Preparation and characterization of conjugates of monoclonal

Akhouri A. Sinha , James L. Sackrison , Onofrea F. DeLeon , Michael J. Wilson , Donald F. Gleason. The Anatomical Record 1996 245 (4), 652-661 ...
1 downloads 0 Views 4MB Size
Bloconjugate Chem. 1993, 4, 455-466

455

Preparation and Characterization of Conjugates of Monoclonal Antibodies and Staphylococcal Enterotoxin A using a New Hydrophilic Cross-Linker Eva Akerblom,'J Mikael Dohlsten,tlg Charlotte Bryno,? Maria Mastej,? Ingrid Steringer,' Gunnar Hedlund,*lt Peter Lando,t and Terje Kallandtpa Kabi Pharmacia AB, S-751 82 Uppsala, Sweden. Received April 27, 1993"

Conjugates between monoclonal antibodies recognizing human cancer cells and the superantigen staphylococcal enterotoxin A (mAb-SEA) represent a potential novel approach to tumor therapy. Such mAb-SEA conjugates direct T-cells to lyse colon carcinoma cells in vitro. The synthesis of mAb-SEA conjugateswhich were prepared by introducing thiol groups on SEA and iodoacetyl or maleimide groups on mAb forming a stable thioether linkage between SEA and mAb is described. A hydrophilic spacer, composed of repeated ethylene oxide units, was constructed to increase the distance between SEA and mAb, preserving biological activity of both proteins. The degree of modification of mAb with SEA was determined with SDS-PAGE. Variables influencing the composition of the conjugates and their effect on the tumor-cell cytotoxicity were studied and optimal conditions for the synthesis were established. Functionally active mAb-SEA conjugates were prepared from a panel of different mAb and T-celldependent cytotoxicity against several human cancer types including colon, ovarial, breast, and renal cancer was obtained. This suggests that mAb-SEA conjugates may be of value in the treatment of human neoplastic disease.

1. INTRODUCTION

Monoclonal antibodies (mAb) have been developed against a variety of tumor types and utilized for experimental as well as clinical studies (I). The use of mAb in the therapy of solid tumors has however met with limited success (2) and a number of critical factors determining the therapeutic outcome have been identified. The amount of mAb accumulating in the tumor area is generally low (3)and the effector mechanisms responsible for tumorcell destruction, e.g. Fc-receptor cell mediated cytotoxicity or complement dependent lysis, seem to be insufficient. To overcome these limitations various mAb conjugates have been developed to increase the therapeutic efficacy. They include mAb conjugated to cytotoxic plant or bacterial toxins, radioactive isotopes, and cytotoxic agents as well as heteroconjugates simultaneously binding to the tumor cell and to activation structures on cytotoxic lymphocytes (4). Staphylococcal enterotoxin A (SEA) belongs to a family of bacterial and viral superantigens that at picomolar concentrations activate a large frequency of T-cells (5). Superantigens bind with high affinity to MHC class I1 molecules and subsequently interact with subgroups of T-cells based on their @-chainvariable region expressed within their T-cell receptor (TCR). We recently described that conjugates of mAb against colon carcinomas and SEA bind to tumor cells expressing the relevant antigen and subsequently direct cytotoxic T-cells to lyse the tumor cells irrespective of the MHC class I1 expression (6). This ~~~~

~

* Author for correspondence.

+ Department of Bioorganic Chemistry, Kabi Pharmacia AB,

Uppsala.

* Department of Immunology, Kabi Pharmacia AB, Lund.

8 Department of Tumor Immunology, The Wallenberg Laboratory, University of Lund. Abstract published in Advance ACS Abstracts, September

suggests that the TCR might be directly trigged by tumorbound mAb-SEA proteins. This paper describes the synthesis of such mAb-SEA conjugates and the variables which influence the composition of the conjugates and their capacity to induce T-cellmediated tumor cytotoxicity. A thioether linkage which is stable and resistant to metabolic degradation was used to couple SEA to mAb to minimize the potential degradation of mAb-SEA conjugates in vivo (7).The mAbSEA conjugates were prepared by introducing thiol groups on SEA and iodoacetyl groups or maleimide groups on mAb. The efficacy of antibody-ricin conjugates has been shown to be increased by the introduction of a flexible peptide spacer in the conjugates (8). We have constructed a hydrophilic spacer, composed of repeated ethylene oxide units, in order to examine if such a spacer might facilitate the interaction of mAb-SEA conjugates with the membranes of colon cancer cells and CTLs, and thus increase the activity of the conjugates. Such hydrophilic spacers are likely to be useful in any system where it is desirable to increase the distance between two active constituents of a conjugate or when an active compound is coupled to a solid phase (9). 2. EXPERIMENTAL PROCEDURES

Material. SEA prepared with a recombinant technique (rSEA) was expressed in Escherichia coli strain W 3110 and secreted to the periplasmic space.l Frozen cells were diluted with buffer, thawed, and centrifuged. The supernatant was applied to a Q-Sepharose column. The eluted crude rSEA fraction was after concentration fractionated on a Sephadex G-75 column and the final purification was performed on a Q-Sepharose column. The different mAbs used were obtained from the following sources and were reported to have the specificities listed in Chart I. Abrahamsh et al. To be published.

15, 1993. 1043-1a

o ~ ~ ~ ~ t ~ ~ o ~ - 0o 1993 ~ ~ American ~ ~ o ~ Chemical . o o ~Society o

456

Akerblom et al.

Bloconjugate Chem., Vol. 4, No. 6, 1993

Chart I mAb Ell G250 D612 B3 SM-3 Ma642 Ma710 IB16-6 MOv-18 (2215 C242 Anti-Thy-1.2

source Claes Juhlin, University Hospital, Uppsala, Sweden Centocor.Levden. The Netherlands J. Schlom, Bithesda, MD Ira Pastan, Bethesda, MD ICRF, London Kabi Pharmacia AB, Uppsala, Sweden Kabi Pharmacia AB, Uppsala, Sweden R. Barth, Columbus, OH Centocor,Leyden, The Netherlands Kabi Pharmacia AB, Uppsala, Sweden Kabi Pharmacia AB, Uppsala, Sweden Sigma, St. Louis, MO

