The methylation of CpG islands in the promoters is associated with loss of protein via repression of gene transcription. Several studies have demonstrated that tumour suppressor and DNA repair genes are often aberrantly hypermethylated in colorectal cancer. The present study was conducted to examine whether the methylation profile of p16INK4a and hMLH1 (human mutL homologue 1) promoters was associated with clinical features and patients’ survival in CRC (colorectal carcinoma). Aberrant methylation of p16INK4a and hMLH1 promoters was found in 47.2 and 53.4% of tumours respectively. For adjacent non-tumoral mucosa, p16INK4a was fully unmethylated in 30% of the cases, whereas hMLH1 was predominantly unmethylated (76%). Methylation of p16INK4a correlated with gender and tumour size (P=0.005 and 0.035 respectively), whereas those of hMLH1 significantly correlated with overall survival (P log rank = 0.007). Concomitant methylation of p16INK4a and hMLH1 was associated with TNM (tumour, lymph node and metastases) stage and tumour size (P=0.024 and 0.021 respectively). Our data show that loss of hMLH1 expression through aberrant methylation could be used as a marker of poor prognosis in CRC.

INTRODUCTION

CRC (colorectal cancer) remains the most prevalent gastrointestinal cancer all over the world [1]. The incidence of CRC in Tunisia was in the range of 2.5–4.5 per 100000 in 2002 [2]. In colorectal carcinogenesis, the progressive accumulation of genetic and epigenetic alterations contributes to malignant transformation of normal colonic mucosa into cancer [3,4]. Chromosomal and MSI (microsatellite instability) pathways constitute the major genetic instability events in colorectal cancer [4,5]. Hypermethylation of gene promoter regions is often observed in colorectal cancer and contributes to the silencing of a number of genes, including tumour suppressor genes, e.g. p16INK4a, hMLH1 [human MLH1 (mutL homologue 1)], MGMT (methylguanine-DNA methyltransferase) and APC (adenomatous polyposis coli) [68]. In recent years, the role of this silencing mechanism in the aetiology of colorectal cancer has increasingly been recognized. Toyota et al. [6] originally proposed that a subset of sporadic colorectal cancers displays a promoter CpG island methylator phenotype.

In the present study, we focused on two selected genes (those encoding p16INK4a and hMLH1) on the basis of their key cellular functions and involvement in carcinogenesis including CRC.

P16INK4a is a CDK4 (cyclin-dependent kinase 4) inhibitor that specifically acts at the G1/S phase of the cell cycle by negatively controlling the Rb (retinoblastoma) phosphorylation status [9]. The methylation of p16INK4a has been reported to occur at a frequency of 7–53% in malignant colorectal tumours [1013]. Other reports implicated p16INK4a methylation as being an early event during colorectal carcinogenesis, whereas others suggest that it is a late event [1416].

In addition, epigenetic alterations affect genes that are critical to the maintenance of DNA stability, such as that encoding hMLH1 [17]. Approx. 15% of all colorectal cancers show a high level of MSI, reflecting dysfunction of the post-replicative DNA mismatch repair system, mainly through the CpG methylation-mediated silencing of the hMLH1 gene [18,19].

The present study was conducted with the aim of studying the methylation frequency of p16INK4a and hMLH1 gene promoters in sporadic CRC of Tunisian patients and to investigate whether the methylation profile correlated with clinical behaviour and prognosis.

MATERIALS AND METHODS

Patient characteristics

A total of 72 primary sporadic adenocarcinomas were collected between January 2003 and December 2007 from patients who underwent radical surgical resection at the Department of Digestive Surgery of Habib Bourguiba University Hospital (Sfax, Tunisia). All patients gave informed consent prior to specimen collection according to institutional guidelines. None of the patients had preoperative or postoperative chemotherapy. Tissues were also taken from the neighbouring non-cancerous mucosa (at least 5–10 cm away from the tumorous lesions) of 20 patients. At the time of surgery, the age of patients ranged from 25 to 85 years (mean: 62.9 years) and the sex ratio was 1.5. The histological subtypes were classified using the WHO (World Health Organization) criteria [20]. The carcinomas were staged according to the TNM (tumour, lymph node and metastases) classification adopted by the American Joint Committee on Cancer [21].

DNA extraction and MSP (methylation-specific PCR)

DNA was extracted from formalin-fixed, paraffin-embedded tissues. After tumour identification on H/E (haematoxylin and eosin)-stained slides, tumoral areas were scraped from 10-mm-thick paraffin sections. The collected materials were de-waxed by washing with xylene and rinsed in ethanol. Dried tissues were digested with proteinase K in the presence of SDS at 55°C overnight, followed by phenol/chloroform extraction as described in [22].

The quantity of DNA was checked by a spectrophotometer and stored at −20°C for further use. For MSP assays, DNA samples (1–2 μg) were subjected to sodium bisulfite treatment, which specifically converts the unmethylated cytosine residues into uracil using a Methyl Detector kit as recommended by Active Motif. The bisulfite-treated DNA was then amplified using specific primers for methylated and unmethylated alleles as described in [23]. The sequences of the primers, annealing temperature and product size are listed in Table 1. PCRs were performed in 25 μl containing 0.2 μM of each primer pair, 200 μM dNTP, 2 mM MgCl2, 1× PCR buffer and 1 unit of Taq DNA polymerase (Fermentas). Amplification was carried out in a thermal cycler (Applied Biosystems) with initial denaturation at 95°C for 10 min, followed by 35 cycles (30 s at 95°C, 30 s at optimal temperature and 45 s at 72°C) and a final cycle of 10 min extension at 72°C. Fully methylated DNA (provided by Active Motif) and lymphocyte DNA from healthy individuals were used as positive controls for the methylated and unmethylated reactions respectively. A blank control without DNA was included in each PCR assay. The PCR products were electrophoresed on a 2.5% agarose gel and visualized under UV illumination after ethidium bromide staining.

