Preservation of Lymphopoietic Potential and Virus Suppressive Capacity by CD8⁺ T Cells in HIV-2–Infected Controllers Free

In the absence of an effective vaccine that prevents the HIV-1 acquisition, the long-term suppression of viral replication in infected individuals remains a major objective to prevent further disease progression and dissemination. Although progression to AIDS can now be significantly delayed by effective combinatorial antiretroviral therapy (cART), HIV remission, which is observed in HIV-1 controllers (HIC1) (1) and rare patients who gain independence from ART after early initiation (2), remains a priority. Much effort, therefore, has been given to the identification of immune correlates of natural HIV control in an attempt to decipher the immunological mechanisms that underlie effective control of HIV-1 replication in the absence of ART. A better understanding of these correlates would provide insights into therapeutic approaches that would modulate the immune system to mimic the immune parameters of HIV-1 controllers, ultimately leading to a functional cure of HIV-1 infection.

Although HIV-2 shares the same modes of transmission and intracellular mechanisms of replication as HIV-1, it provides a unique model of attenuated infection by a HIV. Compared with HIV-1 infection, HIV-2 infection is predominantly characterized by a slower decline of CD4⁺ T cells and lower levels of immune activation (1–3). Although HIV-2–infected patients who progress toward disease present similar clinical manifestations and AIDS severity as those infected with HIV-1, they remain a minority [∼15–20% of those infected (4)]. Many HIV-2–infected individuals spontaneously control their infection and remain asymptomatic while maintaining undetectable viral loads, signifying a high prevalence of HIV-2 controllers (HIC2) [9.1% of HIC2 in the ANRS CO5 HIV-2 cohort (5) versus 0.22% HIC1 in the ANRS CO4 French Hospital Database on HIV (6)]. Deciphering the immune correlates of HIV-2 control and contrasting these findings with those from HIV-1 controllers may illuminate the key features of HIV pathogenesis and provide critical information on what is needed to establish natural control of HIV.

It is well established that CD8⁺ T cell–mediated immunity is critical in the control of HIV-1 replication from the earliest days of acute HIV-1 infection. However, the HIV-specific T cell response declines over time and is often lost in the later phases of chronic infection (7). Moreover, HIV-1–infected individuals with protracted infection usually demonstrate reduced thymic output and lower naive T cell numbers (8–10). This situation, which is reminiscent of immune aging, likely accounts for the loss of CD8⁺ T cell regenerative capacity over time and the failure of the T cell response to adapt to the emergence of escape variants. Only rare HIC1 [often possessing specific “protective” HLA class I alleles (11)] are able to maintain viremia spontaneously and durably at extremely low levels (12). Furthermore, this level of control has been associated with their capacity to develop and sustain the production of HIV-specific CD8⁺ T cells that are endowed with the capacity to efficiently suppress HIV-1 infection of autologous CD4⁺ T cells (13).

In the present work, we focus on the potential role of cellular immunity as a key player in controlling HIV-2 infection. Strong Gag-specific CD8⁺ T cell responses have been previously observed in HIV-2–infected patients, and their magnitude was inversely associated with viremia (14, 15). Of note, these HIV-2 Gag–specific CD8⁺ T cells were described as polyfunctional, although they did not contain high levels of cytolytic markers (16). Persistent CD8⁺ T cell renewal is therefore likely to be essential for maintaining long-term immune efficacy against HIV. We embarked in this study on the fine characterization of HIV-2–specific CD8⁺ T cells in HIC2 from the French ANRS CO5 HIV-2 cohort, with particular focus on the production of HIV-specific CD8⁺ T cells and their HIV-2 suppressive capacity. To our knowledge, we describe for the first time that CD8⁺ T cells from HIC2 demonstrate strong capacity to suppress HIV-2 infection in autologous CD4⁺ T cells ex vivo. Importantly, HIC2 retain robust lymphopoiesis in general and CD8⁺ T cell production in particular, suggesting that the marked premature immune aging phenotype seen in HIV-1 infection is not present in these individuals, which is likely to be key to the maintenance of a durable and efficient CD8⁺ T cell response. Overall, our data suggest that the immune correlates of effective control in HIV-2–infected individuals may derive from the capacity of the immune system to maintain the potent HIV suppression capacity that is mediated by CD8⁺ T cells.

