This study shows that Ad35-GRIN/ENV is safe and immunogenic when administered twice to Ad35-seronegative healthy volunteers with doses from 2×109 to 2×1011 vp. Reactogenicity was dose-related with frequent moderate to severe systemic symptoms consisting of headache, malaise and chills that began within 12–24 hours of initial vaccination. All reactions were self-limited and resolved within hours (systemic symptoms) to a few days (injection site symptoms).
Ad35-GRIN/ENV at all dose levels was found to be immunogenic, producing IFN-γ ELISPOT responses in 90% of all vaccinated individuals although the magnitude was relatively modest to individual peptide pools. Anti-gp140 binding antibodies were seen in all vaccine recipients in Groups A–C. The median and range of Env and Nef ELISPOT responses were lower than those seen to Gag, RT and Int and likely because the PFC assay is less sensitive than the ELISPOT assay, we saw fewer responses to Env and Nef. The IAVI-HIL ELISPOT assay has a positive cut off value of 38 SFC per million PBMC. Assuming 50% of PBMC are T cells, the 38 SFC value would equate to 0.0076% of CD3+ T cells producing IFN-γ. The median range of Env ELISPOT was 60–108 SFC per million PBMC across groups A–C which equates to 0.012 to 0.0216% of CD3+ T cells responding. Such values are below the limit of detection of our current ICS assay. Because of the lower sensitivity of the PFC assays, responses at 2 weeks post-2nd vaccination were less frequently positive than the IFN-γ ELISPOT assays, but indicated that the cellular response was primarily mediated by CD8+ T lymphocytes and was typically polyfunctional, with TNF-α and CD107 detected in addition to IFN-γ. A trend of a dose-response effect in the proportion of subjects with positive IFN-γ ELISPOT assays was seen after a single injection (the small size of the study groups precluded an assessment of statistical significance); however, the booster vaccination at 6 months induced responses in nearly all the participants in all groups. There was no clear indication of an increase in magnitude or breadth (the latter by the number of peptide pools recognized) in IFN-γ ELISPOT responses by dose or as a result of the 6-month booster immunization.
HIV gp140-specific binding antibodies were frequently seen after the first immunization among recipients of the Ad35-GRIN/ENV with a titer significantly increased with the 6-month boost in all groups (p<0.0004), but declined appreciably within the following 6 months. In addition, anti-Ad35 neutralizing antibodies were induced at a low frequency and low titer after the first immunization, but the 6-month boost increased the proportion of responders significantly, particularly among those who received the 2×1011 vp dose.
The presence of pre-existing Ad neutralizing antibodies and the impact of such antibodies on reducing HIV-specific immune responses after multiple immunizations remains a major concern for the use of Ad vectors. In this study where volunteers enrolled were all Ad35-specific neutralizing antibody negative at baseline, we found that even after a second immunization with Ad35 vectors, Ad35-specific neutralizing antibody titers remained low, regardless of the dose. HIV Env- and Gag-specific antibodies were significantly boosted after the second immunization with modest increases in cellular responses, suggesting that HIV-specific cellular and humoral responses were not impacted by Ad35-specific neutralizing antibodies induced by the first Ad35 vector immunization. We cannot however exclude the interference of cross-reactive preexisting Ad-specific T-cell immune responses on the vaccine-induced T-cell responses 
. The low prevalence and titers of Ad35 compared with Ad5 neutralizing antibodies as well as the induction of adequate cellular and humoral responses shown in this study supports the use of Ad35 vectors in heterologous or homologous prime-boost strategies 
This study also explored the potential phenomenon of an Env immunodominance as suggested in animals 
and other human studies 
by comparing IFN-γ ELISPOT responses with GRIN peptide pools between Group B and Group D (who received 2×1010
vp of Ad35-GRIN/ENV or 1×1010
vp of Ad35-GRIN, respectively). Although the small size of the groups precluded firm conclusions, the frequency of positive IFN-γ ELISPOT assays in Group D appeared higher after the first injection relative to Group B. Furthermore, overall, the magnitude of positive responses was significantly higher in Group D (vaccinated in the absence of Env genes) compared to Group B for Gag, Pol/Int and Nef pools. There was insufficient evidence from this small study to evaluate the mechanism of possible Env interference, however as indicated above with the low Ad35 neutralization titers even after the second boost and at the highest dose, we could not see an impact on insert specific immune responses. Understanding how best to administer antigens to avoid immunodominance, competition and the effects of pre-existing or post immunization vector specific responses is an ongoing issue for development of HIV, malaria, Cancer, TB and other vaccines 
Interestingly, one study participant who received placebo in this trial demonstrated IFN-γ ELISPOT responses and anti-gp140 binding antibody at baseline and all timepoints throughout the trial in spite of testing negative in the licensed diagnostic kit Abbott Architect HIV Ab/Ag Combo EIA at the University of Rochester at study entry and negative HIV RNA levels. This person did not disclose any behavior associated with high-risk for HIV infection and had not participated in previous HIV vaccine clinical trials. The reasons for these findings are still under investigation.
