Before there was a defined T cell receptor, Peter Doherty and Rolf Zinkernagel deciphered the rules that governed its recognition of infected target cells.
Immunological compatibility in the 1970s was defined by transplantation antigens—now known as the molecules encoded by the major histocompatibility complex (MHC). These proteins dictated the compatibility of transplanted tumors among different strains of mice. They also set the stage for Rolf Zinkernagel and Peter Doherty to discover the phenomenon of MHC restriction of antiviral responses. For this discovery they were awarded the 1996 Nobel Prize in Physiology or Medicine.
In 1973, Zinkernagel and Doherty combined forces on studies of lymphocytic choriomeningitis virus (LCMV), which in mice causes a fatal neurologic disease with T cell invasion of the central nervous system (CNS). By combining their talents—Doherty doing spinal taps on mice, and Zinkernagel the 51Cr-release cytotoxicity assays—they demonstrated that the T cells recovered from the cerebrospinal fluid of infected mice were antigen-specific cytotoxic T cells (CTLs) that could destroy LCMV-infected L cells (mouse fibroblast cells) (1). This supplied the test system used in the pivotal MHC restriction study.
Michael Oldstone and colleagues had suggested that different strains of mice showed subtle differences in susceptibility to LCMV infection (2). Intrigued, Zinkernagel and Doherty set out to determine whether these differences were a function of cytolytic T cell activity in the CNS. They collected the available strains of mice—a grand total of four—and tested their CNS-invading T cells for the ability to destroy infected L cells. All the mice eventually died, but surprisingly only some of the strains developed cytolytic T cell responses in vitro (3).
Genetic background and fortuity explained this finding. The available reagents had been a mouse strain (CBA) and a cell line (L cells) that both expressed MHC molecules of the H-2k haplotype. It was this combination that showed an effective cytolytic T cell response. Thus, the ability of T cells to recognize target cells was a function of both antigen and MHC haplotype. They proved this formally by replacing the L cells with macrophages from each of the mouse strains. Immune T cells could only recognize infected target cells that shared the same H-2 haplotype (3). Although this made sense based on prior reports of MHC requirements for T–B cell collaborations, the results were unexpected. “Nobody had thought of such a limitation on the interaction between T cells and virus-infected cells,” notes Zinkernagel.
The duo fine-tuned the rules governing MHC restriction using congenic and recombinant mouse strains to pinpoint the restriction element to the K and D regions (class I) of the MHC complex and to rule out involvement of the I region (class II) (4, 5). These studies were published in a series of articles in the Journal of Experimental Medicine.
A contentious model
There were two possible explanations for the influence of the MHC on T cell recognition of infected cells. The first was that T cells express two receptors, one for identification of self cells that was dictated by like-like MHC interactions, and another (T cell receptor) for recognition of antigen. The second possibility—the one they favored—was that the T cell receptor recognizes a virus-induced modification of the MHC product—the “altered-self” hypothesis.
The altered-self model was not popular among many of Zinkernagel and Doherty's contemporaries. But their later experiments showed that T cells from the F1 (AxB) offspring of two MHC-disparate mice recognized only the altered self of MHC-A target cells or MHC-B target cells, but not both (6). This result cast doubt on the dual receptor model in which F1 T cells, having both A and B MHC molecules, should have recognized both targets.