Purpose: ID4 gene is a member of the inhibitor of DNA binding (ID) family proteins that inhibit DNA binding of basic helix-loop-helix transcription factors. The epigenetic inactivation of ID4 gene on colorectal cancer (CRC) development and its clinical significance was assessed.
Experimental Design: In CRC cell lines, ID4 methylation status of the promoter region was assessed by methylation-specific PCR and bisulfite sequencing. The mRNA expression level was assessed by quantitative real-time reverse transcription-PCR. The methylation status of 9 normal epithelia, 13 adenomas, 92 primary CRCs, and 26 liver metastases was assessed by methylation-specific PCR. ID4 protein expression was assessed by immunohistochemistry analysis of tissue specimen.
Results: CRC cell lines were shown to be hypermethylated, and mRNA expression was suppressed and could be restored by 5-aza–cytidine treatment. In clinical specimens from normal epithelia, adenomas, primary CRCs, and liver metastases, the frequency of ID4 hypermethylation was 0 of 9 (0%), 0 of 13 (0%), 49 of 92 (53%), and 19 of 26 (73%), respectively, with a significant elevation according to CRC pathological progression. Methylation status of primary CRCs significantly correlated with histopathological tumor grade (P = 0.028). Immunohistochemistry analysis showed ID4 expression of normal colon epithelia, adenomas, and unmethylated primary CRCs but not hypermethylated CRC specimens. Among 76 American Joint Committee on Cancer stage I to IV patients who had undergone curative surgical resection, overall survival was significantly poorer in patients with hypermethylated ID4 bearing tumors (P = 0.0066).
Conclusions: ID4 gene is a potential tumor suppressor gene for which methylation status may play an important role in the CRC progression.
In the development of colorectal cancer (CRC), tumor-suppressor genes such as APC, p53, and genes on chromosome 18q (DCC, SMAD2, and DPC4/SMAD4) are inactivated by mutations or by chromosomal deletions (1 , 2) . Some CRCs have microsatellite instability caused by inactivation of mismatch repair genes such as hMSH2 or hMLH1 (3) . In addition, epigenetic inactivation by hypermethylation of promoter regions of various tumor suppressor genes such as p16, APC, VHL, and hMLH1 have been found in CRCs (4, 5, 6, 7, 8, 9, 10) and used as molecular markers of CRC (11 , 12) . Methylation of cytosines in CpG islands in the promoter region affects promoter activity and can down-regulate gene transcription (5) . Because the promoter hypermethylation of genes in cancer cells is as significant as deletions or mutations (13, 14, 15) , hypermethylation of key regulatory genes can play a significant role in transformation and colon tumor progression of colon epithelium. Progression of transformed cells requires regulatory gene inactivation that promotes growth, dedifferentiation, invasion, and/or metastasis.
Transcription factors containing a basic helix-loop-helix (bHLH) motif regulate the expression of certain tissue-specific genes (16) and have important roles in cell differentiation and embryonic developmental processes. DNA-binding activity of the bHLH proteins is dependent on formation of homo- and/or heterodimers. ID family proteins, which are distinct members of the helix-loop-helix (HLH) protein family, contain the HLH-dimerization domain but lack the DNA-binding basic domain. Consequently, ID proteins dominantly inhibit binding to DNA and transcriptional transactivation by forming heterodimers with bHLH proteins and modulate various key developmental processes (17) . Currently, four known human ID proteins have been identified. Expression studies have shown that ID proteins play critical roles in early embryonic development (18, 19, 20) . They are also involved in angiogenesis, lymphocyte development, cell cycle control, and cellular senescence (21, 22, 23) . The involvement of ID proteins in neoplastic processes has been suggested. Increased ID1 and ID2 expression has been reported in various tumor types, including adenocarcinomas arising from the colon and pancreas (24 , 25) . Transgene expression of ID1 and ID2 in mice has resulted in tumor formation in the intestinal epithelium and lymphoid organs, respectively (26 , 27) . Expression of ID3 has been more variable; studies report both up-regulation (24 , 28) and down-regulation (29 , 30) in different tumor types.
