CELLULAR IMMUNOLOGY 134, 85-95 (199 i)

In Vitro Studies of the Effect of MAb NDA 4 Linked to Toxin on the Proliferation of a Human EBV-Transformed Lymphoblastoid

B Cell Line and of Gibbon MLA Leukemia Cell Line

PAUL HARRIS,*,’ ELAINE REED,* DONALD WEST KING,? AND NICOLE SUCIU-FOCA*

The rejection ofallografis is mediated by cytolytic T cells and antibody-secreting B cells. Selective ablation of these activated cells from peripheral blood lymphocytes may offer a method of con- trolling allograft rejection. An immunotoxin was prepared from the monoclonal antibody (mAb) NDA 4. which recognizes a differentation antigen (NDA 4) common to activated B and T cells. MAb NDA 4 was conjugated to the ribosome-inhibiting protein gelonin via a cleavable disultide bond provided by a crosslinking reagent. The purified immunotoxin was evaluated for in vifro cytotoxicity on NDA 4 positive T and B cell lines. Conjugation of mAb NDA 4 to gelonin increased the in vim cytotoxicity by a concentration factor of 1000, compared to gelonin alone. The specificity and saturability of mAb NDA 4 binding, as well as the number of antigenic sites per cell on resting versus activated T lymphocytes, were also evaluated. Resting T cells expressed 400-800 sites per cell. PHA-activated T cells and the MLA T cell leukemia expressed 10,000 to 80,000 sites per cell. Peripheral blood mononuclear cells obtained from allografted baboons in quiescence or undergoing rejection were compared for NDA 4 expression by flow cytometry. Lymphocytes obtained from baboons rejecting a heart allograft expressed NDA 4. whereas trans- plant recipients in quiescence showed no detectable NDA 4. These results suggest that mAb NDA 4-derived immunotoxins may be valuable for the selective depletion of activated lymphocytes while sparing the resting population. kti I99 I Academic Press. Inc.

INTRODUCTION

Recently, rapid progress has been made in defining and characterizing cell surface antigens associated with the growth, differentiation, and maturation of human leu- cocytes ( 1). Some of these antigens are expressed only by activated T and B lymphocytes and as such they may serve as targets for immunosuppressive therapy of autoimmune diseases and transplantation. Anti-interleukin 2 receptor mAb, recognizing alloactivated T lymphocytes, is a potent immunosuppressant when used in conjunction with cy- closporin A (2). Monoclonal antibodies linked to toxins enhance their effect by targeting the toxin to specific cells. Such immunotoxins may be useful for ablating selected population of cells without causing systemic damage of immune reactivity (3).

’ To whom correspondence should be addressed

85

0008-8749/9 I $3.00 Capynght G 1991 by 4cademjc Press. Inc. 411 nglm of reproductmn I” any form reserved.

86 HARRIS ET AL.

In previous studies we have described the functional and structural characteristics of cell surface antigens involved in T helper function (4), production of interferon-y and B cell growth factor (5), and lymphocyte proliferation and maturation (6, 7). These differentiation antigens were identified and studied by use of murine mAbs raised by immunizing mice with alloreactive human T cell clones.

One of these mAbs is of potential interest as an immunosuppressive agent since it recognizes a new differentiation antigen named NDA 4, which is expressed both by activated T and B lymphocytes and is associated with their proliferation (6). We have prepared an immunotoxin by conjugating the mAb which recognizes NDA 4 (mAb, NDA 4, an IgGl) to the ribosomal inhibitor protein gelonin. In this report, we describe the in vitro evaluation of the cytotoxicity exhibited by this immunotoxin. An important determinant for the success of selective immunotherapy is the relative expression of the target molecule, NDA 4 in this case, on resting versus activated cells. We have measured NDA 4 expression by Scatchard analysis of mAb NDA 4 binding to pop- ulations of resting, mitogen-activated, and transformed lymphoid cells in an attempt to anticipate the possible in vivo effects of the immunotoxin. To evaluate the potential usefulness of mAb NDA 4 for monitoring and treatment of allograft rejection, we have measured the expression of NDA 4 on in vivo activated lymphocytes obtained from baboons undergoing chronic rejection of heart allografts. We now report that NDA 4 is conserved across primate species (human, baboon, and gibbon), and that it serves as a useful marker for targeting immunotoxins to activated lymphocytes.

