Bioconjugate Chem. 1992, 3, 302-307
302
Cytotoxicity of Chimeric (Human-Murine) Monoclonal Antibody BR96 IgG, F(ab’)z, and Fab’ Conjugated to Pseudomonas Exotoxin Clay B. Siegall,’ Susan L. Gawlak, Jeffrey J. Chin, Mary E. Zoeckler,+ Kathleen F. Kadow,t Joseph P. Brown, and Gary R. Braslawskyt Cellular and Molecular Biology Department and Immunology Department, Bristol-Myers Squibb Pharmaceutical Research Institute, 5 Research Parkway, Wallingford, Connecticut 06492. Received February 6, 1992 We have made antigen-specific cytotoxic reagents by conjugating the chimeric antibody BR96 (chiBR96) to Pseudomonas exotoxin A (PE), as either native P E or a truncated form (LysPE40) devoid of the cell-recognition region (domain I). P E kills cells by ADP-ribosylation of elongation factor 2, thereby inhibiting protein synthesis. Chimeric BR96 immunotoxins were constructed by chemical conjugation of the toxin to Fab‘, F(ab‘)z, and intact IgG and purified by anion-exchange and gel-filtration chromatography. Chimeric BR96 [IgG and F(ab’)zl immunotoxins were cytotoxic against tumor cell lines displaying the BR96 antigen, with ECw values ranging from 0.1 to 110 pM. Immunotoxins constructed with chiBR96 Fab’ were 50-100-fold less cytotoxic. Competition analysis showed that the immunotoxins were specifically active through their BR96 antigen-binding ability. The binding of chiBR96-PE and chiBR96-LysPE40 to antigen was equivalent to that of BR96 itself and these immunotoxins were found to internalize very rapidly, displaying 90% of their cytotoxicity within 1 h. Binding assays determined that chiBR96 F(ab’)z-LysPE40 bound as well as chiBR96-LysPE40; however, chiBR96 Fab’-LysPE40 bound 20-fold less efficiently. The chiBR96 Fab’-LysPE40 internalized similarly to the F(ab’)2 or the IgG immunotoxins. Therefore, the chiBR96 Fab’-LysPE40 immunotoxin is less cytotoxic toward target cells because of reduced antigen binding. This is may be due to the monovalent nature of chiBR96 Fab’-LysPE40. This study shows that the monoclonal antibody chiBR96-Pseudomonas exotoxin A immunotoxins can be effective at inhibiting protein synthesis in target cells.
INTRODUCTION
Immunotoxins have been investigated as a new approach for treating tumors and disease in man (1-3). By coupling a potent but nonselective bacterial or plant protein toxin with an antibody directed against tumor-associated membrane antigens, cytotoxicity of the toxin can be preferentially directed toward neoplastic cells. Immunotoxins have shown promise in preclinical models using human tumor xenografts in nude mice (4-6). The studies outlined here focus on immunotoxins prepared by conjugating Pseudomonas exotoxin A (PE) and a 40-kDa (LysPE40) form of the toxin conjugated to the chimeric (human-murine) antibody BR96. BR96 binds to Lewis Y antigen present on a large number of breast, colon, lung, and ovarian carcinomas (7). The antibody has low crossreactivityto normal tissue, limited to the upper gastrointestinal tract. More importantly, the unconjugated antibody is rapidly internalized after binding antigen, a requirement that has been demonstrated for antibody-directed cytotoxicity in several studies of antibody-toxin conjugates (8) and antibody-drug conjugates (9, 10). Chimeric BR96 was constructed by a two-step homologous recombination method, in which DNA encoding human IgGl heavy chain and human K light chain were attached to the murine BR96 variable region (11). In this study, we have tested immunotoxins using the whole antibody molecule as well as antibody fragments [F(ab’)2 and Fab’] as a first step toward designing optimal recombinant molecules. PE, produced by Pseudomonas aeruginosa, is a polypeptide comprising three domains (12). Domain I encodes the cell-binding ability, domain I1 encodes the proteolytic sensitivity site and the membrane translocation ability, and domain I11 encodes the
* To whom correspondence should be addressed. +
Immunology Department.
