From www.bloodjournal.org by on December 14, 2010. For personal use only.

1993 81: 2644-2651

The roles of von Willebrand factor and factor VIII in arterial thrombosis: studies in canine von Willebrand disease and hemophilia A
TC Nichols, DA Bellinger, RL Reddick, SV Smith, GG Koch, K Davis, J Sigman, KM Brinkhous, TR Griggs and MS Read

Information about reproducing this article in parts or in its entirety may be found online at: http://bloodjournal.hematologylibrary.org/misc/rights.dtl#repub_requests Information about ordering reprints may be found online at: http://bloodjournal.hematologylibrary.org/misc/rights.dtl#reprints Information about subscriptions and ASH membership may be found online at: http://bloodjournal.hematologylibrary.org/subscriptions/index.dtl

Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036. Copyright 2010 by The American Society of Hematology; all rights reserved .

From www.bloodjournal.org by on December 14, 2010. For personal use only.

The Roles of von Willebrand Factor and Factor VI11 in Arterial Thrombosis: Studies in Canine von Willebrand Disease and Hemophilia A
By Timothy C.Nichols, Dwight A. Bellinger, Robert L. Reddick, Scott V. Smith, Gary G. Koch, Keir Davis, Jeff Sigman, Kenneth M. Brinkhous, Thomas R. Griggs, and Marjorie S . Read
We have studied the roles of von Willebrand factor (vWF) and factor Vlll in arterial thrombosis in four canine phenotypes: normal (n = 6). hemophilia A (n = 1l ) , von Willebrand disease (vWD) (n = S), and hemophilia A/vWD (n = 1).vWF activity was determined by botrocetin-induced agglutination of fixed human platelets and vWF antigen (vWF:Ag) by Laurel1 electroimmunoassay and crossed immunoelectrophoresis. Plasma from normal dogs and those with hemophilia A had vWF activity, vWF:Ag, and a full range of vWF:Ag multimers on gel electrophoresis equivalent t o normal canine plasma pool. Platelet cytosol contents were isolated by freezing and thawing, triton X-100 solubilization, or sonication of washed platelets with and without protease inhibitors and inhibitors of platelet activation. Washed platelets were also stimulated with calcium ionophore and MgCI2. There was no measurable vWF activity or vWF:Ag in platelet lysates or releasates in any dog regardless of phenotype. All dogs were studied using a standard arterial stenosis and injury procedure to induce arterial thrombosis. Thromboses were detected by cyclic reductions in Doppler blood flow velocity. Vessels were examined by light and scanning electron microscopy. Thrombosis developed in the arteries of normal (9 of 10) and hemophilia A dogs (16 of 16) but in none of the vWD dogs (0 of 10). Infusion of canine vWF cryoprecipitate into vWD dogs markedly shortened bleeding time but did not support thrombosis as seen in dogs with vWF in the plasma and subendothelium. Thrombosis, then, fails to occur when vWF is absent from the plasma and subendothelial compartments or present only in the plasma compartment. These data are consistent with the hypothesis that vWF in the plasma and subendothelium supports thrombosis. Neither plasma FVlll nor platelet vWF is essential for thrombosis in this model. 0 1993 by The American Society of Hematology.

ON WILLEBRAND FACTOR (vWF)is a large adhesive glycoprotein present in endothelial cells and subendothelium in humans,'-$ pigs,6and vWF is also present in human plateletsg'" and porcine platelets.12,13 There are conflicting reports about whether or not canine platelets contain VWF.',~,'~,'~ A decrease in the degree of expression of plasma or platelet vWF in humans can produce a bleeding diathesis.'6-20 In addition, high molecular weight vWF multimers present in human endothelial cells and platelets may be responsible for part of the biologic activity of vWF ex Variations in both the compartmentalization and multimeric distribution of vWF within these compartments, then, can alter the phenotype of a given species and may be important for the function(s) of the vWF protein. vWF has several complex roles in hemostasis and thrombosis. At sites of exposed subendothelium, vWF supports the initial adhesion of platelets functioning as a ligand between platelet membrane glycoproteins and the subendothelium. This adhesive property of vWF is particularly evident at high

From the Departments of Medicine and Pathology and the Center for Thrombosis and Hemostasis, School ofhfedicine, the Department of Biostatistics, School of Public Health, and the Division of Laboratory Animal Medicine, University of North Carolina at Chapel Hill, NC. Submitted August 10, 1992; accepted December 30, 1992. Supported in part by National Instilutes of Health Grants No. HL01648 and HL26309 and Navy Medical Research and Development Command, Naval Research 14-89-J-1712. Address reprint requesls to Timothy C. Nichols, MD, Departmenl of Medicine. University of North Carolina School of Medicine, C B #7075, 349 Burnett- Womack Bldg, Chapel Hill, NC 27514. The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. section 1734 solely to indicate lhis fact. 0 I993 by The American Society of Hematology. 0006-4971/93/8110-0017$3.00/0
2644

shear rates.25Once platelets attach to the subendothelium, vWF supports platelet spreading in If the vessel is injured at a stenotic site, vWF is required for the development of occlusive thrombo~is.~~"' The relative contribution of vWF in plasma, platelets, endothelium, and subendothelial matrix in platelet attachment and spreading and arterial thrombosis is only partly understood and may differ between species. vWF serves as a carrier for factor VI11 (FVIII) in plasma in humans, pigs, and dogs. This carrier function may provide for the delivery of FVIII to sites of arterial injury. Such a series of events could both localize FVIII activity where vWF attaches to exposed subendothelium and promote thrombosis. However, a monoclonal antibody raised to purified porcine vWF, BB3-BD5, both prevents and aborts thrombosis by neutralizing vWF when infused into normal pigs but does not affect FVIIl activity.29These data suggest that vWF supports thrombosis independent of its function as a carrier for FVIII. Studies have not been done in vivo in the absence of FVIII to confirm this finding. The purpose of this study was to examine the relative roles of plasma and platelet vWF and FVIII in induction of arterial thrombosis in dogs. These studies were performed on the Chapel Hill strains of dogs with normal, hemophilia A, von Willebrand disease phenotypes, and a newly developed strain of bleeder dogs with combined defects, hemophilia A/vWD.~* No platelet vWF was detected in these dogs regardless of phenotype. Thrombosis developed in 90% of stenosed and injured arteries in normal dogs but in none ofthe vWD dogs even when they were infused with vWF concentrates. Replacement of plasma vWF by infusion, then, does not support thrombosis in vWD dogs as seen in dogs that have vWF within the plasma and subendothelium. Thus, thrombosis fails to occur when vWF is absent from the plasma and subendothelial matrix or present only in the plasma compartment. Moreover, thrombosis occurred in hemophilia A dogs as readily as in normal dogs. vWF, then, supports thrombosis independent of its association with FVIII and when it is absent
Blood, Vol 81, No 10 (May 15). 1993: pp 2644-265 1

