Portal de Periódicos da CAPES

Recombinant human microplasmin: production and potential therapeutic properties

2003; Elsevier BV; Volume: 1; Issue: 2 Linguagem: Inglês

10.1046/j.1538-7836.2003.00078.x

1538-7933

N. Nagai, E Demarsin, B. Van Hoef, S. Wouters, Doriano Cingolani, Yves Laroche, Désiré Collen,

Intracranial Aneurysms: Treatment and Complications

SummaryThe effect of recombinant human microplasmin was studied in ischemic stroke models in mice and in an extracorporeal loop thrombosis model in rabbits. Human microplasminogen (µPlg), which lacks the five ‘kringle’ domains of plasminogen was expressed with high yield in Pichia pastoris. It was purified, converted to microplasmin (µPli) and equilibrated with 5 mmol L−1 citrate, pH 3.1, yielding a stable preparation. In mice with middle cerebral artery (MCA) ligation, an intravenous (i.v.) bolus of 5.0 mg kg−1µPli reduced infarct size at 24 h from 27 (26–30) to 25 (21–28) mm3 (median and range, n= 16 each, P= 0.0001), whereas 4.0 mg kg−1 rt-PA and 40 mg kg−1µPlg had no effect. Infarct reduction was observed with administration at 4 h after occlusion. In mice with MCA, infarct size at 24 h was reduced from 20 (14–30) to 9.1 (3.1–25) mm3 with 5.0 mg kg−1µPli (n = 15 each, P < 0.002) and to 11 (5.2–27) mm3 with 4.0 mg kg−1 rt-PA (n = 6; P= 0.02). Infarct reduction was still observed at 10 h after occlusion with µPli but not with t-PA. In rabbits with radiolabeled clots in an extracorporeal arteriovenous loop, local infusion of 2.5 mg kg−1µPli over 2 h, induced 51 ± 15% lysis (mean ± SD, n= 11) vs. a control value of 23 ± 5.5%. µPli did not prolong template bleeding times, whereas equipotent doses of rt-PA were associated with extensive rebleeding. The potency of µPli in both models was similar to that of intact plasmin. These findings indicate that recombinant µPli may be useful for treatment of ischemic stroke and arterial thrombosis. The effect of recombinant human microplasmin was studied in ischemic stroke models in mice and in an extracorporeal loop thrombosis model in rabbits. Human microplasminogen (µPlg), which lacks the five ‘kringle’ domains of plasminogen was expressed with high yield in Pichia pastoris. It was purified, converted to microplasmin (µPli) and equilibrated with 5 mmol L−1 citrate, pH 3.1, yielding a stable preparation. In mice with middle cerebral artery (MCA) ligation, an intravenous (i.v.) bolus of 5.0 mg kg−1µPli reduced infarct size at 24 h from 27 (26–30) to 25 (21–28) mm3 (median and range, n= 16 each, P= 0.0001), whereas 4.0 mg kg−1 rt-PA and 40 mg kg−1µPlg had no effect. Infarct reduction was observed with administration at 4 h after occlusion. In mice with MCA, infarct size at 24 h was reduced from 20 (14–30) to 9.1 (3.1–25) mm3 with 5.0 mg kg−1µPli (n = 15 each, P < 0.002) and to 11 (5.2–27) mm3 with 4.0 mg kg−1 rt-PA (n = 6; P= 0.02). Infarct reduction was still observed at 10 h after occlusion with µPli but not with t-PA. In rabbits with radiolabeled clots in an extracorporeal arteriovenous loop, local infusion of 2.5 mg kg−1µPli over 2 h, induced 51 ± 15% lysis (mean ± SD, n= 11) vs. a control value of 23 ± 5.5%. µPli did not prolong template bleeding times, whereas equipotent doses of rt-PA were associated with extensive rebleeding. The potency of µPli in both models was similar to that of intact plasmin. These findings indicate that recombinant µPli may be useful for treatment of ischemic stroke and arterial thrombosis. Intravenous plasmin for thrombolytic therapy has been investigated in humans since the 1950s [1Ambrus J.L. Ambrus C.M. Bock N. et al.Clinical and experimental studies on fibrinolytic enzymes.Ann New York Acad Sci. 1957; 68: 97Crossref PubMed Scopus (42) Google Scholar, 2Clifton EE. The use of plasmin