CHEMelL
`
`Volume 00, Number 8
`
`December 1998
`
`Hyaluronan: Preparation, Structure, Properties, and Applicationst
`
`Lubomir Lapcik, Jr.* and Lubomir Lapcik
`
`Faculty orCtemisuy, Tecnnical Universiiy or &no, Purkyflova 118, CZ-812 00 Brno, Czech Republc
`
`Stefaan De Smedt•,§.u and Joseph Demeester§
`
`Faculty of Pharmacy, Unive,sily d Ghent, Harebekestraat 72, B-9000 Ghent, Belgirn
`
`Peter Chabrecek
`
`Cerlral Research l.abaalories. Ciba-Geigy, A.G., Postfadl, CH-4002 Basel, Swilzerland
`
`Reaiived Api/128, 1997(/,evised Manu!aipl Received Se,:terrber 3, :998)
`
`Contents
`1 · lntroelJalon
`2. Macromdecuar Character
`2.1. Polyelectro~e Properties and CC1lf00llation
`2.1.1. Polyelectrolyte Properties
`2.1.2. Conformaion
`2.2. Hydrodyranic Behavior and I nteracti:lns
`2.2.1. H~rodynamic Behavior
`2.2.2. Ir1eractions t>av.een HA Chans
`2.2.3. Interactions between HA Chans and
`01tler Macromdecules
`2.3. Molecular Weig-it
`2.3.1. Studes on the Average Molecuar Weig-it
`and Mdecular WeiJht Distribution
`2.3.2. Separation of HA Oligo.5acchari:les
`3. Degradation
`3.1. Ulrasoric Degradation
`3.2. pH-Deperoert and Thermal Degradation
`3.3. [)aJiadation by Free Radicals
`3.4. Enzymatic DegradatiCJl
`4. Ch8ncal Derivatization of HA
`4.1. H~ns
`4.2. Other HA Derivatives
`
`2663
`2665
`2665
`2600
`2600
`2008
`2008
`2008
`2670
`
`2671
`2671
`
`2672
`2673
`2673
`2673
`2673
`2674
`2674
`2674
`2674
`
`t This paper is dNlicated to l'ro f. Dr. A. La uwe.-s who has iuipired
`"" wit.h hi• sci<>nt.ific Appt-OAch, hi" h OllP1'1Y, an,I h i• hnnt a n
`wa,,111h .
`.. Correspondence auchor.
`* Abo wit.Ii tlH' ~' acuity of Ttx:l moloity i11 i 1i11, (}111ich Rt!11ublic.
`f Aul.hors who equally oontribut.ed as the first author.
`"Po.tdoct.oral fdlow of P.W .O.-Vla1111Jere11.
`
`6.
`
`5.
`
`Establis,ed Medi:al and Pharmaceutical
`Applicatons of HA
`Medcal aM Pharmaceuieal Reooaa::h toward
`Fuure Applications
`6.1. Perspectives br Future Applications of
`Noomodified HA
`6.1.1. Ophthalmol~y
`6.12. Nasal Drug Delivery
`6.1.3. Parenteral Drug DeliveJY
`6.1.4. Lymphatic Drug Delivery
`6.1.5. Dermatdogy
`6.1.6. Surgery
`6.1.7. lm~lams
`6.2. Perspectives tor Future Apjjicatims of
`Modified HA
`62.1. H~ans
`62.2. HA Estors
`7. Ackn~edgments
`8. References
`
`2675
`
`2676
`
`2677
`
`2677
`2677
`2677
`2678
`2678
`2678
`2678
`2678
`
`2678
`2679
`2681
`2681
`
`1. Introduction
`Hyaluronan (HA, :t'igure 1), a high molecular
`weigh t biopolysaccharide, was discovered by Meyer
`and Palmer in 1934 in t he vitreous humor of cattle
`eyes. 1 HA is a member of a group of simi lar po lysac(cid:173)
`chatides that have been termed "connective tissue
`polysaccharides", "mucopo]ysacchai;des", or "g]y-
`t
`cosaminog ycans". These polysacchn rides include
`chondroit in su]fat.e, dermatan su]fate, keratan sul-
`fute, hcporon aulfo.te, ond heporin.2
`
`10.1021/cr941199z CCX::: S30.00 ~ 1998 American Chemical Society
`Published on Web 11/26'1998
`
`ALL 2021
`PROLLENIUM V. ALLERGAN
`IPR2019-01505 et al.
