E2F1 Overexpression Correlates with Thymidylate Synthase and Survivin Gene Expressions and Tumor Proliferation in Non–Small-Cell Lung Cancer

Lung cancer is a major cause of cancer-related deaths, and non–small-cell lung cancer (NSCLC) comprises ∼75% of all lung cancers (1). It is important to design an optimal therapeutic strategy according to tumor biology to improve the treatment of NSCLC (2). The selection of effective chemotherapies based on the evaluation of biomarkers (i.e., “tailor-made chemotherapy”) can improve the clinical outcome of NSCLC patients (3). For example, 5-fluorouracil (5-FU)–derived agents would be useful for tumors with a low expression of thymidylate synthase (TS; refs. 4, 5). Gefitinib and erlotinib would be effective for tumors with epidermal growth factor receptor (EGFR) mutations or increased EGFR gene copy numbers (6). Furthermore, the apoptotic index is also a useful indicator to predict the efficacy of chemotherapy for NSCLC patients (7).

The E2F1 transcription factor plays a key role in G₁-to-S phase transition by attracting numerous upstream signals (8). Experimental studies showed that E2F1 induces various genes encoding S phase–activating proteins, including TS (9, 10). Furthermore, a recent experimental study revealed that E2F1 induced the gene expression of Survivin (11), a member of the inhibitor of apoptosis protein family (12). It is important to clarify the molecular mechanisms of these biomarkers in NSCLCs to develop effective cancer treatments. Therefore, we conducted a clinical study in NSCLC patients to investigate E2F1 gene expression in relation to TS and Survivin gene expressions.

Furthermore, E2F1 modulates diverse cellular functions such as DNA synthesis, mitosis, and apoptosis (8). With regard to tumorigenesis, in vitro studies have reported that E2F1 can act either as an oncogene (13, 14) or as a tumor suppressor gene (15, 16). Previous clinical studies in human cancers also showed that E2F1 has a tumor-promoting effect in many cancers including lung cancer (17, 18), breast cancer (19), and thyroid cancer (20), whereas the opposite seems to be the case for other cancers such as colon cancer and bladder cancer (21, 22). Therefore, to investigate the clinical significance of E2F1 in NSCLCs, we also evaluated E2F1 gene expression in relation to tumor proliferation using the Ki-67 index (23) and tumor apoptosis using the apoptotic index (7).

Patient characteristics. We studied consecutive NSCLC patients who underwent surgery at the Second Department of Surgery of Kagawa University from January 2000 to June 2003. This study was approved by the institutional review board of Kagawa University (14-7, a clinical study of biological markers in NSCLCs) and signed informed consent was obtained from each patient. Tumor-node-metastasis staging designations were made according to the postsurgical pathologic international staging system (24). The lymph node status was pathologically evaluated using specimens resected by either thoracotomy or mediastinoscopy. In total, 127 patients with lung cancer up to stage IIIB, including 63 patients with adenocarcinoma, 58 patients with squamous cell carcinoma, and 6 patients with large cell carcinoma, were investigated (Table 1). The patients' clinical records and histopathologic diagnoses were fully documented. This report includes follow-up data as of October 31, 2006. The median follow-up period for all patients was 62.2 months.

With regard to the methods of surgical resection, a pneumonectomy was done in 15 patients with stage II to III NSCLC. A lobectomy was done in 103 patients: 48 patients with stage I NSCLC, 25 patients with stage II NSCLC, and 30 patients with stage III NSCLC. A segmentectomy was done in five patients with stage I NSCLC, and a wedge resection was done in four patients with stage I NSCLC. Systemic chemotherapy using mitomycin/vinblastin/cisplatin or carboplatin/paclitaxel was done in all patients with stage II to III NSCLC: neoadjuvant chemotherapy in 36 patients and postoperative adjuvant chemotherapy in 34 patients with nodal metastases. Radiation therapy was done in 20 patients: 9 patients with T₃ or T₄ status and 11 patients with mediastinal lymph node metastases.

