Size selectivity of hyaluronan molecular sieving by extracellular matrix in rabbit synovial joints
1 Physiology, Basic Medical Sciences, St. George's Hospital Medical School, Cranmer Terrace, London SW17 0RE, UK
2 Division of Medicine, Imperial College, Hammersmith Hospital, London W12 0NN, UK
Abstract
In joint fluid the polymer hyaluronan (HA) confers viscous lubrication and greatly attenuates trans-synovial fluid loss (outflow buffering). Outflow buffering arises from the molecular sieving (reflection) and concentration polarization of HA at the synovial membrane surface. Outflow buffering declines if HA chain length is reduced, as in arthritis, and this has been attributed to reduced HA reflection. This was tested directly in the present study. Infused solutions of HA of 2200 kDa (HA2000, 0.2 mg ml–1) or 500 kDa (HA500, 0.2 mg ml–1) or 140 kDa (HA140, 0.2–4.0 mg ml–1) were filtered across the synovial lining of the knee joint cavity of anaesthetized rabbits at a constant rate, along with a freely permeating reference solute, 20 kDa fluorescein–dextran (FD20). After a priming period the femoral lymph was sampled over 3 h. Mixed intra-articular (I.A.) fluid and subsynovial fluid were sampled at the end. Fluids were analysed by gel exclusion chromatography. The trans-synovial concentration profile was found to depend on polymer size. The I.A. concentration of HA2000 increased substantially relative to infusate and the subsynovial and lymph concentrations fell substantially. For HA500 and HA140 the trans-synovial concentration gradients were less pronounced, and absent for FD. The reflected fractions for HA2000, HA500 and HA140 across the cavity-to-lymph barrier were 0.65 ± 0.05 (n= 10), 0.43 ± 0.09 (n= 3) and 0.19 ± 0.05 (n= 7), respectively, at matched filtration rates (P < 0.0001, analysis of variance). Reflected fractions calculated from HA I.A. accumulation or subsynovial dilution showed the same trend. The results demonstrate size-selective molecular sieving by the synovial extracellular matrix, equivalent to steric exclusion from cylindrical pores of radius 33–59 nm. The findings underpin the concentration polarization-outflow buffering theory and indicate that reduced HA chain length in arthritis exacerbates lubricant loss from a joint.
, 百拇医药
Introduction
In a synovial joint the cells that line the joint cavity form a discontinuous layer. Consequently, the intervening interstitial matrix is in direct contact with the synovial fluid of the joint cavity. This interstitial pathway constitutes the pathway for fluid drainage out of the joint cavity. A subpopulation of the lining cells, called fibroblast-related synoviocytes or type B cells, secrete hyaluronan and lubricin into the joint fluid, creating the viscous, lubricating synovial fluid.
, 百拇医药
Hyaluronan (HA) is a long, unbranched chain of repeating N-acetyl-D-glucosamine-D-glucuronic acid disaccharides (Fraser & Laurent, 1996). Its concentration in synovial fluid is higher than in any other adult tissue, being 2–4 mg ml–1 in healthy rabbit and human knees. Its weight-average molecular mass (Mw) is 2000 kDa in rabbit synovial fluid (Levick et al. 1996; Price et al. 1996). In arthritis both the concentration and chain length are reduced. In aqueous solution, hydrogen bonds stiffen the HA chain, which causes it to adopt an expanded coil configuration with an exceptionally large molecular domain (radius of gyration 100–200 nm). As a result, adjacent molecular domains overlap at 1 (mg HA) ml–1, creating marked chain–chain interaction and a quasi-infinite, dynamic network of loosely linked polymer chains (Day & Sheehan, 2001; Hardingham, 2004). This enables HA to act as a viscous, hydrodynamic lubricant of surfaces under low load, such as synovium on cartilage and synovium on synovium.
, 百拇医药
As well as its lubricating role, HA has a profound buffering effect on fluid drainage from the joint cavity into the subsynovial lymphatic system. When intra-articular fluid pressure is raised (e.g. by joint flexion), fluid begins to drain from the cavity through the interstitial spaces of the synovial lining, but this is soon countered by the osmotic pressure of HA that is sieved out and retained at the membrane surface. As a result, rises in joint pressure above 5 cmH2O produce remarkably little increase in trans-synovial fluid loss when HA is present (McDonald & Levick, 1995). This phenomenon, called outflow buffering, is physiologically important for three reasons. It conserves the tiny volume of synovial fluid (50 μl in a rabbit knee) during periods of sustained high intra-articular fluid pressure such as a maintained flexion; it prolongs by an order of magnitude the intra-articular working life of long-chain HA as a lubricant; and, correspondingly, it reduces by an order of magnitude the rate of HA biosynthesis needed to replace lost HA.
, 百拇医药
The concentration polarization hypothesis has been developed into quantitative steady state and non-steady state models to explain outflow buffering (Coleman et al. 1999; Lu et al. 2005). It was argued that HA accumulates in the fluid adjacent to the partially reflecting membrane (synovial lining interstitium), and the osmotic pressure of this layer opposes filtration. There is experimental support for two key postulates of this hypothesis. First, molecular sieving of HA has been confirmed by a sharp drop in HA concentration between intra-articular fluid and subsynovial fluid during trans-synovial filtration (Sabaratnam et al. 2003). Additional, supporting observations are that native HA has an exceptionally long residence half life in the joint cavity, 14–32 h (Denlinger, 1982; Brown et al. 1991; Coleman et al. 1997), and that HA accumulates in the joint cavity when a solution is filtered experimentally through the synovial lining (Scott et al. 1998a). Second, concentration polarization build up at the interface was confirmed by demonstrating its predicted consequence, namely a negative relation between reflected fraction and filtration rate (Lu et al. 2004; Sabaratnam et al. 2004).
, http://www.100md.com
Although HA can slowly permeate an interstitial matrix, both in joints and in the dermis, to reach the lymph (Brown et al. 1991, 1999; Fraser & Laurent, 1996), it is clear from the above findings that the synovial matrix creates sufficient steric hindrance to HA transport relative to water to cause molecular sieving. The matrix itself is a complex network of heterogeneous biopolymers, including sulphated and non-sulphated glycosaminoglycans, proteoglycans, glycoproteins and microfibrils (for review see Levick et al. 1996). The hydraulic drag of the biopolymers generates a high resistance to fluid escape (Levick, 1987; Scott et al. 1998b), and as a result the hydraulic resistance of the synovial lining is 20-fold higher than that of the areolar connective tissue in the subsynovial space (Scott et al. 2003). The subsynovial compartment contains a network of lymphatic capillaries that do not penetrate the synovium itself (Yamashita & Ohkubo, 1993; Xu et al. 2003) but collect the escaped trans-synovial filtrate and transport it into femoral lymphatic trunk vessels (Davies, 1946; Nagai, 1987; Reimann et al. 1989).
, http://www.100md.com
Experiments showed that the buffering of joint fluid loss depends on HA chain length and concentration. Buffering is lost if the bulk-phase HA concentration falls below 1 mg ml–1 (Scott et al. 2000a) or if the molecular mass falls below 500 kDa (Coleman et al. 2000) or if the molecular domain radius is reduced for a fixed polymer molecular mass (Scott et al. 2000b). Conversely, outflow buffering can be conferred on a non-buffering HA solution (0.75 mg ml–1) by increasing the effective HA domain size through chemical enhancement of the chain–chain interactions (Sabaratnam et al. 2002a).
, 百拇医药
The present study investigates the mechanism by which HA chain length affects outflow buffering. The concentration polarization theory predicts that outflow buffering will decline if HA reflection is a function of HA chain length. Preliminary evidence for the latter was obtained during the study of outflow buffering versus chain length. It was found that less HA accumulated in the joint cavity during infusions of 530 kDa, 300 kDa and 90 kDa HA than 2000 kDa HA (Coleman et al. 2000). The protocol, however, was designed primarily to study outflow buffering rather than sieving, and was not ideal for the latter purpose. In particular, the filtration rates were higher for low Mw HA than high Mw HA, and were necessarily changed many times during each experiment. This is a potentially confounding factor in a sieving study because, as emerged later, the reflected fraction for HA of constant chain length falls as filtration rate is increased (Sabaratnam et al. 2004; Lu et al. 2004). Moreover no measurements have been made of the sieving ratio, i.e. downstream (filtrate) HA concentration relative to upstream (filtrand) concentration.
, http://www.100md.com
The present aim was therefore to measure the effect of HA chain length on molecular sieving between the joint cavity and the subsynovial fluid or the lymphatic system at constant, matched filtration rates. The results have important implications for fluid and HA retention in arthritic joints, where HA chain length is reduced.
Methods
Overview
The joint cavity of a rabbit knee was cannulated and infused with HA solution at a controlled intra-articular pressure to generate a sustained, constant filtration across the synovial lining over several hours. A single filtration rate was studied per preparation. The Mw of the infused HA was 140 kDa or 500 kDa or 2230 kDa (referred to as HA140, HA500, HA2000, respectively). A reference solute, fluorescein–dextran (FD20), was included to determine the amount of joint lymph present in femoral lymph. After a priming period, femoral lymph was collected over 3 h and analysed for HA and FD20. The cavity-to-lymph sieving coefficient (transmitted fraction) was calculated from the infusate and lymph [HA]/[FD20] ratios. At the end, fluid was also aspirated from the subsynovial compartment and from the joint cavity, to assess the cavity-to-subsynovium sieving coefficient and the degree of intra-articular HA accumulation. The method has been described and evaluated previously (Sabaratnam et al. 2002b, 2003) and is summarized below.
, 百拇医药
Materials
Rooster comb HA (HA2000, 0.2 mg ml–1, Mw2230 kDa, radius of gyration 101–181 nm, Coleman et al. 1999) and fluorescein–dextran (30 μg ml–1, 20 kDa, Stokes–Einstein radius 3.1 nm) were purchased from Sigma Chemical Co. (Poole, UK). Shorter HA chains were prepared as described later. The HA and FD20 were co-administered in Baxter Ringer solution (mM: 147 Na+, 4 K+, 2 Ca2+, 156 Cl–; Baxter Healthcare Ltd, Thetford, Norfolk, UK).
, http://www.100md.com
A hyaluronan concentration of 0.2 mg ml–1, the lowest concentration reported for rheumatoid joints (Dahl et al. 1985), was used for all chain lengths because it is well below the critical concentration for HA2000 molecular domain overlap (0.8–1.3 mg ml–1, Coleman et al. 1999; Scott et al. 2000a and Results) and does not buffer outflow. The latter condition was necessary to achieve matched trans-synovial filtration rates for the various sizes of HA (34–37 μl min–1) and to enable filtration rates that yielded good harvests of lymph. It is impossible to achieve matching filtration rates at >1 mg ml–, because this concentration buffers the filtration rate to a very low level in the case of HA2000 (Scott et al. 2000a). In the case of HA140 some studies were also carried out at a higher concentration (4.0 mg ml–1, filtration rate again 37 μl min–1) for technical reasons, namely to facilitate the detection and separation of the 140 kDa HA peak from an adjacent peptide/papain peak during size-exclusion chromatography.
