Keystone 2006

Steady progress presented and discussed at the annual Keystone Symposia on HIV

By Kristen Jill Kresge and Simon Noble, PhD

This is a milestone year in HIV/AIDS. Although HIV has been within the human population since the 1930s, 2006 marks the 25th year since AIDS was first medically recognized. Tristram Parslow of Emory University, one of the co-organizers of the Keystone Joint Symposia on HIV Pathogenesis and HIV Vaccines, opened the symposia by referring to the signature year, saying that it was 25 years since HIV "emerged" and ushered in the "dawn of the HIV epidemic," and reminding the audience that 40 million people are now infected worldwide and that 35 million of those don't know it. This year also marks the 10th year since the advent of HAART, and he pointed out that in those 10 years HAART has added, on average, 13 years to each patient's life, which equates to 3 million life-years saved. Despite this success, that's no comfort to the large majority of the world's infected who still don't have access to ARVs.

The Keystone HIV Symposia are one of the primary showcases for researchers from all corners of the HIV research field to come together and share data, ideas, plans, experiences, and, importantly, develop research collaborations. Presentations range from the intricate molecular biology that HIV employs to go through its lifecycle, to the complex immune effects it elicits as it interacts with its host species, to new findings uncovered as researchers look for ways to prevent infection with vaccine candidates.

Immunology in real time

One of the first presenters began by joking that, since he has never worked on HIV, he originally thought he had been invited by mistake to give one of the keynote opening seminars. Ulrich von Andrian of Harvard Medical School uses intravital microscopy to view immunological processes in living animals in real time. His research group has used 2-photon microscopy to look first-hand at homing and cellular trafficking in the lymphoid tissues, particularly lymph nodes and Peyer's patches in the intestine. After microsurgical exposure of the lymph nodes behind the knee of a mouse that leaves intact the microcirculation of lymph in afferent and efferent lymphatics and blood in high endothelial venules, different immune cell types that have been stained with differently colored fluorophores are introduced.

Von Andrian presented captivating movies showing how these immune cells interact in their natural microenvironment, and revealed just how dynamic and busy a place these tissues are, providing immunologists with the opportunity to see what they usually only get to imagine or represent in cartoon form after inference from indirect experiments. As well as being a realization of what has been previously left to the imagination, his studies provide biological insight and clues to pursue further. He showed cytotoxic T lymphocytes (CTLs) recognizing and interacting with B cells that presented a cognate antigen, the B cells then undergoing lysis, indicated by a color intensity change due to the fluorophore leaking out after cell membrane perforation. He referred to this CTL-B cell interaction as "the kiss of death" and showed that CTL contact suppressed the target B cell's motility prior to lysis. It also seems that CTL can distinguish between viable and dead target cells. The molecular mechanisms behind these two processes will be areas for more traditional analyses.

Von Andrian has also used similar imaging techniques to look at regulatory T (Treg) cells and their interaction with CTLs. His group has found that CTLs under the influence of Tregs fail to degranulate and deliver their lethal payload of granzymes and perforin, and that this defect requires CTL responsiveness to the cytokine TGF-b. They've also found that removal of the Treg cells soon reverses the defect and CTLs regain their lytic potential, raising the possibility that if Treg cells are important suppressors of CTL action in HIV infection (research is ongoing and the jury is still out) then this could form the basis of a corrective therapy.

Antiviral host factors

Host proteins that provide protection from viruses, especially retroviruses, are currently a hot research topic. It's hoped that by learning precisely how these proteins work to prevent infection that similar strategies can be developed for use in prevention or treatment. Joe Sodroski of the Dana Farber Cancer Institute at Harvard Medical School presented a keynote seminar updating current understanding of TRIM5a, an antiretroviral host factor that acts intracellularly to snuff out HIV or SIV replication (see Making a monkey out of HIV). The protein has been identified in a number of mammalian species, most notably humans and other primates, old and new world monkeys, and cows. As Sodroski noted, "human TRIM5a does not restrict HIV, and that's why we now have an HIV epidemic." New comparisons of TRIM5a in different species show that primates and cows have independently evolved their retroviral restriction factors in an example of convergent evolution, which suggests the function of TRIM5a is a vital one.