Quantitative reverse transcription-PCR. Total RNA was extracted from frozen tissue specimens by the acid guanidinium thiocyanate procedure. First-strand cDNA synthesis was done with 5 μg of total RNA using a cDNA synthesis kit (Pharmacia). To quantify E2F1 and TS gene expression, real-time quantitative PCR was done with the ABI PRISM 7700 Sequence Detection System (Applied Biosystems). The primers and probes were from the Assays-on Demand Gene Expression Assay Mix (E2F1 assay ID Hs00153451_m1, PCR product size 84 bp; TS assay ID Hs00426591_m1, PCR product size 87 bp; Applied Biosystems). The PCR cycling conditions were as follows: 50°C, 2 min for AmpErase UNG activation; 95°C, 10 min for AmpliTaq Gold activation; and 40 cycles for the melting (95°C, 15 s) and annealing/extension (60°C, 1 min) steps. Real-time PCR assays were conducted in triplicate for each sample, and each PCR experiment included three nontemplate control wells. The standard curves for serial dilutions of cDNA of the human lung cancer cell line A549, which had positive expressions of E2F1 and TS, were similarly generated. The comparative threshold cycle method (Applied Biosystems) was used to calculate the gene expression ratio in each sample relative to the value observed in the control A549 cells, using GAPDH (Assays-on-Demand Gene Expression system, assay ID Hs99999905_m1, PCR product size 122 bp; Applied Biosystems) as a control for normalization among the samples.

Because Survivin has splice variants with different functions (25), real-time quantitative reverse transcription-PCR is not an appropriate method to discriminate wild-type Survivin from other splice variants. We carried out semiquantitative reverse transcription-PCR assays to evaluate the gene expression of wild-type Survivin, as described preiously (26). For wild-type Survivin gene expression, the primers used were 5′-CCACCGCATCTCTACATTCA-3′ (sense) and 5′-TATGTTCCTCTATGGGGTCG-3′ (antisense), and the PCR product size was 185 bp. The reaction mixture was subjected to 36 PCR amplification cycles of 60 s at 94°C, 60 s at 60°C, and 90 s at 72°C. β-actin DNA amplification was used as the internal PCR control. The primers used were 5′-GATATCGCCGCGCTCGTCGTCGAC-3′ (sense) and 5′-CAGGAAGGAAGGCTGGAAGAGTGC-3′ (antisense), and the PCR product size was 792 bp. Preparations of A549 cells were used as positive controls. The amplified DNA samples were run on a 1% agarose gel with ethidium bromide, and the densitometric value obtained for a Survivin band in a given tumor sample was divided by the value of the β-actin, the internal control. Then the expression ratio for a given tumor sample was divided by the expression ratio of A549 to obtain the normalized Survivin gene expression ratio.

Immunohistochemistry. The following antibodies were used, along with isotype antibodies as negative controls: a mouse monoclonal antibody for E2F1 (KH95, Santa Cruz Biotechnology) diluted at 1:100, a rabbit polyclonal antibody for TS diluted at 1:500 (5), a mouse monoclonal antibody for Survivin (D-8, Santa Cruz Biotechnology) diluted at 1:50, and a mouse monoclonal antibody for the Ki-67 antigen (MIB-1, DAKO) diluted at 1:40. Formalin-fixed paraffin-embedded tissue was cut into 4-μm sections and mounted on poly-l-lysine–coated slides. After deparaffinization and rehydration, the slides were heated in a microwave for 30 min in a 10 μmol/L citrate buffer solution at pH 6.0 and were cooled to room temperature for 20 min. After quenching the endogenous peroxidase activity with 0.3% H₂O₂ (in absolute methanol) for 30 min, the sections were treated with 5% bovine serum albumin to block nonspecific staining. Duplicate sections were incubated overnight with the primary antibodies. The slides were then incubated for 1 h with biotinylated secondary antibodies (Vector Laboratories). The sections were incubated with the avidin-biotin-peroxidase complex (Vector Laboratories) for 1 h, and the antibody binding was visualized with 3,3′-diaminobenzidine tetrahydrochloride. Finally, the sections were lightly counterstained with Mayer's hematoxylin. The sections of lung tumors known to express E2F1, TS, or Survivin were used as positive controls.

