Direct delivery of plasmin using clot-anchoring thrombin-responsive nanoparticles for targeted fibrinolytic therapy

Abstract

Background:

Fibrin-rich clot formation in thrombo-occlusive pathologies is currently treated by systemic administration of plasminogen activators (e.g. tPA), to convert fibrin-associated plasminogen to plasmin for fibrinolytic action. However, this conversion is not restricted to clot site only but also occurs on circulating plasminogen, causing systemic fibrinogenolysis and bleeding risks. To address this, past research has explored tPA delivery using clot-targeted nanoparticles.

Objectives:

We designed a nanomedicine system that can (1) target clots via binding to activated platelets and fibrin, (2) package plasmin instead of tPA as a direct fibrinolytic agent, and (3) release this plasmin triggered by thrombin for clot-localized action.

Methods:

Clot-targeted thrombin-cleavable nanoparticles (CTNPs) were manufactured using self-assembly of peptide-lipid conjugates. Plasmin loading and its thrombin-triggered release from CTNPs were characterized by UV-visible spectroscopy. CTNP-targeting to clots under flow was studied using microfluidics. Fibrinolytic effect of CTNP-delivered plasmin was studied in vitro using BioFlux imaging and D-dimer analysis and in vivo in a zebrafish thrombosis model.

Results:

Plasmin-loaded CTNPs significantly bound to clots under shear flow and showed thrombin-triggered enhanced release of plasmin. BioFlux studies confirmed that thrombin-triggered plasmin released from CTNPs rendered fibrinolysis similar to free plasmin, further corroborated by D-dimer analysis. In the zebrafish model, CTNP-delivered plasmin accelerated time-to-recanalization, or completely prevented occlusion when infused before thrombus formation.

Conclusion:

Considering that the very short circulation half-life (<1 second) of plasmin prevents its systemic use but also makes it safer without off-target drug effects, clot-targeted delivery of plasmin using CTNPs can enable safer and more efficacious fibrinolytic therapy.

1 |. INTRODUCTION

Thrombo-occlusive vascular diseases including atherosclerosis, myocardial infraction, stroke, deep vein thrombosis, pulmonary embolism, etc. continue to be the leading causes of morbidity and mortality in the world [1, 2]. Beyond the already existent global burden of such pathologies, the emergence, and sustained effects of the COVID-19 pandemic has given rise to new challenges of thrombosis and thromboinflammation in the acute as well as chronic phases of patients with COVID-19 [3–5]. Irrespective of etiology, one common end-point of such thrombotic pathologies is the formation of insoluble crosslinked protein fibrin as a major component of the occlusive blood clot [6–10]. Therefore, traditional and emerging therapeutic approaches remain focused on the utilization of drugs that can reduce fibrin formation (e.g., anticoagulants) and that can degrade already formed fibrin (e.g., fibrinolytics) [11–16]. These drugs are administered systemically (oral or intravenous) in a highly regulated regimen, and persistently present challenges of: (i) suboptimal circulation time and bioavailability of the drugs because of plasma-induced deactivation by circulating inhibitors, and (ii) off-target action of the drugs causing harmful side-effects including coagulaopathy and hemorrhage [17–24]. These issues can be potentially resolved by spatiotemporally localizing the delivery and action of the drugs at the vascular thrombotic site. This is where “nanomedicine”-based approaches present significant promise, via the utilization of nanoparticle platforms that can encapsulate drug molecules for disease site-specific delivery and release for localized action [19, 25–30]. The utilization of this principle has been studied in the past, including our own research, by developing nanoparticles that can actively anchor to the thrombotic site via binding to activated platelets or to fibrin [28, 31–35]. For release of the drug payload from such clot-localized nanoparticles, prior studies have utilized diffusion or ultrasound-induced particle destabilization or enzyme-induced (e.g., phospholipase, thrombin, etc.) particle degradation [36]. The drug payload that has been predominantly delivered in such studies is tissue plasminogen activator (tPA) because it is the currently approved drug for fibrinolytic therapy in the United States and Europe.

