Structure-activity relationship of

antithrombotic polysaccharide
derivatives

G. Franz and S. Alban

University of Regensburg, 93040 Regensburg, Germany
Received 12 October 1994; revised 16 January 1995

Heparin has been the drug of choice in clinical pre-surgical and post-surgical prophylaxis
of thrombotic events. However, because of its side-effects, such as bleeding and other
disadvantages (i.e. chemical inhomogeneity and variability of its physiological activities),
alternatives to heparin are an important field of research. A necessary procedure in the
development of new drugs is the evaluation of structure-activity relationships. Genuine
neutral polysaccharides were chemically modified and examined for their anticoagulant
activities. The linear fl-1,3-glucan curdlan, an easily available bacterial polysaccharide, served
as the basic polymer. It could be established that the anticoagulant activity was dependent
on the degree of sulfation and the molecular weight. For heparin, the sulfation pattern, i.e.
the actual location of the sulfate groups along the heparin chain, was of importance in
addition to the degree of sulfation. Therefore, we investigated whether there was also a
relationship between the substitution pattern of the curdlan sulfates and their anticoagulant
activity. For determination of the substitution pattern of the sulfated polysaccharides, a
method was developed that is based on synthesis of the partially alkylated alditol acetates
of the polymer and examination of these derivatives using combined gas chromatography-
mass spectrometry. In addition to the analytical data, the structure-activity relationship of
anticoagulative curdlan sulfates is presented.
Keywords: heparinoids; fl-l,3-glucan; curdlan

In recent years, a broad series of polysaccharides has
emerged as an important class of bioactive natural
products. However, most scientists may think of poly-
saccharides primarily as an energy source or as purely
structural polymers without any physiological effects in
plants, animals and humans. It was indeed surprising
during recent years to learn about possible different
effects of a variety of carbohydrate polymers in treating
a broad range of diseases, such as immune disorders
and tumours, and their activities as anti-inflammatory,
anti-ulcer, cholesterol-lowering, blood-glucose-lowering
or antithrombotic agents 1. Some of these new poly-
saccharide applications have been approved by the
scientific community, while others are still matters of
controversy and hence have not yet been accepted for
clinical use.
An essential prerequisite, when distinct physiological
properties are to be attributed to a biological molecule,
is knowledge of the structural parameters of such
polymers 2. Since the physiological activity of a poly-
saccharide depends on its purity, it is essential to utilize
the most sophisticated techniques for their isolation,
purification, structural modification and structure deter-
mination.
The biological function of glycosaminoglycans
Glycosaminoglycans are complex polysaccharides with
several types of functional groups, such as sulfate ester,
carboxylate, acetamide, and primary and secondary
hydroxyl groups. The biological function of these
compounds is at least in part related to precise
distributions of these functional groups on distinct
oligosaccharide sequences, as for example in the case of
heparin when binding to antithrombin 3. Other functions
of these acidic and sulfated polysaccharides have not yet
been assigned to precise structural domains, such as the
antiproliferative activity of heparin, smooth muscle cell
growth inhibition, or its antiviral activity (dextran
sulfate).
The functional groups can either be modified or
substituted to introduce structural changes to glycos-
aminoglycans with the consequence of altered physio-
logical/pharmacological properties. Thus, the anticoagulant
activity of genuine heparin can be abolished by chemical
degradation of the specific antithrombin binding site using
periodate oxidation, while the antiproliferative activity
is preserved. Chemical modification or partial synthesis
of sulfated polysaccharides provides an opportunity

0141-8130/95/$09.50© ElsevierScienceB.V.All rights reserved Int. J. Biol. Macromol. Volume 1 7 Number 6 1 995 311
Antithrombotic polysaccharide derivatives: G. Franz and S. Alban
to obtain new pharmacological agents with possible
therapeutic uses.