ID4 gene has a relation with growth and differentiation of cells, as reported with oligodendrocytes (31) . Recently, it was reported that ID4 promoter is hypermethylated in 30% of primary gastric cancers, and expression is down-regulated in most gastric cancer cell lines by hypermethylation of the promoter region (32) . Despite the structural similarity, ID4 is known to have some differences from the other three known ID gene members. Unlike ID1, ID2, and ID3, the immunoreactivity of which is significantly elevated in CRCs compared with normal epithelium (24) , ID4 has a more restricted expression pattern during murine and avian embryogenesis and is expressed at more advanced stages of differentiation in tissues (18, 19, 20) . In the development of murine stomach, ID4 expression is restricted to the ventral part where cells grow slower, whereas other ID members are expressed in the dorsal part of the stomach where cells proliferate faster (18) . Information about ID4 function, expression, and regulation of tumor progression is very limited, and there are no published major studies of ID4 in CRC to date.
On the basis of these findings, we hypothesized that ID4 gene may be expressed in normal colon epithelium and have a putative tumor suppressive role in CRC, contrary to other ID members. To examine this hypothesis, we assessed the methylation status of CRCs. We found that ID4 gene transcription is silenced during CRC development and that hypermethylation of the ID4 promoter region is one of the main mechanisms of inactivation.
MATERIALS AND METHODS
Three CRC cell lines SW480, DLD1, and LOVO (American Type Culture Collection, Manassas, VA) were analyzed in this study. Genomic DNA was extracted from cells as described previously (33) . Total RNA was extracted with TRI Reagent (Molecular Research Center, Inc., Cincinnati, OH) according to the manufacturer’s protocol. For ID4 expression restoration study, SW480 and DLD1 were treated with DNA demethylation agent 5-aza-cytidine (5Aza), a known inhibitor of methylation, as described previously (33, 34, 35) . Cells were seeded at 7 × 105/T-25 flask on day zero; the culture medium was changed on day two, and cells were then treated with 5Aza at final concentrations of 5 μg/mL for 48 hours (SW480 and DLD1) and 10 μg/mL for 72 hours (DLD1). After treatment, cells were harvested for DNA and RNA as described above.
Tissue Specimens and Clinicopathological Information.
For the analysis of methylation status at ID4 promoter region, we studied 131 colorectal tumors (13 adenomas, 92 primary CRCs, and 26 liver metastases) from 122 patients randomly selected by the database coordinator from those who underwent colectomy or proctectomy between 1996 and 2001 at Saint John’s Health Center (Santa Monica, CA). Nine normal colorectal epithelial tissues were obtained simultaneously from patients with primary CRC. All patients in this study were consented according to the guidelines set forth by JWCI Institutional Review Board committee. Tumors were classified and staged according to the revised guidelines set by the American Joint Committee on Cancer (36) . Clinicopathological data from the tumor registry were obtained after Institutional Review Board approval for all of the patients.
DNA Extraction from Tissue Specimens and Bisulfite Modification.
Several 5-μm sections were cut with a microtome from formalin-fixed, paraffin-embedded blocks under sterile conditions as described previously (37) . One section for each tumor was stained with hematoxylin after deparaffinization, and the tumor tissues were precisely microdissected under a microscope. Dissected tissues were digested with 50 μL of proteinase K containing lysis buffer at 50°C for 5 hours, followed by heat deactivation of proteinase K at 95°C for 10 minutes.
Sodium bisulfite modification was applied on extracted genomic DNA of tissue specimens or cell lines for methylation-specific PCR or bisulfite sequencing as described previously (33) .
Detection of Hypermethylation.
