Scanning the Scientific Horizon

An array of fundamental developments in understanding HIV infection and pathogenesis, and the interpretation of immune responses against the virus, headlined this year's CROI meeting

By Richard Jefferys*

Over the past decade, new technologies have emerged that allow researchers to analyze the specificity and function of antigen-specific T-cell responses in exquisite detail. One approach that has been popularized recently by researchers at the Vaccine Research Center (VRC), part of the US National Institute of Allergy and Infectious Diseases (NIAID), is using multi-parameter flow cytometry to analyze the ability of antigen-specific T cells to perform multiple functions simultaneously—the buzzword that has emerged in the field is “polyfunctionality,” although this should not be taken to mean that execution of the function automatically equates with in vivo efficacy. For CD8+ T cells, five potential functional markers are typically evaluated: production of the cytokines interleukin (IL)-2, TNF-α, and interferon (INF)-γ, and the chemokine MIP-1ß, along with expression of a cell surface molecule called CD107a, which is a marker for cytotoxic (cell-killing) potential. Data is then reported based on the number of these markers an individual T cell displays. The growing enthusiasm for this approach is largely driven by data demonstrating an association between polyfunctional HIV-specific T cells and control of viral replication and non-progression in infected individuals (Blood 107, 4781, 2006).

But application of new technologies can sometimes lead to challenges in data interpretation. At CROI, Hendrik Streeck of the Partners AIDS Research Center in Boston delivered a cautionary tale about interpreting data on T cell polyfunctionality in HIV infection. To date, Streeck noted, studies have evaluated polyfunctional CD8+ T-cell responses to multiple pooled peptides, as opposed to individual epitopes. In a cohort of individuals with acute HIV infection, Streeck decided to take a different tack. He analyzed the profiles of CD8+ T cells targeting known HIV epitopes and then looked at the relevant genetic sequence of the virus for any evidence of immune escape mutations. Streeck’s analysis revealed that CD8+ T cells with putatively broader functionality (displaying three or more of the above-listed functional markers) were more likely to be targeting epitopes in which escape mutations had developed. These CD8+ T cells were therefore not being stimulated because the epitope they were targeting was no longer present in the virus. In contrast, epitope-specific CD8+ T cells displaying only one of the above functional markers were targeting epitopes that had not escaped. Streeck emphasized that these data do not necessarily mean that all polyfunctional responses are unimportant, but rather supports careful interpretation of studies employing this approach for evaluating T-cell responses.

Surveying the genome

David Goldstein from Duke University and the Center for HIV/AIDS Vaccine Immunology (CHAVI) provided an update session on ongoing efforts to uncover host genetic influences on control of HIV replication and disease progression, following work published last year with collaborator Amalio Telenti of the University of Lausanne (Science 317, 944, 2007). These studies exploit a new method of genetic analyses that allows investigators to assess the impact of common variations in the human genome—called single nucleotide polymorphisms or SNPs—on the outcome of interest.

Goldstein initially focused on viral load set points, and in a study of 486 individuals, three associations were uncovered that remained significant, even after correcting for the staggering half a million different SNP analyses conducted. The first association, which accounted for close to 10% of the variation in viral load set point, was with an SNP located in the HLA complex P5 (HCP5) gene. This gene is known to be linked to the HLA B*5701 allele, which is a well-described correlate of viral control and non-progression in HIV-infected individuals, suggesting that the HCP5 SNP mediates its effect by affecting the HLA B*5701 allele in some way. The HLA B*5701 allele encodes a CD8+ T-cell receptor reported to be particularly adept at recognizing HIV epitopes. However, in the original Science paper, Goldstein and colleagues noted that HCP5 is an endogenous retroviral element and so could conceivably have direct effects on HIV replication, perhaps via an antisense mechanism. At CROI, Goldstein reported that extensive in vitro studies failed to demonstrate any inhibition of HIV, leading him to conclude that an independent effect of HCP5 is unlikely. From the perspective of vaccines this may be somewhat encouraging, as it suggests that the association between the HCP5 SNP and viral load set point is likely related to the CD8+ T-cell response to HIV.