Building on such studies that demonstrate the promise of “targeted delivery” and “stimuli-triggered release” for tPA therapy, we sought to explore whether such a nanomedicine approach can enable direct delivery of plasmin (instead of tPA) for rapid fibrinolytic effect. Direct systemic administration of plasmin has poor efficacy because it is rapidly (within < 1 second) and irreversibly inhibited by inhibitors (e.g., antiplasmin) in circulation [37]. Thus, the clinical standard of care has focused on intravenous administration of plasminogen activators (previously Streptokinase, and currently tPA that was approved in the United States in 1996), because such a plasminogen activator molecule can stay in circulation for longer time (e.g., tPA half-life in circulation is 5–10 min) and thereby can enable conversion of clot-associated plasminogen to plasmin for fibrinolytic action [37,38]. However, tPA conversion of plasminogen to plasmin is not restricted to clot site only, but can occur in circulation on fibrinogen-bound plasminogen resulting in systemic fibrinogenolysis, and this has been implicated as a major cause of bleeding risks associated with tPA therapy [22–24]. In fact, such findings have prompted the nanomedicine research on clot-targetd delivery of tPA as described above. Compared with tPA, plasmin has an improved safety profile because its rapid neutralization by antiplasmin in circulation minimizes off-target risks, but this rapid neutralization is also the reason why plasmin cannot be used as a direct intravenous fibrinolytic therapy. Interestingly, catheter-mediated delivery of plasmin proximal to the clot site has shown fibrinolytic efficacy [39, 40], but such procedures are complicated. Considering these challenges and opportunities, we explored the feasibility of developing a liposome-based nanomedicine system that: (i) Can be systemically administered, whereas allowing specific anchorage onto clots via heteromultivalent binding of activated platelets and fibrin, (ii) Undergo thrombin-triggered degradation of the lipid membrane to enable clot site-responsive particle destabilization for payload release, and (iii) Encapsulate plasmin to protect it from neutralization in circulation, whereas enabling its release specifically at the clot site for localized fibrinolytic action. Figure 1 depicts the design concept and mechanism of action for these plasmin-loaded heteromultivalently clot-targeted thrombin-cleavable nanoparticles (CTNPs).

FIGURE 1.

Design and mechanism of action of plasmin-loaded CTNPs for targeted fibrinolysis; CTNPs anchor to clots via PBP binding to αIIbβ3 on activated platelets and FBP, and undergo thrombin-triggered destabilization for site-localized release of plasmin for fibrinolysis. CTNP, clot-targeted thrombin-cleavable nanoparticle; FBP, fibrin-binding peptides; PBP, platelet-binding peptides.

2 |. MATERIALS AND METHODS

2.1 |. Materials

For liposomal nanoparticle fabrication, distearoyl phosphotidyl choline (DSPC), methoxy polyethylene glycol-conjugated distearoyl phosphotidyl ethanolamine (DSPE-mPEG₂₀₀₀), and maleimide-terminated polyethylene glycol-conjugated distearoyl phosphotidyl ethanolamine (DSPE-PEG₂₀₀₀-Mal) were purchased from Avanti Lipids (Alabaster, USA). N-succinimide-terminated polyethylene glycol-conjugated distearoyl phosphotidyl ethanolamine (DSPE-PEG₂₀₀₀-NHS) was purchased from Nanosoft Polymers (Winston-Salem, USA). Platelet-binding peptide (PBP) sequence CGSSSGRGDSPA that binds to activated platelet surface integrin αIIbβ3 [41–44], fibrin-binding peptide (FBP) sequence cyclo-AC-Y(DGI)C(HPr)YGLCYIQGK-Am [34,45], and thrombin cleavable peptide (TCP) sequence DVTPRC [46] were custom synthesized by Genscript (Piscataway, USA). Polycarbonate membrane filters with 200 nm pore distribution, calcium chloride, stearylamine, diethyl ether, dimethyl sulfoxide, Sephadex G-100, and cholesterol were purchased from Sigma Aldrich (St. Louis, USA). Rhodamine-B-dihex-adecanoyl-sn-glycero-3-phosphoethanolamine (DHPE-RhB, red fluorescence, λex = 561, λem = 582) was purchased from Setareh Biotech (Eugene, USA). Thrombin, plasmin, and α2-antiplasmin were purchased from Hematological Technologies (Essex Junction, USA). Phosphate buffered saline, tris(hydroxymethyl)aminomethane hydrochloride (Tris-HCl), succinimidyl 3-(2-pyridyldithio)propionate (SPDP), fluted filter paper, collagen type I from rat tail, Chrono Log Corporation collagen type I, and AlexaFluor 647-conjugated fibrinogen were purchased from Fisher Scientific (Waltham, USA). Calcein AM was purchased from Thermo Fisher (Waltham, USA). Liposome extruder for nanoparticle manufacture was purchased from Evonik (Essen, Germany). For microfluidic studies, the parallel plate flow chamber (PPFC) was purchased from Glycotech (Gaitersberg, USA), and BioFlux flow controller and microfluidic plates were purchased from Fluxion Biosciences (Alameda, USA). All human blood and plasma for in vitro studies were obtained either from healthy donors using protocol approved by Case Western Institutional Review Board (Case IRB STUDY20191092) or from the Case Western Hematopoietic Biorepository and Cellular Therapy Shared Resource core facility that provides de-identified human blood for research purposes.

