors (CCR1, CCR5, and CXCR4), and different types of cytokine receptors (IFNAR2, TNFRSF10A) suggested their higher potential for proinflammatory reactivity against external stimuli (Fig. 6I). Functional analysis also indicated extensive enrichment in the myeloid leukocyte activation, regulation of innate immune response, and adaptive immune system categories in DC3s, as well as a lower enrichment in the T cell activation and myeloid leukocyte activation categories in DC2s (Supplemental Fig. 4C).

The maturation process seemed weaker in UMDCs, with smaller average fold changes of maturation genes (Supplemental Fig. 4D). We also observed no upregulation in the expression of CCR1, CCR5, CXCL9, IL15, and SOCS1, all expected to be upregulated in the context of DC maturation (Supplemental Fig. 4E) (31). Instead, these cells were characterized by the upregulation of immune tolerance-associated genes, including CCL18, CD63, IL7R, and C1Q members (Supplemental Fig. 4E) (13, 17, 32, 33).

Although moDCs are extensively used both in experimental studies on DC biology and in immunotherapy clinical trials, there is a paucity of information on their own heterogeneity as well as on potential differences in comparison with human peripheral DCs. To fill this gap in knowledge, we studied the transcriptomes of moDCs and human peripheral DCs using single-cell RNA-seq. We identified five subtypes corresponding to cDC2s and two CLEC10A+CD127+ moDC-specific subtypes that are not, to our knowledge, similar to any known peripheral DC subpopulations. We further characterized each subpopulation based on their transcriptomic features and trajectory analyses. Importantly via the integration of data from human peripheral DCs, we revealed five similar cDC2 subtypes, renewing the current classification, and further unveiled the transcriptomic difference between moDCs and peripheral cDC2s. We also studied the transcriptomic changes of moDCs after maturation, defined several maturation markers, and demonstrated CLEC10A and SIGLEC10 as markers for immature DCs.

moDCs have been the most widely used models to investigate human DC biology and function, highlighting their indispensable contribution and irreplaceable roles in DC-associated experiments. Moreover, moDCs represent the model of choice for the design of DC-based vaccines and immunotherapies targeting Ag-specific immune responses with, however, no consensus on filtering or sorting steps (34). In this study, we unveiled the heterogeneity within moDCs and explored the transcriptomes of each subtype. Remarkably, we found that five of the seven identified moDC subtypes matched cDC2s, three of them aligning with noninflammatory DC2s (DC2-1, DC2-2, and DC2-3) and the other two aligning with inflammatory DC3s (DC3-1, DC3-2). Both the DC2-1 and DC3-1 subpopulations seemed to perform conventional noninflammatory (DC2) or inflammatory (DC3) functions. Moreover, we believe that the DC2-2 subpopulation has mainly secretory and regulation-associated roles, with the upregulation of cytokines and chemokines involved in immune cell recruitment. Our data further revealed a transitional state (DC2-3) along with the terminal state (DC3-2), possibly revealing the hidden differentiation or transformation pathway from conventional noninflammatory DC2-1 to DC2-3, then to inflammatory DC3-1, and finally to terminal DC3-2. This hypothesis was also supported by the progressive downregulation of CLEC10A and SIGLEC10 and the upregulation of CCL22 and FABP5. Importantly, trajectory analysis identified DC2-1 as the initiation point and highlighted a differentiation route passing through DC2-3 and ending at DC3-2. Taken together, both gene expression patterns and trajectory analysis indicated a potential differentiation pathway or the possible interrelationships between DC2s and DC3s.

