Selective elimination of lymphocyte subpopulations by monoclonal antibody-enzyme conjugates

Journal oflmmunologicalMethods, 79 (1985) 307-318 Elsevier

307

JIM03501

Selective Elimination of Lymphocyte Subpopulations by Monoclonal Antibody-Enzyme Conjugates Renu B. Lal 1,,, Elinor M. Brown 1, Bruce E. Seligmann 2, Linette J. E d i s o n I a n d T h o m a s M. Chused I Laboratory of Microbial Immunity, and 2 Laboratory of Clinical Investigation, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD 20205, U.S.A.

(Received 30 October 1984, accepted 7 February 1985)

A novel method for the selective depletion of lymphocyte subpopulations has been developed. Conjugates of glucose oxidase (GOx) and phospholipase-C (PL-C) coupled to a monoclonal mouse anti-rat IgG (MAR) were shown to be selectively cytotoxic for targeted lymphocyte subsets in the presence of various rat monoclonal antibodies directed toward murine cell surface antigens. The ability of both conjugates to bind specifically to rat monoclonal antibodies was demonstrated by flow cytometry. The PL-C-MAR conjugate was more stable than the GOx-MAR conjugate. The PL-C conjugate, in conjunction with primary rat anti-mouse monoclonal antibodies, produced selective killing of T or B cells as detected by a loss of proliferative capacity in response to mitogens and by specific cell depletion demonstrated by flow cytometry. Normal mouse serum protected against the cytotoxicity of free enzymes but had no protective effect on enzyme conjugates. Because the substrates of these enzymes are abundant in vivo and serum did not interfere with their cytotoxicity, these enzyme-antibody conjugates may be valuable for selective lymphocyte depletion in vivo. Keywords: immunotoxins - antibody-enzyme conjugates - glucose oxidase- phospholipase

Introduction T h e use o f t h e specificity a n d u n i f o r m i t y of m o n o c l o n a l a n t i b o d i e s for t h e i m m u n o t h e r a p e u t i c e l i m i n a t i o n o f s e l e c t e d p o p u l a t i o n s o f cells b o t h in v i t r o a n d in * Correspondence address: Dr. Renu B. Lal, Laboratory of Microbial Immunity, Building 5, Room 226, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD 20205, U.S.A. Abbreoiations: Con A, concanavalin A; FC, flow cytometry; GOx, glucose oxidase; Hepes, N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid; LPS, lipopolysaccliaride; MAR, mouse anti-rat-kappa18.5; PL-C, pbospliolipase-C; PBS, pliospliate-buffered saline; PHA, pbytohemagglutinin; SMCC, succinimidyl-4-(N-maleimide methyl cyclohexane)-l-carboxylase; SPDP, succinimidyl-3-(pyridyl-dithiopropionate); SOD, superoxide dismutase. 0022-1759/85/$03.30 © 1985 Elsevier Science Publishers B.V. (Biomedical Division)

308 vivo could be widely applicable. Several reports of the development and use of immunotoxins have been published (Philpott et al., 1974; Krolick et al., 1982; Raso et al., 1982; Bumol et al., 1983; Columbatti et al., 1983; Mew et al., 1983; Perek et al., 1983; Vitetta et al., 1983; Hashimoto et al., 1984; Quinones et al., 1984; Vallera et al., 1984). Most have used antibody conjugated to the A chain of ricin, a toxic lectin obtained from castor beans (Ricinus communis) (Olsnes and Pihl, 1973). This immunotoxin has been used for in vivo tumor elimination (Krolick et al., 1982), in vitro depletion of immunocompetent cells before bone marrow transplantation (Quinones et al., 1984; Vallera et al., 1984) and immunoselection procedures (Colombatti et al., 1983). However, the use of ricin has 2 limitations. First, the A chain, a toxic subunit of ricin that inhibits protein synthesis must be dissociated from the B chain to avoid the non-specific binding of the toxin to the cell surface. Secondly, the antibody-ricin A chain conjugate must be internalized by the cell into an endocytic vesicle or phagolysosome and then escape from such a vesicle into the cytoplasm before reaching the susceptible ribosomes. Thus, the efficacy of such an immunotoxin depends on the endocytosis of the conjugate and its subsequent release into the cytoplasm without degradation. An alternate approach has been explored in the current study which uses monoclonal antibody conjugated to enzymes (glucose oxidase (GOx) and phospholipase-C (PL-C)) whose products damage the cell membrane, thus eliminating the requirement that the toxic conjugate be internalized. Enzymes whose substrate is present in vivo in extracellular milieu were selected for study. We have used a sequential localization system, in which monoclonal rat antibody against normal murine differentiation antigens were bound to the cell surface followed by incubation with a monoclonal anti-rat Ig antibody conjugated either to GOx or PL-C. The advantage of this indirect system is that the specificity of the conjugate can be directed towards any surface differentiation antigen of any cell type. Our results indicate that both the conjugates maintain their antigen binding capacity and that small amounts of the conjugates are effective in specifically killing target cells. Although antibody-PL-C conjugate appears less efficient than antibodyGOx conjugate as determined by cell survival, both types of conjugates could be successfully used to deplete targeted cell populations by in vitro immune selection.

