Tissues obtained from hysterectomies were separated by endometrium (EM), endocervix (END) and ectocervix (ECT) by pathologists and transferred to the laboratory post-surgery. Tissues were processed as described previously (10, 31, 42). Briefly, tissues were minced into 1-2 mm fragments in Roswell Park Memorial Institute (RPMI) medium (Gibco) containing enzymes from Tumor Dissociation Kit, human (Miltenyi Biotec) and 0.01% DNAse (Worthington Biochemical) and transfered into sterile gentleMACS™ C Tubes (Miltenyi Biotec). Enzymatic digestion was performed on genlteMACS™ Dissociator (Miltenyi Biotec) using “37C_h_TDK_1” program. Digested tissue was filtered through 100μm, 70μm and 30μm MACS™ SmartStrainers (Miltenyi Biotec) to generate stromal single cells suspensions.

Mixed single-cell suspensions from tissues were washed in Phosphate-Buffered Saline (PBS; Gibco). Subsequently, cells were incubated in the dark at room temperature with LIVE/DEAD Blue (ThermoFisher) dye for 10 minutes. Anti-CD16 (BD Biosciences) was then added and incubated for 5 minutes, followed by anti-CCR7 (BioLegend), anti-CXCR4 (BioLegend), anti-CCR5 (BioLegend) and anti-CX3CR1 (BioLegend) for another 5 minutes. Finally, cells were incubated with an antibody cocktail containing remaining antibodies ( Supplementary Table S2 ) at room temperature for 15 minutes, washed with MACS buffer, fixed in 2% paraformaldehyde (PFA;Thermo Fisher), and analyzed for surface marker expression on Cytek Aurora (5 laser, 64 detector configuration; Cytek Biosciences). Expression of surface markers were quantified using OMIQ (Dotmatics).

HIV-BaL (R5) isolates obtained from the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH (43) were propagated through infection of peripheral blood mononuclear cells (PBMCs) activated with phytohemagglutinin (PHA) (2.5 μg/mL; Sigma, St. Louis, MO) and IL-2 (50 U/mL; AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH: Human rIL-2 from Dr. Maurice Gately, Hoffmann – La Roche Inc) for 6-8 days. Stocks were harvested when p24 concentrations reached 100 ng/mL. Titration of viral stocks were performed using PHA and IL-2 activated PBMCs (44).

Mixed single-cell suspensions from hysterectomy samples obtained from 4 healthy female donor, consisting of 4 endometrium (EM), 2 endocervix (END) and 4 ectocervix (n =10 tissues), were enriched for immune cells by magnetic bead removal of CD3⁺ (CD3 MicroBeads, human; Miltenyi Biotec), CD19⁺ (CD19 MicroBeads, human; Miltenyi Biotec), CD235a⁺ (CD235a (Glycophorin A) MicroBeads, human; Miltenyi Biotec) red blood cells and fibroblasts (Anti-fibroblast MicroBeads, human; Miltenyi Biotec). For homeostatic conditions, cells were washed thoroughly with PBS containing 5% HS to remove any excess magnetic beads and incubated with oligo-conjugated antibodies (AbSeq; Supplementary Table S2 ) and barcoded sample tags (BD™ Hu Single Cell Sample Multiplexing Kit; BD Biosciences), to differentiate between tissue sites, for 20 minutes at room temperature in PBS containing 0.5% HS. For HIV stimulation experiments, cells were incubated with HIV-BaL (MOI = 0.5) in XVIVO-15 or media alone for 30 minutes. Cells were washed thoroughly with PBS+0.5% HS to remove unbound virus and subsequently incubated with oligo-conjugated antibodies and barcoded sample tags as mentioned above. Cells were subsequently washed twice in excess PBS+0.5% HS to remove unbound antibodies and sample tags. Cells were counted and 10,000 cells from each tissue were combined to obtain a total of 30,000 cells each for control and HIV conditions. Cells were then subsequently loaded onto separate BD Rhapsody™ Cartridge (BD Rhapsody™ Cartridge Kit; BD Biosciences) followed by RNA capture on polyA tail capture beads containing unique molecular identifiers (UMIs) (BD Rhapsody™ Enhanced Cartridge Reagent Kit V3; BD Biosciences) on BD Rhapsody™ Express Single-Cell Analysis System (BD Biosciences) followed by cDNA, whole-transcriptome amplification and single-cell indexing libraries according to manufacturer’s protocol (Protocol-Rhapsody WTA+AbSeq+ST; BD Biosciences).