The heterogeneity of human peripheral cDC2s was first studied by Villani et al. (3), demonstrating a noninflammatory subtype, DC2, and an inflammatory subtype, DC3. Another study based on protein analyses partially proved heterogeneity, revealing surface markers CD5 (for DC2s) and CD163 and CD14 (for DC3s), and implied these CD14+ DC3s to be highly proinflammatory, based on scanning-electron microscopy (4). More recently, researchers have proposed a different cDC2 classification strategy in both mice and humans, based on the transcription regulators T-bet and RORγt (5). Via the integration of data obtained with moDCs and peripheral DCs, we not only defined similar DC2 and DC3 subsets in moDCs and cDC2s but also updated the current DC classification, extending the two large cDC2 clusters considered today to five subtypes. Corresponding to previous study, the DC3s exhibited higher proinflammatory potential revealed by transcriptional features. Given the possible differentiation path identified on moDCs, we propose that peripheral cDC2s may also follow a similar differentiation trajectory from DC2-1 to DC3-2. Altogether, our findings provide deeper insight into cDC2s, stressing their heterogeneity and potentially evidencing a differentiation trajectory. The two moDC-specific subpopulations (CLEC10A+CD127+ cells), UMDC-1 and UMDC-2, did not correspond to any of the defined DC subpopulations to date. Based on their shared features, we propose that these cells are CD1c+ DCs, possibly associated with immune regulation. Notably, the smaller UMDC-1 subpopulation exhibited surprisingly strong secretory potential, with enrichment in the expression of chemokine-encoding genes (e.g., CCL8, and CCL18) that were reported to induce immune tolerance and prevent autoimmunity (1618). Additionally, the UMDC-2 cells may also have tolerogenic functions, supported by the upregulation of immunosuppressive genes as well as of CCL22, which has been suggested to promote regulatory T cell functions (35). Importantly these new, to our knowledge, moDC subsets made us reconsider the effects of moDCs in previous experimentally based studies. Although these two subsets only accounted for 25% of all moDCs, their potential regulatory functions can strongly influence the outcomes of moDC-based experiments and immunotherapeutic clinical trials. Thus, we suggest that these new, to our knowledge, CLEC10A+CD127+ cells should be removed in the context of experiments on DC biology and function. Further research is needed to explore new applications based on their regulatory potential. Additionally, comparing the other five subtypes in moDCs to human cDC2s moDCs showed weaker Ag processing and presentation potential along with stronger endocytosis and complement production functions, based on gene expression analysis. We also identified different preferential expression patterns of genes associated with DC migration, recruitment, maturation, and immune regulation, indicating complex transcriptomic differences between them. Thus, the transcriptomic difference between moDCs and human cDC2s, as well as the existence of moDC-specific cells, should be taken into consideration during moDC-based experiments and immunotherapies.

Upon encounter with Ags, DCs switch to a mature state, exhibiting strong phenotypic and functional changes. The phagocytic scavenging potential is decreased, and more sophisticated APC-functions are switched on. In this study, after maturation, all profiled moDCs not only exhibited an evident upregulation of maturation markers, such as LAMP3 and CCR7, but also showed a robust downregulation of CLEC10A and SIGLEC10 (1, 36). In fact, both of them almost disappeared in fully matured moDCs; therefore, we propose they could serve as markers for immature DCs. Interestingly, several studies indicated that all peripheral cDC2s should express CLEC10A, a marker gene for noninflammatory DC2s (6, 7). A more recent research confirmed this conclusion but successfully identified CLEC10A-DC3s in human spleen. Taking previous trajectory analysis into account, we infer that DC3s are semimature cDC2s, thus attending more inflammatory functions compared with CLEC10A+ DC2s. We also identified the robust upregulation of SLAMF7 (CD319), IL15RA (CD215), M4A7 (CD20L4), and CD70; in fact, we believe that all of them can potentially serve as surface markers of mature DCs. Of note, the above findings further support the proposed differentiation route from DC2-1 to DC3-2. Furthermore, going into further detail regarding cDC2-like moDCs, DC2-like and DC3-like moDCs exhibited distinct transcriptomic alteration features during maturation. DC3s seemed to demonstrate stronger T cell priming, migration, and leukocyte recruitment abilities, supported by large fold changes in the expression of genes coding for MHC molecules, costimulatory factors, and cytokines. Besides the inflammatory and noninflammatory identities of immature DC2s/DC3s, our findings extended the functional characterization of these two cDC2 subtypes to the maturation state, suggesting their different capacities in Ag presentation, leukocyte recruitment, migration, T cell priming, and immune regulation. As for moDC-specific subpopulations, most conventional maturation-related genes exhibited weaker changes, and several inhibitory and regulatory genes were upregulated in moDCs, especially CCL18. Interestingly, these moDC subsets continued to participate in regulatory roles even after maturation. Overall, in this study, we delineated the transcriptomic landscape of mature cDC2s, identified markers of both mature and immature DCs, revealed the detailed transcriptomic alterations in the context of each subtype, and provided references for further research on the maturation of peripheral DCs.

