Targeting Epigenetics to Prevent Obesity Promoted Cancers

Obesity is associated with increased and worse prognosis for multiple malignancies (1, 2). Even early-age obesity may elevate the risk of subsequent adult malignancy (3, 4). Although multiple metabolic mechanisms may mediate the obesity–cancer linkage (5), the sustained, latent effects of obesity suggest the possibility that it may lead to more durable genetic and/or epigenetic alterations that promote cancer development over extended periods of time. These epigenetic effects, which are mediated by covalent modification of chromatin DNA and proteins, that alter genetic readout without changing nucleotide sequence, have now become targets for both cancer chemotherapy and cancer chemoprevention (6). Because it is theoretically possible to interfere with epigenetic regulation without cytotoxic consequences, this approach to chemoprevention is especially of interest for disrupting the obesity–cancer linkage. This issue of Cancer Prevention Research contains articles from the Bernard and the Liby laboratories at Michigan State University (East Lansing, MI), reporting that an epigenetic targeting agent, I-BET-762, can attenuate and may eventually contribute to prevention of obesity-mediated carcinogenesis in a mouse model (7, 8).

This strategy of targeting epigenetic changes to prevent obesity-promoted cancer represents the convergence, over 70 years, of multiple research disciplines including:

• (i) Epidemiologic studies showing that nutritional stress can impact delayed and even heritable traits.

• (ii) Chromatin research demonstrating multiple molecular alterations and mediators, which can change gene expression without changing base sequence, sometimes in a heritable epigenetic process.

• (iii) Intensive structure–activity pharmacologic research to develop agents targeting epigenetic processes; and

• (iv) Clinical and molecular studies identifying cancer-promoting epigenetic effects of obesity and interventions to disrupt these linkages.

Multiple epidemiologic and laboratory model studies now indicate that epigenetic effects of prenatal and early childhood nutrition can impact age-related disease occurring long after the initial exposure. Studies of preconception and perinatal nutrition experiences, including the Överkalix Famine 1836, the Dutch Hunger Winter 1944–1945, and The Great Chinese Famine 1958–1961, showed associations between early childhood energy availability, adult obesity, and other diseases occurring both in adulthood and in subsequent generations (9, 10). Another study, the Boyd Orr Cohort, which characterized energy consumption among 1,352 families in prewar Britain, 1937–1939, and then followed cause of death in 3,834 subjects over six decades, revealed a positive association between dietary energy consumption during childhood and nonsmoking cancer-related mortality in adulthood (11).

Studies of chromatin dating back to the 1960s showed that interaction of DNA with chromosomal proteins, including histone and nonhistone proteins, was critical to regulating both the tertiary structure and function of DNA and the processes controlling genetic readout or transcription through RNA synthesis (12). The application of chromatin immunoprecipitation (ChIP) and parallel DNA sequencing (ChIP-Seq) revealed sites of DNA where associated proteins altered the genetic readout. These processes controlling gene transcription were shown to be associated with a host of covalent modifications, including methylation of DNA bases and methylation, acetylation, phosphorylation, and other covalent modifications of histone proteins, which affect transcription factor–binding affinity and specificity, thereby altering the genetic readout without changing the nucleotide sequence (6). Further studies showed that these epigenetic effects alter the tertiary structure of DNA at gene-regulatory elements, including promoters and enhancers (13, 14). In some cases, these epigenetic events lead to activation of otherwise quiescent oncogenes or silencing of tumor suppressors (13, 14). These epigenetic signaling processes are controlled by transcription factors and multiple enzymes that attach, remove, recognize, and respond to “DNA and protein marks” descriptively referred to as “writers,” “erasers,” “readers,” “remodelers,” etc. (6).