FASTQ files containing unaligned reads were generated on NovaSeq6000 sequencing system (Illumina). Gene counts were generated by aligning and annotating reads to the human genome (GRCh38.p12 v29). To assess capture of HIV, reads from respective experiments were also aligned to the HIV genome FASTA file. Count tables were subsequently uploaded to Partek Flow (Partek, an Illumina company) for downstream quantification and visualization of data. Cells with mitochondrial gene expression >25% number of detected features per cell less than 200 or greater than 4300 were excluded from analysis, according to recent studies published by us (31) and others (45). RNA and protein data were split and normalized respectively. To identify DCs, we first selected immune cells expressing PTPRC genes, which encodes CD45 protein. Next, we used protein information to exclude remaining T and B cells by selecting CD3^-CD19⁺ cells, followed by selection of CD11c+HLA-DR-DP-DQ+ cells. Since genital NK cells and neutrophils can express CD11c and HLA-DR (46, 47), we excluded neutrophil contamination by selecting CD15- cells within this population, and NK cell contamination by gating out cells expressing NCAM1 RNA ( Figure 1A ). PCA was performed on normalized data followed by UMAP for visualization. Additionally, we also performed weighted-nearest neighbor analysis to integrate protein and RNA data. Subsequently, we performed k-nearest neighbor analysis and graph-based analysis to identify cell clusters and used “compute biomarker” function in Partek Flow (Partek – an Illumina company) to generate biomarkers associated with the clusters ( Supplementary Data Sheet S1 ). To identify differentially expressed genes (DEGs) between DC subsets, we performed non-parametric ANOVA to identify significantly upregulated and downregulated genes with significance of p ≤ 0.05 and Log₂(FoldChange) ± 1.2. GO Biological processes (48, 49) and Reactome (50) were generated using gene lists of upregulated and downregulated genes. To identify unique expression of genes within each cluster, “compute biomarker” function in Partek Flow Analysis. Curated gene lists were used to assess expression of CD markers, cytokines, chemokines, PRRs, antimicrobial proteins and gene lists associated with GO processes for AUCell (51) ( Supplementary Data Sheet S2 ) were used from GSEA analysis available in Partek Flow (Partek – an Illumina company).

CD14⁺ DCs were isolated from mixed-cell suspensions using magnetic bead separation (CD14 MicroBeads, human; BD Biosciences) as per previous studies (25, 30, 31, 42). Isolated cells were incubated at 37°C, 5% CO₂ with HIV-1-BaL (R5) isolates (MOI = 0.5) for 3 hours. Supernatants were collected and stored at -80C until multiplex assay analysis to determine secreted protein levels.

Supernatants from unstimulated controls and HIV exposed CD14+ DCs were quantified using Millipore human cytokine multiplex kits according to manufacturer’s instructions. Signal was measured using Magpix software with five-parametric-curve fitting for data analysis. Molecules measured included IL-1β, IFNγ, IL-5, IL-13, IL1Rα, GM-CSF, G-CSF, CCL11, CCL22, CXCL1, CXCL10 and CX3CL1.

To identify differentially expressed genes, significance threshold was set as (p<0.05; -1.2 <FC>1.2) on non-parametric ANOVA. Statistics for flow cytometry analysis were done using Friedman’s multiple comparison, (p ≤ 0.05 *; p ≤ 0.01 **; p ≤ 0.001 ***; p ≤ 0.0001 ****). Hierarchical clustering heatmaps were performed on significant, differentially expressed genes, by generating pseudo-bulk expression of groups (eg. HIV vs control; cluster 1, cluster 2, cluster 3, cluster 4 comparison) and visualized using bubble plots, with size of the bubble referring to percentage of cells expressing the particular gene, shade and color referring to Z-score expression of the gene. For Reactome and GO analysis, terms with FDR ≤ 0.05 were chosen as significant values and bubble plots were visualized using GraphPad Prism.

Written informed consent was obtained before surgery from HIV-negative women undergoing hysterectomies at Tufts Medical Center (Boston, MA, USA). Studies were approved by Tufts University Institutional Review Board and the Committee for the Protection of Human Subjects. Surgery was performed to treat benign conditions including fibroids, prolapse, and menorraghia. Trained pathologists selected tissue samples from endometrium (EM), endocervix (END), and ectocervix (ECX), free of pathological lesions and distant from the sites of pathology. Women were HIV− and HPV− but no additional information regarding other genital infections was available.

Prior studies have identified different genital DC and mononuclear phagocyte populations within the anogenital mucosa (14, 15, 20, 24, 25, 31, 39, 42, 52). However, transcriptional, phenotypical, and functional characterization of the heterogenous CD14-expressing populations in the genital tract remains poorly defined.

