Heparin: Do We Understand Its Antithrombotic Actions?

Abstract

Over the past 30 years, heparin has become the drug most commonly used to treat a variety of acute thrombotic disorders. Heparin reduces the incidence of recurrence while causing relatively few adverse reactions. However, the protection afforded by heparin against recurrent thrombosis is not complete, and the amount of the drug which can be administered is limited by the risk of serious bleeding. Moreover, occasional patients develop paradoxic thrombotic complications associated with heparin therapy itself. Recently, a number of investigators have questioned the prevailing view that heparin prevents thrombosis primarily by acting as an anticoagulant. A more complete understanding of the mechanism by which heparin inhibits thrombus formation and the mechanism by which heparin initiates bleeding, may help in the formulation of newer drugs with greater therapeutic potential.Molecular HeterogeneityHeparin is prepared commercially from beef or porcine intestine or lung, where it is present in trace amounts. The methods used by different manufacturers for its extraction remain unpublished. Unlike most drugs, heparin is neither chemically nor biologically homogeneous. Rather, heparin is composed of multiple, chemically-related sulfated mucopolysaccharides which vary in primary sugar sequence, molecular weight, and charge.1 In commercial heparin, anticoagulant activity is present in fractions with molecular weights between approximately 4,000 and 40,000 daltons2–4 and over 20 molecular species differing in charge can be identified in a single preparation by electrophoretic analysis.5 Moreover, there may be differences in the chemical and biophysical properties of heparin prepared by different manufacturers, and in heparin prepared by the same manufacturer at different times.6,7 These distinct molecular species are all called heparin because they share a measurable net “anticoagulant” activity defined by specific in vitro assays. However, they may differ from each other in a variety of other biologic activities, as well as in the mechanism by which they act to prevent thrombosis.Standardization of Commercial Heparin for Clinical UseCommercial heparins are standardized for clinical use by comparing the anticoagulant activity of a particular product with one of a number of standard heparin preparations. For example, the International Heparin Unit (IU) is defined as the activity present in 1/130 mg of the International Standard Heparin. In this country, the units of the United States Pharmacopea are defined by the extent to which a standard heparin preparation inhibits the clotting of recalcified sheep plasma.8 Commercial heparins must have an established minimum activity per milligram of heparin salt for clinical use, but there is significant variation between available preparations.9The choice of this standard for comparing commercial products has been based on the following two assumptions: (1) that each preparation of heparin shares a common mechanism of action; and (2) that the potency of the in vivo antithrombotic activity of heparin can be measured by a single in vitro test of anticoagulant activity. Over the past several years, questions have been raised about each of these assumptions. For example, only 30 to 50 percent of the molecules in commercial heparin have measurable activity in the plasma clotting assay system used to standardize heparin.10 Second, the WHO Collaborative Study found that heparin prepared in different laboratories or in the same laboratory at different times varied by as much as 40 to 50 percent in anticoagulant activity using this single assay.6 Third, the potency of various heparin preparations standardized in USP units may differ when other sources of plasma and procoagulant materials are used to measure anticoagulant activity.7,11 Nevertheless, despite these in vitro differences, it is remarkable that heparins prepared by different manufacturers appear to show little variation in their capacity to prevent thrombosis or in the incidence of bleeding which follows their use.Mechanism of Anticoagulant ActivityThe anticoagulant properties of heparin are mediated primarily through its interaction with the plasma protein anti-thrombin III (AT-III). Following the binding of heparin, the AT-III molecule undergoes a conformational change which accelerates the rate of formation of inactive complexes between AT-II? and several serine proteases of the coagulation system, including thrombin, Xa, IXa, plasmin, and possibly XIa and XIIa.12–17Several steps are required for these coagulation proteins to be inactivated by heparin at the maximal rate. First, heparin must bind to AT-III. Fractions of heparin which show little binding to AT-III demonstrate little anticoagulant activity in vitro or in vivo.10,18,19 Binding requires a specific decasaccharide (or pentasaccharide) sequence which is found on less than 50 percent of the molecules in commercial