Role of platelet count in a murine stasis model of deep vein thrombosis

Abstract

Platelets are core components of thrombi but their effect on thrombus burden during deep vein thrombosis (DVT) has not been fully characterized. We examined the role of thrombopoietin-altered platelet count on thrombus burden in a murine stasis model of DVT. To modulate platelet count compared to baseline, CD1 mice were pretreated with thrombopoietin antisense oligonucleotide (THPO-ASO, 56% decrease), thrombopoietin mimetic (TPO-mimetic, 36% increase), or saline (within 1%). Thrombi and vein walls were examined on postoperative days (POD) 3 and 7. Thrombus weights on POD 3 were not different between treatment groups (p = .84). The mean thrombus weights on POD 7 were significantly increased in the TPO-mimetic cohort compared to the THPO-ASO (p = .005) and the saline (p = .012) cohorts. Histological grading at POD 3 revealed a significantly increased smooth muscle cell presence in the thrombi and CD31 positive channeling in the vein wall of the TPO-mimetic cohort compared to the saline and THPO-ASO cohorts (p < .05). No differences were observed in histology on POD 7. Thrombopoietin-induced increased platelet count increased thrombus weight on POD 7 indicating platelet count may regulate thrombus burden during early resolution of venous thrombi in this murine stasis model of DVT.

Plain Language Summary

Deep vein thrombosis (DVT) is a pathology in which blood clots form in the deep veins of our body. Usually occurring in the legs, these clots can be dangerous if they dislodge and travel to the heart and are pumped into the lungs. Often these clots do not travel and heal where they formed. However, as the body heals the clot it may also cause damage to the vein wall and predispose the patient to future clots, i.e., the biggest risk factor for a second clot is the first clot. DVT can also cause symptoms of pain, swelling, and redness in the long-term, leading to post-thrombotic syndrome where the initial symptoms of the clot persist for a long time. All blood clots have common components of red blood cells, white blood cells, platelets, and fibrin in varying concentrations. Humans maintain a platelet count between 150 and 400 thousand platelets per microliter of our blood. However, diseases like cancer or medications like chemotherapy can cause a change in our body’s platelet count. The effect of a changing platelet count on the size (clot burden) of DVT clot and how platelet count could affect DVT as the clot heals is not fully understood. Studying this might help us develop better targets and treat patients with a wide range of platelet counts who experience DVT. In this study, we intentionally decreased, left unchanged, and increased platelet counts in mice and then created a DVT to study what the effect of low, normal, and high platelet counts, respectively, would be on the clot burden. We observed that mice with higher platelet counts had a higher clot burden during the early part of the healing process of the clot. Within this study, we can conclude that higher platelet counts may lead to higher clot burden in DVT which furthers our understanding of how platelet count affects clot burden during DVT.

Keywords:

• Deep vein thrombosis
• platelets
• postthrombotic syndrome
• thrombus
• veins

Introduction

Venous thromboembolism (VTE) disorders rank as the third most common cause of death from adverse cardiovascular incidents following myocardial infarction and stroke.¹ Deep vein thrombosis (DVT) incidence increases with age, and mortality arises from consequent pulmonary embolism (PE) with up to 100,000 deaths in the US annually.² Morbidity from DVT occurs from postthrombotic syndrome (PTS), where chronic inflammation, pain, and fibrosis lead to a lowered quality of life for patients.^3–7 It is estimated that over 200,000 people in the US will develop DVT this year and over 43% of them will develop PTS. Current treatment options, including catheter-directed thrombolysis in addition to anticoagulation, have failed to prevent PTS.⁸ Therefore, a deeper investigation is warranted into DVT-PTS progression to develop targeted and effective therapies.

Morbidity from DVT differs with degree of occlusion, yet thrombus burden in clinical DVT is not well quantified in health records.^1,9 While the extent of thrombotic occlusion is more consistently noted, exact measurements of the physical parameters of thrombi at diagnosis and follow-up are inconsistent. This obscures the evaluation of clinical thrombus burden in DVT. While increased platelet count has been associated with increased risk of VTE incidence,¹⁰ the role of platelet count on thrombus burden in DVT remains unclear due to the lack of thrombus size measurements.

