Skip main navigation

Chymase Inhibition Resolves and Prevents Deep Vein Thrombosis Without Increasing Bleeding Time in the Mouse Model

Originally published of the American Heart Association. 2023;12:e028056



Deep vein thrombosis (DVT) is the primary cause of pulmonary embolism and the third most life‐threatening cardiovascular disease in North America. Post‐DVT anticoagulants, such as warfarin, heparin, and direct oral anticoagulants, reduce the incidence of subsequent venous thrombi. However, all currently used anticoagulants affect bleeding time at various degrees, and there is therefore a need for improved therapeutic regimens in DVT. It has recently been shown that mast cells play a crucial role in a DVT murine model. The underlying mechanism involved in the prothrombotic properties of mast cells, however, has yet to be identified.

Methods and Results

C57BL/6 mice and mouse mast cell protease‐4 (mMCP‐4) genetically depleted mice (mMCP‐4 knockout) were used in 2 mouse models of DVT, partial ligation (stenosis) and ferric chloride–endothelial injury model of the inferior vena cava. Thrombus formation and impact of genetically repressed or pharmacologically (specific inhibitor TY‐51469) inhibited mMCP‐4 were evaluated by morphometric measurements of thrombi immunochemistry (mouse and human DVT), color Doppler ultrasound, bleeding times, and enzymatic activity assays ex vivo. Recombinant chymases, mMCP‐4 (mouse) and CMA‐1 (human), were used to characterize the interaction with murine and human plasmin, respectively, by mass spectrometry and enzymatic activity assays. Inhibiting mast cell–generated mMCP‐4, genetically or pharmacologically, resolves and prevents venous thrombus formation in both DVT models. Inferior vena cava blood flow obstruction was observed in the stenosis model after 6 hours of ligation, in control‐ but not in TY‐51469–treated mice. In addition, chymase inhibition had no impact on bleeding times of healthy or DVT mice. Furthermore, endogenous chymase limits plasmin activity in thrombi ex vivo. Recombinant mouse or human chymase degrades/inactivates purified plasmin in vitro. Finally, mast cell–containing immunoreactive chymase was identified in human DVT.


This study identified a major role for mMCP‐4, a granule‐localized protease of chymase type, in DVT formation. These findings support a novel pharmacological strategy to resolve or prevent DVT without affecting the coagulation cascade through the inhibition of chymase activity.

Nonstandard Abbreviations and Acronyms


human chymase


ferric chloride


inferior vena cava


mouse mast cell protease‐4


recombinant human chymase


recombinant mouse mast cell protease‐4


wild type

Clinical Perspective

What Is New?

  • The chymase‐specific inhibitor TY‐51469 resolves or prevents venous thrombi in a stenosis model and reduces ferric chloride–induced thrombus formation in the mouse inferior vena cava.

  • Inhibition of chymase does not impact bleeding times or coagulation in wild‐type mice.

  • Chymase reduces plasmin activity within thrombi of deep vein thrombosis mice.

  • Human chymase immunoreactivity is localized in thrombi from patients with deep vein thrombosis, and recombinant human chymase hydrolyzes specific peptide bonds within the active site of plasmin, thus inhibiting its activity in vitro.

What Are the Clinical Implications?

  • Bleeding observed with current anticoagulant therapies remains a common complication.

  • The present study suggests that inhibition of intrathrombus‐located chymase is a new profibrinolytic strategy with no impact on hemostasis.

  • This is also the first nontissue plasminogen activator–dependent strategy to pharmacologically resolve already formed thrombi in an experimental model of inferior vena cava stenosis.

Deep vein thrombosis (DVT), a major cause of venous thromboembolism, is commonly associated with “flow‐impeding” clots located predominantly in major veins.1, 2 DVT in humans requires medical consultation with rapid diagnosis and potential therapy, to reduce the risk of complications, such as pulmonary embolism or post‐thrombotic syndrome.3 Numerous risk factors, such as cancer or cancer therapy, aging, obesity, diabetes, smoking, pregnancy, birth control medications, defective coagulation system, and family genetics, are involved in the incidence and severity of the disease.1, 3, 4 Health care costs associated with DVT and pulmonary embolism annually exceed $10 billion in the United States.5

A suspicion of a DVT by health care professionals is commonly confirmed by ultrasound imaging of the affected area and by measuring plasma concentrations of D‐dimer, a marker for polymeric fibrin degradation complexes.6 Appropriate treatment strategies are often implemented according to a risk assessment model to adapt a treatment regimen to the risk level of the patient. It can be either limited to 3 months to reduce the existing thrombus burden or indefinitely prolonged to reduce the risk of recurrent events as well as the risk of more severe complications, such as pulmonary embolism.6 Low‐molecular‐weight heparins, such as enoxaparin, vitamin K antagonists, such as warfarin, and direct oral anticoagulants, such as rivaroxaban or apixaban, are indicated and recommended for the treatment or prevention of DVT, depending on different factors.2, 6, 7 However, these medications share important limitations. As anticoagulants, they are thought to act indirectly on the degradation of fibrin by shifting the balance between coagulation and fibrinolysis in favor of the latter, which requires extended time and is less efficient in comparison with direct resolution of clot material. Irrespective of their exact mechanisms of action however, intake of all of these anticoagulants leads to increased bleeding risks.2, 7, 8

Although thrombolytics, such as tPA (tissue‐type plasminogen activator), may dissolve clots directly and more effectively, increase the patency of the veins, and reduce the incidence of post‐thrombotic syndrome in clinical studies, higher risks for bleeding complications are prevalent, strongly limiting the target population to patients fulfilling strict eligibility criteria.9, 10 In addition, treatment with thrombolytics is limited to the hospital setting. Thus, there is still a high medical need for efficacious and safe dissolution strategies of thrombi in veins.

