HIV vaccine development: A long and winding road

Within just two years of sentinel reports in 1981 of AIDS in gay men, Robert Gallo and Luc Montanier, separately, made the seminal discovery of the causative agent, a new retrovirus, subsequently named HIV.

Work began almost immediately on large-scale production of HIV in cell cultures, which then enabled low cost, highly sensitive and specific diagnostic tests to secure the blood supply, plot the course of the epidemic and define the immunology of infection. Add to this rapid progress in elucidating HIV’s molecular structure and sequencing its essential genes, and translational science seemed to be on a roll, leading inevitably to a protective vaccine. Indeed, in 1984, a top United States health official famously predicted an epidemic-ending vaccine within 18 months.

The hope at the time was that HIV vaccine development would repeat the remarkable recent successes of vaccines against the epidemiologically similar hepatitis B virus, in which a variety of surface antigen (HBsAg) preparations, when injected intramuscularly in multidose series, elicited better than 95% anti-HBs seroconversion and durable protection. There were no hepatitis B vaccine efficacy trial failures.

Although an inactivated whole HIV vaccine might be the most direct approach to inducing a broadly protective immune response, the unknowable risks of vaccine failure ending in HIV infection caused efforts to shift to engineering recombinant HIV envelope glycoproteins (eg, rgp120 and rgp160), the most promising targets for neutralizing antibodies. The concept here was to destroy the invading HIV on the mucosal “beachhead,” blocking viral entry into susceptible host immunocytes.

This turned out to be the easy part, as a number of biotechnology companies and other research institutions soon produced prototype rgp120 and rgp160 envelope subunit vaccines for further testing. Although most envelope vaccines were antigenic in humans and other animal models and elicited vigorous binding antibody responses to laboratory adapted strains, especially those strains from which the vaccines were made, by 1994, there was increasing evidence that binding antibody contained little or no true neutralizing activity against primary isolates from new HIV-1 infections.

Nonetheless, there also were laboratory data and limited chimpanzee studies to support possible efficacy and vaccine clinical trial activists, led by Donald Francis, to conclude that the critical questions about envelope subunit vaccines could not be definitively answered in the laboratory in the foreseeable future. They, therefore, elected to move Genentech’s rgp120 to a new company (VaxGen) for the sole purpose of conducting double blind, randomized, placebo controlled efficacy trials in humans, and if successful, to begin commercial vaccine production.

The first study (VAX004) enrolled 5,108 high-risk men who have sex with men in 61 clinics in the United States, Canada and the Netherlands, between June 1998 and October 1999. Despite inducing neutralizing and CD4-blocking antibody in all assessed vaccinees, an intramuscular seven-dose series of 300 µg each of MN and GNE 8 gp120/ HIV–1 (B/B) was ineffective in preventing HIV-1 infection or in modifying post-infectious markers of disease progression. Likewise, a companion, double blind, placebo-controlled efficacy trial of bivalent (B/E) gp120 HIV-1 vaccine among injecting drug users (IDU) in Bangkok, Thailand, showed no efficacy. The conclusion for both studies was that these gp120 HIV-1 vaccines failed to protect because they failed to induce broadly reactive anti-envelope antibodies capable of neutralizing genetically diverse primary isolates.

Throughout the humoral immunity research years, complementary work was proceeding on vaccines to activate the cellular arm of the immune system against cells expressing key HIV-1 gene products. Strategies included vaccines using naked DNA, plasmid DNA and recombinant viral vectors, most importantly canarypox (eg, the ALVAC series, Sanofi-Aventis) and replication defective adenovirus serotype 5 (rAd5).

Although there never has been any evidence in humans or non-human primates that cell-mediated immunity (CMI) responses alone can prevent HIV infection, there is a substantial body of research indicating that CMI can control HIV replication and disease progression in human long-term non-progressors and in non-human primate challenge models. Of the many CMI vaccines studied in phase 1 clinical trials, adenovirus type 5 (Ad5) recombinant vector-based vaccines proved to be the most immunogenic and a simian immunodeficiency virus Ad5 protype controlled viremia in some simian challenge experiments.

This experience provided the justification to undertake a double-blind, phase 2, test-of-concept, multisite study (the Step Study) of three injections of Merck’s MRK Ad5 HIV-1 gag/pol/nef vaccine (n=1,494) or placebo (n= 1,506) in predominantly MSM volunteers. Enrollment began in December 2004, but further immunizations were stopped in September 2007 when the Data Safety Monitoring Board determined that the study unexpectedly met prespecified futility boundaries.

Among the findings: the HIV infection rate in MRK Ad5 vaccine recipients was actually higher than in placebo recipients but never adequately explained; the vaccine elicited interferon-γ ELISPOT responses in 75% of a 25% random sample but apparently without benefit; and the vaccine did not reduce early HIV-1 plasma levels (4.61 vs. 4.41 log₁₀ copies per mL). A chronic challenge when evaluating CMI vaccines is how to measure specific functional cytotoxic immune responses. Interferon-γ ELISPOT assays on cryopreserved peripheral blood mononuclear cells, the current standard, are positive in less than 75% of vaccine recipients, use arbitrary cut-off points, tend to decline with time, and often do not correlate well with HIV control.

