Where the Wild Things Are: Pathogenesis of SIV Infection in African Nonhuman Primate Hosts

More than 40 SIVs naturally infect wild African NHPs, with high prevalence [3]. In spite of being designated SIV, by analogy with HIV, these viruses generally do not induce immune suppression in their hosts [2]. This fundamental difference is equally fascinating and intriguing. Obviously, there is an important question raised by the lack of disease progression of SIV infection in natural hosts: are these highly divergent, apparently harmless viruses that infect distant monkey species instructive for HIV infection? Rapidly accumulating data on the pathogenesis of SIV infection in natural hosts and in new hosts upon cross-species transmission indicate that these “exotic” models may actually hold clues regarding the immune correlates of the control of SIV/HIV disease progression [2].

First, the viruses infecting African NHPs retained their pathogenic potential. This is illustrated by their ability to induce immunosuppression and disease progression in new hosts upon cross-species transmission [3]. HIV-1 and HIV-2, the two viruses causing AIDS in humans, resulted from cross-species transmissions from chimpanzees/gorillas and sooty mangabeys (SMs), respectively [3]. Also, accidental transmission of SIVsmm, which naturally infects SMs, to different species of Asian macaques resulted in progressive infections in the new hosts [4]. Moreover, rare AIDS cases were reported in several African NHP species infected with their species-specific or cross-species transmitted SIVs [5]. These cases were described during long-term follow-up after extensive incubation periods in captive animals that outlived the normal species lifespan in the wild. They recapitulate all virologic, immunologic, and clinical conditions that usually define AIDS in humans [5].

Second, the benign coexistence of SIVs and their African NHP hosts resulted from their interactions over millennia [3], and a long-term adaptation of the primate hosts to prevent the deleterious consequences of lentiviral infection. Although the mechanisms through which the nonpathogenicity was achieved are still under debate, this adaptation resulted in a pattern of chronic, persistent, life-lasting infection with an incubation period that normally exceeds the lifespan of the species [5]. Because this is the target of current therapeutic approaches in HIV infection, the study of SIV infection in natural hosts can provide invaluable clues with regard to which aspects of HIV pathogenesis have to be manipulated to successfully control disease progression.

The long-term virus–host coevolution left immunologic imprints in the natural hosts of SIVs.

The CCR5 expression (protein and mRNA) by CD4⁺ T cells from periphery, lymphoid, and mucosal tissues was found to be significantly lower in uninfected natural hosts, such as African green monkeys (AGMs), mandrills, SMs, solatus monkeys, and patas monkeys, compared with progressive hosts, such as rhesus macaques (RMs), pigtailed macaques, and humans [6]. Interestingly, the CD8⁺ T cells isolated from natural hosts express normal levels of CCR5, thus indicating that the finding is not a technical artifact [6]. Furthermore, CCR5 expression on the CD4⁺ T cells was particularly low in newborn AGMs and mandrills, compared with adults of the same species. Differently from RMs, in which CCR5 levels rapidly increase in infants of 6 months of age, reaching levels similar to those expressed by adults, such increases could not be observed in infant and juvenile African monkeys less than 3 years old [7•].

The red-capped mangabey (RCM) is an extreme example of a natural host that totally lacks CCR5 expression, as a result of a high prevalence (83%) in this species of a delta-24a deletion [8] that prevents CCR5 gene transcription. This situation recalls the minority of humans who have a delta-32 deletion in the CCR5 gene structure that renders them resistant to HIV infection or disease progression [9]. Differently from HIV-1, which mainly uses CCR5 as a coreceptor, SIVrcm (the virus that naturally infects RCMs without clinical consequences) adapted to use a CCR2 or CCR4 coreceptor to ensure persistence in the RCM host [10].

These observations led to our original hypothesis that the changes in CCR5 expression in natural hosts reflect the coevolution of the viruses with their hosts [6]. This hypothesis is supported by the observation that, in baboons, CCR5 expression on CD4⁺ T cells is similar to that observed in humans and macaques [6]. Baboons are not natural hosts for SIV and progress to AIDS upon SIV/HIV-2 infections [5].

In chimpanzees, CCR5 expression on CD4⁺ T cells is at intermediate levels between humans and natural NHP hosts [6], in line with molecular analyses that showed that chimpanzees acquired SIVcpz through cross-species transmissions from other monkey species, thus emerging more recently than the SIVs infecting African monkey species [3]. Moreover, SIVcpz is pathogenic in chimpanzees [11•].

To conclude, CCR5 expression by CD4⁺ T cells correlates with the ability of NHP species to develop or resist disease progression when infected with HIV/SIV.