Methylation status of ID4 promoter region was analyzed by methylation-specific PCR and bisulfite sequencing. Because the “minimal promoter region” (−48 to +32) of ID4 gene had been shown previously by deletion analysis for promoter determination (38) , we were able to design highly specific methylation-specific PCR primer sets (Fig. 1)⇓ . Forward primers for methylation-specific PCR covered the TATA box, E-box, and three CpG sites in the ID4 minimal promoter region; reverse primers covered three CpG sites. The methylation-specific primer set was as follows: forward, 5′-D4-TTTTATAAATATAGTTGCGCGGC-3′; and reverse, 5′-GAAACTCCGACTAAACCCGAT-3′. The unmethylation-specific primer set was as follows: forward, 5′-D3-TTTTATAAATATAGTTGTGTGGTGG-3′; and reverse, 5′-TCAAAACTCCAACTAAACCCAAT-3′. PCR amplification was done in a 10-μL reaction volume with 1 μL template for 40 cycles of 30 seconds at 94°C, 30 seconds at 58°C, and 30 seconds at 72°C, followed by a 7-minute final extension at 72°C. Mg2+ concentration was 1.5 mmol/L for methylation-specific primer set and 2.5 mmol/L for unmethylation-specific primer set. Primer concentration was 0.1 μmol/L for methylation-specific primer set and 0.4 μmol/L for unmethylation-specific primer set. PCR products were detected and analyzed by CEQ 8000XL capillary array electrophoresis system (Beckman Coulter, Inc., Fullerton, CA) with CEQ 8000 software version 6.0 (Beckman Coulter) as described previously (35) . Methylation status was determined by the ratio of the signal intensities of methylated and unmethylated PCR products; samples with methylated to unmethylated ratio larger than 0.2 were determined as methylated. Each primer set was confirmed not to yield amplification on DNA without bisulfite treatment.
For bisulfite sequencing, the primer set was as follows: forward, 5′-TTTTATTYGGGTAGTYGGATTTTTYGTTTTTTAGTAT-3′; and reverse, 5′-CCCACCCRAATATCCTAATCACTCCCTTC-3′, with Y = C or T, R = G or A, as described previously (32) . PCR amplification was done in a 50-μL reaction volume with 2 μL template for 40 cycles of 30 seconds at 94°C, 30 seconds at 60°C, and 30 seconds at 72°C, followed by a 7-minute final extension at 72°C with the use of 2.5 mmol/L of Mg2+. Purified PCR products were directly sequenced with CEQ DYE Terminator Cycle Sequencing kit (Beckman Coulter, Inc.). Cycling program includes 30 cycles of 20 seconds at 95°C, 40 seconds at 55°C, and 4 minutes at 60°C. Sequences were read by CEQ 8000XL capillary array electrophoresis system (Beckman Coulter) with CEQ 8000 software version 6.0 (Beckman Coulter) as described previously (37) .
Analysis of mRNA Expression Level.
Reverse-transcriptase reactions were done on 1.0 μg of extracted total RNA with Moloney murine leukemia virus reverse-transcriptase (Promega, Madison, WI) with oligodeoxythymidylic acid primers, as described previously (39) . Quantitative real-time reverse transcription-PCR assay was done on the iCycler iQ Real-Time thermocycler detection system (Bio-Rad Laboratories, Hercules, CA; ref. 39 ). For each PCR, the reaction mixture consisted of cDNA template synthesized by reverse-transcription from 250 ng of total RNA, 0.2 μmol/L of forward primer (5′-CGCTCACTGCGCTCAACAC-3′), 0.2 μmol/L of reverse primer (5′-TCAGGCGGCCGCACACCT-3′), and 0.6 μmol/L of fluorescence resonance energy transfer probe (5′-FAM-CATTCTGTGCCGCTGAGCCG-BHQ-3′). PCR amplification was done in a 20-μl reaction volume for 45 cycles of 30 seconds at 94°C, 30 seconds at 58°C, and 30 seconds at 72°C with 3 mmol/L of Mg2+. Absolute copy numbers were determined by a standard curve with serial dilutions (108 to 101 copies) of DNA containing ID4 or GAPDH cDNA sequence. Analysis without templates was done as a negative control in each study. PCR products were electrophoresed on 2% agarose gels to confirm correct product size and absence of nonspecific bands. The expression level of the housekeeping gene GAPDH was measured as an internal reference with a standard curve to determine the integrity of template RNA for all of the specimens. The ratio of ID4 and GAPDH mRNA level was calculated as follows: (absolute copy number of ID4)/(absolute copy number of GAPDH) as described previously (39) .