From www.bloodjournal.org by on December 14, 2010. For personal use only.

THROMBOSIS: CANINE vWF AND FVlll

2645

from platelets. These data are consistent with the hypothesis that vWF in the plasma and subendothelium supports thrombosis. These data have been reported in part.33
MATERIALS AND METHODS

Experimental animals. Normal (n = 2), hemophilia A (n = I l), vWD (n = 6), and hemophilia A/vWD (n = 1) dogs came from the closed colony at the Francis Owen Blood Research Laboratory at the University of North Carolina, Chapel Hill. Four additional normal dogs were purchased from the Division of Laboratory Animal Medicine, University of North Carolina. All animalswere treated according to the standards set in Guide for the Care and Use of Laboratory Animals (National Institutes of Health publication no. 85-23). No animal had received blood products within 2 weeks of these studies. Dogs with hemophilia A had lessthan 1%FVIII activity but normal vWF antigen and activity levels and normal saline bleeding times.34 vWD dogs had less than 1% vWF activity and vWFAg, depressed levels of FVIII activity, and bleeding times longer than 15 minutes. The dog with combined hemophilia A/vWD had no detectableFVIII activity,vWF activity nor vWFAg, and bleeding time of greater than 15 minutes. Dogs were classified as phenotypically normal if they had ( I ) FVIII activity levels and plasma vWF activity and vWF antigen (vWAg) levels 260% ofthat found in normal canine pooled plasma, and (2) normal bleeding times. FVIII assays. FVIII coagulantactivity was performed by a modified one-stage method using a kaolin-activated MIIdeficient canine plasma substrate.35 Plasma and platelet v WF assays. Platelet-poor plasma (PPP) was prepared by centrifugation of blood anticoagulated with 3.2% sodium citrate, pH 6.8 (1/9: vol/vol, 2,OOOg, 5 minutes). PPP was either tested immediately or frozen at -70C until assayed. Platelet-richplasma (PRP) was prepared by centrifugation of blood anticoagulated with acid-citrate-dextrose(ACD: 100 mmol/L sodium citrate, 0.07 mol/ L citric acid, 0.1 1 mol/L dextrose, pH 4.5, 7J42.5: vol/vol, 500g 8 minutes). The platelets were washed three times by resuspension in citrated saline (0.9% NaC1, 5 mmol/L sodium citrate, pH 6.5) and repeat centrifugation.In some experiments, PRP was prepared from blood anticoagulatedwith EDTA (20 mmol/L) containing leupeptin (10 pmol/L) and prostaglandin E-1 (PGE-I) (50 ng/mL). In these experiments, the platelets were washed in citrated saline containing EDTA, leupeptin, and PGE-I at the same concentrations as above. Platelet lysates were prepared by three different methods. First, washed platelets were frozen and thawed three times. Second washed platelets were solubilized by the addition of 20% triton X-100.'5.36 Third, washed platelets were disrupted by sonication (Branson Sonifier Cell Disrupter 200, Danbury, CT). With all three techniques, the platelet debris was separated from the lysate by centrifugation (2,000g. 10 minutes, 23C). Some lysates were prepared in the presence of protease inhibitors and inhibitors of platelet activation (ACD with aprotinin 400 KIU/mL and 300 mmol/L PGE- I or EDTA, leupeptin, and PGE-1 at the concentrations stated above). Normal pig and human platelets were used as positive controls for these methods. In some experiments, washed canine platelets were stimulated to undergo release by the addition of calcium ionophore (A23187, 1 rmol/L and 100 nmol/L) and MgCl, (0.02 mol/L). The stimulated plateletswere removed by centrifugation(2,00Og, 10 minutes, 23C) and the supernatant was assayed for vWF activity and antigen. Aliquots of the washed platelets were ked in glutaraldehyde before and just after producing platelet lysates or releasates for examination by transmission electron microscopy (TEM) to confirm platelet disruption or degran~lation.~~ Plasma, platelet lysates, and platelet releasates were assayed for vWF activity using the macroscopic tap tube procedure with botrocetin and paraformaldehyde-fixed, lyophilized human platelets.374

Fixed human platelets are the standard platelet reagent for vWF assay in our laboratory. We have obtained comparable results with fixed human or fixed canine platelets in this assay." Appropriate dilutions were done to exclude an inhibitor of vWF activity. vWFAg was assayed by Laurel1 electroimmunoassay?' The vWFAg multimer analysis was performed by sodium dodecyl sulfate (SDS) gel electrophoresis as described." The dry gels were scannedby laser densitometry for the distribution of vWF multimers with the LKB Ultroscan XL (Pharmacia LKB, Piscataway, NJ). Crossed immunoelectrophoresis (CIE) of plasma and platelet lysates was performed on 1%agarose gels with and without denaturing agent^.'^^^^ Rabbit anti-canine v WF antibody. A rabbit anti-caninepolyclonal antibody was raised against canine vWF by the method of Lamb et al." Canine cryoprecipitate from the plasma of hemophilia A dogs was chromatographedon 4% agarose (Biogel A- 1 5m, 200-400 mesh; Bio-Rad Laboratories, Richmond, CA). The void volume fractions were pooled and used as antigen for immunizations. The antiserum was made monospecific for vWF by adsorption of 1 vol of antiserum with cryoprecipitate from 2 vol of plasma from a dog homozygous for vWD. This monospecific antibody did not inhibit canine FVIII activity (data not shown) and was used to detect vWFAg in all assays and to stain gels. Canine plasma and plateletfibronectin. Canine fibronectin was measured in platelet lysates, platelet releasates, and PPP by enzymelinked immunosorbent assay (ELISA) with a polyclonal rabbit anticanine antibody.43Fibronectin in the platelet lysates and releasates was used as a control for recovery of platelet a-granule contents. Bleeding time. Saline ear bleeding times were performed on all

