Linköping University Medical Dissertations No 1261
The role of platelet thrombin
receptors PAR1 and PAR4 in
health and disease
Martina Nylander
Division of Clinical Chemistry
Department of Clinical and Experimental Medicine Linköping University, Sweden
Linköping 2011
© Martina Nylander, 2011
Published papers are reprinted with the permission from the copyright holder.
The role of platelet thrombin receptors PAR1 and PAR4 in health and disease
Cover: A drawing made by the author, illustrating PAR1 & PAR4 cell signaling. Printed in Sweden by LiU-tryck, Linköping, Sweden, 2011
ISBN: 978-91-7393-067-3 ISSN: 0345-0082
”Life is a mystery”
-Julien Offray de La Mettrie
ABSTRACT
Blood cells are continuously flowing in our systems maintaining haemostasis in the arteries and veins. If a vessel is damaged, the smallest cell fragments in the blood (platelets) are directed to cover the wound and plug the leakage to prevent blood loss. Most of the time platelets stop the blood leak without any difficulties. During other, pathological, circumstances, platelets continue to form a thrombus, preventing the blood flow and may cause myocardial infarction or stroke.
Thrombin is the most potent platelet agonist and is a product created in the coagulation cascade. This thesis is focused on the interactions between the two platelet thrombin receptors; protease activated receptors 1 (PAR1) and PAR4 in vitro. We have investigated potential differences between these receptors in several situations associated with cardiovascular disease.
First we studied interactions between PAR1 and PAR4 and the oral pathogen Porphyromonas
gingivalis (which secretes enzymes, gingipains, with properties similar to thrombin). Here we
showed that P. gingivalis is signaling mainly, but not exclusively, via PAR4. Our second study showed that the cross-talk between the stress hormone epinephrine and thrombin occur exclusively through PAR4 if the key-substance ATP is present and cyclooxygenase-1 inhibited by aspirin. The third study investigated platelet secretion, with focus on the protein plasminogen activator inhibitor 1(PAI-1), an inhibitor of the fibrinolytic process responsible for dissolving a formed clot. Here we showed that PAI-1 secretion and synthesis was more sensitive to stimulation through PAR1 than PAR4. Finally this thesis describes differences between PAR1 and PAR4 in cell-signaling pathways regulating the stability of a platelet aggregate, where PAR4 seems to be of importance to create stable platelet aggregates and that this stability is dependent on ADP activation via P2Y12 and cell signaling via PI3-kinase.
Until now, PAR1 has been considered to be the most important thrombin receptor, due to its high affinity for thrombin. However, there must be a reason why platelets express two different thrombin receptors. This thesis highlights several situations where PAR4 plays a complementary and important role in platelet signaling and haemostasis.
In conclusion, this thesis suggests that PAR4 plays a major role in calcium signaling and the induction of sustained aggregation, while PAR1 shows a more prominent role in platelet
secretion and synthesis. This thesis also reveals new interactions between platelet thrombin receptors and the ADP-, ATP- and epinephrine receptors. The results described in this thesis contribute to an increased knowledge of the platelet thrombin receptors and their interplay in situations such as infection, stress, fibrinolysis, and platelet aggregation.
POPULÄRVETENSKAPLIG SAMMANFATTNING
I människokroppen flyter blodet genom kärlen för att upprätthålla homeostasen och andra jämvikter. Ibland händer det att en skada uppstår i kärlväggen, och det första som händer är att blodplättarna (trombocyterna) samlas vid skadan och fäster samt sprider ut sig och bildar en plugg för att stoppa blödningen. När skadan väl är tilltäppt så kommer läkningsprocessen att börja återställa alla kärlväggsceller och blodflödet igen. Ibland händer det dock att processen inte balanseras på rätt sätt, och följden kan då bli att trombocytpluggen blir större och till slut bildar en trombos, en blodpropp, som i sin tur kan orsaka hjärtinfarkt eller slaganfall..
Trombin är ett enzym som bildas när koagulationskaskaden aktiveras, och dess funktion är att aktivera trombocyterna och aktivera koagulationen ytterligare för stabilisering av den pågående temporära trombocytpluggen. Trombocyten i sig har två specifika receptorer som trombin binder in till, PAR1 och PAR4, och det är skillnaden mellan dessa receptorer som undersökts i avhandlingen.
Avhandlingen beskriver trombocyten och dess trombinreceptorer PAR1 och PAR4 och visar på ett antal situationer där dessa spelar olika roller i trombocyten. Trombocyter tillsammans med bakterien Porphyromonas gingivalis samt stresshormonet adrenalin visar sig ge upphov till en trombocytaktivering främst via PAR4, och till en mindre del via PAR1. I trombocyter som behandlats med aspirin visar sig adrenalin också kunna ge upphov till trombocytaggregation i närvaro av mycket låga doser av trombin, denna gång via PAR4 och inte alls via PAR1.
Avhandlingen tar även upp skillnader mellan PAR1 och PAR4 för frisättningen av ett viktigt protein som förhindrar nedbrytningen av blodkoagel i kärlet, PAI-1. Här visar vi att PAI-1-frisättning är känsligare för aktivering via PAR1 än PAR4. Slutligen visar denna avhandling att det finns tydliga skillnader mellan trombinreceptorerna vad gäller trombocytaggregatens stabilitet där PAR4 verkar spela den största rollen för att stabilisera och bibehålla ett bildat aggregat.
Om man betraktar det fylogenetiskt och patofysiologiskt så borde det finnas en orsak till varför trombocyten har två trombinreceptorer. Tidigare har det föreslagits att PAR1 skulle vara den enda trombinreceptor som är av betydelse inom hemostas- och trombosområdet. I
denna avhandling visar vi att den andra trombinreceptorn, PAR4, också kan vara av betydelse för trombocytsignalering, trombocytaktivering och hemostas.
Avhandlingens slutsats är att trombinreceptorn PAR4 kan spela en större roll för trombocytaktivering och proppbildning än vad man tidigare trott. Avhandlingen beskriver även nya skillnader mellan PAR1 och PAR4 och deras interaktioner med ADP-, ATP- och adrenalin-receptorerna. Resultaten från avhandlingen bidrar till mer förståelse för trombocyten och dess trombinreceptorer vid situationer såsom infektion, stress, fibrinolys och trombocytaggregation.
TABLE OF CONTENTS
ABSTRACT ………. 5 POPULÄRVETENSKAPLIG SAMMANFATTNING ……… 7 ABBREVISTIONS ……….. 10 LIST OF PAPERS ………... 13 INTRODUCTION ……….. 15 Haemostasis ………... 15 Platelets ……….. 17 Protease-activated receptors ……… 18 Purinergic receptors………. 25 Adrenergic receptors………... 26Haemostasis & Thrombosis ……….. 26
AIMS OF THE THESIS ………. 29
METHODOLOGY……….. 31
Materials ……… 31
Isolation of human platelets……….. 32
Measurement of platelet aggregation ……….. 32
Culture and preparation of Porphyromonas gingivalis ……….. 32
Measurement of cytosolic calcium ………... 33
Measurement of dense granule secretion ……… 33
Western Blot ……….. 33
Fluorescence Microscopy ………. 35
Enzyme-linked immunosorbent assay ……… 35
mRNA assay ………... 35
Experimental design ………. 36
RESULTS AND DISCUSSION ……….. 39
PAPER I ... 39 PAPER II ... 40 PAPER III ... 41 PAPER IV ……….. 42 GENERAL DISCUSSION ………. 43 PRINCIPAL FINDINGS ……… 51 ACKNOWLEDGEMENTS ………... 53 REFERENCES………. 55
ABBREVIATIONS
AC adenylyl cyclase
Ab antibody
ACD Acid citrate dextrose ADP adenosine 5’-diphosphate ASA acetyl salicylic acid ATP adenosine 5’-triphosphate bFGF basic fibroblast growth factor cAMP cyclic adenosine monophosphate
COX cyclooxygenase
DAG 1,2 diacylglycerol DVT deep vein thrombosis
ELISA Enzyme-linked Immunosorbent Assay GAPDH glyceraldehyde 3-phosphate dehydrogenase
GP glycoprotein
GPCR g-protein coupled receptor
GP glycoprotein
HEPES N-[2-hydroxyethyl]-piperazine-N’-[2-ethanene-sulfonic acid] IP3 inositol 1,4,5-triphosphate
Kgp lysine-specific protease KRG Krebs-Ringer glucose solution mRNA messenger Ribonucleic acid NCCE non-capacitative Ca2+-entry
NO nitric oxide
OMV outer membrane vesicles PAI-1 plasminogen activator inhibitor 1 PAR1 protease-activated receptor 1
PAR1-AP PAR1-activating peptide (amino acid sequence: SFLLRN) PAR4 protease-activated receptor 4
PAR4-AP PAR4-activating peptide (amino acid sequence: AYPGKF) PDGF platelet-derived growth factor
PGI2 prostacyclin = prostaglandin I2
PI3-K phosphatidyl inositol 3 kinase
PKC protein kinase C PRP platelet rich plasma
PS phosphatidylserine
Rgp arginine-specific proteases SDS sodium dodecyl sulfate SOCE store-operated calcium entry tPA tissue plasminogen activator TRAP thrombin receptor activating peptide TXA2 thromboxane A2
VEGF vascular endothelial growth factor vWF von Willebrand factor
LIST OF PAPERS
This thesis is based on the following papers, which will be referred to by their roman numbers.
