ABSTRACT
A robust simian-human immunodeficiency virus (SHIV)-macaque model of latency is critical to investigate eradicative and suppressive strategies that target HIV-1 Env. To this end, we previously reported a novel strategy for constructing SHIVs that bear primary or transmitted/founder (TF) Envs with modifications at Env residue 375 that enable efficient replication in Indian rhesus macaques (RM). Such TF SHIVs, however, have not been examined for their suitability for HIV-1 latency and cure research. Here, we evaluate two promising TF SHIVs, SHIV.D.191859 and SHIV.C.CH848, which encode TF subtype D and C HIV-1 Envs, respectively, for their viral kinetics and persistence during suppressive combination antiretroviral therapy (cART) and treatment interruption in RM. Our results suggest that the viral kinetics of these SHIVs in RM during acute, early, and chronic infection, and upon cART initiation, maintenance and discontinuation, mirror those of HIV-1 infection. We demonstrate consistent early peak and set point viremia, rapid declines in viremia to undetectable plasma titers following cART initiation, infection of long-lived cellular subsets and establishment of viral latency, and viral rebound with return to pretreatment set point viremia following treatment interruption. The viral dynamics and reservoir biology of SHIV.D.191859, and to a lesser extent SHIV.C.CH848, during chronic infection, cART administration, and upon treatment interruption suggest that these TF SHIVs are promising reagents for a SHIV model of HIV-1 latency and cure.
IMPORTANCE Simian-human immunodeficiency viruses (SHIVs) have been successfully used for over 2 decades to study virus-host interactions, transmission, and pathogenesis in rhesus macaques. The majority of Env trimers of most previously studied SHIVs, however, do not recapitulate key properties of transmitted/founder (TF) or primary HIV-1 isolates, such as CCR5 tropism, tier 2 neutralization resistance, and native trimer conformation. Here, we test two recently generated TF SHIVs, SHIV.D.191859 and SHIV.C.CH848, which were designed to address these issues as components of a nonhuman primate model of HIV-1 latency. We conclude that the TF SHIV-macaque model reflects several hallmarks of HIV and SIV infection and latency. Results suggest that this model has broad applications for evaluating eradicative and suppressive strategies against the HIV reservoir, including Env-specific interventions, therapeutic vaccines, and engineered T cells.
INTRODUCTION
Chimeric simian-human immunodeficiency viruses (SHIVs), which encode an HIV-1 envelope glycoprotein (Env) within the background of a macaque-adapted simian immunodeficiency virus (SIVmac) backbone, have wide-ranging applications for evaluating HIV pathogenesis, vaccine, and cure strategies (1). SHIVs have been successfully used for over 2 decades to study virus-host interactions, transmission, and pathogenesis in widely used nonhuman primate (NHP) models, including rhesus macaques (RM) (1–4). SHIVs have particular importance as reagents in studies testing strategies that engage Env, such as broadly neutralizing antibodies (bNAbs) and antibody-based vaccines, and in dissecting the immunology behind antibody elicitation (5–8).
Previous iterations of SHIVs, however, have several limitations. Many SHIVs do not accurately recapitulate key properties of transmitted/founder (TF) or primary HIV-1 isolates from early infection, such as CCR5 tropism (1, 9–11), infection and depletion of relevant long-lived T cell subsets (3, 4), and induction of consistent autologous tier 2 antibody responses (3, 4, 12, 13). Furthermore, most SHIVs do not fully reflect the original conformation and neutralization sensitivity of native HIV-1 Env as a result of adaptations acquired through extensive in vitro and in vivo passage required to enable efficient infection of RM cells (1–3). A small number of SHIVs encoding primary Envs have been shown to fortuitously replicate in RM without adaptation; none of these SHIVs, however, have been characterized for persistence before, during, and after antiretroviral therapy (14).
Recently, we reported the development of a novel strategy to generate designer SHIVs encoding native TF or primary Envs with tier 2 neutralization sensitivity, referred to here as “TF SHIVs” (15). In this method, a single amino acid substitution at Env position 375 mediates increased affinity of HIV-1 Env for rhesus CD4, enabling efficient in vitro replication in rhesus cells and in vivo infection of RM while preserving the antigenic properties of the parental Env. TF SHIVs consistently confer productive infection through multiple inoculation routes, high peak viremia, and desirable early viral kinetics. Importantly, TF SHIVs are CCR5 tropic and have been shown to reproducibly induce autologous antibody responses in infected RM, thereby recapitulating key qualities of primary and TF HIV-1 isolates (15, 16). To date, TF SHIVs have not been evaluated for their suitability for studies of HIV cure and latency.
Here, we evaluate two genetically and antigenically diverse TF SHIVs, SHIV.D.191859 and SHIV.C.CH848, which encode TF subtype D and C HIV-1 Envs, respectively, for their viral kinetics and persistence during suppressive combination antiretroviral therapy (cART) and treatment interruption in RM. Our results indicate that the kinetics of SHIV.D.191859 and SHIV.C.CH848 infection and replication in RM during acute, early, and chronic infection and upon cART initiation and discontinuation mirror those of HIV-1 infection. The findings presented thus provide an experimental framework with which to design future preclinical studies of HIV-1 latency, cure, and immunopathogenesis.
RESULTS
SHIV.D.191859 and SHIV.C.CH848 early viral kinetics.Two TF SHIVs with desirable viral kinetics, SHIV.D.191859 and SHIV.C.CH848, here referred to as SHIV.D and SHIV.C, were evaluated for their suitability for latency and cure studies. SHIV.D encodes a subtype D TF Env identified from an acutely infected Ugandan woman in 2008 (17). The HIV-1 TF virus was shown to be CCR5 tropic, with a tier 2 neutralization profile and the capacity to replicate in both CD4 T cells and macrophages (17). SHIV.C encodes a clade C Env identified in an acutely infected Malawian man in 2008 (18). Subsequently, infection with this virus was shown to elicit N332-dependent V3-targeting bNAbs after 3.5 years of infection. A strategy for eliciting CH848-like bNAbs in humans via vaccination has been described (18); furthermore, similar bNAbs have been elicited in SHIV.C.CH848-infected RM (16). Prior to the present study, SHIV.D and SHIV.C had each been used to infect 6 non-CD8-depleted RM, with mean peak viral loads (VLs) of 106–7 copies/ml for both SHIVs; 10 of 12 SHIV.C- and SHIV.D-infected RM maintained viremia of >103 copies/ml through 24 weeks postinfection (WPI) (15).
SHIV.D.191859 persists on suppressive cART.To evaluate SHIV.D infection after early time points and its response to cART, 6 RM were intravaginally challenged with increasing doses of SHIV.D until infection was confirmed by repeated positive plasma viral load (VL) measurements. Two RM demonstrated infection after the first challenge and 2 additional RM became productively infected after the third challenge (Fig. 1). Two RM remained uninfected after 4 intravaginal challenges and were subsequently inoculated intravenously, demonstrating productive infection at day 7. Peak viremia ranged from 9.7 × 105 to 4.2 × 107 copies/ml, and levels did not appear substantially different between animals infected by mucosal versus intravenous routes (mean peak VL = 1.4 × 107 copies/ml versus 2.1 × 106 copies/ml, respectively).