Arterial stenosis and injury procedure: Animal preparation, histologic studies, and vessel morphometry. The stenosis and injury procedure was performed under general anesthesia and with physiThis model was modified from ologic monitoring as that reported by Folts and U ~ h i d a . Briefly, 4 ~ ~ ~ general ~ anesthesia was maintained with halothane (1%to 3%)from a vaporizer (Fluomatic Vaporizer, Foregger, Smithtown,NY). The vaporizer was periodically checked for calibration by a Type 18 Riken G a s Analyzer (A.M. Bickford, Wale City, NY). Stenosisand injury were then applied to one or both ofthe carotid arteries and/or a coronary artery (left circumflex or left anterior descending) with a Goldblatt clamp. Thrombosis was detected by cyclic flow reductions in arterial blood flow velocity (Doppler, 20 MHz) following stenosisand injury. After each dog was killed with an overdose of pentobarbital, the heart was pressure perfused and placed in fixatives with the carotid arteries. The degree of arterial stenosis and injury was determined by computer-assistedplanimetry of light microscope arterial cross sections (Ziess Videoplan). Coronary and carotid arterial injury was assessed by the amount of smooth muscle cell damage and disruption of the internal elastic The presence or absence of thrombosis was also determined from examination of arterial sections with light microscopy. Isolation and infusion o f v WF. Cryoprecipitatecontaining vWF but lacking FVIII (FVIIIdeficient cryoprecipitate) was prepared from PPP of hemophilia A dogs?7 The cryoprecipitate was stored at -70C and dissolved in a minimal volume of 0.9% NaCl just before infusion. Fresh normal canine PPP was prepared for infusion into one vWD dog. vWF activity, antigen concentration,and multimeric distribution were determined for each infusate. One unit of vWF activity is defined as that equal to the amount present in 1 mL of normal canine pooled plasma. The products were infused over approximately 60 minutes after the surgical preparation had been completed and all baseline samples had been obtained but before the stenosisand injury procedure. Following vWF product infusion, the stenosis and injury procedure was

From www.bloodjournal.org by on December 14, 2010. For personal use only.

2646
NICHOLS ET AL

Table 1. Effect of Canine Phenotype, FVIII, and Platelet and Plasma vWF Activity and Antigen on the Production of Arterial Thrombosis
BIeeding
Phenotype (n) lnfusate Time (min) FVlll (%) Platelet Count (mm3x I@) Platelet. vWF Activity (%) vWF:Ag Plasma vWF Activity Thrombosis per No. of Aneriest

(%I
<1

(%I

vWF:Ag (%)

Normal (6) vWD (6) Hemophilia A (1 1) vWD (2) vWD (1) Hemophilia A/vWD (1)

None None None FVIII-deficient canine cryoprecipitate Normal canine plasma FV111-deficientcanine cryoprecipitate

2.3 94 (1.5.2.6) (66, 122) >15 15.5 80) (7, 2.7 <1 (1.3.4.4) 4.1 17 12 >15 22 <1

297.5 (135, 365) 265 (225, 275) 270 (230, 260) 298 (295, 300) 305 355

4 1

<1

<1 <1
<1

99 (61, 158) <1 135 (87, 160) 103 (44, 162) 54 15

85 (68, 124) <l

110 (76, 152) 41.5 (36, 44) 67 44

1 0 1 11 t

0/10t 16/16t 014 012 012

<l
<1 <1 <1

<1 <1

vWF activity is expressed as a percentageof normal canine plasma pool. In dogs that received a vWF-containing product, the plasma vWF values given were those obtained after infusion and during the stenosis and injury procedure. Median values are listed with maximum and minimum values listed below in parentheses. * Platelet vWF was determined in a subset of all groups: normal (n = 2). vWD (n = 5 ) .hemophilia A (n = 6).hemophilia A/vWD (n = 1). t The ratio indicatesthe number of arteries that developed thrombosis per total number of arteries tested per group. For example, 1 0 of 1 1 arteries thrombosed in the six normal dogs. t P < .01, Wilcoxon rank sum.

begun. Bleeding time was repeated at 30-minute intervals during the stenosis and injury experimental procedure (2.5 to 3.0 hours). In addition, blood samples were also collected every 30 minutes for measurement o f circulating vWF antigen. vWF activity. and the multimeric distribution o f v W F A g at the time o f the stenosis and injury procedure. Results o f the bleeding time and vWF measurements reported for these animals are those obtained following product infusion and during the stenosis and injury protocol. Slulislicul metliod.~ (?f unuljris. All measured variables are reported as median. minimum. and maximum. The Wilcoxon rank sum test was used to compare the normal, vWD, and hemophilia A dogs in a pairwise manner for variables with ordinal or interval levels o f measurement, and its Kruskal-Wallis extension was used for overall comparisons. Unless otherwise noted, P values are from these tests. Fishers exact test was used for comparisons concerning categorical variables. The association o f measured variables with one another was evaluated with linear regression methods and Spearman rank correlation coefficients.