`
`

`

`2664 Chemical Reviews, 1998, Vol. 98, No. 8
`
`Lapcik et al.
`
`Lubomfr Lapcfk was born in Topolna. Czech Republic, in 1937. In 1961
`he received his M.S. degee in chemical technology atthe Slovak Technical
`University in Bra1islava. He obtained his Ph.D. and D.Sc. degrees in
`1967 and 1987 from the same university. He became a professor in
`1987. In 1968-1969, he was granted a postdoctoral fellowship at the
`Institute of Physical Chemistry, University of Uppsala, Sw€<1en , under Proi.
`S. Claesson. Hok!ing 24 industrial patents and author of more than 120
`publications, he has lectured at several European universlies. His current
`interest is in the photochemistry and physical chemistry of polymers. He
`is a member of the Czech Chemical Society.
`
`Lubcmir Lapcfk, Jr. was born in Bratislava, Slovakia, in 1963. He studied
`physical chemistry at the Faculty o! Chemical Technology of the Slovak
`Technical University in Bratislava, where he received his M.S. degree in
`1986. In 1991, he received his Ph.D. degree in physical chemistry from
`the same university with a thesis titled "Hyaluromn - Structure, Properties
`and Applications" (scientific mentors, Prof. V. Kello and Prof. A. Stasko).
`In 1995, he became Associate Professor of Physical Chemistry of the
`Faculty of Chemistry, Technical University of Brno, Czech Republic, where
`he is currently lecturing on colloidal and surface chemistry and chemical
`kinetics. In 1991-1992, he was granted a postdoctoral fellowship under
`Prof. T. G. M. van de Ven and Prof. D. Argyropoulos in the Pulp and
`Paper Research Center, McGill Unrversity, Montreal, Canada, where he
`worked on problems associated with the stability of colloidal suspensions
`and solid-state 13C CP MAS NMR spectroscopy. In 1994, he spent six
`months as a visiting scientist on the Faculty of Pharmaceutical Sciences
`in Ghent, Belgium, where he worl<ed on problems of photochemical cross(cid:173)
`linking of hyaluronan under Prof. J. Demeester. Prof. A. Lauwers. and
`Pro!. D. De Keukeleire. He is a member of the American Chemical
`Society, the Spectroscof:ic Society ol Jan Marcus Marci, and the American
`Water Works Association. His research activ~ies have resulted in 23
`scientific publications and numerous poster and oral presentations at
`conferences. His current scienli!ic interest is !ocused on thin layer
`plasmacherncal modification and the characterization o! oriented solid
`biopolymer surfaces.
`
`Figure I. A tetrasaccharide from a HA chain. The torsion
`angles of the glycosidic linkages (<I> and '-P) are defined in
`the text. (Reproduced from ref 4. C-Opyright 1994 Ameri can
`Chemical Society.)
`
`HA is a linear, unbranched pol y m er. By chemical
`and enzymatic methods, Meyer and co-workers found
`HA to be composed of a repeating disaccharide t h at
`consists of N -acety J-o-gJucosamine (GJcNAc) and o(cid:173)
`g]ucuronic acid (GJcA) linked by a fJ 1-4 glycosi di c
`b ond.3 The disacch.arides ar e linked by fJ 1-3 bonds
`t o form the H A chain.
`In adclition to its presence in the v i treous b ody, HA
`occurs i n many l iving substt-ata su ch as the extra(cid:173)
`ceJluJar matrix and synovia] fluids. 5- 8 The isolation,
`purification, and i dentification of nearly pure HA has
`been the center of scientific interest for m any de(cid:173)
`cades. The procedure devel oped by Balazs was the
`first i ndustri ally applied extr acti on met h od for the
`iso]ation and purification of pharmaceutical grade
`HA.0 U mbilical cords an d rooster combs were frozen
`
`Ste!aan De Smedt was born in Geraardsbergen, Belgium, il 1967. He
`studied phannacy at the University of Ghent, Belgium, and received his
`M.S. degree in pharmaceutical sciences in 1990. As a scholar ol the
`Belgian Institute for the Encouragement of Scientttic Research in Industry
`and Agriculture. he enrolled in a Ph.D. program at the University of Glent,
`under the direction of Prof. J. Demeester. He studied meology at the
`Catholic University of Leuven. He received the Scott Blair Biorheology
`Award in 1993-1995 for his work on the structural characterization of
`hyaluronan solutions. To study diffusion phenomena in polymer solutions,
`he collaborated wi h Prof. Y. Engelborghs at the Laboratory o! Biomolecular
`Dynamics of the Catholic University of Leuven. In 1995, he joined the
`pharmaceutical development group of Janssen Research Foundation.
