von Willebrand Factor Homeostasis: ADAMTS13 and Shear Stress
von Willebrand factor (VWF) is an adhesive glycoprotein that exists in normal plasma as a series of multimers with molecular weight ranging up to >20 × 106 Da. Genetic deficiency of VWF causes bleeding diathesis in patients with von Willebrand disease (VWD), which is broadly classified into three types: type 1 VWD, the most common type, causes partial quantitative deficiency of VWF; type 2 VWD causes functional defect in VWF; and type 3 VWD causes nearly complete deficiency of VWF. Type 2 VWD is further divided into 2A, in which the largest multimers are missing; 2B, in which the VWF is hyperactive; type 2M, in which the VWF is inactive but has normal multimer distribution; and type 2N, in which the mutation affects the binding of VWF with factor VIII, resulting in low factor VIII levels.1
The types of VWF defects affect their clinical severity of the disease. For example, mild type 1 VWD causes mild or minimal bleeding symptoms but may cause profuse bleeding under severe hemostatic challenges. Various types of inflammatory conditions such as surgery, infection, inflammation, menstruation, oral contraceptives, and pregnancy may raise the VWF levels, thereby obscuring the diagnosis of type 1 VWD. Desmopressin increases the VWF levels in type 1 patients and has been used effectively to improve the hemostasis for invasive procedures. In contrast, desmopressin, as well as various conditions such as infection, surgery, or pregnancy, may aggravate the binding of type 2B VWF with platelets, resulting in worsening thrombocytopenia and further VWF deficiency. Desmopressin is ineffective for type 3 VWD.
Extensive clinical observations and laboratory studies have revealed three peculiar features of VWF: 1) VWF, among the adhesive proteins, is uniquely essential for hemostasis in the microvasculature, where the shear stress is at its highest levels in the circulation. In vitro, it has also been observed in various experimental systems that shear stress promotes rather than reduces the VWF-mediated platelet adhesion and aggregation; 2) although VWF exists as a series of multimers in the plasma, the same multimers are not present in vascular endothelial cells or megakaryocytes, the only two types of cells known to synthesize VWF; and 3) the VWF adhesive activity is size dependent, with the small multimers essentially ineffective for hemostasis, although their molecular masses, >1 × 106 Da, are higher than any other known proteins.
The VWF mystery only began to unravel with the discovery of ADAMTS13.2 ADAMTS13 is a metalloprotease of the ADAMTS zinc-dependent protease family that is synthesized primarily in the stellate cells of the liver, and VWF is its only known substrate. Although ADAMTS13 coexists with VWF in the plasma and its cleavage site on VWF was identified in 1990,3 the existence of the protease and its antithrombotic function remained unknown until recent years because proteolysis of VWF does not occur when a plasma sample, which contains both VWF and ADAMTS13, is incubated in a test tube. Cleavage of VWF by ADAMTS13 requires shear stress, explaining why the proteolysis occurs in the circulation but not in a test tube.4
For most proteases, the regulation of proteolysis is accomplished by one or more of three mechanisms: compartmentalization of the protease in storage granules, synthesis of the protease as a zymogen that requires activation, and inhibition by protease inhibitors. The coexistence of VWF and active ADAMTS13 in normal plasma, requiring shear stress for their interaction, represents a novel mode of regulation that appears uniquely adapted for high shear stress conditions.
von Willebrand factor is secreted from vascular endothelial cells as a large disulfide-bonded polymer that is resistant to cleavage by ADAMTS13 unless it is exposed to shear stress.5,6 In a capillary tube, wall shear stress at approximately 50–60 dynes/cm2 for as brief as 15 seconds is sufficient to cause proteolysis of VWF. The response to shear stress is observed with both endothelial VWF and large VWF forms of normal plasma. Importantly, the proteolytic products generated by ADAMTS13 in the presence of shear stress are the same as those found in normal plasma. Thus, the plasma VWF multimers are not polymers of the basic VWF polypeptide, rather they are the products of the endothelial VWF polymer at various stages of proteolysis by ADAMTS13 (Figures 1 and 2).
