B-Domain Deleted Recombinant Coagulation Factor VIII Modified with

Bioconjugate Chem. 2000, 11, 387−396

387

B-Domain Deleted Recombinant Coagulation Factor VIII Modified with Monomethoxy Polyethylene Glycol Johanna Ro¨stin,† Anna-Lisa Smeds,‡ and Eva A° kerblom*,§ Recombinant Factor VIII, R&D, Pharmacia & Upjohn, S-112 87 Stockholm, Sweden. Received October 18, 1999

Recombinant coagulation factor VIII (r-VIII SQ) was chemically modified with monomethoxy poly(ethylene glycol) (mPEG). Three mPEG derivatives were used for coupling to the r-VIII SQ lysines, a mixed anhydride of monomethoxy poly(ethylene glycol) succinic acid (mPEG-SAH), monomethoxy poly(ethylene glycol) succinimidyl succinate (mPEG-SS), and monomethoxy poly(ethylene glycol) tresylate (mPEG-TRES). A consequence of the modification with all derivatives was a substantial reduction in coagulant activity, even at very low degrees of modification. A method was developed with the purpose of avoiding conjugation at certain important biological sites on the factor VIII and thereby producing conjugates with better retained activity. This was achieved by immobilizing the protein onto a solid matrix during the modification reaction. Characterization of conjugates by SDS-PAGE, western blots, interaction with von Willebrand factor (vWf), and thrombin activation/inactivation analyses was undertaken. The SDS-PAGE and western blots revealed coupling heterogeneity regarding degree of modification. The amount of factor VIII able to bind to vWf decreased with the conjugation. Thrombin activated the modified factor VIII to essentially the same extent as the reference preparation of r-VIII SQ. Inactivation of the modified factor VIII was, however, slower than inactivation of the unmodified protein. Finally, an in vitro study was performed to evaluate the influence of the mPEG modification on the protein stability in extract of porcine tissue. Despite that conjugates with low degrees of modification were included in the study, the coagulant activity was preserved to a significantly higher extent in all incubation mixtures containing conjugates compared to that with unmodified protein.

INTRODUCTION

Factor VIII is a protein, which plays a significant role as a cofactor in the blood coagulation process. Deficiency of or defects in this protein results in the bleeding disorder Hemophilia A. Hemophilia A patients are presently treated by intravenous administration of factor VIII, given on demand or preferably as prophylactic treatment several times a week (1). It would be more convenient for the patients if the stability of the substance in vivo could be increased to achieve a longer circulation time and a consequent reduction in frequency of administration. About 10-20% of hemophilia patients develop inhibiting antibodies to factor VIII (1), and a reduction of immunogenicity would greatly improve the efficacy of the medicine in these patients. A well-known technique for improving pharmacokinetic properties of therapeutic proteins in vivo is covalent attachment of monomethoxy poly(ethylene glycol) (mPEG)1 to the proteins. Increased in vivo half-life, increased bioavailability after subcutaneous administration, and reduced immunogenicity are properties that have been enhanced by pegylation of polypeptides for use in a number of therapeutic areas (2, 3). However, an impaired function may also be a result of the modification, an effect that has been reported to increase with the degree of * To whom correspondence should be addressed. Telephone: +46 18 471 4337. Fax: +46 18 471 4474. E-mail: eva.akerblom@ orgf3.bmc.uu.se. † Recombinant Factor VIII. ‡ Present address: Medical Product Agency, Box 26, S-751 03 Uppsala, Sweden. § Present address: Department of Organic Pharmaceutical Chemistry, Uppsala Biomedical Center, Box 574, S-751 23 Uppsala, Sweden.

modification (4). A number of coupling reagents is described, most of them with reactivity toward the -amino groups of lysines (for a review, see ref 5). As polypeptides often contain several lysines, attempts to derivatize more selectively to avoid the loss of activity have been made. Site-directed mutagenesis of recombinant interleukin-2 (rIL-2) is one example where a unique modification site was introduced by conversion of a Thr to a Cys (6). Another approach was used by Caliceti et al. (7). To protect functional sites of trypsin, the enzyme was reversibly bound to an inhibitor immobilized on a solid matrix during the procedure of mPEG conjugation. Factor VIII functions in the blood coagulation system as an essential cofactor for activation of factor X to factor Xa. The molecule circulates in plasma as a heterodimer with structural domains A1-A2-B and A3-C1-C2 (8) as shown in Figure 1. The heavy chain consists of the A1-A2-B domains with a molecular mass ranging from 90 to 200 kDa due to variability of the B domain (8, 9). The 80 kDa light chain consists of the A3-C1-C2 domains. The two chains are held together by a divalent 1 Abbreviations: APC, activated protein C; CHO, Chinese hamster ovary; FF, fast flow; FPLC, fast-protein liquid chromatography; GPC, gel permeation chromatography; HIC, hydrophobic interaction chromatography; HR, high resolution; IU, international unit; kDa, kilodalton; mAb, monoclonal antibody; mPEG, monomethoxy poly(ethylene glycol); mPEG-SA, mPEG succinic acid; mPEG-SAH, mixed anhydride of mPEG succinic acid; mPEG-SS, mPEG succinimidyl succinate; mPEG-TRES, mPEG tresylate; NMR, nuclear magnetic resonance; PAGE, polyacrylamide gel electrophoresis; PG, prep grade; rIL-2, recombinant interleukin 2; r-VIII SQ, recombinant factor VIII SQ; SDS, sodium dodecyl sulfate; SQ, Ser Gln; std, standard deviation; VIII:C, factor VIII activity; vWf, von Willebrand factor.

10.1021/bc990137i CCC: $19.00 © 2000 American Chemical Society Published on Web 04/04/2000

388 Bioconjugate Chem., Vol. 11, No. 3, 2000

Ro¨stin et al.

