Adenovirus 5 serotype vector-specific immunity and HIV-1 infection: a tale of T cells and antibodies

Introduction The Step vaccine trial was the first human efficacy trial completed to explore whether an HIV-1 prophylactic vaccine aimed at inducing cellular immune responses could prevent infection rates or reduce viremia following infection. The vaccine, designed and manufactured by Merck Research Laboratories (Merck Research Laboratories, West Point, Pennsylvania, USA), was composed of recombinant adenovirus serotype 5 (rAd5) vectors expressing HIV-1 clade B Gag, Pol, and Nef. The high-risk populations enrolled in the Step trial were predominantly men who have sex with men (MSM) as well as heterosexual women in North and South America and Australia, and heterosexual women and men in the Caribbean. A sister trial, named Phambili, tested the same vaccine regimen in heterosexual men and women in South Africa. The rAd5 vaccine used in the Step trial was the first HIV-1 vaccine product to elicit interferon-γ ELISpot responses comparable to those raised by natural HIV infection, yet an interim review of study data revealed that the vaccine did not reduce the risk of HIV acquisition or viral load upon infection [1]. In addition, a higher number of HIV infections occurred in the vaccine group compared to placebo recipients. Enrollment in Phambili and vaccinations in both trials were halted upon the release of these findings, and all the participants were unblinded to their vaccination status. Early exploratory analyses in the Step trial revealed that the increased number of infections appeared to be restricted to MSM who had positive Ad5 neutralizing antibody (nAb) titers at baseline, or were uncircumcised, or both [1]. Vaccine participants who were seronegative and circumcised at baseline experienced comparable HIV incidence versus placebo [1]. The number of HIV infections observed in placebo recipients was similar between Ad5-seropositive and Ad5-seronegative individuals [1]. See Table 1 for a timeline of key events associated with Step and Phambili [1,2,3,4,5].Table 1: Step and Phambili studies: event history.Potential explanations for an increased number of infections in Step trial Two prominent hypotheses have emerged to explain the observed trend of increased HIV infections among some vaccinated Step participants: the first suggests that rAd5 activates memory Ad5-specific CD4 T cells in Ad5-seropositive individuals, expanding the potential targets for incoming HIV virions; the second suggests that preexisting nAb to Ad5 can form immune complexes with an rAd5 vaccine vector and promote infection of target CD4 T cells with HIV. Upon release of the Step trial results, it was thought that preexisting Ad5 nAb titers provide a surrogate measure for Ad5-specific T cells and that immunization with rAd5 vector could lead to activation and expansion of memory Ad5-specific CD4 T cells in Ad5-seropositive individuals. This hypothesis has now been refuted by several groups who have shown a lack of correlation between Ad5 nAb titers and both CD4 and CD8 T-cell immunity prior to vaccination, at least when cellular responses are analyzed by ELISpot or intracellular cytokine staining using Ad5 peptides or empty vectors as stimulants [6,2,3,4]. Indeed, Ad5-specific CD4 T-cell responses are present in the majority of Ad5-seronegative individuals, and CD8 T-cell responses, though less frequent, are also easily detectable in this group [6,2]. Therefore, Ad5 nAb titers do not represent a surrogate marker for Ad5-specific CD4 or CD8 T-cell responses in unvaccinated individuals on the basis of exogenous peptide-based cytokine responses. This finding is also apparent in the comparable magnitude of Ad5-specific T-cell responses in Ad5 nAb-seropositive individuals versus Ad5 nAb-seronegative individuals [6,2,3,4]. It is possible that more sensitive assays may show a correlation of Ad5 nAb titers with T-cell responses, but extended time-dependent assays are prone to artifacts of in-vitro manipulation and difficult to validate for use with clinical specimens. With regard to the second hypothesis, a recently published paper found that preexisting nAb to Ad5 can form immune complexes with the rAd5 vaccine vector [7] and promote infection of target CD4 T cells with HIV [8]. These immune complexes are significantly more efficient in activating dendritic cells than either the serum or Ad5 alone and the dendritic cells activated with these immune complexes render CD4 T cells more susceptible to HIV replication in vitro[7]. While the biological significance of the observed three-fold increase in p24 production