a Publication of the HIV Vaccine Trials Network Volume 5, issue 1 OCTOBER, 2013 CLOSING IN ON Probing scientific questions to advance HIV vaccine development Tracey Day and Cecilia Morgan Several HVTN trials have gone beyond the standard safety and immunogenicity assessments to address important scientific questions for the field. These studies provide a means to accelerate vaccine development by obtaining critical information applicable to multiple vaccine concepts. Here, we describe some of the key scientific questions that recent HVTN studies have probed and provide a sneak peak at the emerging data. Maximizing T-cell response breadth (HVTN 083, 085) Considerable evidence indicates that HIV specific T- cell responses play a key role in controlling the levels of virus detected in the blood of infected patients. 1 Although the Merck recombinant adenovirus serotype 5 (rad5)-vectored vaccine elicited high T-cell response rates, it did not protect against infection or reduce viral loads in infected participants in the Step Study. 2,3 One possible explanation that has been postulated is that the T-cell responses elicited by this vaccine were too narrowly focused to provide adequate control of viral replication. 4-6 Few pathogens exhibit the level of genetic sequence variability that is observed in HIV strains. Eliciting immune responses capable of recognizing such a diverse range of variants is one of the most significant challenges for HIV vaccine development. Vaccines that induce greater T-cell response breadth, defined IN THIS ISSUE 11 18 22 Innovative Research Arising from HVTN Award Programs HVTN Protocols -- Enrolling or in Follow Up HVTN Efficacy Trials Results 24 35 38 Signs of the Time: Technology Updates at the HVTN The HIV Vaccine Trials Network at AIDS VACCINE 2013 HVTN Annual Network Award Winners CALENDAR [back cover]
Probing scientific questions to advance HIV vaccine development Probing scientific questions to advance HIV vaccine development as the number of HIV-specific epitopes recognized, may block viral escape pathways and lead to earlier control of replication or slower disease progression for infected individuals. Some epitopes are conserved, or shared, across many HIV variants. Increasing the number of conserved epitopes that are recognized by vaccine-induced T cells, in particular, may further optimize protection from diverse HIV variants. HVTN 083 and HVTN 085 are 2 studies that are being conducted to evaluate multiple factors that may improve the quality of vaccine elicited cellular responses. Rather than validating the specific adenoviral (Ad)-vectored products used, these trials aimed to identify general approaches that may improve the magnitude and breadth of T-cell responses induced by vaccination. Data from these studies have recently become available. HVTN 083 is a phase 1 clinical trial that evaluated prime/boost regimens comprising heterologous vectors and/or heterologous inserts as a means to improve the breadth, magnitude, and/or character of the T-cell response (Figure 1). The study was designed to answer 2 main questions: 1. Do priming and boosting with inserts containing overlapping, non-identical sequences of the same gene focus the T-cell response to conserved regions present in both inserts? 2. Does a prime/boost regimen containing heterologous vectors improve cellular immune response, compared to a regimen containing homologous vectors? To answer the first question, 2 HIV envelope glycoprotein (Env) inserts, one from clade A (Env A) and the other from clade B (Env B), were used with rad vectors in the study. T-cell responses targeting epitopes that were either shared between the Env A and Env B inserts or more generally conserved across HIV strains would theoretically recognize a larger number of distinct HIV isolates. For the second question, the study looked at vectors of different rad serotypes (rad5 and rad35), with the same Figure 1. HVTN 083: Do prime/boost regimens containing heterologous inserts or heterologous vectors enhance vaccine-induced T-cell response breadth? To answer this question, the targets of HIV Env-specific T cells elicited in each study arm were determined via epitope mapping. T cell IFN-γ/ IL-2 intracellular cytokine staining assays were performed with specimens collected 2 weeks after the final vaccination. Peripheral blood mononuclear cells were stimulated with pools of overlapping peptides spanning the Env protein sequence. Positive samples were further tested with individual peptides. In the graph on the left, the frequencies of positive samples (left panel, x-axis) are shown for each peptide relative to the corresponding Env protein sequence (left panel, y-axis). Color coded dots indicate results from each study arm, as illustrated in the study schema to the right. This figure was adapted from a slide presented by Dr. Stephen Walsh at the spring HVTN Full Group Meeting. In this preliminary analysis, it was observed that some epitopes were more frequently targeted and may represent immunodominant hotspots (for example, around residues 50 and 450). Shaded ovals highlight regions where an increase in T-cell response breadth was observed in heterologous insert groups (group B, orange dots; group E, green dots). or different inserts, compared to 2 administrations of the same rad serotype vector. Previous studies have indicated that heterologous vector prime/boost regimens result in higher magnitude T-cell responses than homologous vector regimens. 7-9 This may be due to rapid generation of anti-vector immunity following re-administration of a homologous vector. HVTN 083 enrolled 180 subjects who lacked preexisting neutralizing antibodies specific for Ad35 or Ad5 or both. The 3 vaccines used in the study (rad35 Env A, rad5 Env A, and rad5 Env B) were developed by the Vaccine Research Center (VRC) at the National Institute of Allergy and Infectious Diseases (NIAID). Participants were randomized to one of 5 groups evaluating different combinations of heterologous inserts, homologous inserts, heterologous vectors, and homologous vectors administered in prime/boost regimens as indicated in Figure 1. Preliminary HVTN 083 immunogenicity data from the peak response time point were presented by Dr. Stephen Walsh at the spring HVTN Full Group Meeting in Washington, DC. All participants were exposed to Env A as part of their priming immunization and the majority developed Env A-specific T-cell responses as indicated by positive interferon-γ (IFN-γ) enzyme-linked immunosorbent spot (ELISpot) assay responses to Env A peptide pools. Env B response rates were lower than Env A overall. Groups that received Env B in the boost vaccine generated the highest Env B response rates. A similar pattern was observed from IFN-γ and IL-2 intracellular cytokine staining (ICS) assays. Epitope mapping was performed to compare T-cell response