, http://www.100md.com
Animal preparation, joint cannulation and trans-synovial filtration
New Zealand white rabbits weighing 2–3 kg were anaesthetized with 30 mg kg–1 sodium pentobarbitone plus 500 mg kg–1 urethane I.V. and tracheostomized. Anaesthesia of sufficient depth to abolish the corneal blink reflex was maintained by 15 mg sodium pentobarbitone plus 250 mg urethane I.V. every 30 min. Only one knee joint was studied per animal. Following the procedure of Coleman et al. (1999), an intra-articular cannula was connected to a pressure transducer to record intra-articular fluid pressure Pj (±0.1 cmH2O). A second intra-articular cannula was connected to an infusion reservoir, the height of which regulated Pj, which in turn regulated the trans-synovial filtration rate. Flow from the reservoir into the joint cavity when intra-articular volume is constant depends on the rate at which fluid is draining away through the synovial lining. The rate of fluid uptake was measured using a photoelectric drop counter (5.6 μl) and chart recorder. A small correction was applied for viscoelastic creep of the cavity walls as previously described. Procedures conformed to UK animal legislation, and animals were killed by I.V. sodium pentobarbitone overdose at the end of the study.
, 百拇医药
Collection of mixed femoral lymph
Knee joint lymph drains into 2–3 major lymphatic trunk vessels in the femoral triangle. There are no intervening lymph nodes. The joint was cannulated and trans-synovial filtration initiated prior to lymphatic dissection to fill the lymphatics, then the femoral lymphatics were dissected clear of the adjacent femoral blood vessels and ligated proximally. After a further, undisturbed 1 h priming interval at a constant trans-synovial filtration rate, the largest lymphatic was transected and cannulated by the inside-out method of Sabaratnam et al. (2002b). Lymph was aspirated into a fluid trap. The trap was emptied and the lymph weighed (±1 mg) every 15 min for 2–3 h. The lymph output is well maintained over the 3 h period (Sabaratnam et al. 2002b).
, http://www.100md.com
Volume fraction of joint lymph in femoral lymph (VV)
We have shown previously that the amount of joint lymph in femoral lymph, expressed as the volume fraction Vv, is given by Vv=Ljoint/Lfemoral=Cfemoralref/Cjointref, where Ljoint and Lfemoral are joint and femoral lymph flows, respectively, and Creffemoral and Crefjoint are reference solute (FD20) concentration in femoral lymph and joint, respectively (Sabaratnam et al. 2002b). The subsynovial, intra-articular and infused FD20 concentrations do not differ significantly, showing that FD20 permeates the joint lining freely (Sabaratnam et al. 2003).
, 百拇医药
Hyaluronan transmitted fraction from joint to lymph (1 –Rlymph)
If a fraction Rlymph of the HA molecules in the bulk filtrand (infusate) is reflected by the synovium-to-lymph pathway, the transmitted fraction is (1 –Rlymph). The transmitted fraction is often referred to as a ‘sieving coefficient’ or ‘sieving ratio’ in microvascular and renal physiology. We have shown previously that the transmitted fraction can be calculated from the measured concentrations as CHAfemoral/(CHAinfusatexVV), where CHAfemoral and CHAinfusate are, respectively, the femoral lymph and infused HA concentrations (Sabaratnam et al. 2003).
, 百拇医药
Hyaluronan transmitted fraction from joint to subsynovium (1 –Rsyn)
Transport from the joint cavity to lymph involves two membranes in series, namely the synovium and the lymphatic capillary endothelium. A sample of the intervening subsynovial fluid was aspirated post mortem and analysed to verify that the molecular sieving occurs across the synovial lining, as shown in previous studies (Sabaratnam et al. 2003). The animal was killed by I.V. pentobarbitone, 1 ml Evans blue solution was injected into the knee cavity to visualize its boundaries, and the peri-articular tissue was dissected away to within a millimetre or so of the cavity border. A sample of the accumulated trans-synovial filtrate was aspirated through a catheter. The reflected fraction across the synovial lining, Rsyn, was calculated from the HA/FD20 concentration ratios in the subsynovial fluid and infusate as before (Sabaratnam et al. 2003).
, 百拇医药
HA reflection assessed by accumulation in joint cavity aspirate (Rasp)
Molecular sieving should cause upstream HA accumulation as well as downstream HA dilution. To assess this, the intra-articular fluid was mixed by 10 flexion–extension cycles at the end of the experiment and aspirated for analysis. Upstream reflection (Rasp) was calculated as described by Scott et al. (1998a), namely mass of HA reflected and retained in the cavity divided by the mass of HA presented to the membrane in the cumulative filtrand volume. Thus Rasp equals the increase in the intra-articular HA concentration x intra-articular fluid volume (i.e. the HA mass reflected and retained in the joint cavity) divided by the cumulative volume of fluid filtered during the experiment x infusate concentration (i.e. the HA mass in the total filtrand volume). The assumption that the 10 mixing cycles adequately dissipate the intra-articular concentration polarization layer was supported by the finding that Rasp and Rlymph did not differ significantly (Sabaratnam et al. 2003, 2004). The lymph method has the advantage of circumventing assumptions inherent in the Rasp method.
, http://www.100md.com
Analysis by high-performance liquid chromatography
Samples were centrifuged and diluted to 200 μl in Ringer solution. To prevent partial masking of small HA bands by endogenous albumin, samples were digested with 5.6 units papain (Sigma, UK) at 60°C for 1 h. This does not alter the HA molecular size (Coleman et al. 1997). The HA was quantified by size exclusion, high-performance liquid chromatography (HPLC) using a Waters 2690 separation module (Waters Ltd, Watford, UK), a TosoHaas TSK G6000 PWXL column (Anachem Ltd, Luton, UK) of nominal resolution 40–8000 kDa, and a Waters 486 ultraviolet absorbance detector set at 206 nm for HA analysis (Coleman et al. 1997). The injection volume was 50–100 μl and column flow 1 ml min–1 Ringer solution. Calibration curves for concentration were linear from 3 μg ml–1 to 400 μg ml–1 for each Mw. Sample Mw was estimated from the mean retention time, which was calibrated using HA standards of Mw 210 kDa to 5500 kDa generously donated by Dr O. Wik (New Pharmacia, Uppsala, Sweden) and characterized by laser light scattering. The HA calibration curves have been published previously (Coleman et al. 1997). Fluorescein–dextran was analysed using the same HPLC column and an in-line Waters 474 SATIN fluorimeter set to an excitation wavelength of 475 nm and emission wavelength 530 nm. Minimum detection level was <0.3 μg ml–1.
, 百拇医药
Preparation of HA samples of reduced chain length
The commercial rooster comb HA had a Mw of 2230 ± 50 kDa as indicated by HPLC retention time (Fig. 1A). Its polydispersity has previously been characterized using the ratio Mw/Mn= 2.3, where Mn is number-average molecular mass (970 kDa by osmometry, Coleman et al. 1999). Samples of Mw140 kDa were produced by sonicating 35 ml aliquots of 4 mg ml–1 rooster HA2000 for 16 min at 10 μm amplitude in a Soniprep (MSE Scientific Instruments, Crawley, UK), following the protocol of Coleman et al. (2000). Evaporation during sonication was measured by weight change and replaced by water addition. The HPLC retention time increased to 8.86 ± 0.03 min (n= 12), corresponding to a Mw of 140 ± 8 kDa (Fig. 1A). As shown in previously published chromatographs, the sonication process has little effect on the spread of the HPLC peak, indicating that polydispersity is not altered significantly (Coleman et al. 2000). Since the unit disaccharide has a mass of 379 Da and length 0.95 nm, the average HA140 chain comprised 369 disaccharides and was 0.35 μm long, in contrast to 5884 disaccharides and 5.6 μm long for HA2000. A molecular mass of 140 kDa was chosen because we showed previously that sonicates of 88–305 kDa do not buffer the trans-synovial filtration rate (Coleman et al. 2000). Likewise, 500 kDa at 0.2 mg ml–1 does not buffer outflow, thought it does at 3 mg ml–1.
, 百拇医药
A, HPLC chromatographs for HA2000, HA500 and HA140 at 0.2 mg ml–1. The cited retention times are the means for the series (±S.E.M.). B, plots of reduced viscosity (see Methods) as a function of concentration to determine the intrinsic viscosity [] of each hyaluronan preparation. Lines fitted by linear regression analysis.
Samples of intermediate Mw can be produced by shorter sonication periods, but this proved unnecessary. A commercial sample of hyaluronan that had been stored for some time proved to have a Mw of 500 ± 60 kDa, as determined by the HPLC retention time of 8.30 ± 0.05 min (n= 12) (Fig. 1A). This was ideal for the study because its Mw was approximately four times that of the smaller preparation and one-quarter that of the larger preparation.
, http://www.100md.com
Viscometric characterization of size of hyaluronan molecular volume domains
Samples were analysed by viscometry to determine the effective molecular domain volume and the concentration at which adjacent molecular domains overlap (C*). Intrinsic viscosity, [], the volume occupied by 1 g solute at infinite dilution (ml g–1), is a sensitive index of molecular domain size. Intrinsic viscosity was determined by linear extrapolation of a plot of the logarithm of reduced viscosity redversus concentration C; red is (– 1)/C, where is relative viscosity. Viscosity relative to the solvent (Ringer solution) was measured using an Ostwald viscometer for solutions of low viscosity and negligible shear dependence, or a rotational rheometer for solutions of high viscosity with a marked shear-rate dependence (HA2000 at >1 mg ml–1). The Haaske RS150 cone-on-plate rheometer (Carl Stuart, Leek, UK) measured apparent viscosity (shear stress/shear rate) at 25°C via a set of automated step shear rates from 0.12 s–1 to 500 s–1.
, 百拇医药
The quasi-spherical domain of solvent occupied by a single polymer chain is conventionally characterized by the radius of gyration, Rg (root mean square of polymer segment distance from the molecular centre of gravity), which is related to []. For a neutral, flexible polymer the self-avoiding random walk model of Flory (1971) gives Rg3=MW[]/8.84NA, where NA is Avogadro's number. An alternative, empirical expression specific to HA, which is not neutral but carries one negative charge per disaccharide, is Rg3= 0.025MW0.6S–0.08, where S is the molar salinity of the solvent (Johnson et al. 1987).
, 百拇医药
de Gennes, (1979) defined the ‘critical concentration’C* for a polymer as the concentration at which the molecular domains overlap to produce a statistically homogenous, entangled network (the semidilute regime). This leads to the expression C*[]= 2.1 (McDonald & Levick, 1995), which was used to estimate C* for each preparation. A direct experimental assessment of domain interaction was also obtained by plotting the logarithm of specific viscosity (– 1) versus the logarithm of HA concentration; a sharp increase in the slope of this plot marks the onset of significant chain–chain interaction (Wik & Wik, 1998; Scott et al. 2000a).
, 百拇医药
Statistical analysis
Means are followed by S.E.M. The t test was used for paired, unpaired and one-sample comparisons as appropriate. One-way ANOVA with Tukey's post hoc test was used for comparison of three or more sets of results, and two-way ANOVA was used to compare results subject to two variables (e.g. time and polymer size). Lines were fitted by linear regression analysis. All tests were as implemented in Graphpad Prism (San Diego, CA). Significance was accepted at P 0.05.
, http://www.100md.com
Results
Viscometric assessment of HA domain sizes
For HA140 the relative viscosity was 1.11 at 0.2 mg ml–1 and 1.94 at 4 mg ml–1. The intrinsic viscosity [] extrapolated from a ln(red) versus C plot over the range 0.0063–4.000 mg ml–1 was 334 ± 42 ml g–1 (Fig. 1B). Corresponding values for HA500 were = 1.17 at 0.2 mg ml–1 and []= 646 ± 49 ml g–1. The HA2000 values were = 1.53 at 0.2 mg ml–1 (shear rate 103 s–1) and []= 2455 ± 318 ml g–1 (Fig. 1B). Table 1 summarizes these and other key biophysical parameters for the three HA preparations.
