A minority of liver-resident CD8 T cells searching for Plasmodium-infected hepatocytes demonstrate difficult-to-detect attraction

Malaria is a disease caused by parasites from genus Plasmodium that causes over 200 million infections and kills over 400,000 people every year. A critical step of malaria infection is when mosquito-injected sporozoites travel to the liver and form liver stages. Several malaria vaccine candidates induce high levels of Plasmodium-specific CD8 T cells which are able to eliminate all liver stages, thus providing sterilizing immunity against the disease. However, how CD8 T cells locate the site of infection is not well understood. We generated and analyzed data from intravital microscopy experiments in mice in which movement of T cells relative to the liver stage was recorded in several different settings. To detect attraction of T cells towards the infection site, we developed a novel metric based on the Von Mises-Fisher (VMF) distribution, which is more powerful than previously used metrics. We found that the majority (85-95%) of Plasmodium-specific CD8 T cells and T cells of irrelevant specificity did not display attraction towards the parasite when the parasite was not found by T cells, which was consistent with the random search for infection. In contrast, when some T cells located the parasite and formed a cluster, a minority of other T cells did display strong attraction towards the infection. Interestingly, the speed of T cell movement (and small turning angles) correlated with the bias of T cell movement towards the infection site (while many other parameters do not), suggesting that a deeper understanding of what determines the speed of T cell movement in the liver may help with improving T cell vaccine efficacy. Stochastic simulations suggested that a small movement bias towards the parasite dramatically reduces the number of CD8 T cells needed for the complete elimination of all malaria liver stages, and yet, to detect such attraction by individual cells requires extremely long imaging experiments which may not be currently feasible. Our developed methodology can be allied to detect weak attraction of moving agents in other conditions.