in vitro remains to be determined, no alternative explanations for a direct effect of nAbs on HIV infection rates have been published to date. The difference between vaccine-induced antibodies and nAbs induced following natural exposure to Ad5 in this context requires further investigation [9]. If such types of mechanisms operate, this would significantly challenge the future use of rAd viruses and potentially other viral vectors as vaccine or therapeutic interventions within HIV susceptible populations [10]. Adenovirus serotype 5-seropositive status does not impact HIV-1 incidence in placebo recipients Nearly all Ad5-seronegative Step participants became Ad5-seropositive after rAd5 vaccination, yet the incidence rate of HIV infection in baseline Ad5 seronegatives did not increase after vaccine administration. As expected with a potent vector, antivector nAb titers increased after one [2], two [4], or three [3] vaccine administrations. This increase was seen in participants who were either Ad5-seronegative at baseline or those with preexisting humoral immunity, but it was more pronounced in the latter group [4]. This finding remained constant over a 78-week period of follow-up in both vaccine and placebo recipients [8]. It was also observed that placebo recipients with high nAb (>1000 titer) had a lower incidence of HIV-1 infection compared to placebo recipients with low nAb (<18 titer). Incidence ranged from 4% a year in those with the lowest Ad5 nAb level to 1.2% a year in those with the highest nAb level [1]. However, other baseline characteristics of the Ad5-seropositive population differed substantially from those of the Ad5-seronegative population: Ad5-seronegative participants were recruited primarily from North American cohorts, who reported riskier sexual practices and substance use that may associate with an increased risk of HIV-1 acquisition. Interestingly, the specificity of the vaccine-induced nAbs seems to be qualitatively different from the preexisting antibodies: naturally elicited antibodies are targeted primarily to the Ad5 fiber alone or to other viral capsid proteins in addition to the fiber, whereas exposure to rAd5 through vaccination elicited antibodies primarily to capsid proteins [9]. Vaccination does not lead to preferential expansion of adenovirus serotype 5-specific CD4 T cells in adenovirus serotype 5-seropositive trial participants Recent published studies show that regardless of the source of rAd5 vector, vaccination with rAd5 vector leads to expansion of Ad5-specific CD4+ T cells in Ad5-seronegative trial volunteers, whereas no increases of Ad5-specific CD4 T-cell responses were detected in seropositive participants using peptide-based cytokine quantitation assays [6,2,4]. This observation reinforces the early analysis by intracellular cytokine staining of systemic Ad5-specific T-cell responses from Step specimens, which showed (although in small numbers) that Ad5-specific T-cell responses occurred more frequently in matched controls than in the HIV-infected cases [5]. These data are contrary to the hypothesis that vaccination with rAd5 vectors specifically boosts Ad5-specific CD4 T cells in individuals with preexisting immunity, thereby providing an expanded susceptible target cell population for HIV-1. A recently published report showed a significant in-vitro expansion of CD4 T cells after Ad5 vector stimulation of lymphocytes obtained from healthy volunteers. These expanded cells exhibited elevated expression of the gut-homing molecules α4β7 and CCR9, with significantly higher α4β7 expression in individuals with preexisting Ad5 immunity. Moreover, CD4 T-cell proliferation and interferon-γ production in response to Ad5 vector stimulation correlated with Ad5 antibody titers, though there was no correlation between Ad5 serostatus and Ad5-stimulated cytokine production in a peptide ELISpot assay [11]. The authors hypothesize that in the Step trial, individuals with preexisting humoral Ad5 immunity would respond with elevated expression of mucosal homing integrins to the rAd5 vaccine vector and that these primed T cells would migrate to mucosal surfaces, where they would serve as targets for HIV. An earlier published report also showed an association between CD4 T-cell proliferation and Ad5 nAb in healthy donors [12]. These papers are contrary to the reports from the Barouch and Betts laboratories that investigated specimens from rAd5 recipients and showed that it was Ad5-seronegative, not Ad5-seropositive participants who demonstrated the largest increase in Ad5-specific T cells. They also showed that types of cytokines and chemokines produced