Although cells can respond to multiple internal and external signals by rapidly changing and restoring metabolic processes, epigenetic changes may be more durable, can be passed through multiple rounds of cell division, and can even become transgenerational. Thus, applying modern techniques to probe the chromatin of long-term survivors of perinatal and early childhood extremes in nutritional energy consumption shows persistent alterations, sometimes decades later, in epigenetic marks (9–11). Laboratory studies in cell culture and animal models have likewise shown significant differences in epigenetic modifications of DNA and chromatin proteins in response to multiple nutrients, including polyphenols, resveratrol, fatty acids, folic acid, and vitamin C (15, 16).

Pharmacologic approaches to disrupt epigenetic processes involved in cancer, target both DNA and histone protein modifications. Cytosine base methylation in DNA, particularly in CG dinucleotides, has been shown to be important in silencing tumor suppressors and allowing overexpression of oncogenes (17). Thus, the long pharmacologic quest for synthesis of nucleoside analogues to interfere with cancer cell metabolism and replication led to 5-aza cytidine and 5-aza-2′-deoxycytidine, either of which when incorporated into DNA in place of deoxycytidine is refractory to methylation because the carbon atom in the 5 position of cytosine is replaced by nitrogen, leading to trapping and degradation of the DNA methyltransferase. These two, well-characterized, DNA methyltransferase inhibitors (DNMTi) are currently in widespread clinical use as Vidaza and Dacogen for treatment of hematologic malignancies, including myelodysplastic syndrome and acute myeloid leukemia. Their clinical success has led to an extensive search to develop new anticancer nucleoside and nonnucleoside inhibitors (17).

Therapeutic targeting of chromatin protein modifications was serendipitously initiated in the early 1970s by Charlotte Friend (18), who noted that treatment of murine erythroleukemia cells with DMSO, to enhance viral infectivity, stimulated terminal erythroid differentiation, maturation, and hemoglobin synthesis. Subsequent structure–activity studies led to development and optimization of a pharmacologic intervention employing suberoylanilide hydroxyamic acid, which functions as a histone deacetylase inhibitor and is now clinically employed as vorinostat for the treatment of cutaneous T-cell leukemia (19). This search has expanded such that there are now multiple inhibitors of histone modification, being tested in clinical trials for antitumor effects (6).

Studies of the “readers” that recognize and respond to the epigenetic marks, have led to identification of unique protein domains that selectively bind specific DNA or histone modifications. These proteins usually have additional domains, which further bind and organize proteins, to produce a downstream transcription process (20). For example, bromodomain and extraterminal domain (BET) proteins have two protein motifs that recognize and bind acetylated lysine residues, such as those that occur in N-terminal tails of histones. They likewise contain an extraterminal domain protein motif that recruits additional proteins, which function as transcription regulators. Binding of BET proteins to acetylated histone tails contributes to chromatin structure remodeling and frequently upregulates synthesis of the myc oncogene. Accordingly, major efforts have been focused on developing BET domain inhibitors with emphasis on their use as cancer therapeutics (21).

Multiple authors have suggested a potential cancer-preventive role for epigenetic inhibitors (6, 22, 23); however, most studies focus on their use to treat established tumors. Chronic oral administration of zebularine, a nucleoside analogue, DNMTi, with antitumor activity in vitro and in vivo, was previously shown to decrease and delay development of intestinal neoplasms in genetically predisposed APC^min/+ (Min) female but not male mice (22). The current studies from the Bernard and Liby laboratories now demonstrate that a new BET inhibitor, I-BET-762, can attenuate the development of several obesity promoted malignancies (7, 8).

The tumor-promoting effects of obesity have been considered to be mediated, in part, by epigenetic effects for multiple reasons. In some cases, these effects manifest after long latent periods (3, 4, 24). For example, population-based studies indicate that patients undergoing bariatric surgery for severe obesity may still have elevated risk for colorectal cancer that continues to increase with time, even after 10 years following surgery (24). Other recent studies in murine models show that diet-induced obesity, compared with caloric restriction, in female MMTV-HER2NEU TG mice resulted in decreased DNA methylation and downregulation of ERα and ERβ expression, promoting breast cancer with worse prognosis in obese mice (25). In another study, obesity-driven changes in histone H3 acetylation in mouse colon epithelium were associated with remodeling of chromatin enhancer regions to demonstrate loss of patterns integral to normal differentiation and enrichment for transcription factor–binding sites driving colon cancer progression, including downstream components of RAS, PI3K, and JNK signaling pathways (26).

