Mechanism of protection of live attenuated simian immunodeficiency virus: coevolution of viral and immune responses
2010; Lippincott Williams & Wilkins; Volume: 24; Issue: 5 Linguagem: Inglês
10.1097/qad.0b013e328337795a
ISSN1473-5571
Autores Tópico(s)Immune Cell Function and Interaction
Resumo'A good hockey player plays where the puck is. A great hockey player plays where the puck is going to be.' Wayne Gretzky The best evidence that it will be possible to develop an effective HIV vaccine is based on studies of the ability of live attenuated simian immunodeficiency virus (SIV) to provide protection from pathogenic strains of SIV [1,2]. Although safety concerns have shelved plans for using live attenuated HIV vaccines in humans [3–5], attenuated SIVs and the rhesus macaque model remain the best hope for understanding the mechanisms underlying effective HIV immunization. Here, I hypothesize that a surprisingly often-overlooked aspect of attenuated SIV biology is that although attenuated, the virus still retains sufficient replicative capacity to evolve in vivo[5–10] and this has important implications for understanding its mechanism of protection. For if attenuated SIVs evolve and diversify in response to immune pressure, so too will the immune system evolve and diversify in response to viral evolution. The end result of all this immune/viral competition will be a much broader and more balanced immune response than if viral immune escape had not occurred. Therefore, upon challenge with a pathogenic but closely related strain of wild-type SIV, the incoming virus quickly finds itself facing an immune response selected not only for its ability to contain the original virus but also the preferred immune escape variants of that virus. Under these conditions, viral escape is rendered more costly and effective immune control ensues. In contrast, other vaccine strategies rely on repeated immune stimulation with invariant antigen. This may inadvertently promote immunodominant epitopes at the expense of a broader immune response and lead to more easily escaped immune responses. Thus, it may be precisely those features that make live attenuated SIVs too dangerous to use in humans, that is their tendency to mutate and evolve [5], that makes them such effective immunogens. If this hypothesis is correct, then a challenge to HIV research will be to recreate immune responses that anticipate immune escape variants without the use of live retroviruses. Three predictions resulting from this hypothesis are: attenuated SIV undergo immune escape at a significant rate, immune responses toward attenuated SIV are very broad especially when judged relative to the amount of virus present, and HIV/SIV vaccine efficacy might be increased by employing likely virus-escape variants into vaccine protocols. Attenuated simian immunodeficiency virus mechanism of action: what matters? SIV-specific cytotoxic T lymphocytes (CTLs) are critical for immune control of both HIV and SIV [11–15] as well as for the protective efficacy of attenuated SIV [2,16,17].1 Yet, despite strong evidence for the importance of CTLs in controlling HIV/SIV infection, it remains unclear why other vaccine strategies, especially those that elicit CTL frequencies similar to those of attenuated SIV, do not provide anywhere near the level of protection as do live attenuated SIV [2,23]. A number of explanations could account for this. One hypothesis is that because attenuated SIVs traffic to the same anatomical sites and infect the same cells as do wild-type SIV, they better prime immune responses at relevant sites [2,24–27]. Alternatively, attenuated SIV may elicit immune responder cells that are intrinsically different from those of other vaccine protocols, for example, in their affinity or ability to secrete cytokines, expand, or lyse target cells [2,23,28–33]. Yet although these explanations make intuitive sense and probably do account to varying degree for the efficacy of live attenuated SIV, they do not readily explain all observations. In particular, there is a relatively long 'incubation period' following attenuated SIV vaccination during which the protective efficacy of the immune response improves (but does not correlate with the appearance of neutralizing antibodies [34]).2 This prolonged period can vary from nearly 3 months to over a year and depends on the degree of SIV attenuation as well as the nature of the challenge virus, with more attenuated SIVs [34,36] and more heterologous challenge viruses [37,38] requiring the longer time frame (if protection against heterologous strains arises at all). This observation is surprising because in animals infected with attenuated SIV strains, the level of virus declines within weeks of vaccination, often to undetectable