Send Orders for Reprints to [email protected] Current Pharmaceutical Design, 2014, 20, 5779-5793 5779 Mechanisms of Action of Anesthetics for the Modulation of Perioperative Thrombo- sis: Evidence for Immune Mechanisms from Basic and Clinical Studies Toshiharu Azma1,2*, Florin Tuluc3, Taishin Ito2, Chikako Aoyama-Mani1, Shinji Kawahito4 and Hiroyuki Kinoshita5 1Department of Anesthesiology & Pain Medicine, Kohnodai Hospital, National Center for Global Health and Medicine, Kohnodai 1- 7-1, Ichikawa, Chiba 272-8516, Japan; 2Department of Anesthesiology, Saitama Medical University Hospital, Moroyama-cho, Iruma-gun, Saitama 350-0495, Japan; 3The Division of Allergy/Immunology, The Children's Hospital of Philadelphia Research Insti- tute, Philadelphia, United State of America; 4Department of Anesthesiology, Tokushima University Hospital, 3-18-15 Kuramoto, To- kushima 770-8503, Japan; 5Department of Anesthesiology, Aichi Medical University, 1-1 Yazako Karimata, Nagakute, Aichi 480- 1195, Japan Abstract: Thrombotic events occurring in either arteries or veins are the primary causes of fatal perioperative cardiovascular events. Risk factors for deep vein thrombosis, several of which are evidently associated with specific surgical procedures, are quite different from those for arterial thrombosis (e.g., aging or atherosclerotic diseases). Thrombus formed in arteries consists mainly of platelets coated with fibrin (i.e., white thrombus), while venous thrombus formed at relatively lower shear stress consists of all blood components including erythrocytes as well as leukocytes infiltrated with fibrin (red thrombus). Clinical evidence indicates beneficial roles of neuraxial anesthe- sia/analgesia in the prevention of VTE for patients undergoing high risk surgical procedures. To date, mechanisms of action of drugs used for neuraxial anesthesia/analgesia to prevent venous thrombosis are uncertain. However, accumulation of clinical as well as experimental findings points to the involvement of immune cells (especially monocytes) in red thrombus generation and to the interaction of anesthet- ics with these cells. We also suggest that adhesion molecules associated with the formation of monocyte platelet aggregates as well as substance P: neurokinin-1 receptor (SP/NK1R) pathway that involves neurogenic inflammation are crucial. Local anesthetics and NK1R antagonists are candidate drugs that may possess the capability to prevent venous thrombotic disorders in perioperative settings. Keywords: Perioperative thrombosis, neuraxial anesthesia/analgesia, substance P, neurokinin-1 receptor, microparticle, tissue factor, mono- cyte, local anesthetics. 1. INTRODUCTION: AN OVERVIEW OF THE CLINICAL Other arterial thrombotic disorders that involve fetal periopera- IMPACT OF PREVENTION OF PERIOPERATIVE tive events include cerebral infarction. The latter ischemic cere- THROMBOSIS brovascular event is frequently surveyed as a part of stroke which also includes cerebral hemorrhage. Even so, symptomatic pe- Thrombosis that occurs both in arteries and in veins accounts rioperative stroke in patients undergoing noncardiovascular surgery for major causes of fatal perioperative events. In prospective cohort studies conducted in Western countries [1-5], the incidence of pe- appears to occur less frequently (0.3 - 0.4%) [10, 11] than pe- rioperative MI shown above . It is noteworthy that a prospective rioperative myocardial infarction (MI) in patients with coronary survey conducted by Kimura et al. [11], in which 36634 Japanese artery disease was 1.8% - 5.6% [1-5]. Perioperative cardiac death occurred in 0.7% - 1.3% of these patients [1-5]. A more recent pro- patients undergoing non-cardiac or non-cerebrovascular surgery in two tertiary general hospitals located in a city were enrolled, indi- spective cohort study, in which 8351 patients undergoing non- cated that the incidence of perioperative MI (0.33%) and stroke cardiac surgery with risks of atherosclerotic disease were enrolled from multiple centers in 23 Western and Asian countries during a (0.34%) was similar in these populations. This report, taking to- gether with others showing the incidence of perioperative MI in period of 2002 - 2007, indicated that MI occurred in 5% of patients Western countries, points to the geographical and/or racial differ- within 30 days after surgery. The mortality rate during the 30-day perioperative period, was 11.6% among patients with perioperative ences in the occurrence of perioperative arterial thrombotic disor- ders. MI, while 2.2% among those without MI [6]. A randomized con- trolled trial (RCT) scheduled to evaluate the beneficial effects of In contrast to the generation of thrombus in arteries, thrombosis the interventional statin treatment on perioperative cardiac events in in veins per se is not a life-threatening event. However, occlusion of patients undergoing elective non-cardiac vascular surgery showed pulmonary arteries by venous thrombi torn away generally from that the incidence of myocardial ischemia and 30-day mortality deep veins in legs (i.e., pulmonary embolism, PE) provokes dysp- without the use of statin was 19.0% and 10.1%, respectively [7]. By nea and fatal cardiovascular events. The occurrence of perioperative contrast, the incidence of MI after non-cardiac surgery in the gen- deep vein thrombosis (DVT) is influenced by the type of surgical eral population was reported to be up to 0.7% in a review published procedures. A meta-analysis comparing cardiovascular events that in 1990 [8]. A prospective cohort study, in which relatively unse- occurred in patients undergoing abdominal surgery with or without lected 4315 patients undergoing elective non-cardiac surgery were treatment by low molecular weight heparin indicated that the inci- enrolled between 1989 - 1994, indicated that the incidence of pe- dence of DVT and PE in patients without the treatment was 14.5% rioperative MI or cardiac death was 1.1% or 0.3%, respectively [9]. and 0.5%, respectively [12]. Another meta-analysis that compared the incidence related to venous thrombus generation in patients undergoing orthopedic surgery with or without treatment by vita- *Address correspondence to this author at the Department of Anesthesiol- min-K antagonists indicated the incidence of DVT and PE without ogy & Pain Medicine, Kohnodai Hospital, National Center for Global the treatment being 52.5% and 8.6%, respectively [13]. The geo- Health and Medicine, Kohnodai 1-7-1, Ichikawa, Chiba 272-8516, Japan; graphical difference in the occurrence of perioperative DVT has Tel: 81-47-372-3501; Fax: 81-47-375-4709; E-mail: [email protected] 1381-6128/14 $58.00+.00 © 2014 Bentham Science Publishers 5780 Current Pharmaceutical Design, 2014, Vol. 20, No. 36 Azma et al. also been pointed [14]. The incidence of perioperative DVT in pa- arterial or venous thrombotic disorders. In this section, clinical tients undergoing orthopedic hip surgery in Thailand [15], Korea evidence is presented to support different mechanisms of action of [16], and Hong Kong [17] was 4%, 10%, and 53%, respectively. anesthetics that explain their effect of preventing the generation of arterial as well as venous thrombi in perioperative settings. Considering the incidence of perioperative DVT that appears to be higher than that of perioperative arterial thrombotic disorders, Neuraxial anesthesia, including epidural and spinal anesthesia, DVT is a potential risk for perioperative death even for patients provides potent intraoperative analgesic effects in comparison with without risk factors of arterial thrombotic disorders. The mortality general anesthesia [32]. The postoperative neuraxial analgesia also rate of VTE is unclear from the prospective studies because of its provides better pain relief as compared to the systemic administra- low incidence. A retrospective survey, conducted by the Japanese tion of opioids [32]. Besides the conspicuous roles of neuraxial Society of Anesthesiologists in 2004 that involved 1,131,154 cases anesthesia/analgesia in pain control, increasing evidence has been from 642 hospitals where members of the Society served full time accumulated from clinical practice in the last several decades dem- [18], indicated that perioperative VTE occurred in 409 patients and onstrating that the perioperative mortality of patients receiving 89 (21.8%) of these patients died. Even though the real number of neuraxial anesthesia/analgesia was lower than those receiving gen- patients with perioperative PE is not clear from this retrospective eral anesthesia. Rogers et al. [33] performed a meta-analysis to test survey, it indicated the high mortality rate of perioperative PE once this idea by reviewing randomized controlled trials (RCTs) and it occurred. quasi-experimental prospective cohort studies that compared anes- thesia methods in a variety of surgical procedures [33]. To pharmacologically prevent cardiovascular events related to thrombosis, it