Role of Cross-Linking Agents in Determining the Biochemical and

derived from Aspergillus clavatus, to the monoclonal antibody Mgr6 that recognizes an epitope of the .... expression of a new RIP, clavin, from Asperg...
1 downloads 0 Views 141KB Size
Download PDF
372

Bioconjugate Chem. 1998, 9, 372−381

Role of Cross-Linking Agents in Determining the Biochemical and Pharmacokinetic Properties of Mgr6-Clavin Immunotoxins Franco Dosio,† Silvia Arpicco,† Elena Adobati,‡ Silvana Canevari,‡ Paola Brusa,† Rita De Santis,§ Dino Parente,§ Paola Pignanelli,§ Donatella R. M. Negri,‡ Maria I. Colnaghi,‡ and Luigi Cattel*,† Dipartimento di Scienza e Tecnologia del Farmaco, University of Torino, Torino, Italy, Divisione di Oncologia Sperimentale E, Istituto Nazionale per lo Studio e la Cura dei Tumori, Milano, Italy, and Dipartimento di Biotecnologie, Menarini Ricerche Sud, Pomezia, Italy. Received October 31, 1997; Revised Manuscript Received January 27, 1998

Several immunotoxins (ITs) were synthesized by the attachment of clavin, a recombinant toxic protein derived from Aspergillus clavatus, to the monoclonal antibody Mgr6 that recognizes an epitope of the gp185HER-2 extracellular domain expressed on breast and ovarian carcinoma cells. Conjugation and purification parameters were analyzed in an effort to optimize the antitumor activity and stability of the ITs in vivo. To modulate the in vitro and in vivo properties of the immunotoxins, different coupling procedures were used and both disulfide and thioether linkages were obtained. Unhindered and hindered disulfide with a methyl group linkage ethyl S-acetyl 3-mercaptopropionthioimidate ester hydrochloride (AMPT) or ethyl S-acetyl 3-mercaptobutyrothioimidate ester hydrochloride (M-AMPT) were obtained by reaction with recombinant clavin, while the monoclonal antibody Mgr6 was derivatized with ethyl 3-[(4-carboxamidophenyl)dithio]propionthioimidate ester hydrochloride (CDPT). To achieve higher hindrance (a disulfide bond with a geminal dimethyl group), Mgr6 was derivatized with the N-hydroxysuccinimidyl 3-methyl-3-(acetylthio)butanoate (SAMBA) and clavin with CDPT. To evaluate the relevance of the disulfide bond in the potency and pharmacokinetic behavior of the ITs, a conjugate consisting of a stable thioether bond was also prepared by derivatizing Mgr6 with the N-hydroxysuccinimidyl ester of iodoacetic acid (SIA) and clavin with AMPT. The immunotoxins were purified and characterized using a single-step chromatographic procedure. Specificity and cytotoxicity were assayed on target and unrelated cell lines. The data indicate that the introduction of a hindered disulfide linkage into ITs has little or no effect on antitumor activity and suggest that disulfide cleavage is essential for activity; indeed, the intracellularly unbreakable thioether linkage produced an inactive IT. Analysis of IT stability in vitro showed that the release of mAb by incubation with glutathione is proportional to the presence of methyl groups and increases exponentially with the increase in steric hindrance. Analysis of the pharmacokinetic behavior of ITs in Balb/c mice given intravenous bolus injections indicated that ITs with higher in vitro stability were eliminated more slowly; i.e., the disulfide bearing a methyl group doubled the β-phase half-life (from 3.5 to 7.1 h) compared with that of the unhindered, while a geminal dimethyl protection increased the elimination phase to 24 h. The thioether linkage showed its intrinsic stability with a β-phase half-life of 46 h. The thioether linkage also increased the distribution phase from 17 to 32 min. The in vitro characteristics and in vivo stability of Mgr6-clavin conjugates composed of a methyl and dimethyl steric hindered disulfide suggest clinical usefulness.

INTRODUCTION

The potent and highly specific cytocidal activity of conjugates between monoclonal antibodies and toxins has raised the promise of effective immunotherapy for tumors and graft-versus-host disease. ITs1 have particular potential in the treatment of leukemias and lymphomas (1, 2), although some evidence suggests they are effective against solid tumors as well (3, 4). Extremely active ITs have been constructed by covalently linking mAb to bacterial toxins such as Pseudomonas exotoxin or to plant toxins such as the ricin A chain, saporin, and gelonin (5-8). * Correspondence should be addressed to Luigi Cattel, Dipartimento di Scienza e Tecnologia del Farmaco, v. P. Giuria 9, 10125 Torino, Italy. Phone: ++39-11-6707697. Fax: ++39-116707695. E-mail: [email protected]. † University of Torino. ‡ Istituto Nazionale per lo Studio e la Cura dei Tumori. § Menarini Ricerche Sud.

Other interesting ribosome-inactivating proteins (RIPs) with potent antitumor activity such as R-sarcin, restrictocin, and mitogillin have been isolated from fungi and used to prepare ITs (9-11). RIPs of fungi are extensively homologous in their primary structure (>80%) and have the same catalytic activity, i.e., hydrolysis of the phosphodiester linkage between the G residue at position 4325 and the A residue at position 4326 in 28S rRNA (12). 1 Abbreviations: ITs, immunotoxins; AMPT, ethyl S-acetylpropionthioimidate; M-AMPT, ethyl S-acetyl-3-mercaptobutyrothioimidate; CDPT, 3-[(4-carboxamidophenyl)dithio]propionthiomidate; SAMBA, N-hydroxysuccinimidyl 3-methyl-3-(acetylthio)butanoate; SIA, N-hydroxysuccinimidyl ester of iodoacetic acid; RIP, ribosome-inactivating protein; DTNB, Ellman’s reagent; DTT, dithiothreitol; TNB, 5-mercapto-2-nitrobenzoic acid; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; GSH, reduced glutathione; SMPT, 4-[(succinimidyloxy)carbonyl]-R-methyl-R-(2-pyridyldithio)toluene.

S1043-1802(97)00192-4 CCC: $15.00 © 1998 American Chemical Society Published on Web 04/16/1998

Properties of Mgr6−Clavin Immunotoxins

Recently, Parente et al. (13) described the cloning and expression of a new RIP, clavin, from Aspergillus clavatus. Clavin was highly active on free ribosomes, inducing low and transient systemic toxicity and a late, low-level antibody response in mice. These characteristics and its availability in large amounts identify clavin as a good candidate for IT preparation. To selectively target clavin to breast and ovarian carcinoma cells, we used a mAb (Mgr6) that recognizes an epitope of the gp185HER-2 extracellular domain (14). The tissue distribution of gp185HER-2 in normal human tissues is restricted, and overexpression of the HER-2 oncogene is associated with poor prognosis in breast and ovarian carcinoma patients (15, 16). Existing ITs directed to gp185HER-2 have already shown immunotherapeutic promise (17-19). In this study, we analyzed the conjugation and purification parameters that might yield efficient and stable ITs suitable for preclinical studies. We prepared various ITs and characterized them with respect to coupling systems (disulfide or thioether linkage). In fact, most of the ITs prepared to date have a reducible disulfide bond which is required for the intracellular release of the toxic moiety (20). The major in vivo limitations of such ITs rest in the instability of the chemical linkage, which restricts the amount of conjugate that is able to bind to target cells (21, 22), and in the potential competition of the released antibody with the intact conjugate since the mAb persists longer in circulation. Previously released antibody may mask tumor antigens, thereby compromising IT potency during multiple-dose treatment regimens. In an effort to prolong the in vivo half-life of ITs and enhance their antitumor potency, we prepared Mgr6clavin conjugates using the thioimidate cross-linking system which preserves the positive charge on the derivatized proteins and allows modulation of the hindrance on the chemical linkage between clavin and the mAb. EXPERIMENTAL PROCEDURES

