Conceived and designed the experiments: K.Konishi YW J-PI. Performed the experiments: K.Konishi YW LS YG RC K.Kondo WC SA JJ YB ME SM YK FI MI SH. Analyzed the data: K.Konishi YW J-PI. Contributed reagents/materials/analysis tools: K.Konishi YW LS SA JJ SH J-PI. Wrote the paper: K.Konishi J-PI.
The authors have declared that no competing interests exist.
The contribution of DNA methylation to the metastatic process in colorectal cancers (CRCs) is unclear.
We evaluated the methylation status of 13 genes (MINT1, MINT2, MINT31, MLH1, p16, p14, TIMP3, CDH1, CDH13, THBS1, MGMT, HPP1 and ERα) by bisulfite-pyrosequencing in 79 CRCs comprising 36 CRCs without liver metastasis and 43 CRCs with liver metastasis, including 16 paired primary CRCs and liver metastasis. We also performed methylated CpG island amplification microarrays (MCAM) in three paired primary and metastatic cancers.
Methylation of p14, TIMP3 and HPP1 in primary CRCs progressively decreased from absence to presence of liver metastasis (13.1% vs. 4.3%; 14.8% vs. 3.7%; 43.9% vs. 35.8%, respectively) (
Most DNA methylation differences between primary CRCs and matched liver metastasis are due to random variation and an increase in DNA methylation density rather than de-novo inactivation and silencing. Thus, DNA methylation changes occur for the most part before progression to liver metastasis.
Colorectal cancers (CRCs) are the second leading cause of death from cancer and the third most commonly diagnosed cancers in the United States
Most CRCs develop in a multistep manner through the adenoma-carcinoma sequence over many years to decades
The molecular mechanisms responsible for progression to CRC metastasis are largely unknown. An early model postulated that metastasis results from rare molecular events that provide the ability to invade, disseminate and survive at distant sites
We examined 79 sporadic CRCs comprising 36 CRCs without liver metastasis (stage I–III) and 43 CRCs with liver metastasis (stage IV/liver recurrence). A metachronous liver metastasis was defined as a liver metastasis resected at least 12 months after resection of their primary CRCs, otherwise we considered a synchronous metastasis. Among the 43 patients, 16 had both primary CRC and matched liver metastasis available for evaluation. All tissue specimens were obtained from patients who had undergone surgery or endoscopic biopsy at the M.D. Anderson Cancer Center (n = 64) or at the Showa University Hospital (n = 15). We excluded patients who had syndromic familial predisposition (familial adenomatous polyposis or hereditary nonpolyposis colorectal cancer syndrome). Written informed consent was obtained from all study patients. Tissue collection and analyses were approved by the Institutional Review Board of the University of Texas M.D. Anderson Cancer Center and the Showa University School of Medicine.
We used 95 frozen samples (79 primary and 16 liver metastatic tumors) from 79 patients with CRC. Frozen tissue samples were harvested postoperatively or endoscopically and stored at −80°C. Hematoxylin and eosin (H&E) stained slides from frozen tissue blocks were reviewed by senior pathologists to evaluate the distribution of tumor cells. Representative tumor samples contained a minimum of 80% tumor cells. When colonic biopsy specimens were obtained from patients, we used chromoendoscopy with pit pattern classification to accurately distinguish between neoplastic and non-neoplastic area in the lesion
A total of nine colon cancer cell lines (SW48, RKO, SW480, HCT116, LoVo, Caco2, DLD1, and SW620) were obtained from the American Type Culture Collection (ATCC, Manassas, VA). All cells were cultured in recommended medium with 10% fetal bovine serum in a humidified atmosphere containing 5% CO2 at 37°C. Genomic DNA was extracted from these cell lines and tissue samples using a standard phenol-chloroform method.
