To Clot or Not to Clot: Deepening Our Understanding of Alterations in the Hemostatic System

Abstract

The session on the hemostatic system focused on new developments in coagulation and platelet biology as well as how therapeutic agents may affect hemostasis. The classic cascade model of coagulation was compared with the more recent models of cell-based and vascular-based coagulation, which may provide better insight on how the coagulation cascade works in vivo. A review of platelet biology highlighted that, as platelets age, desialylated platelets form and are recognized by Ashwell-Morell receptor (AMR), leading to hepatic uptake and subsequent increase in thrombopoietin (TPO) production. Administration of therapeutics that induce thrombocytopenia was also discussed, including Mylotarg, which is an antibody-drug conjugate that was shown to decrease human megakaryocyte development but had no effect on platelet aggregation. An acetyl co-A carboxylase inhibitor was shown to cause thrombocytopenia by inhibiting de novo lipogenesis, which is critical for the formation of the megakaryocyte demarcation membrane system responsible for platelet production. It was also illustrated how preclinical translation models have been very helpful in the development of adeno-associated virus (AAV) hemophilia B gene therapy and what old and new preclinical tools we have that can predict the risk of a prothrombotic state in people.

Introduction

This article is a synopsis of the presentations given at Session 4 titled “To Clot or Not to Clot—Deepening Our Understanding of Alterations in the Hemostatic System” at the 41st Annual Symposium of the Society of Toxicologic Pathology in Austin, Texas. Drs Armando Irizarry from Eli Lilly and Gregory Krane from Moderna co-chaired the session as substitutes for William Reagan and Karrie Brenneman. Dr Irizarry started off the session by giving the introductory remarks to set the stage for the presentations on toxicologic pathology related to the hemostasis system. In response to vascular injury, coagulation resulting in clot formation keeps hemorrhage in check and operates in a finely tuned balance with the fibrinolytic system to prevent deleterious formation of thrombi/emboli. Administration of proposed therapeutic or toxic substances can disrupt this homeostatic equilibrium directly or indirectly, resulting in hemorrhage, thrombosis, or both. Evaluation of these perturbations and understanding their mechanism of toxicity can be challenging for the clinical and anatomic pathologist alike. This session discussed these challenges in the context of development of lipogenesis inhibitors, gene therapies for hemophilia B, and antibody-drug conjugates as well as recent advances in platelet biology and preclinical biomarkers of hypercoagulable states.

Coagulation Pathways: Fluid Phase, Cell-Based, and Vascular Bed Hemostasis

Dr Marjory Brooks, the Director of the Comparative Coagulation Laboratory at Cornell University, led off the session with the first presentation where she gave an overview of the hemostatic system titled, “Coagulation Pathways: Fluid Phase, Cell-Based, and Vascular Bed Hemostasis”

Hemostasis is the body’s defense mechanism to limit blood loss from vascular injury. Effective hemostasis requires rapid deployment, but must be tightly controlled to prevent pathologic thrombosis. Our understanding of the complex procoagulant and anticoagulant forces that participate in hemostasis has developed over time. Early discoveries of plasma coagulation factors were followed by recognition of the cell membrane localization of thrombin generation and the role of endothelial cells in regulating clot formation.

Coagulation refers to the transformation of fibrinogen, a soluble plasma protein, to the insoluble polymer, fibrin. The procoagulant factors and cofactors required for fibrin formation and the concept of an enzymatic “cascade” were defined by the mid-20th century. ¹ The cascade model describes two independent pathways that converge to generate the terminal procoagulant protease, thrombin. The intrinsic pathway is initiated through autoactivation of Factor XII upon contact with negatively charged surfaces, whereas exposure of tissue factor, an integral membrane protein, triggers the extrinsic pathway. These two pathways are the basis for the traditional coagulation screening tests: activated partial thromboplastin time (APTT) and prothrombin time (PT). The APTT and PT are used in preclinical studies and clinical practice to detect severe factor deficiencies or inhibitors; however, the intrinsic and extrinsic pathways are no longer believed to represent in vivo thrombin generation. ²

The cell-based model of coagulation describes three overlapping phases of factor activation that more accurately represent the central role of thrombin. ³ In the initiation phase, Factor VIIa interacts with its targets, Factor IX (FIX) and Factor X, on tissue factor-bearing cells to generate trace amounts of thrombin. In the subsequent amplification and propagation phases, activated platelets provide a membrane surface for assembly of highly efficient enzyme complexes. The tenase complex consists of FIXa and its cofactor, Factor VIII, whereas the prothrombinase complex consists of Factor Xa and its cofactor, Factor V. The propagation phase culminates in a large burst of thrombin, sufficient to cleave fibrinogen and form an insoluble fibrin clot localized to the activated platelet matrix.