Paper I. Martina Nylander, Tomas L. Lindahl, Torbjörn Bengtsson, Magnus Grenegård.
The Periodontal Pathogen Porphyromonas gingivalis Sensitises Human Blood Platelets to Epinephrine.
Platelets 2008; 19(5): 352-358
Paper II. Magnus Grenegård, Karin Vretenbrant-Öberg, Martina Nylander, Stéphanie
Désilets, Eva G. Lindström, Anders Larsson, Ida Ramström, Sofia Ramström, Tomas L. Lindahl. The ATP-gated P2X1 Receptor Plays a Pivotal Role in
Activation of Aspirin-treated Platelets by Thrombin and Epinephrine.
JBC 2008; 283 (27): 18493-18504
Paper III. Martina Nylander, Sofia Ramström, Abdimajid Osman, Emma Åklint, Anders
Larsson, Tomas L. Lindahl. The role of thrombin receptors PAR1 and PAR4
for PAI-1 storage, synthesis and secretion by human platelets.
Submitted manuscript
Paper IV. Martina Nylander, Knut Fälker, Sofia Ramström, Magnus Grenegård, Tomas L.
Lindahl. Release of ADP or PAR4 Activation is Required to Sustain
Thrombin-induced Platelet Aggregation
INTRODUCTION
Haemostasis
Blood is a very complex solution consisting of different blood cells, plasma, and plasma proteins flowing through the vessels to maintain oxygen levels in tissues and to remove carbon dioxide. Other functions of the blood are to transport essential substances such as glucose and hormones. The word haemostasis comes from Greek words meaning “blood” and “stop” and describes the processes that maintain equilibrium in the blood system and in the blood vessels. The platelets are one of the essential components to ensure haemostasis. Platelets are small cell fragments flowing in the blood, always searching for damages in the vessel wall. They may also be triggered by other situations such as infection, inflammation, high or low concentrations of different substances. From a pathophysiological point of view, insufficient platelet activation and coagulation leads to bleeding, whilst misdirected and powerful platelet activation and coagulation results in thrombosis which can occlude a vessel, stop blood flow and may elicit myocardial infarction or stroke.
Beyond normal haemostasis, different blood disorders and sickness can develop. One of the major health risks in the western world today are cardiovascular disease and thrombosis. There are different risk factors suggested to contribute to cardiovascular disease such as smoking 1, excessive intake of alcohol 2, bad diet 3,4, obesity 5, lack of exercise 6, infection 7,
stress, and elevated levels of PAI-1 8-11. Why development of cardiovascular disease and
thrombosis occur is not yet fully known. No single risk factor can be pointed out as the cause, but adding several risk factors together will significantly increase the risk of circulatory events. Excessive platelet activation is mainly connected to arterial thrombosis and is often triggered by the rupture of an atherosclerotic plaque in the vessel wall. There are still many remaining questions concerning the development of circulatory disease. Therefore, it is important that research in haemostasis and atherothrombosis is progressing and that involving mechanisms are clarified, in order to reduce the incidence and to improve therapy.
It is known that infection is one of the risk factors contributing to the development of cardiovascular disease. There are many reports on pathogens infecting the circulatory system and where the presence of pathogens is correlated to cardiovascular disease. Both viruses and bacteria species have been detected in different groups of cardiovascular high risk patients. Pathogens found within these patient groups include cytomegalovirus (CMV)12,13, herpes
pathogen Porphyromonas gingivalis17. Many pathogens are known to interact with platelets and other blood cells. Bacteria may interact with platelets by i) direct binding of bacterial proteins to platelet receptors, or by ii) bacteria-secreted virulence factors binding and activating platelets, both pathways resulting in thrombus formation 18-20. Porphyromonas
gingivalis (P. gingivalis) is a well-known bacterium causing the chronic periodontal disease periodontitis. P. gingivalis is a Gram-negative, black pigmented, non-motile anaerobic bacterium21,22. This bacterium can be found in the oral cavity, in the gum, in the tooth pockets and the chin 23. In the past few years, P. gingivalis has been found in atherosclerotic plaques24, and may thereby be connected to cardiovascular disease25. P. gingivalis may migrate through
the epithelial cell barrier in order to hide from the immune response of the host22,26. P. gingivalis contains cysteine proteases, gingipains, that cleave proteins to smaller peptides
used for growth27, to facilitate the invasion of host tissue, to regulate the host immune
defense23 or to attach to human erythrocytes to achieve heme28,29. These extracellular proteases are found in high concentrations in the cell surface membrane and in the outer membrane vesicle of P. gingivalis. Arginine-specific R-gingipains (Rgp-A and Rgp-B and lysine-specific K-gingipains (Kgp) are the most common gingipains30. It has been shown that
gingipains exhibit “thrombin-like” activity, mimicking thrombin24. Gingipains seem to have
different roles during colonization. Gingipains can down-regulate the kallikrein system, the complement pathway and also the coagulation cascade. P. gingivalis and gingipains have also been shown to initiate the coagulation cascade by interacting with prothrombin, factor X and protein C31,32.
Increased levels of catecholamines (e.g. epinephrine, which is also called adrenaline) have been associated to cardiovascular disease33-36. During stress, e.g. mental stress and exercise,
elevated levels of catecholamines are found in plasma, and particularly during myocardial infarction37. Epinephrine infusion can enhance thrombin-induced fibrinogen binding and
aggregability of blood platelets34, ,38 39. There is also an increased incidence of thrombosis in patients with atherosclerosis and increased sympathetic system activation40.
After a clot has been formed and the vessel wall is repaired, it is time to start the breakdown of the clot, the fibrinolysis. The protease controlling this action is called tissue plasminogen activator (tPA) and functions by activating the enzyme plasminogen, which will degrade the fibrin network into fibrin fragments, which will dissolve the clot and restore normal blood flow41. tPA has a specific inhibitor in plasma; plasminogen activator inhibitor 1 (PAI-1)41. If there are high levels of active PAI-1 in plasma, tPA will bind to PAI-1 and become inactive,
and fibrinolysis will be prevented42. It has been reported that platelets contain plasminogen activator inhibitor-1 (PAI-1)43,44. A comparison between arterial vessel thrombi and venous
vessel thrombi reveals that arterial thrombi contain more platelets and are also more resistant to fibrinolysis, suggesting that platelets may contribute to fibrinolysis resistance45.
Platelets
Platelets are small anuclear cell fragments derived from its progenitor cells, the megakaryocytes. Platelets have a life span of between seven and eleven days in the blood system. The concentration of platelets in the blood is normally 150-350 x 109 cells/l, and the
size is approximately 2 µm46. Platelets undergo different stages during activation. When
platelets are circulating in the blood in an inactivated stage, they have a discoid shape. Due to their small size and due to the shear forces in the vessel47 they are circulating close to the vessel wall, always searching for damages, and must within a millisecond recognize, attach and anchor to the injury. After this stage, a platelet’s task is to cover the surface by flattening and pseudopodia formation, to secrete their granule contents and finally to recruit more platelets into a platelet aggregate and to form a platelet plug covering the wounded site48.
