It has been clearly established that MHC antigens have a central role in cellular immunity. However, evidence continues to mount that these molecules, particularly class I MHC antigens, have other roles to play as well. The evidence ranges from genetic correlations between MHC haplotype and cell and animal physiology (Edidin, 1988; Walford, 1987) through effects of transfected MHC genes on tumorigenicity (Wallich et al., 1985; Gattioni-Celli et al., 1988; Sunday et al., 1989) to immunochemical and biophysical demonstrations that the class I antigens are physically associated with peptide hormone and other receptors (Samson et al., 1986; Phillips et al., 1986; Schreiber et al., 1984; Due rt al., 1986). These associations, do not appear to be random, though their relevance to cell biology is not at all clear and even the molecular basis for MHCjreceptor interactions remains to be resolved. However, the number of such associations reported compels one to take account of “non-immune” functions of MHC antigens and to attempt to integrate them into a largescale theory of MHC function. At best such a large-scale theory would give us insights into the evolutionary origins of MHC antigens, and into the wider range of functions subsumed by these molecules. At worst, a full underst~ding of the interactions between MHC antigens and receptors would show these up to be epiphenomena-reflecting some molecular property important in the immune response which only incidentally causes associations between MHC molecules and other surface receptors. My laboratory began working in the area of “nonimmune” functions of MHC over 15 years ago. The work of Ivanyi (1978) on MHC and organ weights, as well as some details of the then current view of antigen presentation, stimulated us to look at the possibility that MHC haplotype affected peptide hormone binding and the ensuing hormone-stimulated changes in cell metabolism. We could not be sure which hormone receptors might be important in these studies, and even had we known, there were few radiolabeled ligands available to probe them. Therefore, it was decided, instead, to measure levels of “second messenger”, cyclic AMP, in mouse liver as a function of MHC genotype and phenotype. In liver these levels mainly reflect glucagon and insulin binding. By comparing congenic-resistant mouse strains, differing only at the MHC, and by following segregation of MHC and CAMP levels in genetic crosses, it was shown that liver CAMP levels were strongly linked to MHC genotype (Meruelo and Edidin, 1975; Lafuse et al., 1979). Lafuse also showed that glucagon-stimulated, but not basal, adenylate cyclase activity was similarly linked to the MHC. Lafuse then showed that this MHC association was at the level of glucagon or insulin binding to its receptor (Lafuse and Edidin, 1980). Comparison of congenic-resistant lines differing at only parts of the MHC, together with some developmental studies, led us to the conclusion that it was the class I MHC antigens that affected peptide hormone binding, and that H-2D was more important in this respect than H-2K (Edidin, 1986). Much of this work was done before the development of the modern armamentarium of cell biology and immunology, particularly molecular biology, with its ready access to directed mutation and expression of genes in new host cells, and hybridoma technology, producing well-characterized high affinity monoclonal antibodies. We have made new progress in the analysis of MHC/receptor interactions, by using the new techniques to test and extend the ideas about the interactions that we had developed using classical immunogenetics and biochemistry. We have turned our main interest from glucagon receptors in liver to insulin receptors, especially those found in activated B-cells. This allowed us to take full advantage of a remarkable series of HLA mutant B-lymphoblasts prepared and characterised by Robert DeMars and his laboratory. The derivation of these mutants is summarised in a recent reference (Shimizu et af., 1986). DeMars began by immortalizing HLA-heterozygous B-cells, transforming them with Epstein-Barr virus. The parent line, HLA-Al, B8, C/A2, B5, C, was gamma-irradiated and HLA loss mutants were selected in terms of their resistance to lysis by anti-HLA antibody and complement. The first sorts of mutants selected had deletions of one of the two HLA-bearing regions of chromosome. A