Combined analysis of mRNA expression of ERCC1, BAG-1, BRCA1, RRM1 and TUBB3 to predict prognosis in patients with non-small cell lung cancer who received adjuvant chemotherapy

A total of 85 patients who underwent curative surgery for NSCLC between August 2007 and April 2009 were enrolled into this study, including 85 tumor tissues and 34 adjacent tissues respectively. Among them there were 60 males and 25 females, aged 24-84 (mean 57) years. According to WHO Classification (2000), there were 25 squamous, 60 adenomatous, with 58 moderate and well differentiated (G1-G2) and 27 low differentiated (G3). Because there were only 4 cases of stage IV patients who all had surgery after single brain metastasis resected firstly, and there were no patients of stage IIIb. On account of stage IV patients were too few, so we combined 48 cases as staged I-II and 37 III-IV based on the revised AJC staging for lung cancer (1997). 28 cases had intra-thoracic lymph node metastasis (N1-N2), and 57 were negative lymph node metastasis. Additional information of surgery and chemotherapy status were all showed in (Table 1). The paracancerous tissues (defined as more than 5 cm away from the carcinoma tissue) taken from 34 cases were used as controls. A total of 74 Patients received at least two cycles of adjuvant chemotherapy within three months after surgery in the First Afflicted Hospital of Guangxi Medical University. 34 of 74 patients were received GP (Cisplatin 75 mg/m² on day 1, Gemcitabine 1000 mg/m² on days 1,8), 29 of 74 patients were received NP (Cisplatin 75 mg/m² on day 1, Vinorelbine 25 mg/m² on days 1 + 8), the other 11 patients were received TP (Carboplatin AUC 6 on day 1, Paclitaxel 175 mg/m² on day 1), every 3 weeks. All of the tumor tissue samples were freshly frozen in liquid nitrogen immediately after surgery, and stored at -80 ⁰ C until analysis was available. We took out the specimens from the parenchymal tissues of tumor, and we must as far as possible make the specimens keep away from the necrotic tissue. We also confirmed the HE stain results from the pathology department after surgery, which tumor sections, from the location specimens taken by us, were full of tumor cells (usually more than 60%-70%). Patients who received neoadjuvant chemotherapy or neoadjuvant radiotherapy were excluded. The study protocol was approved by the Ethical Committee of the First Affiliated Hospital of Guangxi Medical University, China. All subjects signed an informed consent before entry into the study.

Fresh frozen specimens of tumor and adjacent tissues were obtained from 85 patients. Collection time from resection to freezing was required 20 minutes or less for all specimens. The fresh frozen specimens were processed for RNA isolation and reverse-transcriptase polymerase chain reaction (RT-PCR) in detecting expression analysis for the ERCC1, BAG-1, BRCA1, RRM1, and TUBB3 genes. Specimens were microscopically examined to assess quality and to verify the histopathology.

Specimens were pulverized by pulp refiner under Trizol reagent (Invitrogen). Total RNA was extracted with Trizol reagent and dissolved in DEPC water. Total RNA were reverse transcribed with RevertAid™ First Strand cDNA Synthesis Kit (Fermentas) for generation of cDNA. Gene expression for ERCC1, BAG-1, BRCA1, TUBB3, RRM1 and β-actin (internal reference gene) were performed using RT-PCR. The preliminary experiment and large sample experiment PCR were carried out as follows: an initial denaturation at 94°C for 3 min 30 s, followed by 30 cycles of denaturation at 94°C for 40 s, annealing at different temperature for different gene for 40 s, and elongation at 72°C for 50 s, then elongation 72°C for 7 min at last. The sequences of the primers used were in Table 2. All of these primers were checked and met a high specificity by BLAST function in NCBI. Confirmative PCR products through gene sequencing were used as positive controls to exclude false negative, and the no template added reaction system used as negative controls to exclude contamination of genomic DNA (Figure 1).

Tumor specimens from 85 patients were available for the analysis of these genes mRNA. The specimens included 85 tumor tissues and 34 adjacent tissues. The positive rate of ERCC1 mRNA in tumor and its adjacent tissues were 58.8% and 55.9% respectively (P = 0.769). BAG-1 were 37.6% and 82.4% (P = 0.000). BRCA1 were 16.5% and 44.1% (P = 0.002). RRM1 were 30.8% and 38.2% (P = 0.105). TUBB3 were 16.5% and 2.9% (P = 0.089). We chose some of the same samples which ERCC1 mRNA expressions were positive in order to validate the results. Expression of ERCC1 proteins was assessed by immunohistochemistry, and expression of the ERCC1 proteins was detected in the nuclei of cancer cells. All the samples selected were ERCC1 positive, including 6 squamous carcinoma and 6 adenocarcinoma (Figure 2).

The expression of five genes in different clinical features were compared and summarized. It showed that the difference of these five genes were only significant between some parts of clinical features. Correlations were observed between ERCC1 expression and TNM stage (P = 0.006), metastasis of lymph node (P = 0.01), and TUBB3 expression and TNM stage (P = 0.004). No Correlation was observed between ERCC1, TUBB3 expression and other clinical features. Besides, No Correlation was observed between BAG-1, BRCA1, RRM1 expression and gender, age, nationality, histology, differentiation of tumor, metastasis of lymph node, TNM stage, chemotherapy status or performance status.