HIV-2 controllers (HIC2) were part of the ANRS CO5 VIH-2 cohort and included in the ANRS IMMUNOVIR 2 study, which focuses on the study of patients with nonprogressive infection. All patients in the current study (Table I) had characteristic features of HIV controllers (i.e., asymptomatic treatment naive individuals, infected for at least 5 y, with a CD4⁺ T cell count > 400 cells/μl and a viral load < 400 RNA copies/ml).

HIC2 were compared with HIV-1–infected individuals: HIV-1 controllers (HIC1) from the ANRS CO21 Codex cohort and two groups of HIV-1 viremic individuals (one with CD4⁺ T cell count < 200 cells/ml and another with CD4⁺ T cell count > 500 cells/ml) from the Pitié Salpêtrière Hospital (France). Mononuclear cells were isolated over a Lymphoprep gradient and then either used directly or cryopreserved. PBMC from uninfected healthy adults from the French Blood Bank (Etablissement Français du Sang) were also analyzed. A summary of clinical attributes of the patients studied is displayed in Table II.

All participants gave their written informed consent. The study was approved by the institutional ethics committee (i.e., Comité de Protection des Personnes of Ile de France XI).

To generate tetramers, immunodominant HIV-2 epitopes were identified by screening HIV-2 p27 overlapping peptides in IFN-γ ELISPOT assays using HLA-typed HIV-2–infected patient PBMC, going on to define the optimal epitope sequences and length (15). HLA-B*5301 p27 TPYDINQML (TL9) and p27 DRFYKSLRA (DA9) tetramers were synthesized as described previously (17). PBMC were stained with pretitrated concentrations of pentamer/tetramer (conjugated to PE), followed by a panel of Abs as described previously (18). Directly conjugated and unconjugated Abs were obtained from the following vendors: BD Biosciences: CD34 (PE), lineage mixture (CD3, CD14, CD16, CD19, CD20, and CD56/FITC), CD3 (PerCP-Cyanine 5.5 or Alexa Fluor 700), CD45RA (FITC), CCR7 (PE-Cyanine 7), CD107α (PE-Cyanine 5), IFN-γ (Alexa Fluor 700), and TNF-α (PE-Cyanine 7); Beckman Coulter: CD28 (PE-Texas Red), CD45RA (PE-Texas Red); Caltag Laboratories (Thermo Fisher): CD8 (Alexa Fluor 405) and granzyme B (PE-Texas Red); R&D Systems (R&D Systems Europe): MIP-1β (FITC); and BioLegend (Ozyme): CD127 (BV-650), and CD27 (Alexa Fluor 700). Cell surface marker stainings were performed using standard methodologies. CD34⁺ cell phenotyping was performed on enriched populations. The immunomagnetic enrichment of CD34⁺ cells was carried out on PBMCs using MACS technology, according to the provider’s recommendations (Miltenyi Biotec). Stainings were analyzed on an LSR Fortessa flow cytometer (BD Biosciences), and data were analyzed using FlowJo v8.2 (FlowJo) and Diva (BD Biosciences) softwares.

Purified PBMC were thawed and rested overnight at 37°C in complete RPMI medium (RPMI 1640 supplemented with 10% heat-inactivated FCS, l-glutamine, and antibiotics); viability was then examined by trypan blue exclusion (typically ≥70% viable). For stimulation, cells were then incubated in the presence of 15-mer overlapping peptides covering the HIV-2 p27 protein (10 μM) or an overlapping peptide pool encompassing clade B HIV-1 Gag (2 μg/ml). p27-stimulated cells were incubated in the presence of anti-CD107α Abs for 1 h at 37°C in a 5% CO₂ incubator, followed by an additional 5 h in the presence of the secretion inhibitors monensin (2.5 μg/ml; Sigma-Aldrich) and brefeldin A (5 μg/ml; Sigma-Aldrich). BD Cytofix/Cytoperm (BD Biosciences) was used for permeabilization of the cells prior to staining for intracellular markers.