2.2 |. Manufacture of CTNPs

The CTNPs are manufactured utilizing the liposomal self-assembly of several different lipid molecules. For clot-targeting capability, a combination targeting activated platelets and fibrin was used because these are the major components of a thrombus [10, 47]. In fact, we have previously demonstrated that such combination targeting utilizing hemetromultivalent decotation of ligand motifs on nanoparticles can enhance the targeting specificity and efficacy under a hemodynamic flow environment [33, 34]. For targeting active platelets, the PBP sequence CGSSSGRGDSPA was conjugated to DSPE-PEG₂₀₀₀-Mal through thioether chemistry utilizing the thiol (−SH) group of the peptide cysteine residue. For targeting to fibrin, the FBP sequence cyclo-AC-Y(DGI)C(HPr)YGLCYIQGK-Am was conjugated to DSPE-PEG₂₀₀₀-NHS through amide chemistry via the amine on the lysine. The conjugation products were confirmed using matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectroscopy. For rendering the thrombin-responsive degradation property of the nanoparticles, a thrombin-cleavable lipopeptide conjugate was formed via two-step reaction. First, stearylamine was conjugated to succinimidyl 3-(2-pyridyldithio)propionate (SPDP) in dimethylsulfoxide via amide chemistry utilizing the amine group on stearylamine and the NHS ester on the SPDP. Subsequently, the TCP DVTPRC was conjugated to the pyridyl end of SPDP via disulfide chemistry using the thiol (−SH) group on the TCP. Excess peptide was removed by vacuum filtration and washing the conjugate with deionized water, and the resultant conjugate was purified by liquid–liquid extraction with deionized water and ethyl ether. The conjugate was characterized by MALDI-TOF mass spectroscopy and lyophilized until further use. The thrombin-cleavable property of the conjugate was characterized by exposing the conjugate to 250 nM thrombin in saline at 37 °C for 30 minutes and confirming degradation of the conjugate by MALDI-TOF mass spectroscopy. Following confirmation of all lipid-peptide conjugates, DSPC (44 mol %), DHPE-RhB (1 mol %), cholesterol (20 mol %), thrombin-cleavable lipopeptide conjugate (Stearylamine-TCP, 30 mol %), and clot-targeting peptide-lipid conjugates (DSPE-PEG₂₀₀₀-PBP and DSPE-PEG₂₀₀₀-FBP, 2.5 mol % each), were dissolved in 1:1 chloroform:methanol, and a mixed lipid film was formed using thin film evaporation. The lipid film was rehydrated with saline (for targeting studies only ) or a plasmin solution in saline (425 μg/mL concentration, for fibrinolysis studies), and was sonicated for 1 hour to form CTNPs. Post sonication and lipid film resuspension, the resultant multilamellar vesicles were extruded through polycarbonate membrane filters with 200 nm-sized pores using a liposome extruder to form unilamellar CTNP vesicles. Control (undecorated) nanoparticles were manufactured similarly, but instead of using the DSPE-PEG₂₀₀₀-peptide components, DSPE-PEG-mPEG₂₀₀₀ was used. CTNPs as well as the control nanoparticles were characterized for size using dynamic light scattering (DLS) and cryo-transmission electron microscopy (cryo-TEM). For plasmin-loaded CTNPs used in fibrinolysis studies, unencapsulated plasmin was separated from the nanoparticles using Sephadex G-100 bead columns. Plasmin-loaded CTNPs were incubated with Triton X-100 to assess “exhaustive release” (100% release because of complete particle disassembly). For characterizing plasmin release from CTNPs via diffusion only (without thrombin trigger) as well as via “diffusion + particle destablization” (with thrombin trigger), CTNPs (3 × 10¹¹ nanoparticles per mL) were incubated in 2 mL Eppendorf tubes for 2 hours at 37 °C, in absence or in presence of thrombin (250 nM) and antiplasmin (1 μM). For both conditions, sample aliquots (100 μL) were taken at various time intervals over 2 hours and analyzed by plasmin-antiplasmin (PAP) ELISA to calculate plasmin released. All data pertaining to diffusion-mediated release and thrombin-triggered particle destabilization-mediated release were normalized to the “exhasutive release” data.

2.3 |. Microfluidics-based evaluation of clot targeting by CTNPs

Platelet-rich plasma (PRP) with calcein-stained platelets and 5% v/v AlexaFluor-647 fibrinogen was mixed with 250 nM thrombin in 0.5 M CaCl₂ was incubated on collagen-coated glass slides to form “platelet + fibrin”-rich clots showing blue platelets and green fibrin. The clot-bearing slides were sealed into a GlycoTech PPFC and was washed with saline for 5 minutes. The clot was exposed to Rhodamine B (RhB)-labeled empty CTNPs (peptide-decorated but no plasmin encapsulation) or control nanoparticles (no peptide decoration), in saline under a shear stress of 25 dyn/cm², and were allowed to flow in a closed loop over the clot for 30 minutes. The clot was washed with saline for 15 minutes to remove unbound particles and imaged under inverted epifluorescence microscope. With the imaged area maintained constant for all studies, surface-averaged RhB (red) fluorescence intensity was recorded for control nanoparticle vs. CTNPs bound to the clot, and analyzed statistically to quantify nanoparticle binding.

