Cancer
Immunity, Vol. 6, p. 5 (9 March 2006)
Submitted: 17 March 2005. Resubmitted: 21 December 2005. Accepted: 22 December 2005.
Communicated by: V Cerundolo
Effects of desialylation of ovine submaxillary gland mucin (OSM) on humoral and cellular immune responses to Tn and sialylated Tn
Kevin P. O'Boyle1, Susan Coatsworth2, Gillian Anthony2, Mark Ramirez2, Edward Greenwald2, Ronald Kaleya2, Jacob J. Steinberg2, Janice P. Dutcher1, and Peter H. Wiernik1
1Our Lady of Mercy Cancer Center, New York Medical College, 600 E. 233rd Street, Bronx, NY 10466, USA 2Montefiore Medical Center, Albert Einstein College of Medicine, 111 E. 210th Street, Bronx, NY 10467, USA
Resected carcinoma patients were immunized 3-5 times with ovine submaxillary gland mucin (OSM) containing predominantly sialylated Tn (sTn), completely desialylated ovine submaxillary gland mucin (dOSM) containing predominantly Tn, or 50% desialylated OSM containing Tn and sTn plus bacillus Calmette-Guerin (BCG) as an immunologic adjuvant. Pre- and postimmunization sera were quantified by ELISA, whole-cell ELISA, and immune stain dot blots. Fifteen of 17 patients produced IgG antibody titers from 40 to 5120 times more reactive with OSM and dOSM postimmunization. More importantly, these IgG antibodies reacted with LS-174T, a human colon carcinoma cell line. Significant DTH-like responses (1-17 cm) were observed in 15 of 17 patients; the strength of these responses was dependent on the presence or absence of sialic acid. Biopsies of these DTH-like reactions revealed infiltration with some CD8+ lymphocytes and mast cells. These results suggest that a single 9-carbon sugar can affect cellular immune responses to mucin antigens. It is thought that these large erythematous, nonindurated cellular reactions are antibody-mediated Arthus-like reactions. OSM, and especially dOSM, were also found to inhibit lymphocyte proliferation.
Introduction
Tumor-specific transplantation antigens identified in syngeneic animal models are responsible for tumor protection when these animals are challenged with chemically induced tumors (1). Autologous vaccines (2) have yielded some promising results in human clinical trials, but sufficient tumor for vaccine production is often not available, and antigens in such vaccines are not well defined or presented in their most immunogenic form. The study of generic vaccines for cancer, including allogeneic, xenogeneic, synthetic, anti-idiotypic, gene-modified, and recombinant vaccines is proceeding rapidly (3, 4, 5). Certain blood group-associated carbohydrate antigens have been identified that are immunogenic in humans and the expression of which has been correlated with prognosis in several carcinomas (6).
Springer (7) demonstrated that T (Thomsen-Freidenreich) antigen and its precursor, Tn, are pancarcinoma antigens that are immunogenic in humans and may provide the basis for immunotherapy. Tn is a cryptic core α-N-acetylgalactosamine (Gal-NAc) O-linked to serine or threonine on mucin molecules. The addition of a terminal β-linked galactose to Tn results in T antigen, which becomes exposed in carcinomas due to incomplete and aberrant glycosylation. Sialylated Tn is the antigen for mAb B72.3 and is expressed in breast, pancreas, colon, gastric, lung, and ovarian cancers (8).
T, Tn, and sTn are expressed in greater abundance on carcinoma mucins than on normal mucins (9). Mucins are highly glycosylated, high molecular weight glycoproteins that protect normal mucosal cells from chemical and physical trauma. They can also protect cancer cells by masking MHC antigens (10). Mucins have long been associated with a poor prognosis in colon cancer (11, 12), although they are associated with a good prognosis in breast cancer (13). Increased expression of several carbohydrate antigens, including T, sTn, and Tn, has been correlated with a poor prognosis in colon, breast, gastric, ovarian, and bladder cancer (14, 15, 16). Circulating levels of autoantibodies to T antigen were shown to decrease as patients with gastrointestinal cancer progressed, presumably as a result of immune complex formation (17).
Mucin genes that encode various peptide backbones have been cloned and their tissue and tumor distributions studied (18, 19, 20, 21). MUC1 (chromosome 1) is overexpressed in breast and pancreas cancer. MUC1, MUC2 (chromosome 11), and MUC3 (chromosome 7) are associated with normal colonic mucosa and colon cancer. MUC1 is expressed in clear cell renal carcinoma, and its expression level has been correlated with tumor progression and nuclear grade in pT1 renal clear cell carcinoma (22). MUC2 and MUC4 (chromosome 3) are present in normal tracheobronchial mucus. Serum MUC5AC is predictive of poor patient outcome in patients with cholangiocarcinoma (23). Posttranslational, transformation-dependent modifications of MUC1 have been found for MUC1-6 in papillary thyroid carcinoma (24). Exfoliated cells in urine from bladder cancer patients express MUC7 gene (25).
These mucin genes include a tandem repeat unit that encodes a peptide sequence of varying length depending on the gene, and this can be repeated up to 100-200 times in the peptide backbone of the mucin. These repeating peptide units are preferentially exposed in tumors due to incomplete glycosylation, and an MHC-unrestricted cytotoxic T-cell response can be elicited to these tandem repeat units (26, 27). Common amino acids that make up these tandem repeat units are serine and threonine to which carbohydrate epitopes are O-linked. Mucin antigens have a restricted distribution in normal tissues and are, therefore, good targets for immunologic manipulation.
