Genome-scale analysis of DNA methylation in colorectal cancer using Infinium HumanMethylation450 BeadChips

Illumina’s Infinium HumanMethylation450 BeadChip arrays were used to examine genome-wide DNA methylation profiles in 22 sample pairs from colorectal cancer (CRC) and adjacent tissues and 19 colon tissue samples from cancer-free donors. We show that the methylation profiles of tumors and healthy tissue samples can be clearly distinguished from one another and that the main source of methylation variability is associated with disease status. We used different statistical approaches to evaluate the methylation data. In general, at the CpG-site level, we found that common CRC-specific methylation patterns consist of at least 15,667 CpG sites that were significantly different from either adjacent healthy tissue or tissue from cancer-free subjects. Of these sites, 10,342 were hypermethylated in CRC, and 5,325 were hypomethylated. Hypermethylated sites were common in the maximum number of sample pairs and were mostly located in CpG islands, where they were significantly enriched for differentially methylated regions known to be cancer-specific. In contrast, hypomethylated sites were mostly located in CpG shores and were generally sample-specific. Despite the considerable variability in methylation data, we selected a panel of 14 highly robust candidates showing methylation marks in genes SND1, ADHFE1, OPLAH, TLX2, C1orf70, ZFP64, NR5A2, and COL4A. This set was successfully cross-validated using methylation data from 209 CRC samples and 38 healthy tissue samples from The Cancer Genome Atlas consortium (AUC = 0.981 [95% CI: 0.9677–0.9939], sensitivity = 100% and specificity = 82%). In summary, this study reports a large number of loci with novel differential methylation statuses, some of which may serve as candidate markers for diagnostic purposes.     

A suite of DNA methylation markers that can detect most common human cancers

Cancer-specific DNA methylation from the tumor derived fraction of cell free DNA found in blood samples could be used for minimally invasive detection and monitoring of cancer. The knowledge of marker regions with cancer-specific DNA methylation is necessary to the success of such a process. We analyzed the largest cancer DNA methylation dataset available—TCGA Illumina HumanMethylation450 data with over 8,500 tumors—in order to find cancer-specific DNA methylation markers for most common human cancers. First, we identified differentially methylated regions for individual cancer types and those were further filtered against data from normal tissues to obtain marker regions with cancer-specific methylation, resulting in a total of 1,250 hypermethylated and 584 hypomethylated marker CpGs. From hypermethylated markers, optimal sets of six markers for each TCGA cancer type were chosen that could identify most tumors with high specificity and sensitivity [area under the curve (AUC): 0.969-1.000] and a universal 12 marker set that can detect tumors of all 33 TCGA cancer types (AUC >0.84). In addition to hundreds of new DNA methylation markers, our approach also identified markers that are in current clinical use, SEPT9 and GSTP1, indicating the validity of our approach and a significant potential utility for the newly discovered markers. The hypermethylated markers are linked to polycomb associated loci and a significant fraction of the discovered markers is within noncoding RNA genes; one of the best markers is MIR129-2. Future clinical testing of herein discovered markers will confirm new markers that will improve minimally invasive diagnosis and monitoring for multiple cancers.     

Genome-wide DNA methylation profiles in noncancerous gastric mucosae with regard to Helicobacter pylori infection and the presence of gastric cancer

Background and Aims: To determine genome‐wide DNA methylation profiles induced by Helicobacter pylori (H. pylori) infection and to identify methylation markers in H. pylori‐induced gastric carcinogenesis. Methods: Gastric mucosae obtained from controls (n = 20) and patients with gastric cancer (n = 28) were included. A wide panel of CpG sites in cancer‐related genes (1505 CpG sites in 807 genes) was analyzed using Illumina bead array technology. Validation of the results of Illumina bead array technique was performed using methylation‐specific PCR method for four genes (MOS, DCC, CRK, and PTPN6). Results: The Illumina bead array showed that a total of 359 CpG sites (269 genes) were identified as differentially methylated by H. pylori infection (p < .0001). The correlation between methylation‐specific PCR and bead array analysis was significant (p < .0001, Spearman coefficient = 0.5054). Methylation profiles in noncancerous gastric mucosae of the patients with gastric cancer showed quite distinct patterns according to the presence or absence of the current H. pylori infection; however, 10 CpG sites were identified to be hypermethylated and three hypomethylated in association with the presence of gastric cancer regardless of H. pylori infection (p < .01). Conclusions: Genome‐wide methylation profiles showed a number of genes differentially methylated by H. pylori infection. Methylation profiles in noncancerous gastric mucosae from the patients with gastric cancer can be affected by H. pylori‐induced gastritis. Differentially methylated CpG sites in this study needs to be validated in a larger population using quantitative methylation‐specific PCR method.     

