Conflict of Interest: Nobuoki Kohno has a personal royalty of KL-6 from a Japanese pharmaceutical company, Eisai Co., LTD. The remaining authors have no conflicts of interest.
By immunizing mice with a lung adenocarcinoma cell line, we previously established a murine IgG1 monoclonal antibody that recognizes a sialylated sugar chain designated Krebs von den Lungen-6 (KL-6). KL-6 is a high–molecular-weight glycoprotein classified as a human MUC1 mucin. The aim of this study was to determine whether KL-6 expression in tumors correlates with circulating KL-6 levels and whether circulating KL-6 has any prognostic value in patients with surgically resected non-small cell lung cancer (NSCLC). Immunohistochemical analysis of KL-6 expression was performed on 103 NSCLC tissues, and its associations with serum KL-6 levels and survival were examined. We also evaluated whether KL-6 expression patterns and/or serum KL-6 levels could predict prognosis in these NSCLC patients. Immunohistochemical analysis of KL-6 in NSCLC tissues showed that a depolarized KL-6 expression pattern was associated with a high level of circulating KL-6 and a poor prognosis in NSCLC patients who underwent curative surgery. Furthermore, a high circulating KL-6 level was associated with both poorer progression-free survival (PFS) and overall survival (OS), and multivariate analyses confirmed its independent prognostic value for both PFS and OS (p = 0.041 and 0.023, respectively). Our data suggest that preoperative serum KL-6 level reflects KL-6 expression patterns in NSCLC tissue, and can serve as a useful prognostic biomarker in NSCLC patients who undergo curative surgery.
Lung cancer is one of the most common malignant tumors in the world, and non-small cell lung cancer (NSCLC) accounts for nearly 80% of reported cases.1 The Tumor-Node-Metastasis (TNM) staging system for NSCLC is widely used for selecting candidates for surgical intervention and adjuvant chemotherapy, and is also useful for predicting prognosis. Surgical resection is the first choice for treating NSCLC patients when the disease is categorized as stage IA to IIIA. However, even after complete resection, ∼30 and 75% of patients with pathological stage IA and stage IIIA, respectively, die within 5 years.2, 3 The presence of occult metastases in these patients at the time of surgery has been proposed, which is supported by a small but significant improvement in the survival rates of NSCLC patients who received chemotherapy following surgical tumor resection.4–6 An increasing number of studies have reported that evaluation of the malignant potential of cancer cells in addition to TNM staging is useful for more precisely estimating prognosis in NSCLC patients who have undergone surgery. If this evaluation can be conducted using molecules or substances present in the circulation, tremendous clinical benefit can potentially be provided.
By immunizing a mouse with a lung adenocarcinoma cell line, we previously established a murine IgG1 monoclonal antibody (mAb) that recognizes a sialylated sugar chain designated Krebs von den Lungen-6 (KL-6).7 KL-6 is classified as a MUC1 mucin, and is known to be expressed in regenerating type II pneumocytes.8–10 As elevated serum KL-6 levels suggest the presence of interstitial pneumonia,11–15 KL-6 is used in Japan as a serum biomarker for this disease. However, recent studies have implied that KL-6 can also serve as a tumor marker.16, 17 The involvement of KL-6 in tumor progression was first proposed in a study that showed that the presence of serum anti–KL-6 antibody in NSCLC patients is correlated with prognosis.18 This observation is supported by the results of our previous study, which demonstrated an antitumor effect of anti–KL-6 mAb on cancer cell lines through the facilitation of E-cadherin–mediated cell-cell interaction, induced by the capping of MUC1 by the anti–KL-6 mAb.19, 20 Furthermore, an immunohistochemical analysis showed that expression of KL-6 is increased in breast, lung, pancreatic, ovarian, colon, and hepatocellular carcinoma tissues,7, 21–28 and that the KL-6 expression levels in colon and pancreatic carcinomas are inversely correlated with patient survival.25–28 These observations also support the involvement of KL-6 in tumor progression, and highlight its potential value as a tumor biomarker. Unlike MUC1, however, no detailed analyses of KL-6, such as determination of subcellular localization in tumor cells, have been conducted. In normal epithelium, MUC1 is present predominantly on the apical surface of epithelial cells.29 Recent observations demonstrated that in various carcinomas, MUC1 is expressed on the circumferential and basal membranes, in the cytoplasm, and on the apical surface of carcinoma cells.30 Furthermore, aberrant localization of MUC1 is associated with a poorer prognosis in NSCLC patients.31 These observations suggest that subcellular localization of MUC1 in carcinoma cells is also associated with tumor progression.