heparin and which comprises only a small portion of its total composition.13,20However, heparin fragments composed only of this binding sequence demonstrate limited anticoagulant activity.13,21 Sequences consisting of approximately 10 to 12 monosaccharides are required for AT-III to show significant anti-Xa activity while heparin sequences containing 16 to 20 monosaccharides are needed for maximum antithrombin activity. It is not known precisely how these additional residues, which lie outside the AT-III binding site, increase the rate with which the activated coagulation proteins are inactivated. It has been suggested that heparin may bind directly to the coagulation proteins and stabilize their interaction with the AT-III molecule.21–24It is also possible that part of the anticoagulant effect of heparin is not mediated directly through AT-III. For example, it has been reported that purified fractions of heparin, which demonstrate high affinity for AT-III, do not possess either the full anticoagulant potency of the native preparation, nor are they as effective as commercial heparin in preventing thrombus formation in experimental animal models.19,25 Although the residual, low affinity fraction lacked both activities by itself it fully restored the anticoagulant and antithrombotic effects of the high affinity fraction. Second, heparin may inhibit directly the function of several serine proteases.22–24 Third, heparin may inhibit the capacity of phospholipids to accelerate coagulation reactions which occur on the surface of activated platelets.26,27 However, the clinical importance of each of these in vitro effects is far from clear since they either require concentrations of heparin which are above the usual therapeutic range (0.2 to 0.5 U/ml) for their maximal expression or these reactions appear to proceed more slowly than neutralization of thrombin and Xa by heparin in the presence of AT-III.Heparin also accelerates the inactivation of thrombin by a distinct plasma protein, heparin co-factor II (HC-II).28 HC-II has been estimated to account for only a small percentage of the anticoagulant activity of heparin in normal plasma at the usual therapeutic concentrations based on rate constants determined using purified proteins.29 However, the recently described association of venous and arterial thrombosis with hereditary deficiency of HC-II suggests this protein may play a more important physiologic role than previously recognized.125,128 Unlike AT-III, the activity of HC-II is also accelerated by dermatan sulfate which is present within vascular tissue. Moreover, unlike the interactions of heparin with AT-III, the effect of dermatan sulfate on HC-II is not neutralized in vitro by tissue factor, which is released during vascular trauma.30 Therefore, HC-II may play a role in inhibiting the actions of serine proteases within vascular tissues.29Monitoring Anticoagulant Therapy and the Risk of BleedingApproximately 1 to 4 percent of patients who receive intravenous heparin develop clinically significant spontaneous bleeding,31 although a much higher incidence has been reported in other studies.32 The risk may be greater in the elderly, in patients with hypertension, following recent trauma or surgery, or in patients with additional hemostatic abnormalities such as platelet dysfunction. This risk of bleeding limits the dose of heparin which can be safely administered.There is no single test which unequivocally identifies the individual patient receiving heparin who is at risk for recurrent thrombosis or the patient at risk for spontaneous hemorrhage. At the present time, heparin therapy is monitored, and therapeutic levels are defined by the extent to which the activated partial thromboplastin time (APTT) is prolonged. However, questions have been raised as to whether the APTT reliably defines an adequate (ie, antithrombotic) level of heparin (see below) or can be used to identify patients at risk for bleeding. First, when heparin is added to blood in vitro, the prolongation of the APTT is reported to be proportional to the log of the heparin concentration.33 Therefore, only a narrow range of heparin concentrations can be assessed, and small changes in the heparin concentration may give either minimal or more marked changes in the APTT depending on its initial concentration. Second, the APTT measures the net activity of a series of coagulation proteins and control proteins activated in vitro in a particular way. Although heparin affects each of the serine proteases in the coagulation system, the extent to which the APTT becomes prolonged does not vary directly with the concentration of residual functioning protein. Deficiencies of factors involved in the contact phase of the intrinsic pathway cause marked prolongation of the APTT but are not associated with clinical bleeding, while deficiencies of factors VIII, IX, and X are associated with more severe bleeding disorders but cause proportionately less prolongation of the