CELLULAR

IMMUNOLOGY

141, 1-9 (1992)

Potentiation of Transmembrane Signaling by Cross-Linking of Antibodies against the p Chain of the T Cell Antigen Receptor of JURKAT T Cells MAHMOOD JEDDI-TEHRANI,* SEK C. CHOW,? IGNACIO J. ANSOTEGUI,* MIKAEL JONDAL.* AND HANS WIGZELL* Departments of *Immunology

and t Toxicology, Karolinska

Institute, Box 60400, Stockholm, Sweden

Received February 5, 1991; accepted December 5, 1991

Threemonoclonalantibodies(mAb)2D1,3B9, and 3B12 were produced by immunizing BALB/ c mice with JURKAT cells. ThesemAb induce comodulation of the TCR/CD3 complex expressed on JURKAT cells, but do not react with the CD3- JURKAT variant, J.RT3.T3.1. Immunoprecipitation studies with detergent-solubilized JURKAT cell lystes indicate that these mAb react with proteins having characteristics of the TCR molecules. Their low reactivity with peripheral blood mononuclear cells(PBMC)and lackof reactivitywith otherCD3+T cell linessuggestthat they may be anti-idiotypic mAb. Results from binding inhibition assays,reactivity with PBMC, and generation of transmembrane signals suggestthat these three anti-TCR mAb recognized different epitopes on the TCR p chain of JURKAT cells. Although the three mAb are capable of inducing the production of inositol phosphatesand cytosohc free Ca’+ increasein JURKAT cells, their stimulatory capacities vary and are lower than that observed by anti-CD3 antibody (OKT3) stimulation. However, crosslinking thesemAb with rabbit antimouse immunoglobuhns potentiates the stimulatory responseto comparable levels induced by OKT3. These mAb could be useful as tools to study Vp8” T cells in relation to antigen-specific activation. a 1992 Academic press, Inc.

INTRODUCTION The human T-cell antigen receptor (TCR) present on mature T lymphocytes consists of two heterodimeric polypeptide chains (a and ,B)noncovalently associatedwith the CD3 molecular complex (l-5). This TCR/CD3 complex is known to play a major role in the initiation of T cell activation (6-10). Ligation of the TCR/CD3 complex has been shown to result in cell proliferation, cytotoxic function, or the production of lymphokines ( 1l- 13). Analysis of antibodies against the TCR/CD3 complex and their activation potentials has led to the characterization of several steps of T-cell activation via the TCR/CD3 complex. One of the earliest events involved in T-cell activation is the hydrolysis of inositol lipids (14). This in turn generates two putative second messengers,inositol 1,4,5-trisphosphate (Ins( 1,4,5)P,) and diacylglycerol. Ins( 1,4,5)P, induces the release of Ca*+ from intracellular stores ( 13, 15) while diacylglycerol activates protein kinase C to phosphorylate cellular proteins ( 16). mAb directed against different epitopes on the TCR/CD3 complex may have different stimulatory potentials (17) and a requirement of accessory ceils (AC) would normally be necessaryto fully activate some T cells. This is especiallytrue for peripheral

0008-8749192$3.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form resewed.

2

JEDDI-TEHRANI

ET AL.

blood mononuclear cells (PBMC). Accessory cells can, however, be replaced by crosslinking the initial stimulating antibody by a second antibody, or by coupling the first antibody on a stationary phase (18). In the present study, three mAb recognizing different although partially overlapping epitopes on the V p region of the TCR in the human leukemic T-cell line (JURKAT) were produced. The mAb were found to have lower stimulatory effect (generation of inositol phosphates and increase in intracellular free Ca2’ concentration, ( [Ca2’]i)), but cross-linking these mAb results in signals comparable to that obtained by activation through the CD3 complex. EXPERIMENTAL