Despite the many failures of recombinant canarypox vector or rgp120 subunit vaccines to demonstrate meaningful CMI or neutralizing antibody responses or to protect against HIV infection or disease progression, in September 2003, the U.S. Military HIV Research Program, the Thai Ministry of Public Health, the NIH, and two vaccine companies elected to launch the most ambitious HIV vaccine efficacy trial yet. In a community-based, randomized, double blind placebo-controlled trial, 16,402 healthy, primarily heterosexual, men and women, ages 18 to 30 years, in two Thailand provinces received either vaccine (four priming injections of a recombinant canarypox vector [ALVAC-HIV vcp 1521] plus two booster injections of rgp 120 subunit vaccine [AIDSVAX B/E]) or placebo. Supporting the decision to go forward was a phase-2 trial in HIV negative Thai Adults of an ALVAC-HIV prime combined with an AIDSVAX B/E boost, which showed induction of pre-specified cellular and humoral responses. Opposing it was a 2004 editorial in Science signed by 22 top AIDS researchers describing the 105 million USD study as a “waste of money.”

Volunteers were monitored for the co-primary endpoints of HIV-1 infection and early HIV-1 viremia at the end of the vaccination series and every six months thereafter for three years. The results provided a ray of hope for HIV-1 vaccine development, as there was a trend toward prevention of HIV-1 infection among vaccine recipients with intention-to-treat analysis efficacy of 26.4% (P = .08), per-protocol efficacy of 26.2% (P = .16) and a “modified intention-to-treat” efficacy of 31.2% (P = .04).

Only the “modified” analysis reached the traditional, albeit arbitrary, measure of statistical significance and this required the problematic adjustment of subtracting five HIV-1 infections, later thought to be present at enrollment, from the vaccine group but only two from the placebo group. Also, as is common when there are many co-variates, randomization did not prevent significant residual between group differences in age, living with a partner and baseline risk factors. Most importantly, no efficacy trial to date has been able to control for the fundamental risk-factor of exposure to HIV-1 infection corrected for viral load.

Assuming that the vaccine was partially effective, what are the possible mechanisms? There was no evidence that vaccine produced “neutralizing” antibody or affected viral load or CD4+ counts in patients with HIV infection. Still, it is possible that newer techniques to better measure high affinity, broadly cross-reacting neutralization antibodies to an array of primary isolates may reveal significant differences in those who became infected compared with those who did not. It will also be important to compare the vaccine HIV strains with the circulating strains that infected the vaccine recipients.

By the publication of the Thai Study (RV144) in December, there had been more than 191 HIV vaccine trials, including 74 using poxvirus vectors, costing several billions, and without a truly promising candidate vaccine in sight.

Lessons learned

Nevertheless, there are reasons to be optimistic, as much has been learned, including what does not work, narrowing the field for future research. Because the cell co-receptors (CD4, CCR5, etc.) required for most HIV cell entry seem to be conserved, there must be a limited repertoire of functionally interacting HIV envelope conformation epitopes (the “keys” to the co-receptor “locks”) to accomplish membrane fusion. Simian models clearly have demonstrated that neutralizing antibody administered before IV or intravaginal simian immunodeficiency virus challenge can prevent infection; and recent studies of antibody responses in humans with long-term HIV infection indicate that some individuals mount a potent humoral response with the in vitro ability to neutralize isolates from unrelated subjects. Are these antibodies the Holy Grail of HIV vaccinology, ie, high affinity and broadly cross-reacting with conserved native envelope epitopes?

Perhaps most encouraging of all were the extraordinary efforts by investigators, sponsoring institutions and community volunteers to faithfully carry out these four demanding community trials. Dropout rates were consistently less than 3% per year, and by far, the most common reason for joining a trial was the altruistic opportunity to help find a vaccine. When better HIV vaccine candidates emerge, we know how to conduct large-scale clinical trials.

Editor’s Note:

Dr. Frank Judson has been a colleague of mine at this medical center for many years. He is a Professor, Department of Medicine (Infectious Disease) and the Colorado School of Public Health, University of Colorado, Denver.

He has had a long-standing interest in vaccine evaluation and was involved in each of the four trials he described in his comments. At present, he also serves as a member of CDC’s Advisory Committee on Immunization Practices. –Theodore C. Eickhoff, MD

For more information:

• Buchbinder SP. Efficacy assessment of a cell-mediated immunity HIV-1 vaccine (the STEP Study): a double-blind, randomized, placebo-controlled, test-of-concept. Lancet. 2008; 372: 1881-93.
• Flynn NM. Placebo-controlled phase 3 trial of a recombinant glycoprotein 120 vaccine to prevent HIV-1 infection. J Infect Dis. 2005; 191: 654-65.
• Pitisuttithum P. Randomized, double blind, placebo-controlled efficacy trial of a bivalent recombinant glycoprotein 120 HIV-1 vaccine among injection drug users in Bangkok, Thailand. J Infect Dis. 2006; 194: 1661-71.
• Rerks-Ngarm S. Vaccination with ALAC and AIDSVAX to prevent HIV-1 Infection in Thailand. N Engl J Med. 2009; 361: 2209-20.