Bioconjugate Chem. 1990, 7, 274-277
274
Antibody-Directed Fibrinolysis: A Bispecific (Fab’)z That Binds to Fibrin and Tissue Plasminogen Activator Marschall S. Runge,+ Christoph Bode,l Christopher E. Savard, Gary R. Matsueda, and Edgar Haber’ Cardiac Unit, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts 02114. Received December 22, 1989
A bispecific (Fab’)z molecule was constructed by linking the monovalent Fab‘ from an anti-fibrin monoclonal antibody to the Fab’ from an anti tissue plasminogen activator (tPA, single chain) monoclonal antibody by means of inter-heavy-chain disulfide bonds. An immunochemical complex composed of the bispecific (Fab’)z molecule bound to tPA [tPA-bispecific (Fab’):! complex] was then generated and purified. Its molecular weight was 170 kDa [less than half the molecular weight of a previously described tPA-bispecific antibody complex containing the entire anti-fibrin and anti-tPA immunoglobulinmolecules; Runge, M. S., et al. (1987) Trans. Assoc. Am. Phys. 100,250-2551. The tPA-bispecific (Fab’)z complex was 8.6-fold more efficient in fibrinolysis than tPA alone and 94-fold more potent than urokinase. This enhancement in the fibrinolytic potency of tPA compares favorably with that observed for the bispecific whole-antibody complex. These results suggest that this smaller, less immunogenic molecule is capable of binding both fibrin and tPA with high affinity and of enhancing the thrombolytic efficiency of exogenous and, perhaps, endogenous tPA.
INTRODUCTION Treating myocardial infarction with plasminogen activators has been shown to result in a reduction in infarct size ( I , 2) and mortality ( 3 , 4 ) . For optimal results, treatment must be initiated as soon as possible after the onset of symptoms (3). This mandates a rapid therapeutic decision and heightens the need for a thrombolytic agent that is free of serious side effects. A bispecific antibody capable of binding both fibrin and single-chain tissue plasminogen activator (tPA)l has been described (5-7). We assembled it by forming a disulfide bond between N-succinimidyl 3-(2-pyridyldithio)propionate modified anti-fibrin monoclonal antibody 59D8 and 2-iminothiolane-modified anti-tPA monoclonal antibody TCL8. The bispecific antibody enhances the fibrinolytic potency of tPA in vitro and in vivo. Here we describe the production and characterization of a molecule assembled by linking an Fab’ fragment from each of the two antibodies. The bispecific (Fab’):! similarly enhances the fibrinolytic potency of tPA, but unlike the bispecific whole antibody, the bispecific (Fab’)z has a well-defined s t r u c t u r e , is much smaller, a n d lacks t h e highly immunogenic Fc portions of its component antibodies. EXPERIMENTAL PROCEDURES The tPA used in these studies was derived from melanoma cells purchased from BioResponse (Hayward, CA). Two-chain, low molecular weight urokinase (Abbokinase) was purchased from Abbott Laboratories and Sepharose 4B-CL was obtained from Pharmacia P-L Biochemicals. The lZ5I-labeled fibrinogen came from Am+ Present address: Cardiology Division, Emory University,Post Office Drawer LL, Atlanta, GA 30322. Present address: Medizinische Klinik I11 (Kardiologie)der Universitat Heidelberg, 6900 Heidelberg, Federal Republic of Germany. 1 Abbreviations: tPA, tissue plasminogen activator (single chain); S-2288, chromogenic substrate H-D-isoleucyl-L-prolyl-L-
arginine-p-nitroanilide dihydrochloride.
