Cancer cells display widespread genetic and epigenetic abnormalities, but the contribution to disease risk, particularly in normal tissue before disease, is not yet established. Genome-wide hypomethylation occurs frequently in tumors and may facilitate chromosome instability, aberrant transcription and transposable elements reactivation. Several epidemiologic case–control studies have reported genomic hypomethylation in peripheral blood of cancer patients, suggesting a systemic effect of hypomethylation on disease predisposition, which may be exploited for biomarker development. However, more recent studies have failed to reproduce this. Here, we report a meta-analysis, indicating a consistent inverse association between genomic 5-methylcytosine levels and cancer risk [95% confidence interval (CI), 1.2–6.1], but no overall risk association for studies using surrogates for genomic methylation, including methylation at the LINE-1 repetitive element (95% CI, 0.8–1.7). However, studies have been highly heterogeneous in terms of experimental design, assay type, and analytical methods. We discuss the limitations of the current approaches, including the low interindividual variability of surrogate assays such as LINE1 and the importance of using prospective studies to investigate DNA methylation in disease risk. Insights into genomic location of hypomethylation, from recent whole genome, high-resolution methylome maps, will help address this interesting and clinically important question. Cancer Prev Res; 5(12); 1345–57. ©2012 AACR.

Epigenetics is the mitotically heritable control of gene expression and chromatin structure through covalent modification by DNA methylation, posttranslational modifications of histone proteins, and control of gene expression by noncoding RNAs (1, 2). DNA methylation is the addition of methyl groups to cytosines within CpG dinucleotides, to form 5-methylcytosine (5meC) and is catalyzed by the DNA methyltransferases (DNMT; refs. 2, 3). DNA methylation is dependent on the 1-carbon metabolism pathway, which requires dietary supply of folate, homocysteine and choline, and other micronutrients (4, 5). The role of DNA methylation as a negative regulator of transcription initiation at CpG islands (CGI), regions of high CpG density found in approximately 60% of genes, is well established, and is required for silencing of developmental and tissue specific genes (2, 6, 7). DNA Methylation facilitates transcriptional repression independently through preventing binding of DNA polymerases and transcription factors, or through interaction with methyl binding domain proteins and other epigenetic mechanisms such as repressive histone marks (8, 9). The functional roles of DNA methylation elsewhere in the genome is less well understood; however, DNA methylation is thought to confer genomic stability and integrity, and high methylation at repetitive elements is proposed to protect against expression of transposable elements (TE) and endogenous retroviruses (10, 11). Epigenetic regulation is often studied in the context of environmental and population health, as widespread DNA methylation patterns are known to be affected by environmental, lifestyle, and demographic factors that affect complex disease risk, such as diet, carcinogen exposure, reproductive factors, and age (12, 13). Furthermore, this “plasticity” or susceptibility to chemical alteration of the epigenome provides promise for therapeutic intervention where abnormal epigenetic patterns occur in disease or in relation to disease risk (14). A current research topic is the potential of DNA methylation variability in normal tissues such as peripheral blood, for development of cancer predisposition biomarkers (15, 16). This review and meta-analysis aims to synthesise all of the current data and to answer the question of whether there is a link between genome-wide DNA hypomethylation and cancer risk.

DNA methylation dysregulation occurs in almost all cancers, and is becoming established as a hallmark of cancer (1, 11, 17). Most research has been concerned with local hypermethylation of promoter CpG islands, as transcriptional repression because of promoter hypermethylation of tumor suppressor genes, such as MLH1 and BRCA1 is implicated in driving several cancers (7, 8). However, the first described cancer-associated epigenetic phenomenon was the lower net percentage of 5meC in tumor tissue compared with equivalent normal tissue, known as “genome-wide” or “global” hypomethylation, that occurs frequently in all cancer types (18, 19). Despite considerable research, it remains unclear whether promoter hypermethylation and genome-wide hypomethylation are mechanistically linked, or are independent processes. (20, 21). Whole-genome DNA methylation (methylome) analysis has so far revealed than an overall increase in methylation variability (hypervariability) may be more prevalent than discreet changes in DNA methylation levels in cancer, suggesting a widespread loss of epigenetic control (9, 17, 22). Recent research has shown that most cancer-associated DNA hypermethylation occurs at “CpG shores,” regions flanking CpG islands, rather than within CpG islands (17, 23). Furthermore, genome-wide hypomethylation is largely confined to large genomic “hypomethylation blocks” (9, 17), that tend to occur in regions that display intermediate methylation levels in normal tissue, termed partially methylated domains (PMD; refs. 17, 22), and regions of low CpG density (23). That hypomethylation is specifically associated with repetitive elements has been called into question, with only a modest enrichment of repeats in hypomethylation blocks (17, 24). Such apparent compartmentalization of hypomethylation may reflect higher order changes in chromatin structure (9). Whereas genome-wide hypomethylation has long been implicated in loss of transcriptional repression, recent evidence suggests that hypomethylation at PMDs or gene bodies may be often associated with gene repression through formation of repressive chromatin (24). Therefore, understanding of cancer epigenetics is rapidly advancing with the advent of whole-methylome sequencing, revealing a greater complexity and organization to the epigenome than was previously appreciated.

The cause of genome-wide hypomethylation remains poorly understood, and several factors have been implicated (11, 21). Dietary insufficiency of folate and other 1-carbon metabolism-dependent micronutrients, or germline or somatic mutations in members of the pathway can induce hypomethylation in tumor, and normal tissues including blood (4, 14, 25, 26). Exogenous exposures, such as carcinogen exposure may influence methylation by inducing DNA damage or by affecting DNMT enzyme activity (4, 11, 27, 28). Furthermore, the efficiency of DNA methylation may be affected by age (12, 29). Finally, neoplasia may induce “field-effect” hypomethylation in cancer and surrounding histologically normal tissues, through sequestration of DNA methylation enzymes and substrates by rapidly proliferating, dividing cancer cells, resulting in failure of DNA methylation maintenance in surrounding tissue (11, 30–35).

Several tumorigenic events may result from genome-wide hypomethylation. These include chromosomal instability, which may facilitate gene-dosage alterations, mutations, genetic recombination, large deletions, or translocations (2, 11, 36). Hypomethylation may facilitate altered expression of oncogenes and/or tumor suppressors, as well as aberrant transcription of noncoding RNAs via transcriptional read-through subsequent to a loss of repression at repetitive DNA (37–39). Expression of repetitive elements is a feature of many cancers, with largely unknown consequences (37, 40). Whether genome-wide hypomethylation represents an early or causative tumorigenic event or a passive consequence of cancer remains unknown (reviewed elsewhere (21)). The detection of hypomethylation in cancer precursor lesions and normal adjacent tissue, and the induction of cancers in animal models with experimentally induced hypomethylation suggest a causative role (30, 41–43). However, the occurrence of many cancers without apparent hypomethylation, and the progression of hypomethylation with cancer stage suggest a passive role (20, 21).

Because of the conventional view that genome-wide hypomethylation is associated with repetitive elements, surrogate assays for genome-wide methylation have been developed specifically targeting consensus sequences within repetitive elements, including LINEs (Long Interspersed Nucleotide Element 1, LINE1), SINEs (Alu), and Satellite repeats (Sat2; refs. 44, 45). LINE1 (hereafter referred to as “L1”) is the only autonomous [capable of independent retrotransposition (transposition via an RNA intermediate)], and most highly expressed TE in the human genome, comprising approximately 17% of the human genome, with more than 500,000 copies (11, 46, 47). A full length L1 element is approximately 6 kb long with a bidirectional, noncanonical promoter, and 2 open reading frames coding for an endonuclease and retrotransposition machinery proteins (11, 46, 47). L1 transcription is largely regulated by DNA methylation of the 5′ promoter; however, most L1 elements are truncated and cannot be transcribed (37, 38, 40). Approximately 100 L1 elements are functionally capable of retrotransposition; however, only a few contribute to the vast majority of retrotransposition events (46). Alu elements, of which there are multiple families, are the most common TE in the human genome, with approximately 1.1 million copies, comprising roughly 11% of the genome (11, 47). Alu elements are nonautonomous and require the L1 transposition machinery for transposition (11). Satellite repeats (including Sat2) are short tandemly repeated noncoding DNA, frequently in centromeric and heterochromatic regions of chromosome 1 (48). Both L1 and Alu elements are heavily methylated in normal somatic tissue, however, hypomethylation of both, especially L1, is often detectable in tumors (20, 30, 49). Whereas transcription of TEs are required for their transposition, transposition-independent consequences of transcription may have functional consequences, as an estimated 7% of the human transcriptome is derived from transcription start sites within L1 elements (37) and hypomethylation induced transcription of L1 elements within host gene introns can effect host gene transcription through RNA interference (39) and through driving host gene ectopic expression from the L1 antisense promoter (50).

We and others have hypothesized that the epigenetic variability may contribute to risk of cancer development (14, 15, 51–53). “Epigenetic epidemiology” refers to the investigation of epigenetic patterns associated with disease risk, and such research holds great potential for cancer prevention and cancer risk-biomarker development, especially where abnormal epigenetic patterns may be detectable in easily accessible tissues such as blood, which is suitable for population screening (15, 16).

Methylation variability may contribute to, or be predictive of, risk of tumorigenesis in other tissues for several reasons. Many innate cancer risk factors, such as age, anthropometric factors, and genetic factors affect epigenetic patterns in blood (54–56). For example, rare epimutations in MLH1 and MSH2, which consist of local DNA hypermethylation events detectable in all tissues and conferring high risk of colorectal cancer, have been shown to occur because of genetic polymorphisms (57, 58). Blood DNA methylation is also affected by exposure to environmental and lifestyle cancer risk factors, such as smoking, alcohol, and other carcinogens (59, 60), therefore, blood methylation may provide useful biomarkers for acquired or environmentally induced cancer risk (12, 13). Finally, inflammation predisposes to risk of many cancers (61, 62), and affects DNA methylation (63), suggesting that blood methylation may reflect immune effects on cancer risk.

Promising for this avenue of research is the identification of a variable methylated region within the ATM gene associated with breast cancer risk (64). Furthermore, hypermethylation of BRCA1 is reportedly associated with increased prevalence of BRCA1-methylated breast tumors (65, 66), suggesting a functional link between blood detectable DNA methylation and disease histology (67–69). High-throughput discovery studies have identified cancer-associated DNA methylation signatures in blood of patients with breast (69), ovarian (68), bladder (67), and head and neck cancers (70), providing potential cancer diagnostic biomarkers.