Materials MATERIALS AND METHODS

Fresh lymphocytes and cell lines. The human Jurkat T-cell, Raji B-cell, SKW B- cell, and gibbon MLA- 144 T-cell lines were obtained from the American Type Culture Collection and maintained in RPM1 1640 media supplemented with 10% fetal calf serum (FCS), 2 mA4 glutamine, 1 mM nonessential amino acids, 1 mM sodium py- ruvate, 50 pg/ml gentamycin (GIBCO, New York), 0.05% (w/v) penicillin, and 0.05% (w/v) streptomycin. These cell lines were analyzed by flow cytometry and shown to be >95% NDA 4 positive. Human and baboon peripheral blood mononuclear cells (PBMC) were prepared by Ficoll-Histopaque gradient centrifugation and suspended in RPM1 1640 medium supplemented with 10% FCS, 2 mM glutamine, 1 mA4 non- essential amino acids, 1 mM sodium pyruvate, and 50 pg/ml gentamycin (GIBCO, New York) for culture. Purified T and B cell suspensions were prepared from PBMC by first depleting the cells of monocytes by plastic adherence and then rosetting the T cell fraction with neuraminidase-treated sheep red blood cells. Some T and B cell suspensions were prepared using Lympho-kwik T and Lympho-kwik B as suggested by the manufacturer (One Lambda, Inc., Los Angeles, CA). Activated T cells were generated by adding phytohemagglutinin (2 /*g/ml) to cultures of T cells for 48 hr. B cells were activated by adding Staphylococcus aureus Cowan strain 1 (SAC2 0.01%) to 48-hr cultures.

’ Abbreviations used: SAC, Staphylococcus aureus cowan strain I; PHA, phytohemagglutin; PWM, Poke- weed mitogen; PBMC, peripheral blood mononuclear cells; LBCL, lymphoblastoid cell lines; BSA, bovine serum albumin; 3H-TdR, tritiated thymidine; EDTA, ethylene diamine tetraacetate; SDS, sodium dodecyl sulfate; FITC, fluorescein isothiocyanate; mAb, monoclonal antibody; HPLC, high performance liquid chro- matography.

IMMUNOTOXIN TO LYMPHOCYTE ACTIVATION ANTIGEN 87

FACS analysis. Baboon PBMC were obtained as a gift from Dr. Craig Smith from the Department of Surgery, College of Physicians & Surgeons of Columbia University. PBMC were prepared by buoyant density centrifugation on Ficoll-Histopaque and washed in Hank’s balanced salt solution (HBSS). The PBMC were then incubated with 2.2 pg/ml mAb NDA 4 or with an isotype control of irrelevant specificity, for 1 hr at 4°C. The cells were washed three times with HBSS and then stained indirectly with fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse immunoglobulin (Ig) (Tago, Burlingame, CA) for 1 hr at 4°C. Cells were washed three times in HBSS and analyzed on a Becton-Dickinson fluorescence automated cell sorter (FACS) Star (Mountain View. CA) Flowcytometer. Fluorescent and forward scatter measurements were made on >20,000 cells and analyzed using the Consort 30 program package.

Pur$cation and radiolabeling of mAb ND,4 4. MAb NDA 4 was purified from hybridoma culture supernatants by protein A affinity chromatography and labeled with lz51 by the Iodogen method to a specific activity of 5 X I O5 cpm/pg as described by the manufacturer. Labeled antibody was separated from free iodine by ultrafiltration (Centricon 100, Amicon) and gel filtration (G-25 Econocolumn. Bio-Rad).

