Abstract
Factor XIII (FXIII) is a thrombin-activated plasma coagulation factor critical for blood clot stabilization and longevity. Administration of exogenous FXIII to replenish depleted stores after major surgery, including cardiopulmonary bypass, may reduce bleeding complications and transfusion requirements. Thus, a model of extracorporeal circulation (ECC) was developed in adult male cynomolgus monkeys (Macaca fascicularis) to evaluate the nonclinical safety of recombinant human FXIII (rFXIII). The hematological and coagulation profile in study animals during and after 2 h of ECC was similar to that reported for humans during and after cardiopulmonary bypass, including observations of anemia, thrombocytopenia, and activation of coagulation and platelets. Intravenous slow bolus injection of 300 U/kg (2.1 mg/kg) or 1000 U/kg (7 mg/kg) rFXIII after 2 h of ECC was well tolerated in study animals, and was associated with a dose-dependent increase in FXIII activity. No clinically significant effects in respiration, ECG, heart rate, blood pressure, body temperature, clinical chemistry, hematology (including platelet counts), or indicators of thrombosis (thrombin:antithrombin complex and D-Dimer) or platelet activation (platelet factor 4 and beta-thromboglobulin) were related to rFXIII administration. Specific examination of brain, heart, lung, liver, and kidney from rFXIII-treated animals provided no evidence of histopathological alterations suggestive of subclinical hemorrhage or thrombosis. Taken as a whole, the results demonstrate the ECC model suitably replicated the clinical presentation reported for humans during and after cardiopulmonary bypass surgery, and do not suggest significant concerns regarding use of rFXIII in replacement therapy after extracorporeal circulation.
Keywords
Introduction
Excessive perioperative bleeding is a significant complication of cardiopulmonary bypass (CPB) surgery associated with increased morbidity and mortality, blood transfusion, prolonged hospitalization, and re-operation in an estimated 2–6% of patients (Dacey et al., 1998; Herwaldt et al., 1998; Munoz et al., 1999). Despite the use of pharmacological agents, improvements in bypass pumps, minimized pump time, optimized heparin therapy, and the use of blood conservation techniques, perioperative bleeding continues to be a significant source of morbidity (Dacey et al., 1998; McCusker and Lee, 1999).
Factor XIII (FXIII) levels decrease to 40–60% of normal levels during CPB surgery (Brody et al., 1986; Shainoff et al., 1994; Godje et al., 1998; Chandler et al., 2001), and decreased FXIII levels are associated with increased blood loss as measured by postoperative chest tube output (Shainoff et al., 1994; Dacey et al., 1998; Godje et al., 1998).
In contrast, blood loss has not been correlated with alterations in other coagulation factors, such as fibrinogen, Factor V, or tissue plasminogen activator activity, suggesting that FXIII levels may be limiting in preventing postsurgical hemorrhage. Administration of rFXIII might be expected to reduce microvascular bleeding both by increasing clot strength and by inhibiting plasmin-mediated fibrinolysis (Muszbek et al., 1999). A reported increase in clot strength ex vivo (Chandler et al., 2001), and decreased postoperative bleeding among CPB patients after administration of plasma-derived FXIII support this hypothesis (Godje et al., 1998). Factor XIII has also been shown to reduce endothelial cell permeability ex vivo (Noll et al., 1999), while cardiac and total body edema in children undergoing CPB surgery has been reduced by treatment with plasma-derived FXIII (Wozniak et al., 2001). Thus, administration of rFXIII may reduce both the requirement for blood products and the rate of re-operation due to diffuse microvascular bleeding postoperatively.
Factor XIII is a zymogen (proenzyme) protransglutaminase that is activated by thrombin (Figure 1). A primary role of activated FXIII is to cross-link fibrin and protect it from plasmin-mediated fibrinolytic degradation, thus stabilizing newly formed blood clots. Factor XIII is present in plasma and in platelets, monocytes, and other cells of monocytic lineage. Plasma FXIII (pFXIII) is composed of 2 FXIII-A and 2 FXIII-B subunits (A2B2). While the enzymatic activity as a transglutaminase resides with the FXIII-A subunit, the FXIII-B subunit acts as a carrier and is normally in excess. The cellular form of FXIII (cFXIII) is a homodimer of FXIII-A subunits (A2). As shown in Figure 1, activation of pFXIII by thrombin occurs during clot formation and involves the proteolytic cleavage of an amino-terminal activation peptide on the FXIII-A subunit, the dissociation of the FXIII-A and -B subunits (in the presence of Ca2+), and the exposure of the active site on the free FXIII-A subunits (McDonagh, 1987). The primary substrates of activated FXIII transglutaminase activity include many components of the clotting/fibrinolytic system (e.g., fibrin, fibrinogen, and α2-plasmin inhibitor) as well as adhesive and contractile proteins. Clot stabilization by FXIII is derived through the covalent incorporation of α2-plasmin inhibitor (α2-PI) into the fibrin clot (McDonagh, 1987).
