Breast Cancer Resistance Protein Impacts Clinical Outcome in Platinum-Based Chemotherapy for Advanced Non-Small Cell Lung Cancer Free

Lung cancer is a major cause of death from cancer worldwide, and non-small cell lung cancer (NSCLC) accounts for ∼75% of all cases of lung cancer. More than half of NSCLCs are advanced stage IIIB or IV disease at presentation, and patients with advanced NSCLC are candidates for systemic chemotherapy. Meta-analyses have demonstrated that cisplatin-based chemotherapy for metastatic NSCLC statistically improves patient survival compared with supportive care alone (1). However, the response rate to chemotherapy has been poor, only 30%, and very few patients have survived for 5 years (2). During the 1990s, five new drugs became available for the treatment of metastatic NSCLC, paclitaxel, docetaxel, vinorelbine, gemcitabine, and irinotecan. Each of them has since been evaluated in combination regimens with cisplatin or carboplatin and produced responses in 40–50% of patients (3). Unfortunately, despite the increasing number of active chemotherapeutic agents, none of these chemotherapeutic regimens has offered a significant advantage over the others in the treatment of advanced NSCLC in randomized studies (4, 5), and advanced NSCLC patients still have a median survival time of <1 year.

Intrinsic or acquired tumor-mediated drug resistance is a major clinical problem that can result in the lack of tumor response to chemotherapy. In vitro selection for resistance to a single chemotherapeutic agent often results in cross-resistance to various structurally and functionally dissimilar drugs, a phenomenon called multidrug resistance (MDR; Ref. 6). Several members of the superfamily of ATP-binding cassette (ABC) transporter proteins can be involved in MDR in human tumor cells (7), including P-glycoprotein (Pgp), MDR protein (MRP) 1, MRP2, and MRP3 (8, 9). Although in vitro studies have revealed mechanisms of cytotoxic drug resistance in cancer cells (10), there is no convincing evidence that the expression of ABC transporter proteins in clinical samples of NSCLC predicts response to chemotherapy or prognosis.

Breast cancer resistance protein (BCRP), a new member of the superfamily of ABC transporter proteins, also known as mitoxantrone resistance protein, ABCP, and ABCG2, was identified in a MCF-7 breast cancer cell subline that selected for resistance to doxorubicin (11). Elevated expression of BCRP in vitro causes resistance to anticancer drugs, including topotecan, irinotecan, mitoxantrone, and doxorubicin (11, 12, 13, 14). In the past, reports on BCRP expression have mainly relied on measurement of mRNA levels in cancer cell lines and acute leukemia samples (14, 15). However, two monoclonal antibodies specific for human BCRP, BXP-34, and BXP-21, have been produced recently, and BCRP has been shown to be expressed mainly at the plasma membrane in MDR cells and to be present in several normal tissues, including placenta, liver canaliculi, small intestine, colon, the bronchial epithelial layer in the lung, and endothelial cells (16, 17, 18). BXP-21 was used to investigate BCRP expression in a panel of 150 untreated human solid tumors comprising 21 tumor types, and the results suggested that BCRP is expressed frequently in chemotherapy-resistant tumor types, such as tumors of the digestive tract (colon, esophagus, and stomach), endometrium, and lung (19). However, whether BCRP expression is involved in the response to chemotherapy and prognosis of human solid cancers, including NSCLC, has remained unknown.

In this retrospective study, we investigated the level of expression of BCRP, in addition to Pgp, MRP1, MRP2, and MRP3, in clinical samples of advanced NSCLC, and investigated whether its expression predicts response to chemotherapy and prognosis.

A total of 167 stage IIIB or IV NSCLC patients received platinum-based combination chemotherapy at the National Cancer Center Hospital East between January 1996 and December 2001 because of being PS 0 or 1 on the Eastern Cooperative Oncology Group scale. Seventy-two of these patients had adequate tumor biopsy specimens obtainable before chemotherapy and were analyzed in this study. All of the tumor specimens were obtained before chemotherapy, by bronchoscopy from 46 patients, by percutaneous needle biopsy from 20 patients, by thoracotomy from 5 patients, and by mediastinoscopy from 1 patient. The histological classification was based on a WHO report. Clinical staging was based on an initial evaluation consisting of a clinical assessment, chest radiography, computed tomography of the chest and abdomen, computed tomography or magnetic resonance imaging of the brain, and bone scintigraphy. The current International Staging System was used for clinical disease staging (20). The clinicopathological characteristics of all of the patients are listed in Table 1. Their median age at diagnosis was 66 years (range, 44–76 years). Six of the 32 stage IIIB patients were women, and 17 of the 40 stage IV patients were women. All of the patients were treated with platinum-based combination chemotherapeutic regimens, which were considered to be standard regimens for patients with metastatic NSCLC (3). The median follow-up time of the 72 patients was 20 months (range, 2–40 months).

After obtaining informed consent in accordance with institutional guidelines, all of the patients underwent tumor biopsy and chemotherapy.