Recently, Ponomaryov et al reported that mast cells are dynamically involved in the formation of DVT in a stenosis mouse model.11 In support for this, mast cell–deficient KitW/W‐v or KitW‐sh/W‐sh mice did not develop DVT after ligation of the inferior vena cava (IVC) when compared with wild‐type (WT) mice.11

Derived from hematopoietic cells in the bone marrow as well as the spleen, mast cells circulate in the blood as precursors before homing to tissues where their maturation is completed. Mast cells may mature from resident precursors preexisting at different tissues, such as the heart,12, 13 lungs,14 and blood vessels.15 Growth factors, such as stem cell factor, are involved in the maturation of mast cells in selected tissues.16 In humans and rodents, connective tissue and mucosal mast cells have been identified.17 Increased mast cell numbers and densities are reported in myocardial failure, atherosclerosis, and ischemia, and mast cells are instrumental in transplant‐related fibrosis.16, 18, 19 Activated mast cells are involved in tissue repair and inflammatory processes, such as wound healing/fibrosis, cardiac remodeling, and angiogenesis16, 20, 21, 22 Mast cells contain cytoplasmic granules, which, on stimulation, release several autacoids, cytokines, proteoglycans, and proteases, the latter including chymase, tryptase, cathepsin G, and carboxypeptidase A3.16 Of relevance for the present study, mast cells are degranulated by hypoxia, which also occurs in DVTs.23, 24 To date, only one single chymase, α‐chymase (human chymase [CMA‐1]), has been reported in humans, whereas in mice, 4 major chymases are expressed, of which mouse mast cell protease‐4 (mMCP‐4) appears to be the functional ortholog to the human chymase (CMA‐1).18

Over the past decades, several selective chymase inhibitors, such as TY‐51469, have been developed and reported useful against cardiovascular diseases, such as atherosclerosis, heart failure, and myocardial infarction in animal models.18, 25 However, whether genetic repression or inhibition of mast cell–derived chymase also represses venous thrombus formation remains an open question. The latter issue will be addressed in this study, by using mouse models of venous thrombosis caused by either partial ligation of the IVC or ferric chloride (FeCl3)–induced injury of the vascular wall.

Our results reveal that chymase inhibition may be a novel therapeutic target for efficacious and safe dissolution and prevention of venous thrombi.


See Data S1 for the Supplemental Methods. The authors declare that all supporting data and methods are available in the article (and its Supplemental Material). Data supporting the findings of this study are available from the corresponding author on reasonable request.


C57BL/6 mice (n=225) were purchased from Charles River (Montréal, QC, Canada), and the colonies were maintained and bred in our animal facilities. The mMCP‐4 knockout mice (n=50) were also bred as previously described.26 The mMCP‐4 knockout mice were backcrossed for >18 generations with C57BL/6 congeners and are therefore highly congenial with the later strain.26 Female and male mice (aged 8–10 weeks; weight, 18–25 g) were used for the experiments. The animals’ conditions were regulated at constant room temperature (23°C) and controlled humidity of 78%. They were also under a controlled cycle of light/dark (6 am to 6 pm) with standard chow and tap water available ad libitum. Animal care and experimentations were performed in accordance with the Guide for the Care and Use of Laboratory Animals, as adopted and promulgated by the US National Institutes of Health, and were approved by the Ethics Committee on Animal Research of the Université de Sherbrooke in accordance with the guidelines of the Canadian Council on Animal Care.


The chymase inhibitor, TY‐51469 (2‐(4‐((5‐fluoro‐3‐methylbenzo[b]thiophene)‐2‐sulfonamido)‐3‐(methylsulfonyl)phenyl)thiazole‐4‐carboxylic acid), was obtained from Toa Eiyo Limited (Osaka, Japan) and was dissolved in a vehicle solution of 0.025 M NaOH diluted in 10 mM PBS (pH 7.4). The plasmin/plasminogen inhibitor, BAY 1214237 (10‐chloro‐2‐oxo‐1,2‐dihydropyrimido[1,2‐b] indazol‐4‐yl) piperidinium chloride, was obtained from Bayer AG (Wuppertal, Germany) and was dissolved in 10 mM PBS (pH 7.4). BAY 1214237 inhibits human purified plasmin hydrolytic activity with a half maximal inhibitory concentration of 4.43 μM (results not shown).