breadth for each study arm. This method involves consecutive rounds of INF-γ ELISpot assays to de-convolute which peptides within a positive peptide pool are being targeted by the T cells in a specimen. Dr. Walsh presented preliminary epitope mapping results suggesting that vaccination with heterologous inserts led to somewhat of an increase in T-cell response breadth (Figure 1). An assessment of the number of shared or conserved epitopes that were targeted by each study arm has since been performed. Responses from heterologous insert groups were found to target more epitopes that were shared between the Env A and Env B inserts than the homologous insert groups. For heterologous vector groups, numbers of targeted epitopes, both insert-specific and shared, were higher overall than for homologous vector groups. Further analyses are ongoing. Neutralizing antibody responses were an exploratory endpoint in HVTN 083. It was hypothesized that if heterologous insert regimens elicited more conserved epitope responses, humoral responses could possibly focus on epitopes more vulnerable to neutralization. Dr. Walsh presented data from neutralizing antibody assays performed in TZM-bl cells in which neutralization activity was detected largely against a single tier 1 clade C isolate known to be relatively easily neutralizable (MW965.26). In contrast to the exploratory hypothesis, neutralizing antibody response rates were lower in heterologous insert groups. The preliminary findings from HVTN 083 show that prime/boost regimens containing heterologous inserts and vectors can increase breadth of T-cell responses and enhance recognition of conserved epitopes. HVTN 085 was another study evaluating parameters that may enhance vaccine-induced cellular responses. In this study, a polytopic vaccine administration (injecting a vaccine in multiple anatomical locations) was evaluated as a means to increase T-cell response magnitude or breadth (Figure 2). The rationale for this study was based on the concept that by spreading the antigenic wealth, so to speak, certain types of narrowly focused (immunodominant) T-cell responses may be decreased. As presented by Dr. Ian Frank during the spring HVTN Full Group Meeting, polytopic vaccine administration has been shown in a murine study to overcome an immunodominant response to a particular Gag epitope, thereby enhancing responses to a subdominant epitope. 10 In addition, it has been postulated that polytopic vaccine administration, which is often used to accommodate appreciable vaccine volumes in relatively smaller animals, could help explain why vaccine-induced responses are often significantly higher in non-human primates compared to humans. 2 OCTOBER 2013 VOLUME 5:1 HVTNEWS www.hvtn.org HVTNEWS VOLUME 5:1 OCTOBER 2013 3
Probing scientific questions to advance HIV vaccine development Probing scientific questions to advance HIV vaccine development The vaccine tested in HVTN 085 was a 4 component rad5-vectored vaccine developed by the VRC (VRC rad5 gag-pol, env A/B/C). The study compared 3 different administration strategies that are illustrated in Figure 2. In the first group, a combined vaccine containing all 4 vaccine components was injected in a single limb (right arm), a typical vaccine administration strategy in humans. In the 2nd group, each vaccine insert component was injected in a different limb (right arm, left arm, right thigh, left thigh). It was hypothesized that by targeting each insert to a different lymph node, this strategy would eliminate insert competition within the same lymph Figure 2. HVTN 085: Does polytopic administration improve T-cell response quality? To answer this question, T-cell responses were compared when a 4-component adenovirus-vectored vaccine was administered in the same or different limbs as shown in this illustration. Top: all 4 vaccine components were administered intramuscularly in a single injection in the right deltoid, a typical administration strategy. Bottom left: each vaccine component was administered in a different limb, a polytopic administration strategy targeting each insert to a different lymph node. Bottom right: all 4 vaccine components were combined and diluted 1:4 and administered in each of 4 limbs, a polytopic administration strategy distributing a quarter dose of all antigens over 4 lymph nodes. This figure was adapted from illustrations presented by Dr. Ian Frank at the spring HVTN Full Group Meeting. node and thereby help overcome any potential immunodominance. As in the murine study, curtailing dominant responses would be expected to increase T-cell response breadth. In the 3rd group, the combined 4-component vaccine was diluted to a 1/4 dose and equal volumes were injected in 4 different limbs. By distributing a fraction of the complete vaccine among multiple lymph nodes, this administration strategy was hypothesized to increase the magnitude of the response. To maintain the study blind by treatment group, placebo vaccinations were given in the remaining 3 limbs in the first group. Dr. Frank presented the T-cell response data (as assessed by IFN-γ/IL-2 ICS) available thus far from the primary immunogenicity timepoint. Overall, there was a trend toward increased number of responders, particularly for HIV-specific CD8 + T cells, and higher magnitude responses against some antigens in the group receiving the complete diluted vaccine in 4 limbs (Arm 3, Figure 2). A preliminary assessment of T-cell response breadth based on number of peptide pools generating a positive ICS response was also presented. A significantly higher number of peptide pools were recognized by CD8 + T cells from the group receiving the complete diluted vaccine in 4 limbs (Arm 3, Figure 2). This suggests that polytopic administration of the complete vaccine may increase breadth for CD8 + T cells. Complete epitope mapping analyses are ongoing and will allow a more thorough evaluation of how different polytopic vaccination strategies affected T-cell response breadth. Findings from HVTN 083 and HVTN 085 may be applicable to other viral vector vaccine platforms, and additional strategies for improving T-cell response breadth will be assessed in future HVTN trials. One example is HVTN 099, which is being collaboratively developed by the HVTN, NIAID, the Center for HIV/AIDS Vaccine Immunology (CHAVI), Los Alamos National Laboratory (LANL), the IPPOX Foundation in Switzerland, and the Bill and Melinda Gates Foundation. This study will compare T-cell response breadth elicited by mosaic and consensus Env immunogens designed by LANL, CHA- VI, and NIAID. 11 Together these data will serve to guide future efforts to optimize T-cell response frequency and quality, in order to maximize control of HIV replication. Effects of protein boosts (HVTN 073E and 088) The RV144 trial demonstrated for the first time that vaccination can reduce HIV infection risk in humans. 