levels [34,36,39,40]. Therefore, it might be expected that the earlier immune responses, closer to both viral and CTL acute-phase peaks and having just fought off a highly similar virus, would be better equipped than the diminished response at over 1 year to control wild-type challenge, but that is not what is found [36]. This observation suggests that the immune response slowly evolves over time, perhaps reflecting the persistent nature of the live attenuated SIV vaccine. Consistent with this interpretation, attenuated SIV vaccines significantly outperform single-cycle SIV vaccines [41] and their protective efficacy declines as their degree of attenuation increases [42]. A second paradox is why does the immune system following attenuated SIV vaccination generally confers complete protection against homologous wild-type SIV challenge, yet often fails to clear or contain the original 'vaccine' attenuated strain (15–20% of monkeys infected with attenuated SIV die of AIDS-like disorders and an even larger proportion show signs of immune dysfunction) [5,43]? Indeed, reversion mutations that restore wild-type Nef function occurs in some animals many weeks after sterilizing immunity to wild-type challenge normally develops [6,37,44] and disease progression may occur a decade or more after initial attenuated virus infection [43,44]. This is perplexing, for if mere persistence accounted for the superior protective efficacy of live attenuated SIV, then it might be expected that the immune responses elicited should more strongly target the attenuated virus against which it was raised and which lacks the immune protective effects of Nef major histocompatibility complex (MHC) downregulation [8,45] rather than the wild-type virus. And although it may be argued that the attenuated SIV persists because it spreads by more efficient cell-to-cell routes, as opposed to the challenge virus that initially infects by a cell-free route, in at least a few cases the protection observed was not sterilizing and still the attenuated virus apparently won out [34,36]. This suggests that in many other cases, wild-type infection probably did take hold but never rose above the threshold of detection and was then out-competed by the attenuated virus. A second problem of attributing the greater protective efficacy of attenuated SIV to viral persistence alone is that at least up until now the protective efficacy of SIV antigens expressed within the context of a persistent viral infection other than SIV has not nearly approached that of live attenuated SIV. For example, a cytomegalovirus (CMV) vector expressing multiple SIV antigens: Gag, Rev-Tat-Nef, and Env (mainly derived from mac239) elicited SIV-specific T-cell response frequencies not vastly different than those of SIVΔnef (the T cells elicited by the CMV vector were largely of the effector memory type thought to be particularly effective for immune containment of HIV/SIV) [31,32], yet only provided moderate protection against low-dose homologous clonal SIVmac239 vaginal challenge, and once the vaginal barrier was breached, vaccinated animals fared no better than unvaccinated controls in terms of viral loads. In contrast, even a highly attenuated SIVΔ4 vector (acute-phase viral load peak ∼7 logs lower than wild-type SIV and ∼5 logs lower than SIVΔnef [39,42]) provided moderate protection against high-dose vaginal challenge of a slightly heterologous viral swarm, SIVmac251 (which is generally harder to protect against than a homologous clonal challenge virus [37,38]), and provided a degree of protection, namely lower viral load set points, when the vaginal barrier was breached in vaccinated animals [39,42]. Moreover, in contrast to the CMV vector, even a nonpersistent vector (Gag-DNA prime; Gag-Sendai virus vector boost) expressing SIV-Gag alone, lowered viral set points in many animals [46]. Therefore, although persistence almost certainly plays a role in the superior protective powers of attenuated SIV, protection does not strictly correlate with persistence alone or as one researcher put it, 'one virus that replicates is not the same as another' [47]. These results suggest that something more central to the lentivirus itself might distinguish attenuated SIV from other vaccines. And one of the most distinguishing features of HIV is the rate at which it evolves in vivo[48]. Evolution of attenuated simian immunodeficiency virus and immune responses The role that HIV diversity and mutation play in limiting the effectiveness of the immune response is well established [14]. Indeed even when attenuated, it is HIV/SIV's propensity to evolve that prevents its use in human vaccine trials [5]. However, as I have argued, that ability may be precisely what is required to broaden and educate