Mouse FeCl3‐Induced Endothelial Insult Model

The FeCl3‐injury model was performed as previously described.27 WT male mice were treated with vehicle, TY‐51469 (10 mg/kg), BAY 1214237 (1 mg/kg), or a coadministration of TY‐51469 and BAY 1214237 (10 and 1 mg/kg, respectively) 1 hour before anesthesia with ketamine/xylazine (87/13 mg/kg intramuscular). A laparotomy was performed, and the guts were gently displaced without injury from the abdominal cavity and conserved in 0.9% NaCl solution (37°C). The IVC was exposed by separation of the vessel between the renal and iliolumbar veins. A 2×2‐mm double‐layered absorbent gauze saturated in 0.37 M (10% w/v in sterile water) FeCl3 (Sigma‐Aldrich, St. Louis, MO) was applied directly on the isolated IVC for 3 minutes and then removed.27 The guts were subsequently reinserted in their original location in the abdomen. Mice were euthanized 30 minutes later to access the thrombotic area in the IVC. IVCs were dissected and thrombi were isolated and weighted with an analytical balance (0.1 mg accuracy; Entris II; Sartorius).

Ligation Surgery

As previously described by Ponomaryov et al and Canobbio et al,11, 28 stenosis of the IVC was performed in anesthetized mice (mixture of 2% isoflurane and 2.5% oxygen). A laparotomy was performed, and the guts were gently displaced without injury from the abdominal cavity and conserved in 0.9% NaCl solution (37°C). The IVC was exposed by isolating the vessel between the renal and iliolumbar veins. For partial ligation, a 30‐G needle was used as a spacer. None of the side and back branches of the IVC was ligated.28 After the ligation with polypropylene 7.0 sutures (FST, Foster City, CA) around both the IVC and the spacer, the latter was removed gently. The latter ensured that the endothelium is not denuded and that a residual flow of 10% in the IVC remains.11 Guts were subsequently moved back in their original location in the abdomen, and the abdominal cavity was closed in layers with a braided 5.0 absorbable suture. The analgesia protocol was maintained for 24 hours after surgery with buprenorphine (0.1 mg/kg subcutaneous every 6–9 hours). The stenosis was evaluated 24 or 48 hours after ligation. The IVCs were dissected after euthanasia, and thrombi were isolated, measured, weighed, and finally stored at −80°C.

Preligation Treatment With TY‐51469

Mice were pretreated with consecutive intraperitoneal doses of 10 mg/kg TY‐51469,11, 29 according to the following scheme: 48 hours, 24 hours, and 30 minutes before IVC ligation. Animals were euthanized 24 hours after ligation. In a separate series of experiments, no differences in thrombus morphometrics (weight and length) were found in non‐treated or vehicle‐treated male (results not shown) female DVT mice (nonvehicle treated: 6.1±0.98 mg and 3.69±0.58 mm, n=10; vehicle pretreated: 5.78±0.77 mg and 3.79±0.67 mm, n=6).

Postligation Treatment With TY‐51469

Single doses of either vehicle or TY‐1469 (0.1, 1, or 10 mg/kg) were administered intraperitoneally, 1, 6, or 24 hours after ligation of the IVC. Animals were euthanized 24 hours (for the 1 and 6 hours after ligation treatment) or 48 hours (for the 24 hours after ligation treatment) later to examine the thrombus formation.

Tail‐Bleeding Time

Mice were anesthetized with intramuscular ketamine/xylazine (87/13 mg/kg). A dose of 300 U/kg of unfractionated heparin (as specified by the manufacturer, Pfizer Canada Inc) or 10 mg/kg of TY‐51469 was injected 1 hour before the test through the jugular vein or intraperitoneally, respectively. DVT mice were evaluated 24 hours after ligation, with or without a single dose of TY‐51469 (10 mg/kg) 1 hour after ligation. Animals were then immobilized in a vertical position with their tail in a 2‐mL tube filled with 1.5 mL Drabkin reagent (Sigma Aldrich Canada Co, Oakville, ON, Canada).30 The tail was cut 3 mm starting from the extremity, and the time of bleeding cessation was noted. After 30 minutes, all animals were euthanized.30

Murine Doppler Ultrasound

A high‐frequency ultrasound imaging system (Vevo 3100; Fujifilm VisualSonics Inc) equipped with MX400 linear array transducer (40 MHz) was used to visualize by Doppler color blood flow mapping of the IVC and abdominal aorta in the stenosis mouse model of DVT. Female mice were anesthetized during the procedure (2% isoflurane; 98% oxygen; 1.5 L/min), and their abdomens were shaved and covered with ultrasound gel. Imaging of the vessels was performed 12 hours before ligation and 6 and 24 hours after ligation in female WT mice with or without 10 mg/kg TY‐51469 1 hour after DVT. Following the last readings, the animals were euthanized and examined for thrombi formation as well.


Immunochemistry was performed on healthy and ligated IVC of WT and mMCP‐4 knockout mice. Thrombus‐containing vessels were fixed with 10% buffered formalin solution, processed, paraffin embedded, and finally sliced (4‐μm slice) at the service of the histopathology platform at Research Institute–McGill University Health Centre (Montréal, QC, Canada). Mast cells were stained with toluidine blue, and mMcpt4 was detected using goat polyclonal anti‐MCPT4 antibody (US Biological Life Science; catalog number M2414‐20A; 1×400 dilution) and biotinylated horse anti‐goat IgG (Vector Laboratories; catalog number BA9500; 1×200 dilution).

