The Journal of Dermatology Vol. 19: 823-826, 1992
WS V-6
Melanoma-associated Antigen Synthesized In vitro for Active Specific Immunotherapy Kazuhito Hayashibe Abstract The immunogenicity of the antigen molecule is a prerequisite for active specificimmunotherapy for melanoma. Since most of the melanoma-associated antigens recognized by the murine immune system are known to be not immunogenic in man, a detection. and analysis system for melanoma-associated antigens is required to reflect in vivo immune responses in patients with melanoma. One of the promising approaches, an attempt -to develop human monoclonal antibodies from B lymphocytes of patients with melanoma, has met with limited success due to the difficulties of producing large amounts of antibodies and using them in immunochemical assays, because most of them belong to the IgM class and have low affinity. Our approach is to utilize the screening of a cDNA expression library constructed from mRNA extracted from cultured melanoma cells with antibodies from patients with melanoma. The cloned cDNA, designated as D-I, had 1029 bp and showed no significant homology with viral and mammalian sequences stored in GENETYX. cDNA D-I hybridized to a 2.0 kb mRNA species from 3 different cell lines of human melanoma, neuroblastoma, erythroleukemia, B lymphoid, and T lymphoid cells, but not from a renal carcinoma cell line, normal peripheral lymphocytes, or normal fibroblasts. The in vivo expression and distribution of mRNA related to cDNA D-I has been examined in tissue specimens by in situ hybridization and shown to be rather restricted on melanoma cells. The polypeptide antigen encoded by cDNA D-I may be a valuable immunogen for implementing active specific immunotherapy in patients with melanoma. Introduction Humoral immune response directed against human melanoma antigens in patients has been demonstrated by the detection of antibodies to certain melanoma surface components (1). Most of this type of research shows that sera from melanoma patients possess multi specific antigen reactivity (2). This ambiguity is partially caused by masking the specific reactivity of the antibodies with melanoma-associated antigens by the reactivity of other antibodies with non-specific antigens expressed by melanoma cells. In order to overcome this problem and identify the targets by human immunosurveillance against melanoma, we have screened an expression cDNA library in lambda gtll constructed from a cultured human melanoma cell line with sera from patients with melanoma. This approach offers an effective strategy for sysReprint requests to: Kazuhito Hayashibe, Department of Dermatology, Kobe University School of Medicine, Kusunoki-cho 7-5-1, Chuo-ku, Kobe 650,Japan.
tematically detecting polypeptide antigens of interest which can be uniformly expressed in and purified from E. coli bacteria. Several investigators have attempted this strategy to detect and isolate reactive proteins with autoantibodies in sera from patients with paraneoplastic cerebellar degeneration (3), systemic lupus erythematosus (4), Raynaud's phenomenon (5), and Epstein-Barr virus (EBV) infection (6). In addition to the above diseases, which distinctly provide autoantibodies such as antiPurkinje cell, anti-La, anti-centromere, and antiEBV antibodies, malignant melanoma is known to be a suitable tumor for investigating host-tumor antigen interactions (7) and may be one of the best candidates among tumors for application of this approach. For example, the ganglioside GM2 may be the best-studied melanoma antigen recognized by autoantibodies from patients with melanoma, as demonstrated by Tai (8) and Livingston (9). Furthermore, Estin (10) reported that a recombinant vaccinia virus, v-p97NY, which expresses the human melanoma-associated antigen p97, could induce humoral and cell-mediated immunity in mice and
824
Hayashibe
Fig. 1. In situ hybridization with digoxigenin-Iabeled D-l eDNA probe on parafonnaldehyde fixed, frozen section of melanoma lesion. After hybridization, the specimen was washed with 2 x sse and incubated with sheep anti-digoxigenin antibodies conjugated with alkaline phosphatase. Positive signals are observed in the cytoplasm of melanoma cells.