ADP-ribosylation activity of the toxin (13,14). P E kills cells by ADP-ribosylating elongation factor 2 and catalytically inhibiting protein synthesis. By removing domain I from PE, a truncated 40-kDa toxin is formed (PE40) (15). PE40 is weakly toxic to cells,since it lacks the binding domain for the PE receptor (15). In order to conjugate this molecule to an antibody, the amino terminus of PE40 has been modified to include a lysine residue, hence LysPE4O (6). In this report, P E and LysPE40 were coupled to the various antibody forms either by thiolation with 2-iminothiolane or by direct attachment to intact antibody by reduction with DTT. Monovalent Fab’ conjugates showed antibody-directed cytotoxicity, although at much higher protein concentrations than those obtained using bivalent IgG and F(ab’)2 toxins. Differences in cytotoxic potential were attributed to binding differences between the various forms of the antibody, and not to internalization differences. EXPERIMENTAL PROCEDURES
Reagents. N-succinimidyl4-(N-maleimidomethyl)cyclohexane-l-carboxylate (SMCC) and 24minothiolane were purchased from the Pierce Chemical Corp. (Rockford, IL). Soluble pepsin was purchased from Sigma Chemical Co. (St. Louis, MO). Na1251and PHIleucine were purchased from New England Nuclear (Boston, MA). Native PE was purchased from Berna Products (Coral Gables, FL). Mono Q columns were purchased from Pharmacia (Uppsala, Sweden). TSK-3000columns were purchased from TosoHaas, Inc. (Philadelphia, PA). Immunoblots were performed using mouse (anti-idiotypic BR96) and rabbit (anti-PE) ABC kits (Vector Laboratories, Burlingame, CA). Rabbit polyclonalanti-PE antibody and mouse 1) were kindly supanti-PE monoclonal antibody (M40/ plied by Drs. Ira Pastan and David FitzGerald, National Institutes of Health (Bethesda, MD). Anti-idiotypic BR96 0 1992 Amerlcan Chemical Society
Chlmerlc Immunotoxlns
antibody (757-4-1)was provided by Dr. Bruce Mixan, Bristol-Myers Squibb (Seattle, WA). Cell Culture and Plasmids. All cells were cultered in RPMI 1640 supplemented with 10% fetal bovine serum except L929, which was cultured in DMEM supplemented with 10% fetal bovine serum. Plasmid pMS8, which encodes the gene for LysPE4O under control of the the T7 promoter, was previously described (6). Expression and Purification of LysPE40. The plasmid pMS8 encoding LysPE40 (6) was transformed into Escherichia coli BL21(XDE3) cells and cultured in Super Broth (Digene, Inc., Silver Spring, MD) containing 75 pg of ampicillin/mL at 37 "C. When absorbance a t 650 nm was 2.0 or greater, isopropyl l-thio-8-D-galactopyanoside was added (1mM), and cells were harvested 90 min later. The bacteria were washed in sucrose buffer (20 % sucrose, 30 mM Tris-HC1(pH 7.4), 1mM EDTA) and osmotically shocked in ice-cold H20 to isolate the periplasm. LysPE40 protein was purified from the periplasm by successive anion-exchange and gel-filtration chromatography with a Pharmacia fast protein liquid chromatography (FPLC) system, as described previously (6). Generation of BR96 F(ab')z and Fab' Fragments. F(ab')z fragments were generated from chiBR96 (4 mg/ mL) by pepsin digestion (25 pg/mL in 0.1 M citrate buffer, pH 4.0). After 6 h a t 37 "C, digestion was stopped by adjusting the pH to 7.2 with phosphate-buffered saline (PBS). Purity of the digest preparation was 90-95% F(ab')z, determined by SDS-PAGE and Coomassie Blue staining. Fab' was prepared from chiBR96 F(ab')z by reduction with cysteine to break the remaining interchain disulfide bonds (16). Briefly, F(ab')z molecules (2-4 mg/mL) in 0.1 M Tris-HC1 (pH 7.5) were incubated at 37 "C for 2 h with cysteine (0.01 M final concentration). Free sulfhydryl groups on the Fab' molecule were alkylated with 0.02 M iodoacetamide (CalBiochem, San Diego, CA) for 30 min at room temperature to prevent recombination of the Fab' to F(ab')2. Thereaction mixture was dialyzed againstPBS. Purity was greater than 85 % as assessed by SDS-PAGE. Immunotoxin Construction and Purification. Chimeric BR96 (6-10 mg/mL) was thiolated by addition of a 3-fold molar excess of 24minothiolane (2-IT) in 0.2 M sodium phosphate buffer (pH 8.0), 1 mM EDTA for 1 h at 37 "C (4). Unreacted 24minothiolane was removed by PD-10 column chromatography (Pharmacia). Alternatively, free thiol groups were generated by reduction with dithiothreitol (DTT). Chimeric BR96 was incubated with a 20-fold molar excess of DTT for 2.5 h a t 42 "C. Excess DTI'was removed by overnight dialysis against PBS under nitrogen. The number of thiol groups on the antibody was