1828 Gangliosides as Targets for Immunotherapy for Pancreatic Adenocarcinoma Kyo U. Chu, M.D. Mepur H. Ravindranath, Ph.D. Alexandra Gonzales, B.S. Kevin Nishimoto, B.S. Wai Yin Tam, B.S. Daniel Soh, B.S. Anton Bilchik, M.D., Ph.D. Nonda Katopodis Ph.D. Donald L. Morton, M.D., BACKGROUND. Pancreatic adenocarcinoma cells express gangliosides and sialyl Presented in part at the American College of Surgeons Surgical Forum, Orlando, Florida, October 25–30, 1998, and the 90th Annual Meeting of the American Association for Cancer Research, Philadelphia, Pennsylvania, April 10 –14, 1999. Lewis (sLe) antigens. It is not known whether these carbohydrate antigens can be targeted by immunotherapy. The authors measured the expression of GM2 and sLe antigens on the surface of pancreatic carcinoma cells and the serum levels of total gangliosides, GM2, and antiganglioside antibodies in patients with pancreatic carcinoma. METHODS. Cell surface GM2 and sLe antigens were measured by cell suspension enzyme linked immunoadsorbent assay (ELISA) in four pancreatic carcinoma cell lines. Sera from 20 pancreatic carcinoma patients and 20 age- and gender-matched healthy volunteers were analyzed for antiganglioside and anti-sLe immunoglobulin (Ig) M titers by ELISA. Serum levels of total gangliosides and GM2 also were measured. RESULTS. All cell lines expressed GM2 and sLe antigens. When compared with ageand gender-matched volunteers, patients had significantly higher serum levels of total gangliosides (25.6 ⫾ 9.0 mg/dL vs. 15.6 ⫾ 2.7 mg/dL; P ⬍ 0.001), GM2 (0.278 ⫾ 0.415 mg/dL vs. 0.013 ⫾ 0.018 mg/dL; P ⫽ 0.02), ELISA units of anti-GM2 IgM antibody (368 ⫾ 95 vs. 155 ⫾ 25; P ⫽ 0.04) and anti-GD1b IgM antibody (351 ⫾ 91 vs. 138 ⫾ 26; P ⫽ 0.03), but not anti-sLex IgM (1389 ⫾ 345 vs. 1081 ⫾ 224; P ⫽ 0.46) or anti-sLea IgM antibody (1097 ⫾ 253 vs. 1200 ⫾ 315; P ⫽ 0.80). Patients with unresectable tumors had higher serum levels of total gangliosides compared with patients with resectable tumors, and a serum level ⬎ 25 mg/dL was found to correlate significantly with poor overall survival (P ⬍ 0.02). CONCLUSIONS. Increased serum levels of total gangliosides and GM2 may reflect shedding or release of gangliosides from the surface of tumor cells. Production of IgM antibody against GM2 and GD1b indicates that these gangliosides are immunogenic antigens that may be potential targets for effective active immunotherapy. Cancer 2000;88:1828 –36. © 2000 American Cancer Society. The authors acknowledge support from NC1T32-CA 09689 and Rogovin-Davidow Foundation (Los Angeles). KEYORDS: pancreatic carcinoma, gangliosides, carbohydrate antigens, immunotherapy. The authors thank Dr. Nobuo Hanoi for providing the mouse monoclonal anti-GM2 antibody for this investigation. A Roy E. Coats Research Laboratories and the Sonya Valley Ghidossi Vaccine Laboratory of the John Wayne Cancer Institute at Saint John’s Health Center, Santa Monica, California. Kyo U. Chu’s current address: Division of Surgical Oncology, Department of Surgery, University of Virginia Health System, Charlottesville, Virginia. Address for reprints: Mepur H. Ravindranath, Ph.D., Director, Laboratory of Glycolipid Immunotherapy, John Wayne Cancer Institute, 2200 Santa Monica Boulevard, Santa Monica, CA 90404. Received June 23, 1999; revision received December 3, 1999; accepted December 3, 1999. © 2000 American Cancer Society ccording to recent statistics from the American Cancer Society, in the United States, the estimated number of new cases of invasive pancreatic carcinoma and the estimated number of deaths due to invasive pancreatic carcinoma each approached 30,000 during 1999.1 The most common type of invasive pancreatic carcinoma is ductal adenocarcinoma, which is associated with a poor prognosis. Surgery is the only potentially curative therapy, but complete resection is possible in less than 10% of cases. Combined radiotherapy and chemotherapy have shown some efficacy after surgery.2–5 New chemotherapeutic agents appear to be promising,6 – 8 but as yet none has produced a significant survival benefit. Overall 5-year survival rate remains less than 10%, and overall median survival is less than 1 year. Serum Gangliosides in Pancreatic Carcinoma/Chu et al. Immunotherapeutic investigations for pancreatic carcinoma have been hindered by the absence of information on immunogenic tumor-associated antigens and by the significant immunosuppression caused by tumor-derived factors.9 –11 Pancreatic adenocarcinoma cells express glycoconjugate antigens that are immunosuppressive. Among these glycoconjugates are gangliosides, the glycolipids containing sialic acids.12 Because gangliosides are a target of active specific immunotherapy in patients with melanoma13 and colon carcinoma,14 we hypothesized that they also might play a role in tumor immunity against pancreatic adenocarcinoma. We examined the level of expression of particular gangliosides on the surface of pancreatic cell lines and the serum levels of gangliosides and antiganglioside antibodies in patients with pancreatic adenocarcinoma. PATIENTS AND METHODS Study Population The study population was drawn from patients treated for pancreatic adenocarcinoma at the John Wayne Cancer Institute between January 1991 and December 