Znt. J . Cancer: 18, 161-167 (1976)
DETECTION OF HUMAN TUMOR-ASSOCIATED ANTIGENS BY THE LEUKOCYTE MIGRATION IN AGAROSE ASSAY A. W. BODDIE,JR., Marshall M. URIST,Darwin 0. CHEE,E. Carmack HOLMES and Donald L. MORTON Surgical Services, Veterans Administration Hospital, Sepulveda, California 91343; and Division of Oncology, Department of Surgery, University of California, Los Angeles, California 90024, USA
SUMMARY
Twenty-three of 36 (64%) lung cancer patients, 19 of 36 (54%) melanoma patients and 18 of 27 (66%) sarcoma patients tested in the leukocyte migration in agarose assay against soluble extracts of histologically similar tumors showed signijicant inhibition of leukocyte migration. Reactivity to extracts of dissimilar tumors was low. Sera of only 1/13 (7%) lung cancer patients, 2/19 (I0%) melanoma patients and 7/21 (33%) sarcoma patients were inhibited by extracts of histologically dissimilar tumors. Only 7-97!, of cancer patients reacted to paired extracts of normal tissue fiom the tumor donors. An average of 13% of sera from normal controls reacted to tumor extracts. Stage of disease and mode of therapy appeared to have little eflect on overall reactivity in this assay, although the number of patients within the various categories was small for purposes of statistical analysis. The leukocyte migration in agarose assay shows a sensitivity and specificity to tumor-associated antigens comparable to that of the older capillary tube method in general use and may facilitateperformance of migration inhibition. This assay may not be useful as aprognostic test due to the lack of correlation with stage of disease and treatment modality. However, its high specificity and economical use of tumor antigen suggest applications in tumor antigen purification. The use of soluble tumor antigen preparations may make it possible to purify these antigens further to increase specificity and reactivity.
1976) support the usefulness of Clausen's assay for the study of human tumor-associated antigens. This report describes our experience with the assay for detecting antigens associated with a variety of human tumors including lung cancer, sarcomas and malignant melanomas. MATERIAL AND METHODS
Antigen sources and preparation
Lung cancer and sarcoma antigens were prepared from specimens obtained directly from surgery by a sterile technique. Lung cancer specimens used for antigen preparation encompassed a variety of histologies including squamous-cell carcinoma, adenocarcinoma and undifferentiated carcinoma of the lung. Sarcoma specimens included a variety of osseous and soft-tissue sarcomas. Soluble melanoma antigens were prepared from a human melanoma cell line specially adapted for growth in serum-free, chemically defined media. Paired extracts of normal tissue from the tumor donors were prepared to serve as partial controls for HLA activity and for other specificities found in crude antigen preparations. In the case of lung cancer and sarcomas, normal lung tissue and normal muscle tissue were used to control for both HLA and organ-associated antigens. In the case of melanoma, normal muscle tissue was used for control purposes because insufficient skin was available from the tumor donor. -
A
Preparation of antigens
Growing interest in the immunology of human neoplasia has necessitated the development of tests for easy identification of tumor-associated antigens and for the study of host responses to these antigens. The leukocyte migration inhibition assay of Sprborg and Bendixen (1967) has shown promise both for detecting tumor-associated antigens and for monitoring cell-mediated reactivity to them (Sprborg, 1968; Cochran et al., 1972, 1973, 1974). Clausen's recent modification of this assay, the leukocyte migration in agarose assay (Clausen, 1971; Astor et al., 1973) has further simplified the technique for detecting migration inhibition. Recent publications (Spitler, 1974; Boddie et al., 1974, 1975a, b ; Chee,
Soluble antigens were prepared according to the technique of Reisfeld et al. (1971). Finely minced fresh tumors and control tissues or pelleted suspensions of tissue culture cells were resuspended to 3 x lo7 viable cells in 3 M KCl at 4"C. The cell suspension was mixed for 16-24 h on a turning rotor at 4"C. The mixture was centrifuged at 136,000 x g for 60 min at 4" C. The resulting supernatant was dialyzed against 200 vol cold PBS with three changes of medium in 24 h. The dialysate was centrifuged at 136,000 Xg for 15 min. The protein concentration of Received: February 16, 1976, and in revised form
June 7, 1976.
