Evaluation of near infrared spectroscopy for sporozoite detection in mosquitoes infected with wild-strain parasites from asymptomatic gametocyte carriers in Kilifi Kenya

Abstract

BackgroundScreening for Plasmodium spp. sporozoite infection in mosquitoes is routinely done using ELISA (enzyme-linked immunosorbent assay). Near infrared spectroscopy (NIRS), a fast and non-destructive method, has recently been shown to distinguish, with 95% accuracy, between uninfected and sporozoite-infected mosquitoes using laboratory strains of Plasmodium falciparum (PfN54). The aim of this present study was to further investigate the reproducibility of NIRS to identify sporozoite infection in mosquitoes infected using field isolates of P. falciparum gametocytes from asymptomatic carriers. MethodsHealthy individuals (aged 5 years and above) were screened for gametocytaemia by thick-smear microscopy in an area of moderate transmission along the Coast of Kenya between May and September 2018. Asymptomatic gametocyte carriers were recruited for mosquito feeding assays, direct membrane feeding (DMFA) and direct skin feeding (DFA), using insectary-reared Anopheles gambiae s.s (Kilifi strain). Mosquitoes were kept for 14-days post feeding after which they were scanned using NIRS and subsequently analysed for sporozoite infection using circumsporozoite-ELISA. Predictive models were explored using partial least square regressions (PLS). ResultsTwo hundred and ninety-nine (299) individuals were screened for malaria parasites, 74 (24.8%) were found with circulating asexual parasites, and 16 (5.4%) with P. falciparum gametocyte stages. Fourteen (14) asymptomatic gametocyte carriers were recruited to the study for mosquito feeding assays. A total of 134 (7%, 134/1881) sporozoite-infected mosquitoes were obtained from 9 successful experiments. Three different training datasets composed of infected and uninfected mosquitoes were analysed. The PLS models were unable to distinguish between sporozoite-infected and uninfected mosquitoes. A predictive model could not be generated. ConclusionsThe results of this study were not consistent with previous published research on NIRS for detection of sporozoite infection in the same mosquito species and may reflect differences between laboratory and field conditions. The current findings indicate that methods for sporozoite detection should be tested on field isolates at an early stage in their development and are informative for future research seeking novel high-throughput methods for parasite detection in mosquitoes.