There are three types of granules in platelets; lysosomes, dense bodies and α-granules. During platelet activation, granule contents are secreted and enhance platelet activation. Dense bodies contain e.g. adenosine 5’-diphosphate (ADP)49-51, adenosine 5’-triphosphate (ATP), calcium, and serotonin. α-granules contain larger proteins such as von Willebrand factor (vWF), fibronectin, and different coagulation factors (f V, VII, XI, XIII)46. It has been reported that
there are at least two different subtypes of granules in platelets, and that one subtype of α-granule contains and releases pro-angiogenetic substances, e.g. platelet derived growth factor (PDGF), and basic fibroblast growth factor (bFGF), while the other type releases anti-angiogenetic substances, e.g. endostatin, and platelet factor-4, depending on activation52. During normal physiological conditions, endothelial cells release platelet inhibitors such as nitric oxide (NO) and prostacyclin (PGI2)53,54 to maintain equilibrium.
Calcium signaling is essential in all cells, as well as platelets. Platelet activation by agonists such as thromboxane A2 (TXA2), ADP, collagen or thrombin leads to a phospholipase C
(PLC)-mediated production of inositol 1,4,5-triphosphate (IP3) and 1,2 diacylglycerol (DAG),
via phosphatidylinositol 4,5-biphosphate (PIP2). IP3 causes a calcium increase by release of
calcium from the intracellular stores in platelets, whilst DAG is involved in calcium intake from the extracellular compartment, and are illustrated in figure 2. There are two major
pathways controlling cytosolic calcium; store-operated calcium entry (SOCE) and non-capacitative Ca2+-entry (NCCE). There are two key regulators in SOCE, STIM1 and Orai1. In
NCCE, one of the key players is P2X1. These together regulate the calcium transport in and
out from intracellular stores and the extracellular compartment (reviewed by Varga-Szabo et
al 55).
In response to calcium mobilization, negatively charged phospholipids will become exposed on the cell surface of the platelet membrane. Phosphatidylserine (PS) is a part of the phospholipids and acts as the link between platelet activation and prothrombin activation. Exposed PS binds to prothrombin and factor Va, which in turn binds to factor Xa and forms the prothrombinase complex which efficiently converts prothrombin to thrombin (reviewed by Lentz et al.56). PS also supports the assembly of a complex named tenase (e.g. “ten”-ase)
which converts factor X to Xa after forming a complex consisting of factor IXa, its co-factor VIII, and factor X57.
Protease-activated receptors
The enzyme thrombin (also called coagulation factor IIa) is a serine protease, which is activated during the coagulation cascade and has a central role in normal haemostasis and thrombus development by converting the soluble plasma protein fibrinogen to an insoluble fibrin gel and also to activate platelets. The activating effect on platelets is detectable at thrombin concentrations in the pico-/nanomolar range. These concentrations are much lower than what is needed for thrombin to generate fibrin58,59. Thrombin is a very potent platelet agonist, acting via glycoprotein (GP) Ibα and protease-activated receptors (PARs) 1 and 448.
In 1991, PAR1 was the first of the PARs to be reported on platelets60, and in 1998 PAR4 was
found to be yet another platelet thrombin receptor61,62. In 1999, PAR1 was stated to be the
most important thrombin receptor and PAR4 to be a weak receptor of less importance63. At
this time, PAR1 had been known in platelets for almost a decade, and the first papers regarding PAR4 had just been released.
Thrombin activates PARs in an intriguing and fascinating way. A part of the N-terminal exodomain of PARs is cleaved off by thrombin64, and the new unmasked N-terminal then
serves as a tethered ligand65 which in turn activates the receptor 66,67, as can be seen in figure
receptors, called PAR-activating peptides, are now frequently used for research. The sequence of the peptide mimicking the new N-terminal of PAR1 is SFLLRN60. The new N-terminal of
PAR4 has the sequence GYPGQV68, but for a more potent PAR4 activation, the (mouse-)
peptide AYPGKF is often used61, ,62 69.
Figure 1: Thrombin cleaving a protease-activated receptor (PAR). Thrombin binds to the N-terminal of the PAR receptor, and cleave off an amino acid sequence between Arg41 and Ser42 in PAR170 and Arg48 and Gly49 in PAR468, creating a new N-terminal acting as a tethered ligand to the receptor. The amino acid sequences of the new N-terminals are SFLLRN and GYPGQV for PAR1 and PAR4, respectively.
Today, there are four known protease-activated receptors; PAR1, PAR2, PAR3 and PAR4, whereof two, PAR1 and PAR4, are found on human platelets60,62. PAR3 is reported to be
expressed in mouse platelets but not in human platelets71. Thrombin cleaves PAR1, PAR3 and
PAR4, but not PAR2, which is cleaved and activated by trypsin72. Platelet thrombin receptors
Figure 2: Schematic illustration showing cell signaling via G-protein coupled receptors in platelets. Three different types of G-protein coupled receptors are presented in this image; G12/13, Gq, and Gi.
The names of receptors are underlined, agonists are lighted in green, and antagonists are high-lighted in red.
There is an agreement that both PAR1 and PAR4 couple directly to G12/13 and to Gq73, but
there are also reports stating that PAR1, but not PAR4, may couple directly to Gi74. It is
reported that adhesion and platelet shape change is both dependent75-79 and not dependent74 of
G12/13 and further downstream Rho-kinase signaling, leading to myosin rearrangement of the
cytoskeleton, enhancement of platelet activation due to a rise in intracellular calcium concentrations through the calcium/calmodulin signaling pathway, and also activation via Rho/Rho-kinase and MLC (myosin light chain)-kinase75, ,76 79-81, as is illustrated in figure 2. Klages et al.79 state that receptor-induced platelet shape change is signaling via G12/13 with its
downstream Rho/Rho-kinase regulating MLC phosphorylation. Voss et al.74 state that
PAR-mediated GPIIb/IIIa activation (as detected by PAC-1 binding) does not require G12/13
-Rho-Kinase activation. On the other hand, Dorsam et al.77 2002 and Woulfe et al.78 2005 state that
both G12/13 and Gq plays a critical role in GPIIb/IIIa activation. It is also stated that p38,
situated downstream of G12/13, is required for actin polymerization82 leading to shape change,
which is the first sign of platelet activation.
PAR1 and PAR4 also couple to Gαq, which signals through activation of phospholipase C β
(PLCβ) and further protein kinase C (PKC) which results in an increase in intracellular calcium. According to Gabbeta et al.83 and Offermanns et al.84, PLCβ (downstream G
q) is
essential for secretion induced by platelet agonists.
Even though PAR1 and PAR4 are signaling through the same GPCR's, there are reported differences between these two thrombin receptors (also illustrated in figure 3). In 1999, Andersen et al. stated that PAR1 is the major thrombin receptor, and that PAR4 is a weak receptor and of less importance for platelet activation63. Another distinct and interesting
difference reported between PAR1 and PAR4 is the affinity for thrombin; PAR1 binds thrombin with a higher affinity than to PAR461, , ,70 80 85. This difference in affinity may be one
of the reasons to why PAR4 was believed to be a weak receptor in platelet activation, but despite this, we have previously reported that PAR4 plays a role in the early phases of platelet-accelerated thrombin generation and coagulation, suggesting a role even at low concentrations of thrombin86.
Marjoram et al. saw an enhanced collagen binding to α2β1 via Gq and PLC-dependent (and not
PI3-K dependent) signaling, induced mainly by PAR4, and to a lesser extent by PAR187. It has
co-factor for PAR1, but not for PAR488,89, the importance for this GPIbα/PAR interaction is still unclear.
