The identification of BRCA1/2 mutations represents a major milestone in understanding breast cancer and its inheritance, but germline mutations are estimated to be present in only 5-10% of cancers. Increasing evidence indicates that epigenetic modifications affecting the differentially methylated regions (DMRs) that regulate imprinted genes, link environmental exposures during early development to adult disease formation, including breast cancer. As cellular differentiation is essentially an epigenetic process that establishes and maintains cell-type specific gene regulation, the loss of cellular identity and runaway growth of cancer cells may be viewed as a loss of epigenetic control. Environmental factors involved in such epigenetic aberrations will vary by geography and lifestyle, and have differing impacts on health through interaction with underlying genetic factors. Investigation of epigenetic variation in relation to these environmental and genetic variants may prove key to understanding health disparities in breast cancer, with regard to incidence, tumor types, progression, and response to treatment regimens.

Genomic imprinting is an epigenetic form of gene regulation that results in diploid genes being rendered functionally haploid by parent of origin specific silencing of one allele. As imprinted genes are markedly enriched for growth effectors that interact in a coordinated way to regulate early growth and development, they form a particularly interesting set of genes in studying complex human diseases, particularly in regard to environmental influences. Imprinted gene expression is downregulated to prevent somatic tissue overgrowth, indicating that imprinted genes function in a co-regulated biological system or network to precisely regulate the onset and termination of growth in developing tissues. Combined with this coordinate control, the functionally haploid status of these genes suggests a one-hit model of deregulation that could cascade through multiple growth regulatory genes. Therefore, defining the human repertoire of imprinted genes and elucidating their epigenetic control mechanisms that maintain monoallelic expression is critical to better comprehending the etiology of complex diseases such as cancer.

The imprinted gene KCNK9 (TASK3) is a member of a gene set computationally predicted to be imprinted, and was one of the first from this set to have monoallelic expression confirmed. The initial monoallelic expression was observed in brain tissue, and we have since found monoallelic expression in breast tissue, which is also of ectodermal origin. This demonstration of KCNK9 imprinting is significant for cancer research, as KCNK9 has been shown to be overexpressed and subject to gene duplication in several cancer types, including colorectal and breast. If, as is suggested by these findings, KCNK9 dosage and expression level is significant, we hypothesize that loss of imprinting, and subsequent biallelic expression, is functionally equivalent to duplication of an active allele of KCNK9, suggesting multiple mechanisms for KCNK9 involvement in cancer formation.

By cross-species homology comparisons and analysis of ENCODE data on regulatory regions and chromatin structure, we identified several regions within the KCNK9 as potential DMRs for KCNK9, with allele specific differential methylation found for one site. The methylation state of the DMR also correlates to open/closed chromatin status at a proximal region that ENCODE data indicates as binding multiple transcription factors. Initial tests on tumor core biopsies showed bialleleic expression of KCNK9 in some samples, and frequent loss of methylation at the KCNK9 DMR.

Given the regulatory potential of this region, we investigated methylation of this DMR in breast cancer patients as well as women at high risk for breast cancer, to determine if changes in methylation levels correlate to cancer type, and if methylation changes occur early in cancer development, prior to tumor formation. From these patients, DNA from white blood cells (WBCs) and random periareolar fine needle aspiration (RPFNA) were collected for analysis. By comparison of methylation between RPFNA and WBC samples, and between right and left breast RPFNA samples, it could be seen whether any methylation alterations were localized or systemic. Localized changes would indicate late or cell-type specific alteration of the epigenetic state, whereas systemic methylation change would indicate an event early in embryonic development that could predispose individuals to develop breast cancer.

Our results demonstrated loss of methylation at the KCNK9 DMR in these patient samples – of 14 women for whom RPFNA samples were available for both breasts, 5 showed complete loss for both breasts, and 5 showed complete loss for one breast. Comparison of KCNK9 DMR methylation loss and cancer sub-type showed differences between Caucasian and African American patients, with Caucasians having methylation loss more frequently in estrogen receptor positive (ER+) cancers and Her2/neu overexpressing (HER2+) cancers. However, in TNBC, 90% of African American patients showed methylation loss, compared to 67% of Caucasians.

African American women have a higher incidence rate of TNBC than Caucasians, with African Americans also having younger average age of occurrence, larger tumors, greater proliferation, more lymph node involvement, and a higher death rate. Further characterization of KCNK9 in different tumor types from different ethnicities will clarify its role in tumor development and progression. The methylation state of the DMR may have use as a diagnostic or prognostic factor, particularly in at risk African Americans.

These results for the DMR of one imprinted gene provide additional impetus for analysis of DMRs of the entire human imprintome in cancer, given the number of known and predicted imprinted genes involved in growth, development, and disease. DMR patterns distinguishing cancer subtypes may be identified, creating new diagnostic and treatment options. The roles of environmental exposures and genetics, and their interaction, in tumorigenesis would be better understood. This would work towards the goal of improving prevention and treatment options to close the disparity gap for individuals disproportionately at risk for these aggressive tumor subtypes.

Citation Format: David Skaar, Michael Gould, Randy Jirtle, Victoria Seewaldt. Altered imprinted gene DMR regulatory methylation in breast cancer. [abstract]. In: Proceedings of the Sixth AACR Conference: The Science of Cancer Health Disparities; Dec 6–9, 2013; Atlanta, GA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2014;23(11 Suppl):Abstract nr CN02-03. doi:10.1158/1538-7755.DISP13-CN02-03