Search

A major process through which the immune system becomes tolerant to self proteins involves the deletion of self reactive cells in the thymus. However, T cells reactive to peripheral tissue-specific proteins can escape this deletion and become tolerized in the periphery by a variety of mechanisms. We report here, contrary to expectation, that the pancreas-specific protein, elastase I, is also expressed at a low level in the thymus, and that this thymic expression contributes to tolerance induction. To study the mechanism of this tolerance induction, we utilized a double transgenic mouse model. In these mice the expression of a model protein, SV40 T antigen, is directed by the elastase I promoter and hence parallels elastase I expression in the pancreas and thymus. These mice were crossed with mice transgenic for a TCR specific for T antigen, so the majority of thymocytes and T cells in these mice express the transgene. In double transgenic mice we find that thymic expression of T antigen results in anergic thymocytes which also show a reduction of Th1 activity with no decrease in Th2 activity. These functional characteristics persist in peripheral T cells, but there is also a depletion in the number of T antigen reactive T cells in lymph nodes. Chimeras were constructed which directly demonstrated that the thymus is the site of tolerance induction and that the tolerizing element is thymic epithelium. We propose that the loss of Th1 activity as a consequence of the thymic epithelium being encountered by tissue-specific proteins results in the functional tolerization of CTL in vivo, despite the fact that CTL are fully functional in vitro. In this way autoimmune destruction is contained. Thymic expression of peripheral proteins may therefore be an additional way in which tolerance to peripheral proteins can be achieved.

In this study, it was observed that the B lymphoid cell line RPMI 8392 produced a tumor cell cytostatic factor that is of lower molecular weight than previously described cytostatic proteins but of greater molecular weight than arachidonate metabolites. This low molecular weight factor was also found to be produced by normal murine lymphocytes upon mitogenic stimulation. Gel filtration chromatography revealed that the molecular weight of this molecule is approximately 6,500 daltons. This is the approximate molecular weight of another tumor cell directed lymphokine which we have previously described, Tumor Migration Inhibition Factor (TMIF). This lymphokine is defined by its ability to inhibit the in vitro migration of a variety of tumor cell types. It was not determined in this study whether or not the proliferation inhibitory activity and the migration inhibitory activity are due to separate molecules. However, experiments presented here demonstrate that these two activities are functionally distinct. In addition, it was found that the cytostatic factor described here, unlike higher molecular weight cytostatic factors, was not associated with cytotoxicity.

Tumor cells can adhere to endothelial cell monolayers in vitro. The kinetics of this reaction are rapid; 50% of maximal binding occurs by 30 min of incubation. In the case of the P815 mastocytoma, the maximal percentage of binding is approximately 70%, suggesting that there are both binding and nonbinding tumor cell populations. Binding is independent of tumor cell dose over a 200-fold range of cell concentrations. Lymphokine-containing preparations were found to markedly suppress the binding of either P815 mastocytoma or Ehrlich ascites cells to endothelium. This effect appeared to be due to both diminished attachment and enhanced dissociation. The activity is found in the same molecular weight range as tumor migration inhibition factor (TMIF), and is not found in preparations lacking TMIF activity. Thus, the factor may prove to be TMIF itself or a lymphokine related to it. Of equal interest is the possibility that it represents a previously undescribed factor.

Trypsinization of P815 mastocytoma cells interferes with their ability to bind to endothelial monolayers in vitro, suggesting the involvement of proteins in cell-cell binding. Binding is also dependent upon divalent cations. Incubation of tumor cells with tunicamycin, which blocks protein glycosylation, or the monosaccharide N-acetyl-D-glucosamine could also inhibit binding. In addition, antibodies prepared against whole P815 cells had the ability to interfere with tumor cell binding. The protein fragments released from P815 cells by trypsin interfered with the binding of fresh P815 cells to endothelium. Moreover, those fragments as well as CHAPS extracts of tumor cell membranes were able to neutralize the inhibitory effect of an anti-P815 antibody on tumor cell-endothelial cell binding. These observations, taken together, strongly suggest the presence of adhesion molecules on the tumor cell. The anti-P815 antibody preparation could also inhibit endothelial binding of an unrelated tumor. Ehrlich ascites (EA). In addition, membrane preparations of EA could neutralize the antibody's effect on intact P815, and membrane preparations of P815 could neutralize the antibody's effect on EA. These observations suggest that there are common epitopes on these tumor cells and raise the possibility that a variety of tumors may share common mechanisms for attachment to endothelium. This, in turn, raises the possibility that monospecific antibodies against such epitopes might be useful agents for interfering with metastasis in vivo.