Figure 1. Structural domains of the heavy (90-200 kDa) and the light (80 kDa) chains of factor VIII, acidic regions, and the thrombin activation cleavage sites.

metal ion bridge (9). Factor VIII has several sites important for its function. Activation of factor VIII starts by thrombin cleavage at Arg 372 and Arg 740 in the heavy chain and at Arg 1689 in the light chain (8, 10, 11). The cleavage sites are located after acidic spacer regions in the polypeptide. Factor VIII is also susceptible to proteolytic cleavage at Arg 336 by activated protein C, APC (12), factor Xa, and thrombin (10), and this results in inactivation. This is important for the downregulation of the coagulation cascade. Factor VIII circulates in human plasma noncovalently linked to von Willebrand factor (vWf) (13). vWf has an important stabilizing effect on factor VIII by protecting it from proteolytic degradation. The human in vivo half-life of factor VIII is 10-15 h (1), and it has been suggested that the half-life is highly dependent on the factor VIII-vWf complex formation (14-16). The vWf-binding site is located on the factor VIII light chain, at the A3 domain (17). Cleavage of factor VIII by thrombin at Arg 1689 releases the vWf together with the interacting sequence during activation of factor VIII. The function of the B domain is not known, and it is not required for cofactor activity (9, 18). Pharmacia & Upjohn has constructed a recombinant factor VIII lacking this region for treatment of Hemophilia A. The deletion derivative, r-VIII SQ, was produced from a single gene construct encoding for an amino acid sequence, where the Ser 743 of the heavy-chain C-terminal was fused to the Gln 1638 of the light-chain N-terminal (19). In this study, the pegylation technique was applied to r-VIII SQ. Coupling was performed by using three mPEG derivatives with reactivity toward the -amino groups of lysines. The factor VIII function was easily disturbed, and the coagulant activity decreased rapidly with degree of modification. The stability of modified factor VIII was investigated in in vitro experiments using porcine extracts representing the subcutaneous environment. The mPEG modification protected the r-VIII SQ from rapid inactivation in the incubation mixtures, even at low degrees of modification. These results prompted us to investigate whether this effect was applicable to in vivo conditions, and a pharmacokinetic study was performed in cynomolgus monkeys. A 4-fold higher bioavailability

was obtained of the conjugates compared to the unmodified protein after subcutaneous injection (20). MATERIALS AND METHODS

mPEG tresylate (mPEG-TRES) and mPEG succinimidyl succinate (mPEG-SS) of molecular weight 5000 were purchased from Shearwater Polymers, Inc., Huntsville, AL. mPEG was obtained from Hoechst and Union Carbide. Hepes and plasma contact activator kaolin (hydrated aluminum silicate) were purchased from Sigma; boric acid and Tween 80 were from Merck; and deuterium oxide was from Aldrich. Diagen (phospholipid) was obtained from Diagnostic Reagents Ltd., U.K., and triethylamine and N-tert-butylurea from Fluka. Human thrombin was purchased from Enzyme Research Laboratories Inc., IN, and antithrombin was produced by Pharmacia & Upjohn. vWf was prepared from a highpurity therapeutic concentrate (Octonativ-M from Pharmacia & Upjohn) by purification on a Sepharose CL-2B, to which polyclonal goat anti-human vWf antibodies were covalently coupled. The following chromatography equipment and media from Amersham Pharmacia Biotech, Sweden, were used: a fast-protein liquid chromatography (FPLC) system, Superose 12 bulk gel, Q Sepharose FF bulk gel, a prepacked Superdex 200 PG XK 16/60 column, and a prepacked Superdex 200 HR 10/30 column. The IR spectra was recorded on a 5% 1,2-dichloroethane solution in a barium fluoride liquid cell with a PerkinElmer model 580 infrared spectrophotometer. Preparation of r-VIII SQ. Briefly, the r-VIII SQ was produced by a cell culture process using Chinese Hamster Ovary (CHO) cells (19) in a serum-free medium. The conditioned medium was continuously harvested, and the factor VIII was purified by four chromatography steps (21): cationic exchange, immunoaffinity using monoclonal antibodies (mAb), anionic exchange, and hydrophic interaction chromatography (HIC). Virus inactivation was performed after the cationic exchange step using the solvent/detergent method according to Horowitz et al. (22). The factor VIII was eluted from the HIC column in a buffer containing 50 mM L-histidine, 0.5 M ammonium acetate, 50 mM calcium chloride, and 0.02% Tween 80,