ersham; the plasma came from the local blood bank. Chromogenic substrate H-D-isoleucyl-L-prolyl-L-arginine-pnitroanilide dihydrochloride (S-2288) was obtained from Helena Laboratories. Human placenta factor XI11 was purchased from Green Cross (Osaka, Japan); the Superose 12 resin for fast protein-liquid chromatography was from Pharmacia. All other chemicals came from Sigma. Antibodies. tPA-specific monoclonal antibody TCL8 and fibrin-specific monoclonal antibody 59D8 were raised and purified as described (5,B). Protein Concentration. Protein concentration was determined by the method of Bradford (9). Preparation of tPA-Sepharose. Twenty milligrams of recombinant tPA was solubilized in 10 mL of water and passed through a Sephadex G-25 column (30 X 2 cm) equilibrated with 0.2 M NaHCOs, 1.5 M sodium chloride, and 0.1 73 Tween-80 (coupling buffer). The protein was immediately incubated with 15 mL of cyanogen bromide activated Sepharose (Pharmacia P-L Biochemicals) and coupled according to the manufacturer’s instructions. After incubation for 24 h a t 4 OC, residual binding sites on the Sepharose were blocked with 10 mL of 1.0 M ethanolamine (titrated with HCl to pH 8.0) for an additional 8 h. Production and Purification of Bispecific (Fab’)n. The bispecific (Fab’):! molecule was prepared by linking the monovalent Fab’s of antibodies 59D8 and TCL8 through the intrinsic heavy-chain cysteines, which usually form disulfide bonds in native (Fab’)~molecules (IO). Each antibody was first digested with pepsin (11): 10 mL of antibody solution (2.0 mg/mL in 0.01 M sodium phosphate, 0.15 M sodium chloride, pH 7.4) was mixed with 1mL of 1.0 M sodium citrate (pH 2.75) and 1mL of pepsin solution (0.3 mg/mL in water), the final pH was adjusted to 3.5, and the mixture was incubated for 2 h a t 37 “C. The reactions were stopped by the addition of 1.0 mL of 3.0 M Tris HCl, pH 8.5. In pilot experiments, these conditions had produced optimal yields of 59D8 Fab‘ and TCL8 Fab’. Pepsin concentrations had varied between 0.05 and 0.5 mg/ mL, the final pH was between 2.5 and 5.0, and the duration of incubation was between 30 min and 8 h.
1043-1802/90/2901 -0274$02.50/0 0 1990 American Chemical Society
Antibody-Directed Fibrinolysis
(Fab’)2 were purified from the 59D8 and TCL8 digests by chromatography against their respective affinity ligands. For the former, 0peptide (Gly-His-Arg-Pro-Leu-Asp-LysCys, corresponding to the seven amino-terminal residues of the 0chain of fibrin) coupled to Sepharose (P peptideSepharose) was used; for the latter, tPA coupled to Sepharose (see Preparation of tPA-Sepharose) was used. After elution from the affinity matrices with 0.2 M glycine (pH 2.8), each (Fab’)~solution was collected into tubes containing a neutralizing amount of 3.0 M Tris HC1 (pH 8.5), dialyzed into 0.1 M sodium phosphate (pH 6.8), and reconcentrated in an ultrafiltration chamber to 2 mg/ mL. Reduction of the 59D8 and TCL8 fragments was then carried out a t room temperature in 1 mM 2-mercaptoethylamine, 1 mM ethylenediaminetetraacetic acid, and 10 mM sodium arsenite, followed by the addition of solid Ellman’s reagent to a concentration of 5 mM. After 3 h a t room temperature, excess reagent was removed from the two Fab’ solutions by gel filtration on a Sephadex G25 column (30 X 2 cm) equilibrated with 0.1 M sodium phosphate (pH 6.8). The thiol form of anti-fibrin 59D8 Fab’ was then regenerated by treatment with 10 mM 2mercaptoethylamine for 30 min, followed by gel filtration as above. After t h a t the anti-fibrin 59D8 Fab‘ was incubated with the thionitrobenzoate derivative of the antitPA TCL8 Fab’ for 16 h at room temperature in 0.1 M sodium phosphate and 1 mM ethylenediaminetetraacetic acid (pH 6.8). The desired (Fab’)~heterodimer was purified from the reaction mixture by sequential affinity chromatography on peptide-Sepharose and tPASepharose. The final yield of bispecific (Fab’)~was 7 mg, or about 25% of the theoretical yield of the coupling of two different Fab’ molecules. Preparation of Immunochemical Complex. An immunochemical complex made up of tPA and the bispecific (Fab’)~[tPA-bispecific (Fab’)2 complex] was formed by mixing 3.5 mg of tPA (0.5 mg/mL) with 5 mg of bispecific (Fab’)~(0.5 mg/mL) for 2 h a t room temperature. After concentration to a volume of 9 mL, chromatography on Sephacryl S-300 [preequilibriated with 0.1 M sodium phosphate, 0.1 M NaCl, 0.01% sodium azide, pH 7.4 (PBSA)] revealed a peak of approximately 170 kDa [tPA bound to the bispecific (Fab’)~]and a second peak of approximately 70 kDa (unbound tPA). On the basis of enzymatic activity (assayed by S-2288, see below), approximately 1 mol of tPA appeared to bind per mole of bispecific (Fab’h. Measurement of Plasminogen Activator Activity. To compare the activity of native tPA with that of tPA as part of an immunochemical complex, the amidolytic activities and molar amounts of tPA in the various preparations were measured in the following manner. Standardized samples of melanoma tPA or recombinant tPA (as aliquots from a freshly resuspended vial of tPA, in international units) were analyzed in the S-2288 assay with a substrate concentration of 1 X mol/L and an enzyme concentration of 8 X mol/L in 0.15 M Tris, 0.15 M NaCl, pH 8.4. For tPA, 1 international unit was assumed to equal 6.3 X 10-6 nmol. The correlation between the reported change in absorbance/minute for the assay preparations2 and our samples was excellent, such that 100 units of tPA (6.3 X 104 mmol) gave an absorbance change a t 405 nm of approximately O.O6O/min. On the basis of these results, the activity (in appropriate units as above) or the molar amount of active enzyme of an unknown sample of urokinase or tPA was determined by diluting the sample until assay with S-2288 as described above 2
KabiVitrum, 1986.
