Abstract
We constructed a single-chain anti-gp240 antibody (designated MEL sFv) and fused this to the recombinant toxin gelonin (rGel). MEL sFv-rGel was produced in bacterial expression plasmid (pET-32), and the protein composition was confirmed by both DNA sequencing and Western analysis. Inhibition of cell-free protein synthesis by the fusion construct demonstrated an IC50 of 100 pm, comparable with that for native gelonin (104 pm). The MEL sFv-rGel fusion toxin bound to antigen-positive but not antigen-negative cells as assessed by ELISA. Internalization into A-375 target cells was demonstrable by 1 h after exposure. Against A-375 cells, MEL sFv-rGel demonstrated an IC50 of approximately 8 nm, which was 250-fold lower than that for free rGel (2000 nm). The cytotoxic effects of the construct did not involve apoptosis because terminal deoxynucleotidyl transferase-mediated nick end labeling assays of treated cells were negative. 125I-labeled MEL sFv-rGel demonstrated biphasic clearance of the construct from plasma (t1/2 α and t1/2 β were 0.46 and 7.2 h, respectively). At 72 h after administration, xenograft studies showed that the tissue:blood ratio was highest for tumor followed by spleen, kidney, and liver. Groups of tumor-bearing nude mice were treated with fusion toxin at either 2 or 20 mg/kg. Compared with saline-treated controls, for which mean tumor burden increased 6-fold, the groups treated with the high and low doses of fusion construct showed no increase or only a 2-fold increase, respectively. These studies suggest that this recombinant fusion construct has potent cytotoxic activity both in vitro and in vivo and is an excellent candidate for clinical development.
INTRODUCTION
MAb 3 -based procedures have traditionally been used in the diagnosis and therapy of human cancers. However, clinical use of these agents has met with limited success due to drawbacks associated with this approach, e.g., heterogeneity of antigen expression, poor tumor penetration into solid tumors due in part to antibody size, and antigenicity of the antibodies (1, 2, 3, 4, 5) . To circumvent these problems, a number of molecular approaches have been applied to reconfigure the conventional antibody structure into mouse:human chimeras, completely human antibodies, or reshaped antibody fragments containing the antigen-binding portions of the original structure in a smaller and simpler (single-chain) format (6, 7, 8, 9, 10) . Single-chain antibodies (sFv) retaining the binding characteristics of the parent immunoglobulin (IgG) consist of the antibody VL and VH domains linked by a designed flexible peptide linker (11 , 12) . Furthermore, sFvs may be preferred in clinical and diagnostic applications currently involving conventional MAbs or Fab fragments thereof, because their smaller size may allow better penetration of tumor tissue, improved pharmacokinetics, and a reduction in the immunogenicity observed with i.v. administered murine antibodies.
Among the few target antigens that are expressed at high levels in melanoma cells compared with normal tissue is the surface domain of a high molecular weight glycoprotein (gp240) found on a majority of melanoma cell lines and fresh tumor samples (13) . Two murine antibodies (designated 9.2.27 and ZME-018) recognizing different epitopes on this antigen have previously been isolated and described (14 , 15) . The murine MAb ZME-018 possesses high specificity for melanoma and is minimally reactive with a variety of normal tissues, making it a promising candidate for further study. Clinical trials examining the ability of this antibody to localize within melanoma lesions have demonstrated selective concentration in metastatic tumors (16 , 17) .
Successful development of tumor-targeted therapeutic agents is dependent, in part, on the site-specific delivery of therapeutic agents and on the biological activity of the delivered agent. MAbs have been used to impart selectivity to otherwise indiscriminately cytotoxic agents such as toxins, radionuclides, and growth factors (18, 19, 20) . One such molecule is gelonin, a 29 kDa ribosome-inactivating plant toxin with a potency and mechanism of action similar to RTA but with improved stability and reduced toxicity (21 , 22) . Previous studies in our laboratory have identified and examined the biological properties of numerous chemical conjugates of the plant toxin gelonin and various antibodies (23, 24, 25) . In previous studies, antibody ZME-018 was chemically coupled to purified gelonin, and this immunoconjugate demonstrated specific cytotoxicity against antigen-positive melanoma cells both in tissue culture and in human tumor xenograft models (26 , 27) . However, because intact murine MAbs have inherent problems of antigenicity in human patients, we have developed a molecular approach designed to overcome some of these limitations. The present study describes our investigation of a recombinant immunotoxin consisting of a single-chain analogue of the antibody ZME-018 (cloned VH and VL domains) genetically fused to a rGel molecule.