Covic and colleagues showed that PAR1 triggers a rapid intracellular calcium increase, whereas PAR4 triggers a slower, but prolonged increase in intracellular calcium80. Similarly,
Mazharian et al. showed that PAR4 induces a sustained calcium mobilization required to achieve full platelet spreading on a fibrinogen surface82. These findings are confirmed by Leger et al.90 and Shapiro et al.91, who show that PAR1 activation induces a rapid but transient rise in calcium, and that PAR4 activation results in delayed but sustained calcium mobilization, Leger et al.90 also propose that PAR1 activation may facilitate PAR4 cleavage
by thrombin. In a very recent report by Harper and Poole92, they show that calcium
originating from the extracellular compartment and not from the intracellular stores is required for thrombin-mediated PS exposure93, they further report that activation of PAR1 but not PAR4 increases calcium signaling and PS-exposure if signaling via PKC is blocked.
In 2002, Covic et al. reported that PAR4 activation of platelets from a patient with Hermansky-Pudlak Syndrome (HPS) could substitute for the stabilizing effect of ADP on platelet aggregation94. This study showed that PAR4 was capable of inducing an irreversible
aggregation without the autocrine feedback loop of ADP activating its own receptor P2Y12. In
2007, Reséndiz et al. showed that activation of PAR1 induced rapid Akt phosphorylation and that PAR4 induced prolonged Akt phosphorylation independently of ADP and PI3-K activation95. Recently, Wu et al. reported that PAR4 and its signaling pathway is important for
stabilization of a platelet aggregate, and that PAR1 requires PI3-kinase signaling to induce an irreversible aggregation96. In contrast to many others, Holinstat et al. stated that
PAR4-induced aggregation did not occur if both P2Y12 and calcium mobilization was blocked,
which was not the case for PAR1. However, both PAR1 and PAR4 were able to induce aggregation independent of calcium mobilization97.
It is also discussed whether platelet thrombin receptors signal through Gi or not. Voss et al.74
reported that PAR1-induced aggregation was abolished when PI3-K was inhibited with inhibitors wortmannin or LY294002. They stated that PAR1 activates PI3-K by a direct coupling to Gi, but that this is not the case for PAR4. On the other hand, Kim et al.98 claims
It is known that PAR1-induced aggregation may be desensitized by homologous activation with PAR1-AP99,100. This is in agreement with our previous findings that PAR1 but not PAR4
is somehow desensitized and PAR1-mediated aggregation abolished during homologous receptor activation101. In this paper, we showed that PAR1 is not desensitized by removal of
receptors from the cell membrane, but that the receptor-signaling is desensitized.
Finally and very contradictive, the presence of a fourth thrombin receptor is suggested by Lova102 et al., who showed that if GPIbα is cleaved by mocharagin and PAR1 and PAR4 desensitized (in platelet rich plasma using high concentrations of PAR-activating peptides, meaning that all presently known thrombin receptors are inhibited, either physically or by enzyme digestion), actin polymerization, cytoskeleton reorganization, and platelet aggregation still occur upon exposure to thrombin, but how is still not known102.
Figure 3: A schematic drawing illustrating previously reported differences between PAR1 and PAR4 signaling in human platelets.
Purinergic receptors
Platelets express three different kinds of purinergic receptors, and they are divided into two groups; P2Y and P2X. The first purinergic receptors to be found and described were the ADP-activated P2Y receptors, designated P2Y1 and P2Y12. The P2Y receptors are G-protein
coupled receptors. P2Y1 activation leads to platelet activation via Gαq, which in turn activates
phospholipase C, leading to intracellular calcium mobilization, shape change, and finally a transient platelet aggregation103. Activation of P2Y
12 leads to platelet signaling via Gαi and
inhibition of adenylyl cyclase (AC) which suppresses cyclic AMP (cAMP), enabling platelet aggregation103-105. To achieve full platelet activation and aggregation by ADP, both P2Y
1 and
P2Y12 must be activated. This is usually achieved by the endocrine feedback loop created
when e.g. ADP and ATP is released from platelet dense granule upon platelet activation by other agonists106,107. Inhibitors of P2Y receptors were proposed as therapeutic drugs capable of preventing thrombosis without major bleeding complications. The two P2Y12 inhibitors
most frequently used today are clopidogrel (Plavix®) in patients and ARC69931MX
(cangrelor) in research108,109.
Other types of purinergic receptors are the P2X calcium ion channels. P2X1 is expressed on
smooth muscle cells and on human blood platelets110-113. Until year 2000, ADP was thought to be the agonist for P2X1 but then Mahaut-Smith et al. were the first to show that ATP was the
agonist for this receptor114. Characteristic of P2X
1 is the rapid desensitization upon
activation115,116. As compared to ADP, the role of ATP in platelet activation is less well
established. The signaling events following ATP stimulation via P2X1 are hard to study due to
the fast desensitization of the receptor, and historically this is why P2X1 was discovered so
late. The breakthrough came when the use of apyrase (ATP-diphosphohydrolase; EC 3.6.1.5) was established, which prevented desensitization of P2X1. When platelets were stimulated
with the ATP-analogue α,βmethyleneATP, a transient shape change could be observed117, but
stimulation and activation of P2X1 could not induce platelet aggregation alone118. However,
activation via P2X1 is suggested to potentiate other platelet activation pathways114, and also
give transient shape change but not aggregation119. Hechler and colleagues have shown in in vitro and in vivo studies that P2X1 is necessary for thrombus formation in blood flowing over
collagen-coated surfaces at high shear rates120. P2X1 could play a role in platelet activation as
ATP is secreted from dense granule121. In paper II, we demonstrate that ATP and its P2X1
receptor is an important player in a crosstalk between PAR4 and α2A-adrenergic receptor
Adrenergic receptor
Epinephrine (also called adrenaline) is a catecholamine which is released upon sympathetic activation, under a myocardial infarction as well as during physical exercise, and which binds to g-protein coupled adrenergic receptors. As a hormone produced by the adrenal medulla, epinephrine acts by binding to α- and β-adrenergic receptors. α2A-adrenergic receptors are
found in many cells and tissues, and also in blood platelets123. Epinephrine is known to be a
weak platelet agonist, since it does not induce rises in [Ca2+]
i in isolated platelet suspensions,
but together with a secondary platelet agonist, epinephrine enhances platelet activation via its α2A-adrenergic receptors33,124. Steen et al. state that epinephrine potentiates the effect of
thrombin125. Platelets express α
2A-adrenergic receptors which are Gαi-coupled receptors. It is
known that platelets are activated by epinephrine in plasma, whereas isolated platelets in buffer are not activated or aggregated by epinephrine alone33. Binding to the Gi-coupled α2A
-adrenergic receptor will result in inhibition of adenylyl cyclase126,127.
Haemostasis & Thrombosis
Thrombus formation is a series of events which are initiated by peripheral circulating platelets adhering very quickly to a damaged site exposing extracellular matrix proteins into the circulation. One of the first adhesion steps is when the subendothelial substance collagen binds to platelet receptor glycoprotein (GP) Ib-IX-V via von Willenbrand factor (vWF) released from endothelial cells and platelets. Further, two other platelet receptors, GPVI and GPIa-IIa (α2β1), binds directly to exposed collagen and are essential for collagen-mediated adhesion and aggregation128. This primary adhesion step induces activation in platelets and release of their granule contents consisting of ADP129, ATP, serotonin, and formation of TXA2,
and when this occur, even more platelets are recruited to the injury. The fibrinogen receptor GPIIb/IIIa is activated upon agonist stimulation and undergoes a calcium-dependent conformational change to reach its active state130,131. GPIIb-IIIa on the platelet surface binds
to both fibrinogen and vWF, causing binding of platelets to the damaged area via vWF, as well as to other platelets via fibrinogen. The final step in the signaling cascade causing platelet aggregation is the inside-out activation of the fibrinogen receptor, to sustain aggregation it must remain in its active confirmation132. To stabilize the thrombus, thrombin is generated and transforms fibrinogen into unsolvable fibrin.
Researchers are developing new methods to look at thrombus formation in vivo. Furie et al. developed a new method to study thrombus formation in live mice. Vessel wall injury in mice
arterioles is induced by a laser or by application of FeCl3 and thrombus development is
visualized by fluorescence microscopy. In this type of mouse model, it is now possible to study how platelets attach to each other and how a thrombus is developing in vivo133.
