Endothelial targeting of a recombinant construct fusing a PECAM-1 single-chain variable antibody fragment (scFv) with prourokinase facilitates prophylactic thrombolysis in the pulmonary vasculature

Abstract

Means to prevent thrombus extension and local recurrence remain suboptimal, in part because of the limited effectiveness of existing thrombolytics. In theory, plasminogen activators could be used for this purpose if they could be anchored to the vascular lumen by targeting stably expressed, noninternalized determinants such as platelet-endothelial-cell adhesion molecule 1 (PECAM-1). We designed a recombinant molecule fusing low-molecular-weight single-chain prourokinase plasminogen activator (lmw-scuPA) with a single-chain variable fragment (scFv) of a PECAM-1 antibody to generate the prodrug scFv/lmw-scuPA. Cleavage by plasmin generated fibrinolytically active 2-chain lmw-uPA. This fusion protein (1) bound specifically to PECAM-1-expressing cells; (2) was rapidly cleared from blood after intravenous injection; (3) accumulated in the lungs of wild-type C57BL6/J, but not PECAM-1 null mice; and (4) lysed pulmonary emboli formed subsequently more effectively than lmw-scuPA, thereby providing support for the concept of thromboprophylaxis using recombinant scFv-fibrinolytic fusion proteins that target endothelium.

Figure 1.

Molecular design, expression, and characterization of anti-PECAM scFv-lmw scuPA. (A) Schematic diagram describing the cloning strategy for the fusion construct pMT-BD1. Variable domains of heavy chain and light chain were linked by a (Gly₄Ser)₃ linker and then fused to the N-terminus of lmw-scuPA by a (Ser₄Gly)₂Ala₃ linker. (B) Variable domains of heavy chain (H chain) and light chain (L chain) of P-390 were amplified and assembled into full-length scFv. M indicates DNA standards. (C) Lmw-scuPA and anti-PECAM scFv were ligated and cloned into SpeI and XhoI sites of the pMT expression vector. XhoI and SpeI digestion of the fusion construct (fusion). (D) Western blot analysis of 40 μL culture medium alone or after induction by 0.5 mM CuSO₄. Purified fusion protein (50 ng and 200 ng) was blotted to compare expression levels. (E) Four percent to 12% gradient SDS-PAGE of purified fusion protein with or without plasmin treatment under unreduced or reduced conditions.

Figure 2.

Specific binding of scFv-uPA fusion protein to cells expressing mouse PECAM. (A) FITC-streptavidin staining of REN/PECAM (left) versus control REN (right) cells after incubation with biotinylated anti-PECAM scFv-scuPA (original magnification × 40). (B) ELISA: binding of anti-PECAM scFv-uPA to REN/PECAM (•) versus REN (○) cells. (C) ELISA: inhibition of binding of fusion protein to REN/PECAM cells by parental anti-PECAM IgG mAb 390. Error bars indicate standard error of the mean (SEM).

Figure 3.

Urokinase activity of free and cell-bound anti-PECAM scFv-uPA. (A) Amidolytic activities of fusion protein generated at different molar ratios of plasmin to scFv-uPA. (B) Fibrinolytic activity using a fibrin plate. From left to right 1:3 serial dilutions of lmw-tcuPA (50 ng), lmw-scuPA (100 ng), and scFv-lmw scuPA (200 ng) were incubated on a fibrin-coated plate at 37°C. Lytic zones were measured after staining fibrin with trypan blue. (C) Amidolytic activity associated with the cell surface of control REN (○) versus PECAM-transfected (•, REN/PECAM) REN cells was determined by conversion of chromogenic substrate after incubation with various amounts of fusion protein. (D) Preincubation of REN/PECAM cells with parental anti-PECAM IgG, mAb 390, reduces binding of enzymatically active scFv-uPA. Kinetics of disappearance of cell-bound scFv-uPA (□) and its amidolytic activity () was determined by using H5V mouse endothelioma cells (E) and human REN/PECAM-1 cells (F). Basal levels of uPA antigen (▪) and amidolytic activity () were determined by using intact cells. Amounts of fusion protein and amidolytic activity anchored to cell surface were significantly different from basal levels at 3 hours in mouse cells (P < .002) and 24 hours in human cells (P < .01). Error bars indicate SEM.

Figure 4.

Biodistribution of anti-PECAM scFv-lmw scuPA and lmw-scuPA in vivo. Ten micrograms fusion protein or equal molar lmw-scuPA was mixed with 0.2 μg radiolabeled tracer protein and injected intravenously into wild-type or PECAM null mice, respectively. One hour later, tissue uptake was measured. (A) Percentage of injected dose per gram tissue (%ID/g). Note that scFv-uPA, but not scuPA, shows preferential uptake in the lungs and other vascularized organs in wild-type (WT), but not in PECAM KO mice. (B) Organ-to-blood ratio for various organs. Broken line indicates blood level; ratio equal to 1.0. (C) Immunospecificity index (ISI), calculated as ratio of organ-to-blood ratios of targeted and untargeted counterpart. The interrupted line shows an ISI of 1.0, reflecting equal tissue levels of targeted and untargeted counterparts. Error bars indicate SEM.

Figure 5.

Blood clearance of anti-PECAM scFv-scuPA and kinetics of in vivo pulmonary targeting. (A) Blood clearance of targeted fusion construct (•) and nontargeted scuPA (○) in %ID/g (percentage of injected dose per gram tissue). (B) Kinetics of the targeted fusion protein levels in lungs (•) and blood (○). Fusion protein exhibited a rapid and prolonged accumulation in lung tissues. Lung-to-blood ratios at indicated time points were calculated (inset). (C) In vivo biodistribution of ¹²⁵I-scFv-uPA (▪) and free uPA () 3 hours after intravenous injection in wild-type mice. Error bars indicate SEM.

Figure 6.

Accumulation of fusion protein in pulmonary vasculature facilitates local fibrinolysis. (A) Dose-response curve of pulmonary thrombolysis. Dissolution of ¹²⁵I-labeled microemboli lodged in mouse pulmonary vasculature by bolus injection of 300, 100, and 30 μg fusion protein, the equivalent amounts of lmw-scuPA (150, 50, and 15 μg) versus a mixture of lmw-scuPA and parental antibody, respectively. Thrombolytic potency was expressed as the percentage of lysis versus the dose administered. Interrupted line indicates spontaneous lysis. Error bars indicate SEM. (B) Simplified model of a proposed strategy for thromboprophylaxis using vascular immunotargeting of genetically engineered anti-PECAM scFv-uPA fusion protein. Anti-PECAM scFv-uPA circulates in a form of a prodrug, single-chain uPA, binds to PECAM-1, and remains anchored on the luminal surface of endothelium for at least several hours. In situ thrombosis or embolism induces initial local conversion of plasminogen (Pg) into plasmin (Pn) by endogenous plasminogen activators (Endo-PA). Plasmin (and perhaps other enzymes) formed in the vicinity of the clot converts the endothelium-bound scFv-scuPA into enzymatically active tcuPA (Figure 3A), which in turn amplifies local formation of plasmin, reinforcing local thrombolysis, preventing clot extension and reocclusion.