J N e u r o s u r g 72:476-481, 1990

Induction of cytotoxicity in human T cells coated with anti-glioma x anti-CD3 bispecific antibody against human glioma cells TAIZO NITTA, M.D., KIYOSHI SATO, M.D., K o OKUMURA, M.D., AND SHOZO ISHII, M.D. Departments of Neurosurgery and Immunology, Juntendo University School of Medicine, Tokyo, Japan ~" A bifunctional hetero-F(ab')2 antibody fragment was developed that contained the Fab portions from antiCD3 and anti-glioma monoclonal antibodies. The antibody simultaneously recognized two different molecules, the CD3 complex on effector T cells and a human glioma-associated antigen; thus, it could cross-link effector and target cells. This bispecific F(ab')2 fragment induced peripheral blood mononuclear cells (PBMC's) from healthy donors to lyse cells of the human glioma cell line, U251MG, which are resistant to natural killer cellmediated cytolysis. The effect of the bispecific antibody on lymphokine-activated killer (LAK) cell activity was tested in patients suffering from malignant glioma. For this study, PBMC's from these patients were preactivated with recombinant interleukin-2 and their killer activity against U251MG cells was investigated in vitro with and without the bispecific antibody. The LAK cell activity of the PBMC's from patients with malignant gliomas was found to be suppressed compared with those of healthy donors. However, after preincubation with bispecific antibody, the patients' LAK cells exhibited marked cytolytic activity against U251MG cells. These fndings suggest that this bispecific antibody may be a useful addition to anti-glioma immunotherapy. KEY W O R D S 9 brain neoplasm lymphokine-activated k i l l e r cell

A

9 glioma

NAPLASTIC glioma and glioblastoma multiforme are refractory to various types of adjuvant therapy following radical resection and radiotherapy. m4'3~ Kitahara, et al., 7 recently treated patients suffering from malignant glioma with lymphokine-activated killer (LAK) cells and cytotoxic T cells. 3'19 G r i m m , et al.,4 originally described L A K cells as interleukin-2 (IL-2)-activated effector cells that appeared to be clearly distinct from natural killer (NK) cells because they could lyse N K cell-resistant t u m o r cells and autologous fresh t u m o r cells without limiting components of the major histocompatibility complex. Kitahara, et al., used an O m m a y a reservoir to administer LAK cells, obtained by incubating peripheral blood mononuclear cells (PBMC's) from glioma patients with h u m a n recombinant IL-2 (rlL-2) into the cavity remaining after t u m o r resection. Although L A K cells can lyse glioma cells in vitro, no clear benefit has been demonstrated from the use of L A K cell therapy in vivo. The failure of L A K cell therapy in the treatment of malignant gliomas m a y be due to the following problems. First, if 476

9 monoclonal

antibody

9 immunotherapy

9

the protocol of Rosenberg, eta/., ~9'2~is used, 101~ ~I L A K cells are required for intracerebral infusion, and it is technically difficult to prepare such a large number of lymphocytes. Second, intracerebral infusion of such large numbers of cells has been reported to cause postinfusional hydrocephalus. 7 A third problem arises if infused L A K cells do not recognize the target cells and disperse throughout the cerebrospinal fluid space, where they can have adverse affects. In view of these problems, there is an obvious need to modify L A K cell therapy in order to enhance the affinity, specificity, and cytotoxicity o f the killer cells. Recent studies have shown that a bispecific antibody, comprising an antibody against a c o m p o n e n t of the Tcell receptor complex or CD3 complex (both are intimately related) cross-linked to an antitumor antibody, can cause T cells to specifically lyse target cells. 6,9A4A5 The bispecific antibody links the target cell directly to CD3 molecules on the effector T cell, thus enhancing the lyric activity and binding o f the T cell to the target cell and facilitating lysis. 26'28'29This study demonstrates J. Neurosurg. / Volume 72/March, 1990

Cytotoxicity induced by bispecific antibody against glioma that this strategy can be used to target peripheral blood T cells from glioma patients against human glioma cells.