the immune response so that it can contain viral immune escape variants. Under this hypothesis, the previously mentioned time frame needed to allow 'maturation' of the immune response [34,36,38] becomes readily explainable as the time needed for the attenuated virus to begin to generate escape variants and for the immune response to begin to adjust to better contain those variants. This hypothesis also accounts for the fact that increasingly attenuated strains of SIV are less efficacious vaccines [42] because as the virus' ability to replicate declines so too will the rate at which it generates escape variants. Additionally, Berry and coworkers [37,38] have found that following attenuated SIVmacC8 infection, protection arose much more rapidly against a clonal, isogenic challenge virus, SIVmacJ5 (∼3 weeks), than it did against challenge with a closely related but slightly heterologous viral swarm, SIVmac251 (∼20 weeks).3 These results are consistent with the proposed hypothesis, as it would be expected that longer durations prior to challenge allow continual immune broadening and protection against a wider range of viral challenges. This temporal increase in vaccine efficacy could manifest either due to quantitative immune broadening (i.e., increasing the number of epitopes targeted over time) and/or qualitative immune broadening (i.e., allowing a broader range of already expanded immune cells to 'mature' over time and thereby acquire more effective antiviral activity). Nevertheless, an objection to the theory may be raised on the grounds of whether significant immune broadening can occur within 3 weeks of infection to account for immune protection against the clonal homolog, SIVmacJ5. However, although no evidence exists on whether the immune response itself broadens this rapidly, immune escape variants can arise and become the dominant population within just a few weeks of infection [49–62]. Moreover, even if such variants are not yet dominant, the relatively high viral loads of attenuated SIV acute-phase infection and a rising immune response [11–13,63] should select for escape variants, 'seeding' them deep into the surviving attenuated SIV population and creating 'pockets' of sites in lymph nodes where subdominant CTLs preferentially expand. In contrast, regardless to the extent to which immune broadening has occurred, the wild-type challenge virus infects a host with a highly developed immune response that will significantly limit its replication. Extremely low wild-type viral loads in turn greatly limit the virus' subsequent ability to generate escape variants. Therefore, by the time of challenge, the attenuated virus although at a fitness disadvantage due to the deletions it carries has a few potential advantages over the wild-type virus as its 'head start' may allow it time to achieve higher viral loads, begin to adapt to immune and other in-vivo pressures (for example to expand the tropism range of the virus), and correct inherent defects in the virus.4 For all the above reasons, it is likely that at the time of challenge, despite its Nef defect, the evolving attenuated SIV 'swarm' may have achieved a greater fitness level (for replication in that particular host) than the incoming wild-type clone. Hence, attenuated SIV will continue to replicate and evolve at a higher rate than the challenge virus and remain the chief engine driving further immune evolution. Further immune evolution in turn will further reduce wild-type SIV's already limited capacity to replicate and evolve. Therefore, adaptations that the attenuated virus achieves in the first few weeks of infection, the wild-type virus may never achieve. This then offers a simple explanation to the apparent paradox of why the immune system is so effective at controlling the wild-type challenge virus, but often not the original infecting attenuated virus: by the time of challenge, the attenuated virus may be more fit to survive in that particular host than wild-type SIV. Thus, the attenuated virus survives and continues to coevolve along with the immune response. This might also explain why recombination frequently occurs between attenuated SIV and heterologous challenge virus [20,40,67,68], as there would be strong selective pressure for the two viruses to swap highly immunogenic regions for less immunogenic ones. There is evidence that a similar immune escape mechanism drives recombination in HIV superinfected individuals [69].5 Recombination and immune-driven escape may also lie behind the phenomenon of 'vaccine virus restimulation', whereby challenge with a heterologous (but not homologous) virus leads to outgrowth of the original 'vaccinating' virus [70] (in general, the identity of the reemerging virus has been determined largely based on