Immunohistochemistry analysis of ID4 protein expression in primary CRCs, adenomas, and normal colon tissue sections was done to determine concordance with methylation-specific PCR results. Immunohistochemistry was done on 3-μm sections of formalin-fixed, paraffin-embedded tissues, which were placed on silane-coated slides and baked at 60°C for 1 hour. Afterward, the slides were deparaffinized, hydrated, and placed in antigen retrieval buffer (DAKO Corporation, Carpinteria, CA) at 95°C for 10 minutes. Endogenous peroxidase activity was quenched by 1% hydrogen peroxide for 10 minutes. After blocking with 1% BSA for 60 minutes, 1:100 dilution of an anti-ID4 polyclonal rabbit IgG antibody, sc-491 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), was applied and incubated for 3 hours at room temperature. After washing in PBS, antibody binding was visualized with DAKO LSAB+ kit (DAKO Corporation) followed by diaminobenzidine staining with DAB substrate kit for peroxidase (Vector Laboratories, Inc., Burlingame, CA) for 2 minutes at room temperature. The sections were lightly counterstained with hematoxylin and then mounted. As negative controls, adjacent sections of each ID4 immunostained section were stained simultaneously without primary antibody.
The relation between methylation status of ID4 gene promoter region and tumor classification was assessed with Fisher’s exact test, χ2 test, and Cochran-Armitage trend test. The relation between ID4 methylation and clinicopathological characteristics was assessed with Fisher’s exact test, χ2 test, and Wilcoxon’s rank-sum test for univariate analysis and logistic regression model for multivariate analysis. For survival analysis grouping with ID4 methylation, Kaplan-Meier analysis was used, and differences between the survival curves were analyzed with the log-rank test. Cox’s proportional hazard regression models were used for univariate and multivariate analyses of clinicopathological characteristics and prognosis. Variables suggested by the univariate analyses (P < 0.10), except for the highly dependent variable of ID4 methylation, were entered into the multivariate analyses. The statistical package SAS JMP version 5.0.1 (SAS Institute Inc., Cary, NC) was used to conduct statistical analyses. A P < 0.05 (two-tailed) was considered as significant.
Cell Line Analysis.
Among the three CRC cell lines studied, SW480 and DLD1 showed only methylation-specific peaks by methylation-specific PCR and were determined as hypermethylated, and LOVO showed both methylation-specific and unmethylation-specific peaks by methylation-specific PCR and was determined as hypermethylated (Fig. 2B)⇓ . Because methylation-specific PCR results depend on the methylation status of only six CpG sites at the primer annealing sites, we did bisulfite sequencing of the promoter region. All of the sequenced CpG sites were methylated in SW480 and most of them were methylated in DLD1 (Fig. 2)⇓ . In LOVO, sequence results at about one-third of CpG sites showed mixture of methylation and unmethylation. For a negative control, DNA from peripheral blood lymphocytes obtained from a healthy donor, which was determined to be unmethylated by methylation-specific PCR, was assessed by bisulfite sequencing. All sequenced CpG sites were unmethylated in peripheral blood lymphocytes (Fig. 2A)⇓ . It was verified that methylation-specific PCR results represent the methylation status of promoter region accurately in our assays.
