09 Nov 2020
Issue #32:The Discovery
Setting it Straight - Issue #32
Written by Nobel Laureate Professor Peter Doherty
Following on from the past three weeks…
Back when I joined the John Curtin School of Medical Research (JCSMR) in late 1973, their animal breeding facility was producing three inbred mouse lines – CBA/H, BALB/CJ and Miss Abbie Lathrop’s C57BL/6J (B6). The terminology for the H-2 system MHC class I (MHCI) genes is arcane and historic (the H-2ᵏ CBA/H strain is H-2KᵏDᵏ). Simplifying that here, we’ll identify the mouse H-2 haplotypes as X, Y and Z and the transplant genes mapping to the MHC class I (MHC1) H-2K and H-2D loci as: X1X2 (CBA/H), Y1Y2 (BALB/CJ) and Z1Z2 (C57Bl6/J). All six of these X1X2 Y1Y2 and Z1Z2 genes are completely different with there also being no shared themes for those at the H-2K versus the H-2D locus.
Elsewhere on campus, the Zoology Department was breeding the ‘natural’ H-2 recombinant A/J (X1Y2) and, in the USA, the geneticists had, in their efforts to define the I-A region (MHCII) developed a whole range of ‘congenic’, (identical for all except the H-2 genes) recombinant inbred strains that were, for instance, Z1Y2 and X1Z2. In addition, Russian researcher Igor Egorov had made a mouse with single point mutation (Z1mZ2) that rejected skin from the original Z1Z2 ‘wild type’ strain. All these were later to prove invaluable.
I’d come to Canberra to learn both the conceptual framework and the techniques that Bob Blanden used to establish the role of T cell-mediated immunity (CMI) in the bacterial infection Listeriosis (caused by Listeria monocytogenes), then in ectromelia (mouse pox). The protocol was to infect a group of X mice, let them recover, then transfer (by intravenous injection) their ‘donor’ immune lymphocytes into a second set of X ‘recipients’ that had been recently infected. The ‘adoptively transferred’ donor T cells ‘cured’ the recipients in three to four days, compared with the seven to 10 days for their own, ‘naïve’ immune response to do that.
Apart from my sheep work with louping-ill virus in Edinburgh, I’d also experimented with mice and hamsters (an optimal species for preclinical studies with SARS-CoV-2). I knew my way around a laboratory animal house, though doing tail vein injections in mice was a new experience! Continuing the meningoencephalitis theme, I’d developed a method for counting cells in mouse cerebrospinal fluid (CSF), with the intent of quantifying the role of CMI in this inflammatory process.
Initial studies with the Semliki Forest alphavirus proved to be not very useful: too much non-specific inflammation. Lymphocytic choriomeningitis virus (LCM), where Bob and Cedric Mims had shown that transferred immune cells can eliminate the virus, provided the ideal experimental system. In the absence of CMI, LCMV causes minimal pathology. By mid-1973, I’d worked out how immune T cells cause the severe meningitis characteristic of LCM and had shown how the host response kills mice in this experimental model of immunopathology.
At this stage, I’d been sharing the lab with Rolf Zinkernagel for several months and we were talking a lot. Rolf had tried in vitro cytotoxic T lymphocyte (CTL) assays earlier in Lausanne, though nothing worked. This involved infecting cells with the bacterium he’d been studying, then overlaying with immune lymphocytes from an immunised animal. In this assay, ‘killing’ is measured hours later by the release of radiolabel Na⁵¹Cr, detected by analysing the supernatant fluid in a gamma-radiation counter. The approach had caught on in the infection/immunity culture. Others, including Bob’s PhD student, Ian Gardner working with ectromelia, were trying CTL assays with LCM and different pathogens.
Aware that the CSF of LCMV-infected CBA/H (X) mice contains lots of ‘activated’ lymphocytes, Rolf and I decided to see if they would kill LCMV infected, Na⁵¹Cr-labeled target cells. Fortuitously, because they were being produced by the department’s ‘core’ cell production laboratory for the virologists, Rolf used L929 cells (L cells), a ‘fibroblast’ line derived from another strain of X mice. We were over the moon when the LCM CSF cells utterly destroyed the LCMV-infected L cells. The wipe-out was way beyond anything anyone had ever seen for a pathogen. Other experiments showed that LCMV-immune spleen cells were also potent killers. The findings were quickly written up and published in the Journal of Experimental Medicine, the top format (apart from Nature and Science) for immunology papers at that time.
Then we read an article co-authored by LCMV-expert Michael Oldstone (Scripps Institute), geneticist Hugh McDevitt at Stanford and Australian (from WEHI) Graham Mitchell who was a postdoctoral fellow with Hugh. They claimed that H-2-different mice showed varying susceptibility to LCM immunopathology. We decided to test that by seeing if we could see different levels of CTL killing from X and Z mice. What we found blew our minds. Though both strains developed equivalent levels of inflammatory pathology, the X immune ‘effector’ T cells killed the LCMV-infected X targets, while the Z-set did nothing. We didn’t have a convenient Z target but Rolf had, in his Listeria studies, been obtaining macrophages by peritoneal lavage. Using LCMV-infected macrophages, we confirmed that the X immune CTLs killed only the infected X targets and the reciprocal was true for Z.
At first, we thought that, in line with McDevitt’s thinking, this was probably an immune response gene (Ir gene) effect mapping to the H-2IA region between the K and D transplant loci. That would have made our findings interesting, but not totally unexpected. But then we ‘borrowed’ a few A/J (X1Y2) mice and did reciprocal experiments with X1X2 and Y1Y2 mice and LCMV-infected targets. There was massive excitement when we found that our CTL killing mapped to the strong transplantation genes (MHCI), not to the MHCII I-region.
Further experiments (many of which are even more complex) confirmed that, both for CTLs in vitro and for CMI in vivo, we saw T cell effector function wherever there was just one, shared MHCI gene (same letter, same number) between the immunised mouse and the infected target cell, with that effect also mapping to Egorov’s Z1m mutant. Suddenly, with a series of brief, 'Letters to Nature', publications in other leading journals and speaking tours in Europe and the USA, we went from total obscurity to fame and, for some, notoriety in the worlds of immunology and transplantation genetics. As I’ll summarise next week, with simple experiments that built on decades of work by others, we’d discovered why we have a transplant system. The MHC1 alleles are not primarily about organ graft rejection. These structural genes are central to ‘immune surveillance of self.’