Vascular bed hemostasis refers to the differential regulation of coagulation throughout the vasculature. ⁴ The local milieu of each vascular bed influences endothelial cell expression profiles through oxygen tension, nutrient composition, fluid shear stress, turbulence, and pressure. The potent anticoagulants, tissue factor pathway inhibitor and activated Protein C, are synthesized by or activated on the endothelium. Endothelial cells also secrete antiplatelet agents, including nitric oxide, prostacyclin, and NTPdase (apyrase), and plasminogen activators that initiate fibrinolysis and restore vascular patency. Pathologic thrombus formation results from perturbations of these regulatory pathways induced by systemic disease or regional endothelial cell dysfunction. ⁵ These insights into the biologic basis of physiologic and pathologic clot formation have been instrumental in developing new diagnostic tests and targeted treatments to identify and manage hemostatic imbalance.

The Sweet Side of Platelet Clearance: Novel Mechanisms of Platelet Clearance and TPO Regulation

Dr Renata Grozovsky is a Research Scientist at the University of Miami. She studied the regulatory mechanisms of platelet clearance at Brigham and Women’s Hospital and presented a summary of this work in the second presentation of the session titled “The Sweet Side of Platelet Clearance: Novel Mechanisms of Platelet Clearance and Thrombopoietin Regulation.”

Glycosylations are posttranslational modifications with many protective, stabilizing, organizational, and barrier functions. In addition, changes in glycan structures can modulate interactions of proteins with one another, limiting or enhancing interactions. ⁶ The platelet membrane is rich in glycoproteins organized as receptors of very different types. Therefore, the notion that glycosylation may regulate platelet function and survival highlights the importance of a molecular understanding of the roles of platelet glycosylation in physiological processes.

Platelets are small, anucleated cells, with a short life span of 4 to 5 days in mice and 8 to 10 days in humans. It is estimated that the human body produces and removes ~1 billion platelets daily. Maintenance of the platelet population number and function is central to regulate tissue homeostasis as dysregulated platelet clearance can have devastating consequences: bleeding due to thrombocytopenia is a major cause of morbidity in clinical disorders such as sepsis or cancer. ⁷ Platelets, unlike other transplantable tissues or cell types, do not tolerate refrigeration and disappear rapidly from the circulation if subjected to chilling before transplantation. ⁸ Efforts to address the problem of platelet refrigeration led to the detection of a new carbohydrate-dependent platelet clearance mechanism: cooling of platelets induced progressive loss of sialic acid, exposing terminal β-galactose residues causing their recognition and clearance by Integrin αMβ2 in macrophages and the Ashwell-Morell receptor (AMR) in hepatocytes.^9,10

Thrombopoietin (TPO) is the main hormone that supports megakaryocyte survival, maturation, and differentiation. Hepatocytes are the major source of circulating TPO; however, the mechanism regulating TPO production has been debated for decades. Using an in vitro system where we cultivated HepG2 cells in the presence of human platelets: control (not treated) or desialylated (treated with sialidase), we showed that within 6 hours there was higher ingestion of desialylated platelets, increased release of TPO protein in the culture media, and increased TPO messenger RNA (mRNA) expression in HepG2 cells incubated with desialylated platelets when compared with control platelets. We also found that JAK2-STAT3 signaling pathway was activated when desialylated platelets were present in the culture. Furthermore, the addition of AZD1480, a JAK2 inhibitor, to the cell culture inhibited TPO mRNA expression. These results suggested that ingestion of desialylated platelets by hepatocytes stimulates TPO production through JAK2-STAT3 signaling.