To assess the mRNA expression level of ID4, we did quantitative real-time reverse transcription-PCR. In the SW480 and DLD1 cell lines, ID4 mRNA expression was not detected. In contrast, the partially hypermethylated LOVO cell line had 1.4 × 105 copies/250 ng of total RNA, and the ID4 to GAPDH expression ratio was 2.5 × 10-2 (Table 1)⇓ .
To ascertain that ID4 transcription down-regulation was caused by promoter hypermethylation, hypermethylated cell lines were treated with the DNA demethylating agent 5Aza. After treatment with 5 μg/mL of 5Aza for 48 hours, SW480 showed an unmethylation-specific peak by methylation-specific PCR analysis (Fig. 2B)⇓ . Bisulfite sequencing revealed that most CpG sites were changed to unmethylated form (Fig. 2C)⇓ , whereas all of the CpG sites of nontreated cells remained methylated. ID4 expression was restored from an undetectable level to 4.7 × 102 copies/250 ng of total RNA (Table 1)⇓ . Because the treatment with 5 μg/mL of 5Aza for 48 hours did not overtly restore ID4 expression in DLD1, the dose and duration of exposure were increased to 10 μg/mL of 5Aza for 72 hours. This treatment produced an unmethylation-specific peak on methylation-specific PCR analysis (Fig. 2B)⇓ , and bisulfite sequencing revealed that CpG sites partially lost their methylation (Fig. 2C)⇓ . ID4 expression was restored from an undetectable level to 4.8 × 102 copies/250 ng of total RNA (Table 1)⇓ .
Tumor Progression of Clinical Specimens and ID4 Methylation Status.
To show the changes in ID4 methylation status during CRC development, we assessed each stage of CRC progression. ID4 hypermethylation was not detected in nine normal colon epithelia or in 13 adenomas, but it was identified in 49 of 92 (53%) primary CRCs and in 19 of 26 (73%) liver metastases. The frequency of hypermethylation was significantly higher in primary carcinomas than in adenomas (P = 0.0002 by Fisher’s exact test). In addition, the frequency of hypermethylation was significantly increased as the tumor progressed from adenoma to primary carcinoma and then to metastatic CRC (P < 0.0001 by χ2 test; P < 0.0001 by Cochran-Armitage trend test; Table 2⇓ ).
Of the 10 primary CRCs analyzed with immunohistochemistry, six were hypermethylated and four were unmethylated. In Fig. 3⇓ , representative immunohistochemistry results are aligned with corresponding methylation-specific PCR results. Normal colonic epithelia and adenomas had diffusely stained cytoplasm, representing high concentration of ID4 protein expression (Fig. 3, A and B)⇓ . In primary CRCs determined as unmethylated by methylation-specific PCR, cell cytoplasm was lightly stained for ID4 protein. A representative microscopic picture of an unmethylated primary CRC, which was well to moderately differentiated carcinoma, is shown in Fig. 3C⇓ . In contrast, cytoplasm of primary CRCs determined as hypermethylated by methylation-specific PCR did not immunostain. Representative microscopic pictures of hypermethylated primary CRCs, which were poorly differentiated carcinoma and mucinous carcinoma, are shown in Fig. 3⇓ , D-1 and D-2, respectively. No nuclear staining was observed in any type of tissues.
Methylation status of primary CRCs was independent of sex, age, tumor location, tumor diameter, American Joint Committee on Cancer tumor-node-metastasis (TNM) scores, and stage (Table 3)⇓ . However, there was a significant correlation with histopathological tumor grade, which represents tumor cell differentiation (P = 0.028, Fisher’s exact test). In a multivariate analysis of hypermethylation status by logistic regression model, histopathological tumor grade and TNM T score were incorporated by stepwise variable selection, but only the histopathological tumor grade was significant for methylation status (P = 0.025).