174Iodoacetylamino)-3,6,9,12,15-pentaoxaheptadecanoic Acid (3a). Isopropyl 17-amino-3,6,9,12,15-pentaoxaheptadecanoate (1.1g, 3.2 mmol) (18)was hydrolyzed in 3 mL of 1M NaOH for 30 min. 1.5 mL of 6 M HC1 was added and the mixture was evaporated to dryness. The product was taken up in dichloromethane. A 545-mg portion of 17-amino-3,6,9,12,15-pentaoxaheptadecanoic acid hydrochloride (1) was obtained after evaporation of the solvent. Compound 1 (460 mg, 1.39 mmol) was dissolvedin 10mL of 1M borate buffer, pH 8.4. A solution of 432 mg (1.52 mmol) of N-succinimidyl 2-iodoacetate (2a) (Sigma) in 5 mL of dioxane was added and the pH was kept at 8.4 by addition of 5 M NaOH. The reaction was run in an Nz atmosphere and was completed in 15 min. The pH of the reaction solution was then adjusted to 3 and the solution was fractionated on a reversed-phase PEP-RPC HR 30/26 (Pharmacia Biotech AB) column using a gradient of 0-13 5% acetonitrile with 0.1 % trifluoroacetic acid (TFA) followed by isocratic separation at 13% acetonitrile, 0.1% TFA. Fractions were pooled and lyophilized, giving 351 mg (76%) of 3a: lH NMR (DzO) 4.23 (8, OCHzCOOH), 3.76 (8, ICHz), 3.71-3.76 (m, OCHzCHzO), 3.65 (t, NHCHzCHzO), 3.41 (t, NHCH2CH20). 174[3-(2-Pyridyldithio)propionyl]amin0]-3,6,9,12,15-pentaoxaheptadecanoic Acid (3b). Compound 1 (183 mg, 0.55mmol) was dissolved in 5 mL of 1M borate buffer, pH 8.4. N-Succinimidyl 3-(2-pyridyldithio)propionate (SPDP, Pharmacia Biotech AB) (190 mg, 0.61 mmol) dissolved in 1.5 mL of dioxane was added and pH was adjusted to 8 with 5 M NaOH. After 30 min of reaction, the pH was adjusted to 4.6 with 5 M HCl and the reaction solutionwas filtered and lyophilized. The reaction mixture was fractionated on a reverse-phase PEP RPC HR 30/26 column using a gradient of 0.17 % acetonitrile with 0.1 % TFA followed by isocratic separation at 17% acetonitrile with 0.1% TFA. Fractions were pooled and lyophilized giving 152 mg (88%) of 3b: 1H NMR (D2.0) 8.5 (6 H, pyridine), 8.0 (3 and 5 H, pyridine), 7.4 (4 H, pyridine), 4.15 (8,OCH2COOH), 3.68-3.72 (m,-OCHzCHzO-), 3.62 (t,NHCH~CHZO), 3.37 (t,-NHCHzCH20-), 3.12 (t, -SCH2CHz-), 2.7 (t, -SCHzCH2-). 174(3-Maleimidopropionyl)amino]-3,6,9,12,15-pentaoxaheptadecanoicAcid (3c). Compound 1 (225 mg, 0.68 mmol) was dissolved in 5 mL of 1 M borate buffer, pH 8.0, and the pH was adjusted with NaOH to pH 8.0. N-Succinimidyl 3-maleimidopropionate (190 mg, 0.71 mmol) was dissolved in 3 mL of warm acetonitrile and added dropwise to the above solution. pH decreased and was adjusted to 8.0 with NaOH. The reaction was finished in 15 min and pH was adjusted to 4.4 with 5 M HC1. The reaction solution was partly evaporated and then lyophilized. The crude product was fractionated on a PEPRPC HR 30/26 column using a gradient of 0-10% acetonitrile with 0.1% TFA followed by isocratic sepa-

subclass IgG 1 IeG 1 I ~ G 2a IgGl IgG 1 IgG 1 IgG 1 IgG 2a IgG 1 IgG 2a IgG 1 IgG 2a

reported reactivity parathyroid tumors renal cell carcinoma colon carcinoma pancarcinoma breast cancer breast cancer breast cancer mouse melanoma ovarian carcinoma colon carcinoma colon carcinoma mouse T-cells

ref 10

11 12 13 14 15 16 17

ration at 10% acetonitrile with 0.1 5% TFA. Fractions were pooled and lyophilized giving 112 mg (37%) 3c. 'H NMR (CDC13) 6.7 (8, HC=CH), 4.18 (9, CHzCOOH), 3.82 (t, NCHz), 3.4-3.8 (m, -CH2CHzO-), 2.58 (t,-CHzCONH-). Preparation of the N-Hydroxysuccinimide Ester of 3a, 3b, and 3c (10,6,and 12). 3a,3b,or 3c (25 pmol) was dissolved in 0.6 mL of dry dioxane and added to 2.9 mg (25 pmol) of hydroxysuccinimide in a reaction vial. A 107-pL sample of a dioxane solution with dicyclohexylcarbodiimide (37.5pmol) was added and the reaction vial was closed. After 4 h the reaction mixture was filtered and analyzed with NMR for content of 10, 6, and 12, respectively. Moisture was excluded to obtain a high yield. Preparation of [3-(2-Pyridyldithio)propionyl]amino-rSEA (4)and (3-Mercaptopropiony1)amino-rSEA (5). SPDP (0.31 mg, 0.98 pmol) dissolved in 54 pL of dioxane was added to a solution of rSEA (13.7 mg, 0.49 pmol) in 2.4 mL of 0.1 M borate buffer, pH 8.0. The reaction was completed in 30 min. A 100-pL portion of the reaction solution was added to a Sephadex G25 NAP5 column (Pharmacia Biotech AB) to remove excess SPDP and the protein fraction was analyzed for content of coupled SPDP groups on rSEA by the method described in ref 19. The pH of the reaction solution was adjusted to 4.8 with 2 M acetic acid and then 7.4 mg of dithiotreitol dissolved in 100 pL of 0.9% sodium chloride was added to reduce the disulfide bond. The reaction was run for 20 min and thereafter the reaction solution was separated on a Sephadex G25MPD 10column (Pharmacia Biotech AB). The protein was eluted with 3.3 mL of 0.1 M phosphate buffer the pH (7.5,6.0 or 5.4) of which was depending on the pH used in the following coupling reaction to mAb. The protein content was determined by amino acid analysis. The solution was stored under a nitrogen atmosphere and used within 3 h. Preparation of [ 17-[[3-(2-Pyridyldithio)propionyl]amino]-3,6,9,12,15-pentaoxaheptadecanoyl]aminor S E A (7) and [17-[(3-Mercaptopropionyl)amino]3,6,9,12,15-pentaoxaheptadecanoyl]amino-rSEA (8). The procedure was the same as for preparation of the compound 5,except that the reduction was performed at pH 2.5 to avoid precipitation. Preparation of (1odoacetyl)amino-mAb (9). A solution of 0.58 mg (2.05 pmol) of N-succinimidyl 2-iodoacetate 2a in 11pL of dioxane was added to a solution of 20 mg (0.13 pmol) of mAb C215 in 10 mL of 0.1 M borate buffer, pH 8.0. The reaction was run for 30 min and thereafter the reaction solution was fractionated on a Superdex 75 HR 16/50 column (Pharmacia Biotech AB) using 0.1 M phosphate buffer, pH 7.5, with 0.9% NaCl as eluent. The protein fraction was collected and concentrated in an Amicon cell through a YM 30 filter to 3.9 mL. The protein content was determined by amino acid analysis to be 4.0 mg/mL. The number of iodoacetyl groups per mAb was determined by iodine analysis to be 10.

Conjugates of mAb and SEA

Bioconjugate Chem., Vol. 4, No. 6, 1993 457

Table I. Synthesis of mAb(rSEA), Conjugates According to Method 1 (Scheme IV) starting materials reaction conditions* products molar mAb % composition of syn: ratio n for m for (spacer),, molar ratio of % yieldc monomer mAb(rSEA),d thesis of 2a or mAb rSEA concn, rSEA(SH),/ reaction mono- dimer + p=4 mAb no. 10/mAb (spacer),,o (SH), mg/mL mAb(spacer),, time, h mer polymer p = 0 p = 1 p = 2 p = 3 (p = 5) C215 1 27 19 1.5 2.3 3.4 19 52 41 33 31 21 12 2 16 10 1.7 2.3 4.3 17 51 49 40 27 17 8 C215 3 28 16.6 1.9 2.6 1.8 16 56 48 30 36 22 11 4 28 16.6 1.9 2.2 2.7 45 43 41 14 43 34 9 3.6 5 28 16.6 1.9 1.8 43 49 78 3 22 34 23 17(6) 3.4 6 15 7.7 1.4 2.9 22 88 36 42 32 17 9 5.3 7 10 5.1 1.6 1.9 19 76 45 30 36 26 10 8 15 9.0 1.6 2.6 4.1 20 74 49 21 28 30 14 1 9 15 8.4 1.6 2.6 3.8 15 94 37 33 33 26 7 1 3.1 C242 10 15 8.5 1.6 1.4 18 88 12 47 39 14 1 11 22 13.9 1.6 2.3 3.2 20 89 37 24 38 30 19 anti-Thy- 12 23 11.5 1.6 1.3 3.2 18 75 32 22 40 27 8 192 13 18 7.1 1.9 1.9 4.4 20 68 32 11 25 30 21 11 Ell 14 20 7.7 1.6 1.3 4 17 85 17 48 35 14 3 B3 15 15 9.6 1.6 2.2 3.3 17 104 21 41 34 19 6 IB16-6 16 17 7.6 1.6 1.6 5.7 21 58 28 28 31 25 16 Compound 9, syntheses 1 and 2; compound 11, syntheses 3-16. Coupling was performed at pH 7.5. The yield of monomer and dimer + polymer respectively was counted as percent protein obtained regardless of that the molecular weight of mAb(rSEA), was higher than that for the starting material mAb. Composition of monomer mAb(rSEA), in the reaction mixture including unconjugated mAb.

