Partial Synthetic Glucan Sulfates as Potential New Antithrombotics: A

The correlation between DS and MW of the glucan sulfates, respectively, and their ..... are more avid ligands for TGF-β1 and more potent antiprolifer...
0 downloads 0 Views 104KB Size
Biomacromolecules 2001, 2, 354-361

354

Partial Synthetic Glucan Sulfates as Potential New Antithrombotics: A Review Susanne Alban* and Gerhard Franz Institute of Pharmacy, University of Regensburg, Universita¨tsstrasse 31, 93040 Regensburg, Germany Received February 8, 2001

Structurally defined sulfated polysaccharides were produced by partial synthesis to develop new antithrombotics as potential heparin alternatives. Glucans of different natural origins were used as starting polymers. The resulting glucan sulfates display pronounced anticoagulant effects; some of them are as active as heparin. According to studies on the structure-activity relationships, besides the molecular weight (MW) and the degree of sulfation (DS), the sulfation pattern and the polysaccharide basic structure are crucial parameters for their anticoagulant potency. Their mode of action differs from that of heparin. Depending on their individual structure, they specifically interfere with various stages of the coagulation process. In vivo, they partly exhibit antithrombotic activity similar to that of heparin. But the in vivo efficacy is not just based on their anticoagulant activity. Their profibrinolytic actions and their strong TFPI-releasing effect may considerably contribute to this overall effect. Due to their manifold interactions with the system of hemostasis, each glucan sulfate shows a structure-dependent, individual action profile. From the investigated glucan sulfates, mainly C2- and C4-sulfated, linear β-1,3-glucan sulfates with DS > 1.0 and MW between 18 and 50 kDa proved to be most suitable for a potential use as heparin alternatives. The results of this study demonstrate the impact of the various structural parameters on the antithrombotic activity of sulfated polysaccharides. However, the biological actions of sulfated polysaccharides are not limited to hemostasis, but they also show manifold modulating effects on other biological systems. Therefore, the approach of using highly sophisticated carbohydrate drug design might be a possibility to obtain new drugs with specific action profiles. 1. Introduction For more than 60 years, the glycosaminoglycan heparin is the drug of choice in the prevention and treatment of thromboembolic disorders. However, there are some welldocumented problems related to its clinical application; e.g., it is ineffective in treating in antithrombin- (AT-) deficient patients, can cause bleeding complications, and can lead to heparin-induced thrombocytopenia.1 Moreover, heparin shows variations in its structural parameters and consequently also in its physiological activities.2 The isolation of heparin from animal tissues may result in severe contamination, e.g., with prions inducing BSE. Further, due to the continuous worldwide increase in the use of heparin, the natural sources may not be sufficient to meet the need in future. Therefore, an important field of research is the development of effective alternatives to substitute for heparin. One possibility is sulfated polysaccharides chemically related to heparin. Another reason that attracts attention to these compounds is the increasing knowledge about their multiple biological activities. Besides their anticoagulant and antithrombotic activities,3 antiatherosclerotic,4 antiproliferative,5 antiadhesive,6 antiangiogenic,7,8 antiinflammatory,9,10 anticomplementary,11 and antiviral effects12 have been described. At * Corresponding author. Telephone: ++49-941-943 4792. Fax: ++49941-943 4762. E-mail: [email protected].

present, there is also a special interest in the question whether and to what extent these additional actions support the therapeutic benefit of heparin. The prolonged survival time of tumor patients treated with heparin is discussed to be correlated with its antiproliferative, antiangiogenic, and antimetastatic effects.7 Furthermore, the anticomplementary and antiinflammatory potential of heparin may be of advantage during extracorporeal circulation.13 Consequently, sulfated polysaccharides having a similar or even improved action profile represent a promising field of research. These compounds are widespread in nature. They occur as components of the extracellular matrix and on the cell surface of vertebrates and are produced by several marine organisms. However, most natural sulfated polysaccharides are complex polydisperse mixtures of macromolecules showing wide variations in their structure and their biological activities. This complicates any drug development utilizing such naturally occurring polymers. Structurally defined sulfated polysaccharides obtained by partial synthesis may be more suitable for studying structure-activity relationships, to establish the optimal structures for heparin alternatives. 2. Starting Polymers for Sulfation An important prerequisite for the partial synthesis of sulfated polysaccharides is the choice of appropriate starting polysaccharides. They should be readily available, inexpensive, and chemically well-defined.

10.1021/bm010032u CCC: $20.00 © 2001 American Chemical Society Published on Web 04/10/2001

Partial Synthetic Glucan Sulfates

Biomacromolecules, Vol. 2, No. 2, 2001 355

Figure 1. Natural glucans used as starting polysaccharides for the partial synthesis of anticoagulant glucan sulfates.

