Lymph node dendritic cells harbor inducible replication-competent HIV despite years of suppressive ART

Abstract

Although gut and lymph node (LN) memory CD4 T cells represent major HIV and simian immunodeficiency virus (SIV) tissue reservoirs, the study of the role of dendritic cells (DCs) in HIV persistence has long been limited to the blood due to difficulties to access lymphoid tissue samples. In this study, we show that LN migratory and resident DC subpopulations harbor distinct phenotypic and transcriptomic profiles. Interestingly, both LN DC subpopulations contain HIV intact provirus and inducible replication-competent HIV despite the expression of the antiviral restriction factor SAMHD1. Notably, LN DC subpopulations isolated from HIV-infected individuals treated for up to 14 years are transcriptionally silent but harbor replication-competent virus that can be induced upon TLR7/8 stimulation. Taken together, these results uncover a potential important contribution of LN DCs to HIV infection in the presence of ART.

Keywords: HIV; HIV cure; HIV persistence; HIV reservoir; HIV single-genome sequencing; T follicular helper cells; dendritic cells; intact HIV; lymph node; replication-competent virus.

Figure 1.. Characterization of LN resident and migratory DCs

(A) Ex vivo transcriptomic uniform manifold approximation and projection (UMAP) profile of LN resident and migratory DCs from aviremic ART-treated HIV-infected individuals (N = 3). (B) Heatmap representing the top 20 differentially expressed genes within LN resident and migratory DCs. (C–H) Violin plots showing the normalized expression level of differentially expressed genes between LN resident and migratory DCs. (I) Percentages of cDC1, cDC2, and DC3-like cells within LN resident and migratory DCs of HIV-infected individuals (N = 3). (J and K) (J) Stimulatory potential of LN resident or migratory DCs as assessed by mixed leukocyte reaction (MLR) assay or using HIV-specific stimulation (K). Individuals are color coded (K). Bars correspond to mean (C–I and K). Red bars correspond to mean ± SEM (I–K). Dashed lines represent the cutoff for positivity (J and K). Red stars indicate statistical significance (*p < 0.05) (C–H). Statistical significance of cDC2-type LN resident (blue stars) or migratory (purple stars) DCs compared with cDC1 and DC3-type LN DCs (*p < 0.05) (I). Statistical significance of LN resident (green stars) or migratory (pink stars) DCs:CD4 T cell MLR cultures at various DC:T cell ratios compared with CD4 T cells alone (*p < 0.05) (J). Statistical significance was assessed with the non-parametric Wilcoxon rank sum test implemented in the ‘‘FindMarkers’’ function of the Seurat package (C–H) or using one-way ANOVA followed by either multiple comparison test (I) or Wilcoxon matched-pairs two-tailed signed rank test (J and K). ‘‘Res DC’’ corresponds to LN resident DCs and ‘‘Mig DC’’ corresponds to LN migratory DCs.

Figure 2.. Phenotypic profile of LN resident and migratory DCs by mass cytometry

(A) t-SNE plots depicting the surface expression patterns of CD4, CCR5, CXCR4, DC SIGN, CD169, and CD163 directly ex vivo on naive CD4 T cells, LN migratory and LN resident DCs from a treated HIV-infected individual (#161). (B) Representative mass cytometry profile of BIRC3 and SAMHD1 expression levels in naive CD4 T cells, LN Tfh cells, LN resident, and LN migratory DCs isolated directly ex vivo from a representative HIV-uninfected, viremic (#1009), and aviremic ART-treated HIV-infected individual (#163). (C) Frequencies of LN resident and LN migratory DCs expressing CD4, CCR5, CXCR4, DC-SIGN, CD169, and CD163 or mean signal intensity (MSI) of SAMHD1 and BIRC3 in naive CD4 T cells, LN resident, and LN migratory DCs isolated directly ex vivo from HIV-uninfected (N = 4), viremic (N = 6) and ART-treated HIV-infected individuals (N = 7). Red bars correspond to mean ± SEM (C). Red stars indicate statistical significance (*p < 0.05) for intra-group comparisons, i.e., LN migratory DCs versus LN resident DCs (C). Green stars indicate statistical significance (*p < 0.05) for intra-group comparisons of LN migratory DCs, LN resident DCs, and naive CD4 T cells (C). Statistical significance was obtained using one-way ANOVA (Kruskal-Wallis test) followed by Wilcoxon matched-pairs two-tailed signed rank test.

