Activated Protein C in Neuroprotection and Malaria

Abstract

Purpose of review:

Activated protein C is a homeostatic coagulation protease with anticoagulant and cytoprotective activities. Focusing on APC’s effects in the brain, this review discusses three different scenarios that illustrate how APC functions are intimately affecting the physiology and pathophysiology of the brain.

Recent findings:

Cytoprotective APC therapy holds promise for treatment of ischemic stroke, and a recently completed trial suggested that cytoprotective-selective 3K3A-APC reduced bleeding in ischemic stroke patients. In contrast, APC’s anticoagulant activity contributes to brain bleeding as shown by the disproportional upregulation of APC generation in cerebral cavernous malformations (CCM) lesions in mice. However, too little APC generation also contributes to maladies of the brain, such as in case of cerebral malaria where the binding of infected erythrocytes to the endothelial protein C receptor (EPCR) may interfere with the EPCR-dependent functions of the protein C pathway. Furthermore, discoveries of new activities of APC such as the inhibition of the NLRP3-mediated inflammasome and of new applications of APC therapy such as in Alzheimer’s disease and graft-versus-host disease continue to advance our knowledge of this important proteolytic regulatory system.

Summary:

APC’s many activities or lack thereof are intimately involved in multiple neuropathologies, providing abundant opportunities for translational research.

The homeostatic protease APC:

Activated protein C (APC) is a coagulation protease with multiple biological functions regulating coagulation, inflammation, and cell survival that are important for maintaining the homeostatic balance during health and disease (Figure 1).^1–7 Activation of protein C zymogen is mediated by the thrombin-thrombomodulin (TM) complex on the endothelial cell surface and facilitated by presentation of protein C by the endothelial protein C receptor (EPCR).^2,8 Upon activation, the serine protease APC can engage with multiple substrates to mediate distinctly different functional effects that are broadly referred to as APC’s anticoagulant and cytoprotective activities. APC anticoagulant effects of involve the proteolytic inactivation of activated blood coagulation factors V and VIII aided by cofactors such as protein S, factor V itself, and various lipids, described in numerous reviews.^9–11 Increased risks for thrombosis associated with impairments of APC’s anticoagulant pathway are well appreciated,^12–14 but recent advances also highlight that disproportional APC generation contributes to bleeding in patients with severe trauma,^15–17 hemophilia,^18,19 or cerebral cavernous malformations²⁰ (see Scenario 2 below). This information illustrates the importance of proper regulation of APC generation and activity, with both too little and too much APC contributing to pathologies. As such, recent interest in therapeutic modulation of APC activity focus on inhibition of APC anticoagulant activity to dampen bleeding.^21–25 However, it is critical to consider potential effects on APC’s cytoprotective activities when evaluating such strategies.

Figure 1: Pathways of the protein C system.

The four major pathways of the protein C system are: (1, bottom left) activation of protein C and generation of APC, (2, bottom right) the cytoprotective pathway, (3, top right) the anticoagulant pathway, and (4, top left) inactivation of APC by serine protease inhibitors (SERPINs). (1, bottom left) Activation of protein C is mediated by thrombomodulin (TM)-bound thrombin and facilitated by presentation of protein C bound to EPCR on the endothelial cell surface. (2, bottom right) After activation when APC is bound to EPCR, APC can induce cytoprotective activities via the non-canonical activation of PAR1 at Arg46 and of PAR3 at Arg41. Cytoprotective activities may include anti-apoptotic activities, anti-inflammatory activities including inhibition of NLRP3-mediated inflammasomes and inhibition of neutrophil extracellular trap formation, stabilization of endothelial barriers, and alterations of gene expression profiles. Not shown here is APC’s ability to inactivate extracellular histones, which does not require APC-mediated cell signaling but involves proteolytic degradation by APC. Also, not depicted, are other receptors required for some of APC’s cytoprotective effects, including Mac-1, S1P receptor 1, Tie2, and others. (3, top right) Dissociation of APC from EPCR and binding of APC to negatively charged lipid membranes such as on activated platelets permits anticoagulant activity of APC. Proteolytic inactivation of the activated clotting factors Va and VIIIa by APC is supported by various cofactors such as protein S, factor V, and various lipid cofactors. (4, top left) Inactivation of APC by SERPINs limits the activity of APC in plasma and is primarily responsible for the half-live of ~16 min of APC in vivo. Abbreviations: PC, protein C; APC, activated protein C; IIa, thrombin; TM, thrombomodulin; EPCR, endothelial protein C receptor; PAR, protease activated receptor; FVa, activated factor V; FVIIIa, activated factor VIII; SERPINs, serine protease inhibitors. This figure is modified from Griffin et al, ATVB, 2016.¹²⁴

