Biorheology, Votlll, No. lI, pp.

2l1S-244, 1994
~ Pergamon e
Copyright 1994 Ebevier Science Ltd
Printed in the USA. All rights reserved
OOOf).lI55X/94 $6.00 + .00
0006-355X(94)EOOl6-N

VISCOELASTICITY OF HYALURONIC ACID WITH

DIFFERENT MOLECULAR WEIGHTS
Y. KOBAYASHI, A. OKAMOTO AND K NISHINARIt

Research Center, Denki Kagaku Kogyo Co., Ltd, Asahimachi, Machidashi,

Tokyo 194, JAPAN
tDepartment of Food and Nutrition, Faculty of Human Life Science, Osaka
City University, Sugimotocho, Sumiyoshiku, Osaka 558, JAPAN

ABSTRACT Storage and loss moduli of hyaluronic acid solutions with

different molecular weights were observed as a function of frequency in the
presence of sugars and salts. The hyaluronic acid solutions of higher
molecular weight (Mw > 17 x 106) fractions showed entanglement, whereas
lower molecular weight fractions did not. For inducing the entanglement of
molecular chains of hyaluronic acid, increasing the molecular weight of
hyaluronic acid was more effective than increasing the concentration of the
lower molecular weight fractions (Mw = 7. 8 x 10 5 ). Glucose, fructose,
galactose, and sucrose increased both storage and loss moduli, while NaCl,
KCl and CaCl2 decreased both moduli. It is suggested that sugars create
hydrogen bonds and strengthen the transient network. Cations shield the
electrostatic repulsion of hyaluronic acid molecules, and the polymer chains
are shrunk into compact coils from expanded stiffened coils.

Introduction
Hyaluronic acid (HA) is one of the main components of synovial fluid, acting
as a lubricant and shock absorber (Myers et al., 1966). It has attracted much
attention in the biomedical field because it is used in viscoelastic surgery
(Balazs et at., 1989). Viscoelasticity of HA solutions has been extensively
studied by many research groups (Schurz, 1967; Gibbs et al., 1968;
Morris et at., 1980; Tirtaatmadja et al., 1984; Bodmer et al., 1987;
Yanaki et al., 1990; De Smedt et al., 1993). It has been shown that synovial fluid
acts as a viscous liquid in low-frequency regions (corresponding to the slowly
moving joint) but shows an elastic behavior in high frequency regions
(corresponding to the rapidly moving joint). Synovial fluid's relaxation time
falls in the range between the normal motion of a knee, for example, and
motion which is rapid enough to induce trauma. Viscoelastic parameters are
decreased in the synovial fluids of patients with rheumatoid arthritis, which has
been ascribed to a decreased amount of high molecular weight HA, but this
was questioned by Dahl et al. (1985) and the decreased viscoelastic parameters
were attributed to the presence of a lower concentration of HA. On the other
hand, considerable variations in molecular weight distribution between

KEYWORDS: Hyaluronic acid; solution; molecular weight; viscoelasticity;

entanglement; sugars; salts

235
 236 Viscoelasticity of hyaluronic acid Vol. 31, No.3

individual synovial fluids from arthritic patients were found (Bjelle et al., 1982),
while other research groups (Stafford et al., 1964) found no correlation
between the intrinsic viscosity of HA in synovial fluids or the concentration of
HA and the clinical status of the diseases in the same group as rheumatoid
arthritis. Caster et al. (1966) suggested that decreased HA concentration
reflects the vascular aspects of synovial inflammation and that a decreased
molecular weight of HA is more directly related to altered function of the living
synovial cells. Therefore, this problem seems still to be in the stages of debate;
to help resolve this debate, it is important to know quantitatively the effects of
molecular weight and concentration on the viscoelastic parameters of HA
solutions.
Conformation of HA has been studied by light-scattering (Cleland, 1977;
Terbojevich, 1986) viscometry (Cleland, 1970), NMR (Welti et al., 1979), and is
believed to be a stiffened random coil, i.e., all the segments in the HA chain
are in continuous equilibrium between a stiff state and a flexible state. Some
local conformational ordering is maint.ained by hydrogen bonds that are
continuously forming and breaking. Effects of sugars and salts on the
viscoelasticity of HA solutions were studied in the present work.