in humans.Ann New York Acad Sci. 1957; 68: 209-29Crossref PubMed Scopus (35) Google Scholar, 3Jensen VJ (inventor). US patent 3,950,513. Process of stabilizing therapeutically useful plasmin solutions. Novo Terapeutisk Laboratorium (assignee). Issued 13 April, 1976.Google Scholar]. Administration of a large amount of plasmin, unlike certain other proteolytic enzymes, appeared to be well tolerated, but the development of plasmin for treatment was discontinued in the 1970s. Nagai et al. [4Nagai N. De Mol M. Van Hoef B. Verstreken M. Collen D. Depletion of circulating α2-antiplasmin by intravenous plasmin or immunoneutralization reduces focal cerebral ischemic injury in the absence of arterial recanalization.Blood. 2001; 97: 3086-92Crossref PubMed Scopus (44) Google Scholar] reported a reduction in infarct size in mice with a permanent ligation of the middle cerebral artery, with a bolus of human plasmin, which was associated with neutralization of plasma α2-antiplasmin [5Nagai N. De Mol M. Lijnen H.R. Carmeliet P. Collen D. Role of plasminogen system components in focal cerebral ischemic infarction. A gene targeting and gene transfer study in mice.Circulation. 1999; 99: 2440-4Crossref PubMed Scopus (212) Google Scholar]. Marder et al. [6Marder V.J. Landskroner K. Novokhatny V. Zimmerman T. Koig M. Kanouse J.J. Plasmin induces local thrombolysis without causing hemorrhage: a comparison with tissue plasminogen activator in the rabbit.Thromb Haemost. 2001; 86: 739-45Crossref PubMed Scopus (96) Google Scholar] recently demonstrated that human plasmin was comparable to tissue plasminogen activator (t-PA) for local thrombolysis in a rabbit abdominal aorta thrombosis model, but associated with less bleeding. Thus, provided plasmin can be adequately produced and formulated for pharmaceutical use, it could be useful for treatment of ischemic stroke and peripheral arterial thromboembolic disease. Plasminogen is easily obtained from human plasma or plasma fractions by affinity chromatography on lysine-Sepharose with yields of 0.25 g L−1. Plasminogen cannot readily be expressed in eukaryotic expression systems but has been obtained from the baculovirus/insect cell system [7Whitefleet-Smith J. Rosen E. McLinden J. Ploplis V.A. Tomlinson J.E. McLean J.W. Castellino F.J. Expression of human plasminogen cDNA in a baculovirus vector-infected insect cell system.Arch Biochem Biophys. 1989; 271: 390-9Crossref PubMed Scopus (37) Google Scholar], which is however, not suitable for large scale production. Plasmin is very unstable in neutral solutions, but can be stabilized with specific amino acids (such as lysine, 6-aminohexanoic acid or tranexamic acid), acid pH (range 2–4) or glycerol (10–50%) [3Jensen VJ (inventor). US patent 3,950,513. Process of stabilizing therapeutically useful plasmin solutions. Novo Terapeutisk Laboratorium (assignee). Issued 13 April, 1976.Google Scholar]. The present study deals with the high yield production, in the yeast Pichia pastoris, of recombinant human microplasminogen (µPlg) and the use of quantitatively activated microplasmin (µPli), stabilized with a dilute citrate buffer at pH 3.1. Microplasmin is a derivative of plasmin which lacks the five kringle domains [8Shi G.Y. Wu HL. Isolation and characterization of microplasminogen. A low molecular weight form of plasminogen.J Biol Chem. 1988; 263: 17071-5Abstract Full Text PDF PubMed Google Scholar]. Recombinant microplasminogen has previously been expressed in up to 15 mg L−1 culture medium [9Reich E, Easton T G (inventors). US patent 5,288,489. Fibrinolysis and fibrinogenolysis treatment. Orion Therapeutic Systems (assignee). Issued 22 February, 1994.Google Scholar]. The second order rate constant of the inhibition of microplasmin by α2-antiplasmin is 2 × 105 mmol L−1 s−1, which is about 100 times slower than the inhibition rate of intact plasmin by α2-antiplasmin (2–4 × 107 mmol L−1 s−1), due to the absence of the lysine binding site in microplasmin. This lower second order rate constant, corresponds to a half life of microplasmin in circulating blood of 4 s, as compared to a half-life of 0.02 s [10Wiman B. Collen D. On the kinetics of the reaction between human antiplasmin and plasmin.Eur J Biochem. 