-Studies in this laboratory have established that trophoblast cells have developed two novel mechanisms of silencing MHC class II genes: an upstream negative-regulatory element controlled by two trophoblast specific DNA binding proteins, and repression of transcription of the class II transactivator (CIITA). Recently, some human tumor cells have been shown to have a trophoblast phenotype, in that CIITA and class II are not inducible with g-IFN or other cytokines. In collaboration with A. Latif Kazim and Carleton Stewart, RPCI, we have shown that tumor cell lines that are not responsive to g-IFN despite an intact g-IFN receptor and Jak-Stat signaling pathway, can be induced to express mRNA and cell surface class II with agents that acetylate histones and remodel chromatin structure. Unexpectedly, we found that class II induction by inhibitors of histone deacetylase occurred in the apparent absence of transcription of CIITA. Recent studies with mutant cell lines defective in established class II regulatory genes suggest the existence of a pathway independent of CIITA, the presumed master regulator of MHC class II. During these experiments, we also showed that the expression of two other molecules of immunological importance MHC class I and CD40 also were activated by histone deace-tylase inhibitors. Dr. William Magner has joined my laboratory recently and, together with Dr. Elizabeth Repasky, we extended these studies to different mouse tumor cell lines, and demonstrated by flow analysis the induction of MHC class I and II genes as well as the costimulator molecule CD40. We are exploring the spectrum of antigens expressed on tumor cells treated with histone deacetylase inhibitors including TGF-b receptors and tumor associated antigens. We are determining whether treated tumor cells administered in vivo show enhanced immunity and decreased tumorigenicity to subsequent challenge with wild type cells. In this regard, as first shown by Ostrand-Rosenberg, et al., mouse sarcoma I cells, when transfected with MHC class II, produce solid and long-lasting immunity to inoculation with wild type cells. Thus, these findings have significance in tumor escape mechanisms and the development of vaccines.

We have explored the possibility that repressors of CIITA transcription exist in trophoblast and tumor cells, thus inhibiting class II expression and its induction by g-IFN. TGF-b1 and IL-10 are produced by all of the tumor cell lines we examined that were deficient in class II expression. Since these cytokines repress class II transcription by inhibiting CIITA transcription, the possibility of autocrine inhibition is suggested. The mechanisms of TGF-b repression may involve recruitment of deacetylase enzymes to the class II promoter site by CREB-CBP-CIITA complexes. This possibility will be explored using chromatin immunoprecipitation techniques. Studies have been initiated on the role of transcriptional cofactors with histone acetyltransferase activity, as well as the role of ATP-dependent multiprotein chromatin remodeling complexes in regulating MHC gene expression.

Our work also has demonstrated that TGF-b can bind to receptors on dendritic cells and inhibit antigen presentation. Moreover, serum TGF-b1 levels are elevated in AIDS and DC cells derived from the blood of HIV patients who have high levels of cell surface TGF-b1. Treatment with anti-TGF-b, or anti-type III TGF-b receptor antibodies, or the addition of GM-CSF reverses the defect in antigen presentation in AIDS-PBL in vitro. Prelimin-ary clinical studies in patients with advanced AIDS, in collaboration with the Department of Medicine, have shown that GM-CSF reverses the antigen presentation defect and significantly raises CD4 cell numbers in many, but not all, AIDS patients.

To test the hypothesis frequently quoted in the literature that aberrant expression of class II antigens on fetal trophoblast cells is responsible for most spontaneous abortions, transgenic animals are being prepared that will express class II on placental trophoblast cells at a specific time during gestation. We have cloned 300 bp of the proximal promoter of a trophoblast specific gene, 4311, and placed this in an expression plasmid upstream of the full-length CIITA gene. This construct was shown to be active when transfected into freshly isolated mouse spongiotrophoblast cells. Thus, mice bearing this transgene should begin to transcribe CIITA at day 6 of gestation when the 4311 gene is first expressed. If the hypothesis is correct, then MHC class II should be expressed by day 7-8, and abortion would occur. A dominant negative of CIITA has been produced by deletion of part of the activation domain of CIITA, and this could potentially be expressed in trophoblast cells of strains with high abortion rates. If as postulated by others, fetal loss in these abortogenic strains is due to aberrant expression of CIITA and subsequently class II, the dominant negative should inhibit abortions in these mice.