As one example, the β-1,3-glucan curdlan (Figure 1) meets all these requirements. This long-chain polymer is a bacterial exopolysaccharide originally discovered from the soil bacterium Alcaligenes faecalis var. myxogenes.14 By fermentation methods, large amounts are produced for the food industry and for other industrial purposes.14 Besides curdlan, we have chosen the short-chain (5.6. kDa) brown algae β-1,3glucan laminarin (Figure 1). This reserve polysaccharide amounts up to 50% of the algae dry weight. The linear, neutral R-1,4/1,6-glucan pullulan (Figure 1) from the black yeast Aureobasisidum pullulans was included in the studies to evaluate the influence of the type of the glycosidic bonds on the biological activities. Like curdlan, this exopolysaccharide represents an important polymer for manifold applications in the food, cosmetic, and pharmaceutical industries.15 For comparing the effects of linear sulfated polysaccharides with those of branched derivatives, curdlan was synthetically branched in position 6 with glucose, gentiobiose, rhamnose, and arabinose, respectively.3,16 Furthermore, two well-known naturally branched glucans were used as starting polymers: the long-chain bacterium polysaccharide xanthan (Xanthomonas campestris) (Figure 1), whose β-1,4glucan main chain is substituted by a trisaccharide unit (consisting of R-D-mannose-6-acetate, β-D-glucuronic acid and β-D-mannose-4,6-pyruvate ketal) on each second glucose unit,17,18 and further, Glyloid 3S (Dainippon Pharmaceutical, Osaka, Japan), a short-chain highly branched xyloglucan (Figure 1) isolated from tamarind seed.19 The molecular weight (MW) of the produced glucan sulfates was determined as hydrodynamic volume using gel permeation chromatography (GPC). It was dependent not only on the chain length of the starting polymer, but also on the conditions of the sulfation reaction. In addition, it was varied by thermal20 or ultrasonic21 degradation of the starting

polymer. Polydisperse products were fractionated by GPC to obtain compounds with a narrow MW range. 3. Synthesis of Glucan Sulfates The sulfation of the starting glucans was typically carried out with the SO3-pyridine complex dissolved in DMF. This complex was added stepwise to the polysaccharide dissolved or finely suspended in DMF containing pyridine equimolar to the sulfation reagent. This procedure based on the method described by Larm et al.22 had been modified and optimized resulting in reproducible, highly sulfated, nondegraded polysaccharide derivatives.23 The degree of sulfation (DS, i.e., sulfate groups per glucose monomer) of the glucan sulfates proved to be dependent on the temperature, the reaction time and the molar excess of the sulfation reagent.16,23,24 A crucial point is the pretreatment, i.e., the activation, of the polysaccharides such as the formation of the sodium complex of curdlan. Further, the basic polysaccharide structure plays an important role for the resulting DS. For instance, an identical sulfation procedure results in laminarin sulfates with higher DS compared to corresponding derivatives of the long-chain curdlan. The β-1,3-glucans turned out to be more accessible to sulfation than pullulan which has less free primary OH groups due to its proportion of R-1,6-glycosidic bonds. By applying this sulfation method, mainly the primary OH group in position 6 of the glucose substituted. To examine if apart from the DS also the position of the sulfate groups within the glucose units influences the biological activity, β-1,3-glucan sulfates with different sulfation patterns were produced. To reduce the predominant sulfation of the primary in favor of the secondary OH groups in positions 2 and 4, position 6 was protected prior to the sulfation. Comparing various protection groups,16 the large adamantoyl residue was

356

Biomacromolecules, Vol. 2, No. 2, 2001

Alban and Franz

5. Structure-Activity Relationships

Figure 2. Simplified classic scheme of the coagulation cascade with the coagulation tests recording different stages within this process (to avoid complexity, the positive feedback loops, e.g. activation of FXI, cofactors V and VIII by thrombin, and the interactions between the classically differentiated intrinsic and extrinsic pathway are not shown; for details see a review65).

found to be the most suitable25 resulting in products presenting up to 97% of the sulfate groups in positions 2 and 4.19 4. Anticoagulant Activity of Glucan Sulfates For examining the anticoagulant activity of the new glucan sulfates, they were tested in the classical coagulation tests26 prothrombin time (PT), activated partial thromboplastine time (APTT), and thrombin time (TT) and an anti-factor Xa clotting test (Heptest). Unfractionated heparin was used as reference, and the fourth International Standard for Heparin (4th I.St) was used for the determination of the specific anticoagulant activity (U/mg). Since the various coagulation assays record interactions with different stages of the coagulation process (Figure 2), they provide basic information about the mode of action of anticoagulants.27 The PT determines interference with the extrinsic- and the APTT with the intrinsic coagulation pathway. The TT is the test for the last coagulation step, the thrombin-mediated fibrin formation. The Heptest, especially developed for heparin, measures the inhibition of factor Xa. In contrast to the inactive nonsulfated glucans, the sulfated derivatives exhibit anticoagulant effects in all of these assays. Their respective activities range from less than 1% up to 135% of the heparin activity3,16,24 depending on their individual structure. In the APTT and thrombin time, the glucan sulfates display the highest activity. Heparin mainly inhibits blood coagulation by catalyzing the factor Xa and thrombin neutralization by the endogenous coagulation inhibitor AT.28 In the Heptest, the glucan sulfates reach only 10% of the heparin activity. Consequently, with regard to factor Xa, a heparin-like mechanism can be excluded for the glucan sulfates. Their still slight activity in the Heptest is due to lack of specificity of this assay. The Heptest does not exclusively measure the anti-factor Xa activity but also the anti-thrombin activity and influences on the intrinsic coagulation to some extent.29 Like heparin,30 the glucan sulfates exhibit only modest PT activity indicating that the extrinsic part of coagulation is not an important site of action for these glucan derivatives.