Human venous thrombi premade slides were obtained from Tissue for Research Ltd (Ellingham, Suffolk, UK). Mast cell chymase was detected using goat polyclonal IgG to human MCPT4 (Abcam; catalog number ab111239; 1×200 dilution).

Plasmin Extraction ex Vivo

Plasmin‐containing homogenates were extracted from thrombi of ligated WT mice. Thrombi were homogenized in 10 volumes (w/v) of 20 mM PBS (pH 7.4) supplemented with 1.5 mg/mL of BSA using a glass‐Teflon homogenizer. The homogenates were transferred to 1 mL ultracentrifuge tubes and centrifuged at 100 000g for 20 minutes at 4°C.31 Supernatants were kept at 4°C until they were tested for plasmin enzymatic activity the same day.

Chymase Extraction ex Vivo

Chymase‐containing homogenates were extracted from thrombi of ligated WT mice. Thrombi were homogenized in 10 volumes (w/v) of 20 mM PBS (pH 7.4) supplemented with 0.5 mg/mL of BSA using a glass‐Teflon homogenizer.32 The homogenates were centrifuged at 18 000g for 30 minutes at 4°C, and the supernatants were discarded. Pellets were then resolubilized in the same initial volume of PBS. These last 2 steps were repeated 2 more times, except after the last centrifugation, supernatants were conserved.32 Supernatants were subsequently maintained at 4°C until they were tested for chymase enzymatic activity on the same day following extraction.

Measurement of Intrathrombus Plasmin or Chymase Enzymatic Activity

A total of 30 μL of supernatants from thrombus extracts was used and placed directly in white 96‐well plates. No variations in protein concentration per volume unit of thrombi were found using the Bradford protein assay (result not shown). Some homogenates were treated with either vehicle or 10 μM of TY‐51469 (volume of 5 μL) immediately before the fluorescence readings. A total of 65 to 70 μL of 20 mM PBS (pH 7.4) supplemented with 1.5 mg/mL of BSA (for plasmin) or 0.5 mg/mL of BSA (for chymase) was added to each well to achieve a final volume of 100 μL. Plasmin activity was determined by the hydrolysis rate of 50 μM of the fluorogenic substrate, D‐Ala‐Leu‐Lys‐7‐amino‐4‐methylcoumarin (Sigma Aldrich) at 37°C for 1 hour. Chymase activity was determined by the hydrolysis rate of 10 μM of the fluorogenic substrate, Suc‐Leu‐Leu‐Val‐Tyr‐7‐amino‐4‐methylcoumarin (Peptide Institute Inc, Osaka, Japan) at 37°C for 1 hour. The fluorescence released was measured with an Infinite M1000 spectrophotometer (Tecan Austria GmbH, Grödig, Austria) with λex=370 nm and λem=460 nm. A standard curve of 7‐amino‐4‐methylcoumarin was also performed to quantify the concentration (nM) of the fluorogenic substrate cleavage.

Statistical Analysis

All data are presented as the mean±SEM. All statistical analyses were conducted using the GraphPad Prism 9 software (GraphPad Software, La Jolla, CA). Statistical significance was determined on the basis of 1‐way ANOVA, followed by Bonferroni analysis (for multiple groups) or Mann‐Whitney test (for comparison between 2 groups). Statistical significance was reached when the P value was <0.05, and 1 symbol (*) denotes P<0.05, 2 symbols (**) denote that P<0.01, and 3 symbols (***) denote that P<0.001.


Short‐Term Treatment With a Specific Chymase Inhibitor Prevents DVT Formation in WT Mice

Single‐dose administration of TY‐51469, a specific chymase inhibitor,18 1 hour post‐ligation in the stenosis model reduced thrombus formation in a dose‐dependent manner (Figure 1A and 1B; see experimental design in Figure S1). Both the weights and lengths of thrombi were reduced by >95% in TY‐51469–treated male and female DVT mice at the dose of 10 mg/kg (Figure 1A and Figure S2A).

Figure 1. Short‐term single dose of TY‐51469 protects against deep vein thrombosis (DVT) formation in a dose‐dependent manner without bleeding adverse effects.

A, Dose‐dependent effect on the weight (mg; i) and length (mm; ii) of thrombi obtained 24 hours after DVT from inferior vena cava–ligated male wild‐type (WT) mice treated with 0.1, 1, and 10 mg/kg chymase inhibitor TY‐51469 1 hour post‐ligation. B, Dose‐dependent impact of chymase inhibitor TY‐51469 on the thrombus incidence in WT mice. C, Bleeding times in male mice, as assessed by a tail‐bleeding assay, reveal no difference between WT mice with or without 10 mg/kg TY‐51469, in mouse mast cell protease‐4 (mMCP‐4) knockout (KO) mice, or DVT WT mice with or without 10 mg/kg TY‐51469. The 300 U/kg heparin‐treated WT mice were used as positive controls. Each group of mice represents the mean±SEM (n=6–13). *P<0.05, **P<0.01, and ***P<0.001 compared with WT control group (asterisks directly placed over the error bar of each column); comparisons were also made between treated groups.

