The multicenter Step Study evaluated a Merck Ad5/HIV trivalent subtype B vaccine in HIV-1-uninfected persons reporting high-risk activities and residing in regions where HIV-1 subtype B predominates. No significant reduction in the vaccine group, compared to the placebo group, was observed in HIV-1 acquisition rates [
1]. In fact, the initial analyses revealed an increased infection rate in uncircumcised male vaccine recipients with pre-existing adenovirus serotype 5 neutralizing antibodies [
1], but detailed regression analyses with additional data indicate that circumcision status in the vaccinees was the significant baseline risk factor associated with higher infection rates (S. Buchbinder, unpublished). Moreover, while the T cell-based vaccine regimen was assumed to have a greater impact on control of viremia post-infection than on prevention of infection, the vaccine failed to lower set point viremia despite eliciting remarkably high HIV-specific CD8+ T cell response rates [
3].
A wide range of investigations ensued in an attempt to explain the lack of vaccine efficacy, and a few key issues have emerged as relevant to control of acute viremia. A chief consideration has been the incorporation of the appropriate HIV-1 genes or proteins into the vaccine candidate. The selection of vaccine antigens for HIV-1 has been largely empirical, but common approaches include expression of HIV-1 structural genes, particularly
gag for those vaccines intended to induce primarily CD8+ T cells and
env for those intended to induce neutralizing antibodies. The MRKAd5/HIV trivalent vaccine, intended to induce a T cell response, encoded HIV-1 Gag, Pol and Nef, but did not encode Env because of its high variability and the desire to exclude functional antibodies in this test of concept study. While HIV-specific CD8+ T cells are able to exert control of viral replication during acute infection in the absence of Env-specific neutralizing antibodies [
4–
6], it remains unclear if the combination of effector responses, previously primed by vaccination, may offer benefit over either response alone in containing initial bursts of infection in mucosal epithelium. In progress is an expanded Phase II vaccine trial (HVTN 505) evaluating an envelope-containing vaccine regimen using a HIV-1 DNA prime and replication-incompetent Ad5/HIV boost, and results may yield further insights into the role of Env antigens in HIV-1 protection.
An additional consideration is the understanding of the epitope specificities that are important to elicit in a protective HIV-1 vaccine. Detailed studies of immune responses in acute HIV-1 infection have revealed the emergence of circulating HIV-1-specific CD8+ T cells as viremia peaks, and rapid viral escape mutations can occur within or upstream of the targeted epitopes [
6–
9]. Thus, viral epitopes first recognized in acute infection are unlikely to represent those commonly seen in chronic infection. Advances in the early diagnosis of HIV infection and implementation of these screening tests into large scale vaccine trials will permit better understanding of whether a vaccine is exerting pressure on incoming virus or whether escape is occurring very early within the epitopes originally recognized. Furthermore, although Gag-specific CD8+ T cells have been associated with lower levels of plasma HIV-1 RNA [
10], the range and specificities of these responses may be important in generating effective vaccine-induced protection. In the Step Study, participants who acquired HIV-1 infection (cases) mounted similar frequencies of Gag-specific CD8+ T cell responses following vaccination compared to matched controls who did not acquire infection (40% vs. 42%, respectively) [
3]. Some cases mounted vaccine-induced CD8+ T cells recognizing as many as six Gag epitopes, but on average only one Gag epitope was recognized, and these were located within both variable and conserved regions of the protein (McElrath, unpublished findings). Although no vaccine effect was demonstrated on post-infection viral load, viral sequence analysis suggests that immune pressure was exerted by the vaccine on the early founder viruses [
11]. Taken together, it is likely that T cell-inducing vaccines will need to either elicit broader responses or target certain antigen specificities associated with conserved or functional sequences in order to contain early bursts of viremia and rapid viral escape mutation. In this regard, new vaccine designs employing conserved sequences and antigenic mosaic sequences may have utility, as they have the potential to induce T cell responses to high frequency viral strains, a greater number of epitopes and more variants within a given epitope [
12–
14]. Clinical trials will soon evaluate these approaches.
Understanding correlates of protection with T-cell inducing vaccines may require more extensive analysis of the anti-viral functional properties of the T cells elicited. In recent years, extensive efforts to standardize and validate immunogenicity measurements have been undertaken in order to clearly define response rates, magnitude and antigen specificity of the responding T cells [
15,
16]. These assays are robust, highly sensitive and reliable when applied in early trials to screen vaccine candidates and to compare vaccine regimens, and they have utility for bridging immunogenicity results from the previous Phase I/II studies to the larger trials. However, they may not be sufficient as the benchmark for predicting vaccine efficacy. Additional approaches to be incorporated into future studies include, alone or in combination, analyses of T cell suppression of viral replication, proliferation and cytolysis, and broader immunologic functions at the transcriptional and proteomic levels [
3,
17–
19]. Key factors in moving these platforms forward will be scaling them for performance in small specimen numbers, making them cost-effective, and building the capacity to manipulate the resulting large databases for timely comparative analyses.
The publisher's final edited version of this article is available free at 