To address these gaps, we optimized a protocol to define DC subsets in the FGT phenotypically and transcriptionally at the single-cell level. We generated single cell suspensions from human hysterectomy samples and enriched the target immune cells as detailed in methods. We labeled the enriched cells with oligo-conjugated antibodies targeted towards surface proteins ( Supplementary Table S1 ), lysed the cells to release RNA, and sequenced the RNA to simultaneously determine the whole transcriptome profile and surface protein expression at the single-cell level. To identify the genital resident DCs, we took advantage of CITE-seq’s combined surface protein information and gene expression and developed a gating strategy to select the CD11c⁺HLA-DR⁺ population ( Figure 1A ), which contains DCs. This approach eliminates the need for lengthy fluorescence activated cell-sorting (FACS) prior to sequencing, thereby limiting processing steps that can alter primary tissue resident immune cells (53).

Unbiased clustering analysis discriminated four distinct clusters of CD11c⁺HLA-DR⁺ cells within the genital mucosa ( Figure 1B ). We determined the top RNA and surface protein (Abseq) expression unique to each cluster by using “compute biomarkers” function ( Figure 1B ). Cluster 1 was characterized by RNA expression of VCAN, S100A8, S100A9 and surface protein expression of CD14, CD11b and CD64 ( Figure 1C ), confirming the presence of a CD14⁺ monocyte-derived DC population in the genital mucosa as previously described by us and others (20, 25, 30, 31, 52, 54). Cluster 2 was characterized by surface protein expression of CD1c, the DC maturation marker CD83, and the pro-survival and tissue homing marker CXCR4 (55, 56), suggesting enrichment of cDC2s in this cluster ( Figure 1C ). Clusters 3 and 4 displayed phenotypes similar to monocyte populations. Cluster 3 expressed elevated CD16 and CD127, whereas cluster 4 displayed elevated expression of CD64, HLA-DR and low levels of CLEC12A protein expression ( Figure 1C ). Additionally, cluster 3 expressed RNA transcripts such as FCGR3A, SMIM25 and SPN ( Figure 1B ), consistent with non-classical monocyte (NCM) signature expression (57). Cluster 4 expressed elevated levels of complement encoding RNA transcripts C1QC and C1QB, consistent with inflammatory monocytes/macrophages (infMons) (58). These two clusters indicated the presence of two distinct monocyte populations within CD11c⁺HLA-DR⁺ cells.

Recent studies identified a novel DC population in peripheral blood named DC3 that co-expresses CD14 and CD1c and displays and intermediate phenotype and function between monocytes and classical myeloid DCs (cDC2s) (36, 59). Interestingly, we and others have previously described the presence of CD14⁺CD1c⁺ cells in the FGT (20, 25, 52). To determine whether the DC3 population was present in the FGT, we analyzed our CITE-seq data using AUCell (51) with the top 50 genes described by Villani et al. in their peripheral blood DC3 cluster ( Supplementary Data Sheet S2 ) (35). We observed an enrichment of the DC3 signature in cluster 1 ( Figure 1D ), indicating the presence of DC3s within the CD14⁺ DC cluster.

Overall, our multi-omics data indicates the presence of four distinct CD11c⁺HLA-DR⁺ populations in the FGT, consisting of two DC and two monocyte populations, with unique transcriptional and phenotypic signatures. Furthermore, we observed the presence of novel DC3 gene signature in the CD14⁺ DC cluster, suggesting inherent heterogeneity of CD14⁺ DCs within the FGT.

Next, we investigated whether our phenotypical clustering was associated with distinct subset-specific functions for each subset under homeostatic conditions.