the presence or absence of nef deletions and therefore recombination elsewhere in the virus may have been missed). An alternative but not necessarily mutually exclusive theory to explain why protection against a homologous wild-type clone arises much faster than that against a closely related viral swarm [37,38] postulates that protection arises not from immunity but from viral interference-target cell depletion [37,70,71]. However, although viral interference-target cell depletion probably plays some role in limiting challenge virus replication in at least some animals, it seems unlikely to be the sole explanation of protection due to the following: (1) Unlike wild-type SIV infection, SIVmac239Δnef infection does not lead to overt CCR5+CD4+ target T-cell depletion [24,72–75], and even less target cell depletion is expected following infection with the more attenuated SIVmac239Δ3 strain (although the ability of attenuated SIV to deplete hereto unrecognized critical subpopulations of target cells or block challenge virus replication by other means, such as Env shedding that may block SIV receptors on uninfected cells, cannot yet be excluded). (2) Even during wild-type HIV infection, there is evidence that superinfection [76] and viral recombination [77–79] occur throughout infection at surprisingly high rates. Therefore, it is unclear how effective interference and target cell depletion are at blocking competing viruses. (3) The viral tropism (based on coreceptor usage) of the SIVmac251 breakthrough virus in unprotected SIVmac239Δnef vaccinated animals is similar to that of SIVmac239 [80]. Therefore, it is unclear whether tropism accounts for the earlier protection against SIVmac239 than against SIVmac251 in SIVmac239Δnef vaccinated animals (although conceivably differences in receptor affinity or in other SIV proteins might still create tropism differences between the viruses). (4) Attenuated simian-human immunodeficiency viruses (SHIVs) of a particular tropism (e.g., CXCR4 tropic) provide protection against challenge of SIV of a different tropism (e.g., CCR5 tropic) [81–84]. (5) The attenuated virus generally continues to replicate in vivo. So, either target cell depletion is not complete, or cell-to-cell transmission of the attenuated virus gives it an advantage over 'cell-free' clonal homologous challenge virus (but not over the slightly heterologous SIVmac251 viral swarms?), or the attenuated SIV has 'evolved' in vivo to expand its cell tropism. (6) The occurrence of nonsterilizing immunity following challenge [17,40,42,85] suggests target cell depletion/viral interference probably is not the sole mechanism of protection following attenuated SIV infection. (7) Macaques with protective MHC alleles vaccinated with live attenuated SIV resist heterologous challenge better than vaccinated macaques lacking protective MHC alleles [40]. (8) Protection against heterologous challenge (SIVsmE660) does not appear to be vastly different in attenuated relative to wild-type-infected animals [85,86], despite the viruses' vastly different ability to deplete target cells and viral loads [24,72–75]. (9) Depletion of CD8+ lymphocytes at the time of wild-type SIV challenge generally leads to higher postchallenge viral loads in SIVΔ3 vaccinated animals compared with vaccinated controls [17]. In contrast, the best evidence that target cell depletion or viral interference does play a role in protection comes from Stebbings et al. [70], who found that animals infected with attenuated SIV and CD8-depleted throughout acute infection were still protected from isogenic wild-type challenge at day 20. This suggests that CD8+ lymphocytes are not critical for acute-phase protection. However, it should be noted that in their system, upon CD8+ lymphocyte depletion, during acute infection the attenuated SIV replicated to nearly wild-type levels [70]. As wild-type SIV infection typically depletes CCR5+CD4+ T cells [24,72–75], target cell depletion likely occurs in this system as well. Hence, to varying degrees, target cell depletion/viral interference, innate immunity, and an expanded humoral response may all account for the protection in this system prior to the return of CD8+ T cells [70] and may contribute to protection under other conditions where a significant degree of attenuated virus replication occurs. However, in other systems (e.g., those employing more attenuated viruses, with intact CD8+ lymphocytes, and delaying challenge until long after vaccination), immune protection likely plays a more significant role.6 Evidence for the theory A critical prediction of the hypothesis that mutation is essential for the efficacy of attenuated SIVs is a demonstration that despite attenuation, attenuated SIVs still undergo significant