Structure and anticoagulant properties of sulfated glycosaminoglycans from primitive Chordates

Dermatan sulfates and heparin, similar to the mammalian glycosaminoglycans, but with differences in the degree and position of sulfation were previously isolated from the body of the ascidian Styela plicata and Ascidia nigra. These differences produce profound effects on their anticoagulant properties. S. plicata dermatan sulfate composed by 2-O-sulfated α-L-iduronic acid and 4-O-sulfated N-acetyl-β-D-galactosamine residues is a potent anticoagulant due to a high heparin cofactor II activity. Surprisingly, it has a lower potency to prevent thrombus formation on an experimental model and a lower bleeding effect in rats than the mammalian dermatan sulfate. In contrast, A. nigra dermatan sulfate, also enriched in 2-O-sulfated α-Liduronic acid, but in this case sulfated at O-6 of the N-acetyl-β-D-galactosamine units, has no in vitro or in vivo anticoagulant activity, does not prevent thrombus formation but shows a bleeding effect similar to the mammalian glycosaminoglycan. Ascidian heparin, composed by 2-O-sulfated α-L-iduronic acid, Nand 6O-sulfated glucosamine (75%) and α-L-iduronic acid, Nand 6-O-sulfated glucosamine (25%) disaccharide units has an anticoagulant activity 10 times lower than the mammalian heparin, is about 20 times less potent in the inhibition of thrombin by antithrombin, but has the same heparin cofactor II activity as mammalian heparin.


INTRODUCTION
Heparin and dermatan sulfate polymers have recently been isolated from test cells and tissues of ascidians (Chordate -Tunicate), and characterized in our laboratory.These glycosaminoglycans possess unique structural features, when compared to their mammalian counterparts.The ascidian heparin has the same basic structure of mammalian heparin, but with less 2-O-sulfation.On the other hand, the ascidian dermatan sulfate contains the same backbone structure but with different patterns of sulfation.All of the ascidians dermatan sulfates had a high content of iduronic acid 2-sulfate residues, but they differed in the pattern of sulfation of the Nacetylgalactosamine units.These differences produce profound effects on the anticoagulant and antithrombotic activities of the ascidian glycosaminoglycans.In the present review we describe our findings about the structure and the anticoagulant prop-MAURO S. G. PAVÃO erties of these compounds.

MAMMALIAN GLYCOSAMINOGLYCANS
The ability of certain glycosaminoglycans to interfere with blood coagulation has been known for over 50 years (McLean 1916), as illustrated by the extensive use of heparin as an antithrombotic agent (Rodén 1989).The effect of heparin is to accelerate the inactivation of plasma coagulation enzymes by the serpins (serine protease inhibitor) antithrombin (Bourin and Lindahl 1993) and heparin cofactor II, another serpin, which is also activated by dermatan sulfate, and which selectively inhibits thrombin (Tollefsen et al. 1982).

Heparin
Heparin is a highly sulfated glycosaminoglycan composed by disaccharide repeats of hexuronic acid (α-L-iduronic acid or β-D-glucuronic acid) linked 1,4 to D-glucosamine.The heparin molecules isolated from highly vascularized tissues, such as beef lung and hog mucosa are made up of a heterogeneous mixture of polymers with a similar backbone.The heterogeneity results from variations of sulfation on the D-glucosamine (N-acetylated, N-sulfated, Osulfated at C6 and/or C3) and on the uronic acid residue (O-sulfated at C2).Because of its unique binding to antithrombin (ATIII), involving a specific pentasaccharide sequence, heparin is endowed of a potent anticoagulant activity (McLean 1916).In the presence of heparin, the rates of inhibition of thrombin, factor Xa, and factor IXa by AT III are increased ∼ 1000-fold (Jordon et al. 1980) so that inhibition is essentially instantaneous.AT III inhibits the target protease by forming a 1:1 stable complex with the protease (Rosenberg and Domus 1973).The AT III-binding pentasaccharide has been identified as GlcNAc(6SO 4 )-GlcA-GlcNS(3-SO 4 )-IdoA(2-SO 4 )-GlcNS(6-SO 4 ) (Conrad 1998).The O-sulfate group at C3 of the glucosamine residue is always present, and represents a unique structure found only in the ATIII-binding sequences of heparin.The induction of a conformational change in ATIII appears to be sufficient for the acceleration of the inactivation of factor Xa (Rosenberg and Domus 1973).In contrast, acceleration of the thrombin inhibition reaction by AT III requires an interaction between heparin and thrombin as well as heparin and ATIII (Griffith 1983, Oosta el al. 1981, Laurent et al. 1978).