APTT. Third, there are extensive data indicating that certain low molecular weight fractions of heparin which prevent thrombosis and neutralize Xa activity cause little or no change in the APTT.34–36 Therefore, prolongation of the APTT can result from a variety of different individual and interrelated coagulation abnormalities, some of which may place the patient at greater risk for bleeding than others.Fourth, and perhaps more importantly, heparin affects multiple steps in the hemostatic process, some of which do not cause changes in the APTT at all. For example, heparin prolongs the bleeding time, although the mechanism(s) by which this occurs is unknown.37 Heparin fractions, especially those of high molecular weight and low affinity for AT-III, bind to platelets38–40 causing a variety of functional abnormalities including inhibition of collagen-induced aggregation.30,40 As detailed below, heparin also binds to vascular tissue and may inhibit the adherence of platelets to the vessel wall under some circumstances.41,42 Furthermore, as noted below, heparin also increases serum fibrinolytic activity. However, it is unknown if prolongation of the bleeding time, the effect of heparin on platelet function, or the increase in serum fibrinolytic activity increases the risk of bleeding in certain patients receiving heparin. Although preparations of commercial heparin have a variable composition and could thereby affect each of these systems to a greater or lesser extent, the drug is only standardized for clinical use by its effect on the clotting time, and heparin therapy is monitored clinically only by the prolongation of the APTT. This single test may not be sufficient to fully evaluate the overall effect of the drug on hemostasis, and may thereby lead to underestimating the risk of bleeding in an individual patient.Antithrombotic Mechanism of Heparin: Development of Low Molecular Weight HeparinExtensive clinical and laboratory evidence indicates that heparin prevents thrombosis by acting as an anticoagulant, ie, by inhibiting the formation or actions of thrombin in the circulation. Recently, however, the mechanism by which heparin prevents thrombus formation has been examined in greater detail with a view to developing chemically related compounds which are not anticoagulants and which, therefore, may be less likely to cause bleeding. These studies have been interpreted as suggesting that heparin inhibits thrombosis through mechanisms which are independent of its known anticoagulant actions.For example, dissociation between the antithrombotic and anticoagulant effects of heparin is suggested by the work of Chiu et al43 who measured the accumulation of 125I-labeled fibrinogen within ligated segments of rabbit jugular veins following infusion of thrombin. As expected, less deposition of fibrinogen occurred in animals treated with heparin than in animals receiving saline as a control. The degree of protection correlated with both the plasma heparin level (measured by protamine neutralization) and the extent of anticoagulation (measured by the APTT). However, if the heparinized animals were subsequently infused with sufficient cryoprecipitate to lower the APTT towards normal, the accumulation of fibrinogen was still inhibited proportional to the amount of heparin in the plasma. Although these results are restricted to the particular model of inducing and measuring thrombus formation, they suggest that heparin may have important inhibitory effects on other aspects of thrombus development. Furthermore, discarded heparin fractions considered inactive by virtue of possessing little effect on the AT-III-thrombin interaction as measured by the APTT may possess clinically important antithrombotic activity.Prolongation of the APTT also does not prevent the development of thromboses in patients with heparinassociated thrombocytopenia. Almost 40 percent of the patients who develop thrombocytopenia after heparin administration have been reported to develop arterial or venous thromboses composed of platelets and fibrin.44,45 Some of these patients have shown resistance to the anticoagulant effects of heparin, while in other cases, the APTT has remained in the therapeutic range at the time the thrombosis was noted clinically.46,47 Despite prolongation of the APTT, recurrent thromboses develop if heparin is continued, but new thromboses do not develop if heparin is withdrawn.Based on these observations, compounds related to heparin have been developed which are reported to possess greater “antithrombotic” effectiveness while producing less anticoagulation than standard commercial heparin. A number of low molecular weight (LM Wt) derivatives of commercial heparin have been prepared by limited chemical depolymerization, gel filtration, or solvent extraction.48,52 These approaches have been based on the knowledge that prolongation of the APTT by heparin results primarily from the capacity of the higher molecular weight fragments of commercial