There are two main different types of thrombosis; deep vein thrombosis (DVT) and arterial thrombosis. DVT may occur in veins where the blood flow rate is decreased and also due to damage in the vessel wall. The second type of thrombosis, arterial thrombosis takes place in the arteries, and in most of the cases the thrombosis is due to rupture of an atheroma, (a core of cells and lipids accumulating under the endothelium layer, causing a narrowing of the vessel lumen), therefore arterial thrombosis is also referred to as atherothrombosis. Examples of arterial thrombotic diseases are stroke, peripheral vascular disease, and myocardial infarction.
AIMS OF THE THESIS
As described in the previous chapters, several reports are indicating differences between thrombin receptors PAR1 and PAR4 in platelets. In this thesis, I have sought to investigate this matter further.
My major aim has been to study and clarify potential differences between platelet thrombin receptors PAR1 and PAR4 from a physiological and pathological perspective. In the different papers, I have investigated the role of PAR1 and PAR4 in several aspects of platelet function:
- Pathological interaction between platelets and bacteria (Paper I) - The co-existence and interplay with other platelet activators (Paper II) - Platelet granule mechanisms for synthesis and secretion (Paper III) - The platelet aggregation response (Paper IV)
METHODOLOGY
MaterialsCathepsin B inhibitor II was from Calbiochem (San Diego, California, USA). In paper I-III we used the specific PAR1 thrombin-receptor-agonist peptide (SFLLRN) and the PAR4 thrombin-receptor-agonist peptide (AYPGKF), both synthesized by the Biotechnology Centre of Oslo (Oslo University, Norway). In paper 4, PAR-activating peptides with the same sequences were synthesized by JPT (Berlin, Germany). The PAR4-blocking polyclonal chicken antibody was raised using a peptide with the sequence GGDDSTPSILPAPRGYPGQVC, which spans the thrombin cleavage site of PAR4. The peptide was synthesized by the Biotechnology Center (Oslo, Norway) and used to immunize chicken86. The cell-penetrating pepducin P4pal-i1 (palmitate-NH-ATGAPRLST), which resembles the first intracellular loop of PAR4 and has been claimed to selectively inhibit PAR4 activation by interfering with binding of the G-protein90 was synthesized by Innovagen
(Lund, Sweden). The PAR1 antagonist SCH79797 dihydrochloride (N3-Cyclopropyl-7-[4-(1m-ethylethyl) phenyl] methyl-7H-pyrrolo [3,2-f]quinazoline-1,3-diamine dihydrochloride), a potent non-peptide PAR1 antagonist, was obtained from Tocris Cookson Ltd. (Bristol, UK). Mouse monoclonal VEGF Ab-7 (VG1) was from Neomarkers (Fremont, CA, USA), and secondary goat anti-mouse IgG-HRP, Sc-2005, was obtained from Santa Cruz (CA, USA). The monoclonal mouse antibody (MAI-12) directed against human PAI-1 and tPA/PAI-1 complexes was from BioPool (Umeå, Sweden). To detect phosphorylated Akt, a mouse antibody directed against Ser437, or a rabbit antibody directed against Thr308 was used in combination with a secondary horseradish peroxidase-conjugated antibody, all purchased from Cell Signaling Technology (Danvers, MA, USA). Zenon® Alexa Fluor Labeling kit, MOPS buffer, NuPAGE gels, and Magic Marker were all obtained from Invitrogen (Eugene, Oregon, USA). Polyvinylidene difluoride (PVDF) membranes, ImmobilonTM Western (Chemiluminescent HRP Substrate), and mouse monoclonal anti-endostatin antibodies were obtained from Millipore Corporation (Billerca, MA, USA). Reference proteins, Dual color Precision Plus Protein Standards and fat free milk powder were obtained from BioRad (Hercules, CA, USA). Cangrelor (formerly AR-C69931MX; N6
-(2-methyl-tioethyl)-2-(3,3,3-trifluoro propylthio)-β, γ-dichloromethylene ATP tetrasodium salt) was kindly provided by the Medicines Company (Parsippany, MA, USA). Medium M199 supplemented with L-arg was obtained from Gibco BRL, Life Technology (Paisly, UK). α,β-Me-ATP, ADP, aspirin, apyrase, epinephrine, fura-2, ionomycin, leupeptin, luciferin/luciferase bioluminescence kit,
MRS2159, MRS2179, NF449, prazosine, Ro318220, triton X-100, thrombin from bovine plasma (T4648) and yohimbine, as well as the chemicals for the buffers; Krebs-Ringer Glucose (KRG) (20 mM NaCl, 4.9 mM KCl, 1.2 mM MgSO4, 1.7 mM KH2PO4, 8.3 mM
Na2HPO4, 1 mM CaCl2 and 10 mM glucose, pH 7.3), HEPES buffer (pH 7.4) composed of
145 mM NaCl, 5 mM KCl, 1 mM MgSO4, 10 mM HEPES, and 10 mM glucose, and acid citrate dextrose (ACD) (85 mM Trisodium citrate, 71 mM citric acid, and 111 mM glucose) were obtained from Sigma Chemicals Co (St. Louis, MO, USA). PI3-kinase inhibitor LY294002 was obtained from Tocris (MO, USA).
Isolation of Human Platelets (paper I-IV)
Venous heparinized blood was collected from blood donors at the local blood centre. The blood donors were informed about the purpose of the study and gave informed consent. The blood collection protocol was approved by the Ethics Committee at Linköping University Hospital. The blood was mixed (1/5; v/v) with an acid citrate dextrose solution (ACD, 85mM Trisodium citrate, 71mM citric acid, and 111mM glucose) and then centrifuged at 220 x g for 20 min. The resulting platelet-rich plasma (PRP) was collected and then incubated at room temperature, with or without aspirin (100 mM) and apyrase (0.5 U/ml or 0.05 U/ml). The PRP was subsequently centrifuged again at 480 x g for 20 min, and the pellet containing platelets was re-suspended in HEPES buffer or KRG buffer supplemented with apyrase (1 U/ml or 0.05 U/ml). The platelet suspensions were kept in plastic tubes and were used within 3 h. Extracellular Ca2+ concentration was adjusted to 1mM immediately before each measurement.
Measurement of Platelet Aggregation (paper I, II, and IV)
Aliquots (0.2 - 0.5 ml) of platelet suspensions (2.5 x 108 platelets/ml) were pre-incubated at 37°C for 2 min, 900 RPM. Thereafter, platelet aggregation was induced by adding antagonists and agonists according to the experimental designs (described in each paper). Changes in light transmission were recorded using a Chronolog Dual Channel lumi-aggregometer (Model 560, Chrono-Log, Haverston, PA, USA).
Culture and Preparation of Porphyromonas gingivalis (paper I)
The facultative anaerobic bacteria P. gingivalis (ATCC 33277) was cultured in fastidious anaerobe broth pH 7.2 (29.7 g/l, Lab M, Lancashire, UK) and on fastidious anaerobe agar at pH 7.2 (46.0 g/l agar, with added L-tryptophan 0.1 g/l, Lab M, Lancashire, UK) in an
atmosphere containing CO2, N2, and H2 (80:10:10; Concept 400 Anaerobic Work Station,
Ruskinn Technology Limited, Leeds, UK). Suspensions of P. gingivalis and fastidious anaerobe broth cultured for 48–72 hours were centrifuged and washed twice at 6000 x g for 30 min (4°C), the supernatant was removed and the pellet re-suspended in KRG. The washed bacteria were diluted in KRG to achieve an optical density (OD) of 1.7 at 600 nm, which corresponded to 1.5–2 x 109 Colony Forming Units per ml (CFU/ml), as determined by viable
counting.
