INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

MUC1 is a heavily glycosylated large glycoprotein found on the apical surface of a variety of glandular epithelial cells (1). Its expression is increased in breast, lung, ovary, and colon carcinomas (2), suggesting that this aberration of MUC1 expression is a common property of adenocarcinomas. We previously identified a mouse IgG1 monoclonal antibody (mAb) directed against a lung adenocarcinoma cell line, which recognizes a sialylated sugar chain, designated KL-6 (6). KL-6 is classified as "Cluster 9 (MUC1)" of lung tumor and differentiation antigens, according to the findings of immunohistochemical and flow cytometry studies (Third International Workshop on Lung Tumor and Differentiation Antigens, 1993, Zurich, Switzerland) (7). The molecule consists of multiple heterogenous submolecules (8).

Recently, we found that anti-KL-6/MUC1 antibody (Ab) is produced in the patients with non-small cell lung cancer (NSCLC). Several studies have reported that humoral anti-MUC1 reactions were detected in a number of patients with ovarian and breast carcinomas (11, 12). Abundant evidence is available for the existence of natural autoantibodies in normal serum. Reactivities of these natural autoantibodies have been extensively studied using a panel of self- and nonself-antigens (13, 14). These antigens include cytoskeletal proteins, surface proteins, DNA, thyroglobulins, bacterial antigens such as lipopolysaccharide, and various chemical haptens, but their physiologic functions remain unknown (15). Thus, in the present study we examined patients with NSCLC and normal individuals for the presence of anti-KL-6/MUC1 Ab and assessed the association between the titer of this Ab and clinical characteristics, response to treatment, and survival.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Clinical Characteristics of the Patients

Serum was obtained from 60 healthy individuals (50 male, 10 female; mean age, 60 yr; range, 35 to 72 yr; current smoker 32, ex-smoker 15, nonsmoker 13). Sera from healthy individuals were provided from two health care centers (Saijoh Central Hospital and Health Care Center, Daio Paper Producing Corp.). The healthy individuals had no abnormality in complete blood cell counts, C-reactive protein, erythrocyte sedimentation rate, liver function tests, renal function tests, urinalysis, fecal examination, chest X-ray, and electrocardiogram. Serum was also obtained from 30 patients (25 male, 5 female; mean age, 64 yr; range, 37 to 77 yr) consecutively admitted to the Second Department of Internal Medicine, Ehime University Hospital between January 1990 and June 1995. Patients were selected based on the following characteristics: (1) newly diagnosed, previously untreated, (2) cytologically or histologically proven NSCLC and classified in Stage IIIB or IV by whole body examinations, (3) and treated with chemotherapy. They consisted of 14 patients with adenocarcinoma, 10 patients with squamous cell carcinoma, and six patients with large cell carcinoma. Their clinical records and histopathological diagnoses were fully documented. The sera of all the patients were obtained at the time of diagnosis and stored at 80° C until use.

Determination of disease stage of all patients was supported by a computed tomography (CT) scan of the chest, CT scan or ultrasound examination of the abdomen, bone scintigraphy, and CT scan or magnetic resonance imaging of the head.

All of the patients received standard chemotherapy (100 mg per square meter platinum on Day 1, 3 mg per square meter vindesine on Days 1 and 8, and 8 mg per square meter mitomycin on Day 1). Standardized response criteria for chemotherapy, based on objective measurements of tumor size, included partial response (PR) (50% or greater reduction in the sum of the products of the greater and lesser diameters of all measured lesions lasting at least 1 mo, and an absence of any new lesions during treatment), progressive disease (PD) (25% or greater increase in the sum of the products of the greater and lesser diameters of all measured lesions or the development of new lesions), and no change (NC) (the sum of the products of the above tumor measurements neither increases by more than 25% nor decreases by more than 50%). Thirty-three percent of all patients achieved PR, 37% exhibited NC, and 30% exhibited PD. A follow-up rate of 100% was achieved as of June 1997, and 7% of the patients were alive at the end of the study; the median follow-up time for all patients alive at the last follow-up was 54.0 mo (24.1 to 81.5 mo). For all patients, survival was calculated from the start date of therapy until the date of death or last clinical examination.