Murine models of DVT allow for measurement of thrombus size and burden but vary widely in thrombus incidence rates and their ability to model the anatomy and degree of occlusion present in clinical DVT.¹¹ The murine stasis model of DVT, also known as the infrarenal inferior vena cava (IVC) ligation model, completely obstructs blood flow in the IVC and produces a relatively higher and consistent incidence of thrombus compared to other models. The constriction model of DVT, also known as the IVC stenosis model, constricts the infrarenal IVC and does not completely obstruct blood flow which results in inconsistent thrombus formation. Previous murine DVT studies have indicated a protective effect of severe thrombocytopenia (99% reduction of platelet count) on thrombus burden at 48 h or post-operative day (POD) 2 in IVC stenosis models; however, further investigation is needed to deepen our understanding of the role of platelet count on thrombus burden during DVT progression in additional models, such as in murine stasis models recapitulating acute and fully occlusive DVT.^12–14

In this study, we investigated the role of thrombopoietin-altered platelet count on thrombus burden during DVT. Using the murine stasis model of DVT, we demonstrate that increased platelet count significantly increases thrombus burden during early DVT resolution.

Methods

Animals

CD1 mice were purchased from Charles River Laboratories (Wilmington, MA) and housed in a 12-h light/dark cycle with adlib access to food, water, and enrichment. A total of 36, 8–12-week-old CD1 mice (mean weight ± standard deviation: 27 ± 3 g) were used. All experiments were done in accordance with the national standards outlined by the National Institutes of Health (NIH) and Institutional Animal Care and Use Committee (IACUC) policies at the Oregon Health and Science University.

Treatment groups

Mice were separated into three treatment cohorts (Figure 1). To reduce platelet count, mice were subcutaneously injected both once a week for 2 weeks prior to IVC ligation and on the day of surgery with a second-generation triantennary N-acetyl galactosamine (GalNAc₃) conjugated anti-sense oligonucleotide (ASO) to hepatic thrombopoietin (THPO-ASO) at 10 mg/kg. GalNAc₃ conjugation facilitates specific delivery and uptake of ASOs via high-affinity binding to hepatocyte-specific asialoglycoprotein receptor, causing up to a 30-fold increase in potency over first-generation unconjugated ASOs.^15–17

Figure 1. Timeline for treatments, blood draws, surgery, and euthanasia. Antisense oligonucleotide treatment against thrombopoietin (THPO-ASO, blue triangle) and saline (red circle) were administered once a week to decrease or control for platelet count, respectively. A mimetic to thrombopoietin (TPO-mimetic, green triangle) was administered using oral gavage to mice 5 days a week and continued from surgery to euthanasia to maintain an elevated platelet count. Blood draws were analyzed weekly through complete blood count with differential. Separate cohorts of mice underwent euthanasia at post-operative day (POD) 3 and 7 to evaluate the effects of treatments on deep vein thrombus burden. Created with Biorender.com.

In a parallel regimen, the control cohort received subcutaneous injections of 0.9% normal saline once a week. To increase platelet count, mice underwent oral gavage treatment with a thrombopoietin mimetic (TPO-mimetic, eltrombopag) for 5 days a week beginning 2 weeks prior to IVC ligation and throughout the study period at 75 mg/kg (Figure 1).

Hematological analysis

Animals were anesthetized with 5% isoflurane vapor, and whole blood was drawn from the retro-orbital sinus once weekly. Blood was diluted with an equal volume of 10 mM EDTA, and raw blood cell counts and volumes were analyzed using a Hemavet 950FS instrument (Drew Scientific, Miami Lakes, FL). A complete blood count (CBC) with differential was performed on all whole blood samples to measure platelet count (Figure S1). The remaining plasma was frozen at −80°C for future analyses. Circulating fibrin degradation products in the plasma were quantified at the terminal timepoints using a colorimetric D-dimer assay (Novus Biologicals, Product NBP3-08100).