When using a 27 times molar excess of 2a, the number iodoacetyl groups per mAb was 19. Preparation of [ 1 7 4(Iodoacetyl)amino]-3,6,9,12,15pentaoxaheptadecanoyl]amino-mAb (1 1). mAb was reacted with reagent 10 and the reaction conditions and the purification procedure was the same as described in the proceeding example for 9. The molar excess of reagent 10 per mAb used for different synthesis is given in Table I. The number of iodoacetyl spacers substituted to mAb was determined by amino acid analysis using compound 1 as reference, and the values are given in Table I. Preparation of [ 174(3-Maleimidopropiony1)amino]-3,6,9,12,15-pentaoxaheptadecanoyl]amino-mAb (13). Reagent 12 was reacted with mAb under the following different conditions. (a) Reagent 12 was reacted with mAb at pH 8.0 for 30 min and then the reaction solution was fractionated on a Superdex 75 HR 16/50 column using 0.1 M phosphate buffer, pH 7.5, as eluent. The protein fractions were pooled and concentrated in an Amicon cell through a YM 30 filter. The purification procedure was completed in 4-6 h (syntheses 17-20, Table 11). (b) Reagent 12 was reacted with mAb at pH 8.0 or 7.5 for 15min and then fractionated on a Superdex 75 column using as eluent phosphate buffer, pH 7.0, 6.5, or 6.0, or acetate buffer, pH 5.4 (syntheses 21-38, Table 11). The molar excess of reagent 12 per mAb used for the different synthesis is given in Table 11. The number of maleimido spacers substituted on to mAb was determined by amino acid analysis and the values are given in Table 11. General Procedure for Preparing Conjugate 14,15, and 17. (3-Mercaptopropiony1)amino-rSEA( 5 ) was added to a solution of (iodoacety1)amino-mAb (9), [17-(iodoacetyl)amino-3,6,9,12,15-pentaoxaheptadecanoyll amino-mAb (ll),and [17-[(3-maleimidopropionyl)amino]3,6,9,12,15-pentaoxaheptadecanoyll amino-mAb (13), respectively, and the reaction was run under an Nz atmosphere at room temperature. Details of the individual syntheses are given in Tables I and 11. After the reported reaction time the reaction solution was analyzed with SDSPAGE and then fractionated on a Superdex 200 HR 16/60 column (Pharmacia Biotech AB) (Figure 2).

Fractions containing the monomer conjugate were pooled in different ways, giving conjugates of different composition. Syntheses 1-16 (Table I) were, before fractionation, treated for 1h with a 2 times molar excess of mercaptoethanol based on the number of iodoacetyl groups on mAb for reaction with unsubstituted iodine groups. Preparation of Conjugate 16 with a 42-Atom-Long Spacer Arm. [17-[(3-Mercaptopropionyl)aminol-3,6,9,12,15-pentaoxaheptadecanoyllamino-rSEA(2.6 mg, 0.095 pmol) (8) was added to 7.0 mg (0.045 pmol) of iodoacetyl spacer modified mAb (11). The reaction vessel was filled with Nz and closed. After 18 h, 0.74 pmol of mercaptoethanol was added to react with an excess of iodine groups for 1 h. The reaction solution was then fractionated on a Superdex 200 HR 16/60 column. The monomer peak was divided into two parts. The high molecular weight part gave conjugate 16a (Figure 6d). Removal of Unconjugated mAb from mAb(rSEA), Conjugates. Affinity Chromatography on a Sepharose 4B-anti-r-SEA Column followed by Size-Exclusion Chromatography on Superdex 200. Rabbit antibodies against SEA were coupled to CNBr-activated Sepharose 4B according to the instructions supplied by the manufacturer (Pharmacia Biotech AB). A Sepharose 4B-anti-r-SEA column (16 X 30 mm) was packed and equilibrated with 0.1 M phosphate, pH 7.5, containing 0.5 M NaCl at a flow rate of 60 mL/h. Purified mAb(rSEA), conjugate also containing unconjugated mAb (in 0.1 M phosphate buffer, pH 7.5,0.5 M NaC1) was applied to the column. The effluent volume was continuously monitored at 280 nm using a UV detector. Unspecific binding to the adsorbent was prevented by adding extrasalt (0.5 M NaC1) to the buffer. After careful washing of the column, the mAb(rSEA), conjugate was eluted backward with 0.2 M glycine, pH 2.8, at a flow rate of 30 mL/h. The fractions collected were immediately neutralized to pH 7.5 with 1 M Tris-HC1, pH 8.5. The fractions containing mAb(rSEA), were pooled. A Superdex 200 column (50 X 100 mm) (Pharmacia Biotech AB) was prepared for gel filtration of the mAb(&EA), conjugate. The column was equilibrated with 20

458

Bioconjupte Chem., Vol. 4, No. 6, 1993

Akerblom et el.

'Ac

Conjugates of mAb and SEA

mM phosphate buffer, pH 7.5,0.9% NaCl at a flow rate of 300 mL/h. Pooled material was applied to the column and eluted with the same buffer. Separation of the monomer and dimer form of mAb(rSEA), was obtained and the fractions with monomer mAb(rSEA), were pooled and further concentrated by membrane filtration. Determination of Dimer and Polymer by SizeExclusion Chromatography (SEC). The content of dimers and polymers in mAb(rSEA), conjugates was determined by SEC using a Superdex 200 HR 16/60 column. A 150-pg sample of protein was applied to the column. 20 mM phosphate, 0.9 % NaCl, pH 7.5 was used as an eluent and a UV detector at 280 nm was used for monitoring. The dimer, polymer, and monomer form of the mAb(rSEA), conjugate were determined gravimetric and the dimer and polymer form calculated as percent of the total content. Phast Analysis of the Composition of mAb(rSEA), Conjugates. SDS-PAGE. Phast System from Pharmacia Biotech AB was used for electrophoretic separation of the mAb(rSEA),, staining, and destaining. PhastGel gradient 4-15 was used for analysis under non-reducing conditions in the presence of 2.5% SDS. Each sample was boiled for 3 min. Separation and staining procedures were carried out according to the instructions given in file no. 130,200,and 210supplied by the manufacturer. Some modifications were made by running the electrophoresis for 80 V h and using an extra washing step in the CBB staining. One microliter of mAb(rSEA), conjugate containing 1-4 mg of protein/mL for CBB staining and 0.10.2 mg of protein/mL for silver staining was subjected to electrophoresis. Pharmacia calibration kit proteins for high molecular weight measurements, with molecular weights between 14400 and 94000 Da, were used as standards. Evaluation with PhastImage. PhastImage (Pharmacia Biotech AB) was used for the evaluation of the PhastGel media. A 613-nm filter was used when scanning CBB-stained gels and a 546-nm filter for silver stained gels in the gel analyzer. An Olivettipersonal computerwith PhastImage software controlled the gel analyzer and also evaluated the onedimensional separations. PhastImage defined contours around the bands and transferred them into a peak data curve which was possible to integrate. The distribution of mAb(rSEA), (p = 0, 1, 2...) was calculated. Each area was compared with the total area for all peaks/bands. Cell Lines. The colon carcinoma lines SW620, COLO 205, and LoVo and the ovarian carcinoma line HeLa were obtained from the American Type Culture Collection and cultured in R medium [RPMI 1640 medium (GIBCO) supplemented with 10% fetal calf serum, 1 mM nonessential amino acids, 50 pM 2-mercaptoethanol, and 1mM pyruvate (FlowLaboratories)]. The mammary carcinoma line MDA-MB 361was obtained from CanAg and cultured as described above. T-cell lines were established by stimulation of human peripheral blood lymphocytes with SEA (1ng/mL) as detailed earlier (20). These T-cell lines were all >99% CD3+ as determined by flow-cytometric analysis. Cytotoxicity Assay. Cytotoxicitywas measured at an effector/target cell ratio of 3 0 1 in a standard 4-h 51Crrelease assay as described (20). SEA, mAb, or conjugates were added at various concentrations directly into the assay. The percentage specific cytotoxicitywas computed according to the formula: % specific cytotoxicity = 100

