WEDGE: An anticoagulant thrombin mutant produced by autoactivation

Summary

Background

Production of therapeutically relevant proteases typically involves activation of a zymogen precursor by external enzymes, which may raise regulatory issues about availability and purity. Recent studies of thrombin precursors have shown how to engineer constructs that spontaneously convert to the mature protease by autoactivation, without the need of external enzymes.

Objectives

Autoactivation is an innovative strategy that promises to simplify the production of proteases of therapeutic relevance, but has not been tested in practical applications. This study aims to provide a direct test of this strategy.

Methods

An autoactivating version of the thrombin mutant W215A/E217A (WE), currently in pre-clinical development as an anticoagulant, is engineered.

Results and Conclusions

The autoactivating version of WE can be produced in large quantities, like WE made in BHK cells or E. coli, and retains all significant functional properties in vitro and in vivo. The results serve as proof of principle that autoactivation is an innovative and effective strategy for production of trypsin-like proteases of therapeutic relevance.

Introduction

The application of therapeutic proteases to diseases of coagulation pathways is well documented [1]. In properly regulated cellular pathways, proteases are synthesized as inactive precursors which are then activated in response to metabolic state of the cell, as well as in response to extracellular events, such as vascular injury [2]. This strategy from biology has been successfully applied for protease production in recombinant organisms by overexpression as an inactive precursor, followed by an activation step. Among the first proteins manufactured by recombinant engineering were proteases of the coagulation cascade for hemophilia treatment [1]. Factor IX zymogen is overexpressed as a secretion construct in CHO cells, where activation occurs by furin cleavage of the proenzyme [3]. Production of activated protein C was also accomplished by such a strategy, in which protein C zymogen is secreted into recombinant mammalian cell culture medium and subsequently activated by thrombin to the active protease [1].

The need for exogenous enzymes for the activation step in production of therapeutic proteases presents challenges upon scale up that include potential safety concerns associated with proteins isolated from tissues or blood products, maintenance of consistent quality and availability of the proteases, and increased cost of production. An alternative strategy for zymogen activation has emerged recently from structural biology of thrombin precursors. Prethrombin-2 shows R15 in the site of proteolytic activation by prothrombinase or ecarin in electrostatic interaction with E14e, D14l and E18 [4]. Disruption of these interactions by mutagenesis produces derivatives that spontaneously convert to thrombin, without appreciable perturbation of the functional properties of the enzyme [5]. The reaction is started by the zymogen itself and is abrogated by inactivation of the catalytic S195. This suggests a convenient strategy for the production of protein therapeutics with desired pharmacodynamic properties that obviates the need of external enzymes and associated potential regulatory hurdles. Here we describe a large scale production strategy that exploits autoactivation for the thrombin mutant WE [6], currently in pre-clinical development due to its compelling profile of efficacy and safety as an anticoagulant/antithrombotic and anti-inflammatory agent in vivo documented by several pre-clinical studies in rodent and non-human primate models [7–12]. The strategy offers a suitable alternative to existing protocols for production of thrombin from activation of prethrombin-1 by prothrombinase [13, 14] or prethrombin-2 by ecarin [15, 16], thereby obviating the need, costs and possible contaminations from external proteases.

Methods

Purification of WEDGE was similar to that of WE expressed in E. coli [4, 16], with modifications for larger scale production. Inclusion bodies were produced using a fed-batch fermentation employing 2×M9 yeast extract glucose medium and a glycerol plus yeast extract feed solution. Refolding was initiated by addition of reduced, denatured inclusion bodies into rapidly stirring refolding buffer. Concentration and diafiltration of refold reactions prior to heparin-Sepharose chromatography was carried out using a hollow-fiber ultrafiltration cartridge. Autoactivation was allowed to proceed at room temperature and pH 8.0 after concentration of the heparin-Sepharose pool to 2–3 mg/mL. Progress of autoactivation was monitored by reversed-phase HPLC separation of thrombin A and B chains, and other intermediate forms, using a Vydac C4, 2.1×50 mm column. After sample reduction in 3 M guanidine HCl by 10 mM DTT, acidified samples were chromatographed using gradient elution from 15% to 65% acetonitrile in 0.1% trifluoroacetic acid at a flow rate of 0.20 mL/min with UV detection at 214 nm. After autoactivation had reached completion, buffer exchange using a Sephadex G-25 column was followed by cation exchange chromatography on SP-Sepharose in MES buffer, pH 6.0 with elution of active WEDGE with a gradient of NaCl before storage of bulk WEDGE at −20 °C. Overall yield from 14 g of inclusion bodies derived from 5 L fermentation to approximately 33 mg of bulk product was 1.2%, quite comparable to the 1.1% yield of WE produced in E. coli. This yield is typical of proteins refolding with multiple disulfide bonds. Samples for in vivo studies were treated with Detoxigel™ (Thermo-Fisher) before use to remove residual endotoxin. Studies with baboons were approved by the Institutional Animal Care and Use Committee of Oregon Health & Science University. Baboons were given a single intravenous bolus dose of WEDGE in 1 ml volume of saline. On day 1, baboon #1 received a single dose of 2.5 µg/kg WEDGE and baboon #2 received a single dose of 1.0 µg/kg WEDGE. On day 2, baboon #2 received a single dose of 2.5 µg/kg WEDGE. APTT was measured after collecting blood samples into 1/10 vol of 3.2% citrate processed to plasma prior to treatment and at 5, 15, 30, 60, 120 min post treatment. APTT was read on a KC-1 using standard protocols. The effects of WEDGE was also assessed using the standard template, skin bleeding time test (Surgicutt; International Technidyne, Piscataway, NJ) at 15, 30, 60 min after the start of treatments.