Quantification of Circulating KL-6/MUC1 Antigen and Anti-KL-6/MUC1 Antibody

The concentration of KL-6/MUC1 antigen in serum was measured using a sandwich enzyme-linked immunosorbent assay (ELISA) using mouse anti-human KL-6/MUC1 mAb (IgG1) and horseradish peroxidase (HRP)-labeled KL-6 Ab, as described previously (6). Anti-KL-6/MUC1 Ab was measured by a heterogeneous ELISA using purified KL-6/MUC1 (16) as a catcher molecule and goat anti-human IgG polyclonal Ab (Chemicon, Temecula, CA) as a tracer antibody. In brief, each well of a 96-well microtest plate (Costar, Cambridge, MA) was sensitized with 100 µl of 100 unit per milliliter of purified KL-6/ MUC1 in Dulbecco's phosphate-buffered saline (DPBS) (10 mM phosphate-140 mM sodium chloride, pH 7.4) at room temperature for 1 h. The KL-6/MUC1 was purified from ascites obtained from a patient with adenocarcinoma of the lung by an immunoaffinity column labeled with the mouse anti-KL-6/MUC1 mAb and by a gel-filtration as described (16). The initial concentration of KL-6/MUC1 was 63,000 U/ml in the ascites. The final yield of KL-6/MUC1 was 6.2% and the concentration was 1.57 × 10⁷ U/mg protein (16). After washing with phosphate-buffered saline (PBS) (50 mM phosphate-100 mM sodium chloride, pH 6.5) containing 0.05% Tween 20 and 0.1% bovine serum albumin (BSA), the wells were incubated overnight at 4° C with dilution buffer (10% fetal calf serum-1% BSA-PBS) to block nonspecific binding sites. The wells were washed three times using DPBS containing 0.05% Tween 20 and 0.5% BSA, and then 100 µl of human serum (1:20 or 1:100 in dilution buffer) was applied per well. After 1 h incubation at room temperature, the plates were washed three times using DPBS containing 0.05% Tween 20 and 0.5% BSA. After the last washing step, 100 µl of HRP-labeled goat anti-human IgG polyclonal Ab (1:3,000 in dilution buffer) were added to each well and incubated at room temperature for 1 h. The plates were washed, and 100 µl per well of o-phenylenediamine dihydrochloride solution (0.3% o-phenylenediamine dihydrochloride-0.02% hydrogen peroxide-150 mM citrate-phosphate buffer, pH 4.9) was added and allowed to react for 30 min. The reaction was then stopped with 2 M hydrochloride and the absorbance (A₄₉₂) was measured. The experiment was performed in duplicate. For quantification of anti-KL-6/MUC1 Ab, a stocked pleural effusion from a patient with pulmonary adenocarcinoma was used as a standard reference sample. A four-point standard curve was made for each plate using this reference sample. This reference pleural effusion contained 70 units per milliliter of anti-KL-6/MUC1 Ab and the value of the samples tested was calculated within each plate in relation to the standard curve by least-squares regression analysis.

The analytical performance of the assay was assessed by evaluating intra-assay imprecision; interassay imprecision, and the linearity on dilution. The intra-assay coefficients of variation (CV) (three samples at variable anti-KL-6/MUC1 Ab concentrations tested 9 times in duplicate) of the present ELISA system ranged from 1.4 to 5%, and the interassay CV (the standard curve in 10 consecutive working days) ranged from 5.7 to 14.7%. Correlation coefficients for serially diluted serum samples ranged from 0.999 to 0.950.

Purification of Human Anti-KL-6/MUC1 Autoantibody

Human anti-KL-6/MUC1 Ab was purified from serum obtained from normal individuals. Purified KL-6/MUC1 antigen was applied to a mouse anti-human KL-6/MUC1 Ab-coupled immunoaffinity column (16). The column was washed with 10 column volumes of 100 mM Tris-hydrochloride buffer (pH 8.0) and salt fractionated serum was applied. The column was further washed using Tris buffer, and the anti-KL-6/MUC1 Ab retained on the column was eluted with 100 mM Tris-hydrochloride buffer (pH 8.0) containing 2 M magnesium chloride and dialyzed against 100 mM Tris-hydrochloride buffer (pH 8.0). The eluate was concentrated by ultrafiltration (Model Centricon plus-20; Millipore, Bedford, MA).

