Adenoviruses are common in nature and are a major cause of severe colds. During the past decade, attenuated forms of adenovirus have become increasingly utilized as experimental vaccine vectors due to their targeting of dendritic cells and ability to induce CD8 T cell responses in the majority of recipients (no other vector to date has shown comparable CD8 T cell immunogenicity). Adenoviruses are categorized into a number of different serotypes, with serotype 5 (Ad5) being among the most prevalent. An HIV vaccine candidate that used Ad5 as a vector was developed by Merck, but failed to show efficacy and significantly enhanced HIV acquisition risk in a subset of trial participants with pre-existing antibody responses against natural Ad5.
Since the Merck data were published, there has been debate about the mechanism by which the Ad5 vector could have increased the risk of HIV acquisition, and even claims that the result was a statistical fluke or represented a confounding effect related to circumcision status (the former claim is possible, the latter was contradicted by a multivariate analysis showing that both lack of circumcision and receipt of the Ad5 vaccine were associated with an increased likelihood of HIV acquisition). Perhaps most controversially, it has been suggested that perhaps the Merck vector boosted Ad5-specific CD4 T cell responses, thus providing more activated CD4 T cell targets for HIV infection. Two studies that measured Ad5-specific CD4 T cell responses in recipients of the Merck vaccine based on cytokine production have argued against this possibility (see here and here), while another that assessed Ad5-specific CD4 T cell responses based on their ability to proliferate suggested that CD4 T cell-mediated enhancement could have played a role.
Lurking on the sidelines of this contentious debate has been the issue of natural adenovirus infection, about which relatively little is known. Now a study by Larry Corey’s research group, published recently in PLoS One, has for the first time evaluated natural adenovirus infections among one of the populations enrolled in the Merck HIV vaccine trial, Peruvian men who have sex with men (MSM).
The results are perhaps surprising: the majority of the 20 men studied had detectable adenovirus DNA in rectal swabs during follow-up; the average rate of detection at any given timepoint was in 30% of samples. A variety of adenovirus serotypes were represented, including Ad5, Ad26 and Ad48, all of which are in development as vaccine vectors. The study authors note that: “shedding typically occurred on consecutive days in clustered episodes lasting a median of 4 days (range 1 to 9 days) separated by periods without shedding, suggesting frequent new infections or reactivation of latent infections over time.”
The study raises a number of issues. First and foremost, it suggests that when a pathogen is being developed as a potential vaccine vector, a thorough understanding of the natural infection should be obtained early on in the research process, and not after the fact. In the case of adenovirus vectors, their success in solving the difficult problem of inducing CD8 T cell responses may have caused the lack of information on natural infections to be somewhat overlooked. In terms of specific relevance to the Merck trial, the study implies that interactions between adenovirus-specific immunity and natural adenovirus infections are likely to be far more dynamic and ongoing than had previously been surmised. The boosting of Ad5-specific CD4 T cell responses by vaccination might therefore have different consequences depending on whether natural adenovirus infection is present, particularly given that the virus localizes in the mucosa and Ad5-specific CD4 T cell responses cross-react with a wide variety of serotypes. Additional studies will now be needed to investigate the impact of adenovirus-based vaccines on mucosal virus-specific CD4 T cell responses in the presence and absence of natural infection.
PLoS One. 2010 Jun 25;5(6):e11321.
Curlin ME, Huang ML, Lu X, Celum CL, Sanchez J, Selke S, Baeten JM, Zuckerman RA, Erdman DD, Corey L.
Fred Hutchinson Cancer Research Center, Seattle, Washington, United States of America.
BACKGROUND: The association between baseline seropositivity to human adenovirus (HAdV) type 5 and increased HIV acquisition in the Step HIV Vaccine Study has raised questions concerning frequency of acquired and/or persistent Adenovirus infections among adults at high risk of HIV-1 infection. METHODOLOGY: To evaluate the frequency and pattern of HAdV shedding from the lower GI tract, we retrospectively tested rectal swabs for HAdVs in a cohort of 20 HSV-2 positive HIV-positive Peruvian men who have sex with men (MSM) undergoing rectal swabbing three times/week for 18 consecutive weeks, in a prospective study of HSV-2 suppression in HIV infection. Viral DNA was extracted and amplified using a sensitive multiplex PCR assay that detects all currently recognized HAdV types. Molecular typing of viruses was performed on selected samples by hexon gene sequencing. Baseline neutralizing antibody titers to HAdVs -5, -26, -35 and -48 were also assessed. PRINCIPAL FINDINGS: 15/20 individuals had HAdV detected during follow up. The median frequency of HAdV detection was 30% of samples (range 2.0% to 64.7%). HAdV shedding typically occurred on consecutive days in clustered episodes lasting a median of 4 days (range 1 to 9 days) separated by periods without shedding, suggesting frequent new infections or reactivation of latent infections over time. 8 of the 15 shedders had more than one type detected in follow-up. 20 HAdV types from species B, C, and D were identified, including HAdV-5, -26 and -48, HAdV types under development as potential vaccine candidates. 14/20 subjects were seropositive for HAdV-5; 15/20 for HAdV-26; 3/20 for HAdV-35; and 2/20 for HAdV-48. HAdV shedding did not correlate with CD4 count, plasma HIV-1 viral load, or titers to HAdV-5 or HAdV-35. The sole individual with HAdV-5 shedding was HAdV-5 seropositive. CONCLUSIONS: HAdV shedding was highly prevalent and diverse, including types presently under consideration as HIV vaccine vectors. Subclinical HAdV infection of the GI tract is common among MSM in Peru; the prevalence of HAdV in the enteric tract should be evaluated in other populations. The association between ongoing recent enteric HAdV and the immune response to recombinant HAdV vaccines should be evaluated.