The precise mechanism of action of TRIM5a is still being worked out. Sodroski's group has been conducting a mutational analysis to discern which domains within the protein are most important for its function. Comparisons of TRIM5a from different monkey species show that there are four variable regions (V1 to V4) within the SPRY domain, and mutational analyses indicate that the amino acid residue at position 332 within V1 is particularly important; substitution of the positively charged arginine residue there results in a TRIM5a that potently restricts HIV infectivity. Sodroski said that "the problem with human TRIM5a is that it has an arginine at position 332; if it had just about any other residue there it would be a pretty good inhibitor of HIV." By extension from his earlier statement, that arginine residue at position 332 in TRIM5a goes a long way to explaining why humans are susceptible to HIV and why there is now a pandemic.

The viral capsid protein is the major determinant of sensitivity to restriction by monkey TRIM5a proteins, and recognition and binding correlates with the degree of restriction. Structural studies suggest to Sodroski that the pseudotrimeric symmetry of the capsid protein may form a target for the trilobular coiled-coil domain within TRIM5a. In questions after his talk, Sodroski was asked to comment on the fate of the capsid in TRIM5a-expressing cells, and he thinks that "the evidence is pointing to acceleration of uncoating." Why this should derail HIV's replication is not clear, but he emphasized the need for further research to fully understand the molecular steps in retroviral uncoating.

That potentiation of restriction correlates with an increase in capsid binding leads him to believe that pharmacological intervention may be attainable. Jonathon Stoye of the National Institute for Medical Research, London, pointed out later in his own presentation, that the restriction factors themselves are unlikely to represent useful drugs—"for a start, they're very large proteins, which disqualifies them." He suggested some conceivable alternatives: the use of gene therapy to deliver the restriction factors, developing small molecule drugs that mimic restriction factors, or developing drugs that can either re-target or stimulate expression of endogenous restriction factors.

Nucleic acid scrambler

Another important, recently-discovered host restriction factor is APOBEC3G (A3G), a ribonucleoprotein complex with cytidine deaminase activity that can effectively scramble RNA or DNA and render it nonsensical, thereby stopping HIV replication in its tracks (see Guardian of the genome). This cellular protein is the long sought-after target of the HIV protein Vif (viral infectivity factor) that is essential for virus infectivity and acts by binding to A3G and tagging it for degradation by the proteasomal pathway. Although A3G is a cellular protein, it is packaged within virions and acts soon after uncoating when the virion enters a new cell.

Warner Greene of the Gladstone Institute at the University of California, San Francisco, gave an update on new insights into A3G. The biology of A3G is intrinsically linked to its ability to exist in a high- (HMM) or low-molecular mass (LMM) complex. In resting CD4+ T cells and monocytes that are not permissive for HIV replication, the enzymatically-active LMM form of A3G is expressed and it acts as a potent post-entry restriction factor for HIV, whereas in activated CD4+ T cells and macrophages the enzymatically-inactive HMM form of A3G is expressed and the HIV replication cycle can proceed.

Greene and his research group have looked at virion-packaged A3G. Surprisingly, this virion A3G is in an enzymatically-inactive HMM form, although distinct from the cellular A3G HMM complex. They've also determined that it is bound to HIV RNA and requires HIV's RNaseH enzyme to trigger its cytidine deaminase activity. So HIV's RNaseH activity is required for both the generation of the ssDNA template that is the target of A3G activity, and for the activation of HMM virion-packaged A3G activity. Greene noted that a host restriction factor (A3G) that requires activation by a virus enzymatic activity (HIV RNaseH) is an example of "an unusual host-pathogen relationship."