Based on these observations, we hypothesized that subcellular localization of KL-6 might be associated with malignant potential of NSCLC and/or elevated circulating KL-6 levels. To test this hypothesis, we analyzed the relationships between the subcellular localization of KL-6 in tumors, KL-6 serum levels, and the various clinical features of patients with surgically resected NSCLC. Furthermore, compared to other forms of MUC1, KL-6 has a high probability of being detected in the circulation; therefore, to assess the clinical significance of circulating KL-6 levels in NSCLC patients, the correlation between serum KL-6 levels and prognosis were evaluated in the same study group.
Material and Methods
The study group consisted of 103 patients with NSCLC (38 females and 65 males) who underwent curative surgery at Hiroshima University Hospital (Hiroshima, Japan) between 2000 and 2008. The inclusion criteria were almost identical to those described in our previous studies.32–36 In brief, inclusion criteria were: (i) age ≥20 years, (ii) no significant abnormalities in liver or kidney function, and (iii) absence of active interstitial lung disease. The final diagnosis of NSCLC was made histologically on surgically excised tissues using WHO criteria.37 The postsurgical pathologic TNM stage was determined according to the guidelines of the American Joint Committee on Cancer.38 All of the patients underwent curative surgery without preoperative chemotheory or radiotherapy. For postoperative adjuvant chemotherapy, orally administered UFT (tegafur and uracil) was selected for patients with pathologic stage IB adenocarcinoma, and platinum-based chemotherapy was selected for patients with histologically confirmed lymph node metastasis. The ages of the patients ranged from 43 to 91 years (median age, 67 ± 10.3 years). The histological types of the 103 NSCLC patients were as follows: 78 adenocarcinomas (ADC), 16 squamous cell carcinomas (SCC), 1 large cell carcinoma (LCC), and 8 adenosquamous cell carcinomas (ASC). Local recurrence was defined as recurrence within the bronchial stump, staple line, ipsilateral hilar, or mediastinal lymph nodes as previously described.39 Distant recurrence was defined as the appearance of tumors in the ipsilateral lung, contralateral lung, or other organs such as adrenal glands, brain, and bone. In addition, 68 patients with interstitial lung diseases (ILDs) and 102 healthy volunteers (61 smokers and 41 nonsmokers) were enrolled as controls in this study. Serum samples were obtained with informed consent from all 103 NSCLC patients just before surgery, the patients with ILDs, and the healthy volunteers at enrollment and were stored at −80°C. This study and the use of all clinical materials mentioned were approved by the individual institutional Ethical Committees.
Purification of anti–KL-6 monoclonal antibody
The anti–KL-6 mouse IgG1 monoclonal antibody (mAb) was purified from the ascites collected from mice bearing anti–KL-6 mAb-producing hybridomas as previously described,7 using a protein A affinity column (Affi Gel Protein A MAPS II Kit; Bio-Rad, Hercules, CA).