`Since 1997 he has been a postdoctoral !ellow of F.W.0.-Vlaanderen at
`the Laboratory o! General Biochemistry and Physical Pharmacy ol the
`University of Ghent. He is a member of the Controlled Release Societ~.
`the European Federation tor Pharmaceutical Sciences, the Belgian Society
`for Pharmaceutical Sciences, the European Society of Rheology, the
`Belgian Biophysical Society, and the Polymer Networks Group. He is a
`consultant to the Journal of Controlled Release and to Pharmaceutical
`Research. His current research interests include the mobility and
`interactions of ma::romolecular drugs in pharmaceutical polymer matrixes
`and oological polymer systems.
`
`t,0 destroy the cell membranes, and HA was extracted
`with water and precipi tated in organic so]vents such
`as, e.g., ethanol, chloroform, or cetyl pyridin.ium chlo(cid:173)
`ride. After purificati on of the extract, 0.5% protein
`impur i t ies remained, and the yi eld was 0.9 grams of
`HA per kilogram of the oiiginaJ material. Other
`
`

`

`Chemical Reviews, 1998, Vol. 98, No. 8 2665
`
`The bacteria] production of HA by Streptococcus
` and Streptococcus zooepidemicus 14 enabled it
`equi 1
`to be produced in larger quantities than could be
`achieved with the extraction methods. HA produced
`by S. equi has a lower molecular weight (MW) than
`does HA produced by S. zooepidemicu.s, which has a
`MW of about 1.8 to 2 x 106 Da with a yield of around
`4 grams of HA per liter of the cultivated solution. A t
`present, HA from various sources, w:ith different
`degrees of purity and molecular weights, is available
`for medical applications (section 5). The main im(cid:173)
`purities, depending on the source and purification
`method, are bacteria] endotoxines, chondroitin sul(cid:173)
`fates, proteins, nucleic acids, sodium chloride, and
`heavy metals. Water is usuaJly present between 5
`and 10% in the very hygroscopic powder or fibrous
`aggregate.
`No official requirements for HA used in pharma(cid:173)
`ceutical applications h ave as yet been established.
`It is hoped this review will serve the scientific
`committees that are developing pharmaceutical mono(cid:173)
`graphs. Attention will have to be paid to the devel(cid:173)
`opment ofworldw:ide accepted physicochernical meth(cid:173)
`ods to identify HA. Since the first conventional
`infrared (IR) spectroscopic rneasuxements on HA, 1"
`little attention has been given to the use of IR
`spectroscopy for its identification, although later
`studies showed that Fourier transform IR spectros(cid:173)
`copy might be a useful way to do this. 16- 18
`In
`addition to identi fica Liou methods, genera1ly accepted
`physicochemica] methods will be necessary to char(cid:173)
`acterize the macromolecular properties of HA batches.
`Although, ideally, the complete molecular weight
`distribution should be determined, the chai·acteriza(cid:173)
`tion of the macromolecular properties of HA batches
`might become possible with a standardized determi(cid:173)
`nation of the intrinsic vfacosity. 19
`It is beyond th e scope of this paper to review the
`extensive clinical and biological research that has
`been conducted on HA The main aim here is to
`describe the chemical and physicochemical features
`of tlus unique po]ysaccharide. Due to the medical
`interest in this polymer, this publication also reviews
`the physico-pharmaceuticaJ and medically applied
`HA research. By this "double view" on HA we hope
`to create closer links between the fundamental and
`the application-oriented HA research of the future.
`This might reveal new perspectives20 for this
`biopolymer, which is still expensive but, due to its
`exceptional hydrodynamic properties and its biocom(cid:173)
`patibiJity, hardly replaceable by other polymers. A
`secondary aim of this paper is to describe the main
`topics in the chemical reseiuch on HA derivatives.
`1'bis review considers the promising future of HA
`applications and how it may be based on chemical
`derivatives of HA.
`
`=1
`
`2. Macromolecular Character
`
`2.1. Polyelectrolyte Properties and Conformation
`The importance of the conformation and the inter(cid:173)
`actions of HA in solution led to basic research on this
`polymer in these areas. In the 1940s, Blix and
`SneUman studjed the size and shape of HA chains
`
`Hyaluronan
`
`Joseph Demeester was born in Ghent, Belgium. in 1951. He received a
`M.S. degree in pharmaceutical sciences from the University of Ghent in
`1974 ard earned a Ph.D. degree in pharmaceutical sciences in 1980
`under Prof. A. Lauwers. He became a professor at the same university
`in 1989 in the Laboratory of General Biochemistry and Physical Pharmacy.
