Series of Superinfection Reports Leave Researchers Debating Not “If” but “How” — And What It All Means
“Extrapolating Implications”
The possibility of multiple HIV infections in a given individual has concerned both people living with HIV and researchers for many years. While the theoretical case for such an occurrence has been argued, convincing evidence has until recently been notoriously elusive. The report this past July at the Barcelona AIDS conference of a re-infection, or “superinfection,” in study patient of a Harvard cohort was initially greeted with more than a little circumspection. Then, in August and September, came three additional reports — reviewed and published in none other than the
New England Journal of Medicine and Journal of Virology. Suddenly re-infection is for real, and it has vaccine researchers and public health officials particularly dismayed. Richard Jefferys dissects the data and explains why our worst fears may not (yet) be realized.
Amidst a cacophony of information about the ever-expanding worldwide HIV epidemic, perhaps the most widely covered story from the XIV International AIDS Conference in Barcelona revolved around a single individual. Patient AC-06, a participant in a small Boston-based study of treating acute HIV infection with antiretroviral drugs, was reported to have been re-infected (or “superinfected”) by a second, slightly divergent HIV strain (belonging to the same B subtype, or “clade,” as the original infecting virus) during the trial. Principal investigator Bruce Walker divulged the data during a morning session on immunity to HIV infection, prompting French doyen of immunology Brigitte Autran to exclaim: “Bruce, this is terrible news.”
The theoretical possibility of dual (or even multiple) HIV infections has concerned both people living with HIV and researchers for many years, but convincing evidence that such superinfections can occur in humans has been notoriously hard to come by. The last time a potential case hit the headlines was in the year 2000, when a research team from Canada reported (at the annual Retrovirus Conference) that an asymptomatic individual with HIV had picked up a more virulent virus from his partner, resulting in a sudden immunologic decline and disease progression. This case, however, never made it into the medical literature. And it later transpired that the results were erroneous and most likely due to a laboratory contamination. The story fizzled out so quietly that many people never even realized that the data had been debunked.
Given this history, it is perhaps not surprising that new reports of re-infection are typically greeted with a dose of skepticism, and Bruce Walker’s case is no exception (“they need to check those results with a good virologist” was one comment heard in the conference corridors). However, synchronous with Walker’s data come two additional published reports of possible re-infections, the very first batch to make it through peer review and into the medical literature. The first paper, authored by Thai and U.S. collaborators, was published in the August issue of the Journal of Virology and reports on two individuals in a Thai cohort of intravenous drug users that acquired a second strain of HIV within months of their initial infection. The second report (also presented in Barcelona, and subsequently published in the New England Journal of Medicine September 5) comes from investigators involved in a European study of acute infection treatment known as QUEST, and comprises data on a single individual who became infected with a second HIV strain while participating in the trial (for details of these two published studies, including the strains of HIV involved, see “Re-Infection Data Summarized” in this issue). This accumulation of new evidence is giving even hardened re-infection skeptics pause, and the discussion has begun to shift from questioning the data to debating its implications for both vaccine research and the public health messages being delivered to individuals living with HIV.
In terms of vaccine research, the immediate question appears simple: if infection with HIV cannot induce sufficient immunity to protect against other HIV strains — even those from the same genetic subtype — how on earth is a vaccine going to achieve broad protection against HIV infection? For some, the new re-infection data swigs another gulp from a glass already seen as half-empty. Dissecting out legitimate questions regarding potential differences between virus- and vaccine-induced immune responses is difficult, particularly since the underlying immunology is complex and still poorly understood. Nevertheless, there may be reasons to be cautious about extrapolating implications for vaccines from these reports.
Specifically, there are two key issues relating to immune responses in the individuals that became re-infected. One relates to quality: how well were the HIV-specific immune responses actually functioning? The second relates to targeting: which parts of the virus were being targeted by the HIV-specific immune response?
Immune Response: Quality
In the case of Walker’s patient AC-06, high levels of HIV-specific CD8 T cells could be detected prior to the apparent re-infection. These responses were assessed using an ELISpot assay that measures production of the cytokine interferon-gamma by CD8 T cells in response to stimulation with epitopes (small slices of protein) derived from HIV. Prior studies have demonstrated that responses measured using this technique do not always correlate with control of viral load, but in the case of patient AC-06 viral load was below 5,000 copies before the second virus broke through, suggesting that his HIV-specific CD8 T cells were functioning relatively well.
Only limited information regarding HIV-specific CD4 T cells was reported (proliferative responses to p24 were detectable), and the quality of these responses in patient AC-06 is less clear. The ability of HIV to target HIV-specific CD4 T cells as they attempt to transition from an initial “naive” state into memory cells suggests that live, pathogenic HIV may not be the ideal candidate for priming effective memory CD4 T-cell responses. Whether vaccines could help circumvent this problem by priming HIV-specific memory CD4 T cells prior to exposure is a question for future research, but it is not outside the realm of possibility.
Immune Response: Targeting
What about targeting? At the time of the re-infection, patient AC-06 showed CD8 T-cell responses targeting eight different epitopes from HIV. After the rebound in viral load caused by the second HIV infection, HAART was reinitiated for several months followed by another interruption. Viral load once again rebounded, and HAART was restarted. Data from this timepoint indicated that the CD8 T-cell response had broadened to target twelve epitopes.
Several months later, HAART was again interrupted, and control of viral replication appeared to have improved in association with a further broadening of the CD8 T-cell response which now targeted 30 different HIV epitopes (this information was presented by Bruce Walker at the 2001 AIDS vaccine conference in Philadelphia, prior to the realization that a superinfection had occurred). Although additional data have not yet been reported, this preliminary observation suggests that infection with a single HIV strain may not necessarily induce broadly targeted CD8 T-cell responses. This may at least offer a glimmer of hope for vaccines, particularly “multivalent” constructs that include epitopes from a number of different HIV subtypes.
In terms of public health messages, it is not possible to dismiss all four of these reported cases as erroneous. The true incidence of re-infection remains entirely unknown but it must be acknowledged that the risk is real. Many difficult questions persist, including the degree of risk, whether the duration of infection plays a role in susceptibility (all four of the reported cases occurred relatively early after the initial infection), whether HAART is protective against reinfection and whether long-term non-progressors — whose HIV-specific immune responses are thought to be highly functional — are also susceptible to re-infection.