Introduction: Regulatory T cells (Treg) are being considered for evaluation as a cellular therapy in clinical transplantation and most suggested protocols will examine polyclonally-expanded naturally occurring populations (nTreg). Whilst both mouse and human nTreg can regulate rejection in adoptive transfer mouse models, it is less clear whether they have the same impact in immunocompetent recipients, a question of considerable relevance to the clinical situation. Using an approach that enriches for alloantigen-experienced Treg in vitro, we have been able to compare directly the effectiveness of freshly isolated nTreg, expanded nTreg and alloantigen-driven adaptive Treg in immunocompetent mouse heart allograft recipients. Methods: Recipient (CBA H-2k) splenic CD4+CD25+ nTreg were isolated by MACS (purity >95% CD25+) and were used either after isolation or expanded with plate-bound anti-CD3 and anti-CD28 antibody plus IL-2 for 5 days. Antigen-experienced Treg (IFN-γ Treg) were enriched in vitro by stimulating naïve CBA CD4+ T cells with modulated donor-strain (C57BL/6 H-2b) antigen-presenting cells in the presence of interferon-γ. All Treg populations were examined phenotypically for expression of FoxP3 and evaluated for their ability to control rejection of fully MHC-mismatched heterotopic cardiac allografts. Treg were given on day -1 i.v. together with co-stimulation blockade deliberately designed to be sub-optimal (CTLA-4 Ig, 25μg day 0, 1, 5), in order to reveal any Treg-dependent impact on graft outcome. Results: Untreated CBA recipients rejected C57BL/6 hearts acutely (median survival time, MST, 8 days n=10). Sub-optimal CTLA-4 Ig treatment prolonged graft survival to an MST of 50 days, (n=8). When given in combination with sub-optimal CTLA-4 Ig, neither 2×106 freshly isolated nTreg (90-95% FoxP3+), nor 2×106 expanded-nTreg (90% FoxP3+) prolonged graft survival long-term and in fact, freshly isolated nTreg appeared to partially over-ride the effect of CTLA-4 Ig (MST=24 days n=8; MST=40 days n=5, fresh and expanded nTreg respectively). In clear contrast, delivery of 2×106 IFN-γ Treg led to a median survival time of 95 days (n=6), with 50% graft survival >100 days. The effect was dose-dependent (MST=66 days n=4; MST=100 days n=3, 0.5×106 and 4×106 IFN-γ Treg, respectively) and occurred despite the fact that this non-selected population was only between 40-50% FoxP3+. Furthermore, preliminary results indicate that the optimal time of IFN-γ Treg delivery is before rather than after transplant (MST 100 days, n=3 vs MST 18 days, n=4; delivery on day -3 vs. day +3 respectively) suggesting that early homing to the graft may be critical determinant in Treg efficacy. Conclusions: Although initial attempts to utilise Treg cell therapy in clinical transplantation will focus predominantly on expanded nTreg, our data indicate that efforts to develop alloantigen-driven adaptive Treg should not be overlooked. Whilst the generation of alloantigen-driven Treg will be challenging in terms of logistics, the data suggest that such cells may provide the most effective populations for cellular therapy and may offer a route to drug-minimisation without compromising transplant outcome.