Mechanisms of action for APC cytoprotective effects:

Independent of APC’s anticoagulant activities are APC’s effects on cells that are collectively referred to as APC’s cytoprotective effects (Figure 1). Multiple receptors contribute to the cellular effects of APC on different cell types, including CD11b/CD18 (MAC-1), ApoER2, Tie2, protease activated receptor (PAR) 2, sphingosine-1-phosphate receptor 1 (S1P1), but most often a combination of EPCR and PAR1 and/or PAR3 are required.^2–6,26 Beneficial effects of APC that require its cytoprotective activities have been reported in many different disease models of different organs including the brain, lung, and kidney that manifest as anti-apoptotic and anti-inflammatory effects, endothelial and epithelial cell barrier protective effects, and/or regenerative and wound healing effects.^2–7,27 While signaling pathways induced by APC show some cell-type specificity, generally, APC is understood to induce PAR1-mediated biased signaling that encompasses β-Arrestin-2-mediated signal transduction in caveolin-1-enriched caveolae.^28–31 In endothelial cells, PAR1-dependent APC signaling manifests as activation of the PI3K-Akt hub and Rac-1, after which downstream pathways provide anti-apoptotic and endothelial barrier protective effects. These signaling pathways contrast with those induced by thrombin-mediated PAR1 agonism on endothelial cells that instead involve the activation of ERK1/2 and Rho-A and manifest as proliferative and endothelial barrier disruptive responses.³¹

PARs do not rely on external ligands for activation but instead carry their own encrypted tethered ligand to be exposed after proteolytic cleavage, such as the canonical cleavage of PAR1 by thrombin at Arg41 exposing the classical “SFLLRN…” tethered-ligand that induces G-protein-dependent signaling.^32–34 The notion that some proteases activate PAR1 at other sites, resulting in different tethered-ligands with different activities, provides an explanation for the conundrum how different proteases can use the same receptor to induce sometimes very different and contrasting effects (Table 1).^3,33,34 In particular, APC induces biased agonism of PAR1 by activating PAR1 at non-canonical Arg46 thereby generating the tethered-ligand “NPNDKY…” that promotes PAR1 conformations associated with β-Arrestin-2-mediated biased signaling.³¹ Peptides (e.g. TR47) representing this N-terminal starting at Asn47 mimic APC-mediated PAR1 signaling.^6,31,35 In addition, mice with PAR1 mutations at either canonical Arg41 or non-canonical Arg46, thereby rendering them resistant to activation by thrombin or APC, respectively, demonstrated that beneficial protective effects of APC in stroke and sepsis models require PAR1 activation at Arg46.³⁶ Thus, providing in vivo proof for the non-canonical activation of PAR1 by APC and APC-induced biased signaling. Parmodulin 2, a small molecule binding to the cytoplasmic side of PAR1, also acts as a biased agonist, presumably by stabilizing PAR1 conformations associated with β-Arrestin-2-mediated signaling, and recapitulates PAR1-dependent cytoprotective signaling.³⁷

Table 1:

Canonical agonist and biased agonist peptides of PAR1 and PAR3 generated by APC and thrombin.

Peptide	PAR	Generated by	Sequence	Biological effect
TR42 (TRAP)	PAR1	thrombin	42-SFLLRNPNDKYEPFWEDEEKNESGL-66	platelet activator, proinflammatory, endothelial barrier disruptive
TR47	PAR1	APC	47-NPNDKYEPFWEDEEKNESGL-66	cytoprotective, anti-inflammatory, endothelial barrier protective
P3K	PAR3	thrombin	39-TFRGAPPNSFEEFPFSALEGWTGATIT-65	enhances PAR1-dependent ERK1/2 signaling by thrombin
P3R	PAR3	APC	42-GAPPNSFEEFPFSALEGWTGATIT-65	endothelial barrier protective, activates Tie2

