Background: Global DNA hypomethylation may result in chromosomal instability and oncogene activation, and as a surrogate of systemic methylation activity, may be associated with breast cancer risk. Methods: Samples and data were obtained from women with incident early-stage breast cancer (I–IIIa) and women who were cancer free, frequency matched on age and race. In preliminary analyses, genomic methylation of leukocyte DNA was determined by measuring 5-methyldeoxycytosine (5-mdC), as well as methylation analysis of the LINE-1-repetitive DNA element. Further analyses used only 5-mdC levels. Logistic regression models were used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for risk of breast cancer in relation to amounts of methylation. Results: In a subset of samples tested (n = 37), 5-mdC level was not correlated with LINE-1 methylation. 5-mdC level in leukocyte DNA was significantly lower in breast cancer cases than healthy controls (P = 0.001), but no significant case–control differences were observed with LINE-1 methylation (P = 0.176). In the entire data set, we noted significant differences in 5-mdC levels in leukocytes between cases (n = 176) and controls (n = 173); P value < 0.001. Compared with women in the highest 5-mdC tertile (T3), women in the second (T2; OR = 1.49, 95% CI = 0.84–2.65) and lowest tertile (T1; OR = 2.86, 95% CI = 1.65–4.94) had higher risk of breast cancer (P for trend ≤0.001). Among controls only and cases and controls combined, only alcohol intake was found to be inversely associated with methylation levels. Conclusion: These findings suggest that leukocyte DNA hypomethylation is independently associated with development of breast cancer.

We examined the function of two key DNA methyltransferase (DNMT) enzymes in epigenetic regulation of Xlinked cancer/germline (CG-X) antigen genes in human cancer cells, using MAGE-A1, NY-ESO-1, and XAGE-1 as models. In HCT116 cells, genetic knockout of DNMT1 caused moderate activation of CG-X genes, DNMT3b knockout had a negligible effect, and double knockout of both enzymes caused robust gene induction. Similarly, dual DNMT knockout caused dramatic hypomethylation of the MAGE-A1 and NY-ESO-1 promoters, DNMT1 knockout showed moderate hypomethylation, and DNMT3b knockout elicited only slight methylation changes. In contrast, both single and double knockout cells showed significant hypomethylation of the XAGE-1 promoter. RNA interference (RNAi) targeting of DNMT1 in HCT116 cells validated the results seen using genetic knockout cells; however, RNAi targeting of DNMT1 in a different colorectal cancer cell line revealed a greater independent role for DNMT1 in mediating CG-X gene repression and promoter methylation in other cell types. Notably, the histone H3 modification pattern at CG-X promoters was altered following DNMT knockout. DNMT1 or DNMT3b knockout reduced dimethylated lysine-9 (diMe-H3K9) levels, but did not significantly affect dimethylated lysine-4 (diMe-H3K4) or acetylated lysine-9 (Ac-H3-K9) levels. In contrast, dual DNMT1/3b knockout reduced the level of diMe-H3K9 and dramatically increased the levels of diMe-H3K4 and Ac-H3K9 at CG-X gene loci. In summary, DNMT1 and DNMT3b were found to perform both redundant and independent functions in epigenetic regulation of CG-X antigen genes in human cancer cells.

Purpose: Cancer germline (CG) antigens are frequently expressed and hypomethylated in epithelial ovarian cancer (EOC), but the relationship of this phenomenon to global DNA hypomethylation is unknown. In addition, the potential mechanisms leading to DNA hypomethylation, and its clinicopathologic significance in EOC, have not been determined.Experimental Design: We used quantitative mRNA expression and DNA methylation analyses to determine the relationship between expression and methylation of X-linked (MAGE-A1, NY-ESO-1, XAGE-1) and autosomal (BORIS, SOHLH2) CG genes, global DNA methylation (5mdC levels, LINE-1, Alu, and Sat-a methylation), and clinicopathology, using 75 EOC samples. In addition, we examined the association between these parameters and a number of mechanisms proposed to contribute to DNA hypomethylation in cancer.Results: CG genes were coordinately expressed in EOC and this was associated with promoter DNA hypomethylation. Hypomethylation of CG promoters was highly correlated and strongly associated with LINE-1 and Alu methylation, moderately with 5mdC levels, and rarely with Sat-a methylation. BORIS and LINE-1 hypomethylation, and BORIS expression, were associated with advanced stage. GADD45A expression, MTHFR genotype, DNMT3B isoform expression, and BORIS mRNA expression did not associate with methylation parameters. In contrast, the BORIS/CTCF expression ratio was associated with DNA hypomethylation, and furthermore correlated with advanced stage and decreased survival.Conclusions: DNA hypomethylation coordinately affects CG antigen gene promoters and specific repetitive DNA elements in EOC, and correlates with advanced stage disease. The BORIS/CTCF mRNA expression ratio is closely associated with DNA hypomethylation and confers poor prognosis in EOC. Clin Cancer Res; 17(8); 2170-80. Ó2011 AACR.