F (ab')=

Materials and Methods

I Reduction Fab'-thiol(DTT)

Preparation of Monoclonal Antibody For the anti-human CD3 monoclonal antibody (MAb), we used OKT3, which is a mouse immunoglobulin (Ig)G2, subclass. To obtain the OKT3 MAb, Pristane (2,6,10,14-tetramethylpentadecane)-primed Charles River mice (male BALB/C), aged 4 to 6 months, were injected with OKT3 hybridoma cells (1 • 10 7 cells). One or 2 weeks later, the mice were sacrificed and ascitic fluid was aspirated. Purified MAb was isolated by a m m o n i u m sulfate precipitation followed by Sepharose protein A chromatography. The anti-glioma MAb (LU 246),* which binds a 145-kD glycoprotein cell-surface antigen expressed on human glioma and small-cell lung cancer lines and also a mouse IgGl subclass, was received in ascitic fluid? It was purified from the ascitic fluid by the same procedure as was used for OKT3.

Effector and Target Cells for Cytotoxicity Assay The human glioma cell lines A172 and U251MG were used as target cells. Both cell lines were grown in culture medium (RPMI 1640 containing 10% fetal calf serum, penicillin (100 U/ml), streptomycin (100 ~g/ ml), 10 m M HEPES, and 25 gM 2-mercaptoethanol). H u m a n PBMC's were isolated by Ficoll-Hypaque gradient centrifugation from the heparinized blood of healthy donors or from t u m o r patients. The tumors were histologically verified to be anaplastic astrocytoma and/or glioblastoma multiforme. The cells were suspended in culture medium at a density of 1 • 106 cells/ ml. To induce L A K cell activity, PBMC's prepared as above were incubated in culture medium containing 100 U/ml of rlL-2 for 3 days at 37"C, in a 5% CO2 incubator.

Preparation of Bispecific F(ab')2 Fragments Heteroconjugation of two different MAb's (OKT3 and L U 246) was performed using the F(ab')2 fragments of each, as described previously ~5 (Fig. 1). The F(ab')2 fragments were composed of dimeric Fab devoid of the Fc portion of IgG and had a molecular weight of 100 kD. The F(ab') fragments of L U 246 were obtained by digesting intact IgG~ with preactivated papain according to the method of ParhamJ 6 The F(ab')2 fragments of OKT3 were prepared by digesting intact IgG2~ with pepsin rather than with papain (which digests mouse IgG2~ completely into Fab fragments). After digestion, 100-kD material was isolated by highperformance liquid chromatography (HPLC) using a

* Anti-glioma MAb was a kind gift from Dr. S. Hirohashi, National Cancer Institute, Japan.

J. Neurosurg. / Volume 72/March, 1990

F (ab')=

derivatives

I Reduction (DTT)

bispecific antibody

FIG. 1. Reaction scheme by bispecific hetero-F(ab')2 from two different F(ab')2 fragments. F(ab')2 fragments (OKT3 and LU 246), digested by pepsin or papain, were both reduced to Fab-SH fragments. One Fab-SH fragment was incubated with nitrobenzoic acid to Fab-S-NB, and then both were reacted together. DTT = dithiothreitol; DTNB = 5,5'-dithio bis 2-nitrobenzoic acid. TSK 3000 S W X L column. To examine the purity of the F(ab')2 fragments, their molecular weight was ascertained by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) under nonreducing conditions. 8 Hetero-F(ab')2 fragments were produced according to a previously described procedure. In brief, each F(ab')2 fragment was reduced to the Fab-SN state with 0.5 m M dithiothreitol in 0.1 M phosphate-buffered saline (PBS) for 30 minutes at room temperature (Fig. 1). One Fab-SN fragment was then further incubated with 5 m M 5,5'-dithio bis 2-nitrobenzoic acid (DTNB) 2,15in PBS for 30 minutes at room temperature to the Fab-S-NB stage. This Fab-S-NB fragment and the other reduced Fab-SH fragment were then mixed together at a ratio of 1:1 and were incubated for 4 hours at 37"C. The hetero-F(ab')2 antibodies were then purified by size-exclusion HPLC.