2.4 |. Fibrinolysis with plasmin-loaded CTNPs in vitro

The fibrinolytic efficacy of plasmin-loaded CTNPs was first evaluated in vitro in a well–plate assay. For this, clots were formed in well plates by incubating PRP (50 μL) with 250 nM thrombin in calcium for 2 hours. The clots were washed with saline to remove residual plasma, and then incubated with free plasmin, plasmin-loaded CTNPs without thrombin or plasmin-loaded CTNPs with thrombin (250 nM), or Tris-HCl buffer (pH 8) for 30 minutes at 37 °C. Sample aliquots (100 μL) were taken at various time intervals (5, 15, and 30 minutes time points), and D-dimer ELISA was performed on the samples to determine the amount of clot degradation. For each treatment group, 3 replicates of this experiment were used for analysis. The targeted fibrinolytic efficacy of the plasmin-loaded CTNPs was then evaluated under flow using the Bioflux microfluidic device. Microfluidic channels were coated with equine collagen type IV and von Willabrand Factor (vWF) by incubating for 1 hour. The channels were washed with saline, and PRP with calcein AM-stained platelets and AlexaFluor-647 fibrinogen was flowed in the channel for 10 minutes at 60 dyn/cm² to form a fibrin-rich clot. The clot was washed with saline, and plasmin-loaded CTNPs with or without clot-targeting capabilities were flowed over the clot for 30 minutes at 25 dyn/cm². Free plasmin in saline as well as plasmin introduced with antiplasmin were used as positive and negative controls, respectively. The clot was imaged using an inverted fluorescent microscope over 0–30 minutes, the images were analyzed for surface-averaged fibrin (green) fluorescence intensity in the fixed channel area, and clot degradation kinetics was measured as a decrease in this fibrin fluorescence intensity over 30 minutes normalized to the fibrin fluorescence at t = 0 min. Additionally, at the 30 minutes time-point for each experiment, the fluid in the outlet well was collected and analyzed by D-dimer ELISA to further quantify fibrin degradation. The studies were run in triplicate for statistical data analysis.

2.5 |. Fibrinolysis with plasmin-loaded CTNPs in vivo in a zebrafish thrombosis model

Zebrafish (Danio rerio) were raised in accordance with animal care guidelines as approved by the University of Michigan Animal Care and Use Committee. Fish strains were acquired from the Zebrafish International Resource Center and all experiments were performed in an AB X TL hybrid. Laser-mediated endothelial injury [48–50] was used to produce occlusive thrombosis at 3 days post-fertilization (3 dpf) in the zebrafish posterior cardinal vein (PCV). Larvae were first anesthetized in tricaine and mounted in 0.8% low melting point agarose on glass coverslips. The agarose around the head of the larvae was removed and 3 nl of CTNPs or 1.75 nl free plasmin (72.9 μg/mL) were infused retro-orbitally via the anterior cardinal vein. For time-to-occlusion (TTO) studies, this treatment administration was performed first, followed by laser injury of the PCV endothelium 5 somites caudal to the anal pore using 99 pulses at power level 18 (MicroPoint Pulsed Laser System, Andor Technology). TTO of the PCV was observed up to 2 minutes and then larvae were checked for successful infusion of CTNPs by the presence of fluorescence in circulation in all assays except for free plasmin. For time-to-recanalization (TTR) studies, venous laser injury was performed first, followed by infusion of CTNPs into the anterior cardinal vein, and TTR observed up to 30 minutes. For control conditions, TTO and TTR experiments were performed without or with empty nanoparticles, as well as with plasmin-loaded clot-targeted nanoparticles (CNPs) that do not have the thrombin-cleavable component.

2.6 |. Statistical analysis

For in vitro clot-binding studies of nanoparticles, paired t-test was performed for statistical analysis based on quantification of nanoparticle RhB intensity bound to the clot, and significance was considered for p < .05. For well plate-based fibrinolysis assays, one-way analysis of variance tests were used for the statistical analysis, and significance was considered for p < .05. Comparisons were carried out between free plasmin versus plasmin-loaded CTNPs with or without thrombin versus Tris-HCl buffer. For BioFlux microfluidic-based fibrinolysis studies and associated D-dimer ELISA, one-way analysis of variance tests were used for the statistical analysis, and significance was considered for p < .05. Comparisons were carried out between free plasmin versus plasmin-loaded CTNPs with or without thrombin versus free plasmin incubated with antiplasmin. For zebrafish thrombosis TTO and TTR studies, statistical comparison of treatment groups were performed using the Mann-Whitney U test.