Materials and Methods

High molecular weight HA samples extracted from the culture medium of
Streptococcus equi were purified in the laboratory of Denki Kagaku Kogyo Co.,
Ltd. Lower molecular weight HA samples were prepared from the high
molecular weight HA by the ultrasonic degradation method of Chabrecek et al.
(1991; Orvisky et ai., 1993). Molecular weights of the samples used in the
present work are 5.6 x 10 5 , 7.6 x 105 ,7.8 x 10 ,1.04 X 10 6 , 1.7 x 106 , 1.93 x 106 ,
and 1.97 x 10 6 . Molecular weights were estimated from intrinsic viscosity using
the Mark-Houwink parameters reported by Laurent et al. (1960).
In terms of viscoelastic measurements, the complex shear modulus,
G* = G' + iG", of HA solutions was observed as a function of angular frequency
by a CSL500 Controlled Stress Rheometer (Carrimed, England). The diameter
and angle of the cone were 4 cm and 2°, respect.ively. The temperature was
fixed at 5°, 20° and 25° C. The angular frequency was changed from 0.0065 to
80 rad/s. The effects of various sugars and salts on G' and G" were examined.

Results and Discussion

Storage modulus, G', and loss modulus, G", for 1% HA solutions as a function
of angular frequency are shown for different molecular weights in Fig. 1. As
observed by many research groups, the higher molecular weight fractions
(> 1.04 x 106 ) show a plateau region at higher frequency region where G'
predominates over G", which has viewed as the result of a cross-linked network.
Since molecular chains cannot disentangle during the short period of
oscillation, they behave as a temporary cross-linked network. At lower
frequencies, however, G" predominat.es over G', indicating that the cross-
linking is transient and not formed by covalent or permanent bonds but by
entanglement, ancI that molecular chains can disentangle and rearrange during
the long period of oscillation. Recently, the viscometry was carried out for HA
with weight average molecular weight 6.8 x lOS, and the zero shear viscosity was
plotted against the HA concentration (de Smedt et at., 1993). The relation
showed a st.eep upturn around 0.1 wt% which was taken as coil overlap
concentration. This is in agreement with our observation that HA molecular
chains do entangle at 1 % in all the HA fractions except those with the lowest
molecular weight (Mw = 5.6 x 105 )
Vol. 31, No.3 Viscoela.sticity of hyaluronic acid 237

a)
10'

10'

10'

"
~ 10'
(,
10. 1

10.2

10·~O·3 10'\ 10' 10' 10'

b) OJ I rad s·'
10' ---

10'

10'

~" 10'
Cl

- ~'~'0"7,-~'07.- --,...,0'----",0'

w Irad s·1
Fig. 1. Storage (a) and loss (b) shear moduli of 1 % hyaluronic
acid (HA) solutions with different molecular weights at 5° C as a
function of angular frequency. Numbers beside each curve
represent the molecular weight of HA.

10 2 . - - - - - - - - . - - - - - - - - - . - - - - - - - - - .

-
r,Q
c:
ca

100~------~--------~--------~
50 100 150 200
Mw/104
Fig. 2. Loss tangent G"/G' of 1 % HA solutions at ro = 0.7211
rad/s and at 5° C as a function of molecular weight.
238 Viscoelasticity of hyaluronic acid Vol. 31, No.3

a)

10'

10 '

10

i
10
1
10
"_, ~ ... -" ___ . _______. ._-'-_ ,_,_.,J
10:' ,0 ~ 10 10° 10 10'
b) /rad 5- 1
10 3

10)

~
. 10'
Cl

10°

10.2 10' 10° 10' 10'

OJ Irad 5- 1
Fig. 3. Storage (a) and loss (b) shear moduli of HA solutions as
a fUIlction of angular frequency at 20° C: 0, 3% HA
(Mw = 7.8 x 10 5 ); +, 2% HA solutions (Mw = 7. 8 x 10 5 );
0, 1.2% HA (Mw = 7.8 x 10 5 ); x, 1.0% HA (Mw = 1.93 x 10 6 ).
a)
10 3 ••... _ - _

