Life in the MHC

I began to read about the MHC in a serious way when I was thinking about what to do in my postdoctoral period. I thought that genetics was cool, and discovering new genes even cooler. Our idea was to make mice that had identical genetic backgrounds but differed in only one phenotype: the ability to stimulate a mixed lymphocyte response. We selected mice that, based on our serological typing, were identical but could still stimulate a response in the mixed lymphocyte test. We reasoned they if they could stimulate this response, genetically encoded differences must exist in the region located between K and D. If we then cross-immunized these mice with lymphocytes, we should be able to raise Abs that could distinguish those strains. We did exactly that experiment with strains A.TH and A.TL (1). These strains were critical not only in the production of Abs that described what came to be known as the Ia Ags or class II Abs, but they also set the course for immunology in the last flowering of the pre-MTA world. These strains were rapidly passed from laboratory to laboratory around the world. They allowed the rapid mapping of a great many immune traits to either class I or class II. Indeed, they were critical in the Nobel Prize-winning work of Doherty (2) and Zinkernagel (3) for deciphering MHC restriction and, indirectly, T cell selection.

Nonetheless, alas, ours was not the only lab to have the idea of making these interesting Abs. Along with us, three other groups published papers using the same general idea within a year (4–6). It is important to remember that the people in these labs were already friends, they had been graduate and postdoctoral students in the same labs, and the principal investigators all knew each other well. Indeed, the competing labs had all collaborated in the past and would do so again in the future. The identification of class II opened a huge new area. Personally, it hooked me on immunology forever; I wanted to understand how the MHC worked. A full understanding took another 15 years, with the eureka moment coming with the structure of peptide bound in the class I protein (7). Until then, the MHC kept many people busy. As an example, the first MHC meeting in Bar Harbor, ME, had about 30 attendees. The last MHC meeting had more than 600.

My lab made contributions to the growth of understanding of the MHC. In collaboration with Lee Hood's group, we showed that MHC genes could be identified by their reactivity with Abs following transfection (8). We pioneered the use of transfected cell lines as a way to study the function of different H2 genes and alleles (8–10). We were able to show that different MHC alleles had considerable dominance. For example, we showed that nearly the entire lymphocytic choriomeningitis virus-specific CTL response of BALB/c mice was restricted by the L^d molecule (11). Responses to K^d and D^d were not detectable using target cells transfected with those genes.

The problem of peptide specificity intrigued us. How is the specificity of the immune response impacted by the process of peptide generation and peptide binding? Clearly, degenerate binding existed, but how was this degeneracy created and how did it affect T cell recognition? We approached this problem in two ways. First, we asked how the MHC structure really changed the peptide and T cell recognition by mutagenesis of the class I protein itself (12–14). We then showed that “transplantation” of the binding pocket residues could transform the peptide-binding requirements of HLA-B27 into HLA-A2, but substitution of the peptide residues in a flu epitope could restore recognition by B27-restricted CTLs (15). Thus, the binding constraints and the T cell recognition constraints were separate.

These studies led us to consider whether improving peptide binding by altering the MHC contact residues, but not TCR residues, would improve immunogenicity and T cell recognition. We initially showed this using an HIV epitope but extended it to the breast cancer epitope Her2 (16–18). In this case, we demonstrated much improved immunogenicity in animal studies and ultimately conducted a phase 1 trial of the concept, using autologous dendritic cells (19).

In the past few years, our lab has focused on two areas: infectious disease immunology and autoimmunity. Both were driven in large part by my wonderful colleagues at the University of North Carolina, who keep telling me how important and interesting these areas are.

The pathogenesis of infectious disease has traditionally been studied quite separately by microbiologists and immunologists. For microbiologists, the immune system is viewed as a constant inhibitory pressure against proliferation of the microbe. This position makes immunologists cringe, as we all know that the kind, not simply the amount, of immune response has a critical impact on the host and the outcome of infection. Even when outcomes have been plainly demonstrated, such as the Th1 and Th2 response to Mycobacterium leprae (20), microbiologists ignore that as confusing immunology. I am embarrassed to say immunologists are not much better; we have a tendency to view microbes as convenient self-replicating Ags (I attribute this remark to Bob Johnston at the University of North Carolina).

