Might male circumcision be more protective against HIV in the highly exposed? An immunological hypothesis

Observational studies have suggested that male circumcision is associated with a reduced risk of HIV acquisition in men [1–5]. This is biologically plausible, because the intact foreskin is vulnerable to trauma during intercourse and is susceptible to genital ulceration and inflammatory conditions (such as balanitis), which may act as co-factors for HIV acquisition [6–8]. Also, the inner preputial mucosa of the intact foreskin is rich in HIV target cells [9]. In contrast, the penile shaft and prepuce are keratinized [6] and contain lower numbers of HIV target cells [9]. After circumcision, the only remaining mucosa is the urethral meatus. The anatomical reduction of vulnerable mucosa may thus reduce the risk of HIV. Paradoxically, observational data have suggested that circumcision may be more highly protective against HIV acquisition among men at high risk of repeated exposure, compared to men with less frequent HIV exposures. In a meta-analysis, the adjusted rate ratio (RR) of HIV acquisition in circumcised compared with uncircumcised men was 0.29 [95% confidence interval (CI) 0.20–0.41] in men with high-risk behaviors, compared with an adjusted RR of 0.56 (95% CI 0.44–0.70) in general male populations [1]. In Rakai, Uganda, the adjusted RR of male HIV acquisition associated with circumcision was 0.53 (95% CI 0.33–0.87) in the general population, but among frequently exposed men in HIV-discordant couples, there were no seroconversions in 50 circumcised men followed for an average of 2.1 years, compared with an incidence rate of 16.7 per 100 person-years in 137 uncircumcised men (P = 0.0004) [2]. Substantial circumcision-associated reductions in HIV risk have also been observed in frequently exposed male sexually transmitted disease patients in Pune, India (RR 0.12, 95% CI 0.03–0.48) [3] and Kenyan truck drivers (RR 0.25, 95% CI 0.12–0.52) [4,5]. Circumcision may potentially be more protective in frequently exposed men because the procedure reduces the risk of genital ulcer disease, a co-factor for HIV acquisition [7,8]. However, we propose that immune responses may also contribute. Because the urethral meatus represents a small area of vulnerable mucosa, the amount of HIV exposure per coital act is likely to be lower in circumcised compared with uncircumcised men. In circumcised men experiencing recurrent exposure to low doses of HIV, antigenic stimulation by repeat subinfectious inoculums may induce a mucosal immune response, enhancing protection over and above the reduced risk afforded by the removal of the foreskin per se. The induction of mucosal immunity has been observed in highly exposed but uninfected members of HIV-discordant couples [10] and in commercial sex workers, among whom there were enhanced mucosal CD8 T-cell responses [11–13] and HIV-1 specific IgA [14,15]. The interruption of sex work was associated with a loss of CD8 HIV-specific responses and an increased risk of HIV infection [16]. Other studies of persistently seronegative female sex workers have also reported HIV-specific antibody responses [17,18], and increased α-defensin production and α-defensin-expressing CD8 lymphocytes [19]. Mucosal alloimmunization has also been suggested as a protective factor in HIV acquisition [20]. It is thus possible that circumcised men with frequent HIV exposures experience an induction of mucosal immunity and greater protection against HIV acquisition, compared with circumcised men with infrequent HIV antigen exposure. Furthermore, the effects of induced immunity in highly exposed circumcised men may be more pronounced than in highly exposed women, because the surface area of vulnerable mucosa in the former is substantially smaller than in the female genital tract, and the infectious inoculum per act of intercourse may be less likely to overwhelm the effects of partial protection. In summary, circumcision may protect men from HIV by three possible mechanisms: (i) an anatomical mechanism consequent on the removal of vulnerable foreskin mucosa; (ii) the reduction of co-factors such as genital ulcer disease; and (iii) the induction of a mucosal immune response in the presence of repeated antigen stimulation. This hypothesis can potentially be assessed by analysing effects in highly exposed versus less exposed participants in three randomized trials of male circumcision for HIV prevention, conducted in Kisumu, Kenya, Rakai, Uganda (both ongoing), and South Africa (recently completed). Futhermore, foreskin tissue is being preserved in the Kenyan and Ugandan trials, and immunohistochemistry may yield further insights into questions of mucosal immunity. These issues are not merely of theoretical concern. If the trials show that the magnitude of circumcision efficacy varies with the intensity of HIV exposure, the finding would have implications for the effectiveness and cost-effectiveness of circumcision as a method of HIV prevention in diverse populations. In addition, if the relative effects of male circumcision differ by the degree of HIV exposure, this will present unique challenges for the provision of circumcision-related HIV health education, in order to prevent misinterpretation, and increased risk behaviors.