Mice immunized with completely dOSM, containing predominantly Tn, were protected against challenge from a very tumorigenic, syngeneic mouse mammary carcinoma cell line, TA3Ha (28). High serum anti-Tn titers of both IgM and IgG isotypes were detected in these mice. More importantly, there was evidence of T-cell immunity directed against the carbohydrate Tn epitope. DTH responses, measured as footpad swelling, were elicited to dOSM and irradiated TA3-Ha cells in this model. Lymphocytes were found to proliferate in response to dOSM but not to deglycosylated dOSM, suggesting that the Tn epitope was required for the response. The phenotype of the reactive lymphocytes was determined to be CD4+/CD8-. Recombinant cytokines, including IL-2, have been shown to boost the specific antibody response to OSM (29).
Previously, 20 colon cancer patients at risk for recurrence were immunized with partially desialylated OSM (containing approximately equivalent amounts of Tn and sTn), and this resulted in production of IgM and IgG antibodies reactive with Tn and sTn (30). Administration of an immunologic adjuvant such as BCG was absolutely essential for the production of antibody to modified OSM. Human IgM mAbs generated from one of the patients immunized in this trial were shown to react with Tn and sTn, with a mucinous human colon cancer cell line expressing Tn and sTn, and with paraffin-embedded colon carcinomas (31).
In order to test whether carbohydrate antigens can induce cellular as well as humoral immunity, we initiated another phase I study in which patients at risk of recurrence were immunized with OSM vaccine desialylated to different degrees. Recent evidence suggests that under special circumstances it may be possible to generate cell-mediated immunity to carbohydrate antigens. Ganglioside GM3 encapsulated into liposomes induced cytotoxic T lymphocyte activity against GM3-expressing melanoma (32). T-cell lines generated from the cerebrospinal fluid of patients with multiple sclerosis were found to recognize purified gangliosides GM1, GD1a, and GD1b (33). Recent studies (34, 35, 36, 37) have shown that many glycopeptide determinants can induce T-cell immunity that is glycopeptide-specific. T-cell recognition of the glycopeptide is lost when the carbohydrate is removed.
We chose OSM desialylated to varying degrees as our vaccine for a number of reasons. As a result of our experience immunizing mice with a synthetic T-KLH (T-keyhole limpet hemocyanin) glycoconjugate that yielded high affinity IgG1 monoclonal antibodies reactive with synthetic T-human serum albumin but not desialylated glycophorin (containing a natural source of T antigen) (38), we prefer a natural source of Tn and sTn antigen over synthetic glycoconjugates. Desialylation of OSM allows the optimal ratio of Tn:sTn for the vaccine to be determined and makes it possible to analyze the effect of removal of sialic acid on the immunogenicity of the vaccine (39).
BCG was chosen as the adjuvant for this vaccine study because the only evidence of a cell-mediated immune response found in our initial study with partially desialylated OSM was obtained with patients who received BCG. BCG is a lyophilized preparation of live, attenuated organisms derived from Mycobacterium bovis. It is very well tolerated. The major toxicity is local inflammation and ulceration at the vaccination site. The dose of BCG was reduced by a factor of 3 with each subsequent immunization if skin toxicity was grade 3 or greater.
In this clinical trial, patients were not pretreated with cyclophosphamide to inhibit suppressor T cells because, despite evidence in animal models that it augments humoral and cellular immunity, the role of cyclophosphamide in clinical vaccine trials has not been proven (40). The use of cyclophosphamide exposes patients to unnecessary toxicity and may also prevent them from entering other clinical trials with investigational agents should their cancer recur. We chose to immunize patients without advanced cancer, but with a significant risk of recurrence, because they are immunocompetent, and because a vaccine has the best chance of working when there is minimal residual disease. All patients had a pathologically confirmed diagnosis of carcinoma, a good performance status, were documented to be free of disease after surgery, and were not concurrently receiving chemotherapy or radiation.
Tn and sTn are expressed in minimal amounts by isolated secretory and ductal cells in the salivary glands, colon, pancreas, and esophageal lining; otherwise, their normal tissue distribution is highly restricted (41). Autoimmune hypersensitivity reactions to components of the vaccine or to skin test antigens were a theoretical concern. The only toxicity noted in the previous study was local skin toxicity attributed to the immunologic adjuvants, such as BCG, given with the OSM vaccine.
Since the patients entered into this phase Ib adjuvant study do not share a single tumor type, it is not possible to determine whether this vaccine approach protected these patients from recurrence. However, this pilot trial should provide evidence of antibody production and induced cell-mediated immunity to the carbohydrate moieties. This will allow us to decide whether a larger therapy trial, either to treat advanced disease or to prevent recurrences, is warranted.
This trial tests whether it is possible to elicit a cell-mediated immune response (considered important for tumor protection) to two of the most prevalent carbohydrate epitopes in human carcinomas, Tn and sTn. The role of carbohydrates in cell-mediated immunity has never been thoroughly studied, and this protocol provides an opportunity to analyze the cellular immune response of patients immunized with a xenogeneic mucin that expresses two common carbohydrate human tumor-associated antigens.
Results
When OSM was heated for 1 s in 0.1 N sulfuric acid and dialyzed against distilled water, greater than 99% of the sialic acid residues remained on OSM according to the periodate thio-barbituric acid test. This sialylated form of OSM was used as OSM. When OSM was heated for 4 min in 0.1 N sulfuric acid and dialyzed, 50% of the sialic acid residues remained on OSM according to the thio-barbituric acid method, and this was used as the partially desialylated form of OSM (containing a 50:50 mix of sTn and Tn). When OSM was heated for 2 h in 0.1 N sulfuric acid and dialyzed, 99% of the sialic acid was removed from the OSM, as verified by the thio-barbituric acid method, and this was used as dOSM. The removal of sialic acid from OSM in the dOSM preparation was also confirmed by loss of reactivity with mAb B72.3 in an ELISA assay.