Materials. Anhydrous dichloromethane, Ellman’s reagent (DTNB), and all other reagents were from Aldrich (Milwaukee, WI). Diethyl ether was distilled from lithium aluminum hydride, and dry tetrahydrofuran was obtained by distillation from sodium. General Procedures. Melting points were determined with a Reichert Kofler apparatus and are uncorrected. 1H NMR spectra were recorded on a JEOL PMX60 spectrometer, operating at 60 MHz, with tetramethylsilane as the internal standard. IR spectra were obtained as KBr disks on a Shimadzu FT-IR 8101 M spectrophotometer; wavelengths are given in inverse centimeters. Mass spectra were obtained with a FINNIGAN-MAT TSQ-700 or with a VG Analytical 70-70 EQHF spectrometer. Ultraviolet spectra were recorded on a Beckman DU-70 spectrophotometer. Reactions were checked on F254 silica gel precoated sheets (Merck, Milan, Italy). Purification was carried out by column flash chromatography on silica gel 60 (Merck, 230-400 mesh). Synthesis of Cross-Linkers. Preparation of Thioimidate Ester Hydrochlorides (CDPT, AMPT, and M-AMPT). Hydrogen chloride gas was bubbled through ice-cold ethanethiol (3.25 mL, 0.0435 mol) for 1 h. The nitriles 4-[(2-cyanoethyl)dithio]benzamide, 3-(acetylthio)propionitrile, and 3-(acetylthio)butyronitrile, synthesized

Bioconjugate Chem., Vol. 9, No. 3, 1998 373

as described by Arpicco (23), were diluted in anhydrous diethyl ether or dry tetrahydrofuran and quickly added to the cold solution with stirring, and the reaction mixture was left overnight at 0 °C. Anhydrous cold diethyl ether was then added, and the reaction mixture was left at -20 °C until a crystalline solid formed. The supernatant was decanted, and the precipitate was washed with anhydrous diethyl ether under argon and dried under reduced pressure at room temperature. Ethyl 3-[(4-carboxamidophenyl)dithio]propionthioimidate ester hydrochloride (CDPT): yield 0.64 g (65%); mp 110115 °C; 1H NMR (CD3OD) δ 7.85 (q, 4H, Ar-H), 3.4-3.15 (m, 6H, SCH2CH3 and CH2CH2), 1.47 (t, 3H, SCH2CH3); IR (KBr) 3500-2200 (NH2 and NH2+), 1650 (CdO), 1620 (CdN); MS (FAB+) 301 (M+ + 1). Ethyl S-acetyl 3-mercaptopropionthioimidate ester hydrochloride (AMPT): yield 0.87 g (88%); mp 64-66 °C; 1 H NMR (CDCl3) δ 3.48 (q, 2H, SCH2CH3), 3.26 (m, 4H, CH2CH2), 2.38 (s, 3H, SAc), 1.45 (t, 3H, SCH2CH3); IR (KBr) 3300-2400 (NH), 1690 (CdO), 1620 (CdN), 1360 (SAc). Ethyl S-acetyl 3-mercaptobutyrothioimidate ester hydrochloride (M-AMPT): yield 0.79 g (75%); mp 112 °C; 1 H NMR (CDCl3) δ 3.4-3.2 (m, 5H, SCH2CH3, CH, and CH2), 2.43 (s, 3H, SAc), 1.6 (d, 3H, CH3), 1.43 (t, 3H, SCH2CH3); IR (KBr) 3300-2400 (NH), 1690 (CdO), 1620 (CdN), 1360 (SAc); MS-EI m/z (relative intensity) 206 (M+, 7), 162 (100), 145 (53), 130 (20), 102 (65), 89 (20), 75 (33), 61 (52), 43 (100). Preparation of N-Hydroxysuccinimidyl 3-Methyl-3(acetylthio)butanoate (SAMBA). The method of Carrol (24) was used with minor modifications. Briefly, 3-methyl-3-(acetylthio)butanoic acid (1.5 g, 0.0085 mol) dissolved in 12 mL of dry dichloromethane was mixed with 1.3 g of NHS (0.011 mol) in 5 mL of dichloromethane, and dicyclohexylcarbodiimide (2.3 g, 0.011 mol) dissolved in 1 mL of dry dichloromethane was added dropwise. The reaction mixture was stirred at room temperature for 22 h. After filtration and evaporation, the mixture was purified by flash chromatography with elution in hexane/ EtOAc (80/20). The ester was obtained as a pale yellow oil that crystallized rapidly at room temperature: yield 1.6 g (70%); mp 63 °C; TLC (80/20 hexane/EtOAc) Rf ) 0.15; 1H NMR (CDCl3) δ 3.23 (s, 2H, CH2), 2.80 (s, 4H, NHS ester), 2.25 (s, 3H, SAc), 1.6 (s, 6H, CH3); MS-EI m/z (relative intensity) 273 (10), 159 (90), 117 (30), 75 (15), 43 (55). The N-hydroxysuccinimidyl ester of iodoacetic acid (SIA) was prepared as described (25). Clavin and Mgr6 Production. For clavin production, a 100 mL culture of Escherichia coli HB101 cells carrying the pMRS38 plasmid was used to start the final cultivation step in a Chemap bioreactor with a working volume of 1.5 L. The cultivation was performed in complex medium [20 g/L glucose, 25 g/L yeast extract, 40 g/L casamino acid (Difco), 0.5 g/L NaCl, 5 g/L KCl, 2.6 g/L K2SO4, 0.86 g/L MgCl2‚6H2O, 6.6 mg/L CaCl2‚ 6H2O, and trace amounts of oligo elements], supplemented with 100 mg/mL ampicillin. Cultivation parameters such as temperature, pH, and oxygen dissolved in tension (DOT) were computer-controlled. The culture supernatant was recovered by centrifugation and filtered on a Millipore 0.22 µm filter. Clavin was purified as described (13). Mgr6 hybridoma, directed against the extracellular domain of gp185HER-2, was obtained by immunization of Balb/c mice with the adenocarcinoma cell line CaLu3, and its characterization is described elsewhere (14). mAb Mgr6 was obtained by cultivating hybridoma cells in a