RESULTS

Caninc plasma vWF, FVIII, and,frhroneciin. The results of measurements of plasma vWF and FVIlI are shown on

Table I for all dogs. Plasma from vWD dogs had less than I % vWF activity and antigen. FVlll activity was reduced in the vWD dogs. Plasma from the hemophilia A dogs had less than I % FVllI activity but normal levels of vWF activity. and antigen. Plasma from the hemophilia A/vWD dog had less than 1% vWFAg. vWF activity. and FVIII activity. The plasma of all normal dogs contained vWF activity. vWFAg. and FVlll equivalent to that found in normal canine pooled plasma. In addition, the distribution of vWFAg multimers was comparable to that in normal pool (Fig I). Plasma fibronectin was measured in dogs ofthe following phenotypes: normal (258 pg/mL. n = I), hemophilia A (420 pg/mL. n = I). and vWD dog (420 pg/mL, n = I). These plasma fibronectin values are within the range reported for d0gs.4~ Caninc. plu~ch~i \-A Fund,fihroneciin. Platelet lysates and releasates were prepared by different physical and chemical procedures chosen to minimize the chance of destroying any extant platelet vWF during p r o ~ e s s i n g . ~ ~ Platelet . ~ * - ~ lysates ~ were prepared from blood anticoagulated with citrate or EDTA and in the presence of protease inhibitors (aprotinin

Fig 1. Multimeric distribution of vWF:Ag in the ) .heplatelets and plasma of normal (lanes 1 and 2 mophilia A (lanes 3 and 4). and vWD (lanes 5 and 6) dogs. Platelet lysates are shown in lanes 1, 3, and 5.Plasma samples are shown in lanes 2,4, and 6.Normal canine pool plasma is shown in lane 7 for comparison.

From www.bloodjournal.org by on December 14, 2010. For personal use only.

THROMBOSIS: CANINE vWF AND FVlll

2647

or leupeptin) and a platelet membrane stabilizer (PGE- 1). None of these procedures yielded detectable canine platelet vWF antigen or activity (Table l), nor were any vWF multimers detected (Fig I). Dilution studies failed to detect an inhibitor of vWF activity in canine platelets. In addition, no precipitin arc was seen on CIE of the platelet lysates of any canine phenotype (data not shown). This finding was in contrast to the precipitin arc detected on CIE of plasmas from normal dogs and hemophilia A dogs that was comparable to that found in normal canine pooled plasma (data not shown). Platelet lysates prepared by these methods from normal pig and human platelets contained normal amounts of vWF activity and antigen and a normal distribution of multimers (data not ~ h o w n ) . ~ ' . ~ ' Complete disruption of platelets by each method (ie, freezing and thawing, Triton X-100, and sonication) was confirmed by transmission electron micrograph (TEM). Platelet degranulation following stimulation with MgC12 and two concentrations of A23 187 was confirmed by TEM taken before and after stimulation (data not shown). Finally, there was no detectable vWF functionally or antigenically in any of the platelet washes before stimulation or disruption. The platelet lysates and releasates from all procedures yielded detectable fibronectin with an average of 5.2 k 1.3 pg/mL X 10" platelets. These data suggest that platelet agranule contents were successfully released and recovered by these physical and chemical methods. Incidence o f thrombosis by phenotype. The incidence of thrombosis in the four canine phenotypes is listed on Table 1. Ten of 1 1 arteries (6 coronary, 5 carotid) tested in 6 normal dogs developed cyclic flow reductions (CFR) indicating thrombosis was induced by stenosis and injury. In addition, CFR occurred in all 16 arteries (5 coronary, 1 1 carotid) tested in 11 hemophilia A dogs (Fig 2). In contrast, no CFRs were detected in any of 10 arteries (2 coronary, 8 carotid) in six vWD dogs (Fig 2). We have previously found that the incidence of thrombosis in these two arteries relative to the presence or absence of vWF is eq~ivalent,~' and all dogs in this present study with assessment for both carotid and coronary arteries had the same outcome. This finding has allowed the grouping of the two circulations for statistical analysis. The difference between the incidence of thrombosis between phenotypes in the present study is significant ( P < .01 for percentage of tested arteries with thrombosis per dog, Wilcoxon rank sum); this difference is also significant for carotids alone for which sufficiently many animals were assessed to allow separate testing. Plasma v WF, bleeding time, and incidence of thrombosis after infusion o f v WF. FVIII-deficient cryoprecipitate was infused into two vWD dogs (336 U of vWF or 19.9 U/kg, and 1260 U or 93 U/kg, respectively) and a hemophilia A/ vWD dog (36 U or 4.6 U/kg). In the two vWD dogs, plasma vWF activity increased from less than 1% to a median of 103% (Table 1). The plasma vWF:Ag increased from less than 1% to a median of 41.5%. The saline bleeding time (SBT) was corrected in both vWD dogs from more than 15 to 4.1 minutes. In the hemophilia A/vWD dog, the vWF antigen and activity increased to 15%and 44%, respectively, after infusion. The prolonged bleeding time was not corrected

in this double-deficient dog. One additional vWD dog was given fresh normal canine PPP (850 U vWF or 32 U/kg). Although vWF activity levels of 54% and vWFAg of 67% were measured, bleeding time was only partially corrected to 12 minutes. A full range of multimers of vWFAg appeared in the plasma of these vWD and hemophilia A/vWD dogs after infusion of the FVIII-deficient cryoprecipitate and normal plasma. The distribution of high molecular weight multimers was comparable by laser densitometry to that present in normal canine pool plasma and FVIII-deficient cryoprecipitate infusate (data not shown). Despite the presence of vWF functionally and antigenically, none of eight arteries (4 coronary, 4 carotid) in these four dogs developed evidence of thrombosis following stenosis and injury (Table 1). Morphometry of stenosis and injury among phenotypes. The degree of stenoses induced by the Goldblatt clamp was similar when phenotypic groups were compared (data not shown). Most arteries were occluded to 80% or more of the cross-sectional area of the artery. Also, each group had one or more arteries with less stenosis. The degree of arterial injury as assessed by the amount of smooth muscle cell damage and disruption of the internal elastic lamina varied over a wide range but was similar across the phenotypic groups. Hemodynamic values, hematocrit, and platelet counts. No significant differences were found in systolic blood pressure, heart rate, or double product (systolic blood pressure X heart rate) when normal, vWD, hemophilia A, and hemophilia A/vWD dogs were compared (data not shown). The diastolic pressure tended to be somewhat lower in the vWD dogs (normal = 69 k 4.9, hemophilia A = 74 k 4.1, vWD = 56 . t 6.8, P < .05, Kruskal-Wallis). Other investigators have shown that cyclic flow reduction in dogs continue unabated by lowering systolic blood pressure from 122.5 to 75.0 mm Hg.52By necessity, a profound reduction in diastolic blood pressure would have also occurred. Given this finding, it seems unlikely that the lower diastolic blood pressure noted in our vWD dogs would explain the difference in the incidence of thrombosis. There were no significant differences between groups in hematocrits or platelet counts (Table 1).
DISCUSSION