MATERIALS AND METHODS
The cDNA encoding antibody ZME-018 was amplified from hybridoma RNA obtained from hybridoma cells expressing the murine antibody using kits from Novagen (Madison, WI) and Invitrogen Corp. (Carlsbad, CA). The PCR reagents were obtained from Fisher Scientific (Pittsburgh, PA), and the molecular biology enzymes were purchased from either Boehringer Mannheim (Indianapolis, IN) or New England Biolabs (Beverly, MA). Bacterial strains and pEt bacterial expression plasmids were obtained from Novagen, and growth media were purchased from Difco Laboratories (Detroit, MI). The in situ Cell Death Detection Kit and AP and Fast Red tablets were purchased from Roche Molecular Biochemicals (Indianapolis, IN).
All other chemicals and reagents were from either Fisher Scientific or Sigma Chemical Co. (St. Louis, MO). Metal affinity resin (Talon) was obtained from Clontech Laboratories (Palo Alto, CA). Other chromatography resins and materials were from Pharmacia Biotech (Piscataway, NJ). Tissue culture reagents were from Life Technologies, Inc. (Gaithersburg, MD). All DNA sequencing was performed at the M. D. Anderson Cancer Center Core Facility.
Cloning of the VH and VL Domains of Antibody ZME-018.
mRNA from murine hybridoma FMT 112 P2 expressing antibody ZME-018 (IgG2A) was isolated using the Invitrogen Fast Track kit and transcribed to cDNA with the Invitrogen Copy Kit using the specified conditions. Amplification of antibody light- and heavy-chain variable regions was carried out using the Novagen Ig-Prime kit with the mouse Ig-primer set. The PCR profile for light-chain amplification was as follows: 30 cycles of 94°C × 1 min, 60°C × 1 min, and 72°C × 1 min terminated by a 5-min incubation at 72°C. For heavy-chain reactions, identical conditions were used, except that the annealing temperature was 50°C instead of 60°C. DNA amplified using this procedure was then cloned into the Invitrogen T/A cloning vector pCR II without further purification, transformed into Escherichia coli XL1-Blue, and identified using blue-white screening procedures. Positive clones (five each from the heavy- and light-chain libraries) were sequenced using the T-7 and SP6 promoter primers, and antibody domains were identified by homology to other immunoglobulin sequences.
Construction of Genes Encoding the Single-chain Antibody MEL sFv and the Immunotoxin MEL sFv-rGel.
A two-step splice-overlap extension PCR method (28) was used to construct the single-chain antibody MEL sFv using light- and heavy-chain DNA clones as templates. Light-chain sequences were amplified using primers A (5′-GCTGCCCAACCAGCCATGGCGGACATTGTGATG-3′) and C (5′-GCCGGAGCCTGGCTTGC(A/C)GCTGCCGCTGGTGGAGCCTTTGATC(A/T)CCAG-3′), whereas heavy-chain DNA was amplified with the primers B (5′-AAGCCAGGCTCCGGCGAAGGCAGCACCAAAGG CGAAGTGAAGGTT-3′) and D (5′-GCCACCGCCACCACTAGTTGAGGAGACTGT-3′). The PCR profiles for each set of reactions were as follows: 30 cycles of 1-min denaturation at 94°C, 1-min annealing at 50°C, and a 1-min extension at 72°C, followed by a final 5-min incubation at 72°C. A portion (0.01 ml) of each of these reactions was combined and used directly in a second PCR with only primers A and D following the same reaction profile as before. The final product was purified using Geneclean II (Bio 101, Vista, CA) and then was fused together using the splice-overlap extension PCR method with gelonin DNA as templates and primers NbsphZME (5′-GGCGGTGGCTCCGTCATGACGGACATTGTGATGACCCAGTCTCAAAAATTC-3′), Primer 3 (5′-CCGGAGCCACCGCCACCGCTAGCTGAGGAGAC(T/G)GTGA-3′), NTCOM (5′-GGTGGCGGTGGCTCCGGTCTAGATACCGTTAGC-3′) COMBAC (5′-AAGGCTCGTGTCGACAAGCTTTCATTATTCCGGGTCTTTCTCGAG-3′; Fig. 1 ⇓ ). Purified PCR products were then purified with digested BSPH1 and HindIII as described previously and cloned into vector pET-32a. Sequenced DNA clones were subsequently transformed into E. coli strain AD494(DE3) pLys S obtained from Novagen for expression of the fusion toxin.
Complete DNA sequence analysis of the MEL sFv-rGel fusion construct.
Protein Expression in E. coli.