Materials and Methods

Animals N F S / N mice were obtained from the Small Animal Section, Veterinary Research Resources Branch, NIH, Bethesda, MD.

Antibody reagents Cell lines secreting monoclonal anti-Thy-l.2 (clone Jij, rat IgM (Bruce et al., 1981)) was kindly provided by Dr. J. Sprent and anti-L3T4a (clone GK1.5, rat IgGza (Dialynas et al., 1983)) by Dr. F. Fitch. Anti-B cell surface glycoprotein B220 (clone

309 RA3-3A1/6.1, rat IgM (Coffman and Weissman, 1981); clone 14.8, rat-IgG2a (Kincade et al., 1981)); anti-IgM (clone 331.12, rat IgG2a (Kincade et al., 1981)); and M1/69.16.11.HL, rat IgG2b (Springer et al., 1978), which recognizes a heat stable antigen on mouse red blood cells (RBC), thymus, mature B lymphocytes, granulocytes, and monocytes were obtained from the American Type Culture Collection. Monoclonal anti-Lytl (clone 53-7.3, rat IgG2~) and anti-Lyt2 (clone 53-6.7, rat IgG2~) were harvested from culture supernatants of cell lines kindly provided by Dr. J. Ledbetter (Ledbetter and Herzenberg, 1979). The antibodies were isolated from culture supernatants by ammonium sulfate precipitation followed by ion exchange and gel filtration chromatography. Anti-Thy1.2, Lytl, Lyt2, L3T4a, and anti-IgM were conjugated with fluorescein isothiocyanate (Research Organics, Cleveland, OH). The molar fluorescein protein ratio was between 1 and 4. Fluorescein-conjugated goat anti-mouse IgG (Fc-specific; Cappel Laboratory, Cochranville, PA) was absorbed with murine splenocytes until no staining was detectable by flow cytometric analysis. Staining reagents were centrifuged at 130,000 × g for 20 rain in a Beckman Airfuge to remove aggregated IgG. Monoclonal mouse anti-rat kappa antibody (clone MAR 18.5), provided by Dr. L. Lanier (Lanier et al., 1982), was purified from the culture supernatant by absorption and elution from a Staphylococcus protein A-Sepharose affinity column (Pharmacia Fine Chemicals, Uppsala). Two peaks were eluted at pH 5.5 and pH 4.5, the later having the antibody activity. This was concentrated and dialyzed against phosphate-buffered saline (PBS).

Enzymes Xanthine oxidase, choline oxidase, and phospholipase-D were obtained from Calbiochem, while phospholipase-A 2, phospholipase-C, and glucose oxidase were purchased from Sigma. GOx and PL-C were dissolved in 20 mM sodium phosphate buffer, pH 7.0, and further purified by gel filtration on a Bio-gel A-0.5 m (1 × 100 cm) at a flow rate of 5 ml/h at 4°C. GOx eluted as a single peak, while PL-C had 2 peaks. The enzyme activity of PL-C was found in the second peak which was pooled and concentrated by vacuum dialysis. Both purified enzymes gave a single band on SDS-gel electrophoresis performed by using 5% polyacrylamide gel in 0.05 M Tris-glycine buffer at pH 8.6 (Laemmli, 1970). Conjugation of MAR to GOx and J'L-C Maleimide groups were introduced into GOx or PL-C by succinimidyl-4-(Nmaleimide methyl cyclohexane)-l-carboxylase (SMCC, Pierce Chemicals) (Yoshitake et al., 1979). To purified GOx (4.18 mg/ml) and PL-C (1.98/mg/ml), 10 /~1 of SMCC in dioxane (20 mg/ml) were added dropwise over a period of 5 min with continuous stirring. Incubation was carried out at room temperature for 1 h, and the protein was separated from excess SMCC by gel filtration on Sephadex G-25 in 20 mM sodium acetate, pH 4.5, containing 0.1 M NaCI. Succinimidyl-3-(pyridyl-di-thiopropionate) (SPDP, Pharmacia Fine Chemicals) (Carlsson et al., 1978) was used to introduce 2-pyridyl disulfide groups in MAR.

