Epigenetic modifications in cancer

Epigenetic modifications in cancer

In human cancers, aberrant epigenomics are known to contribute to various phases of neoplastic development including initiation, promotion, invasion, metastases and chemotherapy resistance. It has been proposed recently that more than 300 genes and gene products are epigenetically altered in various human cancers (4). A link between DNA methylation and cancer was first shown in 1983, when it was shown that the genomes of cancer cells are hypomethylated relative to their normal counterparts (29). In terms of DNA methylation, cancer cells show genome-wide hypomethylation and site-specific CpG island promoter hypermethylation (30, 31). Hypermethylation at specific genes typically affects promoter CpG-islands inactivating transcription. Hypermethylation is observed at specific CpG islands. The transcriptional inactivation caused by promoter hypermethylation affects genes involved in the main cellular pathways: DNA repair [hMLH1 (mismatch repair gene 1), MGMT (O6-methylguanine–DNA methyltransferase), WRN (Werner syndrome, RecQ helikase like), BRCA1 (breast cancer 1)], cell cycle control (p16 ^INK4a, p15 ^{I NK4b}, RB), Ras signaling {RASSF1A [Ras association (RalGDS/AF-6) domain family member 1], NOREIA}, apoptosis [TMS1 ( target of methylation-induced silencing 1), DAPK1 (death-associated protein kinase), WIF-1, SFRP1 ], metastasis [cadherin 1 (CDH1), CDH13, PCDH10 ], detoxification [GSTP1 (glutathione S-transferase pi 1)], hormone response (ESR1, ESR2 ), vitamin response [RARB2 (retinoic acid receptor b2), CRBP1] and p53 network [p14 ^ARF , p73 (also known as TP73 ), HIC-1 ] among others. This provides tumor cells with a growth advantage and increases their genetic instability and aggressiveness (22, 30). Hypomethylation in tumor cells is primarily caused by the loss of methylation from repetitive regions of the genome and the resulting genomic instability is a hallmark of tumor cells. In addition, imprinting patterns are often disturbed (32). In cancers, hypomethylation is often associated with oncogenes. c-Myc, a transcription factor that acts as an oncogene, is one of the widely reported hypomethylated genes in cancers. Hypomethylation at specific promoters can activate the aberrant expression of oncogenes and induce loss of imprinting (LOI). The most common LOI event due to hypomethy-lation is insulin-like growth factor 2 (IGF2 ), which has been reported in a wide range of tumor types, including breast, liver, lung and colon cancer (33, 34). S100P in pancreatic cancer, SNCG in breast and ovarian cancers and melanoma-associated gene (MAGE ) and dipeptidyl peptidase 6 (DPP6 ) in melanomas are well-studied examples of hypomethylated genes in cancer (12, 35).

In addition to changes in DNA methylation, histone modification patterns are also altered in human tumors. Recent studies have shown that histone modification levels are predictive for gene expression. Actively transcribed genes are characterized by high levels of H3K4me3, H3K27ac, H2BK5ac and H4K20me1 in the promoter and H3K79me1 and H4K20me1 along the gene body (26). Loss of acetylation is mediated by HDACs that have been found to be over-expressed or mutated in different tumor types. Aberrant expression of both HMTs and HDMs are observed in various cancer types (36). A recent study has described inactivating mutations in the histone methyltransferase SETD2 and in the histone demethylase UTX and JARID1C in renal carcinomas (36, 37). H3 acetylation and H3K9 dimethylation can discriminate between cancerous and nonmalignant prostate tissue and H3K4 trimethylation can predict occurrence of prostate-specific antigen serum level elevation after prostatectomy for cancer (28). EZH2 (enhancer of zeste homolog 2) expression is an independent prognostic marker that is correlated with the aggressiveness of prostate, breast and endometrial cancers (38).

Because abnormal cell proliferation is a hallmark of human cancers, it seems possible that miRNA expression patterns might denote the malignant state. Indeed, altered expression of a few miRNAs has been found in some tumor types (39, 40). The first association between miRNA and cancer development was described in chronic lymphocytic leukemia with chromosome 13q14 deletion. This deletion deregulates miRNA-15 and miRNA-16 (41). Most of the targets of these two miRNAs are involved in cell growth and cell cycle. The let-7 is one of the most widely studied miRNA families in cancer. Alterations of let-7 function have been described in several human cancer types, including carcinomas of the head and neck region, lung, colon, rectum and ovary. It acts mainly as a tumor-suppressor miRNA (28). miRNA-145 is a well-known tumor-suppressor miRNA downregulated in many human cancers owing to aberrant DNA methylation of its promoter and/or p53 mutations (42). This miRNA is a pluripotency repressor which regulates silencing of OCT, SOX2 and KLF4 in human embryonic stem cells; these genes are required for cell self-renewal and pluripotency maintenance (43). Interestingly, it is becoming apparent that the expression of epigenetic regulatory enzymes such as DNMT, HATs, and HMTs can be controlled by miRNAs (44). In particular, the miRNA-29 family can directly regulate the expression of DNMTs such that downregulation of this family of miRNAs in small-cell lung cancer results in increased expression of DNMT3A and 3B causing a global genomic hypermethylation and specific methylation-induced silencing of tumor-suppressor genes such as FHIT and WWOX (45, 46).

