Low Expression of Bax Predicts Poor Prognosis in Patients with Locally Advanced Esophageal Cancer Treated with Definitive Chemoradiotherapy

Surgical resection is still the mainstay of the curative treatment for locally advanced esophageal cancer (1–3). However, surgical resection is feasible in only one-third of patients due to unresectable disease or comorbidities (4). Although radiotherapy had been the standard treatment option for inoperable esophageal cancer, recently, concurrent chemoradiotherapy has been shown to be more effective than radiotherapy alone in patients with locally advanced esophageal cancer (1–3). A phase III trial from the United States (Radiation Therapy Oncology Group trial 85-01) showed that adding 5-fluorouracil and cisplatin (FP) combination chemotherapy to radiotherapy is superior to radiotherapy alone in patients with locally advanced esophageal cancer (1–3, 5, 6). In addition, in a systematic review that included 898 patients from nine randomized trials, concurrent chemoradiotherapy was associated with a 7% absolute reduction in both 1- and 2-year mortality rates compared with radiotherapy alone (3). Moreover, two recently published phase III trials comparing surgery after chemoradiotherapy with chemoradiotherapy alone suggested that definitive chemoradiotherapy could be considered as an alternative treatment in patients with resectable esophageal cancer (7, 8). Therefore, definitive chemoradiotherapy using the FP regimen could be considered as the standard of care in patients with both resectable and unresectable locally advanced esophageal cancer (1–3). On the other hand, FP-based concurrent chemoradiotherapy is an essential component of neoadjuvant chemoradiotherapy, which is a useful treatment option for resectable esophageal cancer, considering the survival benefit in three out of five meta-analyses (1–3, 9–11).

Because chemoradiotherapy with FP is an important component in the management of both resectable and unresectable locally advanced esophageal cancer, the determination of parameters that may identify those patients who would benefit from chemoradiotherapy and who would not has strong clinical implications. Apoptosis-related proteins are important candidates for such parameters because apoptosis plays an important role in the cytotoxic effects of chemotherapy and radiotherapy on cancer cells (12). We evaluated the prognostic significance of apoptosis-related proteins, p53, Bcl-2, Bax, and galectin-3 in patients with locally advanced esophageal cancer treated with chemoradiotherapy using the FP regimen.

Patients. All patients were required to have the following to be a candidate for chemoradiotherapy: (a) histologically documented esophageal cancer; (b) previously untreated; (c) locally advanced stages II-IV according to the American Joint Committee on Cancer (13); (d) Eastern Cooperative Oncology Group (ECOG) performance status 0-2; and (e) no significant medical disease. Patients with distant metastasis except for supraclavicular and celiac lymph nodes were excluded. Each patient underwent the following staging procedures: physical examination, esophagogastroduodenoscopy, chest radiography, computed tomography (CT) of chest and upper abdomen, esophagography, and hematologic and biochemical profiles. This research protocol was approved by the Institutional Review Board of the Ajou University Medical Center, Suwon, Korea.

Chemotherapy. After completion of diagnostic workup, chemotherapy and radiotherapy were started on the same day. Cisplatin 25 mg/m²/day and 5-fluorouracil 1,000 mg/m²/day were administered as a continuous i.v. infusion on days 1 to 3. Two cycles of chemotherapy were done during radiotherapy at 4-week intervals. Three or four weeks after completion of radiotherapy, two more cycles of FP chemotherapy with the same dose were given at 3- or 4-week intervals. In cases of partial or complete response (CR) after planned chemoradiotherapy, additional cycles of chemotherapy were administered according to the decision of the physicians. In cases of resectable disease at the time of diagnosis with response after concurrent chemoradiotherapy using two cycles of FP, a proportion of the patients underwent esophagectomy 4 to 6 weeks after completion of chemoradiotherapy according to the discretion of the physicians. In these patients, adjuvant chemotherapy with FP was done after surgery if possible.

Radiotherapy. All patients received external beam radiotherapy using 6 or 15 MV LINAC (CLINAC 2100CD, Varian Medical Systems). The initial treatment volume included the primary tumor with a radial margin of 1.5 to 2 cm and a proximal and distal margin of 5 cm and enlarged lymph nodes. The celiac trunk was generally irradiated for lower thoracic esophageal lesions. Two-dimensional or three-dimensional treatment plans using CT scans were done. A total radiation dose of 45 to 64.4 Gy (1.8 to 2 Gy/fraction, 5 days a week) was delivered with 3- or 4-field technique. The dose to the spinal cord was limited within 40 to 45 Gy, and the treatment field was reduced after 45 to 50.4 Gy.