10'
-1
b) w Irad 5
10 3

10'

10'

~'"
Cl 10 ~ 1.0 ,
05
00

10 '

10.2
10 1 1 0. 2 10"
--"-1~O 1 0'
10
OJ Irad 5. 1

Fig. 4. Storage. (a) and loss (b) shear moduli of 1% HA

(Mw = 1.70 x 1( 6 ) solutions with and without glucose as a
function of frequency at 5° C. Numbers beside each curve
represent the concentration of glucose in mol/I.
Vol. 31, No.3 Viscoelasticity of hyaluronic acid 239

fraction, in which Gil predominates over G' at all the frequencies examined and
no plateau region of G' was observed.
The loss tangent, tan 8 = G"/G', for 1% HA solutions with different
molecular weights is shown in Fig. 2 as a function of molecular weight at a
fixed angular frequency, CO = 0.7211 rad/s, which corresponds to joint
movements at ordinary walking speed. The value of tan 8 decreased
monotonically with increasing molecular weight, indicating that HA tends to a
more solid-like state with increasing molecular weigh t. The increase in
molecular weight gives rise to the increase in the probability of contact between
different molecular chains and also between segments in a single chain so that
the number of elastically active chains increases; hence, G' increases and tan 8
decreases.
Fig. 3 shows the angular frequency dependence of storage and loss moduli
for HA solutions of different concentrations with a lower molecular weight
fraction (Mw = 7.8 x 10 5 ) and 1.0% HA solutions with a high molecular weight
fraction (Mw = 1. 93 x 10 6 ). Curves of storage and loss moduli for different
concentrations did not cross each other, and increased with increasing
concentration of HA, while the curves for G' and Gil of 1.0% HA solution with
a higher molecular weight intersected those with a lower molecular weight
fraction. The difference in behavior between lower molecular weight fractions
and higher molecular weight fractions may be induced by a difference in the
entanglement of HA molecular chains. Even if the concentration is 3.0
(= 3. 0/1.0) times lower, a difference of 2.5 (= 1.93/0.78) times in molecular
weight is more effective for increasing the number of elastically active chains
and for promoting the formation of transient network. As is well known, curves
of double logarithmic plot for the storage modulus and angular frequency are
parallel for polymers with random coil conformation, such as polymethyl
methacrylate (Masuda et at., 1970) and polystyrene (Onogi et at., 1970), which
have different molecular weights.
Fig. 4 shows the angular frequency dependence of the storage and loss
modulus for 1% HA (Mw = 1.70 x 10 6 ) solutions with and without glucose.
Storage modulus, G', is smaller than loss modulus, Gil, at lower frequencies,
while G' is larger than Gil at higher frequencies. The addition of glucose
increased both G' and Gil at all the frequencies examined. The frequency
beyond which G' predominates over Gil shifted to lower frequencies with the
increasing concentration of glucose, indicating that the transient network
formation is promoted by the addition of glucose.
The specific intermolecular interactions in HA solutions were studied
using the competitive inhibition approach (Welsh et at., 1980). It was found
that on adding an equal concentration of hyaluronate segments (- 60
disaccharide units), all evidence of coupling was lost and the solution rheology
closely approximated that typical of isolated chains. It was also found that no
inhibition was observed with very short chain segments (4 disaccharide units).
We found that sugars promote the network formation. The strengthening
of the HA network by the addition of glucose might be realized by newly
created hydrogen bonds through glucose molecules or by the increase in
effective HA concentration through the immobilization of water molecules by
glucose molecules. The bound water determined by ultrasonic velocity
measurements for monosaccharides or oligosaccharides is about 0.3 ml/g
(Uedaira, 1989) while that for HA is about 0.77 ml/g (Suzuki et (Lt., 1970). The
immobilization of water molecules by added sugars is, therefore, not as
important as that by HA molecules. The newly created hydrogen bonds are
240 Viscoelcuticity of hyaluronic acid Vol. 31, No.3

a)
10'

10'

co
e:
(,
10'

10'

10';0.3 10.2 10" 10' 10' 10'

w Irads· 1
b)
10'~--~-~--------

10'

::1 ____._ _ _ _ ~~ ___~ -~

10 3 10. 2 10 ' 10° 10' 10

W Irad s·1

Fig. 5. Storage (a) and loss (b) shear moduli of 1% HA

(Mw = 1. 70 x 10 6 ) solu tions as a function of frequency at 50 C in
the presence of 2.0 M glucose (x), 2.0 M fructose (0), 2.0 M
galactose (+), and 1.0 M sucrose (0).
102r---------r----------,----------~--------_ __;

-
Il.
( !)