1978; 84: 573-8Crossref PubMed Scopus (236) Google Scholar]. The present study evaluates the effect of recombinant human microplasmin on the reduction of focal cerebral ischemic infarction (ischemic stroke) in two mouse models and on the dissolution of blood clots in an arteriovenous loop thrombosis model in rabbits. The pPICZαA vector (Invitrogen Corporation, Carlsbad, CA, USA) was used for expression and secretion of recombinant human microplasminogen (µPlg) in Pichia pastoris, under a license agreement between Thromb-X NV (Leuven, Belgium) and Research Corporation Technologies Inc (Tucson, AZ, USA). The production was carried out under good laboratory practice (GLP) conditions, at Eurogentech SA (Seraing, Belgium), as described elsewhere [11Collen D, Nagai N, Laroche Y. Stabilized recombinant plasmin. Patent Appl UK0116702.2, 11 July, 2001.Google Scholar]. Briefly, the vector contained alcohol oxidase elements to allow methanol inducible expression of transgenes, and a secretion signal for secretion of the heterologous protein in the medium. The µPli cDNA was isolated from the vector Fmyc-µPli [12Lasters I. Van Herzeele N. Lijnen H.R. Collen D. Jespers L. Enzymatic properties of phage-displayed fragments of human plasminogen.Eur J Biochem. 1997; 244: 946-52Crossref PubMed Scopus (16) Google Scholar] by PCR amplification and cloned flush with the secretion signal to obtain expression of µPlg with a native NH2 terminus. A zeocin resistant plasmid clone pPICZα-MPLG1 (clone #5) containing an insert of the expected size was sequenced to confirm correct insertion of the µPlg coding region (Fig. 1). Compared with the published sequence of human plasminogen [13Forsgren M. Raden B. Israelsson M. Larrson K. Heden L.O. Molecular cloning and characterization of a full-length cDNA clone for human plasminogen.FEBS Lett. 1987; 213: 254-60Crossref PubMed Scopus (154) Google Scholar], the nucleotide sequence differed in 10 positions, but the amino acid sequence was identical. Competent Pichia pastoris X33 cells (Invitrogen Corporation) were transformed with pPICZα-HPLG1 and a strain with high methanol inducible expression of µPlg (clone X33-MPLG1#5) was selected. Fermentation at a 50 L scale was carried out in four steps and µPlg was purified from the culture broth in a three step process, as described elsewhere [11Collen D, Nagai N, Laroche Y. Stabilized recombinant plasmin. Patent Appl UK0116702.2, 11 July, 2001.Google Scholar]. The activation of µPlg to µPli was performed at room temperature, for 30 min, at a molar ratio of 0.5% of staphylokinase (variant SY162 with reduced immunogenicity, comprising 12 amino acid substitutions). The activator was removed (over 99%) from µPli by hydrophobic chromatography on Phenyl Sepharose 6 Fast flow in the presence of 0.1 mol L−1 tranexamic acid, which was then removed and µPli equilibrated with 5 mmol L−1 citric acid, pH 3.1 by tangential ultrafiltration. The final yield of purified material approached 0.4 g L−1 culture broth. Human plasmin was prepared from plasma as described elsewhere [4Nagai N. De Mol M. Van Hoef B. Verstreken M. Collen D. Depletion of circulating α2-antiplasmin by intravenous plasmin or immunoneutralization reduces focal cerebral ischemic injury in the absence of arterial recanalization.Blood. 