B-Domain Deleted Recombinant Coagulation Factor VIII

pH 6.8. The purity was very high, resulting in a mean specific activity of 15 000 IU/mg of protein. Synthesis of mPEG-SA. mPEG (MW 3000, 120 g, 45 mmol) was added to succinic anhydride (8 g, 80 mmol) in 240 mL of dichloromethane. The solution was refluxed for 20 h. Excess succinic anhydride was then converted to monoethyl succinate by addition of 33 mL of ethanol, and the reaction mixture was refluxed for another 6 h. After evaporation in a vacuum, the residue was repeatedly washed with ether, subsequently dissolved in acetone, and precipitated in petroleum ether. The mPEGSA was analyzed by titration and shown to contain 0.32 mmol of COOH/g. 1H NMR (CDCl3): δ 2.57-2.66 (m, 4H, CH2CH2CO), 3.35 (s, 3H, CH3O), 3.44 (t, 2H, CH2OCO), 2.52 (t, 2H, CH3OCH2), 3.55-4.00 [br, (CH2CH2O)n], 3.79 (t, 2H, CH3OCH2CH2), 4.23 (t, 2H, CH2CH2OCO). 13C NMR (CDCl3): δ 29.0, 29.5 (CH2CH2CO), 59.1 (CH3O), 63.9 (CH2OCO), 69.1 (CH3OCH2), 70.7 [(CH2CH2O)n], 72.0 (CH3OCH2CH2), 77.4 (CH2CH2OCO), 172.2 (OCO), 173.7 (COOH). The peaks were identified by comparing with 1H and 13C NMR spectra of PEG and mPEG. IR (ClCH 2 CH2Cl): ν 1740 cm-1 (ester and carboxylic acid carbonyl). Activation of mPEG-SA (mPEG-SAH). mPEG-SA (77.3 g, 27.7 mmol) was dissolved in 700 mL of dichloromethane. Triethylamine (7.8 mL, 55.3 mmol) and isobutyl chloroformate (7.7 mL, 55.3 mmol) were subsequently added. The reaction solution was stirred for 30 min at room temperature and evaporated in a vacuum. The solid residue was washed seven times with 350 mL of petroleum ether by shaking the flask. After the last wash, the petroleum ether was decanted and the rest of the solvent was removed by evaporation in a vacuum. The mPEG-SAH was stored in a freezer. 1H NMR (CDCl3): δ 0.92 [d, 6H, CH(CH3)2], 1.93-2.03 [m, 1H, CH(CH3)2], 2.67 (t, 2H, CH2CH2CO), 2.77 (t, 2H, CH2CH2CO), 3.34 (s, 3H, OCH3), 3.42 (t, 2H, CH2OCO), 3.51 (t, 2H, CH3OCH2), 3.55-3.75 [br, (OCH2CH2)n], 3.78 (t, 2H, CH3OCH2CH2), 4.00 [d, 2H, OCH2CH(CH3)2], 4.22 (t, 2H, CH2CH2OCO). 13C NMR: δ 18.9 [CH(CH3)2], 27.7, 28.5, 29.3 [CH(CH3)2; CH2CH2CO], 59.1 (CH3O), 64.2 (CH2OCO), 69.0 (CH3OCH2), 70.6 [(CH2CH2O)n], 72.0 (CH3OCH2CH2), 75.6 (CH2CH2OCO), 166.9 (COOCO), 171.6 (OCO). In the 1H and 13C NMR spectra, there were peaks indicating an impurity of triethylamine HCl in mPEGSAH. IR (ClCH2CH2Cl): ν 1825 cm-1 (anhydride carbonyl), 1740 cm-1 (ester carbonyl). The ratio between the heights of the peak at 1825 and the peak at 1740 cm-1 was used for studying the stability of mPEG-SAH. mPEG-SAH was stable for at least one year in a freezer. Random Coupling of Activated mPEG Derivatives to Factor VIII. A buffer exchange of the r-VIII SQ HIC eluate was performed by gel permeation chromatography on a Superose 12 column, previously equilibrated with the coupling buffer: 0.25 M Hepes, 3.7 mM calcium chloride, pH 7.8. As a control experiment to determine the influence of mPEG on factor VIII, nonreactive mPEG was first added to aliquots of the factor VIII solution at a molar ratio of mPEG to protein as indicated in Table 1. Factor VIII activity (VIII:C) was determined by the chromogenic assay. Three mPEG reagents were used to modify the r-VIII SQ lysines randomly (Figure 2). Activated mPEG was added in a solid state to the factor VIII solution (containing 0.4-0.7 mg/mL protein), and the mixture was incubated under rotation for 1.5 h at room temperature. The degree of modification was controlled by the molar excess of the mPEG added. To remove the excess of the activated mPEG and to separate the fractions of modified factor VIII according to size, gel permeation chromatography

Bioconjugate Chem., Vol. 11, No. 3, 2000 389 Table 1. Factor VIII Activity in an mPEG Containing Environment molar excess mPEG (mol of mPEG/mol of r-VIII SQ)

VIII:C IU/mL

remaining activity %

0 100 200 400 2000 2800

860 880 960 890 960 990

100 >100 >100 >100 >100 >100

a

b

c

d

Figure 2. (a) Attachment of mPEG(3000)SAH to r-VIII SQ. (b) Attachment of mPEG(5000)SS to r-VIII SQ. (c) Attachment of mPEG(5000)TRES to r-VIII SQ by alkylation. (d) Attachment of mPEG(5000)TRES to r-VIII SQ by acylation.

was performed. A Superdex 200 PG column equilibrated with a buffer of 14.7 mM L-histidine, 0.4 M sodium chloride, and 3.7 mM calcium chloride, pH 6.9, was used. The conjugates included in Table 2 a refer to protein fractions corresponding to the first and second half of the protein peaks from each coupling reaction. The fractions were analyzed separately and denoted pool 1 and pool 2. The specific activities were calculated from VIII:C determined by the chromogenic assay and the protein content determined by amino acid analysis. Coupling of Activated mPEG to Factor VIII Immobilized on a Solid Phase. The factor VIII was adsorbed on an anionic exchange gel (approximately 2 mg of protein/g of dry gel) during the modification reaction. A C-10 column packed with Q Sepharose FF, equilibrated with a buffer of the composition 50 mM L-histidine, 0.15 M sodium chloride, and 4 mM calcium chloride, pH 6.5, was used to adsorb the factor VIII. Prior to loading the protein onto the column, the HIC eluate of r-VIII SQ was conformed to the ionic strength and pH of the equilibration buffer by a 3-fold dilution using a buffer of 50 mM L-histidine and 50 mM calcium chloride, pH 6.4. The factor VIII was then loaded onto the column at a linear flow rate of 15 cm/h, and the column was subsequently equilibrated with the coupling buffer (0.25


Ro¨stin et al.

Table 2. Determination of Specific Activities and Degrees of Modification of Factor VIII (a) Modified with mPEG Reagents Randomly

conjugate

mPEG reagent r-VIII SQ blanka mPEG(3000)SAH

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

mPEG(3000)SAH mPEG(3000)SAH mPEG(5000)SS mPEG(5000)SS mPEG(5000)TRES mPEG(5000)TRES mPEG(5000)TRES

pool

molar excess mPEG added (mol of mPEG/ mol of r-VIII SQ)

degree of modification (mol of mPEG/ mol of r-VIII SQ)

specific activity (kIU/mg)

specific activity (% of r-VIII SQ)

1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2

0 0 35 35 70 70 110 110 30 30 60 60 50 50 100 100 200 200

0 0 2.5 1.2 2.4 na 4.8 3.8 3.2 2.7 6.6 5.8 3.6 2.4 3.6 2.0 5.4 4.2

15.0 14.0 4.2 6.8 2.0 4.7 1.3 2.7 3.9 7.3 0.5 1.2 7.1 10.3 9.6 10.3 2.9 5.8

100 93 28 45 13 31 9 18 26 49 3 8 47 69 64 69 19 39

(b) Modified with mPEG(3000)SAH under Immobilization

conjugate

molar excess mPEG added (mol of mPEG/ mol of r-VIII SQ)

degree of modification (mol of mPEG/ mol of r-VIII SQ)


specific activity (% of r-VIII SQ)

17 18 19 20 21 22 23 24 25 26 27 28b 29b

160 170 200 240 250 260 260 310 330 400 480 370 530

3.5 1.2 2.6 3.4 0.8 4.2 2.1 5.0 2.4 5.9 na 4.4 9.5

11.4 12.6 8.3 8.2 13.3 7.5 10.6 5.5 8.5 4.3 5.2 7.8 2.8

76 84 55 55 89 50 71 37 57 29 35 52 19

a The r-VIII SQ was subjected to the coupling procedure without addition of mPEG. b pH of the protein load was not adjusted prior to adsorption onto the solid phase. The conjugates are not included in Figure 5.