Heparin-a classical biopolymer
Genuine heparin, 75 years after its discovery, remains an
important tool in medicine for the prevention of
post-operative thrombosis, the treatment of acute venous
thrombosis and the prevention of clot formation in the
heart-lung system. The year 1935 was a remarkable
landmark in the history of this glycosaminoglycan. By
then it was obvious that this biopolymer was not a
phospholipid as had originally been assumed 4. It was
shown that this biopolymer was in fact a carbohydrate
derivative containing uronic acid. The newly developed
Elson-Morgan procedure provided evidence for the
presence of hexosamines in heparin, and the high sulfate
content was determined by analysis of the ash. It was
concluded that this molecule was the most highly charged
polyanion in nature. In 1936, five monosaccharides were
identified as components of the heparin molecule:
D-glucosamine, D-glucuronic acid, L-iduronic acid, D-
galactose and D-xylose. The definitive monosaccharide
composition was not obtained until 1964. In 1950, it was
further established that heparin in situ was covalently
linked to protein via serine residues. In the 1940s, it was
assumed that the pronounced anticoagulation activity
of heparin was in part due to its high negative-
charge density. This hypothesis was supported by the
finding that chemical sulfation of otherwise neutral
polysaccharides enabled them to act in a similar way.
These components with a carbohydrate backbone and
containing carboxyl- and sulfate functions were named
'heparinoids' which are synthesized and biologically
examined at a large scale.
Drawbacks of heparin Heparin has remained the anti-
coagulant of choice for over three decades s. It is usually
obtained from mammalian tissue such as lung or mucosa.
Genuine heparin shows a broad molecular heterogeneity,
consisting of various chains of sulfated polysaccharides
or oligosaccharides with molecular weights ranging from
2000- 50000. The mixtures can be fractionated into
low- and high-molecular-weight compounds differing in
their physiological and pharmacokinetic behaviour 6.
Low-molecular-weight heparin (LMWH) is usually
obtained by chemical or enzymatic digestion of isolated
heparin. The LMWHs offer several advantages over
genuine heparin: better bioavailability, lower risk of
bleeding, and the requirement of only one injection per
day for prophylactic action. Another possibility could be
synthesis of better defined heparin. It has been shown
that heparin sequences can be synthesized a priori in a
multi-step approach under very uneconomic conditions.
However, a five-membered pentasaccharide has been
obtained by Petition et al. in a 75-step synthesis and has
gained some importance 7. Biotechnological methods for
controlled production have been tried but with little
success. Finally, it was discovered recently that heparin
extracted from mammalian sources could be contaminated
with highly infectious bovine spongiform encephalopathy.
312 Int. J. Biol. Macromol. Volume 17 Number 6 1995
New anticoagulants replacing heparin:
the heparinoids
In the past, attempts to replace heparin with other
anticoagulants have not always been successful. Many
glycosaminoglycan-derived agents such as heparan,
dermatan and chrondroitin sulfate have been examined 8.
Mostly these agents show weaker anticoagulant action
and are very often devoid of anti-platelet actions. In order
to eliminate these differences, attempts were made to
establish new sulfate polymers in which the critical
features could be controlled, such as chain length, type
of glycosidic linkage, degree of sulfation and position of
sulfate groups. Another possibility is examination of
naturally occurring sulfated polysaccharides, mainly of
algal origin, such as fucoidans and carrageenaans 9. Some
showed weak biological activity, others were completely
devoid of any anticoagulant activity, and others were
simply too toxic for any application in humans. However,
there was still almost no knowledge concerning a
possible relationship between the sulfate substitution
pattern, the chain length, conformation and the resulting
physiological activity 1°.
This was the background against which we started to
produce a series of new heparinoids utilizing neutral,
unsubstituted, structurally well-defined polysaccharides,
i.e. fl-l,3-glucans, such as laminarin and curdlan.
Laminarin has a slight degree of branching at C6 while
curdlan is strictly linear. Both differ in their respective
molecular weight range ~1.
Experimental studies
The first aspect of our studies was concerned with
adequate methods for non-destructive polysaccharide
sulfation. Furthermore, a method had to be established
to allow position-specific introduction of sulfate groups
on the glucose monomers in addition to permanent
control of the overall degree of sulfation (DS) and the
sulfation pattern.
Prior to sulfation, treatment of these glucans with
N,N-dimethylformamide (DMF) was shown to be very
important. The highly polar D M F associates with the
hydroxyl groups of the substrate and makes them more
accessible to the SO3/pyridine complex which is utilized
as a sulfate donor. The concentration of the SO3/pyridine
complex in D M F as well as the ratio of the glucan and
the sulfation reagent were varied in order to obtain
derivatives of different DS.
