Optimization of Conditions for Formation and Analysis of Anti-CD19

Jan 1, 1997 - Anti-CD19 scFv FVS191cys was constructed by engineering a ... has a CD19 binding ability similar to that of its parental mAb B43 and is ...
0 downloads 0 Views 161KB Size
Download PDF
Bioconjugate Chem. 1997, 8, 64−70

64

Optimization of Conditions for Formation and Analysis of Anti-CD19 FVS191 Single-Chain Fv Homodimer (scFv′)2 Duo Wang,† Erica Berven,† Quanzhi Li,† Fatih Uckun,‡ and John H. Kersey*,†,§,|,⊥ University of Minnesota Cancer Center, Biotherapy Institute, and Departments of Laboratory Medicine/ Pathology and Pediatrics and Therapeutic Radiology, University of Minnesota, Minneapolis, Minnesota 55455. Received August 29, 1996X

In this report, we present the production of a dimeric form of anti-CD19 scFv, the FVS191cys (scFv′)2. Anti-CD19 scFv FVS191cys was constructed by engineering a cysteine residue at the C terminus of the VL domain of scFv FVS191. FVS191cys (scFv′)2 was formed through a disulfide bond between two FVS191cys molecules. To optimize the yield of FVS191cys (scFv′)2, the effects of oxidation time, buffer pH, and temperature on the formation of dimeric scFv were analyzed. Our study indicates that the formation of FVS191cys (scFv′)2 is oxidation time- and buffer pH-dependent; a high pH buffer facilitates the formation of disulfide-linked (scFv′)2. The maximum yield of FVS191cys (scFv′)2 can be achieved when FVS191cys is air-oxidized at 4 °C, in buffer with a pH of 8.5-9. The biological activity of FVS191cys (scFv′)2 was analyzed by ELISA and an internalization assay. FVS191cys (scFv′)2 has a CD19 binding ability similar to that of its parental mAb B43 and is internalized by CD19 positive Nalm 6 cells. This study indicates that FVS191cys (scFv′)2 is a potential candidate for tumor diagnosis or therapy.

INTRODUCTION

(mAbs)1

Monoclonal antibodies have been useful for antigen-specific targeting of tumor cells; however, the application of mAbs has demonstrated limited diffusion of mAbs from the vasculature into the tumor. The limited diffusion of intact antibodies is due to the large size of mAbs and the host effect elements of their Fc domains (1, 2). A recombinant single-chain Fv (scFv) composed of the minimal antigen binding domains, variable heavy chain (VH) and variable light chain (VL), is one-sixth of the size of a mAb (28 kDa). ScFvs have exhibited improved tumor penetration with higher tumor to normal tissue ratios than corresponding IgG or Fab and faster plasma clearance rates (3-5). However, scFvs have only one antigen binding arm which may decrease antigen binding affinity (4). Previous studies have demonstrated that homodimers of scFv [(scFv′)2], such as an anti-c-erb2 (scFv′)2, have exhibited divalent binding and increased retention in tumors as compared with the corresponding scFv monomers (6). Also, radiolabeled (scFv′)2 was found to have improved tumor imaging compared to Fab, monomeric scFv, and intact antibody (7). These studies suggest that dimers of scFvs may be useful candidates for tumor diagnoses and therapy. * To whom requests for reprints should be addressed at Box 86 UMHC, 420 Delaware St. SE, Minneapolis, MN 55455. Telephone: 612-625-4659. Fax: 612-624-8965. E-mail: kerse001@ maroon.tc.umn.edu. † University of Minnesota Cancer Center. ‡ Biotherapy Institute. § J.H.K. is a recipient of an Outstanding Investigator Grant Award (CA 49721) from the National Cancer Institute. | Department of Laboratory Medicine/Pathology. ⊥ Department of Pediatrics and Therapeutic Radiology. X Abstract published in Advance ACS Abstracts, January 1, 1997. 1 Abbreviations: mAb, monoclonal antibody; scFv, singlechain Fv; scFv′, single-chain Fv with C-terminal cysteine; (scFv′)2, dimeric single-chain Fv; BMH, 1,6-bis(maleimido)hexane; MCA, N,N-bismaleimidocaproyl amino acid; SDSPAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