SHIV.D.191859 viremia over time. (A) Plasma viremia (copies/ml, y axis) is depicted longitudinally with week postinfection indicated on the x axis. Mucosal inoculation is indicated by black arrows; intravenous inoculation is indicated by a red arrow. Euthanasia is indicated by black crosses. Time of cART administration (24 weeks) is shown as a shaded gray box. (B) Time to virus suppression after cART initiation and time to detectable plasma viremia after ATI are shown for each RM.
RM were followed longitudinally through early infection and establishment of a relatively stable viral set point (Fig. 1). In all 6 RM, viremia of >103 copies/ml was maintained for at least 24 weeks, with set point viral loads ranging from 103 to 106 copies/ml. After 42 WPI, RM FT42 spontaneously controlled plasma viremia to below the assay limit of quantification (LOQ; 83 copies/ml) and was excluded from further analysis. RM DE33 was euthanized at 66 WPI after viremia rose to greater than 106 copies/ml, coincident with signs and symptoms of severe illness; this RM was positive for the Mamu A*01 allele, which is significantly associated with frequency of spontaneous viremic control of SIV infection (19). In this small cohort, the mean set point VL did not appear to be substantially different between animals that were productively infected via mucosal or intravenous inoculation (mean set point VL = 3.0 × 105 copies/ml versus 1.6 × 105 copies/ml, respectively).
We evaluated persistence of SHIV.D on suppressive cART in the 4 remaining viremic RM using daily injectable combination antiretroviral therapy (cART) consisting of dolutegravir, emtricitabine, and tenofovir. Viral suppression of <83 copies/ml was achieved between 7 and 21 days of cART initiation and durably maintained in all RM. One incidence of detectable viremia to 804 copies/ml occurred in EJ94 16 weeks post-cART initiation. cART was discontinued after 24 weeks. Rebound plasma viremia was detected between 7 and 19 days posttreatment interruption (Fig. 1). Plasma virus rebounded to at or near the pre-cART set point VL in all RM (103 to 107 copies/ml by 4 weeks postrebound).
We next quantified the levels of cell-associated viral DNA (ca-DNA) and RNA (ca-RNA) in peripheral blood mononuclear cells (PBMCs) sampled during chronic infection in SHIV.D-infected RM prior to cART initiation (Fig. 2). The mean ca-DNA measurement in PBMC was 9.9 × 102 copies/106 cells and remained detectable after 4 months of cART suppression in all four RM, with a mean 1.9-fold decrease. The mean ca-RNA level was 5.4 × 103 copies/106 cells and similarly decreased 2.0-fold, on average, after 4 months of cART. Measurements of ca-DNA and RNA were similar after 4 and 6 months of cART.
SHIV.D.191859 quantification over time. Cell-associated DNA levels in PBMCs (A), cell-associated RNA levels in PBMCs (B), and matched plasma viral load measurements (C) are shown for all SHIV-infected RM. Pre-cART samples were obtained during chronic infection (>33 WPI); second and third sets of samples were obtained after 16 (month 4) and 23 (month 6) weeks of cART suppression, respectively. Limit of quantification (LOQ) is denoted by the dashed line. Time of cART administration (24 weeks) is shown as a gray box.
SHIV.D.191859 diversity pre- and post-cART.We next characterized the diversity of SHIV populations by longitudinal single-genome sequencing (SGS) of gp160 env sequences; 284 total sequences (median, 31 per animal time point) were generated. To characterize perimaximal SHIV diversity, plasma virus was sequenced immediately prior to cART-initiation, when the four RM had been productively infected for between 33 and 81 weeks (Fig. 3), revealing a maximum within-animal env pairwise diversity between 1.0% and 2.5%.
Phylogenetic trees of longitudinal SHIV.D.191859-infected RM env sequences. The TF SHIV.D sequence is shown in pink, pre-cART sequences are shown in black, sequences from the first and second weeks of detectable viremia after ATI are shown in blue and orange, respectively. Genetic distance is indicated by scale bar, indicating 10 nucleotides, or ∼0.3% difference. Sequences are shown from RM EJ94 (A), RM FE43 (B), RM FR55 (C), and RM GA67 (D).
To characterize the rebound virus populations reactivating from latency upon treatment interruption, sequences were generated from the first time point where plasma VL was greater than 500 copies/ml. Maximum likelihood phylogenetic trees comparing the TF SHIV.D env sequence and pre-cART and rebound plasma virus populations are shown in Fig. 3. Maximum within-animal pairwise diversity values at the time of first detectable rebound were not significantly different from pre-cART values, ranging from 0.3% to 1.9% (P = 0.25, n = 4, Wilcoxon signed-rank test).
To enumerate the minimum number of virus-infected cells that contributed to recrudescent viremia, we used a conservative estimate of the amount of genetic diversity that could accrue in the elapsed time posttreatment interruption. For these 4 RMs, the maximum diversity estimate discriminating between variants ranged between 3 and 19 nucleotides (nt) per env sequence, based on differences in time to rebound (see Materials and Methods). For example, RM EJ94 had the fastest time to rebound of the four RM (7 days). Given an estimated mutation rate of 0.76 nt per amplified viral genome per day, time to rebound, and an approximate terminal drug washout period of 4 days, we infer that the maximum divergence one reactivating lineage could accrue within env between reactivation and our sampling is two nucleotide substitutions. Thus, virus lineages that differ in env by more than two nucleotides are likely representative of distinct reactivating cells. FE43, in contrast, had the slowest time to rebound (19 days). For this RM, our model likely overestimates the maximum divergence one reactivating lineage could accrue, as virus is unlikely to be consistently replicating for 15 days prior to systemic detection. Using this diversity cutoff, rebound virus populations in the four RM grouped into 7 (FE43), 10 (FR55), and 13 (EJ94) low-diversity lineages at first detectable viral rebound. In GA67, sequences fell into either one or two closely related lineages at day 12 after analytical treatment interruption (ATI). One week later, at day 19 post-ATI, six genetically distinct low-diversity lineages were sampled in RM GA67. Thus, in all four RM, rebound viremia was founded by multiple virus-infected cell reactivations.