12 This trial evaluated a canarypox vector prime (ALVAC, Sanofi Pasteur) and an Env protein boost (AIDSVAX B/E, VaxGen) vaccine regimen in Thailand. The study reported an estimated vaccine efficacy of 31.2% and although only partial efficacy was observed, this result invigorated the field. In a subsequent immune correlates analysis, immune responses correlating with infection risk were identified. 13 Chief among these were Env-specific antibody responses that correlated with decreased risk of HIV infection. These findings resulted in a surge of interest in vaccine-induced antibody responses. One way to promote antibody responses is through protein vaccine boost immunizations, such as the Env protein boost administered as part of the RV144 trial regimen. To begin further exploration of the effects of protein boosts, HVTN researchers conducted 2 studies in which Env protein boosts were administered to study subjects that had been primed as part of previous or ongoing HVTN trials. One of these studies, HVTN 073E, was an extension of HVTN 073/SAAVI 102. HVTN 073/SAAVI 102 was a phase 1 trial conducted in South Africa and the United States. The trial evaluated a DNA vaccine (SAAVI DNA C2) prime followed by a modified vaccinia Ankara (MVA) viral vector vaccine (SAAVI MVA C) boost. Although this regimen differed from RV144, both regimens had in common the use of a poxvirus vector vaccine. Administering an Env protein boost to HVTN 073/SAAVI 102 participants was an efficient way to examine how a protein boost could enhance antibody responses following a viral vector prime. Thus, a study extension (HVTN 073E) was developed to evaluate the safety and immunogenicity of Novartis Sub C gp140 vaccine with MF59 adjuvant as a late boost following a DNA/MVA prime/boost regimen. The first protein dose was administered around 2 years 4 OCTOBER 2013 VOLUME 5:1 HVTNEWS www.hvtn.org HVTNEWS VOLUME 5:1 OCTOBER 2013 5
Probing scientific questions to advance HIV vaccine development Probing scientific questions to advance HIV vaccine development after the participant s first (DNA or placebo) injection in HVTN 073/SAAVI 102. A second protein dose was given 3 months after the first. HVTN 073E enrolled 27 participants and data available thus far include neutralizing antibody responses from specimens collected 2 weeks after each protein administration. Preliminary results indicate that boosting with protein improves neutralizing antibody responses for DNA/MVA primed subjects. Neutralizing antibody response rates were significantly higher in those receiving the DNA/MVA prime and gp140 boost than in those who received the DNA/MVA prime and were not boosted with gp140. This was true for neutralization assays performed in both TZM-bl and A3R5 cells, in which neutralizing activity was detected against tier 1 and tier 2 viruses. Notably, neutralizing antibody responses, which were not detected in the main HVTN 073/SAAVI 102 dataset, were detected after a single protein boost in the study extension. In the absence of priming, several protein administrations are typically required to elicit neutralizing antibody responses. That these responses were detected after a single dose indicates efficient priming by the DNA/MVA regimen and points to protein boosts as an effective way to boost such responses. Whether binding antibody responses were also enhanced by the protein boosts will also be of interest. These analyses, as well as T-cell response assessments, are ongoing. HVTN 073E represents the first time that a protein was administered as a boost to a poxvirus vector vaccine in South Africa, setting the stage for further exploration of pox/protein combinations in the region. HVTN 088 is another study investigating whether boosting previously primed study subjects with protein could enhance antibody responses. This time the boosting occurred after a significant rest of 5-7 years or more. Providing a longer rest was considered a potential way to allow more time for B cell maturation and thus, possibly more effective antibodies to develop. In addition, HVTN 088 examined whether boosting with an immunogen specific for a different virus clade would result in broadening the virus cross-reactivity of the boosted immune responses. In this case, participants were previously primed with clade B immunogens and, as part of HVTN 088, they would be boosted with a clade C Env protein vaccine, Sub C gp140/mf59, the same vaccine used in HVTN 073E. The concept of using a divergent protein boost immunogen to broaden primed antibody responses was based on promising results from an influenza vaccine study. In that study, priming and boosting with hemagglutinin antigens from different influenza strains resulted in the induction of high titer antibody responses that were cross-reactive for multiple virus strains. 14 HVTN 088 aimed to test whether this would hold true for HIV Env antigens as well. Some of the HVTN 088 participants came from older studies containing clade B protein immunogens. These studies were conducted by the HVTN predecessor network, the AIDS Vaccine Evaluation Group (AVEG). The majority of participants, however, were primed as part of HVTN 049, a protocol that tested a clade B gag, env DNA vaccine (DNA/polylactide coglycolide microparticles, Chiron) followed by a clade B Env protein vaccine (clade B recombinant gp140 protein with MF59, Novartis, formerly Chiron). This regimen in HVTN 049 elicited good titers of neutralizing antibodies in 100% of recipients. But, virus cross-reactive neutralization breadth, as measured by the TZM-bl assay, was limited. An update on HVTN 088 results, based on the data available thus far, was presented by Dr. Paul Spearman at the spring HVTN Full Group Meeting. As presented by Dr. Spearman, even before boosting, a significant portion of HVTN 049 participants had detectable T-cell and antibody responses to the HVTN 049 clade B HIV antigens. This indicates that the HVTN 049 regimen elicited memory responses that could persist for at least 5-7 years. Administering the clade C protein vaccine resulted in a rapid and strong boosting of these responses after a single dose with little change after the second dose. Neutralizing antibody response data obtained in both TZM-bl and A3R5 assays indicated that the boosted neutralizing antibody responses were only slightly higher for clade B viruses than the HVTN 049 peak response levels. And surprisingly, even after the 2nd boost with a clade C protein, the neutralizing antibody responses for clade C viruses were lower, rather than higher, than the peak HVTN 049 levels. Thus, in contrast to the hypothesis that priming and boosting with divergent antigens could broaden neutralizing antibody responses, based on the HVTN 088 data thus far, response breadth seemed to decrease. These unexpected results inform researchers that