Abstract

Colorectal cancer (CRC) presents as a very heterogeneous disease which cannot sufficiently be characterized with the currently known genetic and epigenetic markers. To identify new markers for CRC we scrutinized the methylation status of 231 DNA repair-related genes by methyl-CpG immunoprecipitation followed by global methylation profiling on a CpG island microarray, as altered expression of these genes could drive genomic and chromosomal instability observed in these tumors. We show for the first time hypermethylation of MMP9, DNMT3A and LIG4 in CRC which was confirmed in two CRC patient groups with different ethnicity. DNA ligase IV (LIG4) showed strong differential promoter methylation (up to 60%) which coincided with downregulation of mRNA in 51% of cases. This functional association of LIG4 methylation and gene expression was supported by LIG4 re-expression in 5-aza-2′-deoxycytidine-treated colon cancer cell lines, and reduced ligase IV amounts and end-joining activity in extracts of tumors with hypermethylation. Methylation of LIG4 was not associated with other genetic and epigenetic markers of CRC in our study. As LIG4 is located on chromosome 13 which is frequently amplified in CRC, two loci were tested for gene amplification in a subset of 47 cases. Comparison of amplification, methylation and expression data revealed that, in 30% of samples, the LIG4 gene was amplified and methylated, but expression was not changed. In conclusion, hypermethylation of the LIG4 promoter is a new mechanism to control ligase IV expression. It may represent a new epigenetic marker for CRC independent of known markers.

INTRODUCTION

Worldwide, more than 1 million people are diagnosed with colorectal cancer (CRC) each year (1). On the molecular level, sporadic CRCs are a heterogeneous disease and can be classified into specific phenotypes based on genetic instability and DNA methylation. Frequently observed mutations affect cancer associated genes like APC, KRAS, TP53 and BRAF. In addition, chromosomal instability (CIN) or microsatellite instability (MSI) is found. MSI is characterized by an accumulation of mutations in microsatellite regions in the DNA caused by defective DNA mismatch repair (2,3). Approximately 15% of CRCs display MSI (4) with the majority showing repression of the mismatch repair gene MLH1 by means of promoter hypermethylation (5). CRCs without MSI present a heterogeneous group that can be further differentiated into cases showing CIN with amplifications, deletions and translocations of whole chromosomes or parts of chromosomes (6) and a smaller group of cases showing no CIN (7,8). In summary, alterations in mechanisms involved in maintaining genomic stability such as DNA repair seem to be strongly involved in colon cancer development.

A common phenomenon of CRC is a concordant hypermethylation of multiple CpG islands (CGIs) known as CGI methylator phenotype (CIMP) (9). Today, different marker panels for identification of CIMP have been established including the distinction between two different CIMP types: CIMP-high (CIMP-H) and CIMP2 (also termed CIMP-L) (10–12). MSI and CIMP were shown to be inversely correlated with CIN in sporadic CRC (13). This indicates the presence of independent tumorigenic pathways and highlights the important role of aberrant DNA methylation in colon tumorigenesis. However, the mechanisms underlying CIN are poorly defined and seem to be more complex than the association between defective mismatch repair and MSI.

Thus, our aim was to identify new genes showing aberrant methylation in CRC that might be involved in tumorigenesis. We used methyl-CpG immunoprecipitation (MCIp), which enriches for methylated DNA (14). We screened for differentially methylated regions in 16 matched pairs of colon cancer and normal tissues by analysis of the enriched methylated DNA on CGI microarrays. Since defects in the DNA repair machinery are likely to play a central role in the development of genetic instability, a DNA repair pathway-specific analysis of MCIp data was performed. The epigenetic regulation of one of the identified genes, the DNA, ATP-dependent ligase IV (LIG4), was characterized in more detail. Ligase IV is the enzyme finalizing the non-homologous end-joining (NHEJ) of DNA double-strand breaks (DSBs) during DNA repair and V(D)J recombination in the immune response. Our results highlight that LIG4 expression deregulated by DNA methylation has a functional impact on DSB end-joining in a colon cancer cell model and in colon tumor tissue.

RESULTS

Identification of new DNA repair genes hypermethylated in colon cancer

A genome-wide DNA methylation analysis was performed in DNA of tissues from 16 CRC patients (Supplementary Material, Table S2) [Department of Pathology, Hong Kong University (HKU)] using MCIp and hybridization to CGI micorarrays. For a DNA repair pathway-related evaluation, a compilation of 231 genes (Supplementary Material, Table S1) was used which consisted of genes involved in the core DNA repair pathways and DNA damage signaling (15) as well as genes participating in replication and chromatin remodeling (16). For selecting differentially methylated candidate CGIs, a significance analysis of microarrays (SAMs) analysis of all probes and all 16 samples with a false discovery rate (FDR) of 5% revealed 44 920 significantly hypermethylated probes. These included 258 probes annotated with DNA repair-related genes. A further selection step concentrated on hypermethylated regions which required at least two significantly hypermethylated probes within a 500 bp genomic region. This resulted in 37 CGIs associated with 28 different genes from our list which were hypermethylated in colon cancer (Table 1).

Table 1.

List of differentially methylated regions and related genes selected by SAM

Gene Chromosome Location of differentially methylated regiona FDR q value (%)b 
BRCA2 13 TSS 0.1008 
DNMT3A Intragenic 0.1008 
DNMT3A TSS 0.1940 
DOT1L 19 Intragenic region A 0.0000 
DOT1L 19 Intragenic region B 0.0289 
DOT1L 19 Intragenic region C 0.6816 
DOT1L 19 Intragenic region D 1.2200 
DVL3 Intragenic 0.1008 
G3BP1 TSS 0.3645 
H2AFZ TSS 2.1295 
HDAC4 Intragenic region A 0.0102 
HDAC4 Intragenic region B 0.0102 
HDAC4 Intragenic region C 0.0516 
HDAC4 Intragenic region D 0.1940 
HDAC4 TSS 2.1295 
HLTF TSS 0.1008 
LIG1 19 Intragenic 1.2200 
LIG4 13 TSS 0.0516 
MBD3 19 Intragenic 0.0000 
MLH1 TSS 3.5372 
MMP9 20 Intragenic region A 0.0102 
MMP9 20 Intragenic region B 0.6816 
MSH6 TSS 0.0516 
PLK1 16 TSS 1.2200 
PNKP 19 Intragenic 0.3645 
PPP2R5C 14 Intragenic 0.0000 
PRMT1 19 Intragenic 0.0000 
RAD52 12 TSS 0.0000 
RECQL4 Intragenic 0.6816 
SETD7 TSS 0.3645 
SIX4 14 Intragenic 2.1295 
SMARCA4 19 Intragenic 0.0289 
SNRPF 12 TSS 0.0000 
SOX4 Intragenic 1.2200 
TGIF2 20 TSS 0.1008 
TSTA3 Intragenic 0.6816 
UNG 12 TSS 2.1295 
Gene Chromosome Location of differentially methylated regiona FDR q value (%)b 
BRCA2 13 TSS 0.1008 
DNMT3A Intragenic 0.1008 
DNMT3A TSS 0.1940 
DOT1L 19 Intragenic region A 0.0000 
DOT1L 19 Intragenic region B 0.0289 
DOT1L 19 Intragenic region C 0.6816 
DOT1L 19 Intragenic region D 1.2200 
DVL3 Intragenic 0.1008 
G3BP1 TSS 0.3645 
H2AFZ TSS 2.1295 
HDAC4 Intragenic region A 0.0102 
HDAC4 Intragenic region B 0.0102 
HDAC4 Intragenic region C 0.0516 
HDAC4 Intragenic region D 0.1940 
HDAC4 TSS 2.1295 
HLTF TSS 0.1008 
LIG1 19 Intragenic 1.2200 
LIG4 13 TSS 0.0516 
MBD3 19 Intragenic 0.0000 
MLH1 TSS 3.5372 
MMP9 20 Intragenic region A 0.0102 
MMP9 20 Intragenic region B 0.6816 
MSH6 TSS 0.0516 
PLK1 16 TSS 1.2200 
PNKP 19 Intragenic 0.3645 
PPP2R5C 14 Intragenic 0.0000 
PRMT1 19 Intragenic 0.0000 
RAD52 12 TSS 0.0000 
RECQL4 Intragenic 0.6816 
SETD7 TSS 0.3645 
SIX4 14 Intragenic 2.1295 
SMARCA4 19 Intragenic 0.0289 
SNRPF 12 TSS 0.0000 
SOX4 Intragenic 1.2200 
TGIF2 20 TSS 0.1008 
TSTA3 Intragenic 0.6816 
UNG 12 TSS 2.1295 

TSS, transcription start site.

aFor some genes several differentially regions were identified.

bThe most significantly differentially methylated probe in the CGI.