Dermatan Sulfate
Dermatan sulfate is an anticoagulant glycosaminoglycan consisting of alternating disaccharide units of hexuronic acid (iduronic acid or glucuronic acid) 1, 3 linked to N-acetylgalactosamine.Variations in the degree of sulfation on both hexuronic acid (2-Osulfated) and N-acetylgalactosamine (4-O-or/and 6-O-sulfated) are responsible for the extensive heterogeneity of this polymer (Kjellén and Lindahl 1991).The anticoagulant activity of dermatan sulfate is due to the binding of a specific region within the molecule, consisting of [IdoA(2SO 4 )-GalNA(4SO 4 )] n (n ≥ 3) (Scully et al. 1988, Maimone and Tollefsen 1990, Pavão et al. 1998) to a glycosaminoglycan binding site in heparin cofactor II (HCII), a serine protease inhibitor from plasma (Tollefsen et al. 1982).The binding of dermatan sulfate increases by 1000 fold the specific inhibition of thrombin by HCII (Tollefsen et al. 1983).Since [IdoA(2SO 4 )-GalNA(4SO 4 )] comprises only 5% of the disaccharides present in dermatan sulfate, clustering of this subunit to form the high affinity hexasaccharide must not be a random biosynthetic event (i.e., the random probability of three such disaccharides occurring together would be 0.000125).

Antithrombotic Use of Mammalian Glycosaminoglycans
Heparin has been used clinically for almost 70 years.It is an important therapeutic agent in the treatment of patients with thrombosis or patients at risk to develop it.Heparin is highly used in the prevention of thromboembolic events frequently observed after some kind of surgery, especially pelvic and orthopedics (Leyvraz et al. 1983, Poller 1985).How-ever, the unrestricted use of heparin as an antithrombotic agent is limited due to its various side effect, such as thrombocytopenia, osteoporosis, and hemorrhagic complications (Ockelford et al. 1982, Carreras 1980, Mätzsch et al. 1986).More recently, heparin has been gradually replaced by its low molecular weight derivative (less than 3,000 and ∼ 9,000, comparing to ∼ 13,000 of native heparin), obtained by partial fragmentation and fractionation of native heparin (Hamulyák et al. 1995, Meyer et al. 1995).Fragments below 16 to 20 monosaccharide units per heparin molecule (MW < 5, 000), while containing the specific pentasaccharide sequence essential for binding AT III, are not long enough to permit binding to thrombin; they therefore inhibit only activated factor X (Choay et al. 1981).Low molecular weight heparins prolong moderately the clotting time (indicating no thrombin inhibition) but are still capable of potentiating the inhibition of activated factor X. This fact at first raised the hope of dissociating the antithrombotic property (anti-Xa) from the anticoagulant (inhibition of thrombin), which then would avoid the hemorrhage-inducing effect of unfractionated heparin.However, it has been shown more recently in animal experiments, that anti-Xa activity is a prerequisite, although not sufficient in itself, for a thrombosis-preventing effect.Heparin molecules, large enough to retain some thrombinblocking action are indeed also necessary (Holmer et al. 1982, Thomas et al. 1982, Thomas et al. 1989).
Dermatan sulfate was initially shown to have antithrombotic activity in the rabbit jugular venous stasis model (Fernandez et al. 1986).The amount of bleeding from a standardized incision was much greater with heparin than with dermatan sulfate at equivalent antithrombotic doses, suggesting that dermatan sulfate might be a safer therapeutic agent.The relative hemorrhagic effects of the two glycosaminoglycans appeared to be correlated with the ability to inhibit collagen-induced platelet aggregation in vitro.Recent studies suggest that dermatan sulfate is a more potent inhibitor than heparin of clotbound thrombin (Fernandez et al. 1986, Buchanan et al. 1986) and that dermatan sulfate in combination with low molecular weight heparin may produce synergistic antithrombotic effect (Cosmi et al. 1993).Human clinical trials have shown that dermatan sulfate prevents thrombosis during hemodialysis (Lane et al. 1992) and that in the post-operative setting (Agnelli et al. 1992, Prandoni et al. 1992).However, the mechanism of antithrombotic activity of dermatan sulfate has not been clearly established yet.The antithrombotic mechanism of dermatan sulfate is a matter of controversy.Fernandez et al. (1987) showed that there is a good correlation between antithrombotic potency and HCII activity of dermatan sulfate.However, Dol et al. (1988) reported that enhance of the catalytic activity of dermatan sulfate fail to increase its antithrombotic activity.More recently, Sie et al. (1993) showed that the antithrombotic effect of dermatan sulfate is essentially dependent of HCII binding and activation.