, 百拇医药
The Flory radius of gyration was 21 nm for HA140, 39 nm for HA500 and 101 nm for native rooster comb (HA2000). Corresponding values from the Johnson et al. (1987) salinity formula (see Methods) were 34 nm, 72 nm and 177 nm radius.
The de Gennes (1979) overlap concentration C* was 6.29 mg ml–1 for HA140, 3.35 mg ml–1 for HA500 and 0.86 mg ml–1 for HA2000. All the infusates in the present study were below C*, and thus in the dilute regime where molecular domains do not overlap. Plots of log (specific viscosity) versus log (concentration) confirmed this. For HA2000, the slope increase denoting onset of molecular interaction occurred at 1.30 mg ml–1, a result close to the value of 1.35 mg ml–1 illustrated in Scott et al. (2000a). HA140 showed no slope change up to the highest concentration explored, 4 mg ml–1. With HA500, any slope change was at 2 mg ml–1.
, 百拇医药
Lymph composition versus time for HA and FD
Figure 2 illustrates the near constancy of femoral lymph composition over the 3 h collection period. The changes in HA140 concentration and FD20 concentration with time were small, and the regression slopes were not statistically significant (P 0.38). We have previously demonstrated similar findings for HA2000 (Sabaratnam et al. 2003). The mean HA140 concentration in the lymph, 1.83 ± 0.13 mg ml–1, was 45.8%± 3.4% of the infused concentration (n= 7), whereas the FD20 concentration, 19.7 ± 1.2 μg ml–1, was proportionately higher at 65.6%± 4.1% of the infused concentration (n= 7). Two-way ANOVA of these results confirmed that the percentage of FD20 in the lymph was significantly higher than the percentage of HA140 (P < 0.001), and that time did not significantly influence the values over 3 h (P= 0.83). These results show that HA undergoes some degree of selective molecular sieving, even at the reduced chain length of 140 kDa.
, 百拇医药
A, lymph HA140 concentration versus time; slope of relation not significantly different from zero (0.0036 mg ml –1 min–1, P= 0.37). Broken line shows infused concentration. B, corresponding FD20 concentrations; slope of relation not significantly different from zero (0.033 μg ml –1 min, P= 0.38). C, ratio of mean HA140 concentration to mean FD concentration versus time: concentrations were normalized as percentage of the infused concentration. Linear regression line ±95% confidence intervals; slope 0.0003 min–1 not significant (P= 0.36). The mean ratio, 0.69 ± 0.01 (n= 8), was significantly less than 1 (P < 0.0001, one-sample t test), demonstrated selective molecular sieving of HA140 across the cavity-to-lymph barrier.
, 百拇医药
The percentage FD20 value represents Vv, the joint lymph volume fraction in femoral lymph under the experimental conditions. Vv is a positive function of trans-synovial filtration rate (Sabaratnam et al. 2002b). In the present study the filtration rates did not differ significantly for the three molecular sizes investigated, being 37.3 ± 1.9 μl min–1 (range 28–43 μl min–1) for HA140 at 4 mg ml–1; 37.7 ± 2.0 μl min–1 (range 33–40 μl min–1) for HA140 at 0.2 mg ml–1; 36.9 ± 1.8 μl min–1 (range 35–40 μl min–1, n= 3) for HA500; and 34.4 ± 2.4 μl min–1 (range 28–44, n= 9) for HA2000.
, 百拇医药
Concentration drop across the synovial lining is related to molecular size
The effect of Mw on the concentration of biopolymer in the four analysed fluid compartments (infusion line; mixed aspirate from the cavity at the end; subsynovial fluid at the end; mean femoral lymph) is plotted out in Fig. 3A. Concentrations were expressed as a percentage of the infused concentration to facilitate comparison with FD20. Figure 3B shows the same results but with anatomical location rather than Mw as the horizontal axis, to demonstrate the striking change in the concentration profile across the joint lining as HA chain length increases.
, 百拇医药
Concentrations are normalized as a percentage of the infused level. FD values are the mean ±S.E.M. for all experiments; for HA140 n= 7 joints; for HA500 n= 3 joints; for HA2000 n= 9 joints. A, graded relation between concentration of polymer in a given fluid compartment and the weight-average molecular mass of the HA, Mw. B, change in concentration profile across joint lining for polymers of increasing molecular size. Same results as above. In, infusate; Cav, joint cavity; Ss, subsynovium; L, femoral lymph.
, http://www.100md.com
The concentration profile showed an obvious, graded dependence on molecular size. Concentration fell sharply across the synovial lining for the largest polymer, HA2000; it fell by intermediate amounts for HA500 and HA140, and not at all for FD20. For HA2000 the concentration in the cavity increased substantially, reaching 253%± 18% of the infused level in the mixed terminal aspirate (n= 9), while its concentration in the subsynovial fluid and lymph fell to 23.9%± 9.8% and 15.5%± 2.3% of the infused level, respectively. The profile demonstrated marked molecular sieving of HA2000. For HA500 the change in concentration across the membrane was less marked: cavity aspirate 176%± 31%; subsynovial fluid 76.0%± 6.6%; femoral lymph 34.3%± 13.6% (n= 3). For HA140 the gradient was even smaller (cavity aspirate 124%± 6%; subsynovial fluid 92.5%± 7.9%; femoral lymph 45.9%± 7.1%; n= 7). For FD20 the profile across the synovial lining was flat, as expected for a freely permeating reference solute. The end-aspirate FD20 concentration showed no significant increase, being 29.59 ± 0.41 μg ml–1 or 98.6% of infused concentration in the seven HA140 experiments, and 100.6%± 2.3% for the entire study. Similarly, the subsynovial FD20 concentration was not significantly reduced (95.8%± 4.8%; P > 0.05, one-way ANOVA). Only in femoral lymph was the FD20 concentration reduced significantly, namely to 52.6%± 6.2% (whole series, P= 0.003, one-way ANOVA). This defined Vv, the dilution of the joint lymph by skin and muscle lymph in the femoral lymphatics (see Methods).
, http://www.100md.com
The differences between the terminal intra-articular concentrations of HA2000, HA500 and HA140 were statistically significant, thereby demonstrating that intra-articular HA accumulation during trans-synovial filtration increases with chain length (P < 0.0001, one-way ANOVA). The differences between subsynovial concentrations were likewise significant, showing that subsynovial HA concentration declines with increasing chain length (P < 0.0001, one-way ANOVA). The same was true for lymph HA concentration (P < 0.005, one-way ANOVA). The results thus showed a clear, consistent pattern; the greater the chain length, the greater the upstream accumulation of polymer and the lower the downstream concentration. It can be concluded that HA of reduced molecular mass is less well retained in the joint cavity than HA of high molecular mass.
, 百拇医药
Magnitude of reflected fractions for different hyaluronan chain lengths
The HA reflected the fraction across the composite cavity-to-lymph barrier, Rlymph, was calculated from the lymph HA and FD20 concentrations; see ‘Methods’. For HA2000 Rlymph averaged 0.65 ± 0.05 (n= 10 rabbits) at the concentration and filtration rate used in this study. (Different values are obtained under different filtration conditions; Sabaratnam et al. 2004; Lu et al. 2004). For HA500 Rlymph fell to 0.43 ± 0.09 (n= 3), and for HA140 it fell further to 0.19 ± 0.05 (n= 7) (Fig. 4). The reflected fraction was thus >3 times bigger for HA2000 than HA140 (P < 0.0001, one-way ANOVA). Even for HA140, however, Rlymph was significantly greater than zero (P < 0.05, one sample t test).
, 百拇医药
See text for definitions. Mean ±S.E.M.; n as for Fig. 3. The effect of chain length on reflected fraction is statistically significant in each case (see text).
A similar, graded dependence of reflected fraction on molecular size was demonstrated by Rsyn, which is calculated from the subsynovial HA and FD20 concentrations (Fig. 4) (P= 0.0001, ANOVA). The graded effect of molecular size on reflection was also confirmed by Rasp, which is calculated from upstream HA accumulation rather than downstream reduction (P < 0.0001, ANOVA). The numerical values for the three estimates of HA reflection, Rlymph, Rsyn and Rasp are summarized in Table 1, along with the global averages of the three reflected fractions. The results demonstrate a positive, monotonic relation between reflected fraction and the radius of gyration of the hydrated molecular domain; see Discussion. The tendency of Rasp to be a little lower than Rlymph or Rsyn could be partly due to retention of a small amount of sieved HA at the synovial surface despite the terminal mixing cycles.
, 百拇医药
For technical reasons (Methods) the infused concentration was set at 4 mg ml–1 in the main seven experiments with HA140, whereas it was lowered to 0.2 mg ml–1 for HA500 and HA2000 to prevent outflow buffering from interfering with the matching of filtration rates. To test whether HA140 concentration might have influenced the results, HA140 was infused at 0.2 mg ml–1 in three further studies. It was not possible to analyse the filtrate HA140 concentrations reliably in these three studies for technical reasons (overlap of the small HA140 chromatograph peak with other peaks), but it was possible to measure the accumulation of reflected HA140 in the joint cavity and thus calculate Rasp. The HA140 concentration in the terminal aspirate increased to 118%± 12% of the infused concentration, an increase similar to that observed for HA140 infusates at 4 mg ml–1 (124%± 6%). Rasp averaged 0.028 ± 0.017 for HA140 at 0.2 mg ml–1 (n= 3), compared with 0.052 ± 0.012 at 4.0 mg ml–1 (n= 7) (P= 0.31, unpaired t test). These results demonstrate that the low reflected fraction for HA140 relative to HA500 and HA2000 was not due to differences in the infused HA140 concentration.
, 百拇医药
Chromatogram retention times for reflected and transmitted HA chains
The retention time of the HA chromatogram peak, tret, is negatively related to average molecular mass over the range of interest (Fig. 1A and Coleman et al. 1997). Since the infused preparation was polydisperse (see Methods), tret was measured in the sieved and reflected samples, to assess whether the longer HA chains in the mixture might be selectively separated from the shorter ones during trans-synovial filtration. The HA2000 experiments provided the most favourable conditions for detecting within-sample selective molecular sieving, because reflection was most pronounced with HA2000 (Figs 3 and 4). In every joint infused with HA2000, tret for HA in the end-experiment aspirate from the joint cavity was smaller than that of the paired infusate tret (Fig. 5). This demonstrates a preferential retention of the longer chains in the joint cavity (P < 0.001, paired t test). In keeping with this, lymph tret increased (indicating a greater proportion of shorter chains), but the change did not reach statistical significance due to an increase in variance. The tret values were infusate 7.37 ± 0.05 min; end-experiment aspirate 7.21 ± 0.06 min; subsynovial fluid 7.44 ± 0.19 min; lymph 7.51 ± 0.14 min (n= 9 animals). A similar pattern, namely reduced aspirate tret and increased lymph tret, was noted by Sabaratnam et al. (2003).
, http://www.100md.com
The reduced HPLC retention time of HA2000 aspirated from the joint cavity at the end of the filtration period indicates selective intra-articular retention of the longer chains of the polydisperse HA2000 preparation (n= 9). Size-related separation was not detected within the HA140 preparation (n= 10) or HA500 preparation (n= 3).
In the case of HA140, which experiences relatively little molecular sieving (Figs 3 and 4), tret did not differ significantly between infusate, end-experiment aspirate, subsynovial fluid and lymph (Fig. 5) (P= 0.18, one way ANOVA, n= 10 preparations). The tret for HA500 likewise showed no significant differences between compartments. tret for FD20 was essentially identical in all four fluid compartments, as expected for a freely permeating solute (infusate 11.07 ± 0.03 min, end-experiment joint aspirate 11.02 ± 0.05 min, subsynovial fluid 11.03 ± 0.06 min, lymph 11.09 ± 0.02 min; pooled results from all experiments, P= 0.80, one-way ANOVA).