No Changes in Hemostasis Are Observed in Chymase‐Deficient Mice or WT Treated Mice With Chymase Inhibitor, TY‐51469, in Vivo and in Vitro

Considering that the chymase inhibitor TY‐51469 resolves or prevents IVC ligation–induced DVT, effects of the chymase inhibitor or genetic chymase deficiency on hemostasis parameters were investigated in both male and female mice. Bleeding times were not increased in WT mice treated with the chymase inhibitor TY‐51469 (10 mg/kg) or in mMCP‐4 knockout mice of either sex (Figure 1C and Figure S2C). As a positive control, the bleeding times in mice treated with unfractionated heparin (300 U/kg) increased by 5‐fold (25±1.2 minutes) when compared with vehicle‐treated WT control male (3.6±0.8 minutes) (Figure 1C) or female mice (Figure S2C). Finally, prothrombin times as well as standard thromboelastographic parameters remained unchanged in blood recovered from WT mice treated with TY‐51469, as well as in mMCP‐4 knockout mice (Figures S2C and S3).

Chymase Inhibition Allows Complete Venous Blood Flow in Ligated IVC of WT Mice

Next, were correlated, by laser Doppler‐color mapping and blood velocity monitoring, the TY‐51469–dependent reduction of thrombus size (Figure 1) with the restoration of IVC blood flow in the mouse stenosis model (Figure 2). DVT was associated with a marked increase in abdominal aortic flow33 and full stasis of vena cava blood flow, from 6 to 24 hours after ligation (Table 1). In contrast, a single dose of TY‐51469 (10 mg/kg, 1 hour after ligation) fully restored IVC blood flow in a sustained manner for at least 24 hours post‐ligation in DVT mice (Figure 2 and Table 1). Furthermore, no changes were observed in the abdominal aortic flow in TY‐51469–treated mice within the 24‐hour time frame (Figure 2 and Table 1).

Figure 2. Short‐term chymase inhibition reinstates vena cava blood flow in deep vein thrombosis (DVT) mice.

Doppler color blood flow mapping of the inferior vena cava (IVC) and adjacent abdominal aorta (AA) in 8‐week‐old female mice. TY‐51469 (10 mg/kg) was administered intraperitoneally 1 hour after ligation. The image illustrates the apical view of the IVC and AA with the ultra‐high‐frequency linear array transducer (or scan head) tilted at an angle of 35±10 degrees from the posterior end of the mouse. DVT‐induced IVC blood stasis in control WT mice 6 hours after ligation. WT mice treated with TY‐51469, 1 hour after ligation, show no variation in IVC and AA blood velocities when compared with the same animals before ligation. Control WT, unlike TY‐51469–treated DVT mice, also showed compensatory increases in AA blood flow, 6 and 24 hours after ligation. The images are representative of 6 to 8 independent experiments. Bar=1 mm. WT indicates wild type.

Table 1. Pulse‐Wave Doppler Velocity Analysis

Time after ligation, hAbdominal aorta blood velocity, mm/sInferior vena cava blood velocity, mm/s
WTNo.+10 mg/kg TY‐51469No.WTNo.+10 mg/kg TY‐51469No.

Each value represents the mean±SEM of 6 to 8 experiments. Statistical analysis: Mann‐Whitney test. No, indicates number of experiments; and WT, wild type.

*P<0.05, P<0.001 compared with the WT control group.

mMCP‐4 Knockout Mice Are Resistant to DVT Formation in the Stenosis Model

mMCP‐4 is the main chymase expressed by connective tissue–type mast cells and is the functional ortholog to human chymase.26 Next, we assessed whether genetic deficiency of chymase mMCP‐4, similar to pharmacological chymase inhibition, affected DVT formation. Indeed, mMCP‐4 knockout mice were found to be highly resistant to DVT formation after ligation in both male and female congeners (Figures 3A and S2A). Notably, only 5 of 31 tested male and female mMCP‐4 knockout mice subjected to DVT developed morphometrically measurable thrombi.

Figure 3. Chymase inhibition protects against and resolves deep vein thrombosis in male mice in 2 murine models.

Male wild‐type (WT) mice were treated with 10 mg/kg TY‐51469 (intraperitoneally) (48 hours before and 6 or 24 hours after ligation), and mouse mast cell protease‐4 (mMCP‐4) knockout (KO) mice were subjected to an inferior vena cava stenosis for 24 (A) or 48 (B) hours. Morphometric parameters of thrombi were measured as weight (mg; i) and length (mm; ii). C, Weight of thrombi obtained after the ferric chloride insult model, 30 minutes after damage, from male WT mice treated with 10 mg/kg TY‐51469 and mMCP‐4 KO mice. Each group represents the mean±SEM (n=6–12). ***P<0.001 compared with the WT control group.