M Hepes and 3.7 mM calcium chloride, pH 7.8). The gel was transferred to a reaction vessel, and the activated mPEG was added in a solid state to the slurry at a molar ratio to factor VIII as indicated in Table 2b. The mixture was incubated under rotation for 1.5 h at room temperature. The gel was then repacked in the column and was extensively washed, and the modified factor VIII was eluted with a buffer of 50 mM L-histidine, 0.6 M sodium chloride, 4 mM calcium chloride, and pH 6.8 at 15 cm/h. The specific activities of the fractions were calculated from VIII:C determined by the chromogenic assay and the protein content was determined spectrophotometrically. Fractionation of mPEG Modified Factor VIII. Prior to SDS-PAGE, western blots, vWf interaction, and thrombin activation analyses, the material produced by the coupling procedures above was further fractionated by gel permeation chromatography. A Superdex 200 HR 10/30 column equilibrated with 14.7 mM L-histidine, 0.4 M sodium chloride, 3.7 mM calcium chloride, and 0.02% Tween 80, pH 6.9, was used. Protein Concentration and Amino Acid Content. Approximate factor VIII concentrations were determined spectrophotometrically with a Hitachi U-3200 spectrophotometer at 280 nm; the calculations were based on ) 306 × 103 (M-1 cm-1) and a molecular mass of 165 306

Da for the r-VIII SQ, respectively. The amino acid content was determined by first hydrolyzing the protein in 6 M HCl at 155 °C for 45 min and then analyzing the sample on an ion-exchange chromatography column where the amino acids were detected with ninhydrin. A Beckman 6300 amino acid analyzer equipped with a System Gold data handling system (Beckman) was used. Determination of Degree of Modification. The mPEG content was determined by NMR, and the protein content by amino acid analysis. From these analyses, the degree of modification was calculated as moles of mPEG per moles of protein. The NMR determinations were performed either on a Varian Unity 500 or a Bruker 500 DRX instrument. Prior to the NMR analysis, a buffer exchange of the pegylated protein was performed by gel permeation chromatography using a Superose 12 column equilibrated with a buffer of the following composition: 0.1 M boric acid and 0.2 M sodium chloride, pH 7.8. The samples were repeatedly lyophilized and redissolved in deuterium oxide to remove traces of water before the NMR analysis. A standard curve using mPEG(3000) with concentrations ranging from 0.003 to 0.250 mg/mL was prepared. N-tert-Butylurea (0.0125 mg/mL) was used as an internal standard. SDS-PAGE. SDS-PAGE was performed according to Laemmli (23). The gel thickness was 0.75 mm. The total


polyacrylamide concentration was 4% in the stacking gel and 7.5% in the separating gel and the cross-linkage was 3.3%. Samples of 0.7 µg were reduced by 5% (vol) 2-mercaptoethanol in 2% (mass/vol) SDS. The mixtures were heated for 5 min at 90 °C. The PAGE was run at 20 mA/gel for 4 h and the electrophoresis was followed by silver staining. Western Blot Analysis. Western blot analysis was performed using polyclonal antibodies against plasma factor VIII as described previously (9). Two primary anti(factor VIII) antibodies, produced as described in ref 19, were used. The antibodies were directed against the C-terminus of the 90 kDa chain (Tyr 729 to Arg 740) and the C-terminus of the 80 kDa chain (Ala 2318 to Tyr 2332), respectively. Factor VIII Activity. The factor VIII activity, VIII: C, was determined by a chromogenic substrate method (24), using the Coatest factor VIII kit (Chromogenix AB, Mo¨lndal) and by a one-stage clotting method (25) where the VIII:C was determined using diagen platelet substitute diluted 1:50, containing 5 mg/mL kaolin. The definition of 1 IU corresponds to the coagulant activity of 1 mL of normal plasma as determined using a reference standard calibrated against the current international standard. Interaction with Thrombin. The VIII:C was determined by the one-stage clotting method. mPEG-modified factor VIII conjugates were diluted to 1 and 10 IU/mL, respectively, in 50 mM Tris buffer, pH 7.4, containing 0.15 M sodium chloride and 4 mM calcium chloride. Thrombin was added to 1 and 10-1 NIH U/IU factor VIII, respectively. The changes in factor VIII activity were followed for 20 min by immediate assay of samples from the incubation mixture using the one-stage clotting method. Interaction with vWf. Determination of the concentration of factor VIII able to bind to vWf was performed by using a BIAcore instrument (BIAcore Amersham, Uppsala, Sweden) (26, 27). The binding of factor VIII to surface immobilized vWf was continuously monitored. Three different dilutions of each sample were analyzed against a reference preparation of r-VIII SQ. The content of factor VIII able to bind to vWf was calculated by dividing the concentration of interacting factor VIII with the total protein concentration. In Vitro Test of Protein Stability in Extract of Porcine Tissue. After anestethizing the animal, skin samples were taken from its back or side immediately before it was sacrificed. A buffer of the following composition was made: 14.7 mM L-histidine, 0.4 M sodium chloride, 4 mM calcium chloride, 0.02% Tween 80, 13 mM tri-sodium-citrate, and 17 units of antithrombin/mL and pH was adjusted to 6.8 with 5 M hydrochloric acid. The buffer was added to the porcine tissue in a volume of 0.6 mL/g tissue. The mixture was incubated on ice for 2 h and stirred. The tissue was then removed. Different mPEG conjugates of factor VIII were incubated in the extract at 37 °C, and aliquots were taken at 0, 2, 4, 6, and 24 h and subsequently analyzed for coagulant activity by chromogenic assay. RESULTS

Prior to the mPEG modification experiments, the stability of factor VIII was assessed in the buffer compositions applied. It was found that the VIII:C did not decrease during 50 h in the buffers used, at room temperature (data not shown). Control experiments were also made to verify that factor VIII was stable in an

Bioconjugate Chem., Vol. 11, No. 3, 2000 391

Figure 3. Relation between excess of reagent and degree of modification during random coupling with mPEG-SAH. The lower values of degree of modification, below 5 mPEG/r-VIII SQ, are “pool 1” values reported in Table 2a.