To assess the capacity of CD8⁺ T cells to suppress HIV-2 infection of autologous CD4⁺ T cells in vitro, we adapted our previously published HIV-1 suppression assay (19). Briefly, PBMC were isolated from peripheral blood by density centrifugation. CD4⁺ and CD8⁺ T cells were then isolated by, respectively, positive and negative magnetic bead sorting (StemCell Technologies). CD4⁺ T cells were activated in complete RPMI 1640 medium supplemented with PHA (1 μg/ml) and IL-2 (100 IU/ml) for 3 d. CD8⁺ T cells were cultured in nonsupplemented culture medium. Activated CD4⁺ T cells were infected with HIV-2 SBL (20) using a spinoculation protocol (21) and cultured alone or with autologous CD8⁺ T cells at a 1:1 ratio during 14 d in IL-2 (100 IU/ml) supplemented culture medium. Viral replication was measured by p27 production in culture supernatants every 3–4 d as determined by ELISA (Gentaur). The capacity of CD8⁺ T cells to suppress HIV infection was calculated at the peak of viral replication as the log decrease in p27 production when superinfected CD4⁺ T cells were cultured in the presence of CD8⁺ T cells (log[p27 in CD4 T cell culture/p27 in CD8:CD4 1:1 cocultures]).

Thymic function was estimated by quantification of signal joint T cell reception excision circles (sjTREC) as described previously (22). Briefly, PBMC lysates were subjected to multiplex PCR for 22 cycles using sjTREC and CD3γ-chain outer primer pairs. Each of the amplicons was then quantified using Light Cycler technology (Roche Diagnostics), performed on 1/100th of the initial PCR, in independent experiments, but on the same first-round, serially diluted standard curve. This highly sensitive, nested quantitative PCR assay allows detecting one copy of sjTREC in 10⁵ cells. The sjTRECs were quantified in triplicate.

Statistical analyses were performed using Prism software (GraphPad). Nonparametric tests of significance were performed throughout all analyses, using Kruskal–Wallis and Mann–Whitney testing for intergroup comparisons and Spearman rank test to determine correlations. A p value > 0.05 was considered not significant.

We studied 20 HIV-2–positive individuals with controlled viremia in the absence of cART. The detailed characteristics of the patients are reported in Table I and Table II. The proportion of female patients was 60% (n = 12); the median age was 49 y (interquartile range [IQR]: 42–52). At inclusion, the median time since HIV-2 infection diagnosis was 13.6 y (IQR: 9.7–19.3), and the median CD4⁺ T cell counts were 893 cells/μl blood (IQR: 707–1170 cells/μl). Two individuals had detectable, albeit low, HIV-2 viral load (54 and 117 RNA copies/ml). Three individuals were born in France, 1 in Colombia, and 16 in West African countries (Table I).

Only two of the HIV-2–infected patients studied carried the “HIV-1 protective” B*57 allele and just one the B*27 allele. The most common class I HLA allele was C*04, which was carried by nine individuals (45%), of whom five were homozygotes (Table I). All five C*04/C*04 individuals were born in Ivory Coast. C*04 is a relatively common gene in France and West Africa. Overall, 11 (55%) individuals expressed homozygous alleles, suggesting that these individuals came from bottleneck populations with limited genetic diversity.

