SN-38, an active metabolite of CPT-11, was provided by Yakult (Tokyo, Japan). SN-38 was dissolved in dimethyl sulfoxide at a concentration of 1 μM and stored at −20°C, before further dilution in phosphate-buffered saline (PBS) and filter sterilization immediately prior to use. DAC was obtained from Sigma (St. Louis, MO, USA), dissolved in PBS and filter-sterilized.

The human colon adenocarcinoma cell lines HCT-15 and HT29 were provided by the Cell Resource Center for Biomedical Research, Tohoku University (Miyagi, Japan) and the American Type Culture Collection (Manassas, VA, USA). HCT-15 cells were maintained in RPMI-1640 (Sigma) containing 10% heat-inactivated fetal bovine serum (FBS), 100 U/ml penicillin, 100 μg/ml streptomycin, 10 mM HEPES (Gibco-BRL, Gaithersburg, MD, USA) and 1.0 mM sodium pyruvate (Gibco). HT29 cells were maintained in Mac Coy’s medium containing 10% heat-inactivated FBS, 100 U/ml penicillin and 100 μg/ml streptomycin. The cell lines were cultured at 37°C in 5% CO₂. Cells (5×10³ per well) were plated in 24-well plates on day 1 of the culture. On day 3, the medium was removed and new medium containing DAC was added. The cells were treated with 0.5 μM DAC (or PBS as a control) for 72 h. On day 6, the cells were rinsed twice with FBS-free medium, and fresh medium containing 0.0015 μM SN-38 (or PBS) was added. The dose of each drug was administered on the basis of its pharmacological dose as previously reported (14), and our preliminary experiments (data not shown). The cells were harvested following incubation with trypsin-EDTA on day 9, stained with trypan blue and counted. Total mRNA or genomic DNA was extracted on day 6 and utilized in the analysis of mRNA expression or methylation status following DAC treatment.

Total RNA was extracted from cultured cells using the RNeasy kit (Qiagen, Hilden, Germany). The integrity of the RNA was assessed using the Agilent 2100 BioAnalyzer (Agilent Technologies, Palo Alto, CA, USA), and the samples were confirmed to have an RNA Integrity Number >5.0 prior to gene expression analysis. Contaminant DNA was removed via digestion with RNase-free DNase (Qiagen). Using 2 μg total RNA, cRNA was prepared using the one-cycle target labeling and control reagents kit (Affymetrix, Santa Clara, CA, USA). Hybridization and signal detection of HG-U133 plus 2.0 arrays (Affymetrix) were performed following the manufacturer’s instructions. The 4 microarray datasets obtained were normalized using the robust multi-array average (RMA) method (16) and R 2.4.1 statistical software (17) together with a BioConductor package. Normalized gene expression levels were log2-transformed, and 62 control probe sets were removed for further analysis. For each of the 54,613 probe sets, a fold change analysis was performed using each pair of treated and untreated HCT15 and HT29 cells. Probe sets that showed >1.5-fold increase or decrease were identified as differentially expressed genes.

Total cellular RNA was extracted from the treated cells using the RNeasy Mini Kit (Qiagen, Valencia, CA, USA) and incubated with DNase as previously reported (14). cDNA was then synthesized from 10 ng extracted RNA using the High-Capacity cDNA reverse transcription kit (Applied Biosystems). PCR was performed as previously described (14), using the conditions: initial denaturation for 96°C for 2 min, followed by 26 cycles of denaturation for 1 min at 96°C, annealing for 1 min at 57–60°C and extension for 1 min at 72°C. As an internal control for RT-PCR analysis, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) transcripts were amplified from the same cDNA samples. Primer sequences were: GAPDH, 5′-CAACAGCCTCAAGATCATCAGC-3′ (forward) and 5′-TCCTAGACGGCAGGTCAGGTC-3′ (reverse); BNIP3, 5′-GGATGCAGGAGGAGAGCCT-3′ (forward) and 5′-CGAG GTGGGCTGTCACAGT-3′ (reverse); and SOCS3, 5′-GACCA GCGCCACTTCTT-CAC-3′ (forward) and 5′-ACTGGATGCG CAGGTTCTTG-3′ (reverse). Following amplification, PCR products were separated on 2% agarose gels, stained with ethidium bromide and visualized under ultraviolet illumination.