~
ca
10 ' G"

100~L--~~~--------L------~--------~
10.2 10·' 10° 10' 10 2
W Irad S-1
Fig. 6. Storage and loss shear moduli of 1 % HA solutions with
and without NaCl as a function of frequency at 50 C.
Concentration of NaCl: 0 , 0.01 M; +, (J.()5 M;
0, (1.l0 M; x, without NaC1.
Vol. 31, No.3 Viscoelasticity of hyaluronic acid 241

102r----------.----------.----------.----------~

100~~----LL~----------~--------~----------~
10.2 10" 10 0 10' 10 2
W Irad s -1
Fig. 7. Storage and loss shear moduli of 1 % HA solutions as a
function of frequency in the presence of 0.01 M NaGI (x),
0.05 M NaGI (0),0.01 M KGl (<», 0.01 M GaGl2 (0),
0.05 M GaGl2 (+) at 5° G.

possibly more important for strengthening the transient network of HA. It is

well known that sugars increase the storage modulus of agarose and 1(-
carrageenan gels (Watase et al., 1986; Nishinari et al., 1986; Nishinari et al.,
1992) probably because junction zones made by aggregated double helices are
stabilized in these gels by newly created hydrogen bonds. The transient
network structure of HA solutions is stabilized on adding glucose by newly
created hydrogen bonds which increase the number of elastically active
network chains.
Fig. 5 shows the frequency dependence of G' and Gil for 1 % HA
(Mw = 1.70 x 10 6 ) solutions with various sugars-glucose, fructose, galactose,
and sucrose. All these sugars increased G' and Gil at all the frequencies
examined. The order of increasing G' and Gil was as follows:
Galactose> sucrose> glucose> fructose, at lower frequencies. This order is in
agreement with the dynamic hydration number (nDHN( (galactose) = 16.6;
nDHN (sucrose) = 25.2; nDHN (glucose) = 18.6; nDHN (fructose)= 16.5), or the
number of equatorial hydroxyl groups in each sugar except for galactose
(Uedaira, 1989; 1990). This finding should be explored further.
The frequency dependence of G' and Gil for 1 % HA solutions with and
without NaGI is shown in Fig. 6. G' and Gil decreased at all the frequencies
examined with an increasing concentration of NaGI higher than 0.05 M.
However, a small amount of NaGI (0.01 M) increased both G' and Gil.
Hydrated water molecules surrounding anionic groups in HA molecules were
removed by sodium ions; as a result, the electrostatic repulsion between
anionic groups in HA molecules was enhanced. Hence, HA in the presence of
0.01 M NaGI showed larger G' and Gil values. This interpretation should,
however, be confirmed by more detailed studies in the future. The decrease in
G' and Gil with the increasing concentration of the added NaGI has been
242 Viscoelasticity of hyaluronic acid Vol. 31, No.3

attributed to the shielding of electrostatic repulsion between anionic groups in

HA molecules, which leads to the contracted forms of expanded coils. As is
well known (Morris et al., 1980; Watase et al., 1982), the addition of cations
increases storage and loss moduli for K-carrageenan gels where junction zones
are believed to be made by the aggregation of double helices. In this case, the
cations shield the electrostatic repulsion of sulfate groups in K-carrageenan
molecules, and as a result, the formation of helices and the aggregation of
helices is promoted. In hyaluronic acid, however, since the transient network
is made up of entangled stiffened coils, the contraction of coils induced by
cations leads to the decrease in shear modulus.
The effects of the addition of various salts on G' and Goo are compared in
Fig. 7. The order of decreasing G' and Goo was as follows:
CaC12 > KCl > NaCl. Calcium ions are structure-disordering ions, while
sodium ions are structure-ordering ions, from the viewpoint of the structure of
water (Uedaira, 1989; 1990). Calcium ions, therefore, shield the electrostatic
repulsion more directly than do sodium ions.