Bioconjugate Chem., Vol. 1, No. 4, 1990
600
401
/“
275
276
-
Bioconjugate Chem., Vol. 1, No. 4, 1990 1
kDa 94--01.*
2
3
1
2
Runge et al. 3
4
5
-
30--r,
20-
14.4--341O
>j : 1 I6
$
A
B
Figure 2. SDS-PAGE of tPA-bispecific (Fab’)~complex. Ten percent polyacrylamide gels were run under nonreducing (panel A) and reducing (panel B) conditions. (A) numbers on the left correspond to molecular weight standards (Pharmacia): lane 1, tPA-bispecific (Fab’)~complex; lane 2, single-chainrecombinant tPA; and lane 3, anti-tPA antibody. (B) Lanes 1-3 as in A; lane 4, antifibrin 59D8 (Fab’)2; Lane 5, anti-tPA TCL8 (Fab’)2.
PLASM/NOG€N AC77VATOR funits)
Figure 3. Release of radioactive peptides from 12%labeled fibrinSepharose by the tPA-bispecific (Fab’)2 complex (filled circle), by tPA alone (open triangle), and by urokinase alone (filled square). Lysis is expressed as the quotient of released radioactivity over total radioactivity. Each point represents the mean of three determinations. A three-parameter inverse logit function was used to fit the curves to the original triplicate points. the other to the size of a tPA molecule. Under reducing conditions (panel B), bands corresponding in size to the two chains of the Fab’ molecules and to single-chain t P A could be seen, and there was evidence of traces of twochain tPA. The tPA-bispecific (Fab’)2 complex was compared with tPA alone and with urokinase in the quantitative fibrinolysis assay (Figure 3). In these experiments, the tPAbispecific (Fab’):! complex was 8.6-fold more efficient in fibrinolysis than tPA alone ( p < 0.0001), and 94-fold more potent than urokinase (p < 0.0001). Here tPA was 11fold more efficient than urokinase, a finding similar to that of previous experiments (5). It is of particular interest that t h e tPA-bispecific (Fab’)2 complex showed a n enhancement in the fibrinolytic activity of tPA similar to that of the tPA-bispecific antibody complex (5). DISCUSSION Here we describe a bispecific (Fab’)z that has affinity for both fibrin and tPA. The rationale for producing this bispecific (Fab’)2 was based on promising results obtained when intact anti-fibrin and anti-tPA antibodies were coupled. We previously demonstrated (5-7) that a bispecific antibody containing intact anti-fibrin and antitPA antibodies is capable of concentrating t P A and enhancing the fibrinolytic potency of tPA in vitro, in human plasma, and in vivo in the rabbit jugular vein model. In model experiments, the bispecific antibody was able to
concentrate the low amounts of t P A normally present in human plasma and effect fibrinolysis. This increased potency is probably related to the fact that the antifibrin antibody has an affinity for fibrin 1800 times greater than that of tPA: the K D of tPA for fibrin is 0.14 X lo+ M (15),whereas that of antibody 59D8 is 0.77 X 10-lo M.3 However, one limitation to the clinical utility of the bispecific whole antibody is that the coupling chemistry used to create it does not yield a homogeneous product. The bispecific (Fab’)2 is a better-defined molecule than the bispecific whole antibody because the linkage between the component Fab’s in the fragment antibody occurs only at the disulfide bonds of the hinge region, with the practical result that only (Fab’)2 molecules are present after the coupling reaction. T o ensure t h a t only molecules possessing both fibrin and tPA binding properties (in a 1:l molar ratio) were present in the final reaction mixture, we subjected the products of the coupling reaction to sequential affinity chromatography against the two antigens. The gel-filtration data presented in Figure 2 suggest that tPA binds to the bispecific (Fab’):! in a 1:l complex. Our calculations indicate that almost all the bispecific (Fab’)2 reacted with the tPA. This suggests that the purification process selects for functionally active molecules. The ability of the bispecific (Fab’)2 to enhance the fibrinolytic potency of tPA is apparently identical with that of the bispecific