S1043-1802(96)00077-8 CCC: $14.00

Covalently linked (scFv′)2 have been generated by several different approaches: (i) forming a disulfide bond between carboxyl-terminal cysteine residues (6, 8), (ii) cross-linking with chemical linkers (6, 9), and (iii) forming covalent bundle helices or leucine zippers at the C terminus (10). The first method is carried out by introducing one cysteine at the 3′ end of the scFv and forming a C-terminal disulfide bond through air oxidation. Chemical linking uses BMH or peptide-bridged MCA to link two scFvs (6). This method is complicated by the multiple steps involved in the chemical modification. Finally, engineering small peptides at the C terminus of scFv for formation of helix bundles or leucine zippers may be useful, but it may increase the immunogenicity of the molecules. In this report, we studied the formation of dimeric forms of FVS191cys. FVS191 is an anti-CD19 scFv produced in our laboratory from the B43 hybridoma (11). FVS191cys was constructed by engineering a cysteine residue at the C terminus of FVS191. We focused on the production of a disulfide-linked (scFv′)2 because this species was relatively easy to generate and was stable in solution as chemically linked (scFv′)2 (6). It was observed that FVS191cys (scFv′)2 could be formed by air oxidation in basic buffers without any chemical modifications or manipulations. To optimize the conditions for formation of disulfide-linked (scFv′)2, we used FVS191cys as a model molecule to study the effects of oxidation time, buffer pH, and temperature on formation of FVS191cys (scFv′)2. In addition, the CD19 binding ability of FVS191cys (scFv′)2 was evaluated by ELISA. The internalization of FVS191cys (scFv′)2 by CD19 positive cells was also investigated. MATERIALS AND METHODS

Construction of the Plasmid for the Expression of FVS191cys. The expression plasmid of FVS191cys was derived from pFVS191 (11). To insert an additional cysteine residue at the 3′ end of the VL gene, the pFVS191 plasmid was restricted with BglII and EcoRI and ligated with the annealed, complementary oligonucleotides of HLCys1 and HLCys2. The correct insertion of the oligonucleotides HLCys1/HLCys2 was verified by nu© 1997 American Chemical Society

Dimers of Anti-CD19 scFV FVS191

cleotide sequencing. The construction of pFVS191cys is shown in Figure 1. Expression and Purification of FVS191cys. The pFVS191cys plasmid was transformed into Escherichia coli BL21(DE3) (Novagen, Madison, WI). The transformed bacterial cells were grown in 1 L of SOB medium (20 g of tryptone, 5 g of yeast extract, 0.5 g of NaCl, and 5 g of MgSO4‚7H2O per liter) at 37 °C. When the absorbance of A600 of the bacterial culture reached 0.65, recombinant protein production was induced with 1 mM isopropyl β-D-thiogalactopyanoside (IPTG) for 1.5 h at 37 °C. The inclusion bodies of FVS191cys were isolated using the following procedures. The harvested cell pellets were mixed with 50 mL of inclusion body separation (IBS) buffer (0.1 M KCl, 0.02 M Tris-HCl, 5 mM EDTA, and 0.1% Nonidet P-40 at pH 8.0) and sonicated. Lysozyme was added to the sonicated cell mixture to a final concentration of 0.2 mg/mL. After incubation at room temperature for 1 h, the cell lysate was frozen at -80 °C. Cell lysates were then thawed at room temperature, sonicated, and centrifuged at 17000g for 30 min at 4 °C. The supernatant was discarded, and the pellet was suspended in 50 mL of IBS buffer. Sodium deoxycholate (10%, w/v) was added to the suspensions for a final concentration of 2%. The mixture was stirred at room temperature for 1 h and centrifuged at 17000g for 30 min at 4 °C. The inclusion body pellet was washed by suspending in IBS buffer and centrifuging at 17000g for 30 min. The wash was repeated once with water, and the inclusion bodies were stored at -80 °C. Refolding of FVS191cys. FVS191cys was denatured and refolded according to Buchner’s method (12) with some modifications. The inclusion bodies of FVS191cys were dissolved in a denaturing buffer [0.1 M Tris, 6 M guanidine hydrochloride, 0.3 M dithioerythritol (DTE), and 0.002 M EDTA at pH 8] at room temperature for 2 h. Insoluble materials were removed by centrifugation at 30000g for 30 min. The soluble protein concentration was determined using Coomassie Plus Protein Assay Reagent (Pierce) and bovine serum albumin (BSA) as standards. The final concentration of protein in denaturing solution was adjusted with denaturing buffer up to 20 mg/mL. Renaturation of FVS191cys was carried out by a rapid 100-fold dilution of the denatured protein into refolding buffer [100 mM Tris-HCl, 500 mM L-arginine, 8 mM oxidized glutathione (GSSG), and 2 mM EDTA at pH 8] at 10 °C and incubation for 48 h. Purification of Refolded FVS191cys. After renaturation, the FVS191cys was concentrated and dialyzed against phosphate-buffered saline (PBS) buffer using an Amicon RA2000 concentrator with a YM10 cartridge (Amicon, Beverly, MA). FVS191cys proteins were purified by FPLC using a Superdex 75 (16 × 60 mm) gel filtration column (Pharmacia). Purified FVS191cys monomers were concentrated by ultrafiltration through a YM10 membrane (Amicon). Formation of the FVS191cys Dimer. FVS191cys (scFv′)2 was formed by air oxidation of FVS191cys at 4 °C. To study the oxidation time and pH effects on FVS191cys (scFv′)2 formation, 500 µL of 400 µg/mL purified FVS191cys was injected into Slide-A-Lyzer 10K Dialysis Cassettes (Pierce) and dialyzed against Tris/ EDTA/NaCl (TEN) buffer (10 mM Tris-HCl, 2 mM EDTA, and 100 mM NaCl at pH 6.5-12). After dialysis, the proteins were transferred into FALCON 2059 tubes and air-oxidized for 1, 5, 12, 16, 20, and 29 days at 4 °C. To examine the formation of FVS191cys (scFv′)2, 50 µL of the protein sample was run on a 12% nonreducing SDSPAGE (Bio-Rad, Hercules, CA) and visualized with