PROCEDURES

Materials The calcium indicator Fura 2-AM and biotin were from Sigma Chemical Co. myo[3H]Inositol was from American Radiotabeled Co. and 1251was from Du Pont Scandinavian AB. Streptavidin-FITC, rabbit anti-mouse immunoglobulins, and rabbit antimouse F(ab’)2FITC were from Dakopatts. Prestained molecular weight standards were from Bio-Rad. RPM1 1640 and fetal calf serum (FCS) were from GIBCO, UK. Inositol-free RPM1 1640, Ficoll-Hypaque, and protein A-Sepharose beads were from Pharmacia. The antibody VP8 was a gift from Dr. C. H. Janson, Department of Immunology, Karolinska Institute, Stockholm, Sweden. All other reagents used were of analytical grade. Cell Lines and Purification of PBMC JURKAT.CD3+, J.RT3.T3.1. a CD3- JURKAT variant, and E6-1 .CD3+ (IL-2 producing JURKAT variant) were obtained from ATCC. Molt-4.CD3-, HPB-ALL.CD3+, HPB-MLT.CD3+, Molt-16.CD3+, and H9.CD3+ were from Dr. Jun Minovada, Fuji&i Cell Center (Japan). Normal JURKAT clones and all cell lines were maintained in suspension culture using RPM1 1640 medium supplemented with 5% FCS and antibiotics ( 100 U/ml penicillin and 100 &ml streptomycin) in a humidified incubator (37°C) under an atmosphere of 5% CO2 in air. PBMC were obtained by Ficoll-Hypaque gradient centrifugation of peripheral blood from healthy blood donors. Production of anti-TCR mAb The procedures were as described previously (19). In brief, BALB/c mice were immunized intraperitoneally with 20 X lo6 JURKAT cells. After 3 weeks, the mice received another injection with the same amount of cells. Three days later, the spleen cells of the mouse were fused with the mouse myeloma cell line SP2/0. Hybridomas obtained were selected against HAT medium. Supematants from hybridomas were tested for their reactivity and ability to modulate the TCR/CD3 complex from JURKAT cells using indirect immunofluorescence test (IFL) (19). Comodulation of TCR/CD3 Complex Hybridoma supernatants reactive with JURKAT cells in IFL were incubated with 1 X lo6 JURKAT cells for 18 hr at 37°C. Thereafter cells were washed and analyzed

ANTIBODY

TRANSMEMBRANE

SIGNALING

using IFL with FITC-conjugated anti-Ti mAb or biotinylated streptavidin-FITC as the conjugate.

3 OKT-3, using

Iodination of mAb and Cell Surface Proteins Anti-TCR mAb were labeled with 1251by the iodogen method (20), followed by separation over a G25 superfine Sepharose (Pharmacia, Uppsala, Sweden) column. The fractions with the highest amount of radioactivity were pooled and frozen at -20°C until used. Iodination of cell surface proteins was performed using the lactoperoxidase method (21). Labeled cells were washed with PBS and lysed in NP-40 or digitonin lysing buffers (0.5% NP-40 in 0.1 M NaP04, pH 8, or 1% digitonin in 0.1 A4 NaP04, pH 8, respectively) with protease inhibitors (aprotinin 0.06 TIU/ml and 1% 6-aminocapronic acid). After 30 min incubation on ice, the nuclei and cell debris were separated by centrifugation (20,OOOgfor 30 min). The lysates were preabsorbed for 20 hr by mixing with protein A-Sepharose beads (Pharmacia, Sweden) at 4°C before used. Immunoprecipitation and Separation of Cell Surface Proteins Anti-TCR mAb were added to protein A-Sepharose and incubated for 1 hr at 4°C. Thereafter, the beads were washed four times with 0.1 A4 NaP04 at pH 8 and then once with the appropriate lysing buffers. Preabsorbed cell lysates (see above) were added to the beads and mixed continuously for 2 hr at 4°C in Eppendorf tubes. The beads were then washed six times with the appropriate lysing buffers. The bound proteins were eluted with sample buffer, and separated using 12% SDS-PAGE under reducing and nonreducing conditions. Bio-Rad prestained molecular weight standards were used for calibration. The gels were then dried and autoradiographed. Blocking Assays JURKAT cells were first incubated with anti-CD3/TCR mAb or medium for 30 min on ice. 1251-labeledanti-TCR mAb or biotinylated OKT-3 were then added to the cells and then further incubated on ice for 30 min. Thereafter, cells were washed three times with ice cold PBS. The reactivities of radiolabeled or FITC-conjugated mAb were determined using gamma counters and flow cytometry, respectively. Measurement of Cytosolic Free Cazt Concentration The method used was the same as previously described (22). In brief, Fura 2-loaded JURKAT cells (3 X 1O6cells) were suspendedin 2 ml modified Krebs-Henseleit buffer in a quartz cuvette maintained at 35°C in a Sigma ZFP 22 dual wavelength spectrofluorometer (Sigma Instruments, Berlin West, FRG) with excitation wavelengths 334 and 366 nm (emission was measured using a 500 nm cutoff filter). To minimize leakage, 35°C was used instead of 37°C (23). Changes in cytosolic free Ca2’ were monitored by recording changesin the Fura 2 fluorescence ratio signal, calibrated in terms of [Ca2’]i as described (24) using 225 nA4 as the Kd for Fura 2. Production and Separation of Radiolabeled Inositol Phosphates This was essentially the same as previously reported (22). In brief, myo-[3H]inositol labeled JURKAT cells were stimulated with mAb alone or 5 set later with rabbit anti-