In vivo, we showed that mice lacking subunit2 of the AMR (Asgr2^−/−) had increased platelet count and survival time when compared with wild type (WT) mice. Surprisingly, Asgr2^−/− derived platelets showed increased loss of sialic acid and the TPO mRNA in the liver of Asgr2^−/− mice was ~40% reduced compared with WT. In contrast, St3Gal4^−/− mice, which have increased desialylated platelet uptake by the AMR, have significant increase (+30%) in liver TPO mRNA levels. Remarkably, transfusion of desialylated platelets derived from Asgr2^−/− and St3Gal4^−/− mice into WT mice significantly increased liver TPO mRNA expression followed by increase in plasma TPO when compared with mice transfused with WT-derived platelets. As we observed in vitro, treatment of WT with AZD1480 prevented liver TPO mRNA increase after transfusion of desialylated platelets. In summary, our work showed that platelets lose sialic acid as they circulate and desialylated platelets are recognized by hepatic AMR triggering JAK2-STAT3 signaling and increasing TPO production, providing a novel physiological feedback mechanism of TPO production in vitro and in vivo. ¹¹

This work also highlights the importance of understanding the regulation of platelet hemostasis in all its nuances. While we shed light into the mechanism leading to thrombocytopenia commonly observed in patients treated with JAK inhibitors, we ask new questions into other clearance mechanisms yet to be elucidated, such as the relationship between desialylation and the intrinsic apoptotic pathway.

The Effects of Lipogenesis Inhibitors on Platelet Production in Humans and Nonclinical Species

Dr William J. Reagan, a Research Fellow in Global Drug Safety Research and Development organization at Pfizer, presented the third talk on “The Effects of Lipogenesis Inhibitors on Platelet Production in Humans and Nonclinical Species.”

Acetyl CoA carboxylase inhibitors (ACCi) are being developed for the treatment of nonalcoholic steatohepatitis (NASH). Acetyl CoA carboxylase is a key enzyme for de novo lipogenesis (DNL). ¹² Inhibition of this enzyme will not only decrease DNL but also increase the β-oxidation of fats, which together will decrease lipid accumulation in the liver. The Acc inhibitor, PF-05175157, was in development for the treatment of NASH and, during a multiple ascending dose study in humans, induced an unexpected reversible thrombocytopenia with no evidence of bleeding. ¹³ This was not seen in the nonclinical studies, including 4-month dog and rat studies. The development of this compound was put on hold to better understand the pathogenesis of this finding and to try to develop a more liver-specific compound.

We showed that PF-05175157 could induce thrombocytopenia in a 2-week nonhuman primate (NHP) study. These animals also had decreased white and red blood cells, which were not seen in people. There was also decreased myeloid and erythroid cellularity in bone marrow in NHP, suggesting decreased production of white and red blood cells, respectively. Megakaryocytes were present in adequate to increased numbers, and there was no evidence of platelet consumption based on a lack of platelet activation (assessed by P-selectin expression) or significant coagulopathy (no major changes in APTT, PT, and/or fibrinogen). Ploidy analysis was also done and did not show major differences compared with the concurrent control group, suggesting that megakaryocytes were undergoing normal endomitosis. Bone marrow transmission electron microscopy indicated a decrease in the development of the megakaryocyte demarcation membrane system (DMS), which is critical for platelet production.

We showed that, by Airyscan confocal imaging, human megakaryocyte cultures treated with ACCi had a less developed DMS. In addition, flow cytometric assessment showed a decrease in large surface complex megakaryocytes, which is an indicator of the development of the DMS. The decrease in the DMS led to decreased platelet production in vitro. Serial lipogenic gene analysis was also done and showed that expression of lipogenic genes (acetyl CoA carboxylase, fatty acid synthase, and sterol regulatory element-binding protein 1) occurred mainly in early stages of megakaryocyte development; if the cells were exposed to ACCi in this early stage, they had the greatest suppressive platelet production effects compared with controls. Finally, we also demonstrated that decreased platelet production occurred in the presence of several different structurally distinct ACCis, indicating that this was likely a mechanistic effect and not chemotype-specific. Collectively, our data suggest that DNL is essential for formation of the megakaryocyte DMS in humans and NHPs, and inhibiting this process leads to a decrease in platelets. To try to circumvent this issue, another ACCi (PF-05221304) that was more liver-specific was developed and when tested in NHPs still did produce a thrombocytopenia but did not induce other cytopenias. With long-term studies in NHPs up to 16 weeks, the thrombocytopenia did not progress, and when tested in humans, effective inhibition of DNL occurred without induction of thrombocytopenia. ¹⁴ This compound is moving forward in development.