The most widely studied putative epigenetic risk marker for cancer is genome-wide or “global” DNA methylation in blood. The studies investigating this, their findings, and the research approaches used, will be discussed further.

A challenge in investigating genome-wide DNA methylation in population studies is the lack of cost-effective, high-throughput assays that have both wide genome coverage and high resolution (45). Early studies used methods that measure genomic 5meC content using 5meC-specific antibodies (71), methyl-acceptance assays (72) and high-performance liquid chromatography (HPLC) (73). Such methods, however, give no information about the spatial arrangement or genomic location of DNA methylation and require large amounts of input DNA, making them unsuitable for population studies using precious patient samples. Widespread DNA methylation patterns can be measured using methylation sensitive restriction enzymes with relatively high resolution; however, methylation analysis is biased toward regions of high CpG density, as the enzyme restriction sites occur at sequences such as CCGG, and CGCG for the HpaII and Hha1 restriction enzymes, respectively (45, 74). Most DNA methylation assays are based on bisulphite conversion of DNA, where incubation of DNA with sodium bisulfite causes deamination of unmethylated, but not methylated, cytosines to uracil (45). Sequencing of bisulphite converted DNA represents the “gold standard” for DNA methylation analysis (16, 75); however, real-time PCR, restriction enzyme based [combined bisulphite restriction analysis (COBRA; ref. 76)], and microarray-based methods (75) are also used. Most popular and convenient for population studies is the measurement of methylation at the repetitive elements such as L1, Alu, and Sat2, which are distributed at high frequency throughout the human genome (11, 44). These are considered “surrogate” measures of genome-wide DNA methylation as their methylation is thought to reflect genome-wide methylation levels (77). However, the efficacy of these surrogate assays for genome-wide methylation has been questioned (refs. 14, 78; discussed below), and a reinterpretation of the results reported using these assays is warranted.

Twenty-three publications have reported population based cancer case–control studies investigating blood genomic DNA methylation in relation to cancer risk (Table 1). It is important to note that the comparability of these studies is limited by several differences between each, including cancer type, assay, study design (prospective/retrospective), length of time between blood-draw and diagnosis, sample size, sex (male/female/mixed), ethnicity, cancer treatment exposure, and analytical/statistical methods. Furthermore, populations at different cancer risk are included, for instance, one report included elderly men at high cancer risk (79), whereas another included Asian women at lower risk of breast cancer (80). Some reports included more than 1 “study,” because of use of different assays, different patient populations, or different study designs, so altogether there were 34 individual studies of genome-wide methylation and cancer incidence/prevalence (Table 1). The greatest limitation to comparison between studies was reporting of data analyses. Twenty-six studies reported an odds ratio (OR) for cancer associated with DNA methylation, generated by categorical analysis. Categorical analysis compares the OR for cancer for individuals displaying methylation within the lowest category (tertile, quartile, or decile) of methylation, compared with individuals within the highest (reference) category. The choice of categorical “split” for methylation affects the OR, with fewer categories providing more conservative analysis, and this varied between studies (Table 1). OR was the most consistently reported, comparable, and representative (of overall results) factor between studies, and was, therefore, used for meta-analysis. Therefore, 8 studies reporting risk analysis at the mean/median level only (none of which showed significant case–control differences) could not be included, resulting in an overestimate of any positive effects (refs. 69, 73, 79, 81–83; Table 1). We have recently reported methylation analysis of blood L1 methylation in 2 population-based prospective case–control studies for breast cancer risk, but did not report categorical analysis (78). However, for the purpose of comparison with other studies, we have included categorical analyses of these in this review. In addition, as 4 studies (3 reports; refs. 31, 84, 85) used the lowest, rather than the highest methylation category as the reference category, inverse ORs were calculated for these studies. Importantly, use of different methylation “split” (tertile, quartile, quintile, or decile) for categorical analysis between reports is a potential confounding factor; however, this cannot be easily corrected without obtaining the raw data for each study. A meta-analysis of many of the relevant studies was recently reported by Woo and colleagues (86), which did not include 5 relevant and recent reports (64, 84, 87–89). Therefore, we report a revised meta-analysis including these studies (Fig. 1). A thorough literary search was carried out to identify all case–control studies relationship between genomic DNA methylation in blood, and either cancer incidence or prevalence (Table 1). Included were studies investigating any cancer type and using any quantitative measurement of genome-wide DNA methylation, including 5meC measures and surrogate assays. Excluded were studies measuring methylation at single-locus sites, or measuring methylation is tissues other than blood. Summary estimates were weighted by sample size, using random-effects models because of significant interstudy heterogeneity. (Full details of search strategy and meta-analysis are provided in supplementary methods.). The overall summary estimate OR for all using a random effects model was OR, 1.4 (0.9–1.9), suggesting no overall association with cancer risk. Summary ORs for studies using L1, Alu, and Sat2 repetitive elements were not significantly associated with cancer risk. It appears that total 5meC genomic content is the most consistent association with cancer prevalence, as all 5 studies identified significant associations between hypomethylation and cancer prevalence in categorical analysis, 4 of which also found significant association at the mean level. Though the effect sizes are variable, with smaller studies showing the largest ORs, it appears evident that genomic 5meC content is reduced in blood of cancer patients, suggesting potential for biomarker development. Interestingly, these studies represent the earliest published reports investigating blood genomic methylation in relation to cancer risk, and no similar study has attempted to replicate these findings since. Given that most studies using surrogate assays are null for cancer risk, it may be worthwhile returning to 5meC measures in an effort to reproduce this association. Interestingly, the only study (84) investigating widespread genomic methylation, using the restriction enzyme and bisulphite-pyrosequencing–based luminometric methylation assay (LUMA), identified a strong protective effect of genomic DNA hypomethylation on breast cancer risk. This inconsistent finding was likely influenced by the use of restriction enzymes, which measure methylation mainly at CGIs, which tend to be unmethylated, meaning that the only detectable change would be an increase in methylation (84).

Figure 1.

Meta-analysis of ORs reported in studies investigating Genome-wide DNA methylation in peripheral blood DNA for cancer risk. Test for heterogeneity showed highly significant heterogeneity across all studies (P < 0.001), and specifically in the analysis of 5meC (P < 0.001) and LINE1 (P < 0.001), but not significant for Alu (P = 0.121) or Sat2 (P = 0.827). Squares represent the size of the study and are centered on the OR with whiskers representing the 95% CIs. Random effects (RE) model was used for all summary analyses and the width of the diamond represents the confidence intervals. CRC, colorectal cancer; CRA, colorectal adenoma; BGS, breakthrough generations study; EPIC, European Prospective Investigation into Cancer and Nutrition.

Figure 1.

Meta-analysis of ORs reported in studies investigating Genome-wide DNA methylation in peripheral blood DNA for cancer risk. Test for heterogeneity showed highly significant heterogeneity across all studies (P < 0.001), and specifically in the analysis of 5meC (P < 0.001) and LINE1 (P < 0.001), but not significant for Alu (P = 0.121) or Sat2 (P = 0.827). Squares represent the size of the study and are centered on the OR with whiskers representing the 95% CIs. Random effects (RE) model was used for all summary analyses and the width of the diamond represents the confidence intervals. CRC, colorectal cancer; CRA, colorectal adenoma; BGS, breakthrough generations study; EPIC, European Prospective Investigation into Cancer and Nutrition.

Close modal
Table 1.

Study details abstracted from all reports included in meta-analysis and all relevant studies excluded

