Adenovirus Serotype 5-Specific Neutralizing Antibodies Target Multiple Hexon Hypervariable Regions

ABSTRACT The immunogenicity of adenovirus serotype 5 (Ad5) vectors has been shown to be suppressed by neutralizing antibodies (NAbs) directed primarily against the hexon hypervariable regions (HVRs). We previously reported that replacing all seven HVRs with those from the rare serotype virus Ad48 resulted in a chimeric Ad5HVR48(1-7) vector that largely evaded preexisting Ad5 immunity in mice and rhesus monkeys. In this study, we evaluated the extent to which Ad5-specific NAbs are directed against various HVRs. We constructed partial HVR-chimeric Ad5 vectors with only a subset of HVRs exchanged, and we utilized these vectors in both NAb assays and murine immunogenicity studies with and without baseline Ad5 immunity. Our results demonstrate that Ad5-specific NAbs target multiple HVRs, suggesting that replacing all HVRs is required to optimize evasion of anti-Ad5 immunity. These data have important implications for the development of novel vectors for both vaccines and gene therapy.

R ecombinant, replication-incompetent adenovirus serotype 5 (Ad5) is a highly immunogenic vector that elicits vigorous immune responses against foreign transgenes from multiple pathogens (17,18). However, preexisting Ad5 immunity in human populations has been shown to suppress the immunogenicity of Ad5-based vectors in both preclinical studies (13,16,19,20) and clinical trials (3,11). The majority of Ad5-specific neutralizing antibodies (NAbs) are directed primarily against the hexon protein, although NAb responses have also been reported against the fiber protein (7,8,19,22). Each virion consists of 240 hexon homotrimers (14), and recently determined high-resolution cryoelectron microscopy (cryo-EM) and X-ray crystal structures reveal interhexon associations as well as interactions with minor capsid proteins (9,12). Sequence variability among adenovirus serotypes is concentrated in the seven loops located at the solventexposed surface of the hexon, termed hypervariable regions (HVRs) (5), which have been shown to harbor major neutralizing determinants (19,22).
We previously reported that replacing all seven hexon HVRs in Ad5 with those from a rare human adenovirus serotype, Ad48, resulted in a chimeric vector, Ad5HVR48 (1)(2)(3)(4)(5)(6)(7), that evaded the majority of preexisting Ad5 immunity in preclinical studies in mice and rhesus monkeys (13). However, the relative importances of the seven individual HVRs as NAb epitopes remain incompletely understood, and recent studies have suggested that Ad5 NAb responses may actually be focused primarily on one specific HVR, such as HVR1 or HVR5 (1,15). In this study, we characterized the contribution of individual hexon HVRs as Ad5 NAb epitopes. We constructed chimeric Ad5 vectors in which only subsets of HVRs were exchanged and evaluated these vectors in both NAb assays and immunogenicity studies.
In mice with high baseline Ad5 NAb titers (median log NAb titer, 4.0), which model the upper bound of Ad5 NAb titers in Africa (2), the immunogenicity of Ad5-SIV Gag was abrogated, as expected, while the immunogenicity of Ad5HVR48(1-7)-SIV Gag was largely preserved (Fig. 2B), consistent with our previously reported findings (13). Importantly, the immunogenicity of both Ad5HVR48(1)-SIV Gag and Ad5HVR48(1-3)-SIV Gag was also suppressed (Fig. 2B). These data indicate that exchanging just HVR1 to HVR3 was insufficient to circumvent high levels of preexisting Ad5 immunity, consistent with the intermediate reduction of NAb titers to the Ad5HVR48(1-3) vector in virus neutralization assays in vitro ( Fig. 1B and C). These data suggest that Ad5 NAb epitopes that are relevant in vivo for suppression of vector immunogenicity likely include multiple HVRs and that Ad5HVR48(1-3) does not adequately evade Ad5 NAb responses in vivo.