This study adds to the accumulating literature of safe and immunogenic candidate vaccines that could be included in combination vaccine regimens. Similar to the Ad5 vector tested in the Step Study, the Ad35 vector induced a high frequency of HIV-specific CD8+ T-cell responses, however, our data suggests that the Ad35-GRIN/ENV candidate induced a broader response in terms of the number of protein regions recognized. In the original Step study, 218 of 354 (62%) of individuals recognized two to three HIV proteins with a median of 1 protein per person 
. Further assessment of the breadth of response was performed in 72 subjects who participated in an earlier Phase I trial of the same vaccine utilized in the Step study. In the latter study, using mini-pools containing eight 9-amino acid (aa) peptides spanning 16 aa of the vaccine sequence, the median (and mean) of positive mini-pool responses per subject was 1 (1) for Gag, 1 (1) for Nef and 2 (3) for Pol 
. It is unclear what this may mean in terms of potential vaccine efficacy, but the Step Study did show evidence that vaccine-elicited T cells had an impact on the HIV-1 strains that established infection (a sieve effect) 
and thus it is reasonable to conclude that a broadened antiviral effect could be beneficial. In addition, this study suggests that improved non-envelope IFN-γ ELISPOT responses may be induced if vaccination with Env proteins or Env encoding vectors are separated in time or space.
As noted above, the community-based trial conducted in Thailand (RV144) that employed a combination regimen of a recombinant canarypox vector expressing HIV-1 gag, protease, env
genes boosted by a recombinant HIV-1 gp120 envelope protein subunit showed protection against HIV acquisition however without a measurable effect on viremia or the CD4+ T-cell count in vaccinated and infected subjects. The vaccination regimen induced modest levels of HIV-specific cellular immune responses, mediated primarily by CD4+ T cells and high titers of HIV-envelope specific binding antibodies 
. The Step and RV144 results reinforce the need to develop new vaccines or vaccine combinations able to induce more effective immune responses, in particular at the mucosal level where HIV transmission events occur 
. Combination regimens using heterologous vectors in prime-boost and inserts aiming at broadening CD4+ and CD8+ T-cell responses such as mosaics 
and conserved sequences 
are promising avenues. Indeed, three other phase I clinical trials, one with Ad35-GRIN in combination with an adjuvanted core HIV protein vaccine (IAVI B002; NCT01264445), a second with Ad35-ENV in combination with Ad26-ENVA (IAVI B003-IPCAVD004-HVTN091; NCT01215149), both of which have completed vaccinations, and a third with Ad35-GRIN/ENV in prime-boost regimens with DNA administered by electroporation and adjuvanted with IL-12 have been initiated (IAVI B004; NCT01496989). These trials are designed to increase the breadth in terms of epitope coverage, improving the CD4+ and CD8+ T-cell responses and functional antibody responses. In spite of the lack of HIV neutralizing antibodies, additional studies on the fine specificity of the Env antibody response induced by Ad35-GRIN/ENV are also warranted following the recent findings that IgG binding antibody to a scaffolded HIV-1 gp120 V1V2 protein encompassing the gut homing marker α4β7 integrin binding site was identified in a post-hoc
analysis as a correlate of risk for acquisition of HIV infection in RV144 
. In conclusion, the Ad35-GRIN/ENV candidate was generally well tolerated and immunogenic at a dose of 2×1010
vp and has moved forward into additional studies to further assess safety and immunogenicity and should also be considered as a potential component of a regimen tested in efficacy trials.