Recombinant FXIII (rFXIII) is produced in Saccharomyces cerevisiae as a nonglycosylated FXIII A2 homodimer that is virtually identical to human cFXIII. The rFXIII readily forms heterotetrameric rA2B2 complexes in the presence of human B subunit (Radek et al., 1993), and demonstrates appropriate transglutaminase activity (Edwards et al., 1993; Procyk et al., 1993; Kurochkin et al., 1995). Recombinant FXIII demonstrates expected cross-linking activity (i.e., formation of fibrin gamma dimers during plasma clotting) in human and cynomolgus monkey plasma, formation of heterotetramer with human or monkey FXIII-B subunit, and pharmacokinetics as evidenced by a circulating half-life in humans and cynomolgus monkeys comparable to that of endogenous FXIII in humans (Ponce et al., 2005; Reynolds et al., 2005).
The clinical evidence demonstrating an association between decreased FXIII activity after CPB and postsurgical blood loss, and the biologic activity of FXIII as an effective fibrinolytic inhibitor provide a rationale for its therapeutic use post-CPB. Thus, studies were conducted to develop a cynomolgus monkey model of CPB and then to evaluate the safety of rFXIII in the model. Results presented here demonstrate a model in cynomolgus monkeys that appropriately reflects the human clinical presentation during and after cardiopulmonary bypass, including blood loss, hemodilution, platelet activation, and coagulation system activation. Administration of rFXIII in this model was associated with markedly increased FXIII activity without apparent pathologic sequelae over a 6-hour postsurgery observation period.
Methods
All procedures were conducted according to a written protocol approved by an institutional animal care and use committee, and with strict compliance to accepted animal welfare standards, including the Public Health Service Policy on Humane Use of Laboratory Animals, the ILAR Guide for the Care and Use of Laboratory Animals, and the Animal Welfare Act. In addition, the facility operates under standard laboratory safety procedures for the safe handling and disposal of the hazardous substances and biological tissues.
Study Animals
Purpose-bred, experimentally naïve male cynomolgus monkeys (Macaca fascicularis) obtained from Covance Research Products (Alice, TX) were used for this study. Animals were approximately 4–8 years of age and weighed 6 to 10 kg. The study animals underwent a comprehensive health screening prior to assignment to study groups, including a complete physical examination by a staff veterinarian with abdominal palpation and observations of the condition of integument, respiratory, and cardiovascular systems. The prestudy determination of health status also included evaluation of a standard panel of serum chemistry, hematology, and coagulation parameters, as well as an examination of fecal samples for ova and parasites.
Test and Control Articles
The test article (rFXIII) manufactured by ZymoGenetics, Inc. (Seattle, WA) has been previously been described (Ponce et al., 2005). Briefly, the rFXIII was formulated as 700 Units/mL (5.0 mg/mL) rFXIII in 3% sucrose, 20 mM L-histidine, 200 mM glycine, and 0.01% (w/v) polysorbate 20, at pH ≈ 8.0. The units of enzymatic activity (U/kg) can be expressed as a concentration of recombinant FXIII (mg/kg) by using the following conversion: 1 mg rFXIII = 142 U, where 1 U equals the amount of transglutaminase activity in 1 mL of normal human plasma. The composition and purity of the test article was characterized by a battery of analyses including size exclusion high-performance liquid chromatography (for content, identity, purity), anion exchange high-performance liquid chromatography (for rFXIII-related charge heterogeneity), potency (for total activity, % activated rFXIII), sterility and limulus amebocyte lysate (for endotoxin). The control article (placebo for rFXIII) consisted of 3% sucrose, 20 mM histidine, 200 mM glycine, and 0.01 (w/v) polysorbate 20, at pH ≈ 8.0. The test and control articles were stored frozen (−10 to −30°C) until use. Dose administration and clinical observations were conducted in a blinded fashion.
