Mucosal IgA in exposed, uninfected subjects: evidence for a role in protection against HIV infection

We read with interest the well-conducted study by Dorrell et al. [1], in which the authors did not detect HIV-1-specific mucosal antibody responses in exposed, uninfected Gambian sex workers. Although we agree with their conclusion that ‘the presence of IgA responses against HIV is not a necessary requirement for resistance', we do not agree that it is ‘unlikely that mucosal IgA plays an important role in resistance against HIV in exposed uninfected individuals'. HIV-1-specific IgA has been described in the genital tracts of between 38 and 82% of subjects in three published exposed, uninfected cohorts [2–4], and in a majority of exposed, uninfected Cambodian sex workers (M. Bomsel, in preparation). Dorrell et al. [1] correctly point out that Belec et al. [5] did not find HIV envelope-specific IgA in exposed, uninfected women from the Central African Republic, but they did demonstrate IgA and IgG directed against HIV-1 core antigens, supporting a role for specific mucosal antibodies in HIV-1 resistance. It does seem unlikely that IgA is the only mechanism of immune-mediated resistance, because HIV-1-specific cellular responses were also commonly found in those studies [3,4], and were sometimes independent of IgA. However, it does not follow that IgA plays no role in resistance. Why might IgA be found in some exposed, uninfected cohorts, but not at all in that studied by Dorrell et al. [1] ? One factor may be cohort comparability, which is important when examining HIV-specific IgA in exposed, uninfected individuals, because these responses may wane in the absence of ongoing HIV exposure [6]. Although Dorrell et al. [1] did not present correlates of HIV exposure, in a 1992 survey Gambian sex workers used condoms 80% of the time, worked 4 days a week and had an average of two to three clients a night [7]. Given an HIV-1 prevalence among Gambian sexually transmitted disease (STD) clinic attendees of 4.7% in 1991 [8], one would predict 0.4 unprotected exposures to HIV per month. Sex workers from Nairobi [3] had 2.7 clients per day and used condoms with 59.9% of clients. Using 1991 data so as to maintain comparability with the Gambian survey, HIV-1 prevalence among STD clinic attendees was 28.8%. These women would therefore be predicted to have 8.7 unprotected exposures to HIV per month, approximately 20-fold more than the Gambian women. A mean of 2.0 clients per day, with an HIV-1 prevalence in Thai STD clinic attendees of 15–20%, also suggested higher levels of exposure in the study of Beyrer et al. [2]. Reduced HIV exposure in the Gambian sex workers would be expected to result in a lower HIV infection pressure, and so protective immune responses in this cohort may not be selected for in the same way. Dorrell et al. [1] suggested that a high level of sexual exposure could act as a confounder, because IgA from residual HIV-infected semen may be detected in the genital tract. However, the simultaneous detection of HIV-specific IgA in the blood of subjects from the Nairobi, Thai and Italian studies [2,3,6] makes this unlikely. Although IgA was present at a low titre in all studies, Beyrer et al. [2] were able to detect IgA at a 1 : 8 dilution, and in all the studies there is a considerable dilution factor involved in mixing a small volume of genital secretions into a much larger amount of phosphate buffered saline. Even if IgA is present at a low level, it does not follow that cellular defences play a greater role in immunity against HIV, because cytotoxic T lymphocytes are also present at a considerably lower level in HIV-exposed than infected individuals [9,10]. IgA detected in other exposed, uninfected cohorts was specific for HIV gp160, and has been mapped to epitopes within HIV gp41 for seronegative partners of infected individuals [11] and for a cohort of exposed, uninfected Cambodian prostitutes (M. Bomsel, in preparation). Furthermore, IgA from exposed, uninfected donors can neutralize both T-tropic [6] and M-tropic virus [12], and has been shown to inhibit HIV epithelial transcytosis [13]. It does not seem logical to discount this IgA response on the basis that it does not recognize gp120, which harbours two out of the three known primary isolate neutralizing epitopes. The lack of existing HIV neutralizing epitopes is a major problem in vaccine design [14]. As mucosal IgA from both HIV-infected [15] and exposed, uninfected [12] donors can neutralize HIV, mapping the specificity of these novel IgA antibodies may have great potential relevance for HIV vaccine studies. In conclusion, we feel that the role of IgA in protection against HIV-1 remains a promising avenue of investigation, despite the well-conducted negative findings of Dorrell et al. [1]. We really do hope that this excellent research group will continue to explore the role of IgA in HIV resistance, rather than to allow their negative result to dissuade them from further research in this area.