, http://www.100md.com
The tret results thus provided limited evidence for selective synovial permeability to chains of differing size within a heterodisperse HA sample of physiological average molecular size.
Discussion
The principal new finding was that the permeation of HA through synovial interstitial matrix is a graded function of HA molecular size. We discuss below the earlier work in this area, the mechanisms involved, the relevance of the findings to the concentration polarization hypothesis, and their pathophysiological significance.
, 百拇医药
Previous evidence indicating synovial molecular sieving
The quantitative results here are in broad agreement with a histological assessment of HA permeation by Asari et al. (1998). They found that intra-articular fluorescein-labelled HA of Mw 2300 kDa hardly penetrated the synovial lining of the dog knee, whereas 840 kDa HA penetrated it more readily. Size-selective molecular sieving by synovium is also indicated by the observation that the clearance of large proteoglycans and radio-colloid from the joint cavity is slower than that of albumin (Page-Thomas et al. 1987; Reimann et al. 1989).
, http://www.100md.com
Coleman et al. (2000) studied the relation between intra-articular fluid pressure and trans-synovial volume flow in the presence of 2000 kDa, 530 kDa, 300 kDa and 90 kDa HA. Although they did not collect subsynovial fluid or lymph, they noted that HA accumulated in the joint cavity over the course of the study in proportion to its molecular size, as found here (Fig. 3). In this earlier work, however, filtration rate varied both during each study and between classes of HA. The Rasp values reported by Coleman et al. (2000) were 0.79 for HA2000 at a mean filtration rate of 3 μl min–1, 0.25 for HA 530 at 10 μl min–1 and 0.12 for HA90 at 15 μl min–1. Since it was later shown that the HA reflected fraction falls when filtration rate is increased (Sabaratnam et al. 2004), the findings of Coleman et al. (2000) were potentially open to an alternative interpretation. In the present study, however, the filtration rate was held constant throughout the experiment and was matched for each HA preparation. The new results confirm that the increase in Rasp with increasing Mw is independent of the effect of filtration rate on Rasp.
, 百拇医药
Quantitative relation between reflected fraction and molecular domain size; equivalent pore size of extracellular matrix
The relation between HA reflected fraction and molecular domain radius is plotted in Fig. 6, with additional results for HA, 2000 kDa dextran and 67 kDa albumin from previous work (see figure legend). The simplest explanation for the relation is that large solutes experience partial steric exclusion from the water space in the extracellular matrix. The greater the solute domain size relative to the effective pore size in the matrix, the greater the solute exclusion and hence the greater the reflection of the solute during filtration. This hypothesis can be assessed quantitatively as follows.
, 百拇医药
Molecular radius is expressed as the Flory Rg value except in the case of albumin (Stokes–Einstein radius 3.6 nm). with S.E.M. bars are mean of HA reflected fractions from current study. without S.E.M. are Rasp values from Coleman et al. (2000), measured at lower filtration rates and a higher HA concentration (3.6 mg ml–1). is the reflected fraction for plasma albumin (Coleman et al. 2000). is reflected fraction for 2000 kDa dextran (Scott et al. 2000b). is reflection coefficient for HA2000 at 0.2 mg ml–1 extrapolated from the relation between reflected fraction and filtration rate; bars are 95% confidence limits (Sabaratnam et al. 2004). Dashed line is the theoretical relation between reflection coefficient and solid sphere radius rs (lower, offset x scale) based on steric exclusion in cylindrical pores of radius 33 nm (Anderson & Malone, 1974). To plot the polymers it is assumed that, due to their chain flexibility, effective polymer rs= 0.26Rg (see text). Dotted line is theoretical based on steric exclusion in a randomly orientated molecular fibre matrix (Ogston, 1970) of polymer concentration 11.5 mg ml–1 (Scott et al. 2003).
, 百拇医药
A simple approach to characterizing the irregular aqueous channels through an extracellular biopolymer matrix in terms of a single parameter is to work out the radius of a cylindrical, water-filled pore that would have the same excluding properties – the ‘equivalent cylindrical pore’ model. For a pore of radius rp and a solid, neutral spherical solute of radius rs, the solute reflection coefficient can be related to rs/rp through the solute partition coefficient , which is the complement of the steric exclusion fraction for solute in a narrow pore (Anderson & Malone, 1974; Curry, 1984) -
, 百拇医药
where = (1 –rs/rp)2. The term 1 –rs/rp describes the radial space available to the solute centre of mass relative to water. Second-order exclusion through charge interactions could increase the degree of exclusion but are not considered here. Eqn (1) was applied to dextran reflection data by Scott et al. (2000b), who estimated the upper limit of synovial rp to be = 87 nm. A further estimate can be made from the reflection coefficient = 0.91 for HA2000 (Sabaratnam et al. 2004), though this entails estimating the equivalent solid sphere radius rs of HA2000. This is much smaller than Rg (101–177 nm, Table 1), because the HA chain is flexible and can deform to access pores smaller than Rg. Munch et al. (1979) showed that linear polyelectrolytes behave hydrodynamically as an equivalent solid sphere whose radius is less than half Rg. Their data for the molecular sieving of 3000 kDa hydrolysed polyacrylamide across a Nucleopore membrane with 100 nm radius pores indicates that rs is 0.26Rg at = 0.91. Substitution of rs= 0.26Rg and = 0.91 for HA2000 into eqn (1) provides a new estimate of synovial matrix rp, namely 33–59 nm. From this we can assess whether steric exclusion can explain the experimental results, as follows.
, http://www.100md.com
The dashed line in Fig. 6 shows theoretical synovial reflection coefficients for rp= 33 nm (eqn (1)). For HA140 (Rg= 21–34 nm; rs= 0.26Rg), the predicted is 0.08–0.09, which is close to the average observed reflected fraction, 0.088 ± 0.033 (Table 1). For HA500 (Rg= 39–72 nm; rs= 0.26Rg), the predicted is 0.27–0.29, which is close to the average observed reflected fraction, 0.26 ± 0.06. Steric exclusion can thus account reasonably well for the size dependence of HA reflection. It should be noted that the value adopted for solid sphere scaling (here rs/Rg= 0.26) does not materially affect the above conclusion, because a bigger scaling factor increases not only the value of rs for a given Rg, but also the estimate of rp. Concentration polarization theory predicts that the HA2000 reflected fraction in the present study should fall well below the theoretical at raised filtration rates, which it does (Fig. 6) (Sabaratnam et al. 2004; Lu et al. 2004). This aspect is discussed further below.
, 百拇医药
An alternative approach to steric exclusion is to relate it to a more realistic structural model (Ogston, 1970). If the molecular chains (glycosaminoglycans, proteoglycan core proteins, glycoproteins) of the extracellular, extrafibrillar matrix are randomly distributed, the fraction of the water space Kav available to a solid spherical solute of radius rs is given by:
where C is polymer concentration, v is effective specific volume (0.65 ml g–1) and rf is chain radius. Rabbit synovial extrafibrillar space contains 3.9 mg ml–1 glycosaminoglycan (Price et al. 1996) and an estimated 7.6 mg ml–1 of glycoprotein and proteoglycan core protein, giving a total C= 11.5 mg ml–1 (Scott et al. 2003). Fibre radius rf is approximated here as 0.87 nm, which is a weighted average for glycosaminoglycan rf (0.6 nm) and protein rf (1.0 nm). Theoretical for these values, based on eqns (1) and (2), are shown as a dotted line in Fig. 6. The fit appears less good than with the simple equivalent pore model, but both models show that steric exclusion has the potential to explain the experimental findings.
, http://www.100md.com
Relation of findings to concentration polarization and attenuation of synovial fluid loss
The volume of synovial fluid in a normal joint is very small (rabbit knee 50 μl, human knee 500–1000 μl), and a major role of HA is to prevent this vital fluid from being squeezed out of the joint cavity during a sustained flexion (outflow buffering, see Introduction). The effectiveness of outflow buffering is a graded function of HA molecular size (Coleman et al. 2000), and buffering is lost as Mw is reduced, i.e. fluid drains away more rapidly. According to the current concentration polarization hypothesis, the outflow buffering depends on HA reflection (Coleman et al. 1999; Lu et al. 2004). Reflection raises the HA concentration at the membrane surface and the resulting osmotic pressure attenuates the water outflow. Loss of outflow buffering with reduced HA Mw was attributed to a fall in the HA reflected fraction. The new findings in Figs 3 and 4 support this explanation, because they show that the HA reflected fraction does indeed fall as HA chain length is reduced.
, 百拇医药
Both concentration polarization theory and experimental observation indicate a negative relation between HA reflected fraction and trans-synovial filtration rate (Sabaratnam et al. 2004: Lu et al. 2004). This accounts for the finding that the HA2000 reflected fraction (0.63 ± 0.04 at filtration rate 35 ml min–1) is smaller than the measured or theoretical HA2000 reflection coefficient (0.91, in Fig. 6). The difference between measured reflected fraction and theoretical is less marked for the smaller HA preparations, because a fall in Mw increases the HA diffusivity and synovial permeability, reducing the degree of concentration polarization and thus allowing the reflected fraction to approach closer to the reflection coefficient at a given filtration rate.
, 百拇医药
Patho-physiological significance for arthritis
HA chain length is reduced in arthritic effusions and can be as little as 28% of normal size (Balazs et al. 1967; Bjelle et al. 1982; Dahl et al. 1985; Yingsung et al. 2003). The reduced chain length is probably due mainly to degradation by free oxygen radicals in the inflamed joint (McCord, 1974; Greenwall & Moak, 1986; Baker et al. 1989; Halliwell, 1995; Haubeck et al. 1995; Schenck et al. 1995). It is also possible that pro-inflammatory cytokines alter the expression of hyaluronan synthase (HAS) isoforms (Castor & Dorstewitz, 1966; Vuorio et al. 1982). Of the three mammalian isoforms, HAS2 is most commonly expressed and synthesizes large-Mw HA, whereas HAS3 synthesizes chains of only 200–300 kDa. HAS3 is up-regulated by cytokines found in inflamed joints, such as IL-1and TNF (Spicer & Nguyen, 1999; Ohno et al. 2001). Dahl & Husby (1985) reported HA chains as small as 50 kDa in the supernatant of cultured rheumatoid cells. The present findings show that such chains would quickly leak out of the joint cavity, leaving behind the longer chains. In rheumatoid synovial fluid the hyaluronan loss may be offset to some degree by the presence of SHAP–hyaluronan complexes, but their quantitative significance is not yet clear (Yingsung et al. 2003).
, 百拇医药
The reduced HA chainlength in arthritis has a number of functional consequences. First, the fall in intrinsic viscosity (Table 1) reduces the hydrodynamic lubricating ability of the synovial fluid, as long recognized (Balazs et al. 1967). Second, in arthritis joint movements are limited and effusion pressures are always positive, as in the present study. Our results indicate that a fall in HA reflected fraction will reduce the build-up of a concentration polarization layer at the synovial surface during fluid drainage. There will therefore be less osmotic buffering of outflow from the arthritic joint. This can be viewed as a useful biological response in that it facilitates the resolution of the intra-articular effusion that inevitably accompanies arthritis. The reduced reflection also has a downside, however, because the increased loss of HA will contribute to the known fall in its intra-articular concentration, which will reduce further the lubricating power of the intra-articular fluid. Increased loss of HA also calls for a faster synthesis rate to replace it. The latter is a known effect of some pro-inflammatory cytokines, but it increases the metabolic burden on the synovial lining cells, which are in a relatively hostile, acidotic and hypoxic environment in arthritis.