`He was a laureate of the Belgian Royal Academy of Sciences in 1980
`and 1irst laureate of the Travel Grant of the Ministry of Education in 1981.
`He did postdoctoral research on light scattering and rheology at the Institute
`of Physical Chemistry of the University o1 Graz, Austria, w~h Prof. J.
`Schurz in 1985. He is a member of many scientific organizations, including
`the Biochemical society, and has been vice-president ot the Belgian
`Biophysical Society since 1994.
`In 1994, he became president of the
`International Center for Standards of the International Pharmaceutical
`Federation. His current research interests include the study of the action
`mechanism of different po1ymer-<1egradlng pharmaceutical enzymes such
`as hyaluronidases, proteases, and cellulases, the characterization of
`polymers such as hyaluronates and proteoglycans, and the controlled
`delivery of macromolecular drugs using biodegradable polymer hydrogels
`combined with enzymes. He is married to Riet Debruyne and has 1our
`children.
`
`Peter Chabrecek was born in Rakova, Slovakia, in 1955. He studied
`organic chemistry at the Comenius University in Bratislava where he
`received his M.S. degree in 1981. In 1986, he received his Ph.D. degree
`at the same university with a thesis titled ''Synthesis and Studies o1
`Benzothiazole Derivatives". In 1987-1988, he worked at the Research
`Institute o1 Preventive Medicine in Bratislava on the analytical characteriza(cid:173)
`tion of pesticide residues and metabolites. In 1989, he joined the research
`group of Prof. A. Blazej at the Institute of Biotechnobgy, Slovak Technical
`University, Bratislava, where he worked on the isolation, modification, and
`characterization of biopolymers. From Apri 1992 to June 1996, he was
`a postdoctoral fellow at the Central Research Laboratories of Ciba-Geigy
`in Basel, Switze~and. His work there focused on the surface modi1ication
`and characterization of polymeric materials for biological use, primarily
`for contact lenses. Currenfly. he is workirYJ for CSIRO, Australia, as a
`visiting scientist His research activities have resulted in 18 scientific
`publk:ations, 20 patents, and numerous conference presentations.
`isolation and purification methods have been de(cid:173)
`scribed by Galatik et al., Soltes et al., and Della Valle
`et a] . 10-12
`
`

`

`2666 Chemical Reviews, 1998, Vol. 98, No. 8
`
`from viti·eous humor by streaming birefringence.21
`They observed that HA chains were polydisperse
`molecules with a long ''particle length". In the early
`1950s, Ogston aud Stainer described how HA in
`solution behaved hydrodynamically like a large sol(cid:173)
`vated sphere containing a thousand times more water
`than organic materiai.22- 21 Despite the very s imple
`structure of the repeating disaccha1·ide (Figure 1) and
`about 60 years of intensive research on the properties
`of HA solutions, the conformation of HA in solution
`has been very djfficult to determine. As described
`in this section the conformation and the interactions
`of HA in the dissolved state are still controversial.
`
`2. 1.1. Polyelectrolyte Properties
`A typical polyelectrolyte pattern of viscosity was
`pointed out by Balazs and Laurent in the 1950s.25
`Upon complete ionization of the carboxylic groups
`within o-glucuronic acid, the charges are about 1 nm
`from each other. These charges ore influenced by the
`ionic strength and pH of the environment and, in
`turn, influence the shape of the chains and their
`interactions with surrounding molecules. In 1957,
`Laurent compared static light scattering and viscos(cid:173)
`ity results of sodium hyaluronate in water and
`cetylpyridinium hyaluronate in methanol. 26 He
`showed tht1t the radius of gyration, which was 200
`nm in the former solution and 120 nm in the latter,
`depends on the solvent, and he argued that the
`decrease of the raruus of gyration in cetylpyridinium
`hyaluronate was due to a collapse of the chain as the
`charges become neutralized. Cleland showed that
`HA chains contract with ·increasi ng ionic strength
`and decreasing pH, which indicates thei1· polyelec(cid:173)
`trolyte behavior.27 More recently, Fouissac etal. and
`Hayashi et al. studied the influence of the ionic
`content on the radius of gyration and the persistence
`length of HA with different molecular weights. 28•29
`Fouissac et al. showed the electrostatic expansion of
`HA chains could be wel1 described w:ithin the frame(cid:173)
`work of Odijk's modeJ3° by assuming a wormlike
`chain. However, Hayashi et al. indicated that the
`electrostatic contribution to the persistence length at
`a lower salt concentration is much larger than would
`be predicted from Odijk's model. In the presence of
`salts the dissociation constant (.K) of o-glucuronic
`acid 'on HA increases linearly as a fru1ction of Lhe
`degree of ionization (a.).31 While the pK of the
`polymer, as obtained by extrapolation to a.= 0, was
`estimated to be 2.9, the pK of the monomer n(cid:173)
`glucuronic acid is 3.23.32 The difference in pK was
`attributed to effects of substitution at carbon 4.