We previously showed that exhaustion of lymphopoiesis is a major correlate of disease progression in HIV-1 infection (9, 23). Our analyses highlighted a marked decline of primary immune resources in HIV-1 progressors, including CD34⁺ hematopoietic progenitor cells (HPC) and naive CD4⁺ and CD8⁺ T cell numbers. These alterations, which are the hallmark features of advanced immune aging, were not observed in HIC1. We therefore performed similar analyses in HIC2, directly analyzing the frequency and phenotype of CD34⁺ HPC (Fig. 1A, 1B), as well as the frequency of naive CD8⁺ T cells from blood (Fig. 1C). The frequency of circulating HPCs (CD34⁺Lin⁻) in HIC2 was high compared with levels found in untreated HIV-1–infected patients with low CD4⁺ T cell counts and equivalent to those observed in HIC1 (Fig. 1A). In a similar manner, HIC2 displayed a high proportion of lymphoid precursor progenitor cells that were identified by the CD117⁻CD45RA⁺ phenotype among CD34⁺ cells (Fig. 1B). Moreover, we found high levels of naive lymphocytes within the CD8⁺ T cell compartment of these individuals (Fig. 1C). Our results demonstrate the maintenance of effective lymphopoiesis in HIV-2 infection, which is likely to be linked with a capacity to sustain the long-term production of naive CD8⁺ T cells.

In HIV-1 infection, the progressive decrease in naive T cell proportions in HIV-infected adults likely results from the Ag-driven maturation of these cells in the absence of adequate T cell renewal capacity because of the declining thymic output of new cells (8, 10, 24). We and others previously reported that thymic function is preserved in HIV-2–infected individuals with low viral load (25). It was also recently shown that HIV-2 infection in thymocytes is impaired (26, 27), which may contribute to the maintenance of high CD4⁺ T cell counts in HIV-2–infected patients. In this study, we measured sjTREC levels in total PBMC to estimate thymic function in HIC2. Naive CD8⁺ T cell counts directly correlated with the amount of sjTREC per milliliter (Fig. 1D), suggesting that the maintenance of these cells was related to sustained thymic activity. This is in line with the correlation between naive CD8⁺ and naive CD4⁺ T cell counts in blood taken from the same individuals (Supplemental Fig. 1).

In addition, we assessed the levels of recent thymic emigrants using CD31 expression (28), which primarily provides information on the thymic history of naive CD4⁺ T cells. In line with the TREC data, a strong correlation was observed between the frequency of recent thymic emigrants (i.e., CD31⁺naive CD4⁺ T cells) and the naive CD8⁺ and CD4⁺ T cell counts (Fig. 1E, Supplemental Fig. 1). Taken together, these data support the maintenance of a robust lymphopoietic capacity and thymic output in HIC2 and signifies robust preservation of the naive CD8⁺ T cell compartment. The latter is central to the mounting of de novo responses, as recently shown in elderly subjects (29) and old vaccinated primates (30), and the ability to sustain effective T cell responses against a virus with high evolutionary rates, as is the case with HIV.

We next performed a comprehensive characterization of HIV-2–specific CD8⁺ T cell responses. Because the IFN-γ response of Gag-specific CD8⁺ T cells was found to correlate inversely with viremia In HIV-2–infected individuals, suggesting an important role of these cells in controlling HIV-2 infection (15), we decided to focus on phenotyping CD8⁺ T cells that were specific for p27, which is the major component of HIV-2 Gag. To assess the overall magnitude of HIV-2 Gag–specific CD8⁺ T cells in our patients, we performed intracellular IFN-γ staining on CD8⁺ T cells from PBMC that were stimulated with 15-mer overlapping peptides that spanned the proteome of the HIV-2 p27 protein. Robust p27-specific CD8⁺ T cell responses were detected in most patients (Fig. 2A). These responses were of similar magnitude to those observed in HIC1, although a direct comparison cannot be made, considering that a different Ag would be used to stimulate HIV-1–specific responses. These cells also exhibited other effector functions, as revealed by costaining for TNF-α, MIP-1β, and CD107α (Supplemental Fig. 2), highlighting a polyfunctional profile that has been described previously (31). We furthered our investigation by using recombinant MHC class I–peptide tetrameric complexes, which more precisely capture the memory differentiation phenotype of p27-specific CD8⁺ T cells (Fig. 2B). We analyzed in particular CD8⁺ T cells that were restricted by HLA-B*5301 because B*53 was the most common allele in our group of patients. We were able to generate HLA-B*5301 tetramers with the immunodominant p27 epitope TPYDINQML (TL9). All of the seven HLA-B*53 patients tested positive for TL9-specific CD8⁺ T cells. We also detected HLA-B*1401–restricted p27 DA9 (DRFYKSLRA)–specific CD8⁺ T cells in two of two HLA-B*14 patients.