162
BODDIE ET
the final supernatant was determined by the Biuret method. Fresh tissues used for antigen preparation were cultured to exclude bacterial antigen contamination. Tissue culture lines used for antigen extraction were regularly screened for PPLO and viral contamination. Soluble antigen extracts were screened in Ouchterlony gel diffusion against 35 antisera to common bacterial anti gens. Leukocyte migration inhibition
Leukocyte migration in agarose was performed according to the technique of Clausen (1971) as adapted to the study of human tumor antigens (Spitler, 1974; Boddie et al., 1974, 1975a). Since the technique has been described in previous publications, it will be discussed only briefly here. Buffy-coat leukocytes were separated from heparinized whole blood by gravity sedimentation, washed four times and resuspended to a final concentration of 2.2 x los viable WBCs in medium 199 with 10% horse serum, 66 IU penicillin per ml, 66 pg streptomycin per ml and 1.25 g NaHCO, per 1. Five pl of the test antigen, the normal tissue antigen and a saline control were added to separate test tubes, then mixed with 50 pI of cell suspension and incubated for h. After incubation, the contents of each tube were resuspended and pipetted into triplicate wells cut into a specially prepared agarose plate. The plates were placed in moist chambers in a 5 % COz 37" C incubator for 18 h. After 18 h, the plates were removed, the agarose was melted off and the cells were heat-fixed to the plastic surface of the plate. Migration diameters for triplicate samples were measured with a Vernier caliper, the corresponding areas determined by the formula for the area of a circle (A = nr2) and a mean area calculated. Migration indices were determined as follows: average area of migration cellsSantigen Migration index average area of migration cells+saline
AL.
nator were compared by Student's t-test with correction for unequal variance, and thus we considered migration indices of t 0 . 7 to show significant reactivity to the test antigen. Patient selection
A total of 125 patients with cancer were tested, including 36 with lung cancer, 36 with malignant melanoma, 27 with osseous and soft-tissue sarcomas and 26 with assorted tumors including cancers of breast, colon and head and neck. These patients include subjects reported in previous papers as well as additional subjects tested since publication. Normal controls consisted of volunteers of comparable age with no history of malignancy. No attempt was made to preselect patients for testing on the basis of stage of disease or prior therapy. Patients who were receiving immunotherapy or chemotherapy were included as well as patients who had only surgical therapy and others who were long-term survivors at the time of initial testing. A breakdown of patients by stage (where available) and treatment groups is given in Tables I and 11. RESULTS
The protein concentrations at which individual antigen preparations were active varied considerably, but generally fell in the range of 2-5 mg/ml. After mixture with leukocyte suspension, final antigen concentration varied from 181 pg/ml to 454 pg/ml. Only infrequently were unpurified antigen preparations active at lower concentrations. Several
TABLE I NUMBER OF LUNG CANCER, MELANOMA A N D SARCOMA PATIENTS POSITIVE TO ANTIGENS PREPARED FROM HISTOLOGICALLY SIMILAR TUMORS BY STAGE OF DISEASE Population
Migration indices of
DETECTION OF HUMAN TUMOR-ASSOCIATED ANTIGENS BY THE LEUKOCYTE MIGRATION IN AGAROSE ASSAY A. W. BODDIE,JR., Marshall M. URIST,Darwin 0. CHEE,E. Carmack HOLMES and Donald L. MORTON Surgical Services, Veterans Administration Hospital, Sepulveda, California 91343; and Division of Oncology, Department of Surgery, University of California, Los Angeles, California 90024, USA
SUMMARY
Twenty-three of 36 (64%) lung cancer patients, 19 of 36 (54%) melanoma patients and 18 of 27 (66%) sarcoma patients tested in the leukocyte migration in agarose assay against soluble extracts of histologically similar tumors showed signijicant inhibition of leukocyte migration. Reactivity to extracts of dissimilar tumors was low. Sera of only 1/13 (7%) lung cancer patients, 2/19 (I0%) melanoma patients and 7/21 (33%) sarcoma patients were inhibited by extracts of histologically dissimilar tumors. Only 7-97!, of cancer patients reacted to paired extracts of normal tissue fiom the tumor donors. An average of 13% of sera from normal controls reacted to tumor extracts. Stage of disease and mode of therapy appeared to have little eflect on overall reactivity in this assay, although the number of patients within the various categories was small for purposes of statistical analysis. The leukocyte migration in agarose assay shows a sensitivity and specificity to tumor-associated antigens comparable to that of the older capillary tube method in general use and may facilitateperformance of migration inhibition. This assay may not be useful as aprognostic test due to the lack of correlation with stage of disease and treatment modality. However, its high specificity and economical use of tumor antigen suggest applications in tumor antigen purification. The use of soluble tumor antigen preparations may make it possible to purify these antigens further to increase specificity and reactivity.