Surgical model of DVT

The murine stasis model of DVT was used to produce consistent venous thrombi in mice, as described previously.^11,18 Briefly, the animals were prepped in sterile fashion. Five percent isoflurane vapor was used for induction and reduced to 2.5% for maintenance of anesthesia during the procedure. A laparotomy was performed to expose the abdomen. Displacement of the intestines provided exposure of the retroperitoneal space. The fascia between the infrarenal aorta and the IVC was blunt dissected to allow a 7-0 prolene suture to pass between the two vessels. The infra-renal IVC was ligated with the suture, and the endothelium was injured by clamping the IVC wall, caudal to the ligation, twice for 15 s. Side branches were cauterized to prevent collateral flow and increase the consistency of thrombus formation. The laparotomy was then closed in two layers with 4–0 vicryl suture. Mice were euthanized at either POD 3 or 7 (n = 6/treatment group/day). At the time of euthanasia, blood was collected via cardiac puncture and the infrarenal IVC was excised from the point of ligation to the iliac bifurcation for further analysis of the thrombus and the vein wall.

Thrombus analyses

Thrombus weight

Specimen consisting of both thrombus and vein wall was excised from the point of ligation to the iliac bifurcation and weighed. The thrombus and vein wall were weighed together in a consistent and reproducible manner to allow for direct comparison among specimens of different groups as shown in previous studies.^13,18–22 Excess fluid was removed by blotting on delicate task wipes, and the IVC and thrombus were weighed on a digital analytical balance.

Thrombus and vein wall composition

The specimen consisting of both thrombus and vein wall was formalin fixed, cut along the transverse midplane of the thrombus, processed and paraffin embedded in cassettes, and cut in 5 μm sections from both open cross sections for histological analyses. A separate slide section – that included the two open cross sections – was used for each subject and for each stain. Histological analyses were performed using standard techniques as previously described.²³ To stain for collagen (blue) and muscle/keratin (red), a Masson’s Trichrome stain (Aniline Blue Kit, Newcomer Supply, Middleton, WI) was used. Hematoxylin and eosin (H&E) stain was used to grade erythrocyte, platelet, and acellular plasma protein and fibrin-rich thrombus regions as described previously.^24,25 Collagen content and nucleated cell presence were used to determine maturity and organization of the thrombus. For immunohistochemistry analyses, sections were blocked with 3% milk and incubated with primary antibodies to Ly6G to indicate neutrophil presence (rat anti-mouse, 1:200, BD Biosciences, NJ, USA), Iba1 to indicate macrophage presence (rabbit polyclonal, 1:5000, Fujifilm Wako Chemicals, Richmond, VA), endothelial cell CD31 to indicate endothelial cells and neovascularization if distinct channel structures were observed (1:250, Abcam, Cambridge, MA), and smooth muscle actin to indicate presence of smooth muscle cells (SMA; anti-mouse, 1:1000, DAKO, Agilent Technologies, Santa Clara, CA). Immunostained slides were imaged and scored in a blinded fashion, i.e., the scorer was unaware to which treatment group the slides being graded belonged. The positively stained percent area was estimated and ranked as 1 = no staining, 2 = <25%, 3 = 25–50%, 4 = 50–75%, 5 = 75–100%.

Data processing and statistical analysis

The platelet counts, thrombus measurements, and histology grading were collected cumulatively. Data were analyzed using the R programming software version 4.2.1. An ANOVA with Tukey’s post hoc testing was used to compare continuous variables. A Kruskal–Wallis test with a Dunn’s post-hoc test with Benjamini–Hochberg correction was used to compare ordinal variables. A Fisher’s exact test was used to compare proportions of categorical variables. Significance was set at the 0.05 level.

Results

Platelet count modulation: hematological analysis

At the time of IVC ligation, the mean decrease in platelet count was 56.49 ± 14.88% in THPO-ASO treated mice, the mean increase in platelet count was 36.35 ± 24.36% in the TPO-mimetic treated mice, and the mean changed was 0.71 ± 18.02% in the saline treated mice compared to their baseline values (Figure 2 and Figure S1). The mean decrease in platelet count in the THPO-ASO treated mice was significantly different than the saline (Figure 2, p = .005, Tukey’s test) and the TPO-mimetic (Figure 2, p < .001, Tukey’s test) treated mice. The mean increase in the TPO-mimetic treated mice was significantly higher than mice treated with saline (Figure 2, p = .02, Tukey’s test).

Figure 2. Platelet count. Percent change in platelet count from baseline to surgery as a result of THPO-ASO, saline, and TPO-mimetic treatment. (N = 12 per treatment group, *** indicates p < .001, and **** indicates p < .0001, ANOVA with Tukey post-hoc).