Fifteen of 17 patients analyzed by ELISA produced IgG antibody titers from 40 to 5120 times more reactive with OSM and dOSM postimmunization. Greater antibody titers were seen for dOSM compared to OSM for all vaccine groups, suggesting that it was more immunogenic than OSM. Time courses of some of the IgG antibody responses to OSM and dOSM are shown in Figures 1-4. These IgG antibodies were shown to react with bovine submaxillary mucin (BSM, containing sTn), desialylated bovine submaxillary mucin (dBSM, containing Tn), Tn-human serum albumin (Tn-HSA), and sTn-human serum albumin (sTn-HSA, obtained from Biomira, Inc., Edmonton, Canada) in addition to OSM and dOSM (Table 1). More importantly, whole-cell ELISA showed these IgG antibodies reacted with LS-174T, with a mucinous human colon carcinoma cell line containing MUC2, and with sTn and Tn, with antibodies titers up to 640 times greater than preimmunization. No reactivity was seen with the Raji human B-cell lymphoma line used as a negative control. Desialylation reduced antibody titers to OSM to some degree, but did not alter production of IgG antibodies to sTn and Tn as significantly as expected. Immune stain dot blots confirmed the IgG results but showed greater reactivity with OSM than with dOSM for all vaccine groups (Table 2). This suggests that there is cross-reactivity between antibodies to sTn and Tn.
Pre- and post-IgG immune blot results for some antigens
Cellular immunity was quantified by measuring DTH-like reactions to OSM, dOSM, BSM, and dBSM (Table 3). Positive DTH-like reactions to OSM were punch-biopsied, and hematoxylin and eosion (H and E) staining and fresh-frozen immunohistochemistry using anti-T-cell antibodies were performed. Significant erythematous, nonindurated DTH-like reactions (1-17 cm) were observed in 15 of 17 immunized patients (Figure 5). At least three immunizations were required before significant cellular immune responses were observed. The occurrence of DTH-like responses correlated with the development of high-titer IgG responses. If patients were immunized with OSM, their DTH responses were greater to OSM than to dOSM; if patients were immunized with dOSM, their DTH responses were greater to dOSM than to OSM. The DTH-like reactions were also elicited by BSM and dBSM in some cases. BSM presumably contains a different protein backbone, but also expresses sTn. These results suggest that the DTH-like reactions were at least in part carbohydrate-dependent. The presence or absence of sialic acid could turn the DTH-like reactions on or off, partially or completely.
Erythematous, nonindurated DTH-like reactions 2 wk after the fourth vaccine dose, on the arm of a patient in the dOSM vaccine group.
Biopsies of positive DTH-like reactions revealed infiltration with mast cells (identified on H and E stains) and some CD8+ T lymphocytes. Lymphocyte proliferation assays yielded entirely negative results. Presumably, the OSM and dOSM added to stimulate the lymphocytes inhibited the activation and proliferation of lymphocytes in these assays, since the addition of these mucins was found to inhibit the control PHA-activation and proliferation of lymphocytes isolated from these patients, as well as from normal controls (Figure 6). Inhibition of lymphocyte proliferation could be due to carbohydrate epitopes present on the mucins (42) in a manner similar to ganglioside inhibition of lymphocyte activity (43), or it could be due to the protein apomucin portion, as was thought to be the case with MUC1 (44, 45). The fact that dOSM inhibited lymphocyte function even more than OSM is peculiar, since the DTH-like reactions were greater for dOSM. This suggests that the DTH-like reactions may be mediated by something other than T cells, such as antibodies.
Lymphocyte proliferation (measured by tritiated thymidine uptake) for a normal donor with PHA-stimulation, and inhibition of PHA-stimulation with OSM and dOSM.
This data suggests an important role for sialic acid in eliciting DTH-like reactions to OSM and dOSM. Toxicity included local inflammatory reactions at the sites of vaccine injection. There was no significant evidence of autoimmunity, except in one patient immunized with four injections of dOSM who developed generalized livedo reticularis, after which her vaccinations were stopped. Two patients, one of whom was the patient who developed livedo reticularis, developed transient numbness and tingling in their hands. The other patient who developed transient tingling and numbness in her hands was immunized with partially desialylated OSM (50/50 mix). These results suggest that OSM and dOSM are immunogenic and can be administered safely as a vaccine.
Discussion
Unlike an earlier study, this study found OSM to be very immunogenic when administered as one large, intradermal dose without cyclophosphamide and given with BCG (31). IgG antibody was induced to both sTn (OSM) and Tn (dOSM) in all three patient groups. The antibody reacted in many cases with a human mucinous colon carcinoma cell line, LS-174T, expressing Tn and sTn. The effect of desialylation of OSM reduced the humoral immune response to OSM and enhanced the antibody response to dOSM. Greater antibody titers were seen for dOSM than for OSM, suggesting that the presence of sialic acid might block antibody binding to sTn and Tn. The other possibility is that dOSM binds preferentially to the ELISA plates. Superior binding to sTn-HSA could be a reflection of epitope density, which is greater for OSM than for sTn-HSA. It may be that desialylation of OSM enables antibodies reactive with sTn and Tn to bind more easily to Tn and to residual sTn epitopes in dOSM. Enhanced affinity of these antibodies for dOSM may result from removal of the predominant negative charge or from steric hindrance blocking antibody binding to OSM.