immune-driven evolution in vivo. Unfortunately, although there is a wealth of evidence that attenuated SIVs evolve in vivo[5–10], little links that evolution to immune pressure.7 However, based on the limited sequence data currently available from the Sydney Blood Bank Cohort [88,89], preliminary analysis suggests evidence of HLA 'footprints [90]' and frequent CTL escape mutations in Gag. Indeed, in one individual, the majority of the 42 amino acid differences between donor and recipient Gag likely derive from CTL pressure (C. Brander, manuscript in preparation). Although these data are consistent with the hypothesis proposed here, as blood samples derive many years after initial infection, questions of whether such mutations kinetically can account for relatively rapid immune protection await further studies. In that regard, preliminary sequencing results suggest that during initial SIVmac239Δnef-infection, the virus undergoes immune escape at multiple epitopes and with kinetics approximating that of the wild-type virus (D. O'Connor, manuscript in preparation). Nevertheless, in the absence of more complete data, I will make the case why selection of escape variants is likely to occur during attenuated SIV infection. First, wild-type HIV/SIV undergoes rapid and extensive immune evasion during primary-phase infection. Indeed, the majority of early HIV/SIV evolution is driven in response CTL pressure [60,91,92]. The fastest occurring of these mutations is the SIV Tat SL8 epitope in which viral variants can become the dominant population within 2–3 weeks of infection, but other escape variants arise shortly thereafter [50–52,54,56–62]. Significantly, fast escaping epitopes are disproportionately restricted by protective HLA alleles, suggesting that these escape variants incur a cost in viral fitness [40,60,61,93–96]. Yet despite the abundance of evidence that immune escape occurs frequently during HIV/SIV infection, the lower viral loads observed during attenuated SIV infection may mitigate production of escape variants at a high rate (the acute-phase peak for SIVΔnef is typically 50–100-fold lower than that of wild-type SIV) [34,39,40]. Still, a considerable amount of viral replication occurs during acute attenuated SIV infection and selection of immune escape variants is commonly observed in other systems with low viral loads, for example, in elite controllers [14,97–99]. Additionally, because HIV/SIV Nef downregulates MHC class I surface expression, absence of Nef in commonly used attenuated SIV strains, SIVΔnef and SIVΔ3, may increase cell surface expression of viral peptide–MHC complexes and heighten immune pressure for escape variants [8,45]. Furthermore, in the Sydney Blood Bank cohort, the infecting HIVΔnef strain has sufficient replicative capacity to continuously evolve over two decades of infection [9]. The hypothesis that immune escape variant generation is critical to the success of attenuated SIV vaccines further assumes that when nearly identical strains of SIV infect an individual, preferred escape routes exist and, therefore, immune responses and viral evolution follow predictable courses. This assumption seems justified based on murine models of viral/bacterial infection (done in inbred mice) [100–103] and studies of HIV-infected human homozygous twins [104,105] or MHC-matched macaques infected with SIV [106], where disease, viral evolution, and immune responses do follow fairly predictable trajectories [107]. Furthermore, in the studies cited, viral and immune responses evolve in separate individuals and, therefore, would be subjected to both stochastic and other forces (e.g., exposure to heterologous viruses [108]) that normally tend to diverge immune responses across genetically identical individuals. In contrast, in attenuated SIV-infected macaques, the challenge virus infects the same individual and faces those same adaptive responses expanded by the attenuated virus. Thus, the immune pressures exerted on both attenuated and challenge virus in that animal are expected to be even more similar than in identical twins. Finally, we may speculate as to how prior immune exposure to escape variants elicits a more effective immune response. First, the escaped epitope, if it still binds MHC, may selectively expand CTLs that recognize the escape variant [97,98,109–113]. Second, even if the escape variant is not recognized by CTL, immunodominant responses, probably through competition for antigen-presenting cells, can squelch subdominant responses and thereby limit immune diversity [114–117]. Therefore, the loss of immunodominant responses may allow subdominant responses to expand or mature in ways they ordinarily might not in absence of immune evasion and as immunodominance does not necessarily