Chymase Inhibition Resolves Venous Thrombi in the Mouse Stenosis Model

To extend these findings, the impact of chymase inhibition as a preventive therapy for DVT was assessed by using a repetitive dose of 10 mg/kg of TY‐51469 each day for 2 days, and on the same day as the ligation surgery (see experimental design, Figure S1). By using this approach, similar results as for mMCP‐4 knockout mice (either males or females) were obtained (Figures 3A and S2A). Furthermore, whether a single dose of 10 mg/kg of TY‐51469 reduces mature venous thrombi was subsequently assessed. Indeed, WT mice treated with TY‐51469, either 6 or 24 hours after ligation, displayed resolved thrombi, in both male and female mice (Figures 3A and 3B and S2A and S2B).

Similar to the stenosis model, it was observed in a FeCl3‐induced DVT model that a single dose of TY‐51469 given 1 hour before venous damage in male WT mice reduced thrombus formation by 70% (Figure 3C). In agreement with the latter result, a similar reduction of thrombus size was observed in mice lacking mMCP‐4 (Figure 3C).

Mast Cell–Derived mMCP‐4 Is Located Inside the Thrombus of WT Ligated IVC

To provide further insight into the role of chymase in regulation of thrombus formation during DVT, we next stained for mast cell (toluidine blue) and chymase immunoreactivity in IVCs of healthy and DVT mice. Toluidine blue–stained mast cells were identified in the adventitia of IVCs derived from both WT and mMCP‐4 knockout mice (Figure 4). At 24 hours post‐ligation, mMCP‐4–specific immunoreactivity was localized in the periphery as well as within the occlusive thrombus in the IVC of WT mice (Figure 4) but not in thrombus‐free veins isolated from IVC‐ligated mMCP‐4 knockout mice (Figure 4). Finally, toluidine blue and mMCP‐4–specific immunoreactivity scores were quantified (Table S1) from the images of the 6 mouse vessels shown in Figure 4.

Figure 4. Mouse mast cell protease‐4 (mMCP‐4) immunoreactivity and toluidine blue (TB) mast cell staining in the inferior vena cava (IVC) of deep vein thrombosis mice.

A, TB staining of a nonligated IVC from a wild‐type (WT) mouse. B, mMCP‐4 immunoreactivity of a nonligated IVC from a WT mouse. C, TB staining of a ligated vein from a WT mouse. D, mMCP‐4 immunoreactivity in a ligated IVC from a WT mouse (vessel lumen). E, TB staining of a ligated IVC from an mMCP‐4 knockout mouse. F, Absence of mMCP‐4 immunoreactivity in the same ligated IVC as in E. Dilution of the mMCP‐4 antibody: 1×400. Red arrows: mast cells; black arrows: mMCP‐4 immunoreactivity. The images are representative of 3 independent experiments. Bar=50 μm.

Mouse Chymase Inhibits Plasmin Activity in Vitro

Plasmin is the main fibrinolytic enzyme involved in the degradation of fibrin complexes in preformed thrombi.34 We next assessed whether chymase can directly affect plasmin activity. Indeed, recombinant mMCP‐4 induced a significant reduction of plasmin activity (measured as hydrolysis of a plasmin‐specific fluorogenic substrate), and this effect of the murine chymases was inhibited in a concentration‐dependent manner by TY‐51469 (Figure 5A).

Figure 5. Mouse mast cell protease‐4 (mMCP‐4) reduces plasmin activity within thrombi of deep vein thrombosis mice.

A, Plasmin‐like activity was measured in vitro using mouse purified plasmin incubated for 24 hours at 37°C with recombinant mMCP‐4 with or without increasing concentrations of the chymase inhibitor TY‐51469. B and C, Ex vivo enzymatic activity in thrombi harvested 24 or 48 hours after ligation from wild‐type (WT) mice. WT thrombi were treated with vehicle or 10 μM TY‐51469 before the reading. B, Plasmin activity was measured as 7‐amino‐4‐methylcoumarin (AMC)–specific cleavage (nM) of the fluorogenic substrate D‐Ala‐Leu‐Lys‐AMC. C, Chymase activity was measured as AMC‐specific cleavage of the fluorogenic substrate Suc‐Leu‐Leu‐Val‐Tyr‐AMC. Each bar represents the mean±SEM (n=6–8 independent experiments). **P<0.01, ***P<0.001 compared with the plasmin only group or WT nontreated group. Max. indicates maximum.

Intrathrombus Chymase Inactivates Plasmin Activity in a TY‐51469–Sensitive Manner, ex Vivo

Enzymatic activity analyses showed that chymase is highly active in 24‐ and 48‐hour‐old thrombi, albeit to a lesser extent in the latter (Figure 5B and 5C). TY‐51469 (10 μM) abolished the chymase‐like activity found in thrombi ex vivo (Figure 5C). In WT thrombi, an inverse correlation between chymase and plasmin enzymatic activities was observed at 24 and 48 hours post‐IVC ligation. Moreover, TY‐51469 (10 μM) abolished chymase activity but conversely caused a robust enhancement of plasmin activity (3.5‐fold) in thrombi of WT mice (Figure 5B and 5C).