Using the following criteria, three genes were selected for validation in two independent sample sets: MMP9 was the gene with the highest number of differentially methylated probes, DNMT3A as a methyltransferase might play a role in tumor-associated hypermethylation and LIG4 represented a major component of DSB repair by NHEJ. Differential methylation was confirmed by the quantitative MassARRAY methylation assay. In Validation set 1 (Fig. 1A and Supplementary Material, Table S2), the median methylation values were higher in tumor tissue than in controls for MMP9 (65 versus 37%), DNMT3A (51 versus 35%) and LIG4 (36 versus 15%). All differences were statistically significant (P < 0.001, Mann–Whitney test). Similar methylation values for these genes were found in Validation set 2 (Fig. 1B). In summary, methylation differences discovered by MCIp analysis were confirmed by quantitative methylation analysis in two independent sample sets of CRC patients of different ethnic origin.

Figure 1.

Pathway-specific evaluation of a genome-wide methylation screen revealed MMP9, DNMT3A and LIG4 as differentially methylated genes which could be confirmed with independent quantitative methylation analysis. (A) Validation set 1 with patients from Hong Kong, (B) Validation set 2 with patients from Heidelberg (T, tumor; N, normal; ***P < 0.001, **P < 0.01, Mann–Whitney test; bars indicate median).

Figure 1.

Pathway-specific evaluation of a genome-wide methylation screen revealed MMP9, DNMT3A and LIG4 as differentially methylated genes which could be confirmed with independent quantitative methylation analysis. (A) Validation set 1 with patients from Hong Kong, (B) Validation set 2 with patients from Heidelberg (T, tumor; N, normal; ***P < 0.001, **P < 0.01, Mann–Whitney test; bars indicate median).

Based on mRNA expression in colon cancer cell lines and on the novelty as a differentially methylated marker gene in colon cancer, further analyses focused on LIG4. MMP9 and DNMT3A were not further analyzed as mRNA expression in CRC cell lines (HCT116, HT29 and SW48) was below the detection limit and expression in colon mucosa was reported to be low (BioGPS gene expression database www.biogps.org). The UCSC genome browser revealed three transcripts of LIG4 that all encode the same protein (Fig. 2A). Both the long and the two short transcripts of LIG4 were well detectable and expressed in the cell lines (Fig. 3C). The differentially methylated CGI was located at the internal promoter site and associated with open chromatin marks suggesting a possible role in transcriptional regulation. The importance of correct ligase IV function is supported by reports that LIG4 plays an important role in DSB repair via NHEJ, that a loss of ligase IV function by mutations causes severe immunological and developmental defects (LIG4 syndrome) (17), and that LIG4 is involved in the pathogenesis of Myelodysplastic Syndromes (MDS) (18). Furthermore, LIG4 deficient, heterozygous mice have shown increased cytogenetic aberrations (19) and dramatic CIN (20). However, there are no reports on gene regulation so far, especially on differential methylation in tumors. Thus, we hypothesized that a reduced LIG4 expression due to hypermethylation might play a role in CRC development.

Figure 2.

DNA methylation of LIG4 CGIs in CRC. (A) Schematic representation of the LIG4 gene located on chromosome 13q33–34 together with transcripts and H3K4Me1 and H3K4Me3 histone marks (black bars: exons; open bars: CGIs; gray bars: amplicons used for methylation analysis; arrows indicate transcription start sites). (B) DNA methylation analysis of the CGI associated with the two short LIG4 transcripts 2 and 3 in Validation sets 1 and 2 (LIG4_a1 amplicon; T, tumor; N, normal; ***P < 0.001, **P < 0.01, Mann–Whitney test; bars indicate median). (C) DNA methylation analysis of the CGI associated with the long LIG4 transcripts 1 in Validation set 1.

Figure 2.

DNA methylation of LIG4 CGIs in CRC. (A) Schematic representation of the LIG4 gene located on chromosome 13q33–34 together with transcripts and H3K4Me1 and H3K4Me3 histone marks (black bars: exons; open bars: CGIs; gray bars: amplicons used for methylation analysis; arrows indicate transcription start sites). (B) DNA methylation analysis of the CGI associated with the two short LIG4 transcripts 2 and 3 in Validation sets 1 and 2 (LIG4_a1 amplicon; T, tumor; N, normal; ***P < 0.001, **P < 0.01, Mann–Whitney test; bars indicate median). (C) DNA methylation analysis of the CGI associated with the long LIG4 transcripts 1 in Validation set 1.

Figure 3.

LIG4 is downregulated in CRC by DNA methylation. (A) LIG4 mRNA expression of transcripts 2 and 3 relative to GAPDH in tumor and adjacent normal tissues (N, normal; T, tumor; ***P < 0.001, Mann–Whitney test; whiskers indicate maximum and minimum values). The 47 samples that were used for methylation analysis are shown separately as points in the right column (bar indicates median). (B) Association of reduced LIG4 expression and hypermethylation in 41 CRC patients. Methylation values higher than the 75-percentile of the normal tissues (>19%) were defined as LIG4 hypermethylated (n = 36); relative mRNA expression lower than in the mean of the normal tissues (<0.0019) was defined as LIG4 downregulated (n = 24). (C) Expression of LIG4 transcripts 1 and 2 and 3 in the colon cancer cell lines HCT116, HT29 and SW48 relative to ACTB, GAPDH and HPRT as reference genes. D and E) Treatment of cell lines with 0.5 µm 5-aza-dC for 96 h leads to (D) a decrease in global methylation (measured using LINE-1 as surrogate), (E) a decrease in LIG4 methylation (upper panel), and an increase in LIG4 mRNA expression (lower panel). (F) LIG4 methylation and expression in HCT 116 wild-type and DNMT1/DNMT3b knockout cells.

Figure 3.

LIG4 is downregulated in CRC by DNA methylation. (A) LIG4 mRNA expression of transcripts 2 and 3 relative to GAPDH in tumor and adjacent normal tissues (N, normal; T, tumor; ***P < 0.001, Mann–Whitney test; whiskers indicate maximum and minimum values). The 47 samples that were used for methylation analysis are shown separately as points in the right column (bar indicates median). (B) Association of reduced LIG4 expression and hypermethylation in 41 CRC patients. Methylation values higher than the 75-percentile of the normal tissues (>19%) were defined as LIG4 hypermethylated (n = 36); relative mRNA expression lower than in the mean of the normal tissues (<0.0019) was defined as LIG4 downregulated (n = 24). (C) Expression of LIG4 transcripts 1 and 2 and 3 in the colon cancer cell lines HCT116, HT29 and SW48 relative to ACTB, GAPDH and HPRT as reference genes. D and E) Treatment of cell lines with 0.5 µm 5-aza-dC for 96 h leads to (D) a decrease in global methylation (measured using LINE-1 as surrogate), (E) a decrease in LIG4 methylation (upper panel), and an increase in LIG4 mRNA expression (lower panel). (F) LIG4 methylation and expression in HCT 116 wild-type and DNMT1/DNMT3b knockout cells.

Quantitative characterization of methylation differences in two CGIs associated with LIG4

We therefore comprehensively characterized aberrant DNA methylation of LIG4 in CRC samples of both validation sets. The LIG4 gene harbors two CGIs (Fig. 2A). The hypermethylation found in colon tumors by MCIp is located in the downstream CGI that has a length of 300 bp. Amplicon LIG4_a1 covers 10 CpG sites, seven of which gave a read-out in the MassARRAY assay and revealed a highly significant hypermethylation in CRC tissues in the two independent sample sets (Validation set 1: P = 0.0001, set 2 P = 0.004; Mann–Whitney test; Fig. 2B). Validation set 1 included DNA from 11 matched pairs of tumor and normal colon tissue and methylation was increased in seven of the matched tumors (Supplementary Material, Fig. S1, P < 0.01, Wilcoxon matched pair test). In addition, we confirmed increased DNA methylation by clonal bisulfite sequencing in eight tumor and two normal samples from Validation set 1 (Supplementary Material, Fig. S2).

A second CGI covers the transcription start site of the longest LIG4 transcript and is shared by ABHD13, a gene which is transcribed in the opposite direction. This CGI is 889 bp long and the 303 bp amplicon named LIG4_a2 covers 29 CpG sites, 19 of which were analyzed by the MassARRAY assay. This region is completely unmethylated in normal and tumor colon tissues (Fig. 2C) suggesting that it serves as a bidirectional promoter for ABHD13 and LIG4, and that transcriptional regulation of LIG4 via DNA methylation occurs at the downstream CGI.

To further support the role of LIG4 methylation in CRC development, we analyzed Validation set 1 for associations between LIG4 hypermethylation and further molecular (MSI and CIMP status, mutations in KRAS, BRAF, TP53 and APC) or clinical parameters (age, sex and Duke's stage) (21). None of the variables was significantly associated with LIG4 hypermethylation (Table 2). In particular, there was no increase in methylation according to CIMP status suggesting LIG4 hypermethylation as a novel marker in CRC, independent from so far known markers. This association analysis of methylation and clinical parameters was not repeated in Validation set 2 because of missing data.