4

JEDDI-TEHRANI

ET AL.

TABLE I Reactivity of Anti-Ti mAb on Different T Cell Lines and PBMC Anti-Ti mAb Cells

Characteristics

JURKAT” JURKAT E6- 1 J.RT3.T3.1_ Molt-4 Molt- 16 H9 HPB-ALL HPB-MLT PBMC’

Wild type IL-2 producing CD3- variant CD3CD3+ CD3+ CD3+CD4+CD8+ CD3+CD4+CD8+

68.7’ 80.4 1.1 0.3 0.3 I.0 0.6 0.3 0.5-2

77.3 86.6 2.0 0.3 0.4 0.9 0.5 0.2 l-4

74.1 85.0 1.5 0.4 0.3 0.7 0.7 0.3 0.5-2

79.8 88.9 2.7 15.8 100.0 99.8 99.3 99.7 76.9

Note. Different T cell lines and PBMC were first incubated with optimal concentrations of the anti-Ti mAb (OKT3, 2D1, 3B9, and 3B12), and then with the FITC-conjugated F(ab’)* fragment of rabbit antimouse immunoglobins as outlined under Experimental Procedures.Cells were analyzed using flow cytometry (FACSCAN). u Do not produce IL-2. b Percentageof positive cells. ’ PBMC from 15 different donors were analyzed.

mouse immunoglobulins. Incubations were terminated after 10 min with ice cold 10% perchloric acid and the supernatantswere processedas describedby Downes et al. (25). RESULTS AND DISCUSSION Spleen cells from BALB/c mice immunized with JURKAT cells were fused with SP2/0 myeloma cells to generatehybridomas. Using IFL and CD3 comodulation tests, three mAb (2D1, 3B9, and 3B12) produced by the hybridomas were found to react with the JURKAT cells and comodulate the TCR/CD3 complex from the cell surface. TABLE 2 Comodulation of Ti and CD3 Complex on JURKAT Cells Percentagereactivity of mAb Modulating mAb

Medium

2Dl

3B9

3B12

OKT3

Medium 2Dl 3B9 3B12 OKT3

0.4 0.6 0.7 0.8 0.5

73 30 31 26 25

61 9 7 11 4

58 9 13 15 8

68 5 14 I 0.5

Note. JURKAT cells were incubated for 18 hr at 37°C with saturating amounts of mAb (2D 1, 3B9, 3B 12, and OKT3). Thereafter, cells were stained with FITC-conjugated 2D1, 3B9, and 3B12, and biotinylated OKT3 (with streptavidin-FITC as fluorescent conjugate) and analyzed using flow cytometry (FACSCAN) as outlined under Experimental Procedures.