ReferenceAuthorPublication dateJournalCancer typeMeasurementAssayCategorical split*sexFirst authorReported restricted ethnicityCountry/continentn (case/ctrl)Prospective/retrospectiveNewly diagnosedPre/postchemotherapyCases (description)Controls (description)
72 Pufulete 5/2/2003 GASTROENTEROLOGY Colorectal 5meC Methyl acceptance Tertiles Mixed Pufulete No UK 28/76 Retrospective Yes Pretreatment Cancer screening Unaffected on screening 
72 Pufulete 5/2/2003 GASTROENTEROLOGY Colorectal adenoma 5meC Methyl acceptance Tertiles Mixed Pufulete No UK 35/76 Retrospective Yes Pretreatment Cancer screening Unaffected on screening 
31 Lim 1/1/2008 GASTROENTEROLOGY Colorectal 5meC HPLC and MS Tertiles Female Lim No US 115/115 Retrospective Prospective Pretreatment Cancer screening Unaffected on screening 
106 Moore 4/9/2008 LANCET ONCOL Bladder 5meC HPCE and hpaII* Quartiles Mixed Moore No Spain 775/397 Retrospective Yes Unreported Hospital based Hospital based 
73 Choi 7/7/2009 CARCINOGENESIS Breast 5meC Mass spectrometry Tertiles Female Choi No U.S. 176/173 Retrospective Yes Pretreatment Hospital-based Patients (unrelated conditions) 
99 Hou 10/15/2010 INT J CANCER Gastric Alu Pyrosequencing Tertiles Mixed Hou Caucasian Poland 302/421 Retrospective Yes Mixed Hospital based Cancer registry 
79 Zhu 3/22/2011 CANCER CAUSE CONTROL All (combined) Alu Pyrosequencing Quartiles Male Zhu No Boston 205/486 Retrospective No Posttreatment Baseline screening of prospective population cohort Unaffected at baseline and follow up 
97 Gao 1/31/2012 BRIT J CANCER Gastric Alu Pyrosequencing Quartiles Female Gao No Shanghai 192/381 Prospective Prospective Pretreatment Incident cases from prospective cohort Unaffected on follow up 
48 Wu 6/7/2012 CARCINOGENESIS Breast Alu MethyLight Quartiles Female Wu No US/Australia 266/334 Retrospective No Posttreatment High-risk families Unaffected sisters 
74 Xu 2/27/2012 FASEB J Breast LUMA CCGG* Quintiles Female Xu No US 1055/1101 Retrospective Yes Posttreatment Cancer registry, first primary breast cancer Randomly selected (digit dialing) 
101 Wilhelm 3/1/2010 CLIN CANCER RES Bladder LINE1 Pyrosequencing Deciles Mixed Wilhelm No US 285/465 Retrospective Yes Unreported Cancer registry (hospitals) From previous study 
88 Mirabello 3/17/2010 BMC MED GENET Testicular LINE1 Pyrosequencing Tertiles Male Mirabello No US 114/151 Retrospective No Posttreatment High-risk families Unaffected relatives 
99 Hou 10/15/2010 INT J CANCER Gastric LINE1 Pyrosequencing Tertiles Mixed Hou Caucasian Poland 300/419 Retrospective Yes Mixed Cancer registry (hospitals) Randomly selected (registry) 
79 Zhu 3/22/2011 CANCER CAUSE CONTROL All (combined) LINE1 Pyrosequencing Quartiles Male Zhu No Boston 30/487 Prospective Prospective Pretreatment Incident cases from prospective elderly male cohort Unaffected at case selection 
79 Zhu 3/22/2011 CANCER CAUSE CONTROL All (combined) LINE1 Pyrosequencing Quartiles Male Zhu No Boston 204/478 Retrospective No Posttreatment Baseline screening of prospective population cohort Unaffected at case-selection 
104 Rusiecki 4/3/2011 EPIGENOMICS Gastric LINE1 Pyrosequencing Tertiles Mixed Rusiecki No Oman 110/104 Retrospective Yes Unreported Hospital based Donors/government workers 
85 Liao 6/11/2011 PLOS ONE Renal cell LINE1 Pyrosequencing Quartiles Mixed Liao Caucasian Europe 328/654 Retrospective Yes Unreported Hospital based Patients (unrelated conditions) 
87 Di 10/12/2011 J ZHEJIANG UNIV-SC B Hepatocellular LINE1 Pyrosequencing Quartiles Mixed Di No China 315/356 Retrospective Yes Pretreatment Hospital based Unaffected at case selection 
97 Gao 1/31/2012 BRIT J CANCER Gastric LINE1 Pyrosequencing Quartiles Female Gao No Shanghai 192/382 Prospective Prospective Pretreatment Incident cases from prospective female cohort Unaffected on follow up 
74 Xu 2/27/2012 FASEB J Breast LINE1 Pyrosequencing Quintiles Female Xu No US 1064/1100 Retrospective Yes Posttreatment Cancer registry, first primary breast cancer Randomly selected (digit dialing) 
80 Cash 3/1/2012 INT J CANCER Bladder LINE1 Pyrosequencing Tertiles Mixed Cash Han Chinese China 510/528 Retrospective No Posttreatment Cancer registry Recruited from registry 
15 Brennan 5/1/2012 CANCER RES Breast LINE1 Pyrosequencing Quartiles Female Brennan White UK 252/255 Prospective Prospective Pretreatment Incident cases from prospective female cohort Unaffected on follow up 
15 Brennan 5/1/2012 CANCER RES Breast LINE1 Pyrosequencing Quartiles Female Brennan No Europe 263/257 Prospective Prospective Pretreatment Incident cases from prospective female cohort Unaffected on follow up 
89 Wu 5/12/2012 CARCINOGENESIS Hepatocellular LINE1 Pyrosequencing Quartiles Mixed Wu No Taiwan 305/1254 Prospective Prospective Pretreatment Incident cases from prospective cohort Unaffected on follow up 
47 Wu 6/7/2012 CARCINOGENESIS Breast LINE1 Pyrosequencing Quartiles Female Wu No US/Australia 279/340 Retrospective No Posttreatment High-risk families Unaffected sisters 
76 Hsiung 1/16/2007 CANCER EPIDEM BIOMAR Head and neck LRE1 COBRA* Tertiles Mixed Hsiung No Boston 278/526 Retrospective Yes Posttreatment Hospital based Recruited from registry 
89 Wu 5/12/2012 CARCINOGENESIS Hepatocellular Sat2 MethyLight Quartiles Mixed Wu No Taiwan 305/1254 Prospective Prospective Pretreatment Incident cases from prospective cohort Recruited from registry 
47 Wu 6/7/2012 CARCINOGENESIS Breast Sat2 MethyLight Quartiles Female Wu No US/Australia 266/333 Retrospective No Posttreatment High-risk families Unaffected sisters 
26 Cho 12/30/2010 ANTICANCER RES Breast Alu MethyLight NA Female Cho No Turkey 40/40 Retrospective Yes Pretreatment Hospital based Healthy volunteers 
69 Widschwendter 7/16/2008 PLOS ONE Breast Alu MethyLight NA Female Widschwendter No Germany 169/180 Retrospective Prospective Pretreatment Hospital based Unaffected on routine health check 
73 Choi 7/7/2009 CARCINOGENESIS Breast LINE1 Pyrosequencing NA Female Choi No US 19/18 Retrospective Yes Pretreatment Hospital based Unaffected relatives/visitors 
26 Cho 12/30/2010 ANTICANCER RES Breast LINE1 MethyLight NA Female Cho No Turkey 40/40 Retrospective Yes Pretreatment Hospital based Healthy volunteers 
98 Piyathilake 5/27/2011 NUTRITION CIN* LINE1 Pyrosequencing NA Female Piyathilake No US 103/273 Prospective Prospective Pretreatment Incident CIN cases from prospective cohort Women with 1 or less CIN 
83 Patchsung 5/15/2012 PLOS ONE Bladder LINE1 COBRA NA Mixed Patchsung No Thailand 61/45 Retrospective Yes Unreported Hospital based Healthy elderly volunteers 
82 Pobsook 11/12/2010 CLIN CHIM ACTA Oral LINE1 COBRA NA Mixed Pobsook No Thailand 34/43 Retrospective Yes Pretreatment Hospital based Volunteers 
26 Cho 12/30/2010 ANTICANCER RES Breast Sat2 MethyLight NA Female Cho No Turkey 40/40 Retrospective Yes Pretreatment Hospital based Healthy volunteers 
ReferenceAuthorPublication dateJournalCancer typeMeasurementAssayCategorical split*sexFirst authorReported restricted ethnicityCountry/continentn (case/ctrl)Prospective/retrospectiveNewly diagnosedPre/postchemotherapyCases (description)Controls (description)
72 Pufulete 5/2/2003 GASTROENTEROLOGY Colorectal 5meC Methyl acceptance Tertiles Mixed Pufulete No UK 28/76 Retrospective Yes Pretreatment Cancer screening Unaffected on screening 
72 Pufulete 5/2/2003 GASTROENTEROLOGY Colorectal adenoma 5meC Methyl acceptance Tertiles Mixed Pufulete No UK 35/76 Retrospective Yes Pretreatment Cancer screening Unaffected on screening 
31 Lim 1/1/2008 GASTROENTEROLOGY Colorectal 5meC HPLC and MS Tertiles Female Lim No US 115/115 Retrospective Prospective Pretreatment Cancer screening Unaffected on screening 
106 Moore 4/9/2008 LANCET ONCOL Bladder 5meC HPCE and hpaII* Quartiles Mixed Moore No Spain 775/397 Retrospective Yes Unreported Hospital based Hospital based 
73 Choi 7/7/2009 CARCINOGENESIS Breast 5meC Mass spectrometry Tertiles Female Choi No U.S. 176/173 Retrospective Yes Pretreatment Hospital-based Patients (unrelated conditions) 
99 Hou 10/15/2010 INT J CANCER Gastric Alu Pyrosequencing Tertiles Mixed Hou Caucasian Poland 302/421 Retrospective Yes Mixed Hospital based Cancer registry 
79 Zhu 3/22/2011 CANCER CAUSE CONTROL All (combined) Alu Pyrosequencing Quartiles Male Zhu No Boston 205/486 Retrospective No Posttreatment Baseline screening of prospective population cohort Unaffected at baseline and follow up 
97 Gao 1/31/2012 BRIT J CANCER Gastric Alu Pyrosequencing Quartiles Female Gao No Shanghai 192/381 Prospective Prospective Pretreatment Incident cases from prospective cohort Unaffected on follow up 
48 Wu 6/7/2012 CARCINOGENESIS Breast Alu MethyLight Quartiles Female Wu No US/Australia 266/334 Retrospective No Posttreatment High-risk families Unaffected sisters 
74 Xu 2/27/2012 FASEB J Breast LUMA CCGG* Quintiles Female Xu No US 1055/1101 Retrospective Yes Posttreatment Cancer registry, first primary breast cancer Randomly selected (digit dialing) 
101 Wilhelm 3/1/2010 CLIN CANCER RES Bladder LINE1 Pyrosequencing Deciles Mixed Wilhelm No US 285/465 Retrospective Yes Unreported Cancer registry (hospitals) From previous study 
88 Mirabello 3/17/2010 BMC MED GENET Testicular LINE1 Pyrosequencing Tertiles Male Mirabello No US 114/151 Retrospective No Posttreatment High-risk families Unaffected relatives 
99 Hou 10/15/2010 INT J CANCER Gastric LINE1 Pyrosequencing Tertiles Mixed Hou Caucasian Poland 300/419 Retrospective Yes Mixed Cancer registry (hospitals) Randomly selected (registry) 
79 Zhu 3/22/2011 CANCER CAUSE CONTROL All (combined) LINE1 Pyrosequencing Quartiles Male Zhu No Boston 30/487 Prospective Prospective Pretreatment Incident cases from prospective elderly male cohort Unaffected at case selection 
79 Zhu 3/22/2011 CANCER CAUSE CONTROL All (combined) LINE1 Pyrosequencing Quartiles Male Zhu No Boston 204/478 Retrospective No Posttreatment Baseline screening of prospective population cohort Unaffected at case-selection 
104 Rusiecki 4/3/2011 EPIGENOMICS Gastric LINE1 Pyrosequencing Tertiles Mixed Rusiecki No Oman 110/104 Retrospective Yes Unreported Hospital based Donors/government workers 
85 Liao 6/11/2011 PLOS ONE Renal cell LINE1 Pyrosequencing Quartiles Mixed Liao Caucasian Europe 328/654 Retrospective Yes Unreported Hospital based Patients (unrelated conditions) 
87 Di 10/12/2011 J ZHEJIANG UNIV-SC B Hepatocellular LINE1 Pyrosequencing Quartiles Mixed Di No China 315/356 Retrospective Yes Pretreatment Hospital based Unaffected at case selection 
97 Gao 1/31/2012 BRIT J CANCER Gastric LINE1 Pyrosequencing Quartiles Female Gao No Shanghai 192/382 Prospective Prospective Pretreatment Incident cases from prospective female cohort Unaffected on follow up 
74 Xu 2/27/2012 FASEB J Breast LINE1 Pyrosequencing Quintiles Female Xu No US 1064/1100 Retrospective Yes Posttreatment Cancer registry, first primary breast cancer Randomly selected (digit dialing) 
80 Cash 3/1/2012 INT J CANCER Bladder LINE1 Pyrosequencing Tertiles Mixed Cash Han Chinese China 510/528 Retrospective No Posttreatment Cancer registry Recruited from registry 
15 Brennan 5/1/2012 CANCER RES Breast LINE1 Pyrosequencing Quartiles Female Brennan White UK 252/255 Prospective Prospective Pretreatment Incident cases from prospective female cohort Unaffected on follow up 
15 Brennan 5/1/2012 CANCER RES Breast LINE1 Pyrosequencing Quartiles Female Brennan No Europe 263/257 Prospective Prospective Pretreatment Incident cases from prospective female cohort Unaffected on follow up 
89 Wu 5/12/2012 CARCINOGENESIS Hepatocellular LINE1 Pyrosequencing Quartiles Mixed Wu No Taiwan 305/1254 Prospective Prospective Pretreatment Incident cases from prospective cohort Unaffected on follow up 
47 Wu 6/7/2012 CARCINOGENESIS Breast LINE1 Pyrosequencing Quartiles Female Wu No US/Australia 279/340 Retrospective No Posttreatment High-risk families Unaffected sisters 
76 Hsiung 1/16/2007 CANCER EPIDEM BIOMAR Head and neck LRE1 COBRA* Tertiles Mixed Hsiung No Boston 278/526 Retrospective Yes Posttreatment Hospital based Recruited from registry 
89 Wu 5/12/2012 CARCINOGENESIS Hepatocellular Sat2 MethyLight Quartiles Mixed Wu No Taiwan 305/1254 Prospective Prospective Pretreatment Incident cases from prospective cohort Recruited from registry 
47 Wu 6/7/2012 CARCINOGENESIS Breast Sat2 MethyLight Quartiles Female Wu No US/Australia 266/333 Retrospective No Posttreatment High-risk families Unaffected sisters 
26 Cho 12/30/2010 ANTICANCER RES Breast Alu MethyLight NA Female Cho No Turkey 40/40 Retrospective Yes Pretreatment Hospital based Healthy volunteers 
69 Widschwendter 7/16/2008 PLOS ONE Breast Alu MethyLight NA Female Widschwendter No Germany 169/180 Retrospective Prospective Pretreatment Hospital based Unaffected on routine health check 
73 Choi 7/7/2009 CARCINOGENESIS Breast LINE1 Pyrosequencing NA Female Choi No US 19/18 Retrospective Yes Pretreatment Hospital based Unaffected relatives/visitors 
26 Cho 12/30/2010 ANTICANCER RES Breast LINE1 MethyLight NA Female Cho No Turkey 40/40 Retrospective Yes Pretreatment Hospital based Healthy volunteers 
98 Piyathilake 5/27/2011 NUTRITION CIN* LINE1 Pyrosequencing NA Female Piyathilake No US 103/273 Prospective Prospective Pretreatment Incident CIN cases from prospective cohort Women with 1 or less CIN 
83 Patchsung 5/15/2012 PLOS ONE Bladder LINE1 COBRA NA Mixed Patchsung No Thailand 61/45 Retrospective Yes Unreported Hospital based Healthy elderly volunteers 
82 Pobsook 11/12/2010 CLIN CHIM ACTA Oral LINE1 COBRA NA Mixed Pobsook No Thailand 34/43 Retrospective Yes Pretreatment Hospital based Volunteers 
26 Cho 12/30/2010 ANTICANCER RES Breast Sat2 MethyLight NA Female Cho No Turkey 40/40 Retrospective Yes Pretreatment Hospital based Healthy volunteers 