Table 2.

Clinical and molecular characteristics of CRC patients in the HK validation set and the association of these characteristics with LIG4 hypermethylation

Characteristics All patients Patients with LIG4 hypermethylationa
 
P-valueb 
n n 
Total 65 49 75  
Sex 
 Male 31 24 77  
 Female 34 25 74 0.78 
Age 
 ≤50 years 29 23 79 0.57 
 >50 years 36 26 72 
Duke’s stage 
 A 75 0.17 
 B 26 23 88 
 C 25 18 72 
 D 56 
 No info.  
MSI status 
 MSI 22 16 73 0.77 
 MSS 43 33 77 
CIMP status 
 CIMP 45 36 80 0.22 
 Non-CIMP 20 13 65 
KRAS mutation 
 Mutant 32 24 75 1.00 
 Wild-type 33 25 76 
BRAF mutation 
 Mutant 83 1.00 
 Wild-type 59 44 75 
TP53 mutation 
 Mutant 27 20 74 0.48 
 Wild-type 22 19 86 
 No info. 16 10  
APC mutation 
 Mutant 12 10 63 1.0 
 Wild-type 13 10 83 
 No info 40 29  
Characteristics All patients Patients with LIG4 hypermethylationa
 
P-valueb 
n n 
Total 65 49 75  
Sex 
 Male 31 24 77  
 Female 34 25 74 0.78 
Age 
 ≤50 years 29 23 79 0.57 
 >50 years 36 26 72 
Duke’s stage 
 A 75 0.17 
 B 26 23 88 
 C 25 18 72 
 D 56 
 No info.  
MSI status 
 MSI 22 16 73 0.77 
 MSS 43 33 77 
CIMP status 
 CIMP 45 36 80 0.22 
 Non-CIMP 20 13 65 
KRAS mutation 
 Mutant 32 24 75 1.00 
 Wild-type 33 25 76 
BRAF mutation 
 Mutant 83 1.00 
 Wild-type 59 44 75 
TP53 mutation 
 Mutant 27 20 74 0.48 
 Wild-type 22 19 86 
 No info. 16 10  
APC mutation 
 Mutant 12 10 63 1.0 
 Wild-type 13 10 83 
 No info 40 29  

aTumors showing a higher methylation in the LIG4_a1 region than the 75-percentile of the normal tissues (19%) were defined as hypermethylated (see Fig. 2A and B). % indicate the amount of hypermethylated cases in relation to all samples with a specific characteristic.

bFisher's exact test compared patients with normal methylation and those with hypermethylation of LIG4.

LIG4 expression is downregulated in CRC

To investigate a possible functional impact of LIG4 methylation in CRC, mRNA expression of LIG4 was analyzed by quantitative RT–PCR in 393 CRC and 61 paired normal colon (Validation set 3, Supplementary Material, Table S2). Analysis focused on an amplicon covering the two transcripts LIG4_2 and LIG4_3 because their transcription start sites were located within or close to the differentially methylated CGI (Fig. 2A). A marked downregulated expression of LIG4 was detected in the tumor samples (P < 0.001; Fig. 3A). Validation set 3 included 47 tumor samples of Validation set 1 with methylation data. This subgroup showed an identical distribution of LIG4 expression as the complete cohort (Fig. 3A). About half of the samples (24/47) had lower relative LIG4 expression than the 25-percentile of the normals (<0.0019), therefore they were defined as LIG4 downregulated. The majority of the LIG4 downregulated tumors (19/24; 79%) showed LIG4 hypermethylation (Fig. 3B) suggesting a functional link between DNA hypermethylation at the internal promoter site of LIG4 and reduced gene expression.

To corroborate this link, colon cancer cell lines (HCT116, HT29 and SW48) were treated with the demethylating agent 5-aza-2′-deoxycytidine (5-aza-dC), and methylation state of the LIG4 promoter and re-expression of the gene were measured. Decrease of LINE-1 methylation reflected the effect of the treatment on global methylation (Fig. 3D). LIG4 methylation levels also decreased upon treatment: in HCT116 cells, average LIG4 methylation was reduced from 68 to 35%, in HT29 from 53 to 22%, and in SW 48 from 92 to 33% (Fig. 3E, upper graph). Accordingly, LIG4 mRNA expression was induced after the 5-aza-dC treatment: in HCT116 LIG4 expression increased 1.6-fold upon the demethylation, in HT29 3.5-fold and in SW48 1.7-fold (Fig. 3E lower graph, for LIG4 expression levels without 5-aza-dC, see Fig. 3C). The link between promoter methylation and expression was further supported when HCT116 wild-type cells were compared with their isogenic double-knockout counterpart (DNMT1−/−DNMT3b−/−) cells (Fig. 3F). In fact, decreased LIG4 methylation (upper graph) and increased LIG4 mRNA expression (lower graph) was measured in the deficient cell line compared with the wild type. These results support our hypothesis on transcriptional regulation of LIG4 expression via DNA methylation.

Association of LIG4 amplification and methylation

LIG4 is located on 13q33-34, a chromosomal region known for frequent amplification in CRC [e.g. the Integrative OncoGenomics database, www.intogen.org (22)]. Amplification could affect gene expression or methylation status. Thus, we determined amplification of LIG4 in Validation set 1 in order to correlate it with expression or methylation. A quantitative RT–PCR assay was applied to genomic DNA using primers that detected either LIG4 as a target or the single copy gene Albumin (ALB) as a reference gene. Data revealed that all normal tissues had two LIG4 copies, while 52% (34/65) of the CRC samples contained three or more copies of LIG4 (Fig. 4A). Analysis of an additional locus on chromosome 13 confirmed LIG4 amplification status in 17 of 22 samples (77%). Frequency of amplification and the fact that only microsatellite stability (MSS) but no MSI samples showed amplification is in line with published data (Supplementary Material, Fig. S3) (7).

Figure 4.

LIG4 amplification, methylation and mRNA expression in CRC. (A) LIG4 copy numbers determined by quantitative PCR in CRC tumors and adjacent normal tissues (T, tumor; N, normal; ***P < 0.001, Mann–Whitney test; bars indicate median). All samples with a copy number >3.0 were categorized as amplified. (B) Overlap between LIG4 hypermethylation (n = 36), downregulation (n = 24) and amplification (n = 27) in a subset of 44 cases (for definition of hypermethylation and downregulation, see Fig. 3).

Figure 4.

LIG4 amplification, methylation and mRNA expression in CRC. (A) LIG4 copy numbers determined by quantitative PCR in CRC tumors and adjacent normal tissues (T, tumor; N, normal; ***P < 0.001, Mann–Whitney test; bars indicate median). All samples with a copy number >3.0 were categorized as amplified. (B) Overlap between LIG4 hypermethylation (n = 36), downregulation (n = 24) and amplification (n = 27) in a subset of 44 cases (for definition of hypermethylation and downregulation, see Fig. 3).

It was striking that none of the tumors showed an mRNA expression higher than in normal tissues (Fig. 3A), although, when extrapolating from our amplification data, about half of the tumors should have an increased number of gene copies. We verified this observation in a subset of 44 tumors, for which LIG4 methylation, mRNA expression and copy number data were available. Within this sample set, 36 samples were hypermethylated, 24 showed lower LIG4 expression and 27 were carrying three or more LIG4 copies as compared with normal tissue. When assessing the overlap between these features, it became evident that the majority of hypermethylated cases (32/36) was associated with either decreased LIG4 expression (n = 19) or amplification and normal expression (n = 13) (Fig. 4B). Only four cases showed hypermethylation without amplification or a change in gene expression. This underlines the importance of DNA methylation in the transcriptional regulation of LIG4 expression.

Functional consequences of impaired ligase IV protein expression in CRC

In order to study consequences of the observed hypermethylation and downregulation of LIG4, we monitored DSB repair after a 10-Gy irradiation by histone H2AX phosphorylation (γH2AX assay) in the colon cancer cell line HCT116 with and without LIG4 siRNA knockdown (Fig. 5A and B). Cells only treated with transfection reagent reached the strongest γH2AX signal intensity 0.5 h after irradiation and removed ∼50% of the signal within 22 h after irradiation indicating efficient DNA DSB repair (Fig. 5C). Cells with LIG4 knockdown showed a similar intensity of γH2AX signals after radiation, but they did not reduce the signal within 22 h. Results indicate that a loss of LIG4 expression leads to a severe DSB repair defect in a colon cancer cell line.

Figure 5.

