Quantitative analysis predicts the relative therapeutic efficacy of different forms of CTLA4Ig.

Modulating the activities of costimulatory molecules controlling immune responses holds considerable promise for immunotherapy. CTLA4Ig (abatacept), a soluble version of the T cell-expressed membrane receptor CTLA-4, is approved for the treatment of rheumatoid arthritis. Like natural CTLA-4 molecules, CTLA4Ig ligates B7-1 and B7-2 on antigen presenting cells, preventing CD28-mediated costimulation of T cells. However, CTLA4Ig can also prevent ligation of CTLA-4, potentially blocking vital inhibitory signals, thereby augmenting immunity. There have been no quantitative analyses of the likely effects of CTLA4Ig on costimulatory interactions at the immunological synapse. We present a mathematical model, based on rigorous biophysical and expression data, for simulating the effects of abatacept and a mutated derivative, LEA29Y, on the synaptic interactions of CD28 and CTLA-4. The simulations reveal an unexpectedly large window within which CD28, but not CTLA-4, ligation is blocked by CTLA4Ig, perhaps explaining the efficacy of abatacept at the recommended therapeutic dose (10mg/kg) and its relative safety. However, the simulations suggest that the present dosing regimen is close to the maximum theoretically safe dose. The simulations also show that, within the therapeutic window, LEA29Y enhances the interaction of CTLA-4 with the more potent of its two native ligands, B7-1. They also suggest that CTLA-4 ligation by B7-1 could, in principle, be enhanced by further decreasing the off-rate of CTLA4Ig for binding to B7-2. Our findings therefore offer molecular explanations for why LEA29Y might prove to be more effective than abatacept in a clinical setting, and suggest ways in which its therapeutic efficacy could be further optimised.