Table 1
Summary of primer sequences, annealing temperature, product size and number of cycles used in the MSP assay

Abbreviations: M, methylated DNA; U, unmethylated DNA; F, forward; R, reverse; Tm, melting temperature.

GeneSequences (5′–3′)Tm (°C)Size (bp)PCR cycles
P16INK4a (U) TTATTAGAGGGTGGGGTGGATTGTCA 60 151 35 
 ACCCCAAACCACAACCATAA    
p16INK4a (M) TTATTAGAGGGTGGGGCGGATCGC 65 150 35 
 GACCCCGAACCGCGACCGTAA    
hMLH1 (U) TTTTGATGTAGATGTTTTATTAGGGTTGT 60 114 35 
 ACCACCTCATCATAACTACCCACA    
hMLH1 (M) ACGTAGACGTTTTATTAGGGTCGC 58 110 35 
 CCTCATCGTAACTACCCGCG    
GeneSequences (5′–3′)Tm (°C)Size (bp)PCR cycles
P16INK4a (U) TTATTAGAGGGTGGGGTGGATTGTCA 60 151 35 
 ACCCCAAACCACAACCATAA    
p16INK4a (M) TTATTAGAGGGTGGGGCGGATCGC 65 150 35 
 GACCCCGAACCGCGACCGTAA    
hMLH1 (U) TTTTGATGTAGATGTTTTATTAGGGTTGT 60 114 35 
 ACCACCTCATCATAACTACCCACA    
hMLH1 (M) ACGTAGACGTTTTATTAGGGTCGC 58 110 35 
 CCTCATCGTAACTACCCGCG    

MSI analysis

Genomic DNA extracted from normal and tumour tissue samples was amplified by PCR at the five-microsatellite Bethesda panel (D2S123, D5S346, D17S250, BAT26 and BAT25).

PCRs were performed in a total volume of 25 μl containing 1 × buffer, 0.125 mM deoxynucleoside triphosphate, 0.2 μM of each primer and 0.25 unit of Taq DNA polymerase (Fermentas). The PCR conditions were described previously [24]. After PCR, 2 μl of the product was mixed with 8 μl of loading buffer containing formamide and Rox size standards. This mixture was denatured at 95°C for 2 min and cooled on ice before loading on to an ABI PRISM 3100 sequencer (Applied Biosystems). The results were analysed using GeneScan software (Applied Biosystems).

The tumours were classified into the following [25]: MSI-H (high degree of MSI) if two or more of the five markers showed MSI, MSI-L (low degree of MSI) if only one of the five markers did and MSS (microsatellite stable) if any marker showed instability. For the purposes of the present study, the MSS and MSI-L cases were considered together.

BRAF (encoding B-Raf kinase) V600E mutation detection

To determine the BRAF V600E mutations, exon 15 of the BRAF gene was amplified using the forward primer 5′-TCATAATGCTTGCTCTGATAGGA-3′ and the reverse primer 5′-CTTTCTAGTAACTCAGCAGC-3V. Amplifications were carried out using Taq DNA polymerase (Fermentas) and a PCR profile consisting of a 10 min initial denaturation at 95°C followed by 40 cycles of 20 s at 94°C, 20 s at 60°C and 30 s at 72°C with a 10 min final extension at 72°C. Mutations were verified by sequencing.

RNA extraction and RT–PCR (reverse transcription–PCR)

Total RNA was isolated from fresh frozen tissues with TRIzol® according to the manufacturer's protocol (Invitrogen). First-strand cDNA synthesis was performed on total RNA, previously treated with DNase I (Amersham Biosciences), using 2 μg of RNA, 0.5 μg of oligo(dT), 2 mM dNTP, 0.5 unit/μl RNase inhibitor (Amersham Biosciences), 4 μl of 5 × RT buffer and 200 units of MMLV (Moloney-murine-leukaemia virus) reverse transcriptase (Invitrogen). The reaction mixture was incubated at 37°C for 1 h, followed by 70°C for 10 min. cDNA (3 μl) was used as a template in PCR using specific primers for p16INK4a, forward: 5′-GGGGGCACCAGAGGCAGT-3′, reverse: 5′-GGTTGTGGCGGGGGCAGTT-3′ (160 bp); and hMLH1, forward: 5′-GTGCTG GCAATCAAGGGACCC-3′, reverse 5′-CACGGTTGAGGCATTGGGTAG-3′ (214 bp) and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) as an endogenous control: forward, 5′-ACCCACTCCTCCACCTTT-3′ and 5′-ACCCACTCCTCCACCTTTG-3′ (188 bp). After 5 min of initial denaturation, 35 cycles of 94°C for 45 s, 58°C for 45 s and 72°C for 1 min were performed. The PCR products were analysed on a 2% agarose gel, stained with ethidium bromide and visualized under UV light.

Statistical analysis

Statistical analyses were performed using SPSS 13.0 statistical software for Windows. The two-sided χ2 test or Fisher's exact test was used to determine associations between the methylation status of p16INK4a and hMLH1 with various clinico-pathological features. Correlation with overall survival was performed using the Kaplan–Meier survival plots and the significance was tested using the log-rank test. The Cox proportional hazards model was used for multivariate analysis of prognostic factors. P<0.05 was considered statistically significant.

