A total of 205 patients with CRC (121 males and 84 females; mean age, 62.6 years; range, 25.4–90.1 years) from the Gastrointestinal Department of Cathay General Hospital (Taipei, Taiwan) were enrolled from January 2006 to December 2008 in the current study. Survival data were acquired from 176 patients and others were lost to follow-up due to referral. The mean follow-up time was 17.0±15.6 months (median, 10.5 months). Suspicious growths in patient colonic tissues were sampled with small biopsy forceps inserted through a colonoscope. The tissues were formalin-fixed, paraffin-embedded and cut into slices of 4–5-µm thickness for immunohistochemical staining, or immersed in RNAlater^® solution (Thermo Fisher Scientific, Inc., Waltham, MA, USA) for genomic DNA preparation, according to the manufacturer's protocol. Presence of distant metastasis was routinely confirmed by abdominal computed tomography. In addition, blood samples were collected from each patient to serve as controls when determining the microsatellite status. The study protocol was approved by the Institutional Review Board of Cathay General Hospital, and informed consent was obtained from all patients prior to obtaining tissue samples.

Cells from the human colorectal carcinoma HCT116 [American Type Culture Collection (ATCC) no. CCL-247; MSI) and SW480 (ATCC no. CCL-228; MSS) cell lines were purchased from the ATCC (Manassas, VA, USA) and maintained as recommended by their guidelines (www.atcc.org) (22,23). All cultured cells used in the current study were washed in ice-cold PBS (pH 7.4), scraped from culture dishes on ice using a plastic cell scraper and collected in 1.5-ml microcentrifuge tubes in 1 ml ice-cold PBS. Following a short centrifugation step (1,000 × g for 3 min at 4°C) to pellet the cells, the supernatants were removed from each sample, and the cell pellets were resuspended in 900 µl ice-cold 0.1% NP-40 (Merck Millipore, Darmstadt, Germany) in PBS. Subsequently, a reagent-based protocol that enabled the stepwise lysis of cells, separation of cytoplasm from intact nuclei and extraction of nuclear proteins was performed, according to the protocol of the NE-PER™ Nuclear and Cytoplasmic Extraction kit (Thermo Fisher Scientific, Inc.). The cytoplasmic (10 µg) and nuclear (10 µg) fractions underwent 10% SDS-PAGE and were subsequently transferred to polyvinylidene difluoride membranes. Membranes were then blocked with 10% skimmed milk and 3% bovine serum albumin in TBS buffer containing Tween-20 [TBST; 20 mM Tris (pH 7.4), 150 mM NaCl and 0.05% Tween 20] for 1 h to prevent non-specific binding of antibodies. Mutated KRAS and MSH2 were immunoblotted with anti-KRAS antibody (1:2,000 in TBST; 05–516; Merck Millipore) and anti-MSH2 antibody (1:500 in TBST; ab52266; Abcam, Cambridge, MA, USA) for 1 h, respectively. In addition, membranes were incubated with an anti-α-tubulin antibody (1:500 in TBST; sc-5286; Santa Cruz Biotechnology, Inc., Dallas, TX, USA) as a cytoplasmic marker, or with anti-lamin A/C antibody (1:500 in TBST; sc-7292; Santa Cruz Biotechnology, Inc.) as nuclear marker, for 1 h each. Protein signals were detected with Western Lightning Chemiluminescence Reagent Plus (PerkinElmer, Inc., Waltham, MA, USA) upon incubation with an appropriate secondary antibody conjugated with peroxidase (1:5,000 in TBST; Dako, Glostrup, Denmark) for 1 h. All incubations were performed at room temperature on a rocking platform and were followed by several washes with TBST buffer to remove the residual solution. Protein bands were quantified by densitometry using image processing AlphaView software version 3.2.2.0 for the FluorChem FC2 system (Cell Biosciences, Inc., Santa Clara, CA, USA). Relative protein levels were calculated and determined by normalizing their expression to that of a-tubulin or lamin A/C.