Surgical Procedures
Pilot study
Five animals were divided into 2 groups to compare the clinical presentation and outcome after 2 hours of extracorporeal circulation either with (n = 3) or without (n = 2) sternotomy (Table 1). For 12–15 h prior to surgery on Day 1, each animal was fasted but allowed ad libitum access to water. On the day of surgery, the animal was immobilized with ketamine hydrochloride (10–15 mg/kg, IM). The ventral neck, abdomen and both inguinal regions were shaved, and the animals were intubated and placed on isoflurane inhalation anesthesia for the duration of the study. A urinary catheter was placed and the urinary output was periodically monitored during the study.
A sternotomy was conducted on selected animals, a chest tube was placed to facilitate drainage, and the sternotomy was closed. Using standard technique, appropriately sized silastic or polyurethane catheters were placed into the jugular vein and the carotid artery. The venous line supplied blood via gravity feed to the ECC bypass circuit. (At the discretion of the perfusionist, this venous supply was augmented by autologous blood from the venous reservoir during ECC to maintain blood pressure.) In addition to the carotid arterial catheter used for return of oxygenated blood to the animal, an arterial catheter was placed in the femoral artery to facilitate blood sampling and blood pressure measurements. The animals were instrumented for electrocardiography (ECG) and continuous body temperature. Animals were heparinized (100–300 IU/kg, Baxter Healthcare, Deerfield, IL) to achieve an activated clotting time (ACT) of at least 2–3× over baseline (>350 s).
The arterial-venous loop used a ¼″ venous line and a 3/16″ arterial line of polyvinyl chloride tubing. Blood oxygenation occurred through a Cobe Cardiovascular Lilliput 1 hollow fiber membrane oxygenator with an integrated heat exchanger and venous reservoir (Cobe Cardiovascular, Arvada, CO). For each animal, the ECC bypass circuit was primed with autologous blood and lactated Ringer’s solution containing 1000–1200 IU heparin. The temperature of blood in the ECC circuit was maintained with a heated water jacket surrounding the membrane oxygenator. Additional heparin was administered as needed during the ECC procedure to maintain an appropriately prolonged ACT. Animals were weaned off the ECC circuit after 2 h and protamine sulfate (American Pharmaceutical Partners, Schaumburg, IL, 1 mg/100 IU heparin given) was administered to reverse the anticoagulation by heparin.
Twenty min after protamine administration (and upon confirmation of a normalized ACT), a single slow bolus injection of placebo (n = 5) was administered via the saphenous or femoral vein to each animal. The animals were maintained under anesthesia for 10 h after protamine administration. Body temperature, blood pressure, heart rate, electrocardiography, hematology, blood gases, serum chemistry and hemostasis parameters (including APTT, PT, fibrinogen, D-dimer, thrombin:antithrombin (TAT), FXIII activity) were measured at various times as described below. Animals were observed for 10 h after protamine administration, then euthanized and necropsied.
Main Study
To evaluate the preclinical safety of rFXIII administration after 2 h of ECC, 15 animals were divided into 3 treatment groups (Table 1). Surgical procedures and study methods were similar to those described for pilot study animals without inclusion of sternotomy. Additional specialty coagulation analyses of β-thromboglobulin and platelet factor 4 were conducted on selected animals. Twenty min after protamine administration (and upon confirmation of a normalized ACT), a single-blinded slow bolus injection of placebo (n = 6), 300 U/kg rFXIII (2.1 mg/kg, n = 3) or 1000 U/kg rFXIII (7 mg/kg, n = 6) was administered via the saphenous or femoral vein to each animal (Table 1). Animals were observed for 6 h after protamine administration, then euthanized and necropsied.
Body Temperature, Pulse Oxymetry, ECG, Blood Pressure, and Heart Rate
Heart rate, electrocardiography, ventilatory rate, body temperature, and blood pressure (systolic, diastolic and mean) were continuously measured during the surgery, ECC period, and observation period using a Hewlett Packard 78534B physiological monitor. Arterial oxygen saturation and end-tidal CO2 were determined by using a continuous pulse oximeter (Nellcore; Puritan Bennett, Inc, Pleasanton, CA). A heated water blanket was used during surgery with adjustments made to maintain body temperature.
