Factors affecting the sensitivity of human-derived esophageal carcinoma cell lines to 5-fluorouracil and cisplatin

Effective chemotherapy against esophageal carcinoma is considered achievable with a combination of 5-fluorouracil (5-FU) and cisplatin (CDDP). However, chemo-therapy remains ineffective in certain patients. The aim of this study was to clarify the factors which affect sensitivity to 5-FU and CDDP. The effects of factors known to influence sensitivity to 5-FU and CDDP, namely transporters, DNA repair enzymes and metabolic enzymes, were examined. mRNA levels of four transporters, SLC22A2, SLC23A2, ABCB1 and ABCC2, two DNA repair-related enzymes, Rad51 and MSH2, and one metabolic enzyme, dihydropyrimidine dehydrogenase (DPYD), showed a strong correlation (|r|>0.7) with IC50 values for 5-FU. In addition, the mRNA levels of ABCC2, MSH2 and DPYD showed a strong correlation (|r|>0.7) with the IC50 values for CDDP. Gimeracil, a DPYD inhibitor, enhanced the sensitivity of some cells to 5-FU but decreased the sensitivity of all the cells to CDDP. The inhibitory effects of ABCC2 with MK571 did not correspond to those observed in the correlation analysis. In conclusion, mRNA levels of SLC22A2, SLC23A2, ABCB1, ABCC2, Rad51, MSH2 and DPYD were confirmed to be strongly correlated with IC50 values for 5-FU, and mRNA levels of ABCC2, MSH2 and DPYD were confirmed to be strongly correlated with IC50 values for CDDP. In addition, the inhibition of DPYD appeared to affect the cytotoxicity of CDDP.


Introduction
In Japan, one-third of all mortalities are cancer-related (1). The incidence of lung, colorectal and breast cancer is increasing in Japan as well as worldwide (1). Esophageal carcinoma has a lower incidence than other types of cancer, but 5-fluorouracil (5-FU) and cisplatin (CDDP)-based chemoradiotherapy results in moderately high response and survival rates relative to other types of cancer. In fact, the complete response and 5-year survival rates following 5-FU and CDDPbased chemoradiotherapy have been reported to be 58 and 29%, respectively, among Japanese esophageal carcinoma patients (2). However, chemotherapy remains ineffective in certain patients. Therefore, identifying the factors that affect sensitivity to 5-FU and CDDP is necessary for enhancing the clinical outcome of chemotherapy for esophageal carcinoma.
Certain factors affecting sensitivity to 5-FU or CDDP have previously been revealed, including the molecular mechanisms involved in the cellular kinetics and dynamics of 5-FU and CDDP. For example, overexpression of the ABC transporter superfamily C5 (ABCC5/MRP5) decreases cellular accumulation of 5-FU, resulting in resistance to . In addition, dihydropyrimidine dehydrogenase (DPYD), a 5-FU metabolizing enzyme, has been correlated with clinical response to 5-FU-based chemotherapy among colon cancer patients (4,5). The cytotoxic effects of CDDP are also attenuated by ERCC1, a DNA repair-related enzyme associated with restoration of DNA damage induced by chemotherapeutic agents or UV rays (6)(7)(8). However, there is little information concerning whether the levels of these molecules are predictive of sensitivity to 5-FU or CDDP in esophageal carcinoma.
In the present study, sensitivity to 5-FU and CDDP and mRNA levels of 35 genes, including drug transporters, DNA repair enzymes and metabolic enzymes, were evaluated in 5 human esophageal carcinoma cell lines. Based on these findings, factors affecting the sensitivity of esophageal carcinoma cells to 5-FU and CDDP were examined.