As to the precise nature of the HMM complex itself, co-precipitation studies indicate that A3G associates with a diverse mix of over 75 different proteins and that it contains cellular RNA components, a class of retrotransposon RNAs, Alu, and hY. Alu is one of the most prolific mobile genetic elements in mammalian genomes, and the finding of such elements raises the possibility that the HMM A3G is defending against the "threat within" posed by retrotransposons after T-cell activation. Unfortunately this switch to the HMM A3G complex could open the door to external threats like exogenous retroviruses.

Identification of the proteins in the HMM A3G by Greene's group prompted them to search databases for comparisons and they find similarities to the make-up of Staufen RNA granules (important in RNA localization), Ro-La ribonucleoproteins (which have roles in RNA processing and possibly as chaperones), and pre-spliceosomes (an intermediate in mRNA processing). Indeed Greene went on to speculate that HMM A3G could be a collection of related complexes.

Greene rounded off his talk considering the therapeutic implications of recent findings about A3G, and said it would be nice to identify small-molecule inhibitors of the HMM complex, or ones that promote its dissociation, to preserve the restrictive LMM A3G in activated CD4+ T cells.

Figure 2.

View Larger Image

Alternative models

Because of the limits to what researchers can learn from human subjects infected with HIV-1 or the best monkey models of acquired immunodeficiencies, they are always looking for other models of disease that perhaps hold important lessons for HIV-1 infection in humans. Sarah Rowland-Jones of the UK Medical Research Council Laboratories in the Gambia and in Oxford, UK, has been studying HIV-2 infection, which she considers a neglected model of naturally-attenuated HIV infection in humans."

HIV-2 evolved from SIV found in sooty mangabeys (SIVsm; whereas HIV-1 evolved from SIV found in chimpanzees) and is endemic in West Africa with a seroprevalence around 1%, meaning about 1 million people are infected. HIV-2 infection provides no protection from infection with HIV-1 and is probably a risk factor. A minority of those infected with HIV-2 will develop AIDS and could benefit from ARVs, but the majority, 80-85%, will not progress to AIDS and it is these natural long-term non-progressors (LTNPs) that may provide clues to what is protective in immunodeficiency virus infection.

The CD4+ T-cell depletion in HIV-2-infected individuals is less rapid than in HIV-1 infection, and while most HIV-2-infected people have an undetectable to low plasma viral load, their proviral burden is similar to that seen in HIV-1 infection. Strong HIV-2-specific CTL responses are evident in infected individuals and these frequently show cross-reactivity with HIV-1 epitopes, but the overall magnitude and cytotoxicity of HIV-2-specific CTLs and the magnitude of interferon (IFN)-g responses does not differ markedly from HIV-1 infection.

However Rowland-Jones and her colleagues have found that there are significant differences in natural killer (NK) cell activity (in terms of circulating numbers, cytotoxicity, and cytokine and chemokine secretion) between HIV-1- and HIV-2-infected individuals who have normal CD4+ T-cell counts, with the most striking difference being much higher production of the chemokine MIP-1b by NK cells in HIV-2-infected individuals. These differences are not seen in individuals with lower CD4+ T-cell counts.

Rowland-Jones' team have also found that the magnitude and breadth of CD4+ T-cell responses are greater in healthy HIV-2-infected individuals and that they also have a functionally distinct population of IFN-g+/interleukin (IL)-2+ CD4+ T cells that are not seen in HIV-1 infection, supporting again the notion that T helper cell responses are important in HIV infection. Rowland-Jones postulated that HIV-2-infected individuals with a normal CD4+ T-cell count may represent a distinct LTNP population with broad and sustained CD4+ T-cell and preserved NK-cell responses, and concluded that HIV-2 is a potentially informative model" with regard to HIV-1 infection.