Tumor or normal mucosal tissue was cut into 10-μm sections from tissue blocks, with a change of blade and cleaning of the microtome between specimens. The specimens were deparaffinized and washed, and tumor tissue was manually dissected using a razor blade, using a hematoxylin and eosin (H&E)-stained slide as a guide. Dissected tissues or cultured cell lines were then incubated overnight with proteinase-K in digestion buffer and the genomic DNA was extracted using a standard phenol chloroform method.

Sodium bisulfite conversion and DNA recovery were performed using EpiTect Bisulfite (Qiagen). Following sodium bisulfite conversion, the genomic DNA was analyzed using the Methylight technique, a fluorescence-based RT-PCR (Q-PCR) assay (18) and the ABI Prism 7300 R-T PCR System (Taq Man; Applied Biosystems, Foster City, CA, USA). Two sets of primers and probes, designed specifically for bisulfite-converted DNA were used. One set detected methylation in the gene of interest and the second set served as a reference set for β-actin (ACTB) to normalize for input DNA. The reference primers and probes were designed in a region of the ACTB gene that lacks CpG dinucleotides, allowing for equal amplification regardless of the methylation levels. Primer and probe sequences were: ACTB, 5′-AGGTTGGGGAAGTTT GTTTTTG-3′ (forward), 5′-CCACCACCCAACACACAATA-3′ (reverse) and 5′-TGGGGTGGTGATGGAGGAGGTT-3′ (probe); BNIP3, 5′-TTCGGTCGGAGGAATTTATAGG-3′ (forward), 5′-CCCCAATCGCGACCAA-3′ (reverse) and 5′-ACGACGCGACCGCAAAT-3′ (probe); as well as SOCS3 5′-TCGCGTTTTTTT-TTTCGTAGTTT-3′ (forward), 5′-CGC GACCTCCGCACA-3′ (reverse) and 5′-CGACCGCTACCG CATCCCGA-3′ (probe). SssI-treated HCT-15 DNA was used as a fully methylated positive control (100% methylation ratio). Parallel Taq Man PCR reactions were performed with specific primers for the bisulfite-converted methylated sequence for a particular locus and with the ACTB reference primers. In each case, triplicate threshold cycle (Ct) values were obtained and averaged, and expression levels were then evaluated by the 2-ΔΔCt method (19). As an internal standard, each individual sample was normalized to its ACTB content and compared to the gene expression level of SssI-treated HCT-15 DNA as positive controls (calibration sample) as follows: 2^ΔΔCt, where ΔΔCt = (Ct-target-Ct-reference)-treated sample – (Ct-target-Ct-reference) calibrator sample. The percentage of fully-methylated reference (PMR) was defined as 2^ΔΔCt × 100%.

Patients entered into this study had undergone surgical resection for primary sporadic colorectal cancer at the Department of Surgical Oncology, Tokyo Medical and Dental University (Tokyo, Japan). This study was pre-approved by the institutional review board of Tokyo Medical and Dental University, and written informed consent was obtained from all 112 participants prior to the commencement of the study. Following surgery for both colon and rectal cancer, patients exhibiting stage III cancer received oral or intravenous 5-FU-based adjuvant chemotherapy, and those with stage IV tumor or following recurrence received 5-FU-based systemic chemotherapy. The resected specimens were fixed in 10% neutral buffered formalin and embedded in paraffin. For all cases, archival H&E slides of the primary tumors were retrieved and reviewed to confirm pathological features. In addition to the International Union Against Cancer Tumor-Node-Metastasis (TNM) classification (20), clinicopathological factors such as age, degree of histological differentiation, pathological tumor class and lymph node metastasis at the time of diagnosis in stage I–III patients during the follow-up period were studied. The response rate to chemotherapy including CPT-11 treatment was assessed by Response Evaluation Criteria In Solid Tumors (21).

Statistical analysis was carried out using StatView Software (version 5.0). To estimate the differences between groups, the Chi-square test, Welch’s t-test and log-rank test were used where appropriate. The Kaplan-Meier method was used to estimate survival. Survival was calculated from the date of surgery. Prognostic factors were examined by univariate and multivariate analysis using the Cox proportional hazards model. P<0.05 was considered to be statistically significant.