RESULTS

Methylation profiles of p16INK4a and hMLH1 gene promoters

To investigate the epigenetic changes in p16INK4a and hMLH1, we analysed by MSP the promoter methylation status in 72 CRC specimens and 20 normal adjacent mucosae. Figure 1 shows representative examples of MSP results. It is worth noting that, in some samples, we observed both the methylated and unmethylated bands. This profile, termed ‘hemimethylated’, could be explained either by (i) the presence of non-malignant cells in addition to the tumour cells in the biopsies or (ii) cellular heterogeneity within the same tumour knowing that DNA methylation is a progressive process during carcinogenesis. The p16INK4a and hMLH1 gene promoters were methylated in 47.2 and 53.4% of the cases respectively. In contrast, in normal adjacent mucosa, 14 out of 20 specimens displayed the hemimethylated pattern and six samples were totally unmethylated for p16INK4a, whereas hMLH1 was predominantly unmethylated (13 out of 20 cases). These data suggest that methylation of p16INK4a occurs early during CRC carcinogenesis.

Analysis of p16INK4a and hMLH1 promoter methylation by MSP in colorectal carcinoma (T12–T15) and the corresponding normal colon mucosa (N12–N15)

Figure 1
Analysis of p16INK4a and hMLH1 promoter methylation by MSP in colorectal carcinoma (T12–T15) and the corresponding normal colon mucosa (N12–N15)

PCR products in the lanes marked U showed the presence of the unmethylated templates of each gene, whereas the products in the lanes marked M indicate the presence of methylated templates. Abbreviations: NL, DNA from normal lymphocytes used as a positive control for unmethylated alleles; FM, commercial fully methylated DNA used as a positive control for methylated alleles; NC, PCR negative control; and L, molecular-mass marker (100 bp DNA ladder; Fermentas).

Figure 1
Analysis of p16INK4a and hMLH1 promoter methylation by MSP in colorectal carcinoma (T12–T15) and the corresponding normal colon mucosa (N12–N15)

PCR products in the lanes marked U showed the presence of the unmethylated templates of each gene, whereas the products in the lanes marked M indicate the presence of methylated templates. Abbreviations: NL, DNA from normal lymphocytes used as a positive control for unmethylated alleles; FM, commercial fully methylated DNA used as a positive control for methylated alleles; NC, PCR negative control; and L, molecular-mass marker (100 bp DNA ladder; Fermentas).

Association between methylation and clinicopathological features

Table 2 summarizes the association between the methylation status of p16INK4a and hMLH1 and clinico-pathological features. The most consistent association was found between p16INK4a methylation and gender with a significant higher frequency of methylation in women compared with men (69% versus 34%, P=0.005, Table 2). In addition, a correlation was found with tumour size (P=0.035, Table 2). Aberrant methylation of hMLH1 was more often observed in rectosigmoide than in colon and rectum, but the difference was not significant statistically (P=0.063, Table 2). Among the 72 CRC specimens, 18 (25%) are methylated in both promoters, whereas 18 specimens are totally unmethylated and 36 tumours exhibit the hemimethylated pattern. Concomitant methylation of p16INK4a and hMLH1 correlated with TNM and tumour size (P=0.024 and 0.021 respectively, Table 2).

Table 2
Association between methylation of p16INK4a and hMLH1 gene promoters and clinico-pathological features

Values are proportions (percentage). Abbreviations: M, methylated profile; U, unmethylated profile.