In a last series of in vivo experiments, the plasmin/plasminogen inhibitor BAY 1214237 (1 mg/kg) administered intraperitoneally, 1 hour prior to FeCl3 application, fully reversed the inhibition of thrombus formation induced by TY‐51469 (10 mg/kg, administered also 1 hour before chemical insult) in the mouse IVC (Figure S4).

Chymase‐Positive Mast Cells and CMA‐1 Are Located Inside Thrombi of Patients With DVT

In 3 independent experiments, CMA‐1–positive mast cells were identified within thrombi of human patient left iliac, right femoral, and left groin varicose veins. Figure 6A depicts CMA‐1–positive mast cells in the left groin varicose vein.

Figure 6. Specific chymase immunostaining is found within human venous thrombi, and human recombinant chymase (rCMA‐1) cleaves/inactivates human plasmin.

A, Human deep vein thrombosis (DVT) samples show chymase immunoreactivity (brown) and chymase‐positive mast cells inside the thrombi. Dilution of the CMA‐1 antibody: 1×200; black arrows: CMA‐1 immunoreactivity. The image is representative of independent experiments performed with venous thrombi extracted from 3 patients. Bar=20 μm. B, Plasmin‐like activity was measured in vitro with purified human plasmin incubated for 24 hours at 37°C with rCMA‐1 with or without increasing concentrations of the chymase inhibitor TY‐51469. Each bar represents the mean±SEM (n=8–13). *P<0.05, ***P<0.001. C, Identification of rCMA‐1 cleavage sites in human plasmin using trypsin digestion and cleavage product identification by liquid chromatography–tandem mass spectrometry, represented as arbitrary intensity (4 samples tested). D, Schematic representation of human plasmin/plasminogen. rCMA‐1 cleavage sites are indicated by black arrows. Created with AMC indicates 7‐amino‐4‐methylcoumarin; AU, arbitrary unit; and Max., maximum.

Human Chymase Inhibits Plasmin Activity in Vitro

Similarly to mMCP‐4, human recombinant chymase, CMA‐1, induced a significant reduction of plasmin activity (measured as hydrolysis of a plasmin‐specific fluorogenic substrate), and this effect was inhibited in a concentration‐dependent manner by TY‐51469 (Figure 6B). In addition, 6 cleavage sites of purified human plasmin by recombinant human chymase CMA‐135 (yielding 5 distinct fragments; 4 of these within the catalytic pocket) were identified using trypsin digestion and cleavage product identification by liquid chromatography–tandem mass spectrometry (Table S2). This plasmin degradation was reduced in the presence of the chymase inhibitor, TY‐51469 (Figure 6C and 6D). In agreement with the latter result, recombinant mMCP‐4 also produced similar TY‐51469‐sensitive cleavage patterns for murine plasmin fragments (Figure S5A).


Our study reports, for the first time, that mast cell–derived chymase plays a pivotal role in the stabilization of venous thrombi in mouse models by limiting the fibrinolytic efficacy of plasmin within the clot. Plasmin activity was impaired by chymase but was preserved by the chymase inhibitor TY‐51469 in in vitro settings, both for recombinant human and mouse chymase as well as in mouse thrombi ex vivo. In agreement with this, chymase and chymase‐positive mast cells were found within thrombi recovered from the mouse experiments as well as from patients with DVT. Importantly, mMCP‐4 knockout mice or WT mice treated with TY‐51469 did not display anticoagulant responses in ex vivo clotting time assays, nor was a prolongation in bleeding times observed when compared with control animals.

In the present study, we also demonstrate that chymase inactivates plasmin in the thrombus. Our results show that unlike current therapy for DVT, such as heparin, warfarin, or direct oral anticoagulants,2, 7, 36 inhibition of chymase did not affect the bleeding time in the tail vein model in mice, suggesting that inhibition of the mast cell endogenous serine protease does not influence hemostasis. More important, Ponomaryov et al also demonstrated that mast cell depletion abolishes DVT, with no impact on hemostasis and tail bleeding time in the mouse DVT model.11

Chymase released from mast cells is known to be rapidly neutralized in the blood by plasma protease inhibitors, such as α‐2‐macroglobulin.37 In this study, the tail‐bleeding assay and thromboelastographic experiment demonstrate that circulating chymase at the bleeding site is not active, as previously reported for other large enzymes, such as plasmin.37 However, our experiments suggest that intrathrombus chymase may be protected from protease inhibitors and, thereby, is capable of reducing plasmin activity via proteolysis. Together, our study thus introduces a novel role of chymase in the regulation of plasmin activity, specifically within the thrombotic microenvironment.