Sections were dewaxed in xylene and rehydrated through a graded series of ethanol concentrations (100, 95 and 70%) to water. Antigen retrieval was performed by immersing the slides in BD Retrievagen A (pH 6.5 for MLH1 and pH 6.0 for MSH2; BD Biosciences, San Jose, CA, USA) and heating the slides in a microwave oven for 30 min at 95°C. Sections were treated with 3% hydrogen peroxide for 5 min to block endogenous peroxidase activity. Sections were washed in PBS and subsequently placed in 20% normal goat serum (G9023; Sigma-Aldrich; Merck Millipore) in PBS for 20 min to reduce non-specific staining. Sections were then incubated with monoclonal antibodies raised against human (h) MLH1 (1:10; clone G168-15; BD Biosciences) for 1 h 32 min at 42°C or against MSH2 (1:25; clone 25D12; Thermo Fisher Scientific, Inc.) for 1 h at room temperature. Subsequently, visualization was performed using the EnVision+Dual Link system-HRP (Dako) according to the manufacturer's protocol, using 3,3′-diaminobenzidine as a chromogen. Finally, slides were counterstained with hematoxylin solution. Both MLH1 and MSH2 were scored as either negative or positive staining. Tissue specimens were analyzed by two independent pathologists blinded to the conditions.

Colonic tissues pathologically diagnosed as CRC were subjected to the assessment of microsatellite status and identification of hotspot mutations of KRAS and BRAF. Peripheral blood samples from patients diagnosed with CRC were collected in anticoagulant tubes containing EDTA. Genomic DNA was extracted from blood and tissue samples according to a standard protocol (26). To determine the microsatellite status of the enrolled CRC patients, a reference panel of five fluorescent dye-labeled microsatellite primers (BAT25, BAT26, D2S123, D5S346 and D17S250) purchased from Thermo Fisher Scientific, Inc., was used (27). Primer sets for the BAT25, BAT26 and D5S346 loci, and for the D2S123 and D17S250 loci, were respectively combined for the multiplex polymerase chain reaction (PCR) amplifications in a volume of 5.5 µl containing 20.0 ng genomic DNA. Denatured PCR products were analyzed by capillary electrophoresis in an ABI Prism^® 3100 Genetic Analyzer (Applied Biosciences; Thermo Fisher Scientific, Inc.). Raw data were collected using Data Collection Software version 1.1 (Applied Biosciences), and marker performance was evaluated using GeneScan^® Analysis Software version 3.7 (Applied Biosciences) electropherograms in full-view display for each dye color. For the purpose of prognostic evaluation, microsatellite status was categorized as MSS or MSI according to whether instability was evident for no markers or for ≥1 marker, respectively (28). To detect KRAS mutations at codons 12 and 13, a restriction enzyme-based analysis (BstNI for codon 12 and XcmI for codon 13) following appropriate PCR amplifications was employed, as previously described (29,30). Briefly, the cycling conditions for codon 12 were 35 cycles of amplification (93°C for 35 sec, 56°C for 50 sec and 72°C for 50 sec), and for codon 13 were 40 cycles of amplification (93°C for 60 sec, 52°C for 60 sec and 72°C for 60 sec). The sequences of the PCR primers used were: BstNI forward, 5′-ACTGAATATAAACTTGTGGTAGTTGGACCT-3′ and reverse, 5′-TCATGAAAATGGTCAGAGAAACC-3′; and XcmI, forward, 5′-ACTGAATATAAACTTGTGGTCCATGGA-3′ and reverse, 5′-TATCTGTATCAAAGAATGGTCCTGCACCAG-3′ (30,31). The presence of the BRAF V600E mutation was determined using an allele-specific PCR assay (32). Depending on the genotypes, either allele-specific reaction could amplify the target sequence. The cycling conditions used for PCR were 10 min at 95°C for initial activation and 40 cycles of amplification (92°C for 15 sec and 60°C for 60 sec).

The components of the reaction for the different genotypes were identical except for the respective allele-specific probes, as described below. The primer sequences were: Forward, 5′-CATGAAGACCTCACAGTAAAAATAGGTGAT-3′ and reverse, 5′-TGGGACCCACTCCATCGA-3′. Allele-specific TaqMan^® probes (VIC^®-labeled reporter T allele, 5′-CTAGCTACAG[T]GAAATC-3′ and 6-carboxyfluorescein-labeled reporter A allele, 5′-TAGCTACAG[A]GAAATC-3′) (Applied Biosciences) and TaqMan Genotyping Master Mix (Applied Biosciences) were used in a 7300 Real-Time PCR system (Applied Biosciences; Thermo Fisher Scientific, Inc.) (32). All identified KRAS/BRAF mutations were then validated using a BigDye^® Terminator v3.1 Cycle Sequencing kit (Applied Biosciences; Thermo Fisher Scientific, Inc.).

The association between tumor microsatellite status and patient clinicopathological features was analyzed using a χ² test. Patient survival time was calculated from the date of complete resection of CRC tumors to the date of last follow-up, with the only patient who succumbed to CRC being excluded from counting towards disease-specific survival (DSS). DSS distributions were estimated using a Kaplan-Meier method and compared using a log-rank test. P<0.05 was considered to indicate a statistically significant difference. All calculations were performed using IBM SPSS Statistics for Windows version 22.0 (IBM SPSS, Armonk, NY, USA).