Altogether, our data reveal the heterogeneity of moDCs and highlight two new, to our knowledge, moDC subtypes (CLEC10A+CD127+ cells) that should be removed to acquire unadulterated cDC2 populations. Our findings provide new insights into cDC2 heterogeneity, renewing the current classification. Moreover, we provide evidence for the possible differentiation trajectory from DC2-1 to DC3-2. We delineate the maturation landscapes of DC2-like and DC3-like moDCs, and propose new markers of mature and immature DCs. In this context, our data will serve as a reference for future research on the maturation of peripheral DCs.

We thank the frontline staff at Zhongshan Op for their dedication and excellent patient care.

This work was supported by the National Key Research and Development Program of China (2017YFA0105804).

Y.Z. and W.S. designed the study. H.L. and X.W. conducted the experiments and acquired the data. L.X. assisted in the experiments. Y.G., Z.L., and Z.H. analyzed the single-cell RNA sequencing data. The manuscript was written by Y.G., B.C., and X. Liu and reviewed by all authors. Z.L., Y.G. and X. Lin prepared the figures and performed the statistical analyses. X.W. revised the manuscript. Order of co-first authors is based on the length of time spent on the project.

The sequences presented in this article have been submitted to the National Genomic Data Center (https://bigd.big.ac.cn/gsa-human/browse/HRA000563) under accession number HRA000563.

The online version of this article contains supplemental material.