NOTE: Categorical split represents the level of categorization of DNA methylation used in the study. Country/continent refers to the place of sample collection. Prospective/retrospective and pre/postchemotherapy indicate whether blood samples were collected specifies whether blood samples were collected before or after diagnosis and neo-adjuvant chemotherapy, respectively. “Newly diagnosed” indicates whether retrospectively collected samples were collected from patients reported as “newly” or “recently” diagnosed. Cases and controls give a brief indication of the way in which case and control subjects were recruited and defined. “Reported restricted ethnicity” indicates whether publications for each study reported exclusion of study participants based on race. “No” indicates that no race exclusion was reported. Other values indicate the race to which the sample group was restricted.

There is no overall association between methylation at L1 and cancer risk according to this analysis. Whereas early studies suggested an association of blood L1 hypomethylation with cancer, later studies have failed to reproduce this. Furthermore, the direction of effect on cancer prevalence of L1 hypomethylation was inconsistent between significant studies (85), suggesting that this association may have occurred because of chance. One consideration is that heterogeneity between cancer types may affect the overall findings from this meta-analysis. However, in 3 breast cancer studies there is no association of L1 methylation with either cancer incidence (64) or prevalence (73, 84). Of 5 prospective studies, only 1 (79) identified an association between low L1 methylation and cancer incidence, however, small sample size, combining of multiple cancer types, and selection of a very high-risk population (elderly males) may be confounding factors for this study. Four more recent, larger studies have failed to identify any such association in gastric (90), breast (64), or hepatocellular (89) cancer, suggesting that L1 methylation is not a cancer risk factor.

Study design

Many retrospective studies investigating blood DNA methylation in cancer patients inappropriately use the term “risk” to describe associations between methylation and disease. Association with disease risk of blood DNA methylation variability can only be determined using prospective studies, with blood samples collected several years before disease development, because of the possibility of “reverse causality,” that is, the alteration of blood DNA methylation by presence of active disease (15, 68, 89). Whereas the effects of active cancer on blood DNA methylation are unknown, proliferation of lymphocytes, or depletion of required substrates or enzymes may affect blood DNA methylation (91). In addition, blood methylation may be affected by the presence of circulating tumor cells (92), though the contribution to overall blood methylation of this small fraction of tumor cells may not be detectable. An important factor for prospective studies is latency, that is, the duration between sample collection and cancer diagnosis (15), as methylation may be affected by as yet undetected cancer in samples collected shortly before diagnosis. Furthermore, the degree to which latency affects methylation-risk association may help determine whether methylation variability confers long-term or transient risk, or whether methylation is likely to be an early tumorigenic event. Retrospective studies are useful for investigating the relationship between DNA methylation variability and cancer prevalence, and may be used for development of diagnostic cancer biomarkers (16, 68). A potential confounding factor in some retrospective studies is the possible effect of cancer treatment, as several commonly used neo-adjuvant chemotherapeutic drugs are known affect inhibit activity of the folate-mediated 1-carbon metabolism pathway (93, 94). Whereas this treatment-induced inhibition of DNA methylation could potentially explain the occurrence of genome-wide hypomethylation in posttreatment blood samples, 4 studies identifying significant associations between genomic hypomethylation and cancer prevalence, including 1 L1 study (87), and 3 5meC studies (31, 72, 73) used pretreatment sample only. Nonetheless, the exposure of cases, but not controls to cancer treatment may represent a confounding factor. Sample size is an important factor for case–control studies, and a confounding factor when comparing different population-based case control studies. Studies of small sample size (n < 40 cases; refs. 72, 79), may have little power to detect subtle methylation variability, perhaps leading to a bias toward nonsignificance, or to chance detection (16). Most of the studies relevant to this report were of small sample size. Conversely, larger sample sizes may lead to detection of statistically significant results with a very small overall effect size (95), which can be misleading. The largest study to date included 1,000 cases/control pairs, at alpha = 0.05 this study would have 80% power to detect a difference of 0.12 standard deviations. For the L1 assay, with a standard deviation of 1.8% across the population, this would equal a difference in means of 0.06%, which is well below the technical variability in this assay [0.5%–5% (78, 88, 89, 96–99)] and may be biologically meaningless.

Sample selection

Various patient characteristics are potential modifiers of both cancer risk and DNA methylation. Such factors may represent intermediate factors in the link between DNA methylation and disease risk, but may also represent potential confounding factors within the study. For example, where DNA methylation is affected by smoking, an association between DNA methylation and cancer risk may be apparent because of uneven numbers of smokers in the case and control groups. To reduce the effect of such potential confounders, case–control pairs should be matched on potential confounding patient characteristics, as well as technical factors such as experimental batch (16). Furthermore, statistical adjustment for these factors should be applied during analysis. Age is a strong modifier of DNA methylation for some genes (12, 100), and is the biggest risk factor for most cancers, therefore, adjustment for age is essential (12, 100). L1 methylation is significantly lower in females than males (80, 101, 102), probably because of differences in the X and Y chromosomes (103), therefore, population studies investigating genome-wide DNA methylation must be stratified by gender (79, 80, 101, 104). Whereas gender may represent a confounding factor for some of the included studies (72, 85), statistical adjustment may have helped to reduce this bias. Both L1 methylation and cancer risk are modified by ethnicity and environmental carcinogen exposures (55, 80, 99, 105). Hospital-based studies (106) may include controls confounded by conditions unrelated to the disease under investigation. Ideally, controls samples should represent healthy individuals, matched on all potential confounding factors, both technical and biological. Case–control studies nested within prospective cohorts remain the gold standard for this type of analysis (16, 107).

Assay measurement

Repetitive element surrogate assays provide a practical and cost-effective indicator of genomic hypomethylation in tumor tissues and cell lines, where L1 methylation is highly variable (108). However, in blood, L1 methylation displays little variability (109), and it remains unclear how sensitive such assays are to detect subtle DNA methylation variability. Methylation of TEs, particularly L1, is often reported as genome-wide or “global” DNA methylation. However, L1 pyrosequencing measures methylation at only 3 to 4 CpG sites within a pool of L1 elements based on a consensus sequence (44), and is, therefore, not representative of genome-wide methylation (14). The detection of genome-wide hypomethylation by repetitive-element assays in tumor DNA is likely due the occurrence of a proportion of repetitive elements within hypomethylated domains, rather than to a specific enrichment at repetitive elements (17, 24). Key to the utility of repetitive element methylation as a surrogate for genome-wide methylation is the reported correlation between 5meC measured by HPLC, and methylation of L1 and Alu measured by MethyLight, a methylation-sensitive PCR-based system, in a panel of cell lines and tissues. However, these correlations were moderate (r = 0.66 and 0.70 for L1 and Alu, respectively), and the authors recommended using a composite measure of 2 repetitive elements, ALU-M2 and Sat2-M1 (r = 0.85), rather than individual elements, as a surrogate for genome-wide methylation. The only study to measure the correlation between L1 pyrosequencing and 5meC in blood, or indeed any tissue, failed to find any correlation (r = −0.204), though the sample size was small (n = 27; ref. 73). Whereas L1 methylation measured by pyrosequencing is reduced in cell lines exposed to DNA methylation inhibitors (44), and in many tumors, it is unclear whether L1 methylation reflects subtle genomic methylation variability in blood. This discrepancy may explain the stronger association with cancer risk of blood 5meC than blood L1 methylation.

