Capsules from Keystone

A broad collection of updates at this year’s joint symposia served to inform the development of future AIDS vaccine candidates

By Andreas von Bubnoff and Richard Jefferys

A lot of snow fell the week of March 22-27 in Keystone, Colorado, accompanied by a flurry of updates on vaccine research and development by researchers who gathered for the joint Keystone Symposia on “HIV Immunobiology: From Infection to Immune Control” and “Prevention of HIV/AIDS.”

This year’s meeting marked the 25th anniversary of Keystone Symposia’s first meeting on HIV/AIDS, which was held in 1984, three years after the first HIV infections were described. The speakers at the conference’s opening session left no doubt that 28 years after HIV emerged, there is still much to do. “We probably got rid of the iceberg, but under the water there is a mass of ice and that’s the current AIDS epidemic,” said Didier Trono, of the Ecole Polytechnique Federale de Lausanne, in his introductory address. Nobel laureate Francoise Barre-Sinoussi from the Institut Pasteur said recent developments in the AIDS vaccine field show that “we have to come back to basic science.”

If the plethora of findings presented at this year’s conference is any indication, researchers are already heeding her call. A broad collection of updates, ranging from imaging studies of viral transfer to results with vaccine candidates in animal models and clinical trials, all served to inform the development of future vaccine candidates.

Tracking transmission

Evidence has been accumulating to suggest that the majority of productive clinical HIV infections after heterosexual transmission can be traced back to a single transmitted founder virus (1). This data, collected by George Shaw, a professor in the department of medicine at the University of Alabama at Birmingham, and others, is intriguing to vaccine researchers who have to deal with HIV’s epic diversity (see HIV Transmission: The Genetic Bottleneck, IAVI Report, Nov.-Dec. 2008).

At Keystone, Shaw mentioned additional data that confirm and extend these observations. Of 171 heterosexual HIV transmissions studied, 81% can be traced back to one transmitted virus, he said, while in only 19%, more than one founder virus can be identified (2).

The researchers also used 454 sequencing to analyze more fully an individual with evidence of a single transmitted founder virus—analyzing not just 30, but 600,000 env sequences in that person—and found the same result.

Shaw extended these studies to additional populations with different routes of transmission—men who have sex with men (MSM) and injection-drug users (IDUs). Sequence analyses of HIV in 49 MSMs showed that 40% of them had infections from more than one transmitted founder virus, which is double the percentage compared with heterosexual transmission, Shaw said. He presented data from one MSM whose infection could be traced back to six transmitted founder viruses. Preliminary results from IDUs suggest that 60% (three of the five studied) of infections stemmed from multiple founder viruses. One IDU Shaw evaluated was infected with at least nine transmitted founder viruses. “This is highly relevant for vaccine development because with a vaccine, you would ultimately like to prevent all of these transmissions,” said Shaw.

DNA/Ad5 against rigorous challenge

As findings by Shaw and others shed light on the transmitted virus in humans, researchers are starting to replicate this in animal models. David Watkins, professor at the department of pathology and laboratory medicine at the University of Wisconsin-Madison, collaborated with Shaw and Brandon Keele, then at the University of Alabama at Birmingham, to make a repeated low-dose rectal challenge with the swarm virus SIVsmE660 in macaques more similar to the situation in humans by titrating the challenge so only one to three virus variants would get across the mucosal barrier to cause infection.

The animals were vaccinated with a DNA prime and adenovirus serotype-5 (Ad5) boost regimen containing all SIVmac239 genes except for env. Intramuscular vaccination of the macaques with three DNA primes and a single Ad5 boost induced massive T-cell responses, the strongest of which were to Gag, targeting an average of 20 different epitopes of SIVmac239 (see AIDS Vaccine Researchers STEP Up to the Challenge, IAVI Report, Sep.-Oct. 2008).

Both the vaccinated and control animals became infected with the heterologous E660 after an average of four challenges. But the vaccinated animals had a lower average peak viral load than controls—32,000 copies/ml compared to 2.5 million in unvaccinated controls. Vaccinated animals also had a much lower average set-point viral load (201 copies/ml as compared to 77,000 for controls). This was the first time that a non-replicating T-cell vaccine showed such control of acute viral load, Watkins said.