Preparution ofimmunoto,\-in. Purified gelonin, N-succinimidyl 3-(2-pyridyldithio) propionate (SPDP) and Iodogen were obtained from Pierce (Rockford, IL). Sodium (12’1) iodide and 3H-TdR were obtained from New England Nuclear (Bedford. MA). Conjugates were prepared and purified essentially by the method described by Marsh (8). Gelonin ( 10 mg/ml in 0.12 M borate. pH 8.0) was reacted with a 5 M excess of 2-iminothiolane for 90 min at room temperature as described (9). The thiolated toxin was desalted on G-5 column (Pierce), washed, and concentrated to its original volume in Centricon 30 microconcentration cartridge (Amicon). The number of free thiols (I-2-SH/molecule) was determined by Ellman Reagent. Purified mAb NDA 4 (2.0 ml at 10 pg/ml in 0.12 A4 borate, pH 8.0) was reacted with a 5 M excess of h:-succinimidyl 3-(2-pyridyldithio) propionate (1 pg/ml in dimethyl sulfoxide). After 20 min, the mod- ified proteins were desalted and added to a IO M excess of the thiolated toxin. A small aliquot of the SPDP modified mAb was removed. treated with 1 M dithiothreitol. and the concentration of free pyridyl groups was determined by the absorbance at 403 nm. The SPDP modification of MAb NDA 4 introduced one to two pyridyl groups per antibody molecule. After 1 hr. residual thiols were reacted with iodoacetamide. The conjugate was desalted by gel filtration. washed, and concentrated (Centricon 100). The conjugate was purified by high performance gel filtration on a 2 1.5 X 600- mm TSK-G 3000 column. The major conjugate peak was collected, concentrated, and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS- PAGE). This sample peak was contaminated with ~5%) of unreacted mAb as deter- mined by silver staining. No free gelonin was observed.

SDS-PAGE analysis of‘radiolubelc~d mAb NDA 4 und rnAb ND24 4-gelonin con- jrdggate. The mAb and immunotoxin were labeled with “‘1 and precipitated with tri- chloroacetic acid, The precipitates were washed with ethanol/sodium acetate and dried at room temperature. Precipitated samples and the Bio-Rad high molecular weight standards were incubated for 3 min at 95°C in nonreducing SDS sample buffer and applied to a 5- 10% polyacrylamide gradient gel. Samples were electrophoresed at 10 mA overnight. The gel was dried and autoradiographed overnight at -70°C.

Satzrrabilir!~ of’mAb NDA 4 binding. Cultured Raji cells were washed three times in HBSS, reacted with increasing concentrations of “‘I-labeled mAb NDA 4. and

88 HARRIS ET AL.

incubated for 60 min at 4°C. Cells were washed three times with cold binding buffer and the specifically bound counts were determined. The means of triplicate deter- minations and the standard deviations of the means were calculated.

Specificity of mAb NDA 4 binding. Cultured Raji cells were washed three times in HBSS, and reacted with 3 X lo4 cpm of 12%-labeled mAb NDA4 and with different amounts of unlabeled mAb NDA4 for 60 min at 4°C. Cells were then washed three times with cold binding buffer and the specifically bound counts were determined.

Scatchard analysis of mAb NDA 4 binding. The number of antigenic sites per cell was determined as follows: Cells were washed three times at 4°C in binding buffer and resuspended at lo7 cells/ml. Various concentrations of labeled mAb alone or labeled mAb with unlabeled mAb were added in 0.5 ml to an equal volume of cell suspension. After 4 hr of incubation at 4°C the supernatants were harvested and counted. The cells were washed three times in binding buffer, lysed in 100 ~1 of 0.5 M NaOH, and analyzed in a Packard MultiPrias Gamma Counter. The number of antigenic sites was estimated by Scatchard analysis which extrapolates the binding curve to a B/F ratio of zero.