The roles of plasma and platelet vWF and factor VI11 were studied in dogs with a stenosis and injury model of induced arterial thrombosis. Occlusive thrombosis was consistently induced in normal and hemophilia A dogs, but vWD dogs did not develop occlusive thrombi even with infused canine vWF products present in the plasma. Platelet vWF was not detected in the dogs of any of the four phenotypes included in this study despite the presence of normal platelet a-granule fibronectin and normal platelet a-granule morphology (not shown). In this model, thrombosis fails to occur in the absence of vWF and is not supported by factor VI11 deficient cryoprecipitate. Thus, replacement of plasma vWF does not support thrombosis as seen in dogs with vWF in the plasma and subendothelium. These data support the hypothesis that the plasma and subendothelium compartments of vWF support thrombosis independent of platelet vWF or of its association with FVIII.

From www.bloodjournal.org by on December 14, 2010. For personal use only.

2648
NICHOLS ET AL

Fig 2. Photomicrographs of stenosed and injured carotid arteries from dogs with hemophilia A and vWD. (A) Thrombosis in a carotid artery of a hemophilia A dog that developed at the site of stenosis and pinch injury. There are regions within the thrombus where platelets are predominant and others where erythrocytes are predominant. (B) Carotid artery from a dog with vWD taken from the stenosis and pinch-injury site. There is no detectable thrombus. Hematoxylin-eosin. Bar = 1 mm.

Subjects with hemophilia A lack FVIII and have serious impairment of thrombin generation. Because human arterial thrombi are rich in fibrin in addition to platelets, arterial thrombosis occumng in the presence of vWF without FVIII may seem paradoxical. However, it has been observed that human patients with severe hemophilia A are not protected from myocardial infarction or thrombosis complicating aortic atheroscler~sis.~~ In both humans and dogs, then, vWF supports arterial thrombosis independent of its association with FVIII. The mechanism by which vWF supports arterial thrombosis in dogs may involve vWF in two compartments: plasma vWF and exposed subendothelial matrix vWF. This hypoth-

esis is supported by the finding that infused plasma vWF fails to support thrombosis in the vWD dogs with vWF-deficient subendothelial matrix (Table 1). Infusion of FVIII-deficient cryoprecipitate into vWD dogs provided plasma vWF with a full range of multimers that was readily detectable functionally and antigenically but would not have replaced subendothelial vWF. In dogs, then, vWF in the subendothelial matrix may be needed to support thrombosis in addition to plasma vWF. Studies in vWD pigs have also shown that plasma vWF alone is insufficient to support arterial thrombosis, whereas the combination of plasma and platelet vWF does support thrombosis.**.29.In vWD pigs, then, platelet vWF may interact with plasma vWF in the formation of

From www.bloodjournal.org by on December 14, 2010. For personal use only.

THROMBOSIS: CANINE vWF AND FVlll

2649

thrombosis at sites of arterial stenosis and injury in the absence of subendothelial matrix vWF. Plasma vWF is the common denominator in a two-compartment model of vWFdependent arterial thrombosis in these two vWF-deficient species: plasma and subendothelial matrix vWF in vWD dogs and plasma and platelet vWF in vWD pigs. Considered together, these data support the hypothesis that plasma vWF may activate, be activated by, or function as a cofactor to vWF in another compartment to support arterial thrombosis. Bleeding time shortening following vWF infusion appears to differ among vWF-deficient species. Infusion of vWF concentrates largely corrected bleeding times in vWD dogs in this study but has failed to do so in vWD pig^.*^,^' This difference in bleeding time response to vWF infusion between vWD pigs and dogs was observed even with plasma vWF activity levels greater than 100%in both species. In humans with vWD, a divergence also has been noted between bleeding time and vWF activity following vWF infusion.54It appears in humans with type I11 (severe homozygous) vWD that replacement of plasma and platelet vWF are required to normalize the bleeding time, whereas plasma vWF alone is in~ufficient.~"~~ Our current study in dogs considered with those performed in pigs and humans suggests there may be species differences in the bleeding time shortening in vWF-deficient subjects. Several possibilities could explain these species differences in bleeding time shortening. Clearly, the variability in vWF compartmentalization among dogs, pigs, and humans could affect this test. It also is possible that plasma vWF in dogs corrects prolonged bleeding time by a mechanism that differs from that by which it supports thrombosis. This hypothesis is supported by the observation that infusion of vWF supports vWF-dependent bleeding time correction without supporting vWF-dependent arterial thrombosis. Alternatively, cryoprecipitation may concentrate other plasma constituents that could shorten the bleeding time in dogs independent of vWF. In addition, there may be changes in the molecular conformation of vWF induced by plasma harvesting or cryoprecipitation that would shorten bleeding time but not support arterial thrombosis. This hypothesis is supported by another report from our laboratory that documented a difference in the half-life of FVIII complexed with endogenous and exogenous vWF.'~Regardless of the mechanism, bleeding time shortening by plasma vWF in the absence of platelet or subendothelial matrix vWF in vWD dogs suggests that plasma vWF is the main determinant of the bleeding time in this species. Other investigators have noted that occasionally mongrel dogs fail to develop thrombosis in this stenosis and injury Three such dogs had a vWF antigen of 92% k 11%; however, the administration of deamino-8-D-argenine vasopressin (DDAVP), produced a rise in vWF antigen to 133% f 12% that was accompanied by cyclic flow reductions. No other data on vWF were given in that abstract. Other studies that have examined canine platelets for vWF antigen have yielded variable results, both the presence and absence of platelet V W F . ~ , ~ Platelet , ' ~ , ' ~ vWF activity was not reported in these studies. In the present study, we used several physical and chemical procedures to isolate platelet vWF from

