Proc. Nat. Acad. Sci. USA Vol. 69, No. 6, pp. 1620-1624, June 1972 Fractionation of Antigen Reactive Cells on a Cellular Immunoadsorbent: Factors Determining Recognition of Antigens by T-Lymphocytes (mouse fibroblasts/cell receptors/cell culture) H. WEKERLE*, P. LONAIt, and M. FELDMAN Department of Cell Biology, The Weizmann Institute of Science, Rehovot, Israel Communicated by Gerald M. Edelman, April 10, 1972 ABSTRACT Monolayers of mouse fibroblasts were used as cellular immunoadsorbents to separate rat lymphocytes that recognize specific mouse histocompatibility antigens. Normal lymphocytes were incubated on fibroblasts of strain C3H/eb, and nonadherent cells were separated from adherent cells, then transferred for sensitization onto fresh monolayers of C3H. When tested on 5"Crlabeled target monolayers the nonadherent cells manifested significantly lower cytotoxicity than the adherent cells. However, the nonadherent cells could be sensitized against mouse fibroblasts of an unrelated H-2 phenotype (strain Balb/c). The immune specificity of the adherence was further demonstrated by a stepwise adsorption, which resulted in complete loss of the capacity to acquire specific lytic activity towards C3H antigens. Lymphocytes recognizing strain-specific antigens of the mouse were separated at 370 (not at 00 or 40); separation was inhibited after treatment with dinitrophenol, sodium azide, or trypsin. Prior treatment with neuraminidase rendered the lymphocytes suceptible to separation at 00. The diversity of antibodies is attributed to an underlying diversity among the precursors of antibody-producing cells. This conclusion could be inferred from (a) the specific binding of antigens to lymphocytes (1), (b) the elimination of such lymphocytes after binding to highly radioactive antigens (2), and (c) the specific adherence of antigen-binding cells to antigen-coated bead columns (3). Furthermore, by the use of similar approaches with anti-immunoglobulin antibody, it was shown that the cell receptors for antigen in these cases are most probably immunoglobulin molecules (4). In all these studies, the reacting cells represented lymphocytes derived from bone marrow (B-cells), i.e., the precursors of cells that would actually produce antibodies. No such specific binding, and hence no diversity, could be demonstrated in unstimulated thymus-derived lymphocytes (T-cells), which are involved in carrier recognition as helpers for antibody production (5) or as effectors in cell-mediated immunity (6). To test the diversity of cells participating in cellular immunity, we used an in vitro system developed in our laboratory, in which a primary graft reaction is induced in cell culture (7). Rat lymphocytes are cultured on mouse fibroblast monolayers. After interaction with the mouse fibroblasts, part of the lymphocytes transform to large blastoid cells, multiply, and acquire the capacity to lyse fibroblasts genetically identical with those of the sensitizing * Visiting scientist from the Max-Planck-Institut fur Immunbiologie, Freiburg, Germany. t Present address: Division of Immunology, Stanford University Medical Center, Stanford, Calif. 1620 monolayer. The lytic activity is measured by transfer of aliquots of the in vitro sensitized lymphocytes (after 4-5 days of cultivation) to "Cr-labeled target monolayers and determination of the release of radioactive chromium. The reaction is specific: rat lymphocytes cultured on sensitizing monolayers of a certain mouse histocompatibility phenotype would lyse target monolayers of different phenotypes to a much lesser extent than identical monolayers (8). Furthermore, it was demonstrated that the specificity of the reaction is determined by the antigens interacting with thymus-derived lymphocytes (9), and that the lytic process is brought about by direct contact between effector lymphocytes and target cells (10). No synthesis of immunoglobulin has been demonstrated (11). In a previous report, we described that after plating of lymphocytes on sensitizing monolayers a certain proportion of the lymphocytes adhere firmly to the fibroblasts. We demonstrated that