Bioconjugate Chem., Vol. 4, No. 6, 1993 459

kd

212 170 116 76 53

s t a

b

c

d

e

f

s

t

Figure 1. SDS-PAGE of mAb(rSEA), conjugates and their protein components: (a) mAb C242, (b) C242(rSEA), purified from unconjugated mAb with affinity chromatography, (c) reaction mixture of C242(rSEA), synthesis, (d) C215(rSEA),, (e) C215(rSEA), purified from unconjugated mAb with affinity chromatography, (f) mAb C215, (st)molecular weight standard.

x [experimental release (cpm) - spontaneous release (cpm)/maximum release (cpm) - spontaneous release (cpm)l 3. CHEMICAL RESULTS

3.1. Composition of the mAb(rSEA), Conjugates. The composition of the mAb(rSEA), conjugates was determined with SDS-PAGE analysis and is presented in Tables I and 11. SDS-PAGE analysis under nonreducing conditions gave separate bands for mAb and mAb(rSEA), with 1-5 rSEA and a broad band for the corresponding dimer forms (Figure 1). Scanning the bands with PhastImage gel analyzer gave good reproducibility for the composition of the monomer conjugate when staining with CBB. However, concentrations > 1mg/mL were required. The limited availability of material made it therefore often necessary to use silver nitrate staining. This method has not been documented for quantitative use. However, we found reproducible results and good conformity with CBB staining when concentrations of 0.1-0.2 mg/mL were analyzed (data not shown). Scanning of the broad dimer band gave too low values. Quantification of dimer was therefore performed with SEC analysis. Separation of the bands on SDS-PAGE required heating of the sample solution. This caused some dissociation of light chain in mAb and mAb(rSEA), conjugates. The values in Tables I and I1 were corrected for loss of light chains. 3.2. Purification of mAb(rSEA), Conjugates, A good separation of the mAb(rSEA), conjugate and the monomer and dimer form of rSEA(SH), (5) was obtained by SEC on a Superdex 200 column (Figure 2, peaks 2,3, and 4, respectively). There was also a fairly good separation between the monomer conjugate peak 2 and peak 1 representingmaterial with high molecularweight including complexes of two molecules of antibody or more, together with an unknown number of rSEA molecules. Peak 2 with the monomer mAb(rSEA), conjugate was divided in fractions. However, each fraction showed overlapping species with 1-4 rSEA substituted on to the mAb and also unconjugated mAb. Fractions in the beginning of the peak were enriched in highly substituted

460

Bioconjugate Chem., Vol. 4, No. 6, 1993

Akerblom et al.

Scheme I1 a. rSEA(NHz), + (NO$CHzCHzSS 0

SPDP

0

rSEA(NH$CH2CH2SS 50

60

70

80

90

O

100

Elution time, minutes

Figure 2. Purification of mAb(rSEA), conjugate on Superdex 200 column: peak 1,aggregated mAb(rSEA),; peak 2, monomer mAb(rSEA),; peak 3, dimer of rSEA(SH),; peak 4, monomer of

rSEA(NH$CHzCHzSH),

,)

0 5

4

0 b. rSEA(NH2),

+

NO~CH,(OCH,CH~,NH$CH,CH,SS 0

rSEA(SH),.

C0

O

6

Scheme I HO$CHZ(OCH~CH~),NH~ 0 1

rSEA(NH$CH,(OCH2CH2),NHSCH,CH2sH),

Co

0

0 8

N O$CHzX

Scheme I11

0

0

2a

x=I

3a N 7

2b

X=CH2SS 0

3b

3c

mAb and in the end of the peak low-substituted and unconjugated mAb was found. Unconjugated mAb was removed effectively from mAb(rSEA) conjugates on a rabbit IgG anti-rSEA-Sepharose 6B column (Figure 1). However, due to the high efficiency of the column, only 50% of the conjugate was recovered. Three conjugates were purified from unconjugated mAb whereas the other conjugates were tested for cytotoxic activity without removal of free mAb since presence of the latter showed no interference with the in vitro test results (compare Figure 6a,b). Attempts to remove unconjugated mAb on Red A gel (Amicon Ltd.), Red Sepharose C1-6B (Pharmacia Biotech AB) or hydroxyapatite were unsuccessful (data not shown). 3.3. Coupling Procedure. Two methods were used for coupling rSEA via a thioether linkage to mAb (Schemes IV and V). In both cases thiol groups were introduced on the lysines of rSEA either with the SPDP reagent or the spacer reagent 6 (Scheme 11). A 2-fold molar excess of SPDP per rSEA and a concentration of rSEA around 6 mg/mL and pH 8.0 gave 1.5-1.7 SH groups per rSEA. A 1.25 molar excessof SPDP gave 0.9 SH groups per rSEA. A 4.5 molar excess of reagent 6 per rSEA gave 2.9 SH groups. In method 1 iodoacetyl groups were introduced on the lysines of mAb and in method 2 maleimide groups were introduced (Schemes IV and V, respectively). Both methods gave a stable thioether linkage between rSEA and mAb. The mAb(rSEA)p conjugates formed were a heterogenous mixture of conjugates with 1-4 rSEA and unconjugated mAb (Tables I and 11). 3.3.1. Method 1 . Coupling with Iodoacetyl Groups. Iodoacetyl groups were coupled to mAb at pH 8.0 using a concentration of mAb around 2 mg/mL. By using different excessive amounts of N-succinimidyl iodoacetate 2a and iodoacetyl reagent 10, respectively, a different number of spacers was coupled to mAb. The valid

0

0

-

Scheme IV a. 9

+S

b. 11 + 5 c. 11 + 8

mAb(NHFCH2SCH2CH2~M1-rSEA), 0 14 0 mAb(h31SCH2(OCHzCHz)5~~CH2SCHzCHz~NH-rSEA), 0 1s 0 0

mA b ~ h ~ ~ C H z ( O C H 2 C H z ~ 5 ~ ~ C H z S C H 2 C H z ~ N H ( C H z C H z O ) 5 C H z ~ ~ ~ r S E A 0 0 0 16

Scheme V

13+5

\ m A b ( N H ~ C H z ( O C H 2 C H z ) , M 1 ~ C H z C H z ~=CHCOOH), ~CH 0

0

0

18

correlation had to be determined for each mAb. Figure 3 shows the correlation found for mAb C215. The iodoacetylated mAb products 9 and 11could not be stored frozen, as they were found to aggregate.

Conjugates of mAb and SEA 20.1.

0

*

5

-

.

Bioconjugate Chem., Vol. 4, No. 6, 1993 .