PRP in several anticoagulant regimens. These procedures were successfully used to isolate porcine and human platelet vWF. For this study, we developed a rabbit anti-canine vWF antibody that was raised against v W F from a dog with hemophilia A and rendered monospecific by adsorption with canine vWD cryoprecipitate. This antibody readily detected vWF antigen in canine plasma but not platelet lysates. These findings are in agreement with those of Potter et al, who used rabbit anti-canine vWF antibodies7 There are two reports that identified small amounts of vWF antigen in the platelets of mongrel dogs. In the first study, anti-human and anticanine antibodies were used that had been rendered monospecific by adsorption with human but not canine vWD p1a~ma.I~ This anti-human vWF antibody was shown to be monospecific for human but not canine vWF by a rocket inhibition method against commercial anti-human vWF antisera.60In the second study, an anti-human vWF antibody from a commercial source was used.I5 No antibody characterization data were given in that study. It is possible that these anti-human vWF antibodies that were not rendered monospecific by adsorption with canine plasma or cryoprecipitate may have recognized epitopes other than canine vWF in canine platelets. It may also be that some breeds of dogs fail to express vWF in their platelets, whereas others express small amounts. We used dogs primarily from the closed colony of inbred dogs, thus providing homogeneous groups of normal, hemophilia A, and vWD. The source of normal canine platelets was mongrel dogs in the studies that detected platelet vWF antigen. Dogs, like humans, may have discordance in the expression of platelet and plasma v W F . ' ~ , ' ~ Halothane was used by itself for anesthesia in this study. This agent has been successfully used in studies of vWF in arterial t h r o m b o ~ i s . *Other ~ - ~ ~investigators have shown that halothane in combination with thiamyl interrupted cyclic thrombosis in this model.'' It may be that the antiplatelet effect of barbiturates on platelets accounts for the apparent discrepancy between our report and that of Bertha et al.6' Our findings corroborate those of Barr et al., who found that halothane did not induce an antiplatelet effect in dogs.62 Our study has shown that when vWF is absent, arterial thrombosis does not occur in this model. An interesting alternative hypothesis for the protection from thrombosis afforded by vWD in dogs could be a concomitant lack of other endothelial cell products such as tissue plasminogen activator. Such a combination of defects has been documented in hum a n ~ To . ~ confirm ~ such a possibility would require characterization of the plasminogen activator response in our vWD dogs. This possibility will require further investigation. vWF supports platelet adhesion, platelet spreading, and arterial thrombosis in several species. The contribution of vWF within plasma, endothelial cells, subendothelial matrix, and platelets to each of these processes is complex. The difference in expression of vWF by canine and porcine endothelium and platelets suggests that these two animal species can be used to explore differences in the function of vWF in its various compartments. Also, agents designed to prevent or treat thrombosis or hemorrhage by modifying vWF in plasma or platelets could be tested for differential results in these two species. Additional comparative studies in dogs,

From www.bloodjournal.org by on December 14, 2010. For personal use only.

2650
NICHOLS ET AL

pigs, and humans should continue to clarify the relative contribution of vWF in its different compartments to thrombosis and hemostasis.
ACKNOWLEDGMENT

The authors thank Robin Raymer, Manager of the Francis Owen Blood Research Laboratory, and her staff for excellent care and management of the animals used in this study.
REFERENCES

I . Bloom AL, Giddings JC, Wilks CJ: Factor VI11 on the vascular intima: Possible importance in haemostasis and thrombosis. Nature 241:217, 1973 2. Hoyer L, de 10s Santos R, Hoyer J: Antihemophilic factor antigen, localization in endothelial cells by immunofluorescent microscopy. J Clin Invest 52:2737, 1973 3. Jaffe E, Hoyer LW, Nachman RL: Synthesis of antihemophilic factor antigen by cultured human endothelial cells. J Clin Invest 52: 2757, 1973 4. Sakariassen KS, Bolhuis PA, Sixma JJ: Human blood platelet adhesion to artery subendothelium is mediated by factor VIII/von Willebrand factor bound to the subendothelium. Nature 279:636, 1979 5. Spom L, Marder V, Wagner D: Inducible secretion of large, biologically potent von Willebrand factor multimers. Cell 46: 185, 1986 6. Bahnak BR, Wu QY, Coulombel L, Assouline A, Kerbiriou D, Pietu G, Drouet L, Caen JP, Meyer D: Expression of von Willebrand factor in porcine vessels: Heterogeneity at the level of von Willebrand factor mRNA. J Cell Physiol 138:305, 1989 7. Potter EV, Shaughnessy MA, Green D Some immunofluorescent observations on factor VIII/von Willebrand factor in dogs. Thromb Haemost 4456, 1980 8. Meyers KM, Wardrop KJ, Helmick CM, White F P Immunohistochemical detection of von Willebrand factor in vascular endothelium but not platelets from plasma VI1R:AG deficient and control dogs and species differences in platelet VI1I:RAG. Fed Proc 46: 424, 1987 9. Koutts J, Walsh PN, Plow EF, Fenton JW, Bouma BN, Zimmerman TS: Active release of human platelet factor VIII-related antigen by adenosine diphosphate, collagen, and thrombin. J Clin Invest 62:1255, 1978 10. Zucker M, Broekman M, Kaplan K: Factor VIII-related antigen in human blood platelets: Localization and release by thrombin and collagen. J Lab Clin Med 94:675, 1979 1 I. Sporn LA, Chavin SI, Marder V, Wagner D: Biosynthesis of von Willebrand protein by human megakaryocytes.J Clin Invest 76: 1102, 1985 12. Cramer EM, Caen JP, Drouet L, Breton-Conus J: Absence of tubular structures and immunolabeling for von Willebrand factor in the platelet a-granules from porcine von Willebrand disease. Blood 86:774, 1986 13. Bowie EJW, Solberg LA, Fass DN, Johnson CM, Knutson GJ, Stewart ML, Zoecklein LJ: Transplantation of normal bone marrow into a pig with severe von Willebrand's disease. J Clin Invest 78:26, 1986 14. McCarroll DR, Waters D, Steidley KR, Clift R, McDonald T P Canine platelet von Willebrand factor: Quantification and multimeric analysis. Exp Hematol 16:929, 1988 15. Parker MT, Turrentine MA, Johnson GS: Von Willebrand factor in lysates of washed canine platelets. Am J Vet Res 52:119, 1991 16. Mannucci PM, Lombardi R, Bader R, Vianello L, Federici AB, Solinas S, Mazzucconi MG, Mariani G: Heterogeneity of type