increasing neutralizing antibody response breadth is complex, and highlight a need for further investigation. Detailed B-cell analyses of the HVTN 088 specimens are planned that may help investigators better understand these results. In addition, mucosal immune response analyses are ongoing, which may also provide clues as to how boosting with a divergent protein immunogen may affect antibody responses at mucosal sites, as compared to the responses detected in blood. Why might clade C neutralizing antibody responses not improve with a clade C antigen boost? This remains unknown, but as explained by Dr. Spearman during his presentation, these findings may present evidence for a theory known as original antigenic sin. This theory is based on the observation that in some cases, boosting with a heterologous antigen can result in a dampening of antibody responses to the divergent antigen. The hypothesis would be that the antibody responses generated against the priming clade B antigen dominated and interfered with responses to the clade C boost antigen. If this were indeed the case for HVTN 088, one would expect that peak clade C neutralizing antibody responses would be higher in subjects who received the clade C protein without being primed with the clade B antigens. Such a difference was not observed; however, it is possible that this was because this group received only 2 protein doses and, although response rates were steadily increasing, prior studies have indicated that 3 protein administrations are often required for peak responses. Taken together, the preliminary results from HVTN 073E and HVTN 088 indicate that antibody responses can be enhanced by protein boost administrations. Ongoing analyses will provide a more complete picture of the specific character and durability of these responses, both systemically and within mucosal tissues. Although some of these findings were unexpected, they serve to reveal the complexity of heterologous vaccination strategies and highlight the need for further studies, like these, to improve our understanding of how best to employ such approaches. Going forward, additional ongoing and future studies will further address how to utilize protein vaccines to elicit optimal responses. For more information, see the HVTN Protocols -- Enrolling or in Follow Up table in this issue (pages 18-21), specifically the descriptions of HVTN 086, 096, and 097. Exploring heterologous viral vector prime/boost regimens (HVTN 077 and 078) The use of viral vectors as HIV vaccines is attractive, because they can induce strong T-cell and antibody responses. Two recent phase 1b studies, HVTN 077 and 078, have explored using heterologous viral vector primes and boosts and factors impacting immune responses. Data emerging from these studies are revealing optimal strategies for employing viral vectors to induce robust T-cell and antibody responses. HVTN 078 was a phase 1b study that addressed two important questions regarding the optimal way to combine two different viral vectors in a prime/boost regimen: 1. How does administration order impact responses when 2 different viral vector vaccines are used? 2. What dose of a strongly immunogenic prime works best when paired with a strongly immunogenic boost? Animal studies have shown that, for heterologous viral vector regimens, order can influence the character of the responses to specific inserts. 15 Therefore, HVTN 078 compared immune responses when rad5 was used as a prime followed by a NYVAC boost to the reverse order (a NYVAC prime and a rad5 boost). In addition, 3 different rad5 doses (10 8, 10 9, and 10 10 plaque forming units) were tested to evaluate whether weak, intermediate, or strong priming produced the best responses preceding a NYVAC boost. 6 OCTOBER 2013 VOLUME 5:1 HVTNEWS www.hvtn.org HVTNEWS VOLUME 5:1 OCTOBER 2013 7
Probing scientific questions to advance HIV vaccine development Probing scientific questions to advance HIV vaccine development The study enrolled 80 Ad5 neutralizing antibody seronegative circumcised men in Switzerland. Primary immunogenicity data was assessed 2 weeks after the final vaccination. T-cell response data from HVTN 078 were presented at the AIDS Vaccine 2012 conference in an oral presentation by Dr. Pierre-Alexandre Bart and antibody response data were presented in a poster by Dr. David Montefiori. 16,17 Higher CD4 + T-cell response rates, as assessed via IFN-γ/IL-2 ICS, were observed when rad5 was used as a prime and NYVAC was used as a boost. A similar trend was observed for CD8 + T-cell responses. Different doses of rad5 used for priming did not significantly affect T-cell response frequencies or magnitude. Further analyses of T-cell response characteristics, including polyfunctionality regarding cytokine production, are ongoing. An assessment of mucosal responses is also planned. For humoral responses, the highest binding and neutralizing antibody response rates were observed when recipients were primed with the highest rad5 dose and boosted with NYVAC. Among positive responders, the same effect was observed for response magnitudes. In comparison with other rad5 studies, as noted by Dr. Montefiori, neutralizing capacity against tier 1 viruses from rad5/nyvac was generally better than has been observed with DNA/rAd5 regimens (eg, the regimen evaluated in HVTN 204, Figure 3). 17 Overall, HVTN 078 demonstrated a significant impact of vector order in heterologous vector regimens, with rad5 priming and NYVAC boost eliciting higher T-cell and antibody responses. The highest rad5 priming dose induced the strongest antibody responses, and no dose response was observed for T-cell responses. Thus, these data support the concept of a employing a strong priming immunogen as a potential means to strongly elicit both cellular and humoral responses. One potential limitation to using some viral vectors, however, is the presence of pre-existing immunity from exposure to the wild type virus that can then direct immune responses against the viral vector vaccine. The estimated seroprevalence of Ad5 neutralizing antibodies is 37-99%, and is particularly higher in some of the same regions where HIV incidence is high. For example, in South Africa this seroprevalence has been reported at 88 90%; whereas the neutralizing antibody seroprevalence to other Ad serotypes, such as Ad35, is substantially lower. 18,19 HVTN 077 is a phase 1b trial investigating whether preexisting humoral immunity to Ad5 could be circumvented using an alternative adenoviral serotype vector, Ad35. This study tested a rad35-vectored vaccine in combination with a rad5 or a DNA vaccine, all developed by the VRC and contained HIV clade A Env inserts. The study was designed to address the following questions: 1. How does rad35 perform compared to DNA as a prime for a viral vector with higher pre-existing seroprevalence? 