| Abstract |
The goal of this work was to investigate how the systemic absence of TNF affects CD8 T cell tolerance to a defined autoantigen. Since TNF is known to have pleiotropic effects (pro-inflammatory, pro-apoptotic, modulation of APCs and Tregs), understanding the balance of these mechanisms in vivo is very important. To establish a suitable experimental system, we generated a novel transgenic mouse expressing influenza-NP under the pdx-1 promoter and crossed these to NP-TCR transgenic mice (F5) and TNF/Rag1-deficient backgrounds. One transgenic line, pNP31, expressed the NP and was used for all further studies. These mice expressed NP in multiple organs, notably in the pancreas, thymus, spleen and brain (RT-PCR). Initially, we began to attempt to break self-tolerance in such mice by transferring activated and non activated TcR (F5) transgenic cells recognizing the NP peptide, also in Rag1-/- and TNF-/- backgrounds. Only 2 out of more than 100 recipients developed overt diabetes and impaired glucose tolerance was seen in some mice, more in males than in females. Therefore, new double transgenic mice were generated by directly crossing the pdx1-NP (pNP31) mice with the F5 TCR transgenics. The outcome was interesting and encouraging, since approximately 40% of such double transgenics (F5/pNP31/Rag1-/-) developed diabetes when TNF was absent, whereas their TNF-expressing counterparts remained tolerant with only one out of more than 60 mice developing diabetes. Thus, the experimental model system was well suited to investigate CD8 tolerance in the absence of TNF in direct relevance to the development of autoimmune disease (type 1 diabetes). Those TNF-deficient mice with diabetes also exhibited profound insulitis, as one would expect. The fact that not all mice developed clinical disease is of interest and reminiscent of another spontaneous diabetes model, the NOD mouse. Likely environmental factors, for example inter-individual differences in gut flora and pathogen encounters, will play a role. In addition, minor inter-individual variations in background genes through the sophisticated breeding scheme will also be responsible for the fact, that disease penetrance is not 100%. This feature lends the model more realism for a human disease. Tolerance in F5/pNP31/Rag1-/- double transgenic mice was achieved by peripheral deletion and anergy of autoreactive F5 T cells, characterized by upregulation of CD44, CD69, CD25, downregulation of CD62L and inability to produce interferon-γ upon in vitro challenge with the NP antigen. These observations indicated that anergy of F5 cells was preceded by their initial activation, likely after encountering the NP antigen in the periphery, for example also in lymphoid organs. Thus the presence of NP, but not the absence of TNF lead to activation of T cells. Tolerance was achieved by T cell loss, likely thorough activation-induced cell death (AICD) as well as induction of anergy in the remaining F5 T cells. We next wanted to test, whether NP-antigen specific reactivity was affected in those cells that had not been deleted in pNP31 mice, when TNF was absent. To this purpose, spleenocytes were isolated and their proliferation was tested in the presence of various concentrations of NP peptide in vitro. Interestingly, proliferative capacity was enhanced selectively in the fraction of diabetic TNF deficient mice. We then analyzed the capacity to produce interferon-γ (by FACS) after NP peptide stimulation. In F5/pNP31/Rag1-/-/TNF+/+ mice this capacity is greatly impaired (functional anergy of remaining CD8 cells). In contrast, in F5/pNP31/Rag1-/-/TNF-/- mice, IFN-γ production was restored to levels usually seen in F5/Rag1-/-/TNF+/+ mice. Thus, in the absence of TNF, functional anergy of CD8 T cells is reversed. We also tested the in vivo cytotoxic capacity of CD8 cells by injecting NP-peptide coated CFSE labeled target cells in vivo. In accordance with the findings for proliferation and interferon-γ production, the results show enhanced cytotoxicity (significant) in TNF-deficient F5/pNP31/Rag1-/- mice compared to TNF+/+ counterparts. It is of interest to note, that F5/Rag1-/-/TNF-/- mice exhibited a strong amount of cytotoxicity that was NP specific, which shows that even in the absence of a cognate autoantigen (NP), spontaneous T cell activation of TCR transgenic NP-specific T cells can occur in the absence of TNF, possibly through crossreactivity with gut antigens. Autoimmunity did not result in F5/Rag1-/-/TNF-/-, because the self-antigen was not expressed, of course. Last, we wished to test the hypothesis, whether the absence of TNF from birth would lead to lesser degrees of apoptosis in diabetic F5/pNP31/Rag1-/-/TNF-/- mice, which could account for increased CD8 T cell numbers and lesser degrees of T cell deletion. Indeed, apoptosis was reduced in the diabetic mice, which provides a mechanistic explanation for the increased cellularity in these mice and explains in a quantitative way, why tolerance is lost. Conclusions: NP expression leads to activation of CD8 T cells in F5/pNP31/Rag1-/- mice. These cells usually apoptose in the presence of TNF resulting in reduced cellularity and self-tolerance by deletion. In the absence of TNF, activation-induced cell death is reduced, resulting in lesser apoptosis, lesser deletion and T1D in a significant proportion of animals, if NP is expressed. Interindividual variation correlates interestingly with disease development. Rag1-/- is required for autoimmune disease penetrance, possibly because it is known that lymphopenia can drive autoimmunity. In addition, functional anergy of CD8 cells that were not deleted in the presence of TNF is reversed in the absence of TNF. Thus, TNF is an important checkpoint for maintenance of self-tolerance influencing cellular apoptosis (AICD-deletion) and cellular functional anergy. Collectively, my data show that endogenous TNF can be essential for the induction and maintenance of peripheral CD8 T cell tolerance, and in its absence, organ-specific autoimmunity can occur.
|
Μελέτη μηχανισμών περιφερικής ανοσολογικής ανοχής των CD8 T λεμφοκυττάρων σε νέο διαγονιδιακό μοντέλο TCR/αντιγόνο στο ποντίκι
