HIV-1 Env-specific cytotoxic T-lymphocyte responses in exposed, uninfected Kenyan sex workers

The prospective significance of HIV-specific cytotoxic T lymphocyte (CTL) responses in highly exposed, persistently seronegative populations is unknown. In 1996–1997 we screened for CTL responses against HIV clade BEnvin 39 recently enrolled Kenyan female sex workers, and followed these women prospectively. Annual HIV incidence was 5.8%. CTL were independently associated with age and recent HIV-1 exposure, but were not prospectively associated with protection in a multivariable model that included HIV-1 exposure and duration of sex work. Some individuals remain HIV seronegative and immunologically normal despite repeated HIV-1 exposure. Although a minority harbour trace amounts of HIV-1 genetic material, suggesting productive infection below the limit of detection [1], biological factors also play a role in sterile protection [2]. HIV-1-specific cytotoxic T lymphocytes (CTL) are important in viral control in HIV-1-infected individuals, but are unable to prevent eventual immunosuppression and death [3]. HIV-1-specific CTL have been described in some HIV-1-exposed, persistently seronegative (HEPS) populations, but not in others [4], and it has been suggested that their detection might be a laboratory artifact [5]. Female sex workers (FSW) in the Pumwani slum of Nairobi, Kenya, have over 60 unprotected sexual exposures to HIV-1 per year, resulting in intense HIV-1 infection pressure [6]. Rates of HIV-1 seroconversion are highest during the first 3 years of cohort enrollment, although infection has been observed more than 10 years after enrollment [7]. Various HIV-1-specific immune responses have been detected in HEPS FSW from this cohort, including HIV-1-specific CD8 T cells [8–10] and HIV-1-Env-specific proliferative responses [10,11]. Between 1996 and 1997 we assayed for HIV-1-Env-specific CTL in newly enrolled (< 3 years) HIV-seronegative FSW, and followed these women in order to define the prospective significance of HIV-specific CTL. FSW were confirmed to be HIV-1 uninfected by a polymerase chain reaction system adapted to detect African clades [12]. Bulk CTL assays were performed as previously described [9,10] by laboratory staff blinded to participant clinical and behavioural characteristics. phytohemagglutinin (PHA) blasts were infected with HIV-1IIIB, inactivated with mitomycin C and used as autologous stimulators for bulk autologous peripheral blood mononuclear cells in IL-2 at a dilution of 10% (Lymphocult-T; Biotest, Solihull, United Kingdom). On day 10, autologous Epstein–Barr virus transformed B cells were infected overnight with either vSC8 (wild type vaccinia with the β-galactosidase gene) or vPE16 (HIV-1 clade B Env inserted into vSC8; obtained through the AIDS Research and Reference Reagent Program, NIAID, National Institutes of Health), and a 51Cr release assay was performed in triplicate with effector to target (E : T) ratios of 50 : 1, 25 : 1, and 12.5 : 1. A positive assay was defined as the lysis of vPE16-infected targets 10% or greater than that of the control. All FSW were provided with free condoms, risk reduction counselling and sexually transmitted infection case management. Univariate analysis used one-way analysis of variance or the Mantel–Henszel chi-square test. Independent associations of CTL and HIV-1 acquisition were examined in multivariable models that included those variables associated in univariate analysis (P ≤ 0.1) and in previous studies. Informed verbal consent was obtained from all study participants, and studies were approved by research ethics board (REB) at the University of Manitoba and the University of Nairobi. All 39 newly enrolled seronegative FSW in follow-up during 1996–1997 were assayed, and followed for a total of 156.4 person-years (mean 4.0 years; range 0.1–6.5 years). CTL assays were performed at baseline only. At the time of the assay, participants had already been followed in the clinic for a mean of 1.0 years (range 0–2.0), and had practised sex work in any setting for a mean of 5.2 years (range 0.3–18). Their mean age was 32.2 years (range 21–42). Assuming a 30% HIV-1 seroprevalence in male clients [7], and using self-reported client numbers and condom use, the number of unprotected HIV-1 exposures over the past year was calculated to be 42.3 (range 0.9–202.5). CTL directed against clade B HIV-1 Env were detected in eight out of 39 FSW (21%). CTL-positive FSW were older (35.9 versus 31.3 years; P = 0.02), but had a similar duration of previous sex work (5.1 versus 5.3 years; P = 0.9) and had been followed in the Pumwani cohort for a similar time (1.0 year for both; P = 0.9). There was a trend towards higher HIV-1 exposure over the preceding year in the CTL-positive group (63.9 versus 36.4 exposures; P = 0.2), and CTL-positive FSW were more likely to be taking an oral contraceptive [4/7 versus 5/30; likelihood ratio (LR) 4.5; P = 0.04], and were less likely to be using injectable progesterone (0/7 versus 9/30; LR 4.4; P = 0.04). Stepwise logistic regression (variables included: age, number of HIV-1 exposures over the past year, current use of oral contraception or injectable progesterone) demonstrated that only age and recent HIV-1 exposure were independently associated with CTL [LR 1.9, 95% confidence interval (CI) 1.1–3.6; P = 0.03; and LR 1.04, 95% CI 1.0–1.07; P = 0.05; respectively]. There were nine new HIV-1 infections during 156.4 person-years of follow-up, for an annual HIV-1 incidence of 5.8%. FSW who acquired HIV-1 reported a longer duration of previous sex work (9.5 versus 4.1 years; P = 0.003). The presence of HIV-1-specific CTL at baseline was associated with a reduced subsequent risk of HIV-1 acquisition on univariate analysis (0/8 versus 9/31: LR 4.8; P = 0.03), and there was a trend towards longer HIV-1-free survival in Kaplan–Meier analysis (Fig. 1: log-rank P = 0.1). However, in Cox multivariable regression analysis (variables included: HIV-1 exposure over past year, previous duration of sex work, presence of HIV-1-specific CTL), only an increased previous duration of sex work was independently associated with seroconversion (risk ratio 1.2, 95% CI 1.0–1.3; P = 0.02). There was no independent association between HIV-1 protection and baseline HIV-specific CTL (P = 0.9).Fig. 1.: Time to HIV-1 seroconversion, by cytotoxic T-lymphocyte status at baseline. CTL, Cytotoxic T lymphocyte; ELISA, enzyme-linked immunosorbent assay. –– Baseline CTL positive; – – – baseline CTL negative.The association of recent HIV-1 exposure with the detection of CTL by laboratory staff blinded to exposure data supports previous associations of exposure with HIV cellular immune responses [7,13], and strongly suggests that these responses were not a laboratory artifact. By following FSW prospectively, we attempted to clarify whether the CTL detected were responsible for HIV-1 protection, or were perhaps induced/maintained by the non-productive HIV exposure, with other factor(s) actually mediating protection. Although baseline CTL were associated with a reduced HIV incidence in univariate analysis, this association fell out in a multivariable model that included the duration of previous sex work and levels of HIV exposure. Our sample size was moderate, with 156 person-years of follow-up and a relatively low HIV-1 incidence of 5.8%, perhaps as a result of effective socio-behavioural interventions [14], so a causal association between CTL and HIV protection cannot be excluded. However, an alternative hypothesis is that CTL are the result of non-productive HIV-1 exposure, rather than mediators of protection. This might explain the low magnitude of HIV-specific CD8 responses in HEPS cohorts, compared with those seen in HIV-infected individuals [15]. The only definitive way to resolve the issues of CTL and HIV-1 protection is to induce similar responses in the context of a randomized vaccine trial, which is the aim of several HIV-1 vaccines in development [16,17].

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.