correlate with protective efficacy [118] this may be advantageous. The importance of a broad CTL response in controlling HIV/SIV infection is further supported by studies of subdominant CTL responses during HIV/SIV infection [119,120] and the fact that HLA class I homozygosity is associated with rapid progression to AIDS [121]. It is also important to remember that once dominant responses are escaped, they are not lost from the immune repertoire, but rather will slowly decline in numbers as do other memory responses upon antigen 'clearance'. However, these responses remain available and should rapidly expand during episodic outbreaks of wild-type revertants or wild-type challenge [122]. Finally, loss of CTL responses could lead to compensatory strengthening of humoral responses [17,70]. The requirement for coevolution of viral and immune responses for effective immunity may also lie behind another model of effective vaccination: wild-type SIV infection followed shortly after (∼1 day later) by transient drug treatment [123,124]. Interestingly, although such treatments stimulate protection against both homologous and heterologous challenge virus, so far the treatment has only worked when the 'vaccinating' virus was a viral 'swarm' (E660) but not when it was clonal (SIVmac239) [123–125]. It has been suggested that the success of the quasispecies may depend on the existence of partial drug-resistant variants within the viral swarm that allows it to replicate (at least to some degree) during drug treatment [125]. Moreover, CD8+ lymphocytes are required to fully suppress viral replication during transient drug treatment [126]. Thus, even during active drug treatment, viral replication likely occurs and may be critical for the development of effective protection. It is also important to note that because E660 starts out as a quasispecies, it may more quickly generate a very broad immune response than when protection is attempted employing a 'vaccinating' clone. Thus, the transient drug treatment model might also be consistent with the hypothesis proposed here; that is, vaccine viral diversity begets the strongest immune protection. In summary, an HIV vaccine that uses viral diversity/evolution to 'educate' the immune response has many advantages over vaccines that repetitively stimulate with inviolable antigen. Moreover, even if HIV antigen is placed within a replicating nonlentiviral viral vector, the vector generally does not contain an error-prone polymerase and the antigen is not critical for viral replication. Therefore, there will be no selective pressure to enforce coevolution of HIV antigen and immune responses and a less broad and effective HIV-specific immune response may result (indeed, if such vectors elicit high-affinity T cells against fewer epitopes, they may hasten the pace of immune escape). Finally, immune evolution during attenuated SIV infection occurs in absence of extensive target cell depletion [24,72]. This may allow the immune system to replace escaped CTL responses with other highly effective responses while maintaining functional memory T cells targeting the escaped epitopes [127]. In contrast, significant damage to memory helper T cells occurs early during wild-type HIV/SIV infection, which may permanently impair the immune system's ability to later keep pace with viral evolution [24,72–75]. Therefore, in contrast to wild-type infected animals, or animals vaccinated by other means, attenuated SIV may elicit a much broader and balanced immune response, thus minimizing the selective advantage of each individual immune escape variant. This in many ways recalls the experience of multidrug versus single-drug therapy, in that a multiprong attack is much harder to escape. Indeed, in at least one recent study, the ability of vaccinated and unvaccinated monkeys to achieve low viral set points correlated with neither the initial size of the immune response against a particularly protective epitope nor the avidity of T cells for that epitope, but did so with the diversity of the T-cell receptor (TCR) recognizing that epitope and the ability of those T cells to recognize escape variants [111,128]. How low can you go? In the prior paragraphs, I have argued that immunodominant responses can hinder the development of subdominant responses and that both dominant and subdominant responses are required for effective immune control of HIV. If these assertions hold, then as 'the vast majority of (epitope) competition and affinity maturation is observed after multiple immunizations' [115], an obvious implication is that vaccines that rely on multiple boosts using immutable SIV/HIV antigens may inadvertently lead to a very narrowly focused CTL response.8 This suggests that lowering the strength of dominant responses and increasing the strength of subdominant responses might boost vaccine efficacy. But would lowering dominant responses significantly impair their ability to control HIV? Although the answer is not known in that regard, the experiments of Barouch et al.