All of the immunostained sections were reviewed by two authors (C.H. and M.U.) without knowledge of the patients' characteristics. Cases with discrepancies were jointly reevaluated until a consensus was reached. With regard to E2F1 protein expression, only nuclear staining was considered positive, and the E2F1 status was assessed as the percentage of stained tumor nuclei (17). In the cytoplasmic staining for TS and the nuclear staining for Survivin, all sections were scored in a semiquantitative manner, which reflects both the intensity and the percentage of cells staining at each intensity. The intensity was classified as 0 (no staining), +1 (weak staining), +2 (distinct staining), or +3 (very strong staining). A value designated the “HSCORE” was obtained by using the following algorithm: HSCORE = Σ(I × PC), where I and PC represent staining intensity and the percentage of cells that stain at each intensity, respectively. The percentage of carcinoma cells with positive staining for Ki-67 in a given specimen was scored as the Ki-67 proliferation index (23).

Apoptosis. Apoptotic cells were detected with the terminal deoxyribonucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) method using the In Situ Apoptosis Detection Kit (Takara Biomedicals). After deparaffinization and rehydration, the slides were treated for 15 min with 20 μg/mL proteinase K. After quenching the endogenous peroxidase activity with 3% H₂O₂ for 5 min, the sections were incubated for 90 min at 37°C with the TUNEL reaction mixture, including terminal deoxynucleotidyl transferase. The sections were subsequently incubated for 30 min at 37°C with anti-FITC-horseradish peroxidase conjugate. The staining was developed with 3,3′-diaminobenzidine tetrahydrochloride. Lastly, the sections were lightly counterstained with Mayer's hematoxylin. Sections incubated with the TUNEL reaction mixture without terminal deoxynucleotidyl transferase were used as negative controls. Apoptotic cells were determined based on observations of TUNEL-staining sections and serial H&E-staining sections. TUNEL-staining cells, if they represented the histologic features of necrosis in H&E-staining sections, were not considered to be apoptotic cells. In each case, a total of 10,000 tumor cells were evaluated at high magnification by two authors (J.N. and M.U.) independently without knowledge of the patients' characteristics. Cases with discrepancies were jointly reevaluated until a consensus was reached. The apoptotic index was defined as the number of apoptotic cells per 1,000 tumor cells.

Statistical analysis. Because the distributions of the values, including normalized E2F1 gene expression ratio (P = 0.1631), normalized TS gene expression ratio (P = 0.1606), normalized Survivin gene expression ratio (P = 0.4736), Ki-67 proliferation index (P = 0.9367), and apoptotic index (P = 0.3758), all showed normal distributions (Kolmogorov-Smirnov analysis), the statistical significances of these values were assessed by the t test, ANOVA with Bonferroni/Dunn test, or Pearson's correlation coefficient. In addition, because each cutoff value showed the most significance in relation to the Ki-67 proliferation index, the sample was classified as a high-E2F1 tumor when the normalized E2F1 gene expression ratio was >0.25, a high-TS tumor when the normalized TS gene expression ratio was >0.25, and a high-Survivin tumor when the normalized Survivin gene expression ratio was >1.0. Overall survival was defined as the time from treatment initiation to the date of death from any cause. The Kaplan-Meier method was used to estimate the probability of overall survival as a function of time, and the differences in the survival of subgroups of patients were compared by using Mantel's log-rank test. A multivariate analysis was done using the Cox regression model to study the effects of different variables on survival. All P values were based on a two-tailed statistical analysis and P < 0.05 was considered statistically significant.

E2F1 gene expression in NSCLCs. We initially studied E2F1 gene expression in 10 noncancerous samples. The normalized E2F1 gene expression ratio was low in normal lung tissues (mean, 0.0179 ± 0.0115). In contrast, the normalized E2F1 gene expression ratio varied greatly among the 127 tumor tissues we studied (mean, 0.368 ± 0.547; Table 2). E2F1 protein expression exhibited a nuclear staining pattern with varied cytoplasmic staining (Fig. 1A). The normalized E2F1 gene expression ratio significantly correlated with the percentage of tumor cells with E2F1 protein expression (r = 0.708; P < 0.0001). Seventy-six (59.8%) carcinomas were low-E2F1 tumors and 51 (40.2%) carcinomas were high-E2F1 tumors. With regard to tumor histology, the normalized E2F1 gene expression ratio was 0.301 ± 0.508 in adenocarcinomas and 0.441 ± 0.605 in squamous cell carcinomas. There was no significant difference in E2F1 gene expression according to tumor histology. However, for tumor differentiation, the normalized E2F1 gene expression ratio was 0.119 ± 0.179 in well-differentiated tumors, 0.322 ± 0.430 in moderately differentiated tumors, and 0.559 ± 0.696 in poorly differentiated tumors. The normalized E2F1 gene expression ratio was significantly higher in poorly differentiated tumors in comparison with well-differentiated and moderately differentiated tumors (P = 0.0008). Furthermore, for pathologic stage, the normalized E2F1 gene expression ratio was 0.224 ± 0.363 in stage I tumors, 0.607 ± 0.808 in stage II tumors, and 0.400 ± 0.478 in stage III tumors. The normalized E2F1 gene expression ratio was significantly lower in stage I tumors than in stage II or III tumors (P = 0.0072).