The induction of proteolysis by shear stress suggests that VWF has a flexible conformation that is responsive to shear forces. The conformational change of VWF by shear stress has been demonstrated by atomic force microscopy (Figure 3A) or fluorescent spectroscopy in cone and plate viscometers,7,8 by fluorescent microscopy in a microfluidic device driven by acoustic streaming9 or a parallel plate flow chamber,10 and by small-angle neutron scattering in a rotating cylinder.11 Interestingly, the studies using the small-angle neutron scattering suggest that, when exposed to shear stress or guanidine HCl, conformational changes occurring at the domain level is sufficient to make VWF susceptible to ADAMTS13; conformational stretching to filamentous configurations is not essential for proteolysis.
Previously, it was observed that shear stress enhanced VWF-platelet binding and aggregation in a de-endothelialized, inverted rabbit aortic ring, a cone and plate viscometer, or a parallel plate perfusion chamber.12–15 It was postulated that the VWF-platelet aggregation resulted from shear stress-induced activation of platelets. Nevertheless the shear stress required to induce platelet activation is much higher than the physiological or experimental levels used. The responsiveness of VWF conformation to shear stress prompts the hypothesis that shear stress also augments the adhesive activity of VWF, thereby explaining why shear stress increases VWF-supported platelet adhesion and aggregation6 (Figure 3, B–D). The effects of tensile forces on VWF have been further investigated using laser tweezers on a recombinant A2 domain fragment of VWF in which the Tyr1605-Met1606 is located.16
The recognition of shear stress as a regulator of VWF conformation and the cloning of ADAMTS1317 provides new insights of the VWF homeostasis: 1) high shear stress increases VWF-mediated platelet adhesion and aggregation at sites of vessel injury because it exposes VWFs multiple binding epitopes. Although VWF-glycoprotein 1b (GP1b) binding exhibits a fast on-off kinetics, the presence of multiple binding sites along the direction of flow assures repetitive binding, eventually leading to platelet activation and stable platelet-VWF binding (Figure 4). At sites of vessel injury, VWF may be protected by thrombospondin from cleavage by ADAMTS13.18 Furthermore, the protease may be inactivated by thrombin and plasmin19; 2) Large VWF multimers are hemostatically more effective than small multimers because they are more responsive to shear stress and provide longer stretches of adhesion epitopes for arresting rolling platelets; and 3) ADAMTS13, by cleaving VWF whenever any of its cleavage sites are exposed by shear stress, maintains VWF in its compact, inactive form as it becomes progressively smaller in the circulation (Figure 4).
Clinical Conditions With Deranged VWF Homeostasis
A corollary of these findings is that ADAMTS13, by cleaving VWF whenever its conformation is altered by shear stress, plays a critical role in preventing VWF-platelet aggregation in the circulation and that deficiency of ADAMTS13 may lead to intravascular VWF-platelet aggregation and microvascular thrombosis, as observed in patients with thrombotic thrombocytopenic purpura (TTP) (Figure 5).6 Additionally, the scheme may also explain the VWF changes observed in type 2A VWD and various medical conditions with abnormal shear stresses such as the hemolytic uremic syndrome (HUS), aortic stenosis (AS), and ventricular assist devices (Table 1).
Thrombotic Thrombocytopenic Purpura
Thrombotic thrombocytopenic purpura is a relatively uncommon disorder characterized by widespread VWF- and platelet-rich thrombi in the arterioles and capillaries, where the shear stress is the highest in the circulation. In TTP, VWF-platelet thrombosis is a consequence of severe ADAMTS deficiency, which is caused by autoimmune inhibitors in most cases and by compound heterozygous or homozygous mutations of the ADAMTS13 gene in the much less common hereditary cases.2
Shear stress applied to VWF in normal plasma promotes its proteolysis and decreases its adhesive activity.6 In contrast, the same shear stress applied to VWF in TTP plasma does not induce proteolysis; instead, it increases the adhesive activity of the VWF. It is conceivable that repetition of this process will eventually lead to VWF-platelet binding, resulting in VWF-platelet aggregation and microvascular thrombosis of TTP.