Abbreviations used in this article

cDC

conventional DC

DC

dendritic cell

DEG

differentially expressed gene

MHCII

MHC class II

moDC

monocyte-derived DC

scRNA-seq

single-cell RNA sequencing

TF

transcription factor

UMAP

uniform manifold approximation and projection

UMDC

undefined monocyte-derived cell

1.
Villar
J.
,
E.
Segura
.
2020
.
Decoding the heterogeneity of human dendritic cell subsets.
Trends Immunol.
41
:
1062
1071
.
2.
Collin
M.
,
V.
Bigley
.
2018
.
Human dendritic cell subsets: an update.
Immunology
154
:
3
20
.
3.
Villani
A. C.
,
R.
Satija
,
G.
Reynolds
,
S.
Sarkizova
,
K.
Shekhar
,
J.
Fletcher
,
M.
Griesbeck
,
A.
Butler
,
S.
Zheng
,
S.
Lazo
, et al
2017
.
Single-cell RNA-seq reveals new types of human blood dendritic cells, monocytes, and progenitors.
Science
356
:
eaah4573
.
4.
Chapuy
L.
,
M.
Sarfati
.
2020
.
Single-cell protein and RNA expression analysis of mononuclear phagocytes in intestinal mucosa and mesenteric lymph nodes of ulcerative colitis and Crohn’s disease patients.
Cells
9
:
813
.
5.
Brown
C. C.
,
H.
Gudjonson
,
Y.
Pritykin
,
D.
Deep
,
V. P.
Lavallée
,
A.
Mendoza
,
R.
Fromme
,
L.
Mazutis
,
C.
Ariyan
,
C.
Leslie
, et al
2019
.
Transcriptional basis of mouse and human dendritic cell heterogeneity.
Cell
179
:
846
863.e24
.
6.
Heger
L.
,
S.
Balk
,
J. J.
Lühr
,
G. F.
Heidkamp
,
C. H. K.
Lehmann
,
L.
Hatscher
,
A.
Purbojo
,
A.
Hartmann
,
F.
Garcia-Martin
,
S.-I.
Nishimura
, et al
2018
.
CLEC10A is a specific marker for human CD1c+ dendritic cells and enhances their toll-like receptor 7/8-induced cytokine secretion.
Front. Immunol.
9
:
744
.
7.
Heidkamp
G. F.
,
J.
Sander
,
C. H. K.
Lehmann
,
L.
Heger
,
N.
Eissing
,
A.
Baranska
,
J. J.
Lühr
,
A.
Hoffmann
,
K. C.
Reimer
,
A.
Lux
, et al
2016
.
Human lymphoid organ dendritic cell identity is predominantly dictated by ontogeny, not tissue microenvironment.
Sci. Immunol.
1
:
eaai7677
.
8.
Sallusto
F.
,
A.
Lanzavecchia
.
1994
.
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.
J. Exp. Med.
179
:
1109
1118
.
9.
Dudek
A. M.
,
S.
Martin
,
A. D.
Garg
,
P.
Agostinis
.
2013
.
Immature, semi-mature, and fully mature dendritic cells: toward a DC-cancer cells interface that augments anticancer immunity.
Front. Immunol.
4
:
438
.
10.
Romani
N.
,
S.
Gruner
,
D.
Brang
,
E.
Kämpgen
,
A.
Lenz
,
B.
Trockenbacher
,
G.
Konwalinka
,
P. O.
Fritsch
,
R. M.
Steinman
,
G.
Schuler
.
1994
.
Proliferating dendritic cell progenitors in human blood.
J. Exp. Med.
180
:
83
93
.
11.
Inaba
K.
,
R. M.
Steinman
,
M. W.
Pack
,
H.
Aya
,
M.
Inaba
,
T.
Sudo
,
S.
Wolpe
,
G.
Schuler
.
1992
.
Identification of proliferating dendritic cell precursors in mouse blood.
J. Exp. Med.
175
:
1157
1167
.
12.
Zimmer
A.
,
J.
Bouley
,
M.
Le Mignon
,
E.
Pliquet
,
S.
Horiot
,
M.
Turfkruyer
,
V.
Baron-Bodo
,
F.
Horak
,
E.
Nony
,
A.
Louise
, et al
2012
.
A regulatory dendritic cell signature correlates with the clinical efficacy of allergen-specific sublingual immunotherapy.
J. Allergy Clin. Immunol.
129
:
1020
1030
.
13.
Guimond
M.
,
R. G.
Veenstra
,
D. J.
Grindler
,
H.
Zhang
,
Y.
Cui
,
R. D.
