Clues from CROI

The 15th Annual meeting had its share of sobering news from clinical trials, but researchers also reported many exciting advances

By Andreas von Bubnoff

“Triumph and defeat, success and failure—that’s the theme for this conference.” At least that is how conference vice chair John Mellors of the University of Pittsburgh summed up the 15th Conference on Retroviruses and Opportunistic Infections (CROI) held in Boston February 3-6.

Mario Stevenson of the University of Massachusetts Medical School and conference chair said that the mood coming into this year’s meeting was indeed somber, following the disappointing results of the STEP trial with Merck’s vaccine candidate. This news had an obvious influence on the conference program, which was more focused on AIDS vaccine research and HIV prevention strategies than in previous years. Additional data from multivariate analyses of the STEP trial were presented, as well as news from studies with other HIV prevention methods, including data from additional circumcision studies and the first results from a large randomized clinical trial evaluating whether or not suppression of herpes-simplex virus (HSV)-2 could reduce the risk of HIV infection. There was also a special plenary session devoted to the scientific obstacles that are impeding the discovery of an effective AIDS vaccine.

Stevenson said CROI also showcased “tremendous progress in the basic sciences.” There were a slew of presentations highlighting advances in immunology and biology that open potential avenues of exploration for both HIV prevention and treatment, including the identification of a third cellular antiviral factor, the discovery of a peptide in semen that aids transmission of HIV, as well as a myriad of other fundamental discoveries (seeScanning the scientific horizon).

One STEP forward…

Susan Buchbinder of the University of California in San Francisco, and principal investigator of the STEP trial, presented additional results based on multivariate analyses of data from this Phase IIb test-of-concept trial of Merck’s adenovirus serotype 5 (Ad5)-based vaccine candidate, known as MRKAd5. Immunizations were stopped on September 21 last year after an interim analysis by the data safety monitoring board determined that the vaccine did not offer any protection against HIV infection or reduce viral load in individuals who acquired HIV infection. Subsequent analysis of the data indicated there was a trend toward more infections in the vaccinees than in placebo recipients (see A STEP back?, IAVI Report, September-December 2007).

The first analysis of male volunteers—women were excluded because only one HIV infection occurred in a female volunteer—revealed that vaccinees with higher levels of natural pre-existing Ad5 immunity appeared more susceptible to HIV infection than those with lower Ad5 immunity. The multivariate analyses Buchbinder presented at CROI showed that after adjustment for possible confounding factors such as age, ethnicity, location, and sexual behavior, the apparent increase in susceptibility to HIV infection still appeared to be due to vaccination. “This is still a hypothesis,” Buchbinder cautioned, adding that additional factors needed to be tested such as infection with HSV-2, human leukocyte antigen (HLA) typing, and the possibility of clusters of sexual transmission at certain sites, all of which could have contributed to a spurious result. But, Buchbinder said, “At this point we don’t see any other explanations [apart from the vaccine] for the increased number of infections in the vaccine group.”

Researchers conduct multivariate analyses to analyze several variables simultaneously, while correcting for confounding factors. This type of statistical analysis is used to calculate the potential contribution of each variable on the observed outcome. In this case, researchers used an increasingly complex set of statistical models, each accounting for more variables. Among the vaccinees, the multivariate analyses identified circumcision status as an additional risk factor—vaccinees were over four times more likely than placebo recipients to acquire HIV if they were both uncircumcised and Ad5 seropositive. Buchbinder said it is unclear whether or not the risk of HIV infection was increased for vaccine recipients who had either high levels of Ad5 immunity or were uncircumcised, but not both. In these analyses there was no evidence of enhanced risk of HIV infection in vaccinated volunteers who were circumcised and Ad5 seronegative.

Buchbinder said that the circumcision effect on HIV infection risk seemed to be as strong, if not stronger, than the effect of pre-existing Ad5 immunity. “The interaction for Ad5 was not significant in all the models, but circumcision was significant in most of the models and was of a greater magnitude,” Buchbinder said, referring to the different statistical models used in the analyses. But interpretation of these results is complicated because volunteers were not originally separated based on circumcision status. “What it means is hard to say because we stratified based on Ad5. The circumcision is really a post-hoc analysis and so I think you have to take all of this with a grain of salt.”

