In patients with nonsmall cell lung cancer (NSCLC), several studies have demonstrated a positive correlation between somatic mutation in the epidermal growth factor receptor (EGFR) tyrosine kinase domain and clinical outcomes with the use of EGFR tyrosine kinase inhibitors (TKIs). However, some patients with wild-type (WT) EGFR also responded to EGFR TKIs and remained stable. Recently, amphiregulin (AR) has been suggested as a predictive marker for EGFR TKIs in patients with WT EGFR-positive NSCLC. The objective of the current study was to evaluate the association between AR expression and the efficacy of using EGFR TKIs in the treatment of patients with WT EGFR-positive NSCLC.
Seventy-three patients with WT EGFR-positive NSCLC received treatment with gefitinib or erlotinib between May 2005 and December 2008. AR expression was assessed by immunohistochemistry.
The clinical response to EGFR TKIs was reassessed for all patients as follows: 16 of 73 patients had a partial response (21.9%), 12 patients had stable disease (16.5%), and 45 patients had progressive disease (61.6%). AR expression was positive in 24 of 40 patients (60%). The ability to achieve disease control did not differ significantly between AR-positive patients and AR-negative patients (P = .188). At a median follow-up of 25.4 months (range, 10.5-53.3 months), progression-free survival was 8.1 weeks in AR-positive patients and 4 weeks in AR-negative patients (P = .025), and overall survival was significantly longer in AR-positive patients than in AR-negative patients (12.2 months vs 4.1 months; P = .001).
Nonsmall cell lung cancer (NSCLC) is the major cause of cancer-related deaths worldwide. Platinum-based combination regimens as a first-line treatment have offered a modest survival advantage in patients with advanced NSCLC, but most patients eventually experienced disease progression.1
The epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs) gefitinib and erlotinib, as molecularly targeted agents, are used commonly, and several large, randomized trials have demonstrated a survival benefit for patients with NSCLC who progressed on previous chemotherapy.2-5 Compared with other cytotoxic agents like docetaxel or pemetrexed, EGFR TKIs can be maintained for a long time because of their low toxicity and good tolerability.2, 6 Although the response rate with EGFR TKIs in unselected patients with NSCLC was <10%, disease control was achieved by >40% of patients.4 It is well known that patients who have an activating EGFR mutation (exons 18-21) achieve a higher response rate and prolongation of progression-free survival (PFS) when they receive EGFR TKIs.7-12 Therefore, EGFR mutation is considered a predictive marker for response and PFS. It is noteworthy that some patients with wild-type (WT) EGFR also have a response to EGFR TKIs and maintain stable disease (SD), suggesting that other mechanisms may be involved. Thus, several biomarkers have been investigated in WT EGFR-positive NSCLC. The high gene copy numbers of EGFR, v-erb-b2 erythroblastic leukemia viral oncogene homolog 2 (HER-2), and ErbB-3 are associated with sensitivity to gefitinib.13-15
In contrast to EGFR expression, EGFR ligand expression has not been a focus of study because there are so many ligands for EGFR, such as epidermal growth factor (EGF), transforming growth factor-alpha (TGF-α), amphiregulin (AR), epiregulin, betacelluin, and heparin-binding epidermal growth factor (HB-EGF). Recently, a role for AR has been suggested in tumor growth and survival in which AR stimulates an autocrine loop through EGFR in various cancers.16-18 The binding of AR to EGFR induces autophosphorylation of the EGFR intracellular tyrosine kinase domain, which activates the mitogen-activated protein kinase/extracellular signal-regulated kinase 1 and 2 (MEK/ERK1/2) pathway and the phosphoinositide 3-kinanse/protein kinase B (PI3K/AKT) pathway. Consequently, cancer cells have growth, proliferation, migration, or invasiveness properties after the activation of EGFR.19 Yonesaka et al reported that the autocrine production of AR can predict sensitivity to both gefitinib and cetuximab in WT EGFR cancers.18 Also, increased amphiregulin levels in serum were associated with a poor response to gefitinib in patients with advanced NSCLC.20 However, to date, a definite correlation between AR and response to EGFR-TKIs has not been fully established. Thus, the objective of the current study was to evaluate the association between AR expression and the efficacy of EGFR-TKIs in patients with WT EGFR-positive NSCLC.