To express the immunotoxin, bacterial cultures were incubated at 37°C in 2× YT growth medium with strong antibiotic selection (200 μg/ml ampicillin, 70 μg/ml chloramphenicol, and 15 μg/ml kanamycin) and grown until early log phase (A600 nm = 0.4–0.8). The cultures were then diluted 1:1 with fresh 2× YT medium containing the same concentrations of antibiotics, and target protein expression was induced at 23°C by the addition of 0.1 mm IPTG for 16–23 h. Induced bacterial cultures were then centrifuged and stored frozen at −80°C for later purification.
Immunotoxin Protein Purification.
Frozen bacterial pellets from induced cultures expressing immunotoxin anti-melanoma MEL sFv-rGel (MEL sFv-rGel) were thawed at room temperature and lysed by the addition of 1 mg/ml lysozyme in 10 mm Tris-HCl (pH 8.0) for 30 min at 4°C. The bacterial lysates were then sonicated three times for 10 s each with a cell disruptor and centrifuged at 14,000 rpm for 30 min at 4°C. The supernatant was transferred and saved on ice, and the sonication procedure was repeated with the cell pellet. Supernatants from the two lysates were then combined and ultracentrifuged at 40,000 rpm in a Ti-45 rotor for 45 min at 4°C. The samples containing only soluble protein were then filtered (0.22-μm pores), adjusted to 40 mm Tris-HCl with 1 m Tris-HCl (pH 8.0), and then loaded at room temperature onto a Talon metal affinity column pre-equilibrated with the same buffer. After loading, the column was washed with 3 column volumes of loading buffer, followed by a 5-column volume wash with 40 mm Tris-HCl (pH 8.0), 500 mm NaCl, and 5 mm imidazole. Bound protein was then eluted with 5 column volumes of buffer containing 40 mm Tris-HCl (pH 8.0), 500 mm NaCl, and 100–200 mm imidazole. Fractions containing immunotoxin were combined, quantitated, and dialyzed into 20 mm Tris-HCl (pH 7.4) and 150 mm NaCl before digestion with enterokinase to remove the 6× His tag using the procedure established by Novagen.
ELISA and Western Analyses.
All ELISA incubation steps were at room temperature for 1 h, unless otherwise specified, and between incubations all wells were washed with ELISA wash buffer [10 mm Tris-HCl (pH 8.0), 150 mm NaCl, and 0.2% Tween 20]. Wells of a 96-well microtiter plate were each coated with 50,000 gp240-antigen-positive A-375M melanoma cells and dried. These were then rehydrated and blocked with 3% BSA in wash buffer. Plates were incubated, and the purified immunotoxin samples, rabbit anti-gelonin polyclonal antibody [at 100 ng/ml in dilution buffer (ELISA wash buffer containing BSA at a concentration of 10 mg/ml)], and peroxidase-conjugated goat antirabbit IgG (Sigma; used at a 1:5,000 dilution in dilution buffer) were added. Individual wells were thoroughly washed with wash buffer and then developed for 30 min with 2,2′-azino-bis[3-ethylbenz-thiazoline-6-sulfonic acid] in 0.1 m citrate buffer (pH 4.2), and the signal was measured at 405 nm.
For Western blots, all incubations were performed at room temperature for 1 h, unless otherwise specified. Briefly, proteins were separated by SDS-PAGE and transferred onto nitrocellulose overnight at 4°C in transfer buffer [25 mm Tris-HCl (pH 7.5), 190 mm glycine, and 20% (v/v) high-performance liquid chromatography-grade methanol] at 40 V. The filters were blocked with 5% BSA in Western wash buffer [TBS + 0.5% Tween 20] and then reacted successively with rabbit anti-gelonin polyclonal antibody [at a concentration of 100 ng/ml in Western wash buffer, TBS (pH 7.6), and 0.5% Tween 20] and peroxidase-conjugated goat antirabbit IgG (Sigma; at a dilution of 1:10,000 in wash buffer). The signal was developed using the Amersham enhanced chemiluminescence detection system.
Reticulocyte Lysate in Vitro Translation Assay.
The gelonin-induced inhibition of [3H]leucine incorporation into protein in a cell-free protein synthesizing system after the administration of various doses of immunotoxin was carried out as specified by the manufacturer (Promega) and as described previously (29) .
Internalization and Immunofluorescence Staining.