ANTIBODY

TRANSMEMBRANE

SIGNALING

5

All three mAb failed to react with J.RT3.T3.1 cells, a TCR- JURKAT variant (Table 1). Incubation of JURKAT cells with the three hybridoma supernatants or OKT3 for 18 hr resulted in comodulation of structures recognized by all three mAb (Table 2). Further evidence that these mAb recognized the TCR and not the CD3 complex came from immunoprecipitation studies. The results in Fig. 1 (NP-40 lysates) show that under nonreducing conditions 2D1, 3B9, and 3B12 precipitated a single band at 90 kDa, while OKT3 yielded two bands, 20 and 28 kDa. Under reducing conditions, the proteins precipitated by 2D1, 3B9, and 3Bl2 yielded two bands, 43 and 48 kDa, whereas OKT3 gave results similar to those observed under nonreducing conditions. With digitonin lysates (Fig. 1) and under nonreducing conditions, all mAb (2D 1,3B9, 3B12, and OKT3) yielded three bands, 90,28, and 20 kDa. Under reducing conditions, the proteins brought down by all the mAb yielded four bands, 48,43,28, and 20 kDa. Because NP-40 treatment dissociates the TCR from the CD3 complex, the results show that 2D1, 3B9, and 3B12 recognize structures associatedwith the TCR and not the CD3 complex. Furthermore, OKT3 had little effect on the binding of these three mAb to JURKAT cells, which further indicates that these mAb bind to the TCR and not the CD3 complex. Reactivity of these three mAb with PBMC of normal donors (Table 1) (0.5-2% for 2D 1 and 3B 12, l-4% for 3B9) is consistent with the frequency of expression of different V gene products in normal PBMC (26). When other CD3+ T-cell lines were used, i.e., Molt- 16, H9, HPB-ALL, HPB-MLT, the reactivity was lessthan 1%for all three mAb. On the other hand, OKT3 gave >99% reactivity with the same T-cell lines and 76.9% (n = 15) reactivity with PBMC. The failure of 2D 1,3B9, and 3B 12 to react with other

A

NP-40

Digitonin -

110 84 47

* * -

33 24

* -

12345

12345

B NP-40

Digitonin

12345

12345

1.Immunoprecipitation of anti-ii mAb. Antigens were immunoprecipitated from detergent-solubilized I’z5-labeled JURKAT cells. A, nonreducing conditions; B, reducing conditions. Lane 1 (T4.2, anti-CM), lane 2 (2Dl), lane 3 (3B9), lane 4 (3B12), and lane 5 (OKT3). Proteins were analyzed using 12% SDSPAGE as described under Experimental Procedures. FIG.

6

JEDDI-TEHRANI

ET AL.

TABLE 3 Binding Inhibition Assay with anti-Ti mAb Percentageinhibition by anti-Ti mAb” Labeled mAb

2D1

3B9

3B12

OKT3

2DI 3B9 3B12 OKT3’

19 16 50 0

83 90 84 0

86 89 71 0

-17 22 24 99

Note. JURKAT cells (1 X IO6cells) were first incubated with excesscold or nonconjugated mAb before the addition of “‘1 or biotin-labeled mAb as described under Experimental Procedures. a A representative experiment out of three different experiments. b Biotin conjugated.

TCR/CD3+ T-cell lines strongly suggeststhat these three mAb are strictly clonotypic and recognize the variable region of the TCR in JURKAT cells. Using blocking assays(Table 3), two of the mAb, 2D1, and 3B12 appear to map different epitopes of the V region (based upon their different ability to inhibit the binding of each other to the TCR). 3B9 and 2Dl inhibit the binding of each other to the TCR with similar ability. Their reactivity with PBMC was, however, different, with 3B9 having consistently higher reactivity (0.52% for 2Dl and l-4% for 3B9), suggestingthat they may recognize different epitopes of the V region. This notion was further supported when the two mAb were allowed to bind after JURKAT cells were first exposedto OKT3. The results consistently demonstrate that upon the binding of OKT3, binding of 2Dl increasedwhereas3B9 or 3B12 binding was reduced. Interaction of OKT3 with the CD3 complex may induce conformational changes that increase the affinity of 2Dl to its binding site. Using an antibody (V/38) against the VP8 family, the binding of all three mAb was found to be inhibited (data not shown). This suggests that these three mAb, 2D 1, 3B9, and 3B 12 recognized epitopes on the fi chain of the TCR in JURKAT cells. Induction of inositol lipid hydrolysis and increase in [Ca2+]i after ligation of the TCR/CD3 complex by anti-TCR or CD3 antibodies are well characterized early events in T-cell activation ( 13- 15). This leads to cell proliferation, lymphokines production, and expression of T-cell markers such as IL-2 and transferrin receptors (1 l- 13). To determine whether 2D1, 3B9, and 3B12 could activate JURKAT cells, changes in [Ca2+]i and the production of inositol phosphates were monitored. JURKAT cells loaded with the Ca2+indicator dye, Fura 2 were stimulated with either of the three mAb or OKT3. The results in Fig. 2 (C, E, and G) show that all three anti-TCR mAb are capable of inducing [Ca2+]iincrease in JURKAT cells. The level of [Ca2+]iincrease was, however, considerably lower than the OKT3-induced [Ca2’]i increase (Fig. 2A). The time taken for [Ca2+]i to peak after stimulation by 2Dl and 3B12 was similar to that for OKT3, that is less than 1 min, while 3B9 required at least 2 min to induce a [Ca2+]ipeak. Increasing the antibody concentration of the three mAb did not induce any higher [Ca2+]i increase (results not shown), nor did it decreasethe time for any one of the mAb, especially 3B9, to induce [Ca2+]i to peak. Analysis of the inositol phosphates generated after stimulation by these anti-TCR mAb (Fig. 3) showed that