Immunohistochemical analysis of KL-6 expression was performed on tissue sections prepared from paraffin blocks of surgically resected lung tissue as described previously.32–36 The slides were immersed in Target Retrieval Solution, Citrate pH 6 (Dako Japan, Tokyo, Japan) and boiled at 108°C for 15 min in an autoclave for antigen retrieval. After the blocking of endogenous peroxidase activity with 0.03% H2O2 for 30 min, a mouse anti-human KL-6 mAb was added to the sections. Sections were incubated with a secondary antibody, HRP-labeled anti-mouse IgG, followed by the addition of a substrate-chromogen, and was then counterstained with hematoxylin. Three independent investigators assessed the staining patterns of KL-6 without prior knowledge of the clinicopathological data. The intensity and pattern of staining were evaluated in 10 microscopic fields that were randomly chosen from the area of tumor, or in the entire area if the tumor tissue was comprised of less than 10 fields. The staining intensity of KL-6 was first classified as negative or positive. If positive KL-6 staining was observed in the tumor tissue, the staining pattern was subsequently classified as apical membrane, circumferential membrane, or cytoplasm, as described previously.28, 30, 31
Measurement of serum KL-6 levels
Serum KL-6 levels were measured by a sandwich-type electrochemiluminescence immunoassay (ECLIA) using a Picolumi 8220 Analyzer (Sanko Junyaku, Tokyo, Japan), as previously described.40 The serum sample was incubated with anti–KL-6 antibody-coated magnetic beads prior to separation by a magnetic rack. The ruthenium-labeled anti–KL-6 antibody was then added to the beads as a second antibody following a wash with PBS. The reaction mixture was placed into the electrode, and the photons emitted from the ruthenium were measured with a photomultiplier.
Measurement of the serum carcinoembryonic antigen (CEA) and cytokeratin 19 fragments (CYFRA 21-1) levels
The serum levels of CEA and CYFRA 21-1 were measured using commercially available electrochemiluminescence immunoassay systems for CEA (Abbott Diagnostics, Tokyo, Japan) and for CYFRA21-1 (Roche Diagnostics Corp Indianapolis, IN).
All statistical analyses were performed using a statistical software package (SPSS for Windows, version 12.0; SPSS Inc; Chicago, IL). Differences between groups were analyzed using the Mann-Whitney U-test or Fisher's exact test. To test differences among the variables evaluated prior to and after surgical resection, the Wilcoxon test was used. Progression-free survival (PFS) was defined as the interval starting from the date of surgery to the date of documented disease progression, death from any cause, or last follow-up. Overall survival (OS) was defined as the interval starting from the date of surgery to the date of death from any cause, or last follow-up. Serum KL-6 levels were further analyzed for their ability to predict prognosis by receiver operating characteristic (ROC) curves and Kaplan-Meier curves. The upper left corner coordinate point of the ROC curve was determined as an optimal cut-off value to discriminate survivors from non-survivors. Survival curves and 95% confidence intervals (CIs) were analyzed by the Kaplan-Meier method, and differences between groups were compared with the log-rank test. The risk factors associated with patient prognosis were evaluated using the Cox proportional hazards regression model with a step-down procedure. Only variables that were statistically significant in the univariate analysis were evaluated by multivariate analysis. The criterion for removing a variable was the likelihood ratio statistic, which was based on the maximum partial likelihood estimate (default p-value of 0.05 for removal from the model).