SHIV.D.191859 and CD4 T cell quantification in tissues.To assess the extent of systemic virus dissemination and CD4 T cell depletion in SHIV.D-infected RM, we quantified CD4 T cell percentage, total SHIV DNA, and total SHIV RNA in lymphoid and other tissues following euthanasia due to either clinical indication or study endpoint. RM DE33 and EJ94 were euthanized due to clinical disease progression at 18 and 39 months postinfection (MPI), respectively. RM FR55 and GA67 were resuppressed on cART at 39 and 34 MPI, respectively. RM FR55 was electively euthanized 6 months later at 45 MPI, while RM GA67 was euthanized due to clinical indication at 35 MPI after 1 month of cART resuppression. RMs FE43 and FT42 were euthanized per study protocol: FE43 while viremic at 40 MPI and RM FT42 while virally suppressed off cART at 36 MPI, after 25 months of spontaneous virus control. In the two RM with clinical progression coincident with high viremia, CD4 T cells were depleted and levels of HIV DNA and RNA were high throughout the body. In RM EJ94, analysis of PBMC and all lymphoid tissues, including gut-associated lymphoid tissues (GALT), spleen, and lymph nodes, revealed severe CD4 T cell depletion at necropsy (0.9% to 2.5% of total T cells) (Fig. 4A). In RM DE33, a similarly severe degree of CD4 T cell depletion was observed in GALT, including the colon and ileum (4.7% and 6.0% of total T cells, respectively) as well as in bone marrow (7.0%) (Fig. 4A). In RM with less SHIV-associated disease progression, CD4 T cell depletion was more moderate, and levels of SHIV DNA and RNA were lower throughout. Despite spontaneously suppressing plasma viremia, RM FT42 had detectable ca-DNA and RNA in multiple tissues at necropsy. Across the six RM, SHIV DNA and RNA quantification showed systemic virus dissemination in every tissue sampled from all SHIV.D-infected RM, which included GALT, lymph nodes, tonsil, and spleen (Fig. 4B and C). The mean total SHIV DNA and RNA measurements across all sampled tissues in all RM were 3.4 × 104 and 2.9 × 107 copies/106 cells, respectively (Fig. 4B and C).
SHIV.D.191859 and CD4 T cell quantification in tissues. CD4 T cell percentages (A), total SHIV DNA (B), and total SHIV RNA (C) are shown for tissues from all SHIV.D-infected RM at necropsy. Euthanasia was performed at different time points, for clinical indications or at study completion. RM DE33 and EJ94 were euthanized while viremic for clinical indication at 18 and 39 study months, respectively. RM GA67 and FR55 were euthanized while on cART, FR55 after 6 months of cART by study protocol, and GA67 after 1 month of cART due to clinical deterioration. FE43 was euthanized while viremic at study month 40. RM FT42 spontaneously controlled viremia to below the assay LOQ and was euthanized with suppressed virus off cART. MLN, mesenteric lymph node; ALN, axillary lymph node; BM, bone marrow; PLN, pharyngeal lymph node; BAL, bronchoalveolar lavage. Asterisks indicate tissue not collected or analyzed.
SHIV.C.CH848 infection pre- and post-cART.We next examined SHIV.C infection for persistence on cART. Four RM were intravenously challenged with a high-dose SHIV.C inoculum. All animals became productively infected with peak VLs ranging from 105 to 107 copies/ml (Fig. 5). In all four RM, viremia was maintained through 16 WPI, with set point VLs ranging from 103 to 105 copies/ml. After 16 weeks of infection, cART was initiated. Viral suppression of <83 copies/ml was achieved in all RM within 7 days post-cART initiation and was durably maintained for 24 weeks. cART was discontinued at 40 weeks postinfection, with recrudescent plasma viremia detected within 5 to 22 days of treatment interruption. As with SHIV.D-infected animals, RM with higher pre-cART viremia rebounded earlier (Fig. 5). Postrebound, the four RM reestablished a quasi-set point VL at ∼1 log10 to 1.5 log10 copies/ml below pre-cART levels. For KK54, this led to consistent viremia (∼104 copies/ml). For the remaining three RM, lower levels of systemic virus replication were detected post-ATI, with intermittent declines of viremia to <83 copies/ml. In KH83, the animal with the lowest viremia pre-cART, virus initially rebounded shortly post-ATI but then fell to below the limit of assay detection 1 week later and remained undetectable for 21 weeks until rising to 103 copies/ml through 24 weeks postrebound. The two RM with intermediate viral loads pre-cART had postrebound viremia that ranged from less than 102 to greater than 103 copies/ml through 24 weeks.
SHIV.C.CH848 viremia over time. (A) Plasma viremia (copies/ml, y axis) is depicted longitudinally with week postinfection indicated on the x axis. Intravenous inoculation is indicated with a red arrow. Time of cART administration (24 weeks) is shown as a gray box. (B) Time to virus suppression after cART initiation and time to detectable plasma viremia after ATI are shown for each RM. (C) CD4 T cell percentages over time.
We then quantified the levels of ca-DNA and ca-RNA in SHIV.C-infected PBMCs at peak infection (3 WPI), at chronic infection (12 WPI), and after 12 weeks of cART to characterize the dynamics of PBMC infection throughout the course of the study (Fig. 6). At peak infection, the mean ca-DNA and ca-RNA measurements were 6.0 × 103 and 6.9 × 105 copies/106 cells, respectively (Fig. 6A). During chronic infection, ca-DNA and ca-RNA levels decreased to 7.8 × 102 and 3.9 × 103 copies/106 cells, respectively (Fig. 6A). The levels of ca-DNA and ca-RNA in chronic SHIV.C infection were similar to those in chronic SHIV.D infection (P = 1 and 0.89, respectively, n = 4, Mann-Whitney test). Notably, after 3 months of cART administration, ca-DNA and ca-RNA decreased less than 1 log10 to 3.5 × 102 and 7.3 × 102 copies/106 cells, respectively (Fig. 6A).
SHIV.C.CH848 quantification over time. (A) Cell-associated DNA levels in PBMCs, cell-associated RNA levels in PBMCs, and matched plasma viral load measurements are shown for all RM. Cell-associated DNA levels in PBMC- (B) and LNMC-derived (C) Tcm, Tn, and total resting memory CD4 T cells are shown for all RM. Samples were obtained at baseline (2 weeks prior to inoculation), during peak infection (3 WPI), during chronic infection (15 WPI), and on cART (22 weeks post-cART initiation). Limit of quantification (LOQ) of each assay is denoted by the dashed line. Time of cART administration (24 weeks) is shown as a gray box.
To provide greater understanding of latency measures in long-lived cell populations, we sampled PBMC and lymph node mononuclear cells (LNMCs) and quantified ca-DNA in several T cell subsets over time: central memory CD4 T cells (Tcm), naive CD4 T cells (Tn), and total resting memory CD4 T cells (Fig. 6B and C). In all assayed subsets, the magnitude of infection was highest at peak infection (3 WPI), decreased by 4.6-fold across all subsets during chronic infection (15 WPI), and further decreased by 3.2-fold after 22 weeks of cART (Fig. 6B and C). Tcm ca-DNA was highest in both PBMCs and LNMCs at peak infection (mean = 4.7 × 104 copies/106 cells), decreased by 6.1-fold after 15 WPI (mean = 7.7 × 103 copies/106 cells), and further decreased by 3.9-fold after 22 weeks of cART (mean = 2.7 × 103 copies/106 cells) (Fig. 6B and C). Notably, ca-DNA in this long-lived subset remained detectable in all RM despite more than 20 weeks of cART suppression. In Tn and total resting memory CD4 T cells, decay kinetics upon cART initiation were comparable to those for Tcm (Fig. 6B and C). Total resting memory T cells and Tcm were infected at similar magnitudes at peak and chronic time points, while Tn were infected at a decreased frequency at peak and chronic infection (mean = 6.8 and 4.8 × 103 copies/106 cells, respectively) (Fig. 6B and C).