Figure 3.. LN resident and migratory DCs are susceptible to HIV infection in vitro

(A and B) Representative flow cytometry profiles of LN resident and migratory DCs isolated from a representative HIV-uninfected individual showing their relative susceptibility to (A) CCR5-tropic HIV-derived vector encoding for EGFP in the presence or absence of SAMHD1-modulating factor, i.e., Vpx at day 4 post-exposure or (B) to CXCR4-tropic HIV-derived vector encoding for EGFP. (C) Percentages of EGFP⁺ LN DCs at day 4 post-exposure to either CCR5-tropic (N = 12) or CXCR4-tropic (N = 7) HIV-derived vectors. Activated LN CD4 T cells were added as controls (N = 5). (D) Percentage of EGFP⁺ LN DCs at day 4 post-exposure to CCR5-tropic HIV-derived vector encoding for EGFP in the presence or absence of Vpx (N = 9). (E) Levels of HIV RNA (copies/mL) in culture supernatants of unstimulated LN CD4 T cells, activated LN CD4 T cells, LN migratory DCs, and LN resident DCs of HIV-uninfected individuals exposed to HIV Ba-L (N = 3). (F–I) Levels of HIV RNA (copies/mL) in culture supernatants of unstimulated LN CD4 T cells (F), activated LN CD4 T cells (G), LN migratory DCs (H), and LN resident DCs (I) of HIV-uninfected individuals exposed to HIV Ba-L in the presence or absence of emtricitabine (N = 3). Red bars correspond to mean ± SEM (C–I). Red stars indicate statistical significance (*p < 0.05) of intra-group comparisons (C and D) or of conditions cultured without emtricitabine compared with conditions with emtricitabine (G and I). Black, gray, and violet stars indicate statistical significance (*p < 0.05) of activated CD4 T cells, LN resident, and migratory DCs, respectively, compared with the levels of unstimulated CD4 T cells at the corresponding time points (E). Wells with detectable HIV-1 RNA (≥2,000 HIV-1 RNA copies/mL) were referred to as HIV-1 RNA-positive wells (E–I). Gray dashed lines (E–I) represent the limit of detection. Statistical significance (p values) was obtained using one-way ANOVA (Kruskal-Wallis test) followed by Wilcoxon matched-pairs two-tailed signed rank test (C–E) or following a one-tailed ratio paired t test (F–I). “Res DC” corresponds to LN resident DCs and “Mig DC” corresponds to LN migratory DCs.

Figure 4.. LN resident and migratory DCs are enriched with genome-intact HIV DNA

(A) Maximum-likelihood phylogenetic trees of all HIV-1 sequences obtained from LN resident and migratory DCs of the 3 chronic viremic HIV-infected individuals (#1024, #1025, and #1019; 8–9 kb amplicons) as assessed by FLIP-seq and MIP-seq assays. Sequences from LN PD-1⁺/Tfh and LN PD-1⁻/non-Tfh cells are also depicted. Chromosomal integration site coordinates for the respective sequences are indicated. (B) Pie charts reflecting the relative proportion of intact and defective HIV-1 sequences. (C and D) Frequencies of cells harboring either intact (C) or defective (D) HIV-1 sequences (cells/million) as assessed by both FLIP-seq and MIP-seq assays (2 assessments per individual). HIV-infected individuals are color coded (C and D). LN cell populations are color coded (C and D). Histograms represent minimum to maximum of the range and the line corresponds to the median (C and D). Red bars correspond to 95% confidence interval. “Res DC” corresponds to LN resident DCs and “Mig DC” corresponds to LN migratory DCs.