The second difference in the immune system of the natural versus progressive hosts is that some African NHP species, such as AGMs, patas, and solatus monkeys, have low numbers of CD4⁺ T cells [12]. These low levels of CD4⁺ T cells appear to be characteristic only for African guenons (tribe Cercopithecini), as SMs and mandrills (tribe Papionini) were reported to have higher CD4⁺ T cell counts, similar to RMs, which belong to the same tribe.

Over the years, it was also reported that a large proportion of the CD3⁺ T cells do not express either CD4 or CD8, in natural hosts [13, 14•, 15, 16]. Recently, these double-negative cells where shown to arise through CD4 downregulation, as the cells enter the memory pool. This downregulation is independent of SIV infection. The CD4^neg memory T cells maintain functions attributed to CD4⁺ T cells, including production of interleukin (IL)-2 and IL-17, histocompatibility complex class II restriction, and expression of CD40 ligand and of forkhead box P3 [13, 14•]. In contrast with the double-negative T cells observed in HIV-infected patients (which also occur through downregulation of CD4 [17]), the double-negative cells from AGMs are not SIV infected [14•]. Preservation of a cell subset that maintains CD4⁺ T-cell function while being resistant to SIV infection in vivo may be responsible for the lack of disease progression in natural hosts. The precise impact of the double-negative T-cell population for SIV transmission and pathogenesis remains to be determined.

Recent studies reported differences in the functionality of dendritic cells between pathogenic and nonpathogenic hosts. Thus, two studies reported that AGM and SM plasmacytoid dendritic cells (pDCs) produce markedly less interferon (IFN)-α in response to SIV and other Toll-like receptor (TLR) 7 and 9 ligands ex vivo [18, 19]. Such an observation may suggest that, in contrast with progressive hosts in which chronic stimulation of pDCs by SIV/HIV may drive the unrelenting immune system activation and dysfunction underlying AIDS progression, such a vicious cycle of continuous virus replication and immunopathology is absent in natural hosts.

These particularities of the immune system of natural hosts may have significant impact on the pathogenicity of SIV infection in these species (Table 1). One can imagine that the low levels of target cells (cells expressing CD4 and CCR5) may result in lower rates of SIV transmission or lower viral replication in natural hosts than in new hosts. Also, the low expression of CCR5 on the CD4⁺ T cells and a reduced capacity of the innate immunity to activate the adaptive immune system may result in lower levels of immune activation in the natural host than in progressive infections.

The observation that natural hosts of SIV have low levels of CCR5⁺CD4⁺ T cells at mucosal sites raised the question of whether the virus can be transmitted through mucosal routes in these species. This hypothesis was also suggested by the observation that in offspring and juvenile monkeys, which have even lower levels of CCR5⁺CD4⁺ cells at mucosal sites, breastfeeding transmission is significantly lower than in pathogenic infection [7•]. However, the prevalence of SIV infection in wild populations of natural hosts significantly increases in adult animals, suggesting sexual transmission [3].

Experiments in mandrills failed to demonstrate any breastfeeding transmission of SIVmnd-1, even when mandrill dames were infected on the day of the delivery [7•]. In spite of very high viral load (VL) levels during the acute and chronic infection of the dames, and of high levels of replicating virus in milk during acute infection, no case of SIVmnd-1 transmission to offspring was observed during the lactation period. The design of this study, with dames being infected after delivery, precluded protection of the offspring through passively transmitted maternal antibodies. The lack of breastfeeding transmission in natural hosts was therefore correlated with the low CCR5 expression at mucosal sites [7•].

In a recent experiment, we assessed for the first time the efficacy of mucosal SIV transmission in a natural host by performing a dose-escalation study, in which AGMs and pigtailed macaques were intrarectally exposed to SIVagm (Pandrea, Unpublished data). This experiment demonstrated that natural hosts of SIVs can be infected by mucosal transmission, but at higher doses compared with pigtailed macaques. The efficacy of transmission was lower in juvenile animals than in adult AGMs, also suggesting that transmissibility depends on the availability of target cells at the mucosal sites. All these findings point toward a strong correlation between the levels of CCR5 expression on the CD4⁺ T cells and the rates of SIV transmission. Our studies also suggest that blockage of CCR5 may be an efficient treatment in preventing breastfeeding transmission, in addition to vaginal or rectal transmission.

Despite the low levels of target cells, SIV replicates to high levels in natural hosts [20]. During acute infection, VLs are in the same range as those reported for pathogenic HIV/SIV infections [21]. During chronic infection of natural hosts, the levels of viral replication are slightly higher [20] than those observed in HIV-1–infected patients [22]. In contrast with pathogenic infections, the levels of viral replication are remarkably stable for decades during chronic infection of natural hosts [23, 24].