Bisulfite treatment was performed as previously described
We evaluated the methylation status of 13 genes (MINT1, MINT2, MINT31, MLH1, p16, p14
For most analyses, we treated DNA methylation as a continuous variable in this study. To define CIMP, however, we converted the continuous values to categorical variables (positive/negative) defined by a methylation density greater than 15%. CIMP was defined using six genes (MINT1, MINT2, MINT31, p16, p14 and MLH1) as described previously
Methylated CpG island amplification (MCA) was performed for three primary CRCs and their paired liver metastatic samples randomly selected from the 16 paired primaries and liver metastases. One was stage IV and two had liver recurrence. A detailed protocol for MCA was described previously
Pyrosequencing provides a methylation level (%), which was analyzed as a continuous variable for comparison of each gene with clinicopathologic variables, and we computed mean, ranges, and 95% confidence interval (95% CI). Z-score analysis was used to normalize the methylation data of multiple genes and allow the derivation of a mean methylation score. The Z-score of methylation for each gene was calculated using the following formula: (methylation density of each sample – mean value of methylation density)/SD of methylation density. When analyzing multiple genes, we used the average of the Z-score for each gene. Differences in promoter methylation between two groups and associations between methylation and clinicopathologic characteristics were analyzed by the Mann-Whitney U test. The incidence of CIMP or gene mutation and patient characteristics were compared between tumor groups using the χ2 test or Fisher's exact test when testing small numbers of samples. All tests were two sided, and
Lowess normalization and data analysis of microarray data were performed as described previously
Each dot represents the methylation level of an individual sample. Horizontal lines represent mean methylation levels for each group. *,
Liver metastasis | |||
absence | presence | ||
(N = 36) | (N = 43) | ||
Gender | Male | 23 | 24 |
Female | 12 | 19 | |
Missing | 1 | 0 | |
Mean age | (yrs) | 66.3 | 62.2 |
(range) | (40–81) | (35–82) | |
Location | Proximal | 15 | 14 |
Distal | 15 | 29 | |
Missing | 6 | 0 | |
Stage |
1 | 4 | 0 |
(TNM) | 2 | 22 | 3 |
3 | 8 | 8 | |
4 | 0 | 32 | |
Missing |
2 | 0 | |
Liver metastasis | synchronous | NA | 36 |
metachronous | 7 | ||
Non-liver | lymph node | 8 | 30 |
metastasis | lung | 6 | |
ovary | NA | 2 | |
peritoneum | 23 | ||
brain | 1 |
*, Stage represents initial stage when primary tumors were surgically resected. Eleven cases (three stage 2 and eight stage 3 CRCs) showed liver metastases after surgery for primary tumors.
**, Two cases were known as colorectal cancers without distant metastasis. NA, not applicable.
We next classified tumors as CIMP-positive or CIMP-negative based on methylation of 2 or more CIMP-related genes (MINT1, 2, 31, p16, p14 and MLH1) and we observed no significant difference in the frequency of CIMP between primary CRCs without and with liver metastasis (15/36, 42% vs. 13/43, 30%). When we used Z-score analysis to normalize the data of CIMP-related genes, there was no significant difference in the average of Z-scores for CIMP-related genes between CRCs without and with liver metastasis (1.5 [95%CI, 2.5 to −0.7] and 1.5 [95%CI, 2.2 to 0.7],
We also evaluated DNA methylation and mutation status of primary CRCs with synchronous and metachronous liver metastasis. Only MINT1 methylation was significantly higher in primary tumors with synchronous than those with metachronous liver metastasis (15.8% [95% CI, 22.6% to 8.9%] vs. 4.3% [95% CI, 12.9% to −0.9%];
We measured DNA methylation for 13 genes in 16 paired primary and liver metastasis specimens which resulted in 205 measurement pairs (Data for THBS1 methylation in one primary and two metastatic tumors was not available). The data are shown in
A) DNA methylation status of thirteen cancer-specific or age-related genes/CpG islands in 16 primary CRCs and matched liver metastasis. Each dot represents the methylation level of an individual sample. Horizontal lines represent mean methylation levels for each group. ¶,
We next analyzed in detail the changes in methylation levels between primary and metastatic tumors (
We used MCAM in three paired primary tumors and liver metastasis (
A) The Venn diagram shows the overlap and differences in methylated genes of liver metastasis in three patients. A total number of 6528 genes were analyzed by 18340 microarray probes recognizing promoter CpG islands. B) Dendrogram and heat map overview of unsupervised hierarchical cluster analysis of DNA methylation in liver metastatic cancers of three patients. Each cell represents DNA methylation status of a gene in an individual sample. Red and green in cells reflect high and low methylation level, respectively, as shown in the scale bar (log2-transformed scale) below the matrix.