Competitive binding assay of mAb NDA 4-gelonin immunotoxin. NDA 4 positive Raji cells ( lo6 cells/ml in 0.1 ml HBSS) were incubated with FITC-mAb NDA 4 and increasing amounts of mAb NDA 4-gelonin conjugate (0.02 and 20.0 pg). Cells were washed and the binding of the fluorescent mAb was measured by FACS analysis.

Cytotoxic efects of mAb NDA 4-gelonin conjugate on the SKW lymphoblastoid cell line and on the MLA gibbon leukemia cell line. SKW or MLA-144 cells were plated in 96-well Linbro flat bottom trays at a concentration of 2 X lo4 cells in 0.2 ml of complete medium. Cultures received immunotoxin conjugates, toxin alone, or mAb alone at the indicated concentrations. After 48 hr the cultures were labeled with 3H- TdR and harvested 18 hr later. Thymidine incorporation was determined by scintil- lation counting. In parallel experiments, the cytotoxic effect of the immunotoxin was blocked by the addition of various concentrations of unconjugated mAb to the 96- well cultures. Cells were cultured for 72 hr, then stained with the vital dye fluorescein diacetate and analyzed by FACS. The means and standard errors of cpm and of the percentage viability in triplicate wells were calculated.

RESULTS

Relative Levels of NDA 4 Expression on Activated PBL and Cell Lines

To determine the level of expression of NDA 4 on lymphocyte membranes we first ascertained the specificity and saturability of mAb binding to cell membrane and then calculated the number of binding sites by Scatchard analysis. The radiolabeled mAb NDA 4 was subjected to nonreducing SDS-PAGE to assess the integrity of the antibody molecule. A major band of MWr 160 kDa was observed, consistent with the presence of IgG (Fig. 1). To investigate the binding of radiolabeled mAb NDA4 to SKW and Raji cells we added increased amounts of labeled antibody to the cells. A plateau was reached at 100 pg/ml (Fig. 2). Binding of radiolabeled mAb NDA 4 to the cells was efficiently blocked by adding increasing amounts of nonlabeled antibody to the sus- pensions. The addition of 500 pg nonlabeled antibody inhibited completely the binding of radiolabeled mAb NDA 4 (Fig. 3). The fact that the unlabeled antibody blocked

IMMUNOTOXIN TO LYMPHOCYTE ACTIVATION ANTIGEN

A B

MY

.

205

89

116

FIG. I. SDS-PAGE analysis of radiolabeled mAb NDA 4 and mAb NDA 4-gelonin conjugate. Lane A. mAb NDA 4. Lane B, mAb NDA 46gelonin immunotoxin.

the binding of the labeled antibody in a dose-related manner indicates that radioio- dination did not alter the binding property of mAb NDA 4.

To determine the number of binding sites on the cell membrane of resting, activated, and transformed lymphocytes we incubated cell suspensions with various concentra-

30000 1

.Ol .1 1 10 100 1000

ug 1251 labeled mAb

FIG. 2. Saturability of mAb NDA 4 binding. Cultured Raji cells were reacted with increasing concentrations of ‘251-labeled mAb NDA 4. The specifically bound counts were determined. Each point represents the mean of triplicate determinations. The standard deviations of the means were less than IOq,.