Heparin
A heparin with similar structure and lower anticoagulant properties to the mammalian heparin has been identified in the test cells and tissues of the ascidian Styela plicata (Cavalcante et al. 2000).An extensive structural analysis of the polymer indicated that the invertebrate heparin is composed mainly by the disaccharide ]n is also present.This molecule was shown to occur in cytoplasmic granules of test cells, and also inside granules of epithelial cells forming a layer along the lumen of intestine and pharynx of the ascidian Styela plicata (Cavalcante et al. 2000).
The anticoagulant properties of the ascidian heparin (Table I) show that the polymer has an anticoagulant activity 10 times lower than the mammalian heparin.In addition, it is about 20 times less potent in the inhibition of thrombin by antithrombin, when compared to the mammalian counterpart (Table I).However, it activates heparin cofactor II (HCII) by the same extent, as indicated by the IC 50 for thrombin inhibition in the presence of the inhibitors.

Dermatan Sulfate
Dermatan sulfates, similar to the mammalian glycosaminoglycan, but with differences in the degree and position of sulfation was previously isolated from the body of the ascidian Ascidia nigra (Pavão et al. 1995) and Styela plicata (Pavão et al. 1998) (Table II).The dermatan sulfate in the tissues of the ascidian occurs in the extracellular matrix of intestine and pharynx (Gandra et al. 2000).The A. nigra dermatan sulfate, composed predominantly by [4α-L-IdoA(2SO 4 )-1 → 3-β-D-GalNAc(6SO 4 )-1]n, where n ≥ 3, has no discernible aPTT activity and has very low ability to potentiate heparin cofactor II, as indicated by the IC 50 for thrombin inhibition (Table II).On the other hand, the dermatan sulfate from S. plicata, composed by [4-α-L-IdoA(2SO 4 )-1 → 3-β-D-GalNAc(4SO 4 )-1]n, where n ≥ 3, has an aPTT activity 4 times higher than mammalian dermatan sulfate and is a potent activator of heparin cofactor II, possessing a HCII activity 10 times higher when compared to the mammalian counterpart (Table II).

In Vivo Anticoagulant Activity of Ascidian Dermatan Sulfates
We have started to investigate the effect of ascidian and mammalian dermatan sulfates on coagulation, using an experimental model in rats.4 mg/kg of the dermatan sulfate samples were injected intravascularly as a bolus in the carotid artery of anesthetized Wistar rats.Blood was collected at different times after administration of the glycans for aPTT analysis (Vicente et al. 2001).We observed an increase in the aPTT up to ∼ 3-fold during the first 10 min, following injection of the S. plicata dermatan sulfate, returning to basal levels after 20 min.Bovine lung dermatan sulfate induced a less intense change in the aPTT values, with a maximum increment of 1.5-fold at 15 min, returning to normal values after 20 min.A. nigra dermatan sulfate did not produce any significant change in the aPTT during the 60min experiment.These results show a good correlation with the experiments obtained in vitro, where the aPTT was measured in human plasma.Intravascular administration in rats of the low HCII active A. nigra dermatan sulfate, which showed a discernible aPTT activity in vitro (Table II), also did not produce a significative change in the aPTT in the experiment in vivo.Mammalian dermatan sulfate, which has an intermediate HCII activity, produced a mild pronounced change in the aPTT in the experiments in vivo S. Plicata dermatan sulfate which has the higher HCII activity, produced the higher change in the aPTT in vivo (Vicente et al. 2001).

In Vivo Antithrombotic Activity of Ascidian Dermatan Sulfates
The effect of ascidian and bovine lung dermatan sulfate samples on thrombosis was investigated preliminarily, using an experimental venous thrombosis model.A single injection of S. plicata dermatan sulfate, given 5 min before the thrombogenic stimulus with brain thromboplastin caused a dose-dependent inhibition of thrombus formation.With the dosage of 1.6 mg/kg body weight, a 50% reduction was observed and with 4 mg/kg body weight, thrombosis was not observed in any of the animals.Surprisingly, mammalian dermatan sulfate was more effective than S. plicata dermatan sulfate in dosages up to 3 mg/kg body weight.A 50% reduction was observed at the dosage of 0.9 mg/kg body weight and with 4 mg/kg, mammalian dermatan sulfate totally

