A T cell wanders through the liver like a person without a map in a labyrinth. The infected cell is somewhere nearby, but there are no signposts. It used to be thought that immune cells sense a chemical trail from the parasite and head straight for it. New observations say otherwise: the cell simply tries one corridor after another, and that proves to be enough.
Researchers injected activated T cells into mice and then infected them with the malaria parasite. Under the microscope they simultaneously filmed the movement of the cells, the position of the parasite, and the network of vessels. The vessels were digitized and turned into a graph: branch points became vertices, and segments of vessels became edges weighted by length. Every position of a cell and of the parasite was projected onto the nearest vertex of the graph.
When the movement was analyzed in open space, a small fraction of the cells looked as if they were drawn toward the parasite. But once the structure of the vessels was taken into account, the numbers of attracted and repelled cells evened out. There is no systematic attraction. Simulations of random search, however, showed that a single cell, across the entire observation volume, finds the infected site in 12 hours on average. Over 48 hours the probability of success exceeds 94%.
Even without attraction, ten cells form a cluster around the parasite within six hours. Such a cluster used to be taken as proof of directed attraction. Now it is clear that it arises on its own, as in a labyrinth where several people wandering at random happen to meet by the same wall. This does not change the outcome: the parasite is found and destroyed.
The labyrinth that seemed like a hindrance actually sets up a convenient geometry for the search. A T cell does not need a map; it is enough simply to keep walking the corridors. Understanding this helps estimate how many immune cells are needed for protection, and it may prove useful in developing malaria vaccines.
The fractal dimension of the sinusoid graph turned out to be about 1.7–2.0, less than the three dimensions of the liver. This means that T cells search for a target in a space whose properties are closer to a plane than to a volume, and therefore random search works faster.
Not published on arXiv. Here are its page, the PDF and the formalization: you can check it yourself, not only read our account.