Despite dramatic clinical effects resulting from the advent of highly active antiviral therapy (HAART) for the treatment of HIV-1 infection, it is increasingly clear that viral eradication will be exceedingly difficult to achieve. Some persons, however, have now been infected with HIV for more than 30 years without developing progressive disease, and maintain low to undetectable viral loads in blood, even though they have never been treated with antiviral drugs. These persons seem to have achieved with HIV-1 what is achieved with other chronic viral infections such as Epstein-Barr virus (EBV) and cytomegalovirus (CMV)—viral containment rather than eradication. Our goal is to perform patient-oriented research to understand the immune control of HIV and to use this information to develop interventions to induce immunologic control in persons with otherwise progressive infections.
Correlates of Immune Protection in HIV Infection: Learning from Patients
As an infectious disease specialist and practicing physician, I focus my research on human diseases, and in particular on HIV infection, which was a new and undefined illness when I started my medical career. The major focus of our work has been on defining how the immune system fights HIV and why it usually loses the battle. Our early work identified HIV-1–specific cytotoxic CD4 T lymphocyte (CTL) responses in infected persons (Nature 1987; Science 1988), detected strong virus-specific T helper cell responses in persons whose viremia is controlled without treatment (Science 1997), and demonstrated that early treatment of acute infection leads to increased immunity and the transient ability to control viremia in the majority of infected persons (Nature 2000).
In our ongoing attempts to understand immune control of HIV, we concentrate on two groups of infected persons in whom control of HIV infection is apparent. One group is persons with chronic HIV infection who maintain extremely low levels of plasma viremia without the need for treatment, and thus do not progress to develop HIV-associated disease. We established an international collaborative study that has now recruited a cohort of more than 1,500 infected such persons who control viremia either to undetectable levels (elite controllers, <50 RNA copies/ml plasma) or to levels at which disease progression and transmission are markedly reduced (viremic controllers, 50–2,000 RNA copies/ml), despite never having been treated with antiviral drug therapy. We have used this cohort to define viral, host genetic, and immunologic factors associated with this control.
A second group in which at least transient viral control is apparent is persons in the acute stages of HIV infection. Over the past 10 years, we have identified more than 300 persons with acute HIV infection, usually presenting with a flu-like syndrome and often initially not appropriately diagnosed by health-care providers because the diagnostic HIV antibody test remains negative for a few weeks after infection. We focus these studies on patients we recruit in Boston and patients we have identified at our additional research site at the heart of the HIV and tuberculosis epidemics in KwaZulu-Natal Province, South Africa. There the numbers of new infections remain staggering, with infection rates among pregnant women at our clinical site rising from less than 1 percent at age 15 to more than 50 percent by age 23. Because the initial peak viremia and subsequent containment of viremia vary markedly among acutely infected persons, these cohorts provide an optimal cohort in which to define effective and ineffective immune responses.
In these cohorts, our main hypothesis is that durable immune control of HIV can be achieved and that understanding this control will contribute to the development of vaccines and immunotherapeutics. More specifically, we hypothesize that control is due to a combination of host genetic, immunologic, and viral factors, and that these can be defined by analysis of spontaneous control of viremia.
Host Genomics and HIV Control
Despite years of research, the actual correlates of protection in HIV infection remain unclear. We reasoned that new developments in the field of genomics might provide insights into genetic factors that modulate HIV control. To address this, we established a cohort of HIV controllers and progressors and initiated a genome-wide association scan (GWAS) to determine single-nucleotide polymorphisms (SNPs) that are associated with control or lack of control. We enrolled 1,500 HIV controllers, defined by the ability to maintain viral loads of fewer than 2,000 RNA copies/ml in the absence of therapy. These people were recruited through an international collaborative network involving not just scientists and research clinicians but also physicians and nurses in private practice who care for such patients. Studies were completed in nearly 1,000 of the controllers as well as nearly 3,000 persons with progressive infection and high viral loads, in whom we measured 1.3 million SNPs and performed stepwise logistic regression to define significant SNPs associated with control.