Bioconjugate Chem., Vol. 8, No. 1, 1997 65

Coomassie Brilliant Blue R-250. The relative amounts of monomers and dimers of FVS191cys were quantified using a GS-700 imaging densitometer with the program of Molecular Analyst (Bio-Rad). The proportion of dimeric FVS191cys (% d) was calculated as % d ) dimer/ (monomer + dimer) × 100%. This quantification method was tested with BSA standards, and the result indicated that this method could be used to compare the amounts of proteins. To study the effect of dithiothreitol (DTT) reduction on the formation of dimeric FVS191cys, FVS191cys was reduced with DTT before air oxidation. Purified FVS191cys in PBS buffer was added by 1/100 volume of 200 mM DTT and incubated at room temperature for 1 h. After reduction, DTT was removed by a PD10 column (Pharmacia). Reduced FVS191cys was dialyzed against TEN buffers with the pH at 7, 7.5, 8, 8.5, 9, and 10. The formation of dimeric FVS191cys was examined by 12% nonreducing SDS-PAGE after air oxidation for 1, 5, 12, 16, 20, and 29 days. The visualization and quantification of monomers and dimers of FVS191cys were performed as described above. The effect of temperature on the formation of dimeric FVS191cys was evaluated as follows. FVS191cys in TEN buffer at pH 8 was incubated at 4 °C, room temperature, and 37 °C. After air oxidation for 4 h or longer, the formation of FVS191cys (scFv′)2 was examined by SDSPAGE. Purification of FVS191cys (scFv′)2. FVS191cys (scFv′)2 was purified by FPLC using a Superdex 75 column (HiLoad 16/60) (Pharmacia). PBS buffer was used as an elution buffer; the column was run at a flow rate of 1 mL/min. The fractions containing FVS191cys (scFv′)2 were collected and concentrated by ultrafiltration using a YM10 membrane (Amicon). 125 I Labeling of FVS191cys. Purified FVS191cys (scFv′)2 was labeled with radioiodine using the IODOBEADS method of Pierce (Pierce, Rockford, IL). FVS191cys (scFv′)2 (100 µg) in 100 µL of PBS was incubated with 100 µL of 125I solution containing 1 mCi sodium iodine (NEN, Boston, MA) in a 1 mL Reacti-Vial. The reaction proceeded for 10 min at room temperature. The free iodine was separated from the labeled dimers with a Sephadex G-25 gel filtration column (Pharmacia). The specific activity of labeled FVS191cys (scFv′)2 was approximately 2 µCi/µg. CD19 Antigen Binding Assay. The specific CD19 binding ability of FVS191cys (scFv′)2 was assessed by 125Ilabeled proteins. CD19 positive (Nalm 6) and CD19 negative (Molt 13 and Peer) cell lines were grown in RPMI 1640 tissue culture medium supplemented with 10% fetal bovine serum (FBS). The log phase cells were harvested by low-speed centrifugation, washed once with fresh medium, and adjusted to a final concentration of 2 × 108 cells/mL. The 125I-labeled FVS191cys (scFv′)2 was diluted in fresh tissue culture medium to a final concentration of 200 ng/mL. Pellets of 0.1 mL of cell suspension from each cell line in duplicate were incubated with 400 µL of labeled antibody solution at 4 °C for 1 h. After incubation, the cells were washed three times with 1 mL of fresh tissue culture medium. The radioactivity associated with pellets of each cell line was determined in a γ counter (Beckman). The CD19 binding ability of FVS191cys (scFv′)2 was further assessed by a competitive ELISA. The ELISA method was performed as follows. Purified B43 mAb was conjugated to alkaline phosphatase (B43-AP) using a commercial conjugation kit (Pierce). The conjugates were used directly without further purification. A series of experiments were carried out to determine the optimal

66 Bioconjugate Chem., Vol. 8, No. 1, 1997

Wang et al.

Figure 2. Expression and purification of FVS191Cys. Electrophoretic analysis of samples on a 12% reducing SDS-PAGE: lane 1, total bacterial lysate of uninduced BL21(DE3) carrying pFVS191cys; lane 2, total cell lysate of IPTG-induced BL21(DE3) carrying pFVS191cys; lane 3, inclusion bodies of FVS191cys; and lane 4, refolded FVS191cys purified by FPLC. The sizes of molecular mass markers are indicated at each side of the gel. Figure 1. Cloning strategy used in the construction of FVS191cys. pFVS191 plasmid was restricted with enzyme BglII and EcoRI and inserted with annealed oligonucleotides HLCys1/ HLCys2.