Results of immune stain dot blots confirmed IgG antibody responses seen on ELISA, but differed in that greater reactivity was seen for OSM than for dOSM, and that there were no significant differences in specificity for OSM and dOSM among the different vaccine groups. The lack of specific variation in IgG antibody responses among the three vaccine groups, and the discordance of these results with in vivo immune responses seen for dOSM and dBSM, indicate a possible bias in favor of OSM over dOSM in the dot blot results. This possibility is supported by the fact that, as a result of extensive sialic acid residues, OSM is a highly negatively charged molecule, and this may confer increased affinity of OSM for the highly positively charged nitrocellulose, compared to dOSM.
The effect of desialylation on the specificity of IgG antibody produced was less than expected, suggesting cross-reactivity with both sTn and Tn. That the antibodies produced do not always distinguish between sTn and Tn indicates that the sialic portion of the carbohydrate may not be the most antigenic part of the carbohydrate epitope. This observation, although surprising for the specificity of antibodies, is supported by the fact that most anti-Tn and -sTn mAbs previously studied, including human mAbs (31), react with both Tn and sTn. That most of the antibodies produced were of the IgG class suggests that OSM, in the absence of any conjugation to an immunogenic carrier protein (such as keyhole limpet hemocyanin), was able to induce a strong enough cellular immune reaction to class switch these anticarbohydrate antibodies to IgG.
OSM, given intradermally with BCG, induced a cellular immune reaction, as demonstrated by a significant DTH-like response in most patients. Surprisingly, desialylation of OSM had a pronounced effect on the size and specificity of the DTH-like reaction. The strong DTH-like response to OSM, and especially to dOSM, does not confirm the presence of a strong T-cell mediated immune phenomena. Indeed, the absence of induration and CD4 lymphocytes and the presence of mast cells suggest that this may be an Arthus-like reaction mediated by antibodies. That the size and specificity of the cellular immune reactions correlated with IgG antibody titers and specificity for OSM and dOSM supports this idea. The association of greater antibody response to, and specificity for, dOSM with the size and specificity of the DTH-like reactions is consistent with cellular immune responses being Arthus-like reactions. That DTH-like reactions were elicited with xenogeneic BSM and dBSM, which contain sTn and Tn on a different peptide backbone, also supports a role for antibodies reactive with the carbohydrate epitopes in the etiology of the DTH-like reactions. Only a few CD8+ lymphocytes were present in the skin reactions, evidence that this is not a typical DTH response.
The strong immunogenicity of OSM probably results from its xenogenicity. The idea of using xenogeneic vaccines is not a new one, and it is now being applied to the development of DNA-based vaccines (46). OSM and dOSM are capable of inducing strong IgG antibody responses against some of the most important carbohydrate antigens implicated as tumor-associated antigens. It is not clear that human mucins expressing these same epitopes would be as immunogenic in patients. Recently, recombinant MUC1 has been incorporated into a BCG vector to test the idea of immunizing patients with syngeneic mucins (47). This approach may be optimal for inducing autologous cytolytic T lymphocytes against the apomucin of MUC1; however, it may be preferable to immunize with xenogeneic mucins in order to induce significant IgG antibody responses to shared carbohydrate epitopes.
If these Arthus-like reactions, dependent upon carbohydrate epitopes, could be duplicated at the site of tumors, then this vaccine would have strong therapeutic implications even in the absence of a T-cell response to the human apomucins in tumors. This would be important, since the sTn and Tn epitopes are widely distributed in human tumors and are present on many different peptide apomucins. It would also suggest a role for a vaccine that induces strong antibody responses, and it would allow for the use of xenogeneic carbohydrate-based vaccines, since the need for a peptide match would be absent and, perhaps, not desirable.
OSM, and particularly dOSM, inhibited PHA-stimulated lymphocyte proliferation and therefore may have an immunosuppressive effect, perhaps by interfering with cell-cell interactions (48). Inhibition of NK cell activity by mucins containing sTn was reported previously (42), and others have observed inhibition of lymphocyte proliferation by mucins (44) and induction of apoptosis in T cells after exposure to mucins (45). This immunosuppressive effect is similar to that described for other repeating, carbohydrate-containing molecules, such as gangliosides (43), which may account for the poor prognosis associated with some mucinous tumors, such as colloidal colon cancer (11, 12). It is a paradox that these mucins are capable of inducing IgG antibodies and, at the same time, appear to have an inhibitory function on T-cell function. This data, however, is in line with a similar, previously published report of apparent inverse relationships between DTH responses and IgG antibody production after vaccination with similar carbohydrate antigens in a syngeneic mouse mammary carcinoma model (28). It is possible that completely deglycosylated OSM might have stimulated lymphocytes and elicited true DTH responses, since the T-cell help for class switching the anticarbohydrate antibodies from IgM to IgG probably came from a cellular immune reaction against the peptide backbone of OSM.
Greater inhibition of lymphocyte function and stronger Arthus-like reactions were seen with dOSM than with OSM. This suggests that sialic acid plays a role in the specificity of the cell-mediated reactions. Arthus-like reactions with the employment of complement activation could potentially have stronger antitumor effects than autologous cytotoxic killer cells. They would be able to attack external epitopes on tumor cells and to destroy tumor cells through complement cytotoxicity and other cell-mediated processes, as well as antibody-dependent cellular cytotoxicity, due to the recruitment of cells through the production of anaphylatoxins. Evidence of a strong antitumor effect for dOSM was found in a syngeneic mouse mammary carcinoma model in which immunizations with dOSM led to protection against challenge from a virulent tumor cell line expressing Tn (28).