Materials and methods
Chemicals. 5 Growth rate of esophageal carcinoma cell lines. The growth rate of esophageal carcinoma cells was evaluated with a WST-1 assay utilizing succinate dehydrogenase activity. Cells were seeded onto a 96-well plate (Corning Inc., Corning, NY, USA) at a density of 5x10 3 cells/well/100 µl and cultured in an atmosphere of 95% air and 5% CO 2 at 37˚C. After 0, 6,12,18,24,36,48,72 and 96 h, the culture medium was exchanged for 110 µl of medium containing WST-1 reagent solution (10 µl WST-1 solution and 100 µl culture medium), and 3 h later the absorbance was determined using a microplate reader at 450 nm with a reference wavelength of 620 nm (SpectraFluor™, Tecan Group Ltd., Männedorf, Switzerland). The doubling time for cell growth was calculated from the logarithmic phase of a growth curve (10) as follows: Doubling time = (t 1 -t 0 ) x log 10 2/(log 10 N 1 -log 10 N 0 ). N 0 and N 1 are the number of cells (% of day 0) at t 1 and t 0 , respectively.
Growth inhibitory activity assay. Cells were seeded onto 96-well plates (Corning Inc.) at a density of 5x10 3 cells/well/100 µl on day 0. After incubation for 24 h, the culture medium was exchanged for one containing 5-FU or CDDP at various concentrations (day 1). On day 4, a WST-1 assay was performed as described above.
The effects of gimeracil and MK571 on the growth inhibitory effects of 5-FU and CDDP were also evaluated by WST-1 assay. Cells were incubated for 24 h as described above and the culture medium was exchanged for one containing 5-FU or CDDP at various concentrations with or without gimeracil (100 µM) or MK571 (50 µM). Following incubation for 72 h at 37˚C, the culture medium was replaced with a medium containing WST-1 and the absorbance was measured.
The 50% growth inhibitory concentrations (IC 50 ) were calculated according to the sigmoid inhibitory effect model: E = E max x [1 -C γ /(C γ + IC 50 γ )], using the nonlinear least-squares fitting method (Solver, Microsoft ® Excel). E and E max represent the surviving fraction (% of control) and its maximum, respectively. C and γ are the drug concentration in the medium and the sigmoidicity factor, respectively. Relative sensitivity was calculated as follows: Relative sensitivity = IC 50 (without gimeracil or MK571)/IC 50 (with gimeracil or MK571).
Real-time reverse transcription (RT)-PCR. The mRNA expression levels were measured by real-time RT-PCR. Cells were seeded at a density of 2x10 6 cells/60 mm culture dish and 48 h later, total RNA was extracted from the cells with a GenEluteTM Mammalian Total RNA Miniprep kit (Sigma-Aldrich). Total RNA (1 µg) was used for RT with a PrimeScriptTM RT reagent kit (Takara Bio, Inc., Shiga, Japan) and a thermal cycler (i-Cycler, Bio-Rad Laboratories, Inc., Hercules, CA, USA). The RT reaction was conducted in 40 µl reaction buffer at 37˚C for 15 min and terminated by heating at 85˚C for 5 sec followed by cooling at 4˚C.
Real-time PCR was performed with a 7500 Real-time PCR system (Applied Biosystems, Carlsbad, CA, USA) and SYBR Premix Ex Taq™ (Takara Bio, Inc.). The primer sequences are shown in Table I. PCR was performed at 95˚C for 10 sec, followed by 40 cycles of 95˚C for 5 sec and 60˚C for 34 sec. Dissociation was initiated at 95˚C for 15 sec followed by 60˚C for 1 min and 95˚C for 15 sec. To compare the relative expression of target mRNA levels between the cell lines, the comparative Ct method was used, as previously described (10); β-actin (ACTB) was used as an internal standard. Samples were prepared in duplicate and three independent sample sets were analyzed.

Statistical analyses.
Data are shown as the mean ± standard deviation (SD). Comparisons between 2 and among 3 or more groups were performed with Student's unpaired t-test and repeated one-way analysis of variance (ANOVA) followed by Scheffe's F test, respectively. P<0.05 (two-tailed) was considered to indicate a statistically significant result. The correlation analysis was performed using Pearson's correlation coefficient (r).

Results
Growth rates of esophageal carcinoma cell lines. Table II shows the cell growth doubling times for the 5 esophageal carcinoma cell lines. Doubling times for the cells varied from 20 to 25 h, revealing a significant difference between lines.  (Table III). In the case of CDDP, the IC 50 values were also substantially different among the cell lines (2.17-19.5 µM). The rank order of sensitivity to CDDP was comparable to that for 5-FU.
Correlation analysis of factors affecting drug sensitivity. The level of mRNA expression differed among the esophageal carcinoma cell lines (Table IV). The correlations between the IC 50 values and the mRNA levels of the 35 different genes were analyzed (Table V). SLC22A3 mRNA was not detected in any cells, with the exception of the OE33 cell line. ABCC6 mRNA expression was not observed in KYSE30 and KYSE70 cells.
The mRNA levels of SLC22A2, SLC23A2, ABCB1 and Rad51 showed a strong negative correlation (r<-0.7) with the IC 50 values for 5-FU. ABCC2, MSH2 and DPYD were positively correlated with the IC 50 values for 5-FU (r>0.7; Table V and Fig. 1). In the case of CDDP, a high positive correlation coefficient (r>0.7) was found between the IC 50 values and ABCC2, MSH2 and DPYD mRNA expression (Table V and Fig. 2).
Effects of gimeracil and MK571 on sensitivity of esophageal carcinoma cell lines to 5-FU and CDDP. The sensitivity of KYSE30 cells to 5-FU was enhanced by gimeracil, but in the other cell lines gimeracil had no observable effect (Table VI).
In addition, gimeracil showed a tendency to decrease the sensitivity of all the cell lines to CDDP.