Murphy
,
S. Y.
Kim
,
R.
Na
,
L.
Hennighausen
,
S.
Kurtulus
, et al
2009
.
Interleukin 7 signaling in dendritic cells regulates the homeostatic proliferation and niche size of CD4+ T cells.
Nat. Immunol.
10
:
149
157
.
14.
Shalev
I.
,
H.
Liu
,
C.
Koscik
,
A.
Bartczak
,
M.
Javadi
,
K. M.
Wong
,
A.
Maknojia
,
W.
He
,
M. F.
Liu
,
J.
Diao
, et al
2008
.
Targeted deletion of fgl2 leads to impaired regulatory T cell activity and development of autoimmune glomerulonephritis.
J. Immunol.
180
:
249
260
.
15.
Dean
R. A.
,
J. H.
Cox
,
C. L.
Bellac
,
A.
Doucet
,
A. E.
Starr
,
C. M.
Overall
.
2008
.
Macrophage-specific metalloelastase (MMP-12) truncates and inactivates ELR+ CXC chemokines and generates CCL2, -7, -8, and -13 antagonists: potential role of the macrophage in terminating polymorphonuclear leukocyte influx.
Blood
112
:
3455
3464
.
16.
Nguyen Hoang
A. T.
,
H.
Liu
,
J.
Juaréz
,
N.
Aziz
,
P. M.
Kaye
,
M.
Svensson
.
2010
.
Stromal cell-derived CXCL12 and CCL8 cooperate to support increased development of regulatory dendritic cells following Leishmania infection.
J. Immunol.
185
:
2360
2371
.
17.
Azzaoui
I.
,
S. A.
Yahia
,
Y.
Chang
,
H.
Vorng
,
O.
Morales
,
Y.
Fan
,
N.
Delhem
,
C.
Ple
,
A. B.
Tonnel
,
B.
Wallaert
,
A.
Tsicopoulos
.
2011
.
CCL18 differentiates dendritic cells in tolerogenic cells able to prime regulatory T cells in healthy subjects.
Blood
118
:
3549
3558
.
18.
Vulcano
M.
,
S.
Struyf
,
P.
Scapini
,
M.
Cassatella
,
S.
Bernasconi
,
R.
Bonecchi
,
A.
Calleri
,
G.
Penna
,
L.
Adorini
,
W.
Luini
, et al
2003
.
Unique regulation of CCL18 production by maturing dendritic cells.
J. Immunol.
170
:
3843
3849
.
19.
Leitner
J.
,
D.
Herndler-Brandstetter
,
G. J.
Zlabinger
,
B.
Grubeck-Loebenstein
,
P.
Steinberger
.
2015
.
CD58/CD2 is the primary costimulatory pathway in human CD28-CD8+ T cells.
J. Immunol.
195
:
477
487
.
20.
Lock
K.
,
J.
Zhang
,
J.
Lu
,
S. H.
Lee
,
P. R.
Crocker
.
2004
.
Expression of CD33-related siglecs on human mononuclear phagocytes, monocyte-derived dendritic cells and plasmacytoid dendritic cells.
Immunobiology
209
:
199
207
.
21.
Agod
Z.
,
K.
Pazmandi
,
D.
Bencze
,
G.
Vereb
,
T.
Biro
,
A.
Szabo
,
E.
Rajnavolgyi
,
A.
Bacsi
,
P.
Engel
,
A.
Lanyi
.
2018
.
Signaling lymphocyte activation molecule family 5 enhances autophagy and fine-tunes cytokine response in monocyte-derived dendritic cells via stabilization of interferon regulatory factor 8.
Front. Immunol.
9
:
62
.
22.
Oh
S. A.
,
D.-C.
Wu
,
J.
Cheung
,
A.
Navarro
,
H.
Xiong
,
R.
Cubas
,
K.
Totpal
,
H.
Chiu
,
Y.
Wu
,
L.
Comps-Agrar
, et al
2020
.
PD-L1 expression by dendritic cells is a key regulator of T-cell immunity in cancer.
Nat. Cancer
1
:
681
691
.
23.
Demeure
C. E.
,
H.
Tanaka
,
V.
Mateo
,
M.
Rubio
,
G.
Delespesse
,
M.
Sarfati
.
2000
.
CD47 engagement inhibits cytokine production and maturation of human dendritic cells.
J. Immunol.
164
:
2193
2199
.
24.
Weiss
J. M.
,
J.
Sleeman
,
A. C.
Renkl
,
H.
Dittmar
,
C. C.
Termeer
,
S.
Taxis
,
N.
Howells
,
M.
Hofmann
,
G.
Köhler
,
E.
Schöpf
, et al
1997
.
An essential role for CD44 variant isoforms in epidermal Langerhans cell and blood dendritic cell function.
