Stimulation of different pathways of T-cell functions by syngeneic tumor cells and soluble membrane proteins

CELLULAR

1MMUNOLOGY

Stimulation

78, 13-22 (1983)

of Different Pathways of T-Cell Functions by Syngeneic Tumor Cells and Soluble Membrane Proteins’ M. BERTSCHMANN*

AND E. F. LUSCHER

Theodor Kocher Institute, University of Berne, CH-3000 Bern 9, Switzerland Received September 1. 1982; accepted December 21, 1982 Cells from the draining lymph nodes of DBA/2 mice bearing syngeneic intradermal P-815 tumor represent an excellent responding cell population for secondary stimulation in vitro, despite the virtual absence of response by spleen cells of the same animals. Cytotoxicity is a result of stimulation with intact mitomycin-treated tumor cells. Isolated tumor cell membranes, in the form of small vesicles, stimulated cytotoxicity to a very limited extent and inhibited the development of cytotoxic T lymphocytes over a wide range of concentrations. Membrane proteins solubilized with deoxycholate and papain had only a suppressive effect. Soluble proteins exerted their effect during the induction of cytotoxic T lymphocytes and were ineffective when present during the effector phase of cytotoxic T lymphocytes. The suppressive capacity was shown to reside in a cell population which was sensitive to treatment with monoclonal Lyt-2.1 antibody and complement but not with Lyt-1 .I antibody. This phenotype was compatible with a specific rather than a promiscuous type of suppressor effector cell.

INTRODUCTION Cancer cells developing in or transplanted into an organism induce and then modulate the host’s immune response in multiple ways. Phases of persistence, growth, regression, and progression may reflect these complex interactions between cancer cells and the host’s various populations of lymphoid and other cells. The final dominance of a tumor over its host, i.e., tumor progression despite a tumor-directed immune response, is still an essentially unexplained although most important phenomenon. Suppressor T cells (1, 2) have ken thought to explain the paradox of stimulation and untimely suppression of immune reactivity, as observed in many tumor models. In fact, both specific and nonspecific suppression have been demonstrated to occur during the development of various tumors, including P-815 (36). However, suppressor T cells are not exclusively stimulated by tumor cells but can be detected as normal constituents of every immune response.In some tumor models the highest activity of suppressor T cells is found even in animals which have successfully rejected their tumors (i’), and highly immunogenic tumors are reported to be most active in suppressor T-cell stimulation (8). For the P-8 15 tumor we and others have described suppression when crude membrane preparations were used for immunization of animals (9, 10). On the other side ’ This work was supported by the Swiss National Science Foundation, Grant 3.48679. * To whom all correspondenceshould be addressed:Theodor Kocher Institute, University of Beme, P.O. Box 99, CH-3000 Bern 9, Switzerland. 13 0008-8749183$3.00 Copyright D 1983 by Academic Res, Inc. All rights of reproduction in any form reserved.

14

BERTSCHMANN

AND LijSCHER

it appeared difficult in our hands to modulate the pattern of id3 tumor development by the adoptive transfer of suppressor cells, which are presumably present in tumorbearing animals (11, 20). In the present paper we demonstrate that by secondary stimulation in vitro with DOC/Papsolubilized membrane proteins, suppressor T cells of the Lyt-l-,2+ phenotype can be induced and that the suppressor pathway had priority over the killer T-cell pathway which predominates in the presence of intact P-8 15 cells. Subcellular components released from decaying tumor cells during the regression phase could possibly be involved in the final preference of the suppressor over the killer pathway. MATERIALS