'

10 15 20 Spacer reagent/mAb

25

30

Figure 3. Linear correlation between excess of spacer reagent 2a and 10, respectively, per C215 and degree of substitution of C215(spacer), (9 and 11, respectively). Correlation coefficients squared for both lines were 0.96.

pH 7.5 was used for coupling thiol-substituted rSEAs 5 and 8 to iodoacetylated mAbs 9 and 11. This pH was

chosen to avoid reoxidation of the thiol groups on rSEA to a disulfide dimer of rSEA which occurred if coupling was performed at a higher pH (data not shown). A total of 1.6 thiol groups on rSEA were chosen to avoid excessive cross-linking to high molecular weight conjugates. To optimize the coupling reaction conditions a different number of iodoacetyl spacers on mAb was used and the molar excess of rSEA(SH), 5 was varied (Table I). The reaction rate was studied for synthesis 6 with eight spacers on mAb C215 and a 3.4 molar excess of rSEA(SH), 5. The composition of the formed monomer conjugate was analyzed after 1,4.5, and 21 h with SDS-PAGE (Figure 4a). There was still 42 7% unconjugated mAb in the monomer 100

461

after 21 h. Increasing the reaction time further up to 40 h gave as a rule a lower yield of monomer conjugate as more high molecular material was formed. A more complete reaction was obtained either by increasing the molar excess of rSEA(SH), as in synthesis 8 or increasing the number of spacers as in synthesis 4. Optimal conjugates with mainly one and two rSEA substituted on to the mAb were obtained by using 8-10 iodoacetyl spacers on the mAb and a 4 molar excess of rSEA(SH), 5. The excess of iodoacetyl groups on the mAb(rSEA), conjugate was reacted with mercaptoethanol. Certain mAbs, e.g. C215, were sensitive to treatment with a large molecular excess of mercaptoethanol. SDS-PAGE analysis showed loss of the light chain, indicating that disulfide bands in the mAb had been reduced. 3.3.2. Method 2. Coupling with Maleimide Groups. Maleimide groups were coupled to mAb by reacting different molar excesses of spacer reagent 12 with mAb at pH 7.5 and 8.0. When the coupling reaction was run for 0.5 h at pH 8.0 and purification took 4-6 h at pH 7.5, the following reactions with rSEA(SH), gave incomplete coupling, indicating hydrolysis of the maleimide ring (syntheses 17-20, Table 11). To avoid this side reaction, coupling of 12 to mAb is recommended to be performed at pH 7.5 for 15 min and with a purification time of less than 2 h. To optimize the coupling reaction between 13 and 5 (SchemeV)the number of maleimide spacers on mAb was varied and different amounts of molar excess of rSEA(SH), 5 were used as well as changes of the coupling pH (Table 11). The reaction rate was studied for syntheses 17-28 by analyzing with SDS-PAGE the composition of the formed

I

]b

0

80

I h 4.5 h

I

60 40

20 0 mAb

mAb(rSEA) mAb(rSEA)2 mAb(rSEA)J

0

ed

mAb

mAb(rSEA) mAb(rSEA)2 mAb(rSEA)J mAb(rSEA)4

]d 0,75 h 4.5 h 21h

4.5 h 20h

I I

mAb

mAb(rSEA) mAb(rSEA)2 mAb(rSEA)S

mAb

mAb(rSEA) mAb(rSEA)2 mAb(rSEA)S mAb(rSEA)4

Figure 4. The composition of formed monomer mAb(rSEA1, conjugates measured with SDS-PAGE after different reaction times: (a) synthesis 6 (Table I) reaction between C215 with iodoacetyl spacer (11) and rSEA(SH),, (b) synthesis 24 (Table 11)reaction between C215 with maleimide spacer (13) and rSEA(SH) at pH 6.0, (c) synthesis 26 (Table 11)reaction between C215 with maleimide spacer (13) and rSEA(SH), at pH 5.4, (d) synthesis 29 (Table 11) reaction between (2242 with maleimide spacer (13) and rSEA(SH), at pH 5.4.

482

Bioconjugate Chem., Vol. 4, No. 6, 1993

Akerblom et ai.

Table 111. Stability of mAb(rSEA), Conjugates conjugate from storage time, dimer or synthesis temp, O C months polymeP % 25 3 -20 11 16 4 (15c) -20 10 traces 5 (150 -20 10 6 (15b) -20 0 0.5 1

3 7 (Ma)

-20

0

1

3 8 (15e)

-70

1

12 9 115h)

3 10 1

-70

17

-70

25

-70

3.5 0.3 4

6 a

18 15 10

23 23 traces 1

12 10 9 9

6 5 5

Determined by size-exclusion chromatography.

monomer conjugate after 1,4.5, and 20-21 h. The reaction was rapid at pH 7.0 and 7.5 and completed in 1h (syntheses 17-20) (Data not shown). The reaction was slower at 5.4, 6.0, and 6.5 and needed more than 4 h. Data for syntheses 24, 26, and 29 are given in Figure 4, parts b, c, and d, respectively. Reaction at pH 6.5 and 7.0 is not recommended since rSEA(SH), 5 had limited solubility at these pH and precipitations were observed. The coupling reaction gave significant dimer and polymer products. An effort to minimize these products by using only 0.86 thiol groups on rSEA(SH), was not successful (syntheses 27-31). The amount of dimer and polymer products formed was dependent on the individual mAb used as earlier found for method 1. The optimal coupling conditions for obtaining conjugates with mainly one or two rSEA with method 2 were when coupling occurred at a pH interval of 5.4-6.0 and 4-8 spacers on the mAb were used together with 1.6 SH groups on rSEA and 3 times excess of rSEA(SH), (Table 11)* 3.4. The Stability of mAb(rSEA), Conjugates. The stability of mAb(rSEA), conjugates prepared with both method 1and 2 and stored in 20 mM PBS, pH 7.5,0.9% NaCl at -20 or -70 "C is presented in Table 111. The freezing-thawing process gave traces of dimer + polymer (all data not presented). Storage at -20 "C gave between 15 and 25% dimer polymer on storage up to 11months. Storage at -70 "C was preferable as the dimer + polymer formed was only 1-12%. There was no difference in stability between conjugates prepared with the two different methods.

+

4. RESULTS OF BIOLOGICAL ACTIVITY

4.1. Cytotoxicity. The C215(rSEA),conjugatesinduced strong T-cell-dependent tumor cell lysis at 10 ng/mL and significant lysis at 1ng/mL, whereas unconjugated mAb and SEA itself had a marginal effect even at concentrations of 1000 ng/mL. The coefficient of analytical variation, calculated from the measurements carried out in triplicate, was routinely less than 15% A difference in activity is significant when the concentration of two conjugates, giving similar cytotoxicity, differs 4-fold or more. Variation in the maximal level of cytotoxicity recorded in the different experiments most likely depends on variation in activity from one T-cell line to another and to changes in the T-cell line during the in vitro cultivation time.

.