The platinum-based regimens were vinorelbine 25 mg/m² on days 1 and 8 plus cisplatin 80 mg/m² on day 1 of a 21-day cycle (25 patients), docetaxel 60 mg/m² on day 1 plus cisplatin 80 mg/m² on day 1 of a 21-day cycle (16 patients), irinotecan 60 mg/m² on days 1, 8, and 15 plus cisplatin 80 mg/m² on day 1 of a 28-day cycle (11 patients), gemcitabine 1000 mg/m² on days 1 and 8 plus cisplatin 80 mg/m² on day 1 of a 21-day cycle (10 patients), and paclitaxel 200 mg/m² administered over 3 h on day 1 plus carboplatin dosed with an area under the curve of 6 on day 1 of a 21-day cycle (10 patients). All of the patients received 2 or more courses of chemotherapy or until the appearance of progressive disease. We used the standard response criteria (21) to evaluate the response to chemotherapy. A complete response was defined as the disappearance of all of the clinically detectable disease for at least 4 weeks. A partial response required a minimum of a 50% reduction in the sum of the products of the greatest perpendicular diameters of all of the measurable lesions for a minimum of 4 weeks. Progressive disease was defined as the appearance of new lesions or an increase in disease of >25% measured in the same manner as for partial response. All of the other results were classified as “no change.” The response rate was defined as the total of the complete response cases and partial response cases expressed as a percentage of all of the cases. PFS (progression-free survival) was measured from the start of chemotherapy to the time progressive disease was documented or death occurred.

Immunostaining was performed on 4 μm formalin-fixed, paraffin-embedded tissue sections. The slides were deparaffinized in xylene and dehydrated in a graded ethanol series. Endogenous peroxidase was blocked with 0.3% H₂O₂ in methanol for 10 min. For antigen retrieval, the slides for Pgp, MRP1, MRP2, and BCRP were immersed in 10 mm citric buffer solution (pH 6.0); the slides for MRP3 were immersed in 1 mm EDTA retrieval fluid (pH 8.0). All of the slides were heated to 95°C by exposure to microwave irradiation for 20 min. The slides were then cooled for 1 h at room temperature and washed in water and PBS. Next, nonspecific binding was blocked by preincubation with 2% BSA plus 0.1% NaN₃ for 30 min. Blocking serum was drained off, and the slides were incubated overnight at 4°C with the primary antibodies listed in Table 2. Staining with an irrelevant mouse IgG1 or IgG2a was routinely performed as a negative control procedure. After washing three times in PBS, the slides were incubated with a labeled polymer EnVision+, Peroxidase Mouse (DAKO, Glostrup, Denmark) for 60 min. The chromogen used was 2% 3,3′-diaminobenzidine in 50 mm Tris buffer (pH 7.6) containing 0.3% hydrogen. Slides were counterstained with hematoxylin. Normal liver and lung tissue human was used as a positive control (17).

Staining of all of the antibodies was considered positive if >10% of the tumor cells stained, because the 10% cutoff level has actually been used in several studies on MDR (19, 22). All of the slides were examined and scored independently by two observers without knowledge of the patient clinical data. When the antibody evaluations differed between observers, they discussed it, with or without reevaluating the slides, until an agreement was reached.

The correlations between immunohistochemical expression and the clinical variables and response to chemotherapy were evaluated by the χ² test or Fisher’s exact test, as appropriate. PFS was used as a clinical marker for duration of response to chemotherapy. Overall survival was measured from the start of chemotherapy to the date of death from any cause or the date the patients were last known to be alive. Survival curves were estimated by the Kaplan-Meier method, and differences in PFS and survival between subgroups were compared by using the log-rank test. The Cox proportional hazards model was used for multivariate analysis. Ps <0.05 were considered significant. Two-sided statistical tests were used in all of the analyses. Statistical analysis software (StatView-J Ver.5.0, Macintosh) was used for the analyses.

Twenty-five (35%) of the 72 patients were Pgp positive, 56 (78%) were MRP1 positive, 12 (17%) were MRP2 positive, 10 (14%) were MRP3 positive, and 33 (46%) were BCRP positive. Most of ABC-transporter-protein-positive tumors showed mixed membranous and cytoplasmic staining. Representative immunohistochemical BCRP staining is shown in Fig. 1,A. BCRP in the apical membrane of the bronchial layer was used as an internal control, and the endothelial cells of blood vessels also stained positive (Fig. 1, B and C). The relationship between expression of ABC transporter proteins and clinical variables is shown in Table 3. Pgp expression was significantly greater in females than in males (P = 0.03) and in nonsquamous cell carcinoma than in squamous cell carcinoma (P = 0.02). MRP2 expression was significantly greater in stage IIIB than in stage IV (P = 0.0007). There were no significant correlations between expression of MRP1, MRP3, or BCRP and the clinical variables.