Measurement of Cytosolic Calcium (paper I and II)
Isolated platelets were loaded with fura-2 by incubating PRP with 4 μM fura-2-acetoxymethylester (from a 4 mM stock solution dissolved in DMSO) for 45 min at 20°C, after which they were pelleted and re-suspended as described in “Isolation of human
platelets”. Before each measurement, 2 ml of platelet suspension (1-2 x 108/ml) was incubated
at 37°C for 5 min and then exposed to different drugs according to specific experimental design. Fluorescence signals from platelet suspensions were recorded using a Hitachi F-2000 fluorescence spectrofluorometer especially designed to measure [Ca2+]i. Fluorescence
emission was determined at 510 nm, with simultaneous excitation at 340 nm and 380 nm. [Ca2+]
i was calculated according to the general equation reported by Grynkiewicz et al. 134:
[Ca2+]
i = Kd(R-Rmin)/(Rmax-R) (Fo/Fs). Maximum and minimum ratios were determined by
adding 0.1% Triton X-100 and 25 mM EGTA, respectively.
Measurement of Dense Granule Secretion (paper II)
The amount of liberated ATP in platelet suspensions (0.5-ml aliquots; 2.5 x 108 platelets/ml) was measured using a luciferin/luciferase bioluminescence kit. Secretion of ATP was induced by adding thrombin or PAR-activating peptides, alone or combined with epinephrine. The ATP-dependent increase in bioluminescence was recorded in the Chronolog lumi-aggregometer.
Western Blot (paper II, III)
Platelet suspension:
Isolated platelets (1-2 x 109/ml; 100-μl aliquots) were pre-warmed at 37ºC for 3 min and were
treated according to each experimental design. All reactions (except secretion samples in paper III, further described in experimental design) were stopped by the addition of SDS (1:5
v/v) sample buffer (10 % SDS, 1.43 M β-mercaptoethanol, 20 % glycerol, 26 % urea, 125 mM Tris-HCl, pH 6.8) followed by heating at 95°C for 5 min. Thereafter, the samples were stored at -70°C until analysis.
SDS-PAGE:
Upon thawing, samples were heated once more at 95° C for 5 min, after which they were separated on ice, using 4-12 % loading gels (Invitrogen) in a MOPS-running buffer (Invitrogen), at a constant current of 140 V.
Blotting:
The proteins were transferred from the gels to polyvinylidene difluoride (PVDF) membranes (Millipore Corporation, Billerca, MA, USA) using a Mini Trans-Blot Electrophoretic Transfer Cell (Bio-Rad). To minimize unspecific binding, the membranes were blocked at 4ºC with 5% (w/v) dry milk and 0.1% (v/v) Tween 20 in PBS (10 mM phosphate-buffer and 150 mM NaCl, pH 7.4).
Protein analysis
Paper II: After blocking the membranes, phosphorylated Akt was detected using a polyclonal rabbit antibody against serine437 of Akt. The antibody was incubated with the membranes (at 4 °C) overnight. The following day, the membranes were rinsed and a secondary anti-rabbit horseradish peroxidase-conjugated antibody was added (dilution 1:1000) and incubated at room temperature for 1 hour with gentle orbital shaking.
Paper III: A mouse monoclonal antibody directed towards human PAI-1 (MAI-12, dilution 1:1000; i.e. 0.5 μg/ml) was used for detection of PAI-1 and tPA/PAI-1 complexes. The membranes were incubated with MAI-12 overnight at 4º C, after which a secondary horseradish peroxidase-conjugated antibody was used at a dilution of 1:2000 (0.2 μg/ml) and incubated at room temperature for 1 hour with gentle orbital shaking.
The membranes were rinsed in PBS supplemented with 0.1% (v/v) Tween 20 between incubations, and then analyzed using ECL Western blotting detection reagents in a LAS-1000 Imaging Analyzer (Fuji Photo Film, Tokyo, Japan).
Fluorescence Microscopy (paper III)
Platelets in PRP were incubated in an IBIDI μ-Slide 18 well (Martinsried, Germany) for 15 min to allow adhesion, followed by washing with PBS, permeabilization with Triton X-100 (0.1 %), fixation with PFA (1.85 %), further washing with PBS, and incubation with BSA-blocking buffer (0.1 %) for three hours. Fluorescence labeling of primary antibodies (towards PAI-1, VEGF, or endostatin) was done using the Zenon Alexa fluor labeling kit #1 (Alexa 488, Alexa 546, and Alexa 647) and was performed according to the manufacturer’s manual. Labeled antibodies (5 μg/ml) were incubated with platelets for 20 minutes (protected from light, at room temperature), followed by rinsing in PBS and once again fixation in PFA (1.85 %) for 10 min. Fluorescence microscopy images were captured using a Zeiss AxioObserver D1 with Axiovision® software (Zeiss, Oberkochen, Germany).
Enzyme-Linked Immunosorbent Assay (paper III)
Platelet contents (intracellular fraction) and secretion of different α-granule proteins (PAI-1, VEGF, and endostatin) was measured by Enzyme-linked Immunosorbent Assay (ELISA), using the sensitive TintELIZE® PAI-1 kit (Biopool AB, Umeå, Sweden) and endostatin and
VEGF kits from R&D Systems (Minneapolis, MN, USA) according to the manufacturer’s instructions. For ELISA measurements, isolated platelets were diluted in KRG to 2.5 x 108/ml
and pre-warmed at 37º C for 3 min. Next, the samples were incubated in round-bottom tubes (in an orbital shaker incubator, BioGrant, at 37° C and 600 RPM) with PAR-activating peptides (SFLLRN for PAR1-activation, AYPGKF for PAR4-activation) or thrombin for 10 min, 3 hours or 24 hours. The reaction was stopped by centrifugation (14000 x g, 10 min, 4° C), after which the supernatant was gently removed from the pellet phase and saved for protein secretion measurements. The pellet phase was washed with PBS and further re-suspended in M199 and lyzed in Triton X-100 (0.1%). All samples were stored at -70° C before batch analysis by ELISA. In some designs, supernatant samples were split into three different aliquots for analysis of secretion of PAI-1, endostatin, and VEGF, respectively. The pellet phase was analyzed to determine the intracellular fraction of the proteins.
mRNA Assay (paper III)
Non-irradiated apheresis platelets were collected from 5 healthy blood donors (4 males and 1 female) after obtaining their consent. The platelet suspension, with a volume of 20 ml and a platelet count ranging from 1.0-1.5 × 109 cells/ml, was centrifuged at 800 x g for 8 min and
the supernatant was discarded. The platelet pellet was processed for depletion of residual contaminating leukocytes using dynabeads conjugated with anti-CD45 according to the manufacturer’s instructions (Pan Leukocyte; Invitrogen, Carlsbad, CA). Extraction of mRNA and cDNA synthesis was performed following the method described by Rox et al.135. Platelet mRNAs were detected using 96-format TaqMan® Custom PCR Array on a 7900HT real-time
PCR apparatus (Applied biosystems, Carlsbad, CA, USA) according to the supplier’s instructions. For 20 μl qPCR reaction, 50 ng cDNA was used. Relative transcript quantification was performed using GAPDH and β-actin (ACTB) as reference genes.
Experimental Designs
Paper I: Isolated platelet suspensions were incubated with P. gingivalis and changes in light transmission or fura-2 fluorescence were registered as described above. In some experiments,
P. gingivalis was pre-treated with protease inhibitors for 45 min to establish the role played by
gingipains [5]. Leupeptin (1 mM) and cathepsin inhibitor II (1 mM) was used as Rgp-specific and Kgp-specific gingipain inhibitors, respectively. The importance of α2A- adrenergic
receptors was elucidated by pre-treating platelets for 1 minute with the antagonist yohimbin (1 mM). The significance of PAR1 and PAR4 was evaluated by pre-treating platelets for 3 minutes with the PAR1 antagonist SCH79797 (5 mM) and a PAR4-blocking polyclonal chicken antibody (3 mg/ml), respectively.