Tumor proliferation and apoptotic index of NSCLCs in relation to E2F1 gene expression. To investigate the biological function of E2F1 expression in NSCLC, we evaluated the Ki-67 proliferation index and the apoptotic index in relation to the E2F1 status (Fig. 1B). With regard to tumor proliferation, the normalized E2F1 gene expression ratio showed a significant positive correlation with the Ki-67 proliferation index (r = 0.487; P < 0.0001; Fig. 2A). The Ki-67 proliferation index was significantly higher in high-E2F1 tumors than in low-E2F1 tumors (66.5 ± 22.2% versus 35.0 ± 26.0%; P < 0.0001). However, no correlation was observed between the normalized E2F1 gene expression ratio and the apoptotic index (r = 0.129; P = 0.1488; Fig. 2B). The apoptotic index was 2.04 ± 0.67 in low-E2F1 tumors and 2.02 ± 0.68 in high-E2F1 tumors.

TS and Survivin gene expressions in relation to E2F1 gene expression. We subsequently studied the gene expressions of TS and Survivin, the tumor-associated E2F1 target genes. With regard to TS, the normalized TS gene expression ratio varied greatly among the 127 tumor tissues we studied (mean, 0.668 ± 1.029). The TS protein expression exhibited a cytoplasmic staining pattern (Fig. 1C) and significantly correlated with the HSCORE of TS protein expression in tumor cells (r = 0.603; P < 0.0001). Furthermore, the normalized E2F1 gene expression ratio showed a significant positive correlation with the normalized TS gene expression ratio (r = 0.709; P < 0.0001; Fig. 2C). The normalized TS gene expression ratio was significantly higher in high-E2F1 tumors than in low-E2F1 tumors (1.296 ± 1.369 versus 0.246 ± 0.282; P < 0.0001).

With regard to Survivin, the normalized Survivin gene expression ratio also varied greatly among the 127 tumor tissues we studied (mean, 1.327 ± 0.686). The Survivin protein expression exhibited a pattern of nuclear staining with varied cytoplasmic staining (Fig. 1D). The normalized Survivin gene expression ratio significantly correlated with the HSCORE of the nuclear staining of Survivin in tumor cells (r = 0.448; P < 0.0001). Furthermore, the normalized E2F1 gene expression ratio showed a significant positive correlation with the normalized Survivin gene expression ratio (r = 0.403; P < 0.0001; Fig. 2D). The normalized Survivin gene expression ratio was significantly higher in high-E2F1 tumors than in low-E2F1 tumors (1.585 ± 0.694 versus 1.153 ± 0.626; P = 0.0004).

Clinical significance of TS expression in NSCLCs. With regard to tumor proliferation, the normalized TS gene expression ratio positively correlated with the Ki-67 proliferation index (r = 0.460; P < 0.0001; Fig. 2E). The Ki-67 proliferation index was significantly higher in high-TS tumors than in low-TS tumors (61.3 ± 23.4% versus 31.9 ± 26.8%; P < 0.0001). However, for tumor apoptosis, the apoptotic index was 2.14 ± 0.66 in low-TS tumors and 1.92 ± 0.67 in high-TS tumors. There was no difference in the apoptotic index in relation to TS status.

Clinical significance of Survivin expression in NSCLCs. The normalized Survivin gene expression ratio positively correlated with the Ki-67 proliferation index (r = 0.357; P < 0.0001; Fig. 2F). The Ki-67 proliferation index was significantly higher in high-Survivin tumors than in low-Survivin tumors (58.2 ± 24.8% versus 35.5 ± 28.8%; P < 0.0001).