To determine if the distinct populations would be differently posed to detect and interact with invading pathogens, including HIV, we compared expression levels of PRRs ( Supplementary Data Sheet S2 ) including toll-like receptors (TLRs), C-type lectin receptors (CLRs) and NOD-like receptors (NLRs). Using a list of known PRRs ( Supplementary Table S2 ) we compared expression levels between subsets through hierarchical clustering ( Figure 2A ). We observed no TLR9 expression across all subsets, ruling out the possibility of plasmacytoid DC contribution to the gene signature of each subset. Compared to the other clusters, Cluster 1 (CD14⁺ DCs) showed relatively higher expression of TLRs and CLRs associated with detecting and binding to HIV, including CLEC4A, CLEC4E and TLR4 (60–62). Additionally, in Clusters 1 and 3, we observed shared expression of CLEC12A, which has been linked to improved antigen delivery and cross presentation by DCs (63), and CLEC7A, previously described in CD14⁺ DCs from blood and lungs (64). Additionally, Clusters 1, 3 and 4 also shared expression of TLR2, which binds HIV ( 65 ). Cluster 2 (cDC2s) had no unique gene expression pattern. The highest expression in Cluster 2 was observed for the CLRs CLEC10A and CLEC1A, involved in antigen internalization and presentation (66, 67). When compared to the other clusters, Cluster 3 (NCM) displayed high preferential expression of IFIH1 (encoding MDA5), DDX58 (encoding RIGI) which mediate detection of cytosolic viral RNA (68, 69), and TLR8, which recognizes endolysosomal ssRNA (70). Cluster 4 (infMons) expressed MRC1 (encoding mannose receptor (MR)) and CD209, consistent with a monocyte/macrophage phenotype (71, 72). Cluster 4 also expressed high levels of CLEC5A which forms heterodimers with CD209 and MR upon antigen exposure to enhance viral internalization (73). Further, compared to the other clusters, Cluster 4 had preferential expression of TLR1 and TLR6, known to dimerize with TLR2 for detection of bacterial lipopeptides; TLR7 and TRL3 which recognize ssRNA and dsRNA, respectively (74, 75). Finally, a small percentage of cells in Cluster 4 preferentially expressed AIM2, involved in recognition of cytosolic dsDNA (76). Altogether, our analysis reveals subset specific differences in genital DC recognition of different moieties of pathogens, Cluster 1 being enriched with PRRs associated with membrane HIV recognition and binding, Cluster 2 expressed PRRs for antigen uptake, Cluster 3 and Cluster 4 were enriched for PRRs that detects cytosolic and endolysosomal ssRNA, dsRNA and dsDNA.

Using the same analytical approach, we next focused on cytokine and chemokine expression under homeostatic conditions, to understand functional differences in innate secretory functions involved in maintenance of mucosal barrier integrity, immune cell recruitment and anti-microbial protection. We identified the top 5 cytokine and chemokine gene signatures unique to each cluster ( Figure 2B ; Supplementary Data Sheet S2 ). Cluster 1 (CD14⁺ DCs) was characterized by CCL20, CXCL8 and CXCL2 expression, genes involved in antimicrobial activity (77), chemotaxis for Th17 cells, neutrophils, monocytes and DCs (78–81). Expression of IL1B and its inhibitory receptor IL1RN were also elevated in Cluster 1, suggesting production of IL1β but prevention of autocrine/paracrine actions of IL1β on CD14⁺ DCs. Cluster 2 (containing cDC2s), specialized in expression of genes involved in T cell chemotaxis: IL33, an alarmin that controls tissue homeostasis and type 2 immunity (82); CCL17, a chemoattractant for helper T cells expressing CCR4 (such as Th2 and regulatory T cells) (83) (84); CXCL12, a CXCR4-ligand; CX3CL1 (fractalkine) which play critical roles in menstruation (85); and IL7, suggesting an important role for cDC2s in control of genital T cell populations and favoring Th2 and T regulatory profiles. Cluster 3 (NCMs) expressed high levels of IL16, a cytokine that exclusively binds and signals through CD4 (86); and CXCL16, important for control of MAIT cells and NK cells and expressed by non-classical monocytes in blood (87–89). In addition, a smaller percentage of cells in cluster 3 expressed CCR3 ligands CCL24, CCL26 and CCL13, with antimicrobial properties and potent chemoattractants for eosinophils and basophils (90) (91). Cluster 4 (infMons) was characterized by expression of CCR5 ligands CCL3, CCL4, CCL5 and CXCR4 ligands CXCL12 and CCL4L2, suggesting a potential role in mediating anti-HIV activity. Cluster 4 also expressed high levels of IL18, a proinflammatory cytokine described in tissue-resident macrophages that induces IFN-γ production by T cells in an inflammatory environment, or Th2 differentiation in the absence of inflammatory cytokines (92, 93).

To further understand functional contributions to mucosal homeostasis and defense, we performed AUCell analysis (51) of genes associated with immune protection ( Figure 2C ; Supplementary Data Sheet S2 ). Cluster 1, containing CD14⁺ DCs, was enriched for genes associated with inflammatory and defense response, suggesting enhanced potential to mediate innate immune protection. Furthermore, enrichment of genes associated with positive regulation of myeloid leukocyte migration within this cluster suggests the presence of a migratory DC population. Cluster 2 (cDC2s) and Cluster 4 (infMon) were enriched for genes related to activation of the complement classical pathway, correlating with increased expression of complement associated genes. Lastly, no differences were observed for enrichment of antigen processing and presentation or positive regulation of T cell proliferation genes in genital CD11c⁺HLA-DR⁺ subsets under homeostatic conditions.