is important to consider which type of thrombus Although outcome measures of these studies were not standard- should be targeted by the therapy, i.e., red or white thrombus in ized, data obtained as nominal scale, such as the mortality and the macroscopic histology [19-21] White thrombus consists mainly of incidence of MI, stroke, DVT or PE, can be accurately pooled from platelets infiltrated by fibrin [21]. Arterial thrombus that occurs at a these studies. Furthermore, investigators improved the validity of site with rapid blood stream is typically a white thrombus [19, 20]. data by interviewing authors of the original articles [33]. Pooled Because of such pathophysiological basis, anti-aggregators of plate- data from 141 trials with 9,559 participants indicated that neuraxial lets are effective for the primary prevention of acute coronary syn- anesthesia/analgesia decreased the 30-day mortality as compared to drome (ACS) [22,23]. general anesthesia (2.1% vs. 3.1%). Although subanalysis of pooled On the other hand, in conditions that blood flow is highly de- data failed to indicate the statistical significance that a certain sur- creased, an impregnable red thrombus is generated where every gical procedure favored neuraxial anesthesia/analgesia, these anes- thetic methods tended to reduce the postoperative mortality most blood component including leukocytes and erythrocytes as well as fibrin is infiltrated [21]. Blood coagulation is more important than remarkably in orthopedic procedures [33]. platelet adhesion and aggregation to generate red thrombi [24], as It is also suggested that neuraxial anesthesia/analgesia appeared supported by the clinical evidence of anti-coagulants to be benefi- to decrease the incidence of thrombotic disorders. The occurrence cial for the prevention of arterial fibrillation [25, 26] or DVT [27, of DVT as well as fatal cardiovascular events related to PE was 28]. In a prospective study that examined the histological character- significantly reduced by neuraxial anesthesia/analgesia [33]. It is istics of thrombi in ACS indicated that red thrombus involves MI in also notable that more than 80% of these events were recorded in relatively large area as compared to white thrombus does [21]. orthopedic trials. By contrast, neuraxial anesthesia/analgesia tended These findings also suggested that red thrombus plays a crucial role to but did not significantly reduce the cardiovascular events related even in the acute deterioration of arterial thrombotic disorders [21]. to the generation of thrombus in arteries such as MI or stroke [33]. It is currently suggested for the perioperative prevention of Several other investigators performed meta-analyses to evaluate DVT that anti-coagulants are preferable class of drugs [27, 28], beneficial roles of neuraxial anesthesia/analgesia in patients under- although bleeding tendency caused by the use of these drugs in- going orthopedic procedures. Urwin et al. performed a systematic creases risks related to surgery [27, 28]. For example, current review concerning the clinical outcomes of patients undergoing hip guidelines for the prevention of DVT [27,28] and the left atrial fracture surgery under general versus neuraxial anesthesia [34]. thrombus associated with arterial fibrillation [25,26] recommend From four prospective controlled trials in which 259 patients were the use of warfarin at a target range of prothrombin time interna- enrolled, it appeared that the incidence of DVT was significantly tional normalized ratio (PT-INR) to be 2 - 3. However, guidelines reduced by neuraxial anesthesia. However, the incidence of MI for perioperative blood component therapy recommend the use of (888 participants from 4 trials) as well as cerebrovascular events fresh frozen plasma in a condition that an abnormal bleeding occurs (1,085 participants from 7 trials) was not influenced by the anes- while the value of PT-INR is greater than 2 [29, 30]. From the per- thetic techniques [34], supporting the findings by Rogers et al. [33]. spective for the patient safety during perioperative periods, discov- Mauermann et al. systematically reviewed prospective trials in ery of a new class of drugs for anti-thrombosis that do not directly which patients undergoing elective total hip replacement were en- affect coagulation factors and do not increase the bleeding risks rolled [35]. Such patient selection seemed to be advantageous to would be highly beneficial. reduce statistical bias from preoperative bleeding or pathophysi- Growing body of clinical as well as experimental evidence in- ological changes in patients with hip fracture. From 5 trials in dicates that circulating immune cells, especially monocytes, are which 409 patients were enrolled, the incidence of radiographically- proven DVT was again significantly reduced by neuraxial anesthe- involved in the coagulation, thrombosis, thrombolysis, or recanali- zation of occluded vessels [31]. However, the relationship between sia [35]. A systematic review performed later by Hu et al. selected perioperative thrombosis and immune cells especially in anesthe- only randomized controlled trials which compared the clinical out- comes of patients undergoing elective joint replacement for hip and tized patients has not yet been well studied. In this review article, we will focus on mechanisms of action of immune cells, especially knee under general anesthesia with those under neuraxial anesthesia circulating monocytes, in perioperative thrombosis. [36]. Pooled data from 10 randomized clinical trials with 910 par- ticipants indicated that neuraxial anesthesia significantly decreased 2. CLINICAL EVIDENCE FOR BENEFICIAL ROLES OF the incidence of DVT as compared to general anesthesia [36]. NEURAXIAL ANESTHESIA/ANALGESIA IN VENOUS BUT However, sub-analysis of the pooled data, selecting trials in which NOT ARTERIAL THROMBOSIS patients were applied with anti-coagulants for the prophylaxis against thrombosis, indicated that the occurrence of DVT was not A number of systematic reviews point to relationships between reduced by neuraxial anesthesia combined with use of anti- certain anesthetic methods and cardiovascular events caused by coagulants [36]. Mechanisms of Action of Anesthetics for the Modulation of Perioperative Current Pharmaceutical Design, 2014, Vol. 20, No. 36 5781 On the other hand, relatively large scale multicenter RCTs were crease the incidence related to venous thrombotic disorders includ- conducted to test the hypothesis that neuraxial anesthesia improves ing those of DVT and PE, but not the incidence related to arterial the perioperative outcome in patients undergoing major abdominal thrombotic disorders such as MI and stroke. surgery. Park et al. conducted a RCT in which perioperative pa- The concept, that neuraxial anesthesia/analgesia are recom- tients undergoing major abdominal surgery of multiple surgical mended for the perioperative management of patients who are at procedures were enrolled (1021 patients) [37]. The effects of com- risks of venous thrombosis, should be obsolete because of the pro- bination of epidural anesthesia with general anesthesia followed by phylactic use of anti-coagulants against VTE in the current clinical postoperative epidural analgesia on several outcome measures were practice [27, 39] However, even though clinical guidelines for pre- examined. Patients enrolled in epidural group were intraoperatively vention of VTE recommend the perioperative application of antico- applied with epidural bupivacaine at a concentration of 0.5%. Post- agulants [27, 39], perioperative use of such drugs is not always operatively, intermittent 3 - 6 mg epidural morphine at a concentra- preferable because of adverse outcomes occurred by anticoagula- tion of 0.5 mg/ml was administered to these patients. Patients in tion in some clinical settings, e.g., postoperative paraplegia caused control group were treated with intravenous or intramuscular by epidural hematoma in patients applied with neuraxial anesthesia opioids applied as requested. The incidence of DVT was not meas- or in those undergoing spinal surgery [27, 39]. Therefore, discovery ured, while severe respiratory and/or cardiovascular events diag- of a new class of drugs, that inhibit the growth of venous thrombi nosed as PE was recorded in this study [37]. Even though the occur- without directly affecting blood coagulation, is desired for the ap- rence of PE was recorded for two patients enrolled in a group with- plication to a certain sub-population scheduled for surgery. Thus, out epidural anesthesia and for no patients in a group with epidural, clinical evidence for the beneficial roles of neuraxial anesthe- significant difference in the incidence of PE was not detected [37]. sia/analgesia in perioperative venous thrombosis provides clues to The incidence of cardiovascular events, including those related to clarify mechanisms of action of anesthetics on thrombosis. The