MATERIALS AND METHODS
Patients and Treatment
Between May 2005 and December 2008, 140 patients who received EGFR-TKI treatment and also had adequate tissue samples available for EGFR mutation testing were analyzed. Initially, DNA direct sequencing was used and reconfirmed by peptide nucleic acid-locked nucleic acid polymerase chain reaction clamp. Sixty-seven of 140 patients (47.9%) had an EGFR-activating mutation of exon 19 deletion or L858R. The high EGFR mutation rate is attributed to the enriched patient population, which had favorable clinical parameters. Seventy-three patients (52.1%) had the WT EGFR gene. Patients received either gefitinib or erlotinib at their physician's discretion. All patients had histologically confirmed, advanced/metastatic or recurrent NSCLC. Oral gefitinib was given daily at a dose of 250 mg per day, and oral erlotinib was given at a dose of 150 mg per day on a treatment cycle of 4 weeks. These EGFR TKIs were continued until patients developed disease progression or unacceptable toxicity. Response evaluation was assessed according to Response Evaluation Criteria in Solid Tumors (version 1.0) guidelines.21 Response was evaluated every 6 weeks or 8 weeks during chemotherapy with a chest computed tomography scan. Forty tissue samples from the 73 patients were available for AR expression analysis. Specimens were collected by lymph node biopsy,11 gun biopsy,4 bronchoscopic biopsy,8 pleural biopsy,3 and surgical specimen.14 This analysis was approved by the Samsung Medical Center institutional review board.
Immunohistochemistry for Amphiregulin Expression
Histologic sections were cut from previously formalin-fixed, paraffin-embedded blocks for immunohistochemical staining. These formalin-fixed, paraffin-embedded tissue blocks were sectioned to 4 μm thickness. Then, the tissue sections were deparaffinized in xylene and rehydrated in serially graded alcohol. AR antigen retrieval consisted of heating in 10 mM citrate buffer, pH 6.0, with microwaves (for 15 minutes at 700 W) and cooling to room temperature for 20 minutes. After washing in phosphate-buffered saline (PBS), the slides were preincubated in 5% normal blocking solution (goat serum) for 10 minutes to reduce nonspecific binding. Then, the slides were incubated overnight at 4°C with goat anti-AR (AF262; R&D Systems, Minneapolis, Minn) at a 1:50 dilution in a humidified chamber. The primary antibody was observed with a 2-step, biotin-free polymer-horseradish peroxidase detection system (Golden Bridge International, Mukilteo, Wash). Finally, the sections were washed again in PBS, and color development was achieved using 3,3′-diaminobenzidine tetrahydrochloride (Golden Bridge International). Finally, the sections were counterstained with Mayer hematoxylin.
Evaluation of Immunostaining
Immunohistochemical evaluations were performed independently by pathologists (C.-K.J. and Y.-L.C.) who had no knowledge of the clinical data. AR immunoreactivity levels in each section were assessed under a light microscope. Representative areas of the images were acquired at ×400 magnification for each specimen. Tumor staining intensity was graded on a scale from 0 to 4, and the fraction of cell staining ranged from 0% to 100%. Then, this fraction score was multiplied by the staining intensity to obtain a final “H score.” AR-positive tumors were defined as those with H scores >100.9, 18
Clinical and pathologic variables were compared across groups using Student t tests and chi-square tests or Fisher exact tests for continuous and categorical variables. The duration of overall survival (OS) was calculated from the first date of EGFR-TKI treatment until the date of either death or the most recently documented follow-up. Progression-free survival (PFS) was calculated from the first date of EGFR-TKI treatment to the date when disease progression was recognized or the date of last follow-up. OS and PFS were estimated with the Kaplan-Meier method. Comparisons between groups were performed using the log-rank test for time-to-event variables. P values <.05 were considered statistically significant. Multivariate analysis of the independent predictive or prognostic factors for survival was performed using a backward, stepwise method (likelihood-ratio statistics based on the conditional parameter estimate: conditional likelihood ratio) of the Cox proportional hazard regression model with 95% confidence intervals (CIs).
The median age of the 73 patients was 59 years (range, 33-82 years), and 53.4% of patients were women. Fifty-eight patients had adenocarcinoma (79.5%), and 50 patients were never smokers (68.5%). The treatment lines were as follows: 5 patients (6.9%) received an EGFR TKI as neoadjuvant therapy, 15 patients (20.5%) received an EGF TKI as first-line therapy, 43 patients (58.9%) received an EGFR TKI as second-line therapy, and 10 patients (13.7%) received an EGFR TKI as third-line or later therapy. Fifty-three patients received erlotinib, and 20 patients received gefitinib (Table 1).