Knockdown of LIG4 expression affects DSB repair in HCT116 cells. Efficiency of LIG4 knockdown on (A) mRNA and (B) protein level 72 h after siRNA transfection (data of a representative experiment shown). (C) γH2AX signal intensities 72 h after LIG4 knockdown. Cells were irradiated with 10 Gy and repair was allowed for the indicated time points. (Error bars indicate standard deviation of four independent measurements, *P < 0.05, Mann–Whitney test.)

Figure 5.

Knockdown of LIG4 expression affects DSB repair in HCT116 cells. Efficiency of LIG4 knockdown on (A) mRNA and (B) protein level 72 h after siRNA transfection (data of a representative experiment shown). (C) γH2AX signal intensities 72 h after LIG4 knockdown. Cells were irradiated with 10 Gy and repair was allowed for the indicated time points. (Error bars indicate standard deviation of four independent measurements, *P < 0.05, Mann–Whitney test.)

We also examined the impact of transient ectopic overexpression of LIG4 in the CRC cell lines HCT116 and HT29. Although we found ∼1000-fold overexpression of LIG4 mRNA, protein levels were only moderately increased, especially in HT29 (Supplementary Material, Fig. S4). This LIG4 overexpression did not affect cell viability measured by propidium iodide and trypan blue exclusion as well as calcein fluorescence (Supplementary Material, Fig. S5). In HCT116, however, LIG4 overexpression affected the cell cycle as indicated by a significant decrease of the G1 phase.

In addition, we measured LIG4 methylation, protein amount and end-joining activity in nuclear extracts (see Supplementary Material, Fig. S6) from 10 randomly selected colon tumors and normal adjacent mucosa samples to analyze the correlation between LIG4 promoter methylation and protein activity. As four samples showed protein degradation visible in the western blot, only six sample pairs remained in the analysis. Three of the tumor tissues (P1–P3) showed LIG4 hypermethylation (>19%, the cutoff value determined in Validation set 1, Fig. 6A). Two samples clearly showed reduced ligase IV protein levels (see samples P2 and P6, Fig. 6B), whereas for sample P1 the western blot was ambiguous. For the samples without methylation change (P4 and P5) equal amounts of ligase IV protein and end-joining activity was found. End-joining activity (Fig. 6C) was reduced in tumors of sample pairs P1, P2 and P6, corresponding to the reduced amount of protein in tumors P2 and P6 and to the hypermethylation in tumors P1 and P2. Thus, this small sample set confirms the functional impact of LIG4 methylation status on ligase IV activity; however, sample pairs P3 and P6 indicate that further mechanisms must be active in the regulation of LIG4 expression.

Figure 6.

LIG4 methylation, protein amount and end-joining activity in nuclear extracts from tumors and matched normal mucosa of 6 CRC patients (P1–P6). (A) Methylation analysis; a methylation value >19% (higher than the 75-percentile of the normal tissues in Validation set 1) was defined as hypermethylated. (B) Western blot detection of ligase IV. Actin B served as the reference gene. (C) End-joining activity measured in nuclear extracts; ligation sites were calculated relative to substrate input. (Error bars indicate standard deviations from three technical replicates.)

Figure 6.

LIG4 methylation, protein amount and end-joining activity in nuclear extracts from tumors and matched normal mucosa of 6 CRC patients (P1–P6). (A) Methylation analysis; a methylation value >19% (higher than the 75-percentile of the normal tissues in Validation set 1) was defined as hypermethylated. (B) Western blot detection of ligase IV. Actin B served as the reference gene. (C) End-joining activity measured in nuclear extracts; ligation sites were calculated relative to substrate input. (Error bars indicate standard deviations from three technical replicates.)

DISCUSSION

Our genome-wide MCIp analysis revealed new differentially methylated regions associated with CRC. We scrutinized the methylation status of 231 DNA repair-related genes as altered expression of these genes could drive the genomic instability observed in CRC. DNA methylation-related reduction of mismatch repair gene expression, e.g. of MLH1 and MSH2, is already known in MSI colorectal tumors (23). Molecular causes of MSS tumors, however, which typically present with CIN (24) are not analyzed as detailed so far but further defects in DNA repair genes are plausible. An additional link between DNA repair and CRC comes from reports on exogenous cancer risk factors which are associated with certain subtypes of CRC, especially smoking and CIMP tumors (25,26). As our screen focused on differentially methylated regions with functional impact but not on those with the strongest methylation differences, we carefully controlled our statistical evaluation for detection of false positives by a FDR of 5% and by counting only CGIs with at least two significant probes close to each other. We identified the already known hypermethylation of MLH1 with this strategy and thus consider this stringent data evaluation approach as reliable.

Its success is further underlined as three hypermethylated candidate regions in the MMP9, DNMT3A and LIG4 genes were newly identified as biomarkers for CRC. This was confirmed in independent CRC sample sets with a quantitative methylation assay. The markers MMP9, a member of the matrix metalloproteinase family degrading Type IV and V collagen, and DNMT3A, a DNA methyltransferase, were not further followed up in this study, as for both genes we found only low amounts of mRNA expression in normal colon tissues and CRC cell lines. They could however present biomarkers for diagnosis or prognosis in CRC.

LIG4 encodes an ATP-dependent DNA ligase that is essential for V(D)J recombination and joins DSBs in the NHEJ pathway, thus loss of LIG4 expression is supposed to contribute to genomic instability and tumorigenesis. Ligase IV is the crucial enzyme for completing NHEJ by forming a complex together with XRCC4 and the XRCC4-like factor XLF for final ligation of the break in an ATP-dependent step (27). In mice, homozygous inactivation of LIG4 was found to be embryonic lethal at E16.5 in two independent knockout approaches (28,29). LIG4−/− mouse embryonic fibroblasts generated from 13.5-day-old embryos exhibited growth defects, premature senescence and increased sensitivity to ionizing radiation. LIG4+/− heterozygous mice could not be distinguished from their wild-type littermates, however, LIG4+/− fibroblasts showed increased CIN (30). Mice with a homozygous LIG4 R278H mutation showed also colon adenocarcinomas (31). The LIG4 syndrome is a rare hereditary disorder characterized by growth retardation, developmental delay, microcephaly, unusual facial features, skin anomalies and pancytopenia. Different LIG4 mutations were reported, but all of them are hypomorphic and leave residual ligase activity (32). Patients exhibit marked radiosensitivity, early genome instability, immunodeficiency and bone marrow abnormalities (33). Some of them showed T-cell leukemia, B-cell lymphoma or myelodysplasia (17) implying a higher risk for lymphoid malignancies in this disease.

As a novelty in CRC, we demonstrate LIG4 hypermethylation of a CGI located close to the transcription start site of two LIG4 transcripts in ∼80% of CRC tissues analyzed. Asian patients (Validation set 1) showed a similar frequency and distribution of LIG4 methylation as Caucasian patients (Validation set 2) indicating that there are no differences between ethnic groups. Interestingly, LIG4 hypermethylation was not associated with any of the investigated molecular and clinical markers (age, gender, Duke's stage, MSI status, CIMP status, KRAS, BRAF, TP53 and APC mutation). As a conclusion, LIG4 hypermethylation is not a passenger of the CIMP phenotype suggesting that it could serve as a novel and independent marker in CRC.

Quantification of mRNA and protein levels in CRC tissue, adjacent normal tissues and cell lines showed that LIG4 is moderately expressed in normal colon mucosa and cell lines whereas expression is weaker in ∼50% of the tumors. Thus, the frequent hypermethylation of LIG4 found in CRC tissues was accompanied by a downregulation of the gene on the mRNA and the protein level suggesting a regulatory role of promoter DNA methylation. Demethylation experiments in cell lines using 5-aza-dC strengthened this association between LIG4 hypermethylation and downregulation of expression. In addition, we confirmed that DSB repair was impaired by LIG4 knockdown in a colon cancer cell line and that end-joining activity was compromised in cellular extracts from colon tumors with methylated LIG4. These data point at an important function of LIG4 in the colon; this function seems to be disturbed in a subgroup of colon tumors. No details are known about the transcriptional regulation of LIG4; published investigations on LIG4 regulation mainly covered protein stability and posttranslational modifications (34,35) but not, as shown here, epigenetic regulation of this gene.