[122] are most interesting. They infected Mamu-A*01-positve and Mamu-A*01-negative macaques with SIV carrying a CTL escape mutation in an immunodominant Mamu-A*01 response (Gag CM9). Interestingly, although stable revertants to the wild-type epitope occurred in Mamu-A*01-negative animals, reversions to wild-type in Mamu-A*01-positive macaques appeared only transiently. Concurrent with transient reversion to the wild-type sequence, a small population of Gag CM9-specific CTLs transiently arose (<0.5%). In contrast, during normal wild-type SIV acute infection, responses to the Gag CM9 peptide can exceed 10% and generally are 1–5% during chronic infection [49,122,132]. Yet, despite the relatively small number of Gag-CM9-specific CTLs transiently appearing in the Barouch et al. study, the CTLs exerted significant pressure and forced the reemergence of the escape variant as the dominant population [122]. The success of the Gag-CM9-specific CTLs in the study by Barouch et al. [122] raises two important points: the normally dominant Gag-CM9-specific CTL response is able to arise de novo during an already ongoing immune response and once it arises, relatively low numbers of Gag-CM9-specific CTLs exert significant immune pressure and force reversion to a less-fit variant. In other words, immunodominant responses are dominant for a reason – they arise easily. Moreover, by limiting the strength of subdominant responses, they may inadvertently limit immune protection. This questions whether dominant epitopes in HIV vaccines need to be aggressively boosted following initial priming. The fact that most macaques develop only weak-to-modest immune responses during attenuated SIV/SHIV infection [17,33,35,40,133] yet still develop CD8-dependent protective immunity supports the contention that the size of the vaccine CTL response may not matter beyond a certain point. In fact, the large immune responses typically associated with HIV infection may not be a true measure of immune success but instead reflect antigen stimulation and failure to contain viral replication [134–137].9 With high hope for the future, no prediction in regard to it is ventured Individuals who spontaneously control HIV/SIV infection are extremely rare. In contrast, control of wild-type SIV infection following attenuated SIV vaccination occurs more broadly in the general population, offers some degree of protection against heterologous SIV strains [40,85], and is a source of hope that creating an effective HIV vaccine is feasible. However, of the many HIV/SIV vaccines so far tested, none has approached the protection afforded by attenuated SIVs. As outlined in the introduction, a confluence of factors may explain this [2,23,28–33,70]. In addition to those factors, I have suggested an often-overlooked aspect of attenuated SIV biology is that the virus is a dynamic and not static immunogen and that immune-escape variants and their immune-broadening effects may be critical for immune protection by attenuated SIV 'vaccines'. This hypothesis, as noted earlier, explains away many paradoxes long associated with attenuated SIV and is based on a central feature of HIV/SIV replication: its high mutation rate. In that regard, the hypothesis is not novel, for the ability of HIV viral replication and mutation to drive immune expansion – and the requirement for a broad immune response for an effective HIV vaccine – has long been noted [119,120,134–142]. This is a strength of the hypothesis, for it explains away the superior protective powers of attenuated SIV parsimoniously, using mainstream HIV vaccine thoughts and requiring no new speculation on SIV immunology/biology. What remains novel is the suggestion that immune escape and its immune-broadening effects also lie behind the protective power of live attenuated SIV vaccines. A key prediction of this hypothesis will be to document both the existence of immune-escape variants and other fitness-enhancing mutations during attenuated SIV infection and that the breadth of effective immune responses elicited is indeed broader than other vaccine approaches. The next step then would be to create vaccines that mimic the immune responses induced by attenuated SIV without the use of live retroviruses. In that regard, the MHC-matched SIV/cynomologus macaque model has many advantages [143]. For example, MHC-matched cynomologus macaques could be infected with either wild-type or attenuated SIV and a year after infection the evolved virus harvested, sequenced, and incorporated