, 百拇医药
To summarize, a study of HA transport in vivo showed that the retention of HA in joints depends critically on its chain length. Reduction of hyaluronan chain length, as in arthritis, increases the rate of loss of not only HA but also synovial fluid itself, because there is reduced concentration polarization and hence reduced outflow buffering.
References
Anderson JL & Malone DM (1974). Mechanism of osmotic flow in porous membranes. Biophys J 14, 957–982.
, 百拇医药
Asari A, Miyauchi S, Matsuzaka S, Ito T, Kominami E & Uchiyama Y (1998). Molecular weight-dependent effects of hyaluronate on the arthritic synovium. Arch Histol Cytol 61, 125–135.
Baker MS, Green SP & Lowther DA (1989). Changes in the viscosity of hyaluronic acid after exposure to a myeloperoxidase-derived oxidant. Arthritis Rheum 32, 461–469.
Balazs EA, Watson D, Duff IF & Roseman S (1967). Hyaluronic acid in synovial fluid. I. Molecular parameters of hyaluronic acid in normal and arthritic human fluids. Arthritis Rheum 10, 357–376.
, 百拇医药
Bjelle A, Anderson T & Granath K (1982). Molecular weight distribution of hyaluronic acid of human synovial fluid in rheumatic diseases. Scand J Rheumatol 12, 133–138.
Brown TJ, Alcorn D & Fraser JRE (1999). Absorption of hyaluronan applied to the surface of intact skin. J Invest Dermatol 113, 740–746.
Brown TJ, Laurent UBG & Fraser JRE (1991). Turnover of hyaluronan in synovial joints: elimination of labelled hyaluronan from the knee joint of the rabbit. Exp Physiol 76, 125–134.
, 百拇医药
Castor CW & Dorstewitz EL (1966). Abnormalities of connective tissue cells cultured from patients with rheumatoid arthritis. I. Relative unresponsiveness of rheumatoid synovial cells to hydrocortisone. J Lab Clin Med 68, 306–313.
Coleman PJ, Scott D, Mason RM & Levick JR (1999). Characterization of the effect of high molecular weight hyaluronan on trans-synovial flow in rabbit knees. J Physiol 514, 265–282.
Coleman PJ, Scott D, Mason RM & Levick JR (2000). Role of hyaluronan chain length in buffering interstitial flow across synovium in rabbits. J Physiol 526, 425–434.
, 百拇医药
Coleman PJ, Scott D, Ray J, Mason RM & Levick JR (1997). Hyaluronan secretion into the synovial cavity of rabbit knees and comparison with albumin turnover. J Physiol 503, 645–656.
Curry FE (1984). Mechanics and thermodynamics of transcapillary exchange. In Handbook of Physiology, section 2, The Cardiovascular System, vol IV The Microcirculation, ed Renkin EM & Michel CC, pp. 309–374, American Physiological Society, Bethesda.
, 百拇医药
Dahl LB, Dahl IM, Engstrm-Laurent A & Granath K (1985). Concentration and molecular weight of sodium hyaluronate in synovial fluid from patients with rheumatoid arthritis and other arthropathies. Ann Rheumatic Dis 44, 817–822.
Dahl IMS & Husby G (1985). Hyaluronic acid production in vitro by synovial lining cells from normal and rheumatoid joints. Ann Rheumatic Dis 44, 647–657.
Davies DV (1946). The lymphatics of the synovial membrane. J Anat 80, 21–23.
, http://www.100md.com
Day AJ & Sheehan JK (2001). Hyaluronan: polysaccharide chaos to protein organisation. Curr Opin Struct Biol 11, 617–622.
de Gennes P-G (1979). Scaling Concepts in Polymer Physics. Cornell University Press, Ithaca and London.
Denlinger JL (1982). Metabolism of Sodium Hyaluronate in Articular and Ocular Tissues. Doctoral Dissertation, Université de Sciences et Techniques de Lille, Lille.
Flory PJ (1971). Principles of Polymer Chemistry. Cornell University Press, Ithaca, New York.
, 百拇医药
Fraser JRE & Laurent TC (1996). Hyaluronan. In Extracellular Matrix, Vol 2. Molecular Components and Interactions, ed. Comper WD, pp. 141–199, Harwood Academic Publishers, Amsterdam.
Greenwall RA & Moak SA (1986). Degradation of hyaluronic acid by polymorphonuclear leucocytes. Inflammation 10, 15–30.
Halliwell B (1995). Oxygen radicals, nitric oxide and human inflammatory joint disease. Ann Rheumatic Dis 54, 505–510.
, 百拇医药
Hardingham T (2004). Solution properties of hyaluronan. In Chemistry and Biology of Hyalurona, ed. Garg HG & Hales CA, pp. 1–19, Elsevier, Amsterdam.
Haubeck HD, Kock R, Fischer D-C, Leur EV, Hoffmeister K & Greiling H (1995). Transforming growth factor B1, a major stimulator of hyaluronan synthesis in human synovial lining cells. Arthritis Rheum 38, 669–677.
Johnson M, Kamm R, Ethier CR & Pedley T (1987). Scaling laws and the effects of concentration polarization on the permeability of hyaluronic acid. Physicochem Hydrodynamics 9, 427–444.
, 百拇医药
Laurent TC, Ryan M & Pietrustkiewicz A (1960). Fractionation of hyaluronic acid. The polydispersity of hyaluronic acid from the bovine vitreous body. Biochemica et Biophysica acta 42, 476–485.
Levick JR (1987). Flow through interstitium and other fibrous matrices. Q J Exp Physiol 72, 409–438.
Levick JR, Price FM & Mason RM (1996). Synovial matrix-synovial fluid system of joints. In Extracellular Matrix, Vol. 1. Tissue Function ed. Comper WD, pp. 328–377. Harwood Academic Publishers, Amsterdam.
, 百拇医药
Lu Y, Levick JR & Wang W (2004). Concentration polarization of hyaluronan on the surface of the synovial lining of infused joints. J Physiol 561, 559–573.
Lu Y, Levick JR & Wang W (2005). Synovial fluid retention in pressurised joint cavities is achieved by hyaluronan concentration polarisation. Microcirculation (in press).
McCord JM (1974). Free radicals and inflammation: protection of synovial fluid by superoxide dismutase. Science 185, 529–531.
, 百拇医药
McDonald JN & Levick JR (1995). Effect of intra-articular hyaluronan on pressure–flow relation across synovium in anaesthetised rabbits. J Physiol 485, 179–193.
Munch WD, Zestar LP & Anderson JL (1979). Rejection of polyelectrolytes from microporous membranes. J Memb Sci 5, 77–102.
Nagai S (1987). The basic study of the synovial membrane of the normal rabbit knee by using the induction system of carbon particles. J Tokyo Med Coll 45, 488–498.
, 百拇医药
Ogston AG (1970). On the interaction of solute molecules with porous networks. J Phys Chem 74, 668–669.
Ohno S, Tanimoto K, Fujimoto K, Ijuin C, Honda K, Tanaka N, Doi T, Nakahara M & Tanne K (2001). Molecular cloning of rabbit hyaluronan synthases and their expression patterns in synovial membrane and articular cartilage. Biochim Biophys Acta 1520, 71–78.
Page-Thomas DP, Bard D, King B & Dingle JT (1987). Clearance of proteoglycan from joint cavities. Ann Rheum Dis 46, 934–937.
, 百拇医药
Price FM, Levick JR & Mason RM (1996). Glycosaminoglycan concentration in synovium & other tissues of rabbit knee in relation to hydraulic resistance. J Physiol 495, 803–820.
Reimann I, Vittas D, Nielsen SL & Svalastoga E (1989). Lymphatic transport from normal and synovitic knees in rabbits. Acta Orthop Scand 60, 185–187.
Sabaratnam S, Coleman PJ, Badrick E, Mason RM & Levick JR (2002a). Interactive effect of chondroitin sulphate C and hyaluronan on fluid movement across rabbit synovium. J Physiol 540, 271–284.
, 百拇医药
Sabaratnam S, Mason RM & Levick JR (2002b). Inside-out cannulation of fine lymphatic trunks used to quantify coupling between trans-synovial flow and lymphatic drainage from rabbit knees. Microvasc Res 64, 1–13.
Sabaratnam S, Mason RM & Levick JR (2003). Molecular sieving of hyaluronan by synovial interstitial matrix and lymphatic capillary endothelium evaluated by lymph analysis in rabbits. Microvasc Res 66, 227–236. Erratum to all figures. Microvasc Res (2004) 67, 245–246.
, http://www.100md.com
Sabaratnam S, Mason RM & Levick JR (2004). Filtration rate-dependence of hyaluronan reflection by joint-to-lymph barrier in rabbits; evidence for concentration polarisation. J Physiol 557, 909–922.
Schenck P, Schneider S, Miehlke R & Prehm P (1995). Synthesis and degradation of hyaluronate by synovia from patients with rheumatoid arthritis. J Rheumatol 22, 400–405.
Scott D, Coleman PJ, Abiona A, Ashhurst DE, Mason RM & Levick JR (1998b). Effect of depletion of glycosaminoglycans and non-collagenous proteins on interstitial hydraulic permeability in rabbit synovium. J Physiol 511, 629–643.
, 百拇医药
Scott D, Coleman PJ, Mason RM & Levick JR (1998a). Direct evidence for the partial reflection of hyaluronan molecules by the lining of joints during trans-synovial flow. J Physiol 508, 610–623.
Scott D, Coleman PJ, Mason RM & Levick JR (2000a). Concentration-dependence of interstitial flow buffering by hyaluronan in synovial joints. Microvasc Res 59, 345–353.
Scott D, Coleman PJ, Mason RM & Levick JR (2000b). Action of polysaccharides of similar average mass but differing molecular Volume and charge on fluid drainage through synovial interstitium in rabbit knees. J Physiol 528, 609–618.
, 百拇医药
Scott D, Levick JR & Miserocchi G (2003). Non-linear dependence of interstitial fluid pressure on joint cavity pressure and implications for interstitial resistance in rabbit knee. Acta Physiol Scand 179, 93–101.
Spicer AP & Nguyen TK (1999). Mammalian hyaluronan synthase: investigation of functional relationships in vivo. Biochem Soc Trans 27, 109–115.
Vuorio E, Einola S, Hakkarainen S & Penttinen S (1982). Synthesis of underpolymerized hyaluronic acid by fibroblasts cultured from rheumatoid and non-rheumatoid synovitis. Rheumatol Int 2, 97–102.
, 百拇医药
Wik HB & Wik O (1998). Rheology of hyaluronan. In The Chemistry, Biology and Medical Applications of Hyaluronan and its Derivatives, ed. Laurent TC, pp. 25–32. Portland Press, London.
Xu H, Edwards JC, Banerji S, Prevo R, Jackson DG & Athanasou NA (2003). Distribution of lymphatic vessels in normal and arthritic human synovial tissue. Ann Rheum Dis 62, 1277–1279.
Yamashita S & Ohkubo M (1993). Distribution and three-dimensional reconstruction of lymphatic vessels of the elbow joint capsule of rabbits. Acta Anat Nipponica 68, 513–521.
