MUC1 is a transmembrane-type mucin characterized by a large and linearly extended extracellular domain, most of which consists of tandem repeat sequences.1 The in vitro inducibility of cytotoxic T cells recognizing the tandem repeat peptide first showed the in vivo immunogenicity of MUC1.2 A finding that EBV-immortalized B-cell clones from ovarian cancer patient secreted antibodies reactive with the tandem repeat peptide3 suggested that the humoral immune response against MUC1 could be induced in vivo. We first demonstrated the presence of circulating anti-MUC1 IgG antibodies in ulcerative colitis by Western blot analysis with a recombinant MUC1 protein.4 Those antibodies were also detected in sera from patients with breast, pancreatic or colorectal cancer by Kotera et al.5 and subsequently found in a variety of tumors.6, 7, 8, 9 Importantly, a recent study by von Mensdorff-Pouilly et al.6 showed that a positive test result for both IgG and IgM antibodies in pretreatment serum was associated with a significant benefit for disease-specific survival in Stage I and II breast cancer patients. Although the in vivo functions of circulating anti-MUC1 antibodies in cancer patients remain unknown, this finding suggests that their serum levels might be of use for predicting clinical outcomes. In our study, we measured circulating anti-MUC1 antibodies against a synthetic tandem repeat peptide in pancreatic cancer patients and determined whether they had any impact on survival.
MUC1 is immunogenic in vivo and humoral and cellular immune responses against MUC1 have been detected in cancer patients. Our study explored the association of circulating anti-MUC1 antibodies with clinicopathological parameters or patients' survival of pancreatic cancer. Serum specimens from 36 patients with invasive ductal carcinoma of the pancreas were subjected to enzyme immunoassay for anti-MUC1 IgG or IgM antibodies. Serum levels of anti-MUC1 IgG antibodies were significantly correlated with survival time (p = 0.0004), whereas neither those of anti-MUC1 IgM nor anti-Galα(1,3)Gal IgG antibodies, the latter known as natural antibodies cross-reactive with MUC1, showed a given tendency. Some patients' sera with the higher antibody titer showed the reactivity with MUC1-transfectants of cultured pancreatic cancer cells, but not with MUC1-negative parental cells. When the samples were tentatively divided into 2 groups by the serum level of anti-MUC1 IgG antibodies, the survival of patients was significantly longer in the group with optical density ≥0.3 than in that with optical density <0.3 (p = 0.008). Circulating anti-MUC1 IgG antibody levels remained significant (HR, 0.03; 95% CI, 0.003–0.289; p = 0.0024) after multivariate analysis for pTNM stage, patient age and gender. These data suggest that circulating anti-MUC1-IgG antibody levels may be predictive for survival of pancreatic cancer patients. © 2002 Wiley-Liss, Inc.
MATERIAL AND METHODS
Human serum samples
Serum samples were obtained from 36 patients (23 male, 13 female, mean age 64 years, range: 36–82) histopathologically diagnosed as invasive ductal cell carcinoma (IDC) of the pancreas and surgically treated at Yamaguchi University Hospital between 1990 and 1998. Only patients who survived more than 2 months after operation and whose entire course could be observed were included in the study. All of the patients tested died of cancer during the period of follow-up. Mean survival time was 10.1 months (range: 3–32). Frequency of each pathological (pTNM) stage10 of patients was 2 in Stage I, 10 in II, 8 in III and 16 in IV. Serum samples were obtained from those patients before treatment and preserved at −80°C until measurement.