Antigen-positive (A-375 melanoma) cells were added to polylysine-coated 16-well chamber slides (Nunc) at 104 cells/chamber and incubated at 37°C overnight under 5% CO2 atmosphere. Cells were treated with 50 μg/ml MEL sFv-rGel fusion construct at various times. Cells were then washed briefly with PBS, and then proteins bound to the cell surface were stripped by 10-min incubation with glycine buffer [500 mm NaCl and 0.1 m glycine (pH 2.5)], neutralized for 5 min with 0.5 m Tris (pH 7.4), washed briefly with PBS, and then fixed in 3.7% formaldehyde (Sigma) for 15 min at room temperature, followed by a brief rinse with PBS. Cells were then permeabilized for 10 min in PBS containing 0.2% Triton X-100, washed three times with PBS, and incubated with PBS containing 3% BSA for 1 h at room temperature. After a brief wash with PBS, cells were incubated with rabbit anti-rGel polyclonal antibodies diluted 1: 500 in PBS containing 0.1% Tween 20 and 0.2% BSA for 1 h at room temperature. Cells were washed three times in PBS containing 0.1% Tween 20 for 10 min and blocked for 1 h at room temperature with PBS containing 3% BSA, followed by a 1:100 dilution of FITC-coupled antirabbit IgG (Sigma) containing 2.5/μg ml of propidium iodide (PI). Control cells were incubated only with the secondary FITC-coupled antirabbit IgG (1:100) plus 2.5 μg/ml of PI. After three final washes with PBS containing 0.1% Tween 20, cells were washed once in PBS for 10 min and mounted in DABCO mounting medium containing 1 μg/ml of PI. Slides were then analyzed with a Nikon Eclipse TS-100 fluorescence microscope and a Zeiss LSM510 confocal laser scanning microscope. Each photograph was representative of at least 10 fields for each experiment at ×400 magnification.
In Vitro Cytotoxicity Assay.
Samples were assayed using a standard 72-h cell proliferation assay with log-phase (5000 cells/well) antigen-positive A-375M and antigen-negative Me-180 or SK-OV-3 cell monolayers and using crystal violet staining procedures as described previously (30) .
TUNEL Assay.
Log-phase A-375M cells were plated into 16-well chamber slides (10,000 cells/well) and incubated overnight at 37°C under 5% CO2 atmosphere. Cells were treated with the fusion protein MEL sFv-rGel or rGel at a final concentration of 87 nm for different time periods (24 and 48 h) and washed briefly with PBS. Cells were fixed with 3.7% formaldehyde at room temperature for 20 min, rinsed with PBS, then permeabilized with 0.1% Triton X-100 and 0.1% sodium citrate on ice for 2 min and washed twice with PBS. Cells were incubated with TUNEL reaction mixture at 37°C for 60 min, followed by incubation with Concerter-AP at 37°C for 30 min, and finally reacted with Fast Red substrate solution at room temperature for 10 min. After a final wash step, the slides were mounted in mounting medium and analyzed under a light microscope. Positive controls were induced in each experimental set up. Fixed and permeabilized cells were incubated with 1 mg/ml DNase I for 10 min at 37°C to induce DNA strand breaks.
Radiolabeling MEL sFv-rGel.
One drawback in the use of 125I- or 131I-labeled proteins in vivo is the potential for rapid and extensive dehalogenation. A novel procedure for radioiodination has been described and used for MAbs that incorporates iodine into protein via a metabolically stable linkage (31) . The details of this method have been published previously (32) .
In Vivo Cytotoxicity Studies.
Athymic (nude) mice (4–6 weeks old) were divided into groups of 5 mice/cage. Log-phase A-375 human melanoma cells (5 × 106 cells/mouse) were injected s.c. in the right flank, and tumors were allowed to establish. Once tumors were measurable (∼30–50 mm2), animals were treated (i.v. via tail vein) with either saline (control) or various concentrations of the MEL sFv-rGel fusion toxin for 4 consecutive days. Animals were monitored, and tumors were measured for an additional 30 days.
RESULTS
Design of MEL sFv-rGel Fusion Protein.
The variable region genes for the ZME-018 antibody and the gelonin gene (25) were the templates for the construction of the antimelanoma immunotoxin gene. As a first step, we assembled the immunotoxin in one orientation and assessed its binding and cytotoxicity to antigen-positive A-375M melanoma cells. The genes encoding the antibody and gelonin fragments were linked together using a PCR-based method to construct a fusion in the antibody-gelonin orientations. The immunotoxin gene was also COOH-terminal-tagged with a hexahistidine sequence and expressed in E. coli AD494(DE3) pLysS using the Novagen T-7-based expression vector pET-32a.