Discussion
Combination chemotherapy with 5-FU and CDDP is known to be effective against esophageal carcinoma. However, it remains ineffective in certain patients, and the causes for this have not been clarified. The aim of the present study was to examine the factors affecting the sensitivity of esophageal carcinoma cells to 5-FU and CDDP. The sensitivity of the 5 different esophageal carcinoma cell lines to 5-FU and CDDP differed (Table III). OE33, an adenocarcinoma cell line, showed a high sensitivity to 5-FU and CDDP, whereas the squamous cell carcinoma KYSE30 cells showed low sensitivity to 5-FU and CDDP. In addition, OE33 cells had the longest doubling time (an index of cell growth) of all the cell lines and KYSE30 cells the shortest (Table II), resulting in a trend for lower sensitivity to chemotherapeutic agents among cells with higher growth activity. These findings suggest that sensitivity to 5-FU and CDDP was influenced by the growth activity of cells, although cytotoxic agents such as 5-FU and CDDP are known to be more toxic in cells with higher growth activity. In order to resolve this discrepancy, further studies concerning the correlation between cell growth and sensitivity to 5-FU or CDDP should be performed.
The correlations between sensitivity to 5-FU and CDDP and the mRNA levels of the 35 genes were then examined. The levels of target mRNA expression differed among the cell lines (Table IV). The mRNA levels of ABCC2, MSH2 and DPYD were positively correlated with the IC 50 values of 5-FU (r>0.7; Fig. 1 and Table V). By contrast, a negative correlation between the IC 50 values of 5-FU and the mRNA levels of SLC22A2, SLC23A2, ABCB1 and Rad51 was observed. In the light of the biological roles of these genes, the negative correlation between SLC22A2 and SLC23A2 mRNA expression and sensitivity was considered to be noteworthy. SLC22A2 encodes an organic cation transporter which is responsible for cell uptake of various drugs, including CDDP (11,12). A colon carcinoma cell line exhibiting resistance to 5-FU has been reported to show lower expression of SLC23A2 mRNA than its parent cells (13). ABCC2, MSH2 and DPYD are known to act in detoxifying mechanisms; they are an efflux transporter, DNA repairrelated protein and metabolic enzyme, respectively. Although ABCB1 is a known efflux transporter that contributes to drug resistance, the cytotoxicity of 5-FU was not influenced by the expression of ABCB1 (14). In addition, the overexpres-      (Table III). The mRNA expression levels (2 -∆Ct ) in the cells were evaluated by real-time RT-PCR assay using SYBR ® -Green. The threshold cycle (Ct) values were used to quantify the PCR product, and the relative expression level of the target gene was expressed as 2 -∆Ct . The ∆Ct was calculated by subtracting Ct (β-actin; as an internal standard) from Ct (target gene). 5-FU, 5-fluorouracil; RT, reverse transcription.  (Table III). The mRNA expression levels (2 -∆Ct ) in the cells were evaluated by real-time RT-PCR assay using SYBR ® -Green. The threshold cycle (Ct) values were used to quantify the PCR product, and the relative expression level of the target gene was expressed as 2 -∆Ct . ∆Ct was calculated by subtracting Ct (β-actin; as an internal standard) from Ct (target gene). CDDP, cisplatin; RT, reverse transcription.
sion of DNA-repair related proteins, including Rad51, has been reported to contribute to resistance to DNA damaging agents (15). Although the present findings showing a negative correlation between IC 50 values and ABCB1 and Rad51 mRNA expression levels conflict with previous findings, they may indicate that ABCB1 and Rad51 have no significant impact on sensitivity.
In the case of CDDP, a positive correlation (r>0.7) between the IC 50 values and the mRNA levels of ABCC2, MSH2 and DPYD was identified. The findings for ABCC2 and MSH2 are supported by their functions; the export of CDDP from cells (16) and repair of DNA damaged by CDDP (17), respectively (Table V and Fig. 2). In addition, proliferating cell nuclear antigen-normalized mRNA expression of DPYD Table IV. Expression levels of mRNA in esophageal carcinoma cell lines. has previously been reported to be associated with sensitivity to CDDP in lung cancer tissues (18). Although the correlation between CDDP and DPYD has not been investigated in detail, these previous results may support the present findings.
In conclusion, the mRNA levels of SLC22A2, SLC23A2, ABCB1, ABCC2, Rad51, MSH2 and DPYD were confirmed to be strongly correlated with the IC 50 values for 5-FU, and those of ABCC2, MSH2 and DPYD were also confirmed to be strongly correlated with the IC 50 values for CDDP. These genes have the potential to affect the sensitivity to 5-FU and CDDP. In addition, the inhibition of DPYD was suggested to affect the cytotoxicity of CDDP. These findings provide useful information for improving the clinical outcome of chemotherapy against esophageal carcinoma.