determined by DTNB reduction (Ellman's reagent, Sigma Chemical Co., St. Louis, MO) (17). This procedure routinely gave four thiol groups per BR96 antibody, with no reduction in antibody binding reactivity or protein concentration. The procedure was not used with F(ab')z or Fab' fragments. Thiolated BR96 antibody was condensed with maleimide-modified P E or LysPE4O. A maleimide group was attached to lysine residues on the toxin (PE or LysPElO; 6-8mg/mL) by mixing with a 3-fold molar excess of SMCC in 0.2 M sodium phosphate (pH 7.01, 1 mM EDTA a t room temperature for 30 min and purified on a PD-10 column. Modified toxin and thiolated antibody were mixed in a 4:l molar ratio and incubated a t room temperature for 14-16 h to allow a thioether linkage to form. Immunotoxins were purified by anion exchange (Mono Q)to remove unreacted antibody and gel-filtration chro-
BloconjuQate Chem., Vol. 3,
No. 4, 1992 909
matography (TSK-3000) to remove unconjugated toxin as previously described (4, 6, 15). Inhibition of Protein Synthesis Assay. Tumor cells (lo5cells/mL) in growth media were added to 96-well flat bottom tissue culture plates (0.1 mL/well) and incubated at 37 "C for 16 h. Dilutions of toxin or toxin-conjugates were made in growth media, and 0.1 mL was added to each well (three wells/dilution) for 1 h or 20 h at 37 "C. After the appropriate incubation time unreacted material was removed by washing the monolayer with growthmedia. Cells were incubated in 0.2 mL of growth media for a total of 20 h and pulse labelled with PHIleucine (1pCi/well) for an additional 4 h at 37 "C. The cells were lysed by freezing and thawing a t 37 "C and were harvested using a Tomtec cell harvester (Orange, CT). Cellular protein labeled with PHIleucine was determined by counting the incoporated radioisotope using a LKB Beta Plate liquid scintillation counter. Competition for Cytotoxicity Analysis. Tumor cells were plated and incubated at 37 "C as described above. Chimeric BR96 diluted to 50 pg/mL in growth media was added to the cell monolayer (0.5 mL/well). After incubation a t 37 "C for 1 h, dilutions of chiBR96 immunoconjugates (50pL) were added and incubated an additional 1h and cell supernatants were removed and washed with RPMI. Growth media (0.2 mL) was added to each well, and cells were incubated at 37 "C (20 h) and labeled with L3H1leucine as described above. Binding Activity. Competition Binding. L2987 cells were removed from monolayer culture with 0.2% trypsin and washed with RPMI 1640 containing 2% FCS (wash buffer). Cell suspensions (1.0 X 106 cells/O.l mL) were incubated withO.l mL of fluorescein isothiocyanate (FITC) labeled chiBR96 (13.3 pg/mL final concentration) and 0.1 mL of diluted antibody or immunotoxin 4 "C for 1h and washed, and the amount of cell-bound FITC-labeled chiBR96 was quantified on an EPICS V Model 753 flow cytometer (Coulter Corp., Hialeah, FL). Direct Binding. Two-fold serially diluted immunotoxins or antibody was incubated for 1 h at 4 "C in 0.2 mL of wash buffer containing 1 X lo6L2987 cells. Cells were washed and then incubated in wash buffer containing 1: 40 diluted FITC-labeled goat anti-human K antibody (Bethyl Labs, Montgomery, TX) for an additional 30 min at 4 "C. Cells were washed and analyzed for cell-surface fluorescence on a flow cytometer. Antigenic Modulation and Internalization. The level of membrane-associated immunotoxin bound to target cell populations during antigenic modulation was quantified using an [12511anti-PEmonoclonal antibody M40/1. Previously, the M40/1 epitope was mapped to a 44 amino acid region in P E domain I1 (18). The M40/1 antibody was radioiodinated using NalZ5I(New England Nuclear, MA) and chloramine T (Kodak Chemical Co., Rochester, NY) according to ref 19). Radioiodinated M40/ 1 was separated from unbound iodine by PDlO column chromatography (Pharmacia). Specific activities ranged from 2 to 5 X lo5 cpm/pg. Modulation of intact chiBR96, F(ab')a, or Fab' immunotoxins was assayed on L2987 cells propagated as 9095% confluent monolayer cultures in 96 well microtiter plates prepared as described above. Target cells were pulsed for 1h a t 4 "C with 0.1 mL of 2-fold serially diluted immunotoxin ranging from 1.6 to 3.2 X M antibody protein in binding buffer. Subsequently, cultures were washed using growth medium and individual plates were incubated in complete medium at either 4 or 37 "C. A t specific time points during incubation at 37 or 4 "C,
304 kDa 200-
BiOConjUgate Chem., Vol. 3, No. 4, 1992 1
97-
2
3
4
5
6
kDa7
Siegall et al. 8
9
10
u.