Extensive studies on the structure-activity relationships revealed that the anticoagulant activity of the partial synthetic glucan sulfates depends on several structural parameters. A certain minimum charge density as well as a certain minimum chain length are essential to any anticoagulant effect. These threshold values are conditional on each other and on the basic structure of the respective polysaccharides. Linear, short-chain β-1,3-glucan sulfates, i.e., laminarin sulfates, require DS higher than 0.7 for any anticoagulant activity.23 In contrast to this, linear β-1,3-glucan sulfates with high MW are active with a DS as low as 0.5. As an example, a curdlan sulfate with a DS of 0.54 and a MW of 190 kDa had 28% of the heparin activity in the APTT and TT. A DS lower than 0.7 is also sufficient for activity in branched, short-chain β-1,3-glucan sulfates.16 Accordingly, an arabinose-branched β-1,3-glucan sulfate with a DS of 0.66 and a MW of 15 kDa exhibited 11% of the heparin activity in the APTT and 6% in the TT. The pullulan sulfates require still lower DS for activity.31 Here, a compound with a DS of 0.17 and a MW of 48 kDa had 5% of the heparin activity in the APTT and 8% in the TT. As a general rule, the activity in all the coagulation tests improves with increasing DS of the glucan sulfates.3,23,24,31 A second decisive parameter for the anticoagulant activity is the chain length of the polymer.20 The higher the average MW, the better is the anticoagulant effect. Above approximately 150 kDa, no further activity increase was observed. Especially the thrombin time activity is strongly dependent on this parameter: The minimum MW for any effect in the TT is higher than that for any APTT activity, but the TT activity of compounds with very high MW may even exceed their APTT activity. The APTT activity of a sulfated β-1,3-glucose-octasaccharide with a DS of 1.05 amounted to 20% of the heparin activity, whereas it was completely inactive in the TT assay. In contrast, a curdlan sulfate with the same DS, but a MW of about 250 kDa, had 135% of the heparin in both, TT and APTT assay. As a consequence, the ratio of the TT to the APTT activity, which is 1.0 for heparin, increases with increasing MW of the glucan sulfates. The correlation between DS and MW of the glucan sulfates, respectively, and their anticoagulant activity corresponds to findings reported for other sulfated polysaccharides including fucoidans,32,33 sulfated dextran derivatives,34 and dermatan sulfates.35 In addition to the well-documented DS and MW dependence of the anticoagulant activity of sulfated polysaccharides, the sulfation pattern is important for the activity. This has been demonstrated using “selectively” sulfated β-1,3glucan sulfates.36 Recently similar results were reported for fucoidans and dermatan sulfates.37,38 The distribution of the sulfate groups on the carbon atoms of the glucose units markedly influences the activity and moreover also influences the action profile. Comparing β-1,3-glucan sulfates with identical DS and MW, the anticoagulant efficacy improves with increasing percentage of sulfate groups in position 2 and 4.36 Thus, by modifying the sulfation pattern, activities are reached, which normally require higher DS and MW.

Partial Synthetic Glucan Sulfates

Biomacromolecules, Vol. 2, No. 2, 2001 357

Figure 4. Comparison between linear and branched glucan sulfates regarding their the APTT (9) and TT activities (0) [U/mg] (measured by means of the 4th I.St., unfractionated heparin had an APTT and TT activity of 150 U/mg): (A) long-chain glucan sulfates, curdlan sulfate (CurS) and xanthan sulfate (XanS); (B) short-chain glucan sulfates, laminarin sulfate (LamS) and xyloglucan sulfate (XyglS).

Figure 3. Comparison between mainly C6- (GS-6) and mainly C2and C4-sulfated β-1,3-glucan (GS-2,4) sulfates regarding their the APTT (9), TT (shaded box), and Heptest activities (0). Depending on their DS and MW, the compounds are divided in four groups. To demonstrate the changes of the action profile, the original activities [U/mg] are given as values relative to the APTT activity of the respective GS-6, which is defined as 1.

As shown in Figure 3, the increase in activity is strongly dependent on the DS and the MW. β-1,3-glucan sulfates with a low DS and a MW > 20 kDa profit most from a high percentage of C2- and C4-sulfate groups. Moreover, the DS and the MW determine the extent of activity increase in the various assays. Therefore, the β-1,3-glucan sulfates show different action profiles depending on their DS, MW, and sulfation pattern. Whereas the mainly C6-sulfated compounds with low DS or low MW show no significant effects in the TT assay, those with a predominant C2- and C4-sulfation are active in this assay. In compounds with low DS as well as low MW, the APTT and Heptest activities increase by C2- and C4-sulfation without any significant change of the TT activity. Furthermore, branched β-1,3-glucan sulfates are more potent anticoagulants than linear compounds with the same DS and MW.16,24 The TT activity is generally higher, whereas the change of the APTT activity is dependent on the type of the carbohydrate side chain (e.g., mono- or disaccharide, pentose or hexose, or desoxysugar) as well as the degree of branching (DB). Since glycosidic substitutions enhance the flexibility of the polymer chain, this increase in activity may be based on an improved interaction with the coagulation enzymes.39 The general validity of these findings was proven with naturally branched glucan sulfates. A xanthan sulfate possessing trisaccharide side chains was slightly less active in the APTT, but it was nearly 4 times more potent in the TT than a corresponding linear glucan sulfate (Figure 4). Similar modifications of the activities were observed with β-1,3-glucan sulfates branched by the disaccharide gentiobiose. The slightly reduced APTT and the about 5-fold increased TT activities of highly branched (DB ≈ 0.75) xyloglucan sulfates correspond to the findings with β-1,3glucan sulfates with DB > 0.70.

