Antigen exposure during enhanced CTLA-4 expression promotes allograft tolerance in vivo.

The role of CTLA-4 in tolerance is primarily inferred from knockout and blocking studies. Anti-CD45RB mediates allograft tolerance in mice by inducing CTLA-4 expression on CD4 cells, providing a novel opportunity to determine how therapeutic enhancement of CTLA-4 promotes tolerance. We now show that induced CTLA-4 expression normally resolves by day 17. Although thymectomy prolongs enhanced CTLA-4 expression, long-term engraftment is unaffected. To address the temporal relationship between increased CTLA-4 expression and engraftment, transplantation was delayed for various times after anti-CD45RB treatment. Delaying transplantation for 7 days (when CTLA-4 expression had peaked but treatment mAb was no longer detectable), resulted in long-term engraftment comparable to transplantation with no delay (day 0). Delaying transplantation from 10 to 18 days led to a progressively poorer outcome as CTLA-4 expression returned to baseline. This suggested that Ag exposure while CTLA-4 expression is enhanced is sufficient to induce long-term engraftment. To substantiate this, on day 0, anti-CD45RB-treated mice received BALB/c vs unrelated alloantigen, followed by transplantation of BALB/c islets 10 days later. Whereas recipients exposed to unrelated Ag experienced acute rejection, recipients exposed to donor Ag achieved long-term engraftment. Anti-CD45RB-treated mice exposed to alloantigen exhibited anergic CD4(+)CD25(-) effector cells and regulatory CD4(+)CD25(+) cells. Moreover, CD25 depletion in the peritransplant period prevented anti-CD45RB-mediated engraftment. Thus, exposure of CD4 cells expressing CTLA-4 to donor Ag is necessary and sufficient to induce long-term engraftment which appears to be mediated by both regulation and anergy.

Cutting edge: transplantation tolerance through enhanced CTLA-4 expression.

Knockout and blocking studies have shown a critical role for CTLA-4 in peripheral tolerance, however, it is unknown whether augmenting CTLA-4 expression actually promotes tolerance. Here we demonstrate a specific and requisite role for CTLA-4 and its up-regulation in tolerance through anti-CD45RB. First, long-term murine islet allograft survival induced by anti-CD45RB is prevented by CTLA4-Ig, which interferes with B7:CTLA-4 interactions. Second, anti-CD45RB is ineffective in recipients lacking CTLA-4, B7-1, and B7-2. In contrast, CTLA4-Ig, which targets B7 on allogeneic cells, promotes long-term engraftment in these mice. Moreover, anti-CD45RB was effective in B7-deficient controls expressing CTLA-4. Finally, in wild-type mice, CTLA-4 expression returned to baseline 17 days after receiving anti-CD45RB, and was refractory to further increase. Transplantation and anti-CD45RB therapy at this time could neither augment CTLA-4 nor prolong engraftment. These data demonstrate a specific role for CTLA-4 in anti-CD45RB-mediated tolerance and indicate that CTLA-4 up-regulation can directly promote allograft survival.

The complementarity-determining region-like loops of CD8 alpha interact differently with beta 2-microglobulin of the class I molecules H-2Kb and thymic leukemia antigen, while similarly with their alpha 3 domains.

The murine CD8 glycoprotein interacts with both classical MHC class I molecules and some nonclassical molecules, including the thymic leukemia Ag (TL). TL binds preferentially to CD8alphaalpha homodimers with a 10-fold higher affinity than H-2K(b) class I molecules. To understand the molecular basis for this difference, we created a panel of CD8alpha mutants and tested the ability of the CD8alphaalpha homodimers to bind to H-2K(b) tetramers and TL tetramers. Mutations in three CD8 residues located on the complementarity-determining region-like loops contacting the negatively charged loop in the alpha3 domain of MHC class I greatly reduced binding to both tetramers. Because TL and H-2K(b) class I sequences are highly conserved in the alpha3 domain of MHC class I, this suggests that CD8 contacts the alpha3 domain of TL and H-2K(b) in a similar manner. In contrast, mutations in residues on the A and B beta strands of CD8 that are involved in contact with beta(2)-microglobulin affected interaction with the H-2K(b) tetramer, but not the TL tetramer. Therefore, the orientation of interaction of TL with CD8 appears to be different from that of H-2K(b). The unique high affinity binding of TL with CD8alphaalpha is most likely a result of amino acid differences in the alpha3 domain between TL and H-2K(b), particularly at positions 198 (K to D) and 228 (M to T), which are contact residues in the CD8alphaalpha-H-2K(b) cocrystal.