SHIV.C.CH848 diversity before and after suppressive cART.We next characterized the diversity of SHIV populations by SGS of gp160 env sequences. A total of 341 sequences (median, 27 per animal time point) were generated. To characterize perimaximal env diversity, plasma virus was sequenced 2 weeks prior to cART initiation, at 14 WPI (Fig. 7). SGS sequences revealed a maximum within-animal pairwise diversity of 0.7% to 1.2%. Rebound plasma was sampled twice weekly posttreatment interruption, and sequences were generated from the time of first detectable rebound. Maximum likelihood phylogenetic trees comparing pre-cART and rebound plasma virus env sequences with the TF SHIV.C env sequence are shown in Fig. 7. Maximum within-animal pairwise diversity measurements at the time of first detectable rebound were not significantly different from pre-cART values (range = 0.1% to 0.8%, P = 0.13, n = 4, Wilcoxon signed-rank test). At the first detectable rebound, we sampled multiple genetically distinct virus populations in three of four RM. Rebound virus sequences clustered into between 3 and 11 discrete low-diversity lineages. The exception, RM KH83, exhibited a monoclonal virus population at rebound and had correspondingly low diversity measures (0.1%). While the rebounding virus population was polyclonal in three of four RM, fewer rebound lineages were identified in SHIV.C versus in SHIV.D-infected RM.
Phylogenetic trees of longitudinal SHIV.C.CH848-infected RM env sequences. TF reference sequence is shown in pink, pre-cART sequences are shown in black, sequences from the first, second, and third longitudinal rebound time points following ATI are shown in blue, orange, and green, respectively. Genetic distance is indicated by scale bar, indicating 10 nucleotides, or ∼0.3% difference. All RM were viremic for 16 weeks prior to cART initiation. (A) Rebound sequences from RM KH83 at 29 days post-ATI. (B) Rebound sequences from RM KM11 at days 19, 42, and 49 post-ATI. (C) Rebound sequences from KM65 at days 15, 19, and 56 post-ATI. (D) Rebound sequences from RM KK54 at days 12, 22, and 29 post-ATI.
To determine whether virus populations persisted through periods of control or were cleared and replaced with novel viruses, plasma virus was sequenced longitudinally (Fig. 7). In both KM65 and KM11, we found that the lineages present at first detectable rebound persisted over time despite intermittent periods of undetectable (<83 copies/ml) viremia. In KM65, for example, at the first detectable rebound (15 days post-ATI), three distinct low-diversity lineages were present. Sequences from 19 days post-ATI were identical or closely related (1 to 3 nt different) to those at 15 days post-ATI, with two additional low-diversity lineages sampled. Sequences from day 56 post-ATI, which followed a short period of spontaneous control, formed 1 distinct low-diversity lineage that was closely related (1 to 8 nt different) to lineages sampled at 15 and 19 days post-ATI. Similar results were identified in RM KM11 at days 42 and 49 post-ATI following a period of viral suppression (Fig. 7). In RM KK54, which maintained viremia above the assay LOQ throughout the postrebound period, 4 of 11 virus lineages sampled at the first detectable rebound (12 days post-ATI) persisted at 22 and 29 days post-ATI. An additional six and four distinct lineages were sampled at 22 and 29 days post-ATI, respectively.
Rebound virus population diversity.To assess how virus populations changed in response to cART treatment, we compared pre-ART and rebound virus sequences. Shown in Fig. 8 are logo plots depicting all amino acid positions with greater than 20% change in identity between pre-cART and rebound plasma virus populations. The vast majority of positions, not shown in the logo plots, were relatively fixed from pre-cART to rebound. In positions with diversity in virus replicating at cART initiation, overall diversity modestly decreased. This was most notable in RM with mono- or oligoclonal rebound, e.g., GA67, KH83, and KM65. RM that went on to experience postrebound control generally had fewer rebound variants and thus less virus diversity at rebound. Beyond lower relative diversity, we found no specific sequence signature of postrebound virus control. In the majority of studied RM that experienced polyclonal rebound, pre-cART virus diversity was better retained at rebound, albeit with shifts in relative frequencies at specific positions. At rare positions in RM with greater numbers of rebounding viruses, virus diversity increased at select positions, as either minor or ancestral variants rebounded to replicate systemically.
Rebound virus diversity. Logo plots depicting amino acid positions with greater than 20% change in identity between pre-cART and rebound time points are shown for each RM. Variable positions from pre-cART and first detectable rebound (R1) are shown for each RM. The second time point of rebound (R2) is shown for RM GA67, in which additional rebound lineages arose 1 week after the R1 time point. The third rebound time point is shown for KK54, which experienced persistent viremia post-ATI, and KM65, which had periods of virus control post-ATI.
DISCUSSION
NHP models of HIV/AIDS have been fundamental to elucidating key features of HIV transmission, pathogenesis, persistence, and response to therapy. As the field continues to explore cure strategies, development of an NHP model that mirrors key features of HIV persistence is essential. A model with viruses encoding primary HIV-1 Envs facilitates more direct testing of HIV-1 Env-targeting strategies and is a priority. Here, we tested the capacity of two TF SHIVs to demonstrate key features of HIV-1 replication and persistence on cART. We found that both SHIV.C and SHIV.D reproduced important characteristics of HIV-1 infection.
For an ideal virus in the NHP model system, recapitulation of key components of each stage of infection is important, with acute HIV infection being paramount, as the host-pathogen balance may be established at the earliest stages of infection. In acute HIV infection, peak viral loads can reach greater than 106 to 107 copies/ml with systemic virus dissemination (20, 21). Following the exhaustion of target cells and the onset of adaptive immune responses, a mean set point VL of between 102 and 105 copies/ml is established, albeit with substantial variation by individual (20–22). Both SHIV.C- and SHIV.D-infected RM demonstrated the consistent timing and magnitude of early virus replication previously seen with TF and other SHIVs, with peak viremia of between 105 and 108 copies/ml within 7 to 14 days postinoculation (1–3, 15). Both SHIVs studied here established set point viral loads of between 103 and 106 copies/ml through either 24 weeks postinfection or until initiation of cART in all ten infected NHP, mirroring the viral kinetics of acute and early chronic HIV infection.