90 HARRIS ET AL.

0 1 10 100 1000

ug unlabeled antibody

FIG. 3. Specificity of mAb NDA 4 binding. Cultured Raji cells were reacted with ‘251-labeled mAb NDA 4 and the indicated amount of unlabeled mAb NDA 4.

tions of radiolabeled mAb NDA 4 and measured the number (mean cpm) of bound and free counts in duplicate reactions. Binding curves generated from Scatchard analysis of ‘251-mAb NDA 4 binding to resting B and T cells showed that these cells have 400- 800 sites/cell (Fig. 4). Two days following PHA activation, however, the expression of NDA 4 and on T cells increased to approximately 10,000 to 20,000 sites per cell. The Raji and SKW lymphoblastoid B cell lines, which are constitutively NDA 4 positive, expressed approximately 40,000 to 80,000 sites/cell. The binding of ‘251-mAb NDA 4 to Staphylococcus aureus Cowan strain 1 activated B cells was variable and showed a high level of unsaturable binding, making Scatchard analysis of the number

0.8

1 8 t . ‘0 c

; m

Resting B

Resting T

Activated T

Raji Cell Line

10 100 1000 10000 100000

Bound (pMolar)

FIG. 4. Scatchard analysis of mAb NDA 4 binding. Cells were incubated with various concentrations of ‘251-labeled mAb NDA 4. The specifically bound and free counts were determined. Each point is the mean of duplicate samples.

IMMUNOTOXIN TO LYMPHOCYTE ACTIVATION ANTIGEN 91

ofbinding sites unreliable. Fluorescence measurements of FITC-mAb NDA 4 binding to these cells, however, show that they are strongly positive for NDA 4 expression (6).

Characterization and in Vitro Cjwtoxicitv of’NDA 4 Immunotoxin

The mAb NDA 4-gelonin conjugate was analyzed by nonreducing SDS-PAGE. The molecular weight of the major species. 2 10 kDa, is consistent with the introduction of two gelonin molecules per molecule of mAb NDA4 (Fig. 1). The binding properties of the immunotoxin were also tested in a competitive inhibition assay. NDA 4 positive Raji cells were reacted with FITC-mAb NDA 4 and various concentrations of mAb NDA 4-gelonin immunotoxin and the binding of the fluorescent mAb was measured by FACS analysis. Increasing concentrations of conjugate reduced the binding of FITC- mAb NDA 4-stained Raji cells in a dose-dependent manner demonstrating that con- jugation of gelonin to the antibody did not significantly affect the binding properties of mAb NDA4 (Fig. 5). The in vitro cytotoxicity of the mAb NDA 4-gelonin conjugate was evaluated on the SKW human lymphoblastoid cell line and on the MLA-144 gibbon T cell leukemia cell line. These lines were chosen because of their stable con- stitutive expression of NDA 4 and ease of culture. The proliferative response of the cells was measured after incubation with the mAb NDA 4-gelonin conjugate. and the dose-response curves for each cell line, incubated in the presence of conjugate. toxin, or mAb alone were determined (Figs. 6 and 7). The concentrations ofconjugate. toxin, or mAb NDA 4 required to inhibit ‘H-TdR incorporation by 50% (I&) are shown in Figs. 6 and 7. When mAb NDA 4 was added alone to cultures of SKW or MLA cells lines there was no inhibition of 3H-TdR incorporation over a 48-hr incubation period. Free gelonin reduced incorporation with an lC5,, of 0.1 and 0.5 PM for SKW and MLA cells, respectively. Conjugation of gelonin to mAb NDA 4 conferred a 1 OOO- fold greater cellular toxicity over the toxin alone in both cell lines tested. In a parallel assay, we measured the effects of the conjugate on the actual viability of SKW cells

8000 1

- FITC mAb

2000 - -----.... -.- FITC mAb+lT 2

FITC mAb + IT 2.0

Fluorescence Channel Number

FIG. 5. FACS analysis of specificity of mAb NDA 4-gelonin conjugate binding to Raji cells: competitive binding assay. NDA 4 positive Raji cells were incubated with FITC-mAb NDA 4 and increasing amounts of mAb NDA 4-gelonin conjugate (0. 0.2. and 2.0 pg). The ordinate shows the number of cells binding FITC-conjugated mAb NDA 4.