In addition, non-canonical activation of PAR3 at Arg41 by APC generates the tethered-ligand “GAPPNS…” with barrier protective properties.³⁸ Similar to PAR1, PAR3 peptides (e.g., P3R), representing the non-canonical tethered-ligand generated by APC, mimic effects of APC (Table 1).^38,39 While the APC generated PAR1 (TR47) and PAR3 (P3R) peptides share some functional outcomes, such as protection of endothelial barrier function, it is becoming increasingly clear that these peptides also can induce unique signaling pathways, such as the activation of Tie2 by P3R but not TR47 peptides (Figure 2).³⁹ This suggests that APC uses PAR1 and PAR3 to diversify it signaling repertoire. New expansion of the repertoire of APC’s cytoprotective effects in diabetes and graft-versus-host disease adds to this notion.^40,41

Figure 2: Functional selectivity of PAR1 and PAR3 signaling due to non-canonical activation by APC.

Schematic portrayal of how different PAR1 and PAR3 tethered-ligands are generated that start at either residue 42 or 47 for PAR1 or at residue 39 or 42 for PAR3 depending on whether activation occurred by thrombin or APC, respectively. These canonical and non-canonical tethered ligands induce PAR conformations that selectively promote the activation of distinctly different signaling pathways, thereby providing an explanation for the functional selectivity of thrombin’s and APC’s protease signaling. For instance, the thrombin generated PAR1 tethered-ligand starting at residue 42 promotes the association of conformational subsets of PAR1 with G-proteins, resulting in activation of ERK1/2 and RhoA that manifests in endothelial barrier disruption. The APC generated PAR1 tethered-ligand starting at residue 47 however, promotes association of PAR1 with β-Arrestin, resulting in biased signaling involving activation of Akt and Rac-1 that manifests in endothelial barrier protection. The PAR3 tethered-ligand sequence starting at residue 39 can form heterodimers with PAR1 to promote endothelial disruption via ERK1/2 activation.¹²⁵ Alternatively, PAR3 can promote activation of Tie2 and vascular protective effects when the PAR3 tethered-ligand sequence starts at residue 42 due to cleavage by APC. Integration of PAR3 with PAR1 or PAR2 (not depicted here) signaling after activation by APC may also involve the formation of heterodimers but exact mechanisms remain to be elucidated. Thus, the PAR1-PAR3 activation profile of thrombin and APC represent the two ends of the spectrum for the functional selectivity of protease signaling with distinctly different cellular effects.

Inhibition of inflammasome, a new framework for APC’s anti-inflammatory activity:

Anti-inflammatory effects of APC have long been appreciated, and include inhibition of NFkB activation, changes in gene expression profiles, down regulation of adhesion molecules on endothelial cells, inhibition of neutrophil NETosis, and cleavage of extracellular histones.^42–47

The recent discovery that APC inhibits activation of the NLRP3-mediated inflammasome puts a new framework on APC’s anti-inflammatory activities.⁴⁸ Inflammasomes are intracellular sensors for infectious and sterile stressors to execute the release of inflammatory mediators.^49,50 Especially, the NOD-like receptor NLRP3-mediated inflammasome responds to both endogenous stimuli, e.g., generated by ischemia-reperfusion injury as well as infection-generated danger signals such as LPS, to activate caspase 1 and mediate the proteolytic activation and release of IL1β, and IL-18 from cells in a two-step process.^51,52 In the first priming step, activation of pattern recognition receptors upregulate inflammasome components in reactions that involve NFκB-dependent gene expression pathways.^52,53 Mechanisms for NLRP3 activation in the second step, that involve the formation of the caspase-1 activating macromolecular complex of NLRP3 with adaptor and effector proteins, remain to be fully elucidated but may result from cellular electrolyte imbalance, and/or metabolic, mitochondrial or lysosomal dysfunction.^51,52 Early data suggests that APC has potential to inhibit both the priming step, potentially by inhibition of NFκB activation, and the NLRP3 activation step.^48,53–55 Mechanisms for inhibition of NLRP3 activation by APC remain to be elucidated but may involve dampening of mitochondrial and/or metabolic dysfunction as these are activities of APC that have been previously reported.^48,56–58

Accumulating evidence implicates that inflammasome activation contributes to tissue damage in many cardiovascular diseases, including ischemia-reperfusion injury, atherosclerosis, diabetes and stroke.^50–52 This notion is further supported by the realization that in addition to innate immune cells also many other tissue resident cell types have functional inflammasomes, including endothelial cells, fibroblasts and even platelets. Based on the initial report that APC inhibits the NLRP3-mediated inflammasome,⁴⁸ an obvious question is whether protective effects of endogenously generated APC or pharmacologically administered APC may involve inflammasome inhibition in other cardiovascular diseases where the inflammasome is implicated to contribute to pathology? Research in this area is likely to see considerable developments in the next few years.