After early viral kinetics, a high priority for viruses used in cure studies is their response to cART, in terms of the time to suppression after cART initiation, persistence on longer-term cART, and viral rebound after cART interruption. Upon cART initiation, all TF SHIV-infected NHP rapidly achieved virus suppression below the assay limit of detection, with the majority achieving undetectable levels within 2 weeks. This time frame is slightly more rapid than in HIV-1 infection, where cART containing an integrase strand transfer inhibitor leads to viremia of less than 200 copies/ml within 1 to 2 months of cART initiation in most individuals (23).
All eight TF SHIV-infected RM maintained virus suppression with 24 weeks of cART, with a single RM demonstrating a transient period of breakthrough viremia. Upon cART interruption, the time to virus rebound was akin to that of chronic cART-initiating HIV-positive participants in treatment interruption trials. All TF SHIV-infected RM rebounded within 5 and 22 days (mean of 12 days), which aligns well with the 85% of ATI trial participants who rebound within 4 to 47 days (mean of 11 days) (24, 25). We note the modest variability in the reactivation rate of these animals, which is similar to the heterogeneity in time to rebound in HIV-positive individuals undergoing ATI. In sum, the viral kinetics of TF SHIV infection before, during, and after cART provide a reasonable approximation of HIV-1 infection, with similar average set point VLs and recapitulation of some of the variability seen in HIV-positive individuals.
The kinetics of TF SHIV infection prior to, during, and after cART differ from those of SIVmac239 infection, which is one of the most frequently used NHP model systems. SIVmac239 is a highly virulent clonal strain, which leads to a consistent and pathogenic disease course in RM. By inducing a rapid progressor phenotype (26), characterized by accelerated disease progression, elevated inflammatory markers, and impaired immune responses, SIVmac239 infection of RM can be harnessed for efficient and economical experiments with fewer animals and shorter duration to clinical endpoints. Compared with TF SHIVs and the majority of people living with HIV, SIVmac239 infection leads to higher peak (mean =107–8 copies/ml [19, 27]) and set point (mean = 106 copies/ml [19, 26, 27]) viremia. Furthermore, chronic SIVmac239 infection typically requires more than 4 months of cART administration to achieve virus suppression, which may impose constraints on latency and cure experiments (28, 29).
The mechanisms underlying differences in TF SHIV and SIVmac239 viral kinetics both off and on cART are likely primarily due to the enhanced virulence of SIVmac239. Other related factors, including increased set point viremia, immune dysregulation, microbial translocation, chronic inflammation, and/or infection of cellular subsets with different half-lives and homeostatic programming (19, 28, 30), may contribute to the differences between viruses. We emphasize that SIVmac viruses have been robustly validated in large cohorts of RM. In addition, SIVmac infection results in more consistent viral kinetics and disease progression than SHIV infection, as a percentage of SHIV-infected RM typically spontaneously control infection and do not progress to AIDS. Thus, when selecting a virus for RM studies of HIV latency and cure, key differences in viral kinetics on and off cART as well as pathogenesis and related disease progression should be considered.
One inherent limitation of many NHP experiments, including the study reported here, is the relatively short duration of both active viremia and cART administration. In contrast, the vast majority of people living with HIV and clinical trial participants are viremic for years prior to cART initiation, leading to a greater diversity of replicating and archived virus. Individuals are then treated with suppressive cART for multiple years prior to reservoir evaluation, affording distinct reservoir dynamics with more time for the extinction of shorter-lived infected cells and homeostatic or antigen-driven clonal proliferation (31). Given the importance of understanding what cells constitute the persistent reservoir and developing strategies to counter clonal expansion, this limitation of NHP is critical. Experiments that evaluate the effects of years of cART therapy on SHIV- and SIV-infected RM in the absence of an intervention, although logistically and financially challenging, could address many of these concerns.
Despite relatively short times of virus replication prior to cART initiation in our two experiments, SHIV infection prior to and on cART reflects several characteristics of HIV-1 infection, including viral diversity, polyclonal rebound, and infection of long-lived cells. Both SHIV.D- and C-infected RM revealed significant pre-cART diversity, with a maximum within-animal pairwise diversity of between 0.7% and 2.5%, allowing for differentiation of distinct rebound virus populations. In people living with HIV, recrudescent viremia after treatment interruption results from the systemic reactivation of multiple distinct latently infected cells (32–35). Similarly, TF SHIV rebound was largely polyclonal, with 7 of 8 cART-treated RM demonstrating multiple genetically distinct virus lineages replicating systemically within the first weeks of rebound, supporting a multifocal model of rebound in both systems.
We found that the mean ca-DNA and ca-RNA measurements in chronic SHIV.D- and SHIV.C-infected RM were comparable to those of chronically HIV-infected cART-naive patients (102–3 DNA copies/106 cells and 102–4 RNA copies/106 cells, respectively) (36, 37). ca-DNA and ca-RNA in SHIV.D- and C-infected RM decreased modestly in the first 3 to 6 months following initiation of cART, consistent with the decay kinetics reported in studies of cART-suppressed HIV-positive individuals (23, 36–39). The magnitude of this decay reflects, at minimum, the clearance of plasma virus, productively infected activated T cells with a short half-life (less than 2 days), resting CD4 T cells with unintegrated virus, and likely shorter-lived monocyte- and macrophage-lineage cells (40–42). The persistence of ca-DNA and RNA and the rapid virus rebound seen after ATI indicates that the TF SHIV reservoir both persists with modest diminution in number and transcriptional activity of infected cells after 3 to 6 months of cART and has a slow decay rate on the order of months to years, as seen in HIV-1 infection.
In SHIV.C-infected RM, we had a sufficient quantity of cells to characterize infection of multiple cellular subsets in LNMCs and PBMCs; low cell yields prevented similar analyses in SHIV.D-infected RM. We found that total resting memory CD4 T cells and Tcm in particular were productively infected at a mean frequency of 103–4 copies/106 cells throughout the course of the experiment (Fig. 6B and C). Tcm have been identified as a major contributor to the latent reservoir in HIV-positive cART-suppressed humans (43). We also found that Tn, which given our sorting strategy, includes memory stem T cells, were productively infected in all SHIV.C-infected RM, albeit at a lower magnitude than other subsets (Fig. 6B and C). Our data suggest that TF SHIV infection leads to the establishment of a latent reservoir composed of genetically diverse and long-lived cells, which persist through months of viral suppression to reactivate in the absence of cART.