Of the 92 patients from whom primary CRC tissue was obtained, 16 underwent noncurative resection; all of these patients had American Joint Committee on Cancer stage IV disease with irresectable remote metastases or local invasion. Four of the 16 patients expired from surgery-related causes within 30 postoperative days and were therefore excluded from survival analysis; the remaining 12 had primary CRCs that were hypermethylated (n = 7) or unmethylated (n = 5). Table 4⇓ shows the results of univariate survival analysis for the 88 evaluable patients and for the subgroup of 76 patients who underwent curative surgical resection of CRC. Methylation status, TNM stage, N score, methylated score, histopathological tumor grade, and tumor diameter were significantly correlated with survival in both groups by Cox’s regression analysis.
Among the 88 evaluable patients, those whose primary CRC was unmethylated had significantly better prognosis by Kaplan-Meier analysis (P = 0.019, log-rank test; Fig. 4A⇓ ). This difference was even more significant in the 76 patients who underwent curative surgical resection (P = 0.0066, log-rank test; Fig. 4B⇓ ): 5-year survival rates were 88% and 59% in unmethylated and methylated groups, respectively.
In a multivariate analysis of Cox’s proportional hazard model, methylation status, TNM N and methylated score, and tumor diameter were selected as covariates that had P values under 0.1 in univariate analysis. TNM stage was not selected because of the direct association with the TNM N and methylated score, and histopathological tumor grade was not selected because it was highly dependent on ID4 methylation status. The risk ratio of ID4 hypermethylation was 1.82 (95% confidence interval 1.09–3.43; P = 0.020; Table 5⇓ ). Hypermethylation of the ID4 promoter region was identified as an independent prognostic factor.
This is the first report of ID4 inactivation in CRC and its effect on overall survival. Because ID4 is suggested as one of the key controlling factors for cell differentiation, we hypothesized that epigenetic regulation of ID4 gene might affect tumor differentiation and progression of CRC. The relationship between ID4 promoter hypermethylation and mRNA transcription, protein expression levels, and clinicopathological characteristics was determined.
Two of the three CRC cell lines had fully hypermethylated ID4 genes and consequently ID4 mRNA expression was inactivated. Bisulfite sequencing showed a concordance of methylation-specific PCR results and promoter hypermethylation. Only LOVO was partially methylated; it has both a methylated allele and an unmethylated allele and had been shown to have ID4 mRNA expression. These findings support the hypothesis that ID4 gene can be inactivated by promoter hypermethylation and CRC can have biallelic methylation. ID4 re-expression studies after 5Aza treatment showed that hypermethylation of the promoter region silenced expression of ID4 gene in CRC cell lines. The GAPDH mRNA was decreased after 5Aza treatment. However, this influence was compensated in the ratios of mRNA expression levels of ID4 and GAPDH, and the analysis was not affected.
To show that aberrant hypermethylation down-regulates ID4 protein expression level in CRCs, we did immunohistochemistry study on a subset of specimens that were methylated or unmethylated by methylation-specific PCR. ID4 protein was diffusely expressed in cytoplasm of nonmalignant epithelium, adenomas, and unmethylated CRCs, but it tended to be unexpressed in hypermethylated CRCs. Concordance between immunohistochemistry and methylation-specific PCR results showed that aberrant hypermethylation of ID4 down-regulated protein expression.
Methylation analysis on tissue specimens was done in a blinded fashion without any clinical information. The methylation status was determined by methylation-specific PCR based on the intensity ratio of the methylation-specific and unmethylation-specific peaks by automated capillary array electrophoresis system with analysis software. Because methylation-specific PCR can detect a very small percentage of methylated DNA in abundant unmethylated DNA, the methylation-specific PCR results were methylation-positive, even if only a small part of the microdissected tumor cells was hypermethylated. Aberrant hypermethylation was not found in normal colonic epithelia and adenomas, whereas 53% of primary CRCs and 73% of liver metastases were hypermethylated. The frequency of hypermethylation increased with CRC progression. This trend supports the concept of multistep colorectal tumorigenesis (1 , 40) , in which genetic alterations accumulate during tumor progression. ID4 gene could be a putative tumor suppressor gene of CRC, which is epigenetically inactivated by promoter hypermethylation at a later stage of cancer development. The absence of hypermethylation in adenomas, which are considered to be precursors of CRC (1 , 40) , and the high frequency of hypermethylation in primary CRCs support the concept that ID4 down-regulation is related to the malignant transformation of neoplastic tumor cells.