The cytotoxicity of rSEA seen in the different experiments may be due to low expression of the SE receptor distinct from the MHC class I1 receptor (21). A high concentration of the relevant but not of an irrelevant antibody blocks SADCC induced by the mAb(rSEA), conjugates, suggesting an antigen-specific effect (6). Cell lines expressing the relevant target antigen were sensitive to SADCC while cell lines lackingthe antigen were resistant (unpublished data). Studies of binding were performed on tumor cells expressing the mAb-defined antigen and demonstrated that the C215(rSEA), conjugates displayed target cell binding in a similar range as the native mAb C215 (Kd = 2 X M) (data not shown). 4.1.1. Cytotoxicity of Conjugates with a Different Number of Spacers. The cytotoxicity of C215(rSEA), conjugates (15) with 5,7, and 17 spacers (15a-c) prepared with method 1was compared (Figure 5a,b). All conjugates (15a-c) mediated strong lysis and only marginal differences could be recorded. Thus, mAb C215 allowed substitution with 17 spacers without loss of activity, whereas other mAbs (C242 and anti-Thy 1.2) lost activity at high rates of substitution (data not shown). It is therefore recommended to substitute mAbs with not more than 8-10 spacers when preparing new mAb(rSEA), conjugates with method 1. 4.1.2. Cytotoxicity of Conjugates with a Different Number of r S E A Substituents. We were not able to identify effective separation conditions to isolate mAb@SEA), conjugate with 1,2, or 3 rSEA per mAb. We have therefore compared the activity of C215(rSEA), conjugates with different composition (15d-f) (Figure 5c,d). Conjugate 15d was composed mainly of mAb with one rSEA whereas 15e was a mixture of mAb with 1,2, and 3 &EA. Conjugate 15f was a highly substituted conjugate. The different conjugates mediated similar cytotoxic activity. 4.1.3. Cytotoxicity of Conjugates Prepared with Different Methods. Conjugates 15h and 17a with similar composition but prepared with method 1 and 2, respectively, showed similar cytotoxicity (Figure 6a). Furthermore, conjugates 15h and 17a, with unconjugated mAb removed, showed essentially the same cytotoxic activity as conjugate 15g, where 17% mAb was present. 4.1.4. Cytotoxicity of Conjugates with Different Lengths of Spacers. Conjugates were prepared with 6-, 24-, and 42-atom-long spacer arms (14, 15, and 16, respectively) (Scheme IV), and their cytotoxicity was compared (Figure6b). There was no difference in activity between conjugates prepared with 6- or 24-atom spacer (14a and 15f,h, respectively). The difference in activity of conjugate with a 42-atom spacer (16a) was within the variability of the test method. Conjugates with the short 6-atom spacer gave about the same activity as those with 10- and 19-atom spacers (data not shown). 4.1.5. Cytotoxicity Using a Panel of mAb-SEA Conjugates with Different Tumor Selectivity. To analyze whether active mAb-SEA conjugates could be developed with several different mAb, we conjugated a panel of mAb reactive with different tumor types, i.e. colon cancer, ovarial cancer, breast cancer, and lung cancer. The mAbs C215, C242, and D612 reactive with colon carcinomas induced T-cells to mediate strong lysis of SW620, LoVo, and Colo205 cells, whereas unconjugated mAb or SEA alone had marginal effects (Figure 7a,b). The HeLa cell line was strongly lysed by T-cells in the presence of MOv18(rSEA), and C215(rSEA), conjugates, while no lysis was seen by unconjugated mAb and SEA. Similarly MA642@SEA),, MA710(rSEA),, and C215(rSEA), conjugates mediated strong lysis of the breast cancer cell line MDA-

Conjugates of mAb and SEA 80

Bioconjugate Chem., Vol. 4, No. 6, 1993 463

I SW620

I -- !, 711

A

.. .. ...._,

,01

50

a

. .....- . ..."..

,1

15a 15b 15c rSEA

I

. ....- . ..." . .d

1 10 100 1000 C215(rSEA)p, rSEA (nglml)

15a 5 spacer

10000

15b 7 spacer

1% 17 spacer

C215(rSEA)p p=0,1,2,3

80

70

c

0

.O

40

$

30

c

15 d 15 e 15 f SEA

,01

,1

1 10 100 1000 C215(rSEA)p, rSEA (nglml)

10000

15 d

15 e C215(rSEA)p p=0,1,2,3,4,5

15 f

Figure 5. The capacity of C215(rSEA), conjugates and rSEA to target SEA-responsiveCTLs against the C215+MHCclass 11- colon carcinoma cell lines SW 620 (a) C215(rSEA), conjugates prepared with different number of iodoacetyl spacers on mAb 11 (n = 5, 7, 17 Scheme 3) and (b) their chemical composition; (c) C215(rSEA), conjugates with different degree of substitution of rSEA and (d) their chemical composition. All conjugates were prepared with iodoacetyl spacer 11 method 1.

MD361 (Figure 7d), whereas the tumor cells remained resistant to unconjugated mAb and SEA alone. 5. DISCUSSION

mAb(rSEA), conjugates act by binding of the mAb moiety to the membrane of the tumor cell. The rSEA displayed on the tumor cell surface is recognized by a high frequency of T-cells and the tumor cell is subsequently lysed by the T-cells (6) (Figure 8). Thus, the mAb(rSEA), conjugates work with a mechanism distinct from that of immunotoxins (22) and do not require internalization. The coupling strategy is therefore different from that of the immunotoxins, which usually require a cleavable bond between mAb and toxin to be cytotoxic. The mAb(rSEA), conjugates did not need to be processed to activate T-cells and should have a stable linkage between the mAb and rSEA to minimize the side effects of free rSEA. A stable thioether linkage was therefore chosen instead of the often used disulfide bond. The new hydrophilicspacer reagents 6, 10, and 12 are composed of five repeated ethylene oxide units. Polymers with ethylene oxide units, polyethylene glycols, have several advantages compared to other spacers. They do not interfere with the conformation of peptides or proteins to which they are attached and they have amphiphilic characteristics enhancing solubility in both water and polar organic solvents (23). In aqueous solution they appear to exist as a freely permeable coil. One thiol group per rSEA is ideal to prevent cross-linking of mAb to high molecular weight material. However chemical coupling always results in a mixture of rSEA(SH), products with different number of thiol groups. An average substitution of 1.5-1.7 thiol groups per rSEA was chosen. Most rSEA molecules are expected to have one or two thiol groups and a lesser fraction no or more than two thiol groups.

The amount of high molecular weight material formed was rather high (20-50 % ) and was found to be dependent on the individual mAb used rather than on the number of thiol groups on rSEA. rSEA with 0.89 thiol groups per mAb gave a similar amount of high molecular weight material as rSEA with 1.6 groups (syntheses 27 and 28 compared with syntheses 17-26, Table 11). The number of spacers on the mAb was varied from 2 to 19. The reaction between thiol groups and maleimide groups was more rapid than the corresponding reaction with iodoacetyl groups. A lower number of maleimide spacers than iodoacetyl spacers was therefore needed to obtain satisfactory coupling results. Four maleimide spacers on the mAb and a 3 times excess of rSEA(SH), gave a similar composition of the mAb(rSEA), conjugate as 8-9 iodoacetyl spacers on the mAb and a 4 times excess of rSEA(SH), (synthesis 22, Table 11, and syntheses 8 and 9, Table I, respectively). It should also be noted that the concentration of mAb(spacer). was only 1.1mg/mL in synthesis 22 compared with 2.6 mg/mL in the syntheses 8and 9. In general it is advantageous to use as few spacers as possible to avoid possible interference with antibody binding. The maleimide group is sensitive to hydrolysis (24). Coupling of the spacer to the mAb during 0.5 h at pH 8.0 followed by purification for 6 h at pH 7.5 gave 40% unmodified mAb in the product (syntheses 17-20) independent of the number of spacers and the amount of rSEA(SH), used. This indicates that the maleimide spacer was partly hydrolyzed before coupling to rSEA(SH),. Coupling of the latter to the maleimide spacer at pH 7.0 and 7.5 was very rapid but carries the risk of hydrolysis at pH 7.5. A t pH 6.5-7.0 the rSEA(SH), had limited solubility, which was most obvious in synthesis 23 with 47 % unconjugated mAb in the mAb(rSEA), conjugate. Coupling at pH 5.4-6.0 worked well for a series of different

464

Akerblom et al.