Effect of platelet count modulation on thrombus burden

There were no significant differences in thrombus incidence between treatment groups on POD 3 (p = .294, Fisher’s exact test, Figure S2) or POD 7 (p = 1, Fisher’s exact test, Figure S2). In the murine stasis model of DVT, peak thrombus size is reported between days 2–4^11,26; therefore, to evaluate the effect of platelet count on thrombus burden during DVT formation, thrombus weights were analyzed at POD 3. THPO-ASO treated mice had a mean thrombus weight of 30.5 ± 7.1 mg compared to mice treated with saline with a mean thrombus weight of 32.8 ± 6.5 mg and mice treated with TPO-mimetic with a mean thrombus weight of 29.3 ± 11.2 mg (Figure 3). Thrombus weights (p = .84, ANOVA) were not significantly different between these treatment groups at POD 3. Furthermore, no significant differences were noted in plasma d-dimer concentrations between groups at POD 3 (Figure S3).

Figure 3. DVT outcomes. Thrombus weights were measured to determine thrombus burden at post-operative day (POD) 3 and 7 (*indicated p < .05, ** p < .01, ANOVA with Tukey post-hoc).

In this murine stasis model of DVT, thrombus resolution starts after POD 4 measured by decreasing thrombus weight^11,26; therefore, to evaluate the effect of platelet count on thrombus burden during resolution, thrombus weights were analyzed at POD 7 in a separate cohort. Thrombus weights were significantly different (p = .004, ANOVA) between treatment groups (Figure 3); specifically, the TPO-mimetic treated group (41.4 ± 5.7 mg) had significantly increased mean thrombus weights compared to the mice treated with THPO-ASO (28.3 ± 1.7 mg, p = .005, Tukey’s test) and saline (30.4 ± 5.6 mg, p = .012, Tukey’s test). However, no significant differences were noted in the plasma d-dimer concentration between treatment groups at POD 7 (Figure S3).

Effect of platelet count on thrombus and vein wall composition

The compositions of the thrombi and vein wall were analyzed at POD 3 and POD 7. Significantly higher levels of smooth muscle cell actin staining (Figure 4G-I) were observed at POD 3 in the thrombi of mice treated with TPO-mimetic (1–25%) compared to THPO-ASO (no thrombus SMA observed, Dunn’s test with Benjamini Hochberg correction, p = .002) and saline (no thrombus SMA observed, Dunn’s test with Benjamini Hochberg correction, p = .004). Significantly higher levels of CD31 positive channel staining (Figure S4), indicating neovascularization as part of fibrotic remodeling, were observed at POD 3 in the vein wall of mice treated with TPO-mimetic compared to THPO-ASO (Dunn’s test with Benjamini Hochberg correction, p = .01) and saline (Dunn’s test with Benjamini Hochberg correction, p = .01). No significant differences in collagen content or macrophages were observed in thrombus and vein wall composition at POD 3 or POD 7 among treatment groups (Figures 4 and 5). There were large changes in thrombus organization at POD 7 compared to POD 3 as characterized by increases in collagen content, neovascularization, and cellular infiltration including smooth muscle cells and macrophages.

Figure 4. Immunohistochemistry of thrombi. A–C. Trichrome staining for collagen and thrombus organization. D–F. CD31 staining for endothelial cells and channel formation indicative of neovascularization. G–I. Smooth muscle actin (SMA) staining for smooth muscle cells (** p < .01, Kruskal–Wallis with Dunn’s post-hoc). A star indicates the location of the thrombus, a diamond indicates the location of the vein wall, and an arrow identifies areas of positive staining.

Figure 5. Immunohistochemistry of thrombi. A–C. IBA1 staining for macrophages D–F. LY6G staining for neutrophils. G–I. H&E staining for erythrocyte, platelet, and plasma protein/fibrin content. A star indicates the location of the thrombus, a diamond indicates the location of the vein wall, and an arrow identifies areas of positive staining.