Still, she said, “It is possible that this is really from circumcision and not from pre-existing Ad5 immunity.” More than half of the 1,787 men in the analysis were circumcised and 49 were excluded from the multivariate analyses because their circumcision status was unknown. The role of circumcision in men who have sex with men—the population in the STEP trial—is unclear and is obviously dependent on whether or not the individuals were the receptive or insertive partner during anal sex. “If you are the receptive partner, obviously your foreskin doesn’t matter,” said Michael Robertson of Merck. “We are looking at those kinds of associations as well, but right now we don’t have a clear understanding.”

Robertson presented results of an ongoing characterization of the immune responses induced by MRKAd5 in STEP trial volunteers. Analysis of peripheral blood mononuclear cells (PBMCs) collected at week 30, four weeks after the second vaccination, showed no difference in the potency or breadth of immune responses to MRKAd5 in the volunteers who later became infected with HIV, compared with those who remained uninfected. “It was not [that] the people that became infected were just non-responders to the vaccine,” Robertson said.

One hypothesis to explain the trend in higher susceptibility to HIV infection among vaccine recipients is that the vaccine might have induced activated T cells, either in response to the vector or the HIV immunogens, and therefore provided additional targets for HIV. But for now this hypothesis does not seem to be supported by the data. “It doesn’t match what we observed in the study,” said Robertson, because T-cell responses to the Ad5 vector, as well as to the HIV inserts were higher in people with the lower pre-existing Ad5 immunity.

Robertson said there is, however, an important caveat. “We are only looking in the peripheral blood,” he said. “Maybe you are activating cells that don’t stay in the blood but traffic to mucosal sites. Maybe we don’t see an association because we are looking in the wrong place.” He said one biologically plausible hypothesis is that activated cells move to mucosal sites such as the foreskin. “So if you are circumcised you have less targets to get infected,” added Robertson. “It’s biologically plausible but pure speculation at this point.”

Robertson said it would be of interest to check immune responses at mucosal sites, but there are only a few mucosal samples from STEP trial volunteers, such as semen samples, which were collected from individuals after infection. “We don’t have samples just after vaccination in uninfected people,” Robertson said. So for now, there is still no clear explanation for why there was an increased risk of HIV infection in people with high pre-existing Ad5 immunity, Robertson said. “I think we have more questions than answers.”

Extended benefits of circumcision

In heterosexual men, the protective effect of circumcision has been firmly established by three randomized, controlled clinical trials, which show that surgical removal of the foreskin reduces the risk of HIV acquisition by approximately 60% (see Vaccine Briefs). Additional studies were presented at CROI that looked at the effect of male circumcision on the acquisition of HSV-2, on transmission of HIV from infected men to their female partners, and on acquisition of sexually-transmitted infections (STIs) in female partners.

Aaron Tobian of Johns Hopkins University reported results from a randomized trial in Rakai, Uganda, that enrolled over 3,500 uncircumcised men who were both HIV and HSV-2 uninfected. Half of them were randomly assigned to immediate circumcision, while the other half were offered circumcision at the conclusion of the trial. After two years of follow up, researchers observed that circumcision reduced the risk of acquiring HSV-2 infection by almost 25%. Several observational studies have supported the role of HSV-2 infection in aiding HIV transmission. “This might be part of the reason male circumcision decreases HIV acquisition,” Tobian said.

More than 1,600 HIV-uninfected female partners of men in the first trial were also followed for 12 months to determine the effect male circumcision had on the acquisition of vaginal diseases. Researchers found that the female partners of circumcised men had an approximately 25% lower risk of acquiring genital ulcer disease, an almost 50% lower risk of acquiring Trichomonas vaginalis, and an almost 20% lower risk of acquiring bacterial vaginosis, compared with female partners of uncircumcised men. These results are consistent with previous observational studies, Tobian said. A possible explanation is that in uncircumcised men, the moist cavity created by the foreskin may favor the survival of bacteria.