Key epigenetic changes associated with lung cancer development

Epigenetic alterations are a common event in lung cancer and their identification can serve to inform on the carcinogenic process and provide clinically relevant biomarkers. Using paired tumor and non-tumor lung tissues from 146 individuals from three independent populations we sought to identify common changes in DNA methylation associated with the development of non-small cell lung cancer. Pathologically normal lung tissue taken at the time of cancer resection was matched to tumorous lung tissue and together were probed for methylation using Illumina GoldenGate arrays in the discovery set (n = 47 pairs) followed by bisulfite pyrosequencing for validation sets (n = 99 pairs). For each matched pair the change in methylation at each CpG was calculated (the odds ratio), and these ratios were averaged across individuals and ranked by magnitude to identify the CpGs with the greatest change in methylation associated with tumor development. We identified the top gene-loci representing an increase in methylation (HOXA9, 10.3-fold and SOX1, 5.9-fold) and decrease in methylation (DDR1, 8.1-fold). In replication testing sets, methylation was higher in tumors for HOXA9 (p < 2.2 × 10^?16) and SOX1 (p < 2.2 × 10^?16) and lower for DDR1 (p < 2.2 × 10^?16). The magnitude and strength of these changes were consistent across squamous cell and adenocarcinoma tumors. Our data indicate that the identified genes consistently have altered methylation in lung tumors. Our identified genes should be included in translational studies that aim to develop screening for early disease detection.     

DNA Methylation Profile Distinguishes Clear Cell Sarcoma of the Kidney from Other Pediatric Renal Tumors

A number of specific, distinct neoplastic entities occur in the pediatric kidney, including Wilms’ tumor, clear cell sarcoma of the kidney (CCSK), congenital mesoblastic nephroma (CMN), rhabdoid tumor of the kidney (RTK), and the Ewing’s sarcoma family of tumors (ESFT). By employing DNA methylation profiling using Illumina Infinium HumanMethylation27, we analyzed the epigenetic characteristics of the sarcomas including CCSK, RTK, and ESFT in comparison with those of the non-neoplastic kidney (NK), and these tumors exhibited distinct DNA methylation profiles in a tumor-type-specific manner. CCSK is the most frequently hypermethylated, but least frequently hypomethylated, at CpG sites among these sarcomas, and exhibited 490 hypermethylated and 46 hypomethylated CpG sites in compared with NK. We further validated the results by MassARRAY, and revealed that a combination of four genes was sufficient for the DNA methylation profile-based differentiation of these tumors by clustering analysis. Furthermore, THBS1 CpG sites were found to be specifically hypermethylated in CCSK and, thus, the DNA methylation status of these THBS1 sites alone was sufficient for the distinction of CCSK from other pediatric renal tumors, including Wilms’ tumor and CMN. Moreover, combined bisulfite restriction analysis could be applied for the detection of hypermethylation of a THBS1 CpG site. Besides the biological significance in the pathogenesis, the DNA methylation profile should be useful for the differential diagnosis of pediatric renal tumors.     

High-throughput assessment of CpG site methylation for distinguishing between HCV-cirrhosis and HCV-associated hepatocellular carcinoma

Methylation of promoter CpG islands has been associated with gene silencing and demonstrated to lead to chromosomal instability. Therefore, some postulate that aberrantly methylated CpG regions may be important biomarkers indicative of cancer development. In this study we used the Illumina GoldenGate Methylation BeadArray Cancer Panel I for simultaneously profiling methylation of 1,505 CpG sites in order to identify methylation differences in 76 liver tissues ranging from normal to pre-neoplastic and neoplastic states. CpG sites for ESR1, GSTM2, and MME were significantly differentially methylated when comparing the pre-neoplastic tissues from patients with concomitant hepatocellular carcinoma (HCC) to the pre-neoplastic tissues from patients without HCC. When comparing paired HCC tissues to their corresponding pre-neoplastic non-tumorous tissues, eight CpG sites, including one CpG site that was hypermethylated (APC) and seven (NOTCH4, EMR3, HDAC9, DCL1, HLA-DOA, HLA-DPA1, and ERN1) that were hypomethylated in HCC, were identified. Our study demonstrates that high-throughput methylation technologies may be used to identify differentially methylated CpG sites that may prove to be important molecular events involved in carcinogenesis.     