Results and Discussion

Construction of several mutations in the activation domain of prethrombin-2 carrying the WE substitution revealed a construct carrying the additional three substitutions D14lA/G14mP/E18A (WEDGE) that completely activated to thrombin in 10 h when used at a concentration of 3 mg/mL. This time frame is consistent with protocols for large scale production in a biotechnological/pharmaceutical setting. WEDGE features functional properties toward physiological substrates in vitro (Figure 1) that are practically identical to those reported for WE produced in mammalian BHK cells from prethrombin-1 and activated with prothrombinase [6], or produced in E. coli from prethrombin-2 and activated with ecarin [4, 16]. Hence, differences in glycosylation and introduction of three additional mutations in the activation domain of WEDGE relative to WE are inconsequential on the kinetics of substrate hydrolysis.

Figure 1.

Values of the log of the specificity constant k_cat/K_m (M⁻¹s⁻¹) for the hydrolysis of physiological substrates (fibrinogen, PAR1 fragment ³³ATNATLDPRSFLLRNPNDKYEPFWEDEEKN⁶² and protein C in the presence of 10 nM thrombomodulin and 5 mM CaCl₂) by thrombin wild-type, WE expressed in BHK cells as prethrombin-1 and activated with prothrombinase [6] or expressed in E. coli as prethrombin-2 and activated with ecarin [16], and WEDGE produced by autoactivation. The WE mutation causes a drastic loss of activity toward fibrinogen and PAR1 with only a modest effect on protein C activation in the presence of Ca²⁺ and thrombomodulin. The additional mutations introduced in WEDGE relative to WE to enable autoactivation are inconsequential on the functional properties of the construct. Experimental conditions are: 5 mM Tris, 0.1% PEG8000, 145 mM NaCl, pH 7.4 at 37 °C.

The pharmacodynamic efficacy of WEDGE was tested in non-human primates by analysis of activated partial thromboplastin time (APTT) in plasma. Bolus administration of WEDGE caused prolongation of the APTT (Figure 2) with the greatest effect occurring at 15 min post-injection and returning to baseline by 120 min. WEDGE prolonged the APTT approximately 1.6 fold at 2.5 µg/kg and 1.3 fold at 1.0 µg/kg after 15 min, consistent with the dose-dependent increase in APTT reported previously for WE [10]. There were no WEDGE-related adverse events in the study subjects over the time scale of the experiments. Bleeding times measured at 15 (3.3 min), 30 (2.8 min) and 60 (2.2 min) min post injection were not significantly elevated over baseline (2.5 min).

Figure 2.

APTT values as a function of time following a single bolus injection of WEDGE at 1.0 µg/kg (open red circles) and 2.5 µg/kg (closed red circles), or 2.5 µg/kg in a second baboon (closed green circles). WEDGE causes the APTT to increase significantly and transiently in a dose-dependent manner that reproduces the effect previously reported for WE [9, 10].

The development of an efficient strategy for the production of thrombin from its inactive precursors by autoactivation suggests other applications. The structurally related zymogen protein C has been extensively studied as an anticoagulant and anti-inflammatory agent [17–19]. Production of activated protein C requires thrombin, which must be added to the preparation and then eliminated to avoid potential thrombotic complications. The structural analogy of the activation domain of protein C with prethrombin-2 supports the viability of autoactivation for this natural anticoagulant factor [20] and its large-scale production devoid of external enzymes. In general, the activation sequence of a trypsin-like protease may be re-engineered to promote autoactivation for production of mature proteases of clinical and biotechnological relevance.