310 Thirty-five /xl of 2 mg/ml SPDP in ethanol was added to 1.7 ml of 3.75 mg/ml MAR in 0.1 M sodium phosphate buffer, pH 7.5. After incubating for 2.5 h at room temperature, the pyridyl disulfide-substituted MAR was isolated by chromatography on Sephadex G-25 (PD-10 column, Pharmacia Fine Chemicals) in 20 mM sodium acetate, pH 4.5, containing 0.1 M NaC1. The ratio of 2-pyridyl disulfide to IgG was 2.6. The 2-pyridyl disulfide was reduced by the addition of dithiothreitol (DqT) at a final concentration of 25 mM in sodium acetate, pH 4.5. After incubation for 30 min, the antibody was recovered from Sephadex G-25 in sodium phosphate buffer with 1 mM ethylenediaminetetraacetic acid (EDTA), pH 7.5. Thiolated antibody (850 ~1) was immediately added to 1 ml each of maleimide GOx or maleimide PL-C, adjusted to pH 7.8-8.2 with 1 M Tris, and kept at 4°C for 16 h. Remaining thiol groups were blocked by adding 10 mM sodium iodoacetate. The conjugates were then subjected to gel filtration on a column of Bio-Gel A-0.5 m (1 x 100 cm) using 20 mM sodium phosphate, pH 7.5. For GOx-MAR conjugate, the fraction eluting at approximately 300 kDa, was collected. For PL-C-MAR conjugate, the initial fraction of the peak eluting at approximately 250 kDa was collected and concentrated. The molecular weights of the conjugates were confirmed by sedimentation equilibrium analysis (data not shown), and homogeneity was established by SDS-polyacrylamide gel electrophoresis (data not shown). Bovine serum albumin (0.01%) and magnesium chloride (2/~M) were added to stabilize the enzyme activity and the conjugates were stored at 4°C.

Binding and cytotoxic studies Single cell suspensions were prepared by gently flushing thymuses or halved spleens in Medium-199 (Gibco) containing glutamine (2 raM) and Hepes (10 mM), pH 7.2. Cell debris was removed by settling at room temperature for 10 rain, and the cells were then washed twice and counted with a Coulter counter (Coulter Electronics, Hialeah, FL). To determine non-specific cytotoxicity of free enzymes, one million target cells in 50 ~1 were incubated with free enzyme for 30 min, suspended in 0.4 ml medium, and incubated for 2 h at 37°C. To determine the specific cytotoxicity of enzyme conjugates, one million target cells were incubated with a saturating concentration of rat anti-mouse monoclonal antibody for 30 rain at room temperature and washed twice. Controls were incubated with PBS in the same way. Enzyme conjugate directed against rat antibody was then added, incubated for 30 min at room temperature, washed twice, resuspended in 0.4 ml medium, and incubated at 37°C for various time points in 5% CO 2. Cell survival was determined by flow cytometry (FC). Serum, catalase, or SOD was added before free enzyme or enzyme conjugate was added. For indirect immunofluorescence studies, cells were incubated sequentially with primary antibody, MAR conjugate, and then with fluoresceinated antibody directed against the Fc portion of the MAR conjugate, and examined by FC. Flow cytometry FC was performed on a fluorescence activated cell sorter (FACS II), Becton-Dickinson, Mountain View, CA. The data was analyzed by a PDP 11/34 computer

311 (Digital Equipment, Maynard, MA) using programs developed by T.M. Chused. All data were collected using log amplifiers and displayed on a log10 scale of increasing green and red fluorescence intensity. To determine cell survival, dead cells were rejected from analysis by forward light scatter and propidium iodide (PI) staining by electronic gating. Ten/~1 of PI at 50/~g/ml in 0.1% sodium citrate was added to the cell sample tube just prior to FC analysis. By using a fixed pressure in the flow system, the rate of sample flow was held constant. Thus, the flow rate of viable cells was directly proportional to cell survival under the various experimental conditions. The flow rate of cells not exposed to enzyme conjugate was considered as 100%, and the experimental data was proportionately expressed. Each experiment was performed at least twice. Determinations within each experiment were done in triplicate. Mean values are shown in the figures.