monkeys. This indicates that the recombinant proteins can be immunogenic in vivo despite the lack of glycosylation which often plays a key role in presenting immunogenicity of human melanoma cells. This fact suggests the possibility of the recombinant protein which has been obtained using our approach being used as an active immunotherapeutic reagent against melanoma (11). Results
Following blockage with immunoglobulins isolated from normal human sera, an expression cDNA library constructed from cultured human melanoma cells A375 at EcoR I site of lambda gtl l was screened with biotinylated immunoglobulins isolated from a serum pool of 20 patients with melanoma which had been absorbed by E. coli proteins to reduce a heavy background due to antibodies cross-reacting with E. coli components. The D-I clone was identified after several rounds of rescreening with sera from patients with melanoma and successfully subcloned from lambda gtl l into plasmid Bluescript, containing a 1029 bp insert at EcoR I site. This D-I clone was expressed in E. coli JM 109 with plasmid Bluescript construct, and the D-I fusion protein was purified from bacterial cell lysates by u'ltra-centrifugation. After purification, the D-I fusion protein was analyzed by SDS-PAGE and
Western blotting, which revealed a predominant 150,000 antigenic species representing the intact D-I fusion protein. To measure the reactivity of the D-I fusion protein with sera from patients with melanoma and from healthy donors, DDIA was performed by coating the plates with the murine anti-beta-galactosidase monoclonal antibody (10 pg/ml) to immobilize the D-I fusion protein. The plasmid Bluescript protein preparation (BS) was used for comparison. Individual human sera were 50-fold diluted before application. The reactivity of each serum with the D-I fusion protein was calculated as (a.D. 405 with the D-I minus a.D. 405 with B.S.) to eliminate serum antibody reactivities with the betagalactosidase component and/or E. coli proteins. The mean a.D. 405 ± SD of sera from 100 patients with melanoma and from 100 healthy donors was 0.085 ± 0.081 and 0.015 ± 0.024 for binding to the D-I fusion protein, respectively (II). In addition to this result, since it is clearly demonstrated by Western blotting that serum antibody specificities can hardly react with the betagalactosidase component of the the fusion protein, the result of a fivefold higher-mean value of sera from patients with melanoma implied that most anti-immunogenic melanoma antigen antibodies in sera bound to antigenic determinants expressed on
Melanoma-associated Antigen for Immunotherapy the recombinant D-I fusion protein. To analyze D-I related transcripts, polyadenylated RNAs from some human tumor cells, normal peripheral blood lymphocytes, and normal skin fibroblasts were tested by Northern blot using the D-I fragment as a probe. The D-I probe was shown to hybridize to a 2.0 kilobase mRNA species in the melanoma lines (A375, MeWo and 3S5), the neuroblastoma line (TE671), the erthroleukemia line (K562), the B-Iymphoid line (LG2), and the Tlymphoid line (Molt 4), but not in the renal carcinoma line (SK-RC-29), normal peripheral blood lymphocytes, or normal skin fibroblasts (11). Utilizing in situ hybridization methods, tissue distribution of mRNA corresponding to 0-1 cDNA was examined on paraformaldehyde fixed frozen specimens. As Figure I shows, digoxigenin-Iabeled D-I cDNA probe clearly hybridized with mRNA in the cytoplasm of melanoma cells. Keratinocytes, fibroblasts, and lymphocytes on the same specimen as controls could not positively hybridize with D-I cDNA (to be submitted). The above findings on D-I eDNA and its coding polypeptide indicate the D-I immunogenic protein to be a valuable reagent for developing immunotherapeutic approches to melanoma if additional research shows that this antigen has a restricted heterogeneity in melanoma lesions and limited distribution in normal tissues. Discussion
The requirement for in vitro assessment to predict anti-melanoma effects of immunological reagents in vivo is a recurring theme (12-15), and it seems clear that the lack of such information is proba:bly one of the main factors limiting progress in this area. Crude immunogens, such as melanoma cell extracts, often show significant clinical effects on the course ofpatients with melanoma. However, it is not clear which components, including the ratio with adjuvants, caused the biological events. Among several candidates, including protein and glycoprotein antigens recognized by murine monoclonal antibodies, a few carbohydrate antigens are favourably known to be immunogenic in patients (16). Although the effect of those carbohydrates is limited in the previous reports, it does indicate that human immune system possesses the repertoire for the recognition of melanoma-associated antigens. Our approach could be useful for both the screening of immunogenic antigens and the structural identification of antigen molecules. Similar attempts successfully showed that synthesized poly-
825
peptide antigens were identified by human T cell clones (17). Combined with further advances in studies of antigen presentation mechanisms and T cell and B cell activation, novel reagents for active specific immunotherapy will be constructed in the near future (18). Acknowledgments I thank Noritaka Ishii for technical assistance in in situ hybridization. This work was supported by Grants-in Aid No. 03857127 from the Ministry ofEducation, Science and Culture, japan and by the Shiseido Research Fund for Basic Medicine (1991).