The second association that emerged was with an SNP in the gene encoding HLA-C, which is also involved in the presentation of epitopes to CD8+ T cells. As background, thousands of HLA molecules adorn every cell in the human body, with the exception of red blood cells, and their function is to display fragments of proteins (epitopes) lifted from inside the cell to passing CD8+ T cells. Should any of these epitopes be recognized as foreign (for example if the cell is infected with HIV), a CD8+ T cell can release enzymes that kill the infected cell. The HLA molecules that interact with CD8+ T cells are called class I, and within this class there are three main sub-families: HLA-A, HLA-B, and HLA-C. Studies have shown that cells express far more HLA-A and HLA-B molecules than HLA-C, so presentation of HIV epitopes by HLA-C has generally been less studied. Goldstein’s hypothesis is that this second SNP may be associated with increased expression of HLA-C on cells, potentially facilitating enhanced CD8+ T-cell recognition—an attractive hypothesis given that HIV’s Nef protein is known to downregulate HLA-B molecules from the surface of infected cells as a means of immune escape. He has found some support for this idea in published databases but, so far, efforts to independently confirm the association between the SNP and HLA-C expression (by measuring the effect of the SNP on HLA-C messenger RNA levels in cells) have been stymied by unanticipated variation dependent on which HLA-C allele is used as a probe. Goldstein is now collaborating with Andrew McMichael to directly measure HLA-C protein expression on cells of individuals possessing the beneficial SNP.

The third association discussed by Goldstein is with a set of seven SNPs located in or near a pair of more obscure genes dubbed ZNRD1 and RNF39. ZNRD1 encodes a protein involved in RNA transcription. The plausibility of this association was recently buttressed by the finding that ZNRD1 is one of a dizzying array of host proteins needed by HIV to replicate in human cells (Science 10 Jan 2008, DOI: 10.1126/Science.1152725). The function of RNF39, however, requires further study. In the case of these SNPs, the stronger association was initially seen with disease progression (defined as time for a CD4+ T-cell count to decline to below 350 or initiation of antiretroviral treatment) rather than viral load set point. Importantly, however, Goldstein was able to unveil new data from an additional 1,000 HIV-infected individuals and all of the above-described genetic polymorphisms were associated significantly with viral load set point. Based on these additional analyses, he stressed that “these polymorphisms are most certainly real effects that are here to stay.” Further analyses will aim to reveal the functional relationships between the SNPs, relevant gene products, and control of HIV replication.

To give a broader sense of the profound interplay between host genetics and HIV disease progression, Goldstein evaluated the combined effect of these newly discovered polymorphisms in the HCP5, HLA-C, and ZNRD1/RNF39 genes along with two other favorable genetic polymorphisms in the CCR5 and CCR2 genes (CCR5Δ32 and CCR2 V64I). Possession of one or two favorable mutations in at least four of these genes was associated with a four-times-longer average period before CD4+ T cells declined to less than 350 (from less than 2 years to more than 8).

Goldstein closed by describing the future plans of his CHAVI team, which include a potentially important evaluation of genetic associations with the magnitude of antibody responses generated by participants in the efficacy trials of VaxGen’s gp120 vaccine, AIDSVAX. Although the vaccine failed to protect against HIV infection, an association was seen between the magnitude of an individual’s antibody response to the immunogen and susceptibility to HIV infection; participants who generated high levels of antibodies were less likely than placebo recipients to acquire HIV infection, while those who mounted poor antibody responses were more susceptible to HIV than the placebo group. Goldstein’s study could therefore shed additional light on the genetic influences on both the immune response to vaccines and susceptibility to HIV infection, and might even assist researchers in understanding the surprising association between the magnitude of antibody responses to adenovirus and susceptibility to HIV infection observed among placebo recipients in the STEP trial.

Debating gut depletion

Several years ago, a number of scientists popularized the view that loss of CD4+ T cells from the gut mucosal tissue in the first few weeks of infection was a catastrophic insult from which the immune system never recovers, and the most important augury of the ultimate development of immunodeficiency. But, as described by Cristian Apetrei in a poster discussion talk titled “HIV Pathogenesis: Viral Blitzkrieg or 10 Years’ War?,” emerging data is prompting many researchers to reconsider this theory.

Apetrei provided examples from several different pathogenic and nonpathogenic models of SIV infection illustrating that the impact on gut CD4+ T cells is variable and not predictive of disease progression. He also showed an example of a rhesus macaque infected with a pathogenic SIV isolate that progressed to simian AIDS despite experiencing very little acute depletion of gut CD4+ T cells. Apetrei argued that events during the chronic phase of infection—such as increasing immune activation and anti-SIV antibody responses—are more important drivers of disease progression than the early loss of gut CD4+ T cells.