by the CD4 T cells after vaccination in Ad5-seropositive versus Ad5-seronegative individuals did not differ in their expression profiles and therefore they would not have migrated differentially to mucosal portals [3,4]. Moreover, the investigators did not see an upregulation of mucosal targeting receptors in the systemic samples obtained from Ad5-vaccinated humans. It is possible that the Barouch and Betts studies may not have detected cells expressing the mucosal homing markers as their analysis was confined to peripheral blood samples that may only transiently, at low levels, contain cells expressing mucosal homing markers and thus would remain undetectable by the assays used in these investigations. A second possibility is that expression levels of α4β7 and CCR9 on the systemic samples evaluated in these studies may have been elevated, but returned to baseline values at the protocol-specified blood draw time points. Alternatively, the observations in the study by Benlahrech et al. [11] may be restricted to the in-vitro setting used to assess the changes in the expression of mucosal homing molecules in peripheral blood mononuclear cells (PBMCs) from unvaccinated individuals. Interestingly, circulating CCR5-expressing activated CD4 T cells were more abundant in people with high Ad5 nAb titers, which could have increased target cell susceptibility to HIV-1 after exposure at mucosal sites, though detectable amounts of these cells were not increased in the Step trial vaccine group compared with the placebo group [5]. Clearly, the degree to which a systemically administered vaccine induces immune responses in both the periphery and mucosa at various times after immunization over a prolonged period of time is necessary to resolve these disparate observations and will be a focus of activity during upcoming vaccine clinical trials. Does route of immunization and recombinant adenovirus serotype 5 vector matter? The Step vaccine was given by intramuscular injection to individuals with very low or preexisting immunity to Ad5. In contrast, natural exposure to Ad5 is usually through targeting nasopharynx or gut and can persist at mucosal surfaces over several years [13]. It is conceivable that intramuscular vaccination may perturb the balance between wild-type Ad5 persistence and immune pressure resulting in activation of Ad-specific cells that traffic to mucosal tissues. It is also possible that vaccine recipients with preexisting Ad5 immunity might generate an Ad5-specific immune response that homes to mucosal surfaces, whereas vaccine-induced Ad5-specific T cells in the - preexisting nAb might exhibit a different homing phenotype. Unfortunately, these possibilities cannot be resolved with the existing Step samples as peripheral specimens would fail to detect potential differences between Ad5-specific T cells of Ad5-seronegative and Ad5-seropositive vaccinees present at mucosal surfaces. Interestingly, the effect of rAd5 immunization on the cellular immune response appears to vary depending on the vector used. The MRL vaccine (which has a deletion in the E1 adenovirus gene) invariably leads to significantly increased magnitudes of Ad5-specific CD4 and CD8 (where measured) T-cell responses in trial participants who are Ad5-seronegative at baseline [6,3,4]. When larger numbers of specimens are evaluated [6,4], this increased magnitude is blunted for both CD4 and CD8 T cells in the presence of preexisting nAb titers. Contrary to these observations, no significant increase in the magnitude of Ad5-specific CD4 T-cell responses was observed following vaccination in either seropositive or seronegative participants enrolled in a phase I study of the National Institute of Allergy and Infectious Diseases (NIAID) Vaccine Research Center (VRC) designed rAd5 vector (which includes deletions of the Ad5 E1 and E4 regions with a partial deletion in the E3 region), though a significant increase in hexon-specific CD8 T cells is reported for seropositive participants [2]. This discrepancy between the studies using different vectors could in part be explained by the reduced expression of Ad5 structural and nonstructural proteins in cells infected with the VRC E4-deleted vector compared to an E4 wild-type vector, as well as the investigation of fewer specimens [2]. If recombinant adenovirus serotype 5 platforms are restrictive in their applicability, what next? Preexisting Ad5-specific nAbs have been reported to reduce the immunogenicity of rAd5 vaccines, possibly by restricting efficient vector transduction of target cells and limiting expression of vaccine antigen [8]. This dilemma may be relevant in