50
0
97-
40
60
6843-
24
30
45
20
29
10
10
18-
Figure 1. Nonreducing SDS-PAGE analysis of conjugated and unconjugated chiBR96 IgG, Fab’, F(ab’)2. 4-2096 gels were stained with Coomassie Blue. Lanes: 1,chiBR96IgG; 2, chiBR96 Fab’; 3, chiBR96 Fab’-LysPE40; 4, native PE; 5, LysPE40; 6, chiBR 96 IgGLysPE40; 7, chiBR96 (Fab’)2; 8, chiBR96 F(ab’)zLysPE40; 9, chiBR96 IgGLysPE40; 10, chiBR96 IgG.
100
10
1000
10000
COMPETITOR, (umoles, antibody protein)
Figure 2. Competition of chiBR96-PE and chiBR96-LysPE40 binding. Inhibition of FITC labeled chiBR96 (40 pg/mL) by chiBR96 (a);chiBR96-PE (W), and chiBR96-LysPE40 (A)using L2987 target cells. 0 W L . , ”
triplicate sets of wells were washed twice with wash buffer and pulse-labeled with 0.1 mL ‘“1-labeled M40/ 1antibody (0.5 pg/mL in wash buffer containing 0.2 % sodium azide). After 15min, unbound label was removed from the monolayers, and cell-bound cpm determined by solubilization of the cell monolayer with 0.5 N NaOH. Cell-bound radioactivity was determined using a LKB Model 1272 y counter. Nonspecific binding was determined by incubation of target cells with a similer concentration of unconjugated chiBR96. In certain experiments, unconjugated P E was used to determine background binding levels. 1251-labeledM40/1 antibody did not react with membrane bound antibody or PE. RESULTS
Construction and Purification of chiBR96-Toxin Conjugates. Reduced chiBR96-drug conjugates have been successfully used as (in vivo) antitumor agents. In this study, we have used a similar approach in constructing chiBR96-toxin conjugates. Chimeric BR96, thiolated by condensation with 2-iminothiolane or after mild reduction using DTT, was chemically conjugated to both P E (66 kDa) and LysPE40 (40 kDa). LysPE40 was additionally conjugated to chiBR96 Fab’ and F(ab’)2 antibody forms. The immunotoxins were purified as described in Experimental Procedures. Chimeric BR96 IgGLysPE4O (190 kDa), Fab’-LysPE40 (96 kDa), and F(ab’)zLysPE40 (145 kDa) conjugates were analyzed by nonreducing SDS-polyacrylamide gel electrophoresis to determine the size of the native conjugate (Figure 1). From the Coomassie Blue stained gels, we determined that there was less than 5 % unconjugated antibody after purification. Binding Activity of chiBR96-Immunotoxins. Two methods were used to determine whether there was an alteration in antibody binding activity after conjugation to PE or LysPE40. Competition binding analysis showed that both immunotoxins competed with FITC-labeled chiBR96 as efficiently as unconjugated chiBR96 antibody (Figure 2), indicating that binding affinity for the BR96 antigen was not perturbed after chemical conjugation. Similar results were obtained using the direct binding assay for both P E and LysPE4O conjugates (results not shown). Binding activities of LysPE40-conjugated and unconjugated IgG, F(ab’)2, and Fab’ were also compared by direct binding to L2987 tumor cells. Cell-boundantibody protein was quantitated using FITC-labeled goat anti-human K light chain antibody. Binding of the LysPE40 immunotoxin was similar to that obtained using unconjugated chiBR96 antibody (Figure 3A) and agrees with results
180
-
160140
-
120100
-
2
1 W
807 .1
140
B 1
100 10
1 0 0 1 0 010
.1
1
10
1 0 0 1000
.1
1
10
100 1000
ANTIBODY CONCENTRATION, (uglml)
Figure 3. Direct binding of intact chiBR96-LysPE40, F(ab’)PLysPE40,and Fab’-LysPE40 to L2987 cells. Cells were incubated (4 “C, 1 h) with immunotoxin or antibody as described in Experimental Procedures. Cell-boundantibody was quantitated with FITC labeled anti-human K antibody. The amount of immunotoxin or antibody remaining on the cell surface was determined by measuring cell-suface fluorescence. Legend: ChiBR96 (a), chiBR96-LysPE40 (A), chiBR96 F(ab’)g ( O ) , chiBR96F(ab’)z-LysPE40 (m), chiBR96 Fab’ ( O ) ,chiBR96Fab’LysPE40 (A).