Only four SNPs reached statistical significance, and all were located in the HLA (human leukocyte antigen) region of the major histocompatibility complex on chromosome 6. Subsequent sequence analysis of this region has revealed that the SNPs are tagging specific amino acids in the HLA-B binding groove that impact viral peptide binding for recognition by HIV-specific CD8+ T cells, as well as another amino acid that impacts HLA-C expression (Science 2010). These data reconcile previous work done by us and others showing that HLA-B alleles are associated with differences in disease outcome, and together the data show that host genetics explains 23 percent of the variability in viral load. We are building on these studies to define the impact of the HLA-B polymorphisms on peptide binding and CD8+ T cell functions (Gaiha et al, Immunity 2015), as well as the impact of T cell receptor clonotypes on immune control (Chen et al, Nature Immunology 2012). Current studies are examining the impact of combinations of protective genetic parameters on immune control, and the mechanism by which immune control is achieved. For example, through collaborative studies, we have shown that the mechanism of HLA-C expression modulation of viral load is due to differential regulation of HLA-C expression by microRNA (Kulkarni et al, Nature 2011).
From Genetic to Functional Immunologic Studies of HIV Control
The above genetic studies strongly implicate HIV-specific, HLA class I–restricted CD8+ T cells in modulating viremia. However, even in persons with favorable amino acids in the key positions in the HLA-B peptide binding groove (particularly positions 67, 70, and 97), the majority will not end up being HIV controllers, but rather will experience progressive infection. This cohort of persons with both controlled and progressive infection has provided us with an opportunity to determine why this is, through functional studies of immune function in persons in whom we have sufficient subjects to be able to stratify by host genetic parameters that by themselves influence outcome.
One area of investigation is to define the role of immunoregulatory mechanisms in modulating the effective immune response to HIV. In early studies we reported that the inhibitory immunoregulatory molecule PD-1 is up-regulated in chronic HIV infection (Day et al., Nature 2006), and we are currently investigating the mechanism of this effect, including studies of the effect of PD-1 and other molecules on the immunologic synapse, and the mechanisms allowing experimental blockade of this pathway to restore CD4+ and CD8+ T cell function in vitro. Additional studies focus on CTLA-4, a second inhibitory regulator in the B7-CD28 family, which we found to be selectively up-regulated in virus-specific CD4+ T cells in all categories of infected subjects except for elite controllers. Blockade of this pathway results in reversal of the immune dysfunction (Kaufmann et al., Nature Immunology 2007).
In another area of investigation, we are pursuing the hypothesis that observed differences in control are influenced by differences in the functional ability of CD8+ T cell responses to exert immune selection pressure. Part of this effort focuses on the specificity of the immune response, where we are investigating the antiviral function of CD8+ T cells directed against different viral proteins. These results show that specificity does matter, with responses to the more conserved Gag protein associated with enhanced immune control, and responses to HIV envelope associated with diminished immune control—findings we are expanding, with the goal of immunotherapeutic interventions (Kiepiela et al., Nature 2007).
These studies of immune pressure have also revealed marked differences in sequence evolution in the presence CD8+ T cell selection pressure, and that many epitopes present in the infecting virus do not elicit CD8+ T cell responses, a clear sign that many HIV-specific CD8+ T cell responses measured by interferon-γ production are ineffective. We are assessing yet another dimension of CD8+ T cell function, namely the ability to inhibit HIV replication when autologous CD4+ T cells, the major in vivo target of the virus, are infected in vitro. Here again, we find marked differences in the ability to control HIV in vitro, and we are working with engineers at MIT to develop techniques to assess cell killing at the single-cell level (Varadarajan et al., Journal of Clinical Investigation 2011) and to develop a three-dimensional tissue matrix model to directly observe T cell killing by videomicroscopy (Foley et al, PloS One 2014).
An additional area of investigation into T cell function involves more detailed analysis of immune specificity and function in the earliest stages of acute HIV infection. Because the infection rate is so high at our site in South Africa, we are creating a longitudinal cohort of high-risk uninfected women. We screen these women for acute infection twice weekly, thereby allowing us to obtain blood and tissue specimens at the earliest stages of acute infection, before symptoms occur and before peak viremia is reached. This study, entitled FRESH (females rising through education, support, and health), is a novel approach, partially philanthropically funded, designed as both a pathway out of poverty, one of the main drivers of HIV infection risk, as well as a study of the earliest immunologic events in what we term “hyperacute” infection—the period between the first evidence of plasma viremia until peak viremia. We provide classes to help the FRESH participants develop job skills and to educate them to protect themselves from becoming infected. These studies have revealed that the very earliest stage of HIV infection are associated with massive expansion of HIV-specific CD8 T cells that rapidly undergo apoptosis and fail to upregulate molecules associated with long term memory development (Ndhlovu et al., Immunity 2015). These studies have also provided insights regarding HIV transmission, showing that the use of injectable progesterone-containing contraceptives is associated with a significantly increased risk of HIV acquisition (Byrne et al, Lancet Infectious Diseases 2016) and that the composition of the vaginal microbiome influences transmission (Anahtar et al., Immunity 2015).