Activated Whole Blood Clotting Time (ACT)
Activated clotting times were collected at baseline (prior to heparinization), at 30 min and 2 h of ECC; at 0, 1, 3, and 6 h postprotamine, and otherwise at the discretion of the study director or perfusionist. Activated clotting times were evaluated on whole blood collected from the in-dwelling catheter into tubes containing diatomaceous earth and analyzed using the Hemochron 801 Whole Blood Coagulation System (International Technidyne Corp., Edison, NJ).
Blood Gases
Blood gas analyses were conducted prior to surgery; at 30 min and 2 h of ECC; and 0, 1, 3, and 6 h postprotamine. Blood gas analyses were conducted on arterial whole blood collected into heparinized syringes and analyzed in an IRMA Blood Analysis System (Diametrics Medical, St. Paul, MN).
Clinical Pathology
Comprehensive clinical pathology analyses (hematology, clinical chemistry, and coagulation analyses) were conducted at baseline, prior to surgery, during the ECC period, and at periodic intervals during the recovery period. Comprehensive hematology analyses were conducted on a HemaVET HV950FS multispecies hematology analyzer (Drew Scientific, Dallas, TX) using arterial blood collected into vacutainer tubes containing EDTA. Endpoints measured included red blood cell counts, mean cell hemoglobin, white blood cell and differential cell counts, mean corpuscular volume, hemoglobin concentration, mean corpuscular hemoglobin concentration, hematocrit, platelet counts, reticulocyte counts, red cell distribution width, and blood cell morphology.
Comprehensive serum chemistry analyses were conducted on a VETTEST 8008 Blood Chemistry Analyzer (IDEXX Laboratories, ME). Arterial blood was collected into separator tubes and serum was isolated by centrifugation (1800 × g, 10 min) for analysis. Endpoints measured included sodium, calcium, potassium, phosphorus, chloride, urea nitrogen, carbon dioxide, creatinine, total bilirubin, total protein, alkaline phosphatase, albumin, lactate dehydrogenase, globulin, aspartate aminotransferase, albumin/globulin ratio, alanine aminotransferase, glucose, gamma-glutamyltransferase, cholesterol, C-reactive protein, and triglycerides.
Coagulation analyses of activated partial thromboplastin time (aPTT), prothrombin time (PT), and fibrinogen were conducted on arterial blood collected into sodium citrate anticoagulant tubes and centrifuged (1800 × g, 10 min) to obtain plasma. Samples from initial surgical groups (Table 1) were frozen and shipped on dry ice for subsequent analysis on an AMAX CS-190 Coagulation Analyzer (Sigma Diagnostics, St. Louis, MO). Samples from follow-up surgical groups were maintained at room temperature and analyzed within 4 h of blood draw in order to avoid development of a cryoprecipitate observed following frozen storage of plasma samples containing elevated rFXIII. To more fully explore the effect on plasma coagulation profiles of freezing cynomolgus monkey plasma containing elevated rFXIII concentrations, 3 additional experiments were conducted. In the first experiment, fresh (unfrozen) sodium citrate- or EDTA-treated plasma was obtained from normal cynomolgus monkeys that were treated with placebo or 1000 U/kg rFXIII. In the second experiment, sodium citrate- or EDTA-treated plasma samples were obtained from untreated (rFXIII-naïve) cynomolgus monkeys and spiked with 25 U/mL rFXIII. In the third experiment, sodium citrate and EDTA-treated plasma samples were obtained from rFXIII-treated cynomolgus monkeys (0, 300, or 1000 U/kg) after 2 h ECC. In each case, plasma samples were divided into 2 aliquots and either analyzed fresh (without freezing) or after overnight storage at −70°C for aPTT, PT and fibrinogen concentration to evaluate the effect of freezing, rFXIII concentration and method of anticoagulation on the assay outcome.
Specialty Hemostasis
D-Dimer, TAT, platelet factor 4 (PF4), FXIII activity, and β-thromboglobulin (β-TG) analyses were conducted using platelet-poor plasma. For TAT and D-dimer analyses, plasma was isolated from arterial blood collected into precooled blood collection tubes containing PPACK and sodium citrate. Analyses of β-TG and PF4 were conducted using plasma isolated from arterial blood collected into precooled blood collection tubes containing trisodium citrate, theophylline, adenosine, and dipyridamole. Factor XIII activity was measured using arterial blood collected into precooled blood collection tubes containing sodium citrate. In all cases, plasma was stored frozen at −80 °C until analysis.