obtained using the competition binding assay (Figure 2). Figure 3B compares the binding activity of intact IgG to that of F(ab’)2 and F(ab’)2-LysPE40. There was no loss in immunoreactivity with the F(ab’)2 and F(ab’)z-immunotoxin as compared to chiBR96IgG. Conjugationof PE40 to Fab’ did not affect immunoreactivity (Figure 3C); however, binding of the Fab’ was significantly decreased as compared to intact IgG (Figure 3C), most likely due to the monovalency of the Fab’ molecule. Modulation and Internalization of chiBR96-Immunotoxins. BR96 has been shown to be a rapidly internalizing antibody (7). The ability of chiBR96-PE and chiBR96-LysPE40 to induce antigenic modulation was initially measured by determining the loss of immunotoxin from the cell surface membrane (Figure 4). There was no difference in modulation kinetics between PE or LysPE40 immunotoxins with approximately 50 % of the original cell-bound immunotoxin modulated from the surface membrane 30 min after warming to 37 “C. Cells incubated under conditions which do not allow antigenic modulation (4 “C) showed essentially no loss of cell surface toxin within 6 h (Figure 4a). In order to confirm that the loss of cell-surface immunotoxin was due to endocytosis, cells were incubated with a 1251-labeledimmunotoxin complex for 1 h at 4 OC to permit binding, washed, and subsequently modulated at 37 “C. As demonstrated in Figure 4B, essentially all the radiolabeled immunotoxin remained cell-associated, despite the concomitant loss of immunotoxin from the cell-surfacemembrane (Figure 4A). These findings confirm that most if not all of the membrane-associated BR96 immunotoxins were rapidly
Chimeric Immunotoxlns
A.
Bioconlugate Chem., Vol. 3, No. 4, 1992 305
B.
and L2987 cells are the most sensitive cell lines tested. The I g G and F(ab’)~-LysPE40molecules showed similar cytotoxic activity against MCF-7 cells (ECm = 8-14 pM) while the, Fab’-LysPE40 conjugate was much less active l 60 : p (ECm = 780 pm) (Figure5 ) . Specificityof protein synthesis inhibition activity of Fab’ and F(ab’)z conjugates was also preserved, with little or no inhibitory activity observed against the antigen negative cell lines A2780. 40 20 Specificity of Growth Inhibition by chiBR96-LysPE40. Specificity was confirmed by abrogating the protein 0 - 0 0 1 2 3 4 5 synthesis inhibition by chiBR96-LysPE40 with uncon0 1 2 3 4 5 jugated chiBR96. Addition of excess chiBR96 antibody TIME AT INDICATED TEMP., (hr) (50pg) with chiBR96-LysPE40 immunotoxin, resulted in Figure 4. Endocytosis of cell-surface immunotoxin after moda decrease of in vitro potency (Figure 6). At 20 pM of ulation with chiBR96-PE or chiBR96-LysPE40 immunotoxins. chiBR96-LysPE40, approximately 50 % of its cytotoxic (A) Loss of cell surface immunotoxin was measured by incubating the cell populations under modulating (37 “C) or nonmodulating effect was blocked by the addition of excess unconjugated (4 “C)conditions. The amount of immunotoxin remaining on antibody, while at 4 pM, the excess chiBR96 completely the cell surface at each individual time point was quantitated blocked the cytotoxic activity of chiBR96-LysPE40. with [1sI]M-40/1 (anti-PE)antibody. (B)Internalizationof cellKinetics of Cytotoxicity of chiBR96 Immunotoxbound immunotoxin was measured by incubating the immunoins and Native PE. In part, the effectiveness of immutoxin plus [l26I]M-40/1 complex at 37 or 4 O C . Cell-bound notoxins may depend upon the rate of internalization after radioactivity was determined as described in Experimental Procedures. Chimeric BR96/PE coated cells incubated at 4 O C binding to antigen positive cells. To determine the cy(0) or at 37 “C(a);chiBR96-LysPE40 coated cells incubated at totoxic activity of chiBR96-PE, chiBR96-LysPE40, and 4 “C (A) or at 37 “C (A). PE, we performed a time-course analysis where cells were incubated with toxin for up to 20 h as described in Table I. Internalization of Chimeric BR96-Immunotoxins Experimental Procedures. from the Cell Surface Membrane of L2987 Cells After a 1-h incubation, MCF-7 cells were sensitive to 7% internalization chiBR96-PE and chiBR96-LysPE40, with ECm values of 2.5 h 4.5 h 1 and 60 pM, respectively, but not to the native toxin chiBR96-LysPE40 74.0 85.0 (ECm > 10 000 pM). After 20 h MCF-7 cells were slightly chiBR96 F(ab’)2-LysPE40 72.0 91.6 more sensitive to chiBR96-PE and chiBR96-PE40, but chiBR96 Fab’-PE40 12.0 89.6 much more sensitive to PE; ECm = 200 pM (Figure 7). We also repeated this assay a t 2,4, and 6 h time points (data internalized, and that internalization rates were similar not shown). At each time point, P E was considerably less for P E and LysPE4O immunotoxins. cytotoxic against MCF-7 cells than chiBR96-immunotoxThe capacity of chiBR96 F(ab’)~-LysPE40and Fab’ins. This may be due in part to the mechanism by which LyaPE4O immunotoxins to internalize was also determined the toxin molecule is