Furthermore, there was a significant inverse relationship between the normalized Survivin gene expression ratio and the apoptotic index (r = −0.355; P < 0.0001; Fig. 2G). The apoptotic index was significantly lower in high-Survivin tumors than in low-Survivin tumors (1.80 ± 0.54 versus 2.29 ± 0.72; P < 0.0001).

Overall survival of NSCLC patients. With regard to E2F1 status, among 127 NSCLC patients, the 3-year survival rate was significantly lower in patients with high-E2F1 tumors than in those with low-E2F1 tumors (50.0% versus 79.8%; P = 0.0027; Fig. 3A). Especially among patients with stage II to III NSCLCs, the 3-year survival rate was significantly lower in patients with high-E2F1 tumors than in those with low-E2F1 tumors (40.1% versus 71.6%; P = 0.0188; Fig. 3B). In contrast, no difference was observed on the survival of patients with stage I NSCLCs according to the E2F1 status (Fig. 3C). As shown in Table 3, a multivariate analysis using the Cox regression model showed that the E2F1 status (hazard ratio, 2.052; P = 0.0261) and the pathologic stage (hazard ratio, 2.001; P = 0.0008) were significant prognostic factors for NSCLC patients.

We subsequently studied the overall survival of NSCLC patients in relation to the E2F1 target genes. With regard to TS status, the 3-year survival rate was significantly lower in patients with high-TS tumors than in those with low-TS tumors (59.9% versus 77.6%; P = 0.0356; Fig. 3D). Furthermore, for the E2F1 and TS status, the 3-year survival rates were 83.6% in 53 patients with both low-E2F1 and low-TS tumors, 61.8% in 28 patients with either high-E2F1 or high-TS tumors, and 53.7% in 46 patients with both high-E2F1 and high-TS tumors (Fig. 3E). The survival rate in patients with both low-E2F1 and low-TS tumors was significantly the highest among all of these groups (P = 0.0098). On the other hand, for the Survivin status, no difference was observed between the 3-year survival rate in patients with high-Survivin tumors and that in patients with low-Survivin tumors (67.1% versus 68.8%).

The E2F family plays a pivotal role in cell cycle control and apoptosis (8). E2F1 belongs to a subclass of E2F factors that are thought to act as transcriptional activators, and it is the “final frontier” of the G₁-to-S phase boundary (27). Many upstream stimulatory and inhibitory signals converge on the E2F1-pRB (retinoblastoma gene product) complex (28). Inactive E2F1 is complexed to pRB. On hyperphosphorylation of pRB as part of the transition from the G₁ to S phase, the E2F1 protein is released from the E2F1-pRB complex, and the “free E2F1” becomes available to transcriptionally activate various target genes required for DNA synthesis.

However, because E2F1 modulates diverse cellular functions, such as mitosis and apoptosis (8), in vitro studies on E2F1 reported controversial results about its function associated with tumorigenesis. Several studies have revealed that the E2F1 gene can act as an oncogene by enhancing transformation activity, tumor proliferation, and invasive ability (13, 14). In contrast, other studies reported that the E2F1 gene transfer induced apoptosis and suppressed tumor growth (15, 16). In fact, previous clinical studies also showed that E2F1 has a tumor-promoting effect in many cancers including lung cancer (17–21) whereas the opposite seems to be the case for other cancers (21, 22).

Therefore, to clarify the clinical significance of E2F1 in NSCLCs, we studied E2F1 gene expression in relation to tumor proliferation and tumor apoptosis. The present study showed that E2F1 gene expression positively correlates with tumor proliferation. However, no correlation was observed between E2F1 gene expression and tumor apoptosis. Gorgoulis et al. (17) also reported similar results with E2F1 protein expression in NSCLC. These results in clinical studies might be partly because acquired gene mutations in the apoptotic pathway, such as in p53, can neutralize the apoptotic response of E2F1 (29). It is notable that 44% (56 of 127) of NSCLCs had p53 mutations in the present study (data not shown). In addition, the present study revealed that E2F1 overexpression is associated with a poor prognosis in NSCLC patients, as also reported by Gorgoulis et al. (17). These results revealed that E2F1 overexpression has a tumor-promoting effect in NSCLCs.

Furthermore, the analyses of E2F1 target genes associated with tumorigenesis are important for the clarification of its biological functions in NSCLC. We therefore carried out a clinical study of E2F1 gene expression in relation to its target genes, TS (9, 10) and Survivin (11). Consequently, the present study identified the novel fact that E2F1 gene expression positively correlates with TS and Survivin expressions in NSCLC.