heparin to accelerate the AT-III-thrombin interaction; secondly, amounts of heparin which almost completely inhibit Xa activity are effective in preventing thrombosis but cause minimal or no prolongation of this test.53 The LMWt fractions which are more homogeneous in size have been prepared by additional, more restrictive gel filtration and further purified based on a high affinity for AT-III or other chemical characteristics.49,51,54–56 More recently, oligosaccharide sequences of defined composition and length have been synthesized directly.20,56–58 In addition, there is increased interest in the antithrombotic actions of heparinoids, naturally-occurring mucopolysaccharides which are chemically distinct from heparin, but which show anticoagulant activity in vitro.28,30,56,58–64Over the past several years, the anticoagulant and antithrombotic actions of these newer preparations have been tested and compared with unfractionated heparin. Antithrombotic efficacy has generally been assessed in animal models in which mechanical stasis has been induced or the endothelium mechanically removed from isolated vessels which were then infused with a procoagulant material such as heterologous serum, Russell's viper venom, or thromboplastin. Following infusion of the standard heparin, heparin fraction, or analogue, deposition of fibrinogen in the vessel has been measured directly or estimated by accumulation of radio-labeled fibrinogen. In some studies, the effect of each fraction on the bleeding time or loss of 51Cr-labeled erythrocytes from the vessel has been determined. A limited number of clinical studies in humans have been completed as well.The results of a large number of studies using various heparin preparations58,80 demonstrate the following: (1) low molecular weight fractions of heparin (as low as 4000 to 6000 daltons) as well as synthetic oligosaccharides of the appropriate predicted composition prevent thrombosis in experimental animal models to the same extent as unfractionated heparin.36,52,54,58,60,61,65–67 and (2) subcutaneous administration of LMWt preparations appears to be comparable to standard heparin in preventing deep venous thrombosis and pulmonary embolism in patients undergoing hip replacement or abdominal surgery, although the results of direct comparative studies have not been reported;68,69 (3) the LMWt fractions of heparin and heparinoids have been used successfully in some, but not all, patients who developed heparin-induced thrombocytopenia/thrombosis following standard heparin therapy70–77 and as the sole anticoagulant in patients undergoing hemodialysis;78 (4) in general, at the doses which prevent thrombosis, these heparin fractions or heparin analogues demonstrate relatively little antithrombin effect and therefore produce little “measurable” anticoagulation, assessed by prolongation of the APTT;36,52,54,58,60,61,65–67 (5) the relative antithrombotic potency of these heparin fractions is not defined by either their anti-Xa activity alone or by a ratio of anti-Xa/anti-thrombin activity;35,52,58,60,61,70 and (6) in a limited number of studies, LMWt heparin has been reported to cause less prolongation of the bleeding time and less blood loss in experimental animals than equivalent concentrations of unfractioned heparin.36,50,58,61Taken as a whole, these studies suggest that the LMWt heparin fractions and certain synthetic mucopolysaccharides may demonstrate a greater therapeutic:risk ratio than standard heparin when used in vivo. Therefore, to the extent that the risk of heparin-induced bleeding can be assessed by prolongation of the APTT, it would appear that higher therapeutic drug concentrations might be achieved with these newer preparations. In addition, several of the LMWt fractions show greater bioavailability than unfractionated heparin following subcutaneous administration which may result in higher blood and tissue levels and permit longer intervals between administration.58,80However, a number of important issues remain to be resolved before these preparations replace standard heparin in routine clinical practice. First, the relationship between the experimental models used in these studies and the pathogenesis of thrombosis in man needs to be more critically analyzed. In these models, vessels are surgically isolated, marked stasis or vascular trauma is induced, and procoagulants of defined and restricted potency are infused into the circulation.36,43,58,60,65,79,80 None of these circumstances is identical to the clinical situations which generally precede the development of thrombosis in man. Second, even in these restricted experimental situations, LMWt heparin has been shown to be of greater benefit than unfractionated heparin only under a small number of circumstances with respect to dose, duration, and the timing of the procoagulant stress. Far more extensive clinical testing will be required to identify the precise clinical situations in which these newer formulations show a