the developing world where infections with adenoviruses are common. In one study of young adults in the United States, 55% of individuals had Ad5 antibody titers more than 1: 20 [12]. In sub-Saharan Africa, Ad5 seroprevalence was substantially higher and in certain populations was more than 90% [14,15]. Two strategies to circumvent preexisting immunity include development of vectors from rare human adenovirus serotypes as well as from adenovirus species that infect nonhuman hosts. The first strategy explores the use of serologically distinct rAd vectors that are rare in humans as vectors in heterologous rAd prime-boost vaccine regimens [16–18]. Several serologically distinct rAd vectors are under development, including rAd26, rAd35, and a chimeric rAd5HVR48 [19]. Of these, the rAd26-based vector has just completed enrollment in a phase I study on humans, whereas the phase I rAd35 and the chimeric rAd human trials are ongoing. It will be key to observe whether HIV gene inserts in alternate adenovirus serotypes are as immunogenic in humans as rAd5. Preclinical studies on monkeys have shown that heterologous prime-boost regimen of rAd26/rAd5 elicits improved magnitude, breadth, and quality of T lymphocyte responses compared to a homologous Ad5 regimen [20]. A concern remains that even in Ad5 nAb-seronegative individuals, preexisting nAbs to rare adenovirus serotypes could crossreact with rAd5 vaccine and enhance HIV-1 acquisition after rAd5 vaccination. With the novel and chimeric rAd vaccines, it will be crucial to see whether the safety issue is circumvented. The extensive targeting of Ad5 by T cells in the absence of a documented exposure to this virus (as deduced by the lack of positive nAb titers) may occur because of the extensive homology of the human adenovirus serotypes [21]. Although over 50 distinct human adenovirus serotypes have been defined based on type-specific nAbs, sequence homology, especially within each of the six subgroups A–F, is extremely high. Ad5 belongs to subgroup C, which also comprises the commonly occurring serotypes Ad1 and Ad2 [22]; therefore, infection with Ad1 or Ad2 could explain the presence of adenovirus-specific T cells in Ad5-seronegative hosts. In product development, it will be important to assess the degree of T-cell crossreactivity among various adenovirus serotypes with and without preexisting Ad5 immunity. Such results could help determine the potential utility of novel rAd vectors for HIV-1 vaccine development. Recent data from the HIV Vaccine Trials Laboratory (HVTN) indicate that Ad5-specific T-cell responses are observed against regions of Ad5 that are conserved across multiple adenovirus serotypes [23]. Crossreactive adenovirus-specific T cells have been studied in the past, mostly in the context of broadly protective adoptive immunotherapy in the treatment of individuals undergoing stem cell transplantation [24–27]; but only recent extensive Ad5 epitope mapping studies using Ad5 peptides have shown that the regions targeted by adenovirus-specific T cells are highly conserved across adenovirus serotypes [23]. Hence, cellular immune responses generated against any adenovirus (e.g. the seroprevalent Ad1 or Ad2, but also more distantly related viruses such as Ad26 or Ad35) can be detected using Ad5 peptides as a screening tool. However, these non-Ad5 T-lymphocyte responses do not seem to get boosted following rAd5 vaccination [4]. The second strategy to circumvent preexisting immunity involves the use of adenoviruses derived from nonhuman serotypes as vaccine vectors [25,26]. In a recently published nonhuman primate study, investigators tested two chimpanzee adenovirus (AdC) vectors, AdC6 and AdC7 expressing Gag of HIV-1, in a heterologous prime-boost regimen, in comparison to the human Ad5 vector in a homologous two-dose prime-boost regimen. In both immunization protocols, the nonhuman primates were divided into naive and AdHu5 preexposed cohorts. The results of this study showed that AdC induces potent Gag-specific CD8 T-cell responses and that these responses were substantially increased by booster immunizations with serologically distinct adenovirus vectors as opposed to a homologous vaccine regimen similar to that used in the Step trial [28]. trials with rare vectors are human trials of AdC vectors that and such as a phase I trial of a chimeric multiple that is composed of the antigen to a including T-cell and The vaccine is the by a heterologous The regimen was well in healthy volunteers. This same regimen was recently evaluated in a phase vaccination and challenge study on healthy volunteers. The results of this trial showed