Mitogen response after conjugate treatment Spleen cells (6 X 105 cells) were treated with PL-C-MAR in the presence or absence of primary antibody as described above, incubated at 37°C for 150 min in 5% CO2, and centrifuged (400 x g) for 10 min. Cells were washed once with RPMI 1640 (Whittaker, M.A. Bioproducts, MD) supplemented with 10% fetal calf serum, 10 mM Hepes, 2 mM glutamine, 5 x 10 -5 M 2-mercaptoethanol, penicillin and streptomycin, and cultured in triplicate at 2 × 105 cells/well in 0.2 ml medium in a 96-well flat bottom culture plate. Concanavalin A (Con A, Pharmacia) was added to cultures at 5 /~g/ml and phytohemagglutinin (PHA, Wellcome) was used at 2 /~g/ml. Preliminary experiments using a range of concentration of Con A and PHA determined that these were the respective optimal doses of the mitogens for DNA synthesis response. Cultures were incubated at 37°C for 48 h, pulsed with tritiated thymidine (specific activity 6.7 Ci/mM, New England Nuclear, Boston, MA), 1 /~Ci/well, and harvested 4 h later. Incorporation of radiolabeled thymidine was measured by a liquid scintillation counter.

Results

Screening of potentially toxic enzymes Six enzymes were initially screened for possible usefulness as immunotoxins (Materials and Methods). Among the peroxide-generating enzymes, xanthine oxidase had no cytotoxic effect (data not shown), and its substrate, xanthine, was only partly soluble under experimental conditions. Choline oxidase was only active when 10 mM choline was added to the enzyme; 2.5 x 10 -8 g/ml killed 39% of murine thymocytes. Exposure of thymocytes (Fig. 1A) or spleen cells (Fig. 1B) to GOx for 2 h caused a dose-dependent decrease in cell survival. The addition of 5-10 mM exogenous glucose did not enhance the killing (data not shown), suggesting that glucose was not a limiting factor. Of the hydrolytic enzymes tested, phospholipase-A 2 and phospholipase-D had no cytotoxic effect (data not shown); in the presence of PL-C a decrease in viable cells of both thymocytes (Fig. 1A) and spleen cells (Fig. 1B) was observed after 2 h

312

A

too

1(~

B

5o

0.25

2.5

25

250

ENZYME CONCENTRATION (10 Sg/rnl)

2500

g

2.5

25

250

2500

ENZYME CONCENTRATION (10eg/ml)

Fig. 1. Selective killing of mouse thymocytes (A) and spleen cells (B) in the presence of unconjugated G O x (O) or PL-C (F1).

incubation. Though incubation of GOx or PL-C (2.5 x 10-8 g / m l ) with thymocytes for 2 h resulted in 25% and 28% killing, respectively, the observed cytotoxicity was approximately 75% when GOx and PL-C were added together at the above concentration, indicating a synergistic effect of the 2 enzymes. As a result of this initial screening, GOx and PL-C were selected for further study. To investigate the suitability of these enzymes for in vivo use, the effects of serum on the cytotoxicity of unconjugated GOx and PL-C was examined. The addition of mouse serum to thymocytes before exposure to GOx a n d / o r PL-C protected cells against cytotoxicity. Serum provided a 15%-25% increase in cell survival after exposure to GOx (Fig. 2A), a 40-50% increase in the case of PL-C (Fig. 2B) or the combination of GOx and PL-C (Fig. 2C). Since catalase and SOD are known to scavenge superoxide radicals, their effect on GOx-mediated cytotoxicity was studied. As shown in Fig. 3A, catalase at a concentration as low as 12.5 U / m l provided A

B

100

C o

2= ¢..)

4.

~

o

Izs

SERUM (percent)

Fig. 2. Protective effect of mouse serum on enzyme toxicity of unconjugated enzymes on mouse thymocytes. A: G O x at 2500 X 1 0 - 8 g / m l (e); 250 x 1 0 - s g / m l (11); 25 × 1 0 - 8 g / m l (rq); and 2.5 x 1 0 - 8 g / m l (A). B: PL-C at same concentrations as A. C: GOx and PL-C combined. G O x at 2 5 0 0 x 1 0 -8 g / m I + P L - C at 2 5 0 0 x 1 0 -8 g / m l (11); G O x at 2 5 0 0 x 1 0 -8 g / m l ( 0 ) ; G O x 2 5 0 0 x 1 0 - s g / m I + P L - C 250 X 10-8 g / m l (O); GOx 2.5 x 10-8 g / m l + PL-C 25 × 1 0 - s g / m l (O); PL-C 25 x 10-8 g / m l (D); GOx 2 . 5 x 1 0 - 8 g / m I + P L - C 2 . 5 × 1 0 - 8 g / m l (zx); GOx 2 . 5 × 1 0 -8 g / m l (O); PL-C 2 . 5 x 1 0 -8 g / m l (A). E n z y m e was added immediately after serum.