ANTIBODY 0.7

r

0.4

TRANSMEMBRANE

SIGNALING

A OKT3

Ii\ 0.2 I

I

0.1 i

0.4

2D1

.-s ii z

0.2

0.2

z g 0

0.1 I

0.1

0.7

r

F

E 389

I P

0.7 0.4

:iY

I -4

0.2

0.1 Ii -lo.7

0.4

3812

0.4

FIG. 2. Effect the anti-Ti mAb on of [Ca2+li increase in JURKAT cells. Changes in cytosolic free Ca2’ concentration were monitored using the Ca2+fluoresence dye Fura- as outlined under Experimental Procedures. Where indicated, anti-Ti mAb were added to the JURKAT cells either alone (A, C, E, and G) or followed by rabbit anti-mouse immunoglobulins (R), 5 set later (B, D, F, and H). All anti-Ti mAb were added to the cells in a final dilution of 1:400and rabbit anti-mouse immunoglobulins, a 1:200final dilution.

2Dl and 3B12 induced more inositol phosphate production than 3B9 but less than OKT3. This further supports the difference in epitope recognition by 3B9 and 2D 1. The difference in stimulatory potential of these mAb is in line with previous findings (17) that the biological activity of the anti-TCR V region mAb is determined by the epitope recognized. It is probable that upon the ligation of the TCR/CD3 complex by antibodies, conformational changes occur and result in transduction of signals across the plasma membrane (17). Although the exact mechanism is still unclear, evidence available has suggestedthe involvement of a coupling G protein which in turn activates phospholipase C to hydrolyze the inositol lipids. It is also possible that ligation of the TCR/

8

JEDDI-TEHRANI

N.S.

*

OKT3

ZDI

Monoclonal

ET AL.

389

3B12

antibodies

FIG. 3. Effect of the anti-Ti mAb on the generation of inositol phosphates. myo-[H3]Inositol-labeled JURKAT cells (10 X lo6 cells) were stimulated with the various anti-Ti mAb alone (blank bars) or together with rabbit anti-mouse immunoglobulins (cross-hatchedbars). Anti-Ti mAb were added in a final dilution of 1:400and rabbit anti-mouse immunoglobulins, a 1:200final dilution. Inositol phosphatesgenerated were analyzed as outlined under Experimental Procedures.Results were means + SEM from different batches of cells (n = 7). Statistical analyseswere performed using the Student t test with unpaired values. *P < 0.01; N.S., not significant.

CD3 complex by some soluble mAb alone or the site at which some mAb interact with the TCR is inadequate in inducing total conformational change and therefore lower signals acrossthe plasma membrane. This, however, can be overcome by crosslinking of the antibodies on a matrix or using second antibodies. Crosslinking of the three anti-TCR mAb with rabbit anti-mouse immunoglobulins markedly potentiated the [Ca2+]i increase and total inositol phosphates produced, Fig. 2 (D, F, and H) and Fig. 3, respectively. After crosslinking, 2Dl and 3B12 induced [Ca2+]i increases and inositol phosphate production greater than 3B9 and comparable to that of OKT3. Rabbit anti-mouse Ig on its own did not induce any increase in [Ca2+]ior production of inositol phosphates(results not shown). This is physiologically relevant since antigens presented by AC via MHC class I or II are physiologically crosslinked. Since virgin T cells expressing selectedcombinations of V gene products are of very low frequency, it is important that activation of these cells is performed in the most efficient way, that is by AC. Although the three mAb have different reactivity profiles against the VB8-linked epitopes of the human TCR, they all comodulate the TCR complex without the requirement of additional crosslinking antibodies. This suggeststhat these mAb exhibit a certain degreeof cross-linking themselvesbut it is lessefficient in the absence of secondary antibodies. From the conventional crosslinkagemodel, it can be predicted that F(ab’)2 fragments of the mAb, although bifunctional, may still be capable of activating JURKAT cells, while as monovalent fragments they would be expected to do very little. In fact, they should be expected to act as inhibitors for their bivalent counterparts. The F(ab)l fragments of these mAb could be useful as tools to define regions of the VP TCR

