The release of the results from the phase III efficacy trial sponsored by Merck, HIV Vaccine Trials Network (HVTN) and the NIAID-NIH has shattered the community of scientists engaged in the field of HIV vaccine. The STEP vaccine trial (also referred to as HVTN 502 or Merck V520-023) relied on recombinant adenovirus serotype 5 (rAD5) as a vaccine vector to induce a strong anti-HIV T-cell immunity. The trial included three rAD5 vectors expressing gag, pol or nef coding sequences and was expected to either prevent vaccinees from HIV infection, or to reduce significantly the plasma viral loads at post-infection set point. The STEP study enrolled 3,000 volunteers at sites in Australia, Brazil, Canada, the Dominican Republic, Haiti, Jamaica, Peru, Puerto Rico and the United States. The first set of results, released in September 2007, showed no protection from infection in rAD5-gag, rAD5pol and rAD5nef vaccinated individuals; surprisingly, the vaccinated cohort showed an increased number of new infections compared with the control group 1. In fact, an independent Data and Safety Monitoring Board (DSMB) found 24 cases of HIV infection among the 741 volunteers who received at least one dose of the investigational vaccine compared with 21 cases of HIV infection among the 762 volunteers who were vaccinated with placebo. In volunteers who received at least two vaccinations, the DSMB found 19 cases of HIV infection among the 672 volunteers who received the investigational vaccine and 11 instances of HIV infection among the 691 subjects who received the placebo. Surprisingly, the STEP results underscored that people who got the vaccine were more likely to get infected with HIV. Moreover, the increased susceptibility to HIV was preeminent in subjects with pre-existing immunity (antibodies) to adenovirus type 5 (Ad5) (Table 1). As a consequence, enrollments and vaccinations were discontinued. It is noteworthy that the STEP trial was stopped before completing the immunizations in most of the volunteers and that the reported infections occurred in subjects with incomplete vaccine regimen, thus precluding a fair evaluation of the vaccine protection. In addition, the control groups were not inoculated with empty vector and do not act as true controls to evaluate the potential enhancement of infection. Nevertheless, the available results do suggest that multiple immunizations with a single immunogenic vector may stimulate a harmful anamnestic response.

These results represent a grim setback for the HIV vaccine field; however, they came with no surprise within the vaccine community. Indeed, such an outcome was predicted given the nature of the HIV-1 infection and by previous preclinical trials in monkeys. HIV-1 infection is characterized by an acute phase of infection where the virus integrates into the host CD4-positive cell genome and establishes an early viral reservoir characterized by a HIV latency in memory CD4+ T cells 23. However, upon a stochastic activation of CD4 T cells by multiple stimuli, the HIV latency is overcome by the Tat-mediated transactivation of the LTR, which promotes the virus gene transcription and production of infection-competent viral particles that spread to uninfected T cells thereby reconstituting the viral reservoir 3456. Reactivation of viral infection occurs preferentially in activated memory T cells 789, and results in a limited T-cell repertoire and in a reduced immune competence 10. In this setting, the immune system mounts a robust anti-HIV T cells immune response that decreases significantly the acute plasma viremia to a lower set point; however, shortly after the primary infection, the T cell-mediated immune pressure generates escape mutants that mediate new rounds of infection that result in increased virus spreading and further deletion of CD4⁺ susceptible cells 11121314 (shown in Figure 1, 2).

Antibody-mediated immune response is elicited after few weeks from primary infection; however, the anti-HIV antibodies show a poly-specificity for non-neutralizing, linear epitopes 15. Antibodies specific for conserved, conformational epitopes are elicited at later stage and may play a favorable role by decreasing the plasma viremia 161718. In addition, anti-HIV envelope antibodies also contribute to the containment of infected T cells by triggering antibody-mediated cell cytotoxicity (ADCC; 19). As in the case of virus escape mutants promoted by the T-cell immune pressure, escape mutants are generated in presence of epitope specific antibodies 20.

In this setting, most vaccine trials in humans have been initiated under the assumption that the immune response to HIV-1 differs from the one elicited in non-human primates upon infection with SIV or SHIV strains; by inference, only vaccine trials in humans would validate vaccine candidates. Consistently, a large number of monkey vaccine studies have been overlooked despite the evidence that the monkeys' models of SIV/SHIV infection mimic well the natural HIV-1 infection in human subjects. Indeed, monkeys infected with SIV or SHIVs show levels of viremia comparable with the ones experienced by HIV-infected subjects in the course of the primary infection and at the subsequent set points 2122; CD4 activated T cells are the main susceptible cell type and the involved anatomical districts, mucosal sites and lymphoid tissues, are similar 2; the functional characteristics of the monkeys' immune response do not differ significantly from the one elicited in humans both at the time of primary infection and at the later stage of the immune deficiency 1323; the T cell-mediated and the antibody immune response are comparable with the ones elicited in humans infected with HIV-1 strains and give rise to escape viral mutants, as reported in HIV-1 infected subjects 122324. Relevant to the vaccine field, several monkey studies have documented that passive transfer of neutralizing MAbs induced protection from systemic and mucosal challenge 252627. This evidence indicates a narrow path to a successful HIV vaccine based on the induction of neutralizing antibodies focused on conserved, constrained envelope domains.