Bloconlugate Chem., Vol. 4, No. 6, 1993

mAb (Table 11). Coupling at a pH value near the p l value of the mAb should in general be avoided since there is risk for precipitation. Method 2 using maleimide spacer was more convenient than method 1 using iodoacetyl spacer but gave a lower yield. The stability at -70 "C of mAb(rSEA), conjugates coupled with the two methods was about the same (Table 111). Most experiments were performed with the hydrophilic spacers 3a and 3c but for comparison some couplings were made with iodoacetylated mAb and rSEA(SH),, (Scheme IV, part a). The latter reaction was slower than the corresponding reaction with iodoacetylhydrophilic spacer (Scheme IV, part 6). Synthesis 2 gave a much lower substitution than synthesis 8 and the same was found for synthesis 1if compared with syntheses 3 and 5. The yield was also lower for synthesis 2 compared with syntheses 6-9 (Table I). These results show that the 21-atom spacer arm on mAb facilitates coupling to thiol groups on rSEA. The high amount of unconjugated mAb in all mAb@SEA), conjugates after completed reaction (Table I and 11) was unexpected. The reaction stopped despite the iodine groups left on the mAb and the excess of &EA(SH),. Prolongation of the reaction time from 20 to 45 h gave only slightly higher substitution. All attempts to get complete mAb conjugation resulted in conjugates with higher substitution. We were interested in studying the variables involved in mAb(rSEA), construction and their effect on the cytotoxicity. It has been reported that for ricin immunotoxins the biricin conjugates were 8 times more potent than the monoricin conjugates (25)and that a 29-residue flexible peptide spacer increased the potency 10times (8). SW 620

However, the structure-activity studies of the mAb conjugates were complicated by their heterogeneity. All mAb(rSEA), conjugates prepared were a mixture of conjugates with different number of rSEA substituted. Moreover, the coupling site of rSEA to mAb could vary for conjugates with the same degree of substitution. Another complication was that the SADCC test used for analyzing cytotoxicity does not allow measurements of minor differences in activity. The differences must be 4-fold or greater to be detected. It is reported that the number of spacers which can be coupled to an mAb without interfering with the affinity of the antibody is dependent on the individual mAb (26). mAb C215 allowed a very high number of spacers to be coupled with essentially unchanged antigen binding (6). No difference in cytotoxicity was seen for conjugates with 5, 7, or 17 spacers on C215 (Figure 5a). The maximal degree of substitution for the other mAbs was not examined, but the number of spacers was kept between 8 and 10when using coupling method 1and between 4 and 8 when using coupling method 2 to reduce interference with essential binding residues in the mAb CDR regions. Further we found that increasing the number of SEA molecules in the mAb-SEA conjugates did not improve the efficacy compared to those with one SEA per mAb. This observation suggests no major beneficial effects of multivalency (Figure 5c). No difference in cytotoxicity was seen between conjugates prepared with methods 1 and 2 (Figure 6a) and no improvement in cytotoxic activity was seen when the length of the spacer arm was increased as reported for ricin mAb conjugates (Figure 6c). A significant advantage of the hydrophilic spacer described in this report was that it increased the rate of

*

-

60 h h

-

50

.-

40

o In

b

a

17a rSEA

: 30

0

0

811

Fa2 El3

C

2 .-

P 0

20

cn

10 0

0 ,01

40

,1

10 100 1000 C215(rSEA)p, rSEA (nglml) 1

10000

15 9

15 h C215(rSEA)p p=0,1,2,3

17 a

60

SW 620

C

d

n

A

5

-

30 .-2. 0 c

.-0 X

E

20

0

0 .'c

8

2-

10

_f_

.. . ...... . . .7. . . ~

0 PO1

.-

14a 15h 151 I 16a rSEA

c

,1

1 10 100 1000 C215(rSEA)p, rSEA (nglml)

10000

0

2

30

C

.*s

20

v)

g

10

0

0 14a

151 C215(rSEA)pp=0,1,2,3,4,5

16a

Figure 6. The capacity of C215 (rSEA), conjugates and rSEA to target SEA-responsiveCTCs against the C215+MHC class 11- colon carcinoma cell lines SW 620: (a) The cytotoxicity of C215(rSEA), conjugates 15h and 17a prepared with method 1 and 2, respectively, and purified from unconjugated mAb and conjugate 15g prepared with method 1 and containing 17 ?6 unconjugated mAb and (b) their chemical composition; (c) the cytotoxicity of C215(rSEA), conjugates with different lengths of the spacer arm (14a with a 6-atom spacer, 15h and 15f with a 24-atom spacer, 16a with a 42-atom spacer) and (d) their chemical composition. The chemical composition of 15f is presented in Figure 5d.

Conjugates of mAb and SEA 70

Bioconlugafe Chem., Vol. 4, No. 6, 1993 465 50

4 Col0205

a

I LoVo

A

5

40

.-0 .-

)r

C

;5 20

3

0

0

-j

30 15h c 0 15i c iSEA 0” 20 nAb C215 g

;

2 10

1%

15h rSEA

.-0

pi

10

0

901

9 1

1

10

100

1000

10000

,01

C21YC242(rSEA)p, SEA,C215 (ng/ml)

40 Y

Y

s

4

$

.- 30 x

.-0

.-0 .-

0

0 X

)r

c

C

.

E 20

-g 20

0

0

0 .c .-

0 .c

v)

v)

E lo

-

30

c

X

n eo1

10000

C h

10

1

15h 17d 178 rSEA

I

0 8

1

1

10

100

1000

10000

901

Mov 18/C215(rSEA)p, rSEA (nglml)

1

60,

el

c

-

1 10 100 1000 D612/C215(rSEA)p, rSEA (nglml)

40 ,

HeLa A

?.!

,1

40

,1 1 10 100 1000 C21SmAA6421MA71O(rSEA)p, SEA (ng/ml)

10000

1

m

40

0

30

c

.2

20

v)

f 10 0

0

15h

15i

17b

C215/C242/D612(rSEA)pp=0,1,2,3,4

17 c 17 d 17 e Mov 18mAA71O/MA642(rSEA)p, p=0,1,2,3,4

Figure 7. The cytotoxicity of a panel of mAb-SEA conjugates with different tumor selectivity: (a) The cytotoxicity of C215(rSEA) 15h and C242(rSEA), 15i against colon carcinoma cell line Colo 205, (b) the cytotoxicity of C215(rSEA), 15h and D612(rSEA), 176 against the colon carcinoma cell line LoVo, (c) the cytotoxicity of C215(rSEA), 15h and MovlS(rSEA), 17c against the ovarial carcinoma cell line HeLa, (d) the cytotoxicity of C215(rSEA), 15h, MA642(rSEA), 17d, and 710(rSEA), 17e against the breast cancer cell line MDA-MD361, (e and f) the chemical composition of 15h, 1 5 , 17b, 17c, 17d, and 17e measured with SDS-PAGE.

the reaction between rSEA(SH), and iodoacetyl-modified mAb and probably also with maleimide modified mAb. Less rSEA(SH), was therefore needed and the yield was better. The two methods developed for C215 and described in this paper have been shown to work very well for other mAbs examined (Tables I and 11). Method 2 with maleimide-modified mAb is the method of choice as it is less laborious than method 1and requires less rSEA. mAb(rSEA), conjugates represent a novel class of antitumor agents which has potential as new drugs for cancer therapy. We have conjugated about 15 different tumor-selective mAbs with SEA using the presently described methods. More than 50% of the mAb-SEA conjugates are potent antitumor agents in vitro. The tumor cell spectrum sensitive to mAb-SEA effects has included

colon, ovarial, mammary (Figure 7)and renal cancer (data not shown). This suggests that mAb-SEA conjugates may be of therapeutic value against a broad range of tumors and warrant further investigation as anticancer agents in vivo. ACKNOWLEDGMENT

We thank Ms. Elisabeth Hugosson, Ms. Ingegerd Andersson, Ms. Lill Ivarsson, and Mr. Per Anders Bertilsson for excellent technical assistant and Dr. Per Bjork for purification of some of the antibodies, and Dr. D. Eaker for modification of the amino acid analysis to measure the 17-amino-3,6,9,12,15-pentaoxaheptadecanoic acid. Many thanks to Mr. Goran Forsberg and Ms. Eva Lagerholm for

466

Biocon/ugate Chem., Vol. 4, No. 6, 1993

Akerblom et ai.

