Biochemical and Biophysical Research Communications
ReviewEpigenetic epidemiology of cancer
Introduction
Epigenetic epidemiology includes the study of variation in epigenetic traits and the risk of disease in populations. The merging of these two fields can facilitate insight into which epigenetic marks are associated with cancer, whether some of these marks explain the link between certain exposures and cancer, and how these epigenetic marks can be utilised as biomarkers. Epigenetic epidemiology can therefore serve to promote primary cancer prevention by identifying risk factors and their method of action, secondary prevention by establishing markers of early disease, and tertiary prevention by establishing markers of disease progression and drug resistance. Biomarkers can be identified by the adoption of appropriate study designs, in conjunction with a solid understanding of how any of the three cornerstones of epigenetics may be involved in carcinogenesis: DNA methylation; chromatin and histone modifications; and non-coding RNAs.
In addition to the identification of markers of drug resistance, insight into epigenetic dysregulation of the genome also provides new bases for therapies. For example, azacitidine is a demethylating agent that acts through inhibition of DNA methyltransferases and has been approved for use with myelodysplastic syndromes. It is currently in Phase I and Phase II clinical trials for use with diffuse large B-cell lymphomas, non-small cell lung cancers (NSCLC), breast cancer, pancreatic cancer, and oesophageal cancer. Histone deacetylase inhibitors, such as vorinostat and romidepsin, have been approved for use with the treatment of T-cell lymphomas. While no microRNA-based therapies have yet been approved for clinical usage, Phase I clinical trials are underway investigating the use of an miR-34 mimic, MRX34, with liver cancer and lymphoma patients. It is therefore evident that better understanding the epigenetic basis of cancer is enabling the development of a range of new therapeutic options.
In this review, we will describe how epidemiology studies have related epigenetic variation with environmental factors and have identified diagnostic and prognostic biomarkers that can be applied in the clinical setting. We will describe the challenges in study design, and we will review progress that has been made in identifying biomarkers of disease risk, and especially the efforts in developing non-invasive means of screening patients. We will focus exclusively on work using primary human tissues, and we will pay particular attention to large-scale and prospective studies due to their relative strength in identifying biomarkers and the epigenetic dysregulation associated with cancer.
Section snippets
The suitability of epigenetic biomarkers
Epigenetic traits have the potential to serve as excellent diagnostic and prognostic markers of cancer. In addition to the stability of DNA methylation and the resistance of microRNAs to RNase-degradation, aberrant epigenetic events are frequently observed in early-stage cancers and in adenomas [1], [2], [3], [4], [5], [6], [7]. Significantly increased stochastic variation in DNA methylation has been observed in cervical cells of normal morphology in patients who went on to develop cervical
Epigenome-wide association studies
It has become increasingly common for epigenetic epidemiology studies to use microarray or next-generation sequencing technology in order to assess epigenetic variation on a substantially larger scale, rather than using a surrogate marker of global methylation, such as LINE-1 (Fig. 1). Referred to as epigenome-wide association studies (EWAS), these can offer a cost-effective means of interrogating large numbers of loci without requiring infeasible quantities of starting material, and can bring
Study design
The identification of epigenetic biomarkers requires an appropriate study design. Retrospective case-control studies cannot determine the temporality of epigenetic modifications with regard to disease occurrence. Moreover, in cancer development ‘drivers’ and ‘passengers’ cannot be distinguished, which prohibits identification of early markers of disease. Prospective studies permit temporal associations and are therefore highly preferable for the identification of biomarkers (a summary of
The use of blood as a surrogate tissue
Prospective epidemiology studies are commonly performed with several thousand participants. Since it is not feasible to collect tissue samples, such as biopsies, from target organs in healthy individuals, prediagnostic specimen collection in prospective studies is restricted to blood, urine and other easily accessible tissues. Blood-based biomarkers are also advantageous in the clinical setting as they facilitate easier and cost-effective methods of screening, and they substantially reduce the
Non-invasive sample collection as an alternative to the use of biopsies
Many tissues provide alternative non-invasive means to the use of peripheral blood. As will be described later, exfoliated cells can be obtained from sputum to study the lung, from urine to study the bladder, and from stool to study the colon. Further to these examples, a cytosponge method has been developed to retrieve cells from the oesophagus. A screening approach based upon this method has been estimated by microsimulation modeling to possibly reduce the number of cases of oesophageal