The single-chain antibody was constructed to encode the VL at the NH2 terminus of the protein followed by an 18-amino acid flexible peptide linker [218 linker sequence, GSTSGSGKPGSGEGSTKG (33 , 34)] and the VH COOH terminus. rGel was positioned downstream of the VH following another linker [G4S (glycine-glycine-glycine-glycine-serine), Ref. 35] . We chose this configuration for reasons involving the unhindered flexibility of the antibody-binding site. With the toxin at the NH2 terminus of the fusion protein, a longer peptide would have been required to provide for optimal spatial orientation of the two protein moieties, and construction of this variant is in progress. DNA sequencing studies of the final fusion gene confirmed the sequence of the final product and that no errors had been introduced using this PCR method. In addition, sequencing (Fig. 1) ⇓ also confirmed that the target gene was inserted into the correct reading frame in the pET-32a vector.
Expression and Purification of Fusion Protein.
The plasmid vector pET-32a containing the fusion gene was transformed into E. coli AD494(DE3) pLysS, and the target protein was induced by the addition of IPTG. As shown in the Coomassie Blue-stained gel in Fig. 2 ⇓ , a protein of the expected molecular mass (76 kDa) was induced. This protein was purified using IMAC resin, and the eluate was exposed to recombinant enterokinase (EK) to yield the final native fusion construct migrating as one band at 56 kDa. The fusion construct was also examined by Western blot using both an anti-gelonin antibody and an anti-single-chain antibody. As shown in Fig. 2 ⇓ , the MEL sFv-rGel fusion construct migrating at 56 kDa reacted with both antibodies, thus demonstrating the presence of immunoreactive antibody and toxin components in the fusion construct. Estimated yields of soluble MEL sFv-rGel immunotoxin from the induced bacterial cultures were approximately 700 μg/liter; however, the yield of final, purified fusion toxin was approximately 200 μg/liter. The primary reason for the reduced yield was found to be an inability of the IMAC to completely capture all of the available soluble target protein. Changes made to the binding buffers and conditions as well as changing brands of IMAC capture resin did not improve these results. Our data suggested that the failure of the IMAC resin to efficiently capture the target protein may have been due to folding of the molecule so that the hexahistadine tag was unavailable for interaction with the resin. Alternatively, it is also possible that posttranslational modification of the target protein could have removed or inactivated the His tag.
Coomassie and Western analysis of the MEL sFv-rGel construct. As shown in the Coomassie Blue-stained SDS-PAGE (left panel), addition of IPTG (Induced) resulted in the appearance of a new band migrating at 76 kDa. This protein was purified using immobilized metal affinity resin (IMAC), and the material was exposed to recombinant enterokinase (EK Cleaved) to remove purification tags to yield the final product migrating at the expected molecular mass (56 kDa). Western analysis using either anti-gelonin (middle panel) or anti-MEL sFv-rGel (right panel) antibodies demonstrated that the MEL sFv-rGel fusion construct was recognized by both antibodies.
ELISA Binding of Immunotoxins.
To ensure that the purified fusion protein retained antigen binding ability, the binding of this material was compared with the binding of intact IgG ZME-018-gelonin chemical conjugate in an ELISA-based binding assay (Fig. 3) ⇓ using intact antigen-positive human melanoma cells as the antigen source. The MEL sFv-rGel fusion construct was found to retain antigen binding comparable with that of the chemical conjugate. The fusion protein also demonstrated specific and significant ELISA binding activity to target A-375M melanoma cells with background levels of binding to SK-OV-3 or ME-180 cells (data not shown).
Comparative binding of the parental ZME-rGel chemical conjugate and MEL sFv-rGel fusion construct. Binding to A-375 cells was assessed using ELISA and a polyclonal rabbit anti-gelonin antibody. The binding of both constructs to target cells was similar, although slightly higher binding was observed for the recombinant fusion construct.
Cell-free Protein Synthesis Inhibitory Activity of the MEL sFv-rGel Fusion Toxin.
The biological activity of toxins can be severely compromised when incorporated into fusion constructs. To examine the n-glycosidic activity of the rGel component of the fusion construct, this material was added to an in vitro protein translation assay using [3H]leucine incorporation by isolated rabbit reticulocytes. Inhibition curves for the fusion construct and native rGel were compared, and the IC50 values for the two molecules were found to be virtually identical (100 versus 104 pm, respectively; data not shown).
Binding and Internalization of MEL sFv-rGel by Immunofluorescence.
Immunofluorescence staining followed by confocal imaging was performed on A-375M and SKBR3 (human breast carcinoma) cells treated with MEL sFv-rGel. (Fig. 4) ⇓ . Internalized rGel toxin was detected using rabbit anti-rGel antibody followed by FITC-coupled antirabbit IgG. The rGel moiety of MEL sFv-rGel fusion protein was observed primarily in cytosol after treatment, and the amount of rGel in cytosol increased over time but was maximal by 1 h after exposure to the fusion construct, thus demonstrating that the fusion construct is capable of efficient cell binding and rapid internalization and delivery of the rGel toxin to the cytoplasm after exposure of log-phase cells.