SADCC

/-\/

-

=

qmor ntigen

Ig-SEA con'ug.

Tumor cell

'

w T-cell

Figure 8. A cytotoxicT-cellwith TCR expressingthe appropriate TCR VBsequence binds to the mAb(rSEA), conjugate on a target. The lytic activity of the T-cell is activated and the tumor cell lysed (superantigen-antibody-dependentcell-mediated cytotoxicity, SADCC).

purification of rSEA and to Dr. Peter Lind for preparation of Figure 8. LITERATURE CITED

(1) Houghton, A. N., and Scheinberg, D. S. (1991) Monoclonal antibodies. In Comprehensive Textbook of Oncology (A. K. Moosa, C. Schimpff, and M. C. Robson, Eds.) p 601, Williams & Wilkins, Baltimore. (2) Harris, D. T., and Mastrangelo, M. J. (1989) Serotherapy of cancer. Sem. Oncol. 16,180-198. (3) Jain, R. K. (1987) Transport of molecules in the tumor interstitium: A review. Cancer Res. 47, 3039-3051. (4) Foon, K. A. (1989) Biological response modifiers: The new immunotherapy. Cancer Res. 49, 1621-1639. ( 5 ) Marrack, P., Kappler, J. (1990) The staphylococcal enterotoxins and their relatives. Science 248, 705-711. (6) Dohlsten, M., Hedlund, G., Akerblom, E., Lando, P. A., and Kalland, T. (1991) Monoclonal a n t i b o d y - t a r g e t e d superantigens: A different class of anti-tumour agents. Proc. Natl. Acad. Sci. U.S.A. 88, 9287-9291. (7) Wong, S. S. (1991) Chemistry of protein conjugation and cross-linking, pp 267-288, CRC Press, Inc., Boca Raton, FL. (8) Marsh, J. W., and Neville, D. M., Jr. (1988)A flexible peptide spacer increases the efficacy of holoricin anti-T-cell immunotoxins. J. Immunol. 140, 3674-3678. (9) von Kiederowski, G. (1991) Light-directed parallel synthesis of up to 250 OOO different oligopeptides and oligonucleotides. Angew. Chem. Int. Ed. Engl. 30,822-823. (10) Juhlin, C., Lundgren, S., Johansson, H., Lorenzten, J.,Rask, L., Larsson, E., Rastad, J., Akerstrom, G., and Klareskog, L. (1990) 500-Kilodalton calcium sensor regulating cytoplasmic Ca2+in cytotrophoblast cellsof human placenta. J. Biol. Chem. 265,8275-8279. (11) Oosterwijk, E., Ruiter, D. J., Hoedemaeker, Ph. J., Pauwels, E. K. J., Jonas, U., Zwartenduk, J., and Warnaar, S. 0. (1986) Monoclonal antibody G250 recognizes a determinant present in renal-cell carcinoma and absent from normal kidney. Int. J. Cancer 38,489-494. (12) Muraro, R., Nuti, M., Natali, P. G., Bigotti, A., Simpson, J. F., Primus, F. J., Colcher, D., Greiner, J. W., and Schlom, J. (1989) A monoclonal antibody (D612) with selective reac-

tivity for malignant and normal gastro-intestinal epithelium. Int. J. Cancer 43,598-607. (13) Pastan, I., Lovelace, E. T., Gallo, M. G., Rutherford, A. V., Magnani, J. L., and Willingham,M. C. (1991) Characterization of monoclonal antibodies B1 and B3 that react with mucinous adenocarcinomas. Cancer Res. 51,3781-3787. (14) Burchell, J., Gendler, S., Taylor-Papadimitriou, J., Girling, A., Lewis, A,, Millis, R., and Lamport, D. (1987) Development and characterization of breast cancer reactive monoclonal antibodies directed to the core protein of the human milk mucin. Cancer Res. 47,5476-5482. (15) Tzeng, J. J., Barth, R. F., Johnson, C. W., and Adams, D. M. (1990) Distinct and non-crossreactive epitopes are recognized on B16 melanoma by LAK cells andanti-B16 monoclonal antibodies. Biochim. Biophys. Acta 1026, 69-79. (16) Miotti, S., Canevari,S., Menard, S., Mezzanzanica,D., Porro, G., Pupa, S. M., Regazzoni, M., Tagliabue, E., and Colnaghi, M. I. (1987) Characterization of human ovarian carcinoma associated antigens defined by novel monoclonal antibodies with tumor-restricted specificity. Int. J. Cancer 39,297-303. (17) Opitz, H. G., Opitz, U., Hewlett, G., and Schlumberger, H. D. (1982) A new model for investigations of T-cell functions in mice: Differential immunosuppressive effects of two monoclonal anti-Thy-1.2 antibodies. Immunobiol. 160,438453. (18) Agback, H., Ahrgren, L., Haraldsson, M., and Akerblom, E. (1991) A conjugate substance, a reagent and novel polyethers. International patent appl. PCT/SE91/00497 (Publication no W092/01474). (19) Carlsson, J., Drevin, H., and Axen, R. (1978) Protein thiolation and reversible protein-protein conjugation. N-succinimidyl3-(2-pyridyldithio)propionate,a new heterobifunctional reagent. Biochem. J. 173, 723-737. (20) Dohlsten, M., Lando, P. A., Hedlund, G., Trowsdale, I., and Kalland, T. (1990) Targeting of human cytotoxic T-lymphocytes to MHC class I1 expressing cells by staphylococcal enterotoxins. Immunology 71,96-100. (21) Dohlsten, M., Hedlund, G., Segren, S., Lando, P., Herrman, T., Kelly, A., and Kalland, T. (1991) Human major histocompatibility complex class II-negative colon carcinoma cells p r e s e n t staphylococcal superantigens t o cytotoxic T-lymphocytes: Evidence for a novel enterotoxin receptor. Eur. J. Immunol. 21, 1229-1233. (22) Cumber, A. J., Westwood, J. H., Henry, R. V., Parnell, G. D., Coles, B. F., and Wawrzynczak, E. J. (1992) Structural features of the antibody linkage that influence the activity and stability of ricin A chain immunotoxins. Bioconjugate Chem. 3,397-401. (23) Powers, S. P., Foo, I., Pinon, D., Klueppelberg, U. G., Hedstrom, J. F., and Miller, L. J. (1991) Use of photoaffinity probes containing poly(ethy1ene glycol) spacers for topographical mapping of cholecystokinin receptor complex. Biochemistry 30, 676-682. (24) Kitagawa, T., Shimozono, T., Aikawa, T., Yoshida, T., and Nishimura, H. (1981) Preparation and characterization of hetero-bifunctional crosslinking reagents for protein modifications. Chem. Pharm. Bull. 29,1130-1135. (25) Marsh, J. W., and Neville Jr., D. M. (1986) Kinetic comparison of ricin immunotoxins: Biricin conjugates has potensiated cytotoxicity. Biochemistry 25, 4461-4467. (26) Baldwin, R. W., and Byers, V. S. (1987) Monoclonalantibody targeting of cytotoxic agents for cancer therapy. In Immunology of malignant diseases (V. S . Byers, and R. W. Baldwin, Eds.) pp 44-54, MTP press, Lancaster.