fibroblast monolayers may be used as cellular immunoadsorbents for the separation of cells that recognize one antigenic complex from a nonimmunized lymphocyte population (12). In the present study, we tested whether, in fact, the initial adherence of these cells represents the specific binding of lymphocytes via their receptors to strain-specific antigens on mouse-fibroblasts, as well as the conditions that determine such binding. MATERIALS AND METHODS Animals. Female Lewis rats, 1- to 2-months old, were used as donors of lymph node cells. Cultures of mouse fibroblasts were prepared from C311/eb and Balb/c mice. Fibroblast Monolayers. The preparation of monolayers of mouse fibroblasts has been described in detail (8). Briefly, primary fibroblast monolayers were prepared by trypsinization of embryos at 16 days of gestation. The cultures were incubated in 100-mm plastic petri dishes (Falcon Plastics, Los Angeles) in Waymouth's medium enriched with 5% calf serum. From these cells, secondary cultures were prepared, either as sensitizing cultures (2 X 106 cells per 60-mm plastic dish) or as target cultures (0.75 X 10 cells per 35-mm plastic dish). All experimental monolayers were irradiated with 2000 R of x-rays. Target fibroblasts were labeled 5 days before use by incubation with Na25"CrO4, (100-300 mci/mg of Cr, Radiochemical Centre, Amersham, England) at a concentration of 2 uci per target plate in 1 ml of Waymouth's medium. The radioactive medium was replaced by fresh medium after 15 hr.
Proc. Nat. Acad. Sci. USA 69 (1972) PAT-Lvmiphocyte Fractionation 1621 Lymphocyte Cultures. Rats were killed by heart puncture under ether anesthesia. All superficial and mesenteric lymph nodes were excised and dispersed in loosely fitting homogenizers. The cells were washed twice in cold Eagle's medium, and resuspended in culture medium (85% Eagle's medium- 15% horse serum). Eosin dye exclusion tests usually indicated a proportion of about 90% viable cells. Equal numbers of 50-60 X 106 viable lymphocytes, in volumes of 2 ml each, were plated on sensitizing monolayers, which were later separated. Control cultures not subjected to separation contained half the cell number. Separation Procedure. The cultures to be separated were placed on a rotating shaker for 5 min. All nonadherent or loosely attached lymphocytes were removed by thorough pipetting of the monolayer surface. 4 ml of Eagle's medium was added to each plate to rinse off the remaining loose cells. The cultures containing the adherent lymphocytes were immediately replenished with 4 ml of culture medium per plate. About 15% of the original lymphocyte population remains in the adherent population after separation. The supernatant cells were counted, tested for viability, and plated on fresh sensitizing fibroblast monolayers (30 X 106 lymphocytes per plate). Adherent cells remained tightly bound for at least 48 hr of further incubation. Assay of Lymphocyte-Mediated Cytolysis. Target-cell destruction was estimated by the percentage of total 5"Cr released from target fibroblasts through the action of sensitized lymphocytes (8). We plated 3 X 106 viable lymphocytes on each target monolayer and determined the chromium release after 18-40 hr of incubation. The radioactive samples were measured in a well-type Nal crystal scintillation counter (Packard Instruments Co., Downers Grove, Ill.). Each experimental value represents the mean of three target cultures. Spontaneous release of 5"Cr was also measured in triplicate; depending on the duration of incubation, it varied between 8 and 24%. Neuraminidase Treatment. Neuraminidase of Vibrio cholerae (Behringwerke, Marburg, Germany) was added to lymphocytes suspended in Eagle's medium (108 lymphocytes per ml) to a concentration of 50 units/ml. The enzyme-containing suspensions were incubated at 370 for 15 min. After they were washed in fresh Eagle's medium, the lymphocytes were Hours FIG. 1. Minimal time required for specific adherence of normal rat lymphocytes to mouse fibroblasts. 