Yingsung W, Zhuo L, Morgelin M, Yoneda M, Kida D, Watanabe H, Ishiguro N, Iwata H & Kimata K (2003). Molecular heterogeneity of the SHAP-hyaluronan complex. J Biol Chem 278, 32710–32718., http://www.100md.com(S. Sabaratnam, V. Arunan,)
2 Division of Medicine, Imperial College, Hammersmith Hospital, London W12 0NN, UK
Abstract
In joint fluid the polymer hyaluronan (HA) confers viscous lubrication and greatly attenuates trans-synovial fluid loss (outflow buffering). Outflow buffering arises from the molecular sieving (reflection) and concentration polarization of HA at the synovial membrane surface. Outflow buffering declines if HA chain length is reduced, as in arthritis, and this has been attributed to reduced HA reflection. This was tested directly in the present study. Infused solutions of HA of 2200 kDa (HA2000, 0.2 mg ml–1) or 500 kDa (HA500, 0.2 mg ml–1) or 140 kDa (HA140, 0.2–4.0 mg ml–1) were filtered across the synovial lining of the knee joint cavity of anaesthetized rabbits at a constant rate, along with a freely permeating reference solute, 20 kDa fluorescein–dextran (FD20). After a priming period the femoral lymph was sampled over 3 h. Mixed intra-articular (I.A.) fluid and subsynovial fluid were sampled at the end. Fluids were analysed by gel exclusion chromatography. The trans-synovial concentration profile was found to depend on polymer size. The I.A. concentration of HA2000 increased substantially relative to infusate and the subsynovial and lymph concentrations fell substantially. For HA500 and HA140 the trans-synovial concentration gradients were less pronounced, and absent for FD. The reflected fractions for HA2000, HA500 and HA140 across the cavity-to-lymph barrier were 0.65 ± 0.05 (n= 10), 0.43 ± 0.09 (n= 3) and 0.19 ± 0.05 (n= 7), respectively, at matched filtration rates (P < 0.0001, analysis of variance). Reflected fractions calculated from HA I.A. accumulation or subsynovial dilution showed the same trend. The results demonstrate size-selective molecular sieving by the synovial extracellular matrix, equivalent to steric exclusion from cylindrical pores of radius 33–59 nm. The findings underpin the concentration polarization-outflow buffering theory and indicate that reduced HA chain length in arthritis exacerbates lubricant loss from a joint.
, 百拇医药
Introduction
In a synovial joint the cells that line the joint cavity form a discontinuous layer. Consequently, the intervening interstitial matrix is in direct contact with the synovial fluid of the joint cavity. This interstitial pathway constitutes the pathway for fluid drainage out of the joint cavity. A subpopulation of the lining cells, called fibroblast-related synoviocytes or type B cells, secrete hyaluronan and lubricin into the joint fluid, creating the viscous, lubricating synovial fluid.
, 百拇医药
Hyaluronan (HA) is a long, unbranched chain of repeating N-acetyl-D-glucosamine-D-glucuronic acid disaccharides (Fraser & Laurent, 1996). Its concentration in synovial fluid is higher than in any other adult tissue, being 2–4 mg ml–1 in healthy rabbit and human knees. Its weight-average molecular mass (Mw) is 2000 kDa in rabbit synovial fluid (Levick et al. 1996; Price et al. 1996). In arthritis both the concentration and chain length are reduced. In aqueous solution, hydrogen bonds stiffen the HA chain, which causes it to adopt an expanded coil configuration with an exceptionally large molecular domain (radius of gyration 100–200 nm). As a result, adjacent molecular domains overlap at 1 (mg HA) ml–1, creating marked chain–chain interaction and a quasi-infinite, dynamic network of loosely linked polymer chains (Day & Sheehan, 2001; Hardingham, 2004). This enables HA to act as a viscous, hydrodynamic lubricant of surfaces under low load, such as synovium on cartilage and synovium on synovium.
, 百拇医药
As well as its lubricating role, HA has a profound buffering effect on fluid drainage from the joint cavity into the subsynovial lymphatic system. When intra-articular fluid pressure is raised (e.g. by joint flexion), fluid begins to drain from the cavity through the interstitial spaces of the synovial lining, but this is soon countered by the osmotic pressure of HA that is sieved out and retained at the membrane surface. As a result, rises in joint pressure above 5 cmH2O produce remarkably little increase in trans-synovial fluid loss when HA is present (McDonald & Levick, 1995). This phenomenon, called outflow buffering, is physiologically important for three reasons. It conserves the tiny volume of synovial fluid (50 μl in a rabbit knee) during periods of sustained high intra-articular fluid pressure such as a maintained flexion; it prolongs by an order of magnitude the intra-articular working life of long-chain HA as a lubricant; and, correspondingly, it reduces by an order of magnitude the rate of HA biosynthesis needed to replace lost HA.
, 百拇医药
The concentration polarization hypothesis has been developed into quantitative steady state and non-steady state models to explain outflow buffering (Coleman et al. 1999; Lu et al. 2005). It was argued that HA accumulates in the fluid adjacent to the partially reflecting membrane (synovial lining interstitium), and the osmotic pressure of this layer opposes filtration. There is experimental support for two key postulates of this hypothesis. First, molecular sieving of HA has been confirmed by a sharp drop in HA concentration between intra-articular fluid and subsynovial fluid during trans-synovial filtration (Sabaratnam et al. 2003). Additional, supporting observations are that native HA has an exceptionally long residence half life in the joint cavity, 14–32 h (Denlinger, 1982; Brown et al. 1991; Coleman et al. 1997), and that HA accumulates in the joint cavity when a solution is filtered experimentally through the synovial lining (Scott et al. 1998a). Second, concentration polarization build up at the interface was confirmed by demonstrating its predicted consequence, namely a negative relation between reflected fraction and filtration rate (Lu et al. 2004; Sabaratnam et al. 2004).
, http://www.100md.com
Although HA can slowly permeate an interstitial matrix, both in joints and in the dermis, to reach the lymph (Brown et al. 1991, 1999; Fraser & Laurent, 1996), it is clear from the above findings that the synovial matrix creates sufficient steric hindrance to HA transport relative to water to cause molecular sieving. The matrix itself is a complex network of heterogeneous biopolymers, including sulphated and non-sulphated glycosaminoglycans, proteoglycans, glycoproteins and microfibrils (for review see Levick et al. 1996). The hydraulic drag of the biopolymers generates a high resistance to fluid escape (Levick, 1987; Scott et al. 1998b), and as a result the hydraulic resistance of the synovial lining is 20-fold higher than that of the areolar connective tissue in the subsynovial space (Scott et al. 2003). The subsynovial compartment contains a network of lymphatic capillaries that do not penetrate the synovium itself (Yamashita & Ohkubo, 1993; Xu et al. 2003) but collect the escaped trans-synovial filtrate and transport it into femoral lymphatic trunk vessels (Davies, 1946; Nagai, 1987; Reimann et al. 1989).
, http://www.100md.com
Experiments showed that the buffering of joint fluid loss depends on HA chain length and concentration. Buffering is lost if the bulk-phase HA concentration falls below 1 mg ml–1 (Scott et al. 2000a) or if the molecular mass falls below 500 kDa (Coleman et al. 2000) or if the molecular domain radius is reduced for a fixed polymer molecular mass (Scott et al. 2000b). Conversely, outflow buffering can be conferred on a non-buffering HA solution (0.75 mg ml–1) by increasing the effective HA domain size through chemical enhancement of the chain–chain interactions (Sabaratnam et al. 2002a).
, 百拇医药
The present study investigates the mechanism by which HA chain length affects outflow buffering. The concentration polarization theory predicts that outflow buffering will decline if HA reflection is a function of HA chain length. Preliminary evidence for the latter was obtained during the study of outflow buffering versus chain length. It was found that less HA accumulated in the joint cavity during infusions of 530 kDa, 300 kDa and 90 kDa HA than 2000 kDa HA (Coleman et al. 2000). The protocol, however, was designed primarily to study outflow buffering rather than sieving, and was not ideal for the latter purpose. In particular, the filtration rates were higher for low Mw HA than high Mw HA, and were necessarily changed many times during each experiment. This is a potentially confounding factor in a sieving study because, as emerged later, the reflected fraction for HA of constant chain length falls as filtration rate is increased (Sabaratnam et al. 2004; Lu et al. 2004). Moreover no measurements have been made of the sieving ratio, i.e. downstream (filtrate) HA concentration relative to upstream (filtrand) concentration.
, http://www.100md.com
The present aim was therefore to measure the effect of HA chain length on molecular sieving between the joint cavity and the subsynovial fluid or the lymphatic system at constant, matched filtration rates. The results have important implications for fluid and HA retention in arthritic joints, where HA chain length is reduced.
Methods
Overview
The joint cavity of a rabbit knee was cannulated and infused with HA solution at a controlled intra-articular pressure to generate a sustained, constant filtration across the synovial lining over several hours. A single filtration rate was studied per preparation. The Mw of the infused HA was 140 kDa or 500 kDa or 2230 kDa (referred to as HA140, HA500, HA2000, respectively). A reference solute, fluorescein–dextran (FD20), was included to determine the amount of joint lymph present in femoral lymph. After a priming period, femoral lymph was collected over 3 h and analysed for HA and FD20. The cavity-to-lymph sieving coefficient (transmitted fraction) was calculated from the infusate and lymph [HA]/[FD20] ratios. At the end, fluid was also aspirated from the subsynovial compartment and from the joint cavity, to assess the cavity-to-subsynovium sieving coefficient and the degree of intra-articular HA accumulation. The method has been described and evaluated previously (Sabaratnam et al. 2002b, 2003) and is summarized below.
, 百拇医药
Materials
Rooster comb HA (HA2000, 0.2 mg ml–1, Mw2230 kDa, radius of gyration 101–181 nm, Coleman et al. 1999) and fluorescein–dextran (30 μg ml–1, 20 kDa, Stokes–Einstein radius 3.1 nm) were purchased from Sigma Chemical Co. (Poole, UK). Shorter HA chains were prepared as described later. The HA and FD20 were co-administered in Baxter Ringer solution (mM: 147 Na+, 4 K+, 2 Ca2+, 156 Cl–; Baxter Healthcare Ltd, Thetford, Norfolk, UK).
, http://www.100md.com
A hyaluronan concentration of 0.2 mg ml–1, the lowest concentration reported for rheumatoid joints (Dahl et al. 1985), was used for all chain lengths because it is well below the critical concentration for HA2000 molecular domain overlap (0.8–1.3 mg ml–1, Coleman et al. 1999; Scott et al. 2000a and Results) and does not buffer outflow. The latter condition was necessary to achieve matched trans-synovial filtration rates for the various sizes of HA (34–37 μl min–1) and to enable filtration rates that yielded good harvests of lymph. It is impossible to achieve matching filtration rates at >1 mg ml–, because this concentration buffers the filtration rate to a very low level in the case of HA2000 (Scott et al. 2000a). In the case of HA140 some studies were also carried out at a higher concentration (4.0 mg ml–1, filtration rate again 37 μl min–1) for technical reasons, namely to facilitate the detection and separation of the 140 kDa HA peak from an adjacent peptide/papain peak during size-exclusion chromatography.