delivered to target cells. by measuring the loss of cell surface toxin using raDISCUSSION diolabeled anti-PE antibody. Essentially all of the chiBR96 immunotoxins were completely internalized after We have produced immunotoxins containing the car4.5 h including Fab’-immunotoxin (Table I). However, cinoma-reactive monoclonal antibody chiBR96 and rate differences were observed. At 2.5 h, when 76 % of the Pseudomonas exotoxin A. The antibody was used informs intact IgG toxin and 72 % of the F(ab’)z were internalized, including intact IgG, F(ab’)z, and Fab’. The toxin comonly 129% of the Fab’-immunotoxin was internalized. ponent of the immunotoxin was either native PE or Therefore, I g G , F(ab’)Z-, and Fab’-LysPE40 immunoNLysPE40 (referred to as LysPE40 in this paper), a toxins were modulated from the cell surface membrane, truncated form containing a genetically modified amino but at different rates. terminus that includes a lysine residue for conjugation In Vitro Cytotoxicity of Intact IgGPE Immunopurposes (4). Chimeric BR96-toxin conjugates were found toxins. Having established the binding and internalizing to be cytotoxic to cells which display Lewis Y, the epitope activities of these chimeric molecules, we assayed their recognized by BR96. The most cytotoxic of the conjugates cytotoxicity against cancer cells by comparing inhibition produced was chiBR96-PE, which was 1000-fold more of protein synthesis on antigen-positive and antigenpotent than PE itself against MCF-7 breast carcinoma negative cell lines (Table 11). BR96 antigen-positive cell cells. Chimeric BR96-LysPE40 was also extremely cylines MCF-7, L2987, and RCA were the most sensitive to totoxic toward BR96 antigen-positive cells (1000-foldmore chiBR96-PE with ECm values of 0.14, 0.28, and 1.4 pM, potent than LysPE4O). Both chiBR96-PE and chiBR96respectively. The immunotoxin was also more inhibitory LysPE40 were produced using two procedures to generate than native PE, which had EC60 values of 200, 140, and sulfhydryl groups on the antibody, mild reduction or 380 pM. When tested on antigen-negative cell lines, little treatment with 2-iminothiolane. The former procedure difference in ECm values between PE and the immunogave a greater yield of conjugate, but conjugates produced toxin was observed. Specificity (antibody directed cellby both procedures resulted in chimeric molecules of killing) must also take into consideration the different identical activities. sensitivities of the various cell lines to native PE. Thus, Chimeric BR96-PE and chiBR96-LysPE40 were almost the chiBR96 immunotoxins were 100-500-fold more potent fully active with 1-h incubation, while P E was relatively than native PE against the antigen positive cell lines. inactive (Figure 7). With continued incubation, chiBR96Cytotoxicity of chiBR96 Antibody and Enzymatic immunotoxins increase cytotoxic activity only slightly Fragments Linked to LysPE40 against MCF-7 Cells. We while P E becomes cytotoxic to the MCF-7 cells at later also compared the cytotoxic activity of chiBR96 in Fab’, time points. This rapid efficacy of chiBR96-immunoF(ab’)z, and IgG form linked by LysPE40 (Table 111).As toxins is evidence of the utility of chiBR96 in targeting with the chiBR96-PE immunotoxin (Table II), MCF-7 cell populations for elimination.
1
I
306
Bloconlugete Chem., Vol. 3, No. 4, 1992
Siegall et al.
Table 11. Cytotoxicity of Chimeric BR96-PE on Human Tumor Cells
BR96 DTT-reduced 2-iminothiolane-treated cell line antigen chiBR96-PE chiBR96-PE native PE breast cancer + 0.10 MCF-7 0.14 200.0 L2987 lung cancer + 0.25 0.28 140.0 + 1.2 RCA colon cancer 1.4 380.0 +23.0 A2780 ovarian cancer 23.0 60.0 13.5 mouse fblst L929 14.0 3.0 220.0 epidermoid cancer KB 231.0 227.0 ECm represents the amount of immunotoxin or toxin required to inhibit 50% of the protein synthesis as determined by [3H]leucine incorporation in cellular protein. Table 111. Cytotoxicity of 2-Iminothiolane-Substituted Chimeric BR96-LysPE40 on Human Tumor Cells
cell line tme BR96 antigen IgGPE40 breast cancer + 8 MCF-7 lung cancer + 37 L2987 RCA colon cancer + 84 ovarian cancer +650 A2780 >5000 epidermoid cancer KB a ECm is described in the Table I footnote. ND = not determined. 2
s 10
1
100
1000
10000
100000
IMMUNOTOXIN, pM
Figure 5. Cytotoxicity of various chiBR96 forms conjugated to LysPE40 against MCF-7 cells. Immunotoxins were incubated for 20 h, and protein synthesis was determined as described in Experimental Procedures. Legend: chiBR96-PE40 (a),chiBR96 F(ab’)2-PE40 (a),chiBR96 Fab’-PE40 (A),PE40 ((
2
w I
k
100
,
1
W chiBR96-LysPE40
+ 50 ug chiBR96
0.8
4
20
pM, chiBR96-PE40
Figure 6. Competition of chiBR96-PE40 cytotoxic activity against MCF-7 cells. Chimeric BR96-PE40 was added at 0.8,4, and 20 pM concentrations to MCF-7 cells in presence or absence of 50 pg (333 pM) chiBR96 antibody. Cytotoxicity was determined as described in Experimental Procedures.