I von Willebrand disease: Evidence for a subgroup with abnormal von Willebrand factor. Blood 66:796, 1985 17. Gralnick HR, Rick ME, McKeown LP, Williams SB, Parker RI, Maisonneuve P, Jenneau C, Sultan Y: Platelet von Willebrand factor: An important determinant of the bleeding time in type I von Willebrand's disease. Blood 68:58, 1986 18. Ruggeri ZM, Zimmerman TS: The complex multimeric composition of factor VIII/von Willebrand factor. Blood 57:1140, 1981 19. Rodeghiero F, Castaman G, Ruggeri M, Tosetto A: The bleeding time in normal subjects is mainly determined by platelet von Willebrand factor and is dependent on blood group. Thromb Res 65:605, 1992 20. Eller Th, Brauer P, Albert J, Keller F Significanceand quantitative analysis of von Willebrand factor in human platelets. Thromb Res 65:631, 1992 21. Ruggeri 2, Zimmerman T: Variant von Willebrand's disease characterization of two subtypes by analysis of multimeric composition of factor VIII/von Willebrand factor in plasma and platelets. J Clin Invest 65:1318, 1980 22. Meyer D, Obert B, Pieta B, Lavergne T, Zimmerman T: Multimeric structure of factor VIII/von Willebrand factor in von Willebrand's disease. J Lab Clin Med 95:590, 1980 23. Moake JL, Turner NA, Stathopoulos NA, Nolasco LH, Hellums JD: Involvement of large plasma von Willebrand factor multimers and unusually large vWF forms derived from endothelial cells in shear stress-induced platelet aggregation. J Clin Invest 78: 1456, 1986 24. Spom L, Marder V, Wagner D von Willebrand factor released from Weibel-Paladebodies binds more avidly to extracellular matrix than that secreted constitutively. Blood 69:1531, 1987 25. Weiss HJ: von Willebrand factor and platelet function. Ann NY Acad Sci 614:125, 1991 26. Reddick R, Griggs T, Lamb M, Brinkhous K: Platelet adhesion to damaged coronary arteries: Comparison of normal and von Willebrand diseased swine. Proc Natl Acad Sci USA 795076, 1982 27. Reddick RL, Read MS, Brinkhous KM, Bellinger DA, Nichols TC, Griggs TR: Coronary atherosclerosis in the pig: Induced-plaque injury and platelet response. Arteriosclerosis 1054 I, 1990 28. Nichols TC, Bellinger DA, Johnson TA, Lamb MA, Griggs TR: von Willebrand's disease prevents occlusive thrombosis in stenosed and injured porcine coronary arteries. Circ Res 59:15, 1986 29. Bellinger DA, Nichols TC, Read MS, Reddick RL, Lamb MA, Brinkhous KM, Evatt BL, Griggs TR: Prevention of occlusive coronary artery thrombosis by a murine monoclonal antibody to porcine von Willebrand factor. Proc Natl Acad Sci USA 84:8100, 1987 30. Nichols TC, Bellinger DA, Tate DA, Reddick RL, Read MS, Koch GG, Brinkhous KM, Griggs TR: Von Willebrand factor and occlusivearterial thrombosis: A study in normal and von Willebrand's disease pigs with diet-induced hypercholesterolemia and atherosclerosis. Arteriosclerosis 10:449, 1990 3 I. Bellinger DA, Nichols TC, Reddick RL, Read MS, Koch GG, Brinkhous KM. Griggs T: Role of plasma and platelet von Willebrand factor in a porcine arterial thrombosis model. Circulation 82:III-595a, 1990 32. Smith SV, Reddick RL, Brinkhous KM, Read MS: Immunohistological evidence for a blood vessel wall protein analogous to botrocetin. Blood 78:397a, 1991 33. Nichols TC, Bellinger DA, Read MS, Reddick RL, Koch GG, Brinkhous KM, Griggs T von Willebrand factor, factor VIII, and arterial thrombosis: Studies in canine von Willebrand disease and hemophilia A. Thromb Haemost 65:855a, 1991 34. Graham JB, Buckwalter JA, Hartley LJ, Brinkhous KM: Canine hemophilia: Observations on the course, the clotting animaly, and the effects of blood transfusion. J Exp Med 90:97, 1949

From www.bloodjournal.org by on December 14, 2010. For personal use only.