The only other time this has been observed was with SIVmac239∆nef, he added. But while SIVmac239∆nefvaccinated animals are protected from intravenous (IV) challenge with the homologous SIVmac239 or SIVmac251, Watkins showed that they were not protected from IV E660 challenge. In some animals, SIVmac239∆nef recombined with E660 to give a much more pathogenic virus. “The stock of E660 was difficult to protect against even with our best vaccine after IV challenge,” Watkins concluded.

DNA prime doesn’t pay

Dan Barouch, an associate professor of medicine at Beth Israel Deaconess Medical Center and Harvard Medical School, also reported results from a study in macaques testing a prime-boost regimen of DNA/Ad5 vaccines. Barouch’s adenovirus vaccine, referred to as Ad5HVR48, is composed almost entirely (~98%) of Ad5, except for the hexon protein, which is swapped with the same protein from the less common Ad48 serotype.

The study involved 30 rhesus macaques, evenly divided into five groups, each receiving a different vaccination regimen (see table, below). Macaques with major histocompatibility complex (MHC) genes that have been associated with superior control of SIV replication (Mamu A*01, B*08, and B*17) were specifically excluded.

Group   Vaccination Regimen 
 1  DNA/Ad5HRV48 (SIVmac251 Gag, Pol, Nef, Env)
 2  DNA (SIVmac251 Gag, Pol, Nef, Env) + adjuvants encoding MIP-1α and flt-3 ligand / Ad5HVR48 (SIVmac251 Gag, Pol, Nef, Env)
 3  Ad5HVR48 (SIVmac251 Gag, Pol, Nef, Env)
 4  Ad5HVR48 (SIVmac251 Gag, Pol, Nef)
 5  Placebo

Six months after the last immunization, all animals were challenged intravenously with a high-dose of SIVmac251. Immediately prior to challenge, vaccine-induced SIV-specific T-cell responses were five-fold higher in the macaques that received the DNA prime/Ad5HVR48 boost compared to those that received only Ad5HVR48. But Barouch was surprised to find that post-challenge outcomes were not consistent with this immunogenicity data. Set-point viral loads were lowest in the animals that received only Ad5HVR48, averaging 4.4 logs, whereas in groups one and two, animals that received the DNA/Ad5HVR48 candidates had viral loads similar to those seen in the macaques that received placebo (5.2 logs in group one, 5.8 logs in group two, and 5.5 logs in the placebo group).

Barouch conducted a post-hoc exploratory analysis of the two Ad5HVR48 groups combined, compared to the two DNA/Ad5HVR48 groups combined. In this analysis, prime-boost recipients had higher viral loads (by around 0.75 logs) than those animals given only Ad5HVR48. Although this difference was statistically significant, Barouch stressed that because this was an exploratory analysis, “the result must be viewed as hypothesis-generating rather than conclusive.” After 500 days of follow up, four of 12 animals in the DNA/Ad5HVR48 groups were alive, compared to 10 of 12 in the Ad5HVR48-only groups and one of six among placebo recipients.

In conclusion, Barouch speculated that the DNA/Ad5HVR48 was inferior because the DNA vaccine disproportionately increased CD4+ T-cell response to Env compared to the other antigens—prime-boost recipients showed four- to five-fold higher responses to Gag, Pol, and Nef, but 10-fold higher responses to Env. Barouch said one possible explanation is that Env-specific CD4+ T cells might have provided more targets for SIV, thereby counteracting the protective effect of vaccination seen in the Ad5HVR48 groups. A similar phenomenon was described before in an SIV challenge study involving a varicella zoster virus vector-based vaccine that only induced Env-specific CD4+ T-cell responses (3).