Identification of Anti-KL-6/MUC1 Autoantibody

To identify the anti-KL-6/MUC1 autoantibody, we performed a sodium dodecyl sulfate (SDS)-8% polyacrylamide gel electrophoresis (PAGE), visualized using silver stain (Silver Stain II kit Wako; Wako Pure Chem., Osaka, Japan). To avoid the pitfalls of ELISA including nonspecific reactivity, the reactivity of the purified antibody was compared with previously isolated antibodies, using Western immunoblotting. Purified KL-6/MUC1 antigen (1.5 mg per lane) was resolved on an SDS-4% polyacrylamide gel and transferred to a nitrocellulose membrane (Bio-Rad Laboratories, Hercules, CA) (16). The nitrocellulose strips were incubated with the purified human anti-KL-6/MUC1 Ab (5 mg/ml), mouse antihuman KL-6/MUC1 mAb (1 mg/ml) (16), mouse anti-MUC1 mAb (Ma695:IgG1) (YLEM, Roma, Italy) (17), HMFG2 mAb, which recognizes human MUC1, (IgG1: YLEM) (18), or a normal human IgG (5 mg/ml) (Zymed Laboratories, Inc., San Francisco, CA), followed by an HRP-conjugated goat anti-human IgG Ab (Harlan Sera-Lab Ltd., Sussex, UK) or an HRP-conjugated rat antimouse IgG1 Ab (Zymed Laboratories). The membrane strips were reacted with enhanced chemiluminescence detection reagent (Amersham, Buckinghamshire, UK) for 1 min and exposed to X-ray film at room temperature for 15 min.

Immunocytochemistry

The specificity of antibody reactivity with the KL-6/MUC1 antigen was assessed by immunocytochemistry. YMB-S cells (19), which express high levels of MUC1, were stained with purified human anti-KL-6/MUC1 Ab, mouse antihuman KL-6/MUC1 mAb, HMFG2 Ab, or a normal human IgG using Vectastatin Elite ABC kits (Vector Laboratories, Burlingame, CA).

Statistical Analysis

Frequencies in contingency tables were compared using Pearson's chi-squared test. Student's t test was used to assess the differences in subject age. Survival distributions were computed by the Kaplan-Meier methods and were compared by the log-rank test. Survival was measured in months from the start of chemotherapy. The Cox proportional hazards regression model was used to study the effects of different variables on survival. Seven factors (sex, albumin, carcinoembryonic antigen [CEA], lactate dehydrogenase [LDH], N stage, M stage, and anti-KL-6/MUC1 Ab status) were studied; scores were assigned to each variable for the regression analysis. All analyses were conducted using Statview J-4.11 (Abacus Concepts, Berkeley, CA) and SPSS 6.1 (Statistical Product and Service Solutions, Chicago, IL) for Macintosh computers.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Anti-KL-6/MUC1 Ab was detected in all serum samples from both patients and normal individuals using ELISA (Figure 1). Levels of anti-KL-6/MUC1 Ab in NSCLC patients (32 ± 13 U/ml) were significantly lower than those in normal individuals (108 ± 25.6 U/ml; p < 0.001). When the titers were normalized to serum albumin or IgG, the values were still significantly different (829 ± 315 versus 2,033 ± 405 U per g albumin, p < 0.0001; 1,992 ± 625 versus 7,880 ± 1,920 U per g IgG, p < 0.0001; for patients with NSCLC and normal, respectively). In addition, no significant correlation was observed between the titer of KL-6/MUC1 antigen and Ab (Y = 33.4  0.0027X, p = 0.282).

View larger version (11K): [in this window] [in a new window]   Figure 1.   Levels of anti-KL-6/MUC1 Ab in the sera of healthy control subjects and patients with NSCLC.

