Ecology and behaviour of vectors of Plasmodium knowlesi malaria in Sabah, Malaysian Borneo

• Main Author: Brown, Rebecca Emma
• Published: University of Glasgow 2019
• Subjects: QH301 Biology
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.768711

Over the last decade, the transmission of zoonotic malaria from non-human primates to humans has emerged as a public health problem and possible threat to malaria elimination in Southeast Asia. A major outbreak of the macaque malaria parasite Plasmodium knowlesi in humans began in Malaysian Borneo in 2004 and is now the primary cause of malaria in this region. This simian parasite is transmitted by mosquitoes in the Anopheles leucophyrus species complex. The emergence of P. knowlesi has been tightly linked to land-use change, particularly the widespread deforestation occurring in the state of Sabah where the largest focus of human infection is found. Efforts to combat this disease and understand its emergence and future spread in humans are hindered by limited knowledge of mosquito vector ecology and behaviour; and the risks of exposure to vectors in changing landscapes. This PhD aimed to address these knowledge gaps by carrying out a series of field studies near the epicentre of human P. knowlesi cases in Sabah, Malaysian Borneo, to elucidate P. knowlesi vector ecology, behaviour and transmission potential, verify associations between land-use type and human exposure risk, and characterize the dynamics of transmission within reservoir macaque populations. In combination this information will deepen understanding of P. knowlesi transmission and emergence, and provide insights for the control of this and other emerging zoonotic malarias. My initial study evaluated new sampling methods for collecting resting P. knowlesi vectors. Resting collections are valuable for characterization of mosquito habitat and host species choice, however no standard methodology is currently available for P. knowlesi vectors. I evaluated two simple traps, resting buckets and sticky resting buckets, for sampling resting P. knowlesi vectors within two villages in Kudat District, Sabah. The performance of traps was evaluated, and the relative abundance and host choice of resting mosquito vectors was compared across eight different habitat types representing a gradient of deforestation. In 5748 trap days, a total of 2212 mosquitoes were collected in resting collections, but none were malaria vector species. Culex and Aedes genera dominated collections; with the former being most abundant in resting bucket traps and CDC aspirator catches, and the latter in sticky resting bucket traps. Several other vector species were collected including the sylvatic dengue vector Aedes albopictus, and Culex vectors of filarasis and Japanese encephalitis. Consequently these simple resting traps could be effective for studying the ecology of a range of other important mosquito vectors in Sabah even if not those responsible for malaria. In a following study I investigated associations between habitat and human exposure to P. knowlesi vectors, and tested for associations between vector abundance and human infection risk across a broad geographic range in Sabah. Previous studies indicated that the primary P. knowlesi vector was An. balabacensis. This vector was more abundant in a village than forest site, conflicting with the original hypothesis that humans are at greatest risk of infection in forests and suggested the possibility of peri-domestic transmission. However this inference was drawn from a limited number of sampling sites in only one district, Kudat, within Sabah. To test this hypothesis over a broader geographical scale, I conducted extensive entomological sampling across four districts in Sabah. Human landing catches were performed to measure human biting rates in forest, farm (plantation) and peri-domestic habitats in 11 villages. Prior to entomological sampling survey of human sero-positivity to P. knowlesi was conducted in all villages, carried out as part of a larger research programme. Making use of this data, I tested for associations between vector abundance and human infection risk at the village level. The primary vector An. balabacensis was found in all four districts, but at much lower relative abundance than in pilot work from Kudat. Additionally this vector was more abundant in forest and farm habitats than in peri-domestic settings. Only 1 of the 32 An. balabacensis collected in this study tested positive for P. knowlesi; an individual caught in a forest site. No significant association between the mean abundance of An. balabacensis and human P. knowlesi sero-positivity was detected in this study. However the relatively small sample size of mosquitoes and sites used here meant there was relatively low power to detect such an effect. This study highlights the importance of incorporating geographical heterogeneity and replication when assessing mosquito-habitat associations, and the need for more intensive longer-term sampling to establish potential entomological indicators of P. knowlesi infection in humans. A final study was conducted to investigate the transmission dynamics of P. knowlesi in macaque reservoir populations. Most studies of P. knowlesi vectors have been conducted in or near disturbed forest, where both humans and macaques are in contact. It is unknown whether the same vector species involved in human-macaque infection also mediate transmission between macaques. To investigate this and other aspects of macaque-mosquito interactions, I conducted a field study within the Danau Girang Field Centre in Sabah where there is a large population of long-tailed macaques. First I evaluated the use of Mosquito Magnet Independence Traps (MMIT) as a non-invasive means to sample mosquitoes host seeking near macaque sleeping sites. The MMIT performed well relative to the human landing catch, with both methods collecting An. balabacensis and other malaria vector species. Second, MMITs were used to sample mosquitoes host seeking near trees where macaques were sleeping and at unoccupied control trees. Additionally, macaque faecal samples were tested for malaria as an estimate of infection rate in the reservoir population. Anopheles balabacensis was more abundant at macaque sleeping sites than control trees indicating this vector has a specific propensity for feeding on macaques. Approximately 37% (n = 17/46) of macaque stool samples tested positive for Plasmodium infection but none of these were identified as being P. knowlesi. Two Anopheles vectors tested positive for Plasmodium which was subsequently confirmed as the primate parasite P. inui. Thus P. inui is likely the major source of malaria infection in this primate population. This study indicates that not all macaque populations pose a P. knowlesi risk, but other malaria parasites are common and should be monitored to assess for future spillover. In combination, this research expands knowledge of P. knowlesi transmission in Malaysian Borneo, and has implications for planning surveillance and control. Notably it emphasizes the value of larger-scale surveillance of vector and macaque populations to assess human exposure risk, as and requirement of an integrated One Health approach to tackle zoonotic malaria.