Neuroprotective effects of APC:

APC therapy provides beneficial effects in multiple rodent models of brain disease that include ischemic stroke, amyotrophic lateral sclerosis, multiple sclerosis, and recently, Alzheimer’s disease.^2,59–62 EPCR-dependent translocation of APC across the blood-brain-barrier indicates that APC effects in the brain are not limited to the vascular compartment but can also involve cellular targets within the brain.⁶³ APC’s neuroprotective effects primarily involve its cytoprotective activities that require PAR1, PAR3 and EPCR and manifest as protection of blood-brain-barrier function, inhibition of neuroinflammation, inhibition of neuronal apoptosis, and regenerative effects targeting neuronal stem cells.^2,64–66

Bleeding in the brain and hemorrhagic conversion after thrombolytic therapy is a serious concern in ischemic stroke patients,^67,68 and lessons learned from APC therapy (Xigris, Eli Lilly) in sepsis patients in the early 2000’s indicated that wild type (wt)-APC therapy is limited by its anticoagulant activity and associated bleeding risk.^69–71 Since substrates for APC anticoagulant activity (factor Va and VIIIa) differ from APC cytoprotective activity (PAR1 and PAR3), mutation of exosite residues in APC important for interactions with specific substrates permitted the generation of activity-selective APC variants.^9,72 The cytoprotective-selective APC variant 3K3A-APC with 3 Lys residues replaced with Ala in loop 37 has full cytoprotective activity yet <10% anticoagulant activity (Figure 3) and has been extensively tested in ischemic stroke and other brain disease models demonstrating beneficial effects.^2,72

Figure 3: Cytoprotective-selective 3K3A-APC.

APC, denoted on the left by a ribbon structure, is comprised of an N-terminal GLA domain (green) that binds to negatively charged phospholipids and EPCR, two EGF-like domains (blue), and the protease domain (yellow) containing the active site (red). Four glycosylation attachment sites are indicated in gray. Interaction of APC with its multiple substrates is often mediated by charged residues in surface exposed loops on APC. These so called binding exosites on APC vary for different substrates. Three positively charged lysine residues (KKK191–193) in the 37-loop of APC form a positively charged patch (see highlighted yellow box top right) important for APC’s recognition of factors Va and VIIIa. Mutation of these 3 lysine residues to alanine (3K3A-APC) eliminates this binding exosite for factors Va and VIIIa on the surface of APC (see highlighted yellow box bottom right) and results in a reduction of APC’s anticoagulant activity by 90%. In contrast, the cytoprotective substrates, PAR1, or its other known cell signaling receptors do not require these 3 lysine residues for interaction with APC and consequently 3K3A-APC retains normal cytoprotective activities. While surfaces on APC interacting with PAR1 remain to be fully defined, a potential binding exosite for PAR1 is located in the protease domain of APC (yellow) around the glycosylation attachment site at Asn329 (bottom left). Elimination of the glycosylation at Asn329 increases APC’s cytoprotective activities, whereas mutation of two nearby negatively charged glutamic acid residues at 330 and 333 to alanine eliminates APC’s cytoprotective activities.^126,127 This suggests that exosites for APC’s anticoagulant and cytoprotective substrates are located on different and opposite sides of the protease domain thereby providing an explanation why 3K3A-APC is cytoprotective-selective. This figure is modified from Griffin et al, Blood, 2018.²

Activation of the NLRP3 inflammasome has been implicated to promote disease progression of numerous brain maladies, including neuroinflammation and post-ischemic reperfusion injury after ischemic stroke.^50,73 Inflammasomes have been described in microglia, astrocytes and neurons and some of APC’s effect in the brain may involve inhibition of inflammasome activation. 3K3A-APC efficiently inhibited NLRP3 inflammasome in the heart and kidney and 3K3A-APC therapy reduced brain bleeding associated with thrombolytic therapy in a rodent stroke model,^48,74 similar to attenuation of NLRP3.⁷⁵ Encouraging initial results for NLRP3 inhibition demonstrating neuroprotection in rodent stroke models highlight the potential important contribution of inflammasome activation in the pathogenesis of ischemic stroke and potential for therapeutic approaches.^76,77 However, it is not known whether inflammasome inhibition by APC contributes to the repertoire of its neuroprotective effects.