Enzyme immunoassay for circulating antibodies
An enzyme immunoassay for detecting antibodies was carried out as described previously.11 Briefly, a 100-amino-acid peptide corresponding to the extracellular tandem repeat domain of MUC1 (provided by Dr. O.J. Finn, University of Pittsburgh) or Galα(1,3)Gal-HSA (human serum albumin) (Dextra, Laboratories, UK) were coated onto 96-well microtiter plates (Asahi Techno Glass Corporation, Japan) at 100 μg/ml in phosphate buffered saline (PBS) (pH 7.4) at 4°C for 12 hr. The plates were washed with PBS, and non-specific binding sites were blocked with 3% HAS/PBS at 37°C for 1 hr. The plates were then incubated with patient's sera diluted 1:40 in 1% HSA/PBS at 37°C for 1 hr. After washing with 0.05% Tween-20/PBS, they were incubated with the second antibody, a horseradish peroxidase-conjugated mouse anti-human IgG (Dako Corporation, Carpinteria, CA) diluted 1:5000 in 1% HSA/PBS, and washed with PBS. Substrate reaction using O-phenylenediamine dihydrochloride (Dako) was determined at 492 nm in an autoreader (Labsystems, Helsinki, Finland). When the immunoglobulin subclass IgM of circulating antibodies was detected, a horseradish peroxidase-conjugated anti-human IgM (Dako) was used as the second antibody. An anti-MUC1 MAb E29 (Dako) was used as a positive control. All of the serum samples (n = 36) were simultaneously measured in duplicate using one 96-well plate to compare each optical density (OD) value in each assay of anti-MUC1 IgG, anti-MUC1 IgM or anti- Galα(1,3)Gal antibodies.
Measurement of circulating MUC1 antigen levels
KL-6 antigen was measured with a commercially available kit (Picolumi KL-6, Sanko Junyaku Co. Ltd., Tokyo, Japan) by using the ECLIA (electro-chemiluminescence immunoassay) automatic equipment (Picolumi 8220, Sanko Junyaku Co. Ltd., Tokyo, Japan) according to manufacturer's instructions.
Indirect immunofluorescence assay was carried out as described previously.12 Briefly, 10 μl of patient's sera or mouse anti-MUC1 MAb (E29:DAKO) was added to 106 of stable MUC1 transfectants of human pancreatic carcinoma cell S2-013 cells13 or parental cells and incubated at 4°C for 60 min. Human IgG (DAKO) was used as a negative control. After washing with PBS, the FITC-conjugated secondary antibody (mouse anti-human IgG or rabbit anti-mouse IgG:Dako) was added and incubated at 4°C for 60 min. The cells were resuspended in PBS after washing. The expression of MUC1 on 105 of the live cells was measured using EPICS-XL Flow Cytometer System (Beckman Coulter, USA), and analyzed by the XL System II software (version 3.0).
Fisher's exact test was used to compare the categorical data between the 2 groups divided by antibody titer at an OD of 0.3. Cumulative overall survival of patients was assessed by the method of Kaplan-Meier, and comparison of the survival curves of the two groups by the log-rank test. Cox regression analysis was carried out to determine whether circulating anti-MUC1 antibody was an independent prognostic factor. Hazard ratio (HR) and 95% CI were calculated. A p-value of <0.05 was considered significant.
A significant correlation (r = 0.561, p = 0.0004) was found between the survival time and the serum levels of anti-MUC1 IgG antibodies in 36 patients with pancreatic cancer, whereas there was no such tendency in the cases of anti-MUC1 IgM antibodies and anti-Galα(1,3)Gal antibodies, the latter of which are a kind of natural antibody known to be cross-reactive with the peptide of the extracellular tandem repeat domain of MUC114 (Fig. 1).
When the samples were arbitrarily divided into 2 groups by the mean serum level of anti-MUC1 IgG antibodies (OD ≥ 0.3 vs. OD < 0.3), it was revealed that the survival of patients was significantly longer in the group with OD ≥ 0.3 than in that with OD < 0.3 (p = 0.008) (Fig. 2). No association of anti-MUC1 IgG antibody levels (OD ≥ 0.3 vs. OD < 0.3) with age (≥60 vs. <60), gender (male vs. female), and pTNM stage (I,II vs. III,IV) was found by Fisher's exact test. Circulating anti-MUC1 IgG antibody levels remained significant (HR, 0.03; 95%CI, 0.003–0.289; p = 0.0024) after multivariate analysis for pTNM stage, patient age and gender, indicating that it is an independent prognostic factor in pancreatic cancer.