, http://www.100md.com
Animal preparation, joint cannulation and trans-synovial filtration
New Zealand white rabbits weighing 2–3 kg were anaesthetized with 30 mg kg–1 sodium pentobarbitone plus 500 mg kg–1 urethane I.V. and tracheostomized. Anaesthesia of sufficient depth to abolish the corneal blink reflex was maintained by 15 mg sodium pentobarbitone plus 250 mg urethane I.V. every 30 min. Only one knee joint was studied per animal. Following the procedure of Coleman et al. (1999), an intra-articular cannula was connected to a pressure transducer to record intra-articular fluid pressure Pj (±0.1 cmH2O). A second intra-articular cannula was connected to an infusion reservoir, the height of which regulated Pj, which in turn regulated the trans-synovial filtration rate. Flow from the reservoir into the joint cavity when intra-articular volume is constant depends on the rate at which fluid is draining away through the synovial lining. The rate of fluid uptake was measured using a photoelectric drop counter (5.6 μl) and chart recorder. A small correction was applied for viscoelastic creep of the cavity walls as previously described. Procedures conformed to UK animal legislation, and animals were killed by I.V. sodium pentobarbitone overdose at the end of the study.
, 百拇医药
Collection of mixed femoral lymph
Knee joint lymph drains into 2–3 major lymphatic trunk vessels in the femoral triangle. There are no intervening lymph nodes. The joint was cannulated and trans-synovial filtration initiated prior to lymphatic dissection to fill the lymphatics, then the femoral lymphatics were dissected clear of the adjacent femoral blood vessels and ligated proximally. After a further, undisturbed 1 h priming interval at a constant trans-synovial filtration rate, the largest lymphatic was transected and cannulated by the inside-out method of Sabaratnam et al. (2002b). Lymph was aspirated into a fluid trap. The trap was emptied and the lymph weighed (±1 mg) every 15 min for 2–3 h. The lymph output is well maintained over the 3 h period (Sabaratnam et al. 2002b).
, http://www.100md.com
Volume fraction of joint lymph in femoral lymph (VV)
We have shown previously that the amount of joint lymph in femoral lymph, expressed as the volume fraction Vv, is given by Vv=Ljoint/Lfemoral=Cfemoralref/Cjointref, where Ljoint and Lfemoral are joint and femoral lymph flows, respectively, and Creffemoral and Crefjoint are reference solute (FD20) concentration in femoral lymph and joint, respectively (Sabaratnam et al. 2002b). The subsynovial, intra-articular and infused FD20 concentrations do not differ significantly, showing that FD20 permeates the joint lining freely (Sabaratnam et al. 2003).
, 百拇医药
Hyaluronan transmitted fraction from joint to lymph (1 –Rlymph)
If a fraction Rlymph of the HA molecules in the bulk filtrand (infusate) is reflected by the synovium-to-lymph pathway, the transmitted fraction is (1 –Rlymph). The transmitted fraction is often referred to as a ‘sieving coefficient’ or ‘sieving ratio’ in microvascular and renal physiology. We have shown previously that the transmitted fraction can be calculated from the measured concentrations as CHAfemoral/(CHAinfusatexVV), where CHAfemoral and CHAinfusate are, respectively, the femoral lymph and infused HA concentrations (Sabaratnam et al. 2003).
, 百拇医药
Hyaluronan transmitted fraction from joint to subsynovium (1 –Rsyn)
Transport from the joint cavity to lymph involves two membranes in series, namely the synovium and the lymphatic capillary endothelium. A sample of the intervening subsynovial fluid was aspirated post mortem and analysed to verify that the molecular sieving occurs across the synovial lining, as shown in previous studies (Sabaratnam et al. 2003). The animal was killed by I.V. pentobarbitone, 1 ml Evans blue solution was injected into the knee cavity to visualize its boundaries, and the peri-articular tissue was dissected away to within a millimetre or so of the cavity border. A sample of the accumulated trans-synovial filtrate was aspirated through a catheter. The reflected fraction across the synovial lining, Rsyn, was calculated from the HA/FD20 concentration ratios in the subsynovial fluid and infusate as before (Sabaratnam et al. 2003).
, 百拇医药
HA reflection assessed by accumulation in joint cavity aspirate (Rasp)
Molecular sieving should cause upstream HA accumulation as well as downstream HA dilution. To assess this, the intra-articular fluid was mixed by 10 flexion–extension cycles at the end of the experiment and aspirated for analysis. Upstream reflection (Rasp) was calculated as described by Scott et al. (1998a), namely mass of HA reflected and retained in the cavity divided by the mass of HA presented to the membrane in the cumulative filtrand volume. Thus Rasp equals the increase in the intra-articular HA concentration x intra-articular fluid volume (i.e. the HA mass reflected and retained in the joint cavity) divided by the cumulative volume of fluid filtered during the experiment x infusate concentration (i.e. the HA mass in the total filtrand volume). The assumption that the 10 mixing cycles adequately dissipate the intra-articular concentration polarization layer was supported by the finding that Rasp and Rlymph did not differ significantly (Sabaratnam et al. 2003, 2004). The lymph method has the advantage of circumventing assumptions inherent in the Rasp method.
, http://www.100md.com
Analysis by high-performance liquid chromatography
Samples were centrifuged and diluted to 200 μl in Ringer solution. To prevent partial masking of small HA bands by endogenous albumin, samples were digested with 5.6 units papain (Sigma, UK) at 60°C for 1 h. This does not alter the HA molecular size (Coleman et al. 1997). The HA was quantified by size exclusion, high-performance liquid chromatography (HPLC) using a Waters 2690 separation module (Waters Ltd, Watford, UK), a TosoHaas TSK G6000 PWXL column (Anachem Ltd, Luton, UK) of nominal resolution 40–8000 kDa, and a Waters 486 ultraviolet absorbance detector set at 206 nm for HA analysis (Coleman et al. 1997). The injection volume was 50–100 μl and column flow 1 ml min–1 Ringer solution. Calibration curves for concentration were linear from 3 μg ml–1 to 400 μg ml–1 for each Mw. Sample Mw was estimated from the mean retention time, which was calibrated using HA standards of Mw 210 kDa to 5500 kDa generously donated by Dr O. Wik (New Pharmacia, Uppsala, Sweden) and characterized by laser light scattering. The HA calibration curves have been published previously (Coleman et al. 1997). Fluorescein–dextran was analysed using the same HPLC column and an in-line Waters 474 SATIN fluorimeter set to an excitation wavelength of 475 nm and emission wavelength 530 nm. Minimum detection level was <0.3 μg ml–1.
, 百拇医药
Preparation of HA samples of reduced chain length
The commercial rooster comb HA had a Mw of 2230 ± 50 kDa as indicated by HPLC retention time (Fig. 1A). Its polydispersity has previously been characterized using the ratio Mw/Mn= 2.3, where Mn is number-average molecular mass (970 kDa by osmometry, Coleman et al. 1999). Samples of Mw140 kDa were produced by sonicating 35 ml aliquots of 4 mg ml–1 rooster HA2000 for 16 min at 10 μm amplitude in a Soniprep (MSE Scientific Instruments, Crawley, UK), following the protocol of Coleman et al. (2000). Evaporation during sonication was measured by weight change and replaced by water addition. The HPLC retention time increased to 8.86 ± 0.03 min (n= 12), corresponding to a Mw of 140 ± 8 kDa (Fig. 1A). As shown in previously published chromatographs, the sonication process has little effect on the spread of the HPLC peak, indicating that polydispersity is not altered significantly (Coleman et al. 2000). Since the unit disaccharide has a mass of 379 Da and length 0.95 nm, the average HA140 chain comprised 369 disaccharides and was 0.35 μm long, in contrast to 5884 disaccharides and 5.6 μm long for HA2000. A molecular mass of 140 kDa was chosen because we showed previously that sonicates of 88–305 kDa do not buffer the trans-synovial filtration rate (Coleman et al. 2000). Likewise, 500 kDa at 0.2 mg ml–1 does not buffer outflow, thought it does at 3 mg ml–1.
, 百拇医药
A, HPLC chromatographs for HA2000, HA500 and HA140 at 0.2 mg ml–1. The cited retention times are the means for the series (±S.E.M.). B, plots of reduced viscosity (see Methods) as a function of concentration to determine the intrinsic viscosity [] of each hyaluronan preparation. Lines fitted by linear regression analysis.
Samples of intermediate Mw can be produced by shorter sonication periods, but this proved unnecessary. A commercial sample of hyaluronan that had been stored for some time proved to have a Mw of 500 ± 60 kDa, as determined by the HPLC retention time of 8.30 ± 0.05 min (n= 12) (Fig. 1A). This was ideal for the study because its Mw was approximately four times that of the smaller preparation and one-quarter that of the larger preparation.
, http://www.100md.com
Viscometric characterization of size of hyaluronan molecular volume domains
Samples were analysed by viscometry to determine the effective molecular domain volume and the concentration at which adjacent molecular domains overlap (C*). Intrinsic viscosity, [], the volume occupied by 1 g solute at infinite dilution (ml g–1), is a sensitive index of molecular domain size. Intrinsic viscosity was determined by linear extrapolation of a plot of the logarithm of reduced viscosity redversus concentration C; red is (– 1)/C, where is relative viscosity. Viscosity relative to the solvent (Ringer solution) was measured using an Ostwald viscometer for solutions of low viscosity and negligible shear dependence, or a rotational rheometer for solutions of high viscosity with a marked shear-rate dependence (HA2000 at >1 mg ml–1). The Haaske RS150 cone-on-plate rheometer (Carl Stuart, Leek, UK) measured apparent viscosity (shear stress/shear rate) at 25°C via a set of automated step shear rates from 0.12 s–1 to 500 s–1.
, 百拇医药
The quasi-spherical domain of solvent occupied by a single polymer chain is conventionally characterized by the radius of gyration, Rg (root mean square of polymer segment distance from the molecular centre of gravity), which is related to []. For a neutral, flexible polymer the self-avoiding random walk model of Flory (1971) gives Rg3=MW[]/8.84NA, where NA is Avogadro's number. An alternative, empirical expression specific to HA, which is not neutral but carries one negative charge per disaccharide, is Rg3= 0.025MW0.6S–0.08, where S is the molar salinity of the solvent (Johnson et al. 1987).
, 百拇医药
de Gennes, (1979) defined the ‘critical concentration’C* for a polymer as the concentration at which the molecular domains overlap to produce a statistically homogenous, entangled network (the semidilute regime). This leads to the expression C*[]= 2.1 (McDonald & Levick, 1995), which was used to estimate C* for each preparation. A direct experimental assessment of domain interaction was also obtained by plotting the logarithm of specific viscosity (– 1) versus the logarithm of HA concentration; a sharp increase in the slope of this plot marks the onset of significant chain–chain interaction (Wik & Wik, 1998; Scott et al. 2000a).
, 百拇医药
Statistical analysis
Means are followed by S.E.M. The t test was used for paired, unpaired and one-sample comparisons as appropriate. One-way ANOVA with Tukey's post hoc test was used for comparison of three or more sets of results, and two-way ANOVA was used to compare results subject to two variables (e.g. time and polymer size). Lines were fitted by linear regression analysis. All tests were as implemented in Graphpad Prism (San Diego, CA). Significance was accepted at P 0.05.
, http://www.100md.com
Results
Viscometric assessment of HA domain sizes
For HA140 the relative viscosity was 1.11 at 0.2 mg ml–1 and 1.94 at 4 mg ml–1. The intrinsic viscosity [] extrapolated from a ln(red) versus C plot over the range 0.0063–4.000 mg ml–1 was 334 ± 42 ml g–1 (Fig. 1B). Corresponding values for HA500 were = 1.17 at 0.2 mg ml–1 and []= 646 ± 49 ml g–1. The HA2000 values were = 1.53 at 0.2 mg ml–1 (shear rate 103 s–1) and []= 2455 ± 318 ml g–1 (Fig. 1B). Table 1 summarizes these and other key biophysical parameters for the three HA preparations.