1976) support the usefulness of Clausen's assay for the study of human tumor-associated antigens. This report describes our experience with the assay for detecting antigens associated with a variety of human tumors including lung cancer, sarcomas and malignant melanomas. MATERIAL AND METHODS
Antigen sources and preparation
Lung cancer and sarcoma antigens were prepared from specimens obtained directly from surgery by a sterile technique. Lung cancer specimens used for antigen preparation encompassed a variety of histologies including squamous-cell carcinoma, adenocarcinoma and undifferentiated carcinoma of the lung. Sarcoma specimens included a variety of osseous and soft-tissue sarcomas. Soluble melanoma antigens were prepared from a human melanoma cell line specially adapted for growth in serum-free, chemically defined media. Paired extracts of normal tissue from the tumor donors were prepared to serve as partial controls for HLA activity and for other specificities found in crude antigen preparations. In the case of lung cancer and sarcomas, normal lung tissue and normal muscle tissue were used to control for both HLA and organ-associated antigens. In the case of melanoma, normal muscle tissue was used for control purposes because insufficient skin was available from the tumor donor. -
A
Preparation of antigens
Growing interest in the immunology of human neoplasia has necessitated the development of tests for easy identification of tumor-associated antigens and for the study of host responses to these antigens. The leukocyte migration inhibition assay of Sprborg and Bendixen (1967) has shown promise both for detecting tumor-associated antigens and for monitoring cell-mediated reactivity to them (Sprborg, 1968; Cochran et al., 1972, 1973, 1974). Clausen's recent modification of this assay, the leukocyte migration in agarose assay (Clausen, 1971; Astor et al., 1973) has further simplified the technique for detecting migration inhibition. Recent publications (Spitler, 1974; Boddie et al., 1974, 1975a, b ; Chee,
Soluble antigens were prepared according to the technique of Reisfeld et al. (1971). Finely minced fresh tumors and control tissues or pelleted suspensions of tissue culture cells were resuspended to 3 x lo7 viable cells in 3 M KCl at 4"C. The cell suspension was mixed for 16-24 h on a turning rotor at 4"C. The mixture was centrifuged at 136,000 x g for 60 min at 4" C. The resulting supernatant was dialyzed against 200 vol cold PBS with three changes of medium in 24 h. The dialysate was centrifuged at 136,000 Xg for 15 min. The protein concentration of Received: February 16, 1976, and in revised form
June 7, 1976.
162
BODDIE ET
the final supernatant was determined by the Biuret method. Fresh tissues used for antigen preparation were cultured to exclude bacterial antigen contamination. Tissue culture lines used for antigen extraction were regularly screened for PPLO and viral contamination. Soluble antigen extracts were screened in Ouchterlony gel diffusion against 35 antisera to common bacterial anti gens. Leukocyte migration inhibition
Leukocyte migration in agarose was performed according to the technique of Clausen (1971) as adapted to the study of human tumor antigens (Spitler, 1974; Boddie et al., 1974, 1975a). Since the technique has been described in previous publications, it will be discussed only briefly here. Buffy-coat leukocytes were separated from heparinized whole blood by gravity sedimentation, washed four times and resuspended to a final concentration of 2.2 x los viable WBCs in medium 199 with 10% horse serum, 66 IU penicillin per ml, 66 pg streptomycin per ml and 1.25 g NaHCO, per 1. Five pl of the test antigen, the normal tissue antigen and a saline control were added to separate test tubes, then mixed with 50 pI of cell suspension and incubated for h. After incubation, the contents of each tube were resuspended and pipetted into triplicate wells cut into a specially prepared agarose plate. The plates were placed in moist chambers in a 5 % COz 37" C incubator for 18 h. After 18 h, the plates were removed, the agarose was melted off and the cells were heat-fixed to the plastic surface of the plate. Migration diameters for triplicate samples were measured with a Vernier caliper, the corresponding areas determined by the formula for the area of a circle (A = nr2) and a mean area calculated. Migration indices were determined as follows: average area of migration cellsSantigen Migration index average area of migration cells+saline
AL.
nator were compared by Student's t-test with correction for unequal variance, and thus we considered migration indices of t 0 . 7 to show significant reactivity to the test antigen. Patient selection
A total of 125 patients with cancer were tested, including 36 with lung cancer, 36 with malignant melanoma, 27 with osseous and soft-tissue sarcomas and 26 with assorted tumors including cancers of breast, colon and head and neck. These patients include subjects reported in previous papers as well as additional subjects tested since publication. Normal controls consisted of volunteers of comparable age with no history of malignancy. No attempt was made to preselect patients for testing on the basis of stage of disease or prior therapy. Patients who were receiving immunotherapy or chemotherapy were included as well as patients who had only surgical therapy and others who were long-term survivors at the time of initial testing. A breakdown of patients by stage (where available) and treatment groups is given in Tables I and 11. RESULTS
The protein concentrations at which individual antigen preparations were active varied considerably, but generally fell in the range of 2-5 mg/ml. After mixture with leukocyte suspension, final antigen concentration varied from 181 pg/ml to 454 pg/ml. Only infrequently were unpurified antigen preparations active at lower concentrations. Several
TABLE I NUMBER OF LUNG CANCER, MELANOMA A N D SARCOMA PATIENTS POSITIVE TO ANTIGENS PREPARED FROM HISTOLOGICALLY SIMILAR TUMORS BY STAGE OF DISEASE Population
Migration indices of