References 1) Carey TE, Uoid KO, Takahashi T, Travassos L, Old Lj: Solubilization and partial characterization of the AU cell surface antigen of human malignant melanoma, Proc Natl Acad Sci USA, 76: 2898-2902, 1979. 2) Damato BE, Cambell AM, McGuire Bj, Lee WR, Foulds WS: B-Iymphoeytes from melanoma patients and normal individuals react with melanoma cells but also with irrelevant antigens, Br ] Cancer, 58: 182-185,1988. 3) Dropcho Ej, Chen Y-T,PosnerjB, Old Lj: Cloning of a brain protein identified by autoantibodies from a patient with paraneoplastic cerebellar degeneration, Proc Natl Acad Sci USA, 84: 4552-4556,1987. 4) Chambers jC, Keene jD: Isolation and analysis of cDNA clones expressing human lupus La antigen, Proc Natl Acad Sci USA, 82: 2115-2119, 1985. 5) Earnshaw WC, Sullivan KF, Machlin PS, et al: Molecular cloning of cDNA for CENP-B, the major human centromere autoantigen,] Cell Biol, 104: 817-829,1987. 6) Walls D, Perricaudet M, Gannon F: The analysis of EBV proteins which are antigenic in vivo, Nucl Acids Res, 16: 2859-2872, 1988. 7) Old Lj: Cancer immunology: the search for specificity-G.HA Clowes Memorial Lecture, Cancer Res, 41: 361-375, 1981. 8) Tai T, Paulson jC, Cahan LD, lrie RF: Ganglioside GM2 as a human tumor antigen (OFA-I-l), Proc Natl Acad Sci USA, 80: 5392-5396, 1983. 9) Livingston PO, Natali Ej, Calves Mj, Stockert E, Oettgen HF, Old Lj: Vaccines containing purified GM2 ganglioside elicit GM2 antibodies in melanoma patients, Proc Natl Acad Sci USA, 84: 2911-2915, 1987. 10) Estin CD, Stevenson US, Plowman GD, et al: Recombinant vaccinia virus vaccine against the human melanoma antigen p97 for use in immunnotherapy, Proc Natl Acad Sci USA, 85: 1052-1056, 1988. 11) Hayashibe K, Mishima Y, Ferrone S: Cloning and in vitro expression of a melanoma-associated antigen immunogenic in patients with melanoma,] Immunol,
826
Hayashibe
147: 1098-1104, 1991. 12) Ferrone S, Pellegrino MA: Antigens and antibodies in malignant melanoma, in Waters H (ed): Handbook of Cancer/Immunology, Vol. 3, Garland Publishing, New York, 1978, p 291. 13) GrafLHJr, Ferrone S: Human melanoma-associated antigens, in Litwin SD (ed): Human Immunogenetics. BasicPrinciples and ClinicalRelevance, Marcel Dekker, Inc., New York, 1989, p 643. 14) Anichini A, Mazzocchi A, Fossati G, Parmiani G: Cytotoxic T lymphocyte clones from peripheral blood and from tumor site detect intratumor heterogeneity of melanoma cells. Analysis of specificity and mechanism of interaction, ] Immunol, 142: 3692, 1989. 15) Topalian SL, Kasid A, Rosenberg SA: Immunoselection of human melanoma resistant to specific lysis by
autologous tumor-infiltrating lymphocytes. Possible mechanisms for immunotherapeutic failures, ] Immunol, 144: 4487, 1990. 16) Livingston PO, Ritter G, Srivastava P, et al: Characterization of IgG and IgM antibodies induced in melanoma patients by immunization with purified GM2 ganglioside, Cancer Res, 49: 7045, 1989. 17) Hersey P, MacDonald M, Werkman H: Western blot analysis of antigens on .melanoma cells recognized by cytotoxic T cells, ] Natl Cancer Inst, 80: 826-835; 1988. 18) Livingston PO: Carbohydrate antigen vaccines against melanoma and colon cancer, in Hersey P (ed): Proceedings of InternationalConferena on Biological Treatment of Melanoma and Other Cancers, Newcastle, 1990, p 121.