3 |. RESULTS

3.1 |. Manufacture and characterization of CTNPs

Supplementary Figures S1 and S2 show the peptide-lipid conjugation schematic as well as representative MALDI-TOF mass spectroscopy data for the DSPE-PEG₂₀₀₀-PBP and DSPE-PEG₂₀₀₀-FBP synthesis. Supplementary Figure S3 shows the reaction schematic for the synthesis of the thrombin-cleavable lipopeptide conjugate (TCP-stearate), as well as representative MALDI-TOF mass spectroscopy characterization of the TCP-stearate molecule before and after exposure to thrombin (30 minutes, 250 nM) to confirm thrombin-induced cleavage. The results demonstrate that thrombin could efficiently cleave the conjugate, evidenced by the fact that after thrombin exposure very little of the TCP-stearate conjugate remains (peak at 1046 m/z), whereas 2 new degradation peaks appear (495 and 534 m/z) indicative of the degradation products. Supplementary Figure S4 shows the schematic for plasmin-loaded CTNP manufacture using thin film rehydration and extrusion method, resulting in nanoparticles ~150 to 200 nm in diameter.

Figure 2A shows representative DLS characterization data and cryo-TEM images of the vesicles before extrusion and after extrusion, indicating that after extrusion the unilamellar vesicles had a size of ~150 to 200 nm diameter. Figure 2B shows plasmin loading results across 5 representative CTNP batches, indicating that per batch (3 × 10¹² nanoparticles per mL) the encapsulated plasmin concentration is 443.2 ± 86.3 nM. Figure 2C shows release kinetics of plasmin from such CTNPs, indicating that without exposure to thrombin the diffusive release of plasmin is low (~40% over a 2-hour period), whereas upon exposure to thrombin the release is significantly increased to ~70% during the same period. This enhanced release of plasmin upon exposure of CTNPs to thrombin can be attributed to thrombin-induced cleavage of the TCP-stearate in the CTNP shell, which destabilizes the particles.

FIGURE 2.

(A) Representative dynamic light scattering analysis and cryo-TEM imaging data for CTNPs show that pre-extrusion the multilamellar vesicle diameter is ~1 μm and upon extrusion the unilamellar vesicle diameter is ~150 nm (scale bar in cryo-TEM images is 200 nm); (B) Plasmin encapsulation analysis shows that plasmin can be reproducibly loaded at ~350–500 nM per batch of CTNPs; (C) Release kinetics analysis shows that over a 2-hour period the diffusive release of plasmin from CTNPs is low (~40% release by 120 min), whereas thrombin-triggered release of plasmin from CTNPs is significantly high (~40% by 30 minutes and ~70% by 120 minutes).

3.2 |. Clot-targeting capability of CTNPs in vitro

Supplementary Figure S5 depicts the PPFC microfluidic setup for evaluating the clot-targeting capability of CTNPs under a vascularly relevant shear flow (25 dyn/cm²) environment. Figure 3 shows representative confocal fluorescence images at the experiment end-point (30 min), as well as analyzed quantitative data for surface-averaged RhB fluorescence intensity from these targeting studies. Supplementary Movies M1 (for control nanoparticles) and M2 (for CTNPs) depict representative videos for such targeting studies. As evident from Figure 3, control nanoparticles showed minimal binding to the clots, whereas CTNPs showed substantial binding, co-localizing with both platelets (appearing purple in the image) and fibrin (appearing yellow in the image). Quantitative analysis of the RhB fluorescence intensity clearly indicated that the CTNPs had significantly higher (p ≤ 0.0001) binding to clots, than control nanoparticles. The nonzero binding of control nanoparticles to clots is possibly a result of nonspecific binding as some lipidic nanoparticles can fuse with the platelet cell membrane as well as get physically trapped in the clot mesh. These results indicate that the combination of platelet-binding (via PBP) and fibrin-binding (via FBP) mechanisms renders clot-specific high binding ability of the CTNPs.

FIGURE 3.

Representative fluorescent images and quantitative data showing that compared with control nanoparticles (no peptide decoration), CTNPs can significantly anchor onto clots under flow. Scale bar in fluorescence images: 10 μm; ****p ≤ 0.0001.

3.3 |. Fibrinolysis with plasmin-loaded CTNPs in vitro

Figure 4 shows quantitative data from the D-dimer analysis of the well-plate-based fibrinolysis studies where “platelet +fibrin”-rich clots in wells were incubated for 1-hour with free plasmin (positive control), or saline only (negative control), or plasmin-loaded CTNPs with versus without exogenously added thrombin (to simulate the thrombin-rich environment of clots). Elevated D-dimer is a marker of increased fibrinolysis. The results indicate that over the 30-minute period free plasmin could significantly degrade the fibrin clot (increasing D-dimer level), whereas saline was unable to cause any substantial clot degradation (low D-dimer value). The plasmin-loaded CTNPs rendered only modest degradation of clots when thrombin was not added exogenously to the wells, but in the presence of thrombin the plasmin-loaded CTNPs rendered significantly higher fibrinolysis. This suggests that in absence of thrombin the small amount of diffusively released plasmin from CTNPs can cause a modest extent of fibrinolysis, but the fibrinolytic effect is significantly enhanced when plasmin release from CTNPs is increased by thrombin trigger.

FIGURE 4.

