Abstract
A relatively small number of liver-localized CD8 T cells can provide sterilizing protection against exposure to Plasmodium sporozoites. Previously published mathematical model-based simulations of randomly searching for an infection T cells, assuming that T cells move in open (3D) space or on a regular lattice, suggested that some degree of attraction is needed for cells to find the infection. Yet, in our previous experiments we failed to find evidence of attraction towards the infection in most T cells. Because the movement of liver-localized T cells is confined to sinusoids, the blood vessels of the liver, T cells may in fact be attracted to the infection site via the sinusoids. In new experiments we imaged Plasmodium-specific CD8 T cells, sporozoites, and liver sinusoids simultaneously. We also developed a new method to model imaged sinusoids as graphs and to simulate movement of T cells on graphs, and new tests to detect biased cell movement on graphs. As before, our previously published open space/3D-based metric suggested that few T cells are attracted to the infection site in presence of moderate cluster of T cells. However, with our new graph-based metric we could not detect attraction when constraining T cell movement to sinusoids. Surprisingly, simulations of T cells searching for an infection via liver sinusoids showed that a relatively small number of T cells (1 cell/imaging volume or < 10(5) cell/liver), moving without attraction at experimentally observed speeds, is sufficient to find the infection within 24 hours, well below the expected lifespan of the Plasmodium liver stage in mice. Our results suggest that the constrained environment of the tissue structure makes a random search more effective, and that detecting T cell attraction towards specific areas may arise as an artifact of metrics designed for open space/3D. Our results thus call for a re-evaluation of the implicit assumption that attraction is necessary for T cells to find infections in tissues.