The median number of cycle of EGFR-TKI administration was 2 (range, 1-50 cycles). Seventeen patients (23.3%) received >6 cycles. The clinical response to EGFR TKIs was reassessed for all 73 patients as follows: There were no complete responses, 16 patients (21.9%) had a partial response, 12 patients (16.5%) had SD, and 45 patients (61.6%) had progressive disease. The disease control rate was 38.4% (Table 1). Two patients discontinued therapy because of interstitial pneumonia and general weakness. At a median follow-up of 25.4 months (range, 10.5-53.3 months), the median PFS was 1.3 months (95% CI, 0.5-2.2 months), and the median OS was 11.5 months (95% CI, 8.2-14.8 months) for all 73 patients, and 45 patients (61.6%) had died.
Assessment of Amphiregulin Expression
Forty of 73 patients were available for an immunohistochemical study of AR expression. In these patients, AR expression was localized predominantly to the cytoplasm. All normal pneumocytes were unstained by AR, but bronchial epithelial cells and a small portion of inflammatory cells were stained partially. We evaluated the intensity according to a bronchial epithelial cell staining score of 2+. Representative results are illustrated in Figure 1. No 4+ staining specimens were identified. “AR positivity” was observed in 24 of 40 patients (60%). There were no significant differences in clinicopathologic parameters, such as sex, age, performance status, smoking status, histologic type, EGFR-TKI type, or the rate of disease control, between AR-positive patients and AR-negative patients (Table 2).
Association Between Amphiregulin Expression and Clinical Outcomes
Twelve of 24 patients with AR-positive (50%) achieved disease control (complete responses, partial responses, and SD), whereas only 4 of 16 patients (25%) with AR-negative tumors achieved disease control. It is noteworthy that 8 of 24 patients (33.3%) with AR-positive tumors had SD, whereas none of the 16 patients with AR-negative tumors had SD. However, this difference did not reach statistically difference (P = .188) (Table 2). We analyzed survival according to the different H scores (100 and 200). When we used a cutoff H score of 200, PFS (P = .899) and OS (P = .311) did not differ significantly. However, on univariate analysis using a cutoff H score of 100, PFS was 8.1 weeks for patients with AR-positive tumors and 4 weeks for patients with AR-negative tumors, and the difference was significant (P = .015) (Fig. 2). Other factors that affected the prolongation of PFS were good performance status (P = .019), never smoker (P = .003), adenocarcinoma (P = .013), and the receipt of gefitinib (P = .048) (Table 3). However, in a Cox proportional hazard model, we demonstrated showed that only being a never smoker (hazard ratio [HR], 0.303; 95% CI, 0.138-0.670; P = .003) was an independent factor for the prolongation of PFS (Table 4).
Table 3. Univariate Analysis of Prognostic Factors for Survival
In addition, OS was significantly longer in the AR-positive group than in the AR-negative group (12.2 months vs 4.1 months; P = .001) (Fig. 2). Other factors that affected the prolongation of OS were being a women (P = .010), having a good performance status (P = .000), being a never-smoker (P = .000), having adenocarcinoma (P = .000), and the receipt of gefitinib (P = .014) (Table 3). A Cox proportional hazards model indicated that being in the AR-positive group (HR, 0.345; 95% CI, 0.150-0.793; P = .012), having adenocarcinoma (HR, 0.173; 95% CI, 0.066-0.455; P = .000), and having a good performance status (HR, 0.223; 95% CI, 0.065-0.761; P = .017) were independent prognostic factors for the prolongation of survival (Table 4).
In the current study, the response and disease control rates with EGFR-TKI treatment were 21.6% and 37.8%, respectively, in patients with WT EGFR-positive NSCLC. This result is consistent with our previous study, which indicated a 16.2% rate of response to erlotinib (11 of 68 patients) among carriers of WT EGFR.12 More recently, a randomized phase 3 study of first-line, single-agent gefitinib versus gemcitabine plus cisplatin in never-smokers with adenocarcinoma of the lung (the First-SIGNAL trial) reported a 25.9% response rate (7 of 27 patients) and a 40.7% disease control rate (11 of 27 patients) with gefitinib in patients with WT EGFR.22 Conversely, a large, randomized, phase 3 trial of gefitinib versus carboplatin-paclitaxel in selected patients with NSCLC in an Asian population (the I-PASS trial) demonstrated that, in a molecular biomarker analysis, the response rate for patients with WT EGFR was only 1.1%.11 These discrepant results may be explained by differences in the EGFR mutation detection method. Direct DNA sequencing was used in our study and other studies, but a more sensitive amplification-refractory mutation system (ARMS) in was used in the I-PASS trial. Because the direct DNA sequencing method still considered the gold standard for detecting EGFR mutations, further validation and standardization of EGFR mutation tests are be needed.
