Journal of Autoimmunity (1990) 3 (Supplement), 87-90

Tolerance and Diabetes in Transgenic MiceOver-expressing Class I Histocompatibility

Molecules in Pancreatic PCells

Jacques F. A. P. Miller, Janette Allison and Grant Morahan

The Walter and Eliza Hall Institute ofMedical Research, P.O. Royal MelbourneHospital, Victoria 3050, Australia

The Class I gene, H-2Kb, was linked to the rat insulin promoter and theconstruct inoculated into fertilized mouse eggs to produce lines of trans­genic mice. Mice which expressed the Class I molecule in the pcells of thepancreas developed diabetes and progressive loss of their pancreaticp cells. This occurred whether the transgene product was syngeneic orallogeneic with respect to its host. No lymphocytic infiltration was everseen in transgene expressing mice, even in those deliberately Immunizedwith H-2Kb-bearing cells. When the transgene product was allogeneic,spleen cells from the transgenic mice stimulated in vitro with irradiatedB 1O.A(5R) cells (KbDd ) , could kill H_2 d targets in vitro, but not targets bear­ing H_2Kb • Responsiveness of spleen cells to H-2Kb targets returned withadvancing age, as the severity of diabetes increased. The results indicatethat diabetes in this model occurs independently of the immune system,and point to an extra-thymic mechanism oftolerance induction dependenton the continuous presence ofantigen.

Introduction

The immune system has to face a multitude of foreign antigens but must not reactagainst self-components. This self-tolerance is not preprogrammed in the genomebut is acquired somatically by various mechanisms that purge or silence self-reactivelymphocyte clones. One way in which self-tolerance is established has recently beendocumented for T lymphocytes. The receptor on T cells specific for the Class IImolecule, I-E, can be identified by a particular monoclonal antibody. In miceexpressing this molecule, cells with such a receptor were present among the imma­ture thymus lymphocytes in the cortex, but absent from the mature population andfrom the peripheral T-cell pool [1]. A similar situation pertains to another antigenicsystem where the determinants, known as minor lymphocyte stimulating antigens,act as strong stimuli for the proliferation ofT cells [2, 3]. The best interpretation of

87

0896-8411/901020S87 +04 $03.0010 © 1990 Academic Press Limited

88 J. F. A. P. Miller et al,

these data is in terms of the clonal elimination of self-reactive T lymphocytes duringtheir differentiation phase within the thymus.

If self-tolerance is induced in differentiating T cells to antigens synthesized bythymic stromal cells, a major dilemma has to be faced. Must the entire array of thebody's self components be manufactured in the thymus to enable self tolerance to beimposed in T cells? Alternatively, are unique extrathymic autoantigens shed into theblood stream to ensure that the thymus is constantly bathed in self-peptides? It seemedimportant, therefore, to determine whether T lymphocytes could ever become toler­ant to antigens not expressed in thymus tissue. To this end, we used the transgenicmouse approach [4], as it offers many unique advantages, including the ability to

introduce specific genes into inbred strains, thereby allowing those genes to beexpressed and treated as selfmolecules. Moreover, linking the gene to a heterologouspromoter enables its expression-to, be directed to specific cell types. This allows, forexample, examination of the in vivo effect on the immune system of the expression ofhistocompatibility antigens in defined cell types, in the absence of any trauma andinflammation which always occurwhen tissues are transplanted surgically.