Pre-incubation with the demethylating agent DAC was examined to investigate whether it influenced the sensitivity of the colon cancer cell lines HCT-15 and HT29 to SN-38. On day 9, the total number of HCT-15 cells that had been pre-incubated with DAC prior to exposure to SN-38 was found to be significantly decreased in comparison to cells treated with SN-38 alone (p=0.0049, Fig. 1A). A similar reduction was observed in the HT29 cells (p=0.0038, Fig. 1A). Treatment with DAC or SN-38 alone resulted in no significant effect on the rate of cell proliferation in the two cell lines.

Since inactivation of the apoptotic pathway is associated with chemoresistance (8), identification of the pro-apoptotic genes that may be reactivated by DAC was attempted. Using oligonucleotide microarrays, the global changes in gene expression following the treatment of HCT-15 and HT29 with DAC were analyzed, and the expression changes in treated and untreated cells were compared. Expression levels of 943 genes were found to be increased by >1.5-fold following DAC treatment in HCT-15 cells, and 2,720 genes were increased in the HT29 cells. A total of 234 of these genes were found to be similarly altered in the two cell lines. Among the 234 genes, 10 were identified as having known pro-apoptotic activities (Table I). As the expression levels of BNIP3 and SOCS3 are frequently reduced due to promoter methylation in gastrointestinal cancers (22,23), up-regulation of these genes following DAC treatment was analyzed to determine whether it was caused by DNA demethylation. The BNIP3 mRNA expression was up-regulated and its DNA promoter was demethylated in both the HCT-15 and HT29 cell lines following treatment with DAC (Fig. 1B and C). In contrast, although SOCS3 mRNA expression was enhanced after DAC treatment, primers and probes specific for methylated SOCS3 failed to amplify any product in the two cell lines by Methylight analysis (data not shown). The results indicated that SOCS3 expression may be induced by DAC treatment independently of the methylation status of the DNA promoter.

The median PMR value for BNIP3 methylation in 16 randomly selected patients with normal colon epithelial tissue was 0. The positive methylation was defined as PMR>0. The BNIP3 methylation status was then investigated in 112 patients who had undergone surgical resection for primary CRC between March 2000 and April 2003. Patients were prospectively followed over a median post-operative duration of 42 months and methylation of BNIP3 was detected in 58% of the 112 patients. No significant correlations were found between BNIP3 methylation and clinicopathological factors such as age, gender, tumor depth, vessel invasion, lymphatic invasion, lymph node metastasis and stage (Table II). Patients with BNIP3 methylation also exhibited a significantly shorter overall survival time (OS) compared to those without methylation (p=0.012, Fig. 2A). In stage II/III tumors, patients with BNIP3 methylation also exhibited a significantly shorter OS than those without methylation (p=0.039, Fig 2B). No difference in relapse-free survival (RFS) was observed between patients with or without methylation (p=0.16, Fig. 2C). Regarding stage II/III patients, investigation of the univariate analysis of survival using the Cox proportional hazards model showed that BNIP3 methylation (p=0.043) and lymph node metastasis (p=0.0086) were significantly associated with OS. All of the variables with a significance level at p<0.35 in the univariate analyses were subsequently used for multivariate analyses. Multivariate analysis of the variables failed to show BNIP3 methylation to be an independent prognostic factor (Table III).

BNIP3 methylation was examined to determine whether it was a prognostic marker in primary CRC patients treated with CPT-11. Of the patients who had undergone surgical resection for primary CRC between May 1998 and September 2007, 30 patients (including 5 from 112 patients between 2000 and 2003) who demonstrated recurrence after radical resection or incomplete resection due to distant metastasis, received first-line chemotherapy with CPT-11. Patients were subsequently observed over a median duration of 15 months following the initial administration of CPT-11. One patient had complete response (CR), 10 partial response (PR), 7 stable disease (SD) for >6 months, and 12 progression of disease (PD). When CR and PR patients were defined as responders, the observed response rate was lower in patients with BNIP3 methylation than those without (33.3 vs. 40%), but this difference was not statistically significant (p=0.70, Fig. 3A). Time to progression (TTP) in patients with BNIP3 methylation was slightly shorter than that in patients without methylation, but this difference was not statistically significant (median 196.5 vs. 237 days) (p=0.10, Fig. 3B).