The differentiation phenotype of HIV-2–specific CD8⁺ T cells was assessed based on the surface expression of the receptors CD27, CD28, CCR7, CD45RA, and CD127 (Fig. 2C) and compared with the phenotype of HLA-B*2705–restricted p24 KK10–specific CD8⁺ T cells, the latter selected as the prototype of primary effector cells in HIV-1 infection (18, 32). Overall, both groups of HIV-specific CD8⁺ T cells displayed similar phenotypes, being mainly CD27⁺CCR7⁻CD127⁻, with slight but noticeable differences. For example, HIV-2–specific CD8⁺ T cells trended toward higher CD45RA expression. In HIV-1–infected patients, CD45RA expression was the hallmark of a rare population of resting cells that resemble long-lived memory cells and were particularly noticeable shortly after viral replication was curtailed by cART (33). Furthermore, the expression of CD28 was significantly higher on HIV-2–specific CD8⁺ T cells (Fig. 2D), in line with previous observations made with HLA-B*3501–restricted p27 NPVPVGNIY (NY9)–specific CD8⁺ T cells (34). This indicates that HIV-2–specific CD8⁺ T cells are in general less differentiated (35, 36), displaying an even younger phenotype than highly effective HIV-1–specific CD8⁺ T cells. Accordingly, HIV-2–specific CD8⁺ T cells are likely to retain proliferative potential, as suggested previously (34). Nevertheless, despite their early memory differentiation state, HIV-2–specific CD8⁺ T cells express high levels of the transcription factors T-BET and EOMES (Supplemental Fig. 3), which are key regulators of memory differentiation and the acquisition of effector functions (37). In accordance with the expression of these transcription factors, HIV-2–specific CD8⁺ T cells possessed high intracellular levels of granzyme B, at least relative to what is seen in effective HIV-1–specific CD8⁺ T cells (Fig. 2D). Taken together, these data support the presence of robust Gag-specific CD8⁺ T cell responses in HIC2. These cells are characterized by an early or young differentiation phenotype and display strong effector potential.

Last, we aimed to show that this strong phenotypic functional potential translated into an effective capacity of these cells to suppress HIV-2 infection. We previously showed that HIC1 often possess HIV-specific CD8⁺ T cells with a striking capacity to suppress infection by killing HIV-infected autologous CD4⁺ T cells (38, 39). This capacity was not found in cART-treated or untreated viremic HIV-1 patients, either during the acute or chronic phase of infection (38–41). HIV-1 suppression assays provide the most robust method to distinguish between the functional capacities of CD8⁺ T cells from HIV-1 controllers and noncontrollers (40, 42, 43). We thus sought to adapt this in vitro assay to the study of HIV-2 using an SBL virus strain (20), with supernatant p27 concentration as a readout.

We found that CD8⁺ T cells from HIC2 have a strong capacity ex vivo to suppress HIV-2 infection of autologous CD4⁺ T cells (Fig. 3A), with a median reduction in p27 production of 2.7 logs [IQR: 1.3–3.3]) (Fig. 3B). The viral suppressive capacity shown by the CD8⁺ T cells from HIC2 was remarkably high, comparable to the levels observed with cells from HIC1 from the ANRS CO21 cohort (Fig. 3B) or even among our previous studies (38, 39). Similarly to what we found in HIV-1 controllers (39), the capacity of CD8⁺ T cells from HIV-2 controllers to suppress infection was relatively stable over time (Fig. 3C). There was, however, some heterogeneity in HIV-2 suppressive capacity among patients, with logp27 decrease values ranging from 0.02 to 3.7 (Fig. 3B). Interestingly, in two patients who had the steepest decline in CD4⁺ T cell counts over the 3 y preceding study inclusion (92.5 and 215.7 cells/mm³/y), their CD8⁺ T cells showed no ex vivo capacity to suppress HIV-2 infection in our assay (Fig. 4A). Accordingly, we found a positive correlation between capacity of CD8⁺ T cells to mediate HIV-2 suppression ex vivo and the patients’ CD4 T cell counts (Fig. 4B), which supports the key role of these CD8⁺ T cells in suppressing HIV-2 infection and preventing disease progression. These results are in line with previous studies showing that strong CD8⁺ T cell–mediated HIV-1 suppression ex vivo correlates with better clinical outcome (38, 44, 45) and strengthen the value of the HIV-1/2 suppression assay as a correlate of HIV control.