We began a long collaboration with Tom Kawula, an outstanding bacterial geneticist, in an attempt to understand the interaction between Francisella tularensis and its mammalian host. Our goal from the beginning was to comprehend how the bacterium adapted to the changing host environment and how those changes in turn influenced the host.

Early in our efforts, we noted that cultures with Francisella-infected APCs did not stimulate T cells effectively (21). This turned out not to be a defect in Ag processing but was mediated by secretion of PGE₂ by the infected cells. PGE₂ secretion not only impacted immune responses in vitro but also resulted in diminished Th1 responses in the lungs of infected mice (22). Together with the Kawula lab (and a library supplied by Colin Manoil), we found that most of the genes responsible mapped in the Francisella pathogenicity island. These seem to be a combination of a secretion system and effector molecules. We continue to try to understand the mechanisms in the interactions of Francisella and the host response.

Our interest in autoimmunity was brought to fruition by our collaboration with Roland Tisch's group. Type 1 diabetes is characterized by the destruction of beta cells in the islets of Langerhans, resulting in the inability to regulate glucose levels. The pathogenesis is complex, involving CD4, CD8, and regulatory T cells. Our work has focused on understanding the relationship of the responses in the periphery to the islets as well as developing Ag-specific therapies using toxin-coupled MHC tetramers. We first focused on a simple question: Are the T cells in the periphery the same as those in the islets? The answer, based on TCR sequencing, is that the populations are different, but overlapping (23–26). Indeed, when autoantigen-specific T cells sequenced from the spleen, pancreatic lymph nodes, and islets were examined, they had some shared clones, but many were found only in the islets. Further, we found evidence that proliferation of autoantigen-specific cells in the islets was significant, consistent with an islet's specific population that was independent of the responses in the periphery. The implications of understanding autoimmunity in humans are clear—one cannot monitor the periphery and assume that the same mix of cells is present in the target organ.

Our other interest has been in trying to create platforms for testing and treating type 1 diabetes. In this regard, Paul Hess (27) created saporin-coupled MHC class I tetramers and was able to show that they not only bind to specific T cells but also delete them in vitro and in vivo. Further, as drugs, they can be used to delete autoreactive cells, and their deletion is surprisingly durable, with many treated mice showing no recurrence of T cells of that specificity (28). In the same vein, we developed NOD-scid-γc^null mice transgenic for a complete human HLA class I gene. Expression of the complete gene is important both for survival of transferred CD8 T cells that require engagement of CD8 with class I for survival and for efficient killing of autoantigen-expressing cells. When these mice are reconstituted with PBMCs from diabetic donors, but not normal nondiabetic donors, the mice develop insulitis and have Ag-specific T cells in those islets that produce IFN-γ (29). This method will offer us a way to test new therapies and to evaluate old ones with a large number of donors and treat the same immune system with different therapies to provide a side-by-side evaluation.

Our final new venture is a return to my genetics roots. The collaborative cross is a novel set of inbred strains derived from crossing eight inbred strains, including three wild-derived stains. This platform will allow us to infer, with ease, the genetics of complex traits in the mouse (30). In addition, it will provide a way to confirm the genetics using crosses of these recombinant inbred mice. Finally, these animals will allow us to potentially perform mouse-on-man mapping, using synteny of short regions.

I thank many agencies for the support of our work. These include the Jane Coffin Childs Fund, the American Cancer Society, the Leukemia Society, the Damon Runyon Walter Winchell Fund, the Juvenile Diabetes Research Fund, the American Diabetes Association, the National Cancer Institute, and, most importantly, the National Institute of Allergy and Infectious Diseases.

The author has no financial conflicts of interest.