2001; 97: 3086-92Crossref PubMed Scopus (44) Google Scholar]; rt-PA (Actilyze®) was kindly donated by Boehringer Ingelheim GmbH (Ingelheim, Germany) and the platelet aggregation inhibitor ridogrel was a kind gift from Dr F. Declerck (Janssen Research Foundation, Beerse, Belgium). SDS gel electrophoresis was performed using the Mini-Protean II system (Bio-Rad, Nazareth, Belgium). (Micro)plasmin activity was measured with 0.4 mmol L−1 of the chromogenic substrate S-2403 (<Glu-Phe-Lys-pNA, Chromogenix, Antwerp, Belgium) using a ΔA min−1 at 37 °C, pH 7.4 and I 0.15, of 0.03 per nmol L−1 plasmin as determined by the manufacturer. α2-Antiplasmin levels were measured as described elsewhere [4Nagai N. De Mol M. Van Hoef B. Verstreken M. Collen D. Depletion of circulating α2-antiplasmin by intravenous plasmin or immunoneutralization reduces focal cerebral ischemic injury in the absence of arterial recanalization.Blood. 2001; 97: 3086-92Crossref PubMed Scopus (44) Google Scholar], and fibrinogen using a routine kinetic turbidimetric assay. Staphylokinase in the final µPli preparation was determined using a specific ELISA [11Collen D, Nagai N, Laroche Y. Stabilized recombinant plasmin. Patent Appl UK0116702.2, 11 July, 2001.Google Scholar]. Focal cerebral ischemia was produced by persistent occlusion of the middle cerebral artery (MCA) according to Welsh et al. [14Welsh F.A. Sakamoto T. McKee A.E. Sims R.E. Effect of lactacidosis on pyridine nucleotide stability during ischemia in mouse brain.J Neurochem. 1987; 49: 846-51Crossref PubMed Scopus (116) Google Scholar], as described elsewhere [5Nagai N. De Mol M. Lijnen H.R. Carmeliet P. Collen D. Role of plasminogen system components in focal cerebral ischemic infarction. A gene targeting and gene transfer study in mice.Circulation. 1999; 99: 2440-4Crossref PubMed Scopus (212) Google Scholar]. BALB/c mice were preferred because of the large size and low variability in ischemic cerebral infarction [4Nagai N. De Mol M. Van Hoef B. Verstreken M. Collen D. Depletion of circulating α2-antiplasmin by intravenous plasmin or immunoneutralization reduces focal cerebral ischemic injury in the absence of arterial recanalization.Blood. 2001; 97: 3086-92Crossref PubMed Scopus (44) Google Scholar]. Briefly, mice were anesthetized with 2.5% isoflurane in oxygen and rectal temperature was maintained at 37 °C. The left temporal muscle was transected and the skull exposed. A 1-mm opening was made in the region over the MCA with a hand-held drill under saline superfusion. The meningae were removed, the MCA occluded by ligation with 10–0 nylon thread (Ethylon, Neuilly, France) and transected distally, and the temporal muscle and skin sutured back in place. Study drugs (µPli, µPlg, plasmin, rt-PA or solvent) were given intravenously as a bolus, from 15 min up to 6 h after ligation of the MCA. The surgical operator was blinded with respect to the allocation of animals to control (solvent) or study drug groups. After 24 h, 3 days or 7 days, the animals were sacrificed with Nembutal (500 mg kg−1, Abbott Laboratories, North Chicago, IL, USA) and decapitated. The brain was placed in a matrix for sagittal sectioning in 1 mm segments. The sections were immersed in 2% 2,3,5-triphenyltetrazolium chloride (TTC) in saline, for 30 min at 37 °C, and placed in 4% formalin in phosphate-buffered saline (PBS), whereby the necrotic infarct area remains unstained (white) and distinguishable from stained (brick red) viable tissue. The sections were photographed and subjected to planimetry, and the focal cerebral ischemic infarct size was determined as the sum of the unstained areas of the sections, multiplied with their thickness. Photochemically induced thrombosis was produced with Rose Bengal [15Umemura K. Wada K. Uematsu T. Nakashima M. Evaluation of the combination of a tissue-type plasminogen activator, SUN9216, and a thromboxane A2 receptor antagonist, vapiprost, in a rat middle cerebral artery thrombosis model.Stroke. 