D-dimer ELISA analysis using a well-plate-based assay shows that plasmin-loaded CTNPs can render modest levels of fibrinolysis over 30 minutes when plasmin is diffusively released and this fibrinolytic effect is significantly increased when release of plasmin is enhanced by exogenously added thrombin trigger; Treatment with free plasmin added directly was used as positive control and treatment with saline was used as negative control. *p ≤ 0.05, **p ≤ 0.01.

Building on the above-described demonstration of plasmin-loaded CTNPs to render increased fibrinolysis in presence of thrombin, subsequent studies were performed to evaluate whether this capability of CTNPs is conserved under a simulated vascular flow environment. For this, a BioFlux microfluidic system was used (schematically shown in Supplementary Figure S6) where PRP clots (blue platelets, green fibrin) were formed by 2-hour incubation in “collagen + vWF”-coated microchannels, washed with saline to remove residual liquid, and then exposed to flow of free plasmin (positive control), or “free plasmin + antiplasmin” added together (negative control), or plasmin-loaded CTNPs without versus with exogenously added thrombin (to simulate a thrombin-rich clot environment). Clot lysis was imaged for 30 minutes, and the loss of fibrin fluorescence (green) over time was assessed as an indicator of fibrinolysis. Figure 5A shows representative images from these studies under various treatment groups over the 30-min period, and Figure 5B shows kinetic analysis of “fibrin fluorescence loss” over the 30-min period in the imaged microfluidic area. Supplementary Movies M3, M4, M5, and M6 show representative videos from BioFlux microfluidic studies of flowing “free plasmin, ” “free plasmin + antiplasmin,” “plasmin-loaded CTNPs with thrombin,” and “plasmin-loaded CTNPs without thrombin,” respectively, over clots. For M5 and M6, the videos also show the RhB (red) fluorescence of CTNPs along with platelets (blue) and fibrin (green). Figure S7 shows “fibrin fluorescence intensity” data from image analysis in the channel area in these experiments at various timepoints over the 30-minute period.

FIGURE 5.

(A) Representative fluorescence images of clots (blue: platelets, green: fibrin) under flow of various treatment groups in the BioFlux microfluidic channels show that plasmin-loaded CTNPs in the presence of exogenously added thrombin can rapidly lyse clots (loss of green fibrin over time) at a level similar to free plasmin effect; (B) Kinetic analysis of “fibrin fluorescence loss” confirms that plasmin-loaded CTNPs in the presence of exogenously added thrombin rapidly lyse clots at a level similar to free plasmin effect, whereas plasmin in the presence of exogenously added antiplasmin or plasmin-loaded CTNPs in the absence of exogenously added thrombin render minimal fibrinolysis; (C) D-dimer analysis of lysis products from the outlet well of BioFlux experiments after the 30-min time-point confirm the enhanced fibrinolytic effect of plasmin-loaded CTNPs in presence of exogenously added thrombin. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

As evident from these results, free plasmin was able to rapidly lyse the clot (sharp drop in fibrin fluorescence), and this effect was majorly inhibited when antiplasmin was added to the system. Plasmin-loaded CTNPs without thrombin exposure caused minimal fibrinolysis, whereas on exposure to thrombin these particles were able to render substantial clot lysis, comparable to the effect of free plasmin. As an additional comparison, plasmin-loaded clot-targeted nanoparticles without the thrombin-cleavable lipid component in the particle shell were prepared (termed clot-targeted nanoparticle or CNP) and flowed over clots in presence of exogenously added thrombin, in the BioFlux channels. Thus, “plasmin-loaded CTNPs without thrombin exposure” and “plasmin-loaded CNPs with thrombin exposure” were the 2 complementary groups to assess whether the thrombin-triggered plasmin release property is necessary for enhanced fibrinolysis. The comparison data are shown in Supplementary Figure S8A where the “fibrin fluorescence loss” from treatment of the clots with plasmin-loaded CTNPs without thrombin or plasmin-loaded CNPs with thrombin were both minimal compared with treatment with plasmin-loaded CTNPs plus thrombin. Supplementary Movie M7 shows a representative video from BioFlux studies of flowing plasmin-loaded CNPs with thrombin over clots. These results strongly indicate that the thrombin-triggered enhanced release of plasmin from CTNPs can render significantly higher fibrinolysis compared with when such triggered release is not present and only diffusive release is present.

Additional analyses were done by collecting the lysis products from the outlet well at the 30-minutes time-point for the various treatment groups and analyzing D-dimer concentration, as shown in Figure 5C. These results further indicate that plasmin-loaded CTNPs in the absence of thrombin exposure resulted in low fibrinolysis (low D-dimer values), whereas in the presence of thrombin these CTNPs rendered significantly higher fibrinolysis. Supplementary Figure 8B shows additional comparisons from D-dimer analysis of the lysis products collected from these experiments, further confirming that the low amount of plasmin diffusively released from CTNPs in the absence of thrombin or from CNPs in the presence of thrombin is unable to cause substantial lysis under a flow environment, whereas the enhanced release of plasmin from CTNPs in the presence of thrombin can increase fibrinolysis.