The median follow-up time was 23.3 months (range 2.3-42.6), and the median overall survival and median PFS (progression-free survival) were 27.2 months (range 2.3-42.6) and 26.5 months (range 0.8-42.6), respectively. Figures 3, 4, 5 and 6 showed the Kaplan-Meier survival curves in patients positive and negative for ERCC1 and BAG-1 expression. Patients negative for ERCC1 expression had a significantly longer median progression-free (more than 42.6 vs. 15.4 months. P = 0.001) and overall (more than 42.6 vs. 20.9 months. P = 0.001) survival, compared with those positive for ERCC1 expression. Patients negative for BAG-1 expression had a significantly longer median progression-free survival (more than 42.6 vs. 12.9 months. P = 0.001) and overall survival (more than 42.6 vs. 17.0 months. P = 0.001), than those positive for BAG-1 expression. The relationships between the PFS and BRCA1, RRM1 and TUBB3 were no statistical significance (P = 0.088, P = 0.116 and P = 0.271), and there were also the same results for OS (P = 0.057, P = 0.110 and P = 0.342).

Median value of clinicopathologic factors and expression of genes of tumor samples were used as a cut-off point at univariate analysis. Univariate Cox analysis was carried out to identify the factors that were significantly associated with progression-free and overall survival (Table 3). In the univariate analysis, ERCC1 expression (P = 0.001), BAG-1 expression (P = 0.001), TNM stage (P = 0.007) and metastasis of lymph node (P = 0.006) were prognostic of progression-free survival. For overall survival, ERCC1 expression (P = 0.002), BAG-1 expression (P = 0.001), TNM stage (P = 0.008) and metastasis of lymph node (P = 0.007) were also prognostic. None of the other factors examined were statistically correlated with progression-free or overall survival for ERCC1, BAG-1 and the other three genes.

Multivariate Cox regression analysis was performed to evaluate the influence of these genes on the progression-free survival adjusting for possible confounding factors. From the results of the univariate analysis, TNM stage and metastasis of lymph node, also ERCC1 and BAG-1 were significantly correlated to the progression-free survival (Table 4). After multivariate analysis, ERCC1 was statistically significant (P = 0.018) and the hazard ratio was 0.0427 (95% CI: 0.211-0.864). BAG-1 was also statistically significant (P = 0.017) and the hazard ratio was 0.0474 (95% CI: 0.257-0.874). However, the P-value for TNM stage (P = 0.340, 95% CI: 0.336-1.457) and lymph node (P = 0.217, 95% CI: 0.299-1.315) were not statistically significant.

Multivariate Cox regression analysis was also performed for the overall survival. In addition to ERCC1, BAG-1, TNM stage and metastasis of lymph node were included in the Cox models. The other gene of TUBB3, BRCA1 and RRM1, together with the rest of patients' and clinical characteristics were not included. Negative ERCC1 and BAG-1 expression were independent and significant predictor of favorable outcome for overall survival (P = 0.027 and P = 0.022), with a hazard ratio of ERCC1 was 0.447 (95% CI: 0.219-0.911); for BAG-1, with a hazard ratio of 0.486 (95% CI: 0.262-0.901), whereas TNM stage and metastasis of lymph node had no significant association. The reason that TNM staging and lymph node were not associated with survival in the multivariate analysis might be the statistical significance of the two characteristics with survival contained in the other variables (ERCC1 and BAG-1). The other explanatory reason might be the limit of sample size.

74 of 85 patients received at least two cycles of adjuvant chemotherapy, of whom 66 (89.2%) finished at least 4 cycles. The main chemotherapy regimens included gemcitabine (GEM, 45.9%), vinorelbine (NVB, 39.2%) and paclitaxel (PTX, 14.9%) combined with cisplatin (DDP)/carboplatin (CBP).

In 74 patients treated with the regimen of cisplatin/carboplatin, patients negative for ERCC1 expression had a significantly longer median progression-free (more than 42.6 months vs. 13.0 months, P = 0.001) and overall (more than 42.6 months vs. 19.7 months, P = 0.001) survival, compared with those positive for ERCC1 expression (Figures 7, 8). Patients negative for BAG-1 expression also had a significantly longer median progression-free survival (29.0 months vs. 11.2 months, P = 0.002) and overall survival (32.3 months vs. 15.2 months, P = 0.002), than those positive for BAG-1 expression (Figures 9, 10). Whereas, there was no statistical significance in progression-free and overall survival to patients with BRCA1 expression (P = 0.129 and P = 0.073, respectively). In those treated with the regimen of gemcitabine, there was no statistical significance found in progression-free and overall survival for patients with RRM1 expression (P = 0.310 and P = 0.299, respectively). In the anti-tubulin regimen group of vinorelbine or paclitaxel, no statistical significance was found in progression-free and overall survival between the negative and positive expression of TUBB3 (P = 0.745 and P = 0.742, respectively); in the same measure, no statistical significance was found in progression-free and overall survival between the negative and positive expression of BRCA1 (P = 0.612 and P = 0.389, respectively).

There were 25 cases that expressed both ERCC1 and BAG-1 and 27 cases that expressed neither. As shown in Table 5, the correlation was found between ERCC1 and BAG-1 gene expression (P = 0.042, r = 0.247). All 52 patients of both positive and negative expression were received adjuvant chemotherapy. For both negative mRNA expression had a significantly longer median progression-free (more than 42.6 months vs. 8.8 months, P = 0.000) and overall (more than 42.6 months vs. 17.0 months, P = 0.000) survival, compared with those positive for both ERCC1 and BAG-1 expression (Figures 11, 12).