meaningful clinical advantage over unfractionated heparin. It will also be necessary to study the effects of these newer formulations on platelet35,39,40,81,82 and endothelial cell function,83,84 lipoprotein lipase activity,85 bone resorption, complement activation, and many of the other biologic systems known to be modulated by unfractionated heparin before their routine clinical use can be advocated.Antithrombotic Effect of Heparin: Role of the VasculatureJacques et al11,86,87 have argued that the prevailing view of heparin as an anticoagulant which acts within the circulation and which is subsequently removed by the reticuloendothelial system (RES) is largely incorrect. They have proposed that the heparin is cleared initially from plasma by the endothelium as well as by tissue macrophages. Following chemical modification within the cell, a more active heparin species is slowly secreted. This modified heparin exerts its antithrombotic action primarily by increasing the electronegative potential of the endothelial cell surface, thereby inhibiting the adhesion of platelets and coagulation factors. According to this model, the anticoagulant effects of heparin which we currently monitor for potency and therapeutic efficacy are peripheral to its mechanism of action and are a limiting side effect of heparin fragments which are themselves of limited therapeutic potency.Jacques et al have offered experimental evidence in support of this hypothesis. Following inhalation of radiolabeled heparin via an intratracheal or intrapulmonary cannula, only trace amounts of heparin are found in the plasma of experimental animals. Within the first two hours, approximately half the drug leaves the pulmonary tissues. The radioactive label is found predominantly in tissue macrophages and endothelial cells in organs remote from the lung, where it has been estimated to be concentrated over 100-fold above the plasma level.88,89 Radioactivity (? heparin) continues to be released from these tissue stores for periods up to several weeks, and moderate anticoagulant activity remains in the animal's plasma during this period.88,89 The authors also found that the heparin released from these vascular sites displays a greater specific antithrombotic activity per milligram than the unfractionated heparin used initially to inject the animal.90 They found that when rabbits were injected subcutaneously with heparin which had been largely inactivated by prior acidification, these animals continued to release heparin with anticoagulant activity into their plasma.88,90 The authors reasoned that the injected heparin might have exchanged with more active heparin fragments synthesized by the endothelium or may have undergone a direct metabolic conversion to a more active form within these cells.90 However, the authors did not exclude the possibility that heparin fragments which had less anticoagulant activity were more rapidly removed from the circulation and were catabolized within the RES.In addition to these in vivo studies, there is increasing experimental evidence that the interaction of heparin with the endothelium may play a role in its antithrombotic actions. It is known that endothelial cells synthesize and secrete heparan sulfate, which is chemically and functionally related to heparin.91–95 Heparan sulfate is found both on the luminal surface of the cell and becomes incorporated within the subendothelial basement membrane as well. There is increasing interest in the function of this heparan sulfate which invests the endothelium. During neoangiogenesis following vascular injury, the orderly migration and orientation of endothelial cells and the subendothelial components of the vessel wall may be regulated in part by the composition of the vascular matrix.96 In addition, heparan sulfate has been reported to inhibit the proliferation97 and to modify various synthetic functions of smooth muscle cells.98,99 On the luminal surface, heparan sulfate is a major component of certain anionic microdomains which appear to be involved in the regulation of vascular permeability and protein transport;100 heparin added exogenously can restore the presence and function of these domains which have been lost during experimental injury.101 In addition, two groups of investigators have demonstrated that the endothelium in situ accelerates the capacity of AT-III to neutralize thrombin93–95 and that this vascular cofactor activity is removed by the addition of heparinase.94,95 Finally, it is also possible that the heparan sulfate on the luminal surface may bind inhibitors such as platelet factor 4102,103 histadine-rich glycoprotein,104 and cellular heparinases.105,106Endothelial cells also respond to exogenously administered heparin. Heparin binds to cultured human endothelial cells in vitro in a saturable manner.83,88,107 Both commercial and naturally-occurring heparin promote endothelial cell growth108–110 and activate membrane-bound lipoprotein lipase, an effect which can