No Viral Evolution in the Lymph Nodes of Simian Immunodeficiency Virus-Infected Rhesus Macaques during Combined Antiretroviral Therapy

TEXT

Combination antiretroviral therapy (cART) has transformed human immunodeficiency virus (HIV) infection from an incurable disease to a manageable one. It suppresses the viral burden in patients to undetectable levels (1–3), lowers the chance of viral transmission (4), increases the number of CD4⁺ T lymphocytes (1, 2), reconstitutes immunity (5–7), and extends the life expectancy of patients (8). However, cART does not cure patients because of its inability to eradicate the virus from infected individuals (9), suggesting the existence of a viral reservoir that is refractory to cART. Its identification and eradication are therefore requisites for a functional cure for AIDS. To establish a strategy for eradication of the HIV reservoir, the mechanism of persistence of the virus must be elucidated. Two mechanisms of viral persistence have been proposed: one is ongoing cycles of viral replication despite the presence of antivirals (10), and the other is provirus integration into long-lived cells (11). Whereas previous studies concerning this issue have been extensively conducted with clinical specimens from HIV-1-infected patients, including plasma, peripheral blood mononuclear cells, and gut-associated lymphatic tissues (12–14), lymph nodes, which are epicenters of virus replication in infected individuals not undergoing therapy (15–17), have only rarely been subjected to scrutiny. In animal models of cART, in particular, the simian immunodeficiency virus (SIV)-macaque model, which allows systemic examination, the location of the viral reservoir and the mechanism of viral holding have not been studied in detail.

To elucidate how the virus is maintained during cART in an animal model of anti-HIV chemotherapy, we administered a combination of nucleotide/nucleoside reverse transcriptase inhibitors (azidothymidine, lamivudine, and tenofovir disoproxil fumarate) and protease inhibitors (lopinavir with ritonavir) to four SIV239-infected rhesus macaques for 1 year (18). Although the plasma viral RNA loads of the animals were suppressed to levels below the assay detection limit during the period of chemotherapy, a systemic analysis conducted at the completion of therapy revealed viral RNA present in lymphatic tissues, especially in mesenteric and splenic lymph nodes (MLN and SLN, respectively) at high titers. Reasoning that any possible mode(s) of viral persistence should be in operation in tissues with high levels of viral RNA expression, we investigated viral genes in these tissues.

It is expected that viral genes accumulate nucleotide substitutions in proportion to the time postinfection in individuals not undergoing therapy because of continuous virus replication mediated by the error-prone viral reverse transcriptase. Such mutation rates have indeed been observed in the V3 loop of env, p17 of gag (19), and the C2-to-C5 region of env (20) in HIV-1-infected patients, as well as in the env gene from monkeys experimentally infected with SIV (21, 22). We hypothesized that viral genes would accumulate mutations if the virus was continuously replicating in the reservoir despite the presence of antivirals.

First, to ascertain whether such an accumulation of mutations took place at a detectable magnitude in our experimental system, we used SIV239, a molecularly cloned virus, to infect macaques for 1 year and periodically sampled viral genes from the untreated control animal (MM521). To reveal ongoing expression of viral genes at sampling, total RNA was extracted from plasma samples collected at 8, 18, 42, and 68 weeks postinfection (wpi) and examined. Single-genome amplification (SGA) (23) was used to amplify the viral genes present and to avoid the selective amplification of a particular genotype or recombination between genotypes during PCR. Using a nested PCR method, we amplified the entire env gene, which accumulates nucleotide substitutions in the greatest numbers, following reverse transcription of cDNA from the extracted RNA. The initial PCR cycles were carried out with the following primers: forward, SIV20F (5′-CTC CAG GAC TAG CAT AAA TGG-3′); reverse, SHenv9R (5′-GGG TAT CTA ACA TAT GCC TC-3′). Successive PCR cycles were run with the following primers: forward, SIV21F (5′-CTC TCT CAG CTA TAC CGC CC-3′); reverse, SHenv8R (5′-GCC TTC TTC CTT TTC TAA G-3′). The PCR products from an average of 12 independent reactions per time point were directly subjected to sequencing.

We computed the number of mutations in each SGA clone obtained from plasma samples of an untreated monkey (MM521) through a comparison with that of the inoculum virus (Fig. 1). A linear relationship with a coefficient of 1.25 × 10⁻⁴ (r² = 0.8503, P < 0.0001; GraphPad Prism, La Jolla, CA) was revealed between the number of mutations in the SGA clones and the time postinfection. By using the coefficient, the cumulative number of mutations per annum was determined to be 6.5 × 10⁻³ substitutions/site/year, a value comparable to those of SIV and HIV reported previously (9 × 10⁻³ [21, 22] and 6.0 × 10⁻³ [23] substitutions/site/year, respectively). The accuracy of the “molecular clock” in our experimental setting prompted us to examine viral RNA extracted from the lymph nodes of animals that underwent cART for 1 year.