Flow Cytometry Flow cytometry was used to test whether the heteroF(ab')2 fragments could bind to both CD3 and the human glioma-associated antigen. Hetero-F(ab')2 fragments (1 ug) were incubated with either fresh PBMC's ( 1 x 106) or trypsinized U 2 5 1 M G glioma cells (5 x 105) for 30 minutes at 4~ After being washed twice with PBS, cells were reacted with 1 /~g fluorescein isothiocyanate-conjugated goat anti-mouse IgG, then analyzed by flow cytometry.t T o determine whether this product

t FAC star flow cytometer manufactured by Becton-Dickinson, Mountain View, California. 47'7

T. Nitta, et al. was predominantly hetero-cross-linked, F(ab')2 fragments without contamination of homoconjugated fragments (OKT3 • O K T 3 or L U 246 • LU 246), U 2 5 1 M G cells (1 • 106) were incubated with 1 gg prepared bispecific F(ab')2 for 30 minutes at 4~ followed by centrifugation at 1200 rpm. The supernatant was collected and tested for anti-CD3 activity by flow cytometry. Conversely, reactivity to U 2 5 1 M G cells was checked by using supernatants in which the bispecific antibodies were absorbed by T cells. In Vitro Cytotoxicity A s s a y A standard 4-hour 5~Cr release assay was performed according to a previously published procedure. 15,17,18In brief, target glioma cells (1 • 105 cells/ml) were labeled with 100 uCi of s ~Cr (Na25 ~CrO4) for 60 minutes at 37~ Cells were then washed with m e d i u m and incubated with bispecific antibodies (1 ug) for 30 minutes in a 96well microtiter plate. Each well contained 104 target cells in a 100-ul volume. Appropriate numbers of effector cells were then incubated for 4 hours at 370C in a 5% CO2 incubator. After centrifugation, 100 ul of supernatant was removed and counted for radioactivity. Cytotoxicity was calculated using the following formula: (cpm, experimental) - (cpm, spontaneous) • 100, (cpm, maximum) - (cpm, spontaneous) where m a x i m u m lysis was determined by adding 10 ul of 20% Triton X to the target cells. Each value reported represents the average o f triplicate determinations with a standard error of the m e a n of less than 5 %. Results

Fab-S-NB were fractionated by H P L C gel chromatography. The earlier peak, eluted at 14 minutes, was that of the F(ab')2-sized molecules, which represented between 60% and 70% of the final product. The later peak included the unreacted Fab-sized molecules. Material eluting in the F(ab')2 peak was ascertained by SDS-PAGE to be nearly pure; this can be seen in Fig. 3, in which Lane 4 shows a signal band (100 to 110 kD) with the same size as the F(ab')2 fragments of O K T 3 IgG digested by pepsin (Lane 2). To examine whether the F(ab')2 fragments eluted were pure heterodimers and not homodimers, their reactivity with both CD3 positive T cells and U251MG cells was investigated by indirect immunofluorescence. As Fig. 4 shows, the F(ab')2 fragments were found to be bifunctional: that is, reacting not only with T cells but also with U 2 5 1 M G cells. Next, a target absorption assay was performed to rule out the existence of contaminating h o m o - F ( a b ' ) 2 fragments. This showed that absorption o f the hetero-F(ab')2 by glioma cells concomitantly removed the anti-CD3 activity and vice versa, indicating that the molecules were indeed true hetero-F(ab')2 fragments rather than a mixture of homoconjugated F(ab')2 fragments. The functional capacity of this bispecific antibody was then evaluated using anti-CD3 x anti-glioma (OKT3 • L U 246) hybrid antibody to mediate the lysis of 5~Cr-labeled U 2 5 1 M G cells by h u m a n PBMC's. Figure 5 shows that the bispecific F(ab')2 fragments invoked efficient lysis of the target glioma cells. This was especially noticeable for U 2 5 1 M G cells (Fig. 5 right), in which PBMC's lysed the target cells significantly in the presence of bispecific antibody, but did not mediate

