Simian immunodeficiency virus macaque models of HIV latency

Abstract

Purpose of review

This review will focus on recent developments in several nonhuman primate models of AIDS. These models are being used to address viral latency and persistence during antiretroviral therapy in studies that are not feasible in humans.

Recent findings

Further characterization of the various macaque models of AIDS has demonstrated that several aspects of viral persistence during antiretroviral therapy model HIV-1 infection in humans, including viral decay kinetics. Widespread distribution of viral RNA and viral DNA has been detected in many tissue organs. In addition, the brain has been identified as a site of persistent viral DNA.

Summary

The macaque models of AIDS are well suited for addressing viral persistence during antiretroviral therapy, including viral latency, residual replication, and tissue organ distribution.

Introduction

Human immunodeficiency virus type 1 (HIV-1) infection persists despite tremendous advancements in highly active antiretroviral therapy (HAART). HAART, consisting of combinations of three or more antiretrovirals from two or more classes, can reduce HIV-1 RNA levels in plasma to below the level of detection of standard clinical assays and prevent the progression to AIDS [1]. However, the development of more sensitive virus load assays has demonstrated that low-level viremia persists. In addition, virus loads rebound upon cessation of HAART. Although the mechanisms of viral persistence during HAART are not known, some viral reservoirs have been identified. Resting memory CD4⁺ T lymphocytes are a major reservoir that was discovered upon suppression of plasma virus loads to low levels with combination therapy [2–4]. These latently infected cells are established during primary infection, and it has now been demonstrated that they persist for years despite viral load suppression during HAART, expressing little viral RNA (vRNA) until cellular activation [3,5,6]. Other cellular reservoirs might exist, consisting of long-lived cell types, such as macrophages (reviewed in [7]). Further complicating the issue, there is some evidence for ongoing viral replication despite HAART [8–10]. This residual replication could occur in tissues with limited drug access or in cell-types that fail to metabolically activate the nucleoside reverse transcriptase inhibitors (NRTIs). Any residual replication could reseed the cellular reservoirs, thereby extending their rate of decay and preventing eradication. Nonhuman primate animal models of HIV-1-infection have been developed that have the potential to address aspects of viral persistence, including residual replication and viral latency. These models can be used to perform controlled studies that are not feasible in humans. This review will focus on the most recent developments in macaque models being used to understand viral persistence during HAART and their implications for HIV-1 persistence and latency.

Simian immunodeficiency virus

Asian macaques infected with simian immunodeficiency virus (SIV) develop an AIDS-like disease similar to that observed in HIV-1-infected humans, defined by vRNA in plasma, loss of CD4⁺ T lymphocytes, and opportunistic infections (reviewed in [11]). Several strains of SIV have now been developed as animal models to study various antiretroviral effects on infection (recently reviewed in [12]). SIV models are being used to evaluate vaccine strategies and immunological parameters that are beyond the scope of this review. The two primary species of macaques being used in the antiretroviral studies are rhesus (Macaca mulatta) and pig-tailed (Macaca nemestrina). The different strains of SIV can be segregated by pathogenesis based upon the levels and rates of viral replication and the rates of disease progression. Pathogenic strains of SIV, such as SIVmac239, are used in several models because the high level of viral replication and rapid disease progression lead to a reproducible, robust model and a shorter evaluation time (reviewed in [11]). Because macaques are similar to humans anatomically, physiologically, and immunologically, the SIV-infected macaque is suitable as a model for many aspects of HIV-1 infection.