Figure 5.. LN DCs of viremic individuals are HIV-infected, transcriptionally active, and produce HIV upon TLR7/8 in vitro stimulation

(A–F) LN resident and migratory DCs were isolated from untreated viremic HIV-infected individuals. (A) Levels of integrated HIV DNA (levels/million cells) within LN resident and migratory DCs (N = 5). (B) Levels of cell-associated unspliced HIV gag RNA within LN resident and migratory DCs (N = 5). (C and D) (C) Frequencies of cells containing unspliced HIV gag RNA (cells/million) (C) and multi-spliced tat-rev RNA (cells/million) (D) in LN resident and migratory DCs (N = 5). (E and F) (E) Levels of HIV production in LN resident and migratory DC culture supernatants at day 7 post-stimulation with TLR7/8 agonist as assessed by HIV-RNA levels (copies/mL) (E) or HIV P24 levels (pg/mL) (F) (N = 6). Autologous LN PD-1⁺/Tfh and LN PD-1⁻/non-Tfh cells were used as controls (A–F). LN PD-1⁺/Tfh and LN PD-1⁻/non-Tfh cells were stimulated or not with anti-CD3/anti-CD28 mAbs for 3 days (E and F). Dashed line represents the limit of detection (A–F). HIV-infected individuals are color coded (A–D). Red bars correspond to mean ± SEM (A–F). Red stars indicate statistical significance (*p < 0.05). Statistical significance (p values) was obtained using one-way ANOVA (Kruskal-Wallis test) followed by Wilcoxon matched-pairs one-tailed signed rank test (E and F). ‘‘Res DC’’ corresponds to LN resident DCs and ‘‘Mig DC’’ corresponds to LN migratory DCs.

Figure 6.. HIV-infected LN DCs containing replication-competent and infectious HIV are still detectable despite years of ART

(A–N) LN resident and migratory DCs were isolated from treated aviremic HIV-infected individuals. (A) Frequencies of LN DCs harboring integrated HIV DNA (cells per million) directly ex vivo (N = 4). (B) Frequencies of LN DCs harboring cell-associated unspliced HIV gag RNA or multi-spliced tat-rev HIV RNA (cells per million) directly ex vivo (N = 3). (C) Levels of HIV RNA (copies/mL) produced in culture supernatants of unstimulated VOA conditions (N = 3). (D) Levels of HIV RNA (copies/mL) produced in culture supernatants of stimulated VOA conditions (N = 9). (E) Proportion of HIV-infected individuals with detectable HIV RNA (≥20 HIV RNA copies/mL) in stimulated VOA conditions at day 14 (N = 9). (F) Levels of HIV RNA (copies/mL) produced in the culture supernatants at day 14 in the stimulated VOA conditions (N = 9). (G) Correlation between duration of ART (in years) and levels of HIV RNA (copies/mL) produced in the stimulated VOA culture supernatants at day 14 (N = 9). (H–J) (H) Estimated decay slopes for HIV-RNA levels in stimulated VOA culture supernatants of LN CD4 T cells (H), LN migratory DCs (I), or LN resident DCs (J) with the duration of suppressive therapy. (K) Proportion of HIV-infected individuals with detectable HIV-1 RNA (≥20 HIV RNA copies/mL) in LN cell populations at day 14 of the in vitro HIV-1 infection assay (N = 9). (L) Levels of HIV RNA (copies/mL) in culture supernatants at day 14 in the in vitro infection assay (N = 9). (M) Schematic representation of the QVOA preformed on LN resident and migratory DCs. HIV-infected individuals were considered as replicates at 40,000 cell concentration (N = 9). (N) Estimated mean frequencies of LN resident and migratory DCs harboring inducible replication-competent virus/RNA units per million (RUPM) cells. Dashed line represents the limit of detection (A–D), (F), and (L). Undetectable values were arbitrarily defined as 10 HIV RNA copies/mL (C), (D), (F–J), and (L). HIV-infected individuals were color coded (A), (B), (F), and (L). Histograms correspond to the estimated mean (E), (F), (K), (L), and (N). Red bars correspond to mean ± SEM (D), (F), and (L) or to lower and upper confidence interval at 0.95 (N). Red stars indicate statistical significance (*p < 0.05) (D–F) and (K). Statistical significance (p values) was obtained using one-way ANOVA (Kruskal-Wallis test) followed by Wilcoxon matched-pairs one-tailed signed rank test (D) and (F) or using Spearman rank test for correlations (G) or using chi-square test for proportion of positive individuals (E) and (K). ‘‘Estimate’’ refers to the decay slopes for LN CD4 T cells (H), LN migratory DCs (I), or LN resident DCs (J). Red color wells correspond to HIV-RNA-negative wells, whereas green wells correspond to HIV-RNA-positive wells (M). ‘‘Res DC’’ corresponds to LN resident DCs and ‘‘Mig DC’’ corresponds to LN migratory DCs.