Overall, our data demonstrates distinct functional specialization of genital CD11c⁺HLA-DR⁺ subsets under homeostatic conditions.

Data in this study highlights the inherent heterogeneity of DC subsets within the mucosa, with unique transcriptional and phenotypic properties under homeostatic conditions. To validate our CITE-seq findings in a larger number of patients and further delineate genital DC populations, we developed a multi-parameter spectral flow cytometry panel ( Supplementary Table S2 ). We gated on CD11c⁺HLA-DR⁺ cells ( Figure 3A ) and identified four different populations based on CD1c and CD14 surface expression ( Figure 3B ), main discriminatory markers identified in our CITEseq data ( Figure 1C ). As seen in Figure 3C , CD14^highCD1c^- cells (CD14⁺ DCs) were the most abundant subset (53%), followed by CD14^lowCD1c- (29.1%), CD14⁺CD1c⁺ DCs (DC3s) (7.44%) and CD14^-CD1c⁺ (cDC2s) cells (6.5%). To determine if CD11c⁺HLA-DR⁺ subset distribution varies throughout the genital tract, we compared the abundance of each subset in the endometrium (EM), endocervix (END), and ectocervix (ECX) ( Figure 3D ). We observed an increased abundance of CD14⁺ cells in the ECX compared to the EM, whereas the opposite was true for the CD14^lowCD1c- subset, with no significant changes observed in CD1c⁺ and CD14⁺CD1c⁺ cells.

In our CITE-seq data set we identified two distinct CD14^-CD1c^- populations with differential expression of CD16 and CLEC12A ( Figure 1C ). Therefore, we further characterized the CD14^-CD1c^- population to determine if this group was composed of two different cell subsets. Consistent with our CITE-seq findings, the CD14^lowCD1c^- population contained two different subsets: CD14^-CD16⁺ and CD14^-CD16^- ( Figure 3E ). Furthermore, these two populations could be distinguished by CLEC12A expression ( Figure 3E , histogram), suggesting CD14^-CD16⁺CLEC12A⁺ cells are similar to NCMs, whereas lack of CLEC12A expression in CD14^-CD16^- cells corresponds to the inflammatory/activated monocyte population (infMon).

Next, we analyzed expression of surface markers associated with DC activation status and function ( Figure 3F ). First, we assessed maturation status by gating on HLA-DR^high cells consistent with our previous publication (25). We observed that a large proportion of CD14⁺CD1c⁺ and CD1c⁺ cells were HLA-DR^high, significantly higher than other CD11c⁺HLA-DR⁺ subsets. Compared to the other subsets, CD14⁺CD1c⁺ and CD14⁺ cells showed higher expression of CD11b and CD64, associated with cell adhesion and antigen uptake. CD14⁺CD1c⁺ cells expressed the highest levels of CD54, a molecule involved in enhanced immune synapse formation between DCs and T cells (94). CD14^-CD16⁺ cells displayed the highest levels of CLEC12A (a myeloid inhibitory receptor), and along with CD14⁺CD1c⁺ cells, showed high expression of CX3CR1, the receptor for CX3CL1 (fractalkine).

Previous studies from our group and others have shown that CD14-expressing FGT DCs preferentially capture HIV viral-like particles (15, 18, 20, 25), indicating subset-specific interactions with HIV. Based on this, we evaluated expression of receptors and coreceptors for HIV in the different subsets ( Figure 3G ). We observed that all three DC subsets expressed significantly higher levels of CD4 and CXCR4 compared to CD14^-CD16^- monocytes. Importantly, CD14⁺CD1c⁺ DCs expressed the highest levels of CCR5, which is crucial for HIV acquisition in the mucosa (95). In addition to the classical co-receptors, CD49d (α4β7 integrin) expression was significantly higher in CD14⁺CD1c⁺ DCs and CD14^-CD16⁺ monocytes compared to other subsets. CD49d acts as a tissue homing marker and non-classical HIV co-receptor important for mucosal infection (96).

Altogether, our flow cytometry data demonstrates the presence of phenotypically distinct genital DC populations and further supports the presence of CD14+ DCs and DC3s within the CD14+ population.We demonstrate that genital DC3s expressed levels of activation markers similar to conventional cDC2s (CD1c⁺ DCs) and share expression of classical monocyte-derived DC markers with CD14⁺ DCs. Additionally, we demonstrate elevated expression of HIV coreceptors in DC3s compared to other DC populations, suggesting predisposed differential response to HIV within the genital mucosa.