1997. The 20 patients chosen for this study were those whose follow-up was accurately documented and from whom sera had been preoperatively obtained and cryopreserved. Charts for these patients were retrospectively reviewed, and clinical information was entered into a computer database. Table 1 shows the treatments received by these patients. The control population was drawn from healthy volunteers who donated blood at Saint John’s Health Center during a 3-month interval (January through March) in 1998. Twenty control subjects were selected from this group by computer matching their age and gender with those of the study patients. None of these donors had an active medical problem or was taking any medications. Informed consent was obtained from the subjects to use their sera for the research investigation. This study was approved by the joint institutional review board of Saint John’s Health Center and John Wayne Cancer Institute (Protocol number: RavM-Cellsurf-1196). Measurement of Serum Total Gangliosides The most frequently used assay for measuring the level of serum gangliosides (expressed as lipid-bound sialic acid) is described by Katopodis et al.15 Approximately 30% of the sialic acids extracted with this technique may be derived from sialoproteins, or proteolipids with sialic acid.16,17 However, some of these sialic acid-containing proteins may include ganglioside binding transport proteins17 or immunoglobulins complexed with gangliosides. For routine analysis, the 1829 TABLE 1 Patient Characteristics and Treatments Received Patient no. Age (yrs) Gender Adjuvant treatment 1 71 Male 2 64 Male 3 4 5 7 6 8 9 10 11 12 13 14 15 16 17 18 19 20 87 59 65 70 54 48 68 58 68 47 52 63 56 58 78 63 66 42 Male Female Female Male Female Female Male Female Male Male Male Male Male Male Male Male Female Female Systemic chemotherapy (5-FU); radiation Systemic chemotherapy; hyperthermia; radiation None Radiation Regional chemotherapy; hyperthermia; None None Systemic chemotherapy; hyperthermia; Chemotherapy; radiation None Systemic chemotherapy; radiation Systemic chemotherapy; hyperthermia BCG/systemic chemotherapy Chemotherapy; hyperthermia None Systemic chemotherapy; brachytherapy Brachytherapy None None Systemic chemotherapy; hyperthermia; 5-FU: 5-fluorouracil; BCG: bacillus Calmette–Guerin. serum level of total gangliosides can be measured with an assay that detects lipid-associated sialic acids (LASAs) (Dianon Systems, Inc., Stratford, CT). We checked the reliability and reproducibility of this LASA assay by a double blind analysis. In addition, the normal values reported in this study were comparable to those reported by other investigators. Schutter et al.18 reported that the sensitivity of serum ganglioside level as a marker of various human cancers may vary from 30% to 100% using a cutoff level of 20 mg/dL. The LASA assay appears to have a particularly high positivity rate in leukemia, Hodgkin disease, melanoma, sarcoma, advanced ovarian carcinoma, and oropharyngeal tumors. Briefly, for the LASA assay, 100 L of serum was extracted with 100 L of a chloroform:methanol (2:1 v/v) mixture. After thorough vortexing and centrifuging of the mixture, 100 L of the supernatant was transferred to a clean test tube. The concentration of sialic acid in the supernatant was determined using a resorcinol method described by Svennerholm.19 The resorcinol/HCl reagent was mixed with supernatant and then heated at 100 °C for 15 minutes. After cooling at room temperature, a butylacetate:n-butanol (85:15 volume/volume [v/v]) mixture was added, and the organic layer containing the chromogen was read at an absorbency of 580 nm. Using a standard curve 1830 CANCER April 15, 2000 / Volume 88 / Number 8 obtained with commercial N-acetylneuraminic acid (NeuAc; Sigma, St. Louis, MO), the serum level of total gangliosides was expressed as milligrams per deciliter. Enzyme-Linked Immunosorbent Assay for Serum Anticarbohydrate Antibody Sera from 20 pancreatic carcinoma patients were screened for immunoglobulin (Ig) G and IgM antibodies against GM1, GM2, GM3, GD1a, GD1b, GD2, GD3, and GT1b by enzyme-linked immunosorbent assay (ELISA). Serum titers of antiganglioside antibodies were measured by a method described previously.20 Microtiter plates (Falcon Probind 3915) were coated with an ethanolic suspension of gangliosides (3 nmol/ 100 L/well; Sigma) and dried in a vacuum desiccator for 2 days. The plates were blocked with phosphatebuffered saline (PBS), 4% human serum albumin (HSA), pH 7.2, for 90 minutes. Sera were diluted to 1:100 and incubated in a water bath for 30 minutes at 37 °C. Sera were further serially diluted up to 1:12,800, overlaid on the plates, and incubated at 37 °C for 2 hours. The plates were manually washed five times with washing buffer (PBS 0.1% HSA, 0.1% Tween 20). Antihuman IgM and IgG antibodies coupled to peroxidase (Jackson Immunoresearch Laboratories, Inc., Pittsburgh, PA) were used as the second antibody at a dilution of 1:5000, and the plates were incubated with second antibody (100 L/well) for 1 hour at 37 °C. The plates were washed as before. The substrate, o-phenylenediamine dissolved in citrate–phosphate buffer (pH 5.0) and hydrogen peroxide (4.2 L/10 mL), was added to the plates (100 