PRAME is a cancer-testis antigen (CTA) and potential immuno-therapeutic target, but has not been well-studied in epithelial ovarian cancer (EOC) or its high grade serous (HGSC) subtype. Compared to normal ovary, PRAME expression was significantly increased most EOC, regardless of stage and grade. Interestingly, PRAME mRNA expression was associated with improved survival in the HGSC subtype. The PRAME locus was a frequent target for copy number alterations (CNA) in HGSC but most changes were heterozygous losses, indicating that elevated PRAME expression is not typically due to CNA. In contrast, PRAME promoter DNA hypomethylation was very common in EOC and HGSC and correlated with increased PRAME expression. PRAME expression and promoter hypomethylation both correlated with LINE-1 hypomethylation, a biomarker of global DNA hypomethylation. Pharmacologic or genetic disruption of DNA methyltransferase (DNMT) enzymes activated PRAME expression in EOC cells. Immunohistochemistry (IHC) of PRAME in EOC revealed frequent, but low level, protein expression, and expression was confined to epithelial cells and localized to the cytoplasm. Cytoplasmic PRAME expression was positively associated with PRAME mRNA expression and negatively associated with promoter methylation, but the latter correlation was not statistically significant. PRAME protein expression did not correlate with EOC clinicopathology or survival. In summary, PRAME is frequently expressed in EOC at the mRNA and protein levels, and DNA methylation is a key mechanism regulating its expression. These data support PRAME as an immunotherapy target in EOC, and suggest treatment with DNMT inhibitors as a means to augment PRAME immunotherapy.

The H3K9me2 histone methyltransferases G9a and GLP repress Mage-a class cancer germ-line (CG) antigen gene expression in murine embryonic stem (ES) cells, but the role of these enzymes in CG antigen gene regulation in human cancer cells is unknown. Here we show that whereas independent or dual knockdown of G9a and GLP in human cancer cells leads to reduced global and CG antigen promoter-associated H3K9me2 levels, it does not activate CG antigen gene expression. Moreover, CG antigen gene repression is maintained following pharmacologic targeting of G9a or treatment of G9a knockdown cells with the histone deacetylase inhibitor trichostatin A. However, G9a knockdown cells display increased sensitivity to CG antigen gene activation mediated by the DNA methyltransferase inhibitor decitabine. To account for these findings, we examined DNA methylation at CG antigen gene promoters in both cell types. We found robust DNA hypomethylation in G9a/GLP targeted murine ES cells but a lack of DNA methylation changes in G9a/GLP targeted human cancer cells; intriguingly, this distinction also extended to markers of global DNA methylation. These data reveal that G9a/GLP is required for DNA methylation of CG antigen genes and genomic DNA in murine ES cells, but not human cancer cells, and implicate DNA methylation status as the key epigenetic mechanism involved in CG antigen gene repression. (Mol Cancer Res 2009;7(6):851-62)