374 Bioconjugate Chem., Vol. 9, No. 3, 1998

hollow fiber bioreactor (Acusyst R, Endotronics) using a serum-free medium. Medium, prepared in our facilities, consisted of a basal mixture (1/1) of RPMI 1640 and DMEM supplemented with transferrin, insulin, albumin, and Ex-Cite (Miles Inc.) as a lipid source. Additional ingredients were 2-mercaptoethanol, ethanolamine, and sodium selenite. Cell cultivation was performed as continuous perfusion for approximately 1 month, during which parameters such as glucose and lactic acid concentrations, pH, oxygen and carbon dioxide levels, and antibody production were monitored along with operator responses. Conditioned media harvested during cultivation were pooled after thawing and adjusted to pH 5.5. Material was diluted and filtered on 4.5 µm filters (Millipore Co., Bedford, MA) before being loaded on a Bakerbond Abx column (Mallinckrodt Baker B. V., Deventer, The Netherlands). The eluted protein peak was desalted on a Sephadex G-25 (Pharmacia-Biotech, Uppsala, Sweden) column and loaded onto a Q-Sepharose FF column (Pharmacia-Biotech). The eluted peak was desalted as indicated above and subjected to the final purification step on a S-Sepharose FF column (Pharmacia-Biotech). Elution was obtained by applying a linear gradient from 20 to 200 mM NaCl in 20 mM phosphate buffer. Preparation of Immunotoxins. Disulfide Bridge. In a standard preparation, a recombinant clavin solution (357 µM, 1 mL) was stirred with AMPT (122 mM, 30 µL dissolved in ethanol), M-AMPT (98 mM, 30 µL dissolved in ethanol), or CDPT (139 mM, 40 µL in dry dimethylformamide) for 30 min at 25 °C to incorporate an average of 1-1.3 linkers per mole of clavin. The mixture was purified by gel filtration on a 1 × 20 cm Bio-Gel P6-DG column (Bio-Rad, Hercules, CA) eluted with HBS (0.1 M, 0.2 M NaCl and 1 mM EDTA disodium salt at pH 7.4) at 20 °C. The protein solution was concentrated to 1 mL with an Amicon concentrator (Amicon, Beverly, MA) equipped with a Y10 membrane. mAb Mgr6, dissolved in HBS, was separately reacted with two different cross-linkers, CDPT and SAMBA. The reaction of Mgr6 with CDPT proceeded as follows: CDPT in dry dimethylformamide (13 mM, 40 µL) was added to the mAb in solution (33 µM, 1 mL), and the mixture was stirred for 30 min at 25 °C. The Mgr6/CDPT molar ratio was 1/1.1. Derivatization of Mgr6 with SAMBA (11 mM, 30 µL) dissolved in ethanol was carried out for 30 min at 25 °C to introduce two acetylthio groups. In both cases, the mixture was purified by gel centrifugation on a 15 × 55 mm Bio-Gel P6-DG column preequilibrated in HBS at 20 °C. The number of thioacetylated groups linked to the protein was calculated spectrophotometrically by reaction of the sample with the deacetylating reagent hydroxylamine hydrochloride (0.5 M, 12.5 mM EDTA, pH 7.4) followed by thiol disulfide exchange with DTNB as described (26). The number of aryldithio groups linked to Mgr6 or clavin was determined following the release of the thiolated anion at 313 nm, after incubation of the protein sample (1 mL) with 2-mercaptoethanol in PBS/EDTA (11 mM sodium phosphate and 3 mM EDTA, 50 µL) and NaOH (1 M, 40 µL), to a final pH of 8.8-9.4. The molar absorptivity value of the 4-carboxamidophenylthiolate anion under these conditions at 313 nm was 15 200 ( 300 M-1 cm-1. All conjugations were performed by mixing the derivatized mAb and RIP in the presence of a solution of hydroxylamine (1/10, v/v). Reactions were allowed to proceed for 5 h at 25 °C followed by 18-35 h at 4 °C; at

Dosio et al.

the end of the reaction, a solution of N-ethylmaleimide (20 mM, 20 µL) was added to block free thiol groups. Thioether Bridge. Clavin (357 µM, 1 mL) was derivatized with AMPT as described above; Mgr6 (33 µM, 1 mL) was reacted with SIA (5.7 mM, 30 µL) dissolved in ethanol, for 30 min at 25 °C. The iodoacetyl groups inserted on Mgr6 were identified by reaction with the TNB (5-mercapto-2-nitrobenzoic acid) reagent previously prepared; briefly, 25 mg of dithiothreitol (DTT) was added to 50 mg of DTNB dissolved in 300 µL of 1 M aqueous NaOH under an inert atmosphere. After 30 min, the reaction mixture was acidified with 1 M HCl and the mixture was extracted with diethyl ether under an inert atmosphere. The organic layers were then anhydrificated and evaporated under reduced pressure. The residue was washed several times with ethanol to obtain TNB as a dark yellow product. The Mgr6 derivatization degree was determined by reaction of mAb (3.3 µM, 500 µL) with TNB dissolved in ethanol (3.4 mM, 10 µL) for 90 min at 25 °C. After purification by gel centrifugation, the number of iodoacetyl groups was determined spectrophotometrically (TNB molar absorption coefficient was 8800 M-1 cm-1 at 339 nm). The mean derivatization degree was 1.7. The derivatized mAb and RIP were mixed in the presence of a solution of hydroxylamine (1/10, v/v). The conjugation reactions proceeded for 18 h at 4 °C; at the end of the reaction, a solution of N-ethylmaleimide (20 mM, 20 µL) was added to block free thiol groups. Immunotoxin Purification. Conjugates were purified from unconjugated clavin and Mgr6 in a one-step procedure using CM MemSep cartridges (1000 or 1010) (Millipore) for analytical or preparative application and a Merck-Hitachi 655A-12 HPLC gradient system equipped with an L-5000 LC controller. The eluting fractions were monitored at 280 nm using a L4000 UV detector. Peak heights were recorded and processed on a CBM-10A Shimadzu interface. The mobile phase was sodium acetate buffer (20 mM at pH 5.5) flushed at 1 mL/min. Fractions containing immunoconjugates were dialyzed and concentrated. The purity of the immunotoxins was monitored by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using 4-15% precast gels (Bio-Rad) under nonreducing conditions and Coomassie blue staining. In Vitro Evaluation of Disulfide Bond Stability. The in vitro stability of the bond in the various conjugates was evaluated as follows. Samples (1 mg/mL, 3 µL) were incubated for 1 h at 37 °C with solutions of reduced glutathione (GSH, 3 µL) in increasing excesses (from 3 to 10000-fold), and the reaction was stopped by addition of excess iodoacetamide. Following SDS-PAGE under nonreducing conditions, the 7.5% precast gels (Bio-Rad) were stained with Coomassie blue, dried, and scanned on a Compact 4800 flatbed scanner using Twain compatible software. Band densities were analyzed using ImagePC (Scion Co., Frederick, MD) to calculate the amount of Mgr6 released. Cell Lines. Human breast carcinoma cell line SKBr3 (ErbB2+) was purchased from ATCC (Rockville, MD), and human melanoma cell line MeWo (ErbB2-) was kindly provided by the late J. Fogh (Memorial SloanKettering Cancer Center, New York). Both cell lines were grown in RPMI 1640 containing 10% fetal calf serum and gentamicin (100 µg/mL). Binding Inhibition Assay. mAb activity after conjugation was assayed as the ability to inhibit binding of [125I]Mgr6 to adherent glutaraldehyde-fixed SKBr3 cells. A fixed amount of [125I]Mgr6 (1 nM) was mixed with serial