AND METHODS

Mice. DBA/2 0 were purchased from Bomholtgard, Ry, Denmark, and used at the age of 12-16 weeks. Tumors. P-8 15 tumor cells of DBA/2 origin were banked in liquid nitrogen and cultivated in DMEM supplemented with 10% HS for P-8 1S/Be and 10% NCS for P815/SK. Depending on the experiment they were also collected from the peritoneal cavity of DBA/2 mice 1 week after the injection of lo7 cells. Secondary stimulation in vitro. DBA/2 animals were primed with lo6 P-815/SK per mouse by injecting the cells id into the right shaved flank, as described before ( 12). Briefly, in vivo- or in vitro-grown cells washed three times in HBSS were injected into anesthetized animals (50 pg of Nembutal/g of body wt, ip) under a stereoscopic microscope in a volume of 10 ~1. On Day 8, when tumors had grown to a size of 4-5 mm, cells of the draining (axillary) lymph node were separated,washed in HBSS, suspended in DMEM, and used as responder cells, generally at a constant number of 2 X 105/well of a flat-bottomed microtiter plate (Nunc Denmark, Cat. No. l67008). Cells of the spleen and pooled cells of distal nodes served as negative controls, as did lymph node cells from unprimed animals. P-8 15 cells treated with 50 &ml of mitomycin C (Kyowa Hakko Kogyo Co, Ltd, Tokyo, Japan) for 1 hr at 37°C washed three times in HBSS, and suspended in DMEMserved as nonproliferating stimulator cells. They were generally used over a range of 105-2.5 X lo3 cells/well. Isolated tumor cell membranes and soluble membrane (glyco) proteins were tested for activity in a similar procedure. Antigens were usually tested in serial double dilutions. Responder/stimulator mixtures were incubated in 200 ~1 of DMEM/well, supplemented with 10% FCS and 5 X lOAsM 2-ME for 5 days, in an atmosphere of 95% air and 5% C02. Tests were set up in quadruplicates. Cell-mediated cytotoxicity. After 5 days of incubation, microplates containing the responder/stimulator mixtures were centrifuged, the supematants discarded, and target cells which were suspended in fresh DMEM containing 10% FCS ( IO4 cells in 200 ~1 of DMEM/well) were added. This procedure measured the cytotoxicity developed in one well without taking into account the actual number of effector cells. 3Abbreviations used: P-815/Be, P-8 15, line Beme, extensive metastasesformation; P-8 15/.X, P-8 15, line of the Sloan Kettering Institute for Cancer Research, N.Y. (through the courtesy of J. H. Burchenal), lower metastasesformation; HBSS, Hanks’ balanced salt solution; DMEM, Dulbecco’s modification of the minimal essential, medium; FCS, fetal calf serum; NCS, newborn calf serum; HS, horse serum; 2-ME, 2mercaptoethanol; CTL, cytotoxic T lymphocyte; id, intradermal; [‘H]TdR, tritiated thymidine; DOC, sodium deoxycholate; Pap, papain; P-8 1S/Mit, mitomycin-treated P-8 15 cells; DLN, draining lymph node.

SOLUBLE MEMBRANE

PROTEINS SUPPRESS CTLs

15

In some experiments, however, effecters were titrated in serial double dilutions with half the stimulator cell number as the starting effector population. Target cells were labeled with i3H]TdR (0.5 pCi/ml, sp act 24 Ci/mMol. Amersham/Searle) for 15- 16 hr. In order to guarantee the release of the radioactive label from damaged target cells, a short treatment with trypsin followed the 24-hr incubation period of effecter/target cell mixtures (25% trypsin (Difco) 1:250 in HBSS (IS)). This treatment did not raise the level of the spontaneous release,which usually was between 5 and 15% after 24 hr. Total release of the marker from all cells was obtained by treatment with 0.5% DOC. Supematants of experimental wells and controls were counted in a scintillator fluid (Permablend III (Packard), Triton X-100, and Xylol) for @emission. Specific cytotoxicity was expressed as specific cytotoxicity =