Immunohistochemical staining pattern of KL-6 in tumor tissues and association between clinical outcomes and prognosis in NSCLC patients
To determine whether KL-6 is expressed in NSCLC tissues and whether KL-6 expression is associated with clinical characteristics in NSCLC patients, an immunohistochemical analysis of KL-6 expression was performed on 103 surgically resected NSCLC tissues. Interestingly, KL-6 expression was observed at the apical membrane (Fig. 1a), circumferential membrane (Fig. 1b), and/or cytoplasm (Fig. 1c) in all 103 NSCLC tissues. Regarding the staining intensity, we could not visually detect apparent differences between the sites of KL-6 expression or the types of histology (data not shown). As shown in Table 1, KL-6 expression patterns were divided into two categories: polarized expression was defined as apical staining only, while depolarized expression was defined as either circumferential membrane or cytoplasm staining. The category of KL-6 expression pattern defined by three independent investigators matched in 97 of 103 (94.1%) studied cases. The unmatched or equivocal cases were re-examined and a consensus category was decided by the three investigators. In total, 41 (39.8%) and 62 (60.2%) NSCLC tissues were categorized as having polarized expression and depolarized expression, respectively. The depolarized KL-6 expression pattern was significantly associated with gender (higher in males; p = 0.021 by Fisher's exact test), histological type (difference among ADC, SCC, and others; p < 0.001 by Chi-square test), T factor (higher in T2+3; p = 0.009 by Fisher's exact test), N factor (higher in N1+2; p = 0.002 by Fisher's exact test) and recurrence pattern (higher in distant metastasis; p = 0.015 by Fisher's exact test; Table 2). As shown in Figure 2a and b, PFS and OS of NSCLC patients with a depolarized KL-6 expression pattern were significantly poorer than those of patients with polarized KL-6 expression (p < 0.001 and p = 0.005, respectively, by log-rank test).
Table 1. Subcellular KL-6 expression pattern in NSCLC tissues
Table 2. Association between KL-6 expression pattern and clinical characteristics in the 103 patients with surgically resected NSCLC
Association between serum KL-6 level and tumor tissue KL-6 staining pattern
To determine whether serum KL-6 levels were elevated in NSCLC patients, serum KL-6 levels were compared among the NSCLC patients, the ILD patients, the nonsmoker controls, and the smoker controls. The serum KL-6 levels (mean ± SD) were significantly higher in the NSCLC patients (325 ± 189 U/mL) than those in the nonsmoker (236 ± 86 U/mL) and smoker controls (232 ± 79 U/mL; p = 0.043 and p = 0.011, respectively), whereas the serum levels of KL-6 were significantly higher in the ILD patients (1600 ± 1067 U/mL) than those in the NSCLC patients, the nonsmoker controls, and the smoker controls (p < 0.001, p < 0.001, and p < 0.001, respectively). However, there was no significant difference between the nonsmokers and smokers.
The optimal cut-off serum levels for KL-6, CEA, and CYFRA21-1 to discriminate survivors from non-survivors were determined as 400 U/mL, 3.5 ng/mL, and 2.0 ng/mL, respectively, using the upper left corner coordinate point of each ROC curve drawn for KL-6, CEA, and CYFRA21-1. Based on this value, patients were divided into a high KL-6 group (serum KL-6 level ≥400 U/mL) and a normal KL-6 group (serum KL-6 level <400 U/mL). High serum KL-6 levels were observed in 23 (22.3%) of the total 103 patients and in 19 (22.1%) of the 86 patients with node-negative NSCLC. Among clinicopathological variables, histological type (difference among ADC, SCC, and others; p = 0.036 by Chi-square test) and recurrence pattern (higher in distant metastasis; p = 0.010 by Fisher's exact test) were significantly associated with serum KL-6 level (Table 3).
Table 3. Association between serum levels of KL-6 expression and clinical characteristics in the 103 patients with surgically resected NSCLC
To confirm that circulating KL-6 was derived from tumors, the preoperative and postoperative (2 months after surgery) serum KL-6 levels were compared in 25 patients whose serum samples were available at both time points. As shown in Figure 3a, circulating KL-6 levels significantly decreased after surgical tumor resection (p < 0.001 by Wilcoxon test). In addition, a statistical analysis revealed that the depolarized KL-6 expression pattern in tumors was associated with a higher level of circulating KL-6 (p = 0.003 by the Mann-Whitney U test; Fig. 3b).