concentration of coating antigen and B43-AP for the ELISA. The CD19 antigens used in the ELISA were isolated from CD19 positive Daudi cells using the methods of Siegall et al. (13). ELISA plates (Immuron 4, Dynatech Laboratories Inc., Chantily, VA) were coated with 50 µL of isolated membrane proteins at a concentration of 80 µg/mL. The plates were dried in a vacuum desiccator at 4 °C overnight and washed three times with PBS before use. Varying amounts of competing FVS191cys (scFv′)2 and a fixed amount of B43-AP (total volume of 100 µL) were added to the wells and incubated at room temperature for 1 h. The final dilution of B43AP was 1:2000. The plates were washed four times with PBS and incubated with 100 µL of substrate solution at 4 °C overnight. The optical density of the wells was determined by an ELISA reader at 405 nm. Internalization Assay. CD19 positive Nalm 6 cells were grown in RPMI 1640 medium supplemented with 10% FBS. The cells were harvested by centrifugation. 125 I-labeled FVS191cys (scFv′)2 (10 mg) was incubated with 1 × 108 cells in a volume of 5 mL and incubated on ice for 1 h. B43 mAb-blocked Nalm 6 cells were used as a control to assess nonspecific binding. After incubation, the cells were washed four times with cold RPMI 1640 without serum. Aliquots of 5 × 106 cells were transferred to 1 mL test tubes and incubated at 37 °C in a tissue culture incubator for various time intervals. Percentages of cell surface-bound, dissociated, internalized, and degraded FVS191cys (scFv′)2, based on the total cellassociated radioactivity at time 0, were determined according to the methods of Press et al. (14). Supernatants were separated from cell pellets by centrifugation. Precipitation of the supernatants by TCA further defined two portions of the labeled FVS191cys (scFv′)2: degraded FVS191cys (scFv′)2 (TCA nonperceptible) and dissociated FVS191cys (scFv′)2 (TCA perceptible). Cell pellets were treated with acid-papain; the radioactivity released by acid-papain treatment was regarded as cell surface-bound FVS191cys (scFv′)2, and the radioactivity associated with cell pellets after the treatment was seen as the internalized FVS191cys (scFv′)2. RESULTS

FVS191cys Expression and Purification. A cysteine residue was introduced into the anti-CD19 scFv, FVS191, fragment by genetically engineering an additional codon at the 3′ end of the gene encoding the VL domain (Figure 1). The sulfhydryl group introduced to

Figure 3. Formation and separation of FVS191cys (scFv′)2. Protein samples were analyzed by electrophoresis: lane 1, airoxidized FVS191cys on a 12% nonreducing SDS-PAGE; lane 2, air-oxidized FVS191cys on a 12% reducing SDS-PAGE; lane 3, FVS191cys (scFv′)2 (peak II of FPLC) on a 12% nonreducing SDS-PAGE; and lane 4, FVS191cys monomers (peak III of FPLC) on a 12% nonreducing SDS-PAGE. The sizes of molecular mass markers are indicated on the left side of gels.

the 3′ end of FVS191 served as a specific site for the formation of disulfide-bonded dimeric FVS191cys. FVS191cys was produced as insoluble inclusion bodies by pFVS191cys-transformed BL21(DE3) E. coli cells. Figure 2 shows the expression and purification of FVS191cys on a SDS-PAGE. An average of 40 mg of FVS191cys inclusion bodies was isolated from 1 L of bacterial cell culture. The isolated FVS191cys was refolded in DTE-GSSG redox buffer. The final concentration of FVS191cys in refolding buffer influenced the refolding of FVS191cys. If the concentration of FVS191cys in refolding buffer were more than 20 mg/L, the amount of protein aggregates would increase. The refolded FVS191cys was purified by FPLC using a Superdex 75 column (Figure 2, lane 4). Formation of the Disulfide-Bonded Dimer. FVS191cys scFv′ formed disulfide-linked homodimers through air oxidation. This was demonstrated by the formation of 56 kDa proteins when 28 kDa FVS191cys monomers were incubated at 4 °C (Figure 3, lane 1). The 56 kDa protein was reduced to a 28 kDa species in the presence of DTT (Figure 3, lane 2). After removal of DTT, the 28 kDa FVS191cys reformed the 56 kDa proteins (data not shown). This indicates that the 56 kDa protein is a disulfide-linked homodimer of FVS191cys. FVS191cys (scFv′)2 was separated from monomers of FVS191cys by gel filtration chromatography. Figure 4 shows that FVS191cys (scFv′)2 was eluted from the Superdex 75 column in the second peak and FVS191cys monomers in the third peak. Fractions in the first peak contained high-molecular mass aggregates. The separated dimers and monomers of FVS191cys were examined by a nonreducing SDS-PAGE (Figure 3, lanes 3 and 4). Time and pH Dependence of Dimer Formation. The formation of FVS191cys (scFv′)2 by air oxidation

Bioconjugate Chem., Vol. 8, No. 1, 1997 67

Dimers of Anti-CD19 scFV FVS191

Figure 4. FPLC separation of monomeric (peak III), dimeric (peak II), and aggregates of FVS191cys (peak I) using a Superdex 75 column. Conditions for separation are described in Materials and Methods.