Plasma exchange with fresh normal plasma has been used quite effectively, albeit without a rational pathophysiological knowledge, for more than 30 years in the treatment of acquired TTP, antedating the discovery of ADAMTS13 by more than 20 years. It is now believed that plasma therapy resolves microvascular thrombosis of TTP by replenishing the missing ADAMTS13 and also by removing the autoimmune inhibitors of the protease, decreasing the mortality rate of acquired TTP from >90% to <20%. However, because plasma exchange therapy does not address the underlying autoimmunity to ADAMTS13, relapse of thrombotic complications is common during or after plasma exchange therapy. Occasionally, when the levels of autoimmune ADAMTS13 inhibitors are persistent, the patient cannot be weaned off plasma therapy. Suppression of the autoimmune response with rituximab, a chimeric monoclonal antibody that deplete the B-cells, has been quite effective in inducing remission of persistent TTP and in preventing relapses by decreasing the production of autoimmune inhibitors in acquired TTP.
Patients with ADAMTS13 deficiency due to genetic mutations do not require plasma exchange because they do not have inhibitors and readily respond to periodic small amount of plasma infusion. It is conceivable that recombinant forms of ADAMTS13 may replace plasma therapy in the future.
Type 2A VWD
In some patients with type 2A VWD (group 1), the lack of large VWF multimers results from defective polymerization of the VWF mutants during biosynthesis. In other patients (group 2), the VWF mutants polymerize normally but are constitutively susceptible to proteolysis by ADAMTS13 in the absence of shear stress.20–22 Because large polymers are produced normally in the endothelial cells, infusion of desmopressin to induce VWF secretion may increase the large multimers for patients with type 2A group 2 VWD, providing sufficient hemostasis, at least temporarily, for dental or other brief invasive procedures. In the future, modulation of the ADAMTS13 activity may also be a potential target to restore the missing large multimers in type 2A group 2 VWD.
The Hemolytic Uremic Syndrome
In patients with the HUS after Escherichia coli enterocolitis, microvascular thrombosis is believed to result from the cytotoxic effects of the bacterial shiga toxins (stx) on vascular endothelial cells. The cytotoxicity requires the binding of the toxins to a cell surface receptor, globotriaocylceramide (Gb3, CD77), which is more abundantly expressed in the small vessels of the kidney, explaining why renal failure tends to predominate in shiga toxin-associated HUS (stx-HUS). Recent studies suggest that complement activation may also contribute to endothelial injury in stx-HUS.23,24
In stx-HUS, the ADAMTS13 activity is not decreased. Consequently, the abnormally high levels of shear stress created by microvascular thrombosis lead to excessive proteolysis of VWF, thereby decreasing large VWF multimers and increasing the VWF fragments.25 Similar VWF changes have been observed in patients with other types of microvascular thrombosis but normal or slightly decreased ADAMTS13 levels, such as atypical HUS due to complement dysregulation, allogeneic stem cell transplantation, certain medications, scleroderma renal crisis, and lupus vasculitis.
Aortic Stenosis and Ventricular Assist Devices
A decrease in the large VWF multimers is detectable in most patients with AS, with its severity correlating with the pressure gradient over the aortic valve.26–29 Because exposure to high shear stress is brief, the decrease in large VWF multimers may be subtle in AS, and the conventional VWF analyses such as VWF antigen and activity as measured by the ristocetin cofactor assay may yield normal results. (Ristocetin cofactor assay is a commonly used measurement of VWF activity in which VWF-platelet aggregation is induced by ristocetin, a chemical that binds to the VWF A1 domain, exposing its binding epitope for platelet GP1b.) However, the VWF activity to antigen ratio is decreased and assays of VWF-platelet function under high shear stress conditions, such as bleeding time, platelet retention in a filterometry, or PFA-100 platelet function analyzer, are more likely to yield abnormal results. The cause-effect relationship between AS and VWF defects is supported by the observation that surgical correction of AS is followed by normalization of VWF multimers.28–30
It is tempting to hypothesize that abnormally high shear stress created by AS promotes VWF proteolysis by ADAMTS13, similar to what has been observed in patients with shiga toxin-induced HUS. As yet, there is no direct proof that excessive proteolysis is the sole cause of decreased large multimers in AS. The level of VWF proteolytic fragments, when normalized to that of the intact VWF polypeptide (VWF fragment ratio), is increased in patients with AS.26 However, stoichiometrically this ratio is higher for the small multimers compared to the large multimers. Therefore, selective adsorption or consumption of large multimers, as might occur when the platelets are activated, would also cause an apparent increase in the VWF fragment ratio. Indeed there is evidence that platelets are activated in patients with AS.31 Direct mechanical destruction by high shear stress has also been postulated to account for the loss of large VWF multimers. Whether excessive proteolysis is the major cause of decreased large VWF multimers in AS and in ventricular assist devices remains to be defined.