Influence of epigenetics on cancer genetics

It is evident that discrete genetic alterations in neoplastic cells alone cannot explain multistep carcinogenesis whereby tumor cells are able to express diverse phenotypes during the complex phases of tumor development and progression. In fact, cancer cells have an altered epigenome compared to the tissues from which they arise. Deregulated epigenetic mechanisms may initiate genetic instability, resulting in the acquisition of genetic mutations in tumor-suppressor genes and activating genetic mutations in oncogenes. Moreover, epigenetic disruptions in tumors are generally of a clonal nature, indicating occurrence in early generations of cells (47). It is well known that 5-methylcytosine (m5C) residues are ‘hot spots’ for mutations, which can destabilize gene structure and function. One-third of germ-line point mutations leading to human genetic diseases occur at CpGs and most of these mutations are C→T transitions (48). This is because m5C is highly mutable by deamination, resulting in transitional mutations (i.e. C→T) at CpGs. In view of the symmetry of these CpG motifs, the methylcytosine on the opposite strand may also be affected, leading to (G→A) changes. As a consequence, CpGs are hot spots for mutations, in a variety of genes. (48, 49) G→A transitions are found in 44.8% cases of leukemia and myelodysplasia, and in 60% of colon cancer cases. C→T and tandem CC-TT mutations are found in basal cell and squamous cell carcinomas (50). Methylation increases the rate of hydrolytic deamination and also increases the reactivity of neighboring guanines to electrophiles (50, 51). The oxidation of m5C may contribute to the high frequency of C→T transitions at CpG sequences. Oxygen radicals can react with m5C to oxidize the 5, 6-double bond; the intermediate product, m5C glycol, then deaminates to form thymine glycol (52, 53). Oxidative stress can contribute to tumor development not only through genetic but also through epigenetic mechanisms. As noted earlier, the presence of hydroxyl radicals can cause a wide range of DNA lesions including base modifications, deletions, strand breakage and chromosomal rearrangements. Such DNA lesions have been shown to interfere with the ability of DNA to function as a substrate for the DNMTs, resulting in global hypomethylation (54). The presence of 8-OHdG in CpG dinucleotide sequences has been shown to strongly inhibit methylation of adjacent cytosine residues (55, 56). In addition, 8-OHdG may not be recognized by proof-reading enzymes and thus may persist as a mutation resulting in G→T transversions (57, 58). These studies suggest that oxidative DNA damage can affect patterns of DNA methylation leading to aberrant gene expression and possibly contributing to the development of malignancy.

Hereditary cancer genes and cancer predisposition

Studies of familial cancer have identified a group of genes whose mutational inactivation results in predisposition to a characteristic spectrum of cancers. The tumor-suppressor gene, RB which is mutated in retinoblastoma, was the first hereditary cancer gene to be identified. Subsequently, other tumor-suppressor genes operating through a diverse range of mechanisms were identified in other familial cancers, e.g. adenomatosis polyposis coli (APC) mutations were identified in cases of familial adenomatous polyposis coli and p16 ^INK4a mutations were identified in cases of familial melanoma (59). DNA repair genes such as the BRCA1, MLH1, and MSH2 are also often involved in predisposition to familial cancer (60, 61). Specific allele sequence variants such as single nucleotide polymorphisms (SNPs) affect the probability of CpG islands methylation in the cis region. Cis-acting DNA mutations have been shown to cause constitutional epi-mutations early in cancer development (62). One example of epigenetic silencing caused by sequence alterations in cis is the methylation of expanded CGG repeats within the FMR1 promoter in the fragile X mental retardation syndrome (63). Little is known about what causes variation in methylation levels and how this variation relates to future pathogenesis. Genetic factors may play key roles in maintaining a normal methylation pattern, but environmental factors also play an important role. Polymorphisms in MTHFR, DNMT3b, and MTR genes, involved in methyl metabolism or DNA methylation, are some of the obvious genetic candidates for variation in methylation in both normal and tumor tissues (64).

Epigenetic biomarkers

Methylated DNA sequences represent potential bio-markers for diagnosis, staging, prediction of prognosis and monitoring of response to therapy. Epigenetic markers can be evaluated in resected tumors or in body fluids (65). For example, the frequency of occurrence of hypermethylated CDH13, MYOD1, MGMT, p16 ^INK4a and RASSF1A genes varies significantly among cancer types. These changes can be detected in plasma DNA and urine. Recently, combined hypermethylation assays for small number of genes such as RASSF1A, RARβ2, APC and GSTP1 have been used to discriminate between benign and cancerous changes in the prostate (66, 67). Furthermore, repetitive DNA elements such as SINEs and LINEs and other repetitive sequences are often hypomethylated. Their presence has been used to characterize some human cancers, but their clinical utility remains questionable (68).

Histone modification patterns also provide prognostic and diagnostic information in cancer. Repressive chromatin structures characterized by particular histone modifications such as H3K9, H3K27 and H4K20 methylation may precipitate DNA methylation. Generalized changes in chromatin structure and histone modification, such as increased H3K4 dimethylation and H3K18 acetylation activation, are associated with poor prognosis (69). Whether the extent of these changes correlates with alteration in gene activity is a major limitation.

More recently, miRNA has been proposed as potential epigenetic biomarkers in the diagnosis of cancer. Some of the miRNA such as miR-199a, miR-200a, miR-146, miR-214, miR-221 and miR-222 have been found to be upregulated, whereas miR-100 is down-regulated in human cancers (44). The miRNA let-71 has been recently designated as a tumor-suppressor and miR-429, miR-200a and miR-200b were found to be clustered on a single primary transcript regulated by the epithelial-to-mesenchymal transition (70). Studies have shown that two other miRNAs, miR-21 and miR-181a can be used to identify the presence or absence of a malignant phenotype. A group of 27 miRNA has been shown to be significantly associated with chemotherapy response and proposed as possible prognostic and diagnostic biomarker (71, 72).