Paper II: Platelets were exposed to various concentrations of thrombin (0.21-21 nM) for 3 min, followed by addition of epinephrine (0.1-10 μM). In one experimental design, the time to addition of epinephrine was varied from 15 s. to 5 min. Platelets were also exposed to the PAR1- and PAR4-activating peptides SFLLRN (0.3- 12.5 μM) and AYPGKF (30-300 μM) followed by 10 μM epinephrine. To evaluate the significance of PAR4 and PAR1 activation, platelets were initially incubated with a PAR4-blocking antibody (10 μg/ml), or an unspecific IgY antibody (10 μg/ml), or the PAR4-specific inhibitor pepducin P4pal-i1 (10 μM), or the PAR1 antagonist SCH79797 (5 μM) for 5 min and then exposed to thrombin followed by epinephrine. The role of secondary release of ADP was evaluated by pre-incubating platelets for 3 min with the P2Y1 antagonist MRS2179 (10-20 μM) and/or the P2Y12 antagonist
cangrelor (10-100 nM). Correspondingly, the role of released ATP was studied by pre-treating platelets for 3 min with the P2X1 antagonists MRS2159 (0.01-10 μM) or NF449 (0.1-10 μM).
Paper III: Isolated platelets were incubated in round-bottom eppendorf-tubes (in an orbital shaker incubator, BioGrant, at 37° C and 600 RPM) with PAR activating peptides (SFLLRN for PAR1 activation, AYPGKF for PAR4 activation) or thrombin for 10 min, 3 hours or 24 hours. The reaction was stopped by centrifugation (14 000 x g, 10 min, 4° C), and the supernatant gently removed from the pellet phase and saved for protein secretion measurements. The pellet phase was washed with PBS, further re-suspended with M199 and lyzed in Triton X-100 (0.1%). All samples were stored at -70° C before batch analysis by ELISA or western blot. In some designs, supernatant samples were split into three different aliquots for analysis of secretion of PAI-1, endostatin, and VEGF, respectively.
Paper IV: Aliquots of platelet suspensions (2.5 x 108 platelets/ml) were pre-incubated at 37°C
for 2 min. in an aggregometer cuvette, with stirring at 900 RPM. Thereafter, platelet suspension was incubated with either LY294002 or the solvent, DMSO (final conc. 0.2%) for 10 min, or with cangrelor for 5 min. SFLLRN (3, 5, 10, 30 and 100 µM), AYPGKF (10, 30, 100, 150 and 300 µM) or combinations of peptides were added and aggregation measurements were tracked for 7 min.
RESULTS AND DISCUSSION
Paper IIn this paper we investigated the roles of PAR1 and PAR4 in the interaction between the oral pathogen P. gingivalis and the stress hormone epinephrine. P. gingivalis secrete virulence factors called gingipains (Rgp and Kgp)30, known to have thrombin-like properties24. Epinephrine is known to be a weak platelet agonist, and never triggers aggregation in isolated platelets. Here, platelets were incubated with a P. gingivalis at a platelet/bacteria ratio which did not cause aggregation. However, upon addition of epinephrine (10 µM), full aggregation was achieved, in both aspirin-treated and non-aspirin-treated platelets.
The roles of PAR1 and PAR4 were investigated using PAR1 inhibitor SCH79797 (5 µM) and a PAR4-blocking polyclonal chicken antibody. Here we found that PAR4-inhibition totally abolished P. gingivalis/epinephrine-induced aggregation, while the PAR1-antagonist significantly reduced, but could not prevent aggregation. Due to these results it seems that P.
gingivalis/epinephrine-induced aggregation is more dependent on PAR4 than PAR1. In the
presence of yohimbine, a specific inhibitor of α2A-adrenergic receptors, no platelet
aggregation was detected with P. gingivalis/epinephrine, showing that the α2A-adrenergic
receptor has a key role in this event. To further investigate the role of Rgp- and Kgp-gingipains, we used leupeptin (Rgp-inhibitor) and cathepsin II B (Kgp-inhibitor) and results show that Rgp-gingipain is responsible for P. gingivalis/epinephrine-induced platelet aggregation. Further investigations were made by measuring intracellular calcium levels. Addition of P. gingivalis caused no rise in calcium, but when followed by epinephrine, a rise in intracellular calcium was observed. Inhibition of PAR4 abolished the rise in calcium, as well as the inhibition of Rgp-gingipains, and α2A-adrenergic receptors.
In conclusion, paper I shows that there is a difference between PAR1 and PAR4 in regard to their role in the interaction between the periodontal pathogen P. gingivalis and the stress hormone epinephrine, and that P. gingivalis-derived Rgp-specific gingipains can activate PARs on the surface of platelets. We also conclude that the synergistic action of P. gingivalis and epinephrine occur both in the presence and absence of aspirin.
Paper II
The aim of paper II was to elucidate the roles of PAR1 and PAR4 in the thrombin-epinephrine cross-talk. In this paper we report a new difference between PAR1 and PAR4, as PAR4 was shown to be responsible for the synergistic interaction between thrombin and epinephrine in aspirin-treated platelets. We show that a sub-threshold concentration of thrombin followed by epinephrine caused strong platelet aggregation and calcium mobilization, even though none of these agonists caused any changes by themselves. Further, PAR4 inhibition completely abolished platelet aggregation, whereas PAR1 inhibition did not. Inhibition of the α2A
-adrenergic receptor also abolished platelet aggregation. On the other hand, inhibition of the two ADP-receptors; P2Y12 and P2Y1 had no effect on PAR4/epinephrine-induced aggregation.
We know that secondary granule release in response to the first activation event is important in platelets. Therefore, we also measured ATP release, and could show that ATP is secreted during PAR4/epinephrine-activation of platelets. ATP secretion is coupled to another purinergic receptor, P2X1, which is a rapid calcium ion channel with ATP as an agonist.
Furthermore, when P2X1 was blocked, both aggregation and calcium mobilization induced by
PAR4/epinephrine-stimulation was completely abolished. Thus ATP activation of P2X1 was a
key event for the synergistic interaction between PAR4 and epinephrine. To study the signaling downstream of the receptors, phosphorylation of Akt on Ser473 was analyzed. A low dose of PAR4-AP, which gives no dense granule secretion, induced Akt phosphorylation in platelets, but not a low dose of PAR1-AP, indicating that one of the mechanisms underlying this synergistic action may be PI3-kinase/Akt-signaling. We also inhibited PI3-kinase, which reduced the PAR4/epinephrine-activation.
In conclusion, we found that only PAR4 synergies with the α2A-adrenergic receptor and
induces strong aggregation and calcium mobilization in aspirin-treated platelets, and that secreted ATP is a key player in these events via its P2X1 receptor. We also show that the
cross-talk between PAR4, the α2A-adrenergic receptor and P2X1 circumvents the action of aspirin
and ADP receptor antagonists. Finally we can see that PI3-kinase is involved in the cross-talk between PAR4 and epinephrine.
Paper III
The aim of paper III was to study differences between PAR1 and PAR4 regarding platelet storage, release and synthesis of platelet PAI-1. We focused on the role of PAR1 and PAR4 in platelet secretion and synthesis of PAI-1, a protein playing an important part in fibrinolysis resistance. PAI-1 was studied in parallel with two well-studied α-granule proteins; VEGF and endostatin. Fluorescence microscopy results showed that none of the α-granule proteins PAI-1, VEGF or endostatin are co-localized in platelets, which may indicate the presence of three or more subtypes of α-granule. The secretion of PAI-1 and VEGF showed a similar pattern, being more sensitive to PAR1 activation than PAR4 in the lower concentration range, but secretion was also observed with higher concentrations of PAR4-activating peptides. PAI-1 was secreted in an active form. SERPINE1 (PAI-1) mRNA was found in platelets, and elevated levels of PAI-1 were detected after 24 hours incubation of both unstimulated and stimulated platelets, especially in PAR1-stimulated platelets.
Our conclusions are that PAI-1 and VEGF show a similar secretion pattern, being more sensitive to PAR1 than to PAR4 activation, but the secretion is not exclusively selective. Our results also indicate that platelets synthesize PAI-1 if incubated for 24h, both with PAR1-activation and without PAR1-activation. The lack of co-localization of PAI-1, VEGF and endostatin suggests that more than two subtypes of α-granules may be present in human platelets.