OSM was a superior inducer of antibodies reactive with sTn. Vaccines in which antibody responses might be sufficient for efficacy, such as vaccines for viruses, including papilloma virus and HIV, could be designed using xenogeneic molecules containing the appropriate carbohydrate epitopes. HIV has been shown to express sTn on both gp 120 and gp 160 (49). It is also interesting that sTn is expressed at a low level on a subpopulation of T cells (50). The action of anti-sTn antibodies on T cells in this study is unclear. It is not known if they would be activating or inhibitory. It is also not known if inhibition of a subset of CD4 cells by anti-sTn antibodies is responsible for the unexpectedly strong IgG antibody response to these carbohydrate antigens. A randomized cooperative group study comparing OSM vaccine to the Theratope® sTn-KLH vaccine (51) is warranted. BCG would be a more potent adjuvant than synthetic analogs.
1. Foley EJ. Antigenic properties of methylcholanthrene-induced tumors in mice of the strain of origin. Cancer Res 1953; 13: 835-7. (PMID: 13116120)
2. Hanna MG Jr, Hoover HC Jr, Vermorken JB, Harris JE, Pinedo HM. Adjuvant active specific immunotherapy of stage II and stage III colon cancer with an autologous tumor cell vaccine: first randomized phase III trials show promise. Vaccine 2001; 19: 2576-82. (PMID: 11257395)
3. O'Boyle KP, Cohen A. Rational design of generic vaccines for colorectal cancer immunotherapy. Perspect Colon Rectal Surg 1993; 6: 1-20.
4. Mitchell MS. Cancer vaccines, a critical review--Part II. Curr Opin Investig Drugs 2002; 3: 150-8. (PMID: 12054066)
5. Odunsi K, Jungbluth AA, Stockert E, Qian F, Gnjatic S, Tammela T, Intengan M, Beck A, Keitz B, Santiago D, Williamson B, Scanlan MJ, Ritter G, Chen YT, Driscoll D, Sood A, Lele S, Old LJ. NY-ESO-1 and LAGE-1 cancer-testis antigens are potential targets for immunotherapy in epithelial ovarian cancer. Cancer Res 2003; 63: 6076-83. (PMID: 14522938)
6. Levine P. Illegitimate blood group antigen P1, A, and MN (T) in malignancy: A possible therapeutic approach with anti-Tja, anti-A, and anti-T. Ann NY Acad Sci 1976; 277: 428-35. (PMID: 793484)
7. Springer GF. T and Tn, general carcinoma autoantigens. Science 1984; 224: 1198-206. (PMID: 6729450)
8. Thor A, Ohuchi N, Szpak CA, Johnston WW, Schlom J. Distribution of oncofetal antigen tumor-associated glycoprotein-72 defined by monoclonal antibody B72.3. Cancer Res 1986; 46: 3118-24. (PMID: 3516392)
9. Itzkowitz SH, Yuan M, Montgomery CK, Kjeldsen T, Takahashi TK, Bigbee WL, Kim YS. Expression of Tn, sialosyl-Tn, and T antigens in human colon cancer. Cancer Res 1989; 49: 197-204. (PMID: 2908846)
10. Cooper G, Codington JF, Miller DK, Brown MC. Loss of strain specificity of the TA3-St subline: evidence for the role of epiglycanin in mouse allogeneic tumor growth. J Natl Cancer Inst 1979; 63: 163-9. (PMID: 221714)
12. Halvorsen TB, Seim E. Influence of mucinous components on survival in colorectal adenocarcinomas: a multivariate analysis. J Clin Path 1988; 41: 1068-72. (PMID: 2848059)
13. Adsay NV, Merati K, Nassar H, Shia J, Sarkar F, Pierson CR, Cheng JD, Visscher DW, Hruban RH, Klimstra DS. Pathogenesis of colloid (pure mucinous) carcinoma of exocrine organs: Coupling of gel-forming mucin (MUC2) production with altered cell polarity and abnormal cell-stroma interaction may be the key factor in the morphogenesis and indolent behavior of colloid carcinoma in the breast and pancreas. Am J Surg Pathol 2003; 27: 571-8. (PMID: 12717243)
14. Blasco E, Torrado J, Belloso l, Arocena F, Gutierrez-Hoyos A, Cuadrado E. T-antigen. A prognostic indicator of high recurrence index in transitional carcinoma of the bladder. Cancer 1988; 61: 1091-5. (PMID: 3342368)
15. Springer GF. Tn epitope (N-acetyl-D-galactosamine alpha-O-serine/threonine) density in primary breast cancer: a functional predictor of aggressiveness. Mol Immunol 1989; 26: 1-5. (PMID: 2467192)
16. Itzkowitz SH, Bloom EJ, Kokal WA, Modin G, Hakomori S, Kim YS. Sialosyl-Tn. A novel mucin antigen associated with prognosis in colorectal cancer patients. Cancer 1990; 66: 1960-6. (PMID: 2224793)
17. Bray J, Maclean GD, Dusel FJ, McPherson TA. Decreased levels of circulating lytic anti-T in the serum of patients with metastatic gastrointestinal cancer: a correlation with disease burden. Clin Exp Immunol 1982; 47: 176-82. (PMID: 7094422)