cpm exp - cpm spont x 100. cpm total - cpm spont

Isolation of cell membranes and solubifization of membrane proteins. Cell membranes were isolated and solubilized essentially as described ( 13). In brief, washed P8 15 cells were homogenized by N+avitation and were centrifuged at low speed to remove unbroken cells, nuclei, and larger subcellular components. A crude membrane fraction was then collected as a sediment of a centrifugation step of 100,OOOgfor 1 hr. The initial cell concentration was 10% (w/v). The crude membrane preparation in the form of small vesicles was solubilized for 20 min in 1.2% DOC, centrifuged for 1 hr at 1OO,OOOg, and the supematant was treated with 6 mg/ml of papain (Type II, Sigma Chemical Co.) after activation with cysteine (0.3 mg/ml). The reaction was stopped after 10 min with iodoacetic acid (IO-* M final concentration) and the detergent was removed by extensive dialysis against 5 tiphosphate buffer, pH 8, and later against DMEM without serum. After an additional centrifugation of 1 hr at 1OO,OOOg, the supematant was used for secondary stimulation experiments in vitro. Treatment with anti-Lyt antibodies and complement. Monoclonal anti-Lyt-1.1 and Lyt-2.1 antibody in combination with low-toxicity rabbit complement (both from Cederlane Lab., Ontario, Canada) were used in a two-step procedure essentially as indicated by the manufacturer for depletion of T-cell subpopulations.

RESULTS Cytotoxicities of P-8 1S/SK-primed DBA/2 lymph node cells against P-8 15/Be target cells as a function of the number of P-8 15/Be cells used as second-order stimulators for in vitro stimulation are shown in Fig. 1. A slightly depressed stimulation was often but not regularly observed at the highest stimulator-cell concentration (10’ cells/ well). Generally, 5 X lo4 to 1.25 X lo4 stimulators/well, corresponding to responder/ stimulator ratios of 20:5 to 20: 1.25, induced optimum cytotoxicity, and with lower stimulator-cell numbers the capacity to stimulate a measurable cytotoxic reaction was gradually lost. The extent of stimulation and the optimum responder/stimulator ratio, however, varied considerably from one experiment to the other. In the absence of stimulator cells cy-totoxicity was low or not detectable. Spenocytes from the same tumor-bearing animals gave either no or a very weak response,as did cells from distal nodes or from unprimed animals (Fig. 1). In order to test for total cytolytic capacity inherent in one responder/stimulator

16

BERTSCHMANN

AND LiiSCHER

l 0

1

2

3

4 antigen

5

6

dilution

FIG. I. Cytotoxicity of lymphocytes from DLN and the spleen of P-81S/SK-primed DBA/Z animals after secondary stimulation in vitro with P-815/Be/Mit. Responder cells (2 X 105/well) from 0, DLN, A, spleen. Antigen: P-8 1S/Be/Mit, initial number I05/well. Controls: 0, DLN cells alone; A, spleen cells alone; 0, responder cells from distal lymph nodes, stimulated with IO5 P-815/Be/Mit; A, responder cells from unprimed lymph nodes stimulated with 10’ P-8 1S/Be/Mit.

mixture, effector cell suspensionswere also used in serial double dilutions. Figure 2B reveals the total cytolytic capacity contained in one well. Note that despite the ap parent incomplete lysis (Fig. 2A), a wide reserve of killing potency is present in the stimulated responder suspensions (Fig. 2B). Secondary stimulation with membrane vesicles. Crude membrane preparations showed only a very limited capacity to stimulate CTLs in vitro in the syngeneic P8 11%DBA/2system. Figure 3 depicts a typical experiment: With relatively low membrane concentrations a weak stimulation was regularly obtained; higher membrane concentrations, however, were ineffective. In experiments where a residual cytotoxicity was measured in the absence of secondary in vitro stimulation, it appeared that this residual cytotoxicity was suppressedwhen tumor cell membranes were added to the cultures over a wide range of concentrations (Fig. 3). Efict of soluble membrane proteins. DOG/Pap-treated P-815/Be membranes had a similar suppressive activity on the residual cytotoxicity as crude membrane prep arations, but had no stimulatory capacity (Fig. 3). A mock preparation, i.e., buffer containing DOG/Pap and passed through the same procedures as the membrane proteins, proved to suppressthe residual cytotoxicity only at the highest concentration used. Lower concentrations ( 1:2 and lower) were no longer suppressive. The suppression of residual cytotoxicity exerted by the most concentrated mock DOG/Pap preparation could, after microscopic evaluation of the developing effector cell population, but attributed to direct killing of a large percentage of responder cells. Such direct toxicity, most likely due to the presence of residual amounts of detergent, should,