Properties of Mgr6−Clavin Immunotoxins

dilutions of cold Mgr6 or IT, starting from a 100-fold molar excess. The mixture was added to SKBr3 fixed cells (adherent in 96-well plates) and incubated for 3 h at 37 °C. Cells were washed 10 times with PBS and incubated with 2 N NaOH (100 µL/well) for 20 min at room temperature. The supernatant was collected and radioactivity determined in a γ-counter (Beckman Instruments, Fullerton, CA). Percent inhibition was calculated as follows:

Bioconjugate Chem., Vol. 9, No. 3, 1998 375 Scheme 1. Structures of Cross-Linkers Used to Produce Mgr6-Clavin ITs

% inhibition ) (1 - Ci/Cni) x 100 where Ci is the average counts per minute in the presence of cold inhibitor and Cni is the average counts per minute without inhibitor. Inhibition of Protein Synthesis. The assay was carried out essentially as described (13). Briefly, SKBr3 and MeWo cells were suspended in culture medium containing the appropriate concentration of IT, toxin, or mAb alone and incubated for 3 h at 4 °C. Control cells were incubated with medium alone. Cells were centrifuged, resuspended in fresh culture medium, and seeded in 96-well plates (3 × 105 cells/well). After incubation at 37 °C for 48 h, [3H]proline (1 µCi/well) was added. After 48 h, cells were washed and [3H]proline incorporation was determined by liquid scintillation in a β-counter. Results are expressed as a percentage of [3H]proline incorporation in control cells. Radioiodination Procedure. ITs, clavin, and Mgr6 were 125I-labeled using the Iodogen method according to the manufacturer’s instructions, to a mean specific activity of 1.5 mCi/mg for ITs, 8.5 mCi/mg for Mgr6, and 4.5 mCi/mg for clavin. The integrity of the radiolabeled protein was tested by ascending paper chromatography in 10% trichloroacetic acid and by SDS-PAGE analysis. Pharmacokinetic Evaluation. Experiments were performed on 6-week-old female Balb/c mice (Charles River, Como, Italy) maintained according to the provisions of the European Economic Community Council directive 86/209 recognized and adopted by the Italian Government. Mice received a Lugol solution (0.02% I2) and 0.6 mg/ mL KClO4 in their drinking water 3 days before administration of radioiodinated ITs and throughout the experiments to block free iodine uptake by the thyroid gland and the stomach mucosa. Animals were injected intravenously with a single dose of 125I-labeled IT, ranging from 1.7 to 2.5 µg/mouse. Blood samples were drawn from the retro-orbital sinus at fixed times after injection. Pharmacokinetic analysis was carried out on serum samples. Each serum sample was diluted 1/20 and run in SDS-PAGE under nonreducing conditions. Gels were fixed, dried, and exposed in a PhosphorImager screen (Molecular Dynamics, Sunnyvale, CA). Densitometric analysis, performed using Image Quant software (Molecular Dynamics), was used to determine the percentage of intact immunoconjugates in the serum at each time point. Pharmacokinetic parameters were calculated using PCNONLIN software (SCI Software, Lexington, KY). RESULTS

Chemistry. The cross-linkers, prepared as described (23-25), were characterized by NMR, IR, and MS. Before conjugation, the reactivity of the linkers toward the lysyl -amino groups of IgG as a model protein was evaluated to insert 1-2 groups. Scheme 1 shows the structures of the cross-linkers.

Preparation of Immunotoxins. Four different coupling procedures (Schemes 2 and 3) were followed to compare the relative activity, stability, and in vivo fate of the immunotoxins. Disulfide Bridge Immunotoxins. Linkers AMPT and M-AMPT were reacted with recombinant clavin to introduce 1-1.3 S-acetylthio groups per protein molecule, whereas a mean value of 1.1 aryldithio residues was introduced into Mgr6 with CDPT; the derivatized proteins were coupled in the presence of hydroxylamine to deprotect the acetylthio groups (IT-1 and IT-2 in Scheme 2). Clavin modified with CDPT was reacted with the Mgr6 previously reacted with the hindered N-hydroxysuccinimidyl linker SAMBA (IT-4) (Scheme 3). All coupling reactions were allowed to proceed for 5 h at 25 °C and for 18-35 h at 4 °C; aliquots of the reaction mixtures were analyzed by SDS-PAGE. Thioether Bridge Immunotoxin. To obtain a thioether linkage, Mgr6 was derivatized with SIA (1.7 iodoacetyl groups) and clavin with AMPT as described above. After addition of hydroxylamine, the reaction proceeded for 18 h at 4 °C (IT-3 in Scheme 2). The yield of conjugation for the thioether-linked IT was 12%, while for the disulfide-ITs, the yield was approximately 10% for IT-1 and IT-2 and 5% for IT-4. This suggested an inverse correlation between disulfide bond stability and conjugation efficiency. The low yield can be attributed in part to the low (range of 1-1.7) number of reactive groups per protein. Purification of Immunotoxins. All ITs were purified from the reaction mixture in a single step by HPLCion exchange chromatography using a 1010 CM MemSep cartridge. Figure 1A shows an example of the elution profile of the crude mixture eluted using a discontinuous gradient of sodium chloride ranging from 0 to 250 mM. Fractions from each of the four regions of the elution peaks were separately pooled and subjected to nonreducing SDS-

376 Bioconjugate Chem., Vol. 9, No. 3, 1998

Dosio et al.

Scheme 2. Preparation of Mgr6-(CDPT)SS(AMPT)-Clavin (IT-1), Mgr6-(CDPT)SS(M-AMPT)-Clavin (IT-2), and Mgr6-(SIA)CS(AMPT)-Clavin (IT-3)

Scheme 3. Preparation of Mgr6-(SAMBA)SS(CDPT)-Clavin (IT-4)

PAGE analysis (Figure 1B). By comparison with purified Mgr6 and clavin electrophoresed in parallel, peak 1 was composed only of unreacted Mgr6, peak 2 was 80% IT with a molecular mass of 167 kDa, peak 3 was >99% IT (167 kDa), and peak 4 contained only unreacted clavin (17 kDa). Disulfide Bond Stability of Conjugates in Vitro. To test the in vitro stability of the immunotoxins, samples were incubated with reduced GSH in different excesses, subjected to SDS-PAGE, and assessed densitometrically (Figure 2). The profile of the disulfide rupture as a function of GSH excess used is shown in Figure 3. An increase in disulfide stability in relation to the steric hindrance around the disulfide of the conjugates was observed. In comparison, the presence of a thioether bond precluded the breakage by GSH. The increase in stability was exponential, i.e., 10-fold for one methyl group and 100-fold for a geminal dimethyl group. When GSH was used in an excess higher than 7000-fold,

cleavage of the mAb disulfide bond was measurable (data not shown). In Vitro Activity of ITs. To determine whether Mgr6 retained its binding capacity after derivatization and conjugation, competitive inhibition experiments were conducted with the four ITs. A small, reproducible drop in binding capacity was observed after derivatization that was independent of the cross-linker used (see Figure 4, Mgr6-CDPT curve). All ITs competed with the native [125I]mAb, with a titration similar to that of cold derivatized Mgr6, and no significant differences due to the linkage method adopted were evident (Figure 4). Each IT was tested for its ability to inhibit protein synthesis in ErbB2-positive SKBr3 cells compared with the inhibition by unconjugated clavin. All ITs specifically inhibited protein synthesis in the SKBr3 cells, whereas none did so in ErbB2-negative MeWo cell. As shown in Figure 5, the IC50s of the conjugates containing a disulfide bridge were similar and ranged between 1 and 4 nM; the