Figure 5. Comparison between β-1,3- and R-1,4/1,6-glucan sulfates with identical DS and MW each regarding their the APTT (9) and TT activities (0) [U/mg] (measured by means of the 4th I.St.; unfractionated heparin had an APTT and TT activity of 150 U/mg).

The activities of the branched glucan sulfates suggest that in addition to charge density, polymer size, and substitution pattern, the chain flexibility and the three-dimensional structure are essential determinants for any interactions between sulfated polysaccharides and coagulation enzymes. Accordingly, the type of the glycosidic bonds of the polysaccharide main chain, which determine its conformation, also seems to influence the activity. As shown in Figure 5, pullulan sulfates are superior to curdlan sulfates in their APTT and TT activities. Moreover, the ratio of their TT to their APTT activities is higher than that of the activities of curdlan sulfates.31 These data suggest that the R-1,4-/1,6glucan structure interferes to a greater extent with the thrombin-mediated fibrin formation than does the β-1,3glucan structure. The established structure-activity relationships demonstrate that the anticoagulant activity of the partial synthetic glucan sulfates changes both quantitatively and qualitatively depending on structure. It can be concluded that these new heparinoids do not inhibit the coagulation in a nonspecific

358

Biomacromolecules, Vol. 2, No. 2, 2001

Alban and Franz

way due to their negative charges, but that they specifically interfere with distinct stages of the coagulation cascade, whereby these interactions obviously have different requirements on the chemical structure. 6. Mode of Action Investigation of the mode of action of the new partial synthetic glucan sulfates revealed that their anticoagulant mechanisms considerably differ from that of heparin. In amidolytic assays with purified enzymes and chromogenic substrates, the glucan sulfates show neither a significant direct nor AT-mediated inhibition of factor Xa.40 In corresponding anti-thrombin assays, their effects are more pronounced, but the IC50 of the most active β-1,3-glucan sulfate is still about 15 times higher than that of heparin.40 Since their TT activity can be as high as that of heparin, the ATmediated thrombin inhibition may contribute to their anticoagulant activity but is not principally responsible for it. As a practical consequence, these compounds would be also effective in AT-deficient patients. An important site of action of the glucan sulfates is the intrinsic factor Xa generation, i.e., the activation of factor X by the tenase complex consisting of factor IXa, cofactor VIIIa, phospholipids, and Ca2+ (see Figure 2).40 The inhibition of this step may explain the high APTT activity of the glucan sulfates. Similar to the APTT activity, the inhibitory potency is more dependent on the DS than on the MW. As an example, highly sulfated, short-chain β-1,3-glucan sulfates having high APTT but negligible TT activity, are also efficient inhibitors of the intrinsic factor Xa generation. These findings are similar to the anticoagulant action profile of lactobionic acid derivatives, which represent highly sulfated synthetic carbohydrate derivatives with MW smaller than 3 kDa.41 The sulfated lactobionic acids also exhibit high activity in the APTT and the intrinsic factor Xa generation assay, but show only moderate effects in the TT assay. Further studies on the mechanism of the inhibitory effect of the glucan sulfates on of the intrinsic factor Xa generation showed that these compounds inhibit cofactor VIIIa or its activation by thrombin, respectively, and consequently prevent the formation of the FX activating tenase complex.24 The potency of the glucan sulfates was considerably weaker in the assay for the extrinsic factor Xa generation, i.e., the activation of factor X by the tissue factor/factor VIIa complex. The observed inhibition is attributed to the so-called “Josso loop”, i.e., the connection between the extrinsic and intrinsic coagulation system,42 rather than to a direct interference with the extrinsic factor Xa generation. Factor Xa performs its physiological effect by forming a complex with cofactor Va, phospholipids, and Ca2+.43 This so-called prothrombinase transforms prothrombin into active thrombin. The glucan sulfates only moderately inhibit the prothrombinase activity.40 Nevertheless, these effects are still stronger than those for free factor Xa. All these actions of the glucan sulfates result in a reduced thrombin generation, which represents the key process in the thrombus formation. But also the action of thrombin itself is antagonized44 as reflected by the TT activity. The structural requirements for a pronounced TT activity differ from those

Figure 6. Sites of action of the partial synthetic glucan sulfates on the thrombin-mediated fibrin formation. Depending on the individual structural parameters of the glucan sulfates, the various mechanisms contribute more or less to the inhibition of this last step of the coagulation cascade. The thrombin action (bold arrows) is differentiated in the amidolysis of a small chromogenic substrate (in vitro) and the cleavage of its physiological substrate fibrinogen resulting in fibrin formation (in vitro and in vivo). Antithrombin (AT) and heparin cofactor II (HCII) inhibit thrombin. The actions of the glucan sulfates are shown by double lines: (f) activation; (∼) binding: (-|) inhibition. Glucan sulfates do not directly impair the amidolysis of small chromogenic substrates.