Internalization of MEL sFv-rGel. A375-M and SKBR3 cells were treated for various times with the MEL sFv-rGel fusion construct. Surface-bound material was removed by an acid wash, and the cells were permeabilized and stained using a rabbit polyclonal antibody against the RGEL component of the construct. Cytoplasmic localization of the construct was detected within 1 h of exposure.
In Vitro Cytotoxic Activity of Immunotoxins.
The MEL sFv-rGel purified fusion protein and the original ZME-rGel chemical construct were tested for specific cytotoxicity against an antigen-positive (A-375M) and an antigen-negative (SK-OV-3) cell line. As shown in Fig. 5 ⇓ , both the chemically produced construct and the fusion construct both demonstrated IC50 values of approximately 8 nm. In contrast, IC50 values for the rGel toxin were approximately 250-fold higher (approximately 2000 nm). The cytotoxic effects of the immunotoxins against antigen-negative SK-OV-3 cells was similar to that of the gelonin alone (data not shown).
Comparative in vitro cytotoxicity of the parental ZME-rGel chemical conjugate and MEL sFv-rGel fusion construct on antigen-positive A-375 human melanoma cells. Cells were plated and then treated for 72 h with various doses of MEL sFv-rGel fusion construct, ZME-rGel chemical conjugate, or free rGel. IC50 values for both immunoconjugates were approximately 8 nm, whereas the IC50 for the rGel was several orders of magnitude higher at approximately 2 × 103 nm. ▪, ZME-rGel; ▴, MEL sFv-rGel; ▾, rGel.
Coadministration of free ZME IgG antibody with the MEL sFv-rGel immunotoxin (Fig. 6A) ⇓ showed a shift in the dose-response curve so that it is essentially coincident with the nontargeted rGel control, demonstrating a dependence of surface antigen recognition for the development of cellular toxicity of the fusion construct. As expected from the binding data, the fusion construct appeared to be several hundred-fold more cytotoxic to antigen-positive A-375M cells than to antigen-negative SKBR3 cells (Fig. 6B) ⇓ .
A, competitive inhibition of MEL sFv-rGel immunotoxin with ZME antibody. Various concentrations of the recombinant immunotoxin were added to A-375 human melanoma cells in log-phase culture in quadruplicate. To another set of wells, a fixed concentration of antibody ZME (50 μg/ml) was admixed with various doses of MEL sFv-rGel immunotoxin and incubated for 72 h. Addition of free ZME antibody resulted in approximately a 3-fold reduction in immunotoxin cytotoxicity. B, relative cytotoxicity of MEL sFv-rGel against antigen-positive (A-375) and antigen-negative (SKBR3) cells. Various concentrations of the recombinant immunotoxin were added to log-phase cells in quadruplicate. As shown, the IC50 values for the immunotoxin were 8 and 600 nm for A-375 and SKBR3 cells, respectively.
TUNEL Assay for Apoptosis.
The cytotoxic effects of the MEL sFv-rGel were assessed by TUNEL assay to determine whether these effects are mediated by DNA fragmentation or apoptosis. As shown in Fig. 7 ⇓ , treatment of cells for 24 or 48 h with an IC50 dose of MEL sFv-rGel or an equivalent molar dose of rGel showed no effect on DNA fragmentation compared with the DNase I-positive control. This suggests that the observed cytotoxic effect of this fusion construct does not appear to be mediated through an apoptotic mechanism.
Assessment of apoptosis in cells treated with the MEL sFv-rGel fusion construct. A-375 cells were treated with an IC50 concentration of the MEL sFv-rGel construct and then stained for DNA fragmentation at 24 and 48 h after treatment. As shown, neither the MEL sFv-rGel construct nor the rGel toxin itself affected DNA fragmentation.
Pharmacokinetics and Tissue Disposition of 125I-labeled MEL sFv-rGel.
The MEL sFv-rGel fusion construct was labeled with the N-succinimidyl 4-iodobenzoate reagent, which results in a labeled protein resistant to metabolic dehalogenation (31) . This material was administered i.v. to BALB/c mice, and at various times after administration, groups of three mice were sacrificed, and blood was obtained. As shown in Fig. 8 ⇓ , clearance of TCA-precipitable counts in plasma followed a biphasic profile. Pharmacokinetic analysis demonstrated α and β phase half-lives of 0.46 and 7.2 h, respectively. Plasma contained approximately 10% acid-soluble counts at all times examined, suggesting some metabolic degradation of the protein structure.
