Mosquito-based xenosurveillance reveals circulation of Anaplasma and Neoehrlichia in livestock at a human-wildlife interface in Kenya.

backgroundXenosurveillance, which uses blood-feeding insects as biological samplers, is an emerging non-invasive approach for monitoring pathogens circulating among humans, livestock, and wildlife. However, its application to livestock-associated bacterial pathogens at human-animal-wildlife interfaces remains underexplored. We investigated whether mosquito blood meals could be used to detect tick-borne bacterial pathogens circulating in livestock in Kenya.

methodsWe collected 4673 mosquitoes, belonging to Culex, Anopheles, Aedes, Mansonia, and Coquillettidia genera around livestock enclosures in Kajiado and Naivasha counties, Kenya, using CO₂-baited CDC miniature light traps. Traps were selected to maximise species diversity, and as light traps capture fewer engorged mosquitoes than resting traps, only 56 blood-fed individuals (1.2%) were collected and processed as whole-specimen homogenates and analysed for vertebrate blood-meal sources using cytochrome b sequencing. In total, 303 mosquito pools, including blood-fed individuals processed as single-mosquito pools, were screened for Anaplasma, Ehrlichia, Rickettsia, Theileria, and Babesia using PCR-high-resolution melting analysis and confirmatory sequencing.

resultsAll pathogen-positive detections were exclusively from blood-fed individuals. We detected Anaplasma marginale in Culex pipiens (1/150; 0.7%) and Aedes hirsutus (2/11; 18.2%), Anaplasma sp. in Cx. pipiens (2/150; 1.4%) and Ae. hirsutus (1/11; 9.1%), and Candidatus Neoehrlichia mikurensis in Mansonia africana (2/50; 4%). Pathogen detections showed strong host concordance, where A. marginale was associated with cattle-derived blood meals and Anaplasma sp. with goat-derived blood meals, while Candidatus Neoehrlichia mikurensis was detected in Mn. africana that had fed on cattle and on a host that could not be determined.

conclusionsOur results provide preliminary evidence that mosquito-based xenosurveillance can detect tick-borne bacterial pathogens circulating in livestock at human-wildlife interfaces in Kenya. The strong concordance between pathogen identity and vertebrate host in blood-fed mosquitoes supports the biological plausibility of this approach. Notably, detection of Ca. Neoehrlichia mikurensis represents the first report of this zoonotic pathogen in mosquitoes in Africa, highlighting the One Health relevance of xenosurveillance in identifying settings, where pathogen circulation and cross-interface feeding coincide. Mosquito-derived sampling has potential to complement existing surveillance tools in agro-pastoral systems, where direct host sampling is difficult or costly.