p16INK4ahMLH1p16INK4a/hMLH1
ParameterTotal (n)MUMUMU
Subjects 72 34 (47.2) 38 (52.8) 38 (52.8) 34 (47.2) 18 (25) 54 (75) 
Gender        
 Men 46 16 (34.8) 30 (65.2) 23 (50) 23 (50) 9 (19.6) 37 (80.4) 
 Women 26 18 (69.2) 8 (30.8) 15 (57.7) 11 (42.3) 9 (34.6) 17 (65.4) 
P value  0.005  0.530  0.157  
Age        
 <60 years 20 9 (45) 11 (55) 10 (50) 10 (50) 4 (20) 16 (80) 
 ≥60 years 52 25 (48.1) 27 (51.9) 28 (53.8) 24 (46.2) 14 (26.9) 38 (73.1) 
P value  0.815  0.770  0.543  
TNM        
 I and II 36 21 (58.3) 15 (41.7) 18 (50) 18 (50) 14 (38.9) 22 (61.1) 
 III 20 7 (35) 13 (65) 11 (55) 9 (45) 2 (10) 18 (90) 
 IV 16 6 (37.5) 10 (62.5) 9 (56.3) 7 (43.8) 2 (12.5) 14 (87.5) 
P value  0.166  0.892  0.024  
Location        
 Rectum 27 13 (48.1) 14 (51.9) 10 (37) 17 (63) 5 (18.5) 22 (81.5) 
 Colon 39 18 (46.2) 21 (53.8) 23 (59) 16 (41) 10 (25.6) 29 (74.4) 
 Rectosigmoide 3 (50) 3 (50) 5 (83.3) 1 (16.7) 3 (50) 3 (50) 
P value  0.977  0.063  0.271  
Tumour grade        
 Well 32 16 (50) 16 (50) 19 (59.4) 13 (40.6) 11 (34.4) 21 (65.6) 
 Moderately 37 17 (45.9) 20 (54.1) 16 (43.2) 21 (56.8) 6 (16.2) 31 (83.8) 
 Poor 1 (33.3) 2 (66.7) 3 (100) 0 (0) 1 (33.3) 2 (66.7) 
P value  0.837  0.101  0.209  
Tumour size        
 <5 cm 35 21 (60) 14 (40) 18 (51.4) 17 (48.6) 13 (37.1) 22 (62.9) 
 ≥5 cm 37 13 (35.1) 24 (64.9) 20 (54.1) 17 (45.9) 5 (13.5) 32 (86.5) 
P value  0.035  0.824  0.021  
MSI        
 MSI-H 14 5 (35.7) 9 (64.3) 10 (71.4) 4 (28.6) 5 (35.7) 9 (64.3) 
 MSI-L, MSS 27 13 (48.1) 14 (51.9) 12 (44.4) 15 (55.6) 4 (14.8) 23 (85.2) 
P value  0.447  0.100  0.125  
BRAF        
 Mutant 2 (40) 3 (60) 3 (60) 2 (40) 2 (40) 3 (60) 
 Wild-type 61 28 (45.9) 33 (54.1) 34 (55.7) 27 (44.3) 15 (24.6) 46 (75.4) 
P value  0.799  0.854  0.449  
p16INK4ahMLH1p16INK4a/hMLH1
ParameterTotal (n)MUMUMU
Subjects 72 34 (47.2) 38 (52.8) 38 (52.8) 34 (47.2) 18 (25) 54 (75) 
Gender        
 Men 46 16 (34.8) 30 (65.2) 23 (50) 23 (50) 9 (19.6) 37 (80.4) 
 Women 26 18 (69.2) 8 (30.8) 15 (57.7) 11 (42.3) 9 (34.6) 17 (65.4) 
P value  0.005  0.530  0.157  
Age        
 <60 years 20 9 (45) 11 (55) 10 (50) 10 (50) 4 (20) 16 (80) 
 ≥60 years 52 25 (48.1) 27 (51.9) 28 (53.8) 24 (46.2) 14 (26.9) 38 (73.1) 
P value  0.815  0.770  0.543  
TNM        
 I and II 36 21 (58.3) 15 (41.7) 18 (50) 18 (50) 14 (38.9) 22 (61.1) 
 III 20 7 (35) 13 (65) 11 (55) 9 (45) 2 (10) 18 (90) 
 IV 16 6 (37.5) 10 (62.5) 9 (56.3) 7 (43.8) 2 (12.5) 14 (87.5) 
P value  0.166  0.892  0.024  
Location        
 Rectum 27 13 (48.1) 14 (51.9) 10 (37) 17 (63) 5 (18.5) 22 (81.5) 
 Colon 39 18 (46.2) 21 (53.8) 23 (59) 16 (41) 10 (25.6) 29 (74.4) 
 Rectosigmoide 3 (50) 3 (50) 5 (83.3) 1 (16.7) 3 (50) 3 (50) 
P value  0.977  0.063  0.271  
Tumour grade        
 Well 32 16 (50) 16 (50) 19 (59.4) 13 (40.6) 11 (34.4) 21 (65.6) 
 Moderately 37 17 (45.9) 20 (54.1) 16 (43.2) 21 (56.8) 6 (16.2) 31 (83.8) 
 Poor 1 (33.3) 2 (66.7) 3 (100) 0 (0) 1 (33.3) 2 (66.7) 
P value  0.837  0.101  0.209  
Tumour size        
 <5 cm 35 21 (60) 14 (40) 18 (51.4) 17 (48.6) 13 (37.1) 22 (62.9) 
 ≥5 cm 37 13 (35.1) 24 (64.9) 20 (54.1) 17 (45.9) 5 (13.5) 32 (86.5) 
P value  0.035  0.824  0.021  
MSI        
 MSI-H 14 5 (35.7) 9 (64.3) 10 (71.4) 4 (28.6) 5 (35.7) 9 (64.3) 
 MSI-L, MSS 27 13 (48.1) 14 (51.9) 12 (44.4) 15 (55.6) 4 (14.8) 23 (85.2) 
P value  0.447  0.100  0.125  
BRAF        
 Mutant 2 (40) 3 (60) 3 (60) 2 (40) 2 (40) 3 (60) 
 Wild-type 61 28 (45.9) 33 (54.1) 34 (55.7) 27 (44.3) 15 (24.6) 46 (75.4) 
P value  0.799  0.854  0.449  

Association of p16INK4a and hMLH1 methylation with patients’ survival

The survival rate was available for only 38 patients and the follow-up time ranged from 0.5 to 72 months. As shown in Figure 2, patients with the methylated p16INK4a gene promoter have a slightly prolonged survival rate when compared with those with unmethylated profile (P log rank = 0.185, Figure 2A), whereas the Kaplan–Meier analysis shows that the methylation status of hMLH1 is predictive of poor survival. Indeed, patients with the unmethylated hMLH1 gene promoter have a significantly prolonged survival rate when compared with those with a methylated profile (P log rank = 0.007, Figure 2B), suggesting that it could be an important prognostic factor of CRC; nevertheless, this finding should be confirmed on larger series. Furthermore, in the multivariate analysis, factors such as age, gender, tumour size, lymph node metastasis and methylation profile of p16INK4a and hMLH1 were included in the Cox proportional hazards model. The results indicated that, in addition to tumour size, the methylation of hMLH1 was an independent prognostic factor (Table 3).

Kaplan–Meier survival curves correlating overall survival with p16INK4a (A) and hMLH1 (B) methylated and unmethylated profiles and MSI status (C)

Figure 2
Kaplan–Meier survival curves correlating overall survival with p16INK4a (A) and hMLH1 (B) methylated and unmethylated profiles and MSI status (C)
Figure 2
Kaplan–Meier survival curves correlating overall survival with p16INK4a (A) and hMLH1 (B) methylated and unmethylated profiles and MSI status (C)
Table 3
Multivariate survival analysis: the Cox proportional hazards model

Abbreviations: M, methylated profile; U, unmethylated profile; HR, hazard ratio; CI, confidence interval.