We note some key differences between SHIV.D- and SHIV.C-infected RM, suggesting that SHIV.C lacks some characteristics that are desirable for a robust model of HIV-1 latency. Most importantly, SHIV.C-infected animals spontaneously controlled viremia to <103 copies/ml postrebound relatively frequently. In our study, rebound viremia rose to 103–7 copies/ml within 4 weeks post-ATI in all TF SHIV-infected RM and then fell to a postrebound set point, similarly to SHIVs that demonstrate postrebound virologic control in some RM (6, 7). The emergence of adaptive (antibody [13, 19, 26, 44, 45] and CD8 T cell-mediated [46–48]) immune responses has been shown to contribute to virologic control and exert significant selection pressure in both SIV and HIV-1. Furthermore, CD8 T cell responses were reported to be integral for bNAb-mediated virologic control in two independent SHIV studies (5, 7). In other TF SHIV experiments, we have found that CD8 T cell depletion via administration of an anti-CD8 Ab leads to increases in plasma viremia in some, but not all, spontaneous controller RM by 1 to 2 log10 copies/ml (data not shown), suggesting that CD8 T cells may contribute to TF SHIV virologic control, although off-target effects of T cell depletion may also be a factor. As an alternate explanation, SHIV.C infection may lead to lower levels of systemic virus replication postrebound than SHIV.D due to differences in intrinsic viral replicative capacity. The pre-cART VL was 103–4 copies/ml in three of four SHIV.C-infected animals. The postrebound set point VL was approximately 1 log10 copies/ml lower and remained near the assay LOD of 83 copies/ml. In parallel, we identified fewer distinct low-diversity lineages upon rebound in SHIV.C in comparison to that in SHIV.D, which may be attributable to the limited duration of viremia (16 weeks) prior to cART initiation and/or to the virus-host interactions discussed above. Sequences from postrebound time points in RM with control revealed that lineages present at the first detectable rebound persisted over time in RM KM11 and KM65 despite periods of spontaneous viremic control, supporting the hypothesis that virus replication continued throughout this period at a level near our assay LOQ. At our sampling depth, we did not find evidence of clearance and subsequent replacement of the virus populations sampled at the first detectable rebound in RM KM11 and KM65. Comparisons of rebound and pre-cART sequences did not identify specific motifs associated with virus control or persistence. Many RM that went on to experience postrebound control had less virus diversity at rebound, but it is unclear whether the decreased virus diversity is a cause or an effect of enhanced immune-mediated virus control.
In contrast, SHIV.D demonstrated many qualities that are desirable for an NHP model of HIV-1 persistence on cART. SHIV.D infection here and in previous studies (15) reproducibly led to high levels of systemic viremia over time. The viral dynamics and reservoir biology of SHIV.D before, during, and after cART suggest that SHIV.D is a promising NHP model for HIV-1 latency. SHIV.D encodes a well characterized and CCR5-tropic (17) subtype D Env, representing an epidemiologically relevant (49) and pathogenically important component of the HIV-1 pandemic. Recognizing that subtype D viruses have been associated with rapid disease progression and frequent emergence of a CXCR4 or dual-tropic virus population (17), we analyzed sequences at the time point immediately preceding cART for genotypes predictive of CXCR4 tropism. Using the geno2pheno coreceptor algorithm, the majority of sequences in all RM were predicted to be equally or more CCR5 tropic than the TF SHIV (data not shown), suggesting a lack of progression to CXCR4 usage. Furthermore, we did not see an enrichment of predicted CXCR4-tropic viruses in RM DE33 and EJ94, which exhibited higher set point VLs and clinical disease progression. Importantly, we previously tested isolates from SHIV.D-infected RM that succumbed to AIDS using phenotypic coreceptor tropism assays and found that CCR5 tropism was retained (15).
Two of the six SHIV.D-infected animals had coincident viral load rise and clinical deterioration. At necropsy, pathological analysis revealed profound systemic CD4 T cell depletion and high viral burden in various lymphoid tissues in both RM and in PBMCs from RM EJ94 (Fig. 4A to C). Severe CD4 T cell depletion was identified in in gut-associated lymphoid tissue (GALT) from both RM (Fig. 4A); this tissue has been shown to harbor the largest number of CD4 T cells of any site in the body (50). To date, TF SHIV pathogenesis has been documented in acute and early infection, where gut and lymphoid CD4 T cells are diminished, while peripheral CD4 T cells are relatively preserved (51). This degree of CD4 T cell depletion within 1 to 2 years of infection in two of six RM suggests rapid pathogenesis of SHIV.D infection.
Our study has several limitations in addition to those described above. First, one of six SHIV.D-infected RM (FT42) spontaneously controlled infection after 42 weeks of infection. Furthermore, while all SHIV.C-infected RM rebounded, three of four did not maintain persistent plasma viremia following ATI. As delineated in Fig. 8, virus control postrebound was associated with fewer variants and less diversity at rebound, which likely reflects the inability of the virus to replicate sufficiently to escape from host control. The SHIV.C experiment also differed from the SHIV.D experiment in that the SHIV.C-infected RM were actively viremic for a shorter duration (16 weeks versus 37 to 85 weeks) and were inoculated with a stock that was 10-fold less infectious. These factors may have contributed to the frequency of postcontrol rebound in the SHIV.C-infected RM. Altogether, the frequency of pre- and post-cART viremic control of TF SHIVs, in particular, SHIV.C-infected RM, is a limitation for their use in cure studies.
Other study limitations include the sample sizes in both our SHIV.D and SHIV.C experiments, which were small (six RM and four RM per study, respectively), limiting the statistical power of our analyses. Additional studies in larger cohorts will be important to validate these findings. In addition, the duration of cART during our studies was limited to 24 weeks and therefore was insufficient to observe reservoir decay in longer-lived cellular reservoirs, such as resting memory CD4 T cells, which have a half-life of months to years (52). As we did not longitudinally sample tissues from the central nervous system and other possible anatomical reservoirs such as lymphoid tissues, which harbor the majority of CD4+ T cells in HIV-1-infected individuals, we were unable to characterize the viral reservoir dynamics at these sites (42, 53). Moreover, the quantity of cells collected after 3 to 6 months of cART in both SHIV.D- and SHIV.C-infected RM were inadequate to conduct a more complete reservoir analysis, such as characterizing the genetic intactness and replication competence of reservoir viruses using whole-genome sequencing or viral outgrowth assays (52).
In summary, we have validated two novel TF SHIVs as relevant viruses in the RM model of HIV-1 latency. The characteristics of SHIV.D.191859, and to a lesser extent, SHIV.C.CH848, during chronic infection, cART administration, and upon treatment interruption suggest that they are promising reagents for a SHIV model of HIV-1 latency and that they are suitable for studies of HIV-1 reservoir biology and pathogenesis. Our findings demonstrate that the TF SHIV-macaque model has broad applications for elucidating mechanisms of HIV persistence and evaluating a range of eradicative and suppressive strategies.
