Abstract
Pre-clinical assessment of vaccines for neurotoxicity and neurovirulence is mandatory for safe human administration especially for vaccines developed to target neurotropic viruses. Several recent studies suggest that some vaccine candidates, for example yellow fever vaccine tested to be neuroattenuated in monkeys were later found to be neurovirulent in vaccinated population. In this study, we used a stem cell composed human microphysiological system configuration as an in vitro platform and infected them with four strains of DENV (Dengue Virus) followed by capturing the micrographs that were analyzed by AI/ML enabled software to generate neurovirulence score. Additionally, the host gene expression studies done with infected microphysiological system resulted in establishing the signatures specific to the host system against viral strains. Varying degrees of neurovirulence risk scores were recorded by the prediction model. Our approach is the first of its kind, showcasing the use of a healthy human Microphysiological System complemented with digital tools to generate neurovirulent gene signatures represented by HLA-B, C1QB, TIMP4, CD63 and RANTES, which may play a role in host protection or as a result of a pathological response against DENV infection.