A Post-Marketing Investigation of Thrombocytopenia and Hemorrhagic Events in Patients Administered Mylotarg (Gemtuzumab Ozogamicin)

Dr Allison Vitsky is a Research Fellow in Global Drug Safety Research and Development organization at Pfizer where she concentrates on oncology drug development. She presented the fourth talk on “A Post-Marketing Investigation of Thrombocytopenia and Hemorrhagic Events in Patients Administered Mylotarg (Gemtuzumab Ozogamicin).” The full article on this topic can be found in this issue.

Gene Therapy for Hemophilia: Learnings From Hemophilia B

Debra Pittman is Senior Research Director in the Rare Disease Research Unit in the Pfizer Global Research and Development organization. She presented the fifth talk on “Gene Therapy for Hemophilia: Learnings from Hemophilia B.”

Hemophilia B is a rare X-chromosome-linked bleeding disorder characterized by inadequate levels of the intrinsic coagulation FIX. FIX deficiency prevents sufficient thrombin generation for the conversion of fibrinogen to fibrin for the development of a stable clot leading to uncontrolled bleeding. FIX is synthesized in the liver and the disease phenotype is characterized by FIX activity in the plasma with a severe phenotype of FIX activity of <1%. ¹⁵ The current standard of care for hemophilia, protein replacement, requires repeated intravenous injections. Single-dose administration of liver-directed, nonreplicating, recombinant adeno-associated virus (AAV) vectors is in clinical development for the gene therapy treatment of hemophilia. Naturally occurring AAVs are nonpathogenic and replication-deficient members of the Parvoviridae family and have become the vector of choice for hemophilia gene therapy.

The early studies in hemophilia B gene therapy have provided incremental learnings. Humoral and cell-mediated immune responses to the recombinant AAV (rAAV) vector are challenges in gene therapy. In the first AAV2 FIX liver-directed trial, circulating FIX levels were detected; however, the FIX expression was transient, presumably due to cell-mediated immune response to the transduced hepatocytes. ¹⁶ In a subsequent study, the intravenous administration of AAV8 vector in severe hemophilia and the management of the immune response resulted in persistent expression of FIX, improving the hemophilia B to mild or moderate phenotype.^17,18 Neutralizing antibodies to the capsid are present in the population and can limit cellular transduction, and AAV studies currently screen for the presence of neutralizing antibodies.

Persistent transgene expression following AAV gene therapy may require more potent vectors or blunting an immune response. A naturally occurring human FIX variant, arginine 338 changed to leucine (R338L), was identified in Padua, Italy, in a man with thrombophilia. ¹⁹ The FIX variant exhibited eight-fold increase in specific activity compared with wild-type (WT) FIX. A liver-directed nonreplicating recombinant AAV vector encoding FIX-R338L (fidanacogene elaparvovec) is in development for the treatment of hemophilia B. There are well-characterized large and small animal hemophilia disease models that have enabled the nonclinical translation studies of AAV vectors encoding the FIX-R338L variant. In studies in hemostatically normal WT mice dosed with AAV-R338L, ~7X to 10X higher FIX levels were observed compared with FIX-WT, with no significant changes in markers of thrombosis. ²⁰ In hemophilia B dogs, which exhibit a severe spontaneous bleeding phenotype, a single administration of an AAV-canine FIX-R338L showed phenotypic correction of coagulation. No elevations in thrombin-antithrombin complexes were detected at different timepoints, consistent with a nonpathological activation of coagulation. ²⁰ In the long-term follow-up phase 1/2a study of fidanacogene, elaparvovec at a dose of 5 × 10¹¹ vg/kg resulted in persistent expression of FIX associated with low rates of bleeding and infusions for a period of at least 5 years with good tolerability. ²¹ Preclinical translation models used to examine vector construct design, biodistribution, integration, expression levels, hemophilic coagulopathy correction, and safety have helped to advance the AAV hemophilia B gene therapy into clinical trials.