Two prospective cancer risk studies detected blood hypomethylation at Alu (97) and Sat2 (89), but not at L1 in the same samples, and methylation of L1 and Alu are not correlated (99). This is inconsistent with these assays detecting “genome-wide methylation,” but suggests that hypomethylation may be restricted to specific genomic sequences. Widely apparent is the greater variability of repetitive element methylation between different studies than between cases and controls within individual studies (14). L1 methylation varies with ethnicity (55, 105), likely because of L1 genetic variability (14, 90). L1 sequence heterogeneity may also cause cellular and allelic heterogeneity in L1 methylation (110), and may to lead to underestimation of methylation levels (77). L1 pyrosequencing, however, displays little technical variation (108), and it will be important to determine whether measures of 5meC are as technically reproducible. L1 elements are differentially methylated at different genomic loci (50, 103, 111), and at different CpG loci within the L1 consensus sequence (98), therefore, assays preferentially amplifying different L1 elements or measuring methylation at different CpG sites may produce inconsistent results.

Finally, methylation variability at individual “functional” repetitive elements may provide greater biomarker potential than methylation of “global” repetitive elements, as hypomethylation of an aberrantly transcribed L1 within the MET oncogene, but not global L1 methylation, was detected in normal bladder in patients with bladder cancer (50).

Statistical analysis

Categorical analysis of DNA methylation using OR is the most frequently used measure of disease risk for case control epigenetic epidemiology studies (16). Furthermore, categorical analysis is useful for investigating potentially nonlinear relationships between methylation and disease risk and retains more power to detect significant differences compared with linear regression. However, different approaches may introduce bias, for instance, a prominent inconsistency between studies included in this report is the use of different levels of methylation categories, including tertile, quartile, quintiles, and deciles, when a prespecified analysis plan and statistical power should have been described and used for the study. The narrow ranges of DNA methylation reported, especially at L1, would mean that the difference in percentage methylation between categories would be far below the detection sensitivity of the assay, for instance the technical variation for pyrosequencing, one of the most quantitative assays currently available, is around 2% and 3% (64). In our study, we observed an intraclass correlation coefficient for LINE1 in blinded duplicate samples of 0 (95% CI, 0–0.61) that suggests higher within individual variability than between individual variability. A major problem with developing small methylation differences as biomarkers is the overlap between methylation of cases and controls or lack of specificity (14).

Publication factors

According to the STROBE-ME (Strengthening the reporting of observational studies in epidemiology-molecular epidemiology) guidelines, all basic statistics of a biomarker measure distribution (mean, median, range, and variance) and details of all other analyses should be reported in studies investigating molecular biomarkers for disease risk (112); however, many reports do not include these, making direct comparison between studies difficult. Publication bias frequently affects meta-analyses of observational studies (113), and there is some evidence that it affects the current analysis.

All of the included reports showing significant association between L1 methylation and cancer included investigations of L1 methylation only, however all reports showing negative results for L1 also included studies showing significant associations between another methylation marker and cancer incidence/prevalence, consistent with a bias toward publication of significant associations. There appears to be a trend toward lower effect size with later publication date, as OR is significantly correlated with publication year among L1 methylation studies (spearman r = 0.49, P = 0.05). This appears to be independent of sample size, as the correlations between OR and sample size (r = −0.19), and sample size and year of publication (r = 0.2), are not significant. A funnel plot including L1 studies only (Supplementary Fig. S1) does not appear to show asymmetry or publication bias; however, the ability of this plot to indicate publication bias may be limited by the small number of studies (n = 14), and variable direction of effect in studies showing significant associations. A reporting bias exists whereby categorical analysis of methylation-risk associations tends only to be reported if it reveals statistically significant results, and if the authors deem categorical analysis appropriate. The inclusion of categorical analysis of our previously reported L1 studies represents an attempt to address this bias; however, this could only be estimated for other studies that reported no association with cancer risk (Supplementary Fig. S2). As expected, inclusion of our studies, and other recent null studies, attenuated the significant association between L1 methylation and cancer risk reported by a recent meta-analysis (86). Inclusion of all relevant reports is critical to the accuracy of meta-analyses (95), therefore, publication and reporting biases may affect our ability to identify the true effect size.

A concern for all studies investigating blood DNA methylation in relation to cancer in other tissues is the tissue specificity of DNA methylation (91, 100). Several studies have indicated that methylation of repetitive elements is tissue specific, most variable in tumor tissue, and not correlated between tumor and blood (81, 98, 114). Consistently, evidence suggests that genomic hypomethylation in tumor and normal adjacent tissue of bladder and colon cancer was not detectable in blood (30, 50), suggesting that hypomethylation is restricted to the disease-affected tissue. Furthermore, approximately 36% of L1 elements display tissue-specific expression (37). Methylation variability between blood cell types is an important consideration for retrospective studies, as differential methylation may reflect cellular proliferation. A recent microarray-based study found that approximately 39% of CpG sites analyzed were differentially methylated between leukocyte subgroups (91). However, a recent report did not find any association between L1 methylation and blood cell count (115). Another potential caveat for all cancer biomarker studies is tumor subtype heterogeneity, where the biomarker under investigation is a feature of only 1 disease subtype. For example, in tumor tissue, L1 hypomethylation frequently occurs in cancers displaying chromosome instability, but rarely in cancers with microsatellite instability (20, 36, 116). Consistently, L1 hypomethylation is detectable in some colon tumor samples, but not others (20). Therefore, if blood L1 hypomethylation is associated with colorectal cancer risk, it may predict risk of some, but not all cancers.

Epigenetic epidemiology holds great promise for identifying biomarkers of cancer risk and understanding cancer etiology. It is important to investigate methylation variability at individual's loci in the context of widespread methylation changes associated with cancer; therefore, understanding the contribution of genome-wide DNA methylation to cancer susceptibility and development is needed. The challenges of investigating a genome-wide epigenetic phenomenon in population studies are formidable, because of the requirements of high genome coverage, highly quantitative DNA methylation measurement, and large sample size (16, 45). We conclude from our meta-analysis that genome-wide DNA methylation, as measured by the surrogate L1 methylation assay, is not associated with cancer risk, shown most appropriately by several prospective cohort studies. However, there appears evidence for an association of genomic hypomethylation, measured using more representative assays, such as HPLC or methyl-acceptance assays, with cancer prevalence, though this remains to be validated in prospective studies.

A thorough investigation of the relationship between genome-wide DNA methylation and cancer risk would require whole-genome bisulphite sequencing of large cohorts of prospectively collected blood samples, with careful sample selection, case–control matching, and deep sequencing to achieve the sensitivity required to detect subtle DNA methylation variability (16, 45). Such an approach would provide information about genome-scale methylation changes as well as local methylation variability, thus providing candidate regions for risk biomarker development. An equivalent study using retrospectively collected blood samples may provide candidate diagnostic biomarkers. Whole-genome bisulphite sequencing is extremely costly, however, technological improvement will make this feasible in the coming years (45, 75). Whereas current research using whole-genome sequencing of small numbers of samples cannot provide information about methylation variability across populations, further classification of the regions that consistently undergo methylation changes in cancer may provide candidate regions for a sequence capture approach in population studies. New primer technology will enable the investigation of epigenetic regulation at specific “functional” repetitive elements, such as those L1 elements capable of retrotransposition (47, 50), or driving ectopic expression of neighboring genes (37, 50). Furthermore, availability of the first microarrays specifically covering repetitive elements will provide high-throughput methods of investigating repetitive element methylation on a genome scale (117). The most popular tool for current investigation of DNA methylation variability in population studies is the Illumina 450K methylation beadchip, which measures methylation at individual CpG sites across the entire human genome. These arrays are highly quantitative, high-throughput, and largely unbiased in terms of genomic coverage, making them suitable for biomarker discovery and epigenome-wide association studies (16). Use of these assays will likely provide interesting evidence for the implications of DNA methylation variability in disease susceptibility and development in coming years.

No potential conflicts of interest were disclosed.

Conception and design: K. Brennan, J.M. Flanagan

Development of methodology: K. Brennan

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): K. Brennan, J.M. Flanagan

Writing, review, and/or revision of the manuscript: K. Brennan, J.M. Flanagan

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K. Brennan

Study supervision: J.M. Flanagan

The authors thank Prof. Robert Brown for his critical review of this manuscript.

This work was funded by Breast Cancer Campaign fellowship to J.M. Flanagan. K. Brennan and J.M. Flanagan are funded by Breast Cancer Campaign.

