HIV-1 replicates and persists in vaginal epithelial dendritic cells

HIV-1 acquisition occurs most commonly after sexual contact. To establish infection, HIV-1 must infect cells that support high-level replication, namely CD4+ T cells, which are absent from the outermost genital epithelium. Dendritic cells (DCs), present in mucosal epithelia, potentially facilitate HIV-1 acquisition. We show that vaginal epithelial DCs, termed CD1a+ VEDCs, are unlike other blood- and tissue-derived DCs because they express langerin but not DC-SIGN, and unlike skin-based langerin+ DC subset Langerhans cells (LCs), they do not harbor Birbeck granules. Individuals primarily acquire HIV-1 that utilizes the CCR5 receptor (termed either R5 or R5X4) during heterosexual transmission, and the mechanism for the block against variants that only use the CXCR4 receptor (classified as X4) remains unclear. We show that X4 as compared with R5 HIV-1 shows limited to no replication in CD1a+ VEDCs. This differential replication occurs after fusion, suggesting that receptor usage influences postentry steps in the virus life cycle. Furthermore, CD1a+ VEDCs isolated from HIV-1–infected virologically suppressed women harbor HIV-1 DNA. Thus, CD1a+ VEDCs are potentially infected early during heterosexual transmission and also retain virus during treatment. Understanding the interplay between HIV-1 and CD1a+ VEDCs is important for future prevention and cure strategies.

Viruses. Plasmids containing infectious molecular clones (IMC) (YU-2, NL4-3, RHPA, and WARO), expression plasmid (VSV-G, CCR5-tropic HIV gp160 (Bal Envelope), and the CXCR4tropic HIV gp160 (Lai Envelope)), pMM310, containing the BlaM-Vpr construct, HIV-1 Lai∆envluc (Envelope deficient HIV-1 Lai containing a luciferase reporter gene in place of the nef), and pSIV3 + (SIVmac239 Vpx encoding SIV packaging vector) were obtained from the NIH AIDS Reference Reagent Program or described previously (4)(5)(6). SF2 virus stocks were obtained from NIH AIDS Reference Reagent Program (7,8). Virus stocks were generated by transient transfection of 293T cells either with an IMCs or envelope deleted plasmid along with an envelope expression plasmid. BlaM-Vpr containing virus stocks were generated by also cotransfecting pMM310. Vpx-containing virus like particles were produced by the co-transfection of 293T cells with pSIV3 + and VSV-G expression plasmid. All virus containing supernatants were clarified using low speed centrifugation and filtered through a 0.45 um pore filters. BlaM-Vpr containing viruses and luciferase reporter viruses were pelleted through sucrose cushion gradients. The number of infectious particles was estimated on TZM-bl cells and capsid content of HIV-1 was determined by a p24 gag ELISA as previously described (9)(10)(11).
Infections. CD1a+ VEDCs, skin derived LCs and activated or non-activated lamina propria cells were exposed to virus at a multiplicity of infection (MOI) ranging from 0.1 -15 as indicated. In the infections with protease inhibitor (Saquinavir 1µM), the protease inhibitor was maintained in cultures at all times. Exposed cells were washed a minimum of 4 times to remove unbound virus. CD1a+ VEDCs and skin derived LCs were washed after 72 hours and lamina propria cells were washed after 2 hours of exposure. CD1a+ VEDCs, skin derived LCs and nonactivated lamina propria cells were cultured in fresh RPMI complete, while activated lamina propria cells were cultured in RPMI complete with 20 u/ml of recombinant IL-2 (r-IL2). For the co-cultures, exposed and subsequently washed CD1a+ VEDCs were co-cultured with activated lamina propria cells in RPMI complete with 20 u/ml of r-IL2. Infections were monitored by removing around 50 ul of culture supernatant every 2 days and replacing with fresh media.
Infectious virus concentration was estimated by infecting 1 X 10 4 TZM-bl cells with 8 serial twofold dilutions of supernatant culture starting at 50 ul. All infections were done in duplicate in a 96 well format. Two days post-infection, TZM-bls were examined for beta-galactosidase production using Galacto-Light Plus System (Applied Biosystems). In every experiment, background level was estimated as the average of the relative light unit (RLU) from wells with TZM-bl alone plus 1 standard deviation. This background was subtracted from all wells, and wells with negative RLU were assigned a value of 0. Infections with protease inhibitor were monitored by estimating HIV RNA in culture supernatants. Briefly, RNA was isolated from culture supernatants using viral RNA isolation kit (Qiagen). HIV RNA was quantified using a previously published sensitive quantitative PCR assay with minor modifications (12).