Sooty mangabeys

In a similar vein, Mark Feinberg of Emory University is studying host-virus relationships in non-pathogenic SIV infection of reservoir hosts to learn more about the nature of successful immune responses to immunodeficiency virus infections. His research team is grappling with some fundamental questions about AIDS pathogenesis, specifically asking whether chronic immune system activation is the primary mechanism driving CD4+ T-cell depletion, and therefore AIDS progression, in HIV-1 infection. He noted that chronic HIV infection is associated with chronic immune activation, while other chronic viral infections of humans (such as hepatitis B and C) are not.

Feinberg's team is intrigued by the absence of disease in SIV-infected sooty mangabeys (SMs), the natural reservoir species for SIVsm that is commonly naturally-infected at sexual maturity both in the wild and in captivity. SMs show no signs of immunodeficiency, neuropathology, or wasting syndromes despite high levels of chronic viremia, and they have normal levels of naïve CD4+ and CD8+ T cells, no increase in CD8+ T-cell proliferation or evidence of pathologic CD8+ T-cell activation, suggesting that perhaps this attenuated immune activation may protect SMs from developing AIDS.

Feinberg showed that SMs have lower virus-specific CD8+ T-cell responses than humans with HIV infection, and there is no correlation between the magnitude of SIV-specific CD8+ T-cell responses and the level of viremia in infected SMs, suggesting that these responses are not all that important.

His group has developed a comparative infection model where they can compare SIVsm infection in SMs and in rhesus macaques, in which SIVsm is pathogenic, to try to find what underpins the different infection outcomes. Feinberg noted that a large difference early in infection between SMs and macaques in their levels of CD8+ T-cell proliferation implicates perhaps an underlying altered innate response, and went on to show that macaque NK cells proliferate more than SM NK cells during primary and chronic SIVsm infection.

They have also found that plasmacytoid dendritic cells (pDCs) in infected SMs are not activated and don't migrate to the lymph nodes that are the key generative sites for antiviral immune responses, "so it's not surprising that an active immune response is not initiated" said Feinberg. That lack of pDC activation appears to be due to a specific defect in TLR signaling; SM pDCs do not produce IFN-a in response to ex vivo TLR7 or TLR9 stimulation, nor do they produce IFN-a in response to ex vivo stimulation with inactivated SIV. This defect is not general though, since SM pDCs do produce pro-inflammatory cytokines like IL-12 and TNF-a after various ex vivo stimuli, and do produce IFN-a following exposure to other stimuli such as influenza virus. Feinberg thinks that this failure of pDC activation in SMs may have a large impact on activation of downstream innate and adaptive immune responses, and said that SMs "see the virus but respond in a different way." He thinks the defect in type 1 IFN production in response to SIV might be particularly important.

To further illustrate the lack of global immune activation in SMs, Feinberg's team has compared gene expression profiles in CD4+ and CD8+ T cells and PBMCs taken from infected and non-infected humans with HIV and SMs with SIVsm. This analysis reveals that SMs have a far more quiescent immune system in terms of upregulation of genes, and that type 1 IFN response genes are among the most strongly upregulated genes in T cells of HIV-infected humans but not in SIV-infected SMs. Feinberg concluded by suggesting that this failure to activate pDCs in response to viral infection, which then avoids aberrant immune activation, is likely the primary mechanism protecting SMs from AIDS. By extension, he thinks that the chronic CD8+ T-cell activation and bystander immunopathology characteristic of human AIDS might be the result of chronic activation of innate host responses, rather than the primary defect in itself.

Elite controllers

Bruce Walker of Massachusetts General Hospital and Harvard Medical School is looking at HIV-1 infection in humans, but he is refining the cohorts of infected individuals he is studying to try to tease out new observations. His research group is interested in determining what accounts for the differences in time from initial HIV infection to AIDS, which on average is about 10 years but can actually be anywhere from 6 months to more than 28 years and counting. They are studying asymptomatic HIV-1-infected individuals who control their virus without the need for ARV intervention, and have further classified individuals to distinguish viremic controllers (VCs), who have plasma HIV RNA levels below 2000 copies/ml blood, from elite controllers (ECs), who have undetectable plasma HIV RNA (<75 copies/ml blood by bDNA or <50 copies/ml blood by ultrasensitive PCR). These criteria must be met for at least one year.