Pharmacokinetics of 125I-labeled MEL sFv-rGel in mice. The radiolabeled fusion construct was administered i.v. to mice, and 3 mice/group were sacrificed at various times after administration. The radioactivity in plasma both as TCA-precipitable and free counts was assessed, and the pharmacokinetics of the acid-precipitable counts was assessed (pk Analyst; Micro-Math, Inc.).
Radiolabeled MEL sFv-rGel was also administered to nude mice bearing human melanoma (A-375) xenografts, and groups of three mice were sacrificed at various times after administration. Samples of tissues including blood were excised, weighed, and counted. Twenty-four h after administration, kidney contained the highest tissue:blood ratio (Table 1 ⇓ and Fig. 9 ⇓ ). Levels in kidney remained constant over time. Concentration of the label in tumor increased over time, and tumor was the highest site of accumulation by 72 h after administration, followed by spleen, kidney, and liver in descending order.
Tissue distribution of 125I-labeled MEL sFv-rGel in nude mice bearing A-375 tumors. Nude mice bearing human melanoma tumors received injection of 125I-labeled MEL sFv-rGel. Three mice were sacrificed at 24, 48, and 72 h after administration. Tissue and blood samples were obtained, weighed, and counted. The tissue:blood ratio in tumors increased over time, and at 72 h, tumor contained the highest concentration compared with normal organs.
Tissue distribution of 131I-labeled MEL sFv-rGel in mice bearing A-375 tumors
Antitumor Activity of MEL sFv-rGel in Xenograft Models.
Mice bearing well-developed A-375 melanoma xenografts were treated with either saline (controls) or MEL sFv-rGel at either 2 or 20 mg/kg for 4 days. As shown in Fig. 10 ⇓ , tumor size in the control group increased from 30 to 220 mm2 (∼700% increase) over the 40-day length of the experiment. In contrast, mice treated with the fusion toxin at 2 mg/kg showed a slight decrease in tumor size followed by an increase to approximately 60 mm2 (100% increase). Mice treated with the 20 mg/kg dose of fusion toxin demonstrated a 50% decrease in tumor size during treatment, followed by a slow recovery of tumor size back to the original tumor size over 40 days (no increase in overall growth). There were no obvious toxic effects of the immunotoxin on mice at these doses, suggesting that the MTD at this schedule had not been reached.
Nude mice bearing well-developed melanoma tumors (A-375) growing in the right flank were treated i.v. with either saline (controls) or MEL sFv-rGel at 2 or 20 mg/kg (total dose) for 4 consecutive days (arrows). Tumor areas were measured for 40 days. The saline-treated control tumors increased from 30 to 220 mm2 over this period. Tumors treated with the lowest immunotoxin dose increased from 30 to 60 mm2. Animals treated with the highest immunotoxin dose showed no overall increase in tumor size from the original 30 mm2.
DISCUSSION
The development of recombinant, single-chain antibodies has markedly affected the field of targeted therapeutics by allowing numerous molecular design approaches to engineer molecules with improved performance characteristics compared with those of the original intact murine or human complete immunoglobulins (36 , 37) . In addition to reducing the size (and thereby improving the penetration into solid tumors), recombinant techniques have also allowed investigators to incorporate other proteins such as toxins or cytokines directly into the antibody structure, thus improving the stability of the molecule compared with chemical coupling approaches and greatly simplifying the manufacturing process (38) .
This initial report of a recombinant, single-chain antimelanoma immunotoxin is unique in that there are few reported recombinant antibody constructs targeting melanoma. This is surprising because the highly metastatic spread of this disease appears to be an attractive target for immunotoxin therapy. Whereas there have been numerous reports of both murine and human antibodies targeting surface antigen domains that appear to be highly associated with melanoma (39, 40, 41) , the gp240 antigen has been consistently demonstrated to be useful as a discriminator of malignant melanoma cells as compared with normal cells (42) . Antibodies targeting the gp240 antigen such as MEL sFv and ZME are of particular interest because of their rapid and efficient internalization into melanoma target cells, making these antibodies excellent targeting carriers of therapeutic agents.