Earlier studies investigated the role of DCs in HIV pathogenesis using monocyte-derived DCs over a time course from 6-48 hours to assess their responses to HIV during viral uptake and infection (22). However, most studies evaluating the role of genital DCs in HIV pathogenesis have investigated events that occur 12 hours or more after viral exposure, focusing on DC susceptibility to HIV infection and trans-infection to CD4⁺ T cells (38) (15, 18, 20). We have previously shown that genital DCs release chemokines and antimicrobials in a rapid manner, within 3 hours after HIV stimulation (25), indicating a role for DCs in triggering the initial mucosal innate response against the virus. However, the overall antiviral response induced by HIV immediately following challenge of genital DCs, prior to viral replication, integration, or productive infection, remains uncharacterized.

To address this gap, we adapted our CITEseq approach to identify early transcriptional responses of genital DCs to HIV. Single cell suspensions generated from human hysterectomy samples were incubated with HIV at an MOI of 0.5 in vitro for 30 minutes, prior to proceeding with the CITE-seq protocol described in detail in methods, to determine whole transcriptome profile and surface protein expression simultaneously at the single-cell level ( Figure 4A ). The time of 30 minutes was chosen to identify pathways involved in viral recognition and immune response independent of viral replication, integration and productive infection. We used the surface protein expression information to select CD11c⁺HLA-DR⁺ cells as done in Figure 1A , which contains the DC and monocytic populations, and performed differential gene expression analysis between HIV and unstimulated control conditions to define transcriptional changes induced by viral exposure. This analysis identified a significant transcriptional shift induced by HIV within 30 minutes, with 333 genes differentially upregulated and 4148 genes downregulated in response to HIV ( Figure 4B ; ( Supplementary Data Sheet S1 ).

To define this HIV response, we first focused on genes associated with DC-mediated immune protection, inflammation, and antiviral response within the top 50 differentially expressed genes based on fold change and p-value ( Figure 4B ; Supplementary Data Sheet S1 ).

Within the upregulated genes, we detected increased expression of the interferon-stimulated genes (ISGs) IRF1 and IFITM2, indicating initiation of innate immune activation and anti-viral response. We also observed a strong upregulation of genes associated with interleukin and chemokine signaling pathways (IL1B, CXCL8, CXCL10, CISH, SOCS3), indicating rapid induction of innate secreted responses. Consistent with increased secretory response, we observed upregulation of serglycin gene expression (SRGN) in response to HIV, involved in storage and secretion of innate molecules in intracellular vesicles in blood monocytes and other cell types (97). Additionally, we detected upregulation of ACOD1, a gene involved in antimicrobial and antiviral responses of innate cells, and a negative regulator of TLR-mediated inflammatory responses (98).

We detected downregulation of genes associated with pro-viral replication of HIV such as GRIP1, EIF4A2, SRRM2, XRN1, and MALAT1 (99–102), suggesting suppression of host mechanisms to prevent HIV replication in genital DCs.

To uncover markers that enable the identification of phenotypic changes induced by HIV in genital DCs, we evaluated differences in expression of genes associated with surface protein expression (CD marker list, Supplementary Data Sheet S2 ), focusing on molecules associated with DC function such as activation, migration, and T cell co-stimulation ( Figure 4C ). We observed that the majority of genes related to CD markers were downregulated, but detected strong upregulation of three genes: CD69, which mediates tissue retention; CD99, involved in diapedesis of monocytes and dendritic cells into inflamed tissue (103); and ANPEP (encoding CD13), which mediates DC cross-presentation in human DC populations (104). Analysis of the downregulated genes revealed suppression of markers associated with classical DC activation and T cell co-stimulatory ligands (CD83, CD86, CD58) and molecules important for migration of DCs (PECAM1, CD37) (105, 106), indicating potential suppression of T cell interaction molecules and retention of genital DCs within the mucosa at early timepoints. We also observed decreased expression of disintegrin and metalloproteinase family of genes (ADAM10, ADAM17) which mediate shedding of TNF-α (107), suggesting suppression of classical pro-inflammatory responses. Additionally, relative to the control, the HIV-stimulated DCs had decreased expression of CLEC12A, and genes encoding the inhibitory leukocyte immunoglobulin-like receptor subfamily B (LILRB1, LILRB2, LILRB3), suggesting a non-classical mechanism of DC activation (108).