To observe the relation of circulating MUC1 antigen levels with anti-MUC1 IgG antibody levels or patients' survival, KL-6 antigen was measured by ECLIA. Neither circulating anti-MUC1 IgG antibody levels nor patients' survival time was significantly associated with KL-6 antigen levels (Fig. 3).
It is important to clarify whether circulating antibodies against synthetic MUC1 peptides can react with native MUC1 molecules for exploring their roles in vivo. We therefore tested the reactivity of patients' sera with MUC1 transfectants of pancreatic cancer cell line S2-013 by flow cytometry. It was found that the patients' sera with the higher antibody titer (Samples 33 and 35) could react with S2-013 cells expressing MUC1, but not with MUC1-negative parental cells (Fig. 4), although a faint reactivity of Sample 33 with parental cells was observed. The optical density (OD) values of circulating anti-MUC1 IgG antibodies in Samples 7, 33 and 35 were 0.05, 0.56 and 0.69, respectively.
Our data suggested that circulating anti-MUC1 IgG antibodies could be a favorable prognostic factor in patients with pancreatic cancer. It is considered that the findings do not merely reflect the decreased immunoglobulin production in advanced pancreatic cancers because the serum level of anti-MUC1 IgG antibodies was neither associated with pTNM stage nor with the serum level of anti-Galα(1,3)Gal antibodies, another circulating IgG antibody. A natural antibody against Galα(1,3)Gal is known to cross-react with synthetic peptides corresponding to the tandem repeat sequence of MUC1.14 We measured their IgG subclass in the same serum samples as used for the anti-MUC1 antibodies. There was no association of the serum level of anti-Galα(1,3)Gal antibodies with that of anti-MUC1 IgG antibodies, nor with patient survival. We showed that some patients' sera had the reactivity with native MUC1 molecules expressed on MUC1-transfectants of pancreatic cancer S2-013 cells, suggesting that circulating anti-MUC1 peptide IgG antibodies could bind to pancreatic cancer cells. This might result in the association of the serum antibody levels with patients' survival. A weak reactivity of Sample 33 with parental cells (Fig. 4) was probably due to the presence of natural antibodies against S2-013 cells.
The reason why no correlation between IgG and IgM antibodies against MUC1 was detected in pancreatic cancer patients remains unknown. Both circulating IgG and IgM MUC1 antibody levels had significant prognostic value in a previous study on breast cancers in early stages.6 A possible explanation is that the serum level of IgM antibodies may be more prominently affected by immune complex formation than that of IgG antibodies as cancer progresses. Richards et al.7 demonstrated that IgG and IgM antibodies to MUC1 were also detected in multiple myeloma, but only anti-MUC1 IgM antibody levels were significantly decreased compared to healthy donors. Natural IgM and IgG antibodies to Thomsen-Friedenreich (T) antigen were detected in the sera of blood donors and of gastric cancer patients, in which only IgM antibody levels were much lower in patients of Lewis(a−b+) phenotypes than in blood donors.15
It is well known that circulating MUC1 antigen levels increase in parallel with the advance of pancreatic cancer. We measured circulating antigen KL-6, which was recognized with the monoclonal antibody against a sialylated epitope of MUC116 and shown to have the incidence of positivity more than 50% in pancreatic cancer.17 No association was seen between circulating KL-6 and anti-MUC1 IgG antibody levels. In addition, there was no association of circulating anti-MUC1 IgG antibodies with pathological stage, suggesting that individual differences in humoral immune response to synthetic MUC1 peptides may be related with patient survival. The true antigens that induce this humoral immune response are still unclear, although Galα(1,3)Gal as a cross-reactive antigen is at least one of them. It was demonstrated that natural MUC1 antibodies from breast cancer patients reacted more strongly with the peptide that N-acetylgalactosamine was attached to than with the naked peptide.18 A glycosylation-induced change in the peptide conformation may also play an important role. Our present findings suggested a prognostic significance of circulating anti-MUC1 IgG antibodies in pancreatic cancer, but further studies in a larger sample size are obviously required to confirm their relation with each clinicopathological parameter.
This work was supported by Grant-in-Aid for Scientific Research on Priority Areas (C) from the Ministry of Education, Culture, Sports, Science and Technology, Japan (12217097).