Vinculación de la familia gitana en el aprendizaje de

Having made an attempt to address the issue of central tolerance to the P91A tum" antigen in transgenic mice an examination of in vivo responses to the P91A tum" antigen was conducted in non-transgenic mice. The expression of the P91A tum" antigen is critical for the successful rejection of the P91 tumour line as indicated by the failure of DBA/2 mice to reject P91A antigen loss variants (Uyttenhove, 1983). Interestingly, rejection appears to occur in the absence of an obvious CD4-f- T helper response. Additionally, PS 15 has no constitutive MHC class II expression and

attempts to induce MHC class II expression on P815 using y-interferon have been unsuccessful (T.Boon, personal communication). Although reprocessing and presentation of tumour antigens by host antigen presenting cells (macrophages or dendritic cells) may occu^the possibility exists that a CTL response may be mounted against the P91A epitope in the absence of T cell help. If this occurs then it is also possible that P91A transgenic DBA/2 skin may be rejected by non-transgenic DBA/2 mice. In other words the P91A epitope may be sufficient to constitute a minor histocompatibility antigen rejection barrier.

Using the P91A transgenic mouse model it has become evident that in itself, the P91A epitope does not constitute a complete minor histocompatibility antigen. This is evidenced by the acceptance of female P91A transgenic DBA/2 skin by non-

transgenic female DBA/2 mice (Antoniou et al, 1996). It has been further shown that this lack of rejection results from the absence of a suitable MHC class II helper determinant expressed in female transgenic donor skin. Since female DBA/2 mice fail to reject male DBA/2 skin an attempt was made to constitute a complete minor

histocompatibility antigen by combining the P91A epitope with the DBA/2 H-Y determinant (DBA/2 mice are an H-Y non responder strain as assessed by the failure of male DBA/2 mice to reject female DBA/2 skin). When this was done, it became clear that female non-transgenic mice could reject male transgenic skin and that the combination of the P91A epitope with a component of the H-Y antigen was sufficient to elicit graft rejection (Antoniou et al, 1996). The fact that H-Y can facilitate the rejection of P91A+ H-Y+ skin indicates that H-Y reactive cells are present in DBA/2 mice. Furthermore, mice treated with an anti CD4 depleting antibody failed to reject P9TA+ H-Y-k skin indicating that the H-Y component is likely to be a MHC class II helper determinant (Antoniou et al, 1996). In all experiments prior i/p immunisation with P91A+ H-Y+ spleen cells was required before any rejection of P91A-I- H-Y-k skin was observed. The protocol for induction of a rejection response is shown in figure 2.10 A.

The P91A model of skin graft rejection can be interpreted as displaying features of the two signal hypothesis of B cell activation (Bretscher and Cohn, 1970). This hypothesis states that a B cell must receive two consecutive signals in order to become activated: signal one via the antigen receptor and signal two from a second cell specific for a component of the same antigen. In this model the second signal is derived from antigen specific T helper cells and can largely be substituted by IL2. The two signal model of B cell activation and more specifically the requirement for CD4 T lymphocyte help has been applied to the activation of naive CD8+ T cells (Stuhler and Walden, 1993; Guerder and Matzinger, 1989 and 1992; Rees et al, 1990; Vandervegt and Johnson, 1993).

In the P91A model of skin graft rejection help is derived from H-Y specific CD4+ T cells and the effector cells are presumed to be the CD8+ P91A reactive T cells. This interpretation assumes that the CD4+ H-Y reactive cells do not directly participate in the rejection response but exert their effect through CD8-I- effector cells. This is likely to be the situation in DBA/2 mice which are known H-Y non responders. However, it has recently been demonstrated that CD8 knock out mice are capable of rejecting allogeneic skin demonstrating that in certain circumstances rejection can be mediated by CD4+ T cells (Krieger et al, 1996; Dalloul et al, 1996a). Furthermore a recent study has demonstrated that adoptive transfer of naive or sensitised CD4+ cells from CD8 knockout mice into nude mice, which have been grafted with skin from mice deficient in MHC class I or class II respectively can reconstitute rejection (Dalloul et al, 1996b). In this case, although MHC class I expressing allografts were rejected, CD4+ cells did not display allo-antigen specific cytotoxic activity, though they did proliferate in vitro in response to allo-antigen . Although CD4+ cells would seem capable of eliciting a rejection response independent of CD8+ cells the converse does not appear to be true of CD8+ cells in CD4 knockout mice (Krieger et al, 1996). Together, these results suggest a critical involvement of CD44- T cells in skin graft rejection.