Three scenarios for APC effects in the brain:

As the activities of APC are involved in multiple reactions that govern homeostasis, its intimate involvement in brain pathophysiology is therefore not surprising. For better or for worse, the next three scenarios highlight recent new insights into the contributions of the protein C pathway in the brain.

Scenario 1: Cytoprotective-selective APC therapy for ischemic stroke.

Clinical observations support a potential protective effect of protein C in stroke. Prospective epidemiologic studies found an inverse association of protein C levels with the incidence of stroke and circulating APC levels are decreased in stroke patients.^78–80 These clinical associations and multiple studies demonstrating neuroprotective effects of APC in rodent stroke models prompted the translation of the cytoprotective-selective 3K3A-APC variant for therapy in ischemic stroke patients. Preclinical studies in Cynomolgus monkey confirmed an approximately 10-fold difference in anticoagulant activity between 3K3A-APC and wild type-APC (Xigris) in vivo.⁸¹ Initial clinical studies in healthy subjects indicated that 3K3A-APC at multiple high doses (up to 540 μg/kg every 12 hours for 5 doses) was well tolerated with minimal prolongation of APTT clotting times.⁸² Administration of 3K3A-APC at 540 μg/kg in a single iv bolus resulted in peak plasma levels of ~4300 ng/ml with an elimination half-life of ~16 min. The reduced anticoagulant activity of 3K3A-APC, permitting high transient levels of cytoprotective-selective APC in the circulation to reset signaling pathways in stressed cell, highlights the key conceptual difference between 3K3A-APC therapy and previous wild type-APC (Xigris) therapy that was targeting anticoagulation.

The recently completed NeuroNEXT (Rhapsody) trial, a phase 2A randomized, controlled, blinded, dose-escalation safety trial for 3K3A-APC in ischemic stroke patients, reported encouraging results.⁸³ When 3K3A-APC (120–540 μg/kg) was administered in a 15 min bolus starting 30–120 min following tPA or mechanical thrombectomy every 12 hr for 5 doses, exploratory analysis suggested that 3K3A-APC reduced intracranial hemorrhage rates in the combined treatment arms compared to placebo from 86.5% to 67.4% (p=0.046) and total hemorrhagic volume from 2.1 ml to 0.8 ml (p = 0.066). While this trend is encouraging, a larger trial is needed to confirm these results.

Scenario 2: Bleeding in the brain due to excessive APC generation.

Organotypic vascular bed heterogeneity is determined by function and multiple humoral and biophysical factors.^84,85 The vascular bed of the brain is programmed to be prohemostatic, presumably to avert hemorrhage, with relatively low endothelial expression of TM and EPCR.^86,87 Recent insights indicate that this status quo is disrupted in vascular malformations in the brain.

Cerebral cavernous malformations (CCM) are hemorrhagic brain lesions due to clusters of low-flow dilated capillaries that lack a proper vessel architecture. Familial forms of CCM are due to loss-of-function mutations in one of three genes: KRIT1 (CCM1), CCM2 or PDCD10 (CCM3), and patients with CCM are at high risk of developing hemorrhagic stroke, seizures and neurological deficits.^88,89 Comprehensive transcriptome analyses of CCM lesions have provided new mechanistic insights.^90–92 Lopez-Ramirez et al. noted that transcripts for TM and EPCR were increased upon loss of KRIT1 or PDCD10 and demonstrated abnormally high endothelial TM and EPCR expression in CCM lesions, suggesting that disproportional APC generation in CCM lesions may promote hemorrhaging.^20,90 Indeed, deletion of endothelial TM and blocking antibodies to TM and EPCR significantly reduced CCM-associated bleeding in mice. These and other studies, e.g., post-ischemic bleeding induced by anticoagulant-selective E149A-APC,⁹³ indicate that a stressed blood-brain-barrier is particularly susceptible to anticoagulant APC-facilitated bleeding.