for the APTT activity. A certain minimum three-dimensional size of the molecule appears to be crucial for any activity. This can be realized by means of a long chain or by glycosidic branched molecules. The additional pronounced influence of the sulfation pattern and the type of the glycosidic bonds argues for several mechanisms contributing to the inhibition of the last step of the coagulation cascade by the glucan sulfates (Figure 6). An important role is played by heparin cofactor II (HCII).45 Similar to AT, HCII is an endogenous thrombin inhibitor which is catalyzed by sulfated polysaccharides. The anticoagulant activity of the glycosaminoglycan dermatan sulfate is typically associated with the HCII-mediated thrombin inhibition.45 The higher the MW and the DS of the glucan sulfates the more they improve the HCII-mediated thrombin inhibition. Long-chain, high-sulfated xanthan sulfates are the most active agents, superior even to the reference compound dermatan sulfate.24 Their IC50 is about 1000 times lower than that of dermatan sulfate and they accelerate the HCIImediated thrombin inhibition more than 10 000-fold. As already mentioned, the glucan sulfates only marginally intensify the AT-mediated thrombin inhibition. This is expected, since they do not contain the AT-binding site, the heparin-specific pentasaccharide sequence. However, there is one exception among the glucan sulfates: the pullulan sulfates are as active as heparin in a purified AT-thrombin containing test system and exhibit considerably lower TT activity in AT-depleted plasma.24 Up to now, no comparable compound is known. According to binding studies, the pullulan sulfates bind to neither AT nor thrombin alone, but they appear to act through a “template” mechanism as has been proposed for the AT-mediated thrombin inactivation by heparin.46,47

Partial Synthetic Glucan Sulfates

Biomacromolecules, Vol. 2, No. 2, 2001 359

The glucan sulfates do not directly inhibit the amidolysis of a chromogenic substrate by thrombin,40 suggesting that they are unable to inactivate thrombin by interference with its active center as do specific thrombin inhibitors such as argatroban.48 Nevertheless, they impair the action of thrombin on the large substrate fibrinogen.40,49 Only mainly C2- and C4-sulfated β-1,3-glucan sulfates are able to bind directly to thrombin and to inhibit it in this way.24 However, above a certain DS and MW, all types of glucan sulfates form complexes with fibrinogen.49 This probably results in a steric inhibition of fibrin polymerization as has been described for fucoidans.50 Since all the outlined mechanisms more or less contribute to the anticoagulant action of the glucan sulfates depending on their individual structures, it might be possible to produce heparinoids with specific action profiles. The activiation of the HCII-mediated thrombin inhibition represents a major advantage of these compounds. The possibility of overdose is unlikely due to the limited physiological concentration of HCII. Consequently, their bleeding risk should be lower than that of compounds, that inhibit thrombin directly or through an AT-mediated mechanism (personal communication by Prof. H. C. Hemker, Maastricht). 7. Antithrombotic Activity The glucan sulfates were tested in several established animal thrombosis models to determine whether these in vitro anticoagulants also exhibit in vivo antithrombotic activity. They proved to be more or less efficacious depending on their structure. A curdlan sulfate (DS of 0.83, MW of 49 kDa) having 75% of its sulfate groups in positions 2 and 4 was as active as heparin when a rat model of clamping induced jugular vein occlusion was used to produce vascular obstruction.51 However, in vitro it was about 3-4 times less active than heparin. This demonstrates that the in vitro anticoagulant effects do not always correspond to the in vivo potency of antithrombotic agents as additional mechanisms may contribute to their action. The mainly C2- and C6-sulfated β-1,3-glucan sulfates were shown to inhibit the thrombin-induced platelet aggregation in addition to plasma coagulation.52 Moreover, the glucan sulfates are able to promote the fibrinolysis in vitro, the antagonist of the coagulation.53 Profibrinolytic activities contribute to the antithrombotic actions of many heparinoids, e.g., dermatan sulfate, pentosan polysulfate or fucoidans.54-56 As with coagulation, several types of structure-dependent interactions between the glucan sulfates and this complex system were observed.57 They activate the contact system, which stimulates in vivo fibrinolysis.58 Further, they advance the tissue-type plasminogen activator- (t-PA-) mediated plasminogen activation and intensify the amidolytic activity of plasmin, the key enzyme of fibrinolysis. However, they do not degrade fibrin or directly activate plasminogen, but instead they require t-PA for developing any activity. Apart from these in vitro phenomena, the glucan sulfates also release t-PA from the vessel wall similar to heparin59 (see next section). This action probably contributes to the overall antithrombotic efficacy of heparin.60

Figure 7. Comparison of the four glucan sulfates XanS (xanthan sulfate) (9), CurS (curdlan sulfate) (light shaded box), LamS (laminarin sulfate) (heavy shaded box), and PulS (pullulan sulfate) (0) (see Table 1 for the structural parameters) regarding their anticoagulant activities in the APTT and thrombin time, their profibrinolytic in the fibrin plate method and their t-PA releasing effects in the isolated perfused pig ear model.

A second action of the glucan sulfates, which could only be demonstrated in vivo, is the strong release of tissue factor pathway inhibitor (TFPI) from the vessel wall. TFPI is an endogenous serine protease inhibitor, that inhibits the tissue factor-mediated coagulation.61 The mobilization of TFPI by heparin after intravenous and subcutaneous injection is thought to be involved in its antithrombotic action.62 A mainly C2- and C4-sulfated laminarin sulfate was superior to heparin in an animal experiment in monkeys. This supports the conclusion that the TFPI releasing effect could contribute to the important antithrombotic activity associated with these β-1,3-glucan sulfates. 8. Comparison between the Anticoagulant Activities in Vitro, the Profibrinolytic Actions in Vitro, and the t-PA Releasing Effects in Vivo The structure-dependent action profile of the glucan sulfates is not limited to their anticoagulant effects, but was also found for other biological activities. This is illustrated by the comparison of four glucan sulfates with regard to their anticoagulant activity in the APTT and TT, their profibrinolytic effect in vitro, and their t-PA-releasing potency (Figure 7). The structural parameters of the four compounds are listed in Table 1. Their anticoagulant activities reflect the complex structureactivity relationships that were discussed in detail in section 5. Due to its comparatively low DS, the pullulan sulfate has the lowest anticoagulant activity. The APTT and TT activities of the long-chain curdlan sulfate and the short-chain laminarin sulfate confirm that the activity improves with increasing MW, although the laminarin sulfate has a higher DS. Due to its high DB, the xanthan sulfate is as active as heparin