L/well), and the plates were incubated for 45 minutes in the dark. Enzymatic oxidation was arrested with 6N H2SO4, and absorbency was measured at 490 – 650 nm in a microplate reader (Molecular Devices Corporation, Sunnyvale, CA). The values were corrected against the background (wells without antigen). The titer was determined at an absorbency of 0.1. After identification of high anti-GM2 and GD1b IgM levels in the 20 patients with pancreatic adenocarcinoma, sera from cancer and control groups were analyzed for IgM antibodies against GM2, GD1b, sialyl Lewisx (sLex), and sLea (the antigenic determinant of CA 19-9). Serum titers of anti-sLe antibodies were measured using the method described earlier.21 Microtiter plates (Falcon Probind 3915) were coated with 100 L streptavidin (0.5 g/well; Sigma) suspended in PBS (pH 7.2) and incubated overnight at 4 °C. After removing excess streptavidin, the plates were coated overnight with biotinylated polyacrylamide sLe polymer (3.3 pmol of sLe; Glycotech, Rockville, MD) in carbonate-bicarbonate buffer (pH 9.6). The plates were blocked with 4% HSA PBS (pH 7.2) for 90 minutes at 37 °C. The sera were serially diluted and tested on the antigen-coated plates by ELISA. Pooled sera from melanoma patients served as a positive control for anti-GM2 and anti-GD1b antibodies, and pooled sera from colon carcinoma patients served as a positive control for anti-sLex and anti-sLea antibodies. Immunoglobulin M levels were expressed as ELISA units (patient antibody titer/positive control titer ⫻100). Measurement of Serum GM2 One hundred microliters of serum in polypropylene Eppendorf tubes was admixed with 1 mL of chloroform:methanol (1:1), extensively vortexed, and centrifuged to determine the serum GM2 level. The supernatant was collected, dried over nitrogen gas, and resuspended in 500 L of ethanol. One hundred microliters of the ethanolic suspension was used to coat the wells. After drying in a vacuum, the plates were blocked with PBS 4% HSA. After removing the blocking agent, 100 L of anti-GM2 IgM monoclonal antibody (KM696, 1.48 mg/ml affinity purified IgM; Kyowa Hakko Kogyo Co., Tokyo, Japan) diluted to 1:500 was added to the plates. The GM2 absorbency of patient sera was compared with that of sera from the control group of healthy volunteers. Pancreatic Carcinoma Cell Lines Pancreatic carcinoma cell lines (CRL-1687 [BxPC-3/ 14223], CRL-1682 [AsPC-1/14736], CRL-1837 [SU-86-86/ 15002], and CRL-1918 [CFPAC-1/11588]) were obtained from American Type Culture Collection (Rockville, MD) and cryopreserved. Cells were thawed22 and cultured in RPMI-1640 medium (JRH Biosciences, Lenexa, KS) supplemented with 10% fetal bovine serum (Gemini Bioproducts, Calabassas, CA) and antibiotic antimycotic solution (Pen-Strep-Fungizone 5240-013; Gibco BRL, Gaithersburg, MD). Lack of homogeneity of the cell population from fresh tumors, poor recovery of tumor cells from individual tumors, and inability to grow the cells under our routine culture conditions limited our investigations on the antigen expression to cell lines from ATCC. Quantitation of Cell Surface Antigens with Cell Suspension ELISA Quantitation of cell surface antigens was conducted using methods described previously.22,23 Trypsin was not used for harvesting the cells.15 All cells were washed once with RPMI-4% HSA and then suspended in cold RPMI-4% HSA. Cells were counted for viability using 0.1% trypan blue. Suspensions of 0.5 ⫻ 106 cells in microcentrifuge tubes containing 60 L of solution were treated with the secondary antibody only (background), with primary and secondary antibody (exper- Serum Gangliosides in Pancreatic Carcinoma/Chu et al. 1831 TABLE 2 Serum Levels of Total Gangliosides and CA 19-9 Group n Total gangliosides (mg/dL ⴞ SD) Healthy donors Patients with pancreatic carcinoma Resectable tumor Unresectable tumor AJCC Stage II/III AJCC Stage IV 20 20 5 15 10 10 15.6 ⫾ 2.7 26.0 ⫾ 9.0 19.1 ⫹ 4.9 28.3 ⫹ 9.0 22.7 ⫹ 7.6 29.3 ⫹ 9.5 P value (two-tailed) ⬍ 0.001 ⬍ 0.05 NS CA 19-9 (mg/dL ⴞ SD) 10 ⫾ 8 33594 ⫾ 113547 362 ⫹ 570 44,671 ⫹ 130,275 1466 ⫹ 3794 65,721 ⫹ 157,829 P value (two-tailed) ⬍ 0.001 NS ⬍ 0.05 SD: standard deviation; NS: not significant; AJCC: American Joint Committee on Cancer; NS: not significant. imental), or with class-matched isotypes of the primary antibodies (negative control). The primary antibodies (120 L) used in this study were KM696 (IgM affinity purified, 1.23 mg/500 L; Kyowa Hakko Kogyo Co.; final dilution, 1:500) for GM2, KM 93 (IgM affinity purified, 100 L/L; Kamiya Biomed, Seattle, WA; final dilution, 1:100) for sLex, and KM 231 (IgG1 affinity purified, 100 L/L; Kamiya Biomed; final dilution, 1:100) for sLea. The fine specificity of these antibodies has been well characterized.24 Localization of Immunogenic Epitopes in Glycoconjugates by Enzyme-Treated Cell Suspension ELISA Although gangliosides are glycolipids, the sugar epitope of the gangliosides also may be found in glycoproteins.25 Therefore, the carrier molecules of GM2 and sLe can be proteins or lipids. Furthermore, gangliosides on the cell surface may be masked by transmembrane glycoproteins. To unmask gangliosides and distinguish their sugar epitopes from the carbohydrate epitopes on proteins, we subjected pancreatic carcinoma