THROMBOSIS: CANINE vWF AND FVlll

2651

35. Langdell RD, Wagner RH, Brinkhous KM: Effect of antihemophilic factor on one-stage clotting tests. J Lab Clin Med 41:637, 1953 36. Rodeghiero F, Castaman G, Tosetto A, Lattuada A, Mannucci P: Platelet von Willebrand factor assay: Results using two methods for platelet lysis. Thromb Res 59:259, 1990 37. Read MS, Shermer RW, Brinkhous KM: Venom coagglutinin: An activator of platelet aggregation dependent on von Willebrand factor. Proc Natl Acad Sci USA 75:45 14, 1978 38. Allain J, Cooper H, Wagner R, Brinkhous K Platelets fixed with paraformaldehyde: A new reagent for assay of von Willebrand factor and platelet aggregating factor. J Lab Clin Med 85:318, 1975 39. Read MS, Smith SV, Lamb MA, Brinkhous KM: Role ofbotrocetin in platelet agglutination: Formation of an activated complex of botrocetin and von Willebrand factor. Blood 74: 1031, 1989 40. Read MS, Potter JY, Brinkhous KM: Venom coagglutinin for detection of von Willebrand factor activity in animal plasma. J Lab Clin Med 101:74, 1983 41. Laurel1 CB: Quantitative estimation of proteins by electrophoresis in agarose gels containing antibodies. Anal Biochem 15:45, 1966 42. Lamb MA, Reisner HA, Cooper HA, Wagner RH: The multimeric distribution of factor VIII-related antigen studied by an improved crossed-immunoelectrophoresis technique. J Lab Clin Med 98:751, 1981 43. ONeill SL, Feldman BF, Boothby JT: Measurement of canine plasma fibronectin by rocket immunoelectrophoresis. Am J Vet Res 46:1175, 1985 44. Mertz E T The anomaly of a normal Duke's and very prolonged saline bleeding time in swine suffering from an inherited bleeding disease. Am J Physiol 136:360, 1942 45. Uchida Y,Yoshimoto N, Murao S: Cyclic fluctuations in coronary blood pressure and flow induced by caronary artery constriction. Jpn Heart J 16:454, 1975 46. Folts JD, Crowell EB Jr, Rowe GG: Platelet aggregation in partially obstructed vessels and its elimination by aspirin. Circulation 54:365, 1976 47. Griggs TR, Potter J, McClanahan SB, Webster WP, Brinkhous KM: Macromolecular factor VI11 complex: Functional and structural heterogeneity observed in von Willebrand swine with transfusion. Proc Natl Acad Sci USA 74:759, 1977 48. Thomas KB, Firkin BG, Howard MA: Analysis of platelet von Willebrand factor antigen. Haemostasis 17:217, 1987 49. Lopez-Fernandez MF, Batlle J, Ruggeri ZM, Zimmerman TS: Secretion of von Willebrand factor from platelets. Methods Enzymol 169:244, 1989 50. Solymoss S, Golden EA, Bovill EG: Measurement of platelet von Willebrand factor is dependent on method of platelet isolation. Am J Clin Pathol 93:400, 1990

5 1. Bellinger DA, Nichols TC, Samama CM, Reddick RL, Bonneau M, Read MS, Bailliart 0, Koch GG, Vaiman M, Brinkhous KM, Pignaud GA, Sigman JL, Gnggs TR, Drouet L Altered phenotypes of normal and von Willebrand disease pigs by bone marrow transplantation: Effects on thrombosis and hemostasis. Blood 78: 179a, 1991 (abstr, suppl 1) 52. Bertha BG, Folts JD, Nugent M, Rusy B F Halothane, but not isoflurane or enflurane, protects against spontaneous and epinephrine-exacerbated acute thrombus formation in stenosed dog coronary arteries. Anesthesiology 71:96, 1989 53. Brinkhous KM, Reddick RL, Griggs T R Arterial thrombosis, atherosclerosis, and the factor VIII/von Willebrand factor complex, in Zimmerman TS, Ruggeri ZM (eds): Coagulation and Bleeding Disorders: The Role of Factor VI11 and von Willebrand Factor, Vol 9. New York, NY, Marcel Dekker, 1989, p 283 54. Blatt PM, Brinkhous KM, Culp HR, Krauss JS, Roberts H R Antihemophilic factor concentrate therapy in von Willebrand disease: Dissociation of bleeding-time factor and ristocetin co-factor activities. JAMA 236:2770, 1976 55. Mannucci P, Moia M, Rebulla P, Altieri D, Monteagudo J, Castillo R: Correction of the bleeding time in treated patients with severe von Willebrand disease is not solely dependent on the normal multimeric structure of plasma von Willebrand factor. Am J Hematol 2555, 1987 56. Smiley R, Titley P, Rock G: Studies on the prolonged bleeding time in von Willebrand's disease. Thromb Res 53:417, 1989 57. Castillo R, Mateagudo J, Escolar G, Ordinas A, Magallon M, Villar JM: Hemostatic effect of normal platelet transfusion in severe von Willebrand disease patients. Blood 77: 1901, 1991 58. Brinkhous KM, Sandberg H, Garris JB, Mattsson C, Palm M, Grigs T, Read MS: Purified human factor VI11 procoagulant protein: Comparative hemostatic response after infusions into hemophilic and von Willebrand disease dogs. Proc Natl Acad Sci USA 82:8752, 1985 59. Folts JD, Coller BS: Role of von Willebrand factor in acute platelet thrombus formation in stenosed dog coronary arteries. Blood 70402a, 1987 (abstr, suppl 1) 60. McCarroll DR, Waters DC, Steidley KR, Clift R, McDonald T P Canine yon Willebrand factor expressesa multimeric composition similar to human von Willebrand factor. Exp Hematol 16:1060, 1987 61. O'Rourke ST, Folts JD, Albrecht RM: Inhibition of canine platelet aggregation by barbiturates. J Lab Clin Med 108:206, 1986 62. Barr SC, Ludders JW, Looney AL, Gleed RD, Erb HN: Platelet aggregation in dogs after sedation with acepromazine and atropine and during general anesthesia and surgery. Am J Vet Res 53:2067, 1992 63. Ludlam CA, Peake IR, Davies BL, Furlong RA, Bloom AL: Factor VI11 and fibrinolytic response to deamino-8-D-argenine vasopressin in normal subjects and dissociate response in some patients with haemophilia and von Willebrands disease. Br J Haematol45: 499, 1980