Figure 4. Specific activity as a function of degree of modification; modification at random sites with mPEG(3000)SAH (0), mPEG(5000)SS (2), and mPEG(5000)TRES (O).

mPEG-containing environment. According to the results in Table 1 the activity was not hampered by the presence of mPEG. Random Coupling of Activated mPEG to Factor VIII. Three mPEG reagents with reactivity to lysines were chosen for coupling to r-VIII SQ (Figure 2). Previous control experiments performed by using refractometer showed that uncoupled mPEG was separated from the conjugate fractions by the type of chromatographic media used in the GPC step (data not shown). A protein yield of 60-80% was obtained in the chromatography step. First, by use of mPEG-SAH reagent, the degree of modification in relation to excess of mPEG added was studied. The degree of modification increased as the amount of excess mPEG added to the protein solutions increased, showing a correlation between these parameters (Figure 3). The specific activity of the conjugates derivatized above 15 mPEG/r-VIII SQ was, however, eliminated, and the proceeding work was performed by derivatizing r-VIII SQ with very low degrees of modification. Those latter conjugates were found to be heterogenic shown in Figures 6 and 7 by the electrophoresis and western blots (below). The GPC fractions of the first and second half of the protein peaks from these coupling reactions, respectively, were therefore analyzed separately (denoted pool 1 and pool 2 in Table 2a; conjugates 1-16, where the mean values of derivatization are reported). As a consequence of the fractionation by size,


Ro¨stin et al.

Figure 5. Specific activity as a function of degree of modification after acylation with mPEG(3000)SAH. Coupling was performed randomly (0) or during site protection (9).

Figure 7. (a) Western blot using polyclonal antibodies raised against the 80 kDa chain of r-VIII SQ. Lane 1, r-VIII SQ; lane 2, conjugate no. 17; lane 3, conjugate no. 22. (b) Western blot using polyclonal antibodies raised against the 90 kDa chain of r-VIII SQ. Lane 1, r-VIII SQ; lane 2, conjugate no. 17; lane 3, conjugate no. 22.

Figure 6. Reduced SDS-PAGE analyses of r-VIII SQ (HIC eluate) and mPEG modified r-VIII SQ. Lane 1, r-VIII SQ HIC eluate; lane 2, conjugate no. 17; lane 3, conjugate no. 22.

the amount of mPEG determined in pool 1 consistently exceeded the amount of mPEG in pool 2. Due to the low content of mPEG in the conjugate fractions, however, the responses in the NMR analysis were in the lower range of the standard curve. To evaluate the influence of the chromatographic procedure on protein activity, a control experiment, “blank”, was performed, subjecting the r-VIII SQ HIC-eluate to the buffers and chromatography steps used. The specific activity of the blank was essentially the same as the starting material (Table 2a). Coupling with mPEG(3000)SAH and mPEG(5000)SS, respectively, resulted in conjugates with a succinic linker (Figure 2, panels a and b). Table 2a (conjugates no 1-10) and Figure 4 show that the specific activity of factor VIII was reduced dramatically, even at low degrees of modi-

fication; approximately 1 mPEG/r-VIII SQ reduced the specific activity to 50%. The degree of activity retention was not influenced by the molecular weight of the mPEG. Coupling with the mPEG(5000)TRES reagent was earlier believed to give alkylation on amino groups (28) (Figure 2c), but later findings have shown that this reagent also reacts by acylation giving a sulfoacetamide linkage to the protein (29, 30) (Figure 2d). A higher specific activity was retained compared to modification by mPEG-SAH and mPEG-SS; at a mean degree of modification of approximately 3 mPEG/r-VIII SQ, 50% of the activity remained (Table 2a; conjugates 11-16 and Figure 4). Coupling of Activated mPEG to Factor VIII Immobilized on a Solid Support. By adsorbing the factor VIII to an anionic exchange resin during acylation with the mPEG(3000)SAH, lysines at sites interacting with the solid phase were protected from derivatization. A Q Sepharose FF gel was used as the solid support during the coupling reaction. At 4 mPEG/r-VIII SQ, 50% activity still remained after modification. As shown in Table 2b and Figure 5, the specific activity of the mPEG(3000)SAH-modified factor VIII was better retained using this solid-phase technique compared to the random acylation. The correlation between excess of mPEG added and the degree of modification obtained was however low. Characterization of the mPEG-Modified Factor VIII. Reduced SDS-PAGE performed on mPEG-modified factor VIII and a reference preparation of r-VIII SQ (HIC


Bioconjugate Chem., Vol. 11, No. 3, 2000 393

Table 3. Interaction of Factor VIII Conjugates with vWf

conjugate


content of protein able to bind to vWf (% of total protein content)

r-VIII SQ 17b 22b

15.0 7.0 5.3

98 ( 4a 43 26

a

Mean of 12 samples ( std. b Fractionated twice by GPC.

eluate) is shown in Figure 6. The bands of the reference r-VIII SQ appeared at 90 kDa (doublet) and 80 kDa (quartet), and trace amounts of the uncleaved single chain appeared at 170 kDa (lane 1). Electrophoretic bands of the conjugates were seen in positions corresponding to a molecular mass of 90 kDa and above (conjugates 17 and 22 in lanes 2 and 3, respectively). The conjugate preparations were prior to the SDS-PAGE analysis fractionated twice by GPC, and minor amounts of protein were seen in positions below 90 kDa, indicating a successful removal of unmodified protein. The blurred bands visualized the heterogeneity of the conjugates, and the separation of the protein bands was too poor for quantification of the degree of modification. To investigate if modification occurred selectively on any of the two factor VIII chains, western blot analyses were performed on conjugates 17 and 22, using polyclonal antibodies directed against the 80 and 90 kDa chains, respectively. In the blots against both 80 and 90 kDa chains (Figure 7), blurred protein bands appeared in positions corresponding to a molecular mass above that of the unmodified protein. It was therefore concluded that the modification sites were located on both factor VIII chains. A BIAcore analysis was used for studies on the binding of factor VIII to vWf. The results in Table 3 show that the content of conjugated factor VIII able to bind to vWf was reduced compared to the reference r-VIII SQ. Thus, the vWf interaction site was disturbed by the pegylation. The one-stage clotting assay was used to analyze the activation and the subsequent inactivation of factor VIII by thrombin. First, the activity response obtained by the one-stage clotting method was compared to the response obtained by the chromogenic assay for each sample. For the conjugates, a discrepancy between the results obtained by the two methods was found; the activities determined by the one-stage clotting method were significantly lower than the activities determined by the chromogenic assay. For the reference r-VIII SQ, however, the results obtained by the two methods coincided. Figure 8 shows the thrombin activation patterns at two levels of VIII:C as determined by the one-stage clotting assay. The conjugates and the reference r-VIII SQ of dilution 1 IU/mL, incubated with 1 NIH U/IU thrombin, were activated at essentially the same rate (Figure 8a). A higher fold of activation was, however, observed for the conjugates of the concentration 10 IU/mL and 10-1 NIH U/IU thrombin, compared to the reference r-VIII SQ (Figure 8b). During the decline phase, the activity of the conjugates decreased at a slower rate than the reference, and complete inactivation did not occur during the time of the analysis. This effect was more pronounced when the samples were monitored at the lower thrombin to factor VIII ratio. The amounts of conjugated protein exceeded the amount of reference r-VIII SQ by approximately 5- and 20-fold for conjugates 17 and 22, respectively. In Vitro Test of Protein Stability in Extract of Porcine Tissue. To investigate the influence of the