cells to a variety of enzymatic treatments. To facilitate the access of enzymes to the peptide core of cell surface glycoproteins, cells first were treated with sialidase from Vibrio cholerae (Boehringer Mannheim, Indianapolis, IN) for 2 hours at 37 °C. N-linked and O-linked oliogosaccharides then were cleaved with N- and O-glycosidases (Boehringer Mannheim). Pronase (Boehringer Mannheim) was used to cleave the peptide core of cell surface proteins. The enzyme activity was arrested with RPMI 4% HSA. Antibody binding was assessed after washing viable cells three times with RPMI-1640 4% HSA. survival differences among patient subgroups. Overall survival was defined as the period between the time of diagnosis and the patient’s death or last follow-up visit. Multivariate analysis using Cox proportional hazards regression model was performed to determine the impact of significant factors on survival. Serum levels of gangliosides, CA 19-9, anti-GM2 antibody, and anti-GD1b antibody were analyzed as continuous variables in multivariate analysis. All statistical analyses were two-tailed and performed with SAS software (SAS Inc., Cary, NC). Results were considered significant if the probability value was less than 0.05. RESULTS Of the 20 patients with pancreatic adenocarcinoma, 5 underwent potentially curative resection, and 15 were found to have unresectable disease either before or during laparotomy. Ten patients had distant metastatic disease (American Joint Committee on Cancer [AJCC] Stage IV26), nine had regional metastases (AJCC Stage III), and one had localized disease (AJCC Stage II). Median survival of the 20 patients was 10 months; mean survival was 14.4 months. Serum CA 19-9 Levels Serum levels of CA 19-9 were elevated in 17 of 20 patients and were significantly (P ⬍ 0.001) higher in the cancer group than the control group (Table 2). Serum levels of CA 19-9 also were significantly higher (P ⬍ 0.05) in Stage IV patients than Stage III patients. CA 19-9 levels did not differ for patients with resectable versus unresectable disease. Serum Total Ganglioside Levels Statistical Analysis Statistical analyses of serum levels of total gangliosides, GM2, and anticarbohydrate antibodies were performed by Wilcoxon rank-sum test. Survival curves were estimated by the Kaplan–Meier method. Univariate analysis by log rank test was used to determine Serum levels of total gangliosides in the 20 healthy volunteers ranged from 12.1 to 22.2 mg/dL; the mean (15.6 ⫾ 2.7 mg/dL) was comparable to that of normal individuals reported by other investigators.18 Serum ganglioside levels were above the normal reference range of 15–20 mg/dL in 16 of 20 patients and in 1 of 1832 CANCER April 15, 2000 / Volume 88 / Number 8 not significantly higher in the cancer group at any stage of disease. Serum GM2 Level FIGURE 1. Kaplan–Meier curves for patients with pancreatic adenocarcinoma and serum ganglioside (SG) levels ⱖ 25 mg/dL versus ⬍ 25 mg/dL are shown. Median overall survival was 14 months for patients with SG levels ⬍ 25 mg/dL and 10 months for patients with SG levels ⱖ 25 mg/dL (log rank test, P ⫽ 0.03). SG: serum ganglioside. 20 normal volunteers. Serum ganglioside levels were significantly higher in the cancer group than in the control group (25.6 ⫾ 9.0 vs. 15.6 ⫾ 2.7 mg/dL; P ⬍ 0.001) (Table 2). Although patients with Stage IV disease had a higher level of serum gangliosides than did patients with Stage II/III disease, the difference was not significant. However, serum ganglioside levels were significantly higher in patients with unresectable disease than in those with resectable disease (28.3 ⫾ 9.0 vs. 19.1 ⫾ 4.9 mg/dL; P ⬍ 0.05). In addition, overall survival was significantly better (P ⬍ 0.02) in patients with low serum levels of total gangliosides (⬍ 25.00 mg/dL) than in those with high levels (ⱖ 25.00 mg/dL) (Fig. 1). Multivariate analysis of prognostic factors including serum levels of gangliosides and CA 19-9, AJCC stage, and patient gender showed that only serum gangliosides had a significant correlation with the overall survival of patients with pancreatic adenocarcinoma (P ⬍ 0.05). IgM Levels against Gangliosides and sLeX and sLeA Sera from patients with cancer had high levels of antiGM2 and anti-GD1b IgM antibodies, very low levels of IgM antibodies to other gangliosides, and very low levels of IgG antibodies to all gangliosides (data not shown). When compared with the control group, the cancer group had higher levels of anti-GD1b IgM (351 ⫾ 91 vs. 138 ⫾ 26; P ⫽ 0.03) and anti-GM2 IgM (368 ⫾ 95 vs. 155 ⫾ 25; P ⫽0.04). However, anti-sLex IgM (1389 ⫾ 345 vs. 1081 ⫾ 224; P ⫽0.46) and anti-sLea IgM titers (1097 ⫾ 253 vs. 1200 ⫾ 315; P ⫽0.80) were Because the cancer group had significantly higher (P ⬍ 0.05) serum titers of anti-GM2 IgM antibodies, we estimated serum GM2 levels by using an anti-GM2 IgM mouse monoclonal antibody (KM696) to ethanolic supernatants of chloroform:methanol extracts. Box titration of varying concentrations of KM696 against varying concentrations of GM2 showed that the ELISA absorbency was directly proportional to GM2 concentration (data not shown). Serum GM2 absorbency ranged from 0 to 0.04 (mean ⫾ standard deviation [SD], 0.013 ⫾ 0.018) in the control group of healthy volunteers. Serum