p-cell dysfunction

The MHC H-2Kb Class I gene [5] was linked to the rat insulin promoter (RIP) andthe DNA construct micro-injected into fertilized mouse eggs of different haplotypes.These were then implanted into the oviducts of pseudo-pregnant mice. As a result,the H-2Kb molecule became expressed in the pancreatic islet pcells of the transgenicoffspring. We could not detect gene expression in the thymus by Northern blotsor by the sensitive Sl nuclease mapping technique. We therefore expected aT-cellreaction against the ~ cells, and hence insulin-dependent diabetes mellitus (IDDM),but only in mice in which the transgene product was allogeneic. T cells differentiat­ing in the thymus could not have encountered such a product and the clone ofT cellsable to respond to it should not therefore have been eliminated. On the other hand, noreaction was expected in mice expressing a syngeneic transgene H-2Kb product, i.e.one identical with a selfMHC molecule. We were therefore surprised to find that thedisease occurred at a very early age, regardless of whether the mice were syngeneic orallogeneic with respect to the transgene product and in the absence ofT-cell involve­ment [6]. Histology of the pancreas from diabetic mice revealed markedly abnormalislets, disorganized in structure and depleted of p cells, but containing normalnumbers of glucagon and somatostatin-producing cells. Analysis of transgeneexpression in the pcells showed expression ofH-2Kb as early as day 16 ofembryoniclife and throughout the period of study until the islets became grossly depleted of ~cells by day 40-50. Studies on cultured pancreatic islets, isolated from lS-day-oldembryos, showed significant reduction in long-term insulin secretion. A markedfunctional impairment ofthe ~ cells were thus preceded and presumably predisposedto the development of diabetes in these mice.

No lymphocytic infiltration was observed at any stage in transgene-expressingmice whether the transgene product was syngeneic or allogeneic with respect to itshost. Furthermore, transgenic mice depleted of T lymphocytes by neonatalthymectomy also developed IDDM [6], as did nude mice transgenic for the H-2Kb

gene. The disease could be controlled, to some extent, by injections of isophaneinsulin and could be cured by grafting normal syngeneic islet tissue.

Tolerance and diabetes in transgenic mice 89

How does the over-expression of Class I MHC molecules in the pcells result infunctional impairment? Among the possible mechanisms to account for ~-cell deathare the following. First, over-expression of any molecule under the direction of theinsulin promoter may reduce insulin biosynthesis by competing at transcriptional,post-transcriptional or translational levels. This does not seem to be the case in twoother transgenic models in which the human insulin [7] or placental lactogen [8]genes were expressed in and secreted from mouse pancreatic ~ cells. Second, anincrease in the production, transport and insertion ofa membrane protein in the ~ cellmay be associated with cytotoxic effects. Expression of the herpes simplex virusglycoprotein D in 13 cells, however, failed to induce diabetes (T. Stewart, personalcommunication). Third, the Class I molecules may bind some protein synthesized bythe ~ cells and thereby interfere with their function. A characteristic of MHC Class Imolecules is their ability to bind peptides [9], to interact non-covalently with mem­brane receptors involved in metabolism and growth [lO]j and to associate with viralglycoproteins [II]. Some interaction could conceivably occur between the abundanttransgene product with some receptor, or even with either proinsulin or insulin itself.This may result in interference with the normal insulin processing and secretionpathway and eventual death of the cell. Ultrastructural studies are in progress todetermine whether there are abnormalities in the secretion pathway of insulin andthe Class I transgene product.

AIlo-class I transgene induces tolerance

The absence oflymphocyte infiltration in mice expressing an allogeneic transgene onthe ~-ce11membrane [6] does not necessarily imply that Class I-reactive T cells havebeen clonaUy eliminated. One possibility is that T cells are not tolerant of the trans­gene product but are unable to react to it due to lack ofasecond signal from the ~ cells.Such signals, possibly including interleukin 1, may be delivered only by certain celltypes, such as macrophages and dendritic cells [12] which did not express the trans­gene. To bypass the requirement for ~-cell co-stimulation, the transgenic mice wereactively immunized with C57BL/6 spleen cells, using procedures known to activateT cells in allogeneic situations. Although this was done at an age when ~ cells werestill present and expressing Class I molecules on the cell membrane, none ofthe mice showed any lymphocyte infiltration in the pancreatic islets. To test fortolerance to the transgene product, spleen cells from 12-day-old transgenic mice ofendogenous H-2' haplotype were stimulated in a mixed lymphocyte reaction withBlO.A(5R) cells (haplotype KbD d

) . Cytotoxicity tests were then performed on51Cr-Iabelled H -2b or H _2d targets. While normal SJL spleen cells killed both targets,cells from the transgenic mice lysed the H_2 d targets but not the H-2b targets. Whenthe same test was performed in mice older than 5 weeks, spleen cells from transgenicmice were able to kill both targets. This was particularly striking when spleen cellswere taken from mice with advanced diabetes. Peripheral lymphocytes (presumablyT cells) from transgenic mice were therefore specifically tolerant of the transgeneproduct, H-2Kb