be mediated by heparin fractions having little affinity for AT-III.85 There is limited indirect evidence that this endothelial cell-bound heparin increases the electronegativity of the vessel wall,11,86,87 but the relative importance of this effect in preventing thrombosis is entirely unclear. Moreover, the disposition of commercial heparin within the endothelium has not been studied, and it is not known whether exogenous heparin causes release of heparan sulfate from the endothelial cell surface (thereby enhancing AT-III activity or removing PF4, etc) as has been reported when cells derived from certain nonvascular tissues have been studied.111The reported effect of heparin on the fibrinolytic system may also result from an interaction of heparin with endothelial cells. Endothelial cells synthesize urokinase-type (u-PA) and tissue-type plasminogen activators (t-PA), as well as fast-acting and latent inhibitors of t-PA and u-PA.112–115 Unfractionated and low molecular weight heparin and certain heparinoids have all been reported to shorten the euglobulin lysis time.84,116 It does not appear that this increase in fibrinolytic activity can be explained by a direct effect of heparin.84,116 Rather, the increased fibrinolytic activity probably results from a release of t-PA from the endothelium or other cellular pools into the serum beginning several hours to days after heparin administration. In one study, heparin failed to reduce fibrin accumulation following a thrombotic stimulus when the fibrinolytic system of the animal had been inhibited with epsilon amino caproic acid or aprotinin.117 However, the relative importance of t-PA and enhanced fibrinolysis in the antithrombotic effect of heparin in man is far from clear. Nor is there evidence that the various heparin fractions or heparinoids differ in their effects on the fibrinolytic system.84,116–120 However, in view of recent reports showing an association between decreased serum fibrinolytic activity and the development of idiopathic venous thrombosis,121–123 the possibility that heparin acts on the endothelial cell surface to release t-PA or to reduce the level of t-PA-inhibitor merits further study.Endothelial cells also produce a number of other factors which influence hemostasis including synthesis of tissue factor, factors VIII antigen (von Willibrand's factor), and factor V. Endothelial cells express specific binding sites in vitro for factors XI, IX, IXa, Xa and possibly fibrinogen and fibrin.124 Endothelial cells synthesize proteins, prostacyclin, thrombospondin, and thrombomodulin, and secrete collagenous and noncollagenous components of the subendothelial matrix. Heparin or fractions of heparin could influence one or more of these functions, resulting in inhibition or acceleration of thrombosis. Although these possibilities remain theoretic, it is clear that heparin affects vascular tissue in ways that cannot be evaluated by the APTT which is the means by which antithrombotic activity of heparin as well as the risk of hemorrhage following heparin therapy is currently assessed.Conclusions and Future ConsiderationsTo date, there are no convincing data which refute the prevailing concept that heparin prevents thrombosis primarily through its anticoagulant activities mediated by AT-III. Nor is there convincing evidence that heparin or heparin fractions prevent thrombosis primarily by directly affecting the function of endothelial cells or other components of the vessel wall. Administration of sufficient heparin to maintain the APTT at 1½ to 2½ times the normal range reduces the risk of recurrent thrombosis to approximately 1 percent and minimizes the risk of bleeding in otherwise healthy individuals.31 To date, neither LMWt heparin nor heparin analogues have been shown to further reduce this risk of thrombosis or bleeding compared with unfractionated heparin.To develop safer and more rational drugs, the causes of idiopathic arterial and venous thromboses need to be more thoroughly elucidated. Currently, neither abnormalities of coagulation proteins, platelets, or the vessel itself have been identified in most of these patients. The recent reports of a reduction in t-PA and/or an increase in t-PA-inhibitor in many such patients raises the possibility that dysfunction of endothelial cells or other vascular cells could play a role in these disorders. Whether heparin analogues or compounds related to t-PA, which act preferentially at the site of thromboses, will further reduce the risk of recurrent thrombosis and further reduce the risk of bleeding remains to be determined.A complete reference list is available upon request from the author. Over the past 30 years, heparin has become the drug most commonly used to treat a variety of acute thrombotic disorders. Heparin reduces the incidence of recurrence while causing relatively few adverse reactions. However, the protection afforded by heparin against recurrent thrombosis is not complete, and the amount of the drug wh