DS sample
Major disaccharide units aPTT HCII activity (units/mg) b IC 50 (µg/ml) c Bovine lung GalNAc(6SO 4 )-1]n 0.5 320 a These glycosaminoglycans have the same backbone structure [4-α-L-IdoA-1 → 3-β-D-GalNAc-1]n, but possess different patterns of sulfation.The repetitive disaccharide units of bovine dermatan sulfate are sulfated at carbon 4 of the galactosamine moiety; small amounts (< 5%) of 2-O-sulfated α-L-iduronic acid residues are also present.The S. plicata dermatan sulfate is sulfated at both 2-O-position of the α-L-iduronic acid and the 6-position of the N-acetyl-β-D-galactosamine units.On the A. nigra dermatan sulfate most of the α-L-iduronic units are 2-O-sulfated and the N-acetylβ-D-galactosamine are 6-O-sulfated.These differences produce profound effects on their anticoagulant properties.The S. plicata dermatan sulfate has the higher aPTT and HCII activities, whereas the A. nigra dermatan sulfate has the lowest anticoagulant properties.Bovine lung dermatan sulfate possesses intermediate aPTT and HCII activities.HCII, heparin cofactor II.b aPTT activity is expressed in units/mg, using a standard curve based on the international Heparin standard (193 units/mg).c Values obtained from Pavão et al. 1998./smallskipprevented thrombus formation.A. nigra dermatan sulfate, up to 4 mg/kg body weight, was totally ineffective in the reduction of thrombosis.These results suggest that HCII activity is required for the antithrombotic activity of dermatan sulfate, but in our case an enhancement of the HCII activity did not increase antithrombotic potency of the polymer, in the experimental model used (Vicente et al. 2001).

Effect of Ascidian Dermatan Sulfates on Platelets and Hemorrhagic Effect
To investigate the effect of dermatan sulfate samples on platelet aggregation, ascidian or bovine lung dermatan sulfates were added to a suspension of human washed platelets, and the platelets in solution were counted by a cell counter.Alternatively, platelet aggregation was evaluated in human platelet rich plasma (PRP) by aggregometry.In control experiments ADP was used as an inductor of plasma aggregation.ADP reduced the single platelet count from ∼ 300×10 6 platelets/ml to ∼ 50×10 6 platelets/ml.Addition of S. plicata, A. nigra or bovine lung dermatan sulfate did not produce a reduction in the single platelet count.After addition of ascidian or mammalian dermatan sulfate samples to a human PRP preparation no aggregation was observed during a 20-min experiment.ADP, on the other hand, induced platelet aggregation almost instantaneously (Vicente et al. 2001).
The hemorrhagic risk of dermatan sulfate samples was assessed based on blood loss from a rat cut-tail, after intravascular administration of the glycans.At the dosage of 4mg/kg body weight, the S. plicata dermatan sulfate did not modify the blood loss compared with rats receiving saline.Surprisingly, the blood loss was increased ∼ 2-fold in rats receiving the same dosage of A. nigra or bovine lung dermatan sulfate (Vicente et al. 2001).
Generally, bleeding effect results from modification of blood coagulation induced by polysaccharides.An increase in plasma aPPT values is usually related with high bleeding effect (Hirsh 1984, Thomas andRoberts 1997).We used a rat cut-tail bleeding experimental model to compare the hemorrhagic effect of the ascidian and mammalian dermatan sulfates.Surprisingly, the S. plicata dermatan sulfate at the dosage where the higher antithrombotic effect is observed (4mg/kg body weight), showed the same hemorrhagic effect as that of the control.In contrast, mammalian and A. nigra dermatan sulfate, increased blood loss almost 2-fold, compared to the control.The high bleeding effect of mammalian and A. nigra dermatan sulfate observed does not seem to be related to an effect of these polysaccharides on platelets, since none of the dermatans induced platelet aggregation (Vicente et al. 2001).

CONCLUSION
The data reported in the present review provide important information about the chemical structures in glycosaminoglycans that produce the anticoagulant, antithrombotic and hemorrhagic effects of these polymers.Thus, the anticoagulant activity of dermatan sulfate is related to sulfation at specific positions at carbon 2 of iduronic acid and at carbon 4 of the N-acetylgalactosamine residues, that are required to bind and activate heparin cofactor II.Binding and activation of heparin cofactor II is necessary for the antithrombotic activity of dermatan sulfate, but an enhancement of the HCII activity does not increase the antithrombotic potency of the polymer.
Moreover, the hemorrhagic effect of these polymers does not seen to be associated with their anticoagulant effect.