SIV is susceptible to the NRTIs (reviewed in [12]). Treatment of SIV-infected macaques with lamivudine (3TC), emtricitabine (FTC), or tenofovir (PMPA) in monotherapy resulted in virus load suppression [13,14], followed by the rapid selection of virus with reduced drug susceptibility that contained mutations similar to HIV-1 mutants [14]. Combination antiretroviral therapy has been used to establish that latent virus persists in resting CD4⁺ T lymphocytes in macaques [15]. Recently, it was also determined that depletion of CD8⁺ T cells during treatment with PMPA and FTC (to block new rounds of infection) did not increase the lifespan of infected cells in macaques [16^••,17^••]. This important finding suggests that direct killing by CD8⁺ T cells is not responsible for the antiviral effect observed in situations of high virus load, despite previous evidence such as CD8⁺ T cell escape mutations and other depletion studies [16^••,17^••]. These results have implications for the induction or activation strategies being tested in humans in attempts to purge the latent reservoir. Upon activation of infected resting CD4⁺ lymphocytes, a mechanism might be required in order to eliminate these cells. These data suggest that further immune system modulation may need to be included in order to eliminate this reservoir.

To fully address viral persistence during treatment, additional antiretroviral agents must be added to the regimens in order to achieve lower levels of viremia and more sustained virus load suppression. An additional class of antiretroviral agents used in some HAART regimens is the protease inhibitors [1]. SIV is susceptible to some of the protease inhibitors at median effective concentrations (EC₅₀) similar to those observed for HIV-1 [18]. Protease inhibitors have been used in a macaque study in combination with a newer class of antiretroviral agents, integrase inhibitors [19^••]. This combination resulted in biphasic decay of plasma vRNA, sustained virus load suppression, and increased CD4⁺ T lymphocytes in various tissue compartments [19^••]. Importantly, this study demonstrated that SIV persists in resting CD4⁺ T lymphocytes from the spleen, peripheral blood mononuclear (PBM) cells, pooled gut lymph nodes, and pooled head lymph nodes, despite virus load suppression by antiretrovirals, thereby identifying potential reservoirs of latently infected cells in HIV-1-infected persons [19^••].

A recently approved integrase inhibitor, Raltegravir (RAL), has also been demonstrated to inhibit SIV in macaques [20^••]. In addition to in vitro characterization of the efficacy of RAL, it was demonstrated to be effective in vivo in short-term monotherapy (10 days) [20^••]. Upon the addition of PMPA and FTC to the regimen, plasma virus loads were suppressed to less than 50 vRNA copies per ml [20^••]. Because RAL is recommended in drug-experienced individuals [1], this model should be particularly useful for studying the persistence of multi-drug resistant virus.

As a result of the observation in humans that neuroinflammation persists during HAART, the SIV-infected macaque model is also being used to study viral persistence in the brain [21^•–24^•]. In order to increase the incidence of SIV encephalitis in these macaques, the studies have used CD8⁺ T cell depletion [22^•,24^•], a neurotropic strain of SIV [21^•], or SIV serially passaged in microglia [23^•]. These studies have demonstrated that reducing plasma virus loads with antiretroviral therapy results in reduced virus loads in the brain, even with antiretroviral agents that have reportedly poor central nervous system penetration, such as PMPA and the protease inhibitors [21^•–23^•]. However, viral DNA (vDNA) levels were not different than in the control, no treatment group, even when treatment was initiated during acute infection, suggesting early establishment of the brain as a reservoir of persistent virus [21^•].

RT-SHIV

Efavirenz, a non-nucleoside reverse transcriptase inhibitor (NNRTI), is recommended in combination with two NRTIs as an initial HAART regimen, providing the most suppressive and stable treatment regimen for HIV-1 in most individuals (reviewed in [1]). Because the NNRTIs are specific for HIV-1, SIV is not susceptible to this class of antiretrovirals. As a result, the SIV macaque model is limited by the inability to analyze viral persistence during treatment with this clinically important HAART regimen.