Knowing the effects of circumcision in HIV-infected men is also important, according to Maria Wawer of Johns Hopkins University. She says that in some communities being uncircumcised may stigmatize men as being HIV infected. To avoid this, some HIV-infected men may seek circumcision. Wawer reported results from another randomized trial in Rakai, Uganda that enrolled discordant couples—HIV-infected men with uninfected female partners—to study the effects of circumcision in HIV-infected men and on HIV transmission rates to women. In this trial, 93 couples were enrolled in the intervention arm, in which the male received immediate circumcision, and 68 couples were enrolled in the control arm. Men in the control group were offered circumcision after the trial was complete. 

After two years of follow up, researchers found that circumcision did offer some benefit to HIV-infected men—rates of genital ulcers were reduced by about 50% in circumcised trial participants compared to uncircumcised controls. But circumcision had no effect on HIV transmission rates to the female partners of HIV-infected, circumcised participants. This result was “unexpected and disappointing,” Wawer said. “In previous observational data we had seen lower HIV rates in women married to HIV-positive circumcised men compared with HIV-positive uncircumcised men.”

Researchers suggested the women were not protected because some of the couples resumed sex too soon after surgery. “If the males resume intercourse early after circumcision, before the wound is fully healed, there might be increased transmission,” said Wawer. In the 18 couples who reported resuming sex before the wound healed completely, 27% of the female partners were infected with HIV in the first six months of the study, compared to only 9.5% of female partners who were infected after waiting to resume sexual activity. Wawer said it is very important that people not resume sex in the early postoperative period, even if they are not HIV infected. “In our trial of negative men the protective effects of circumcision became significant and apparent only after the six-month follow-up period,” she said, referring to an earlier trial of male circumcision (Lancet 369, 657, 2007).

Stopping HSV doesn’t stop HIV

More sobering news on the HIV prevention front came from Connie Celum of the University of Washington in Seattle, who reported the results of a randomized clinical trial (HPTN 039) assessing whether or not administering suppressive therapy for HSV-2 could reduce HIV acquisition rates.

The trial enrolled 3,277 volunteers who were infected with HSV-2 but not HIV. Volunteers included men who have sex with men in the US and Peru and heterosexual women at sites in Zimbabwe, Zambia, and South Africa. All participants randomized to the intervention group received a twice-daily 400 mg dose of acyclovir, an antiviral drug used in the treatment of HSV-2 infection, which is the cause of genital herpes. After 18 months, there was no difference in the number of new HIV infections between the group that received acyclovir and the control group. Several observational studies have shown that HSV-2 infection increases susceptibility to HIV by two- to three-fold, so this was an unanticipated result. “Many people thought this would be a slam dunk,” said Celum. Taking acyclovir did reduce the incidence of genital ulcers by 37%, but even this was much lower than what was observed in previous studies.

“Why didn’t we have an effect on HIV at all?” Celum asked. She thinks it is unlikely that HSV-2 is not a risk factor for HIV, given the epidemiological data that suggests otherwise. Adherence is one factor that could have influenced the results. Reported levels of adherence to acyclovir were high in the study, but Celum said it could have been overestimated since it was based on self-reported behavior. The less-than-expected reduction in occurrence of genital ulcers also varied geographically, suggesting biological reasons may account for why researchers did not see any effect on HIV infection rates, Celum said. For example, there may be differences in how the drug was metabolized or in susceptibility of the virus to the drug.

Tethering the virus

A basic science finding that caused a stir was the recent discovery by Paul Bieniasz’ group at Rockefeller University, and independently by John Guatelli and colleagues at the University of California in San Diego, of the existence of another naturally-occurring antiviral factor in human cells (Nature 451, 425, 2008). This newly-discovered molecule named tetherin is the third cellular antiviral factor that has been identified to date, in addition to TRIM5α and APOBEC3G (see Guardian of the genome, IAVI Report, April-June 2005). As the name suggests, tetherin keeps newly made HIV particles tethered to the surface of an infected cell after budding. “The cell behaves altruistically,” Bieniasz said. “It itself is doomed, but other parts of the body might be saved by this act of the cell to retain the virus to its surface.”