Figure 8. (a) Thrombin activation assay. The concentration of modified factor VIII conjugates no. 17 (b) and 22 (9) was 1 IU/mL and the ratio of thrombin to factor VIII was 1 NIH U/IU. r-VIII SQ (O) of concentration 1 IU/mL was assayed as reference. (b) Thrombin activation assay. The concentration of modified factor VIII conjugates no. 17 (b) and 22 (9) was 10 IU/mL and the ratio of thrombin to factor VIII was 10-1 NIH U/IU. r-VIII SQ (O) of concentration 10 IU/mL was assayed as reference.

pegylation on protein stability, an in vitro test was undertaken by incubating conjugates in extract of porcine tissue representative of the subcutaneous environment. In this crude test system, where the amount of factor VIII as well as the conjugates constituted a negligible amount of protein, the activity of the blank sample was markedly reduced during the first 6 h while the activity was preserved to a high extent in all incubation mixtures containing conjugates (Figure 9). In a further experiment, a preparation of a reference r-VIII SQ (HIC eluate) was incubated in the extract and the activity was reduced to the same extent as the blank (data not shown). The results show that the mPEG modification protected factor VIII from inactivation and that the coagulant activity was preserved significantly longer compared to the unmodified protein. DISCUSSION

A common problem associated with conjugation of proteins with mPEG is a disturbed biological function and decreased specific activity of the protein. It is not surprising that this problem has also appeared during modification of factor VIII, which is a protein with several biologically functional sites. Reaction products obtained by coupling with the mPEG-SAH and mPEG-SS reagents contain a succinic linker between the mPEG chain and the protein amino groups, the positive charges of which


Figure 9. Stability of r-VIII SQ; blank (9) and mPEG modified factor VIII; conjugate no. 1 (2), conjugate no. 28 (O), and conjugate no. 29 (0), in extract of porcine tissue.

are eliminated after the formation of an amide bond (5, 31). Until recently, coupling with mPEG-TRES was thought to give a stable secondary amine (28). Others (29, 30) have, however, showed that the mPEG can attach to the protein lysines through a sulfoacetamide linkage. This linker contains a sulfonic ester instead of a carboxylic ester as compared to the other conjugates. Also, by this coupling route, the lysine charge is eliminated. The specific activity of the conjugates obtained by random modification with all reagents was reduced, probably due to steric effects in combination with the changes of charge. A difference in reactivity between the reagents was observed. The mPEG-TRES reagent was less reactive than the other mPEG derivatives, and a larger excess was needed to modify the protein to the same extent (Table 2). Less reactivity often correlates to higher selectivity (32), which may account for the moderately higher specific activity of conjugates obtained with mPEG-TRES. With the aim of introducing the mPEG chains more selectively, factor VIII was immobilized on an anionic solid support during the coupling reaction. Reactive lysines, protected by the solid surface, were brought to be inaccessible, and the targets for mPEG conjugation were confined to other sites. As shown in Figure 5, the retained coagulant activity after directed modification by solid phase adsorption was higher than that retained after random modification using the mPEG(3000)SAH. Presumably, the acidic regions, important for activation of factor VIII, were protected from conjugation. All coupling reactions were performed at pH 7.8 due to reduced stability of factor VIII above pH 8 (data not shown). The pH used is considered low for lysine modification as the -amino groups of the lysines are mainly protonated at this pH [the pKa of lysine is in free form 10.8 and in a model peptide 10.0-10.2 (33)]. Relatively large molar excess of mPEG was therefore added (Table 2). Heterogeneity in respect of the degree of modification is a well-known problem when modifying proteins chemically (34). r-VIII SQ contains 79 lysines to which activated mPEG may react. At high degrees of modification (>15 mPEG/r-VIII SQ), a large fraction of reactive lysines located on the protein surface was presumably modified. The degree of modification correlated to the amount of excess reagent added to the protein solutions. At lower degrees of modification, however, only a small fraction of the available lysines was derivatized. Although there was a correlation between amount of reagent added and the degree of modification, these coupling reactions were harder to control. Furthermore, the amounts of mPEG were in the lower range of the standard curve of the NMR

Ro¨stin et al.