An ELISA assay was used to compare the CD19 binding ability of FVS191cys (scFv′)2 with that of mAb B43 or FVS191cys. Figure 7 indicates that the IC50 of FVS191cys (scFv′)2 is 1.45 × 10-9, the IC50 of mAb B43 is 1.32 × 10-9, and the IC50 of monomeric FVS191cys is 1.05 × 10-9. Therefore, the three different forms of B43derived proteins have similar CD19 binding abilities. Internalization by CD19+ Cells. FVS191cys (scFv′)2 was internalized by CD19+ Nalm 6 cells, as shown in Figure 8. The amount of internalized FVS191cys (scFv′)2 increased with time, and the amount of cell surface-bound FVS191cys (scFv′)2 decreased correspondingly. The amount of degraded and dissociated FVS191cys (scFv′)2 remained relatively constant. DISCUSSION

Figure 5. Formation of FVS191cys (scFv′)2 with various pHs. After air oxidation for 12 days at 4 °C, 50 µL of protein samples in TEN buffer (pH 6-12) were electrophoresed on a 12% nonreducing SDS-PAGE. Lanes 1-11 are proteins with pH 6, 6.5, 7, 7.5, 8, 8.5, 9, 9.5, 10, 11, and 12, respectively. The sizes of molecular mass markers are indicated on the left side of the gel.

depends on the length of air oxidation time and the pH of the buffer. The amount of FVS191cys (scFv′)2 increased with oxidation time and buffer pH. Figure 5 shows the formation of FVS191cys (scFv′)2 with varied pH after air oxidation for 12 days at 4 °C. From pH 6 to 9, increasing amounts of dimeric FVS191cys were found with the increments of pH (Figure 5, lanes 1-9). These data indicate that pH affects the formation of (scFv′)2 and basic buffer conditions favor dimer formation. When the pH was over 9, some of the FVS191cys degraded and some formed high-molecular mass aggregates. Therefore, the total amounts of dimers and monomers of FVS191cys with a pH of >9 were reduced (Figure 5, lanes 8-11). The formation of FVS191cys (scFv′)2 was studied by using both DTT-nonreduced FVS191cys and DTTreduced FVS191cys. Various reaction times and buffer pHs were investigated for the formation of FVS191cys (scFv′)2. The percentages of FVS191cys dimers under the above conditions were tabulated in Table 1. Temperature Effect on Dimer Formation. FVS191cys (scFv′)2 can be formed at 4 °C, room temperature, and 37 °C as shown in Figure 6. At room temperature and 37 °C, high-molecular mass aggregates of FVS191cys were found (Figure 6, lanes 2 and 3). No aggregation of FVS191cys was visible at 4 °C (Figure 6, lane 1). A greater proportion of dimers was formed at room temperature than at 4 °C; however, at room temperature, FVS191cys formed aggregates that reduced the population of reactive FVS191cys. Therefore, 4 °C is the best condition for formation of (scFv′)2. Specific CD19 Binding by FVS191cys (scFv′)2. A specific antigen binding assay was carried out with 125Ilabeled FVS191cys (scFv′)2. When the radioactivity associated with cell pellets of CD19 positive cells (Nalm 6) was arbitrarily expressed as 100%, the radioactivity associated with Molt 13 and Peer (CD19 negative cells) were determined to be only 6.2 and 7.9%, respectively. These results demonstrate that the FVS191cys (scFv′)2 is highly specific for CD19 antigen positive cells.

In this report, we have described the production of FVS191cys, an anti-CD19 scFv engineered with a Cterminal cysteine, and demonstrated that FVS191cys forms a disulfide-linked homodimer. In this study, the effects of air oxidation time, buffer pH, and temperature on the formation of dimeric FVS191cys were investigated for the purpose of defining the optimal conditions for dimer formation. Our results indicate that formation of disulfide-linked (scFv′)2 is dependent on oxidation time and pH. Specially, the yield of (scFv′)2 increases with time and pH. The time dependence of (scFv′)2 formation is related to the size and tertiary structure of the scFv′ protein. The solubility of thiols affects the reaction rate of the disulfide bond formation between two molecules; the longer the peptides, the harder it is for oxidation to occur (15). The scFv′ has a mass of 28 kDa, which is large for a peptide, and is, therefore, slow to form (scFv′)2. If the C-terminal cysteine of scFv′ is sterically hindered, it would be more difficult to form a disulfide bond. The pH dependence of dimer formation is due to the oxidation process. There are two steps involved in the air oxidation of thiols: the formation of a thiolate anion and the formation of a thiyl radical by electron transfer between the thiolate anion and oxygen (15). Therefore, oxidation is promoted under basic conditions. In our case, the maximum yield of FVS191cys (scFv′)2 was achieved at pH 8.5 or 9. Above pH 9, although the ratio of dimer to monomer of FVS191cys was high, the dimer yield is low because the total amount of FVS191cys was decreased. The reduction of FVS191cys was due to the high pH in which FVS191cys was unstable and was degraded. FVS191cys aggregates at temperatures above 4 °C. Different scFv′s may have different intrinsic tendencies to form multimers or aggregates (16). Those tendencies may be related to the tertiary structures of the scFv′ and the distribution of charged or uncharged amino acids on the surfaces of the scFv′ molecules. At higher temperatures, the disulfide bond of scFv′ was less stable; FVS191cys was more likely to form covalently linked multimers. Therefore, 4 °C was the best condition for formation of (scFv′)2. FVS191cys (scFv′)2 can be formed from DTT-nonreduced FVS191cys; this indicates that DTT reduction is not necessary for dimer formation. However, McCartney et al. (16, 17) suggested that DTT reduction was a necessary step for forming (scFv′)2. Their different conclusion may have resulted from the different method that they used to refold scFv′. They used urea-glutathione redox buffer to refold 741F8 scFv′; the final concentration of oxidized glutathione in the refolding buffer was 100 mM. Their refolding method converted reduced thiol groups of scFv′ to mixed disulfides with