Epigenetic status of cell cycle regulation genes in the placenta of human embryos with chromosomal mosaicism

Epigenetic mechanisms of cell cycle regulation providing for maintenance of genome stability and precise transmission of hereditary information to the daughter cells attract considerable interest. Up-to-date molecular technologies allow the genome-wide methylation pattern to be determined in a single experiment. In this work, we pioneered in determining the epigenetic status of the placental tissues of the human embryos with mosaic karyotype using a genome-wide Illumina Infinium HumanMethylation27 BeadChip array (Illumina, United States), comprising 27578 CpG dinucleotides contained in 14475 genes. The groups of genes associated with the cell cycle and its regulation that contain differentially methylated CpG sites in their promoter regions have been determined. It was demonstrated that the methylation level most frequently changed in oncogenes (ARHGEF1 and FGF5), tumor suppressors (APC2, BRACA2, DCC, GRLF1, RB1, TP73, TSPYL2, and VHL), and the genes involved in the regulation of chromosome segregation (CNTROB, GMNN, PROCR, and TACC1).     

Genome‐wide mapping of cytosine methylation revealed dynamic DNA methylation patterns associated with genes and centromeres in rice

We conducted genome‐wide mapping of cytosine methylation using methylcytosine immunoprecipitation combined with Illumina sequencing. The chromosomal distribution pattern of methylated DNA is similar to the heterochromatin distribution pattern on rice chromosomes. The DNA methylation patterns of rice genes are similar to those in Arabidopsis thaliana, including distinct methylation patterns asssociated with gene bodies and promoters. The DNA sequences in the core domains of rice Cen4, Cen5 and Cen8 showed elevated methylation levels compared with sequences in the pericentromeric regions. In addition, elevated methylation levels were associated with the DNA sequences in the CENH3‐binding subdomains, compared with the sequences in the flanking H3 subdomains. In contrast, the centromeric domain of Cen11, which is composed exclusively of centromeric satellite DNA, is hypomethylated compared with the pericentromeric domains. Thus, the DNA sequences associated with functional centromeres can be either hypomethylated or hypermethylated. The methylation patterns of centromeric DNA appear to be correlated with the composition of the associated DNA sequences. We propose that both hypomethylation and hypermethylation of CENH3‐associated DNA sequences can serve as epigenetic marks to distinguish where CENH3 deposition will occur within the surrounding H3 chromatin.     

Diagnostics of Epigenetic Alterations in Hereditary and Oncological Disorders

The review considers the epigenetic defects and their diagnostics in several hereditary disorders and tumors. Aberrant methylation of the promoter or regulatory region of a gene results in its functional inactivation, which is phenotypically similar to structural deletion. Screening tests were developed for Prader–Willi, Angelman, Wiedemann–Beckwith, and Martin–Bell syndromes and mental retardation FRAXE. The tests are based on allele methylation analysis by methylation-specific or methylation-sensitive PCR. Carcinogenesis-associated genes (RB1, CDKN2A, ARF14, HIC1, CDH1, etc.) are often methylated in tumors. Tumors differ in methylation frequencies, allowing differential diagnostics. Aberrant methylation of tumor suppressor genes occurs in early carcinogenesis, and its detection may be employed in presymptomatic diagnostics of tumors.     