There are several observations that have to be stressed with regard to SIV replication in the natural host. First, similar to humans and macaques, plasma viremia during chronic infection is predictive of disease progression. The levels of chronic viral replication were significantly higher in African monkeys that progressed to AIDS than in those that resisted progression [5]. Note, however, that even these progressor monkeys showed a remarkable resilience to high levels of viral replication, which were maintained for decades prior to disease progression [5]. Second, SIV VLs in natural hosts appear to be dependent on the availability of CD4⁺ T cells. Multiple lines of evidence support this conclusion: 1) In AGMs, a lower baseline level of CD4⁺ T cells associates steady-state levels of viral replication that are 1 log lower than in SMs and mandrills [20]; 2) In SMs and AGMs with severe CD4⁺ T-cell depletion after SIV infection, VLs decreased significantly [25]; 3) Experimental induction of immune activation in chronically infected AGMs resulted in a dramatic increase in the number of CD4⁺ T cells and VLs [26•]; and 4) As expected from previous studies in AGMs, experimental CD4⁺ T-cell depletion in SMs resulted in a significant decrease of VLs [27].

The low CCR5 expression observed in uninfected African NHPs does not appear to have a major impact on virus replication in the natural host. This paradox may be explained through several hypotheses. First, CCR5 expression may increase in response to SIV infection. The CCR5⁺ CD4⁺ T cells are targeted and depleted by the SIV. Therefore, SIV-associated changes in the CD4⁺ CCR5⁺ T-cell pool may reflect the opposing forces of both SIV replication (which reduces this cellular pool) and immune activation (which increases it). In AGMs, CCR5 expression by CD4⁺ T cells (both protein and mRNA levels) increases during acute SIVagm infection, as well as at certain time points during chronic infection. These increases in the CCR5 expression on the CD4⁺ T cells may be sufficient to expand the pool of SIV target cells and maintain the high VLs. Being only transient, these increases in CCR5 expression are not harmful to the hosts by not triggering excessive immune activation. Second, natural hosts may keep the CCR5 expression under control through an active mechanism, which remains to be determined. Finally, the lack of correlation between coreceptor expression and the levels of viral replication may be due to SIVs having other target cells in natural hosts.

Indeed, the paradox of SIV infection in natural hosts defined by association of a paucity of CD4⁺ CCR5⁺ target cells with high VLs also raised the question of whether in natural hosts SIVs could target cells other than CD4⁺ T lymphocytes. This hypothesis is also prompted by the observation that peripheral CD4⁺ T cells are well preserved in natural hosts during chronic infection, almost at preinfection levels. Such alternative potential target cell candidates may be either the macrophages or CD4⁺ T-cell subsets that do not express CCR5.

In some AGMs, the levels of SIVagm in the brain are as high as in humans with HIV-dementia [28], and the SIVagm strains isolated from the brain of infected AGMs are macrophage tropic. In the same animals, the SIVagm strains isolated from the lymph nodes were more restricted to growth in human T-cell lines [28]. These results suggest compartmentalization in the central nervous system without associated neuropathology and provide additional evidence that the nonpathogenic nature of SIV infection in natural hosts is not due to differences in the tissue and cell tropism from those for HIV-1–infected patients or SIV-infected macaques [28].

Similar to pathogenic infections, the bulk of viral replication is supported in natural hosts by lymphocytes rather than macrophages during both acute and chronic infection:

On the other hand, SIVagm.sab and other viruses infecting natural hosts were reported to be dual tropic in vitro and to use both CCR5 and CXCR4 [34], being therefore conceivable that the main target cells of these viruses are other CD4⁺ T-cell subpopulations (ie, those expressing CXCR4). However, in vivo, SIVagm severely depletes the CD4⁺ T cells from the intestine (which more frequently express CCR5) and not from the blood or lymph nodes (that have a higher CXCR4 expression) in both nonprogressive (AGMs) [29] and cross-species transmitted (RMs and pigtailed macaques) infections [29]. This observation suggests that the virus is in fact CCR5 tropic in vivo.

Altogether these data show that SIVs naturally infecting African NHPs have the same biology and similar target cells compared with the lentiviruses that infect humans or macaques.

Studies suggested that SIV infection in natural hosts is independent of immune responses. The main argument is that, differently from elite controllers in which the lack of disease progression is achieved through control of viral replication [35], such control is not observed in natural hosts [20]. Note, however, that the stability of VLs over decades strongly supports an active (albeit partial) control of viral replication in natural hosts.