Platelet factor 4 and β-TG were measured using sandwich enzyme immunoassay kits (Asserachrom, Diagnostica Stago, Asnieres-Sur-Seine, France). Plasma D-dimer concentrations were measured using the advanced D-dimer assay (Dade Behring, Deerfield, IL) performed on the Dade Behring BCS analyzer. Thrombin:antithrombin concentrations were analyzed using the Enzygnost TAT microassay (Dade Behring). The relative concentrations of D-Dimer, TAT, PF4, and β-TG in the test samples were calculated against human standards in assays using antibodies directed against the equivalent human proteins; while the assays may under- or overestimate the absolute concentrations of the factors in cynomolgus monkeys, the data allow observation of relative increases or decreases over time. Factor XIII activity was measured using the Berichrom FXIII activity assay (Dade Behring) on the Dade Behring BCS analyzer and was calculated against pooled normal monkey plasma (n = 10). No discernable effect of rFXIII and frozen storage was observed in the measurements obtained for these analytes.
Because blood volume expansion occurred upon initiation of ECC, post-ECC measurements for D-Dimer, TAT, PF4, and β-TG were corrected to allow comparison with baseline measurements. Briefly, for each analyte a theoretical post-ECC concentration estimate was calculated by multiplying the baseline concentration by a hemodilution correction factor; the hemodilution correction factor was estimated at each time point as the ratio of the post-ECC hematocrit to baseline. The difference between the measured post-ECC concentration and the theoretical estimate was then calculated as a measurement of the gain/loss in concentration that could not be explained by hemodilution alone. This concentration difference was then added to the baseline concentration to derive an estimate of the post-ECC concentration corrected for hemodilution. Results are reported using this correction method as relative concentrations or activities of PF4, β-TG, D-Dimer, and TAT (Tables 2 and 3) and as the percentage of baseline concentration for D-dimer and TAT to allow comparison with published human data collected using different assay systems (Table 4). FXIII activity reflected both endogenous and exogenously administered rFXIII and is reported as the percent of normal activity with no correction.
Necropsy
Animal euthanasia was conducted under deep anesthesia via exsanguination. Animals were then perfused with physiological saline followed by neutral-buffered 10% formalin. For all animals, a complete gross examination was performed and the heart, liver, kidney, brain, and lungs were preserved in 10% neutral-buffered formalin. Fixed tissue samples from a subset of animals from each treatment group were embedded in paraffin blocks, sectioned at 5 μM, stained with hema-toxylin and eosin, and evaluated microscopically by a board certified veterinary pathologist.
Results
Clinical Pathology and Outcome in the ECC Model With and Without Sternotomy
To evaluate whether the ECC model was a suitable surrogate for cardiopulmonary bypass surgery, we conducted a pilot study to evaluate the clinical presentation, clinical pathology profile, and necropsy findings in animals after ECC alone or after ECC with sternotomy (Tables 1 and 2). Relative to the clinical presentation of animals that underwent ECC alone for 2 h, inclusion of the sternotomy was a more invasive surgical procedure that was associated with more complications during surgery and the post-ECC period, but not more pronounced indicators of coagulation system activation as measured by D-dimer concentration and TAT (data not shown). Difficulties encountered with inclusion of sternotomy included a prolonged surgical period and re-bleeding from the sternotomy incision site due to physical manipulations or efforts to control blood pressure. Because the sternotomy model did not demonstrate a markedly different coagulation activity profile relative to the model with ECC alone and because the goal was to create a stable model for use in the safety assessment of rFXIII, the model with ECC alone was used in the studies of rFXIII safety.
Clinical and Anatomic Pathology of ECC Study Animals
Plasma coagulation parameters (e.g., aPTT, PT, and fibrinogen) demonstrated expected alterations consistent with adequate heparinization and hemodilution during the ECC procedure. Heparin administration resulting in a prolonged ACT (generally >400 sec) relative to basal (generally 100–130 s) was also associated with prolonged aPTT (>100 sec) and PT (>100 s) during ECC (Tables 2 and 3). The amount of protamine administration at the end of ECC was titrated by correction of the ACT to baseline levels (generally 90–110% of basal), returning the aPTT and PT times to normal. Heparin rebound was observed in a few animals from all treatment groups, and was associated with a prolongation in the ACT (>150 s) and aPTT, and to a lesser extent the PT (Tables 2 and 3). Whereas administration of additional protamine to animals demonstrating heparin rebound returned the ACT and aPTT to basal levels in some animals, the aPTT and ACT remained prolonged or worsened in other animals. No other clinical or anatomic pathologic sequelae were associated with these prolongations.