2. How does rad35 perform as compared to rad5 as a boost to DNA? 3. What is the effect of pre-existing neutralizing antibodies to one virus serotype on a regimen that includes a DNA prime and a boost with a viral vector of a different serotype? The answers to these questions would provide specific clues for circumventing any potential negative effects of pre-existing vector immunity. The study was conducted in the US and enrolled 192 participants. Absence of detectable Ad35-specific neutralizing antibodies was required for all subjects. Data collected 2 weeks after the last vaccination was presented by Dr. Jonathan Fuchs at the HVTN Full Group Meeting held in October, 2012 and at the AIDS Vaccine 2012 conference. 20 Each regimen evaluated in HVTN 077 elicited high frequency and magnitude binding antibody responses as determined from validated binding antibody multiplex assays (BAMA). Despite the fact that all vaccines contained only Env inserts from HIV clade A, positive BAMA results were observed against Env antigens for Figure 3. Comparison of neutralizing antibody responses in HVTN 078 and HVTN 204. Magnitude-breadth (M-B) curves, as shown here, integrate data on the neutralizing antibody response magnitude and breadth, 2 important criteria for evaluating vaccine-induced humoral responses. Neutralization response magnitude refers to the quantity of neutralizing antibodies against a single virus strain in a serum sample. This value is determined by assessing the ability of serum antibodies to block viral entry into cells in vitro. Results are plotted as the reciprocal titer of sera required to reduce the number of infectious viral particles by 50% (infectious dose 50 or ID 50). Neutralization breadth refers to the extent by which serum antibodies neutralize multiple virus strains. This value is expressed as a positive response rate indicating the fraction of virus strains from a panel that are neutralized for a given ID 50 titer. Here, M-B curves comparing neutralizing antibody responses in HVTN 078 and HVTN 204 are shown. HVTN 078 data are from the group of subjects who received only placebo (Placebo 078, white) and the subjects who received the highest priming rad5 vaccine dose (10 10 PFU) followed by 2 NYVAC boost administrations (078 rad5/nyvac, red). HVTN 204 data are from subjects who received only placebo (Placebo 204, pink) and subjects who received 3 DNA priming vaccinations followed by a rad5 (10 10 PFU dose) boost (204 DNA/rAd5, blue). Neutralization results are shown for a panel of 4 virus strains (MN.3, SF162.LS, Bal.26, and MW965.26) using TZM-bl cells. The neutralization response magnitude was determined from ID 50 values. Thin lines represent estimated subject-specific responses and solid areas represent group averages. A comparison of group averages indicates that neutralizing antibody responses in the above-mentioned treatment group of HVTN 078 were significantly stronger than responses in the HVTN 204 treatment group. This graph was adapted from a poster presented by Dr. David Montefiori at the AIDS Vaccine 2012 conference. 17 8 OCTOBER 2013 VOLUME 5:1 HVTNEWS www.hvtn.org HVTNEWS VOLUME 5:1 OCTOBER 2013 9
Probing scientific questions to advance HIV vaccine development clades B and C, as well as clade A. Furthermore, high levels of V1/V2-specific IgG binding antibody responses that were correlated with reduced HIV infection risk in the RV144 trial were also detected. 13 No significant differences were found between groups for any of the antibody responses evaluated thus far. HIV-specific CD4 + and CD8 + T-cell responses were also potently induced by all regimens, with no significant differences in response rates or response magnitude between groups. A non-significant trend toward improved CD4 + T-cell responses with DNA priming, as compared to rad35 priming, was noted. In addition, somewhat reduced, but not statistically significant, CD4 + T-cell responses to DNA/rAd35 were observed in the Ad5 seropositive group, indicating a potential hint of some cross reactivity between pre-existing immunity to one vector on a vaccine from another serotype. Taken together, the study findings address several questions regarding the use of vectors made from viruses that are less prevalent. Overall, the performance of the rad35-based vector was comparable to rad5. In addition, the rad35-vector was capable of inducing potent cellular and humoral responses even among individuals harboring pre-existing immunity to Ad5. Summary In accordance with its mission, the HVTN strives to fully characterize the safety, immunogenicity, and efficacy of HIV vaccine candidates. The studies presented above also sought to answer specific scientific questions regarding optimization of vaccine responses. Data emerging from these studies point to a variety of factors impacting vaccine response magnitude and breadth. This information may be applicable for multiple vaccine platforms, and serve to guide the field toward our goal of developing a safe and globally effective vaccine to prevent HIV infection. Acknowledgements We thank HVTN 073E, 077, 078, 083, 085, and 088 protocol team members, in particular Mary Allen, Pierre-Alexandre Bart, Carter Bentley, Stephen De Rosa, Marnie Elizaga, Ian Frank, Jonathan Fuchs, Yunda Huang, Nidhi Kochar, David Montefiori, Georgia Tomaras, and Stephen Walsh for their help in preparing this article. In addition, we thank Allan decamp, Lisa Donohue, Greg Finak, Nicole Frahm, and David Friedrich. Tracey Day is HVTN Senior Science Writer, Cecilia Morgan is HVTN Associate Director, Scientific Development. References 1. Deeks SG, Walker BD. Human immunodeficiency virus controllers: mechanisms of durable virus control in the absence of antiretroviral therapy. Immunity. 2007;27(3):406-416. 2. Buchbinder SP, Mehrotra DV, Duerr A, et al. Efficacy assessment of a cell-mediated immunity HIV-1 vaccine (the Step Study): a double-blind, randomised, placebo-controlled, test-of-concept trial. Lancet. 2008;372(9653):1881-1893. 3. Fitzgerald DW, Janes H, Robertson M, et al. An Ad5-vectored HIV-1 vaccine elicits cell-mediated immunity but does not affect disease progression in HIV-1-infected male subjects: results from a randomized placebo-controlled trial (the Step Study). J Infect Dis. 2011;203(6):765-772. 4. McElrath MJ, De Rosa S, Moodie Z, et al. HIV-1 vaccine-induced immunity in the test-of-concept Step Study: a case-cohort analysis. Lancet. 2008; 372(9653):1894-905. 5. Janes H, Friedrich DP, Krambrink A, et al. Vaccineinduced Gag-specific T cells are associated with reduced viremia after HIV infection. J Infect Dis. 2013;Aug 13. [Epub ahead of print]. 6. Frahm N. Defining the targets of HIV-specific T-cell responses (epitope mapping). HVTNews. 2011;3(1):10-11. 7. Goepfert PA, Elizaga ML, Sato A, et al. Phase 1 safety and immunogenicity testing of DNA and recombinant modified vaccinia Ankara vaccines expressing HIV-1 virus-like particles. J Infect Dis. 2011;203(5):610-619. 