• Open in new tab
• Download powerpoint

Fig 1

Time-dependent accumulation of nucleotide substitutions in SIV genomes circulating in an infected and untreated rhesus macaque. The sequences of viral env genes in circulation collected at 8, 18, 42, and 68 wpi from SIV239-infected animals and an untreated animal (MM521) were determined. Tamura-Nei distances (36) of the sequences were computed with the MEGA5 software (37), and the number of nucleotide substitutions per site was plotted against the number of weeks postinfection. Each symbol represents a single genomic amplicon derived from plasma samples collected at the time points designated.

Total RNA was extracted from the MLN of four treated animals and one untreated animal, as well as the SLN of one of the treated animals (MM530), at the completion of the observation period and used as the template for PCR; the products were subjected to sequence analysis as described above. On average, 10 sequences were obtained from each sample (Fig. 2A and Table 1). The number of mutations observed in the env gene from MM521 (untreated) was, on average, 25 of 2,700 bases. In contrast, the number in treated animals was, on average, 1.5 of 2,700 bases (Table 1). The difference in the number of mutations in env between the plasma and MLN samples from the untreated animal, MM521, at 68 wpi (at necropsy) was statistically insignificant (P > 0.05; Fig. 2A), justifying our comparison of these two distinct anatomical compartments. Thus, we proceeded to compare the substitution numbers in plasma at 8 wpi, immediately before the onset of cART, with those from the lymph nodes of animals treated with cART at necropsy (61 to 65 wpi). The number of nucleotide substitutions in the env gene in both the plasma and MLN of the untreated animal (MM521) at 68 wpi was higher than that in plasma at 8 wpi (P < 0.0001). In contrast, those in the MLN of treated animals at the completion of cART were unchanged (MM528 and SLN of MM530) or decreased significantly (MM491, MM499, and MLN of MM530) (Fig. 2A). The results indicated that the virus did not accumulate further mutations beyond those obtained by 8 wpi.

• Open in new tab
• Download powerpoint

Fig 2

Nucleotide substitutions in env genes from SIV239-infected animals. The number of mutations in env from the plasma (PL, at 8 and 68 wpi) and MLN (at 68 wpi) of an SIV-infected but untreated animal (MM521) and from the plasma (at 8 wpi) and MLN (at necropsy, 61, 63, 64, or 65 wpi) and SLN (at necropsy, 65 wpi) from SIV-infected and treated monkeys (MM491, MM499, MM528, and MM530) were assessed as described in the legend to Fig. 1. (A) Numbers of nucleotide substitutions per site are shown. The statistical significance of differences between substitution numbers was evaluated by Student's t test using GraphPad Prism. *, P < 0.05; **, P < 0.001; ***, P < 0.0001; NS, P > 0.05. (B) Numbers of nucleotide substitutions per annum are shown. *, P < 0.001; NS, P > 0.05.

As the samples were collected from animals at various time points postinfection, the numbers depicted in Fig. 2A were converted to substitutions/site/year (Fig. 2B) for further analysis. Comparison of the number of viral mutations in plasma at 8 wpi (median, 5.9 × 10⁻³ substitutions/site/year) with that in the MLN (median, 7.2 × 10⁻³ substitutions/site/year) in the untreated animal, MM521, indicated no statistically significant difference (P = 0.6265), as predicted by the analysis in Fig. 2A. Next, we compared the numbers in animals that underwent chemotherapy. At 8 wpi, the treated animals were equivalent to MM521 (an untreated animal) in terms of therapeutic status, since cART was started after sample collection at 8 wpi. Not unexpectedly, there was no statistically significant difference in the number of substitutions/site/year in plasma between the untreated and treated animals (MM491, 8.5 × 10⁻³; MM499, 9.8 × 10⁻³; MM528, 1.6 × 10⁻²; MM530, 8.5 × 10⁻³; MM521, 5.9 × 10⁻³), except for MM528 (P = 0.0048 compared to the value for the untreated animal). In contrast, the number of mutations per annum in the lymph nodes of treated animals collected at necropsy (median, 3.4 × 10⁻⁴ substitutions/site/year) was significantly lower than that in the plasma of the animals at 8 wpi (median, 9.8 × 10⁻³ substitutions/site/year; P < 0.0001). The number of mutations per year in the lymph nodes also differed significantly between the untreated and treated macaques (P < 0.0001). This supports the hypothesis that ongoing viral replication contributed little, if anything, to viral persistence during cART.

Examination of the nucleotide substitution numbers did not indicate discernible de novo virus replication during cART. Therefore, we next investigated continuous viral replication during cART through phylogenetic analysis of viral env clones. Clones were obtained from the untreated animal (derived from plasma at 8, 18, 42, and 68 wpi and from MLN) and from one of the treated animals (derived from plasma at 8 wpi and from MLN at necropsy) (Fig. 3; see Fig. S1 in the supplemental material). To illustrate the accumulation and specific sites of mutations, Highlighter plot analysis (http://www.hiv.lanl.gov/content/sequence/HIGHLIGHT/help.html) was also performed. Phylogenetic analysis of the viral genes from the untreated animal revealed that (i) env clones from plasma exhibited increasing genetic distance from the inoculum virus with time; (ii) clones obtained at a given time point branched out of the one immediately before, a clear demonstration of viral evolution; and (iii) clones from lymph nodes formed a cluster with those from plasma collected at the same time. In contrast, clones from treated animals, regardless of the tissue origin or time point, formed a cluster with clones derived from the plasma of the untreated animal at 8 wpi and the inoculum virus (Fig. 3; see Fig. S1). The results of the Highlighter plot analysis were consistent with those of the phylogenetic analysis. These results clearly demonstrated that viral evolution did not take place in SIV239-infected rhesus macaques during cART. Analysis of the env genes in the peripheral blood mononuclear cells and gut-associated lymphatic tissues obtained from HIV-1 patients undergoing cART also found no evidence of de novo viral replication (13).