arterial thrombosis, was not influenced by the combination of following items, extracted and interpreted from the clinical evi- epidural anesthesia with general anesthesia [37]. dence, should be paid attention for the drug discovery. Rigg et al. [38] also designed a multicenter RCT in which high- • Neuraxial anesthesia/analgesia appeared to be beneficial for risk perioperative patients undergoing major abdominal surgery the prevention of thrombosis in veins (i.e., red thrombus gen- with multiple procedures (including oesophagectomy) were en- eration). However, these appeared not to be significant for the rolled (915 patients). Patients were enrolled only when they had at prevention of arterial thrombotic disorders. least one of categorized perioperative risks. They recorded a num- • Targets of drugs applied for neuraxial anesthesia/analgesia ber of fatal cardiovascular and respiratory events, while they did not (i.e., virtually equal to epidural anesthesia/analgesia) are dorsal record the occurrence of PE. In contrast to the RCT by Park et al. root ganglions (DRGs) of spinal nerves that transmit pain sen- [37], patient (or physician) controlled intravenous opioid was used sation to the central nervous system (CNS) [32,40,41]. as the control method for postoperative analgesia [38]. The regimen of postoperative epidural analgesia was not fixed except that • Classes of chemical transmitters between DRGs and CNS bupivacaine or ropivacaine was used. Similar to the results obtained already identified are excitatory amino acids and neurokinins in the RCT conducted by Park et al. [37], cardiovascular events [32,42,43]. Especially, it is known that receptors for the latter were not reduced by the combination of epidural anesthesia [38]. neurokinins also express in systemically circulating monocyte [44-46]. Taken these clinical findings gained from meta-analyses of RCTs [33-36] and large scale RCTs [37,38] together (Table 1), it is • Drugs used for the neuraxial anesthesia/analgesia are local strongly suggested that neuraxial anesthesia/analgesia, which anesthetics and opioids [32, 37, 38]. potently reduce intraoperative as well as postoperative pain, de- Table 1. Systematic reviews and multicenter randomized controlled trials that compared the incidence of death, deep vein throm- bosis, or pulmonary embolism in perioperative patients managed with or without neuraxial anesthesia/analgesia. Source Methods Type of surgery Comparison Mortality DVT PE Rogers Meta-analysis Various procedures Intraoperative NA vs. GA Decreased Decreased Decreased 2000 Urwin Meta-analysis Hip fracture surgery Intraoperative NA vs. GA NS Decreased NS 2000 Mauermann Meta-analysis Elective total hip replacement Intraoperative NA vs. GA ND Decreased NS 2006 Hu Meta-analysis Elective total hip replacement Intraoperative NA vs. GA ND Decreased Decreased 2009 or total knee replacement Park Multicenter major abdominal surgery of Intra- and post-operative NA vs. NS ND NS 2001 RCT multiple surgical procedures GA with conventional post- operative systemic opioid Rigg Multicenter major abdominal surgery of Intra- and post-operative NA vs. NS ND ND RCT multiple surgical procedures GA with systemic opioid PCA 2002 DVT, deep vein thrombosis ;GA, general anesthesia; NA, neuraxial anesthesia/analgesia; ND, not determined; NS, not significant; PCA, patient controlled analgesia; PE, pulmonary embolism ;RCT, randomized controlled trial. 5782 Current Pharmaceutical Design, 2014, Vol. 20, No. 36 Azma et al. • Local anesthetics block the conductance of electrical stimuli in B). Viscoelastometry (Sonoclot®, Sienco, Arvada, CO) peripheral nerves (spinal nerves) by inhibiting the functions of PT and APTT examine the coagulation performance, i.e., en- sodium channels [32]. However, it is generally known that lo- zymatic activity of prothrombinase complex to generate the end- cal anesthetics inhibit signal transduction pathways through G product of coagulation cascade fibrin, in the presence of excess protein-coupled receptors at concentrations used for clinical amounts of negatively charged phospholipids [64]. The latter nega- practice [47-49]. tively charged phospholipids is a replacement of PS on the surface • Opioids potently block the noxious stimuli through binding to of platelets [51, 62] (Fig. 1), where coagulation progresses to form their receptors widely spread in CNS and peripheral nerves the clot involving whole blood components, including systemically [32]. Receptors for opioids as well as those for neurokinins circulating cells (i.e., red thrombus). These also indicate that plate- also express in monocytes [45]. lets act as the scaffold of coagulation and influence the quantity of fibrin cross-linkage between these cells. Therefore, amount of 3. DIAGNOSTIC METHODS FOR THE ASSESSMENT OF platelets [54, 65] as well as their functions to express PS [51, 62] THROMBUS GENERATION influences the whole blood coagulation that occurs beyond the pri- Occlusion of pulmonary arteries by blood clots torn away from mary hemostasis in red thrombus. It is further suggested that diag- thrombi formed typically in deep veins of legs is the common sce- nostic devices to measure the strength of clot ex-vivo are useful for nario for the pathogenesis of PE [50]. The floating end of peduncu- the assessment of red thrombus formation. The oscillating-probe lated venous thrombus grows intraluminally without contact to viscoelastometry is a method derived from orthodox thromboelas- vascular endothelium. Thus, the reinforcement of clot by blood tography addressed in the next section. Advantages of oscillating- components (coagulation factors, platelets, erythrocytes, as well as probe viscoelastometry, that measures both viscosity and elasticity leukocytes) [51] has been suggested to be an essential mechanism of samples, include more prompt detection of the commencement for the growth of venous thrombi and that anti-thrombotic property of fibrin polymerization followed by the clot formation as com- of endothelial cells [52,53] do not prevent the generation of the pared to thromboelastography [56]. floating end of such thrombi. These suggested that the ex-vivo di- Blood-derived sample filled in a plastic disposable cuvette agnostic devices to assess the strength of clot as well as the velocity placed on a stage of the device with a thermostat at 37°C is vigor- of clot formation ((cid:2) coagulation) such as those presented in this ously mixed with a coagulation activator by a metal rod in the cu- section (i.e., thromboelastography or viscoelastometry) [54-56] are vette [54-56, 66, 67]. An open-ended plastic tubular probe mounted useful to understand the roles of blood components and the mecha- on a mechanical-electrical transducer oscillates vertically a distance nisms of pharmacological action of various drugs on the thrombus of 1 um at 200 Hz [55]. This probe is immersed to a fixed position generation in veins. In this section, we summarize a series of ex- in the cuvette and the viscosity of sample is continuously monitored perimental and diagnostic methods which are being used to evaluate by a manufacturer-provided software running on a personal com- clot formation in the absence or the presence of blood components. puter [66,67]. The commencement of recording is synchronized to the switch of the magnetic stirrer to mix materials in the cuvette A). Prothrombin time (PT) and the Activated Partial Throm- [66,67]. The mechanical impedance of the sample exerting a vis- boplastin Time (APTT) cous drag on the probe toward the wall of the cuvette was defined The performance of blood coagulation is mainly indexed as PT as the clot signal (arbitrary unit) [55, 56]. At the commencement of [57-59] and APTT [60, 61]. These indices are measured as clotting recording synchronized to the initial mixing step, the clot signal time, a period from the addition of coagulation activators until clot from the probe oscillating in air is zero. Immersion of this probe to (thrombus) forms. Blood samples are collected in tubes containing the sample in cuvette, performed manually by the operator, causes sodium citrate to chelate free calcium ion (Ca2+). Free Ca2+ is essen- small although significant increase in the clot signal because the tial for the bonding of several coagulation factors to phospholipids viscosity of the sample at an aqueous phase impedes the probe to [62-64]. After collection on citrate, samples are centrifuged to re- oscillate [55, 66, 67] The time until a fast increase in the clot signal move cellular components from plasma [60, 61] Thus, these per- indicates the critical point from the aqueous phase to viscoelastic formance indicators reflect the activity of coagulation factors inde- phase. This is the activated clotting time (ACT) [55, 56] that is pendent of functions of cellular components. APTT evaluates the conceptually equal to APTT, except that phospholipids are replaced efficacy of the "intrinsic" (i.e., the