Keeping the barbarian at the gate

Comment on Belyakov et al, page 3258

It is generally accepted that cell-mediated immune responses play a key role in the control of HIV replication in vivo, yet they do not prevent the initial infection. Although these responses do not prevent infection, a vaccine study in rhesus macaques suggests that the responses can contain the virus at the point of entry.

Areport from Belyakov and colleagues in this issue of Blood shows that mucosal vaccination can slow the spread of a model AIDS virus from the site of infection in the gastrointestinal tract to the peripheral tissues in rhesus macaques. This study also suggests that this early control of tissue viral load is inversely related to the frequencies of local, high-avidity CD8⁺ cytotoxic T lymphocytes (CTLs) elicited by the vaccine. These observations come at a time of increasing awareness that local immune responses at the site of infection are likely to play pivotal roles in the efficacy of an AIDS vaccine. This efficacy comes in 2 forms, sterilizing protection, in which no infection is apparent, and nonsterilizing protection, in which infection occurs but with decreased viral loads and progression to disease. Fortunately, both types of protection have been demonstrated for model AIDS viruses. Sterilizing protection has been shown in passive immunization studies using pooled neutralizing antibodies,^1,2  and nonsterilizing protection has been shown in a number of active immunization studies,³  many using vaccines that lack epitopes that can elicit neutralizing antibodies. The latter immunogens are often called “CTL vaccines,” and this is the type of vaccine used by Belyakov and colleagues in the current study. For both types of efficacy, we are largely ignorant of the details of how the humoral and cellular arms of the immune system afford protection, particularly at local sites. Our ignorance extends even to the question of whether (or how) these 2 arms synergize in protection, which is critical to the rational design of a globally effective AIDS vaccine.

To wit, one recent study⁴  found no synergy for sterilizing protection between passive immunization with pooled neutralizing antibodies and active immunization with a CTL vaccine; however, in that study, the CTL vaccine was given systemically and the viral challenge mucosally. The current and previous⁵  work from Belyakov and colleagues argues that such studies are worth revisiting and might have a different outcome if mucosal vaccines are used. In earlier work,⁵  Belyakov and colleagues showed that mucosal immunization affords superior nonsterilizing protection to that elicited by systemic immunization. The current study provides a possible mechanism for this difference in which high-avidity mucosal CD8⁺ CTLs limit the systemic spread of a model AIDS virus. This mechanism is suggested by the delay in peak viremia and the inverse correlation between early viral loads and the presence of high-avidity CD8⁺ CTLs specific for viral epitopes. Strikingly, the inverse correlations were strongest for CTL responses in tissues draining the mucosal site. These observations suggest a previously unexpected ability of high-avidity CD8⁺ CTL responses to keep the replication of a model AIDS virus local, albeit for a relatively short time. The quest now becomes one of divining an immunization strategy to increase the time that the barbarian is kept at the gate. Hopefully, this barbarian can be held there long enough for the full brunt of the immune system to drive it away completely. ▪