Barouch’s prime-boost regimen is similar to one developed at the Vaccine Research Center (VRC) at the National Institute of Allergy and Infectious Diseases (NIAID). The VRC’s DNA/Ad5 prime-boost regimen is slated to undergo testing in a 1,200-person Phase II trial called HVTN 505, which is a smaller version of the originally proposed Phase IIb test-of-concept trial known as PAVE 100. Scott Hammer, the principle investigator of HVTN 505, stressed that Barouch’s results could not be directly extrapolated to the VRC’s DNA/Ad5 prime-boost regimen. There are differences—Barouch’s vector has a different hexon protein and the VRC Ad5 vector also has additional genes deleted (E3 and E4), which reduces the expression of Ad5 proteins (4). Also, the VRC’s DNA vaccine candidate consists of six different plasmids, one for each encoded antigen (Gag, Pol, Nef, and Env from clades A, B, and C), and dividing the vaccine in this way has been shown to reduce the bias toward Env-specific CD4+ T-cell responses by enhancing Gag-specific and Nef-specific CD4+ T-cell responses (5).

Other macaque studies using the VRC DNA/Ad5 candidates have also produced different results than those obtained by Barouch. In the most cited study, the DNA/Ad5 prime-boost was shown to perform comparably to Ad5 alone (6). A poster by Diane Bolton from the ImmunoTechnology Section at the VRC also included a comparison of the VRC’s DNA/Ad5 to just Ad5, albeit on a slightly different schedule (two Ad5 booster shots were given). Again, while DNA/Ad5 was not shown to be superior to Ad5 alone in this study, the two regimens performed equally well and reduced viral load set points by one to two logs compared to placebo. The US Food and Drug Administration is currently reviewing the HVTN 505 protocol.

The shifting treatment landscape: Implications for vaccine trials 

 Scott Hammer, professor of medicine at Columbia University, reviewed the potential impact of antiretroviral therapy (ART) on future vaccine trials as part of the Prevention of HIV/AIDS symposium at this year’s Keystone meeting, held March 22-27.

Hammer explained that with current vaccine candidates aiming to reduce post-infection viral load, early initiation of antiretroviral therapy (ART) has to be considered when designing studies, specifically evaluating how to censor viral load endpoints among participants who start ART. Although there is currently no consensus regarding initiation of treatment during acute infection, Hammer noted that many clinicians favor it, particularly because of studies indicating that gut T cells are rapidly lost during this period.

Treatment guidelines are also starting to shift toward earlier treatment, driven by data from recent studies, which found that earlier treatment significantly reduced the risk of illness and death (7). This will have to be taken into account in vaccine trials that include time to initiation of ART as an endpoint, Hammer concluded.

For pre-exposure prophylaxis (PrEP), significant developments are close on the horizon. Results from the first efficacy trial may come this year and if PrEP is effective, investigators may need to consider whether it should be offered to vaccine trial participants. —RJ



Building a better antigen

Barouch also presented results from a mosaic vaccine study in rhesus macaques designed to achieve optimal coverage of various 9mers of a given protein. In theory, he said, a mosaic vaccine with sequences encoding a given number of mosaic antigens will have better breadth and coverage than a vaccine candidate encoding the same number of consensus antigen sequences or one encoding the same number of naturally occurring antigen sequences (8).

In this study, Barouch used an Ad26 vector with sequences encoding two mosaic antigens for each of the three HIV proteins Gag, Pol, and Env, designed to optimize coverage of the global HIV-1 M group, which represents the vast majority of all global HIV sequences. Macaques were immunized intramuscularly with the mosaic vaccine candidate; an Ad26 vector with one set of consensus antigen sequences for Gag, Pol, and Env; or an Ad26 vector with one set of natural sequences for these proteins that had the broadest 9mer coverage of all clade C HIV sequences in the Los Alamos National Laboratory HIV sequence database.

Barouch found that the mosaic vaccine generated CD8+ and CD4+ T-cell responses directed at about three to four times more epitopes than the other two vaccines, using a pool of global potential T cell epitope peptides developed by the HIV Vaccine Trials Network that represents 85% of the global viral sequences. The mosaic vaccine increased the breadth of the immune responses as well as their depth, which is the simultaneous induction of responses to different versions of the same epitopes. The mosaic vaccine also showed a broader response than the other two vaccines to consensus Gag proteins from clades A, B, and C, and to two natural clade C Gag proteins.