J. Cell Biol.
137
:
1137
1147
.
25.
Tejera
E.
,
V.
Rocha-Perugini
,
S.
López-Martín
,
D.
Pérez-Hernández
,
A. I.
Bachir
,
A. R.
Horwitz
,
J.
Vázquez
,
F.
Sánchez-Madrid
,
M.
Yáñez-Mo
.
2013
.
CD81 regulates cell migration through its association with Rac GTPase.
Mol. Biol. Cell
24
:
261
273
.
26.
Jones
E. L.
,
J. L.
Wee
,
M. C.
Demaria
,
J.
Blakeley
,
P. K.
Ho
,
J.
Vega-Ramos
,
J. A.
Villadangos
,
A. B.
van Spriel
,
M. J.
Hickey
,
G. J.
Hämmerling
,
M. D.
Wright
.
2016
.
Dendritic cell migration and antigen presentation are coordinated by the opposing functions of the Tetraspanins CD82 and CD37.
J. Immunol.
196
:
978
987
.
27.
Lo
T. H.
,
P. A.
Silveira
,
P. D.
Fromm
,
N. D.
Verma
,
P. A.
Vu
,
F.
Kupresanin
,
R.
Adam
,
M.
Kato
,
V. C.
Cogger
,
G. J.
Clark
,
D. N.
Hart
.
2016
.
Characterization of the expression and function of the C-type lectin receptor CD302 in mice and humans reveals a role in dendritic cell migration.
J. Immunol.
197
:
885
898
.
28.
Li
K.
,
H.
Fazekasova
,
N.
Wang
,
P.
Sagoo
,
Q.
Peng
,
W.
Khamri
,
C.
Gomes
,
S. H.
Sacks
,
G.
Lombardi
,
W.
Zhou
.
2011
.
Expression of complement components, receptors and regulators by human dendritic cells.
Mol. Immunol.
48
:
1121
1127
.
29.
Xia
M. Q.
,
G.
Hale
,
H.
Waldmann
.
1993
.
Efficient complement-mediated lysis of cells containing the CAMPATH-1 (CDw52) antigen.
Mol. Immunol.
30
:
1089
1096
.
30.
Seeger
P.
,
T.
Musso
,
S.
Sozzani
.
2015
.
The TGF-β superfamily in dendritic cell biology.
Cytokine Growth Factor Rev.
26
:
647
657
.
31.
Jackson
S. H.
,
C. R.
Yu
,
R. M.
Mahdi
,
S.
Ebong
,
C. E.
Egwuagu
.
2004
.
Dendritic cell maturation requires STAT1 and is under feedback regulation by suppressors of cytokine signaling.
J. Immunol.
172
:
2307
2315
.
32.
Teh
B. K.
,
J. G.
Yeo
,
L. M.
Chern
,
J.
Lu
.
2011
.
C1q regulation of dendritic cell development from monocytes with distinct cytokine production and T cell stimulation.
Mol. Immunol.
48
:
1128
1138
.
33.
Mantegazza
A. R.
,
M. M.
Barrio
,
S.
Moutel
,
L.
Bover
,
M.
Weck
,
P.
Brossart
,
J. L.
Teillaud
,
J.
Mordoh
.
2004
.
CD63 tetraspanin slows down cell migration and translocates to the endosomal-lysosomal-MIICs route after extracellular stimuli in human immature dendritic cells.
Blood
104
:
1183
1190
.
34.
Wculek
S. K.
,
F. J.
Cueto
,
A. M.
Mujal
,
I.
Melero
,
M. F.
Krummel
,
D.
Sancho
.
2020
.
Dendritic cells in cancer immunology and immunotherapy.
Nat. Rev. Immunol.
20
:
7
24
.
35.
Rapp
M.
,
M. W. M.
Wintergerst
,
W. G.
Kunz
,
V. K.
Vetter
,
M. M. L.
Knott
,
D.
Lisowski
,
S.
Haubner
,
S.
Moder
,
R.
Thaler
,
S.
Eiber
, et al
2019
.
CCL22 controls immunity by promoting regulatory T cell communication with dendritic cells in lymph nodes.
J. Exp. Med.
216
:
1170
1181
.
36.
Jin
P.
,
T. H.
Han
,
J.
Ren
,
S.
Saunders
,
E.
Wang
,
F. M.
Marincola
,
D. F.
Stroncek
.
2010
.
Molecular signatures of maturing dendritic cells: implications for testing the quality of dendritic cell therapies.
J. Transl. Med.
8
:
4
.

The authors have no financial conflicts of interest.

Supplementary data