All of the 72 patients could be assessed for response to chemotherapy and were analyzed for survival. The relationships between clinical variables and response to chemotherapy and survival in this study are shown in Table 4. Six of the 32 stage IIIB patients received thoracic radiotherapy after the completion of chemotherapy, and 3 of the patients were women. Only “stage” was significantly associated with both PFS (P = 0.003) and overall survival (P = 0.01). Table 5 shows the relationship between expression of ABC transporter proteins, and response to chemotherapy and survival. No significant associations were found among expression of Pgp, MRP1, or MRP3, and either response to chemotherapy or survival. MRP2 expression was associated with overall survival (P = 0.002) but not with response to chemotherapy and PFS. By contrast, the response rate to chemotherapy of patients with BCRP-negative tumors was 44%, as opposed to 24% in patients with BCRP-positive tumors. Response rate was lower in BCRP-positive tumors, although this difference was not statistically significant (P = 0.08). The patients with BCRP-positive tumors also had a significantly shorter PFS (P = 0.0003) and overall survival (P = 0.004) than patients with BCRP-negative tumors (2.8 versus 5.1 months and 6.8 versus 15.2 months, respectively).

Multivariate analysis was performed by using the Cox proportional hazards model to determine whether the prognostic value of BCRP disappeared when other prognostic factors were considered (Table 6). A multivariate analysis that included histology, stage, PS, Pgp, MRP1, MRP2, MRP3, and BCRP, showed that stage, PS, and BCRP were significant independent variables correlated with PFS (P = 0.042, P = 0.049, and P = 0.001, respectively). The BCRP-positive value for PFS yielded a hazard ratio of 2.85, with 95% confidence interval ranging from 1.48 to 5.50. No correlation between variables and overall survival was found on multivariate analysis.

The PFS and overall survival curves by the Kaplan-Meier method are shown according to BCRP in Fig. 2. Median survival time in the BCRP-negative group was 15.2 months, as opposed to 6.8 months in the BCRP-positive group. On the basis of the Kaplan-Meier method the estimated 1-year and 2-year survival rates in the BCRP-negative group were 64% and 58%, respectively, as opposed to 33% and 18%, respectively, in the BCRP-positive group.

This is the first study to investigate the relationship between the level of expression of ABC transporter proteins, including BCRP, and the clinical response to chemotherapy and prognosis of previously untreated patients with advanced NSCLC. The analyses of the relationships between expression of ABC transporter proteins and clinical outcome showed that only expression of BCRP before chemotherapy was significantly correlated with PFS and overall survival. In addition, pretreatment BCRP status was an independent prognostic factor of PFS in the multivariate analysis, although the number of patients was relatively small, and the results should be interpreted with caution. These results suggest that positive immunostaining for BCRP in tumor specimens from patients with untreated advanced NSCLC may be a predictor of a poor clinical outcome to platinum-based combination chemotherapy.

The newly identified BCRP gene, which encodes a predicted 655-amino acid protein of M_r ∼72,000 having six transmembrane domains, was originally detected almost exclusively in mitoxantrone-exposed tumor cells, and it appeared to be up-regulated after exposure to topotecan or SN-38 (13, 23). Transfection experiments with BCRP cDNA showed that the phenotype of atypical MDR that is not mediated by Pgp and MRP could be transferred to formerly drug-sensitive cancer cells (8, 14). Although the resistance to chemotherapy exhibited by NSCLC is multifactorial and cannot be explained by the MDR phenotype mechanism alone, our findings suggest that BCRP may serve as a molecular target for reducing drug resistance to chemotherapy and the poor prognosis of patients with advanced NSCLC. Therapeutic intervention with BCRP-specific inhibitors has been reported recently, and GF120918 has been shown to be a potent inhibitor of human BCRP (24, 25).

A MDR phenotype with cellular drug resistance has been intensively studied, and it has been found to be characterized by a reduced intracellular drug level and overexpression of individual members of the ABC superfamily of membrane transporters (26). However, few studies of clinical samples of advanced NSCLC had demonstrated a role for the ABC transporter proteins Pgp and MRP as a major determinant of response to chemotherapy and survival (27, 28). Our results are also consistent with past studies in showing that the expression of Pgp, MRP1, MRP2, and MRP3 is not an independent prognostic factor.

A major drawback of this study is that the chemotherapeutic regimen administered to all of the patients was not the same, and the number of patients was small when the analyses were performed after dividing them according to chemotherapeutic regimen. However, our results showed that a significant relationship between BCRP expression and poor response to chemotherapy and prognosis in the group that received the regimen consisting of irinotecan as a topoisomerase I inhibitor plus cisplatin (data not shown). The facts were consistent with experiments in vitro that showing BCRP is an efficient transporter of topotecan and other topoisomerase I inhibitors (12, 13). This suggest that combined chemotherapy with a topoisomerase I inhibitor and a BCRP-specific inhibitor will be highly effective.

In conclusion, our findings indicated that expression of the novel ABC transporter protein BCRP, but not of Pgp, MRP1, MRP2, and MRP3, is a predictive factor of poor clinical outcome in advanced NSCLC patients treated with standard platinum-based combination chemotherapy. Although a prospective study with a larger number of patients is warranted, we believe that these results have important implications for the treatment of NSCLC, because BCRP may serve as a molecular target for reducing drug resistance to chemotherapy in patients with advanced NSCLC.

We thank Chie Okumura and Yoko Okuhara for technical assistance.