313

complete protection from GOx-mediated cytotoxicity. SOD (Fig. 3B), which promotes the dismutation of superoxide to H 202, had a minimal effect, suggesting that H202, not superoxide, mediates the cytotoxic effect. As expected, catalase and SOD had no effect on PL-C-induced cytotoxicity (data not shown).

Cytotoxic effects of enzyme-antibody conjugates GOx and PL-C were covalently conjugated with mouse anti-rat IgG (MAR) by a thioether linkage (see Materials and Methods). Column purified conjugates were used to study the cytotoxic effect in conjunction with various specific monoclonal antibodies, such as Thy-l.2 and mouse IgM. Prior to the investigation of the cytotoxicity of MAR-enzyme conjugates, the specificity of the binding of conjugate to cell surface was examined. Mouse spleen cells were incubated with monoclonal rat anti-mouse Thy-l.2 (Fig. 4A) or anti-mouse IgM (Fig. 4B) followed by incubation with PL-C-MAR conjugate, which recognizes the rat Ig, and finally, with FITC-conjugated goat anti-mouse IgG which can bind the Fc fragment of the mouse IgG component of the PL-C-MAR conjugate but not the surface IgM or IgD of B cells. Fig. 4A and B indicate that the mouse anti-rat antibody (MAR) component of the conjugate was able to specifically bind rat immunoglobulin (anti-Thy-l.2 or anti-/~). The cytotoxic activity of the conjugates was determined for various periods of time and reached a plateau at 150 min (data not shown), which was used for all further experiments. Fig. 5 shows the dose-response curves for GOx and PL-C conjugates. Both conjugates could kill 85%-90% of the thymocytes at 250 × 10 -8 g/ml. However, GOx-MAR was more efficient in cell depletion than PL-C-MAR. The concentration of antibody-enzyme conjugate required for 50% killing of thymocytes was 0.75 × 10 -8 g/ml in the case of GOx-MAR and 62.5 x 10 -8 g/ml for PL-C-MAR. At a higher concentration of GOx-MAR (25 x 10 -8 g/ml), non-specific cytotoxicity was observed even in the absence of primary antibody (i.e., anti-Thy-l.2), whereas no such effect was observed with PL-C.

B

A 1

| ¢3

o 0

12.5 CATALASE (U/ml)

125

i 12.5

#

I

125

SUPEROXlDE DISMUTASE (U/ml)

Fig. 3. Effect of catalase (A) and SOD (B) on GOx toxicity toward thymocytcs. GOx 2500 × 1 0 - s g / m l (O); 250 x 10- s g / m l (11); 25 × 10- s g/nil (A); 2.5 × 10- s g/nil (17).

314 ,.OOFr ~

A

Ii I

o..oFfi °.,°rll / ,,.,

~ 0.00J 1.00 ~ uJ

B

0.80 0.60

!

0.40 0.20 0.00

i

~ 10

100

1,000

CHANNEL NUMBER

1

I

CONCENTRATION (10-Sg/ml)

Fig. 4. Binding of PL-C-MAR conjugate 75x10 -8 g/ml, to mouse spleen cells followed by FITC anti-mouse IgG, (A) in the presence (solid) or absence (dashed) of anti-Thy-l.2; (B) in the presence (solid) or absence (dashed) of anti-IgM and anti-B220 antibody. Fig. 5. Dose-response curve with conjugates GOx-MAR in the absence (e) or presence (O) of anti-Thy-l.2. PL-C-MAR in the absence (rn) or presence (A) of anti-Thy-l.2.

Effect of serum, catalase, and SOD on conjugate-mediated cytotoxicity It was shown earlier (Figs. 2 and 3) that serum and catalase protected against the cytotoxic effects of G O x on thymocytes, whereas only serum showed such an effect with PL-C (Fig. 2). Similar experiments were done to demonstrate the effects of serum, catalase, and SOD on G O x - M A R - and PL-C-MAR-induced cytotoxicity against thymocytes. In contrast to the protective effect of serum against free PL-C or G O x (Fig. 2A, B, C), serum had no protective effect against G O x - M A R or P L - C - M A R (data not shown). This could be due to the greatly increased local concentration of the enzyme at the cell surface provided by antibody-directed targeting. Similarly, catalase and SOD also had no protective effects (data not shown).