With regard to TS, it plays a central role in the biosynthesis of thymidylate, an essential precursor for DNA synthesis (30). Recent research has revealed that TS exhibits oncogene-like activity (31). Experimental studies showed that TS expression was associated with tumor cell proliferation (32). Clinical studies, including the present study, have also revealed that TS expression is positively associated with tumor proliferation (33). Furthermore, TS is a target molecule of 5-FU, one of the commonly prescribed antitumor drugs, and TS expression is related to 5-FU sensitivity (4). Many clinical studies involving human cancers, including NSCLC and gastrointestinal tumors, have shown that high TS expression is associated with 5-FU resistance and a poor outcome in cancer patients (5, 34, 35). The present study also revealed that high TS expression is associated with a poor prognosis in NSCLC patients.

It is also important to clarify the mechanisms of TS regulation. Previous experimental studies reported that E2F1 induces DNA synthesis and G₁-S regulatory genes, including TS (9). In addition, E2F1-transfected cells associated with an up-regulation of TS were more resistant to 5-FU than the cells that had not been transfected with E2F1 (10). Recent clinical studies also reported that E2F1 gene expression correlates with TS gene expression in colon cancer and osteosarcoma (36–38). The present study is the first clinical study showing the positive association between E2F1 gene expression and TS gene expression in NSCLC.

With regard to Survivin, it is the smallest member of the inhibitor of apoptosis protein family (12). Survivin has a single baculovirus inhibitor of apoptosis repeat domain that inhibits the caspase family, and the caspase family mediates apoptosis (39). Experimental studies have shown that the antisense RNA for Survivin can induce apoptosis in tumor cells (40). Furthermore, the nuclear localization of Survivin protein has been reported to affect cell mitosis (41, 42). Survivin expression is therefore predicted to promote tumorigenesis by regulating not only apoptosis but also mitosis. Previous clinical studies have revealed that Survivin is expressed in most human cancers (43–46), whereas its expression is undetectable in normal differentiated tissues (47). Furthermore, the Survivin expression is associated with apoptosis (43–45), tumor proliferation (45, 46), and a poor prognosis in cancer patients (44, 46). The present study has also showed that Survivin gene expression correlates with the enhancement of tumor proliferation and the inhibition of apoptosis in NSCLC. However, no difference was observed in the patient survival according to the Survivin status. This result might be due to the relatively small number of patients we studied.

Notably, a recent experimental study has shown that E2F1 can bind to the Survivin gene promoter and induce Survivin transcription (11). The present clinical study also revealed that E2F1 gene expression correlates positively with Survivin gene expression in NSCLCs. To our knowledge, the present study is the first clinical report showing the association between E2F1 gene expression and Survivin gene expression in human cancer. Furthermore, the up-regulation of Survivin, which can inhibit tumor apoptosis, might be partly responsible for the absence of a correlation between E2F1 overexpression and tumor apoptosis in human cancers, as shown in the present study.

The present study revealed that E2F1 gene expression was significantly higher in the advanced-stage tumors than in the early-stage tumors. Furthermore, E2F1 expression in NSCLC correlates positively with TS expression, Survivin expression, and tumor proliferation. These results indicate that, during the progression of NSCLC, E2F1 overexpression could result in the production of more aggressive tumor cells with a high proliferation rate and chemoresistance.

Although E2F1 overexpression is frequently seen in many human cancers (17–21, 36–38, 48), its biological mechanism is still unclear. Gene amplification of E2F1 has been reported to be a rare event in human cancers including NSCLC and gastrointestinal cancer (17, 18, 48). Suzuki et al. (48) reported that E2F1 protein expression was associated with the increased expression of its gene, rather than the amplification of its gene. The present study also showed that E2F1 gene expression is positively correlated with E2F1 protein expression. Therefore, deregulated E2F1 mRNA synthesis may be a chief mechanism of E2F1 overexpression (49, 50). Additional investigations are required to develop new treatment strategies for patients with E2F1-overexpressing NSCLC.

We thank Dr. Masakazu Fukushima (Taiho Pharmaceutical Co., Tokyo, Japan) for his kind gift of rabbit polyclonal antibody for TS.