1.
Feinberg
AP
,
Tycko
B
. 
The history of cancer epigenetics
.
Nat Rev Cancer
2004
;
4
:
143
53
.
2.
Laird
PW
. 
Cancer epigenetics
.
Hum Mol Genet
2005
;
14
Spec No 1
:
R65
76
.
3.
Jurkowska
RZ
,
Jurkowski
TP
,
Jeltsch
A
. 
Structure and function of mammalian DNA methyltransferases
.
Chembiochem
2011
;
12
:
206
22
.
4.
Vineis
P
,
Chuang
SC
,
Vaissière
T
,
Cuenin
C
,
Ricceri
F
Genair-EPIC Collaborators
et al
DNA methylation changes associated with cancer risk factors and blood levels of vitamin metabolites in a prospective study
.
Epigenetics
2011
;
6
:
195
201
.
5.
Schernhammer
ES
,
Giovannucci
E
,
Kawasaki
T
,
Rosner
B
,
Fuchs
CS
,
Ogino
S
, et al
Dietary folate, alcohol and B vitamins in relation to LINE-1 hypomethylation in colon cancer
.
Gut
2010
;
59
:
794
9
.
6.
Deaton
AM
,
Bird
A
. 
CpG islands and the regulation of transcription
.
Genes Dev
2011
;
25
:
1010
22
.
7.
Appanah
R
,
Dickerson
DR
,
Goyal
P
,
Groudine
M
,
Lorincz
MC
. 
An unmethylated 3′ promoter-proximal region is required for efficient transcription initiation
.
PLoS Genet
2007
;
3
:
e27
.
8.
Bird
A
. 
CpG-rich islands and the function of DNA methylation
.
Nature
1986
;
321
:
209
13
.
9.
Pujadas
E
,
Feinberg
AP
. 
Regulated noise in the epigenetic landscape of development and disease
.
Cell
2012
;
148
:
1123
31
.
10.
Muñoz-Lopez
M
,
Macia
A
,
Garcia-Cañadas
M
,
Badge
RM
,
Garcia-Perez
JL
. 
An epi [c] genetic battle: LINE-1 retrotransposons and intragenomic conflict in humans
.
Mob Genet Elements
2011
;
1
:
122
7
.
11.
Wilson
AS
,
Power
BE
,
Molloy
PL
. 
DNA hypomethylation and human diseases
.
Biochim Biophys Acta
2007
;
1775
:
138
62
.
12.
Rakyan
VK
,
Down
TA
,
Maslau
S
,
Andrew
T
,
Yang
TP
,
Beyan
H
, et al
Human aging-associated DNA hypermethylation occurs preferentially at bivalent chromatin domains
.
Genome Res
2010
;
20
:
434
9
.
13.
Christensen
BC
,
Marsit
CJ
. 
Epigenomics in environmental health
.
Front Genet
2011
;
2
:
84
.
14.
Nelson
HH
,
Marsit
CJ
,
Kelsey
KT
. 
Global methylation in exposure biology and translational medical science
.
Environ Health Perspect
2011
;
119
:
1528
33
.
15.
Brennan
K
,
Flanagan
JM
. 
Epigenetic epidemiology for cancer risk: harnessing germline epigenetic variation
.
Methods Mol Biol
2012
;
863
:
439
65
.
16.
Rakyan
VK
,
Down
TA
,
Balding
DJ
,
Beck
S
. 
Epigenome-wide association studies for common human diseases
.
Nat Rev Genet
2011
;
12
:
529
41
.
17.
Hansen
KD
,
Timp
W
,
Bravo
HC
,
Sabunciyan
S
,
Langmead
B
,
McDonald
OG
, et al
Increased methylation variation in epigenetic domains across cancer types
.
Nat Genet
2011
;
43
:
768
75
.
18.
Feinberg
AP
,
Vogelstein
B
. 
Hypomethylation distinguishes genes of some human cancers from their normal counterparts
.
Nature
1983
;
301
:
89
92
.
19.
Lapeyre
JN
,
Walker
MS
,
Becker
FF
. 
DNA methylation and methylase levels in normal and malignant mouse hepatic tissues
.
Carcinogenesis
1981
;
2
:
873
8
.
20.
Estecio
MR
,
Gharibyan
V
,
Shen
L
,
Ibrahim
AE
,
Doshi
K
,
He
R
, et al
LINE-1 hypomethylation in cancer is highly variable and inversely correlated with microsatellite instability
.
PLoS One
2007
;
2
:
e399
.
21.
Wild
L
,
Flanagan
JM
. 
Genome-wide hypomethylation in cancer may be a passive consequence of transformation
.
Biochim Biophys Acta
2010
;
1806
:
50
7
.
22.
Lister
R
,
Pelizzola
M
,
Dowen
RH
,
Hawkins
RD
,
Hon
G
,
Tonti-Filippini
J
, et al
Human DNA methylomes at base resolution show widespread epigenomic differences
.
Nature
2009
;
462
:
315
22
.
23.
Ruike
Y
,
Imanaka
Y
,
Sato
F
,
Shimizu
K
,
Tsujimoto
G
. 
Genome-wide analysis of aberrant methylation in human breast cancer cells using methyl-DNA immunoprecipitation combined with high-throughput sequencing
.
BMC Genomics
2010
;
11
:
137
.
24.
Hon
GC
,
Hawkins
RD
,
Caballero
OL
,
Lo
C
,
Lister
R
,
Pelizzola
M
, et al
Global DNA hypomethylation coupled to repressive chromatin domain formation and gene silencing in breast cancer
.
Genome Res
2012
;
22
:
246
58
.
25.
Friso
S
,
Choi
SW
,
Girelli
D
,
Mason
JB
,
Dolnikowski
GG
,
Bagley
PJ
, et al
A common mutation in the 5,10-methylenetetrahydrofolate reductase gene affects genomic DNA methylation through an interaction with folate status
.
Proc Natl Acad Sci U S A
2002
;
99
:
5606
11
.
26.
Cho
YH
,
Yazici
H
,
Wu
HC
,
Terry
MB
,
Gonzalez
K
,
Qu
M
, et al
Aberrant promoter hypermethylation and genomic hypomethylation in tumor, adjacent normal tissues and blood from breast cancer patients
.
Anticancer Res
2010
;
30
:
2489
96
.
27.
Feinberg
AP
. 
Phenotypic plasticity and the epigenetics of human disease
.
Nature
2007
;
447
:
433
40
.
28.
Yang
P
,
Ma
J
,
Zhang
B
,
Duan
H
,
He
Z
,
Zeng
J
, et al
CpG site-specific hypermethylation of p16INK4alpha in peripheral blood lymphocytes of PAH-exposed workers
.
Cancer Epidemiol Biomarkers Prev
2012
;
21
:
182
90
.
29.
Teschendorff
AE
,
Menon
U
,
Gentry-Maharaj
A
,
Ramus
SJ
,
Weisenberger
DJ
. 
Age-dependent DNA methylation of genes that are suppressed in stem cells is a hallmark of cancer
.
Genome Res
2010
;
20
:
440
6
.
30.
Suter
CM
,
Martin
DI
,
Ward
RL
. 
Hypomethylation of L1 retrotransposons in colorectal cancer and adjacent normal tissue
.
Int J Colorectal Dis
2004
;
19
:
95
101
.
31.
Lim
U
,
Flood
A
,
Choi
SW
,
Albanes
D
,
Cross
AJ
,
Schatzkin
A
, et al
Genomic methylation of leukocyte DNA in relation to colorectal adenoma among asymptomatic women
.
Gastroenterology
2008
;
134
:
47
55
.
32.
Worthley
DL
,
Whitehall
VL
,
Buttenshaw
RL
,
Irahara
N
,
Greco
SA
,
Ramsnes
I
, et al
DNA methylation within the normal colorectal mucosa is associated with pathway-specific predisposition to cancer
.
Oncogene
2010
;
29
:
1653
62
.
33.
Sunami
E
,
de Maat
M
,
Vu
A
,
Turner
RR
,
Hoon
DS
. 
LINE-1 hypomethylation during primary colon cancer progression
.
PLoS One
2011
;
6
:
e18884
.
34.
Ibrahim
AE
,
Arends
MJ
,
Silva
AL
,
Wyllie
AH
,
Greger
L
,
Ito
Y
, et al
Sequential DNA methylation changes are associated with DNMT3B overexpression in colorectal neoplastic progression
.
Gut
2011
;
60
:
499
508
.
35.
Kamiyama
H
,
Suzuki
K
,
Maeda
T
,
Koizumi
K
,
Miyaki
Y
,
Okada
S
, et al
DNA demethylation in normal colon tissue predicts predisposition to multiple cancers
.
Oncogene
. 
2012
.
[Epub ahead of print]
.
36.
Matsuzaki
K
,
Deng
G
,
Tanaka
H
,
Kakar
S
,
Miura
S
,
Kim
YS
. 
The relationship between global methylation level, loss of heterozygosity, and microsatellite instability in sporadic colorectal cancer
.
Clin Cancer Res
2005
;
11
(
24 Pt 1
):
8564
9
.
37.
Faulkner
GJ
,
Kimura
Y
,
Daub
CO
,
Wani
S
,
Plessy
C
,
Irvine
KM
, et al
The regulated retrotransposon transcriptome of mammalian cells
.
Nat Genet
2009
;
41
:
563
71
.
38.
Slotkin
RK
,
Martienssen
R
. 
Transposable elements and the epigenetic regulation of the genome
.
Nat Rev Genet
2007
;
8
:
272
85
.
39.
Aporntewan
C
,
Phokaew
C
,
Piriyapongsa
J
,
Ngamphiw
C
,
Ittiwut
C
,
Tongsima
S
, et al
Hypomethylation of intragenic LINE-1 represses transcription in cancer cells through AGO2
.
PLoS One
2011
;
6
:
e17934
.
40.
Rangwala
SH
,
Zhang
L
,
Kazazian
HH
 Jr
. 
Many LINE1 elements contribute to the transcriptome of human somatic cells
.
Genome Biol
2009
;
10
:
R100
.
41.
van Hoesel
AQ
,
van de Velde
CJ
,
Kuppen
PJ
,
Liefers
GJ
,
Putter
H
,
Sato
Y
, et al
Hypomethylation of LINE-1 in primary tumor has poor prognosis in young breast cancer patients: a retrospective cohort study
.
Breast Cancer Res Treat
2012
;
134
:
1103
14
.
42.
Flatley
JE
,
McNeir
K
,
Balasubramani
L
,
Tidy
J
,
Stuart
EL
,
Young
TA
, et al
Folate status and aberrant DNA methylation are associated with HPV infection and cervical pathogenesis
.
Cancer Epidemiol Biomarkers Prev
2009
;
18
:
2782
9
.
43.
Gaudet
F
,
Hodgson
JG
,
Eden
A
,
Jackson-Grusby
L
,
Dausman
J
,
Gray
JW
, et al
Induction of tumors in mice by genomic hypomethylation
.