68 Bioconjugate Chem., Vol. 8, No. 1, 1997

Wang et al.

Table 1. Summary of the Percentage of Dimeric FVS191cys (% d) under Different Times and pHsa DTT-nonreduced FVS191cys

DTT-reduced FVS191cys

pH

day 1

day 5

day 12

day 16

day 20

day 29

6 6.5 7 7.5 8 8.5 9 9.5 10

1.4 1.6 2.0 2.8 4.6 7.5 11.5 14.4 14.7

13.0 13.7 14.7 14.7 24.0 21.0 29.4 35.0 44.9

3.3 4.4 9.0 17.0 26.4 32.7 34.0 38.9 46.6

6.1 8.3 14.7 25.8 35.2 39.0 43.1 46.8 56.2

5.1 8.2 16.2 27.4 36.1 33.1 43.3 47.8 50.0

7.4 13.7 22.4 35.0

a

38.4 44.7 51.2

pH

day 1

day 5

day 12

day 16

day 20

day 29

7 7.5 8 8.5 9

5.5 7.7 8.5 9.8 12.7

5.9 3.5 4.7 5.5 7.9

11.0 14.9 16.0 17.8 18.3

24.8 28.7 29.8 28.9 30.6

19.9 25.5 28.7 29.5 29.9

42.4 36.1 32.6 33.4 35.6

24.8

13.9

21.2

34.0

30.2

37.2

10

The calculation of % d for dimers from DTT-nonreduced and DTT-reduced FVS191cys is described in Materials and Methods.

Figure 6. Formation of FVS191cys (scFv′)2 at 4 °C, room temperature, and 37 °C for 4 h. Proteins (50 µL) incubated at the above temperatures were electrophoresed on a 12% nonreducing SDS-PAGE: lane 1, proteins incubated at 4 °C; lane 2, proteins incubated at room temperature; and lane 3, proteins incubated at 37 °C. The sizes of molecular mass markers are indicated on the left side of the gel. Figure 8. Internalization of 125I-labeled FVS191cys (scFv′)2 by Nalm 6 cells. The percentage of radioactivity calculated as the percentage of total CPM was plotted against each time point. At each time point, cell pellets were separated from supernatants by low-speed centrifugation: (4) acid-papain released, cell surface-bound FVS191cys (scFv′)2; (0) internalized FVS191cys (scFv′)2 which could not be released by acid-papain treatment; (O) dissociated FVS191cys (scFv′)2 which was precipitated by TCA from supernatant; and (2) degraded FVS191cys (scFv′)2 which could not be precipitated by TCA from supernatants.

Figure 7. Affinity assay by competitive ELISA. Various amounts of FVS191cys (9), FVS191cys (scFv′)2 (b), and mAb B43 (O) were mixed up with a constant amount of B43-AP. The means of percentage inhibition of B43-AP binding were plotted against log molar concentrations of inhibition proteins. The bars represent a standard deviation.

glutathione; no thiol was available for forming a disulfide bond. Thus, DTT had to be used to eliminate the blocking group of C-terminal cysteines (16), and then the 741F8 (scFv′)2 was formed. In our study, however, we used DTE-GSSG redox buffer to refold FVS191cys; the final concentration of oxidized glutathione was 8 mM. Most FVS191cys compounds produced by our refolding method were not blocked by glutathione (data not shown). Therefore, DTT reduction was not necessary for formation of the FVS191cys (scFv′)2. DTT reduction is probably

only necessary when the refolded scFv′s are blocked by glutathione. Among the two dimer formation methods, our method has the advantages of fewer experimental steps and lower cost since GSSG is an expensive chemical. It was also observed that the yields of FVS191cys (scFv′)2 from different preparations were variable. This phenomenon may be due to the differences in the protein concentration. The relationship between the concentration of scFv and the formation of (scFv)2 was studied by Desplancq et al. (18). They reported that the proportion of associated dimer to monomer was a function of concentration, with an increased concentration of monomer leading to an increased proportion of dimers. Their study indicates that formation of (scFv)2 can be improved by increasing the concentration of scFv. We found that there was an upper limit for the concentration of FVS191cys monomers. If the concentration of FVS191cys is over 1.5 mg/mL, it will aggregate. The optimal concentration of scFv′ for dimer formation has to be studied individually since different scFv′s have different tendencies to aggregate. The CD19 binding assay with 125I-labeled FVS191cys (scFv′)2 revealed the specific antigen binding ability of FVS191cys (scFv′)2. The ELISA assay indicated that the antigen binding ability of FVS191cys (scFv′)2 was similar to that of parental mAb B43 and monomeric FVS191cys.