These results suggest that “the evaluation of rare serotype Ad vectors expressing HIV mosaic antigens optimized for global coverage may be warranted,” said Barouch.

STEP by step

Susan Buchbinder, principal investigator of the STEP trial, provided another update on this now notorious Phase IIb trial of Merck’s Ad5-based vaccine candidate, MRKAd5. Receipt of MRKAd5 was associated with enhanced susceptibility to HIV infection, most significantly among uncircumcised MSM, who had pre-existing immunity to the Ad5 vector. An explanation for these findings has so far been elusive.

Polysynapses formed between an hiv-infected T lymphocyte and three target cells  HIV Gag: red; tubulin: green; target cells: blue. Image by Dominika Rudnicka, Nathalie Sol-Foulon, and Olivier Schwartz, at the Institut Pasteur

HIV can be transmitted as cell-free or cell-bound virus, and researchers are now starting to literally picture how cell-bound virus is transmitted between cells through virological synapses (see Research Briefs, this issue). At Keystone, Olivier Schwartz, head of the virus and immunity unit at the Institut Pasteur in Paris, France, showed that in cultured human cells, the most common mode of cell-bound HIV transfer is between an infected cell contacting one or several (polysynapses), target cells (9). In polysynapses, Schwartz found that viral Gag particles move simultaneously from the infected cell to several target cells. Gag and Env were colocalized at the contact zone between the cells, suggesting that infectious particles are there. Using transmission electron microscopy, he observed viral particles, at various stages of budding and maturation, accumulate in the extracellular cleft between an infected cell and target cells of a polysynapse. Three-dimensional reconstructions of fluorescent images of Gag proteins at virological synapses often showed them arranged in a ring-like structure at the contact zone (see image above). —AvB  


Some scientists have highlighted the finding that placebo recipients in the STEP trial with the highest pre-existing Ad5 antibody titers had the lowest risk of HIV acquisition, and perhaps some unknown factor renders individuals with high anti-Ad5 titers less susceptible to HIV infection. But based on several analyses, Buchbinder said there was no association between baseline Ad5 antibody titers and risk of HIV acquisition. Herpes simplex virus (HSV)-2 seropositivity was associated with a roughly two-fold increased risk of HIV infection among STEP participants, but it did not explain the enhanced risk of infection among vaccine recipients either.

Buchbinder then unveiled new data from the extended follow-up of STEP participants from October 2007 to January 2009, the period after the volunteers were unblinded. Buchbinder said that while there was a slight drop in high-risk sexual activity after unblinding, risk behavior quickly returned to the level previously observed during the trial. The rate of HIV acquisition among MSM remained high, with 48 new infections occurring during this period—26 among MRKAd5 recipients and 22 in the placebo group. Buchbinder plotted the occurrence of these infections over time, and showed that the difference in HIV incidence between vaccine and placebo groups seems to be disappearing. But she emphasized that while this finding may offer some reassurance that the enhancement effect of Ad5 is time-limited, the results must be interpreted cautiously because of the small numbers of volunteers in these groups. Buchbinder also reported that 12 additional infections have occurred among women volunteers (only one had occurred when the first trial results were announced) and these were divided evenly between the vaccine and placebo groups.

In a complementary presentation, Julianna McElrath from the Fred Hutchinson Cancer Research Center discussed ongoing immunological studies. Although no immune parameters have been identified that discriminate between individuals that became infected versus those that did not, some hints of a vaccine effect on viral load have emerged. These findings derive from a very small number of STEP participants with favorable HLA alleles (HLA B*57 and HLA B*27). In this small subset of individuals, it appears that receipt of MRKAd5 led to slightly lower set-point viral loads compared to placebo recipients with the same HLA alleles. A collaboration between Bruce Walker, director of the Ragon Institute, and David Heckerman at Microsoft Research has also suggested that the targeting of certain epitopes by vaccine recipients led to better virus control (see Canvassing CROI, IAVI Report, Jan.-Feb. 2009).

