Organ transplantation is currently dependent on life-time administration of immunosuppressive drugs to non-specifically suppress immune response, resulting in severe side effects, including infection and cancer. Induction of specific tolerance is an ideal approach, but not generally successful, which however, actually naturally occurs. Thus, ˜1/4 liver allografts in farm pigs are spontaneously accepted, and induce donor specific tolerance. In humans, immunosuppression can be weaned off until completely stopped in ˜1/4 liver transplant patients. The underlying mechanisms are not completely understood. We and others surprisingly found that liver transplant tolerance is dependent on IFN-g stimulation. Mouse liver allograft, otherwise spontaneously accepted, was acutely rejected in IFN-g-/- recipients. In this study, the liver from IFN-gR1-/- mice (B6, H-2b) was transplanted into C3H (H-2k) recipients (n=6). None of them survived >15 days (WT grafts survived >100 days), suggesting that liver transplant tolerance requires intact IFN-g signaling pathway in liver graft. Both flow and histochemical analyses demonstrated markedly increased CD8+ infiltrates in IFN-gR1-/- grafts associated with lower expression of B7-H1 and less infiltrating T cell apoptosis, suggesting a role of B7-H1, an end product of IFN-g signaling in regulating immune response. This was supported by the result that the liver grafts from B7-H1-/- mice were acutely rejected in WT allogeneic recipients, associated with markedly less apoptosis (annexin V) in the graft infiltrating T cells. Liver grafts deficient in IFN-gR1 did not affect much T cell activation (CD44hiCD62Llo), as well as generation of myeloid-derived suppressor cells (MDSC) and Tregs in early phase. However, Tregs were markedly enhanced in long-term survived WT grafts, suggesting that elimination of effectors in early phase is crucial, while the suppressor cells may largely work to maintain tolerance. To identify the cell components in the liver graft responsible for these changes, the graft non-parenchymal cells were isolated at various time points for flow analysis. Almost all graft CD45+ cells were quickly (in one-two days) become recipient origin (H-2k, IFN-gR1+), while CD45- cells remain donor type (H-2b, IFN-gR1-), even in long-term survived grafts, suggesting that liver non-hematopoietic cells may be responsible for these changes. We have studied isolated hepatic stellate cells (HSC) (liver stromal cells, well known for their storing vitamin A and participating in fibrosis) that were exceptionally immunosuppressive. Cotransplantation with HSC effectively protected islet allografts without requirement of immunosuppression. Interestingly, cotransplantation with HSC isolated from IFN-gR1- mice failed to protect islet grafts, associated with markedly more infiltrating CD8+ T cells and less MDSC and Treg cells, indicating that the immune regulatory activity of HSC depends on intact IFN-g signaling pathway. Expression of B7-H1 was upregulated following exposure to IFN-g in a dose dependent manner, suggesting B7-H1 is an end product of IFN-g signaling. HSC isolated from B7-H1-/- mice largely lost their ability to protect islet allgrafts. This was associated with more CD8 T cells graft infiltration. In conclusion, IFN-g/B7-H1 signaling pathway in liver stromal cells play a critical role in establishment of liver transplant tolerance.