ANTIBODY

TRANSMEMBRANE

SIGNALING

9

epitopes and to analyze VPS’ T cells in relation to antigen-specific activation or activation by superantigens during T cell activation. ACKNOWLEDGMENTS This investigation was supported by grants from the Karolinska Institute, Swedish Medical Research Council, and the SwedishCancer Society. We thank Mrs. Berit Olssonfor her excellent work in the production of monoclonal antibodies.

REFERENCES 1. Meuer, S. C., Fitzgerald, K. A., Hussey, R. E., Hodgdon, J. C., Schlossman,S. F., and Reinherz, E. L., J. Exp. Med. 157, 705, 1983.

2. Haskins, J., Kubo, R., White, J., Pigeon, M., Kappler, J., and Marrack, P., J. Exp. Med. 157, 1149, 1983.

3. 4. 5. 6. 7. 8.

Samelson, L. E., and Schwartz, R. H., Immunol. Rev. 76,59, 1983. Brenner, M. B., Trowhridge, I. S., and Strominger, J. L., Cell 40, 183, 1985. Weiss, A., and Stobo, J. D., J. Exp. Med. 160, 1284, 1984. Van Wauwe, J. P., DeMey, J. R., and Goossens,J. G., J. Immunol. 124, 2708, 1980. Chang, T. W., Kung, P. C., Gingras, S. P., and Goldstein, G. Proc. Nut/. Acad. Sci. USA 78, 1805, 1981. Welte, K., Andreef, M., Platzer, E., Holloway, K., Rubin, B. Y., Moore, M. A. S., and Mertelsmann, R., J. Exp. Med. 160, 1390, 1984. 9. Meuer, S. C., Hodgdon, J. C., Hussey, R. E., Protentis, J. P., Schlossman,S. F., and Reinherz, E. L., J. Exp. Med. 158,988, 1983.

10. Weiss, A., Wiskocil, R., and Stobo, J., J. Immunol. 133, 1, 1984. 11. Hadden, J. W., Mol. Immunol. 25, 1105, 1988. 12. Meuer, S. C., Hussey, R. E., Fabbi, M., Fox, D., Acute, O., Fitzgerald, K. A., Hodgdon, J. C., Protentis, J. P., Schlossman,S. F., and Reinherz, E. L., Cell36, 897, 1984. 13. Weiss, A., Imboden, J. B., Shoback, D., and Stobo, J. D., Proc. Nutl. Acad. $5. USA 81, 4169, 1984. 14. Imboden, J. B., and Stobo, J. D., J. Exp. Med. 161,446, 1985. 15. Imboden, J. B., Weiss, A., and Stobo, J. D., J. Immunol. 134, 663, 1985. 16. Nishizuka, Y., Nature 308, 693, 1984. 17. Rojo, J. M., and Janeway, C. A., Jr., J. Immunol. 140, 1081, 1988. 18. Dixon, J. F. P., Law, J. L., and Favero, J. J., J. Leukocyte Biol. 46, 214, 1989. 19. Janson, C. H., Jeddi Tehrani, M., Mellstedt, H., and Wigzell, H. Scund. J. Immunol. 26, 237, 1987. 20. Fraker, P. J., and Speck, J. C., Biochem. Biophys. Rex Commun. 80, 849, 1978. 21. David, G. S., and Reisfeld, R. A. Biochemistry 13, 1014, 1974. 22. Chow, S. C., and Jondal, M., J. Biol. Chem. 265,902, 1990. 23. Treves, S., Di Virgilio, F., Cerundolo, V., Zanovello, P., Collavo, D., and Pozzan, T. J. Exp. Med. 166, 33, 1987. 24. Grynkiewicz, G., Poenie, M., and Tsien, R. Y., J. Biol. Chem. 260, 3440, 1985. 25. Downes, C. P., Hawkins, P. T., and Irvine, R. F., Biochem. J. 238, 501, 1986. 26. Kimura, N., Toyonaga, B., Yoshikai, Y., Du. P-P., and Mak, T. W. Eur. J. Immunol. 17, 375, 1987.