MATERIALS AND METHODS
Nonhuman primates.Indian-origin RMs were maintained at Tulane National Primate Research Center (TNPRC) according to the standards of the Association for Assessment and Accreditation of Laboratory Animal Care. All experiments were approved by the Tulane Animal Care and Use Committee. All animals were tested and found to be negative for described SIV controller alleles Mamu-A*01, B*08, and B*17, except for RM DE33, which was A*01 positive. Whole blood from animals was processed using centrifugation as described previously (15). PBMCs were isolated using Ficoll-Paque (GE Healthcare) gradient centrifugation. Plasma was clarified by centrifugation for 15 min at 3,000 rpm, frozen, thawed, and then subject to viral RNA extraction as described below. SHIV VL levels in plasma were determined by the Pathogen Detection and Quantification Core at the TNPRC using quantitative real-time reverse transcription PCR of SIV RNA as previously described; the assay limit of quantification (LOQ) was 83 copies/ml (54).
Construction and viral stock characterization of SHIV.D.191859 and SHIV.C.CH848.SHIV.D.191859 and SHIV.C.CH848 constructs were generated and viral stocks were made as described previously (15). The virion content of viral stocks was quantified in triplicates using the p27Ag 96-well enzyme-linked immunosorbent assay (ELISA) kit from ZeptoMetrix. The infectivity of viral stocks was determined via titration on TZM-bl cells in quadruplicates. Virion content for the rhesus PBMC-derived SHIV.D.191859.375M.dCT and 293T-derived SHIV.C.CH848.375H.dCT stocks were 211 and 2,342 ng p27Ag/ml, respectively. The SHIV.D.191859 and SHIV.C.CH848 infectivity-to-particle ratios as determined on TZM-bl cells were 1.5 × 10−2 and 5.5 × 10−4 infectious units (IU)/particle, respectively, which are similar to other TF SHIV stocks (15).
SHIV.D.191859 and SHIV.C.CH848 inoculation.Six RM were intravaginally challenged with 1 ml of undiluted SHIV.D.191859.dCT containing 3.19 × 10−7 infectious units (IU) as determined on TZM-bl cells. Two RM (FE43 and EJ94) became productively infected following the first challenge. Four RM that remained uninfected were intravaginally challenged with 1 ml of stock dilutions of 1:10 at week 12, 1:4 at week 16, and 1:2 at week 23. After four challenges, two RM (FR55 and GA67) remained uninfected; these RM were intravenously inoculated with 1.00 × 10−6 IU of SHIV.D.191859 at week 48, which resulted in productive infection of both animals. In a separate study, four RM were intravenously challenged with 1 ml of undiluted SHIV.C.CH848.dCT stock containing 1.29 × 10−7 infectious units (IU) as determined on TZM-bl cells; all RM became productively infected.
Viral sequencing.Single-genome gp160 env sequences were generated as described previously (15). Briefly, 20,000 viral RNA copies were extracted from plasma by means of the Qiagen BioRobot EZ1 Workstation with the EZ1 Virus Mini kit v2.0 (Qiagen). Eluted vRNA was subsequently used as a template for cDNA synthesis and reverse transcribed using the reverse primer SHIV.Env.R1 (5′-TACCCCTACCAAGTCATCA-3′) and SuperScript III reverse transcriptase (Invitrogen Life Technologies). cDNA was serially diluted in a 96-well plate (Applied Biosystems) to identify the dilution at which less than 30% of wells contained PCR amplicons of the correct size. The SHIV gp160 env genome was amplified via nested PCR with primers as follows: first round forward primer SHIV.Env.F1 (5′-CGAATGGCTAAACAGAACA-3′), second round forward primer SHIV.Env.F2 (5′-CTACCAAGGGAGCTGATTTTC-3′), first round reverse primer SHIV.Env.R1 (5′-TACCCCTACCAAGTCATCA-3′), and second round reverse primer SHIV.Env.R2 5′-TATTTTGTTTTCTGTATGCT-3′). PCR conditions were used as follows for the first round of nested PCR: 94°C, 2 min; 37 × (94°C, 20 s; 55°C, 30 s; 68°C, 3 min 30 s); 68°C, 10 min. For the second round of nested PCR, the PCR conditions were as follows: 94°C, 2 min; 42 × (94°C, 20 s; 54°C, 30 s; 68°C, 3 min 30 s); 68°C, 10 min. Amplicons were sequenced via the MiSeq platform (Illumina). Raw reads were aligned to the SHIV.D.191859 or SHIV.C.CH848 TF reference using Geneious R9. Sequences that contained mixed bases at a frequency of greater than 25% per nucleotide position were excluded from further analyses.
TF estimation.To approximate the minimum number of virus-infected cells that contributed to recrudescent viremia, we generated a conservative estimate of the amount of genetic diversity that could accrue in the time elapsed post-ATI. We used the median error rate for HIV-2 reverse transcriptase (RT) reported by Rawson et al. (55) (3.1 × 10−4 mutations/nucleotide base), as SIVmac is more closely related to HIV-2 than HIV-1 (56); of note, error rates for both HIV-1 and HIV-2 RT are reported to be very similar (55, 57). We assumed a virus-infected cell generation time of 1.5 days based on previous studies (40, 41, 58). We multiplied HIV-2 RT error rate (3.1 × 10−4 mutations/nucleotide base) by average TF SHIV gp160 env amplicon length (3,693 bp) and divided this number by the virus-infected cell generation time (1.5 days) to estimate the amount of genetic diversity that could accrue in a day. We then multiplied the resulting number by the number of days elapsed post-ATI to estimate the maximum amount of genetic diversity that could accrue in each virus population during this period, subtracting four days from the number of days elapsed post-ATI to account for a conservative terminal drug washout period based on previous pharmacokinetic elimination studies in HIV-1-infected humans (59). The equation used for estimating the maximum accrued diversity post-ATI for each RM is thus summarized as: (3.1 × 10−4 × 3,693 × [days elapsed post-ATI-4])/1.5. Data regarding the complete pharmacokinetic elimination and drug metabolism of tenofovir, emtricitabine, and dolutegravir in HIV-1-infected humans and SIV- and SHIV-infected rhesus macaques are limited, to date (60–62). This estimate was used to determine the maximum divergence one reactivating lineage unaffected by recombination could accrue during ATI, which allowed us to enumerate the minimum number of independent reactivation events during systemic viral rebound at the time of sampling based on divergence from neighboring lineages.
Phylogenetic analyses.Maximum likelihood phylogenetic trees were generated using the HKY85 substitution model and a transition/transversion ratio of four in PhyML version 3.0 and visualized in FigTree (63). Maximum pairwise diversity measurements of pre- and post-cART virus populations were calculated using DIVEIN (64).
Flow cytometric sorting of PBMCs and LNMCs.(1) Antibodies used for CD4 T cell subset sorting. The following antibodies were obtained from BioLegend: CD8 AF647 (clone RPA-T8), CD14 BV650 (clone M5E2), CD16 BV650 (clone 3G8), and CD20 BV650 (clone 2H7). The following antibodies were obtained from BD Biosciences: CD3 allophycocyanin (APC)-Cy7 (clone SP34-2), CD95 phycoerythrin (PE)-Cy5 (clone DX2), and HLA-DR BV421 (clone G46-6). CD28 ECD (clone CD28.2) was obtained from Beckman Coulter. CD4 PE-Cy5.5 (clone S3.5) was obtained from Life Technologies. The Live/Dead Fixable Aqua Dead Cell Stain kit (Invitrogen) was used for viability exclusion.