Science
2003
;
300
:
489
92
.
44.
Yang
AS
,
Estécio
MR
,
Doshi
K
,
Kondo
Y
,
Tajara
EH
,
Issa
JP
. 
A simple method for estimating global DNA methylation using bisulfite PCR of repetitive DNA elements
.
Nucleic Acids Res
2004
;
32
:
e38
.
45.
Laird
PW
. 
Principles and challenges of genome-wide DNA methylation analysis
.
Nat Rev Genet
2010
;
11
:
191
203
.
46.
Beck
CR
,
Collier
P
,
Macfarlane
C
,
Malig
M
,
Kidd
JM
,
Eichler
EE
, et al
LINE-1 retrotransposition activity in human genomes
.
Cell
2010
;
141
:
1159
70
.
47.
Cordaux
R
,
Batzer
MA
. 
The impact of retrotransposons on human genome evolution
.
Nat Rev Genet
2009
;
10
:
691
703
.
48.
Wu
HC
,
Delgado-Cruzata
L
,
Flom
JD
,
Perrin
M
,
Liao
Y
,
Ferris
JS
, et al
Repetitive element DNA methylation levels in white blood cell DNA from sisters discordant for breast cancer from the New York site of the BCFR
.
Carcinogenesis
2012
;
33
:
1946
52
.
49.
Figueiredo
JC
,
Grau
MV
,
Wallace
K
,
Levine
AJ
,
Shen
L
,
Hamdan
R
, et al
Global DNA hypomethylation (LINE-1) in the normal colon and lifestyle characteristics and dietary and genetic factors
.
Cancer Epidemiol Biomarkers Prev
2009
;
18
:
1041
9
.
50.
Wolff
EM
,
Byun
HM
,
Han
HF
,
Sharma
S
,
Nichols
PW
,
Siegmund
KD
, et al
Hypomethylation of a LINE-1 promoter activates an alternate transcript of the MET oncogene in bladders with cancer
.
PLoS Genet
2010
;
6
:
e1000917
.
51.
Flanagan
JM
,
Munoz-Alegre
M
,
Henderson
S
,
Tang
T
,
Sun
P
,
Johnson
N
, et al
Gene-body hypermethylation of ATM in peripheral blood DNA of bilateral breast cancer patients
.
Hum Mol Genet
2009
;
18
:
1332
42
.
52.
Martin
DI
,
Ward
R
,
Suter
CM
. 
Germline epimutation: a basis for epigenetic disease in humans
.
Ann N Y Acad Sci
2005
;
1054
:
68
77
.
53.
Foley
DL
,
Craig
JM
,
Morley
R
,
Olsson
CA
,
Dwyer
T
,
Smith
K
, et al
Prospects for epigenetic epidemiology
.
Am J Epidemiol
2009
;
169
:
389
400
.
54.
Stepanow
S
,
Reichwald
K
,
Huse
K
,
Gausmann
U
,
Nebel
A
,
Rosenstiel
P
, et al
Allele-specific, age-dependent and BMI-associated DNA methylation of human MCHR1
.
PLoS One
2011
;
6
:
e17711
.
55.
Zhang
FF
,
Cardarelli
R
,
Carroll
J
,
Fulda
KG
,
Kaur
M
,
Gonzalez
K
, et al
Significant differences in global genomic DNA methylation by gender and race/ethnicity in peripheral blood
.
Epigenetics
2011
;
6
:
623
9
.
56.
Lienert
F
,
Wirbelauer
C
,
Som
I
,
Dean
A
,
Mohn
F
,
Schübeler
D
, et al
Identification of genetic elements that autonomously determine DNA methylation states
.
Nat Genet
2011
;
43
:
1091
7
.
57.
Hitchins
MP
,
Rapkins
RW
,
Kwok
CT
,
Srivastava
S
,
Wong
JJ
,
Khachigian
LM
, et al
Dominantly inherited constitutional epigenetic silencing of MLH1 in a cancer-affected family is linked to a single nucleotide variant within the 5′UTR
.
Cancer Cell
2011
;
20
:
200
13
.
58.
Chan
TL
,
Yuen
ST
,
Kong
CK
,
Chan
YW
,
Chan
AS
,
Ng
WF
, et al
Heritable germline epimutation of MSH2 in a family with hereditary nonpolyposis colorectal cancer
.
Nat Genet
2006
;
38
:
1178
83
.
59.
Tobi
EW
,
Lumey
LH
,
Talens
RP
,
Kremer
D
,
Putter
H
,
Stein
AD
, et al
DNA methylation differences after exposure to prenatal famine are common and timing- and sex-specific
.
Hum Mol Genet
2009
;
18
:
4046
53
.
60.
Li
S
,
Hursting
SD
,
Davis
BJ
,
McLachlan
JA
,
Barrett
JC
. 
Environmental exposure, DNA methylation, and gene regulation: lessons from diethylstilbesterol-induced cancers
.
Ann N Y Acad Sci
2003
;
983
:
161
9
.
61.
Federico
A
,
Federico
A
,
Morgillo
F
,
Tuccillo
C
,
Ciardiello
F
,
Loguercio
C
. 
Chronic inflammation and oxidative stress in human carcinogenesis
.
Int J Cancer
2007
;
121
:
2381
6
.
62.
Franks
AL
,
Slansky
JE
. 
Multiple associations between a broad spectrum of autoimmune diseases, chronic inflammatory diseases and cancer
.
Anticancer Res
2012
;
32
:
1119
36
.
63.
Baylin
SB
. 
The cancer epigenome: its origins, contributions to tumorigenesis, and translational implications
.
Proc Am Thorac Soc
2012
;
9
:
64
5
.
64.
Brennan
K
,
Garcia-Closas
M
,
Orr
N
,
Fletcher
O
,
Jones
M
,
Ashworth
A
, et al
Intragenic ATM methylation in peripheral blood DNA as a biomarker of breast cancer risk
.
Cancer Res
2012
;
72
:
2304
13
.
65.
Iwamoto
T
,
Iwamoto
T
,
Yamamoto
N
,
Taguchi
T
,
Tamaki
Y
,
Noguchi
S
. 
BRCA1 promoter methylation in peripheral blood cells is associated with increased risk of breast cancer with BRCA1 promoter methylation
.
Breast Cancer Res Treat
2011
;
129
:
69
77
.
66.
Wong
EM
,
Southey
MC
,
Fox
SB
,
Brown
MA
,
Dowty
JG
,
Jenkins
MA
, et al
Constitutional methylation of the BRCA1 promoter is specifically associated with BRCA1 mutation-associated pathology in early-onset breast cancer
.
Cancer Prev Res (Phila)
2011
;
4
:
23
33
.
67.
Marsit
CJ
,
Koestler
DC
,
Christensen
BC
,
Karagas
MR
,
Houseman
EA
,
Kelsey
KT
. 
DNA methylation array analysis identifies profiles of blood-derived DNA methylation associated with bladder cancer
.
J Clin Oncol
2011
;
29
:
1133
9
.
68.
Teschendorff
AE
,
Menon
U
,
Gentry-Maharaj
A
,
Ramus
SJ
,
Gayther
SA
,
Apostolidou
S
, et al
An epigenetic signature in peripheral blood predicts active ovarian cancer
.
PLoS One
2009
;
4
:
e8274
.
69.
Widschwendter
M
,
Apostolidou
S
,
Raum
E
,
Rothenbacher
D
,
Fiegl
H
,
Menon
U
, et al
Epigenotyping in peripheral blood cell DNA and breast cancer risk: a proof of principle study
.
PLoS One
2008
;
3
:
e2656
.
70.
Langevin
SM
,
Koestler
DC
,
Christensen
BC
,
Butler
RA
,
Wiencke
JK
,
Nelson
HH
, et al
Peripheral blood DNA methylation profiles are indicative of head and neck squamous cell carcinoma: an epigenome-wide association study
.
Epigenetics
2012
;
7
:
291
9
.
71.
Weber
M
,
Davies
JJ
,
Wittig
D
,
Oakeley
EJ
,
Haase
M
,
Lam
WL
, et al
Chromosome-wide and promoter-specific analyses identify sites of differential DNA methylation in normal and transformed human cells
.
Nat Genet
2005
;
37
:
853
62
.
72.
Pufulete
M
,
Al-Ghnaniem
R
,
Leather
AJ
,
Appleby
P
,
Gout
S
,
Terry
C
, et al
Folate status, genomic DNA hypomethylation, and risk of colorectal adenoma and cancer: a case control study
.
Gastroenterology
2003
;
124
:
1240
8
.
73.
Choi
JY
,
James
SR
,
Link
PA
,
McCann
SE
,
Hong
CC
,
Davis
W
, et al
Association between global DNA hypomethylation in leukocytes and risk of breast cancer
.
Carcinogenesis
2009
;
30
:
1889
97
.
74.
Xu
X
,
Gammon
MD
,
Hernandez-Vargas
H
,
Herceg
Z
,
Wetmur
JG
,
Teitelbaum
SL
, et al
DNA methylation in peripheral blood measured by LUMA is associated with breast cancer in a population-based study
.
FASEB J
2012
;
26
:
2657
66
.
75.
Beck
S
,
Rakyan
VK
. 
The methylome: approaches for global DNA methylation profiling
.
Trends Genet
2008
;
24
:
231
7
.
76.
Hsiung
DT
,
Marsit
CJ
,
Houseman
EA
,
Eddy
K
,
Furniss
CS
,
McClean
MD
, et al
Global DNA methylation level in whole blood as a biomarker in head and neck squamous cell carcinoma
.
Cancer Epidemiol Biomarkers Prev
2007
;
16
:
108
14
.
77.
Yang
AS
,
Doshi
KD
,
Choi
SW
,
Mason
JB
,
Mannari
RK
,
Gharybian
V
, et al
DNA methylation changes after 5-aza-2′-deoxycytidine therapy in patients with leukemia
.
Cancer Res
2006
;
66
:
5495
503
.
78.
Brennan
K
,
Garcia-Closas
M
,
Orr
N
,
Fletcher
O
,
Jones
M
,
Ashworth
A
, et al
Intragenic ATM methylation in peripheral blood DNA as a biomarker of breast cancer risk
.
Cancer Res
2012
;
72
:
2304
13
.
79.
Zhu
ZZ
,
Sparrow
D
,
Hou
L
,
Tarantini
L
,
Bollati
V
,
Litonjua
AA
, et al
Repetitive element hypomethylation in blood leukocyte DNA and cancer incidence, prevalence, and mortality in elderly individuals: the Normative Aging Study
.
Cancer Causes Control
2011
;
22
:
437
47
.
80.
Cash
HL
,
Tao
L
,
Yuan
JM
,
Marsit
CJ
,
Houseman
EA
,
Xiang
YB
, et al
LINE-1 hypomethylation is associated with bladder cancer risk among nonsmoking Chinese
.
Int J Cancer
2012
;
130
:
1151
9
.
81.
Zhu
ZZ
,
Hou
L
,
Bollati
V
,