McElrath also presented data generated with Mark Connors, chief of the HIV-specific immunity section at NIAID, who developed a new assay that measures the ability of CD8+ T cells to kill HIV-infected CD4+ T cells in vitro (seeResearch Briefs, IAVI Report, Jan.-Feb. 2009). When this assay was used to test the efficacy of CD8+ T-cell responses induced by two doses of MRKAd5 in HVTN 071 (sufficient cells were not available from the STEP trial), the extent of cell killing was poor and generally comparable with what Connors has reported for individuals with progressive HIV infection. McElrath noted that it’s likely that vaccine candidates will instead need to induce the type of efficient killing Connors has documented among elite controllers.

Mucosal protection

Chris Miller, a professor at the School of Veterinary Medicine at the University of California in Davis, and Barbara Felber, chief of the human retrovirus pathogenesis section at the National Cancer Institute, both presented studies suggesting that, in Felber’s words, the dogma that only mucosal vaccination can induce mucosal immune responses, may not be entirely true.

Miller presented studies exploring the possible mechanism for how IV vaccination of macaques with an SIV/HIV hybrid strain known as SHIV 89.6 protects about 60% of macaques from SIVmac239 vaginal challenge (10; 11). Previously he had found that for six months after SIV challenge, the viral load of 60% of SHIV 89.6 vaccinated macaques remained below 10,000 copies/ml, two to three logs lower than unvaccinated controls (12). These animals were protected from progression to AIDS, and did not show progressive CD4+ T cell loss compared with either vaccinees with viral loads higher than 10,000 copies/ml, or controls.

The current studies found no definitive correlates of protection in blood, but in the vagina Miller found that before challenge, 60% of the vaccinated monkeys had SIV Gag-specific CD8+ T-cell responses, and all of them had SIV Gag-specific CD4+ T-cell responses. “The fact that the CD4s were [in] 100% [of the monkeys] and we are only getting protection in 60% suggests that the CD4s aren’t the protective cell type but the CD8s are,” Miller said.

Depletion of CD8+ T cells at the time of challenge resulted in massive virus replication, including enhanced virus levels in the vagina and the cervix compared with unvaccinated controls. This, along with the observation that all vaccinated animals had CD4+ T-cell responses in the vagina but only 60% were protected, suggests to Miller that CD4+ T-cell responses might enhance viral replication in the genital tract. “There is a fine balance between protection and enhanced viral replication” elicited by the vaccine, he said.

These experiments also show that one doesn’t necessarily have to vaccinate via a mucosal route to elicit mucosal immune responses. “We are IV inoculating [these animals] with a replicating virus, and that’s inducing these mucosal T-cell responses,” Miller said. “The more we look at this stuff the more we find that any time we elicit a strong systemic immune response, a lot of it spills off into the mucosa.”

Felber also found that systemic vaccination can lead to mucosal immune responses. She observed that intramuscular electroporation of a DNA vaccine encoding the SIVmac239 proteins Gag, Pol, Env, and Nef-Tat-Vif, induced cellular and humoral immune responses in the broncheo alveolar lavage (BAL) of rhesus macaques as well as cellular immune responses in the rectal mucosa. The cellular immune responses in the BAL were between 2 and 40 times higher than in blood, and were still observed 45 weeks after vaccination, much longer than the two weeks that immune responses are detectable after classical intramuscular injection. In the BAL, the vaccination also induced Gag-specific IgA antibodies.

She said this is the first time that electroporation has been shown to induce mucosal immune responses. “There has been this dogma that in order to get mucosal immune responses, you need to deliver your antigen via dendritic cells or mucosal tissue,” said Felber. “[But with] a very efficient delivery like intramuscular electroporation, we find beautiful mucosal cell and humoral immune responses.”

Antibodies: Better together or alone?