1993; 24: 1077-81Crossref PubMed Google Scholar]. Anesthesia was performed with 2.5% isoflurane, rectal temperature was maintained at 37 °C, and a catheter (2FG, SIMS portex Limited, Kent, UK) for the administration of Rose Bengal was placed in the left jugular vein. The temporal muscle was transected and the skull was exposed. A 1.5-mm opening was made over the MCA with saline superperfusion and meningae and dura were removed. Photoillumination of green light (540 nm wave length) was achieved with a xenon lamp (model L-4887, Hamamatsu Photonics, Hamamatsu, Japan) with heat absorbing and green filters, via an optic fiber with a focus of 1 mm, placed on the opening in the skull. Rose Bengal (20 mg kg−1, Sigma, St. Louis, MI, USA) was injected, and photoillumination performed for 10 min. Administration of study compounds and analysis of cerebral infarct size was carried out as described above. The thrombolytic effect of µPli, plasmin and rt-PA was studied in an extracorporeal arteriovenous loop thrombus model in rabbits [16Hotchkiss A. Refino C. DeGuzman L. Rigter B. Eisert W. A new pan species model for the measurement of in vivo thrombolysis.Thromb Haemost. 1987; 58: 107Google Scholar]. New Zealand white rabbits weighing 2.9 ± 0.33 kg (mean ± SD) were anesthesized with 1.0 mL of 2% xylazine (Rompun, Bayer, Leverkusen, Germany) and 0.5 mL of ketamine (Ketalin, Apharmo BV, Arnhem, the Netherlands) and additional Nembutal [12 mg (intravenous) i.v. per hour]. A femoral vein catheter was introduced for blood sampling and a femoral artery catheter for blood pressure measurement (PDCR 75, Druck Ltd, Leicester, UK). A 300-µL thrombus was formed around a woollen thread introduced longitudinally in each of two adapted insulin syringes, from a mixture of 125I-labeled fibrinogen (400 000 cpm), platelet poor rabbit plasma, and 0.07 mL thrombin solution (100 NIHU mL−1). The clot was aged for 30 min at 37 °C and the syringes were inserted in a silicon tubing connecting a femoral artery with a marginal ear vein. Blood flow was regulated via a peristaltic pump (Pump P1, Pharmacia LKB, Piscataway, NJ, USA). Clot extension was prevented by infusion of heparin (300 U kg−1 bolus and 200 U kg−1 over 2 h) and ridogrel (7.5 mg kg−1 bolus, 30 min before the start of the infusion). Study drugs (6 mL over 2 h) were infused with a pump (Perfuser VI, B. Braun, Penang, Malaysia) via a three way valve, just proximal to the first inserted syringe in the extracorporal loop. Thrombolysis was determined 2.5 h after the start of the infusion, as the difference between the radioactivity originally in the clot and that remaining in the syringes, expressed as percent. Progression of clot lysis was monitored with Geiger counters, linked to a dedicated analysis system (Canberra, Meriden, CT, USA). Two-milliliter blood samples were drawn into trisodium citrate (0.011 mol L−1), for measurements of fibrinogen, α2-antiplasmin [3Jensen VJ (inventor). US patent 3,950,513. Process of stabilizing therapeutically useful plasmin solutions. Novo Terapeutisk Laboratorium (assignee). Issued 13 April, 1976.Google Scholar], and activated partial thromboplasmin time. Bleeding times were performed by applying a Symplate II device (Organon Technica, Durham, NC, USA) to a shaved inner thigh surface. The data are represented as mean ± SD or as median and range of n determinations. The significance of differences was determined using analysis of variance followed by Fisher's PLSD test, using the statview software package or by Student's t-test or Mann–Whitney