ParameterP valueHR95% CI
Gender (men compared with women) 0.231 2.321 0.585–9.217 
Age (<60 compared with ≥60) 0.172 2.503 0.670–9.348 
Tumour size (<5 cm compared with >5 cm) 0.059 3.264 0.95–11.152 
Lymph node (absent compared with present) 0.277 1.960 0.582–6.593 
p16INK4a methylation (M compared with U) 0.144 3.308 0.663–16.497 
hMLH1 methylation (M compared with U) 0.007 0.120 0.026–0.559 
ParameterP valueHR95% CI
Gender (men compared with women) 0.231 2.321 0.585–9.217 
Age (<60 compared with ≥60) 0.172 2.503 0.670–9.348 
Tumour size (<5 cm compared with >5 cm) 0.059 3.264 0.95–11.152 
Lymph node (absent compared with present) 0.277 1.960 0.582–6.593 
p16INK4a methylation (M compared with U) 0.144 3.308 0.663–16.497 
hMLH1 methylation (M compared with U) 0.007 0.120 0.026–0.559 

Relationship between p16INK4a and hMLH1 expressions with the methylation profile

In attempt to validate the effect of aberrant methylation on the expressions of p16INK4a and hMLH1, we performed RT–PCR analysis on 20 available fresh frozen tumours with methylated, unmethymated or hemimethylated profiles using GAPDH as the control gene (Figure 3). The association between promoter methylation status and RT–PCR results did not reach statistical significance probably because of the small number of cases. However, we showed that the two tumours with the p16INK4a methylated DNA failed to express the corresponding mRNA, whereas this transcript was detected in six out of seven cases with the unmethylated pattern. For the samples with partial methylation, the p16INK4a mRNA was detected in seven out of 11 cases, suggesting that the methylation at the CpG islands in the p16INK4a promoter is responsible, at least in part, for the loss of mRNA expression (Figure 3A). On the other hand, among the 20 samples tested, the five tumours displaying the unmethylated hMLH1 DNA did not express the transcript, whereas the hMLH1 mRNA was observed in 12 out of 15 cases in tumours with hemimethylated profile (Figure 3B).

Representative results of MSP and RT–PCR for p16INK4a (A) and hMLH1 (B) in primary CRC

Figure 3
Representative results of MSP and RT–PCR for p16INK4a (A) and hMLH1 (B) in primary CRC

GAPDH was used as an endogenous control. Abbreviation: T, tumour tissues. Samples T7, T8 and T9, unmethylated; T6, methylated; and T5, T10 and T11, hemimethylated. Lanes U and M correspond to unmethylated and methylated DNAs respectively. H2O, negative control for MSP; L, 100 bp DNA ladder (Fermentas).

Figure 3
Representative results of MSP and RT–PCR for p16INK4a (A) and hMLH1 (B) in primary CRC

GAPDH was used as an endogenous control. Abbreviation: T, tumour tissues. Samples T7, T8 and T9, unmethylated; T6, methylated; and T5, T10 and T11, hemimethylated. Lanes U and M correspond to unmethylated and methylated DNAs respectively. H2O, negative control for MSP; L, 100 bp DNA ladder (Fermentas).

BRAF mutation and MSI status

BRAF V600E mutation was detected in five out of 66 specimens (7.5%). No association was found between BRAF mutation and aberrant methylation of individual gene promoters. However, concomitant methylation of p16INK4a and hMLH1 is likely to be more prevalent in tumours harbouring the V600E mutation (40% versus 24.6%), but statistical significance was not reached (Table 2). MSI analysis was carried out on 41 tumours; among them 14 (35.7%) were MSI-H and 27 (48.1%) were MSS. No association was found between the MSI status and aberrant methylation of p16INK4a and hMLH1; however, hypermethylation of hMLH1 was more often observed in MSI-H tumours (Table 2). On the other hand, we assessed the influence of MSI on patients’ survival and found that, compared with patients with MSS tumours, those with MSI-H tumours have a shorter survival rate as shown in Figure 2(C).

DISCUSSION

Aberrant methylation of CpG islands in gene promoters is a frequent event in cancer, and it can affect critical steps in the process of carcinogenesis such as DNA repair and cell cycle control [8,26,27]. In the present study, aberrant methylation of p16INK4a and hMLH1 gene promoters was found in 47.2 and 53.4% of tumours respectively. In CRC, the methylation status of p16INK4a and hMLH1 gene promoters has been extensively studied and the percentage of aberrant methylation varies widely from one study to another [2830]. Methodological differences may be one explanation for such variability, but it could also be attributed to ethnic variations caused by genetic and/or dietary factors. To the best of our knowledge, this is the first report on p16INK4a and hMLH1 methylation profile in Tunisian patients with CRC and their correlation with major clinical features and survival. In the present study, we showed that methylation of p16INK4a correlated with gender and tumour size (P=0.005 and 0.035 respectively). Other authors have found that p16INK4a methylation correlated with mucinous histology, more advanced T stage, lymph node involvement and poor prognosis [13,16,31]. No statistical significance was found between hMLH1 methylation and any of the clinical features except for a tendency of association with tumour location (P=0.063), which is in line with a previous report showing that methylation of hMLH1 is more often seen in proximal than in distal tumours [32].

The correlation between p16INK4a hypermethylation and prognosis in patients with CRC is controversial [3337]. In the present study, the Kaplan–Meier plot showed that patients with methylated p16INK4a promoter have a slightly more prolonged overall survival rate than those exhibiting unmethylated DNA (P log rank = 0.185). In contrast, methylation of hMLH1 correlated significantly with overall survival since patients with unmethylated hMLH1 gene promoter have a significant prolonged survival rate compared with those with the methylated profile (P log rank = 0.007), suggesting that it could be an important prognostic factor of CRC. The Cox regression model revealed that in addition to tumour size, the methylation of hMLH1 could have a prognostic value; nevertheless, this finding should be confirmed since the follow-up data were available for only 38 among 72 patients included in the present study.