It was, however, striking that only half of the LIG4 hypermethylated cases had lower LIG4 mRNA levels suggesting that further mechanisms must be involved in LIG4 gene regulation. We suspected LIG4 amplification to be involved, because the gene is located on chromosome 13q33–34, a region which is frequently amplified in CRC (7). In general, amplification can cause increased gene expression (36). We could verify LIG4 amplification in ∼50% of the cases of Validation set 1. In Validation set 3, where we measured LIG4 mRNA expression in 393 CRC and 61 adjacent normal tissues, distribution of data reveals that none of the CRC samples showed a higher expression than that obtained for the group of normal tissues. Although we did not measure amplification of the LIG4 locus in this sample set, we have to assume that about half of the samples show this amplification. Thus, we conclude that promoter DNA methylation might be a mechanism in CRC to balance LIG4 amplification and to prevent overexpression. This is in concordance with a recent report that the vast majority of genes located on frequently amplified chromosome arms which are frequently amplified in CRC do not show elevated expression levels (37). Interestingly, the strongest LIG4 reexpression after 5-aza-dC demethylating treatment was found in HT29 cells which have been reported to harbor amplification of chromosome 13 (38). We concluded that LIG4 expression is tightly regulated and its overexpression might not be well tolerated by the cell. In line with this, transient LIG4 overexpression in HCT116 cells (without LIG gene amplification) led to a moderate protein overexpression and slight alterations of the cell cycle. Protein overexpression in HT29 was only marginal indicating a strong posttranslational control of protein expression. HT29 shows a strong alteration of the cell cycle (data not shown) which is caused by a defect in p53. Potential changes induced by LIG4 overexpression might therefore not be detectable. Moreover, the limited effects of LIG4 overexpression might be due to the fact that ligase IV, together with XRCC4, is a component of the functional heterodimeric ligation complex and, thus, does not exert its activity on its own (39). A further mechanism controlling the dosage of LIG4 mRNA amounts could be posttranscriptional silencing by miRNAs. We used miRGator to predict LIG4 targeting miRNAs (miRGator v3.0; http://mirgator.kobic.re.kr). Some of them might be expressed in CRC and could regulate LIG4 expression (40). In addition, reduced LIG4 levels and activity in tumors could be caused by somatic gene mutations. To address this point we screened the Cancer Genome Atlas Network database for LIG4 mutations detected in 276 colon and rectal cancers by exome sequencing (41). In fact, LIG4 mutations were found in mismatch repair deficient, hypermutated MSI tissues as well as in non-hypermutated cancers but not at a high frequency. To further corroborate our data, a detailed investigation of the interplay between LIG4 methylation, amplification and expression is required which should also include additional potential regulatory factors such as the strong posttranslational control of ligase IV activity. In addition, there are ligase IV-independent DSB repair pathways via homologous recombination (HR) (42,43) or a back-up NHEJ pathway (B-NHEJ) (44). Both pathways involve DNA ligase III and could compensate for a LIG4 deficiency in NHEJ.

Moreover, further molecular defects of CRC tumors such as mismatch repair or p53 mutation status could provide an additional repair defect. Information on the interaction of these defects is scarce; we do however know that p53 knockout can compensate certain effects of ligase IV deficiency in mice (45), probably because of deficient cell cycle control. Interactions between repair defects could be useful targets for new cancer treatments by providing a rationale for synthetic lethality (46). In summary, our data present evidence that LIG4 promoter hypermethylation contributes to reduced LIG4 expression in CRC. An important question remains whether this differential methylation affects survival of CRC patients.

MATERIALS AND METHODS

Patient cohorts

Frozen tumor and normal colon samples of 16 Chinese CRC patients were used for genome-wide DNA methylation analysis (discovery set, Department of Pathology, HKU). Twelve patients showed CIMP characterized by 12 CIMP markers, four patients were CIMP-negative, two patients showed MSI. All samples except the two MSI cases were patients with an age at diagnosis of 50 years or younger. Differentially methylated sites detected from genome-wide methylation analysis were confirmed in two independent sample sets by quantitative methylation analysis. Validation Set 1 (Department of Pathology, HKU) comprised 65 CRC tissue samples and 30 samples from adjacent normal tissues including 11 matched tissue pairs. Validation set 2 (Department of Applied Tumor Biology, Institute of Pathology, Heidelberg) included DNA from 50 CRC tissue samples and 9 from adjacent normal tissues and was used for an independent validation of the methylation data from the MCIp screen. For analysis of mRNA expression, 393 CRC tumor tissues and 61 matched tissues from adjacent normal tissues (Validation Set 3, Department of Pathology, HKU) were used. The use of archival frozen samples from Hong Kong for the current study was approved by the Institutional Review Board of the University of Hong Kong and the Hospital Authority Hong Kong Western Cluster. The use of samples from Heidelberg was approved by the local institutional ethics committees. Patient characteristics are summarized in Supplementary Material, Table S2. DNA or RNA extraction was performed after cryostat sectioning to confirm the tumor purity in tumor block (>70%) and the absence of tumor contamination in normal block.

Cell culture, 5-aza-dC and siRNA treatment

Colon cancer cell lines HCT116 and the isogenic double-knockout (HCTDKO (47)) cells (DNMT1−/−; DNMT3B−/−) were cultivated in McCoy’s medium, HT29 and SW48 in RPMI. Both culture media were supplemented with 10% FCS and 1% penicillin/streptomycin. All cell lines were maintained at 37°C in a humidified 5% CO2 atmosphere. Cell lines were authenticated and tested for mycoplasma contamination in regular intervals by the DKFZ Genomics and Proteomics Core Facility (48,49).

For treatment with 5-aza-dC, cells were seeded in T75 culture flasks (5 × 105 cells) the day before the treatment. Freshly diluted 5-aza-dC (Sigma-Aldrich) in PBS was added to the culture medium at a final concentration of 0.5 µm. Treatment was repeated every 24 h. Treatment with PBS was used as a negative control and cells were harvested 96 h after starting the treatment.

For transient transfection with siRNA, HCT116 cells were seeded in 24-well plates (2.5 × 104 cells/well) and were kept in antibiotic-free culture medium overnight. A pool of three Ambion Silencer Select siRNAs targeting LIG4 (Life Technologies) was added together with DharmaFECT1 transfection reagent (Thermo Fisher Scientific) at a final concentration of 20 nm to the cells. Transfection reagent only served as negative control. Cells were harvested 72 h after transfection, efficiency of the knockdown was confirmed by mRNA quantification and western blot.

DNA and RNA isolation

Genomic DNA from cell lines was isolated using the QIAamp DNA Mini Kit (Qiagen), total RNA was isolated using the RNeasy Mini Kit (Qiagen). DNA and RNA from frozen tissue samples were isolated using the Allprep DNA/RNA/Protein Mini Kit (Qiagen) or comparable methods.

MCIp and CGI microarray analysis

Methyl-CpG-binding domain (MBD)-Fc protein was produced and MCIp performed as previously described (14,50,51) with minor modifications. In brief, 30 µg of MBD-Fc protein was coupled to 50 µl SIMAG protein A magnetic beads (Chemicell) at 4°C overnight. After binding of 2 µg sonicated DNA, the DNA was eluted with increasing NaCl concentrations (300–1000 mm) using the MagnetoPure-Micro bead separator (Chemicell). Weakly and highly methylated DNA eluted at the lowest and highest concentrations, respectively. Effective DNA enrichment was monitored by quantitative RT–PCR of the imprinted SNRPN gene. Enriched methylated DNA from CRC tissues was labeled with Alexa Fluor 5, DNA from adjacent normal tissues with Alexa Fluor 3. Labeled samples were cohybridized to a 244 K Human CGI microarray (Agilent Technologies). Arrays contain 199 400 probes covering 27 800 CGIs. Data processing and statistical analyses were done within the R statistical environment, v. 2.13.1 (52). Background correction and log 2-ratio transformation were performed according to the NormExp method with offset = 50 (53). To reduce variations between co-hybridized samples, intensity-based LOESS normalization on rank-invariant probes and negative controls was applied (54). One class significance analysis of microarrays (SAM, package samr, version 2.0 (55)), was performed to find significantly hypermethylated probes with FDR = 5%.

Quantitative DNA methylation analysis

Genomic DNA was treated with sodium bisulfite using the EZ DNA methylation Kit (Zymo Research). Quantitative DNA methylation analysis was performed by MassARRAY technique (Sequenom) as previously described (56). Amplicons were designed to cover hypermethylated regions derived from microarray results, for primers see Supplementary Material, Table S3.

Clonal bisulfite sequencing

Bisulfite-treated DNA was PCR-amplified using the same primers and conditions as for MassARRAY analysis. Gel-purified PCR products were ligated into the pCR 2.1 vector using the TOPO TA Cloning Kit (Invitrogen) followed by heat-shock transformation into XL 10 Gold ultracompetent cells (Agilent Technologies). Sequences of 24 colonies per sample were determined (GATC Biotech AG) and analyzed using a BIQ Analyzer (57).

mRNA quantification

cDNA was synthesized from 500 ng total RNA using Superscript III reverse transcriptase (Invitrogen) and oligo(dT) primers. mRNA expression analysis was performed in triplicates on a LightCycler 480 (Roche) using the LightCycler 480 Probes Master and hydrolysis probes from the Universal Probe Library (Roche) as described (www.roche-applied-science.com). Normalized expression ratios were determined for each sample using ACTB, GAPDH and HPRT1 as reference genes; primer sequences see Supplementary Material, Table S3. Relative quantification of LIG4 mRNA expression in Validation set 3 was performed in triplicates on an ABI PRISM 7000 instrument using QuantiFast SYBR Green (Qiagen) and GAPDH as the reference gene.