into a nonretrovirus vaccine vector (replicating or not) or single-cycle SIV [41]. To mimic in-vivo evolution of attenuated SIV, MHC-matched animals would then be vaccinated with wild-type antigen with later boosts employing 'evolved' variant antigen. However, as discussed earlier, non-SIV vectors, even replicating ones, may not faithfully reproduce all attributes of live attenuated SIV. In that case, it may be worthwhile to reexamine highly attenuated SIV vectors, such as SIVΔ4, which provided some protection against vaginal SIV challenge but failed to protect against intravenous challenge [42]. For example, by infecting macaques with a 'swarm' of attenuated SIVΔ4 clones, the immune response elicited might be broadened to increase the vector's protective capacity (whether SIVΔ4 would ever be deemed safe for human trials is another story). Alternatively, following SIVΔ4 'priming' subsequent boosts might employ heterologous SIV antigens expressed in nonlive SIV vectors or in less attenuated SIV strains (e.g. SIVΔ3). Although it may be argued that epitope interference or altered peptide ligands [144] will limit the utility of this approach, I would argue that if live attenuated SIV protects by the mechanisms outlined here then this suggests interference is not an insurmountable obstacle. However, this is also why whole virus or as many whole viral proteins as possible should be included in the vaccine as opposed to using just peptide epitopes, as interference may be more of an issue at high peptide concentrations and it is best to initially mimic conditions of the attenuated virus to avoid confounding influences and assumptions not yet understood (e.g., it is unclear which CTL or even humoral responses do the 'heavy lifting' in terms of limiting viral replication). Indeed, given the diversity of MHC alleles present in the human population, it is expected that the number of epitopes and even relative contribution of cellular and humoral responses required for effective immunity may vary from individual to individual [17]. Finally, a related hypothesis, that antibody immune-driven evolution also occurs in the context of attenuated SIV infection and accounts for some of the attenuated SIVs' vaccine efficacy also needs to be investigated. That immune escape occurs rapidly from neutralizing antibody responses is now well documented [139–141,145]. In addition, there is increasing evidence for a role of nonneutralizing antibodies or perhaps natural antibodies in immune protection [2,17–22]. With the advent of new techniques for rapidly isolating antigen specific B cells, the time is ripe for characterization of: those epitopes undergoing antibody-mediated immune escape during attenuated SIV infection, the antibodies responsible for driving that escape, and the mechanisms by which those antibodies apply immune pressure (neutralization, ADCC, etc.). Lastly, it may be argued that if the proposed mechanism is correct, it may not be applicable to real-world vaccines where HIV and MHC diversity abound, quasispecies evolution remains poorly understood, and the nature of the challenge virus is not identical to that of the vaccine virus. However, the question of real-world relevance confounds other proposed mechanisms of attenuated SIV protection as well, for example, target cell depletion. Therefore, the goal of such studies should be to first understand what accounts for attenuated SIV's superior protective efficacy and worry about their 'real-word' implications later. Indeed, it could be that attenuated SIV are a highly artificial system with little relevance for HIV vaccine development; yet given the amount of effort currently expended on the model [2], that in itself would be a significant finding. Still I remain optimistic that the mechanism proposed here, if correct, would have real-world implications. For if we learn how to better design vaccines against even one SIV strain, that should translate into knowledge of how to protect against a wider range of strains, especially given that attenuated SIV do offer at least some protection against heterologous strains [40,85]. In conclusion, a multitude of factors may explain the protective efficacy of live attenuated SIV. However, a decade of research has established immune evasion/viral diversity to be a fundamental property of HIV/SIV; therefore, it is reasonable to consider that viral evolution and its immune-broadening effects are central as well to the protective efficacy of live attenuated vaccines. Mutation, evolution, and viral diversity may, therefore, lie behind both the power and danger of live attenuated HIV/SIV vaccines. Acknowledgement I would like to thank those scientists, friends, and family who inspired me during my scientific career.
Referência(s)