I-BET-762 is a benzodiazepine derivative shown to have antitumor activity in preclinical models and is currently in clinical trials against NUT midline, breast and lung cancers (27, 28). Interestingly, in this issue of Cancer Prevention Research, the chemopreventive effects, rather than its antitumor effects, were demonstrated. I-BET-762 decreased cell proliferation and downregulated cMyc and Cyclin D1 expression and delayed growth in tissue culture of primary tumor explants, and in breast and lung cancer cell lines (8). Importantly, I-BET-762 delayed, but did not prevent, development of breast tumors in MMTV-PyMT mice. Furthermore, I-BET-762 reduced the number and size of vinyl carbamate induced lung cancers in A/J mice.

Previous studies of mechanisms of obesity-linked carcinogenesis have shown that high-fat diet–fed mice show increased UVB-induced carcinogenesis of nonmelanoma skin cancer, that the effects could be mitigated by surgical removal of visceral adipose tissue (VAT), and that the procarcinogenic effects of VAT could be reconstituted by replacement with conditioned medium prepared from mouse (MFTF) or human (HuFTF) adipose tissue derived from obese donors (7). Interestingly, Chakraborty and colleagues previously showed that treatment of immortalized keratinocytes with conditioned media derived from VAT of mice fed a high-fat diet, or VAT from obese humans, enhanced expression of cMyc. They now show that I-BET-762 attenuates upregulation of cMyc and reduces colony growth in soft agar of immortalized keratinocytes. Importantly, I-BET-762 attenuated JB6 P+ mouse skin epithelial cell transformation in soft agar and reduced FGF2 promoted tumor growth of JB6 P+ cells in immunocompromised mice. The efficacy of I-BET-762 was also demonstrated by decreased expression of cMyc and decreased transformation of MCF 10A and NMuMG nontransformed mammary tumor cell lines.

These studies suggest that adipose tissue factors promote malignant transformation in target tissues by mechanisms involving enhanced C-Myc activity, and this process can be attenuated, in part, by the epigenetic targeting agent I-BET-762. Although it did not completely prevent tumor development, these studies provide important proof of principle that targeting epigenetic processes in general, and more specifically engaging BET protein inhibition, constitutes a noncytotoxic strategy for developing cancer chemoprevention. Moreover, this approach may be particularly useful for disrupting the obesity–cancer linkage.

Because obesity is likely to engage multiple mechanisms of epigenetic regulation, involving different forms of chromatin covalent modification, it is likely that maximal chemopreventive effects will require a combination of agents targeting the different epigenetic modulation processes. Because the epigenetic effects of obesity are probably exerted over long latent periods, chemopreventive agents will need to be administered over long periods following exposure. Accordingly, they should be highly effective upon oral administration, safe and nontoxic at effective epigenetic targeting doses; they should not interfere with growth and development and have no reproductive toxicity.

Recent estimates of the worldwide obesity pandemic indicate that there are 120 million obese children and adolescents and 640 million obese adults on a worldwide basis, who constitute the at-risk population for obesity-linked malignancies (29, 30). Thus, the patient population that might benefit from these chemopreventive agents is enormous and the potential impact critically important.

No potential conflicts of interest were disclosed.

Nathan A. Berger was supported, in part, by NIH grants GI SPORE P50CA150964, BETRNet U54CA163060, and R01 M000969. Peter C. Scacheri was supported, in part, by NIH grants R01CA160356, R01CA193677, R01CA204279, and R01CA143237.