LIG4 copy number analysis

LIG4 copy numbers in CRC patient samples were measured by quantitative PCR on genomic DNA (5 ng per reaction) using primers for LIG4 as the target gene and ALB as the reference gene on a LightCycler 480, primer sequences see Supplementary Material, Table S3. All samples were measured in triplicates and the average Cp value was applied for calculation. Copy numbers were calculated according to Bodin et al. (58). In brief, DNA from adjacent normal tissue without copy number alterations was used as calibrator. LIG4 Cp values of the tumor samples were normalized to ALB as internal reference gene and to the Cp values of the normal tissues as calibrator.

Analysis of DSB repair using γH2AX assay

HCT 116 cells were seeded in 24-well plates (2.5 × 104 cells/well) and adherence was allowed overnight. At 72 h after LIG4 siRNA treatment, cells were irradiated with 10 Gy using a Cs-137 unit (Gammacell® 40 Exactor, Best Theratronics, Ottawa, Canada) at a dose rate of 50 cGy/min to induce DSBs, followed by incubation at 37°C for 0.5, 4 and 22 h. Cells were trypsinized and fixed with 2 ml of 1% paraformaldehyde solution (59). Fixed cells of six wells were combined, centrifuged and resuspended in ice cold 70% ethanol, then 1% Triton X. Cells were labeled using Alexa488 mouse Anti-H2AX-phosphorylated (Ser)139 antibody (Santa Cruz Biotechnology). Labeled cells were analyzed on a FACS Calibur (BD Bioscience) using CellQuest Pro software.

Western blot analysis of ligase IV

Nuclear extracts from cell lines and frozen tissues were prepared with NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific, Rockford, IL, USA). The amount of protein was quantified by the Bicinchoninic Acid Assay (Sigma-Aldrich). Proteins were separated according to size by SDS–PAGE using the NuPAGE Gel System and blotted on a 45 µm polyvinylidene fluoride (PVDF) membrane using the XCell SureLock System. The membrane was blocked with 5% milk in Tris-buffered Saline Tween20 (TBST) for 1 h at room temperature. After washing with TBST, the membrane was incubated with the primary ligase IV antibody [ab26039, abcam, (rabbit polyclonal) diluted 1:1600 in 1% milk in TBST] at 4°C overnight. After washing, the membrane was incubated with the Horseradish peroxidase (HRP)-coupled secondary anti-rabbit antibody sc-2004 (Santa Cruz) for 1 h at room temperature. For protein detection, the Western Lightning Plus-ECL Enhanced Chemiluminescence Substrate system was used. For normalization, β-actin (sc-47778, Santa Cruz) was determined on the same blot.

Cell viability and cell cycle distribution in CRC cells transiently overexpressing LIG4

A clone containing the complete LIG4 open reading frame was obtained from the Genomics and Proteomics Core Facility, DKFZ, Heidelberg. As expression vector, the pDEST-D11 vector was used containing a cytomegalovirus promoter and an attached T2A-green fluorescent protein (GFP) at the C-terminus of the expressed protein. Cells were transiently transfected with either empty vector (EV, pDest-D11-T2A-GFP) or LIG4 overexpressing vector (LIG4, pDest-D11-LIG4-T2A-GFP). As transfecting agent TransIT (Mirus) was used for HCT116 and JetPRIME (Polyplus) for HT29 cells. GFP positive cells were selected by flow cytometry or cell sorting (Imaging and Cytometry Core Facility, DKFZ, Heidelberg). They were considered as efficiently transfected and were used in analyses for cell viability by trypan blue or propidium iodide exclusion or calcein fluorescence. Distribution of cell cycle phases was measured by propidium iodide staining and flow cytometry.

End-joining assay

End-joining activity was measured according to (60) and quantified by RT–PCR (61). pBlueScript II K+S plasmid was linearized by digestion with BamHI (New England Biolabs), purified by PCR Purification Kit (Qiagen) and used as a substrate. The linearized plasmid had a size of 2961 bp and sticky ends. The end-joining reaction was performed at 25°C for 60 min in a total volume of 20 µl including 10 µg of desalted nuclear extract and 110 ng linearized plasmid in ligation buffer (20 mm HEPES–KOH, pH 7.5; 80 mm KCl; 10 mm MgCl2; 1 mm DTT, 1 mm ATP). An extract from HCT116 cells served as a positive control, extracts from HCT116 treated with LIG4 siRNA as a negative control (Supplementary Material, Fig. S4). Ligation sites were quantified by RT–PCR with primers flanking the ligation site (primer-lig-l: TAAAACGACGGCCAGTGAG; primer-lig-r: CCTCGAGGTCGACGGTATC). The total amount of substrate used in the reaction was measured using primers in the center of the plasmid (primer-subst-l: TTGCCGGGAAGCTAGAGTAA; primer-subst-r: AAGCCATACCAAACGACGGAG). The PCR products were detected using the QuantiTect SYBR Green PCR Kit (Qiagen). Evaluation of PCR data was based on the ΔΔCt method. The background amount of ligation product was subtracted from the final amount of ligated sites.

Statistical analyses of quantitative methylation analyses

Comparison of methylation values between two groups (e.g. normal versus tumor) was performed by the Mann–Whitney U-test or by the Wilcoxon matched pair test. Associations between LIG4 hypermethylation and patient characteristics were assessed by Fisher's exact test. Differences were considered as significant at P-values of <0.05. Analyses were performed on GraphPad Prism software.

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG online.

FUNDING

This work was supported by the Ernst-Schering-Foundation and a project-related personal exchange grant of Deutscher Akademischer Austauschdienst (DAAD) and Hong Kong University (Germany/Hong Kong Joint Research Scheme G_HK024/09, General Research Fund (HKU761408M).

ACKNOWLEDGEMENTS

The authors are thankful to all the patients for participation in the study. We thank Professor B. Vogelstein for the kind gift of HCT116 and HCT116 (DNMT1−/−; DNMT3B−/−) double-knockout cells, and R. Gliniorz, O. Mücke, P. Waas and M. Bär for their excellent technical assistance.

Conflict of Interest statement. None declared.