) but only in young mice. The acquisition of immune responsivenessto H-2Kb coincided with the occurrence of diabetes which was accompanied by asevere loss in transgene expressing ~ cells. In vivo tolerance to the transgene productwas therefore maintained only during the period in which Class I MHC molecules

90 J.F. A. P. Miller et al,

were expressed in ~ cells. This occurred in spite of the fact that transgene expressionwas not detectable in the thymus by highly sensitive techniques.

When spleen cells from young transgenic mice were stimulated in vitro with asource of IL-2, the stimulated cells were cytotoxic to both H_2b and H_2d targets.Tolerance in our transgenic model is thus clearly not imposed by intrathymic clonalelimination. Whether it results from T cells being made anergic following recog­nition of the Class I molecule in the absence of IL-2 stimulation remains to bedetermined.

References

1. Kappler, J. W., N. Roehm, and P. Marrack. 1987. T -cell tolerance by clonal eliminationin the thymus. Cell 49: 273-280

2. Kappler, J. W., U. Staerz, J. White, and P. C. Marrack. 1988. Self tolerance eliminates Tcel1s specific for the MIs modified products of the major histocompatibility complex.Nature 332: 35-40

3. MacDonald, H. R., R. Sneider, R. K. Lees, R. C. Howe, H. Acha-Orbea, H. Fetenstein,R. M. Zinkernagel, and H. Hengartner. 1988. T-cel1 receptor Vp use predicts reactivityand tolerance to Mls'-encoded antigens. Nature 332: 40-45

4. Brinster, R. L. and R. D. Palmiter. 1986. Introduction of genes into the germ line ofanimals. Harvey Lectures Series 80: 1-38

5. Weiss, E., L. Golden, R. Zakut, A. Mel1as, K. Fahrner, S. juist, and R. A. Flavell. 1983.The DNA sequence of the H-2Kb gene: evidence for gene conversion as a mechanism forthe generation of polymorphism in histocompatibility antigens. EMBO J. 2: 453-462

6. Allison, J., 1. L. Campbell, G. Morahan, T. E. Mandel, L. Harrison, and J. F. A. P.Miller. 1988. Diabetes in transgenic mice resulting from over-expression of class I histo­compatibility molecules in pancreatic ~ cel1s. Nature 333: 529-533

7. Bucchini, D., M.-A. Ripoche, M.-G. Stinnakre, P. Debois, E. Monthioux, J.Absil, J.-A.Lepesant, R. Pictet, and J. [ami. 1986. Pancreatic expression of human insulin gene intransgenic mice. Proc. Natl. Acad. Sci. USA 83: 2511-2515

8. Alpert, S., D. Hanahan, and G. Teitelman. 1988. Hybrid insulin genes reveal a develop­mental lineage for pancreatic endocrine cells and imply a relationship with neurons. Cell53:295-308

9. Bjorkman, P. J., M. A. Saper, B. Samraoui, W. S. Bennett, J. L. Strominger, and D. C.Wiley. 1987. The foreign antigen binding site and T cel1 recognition regions of Class Ihistocompatibility antigens. Nature 329: 512-518

10. Due, C., M. Simonsen, and L. Olsson. 1986. The major histocompatibility complex classI heavy chain as a structural subunit of the human cell membrane insulin receptor: impli­cations for the range of biological functions of histocompatibility antigens. Proc. Natl.Acad. Sci. USA 83: 6007-6011

11. Hochman, J. H. and M. Edidin. Associations between Class I MHC antigens andviral glycoproteins detected by fluorescence resonance energy transfer. In MajorHistocompatibility Genes and their Role in Immune Function. C. S. David, ed. AcademicPress, in press

12. Lafferty, K. J.,L. Andrus, and S. J. Prowse. 1980. Role oflymphokine and antigen in thecontrol of specific T-cel1 responses. Immunol. Rev. 51: 279-314