So far Walker's group have not found viral or host genetic factors that are strongly associated with an individual's controller grouping. But they have found host immunologic factors, and Walker showed in his presentation that ECs have less robust CD8+ T-cell responses (lower magnitude and to fewer epitopes) than VCs or chronically-infected individuals. ECs do, however, have more focused Gag-specific CD8+ T-cell responses. ECs also have a significantly higher percentage of HIV-specific, and especially Gag-specific, CD4+ T-cells that secrete both IL-2 and IFN-g.

Walker is now looking to expand his EC cohort to 1000 individuals (he has already identified 200) and issued a recruitment call for his Elite Controller Collaborative Project that currently includes more than 45 collaborators from all over the US and some in Europe. He hopes to define haplotype maps of ECs in collaboration with investigators at the Broad Institute, Boston, and harness the power of the Human Genome Project.

Mucosal focus

There have been a flurry of recent publications and presentations on how HIV wreaks havoc on the cells of the gut during the very earliest stages of HIV infection (see Beast in the belly and CROI covers advancements from start to finish). The importance of understanding the effects of the virus on the intestinal tissues and the mucosal immune responses at this site were a recurring theme at Keystone. "It's an important lymphoid compartment and it needs to be looked at, regardless of route of HIV transmission," said Barbara Shacklett from the University of California, Davis.

Information is now building on the loss of CD4+ T cells in the gut of humans during acute infection that was first reported with SIV infection in rhesus macaques. Researchers, including Shacklett, are also studying HIV-specific responses in mucosal tissues of chronically HIV-infected humans. Her group compared the immune responses in peripheral blood samples from 13 HIV-infected individuals with those in the gut, as obtained by rectal biopsy, in several epitope mapping studies. They found the magnitude and antigenic specificity of these responses are quite similar when measured by IFN-g ELISPOT. Although not statistically different, the trend in these individuals was towards a higher breadth of immune response in the gut, which Shacklett said is expected with such a large amount of virus in these tissues.

Her laboratory then looked at the differences in the functionality of the CD8+ T-cell responses in the two compartments in 22 HIV-infected individuals, 6 of whom were taking ARVs, to see if the functional diversity of the mucosal immune responses correlated with clinical status. Analysis was done for different cytokines, including expression of MIP-1b, TNF-a, IFN-g, IL-2, and expression of CD107a, a protein present in the membrane of cytotoxic granules that is transiently expressed as a result of degranulation and therefore reflects their cytotoxic capabilities (J. Immunol. Methods. 281, 65, 2003). Here Shacklett found a significant difference. Although overall the functional profiles of the immune responses in the peripheral blood and gut were similar, the Gag-specific CD8+ T-cell mucosal responses were greater and were characterized by CD107a expression.

There is some debate in the field about the correlation between CD107a and cytoxicity, but this finding suggests that there is an active CTL response in the gut mucosal tissues during chronic infection, and alludes to an important difference between the gut mucosa and blood.

Shacklett also reported that the functionality of CD8+ T cells within the mucosal tissues of the gut during chronic HIV infection varied with viral load. Individuals with higher viral loads (>30,000 viral copies/ml of blood) were more likely to have CD8+ T cells expressing a single cytokine than those with viral loads below 5,000 copies/ml, who were more likely to have CD8+ T-cells in mucosa that expressed multiple cytokines.