GOLUBLE MEMBRANE

17

PROTEINS SUPPRESS CTLs

40 30

30

20

20

10

10 0

2

1

3

antigen

4 control

2

3

dilutjon

4

5

effector

6

7

cell dilution

FIG. 2. Cytotoxicity of lymphocytes from DLN of P-81S/SK-primed DBA/2 animals after secondary stimulation in vitro. (A) Dependency of cytotoxicity from antigen concentration. Antigen: P-8 I S/Be/Mit; initial number, 105/well.(B) Cytotoxicity inherent in every responder/stimulator mixture of (A). Responder/ stimulator rations: 0, 20: 10; n , 20:5; X, 20:2,5; 0, 20:1,25.

however, not influence the results of all further experiments for which DOG/Pap membrane preparations were used as suppression-inducing agents at concentrations of 1:50. Membrane proteins not only suppressed cytotoxicity which developed spontaneously from in Co-primed lymph node responder cells during in vitro cultivation,

E - 70 ,z 3! 60 ‘;0

50 40 30 20 10

1

2

3

4 antigen

5

6

dilution

FIG. 3. Cytotoxicty of Iymphocytes from DLN of P-815/SiGprimed DBA/Z animals after secondary stimulation in vitro. Antigen: 0, P-8 lS/Be/Mit; A, P-815/Be membrane vesicles; n , soluble P-815/Be membrane proteins. 0, Control, responder cells atone.

18

BERTSCHMANN 310 ,.

AND LiiSCHER

A

"P'O> .t j 60.

.E X2 60 ,o s b 50

0

8 0' 50.

40

40-

20

20.

10

10. -

a

2

3

4

2

ag. dilution

3

4

ag. dilution

FIG. 4. Cytotoxicity of lymphocytes from DLN of P-815/SK-primed DBA/2 animals after secondary stimulation in vitro. (A) Soluble membrane proteins present during the induction phase. Antigen: 0, P8 1S/Be/Mit; 0, P-8 1S/Be/Mit plus soluble membrane proteins; 0, responder cells alone. (B) Soluble membrane proteins present during the effector phase. Antigen: V, P-8 I S/Be/Mit; V, P-8 1S/Be/Mit plus DOC/ Pap membranes; V, responder cells only.

but also cytotoxicity induced by secondary stimulation with intact tumor cells (Fig. 4A). In order to exert their suppressing effect, membrane proteins had to be present during the induction phase of cytotoxicity. Presenceof the material during the effector phase did not prevent cell lysis (Fig. 4B). 2 ‘;.e

80

.o

70

'5

60 50 40 30 20 10

Ln r a

b

1" Exp.

c

d

e

2"

f

9

Exp.

FIG. 5. Cytotoxicity resulting from in vitro stimulation of in viv+primed lymph node cells with various agents. Stimulation with (a) P-8 1S/Be/Mit; (b) DOG/Pap membrane proteins (1’ Expt.); (c) cells from (a), restimulated with P-8 1S/Be/Mit; (d) cells from (b), restimulated with P-8 1S/Be/Mit; (e) new responder cells stimulated with P-8 I S/Be/Mit; (f) new responder cells plus cells from (a), stimulated with P-815/Be/Mit; (g) New responder cells plus cells from (b), stimulated with P-81S/Be/Mit (2” Expt.).