92 HARRIS ET AL.

- NDA4Gel

- Gel

Log Dose (uM)

FIG. 6. Cytotoxic effects of mAb NDA 4-gelonin conjugate on the SKW lymphoblastoid cell line. Cells were incubated with the conjugate (+), gelonin (m), or mAb alone (‘Cl) for 48 hr. 3H-TdR incorporation was determined during the last 18 hr of incubation. Points are the means of triplicate measurements. Error bars are the standard deviation of the measurements. The concentrations (NIV) required to inhibit the proliferation of the SKW cell line by 50% were conjugate, 0.001; toxin, 0. I; and mAb > 1000.0.

after a 72-hr incubation period (Fig. 8). The toxicity of the conjugate was abrogated by unconjugated mAb NDA 4, which protects SKW cells against the toxic effects of the conjugate. This indicates that the effect of the conjugate is specific and determined by the antibody binding to the cells. In additional control experiments we measured the viability of resting B and T cells cultured with various concentrations of conjugate. Following 72 hr of incubation, we found the I& of the conjugate to be >500 PM. Thus, resting lymphocytes are significantly less sensitive to the effects of the conjugate compared to lymphoblasts, as expected, since they express significantly lower levels of NDA 4 on their membrane.

40000

E ‘L 2 30000

ii E ; 20000 - NDA4

-El

5 z 10000 m

0 -6 -4

Loi 2Dose (EM) 2 4

FIG. 7. Cytotoxic effects of mAb NDA 4-gelonin conjugate on the MLA gibbon leukemia cell line. The experiment was performed as in Fig. 6. The concentrations (PM) required to inhibit the proliferation of the MLA ceil line by 50% were conjugate, 0.0005; toxin, 0.5; and mAb > 100.0.

IMMLJNOTOXIN TO LYMPHOCYTE ACTIVATION ANTIGEN 93

z 500 7

0 0” 5.002 mg/ml conjugate with 400 - tiarious concentration of mAb

. 300-

t a E 200 -

1 100 - mAb NDA 4 - gelonin toxin conjugate alone

=

6 0 I I I I 00001 .OOOl .OOl 01

Concentration (mc~/ml; 10

FIG. 8. Specificity of mAb NDA 4-gelonin toxicity: rescue ofthe cytotoxic effects of mAb NDA 46gelonin conjugate by unconjugated mAh NDA 4. The ordinate shows the concentration of toxin or mAb NDA 4 added to 72-hr cultures. The abscissa shows the number of wahle cells staining with the vital dye fluorescein diacetate. Each point is the mean of triplicate measurements. Error bars are the standard deviation of the measurements.

Peripheral blood lymphocytes were obtained from normal baboons and baboons receiving orthotopic cardiac allografts and immunosuppressive therapy. Failure of immunosuppressive therapy (steroids and cyclosporin A) in these recipients results in rejection of the graft which is diagnosed by biopsies showing lymphocytic infiltrates and myocardial immunoglobulin deposition (7). We obtained PBL from quiescent (nonrejecting) transplanted baboons receiving cyclosporin A and steroid therapy and from an animal undergoing severe cardiac allograft rejection and measured NDA 4 expression by immunofluorescence flow cytometry. Fluorescence and forward scatter measurements demonstrated that the blast population of PBL from the rejecting ba- boon, representing 25% of PBL, was NDA 4 positive (Fig. 9). A comparison of NDA 4 expression on PBL from normal and quiescent baboons receiving cyclosporin A and steroids showed no difference. An isotype (IgG 1) control mAb of irrelevant spec- ificity was also tested and showed no reactivity with the baboon lymphocytes (data not shown).