Paper IV
In the fourth and last paper in this thesis we wanted to close the circle by investigating some of the differences between PAR1 and PAR4 affecting the platelet aggregation response. During these years we have observed that PAR1-activation results in an unstable, reversible aggregation, whereas PAR4-activation results in stable aggregation. The mechanisms behind this finding were further investigated in this paper. With classical light transmission aggregometry we showed that PAR1-mediated aggregation was unstable and reversible, and even more so when inhibiting P2Y12, whereas PAR4-mediated aggregation was irreversible
both with and without P2Y12-inhibition. We also showed that mimicking thrombin with
PAR1-AP followed by increasing concentrations of PAR4-AP stabilizes the aggregate without any need of ADP (when P2Y12 is inhibited), but that this sustained aggregation can be
reversed by inhibition of PI3-kinase.
The aim of paper IV was to investigate the stability of platelet aggregates formed in response to PAR1 and PAR4-activation, and the role of secondary activation through ADP. We found that platelet aggregation induced by mild PAR1 activation is reversible, but that platelets form a stable aggregate when PAR4 is activated. The P2Y12 antagonist cangrelor inhibits
PAR1-induced platelet aggregation. This inhibition by cangrelor is concentration-dependently abolished by concomitant PAR4 activation in a PI3-kinase-dependent manner. Therefore, our conclusion is that platelet aggregation induced by combined activation of PAR1 and PAR4 is sustained due to an increasing activation of PAR4, and that this sustained aggregation can be abrogated by combined inhibition of P2Y12 and PI3-kinase in a dose-dependent manner.
GENERAL DISCUSSION
During the last decades, a number of studies regarding platelet thrombin receptors PAR1 and PAR4 have been published. This thesis contributes with additional data in this area, and here I will discuss our findings regarding differences between PAR1 and PAR4 and how these are relating to findings by other groups, which finally will lead to a suggestion for future research in this area.
Issues regarding tools and platelet handling.
Handling isolated platelets is a delicate procedure. When whole blood is drawn from vessels platelets will get in contact with foreign materials and may also be activated by vessel wall proteins during the puncture of the vessel, therefore the tube is usually left to rest for some time to reduce “pre-activation”. The use of ACD is also necessary to lower the pH. A 2-step centrifugation will follow, where the first step results in a PRP phase (platelets and plasma), and platelets may also be pre-activated by centrifugation, releasing their granule contents and increasing their P-selectin expression etc. In this step, platelets need to rest again to return to their “resting state”. In this step, platelet receptors are desensitized by a few groups by adding high concentrations of specific agonists96,102. This procedure is controversial according to my experience and knowledge, because when platelets get in contact with agonists, even if it is done in the presence of platelet inhibitors such as prostacyclin and aspirin, they will release granule contents such as ADP, ATP, calcium and other proteins from α-granules, and it is hard to see how platelets could return to a resting state after this step and further believe that agonist stimulation later on will function as well as in non-desensitized platelets. This is exactly what happened with P2X1 since it was rapidly desensitized during centrifugation
when platelets secreted their granule contents activated P2X1 with ATP, and not until apyrase
was introduced to the isolation protocol, ATP was found to be the agonist for P2X1. In the
second and last centrifugation step, platelets are packed very close to each other into a pellet at the bottom of the tube. This step is also crucial as plasma needs to be exchanged to a physiological buffer. Finally platelets need to be gently resuspended, too vigorous mixing or bubble formation can pre-activate platelets again. After isolation, extracellular calcium is added, and platelets must then be used within a few hours, or else spontaneous aggregation or non-responsiveness (sensitivity will be lost) may occur. Another problem is the transportation of patient blood samples, as rough handling of blood tubes can cause pre-activation of platelets. Issues like these, and differences between isolation protocols and procedures may
make it hard to do direct comparisons between studies, and have to be taken into consideration when discussing previous findings.
Another issue regarding research on platelet PAR receptors is the lack of commercially available PAR-inhibitors. We have used one PAR1-antagonist, SCH79797, which seems to be a partial agonist according to “own observations”, which mean that doses above 5 µM can lead to self-aggregation of platelets. Also, SCH79797 is of no use in calcium studies, due to interference with the measurements. Inhibitors of PAR4 are even scarcer. There are a few recently presented, such as YD-3, a PAR4-inhibitor which is not yet commercially available, and which has not been made available to us by the inventing research group so far. In our studies, we have used a PAR4-ab and a PAR4-specific pepducin as PAR4-inhibitors, but they both have their limitations. The PAR4-ab was made by ourselves and is a chicken polyclonal antibody raised against a peptide spanning the PAR4 cleavage site. In high concentrations it is efficiently blocking PAR4 activation by thrombin, but unspecific effects by the antibody or partial cleavage of the receptor despite treatment are hard to rule out in these situations. The pepducin we used resembles the first intra-cellular loop of PAR4 and has been claimed to be a selective inhibitor of PAR490. However, even if we found that it was efficient under the
conditions present in our paper II (isolated platelets, aspirin-treated and in aggregometry), we have also observed that it is not working under any other conditions, making it a less useful tool in general platelet research. The lack of tools to study the PAR receptors may be one factor that makes results from some studies hard to interpret. As another example, there are many commercially available PAR-specific antibodies, but most of them are only working in western blotting and some other visualizing methods, as they are not showing inhibiting effects or even binding to live cells.
As compared to the inhibitors, specific PAR-activating peptides are available, but at the time of discovery of PAR1 and PAR4, PAR4-activating peptides (AP) were not effective at all as compared to PAR1-AP. Some efforts into finding the most efficient PAR4-activating peptide were therefore made at that time, and the peptide proposed by Faruqui et al. in 200069,
AYPGKF, has since then been the choice in most studies published. It has to be taken into account that earlier studies, where less potent PAR4-activating peptides were used, may potentially be underestimating the role of PAR4 for platelet activation.
In summary, the increased knowledge regarding potential problems we are facing in platelet PAR receptor-research will hopefully enable us to produce better and more ”correct” platelet
research. Also when reading platelet papers, this knowledge will help us to use a critical eye, and to design better studies dissecting the underlying mechanisms behind the differences between PAR1 and PAR4.
Platelet secretion and PAR1
In paper III we showed that secretion of active PAI-1 is sensitive to PAR1 activation, but that secretion only occurred in response to very high concentrations of PAR4-activating peptides. Long time storage of platelets increased PAI-1 levels in PAR1-activated platelets, but not in PAR4- or thrombin-stimulated platelets. In paper II, we also showed that PAR1 activation is more highly associated with secretion, when measuring ATP-secretion from dense granule, PAR1-mediated (12.5 µM) ATP-secretion reached approx. 1.2 nmol, as compared to a very high concentration of PAR4-AP (300 µM) which only caused an ATP secretion of approx. 0.4 nmol. More data pointing in the same direction is found in the report by Nylander et al.136 who state that the P-selectin expression did increase even more when ADP was added on top of maximal PAR4-activation. On the other hand, ADP showed no additative effect on top of a maximal PAR1-mediated P-selectin expression, suggesting that PAR1 activation induces maximum secretion upon activation, whereas PAR4 do not.
Contrarily, in paper II we also studied ATP release from another perspective, where we show that ATP is secreted upon PAR4/epinephrine-activation of platelets but not upon PAR1/epinephrine-activation. The explanation could be that we here used very low (sub-threshold) doses of PAR-activating peptides (PAR1-AP 1.2-3 µM and PAR4-AP 30-60 µM, doses that does not cause aggregation), which were followed by a dose of epinephrine which itself is also unable to cause aggregation. This may suggest that a certain fraction of PAR1 receptors needs to be cleaved or activated before secretion of granule contents occur, and it also suggests that PAR1-mediated secretion could be of less importance if PAR4 is involved in cross-talk with other signaling pathways.
Platelet aggregation and PAR4
An increasing number of reports suggest that PAR4 may be an important component for stabilization of the aggregate. In this thesis, paper I, II and IV highlights PAR4 as an important ingredient for sustained platelet aggregation. In paper I, we showed that P.
gingivalis-derived Rgp-specific gingipains activates PARs on the surface of the platelets, with
a preference to PAR4, and that this followed by epinephrine results in a full and strong aggregation137. In paper II, we found that crosstalk between PAR4 (but not PAR1) and the