GM2 absorbency was tested in 17 patients in the cancer group: the GM2 absorbency was high in eight patients (range, 0.120 –1.414; mean ⫾ SD, 0.559 ⫾ 0.472). In the other nine patients, the level of GM2 ranged from 0.013 to 0.086 (mean ⫾ SD, 0.028 ⫾ 0.023). However, overall, the serum GM2 level was significantly higher in the cancer group (0.278 ⫾ 0.415 vs. 0.013 ⫾ 0.018; P ⫽ 0.02) than in the control group. Density of GM2 on Pancreatic Carcinoma Cells Because sera from 8 of 17 pancreatic carcinoma patients had significantly higher levels of GM2 than did sera from control patients, we measured the density of GM2 expression on the surface of cells from four pancreatic carcinoma lines. All pancreatic cell lines expressed GM2, but the level of expression was highest in a cell line developed from a liver metastasis (Figure 2). Expression of GM2 increased after the removal of Olinked or N-linked oligosaccharides associated with cell surface proteins. Expression of GM2 was higher after sialidase and O-glycosidase treatment or after pronase treatment after deglycosylation (Figure 3). Similarly, expression of GM2 was higher after N-glycosidase treatment (data not shown), suggesting that the carbohydrate moiety of glycoproteins in pancreatic carcinomas may be linked to serine and threonine (O-linked) as well as to asparagine (N-linked). The level of expression of GM2 may remain the same or decrease slightly when deglycosylation is followed by treatment with pronase. Interestingly cell surface expression of sLex was markedly reduced after deglycosylation (both N- and O-glycosidase), suggesting that this antigen may be linked to serine/threonine as well as to the asparagine in the protein moiety. DISCUSSION Our results document an apparent correlation between serum ganglioside levels and tumor burden in Serum Gangliosides in Pancreatic Carcinoma/Chu et al. 1833 FIGURE 2. Expression of GM2 on cells of tumor cell lines derived from metastatic lesions of the liver and pancreas and from ascites is shown. The cell surface glycocalyx profile was monitored by cell suspension ELISA by using monospecific monoclonal antibodies, as described in Patients and Methods. ELISA: enzyme-linked immunosorbent assay. pancreatic adenocarcinoma, similar to earlier reports.18 Levels of serum gangliosides are known to increase with the stage of the disease and the extent of metastasis in other malignancies.18 Our study population was too small to reveal a statistically significant increase in the level of serum gangliosides among AJCC Stage IV patients, but the level of serum gangliosides was significantly higher in patients with unresectable disease. Patients with unresectable disease consistently had serum ganglioside levels greater than 25 mg/dL, and multivariate analysis identified serum level of total gangliosides as the only factor that significantly influenced overall survival. These data suggest that patients with serum ganglioside levels less than 25 mg/dL are more likely to be candidates for complete surgical resection and have better survival. Serum ganglioside level as a marker resembles K-ras gene mutation in the plasma DNA of patients with pancreatic carcinoma. K-ras mutation in codon 12 has been observed in pancreatic tissues of patients with pancreatic adenocarcinoma and pancreatitis.27,28 In 9 of 16 patients with K-ras gene mutation-positive tumors, an identical mutation was detected in the plasma DNA.27 Such mutations in plasma DNA are not common in patients with chronic pancreatitis (⬍ 5%)28 or in healthy subjects.27 Plasma K-ras gene mutation, similar to serum ganglioside level, reflected tumor burden, failed to correlate with other clinicopathologic features (age, gender histologic type, mode of invasion, and metastasis),27 correlated with poor survival, and was identified as the only independent prognostic factor.28 It would be worthwhile to investigate serum ganglioside level and K-ras gene mutation in plasma DNA as a combined-marker system to evaluate tumor burden and resectability in patients with pancreatic carcinoma. Because the K-ras gene point mutation at codon 12 changes the amino acid sequence (the normal Gly residue is substituted with either a Val, Asp, or Cys residue), the K-ras products are considered to be unique epitopes for T-cell (CD4⫹ and/or CD8⫹) recognition in cancer immunotherapy.29 Although serum levels of CA19-9 were higher in patients with Stage IV disease, their tremendous variation concealed any significant elevation in patients with unresectable disease, even though these patients are more likely to have a larger tumor burden. This variation makes CA 19-9 less reliable to determine tumor burden. Most of the commercial antibodies for CA 19-9 (121SLE [IgM], Neomarkers, Union City, CA; 1116-NS-19-9 [IgG1], Signet Laboratories, Dedham, MA; M8073022 [IgG1], Fitzgerald, Concord, MA; ZYC09 [IgG1], Zymed, San Francisco, CA; MED-CLA 143, Accurate Chemical and Scientific Corporation, Westbury, NY; NS19-9, International CIS, Cedex, France) are directed against the sugar residues of sLea, which is the antigenic determinant of CA 19-9.30 Sialyl Lewisa is expressed in glycoproteins or glycolipids.30 –35 The glycoproteins expressing sLea include mucin-type glycoproteins belonging to MUC family, such as MUC-1. Tumor cells also may have more than one glycoprotein expressing the epitope of sLea. The heterogeneity of pancreatic tumor cells expressing these sLea-associated antigens and the remarkable variability in the carriers of sLea explain the variability observed in the serum level of CA 19-9. Although our pancreatic carcinoma patients had a high level of IgM antibodies against GM2 and GD1b, they did not have elevated levels of IgM antibodies against sLea and sLex. This suggests that GM2 and GD1b but not sLea or sLex are immunogenic antigens in pancreatic carcinoma patients. 