Johannes Scheid from Rockefeller University presented results from a recently published study that suggests that antibodies in HIV-infected people can work in concert to fight the virus (13). Scheid and colleagues isolated gp140 trimer-binding IgG memory B cells from six HIV-infected individuals with high titers of broadly neutralizing sera against different HIV strains. The researchers found between 22 and 50 independent memory B cell clones in each person, and made monoclonal antibodies from them. These B cell clones did not include any of the four already known broadly neutralizing antibodies, all of which also bind the gp140 trimer.

By themselves, the isolated antibodies had some neutralizing activity, although only rarely against HIV strains that are more difficult to neutralize, Scheid said. They were also much less potent than the four known broadly neutralizing antibodies, he added. Overall, the neutralizing activity of the isolated antibodies was less broad than what was observed for the original sera. In some cases, recombining equal amounts of some of the monoclonal antibodies at high concentrations reconstituted the broad neutralizing activity found in the original serum from the same individual, suggesting that antibodies can work in concert to achieve broad neutralization. Still, Scheid said, the findings do not necessarily mean that the search for one “golden” broadly neutralizing antibody should be abandoned.

And indeed, that search is still on. Dennis Burton, a professor of immunology and molecular biology at the Scripps Research Institute, presented results of an effort by IAVI's AIDS Vaccine Design and Development Laboratory in Brooklyn, New York, and its Neutralizing Antibody Center, in La Jolla, California, which he heads, that led to the identification of two new broadly neutralizing antibodies. As part of its research study, known as protocol G, which seeks to identify new broadly neutralizing antibodies, IAVI screened sera collected from individuals who have been HIV infected for at least three years against panels of viruses. In collaboration with the company Spaltudaq, now called Theraclone Sciences, researchers then isolated IgG memory B cells from the most promising serum samples, made monoclonal antibodies from them, and screened the monoclonal antibodies for neutralization and binding to gp120 and gp41 Env proteins. In the study, antibodies that neutralized certain viruses often did not bind the gp120 protein of the same virus, suggesting that binding doesn’t necessarily preclude neutralization. The fact that the initial screen was for neutralization and not for gp120 binding might have contributed to the project’s success, Burton believes.

Two antibodies did particularly well, even compared with the four already identified broadly neutralizing antibodies. “There is great interest in these antibodies,” Burton said. “They don’t hit everything by any means, but they do hit a lot and what they hit is often very potent.” Burton said the two newly identified antibodies bind to a new epitope on the Env trimer, where there are perhaps fewer problems with accessibility than with some of the known broadly neutralizing antibodies that bind closer to the membrane, or are more sterically obstructed.

More work on antibodies was presented by Peter Kwong, chief of the structural biology section at the VRC. Kwong is trying to understand why antibodies that bind to the CD4 binding site on HIV gp120 do not necessarily neutralize the virus. He found that one such antibody called b13 binds the gp120 monomer at an angle about 15 degrees different from that of the broadly neutralizing antibody b12. This difference, though subtle, induces a conformational change in the gp120 monomer. This b13-induced conformation of the gp120 monomer is not easily compatible with the Env trimer, which means that it likely cannot be induced—or exist—in the context of the trimer, said Kwong. This explains why b13 doesn’t neutralize the virus, even though it binds the gp120 monomer.

So to be able to neutralize, he said, antibodies have to precisely recognize the vulnerable initial site of CD4 attachment on the Env trimer. “You have to be right on, and we have figured out the mechanism of why you have to be right on,” Kwong said, adding that to develop appropriate immunogens to elicit antibodies that effectively target this initial site of CD4 attachment, it’s important to understand what works, but also what doesn’t work. “If you only watch Tiger Woods, you have no idea how difficult it is to hit that ball just right,” he joked.

The studies presented by Scheid and Burton suggest to Kwong that there are two possibilities of how broad neutralization could be achieved: Either there are few broadly neutralizing antibodies that neutralize everything, or many antibodies with weak activity that work together. “Both possibilities might be true,” he said.