360

Biomacromolecules, Vol. 2, No. 2, 2001

Alban and Franz

Table 1. Structural Parameters for Glucan Sulfates Obtained by Sulfation of Xanthan, Curdlan, Laminarin, and Pullulana glucan sulfate xanthan sulfate curdlan sulfate laminarin sulfate pullulan sulfate

basic polysaccharide structure DSb MW,c kDa branched β-1,4-glucan β-1,3-glucan β-1,3-glucan R-1,4/1,6-glucan

1.60 1.30 1.60 0.50

160 120 17 95

a Sulfation was performed with an excess of 6 mol of SO -pyridine/ 3 mol of glucose; for details, see refs 16, 21, 23, and 24. b DS ) sulfate groups per monosaccharide unit. c MW ) molecular weight determined as hydrodynamic volume by GPC using neutral pullulans as standards.

in the TT assay, but its APTT activity is only moderate despite of its high DS and MW. A simple screening assay for the profibrinolytic activity is the traditional fibrin plate method:63 A Petri dish is coated with fibrin by adding thrombin to a solution of human fibrinogen. Aliquots of the dissolved compounds are placed on the plate and the areas of the resulting lysis zones are measured. If highly purified or heat-inactivated fibrinogen is used, the glucan sulfates do not induce any lysis. If the fibrinogen contains some plasminogen and traces of t-PA, they exhibit fibrinolytic activity, whereas heparin is inactive. In contrast to the anticoagulant activity, the effect is independent of the MW, but strongly dependent on the DS. This is shown by the differences between the curdlan and the laminarin sulfate as well as by the low pullulan sulfate activity. In this assay, the basic structure again plays an important role. Derivatives with β-1,3-structure display especially potent effects, whereas the branched xanthan sulfates are not very active. The t-PA releasing potency was tested by means of the isolated perfused pig ear model.64 The ramus intermedius of the arteria auricularis magna of an ear of a freshly slaughtered pig is perfused with tyrode buffer. After 30 min of rinsing, the perfusate is collected in 2 min fractions. After four fractions are obtained for the baseline, the ear is perfused for 4 min with buffer containing the test compound and then again for 12 min with buffer. The t-PA concentration in the fractions is determined in relation to the baseline. Comparing the four glucan sulfates, the activity continuously decreases with decreasing MW independently of the DS and the basic structure. Even the low-sulfated pullulan sulfate is more active than the high-sulfated LamS. The changing order of the presented activities of four glucan sulfates demonstrates the differences in the structureactivity relationships for these activities. In summary, the glucan sulfates display structure-dependent, individual action profiles concerning their interference with the numerous processes of the coagulation and fibrinolysis. 9. Conclusions Partial synthetic glucan sulfates with defined chemical structure were developed and first examined in vitro. Thereby, structure-activity relationships could be established. Investigation of the mode of action revealed that these compounds interfere with several stages of the coagulation cascade in a structure-dependent manner. In vivo, some of them exhibit an antithrombotic activity similar to heparin. The in vivo efficacy, however, is not just based on their