8. Day T, Metch B, Frahm N, Morgan C. Immunogenicity data patterns emerging from cross trial comparisons. HVTNews. 2013;4(2):1-8. 9. Keefer MC, Frey SE, Elizaga M, et al. A phase I trial of preventive HIV vaccination with heterologous poxviral-vectors containing matching HIV-1 inserts in healthy HIV-uninfected subjects. Vaccine. 2011;29(10):1948-1958. 10. Liu J, Ewald BA, Lynch DM, Nanda A, Sumida SM, Barouch DH. Modulation of DNA vaccineelicited CD8+ T-lymphocyte epitope immunodominance hierarchies. J Virol. 2006;80(24):11991-11997. 11. Day T, Morgan C, Kublin JG. Will mosaic vaccine immunogens expand immune response breadth to rival HIV-1 strain diversity? Clinical Investigation. 2013;3(5):413-415. 12. Rerks-Ngarm S, Pitisuttithum P, Nitayaphan S, et al. Vaccination with ALVAC and AIDSVAX to prevent HIV-1 infection in Thailand. N Engl J Med. 2009;361(23):2209-2220. 13. Haynes BF, Gilbert PB, McElrath MJ, et al. Immune-correlates analysis of an HIV-1 vaccine efficacy trial. N Engl J Med. 2012;366(14):1275-1286. 14. Galli G, Hancock K, Hoschler K, et al. Fast rise of broadly cross-reactive antibodies after boosting longlived human memory B cells primed by an MF59 adjuvanted prepandemic vaccine. Proc Natl Acad Sci USA. 2009;106(19):7962-7967. Innovative Research Arising from HVTN Award Programs Tracey Day HVTN award programs are designed to foster innovation, diversity, and collaboration in support of HIV vaccine development. Research conducted in conjunction with 3 different award programs was presented during a plenary session of the spring 2013 HVTN Full Group Meeting, held in Washington, DC. This invigorating session covered a wide range of novel and pertinent HIV vaccine research, which is summarized below. 15. Casimiro DR, Bett AJ, Fu TM, et al. Heterologous human immunodeficiency virus type 1 primingboosting immunization strategies involving replication-defective adenovirus and poxvirus vaccine vectors. J Virol. 2004;78(20):11434-11438. 16. Bart P, Huang Y, Frahm N, et al. rad5 prime/ NYVAC-B boost regimen is superior to NYVAC-B prime/rad5 boost regimen for both response rates and magnitude of CD4 and CD8 T-Cell responses. Retrovirology. 2012;9(Suppl 2):O72 17. Montefiori D, Huang Y, Karuna M, et al. rad5/ NYVAC-B is superior to NYVAC-B/rAd5 and is dependent on rad5 dose for neutralizing antibody responses against HIV-1. Retrovirology. 2012; 9(Suppl 2):P132. 18. Barouch DH, Kik SV, Weverling GJ, et al. International seroepidemiology of adenovirus serotypes 5, 26, 35, and 48 in pediatric and adult populations. Vaccine. 2011;29(32):5203-5209. 19. Morgan C, Bailer R, Metch B, et al. International seroprevalence of neutralizing antibodies against adenovirus serotypes 5 and 35. Presented at: AIDS Vaccine 2005; Sept 6-9, 2005; Montreal, Canada. 20. Fuchs J, Morgan C, Bart P, et al. DNA and recombinant adenovirus serotype 35 and 5 preventive HIV-1 vaccines with Env A inserts elicit cross-clade binding and V1V2 antibodies. Retrovirology. 2012;9(Suppl 2):P136. HVTN Initiative Program The HVTN Initiative Program (HIP) is a new award program supporting innovative HVTN investigatordriven research. The aim is to provide HVTN-procured specimens, clinical, laboratory, and/or behavioral data, and funds for projects addressing scientific questions that support the HVTN research agenda. Winning proposals from this inaugural cohort sought to analyze vaccine 10 OCTOBER 2013 VOLUME 5:1 HVTNEWS www.hvtn.org HVTNEWS VOLUME 5:1 OCTOBER 2013 11
Innovative Research Arising from HVTN Award Programs Innovative Research Arising from HVTN Award Programs Tracking development of the HIV-specific T helper cell responses Given the evidence that HIV-specific antibodies may decrease HIV infection risk, understanding how HIVspecific CD4 + T cells provide help for B-cell responses is of great interest. 5 To gain insight into how HIV-specific T helper cell responses evolve during an immunization series and to identify responses correlating with viral control, Dr. Mark Pilkinton and Dr. Spyros Kalams are directly comparing the CD4 + T-cell receptor repertoires in HIV controllers and HIV vaccine recipients. A hypothesis of the study is that breadth of CD4 + T-cell receptor repertoires may be a marker of a successful immune response. Thus far, a pilot study analyzing the T-cell receptor beta chain sequences of sorted HIVspecific CD4 + T cells in 2 HIV controller subjects has been conducted. HIV-specific CD4 + T cells were isolated using fluorescence activated cell sorting based on either surface activation marker expression (CD69 and CD25) or proliferation capacity upon HIV-specific peptide (Gag and Pol) stimulation. Cells were further sorted into naïve, central memory, or effector memory populations as determined from CD45RO and CCR7 expression. T-cell receptor beta chain sequences from each of these sorted populations were determined using multiplex PCR amplification and deep sequencing methods. This approach generated a large number (typically 1000 100,000) of unique sequences from most of the sorted populations. In comparing the sequences identified in each Dr. Mark Pilkinton memory CD4 + T cell subpopuinduced B- and T-cell response repertoires, and to evaluate biomarkers of response longevity. The conference session presentations provided the Network with a first glimpse of the novel research funded by this program. Specificity and function of vaccine-induced antibodies Advances in recombinant B-cell technology have made it possible to define vaccine-induced antibody repertoires in greater detail than ever before. An in-depth analysis of memory B-cell responses elicited by HIV immunogens is being conducted by Dr. Wilton Williams. In this study, HIV-specific memory B cells from HVTN 204 and HVTN 082 serum samples have been examined on a single cell level. These trials administered a DNA prime and recombinant adenovirus serotype 5 (rad5) boost vaccine regimen containing inserts from HIV clades A, B, and C (gag-pol B, env A/B/C). 1 Using specimens collected 4 weeks after the last vaccination, HIV envelope glycoprotein (Env)- specific memory B cells were captured via antigenspecific B-cell fluorescence-activated cell sorting using Group M consensus and vaccine-matched Env oligomeric proteins. Monoclonal antibodies were investigated from clonal memory B-cell cultures established from 8 trial specimens and characterized for binding specificity and functional profile. Over 90% of the Env-specific antibodies isolated reacted with the gp41 rather than the gp120 subunit of Env. The antibodies predominantly utilized immunoglobulin variable gene subfamily VH1-69 allele. In addition, these antibodies were found to be polyreactive, exhibiting binding not just to Env, but to several other antigens including self antigens and gut flora. Similar properties have been described for gp41- specific antibodies generated in acutely HIV infected patients. 