Dimers of Anti-CD19 scFV FVS191

Thus, dimers formed through C-terminal cysteines maintained the original antigen binding capability. This result is consistent with other reports (8, 9, 19). The ELISA assay with a (scFv′)2 derived from mAb215 showed that the binding constant was quite close to that of the parental mAb and 4-fold higher than those of the scFv′ monomers (8). Dimeric ScAb, a bivalent singlechain antibody fragment against Pseudomonas aeruginosa, had an antigen binding profile similar to that of the parental antibody (19). These results indicate that dimeric scFv′ has the same antigen binding characters as the intact antibody from which it was derived. To our knowledge, FVS191cys (scFv′)2 is the first dimeric scFv′ to be analyzed in an internalization assay. FVS191cys (scFv′)2 can be internalized by CD19+ cells in a manner similar to that of its parental mAb B43 (unpublished data). This result indicates that FVS191cys (scFv′)2 can potentially replace intact mAb as tumor targeting agents on the basis of their antigen binding and internalization ability. ScFvs containing C-terminal cysteines have proved to be very useful molecules. So far, several homodimers of scFvs have been formed through the C-terminal cysteines (6, 8, 17). The other studies have shown that the free sulfhydryl groups of scFvcys can be used to form heterodimers (7) or immunoconjugates with other chemical agents (8). In our laboratory, FVS191cys has been used to form a single-chain immunotoxin containing ricin A chain (unpublished data). Utilization of engineered cysteine to form disulfide-bonded molecules has also been employed to generate bivalent (9, 20) and bispecific (21) recombinant antibodies. An alternative method for producing bivalent scFv is expression of protein in an E. coli system in which the proteins are secreted periplasmically. In periplasmic space, the proteins are refolded and assembled into the dimeric form (22). This method is more attractive when the proteins of interest can be produced in a large amount. In our studies, scFv FVS191 was expressed in such a system, but the yield was low. In conclusion, we have demonstrated a simple and efficient method for producing disulfide-linked dimeric scFv′ through C-terminal cysteines. In this method, scFv′ is refolded in DTE-GSSG buffer and dimeric scFv′ is formed by air oxidation in high-pH buffer. The formation of FVS191cys (scFv′)2 through air oxidation is time- and buffer pH-dependent. We found the optimal conditions for forming dimeric FVS191cys by air oxidation to be 4 °C and a pH between 8.5 to 9. The FVS191cys (scFv′)2 has a CD19 binding ability similar to that of its parental mAb B43 and is internalized by CD19 positive cells. Therefore, FVS191cys (scFv′)2 may be a good candidate for immunotargeting while being one-sixth of the size of the intact mAb. ACKNOWLEDGMENTS

We gratefully thank Ms. Cedith Copenhaver for her help in reviewing and proofreading the manuscript. LITERATURE CITED (1) Jain, R. K. (1987) Transport of molecules in the tumor interstitium: a review. Cancer Res. 47, 3039-3051. (2) Foon, K. A. (1989) Biological response modifiers: the new immunotherapy. Cancer Res. 49, 1621-1639. (3) Yokota, T., Milenic, D. E., Whitlow, M., and Schlom, J. (1992) Rapid tumor penetration of a single-chain Fv and comparison with other immunoglobulin forms. Cancer Res. 52, 3402-3408.