In an attempt to generate a nonhuman primate animal model of HIV-1 that is susceptible to the NNRTIs, a chimeric virus was generated consisting of the SIV-mac239 backbone containing the HIV-1 reverse transcriptase (RT-SHIV) [25,26]. Virus stocks obtained from this clone contain a T-to-C substitution at position eight of the viral tRNA primer binding site, which is required for efficient viral replication in vivo [27,28,29^•]. Infection of rhesus macaques with RT-SHIV resulted in the AIDS-like symptoms and pathology that is observed in SIV infection [26]. Unlike SIV, RT-SHIV was susceptible to nevirapine, an NNRTI [25]. It was later demonstrated that RT-SHIV is susceptible to efavirenz, and that in monotherapy the course of treatment, including the selection of virus with reduced drug susceptibility, is similar to that observed in humans [30].

A sensitive viral load assay was developed for the detection of SIV and RT-SHIV that utilizes TaqMan RT-PCR and has a limit of detection of 50 vRNA copies per ml plasma [31]. The addition of the NRTIs PMPA and 3TC to the efavirenz regimen resulted in the suppression of plasma virus loads below the level of detection of this assay in the treated macaques [28]. Occasional detectable levels of viremia, or ‘blips’, and rebound of viremia upon cessation of HAART demonstrate that RT-SHIV persists in macaques despite long-term virus load suppression with a clinically important HAART regimen [28].

Recently, in an attempt to identify viral reservoirs during HAART, the RT-SHIV-infected rhesus macaque model was used to measure the distribution of vRNA and vDNA in tissues from treated macaques [29^••]. The results demonstrated widespread distribution of both vRNA and vDNA in many tissue compartments, especially in the gastrointestinal tract and several other lymphoid tissues, despite plasma virus levels below 50 vRNA copies per ml [29^••]. Although vDNA was detected in various neurological and reproductive tissues, vRNA levels were below the level of detection in all but two of these samples (prostate and cerebellum in the same animal), suggesting that these tissues might be sites of viral latency [29^••]. In addition to the analysis of lysates of whole tissue pieces, single-cell suspensions were generated from PBMC, mesenteric lymph node, spleen, and jejunum to allow for the enrichment of resting CD4⁺ T cells [29^••]. The resting populations were enriched for vDNA but contained little vRNA, suggesting that these resting CD4⁺ T cells are a latent reservoir, as it has been observed in humans.

Another measure of viral persistence during HAART that has been well studied in humans is the rate of decay of plasma vRNA [32,33]. With the development of a virus load assay with a single-copy limit of detection [34], it was demonstrated that the initial biphasic decline in plasma virus load is followed by a very slow third phase of decay and a fourth phase with little further decay despite years of virus load suppression by HAART, suggesting stable life-long infection [35]. This assay was recently adapted for the RT-SHIV macaque model and demonstrated that plasma vRNA decays in a biphasic manner to a stable, low-level that persists during HAART [36^•]. The study also determined that although the macaques that were studied longitudinally during HAART had similar average virus loads, they varied in terms of the maintenance of virus load suppression [36^•]. These data suggest that parameters other than average virus load will need to be addressed in studies aimed at enhancing HAART to reduce low-level viremia [36^•].

Although the previously mentioned studies using this RT-SHIV in rhesus macaques utilized intravenous inoculation as the route of transmission, a recent study demonstrated mucosal transmission of RT-SHIV in pig-tailed macaques [37]. The results demonstrated that although all macaques became infected following intravaginal challenge, pretreatment with depot medroxyprogesterone acetate (Depo Provera) resulted in plasma virus loads with a higher average set-point and increased the incidence of AIDS-like disease [37]. Together, these results suggest that damaging the vaginal mucosa prior to transmission can affect the course of disease.