Establishment of the extracorporeal circuit was associated with blood loss due to circuit priming and hemodilution due to the use of blood volume expanders. Measures of red blood cell mass typically decreased by 40–60% of basal levels during the ECC interval, increased modestly with re-infusion of blood upon termination of ECC, and remained relatively constant throughout the remainder of the observation period (Table 4). Concomitant decreases in peripheral platelet counts were observed (Table 4) and attributed to blood loss, hemodilution and activation, the latter of which was demonstrated by elevated PF4 and β-TG concentrations during the ECC period (Table 3). Evidence of coagulation system activation was demonstrated by a doubling of TAT complex during the ECC period, despite hemodilution and adequate heparinization of study animals as demonstrated by suitably prolonged ACT values of >400 s. D-dimer concentrations appeared to be relatively insensitive to ECC, despite evidence of coagulation activation, and no remarkable changes were observed during the post-ECC period (Table 3).
Administration of rFXIII After ECC
Intravenous slow bolus injection of 300 or 1000 U/kg rFXIII was well tolerated in adult male cynomolgus monkeys when administered after 2 h of extracorporeal blood circulation. No clinically significant treatment-related effects were noted in observed or measured parameters including respiration, ECG, heart rate, blood pressure, body temperature, clinical chemistry, hematology (including platelet counts), or coagulation measurements (Tables 2–4).
Basal FXIII activity in study animals was approximately 37% of those levels established in a small plasma pool derived from 10 cynomolgus monkeys (Table 3). The apparent low level of activity probably reflects the different sources and ages of animals used to generate the plasma pool from those of the pilot and main study animals. For humans, the normal range of FXIII activity is relatively wide (69–143% of normal; Karpati et al., 2000). In this study, pre- and posttreatment FXIII activity levels were used primarily to confirm correct dosing of animals. Initiation of ECC resulted in decreased FXIII activity levels below the assay limit of detection (<25%). Administration of rFXIII post-ECC led to a dose-dependent increase in FXIII activity. Whereas control animals had measured FXIII activity approximately 34% of normal by the end of the observation period (6 h postprotamine), animals treated with 300 U/kg and 1000 U/kg had 646% and 2060% of normal activity, respectively.
Microscopic evaluation of tissues revealed procedure related changes in the brain, liver, heart, lung, and kidney. All of the changes were minimal to mild in severity and seen in one or a few animals, but not in all animals, and there were no effects associated with administration of rFXIII. A variety of microscopic changes were observed in the lung, including perivascular and peribronchiolar aggregates of pigment attributed to the presence of Pneumonyssus sp. (lung mites), which was also possibly associated with observations of pleural fibrosis and subpleural mononuclear cell infiltrate in several animals. Observation of foreign material aspirate and/or blood and exudate in the bronchi in animals from all treatment groups was accompanied by various associated findings (e.g., acute inflammation, necrosis of bronchial epithelium, focal hemorrhage, or diffuse congestion), and attributed to the prolonged anesthesia and associated procedures. Evidence of vascular damage in animals from all treatment groups included perivascular hemorrhage in the brain, hemorrhage in the heart and lungs, thrombi in capillaries of the alveolar septa of the lung and pulmonary vein, thrombi in the renal glomeruli and red blood cells in the renal tubule. Changes that might be related to poor circulation and ischemia included hepatocellular vacuolization (most like due to lipid), centrilobular hepatocellular necrosis, and increased numbers of leukocytes in the hepatic sinusoids. Acute bronchitis was found in the lung of a few animals, possibly associated with the administration of gas anesthesia.
Cryoprecipitation of Plasma Samples Containing Elevated rFXIII
Development of a gelled material, a prolonged PT and aPTT, and decreased measurable fibrinogen were observed upon thawing and coagulation analysis of previously frozen, sodium citrate-anticoagulated plasma samples from rFXIII-treated cynomolgus monkeys (300 or 1000 U/kg); plasma samples spiked with 25 U/mL rFXIII and evaluated after frozen storage demonstrated similar effects. Evaluation of rFXIII-containing plasma samples for coagulation parameters without prior freezing or after EDTA-anticoagulation returned normal results for aPTT, PT, and fibrinogen. Evaluation of plasma samples from animals not treated with rFXIII gave equivalent coagulation results regardless of storage conditions. These results demonstrate that freezing citrated plasma containing elevated rFXIII concentrations will cause aberrant aPTT, PT, and fibrinogen results, and coagulation analysis of EDTA-treated or fresh (unfrozen) sodium citrate-anticoagulated plasma containing elevated rFXIII yield normal coagulation results. Because of the effect of storage conditions on plasma containing elevated rFXIII, plasma coagulation parameter results were only available from separate rFXIII-treated animals in a follow-up surgical group and not from animals treated with either 300 U/mL or 1000 U/mL rFXIII during initial surgeries.