The protein synthesis-inhibitory activity in cell-free systems of the recombinant fusion toxin compared with that of free rGel suggests that there does not appear to be significant steric crowding of the gelonin active-site cleft due to proximity of the antibody fragment in our designed molecule. Also, because there are no protein cleavage sites within this fusion construct, our data also suggest that gelonin does not necessarily require cleavage from the construct to maintain biological activity. This is in sharp contrast to studies with RTA, which requires release from the protein carrier to recover biological activity (43 , 44) . This is surprising because gelonin and RTA share identical mechanisms of action (21) and also share approximately 30% sequence homology (45) . Thus, rGel joins a small group of proteinaceous toxins that appear to be compatible for inclusion as fusion constructs.
The advantages of this particular toxin over other protein toxins such as RTA, Pseudomonas exotoxin, and diphtheria toxin are that immunotoxins constructed with rGel have plasma clearance kinetics and tumor localization rates almost identical to those of the original antibody. On the other hand, conjugates produced with RTA are rapidly cleared from the circulation and localized within the reticuloendothelial system in the liver and spleen (46 , 47) .
This study also provided important information regarding antibody valency related to immunotoxin function. The recombinant MEL sFv-rGel construct contains 1 binding domain/toxin molecule, whereas the bivalent chemical conjugate ZME-rGel contains 2 binding domains/toxin. Despite this difference and the potential for reduced avidity of the recombinant construct compared with the natural antibody, the molar IC50 values for the two agents were virtually identical (approximately 10 nm). This may suggest that the potential for reduced avidity of the fusion construct could be offset by the potential ability of the construct to deliver more toxin molecules per cellular binding event.
Previous studies of single-chain antibody-Pseudomonas exotoxin fusion constructs have suggested that tumor localization indices (relative to blood) are low when compared with tumor localization of full-length antibody immunoconjugates (48 , 49) . In addition, when compared with full-length immunotoxins, single-chain immunotoxins have a reduced tumor residence time, probably due to their reduced affinity for the target antigen. Our tissue disposition and pharmacokinetic studies suggest that the localization of the construct occurs efficiently and that, by 72 h, concentrations of the fusion construct are highest in tumor tissues. Furthermore, in keeping with other studies of other single-chain antibodies (50) , the tumor:blood ratio of the construct appears to be prolonged and increased over time, due, in part, to the relatively rapid clearance of the agent from the circulation. Tissue:blood ratios of the fusion construct appear to be similar to that found with chemical conjugates of the original intact IgG ZME-018 and rGel (45) . It is important to note that regardless of the absolute delivery of drug to tumor, the most important parameter is the therapeutic efficacy of the agent.
The short half-life observed with this construct suggests that dosing intervals of 24 or 48 h appear to be optimal to achieve maximal concentration of the agent in tumor tissue. Of critical importance to the eventual clinical development of MEL sFv-rGel will be examination of the MTD toxicity profile and efficacy studies of this fusion construct in other well-characterized human melanoma xenograft models at the MTD dose.
Because effective clinical management of solid tumors such as melanoma generally requires prolonged treatment regimens, the immunogenicity of this molecule is of concern. The immunogenicity of antibody-toxin constructs containing full-length antibodies is due, at least in part, to the large size of the antibody itself, the antigenicity of the toxin component, and the long circulation time of the constructs. The single-chain immunotoxin, by comparison, should be significantly less immunogenic because of its small size and more rapid clearance kinetics from the circulation.
Acknowledgments
All sequences were obtained at the DNA Sequencing Facility at the The University of Texas M. D. Anderson Cancer Center supported by Core Grant (CA 16672). We express our sincere appreciation to Julia Merchant for assistance in the preparation of the manuscript. We also acknowledge the Core Grant CA 16672 for support of the Veterinary Medicine Facility at M. D. Anderson Cancer Center.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 Research was conducted, in part, by the Clayton Foundation for Research.
↵2 To whom requests for reprints should be addressed, at The University of Texas M. D. Anderson Cancer Center, Section of Immunopharmacology and Targeted Therapy, Department of Bioimmunotherapy, 1515 Holcombe Boulevard, Unit 44, Houston, TX 77030. Phone: (713) 792-3554; Fax: (713) 794-4261/(713) 745-3916; E-mail: mrosenbl{at}notes.mdacc.tmc.edu
↵3 The abbreviations used are: MAb, monoclonal antibody; VL, variable domain, light chain; VH, variable domain, heavy chain; IPTG, isopropyl-Β-d-galactopyranoside; IMAC, immobilized metal affinity chromatography; TUNEL, terminal deoxynucleotidyl transferase-mediated nick end labeling; RTA, ricin A-chain; MTD, maximal tolerated dose; rGel, recombinant gelonin; TCA, trichloroacetic acid.
- Received July 8, 2002.
- Accepted May 15, 2003.
- ©2003 American Association for Cancer Research.
References
Volume 63, Issue 14
