During 2006–2007, blood samples were obtained from 776 volunteers living in 18 villages bordering Taï National Park. All participants signed informed consent forms and completed questionnaires aimed at determining their exposure to nonhuman primate bushmeat through activities such as hunting of nonhuman primates or consumption of nonhuman primate bushmeat.

To determine effective exposure to HTLV-1/STLV-1, we used an HTLV-1/2 ELISA to test serum samples for reactivity to HTLV-1/2 antigens (10). Of the 776 serum samples, 16 were positive according to the ELISA manufacturer’s criteria; an additional 15 samples had values just below the cutoff. We extracted DNA from all ELISA-reactive samples and performed a search for HTLV-1/STLV-1 sequences by using a tax-specific quantitative PCR (8). Of the 31 samples, 10 were positive and were analyzed by using a multiplex nested/seminested PCR targeting env and long terminal repeat (LTR) sequences (Table 1; Technical Appendix). To identify multiple infections with HTLV-1/STLV-1, this assay was applied on near endpoint dilutions of the 10 DNA extracts (2–6 starting template molecules per reaction; Technical Appendix). For each person, 6–20 env and 2–20 LTR sequences (15–40 sequences per person) were determined by Sanger sequencing. No evidence of multiple infections was found.

Phylogenetic analyses were performed by using Bayesian and maximum likelihood methods on env and LTR datasets (Technical Appendix). Both methods agreed on all essential features of the LTR tree topology (Figure 2) and env tree topology (Technical Appendix Figure). Six of the newly determined HTLV-1 sequences were unambiguously related to HTLV-1A (bootstrap, 94; posterior probabilities, 1) (Figure 2), confirming the predominance of this molecular subtype in Côte d’Ivoire and in western Africa (6). Another 3 HTLV-1 sequences were closely related to STLV-1 sequences found in sooty mangabeys (Cercocebus atys) from Taï National Park (bootstrap, 83; posterior probabilities, 1) (Figure 2; Technical Appendix Table 2), whereas the last 1 was related to STLV-1 sequences from red colobus monkeys (Piliocolobus badius badius) and chimpanzees (Pan troglodytes verus) from Taï National Park (bootstrap, 98; posterior probabilities, 1) (Figure 2; Technical Appendix Table 2) (7,8). Bayesian analyses were run under the assumption of a molecular clock and calibrated. However, reliable divergence dates could not been determined because most shallow nodes of the trees, including those of interest here, were not supported (Technical Appendix). Observed divergences, however, seemed compatible with cross-species transmission events, particularly in the case of study participant Pau009 (divergence to closest STLV-1, 0% in LTR and 0.2% in env) (Table 1).

We investigated the frequency of direct zoonotic transmission of STLV-1 in a rural region of Côte d’Ivoire neighboring Taï National Park and found that only 2 of the STLV-1–related sequences would be compatible with a local human-to-human transmission (Gah050 and Kei005; Figure 2). Therefore, our data support the notion that direct zoonotic transmissions of STLV-1 represent a measurable proportion of HTLV-1 infections, at least in rural regions bordering nonhuman primate habitat. In addition, these results mirror observations made among adult chimpanzees from Taï National Park, which are often infected with retroviruses (i.e., simian foamy viruses and STLV-1) of their prey (Figure 2) (7,11).

Despite the high prevalence of STLV-1, simian foamy virus, and simian immunodeficiency virus infections among red colobus populations (8) and the fact that this nonhuman primate species is the one most frequently hunted by humans (4), most zoonotic transmissions of retroviruses in western Africa seem to originate from sooty mangabeys, as shown here for STLV-1 and previously described for simian immunodeficiency virus of sooty mangabeys, the precursor of HIV-2 (12). It remains to be determined whether these zoonotic transmissions from sooty mangabeys are favored as a result of molecular determinants (e.g., convergent evolution of retroviral receptors) or behavioral determinants (e.g., increased aggressiveness).

Considering the human exposure to nonhuman primate bushmeat in this region (as illustrated by ≈150,000 kg sold per year in markets) (13) and given the high prevalence of STLV-1 among local nonhuman primates (7,8), the observation that zoonotic transmission events are, in absolute terms, exceedingly rare is striking. Yet, the accumulation of genetically distinct HTLV-1/STLV-1 over restricted geographic areas remains possible, as illustrated by the finding of 1 person infected with HTLV-1A and 2 persons infected with putative STLV-1 in a single village, Kéibly (Table 1; Figure 1). Such local accumulations add to the threat represented by direct transmissions of STLV-1 because they can provide an opportunity for recombinant viruses to emerge, even though HTLV-1/STLV-1 biology may be unfavorable to recombination (14).

The analysis of behavioral data reveals generalized exposure of local populations to cooked nonhuman primate bushmeat (Table 2). Exposure to fresh tissues, which can be expected to be more risky in terms of retroviral transmission, is less common (Table 2). Along a gradient of bushmeat freshness, going from hunting to preparation and cooking, a clear reversal of sex-related skew can be observed: only men are hunters, men and women are equally involved in dismembering, and women predominantly prepare and cook nonhuman primate bushmeat (Table 2). Hence, men likely constitute a population at risk. In our study, 75% of persons who were identified as infected with viruses closely related to STLV-1 were men, whereas all HTLV-1A–infected persons were women. Increased surveillance for zoonotic transmission of STLV-1 to humans in areas where such transmission is more likely and increased surveillance of nonhuman primate species with high transmission potential (like sooty mangabeys) will contribute to a better understanding of risk factors.