Tarantini
L
,
Marinelli
B
,
Cantone
L
, et al
Predictors of global methylation levels in blood DNA of healthy subjects: a combined analysis
.
Int J Epidemiol
2010
;
41
:
126
39
.
82.
Pobsook
T
,
Pobsook
T
,
Subbalekha
K
,
Sannikorn
P
,
Mutirangura
A
. 
Improved measurement of LINE-1 sequence methylation for cancer detection
.
Clin Chim Acta
2011
;
412
:
314
21
.
83.
Patchsung
M
,
Boonla
C
,
Amnattrakul
P
,
Dissayabutra
T
,
Mutirangura
A
,
Tosukhowong
P
. 
Long interspersed nuclear element-1 hypomethylation and oxidative stress: correlation and bladder cancer diagnostic potential
.
PLoS One
2012
;
7
:
e37009
.
84.
Xu
X
,
Gammon
MD
,
Hernandez-Vargas
H
,
Herceg
Z
,
Wetmur
JG
,
Teitelbaum
SL
, et al
DNA methylation in peripheral blood measured by LUMA is associated with breast cancer in a population-based study
.
FASEB J
2012
;
26
:
2657
66
.
85.
Liao
LM
,
Brennan
P
,
van Bemmel
DM
,
Zaridze
D
,
Matveev
V
,
Janout
V
, et al
LINE-1 methylation levels in leukocyte DNA and risk of renal cell cancer
.
PLoS One
2011
;
6
:
e27361
.
86.
Woo
HD
,
Kim
J
. 
Global DNA hypomethylation in peripheral blood leukocytes as a biomarker for cancer risk: a meta-analysis
.
PLoS One
2012
;
7
:
e34615
.
87.
Di
JZ
,
Han
XD
,
Gu
WY
,
Wang
Y
,
Zheng
Q
,
Zhang
P
, et al
Association of hypomethylation of LINE-1 repetitive element in blood leukocyte DNA with an increased risk of hepatocellular carcinoma
.
J Zhejiang Univ Sci B
2011
;
12
:
805
11
.
88.
Mirabello
L
,
Savage
SA
,
Korde
L
,
Gadalla
SM
,
Greene
MH
. 
LINE-1 methylation is inherited in familial testicular cancer kindreds
.
BMC Med Genet
2010
;
11
:
77
.
89.
Wu
HC
,
Wang
Q
,
Yang
HI
,
Tsai
WY
,
Chen
CJ
,
Santella
RM
. 
Global DNA methylation levels in white blood cells as a biomarker for hepatocellular carcinoma risk: a nested case-control study
.
Carcinogenesis
2012
;
33
:
1340
5
.
90.
Cash
HL
,
Tao
L
,
Yuan
JM
,
Marsit
CJ
,
Houseman
EA
,
Xiang
YB
, et al
LINE-1 hypomethylation is associated with bladder cancer risk among nonsmoking Chinese
.
Int J Cancer
2012
;
130
:
1151
9
.
91.
Koestler
DC
,
Marsit
CJ
,
Christensen
BC
,
Accomando
W
,
Langevin
SM
,
Houseman
EA
, et al
Peripheral blood immune cell methylation profiles are associated with non-hematopoietic cancers
.
Cancer Epidemiol Biomarkers Prev
2012
;
21
:
1293
302
.
92.
Campan
M
,
Moffitt
M
,
Houshdaran
S
,
Shen
H
,
Widschwendter
M
,
Daxenbichler
G
, et al
Genome-scale screen for DNA methylation-based detection markers for ovarian cancer
.
PLoS One
2011
;
6
:
e28141
.
93.
Hagner
N
,
Joerger
M
. 
Cancer chemotherapy: targeting folic acid synthesis
.
Cancer Manag Res
2010
;
2
:
293
301
.
94.
Xu
X
,
Chen
J
. 
One-carbon metabolism and breast cancer: an epidemiological perspective
.
J Genet Genomics
2009
;
36
:
203
14
.
95.
Thornton
A
,
Lee
P
. 
Publication bias in meta-analysis: its causes and consequences
.
J Clin Epidemiol
2000
;
53
:
207
16
.
96.
Zhu
ZZ
,
Hou
L
,
Bollati
V
,
Tarantini
L
,
Marinelli
B
,
Cantone
L
, et al
Predictors of global methylation levels in blood DNA of healthy subjects: a combined analysis
.
Int J Epidemiol
2012
;
41
:
126
39
.
97.
Gao
Y
,
Baccarelli
A
,
Shu
XO
,
Ji
BT
,
Yu
K
,
Tarantini
L
, et al
Blood leukocyte Alu and LINE-1 methylation and gastric cancer risk in the Shanghai Women's Health Study
.
Br J Cancer
2012
;
106
:
585
91
.
98.
Piyathilake
CJ
,
Macaluso
M
,
Alvarez
RD
,
Chen
M
,
Badiga
S
,
Siddiqui
NR
, et al
A higher degree of LINE-1 methylation in peripheral blood mononuclear cells, a one-carbon nutrient related epigenetic alteration, is associated with a lower risk of developing cervical intraepithelial neoplasia
.
Nutrition
2011
;
27
:
513
9
.
99.
Hou
L
,
Wang
H
,
Sartori
S
,
Gawron
A
,
Lissowska
J
,
Bollati
V
, et al
Blood leukocyte DNA hypomethylation and gastric cancer risk in a high-risk Polish population
.
Int J Cancer
2010
;
127
:
1866
74
.
100.
Christensen
BC
,
Houseman
EA
,
Marsit
CJ
,
Zheng
S
,
Wrensch
MR
,
Wiemels
JL
, et al
Aging and environmental exposures alter tissue-specific DNA methylation dependent upon CpG island context
.
PLoS Genet
2009
;
5
:
e1000602
.
101.
Wilhelm
CS
,
Kelsey
KT
,
Butler
R
,
Plaza
S
,
Gagne
L
,
Zens
MS
, et al
Implications of LINE1 methylation for bladder cancer risk in women
.
Clin Cancer Res
2010
;
16
:
1682
9
.
102.
El-Maarri
O
,
Walier
M
,
Behne
F
,
van Üüm
J
,
Singer
H
,
Diaz-Lacava
A
, et al
Methylation at global LINE-1 repeats in human blood are affected by gender but not by age or natural hormone cycles
.
PLoS One
2011
;
6
:
e16252
.
103.
Singer
H
,
Walier
M
,
Nüsgen
N
,
Meesters
C
,
Schreiner
F
,
Woelfle
J
, et al
Methylation of L1Hs promoters is lower on the inactive X, has a tendency of being higher on autosomes in smaller genomes and shows inter-individual variability at some loci
.
Hum Mol Genet
2012
;
21
:
219
35
.
104.
Rusiecki
JA
,
Al-Nabhani
M
,
Tarantini
L
,
Chen
L
,
Baccarelli
A
,
Al-Moundhri
MS
, et al
Global DNA methylation and tumor suppressor gene promoter methylation and gastric cancer risk in an Omani Arab population
.
Epigenomics
2011
;
3
:
417
29
.
105.
Fraser
HB
,
Lam
LL
,
Neumann
SM
,
Kobor
MS
. 
Population-specificity of human DNA methylation
.
Genome Biol
2012
;
13
:
R8
.
106.
Moore
LE
,
Pfeiffer
RM
,
Poscablo
C
,
Real
FX
,
Kogevinas
M
,
Silverman
D
, et al
Genomic DNA hypomethylation as a biomarker for bladder cancer susceptibility in the Spanish Bladder Cancer Study: a case-control study
.
Lancet Oncol
2008
;
9
:
359
66
.
107.
Buchen
L
. 
Cancer: missing the mark
.
Nature
2011
;
471
:
428
32
.
108.
Irahara
N
,
Nosho
K
,
Baba
Y
,
Shima
K
,
Lindeman
NI
,
Hazra
A
, et al
Precision of pyrosequencing assay to measure LINE-1 methylation in colon cancer, normal colonic mucosa, and peripheral blood cells
.
J Mol Diagn
2010
;
12
:
177
83
.
109.
Ulrich
CM
,
Toriola
AT
,
Koepl
LM
,
Sandifer
T
,
Poole
EM
,
Duggan
C
, et al
Metabolic, hormonal and immunological associations with global DNA methylation among postmenopausal women
.
Epigenetics
2012
;
7
:
1020
8
.
110.
Burden
AF
,
Manley
NC
,
Clark
AD
,
Gartler
SM
,
Laird
CD
,
Hansen
RS
, et al
Hemimethylation and non-CpG methylation levels in a promoter region of human LINE-1 (L1) repeated elements
.
J Biol Chem
2005
;
280
:
14413
9
.
111.
Phokaew
C
,
Kowudtitham
S
,
Subbalekha
K
,
Shuangshoti
S
,
Mutirangura
A
. 
LINE-1 methylation patterns of different loci in normal and cancerous cells
.
Nucleic Acids Res
2008
;
36
:
5704
12
.
112.
Gallo
V
,
Egger
M
,
McCormack
V
,
Farmer
PB
,
Ioannidis
JP
,
Kirsch-Volders
M
, et al
STrengthening the Reporting of OBservational studies in Epidemiology - Molecular Epidemiology (STROBE-ME): an extension of the STROBE Statement
.
PLoS Med
2011
;
8
:
e1001117
.
113.
von Elm
E
,
Altman
DG
,
Egger
M
,
Pocock
SJ
,
Gøtzsche
PC
,
Vandenbroucke
JP
, et al
The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies
.
Lancet
2007
;
370
:
1453
7
.
114.
van Bemmel
D
,
Lenz
P
,
Liao
LM
,
Baris
D
,
Sternberg
LR
,
Warner
A
, et al
Correlation of LINE-1 methylation levels in patient matched buffy coat, serum, buccal cell and bladder tumor tissue DNA samples
.
Cancer Epidemiol Biomarkers Prev
2012
;
21
:
1143
8
.
115.
Zhang
FF
,
Santella
RM
,
Wolff
M
,
Kappil
MA
,
Markowitz
SB
,
Morabia
A
. 
White blood cell global methylation and IL-6 promoter methylation in association with diet and lifestyle risk factors in a cancer-free population
.
Epigenetics
2012
;
7
:
606
14
.
116.
Ogino
S
,
Kawasaki
T
,
Nosho
K
,
Ohnishi
M
,
Suemoto
Y
,
Kirkner
GJ
, et al
LINE-1 hypomethylation is inversely associated with microsatellite instability and CpG island methylator phenotype in colorectal cancer
.
Int J Cancer
2008
;
122
:
2767
73
.
117.
Gilson
E
,
Horard
B
. 
Comprehensive DNA methylation profiling of human repetitive DNA elements using an MeDIP-on-RepArray assay
.
Methods Mol Biol
2012
;
859
:
267
91
.