Properties of Mgr6−Clavin Immunotoxins

Bioconjugate Chem., Vol. 9, No. 3, 1998 377

Figure 1. Elution profile of the IT reaction mixture from the CM-MemSep column (A) and SDS-PAGE analysis of eluted fractions (B). For preparative purposes, a MemSep 1010 cartridge eluted in sodium acetate buffer (20 mM at pH 5.5) was loaded with 400 µL of conjugation mixture, and a step gradient of the same buffer and 1 M NaCl was applied, at a 0.5 mL/min flow rate for the first 12 min and at 2 mL/min thereafter. After 20 min, the gradient was raised over the course of 10 min to 100 mM NaCl and kept constant for another 10 min. To remove unconjugated clavin, the gradient was increased to 250 mM over the course of 10 min and held there for 20 min. The collected fractions (about 3.5 mL) were immediately brought to pH 7.4 by addition of a solution of 1 M PBS (pH 8).

Figure 2. SDS-PAGE analysis of IT-2 incubated with GSH at different concentrations for 1 h at 37 °C.

Figure 4. Competition of [125I]Mgr6 binding to SKBr3 (HER2+) glutaraldehyde-fixed cells with increasing molar concentrations of unlabeled mAb, derivatized mAb, and ITs (SD < 5% of the means).

Figure 3. Release of mAb from the immunoconjugates. The conjugates were incubated with GSH at different concentrations for 1 h at 37 °C. The amounts released were measured by densitometry after SDS-PAGE. Results are the average of three independent experiments (SDs < 10% of the mean).

conjugation to the mAb resulted in a 150-650-fold increase in cytotoxic activity compared to that of clavin alone. No toxicity was observed after treatment with the conjugate containing the thioether bond (IT-3), which cannot be cleaved inside the cell. The specificity of the

cytotoxicity was confirmed by the absence of an effect on MeWo cells (IC50s for all ITs are >1 µM vs an IC50 of 0.8 µM for clavin). Pharmacokinetics. To evaluate the in vivo stability of ITs with a hindered (IT-2 and IT-4) or an unhindered (IT-1) disulfide bond or a thioether bond (IT-3), the pharmacokinetic behavior of each species was evaluated after intravenous bolus administration of 125I-labeled IT in groups of three Balb/c mice. Blood samples were collected at various times, and radioactivity was determined. Plasma samples were analyzed by SDS-PAGE, autoradiography, and densitometry. The unhindered disulfide conjugate (IT-1) broke down to release the free

378 Bioconjugate Chem., Vol. 9, No. 3, 1998

Dosio et al.

Figure 5. Inhibition of [3H]proline incorporation in SKBr3 target cells. Values represent the arithmetic means of three determinations (SDs < 10% of the means).

antibody in vivo at the fastest rate, and after 6 h, only 12% of the intact IT was present as compared with 35% of the IT-2 at the same time point. Plasma samples collected from IT-4 injected mice contained approximately equal amounts of intact immunotoxin and released mAb 47 h after injection. IT-3 appeared to be stable, and by densitometric analysis, the conjugate represented 8690% of the total radioactivity 72 h after injection. Plasma concentrations of IT at each time point were used to construct pharmacokinetic curves for each conjugate (Figure 6). Table 1 lists the parameters obtained from computerized analysis of the clearance data using an open two-compartment pharmacokinetic model. Data for the single components of the ITs are included for comparison. Mgr6 showed a long serum elimination phase, whereas clavin, due to its low molecular mass and relevant charge, was cleared rapidly with only 3% of the initial radioactivity present after 1 h. The clearance values for clavin are shorter than those we reported previously (13) due possibly to the different method used for evaluation. The elimination curves of the ITs were biphasic, with a similar initial rapid R phase (about 17 min) followed by a slower β phase. The immunoconjugates with higher in vitro stability were eliminated more slowly. Indeed, IT-1 was cleared rapidly; IT-2 doubled the β-phase halflife (from 3.5 to 7.1 h), and IT-4 increased the elimination phase to 24 h. The thioether linkage showed its intrinsic stability, with a β-phase half-life of 46 h. The difference in the elimination rate (144 h for the mAb and 46 h for IT-3) is due only to the presence of the clavin molecule. The thioether linkage stability also increased the distribution phase (from 17 to 32 min). DISCUSSION

Immunotoxins have shown promise in therapeutic applications for hematological malignancies (1, 2, 27, 28); however, the therapy of disseminated solid tumors has proven to be a greater challenge, and only few trials have provided encouraging results (3, 29). Several investigators used a reduced size IT Fab linked to the toxic domain

Figure 6. Pharmacokinetic serum curves for ITs, clavin, and Mgr6. Mice were injected intravenously with the proteins, and blood samples were obtained at various time intervals. Concentrations were determined by assay of 125I radioactivity and by densitometric analysis of SDS-polyacrylamide gels. Note the different clearance times in panels A and B.

of diphteria toxin or Pseudomonas exotoxin to allow sufficient tissue permeation; however, this approach was not efficient in the treatment of solid tumors due to faster increased in vivo clearance (30). We and others have used whole ITs (mAb-RIP) with high potency and good pharmacokinetic properties (6, 10, 31, 32). In this study, we chose an anti-erbB2 mAb as the RIP targeting component. A large array of murine mAbs directed to HER2 have been generated (33), and several, including Mgr6, induce down-modulation of the oncoprotein and inhibition of tumor growth. The antitumor effect of these and related mAbs is significantly improved by chemical or molecular coupling with toxins (34, 35). A phase I study in breast cancer patients indicated the clinical feasibility of intravenous administration of up to 80 mg of Mgr6 (our unpublished results). Preliminary data in nine breast cancer patients indicate that Mgr6 has a mean elimination half-life of 37 h and does not

Properties of Mgr6−Clavin Immunotoxins

Bioconjugate Chem., Vol. 9, No. 3, 1998 379

Table 1. Pharmacokinetic Parametersa A (µg/mL) B (µg/mL) t1/2(R) (h) t1/2(β) (h) AUC (µg h mL-1)b Cl (mL/h) Vss (mL) MRT (h)

clavin

Mgr6

IT-1

IT-2

IT-4

IT-3

0.17 ( 0.1 0.015 ( 0.007 0.05 ( 0.02 0.7 ( 0.1 0.03 ( 0.01 27.2 ( 4.3 8.05 ( 2.1 0.28 ( 0.1

0.61 ( 0.06 0.34 ( 0.05 3.76 ( 0.9 144.17 ( 12 73.7 ( 8.2 0.033 ( 0.007 6.74 ( 1 198.9 ( 20