In summary, our results show that, in sporadic colorectal cancers, methylation of p16INK4 is frequently involved in gene silencing. Furthermore, methylation of hMLH1 strongly correlated with shortened patients’ survival. These observations have an important impact on the therapy of colorectal cancer, since drug resistance can be mediated by loss of mismatch repair, and recovering such a key function by the action of demethylated agents might lead to the overcoming of the drug resistance effect.

B-Raf is related to the Ras/Raf/MEK [MAPK (mitogen-activated protein kinase)/extracellular-signal-regulated kinase kinase]/MAPK signal transduction pathway, and oncogenic mutations in BRAF, mainly the V600E mutation, have been reported in colon cancer [38,39]. Previous studies suggested that BRAF mutation is closely associated with sporadic MSI-positive CRCs, but recent data indicate that BRAF mutation is associated with CIMP (CpG island methylator phenotype) rather than MSI [4042]. The present study supports somehow the association of BRAF mutation with promoter methylation, because the BRAF mutation rate is slightly higher in tumours exhibiting concomitant methylation of p16INK4a and hMLH1.

The prognostics implications of the MSI status were investigated in CRC patients. Ogino et al. [42] showed that patients with MSS tumours experienced a significant reduction in colon cancer-specific mortality when compared with those with MSI-H tumours. It was also reported that the MSI-negative subtype associated with CIMP-positive showed worse clinical outcome [43]. In contrast, we showed that patients with MSS tumours had better prognosis than those with MSI-H tumour; however, due to the limited number of patients, these data should be confirmed in further studies.

Abbreviations

     
  • CIMP

    CpG island methylator phenotype

  •  
  • CRC

    colorectal carcinoma

  •  
  • GAPDH

    glyceraldehyde 3-phosphate dehydrogenase

  •  
  • MAPK

    mitogen-activated protein kinase

  •  
  • MLH1

    mutL homologue 1

  •  
  • hMLH1

    human MLH1

  •  
  • MSP

    methylation-specific PCR

  •  
  • MSI

    microsatellite instability

  •  
  • MSI-H

    high degree of MSI

  •  
  • MSI-L

    low degree of MSI

  •  
  • MSS

    microsatellite stable: MMLV, Moloney-murine-leukaemia virus

  •  
  • RT–PCR

    reverse transcription–PCR

  •  
  • TNM

    tumour, lymph node and metastases

AUTHOR CONTRIBUTION

Imen Miladi-Abdennadher performed RT–PCR and MSI experiments, carried out the molecular analysis and wrote the manuscript. Rania Abdelmaksoud-Damak optimized the MSP conditions. Lobna Ayadi, Abdelmajid Khabir and Tahia Sellami-Boudawara performed the histological examination. Foued Frikha, Lamia Kallel and Mounir Frikha recruited patients and helped in consulting the patients’ files. Ali Gargouri and Raja Mokdad-Gargouri participated in the data analysis and the interpretation of results, and helped in drafting the manuscript. All authors have contributed to and approved the manuscript.

We thank the technicians at Centre Hospitalo-Universitaire Habib Bourguiba (Sfax, Tunisia) for assistance.

FUNDING

This work was supported by the Ministère de l’Enseignement Supérieur et de la Recherche Scientifique Tunisien.

References

References
1
Parkin
 
D. M.
Bray
 
F.
 
Global cancer statistics
CA Cancer J. Clin.
2005
, vol. 
55
 (pg. 
74
-
108
)
2
Hsairi
 
M.
Fakhfakh
 
R.
Ben Abdallah
 
M.
Jlidi
 
R.
Sellami
 
A.
Zheni
 
S.
Hmissa
 
S.
Achour
 
N.
Nacef
 
T.
 
Assessment of cancer incidence in Tunisia 1993–1997
Tun. Méd.
2002
, vol. 
80
 (pg. 
57
-
64
)
3
Chung
 
D. C.
 
The genetic basis of colorectal cancer: insights into critical pathways of tumorigenesis
Gastroenterology
2000
, vol. 
119
 (pg. 
854
-
865
)
4
Baylin
 
S. B.
Ohm
 
J. E.
 
Epigenetic gene silencing in cancer – a mechanism for early oncogenic pathway addiction
Nat. Rev. Cancer
2006
, vol. 
6
 (pg. 
107
-
111
)
5
Lengauer
 
C.
Kinzler
 
K. W.
Vogelstein
 
B.
 
Genetic instability in colorectal cancers
Nature
1997
, vol. 
386
 (pg. 
623
-
627
)
6
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. U.S.A.
1999
, vol. 
96
 (pg. 
8681
-
8686
)
7
Kondo
 
Y.
Issa
 
J. P.
 
Epigenetic changes in colorectal cancer
Cancer Med.
2004
, vol. 
23
 (pg. 
29
-
39
)
8
Kim
 
Y. S.
Deng
 
G.
 
Epigenetic changes (aberrant DNA methylation) in colorectal neoplasia
Gut Liver
2007
, vol. 
1
 (pg. 
1
-
11
)
9
Liggett
 
W. H.
Sidransky
 
J. D.
 
Role of the p16 tumor suppressor gene in cancer
J. Clin. Oncol.
1998
, vol. 
16
 (pg. 
1197
-
1206
)
10
Liang
 
J. T.
Chang
 
K. J.
Chen
 
J. C.
Lee
 
C. C.
Cheng
 
Y. M.
Hsu
 
H. C.
Wu
 
M. S.
Wang
 
S. M.
Lin
 
J. T.
Cheng
 
A. L.
 
Hypermethylation of the p16 gene in sporadic T3N0M0 stage colorectal cancers: association with DNA replication error and shorter survival
Oncology
1999
, vol. 
57
 (pg. 
149
-
156
)
11
Wiencke
 
J. K.
Zheng
 
S.
Lafuente
 
A.
Lafuente
 
M. J.
Grudzen
 
C.
Wrensch
 
M. R.
Miike
 
R.
Ballesta
 
A.
Trias
 
M.
 