Viral Evolution and Immune Control
HIV is an extremely variable virus, raising the possibility that differences in the virus itself are impacting immune control. We and others have shown that immune escape from CD8+ T cell responses occurs in vivo (Goulder et al., Nature 2001). Making use of an extensive global collaborative network we developed through our work in Africa, we found that virus evolution is not only predictable but that certain immune escape mutations are becoming fixed in the larger population because of immunodominant CD8+ T cell selection pressure (Kawashima et al., Nature 2009). In other words, HIV is progressively adapting to HLA selection pressure. However, transmission studies we and others conducted show that many mutations selected in a virus donor are unstable, as they revert in the recipient in the absence of ongoing immune selection pressure.
Based on these observations of reverting mutations, we are exploring the hypothesis that immune control is at least in part due to selection for fitness-impairing mutations. To address this, we developed a high-throughput fluorescence-based assay to assess the impact of immune selection pressure in the Gag-protease on viral fitness. We apply this assay to cohorts of HIV controllers and progressors, and have demonstrated that viral fitness is linked to host HLA, with protective alleles associated with the most dramatic reduction in viral fitness (Miura et al., Journal of Virology 2010).
Through computational studies we are conducting in collaboration with computational biologists and physicists at MIT, we have identified a direct link between HLA alleles and T cell function, predicting that protective alleles bind fewer self-peptides and would be more likely to survive thymic selection and also be more cross-reactive with mutants that might arise in vivo (Kosmrlj et al., Nature 2010). To define differences in outcome caused by epitope specificity, we applied random matrix theory, a statistical model used to assess stock market fluctuations, among other things. These results indicate that there are multidimensional constraints on HIV evolution and identified particularly vulnerable sectors that do not tolerate multiple concurrent mutations and that are preferentially targeted by both the CD4+ and CD8+ T cell responses in HIV controllers, but not in progressors (Dahirel et al., PNAS 2011; Fergusson et al., Immunity 2013). These studies form the basis for ongoing studies being conducted to define the overall constraints on HIV evolution and to define immunogens that target those regions of the virus that are most vulnerable to mutation due to loss of fitness.
Augmentation of Effective Immunity in HIV Infection
The existence of persons who are able to control viremia spontaneously, along with the demonstration that effective immune responses against HIV can be augmented by early treatment of acute infection, provides a rationale for attempts to increase immunity to HIV in infected persons. Our recent findings that specificity of responses matters, with responses to Gag associated with enhanced immune control, together with other studies on HIV-specific CD4 T cell function and immune regulation, are paving the way for new initiatives designed to augment effective immunity in persons already infected with HIV.
HIV Vaccine Trials
With the establishment of the Ragon Institute of MGH, MIT and Harvard, we have been able to create a collaborative environment involving scientists and engineers from diverse fields, all committed to crossing disciplines to contribute to the development of an effective HIV vaccine. Through bench research being done in Boston and in Africa, we are defining candidate immunogens, with the goal of inducing effective T and B cell immune responses to HIV. Through our collaborations in developing the HIV Pathogenesis Program at the University of KwaZulu-Natal, the HHMI-funded establishment of the KwaZulu-Natal Research Institute for Tuberculosis and HIV (K-RITH), and collaborations with the KwaZulu-Natal–based Center for the AIDS Program of Research in South Africa (CAPRISA), we expect to move candidate vaccines into clinical trials, not only allowing us to define rules governing vaccine-induced immune responses in humans but also to perform efficacy trials in the quest for an effective AIDS vaccine.
Parts of this work have been supported by the Doris Duke Charitable Foundation, the Bill and Melinda Gates Foundation, the Mark and Lisa Schwartz Foundation, and the National Institute of Allergy and Infectious Diseases.
As of February 15, 2016