Figure 2 shows a typical elution profile obtained when the products of a reaction between Fab-SH and

FIG. 2. Elution profile of heteroconjugated hybrid antibodies (OKT3 x LU 246) by size-exclusion high-performance liquid chromatography on a TSK 3000 SWXL column equilibrated with 0.1 M phosphate, 0.15 M NaC1 (pH 7.5), at a flow rate of 0.5 ml/min. The F(ab')2 fraction shows the bound hetero-F(ab')2 and unbound Fab fraction. 478

FIG. 3. Analysis of sodium dodecyl sulfate-polyacrylamide gel electrophoresis of heteroconjugated antibodies. Lane 1: Intact immunoglobulin G (OKT3); Lane 2:F(ab')2 fragments digested by pepsin; Lane 3: l a b reduced from F(ab')2 by dithiothreitol; and Lane 4:F(ab')2 coupled by incubating FabSH with Fab-S-NB. Numbers at the left indicate positions of standards (kD).

J. Neurosurg. / Volume 72/March, 1990

Cytotoxicity induced by bispecific antibody against glioma PBL

U251MG

350

50

550

.m o

"i

100

,

=ii

10 ~

102 FL. 1

103

Control - - Bispecific F(ab')2 ...... OKT3 (IgG) ......

."

:

i\

......

' ';' ''n "' '~' '"~!"1~ ' '~'~',,,'1 100 10~ 102 103 F'L 1 ...... Control - - Bispecific F(ab')2 ...... LU2460gG)

FXG. 4. Results of flow cytometric analysis using bispecific hetero-F(ab')2. Fresh peripheral blood mononuclear cells (PBL, left) or U251MG (right) were reacted with 1 ~zg of bispecific hetero-F(ab')2 (OKT3 x LU 246) or intact immunoglobulin (Ig)G (OKT3 left, LU 246 right), following reaction with 1 pg fluorescein isothiocyanate-conjugated goat antimouse IgG.

lysis without the antibody. It is known that PBMC's have a capacity to mediate antibody-dependent cell cytotoxicity (ADCC) activity, but this prepared hybrid antibody possessed no Fc fragments so, in this case, ADCC activity could not have mediated lysis. In a second experiment, the concentration of bispecific antibody required to achieve lysis was examined. Figure 6 shows that m a x i m u m lysis was achieved at a concentration of about 1 pg/ml of hetero-F(ab')2 fragments. Even when the effector:target ratio was titrated down to 0.1, the U 2 5 1 M G cells were lysed efficiently by h u m a n P B M C ' s in the presence of this antibody (Fig. 7). It was quite surprising that only one-tenth of the n u m b e r o f effector cells could lyse nearly 40% of the target glioma cells in vitro, a result which was not seen when only L A K cells were used as the effector cells. 4 We next explored the in vitro generation of antiglioma killer activity in PBMC's obtained from six patients with malignant glioma (Fig. 8). The L A K cell

% lysis

20.

7~

o~

/

40 30 20 [

[]

'~ / A~ i

0.05

89

/

-[]

z~/zx-~

A' i

0.I

i

0.5

I

1.0

I

3.0

1

5.0

Bispecific Ab, pg/ml

FIG. 6. Graph showing the concentration of bispecific F(ab')2 fragments titrated in 4-hour 5~Cr release assay against A172 (triangles) and U251MG (squares) glioma cells as described in Fig. 5 legend. Effector:target ratio was 5:1. Ab = antibody.

activity (that is, cytotoxicity) induced by 3 or 4 days of stimulation with IL-2 was suppressed in most patients more than in healthy donors. However, in the presence of the hetero-F(ab')2 antibody, cytotoxicity was improved and lysis o f glioma cells by PBMC's from all six patients was greatly enhanced, especially in Case 3 (Fig. 8).

