Functional Similarity of Anticancer Drugs by MTT Bioassay

We prepared normal or Ha-ras-transformed NIH3T3 cells transfected stably or transiently with various tumorrelated genes. The chemosensitivity of the transfected clones to 16 anticancer drugs was compared to the parental control cells using the MTT assay. The chemosensitivity changes induced by transfected genes were calculated and expressed numerically as the Drug Chemosensitivity Index (DCI). High DCI values (indicating resistance) were frequently observed in cells expressing C/EBP?, C/EBP?, p53, p21, PTEN, dominant-negative MDM2, caspases, HSP90, COUP-TF1 and decorin. In contrast, transfectants expressing ras, src, erbB2 and calpastatin had low DCI values, indicating increased sensitivity. Thus, it may be possible to predict the sensitivity of cancer cells toward anticancer drugs based on the expression levels of these genes. We then performed a regression analysis of DCI values between anticancer drugs. The correlation coefficients (r) were relatively high between cisplatin, camptothecin, mitomycin C and etoposide, suggesting that the mechanisms of action of these drugs are similar. The r values of aclarubicin, vincristine, taxol and cytarabine were low, suggesting that each of these drugs has a different and unique effect. This analysis may provide a rationale for design of combination chemotherapy regimens.


Introduction
Chemotherapy using anticancer drugs is a useful therapeutic method for cancer. Thus far, many effective anticancer drugs have been developed. However, due to the adverse effects of each drug, their application is limited. Thus, combination therapy using smaller amounts of multiple drugs has become more and more common.
The efficacy of anticancer drugs varies among patients. This may be explained by differences in gene expression. For example, overexpression of P-glycoprotein results in prominent resistance to many drugs such as vincristine, etoposide, and paclitaxel [1]. Comprehensive cDNA microarray analysis has been carried out for various cancers to examine alterations of gene expression [2][3][4][5][6][7][8].
In addition, a large diverse panel of cultured human tumor cell lines was tested for sensitivity to anticancer drugs [9]. Although this analysis was efficient to discriminate anticancer drug-responsive tumor cells, the genetic backgrounds of the cell lines were so variable that the precise action mechanism remained unclear. In the present study, we prepared NIH3T3 mouse fibroblasts transfected stably or transiently with varying tumor-related genes, and examined their chemosensitivity to anticancer drugs. Variations in action mechanisms may provide a rationale for combination chemotherapy.

Cell lines and culture
NIH3T3 mouse fibroblasts and Ha-ras-transformed NIH3T3 cells (ras-NIH3T3) [10] were cultured in DMEM supplemented with 5% bovine serum and 100 µg/ml of kanamycin. Cells were transiently or stable transfected with cDNAs in eukaryotic expression vectors such as pcDNA3 and pME18S-FL3. To isolate stable transfectants, cells were transfected with the expression plasmids together with the neo gene using LipofectAMINE reagent (Life Technologies, Carlsbad, CA), and then selected by culture in the presence of G418 (400 µg/ml) for two weeks as described [11][12][13]. A total of 135 stable transfectants and 41 transiently transfected cells were examined for chemosensitivity toward 16 anticancer drugs.

Western-blotting and RT-PCR analyses
The protein expression levels of transfected genes were examined by Western blotting [12,13]. The cells were washed with phosphatebuffered saline (PBS) and lysed in 0.5% Nonidet P-40, 20 mM Tris-HCl (pH 7.5), 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 50 µM leupeptin, 50 µM antipain, 50 µM pepstatin A and 50 µM acetyl-Leucyl-Leucyl-norleucinal for 10 min at 4˚C. The cell lysate was centrifuged at 13,000×g for 10 min and the supernatant was used as cytoplasmic cell extract. The pellet was used as the nuclear fraction. The samples were analyzed by Western blotting using antibodies purchased from Santa Cruz Biotechnology (Santa Cruz, CA), followed by detection using ImmunoStar Reagents (Wako Pure Chemicals).
When the commercial antibodies are not available, the mRNA expression levels were examined by RT-PCT (reverse transcriptionpolymerase chain reaction) as described previously [13,14]. Briefly, total cellular RNA was isolated from the tumor tissues using FastPure RNA Kit (Takara Biochemicals, Kyoto, Japan). Reverse transcription was carried out with oligo(dT) 20 primer using the ThemoScript RT-PCR System (Invitrogen, Carlsbad, CA), according to the manufacturer's instructions. The mRNA expression levels were examined via quantitative realtime RT-PCR using the Universal Probe Library system (Roche, Basel, Switzerland).