test, as appropriate. Fermentation at the 50-L scale yielded approximately 0.4 g highly purified recombinant human microplasmin (µPli) per litre fermentation broth. SDS gel electrophoresis revealed a single band with apparent Mr of 30 kDa under non-reducing and a main band of 30 kDa, and two minor low Mr peptides under reducing conditions (Fig. 2). The material was fully active as determined with the chromogenic substrate S2403, using a Mr of 30 000, a of 1.0 for µPli and a ΔA min−1 at 405 nm of 0.03 per nmol L−1 active plasmin. The residual content of staphylokinase SY162 in the µPli was 0.25% of the original amount added to µPlg (200 parts staphylokinase per million parts µPli). The preparation was stable for at least 1 month at room temperature in solution, with at most a slight increase of low Mr peptide material (Fig. 2). Bolus injection of µPli decreased plasma α2-antiplasmin and fibrinogen levels proportionally to the µPli dose and recovered partially within 2 h (Table 1) and fully within 24 h (not shown), suggesting that α2-antiplasmin depletion was transient during the first hours after the injection of µPli. Bolus injection of 4 mg kg−1 t-PA caused minor α2-antiplasmin reduction without fibrinogen reduction.Table 1Effect of µPli on plasma, α2-antiplasmin and fibrinogen levels in miceAgentDose (mg kg−1)α2-Antiplasmin (%)Fibrinogen (%)0 min15 min1 h2 h0 min15 min1 h2 hµPli2.594 ± 2.173 ± 1675 ± 3.6100 ± 12100 ± 1292 ± 3278 ± 2.475 ± 4.25.0–30 ± 2225 ± 1034 ± 6.7–32 ± 1715 ± 3.464 ± 17t-PA4.0–85 ± 4.1–––100 ± 9.2––The data represent the mean ± SD of groups of at least three animals. Open table in a new tab The data represent the mean ± SD of groups of at least three animals. Ligation of the MCA induced a cerebral infarct with a volume of 29 mm3 (range 27–30) (n = 6) in inbred BALB/c mice (Table 2). Injection of 2.5 mg kg−1µPli had no significant effect on infarct size, whereas 5.0 mg kg−1µPli, or 5.0 mg kg−1 followed 45 min later by another 2.5 mg kg−1, produced a 15% reduction of infarct size. These results are in line with the transient partial reduction of α2-antiplasmin with the lower dose and the more persistent depletion with 5.0 mg kg−1µPli. Infarct size was reduced with µPli injections up to 4 h after MCA occlusion, and when measured up to 3 days after µPli administration. µPlg at a dose of 40 mg kg−1 or rt-PA at a dose of 4.0 mg kg−1 did not reduce infarct size (Table 2).Table 2Effect of systemic administration of µPli, µPlg and rt-PA on focal cerebral ischemic injury in mice with permanent middle cerebral artery (MCA) ligationCompoundDose (mg kg−1)Injection time (h after MCA-O)*MCA-O: middle cerebral artery occlusion;Sacrifice (days after MCA-O)Cerebral infarct sizeControlCompoundP vs. controlµPli2.50.25129 (27–30) (6)29 (27–30) (6)0.745.0**pooled data of 5.0 mg kg−1µPli vs. control 27 (26–30) vs. 25 (21–28) mm3 (median and range, n= 16, P < 0.0001).0.251id26 (21–28) (6)0.045.0 + 2.50.25 and 11id26 (20–28) (6)0.025.0**pooled data of 5.0 mg kg−1µPli vs. control 27 (26–30) vs. 25 (21–28) mm3 (median and range, n= 16, P < 0.0001).0.25127 (26–29) (6)24 (22–27) (6)0.015.011id25 (20–27) (6)0.045.021id24 (22–28) (6)0.025.041id25 (21–28) (6)0.055.061id26 (24–32) (6)0.655.0**pooled data of 5.0 mg kg−1µPli vs. control 27 (26–30) vs. 25 (21–28) mm3 (median and range, n= 16, P < 0.0001).0.25127 (26–29) (6)23 (22–27) (6)0.015.00.25324 (22–28) (6)22 (21–24) (6)0.015.00.25712 (9.0–18) (6)11 (5.2–13) (6)0.15Plasmin7.50.25128 (25–32) (6)24 (20–26) (6)0.01µPlg400.25126 (25–28) (6)26 (24–27) (6)0.87rt-PA4.00.25128 (25–32) (6)26 (22–31) (6)0.11Cerebral infarct size is expressed as median and range of the number of experiments given between brackets;* MCA-O: middle cerebral artery occlusion;** pooled data of 5.0 mg