`
`2. 1.2. Conformation
`Although light scattering3~ and intrinsic viscosity34
`experiments in the 1950s and 1960s suggested that
`HA chains in solution have an expanded "somewhat
`stiff' random coil structure, Cleland showed that the
`size of HA varies with pH and salt concentration as
`would be expected for a fl.exible polyelectroh,te.27 In
`the 1970s, 2-, 3-, and 4-fold (both single and double)
`helical conformations of HA in the solid state were
`discovered from X-ray diffraction.30- 39 It was also
`shown that the helical form of HA in the solid state
`
`Laptik et al.
`
`, y
`..... °½fO
`OH
`'
`OM
`---~~ .. ~ ~o ...... o~ o~ - -(cid:173)
`HO~O~ -/--:7--d, .... tO
`) ······•·Oko
`J 01. ........
`
`0
`
`%
`
`(a)
`
`GICA
`
`GltNAC
`
`GICA
`
`GlcNAt
`
`· · · ·• •• ~CH,J
`
`OH
`
`OH
`'
`OH
`· ···vO
`-- -~~ "~0 : " • ••~~0~ - - (cid:173)
`
`0
`
`CH,
`
`' "t
`HO~O~ -.l-r,7--d .... to\-:--:--u
`j
`
`OH
`
`(b)
`
`1 ·•, •• _ ~O
`w"-et1,·o,~
`If
`Gl,A
`G !c;NAe
`O lcA
`ClcNAc
`Figure 2. Secondary etructure of HA in DMSO (a) and in
`DMSO containing water (b) as proposed by Heatley et al.
`The dotted lines indicate hydrogen bonds. Arrows indicate
`the glycol groups which are resistant to pe1-iodate oxidation
`(see text). (Reproduced with pe1·mission from ref 53.
`Copyright 1988 The Biochemical Society.)
`
`depends on the counterion type, pH, temperature,
`and extent ofhydrntion. The conformation of HA in
`the solid state was reviewed by Amott et al.40 The
`helical conformation of HA in the solid state raised
`the question whether any 01·dered form at all could
`exist under hydrated conditions. As discussed below,
`various conformations have been proposed.
`By nucJear magnetic relaxation (NMR) measure(cid:173)
`ments on IIA solutions, Da1-ke et al. identified two
`types of residues in HA chains.4J The relationship
`between the 1·elaxation times and confo1mational
`mobility showed that the1·e are two types of do~~in
`w:ith rufferent mobility. While one had the mob1hty
`of a flexible polymer, the other was so stiff that it
`had to contain cooperative structure. The stiff part
`represented 55-70% of the HA structure, ~nd_ t~ s
`proportion was not altered by changes 111 1oruc
`strength or temperatm·e, by addition of a denaturant
`such as urea, or by moderate changes in pH. There(cid:173)
`fore, they suggested that the_ stiff chain segrn~nts
`differed from the flexible cham segments by mmor
`covalent features. According to Darke et al. the stiff
`segments were composed of at least 60 disaccha1ide
`uuit-8. This was questioned by Mathews and Decker.42
`From viscosity data they showed that a significant
`degree of stiffness still exists in HA chains, even after
`reduction of the chain composition from 2500 to less
`than 60 disaccharide units.
`In the 1980s, Scott and colleagues continued to
`study intensively the conformational properties of
`dissolved HA oligomers by NMR which provided a
`physical proof of the existence of a structure that had
`been predicted from space-filling molecular models43
`and computer sirnu.1ations44 some years before. The
`HA conformation in solution was considered as an
`ordered structure in wlrich each dfaaccharide unit is
`twisted through 180° compared with those ahead and
`behind in the chain_ A 2-fold single helix was
`proposed as two turns bring back the original
`orientation .45-4.8 In dimethylsulfoxide (DMSO) Scott
`showed that there were hydrogen bonds between
`adjacent sugar units (Figure 2).45 The NMR work
`also showed e,,jdence for the results of Scott and
`Tigwell43 on the pe1iodate oxidation of HA in so~ution.