Evaluation of cytotoxicity of anticancer drugs
Chemosensitivity of transfected cells was examined using the MTT assay according to the method of Mosmann [15] as previously described [16,17]. Cells at exponentially growing phase were used. Five thousand cells per well (100 µl) were plated in 96-well plates in the presence of various concentrations of anticancer drugs, and cultured for three days. The activity of mitochondrial succinic dehydrogenase was measured by incubation for 4 h in the presence of 0.5 mg/ml of MTT (3-(4,5-dimethylthiozol-2-yl)-2,5-diphenyltetrazolium bromide). Absorbance reflects the viable cell number and was measured at 570 nm with a reference wavelength of 650 nm using a microplate reader.
Absorbance reflects the viable cell number and was expressed as a percentage of that of cells cultured in the absence of anticancer drugs.
For the combination treatment, ras-NIH3T3 were added with 5-FU (50 ng/ml), PEP (500 ng/ml) or MTX (50 ng/ml), and immediately after the addition, plated in the presence of varying concentrations of MIT, PEP or VP-16.

Chemosensitivity to anticancer drugs of transfected cells
Mouse fibroblasts, NIH3T3 or ras-NIH3T3, were transfected with various cancer-related genes, and the chemosensitivity to 16 different anticancer drugs was examined by MTT assay. Table 1 summarizes the list of genes of which the transfection induced resistance or sensitization against each anticancer drugs. The DCI values reflect the extent how resistant of the transfected cells were converted by the gene transfection [17][18][19]. The full version of list of DCI values are shown in Supplementary Table S1. High DCI values (indicating drug resistance) were frequently observed in cells expressing C/EBPα, C/EBPβ, wildtype p53, p21, PTEN, mutated MDM2, caspases, HSP90, COUP-TFI and decorin. In contrast, transfectants expressing ras, src, erbB2, calpastatin, mutated p53 and wild-type MDM2 had low DCI values, indicating increased sensitivity. Thus, it may be possible to predict the sensitivity of cancer cells toward anticancer drugs based on the expression levels of these genes. It should be noted that oncogenes such as ras, src, erbB2 and MDM2 increased the chemosensitivity against some, if not all, anticancer drugs whereas tumor suppressor genes such

Regression analysis of DCI values
Because chemosensitivity was similar but not identical among anticancer drugs, we then performed a regression analysis of the DCI values shown in Table 1. The correlation coefficients (r) are summarized in Table 2. The r values were relatively high (significantly correlated) among CDDP, CPT, MMC and VP-16, suggesting that the mechanisms of action of these drugs are similar (Table 2, Figure 2). The r value between MMC and CPT was the highest. The r values of ACR, VCR, taxol and AraC were relatively low (no correlation), suggesting that each of these drugs has a different and unique effect.
Combination chemotherapy is growing more common in cancer chemotherapy. For example, the combination of cyclophosphamide (IFM analogue), VCR and doxorubicin (ACR analogue) is effective for small cell lung cancer [20] and non-Hodgkin lymphoma [21]. The low r value among these drugs means that they works independently. Thus, additive effects can be expected. This was exemplified by the result that the high-r-value combination of MIT and PEP caused less cytotoxic effects than the low r value combination of MIT and 5-FU or MTX ( Figure 3A).
On the other hand, colon cancer and ovarian clear cell adenocarcinoma are treated with a combination of CPT and MMC [22,23], which showed a high r value (Table 2). Both CPT and MMC are effective toward ras-and erbB2-transfected cells (Table 1). Colon cancers are frequently accompanied by mutations in the Ki-ras gene [24,25]. Overexpression of erbB2 in ovarian carcinoma [26,27] may account for the sensitivity against CPT and MMC. Thus, if the target molecules are restricted, synergistic effects focused on the target can be expected. Furthermore, such high-r-value combinations may reduce the side effects caused by a high-dose application of a single drug. Likewise, the high r value combinations of PEP and MTX or VP-16 and MTX showed more suppressive effects than the low r value combinations of PEP and 5-FU or VP-16 and PEP ( Figure 3B and C). Shown are correlation coefficients between two each of anticancer drugs. Coefficients higher than 0.5 or lower than 0.2 are marked.  Figure 3: Results of combination treatment. ras-NIH3T3 cells were added with 5-FU (50 ng/ml), PEP (500 ng/ml) or MTX (50 ng/ml). Immediately after the addition, cells were plated in the presence of varying concentrations of MIT (A), PEP (B) and VP-16 (C). After culture for three days, the viabilities were measured by the MTT assay. The abscissa and ordinate represent the concentration of the latter drugs and the relative viability versus that in the absence of the latter drugs, respectively.
Consequently, this approach using regression analysis of DCI values of anticancer drugs may provide a theoretical basis for design of combination chemotherapy regimens.