40 X 106 Lymphocytes were plated oin fibroblast monolayers. The unattached lymphocytes were separated from the attached ones at various times of incubation. Percentage of large blastoid cells was determined after 5 days of culture, when aliquots of cells were transferrred to 51Crlabeled fibroblasts. Chromium release was measured after 40 hr of incubation. Percent Cr release: A--A, adherent; A/-A, supernatant. Percent blastoid cells: *--*, adherent; O--O, supernatant. suspended in culture medium, and cell number and viability were determined. Trypsinization of Lymphocytes. Lymphocytes (5 X 108) were suspended in Puck's trypsin solution (0.3% in Ca++-, Mlg++-free salt solution) and incubated at 37, ph 7, for 15 miil. The incubation was ended by washing and counting the lymphocytes in the culture medium. Metabolic Inhibitors. Dinitrophenol was dissolved in absolute ethanol, at a concentration of 0.1 M. 0.14-ml 2 C) m 't TABLE 1. Specificity of adsorption % Change Monolayers used Sensitizing Target relative for adsorption monolayer monolayer % Lysis ± SD* to control Control Balb/c Balb/c 31.6 ± 2.1(3) 100 Adherent Balb/c Balb/c Balb/c 46.9 i 0.2(3) 148 Supernatant Balb/c Balb/c Balb/c 4.9 i 0.5(3) 16 Supernatant, Balb/c C3H C3H 26.7 i 3.6(3) 114 Control C3H C3H 23.5 ± 2.7(3) 100 Supernatant (nonadherent) lymphocytes were collected from Balb/c fibroblast monolayers after 2 hr of incubation. Equal numbers of these cells (30 X 106) were transferred to fresh Balb/c and C3H sensitizing monolayers. Lymphocytes sensitized against Balb/c were transferred after 5 days to Balb/c target monolayers, whereas lymphocytes sensitized on C3H were tested on C3H target monolayers. The lytic activity of each group was compared to unseparated lymphocytes sensitized and tested on the respective mouse fibroblasts. Data are reported as specific 65Cr release + standard deviation. Target culture incubation: 38 hr. * Values of n, the number of samples, are given in parentheses.
1622 Immunology: Wekerle et al. Proc. Nat. Acad. Sci. USA 69 (1972) TABLE 2. Stepwise depletion and concentration of specifically reactive lymphocytes Separation Purifica- Adherent Supernatant quotion step fraction* fraction* tientt Unseparated control 23.3 + 5.5(3) 1 Transfer 26.5 ± 0.6(3) 16.7 ± 0.8(3) 1.58 2 Transfers 29.0 + 1.4(3) 12.5 + 1.0(3) 2.32 3 Transfers 48.1 ± 1.8(2) 2.9 ±4 0.1(3) 16.68 60 X 106 Lymphocytes were plated on fibroblast monolayers- At 2-hr intervals, the loosely bound cells were removed from the "adherent" series, to concentrate the specifically adherent cell population. The supernatant lymphocytes were serially transferred to further deplete reactive cells. * Specific 51Cr release ± standard deviation. Target cuilture incubation: 40 hr. Values of n are given in parentheses. t Separation quotient: % specific 65Cr release by "adherent" cultures % specific 65Cr release by "supernatant" cultures Volumes were added to lymphocyte suspensions of 14 ml, which were subsequently plated on sensitizing monolayers. Sodium azide was dissolved in distilled water, to make up a 1 M stock solution. Again, 0.14 ml of stock solution was added to 14 ml of lymphocyte suspension. RESULTS Basic experimental program Normal lymph-node lymphocytes from rats were plated on monolayers of mouse fibroblasts for a short time, usually 0.5-2 hr. The nonadherent (supernatant) lymphocytes were separated from the lymphocytes that were firmly adsorbed on the monolayers, and transferred to fresh monolayers. Both types of cultures were then further incubated to allow reactive lymphocytes to undergo sensitization, manifested by blastoid transformation and proliferation. On the fifth day of sensitization, the lymphoid cell population was harvested from the sensitizing monolayers and equal numbers of viable lymphocytes were transferred onto 5"Cr-labeled target monolayers to assay cytotoxic activity. Specific adherence of normal lymph-node cells to fibroblast monolayers The first set of experiments was designed to measure the minimal time required for rat lymphocytes to specifically adhere to mouse fibroblasts. Fig. 1 shows the results of an TABLE 3. experiment in which we incubated lymphocytes with monolayers of C3H fibroblasts and separated, at different time intervals, the