2 These findings are consistent with the hypothesis that both HIV infection and vaccination with the DNA/rAd5 regimen may induce the expansion of preexisting B-cell precursors that crossreact with the Env antigen. Expansion of a cross reactive memory response generated by non-hiv antigens could explain why Dr. Wilton Williams Env-specific antibodies are often not protective, and may represent an immune evasion strategy employed by HIV. Future studies are planned to investigate this idea further using the Env gp140 immunogen in gnotobiotic mice to examine whether the absence of gut flora will result in a less gp41-dominant response to the vaccine. Biomarkers for predicting vaccine response durability Long-lived immune responses are necessary to provide durable immunity, however currently there is neither a means to predict whether vaccine-induced responses will last nor an understanding of how to generate lasting responses. To investigate how programmed cell death, or apoptosis, of B cells may contribute to waning antibody responses, Dr. Sonya Heath is examining whether markers of apoptosis and cell survival correlate with vaccine induced B-cell response durability in HVTN 065/205 and HVTN 204. HVTN 065 and 205 administered a homologous modified vaccinia Ankara (MVA) viral vector vaccine and a heterologous DNA/MVA prime/boost regimen, whereas HVTN 204 administered the DNA/ rad5 prime/boost regimen described above. B cells from peripheral blood mononuclear cell (PBMC) samples taken after the prime, after the boost, and 6 months after the last vaccination have been analyzed in this study. Flow cytometry was used to characterize memory B cell surface marker phenotype and expression of the proapoptotic marker, cleaved caspase-3, and a prosurvival marker, Bcl-2. Results thus far indicate that high expression of cleaved caspase-3 (proapoptotic) on memory B cells early after priming correlate with a decrease in memory B cell frequencies at 6 months post boost vaccination. In contrast, prevalence of Bcl-2 (prosurvival) expressing memory B cells after priming was associated with maintenance of memory B cell frequencies at 6 months post boost. The data presented by Dr. Heath demonstrate that biomarkers of cell death and survival may be assessed in vaccine trials. Preliminary results also indicated that differences may be detected across trials, Dr. Sonya Heath for example, Bcl-2 (prosurvival) expression was higher in HVTN 205 than HVTN 204 throughout the vaccination schedule. Next steps will include unblinding of vaccine and placebo specimens and correlating these results with antibody responses to determine whether these markers are predictive for antibody response longevity. A long term goal of this study is to define a phenotypic signature of long-lived vaccineinduced B cells. Comparing B-cell profiles in high vs. low risk populations There is evidence that HIV vaccines tested in efficacy trials thus far may have been less effective in groups at high risk for HIV infection than in low risk subjects. 3 Reasons for this remain unknown, but it has been suggested that high risk behavior, such as injection drug use and high risk sexual activity, can alter an individual s immune system properties in ways that result in suboptimal responses to vaccination. Dr. James Kobie To identify what differences may exist in the immune profiles of high risk populations, Dr. James Kobie is comparing the B cell compartment profiles of HVTN 203 participants with high risk and low risk behavior. HVTN 203 evaluated the immunogenicity of ALVAC and AIDVAX B/B combination regimens. 4 To discern potential alterations in B cell homeostasis, samples from 3 months after the last vaccination were subjected to a comprehensive flow cytometry-based phenotypic analysis of global B-cell populations. Preliminary results from bulk B-cell analyses indicate an increased prevalence of memory and activated B cells in high risk as compared to low risk subjects. High risk subjects also had higher frequencies of B cells specific for the Env immunogen (gp120 MN), however, these B cells also reacted with several variable loop deleted Env variants, suggesting a potential for polyreactivity. These observations point to a potential bias toward short-lived or exhausted phenotypes among bulk and Env-specific B cells in high risk subjects. Results from this ongoing study suggest that B cell homeostasis and Env-specific B-cell responses may be altered in high risk subjects, and indicate a need to further investigate immune profiles in this population. Next steps will include linking B cell memory profiles to other immunogenicity data from HVTN 203. In addition, a complete immunoglobulin repertoire analysis and characterization of Env-specific monoclonal antibodies is planned, to investigate the origin of Env-specific B cells and evaluate their potential to give rise to neutralizing antibodies. 12 OCTOBER 2013 VOLUME 5:1 HVTNEWS www.hvtn.org HVTNEWS VOLUME 5:1 OCTOBER 2013 13
Innovative Research Arising from HVTN Award Programs Innovative Research Arising from HVTN Award Programs lation, it was found that HIV-specific T-cell receptors from HIV controller subjects can be localized to specific memory compartments and that a given sequence may exist in multiple compartments (ie, central and effector memory). The next step will be to perform this analysis for HVTN 080 HIV vaccine recipients and compare the results with those found for HIV controllers. HVTN 080 evaluated a DNA vaccine (PENNVAX TM -B) delivered via intramuscular electroporation in combination with a DNA plasmid encoding IL-12 (GENEVAX TM IL-12). 6 Specimens from 2 weeks post final vaccination will be used initially. A longitudinal evaluation of CD4 + T-cell receptor repertoires is planned using prevaccination and post-2nd vaccination specimens. The objective of this proof of concept study is to determine the feasibility of performing such in depth analyses as a means to evaluate quality of vaccine-induced T helper cell responses and their impact on B-cell responses. HVTN Research and Mentorship Program The HVTN Research and Mentorship Program (RAMP) aims to attract and retain HIV vaccine researchers from underrepresented communities. The program links Black and Latino/a medical students with HVTN affiliated investigators, and sponsors short term research projects on topics aligned with HVTN research goals. Topics may focus on one of many HIV vaccine research areas from basic science to clinical to social/ behavioral aspects. The RAMP cohort presenting at the spring conference reported on their work addressing different challenges for participant recruitment and retention in HIV vaccine trials. Testing a long-chain peer referral recruitment method