0.26 ( 0.04 0.21 ( 0.05 0.26 ( 0.06 3.58 ( 0.21 1.197 ( 0.22 1.0 ( 0.18 4.78 ( 0.87 4.77 ( 1.2

0.51 ( 0.09 0.31 ( 0.09 0.28 ( 0.13 7.17 ( 1.6 3.49 ( 0.9 0.46 ( 0.1 4.46 ( 1.2 9.75 ( 2.1

0.39 ( 0.07 0.44 ( 0.03 0.24 ( 0.18 24.1 ( 1.1 15.41 ( 1.7 0.11 ( 0.016 3.81 ( 0.26 34.4 ( 3.1

0.748 ( 0.03 0.495 ( 0.02 0.54 ( 0.07 46.1 ( 3.1 33.53 ( 3.2 0.071 ( 0.007 4.67 ( 0.2 65.3 ( 6.2

a Values were obtained by adopting a two-compartment open pharmacokinetic model. b The trapezoidal rule has been used to calculate the area under the curve values.

induce a human anti-mouse antibody response (manuscript in preparation). As the toxic agent, we used a novel recombinant RIP, clavin, characterized by a low molecular mass (17 kDa), high in vitro activity, and reduced immunogenicity. The latter is especially relevant because almost all of the RIPs used for IT preparation such as PAP, saporin, momordin, and bryodin are highly immunogenic, and the antibody moiety of most IT conjugates elicits an even stronger immune reaction (36). Those findings are consistent with results obtained in patients receiving IT (37). Clavin also shows a very low-level, transient, and reversible systemic toxicity on kidney and liver. This aspect is important because one of the major problems in using RIPs, such as PAP-1, PAP-S, RTA, and saporin, is related to their irreversible liver and kidney damage (38, 39). To be therapeutically effective, ITs composed by type I RIPs must be linked in such a way that their are easily cleaved intracellularly but sufficiently stable in the blood circulation. A stable disulfide bond is essential in maintaining the half-life and potency of ITs, especially for solid tumors. The use of sterically hindered linkers such as SMTP to construct ITs with deglycosylated RTA is widely described, and these ITs are being used in phase I and II clinical studies (1, 2). We therefore chose a series of thioimidate linkers which preserve the positive charge on derivatized proteins and generate both a labile disulfide bond and different steric hindrances. The maintenance of a positive charge is especially important when using toxins such as clavin which is a relatively small molecule with a high positive charge. Using the thioimidate reagents, we prepared three ITs with a disulfide bond and one with a thioether linkage; all were expected to be stable in vivo. A geminal dimethyl hindrance was inserted using the SAMBA reagent because the analogous thioimidate linker (in particular its nitrile precursor) tended to polymerize and appeared to be unsuitable for protein derivatization. Since neither SAMBA nor SIA preserved the surface charge, the protein derivatized with these linkers was the mAb which appeared to be less sensitive to surface charge modification (data not shown). Using a derivatization degree of 1-1.7, we obtained only a 1/1 Mgr6-clavin species of conjugate, reflecting the reactivity of the carboxamidophenylthiolate leaving group presented in the linker CDPT. These conjugation procedures led to a highly homogeneous IT population, which is an advantage in accurately addressing the potency of an IT, but the presence of contaminating unconjugated mAb and RIP remains. The toxin can be removed easily by gel permeation, but this procedure fails to separate IT from free mAb. The amount of free mAb is generally too low to interfere with cytotoxicity assays; however, ITs are cleared more rapidly than unconjugated

mAb from the blood circulation. Moreover, the excess of mAb can compete with the IT for binding to the target antigen and can compromise the effectiveness of the IT in vivo. Several methods for the purification of IT from unconjugated antibody have been reported: ion exchange chromatography (40), high-performance liquid immunoaffinity chromatography (41), chromatography on immobilized Cibacron Blue F3GA (42), and, more recently, a nickel-based resin (Ni-NTA) technique (43) for purifying immunoconjugates containing 10 consecutive histidine residues at the amino terminus of the recombinant toxin. To increase the yield of conjugate and to minimize the number of purification steps, we focused on an HPLC ion exchange procedure exploiting the high ionic charge (pI >10) on the small surface of clavin, which allows interaction with weak ionic exchanger groups. In fact, isoelectrofocusing of the ITs showed an increase in the pI from 6.2 for the free Mgr6 to 8.2 pH units (data not shown). Further, an HPLC procedure with a membrane cartridge system allows the use of an increased elution flow rate, maintaining high peak resolution. The semistepwise linear salt gradient removes the low-molecular mass reactives (N-ethylmaleidimide and leaving groups), and Mgr6 eluted in a sharp peak as well as unconjugated clavin. The IT eluted progressively in different peaks, with the first containing small amounts of Mgr6 and the final peaks containing only one pure species, 1/1 mAbclavin IT. As confirmed by competitive binding assay, the conjugation procedure and the nature of the linkage did not significantly modify the mAb binding capacity. Conjugates with disulfide bonds showed almost the same cytotoxic activity toward the target SKBr3 cell line, whereas they were ineffective on the nontarget MeWo cell line. The intracellular release of free clavin is necessary for its activity, and the IT linked by a thioether bond was not cytotoxic at all. These data are in agreement with previous findings that the introduction of a hindered disulfide linkage into ITs has little or no effect on pharmacological potency (23), suggesting that disulfide cleavage is not the rate-limiting step in the toxic activity of clavin conjugates on the cell. Morever, for RIPs with a ribonuclease action, the maintenance of an intracellular easily breakable linkage between mAb and RIP is essential. This behavior is consistent with that described for gelonin and RTA (44) but in contrast with other findings (45) which suggested that gelonin can be delivered efficently by a thioether linkage with the appropriate anti-R-fetoprotein antibody. Comparison of the stability among unhindered and hindered disulfide- and thioether-linked ITs showed that the release of mAb by incubation with GSH is proportional to the presence of methyl groups and increased exponentially with the increase in steric hindrance. The in vivo stability of the ITs was verified by studing the

380 Bioconjugate Chem., Vol. 9, No. 3, 1998

pharmacokinetic behavior after intravenous bolus administration in Balb/c mice. An exponential correlation was also observed between the clearance rate and steric hindrance for the ITs and Mgr6. These findings underline the importance of inserting a highly stable dimethyl instead of a monomethyl or, obviously, an unhindered disulfide into the clavin IT. This was not so important in the case of RTA ITs (24) where a dimethyl group represented no advantage over a monomethyl, with respect to stability. Moreover, we previously reported (23) a linear correlation in the stability of methyl-substituted disulfide-linked gelonin ITs. We also prepared a nonreducible thioether linkage to demonstrate the theoretical limit of in vivo linkage stability that can be obtained using Mgr6-clavin ITs. Even the presence of 1 mol of clavin per mole of mAb determined a decrease in elimination time (from 144 to 46 h) due to modification of the surface charge in mAb linked to a highly basic toxin. The presence of a thioether linkage also affected the distribution phase (R phase) which was doubled compared to that of the three disulfide-linked ITs. In the latter, all three ITs had a similar R phase, suggesting that the reducibility of the linkage is not a limiting factor during the distribution. We have used the most hindered cross-linking system (SAMBA and CDPT) to obtain an IT which maintains its high potency and specificity and has a slow blood clearance similar to that of RTA-containing ITs used in clinical trials (1, 2, 24). Since even a short half-life IT has been reported to be clinically effective in inducing tumor reduction (29, 30), we speculate that an increase in the half-life to 24 h renders the Mgr6-clavin IT more efficient in inducing in vivo tumor damage. ACKNOWLEDGMENT