nonadherent lymphocytes from those that bound to fibroblasts. An optimal separation effect was reached at 30 min of incubation. The population obtained from adherent lymphocytes was significantly richer in transformed blastoid cells, and produced twice the cytolytic effect as did an equal number of the initially nonadherent lymphocytes. The separation effect might be explained either on the basis of nonspecific adherence of cells necessary for the induction of sensitization, such as macrophages (13), or on the basis of specific binding of lymphocytes capable of recognizing the fibroblast antigens. To differentiate between these two possibilities, we tested the ability of the nonadherent lymphocytes to undergo sensitization against fibroblasts of an antigenic type different from those used for separation Table 1 shows the results of an experiment in which lymphocytes were separated after 2 hr of incubation on Balb/c fibroblasts. The nonadherent lymphocytes were then transferred to fresh sensitizing fibroblast cultures of either Balb/c or C3H/eb genotype. Controls of unseparated sensitizing cultures of both genotypes were prepared. We found that the supernatant lymphocytes separated on Balb/c cells were unable to undergo significant sensitization against Balbic fibroblasts. However, such lymphocytes were sensitized by C3H fibroblasts, and the extent of lysis obtained was similar to that produced by unseparated lymphocytes sensitized on C3H fibroblasts. These findings indicate that the separation effect results from immunospecific adherence of normal lymphocytes to fibroblast antigens, and not from a nonspecific loss of accessory cells. We then tested whether specifically reactive cells could be totally removed from a population of unsensitized lymphocytes, and whether the concentration of reactive cells could be further enriched after adsorption. By repeating the separation procedure three consecutive times, through transfer of the nonadhering cells to fresh monolayers for further separation, we were able to eliminate all the lymphocytes capable of becoming sensitized against the antigens of the adsorbing fibroblasts (Table 2). On the other hand, reactive lymphocytes were markedly enriched by repeated washing of the originally adherent populations, and gentle removal of the loosely attached cells at 2-hr intervals. Energy dependence of specific adherence Antigen recognition by lymphocytes derived from bone marrow (not from thymus gland; B-cells) takes place at tempera- Effect of temperature on specific adherence of lymphocytes Temperature, Separated Separation Expt. 0C after hr Unseparated control Adherent fraction Supernatant fraction quotient 1 37 2 25.6 ± 0.0(3) 46.6 ± 1.0(3) 22.8 ± 3.5(3) 2.05 0 34.8 ± 0.9(3) 24.0 ± 1.6(3) 23.6 ± 1.0(3) 1.01 2 4 2 35.0 42.5 40.0 1.05 6 34.6 34.5 1.00 20 48.8 45.8 0.96 Rat lymphocytes were incubated on mouse fibroblast monolayers at 37, 4, and 00 in 5%o C02-95% air (water-vapor saturated). After separation of nonadherent from adherent lymphocytes, all cultures were transferred to a 370 incubator. Data presented as in Table 2.
Proc. Nat. Acad. Sci. USA 69 (1972) TABLE 4. Inhibition of specific adherence by metabolic inhibitors Untreated Dinitrophenol NaN3 Control 14.8 9.9 5.7 Adherent 18.9 12.2 7.9 Supernatant 6.3 12.9 11.4 Separation quotient 3.00 0.95 0.76 Lymphocytes were cultured on fibroblasts for 1 hr in culture medium containing 1 mm dinitrophenol, or 10 mm sodium azide. After cell separation, all supernatant lymphocytes were washed in Eagle's medium before they were replated onto fresh monolayer. Plates containing adherent lymphocytes were also rinsed with Eagle's medium before being replenished with culture medium Data presented as in Table 2; target culture incubation: 24 hr tures of about 40 (3, 4). We performed experiments to investigate if the same is true for our thymus-dependent immune reaction mediated by T-cells. The separation procedure was attempted at 0 and 40, as compared to control cultures incubated at 370, which showed a marked separation effect. At 0 and 40 the supernatant cells developed cytotoxic activity as did the adherent lymphocytes after subsequent incubation under normal culture conditions (Table 3). These findings suggested that antigen-specific adherence of normal T-lymphocytes requires metabolic energy. This conclusion was confirmed by measurement of the separation effect at 37 in the presence of different metabolic inhibitors. 