We also examined the binding and internalization activities of chiBR96-immunotoxins. Immunoconjugates prepared with intact IgG or its F(ab’)2 or Fab’ enzymatic digest products were not affected in terms of binding by chemical conjugation to LysPE40 (Figure 3) or P E (data not shown). Differences in binding activity between Fab’ and F(ab’)2 or IgG conjugates may be attributed to
F (ab’)z-PE40 14 70 110 2500 ND
Fab’-PE40 780 2700 5000 11000 ND
LysPE40 15000 17500 15000 15000 >25000
~~~
1 HR
20 HR
IMMUNOTOXIN, pM
Figure 7. Protein synthesis inhibition analysis of chiBR96-
PE/LysPE40 vs P E against MCF-7 cells. Immunotoxins were added to the cells a t various concentrations and either removed and replaced with fresh media at 1h or not disturbed. After 20 h, the cells were pulsed with r3H]leucine and harvested as described. Legend: chiBR96-PE (a),chiBR96-PE40 (a),P E (A).
differences in avidity due to the monovalence of the Fab’ molecule. We also cannot exclude the possibility that the enzymatic method used to generate Fab’ did not contribute to the decreased avidity. Of most interest was the comparisonbetween chiBR96LysPE40 and the enzymatic fragment immunotoxins chiBR96 F(ab’)z-LysPE40 and chiBR96 Fab’-LysPE40. This finding is also reflected in the cytotoxicity data (Table 111). The decreased potency of the Fab’-immunoconjugate could not be attributable to the inability to internalize, although minor rate differences were observed between the Fab’ and F(ab’)2 or IgG toxin conjugates. Other groups have used fragments of antibodies to direct toxins toward certain cell populations (20,21). F(ab’)2 and Fab’ fragments are smaller than IgG molecules, and can perhaps penetrate the inside of tumors more readily (22). Our own results (manuscript in preparation) have shown that antibody Fab’ fragments have a shorter serum lifetime and penetrate into the tumor at a faster rate than IgG or F(ab‘)2 fragments. It will be important to determine the serum lifetime and the maximum tolerated dose for these immunotoxin forms. A bivalent BR96 immunotoxin molecule may be necessary to achieve maximal binding efficacy, but a Fab’-immunotoxin molecule may achieve faster penetration into the tumor mass and allow for a greater input of material despite a shorter serum half-life of the molecule.
Chlmerlc Immunotoxins
Chimeric toxins similarto chiBR96-PE have beenshown to be effective as antitumor agents in vivo (5). In this report, we have shown that both intact PE and LysPE40 immunotoxins as well as F(ab'):!and Fab' immunotoxins demonstrate cytotoxic activity in vitro. Potency differences betweenthe various immunotoxinswere most likely attributable to avidity differences and could not be due to the ability of the immunotoxinsto deliver the toxin to the appropriateintracellulartarget. The criticalquestion is whether the pharmacokinetic differencesbetweena monovalent Fab' and a bivalent IgG or F(ab')a immunotoxin will translate into a better therapeutic response. Experiments which will determine the in vivo potency of chiBR96-PE conjugates have been initiated in models of human tumors. ACKNOWLEDGMENT
We thank Drs.I. Pastan, D. FitzGerald, B. Mixan, P. Fell, K. E. Hellstrom, S. McAndrew, and P. Friedman, for reagents and helpful comments, and L. Howe, for expert secretarial assistance. LITERATURE CITED (1) Pastan, I., and FitzGerald, D. (1991) Recombinant toxins for cancer treatment. Science 254, 1173-1177. (2) FitzGerald, D., and Pastan, I. (1991) Redirecting Pseudomonas exotoxin. Semin. Cell Biol. 2, 31-37. (3) Vitetta, E. S., Krolick, K. A., Miyama-Inaba, M., Cushley, W., and Uhr, J. W. (1987) Immunotoxins: A new approach to cancer therapy. Science 644, 650. (4) Batra, J. K., Jinno, Y., Chandhary, V. K., Kondo, T., Willingham, M. C., FitzGerald, D. J., and Pastan, I. (1989) Antitumor activity in mice of an immunotoxin made with antitransferrin receptor and a recombinant form of Pseudomonas exotoxin. Proc. Natl. Acad. Sci. U S A . 