Conclusion
In conclusion, sugars such as galactose, sucrose, glucose, and fructose increase
both the storage and loss moduli of HA solutions, and this is attributed to the
increase in the number of elastically active network chains through the newly
created hydrogen bonds, which strengthen the transient network consisting of
entanglements. However, we found that salts such as CaCI2, KCl and NaCI
decrease bot.h moduli, which is ascribable to the shrinking of expanded
stiffened coils to compact coils. The cations in these salts shield the
electrostatic repulsion of carboxyl groups in hyaluronic acid molecules. When
a small amount of sodium ions was added to HA solutions, an exceptional
behavior was observed. This should be explored in the future.

References
BALAZS, E. A and LESHCHINER, E. A (1989). Hyaluronan, its
crosslinked derivative-hylan-and their medical application.
In: Cellulosics Utilization-Research and Rewards in Cellulosics. H. Inagaki
and G. O. Phillips Eds., London and New York, Elsevier Applied
Science, pp. 233-242.
BJELLE, A, ANDERSON, T. and GRANATH, K. (1982). Molecular
weight distribution of hyaluronic acid of human synovial fluid in
rheumatic diseases. Scand. J. Rheum. 12, 133-138.
BOTHNER, H. and WIK, O. (1987). Rheology of hyaluronate. Acta.
Otolaryngol. (Stockh.) Suppl. 442, 25-30.
CHABRECEK, P., SOLTES, L. and ORVISKY, E. (1991). Comparative
depolymerization of sodium hyaluronate by ultrasonic and enzymatic
treatments. Appl. Polym. Sci. 48, 233-24l.
CLELAND, R. L. (1977). The persistence lengt.h of hyaluronic acid:
An estimate from small angle x-ray scattering and intrinsic viSCOSity.
Arch. Biochem. Biophys. 180, 57-68.
CLELAND, R. L. (1970). Ionic polysaccharides. IV. Free-rotation
dimensions for disaccharide polymers. Comparison with experiment
for hyaluronic acid. Biopolymers 9, 811-824.
DAHL, L. B., DAHL, I. M. S., ENGSTROM-LAURENT, A, and
GRANATH, K. (1985). Concentration and molecular weight of
Vol. 31, No.3 Viscoelasticity of hyaluronic acid 243