Multivariate analysis by Cox’s proportional hazard model identified hypermethylation of ID4 as a significant independent risk factor of poor prognosis after curative surgical resection. Surprisingly, ID4 methylation status had higher impact on prognosis than lymph node metastasis, the most important prognostic risk factor for patients with curatively resected primary CRCs. Many tumor-related genes that are epigenetically inactivated in CRC tumor cells by promoter region hypermethylation, such as p16, APC, VHL, and hMLH1 (4, 5, 6, 7, 8, 9, 10) , have been identified, but the prognostic utility of these methylated genes in primary tumors have not been well described. Our results indicate that ID4 hypermethylation can be used as a prognostic marker for CRC patients.
Univariate analysis and multivariate logistic regression analysis showed that methylation status significantly correlated with histopathological tumor grade. Histopathological tumor grade of CRC is an independent prognostic factor. At present, no regulatory gene for CRC cell differentiation via inhibition of specific transcription factors has been identified, but ID4 protein may have a regulatory function. The mechanisms by which the down-regulation of ID4 protein results in unfavorable prognosis are not clear, but it is likely that ID4 down-regulation promotes dedifferentiation and proliferation of CRC cells. ID4 protein may inhibit DNA binding of bHLH transcription factors that are involved in tumor cell dedifferentiation. Interestingly, some methylated tumors that contained both moderately differentiated and poorly differentiated areas stained positive for ID4 protein in the moderately differentiated area but negative in poorly differentiated area by immunohistochemistry (data not shown). This heterogeneity in the ID4 protein level suggests that protein down-regulation may have induced tumor cell dedifferentiation. It is unknown whether ID4 silencing is directly linked to dedifferentiation or is just a confounding factor of the histologic grade at this time. Additional investigation is needed to reveal the mechanisms by which ID4 expression contributes to tumor cell differentiation in CRC progression.
In this study, it was shown that ID4 of CRC is epigenetically down-regulated. It has been also reported that ID4 is not expressed in certain human breast cancer tissues (41) . However, previous animal studies have reported contrary findings showing that ID4 regulates mammary epithelial cell growth and differentiation and is overexpressed in rat mammary gland carcinomas (42) . These observations suggest that the regulatory mechanisms of ID4 gene function may be differential in malignant tumors.
In conclusion, hypermethylation of the promoter region down-regulates ID4 at the mRNA level in CRC cell lines and at the protein level in clinical specimens. The frequency of hypermethylation is high in primary and metastatic CRCs compared with normal epithelium and adenoma. These results support ID4 as a potential tumor suppressor gene that may play an important role in CRC progression. ID4 transcription inactivation is associated with poorer differentiation of CRC and with unfavorable prognosis. ID4 hypermethylation can be used as a prognostic marker independent of TNM scores or stage of CRC.
Grant support: PO CA 29605 Project II from the National Cancer Institute, NIH, the Rod Fasone Memorial Cancer Fund, and the Roy E. Coates Foundation Laboratory.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Dave S. B. Hoon, Department of Molecular Oncology, John Wayne Cancer Institute, 2200 Santa Monica Boulevard, Santa Monica, CA 90404. Phone: (310) 449-5267; Fax: (310) 449-5282; E-mail:
- Received April 8, 2004.
- Revision received June 1, 2004.
- Accepted June 7, 2004.