Evaluation. Response to treatment was evaluated 4 to 6 weeks after completion of chemoradiotherapy with four cycles of FP using esophageal endoscopic examination with biopsies and chest CT including upper abdomen. In patients with planned surgical resection, a response evaluation was done about 4 weeks after completion of concurrent chemoradiotherapy with two cycles of FP. Then, the patients were evaluated with chest CT ± esophagoscopy every 3 months for 2 years, and every 6 months thereafter.

Response to treatment was evaluated according to the following criteria. CR was defined as the complete resolution of all lesions on the chest CT and esophagoscopy without detection of cancer cells on a biopsy of the primary lesion. Inflammatory or scarring change without cancer cells in the endoscopic biopsy could be included in the criteria for CR if there was no definitive lesion seen on the chest CT. Partial response (PR) was defined as a reduction by 50% or more of the sum of the products of the lesions. Stable disease was indicated by a <50% reduction or <25% increase in tumor size. Progressive disease was defined as an increase of >25% in tumor size or the appearance of new lesions.

Immunohistochemical staining for apoptosis-related proteins. Immunohistochemical staining of formalin-fixed, paraffin-embedded tumor tissue obtained by endoscopic biopsy before chemoradiotherapy was done using mouse anti-human monoclonal antibodies against p53 (DO-7; dilution, 1:20), Bcl-2 (NCL-bcl-2; dilution, 1:50) and galectin-3 (NCL-GAL3; dilution, 1:100; Novocastra Laboratories Ltd.), and rabbit anti-human polyclonal antibody against Bax (dilution 1:1,000; DAKO).

Sections were deparaffinized in xylene and rehydrated in graded alcohols and water. Endogenous peroxidase activity was blocked by treatment with 3% hydrogen peroxide for 10 min. Sections were treated with protein-blocking solution and then with primary antibodies overnight at 4°C. After several rinses in PBS, the sections were incubated in the biotinylated secondary antibody. Bound antibodies were detected by the streptavidin-biotin method with a Cap-Plus detection kit (Zymed Laboratories Inc.). Slides were rinsed in PBS, exposed to diaminobenzidine, and counterstained with Mayer's hematoxylin. The negative controls for these proteins were made by the omission of the primary antibody during the process of immunohistochemical staining. For positive controls for Bax and Bcl-2, lymphocytes in the germinal center and interfollicular area of normal lymph nodes were used, respectively. A tissue section of colon adenocarcinoma known to have high expression of p53 was used as a positive control for p53. The positive control for galectin-3 was a tissue section of the tonsil.

The slides were examined independently by two observers (J.H. Han, J.H. Kim) blinded to both clinical and pathologic data. The evaluation of the expression of apoptosis-related proteins was done as proposed by previous reports with some modifications (14–16). Expression of the apoptosis-related proteins was quantified using a visual grading system based on the extent of staining (percentage of positive tumor cells; graded on a scale of 0 to 4: 0, none; 1, 1-25%; 2, 26-50%; 3, 51-75%; 4, >75%) and the intensity of staining (graded on a scale of 0 to 3: 0, no staining; 1, weak staining; 2, moderate staining; 3, strong staining). For further analysis, we used the extent of staining as well as a product of grades of the extent and intensity of staining to define the cutoff value for high expression of the proteins. Both evaluation methods showed comparable results. Therefore, we presented the extent of staining, which was classified into high (grades 2-4) and low (grades 0 and 1) expression, for all analyses for clarity of data presentation (Fig. 1).

Statistical analysis. Overall survival (OS) was calculated using the Kaplan-Meier method (17). OS was defined as the time from start of treatment to death; data on survivors were censored at the last follow-up. The differences between the survival curves were tested by using the log-rank test. The Cox proportional-hazards regression model was used to determine the joint effects of several variables on survival (18). A comparison of clinicopathologic characteristics was evaluated with the Fisher's exact test. All analyses were done with SPSS for Windows 12.0 software.

Patient characteristics. Between September 1996 and November 2004, 63 patients with locally advanced esophageal cancer were treated with definitive chemoradiotherapy with FP. The clinicopathologic characteristics of the patients are listed in Table 1. All patients had squamous cell carcinoma except for 1 patient with adenocarcinoma, and 61 patients were male. Five patients underwent esophagectomy after concurrent chemoradiotherapy with two cycles of FP. The median delivered dose intensity, defined as actual dose/planned dose in the first four cycles (two cycles in patients with surgical resection), was 71% for both 5-fluorouracil and cisplatin. After chemoradiotherapy with two or four cycles of FP, CR and PR were achieved in 9 patients (14%) and 20 patients (32%), respectively. A planned response evaluation was not done in six patients due to the patients' refusal.