The existence of natural autoantibody for KL-6/MUC1 was confirmed by purification of the Ab detected in our system. SDS-PAGE of the purified anti-KL-6/MUC1 Ab samples identified two bands at estimated molecular weights of 50 kD and 23 kD. Immunostaining of purified KL-6/MUC1 antigen using the purified human anti-KL-6/MUC1 Ab, identified a major high-molecular-weight band, showing a polydisperse pattern (more than 200 kD). Mouse anti-KL-6/MUC1 mAb, anti-MUC1 Ab (Ma695), and HMFG2 Ab exhibited similar reactivity (Figure 2).

View larger version (70K): [in this window] [in a new window]   Figure 2.   SDS-PAGE, Western blotting, and immunostaining. Lane 1: molecular weight standards. Silver stains of purified anti-KL-6/ MUC1 Ab samples (lane 2). Immunostaining of purified KL-6/ MUC1 with purified anti-KL-6/MUC1 Ab (lane 3), normal human IgG (lane 4), mouse anti-KL-6/MUC1 mAb (lane 5), anti-MUC1 Ab (Ma695) (lane 6), and HMFG2 Ab (lane 7 ).

Immunocytochemistry also showed that mouse anti-KL-6/ MUC1 mAb, anti-MUC1 Ab (Ma695), and HMFG2 Ab reacted with MUC1 antigen on YMB-S cells predominantly with the cell membrane and cytoplasm, similarly to the purified human anti-KL-6/MUC1 Ab (Figure 3).

View larger version (150K): [in this window] [in a new window]   Figure 3.   Immunocytochemistry of KL-6/MUC1 expression by YMB-S cells. Mouse anti-KL-6/MUC1 Ab (A), HMFG2 Ab (B), and human anti-KL-6/MUC1 Ab (C ) were used as primary antibodies. Counterstaining was performed with hematoxylin for 1 min. (D), Negative control; the primary antibody was nonimmune human IgG.

A positive correlation was observed between the levels of anti-KL-6/MUC1 Ab and survival time of patients (Figure 4A). On the other hand, serum levels of KL-6/MUC1 antigen did not significantly correlate with survival (Figure 4B), and there were no significant differences among different histologies; the levels were 604 ± 747 U/ml in adenocarcinoma, 405 ± 339 U/ml in squamous cell carcinoma, and 270 ± 223 U/ml in large cell carcinoma. Patients were, furthermore, divided into two groups based on antibody titers. Titers above the NSCLC patient mean (32 U/ml), which corresponds to mean 3 standard deviations of normal levels, were classified as high. Those below the mean were classified as low. No significant differences in the clinical characteristics such as age, sex, histology, TNM stage, and laboratory parameters were observed. However, differences in survival and response between patients whose levels of anti-KL-6/MUC1 Ab were greater than or equal to 32 U/ml, and those less than 32 U/ml were statistically significant (Table 1). Overall survival of patients with high levels of anti-KL-6/MUC1 Ab was significantly higher than those with low levels (survival rate was 90.9% versus 21.1% at 1 yr, 72.7% versus 0.0% at 2 yr, and 36.4% versus 0.0% at 3 yr, respectively, p < 0.001) (Figure 5).

View larger version (16K): [in this window] [in a new window]   Figure 4.   The relationships between survival time and levels of anti-KL-6/MUC1 Ab (A) or KL-6/MUC1 antigen (B).

View larger version (9K): [in this window] [in a new window]   Figure 5.   Kaplan-Meier survival curves for patients with NSCLC whose concentrations of anti-KL-6/MUC1 Ab were  32 U/ml (closed square) or < 32 U/ml (closed circle). Survival time for patients with low levels was significantly lower than patients with high titers (p < 0.001).