Aberrant methylation of p16INK4a in anatomic and gender-specific subtypes of sporadic colorectal cancer
Cancer Epidemiol. Biomark. Prev.
1999
, vol. 
8
 (pg. 
501
-
506
)
12
Trzeciak
 
L.
Hennig
 
E.
Kolodziejski
 
J.
Nowacki
 
M.
Ostrowski
 
J.
 
Mutations, methylation and expression of cdkn2a/p16 gene in colorectal cancer and normal colonic mucosa
Cancer Lett.
2001
, vol. 
163
 (pg. 
17
-
23
)
13
Schneider-Stock
 
R.
Boltze
 
C.
Peters
 
B.
Höpfner
 
T.
Meyer
 
F.
Lippert
 
H.
Roessner
 
A.
 
Differences in loss of p16ink4 protein expression by promoter methylation between left- and right-sided primary colorectal carcinomas
Int. J. Oncol.
2003
, vol. 
23
 (pg. 
1009
-
1013
)
14
Ahuja
 
N.
Mohan
 
A.
Li
 
Q.
Stolker
 
J. M.
Herman
 
J. G.
Hamilton
 
S. R.
Baylin
 
S. B.
Issa
 
J. P.
 
Association between CpG island methylation and microsatellite instability in colorectal cancer
Cancer Res.
1997
, vol. 
57
 (pg. 
3370
-
3374
)
15
Ramirez
 
N.
Bandres
 
E.
Navarro
 
A.
Pons
 
A.
Jansa
 
S.
Moreno
 
I.
Martınez-Rodenas
 
F.
Zarate
 
R.
Bitarte
 
N.
Monzo
 
M.
Garcıa-Foncillas
 
J.
 
Epigenetic events in normal colonic mucosa surrounding colorectal cancer lesions
Eur. J. Cancer
2008
, vol. 
44
 (pg. 
2689
-
2695
)
16
Wettergren
 
Y.
Odin
 
E.
Nilsson
 
S.
Göran
 
C.
Gustavsson
 
B.
 
P16INK4a gene promoter hypermethylation in mucosa as a prognostic factor for patients with colorectal cancer
Mol. Med.
2008
, vol. 
14
 (pg. 
412
-
417
)
17
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
)
18
Thibodeau
 
S. N.
Bren
 
G.
Shaid
 
D.
 
Microsatellite instability in cancer of proximal colon
Science
1993
, vol. 
260
 (pg. 
816
-
819
)
19
Lin
 
S. Y.
Yeh
 
K. T.
Chen
 
W. T.
Chen
 
H. C.
Chen
 
S. T.
Chiou
 
H. Y.
Chang
 
J. G.
 
Promoter CpG methylation of tumor suppressor genes in colorectal cancer and its relationship to clinical features
Oncol. Rep.
2004
, vol. 
11
 (pg. 
341
-
434
)
20
Wittekind
 
C.
Greene
 
F. L.
 
TNM Atlas: Illustrated Guide to the TNM Classification of Malignant Tumours
2005
6th edn
New York
John Wiley & Son
21
Aaltonen
 
L. A.
 
WHO Classification of Tumors: Pathology and Genetics of Tumors of Digestive System
2000
Lyon
IARC Press
(pg. 
105
-
119
)
22
Sambrook
 
J.
Russell
 
D. W.
 
Molecular Cloning. A Laboratory Manual
2001
Cold Spring Harbor, NY
Cold Spring Harbor Laboratory Press
23
Herman
 
J. G.
Graff
 
J. R.
Myöhänen
 
S.
Nelkin
 
B. D.
Baylin
 
S. B.
 
Methylation-specific PCR: a novel PCR assay for methylation status of CpG islands
Proc. Natl. Acad. Sci. U.S.A.
1996
, vol. 
93
 (pg. 
9821
-
9826
)
24
Dietmaier
 
W.
Wallinger
 
S.
Bocker
 
T.
Kullmann
 
F.
Fishel
 
R.
Rüschoff
 
J.
 
Diagnostic microsatellite instability: definition and correlation with mismatch repair protein expression
Cancer Res.
1997
, vol. 
57
 (pg. 
4749
-
4756
)
25
Boland
 
C. R.
Thibodeau
 
S. N.
Hamilton
 
S. R.
Sidransky
 
D.
Eshleman
 
J. R.
Burt
 
R. W.
Meltzer
 
S. J.
Rodriguez-Bigas
 
M. A.
Fodde
 
R.
Ranzani
 
G. N.
Srivastava
 
S.
 
A National Cancer Institute Workshop on microsatellite instability for cancer detection and familial predisposition: development of international criteria for the determination of microsatellite instability in colorectal cancer
Cancer Res.
1998
, vol. 
5
 (pg. 
5248
-
5257
)
26
Momparler
 
R. L.
 
Cancer epigenetics
Oncogene
2003
, vol. 
22
 (pg. 
6479
-
6483
)
27
Grady
 
W. M.
 
Epigenetic events in the colorectum and in colon cancer
Biochem. Soc. Trans.
2005
, vol. 
33
 (pg. 
684
-
688
)
28
Ahuja
 
N.
Li
 
Q.
Mohan
 
A. L.
Baylin
 
S. B.
Issa
 
J. P.
 
Aging and DNA methylation in colorectal mucosa and cancer
Cancer Res.
1998
, vol. 
58
 (pg. 
5489
-
5494
)
29
Goto
 
T.
Mizukami
 
H.
Shirahata
 
A.
Sakata
 
M.
Saito
 
M.
Ishibashi
 
K.
Kigawa
 
G.
Nemoto
 
H.
Sanada
 
Y.
Hibi
 
K.
 
