P653 Introduction: Composite tissue allotransplantation (CTA) is an emerging field requiring an amalgamation of microsurgical technique and immunosuppressive management. Although the technical aspects of CTA have followed logically from replantation, the discipline remains immunologically undefined. Specifically, it has not been established whether CTA rejection follows a predictable pattern through the many transplanted tissues, nor has it been determined whether some tissues are immunodominant or perhaps spared immune attack. Furthermore, of the immunosuppressants available for use in CTA, many have human specificity and cannot be evaluated in small animals. Based on an evaluation of specimens from transplanted human limbs, we have recently introduced a classification system to standardize CTA rejection. We have shown that vasculitis, dermatitis, myositis and perineural involvement can be seen during rejection, but given the sporadic availability of human tissue, the progression of the immune response remains undefined. Aims: To permit systematic study of the progression and resolution of CTA rejection, we have established a non-human primate (NHP) model. NHPs provide anatomical, antigenic and immunological similarity to humans, and can be used to test most human-specific biological agents. Limb transplantation in NHPs could markedly impair their social development. We have therefore designed our model to incorporate all elements of a limb CTA yet leave no functional deficit even in the case of rejection or excision. Methods: Our model involves a transplanted radial forearm flap including bone, muscle, nerve, tendon, artery, vein and skin. To validate this model, 5 animals have undergone autotransplantation and 11 have received allografts with and without immunosuppression (subtherapeutic tacrolimus, mycophenolate and prednisone to allow for rejection to be temporally characterized). Donor-recipient pairs for CTA were selected based on MHC non-identity, blood type and size. Results: The procedure was well tolerated by all animals and had no adverse functional consequences. The model allows for protocol biopsies to be performed without threatening the viability of the flap. Autografts showed no evidence of rejection and served as the baseline for CTA comparison. Non-immunosuppressed allografts showed an infiltrate initially arising within 3 days in the perivenular tissue leading to graft congestion and failure without a prominent dermal infiltrate. Subtherapeutically immunosuppressed animals rejected in 2 weeks with a marked dermal lymphocytic infiltrate similar in magnitude and distribution to human cases seen to date. Conclusions: We have established a NHP model for CTA that is responsive to immunosuppression and allows for the systematic evaluation of CTA histology and therapeutic maneuvers to improve allograft survival.