Stability of conjugates The stability of the conjugates following their conjugation was examined. In a period of 15 days, the G O x - M A R activity decreased by a factor of 63,000 while P L - C - M A R decreased only 3-fold, indicating the instability of G O x - M A R . Due to its greater stability, further experiments were performed using only the P L - C - M A R conjugate.

Selectivity of conjugate cytotoxicity Spleen cells were treated with either anti-Thy-l.2 or anti-IgM and then with

315 100

1.00

_J

0.80

"Z < 0.60

7"

0.40

0.20

o i

0.00

10

PL-C-MAR CONCENTRATION (10"Tg/ml)

100

1000

CHANNELNUMBER

Fig. 6. Dose-response curve of PL-C-MAR on mouse spleen cells; in the absence of primary antibody (e); in the presence of anti-Thy 1.2 (~,) or anti-lgM and anti-B220 (C)). Fig. 7. Analysis of PL-C-MAR conjugate-treated cell mixtures by FC. Spleen cells were incubated with (solid) or without (dashed) anti-IgM and anti-B220 followed by PL-C-MAR, incubated for 30 rain, washed twice, and incubated at 37°C for 150 rain. The cells were then stained with an anti-B cell monoclonal antibody (14.8), which was not inhibited by the previous monoclonal reagents, followed by fluoresceinated MAR. Out of 47% of total B cells stained, only 21% remained after PL-C-MAR conjugate (solid line) treatment compared to the control (dashed line). Most of the IgM bright B cells were depleted.

PL-C-MAR, incubated for 30 min, washed, resuspended in 0.4 ml medium, and incubated at 37°C for 150 min. A substantial fraction of the targeted cells were killed in each instance (Fig. 6). When similarly treated cells were incubated with FITC-conjugated goat anti-rat IgG and analyzed by FC, about 46% of B cells were

2s

zs

PL-C-MAR CONCENTRATION(10qg/ml)

Fig. 8. Mitogen response after PL-C-MAR conjugate treatment. A: Con A (5/zg), and B: ISHA (2/zg) in the absence (C)) or presence (e) of anti-Thy-l.2. Unstimulated control cultures in all conjugate experiments averaged 918 ± 389 cpm; the addition of Con A or P}TA caused a 12-15-fold increase in thymidine incorporation. The data for each point represents the mean values from pooled experiments in which each point is derived by calculating the percentage of control response (i.e., the response of PL-C-MAR-treated cultures divided by the response of control cultures multiplied by 100). In the absence of anti-Thy-l.2, the percent control response was 77% ± 8% and 89% ± 7% for Con A and PHA which decreased to 27% ± 9% and 24% ± 12%, respectively, in the presence of anti-Thy-].2.

316 stained which, in the presence of PI,-C-MAR, were reduced to 21% (Fig. 7), thus indicating a 60% depletion of the targeted population. Similar results were obtained with anti-Thy-l.2-treated cells. Mitogen responsiveness was used to assess the functional capacity of spleen cells exposed to PL-C-MAR (Fig. 8). PL-C-MAR alone (in the absence of primary antibody) reduced the response to Con A or PHA by only about 20%. Anti-Thy-l.2 followed by PL-C-MAR inhibited the Con A (Fig. 8A) or PHA (Fig. 8B) response approximately 80% at concentrations ranging from 2.5 × 10 - 7 g/ml to 250 × 10-7 g/ml. Thus, proliferation of T cells was dramatically inhibited in response to mitogen by the brief exposure to PL-C-MAR prior to culture. When cells were treated with anti-IgM antibody as target, followed by PL-C-MAR, the Con A (5 #g) response was unaffected (94.78 + 8.9% of control). These findings suggest that the specificity of PL-C-MAR conjugate was determined by the antigen-binding region of the antibody. Unconjugated MAR blocked the binding of PL-C-MAR to primary antibody by FC analyses. Furthermore, anti-Thy-l.2 followed by unconjugated MAR did not inhibit either Con A or PHA responses (data not shown). Discussion