Satya Dandekar of the University of California at Davis described results of a new study that approached the issue of gut CD4+ T-cell depletion from a novel angle by looking at a specific subset of gut CD4+ T cells rather than the population overall. Dandekar focused in on Th17 (T helper type 17) CD4+ T cells, a recently discovered lineage characterized by the production of interleukin-17 (IL-17) that is important in the control of microbial pathogens and yet also associated with some pro-inflammatory autoimmune conditions. In a collaboration with Andreas Baumler’s lab, Dandekar conducted studies in rhesus macaques with the microbial pathogen Salmonella typhimurium. When this bacteria was injected directly into the gut tissues of macaques, Dandekar observed an upregulation of IL-17. This response was blunted in SIV-infected macaques, while other cytokine responses, such as IFN- γ, remained unchanged. In addition, in SIV-infected macaques the Salmonella bacteria spread to other tissues, such as the mesenteric lymph node and the spleen. Normally, Salmonella infection remains limited to the gut. “The impact of not having an IL-17 response was that [Salmonella] disseminated to other organs,” Dandekar said. This same phenomenon was also observed in mice experiments. Dandekar said the loss of Th17 CD4+ T cells in the gut, which leads to incomplete clearance and dissemination of microbial pathogens, may be a “major contributor to the chronic immune activation seen in HIV and SIV infections.”

Th17 in the spotlight

Dandekar’s talk was a prelude to a triumvate of similar studies presented at CROI. Mirka Paiardini of Guido Silvestri’s laboratory at the University of Pennsylvania described the differential impact of pathogenic and non-pathogenic infections on the Th17 cell population in the gut. Paiardini noted that CD4+ T cells are now generally characterized as belonging to one of four broad categories: Th1 cells, which primarily produce IFN- γ and are important in cell-mediated immunity; Th2 cells, which produce IL-4 and other cytokines important in supporting antibody (humoral) responses; Treg cells which can produce IL-10 and TGF-ß and have a suppressive, regulatory role; and the newly discovered Th17 cells which mainly produce IL-17 and IL-22 and, as Satya Dandekar highlighted, are important in antimicrobial responses (such as against klebsiella pneumonia and candida albicans), as well as being associated with some autoimmune conditions. Paiardini also cited studies indicating that Th17 cells have an important role in tissue repair, promoting epithelial cell proliferation and thus helping maintain the integrity of the mucosal epithelium.

Paiardini’s study investigated whether Th17 cells have a role in HIV pathogenesis by comparing responses in HIV-infected and uninfected humans and in non-pathogenic SIV infection of sooty mangabeys, a natural host of SIV. He reported that, in humans, no differences were seen in blood samples, but in samples from the gut, Th17 responses were significantly reduced, even in individuals on prolonged antiretroviral therapy. Further analysis revealed that about 60% of Th17 cells in the gut expressed the HIV co-receptor CCR5 and it was this population that appeared to be preferentially depleted. A significant association was also seen between levels of immune activation—in both the gut and the blood—and Th17 cell depletion. Because non-pathogenic SIV infection of sooty mangabeys is associated with a lack of immune activation, Paiardini went on to assess levels of Th17 cells in the gut of these animals, discovering that they remained stable despite the fact that gut CD4+ T-cell numbers overall were significantly lower than in uninfected sooty mangabeys. He also showed that maintenance of gut Th17 cells was associated with a preserved mucosal epithelium, suggesting that this may be an important factor in protecting sooty mangabeys from the pathogenic consequences of SIV infection. This would be consistent with the idea, proposed by Danny Douek and Jason Brenchley of the US National Institutes of Health, that translocation of bacteria across damaged mucosal epithelium contributes to systemic immune activation in HIV infection.

The subsequent presentation from Valentina Cecchinato of the National Cancer Institute provided additional support for Paiardini’s hypothesis. Cecchinato showed that depletion of Th17 cells from the gut of rhesus macaques is a significant predictor of disease progression. Cecchinato also showed that asymptomatic “elite controller” macaques that control SIV replication maintain gut Th17 cell levels.