contact activation) pathway as with other components in the blood-derived sample (e.g., platelets, well as the common coagulation pathway [61]. Activation of the leukocytes or erythrocytes). The increasing rate of the clot signal intrinsic coagulation is commenced by reloading Ca2+ in the pres- reflects the rate of fibrin polymerization occurred mainly on the ence of phospholipids and an activator (e.g., silica, celite, kaolin, surface of activated platelets [54, 55]. Therefore, the number of etc.). The latter materials possess negatively charged surface and platelets as well as their functions influence ACT, indicating that are added to fully activate the intrinsic pathway [56, 60]. the latter reflects procoagulant activity of cellular components in Similarly, a certain kind of phospholipids (e.g., phosphatidyl- the sample [54]. Representative records of the oscillating-probe serine, PS), also possessing negatively charged surface [51, 62, 63] viscoelastometry are shown in (Fig. 2). is added to provide a scaffold of coagulation. The catalytic efficacy C). Thromboelastography of factor Xa, that convert prothrombin (factor II) to thrombin (IIa) in the presence of Ca2+, is boosted more than 1000-fold by the addi- Thromboelastography is the original method for ex vivo diag- tion of phospholipids with negatively charged surface, and 106-fold nosis to assess the strength of clot and the time to commence co- by further addition of factor Va [64]. On the other hand, PT evalu- agulation. Classically, a stainless steel cup, rotated at an angle of ates the efficacy of the "extrinsic" pathway as well as the common 4°45’ with a cycle of 10 s (0.1 Hz), is filled with whole blood, coagulation pathway [57-59]. As compared to APTT, tissue factor while a sensor shaft is placed in a fixed position of the cup. Liquid (TF, factor III) is further added in the absence of negatively charged paraffin is layered onto the blood to prevent the sample from drying activators [58, 59]. Thromboplastin, historically prepared from up. The torque, generated by the torsion of a sensor shaft dragged to placental source, is a combination of phospholipids and TF [57]. the wall of rotating cup by the elasticity of clot, is converted to the Thus the measurement of clotting time in the absence of TF is amplitude of wire that in turn traced serially on a thermal recording termed the activated “partial thromboplastin” time. paper [56, 65, 68]. Mechanisms of Action of Anesthetics for the Modulation of Perioperative Current Pharmaceutical Design, 2014, Vol. 20, No. 36 5783 Thrombin Extrinsic pathway (initiation pathway) Intrinsic pathway F2 (amplification pathway) Prothrombinase complex Outer surface F8 F9 F7 F10 F5 TF Inner surface Fig. (1). Schematic presentation of intrinsic and extrinsic tenase and prothrombinase complex bound to the outer surface of cell membrane or microparticles. Tissue factor (TF) bound to the cell membrane of monocytes as well as microparticles shed from the cells is a specific ligand for coagulation factor VII (F7). The combination of these molecules initiates the blood coagulation. The activated form of F7 directly activates factor X (F10) that is a core enzyme of prothrombinase complex attached to the negatively charged phospholipids phosphatidylserine (PS, closed circles on the outer surface of the cell membrane). Prothrombinase complex consists of activated forms of F10 and its co-enzyme factor V (F5), and the substrate of F10 prothrombin (F2) concentrated on PS in the presence of Ca2+. The activated F7 also cleaves factor IX (F9) to yield the activated form of F9. F9 is the intrinsic tenase, the enzymatic activity of which is enhanced by the combination with its co-enzyme factor VIII (F8) bound on PS in the presence of Ca2+. Thus, this intrinsic pathway acts as the amplifier of extrinsic tenase activity of TF: F7 complex (initiation pathway). The thromboelastograph had been manufactured by Hellige Clot Signal (Freiburg, Germany) and distributed by Haemoscope (Albertson, 100 NY) [68]. Since 1996, “thrombelastograph” and “TEG” became the registered trademarks of the latter company, and were succeeded by Haemonetics (Braintree, MA). In the historical device, mainly three (B) variables are recorded to represent the clot formation [56,65]. The 50 value of “R” denotes the reaction time that is defined as the time from sample was added in the cuvette until the amplitude of the (A) thromboelastography tracing reached 2 mm. The value of “K” de- 0 noting coagulation time is defined as the interval when the trace amplitude reached 20 mm after R. The torque from the rotation of 0 5 10 cup is generated only after the cup and the shaft was bonded by the Min fibrin polymerization in clot formed. Therefore, the amplitude of thromboelastography (= elasticity) is zero before “R”. This is the most different point from the viscoelastometry [56, 66, 67]. Maxi- Fig. (2). Representative traces of oscillating-probe viscoelastmetry obtained mal amplitude (MA) was defined as the greatest amplitude on from human plasma in the absence (A) or the presence (B) of platelets sus- thromboelastography tracing [56, 65]. The time-dependent decrease pended. in MA was also evaluated as the indices of clot retraction or The clot signal (arbitrary unit) indicative of viscoelasticity of plasma was fibrinolysis. The feasibility for the evaluation of the latter fibrino- plotted against time (min). The record was commenced just after reloading of Ca2+ to citrated plasma. At the commencement of recording, the clot lysis had been the special characteristics of thromboelastography. Currently, serial changes in torque generated by torsion is recorded signal from the probe oscillating in air is zero. Immersion of this probe to by a personal computer connected to the thromboelastograph [56]. plasma at an aqueous phase causes small although significant increase in the A representative trace of thromboelastography recorded using the clot signal by the viscosity that impedes the probe to oscillate. The time historical device manufactured by Hellige (Freiburg, Germany) is until a fast increase in the clot signal indicates the critical point from the shown in (Fig. 3). aqueous phase to viscoelastic phase. This is the activated clotting time (ACT, arrows). The increasing rate of the clot signal at the viscoelastic 4. MOLECULAR MECHANISMS FOR THE MODULATION phase reflects the extent of fibrin polymerization that drags oscillating-probe OF PROTHROMBOTIC ACTIVITIES IN PLATELETS to the wall of cuvette during the clot formation. Note that the ACT in the trace B was shorter than that in the trace A, demonstrating the platelet pro- As already mentioned in the previous section, venous thrombus coagulant activity. The clot signal at the viscoelastic phase in trace B was has been traditionally regarded as the composite of red blood cells higher than that in trace A, indicating the crucial roles of platelets as the infiltrated and coated by fibrin glue, with an appearance of red clot scaffold of blood coagulation. [69, 70]. However, early accumulation of platelets at the head of venous thrombus was demonstrated by experimental study using 5784 Current Pharmaceutical Design, 2014, Vol. 20, No. 36 Azma et al. such discrepancy of the results between thromboelastography and oscillating-probe viscoelastometry is due to the difference of these devices to detect initial fibrin polymerization at aqueous phase as the change in viscosity prior to the solid clot formation possessing elasticity (see the Section 3). It is also described in the section 3 that negatively charged MA phospholipid PS expressing in the cell-surface of platelets acts as a R co-factor of blood coagulation that accelerates 106-times the enzy- R+K matic activity of factor Xa to convert a substrate prothrombin to thrombin in the presence of factor V and Ca2+ [51, 64], supporting a concept that aggregated platelets play a role as the scaffold of blood 0 coagulation to build up a red thrombus. However, PS is known to be localized unevenly at the inner surface of cell membrane of inac- tive platelets, indicating that this negatively charged phospholipid is Fig. (3). A representative trace of thromboelastograph obtained from human not always expressed on the cell-surface of platelets when they are whole blood using the historical device manufactured by Hellige (Freiburg, systemically circulating without stimuli [51, 62]. Germany). To maintain such uneven localization of phospholipids, their Reaction time (R) is defined as the time from sample was added in the cu- specific active transporting system in the platelet cell membrane is vette (0) until the amplitude of the thromboelastography tracing reached 2 essential. Two cell-membrane bound enzymes, aminophospholipid mm. Coagulation time (K) is defined as the interval when the trace ampli- translocase and floppase, have been identified as ATP-dependent tude reached 20 mm after R. Maximal amplitude (MA) is defined as the transporters for phospholipids [51, 62]. Aminophospholipid trans- greatest amplitude on the tracing of thromboelastograph. locase actively transports the phospholipids with amine headgroup such as PS from the cell surface to the inner side of the cell mem- radiolabeled platelets [71], suggesting that platelets aggregated and brane [74-76]. This enzymatic reaction of aminophospholipid trans- stimulated following their adhesion to the injured vessel walls play locase is inactivated under a condition that the cytosolic Ca2+ ele- a role as the scaffold of blood coagulation at early phases of red vated up to levels of (cid:2)M [77], indicating that the active transport of thrombus formation. Venous thrombi, generated at the site of low sweeping PS away from the outer membrane is terminated by the shear stress with a floating end of pedunculated shapes, grow in- activation of platelets [51, 77]. traluminally without contact to vascular endothelium. Thus, the Another ATP-dependent active transporter of phospholipids, reinforcement of clot by blood components (coagulation factors, floppase, mobilizes phospholipids from the inner surface of the cell platelets, erythrocytes, as well as leukocytes) is suggested to be an membrane to the outer surface of this lipid bilayer [78]. The enzy- essential mechanism for the growth of venous thrombi [51, 62, 67]. matic activity of floppase is not specific to the class of phospholip- This concept is partly supported by a clinical observation that the ids [51, 78], in contrast to the headgroup-specific active transport occurrence of DVT in patients with hip fracture who are at the by aminophospholipid translocase [51]. Every phospholipid located highest risks for the perioperative DVT/PE was associated with in the inner side is thus conveyed to the cell surface by this trans- hypercoagulability observed by thromboelastography [72]. While porter. However, the rate of active transport by aminophospholipid venous thrombi are mainly composed of erythrocytes glued by fi- translocase is 10-times faster than that by floppase [51]. These two brin, platelets as the scaffold of fibrin polymerization are suggested apparatus for the phospholipid transport maintains the level of PS in to be involved in the growth of thrombi in a similar way that they the cell surface of platelets to be low in a basal (non-stimulated) are crucial for the thrombogenesis at an early phase [69, 70]. Even condition [51, 62]. though it is confirmed in a limited perioperative condition, clinical By contrast, in a condition that cytosolic Ca2+ level in platelets evidence also indicates the significant effects of anti-platelet drugs is elevated by several stimuli, the enzymatic activity of scramblase for the prevention of DVT [28, 73], partly supporting the pivotal actions of platelets in venous thrombus generation. To clarify pre- is increased [51, 62, 76, 79]. Activated scramblase countertrans- ports phospholipids between the inner and the outer leaflets of the cise mechanisms involved in the reinforcement of clot by platelets cell membrane. The shuffling of phospholipids by scramblase to is thus suggested to be important to establish a new strategy for prevention of VTE with low risk-to-benefit ratio in perioperative neutralize the uneven localization of phospholipids causes the cell- surface expression of PS that is performed within several minutes settings. The experimental findings obtained from thromboelasto- [51]. Such PS-related procoagulant activity of platelets can be ob- graphy and/or oscillating-probe viscoelastometry should give sev- eral information for quantitative impact of platelets on the strength served by viscoelastometry [67] as well as thromboelastography [80] using a specific inhibitor of PS, Annexin V. The ex-vivo clot of clot. formation observed by oscillating-probe viscoelastometry is com- By using thromboelastography, Oshita et al. demonstrated that menced within several minutes following the reloading of Ca2+ that the strength of clot (MA) is increased as a function of logarithm of provokes intrinsic coagulation. This commencement of clot forma- the number of platelets suspended in plasma [65]. However, it is tion is accelerated by the presence of platelets in plasma [54, 67], also confirmed that the reaction time (R) is not influenced by the indicating that platelets possess the procoagulant activity (Fig. 2). number of platelets [65]. Saleem et al. also evaluated the effects of Annexin V concentration-dependently prolonged the commence- the number of platelets on the strength of clot as well as ACT by ment of clot formation after reloading of Ca2+ [80], suggesting that using oscillating-probe viscoelastometry [54]. The maximum value platelet-dependent procoagulant activity is at least in part caused by of the clot signal, indicative of the clot strength in oscillating-probe the prompt cell-surface expression of PS that occurs within several viscoelastometry, increased according to the number of platelets in minutes after the reloading of Ca2+. the suspension, supporting the findings using thromboelastography [65]. ACT was reduced according to the platelet number [54], in 5. MOLECULAR MECHANISMS FOR THE MODULATION contrast to the results from thromboelastography demonstrated by OF PROTHROMBOTIC ACTIVITIES IN IMMUNE CELLS Oshita et al. [65]. Azma et al. also demonstrated that the addition of A). Substance P: Neurokinin-1 Receptors (SP/NK1Rs) Pathway platelets to plasma accelerated the clotting time and increased the in Immune Cells clot signal as compared to those in the absence of platelets [67], confirming the early results by Saleem [54]. It is suggested that The target tissue of drugs used for neuraxial anesthe- sia/analgesia is DRG [32]. DRG is the mass of cell bodies of pe- Mechanisms of Action of Anesthetics for the Modulation of Perioperative Current Pharmaceutical Design, 2014, Vol. 20, No. 36 5785 ripheral sensory nerves that provokes neurogenic inflammation in smooth muscles in a condition that the digestive tracts were pre- response to potent noxious stimuli [32, 42]. Currently, neurogenic treated with a competitive muscarinic acetylcholine receptors atro- inflammation is confirmed to be mediated by SP [42]. Because pine and that the tracts insensitive to histamine were used for the NK1Rs, that are preferred receptors for SP and they are also ex- bioassay [87, 88]. pressed in monocytes [44-46], SP is a candidate peptide with inter- It has been speculated in early 1950s that SP should be a cellular signaling functions between sensory nerves and monocytes. chemical transmitter released by the primary afferent nerves since Until the early 1980s, the only known peptide in mammalian tissues the content of SP in the dorsal root ganglion was greater than that in classified as tachykinin was SP [81]. Accumulated evidence for the the ventral roots [89]. It was also confirmed that electrical stimuli involvement of neuropeptides in both of pain medicine and throm- of dorsal roots provoked the release of SP from the isolated spinal bosis is mostly related to SP. In this section, we focus on reviewing cord of the newborn rat [81]. In the late 1990s, several groups re- the pharmacological characteristics of SP and NK1Rs that link the ported observations from mice lacking the gene for SP as well as neurotransmitter properties of SP to its modulatory roles on coagu- NK1R in a rapid succession to confirm a hypothesis of SP to be a lation. neurotransmitter for pain sensation [42, 43, 90]. i. Molecular Characteristics of SP Cao et al. [42] evaluated the behavior against pain in mice lack- SP has been identified in 1971 as a peptide consisting of 11 ing the gene for preprotachykinin A (PPT-A), that encodes SP and a amino acids [81, 82] and later classified as a member of tachykinin related tachykinin neurokinin-A. The proportional localization of family, in which neurokinin A and B are included [81]. Tachykinin neurokinin receptors in brain, spinal cord, and ileum was not influ- is named from a pharmacological property of peptides in this class enced by the knock out of the PPT-A gene. The latency of behavior that constrict smooth muscle cells rapidly [81]. Members in against mid-range thermal pain but not against mild thermal pain tachykinin family possess a common C-terminal sequence of Phe- was significantly reduced in mice lacking the PPT-A gene as com- X-Gly-Leu-Met-NH2, keeping similar biological activities [81, 83]. pared to that in wild-type mice. The frequency of response against On the other hand, the distinct N-terminal sequences of tachykinins nociceptive, inflammatory, as well as visceral painful stimuli was convey receptor specificities [81,83]. Active forms of tachykinins also significantly reduced by knocking out the PPT-A gene. Neuro- are generated from precursor proteins by proteolytic cleavage [81, genic inflammation examined as the plasma extravasation or hind 83]. paw edema provoked