We thank Elena Luison for technical support and Mario Azzini for figure preparation. This work was supported by grants from the Italian Ministry of University and Scientific and Technological Research and the Italian Association for Cancer Research (AIRC), Founds for Italian Cancer Research, and CNR. E.A. is supported by a fellowship from AIRC. LITERATURE CITED (1) Engert, A., Diehl, V., Schnell, R., Radszuhn, A., Hatwig, M. T., Drillich, S., Schon, G., Bohlen, H., Tesch, H., Hansmann, M. L., Schindler, J., Ghetie, V., Uhr, J., and Vitetta, E. (1997) A phase-I study of an anti-CD5 ricin A-chain immunotoxin (RFT5-SMPT-dgA) in patients with refractory Hodgkin’s lymphoma. Blood 89, 403-410. (2) Winkler, U., Barth, S., Schnell, R., Diehl, V., and Engert, A. (1997) The emerging role of immunotoxins in leukemia and lymphoma. Ann. Oncol. 8 (Suppl. 1), 139-146. (3) Lynch, T. J., Lambert, J. M., Coral, F., Shefner, J., Wen, P., Blattler, W. A., Collison, A. R., Ariniello, P. D., Braman, G., Esseline, D., Elias, A., Skarin, A., and Ritz J. (1997) Immunotoxin therapy of small-cell lung cancer: a phase I study of N901-blocked ricin. J. Clin. Oncol. 15, 723-734. (4) Frankel, A. E., Fitzgerald, D., Siegall, C., and Press O. W. (1996) Advances in immunotoxin biology and therapy: a summary of the Fourth International Symposium on Immunotoxins. Cancer Res. 56, 926-932. (5) Pastan, I., Willingham, M. C., and Fitzgerald, D. J. F. (1986) Immunotoxins. Cell 47, 641-644. (6) Falini, B., Bolognesi, A., Flenghi, L., Tazzari, P. L., Broe, M. K., Stein, H., Durkop, H., Aversa, F., Cornell, P., Pizzolo, G., Barbabietola, G., Sabattini, E., Pileri, S., Matelli, M. F.,

Dosio et al. and Stirpe, F. (1992) Response of refractory Hodgkin’s disease to monoclonal anti-CD30 immunotoxin. Lancet 339, 11951196. (7) Blakey, D. C., and Thorpe, P. E. (1988) An overwiev of therapy with immunotoxins containing ricin or its A chain. Antibody, Immunoconjugates, Radiopharm. 1, 1-15. (8) Dosio, F., Brusa, P., Delprino, L., Grosa, G., Ceruti, M., and Cattel, L. (1994) A new approach in the synthesis of immunotoxins: ribosome inactivating protein noncovalently bound to monoclonal antibody. J. Pharm. Sci. 83, 206-211. (9) Conde, F. P., Orlandi, R., Canevari, S., Mezzanzanica, D., Ripamonti, M., Munoz, S. M., Jorge, P., and Colnaghi, M. I. (1989) The Aspergillus toxin restriction is a suitable cytotoxic agent for generation of immunoconjugates with monoclonal antibodies directed against human carcinoma cells. Eur. J. Biochem. 178, 795-802. (10) Orlandi, R., Canevari, S., Conde, F. P., Leoni, F., Mezzanzanica, D., Ripamonti, M., and Colnaghi, M. I. (1988) Immunoconjugate generation between the ribosome inactivating protein restrictocin and an anti-human breast carcinoma mAb. Cancer Immunol. Immunother. 26, 114-120. (11) Rathore, D., and Batra, J. K. (1997) Cytotoxic activity of ribonucleolytic toxin restrictocin-based chimeric toxins targeted to epidermal growth factor receptor. FEBS Lett. 407, 275-279. (12) Endo, Y., and Wool, I. (1982) The site of action of R-sarcin on eukaryotic ribosomes: the sequence at the R-sarcin cleavage site in 28S ribosomal ribonucleic acid. J. Biol. Chem. 257, 9054-9060. (13) Parente, D., Raucci, P., Celano, B., Pacilli, A., Zanoni, L., Canevari, S., Adobati, E., Colnaghi, M. I., Dosio, F., Arpicco, S., Cattel, L., Mele, A., and De Santis, R. (1996) Clavin, a type I ribosome-inactivating protein from Aspergillus Clavatus IFO 8605. cDNA isolation, heterologous expression and recombinant protein characterization. Eur. J. Biochem. 239, 272-280. (14) Centis, F., Tagliabue, E., Uppungumduri, S., Pellegrini, R., Martignone, S., Mastroianni, A., Menard, S., and Colnaghi, M. I. (1992) p185 Her2/neup epitope mapping with murine monoclonal antibodies. Hybridoma 11, 267-276. (15) Di Fiore, P. P., Segatto, O., Taylor, W. G., Aaronson, S. A., and Pierce, J. H. (1990) EGF receptor and erbB-2 tyrosine kinase domains confer cell specificity for mitogenic signaling. Science 248, 79-83. (16) Felip, E., Del Campo, J. M., Rubio, D., Vidal, M. T., Colomer, R., and Bermejo B. (1995) Overexpression of cerbB-2 in epithelial ovarian cancer. Prognostic value and relationship with response to chemotherapy. Cancer 75, 2147-2152. (17) Wels, W., Harwerth, I. M., Mueller, M., Groner, B., and Hynes, N. E. (1992) Selective inhibition of tumor cell growth by a recombinant single-chain antibody-toxin specific for the erbB-2 receptor. Cancer Res. 52, 6310-6317. (18) Batra, J. K., Kasprzyk, P. G., Bird, R. E., Pastan, I., and King, C. R. (1992) Recombinant anti-erbB2 immunotoxins containing Pseudomonas exotoxin. Proc. Natl. Acad. Sci. U.S.A. 89, 5867-5871. (19) Di Lazzaro, C., Digiesi, G., Tecce, R., Lotti, L. V., Torrisi, M. R., and Natali, P. G. (1994) Immunotoxins to the HER-2 oncogene product: functional and ultrastructural analysis of their cytotoxic activity. Cancer Immunol. Immunother. 39, 318-324. (20) Masuho, Y., Kishida, K., Saito, M., Umemoto, N., and Hara, T. (1982) Importance of the antigen-binding valency and the nature of the cross-linking bond in ricin A-chain conjugates with antibody. J. Biochem. 91, 1583-1591. (21) Worrel, N. R., Cumber, A. J., Parnell, G. D., Ross, W. C., and Forrester, J. A. (1986) Fate of an antibody-ricin A chain conjugate administered to normal rats. Biochem. Pharmacol. 35, 417-423. (22) Thorpe, P. E., Blakey, D. C., Brown, A. N. F., Knowles, P. P., Knyba, R. E., Wallace, P. M., Watson, G. J., and Wawrzynczak, E. J. (1987) Comparison of two anti-Thy 1.1-abrin A-chain immunotoxins prepared with different cross-linking agents: antitumor effects, in vivo fate, and tumor cell mutants. J. Natl. Cancer Inst. 79, 1101-1111.

