Costimulation Blockade—A Double-Edged Sword?

Costimulation blockade with CTLA-4Ig is an exciting and recently FDA approved immunomodulatory strategy in treatment of autoimmune disease and kidney transplantation (1). Initially described over 20 years ago, abatacept and the modified form, belatacept are fusions of the extracellular portion of CTLA-4 with an immunoglobulin Fc region (2). The overall clinical effect of CTLA-4Ig administration is to dampen effector T-cell mediated responses. A recent report in American Journal of Transplantation and another in this issue (3,4) add to a growing body of literature complicating this view by showing that CTLA-4Ig can accelerate rejection thus reminding us that costimulation blockade can have opposing functional effects on T-effector versus T-regulatory cells.

Under most circumstances, naive T cells require two signals for full activation; signal one is antigen-specific, occurring when the T cell receptor recognizes proteins of the major histocompatibility complex with peptide and signal two has been described as costimulatory. Costimulation is necessary for proliferation, differentiation, survival and cytokine production. The prototypical costimulatory pathway is mediated by CD28 on the T cell and its ligands CD80/CD86 (B7.1/B7.2) on the antigen-presenting cell (APC). CTLA-4 is a related receptor expressed on activated T cells that also recognizes CD80/CD86, but transmits both cell intrinsic and cell extrinsic negative signals that impair activation and is thus termed coinhibitory. CTLA-4Ig blocks both B7-CD28 and B7-CTLA-4 signaling.

The expression pattern of CD28 differs from CTLA-4. In mice, CD28 is constitutively expressed on all T cells. Conversely, activated but not naive murine T cells express CTLA-4. Tregs constitutively express both CD28 and CTLA-4. CD28 is required for Treg generation and maintenance; CTLA-4 is required for Treg function. Human Tregs and effector T cells have similar expression patterns with the exception of some subsets of human T cells (activated CD8+), which lack CD28 expression, possibly accounting for increased acute rejection seen in clinical trials of belatacept.

Given the wealth of data in other murine models of transplant, it is somewhat surprising to find that administration of CTLA-4Ig accelerates rejection in a MHC Class II mismatch murine model in which graft acceptance is dependent on Tregs. Riella et al. showed that hCTLA-4Ig (abatacept) treatment accelerated rejection of bm12 into B6 heterotopic cardiac allografts, but improved outcomes of Balb/c to B6 allografts (3). Charbonnier et al. used treatment of bm12 into B6 skin transplant recipients with interleukin-2/interleukin-2 receptor (IL2/IL2R) conjugates to increase Treg numbers with resulting prolongation of graft survival (4). This treatment regimen has previously been used in autoimmune models but Charbonnier's report represents the first successful report in a solid organ transplant model. Similar to Riella's studies, coadministration of hCTLA-4Ig significantly attenuated the IL2/IL2R effect, decreasing both Treg numbers and graft survival.

Both reports demonstrate direct effects of hCTLA-4Ig on Tregs. Consistent with earlier reports from several groups, these studies showed that CTLA-4Ig decreased steady-state Treg turnover and numbers. Interestingly, hCTLA-4Ig had an effect on in vitro suppressive activity from Tregs generated by IL2/IL2R administration, but not on those from naïve mice. Moreover, elevated levels of Foxp3, TGFβ and inducible costimulator (ICOS) on IL2/IL2R induced Tregs were antagonized by CTLA-4Ig treatment. Both studies used the abatacept formulation in mice. As belatacept does not efficiently bind murine B7.1/B7.2, extension of these mechanistic findings to the formulation approved for kidney transplant will require nonrodent models.

Should these findings have any impact on the use of belatacept in human kidney transplantation? The short answer is no. Used in appropriate patients and in combination with basiliximab, mycophenylate and steroids, there is proven efficacy as an immunosuppressive agent (1). Studies to determine optimal belatacept combination therapy for immunosuppression or tolerance are ongoing. Treg numbers, but not function, were decreased in patients receiving the combination of basiliximab induction and belatacept, but this effect may be attributed to IL-2R blockade as it was no different than basiliximab in combination with cyclosporine (5). Conversely, the combination of rabbit antithymoglobulin, sirolimus and belatacept increased both Treg numbers and suppressive activity (6). Adoptive Treg therapy for the induction of tolerance in solid organ transplant is nearing clinical trials, which may involve short-term immunosuppression. The findings of Riella and Charbonnier suggest caution may be warranted in using CTLA-4Ig in new drug combinations for immunosuppression and possibly tolerance, especially those strategies dependent on Tregs.

Disclosure

The author of this manuscript has a conflict of interest to disclose as described by the American Journal of Transplantation. J.S.M. has family members with a financial interest in GlaxoSmithKline and Morphotek.