HIV-specific T cells in the gut

Barbara Shacklett, associate professor at the University of California in Davis, presented data on HIV-specific T-cell responses in the gut of HIV-infected individuals. Shacklett’s laboratory has developed expertise in the careful segregation of CD4+ and CD8+ T cells from gut tissue, which allows these populations to be analyzed in detail. In a study led by postdoctoral researcher April Ferre, Shacklett’s group used the approach to compare responses between individuals who control HIV replication in the absence of treatment and those with progressive infection, with or without treatment.

The study included 17 elite controllers (viral loads less than 75 copies/ml), 11 viremic controllers (viral loads between 75 and 2,000 copies/ml), 14 non-controllers (viral loads >10,000 copies/ml), and 10 individuals with undetectable viral loads on antiretroviral therapy (ART). Gag-specific CD8+ T-cell responses were evaluated in rectal mucosa and blood and Shacklett reported that, while there was no difference in blood, mucosal Gag-specific CD8+ T-cell responses were significantly higher in the controllers, both elite and viremic, than in non-controllers and individuals on ART. Similarly, Gag-specific CD8+ T cells expressing multiple cytokines/chemokines were significantly higher in the rectal mucosa of controllers.

A novel finding from this study was that the CD8+ T-cell responses among controllers were associated with particular class II HLA alleles, DRB1*13 and/or DQB1*06, which present antigens to CD4+ T cells. One or both of these alleles was present in 70% of elite controllers, 45% of viremic controllers, and 8% of non-controllers. In a related poster, Ferre showed that these alleles were also associated with the presence of stronger Gag-specific CD4+ T-cell responses in the gut of controllers. Shacklett’s group is now further exploring the role of class II HLA alleles among HIV controllers.

Non-pathogenic SIV infection

Several presentations focused on research into the mysteries of non-pathogenic SIV infection. Guido Silvestri, associate professor of pathology, microbiology, and immunology at the University of Pennsylvania, discussed sooty mangabeys, natural hosts of SIV viruses that are pathogenic in macaques and direct antecedents of HIV-2 in humans. Silvestri explained that the lack of disease progression in mangabeys is not due to control of viral replication—viral loads in these animals are as high or higher than those associated with disease progression in HIV-infected humans or SIV-infected macaques. Rather, what distinguishes non-pathogenic infection is the absence of persistent immune activation. Silvestri spoke about one hypothesis to explain this phenomenon—lack of IFN-γ production by dendritic cells, which is essentially a lack of an innate immune response to SIV.

To evaluate the merits of this, Silvestri designed a study to look at whether acute SIV infection of sooty mangabeys is associated with an absence of immune activation or if activation occurs but is then actively downmodulated. He analyzed gene expression in sooty mangabeys and macaques challenged with either SIVsm or SIVmac239.

Silvestri reported that acute infection of sooty mangabeys was associated with “massive changes” in the transcriptional profile of multiple genes, which were similar to those seen in the macaque groups. The gene expression profiles, however, diverged in chronic infection. Silvestri showed that IFN-stimulated genes were upregulated to the same degree or even higher in sooty mangabeys than in macaques during acute infection, but returned to baseline during chronic infection only in the mangabeys. Silvestri concluded that the data support a model in which there is active downmodulation of immune activation in sooty mangabeys during the transition from acute to chronic infection. Contrary to the idea that the immune system simply ignores the virus, Silvestri stressed that “the host is responding in a very vigorous way to infection.”

Silvestri also presented data indicating that CD4+ T cells express less CCR5, the key co-receptor for virus entry, in sooty mangabeys. This is not the case for CD8+ T cells, where CCR5 expression levels mirror humans. When mangabey CD4+ T cells are activated, CCR5 upregulation is also delayed compared to humans and other monkey species.

Battlefield maps

Ashley Haase, head of the Department of Microbiology at the University of Minnesota, provided an update on his studies of SIV pathogenesis. Of particular interest for vaccine research, Haase has been exploring the facets of an effective SIV-specific CD8+ T-cell response using detailed analyses of tissue samples. Here he introduced a visually compelling innovation he describes as a “battlefield map,” which involves overlaying images of tissue sections stained for viral RNA to identify SIV-infected cells, then stained for SIV-specific CD8+ T cells using a modified version of the tetramer assay. Haase has dubbed the method ISTH, to indicate the combination of in situ tetramer staining to identify the CD8+ T cells and in situ hybridization to locate and quantify virus-infected cells.