`These experiments sh owed the glycol group m the
`
`

`

`Hyaluronan
`
`Chemical Reviews, 1998, Vol. 98, No. 8 2667
`
`glucuronate residues (Figw·e 2) is oxidized 50-100
`times more slowly than the glycol group in similar
`glyco53mjnoglycans. A stable conformation that
`involves hydrogen bonds between the carboxylate,
`acetamido, and hydroxyl groups was postulated for
`the periodate resistance of dissolved HA. They also
`suggested that the extended hydrogen-bonded atTays
`down both "sides" of the HA chains result in consid(cid:173)
`erable rigidity of the polymer, which agreed with the
`earlier observations33,34,41 that HA in aqueous solu(cid:173)
`tions behaves like a rather stiff polvmer. It could also
`explain the dramatic reversible decrease of the
`viscosity of alkaline HA solutions12 as being due to
`the disruption of hydrogen bonds when participating
`protonated groups ionize and Jose their H atoms.
`Alkali-induced ionization of hydroxyl groups in HA
`was also proposed by Welti et al. and Bociek et al.
`based on 1H and 13C NMR spectra.49,00 The same
`view was offered by Ghosh et al. from static light
`scattering experiments performed to study alkali(cid:173)
`induced conformational contraction of HA chains.51
`NMR results reported by Cowman et al. on low
`molecular weight HA in water strongly indicated that
`the acetamide group was wrongly oriented to allow
`a hydrogen bond between the am ide proton and the
`carbo:xyl group of the adjacent uronic acid subunit.52
`Some years later, Scott and co-workers observed that
`the secondary structure ofHA, as established in "dry"
`DMSO (Figure 2), does change upon the addition of
`water.53 They found evidence for the replacement of
`the hydrogen bond between the amide proton and the
`carbo:xyl group by a single water molecule bridging
`both groups (Figure 2). From further investigations
`using molecular models of HA fragments, Scott's
`group revealed that in HA fragments lacking water
`b,·idges two conformations are sterically possible
`having the same type of hydrogen bonding but
`differing in dihedral conformational angles near
`acetamido, glycol, and carboxylate groups bound by
`hydrogen bonds.54 Molecular models of HA second(cid:173)
`ary structmes containing water bridges revealed that
`such bridges can join the acetamido and carboxylate
`groups in four ways which are sterically different.
`Besides extended hydrogen-bonded arrays, Scott
`and colleagues also observed large hydrophobic re(cid:173)
`gions, of about eight CH groups, on alternate sides
`of the single HA helices.45•48•53 Computer simulations
`and energy calculations confirmed that the HA 2-fold
`single helices in solutions may be energetically and
`sterically capable of extensive duplex forma tion
`driven by interactions between the hydrophobic
`"patches" of the HA chains (Figure 3),46.47 In a later
`study, molecular models revealed that hydrophobic
`contacts are possible only between HA chains lacking
`water bridges in the secondary structure.:s-t The
`hydrophobic patches were postulated not only to
`stabilize duplex formation but also to be a basis of
`the network-forming and lateralJy aggregating be(cid:173)
`havior ofHA.46 It was also suggested that they were
`the basis of HA interactions with lipid membranes
`and proteins (section 2.2.3). From 1H NMR spectros(cid:173)
`copy, gel permeation chromatography (GPC), and
`multiangle laser light scatte1ing, Ghosh et al. ob(cid:173)
`served that phosph olipids s uch as ilipalmitoyl phos-
`
`(cid:143)
`
`Figure 3. Scheme of a possible duplex formation between
`two HA chains. 46,47 The two pn1-ticipnting single HA helices
`are antiparnllel to each other. The dotted lines delineate
`each sugar unit, the circles represent acetamido, and
`squaresrepreseni carboxylate groups. The gray dotted ba1·s
`are the hydrop hobic patches s tretching along three suga,·
`units on altemate sides of the polymer chains.