Discussion

Thrombopoietin is produced primarily in the liver but also in the kidney, spleen, and bone marrow.²⁷ Thrombopoietin is the ligand for the myeloproliferative leukemia virus oncogene (MPL) receptor which is essential for megakaryocyte maturation and platelet count regulation.²⁸ Antibody and ASO targeting of thrombopoietin has been demonstrated to effectively lower platelet count in non-human primates and mice.^15,29,30 Using a hepatocyte-specific second-generation THPO-ASO, saline, and a TPO-mimetic, this study investigated the role of thrombopoietin-altered platelet count on thrombus burden in a stasis model of murine DVT.^11,18 Prior to DVT induction platelet counts were successfully modulated from their baseline to be both significantly decreased in the THPO-ASO treated mice and significantly increased in the TPO-mimetic treated mice. The platelet counts remained steady in saline treated mice (Figure 2 and Figure S1). However, at the terminal timepoints (POD 3 and POD 7), the difference in platelet counts between the TPO-mimetic and saline-treated groups was no longer significant. This response may have been a reactive thrombocytosis to the surgical intervention. This transient increase in platelet count occurred only in the saline-treated group, although it is unclear why it was not equally matched in the THPO-ASO and TPO-mimetic treated mice. This phenomenon underscores the complex interplay of platelet dynamics in the context of DVT and highlights the need for further investigation into the nuanced aspects of the role of platelets during thrombus formation and resolution in this murine model.

In this murine stasis model of DVT, the thrombus weight has been reported to increase from POD 1 to POD 4 and the thrombus resolution to occur after POD 4 as measured through a decreasing thrombus weight up to POD 28.^11,26 We chose POD 3 and 7 as the time points to measure the effect of modulating platelet count on thrombus burden during DVT formation and early thrombus resolution, respectively.²⁶

Platelets are the core components of thrombi. During both arterial and venous thrombus initiations, platelets are recruited to the site of injury.³¹ In venous thrombi, they are recruited following endothelial activation and increased selectin expression on an inflamed endothelial monolayer.^32,33 Platelet release of high mobility group box 1 (HMGB1), for which they are the major source, has been shown to be a danger-associated molecular pattern (DAMP) capable of recruiting monocytes and neutrophils and causing feedback activation of platelets. Platelet-derived HMGB1 has been demonstrated to propagate DVT formation by increasing tissue factor expression in monocytes and triggering both neutrophil extracellular trap formation and platelet accumulation through autocrine signaling.³⁴ Therefore, platelet releasate role in venous thrombosis has been demonstrated; however, the role of platelet count on thrombus burden during DVT is unclear.

In murine IVC stenosis models of DVT, targeting the platelet glycoprotein Ibα (GPIbα) to induce severe thrombocytopenia has shown a protective effect on experimental murine thrombogenesis. Achieving a 99% decrease in circulating platelet count in wildtype C57BL/6 mice using an antibody against the GPIbα receptor, Mwiza et al. recorded significantly lower thrombus weights that were a mean of 89% lower at 48 h/POD 2 compared to saline controls post IVC stenosis.^12,13 In this study, we did not observe a significant effect of platelet count on thrombus incidence or burden at POD 3 (Figure 3, Figures S2 and S3). However, this study did not induce severe thrombocytopenia which may have accounted for the lack of decrease in thrombus weight and thrombus burden during DVT formation. Furthermore, we evaluated thrombus burden in a different strain of mice at POD 3 in a murine stasis model of DVT compared to POD 2 in a murine stenosis model of DVT, and we targeted the MPL thrombopoietin receptor ligand versus the GPIbα receptor, which may further account for differences in these findings.