It could be shown that the DS of the newly synthesized
laminarin sulfates mainly depended on the molar
concentration of the SO3/pyridine complex 12. The DS
was determined by conductometric titration of the sulfate
groups. The increase in the molecular weight of the newly
synthesized laminarin sulfates was controlled by gel-
permeation chromatography (GPC) as shown in the case
of LAMS5 with a DS of 1.5. This is further proof for the
non-degrading conditions of the sulfation reaction.
In order to analyse the substitution pattern of the
compounds obtained, a modified methylation method in
combination with gas chromatography-mass spectroscopy
(GC-MS) was established. Partially methylated alditol
acetates (PMAA) of the corresponding polysaccharides
were analysed by GC-MS. The free hydroxyl groups were
transformed to the corresponding methyl ethers. The
 Antithrombotic polysaccharide derivatives: G. Franz and S. Alban
glycosidic linkage was then transformed to the cor-
responding acetic ester. The sulfate groups of the
glucan sulfates were examined and shown to be stable
under alkaline conditions.
The reaction starts with the classical methylation of
the free hydroxyl groups using CHaJ after forming
polyalkoxide ions of the carbohydrates by means of
dimsyl-K. The glycosidic linkages as well as the sulfate
groups were cleaved by treatment with TFA. The
resulting partially methylated glucose monomers were
reduced and acetylated. One further essential step
was necessary for complete methylation of the poly-
saccharide. Due to the presence of the sulfate groups, the
partially methylated polysaccharides are soluble in H 2 0
and thus cannot be extracted in the classical manner with
organic solvents. This problem was overcome by
preparing the soluble pyridinium salt of the corresponding
sulfated polysaccharide.
GC-MS of the sugar derivatives gave the following
results: non-sulfated glucose was present in all laminarin
derivatives with increasing DS; however, the higher the
DS, the lower the number of non-sulfated glucose
residues. In the case of the sulfated glucose moieties, it
was obvious that the reactivity of the primary hydroxyl
groups in position C6 is greater than that of the secondary
hydroxyl groups in positions C2 and C4.
Comparing the products with increasing DS, the
difference between sulfation of the primary and the
secondary hydroxyl groups becomes smaller due to the
increasing ratio of di- and trisulfated glucose units. With
respect to the distribution of the sulfate groups on all the
glucose units in the polymer chain, derivatives up to a
DS of 0.6 are homogeneous. The glucose is either
non-substituted or selectively substituted in position C6
only. With increasing DS, the homogeneity decreases and
more di- and trisulfated glucoses can be detected.
If we assess LAMS3 with a DS smaller than 1, we
find the following substitution pattern: 30% non-sulfated,
50% monosulfated, 10% disulfated and 4% trisulfated
glucose residues. For LAMS4 with a DS above 1, the
sulfation pattern has changed mainly by an increase of
the disulfated glucose residues up to more than 15%.
The next step was the preparation of sulfated
compounds, which vary in their chain length only
and not in the sulfation pattern. These products were
needed to establish a possible correlation between
anticoagulant activity and the relative molecular weight.
A sulfated curdlan with an average DS of 0.6 was
subjected to GPC for a molecular weight range of 2000
up to 380000. Fractions were collected, analysed for
molecular weight and DS. It could be shown that a more
or less uniform DS was obtained in all fractions. The
sulfation pattern showed a preference of C6 sulfation in
the high-molecular-weight fraction, which decreases in
favour of C2 sulfation. In order to obtain derivatives with
low sulfation of the primary hydroxyl group, the C6 was
protected by introducing adamantoyl residues in this
position. This reaction was followed by the normal
sulfation procedure 13.
Testing of the sulfated glucans
Actual antithrombotic drugs represent a wide spectrum
of natural, synthetic, semi-synthetic and even bio-
technologically produced agents with marked differences
in chemical composition, biochemical action and phar-
macological effect. In most cases, only very limited data
on their pharmacological characterization are available.