(ii) Sorting for CD4+ T cell subsets. Cryopreserved PBMCs or LNMCs were thawed, counted, examined for viability, and rested for 2 h at 37°C and 5% CO2 in complete medium (RPMI supplemented with 10% fetal bovine serum [FBS], 2 mM l-glutamine, 100 U/ml penicillin, and 100 mg/ml streptomycin). All incubations were performed in the dark at room temperature. Cells were stained for viability exclusion for 10 min. Undiluted antibody cocktail mix was added and incubated for 20 min. Following surface staining, cells were washed and resuspended in phenol-free RPMI and kept in the dark at 4°C. Cells were sorted on a FACSAria II (BD Biosciences). Sorted cells were washed with phosphate-buffered saline (PBS) and pelleted at −80°C until processing. Tn were identified as CD28+ CD95− CD4 T cells. Tcm were identified as CD28+ CD95+ CD4 T cells. Resting memory CD4 T cells were identified as CD95+ HLA-DR− CD4 T cells.
Quantification of cell-associated SHIV DNA and RNA in PBMCs and tissues.Cellular DNA was purified using the AllPrep DNA/RNA kit (Qiagen) as specified by the manufacturer and normalized to cell equivalents by quantitative PCR (qPCR) using a nonhuman primate-specific TaqMan hRNase P primer-probe assay (Life Technologies). Total cellular SHIV DNA was quantified in a qPCR reaction using primers and probe specific to the SIVmac239 long terminal repeat (LTR) region (forward primer, 5′-TACCCAGAAGAGTTTGGAAGCAAGTCA-3′; reverse primer, 5′-TTGTCAGCCATGTTAAGAAGGCCTCTTG-3′; and probe A, 5′-6-carboxyfluorescein [FAM]-CTGTCAGAGGAAGAGGTTAGAAGAAGGCTAAC-black hole quencher 1 [BHQ1]-3′ [Integrated DNA Technologies]). PCR reaction mixtures each contained 10 μl of 2× TaqMan Universal Master Mix II (Life Technologies), 6 pmol of each primer, 6 pmol of probe, and 5 μl of DNA (total volume of PCR reaction = 20 μl). PCR cycle conditions were as follows: 50°C, 2 min; 95°C, 10 min; 50 × (95°C, 15 s; 60°C, 1 min); PCR was performed on an ABI 7500 FAST machine (Life Technologies). External quantitation standards were prepared using DNA isolated from the SIV1C cell line, which has a single copy of SIV DNA per cell; this was followed by copy number determination using RainDance digital droplet PCR (RainDrop). Specimens were assayed in replicate reaction wells. Copy number was determined by extrapolation against an 8-point standard curve (1 to 100,000 copies), which was performed in triplicates. Due to low viable PBMC and LNMC recovery for some SHIV.C subset samples, the ca-DNA LOQ varied and ranged between 20 and 1,500 copies per sample (Fig. 6B and C). Similarly, ca-RNA measurements for two SHIV.D-infected RM (RM FE43 and GA67) were not detectable at month 6 due to low cell yields and therefore were excluded from Fig. 2B.
qPCR quantification of cellular SIV RNA.Cellular RNA was purified using the AllPrep DNA/RNA kit (Qiagen) as specified by the manufacturer and was normalized to cell equivalents by qPCR using the nonhuman primate-specific TaqMan RPLPO primer-probe assay (Life Technologies). Total cellular SHIV RNA was quantified using a two-step RT-PCR assay. First, 250 ng of RNA was reverse transcribed to produce cDNA using the VILO IV reverse transcriptase enzyme and random hexamer primers (Invitrogen) according the manufacturer’s instructions. cDNA was quantified in a qPCR reaction using primers and a probe specific to the SIVmac239 LTR region (forward primer, 5′-TACCCAGAAGAGTTTGGAAGCAAGTCA-3′; reverse primer, 5′-TTGTCAGCCATGTTAAGAAGGCCTCTTG-3′; and probe A, 5′-FAM CTGTCAGAGGAAGAGGTTAGAAGAAGGCTAAC-BHQ1-3′ [Integrated DNA Technologies]). Each reaction contained 10 μl of 2× Taqman Universal Master Mix II (Life Technologies), 6 pmol of each primer, 6 pmol of probe, and 5 μl of cDNA, with a total reaction volume of 20 μl. PCR cycle conditions were as follows: 50°C, 2 min; 95°C, 10 min; 50 × (95°C, 15 s; 60°C, 1 min); PCR was performed on an ABI 7500 FAST machine (Life Technologies). qPCR runs were carried out using an ABI 7500 FAST Machine (Life Technologies). External quantitation standards were prepared using SIVmac239 RNA followed by copy number determination using the RainDance digital droplet PCR (RainDrop, MA). Specimens were assayed in replicate reaction wells. Copy number was determined by extrapolation against a 6-point standard curve (3 to 300,000 copies), which was performed in triplicates.
Logo plot generation.For each RM, sequence alignments from pre-cART and first detectable rebound were compared. Any amino acid position with greater than 20% change in identity between time points was included in analyses. Logo plots were generated on AnalyzeAlign (https://www.hiv.lanl.gov/content/index).
Statistical analyses.Statistical analyses were performed using Prism 7.0 software (Graphpad). Given the small sample sizes (n = 4 to 6), nonparametric tests were used for all statistical comparisons. The Wilcoxon signed-rank test was used to test for changes in within-animal measures of viral diversity and latency pre- and post-cART administration. The Mann-Whitney test was used to assess the relationship of ca-DNA and ca-RNA levels in PBMCs in RM infected with SHIV.D versus those with SHIV.C.
ACKNOWLEDGMENTS
This work was supported by the Penn Center for AIDS Research Viral and Molecular Core (P30 AI045008), the BEAT-HIV: Delaney Collaboratory to Cure HIV-1 Infection by Combination Immunotherapy (UM1AI126620), and CARE: Delaney Collaboratory for AIDS Eradication (UM1AI126619).
We thank Gilead and ViiV Healthcare for donating tenofovir, emtricitabine, and dolutegravir. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. We declare no conflicts of interest.
A.M.B., H.L., H.X., R.V., and K.J.B. conceived of the project; A.M.B., W.Z., E.L., L.K.-C., H.L., F.-H.L., M.W., H.X., R.V., and K.J.B. designed and performed the experiments; A.M.B., W.Z., E.L., L.K.-C., H.L., and K.J.B. analyzed results; A.M.B. and K.J.B. wrote the manuscript.
FOOTNOTES
- Received 7 October 2019.
- Accepted 7 January 2020.
- Accepted manuscript posted online 22 January 2020.
- Copyright © 2020 American Society for Microbiology.
