Date Published: November 1, 2018
Publisher: Public Library of Science
Author(s): Philip Goulder, Steven G. Deeks, Carolyn B. Coyne.
In this brief review and perspective, we address the question of whether the immune responses that bring about immune control of acute HIV infection are the same as, or distinct from, those that maintain long-term viral suppression once control of viremia has been achieved. To this end, we describe the natural history of elite and post-treatment control, noting the lack of data regarding what happens acutely. We review the evidence suggesting that the two clinical phenotypes may differ in terms of the mechanisms required to achieve and maintain control, as well as the level of inflammation that persists once a steady state is achieved. We then describe the evidence from longitudinal studies of controllers who fail and studies of biologic sex (male versus female), age (children versus adults), and simian immunodeficiency virus (SIV) (pathogenic/experimental versus nonpathogenic/natural infection). Collectively, these studies demonstrate that the battle between the inflammatory and anti-inflammatory pathways during acute infection has long-term consequences, both for the degree to which control is maintained and the health of the individual. Potent and stringent control of HIV may be required acutely, but once control is established, the chronic inflammatory response can be detrimental. Interventional approaches designed to bring about HIV cure and/or remission should be nuanced accordingly.
Identifying the mechanisms by which the host can naturally control HIV or simian immunodeficiency virus (SIV) has long been a priority for immunologists. These mechanisms might be leveraged to develop novel interventions to prevent HIV transmission, control HIV in the absence of therapy (a “remission”), or even fully eradicate the reservoir (a “cure”) . Towards this end, groups around the world have recruited and characterized those rare individuals who maintain near-complete control of the virus in the absence of antiretroviral therapy (ART). Two distinct clinical phenotypes exist: those who naturally control the virus without any treatment (“elite” controllers) and those who do so but only after receiving prolonged ART (“post-treatment controllers”).
The natural history of individuals who are destined to fully control their virus in the absence of therapy (elite control) or after interrupting therapy (post-treatment control) remains poorly defined. This is particularly true during the immediate post-infection or post-interruption period in which the virus likely replicates in the absence of a fully formed host response. Because most controllers are identified long after the acute viremic phase has resolved, the kinetics of HIV replication and the immediate host response are poorly understood.
A clue to the central mechanisms underlying elite control of HIV infection is the high frequency of “protective” HLA molecules such as HLA-B*57 and HLA-B*27 and low frequency of “disease-susceptible” HLA molecules such as HLA-B*35 [8, 39]. In contrast, in the VISCONTI post-treatment controller cohort, HLA-B*35 was highly prevalent, and HLA-B*27 and B*57 were notably infrequent . In the SPARTAC study, the study cohort was a mixture of subtype-B–infected male Caucasians and subtype-C–infected female Africans, and in this heterogeneous group, one or more disease-susceptible HLA alleles (defined as HLA-B*35:01 or HLA-B*07:02 for subtype-B and HLA-B*18:01 or HLA-B*58:02 for subtype-C) were also found in the post-treatment controllers, whereas there was no clear enrichment for the classic protective alleles . Classically defined protective alleles were rare in another recent cohort of post-treament controllers . Among post-treatment controllers, the ability of HIV-specific CD8+ T cells to inhibit viral replication is modest—and similar to that of typical non-controllers—but significantly lower than that of elite controllers. This prompts the question of whether those controlling HIV after treatment are doing so because of an immune response that does not rely on generation and persistence of the immunodominant HLA-B-restricted CD8+ T cell specificities observed in many elite controllers.
The vast majority of mechanistic studies regarding HIV control in people are cross sectional in nature. These studies have well-known limitations. Specifically, it is difficult, if not impossible, to determine cause and effect. The study of host genetics and outcome partially addresses this concern, as the genes were present before the infection and distinguishing cause and effect is straightforward.
Chronic inflammation during elite control has been well-established [58–61]. These inflammatory pathways likely contribute to immune dysfunction  and perhaps the development of cardiovascular disease [63, 64].
The natural history of untreated HIV infection differs by sex. In untreated infection, the viral load set points in females are 0.33 to 0.78 log10 copies RNA/mL lower than in males . Women are 5-fold more likely than men to be elite controllers . Females are also more likely than males to clear hepatitis C virus (HCV) in the absence of treatment . Mechanistically, it has been argued that plasmacytoid dendritic cells (pDCs) in females produce substantially more interferon-alpha in response to stimulation by toll-like receptor 7 (TLR7) ligands such as HIV-1 and other single-stranded RNA viruses  (perhaps, as recently argued, because TLR7 is on the X chromosome and is expressed at higher levels in females than males) . Although type I interferon has potent antiviral activities in the acute setting , too much interferon-alpha, or too much of the wrong subtypes of interferon-alpha, can lead to immune hyperactivation and its well-described detrimental consequences [76–81].
In pediatric HIV infection, the ability of the immune system to control HIV via an effective CD8+ T-cell response is largely neutered by the immunotolerant environment of early life [85, 86]. Consequently, viral loads are higher in children than adults; indeed, the median viral loads are approximately 1.5 log10 higher in young children than in adults at the equivalent time after infection [87, 88]. Elite control, as defined by having an undetectable viral load by standard assays, is rare in perinatally infected children but has been reported [89, 90]. Post-treatment control is also rare but has been reported [35, 91, 92].
SIV infection of monkeys provides additional support for our model. The natural hosts for SIV have low levels of immune activation and disease progression despite persistently high viral loads [98, 99], a phenotype that is similar to human viremic nonprogressors ; indeed, these two share a number of similar biologic profiles [95, 96]. Elite control in the natural hosts for SIV is rare (most animals have viral loads well above 10,000 copies RNA/mL) although not uncommon in experimental models such as rhesus macaques , in which the virus stimulates a profound inflammatory response and causes rapid disease progression.
HLA-B*57:01 is the single human genetic polymorphism most strongly associated with elite control of HIV infection . The mechanism for this protective effect has yet to be fully defined. Many have argued that HLA-B*57, HLA-B*27, and other protective alleles present multiple highly conserved epitopes for CD8+ T-cell recognition, escape from which results in a significant fitness cost to the virus [100–102]. Perhaps the clearest evidence that epitope specificity is important is the observation that distinct protective class I alleles in rhesus macaques (Mamu-B*17) and humans (HLA-B*27:02) target the identical immunodominant epitope (Nef IW9 IRYPKTFGW in SIV and IRYPLTFGW in HIV) [103, 104] despite independent evolution of human and rhesus macaque MHC class I. Similarly, the two protective alleles Mamu-B*08 and HLA-B*27:05 share a similar peptide binding groove and can present the identical epitopes . It has also been reported that enhanced HIV-specific CD8+ T-cell function—as defined by proliferative capacity, cytotoxic activities, and the production of multiple cytokines [5, 6]—has been a consistent predictor of virus control, arguing that the inherent functional capacity of the immune response in addition to the epitopes targeted is a key factor. This raises the possibility that HLA-B*57:01 and other protective alleles somehow stimulate the generation of highly functional CD8+ T cells directly and independent of the epitopes targeted .
As has been consistently demonstrated in studies of SIV and HIV, potent and sustained immune responses to chronic pathogens cause collateral damage. Those who are more likely to achieve elite control are in general more likely to experience disease progression, as shown by our comparisons of women versus men, adults versus children, and natural versus experimental models of SIV (Fig 2). The most robust predictors of virus control in untreated disease are also consistently associated with autoimmunity and other inflammatory disorders. Longitudinally, once elite control is achieved, those individuals who exhibit higher levels of inflammation are more likely to lose control over time.