Next, we performed Gene Ontology (GO) analysis of genes upregulated and downregulated by HIV stimulation (FDR ≤ 0.05) ( Figure 4D ; Supplementary Data Sheet S1 ). Analysis of the upregulated genes revealed an enrichment of terms associated with cytokine mediated signaling pathways, innate sensing, and metabolic processes ( Figure 4D ; Supplementary Data Sheet S1 ). Downregulated genes ( Figure 4D ) were associated to terms related to antigen presentation, T cell responses, regulation of classical inflammatory cytokines, type I IFN responses, and canonical proinflammatory pathways, but also suppression of viral cycle and viral replication. These GO terms suggest suppression of classical inflammation but activation of innate host antiviral responses within 30 minutes of exposure to HIV.

Based on these GO terms, we further explored the secretory response to HIV and compared expression of genes encoding cytokines, chemokines and antimicrobials between control and HIV conditions ( Figure 4E ). In response to HIV we detected upregulation of an innate secretory response, including pro-inflammatory genes IL1B and CXCL10, CXCL8 and CCL2, chemoattractant for neutrophils, DCs and monocytes; and CCL8, a CCR5 ligand with antiviral activity, (109). Compared to the control group, we observed decreased expression of genes encoding inflammatory cytokines IL6 and IL18, which act as Th1 polarizing cytokines; IL16, a CD4⁺ cell specific chemoattractant; CCL28, a CCR10⁺ T cell chemoattractant; and CCL21, which mediates homing to lymphoid tissues, suggesting potential inhibition of CCR7+ DC migration towards lymph nodes. Comparison of antimicrobial peptide gene expression ( (110); Supplementary Data Sheet S2 ) revealed increased expression of RNASE1 and S100A12 in the HIV group compared to control ( Figure 4E ). Interestingly, we observed suppression of SAA1 (encoding serum amyloid A), which contributes to induction of pathogenic Th17 cells (111), known to be preferential targets for HIV infection (6, 112, 113). Additionally, we observed suppression of ADM (encoding the anti-bacterial peptide adrenomedullin), which also acts as suppressor of inflammatory cytokines (114).

Taken together, our data indicates that mononuclear phagocytic populations respond to HIV in a rapid manner within 30 minutes of exposure, with upregulation of endogenous host restriction factors, activation of non-classical inflammatory responses, and increased gene expression of cytokines, chemokines and antimicrobials, suggesting initiation of a localized antiviral response.

Transcriptional changes in genital mononuclear phagocytic populations exposed to HIV indicates rapid activation of endogenous host restriction factors and localized non-classical inflammatory response, but also an overall suppression of gene transcription. Since our results identified distinct DC subsets with differential homeostatic activation and functions, we next determined how each of these subsets contributes to the overall response after HIV challenge. To define cell clusters and determine their responses following HIV exposure, we used the CITEseq libraries generated from single-cell suspensions stimulated with HIV for 30 minutes shown in Figure 4. We performed PCA, followed by unbiased clustering and visualization using UMAP. This analysis revealed four clusters ( Figure 5A ): CD14 DCs, cDC2s, infMons and NCMs, consistent with our findings under homeostatic conditions ( Figure 1B ). Overlay of unstimulated and HIV exposed cells, revealed contribution from control and HIV stimulation conditions to each cluster, except for the NCM cluster, where we found a very limited number of cells from the HIV stimulated condition, and were therefore unable to analyze NCM response to HIV ( Figure 5B ).

First, we utilized the overall transcriptional signature detected after HIV exposure shown in Figure 4B to perform AUCell analysis and determine whether this signature was enriched in specific genital DC subsets ( Figure 5C ). Interestingly, the overall HIV response signature (including upregulated and downregulated genes) was preferentially enriched in cDC2s and infMons. However, we observed enrichment of HIV upregulated genes in CD14 DCs, whereas HIV downregulated genes were enriched in cDC2s and infMons, suggesting that HIV challenge preferentially suppresses transcriptional changes in cDC2s and infMons compared to CD14 DCs. Furthermore, we detected an enrichment of cytokine, chemokine and antimicrobial peptide signatures within the CD14⁺ DC population, but no differences in enrichment of antiviral ISGs ( Figure 5C ), suggesting that CD14⁺ DCs largely mediate innate secretory response to HIV at early stages.