We observed that AR expression was positive in 24 of 40 patients (60%) with WT EGFR-positive NSCLC. In addition, there was a trend toward a higher disease control rate in patients who had AR-positive tumors (50%; 12 of 24 patients) after EGFR-TKI therapy compared with patients who had AR-negative tumors (25%; 4 of 16 patients). This finding indicates that AR expression may be considered as another predictive marker for achieving disease control with EGFR TKIs in patients with WT EGFR NSCLC. However, because of the retrospective nature of our analysis, the results should be interpreted with caution and need to be validated in prospective studies with large numbers of patients.
It is noteworthy that positive AR status was associated with the prolongation of PFS. The PFS of AR-positive patients was 8.1 weeks compared with 4.0 weeks for AR-negative patients. This favorable outcome may be attributed plausibly to a higher proportion of SD in AR-positive patients. Eight of 24 patients (33.3%) who had AR-positive tumors achieved SD compared with no SD among the 16 patients who had AR-negative tumors. Moreover, positive AR tumor expression was an independent factor for OS (P = .001). This is consistent with the observation that patients with high AR tumor expression have more SD compared with patients who have low AR tumor expression,18 It has been suggested, that in WT EGFR NSCLC cell lines with high AR expression, EGFR TKIs cause the inhibition of EGFR downstream signaling not to PI3K/Akt but to ERK1/2, which participates in growth inhibition and G1-S arrest.18, 23 This may explain the mechanism of inducing SD with EGFR TKIs in patients with WT EGFR-positive tumors. Therefore, AR-positive tumors may benefit substantially from EGFR TKIs in the treatment of WT EGFR-positive NSCLC.
It has been reported that AR overexpression is a poor prognostic factor in NSCLC.20, 24 In breast cancer, although overexpression of Her2 itself is a poor prognostic factor, Her2-positive cancer produced better survival than Her2-negative cancer when patients received trastuzumab.25 Regarding WT EGFR tumors, EGFR signaling by EGF ligands decreases over time, unlike in tumors that have an activating EGFR mutation.26 Therefore, the EGFR downstream signal will be sustained under the condition of overexpressed EGFR ligands like AR. In those patients, blocking the pathway with EGFR-TKIs can be useful in controlling WT EGFR tumors. Consequently, ligand expression may be the driving force for the use of EGFR-TKI use in patients with WT EGFR-positive NSCLC.
However, whether increased AR expression can be used to predict responsiveness to EGFR inhibitors in different types of cancer remains controversial. In patients with squamous cell carcinoma of the head and neck, AR secretion was correlated positively with sensitivity to gefitinib.23 In patients with colorectal cancer, increased AR mRNA expression predicted the efficacy of cetuximab,16, 27 and the TNF-α-converting enzyme, which is required for the cell surface proteolytic processing of AR, has been correlated with the efficacy of EGFR inhibitors in colorectal cancer.28 In patients with NSCLC, the detection of AR protein by immunohistochemistry has been associated with sensitivity to gefitinib.18 In contrast, increased levels of AR in serum were correlated with a poor response to gefitinib in patients with advanced NSCLC.20, 29 Other investigators have reported that AR overexpression promotes resistance rather than sensitivity to gefitinib in patients with NSCLC through inhibition of the Bcl2-associated X protein BAX.30 These discrepancies may be explained by the heterogeneous methods that were used to detect AR expression, different cancer types, and differences in AR distribution between local tumors and systemic circulation. Thus, currently, the reasons for this predominance of AR are not clear and need further study.
This study had several limitations. First, although we demonstrated that clinical outcomes, such as the disease control rate, PFS, and OS, were better in patients who had AR-positive tumors compared with patients who had AR-negative tumors among those with WT EGFR-positive NSCLC, the statistical power was too weak to generate any conclusion because of the small number of patients. Second, because our heterogeneous patient cohort was analyzed retrospectively, caution should be exercised in drawing any firm conclusions from the current results. Thus, further prospective studies with more patients will be needed to validate the role of AR expression in patients with WT EGFR-positive NSCLC. In conclusion, the current study indicated that patients with WT EGFR-positive NSCLC may obtain a clinical benefit from EGFR-TKI treatment because of their AR expression status. These results suggest that AR expression may be another potential marker for the selection of patients with WT EGFR-positive NSCLC for EGFR-TKI treatment.