efficacy in vaccinated individuals, with some immunity. there are clinical trials to a prophylactic vaccine for virus is a phase I trial to the safety and immunogenicity of the heterologous with two different adenovirus and a including the nonstructural of with This latter vaccine is and very well at all tested vectors were with the highest resulting in of and extremely potent and CD4 and CD8 T-cell responses An for the potential of HIV-1 acquisition in Ad5-seropositive who the Step vaccine was that following Ad5-specific CD4+ T lymphocyte responses expanded and a potential of target cells for HIV. However, recent from laboratories that Ad5-specific CD4 and CD8 T-cell responses are present in individuals with undetectable levels of Ad5 nAbs In in the Step study, circulating Ad5-specific CD4 T cells were detected in the majority of placebo of their Ad5 neutralizing Ad5-specific responses were detected at baseline in Ad5 trial participants in the other studies it is that Ad5-specific T-cell responses induced by the vaccine were for the increased susceptibility to HIV infection. Moreover, these data also show that Ad5 nAbs at baseline are not a surrogate marker for Ad5-specific cellular immune responses. the Step trial results have raised the vaccine and to will that circumvent the safety by Ad5 and are to the Step regimen. An alternative for the HIV-1 acquisition is that of the Ad5 vector in Ad5-seropositive men who were uncircumcised, or men might an as yet effect on the immune response that leads to an increase in HIV-1 acquisition. the of this vaccine to other HIV vaccines, including those using alternative is not yet adenovirus serotypes and may be than rAd5 vectors for use in humans, yet their future potential may only be by human efficacy strategy to vaccine may be to screening with of HIV-1 viral load at as the to for efficacy In the analysis of all the preexisting immunity to in and of and in the absence of rAd5 vaccine, does not to be associated with an increased risk of HIV infection. The increased numbers of HIV-1 infections are only apparent in vaccine recipients with high nAb titers to Ad5. in the absence of vaccine, prior exposure to Ad5 is to the risk of HIV upon The the of vaccine in men is one hypothesis might be that vaccine-induced T-cell responses traffic from injection to the mucosal of the activated CD4 T cells could dendritic cells and a for HIV infection during the early after vaccination, similar to what has been recently for virus infection participants with high baseline Ad5 titers may generate an immune response to rAd5 vaccine and this may further enhance HIV infection of target cells in the Unfortunately, it is to such a hypothesis as mucosal specimens were not during the Step However, Step study results the to and lymphocytes at mucosal versus blood between participants with and without preexisting Ad5 immunity. mucosal immune responses The increased of mucosal immune responses is important in the of the extensive crossreactivity of adenovirus-specific T as the lack of nAb to the rare serotypes does not an absence of cellular responses to the Although no has been between the presence of adenovirus-specific CD4 or CD8 T cells and HIV the to a biological of increased risk of HIV acquisition the of immune responses for any especially those with the potential to crossreact with commonly occurring that is with and has a which over the year will focus on two and specimens for mucosal immune responses in the lower and and developing to assay the immune responses in these specimens. that this will a and of mucosal cellular immunity using well and assays that are to vaccine and study In the results of the Step study have raised the and among the future use of adenovirus vector rAd5 that has from a highly immunogenic vector with an safety to one the concern is product efficacy studies such as in which the VRC vaccine regimen is a by a rAd5 will only the for which there was no of increased risk for HIV acquisition circumcised men who are Ad5 the for the observed increase in the number of HIV infections in vaccine recipients of the Step trial is clinical trials of rAd5 HIV vaccine include to potential risk of study to a by restricting study enrollment to subgroups with no of increased study for an increased risk of HIV and risk of trial In addition, novel which elicit immune responses that are to the homologous rAd5 regimen and which can be evaluated both systemically and are a high for the HIV-1 vaccine This was by the authors in their as of and Research but the in this paper do not represent those of and Research and for and during the of the has any of to