Association of expression of apoptosis-related proteins with clinicopathologic characteristics. High expression of Bax, p53, Bcl-2, and galectin-3 was observed in 42 (67%), 29 (47%), 15 (24%), and 18 (29%) patients, respectively. The expression status of these apoptosis-related proteins did not correlate with each other. Low expression of Bax was significantly correlated with the lack of clinical CR (P = 0.023). There was no significant association between the low expression of Bax and other clinicopathologic characteristics (Table 2). In addition, high expression of p53, Bcl-2, and galectin-3 did not show any significant association with clinicopathologic characteristics (data not shown).

Association of expression of apoptosis-related proteins with patient outcome. The median follow-up duration of patients was 14 months (range, 1-104 months), and no patient was lost to follow-up. Three patients were alive at the time of analysis. The median OS of total patients was 14 months with 16% of 3-year OS rate (Fig. 2A). In univariate analysis, low expression of Bax was associated with poor OS (median, 8 months versus 16 months; P = 0.0008) in addition to the lack of clinical response to chemoradiotherapy and inadequate chemotherapy or radiotherapy (Fig. 2B and Table 3). High expression of p53, Bcl-2, and galectin-3 was not correlated with patient outcome (Fig. 2C-E). Combination of low expression of Bax with the expression of other apoptosis-related proteins (low Bax/high p53, low Bax/high Bcl-2, low Bax/high galectin-3) did not reveal any prognostic significance (data not shown). In multivariate analysis, low expression of Bax was the most significant independent predictor of poor OS (P = 0.009) followed by low dose intensity of cisplatin and lack of clinical CR (Table 4).

Apoptosis is a predominant mechanism of cancer cell death by chemotherapy as well as radiation (12). Among several proteins involved in apoptosis, the fine interplay between the Bcl-2 family antiapoptotic members and death-promoting members such as Bax and p53 has been suggested as the most important process (12, 19). Bax counteracts the apoptosis-prevention effect of Bcl-2 and initiates a mitochondrial permeability shift transition, leading to the activation of downstream apoptosis signaling pathways (19, 20). In addition, galectin-3, which belongs to a family of galactoside-binding protein-3 and has both structural and functional similarity to Bcl-2, is known to protect cells from apoptosis by a variety of stimuli (21). Therefore, genetic defects in these proteins may result in resistance to the cytotoxic effects of chemotherapy and radiotherapy (22). In esophageal cancer, several studies have investigated with conflicting results the prognostic significance of apoptosis-related proteins, including p53, Bcl-2, and Bax (14–16, 23–35). We evaluated the expression of p53, Bcl-2, Bax, and galectin-3 under the assumption that abnormalities in these apoptosis-related proteins may be associated with resistance to the chemoradiotherapy with FP regimen, ultimately leading to poor survival in patients with locally advanced esophageal cancer.

The most frequently investigated apoptosis-related protein in esophageal cancer is p53 (15, 23–34). Although negative correlation between the mutation of p53 and clinical or pathologic response to chemoradiotherapy has been shown in several studies, the prognostic significance of p53 mutation is controversial (15, 23–34). High expression of Bcl-2 in esophageal cancer has not been associated with poor prognosis in previous studies (15, 35). In the present study, high expression of p53, Bcl-2, and galectin-3 was not associated with poor patient outcome.

The most important finding of the current study was the prognostic significance of the low expression of Bax. Clinical CR to chemoradiotherapy has been reported as the most important predictor of outcome for patients with locally advanced esophageal cancer treated with definitive chemoradiotherapy (36–38). In the present study, low expression of Bax was the most significant predictor of poor OS in multivariate analysis surpassing the lack of clinical CR.

The role of low expression of Bax as a poor prognostic factor has been reported in several malignancies treated with chemotherapy or radiotherapy, such as breast cancer, ovarian cancer, prostate cancer, and head and neck cancer (16, 39–41). In addition, we recently reported that the low expression of Bax was a significant independent predictor of poor OS in patients with nasopharyngeal cancer treated with FP induction chemotherapy followed by concurrent chemoradiotherapy with cisplatin, whereas expression of p53 and Bcl-2 was not associated with patient outcome (39). Therefore, the poor outcome of patients with esophageal cancer with low expression of Bax may have clinical relevance because both radiotherapy and cisplatin, which are integral components of the treatment for esophageal cancer, depend on apoptosis for their cytotoxic effects.