kg−1µPli vs. control 27 (26–30) vs. 25 (21–28) mm3 (median and range, n= 16, P < 0.0001). Open table in a new tab Cerebral infarct size is expressed as median and range of the number of experiments given between brackets; Photoillumination of the MCA produced thrombosis resulting in a cerebral infarct volume of 20 mm3 (range 16–30) (n = 6) in the control group (Table 3), which was reduced to 4.8 mm3 (3.1–22) (n = 6) with 5.0 mg kg−1µPli (P < 0.01) and to 11 mm3 (2.1–22) (n = 6) with 5.0 plus 2.5 mg kg−1 (P = 0.04). Infarct size was reduced with µPli up to 10 h after occlusion and when measured up to 3 days after its administration. Plasmin at a dose of 7.5 mg kg−1 also did, but µPlg at a dose of 40 mg kg−1 did not reduce infarct size significantly. rt-PA reduced infarct size significantly when administered 15 min after illumination but not after 4 h (Table 3).Table 3Effect of systemic administration of µPli, plasmin, µPlg or rt-PA on focal cerebral ischemic injury in mice with photochemically induced MCA thrombosisCompoundDose (mg/kg)Injection time (h after MCA-O)Sacrifice (days after MCA-O)Cerebral infarct size (mm3) Median and rageControlCompoundP vs. controlµPli2.50.25120 (16–30) (6)17 (7.6–29) (6)0.325.00.251id4.8 (3.1–22) (6)<0.015.0 + 2.50.25 and 11id11 (2.1–22) (6)0.045.00.25126 (15–35) (17)14 (4.3–24) (12)<0.015.011id14 (2.4–23) (12)<0.015.021id9.4 (2.7–31) (12)<0.015.041id16 (2.7–25) (12)<0.015.061id16 (4.4–28) (12)<0.015.081id21 (5.7–33) (12) 0.5 g L−1 fermentation broth) in Pichia pastoris, purification to homogeneity, quantitative conversion to microplasmin (µPli), stabilization of the enzyme in a dilute citrate buffer at pH 3.1 and lyophylization. The potential therapeutic efficacy of the homogeneous, fully active material is illustrated in animal models of ischemic stroke (using systemic intravenous administration) and of arterial thrombosis (using local administration). Ischemic stroke due to thrombotic occlusion of a cerebral artery is amenable to therapy with antithrombotic and thrombolytic agents. The use of rt-PA within 3 h of symptom onset improves neurologic outcome, but may cause acute hemorrhage in the brain. In previous studies we observed that α2-antiplasmin gene deficient mice had smaller cerebral infarct size after middle cerebral artery (MCA) ligation [5Nagai N. De Mol M. Lijnen H.R. Carmeliet P. Collen D. Role of plasminogen system components in focal cerebral ischemic infarction. A gene targeting and gene transfer study in mice.Circulation. 1999; 99: 2440-4Crossref PubMed Scopus (212) Google Scholar] and that reduction of α2-antiplasmin with a single bolus of human plasmin, or of an anti-α2-AP Fab fragment reduced infarct size [4Nagai N. De Mol M. Van Hoef B. Verstreken M. Collen D. Depletion of circulating α2-antiplasmin by intravenous plasmin or immunoneutralization reduces focal cerebral ischemic injury in the absence of arterial recanalization.Blood. 2001; 97: 3086-92Crossref PubMed Scopus (44) Google Scholar]. This suggested that reduction of circulating α2-antiplasmin may constitute a new approach to reduce cerebral infarct size in ischemic stroke in the absence of reperfusion, possibly via a neuroprotective mechanism of action. In the present study we have explored the effect of µPli in two mouse models consisting of permanent ligation or photochemically induced thrombosis of the MCA. In the first model, µPli reduced cerebral infarct size by approximately 15%, comparable with that of full length plasmin [4Nagai N. De Mol M. Van Hoef B. Verstreken M. Collen D. Depletion of circulating α2-antiplasmin by intravenous plasmin or immunoneutralization reduces focal cerebral ischemic injury in the absence of arterial recanalization.Blood. 