Researchers have known for a while that HIV lacking its accessory protein Vpu remained stuck to the surface when infecting a certain subset of cells dubbed non-permissive, but had no problem escaping other, so-called permissive cells. While it was unclear what kept the HIV particles tethered to the membrane of non-permissive cells, Bieniasz and colleagues had shown that this property could be induced by treating permissive cells with IFN-α.

“This gave us a handle to try to identify genes that might be responsible,” Bieniasz said. His lab conducted a microarray experiment that compared expression levels of IFN-α-induced genes in permissive and non-permissive cells and looked for a cell surface or secreted protein. They identified tetherin, a membrane protein that co-localizes with accumulating virus particles on the cell surface, as the molecule responsible for the non-permissive phenotype. “[This is] a brand new mechanism which many of us never suspected before,” said John Coffin of Tufts University, who was not part of the study. Researchers have not yet figured out exactly how tetherin actually performs this function or how Vpu counteracts its effect, Bieniasz said, but by inactivating Vpu, it may be possible to mobilize this defense against HIV. “Hopefully pharmaceutical companies will take this on,” Bieniasz added.

Enhancer protein

Another study described the discovery of natural compounds, derived from human body fluids, that inhibit or enhance HIV infection. Frank Kirchhoff of the University of Ulm, in collaboration with Wolf-Georg Forssmann of Hannover Medical School, have established peptide libraries from human body fluids, which have led to the discovery of a natural enhancer of HIV infection.

Kirchhoff’s lab isolated an array of peptides from human semen, a logical place to look for factors affecting the efficiency of HIV infection, Kirchhoff said. From this peptide library, they identified an abundant peptide called prostatic acidic phosphatase (PAP). Initial experiments failed to show any effect of PAP fragments on HIV infection, but after an overnight incubation period, it was discovered that PAP could enhance HIV infection. “At some point we left it overnight and did the experiment the next day and it turned out it was actually becoming active,” said Kirchhoff.

His group found that PAP fragments form amyloid-like fibrils that can capture HIV virions and promote their attachment to target cells, dramatically enhancing HIV infection in cultured cells (Cell 131, 1059, 2007). “When you incubate the fibrils with virus particles,” Kirchhoff said, “they all bind after a few minutes, probably by just electrostatic interaction.” This binding is independent of the sequence of HIV’s Env protein. They also bind to the target cells with high efficiency, he said. The fibrils enhanced the infection rate of diluted HIV added to cultured PBMCs by 100,000-fold. “To be honest I did not believe the data, so we repeated this over and over and the results were pretty consistent,” added Kirchhoff. The fibrils also enhanced HIV infection in relevant human samples such as ex vivo cervical tissue, and in vivo in transgenic rats expressing human transgenes.

Kirchhoff said PAP could be a new target for blocking HIV transmission. “Given that the viral dose during sexual intercourse is normally sub-infectious, blocking an enhancing activity should have a direct impact on the rate of HIV transmission,” said Kirchhoff. He also suggested that a 10- to 20-fold higher dose of microbicides might be required to overcome the fibroid activity of PAP. “When we test microbicides with or without semen and the enhancer we need much higher doses to block the virus,” Kirchhoff said. The fibrils may also enhance other types of infections; his group has already observed this with HSV, hepatitis C virus, and Chlamydia.

Pleas for a Back-to-Basics Approach 

In 2004, after VaxGen’s AIDS vaccine candidate failed in a Phase III clinical trial, CROI featured a session titled, “HIV Vaccines: Clinical Trial Results Call for More Basic Science.” The session closed with a talk by Ron Desrosiers of the New England AIDS Research Primate Center: “Why an HIV-1 vaccine is not currently within our grasp.” Four years later, the failure of Merck’s Ad5-based candidate provided the conference and Desrosiers with an opportunity to revisit this topic. His message was followed by a presentation from Neal Nathanson of the University of Pennsylvania titled, “AIDS Vaccine at the Crossroads.” Mario Stevenson, the conference chair, said both speakers were invited to discuss going “back to basics” in vaccine research because “the CROI committee wanted to deliver this message in stereo.”