SOLUBLE MEMBRANE

PROTEINS SUPPRESS CTLs

19

Mechanism of suppression. Responder cells were incubated during 5 days with a suppressive concentration of solubilized membrane proteins. After centrifugation supernatants were removed and the responder cells, suspended in fresh medium, were admixed to new in viva-primed responder cell populations which were then stimulated with P-8 1S/Mit. Mixtures of new and old responder cells were usually tested in double concentrations (old plus new) and in normal concentrations (l/2 old plus l/2 new). The results of a representative experiment are summarized in Fig. 5 for the optimum responder/stimulator ratio. Responder cells, once treated with DOG/Pap membranes, were no longer capable of being restimulated with intact P-8 lS/Mit, whereas responder cells stimulated a first time with P-8 1S/Mit were perfectly able to respond again. Cell populations treated with DOG/Pap membrane proteins had not only lost the capability to develop CTLs upon restimulation with P-815/Mit, but they also abrogated the capability of the new responder cells to develop cytotoxicity. The new responder cells alone reacted normally to a stimulation with P-815/Mit. Zdentijication of the cell type active in suppression. Lymph node cells treated with DOG/Pap membranes and therefore capable of transferring suppression to new responder cell populations lost this capability when treated with monoclonal anti-lyt2.1 antibody (1:20 dilution) and complement before being added to new responder cell populations. The same treatment with monoclonal Lyt- 1.1 antibody (I:20 dilution) and complement had no influence on the suppressive capacity, thus demonstrating that the population transferring suppression was a Lyt-l-2+ population (Figs. 6A, B).

DISCUSSION Cells from the DLN of animals bearing id P-815/% tumors were used in the present work as a responder cell population for secondary in vitro stimulation of A

40~

30.

20.

loA

,

1

2

ag. dilution

3

1

2

3

ag. dilution

FIG. 6.Cytotoxicity of lymphocytes fromDLN of P-8 1S/SK-primed DBA/Z animals. Effect of admixture to newresponder cells of DOG/Pap membrane-induced suppressor cellsandDOG/Pap membrane-induced suppressor cells after treatment with monoclonal anti-Lytantibodyand complement. (A) Treatment with

anti-Lyt-2.1 antibody, and (B) treatment with anti-Lyt-1 .l antibody. 0, No suppressor cells added; A, untreated suppressor cells added; A, antibody plus complement-treated suppressor cells added.

20

BERTSCHMANN

AND LiiSCHER

CTLs. Lymph node cells were chosen because,when compared to splenocytes, these cells showed a higher cytotoxicity after in vivo priming ( 12). The P-815/SK line was used for in vivo priming because of its lower metastatic capacity and the hi& yield (up to 16 X lo6 cells/node) of responder cells. High cytotoxic reactivity of cells from the DLN and the virtual absence of responsivenessin spleen cell populations were confirmed by secondary in vitro stimulation. After secondary stimulation higher levels of cytotoxicity were obtained than after priming in vivo. The long incubation periods (>6 hr), which were needed in order to observe cytotoxicity after priming in vivo, were necessary for second-order effector cells as well. This possibly reflects the lower immunogenicity of the P-8 15/ Be line as compared to P-815 lines used by others (3, 4, 20). Although total killing of all target cells was never observed, it could be demonstrated that a large excess of killing capacity was present in effector cell populations. The observed refractoriness against lysis of part of the tumor cells may be of practical importance as such cells might represent phenotypes from which progressive tumors can develop. At limited (low) concentrations membrane vesicles induced a very weak cytotoxicity. Reduced immunogenicity as compared to whole cells was likewise observed upon allogeneic and xenogeneic stimulations with membranes or membrane proteins incorporated into liposomes (15, 18). Suppression, however, which was observed in our experiments over a wide range of concentrations of the stimulating material, was not found by other authors (15, 18). An overdose of antigen has been described to induce suppression (16). It is doubtful whether in our tumor model suppression can be attributed to an antigen overdose, since the nature of the structures involved in T-cell stimulation is essentially unknown. It may be cited in this context that two antigens have so far been characterized by our group on the P-8 15 cell with the help of antibodies. One of them was labile and the other one was more stable during the procedure of membrane isolation (13). Their participation in T-cell stimulation remains to be elucidated, however. Considering the exclusive stimulation of the suppressor pathway by soluble membrane proteins (see below), it is suggestedthat the presence in or the release from membrane vesicles of soluble proteins is responsible for the induction of suppression by membrane vesicles. D()C/Pap-solubilized membrane proteins were in no concentration active in CTL stimulation. These findings are in line with earlier reports by our group and by others that isolated membranes or membrane proteins hardly confer protection against a subsequent challenge with viable tumor cells (9, 10, 16). However, successful immunization with subcellular membrane components has also been reported (2 1, 22). For tumor antigens and histocompatibility antigens there also seemsto exist a correlation between antigen particle size and the induction of protection (23, 24). Alaba and Law (25) reported a possible exception from this rule for the highly immunogenic Rauscher leukemia virus-induced lymphoma RBL-5, of which detergent-solubilized membranes induced excellent cytotoxicity in vitro and protection in vivo as well. In the presence of soluble membrane proteins the stimulation of CTLs by whole P-8 15 cells was suppressed,i.e., the pathway of suppression dominated the pathway of cytotoxicity. An analogous in vivo situation has been described by Dye and North (3). The supposed presence in tumor-bearing animals of suppressor T cells abrogated the efficiency of adoptively transfered CTLs, which was pronounced in T-cell-depleted tumor bearers.

