Epigenetic Regulation of Normal and Pathologic CTCF Functions by BORIS

点击翻译按钮获取中文摘要

During the last year, we continued our studies of BORIS (Brother Of the Regulator of Imprinted Sites) - a CTCF-paralogous gene we discovered and published in 2002. Both CTCF and BORIS proteins share a nearly identical 11 Zn-finger (11ZF) DNA binding domain (DBD) that is flanked with dissimilar N- and C-terminal regions. In the mid 90s, this remarkable 11ZF region was defined as a multivalent DBD with respect to the diversity of unusually extended (average 50bp) target sequences recognized in the double-stranded DNA. This DBD region shared by BORIS and CTCF proteins is encoded by sequences displaying an accurate duplication of a conserved genomic region containing all ZF-coding exons of CTCF gene. One of the non-pathological functions that we first proposed for BORIS was based on the apparent need for a different from CTCF factor that could nevertheless recognize CTCF sites during re-establishment in male germ cells of paternal-specific DNA methylation patterns at particular CTCF sites initially mapped and characterized by us and others in the imprinting control region (ICR) of the Igf2/H19 locus. Recent study of BORIS evolution and expression in amniotes confirmed our finding that a testicular-type of BORIS gene expression was absent in birds or in any other lower vertebrates that lack parent-of-origin specific gene imprinting, and uncovered that BORIS expression became gonad-specific in wallaby and cattle, both of which exhibit imprinting of IGF2/H19 genes. This implies that evolution of the testis-restricted BORIS expression in the germline cells does correlate with the evolution of genomic imprinting at IGF2/H19 and other loci exactly as we had first suggested in the original paper describing BORIS. While sharing with CTCF a duplicate of the 11ZF DBD allows BORIS to bind specifically to the same DNA-target sequences that interact with CTCF, in vivo binding of BORIS to a given CTCF site would result in a different functional outcome than binding of CTCF to the same site. Normally, CTCF and BORIS are not expressed in the same cell to avoid functional interference caused by competition for binding to the same spectrum of target sequences they can both recognize. Therefore, we predicted from previous studies of CTCF and imprinting that in spite of the lack of any obvious sequence homologies in the imprinting control regions (ICR) from different imprinted loci, most, if not all ICR sequences would contain meCpG-sensitive CTCF target sites. This prediction turned out to be correct and resulted in several publications describing novel CTCF-driven ICRs in unrelated imprinted loci. BORIS is normally strictly silenced in somatic cells, but activated in cancers together with a particular family of genes, called cancer-testis (CT) genes. To understand the mechanisms governing BORIS expression, we characterized the 5-flanking region of the human gene, and identified three promoters designated A, B, and C, that correspond to 3 transcription start sites. Alternative promoter usage was found to be associated with the generation of at least five alternatively spliced BORIS mRNAs having different half lives determined by varying 5 UTR sequences. While in normal testis BORIS is transcribed from all three promoters, in the majority of cancer cell types ( 80 % of 30 cell lines analyzed) only promoter A and/or promoter C were activated, and in the remaining cell lines, promoters B and C were most active. While DNA methylation contributes to the negative regulation of each promoter, demethylated BORIS promoters are activated by suppression of CTCF expression or by downregulation of p53. Moreover, blocking CTCF in normal cells resulted in demethylation and derepression of BORIS promoters. These results provide a mechanistic basis for understanding two cancer-associated functional connections between decrease or loss of CTCF and loss of functional p53 and relaxation of BORIS promoters silencing. Next, we showed that BORIS appears to be aberrantly expressed with a much higher incidence than other CT genes in many primary tumors of different tissue origins. Two recently published studies of abnormal BORIS gene/protein expresssion in female malignancies were performed in collaboration with Dr. Risinger (NCI) on BORIS in uterine cancers and with Dr. Klenova of the Essex University on BORIS in breast tumors. In the first study we queried the expression of known and putative CT gene transcripts using whole genome gene expression arrays, and examined BORIS transcripts by reverse transcription PCR and quantitative PCR since the transcript was not represented on the array. Microarray analysis detected many CT genes expressed in uterine cancers; however, no individual CT gene was expressed in more than 25% of any cancer. In stark contrast, BORIS was expressed in 73 of 95 (77%) endometrial cancers and 24 of 31 (77%) uterine mixed mesodermal tumors. The high frequency of BORIS expression in uterine cancers suggest again of a great potential of utilizing BORIS as an immunologic or diagnostic target for these malignancies. In the second study, we assessed the relation between BORIS expression levels in breast cancer with clinical/pathological parameters. BORIS was detected in all of 18 breast cell lines tested, but not in a single primary normal breast cell culture. BORIS was expressed in 41 of 58 cases (71%) of primary breast tumors. We also showed that high BORIS levels in breast cancer correlates with high levels of progesterone and estrogen receptor. We showed previously that BORIS, when abnormally expressed in tumors, can induce antibody responses detectable in serum samples from cancer patients (patent pending). We further confirmed this in collaboration with Dr. L.L Hansen (University of Aarhus, Denmark) in a blind study of breast cancer patients and healthy donors, and we are now working to push forward this diagnostic approach to reach its appropriate application in clinical practice through a CRADA agreement. Finally, we have further pursued anti-cancer vaccine development using a mouse models in collaboration with Dr. M. Agadjanyan (Institute for Molecular Medicine, Huntington Beach, CA). Previously, as proof of principle, we showed that BORIS protein is expressed in various mouse cancers and that it is possible to generate immune responses to BORIS using both DNA vaccines and recombinant protein. In a recent study, we compared vaccine strategies employing BORIS recombinant protein formulated in a strong Th1-type adjuvant, QuilA, or DNA encoding this same immunogen in the form of DNA along with plasmids expressing IL12 and IL18 as molecular adjuvants. Both groups of vaccinated mice were examined for induction of tumor-specific immunity (antibody response, T-cell proliferation, cytokine production, T-cell cytotoxicity) as well as the ability to inhibit growth of an aggressive breast cancer cell line and to prolong survival of vaccinated animals. We determined that DNA, but not recombinant protein vaccine, induced potent Th1-like T-cell responses that significantly inhibited tumor growth and prolonged the survival of vaccinated mice. These studies provide a clear guidance for clinical development of a cancer vaccine targeting what appears to be a widely expressed tumor antigen.