whole antibody. Thus monovalent binding to both antigens appears to be sufficient to effect increased potency. Because of its increased diffusibility, the smaller bispecific (Fab’)2, as opposed to the bispecific whole antibody, may provide advantages for later in vivo studies. More importantly, t h e (Fab’)2 can be synthesized in a reproducible manner that does not require the introduction of cross-linking reagents, which have the potential to act as immunogenic epitopes. ACKNOWLEDGMENT M.S.R. is a recipient of a Merck Fellowship from the American College of Cardiology. C.B. is a recipient of a fellowship from Boehringer Ingelheim Fonds and a grant from the Sandoz Foundation. This work was supported by National Institutes of Health Grants HL-19259 and HL28015 and by a grant from the Schering Corp. LITERATURE CITED (1) Collen,D., Topol, E. J., Tiefenbrunn, A. J., Gold, H. K., Weisfeldt, M. L., Sobel, B. E., Leinbach, R. C., Brinker, J. A., Ludbrook, P. A., Yasuda, I., Bulkley, B. H., Robison, A. K., Hutter, A. M., Jr., Bell, W. R., Spadaro, J. J., Jr., Khaw, B. A., and Grossbard, E. B. (1984) Coronary thrombolysis with recombinant human tissue-type plasminogen activator: a prospective, randomized, placebo-controlled trial. Circulation 70,1012-1017. (2) Simoons, M . L., Serruys, P. W., van de Brand, M., Bir, F., de Zwaan, C., Res, J., Verheugt, F. W. A., Krauss, X. H., Remme, W. J., Vermeer, F., and Lubsen, J. (1985) Improved survival after early thrombolysis in acute myocardial infraction. A randomized trial by the Interuniversity Cardiology Institute in The Netherlands. Lancet 2, 578-581. (3) Gruppo Italian0 per lo Studio della Streptochinasinell’Infarto miocardico (GISSI) (1986)Effectivenessof intravenous thrombolytic treatment in acute myocardial infarction. Lancet I, 397-402. (4) The ISAM Study Group (1986) A prospective trial of intravenous streptokinase in acute myocardial infarction (ISAM). N. Engl. J. Med. 314,1465-1471. 3
Rohdan J. Kudryk, personal communication.
Antibody-Directed Fibrinolysis
(5)Bode, C., Runge, M. S., Branscomb, E. E., Newell, J. B., Matsueda, G. R., and Haber, E. (1989)Antibody-directed fibrinolysis: An antibody specific for both fibrin and tissue plasminogen activator. J. Biol. Chem. 264,944-948. (6)Runge, M. S.,Bode, C.,Matsueda, G. R., and Haber E. (1987) Antibody-enhanced thrombolysis: capture of tissue plasminogen activator by a bispecific antibody and direct targeting by an antifibrin-tPA conjugate _ _ in vivo. Trans. Assoc. Am. Phys. 100, 250-255. (7)Runge, M. S.,Quertermous, T., Matsueda, G. R., and Haber, E.(1988)Increasing selectivity of plasminogen activators with antibodies. Clin. Res. 56,501-506. (8) Hui, K. Y., Haber, E., and Matsueda, G. R. (1983)Monoclonal antibodies to a synthetic fibrin-like peptide bind to human fibrin but not fibrinogen. Science 222,1129-1132. (9)Bradford, M. M. (1976)A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72,248254. (10) Brennan, M., Davison, P. F., and Paulus, H. (1985) Preparation of bispecific antibodies by chemical recombination
Bioconjugate Chem., Vol. 1, No. 4, 1990 277
of monoclonal immunoglobulin GI fragments. Science 229, 81-83. (11)Parham, P. (1983)On the fragmentation of monoclonal IgG1, IgG%,and IgGZb, from BALB/c mice. J.Zmmunol. 131,28952902. (12) Bode, C.,Matsueda, G. R., Hui, K. Y., and Haber, E. (1985) Antibody-directed urokinase: a specific fibrinolytic agent. Science 229,765-767. (13)BBN Research Systems (1983)R S / l User's Guide, book 2, release 2,serial V-14342,pp 180-187,Bolt Beranek and Newmann, Cambridge, MA. (14) Bode, C., Runge, M. S., Newell, J. B., Matsueda, G. R., and Haber, E. (1987)Thrombolysis by a fibrin-specific antibody Fab'-urokinase conjugate. J. Mol. Cell. Cardiol. 19,335341. (15)Hoylaerts, M., Rijken, D. C., Lijnen, H. R., and Collen, D. (1982)Kinetics of the activation of plasminogen by human tissue plasminogen activator. Role of fibrin. J. Biol. Chem. 257,2912-2919. Registry No. Plasminogen activator, 105913-11-9.