SOLUBLE MEMBRANE

PROTEINS SUPPRESS CTLs

21

Intradermally developing P-815/SK tumors invariably regress in syngeneic hosts despite their high ip malignancy (M. B., unpublished results). Nonetheless, cells from the DLN of P-815/SK tumor bearers are excellent responder cells for stimulation of CTLs and of suppressor cells as well. This situation illustrates that the mere presence of suppressor or suppressor inducer cells does not allow conclusions about their role in progressive or regressive tumor development in vivo. The difficulties which we encountered in trying to modulate the reactivity against id developing P-815/Be tumors (1 I) and even more so the fact that the most active population of suppressor cells was found by Hellstrom et al. (7) in animals which had successfully rejected their tumors, point in the same direction. Suppression in our experiments was found to be associatedwith a cell population of the Lyt-l-2+ phenotype, whereas Maier et al. (10) found their P-815induced suppression to be associated with Lyt- I ‘2- cells. This discrepancy can possibly best be explained by the identification of suppressor effector or suppressor inducer pop ulations, respectively (26). Suppressor cells of the Lyt-1+2- phenotype were found by Rao et al. (27) in another DBA/2 tumor, the L1210 leukemia. The test which was used by these authors was adoptive transfer, which, however, is perhaps more suited for the detection of suppressor inducers. In our experiments the suppressor population present after the 5-day induction period apparently consisted mainly of suppressor effecters. The Lyt- l-2+ phenotype, as found in our experiments, would rather point to an antigen-specific than to a promiscous suppressor cell type (28) although both types of suppression, specific and nonspecific, have been described for P-815 tumors (3-5, 10). Parish (28) has pointed to the possibility that cytotoxic T cells directed against foreign serum components may be involved when in vitro stimulation is performed. Although nonspecific cytotoxicity or cytotoxicity directed against serum components cannot totally be excluded in our experiments, its contribution is probably negligible in view of the fact that the second-order stimulator cells or target cells were also suitable when cultivated in vivo or in DMEM supplemented with horse serum. In viva-grown P-815 cells were also capable of inducing effective priming and, additionally, the effector phase but not the induction phase of secondary stimulation could be performed in DMEM containing horse serum. Soluble tumor constituents have been shown by Yamauchi et al. (29) to induce in vivo highly specific suppressor T cells capable of recognizing structures different from those that stimulated CTLs. As compared to our experiments, this work shows several differences: first, 3 M KCL extracts of whole cells were used and the tumors were highly immunogenic (sarcomas S1509 and SaI); second, suppressor cells were induced in vivo in unprimed animals; and third, suppressor cells were active during the effector phase of cytotoxicity, whereas in our test it was the induction phasewhich was suppressed. In both cases,however, soluble proteins did not directly influence the CTL effector phase. Taken together our results allow the hypothesis that the termination of the regression phase of id developing P-815/Be tumors (12) might be caused at least in part by the release of soluble membrane components from decaying tumors. ACKNOWLEDGMENTS We thank Dora Etter and Peter Mauderli for excellent technical assistance.