WS V-6
Melanoma-associated Antigen Synthesized In vitro for Active Specific Immunotherapy Kazuhito Hayashibe Abstract The immunogenicity of the antigen molecule is a prerequisite for active specificimmunotherapy for melanoma. Since most of the melanoma-associated antigens recognized by the murine immune system are known to be not immunogenic in man, a detection. and analysis system for melanoma-associated antigens is required to reflect in vivo immune responses in patients with melanoma. One of the promising approaches, an attempt -to develop human monoclonal antibodies from B lymphocytes of patients with melanoma, has met with limited success due to the difficulties of producing large amounts of antibodies and using them in immunochemical assays, because most of them belong to the IgM class and have low affinity. Our approach is to utilize the screening of a cDNA expression library constructed from mRNA extracted from cultured melanoma cells with antibodies from patients with melanoma. The cloned cDNA, designated as D-I, had 1029 bp and showed no significant homology with viral and mammalian sequences stored in GENETYX. cDNA D-I hybridized to a 2.0 kb mRNA species from 3 different cell lines of human melanoma, neuroblastoma, erythroleukemia, B lymphoid, and T lymphoid cells, but not from a renal carcinoma cell line, normal peripheral lymphocytes, or normal fibroblasts. The in vivo expression and distribution of mRNA related to cDNA D-I has been examined in tissue specimens by in situ hybridization and shown to be rather restricted on melanoma cells. The polypeptide antigen encoded by cDNA D-I may be a valuable immunogen for implementing active specific immunotherapy in patients with melanoma. Introduction Humoral immune response directed against human melanoma antigens in patients has been demonstrated by the detection of antibodies to certain melanoma surface components (1). Most of this type of research shows that sera from melanoma patients possess multi specific antigen reactivity (2). This ambiguity is partially caused by masking the specific reactivity of the antibodies with melanoma-associated antigens by the reactivity of other antibodies with non-specific antigens expressed by melanoma cells. In order to overcome this problem and identify the targets by human immunosurveillance against melanoma, we have screened an expression cDNA library in lambda gtll constructed from a cultured human melanoma cell line with sera from patients with melanoma. This approach offers an effective strategy for sysReprint requests to: Kazuhito Hayashibe, Department of Dermatology, Kobe University School of Medicine, Kusunoki-cho 7-5-1, Chuo-ku, Kobe 650,Japan.
tematically detecting polypeptide antigens of interest which can be uniformly expressed in and purified from E. coli bacteria. Several investigators have attempted this strategy to detect and isolate reactive proteins with autoantibodies in sera from patients with paraneoplastic cerebellar degeneration (3), systemic lupus erythematosus (4), Raynaud's phenomenon (5), and Epstein-Barr virus (EBV) infection (6). In addition to the above diseases, which distinctly provide autoantibodies such as antiPurkinje cell, anti-La, anti-centromere, and antiEBV antibodies, malignant melanoma is known to be a suitable tumor for investigating host-tumor antigen interactions (7) and may be one of the best candidates among tumors for application of this approach. For example, the ganglioside GM2 may be the best-studied melanoma antigen recognized by autoantibodies from patients with melanoma, as demonstrated by Tai (8) and Livingston (9). Furthermore, Estin (10) reported that a recombinant vaccinia virus, v-p97NY, which expresses the human melanoma-associated antigen p97, could induce humoral and cell-mediated immunity in mice and
824
Hayashibe
Fig. 1. In situ hybridization with digoxigenin-Iabeled D-l eDNA probe on parafonnaldehyde fixed, frozen section of melanoma lesion. After hybridization, the specimen was washed with 2 x sse and incubated with sheep anti-digoxigenin antibodies conjugated with alkaline phosphatase. Positive signals are observed in the cytoplasm of melanoma cells.