Patient | Gender | Age | Tumor | Size | Histology | Liver | Duration |
Genes methylated at |
(yrs) | location | (mm) | metastasis | (months) | liver metastasis | |||
A | F | 65 | Proximal | 40 | Mod | synchronous | 3 | 307 (4.7%) |
B | M | 73 | Proximal | 51 | Mod-Muc | metachronous | 46 | 716 (10.9%) |
C | M | 60 | Distal | 23 | Mod | metachronous | 12 | 427 (6.5%) |
, Duration between surgical resection for primary cancer and surgical resection for liver metastasis.
, A vs. B; B vs. C; A vs. C;
p<.0001. Mod, moderately differentiated adenocarcinoma; Muc, mucinous carcinoma; NA, not applicable.
To validate the results and determine whether these changes were a result of selection or random drift with time, we selected eight hypermethylated genes that had an average log2 ratio value >1.9 in all 3 tumors and analyzed them by bisulfite-pyrosequencing in 12 paired primary and liver metastases of CRCs. As shown in
Each column represents a separate gene locus indicated on top. Each row represents a primary or metastatic tumor, normal tissue type or colon cancer cell. Average methylation densities of less than 15% are shown in white, 15 to 50% in gray and over 50% in black. PBL, peripheral blood.
Promoter DNA methylation and associated silencing is a frequent and early event in colorectal carcinogenesis
When we looked at the differences in methylation between primary tumors with and without liver metastases, methylation levels of p14, TIMP3 and HPP1 progressively decreased from early-stage to late-stage disease. We have previously found that methylation of p14 and TIMP3 is the markers for predicting CIMP1
Overall, we quantitatively compared the methylation status of 21 genes (13 candidates and 8 from the microarrays) between paired primary and liver metastasis lesions. Of these, only MGMT methylation was consistently higher in the liver metastases than primary tumors. Of 16 pairs studied, five (31%) showed significantly higher MGMT methylation at the metastatic site. Of these five tumor pairs, four pairs demonstrated MGMT methylation at both sites (primary and liver metastatic tumors) with an increase in methylation density. Increased density of methylation could be explained by multiple different factors – increased proportion of methylated cells, switch from monoallelic to biallelic methylation or even differences in the degree of normal cell contamination of the tumor samples. Our data do not allow us to distinguish these possibilities and a larger series with more detailed analysis is needed to confirm our results and address the issue.
MGMT protein stoichiometrically repairs O6-alkylG-DNA adducts
Our genome-wide analysis of hypermethylated genes at the liver metastatic tumor revealed that 7.4% (range, 4.7% to 10.9%) of the genes showed hypermethylation in the metastatic tumors and 1.3% was commonly hypermethylated among three patients. These numbers are quite large at face value, but when we validated the data by bisulfite-pyrosequencing, a change in methylation density was the explanation in most cases. One additional clue to explain this finding came from an analysis of resection time differences between the primary and metastatic lesions. Thus, the percentage of hypermethylated genes at liver metastasis was significantly higher in metachronous metastasis than in synchronous metastasis. In one patient, the time between surgery for the primary tumor and the liver metastasis was 46 months and 10.9% of genes analyzed using MCAM showed differential hypermethylation at the liver metastatic tumor. MCAM data in a patient with synchronous metastasis revealed 4.7% differential hypermethylated genes. Given that population doubling (reflected by patient age) is a prime determinant of methylation in normal and neoplastic colon,
In summary, our results indicate that methylation frequency between primary tumors and matched liver metastasis is similar, suggesting that tumor cells acquire methylation changes before progression to liver metastasis. While we cannot rule out rare consistent changes, it appears that DNA methylation frequency is very stable over time in CRC.
(TIF)
(RTF)
We thank Ms. Nicole Sieffert and Ms. Hisako Nozawa for their assistance in identifying samples for this study.