by capsaicin administration was significantly Three different gene, TAC1 [84], TAC3 [85], and TAC4 [86], lower in mice lacking the gene for PPT-A in comparison with those encoding a variety of tachykinins have been discovered in human. in wild-type mice. These findings confirmed the early hypothesis raised in 1930s that antidromic vasodilation and the afferent im- The peptide sequence of SP is coded in an exon of TAC1 (exon 3) but is synthesized as a portion in multiple isoforms of precursor pulses provoked by the same nerves are transmitted by the same proteins from the TAC1 gene [81, 83, 84]. The other neuropeptides substance [89]. De Felipe et al. [90] reported findings in mice lack- ing the gene for NK1R in the same issue of the journal. Electro- encoded in several other exons in TAC1 includes neurokinin A (substance K, exon 6), neuropeptide K (neurokinin K, exons 3 - 6), physiologically counted number of spikes, indicative of the noci- and neuropeptide (cid:3) (exons 5 -6) [83]. Therefore, neuropeptides (cid:3) ceptive withdrawal reflex, was increased in the wild-type mice but not in the NK1R-gene knockout mice. On the other hand, the and K is a longer and the longest version of NKA, respectively [83]. threshold force of nociceptive mechanical stimuli as well as the latency of behavior to the latter was not influenced by the deletion The multiple isoforms of tachykinin precursors are synthesized of NK1R gene, suggesting that SP did not mediate the signaling of from four splice variant mRNAs that transcript information from acute pain. However, nociceptive behavior after formalin injection several exons in TAC1 gene [83]. The transcript variant (cid:4) is the at both first and second phase was attenuated by the NK1R gene full-length mRNA transcripted from seven exons in TAC1 gene and knock out. These observations were in agreement with those by three mature neuropeptides including those of SP, neuropeptide K, Zimmer et al. [43] who generated mice lacking the Tac1 gene that and NKA are produced [83]. Transcript variant (cid:2) lacks the se- encodes several tachykinins including SP. Stress-induced analgesia quence transcripted from exon 6 which encodes the information for provoked by cold water swimming, that is dependent on NMDA NKA [83]. Only SP encoded in exon 3 is synthesized from this receptors but independent of opioid, was suppressed in mice lack- variant. Transcript variant (cid:3) lacks the sequence transcripted from ing the gene for NK1R, suggesting that SP activates stress-induced exon 4 which encodes a part of information for neuropeptide K analgesia [90]. These findings using mice lacking the gene for [83]. This is the only variant that encodes neuropeptide (cid:3). Tran- NK1R suggested that the pharmacological roles of SP in the re- script variant (cid:5) lacks the sequence transcripted from exon 4 and 6 sponse against pain are complex [90], and that the blockade of SP- thus only encodes SP [83]. In contrast to TAC1, the sole end prod- induced stimuli through NK1R is not quite beneficial for analgesia uct cleaved from precursors of synthesized from TAC 3 is NKB as expected [91]. [83]. TAC4 encodes hemokinin that preferentially activates NK1R iii. NK1Rs [86]. The characteristics of the site cleaved in the precursor proteins appeared to be quite different from other peptides in members of Three human neurokinin receptors NK1R, NK2R and NK3R tachykinin family [83]. Thus the products synthesized from this are known to exhibit preferences for SP, neurokinin-A and -B, re- gene have not yet been fully understood. spectively, acting as specific receptors for these tachykinins [92]. These receptors are encoded by three distinct genes (TACR1, ii. Pharmacological Properties of SP TACR2, TACR3) that possess molecular characteristics of G pro- It has been reported in 1930s that alcoholic extracts of various tein-coupled receptors (GPCR) [93-96]. tissues obtained from horses, including those of minced muscular wall of the small intestine, possess pharmacological activities of The gene for the SP-specific receptor NK1R is encoded in five exons [92, 94] and two naturally occurring transcript variant constricting intestinal smooth muscles in vitro as well as depressing mRNAs, namely the full-length (long) and the truncated (short) arterial blood pressure in vivo [87, 88]. Such crude extracts, pre- pared as “powder” after evaporating the solvent after overnight mRNA, are yielded by alternative splicing [92]. The short isoform of NK1R lacks the C-terminal 96 amino acid residues that is en- precipitation with sulfuric acid, were called substance P (SP) [87, coded in the last exon 5 [92, 95, 97, 98]. This is due to the stop 88]. From early pharmacological study using isolated digestive tracts with a tension bioassay technique, it was confirmed that SP codon encountered by the unremoved intron between 4 and 5 in the truncated transcript variant. neither acted through muscarinic acetylcholine receptors nor through histamine receptors because SP constricted intestinal 5786 Current Pharmaceutical Design, 2014, Vol. 20, No. 36 Azma et al. Fong et al. [97] reported that the binding affinity of the full- thrombus [31]. It is suggested that MPA reflect the interaction be- length isoform of NK1R to SP was 10-fold higher than that of the tween these cells in red thrombosis [31]. Indeed, the risk of graft truncated isoform, and that the full-length NK1R facilitated the occlusion after peripheral vascular surgery is higher in patients with signal transduction as compared by the latter isoform. Electrical high levels MPA in the early postoperative period [101]. MPA are conductance provoked by SP through the human truncated isoform frequently observed in patients undergoing major joint surgery [31], of NK1R expressed in Xenopus oocyte appears to be almost trace while the incidence of VTE in patients with such procedures is the as compared to those through the full length isoform. highest in every surgical procedure [27, 28, 33-36], suggesting that Lai et al. [99] showed that a human monocytic cell line THP-1 MPA should be a biomarker of increased risk for red thrombus constitutively expressed the truncated NK1R but not the full-length generation including VTE [31]. isoform. However, stimulation of THP-1 with a potent activator of Involvement of several adhesion molecules between monocytes protein kinase C (phorbol myristate acetate), which induces macro- and platelets is suggested to be involved in the formation of MPA phage-like differentiation of these cells, provoked the de novo ex- [31]. Cell surface markers of monocytes reported to be associated pression of the full-length isoform [99]. Azma et al. [67] demon- with MPA were including: P-selectin [102], TF [103, 104] and strated in whole blood obtained from healthy volunteers that SP macrophage-1 antigen (Mac-1) [102]. On the other hand, activated enhanced platelet-dependent clot formation through NK1R, sug- platelets also express P-selectin that binds to its receptor, P-selectin gesting that circulating human leukocytes possess functional glycoprotein ligand-1 (PSGL-1), expressed on virtually every types NK1Rs at protein levels [67], while the isoforms expressed were of leukocytes [105]. Platelets also express PSGL-1 as a ligand for not identified. They also examined the relationship between tran- P-selectin on leukocytes. script variant expression for NK1Rs in whole blood and the modu- Both of monocytes and platelets express glycoprotein IV latory effects of SP on clot formation [66]. It appeared that the ex- (CD36) that binds with a number of ligands, including throm- pression of the full-length mRNA for NK1Rs enhanced the clot bospondin [106], collagen [107], and oxidized low density lipopro- formation by SP while the expression of this isoform at the protein tein [108]. Involvement of this ligand with glycoprotein IV to pro- level was not confirmed [67]. voke MPA in an early phase of vascular injury has been reported Even though the signal transduction through the truncated iso- [109]. form of NK1Rs is significantly less potent than that through the full length isoform [97, 99], Lai et al. demonstrated that the pre- C). Tissue Factor (TF) treatment of human monocytic cells with SP enhances the cytosolic TF is historically known as the thromboplastin but that the lat- Ca2+ elevation provoked by CCL5 in these cells that express the ter, derived from placenta, is actually a combination of TF and truncated but not the full-length NK1Rs [99]. These findings sug- negatively charged phospholipids, called the “partial throm- gested that the truncated NK1R is capable of activating signal boplastin” [57]. As already shown in the previous section, nega- transduction mechanisms that enhance stimuli using different sig- tively charged phospholipids, including PS, is a subcomponent of naling pathways [99]. prothrombinase complex (i.e., factor Xa and a co-factor Va bound As historically reviewed in the