Properties of Mgr6−Clavin Immunotoxins (23) Arpicco, S., Dosio, F., Brusa, P., Crosasso, P., and Cattel, L. (1997) New coupling reagents for the preparation of disulfide cross-linked conjugates with increased stability. Bioconjugate Chem. 8, 327-337. (24) Carrol, S. F., Bernhard, S. L., Goff, D. A., Bauer, R. J., Leach, W., and Kung, A. H. C. (1994) Enhanced stability in vitro and in vivo of immunoconjugate prepared with 5-methyl2-iminothiolane. Bioconjugate Chem. 5, 248-256. (25) Higgins, W., and Miles, E. W. (1978) Affinity labeling of the pyridoxal phosphate binding site of the beta-2-subunit of Escherichia coli tryptophan synthase. J. Biol. Chem. 253, 4648-4652. (26) Duncan, R. J. S., Weston, P. D., and Wrigglesworth, R. (1983) A new reagent which may be used to introduce sulfhydyl groups into proteins, and its use in the preparation of conjugates for immunoassay. Anal. Biochem. 132, 68-73. (27) Grossbard, M. L., Gribben, J. G., Freedman, A. S., Lambert, J. M., Kinsella, J., Rabinowe, S. N., Eliseo, L., Taylor, J. A., Blattler, W. A., Epstein, C. L., and Nadler, L. M. (1993) Adjuvant immunotoxin therapy with anti-B4-blocked ricin after autologous bone marrow transplantation for patients with B-cell non-Hodgkin’s lymphoma. Blood 81, 2263-2271. (28) Grossbard, M. L., Lambert, J. M., Goldmacher, V. S., Spector, N. L., Kinsella, J., Eliseo, L., Coral, F., Taylor, J. A., Blattler, W. A., Epstein, C. L., and Nadler, L. M. (1993) Anti-B4-blocked ricin: a phase II trial of 7-day continuous infusion in patients with B-cell neoplasms. J. Clin. Oncol. 11, 726-737. (29) Pai, H. L., Wittes, R., Setser, A., Willingham, M. C., and Pastan I. (1996) Treatment of advanced solid tumors with immunotoxin LMB-1: an antibody linked to Pseudomonas exotoxin. Nat. Med. 2, 350-354. (30) Francisco, J. A., Schreiber, G. J., Comereski, C. R., Mezza, L. E., Warner, G. L., Davidson, T. J., Ledbetter, J. A., and Siegall, C. B. (1997) In vivo efficacy and toxicity of a singlechain immunotoxin targeted to CD40. Blood 89, 4493-4500. (31) Brusa, P., Dosio, F., Pietribiasi, P., Delprino, L., Feraiorni, M., Mariani, M., Bussolati, G., and Cattel, L. (1992) Antitumour activity of a sterically blocked ricin immunotoxin on a human colorectal adenocarcinoma grafted subcutaneously in nude mice. Cancer Immunol. Immunother. 35, 373-380. (32) Lambert, J. M., Goldmacher, V. S., Collinson, A. R., Nadler, L. M., and Blattler, W. A. (1991) An immunotoxin prepared with blocked ricin: a natural plant toxin adapted for therapeutical use. Cancer Res. 51, 6236-6242. (33) Dougall, W. C., and Greene, M. I. (1994) Biological studies and potential therapeutic applications of monoclonal antibodies and small molecules reactive with the neu/c-erbB-2 protein. Cell Biophys. 24-25, 209-218. (34) Tecce, R., Digiesi, G., Savarese, A., Trizio, D., and Natali, P. G. (1993) Characterization of cytotoxic activity of saporin anti-gp185/HER-2 immunotoxins. Int. J. Cancer 55, 122-127.

Bioconjugate Chem., Vol. 9, No. 3, 1998 381 (35) Wels, W., Beerli, R., Hellmann, P., Schmidt, M., Marte, B. M., Kornilova, E. S., Hekele, A., Mendelsohn, J., Groner, B., and Hynes, N. E. (1995) EGF receptor and p185erbB-2specific single-chain antibody toxins differ in their cell-killing activity on tumor cells expressing both receptor proteins. Int. J. Cancer 60, 137-144. (36) Strocchi, P., Barbieri, L., and Stirpe, F. (1992) Immunological properties of ribosome-inactivating proteins and a saporin immunotoxin. J. Immunol. Methods 155, 57-63. (37) Rybak, S. M., and Youle, R. J. (1991) Clinical use of immunotoxins. Monoclonal antibodies conjugated to protein toxins. Immunol. Allergy Clin. North Am. 11, 359-365. (38) Benigni, F., Canevari, S., Gadina, M., Adobati, E., Ferrerei, A. J. M., Ferraris di Celle, E., Comolli, R., and Colnaghi, M. I. (1995) Preclinical evaluation of the ribosome inactivating proteins PAP-1, PAP-S and RTA in mice. Int. J. Immunopharmacol. 17, 829-839. (39) Stirpe, F., Derenzini, M., Barbieri, L., Fararbegoli, F., Brown, A. N. F., Knowles, P. P., and Thorpe, P. E. (1987) Hepatotoxicity of immunotoxins made by saporin, a ribosome inactivating protein from Saponaria officinalis. Virchows Arch. 53, 259-271. (40) Lambert, J. M., Senter, P. D., Yau-Young, A., Blattler, W. A., and Goldmacher, V. S. (1985) Purified immunotoxins that are reactive with human lymphoid cells. J. Biol. Chem. 260, 12035-12043. (41) Cumber, A. J., Henry, V. H., Parnell, G. D., and Wawrzynczak, E. J. (1990) Purification of immunotoxins containing the ribosome-inactivating protein gelonin and momordin using high performance liquid immunoaffinity chromatography compared with Blue Sepharose CL-6B affinity chromatography. J. Immunol. Methods 135, 15-24. (42) Knowles, P. P., and Thorpe, P. E. (1987) Purification of immunotoxins containing ricin A-chain and abrin A-chain using blue sepharose CL-6B. Anal. Biochem. 160, 440-446. (43) Dhanabal, M., Fryxell, D. K., and Ramakrishnan, S. (1995) A novel method to purify immunotoxins from free antibodies using modified recombinant toxins. J. Immunol. Methods 182, 165-175. (44) Wawrzynczak, E. J., and Thorpe, P. E. (1988) Effect of chemical linkage upon the stability and cytotoxic activity of A chain immunotoxins. In Immunotoxins (A. E. Frankel, Ed.) pp 239-251, Kluwer, Boston. (45) Masuda, K., Takahashi, K., Hirano, K., and Takagishi, Y. (1994) Selective antitumor effect of thioether-linked immunotoxins composed of gelonin and monoclonal antibody to alpha-fetoprotein or its F(ab′)2 fragment. Tumor Biol. 15, 175-183.

BC970192W