Epithelial ovarian cancer (EOC) is a highly lethal malignancy due to a lack of early detection approaches coupled with poor outcomes for patients with clinically advanced disease. Cancer-testis (CT) or cancer-germline genes encode antigens known to generate spontaneous anti-tumor immunity in cancer patients. CT45 genes are a recently discovered 6-member family of X-linked CT genes with oncogenic function. Here, we determined CT45 expression in EOC and fully defined its epigenetic regulation by DNA methylation. CT45 was silent and hypermethylated in normal control tissues, but a large subset of EOC samples showed increased CT45 expression in conjunction with promoter DNA hypomethylation. In contrast, copy number status did not correlate with CT45 expression in the TCGA database for EOC. CT45 promoter methylation inversely correlated with both CT45 mRNA and protein expression, the latter determined using IHC staining of an EOC TMA. CT45 expression was increased and CT45 promoter methylation was decreased in late-stage and high-grade EOC, and both measures were associated with poor survival. CT45 hypomethylation was directly associated with LINE-1 hypomethylation, and CT45 was frequently co-expressed with other CT antigen genes in EOC. Decitabine treatment induced CT45 mRNA and protein expression in EOC cells, and promoter transgene analyses indicated that DNA methylation directly represses CT45 promoter activity. These data verify CT45 expression and promoter hypomethylation as possible prognostic biomarkers, and suggest CT45 as an immunological or therapeutic target in EOC. Treatment with decitabine or other epigenetic modulators could provide a means for more effective immunological targeting of CT45.

While the therapeutic activity of the deoxycytidine analogue decitabine is thought to reflect its ability to reactivate methylation-silenced genes, this agent is also known to trigger p53-dependent DNA damage responses. Here, we report that p53-inducible ribonucleotide reductase (p53R2/RRM2B) is a robust transcriptional target of decitabine. In cancer cells, decitabine treatment induces p53R2 mRNA expression, protein expression, and promoter activity in a p53-dependent manner. The mechanism of p53R2 gene induction by decitabine does not seem to be promoter DNA hypomethylation, as the p53R2 5 ¶ CpG island is hypomethylated before treatment. Small interfering RNA (siRNA) targeting of DNA methyltransferase 1 (DNMT1) in wild-type p53 cells leads to genomic DNA hypomethylation but does not induce p53R2, suggesting that DNMT/DNA adduct formation is the molecular trigger for p53R2 induction. Consistent with this idea, only nucleoside-based DNMT inhibitors that form covalent DNA adducts induce p53R2 expression. siRNA targeting of p53R2 reduces the extent of cell cycle arrest following decitabine treatment, supporting a functional role for p53R2 in decitabine-mediated cellular responses. To determine the clinical relevance of p53R2 induction, we measured p53R2 expression in bone marrow samples from 15 myelodysplastic syndrome/ acute myelogenous leukemia (MDS/AML) patients undergoing decitabine therapy. p53R2 mRNA and protein were induced in 7 of 13 (54%) and 6 of 9 (67%) patients analyzed, respectively, despite a lack of methylation changes in the p53R2 promoter. Most notably, there was a significant association (P = 0.0047) between p53R2 mRNA induction and clinical response in MDS/AML. These data establish p53R2 as a novel hypomethylation-independent decitabine gene target associated with clinical response. [Cancer Res 2008;68(22):9358-66]

• Dnmt3b acts as a haploinsufficient tumor suppressor in Myc-induced lymphomas.The drivers of abnormal DNA methylation in human cancers include widespread aberrant splicing of the DNMT3B gene, producing abnormal transcripts that encode truncated proteins that may act as dominant negative isoforms. To test whether reduced Dnmt3b dosage can alter tumorigenesis, we bred Dnmt3b 1/2 mice to Em-Myc mice, a mouse model susceptible to B-cell lymphomas. Em-Myc/Dnmt3b 1/2 mice showed a dramatic acceleration of lymphomagenesis, greater even than that observed in Em-Myc mice that express a truncated DNMT3B isoform found in human tumors, DNMT3B7. This finding indicates that Dnmt3b can act as a haploinsufficient tumor suppressor gene. Although reduction in both Dnmt3b dosage and expression of DNMT3B7 within the Em-Myc system had similar effects on tumorigenesis and DNA hypermethylation, different molecular mechanisms appear to underlie these changes. This study offers insight into how de novo DNA methyltransferases function as tumor suppressors and the sensitivity of Myc-induced lymphomas to DNA methylation.