anticoagulant activity. Their profibrinolytic actions and their strong TFPI-releasing effect may considerably contribute to this effect. Due to their manifold interactions with the system of hemostasis, each glucan sulfate shows a structuredependent, individual action profile. From the investigated glucan sulfates, mainly C2- and C4-sulfated, linear β-1,3glucan sulfates with DS > 1.0 and MW between 18 and 50 kDa proved to be most suitable for a potential use as heparin alternatives. With respect to the availability and the reliably constant structure of the starting polymer, curdlan seems to be superior to any of the other glucans. In contrast to glycosidically branched β-1,3-glucan sulfates, the synthesis of mainly C2- and C4-sulfated, linear derivatives is feasible and can be standardized. Compounds with this structure proved to be most effective in vivo. In addition, they display a convenient action profile. This includes direct interactions with thrombin, pronounced increase of the HCII-mediated thrombin inhibition, mostly AT-independence, marked profibrinolytic activity and strong TFPI-mobilizing effect. In vivo, they proved to be nontoxic and well-tolerated.59 β-1,3-glucan sulfates with MW > 50 kDa are more active in vitro in most cases, but those with lower MW turned out to have a better bioavailability.24,59 The results of this study demonstrate the importance of the various structural parameters for the antithrombotic activity of sulfated polysaccharides. However, the biological actions of sulfated polysaccharides are not limited to hemostasis, but they show manifold modulating effects on diverse biological systems without being toxic. Therefore, the approach of using highly sophisticated carbohydrate drug design might be a possibility to obtain new drugs with specific action profiles. Acknowledgment. S.A. was the holder of the fellowship “Bayerischer Habilitations-Fo¨rderpreis 1996sHans Zehetmair Preis”, and we are grateful for the financial support. References and Notes (1) Weitz, J. Drugs 1994, 48, 485-497. (2) Linhardt, R. J.; Loganathan, D.; al-Hakim, A.; Wang, H. M.; Walenga, J. M.; Hoppensteadt, D.; Fareed, J. J. Med. Chem. 1990, 33, 16391645. (3) Alban, S. In Carbohydrates in Drug Design; Witczak, Z. J., Nieforth, K. A., Eds.; Marcel Dekker: New York, 1997; pp 209-276. (4) Engelberg, H. Semin. Thromb. Hemost. 1991, 17, 5-8. (5) McCaffrey, T. A.; Falcone, D. J.; Borth, W.; Brayton, C. F.; Weksler, B. B. Biochem. Biophys. Res. Commun. 1992, 184, 773-781. (6) Ley, K.; Cerrito, M.; Arfors, K.-E. Am. J. Physiol. 1991, 260, H16671673. (7) Linhardt, R. J.; Toida, T. In Carbohydrates in Drug Design; Witczak, Z. J., Nieforth, K. A., Eds.; Marcel Dekker: New York, 1997; pp 277-341. (8) Zacharski, L. R.; Ornstein, D. L. Thromb. Haemost. 1998, 17, 289297. (9) Winkelhake, J. L. Glycoconj. J. 1991, 8, 381-386. (10) Arfors, K.-E.; Ley, K. J. Lab. Clin. Med. 1993, 121, 201-202. (11) Boisson-Vidal, C.; Haroun, F.; Ellouali, M.; Blondin, C.; Fischer, A. M.; de Agostini, A.; Jozefonvicz, J. Drugs Future 1995, 20, 12371249. (12) DeClercq, E. Anti-HIV activity of sulfated polysaccharides. In Carbohydrates and Carbohydrate Polymers, Analysis, Biotechnology, Modification, AntiViral, Biomedical and other Application; Yalpani, M., Ed.; ATL Press: Shrewsbury, MA, 1993; pp 87-100. (13) Mollnes, T. E. Vox Sang. 1998, 74 (Suppl. 2), 303-307. (14) Harada T. Trends Glycosci. Glycotechnol. 1992, 4, 309-317.

Partial Synthetic Glucan Sulfates (15) Deshpande, M. S.; Rale, B. V.; Lynch, J. M. Enzyme Microb. Technol. 1992, 14, 514-523. (16) Alban, S. Synthese und physiologische Testung neuartiger Heparinoide. Ph.D. Thesis, University of Regensburg, 1993. (17) Melton, L. D.; Mindt, L.; Rees, D. A.; Sanderson, G. R. Carbohydr. Res. 1976, 46, 245-257. (18) Jansson, P. E.; Kenne, L.; Lindberg, B. Carbohydr. Res. 1975, 45, 275-282. (19) Vincken, J.-P.; York, W. S.; Beldman, G.; Varagen, A. G. J. Plant Physiol. 1997, 114, 9-13. (20) Alban, S.; Franz, G. Pure Appl. Chem. 1994, 66, 2403-2406. (21) Schauerte, A. Partialsynthesische Heparinoide aus Polysacchariden natu¨rlichen Ursprungs. Ph.D. Thesis, University of Regensburg, 1998. (22) Larm, O.; Larsson, K.; Scholander, E.; Andersson, L.-O.; Holmer, E.; So¨derstro¨m, G. Carbohydr. Res. 1979, 73, 332-336. (23) Alban, S.; Kraus, J.; Franz, G. Arzneim.-Forsch./Drug Res. 1992, 42, 1005-1008. (24) Alban, S. Von sulfatierten Polysacchariden zu neuen potentiellen Arzneistoffen. Habilitationsschrift, University of Regensburg, 1999. (25) Alban, S.; Jeske, W.; Hoppensteadt, D.; Fareed, J.; Franz, G. Pharm. Pharmacol. Lett. 1995, 5, 70-73. (26) Alban, S.; Franz, M. E.; Franz, G. Pharm. Pharmacol. Lett. 1997, 7, 59-62. (27) Weissbach, G. In Blutgerinnungsdiagnostik in der Klinik: Eine methodische Anleitung, 1st ed., Weissbach, G., Ed.; Verlag Gesundheit: Berlin, 1991; pp 191-217. (28) Hirsh, J.; Warkentin, T. E.; Raschke, R.; Granger, C.; Ohman, E. M.; Dalen, J. E. Chest 1998, 114 (Suppl.), 489-510. (29) Bara, L.; Mardiguian, J.; Samama, M. Thromb. Res. 1990, 57, 585592. (30) Hemker, H. C.; Beguin, S. Semin. Thromb. Hemost. 1989, 17 (Suppl. 1), 29-34. (31) Alban, S.; Schauerte, A.; Franz, G. Carbohydr. Polym. 2001, in press. (32) Nagumo, T.; Nagumo, T. Carbohydr. Res. 1991, 214, 193-197. (33) Nardella, A.; Chaubet, F.; Boisson-Vidal, C.; Blondin, C.; Durand, P.; Jozefonvicz, J. Carbohydr. Res. 1996, 289, 201-208. (34) de-Raucourt, E.; Mauray, S.; Chaubet, F.; Maiga-Revel, O.; Jozefowicz, M.; Fischer, A. M. J. Biomed. Mater. Res. 1998, 41, 4957. (35) Ferrari, G. P.; Marchesini, D.; Maggi, A. P. Carbohydr. Res. 1994, 255, 125-132. (36) Alban, S.; Franz, G. Semin. Thromb. Hemost. 1994, 20, 152-158. (37) Pereira, M. S.; Mulloy, B.; Mourao, P. A. J. Biol. Chem. 1999, 19, 7656-7667. (38) Mulloy, B.; Mourao, P. A.; Gray, E. J. Biotechnol. 2000, 77, 123135. (39) Uryu, T.; Yoshida, T:, Hatanaka. K.; Kaneko, Y.; Mimura. T.; Nakashima, H.; Yamamoto, N. Polym. Sci. I: Symp. Proc. Polym. 1991, 2, 989-996. (40) Alban, S.; Franz, G. Thromb. Res. 2000, 99, 377-388. (41) Klauser, R. J.; Meinetsberger, E.; Raake, W. Semin. Thromb. Hemost. 1991, 17 Suppl. 1, 118-125. (42) Jesty, J.; Nemerson, Y. In Williams Hematology, 5th ed.; Beutler, E., Lichtman, M. A., Coller, B. S., Kipps, T. J., Eds.; McGraw-Hill: New York, 1995; pp 1227-1238.