Association between serum KL-6 level and prognosis in NSCLC patients
The PFS and OS of NSCLC patients with high serum KL-6 levels were also significantly poorer than those of patients with normal serum KL-6 levels (p = 0.003 and p < 0.001, respectively, by log-rank test) (Fig. 3c and d). Therefore, to determine the prognostic importance of clinical characteristics and serum KL-6 levels in patients with surgically resected NSCLC, we performed a Cox proportional hazards regression analysis on the parameters listed in Tables 4 and 5. Univariate analyses revealed that pT stage (odds ratio, 2.338; 95% CI, 1.147−4.765; p = 0.019), pN stage (odds ratio, 2.575; 95% CI, 1.193−5.556; p = 0.016), serum KL-6 level (odds ratio, 2.818; 95% CI, 1.383−5.740; p = 0.004), serum CEA level (odds ratio, 2.506; 95% CI, 1.129−5.561; p = 0.024), and serum CYFRA21-1 level (odds ratio, 2.165; 95% CI, 1.068−4.387; p = 0.032) were significant prognostic factors for PFS (Table 4); and pN stage (odds ratio, 3.768; 95% CI, 1.360−10.440; p = 0.011), serum KL-6 level (odds ratio, 4.789; 95% CI, 1.790−12.814; p = 0.002) and serum CYFRA21-1 levels (odds ratio, 3.778; 95% CI, 1.2011−11.885; p = 0.023) were significant prognostic factors for OS (Table 5). Multivariate analyses demonstrated serum KL-6 level (odds ratio, 2.192; 95% CI, 1.031−4.662; p = 0.041) and pN stage (odds ratio, 2.264; 95%CI, 1.007−5.091; p = 0.048) to be an independent prognostic factor for PFS (Table 4); and, serum KL-6 level (odds ratio, 3.378; 95% CI, 0.774−9.652; p = 0.023) and pN stage (odds ratio, 2.987; 95% CI, 1.024−8.712; p = 0.045) as independent prognostic factors for OS (Table 5).
Table 4. Cox proportional hazards model analysis of progression-free survival in patients with surgically resected NSCLC
Table 5. Cox proportional hazards model analysis of overall survival in patients with surgically resected NSCLC
This study supports the hypothesis that KL-6 can serve as a prognostic biomarker for NSCLC. Immunohistochemical analysis revealed that KL-6 was expressed in all of the NSCLC tissues used in the present study. In addition, a depolarized KL-6 expression pattern in NSCLC tissue was shown to be associated with high circulating KL-6 levels and a poor prognosis in NSCLC patients who underwent curative surgery. Furthermore, a Cox proportional hazards regression analysis demonstrated that serum KL-6 level was an independent prognostic factor for both PFS and OS in patients with surgically resected NSCLC.
First, whether or not the circulating KL-6 observed in NSCLC patients is derived from tumors needs to be addressed. In ILD, the primary cellular source of KL-6 has been shown to be type II pneumocytes.11 KL-6 is immunohistochemically detectable in the epithelial cells of pancreatic and mammary ducts.8 Although our previous studies have already shown that KL-6 levels are elevated in sera from NSCLC patients,7 the cellular origin of circulating KL-6 in NSCLC patients has not been clearly demonstrated. In this study, serum KL-6 levels were found to decrease dramatically after surgical resection of primary tumors. In addition, KL-6 was immunohistochemically detectable in all of the analyzed NSCLC tissues. These results implied that the primary tumor was the origin of circulating KL-6 in NSCLC patients. Furthermore, immunohistochemical analysis of KL-6 revealed two KL-6 staining patterns in NSCLC tissues distinguished by subcellular localization: polarized and depolarized KL-6 expression patterns. In general, MUC1 mucin is present on the apical surface of the normal secretory epithelia. In malignant tissues, however, this apical polarization is frequently lost, resulting in the localization of MUC1 throughout the cell membrane and in the cytoplasm.28–31 Because MUC1 mediates anti-adhesive activity by interfering with cell-to-cell and/or cell-to-extracellular matrix interactions, aberrant subcellular expression of MUC1 facilitate detachment of cancer cells from the primary tumor.41 We believe that this scenario can be applied to KL-6 as well, thereby accounting for the poorer survival observed in patients with NSCLC tumors displaying depolarized KL-6 expression. In addition, we also demonstrated that depolarized expression pattern of KL-6 in the resected tumor was associated with high preoperative serum level of KL-6. As far as we are aware, this is the first study that reports the association between expression pattern of a MUC1-related molecule in tumor and its circulating level.