Fully formed thrombi are generated 3 to 6 hours after ligation in the IVC stenosis mouse model.38, 39 This model causes no direct endothelial damage in ligated veins and allows thrombus formation with a residual blood flow, thus closely mimicking the pathology in the large majority of clinical cases.38, 40 It is striking that the lack of mMCP‐4 or administration of a chymase inhibitor 1, 6, or 24 hours after ligation is sufficient to prevent thrombus formation and, with regard to TY‐51469, even dissolve already existing thrombi in the mouse model. Although this is similar to lytic effects of recombinant tPA (alteplase), chymase inhibition reveals no impact on bleeding time.41, 42 Furthermore, a single dose of TY‐51469 in the stenosis mouse model preserves the IVC blood flow for 24 hours after ligation. In clinical settings, a prompt blood flow restoration is also an indication of a reduced thrombus burden, reduced fibrosis of the vessel, and better outcome in reducing DVT complications in patients.43

In the present study, we also used a mouse model of FeCl3‐induced DVT where the thrombus is provoked by an endothelial insult instead of blood stasis. As the Virchow triad describes, 3 factors are thought to contribute to thrombosis: hypercoagulability, blood stasis, and endothelial dysfunction. In conditions of vascular insult, endothelial dysfunction occurs along with platelet activation, adhesion, and aggregation, followed by fibrin formation.44 Interestingly, the present study shows that mMCP‐4 knockout or WT mice treated with TY‐51469 have reduced thrombus formation also in the FeCl3‐induced endothelial insult model. In addition, the loss of antithrombotic properties of TY‐51469 in IVC‐FeCl3 mice treated with the plasminogen/plasmin inhibitor BAY 1214237 further supports the profibrinolytic properties of chymase inhibitors in this mouse model of venous thrombosis. Thus, this study shows that the impact of the chymase inhibitor TY‐51469 is not limited to a single type of thrombosis pathogenesis, and chymase inhibition has therefore the potential to be effective in most clinical DVT scenarios.

Albeit the present study shows that a chymase inhibitor can resolve venous thrombi, a role for the latter serine protease in arterial thrombosis remains to be investigated. Ponomaryov et al did not report a significant role for mast cells in a mouse model of arterial thrombosis.11 However, the FeCl3‐injury experiments in mesenteric arteries performed by Ponomaryov and colleagues11 were conducted in animals with healthy vessels, whereas it is well established that cardiovascular events in humans occur after plaque rupture or erosion of atherosclerotic arterial wall, which contains substantial amounts of mast cells.45 Although showing large heterogeneity, arterial thrombi obtained from patients with stroke, myocardial infarction postmortem, or after thrombectomy are all known to be strongly dependent on fibrin formation. Therefore, a role for chymase in the pathologic features of arterial occlusion cannot be excluded, and chymase inhibition might have an impact as well. Finally, elevated D‐dimer levels have previously been reported in the IVC stenosis mouse model.46 Whether the antithrombotic effects afforded by mast cell depletion11 or chymase inhibition are associated with reduced D‐dimer levels in the mouse model requires further investigation.

Although we provide evidence that human chymase inactivates human plasmin within human venous thrombi, the strategy of using chymase inhibition in treatment of DVT in humans needs to be validated in clinical settings. Notably, the chymase inhibitor TY‐51469 is not orally bioavailable and thus requires systemic administration. Future studies may require the use of novel orally available chymase inhibitors, such as fulacimstat (BAY‐1142524).18, 47, 48, 49

Altogether, the present in vivo and in vitro studies highlight the potential usefulness of chymase inhibition as a novel strategy for venous thrombus resolution without increasing bleeding risk.

Sources of funding

This project was supported by a grant (MOP‐57883) from the Canadian Institutes for Health Research (Dr D'Orléans‐Juste), the John E. Edwards Cardiology Chair (Dr D'Orléans‐Juste), and the Réseau Québécois de Recherche sur les Médicaments (Drs D'Orléans‐Juste and Sirois). C. Lapointe is the recipient of a doctorate studentship (Gérard‐Eugène‐Plante) from the Université de Sherbrooke. Dr Giguère was the recipient of a Natural Sciences and Engineering Research Council of Canada (Alexander‐Graham‐Bell) graduate scholarship–doctoral award (CGSD–504926). Dr Auger‐Messier is supported by Fonds de Recherche du Québec‐Santé research scholarship–career award (junior: 2–284164, 2020–2022; senior: 313286, 2022–2026). Drs Tinel and Heitmeier are employees of Bayer AG.


C. Lapointe, L. Vincent, Dr Tinel, Dr Heitmeier, Dr Schwertani and Dr D'Orléans‐Juste are inventors on a provisional patent related to this study.


The authors thank Dr François‐Michel Boisvert and Dominique Lévesque (Université de Sherbrooke) for the mass spectrometry experiments; Dr Dany Salvail (IPS‐Thérapeutique, Sherbrooke, QC, Canada), Dr Sébastien Labbé (Immune Biosolutions Inc, Sherbrooke, QC, Canada), and Julie Brodeur (formerly from IPS‐Thérapeutique) for the thromboelastographic experiments; Dr Bin Yu (McGill University, Montréal, QC, Canada) for the immunohistology experiments; Antoine Désilets (Université de Sherbrooke) for the spectrofluorometric experiments; and Dr Robert Day and Roxane Desjardins (Université de Sherbrooke) for the recombinant enzymes, human chymase (CMA‐1) and recombinant mouse mast cell protease‐4.


*Correspondence to: Pedro D'Orléans‐Juste, PhD, Department of Pharmacology and Physiology, Université de Sherbrooke, 3001, 12 e Avenue Nord, Sherbrooke, QC J1H 5N4, Canada. Email:

Supplemental Material is available at

For Sources of Funding and Disclosures, see page 12.



eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.