Fusion Assay. Fusion assay was done as previously described (13). Around 5 X 10 4 to 1 X 10 5 CD1a+ VEDC were infected with BlaM-Vpr-containing virions (200 to 500 ng of p24 Gag ) for 3 h at 37°C in 100 μl of RPMI medium. After infection, cells were then washed once with CO2independent RPMI medium and loaded with CCF2 dye (0.5 mM; Invitrogen). Cells were incubated for 18 h at room temperature in 200 μl of CO2-independent RPMI medium supplemented with 10% FBS and 2.5 mM probenecid. Infected CD1a+ VEDCs were washed once in staining buffer (phosphate-buffered saline [PBS] with 2% FBS) and incubated for 30 min at room temperature with a CD1a directed antibody. The cells were next washed twice with PBS containing 2% FBS. Cells were fixed with 100 µl 4% PFA for 30 min at RT. The change in emission fluorescence of CCF2 after cleavage by the BlaM-Vpr chimera was measured by flow cytometry with a LSRII flow cytometer (Becton Dickinson). Data were collected using FACSDiva software (Becton Dickinson) and analyzed with FlowJo software (TreeStar). The compensation was calculated after data collection on the basis of single-stain controls using FlowJo software.
Measuring HIV DNA intermediates. Late transcription products and integrated proviruses were measured as detailed previously (14). Briefly, DNA from virus exposed cells was isolated using Qiagen DNA Isolation kit. Late reverse transcription and integration products were quantified using quantitative PCR with primers and probes that anneal to LTR -gag and human Alugag regions respectively as described previously (14,15). The number of HIV copies was determined using pNL4-3 as a standard. HIV molecules were normalized to copies per cell by albumin quantitative PCR. All samples were measured 3 independent times.
Transcription assay. CD1a+ VEDCs were exposed to 200ng of p24 gag virus inoculums in the presence or absence of 10uM Maraviroc (MVC), a CCR5 inhibitor or 10uM AMD3100, a CXCR4 blocker. Pseudovirions incorporated the various envelopes with envelope deleted Lai and luciferase gene in the place of viral accessory gene nef. After 3 days, luciferase activity was measured, in duplicate, using Bright-Glo Luciferase Assay System (Promega) in the cultures.
HIV DNA quantitation and envelope sequences. Cells were isolated from HIV-1 infected vaginal tissue and lamina propria using similar techniques as above. Peripheral blood mononuclear cells (PBMCs) were isolated using Ficoll Hypaque density centrifugation.
Quantitative PCR was used to estimate number of HIV-1 DNA copies and number of cells as previously described (12). All estimations were done in triplicate a minimum of 2 independent times. Envelope sequences were isolated using single genome amplification as described previously (16). Sequences were aligned using HIVAlign (https://www.hiv.lanl.gov/content/sequence/VIRALIGN/viralign.html) and alignments were manually edited in BioEdit. Maximum likelihood tree was generated using PhyML (https://www.hiv.lanl.gov/content/sequence/PHYML/interface.html) with HKY85 model. Color and symbol coded phylogenetic tree was generated using RainbowTree (https://www.hiv.lanl.gov/content/sequence/RAINBOWTREE/rainbowtree.html). All unique sequences reported in this publication have been submitted to Genbank (accession numbers MG940909-MG940950). Envelopes were incorporated into an envelope deficient NL4-3 HIV-1 backbone using yeast gap-repair homologous recombination as described previously (16).
Virus stocks were generated from 293T transfection and subsequent PBMC passage. Coreceptor usage was determined as described previously (16). Briefly, TZM-bls were exposed to virus stocks in the presence or absence of co-receptor inhibitors. Statistically significant difference in infection levels in the presence as compared to absence of inhibitors was used to determine the co-receptor usage. All conditions were tested in triplicate, and all assays were conducted with internal controls YU2 (R5), NL4-3 (X4), and 89.6 (R5X4).
T cell receptor re-arrangement quantitative assay. Cell preparations were assessed for frequency of re-arranged TCRG using a previously described quantitative PCR assay (17).