Discussion Fresh peripheral T cells without any antigenic or IL2 stimulation lack cytolytic activity; however, binding of the T-cell receptor-CD3 complex with its relevant MAb results in de novo generation of cytotoxicity and concomitant synthesis of h u m a n perforin. 12J9"3~ Thus,

80

% lysis

50

FIG. 5. Bispecific hybrid antibody-dependent lysis of hum a n glioma cells. Target cells: glioma cell lines A172 (left) and U 2 5 1 M G (right).Effector cells: Fresh peripheral blood mononuclear cells collected from a healthy donor by Conray-

Hypaque density gradient centrifugation (control, white bar) and bispecific F(ab')2 (OKT3 x LU 246, black bar) were added to each well at a final concentration of 1 pg/ml. Effector:target ratio was 10:1.

J. Neurosurg. / Volume 72/March, 1990

_& 6 o 0 0

~_ 4 o to

2O

0.1

0.5

1

5

10

20

Effector to Target ratio

FIG. 7. Graph showing titration of effector: target ratio in cytotoxicity as described in Fig. 5 legend, with (circles) and without (triangles) 1 ug/ml of hetero-F(ab')2 fragments. U251MG cells were used as target cells. 479

T. Nitta, et al. anti-CD3 or anti-T-cell receptor MAb's have an antigen-mimicking effect, with the CD3 complex behaving as a trigger molecule in T cell-mediated cytotoxicity. ~3.1a It has been shown that h u m a n T cells from healthy donors can actively lyse target ceils without major histocompatibility complex restrictions when they are pretreated with bispecific antibodies (anti-CD3 • antitarget). 17 Like the CD3 complex in T cells, the CD16 (Fcr receptor III) complex in N K cells plays a key role in ADCC activity. Bispecific antibodies (anti-CD16 • anti-target) can also render N K cells cytotoxic to target cells. 27 To date, results with adoptive immunotherapy using LAK cells against glioma have been disappointing. Therefore, we prepared bispecific antibodies composed of anti-CD3 and anti-glioma F(ab')2 fragments in an attempt to improve the effectiveness of LAK cell therapy. This bispecific antibody induced human PBMC's to specifically lyse malignant glioma cells in vitro, even in conditions with a low effector:target ratio. Other authors have used the anti-transferrin receptor or anti-hormone receptor MAb's instead of anti-tumor associated antigens, but the cross-linking of effector to target cells that occurred was similar. 11'32 In some reports, N-succinimidyl 3-(2-pyridyldithio) propionate (SPDP) was used as a cross-linking reagent for the production of bispecific antibodies, but products prepared by SPDP are mainly polymers of more than 330 kD that are cleared more rapidly by the reticuloendo-

% specific lysis

10 Control Patient

20

30

40

50

I

1

4

1

] m

none

[ - - ] Anti CD3X Anti-Glioma Ab

5

6

FIG. 8. Comparison of cell cytotoxicity mediated by lymphocytes of glioma patients with and without bispecific heteroF(ab')2 fragments. Effector cells were peripheral blood mononuclear cells (PBMC) obtained from glioma patients and cultured for 3 days in culture medium containing 100 U/ml of recombinant interleukin-2. U251MG cells were used as target cells. Lysis was measured at an effector:target ratio of 5:1 and the hetero-F(ab')2 fragments were added at a final concentration of 1 ug/ml. Control study: cells from a healthy donor; Patients 1 to 6: PBMC from six glioma patients. Ab = antibody. 480

Acknowledgment

We are indebted to Dr. David M. Segal of the National Cancer Institute for his invaluable help in preparation of the manuscript. References