Shacklett's group also looked at secretion of perforin, a protein associated with cytoxicity, by CD8+ T cells in the gut of rhesus macaques during acute SIV infection and found that the quantity peaked in the gut around 21 days after intravaginal infection. But this is "too little, too late" since the bulk of the damage to this compartment occurs within 14 days. After around 180 days the levels of perforin dropped to the same levels as those in uninfected control animals. "There are still CD8+ T cells in the gut, but they aren't able to secrete perforin," said Shacklett. There is also evidence that CD8+ T cells in lymph nodes are low in perforin during chronic infection.

The low levels of this protein in the gut may be the result of a mechanism that evolved to protect the gut from what Shacklett refers to as "friendly fire." Perforin's ability to poke holes in cell membranes could be potentially damaging to the thin intestinal lining and reducing quantities of the protein may be one way the immune system keeps it in check. However it could inadvertently offer an advantage to HIV. "This is one of the things that could help HIV survive as a chronic infection," Shacklett suggested.

Considering the importance of mucosal immunity, the development of AIDS vaccine candidates that stimulate strong immune responses at these tissues remains a priority. To this end several researchers are working on both adjuvants and vaccine vectors that may augment mucosal immune responses.

To date there are no mucosal adjuvants that are approved for human use. Of the known mucosal adjuvants, cholera toxin is by far the most potent but its safety profile makes it unsuitable for evaluation in human volunteers. Therefore many groups have been hard at work modifying bacteria to make them safe while preserving their capabilities to stoke a mucosal immune response. Susan Barnett from Chiron presented work on the company's mucosal adjuvant, known as LTK63, which is a non-toxic mutant of heat-labile enterotoxin (LT) from Escherichia coli. This LT mutant carries a single point mutation and in extensive preclinical studies in mice and rabbits it was found to be safe and immunogenic, producing strong serum and mucosal antibodies (serum IgG and vaginal IgA).

This adjuvant is now being evaluated in a clinical trial for an influenza vaccination, where it is being administered intranasally. Chiron is also studying the immunogenicity of the LTK63 adjuvant with both mucosal and parenteral immunizations in female rhesus macaques in collaboration with Chris Miller at the University of California, Davis.

Other methods investigators are exploring to induce robust immune responses at mucosal tissues is the type of vaccine vector or route of delivery. Intramuscular immunization is generally considered an ineffective way to provoke mucosal immunity, according to Stephen Udem of Wyeth, but results presented by the company at Keystone indicate otherwise. Researchers at Wyeth have found that intramuscular delivery of their recombinant vesicular stomatitis virus (rVSV) vaccine encoding HIV Gag may be a promising inducer of immune responses at mucosal tissues (see Renewed Promise).

Although this viral vector has been used in research for years, "nobody has really looked at VSV as a mucosal immunogen," said Udem. "There's extraordinarily little already known."

In a poster at Keystone, researchers detailed the HIV-specific T cell immunity induced at three different mucosal tissues. Mice immunized intramuscularly with a prime and boost of rVSV vector expressing HIV Gag developed strong CD8+ T-cell responses to both HIV Gag and the N protein of VSV in splenocytes and lamina propria lymphocytes. These CD8+ T cells secreted IFN-g  when exposed to antigenic peptides and persisted for at least a month after the booster immunization. The vaccine faired less well in intraepithelial lymphocytes where only weak immune response were observed.

Wyeth is planning to evaluate the mucosal responses induced by rVSV more closely in rhesus macaques given the vaccine intramuscularly. Even more interesting data might come from intranasal administration, a route thought to induce more potent responses at mucosal tissues, but according to Udem this approach is currently not being studied. Wyeth's priority is establishing the safety of live, but attenuated VSV-based vaccines. "So much of what we're doing right now is trying to get a sufficient level of comfort with regulatory agencies," he says.

The US Food and Drug Administration has agreed to a Phase I clinical trial with an intramuscular administration of Wyeth's VSV-based AIDS vaccine and Udem expects this will start in early 2007. This trial will be conducted with the HIV Vaccine Trials Network (HVTN).