Colorectal cancer
Perhaps the most widely-studied cancer with regard to epigenetic changes is that of the colorectum. There is evidence that epigenetic silencing of genes may be an early event in the adenoma-carcinoma sequence, as suggested by the Vogelstein model [46]. The MGMT promoter has been reported to be methylated in 34% of adenomas [3], while Psofaki et al. [47] reported aberrant methylation of the CDKN2A/p16 and MGMT promoters in 67% and 76% tubulovillous and villous adenomas respectively. Together,
Lung cancer
There is evidence that the leading risk factors for lung cancer, tobacco smoke and air pollution, can influence DNA methylation in the lungs and in blood. A study nested within the EPIC cohort reported hypomethylation of F2RL3, AHRR and two intergenic regions in the blood of smokers in comparison to non-smokers [92]. A follow-up study from the same group demonstrated that four loci, including AHRR, may serve as blood-based markers of tobacco smoke exposure [93]. While many exposure studies
Breast cancer
Breast cancer subtypes are epigenetically distinct, with tumours displaying differing LINE-1 methylation profiles according to the expression of the oestrogen and progesterone receptors [115], [116], [117], [118], [119] and BRCA1 inactivation [119]. The identification of these differences in methylation profiles may suggest that different breast cancer subtypes may be candidates for azacitidine treatment, just as expression of the hormone receptors identify patients who will benefit from
Bladder cancer
Exposure to arsenic and tobacco smoke have been established as leading risk factors for bladder cancer, and these have been demonstrated to be associated with epigenetic changes in healthy populations. A Spanish study of 892 healthy individuals observed that LINE-1 methylation levels in blood are associated with toenail arsenic levels and smoking blond tobacco [38], and the former is an observation that has also been made elsewhere in healthy individuals [149]. However, while statistically
Insights into other cancers
There have been promising developments into the identification of epigenetic biomarkers for a range of other cancers. In particular, the hypermethylation of APC, GSTP1, RARβ and RASSF1A have shown promise in clinical trials aimed at identifying diagnostic markers for prostate cancer [156], [157] and may be applicable to tests using urine from patients [158]. The hypermethylation of the MGMT promoter appears to be an excellent marker with gliomas, being associated with survival [159] and the
Future prospects
The prospects for the application of epigenetic biomarkers in the clinical setting are excellent. Increasing numbers of potential diagnostic and prognostic markers are being identified for a range of cancers, although many promising leads are still to be validated in prospective studies. The advent of increasingly sensitive technologies, and with these becoming ever more affordable, is facilitating the development of sensitive tests based upon a panel of genes. Furthermore, there have been
Conclusions
Epigenetic epidemiology is providing important insight into how epigenetic alterations are associated with the development of cancer. This is being utilised to benefit patients through a better understanding of risk factors, and by the identification of novel biomarkers that enable early detection of disease and which can inform upon appropriate treatment strategies. The cost and duration of prospective studies can be prohibitive, but they remain the ‘gold standard’ for the identification and
References (161)
- et al.
Genomic methylation of leukocyte DNA in relation to colorectal adenoma among asymptomatic women
Gastroenterology
(2008) - et al.
LUMA (LUminometric Methylation Assay) – a high throughput method to the analysis of genomic DNA methylation
Exp. Cell Res.
(2006) - et al.
Safety and efficacy of endoscopic mucosal therapy with radiofrequency ablation for patients with neoplastic Barrett’s esophagus
Clin. Gastroenterol. Hepatol.
(2013) - et al.
A genetic model for colorectal tumorigenesis
Cell
(1990) - et al.
Folate status, genomic DNA hypomethylation, and risk of colorectal adenoma and cancer: a case control study
Gastroenterology
(2003) - et al.
Microsatellite instability as a marker of prognosis and response to therapy: a meta-analysis of colorectal cancer survival data
Eur. J. Cancer
(2010) - et al.
Global histone modification of H3K27 correlates with the outcomes in patients with metachronous liver metastasis of colorectal cancer
Eur. J. Surg. Oncol.
(2013) - et al.
Use of 5-fluorouracil and survival in patients with microsatellite-unstable colorectal cancer
Gastroenterology
(2004) - et al.
Epigenetic status of LINE-1 predicts clinical outcome in early-stage rectal cancer
Br. J. Cancer
(2013) - et al.
Association of RASSF1A and p63 with poor recurrence-free survival in node-negative stage I-II non-small cell lung cancer
Clin. Cancer Res.
(2013)