1

2 3

thelial system than are F(ab')2 monomeric products. 1~ The reactivity of the bispecific heteroF(ab')2 antibody with T cells or U251MG cells seemed to be depressed when compared with the relevant intact IgG by flow cytometry. This is probably because the hetero-F(ab')2 fragment has only a single paratope, while intact IgG or F(ab')2 fragments have two paratopes. Also, the hetero-F(ab')2 antibody lacks the Fc component which may depress its activity compared with native IgG. However, this does not indicate any impairment of binding capacity to the relevant epitopes. 16 Therefore, pure F(ab')2 fragments prepared with DTNB may have clinical applicability. Also, highyield production of pure F(ab')2 m o n o m e r antibodies can be achieved with D T N B in contrast to the use of SPDP (20% to 30%). j5 In cancer-bearing hosts, induction of LAK cell activity from PBMC's has been shown to be impaired because of cancer cell-derived suppressive factors or an increase in the population of suppressive macrophages. 21-23 We found that L A K cell activity was impaired in two-thirds of our patients suffering from malignant glioma, but targeted cytotoxicity using bispecific antibody produced a significant improvement in all cases. These in vitro results suggest that, after intracranial gliomas are surgically removed and are confirmed to react with anti-glioma MAb's, this specific targeting therapy may be a more effective adjuvant immunotherapy than L A K cell therapy alone.

1. Bullard DE, Gillespie GY, Mahaley MS, et al: Immunobiology of human gliomas. Semin Oncol 13:94-109, 1986 2. Ellman GL: Tissue sulfhydryl groups. Arch Biochem Biophys 82:70-77, 1959 3. George RE, Loudon WG, Moser RP, et al: In vitro cytolysis of primitive neuroectodermal tumors of the posterior fossa (medulloblastoma) by lymphokine-activated killer cells. J Neurosurg 69:403-409, 1988 4. Grimm EA, Ramsey KM, Mazumder A, et al: Lymphokine-activated killer cell phenomenon. I1. Precursor phenotype is serologically distinct from peripheral T lymphocytes, memory cytotoxic thymus-derived lymphocytes, and natural killer cells. J Exp Med 157:884-897, 1982 5. Hirohashi S, Kuuni T, Hirano T, et al: 145Kd cell membrane Ag on SCLC. Lung Cancer 4:103-104, 1988 6. Kappler J, Kubo R, Haskins K, et al: The major histocompatibility complex-restricted antigen receptor on T cells in mouse and man: identification of constant and variable peptides. Cell 35:295-302, 1983 7. Kitahara T, Watanabe O, Yamaura A, et al: Establishment of interleukin 2 dependent cytotoxic T lymphocyte cell line specific for autologous brain tumor and its intracranial administration for therapy of the tumor. J Neurooncoi 4:329-336, 1987 J. Neurosurg. / Volume 72/March, 1990

Cytotoxicity induced by bispecific antibody against glioma 8. Laemmli UK: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680-685, 1970 9. Leeuwenberg JF, Spits H, Tax WJM, et al: Induction of nonspecific cytotoxicity by monoclonal anti-T3 antibodies. J Immunol 134:3370-3375, 1985 10. Liu MA, Kranz DM, Kurnick JT, et al: Heteroantibody duplexes target cells for lysis by cytotoxic T lymphocytes. Proe Natl Acad Sci USA 82:8648-8652, 1985 11. Liu MA, Russbaum SR, Eisen HN: Hormone conjugated with antibody to CD3 mediates cytotoxic T cell lysis of human melanoma cells. Science 239:395-397, 1988 12. Martin DE, Zalman LS, M/~ller-Eberhard H J: Induction of synthesis of the cytolytic C9 (ninth component of complement)-related protein in human peripheral mononuclear cells by monoclonal antibody OKT3 or interleukin 2: correlation with cytotoxicity and lymphocyte phenotype. Proc Natl Acad Sci USA 84:2946-2950, 1987 13. Meuer SC, Fitzgerard KA, Hussey RE, et al: Clonotypic structures involved in antigen-specific human T cell function. Relationship to the T3 molecular complex. J Exp Med 157:705-719, 1983 14. Meuer SC, Hussey RE, Cantrell DA, et al: Triggering of the T3-Ti antigen-receptor complex results in clonal T-cell proliferation through an interleukin 2-dependent autocrine pathway. Proc Natl Acad Sci USA 81: 1509-1513, 1984 15. Nitta T, Yagita H, Azuma T, et al: Bispecific F(ab')s monomer prepared with anti-tumor monoclonal antibodies is most potent in induction of cytolysis of human T cells. Eur J Immunol 19:1437-1441, 1989 16. Parham P: On the fragmentation of monoclonal IgGl, IgG2a, and IgG2b from BALB/C mice. J Immunol 131: 2895-2902, 1983 17. Perez P, Hoffman RW, Shaw S, et al: Specific targeting of cytotoxic T cells by anti-T3 linked to anti-target cell antibody. Nature 316:354-356, 1985 18. Perez P, Hoffman RW, Titus JA, et al: Specific targeting of human peripheral blood T cells by heteroaggregates containing anti-T3 crosslinked to anti-target cell antibodies. J Exp Med 163:166-178, 1986 19. Rosenberg SA: Immunotherapy of cancer using interleukin 2. Immunol Today 9:58-62, 1988 20. Rosenberg SA, Lotze MT, Muul LM, et al: Observations on the systemic administration of autologous lymphokine-activated killer cells and recombinant interleukin2 to patients with metastatic cancer. N Engl J Med 313: 1485-1492, 1985 21. Roszman TL, Brooks WH, Elliott LH: Immunobiology of primary intracranial tumors. VI. Suppressor cell function and lectin-binding lymphocyte subpopulations in