Although the immune dynamics responsible for the slow progression of HIV-2 infection has remained puzzling for years, we are gaining momentum in piecing together the precise picture of events that are responsible for this largely nonprogressive HIV infection. At equivalent plasma viral loads, proviral DNA loads are very similar between HIV-1 and HIV-2 infections (1, 46–48). When left unchecked by cART, the survival of HIV-2–infected patients is strongly related to both CD4⁺ T cell counts and plasma viral load. However, in contrast to HIV-1, HIV-2 is often controlled to undetectable plasma viral loads in the absence of cART, which most likely involves a complex immune-related mechanism.

Owing to the maintenance of a strong lymphopoietic capacity, T cell–mediated immune responses appear to be better preserved in HIV-2 infection and HIC2 can sustain the robust suppression of viremia over several years or even decades of chronic infection. HIV-2–specific CD8⁺ T cells display an early or young differentiated phenotype, likely reflecting their potential for T cell renewal. They also harbor particularly potent effector functions, highlighted by their exceptional capacity to suppress the virus in autologous CD4⁺ T cells. These CD8⁺ T cell characteristics are similar to those observed in HIC1 in many aspects and are likely to be part of the cause or consequence of effective control over the virus. This robust lymphopoietic potential is not preserved or restored in the CD8⁺ T cells of HIV-infected patients undergoing antiretroviral treatment even when initiated during primary HIV-1 infection (41), which argues more in favor of causality.

Although overall, strong CD8⁺ T cell–mediated HIV-2 suppression is observed in HIC2 patients, some heterogeneity, as in HIC1, is observed in these patients (39, 49–51). The absent or low CD8⁺ T cell–mediated HIV-2 suppression in four patients may be due to several factors. First, we cannot exclude the possibility that strong CD8⁺ T cell viral suppression was not elicited in these individuals because of other immunodominant epitopes that were not encoded by the HIV-2 SBL virus strain used in our experiments. However, we have previously shown that suppressive CD8 T cell activity in HIV-1 controllers remained low in some donors even when autologous virus was used (39). Alternatively, the absence of suppressive activity may be a consequence of the stringent control of viremia, resulting in the contraction of the HIV-2–specific CD8⁺ T cell response to a small pool of high-quality memory cells that fall below the detection threshold of our assay (49, 52, 53). Moreover, we were only able to analyze the CD8⁺ T cell response in blood and would not be able to measure more robust responses that were active locally in patient tissues (54). Finally, other mechanisms such as innate or humoral responses, or infection with unfit virus, may exert a greater role than the activity of CD8⁺ T cells in controlling viremia in these patients.

The effective suppression of viral replication in HIC2 will prevent the development of chronic immune activation and further damage to the lymphopoietic system, thus allowing the maintenance of an effective CD8⁺ T cell immune response that provides long-term control of the virus. This equilibrium between HIV-2 and host immunity may be referred to as a virtuous cycle, in contrast to the vicious cycle in HIV-1 infection, where apart from exceptional HIC1 cases, poorly controlled viral replication results in elevated chronic immune activation. The intensity of prolonged HIV-1 replication contributes to a process of premature immune aging that exerts its toll and further weakens cellular immunity in infected individuals.