With the use of newer biochemical and pharmacological
screening methods, a valid screening can be used to
develop these polymers for specific clinical indications. A
cautious approach should be made, however, when
interpreting data from experimental models and extrapol-
ating these to the clinical setting. Because of these
complexities, the heparinoids and related substances
cannot be standardized as a single group. Each compound
has to be examined as a separate entity. Heparin,
glycosaminoglycans and heparinoids effect their bio-
logical action through multiple sites and therefore cannot
be assayed by one global anticoagulant test system.
If we take a closer look at the coagulation cascade, it
can be seen that three major reaction complexes
participate and hence can be utilized for an anticoagulant
assay. They interact at different sites and can be used
independently of each other.
The intrinsic pathway comprises the left side of the
reaction sequence. The activated thromboplastin time
(APTT) is the most widely used clinical laboratory test
for monitoring the heparin effect. By measuring the
APTT, mainly the factors XII to VIII are measured. For
the thrombin time (TT), the clotting effect is determined,
when, by interaction with thrombin, fibrinogen is
transformed to fibrin. Finally, the HEPTEST measures
the inhibition of exogenous factor Xa by the plasma factor
AT III, and can be measured and calculated as
prolongation of the recalcification time. Hereby mostly
the extrinsic pathway reaction sequence is determined.
By utilizing all three test systems for all the compounds
mentioned before, we can obtain an overall picture of
the biological activities of the sulfated glucans, i.e.
indications for a possible structure/activity relationship.
As a first example, we have examined the sulfated
laminarins using all three test systems. These laminarins
differ mainly in their respective DS values and only
slightly in their molecular weight.
It is obvious that a minimum degree of sulfation is
essential to obtain any anticoagulant activity. LAMS1
and LAMS2 were inactive in all three test systems. A
certain DS, i.e. above 0.6, is needed. With increasing DS
to 1.5, the biological effect rises, after which one
can observe a decrease by all three systems. It was
further obvious that the APTT test system gave the
most pronounced reactivity, an indication that these
compounds interact at a relatively early stage of the
coagulation cascade 14.
However, this reactivity was altered considerably when
sugar side chains were introduced in the main chain by
different chemical methods. Addition of either arabinose,
glucose, rhamnose or the disaccharide gentiobiose
considerably changed the architecture of the molecule
with the consequence that the onset of the reactivity was
altered. This was mainly the case for the arabinosyl
derivative, where, even at very low DS, a steep rise in the
anticoagulant activity was obvious.
In a subsequent series of experiments, the dependence
of the different biological activities in relation to the chain
length was studied. The fractionated curdlan sulfates, one
group with a regular DS of 0.6 and a second group with
a DS of 1.2, were examined for a molecular weight range
of 12 000 to 160 000. A difference in reactivity by all three
Int. J. Biol. Macromol. Volume 17 Number 6 1995 313
Antithrombotic polysaccharide derivatives: G. Franz and S. A/ban
systems can again be seen, with reactivity for the low DS
values being most pronounced by the APTT test system
and that for the high DS values being most pronounced
by the TT test system.
We can see that both the degree of sulfation as well
as the chain length drastically influence the biological
activity. In the following tests, the specifically sulfated
/~-l,3-glucans were subjected to all three test systems. We
compared three groups of sulfated curdlans which
comprise the molecular weight ranges of 10, 25 and 50 000,
respectively, and which further contain one member
which is non-C6-protected and one member which is
protected at the primary OH on C6. These curdlans
contain more than 50% of the respective sulfate groups
in positions 2 and 4.
The result, i.e. the variation in the biological effect, is
obvious: the higher the ratio of secondary hydroxyl group
sulfate esters, the greater the anticoagulant activity. In
correlation with both DS and MW, the individual tests
were influenced in a different manner, as reflected by the
aFXa/aFIIa values is.
Consequently, uniform distribution of the sulfate ester
groups on the glucose units improves the anticoagulant
activities considerably. Derivatives with a comparable
activity are only obtained for compounds with higher
molecular weifllit and elevated DS.