J. Neurosurg. / Volume 72/March, 1990

22. 23.

24. 25. 26. 27.

28.

29.

30.

31.

32.

patients with cerebral tumors. Cancer 50:1273-1279, 1982 Roszman TL, Brooks WH, Elliot LH: Inhibition of lymphocyte responsiveness by a glial tumor cell-derived suppressive factor. J Neurosurg 67:874-879, 1987 Schwyzer M, Fontana A: Partial purification and biochemical characterization of a T cell suppressor factor produced by human glioblastoma cells. J Immunol 134: 1003-1009, 1985 Shapiro WR: Treatment of neuroectodermal brain tumors. Ann Neurol 12:231-237, 1982 Staerz UD, Kanagawa O, Bevan M J: Hybrid antibodies can target sites for attack by T cells. Nature 314: 628-631, 1985 Ting CC, Hargrove ME, Yun YS: Augmentation by antiT3 antibody of the lymphokine-activated killer cell-mediated cytotoxicity. J Immunol 141"741-748, 1988 Van de Griend RJ, Bolhuis RLM, Stoter G, et al: Regulation of cytolytic activity in C D 3 and CD3 + killer cell clones by monoclonal antibodies (anti-CD16, anti-CD2, anti-CD3) depends on subclass specificity of target cell IgG-FcR. J Immunol 138:3137-3144, 1987 Van Seventer GA, Kuijpers KC, Van Lier RA, et al: Mechanism of inhibition and induction of cytolytic activity in cytotoxic T lymphocytes by CD3 monoclonal antibodies. J Immunol 139:2545-2550, 1987 Van Wauwe JP, Goosens JG, Beverley PCL: Human T lymphocyte activation by monoclonal antibodies: OKT3 but not UCHT 1 triggers mitogenesis via an interleukin2-dependent mechanism. J Immunol 133:129-132, 1984 Walker MD, Green SB, Byar DP, et al: Randomized comparisons of radiotherapy and nitrosoureas for the treatment of malignant glioma after surgery. N Engl J Med 303:1323-1329, 1980 Zalman LS, Martin DE, Miiller-Eberhard H J: The cytolytic protein of human lymphocytes related to the ninth component (C9) of human complement: isolation from anti-CD3-activated peripheral blood mononuclear cells. Proe Natl Acad Sci USA 84:2426-2429, 1987 Zovickian J, Johnson VG, Youle RJ: Potent and specific killing of human malignant brain tumor cells by an antitransferrin receptor antibody-ricin immunotoxin. J Neurosurg 66:850-861, 1987

Manuscript received February 27, 1989. Accepted in final form September 11, 1989. Address reprint requests to: Ko Okumura, M.D., Department of Immunology, Juntendo University School of Medicine, 2-1-1 Hongo, Bunkyo-ku, Tokyo 113, Japan.

481