86, 8545-8549. (5) Pai, L. H., Batra, J. K., FitzGerald, D. J., Willingham, M. C., and Pastan, I. (1991) Anti-tumor activities of immunotoxins made of monoclonal antibody B3 and various forms of Pseudomonas exotoxin. R o c . Natl. Acad. Sci. U.S.A.88,335& 3362. (6) Debinski, W., Karlsson, B., Lindholm, L., Siegall, C. B., Willingham, M., FitzGerald, D., and Pastan, I. (1991) Monoclonal antibody C242-Pseudomonas exotoxin A A specific and potent immunotoxin with antitumor activity. J. Clin. Invest. in press. (7) Hellstrom, I., Garrigues, H. J., Garrigues, U.; and Hellstrom K. E. (1990) Highly tumor-reactive, internalizing, mouse monoclonal antibodies to L@-relatedcell-surfaceantigens. Cancer. Res. 50,2183-2190. (8) Carriere, D. Casella, P., Richer, G., Gros, P., and Jansen, F. (1985) Endocytosis of an antibody-ricin A chain conjugate (immuno-Atoxin) adsorbed on colloidal gold. Exp. Cell. Res. 156, 327-340.
Bloconjugate Chem., Vol. 3, No. 4, 1992 307
(9) Nio, Y., Shiraishi, T., Imai, S., Ohgaki, K., and Tobe, T. (1989) Binding and internalization of human immunoglobulin G conjugated with melphalan (K18) to human tumor cell lines. Anticancer Res. 9, 59-64. (10) Braslawsky, G. R., Kadow, K., Knipe, J., McGoff, K., Edson, M., Kaneko, T., and Greenfield, R. S. (1991) Adriamycin (Hydrazone) antibody conjugates require internalization and intracellular acid hydrolysis for antitumor activity. Cancer Immunol. Immunother. 33, 367-374. (11) Fell, H. P., Yarnold, S., Hellstrom, I., and Hellstrom, K. E. (1989)Homologous recombination in hybridoma cells: Heavy chain chimeric antibody produced by gene targeting. Proc. Natl. Acad. Sci. U.S.A. 86, 8507-8511. (12) Allured,V., Collier, R. J,., and McKay, D. B. (1986)Structure of exotoxin A of Pseudomonas aeruginosa at 3.0-angstrom resolution. Proc. Natl. Acad. Sci. U.S.A. 83, 1320-1324. (13) Hwang, J., FitzGerald, D. J. P., Adhya, S., and Pastan, I. (1987)Functional domains ofPseudomonasexotoxin identified by deletion analysis of the gene expressed in E. coli. Cell 48, 129-136. (14) Siegall, C. B., Chaudhary, V. K., FitzGerald, D. J., and Pastan, I. (1989) Functional analysis of domains 11,Ib and I11 of Pseudomonas exotoxin. J.Biol. Chem. 264,14256-14261. (15) Kondo, T., FitzGerald, D., Chaudhary, V. K., Adhya, A., and Pastan, I. (1988) Activity of immunotoxins constructed with modified Pseudomonas exotoxin A lacking the cell recognition. J. Biol. Chem. 263, 9470-9475. (16) Parham, P. (1986) Preparation and purification of active fragments from mouse monoclonal antibodies. In Handbook of Experimental Immunology, (Weir, D. M., Ed.) pp 1-23, Blackwell Scientific Publishers, New York. (17) Deakin, H., Ord, M. G., and Stocken, L. A. (1963) "Glucose 6-phosphate-dehydrogenase" activity and thiol content of thymus nuclei from control and X-irradiated rats. Biochem. J. 89, 296-304. (18) Ogata, M., Pastan, I. and FitzGerald, D. (1991) Analysis of Pseudomonas exotoxinactivation and conformationalchanges by using monoclonal antibodies as probes. Infec.Immun. 59, 407-414. (19) McConahey, P. J., and Dixon, F. J. (1966) A method for trace iodination of proteins for immunologicalstudies. Arch. Allergy Appl. 29, 185-188. (20) Roffler, S. R., Yu, M.-H., Chen, B. M., Tung, E., and Yeh, M-Y. (1991) Therapy of human cervical carcinoma with monoclonal antibody-Pseudomonas exotoxin conjugates. Cancer Res. 51, 4001-4007. (21) 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.) Kluwer Academic Publishers, New York. (22) Jain, R. K. (1989) Delivery of novel therapeutic agenta in tumors: Physiological barriers and strategies. J.Natl. Cancer Inst. 81, 570-576.