Desrosiers started by asking whether an HIV vaccine is feasible and, as was the case four years ago, he thinks the answer is no. He stated that the naturally occurring immune response to HIV typically fails to control viral replication and does not reliably protect against superinfection with other HIV strains. Desrosiers also cited the daunting diversity of HIV and emphasized that what constitutes a protective immune response against HIV remains unknown.

Researchers, including Desrosiers, have tried to gain information on what type of immune responses are protective by conducting challenge studies in nonhuman primates with simian immunodeficiency virus (SIV). Desrosiers said the best results obtained so far have involved meager viral load reductions of 1-1.5 logs in rhesus macaques challenged with SIV—even under idealized circumstances in which the vaccine and challenge strains are genetically matched. A macaque study with Merck’s Ad5 construct encoding SIV Gag failed to protect any animal from infection with SIVmac239 and also had no effect on post-infection viral loads (J. Virol. 79, 15547, 2005). Based on this now infamous study, Desrosiers declared that the failure of MRKAd5 was predictable. He also believes that no other T cell-based vaccine candidate currently being investigated has a “reasonable chance” of demonstrating efficacy.

The negative results of the STEP trial may also erode funding for both clinical trials and basic science, according to Derosiers. He also said potential volunteers may be alienated and unwilling to enroll in future trials if they hear news of more unsuccessful candidates. Although success of a particular approach cannot be fully assessed without conducting some type of efficacy trial in human volunteers, Desrosiers advised the field to be more selective about which candidates to evaluate in clinical trials and advance only those that have shown significant promise in the SIV model.

Saving perhaps the most controversial aspect of his talk for last, Desrosiers asked, “Has the NIH lost its way?” He explained that focusing on clinical development is a departure from the NIH’s traditional role of supporting basic science and leaving industry to explore promising leads in costly clinical trials. Desrosiers acknowledged that so far the pharmaceutical industry has been reluctant to focus on AIDS vaccine development but he claimed, “When an HIV vaccine does become feasible, pharma will jump on product development, manufacturing, and clinical testing.” To emphasize what he believes are the skewed priorities of the current NIH vaccine portfolio, Desrosiers showed a slide indicating that 33% of NIH funding supports clinical trials. In 2007, he said, the NIH spent US$189 million on AIDS vaccine clinical trials.

Desrosiers asserted that instead funding should support efforts to understand the induction of broadly neutralizing antibodies against HIV, the mechanism of protection with live attenuated SIV vaccines in nonhuman primates, and the causes of long-term nonprogression in humans. He also urged comparative testing of vaccine candidates in the SIV/macaque model and consideration of novel ideas, such as the hybrid vaccine/gene therapy approach developed by Phil Johnson of the Children’s Hospital of Philadelphia that uses an adeno-associated virus (AAV) vector to persistently expresses monoclonal neutralizing antibodies. Desrosiers described the as-yet-unpublished results of Johnson’s SIV challenge studies as “dynamite.”

Nathanson followed by echoing Desrosiers’s call for more basic discovery, starting with the quest to identify neutralizing antibodies against HIV. Nathanson’s hypothesis is that there may be a finite number of sites in HIV’s envelope that can mutate to avoid antibodies targeting the CD4 binding site. He therefore suggested that a basic research priority should be answering the question of whether all possible mutations in the CD4 binding site of HIV’s envelope can be characterized and targeted by a multivalent immunogen.

Shifting gears to monkey models, Nathanson diverged somewhat from Desrosiers’ view that SIV infection of macaques is the sin qua non for vaccine research. Instead, he stressed the importance of evaluating multiple different challenge models, with varying levels of stringency (heterologous versus homologous challenges using a variety of SIV strains and hybrid SIV/HIV isolates known as SHIV) to obtain a comprehensive view of a vaccine candidate’s potential. Nathanson closed by concurring with Desrosiers and suggesting that the NIH’s vaccine research portfolio needs to be reassessed. The format for this CROI session did not allow time for questions or comments at the end of the talks by Desrosiers and Nathanson. —Richard Jefferys