To further evaluate the role of platelet count on thrombus burden during DVT, we measured thrombus weight at POD 7, representing an early resolution timepoint.²⁶ We observed significantly higher thrombus weights in mice with increased platelet count (TPO-mimetic) at POD 7 compared to mice with a decrease (THPO-ASO) or no change (saline) in their platelet count (Figure 3). Thus, increased-platelet-count was associated with a higher thrombus burden during early resolution, which may have resulted from a complex interplay between the thrombotic, inflammatory, and coagulation components. Platelets have been shown to have an intermediary role between these systems. Von Brühl et al. demonstrated that mice lacking the external domain of platelet adhesion receptor GPIbα decreased thrombus burden due to low levels of leukocyte recruitment and neutrophil extracellular trap formation in an IVC stenosis model of DVT.¹³ Mwiza et al. used dual antiplatelet therapy targeting P2Y₁₂ and thromboxane A₂ receptors as well as ibrutinib to target the Bruton’s tyrosine kinase signaling pathway, demonstrating significantly lower thrombus weights at POD 2 in an IVC stenosis model.¹² Additionally, platelet regulation of plasmin activity affects DVT progression as evidenced by Diaz et al., demonstrating lower thrombus weights at POD 2 in plasminogen activator inhibitor-1 knockout mice.³⁵ Furthermore, Baldwin et al. used the stasis IVC ligation model in mice to demonstrate lower thrombus weights with higher plasmin activity, while Dewyer et al. demonstrate higher thrombus weights with inhibition of plasminogen activator inhibitor-1 in a rat IVC ligation model.^36,37 While the thrombus sizes were significantly increased in the TPO-mimetic group at POD 7, the d-dimer concentration was unchanged (Figure S3). However, d-dimer tests are highly variable and may remain elevated during clot formation and resolution, thereby likely unable to measure changes in thrombus burden during early thrombus resolution.³⁸ Therefore, to measure the kinetics of thrombus resolution with increased granularity, ultrasound measurements to quantify thrombus size during formation and resolution would be necessary. To determine the effect of thrombopoietin altered platelet count on the later stages of DVT progression, additional time points (POD 10, 14, etc.) are warranted in future studies. In this model, additional timepoints would elucidate whether TPO-mimetic treated mice have an extended period of thrombus formation or a delay in their resolution process.^11,26 Investigation of platelet functionality in the context of varying platelet count and their role as an intermediary between thrombosis, inflammation, and coagulation is warranted to further investigate their effect on thrombus burden during DVT resolution in this model. Introducing an MPL-receptor agonist (TPO-mimetic) demonstrated higher thrombus burden in this stasis model at POD 7 while treating with THPO-ASO did not increase thrombus burden at either POD 3 or POD 7, compelling further investigation of the MPL pathway in regulating thrombus burden during DVT progression.

To further investigate the constituent changes underlying thrombus burden in this model, we histologically analyzed the thrombus sections (Figures 4 and 5). At POD 3, the TPO-mimetic treated mice with increased platelet count had a significantly higher grade of smooth muscle cells present in the thrombus (Figure 4G–I) and CD31 positive channeling in the vein wall compared to both the THPO-ASO and saline treated mice with lower platelet count (Figure S4). H&E staining did not reveal significant differences in erythrocyte, platelet, or acellular (plasma proteins, fibrin) content of thrombi at POD 3 or POD 7 (Figure 5G–I). Trichrome staining did not reveal changes in collagen content as an indicator of fibrotic remodeling to the thrombus or the vein wall at either time point (Figure 4A–C). No differences between the groups were noted in LY6G or IBA1 staining, indicative of neutrophil and macrophage presence in the thrombi or vein walls (Figure 5A–F). Thus, differences in inflammatory cell presence during DVT development and early resolution were not evident between treatment groups in this murine model as a result of thrombopoietin mediated platelet count modulation. Platelet derived growth factor has been implicated in inducing a proliferative phenotype of venous smooth muscle cells³⁹; therefore, increased platelet count in TPO-mimetic treated mice may have caused a proliferative phenotype for smooth muscle cells relative to the THPO-ASO and saline treated groups. While increased SMA and CD31-positive-channel staining were only observed in the POD 3 group, a proliferative smooth muscle cell phenotype and increased CD31 positive channeling in the vein wall suggest a higher fibrotic response and may account for the increased thrombus burden at POD 7. Further studies are warranted investigating the role of thrombopoietin altered platelet count on thrombus and vein wall fibrosis and its contribution to thrombus burden in this stasis model of DVT.