Black men who have sex with men (MSM) and male to female transgender women (TW) are 2 groups that are disproportionately affected by the US HIV/AIDS epidemic and yet are underrepresented in HIV vaccine trials. New recruitment approaches are needed to ensure that sufficient efficacy data are collected for populations in need of a vaccine. Long-chain peer referral is a recruitment method that has previously been effective in recruiting hard to reach populations. Angela Coombs conducted a mixed methods study to evaluate the feasibility of this approach to recruit black MSM and TW for HIV vaccine trials in San Francisco. To implement this approach, each person screened for the study was offered the opportunity to recruit 5-7 peers via a confidential coupon referral system. Formative qualitative work was performed in which several significant barriers and some motivators to HIV vaccine trial participation were identified. A key finding was that unclear eligibility criteria could result in skepticism around the trials. In terms of number of participants enrolled in a vaccine trial, this approach had limited success. Although a large number of initial seed participants were enlisted, the long chains of peer referral were not propagated. It was suggested that, unlike other HIV prevention studies that have used this approach, vaccine trials require longer term participation and also involve delivery of experimental products, which pose significant challenges. Despite the limited success of long chain peer referral observed Angela Coombs in this study, it was concluded that adaptations could be made to better accommodate the challenges inherent in vaccine trial recruitment. A modification was proposed involving 2 stage processes. In the first stage, all black MSM and TW could be recruited for multiple studies. A second stage would then recruit for specific vaccine trials among this larger pool of participants. Embedding this recruitment approach within a broad HIV vaccine trial education effort in the Black community was also recommended. Factors impacting Atlanta-area recruitment of Blacks Tequilla Pryor s project addressed the problem of the HIV/AIDS epidemic in the state of Georgia, which is ranked 5th in the US for new HIV infections, with the majority occurring in Blacks. The objective of her study was to examine cultural, spiritual, and vaccine-related factors impacting the recruitment of Blacks into HIV vaccine trials. Data were collected using semistructured focus groups with Black men and women from the Atlanta metropolitan area. Focus groups were conducted and transcriptions were coded for themes and subthemes. Demographic data were obtained from a questionnaire. Highlights from the focus group domains include the impression that the severity of the HIV/AIDS epidemic has decreased, and the continued impact of negative stigma. Cultural factors affecting recruitment were iden- tified, including a longstanding mistrust of the medical/ scientific community, lack of knowledge of HIV-related topics, and the absence of a sense of community among Blacks. Spirituality was not found to significantly affect recruitment, however several Tequilla Pryor vaccine-related factors were. These included a fear that the vaccine contains HIV, as well as general fears about side effects, complications, and safety. Participants also made suggestions for increasing the enrollment of Blacks in clinical research. Across all focus groups, similar ideas were based on providing a fun and consistent presence in the community through family-oriented community events, outreach, and social media. These findings underscore the need to persistently engage the communities most affected by the epidemic, in order to build trust and understanding regarding clinical research in general and PROGRAM NAME HOME INSTITUTION NHP ESI Carolina Herrera Imperial College, London NHP ESI Lu-Ann Pozzi Boston Children's Hospital RAMP Angela Coombs Tufts University School of Medicine RAMP Tequilla Pryor Georgia Health Sciences University RAMP HIP Aisha Scherr- Williams Spyros Kalams (Research presented by Mark Pilkinton) Keck School of Medicine of the University of Southern California HIP Sonya Heath University of Alabama at Birmingham HIP Wilton Williams Duke Human Vaccine Institute HIP James Kobie University of Rochester PROJECT LOCATION St. George's, University of London New England Primate Research Center San Francisco Department of Public Health HIV vaccine research in particular. Barriers to HIV testing and trial recruitment in Lima Studies indicate a high HIV prevalence among MSM and TW in Lima, Peru. Aisha Scherr-Williams s project focused on understanding barriers to HIV testing in these populations. Because HIV testing is frequently the entry point to treatment and preventative services, this knowledge may promote access to such services. In addition, information regarding successful and unsuccessful strategies for increasing HIV testing among MSM and TW will facilitate future recruitment of these high risk individuals into HIV vaccine efficacy trials. As a secondary analysis of a test and treat study, data was collected from 3 in-depth interviews and 2 focus groups involving HIV negative MSM and TW participants. Qualitative data was analyzed using grounded theory in which transcript codes were not prespecified, but rather were created as themes emerged from the data. Several key operational and psychosocial barriers that may limit HIV SCHOLAR MENTORS Jerome Kim and Ronald Veazey Xu Yu and Amitinder Kaur Jonathan Fuchs and Michele Andrasik PROJECT TITLE Use of Tissue Explants to Evaluate Mucosal Immune Response in NHPs and Humans An In Vivo Cytolytic CD8 + T Lymphocyte Assay for Evaluating AIDS Vaccine Efficacy Assessing the Feasibility of Long-Chain Referral for the Recruitment of Black MSM and Black Transwomen in HIV Vaccine Studies Atlanta Mark Mulligan Investigating the Motivators and Barriers to Recruitment and Retention of African Americans in Atlanta-based HIV Vaccine Trials IMPACTA, Lima Jorge Sanchez, Patricia Segura, and Jonathan Fuchs Short Message Services as an HIV Vaccine Recruitment Technique Vanderbilt University Vanderbilt University N/A Detailed Characterization of CD4 T Helper Function and T-cell Receptor Repertoire in HIV Vaccine Recipients University of Alabama at Birmingham Duke Human Vaccine Institute University of Rochester N/A N/A N/A Apoptosis of HIV Vaccine Elicited B Cell Responses as a Marker of Vaccine Durability Characterization of the Memory B Cell and Plasma Cell Repertoires with Functional Analysis of the Antibodies Generated in Response to a HIV Envelope DNA Prime and rad5 Vector Boost Vaccination Regimen Impact of High-risk Behavior on the Composition and Engagement of the HIV Specific B cell Repertoire 14 OCTOBER 2013 VOLUME 5:1 HVTNEWS www.hvtn.org HVTNEWS VOLUME 5:1 OCTOBER 2013 15