Discussion
The clinical presentation of humans during and after CPB and coronary artery bypass graft is well described, and consists of an acute decrease in measures of red blood cell mass, platelets, and coagulation factors, and an increase in indicators of thrombin activity and fibrinolysis (Rinder, 2000). Contact of blood with the artificial surfaces of the extracorporeal circuit and tissue injury results in activation of platelets and both the extrinsic and intrinsic arms of the coagulation cascade (Edmunds, 1993; Rinder, 2000). Alterations to the hemostatic balance also occur through acute hemodilution from establishment of the extracorporeal circuit and blood volume expansion with colloids and saline (Abdelbaky and Atallah, 1994), consumption of coagulation factors and platelets (Despotis and Goodnough, 2000), and administration of pharmacological agents or blood products to manage bleeding or thrombosis during and after CPB surgery (Despotis et al., 1999). Ischemia-reperfusion and complement activation drive an acute inflammatory response during CPB, resulting in activation of neutrophils, platelets, and endothelial cells that release nitric oxide, oxygen free radicals, arachidonic acid metabolites, and cytokines (Edmunds, 1993; Wan et al., 1997). In aggregate, these effects influence the hemostatic balance between coagulation and fibrinolysis and can contribute to postsurgical tissue anoxia and injury, alterations to vascular tone, bleeding, thrombosis, and microembolization (Levy, 2001).
A model of CPB consisting of 2 h of ECC without sternotomy appears to mimic many of the hemostatic alterations reported during human CPB/coronary artery bypass graft surgical procedures, including anemia, thrombocytopenia, and activation of coagulation (as evidenced by elevated TAT) and platelets (as evidenced by elevated β-TG and PF4, Table 4). Some pilot study animals had a more severe blood loss compared to the average reported for humans undergoing ECC/CPB. Minor modifications to the procedure were undertaken during the pilot study to conserve blood, including the use of autologous blood to prime the extracorporeal circuit and a reduced length and diameter of the bypass lines. These efforts resulted in a marked improvement in hematocrit and other measures of red cell mass to levels similar to those reported for humans.
We have corrected the measured concentration of TAT, D-Dimers, β-TG, and PF4 for the effects of hemodilution using a correction factor based on changes in hematocrit (Hunt et al., 1998). The generation/elimination of each analyte was estimated by comparing a theoretical, hematocrit-adjusted postdilution concentration to the actual measured concentration at each measured time point; the gain/loss term was then used to calculate relative changes in the analyte concentration over time. Use of hematocrit as a correction factor assumes that hemodilution by blood volume expansion proportionally affects the concentration of these analytes and red blood cells.
Heparin rebound detected as a re-prolongation in the ACT and aPTT was observed during the observation period in a few animals from all treatment groups. While additional protamine administration returned the ACT and aPTT times of some animals to normal levels, other animals experienced poorer ACT and aPTT after receiving additional protamine. These results are consistent with a weak anticoagulant activity of protamine when given in excess (Mochizuki et al., 1998), and similar anticoagulation by excess protamine has been observed in a baboon model of CPB (Bernabei et al., 1995). The frequency of heparin rebound in humans varies widely (Jobes et al., 1981), but has been reported in 30–50% of human cases after CPB in 2 studies, and typically occurs within 5 h of the initial protamine reversal (Ellison et al., 1974; Subramamiam et al., 1995). The aPTT is extremely sensitive to heparin, and may be more highly correlated with heparin levels than the ACT (Smythe et al., 2002). Heparin rebound occurs upon redistribution of heparin into blood from extravascular sources (Kaul et al., 1979) and from metabolism/elimination of protamine (Ellison et al., 1974).