Haase shared battlefield maps from a study of female macaques challenged intravaginally with SIVmac239. The decline in SIV viral load from peak levels at around days 10-14, to day 21, was correlated with the detection of conjugates of SIV-specific CD8+ T cells and SIV-infected cells.

In addition to the images, Haase generated quantitative data by calculating effector-to-target (E:T) ratios. Using this technique, he found that the E:T ratio was correlated with the SIV viral load decline in acute infection. In all but one animal, the highest E:T ratios were attained in the cervical tissues, where initial exposure had taken place.

Haase also outlined results from a collaboration with immunologist Rafi Ahmed at Emory University using the murine lymphocytic choriomeningitis virus (LCMV) model. Their study involved two LCMV variants, the Armstrong strain and clone 13; the former only causes an acute infection, which the immune response rapidly clears, while the latter establishes a chronic infection. Haase used the ISTH technique to demonstrate that these differences are driven by the expanded target cell range of clone 13, which infects many more cells than the Armstrong strain. As a result, the E:T ratio is insufficient to contain clone 13. With the Armstrong strain, the number of effector cells quickly exceeds the number of infected cells and the virus is cleared.

Haase said the ultimate goal of this work is to gain an understanding of how many effector T cells need to be induced by an HIV vaccine to mediate viral clearance or control; in Haase’s words, to achieve “enough, and soon enough” (14).

HSV-2 Infection and HIV risk

The finding that infection with herpes simplex virus (HSV)-2 raises the risk of acquiring HIV by about two- to three-fold, dependent on sex and route of exposure (AIDS 20, 73, 2006), led directly to large randomized trials to evaluate the impact of HSV-2 suppression with acyclovir on HIV infection. The hope was that acyclovir, by suppressing HSV-2 reactivation, would decrease the incidence of HIV infection, but this strategy proved ineffective (see Clues from CROI, IAVI Report, Jan.-Feb. 2008).

Larry Corey, principal investigator of the HVTN, set out to try to understand why by conducting a study with 15 HSV-2 infected individuals, nine untreated and six receiving chronic suppressive therapy with acyclovir. He found that even in the presence of acyclovir, HSV-2 infection was associated with “very large nests of CD4+ T cells” under the dermis in areas of prior lesions—two- to 32-fold more CD4+ T cells compared to controls. The proportion of CCR5-expressing CD4+ T cells was also higher at these sites in 14 of the 15 study participants. These CD4+ T cells were largely HSV-2-specific and appeared to be engaged in active and effective immune surveillance, according to Corey.

He also reported that DC-SIGN-expressing dendritic cells were enriched in these samples, clustered with the CD4+ T cells. This indicates that while acyclovir can clearly prevent the incidence of symptomatic HSV-2 reactivation, it does not abrogate the need for local immune control of HSV-2. As a consequence, HSV-2 infection causes an increased mucosal presence of CCR5-expressing CD4+ T cells, which are optimal targets for HIV infection. Corey concluded that the ideal approach to reducing the affect of HSV-2 on HIV acquisition would be to prevent HSV-2 infection altogether.

1. Proc. Natl. Acad. Sci. 105, 7552, 2008
2. J. Virol. 83, 3556, 2009
3. Proc. Natl. Acad. Sci. 101, 13026, 2004
4.J. Virol. 83, 6318, 2009
5.Vaccine 25, 4085, 2007
6.Science 312, 1530, 2006
7. N. Engl. J. Med. 360, 1815, 2009
8.Nat. Med. 13, 100, 2007
9. J. Virol. doi:10.1128/JVI.00282-09
10. J. Virol. 82, 11181, 2008
11. Mucosal Immunol. 1, 219, 2008
12. J. Virol. 77, 3099, 2003
13. Nature 458, 636, 2009
14. Science 323, 1726, 2009