1 mm dinitrophenol and 10 mm sodium azide consistently prevented a separation effect (Table 4). Lymphocyte viability was not affected at these concentrations. Effect of neuraminidase on lymphocyte separation at 00 Neuraminidase of Vibrio cholerae exposes certain surface proteins of lymphoid and other cells, thus rendering them accessible to antibodies (14), or increasing their antigenicity (15). We tested whether prior treatment of lymph node cells might affect the specific adsorption, possibly by increasing the exposure of cell receptors. Table 5 records two experiments in which lymphocytes were treated with neuraminidase. These lymphocytes were incubated with C3H sensitizing T-Lymphocyte Fractionation 1623 monolayers for 1 hr at 00, and simultaneously with untreated lymphocytes incubated at 37 and 00. No significant separation effect was obtained when untreated lymphocytes were incubated with fibroblast monolayers at 00. In contrast, the neuraminidase-treated lymph node cells showed an adherence effect at 00, which was slightly greater than that of untreated lymphocytes incubated at 37. Effect of trypsin treatment of lymphocytes Certain protein structures on the surface of lymphocytes, e.g., histocompatibility antigens (16) or surface immunoglobulins (17), are sensitive to trypsin digestion. If such a surface protein plays an essential role in the specific adsorption process, then its removal by trypsinization from the lymphocyte surface should abolish specific separation. We treated lymph node lymphocytes with trypsin before plating them on C3H sensitizing monolayers. Table 6 shows that no separation effect was obtained after an incubation period of 1 hr, nor did the capacity to adhere specifically reappear after 24 hr. DISCUSSION Our results indicate that T-lymphocytes participating in cell-mediated immune lysis of mouse cells constitute a diverse population. The diversity is inferred from the capacity of mouse fibroblasts of a given H-2 phenotype to adsorb lymphocytes with specific reactivity towards that same phenotype, while cells reactive against unrelated H-2 phenotypes are retained in the supernatant. Specific binding of immunized mouse lymphocytes to allogeneic target cells was described by Brondz (18). We have recently used nonimmunized mouse lymphocytes in an allogeneic system similar to the xenogeneic one described in this paper, and obtained separation of H-2- reactive cells (19, 20). Unlike the adsorption of B-lymphocytes on antigen columns that takes place at 40, the immunoadsorption described here took place only at 37. The energy dependence of specific adherence might be related to surface migration and concentration of membrane receptors. Such energy-requiring processes have been demonstrated to occur with "cap formation" when lymphocyte surface immunoglobulins are treated with bivalent anti-immunoglobulin antibodies (21). Another explanation might be the exposure of partly hidden receptors. TABLE 5. Specific adherence at 00 after treatment with neuraminidase No enzymatic treatment Neuraminidase 370 00 00 Unseparated control 10.7 + 0.4(3) 10.1 +1 0.6(3) 13.7 i 0.6(3) Adherent fraction 23.9 i 0.9(3) 12.8 i 0.4(3) 19.6 i 1.3(3) Supernatant fraction 9.9 + 0.4(3) 8.0 + 0.7(3) 6.5 ± 1.0(3) Separation quotient 2.43 1.55 3.01 Unseparated control 13.7 i 2.4(3) 17.8 + 1.2(3) 15.3 i 2.3(3) Adherent fraction 20.0 4-2.3(3) 12.1 i4 2.5(3) 14.6(1) Supernatant fraction 14.2 + 2.1(3) 11.6 i4 0.8(3) 5.8 ± 1.0(3) Separation quotient 1.76 1.04 2.50 Lymphocytes were incubated with neuraminidase diluted with Eagle's medium (50 units/108 lymphocytes per ml) for 15 min at 37. After they were washed in Eagle's medium, the cells were resuspended in culture medium at 00 and plated onto cooled fibroblast monolayers. The cultures were incubated on ice in 5% CO2-95' air for 1 hr. After separation of the adherent from the nonadherent lymphocytes, all cultures were transferred to 370 incubators. Data presented as in Table 2; target culture incubation: 24 hr.