Although several studies have investigated the prognostic significance of Bax in esophageal cancer, only one study by Sturm et al. using tumor specimens from 53 patients who underwent curative resection showed the role of low expression of Bax as an independent predictor of poor OS (14–16, 23, 24, 29). In addition, in another study of 141 patients who underwent surgical resection, low expression of Bax was associated with poor outcome in a subset of 57 patients who underwent postoperative chemoradiotherapy (24). On the other hand, in two other studies, low expression of Bax was not associated with the poor prognosis of patients who underwent surgical resection (14, 29). Moreover, in two studies, low expression of Bax did not predict the poor survival of patients treated with chemoradiotherapy using FP-based regimens with or without surgical resection (15, 23). However, because the numbers of patients were relatively small (38 and 54 patients), and surgical resection after chemoradiotherapy was done in 45% and 93% of the study population, respectively, in the two studies, direct comparison with the current study is difficult (15, 23).

To our knowledge, the present study is the first report demonstrating the prognostic significance of low expression of Bax in patients with locally advanced esophageal cancer treated with definitive chemoradiotherapy. Moreover, low expression of Bax was also significantly associated with poor response to chemoradiotherapy. Although the median survival of the patients in the current study was almost identical to that of the Radiation Therapy Oncology Group (RTOG) 8501 trial (14.0 months versus 14.1 months), the 3-year OS rate was relatively low (16% versus 30%; refs. 5, 6). The inclusion of a significant number of patients with more advanced disease (stage IV, 36%) compared with the RTOG 8501 trial (up to stage III) and chemoradiotherapy in the practice setting in the present study may be possible explanations for the difference in survival (5, 6). The OS of patients with low expression of Bax was very poor (median, 8 months; 2 years, 5%), which is similar to that of the radiotherapy alone group in the RTOG 8501 trial (5, 6).

The poor survival of patients with low expression of Bax may be attributable to the intrinsic aggressiveness of tumor, considering the results of Sturm et al., which showed the poor outcome of patients who underwent resection of the esophageal cancer with low expression of Bax (16). Alternatively, poor response to chemoradiotherapy in tumors with low expression of Bax could cause the poor survival of the patients because there was no clinical CR in patients with low expression of Bax compared with 21% in those with high expression in the current study.

The present study has several potential limitations despite the demonstration of the prognostic significance of low expression of Bax in patients with locally advanced esophageal cancer treated with chemoradiotherapy. First, this study is a retrospective analysis with a relatively small sample size. Second, a proportion of the patients did not undergo the planned response evaluation due to treatment in practice setting rather than clinical trial, although the follow-up of survival status was complete in all patients. Third, five patients received surgical resection after concurrent chemoradiotherapy instead of definitive chemoradiotherapy alone. However, low expression of Bax was still a significant independent predictor of poor prognosis in 58 patients with definitive chemoradiotherapy alone (data not shown). Finally, there is a possibility that the immunohistochemical staining results of the pretreatment endoscopic biopsy specimen, as in the present study, do not correlate with those of the surgically resected specimen.

Although further prospective studies with large numbers of patients are warranted to confirm the role of low expression of Bax as a prognosticator of esophageal cancer, the present results have several clinical implications. The results of the current study suggest that the prognosis of patients with low expression of Bax would be poor after chemoradiotherapy with FP, which plays a central role in the management of locally advanced esophageal cancer. In addition, concurrent chemoradiotherapy with FP is usually associated with significant toxicity. Therefore, it is reasonable to consider an alternative treatment strategy for patients with low expression of Bax. Although immediate surgical resection, preferably followed by adjuvant chemotherapy or radiotherapy, seems to be an appropriate approach in patients who can undergo resection, surgical resection in the case of resectable PR after chemoradiotherapy or chemoradiotherapy incorporating new chemotherapeutic agents with high activity in esophageal cancer, such as taxanes and irinotecan, could be treatment options for patients who are unable to undergo resection (1, 3).

In conclusion, low expression of Bax was significantly associated with the poor survival of patients with locally advanced esophageal cancer treated with chemoradiotherapy using FP. If the prognostic significance of the low expression of Bax is validated by further prospective studies with larger numbers of patients, a relatively simple immunohistochemical staining for Bax with a pretreatment endoscopic biopsy specimen may provide valuable information to the oncologist for the selection of patients for combined modality therapy incorporating FP-based chemoradiotherapy or alternative treatment.

The authors are grateful to Geum Sook Jeong for administrative assistance in preparing and submitting the manuscript.