To explain the occurrence of angiodysplasia bleeding in patients with AS, it has been proposed that angiodysplasia and AS occur separately as a consequence of the aging process. Aortic stenosis causes bowel ischemia and impairs hemostasis by decreasing the large VWF multimers, thereby increasing the risk of gastrointestinal bleeding due to angiodysplasia.32,33 However, this explanation may be inadequate to account for the high frequency of gastrointestinal bleeding among patients with AS. In healthy individuals older than 50 years undergoing colonoscopic screening, angiodysplasia is detected with a prevalence of 0.8%. None of the patients experience bleeding complications within 3 years.34 Due to the infrequency of angiodysplasia among normal individuals, it is possible that AS contributes to the development of angiodysplasia as well as the VWF abnormalities in the affected individuals.
Gastrointestinal bleeding from angiodysplasia occurs only rarely in patients with type 2A VWD, suggesting that loss of large multimers is not directly involved in the pathogenesis of gastrointestinal angiodysplasia.
In patients with ventricular assist devices, particularly of the nonpulsatile type, a decrease in large VWF multimers similar to what has been observed in AS is detected in essentially all patients.35,36 Furthermore, patients with nonpulsatile devices such as HeartMate II (Thoratec, Pleasanton, CA) are also prone to serious gastrointestinal bleeding from angiodysplasia, which may affect as many as 25–50% of the patients within 2 months after LVAD placement and may occur diffusely and recurrently.37–39 It is estimated that patients with pulsatile ventricular assist devices such as Heartmate XVE have only one tenth the risk of bleeding from angiodysplasia.
The frequent occurrence of bleeding angiodysplasia in patients with either AS or nonpulsatile ventricular assist devices prompts a search for their common pathophysiology. Because both AS and nonpalsatile devices are characterized by narrow pulse pressures, it is not unreasonable to postulate that low pulse pressures may be the common culprit. A normal pulse pressure may be critical for normal vasomotor regulation of the microvasculature in the gastrointestinal tract. A decrease in pulse pressures may disrupt the autonomic regulation of the precapillary sphincter, resulting in direct arteriole-venule communication and distension of mucosal and submucosal capillaries and postcapillary venules characteristic of angiodysplasia. This scheme is consistent with the morphologic features of angiodysplasia comprising a distended capillary/venule unit draining to a central small vein and may also explain why decreasing the device flow rate to increase the pulse pressure may help alleviate gastrointestinal bleeding.40,41 Furthermore, the intervention also lowers the shear stress, thereby minimizing the loss of large VWF multimers. Correction of AS, decrease in device flow rate, or transition to a heart transplant is often effective in decreasing gastrointestinal bleeding. Whether these interventions also lead to regression of angiodysplasia remains to be determined.
A unifying scheme of how AS or ventricular assist devices may degrade VWF and lead to gastrointestinal bleeding is proposed in Figure 6.
Ventricular assist devices offer new hopes for patients with advanced congestive heart failure that otherwise face a dismal prognosis. With advances in technical designs including the use of rotary pumps, the devices are increasingly used as a destination as well as bridge therapy. This new application is hampered by the unexpectedly high incidence of gastrointestinal bleeding, which imposes particular grave risk for patients with the devices. Recent advances in VWF homeostasis demonstrate how it may bring new perspectives to seemingly unrelated disorders such as VWD, TTP, HUS, AS, and ventricular assist devices. Future interdisciplinary collaboration may foster new directions of therapeutic intervention and new hardware designs.
Supported by grants R01HL062136 and R21HL109692 of the National Heart, Lung and Blood Institute of National Institutes of Health.