2001; 97: 3086-92Crossref PubMed Scopus (44) Google Scholar]. In the photoactivation induced thrombosis model, the extent of cerebral infarct size reduction was much larger (approximately 50%), possibly due to a combined effect of reperfusion-dependent and reperfusion-independent mechanisms. Microplasmin lacks fibrin affinity and has short half-life in blood, suggesting that its thrombolytic potency might primarily result from enhanced endogenous thrombolytic after reduction of α2-AP [17Sakata Y. Eguchi Y. Mimuro J. Matsuda M. Sumi Y. Clot lysis induced by a monoclonal antibody against alpha 2-plasmin inhibitor.Blood. 1989; 74: 2692-7Crossref PubMed Google Scholar, 18Reed G.L. Rd Matsueda G.R. Haber E. Inhibition of clot-bound alpha 2-antiplasmin enhances in vivo thrombolysis.Circulation. 1990; 82: 164-8Crossref PubMed Scopus (35) Google Scholar]. The beneficial effect of µPli was observed when administered for 4–10 h after occlusion whereas the effect of rt-PA had disappeared after 4 h, when infarct expansion was observed. This may be explained by the recent observation that, in the absence of reperfusion, rt-PA causes a dose-related infarct expansion [19Nagai N. Vanlinthout I. Collen D. Comparative effects of tissue-type plasminogen activator, streptokinase and staphylokinase on cerebral ischemic infarction and pulmonary clot lysis in hamster models.Circulation. 1999; 100: 2541-6Crossref PubMed Scopus (42) Google Scholar], which may be due to its neurotoxic effect via plasminogen activator mediated proteolysis and activation of the NMDA receptor [20Nicole O. Docagne F. Ali C. Margaill I. Carmeliet P. Mackenzie E.T. Vivien D. Buisson A. The proteolytic activity of tissue-plasminogen activator enhances NMDA receptor-mediated signaling.Nat Med. 2001; 7: 59-64Crossref PubMed Scopus (612) Google Scholar]. Consequently, provided these observations can be extrapolated to patients with ischemic stroke, microplasmin might constitute a safer alternative to rt-PA. In patients with peripheral artery occlusion disease, intra-arterial thrombolysis may prevent complex surgical procedures and allow simpler elective procedures to correct unmasked lesions. The most feared complication of intra-arterial thrombolysis with plasminogen activators is, however, intracranial bleeding which occurs in 1–2% of treated patients. Local infusion of plasmin obtained from plasma plasminogen has recently been shown in a rabbit model to dissolve arterial clots with less bleeding time prolongation than rt-PA [6Marder V.J. Landskroner K. Novokhatny V. Zimmerman T. Koig M. Kanouse J.J. Plasmin induces local thrombolysis without causing hemorrhage: a comparison with tissue plasminogen activator in the rabbit.Thromb Haemost. 2001; 86: 739-45Crossref PubMed Scopus (96) Google Scholar]. The present study confirms and extends these observations to recombinant human microplasmin. In conclusion, a preparation of µPli was developed, which appears to be suitable for pharmaceutical use. Being a truncated derivative of plasmin lacking the lysine binding sites, it has no affinity for fibrin and it is inhibited more slowly by α2-antiplasmin than intact plasmin. In in vivo models of ischemic stroke and peripheral arterial occlusion, µPli appeared to be comparable in potency to full length plasmin. In line with its lack of fibrin affinity and its short plasma half life, µPli did not cause a bleeding tendency nor hemostatic plug dissolution at distant sites. The present results warrant further investigation of µPli both for the treatment of ischemic stroke by systemic administration and of arterial thromboembolic disease by local catheter delivery. This study was performed under a research agreement between the University of Leuven, Belgium, and Thromb-X, NV in which D. Collen has an equity interest.