sodium hyaluronate in fluid from patients with rheumatoid arthritis

and other arthropathies. Ann. Rheum. Dis. 44, 817-822.
DE SMEDT, S. C., DEKEYSER, P., RIBITSCH, V., LAUWERS, A, and
DEMEESTER,]. (1993). Viscoelastic and transient network properties
of hyaluronic acid as a function of the concentration.
Biorheology 30, 31-4l.
GIBBS, D. A., MERRILL, E. W., SMITH, K A., and BALAZS, E. A
(1968). Rheology of hyaluronic acid. Biopolymers 6, 777-791.
LAURENT, T. C., RYAN, M. and PIETRUSZKIEWICZ, A (1960).
Fraction of hyaluronic acid. The polydispersity of hyaluronic acid
from the bovine vitreous body. Biochim. Biophys. Acta 42, 476-485.
MASUDA, T., KITAGAWA, K and ONOGI, S. (1970). Viscoelastic
properties of poly (methyl methacrylates) prepared by anionic
polymerization. Polymer J 1, 418-424.
MORRIS, E. R., REES, D. A and WELSH, E.]. (1980). Conformation
and dynamic interactions in hyaluronate solutions.
J Mol. Biol. 138, 383-400.
MORRIS, E. R. , REES, D. A and ROBINSON, G. (1980). Cation-
specific aggregation of carrageenan helices: Domain model of polymer
gel structure. J Mol. Biol. 138, 349-362.
MYERS, R. R., NEGAMI, S. and WHITE, R. H. (1966). Dynamic
mechanical properties of synovial fluid. Biorheology 3, 197-209.
NISHINARI, K, WATASE, M., WILLIAMS, P. A., and
PHILLIPS, G. O. (1990). K-carrageenan gels: Effect of sucrose,
glucose, urea, and guanidine hydrochloride on the rheological and
thermal properties. J Agric. Food Chem. 38, 1188-1193.
NISHINARI, K, WATASE, M., KOHYAMA, K, NISHINARI, N.,
KOIDE, S., OGINO, K, OAKENFULL, D., WILLIAMS, P. A, and
PHILLIPS, G. O. (1992). The effect of sucrose on the thermo-
reversible gel-sol transition in agarose and gelatin.
PolymerJ 24, 871-877.
ONOGI, S., MASUDA, T. and KITAGAWA, K (1970). Rheological
properties of anionic polystyrenes. I. Dynamic viscoelasticity of
narrow-distribution polystyrenes. Macrornol. 3, 109-116.
ORVISKY, E., SOLTES, L., CHABRECEK, P., NAVAK, I., and
STANCIKOVA, M. (1993). Size exclusion chromatographic
characterization of sodium hyaluronate fractions prepared by high
energetic sonication. Chromatographia 37, 20-21.
SCHURZ, J. (1967). Empirische Zusammenhange zwischen den
FlieJ3kurven hochmolekularer Losungen und den Eigenschaften der
gel6sten Molekule, Teil II. u. Ill. Kolloid. Ztschrft. 155, 45-64.
STAFFORD, C. T., NIEDERMEIER, W., HOLLEY, H. L., and
PIGMAN, W. (1964). Studies on the concentration and intrinsic
viscosity of hyaluronic acid in synovial fluids of patients with rheumatic
diseases. Ann. Rheum. Dis. 23, 152-157.
SUZUKI, Y. and UEDAIRA, H. (1970). Hydration of potassium
hyaluronate. Bull. Chern. Soc. Japan 43, 1892-1894.
244 Viscoelasticity of hyaluronic acid Vol. 31, No.3

TERBOjEVICH, M., COSANI, A., PALUMBO, M., and

PREGNOLATO, F. (1986). Structural properties of hyaluronic acid in
moderately concentrated solutions. Carbohydr. Res. 149, 363-377.
TIRTAATMADjA, V., BOGER, D. V. and FRASER, I. R. E. (1984).
The dynamic st.eady shear properties of synovial fluid and of the
components making up synovial fluid. Rheol. Acta 23, 311-321.
UEDAIRA, H. (1989). Water in Biological Systems (in japanese),
pp. 69, Kodansha, Tokyo.
UEDAIRA, H. (1990). Hydrogels and water. In: The Science of Food
Hydrocolloids (in japanese), K. Nishinari and T. Yano, Eds., Tokyo,
Asakura Shoten, pp. 7-22.
WATASE, M., NISHINARI, K., WILLIAMS, P. A., and
PHILLIPS, G. O. (1986). Agarose gels: Effect of sucrose, glucose,
urea, and guanidine hydrochloride on the rheological and thermal
properties. J. Agric. Food Chem. 38, 1181-1187.
WATASE, M. and NISHINARI, K. (1982). Effect of alkali metal ions
on the viscoelasticity of concentrated lC-carrageenan and agarose gels.
Rheol. Acta 21, 318-324.
WELSH., E. j., REES, D. A., MORRIS, E. R., and MADDEN, j. K.
(1980). Competitive inhibition evidence for specific intermolecular
interactions in hyalauronate solutions. J. Mol. Biol. 138, 375-382.
WELT!, D., REES, D. A. and WELSH, E. J. (1979). Solution
conformation of glycosaminoglycans: Assignment of the 300-MHz
1H-magnetic resonance spectra of chondroitin 4-sulphate, chondroitin
6-sulphate and hyaluronate, and investigation of an alkali-induced
conformation change. Eur. J. Biochem. 94, 505-514.
YANAKI T. and YAMAGUCHI, T. (1990). Temporary network
formation of hyaluronate under a physiological condition.
I. Molecular-weight dependence. Biopolymers 30, 415-425.

Received 2 December 1993; accepted 14 January 1994.