Aberrant methylation of the p16 gene is frequently detected in advanced colorectal cancer
Anticancer Res.
2009
, vol. 
29
 (pg. 
275
-
277
)
30
Shannon
 
B. A.
Iacopetta
 
B. J.
 
Methylation of the hMLH1, p16, and MDR1 genes in colorectal carcinoma: associations with clinicopathological features
Cancer Lett.
2001
, vol. 
167
 (pg. 
91
-
97
)
31
Sanz-Casla
 
M. T.
Maestro
 
M. L.
Vidaurreta
 
M.
Maestro
 
C.
Arroyo
 
M.
Cerdán
 
J.
 
p16 gene methylation in colorectal tumors: correlation with clinicopathological features and prognostic value
Digest. Dis.
2005
, vol. 
23
 (pg. 
151
-
155
)
32
Deng
 
G.
Peng
 
E.
Gum
 
J.
Terdiman
 
J.
Sleisenger
 
M.
Kim
 
Y. S.
 
Methylation of hMLH1 promoter correlates with the gene silencing with a region-specific manner in colorectal cancer
Br. J. Cancer
2002
, vol. 
86
 (pg. 
574
-
579
)
33
Cheng
 
A. L.
 
Hypermethylation of the p16 gene in sporadic T3N0M0 stage colorectal cancers: association with DNA replication error and shorter survival
Oncology
1999
, vol. 
57
 (pg. 
149
-
156
)
34
Esteller
 
M.
González
 
S.
Risques
 
R. A.
Marcuello
 
E.
Mangues
 
R.
Germà
 
J. R.
Herman
 
J. G.
Capellà
 
G.
Peinado
 
M. A.
 
K-ras and p16 aberrations confer poor prognosis in human colorectal cancer
J. Clin. Oncol.
2001
, vol. 
19
 (pg. 
299
-
304
)
35
Krtolica
 
K.
Krajnovic
 
M.
Usaj-Knezevic
 
S.
Babic
 
D.
Jovanovic
 
D.
Dimitrijevic
 
B.
 
Comethylation of p16 and MGMT genes in colorectal carcinoma: correlation with clinicopathological features and prognostic value
World J. Gastroenterol.
2007
, vol. 
131
 (pg. 
187
-
194
)
36
Maeda
 
K.
Kawakami
 
K.
Ishida
 
Y.
Ishiguro
 
K.
Omura
 
K.
Watanabe
 
G.
 
Hypermethylation of the CDKN2A gene in colorectal cancer is associated with shorter survival
Oncol. Rep.
2003
, vol. 
10
 (pg. 
935
-
938
)
37
Yi
 
J.
Wang
 
Z. W.
Cang
 
H.
Chen
 
Y. Y.
Zhao
 
R.
Yu
 
B. M.
Tang
 
X. M.
 
p16 gene methylation in colorectal cancers associated with Duke's staging
World J. Gastroenterol.
2001
, vol. 
7
 (pg. 
722
-
725
)
38
Davies
 
H.
Bignell
 
G. R.
Cox
 
C.
Stephens
 
P.
Edkins
 
S.
Clegg
 
S.
Teague
 
J.
Woffendin
 
H.
Garnett
 
M. J.
Bottomley
 
W.
, et al 
Mutations of the BRAF gene in human cancer
Nature
2002
, vol. 
417
 (pg. 
949
-
954
)
39
Rajagopalan
 
H.
Bardelli
 
A.
Lengauer
 
C.
Kinzler
 
K. W.
Vogelstein
 
B.
Velculescu
 
V. E.
 
Tumorigenesis: RAF/RAS oncogenes and mismatch-repair status
Nature
2002
, vol. 
418
 pg. 
934
 
40
Oliveira
 
C.
Pinto
 
M.
Duval
 
A.
Brennetot
 
C.
Domingo
 
E.
Espín
 
E.
Armengol
 
M.
Yamamoto
 
H.
Hamelin
 
R.
Seruca
 
R.
Schwartz
 
S.
 
BRAF mutations characterize colon but not gastric cancer with mismatch repair deficiency
Oncogene
2003
, vol. 
22
 (pg. 
9192
-
9196
)
41
Cheng
 
Y. W.
Pincas
 
H.
Bacolod
 
M. D.
Schemmann
 
G.
Giardina
 
S. F.
Huang
 
J.
Barral
 
S.
Idrees
 
K.
Khan
 
S. A.
Zeng
 
Z.
, et al 
CpG island methylator phenotype associates with low-degree chromosomal abnormalities in colorectal cancer
Clin. Cancer Res.
2008
, vol. 
14
 (pg. 
6005
-
6013
)
42
Ogino
 
S.
Nosho
 
K.
Kirkner
 
G. J.
Kawasaki
 
T.
Meyerhardt
 
J. A.
Loda
 
M.
Giovannucci
 
E. L.
Fuchs
 
Ch.S.
 
CpG island methylator phenotype, microsatellite instability, BRAF mutation and clinical outcome in colon cancer
Gut
2009
, vol. 
58
 (pg. 
90
-
96
)
43
Kim
 
J. H.
Shin
 
S. H.
Kwon
 
H. J.
Cho
 
N. Y.
Kang
 
G. H.
 
Prognostic implications of CpG island hypermethylator phenotype in colorectal cancers
Virchows Arch.
2009
, vol. 
455
 (pg. 
485
-
494
)