1. Sadler JE. New concepts in von Willebrand disease. Annu Rev Med. 2005;56:173–191
2. Tsai HM. Pathophysiology of thrombotic thrombocytopenic purpura. Int J Hematol. 2010;91:1–19
3. Dent JA, Berkowitz SD, Ware J, Kasper CK, Ruggeri ZM. Identification of a cleavage site directing the immunochemical detection of molecular abnormalities in type IIA von Willebrand factor. Proc Natl Acad Sci U S A. 1990;87:6306–6310
4. Tsai HM, Sussman II, Nagel RL. Shear stress enhances the proteolysis of von Willebrand factor in normal plasma. Blood. 1994;83:2171–2179
5. Tsai HM, Nagel RL, Hatcher VB, Sussman II. Multimeric composition of endothelial cell-derived von Willebrand factor. Blood. 1989;73:2074–2076
6. Tsai HM, Lian EC. Antibodies to von Willebrand factor-cleaving protease in acute thrombotic thrombocytopenic purpura. N Engl J Med. 1998;339:1585–1594
7. Siedlecki CA, Lestini BJ, Kottke-Marchant KK, Eppell SJ, Wilson DL, Marchant RE. Shear-dependent changes in the three-dimensional structure of human von Willebrand factor. Blood. 1996;88:2939–2950
8. Themistou E, Singh I, Shang C, Balu-Iyer SV, Alexandridis P, Neelamegham S. Application of fluorescence spectroscopy to quantify shear-induced protein conformation change. Biophys J. 2009;97:2567–2576
9. Schneider SW, Nuschele S, Wixforth A, et al. Shear-induced unfolding triggers adhesion of von Willebrand factor fibers. Proc Natl Acad Sci U S A. 2007;104:7899–7903
10. Ruggeri ZM, Orje JN, Habermann R, Federici AB, Reininger AJ. Activation-independent platelet adhesion and aggregation under elevated shear stress. Blood. 2006;108:1903–1910
11. Singh I, Themistou E, Porcar L, Neelamegham S. Fluid shear induces conformation change in human blood protein von Willebrand factor in solution. Biophys J. 2009;96:2313–2320
12. Weiss HJ, Turitto VT, Baumgartner HR. Effect of shear rate on platelet interaction with subendothelium in citrated and native blood. I. Shear rate–dependent decrease of adhesion in von Willebrand’s disease and the Bernard-Soulier syndrome. J Lab Clin Med. 1978;92:750–764
13. Moake JL, Turner NA, Stathopoulos NA, Nolasco L, Hellums JD. Shear-induced platelet aggregation can be mediated by vWF released from platelets, as well as by exogenous large or unusually large vWF multimers, requires adenosine diphosphate, and is resistant to aspirin. Blood. 1988;71:1366–1374
14. Goto S, Ikeda Y, Saldívar E, Ruggeri ZM. Distinct mechanisms of platelet aggregation as a consequence of different shearing flow conditions. J Clin Invest. 1998;101:479–486
15. Savage B, Saldívar E, Ruggeri ZM. Initiation of platelet adhesion by arrest onto fibrinogen or translocation on von Willebrand factor. Cell. 1996;84:289–297
16. Zhang X, Halvorsen K, Zhang CZ, Wong WP, Springer TA. Mechanoenzymatic cleavage of the ultralarge vascular protein von Willebrand factor. Science. 2009;324:1330–1334
17. Levy GG, Nichols WC, Lian EC, et al. Mutations in a member of the ADAMTS gene family cause thrombotic thrombocytopenic purpura. Nature. 2001;413:488–494
18. Bonnefoy A, Daenens K, Feys HB, et al. Thrombospondin-1 controls vascular platelet recruitment and thrombus adherence in mice by protecting (sub)endothelial VWF from cleavage by ADAMTS13. Blood. 2006;107:955–964
19. Crawley JT, Lam JK, Rance JB, Mollica LR, O’Donnell JS, Lane DA. Proteolytic inactivation of ADAMTS13 by thrombin and plasmin. Blood. 2005;105:1085–1093
20. Tsai HM, Sussman II, Ginsburg D, Lankhof H, Sixma JJ, Nagel RL. Proteolytic cleavage of recombinant type 2A von Willebrand factor mutants R834W and R834Q: inhibition by doxycycline and by monoclonal antibody VP-1. Blood. 1997;89:1954–1962
21. O’Brien LA, Sutherland JJ, Hegadorn C, et al. A novel type 2A (Group II) von Willebrand disease mutation (L1503Q) associated with loss of the highest molecular weight von Willebrand factor multimers. J Thromb Haemost. 2004;2:1135–1142
22. Hassenpflug WA, Budde U, Obser T, et al. Impact of mutations in the von Willebrand factor A2 domain on ADAMTS13-dependent proteolysis. Blood. 2006;107:2339–2345