Another strain of RT-SHIV has also been used to study NNRTI-based HAART regimens in pig-tailed macaques. This virus, designated RT-SHIV_mne, also requires the T-to-C substitution in the viral tRNA primer binding site for rapid in vivo replication [38]. RT-SHIV_mne is sensitive in vitro to several NRTIs as well as the NNRTIs efavirenz and nevirapine [38]. In a study designed to model treatment to prevent mother-to-infant transmission, it was demonstrated that a single-dose of orally delivered nevirapine did not inhibit RT-SHIV_mne in pig-tailed because it failed to achieve effective systemic concentrations [39]. Also, as was observed in RT-SHIV-infected rhesus macaques, efavirenz-based monotherapy in RT-SHIV_mne-infected pig-tailed macaques selected for virus containing mutations similar to those identified in HIV-1 that are associated with reduced drug susceptibility [39]. A genotypically well characterized stock of RT-SHIV_mne was recently used to study population dynamics in macaques [40^••]. This study analyzed viral evolution, including resistance to antiretrovirals, by tracking viral genomes during efavirenz monotherapy followed by the addition of PMPA and FTC [40^••]. Genotypic tracking of virus in this model might determine tissue sites of persistent virus, providing details of residual replication and viral latency that will be critical for HIV-1 eradication.

A simian-tropic strain of HIV-1 (stHIV-1) is being developed in pig-tailed macaques [41]. This chimeric virus consists of the HIV-1_NL4–3 backbone containing a substituted macaque-adapted HIV-1 env (pSHIV-KB9) in addition to an SIV vif gene [41]. Although this virus-infected pig-tailed macaques, verified by peak viremia, the animals in the study were able to control viral replication in the absence of antiretroviral agents [41]. Further development of stHIV-1 might allow for increased in vivo replication and pathogenesis.

Conclusion

The mechanisms of viral persistence are not completely understood, although a long-lived viral reservoir of cellular latency in resting CD4⁺ T lymphocytes has been demonstrated. Whether additional reservoirs persist, such as infected macrophage or other long-lived cells, is not known. Another possible mechanism of viral persistence is residual replication in cells or tissues due to restricted drug access or limitations in metabolic activation of antiretrovirals. The various nonhuman primate models of HIV-1 are currently being used to address these critical questions concerning viral persistence. An appropriate model of HIV-1 should be susceptible to current HAART regimens, share cellular tropism, and have similar disease manifestations as HIV-1 in humans. Several models have been developed and each has particular strengths (Table 1). These models have the advantage of allowing studies that are not feasible in humans. Through in-depth tissue analysis, the models have demonstrated widespread vDNA and vRNA in multiple tissues during HAART. Because of the capability to infect animals with a defined viral population, the models allow for tracking of viral genomes throughout the course of infection, including the inoculation or selection of virus with reduced drug susceptibility. This viral sequencing approach might help to identify the tissue reservoirs and cell types responsible for the observed residual viremia. Several studies in macaques have now demonstrated that virus persists in resting CD4⁺ T lymphocytes despite treatment with different combinations of antiretrovirals. These models are particularly well suited for the study of high-risk regimens aimed at eliminating the latent reservoir and preventing any residual replication. A better understanding of viral persistence using SIV Macaque models could play a critical role in the eradication of HIV-1.

Table 1.

Summary of the macaque models of HIV-1 infection, therapy and latency

	SIV	RT-SHIV	RTEnvSHIV	RT-SHIVmne	stHIV-1
Acute Infection
Rhesus	+	+	+	ND	ND
Pig-tailed	+	+	ND	+	+
Chronic infection/AIDS
Rhesus	+	+		ND	ND
Pig-tailed	+	+	ND	+
Mucosal Transmission	+	+	+	ND	ND
ART susceptibilities
NRTI	+	+	+	+	+
NNRTI		+	ND	+	+
PIa	+/−	+/−	ND	+/−	ND
INSTI	+	+	ND	+	ND
Entry/fusion inhibitor	-	-	+	ND	ND

ART, antiretroviral therapy; INSTI, integrase strand transfer inhibitor; ND, not determined; NNRTI, nonnucleoside reverse transcriptase inhibitor; NRTI, nucleoside reverse transcriptase inhibitor; PI, protease inhibitor; SIV, simian immunodeficiency virus; stHIV-1, simian-tropic strain of HIV-1.

^aSIV is susceptible to some PI in monotherapy. Several studies have demonstrated that SIV is inhibited by combinations of antiretroviral agents that include PI.