`
`phatidylcholine (DPC) bind to HA.55 They suggested
`that DPC bindjng occurs by competition for the
`hydrophobic centers along the H A chains, as pro(cid:173)
`posed by Scott et al. 53
`Scott et al. showed that electrostatic repulsion
`between the negative charges may be countered not
`only by hydrophobic interactions but also by hydro(cid:173)
`gen bonding between t he HA chains.54 While most
`polar groups form intramolecular hydrogen bonds,
`two groups, namely t he hydroxymethyl and the
`oxygen a tom of the carboxylate group, are free
`(Figure 2). These groups could mediate intermolecu(cid:173)
`lar hydrophilic interactions in assemblies containing
`large n umbers of HA molecules. Molecular modeling
`showed that hydrogen bon ds between hydroxymethyl
`and carboxylate groups are possible only between
`antiparallel HA chains.54 Each disaccharide residue
`can form two hydrogen bonds, so that bonds on one
`side of the HA molecule alternate with analogous
`bonds on the other side. Such hydrogen bonds can
`join antiparallel HA molecules into sheets wruch are
`planar or curved. Based on hydrophilic and hydro(cid:173)
`phobic interactions, Scott's group proposed that
`several kinds of lateral contact may exist between
`such sheets which may result in the format.ion of
`highly ordered structures.
`In the 1990s, NMR work on th e repeating disac(cid:173)
`charide of HA,oG.57 HA oligomers,58 and high molec(cid:173)
`ular weight HA59,60 continued. The NMR results on
`HA oligosaccharides r eport,ed by Toffanin et al . did
`not suggest u significant role for cooperative hydro(cid:173)
`gen bonding involving the acetamido group in the
`determination of the HA conformation in water.58
`From 13C NMR experiments, Cowman and co-work(cid:173)
`ers confirmed evidence of significant conformational
`
`

`

`2668 Chemical Reviews, 1998, Vol. 98, No. 8
`
`differences between HA in the solid state and dis(cid:173)
`sohed HA.fiH,fiO Changes in the NMR spectra, upon
`dissolution of HA, were partia11y attributed to a
`change in the orientation of the acetamide group,
`possibly accompanied by a rotation about the /3 1-3
`and /3 1-4 glycosidic linkages. In aqueous media, each
`disaccharide showed two intramolecula r hydrogen
`bonds. Their data did not provide direct support for
`a stable water-mediated hydrogen bond across the fJ
`1-4 linkage as was obser ved by Scott and co-work(cid:173)
`ers.53 Cowman et al. also observed that greater
`conformational changes occur in t h e (3 1-3 lin kage
`than in the ,8 1-4 one.60 These conclusions disagreed
`with Holmbeck et al., who studied the glycosidic
`angles of HA octasaccharides and found that the fJ
`1-4 linkage shows a greater Oeicibility t han the f) 1-3
`linkage.4 On the basis of nuclear Overhauser en(cid:173)
`hancement (NOE) NMR combined with molecular
`modeling, they proposed that the confo1mation of an
`octasaccharide of HA in aqueous solutions is defined
`by the torsion angles of the glycosidic linkages and
`by the orientation of the side groups. ¢ 13 = 46° and
`'1113 = 24°, which are the inteq~lycosidicangles of the
`/3 1-3 linkage (Figure 1), were obtained fr:om r e(cid:173)
`strained molecular dynamics calculations. For the
`/3 1-4 linkage, two sets of values were consistent with
`the NOE restraints (<J,11 = 24° and W11 = -53° or <JJ14
`= 48° and 4J 14 = 8°).
`
`2.2. Hydrodynamic Behavior and Interactions
`2.2.1. Hydrodynamic Behavior
`The high viscosity of HA solutions and its impor(cid:173)
`tance in synovial fluid as well as in many medica)
`applications have for years stimulated the investiga(cid:173)
`tions on the hydrodynamic properties and mo)ecular
`interactions in HA solutions. The hydrodynamic
`behavior of HA solutions is so exceptional that most
`other polymer solutions fail to duplicate its viscoelas(cid:173)
`tic behavior.61
`Early studies ou the rheological properties of HA
`solutions and synovial fluid,22- 24,62- 66 which were
`reviewed by Balazs in 1974,67 showed that the
`viscosity is a complex function of polymer and salt
`concentration, pH and shear rate. With increasing
`shear rate, even at low shear rates and low concen(cid:173)
`trations (c), HA solutions exhibit a shear thimung
`behavior (Figure 4): the molecules deform and align
`in the streamlines of fl ow, which results in a decrease
`in the viscosity. This beh avior occurs when pushing
`HA solutions throu gh medical needles, for examp)e,
`which makes them easy to handle. 68
`Recently, Rin.ando's group studied the hydrody(cid:173)
`namic behavior of HA solu tions by focusing on the
`relation between the zero shear rate viscosity (rJo) and
`the concentration and molecular weight of bacterial
`HA. 69,70 Bacterial HA was used to avoid protein(cid:173)
`mediated aggregation of dissolved HA chains, which
`might occm- if an imal HA were used. They observed
`a greater dependency of rJo on the molecular weight;
`·110 ~ MW4 was found instead of 170 ~ MW"3.4 as would
`be expected for fl exible polymers. They also stressed
`that the critical concentration where the chains start
`to entangle was about 10 times lower than values
`
`~ , p
`105 "-......_.....__._._....,,,.l..L........,,....__L..UJ ........ _...... ........ LL. ......... -
`
`......... ~
`
`Lapcik et al.