22

BERTSCHMANN

AND LiiSCHER

REFERENCES 1. Gershon, K., TranspIant. Rev. 26, 170, 1975. 2. Cantor, H., Boyse, E., In “Contemporary Topics in Immunobiology” (0. Stutman, Ed.), Vol., 7, pp. 47-65. Plenum, New York, 1977. 3. Dye, E. S., and North, R. J., J. Exp. Med. 154, 1033, 1981. 4. Tilkin, A. F., Schaaf-Lafontaine, N., Van Acker, A., Boccadoro, M., and Urbain, J., Proc. Nat. Acad. Sci. USA 78, 1809, 1981. 5. Kamo, I., Patel, C., Patel, N., and Friedman, H., J. Immunol. 115, 382, 1975. 6. Argyris, B. F., J. Nat. Cancer Inst. 64, 959, 1980. 7. Hellstrom, I., Hellstrom, K. E., and Bernstein, I. D., Proc. Nat. Acad. Sci. USA 76, 5294, 1979. 8. Hellstrom, K. E., Hellstrom, I., Kant, J. A., and Tamericus, J. D., J. Exp. Med. 48, 799, 1978. 9. Bertschmann, M., and Liischer, E. F., In “Microenvironmental Aspects of Immunity” (B. D. Jankovic and K. Isakovic, Eds.), Vol29, pp. 5 13-518, Plenum, New York, 1973. 10. Maier, T., Levy, J. G., and Kilbum, D. G., Cell Irnmunol. 56, 392, 1980. 11. Bertschmann, M., Schiiren, B., and L&her, E. F., Eur. J. Cancer 18, 405, 1982. 12. Bertschmann, M., Schiiren, B., and L&her, E. F., Immunobiol. 156, 382, 1979. 13. Clemetson, K. J., Gerber, A., Bertschmann, M., Widmer, S., and L&her, E. F., Eur. J. Cancer 12, 263, 1976. 14. MacDonald, H. R., ImmunoI. Today 3, 183, 1982. 15. Engelhard, V. H., Strominger, J. L., Mescher, M., and Burakoff, S., Proc. Nat. Acad. Sci. USA 75, 5688, 1978.

16. Price, M. R., Preston, V. E., Robins, R. A., Ziiller, M., and Baldwin, R. W., Cancer Immunol. Immunother. 3, 247, 1978. 17. Clemetson, K. J., Bertschmann, M., and Liischer, E. F., Mol. Immunol. 18, 135, 1981. 18. Lemonnier, F., Mescher, M., Sherman, L., and Burakoff, S., J. Immunol. 120, 1114, 1978. 19. Warren, H. S., and Lafferty, K. J., &and. J. Immunol. IO, 349, 1979. 20. Takei, F., Levy, J. G., and Kilbum, D. G., J Immunol. 116, 288, 1976. 21. Natori, N., Law, L. W., and Appella E., Cancer Res. 37, 3406, 1977. 22. Pellis, N. R., Mokyr, M. B., Babcock, J. R., and Kahan, B. D., Immunol. Commun. 7, 431, 1978. 23. Kahan, B. D., In “Transplantation Antigens” (B. D. Kahan and R. A. Reisfeld, Eds.), pp. 31 l-338. Academic Press, New York, 1972. 24. Fast, L. D., and Fan, D. P., J. Immunol. 120, 1092, 1978. 25. Alaba, O., and Law, L. W., J. Exp. Med. 148, 1435, 1978. 26. Cantor, H., and Gershon, R. K., Fed. Proc. 38, 2058, 1979. 27. Rao, S. K., Bennett, J. A., Shen, F. W., Gershon, R. K., and Mitchell, M. S., J. Immunol. 125, 63, 1980. 28. Parish, C. R., Immunol. 33, 597, 1977. 29. Yamauchi, K., Fujimoto, S., and Tada, T., J. Immunol. 123, 1653, 1979.