analysis. By introducing a solid phase, during derivatization of r-VIII SQ in the immobilized state, the correlation was substantially reduced as more factors affected the coupling, e.g., the diffusion into the solid-phase particles and the masking of sites on the protein by adsorption. A considerable heterogeneity regarding molecular mass of the low derivatized conjugates was seen. This was revealed by the electrophoresis and western blots and by the fact that the mPEG content in two pools of gel permeation chromatography fractions obtained from one reaction mixture clearly differed. Separation techniques of higher resolution are necessary to further fractionate the different conjugates. The human half-life of factor VIII is 10-15 h for the full-length as well as the B-domain-deleted recombinant variant in the presence of vWf (35). In the absence of vWf, the clearance rate is substantially higher and half-lives of 1-2 h have been reported (14, 16). Therefore, the interaction of the modified factor VIII with vWf was considered a key issue, when characterizing the conjugates in the present study. According to the results obtained by the BIAcore assay, the interaction was hampered by the mPEG conjugation. This could of course influence the in vivo function and should be studied further. Proteolytic cleavage by thrombin activates and subsequently inactivates factor VIII. The activation and inactivation of pegylated factor VIII diluted to 1 IU/mL showed essentially the same kinetics as obtained with the reference r-VIII SQ. The activation of the pegylated factor VIII diluted to 10 IU/mL and incubated with thrombin at 10-1 NIHU/IU factor VIII also showed a qualitatively similar pattern during activation as the reference. Inactivation of the conjugates, at this concentration and thrombin to factor VIII ratio, was however slower. The discrepancies between the two methods of determining VIII:C made the interpretation of the results of this study difficult. The one stage clotting method gave a lower response for modified factor VIII per mole of protein compared with the chromogenic assay. Consequently, a higher protein molar concentration of the modified molecule compared with the reference r-VIII SQ was required for a given activity response. Accordingly, the higher activation fold and the slower decline of activity could be an effect of lower thrombin to modified protein ratio than in the case of the reference protein. The protecting effect of the mPEG often depends on the degree of modification (36). A surprisingly good retention of coagulant activity of the conjugates was achieved at incubation in the tissue extract even at these low degrees of modification. Since coagulant activity was maintained for 24 h in the incubation mixtures containing conjugates obtained from both random and directed modification with the mPEG(3000)SAH, the procedure of directing the mPEG coupling from the acid regions did not seem to influence the conjugates to withstand the process of inactivation in this particular environment. Neither was any correlation between the degree of modification and preservation of activity found, since the conjugate with 2.5 mPEG/r-VIII SQ was equally active as compared to the conjugates with higher degrees of modification, after 24 h. CONCLUSION

In conclusion, r-VIII SQ was chemically modified by covalent attachment of mPEG with a concomitant reduction of coagulant activity. By immobilizing the factor VIII onto a solid matrix, at least one functional site was


shielded from modification, and the specific activity was retained to a higher extent as compared with random modification after acylation. The support chosen was an anionic-exchange chromatography gel, normally used for protein purification. In the field of immobilization techniques and in the context of pegylation, protection of functional sites by immobilization of the target protein onto a specific affinity support has previously been described (7). From the standpoint of the pharmaceutical industry, however, adsorbents with ligands manufactured by chemical synthesis is preferable. The use of complex biological substances in the manufacturing process is hereby increasingly avoided. This is important both due to the safety of the patients, which is a major concern, and the regulatory documentation which is a matter of significant cost. Despite the low degrees of modification applied, for the purpose of a satisfactory retention of coagulant activity, the mPEG markedly protected the factor VIII from inactivation in an in vitro environment representative of the subcutaneous space. In the pharmacokinetic study performed in cynomolgus monkeys (20), conjugates 17 and 22 were administered subcutaneously. The bioavailability was equal for both conjugates and exceeded that of the unmodified protein by approximately 4-fold. This implies a high potential of the modification technique as applied on this protein. ACKNOWLEDGMENT

We thank A° ke Ljungquist and collaborators for providing the r-VIII SQ HIC eluate. Ulla Oswaldsson and collaborators are acknowledged for skillful analyses of factor VIII activity. Helena Sandberg, Annelie Almstedt, and collaborators are acknowledged for contribution to the biochemical part of the work. We also thank SvenOlof Larsson for performing the NMR analyses and Kristina Zachrisson for performing the amino acid analyses. Prof. Karl Hult is acknowledged for valuable discussions and advice on the manuscript. LITERATURE CITED (1) Nilsson, I. M. (1994) Replacement therapy. Hemophilia, Pharmacia Plasma Products, Sweden. (2) Davis, F. F., Kazo, G. M., Nucci, M. L., and Abuchowski, A. (1991) Reduction of immunogenicity and extension of circulating half-life of peptides and proteins. In Peptide and Protein Drug Delivery (V. H. L. Lee, Ed.) pp 831-864, Marcel Dekker, Inc., New York. (3) Katre, N. (1993) The conjugation of proteins with poly(ethylene glycol) and other polymers. Adv. Drug Deliv. Rev. 10, 91-114. (4) Inada, Y., Furukawa, M., Sasaki, H., Kodera, Y., Hiroto, M., Nishimura, H., and Matsushima, A. (1995) Biomedical and biotechnological applications of PEG- and PM-modified proteins. Trends Biotechnol. 13, 86-91. (5) Zalipsky, Z. (1995) Functionalized poly(ethylene glycol) for preparation of biologically relevant conjugates. Bioconjugate Chem. 6, 150-165. (6) Goodson, R., and Katre, N. (1990) Site-directed pegylation of recombinant interleukin-2 at its glycosylation site. Bio/ Technology 8, 343-346. (7) Caliceti, P., Schiavon, O., Sartore, L., Monfardini, C., and Veronese, F. (1993) Active site protection of proteolytic enzymes by poly(ethylene glycol) surface modification. J. Bioact. Compat. Polym. 8, 41-50. (8) Vehar, G. A., Keyt, B., Eaton, D., Rodriguez, H., O’Brien D. P., Rotblat, F., Oppermann, H., Keck, R., Wood, W. I., Harkins, R. N., Tuddenham E. G. D., Lawn, R. M., and Capon, D. J. (1984) Structure of human factor VIII. Nature 312, 337-342. (9) Andersson, L.-O., Forsman, N., Huang, K., Larsen, K., Lundin, A., Pavlu, B., Sandberg, H., Sewerin, K., and Smart,