Intravenous slow bolus injection of 300 U/kg (2.1 mg/kg) or 1000 U/kg (7 mg/kg) rFXIII after 2 h of ECC was well tolerated in study animals. No clinically significant effects on respiration, ECG, heart rate, blood pressure, body temperature, clinical chemistry, hematology (including platelet counts), or specialty evaluation of indicators of thrombosis (thrombin:antithrombin complex and D-Dimer) or platelet activation (platelet factor 4 and beta-thromboglobulin) were related to rFXIII administration.
Histopathological evaluation of the brain, liver, kidneys, heart, and lung were conducted because these organs appeared to be sensitive targets of occlusive coagulopathy caused by high circulating levels of rFXIII (in excess of FXIII-B) in previous studies conducted in normal cynomolgus monkeys (Ponce et al., 2005). The current macroscopic and microscopic findings in both placebo- and rFXIII-treated animals were considered to be either associated directly or indirectly with the bypass procedure, existing presurgical lesions, or were considered incidental findings occurring in a single animal or commonly occurring in the species used in this study. In particular, there was no evidence of overt thrombosis, microembolization, or hemorrhage in any of the organs and tissues examined in this study.
Gelled material, a prolongation in aPTT and PT, and a decrease in measured fibrinogen were observed after frozen storage of sodium citrate anticoagulated, but not EDTA-anticoagulated, plasma samples containing elevated rFXIII. No gelation and normal coagulation parameters were observed for fresh (unfrozen) plasma samples obtained from rFXIII-treated animals. A variety of conditions are described in which cold/freezing can affect plasma coagulation. Cryofibrinogenemia is one such condition, and is a rare disorder arising from the presence of cold-precipitable plasma proteins associated with malignancies, inflammatory processes, and thrombohemorrhagic phenomena. Cryofibrinogen is composed of two proteins, fibrinogen and cold-insoluble globulin that form stable oligomers interlinked through gamma-gamma dimerization (Stathakis et al., 1981), suggesting FXIII transglutaminase activity has occurred. Diagnosis of cryofibrinogenemia in patients is based on clinical cutaneous manifestations, histopathology, and laboratory detection of cryofibrinogen precipitation. In the current study, no clinical or pathological evidence was observed to suggest that this phenomenon occurred in the study animals in vivo. Rather, the lack of plasma gelation in EDTA-treated samples demonstrates the absence of detectable rFXIII activity prior to sample draw, and demonstrates that the observed gelation occurred only during cold storage as a calcium-dependent process. This conclusion is consistent with the significantly higher affinity of EDTA for calcium relative to sodium citrate and the requirement of calcium for FXIII activity.
The toxicity of rFXIII in normal adult cynomolgus monkeys has been evaluated (Ponce et al., 2005). In these studies, acute morbidity is observed after a single rFXIII intravenous injection of >3200 U/kg (22.5 mg/kg) and may occur as soon as 4–6 hours after dosing. Acute rFXIII toxicity presents with characteristic alterations in serum chemistry parameters (i.e., increased blood urea nitrogen, creatinine, lactate dehydrogenase, aspartate aminotransferase, alanine aminotransferase levels), progressive thrombocytopenia, hypotension, and gross and microscopic evidence of systemic thrombosis and hemorrhage. The toxicity of rFXIII in normal animals is attributed to formation of cross-linked fibrinogen oligomers and higher order protein aggregates in plasma that ultimately interfere with blood flow and cause tissue ischemia. The detection of such cross-linked protein aggregates in normal cynomolgus monkeys after administration of very high levels of rFXIII supports the hypothesis that the gelled material in plasma observed after frozen storage consists of cross-linked fibrinogen and other FXIII substrates that develop in the presence of a calcium-dependent activation of rFXIII and precipitate at low temperatures. Because each of the measures coagulation tests rely on the formation of a clot in solution, the lack of fibrinogen after gel formation would result in the observed coagulation parameter abnormalities.
No clinical or anatomic pathology evidence of thrombosis or coagulopathy was observed in ECC normal study animals that achieved a median FXIII activity of 2060% of normal levels following injection of 1000 U/kg (7 mg/kg) rFXIII. These results suggest a 10-fold margin of safety for patients receiving therapeutic rFXIII doses up to 100 U/kg (0.7 mg/kg) or for patients achieving 200% of normal FXIII activity. Although the current studies were conducted in normal, healthy young adult cynomolgus monkeys rather than animals with heart disease and associated comorbidities, such safety margins support the clinical evaluation of rFXIII in patients that have undergone CPB.