previous section, proinflamma- to phospholipids in the presence of Ca2+) [51, 110] (Fig. 1). The tory activity of SP through NK1Rs had been gained the attention of core enzyme of prothrombinase complex factor X (prothrombinase basic scientists [87,88], while clinicians are mostly interested in the itself) is hydrolyzed to be activated by factor IX and factor VII, pharmacological properties of SP/NK1R pathway as the neuro- known as intrinsic and extrinsic tenase (Xase), respectively [63] transmitter for pain [42,43,90]. Noxious stimuli through the activa- (Fig. 1). Likewise, the enzymatic reaction of the intrinsic tenase, tion of polymodal receptors, i.e., capsaicin-sensitive transient po- factor IXa, is accelerated by the cooperation of a co-factor VIIIa tential receptor (TRP) ion channels [100], provoke the accumula- bound to phospholipids in the presence of Ca2+, similar to the tion of SP in afferent sensory peripheral nerves [42]. The accumula- prothrombinase complex [62]. The enzymatic activation of the ex- tion of SP in turn caused the release of SP to provoke neurogenic trinsic tenase, factor VII, provoked by several proteases (thrombin inflammation around the site with noxious stimuli [42]. Taken these or factors Xa, IXa, and VIIa: TF itself) [111] is initiated by the findings together with the clinical evidence for the prevention of molecular bonding with TF [110]. Similar to the prothrombinase VTE by neuraxial anesthesia/analgesia [33-36], it is suggest that complex as well as the intrinsic tenase, the TF: FVIIa bonding is pharmaceutical interventions to alleviate the SP/NK1R pathway enhanced by negatively charged phospholipids in the presence of possess clinical impact to regulate the perioperative inflammatory Ca2+ [110]. disorders including venous thrombotic disorders enhanced by im- TF is a membrane-bound glycoprotein [112] encoded by 6 ex- mune cells. ons [113]. Except a spliced variant form of TF (soluble form of It is also important for the prevention to select a subpopulation TF), the mRNA of which lacks the sequence transcripted from the that possesses increased risks of venous thrombosis. exon 5 as compared to the full-length mRNA [114], the cell mem- brane-bound structure of TF is essential for the molecular function B). monocyte-platelet Aggregates (MPA) and Adhesion Mole- of this glycoprotein. In a physiological condition, TF is expressed cules on the surface of cells which are not exposed to the inner side of blood vessels such as sub-endothelial smooth muscle cells or fibro- As already reviewed in the previous section, platelets play im- portant roles in the generation of white thrombus [20, 21]. How- blasts existing in interstitial spaces [113]. Thus, TF acts as the re- ever, they also play pivotal roles in the generation of red thrombus ceptor for factor VII and triggers coagulation at the extravascular matrix. in which every whole blood component is involved [20, 21]. Plate- lets in white thrombus that covers the damaged vessels are activated However, in a pathophysiological condition, several cells that by several stimuli, thus express PS on their cell surface [51] . Rapid contact to the blood stream express TF on their cell surface, activate blood stream with high shear stress prevent the further generation factor VII to cause intravascular coagulation without exposure of of red thrombus while circumstances with low shear stress promote extracellular matrix to the blood stream [113]. For example, several the generation of red thrombus involving every whole blood com- inflammatory cytokines including tumor necrosis factor-(cid:2) induce ponents [20, 21]. Ex-vivo observation indicated that the strength of expression of TF in endothelial cells that cover the inner side of clot is increased by the function of platelet cell count [54, 65]. blood vessels [113]. Systemically circulating monocytes also ex- Modulatory effects of leukocytes on the platelet-dependent rein- press TF in response to a number of stimuli, including LPS, CRP, forcement of clot were also observed [66,67]. The cell-to-cell inter- proinflammatory cytokines (TNF-(cid:2), IL-1, -4, -10) [113]. actions of these cells should be involved in the generation of red Mechanisms of Action of Anesthetics for the Modulation of Perioperative Current Pharmaceutical Design, 2014, Vol. 20, No. 36 5787 D). Thrombomodulin and Activated Protein C (APC) It has been shown that microparticles released may have a vari- Monocytes, that potentially express TF on the cell surface while ety of effects on other cells. Microparticles are major transport ve- circulating in the blood stream [113], also possess anti-thrombotic hicles for microRNA molecules [134]. Indeed, since RNAases are abundant in blood, in order for microRNA molecules to be deliv- mechanisms. These cells as well as endothelial cells express mole- cules capable of gathering circulating protein C [115], a key ele- ered at sites distant from the cells where they are produced it is ment of anti-thrombosis in human body. necessary to protect them from coming in direct contact with the blood. Thus microparticles appear to be a very effective method of The mature form of protein C is a heterodimer of two peptide packaging these molecules for transport to the target tissue. Using chains that is cleaved from a single precursor molecule synthesized Next Generation Sequencing (NGS) and quantitative reverse tran- in liver and connected by a disulfide bond [115]. The heavy chain scription-polymerase chain reaction (qRT-PCR) to systematically of the heterodimer protein C contains a trypsin-like serine protease compare miRNA profiles of microparticles with the profiles of their domain connected to an activation peptide at the N-terminus [115]. maternal cells in their non-stimulated and stimulated states, Diehl et Thrombin enzymatically cleaves this activation peptide to form al. found large differences between the microRNA profiles ex- activated protein C (APC) [115]. APC, in turn, cleaves several pep- tracted from microparticle generated from stimulated vs. non- tide bonds in factor Va [116] to inhibit the binding of the latter to stimulated monocytic THP-1 cells or endothelial HUVEC cells factor Xa [117]. This drastically decreases the enzymatic activity of [134]. Furthermore, the microRNA profiles of microparticles were prothrombinase complex. It is also known that APC inactivates significantly different from the profiles of the parental cells, further factor VIIIa, a cofactor for the core enzyme of intrinsic tenase fac- confirming that the incorporation and release of microparticles from tor IXa [118] (Fig. 1). cells is a highly regulated process. The same study showed that Thus, factors Va and VIIIa are inactivated by systemically cir- plasma microRNA was specifically located in microparticles. Thus culating protein C which is activated by thrombin. However, enzy- it is clear that microparticles are complex structures that actively matic chain reaction of thrombin to yield APC is further accelerated participate in the regulation of multiple biological processes. by the bonding of thrombin to thrombomodulin [119]. Thrombo- Microparticles shed from variety of cells by stimuli that pro- modulin, expressed on the surface of monocytes as well as endothe- voke apoptosis [135] are suggested to be prothrombotic because of lial cells, forms a 1: 1 stoichiometric complex with thrombin to the cell surface expression of PS [136], a well-known marker for neutralize the enzymatic activity of thrombin that converts fibrino- apoptosis at early phases [51]. It is suggested that monocyte derived gen to fibrin [119]. On the other hand, monocytes and endothelial microparticles are more thrombogenic because of the property of cells possess endothelial protein C receptors (EPCRs) [120]. The these cells for TF expression [103, 136]. Even though monocytes preference of EPCR to APC is equal to that to inactive form of express molecules with anti-coagulant activities as reviewed in the protein C [121]. previous section, those may not favor microparticles shed from The expression of EPCR and thrombomodulin on the surface of intact cell body. monocytes [120] effectively blocks the prothrombotic property of Poitevin et al. [103] reported that the increase in thrombin gen- these cells. This also indicates that certain critical mechanisms are eration induced by Lipopolysaccharide-stimulated monocytes is required to evoke prothrombotic action of monocytes. caused by up-regulation of TF expression and TF-bound micropar- ticle release and that such procoagulant activity is inhibited by en- E). Microparticles dogenous IL-10. We have recently observed that procoagulant ac- Membrane microparticles are cellular fragments with diameters tivity of human monocytic cells (THP-1) stimulated with a Ca2+ below 1 micrometer which are shed into extracellular space from ionophore A23187 is associated with microparticles, but not with cells usually under conditions