One patient (age, 58.4 years; female) out of the 205 enrolled patients met the clinical diagnostic criteria for HNPCC (33,34). Immunohistochemical results indicated that staining for hMLH1 was negative for this patient (Fig. 1). As presented in Table I, 12.7% of the tumors (26 of 205) exhibited MSI, including the tumor in the patient with HNPCC. The mean age of the 26 patients with MSI (58.2 years; range, 29.0–81.6 years) was lower than that of the other 179 patients with MSS (63.3 years; range: 25.4–90.1 years). Furthermore, a KRAS or BRAF mutation (5 for BRAF V600E, 24 for KRAS codon 12 and 4 for KRAS codon 13) was detected in 31.2% of patients (64 of 205). A total of 79.7% (51 of 64) of patients with KRAS/BRAF mutations had MSS (P=0.027), demonstrating that a correlation exists between MSS and KRAS/BRAF mutations (Table II). When comprehensively considering the microsatellite status of patients and the site of the tumors, tumors at distal/rectal areas were significantly diagnosed in patients with MSS (90.4%, 142 of 157; P=0.015).

Significantly different mutation rates of KRAS or BRAF were observed in patients of different ages (P=0.013; Table III). A higher proportion of elderly patients (≥62.6 years) had mutated KRAS or BRAF (39.0%, 41 of 105), whereas only 23.0% (23 of 100) of younger patients (<62.6 years) harbored these mutations. The odds ratio for KRAS or BRAF mutations in the elderly patients was 2.15 (95% confidence interval, 1.17–3.94). In addition, of the 172 patients with available data for tumor differentiation, 15 were diagnosed with poorly differentiated tumors, and 53.3% (n=8) of these patients harbored a KRAS or BRAF mutation (P=0.036).

In the present study, survival data were available for 176 patients with CRC (154 patients with MSS and 22 patients with MSI). Patients with MSS tended to have poorer DSS (57.4±7.0%) compared with patients with MSI, although this difference was not significant (P=0.065; Fig. 2). In addition, it has been demonstrated that BRAF and KRAS mutations are frequently associated with CRC and influence prognosis (35). Therefore, the current study assessed the survival rates of patients with MSS according to their different KRAS/RAS mutations. In the subgroup of 154 patients with MSS, patients without KRAS or BRAF mutations (n=110) had better DSS (58.8±9.4%) than those with KRAS or BRAF mutations (n=44; 50.6±11.0%; P=0.043; Fig. 3).

Due to the clinical significance of microsatellite status and KRAS/BRAF mutations in CRC, MSS (SW480 cells; Fig. 4) and MSI (HCT116 cells; Fig. 5) cell lines were employed in the present study (22,23). Both SW480 and HCT116 cells had mutated KRAS but wild-type BRAF (36). In SW480 cells treated with 17.5 µM 5-FU for 2 days or 150 mM oxaliplatin for 3 days, the levels of KRAS detected did not significantly differ between the nucleus and the cytoplasm (Fig. 4A and B). However, under the same treatment conditions, the expression of the DNA MMR protein, MSH2, varied in SW480 cells (Fig. 4C). MSH2 expression increased in the cytoplasm and decreased in the nucleus of SW480 cells following treatment with 5-FU (both P<0.001). By contrast, MSH2 expression decreased in the cytoplasm and increased in the nucleus of SW480 cells treated with oxaliplatin (both P<0.001; Fig. 4B and D).

In addition, distinct expression patterns for KRAS and MSH2 in HCT116 cells treated with 76.9 µM 5-FU for 2 days or 1.4 µM oxaliplatin for 3 days were observed (Fig. 5). The levels of cytoplasmic and nuclear KRAS significantly decreased in HCT116 cells following 5-FU treatment (P=0.050 for cytoplasm and P=0.030 for nucleus; Fig. 5A). However, in cells with treated with oxaliplatin, the levels of nuclear KRAS decreased significantly (P<0.001), while the levels of cytoplasmic KRAS increased but not significantly (P=0.216; Fig. 5B). MSH2 expression decreased in the cytoplasm (P<0.001) and increased in the nucleus (P=0.027) of 5-FU-treated HCT116 cells compared with control cells (Fig. 5C). However, oxaliplatin treatment did not increase the expression of MSH2 in the nucleus of HCT116 cells (Fig. 5D).