`
`<1 • 0'°410\o •o .. •
`
`0
`
`1
`• •• ••
`
`• • • • • • • • -
`
`0
`
`. .
`, .....
`·.·.· ·.· .o:
`.... .. ... ' ... ,. .. ..
`. . . . . . . . .. . . .. . . . . .
`... . ... ; ; ; ~:;·:~~\(\/
`
`to
`
`1000
`
`10
`
`0.1
`
`c.o.ocxm...,
`c.o.oooe .....
`c.oJJClOS'o'nt
`C.0.00011lt'l"
`C-0.00,,S(IWI
`C.0.IJ0112 ....
`
`c-o=..(cid:173)
`~ ....
`
`c ........ ..,.
`C,O.OOOO,'¥fl'
`c.o,cG18 c,t'!'II
`
`.. ,
`
`0.01
`
`100
`i<•'1
`Figure 4. I nfluence of sh ear r a~ (y) on the specific
`viscosity (17,p) of HA solutions at different polymer concen(cid:173)
`trations. (Reproduced from ref 69. Copyright 1993 Ameri(cid:173)
`can Chemical Society.)
`
`found for flexible chains. Both phen omena were
`attributed to the stiffness of the HA chains. The
`dependency of t70 on the molecular weight and
`concentration differed sJightJy from the results of Wik
`et al.,71 who used animal HA. Rinaudo and co(cid:173)
`workers attributed this to a higher amount of protein
`in the animal HA. Experiments using bt1cterial HA
`"contaminated" with added pr oteins were not per(cid:173)
`formed and wou)d be u seful for confirming this
`position.
`The hydrodynamic and rheological behavior of HA
`mixtures composed of chains with different molecular
`weights was also studied.72 This study arose from
`dilliculties in obtaining bacterial HA with the same
`MW from different fermentations, which in turn
`preven ts the preparation of bacterial HA solution s
`having a reprodu cible viscosity. Given the viscosity,
`the HA concentration, and the solvent conditions,
`Rinaudo et al. studied how to mix I·IA batches with
`different MW. 72 They showed that the viscosity for
`HA mixtures at a given ionic strength :md temper(cid:173)
`ature is a function of the overlap parameter cl·17], f 17)
`being the intrinsic viscosity of the HA mixture which
`can be estimated from the addjtivity law
`
`(1)
`
`where i represents th e HA species with intrinsic
`viscosity (17]; and (JJ ; is the mass fraction of species i
`in the mixture.
`
`2.2.2. Interactions between HA Chains
`As represented in section 2.1.2, in the past decade
`important features were discovered on the interac(cid:173)
`tions between HA chains from NMR experiments and
`molecular modeling. Other views which were pos(cid:173)
`t ulated on the nature of the interactions between
`dissolved HA chains, which strongly influence the
`hydl'Odynamic properties of HA solutions, are repre(cid:173)
`sen ted in this section.
`In an early study on the viscoelastic properties of
`HA solutions, Gibbs et al. found that at pH 2.5 HA
`solutions showed an entirely different dynamic rheo(cid:173)
`logical behavior than at pH 1.5 and pH 7.64 At pH
`
`

`

`Hyaluronan
`
`Chemical Reviews, 1998, Vol. 98, Mo. 8 2669
`
`I I \
`/
`
`//
`
`< 7
`
`7 -20
`
`20-40
`
`> 40
`
`Figure 6. Model of Turner et a l. for self-association of HA
`segments as a function of the length of the segments.77
`
`concentration of HA segments consisting of 3-600
`disaccharide units to high molecular wei