Bioconjugate Chem., Vol. 8, No. 1, 1997 69 (4) Milenic, D. E., Yokota, T., Filpula, D. R., Finkelman, M. A. J., Dodd, S. W., Wood, J. F., Whitlow, M., Snoy, P., and Schlom, J. (1991) Construction, binding properties, metabolism, and tumor targeting of a single-chain Fv derived from the pancarcinoma monoclonal antibody CC49. Cancer Res. 51, 6363-6371. (5) Colcher, D., Bird, R., Roselli, M., Hardman, K., Johnson, D., Pope, S., Dodd, S. W., Pantoliano, M. W., Milenic, D. E., and Schlom, J. (1990) In vivo tumor targeting of a recombinant single-chain Fv antigen-binding protein. J. Natl. Cancer Inst. 82, 1191-1197. (6) Adams, G. P., McCartney, J. E., Tai, M., Oppermann, H., Huston, J. S., Stafford, W. F., III, Bookman, M. A., Fand, I., Houston, L. L., and Weiner, L. M. (1993) Highly specific in vivo tumor targeting by monovalent and divalent forms of 741F8 anti-c-erbB2-2 single-chain Fv. Cancer Res. 53, 40264034. (7) Tai, M.-S., McCartney, J. E., Adams, G. P., Jin, D., Hudziak, R. M., Oppermann, H., Laminet, A. A., Bookman, M. A., Wolf, E. J., Liu, S., Stafford, W. F., III, Fand, I., Houston, L. L., Weiner, L. M., and Huston, J. S. (1995) Targeting c-erbB2 expression tumors using single-chain Fv monomers and dimers. Cancer Res. 55, 5983-5989. (8) Kipriyanov, S. M., Dubel, S., Breitling, F., Kontermann, R. D., and Little, M. (1994) Recombinant single-chain Fv fragments carrying C-terminal cysteine residues: production of bivalent and biotinylated miniantibodies. Mol. Immunol. 31 (14), 1047-1058. (9) Cumber, A. J., Ward, E. S., Winter, G., Parnell, G. D., and Wawrzynczak, E. J. (1992) Comparative stabilities in vitro and in vivo of a recombinant mouse antibody FvCys fragment and a bisFvCys conjugate. J. Immunol. 149 (1), 120-126. (10) Pack, P., and Pluckthun, A. (1992) Miniantibodies: use of amphipathic helices to produce functional, flexibly linked dimeric Fv fragments with high avidity in Escherichia coli. Biochem. 31 (6), 1579-1584. (11) Bejeck, B., Wang, D., Berven, E., Pennell, C. A., Peiper, S. C., Poppema, S., Uckun, F., and Kersey, J. H. (1995) Development and characterization of three recombinant single chain antibody fragments (scFv) directed against the CD19 antigen. Cancer Res. 55, 2346-2351. (12) Buchner, J., Pastan, I., and Brinkmann, U. (1992) A method for increasing the yield of properly folded recombinant fusion proteins: single-chain immunotoxins from renaturation of bacterial inclusion bodies. Anal. Biochem. 205, 263270. (13) Siegall, C. B., Chace, D., Mixan, B., Garrigues, U., Wan, H., Paul, L., Wolff, E., Hellstrom, I., and Hellstrom, K. E. (1994) In vitro and in vivo characterization of BR96 scFvPE40. J. Immunol. 152 (2), 2376-2484. (14) Press, O. W., Farr, A. G., Borroz, K. I., Anderson, K. A., and Martin, P. J. (1989) Endocytosis and degradation of monoclonal antibodies targeting human B-cell malignancies. Cancer Res. 49, 4906-4914. (15) Ohno, A., and Oae, S. (1977) in Organic Chemistry of Sulfur (S. Oae, Ed.) pp 155-156, Plenum Press, New York. (16) McCartney, J. E., Tai, M.-S., Hudziak, R. M., Adams, G. P., Weiner, L. M., Jin, D., Stafford, W. F., III, Liu, S., Bookman, M. A., Laminet, A. A., Fand, I., Houston, L. L., Oppermann, H., and Huston, J. S. (1994) Engineering disulfide-linked single-chain Fv dimers [(sFv′)2] with improved solution and targeting properties: anti-digoxin 26-10 (sFv′) and anti-c-erbB-2 741F8 (sFv′) made by protein folding and bonded through C-terminal cysteinyl peptides. Protein Eng. 8 (3), 301-314. (17) McCartney, J. E., Tai, M. S., Oppermann, H., Jin, D., Warren, F. D., Weiner, L. M., Bookmann, M. A., Stafford, W. F., III, Houston, L. L., and Huston, J. S. (1993) Refolding of single-chain Fv with C-terminal cysteine (sFv): formation of disulfide-bonded homodimers of anti-c-erbB-2 and antidigoxin sFv. Miami Short Rep. 3, 91. (18) Desplancq, D., King, D. J., Lawson, A. D. G., and Mountain, A. (1994) Multimerization behavior of single chain Fv variants for the tumor-binding antibody B72.3. Protein Eng. 7, 1027-1033. (19) McGregor, D. P., Molloy, P. E., Cunningham, C., and Harris, W. J. (1994) Spontaneous assembly of bivalent single

70 Bioconjugate Chem., Vol. 8, No. 1, 1997 chain antibody fragments in Escherichia coli. Mol. Immunol. 31, 219-226. (20) Carter, P., Kelley, R. F., Rodrigues, M. L., Snedecor, B., Covarrubias, M., Velligan, M., Wong, W. L. T., Rowland, A. M., Kotts, C. E., Carver, M. E., Yang, M., Bourell, J. H., Shepard, H. M., and Henner, D. (1992) High level Escherichia coli expression and production of a bivalent humanized antibody fragment. Bio/Technology 10, 163-167. (21) Shalaby, M. R., Shepard, H. M., Presta, L., Rodrigues, M. L., Beverley, P. C. L., Feldmann, M., and Carter, P. (1992)

Wang et al. Development of humanized bispecific antibodies reactive with cytotoxic lymphocytes and tumor cells overexpressing the HER2 protooncogene. J. Exp. Med. 175, 2217-225. (22) Zhu, Z., Zapata, G., Shalaby, R., Snedecor, B., Chen, H., and Carter, P. (1996) High level secretion of a humanized bispecific diabody from Escherichia coli. Bio/Technology 14, 192-196.

BC9600776