• Open in new tab
• Download powerpoint

Fig 3

Phylogenetic relationship of env sequences from treated (MM530) and untreated (MM521) SIV-infected animals. Sequences of the entire env gene from both animals were subjected to phylogenetic analysis. The phylogenetic tree was constructed by the maximum-likelihood method (38). Open circles, sequences in the plasma of MM530 at 8 wpi; closed circles, those from the MLN of MM530 at 65 wpi; open triangles, those from plasma of MM521 at 8 wpi; closed triangles, those from plasma of MM521 at 18 wpi; open inverted triangles, those from plasma of MM521 at 42 wpi; closed inverted triangles, those from plasma of MM521 at 68 wpi; closed rectangles, those from MLN of MM521 at 68 wpi. The scale represents a genetic distance equivalent to 0.002 substitution/site. The corresponding sequence of SIVmac251 32H (GenBank accession no. D01065) was used as the outgroup.

In contrast, other studies have reported continuous virus replication during combined chemotherapy (10, 12). One possible explanation for this discrepancy is the thorough suppression of the plasma viral burden, <20 copies/ml at necropsy, that was achieved in this study (18). De novo virus replication was detected in HIV-1-infected patients whose plasma viral RNA burdens ranged from 20 to 400 copies/ml but not in those with <20 copies/ml (24). Our findings also indicate that the cART regimen we used (18) was robust enough to halt viral evolution nearly completely in animals.

Our sample size, an average of 10 sequences from each specimen, may conceivably have limited our ability to detect minor populations with signs of ongoing replication. An analysis of four animals, however, did not reveal the genotypes detailed in the present study. Therefore, while our results cannot rule out possible de novo viral replication during cART, the data indicate that it is not a major mode of viral persistence in individuals whose virus replication levels are thoroughly suppressed by cART.

The locations of other potential viral reservoirs, in addition to resting CD4⁺ T lymphocytes, an already established HIV/SIV reservoir found to be present in blood (25–28), lymph nodes (25), and the spleen (29), remains elusive. While the cART regimen we developed suppressed viral RNA levels nearly completely in the circulation and fairly well in effector sites, such as the gastrointestinal tract and lungs, viral RNA expression levels in lymph nodes were not contained effectively (18), suggesting that the viral reservoir consists of cells present in lymph nodes. We also detected CD3-positive cells, most likely CD4⁺ T lymphocytes, expressing Nef protein in the follicles of the MLN of an SIV-infected animal that exhibited a viral rebound upon the cessation of cART (18). On the basis of their location, these might be Tfh cells, which are of the memory phenotype (30–32). The results of the present study have further narrowed the location of the viral reservoir from our previous study (18) to cells with longer half-lives that retain provirus for at least 1 year. Since resting CD4⁺ T cells possess long half-lives (33), these cells satisfy this criterion for a viral reservoir during cART. It is conceivable that resting CD4⁺ T cells functioned as the predominant viral reservoir in the SIV239-rhesus macaque model for patients undergoing cART employed in our study, as in preceding studies concerning the issue in the context of HIV and SIV infections.

Lymph nodes serve as a major HIV reservoir throughout the course of infection without intervention by cART (15–17). During clinical latency, the virus persists as an intact provirus, which can produce infectious viral particles upon cell activation, in a miniscule fraction of the resting CD4⁺ T lymphocytes in lymph nodes (25). An extensive examination of lymph node specimens from HIV patients undergoing cART revealed an infinitesimal amount of viral RNA-positive cells by in situ hybridization (34). Hockett et al. (34) revealed that cART lowers the number of viral RNA-positive cells in lymph nodes but that the number of viral copies in each infected cell is constant, regardless of the viral burden in the circulation, suggesting the existence of virus-infected cells actively transcribing viral genes during cART, as we found previously in the lymph nodes of SIV239-infected animals undergoing cART (18). Our present observations, together with those of Hockett et al. (34), indicate that the viral RNA-positive cells present in lymph nodes during cART may represent cells infected with virus prior to the initiation of cART and transcribing viral RNA from integrated provirus during therapy.

Current cART is unable to eradicate the viral reservoir or, more precisely, provirus integrated in the reservoir. On the basis of our results, it is important to establish strategies to target specifically long-lived cells that harbor intact provirus while unlocking the dormant state of the provirus, perhaps by using histone deacetylase (35), to achieve a functional cure for AIDS.