monkeys. This indicates that the recombinant proteins can be immunogenic in vivo despite the lack of glycosylation which often plays a key role in presenting immunogenicity of human melanoma cells. This fact suggests the possibility of the recombinant protein which has been obtained using our approach being used as an active immunotherapeutic reagent against melanoma (11). Results
Following blockage with immunoglobulins isolated from normal human sera, an expression cDNA library constructed from cultured human melanoma cells A375 at EcoR I site of lambda gtl l was screened with biotinylated immunoglobulins isolated from a serum pool of 20 patients with melanoma which had been absorbed by E. coli proteins to reduce a heavy background due to antibodies cross-reacting with E. coli components. The D-I clone was identified after several rounds of rescreening with sera from patients with melanoma and successfully subcloned from lambda gtl l into plasmid Bluescript, containing a 1029 bp insert at EcoR I site. This D-I clone was expressed in E. coli JM 109 with plasmid Bluescript construct, and the D-I fusion protein was purified from bacterial cell lysates by u'ltra-centrifugation. After purification, the D-I fusion protein was analyzed by SDS-PAGE and
Western blotting, which revealed a predominant 150,000 antigenic species representing the intact D-I fusion protein. To measure the reactivity of the D-I fusion protein with sera from patients with melanoma and from healthy donors, DDIA was performed by coating the plates with the murine anti-beta-galactosidase monoclonal antibody (10 pg/ml) to immobilize the D-I fusion protein. The plasmid Bluescript protein preparation (BS) was used for comparison. Individual human sera were 50-fold diluted before application. The reactivity of each serum with the D-I fusion protein was calculated as (a.D. 405 with the D-I minus a.D. 405 with B.S.) to eliminate serum antibody reactivities with the betagalactosidase component and/or E. coli proteins. The mean a.D. 405 ± SD of sera from 100 patients with melanoma and from 100 healthy donors was 0.085 ± 0.081 and 0.015 ± 0.024 for binding to the D-I fusion protein, respectively (II). In addition to this result, since it is clearly demonstrated by Western blotting that serum antibody specificities can hardly react with the betagalactosidase component of the the fusion protein, the result of a fivefold higher-mean value of sera from patients with melanoma implied that most anti-immunogenic melanoma antigen antibodies in sera bound to antigenic determinants expressed on
Melanoma-associated Antigen for Immunotherapy the recombinant D-I fusion protein. To analyze D-I related transcripts, polyadenylated RNAs from some human tumor cells, normal peripheral blood lymphocytes, and normal skin fibroblasts were tested by Northern blot using the D-I fragment as a probe. The D-I probe was shown to hybridize to a 2.0 kilobase mRNA species in the melanoma lines (A375, MeWo and 3S5), the neuroblastoma line (TE671), the erthroleukemia line (K562), the B-Iymphoid line (LG2), and the Tlymphoid line (Molt 4), but not in the renal carcinoma line (SK-RC-29), normal peripheral blood lymphocytes, or normal skin fibroblasts (11). Utilizing in situ hybridization methods, tissue distribution of mRNA corresponding to 0-1 cDNA was examined on paraformaldehyde fixed frozen specimens. As Figure I shows, digoxigenin-Iabeled D-I cDNA probe clearly hybridized with mRNA in the cytoplasm of melanoma cells. Keratinocytes, fibroblasts, and lymphocytes on the same specimen as controls could not positively hybridize with D-I cDNA (to be submitted). The above findings on D-I eDNA and its coding polypeptide indicate the D-I immunogenic protein to be a valuable reagent for developing immunotherapeutic approches to melanoma if additional research shows that this antigen has a restricted heterogeneity in melanoma lesions and limited distribution in normal tissues. Discussion
The requirement for in vitro assessment to predict anti-melanoma effects of immunological reagents in vivo is a recurring theme (12-15), and it seems clear that the lack of such information is proba:bly one of the main factors limiting progress in this area. Crude immunogens, such as melanoma cell extracts, often show significant clinical effects on the course ofpatients with melanoma. However, it is not clear which components, including the ratio with adjuvants, caused the biological events. Among several candidates, including protein and glycoprotein antigens recognized by murine monoclonal antibodies, a few carbohydrate antigens are favourably known to be immunogenic in patients (16). Although the effect of those carbohydrates is limited in the previous reports, it does indicate that human immune system possesses the repertoire for the recognition of melanoma-associated antigens. Our approach could be useful for both the screening of immunogenic antigens and the structural identification of antigen molecules. Similar attempts successfully showed that synthesized poly-
825
peptide antigens were identified by human T cell clones (17). Combined with further advances in studies of antigen presentation mechanisms and T cell and B cell activation, novel reagents for active specific immunotherapy will be constructed in the near future (18). Acknowledgments I thank Noritaka Ishii for technical assistance in in situ hybridization. This work was supported by Grants-in Aid No. 03857127 from the Ministry ofEducation, Science and Culture, japan and by the Shiseido Research Fund for Basic Medicine (1991).