18. Gendler S, Lancaster C, Taylor-Papadimitriou J, Duhig T, Peat N, Burchell J, Pemberton L, Lalani EN, Wilson D. Molecular cloning of human tumor-associated polymorphic epithelial mucin. J Biol Chem 1990; 265: 15286-93. (PMID: 1697589)
19. Lan MS, Batra SK, Qi WN, Metzgar RS, Hollingsworth MA. Cloning and sequencing of a human pancreatic tumor mucin cDNA. J Biol Chem 1990; 265: 15294-99. (PMID: 2394722)
20. Gum JR, Byrd JC, Hicks JW, Toribara NW, Lamport DT, Kim YS. Molecular cloning of human intestinal mucin cDNAs. J Biol Chem 1989; 264: 6480-7. (PMID: 2703501)
21. Porchet N, Van Cong G, Duffosse J, Audie, JP, Guyonnet-Duperat V, Gross MS, Denis C, Degand P, Bernheim A, Aubert JP. Molecular cloning and chromosomal localization of a novel human tracheo-bronchial mucin cDNA containing tandemly repeated sequences of 48 base pairs. Biochem Biophys Res Commun 1991; 175: 414-22. (PMID: 1673336)
22. Leroy X, Zerimech F, Zini L, Copin MC, Buisine MP, Gosselin B, Aubert JP, Porchet N. MUC1 expression is correlated with nuclear grade and tumor progression in pT1 renal clear cell carcinoma. Am J Clin Pathol 2002; 118: 47-51. (PMID: 12109855)
23. Boonla C, Wongkham S, Sheehan JK, Wongkham C, Bhudhisawasdi V, Tepsiri N, Pairojkul C. Prognostic value of serum MUC5AC mucin in patients with cholangiocarcinoma. Cancer 2003; 98: 1438-43. (PMID: 14508831)
24. Magro G, Schiappacassi M, Perissinotto D, Corsaro A, Borghese C, Belfiore A, Colombatti A, Grasso S, Botti C, Bombardieri E, Perris R. Differential expression of mucins 1-6 in papillary thyroid carcinoma: evidence for transformation-dependent post-translational modifications of MUC1 in situ. J Pathol 2003; 200: 357-69. (PMID: 12845632)
25. Okegawa T, Kinjo M, Horie S, Nutahara K, Higashihara E. Detection of mucin 7 gene expression in exfoliated cells in urine from patients with bladder tumor. Urology 2003; 62: 182-6. (PMID: 12837463)
26. Barnd DL, Lan MS, Metzgar RS, Finn OJ. Specific MHC-unrestricted recognition of tumor associated mucins by human cytotoxic T cells. Proc Natl Acad Sci U S A 1989; 86: 7159-63. (PMID: 2674949)
27. Ciborowski P, Finn OJ. Recombinant epithelial cell mucin (MUC-1) expressed in baculovirus resembles antigenically tumor associated mucin, target for cancer immunotherapy. Biomed Pept Proteins Nucleic Acids 1995; 1: 193-8. (PMID: 9346852)
28. Singhal A, Fohn M, Hakomori S. Induction of alpha-N-acetylgalactosamine-O-serine/threonine (Tn) antigen-mediated cellular immune response for active immunotherapy in mice. Cancer Res 1991; 51: 1406-11. (PMID: 1997178)
29. Piera M, De Bolos C, Castro R, Real FX. Cytokines as adjuvants: effect on the immunogenicity of NeuAc alpha 2-6GalNAc alpha-O-Ser/Thr (sialyl-Tn). Int J Cancer 1993; 55: 148-52. (PMID: 8344745)
30. O'Boyle KP, Zamore R, Cohen A, Kemeny N, Welt S, Lloyd KO, Oettgen HF, Old LJ, Livingston PO. Immunization of colorectal cancer patients with modified ovine submaxillary gland mucin and adjuvants induces IgM and IgG antibodies to sialylated Tn. Cancer Res 1992; 52: 5663-7. (PMID: 1394190)
31. O'Boyle KP, Wright K. Anti-Tn human monoclonal antibodies generated following active immunization with partially desialylated ovine submaxillary mucin. Hum Antibodies Hybridomas 1994; 5: 25-31. (PMID: 7532023)
32. Harada Y, Sakatsume M, Nores GA, Hakomori S, Taniguchi M. Density of GM3 with normal primary structure determines mouse melanoma antigenicity; a new concept of tumor antigen. Jpn J Cancer Res 1989; 80: 988-92. (PMID: 2533193)
33. Bellamy A, Davidson AN, Feldman M. Derivation of ganglioside-specific T cell lines of suppressor or helper phenotype from cerebrospinal fluid from multiple sclerosis patients. J Neuroimmunol 1986; 12: 107-20. (PMID: 2426301)
34. Ishioka GY, Lamont AG, Thomson D, Bulbow N, Gaeta FCA, Sette A, Grey HM. MHC interaction and T cell recognition of carbohydrates and glycopeptides. J Immunol 1992; 148: 2446-51. (PMID: 1560200)
35. Harding CV, Kihlberg J, Elofsson M, Magnusson G, Unanue ER. Glycopeptides bind MHC molecules and elicit specific T cell responses. J Immunol 1993; 151: 2419-25. (PMID: 8360471)
36. Haurum JS, Arsequell G, Lellouch AC, Wong YC, Dwek RA, McMichael AJ, Elliot T. Recognition of carbohydrate by major histocompatibility complex class I-restricted, glycopeptide-specific cytotoxic T lymphocytes. J Exp Med 1994; 80: 739-44. (PMID: 8046349)