1624 Immunology: Wekerle et al. Proc. Nat. Acad. Sci. USA 69 (1972) TABLE 6. Inhibition of specific adherence by prior treatment of lymphocytes with trypsin Trypsinized Treatment coincubation Untreated 1 hr 1 hr 24 hr Unseparated control 11.2 ± 1.4(3) 19.5 i 2.5(3) Adherent fraction 35.0 16.4 24.4 ± 0.5(3) Supernatant fraction 8.2 i 2.2(3) 19.2 i 2.0(3) 26.1 + 0.6(3) Separation quotient 4.23 0.86 0.93 6 X 108 Lymphocytes were suspended in 20 ml of 0.3% trypsin solution at 370, ph 7, for 15 min. After they were washed, cells were ncubated at 370 on fibroblast monolayers, and separated after 1 hr and 24 hr. Data presented as in Table 2. The fact that prior treatment of lymphocytes with neuraminidase enables them to recognize the histocompatibility antigens, even at 00, strongly supports the second possibility. It would be of interest to test whether the binding of T-cells to other antigens, for example to carrier proteins of a haptencarrier immunogen, depends also on a temperature of 370; a result that would explain the failure to obtain such binding in previous studies in which only low temperatures were applied. Evidently, the experiments reported here do not indicate the chemical nature of the T-lymphocyte receptors for strainspecific antigens of mice. Protein structures appear to be critical components for the binding process, since trypsin treatment of lymphocytes prevented their specific adsorption. The molecular nature of T-cell receptors is unknown. Studies from other laboratories report a certain reducing effect of anti-l-chain antibodies on graft-versus-host reaction (22), mixed lymphocyte reaction (23), and on helper activity of carrier sensitive lymphocytes (24). We are unable to demonstrate any inhibitory influence of anti-l-chain (Fab'-fragment) on specific adherence in our system (25). We thank Mrs. Varda Segal and Mrs. Ahuva Kapon for skilled technical assistance. This work was supported by a grant from the National Institutes of Health, Agreement no. 06-005-1, under contract P1-480, and by a grant from the Leukemia Research Foundation, Chicago, Ill. 1. Naor, D., & Sulitzeanu, D. (1967) Nature 214, 687-688. 2. Ada, G. L. & Byrt, P. (1969) Nature 222, 1291-1292. 3. Wigzell, H. & MAkeld, 0. (1970) J. Exp. Med. 132, 110-126. 4. Walters, C. S. & Wigzell, H. (1970) J. Exp. Med. 132, 1233-1249.,5. Kunin, S., Shearer, G. M., Segal, S., Globerson, A. & Feldman, M. (1971) Cell. Immunol. 2, 229-238. 6. Lonai, P., Clark, W. R. & Feldman, M. (1971) Nature 229, 566-567. 7. Ginsburg, H. & Sachs, L. (1965) J. Cell Comp. Physiol. 66, 199-219. 8. Berke, G., Ax, W., Ginsburg, H. & Feldman, M. (1969) Immunology 16, 643-657. 9. Lonai, P. & Feldman, M. (1970) Transplantation 10, 372-381. 10. Cohen, I. R. & Feldman, M. (1970) Cell. Immunol. 1, 521-535. 11. Clark, W. R., Berke, G., Feldman, M. & Sarid, S. (1971) Immunochemistry 8, 487-498. 12. Lonai, P., Wekerle, H. & Feldman, M. (1972) Nature New Biol. 235, 235-236. 13. Lonai, P. & Feldman, M. (1971) Immunology 21, 861-867. 14. Schlesinger, M. & Gottesfeld, S. (1971) Transplant. Proc. 3,1151-1155. 15. Currie, G. A. & Bagshawe, K. D. (1968) Brit. J. Cancer 22, 843-853. 16. Wallach, D. F. H. (1972) Biochim. Biophys. Acta 265, 61-83. 17. Rabellino, E., Colon, S., Grey, H. M. & Unanue, E. R. (1971) J. Exp. Med. 133, 156-167. 18. Brondz, B. D. & Snegir6va, A. E. (1971) Immunology 20, 457-468. 19. Altman, A., Wekerle, H., Cohen, I. R. & Feldman, M., Is. J. Med. Sci., in press. 20. Eliraz, A., Lonai, P. & Feldman, M., Is. J. Med. Sci., in press. 21. Taylor, R. B., Duffus, W. P. H., Raff, M. C. & depetris, S. (1971) Nature New Biol. 233, 225-229. 22. Mason, S. & Warner, N. L. (1970) J. Immunol. 104, 762-765. 23. Greaves, M. F., Torrigiani, G. & Roitt, I. M. (1971) Clin. Exp. Immunol. 9, 313-328. 24. Lesley, J. F., Kettman, J. R. & Dutton, R. W. (1971) J. Exp. Med. 134, 618-629. 235. Wekerle, H., Lonai, P. & Feldman, M., Is. J. Med. Sci., in press.