References 1) Carey TE, Uoid KO, Takahashi T, Travassos L, Old Lj: Solubilization and partial characterization of the AU cell surface antigen of human malignant melanoma, Proc Natl Acad Sci USA, 76: 2898-2902, 1979. 2) Damato BE, Cambell AM, McGuire Bj, Lee WR, Foulds WS: B-Iymphoeytes from melanoma patients and normal individuals react with melanoma cells but also with irrelevant antigens, Br ] Cancer, 58: 182-185,1988. 3) Dropcho Ej, Chen Y-T,PosnerjB, Old Lj: Cloning of a brain protein identified by autoantibodies from a patient with paraneoplastic cerebellar degeneration, Proc Natl Acad Sci USA, 84: 4552-4556,1987. 4) Chambers jC, Keene jD: Isolation and analysis of cDNA clones expressing human lupus La antigen, Proc Natl Acad Sci USA, 82: 2115-2119, 1985. 5) Earnshaw WC, Sullivan KF, Machlin PS, et al: Molecular cloning of cDNA for CENP-B, the major human centromere autoantigen,] Cell Biol, 104: 817-829,1987. 6) Walls D, Perricaudet M, Gannon F: The analysis of EBV proteins which are antigenic in vivo, Nucl Acids Res, 16: 2859-2872, 1988. 7) Old Lj: Cancer immunology: the search for specificity-G.HA Clowes Memorial Lecture, Cancer Res, 41: 361-375, 1981. 8) Tai T, Paulson jC, Cahan LD, lrie RF: Ganglioside GM2 as a human tumor antigen (OFA-I-l), Proc Natl Acad Sci USA, 80: 5392-5396, 1983. 9) Livingston PO, Natali Ej, Calves Mj, Stockert E, Oettgen HF, Old Lj: Vaccines containing purified GM2 ganglioside elicit GM2 antibodies in melanoma patients, Proc Natl Acad Sci USA, 84: 2911-2915, 1987. 10) Estin CD, Stevenson US, Plowman GD, et al: Recombinant vaccinia virus vaccine against the human melanoma antigen p97 for use in immunnotherapy, Proc Natl Acad Sci USA, 85: 1052-1056, 1988. 11) Hayashibe K, Mishima Y, Ferrone S: Cloning and in vitro expression of a melanoma-associated antigen immunogenic in patients with melanoma,] Immunol,
826
Hayashibe
147: 1098-1104, 1991. 12) Ferrone S, Pellegrino MA: Antigens and antibodies in malignant melanoma, in Waters H (ed): Handbook of Cancer/Immunology, Vol. 3, Garland Publishing, New York, 1978, p 291. 13) GrafLHJr, Ferrone S: Human melanoma-associated antigens, in Litwin SD (ed): Human Immunogenetics. BasicPrinciples and ClinicalRelevance, Marcel Dekker, Inc., New York, 1989, p 643. 14) Anichini A, Mazzocchi A, Fossati G, Parmiani G: Cytotoxic T lymphocyte clones from peripheral blood and from tumor site detect intratumor heterogeneity of melanoma cells. Analysis of specificity and mechanism of interaction, ] Immunol, 142: 3692, 1989. 15) Topalian SL, Kasid A, Rosenberg SA: Immunoselection of human melanoma resistant to specific lysis by
autologous tumor-infiltrating lymphocytes. Possible mechanisms for immunotherapeutic failures, ] Immunol, 144: 4487, 1990. 16) Livingston PO, Ritter G, Srivastava P, et al: Characterization of IgG and IgM antibodies induced in melanoma patients by immunization with purified GM2 ganglioside, Cancer Res, 49: 7045, 1989. 17) Hersey P, MacDonald M, Werkman H: Western blot analysis of antigens on .melanoma cells recognized by cytotoxic T cells, ] Natl Cancer Inst, 80: 826-835; 1988. 18) Livingston PO: Carbohydrate antigen vaccines against melanoma and colon cancer, in Hersey P (ed): Proceedings of InternationalConferena on Biological Treatment of Melanoma and Other Cancers, Newcastle, 1990, p 121.