Investigating the role of thrombopoietin-altered platelet count in experimental DVT furthers understanding of platelet role in venous thrombosis. Translationally, these data suggest a role of the MPL pathway mediated platelet count in DVT. Patients may present with varying clinical contexts affecting their platelet count from hematologic malignancies and chemotherapy to bone marrow transplant recipients. This study associates thrombopoietin mediated platelet count with DVT burden. Mechanistic studies of DVT in MPL knockout mice are needed to further elucidate the role of the thrombopoietin mediated platelet production on thrombus burden in DVT and its role in PTS development. Over 200,000 people are expected to develop DVT each year in the US and current clinical therapy fails to prevent progression to PTS within 2 years in 40% of these cases,^1–5,8 warranting further investigation into the mechanisms of formation and resolution of DVT. Treatment with anticoagulation, commonly direct oral anticoagulants, is indicated or IVC filters when anticoagulation is contraindicated. In select cases with severe or progressive symptoms, venogram and direct thrombolysis may be indicated in patients with reasonable bleeding risks.^40,41 Whether there is a role for targeting the MPL pathway or platelet count targeted therapy in venous thromboembolism and its sequelae warrants further investigation.^42,43

A limitation of using the murine stasis model of DVT is the complete obstruction of flow, while in human DVT, there are variations in the extent and rate of obstruction of flow which may affect outcomes.⁹ In patients, acute DVT typically presents with complete occlusion of the affected vessel which is demonstrated in this murine stasis model of DVT. Variations in the degree of luminal occlusion are recapitulated in different murine DVT models.¹¹ These murine models vary in the context of thrombus formation and degree of occlusion through IVC constriction instead of ligation, optional IVC clamping (no endothelial injury), and optional side branch cauterization which retains collateral flow. While these alternatives model non-occlusive DVT, they do not demonstrate the high incidence of thrombosis provided by the stasis model which benefits from complete occlusion of flow and endothelial injury. Future studies into the role of platelet count on DVT incidence and resolution in these additional models will deepen and broaden our understanding of platelet role in DVT formation. Further limitations include the lack of an untargeted control ASO; however, our saline control demonstrated a steady platelet count profile compared to the THPO-ASO and TPO-mimetic groups.²⁹ Future studies may benefit in statistical power with higher sample sizes per treatment group. Our study minimized sample size to ethically conserve the number of animals needed in this study.

Conclusion

TPO-mimetic treatment increased thrombus burden during early resolution of DVT in this murine stasis model of DVT. These data warrant further investigation of platelet count as a factor regulating thrombus burden during DVT formation and progression.

Author contributions

RM performed the experiments, aided in study design, analyzed, and interpreted data, wrote, and revised the manuscript, and gave final approval of the manuscript. NS performed the experiments and aided in study design. ALC performed the experiments and aided in study design. AK analyzed and interpreted data, provided critical review, and final approval of the manuscript. JML provided critical review and final approval of the manuscript. RLW aided in study design, provided critical review, and final approval of the manuscript. CUL provided critical reagents, aided in study design, provided critical review, and final approval of the manuscript. AR provided critical reagents, critical review, and final approval of the manuscript. MH conceived and designed the study, analyzed, and interpreted the data, provided critical review, and final approval of the manuscript. KN conceived and designed the study, performed the experiments, analyzed, and interpreted data, provided critical review, and gave final approval of the manuscript.

Acknowledgments

The authors would like to acknowledge the technical assistance of Hillary Le, Rebecca J. Ditmore, and Jared Stoller. Several authors are or were Veterans Affairs employees [ALC, Research Assistant, VA Portland Health Care System (VAPORHCS), Portland, OR; JML, Research Scientist, VAPORHCS, Portland, OR; KPN, Staff Physician, VAPORHCS, Portland, OR]. The contents do not represent the views of the United States Department of Veterans Affairs or the United States Government.

Disclosure statement

Dr Lorentz and the Oregon Health & Science University hold financial interest in Aronora, Inc., a company that holds commercial interest in the results of this research. This potential conflict of interest has been reviewed and managed by the Oregon Health & Science University Conflict of Interest in Research Committee. Dr Revenko is an employee and stakeholder of Ionis Pharmaceuticals. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Data availability statement

Data generated for this study are available through request to the corresponding authors.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/09537104.2023.2290916.

Additional information

Funding

This work was supported by a VA CSR&D Career Development Award [Grant Number: IK2CX001720] and National Institutes of Health funding [5T32GM109835-04, 2T32HL083808, R44HL162386, R01HL144113, R01HL168696, and F30HL163918]. This material is the result of work supported, in part, by the United States Department of Veterans Affairs Biomedical Laboratory Research and Development Merit Review Program [JML under grant #I01BX002061].