To evaluate in greater detail the role of HVR4 to HVR7 as neutralization determinants, we attempted to produce the Ad5HVR48(4-7) vector, but this vector proved nonviable (Fig. 1A). We then attempted to generate Ad5HVR48(4), Ad5HVR48(5), Ad5HVR48(6), and Ad5HVR48 (7) vectors in which only single HVRs were exchanged (Fig. 3A). Of these, only the Ad5HVR48(4) and Ad5HVR48(5) vectors could be produced to high titers (data not shown). We evaluated the immunogenicity of Ad5HVR48(4)-SIV Gag and Ad5HVR48(5)-SIV Gag vectors in studies similar to those described in the legend to Fig. 2. Naïve and Ad5-preimmunized C57BL/6 mice (n ϭ 8/group) were immunized once with 10 9 vp of Ad5-SIV Gag, Ad5HVR48(4)-SIV Gag, and Ad5HVR48(5)-SIV Gag vectors. These vectors proved comparably immunogenic in naïve mice (Fig. 3B), but all these vectors were suppressed in mice with anti-Ad5 immunity (Fig. 3C), indicating that the Ad5HVR48(4) and Ad5HVR48(5) vectors were unable to cir-cumvent preexisting Ad5 immunity. Although we were not able to test the roles of HVR6 and HVR7 individually, it is unlikely that Ad5-specific NAbs are directed entirely against these HVRs, given the partial evasion of Ad5 NAbs in vitro by the Ad5HVR48(1-3) vector. Thus, these data are consistent with a model in which Ad5-specific NAbs are directed against multiple HVRs, suggesting that simultaneous exchange of all HVRs is optimal to evade preexisting Ad5 immunity.
The seroprevalence of AdC68 has been reported to be low in humans (4,6,21).
Ad5-specific antibodies have been shown to be directed against multiple capsid proteins, including fiber, penton, and hexon (8,19), but dominant NAb responses that are critical for suppressing Ad5 vectors appear to be directed largely against the hexon HVRs (19,22). However, the precise targets of Ad5 NAbs remain unclear. In the present study, we assessed the contributions of the seven HVRs as targets for Ad5 NAbs. We built on prior studies from our group that showed that an Ad5HVR48(1-7) vector in which all seven HVRs were exchanged with those from the rare shown as black bars. Ad5HVR48(4) and Ad5HVR48(5) were viable and could be produced to high titers, whereas Ad5HVR48(6) and Ad5HVR48(7) proved nonviable. Naïve (B) or Ad5-preimmunized (C) C57BL/6 mice were immunized i.m. with 10 9 vp of the chimeric vectors Ad5, Ad5HVR48(4), and Ad5HVR48(5) expressing SIV Gag. AL11-specific tetramer responses at multiple time points, IFN-␥ ELISPOT assays using splenocytes on day 28, and ICS assays using splenocytes on day 28 were utilized to measure SIV Gag-specific cellular immune responses. serotype vector Ad48 largely evaded Ad5 immunity in both mice and rhesus monkeys (13). In the present study, we observed that swapping only HVR1 to HVR3, HVR4, or HVR5 individually was insufficient to evade preexisting Ad5-specific immunity in mice, whereas swapping all seven HVRs with those from the coxsackievirus and adenovirus receptor (CAR)-binding virus AdC68 recapitulated the phenotype observed with Ad5HVR48(1-7). Consistent with these results, NAb titers to Ad5HVR48(1-3) were significantly lower than those to Ad5 but were higher than those to Ad5HVR48(1-7) in sera from vaccinated mice and humans with natural Ad5 immunity. These data indicate that Ad5-specific NAbs target multiple HVRs, including epitopes located in both HVR1 to HVR3 and HVR4 to HVR7. Thus, swapping all HVRs simultaneously is likely required for optimal evasion of antivector immunity.
A technical challenge in the present study was the fact that many chimeric vectors proved nonviable. Recent studies have reported the crystal and cryo-EM structures of the Ad5 capsid at 3.5-Å resolution, providing a detailed picture of the structural constraints of the hexon HVRs, including substantial contacts with minor capsid proteins (9,12). These data suggest that there are both structural and biochemical constraints on the HVRs that limit the ability to manipulate these regions, and high-throughput methods to determine the viability of HVR-exchanged Ad vectors are lacking. Development of such techniques would accelerate the generation of HVR-chimeric vectors and prime-boost regimens.
In summary, our findings indicate that multiple HVRs are targeted by both natural and vaccine-elicited Ad5 NAbs and that vectors with complete HVR swaps outperform those with partial HVR swaps in terms of evasion of vector-specific NAbs. Moreover, we show that a novel Ad5HVRC68(1-7) vector recapitulates the ability of Ad5HVR48(1-7) to evade Ad5 immunity in mice. These findings add to our understanding of the humoral immune response to adenoviruses and contribute to our understanding of capsid-chimeric vectors and prime-boost regimens for both vaccine development and gene therapy.