Aberrant DNA methylation of WNT pathway genes in the development and progression of CIMP-negative colorectal cancer

The WNT signaling pathway has an essential role in colorectal carcinogenesis and progression, which involves a cascade of genetic and epigenetic changes. We aimed to analyze DNA methylation affecting the WNT pathway genes in colorectal carcinogenesis in promoter and gene body regions using whole methylome analysis in 9 colorectal cancer, 15 adenoma, and 6 normal tumor adjacent tissue (NAT) samples by methyl capture sequencing. Functional methylation was confirmed on 5-aza-2′-deoxycytidine-treated colorectal cancer cell line datasets. In parallel with the DNA methylation analysis, mutations of WNT pathway genes (APC, β-catenin/CTNNB1) were analyzed by 454 sequencing on GS Junior platform. Most differentially methylated CpG sites were localized in gene body regions (95% of WNT pathway genes). In the promoter regions, 33 of the 160 analyzed WNT pathway genes were differentially methylated in colorectal cancer vs. normal, including hypermethylated AXIN2, CHP1, PRICKLE1, SFRP1, SFRP2, SOX17, and hypomethylated CACYBP, CTNNB1, MYC; 44 genes in adenoma vs. NAT; and 41 genes in colorectal cancer vs. adenoma comparisons. Hypermethylation of AXIN2, DKK1, VANGL1, and WNT5A gene promoters was higher, while those of SOX17, PRICKLE1, DAAM2, and MYC was lower in colon carcinoma compared to adenoma. Inverse correlation between expression and methylation was confirmed in 23 genes, including APC, CHP1, PRICKLE1, PSEN1, and SFRP1. Differential methylation affected both canonical and noncanonical WNT pathway genes in colorectal normal-adenoma-carcinoma sequence. Aberrant DNA methylation appears already in adenomas as an early event of colorectal carcinogenesis.     

Genome-wide DNA methylation profiling of chronic lymphocytic leukemia allows identification of epigenetically repressed molecular pathways with clinical impact

We performed a genome-wide analysis of aberrant DNA methylation in chronic lymphocytic leukemia (CLL) using methylated CpG island amplification (MCA) coupled with a promoter microarray. We identified 280 potential targets of aberrant DNA methylation in CLL. These genes were located more frequently in chromosomes 19 (16%, p=0.001), 16 (11%, p=0.001), 17 (10%, p=0.02) and 11 (9%, p=0.02) and could be grouped in several functional networks. Methylation status was confirmed for 22 of these genes (SOX11, DLX1, FAM62C, SOX14, RSPO1, ADCY5, HAND2,SPOCK, MLL, ING1, PRIMA1, BCL11B, LTBP2, BNC1, NR2F2, SALL1, GALGT2, LHX1, DLX4, KLK10, TFAP2 and APP) in 78 CLL patients by pyrosequencing. As a proof of principle, we analyzed the expression of 2 genes, PRIMA1 and APP, in primary cells and of GALGT2, TFAP2C and PRIMA1 in leukemia cells. There was an inverse association between methylation and gene expression. This could be reversed by treatment with 5-aza-2’-deoxycytidine in cell lines. Treatment in a clinical trial with 5-azacitidine resulted in decreased methylation of LINE, DLX4 and SALL1 in the peripheral blood B-cells of patients with CLL. IgVH mutational status or ZAP-70 expression were not associated with specific methylation profiles. By multivariate analysis, methylation of LINE and APP was associated with shorter overall survival (p = 0.045 and 0.0035, respectively). This study demonstrates that aberrant DNA methylation is common and has potential prognostic and therapeutic value in CLL.     

DNA methylation profiles in primary cutaneous melanomas are associated with clinically significant pathologic features

DNA methylation studies have elucidated a methylation signature distinguishing primary melanomas from benign nevi and provided new insights about genes that may be important in melanoma development. However, it is unclear whether methylation differences among primary melanomas are related to tumor pathologic features with known clinical significance. We utilized the Illumina GoldenGate Cancer Panel array to investigate the methylation profiles of 47 primary cutaneous melanomas. Arraywide methylation patterns revealed a positive association of methylation with Breslow thickness and mutated BRAF, a negative association with mitotic rate, and a weak association with ulceration. Hierarchical clustering on CpG sites exhibiting the most variable methylation (n = 235) divided the melanoma samples into three clusters, including a highly methylated cluster that was positively associated with Breslow thickness and an intermediately methylated cluster associated with Breslow thickness and mitotic rate. Our findings provide support for the existence of methylation‐defined subsets in melanomas with increased methylation associated with Breslow thickness.     

Methylomic analysis identifies frequent DNA methylation of zinc finger protein 582 (ZNF582) in cervical neoplasms

Background

Despite of the trend that the application of DNA methylation as a biomarker for cancer detection is promising, clinically applicable genes are few. Therefore, we looked for novel hypermethylated genes for cervical cancer screening.