of stress or injury. The term mi- their origin THP-1, as demonstrated by flow cytometry using FITC- croparticle is usually interchangeably used for a variety of struc- conjugated fibrinogen dissolved in fresh frozen plasma (unpub- tures that originate from cells. At least three major types of mi- lished observation). A series of study conducted by Douglas and croparticles have been described so far: typical membrane mi- colleagues [124, 137, 138] demonstrated that stimulation of various croparticles, exosomes, and apoptotic bodies. The term microvesi- cells by SP/NK1R pathway provokes cell-membrane blebbing and cle is also widely used, especially to refer to mixtures of micropar- microparticle release. ticles or to microparticles that are not well characterized. The molecular mechanisms suggested to be involved in the Typical microparticles consist of membrane proteins and cyto- modulation of prothrombotic activities in monocytes are summa- solic material derived from the cell from which they originate and rized in a scheme (Fig. 4). usually microparticles have diameters ranging between 0.1 - 1.0 (cid:6)m [122, 123] although larger microparticles generated from cancer 6. EFFECTS OF ANESTHETICS ON PROTHROMBOTIC cells have been reported [124]. Microparticle formation is a com- ACTIVITIES OF MONOCYTES mon feature among a large variety of cells and, at least in principle, Several groups of investigators examined the effects of local any cell type may be able to generate microparticles [125]. A large anesthetics on ex-vivo clot formation in whole blood using throm- number of reports focused on microparticles found in circulation, boelastography. The experimental results among these groups are which are generated from endothelial cells, platelets, leucocytes consistent with the concept that the strength of clot formed ex vivo [122, 126] or erythrocytes [127]. However, numerous other cell is inhibited by local anesthetics at concentrations used for neuraxial types have been shown to be capable of generating microparticles: anesthesia/analgesia but that these agents failed to influence the clot smooth muscle cells [128], cardiomyocytes [129] and podocytes strength at concentrations detected in human plasma during sys- [125], as well as cancer [130] and progenitor cell populations [131]. temic local anesthetic administration [139-142]. It is noteworthy Although initially it was believed that microparticle are gener- from these experimental studies that the effective concentrations of ated mainly from dying cells, it became evident that they are re- local anesthetics required to decrease the clot strength (i.e., MA in leased from perfectly viable cells, particularly under conditions of thromboelastography; see, Fig. 3) were lower than those to inhibit stress and cell activation [125]. the clotting time (R in thromboelastography; Fig. 3) [140, 141]. As Microparticles contain most of the structural components of the described in detail in the previous section, platelets and leukocytes cells from which they derive: proteins and membrane lipids, some influence the value of MA but not that of R in thromboelastogra- of the organelles and molecule with key cellular functions: proteins, phy, it is likely that local anesthetics influenced the ex-vivo clot mRNA, microRNA [127, 132, 133]. formation through inhibiting functions of these cells but not through 5788 Current Pharmaceutical Design, 2014, Vol. 20, No. 36 Azma et al. Monocyte-derived microparticles TF TF SP release from (4) peripheral sensory (3) F7a F2 F2a (5) nerves activated by TF TF noxious stimuli F10a CCR1 Platelet F5a CCL5 (6) P- selectin (2) PSGL1 APC SP Activation of Truncated TM F2a signal NK1R transduction (9) pathway EPCR PC SP (1) Mac-1 Full length NK1R (7) Cross linkage with fibrinogen CD34 (8) GP IIb/IIIa Platelet Cross linkage with thrombospondin CD34 Platelet Fig. (4). Molecular mechanisms for the modulation of perioperative thrombosis by immune cells. Noxious stimuli occurred by tissue injury at surgical area provoke the release of pain transmitter substance P (SP) from sensory nerves to their peripheral end- ings as well as to the synaptic cleft toward central nervous system. (1) SP activates monocytes solely through the full-length isoform of neurokinin-1 receptor (NK1R) or; (2) through the truncated isoform of NK1R in cooperation with other stimuli such as chemokine (C-C motif) ligand 5 (CCL5) that binds to its specific receptor C-C chemokine receptor type-1 (CCR1). (3) Activation of monocytes by SP through NK1Rs increases the expression of tissue factor (TF) on the cell surface, and the release of microparticles shed from these cells. (4) Binding of coagulation factor VII (F7) to TF causes the activation of F7 (F7a); (5) that provokes the activation of factor X (F10a), a core enzyme of the prothrombinase complex. Prothrombinase complex located on the negatively charged phospholipids such as phosphatidylserine (PS) on the cell surface of platelets or microparticles yields thrombin (F2a) from prothrombin (F2). Several adhesion molecules promote the formation of monocyte-platelet aggregate (MPA) that concentrates the molecules for thrombin generation toward monocytes. Adhesion molecules suggested to be involved include: (6) P-selectin expressed on platelets bound with P-selectin glycoprotein ligand-1 (PSGL1) on monocytes; (7) macrophage-1 antigen (Mac-1) on monocytes and glycoprotein (GP) IIb/IIIa on platelets bound with fibrinogen; and (8) CD34 on either cells bound with thrombospondin. (9) On the other hand, intact monocytes possess several anti-thrombotic molecules. Protein C (PC) binds endothelial protein C receptor (EPCR) on the surface of monocytes. Thrombomodulin (TM) binds thrombin (F2a) to yield activated protein C (APC) that irreversibly inactivates F5a. inhibiting enzymatic reaction of coagulation factors at concentra- concentration of lidocaine was again higher than those detected in tions used for neuraxial anesthesia/analgesia. human plasma following neuraxial anesthesia [144]. More recent experimental studies focused on the platelet- These findings reported by several investigators from separate leukocyte interactions to clarify the mechanisms of action of anes- groups suggested that local anesthetics at concentrations adminis- thetics in thrombus formation. Wacker et al. [143] examined the trated to epidural space for neuraxial anesthesia/analgesia appeared effects of sevoflurane on granulocyte-platelet aggregation and clot to inhibit prothrombotic activities of platelets, leukocytes and their formation at 1 h and 24 h after inhalation of this anesthetic to interaction to form aggregates, while these agents at concentrations healthy human volunteers. Remarkable changes were observed at relevant to be detected in human plasma during epidural administra- 24 h after inhalation of this agent. Sevoflurane reduced ADP- tion failed to influence such activities of these cells. It is unlikely, induced expression of P-selectin in platelets and the granulocyte- therefore, that such direct effects of local anesthetics against the platelet aggregation. Sevoflurane also decreased the strength of clot thrombogenic changes in cellular components of whole blood ex- formed ex-vivo as compared to control (i.e., without inhalation of plain the anti-thrombotic property of neuraxial anesthesia/analgesia sevoflurane). It appeared that the effects of sevoflurane on ex-vivo that prevent perioperative DVT occurred at a remote area from clot formation were comparable with those on the expression of P- neuraxis. selectin in platelets as well as granulocyte-platelet aggregation. By contrast, Herroeder et al. conducted a RCT in which effects However, these findings [143] do not support a hypothesis from of systemic administration of lidocaine on levels of pro- clinical evidence shown in the section-2 that anesthetics used for inflammatory cytokines, expression of P-selectin, and of platelet- neuraxial anesthesia/analgesia inhibit thrombus generation, since leukocyte aggregates in patients undergoing colorectal surgery sevoflurane is a general, not local anesthetic classified as sedatives [145]. It appeared that lidocaine decreased levels of these outcome with insignificant analgesic effects. Huang et al. conducted an in- measures as compared to control. Because lidocaine is also used as vitro study to evaluate the effects of lidocaine on the ADP- or a local anesthetic during neuraxial anesthesia, these findings may lipopolysaccharide-induced expression of P-selectin in platelets and support the clinical evidence that favors neuraxial anesthesia for the the platelet-leukocyte aggregation [144]. Lidocaine concentration- prevention of venous thrombotic disorders in perioperative settings. dependently inhibited such outcome measures, while the threshold It is not clear why effective concentrations of lidocaine on the ex- pression of P-selectin and of platelet-leukocyte aggregates are dif-