A Critical Role for STAT-3 Dependent Induction of Indoleamine 2, 3 Dioxygenase in Histone Deacetylase Inhibitors Mediated the Regulation of GVHD.

Histone deacetylase (HDAC) inhibitors reduce experimental acute graft-versus-host disease (GVHD) and recent work by us and others suggest that HDAC inhibitors regulate dendritic cell (DC) function. However, the critical cellular and molecular mechanisms underpinning these observations are not known. We investigated the mechanisms by utilizing two HDAC inhibitors, suberoylanilide hydroxamic acid (SAHA) and ITF 2357. Pretreatment of murine bone marrow (BM) and human peripheral blood mononuclear cell derived DCs with either HDAC inhibtors and stimulated with TLR ligands such as LPS caused a significant reduction in the secretion of TNF-α compared to the untreated controls (P< 0.01). Pre-treatment also significantly reduced the DC mediated in vitro and in vivo stimulation of allogeneic T cell proliferation (P<0.05). SAHA and ITF 2357 increased expression of indoleamine 2, 3-dioxygenase (IDO) at both mRNA and protein levels. Blockade of IDO induction with specific small interfering RNA (siRNA) in the wild type (WT) DCs and those derived from IDO deficient (IDO−/−) animals confirmed a functional role for IDO in the HDAC inhibitor mediated regulation TNFα secretion and allo-T cell proliferation. DNA-protein interaction analysis with ChIP assay demonstrated that both acetylated histone(H) 4 and STAT3 bound to murine IDO promoter. Using TESS DNA soft-wear analysis we found two potential STAT3 binding Gamma Activated Sites (GAS sites) in the IDO promoter and it was recently reported that acetylation of STAT3 is sufficient for its activation (Yuan, et al. Science 2005). We therefore sought to determine whether direct acetylation of STAT3 by the HDAC inhibitors is critical for the induction of IDO in DCs. SAHA or ITF2357 treatment induced acetylation, activation and dimerization of STAT-3 as determined by protein-protein interaction studies. Co-culture studies with JSI-124, an inhibitor of STAT3 signaling, demonstrated that STAT3 is critical for induction of IDO by the HDAC inhibitors. Functional relevance was confirmed by the lack of HDAC inhibitor induced suppression of DC function when co-treated with JSI-124. We next cloned 1500bp DNA fragment upstream of mouse IDO gene start codon and attached it to luciferase gene and peformed mutagenesis studies to evaluate for luciferase activity. Deletion of the GAS regions or treatment with JSI-124 impaired luciferase activity of the IDO promoter constructs demonstrating that STAT3 is necessary and sufficient for transcriptional induction of IDO by the HDAC inhibitors. To specifically address the in vivo relevance of IDO induction by HDAC inhibitors in only the host type DCs, we generated [B6 → B6] and [IDO−/−B6 → B6] BM chimeras and utilized them in a well characterized [BALB/c B6] mouse model of acute GVHD. [B6 → B6] and [IDO−/− B6 → B6] animals received 800 Gy on day −1 and were treated orally with of 50mg/kg of ITF 2357 or diluent on days −1 to +2. Mice were transplanted on day 0 with 3 x 10⁶ T cells and 5 x 10⁶ BM from either syngeneic B6 or allogeneic BALB/c donors. Treatment with ITF 2357 resulted in significantly better survival in the allogeneic [B6 → B6] animals (80% vs. 40%, P < 0.02) but did not confer any survival benefit to the [IDO−/− B6 → B6] animals when compared with diluent treated recipients [20% vs. 30%, P = NS]. Our data thus demonstrate a novel molecular pathway in modulation of GVHD through a STAT3 dependent induction of IDO in the host DCs.