23. Ståhl AL, Sartz L, Karpman D. Complement activation on platelet-leukocyte complexes and microparticles in enterohemorrhagic Escherichia coli-induced hemolytic uremic syndrome. Blood. 2011;117:5503–5513
24. Morigi M, Galbusera M, Gastoldi S, et al. Alternative pathway activation of complement by Shiga toxin promotes exuberant C3a formation that triggers microvascular thrombosis. J Immunol. 2011;187:172–180
25. Tsai HM, Chandler WL, Sarode R, et al. von Willebrand factor and von Willebrand factor-cleaving metalloprotease activity in Escherichia coli O157:H7-associated hemolytic uremic syndrome. Pediatr Res. 2001;49:653–659
26. Pareti FI, Lattuada A, Bressi C, et al. Proteolysis of von Willebrand factor and shear stress-induced platelet aggregation in patients with aortic valve stenosis. Circulation. 2000;102:1290–1295
27. O’Brien JR, Tsai HM, Etherington MD. Defective von willebrand factor activity detected by the filterometer in three clinical conditions. Platelets. 2000;11:388–394
28. Vincentelli A, Susen S, Le Tourneau T, et al. Acquired von Willebrand syndrome in aortic stenosis. N Engl J Med. 2003;349:343–349
29. Yoshida K, Tobe S, Kawata M, Yamaguchi M. Acquired and reversible von Willebrand disease with high shear stress aortic valve stenosis. Ann Thorac Surg. 2006;81:490–494
30. Love JW. The syndrome of calcific aortic stenosis and gastrointestinal bleeding: resolution following aortic valve replacement. J Thorac Cardiovasc Surg. 1982;83:779–783
31. Natorska J, Bykowska K, Hlawaty M, Marek G, Sadowski J, Undas A. Increased thrombin generation and platelet activation are associated with deficiency in high molecular weight multimers of von Willebrand factor in patients with moderate-to-severe aortic stenosis. Heart. 2011;97:2023–2028
32. Warkentin TE, Moore JC, Anand SS, Lonn EM, Morgan DG. Gastrointestinal bleeding, angiodysplasia, cardiovascular disease, and acquired von Willebrand syndrome. Transfus Med Rev. 2003;17:272–286
33. Sucker C. The Heyde syndrome: proposal for a unifying concept explaining the association of aortic valve stenosis, gastrointestinal angiodysplasia and bleeding. Int J Cardiol. 2007;115:77–78
34. Bhutani MS, Gupta SC, Markert RJ, Barde CJ, Donese R, Gopalswamy N. A prospective controlled evaluation of endoscopic detection of angiodysplasia and its association with aortic valve disease. Gastrointest Endosc. 1995;42:398–402
35. Budde U, Bergmann F, Michiels JJ. Acquired von Willebrand syndrome: experience from 2 years in a single laboratory compared with data from the literature and an international registry. Semin Thromb Hemost. 2002;28:227–238
36. Velik-Salchner C, Maier S, Innerhofer P, et al. An assessment of cardiopulmonary bypass-induced changes in platelet function using whole blood and classical light transmission aggregometry: the results of a pilot study. Anesth Analg. 2009;108:1747–1754
37. Crow S, Milano C, Joyce L, et al. Comparative analysis of von Willebrand factor profiles in pulsatile and continuous left ventricular assist device recipients. ASAIO J. 2010;56:441–445
38. Crow S, John R, Boyle A, et al. Gastrointestinal bleeding rates in recipients of nonpulsatile and pulsatile left ventricular assist devices. J Thorac Cardiovasc Surg. 2009;137:208–215
39. Uriel N, Pak SW, Jorde UP, et al. Acquired von Willebrand syndrome after continuous-flow mechanical device support contributes to a high prevalence of bleeding during long-term support and at the time of transplantation. J Am Coll Cardiol. 2010;56:1207–1213
40. Hayes HM, Dembo LG, Larbalestier R, O’Driscoll G. Management options to treat gastrointestinal bleeding in patients supported on rotary left ventricular assist devices: a single-center experience. Artif Organs. 2010;34:703–706
41. Frazier OH, Myers TJ, Westaby S, Gregoric ID. Clinical experience with an implantable, intracardiac, continuous flow circulatory support device: physiologic implications and their relationship to patient selection. Ann Thorac Surg. 2004;77:133–142