§ The Leading Article
Reading the border
A single layer of cells separates the body from the largest microbial population it will ever host. How the CD4 T cell learns to read that border — and how it remembers what it read — is the question this laboratory has followed for three decades.
Every day the lining of the intestine performs a task no other tissue attempts at the same scale. Across a sheet one cell thick, it must admit what is useful and refuse what is dangerous, distinguishing a hundred trillion resident microbes — most of them harmless, many of them necessary — from the rare pathogen that would breach the wall. Get the balance wrong in one direction and infection follows; get it wrong in the other and the tissue turns on itself. The immune system's arbiter in that negotiation is the CD4 T cell.
For most of the twentieth century, the CD4 compartment was understood as a pair: T helper 1 cells for intracellular pathogens, T helper 2 cells for parasites and allergy. In 2005 this laboratory reported a third lineage — effector cells defined not by interferon-γ or interleukin-4 but by interleukin-17, and arising by a developmental route distinct from either paradigm.1 The finding reframed how immunologists think about inflammation, and it opened the question the lab has pursued since: what are these cells actually for?
The answer, increasingly, is that they are a barrier program. Th17 cells and their relatives garrison mucosal surfaces, and the cytokines they secrete are addressed less to other immune cells than to the epithelium itself — instructions to seal, to proliferate, to secrete antimicrobial protein. Interleukin-22 from T cells proved to be non-redundant in the defense of colonic crypts, a role no innate source could substitute for. These are not cells that merely raise an alarm; they issue orders to a tissue.
That reciprocity has forced a second reconsideration, this time of the epithelium. It is easy to picture the gut lining as a passive wall with immune cells stationed behind it. The evidence does not support the picture. Colonocytes targeted by Citrobacter rodentium — a natural mouse pathogen that attaches to the distal colon2 — turn out to recruit T-cell help to precisely the cells under attack, which implies a conversation far more specific than diffusion would allow. Whether the epithelium speaks to CD4 T cells directly, presenting antigen on its own class II molecules, is among the questions now open in the laboratory.
Then there is memory. A tissue that has met a pathogen is not the tissue it was before; some of the cells recruited to the fight stay behind, resident, poised. Following those cells requires watching the same animal across weeks rather than sampling a single day, and it requires reporter and fate-mapping models that record where a cell has been and what it once expressed. Much of the lab's method is built around that requirement: to ask not only what a cell is, but how it got there and what it will do next.
The stakes are not confined to infection. The same programs that defend the colon drive disease when their control is lost — in inflammatory bowel disease, in the passage from chronic inflammation to colorectal cancer, in the dysbiosis that precedes late-onset sepsis in newborns. The ambition is not to silence these responses but to understand them well enough to redirect them: to know which signal, at which moment, decides whether a T cell defends a border or destroys it.