37. Wu DY, Segal NH, Sidobre S, Kronenberg M, Chapman PB. Cross-presentation of disialoganglioside GD3 to natural killer T cells. J Exp Med 2003; 198: 173-81. (PMID: 12847141)
38. O'Boyle KP, Markowitz AL, Khorshidi M, Lalezari P, Longenecker BM, Lloyd KO, Welt S, Wright KE. Specificity analysis of murine monoclonal antibodies reactive with Tn, sialylated Tn, T, and monosialylated (2-6) T antigens. Hybridoma 1996; 15: 401-8. (PMID: 8985750)
39. Van Rinsum J, Smets LA, Van Roov, Van Den Eijnden DH. Specific inhibition of human natural killer cell-mediated cytotoxicity by sialic acid and sialo-oligosaccharides. Int J Cancer 1986; 38: 915-22. (PMID: 3793265)
40. Oratz R, Dugan M, Roses DF, Harris MN, Speyer JL, Hochster H, Weissman J, Henn M, Bystryn JC. Lack of activity of cyclophosphamide on the immunogenicity of a melanoma vaccine. Cancer Res 1991; 51: 3643-7. (PMID: 2065322)
41. Yonezawa S, Tachikawa T, Sadahito S, Sato E. Sialosyl-Tn antigen. Its distribution in normal tissues and expression in adenocarcinomas. Am J Clin Pathol 1992; 98: 167-74. (PMID: 1510031)
42. Ogata S, Maimonis PJ, Itzkowitz SH. Mucins bearing the cancer-associated sialosyl-Tn antigen mediate inhibition of natural killer cell cytotoxicity. Cancer Res 1992; 52: 4741-6. (PMID: 1511439)
43. McKallip R, Li R, Ladisch S. Tumor gangliosides inhibit the tumor-specific immune response. J Immunol 1999; 163: 3718-26. (PMID: 10490967)
44. Agrawal B, Krantz MJ, Reddish MA, Longenecker BM. Cancer-associated MUC1 mucin inhibits human T-cell proliferation, which is reversible by IL-2. Nat Med 1998; 4: 43-9. (PMID: 9427605)
45. Gimmi CD, Morrison BW, Mainprice BA, Gribben JG, Boussiotis VA, Freeman GJ, Park SY, Watanabe M, Gong J, Hayes DF, Kufe DW, Nadler LM. Breast cancer-associated antigen, DF3/MUC1, induces apoptosis of activated human T cells. Nat Med 1996; 2: 1367-70. (PMID: 8946837)
46. Wolchok JD, Gregor PD, Nordquist LT, Slovin SF, Scher HI. DNA vaccines: an active immunization strategy for prostate cancer. Semin Oncol 2003; 30: 659-66. (PMID: 14571413)
47. Chung MA, Luo Y, O'Donnell M, Rodriguez C, Heber W, Sharma S, Chang HR. Development and preclinical evaluation of a Bacillus Calmette-Guerin-MUC1-based novel breast cancer vaccine. Cancer Res 2003; 63: 1280-7. (PMID: 12649188)
48. O’Boyle KP, Chen T, Kozlowski S. Mucin inhibition of lymphocyte function does not require specific mucin-ligand interactions. Scand J Immunol 2000; 52: 46-52. (PMID: 10886783)
49. Hansen JES, Clausen H, Hu SL, Nielsen JO, Olofsson S. An O-linked carbohydrate neutralization epitope of HIV-1 gp 120 is expressed by HIV-1 env gene recombinant vaccinia virus. Arch Virol 1992; 126: 11-20. (PMID: 1381907)
50. Muroi K, Suda T, Nakamura M, Okada S, Nojiri H, Amemiya Y, Miura Y, Hakomori S. Expression of sialosyl-Tn in colony-forming unit-erythroid, erythroblasts, B cells, and a subset of CD4+ cells. Blood 1994; 83: 84-91. (PMID: 7903875)
51. Miles D, Papazisis K. Rationale for the clinical development of STn-KLH (Theratope) and anti-MUC-1 vaccines in breast cancer. Clin Breast Cancer 2003; 3 Suppl 4: S134-8. (PMID: 12620150)
52. Warren L. The thiobarbituric acid assay of sialic acids. J Biol Chem 1959; 234: 1971-5. (PMID: 13672998)
Materials
and methods
Pretreatment evaluation
Patients were entered into this study if they had resectable epithelial carcinomas, were at significant risk of recurrence, and were free of disease after surgery. They were monitored for recurrence by their private physicians. None recurred during the immunization process. Due to the heterogeneity of the vaccine, the different types of carcinoma, the variability in clinical stage, and the small number of patients, it was not possible to assess the clinical efficacy of the vaccine in this study. Patients had a diagnosis of epithelial carcinoma and had either been operated on at Montefiore or Weiler hospital, or reviewed by the Montefiore or Weiler hospital pathology department (NY, USA). All patients signed an informed consent form approved by the internal review boards. A thorough physical examination was performed at either Montefiore or Weiler hospital, and chest x-ray, serum creatinine, and liver function tests were performed within 4 wk of treatment. Patients who were pregnant or positive for HIV were excluded from this study. Patients with abnormal liver function tests or chest x-ray results were accepted if further tests (for example, computed tomography scans) showed no evidence of recurrent carcinoma.