A novel finding from this study is that preoperative circulating KL-6 level can serve as a prognostic biomarker in NSCLC patients who undergo curative surgery. The optimal cut-off level that could discriminate survivors from non-survivors was established by using ROC curves, and patients with serum KL-6 levels below 400 U/mL were shown to have significantly favorable PFS and OS compared to patients with serum KL-6 levels above 400 U/mL. Furthermore, the Cox proportional hazards regression analysis demonstrated serum KL-6 level to be an independent prognostic factor for both PFS and OS in NSCLC patients who underwent curative surgery. On the other hands, neither CEA nor CYFRA21-1 was a significant independent prognostic factor for PFS and OS, suggesting that KL-6 is a more sensitive serum biomarker than CEA and CYFRA21-1 for predicting clinical outcome of NSCLC patients who undergo curative operation. Regarding the association between serum KL-6 level and prognosis in NSCLC patients, we previously reported that a pretreatment serum level of KL-6 serves as an independent prognostic factor in advanced or refractory NSCLC patients treated with orally active epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (EGFR-TKIs).17 Although 6 of 103 patients were treated with EGFR-TKIs following postoperative recurrence, none of the patients received EGFR-TKI therapy as postoperative adjuvant chemotherapy. Since survival of NSCLC patients following curative surgery primarily depends on the presence or absence of occult metastases, NSCLC patients with depolarized KL-6 expression in tumors and high levels of circulating KL-6 are therefore more likely to have occult metastases at the time of surgery. In fact, the depolarized KL-6 expression and high serum KL-6 levels were associated with distant recurrence after the surgery in the present study. These results also suggest that KL-6 expression pattern in tumors and serum KL-6 level may be able to be used for selecting patients who should receive adjuvant chemotherapy.
Compared to detecting biomarkers in tumors by immunohistochemical analysis or polymerase chain reaction (PCR), measurement of circulating biomarkers can be more rapid, reproducible, and inexpensive. Several lines of evidence suggest that overexpression of MUC1 in tumors is associated with shorter survival in patients with surgically resected NSCLC.31, 42, 43 However, no circulating MUC1-related biomarkers other than KL-6 have been reported. Therefore, KL-6 can be regarded as a novel MUC1-associated serum biomarker for NSCLC. The sensitivity of KL-6 as an indicator of NSCLC was poorer compared to conventional serum NSCLC biomarkers, such as CEA and cytokeratin 19 fragments (CYFRA21-1; data not shown). This suggests that circulating KL-6 is not valuable as a diagnostic biomarker. However, a high preoperative KL-6 serum level was demonstrated to be an independent prognostic factor for OS and PFS in NSCLC patients. These results strongly suggest that KL-6 is a more useful prognostic factor for NSCLC patients who undergo curative surgery.
Although promising results were obtained, we are aware that this study has some limitations. First, the number of patients was not sufficient to perform a valid statistical analysis. Second, the association between KL-6 expression and particular molecular features of NSCLC such EGFR mutation was not evaluated. To investigate the significance of KL-6 expression and its association with various clinical variables including EGFR mutation status, we believe that a prospective study in a larger number of patients with surgically resected NSCLC is required.
In conclusion, the results of the present study indicate that both tumor KL-6 expression pattern and circulating KL-6 level are useful for predicting survival of NSCLC patients who have undergone curative surgery.
We thank Kazunori Fujitaka, Junya Inata, Shinichiro Ohshimo, Koji Yoshioka, Shigeo Kawase, Shin Akita and Yasushi Horimasu for their support.