Preclinical Markers of Hypercoagulability: Practical Predictors or Pipe dream?

The final session was presented by Dr Marjory Brooks. She gave a summary on the translatability of biomarkers to predict thrombotic risk. The talk was titled “Preclinical Markers of Hypercoagulability: Practical Predictors or Pipe Dream?”

Cardiovascular toxicities represent an important cause of drug attrition that is often recognized only during late phase trials or after drug approval. In particular, risk assessment for drug-related thrombotic complications remains a major challenge across different drug classes and drugs used for a variety of indications. ²² The clinical manifestations of thrombosis are heterogeneous, including heart attack and stroke, deep vein thrombosis and pulmonary embolism, and vasculitides with small-vessel occlusion. Comorbidities in patient populations, such as age, obesity, metabolic syndrome, cancer, and hereditary thrombophilias, exacerbate thrombotic risk. Preclinical hemostasis testing typically consists of platelet count and coagulation screening tests. ²³ These tests will identify hemorrhagic risk due to coagulation factor deficiencies and thrombocytopenia, but are ill-suited to detect procoagulant imbalance. Furthermore, limiting the predictive value of preclinical testing is the lack of well-defined and translational biomarkers of prethrombotic states denoting hypercoagulability.

Mechanistically, hypercoagulability can result from uncontrolled procoagulant forces, or ineffective anticoagulant or fibrinolytic response. ²⁴ Thrombin, the terminal coagulation protease, plays a central role in hemostasis. Thrombin not only acts on fibrinogen to form the fibrin clot, but it is also a potent platelet agonist driving platelet aggregation and activates the plasma carboxypeptidase, thrombin activatable fibrinolysis inhibitor (TAFI). Thrombin concentration influences the rate of fibrin clot formation and the strength and stability of the mature clot. Thrombin’s actions are conserved across species; thus, detection of thrombin excess represents a translational marker of procoagulant hemostatic imbalance.

The potential capacity to generate thrombin can be measured in kinetic assays that monitor cleavage of a fluorogenic thrombin substrate. ²⁵ Thrombin generation assays (TGA) are configured for microtiter plate formats with low sample volumes and are typically performed using commercially available reagents and standards. Results of TGA include numeric parameters and qualitative thrombin generation curves, or “thrombograms.” With the use of consistent preanalytic processing techniques, TGA hold promise for preclinical studies and comparisons across testing laboratories.

Additional assays to detect conditions of in vivo thrombin excess include quantitative measurement of thrombin-antithrombin complexes (TAT) and the terminal fibrin degradation product, D dimer. Screening tests of impaired fibrinolysis for multispecies studies include kinetic turbidimetric assays and specific measurement of fibrinolytic pathway proteins such as plasminogen, antiplasmin, and plasminogen activator inhibitor. ²⁶ While aggregometry is considered the gold standard test of platelet function, it is technically complex and requires large volume samples. Flow cytometric assays overcome some of these limitations; however, the predictive utility of cytometric markers of platelet hyperreactivity is not yet well defined in any species. ²⁷

A multipronged approach combining global tests of hemostatic imbalance, qualifying assays for multispecies, and selecting preclinical study subjects to more closely match patient populations can improve the screening process to detect drug-induced thrombotic complications.

Research Involving Human Subjects and Animals

Clinical trial studies involving human participants were conducted in compliance with ethical principals originating in or derived from the Declaration of Helsinki and in compliance with all International Conferences on Harmonization of Good Clinical Practice Guidelines. Protocols were reviewed by an Independent Review Board.

Animal studies were conducted in accordance with the current guidelines for animal welfare (National Research Council Guide for the Care and Use of Laboratory Animals, 2011; Animal Welfare Act [AWA], 1966, as amended in 1970, 1976, 1985, and 1990, and the AWA implementing regulations in Title 9, Code of Federal Regulations, Chapter 1, Subchapter A, Parts 1-3). The procedures used in these studies have been reviewed and approved by the Institutional Animal Care and Use Committee.