According to these findings, sulfation of the primary
hydroxyl group at position C6 is not essential for an
anticoagulant effect. Further, it can be concluded
that the ratio of the specific activities is governed by the
combined variations of ~ molecular weight and the
sulfate group content on secondary hydroxyl groups. The
questions remain open, whether such sulfated poly-
saccharides with a relatively low DS and molecular
weight and good in vitro anticoagulant activities will
operate in vivo as potent antithrombotic agents that are
useful for clinical purposes and whether they will prove
to have a reduced risk in terms of acute and long-term
toxicity.
Sulfated polysaccharides to prevent angiogenesis
Angiogenesis, the growth of capillary blood vessels,
plays an important role in normal development and in
physiological functions. Angiogenesis is essential to
the repair of wounds, peptic ulcers and myocardial
infarctions. In these physiological and repair conditions,
angiogenesis is regulated and switched on and off at
predictable times. Tumour growth and metastases are
also angiogenesis-dependent. Continuous induction of
neovascularization is necessary for progressive tumour
development; increasing neovascularization of certain
tumours correlates with increasing metastatic potential.
Over the last two decades, it has been dearly
314 Int. J. Biol. Macromol. Volume 17 Number 6 1995
demonstrated that sulfated glycosaminogiycans such as
heparan sulfate and heparin are involved in the
angiogenetic process 16. It is now known that this specific
activity of genuine heparin is interlinked with the basic
fibroblast growth factor (b FGF). It is possible that other
growth factors may also interact with heparin with the
consequence of mobilization of the F G F from the
extracellular matrix.
Several sulfated polysaceharides can be substituted for
heparin as a regulator of angiogenesis. This led us to
examine a series of known polysaccharide sulfates
which should have only anti-angiogenetic but no
antithrombotic effects. Optimal results were obtained
with genuine carrageenaans which we purified and
standardized. A good test system for examination of the
anti-angiogenesis effect is the chick embryo, where, after
application of the substrate, a concentration-dependent
inhibition of the vascularization can be seen. It was
obvious that carrageenaans were able to stabilize b F G F
and other growth factors in a similar manner to that
shown by heparin.
It is likely that sulfated polysaccharides have a series
of more fundamental regulatory roles, as in the effect on
growth factors, especially in the complex process of
angiogenesis.
Acknowledgements
This work was supported by the 'Fonds der Chemischen
Industrie'.
References
1 Franz,G. Planta Med. 1989,55, 493
2 Blaschek,W. 'Polysaccharides',¢d G. Franz, Springer Verlag,
Berlin, 1991
3 Cannon, C.P. 'The Pharmacology of Anticoagulation',¢xl R.
Pifarrr, Hanleyand BelfusInc, Philadelphia,PA, USA, 1993
4 Casu,B. Adv. Carbohydr. Chem. Biochem. 1985,43, 51
5 Linhardt,R.J. Chem. Ind. London 1991, 1, 45
6 Hemker, H.C., Brguin, S., Bendetowicz,A.V. and Wielders, S.
Thromb. Haemost. 1991,65, 845
7 Petitou,M., Lormeau,J.C. and Choey J. Nature 1991,350, 30
8 Thomas, D.P., Gray, E. and Merston, R.E. Thromb. Haemost.
1990, 64, 290
9 Sveda,S., Sakaguchi,S., Shimano,H. and Nagamatsu,A.Biochem.
Pharmacol. 1992,43, 1853
10 Bode,V. and Franz, G. Arch. Pharm. 1991,324, 363
11 Alban,S. and Franz, G. Drug Res. 1992,42, 1005
12 Alban,S. and Franz,G. Synthes¢und antithrombotischeAktivit/it
yon neuartigenHeparinoiden.In: 'Entziindungenmad verwandte
Reaktionen- neueWirkstoffe',JonapharmPublishers,Jena, 1993
13 Alban,S. and Franz, G. Thromb. Haemost. 1994, 20, 152
14 Alban,S. and Franz, G. Pure Appl. Chem., in press
15 Alban,S. phD Thesis, Universityof Regensburg, 1993
16 Folkman, J. and Ingber, D.E. in: 'Hoparin, Chemical and
Biological Properties', eds A. Lane and I3. Lindahi,CRC Press,
Boca Raton, FL, USA, 1989,p 317