Numerous studies determined that, in natural hosts, the levels of anti-SIV binding and neutralizing antibodies are either low or in the same range as those reported for HIV-1 infection [36]. Natural hosts usually lack anti-Gag antibodies [36, 37]. The significance of this observation is still debated, as the Gag antigens appear to be expressed, and the Gag-specific T-cell responses are comparable to those observed in pathogenic infections [37]. Neutralizing antibodies of variable potency can be found in the majority of SIV-seropositive AGMs [38, 39], and their detection is dependent on the virus strain and the cells used for the assay [38]. Most SIVsmm strains are resistant to neutralization [40]. These observations suggest that the low levels of antibodies or the weak neutralization activity may be responsible for the high viremia in these species [36]. Given the contradictory results of different studies, the potential role of humoral immune responses for the control of viral replication in natural hosts was assessed in AGMs by ablating anti-SIV antibody production through a long-term depletion of B cells. This study demonstrated that humoral immune responses do not play a determinant role in controlling viral replication in the natural host [39]. This is also true for SIVsmm-infected macaques that did not show significant changes in the levels of viral replication or disease progression when depleted of B cells and infected with a neutralization-sensitive SIVsmm strain [40].

In contrast, cellular immune responses observed in natural hosts seem to be involved in the control of viral replication [37, 41, 42], with a direct correlation between the levels of cellular immune responses and plasma VLs being reported in SMs [43]. However, cellular immune responses do not appear to be superior to those observed during SIV infection of RMs [43].

Recently, several groups performed experimental in vivo CD8 cell depletion studies in natural hosts and reported that administration of anti-CD8 monoclonal antibodies during both acute or chronic SIV infection resulted in increases in the levels of virus replication in both AGMs [44] and SMs [45]. During acute infection, CD8 cell depletion did not result in a significant increase of peak viremia, but in a plateau of high VLs after the peak, which were controlled with the recovery of CD8⁺ T cells. During chronic SIV infection, CD8 cell depletion in vivo also resulted in an increase of viral replication in natural hosts. Dual CD8⁺ and CD20⁺ cell depletion studies also pointed to cellular immune responses as the main factors involved in the control of viral replication in natural hosts [44]. None of the in vivo immune cell depletion studies reported any case of disease progression. This is in line with the above observation that even in the face of higher levels of viral replication, progression to AIDS occurs only after very long incubation periods in natural hosts [5].

The assessment of the various contributing factors that enable natural hosts of SIV to resist progression to AIDS thus revealed that the lack of disease progression in natural hosts is not due to 1) lower levels of SIV replication; 2) preservation of peripheral or mucosal CD4⁺ T cells during acute infection; 3) replication in cell populations that are different in natural versus pathogenic SIV infections; or 4) exquisite SIV-specific cellular or humoral immune responses.

The main difference identified between natural SIV infections and pathogenic HIV/SIV infections relied on the normal levels of immune activation and inflammation observed in natural hosts. Generalized immune activation is an important driver of viral replication disease progression in pathogenic HIV/SIV infections [1]. Chronic SIV infection in natural hosts is associated with 1) low levels of CD4⁺ and CD8⁺ T-cell activation and proliferation in the periphery, as well as in lymphoid and mucosal tissues [29, 46]; 2) no significant increased levels of CD4⁺ T-cell apoptosis in vivo or in vitro [29, 47]; 3) absence of cell-cycle dysregulation associated with pathogenic HIV/SIV infections [48]; 4) no significant increase in the levels of proinflammatory cytokines compared with uninfected monkeys [26•, 39]; and 5) preserved function of the T-cell regenerative compartment. During acute SIV infection of natural hosts, the levels of T-cell activation and proliferation, although transiently increased, are blunted compared with those observed during pathogenic SIVmac infection of RMs, with a lower peak and a more rapid and pronounced postpeak decline of immune activation [29].

The low-level immune activation in natural hosts does not appear to represent ignorance of, or tolerance to, SIV antigens because SIV-specific T cells are present during natural SIV infections and exert immune pressure on the virus [41]. Moreover, an anti-inflammatory milieu was reported to be rapidly established in AGMs after SIVagm infection, characterized by an early increase in numbers of Tregs, paralleled by an increase in transforming growth factor-β and IL-10 levels in the plasma very early after infection [49]. This anti-inflammatory milieu may actively attenuate the immune activation during natural SIV infections. These observations conflict with a recent study in SMs that reported that natural hosts intrinsically lack the ability to respond to TLR stimulation in vitro [18]. However, in contrast with the in vitro data showing the incapacity of pDCs to produce IFN-α, numerous recent in vivo studies reported a transient increase of IFN-α production during acute infection in AGMs and SMs and an increased capacity of pDCs to produce IFN-α after infection. These increases, however, are lower than those observed in SIVmac-infected RMs and of very short duration, with IFN-α production returning to baseline levels by the time viral replication is controlled.

Multiple mechanisms may be responsible for the lack of immune activation in the natural host (Fig. 1):