The third Th17 talk was delivered by David Favre of the University of California in San Francisco (UCSF). Favre reported results from studies comparing infection of pigtailed macaques and African green monkeys (AGMs) with the same AGM-derived SIV isolate, showing that while immune activation was persistent in the former species, it is transient in the latter. Favre looked at Th17 cells based on the hypothesis that these pro-inflammatory cells may contribute to immune activation in pigtailed macaques but, like the prior studies, found the opposite. Th17 cells were preserved in AGMs, but depleted from all tissues studied in the pigtailed macaques, including the colon. Since previous studies have suggested that a balance between Th17 and Treg cells may be important, Favre went on to look at Treg cells in the two monkey species, finding that the curtailment of immune activation in AGMs was associated with an early increase in Tregs in both the lymph nodes and blood. This effect was sustained over time. In contrast, only a transient increase in Tregs was seen in pigtailed macaques and it occurred much later in the course of disease. Favre suggested that the maintenance of the balance between Th17 and Treg cells may be particularly important in preventing disease in AGMs (a species that, like sooty mangabeys, does not develop immunodeficiency as a result of SIV infection). He also showed data from a study conducted in collaboration with Peter Hunt and Steve Deeks, both of UCSF, which found that the degree of imbalance between Th17 and Treg cells in rectal biopsies from HIV-infected people correlated strongly with the level of systemic immune activation.

Breaching the vaginal barrier

Scott McCoombe of Northwestern University presented new findings from imaging studies using fluorescent-labeled HIV to illuminate the early interactions between the virus and the vaginal epithelium. McCoombe showed that virions can penetrate several layers deep into the intact epithelial wall, seemingly through the junctions between epithelial cells which were previously thought too narrow to allow HIV to penetrate—the average virion size is about 80-100 nanometers, whereas the epithelium was thought impervious to particles much above 30nm. McCoombe suggested that this may facilitate heterosexual transmission by allowing the virus to get far enough to be taken up by immune system cells (such as dendritic cells, T cells and macrophages) present beyond the epithelial surface. He also noted that any factors that increase the quantity of immune cells present (such as inflammation) and/or impair the integrity of the epithelium (such as ulcers, trauma, or hormone-related thinning) would greatly increase the risk of HIV transmission. Additional work is ongoing to better understand precisely how HIV penetrates the surface of the intact epithelium.

A Twist in the Tale 

Robert “Chip” Schooley of the University of California in San Diego presented data from a therapeutic trial of a prototype of Merck’s Ad5-based AIDS vaccine candidate encoding only HIV Gag. Given the much-publicized failure of the final vaccine construct (which encoded HIV Gag, Pol, and Nef) in a large preventive trial, Schooley’s presentation offered an odd twist.

The therapeutic vaccine trial enrolled 114 HIV-infected individuals on antiretroviral therapy, who had CD4+ T-cell counts over 500 and viral loads less than 500 copies/ml for at least two years prior, and at or below 50 copies/ml at study entry. Individuals with high anti-Ad5 antibody titers (over 1:200) were excluded. Participants received three doses of either the Merck prototype vaccine or placebo. Three months after the final immunization, antiretroviral therapy was interrupted for 16 weeks. During this period therapy was reinitiated if certain safety thresholds were breached or at a participant’s request. The co-primary endpoints of the trial were the area under the curve (AUC) of all viral load measurements during the 16-week treatment interruption—AUC is a method for calculating how much virus each individual was exposed to as a function of time—and viral load set point (the average of viral load levels at weeks 12 and 16 post-interruption).

Reviewing the results, Schooley reported that a meager reduction in the viral load AUC was observed in vaccinees compared to placebo recipients but statistically this result only represented a strong trend. HIV-specific CD4+ and CD8+ T-cell responses appeared to be bolstered by vaccination and, interestingly, there was a statistically significant inverse correlation between the magnitude of the HIV-specific CD4+ T-cell response, as measured by IFN-γ ELISPOT assay, prior to interruption and the post-interruption viral load set point. No correlation was seen with the HIV-specific CD8+ T-cell response. Schooley concluded that the data support pursuing the goal of enhancing viral load control immunologically, but he believes there is not an immunogen available that is potent enough to improve upon these results. Immunologist Mike Lederman, the vice chair of the protocol for this study, was a little more sanguine. “With one antigen derived from a single HIV sequence, I was encouraged we saw any positive signal at all.” –R.J.