On lawnmowers and lay‐down misères
L. Smallwood,A. Baxter
DOI: https://doi.org/10.1111/J.0019-2805.2004.01840.X
2004-03-01
Immunology
Abstract:When one of us (A.G.B.) was rather younger than he is today, he acquired the solemn responsibility for mowing the family lawn with a somewhat battered two-stroke motor mower. Each week, he would wheel it out, prime the carburettor and pull the starter cord. Occasionally it would start, but most weeks it would not. Out came the spanners, screwdrivers and a handy hubcap to keep all the bits, as the motor was stripped down and rebuilt. Finally, it would be repaired and the lawn mowed.1 But a paradox remained, for every time the mower was stripped down and rebuilt, a few pieces were left over: a split washer, a few screws, a nut or a strangely shaped piece of metal that seemed to fit… nowhere. After some years of this treatment, the question had to be asked: How few pieces can a lawnmower have and still run?
The contemporary history of experimental models in immunology has taken something of an opposite path. Once the T-cell receptor (TCR) was cloned2–4 and the structure of the major histocompatibility complex (MHC)-encoded ligands determined5 (and that nasty business with I-J swept under the carpet), there was a general feeling that the vertebrate immune system did not need many parts. It should be possible (it was reasoned) to create an experiment that would demonstrate at what point tolerance was induced and what cell or factor was responsible for its production. And so the three pieces of the machine were assembled: TCR; peptide antigen; and restriction molecule. The critical tool in the assembly was the use of transgenic technology.
In order to be able to track the destinies of the T cells in question, some mechanism of examining T cells with high precursor frequencies was required. One of the earliest attempts was by Allison et al., who expressed the class I histocompatibility gene, H-2Kb, in mice under the control of the rat insulin promoter.6 Being a major alloantigen, the precursor frequency for cells specific for H2-Kb on a B10.BR (H2k) background should have been of the order of a few per cent. A simple enough system. The result was also clear enough – with an early onset of insulin-dependent diabetes mellitus, the islets were destroyed. But there was no evidence of immune activation. The pathology appeared to be an artefact associated with the overexpression of H2-Kb in the islet beta cells.
An alternative approach was to make TCR transgenic mice carrying a large cohort of potentially self-reactive T cells. For example, Kisielow et al. expressed a TCR specific for the male H-Y antigen in male mice.7 The result was not a model of autoimmunity, because the antigen was expressed in abundance in the thymus and resulted in the deletion of high-avidity cells.
Finally, the master stroke: both antigen (H-2Kb expressed in islet cells again) and the antigen-specific TCR controlled through transgenic expression.8 The result was a surprise; there were large numbers of potentially self-reactive T cells circulating that were tolerant to the H2-Kb expressed on the islets, or even on transplanted skin grafts. Evidence for a peripheral mechanism of tolerance induction? Well, perhaps not, as Smith et al.9 have shown that the rat insulin promoter can drive transgene expression in the thymus.
Two groups developed a similar model, but with a twist. Ohashi et al.10 expressed a transgene for the glycoprotein antigen from lymphocytic choriomeningitis virus (LCMV) in the pancreatic islets and then expressed a TCR specific for this antigen, in the context of H-2Db, on T cells. The double-transgenic mice tolerated the viral antigen in their islets until they were infected with live virus. The LCMV was rapidly cleared, and so were the viral antigen-expressing pancreatic cells. Oldstone et al.11 found that LCMV infection of mice transgenically expressing the viral nucleoprotein in their islets also triggered diabetes in previously healthy individuals, even without any TCR transgene; the immune activation and clonal expansion in response to the viral infection was enough to cause tissue destruction in its absence.
These findings led to the demonstration that it was possible to break T-cell tolerance in transgenic models of peripheral foreign antigen expression by the local expression of costimulators or inflammatory mediators.12,13 But the issue was not such a black-and-white matter when it came to the central nervous system. Transgenic mice expressing TCR specific for the central nervous system antigen myelin basic protein occasionally developed spontaneous demyelinating disease when housed in conventional conditions.14 Unfortunately, no pathology was seen in the same transgenic line housed in clean conditions – so it is possible that in this case too, autoimmune disease was caused by LCMV, which is a common mouse pathogen.
In this issue of Immunology, de Jersey et al.15 further characterize their previously described model of pituitary autoimmunity.16 The expression, in the anterior pituitary, of influenza nucleoprotein under the control of the human growth hormone locus control region, together with a TCR specific for a nucleoprotein epitope within the H-2Db restriction factor, resulted in spontaneous autoimmunity. The ensuing hypophysitis decreased growth hormone production to almost undetectable levels and caused severe growth retardation. Now, they revisit Oldstone's experiment and treat the healthy nucleoprotein transgenic, but TCR wild-type, mice with the appropriate virus and found that they, too, develop autoimmune destruction of the anterior pituitary. The importance of this finding, they suggest, is that contraception through immunization with human chorionic gonadotropin (hCG)-based vaccines might lead to autoimmune disease of the anterior pituitary. As this might significantly affect the production of thyroid-stimulating hormone and adrenocorticotrophic hormone, their concerns appear valid.
From the immunological point of view, their observation that full penetrance of the disease required the transgenes to be expressed on the recombinase deficient (Rag–/–) background was of greater interest. This was not the result of clonal expansion secondary to homeostatic proliferation, because the Rag wild-type TCR transgenic mice had a greater number of TCR transgenic T cells than those on the Rag-deficient background. It seems that all those extra non-transgenic T cells might be doing something after all.
And what eventually happened to the lawnmower? Why, it stopped working for good, of course. The strange thing was that the critical part, the one that was absolutely required to run, looked just like all the others – it was a nondescript nut.