Bioconjugate Chem., Vol. 11, No. 3, 2000 395 J. (1986) Isolation and characterization of human factor VIII: Molecular forms in commercial factor VIII concentrate, cryoprecipitate, and plasma. Proc. Natl. Acad. Sci. U.S.A. 83, 2979-2983. (10) Eaton, D., Rodriguez, H., and Vehar, G. A. (1986) Proteolytic processing of human factor VIII. Correlation of specific cleavages by thrombin, factor Xa, and activated protein C with activation and inactivation of factor VIII coagulant activity. Biochemistry 25, 505-512. (11) Pittman, D., and Kaufman, R. (1988) Proteolytic requirements for thrombin activation of anti-hemophilic factor (factor VIII). Proc. Natl. Acad. Sci. U.S.A. 85, 2429-2433. (12) Fay, P. J., Smudzin, T. M., and Walker, F. J. (1991) Activated protein C-catalyzed inactivation of human factor VIII and factor VIIIa. J. Biol. Chem. 266, 20139-20145. (13) Lollar, P. (1991) The association of factor VIII with von Willebrand factor. Mayo Clin. Proc. 66, 524-534. (14) Morfini, M., Mannucci, P. M., Tenconi, P. M., Longo, G., Mazzucconi, M. G., Rodeghiero, F., Ciavarella, N., De Rosa, V., and Arter A. (1993) Pharmacokinetics of monoclonally purified and recombinant factor VIII in patients with severe von Willebrand disease. Thromb. Haemost. 70, 270-272. (15) Fijnvandraat, K., Peters, M., and ten Cate, J. W. (1995) Inter-individual variation in half-life of infused recombinant factor VIII is related to pre-infusion von Willebrand factor antigen levels. Br. J. Haematol. 91, 474-476. (16) Tuddenham, E., Lane, R., Rotblat, F., Johanson, A., Snape, T., Middleton, S., and Kernoff, P. (1982) Response to infusions of polyelectrolyte fractionated human factor VIII concentrate in human haemophilia A and von Willebrand’s disease. Br. J. Haematol. 52, 259-267. (17) Lollar, P., Hill-Eubanks, D. C., and Parker, C. G. (1988) Association of the factor VIII light chain with von Willebrand factor. J. Biol. Chem. 263, 10451-10455. (18) Pittman, D., Alderman, E., Tomkinson, K., Wang, J., Giles, A., and Kaufaman, R. (1993) Biochemical, immunological, and in vivo functional characterization of B-domain-deleted factor VIII. Blood 81, 2925-2935. (19) Lind, P., Larsson, K., Spira, J., Sydow-Ba¨ckman, M., Almstedt, A., Gray, E., and Sandberg, H. (1995) Novel forms of B-domain-deleted recombinant factor VIII molecules: Construction and biochemical characterization. Eur. J. Biochem. 232, 19-27. (20) Ro¨stin, J., Smeds, A.-L., and Mikaelsson, M. (2000) Bioavailability of B-domain deleted recombinant factor VIII conjugated with monomethoxy poly(ethylene glycol) (manuscript submitted for publication). (21) Smeds, A.-L., Ljungqvist, A., Ostlin, A., Ljungquist, C., Gaal, A., Sandberg, H., Rosenquist, J., and Mikaelsson, M. (1995) The purification process for recombinant factor VIII SQ. Thromb. Haemost. 73 (abstract no. 442). (22) Horowitz, B., Wiebe, M. E., Lippin, A., and Stryker, M. H. (1985) Inactivation of viruses in labile blood derivatives. I. Disruption of lipid-enveloped viruses by tri (n-butyl)-phosphate detergent combinations. Transfusion 25, 516-522. (23) Laemmli, U. K. (1970) Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature 227, 680-685. (24) Rose´n, S. (1984) Assay of factor VIII:C with a chromogenic substrate. Scand. J. Haematol. 33 (Suppl. 40), 139-145. (25) Mikaelsson, M., and Oswaldsson, U. (1984) Standardization of VIII:C assays: a manufacturers view. Scand. J. Haematol. 33 (Suppl. 41), 79-86. (26) Jo¨nsson, U., and Malmqvist, M. (1992) Real Time Biospecific Interaction Analysis. The Integration of Surface Plasmon Resonance Detection, General Biospecific Interface Chemistry and Microfluidics into One Analytical System. Adv. Biosensors 2, 291-336. (27) Karlsson, R., Fa¨gerstam, L., Nilshans, H., and Persson, B. (1993) Analysis of active antibody concentration. Separation of affinity and concentration parameters. J. Immunol. Methods 166, 75-84. (28) Nilsson, K., and Mosbach, K. (1984) Immobilization of ligands with organic sulfonyl chlorides. Methods Enzymol. 104, 56-69.

396 Bioconjugate Chem., Vol. 11, No. 3, 2000 (29) Gais., H.-J., and Ruppert, S. (1995) Modification and immobilization of proteins with poly(ethylene glycol) tresylates and polysaccharide tresylates: Evidence suggesting a revision of the coupling mechanism and the structure of the polymer-polymer linkage. Tetrahedron Lett. 36, 3837-3838. (30) King, J., and Gill., M. (1996) Alkyl 2,2,2-trifluoroethanesulfonates (tresylates): Elimination-addition vs bimolecular nucleophilic substitution in reactions with nucleophiles in aqueous media. J. Org. Chem. 61, 7250-7255. (31) Wie, S. I., Wie, C. W., Lee, W. Y., Filion, L. G., Sehon, A. H., and Åkerblom, E. (1981) Suppression of reagenic antibodies with modified allergens. Int. Arch. Allergy Immunol. 64, 84-99. (32) Zalipsky, Z., Seltzer, R., and Menon-Rudolph, S. (1992) Evaluation of a new reagent for covalent attachment of poly(ethylene glycol) to proteins. Biotechnol. Appl. Biochem. 15, 100-114.

Ro¨stin et al. (33) Wong, S. S. (1991) Chemistry of protein conjugation and cross-linking, p 15, CRC Press, Inc., Boca Raton, FL. (34) Kunitani, M., Dollinger, G., Johnson, D., and Kresin, L. (1991) On-line characterization of poly(ethylene glycol)-modified proteins. J. Chromatogr. 588, 125-137. (35) Fijnvandraat, K., Berntorp, E., ten Cate, J. W., Johnsson, H., Peters, M., Savidge, G., Tengborn, L., Spira, J., and Stahl, C. (1997) Recombinant, B-domain deleted factor VIII (r-VIII SQ): Pharmacokinetics and initial safety aspects in hemophilia A patients. Thromb. Haemost. 77, 298-302. (36) Hooftman, G., Herman, S., and Schacht, E. (1996) Review: Poly(ethylene glycol)s with reactive endgroups. II. Practical consideration for the preparation of protein-PEG conjugates. J. Bioact. Compat. Polym. 11, 135-159.

BC990137I

B-Domain Deleted Recombinant Coagulation Factor VIII Modified with

Recommend Documents