The results coming from previous studies in monkeys, including the ones promoted by Merck in collaboration with academic laboratories, may have been largely ignored. For example consider the world's first 2 phase 3 of HIV-1 vaccine efficacy trials that involved two large cohorts of HIV-1-uninfected volunteers and were completed in 2003 28. The first trial (VAX004) included 5403 individuals from North America and The Netherlands; the cohort included both men who have sex with men and 308 women at high risk for heterosexual transmission. The second trial (VAX003) was conducted in 2527 infection drug users subjects in Thailand 2930. Both the studies relied on recombinant gp120 envelope protein as the vaccine immunogen. Results from both the studies indicated that despite a robust antibody response elicited in vaccinees, no significant protection from HIV-1 infection was detected in the placebo and in the vaccine inoculated groups 28. These negative results were anticipated by monkeys studies designed to test a variety of envelope proteins as anti-HIV immunogens 31. Among these pre-clinical studies, the 2001 work of Cho et al. 32 analyzed the efficacy of mono- or polyvalent envelope proteins from distinct HIV clade B strains. This study reported that despite the induction of a strong anti-envelope antibody in vaccinated monkeys, no significant protection from the SHIV challenge was detected 32.

In the case of the Merck's STEP HIV vaccine trial, the rationale was based on a preclinical vaccine study in rhesus macaques monkeys (Macaca mulatta) primed with a plasmid expression vector followed by a boost with Ad5, a replication-incompetent adenoviral vaccine vector, which both delivered the SIVmac239 gag gene; the monkeys were then challenged with SHIV-89.6P, a pathogenic strain 33. The reported results showed high levels of CD8+ T-cell response by tetramer staining at around week 35 post-challenge and a preserved CD4 T-cells up to day 180. Peak viremia was in the range of 10⁶–10⁷ at about week 3, one log lower than the naïve animals; animals were followed up to 120 days 33. In a distinct study, Barouche et al. reported the case of a monkey immunized with a plasmid expressing SIV gag plus IL2-Ig protein and challenged with pathogenic SHIV89.6P 34; this animal, which showed low levels of plasma viremia at week 10–20, developed an escape mutant of the Gag p11C immunodominant epitope and showed a rapid disease progression and death from AIDS-related syndrome 11. These reports are fully consistent with the dynamics of immune escape of HIV in infected subjects 417 and highlight the need for a vaccine capable of inducing a drastic decrease of the viremia at the primary infection to prevent the development of pathogenic escape mutants and the impairment of the T-cell repertoire.

These representative studies raise the question of the quality and the extent of the induced immune response. Indeed, both in monkeys and in human vaccine trials, positive results were expected from the presence of either a strong anti-envelope antibody response 28, or a robust T-cell mediated immunity 1. Clearly, the main correlates of protection, anti-envelope antibodies and T-cell induced activation, need to be re-evaluated. In the case of the antibody response, previous studies have assessed the specificities and concentrations of neutralizing Abs required to achieve a complete protection from systemic or mucosal infection 2526; however, the major challenge in the field is to devise novel immunogens to induce antibodies that match the epitope specificities and affinity of the few available neutralizing antibodies 35. In the HIV vaccine field, the T cell-mediated immune response is assessed by ELISPOT and polychromatic FACS analysis to detect cytokine-expressing cells 36 without a functional assessment of T-cell killing activity toward HIV-infected cells. Thus, the functional parameters of successful T-cell activities are still missing and may hamper the predictive value of the ongoing phase 1–2 immunogenicity studies 373839.

In the context of the above-mentioned studies, clinical trials have flourished upon societal and economic pressure from geographic regions heavily affected by AIDS, and fueled by a large financial support from public and private institutions 4041. However, vaccine trials with the available immunogens in different vector systems may, at best, only moderately decrease the levels of plasma viremia and are not expected to improve the protection from infection. Likewise, any trial with a suboptimal immunogen in a high HIV incidence areas may create a significant number of vaccine breakthroughs and would raise ethical and financial issues concerning the treatment of the infected volunteers 42. In this context, we should learn from the STEP trial that preclinical studies in monkeys are feasible and may anticipate future results in human subjects. Preclinical studies in non-human primates should be expanded in terms of number of animals, study designs to assess potential negative effects and length of the post-challenge follow-up. Several lessons are coming form the Merck vaccine study:

a. A vaccine aimed at inducing T-cell immunity will not prevent primary infection and the establishment of the viral reservoir;

b. Studies aimed at inducing neutralizing antibodies should be pursued with even greater efforts toward a rational development of structurally defined conformational epitopes, and tested in suitable monkey models;

c. The host immune system will react to both the vector and the immunogen in an unpredictable way, by activating a complex immune response that includes the innate and the epitope-specific immunity;

d. Studies aimed at modeling human behavior in the context of HIV exposure during vaccination should be developed and used to test for potential vaccine-induced enhancement of infection.

The STEP vaccine has not provided a successful road to an effective HIV vaccine; however, it has shown to us that we still know too little and should learn more.

The authors declare that they have no competing interests.

EI, MS and GF selected the relevant literature. IQ and GS wrote the commentary. All the authors read and approved the final manuscript.