P91A+H-Y+ skin graft rejection appears to conform to a model in which CD4+ and CD8+ T cells collaborate to effect rejection and as such are consistent with a two signal model of T cell activation in which CD4+ H-Y specific T cells are required in the activation of P91A specific CD8+ T cells. A further prediction of the two signal hypothesis is that a state of T cell tolerance may be induced if signal one is

encountered in the absence of signal two (Guerder and Matzinger, 1989, 1992; Rees et al, 1990). The characteristics of P91A+H-Y+ skin graft rejection allow the two signal model of peripheral T cell tolerance to be tested.

2.2.2.1

demonstration of peripheral tolerance to the P91A tum"

epitope

In order to test out the two signal hypothesis of peripheral T cell tolerance female DBA/2 mice were given an initial i/p injection of female transgenic or male non- transgenic spleen cells (representing signal I or signal II). After two weeks all mice were given a second injection of male transgenic spleen cells (representing combined signals I and II). A further two weeks later all mice were grafted with male transgenic skin and subsequently examined for signs of acute or chronic rejection. The

experimental protocol is outlined in figure 2.1 OB.

Prior immunisation with either the class I CTL epitope P91A or the class II helper epitope H-Y results in long term survival of P91A+ H-Y+ skin with a median survival time of over 100 days (MST > 100 days), table 2.1 A and B. By contrast, mice which did not receive an initial injection of P91A4- or H-Y+ cells reject P91A+ H-Y+ skin with a MST of 48 days, table 2.1 C. This result supports and extends the initial predictions of the two signal model of peripheral T cell tolerance and indicates that presentation of the MHC class II epitope H-Y to the immune system in the absence of the class I epitope P91A induces long term tolerance to grafted P91A+ H-Y+ skin.

H-Y specific CD4+ T cells appear to be rendered tolerant by an initial exposure to H- Y in the absence of P91 A. This suggests that H-Y specific CD4+ T cells are regulated in their response by the simultaneous recognition of the P91A epitope by CD8+ T cells. CD8-I- T cells have previously been implicated in the regulation of in vivo and in vitro immune responses and are capable of secreting IL2 and y- interferon (Inaba et al, 1987; Gill, 1993; Kemeny et al, 1994; Stuhler and Walden, 1993).

A further possibility which could explain non responsiveness to P91A H-Y after initial exposure to H-Y is the development of a post activation refractory period in H- Y responsive T cells. This seems unlikely since 14 days is the standard period

between P91A+ H-Y+ spleen cell immunisation and P91A+ H-Y+ skin grafting. A protocol which is known to induce graft rejection, table 2.1 C.

Non professional antigen presenting cells can tolerise naive T cells, (Fuchs and Matzinger, 1992). Mixed spleen cells, which contain only ~1% professional APCs, are used as the immunising population in the present experiment. It is therefore highly likely that the majority of H-Y responsive CD4-I- T cells will encounter antigen on non professional APCs and become tolerised. The presence of multiple non

professional APCs may tend to decrease the probability of H-Y reactive CD4-k T cells interacting with antigen on a professional APC. This in tum may have the same effect as decreasing the precursor frequency of H-Y reactive cells, since interactions with non professional APCs will anergise the responding cell. It is possible that only a very small percentage of H-Y specific CD4+ T cells become fully activated upon initial exposure to H-Y+ male spleen cells and that a situation of near mass tolerance is induced. Such an outcome may possibly be avoided in the presence of P91A specific CD8+ T cells. It is now known that CD8+ T cells may become activated in the absence of specific T cell help if there is CD28/B7 co-receptor engagement and may even be activated in the absence of CD28/B7 interaction if there is persistence of antigen (Harding et al, 1992; Kündig et al, 1996). It is possible that P91A specific

CD8+ T cells gather in conjunction with H-Y specific CD4+ T cells round non professional APCs and via the secretion of cytokines or induction of costimulatory molecules on the antigen presenting cell support the activation or prevent the tolerisation of the CD4+ cells.

2.2.2.2

demonstration of tolerance to the P91A tum" epitope in