3.4 |. Fibrinolysis with plasmin-loaded CTNPs in vivo in the zebrafish model

Figure 6A shows the general setup for the zebrafish studies where laser-induced endothelial injury in the PCV was used to create occlusive venous thrombi. It is well-reported that such laser-induced venous thrombosis in zebrafish is fibrin-rich [51, 52], and thus, it was considered an appropriate model to study fibrinolytic effect of plasmin-loaded CTNPs. For TTO studies, the treatments were administered before laser injury, with the rationale that the fibrinolytic effect of the treatment, if any, would prevent (or delay) vessel thrombo-occlusion. For TTR studies, the treatments were administered after endothelial injury and clot formation, testing the ability of the treatment to render fibrinolysis and subsequent recanalization. Figure 6B shows results from TTO studies and Figure 6C shows results from TTR studies in the zebrafish tyhrombopsis model. As evident from Figure 6B, larvae without treatment or with “control nanoparticle treatment” exhibited rapid vessel occlusion (in <25 seconds). Interestingly, this was also observed for “free plasmin treatment” group, likely because the plasmin is rapidly inhibited on administration thereby preventing any therapeutic action. In contrast, treatment with plasmin-loaded CTNPs were able to significantly prevent clot formation such that there was no vessel occlusion observed for 120 sec (final time-point of experiment).

FIGURE 6.

(A) Experimental setup for plasmin-loaded CTNP evaluation in the zebrafish venous thrombosis model, with representative brightfield images showing clot formation after laser injury; (B) Time-to-occlusion (TTO) study results showing that without treatment or with “control nanoparticle” treatment the vessel occluded rapidly (< 25 sec) and this was also observed for “free plasmin” treatment likely because plasmin is rapidly inhibited on administration thereby preventing any therapeutic action; in contrast, treatment with plasmin-loaded CTNPs was able to significantly prevent clot formation such that vessel occlusion was not observed up to 120 sec; (C) Time-to-recanalization (TTR) study results showing that without treatment or with “control nanoparticle” treatment no recanalization was observed for 30 min, whereas with plasmin-loaded CTNP treatment the vessel was recanalized in <20 minutes because of effective fibrinolysis. ***p ≤ 0.001; ****p ≤ 0.0001.

These results indicate that encapsulating the plasmin within CTNPs protected it from rapid inhibition, and plasmin was released from CTNPs as the particles started binding to the clot site and undergoing thrombin-triggered destabilization. Supplementary Movies M8, M9, M10, and M11 depict representative videos for TTO studies in the zebrafish model, for “free plasmin treatment,” “no treatment,” “plasmin-loaded CTNP treatment,” and “control particle treatment” conditions, respectively. As evident from Figure 6C, larvae without treatment or with “control nanoparticles” showed no sign of recanalization within 30 min, whereas larvae with plasmin-loaded CTNPs recanalized in <20 minutes. This suggests that plasmin released from clot-anchored CTNPs was able to render effective fibrinolysis to recanalize the vessel. Supplementary Movies M12, M13, and M114 depict representative videos for TTR studies in the zebrafish model, for “no treatment,” “plasmin-loaded CTNP treatment,” and “control particle treatment” conditions, respectively. Additional TTO and TTR study results comparing plasmin-loaded CTNPs (thrombin-triggered mechanism present) versus plasmin-loaded CNPs (thrombin-triggered mechanism absent) in the zebrafish model are shown in Supplementary Figures S8C and S8D. These data confirm that the thrombin-triggered enhanced plasmin release mechanism is necessary to achieve fibrinolysis, compared with when low amounts of plasmin are relased only by diffusion.

4 |. DISCUSSION

Current clinical approaches for fibrinolytic therapy use tissue plasminogen activators (alteplase, reteplase, etc.), which work by converting plasminogen to plasmin that can then break down fibrin. Although the plasminogen to plasmin conversion by such drugs is significantly enhanced for fibrin-associated plasminogen, they can also activate circulating plasminogen leading to systemic fibrinogenolysis, resulting in significant hemorrhagic risks. A safer and more efficacious approach could be direct use of plasmin, but the biggest challenge here is the very short (< 1 second) circulation lifetime of plasmin. Based on these considerations, here we explored whether plasmin can be encapsulated within an intravenously injectable nanoparticle system (thus protected from rapid neutralization in circulation), and delivered specifically at the clot site by engineering the nanoparticle to be capable of (1) specifically binding to activated platelets and fibrin (for clot-anchorage), and (2) then destabilizing under the action of thrombin (upregulated at thrombus site) for clot-localized payload release. Of relevance, we have successfully demonstrated such “clot-targeted enzyme-responsive drug delivery” approach in the past by engineering nanoparticles that can bind to clot-associated platelets and deliver drugs like Streptokinase (for fibrinolysis) or tranexamic acid (for fibrinolysis inhibition) via clot-associated phospholipase A2(PLA₂)-triggered particle destabilization [33, 53].