The usefulness of 2 covalently linked enzyme-monoclonal antibody conjugates for the selective depletion of lymphocyte subset as selective immunotoxins was studied. GOx and PL-C were coupled to MAR, a monoclonal antibody directed against rat Ig by a thioether bond which is more stable than the conventional disulfide bond (Carlsson et al., 1978). Both conjugates retained their ability to specifically bind to rat monoclonal antibody directed against various murine differentiation antigens and were selectively cytotoxic in vitro as judged by flow cytometry and mitogen response. The specificity of the cytotoxicity was demonstrated both by the ability of conjugate to deplete specific lymphocyte subsets and by its ability to block the proliferative response to T cell mitogen in the presence of appropriate primary antibody. In the present study we have used GOx because its substrates, glucose and oxygen, are abundant in the body fluids, and it catalyzes the oxidation of exogenous glucose to gluconic acid with the formation of H202 (Keilin and Hartree, 1948). The GOx-lactoperoxidase and iodide system has been cytotoxic in several tumor cells and parasite models (Nathan, 1979). PL-C hydrolyzes cell surface phospholipids to form diglycerides and an organophosphorous compound. The increased diglyceride concentration in the plasma membrane causes the hemolytic and cytocidal effects of PL-C (Sabban et al., 1972). There has been an earlier report (Moolten et al., 1976) where conjugates of PL-C with antibodies against Forssman antigen of sheep red blood cells (SRBC) could lyse the SRBC, but those conjugates lacked potency. The GOx-mediated cytotoxicity was markedly inhibited by catalase, which scavenges H202 (Fig. 3A), but not by (SOD) (Fig. 3B), which dismutases the superoxide anion to H202. Thus, the observed cytotoxicity of the GOx conjugate system is probably mediated by H202, not superoxide. This is consistent with earlier findings (Weiss et al., 1981) which demonstrated that while superoxide itself is not

317 cytotoxic, the product of its dismutation, H202, plays a vital role in destroying cultured endothelial cells by human neutrophils. Furthermore, catalase has been known to protect against the heavy oxidative stress generated extracellularly (Voetman and Roos, 1980). Catalase and SOD had no protective effect against PL-C cytotoxicity. Similarly, catalase, but not SOD, inhibited the cytotoxic effects of GOx-MAR and had no effect on PL-C-MAR. To investigate the suitability of these enzyme conjugates in vivo, we then examined the effect of serum on GOx- and PL-C-induced cytotoxicity. Serum is known to scavenge H202 through endogenous catalase and also binds to and neutralizes the detergent effect of lytic enzymes like PL-C. We found that serum protected against the cytotoxicity of free GOx or PL-C. However, the cytotoxicity of covalently coupled conjugates was not diminished by serum, when sandwiched to target cells through a primary antibody, which recognized a specific surface antigen. This could be due to a greatly increased local concentration of enzyme at the cell surface and the inability of serum containing low levels of endogenous catalase to scavenge peroxide radicals or neutralize the lytic effect of lipases. The ability of the MAR antibody to bind to rat antibodies was well preserved in the conjugates. When thymocytes labeled with the appropriate rat antibody were exposed to GOx-MAR or PL-C-MAR, the percentage of surviving cells decreased exponentially with time in a dose-dependent manner. To increase the cytotoxicity, we tested the effect of the mixture of 2 conjugates and obtained a synergistic effect. At higher concentrations of GOx-MAR, non-specific killing in the absence of primary rat antibody was observed, whereas for PL-C-MAR it was minimal. Also, within a period of 15 days, GOx-MAR activity was considerably decreased. This is contrast to an earlier study of Yoshitake et al. (1979) who have reported that both the cross-link and GOx activity in Fab' conjugate were stable at pH 6-7 at 4°C for at least 6 months. The reason for this discrepancy is unclear. From our study it appears that PL-C-MAR is a more useful immunotoxin than GOx-MAR because of its stability and lack of non-specific cytotoxicity. The specificity of the conjugates was determined by FC and mitogen responsiveness after conjugate treatment. PL-C-MAR inhibited the Con A or PHA response of splenocytes by 20% in the absence of anti-Thy-l.2, but by 80% in the presence of anti-Thy-l.2. The incomplete inhibition of the T cell proliferative response might be due to a limiting concentration of PL-C bound to the cell surface. Since the combination of the 2 conjugates was synergistic, their combined use may result in complete elimination of the intended subset. The results presented in this paper have shown that antibody-enzyme conjugate enables the selective depletion of a particular lymphocyte subset. The cytotoxicity was quite selective with PL-C-MAR, but the proliferative response of the target cells was not completely inhibited. Specificity is conferred by the use of various primary antibodies directed against specific cell surface antigens. The system, particularly PL-C-MAR, may be feasible for use in vivo, since serum does not have an inhibitory effect on antibody-enzyme induced cytotoxicity, and the substrates for enzymes are found either in the extracellular milieu or in the plasma membrane of the target cells. Studies of combinations of toxic enzyme-antibody conjugates appear warranted.