On the other hand, enhanced generation of APC at the site of lesions may serve to restore vascular integrity via its cytoprotective activity that require PAR1 and PAR3, especially since endothelial tight-junctions have deteriorated in CCM lesions.⁹⁰ The enhanced expression of TM was induced by Krüppel-like transcription factors KLF2 and KL4,^20,94 which are known to be upregulated in CCM and implicated in the pathogenesis.^90,95 These flow-regulated transcription factors normally induce an anti-inflammatory, anti-thrombotic, quiescence state of the endothelium and promote endothelial barrier stabilization,^96–98 but in their quest for endothelial quiescence also downregulate expression of PAR1, PAR2 and PAR3, presumably to minimize proinflammatory effects of thrombin and tissue factor-mediated signaling.^{90,94,99–101} Whether the KLF2/4-mediated downregulation of PAR1 and PAR3 diminishes APC’s ability to induce cytoprotective effects remains to be determined, but it is possible that upregulation of KLF2/4 in CCM lesions facilitate an (unintended) shift in the balance between anticoagulant and cytoprotective effects of APC.

Scenario 3: Cerebral malaria due to an acquired protein C pathway defect.

Vascular sequestration of infected erythrocytes (IEs) is the hallmark of cerebral malaria.^102–104 Sequestration is mediated by the P. falciparum erythrocyte membrane protein 1 (PfEMP1), a parasite-encoded protein family expressed on IEs that mediates the binding to various vascular receptors thereby providing a protected environment for the parasitic biomass to multiply without splenic clearance. An unexpected link between severe malaria and the protein C pathway came from the discovery that some PfEMP1 variants bind to EPCR and that these infections were associated with the most severe forms of malaria, including cerebral malaria.^105–111 Increased awareness that vascular dysfunction underlies the symptomatic manifestations of severe malaria were fueled by observations that expression of TM and EPCR is deceased at sites of IE sequestration in the brain and that brain edema is associated with mortality in children with cerebral malaria suggesting a deterioration of blood-brain-barrier function.^106,112 It is important to note that expression of EPCR-binding PfEMP1 on IEs is restricted to P. falciparum parasites. Other Plasmodium strains that infect humans or mice do not express PfEMP1 molecules and therefore do not encompass the binding of IEs to EPCR in their pathogenesis.¹¹³

The PfEMP1 domains that are responsible for binding to EPCR (i.e., Cysteine-rich interdomain regions (CIDRα1)) block protein C and APC binding to EPCR and inhibit EPCR-facilitated protein C activation, activation of PAR1 and PAR3 by APC, and APC’s ability to induce EPCR-dependent cytoprotective effects, including endothelial barrier protection.^105,114,115 EPCR has important functions in the protein C system and the activities of APC are important to maintain homeostasis in the brain.^2,8,116 Human protein C deficiency, when left untreated, results in neurological defects,^117,118 and the brain is very susceptible to endotoxemia-induced vascular permeability when the expression of EPCR is compromised.¹¹⁹ Furthermore, transcripts of a EPCR-binding CIDRα1 domain that inhibited APC’s barrier protective effects were highly enriched in children with cerebral malaria and brain swelling.¹²⁰ Thus, there are multiple mechanistic and pathologic indications supporting a defective protein C system in severe malaria and that this may contribute to the pathogenesis of cerebral malaria.^{116,121–123}

Conclusion:

These scenarios illustrate that the multiple activities of APC, for better or for worse, are intimately involved in the normal physiology and pathophysiology of the brain. This provides many opportunities for relevant translational research focusing either on APC’s anticoagulant activity to mitigate bleeding or on APC’s cytoprotective activity to restore normal brain function. The continuous discovery of new molecular mechanisms for APC’s cytoprotective activities, new applications for APC therapy, and preliminary beneficial effects of APC therapy in ischemic stroke patients, are encouraging that cytoprotective-selective APC therapy may become ultimately successful and may expand to other diseases of the brain in addition to ischemic stroke.

Key points:

• APC is a homeostatic coagulation protease with anticoagulant and cytoprotective activities that are intimate involved in multiple neuropathologies.

• APC therapy with the cytoprotective-selective 3K3A-APC variant showed encouraging results in ischemic stroke patients.

• Disproportional upregulation of APC anticoagulant activity can contribute to bleeding associated with cerebral cavernous malformations.

• A dysfunction of the protein C pathway due to targeting of the endothelial protein C receptor (EPCR) by infected erythrocytes likely contributes to the pathogenesis of cerebral malaria.