The variables used in the Cox regression analysis are shown in Table 2. The estimated prognostic value of each variable in relation to overall survival is expressed as a p value. Three variables (anti-KL-6/MUC1 Ab status, albumin, and N stage) significantly predicted survival. Sex, LDH, CEA, and M stage were not significant predictors of survival (p > 0.1).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We found that circulating IgG Ab directed against KL-6/ MUC1 is present in all healthy individuals. Reactivity of purified Ab to KL-6/MUC1, examined by Western blot analysis (Figure 2) and immunostaining of a cell line expressing MUC1 (Figure 3), confirmed that the autoantibody was naturally occurring (15). The Ab was also present in serum from patients with NSCLC but its titer was significantly lower than in healthy individuals. Interestingly, the degree of decrease in Ab concentrations predicted patient prognosis. These results indicate that anti-MUC1 immune reactions, including anti-MUC1 IgG Ab, may contribute to the occurrence and progression of NSCLC.

Patients with breast or ovarian carcinoma had circulating Ab to MUC1 molecule (12, 20). However, according to these reports, this Ab was detected only in 8 to 17% of the patients. In contrast, we detected the Ab in all samples, including those of healthy individuals (Figure 1). Only one report (21) described that IgM Ab directed against synthetic peptides of MUC1 could be detected in all healthy women. Similarly to our results, the Ab concentration was lower in patients with ovarian cancer. However, it was not a significant independent prognostic indicator. These discrepancies may be caused by differences in MUC1 epitopes used for detection. Our Ab recognized the sialylated sugar chain KL-6, whereas previous studies targeted various synthetic peptides of the MUC1 core protein. The KL-6/MUC1 antigen was easily detected in blood from healthy volunteers and markedly increased in patients with interstitial pneumonia, rendering it a useful marker for such patients (22). Sugar chains on MUC1 may be the dominant target for MUC1 autoantibodies.

Reasons for the significant decrease in KL-6/MUC1 Ab in NSCLC patients compared with normal individuals are not clear at present. Nutritional status could not explain the difference. Levels of the Ab were not correlated with age, performance status, concentration of albumin, IgG, or different histologies. Even when normalized to albumin or IgG, significant differences still existed between healthy volunteers and patients with NSCLC. Such a decrease could be explained by a complex formation with circulating KL-6/MUC1 antigen, which is expected to increase in patients with NSCLC, as is seen in patients with breast carcinoma (11). However, there was no correlation between concentrations of anti-KL-6/MUC1 Ab and those of KL-6/MUC1 antigen, which includes both complexed and uncomplexed antigen. Although MUC1 is abundant in adenocarcinoma, there were no significant differences in the concetrations of the antibody among patients of different histologic types.

Guddo and coworkers reported that depolarized expression of MUC1 was associated with poor outcome in early stage NSCLC (23). Anti-KL-6/MUC1 Ab may react with cancer cells even in vivo as is shown in vitro in this report. Immune complex formation on the cell surface might induce capping of MUC1. As a result, cell adhesion could occur, cancer invasion could be limited (24), and unmasked cancer cells could be more susceptible to immune recognition by cytotoxic cells (25). The Ab may also facilitate Ab-dependent or complement-dependent cytolysis of cancer cells (26). Soluble MUC1 inhibits human T-cell proliferation, although this effect is reversed by interleukin-2 (27). Anti-MUC1 Ab may exhibit antitumor effects similar to cytokine. It has been reported that vaccination with MUC1 has beneficial effects on patients with breast cancer, leading to large amounts of anti-MUC1 IgG Ab, T cell proliferation, and cytotoxic lymphocyte response (28, 29).

As a prognostic factor, this natural IgG autoantibody against KL-6/MUC1 has promise in clinical use. Histologically proven antigens such as H/Le^y/Le^b (30), amphiregulin (31), and cyclin A (32) have also been reported to be of prognostic value. However, because their expression can only be examined using resected tumor tissues, the prognosis for patients with inoperable advanced NSCLC is difficult to establish. For advanced stage NSCLC patients, circulating antigens such as Le^x-i (33), interleukin-6 (34), and intercellular adhesion molecule-1 (35) can be valuable. These histologically proven antigens and circulating antigen levels reflect cancer tissue structures, and cellular proliferative, invasive, and metastatic potentials. Conversely, the decrease in circulating anti-KL-6/MUC1 natural autoantibody may indicate the deterioration of host-versus-cancer immune reactions. Studies of the pathobiological importance of this novel antibody may lead to the development of tools for controlling the progression of human cancer.