1834 CANCER April 15, 2000 / Volume 88 / Number 8 FIGURE 3. Expression of GM2 on four human pancreatic carcinoma (PC) cell lines after deglycosylation. There was a noteworthy increase in the expression of GM2 after deglycosylation or pronase treatment. The cell surface profile after deglycosylation was monitored by enzyme-linked immunoadsorbent assay (ELISA) cell suspension using monospecific monoclonal antibodies, as described in Patients and Methods. The vertical bars represent standard error of five analyses. sLex: sialyl Lewisx. The presence of anti-GM2 IgM antibody in patients with pancreatic carcinoma is noteworthy in view of the observed correlation between elevated anti-GM2 IgM antibody titer and improved survival of patients with melanoma.36 –38 Because anti-GM2 IgM antibodies can fix complement and clear antigens from the circulation, we assayed the levels of GM2 in sera and tumor tissue. Eight of 17 patients in the cancer group had significantly higher serum levels of GM2 than did age- and gender-matched healthy volunteers. Our findings suggest that GM2 on pancreatic carcinoma cells can be a potential target for immunotherapy. Deglycosylation significantly augments expression of GM2, possibly due to deletion of the oligosaccharide moiety of glycoproteins on the cell surface. Interestingly, deglycosylation results in loss of sLex expression and further facilitates access to GM2. The use of factors such as cytokines and drugs to down-regulate glycoprotein expression on the cell surface deserves further study. The apparent inverse correlation between a serum level of total gangliosides greater than 25 mg/dL and survival may be related to increased tumor burden or more advanced stage of disease. It also may reflect the immunosuppressive effect of circulating gangliosides. Gangliosides shed from a growing tumor39 can suppress expression of CD4 on human helper T lymphocytes,40 expression of major histocompatibility complex Class I and Class II genes,41 and mitogen-induced lymphoblastic transformation. They can also induce T-suppressor activity. Circulating gangliosides can reduce the number of macrophages and CD4⫹ and CD8⫹ T lymphocytes,42 and they can inhibit interleukin-2– dependent lymphocyte proliferation43 by preventing the interaction of this cytokine with its cell surface receptors. If ganglioside-induced immunosuppression is linked with poor prognosis, then clearance of tumor-derived gangliosides from the circulation may be of clinical benefit for patients with unresectable disease. The immunosuppressive role of GM2, particularly its ability to mimic the receptor for interleukin-2, warrants its elimination from circulation. The potential efficacy of anti-GM2 IgM antibody for clearing the serum of immunosuppressive GM2 and for killing tumor cells by complement-mediated cytotoxicity or apoptosis44 makes augmentation of anti-GM2 IgM antibodies a promising immunotherapeutic strategy. The level of anti-GM2 IgM can be augmented by passive administration of a murine antiGM2 IgM such as KM696; however, the development of anti-mouse antibody responses45 limits the efficacy of repeated administration. Nishinaka et al.46 developed a human anti-GM2 monoclonal antibody that can cause complement-dependent cytotoxicity, but its Serum Gangliosides in Pancreatic Carcinoma/Chu et al. production is not cost-effective. Nakamura et al.44,47 constructed mouse/human chimeric anti-GM2 IgM KM966 (mouse residue 33%; human residue 67%) and humanized chimeric anti-GM2 IgM KM8969 (mouse residue 10%; human residue 90%) with therapeutically higher complement-mediated cytotoxicity, antibodydependent cytotoxicity, and apoptosis. However, Bator and Reading45 reported that anti-idiotypic and antiallotypic responses against human or humanized monoclonal antibodies may negate any therapeutic benefit of repeated administration. Suggested solutions to these problems may not be cost-effective for treatment purposes. An alternative to passive immunotherapy using various anti-GM2 antibodies is active specific immunotherapy using purified GM2. Livingston48 augmented the immunogenicity of purified GM2 by conjugating it to hemocyanin, a xenogenic protein with a molecular weight greater than 400 kilodaltons. However, the humoral and cellular immune responses induced by hemocyanin may prohibit effective use of the vaccine. The poor immunogenicity of purified GM2 can be improved by presenting this ganglioside in the context of the cellular membrane.36 In human melanoma, vaccination with tumor cells expressing GM2 can augment serum levels of anti-GM2 IgM, and elevated anti-GM2 IgM levels have been correlated with improved survival.49 Therefore, use of a GM2based carcinoma cell vaccine may be a clinically effective strategy for patients with pancreatic carcinoma. In summary, we have demonstrated that serum gangliosides are elevated in patients with pancreatic adenocarcinoma, suggesting their overexpression and release by pancreatic adenocarcinoma cells. Our study also shows that GM2 and GD1b are immunogenic antigens and thus potential targets for immunotherapy in patients with pancreatic adenocarcinoma. In our study, serum gangliosides were more reliable than CA19-9 as an indicator of tumor burden. Serum ganglioside levels also may be useful for monitoring disease progression and response to surgery and postoperative adjuvant therapy. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. REFERENCES 1. 2. 3. Landis SH, Murray T, Bolden S, Wingo PA. Cancer statistics, 1999. CA Cancer J Clin 1999;49:8 –31. Pisters PW, Abbruzzese JL, Janjan NA, Cleary KR, Charnsangavej C, Goswitz MS, et al. Rapid-fractionation preoperative chemoradiation, pancreaticoduodenectomy, and intraoperative radiation therapy for resectable pancreatic adenocarcinoma. J Clin Oncol 1998;16:3843–50. Yeo CJ, Abrams RA, Grochow LB, Sohn TA, Ord SE, Hruban RH, et al. Pancreaticoduodenectomy for pancreatic adenocarcinoma: postoperative adjuvant chemoradiation im- 18. 19. 1835 proves survival. A prospective, single-institution experience. Ann Surg 1997;225:621–36. Demeure MJ, Doffek KM, Komorowski RA, Redlich PN, Zhu YR, Erickson BA, et al. Molecular metastases in stage I pancreatic cancer: improved survival with adjuvant chemoradiation. Surgery 1998;124:663–9. Ishii H, Okada S, Tokuuye K, Nose H, Okusaka T, Yoshimori M, et al. Protracted 5-fluorouracil infusion with concurrent radiotherapy as a treatment for locally advanced pancreatic carcinoma. Cancer 1997;79:1516 –20. Glimelius B. Chemotherapy in the treatment of cancer of the pancreas. J Hepatobiliary Pancreat Surg 1998;5:235– 41. Burris HA 3rd, Moore MJ, Andersen J, Green MR, Rothenberg ML, Modiano MR, et al. Improvements in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreas cancer: a randomized trial. J Clin Oncol 1997;15:2403–13. Todd KE, Gloor B, Lane JS, Isacoff WH, Reber HA. Resection of locally advanced pancreatic cancer after downstaging with continuous-infusion 5-fluorouracil, mitomycin-C, leucovorin, and dipyridamole. J Gastrointest Surg 1998;2:159 – 66. Weiner LM, Harvey E, Padavic-Shaller K, Willson JK, Walsh C, LaCreta F, et al. Phase II multicenter evaluation of prolonged murine monoclonal antibody 17-1A therapy in pancreatic carcinoma. J Immunother 1993;13:110 – 6. Brown TD, Goodman P, Fleming T, Macdonald JS, Hersh EM, Braun TJ. A phase II trial of recombinant tumor necrosis factor in patients with adenocarcinoma of the pancreas: a Southwest Oncology Group study. J Immunother 1991;10: 376 – 8. Buchler M, Friess H, Schultheiss KH, Gebhardt C, Kubel R, Muhrer KH, et al. A randomized controlled trial of adjuvant immunotherapy (murine monoclonal antibody 494/32) in resectable pancreatic cancer. Cancer 1991;68:1507–12. Fredman P, Nilsson O, Svennerholm L, Myrvold H, Persson B, Pettersson S, et al. Colorectal carcinomas have a characteristic ganglioside pattern. Med Biol 1983;61:45– 8. Morton DL, Ravindranath MH, Irie RF. Tumor gangliosides as targets for active specific immunotherapy of melanoma in man. Prog Brain Res 1994;101:251–75. Ravindranath MH, Bilchik AJ, Chu K, Soh D, Shen P, Morton DL. Anti-GM2 IgM titer inversely correlates with the serum ganglioside level in colon cancer patients after vaccine therapy. Proc Am Assoc Cancer Res 1999;90:3801. Katopodis N, Hirschaut Y, Geller NL, Stock CC. Lipid-associated sialic acid test for the detection of human cancer. Cancer Res 1982;42:5270 –5. Skipski VP, Barclay M, Archibald FM, Lynch TP Jr., Stock CC. A new proteolipid apparently associated with cancer. Proc Soc Exp Biol Med 1971;136:1261– 4. Tonegawa Y, Hakamori S-I. “Ganglioprotein and globoprotein”: the glycoprotein reacting with anti-ganglioside and anti-globoside antibodies and the ganglioprotein change associated with transformation. Biochem Biophys Res Commun 1977;76:9 –17. Schutter EMJ, Visser JJ, van Kamp GJ, Mensdorff-Pouilly S, van Dijk W, Hilgers J, et al. The utility of lipid bound sialic acid (LASA or LSA) as a serum marker for malignancy. A review of literature. Tumor Biol 1992;13:121–32. Svennerholm L. Quantitative estimation of sialic acid. II. A colorimetric resorcinol-hydrochloric acid method. Biochim Biophys Acta 1957;24:604 –11. 1836 CANCER April 15, 2000 / Volume 88 / Number 8 20. Ravindranath MH, Ravindranath RMH, Morton DL, Graves MC. Factors affecting the fine specificity and sensitivity of serum antiganglioside antibodies in ELISA. J Immunol Methods 1994;169:257–72. 21. Ravindranath MH, Amiri AA, Bauer PM, Kelley MC, Essner R, Morton DL. Endothelial-selectin ligands sialyl Lex and sialyl Lea are differentiation antigens immunogenic in human melanoma. Cancer 1997;79:1686 –97. 22. Ravindranath MH, Amiri AA, Bauer PM. A sensitive cell suspension ELISA (cs-ELISA) for measuring the density of cell-surface ganglioside and carbohydrate antigens. In: Townsend R, editor. Techniques in glycobiology. New York: Marcel Dekker, 1997:137– 65. 