Methods and Findings

At the discovery phase, we analyzed the methylation profiles of human cervical carcinomas and normal cervixes by methylated DNA immunoprecipitation coupled to promoter tiling arrays (MeDIP-on-chip). Methylation-specific PCR (MSP), quantitative MSP and bisulfite sequencing were used to verify the methylation status in cancer tissues and cervical scrapings from patients with different severities. Immunohistochemical staining of a cervical tissue microarray was used to confirm protein expression. We narrowed to three candidate genes: DBC1, PDE8B, and ZNF582; their methylation frequencies in tumors were 93%, 29%, and 100%, respectively. At the pre-validation phase, the methylation frequency of DBC1 and ZNF582 in cervical scraping correlated significantly with disease severity in an independent cohort (n = 330, both P<0.001). For the detection of cervical intraepithelial neoplasia 3 (CIN3) and worse, the area under the receiver operating characteristic curve (AUC) of ZNF582 was 0.82 (95% confidence interval  = 0.76–0.87).

Conclusions

Our study shows ZNF582 is frequently methylated in CIN3 and worse lesions, and it is demonstrated as a potential biomarker for the molecular screening of cervical cancer.     

DNA methylation changes associated with risk factors in tumors of the upper aerodigestive tract

Cancers of the upper aerodigestive tract (UADT) are common forms of malignancy associated with tobacco and alcohol exposures, although human papillomavirus and nutritional deficiency are also important risk factors. While somatically acquired DNA methylation changes have been associated with UADT cancers, what triggers these events and precise epigenetic targets are poorly understood. In this study, we applied quantitative profiling of DNA methylation states in a panel of cancer-associated genes to a case-control study of UADT cancers. Our analyses revealed a high frequency of aberrant hypermethylation of several genes, including MYOD1, CHRNA3 and MTHFR in UADT tumors, whereas CDKN2A was moderately hypermethylated. Among differentially methylated genes, we identified a new gene (the nicotinic acetycholine receptor gene) as target of aberrant hypermethylation in UADT cancers, suggesting that epigenetic deregulation of nicotinic acetycholine receptors in non-neuronal tissues may promote the development of UADT cancers. Importantly, we found that sex and age is strongly associated with the methylation states, whereas tobacco smoking and alcohol intake may also influence the methylation levels in specific genes. This study identifies aberrant DNA methylation patterns in UADT cancers and suggests a potential mechanism by which environmental factors may deregulate key cellular genes involved in tumor suppression and contribute to UADT cancers.     

Genome-wide aberrant DNA methylation of microRNA host genes in hepatocellular carcinoma

Mature microRNAs (miRNAs) are a class of small non-coding RNAs involved in posttranslational gene silencing. Previous studies found that downregulation of miRNAs is a common feature observed in solid tumors, including hepatocellular carcinoma (HCC). We employed a genome-wide approach to test the hypothesis that DNA methylation alterations in miRNA host genes may cause deregulated miRNA expression in HCC. We analyzed tumor and adjacent non-tumor tissues from 62 Taiwanese HCC cases using Infinium HumanMethylation27 DNA Analysis BeadChips that include 254 CpG sites covering 110 miRNAs from 64 host genes. Expression levels of three identified miRNAs (miR-10a, miR-10b and miR-196b) were measured in a subset of 37 HCC tumor and non-tumor tissues. After Bonferroni adjustment, a total of 54 CpG sites from 27 host genes significantly differed in DNA methylation levels between tumor and adjacent non-tumor tissues with 53 sites significantly hypermethylated in tumor tissues. Among the 54 significant CpG sites, 15 sites had more than 2-fold tumor/non-tumor changes, 17 sites had differences > 10%, and 10 sites had both features [including 8 significantly hypermethylated CpG sites in the host genes of miR-10a, miR-10b and miR-196b (HOXB4, HOXD4 and HOXA9, respectively)]. Significant downregulation of miR-10a was observed in tumor compared with non-tumor tissues (0.50 vs. 1.73, p = 0.031). The concordance for HOXB4 methylation alteration and dysregulation of miR-10a was 73.5%. No significant change was observed for miR-10b expression. Unexpectedly, miR-196b was significantly upregulated in tumor compared with non-tumor tissues (p = 0.0001). These data suggest that aberrant DNA methylation may lead to dysregulation of miR-10a in HCC tumor tissues.