DISCUSSION

The Scatchard analysis of mAb NDA 4 binding has shown that activated T cells have a lOOO-fold higher expression of NDA 4 than resting T cells. The in vitro cyto- toxicity studies of the immunotoxin prepared from mAb NDA4 and gelonin have demonstrated that high expression of NDA 4 renders a cell sensitive to the toxic effects of the conjugate. Cells which express very low levels of NDA 4 (resting T and B cells) were insensitive to the toxic effects of the conjugate. During allograft rejection in a primate model, activated lymphocytes expressing NDA 4 were detected by immu- nofluorescent flow cytometry. This indicates that elevated expression of NDA 4 is consistent with rejection and that immunotoxins reactive with NDA 4 may be useful for selectively depleting activated B and T cell populations participating in graft re- jection. The in vitro cytotoxicity experiments suggest that resting B and T cells are less susceptible than activated B and T cells to the toxic effect of the conjugate, and. thus, likely to be spared during in vivo immunotherapy.

94 HARRIS ET AL.

’ 0 32 64 96 126 160 192 224 256

0 CSA+St Con

6

0 32 64 ‘)6 ’ 28 160 192 224 256

Fluorescence Channel Number

FIG. 9. Expression of NDA 4 during heart allograft rejection. Peripheral blood mononuclear cells from normal, ahotransplanted cyclosporin-steroid treated quiescent or rejecting baboons were stained with FITC- labeled mAb NDA 4 and analyzed by flow cytometry.

It is possible that such an approach will stimulate natural suppression by ablating the population of activated lymphokine producing cells and of effector cells. A decrease in the size of the immunoreactive populations may allow induction of tolerance to antigens during mAb-mediated immunosuppression. If applicable to autoantigens which stimulate autoimmune diseases and to alloantigens that stimulate rejection, this could represent a useful approach to reestablish tolerance to self and to allogeneic MHC, respectively (10). Although we have not explored yet the range of T and B malignancies which express NDA 4, it is likely that some of them will turn out to be positive and, thus, susceptible to the immunotoxin.

Understanding of mechanisms of NDA 4-triggered signal transduction in T and B lymphocytes (by specific kinases, phosphatases, and G proteins) will be required, how- ever, before considering this antibody as a candidate for immunointervention in human autoimmune diseases, transplantation, or lymphoid malignancies.

ACKNOWLEDGMENTS

This work was supported by the following grants from the PHS: 5-ROl-A125210, 5-POI-HL36581, and l-ROI-HD22920-OlA. The authors thank Dr. FIoria Lupu for her help in preparing the manuscript and Elly Cristea for her technical assistance.

IMMLJNOTOXIN TO LYMPHOCYTE ACTIVATION ANTIGEN 95

REFERENCES

1. Knap, W., Rieber, P., Dorken, B., Schmidt, R. E., Stein. H., and Borne, V. D.. Immune/. 7’oduy 10, 253. 1989.

2. Chantenoud, M.. and Bach, J. F., Curr. @in. Immunol. 1, 235. 1988. 3. Mandel, T. E.. Immunol. Toduy 10, I, 1989. 4. Suciu-Foca, N.. Reed, E., Rubinstein, P.. MacKenzie, W., Ng. A. K. and King, D. W.. Natuw (London)

318,465, 1985. 5. Suciu-Foca. N.. Rosochacki, S. J., Cai. J., Reed, E.. Rubinstein, P., and King, D. W., Cell. Immunol.

110,265, 1987. 6. Suciu-Foca. N.. Cai. J.. Gutierrez, C., Reed, E.. Morrison, S.. Pernis. B., and King. D. W., J. Immunol.

140, 395, 1988. 7. Harris. P., Strba-Checova. K.. Rubinstein, P., Mann. D., King, D. W.. and Suciu-Foca, N. J.. Clm.

Inwxt. 85, 746. 1990. 8. Marsh. J. J.. Rio/. Chem. 263, 15.993. 1988.

9. Stirpe. F.. Olsnes. S., and Pihl, A.. J. Biol. Chem. 256, 6947. 1980. 10. Adorini. L.. Barnaba, V., Bona. C., Celada, F., Lanzavecchia. A., Sercarz. E., Suciu-Foca, N., and

Wekerle. H.. Immunol. Toduy 11, 383. 1990.