Biomacromolecules, Vol. 2, No. 2, 2001 361 (43) Be´guin, S.; Mardiguian. J.; Lindhout. T.; Hemker, H. C. Thromb. Haemost. 1989, 61, 30-34. (44) Alban, S.; Gastpar, R.; Schauerte, A.; Franz, G. Ann. Hematol. 1998, 76 Suppl. I, P264. (45) Tollefsen, D. M. In Heparin - Chemical and Biological Properties, Clinical Applications; Lane, D. A., Lindahl, U., Eds.; CRC Press: Boca Raton, FL, 1989; pp 257-273. (46) Bode, W.; Stubbs, M. T. In Ha¨ mostaseologie: Molekulare und zellula¨ re Mechanismen, Pathophysiologie und Klinik; Mu¨llerBerghaus, G., Po¨tzsch, B., Eds.; Springer-Verlag: Berlin, Heidelberg, Germany, and New York, 1998; pp 269-284. (47) Griffith, M. J. J. Biol. Chem. 1982, 257, 7360-7365. (48) Schro¨r, K. In Ha¨ mostaseologie: Molekulare und zellula¨ re Mechanismen, Pathophysiologie und Klinik; Mu¨ller-Berghaus, G., Po¨tzsch, B. Eds. Springer-Verlag: Berlin, Heidelberg, Germany, and New York, 1998; pp 647-662. (49) Gastpar, R.; Franz, G.; Alban, S. Ann. Hematol. 1998, 76, (Suppl. I), V11 (50) Nishino, T.; Aizu. Y.; Nagumo T. Thromb. Res. 1991, 64, 723-31. (51) Alban, S.; Jeske, W.; Welzel, D.; Franz, G.; Fareed, J. Thromb. Res. 1995, 78, 201-210. (52) Alban, S.; Jeske, W.; Hoppensteadt, D.; Fareed, J.; Franz, G. In NonAnticoagulant Actions of Glycosaminoglycans; Harenberg, J., Casu, B., Eds.; Plenum Publishing Corporation: New York, 1996; pp 235242. (53) Schauerte, A.; Alban, S.; Franz, G. Ann. Hematol. 1997, 74 (Suppl. II), 300. (54) Barbanti, M.; Calanni, F.; Milani, M. R.; Marchi. E.; Semeraro. N.; Colucci M. Thromb. Haemost. 1993, 69, 147-151. (55) Vinazzer, H. Semin. Thromb. Hemost. 1991, 17, 375-378. (56) Soeda, S.; Sakaguchi, S.; Shimeno, H.; Nagamatsu, A. Biochem. Pharmacol. 1992, 43, 1853-1858. (57) Alban, S.; Schauerte, A.; Franz, G. Ann. Hematol. 1997, 74 (Suppl. II), 271. (58) Mannhalter, C. In: Ha¨ mostaseologie: Molekulare und zellula¨ re Mechanismen, Pathophysiologie und Klinik; Mu¨ller-Berghaus, G., Po¨tzsch, B., Eds.; Springer-Verlag: Berlin, Heidelberg, Germany, and New York, 1998; pp 122-131. (59) Markwardt, F., Klo¨cking, H.-P. Haemostasis 1977, 6, 370-374. (60) Haas, S.; Haas, P. In Niedermolekulare Heparine: die Anwendung in Klinik und Praxis; Zett-Verlag: Steinen, Germany, 1996; pp 5051. (61) Alban, S.; Schauerte, A.; Hoppensteadt, D.; Fu. K.; Yang, L.; Iqbal. O.; Fareed, J.; Franz. G. Thromb. Haemost. 1997, Suppl., PS1676. (62) Abildgaard, U. Haemostasis 1993, 23 (Suppl. 1), 103-106. (63) Astrup, T.; Muellertz. S. Arch. Biochem. Biophys. 1952, 40, 346352. (64) Markwardt, F.; Klo¨cking, H.-P. Thromb. Res. 1976, 8, 217-233. (65) Leblond, L.; Winocour P. D. In Handbook of Experimental Pharmacology, Vol. 132 Antithrombotics; Uprichard, A. C. G., Gallagher, K. P., Eds.; Springer-Verlag: Berlin and Heidelberg, Germany, 1999; pp 1-39.

BM010032U