REFERENCES

1
Ferlay
J.
Shin
H.R.
Bray
F.
Forman
D.
Mathers
C.
Parkin
D.M.
Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008
Int. J. Cancer.
 , 
2010
, vol. 
127
 (pg. 
2893
-
2917
)
2
Ionov
Y.
Peinado
M.A.
Malkhosyan
S.
Shibata
D.
Perucho
M.
Ubiquitous somatic mutations in simple repeated sequences reveal a new mechanism for colonic carcinogenesis
Nature
 , 
1993
, vol. 
363
 (pg. 
558
-
561
)
3
Thibodeau
S.N.
Bren
G.
Schaid
D.
Microsatellite instability in cancer of the proximal colon
Science
 , 
1993
, vol. 
260
 (pg. 
816
-
819
)
4
Boland
C.R.
Goel
A.
Microsatellite instability in colorectal cancer
Gastroenterology
 , 
2010
, vol. 
138
 (pg. 
2073
-
2087, e2073
)
5
Kane
M.F.
Loda
M.
Gaida
G.M.
Lipman
J.
Mishra
R.
Goldman
H.
Jessup
J.M.
Kolodner
R.
Methylation of the hMLH1 promoter correlates with lack of expression of hMLH1 in sporadic colon tumors and mismatch repair-defective human tumor cell lines
Cancer Res.
 , 
1997
, vol. 
57
 (pg. 
808
-
811
)
6
Lengauer
C.
Kinzler
K.W.
Vogelstein
B.
Genetic instability in colorectal cancers
Nature
 , 
1997
, vol. 
386
 (pg. 
623
-
627
)
7
Chan
T.L.
Curtis
L.C.
Leung
S.Y.
Farrington
S.M.
Ho
J.W.
Chan
A.S.
Lam
P.W.
Tse
C.W.
Dunlop
M.G.
Wyllie
A.H.
, et al.  . 
Early-onset colorectal cancer with stable microsatellite DNA and near-diploid chromosomes
Oncogene
 , 
2001
, vol. 
20
 (pg. 
4871
-
4876
)
8
Georgiades
I.B.
Curtis
L.J.
Morris
R.M.
Bird
C.C.
Wyllie
A.H.
Heterogeneity studies identify a subset of sporadic colorectal cancers without evidence for chromosomal or microsatellite instability
Oncogene
 , 
1999
, vol. 
18
 (pg. 
7933
-
7940
)
9
Toyota
M.
Ahuja
N.
Ohe-Toyota
M.
Herman
J.G.
Baylin
S.B.
Issa
J.P.
CpG island methylator phenotype in colorectal cancer
Proc. Natl. Acad. Sci. USA
 , 
1999
, vol. 
96
 (pg. 
8681
-
8686
)
10
Shen
L.
Toyota
M.
Kondo
Y.
Lin
E.
Zhang
L.
Guo
Y.
Hernandez
N.S.
Chen
X.
Ahmed
S.
Konishi
K.
, et al.  . 
Integrated genetic and epigenetic analysis identifies three different subclasses of colon cancer
Proc. Natl. Acad. Sci. USA
 , 
2007
, vol. 
104
 (pg. 
18654
-
18659
)
11
Weisenberger
D.J.
Siegmund
K.D.
Campan
M.
Young
J.
Long
T.I.
Faasse
M.A.
Kang
G.H.
Widschwendter
M.
Weener
D.
Buchanan
D.
, et al.  . 
CpG island methylator phenotype underlies sporadic microsatellite instability and is tightly associated with BRAF mutation in colorectal cancer
Nat. Genet.
 , 
2006
, vol. 
38
 (pg. 
787
-
793
)
12
Yagi
K.
Akagi
K.
Hayashi
H.
Nagae
G.
Tsuji
S.
Isagawa
T.
Midorikawa
Y.
Nishimura
Y.
Sakamoto
H.
Seto
Y.
, et al.  . 
Three DNA methylation epigenotypes in human colorectal cancer
Clin. Cancer Res.
 , 
2010
, vol. 
16
 (pg. 
21
-
33
)
13
Goel
A.
Nagasaka
T.
Arnold
C.N.
Inoue
T.
Hamilton
C.
Niedzwiecki
D.
Compton
C.
Mayer
R.J.
Goldberg
R.
Bertagnolli
M.M.
, et al.  . 
The CpG island methylator phenotype and chromosomal instability are inversely correlated in sporadic colorectal cancer
Gastroenterology
 , 
2007
, vol. 
132
 (pg. 
127
-
138
)
14
Kuhmann
C.
Weichenhan
D.
Rehli
M.
Plass
C.
Schmezer
P.
Popanda
O.
DNA methylation changes in cells regrowing after fractioned ionizing radiation
Radiother. Oncol.
 , 
2011
, vol. 
101
 (pg. 
116
-
121
)
15
Wood
R.D.
Mitchell
M.
Lindahl
T.
Human DNA repair genes, 2005
Mutat. Res.
 , 
2005
, vol. 
577
 (pg. 
275
-
283
)
16
Lord
C.J.
McDonald
S.
Swift
S.
Turner
N.C.
Ashworth
A.
A high-throughput RNA interference screen for DNA repair determinants of PARP inhibitor sensitivity
DNA Repair (Amst.)
 , 
2008
, vol. 
7
 (pg. 
2010
-
2019
)
17
Toita
N.
Hatano
N.
Ono
S.
Yamada
M.
Kobayashi
R.
Kobayashi
I.
Kawamura
N.
Okano
M.
Satoh
A.
Nakagawa
A.
, et al.  . 
Epstein-Barr virus-associated B-cell lymphoma in a patient with DNA ligase IV (LIG4) syndrome
Am. J. Med. Genet. A
 , 
2007
, vol. 
143
 (pg. 
742
-
745
)
18
Economopoulou
P.
Pappa
V.
Kontsioti
F.
Papageorgiou
S.
Foukas
P.
Liakata
E.
Economopoulou
C.
Vassilatou
D.
Ioannidou
E.D.
Chondropoulos
S.
, et al.  . 
Expression analysis of proteins involved in the nonhomologous end joining DNA repair mechanism, in the bone marrow of adult de novo myelodysplastic syndromes
Ann. Hematol.
 , 
2010
, vol. 
89
 (pg. 
233
-
239
)
19
d'Adda di Fagagna
F.
Hande
M.P.
Tong
W.M.
Roth
D.
Lansdorp
P.M.
Wang
Z.Q.
Jackson
S.P.
Effects of DNA nonhomologous end-joining factors on telomere length and chromosomal stability in mammalian cells
Curr. Biol.
 , 
2001
, vol. 
11
 (pg. 
1192
-
1196
)
20
Ferguson
D.O.
Sekiguchi
J.M.
Chang
S.
Frank
K.M.
Gao
Y.
DePinho
R.A.
Alt
F.W.
The nonhomologous end-joining pathway of DNA repair is required for genomic stability and the suppression of translocations
Proc. Natl.Acad. Sci. USA
 , 
2000
, vol. 
97
 (pg. 
6630
-
6633
)
21
Suehiro
Y.
Wong
C.W.
Chirieac
L.R.
Kondo
Y.
Shen
L.
Webb
C.R.
Chan
Y.W.
Chan
A.S.
Chan
T.L.
Wu
T.T.
, et al.  . 
Epigenetic–genetic interactions in the APC/WNT, RAS/RAF, and P53 pathways in colorectal carcinoma
Clin. Cancer Res.
 , 
2008
, vol. 
14
 (pg. 
2560
-
2569
)
22
Gundem
G.
Perez-Llamas
C.
Jene-Sanz
A.
Kedzierska
A.
Islam
A.
Deu-Pons
J.
Furney
S.J.
Lopez-Bigas
N.
IntOGen: integration and data mining of multidimensional oncogenomic data
Nat. Methods
 , 
2010
, vol. 
7
 (pg. 
92
-
93
)
23
Lao
V.V.
Grady
W.M.
Epigenetics and colorectal cancer
Nat. Rev. Gastroenterol. Hepatol.
 , 
2011
, vol. 
8
 (pg. 
686
-
700
)
24
Jass
J.R.
Classification of colorectal cancer based on correlation of clinical, morphological and molecular features
Histopathology
 , 
2007
, vol. 
50
 (pg. 
113
-
130
)
25
Samowitz
W.S.
Albertsen
H.
Sweeney
C.
Herrick
J.
Caan
B.J.
Anderson
K.E.
Wolff
R.K.
Slattery
M.L.
Association of smoking, CpG island methylator phenotype, and V600E BRAF mutations in colon cancer
J. Natl. Cancer Inst.
 , 
2006
, vol. 
98
 (pg. 
1731
-
1738
)
26
Limsui
D.
Vierkant
R.A.
Tillmans
L.S.
Wang
A.H.
Weisenberger
D.J.
Laird
P.W.
Lynch
C.F.
Anderson
K.E.
French
A.J.
Haile
R.W.
, et al.  . 
Cigarette smoking and colorectal cancer risk by molecularly defined subtypes
J. Natl. Cancer Inst.
 , 
2010
, vol. 
102
 (pg. 
1012
-
1022
)
27
van Gent
D.C.
van der Burg
M.
Non-homologous end-joining, a sticky affair
Oncogene
 , 
2007
, vol. 
26
 (pg. 
7731
-
7740
)
28
Frank
K.M.
Sekiguchi
J.M.
Seidl
K.J.
Swat
W.
Rathbun
G.A.
Cheng
H.L.
Davidson
L.
Kangaloo
L.
Alt
F.W.
Late embryonic lethality and impaired V(D)J recombination in mice lacking DNA ligase IV
Nature
 , 
1998
, vol. 
396
 (pg. 
173
-
177
)
29
Barnes
D.E.
Stamp
G.
Rosewell
I.
Denzel
A.
Lindahl
T.
Targeted disruption of the gene encoding DNA ligase IV leads to lethality in embryonic mice
Curr. Biol.
 , 
1998
, vol. 
8
 (pg. 
1395
-
1398
)
30
Karanjawala
Z.E.
Grawunder
U.
Hsieh
C.L.
Lieber
M.R.
The nonhomologous DNA end joining pathway is important for chromosome stability in primary fibroblasts
Curr. Biol.
 , 
1999
, vol. 
9
 (pg. 
1501
-
1504
)
31
Rucci
F.
Notarangelo
L.D.
Fazeli
A.
Patrizi
L.
Hickernell
T.
Paganini
T.
Coakley
K.M.
Detre
C.
Keszei
M.
Walter
J.E.
, et al.  . 
Homozygous DNA ligase IV R278H mutation in mice leads to leaky SCID and represents a model for human LIG4 syndrome
Proc. Natl. Acad. Sci. USA
 , 
2010
, vol. 
107
 (pg. 
3024
-
3029
)
32
Girard
P.M.
Kysela
B.
Harer
C.J.
Doherty
A.J.
Jeggo
P.A.
Analysis of DNA ligase IV mutations found in LIG4 syndrome patients: the impact of two linked polymorphisms
Hum. Mol. Genet.
 , 
2004
, vol. 
13
 (pg. 
2369
-
2376
)
33
Chistiakov
D.A.
Voronova
N.V.
Chistiakov
A.P.
Ligase IV syndrome
Eur. J. Med. Genet.
 , 
2009
, vol. 
52
 (pg. 
373
-
378
)
34
Wang
Y.G.
Nnakwe
C.
Lane
W.S.
Modesti
M.
Frank
K.M.
Phosphorylation and regulation of DNA ligase IV stability by DNA-dependent protein kinase
J. Biol. Chem.
 , 
2004
, vol. 
279
 (pg. 
37282
-
37290
)
35
Kim
S.T.
Lim
D.S.
Canman
C.E.
Kastan
M.B.
Substrate specificities and identification of putative substrates of ATM kinase family members
J. Biol. Chem.
 , 
1999
, vol. 
274
 (pg. 
37538
-
37543
)
36
Tang
Y.C.
Amon
A.
Gene copy-number alterations: a cost-benefit analysis
Cell
 , 
2013
, vol. 
152
 (pg. 
394
-
405
)
37
Platzer
P.
Upender
M.B.
Wilson
K.
Willis
J.
Lutterbaugh
J.
Nosrati
A.
Willson
J.K.
Mack
D.
Ried
T.
Markowitz
S.
Silence of chromosomal amplifications in colon cancer
Cancer Res.
 , 
2002
, vol. 
62
 (pg. 
1134
-
1138
)
38
Knutsen
T.
Padilla-Nash
H.M.
Wangsa
D.
Barenboim-Stapleton
L.
Camps
J.
McNeil
N.
Difilippantonio
M.J.
Ried
T.
Definitive molecular cytogenetic characterization of 15 colorectal cancer cell lines
Genes Chromosomes Cancer
 , 
2010
, vol. 
49
 (pg. 
204
-
223
)
39
Cottarel
J.
Frit
P.
Bombarde
O.
Salles
B.
Negrel
A.
Bernard
S.
Jeggo
P.A.
Lieber
M.R.
Modesti
M.
Calsou
P.
A noncatalytic function of the ligation complex during nonhomologous end joining
J. Cell Biol.
 , 
2013
, vol. 
200
 (pg. 
173
-
186
)
40
Luo
X.
Burwinkel
B.
Tao
S.
Brenner
H.
MicroRNA signatures: novel biomarker for colorectal cancer?
Cancer Epidemiol. Biomarkers Prev.
 , 
2011
, vol. 
20
 (pg. 
1272
-
1286
)
41
Comprehensive molecular characterization of human colon and rectal cancer
Nature
 , 
2012
, vol. 
487
 (pg. 
330
-
337
)
42
Thompson
L.H.
Schild
D.
Homologous recombinational repair of DNA ensures mammalian chromosome stability
Mutat. Res.
 , 
2001
, vol. 
477
 (pg. 
131
-
153
)
43
Helleday
T.
Lo
J.
van Gent
D.C.
Engelward
B.P.
DNA double-strand break repair: from mechanistic understanding to cancer treatment
DNA Repair (Amst)
 , 
2007
, vol. 
6
 (pg. 
923
-
935
)
44
Iliakis
G.
Backup pathways of NHEJ in cells of higher eukaryotes: cell cycle dependence
Radiother. Oncol.
 , 
2009
, vol. 
92
 (pg. 
310
-
315
)
45
Frank
K.M.
Sharpless
N.E.
Gao
Y.
Sekiguchi
J.M.
Ferguson
D.O.
Zhu
C.
Manis
J.P.
Horner
J.
DePinho
R.A.
Alt
F.W.
DNA ligase IV deficiency in mice leads to defective neurogenesis and embryonic lethality via the p53 pathway
Mol. Cell
 , 
2000
, vol. 
5
 (pg. 
993
-
1002
)
46
Srivastava
M.
Nambiar
M.
Sharma
S.
Karki
S.S.
Goldsmith
G.
Hegde
M.
Kumar
S.
Pandey
M.
Singh
R.K.
Ray
P.
, et al.  . 
An inhibitor of nonhomologous end-joining abrogates double-strand break repair and impedes cancer progression
Cell
 , 
2012
, vol. 
151
 (pg. 
1474
-
1487
)
47
Rhee
I.
Bachman
K.E.
Park
B.H.
Jair
K.W.
Yen
R.W.
Schuebel
K.E.
Cui
H.
Feinberg
A.P.
Lengauer
C.
Kinzler
K.W.
, et al.  . 
DNMT1 and DNMT3b cooperate to silence genes in human cancer cells
Nature
 , 
2002
, vol. 
416
 (pg. 
552
-
556
)
48
Castro
F.
Dirks
W.G.
Fahnrich
S.
Hotz-Wagenblatt
A.
Pawlita
M.
Schmitt
M.
High-throughput SNP-based authentication of human cell lines
Int. J. Cancer
 , 
2013
, vol. 
132
 (pg. 
308
-
314
)
49
Schmitt
M.
Pawlita
M.
High-throughput detection and multiplex identification of cell contaminations
Nucleic Acids Res.
 , 
2009
, vol. 
37
 pg. 
e119
 