Vaccine preparation
Purified OSM was purchased from Accurate Biochem, Inc. (Westbury, NY, USA). It contains 50% protein (by weight) and 50% carbohydrates that are divided among sTn (94%), Tn (4%), and other related carbohydrate chains (2%). To increase the nonsialylated Tn component of this sialomucin, it was heated in 0.1 N sulfuric acid at 80°C for different time intervals (1 s, 4 min, and 2 h). Samples of the different OSM vaccine preparations were dialyzed against distilled water to separate out the free sialic acid groups, and subsequently concentrated and reheated with 1 N hydrochloric acid at 100°C to release the remaining attached sialic acid residues. The residual amount of sialylation on the OSM samples was quantified using the periodate thio-barbituric acid procedure (52). The remainder of the vaccine preparations were then sterilely dialyzed against distilled water, aliquoted, and stored at -20°C. The OSM and desialylated forms of OSM used for this protocol were tested for sterility by standard culture techniques in the bacteriology laboratory, for pyrogenicity in rabbits, and for safety in mice and rats. On the first day of immunization, 200 µg of OSM, dOSM, or partially desialylated OSM (a 50/50 mix of sTn and Tn) was mixed with 107 units of BCG (Organon Teknika, Rockville, MD, USA) and diluted to a total volume of 0.25 ml. On subsequent days of immunization, the dose of BCG was reduced by a factor of 3 if there was significant toxicity. These studies were approved by the hospitals' internal review boards.
Vaccinations
The first 3 immunizations were given at 2-wk intervals beginning at least 4 wk after surgical resection. Each vaccine dose consisted of 0.25 ml sterile normal saline containing 200 µg of OSM (containing decreasing amounts of sTn and increasing amounts of Tn, depending upon the amount of desialylation) plus BCG (starting at 107 units and reduced each time by a factor of 3 for local toxicity) injected intradermally into one site on either thigh. Two additional booster vaccinations were administered at 6-wk intervals after the third vaccination. A total of 17 patients were immunized.
Serological responses
Peripheral blood (30 ml) was drawn immediately before each vaccination and every 2 wk thereafter until 6 wk after the fifth immunization; then every 6 wk for 1 yr. Patients' sera were tested by ELISA and dot blot immune stains for antibodies reactive with OSM, dOSM, BSM (also containing a natural source of sTn), dBSM (also containing a natural source of Tn), sTn-HSA, and Tn-HSA (the latter two reagents were generously provided by Dr. B. M. Longenecker of Biomira, Edmonton, Canada) to establish the specificity of the humoral response. ELISAs were performed using alkaline phosphatase conjugated goat antihuman IgM and IgG secondary antibodies (Kirkegaard and Perry laboratories, Gaithersburg, MD, USA) as previously described (30). The antibody titer of serum was defined as the reciprocal of the highest dilution giving an ELISA absorbance greater than 0.1 units. We concentrated on measuring IgG titers in this study because our previous report dealt mostly with IgM antibody production. In this study, stronger IgG antibody and lower IgM titers were noted, consistent with a class-switching T-cell help stimulus. Postimmunization titers were compared with preimmunization titers, and a time course of the humoral immune response was plotted against the immunization schedule. The IgG preimmunization titers were zero in all cases, but were arbitrarily placed at "1" for comparison. Dot blot immune stains were performed as previously described (30) with 1 µg of each antigen dried on nitrocellulose strips, incubated with serum diluted 1:10 for 1 h, and, after washing, incubated with peroxidase-conjugated goat antihuman IgM and IgG purchased from TAGO (Burlingame, CA, USA). Dot blots were read and assigned the following numbers: 1 (weak staining); 2 (moderate staining); 3 (strong staining); and 4 (very strong staining). Patients with titers of 40 or higher by ELISA and positive dot blot immune stains on a panel of related antigens were considered to be serologic responders. Whole-cell ELISA was performed on high-titer sera to test for reactivity with human cancer cell lines expressing sTn and Tn, such as the colon carcinoma cell line LS-174T, as previously described (32). Raji, a human B-cell lymphoma cell line, was used as a negative control.
DTH-like reactions
Skin tests for DTH-like reactions against Tn and sTn were performed at the time of the first immunization and 2 wk after the third, fourth, and fifth immunizations. Antigens administered were 200 µg of OSM (predominantly sTn), dOSM (predominantly Tn), BSM, and dBSM. BSM and dBSM provided an alternative source of animal mucins expressing sTn and Tn. Skin tests were performed intradermally, similarly to the vaccinations, except they did not include BCG and they were injected into the upper arms instead of the thighs. The DTH-like responses were read 48 h after injection. Punch skin biopsies were performed at least once for all patients with positive DTH-like responses. These skin biopsies were frozen in cryomolds and stored at -80°C until stained for immunohistochemistry as previously described (38) with mAbs against T-cell antigens, including CD3, CD4, and CD8.
Lymphocyte proliferation assays
PBLs were purified by Ficoll-density gradient centrifugation using standard procedures from 30 ml of heparinized blood collected prior to vaccination and 2 wk after the third, fourth, and fifth vaccinations. Lymphocyte proliferation was quantified by placing 100,000 cells per well in 96-well plates (Costar, Cambridge, MA, USA) in RPMI 1640, 10% FCS, 1 mmol glutamine, 100 µg/ml penicillin-streptomycin (Gibco). Lymphocytes were stimulated with 200 µg/ml of each antigen and 100 µg/ml of PHA as a control. Wells were performed in triplicate, and tritiated thymidine (1 µCi/ml, with a specific activity of 7 µCi/mmol for thymidine) was added on day 6 and lymphocytes were harvested onto glass-fiber filters and counted in a liquid scintillation counter the following day. Proliferation results at least twice background were considered to be positive.
Contact
Address
correspondence to:
Kevin O’Boyle, MD
80-26 166th Street
Hillcrest, NY 11432
USA
Tel.: + 1 718 380-5461
Fax: + 1 718 969-5452
E-mail:
