From 27-29 October 2002, about 200 scientists gathered in Annecy, France for the “13th Cent Gardes Symposium on HIV and AIDS Vaccines.” Meeting for the first time in this scenic Alpine town, the symposium still bears the name of its original home outside Paris—the historic building where Napoleon III once housed his elite troop of bodyguards, the “Cent Gardes,” and Louis Pasteur later maintained his animal laboratories.

Aside from its new location, the conference opening featured another novelty: a welcome letter to participants from former US President Bill Clinton, expressing his strong support for increased efforts in the fight against AIDS and applauding the commitment of scientists in leading the battle.

Its scientific sessions also began on an unusual note, turning the spotlight on renewed efforts to make vaccines that induce broadly neutralizing antibodies (NAbs)—following years of discouragement about the prospects for success, and a shift of focus onto vaccines targeting cellular immunity. “Neutralizing antibodies are coming out of the shadows, where they’ve been for the past several years,” said Susan Zolla-Pazner (New York University), who chaired the first NAb session. In addition, the conference featured updates on the widening pipeline of vaccine candidates and continuing studies on how to best induce cellular immunity, concluding with a sobering session on the enormous challenges still ahead in conducting large-scale trials in developing countries where the epidemic is raging.

The “comeback” of neutralizing antibodies

The renewed focus on antibodies comes at a time of growing awareness that HIV vaccines based on cellular immunity alone will, at best, protect against disease—not infection—and may work for only a limited duration. The hope is that vaccines which induce antibodies capable of neutralizing diverse primary HIV strains will go further, by blunting or completely preventing initial infection. But the field has long been at an impasse in designing such immunogens, with those developed so far inducing only NAbs of narrow strain specificity.

However, two developments in the past few years have fueled a sense that the problem is solvable and that NAbs may indeed bring some of the hoped-for benefits. First was the isolation of rare human antibodies (as monoclonal antibodies, or MAbs) that target conserved epitopes of the HIV envelope protein (Env) and show broad neutralizing ability, proving that the human immune system can produce such Abs. Second was the demonstration, initially by the groups of Ruth Ruprecht (Harvard Medical School, Boston) and John Mascola (Vaccine Research Center, Bethesda), that passive immunization of macaques with cocktails of these MAbs can protect against subsequent challenge with pathogenic SHIVs.

At Cent Gardes, the sessions on NAbs gave an excellent overview of how the field is attempting to build on these findings, and of the overall state-of-the-art.

Neutralizing antibodies in HIV infection

The talks began with a presentation by George Shaw (University of Alabama, Birmingham) on the role of NAbs in controlling HIV during infection, and at how virus manages to evade these immune responses. Describing a study that followed 20 people for two years from the time of acute infection, Shaw reported that all 20 made NAbs (detectable in the first patients at about 10 weeks) which were initially very effective in neutralizing the patients’ own (autologous) virus. But over time, the early viral population lost its sensitivity to neutralization and was eventually replaced by completely neutralization-resistant strains—suggesting that NAbs exert strong selective pressure during infection, and play an under-recognized role in suppressing viral replication.

Analyzing the sequence changes in Env to determine how this resistance arose, Shaw found that only 4 of the 15 most common mutations fell within known neutralizing epitopes—in contrast to HIV escape from cellular immune responses, which usually reflects mutations in epitopes targeted by the responding CD8 T-cells. Instead, 7 of the 15 mutations were at glycosylation sites—suggesting that HIV often escapes from antibodies by modifying its already-dense “glycan shield” to better obstruct NAb binding. Site-specific mutagenesis studies of Env supported this model: by changing amino acids known to bind carbohydrates, viral resistance to neutralization could be increased up to 100-fold.

Similar findings were presented by Doug Richman (VA San Diego Healthcare System), who also followed NAb responses and Env sequence evolution in infected people over time. In his study, 12/14 patients had substantial NAb titers to autologous virus starting at about week 8—but by one year virus had escaped from these NAbs, which showed poor recognition of HIV strains other than the initial one (even those from the same clade). But Richman closed on a more optimistic note, pointing out that HIV’s apparent ease in evading NAb responses might not extend to vaccine scenarios. In a vaccinated person, the immune system would have a head-start following exposure to HIV; without vaccine, it starts from behind and must play catch-up with the rapidly-replicating virus.

Passive immunization and protection by NAbs

Several speakers presented work looking more closely at passive protection of macaques by broadly neutralizing antibodies. Harvard’s Ruth Ruprecht reported that her group’s studies with cocktails of 3 or 4 neutralizing MAbs have shown protection in 22 out of 31 newborn monkeys challenged orally with pathogenic SHIV, leading her to suggest that the well-conserved epitopes targeted by these MAbs might be good candidates for inclusion in vaccines.

In a conceptually fascinating experiment, John Mascola presented a follow-up study to his earlier demonstration of passive protection via MAbs—although the results were ultimately inconclusive, given limitations on the numbers of monkeys available, and therefore on the different dosage variables that could be tested. The idea was to combine MAbs and cellular immunity, asking whether a sub-optimal MAb dose (which on its own protects some but not all animals) blocks infection completely in rhesus macaques previously immunized with a DNA vaccine known to protect against disease (derived from that made by Dan Barouch and Norman Letvin and carrying the same IL-2 and HIV-gag genes but a different env).

The experiment was done with 20 animals divided into four groups of five, given either: (1) DNA vaccine (4 immunizations) plus HIV-MAbs (2F5 and 2G12, at one-third the fully protective dose); (2) DNA vaccine plus an irrelevant MAb; (3) sham DNA plus HIV-MAbs; or (4) sham DNA plus irrelevant MAb (controls). One day after passive immunization, all animals were challenged vaginally with SHIV89.6PD. The findings: no difference in infection rates between the MAb-only and the vaccine-plus-MAb groups (both treatments protected 2/5 animals against infection), indicating that under these experimental conditions, cellular responses did not contribute to sterilizing immunity. All control and DNA-immunized animals became infected, with the latter showing robust CD8 responses and good control of both viremia and CD4 T-cell decline over 24 weeks.

But the study leaves open the larger question of whether some other combination of NAbs and cellular immunity to HIV might yield better overall protection (against either infection or disease progression) than candidates which induce only cellular responses. While a combination strategy is widely viewed as the most promising path to highly effective vaccines, the model is extremely difficult to test, says Mascola, given the existing shortages of monkeys and the large study that would be required.

Also addressing passive protection by antibodies, Malcolm Martin (National Institutes of Health, Bethesda) updated his lab’s work on analyzing how much antibody, and what regimen, is needed to confer sterilizing immunity. His model uses polyclonal IgG from HIV-infected chimps (rather than MAbs) for passive immunization of pigtail macaques, and a high-dose, intravenous challenge one day later with SHIV-DH12, which carries the same Env protein as the HIV strain infecting his chimps. With this system, his group calculated that a fully protective dose corresponded to 149 mg IgG per kg body weight, or an antibody titer of 1:38 (the last dilution that completely blocks HIV infection of cultured cells—a more stringent criteria than that used by most HIV labs) (J. Virol. 76: 2123;2002). He also reported that IgG given six hours after challenge still blocks infection—but that by 24 hours, it’s too late, although these animals controlled virus (viral RNA was first detected at week 5).

Speaking afterwards with the IAVI Report, Martin said that the antibody levels required to achieve sterilizing immunity in his model (with its high challenge dose and stringent route) are completely “within the range” of those attainable by vaccination, citing the example of a vaccinia-gp160 plus gp120 prime-boost regime in monkeys (J. Virol. 75:2224;2001), and are “routinely generated” in natural infection (J. Virol. 74:6935;2000). To his mind, the bigger difficulties in eliciting protective NAb by vaccination will be to get sufficiently broad responses, and whether high enough levels can be recalled fast enough when a vaccinated person is exposed to HIV.

Defining immunogens that induce broad NAbs

Turning to the question of what antigens might induce potent, broad NAbs, talks ranged from studies of defined domains within Env to rational approaches towards designing new immunogens.

Focusing on a well-known antigen, Susan Zolla-Pazner argued for a new look at the long-dismissed, hypervariable V3 loop of Env as a potentially useful immunogen. While early studies in animals detected mostly type-specific NAbs to V3, antisera from HIV-infected people usually show broad anti-V3 NAbs (although these are often poor at neutralizing primary HIV strains). In re-examining V3 as a vaccine antigen, she began from the premise that it is involved in gp120 binding to CD4—which is not clade-restricted—and must therefore have some conserved features which could be exploited for vaccines. Close analysis revealed several well-conserved structural properties despite V3’s hypervariable protein sequence, suggesting that NAbs which recognize 3D conformation rather than linear peptides might be more potent neutralizers of diverse primary HIV strains.

To test this idea, her group derived MAb-producing cell lines from people infected with HIV clade B or A, using a procedure to select for V3 MAbs that recognize conformational rather than linear epitopes. Data from 13 lines suggest that these MAbs do show binding and neutralization across HIV-1 clades, but not all—suggesting that V3 may exist in more than one immunologically-relevant shape.

In a very different approach he calls “reverse vaccinology,” Dennis Burton (Scripps Research Institute, La Jolla) described his team’s efforts to elucidate at the molecular level why certain MAbs are broadly neutralizing—by defining the epitopes they recognize and the features of those epitopes which are essential for MAb binding. The goal is to use this information as a guide in designing immunogens that might induce similar antibodies. Burton described work on two MAbs. The first, called b12, recognizes a CD4-binding domain within Env, and detailed studies of their binding have revealed the regions, 3D shapes and probable mechanisms involved. These results, in turn, paved the way to use site-specific mutagenesis to create a gp120 immunogen that binds b12, but not epitopes for other antibodies that might divert the immune response away from the b12 epitope. Another broadly neutralizing MAb called 2G12 works in a completely different way, recognizing an epitope made of sugar residues attached to specific amino acids within Env. Structural analysis and modeling studies are being used to design an immunogen that presents sugar residues in the same way.

In terms of specific vaccine candidates targeting humoral immunity, there is little in the clinical trials pipeline other than VaxGen’s gp120-based vaccines and the new GlaxoSmithKline Env-containing product. Two speakers presented new immunogens in pre-clinical development—George Lewis (Institute for Human Virology, Baltimore), who described gp120-CD4 complexes (IAVI Report, Jul-Sep 2002, p. 15) and Susan Barnett (Chiron Corp, Emeryville), who discussed Chiron’s native gp140 molecule (IAVI Report, Apr-Jun 2000, p.6). Updates on these and other NAb-inducing immunogens will be included in our reports on conferences in early 2003.

Pulling together the many different threads on NAbs, the overall sense at Cent Gardes was of a field facing enormous scientific challenges—but gaining a more solid foundation for moving forward. “We’re still just teasing out a conceptual understanding of why it’s so difficult to induce broad, potent NAbs, and of how to design the right immunogens,” said John Mascola. “But I think it should eventually be feasible, based on the fact that potent MAbs do exist in humans.”

VSV-based vaccine candidates

Many of the remaining talks at the meeting focused on the more familiar terrain of HIV vaccine strategies that target cellular immunity—an area where questions are much more sharply focused on evaluating specific vectors, antigens, immunization regimes and routes, both in monkeys and clinical trials.

One of the newer candidates in this category is the VSV-based candidate described by John Rose of Yale University (New Haven). VSV is a virus that infects livestock, where it a causes self-limiting infection but is not pathogenic in humans. Building on published protection data in macaques, Rose presented encouraging results from intranasal vaccination, along with studies comparing two different types of boosts.

Last year, Rose’s team showed that a vaccine containing HIV gag and env in an attenuated VSV vector protected 7/7 monkeys against disease after challenge with SHIV89.6P, given 3 or 6 months after the last boost (Cell 106:539;2001). (The challenge and vaccine strains differed at 14 positions in Env.) All 8 control animals became infected and progressed to AIDS. After two years or more of follow-up, the 7 vaccinated animals remain healthy and continue to control virus: 6/7 have undetectable loads, while the seventh maintains a stable load of 4,000.

At Cent Gardes, Rose presented a follow-up study done at Wyeth Vaccines, which has licensed the VSV platform and is developing the HIV vaccines further. The 3-arm study looked at animals immunized intranasally (i.n.) or intramuscularly (i.m.) with VSV carrying env, gag and pol (3 animals per group), plus 4 controls. All monkeys were challenged 5 weeks later with the homologous SHIV89.6 strain.

Using three different assays for measuring HIV-specific CD8 T-cells (tetramer staining, Elispot, CTL killing), the researchers found that i.n. immunization consistently evoked stronger CD8 responses (about 3-fold by tetramer staining). NAb levels were similar in the i.n. and i.m. groups. But despite differences in pre-challenge CD8 responses, protection was similar in the two groups: all animals achieved undetectable viral loads and preserved their CD4 counts. However, clear vaccine-induced protection could not be shown in this study, since 3 of the 4 control animals also suppressed viremia, although not CD4 decline. (The vaccinated animals, but not controls, were all MamuA*01.)

Last, Rose described an ongoing monkey study suggesting that it is more effective to boost VSV-primed animals with a different vector. The study compares two groups of 4 animals, one boosted i.m. with a second VSV-based vaccine (carrying a glycoprotein from a different serotype but otherwise identical to the prime) and the other, an intradermal (i.d.) boost with MVA carrying the same HIV genes. The MVA boost led to higher levels of HIV-specific CD8 cells than VSV (an average of 1.5% vs. 0.5% tetramer-positive cells, and these remained elevated five weeks later—by which time the VSV-boosted response had completely waned.

In the next phase of this NIAID-supported “vaccine development team,” Wyeth is preparing for Phase I trials through the HIV Vaccine Trials Network (HVTN). Since the VSV vector replicates in humans, livestock and the cell cultures of many mammalian species, this will require extensive testing of its tissue distribution in vaccinated animals, as well as further host range and virulence testing, according to Rose. Later trials “would probably also involve a larger study in herd animals,” he said. Preliminary data showed no significant pathology in 4 vaccinated cows (although the animals initially developed small lesions at the inoculation site) and no discernible pathology in 42 macaques vaccinated with VSV vectors.

Clinical trials

Numerous vaccine candidates now in clinical trials in the US and Europe also featured heavily on the Cent Gardes agenda. These included the Merck vaccines, canarypox, GlaxoSmithKline’s NefTat/gp120 combination and the lipopeptides developed as boosts through the French ANRS, all of which have been covered in recent issues of the IAVI Report. Gary Nabel (Vaccine Research Center, Bethesda) discussed a multi-clade DNA vaccine that just entered Phase I studies (see Vaccine Briefs, Multi-Clade Trial Begins). Another was described by Mary Marovich of the US Military HIV Research Program (Rockville), who presented a still-blinded study testing whether volunteers’ own dendritic cells “loaded” with a canarypox-based vaccine construct (ALVAC vCP205) are more immunogenic than standard i.d. or i.m. immunization with the same construct. (For a complete list of ongoing vaccine trials, see IAVI.org).

In the meeting’s final session, speakers addressed the complexities of preparing and conducting clinical trials in developing countries (see IAVI Report, Jul-Sep 2002, p.2). Glenda Gray (Chris Hani Baragwanath Hospital, Soweto) spoke about the regulatory, ethical and logistical challenges facing South Africa’s vaccine trial preparedness efforts, and highlighted issues arising from the fact that youth, especially young women, are at such high risk for HIV infection (see also interview, p. 8).

Gray said that about 26% of South African women who visited antenatal clinics in 2001 were HIV-infected. She also presented hard-hitting new data estimating that about 50% of South African girls are sexually active by age 15, and 10% by age 12—with coercion, violence and poverty as major factors behind these numbers (data from the Kaiser Family Foundation and South African National Youth Survey). To Gray, these findings underscore the need to greatly intensify prevention efforts in very young age groups, and to resolve issues (such as informed consent) hindering the inclusion of adolescents in vaccine cohorts . She also presented survey data from Soweto residents showing that 68% of the respondents said they were “definitely willing” to participate in a vaccine trial; 16% were “definitely not” willing, and the remainder were undecided.

Echoing many of these same themes, Pontiano Kaleebu (Uganda Virus Research Institute, Entebbe) described the strides his country has made since conducting Africa’s first (NIH-sponsored) HIV vaccine trial in 1999. Through that experience, Uganda has established clear procedures for scientific, ethical and legal review of trial protocols, as well as good laboratory capacity. As it prepares to launch a DNA/MVA study in 2003, media, local community and the general public also have more knowledge, and are more accepting, of HIV vaccine trials.

------------
^*Patricia Kahn is editor of the IAVI Report.