For plasmin-loaded CTNP construction, we utilized heteromultivalent surface-decoration of liposomes with 2 peptides, one binding to activated platelet integrin αIIbβ3 and the other binding to fibrin. Furthermore, we developed a unique lipopeptide by conjugating a TCP to stearylamine (thus forming TCP-stearate), and this was combined with the clot-targeting lipid-peptide components for the final CTNP construction. The design rationale here is that stearylamine has an equivalent length of hydrocarbon tail as the other 2 major lipids (DSPC and DSPE) of our liposomal construct and thus it can assemble efficiently within the liposomal membrane with the hydrophilic TCP motif remaining close to the particle surface for accessibility by thrombin. As the CTNPs bind to the clot site via PBP- and FBP-mediated anchorage, the thrombin can cleave the TCP motif off TCP-stearate, rendering the nanoparticle membrane unstable (due to loss of amphiphilicity) and thus release the payload (plasmin) locally. Our studies demonstrated that plasmin-loaded CTNPs can be manufactured reproducibly. Our studies further showed that the diffusive release of plasmin from the CTNPs over 2-hour is low, whereas thrombin-triggered destabilization of CTNPs significantly enhances this release. Our in vitro microfluidic studies demonstrated that CTNPs can actively anchor onto “platelets + fibrin”-rich clots under shear flow, and the thrombin-triggered plasmin release from CTNPs can enhance the “targeted fibrinolysis” capability. Subsequently, therapeutic studies in a zebrafish model further demonstrated that the fibrinolytic capability of the CTNP-delivered plasmin is maintained in vivo for effective clot prevention (preventing occlusion) and clot dissolution (accelerating recanalization).

Our current studies have a few potential limitations that need to be addressed in the future. First, for our studies, we have used only a single dose of CTNPs. Although our studies have shown efficacy with this dose, future studies will need to evaluate the pharmacology and toxicology profile of these nanoparticles at escalating and recurrent doses in appropriate animal models. Second, future studies will need to assess potential immunogenicity of CTNPs, at recurrent and escalated doses for translational advancement. Of note, several liposome-based nanoparticle formulations have been clinically translated successfully [54], and in other studies, small peptides have been shown to elicit reduced immunogenicity [55]. Therefore, a technology like CTNP that utilizes liposomal platform decorated with small peptides, may have exciting translational potential. Third, detailed evaluation will also need to be done to confirm that such recurrent or escalated doses do not pose any systemic bleeding risk similar to that posed by current tPA therapy. Fourth, further evaluation will also be needed regarding potential effects of the delivered plasmin on other tissues beyond the intended fibrinolytic effect. Besides fibrin, plasmin can act on other substrates including coagulation factors, complement, cell surface, and extracellular matrix proteins, with potential effects on immune and inflammatory mechanisms, wound healing, tissue remodeling, etc. [56, 57]. It is important to note here that the nonspecific diffusional release of plasmin from CTNPs is quite low and this diffusively released plasmin will likely be neutralized very rapidly by antiplasmin in circulation and thereby not have the ability to render nonspecific effects. A much higher extent of encapsulated plasmin is released from CTNPs only upon thrombin-triggered destabilization, and we have shown that this clot-localized release of plasmin is the dominant mediator of fibrinolysis rather than nonspecific effects. Nonetheless, the diverse effects of released plasmin may need to be further analyzed for the CTNPs. Considering the fact that several lipid nanoparticle-based therapeutic formulations (including recent Moderna and Pfizer vaccines for COVID-19) continue to be translated to clinical use [58–60], it can be envisioned that a nanomedicine formulation like plasmin-loaded CTNP could potentially advance as a safe and effective approach for targeted fibrinolytic applications.

5 |. CONCLUSION

We demonstrate the feasibility of encapsulating and delivering plasmin as a direct fibrinolytic agent by constructing a nanomedicine platform that can actively target to clot site and deliver an encapsulated payload in a thrombin-responsive manner. Our in vitro studies establish the reproducibility of manufacture and the mechanism of action of this plasmin-loaded CTNP system. Our in vitro and in vivo studies demonstrate that encapsulation within CTNPs can provide a way to protect plasmin from rapid neutralization in circulation and then to deliver this plasmin specifically to a clot for thrombin-triggered release to render localized fibrinolysis. If successfully translated, such an approach can provide an efficacious way of directly using plasmin for fibrinolytic therapies, and potentially avoiding the systemic side-effects of current fibrinolytic approaches involving plasminogen activators.

Essentials.

• Clot-targeted direct delivery of plasmin can improve safety and efficacy of fibrinolytic therapy.

• Plasmin was loaded in clot-targeted thrombin-cleavable nanoparticles (CTNPs) for evaluation.

• In vitro studies with CTNPs showed thrombin-triggered plasmin release for efficient fibrinolysis.

• In vivo studies with CTNPs in zebrafish clot model showed successful therapeutic feasibility.