50
Schilling
E.
Rehli
M.
Global, comparative analysis of tissue-specific promoter CpG methylation
Genomics
 , 
2007
, vol. 
90
 (pg. 
314
-
323
)
51
Sonnet
M.
Baer
C.
Rehli
M.
Weichenhan
D.
Plass
C.
Enrichment of methylated DNA by methyl-CpG immunoprecipitation
Methods Mol. Biol.
 , 
2013
, vol. 
971
 (pg. 
201
-
212
)
52
RDCT
A Language and Environment for Statistical Computing
 , 
2011
Vienna, Austria
R Foundation for Statistical Computing
53
Ritchie
M.E.
Silver
J.
Oshlack
A.
Holmes
M.
Diyagama
D.
Holloway
A.
Smyth
G.K.
A comparison of background correction methods for two-colour microarrays
Bioinformatics
 , 
2007
, vol. 
23
 (pg. 
2700
-
2707
)
54
Tseng
G.C.
Oh
M.K.
Rohlin
L.
Liao
J.C.
Wong
W.H.
Issues in cDNA microarray analysis: quality filtering, channel normalization, models of variations and assessment of gene effects
Nucleic Acids Res.
 , 
2001
, vol. 
29
 (pg. 
2549
-
2557
)
55
Tusher
V.G.
Tibshirani
R.
Chu
G.
Significance analysis of microarrays applied to the ionizing radiation response
Proc. Natl. Acad. Sci. USA
 , 
2001
, vol. 
98
 (pg. 
5116
-
5121
)
56
Chaisaingmongkol
J.
Popanda
O.
Warta
R.
Dyckhoff
G.
Herpel
E.
Geiselhart
L.
Claus
R.
Lasitschka
F.
Campos
B.
Oakes
C.C.
, et al.  . 
Epigenetic screen of human DNA repair genes identifies aberrant promoter methylation of NEIL1 in head and neck squamous cell carcinoma
Oncogene
 , 
2012
, vol. 
31
 (pg. 
5108
-
5116
)
57
Bock
C.
Reither
S.
Mikeska
T.
Paulsen
M.
Walter
J.
Lengauer
T.
BiQ Analyzer: visualization and quality control for DNA methylation data from bisulfite sequencing
Bioinformatics
 , 
2005
, vol. 
21
 (pg. 
4067
-
4068
)
58
Bodin
L.
Beaune
P.H.
Loriot
M.A.
Determination of cytochrome P450 2D6 (CYP2D6) gene copy number by real-time quantitative PCR
J. Biomed. Biotechnol.
 , 
2005
, vol. 
2005
 (pg. 
248
-
253
)
59
Blattmann
C.
Oertel
S.
Thiemann
M.
Weber
K.J.
Schmezer
P.
Zelezny
O.
Lopez Perez
R.
Kulozik
A.E.
Debus
J.
Ehemann
V.
Suberoylanilide hydroxamic acid affects gammaH2AX expression in osteosarcoma, atypical teratoid rhabdoid tumor and normal tissue cell lines after irradiation
Strahlenther. Onkol.
 , 
2012
, vol. 
188
 (pg. 
168
-
176
)
60
Wang
H.
Zeng
Z.C.
Perrault
A.R.
Cheng
X.
Qin
W.
Iliakis
G.
Genetic evidence for the involvement of DNA ligase IV in the DNA-PK-dependent pathway of non-homologous end joining in mammalian cells
Nucleic Acids Res.
 , 
2001
, vol. 
29
 (pg. 
1653
-
1660
)
61
Shao
L.
Feng
W.
Lee
K.J.
Chen
B.P.
Zhou
D.
A sensitive and quantitative polymerase chain reaction-based cell free in vitro non-homologous end joining assay for hematopoietic stem cells
PLoS ONE
 , 
2012
, vol. 
7
 pg. 
e33499
 

Author notes

These authors contributed equally to this work.

Supplementary data