We next analyzed differential gene expression within each subset to identify transcriptional changes induced by HIV exposure ( Figure 5D ; Supplementary Data Sheet S1 ). All three subsets shared upregulation of AF033819.3 (the viral transcript from viral input), SOCS3, and CXCL10 ( Figure 5D ). CD14 DCs and cDC2s shared expression of anti-viral genes IRF1, IFITM2 and ACOD1 and the cytokine signaling genes CISH, IL1A and IL1B. CD14 DCs and infMons shared expression of zinc finger proteins and non-protein coding genes, except for CSF3 (encoding the granulocyte colony stimulating factor (G-CSF)) ( Supplementary Data Sheet S1 ). cDC2s and infMons both upregulated CXCL8 expression upon HIV stimulation.

Next, we analyzed genes uniquely upregulated by each subset in response to HIV. We observed that CD14 DCs upregulated TIFA and FOSL1, genes associated with initiation of inflammation; EMP1, previously described in type I interferon stimulated DCs (115); and CDKN1A, a gene involved in p53 transcription previously shown to be protective at early stages of HIV infection (116). cDC2s uniquely upregulated ISGs IFIT2 and IFIT3, both possessing antiviral functions, CCL3, CCL4 and CCL4L2, encoding chemokines capable of binding HIV receptors (133, 134) and NLRP3, encoding inflammasome protein and indicating rapid induction of inflammation. infMons upregulated RNASE1 (encoding RNA degrading secreted protein) (117), and LIFR, which encodes CD118 described to function as an alternative receptor for HIV (118).

Downregulated genes were more abundant, and we focused our analysis on genes associated with DC function and HIV pathogenesis ( Figure 5D ; Supplementary Data Sheet S1 ). All 3 subsets shared lowered expression of genes associated with HIV transcription such as GRIP1 and SRRM2, which promote HIV replication, and interestingly SAMHD1, which plays a role in preventing replication of HIV. Additionally, all three subsets showed decreased expression of inflammation associated genes JUN, CSF3R and IL6R. CD14 DCs and infMons shared downregulation of HSPA1L (encoding a heat shock protein) and genes associated with adaptive immunity such as CD36, ITGAL, LILRA1 and SIRPB1. CD14 and cDC2s shared suppression of HIV-replication promoting gene XRN1, and interferon signaling associated genes TRIM2, TRIM4, CIITA and MX1. infMons and cDC2s shared downregulated expression of genes associated with unfolded protein response, such as EDEM1, HYOU1, KLHDC3 and KAT2A (119). CD14 DCs uniquely downregulated inhibitory genes CLEC12A, LILRB1, and classical DC activation marker CD83, suggesting non-canonical activation of this subset. We also observed downregulation of HIV binding and nuclear localization protein encoding gene DDX21 (120), suggesting inhibition of nuclear import. cDC2s downregulated genes associated with promoting HIV replication such as EIF4A1, EIF4A2 (HIV translation initiation) (121) and RNA polymerase II encoding genes POLR2F, POLR2G (122). We also observed downregulation of the TNFα degrading gene ADAM17 and A2M encoding alpha2-macroglobulin, that binds to inflammatory interleukins with high affinity, suggesting inhibition of HIV replication and promoting inflammatory environment. infMons downregulated expression of NOD2, CD80, and IFNAR indicating decreased cell activation.

We performed pathway analysis using Reactome to gain information about biological interactions (50) ( Figure 5E ; Supplementary Data Sheet S1 ). Analysis of upregulated genes revealed shared involvement of CD14 DCs and cDC2s in cytokine signaling pathways (IL-10, IL-4/IL-13 pathways). cDC2s uniquely upregulated Type I IFN and IFN-γ signaling, while CD14 DCs uniquely upregulated transcriptional activation of p21, a known restriction factor against HIV infection in monocyte-derived DCs (123). Analysis of downregulated genes revealed shared signatures between CD14⁺ DCs and infMons for heat shock response, and a unique signature for infMons related to downregulation of viral replication.

Finally, to validate our findings of a secreted response at the protein level, we purified genital CD14 DCs by magnetic bead selection and incubated them with HIV as described in methods. Supernatants harvested 3 hours post stimulation from unstimulated and HIV exposed cells were assayed to quantify secreted proteins ( Figure 5F ). In response to HIV, we detected increased secretion of cytokines and chemokines (IFN-γ, IL-5, IL-13, IL1Rα, GM-CSF, CXCL1, CXCL10 and CCL22). These results confirm rapid innate secretory response in response to HIV.

Overall, we demonstrate subset-specific responses to HIV, with a preferential suppression of genes in cDC2s and infMons compared to CD14 DCs, with CD14⁺ DCs largely mediating the secretory innate response after HIV stimulation.