318

Acknowledgements W e w i s h to t h a n k Dr. K a t h l e e n M c C o y for r e v i e w i n g the m a n u s c r i p t a n d Ms. M a y a H a d a r for e x c e l l e n t e d i t o r i a l assistance.

References Bruce, J., F.W. Symington, T.J. McKearn and J. Sprent, 1981, J. Immunol. 127, 2496. Bumol, T.F., Q.C. Wang, R.A. Reisfeld and N.O. Kaplan, 1983, Proc. Natl. Acad. Sci. U.S.A. 80, 529. Carlsson, J., H. Drevin and R. Axen, 1978, Biochem. J. 173, 723. Coffman, R.L. and I.L. Weissman, 1981, Nature (London) 289, 681. Colombatti, M., M. Nabholz, O. Gros and C. Bron, 1983, J. Immunol. 131, 3091. Dialynas, D.P., Z.S. Quan, K.A. Wall, A. Pierres, J. Quintans, M.R. Loken, M. Pierres and F.W. Fitch, 1983, J. Immunol. 131, 2445. Hashimoto, N., K. Takatsu, Y. Masuho, K. Kishida, T. Hara and T. Hamaoka, 1984, J. Immunol. 132, 129. Keilin, D. and E.F. Hartree, 1948, Biochem. J. 42, 221. Kincade, P.W., 1981, Adv. Immunol. 31, 177. Kincade, P.W., G. Lee, L. Sun and T. Watanabe, 1981, J. Immunol. Methods 42, 17. Krolick, K.A., J.W. Uhr, S. Slavin and E.S. Vitetta, 1982, J. Exp. Med. 155, 1797. Laemmli, U.K., 1970, Nature (London) 227, 680. Lanier, L.L., G.A. Gutman, D.E. Lewis, S.T. Griswold and N.L. Warner, 1982, Hybridoma 1, 125. Ledbetter, J.A. and L.A. Herzenberg, 1979, Immunol. Rev. 47, 63. Mew, D., C.-K. Wat, G.H.N. Towers and J.G. Levy, 1983, J. Immunol. 130, 1473. Moolten, F., S. Zajdel and S. Cooperband, 1976, Ann. N.Y. Acad. Sci. 277, 690. Nathan, C.F., 1979, in: Immunobiology and Immunotherapy of Cancer, eds. W.D. Terry and Y. Yamamura (Elsevier/North-Holland, New York) p. 59. Olsnes, S. and A. Pihl, 1973, Biochemistry 12, 3121. Perek, Y., E. Hurwitz, D. Burowski and J. Haimovich, 1983, J. lmmunol. 131, 1600. Philpott, G.W., R.J. Bower, K.L. Parker, W.T. Shearer and C.W. Parker, 1974, Cancer Res. 34, 2159. Philpott, G.W., A. Kulczycki, Jr., E.H. Grass and C.W. Parker, 1980, J. Immunol. 125, 1201. Quinones, R.R., R.J. Youle, J.H. Kersey, E.D. Zanjani, S.M. Azemove, C.C. Soderling B., T.W. LeBien, P.C.L. Beverley, D.M. Neville, Jr. and D.A. Vallera, 1984, J. Immunol. 132, 678. Raso, V., J. Ritz, M. Basala and S.F. Schlossman, 1982, Cancer Res. 42, 457. Sabban, E., Y. Laster and A. Loyter, 1972, Eur. J. Biochem. 28, 373. Springer, T., G. Galfr~, D.S. Secher and C. Milstein, 1978, Eur. J. Immunol. 8, 539. Vallera, D.A., R.C. Ash, E.D. Zanjani, J.H. Kersey, T.W. LeBien, P.C.L. Beverley, D.M. Neville and R.J. Youle, 1984, Science 222, 512. Vitetta, E.S., K.A. Krolick, M. Miyama-Inaba, W. Cushley and J.W. Uhr, 1983, Science 219, 644. Voetman, A.A. and D. Roos, 1980, Blood 56, 846. Weiss, S.J., J. Young, A.F. Lobuglio, A. Slivka and N.F. Nimeh, 1981, J. Clin. Invest. 68, 714. Yoshitake, S., Y. Yamada, E. Ishikawa and R. Masseyeff, 1979, Eur. J. Biochem. 101,395.