23. Ravindranath MH, Bauer PM, Cornillez-Ty C, Garcia J, Morton DL. Quantitation of the density of cell-surface carbohydrate antigens on cancer cells with a sensitive cell-suspension ELISA. J Immunol Methods 1996;197:51– 67. 24. Hanai N, Shitara K, Furuya A, Yoshida H, Dohi T, Nudelman E, et al. Detailed characterization of reactivities of antigastric cancer monoclonal antibodies to carbohydrate antigen. Anticancer Res 1990;10:1579 – 86. 25. Hiraiwa N, Tsuyuoka K, Li YT, Tanaka M, Seno S, Okubo Y, et al. Gangliosides and sialoglycoproteins carrying a rare blood group antigen determinant, Cad, associated with human cancers as detected by specific monoclonal antibodies. Cancer Res 1990;50:5497–503. 26. Fleming ID, Cooper JS, Henson DE, Hutter RVP, Kennedy BJ, Murphy GP, et al., editors. AJCC cancer staging manual. 5th ed. Philadelphia: Lippincott-Raven, 1997. 27. Yamada T, Nakamori S, Ohzato H, Oshima S, Aoki T, Higaki N, et al. Detection of K-ras gene mutations in plasma DNA of patients with pancreatic adenocarcinoma: correlation with clinicopathological features. Clin Cancer Res 1998;4: 1527–32. 28. Castells A, Puig A, Mora J, Boadas J, Boix L, Urgell E, et al. K-ras mutations in DNA extracted from the plasma of patients with pancreatic carcinoma: diagnostic utility and prognostic significance. J Clin Oncol 1999;17:578 – 84. 29. Abrams SI, Hand PH, Tsang KY, Scholom J. Mutant ras epitopes as targets for cancer vaccines. Semin Oncol 1996; 23:118 –34. 30. Magnani JL, Brockhaus M, Smith DF, Ginsburg V, Blaszczyk M, Mitchell KF, et al. A monosialoganglioside is a monoclonal antibody-defined antigen of colon carcinoma. Science 1981;212:55– 6. 31. Magnani JL, Nilsson B, Brockhaus M, Zopf D, Steplewski Z, Koprowski H, et al. A monoclonal antibody-defined antigen associated with gastrointestinal cancer is a ganglioside containing sialylated lacto-N-fucopentaose II. J Biol Chem 1982; 257:14365–9. 32. Gupta MK, Arciaga R, Bocci L, Tubbs L, Bukowski R, Deodhar SD. Measurement of a monoclonal-antibody-defined antigen (CA19-9) in the sera of patients with malignant and nonmalignant disease: comparison with carcinoembryonic antigen. Cancer 1985;56:277– 83. 33. Ho JJL, Siddiki B, Kin YS. Association of sialyl Lewisa and sialyl Lewisx with MUC-1 apomucin in a pancreatic cancer cell line. Cancer Res 1995;55:3659 – 63. 34. Baeckstrom D, Nilsson O, Price MR, Lindholm L, Hansson GC. Discrimination of MUC1 mucins from other sialyl Lea carrying glycoproteins produced by colon carcinoma cells using a novel monoclonal antibody. Cancer Res 1993;53:755– 61. 35. Ho JJL, Kim YS. Serological pancreatic tumor markers and the MUC1 apomucin. Pancreas 1994;9:674 –91. 36. Ravindranath MH, Morton DL. Immunogenicity of membrane-bound gangliosides in viable whole-cell vaccines. Cancer Invest 1977;15:491–9. 37. Livingston PO, Wong GY, Adluri S, Tao Y, Padavan M, Parente R, et al. Improved survival in stage III melanoma patients with GM2 antibodies: a randomized trial of adjuvant vaccination with GM2 ganglioside. J Clin Oncol 1994;12: 1036 – 44. 38. Ravindranath MH, Hsueh E, Jones RC, Morton DL. Anti-GD3 IgM antibody induced by polyvalent melanoma cell vaccine is significantly correlated with survival of patients with regional metastatic melanoma. Proc Am Assoc Cancer Res 1998;89:369. 39. Valentino L, Moss T, Olson E, Wang HJ, Elashoff R, Ladisch S. Shed tumor gangliosides and progression of human neuroblastoma. Blood 1990;75:1564 –7. 40. Offner H, Thieme T, Vandenbark AA. Gangliosides induce selective modulation of CD4 from helper T lymphocytes. J Immunol 1987;139:3295–305. 41. Massa PT. Specific suppression of major histocompatibility complex class I and class II genes in astrocytes by brainenriched gangliosides. J Exp Med 1993;178:1357– 63. 42. Robb RJ. The suppressive effect of gangliosides upon IL 2-dependent proliferation as a function of inhibition of IL 2-receptor association. J Immunol 1986;136:971– 6. 43. Chu WK, Sharom FJ. Gangliosides inhibit T-lymphocyte proliferation by preventing the interaction of interleukin-2 with its cell surface receptors. Immunology 1993;79:10 –7. 44. Nakamura K, Hanibuchi M, Yano S, Tanaka Y, Fujino I, Inoue M, et al. Apoptosis induction of human lung cancer cell line in multicellular heterospheroids with humanized antiganglioside GM2 monoclonal antibody. Cancer Res 1999;59:5323–30. 45. Bator JM, Reading CL. Antibody-mediated cancer diagnosis and therapy. In: Borrebaeck CAK, Larrick JW, editors. Therapeutic monoclonal antibodies. Chapter 3. New York: Stockton Press, 1990: 35–56. 46. Nishinaka Y, Ravindranath MH, Irie RF. Development of a human monoclonal antibody to ganglioside GM2 with potential for cancer treatment. Cancer Res 1996;56:5666 –71. 47. Nakamura K, Koike M, Shitara K, Kuwana Y, Kiuragi K, Igarashi S, et al. Chimeric anti-ganglioside GM2 antibody with antitumor activity. Cancer Res 1994;54:1511– 6. 48. Livingston PO. Approaches to augmenting the immunogenicity of melanoma gangliosides: from whole melanoma cells to ganglioside-KLH conjugate vaccines. Immunol Rev 1995;145:147– 66. 49. Morton DL, Foshag LJ, Hoon DSB, Nizze JA, Famatiga E,Wanek LA, et al. Prolongation of survival in metastatic melanoma after active specific immunotherapy with a new polyvalent melanoma vaccine. Ann Surg 1992;216:465– 82.