We do not yet have a complete picture to explain the virtuous cycle of HIV-2 infection; although many epidemiological studies have found an association between specific HLA class I alleles and HIV-1 disease outcome [i.e., rapid or delayed progression to AIDS (11)], the role of HLA polymorphism in HIV-2 is still unclear. Evidence exists that HLA-B*35 (55) and B*1503 (56) alleles are associated with HIV-2 disease progression; nevertheless, several individuals bore HLA-B*35 alleles in our study yet controlled their infection. Similarly, in the context of HIV-1 infection, HLA class I homozygosity is associated with rapid progression to AIDS (57, 58); however, most of the HIC2 in our study are homozygotes. Interestingly, no HLA allele associated with control of infection has been found in HIV-2 infection. It was suggested by Yindom et al. (56) that this might be because HIV-2 infection is easier to control compared with HIV-1 and therefore that less stringent requirements for control may not have led to the selection of “protective” epitopes that are restricted by specific HLA molecules during HIV-2 infection. In addition, unlike the partial control exerted by CD8⁺ T cells in HIV-1 infection (59), the robust CD8⁺ T cell control of HIV-2 could limit the emergence of escape variants and therefore abrogate the requirement for HLA class I molecules that possess sufficient flexibility to adapt to new variants.

Although still hypothetical at this stage, a greater sensitivity of HIV-2, compared with HIV-1, to intracellular restriction factors may play a crucial role in constraining the initial replication of HIV-2 in CD4⁺ T cells. It has been shown that HIV-2 is more susceptible to restriction by tripartite motif protein isoform 5 α (TRIM5a) than HIV-1 (60). Subsequently, it was found that tetherin interacts differentially with the two viruses depending on whether it is counteracted by Vpu or Env, as is the case in HIV-1 and HIV-2, respectively (61, 62). This difference might reflect a different sensitivity of HIV-1 and HIV-2 to tetherin control at the cell membrane. cART-naive HIV-2–infected patients from the ANRS CO5 HIV-2 cohort also demonstrated high levels of APOBEC3F/G (apolipoprotein B MRNA editing enzyme catalytic subunit 3F/G) editing activity in a previous study (63); however, immunovirological parameters were not found to be associated with the latter.

Finally, it has been proposed that the presence of the HIV-2/SIVsmm–specific Vpx, an accessory protein that inhibits the action of SAMHD1, enhances the sensing of HIV-2 by dendritic cells, which may promote the induction of more potent CD8⁺ T cell responses (64). This requires further investigation because other groups have reported that dendritic cells are refractory to HIV-2 infection, which could in turn, restrict HIV-2 replication (65, 66). Overall, an enhanced sensitivity of HIV-2 to host restriction factors may be crucial in limiting its replication initially, preventing further damage to the host immune system, providing the necessary window of opportunity for the subsequent induction and enactment of an effective T cell response that establishes and maintains control of the virus. Once activated, this timely bottleneck would prevent the overexertion and subsequent exhaustion of immune resources—usually related to the maintenance of the CD4 T cell pool and the consumption of HIV-specific immune cells—together with the presence of the associated hyperinflammatory status often seen in HIV-1 infection, ultimately preserving the immune functions of the infected person. Combined together, this synergy in innate and adaptive factors may provide the necessary conditions for to establish the HIV-2 virtuous circle and therefore exemplify a holistic mechanism for natural control of the virus.

We thank Valérie Monceaux for technical assistance and Assia Samri for help with inclusion of patients. We thank Ioannis Theodorou, Sabine Canivet, and Marie-Line Moussalli for HLA typing. The authors are indebted to Tao Dong and Yanchun Peng for providing tetramer reagents. The authors also thank Marie-Anne Rey-Cuille, Livia Pedroza-Martins, and Sandrine Couffin-Cadiergues for their support.

Victor Appay, Brigitte Autran, Amel Besseghir, Françoise Brun-Vezinet, Nathalie Chaghil, Charlotte Charpentier, Sandrine Couffin-Cardiergues, Rémi Cheynier, Diane Descamps, Anne Hosmalin, Gianfranco Pancino, Nicolas Manel, Lucie Marchand, Sophie Matheron, Marine Naudin, Livia Pedroza-Martins, Marie-Anne Rey-Cuille, Asier Sáez-Cirión, Assia Samri, Rodolphe Thiebaut, and Vincent Viellard

The authors have no financial conflicts of interest.