, 百拇医药
The Flory radius of gyration was 21 nm for HA140, 39 nm for HA500 and 101 nm for native rooster comb (HA2000). Corresponding values from the Johnson et al. (1987) salinity formula (see Methods) were 34 nm, 72 nm and 177 nm radius.
The de Gennes (1979) overlap concentration C* was 6.29 mg ml–1 for HA140, 3.35 mg ml–1 for HA500 and 0.86 mg ml–1 for HA2000. All the infusates in the present study were below C*, and thus in the dilute regime where molecular domains do not overlap. Plots of log (specific viscosity) versus log (concentration) confirmed this. For HA2000, the slope increase denoting onset of molecular interaction occurred at 1.30 mg ml–1, a result close to the value of 1.35 mg ml–1 illustrated in Scott et al. (2000a). HA140 showed no slope change up to the highest concentration explored, 4 mg ml–1. With HA500, any slope change was at 2 mg ml–1.
, 百拇医药
Lymph composition versus time for HA and FD
Figure 2 illustrates the near constancy of femoral lymph composition over the 3 h collection period. The changes in HA140 concentration and FD20 concentration with time were small, and the regression slopes were not statistically significant (P 0.38). We have previously demonstrated similar findings for HA2000 (Sabaratnam et al. 2003). The mean HA140 concentration in the lymph, 1.83 ± 0.13 mg ml–1, was 45.8%± 3.4% of the infused concentration (n= 7), whereas the FD20 concentration, 19.7 ± 1.2 μg ml–1, was proportionately higher at 65.6%± 4.1% of the infused concentration (n= 7). Two-way ANOVA of these results confirmed that the percentage of FD20 in the lymph was significantly higher than the percentage of HA140 (P < 0.001), and that time did not significantly influence the values over 3 h (P= 0.83). These results show that HA undergoes some degree of selective molecular sieving, even at the reduced chain length of 140 kDa.
, 百拇医药
A, lymph HA140 concentration versus time; slope of relation not significantly different from zero (0.0036 mg ml –1 min–1, P= 0.37). Broken line shows infused concentration. B, corresponding FD20 concentrations; slope of relation not significantly different from zero (0.033 μg ml –1 min, P= 0.38). C, ratio of mean HA140 concentration to mean FD concentration versus time: concentrations were normalized as percentage of the infused concentration. Linear regression line ±95% confidence intervals; slope 0.0003 min–1 not significant (P= 0.36). The mean ratio, 0.69 ± 0.01 (n= 8), was significantly less than 1 (P < 0.0001, one-sample t test), demonstrated selective molecular sieving of HA140 across the cavity-to-lymph barrier.
, 百拇医药
The percentage FD20 value represents Vv, the joint lymph volume fraction in femoral lymph under the experimental conditions. Vv is a positive function of trans-synovial filtration rate (Sabaratnam et al. 2002b). In the present study the filtration rates did not differ significantly for the three molecular sizes investigated, being 37.3 ± 1.9 μl min–1 (range 28–43 μl min–1) for HA140 at 4 mg ml–1; 37.7 ± 2.0 μl min–1 (range 33–40 μl min–1) for HA140 at 0.2 mg ml–1; 36.9 ± 1.8 μl min–1 (range 35–40 μl min–1, n= 3) for HA500; and 34.4 ± 2.4 μl min–1 (range 28–44, n= 9) for HA2000.
, 百拇医药
Concentration drop across the synovial lining is related to molecular size
The effect of Mw on the concentration of biopolymer in the four analysed fluid compartments (infusion line; mixed aspirate from the cavity at the end; subsynovial fluid at the end; mean femoral lymph) is plotted out in Fig. 3A. Concentrations were expressed as a percentage of the infused concentration to facilitate comparison with FD20. Figure 3B shows the same results but with anatomical location rather than Mw as the horizontal axis, to demonstrate the striking change in the concentration profile across the joint lining as HA chain length increases.
, 百拇医药
Concentrations are normalized as a percentage of the infused level. FD values are the mean ±S.E.M. for all experiments; for HA140 n= 7 joints; for HA500 n= 3 joints; for HA2000 n= 9 joints. A, graded relation between concentration of polymer in a given fluid compartment and the weight-average molecular mass of the HA, Mw. B, change in concentration profile across joint lining for polymers of increasing molecular size. Same results as above. In, infusate; Cav, joint cavity; Ss, subsynovium; L, femoral lymph.
, http://www.100md.com
The concentration profile showed an obvious, graded dependence on molecular size. Concentration fell sharply across the synovial lining for the largest polymer, HA2000; it fell by intermediate amounts for HA500 and HA140, and not at all for FD20. For HA2000 the concentration in the cavity increased substantially, reaching 253%± 18% of the infused level in the mixed terminal aspirate (n= 9), while its concentration in the subsynovial fluid and lymph fell to 23.9%± 9.8% and 15.5%± 2.3% of the infused level, respectively. The profile demonstrated marked molecular sieving of HA2000. For HA500 the change in concentration across the membrane was less marked: cavity aspirate 176%± 31%; subsynovial fluid 76.0%± 6.6%; femoral lymph 34.3%± 13.6% (n= 3). For HA140 the gradient was even smaller (cavity aspirate 124%± 6%; subsynovial fluid 92.5%± 7.9%; femoral lymph 45.9%± 7.1%; n= 7). For FD20 the profile across the synovial lining was flat, as expected for a freely permeating reference solute. The end-aspirate FD20 concentration showed no significant increase, being 29.59 ± 0.41 μg ml–1 or 98.6% of infused concentration in the seven HA140 experiments, and 100.6%± 2.3% for the entire study. Similarly, the subsynovial FD20 concentration was not significantly reduced (95.8%± 4.8%; P > 0.05, one-way ANOVA). Only in femoral lymph was the FD20 concentration reduced significantly, namely to 52.6%± 6.2% (whole series, P= 0.003, one-way ANOVA). This defined Vv, the dilution of the joint lymph by skin and muscle lymph in the femoral lymphatics (see Methods).
, http://www.100md.com
The differences between the terminal intra-articular concentrations of HA2000, HA500 and HA140 were statistically significant, thereby demonstrating that intra-articular HA accumulation during trans-synovial filtration increases with chain length (P < 0.0001, one-way ANOVA). The differences between subsynovial concentrations were likewise significant, showing that subsynovial HA concentration declines with increasing chain length (P < 0.0001, one-way ANOVA). The same was true for lymph HA concentration (P < 0.005, one-way ANOVA). The results thus showed a clear, consistent pattern; the greater the chain length, the greater the upstream accumulation of polymer and the lower the downstream concentration. It can be concluded that HA of reduced molecular mass is less well retained in the joint cavity than HA of high molecular mass.
, 百拇医药
Magnitude of reflected fractions for different hyaluronan chain lengths
The HA reflected the fraction across the composite cavity-to-lymph barrier, Rlymph, was calculated from the lymph HA and FD20 concentrations; see ‘Methods’. For HA2000 Rlymph averaged 0.65 ± 0.05 (n= 10 rabbits) at the concentration and filtration rate used in this study. (Different values are obtained under different filtration conditions; Sabaratnam et al. 2004; Lu et al. 2004). For HA500 Rlymph fell to 0.43 ± 0.09 (n= 3), and for HA140 it fell further to 0.19 ± 0.05 (n= 7) (Fig. 4). The reflected fraction was thus >3 times bigger for HA2000 than HA140 (P < 0.0001, one-way ANOVA). Even for HA140, however, Rlymph was significantly greater than zero (P < 0.05, one sample t test).
, 百拇医药
See text for definitions. Mean ±S.E.M.; n as for Fig. 3. The effect of chain length on reflected fraction is statistically significant in each case (see text).
A similar, graded dependence of reflected fraction on molecular size was demonstrated by Rsyn, which is calculated from the subsynovial HA and FD20 concentrations (Fig. 4) (P= 0.0001, ANOVA). The graded effect of molecular size on reflection was also confirmed by Rasp, which is calculated from upstream HA accumulation rather than downstream reduction (P < 0.0001, ANOVA). The numerical values for the three estimates of HA reflection, Rlymph, Rsyn and Rasp are summarized in Table 1, along with the global averages of the three reflected fractions. The results demonstrate a positive, monotonic relation between reflected fraction and the radius of gyration of the hydrated molecular domain; see Discussion. The tendency of Rasp to be a little lower than Rlymph or Rsyn could be partly due to retention of a small amount of sieved HA at the synovial surface despite the terminal mixing cycles.
, 百拇医药
For technical reasons (Methods) the infused concentration was set at 4 mg ml–1 in the main seven experiments with HA140, whereas it was lowered to 0.2 mg ml–1 for HA500 and HA2000 to prevent outflow buffering from interfering with the matching of filtration rates. To test whether HA140 concentration might have influenced the results, HA140 was infused at 0.2 mg ml–1 in three further studies. It was not possible to analyse the filtrate HA140 concentrations reliably in these three studies for technical reasons (overlap of the small HA140 chromatograph peak with other peaks), but it was possible to measure the accumulation of reflected HA140 in the joint cavity and thus calculate Rasp. The HA140 concentration in the terminal aspirate increased to 118%± 12% of the infused concentration, an increase similar to that observed for HA140 infusates at 4 mg ml–1 (124%± 6%). Rasp averaged 0.028 ± 0.017 for HA140 at 0.2 mg ml–1 (n= 3), compared with 0.052 ± 0.012 at 4.0 mg ml–1 (n= 7) (P= 0.31, unpaired t test). These results demonstrate that the low reflected fraction for HA140 relative to HA500 and HA2000 was not due to differences in the infused HA140 concentration.
, 百拇医药
Chromatogram retention times for reflected and transmitted HA chains
The retention time of the HA chromatogram peak, tret, is negatively related to average molecular mass over the range of interest (Fig. 1A and Coleman et al. 1997). Since the infused preparation was polydisperse (see Methods), tret was measured in the sieved and reflected samples, to assess whether the longer HA chains in the mixture might be selectively separated from the shorter ones during trans-synovial filtration. The HA2000 experiments provided the most favourable conditions for detecting within-sample selective molecular sieving, because reflection was most pronounced with HA2000 (Figs 3 and 4). In every joint infused with HA2000, tret for HA in the end-experiment aspirate from the joint cavity was smaller than that of the paired infusate tret (Fig. 5). This demonstrates a preferential retention of the longer chains in the joint cavity (P < 0.001, paired t test). In keeping with this, lymph tret increased (indicating a greater proportion of shorter chains), but the change did not reach statistical significance due to an increase in variance. The tret values were infusate 7.37 ± 0.05 min; end-experiment aspirate 7.21 ± 0.06 min; subsynovial fluid 7.44 ± 0.19 min; lymph 7.51 ± 0.14 min (n= 9 animals). A similar pattern, namely reduced aspirate tret and increased lymph tret, was noted by Sabaratnam et al. (2003).
, http://www.100md.com
The reduced HPLC retention time of HA2000 aspirated from the joint cavity at the end of the filtration period indicates selective intra-articular retention of the longer chains of the polydisperse HA2000 preparation (n= 9). Size-related separation was not detected within the HA140 preparation (n= 10) or HA500 preparation (n= 3).
In the case of HA140, which experiences relatively little molecular sieving (Figs 3 and 4), tret did not differ significantly between infusate, end-experiment aspirate, subsynovial fluid and lymph (Fig. 5) (P= 0.18, one way ANOVA, n= 10 preparations). The tret for HA500 likewise showed no significant differences between compartments. tret for FD20 was essentially identical in all four fluid compartments, as expected for a freely permeating solute (infusate 11.07 ± 0.03 min, end-experiment joint aspirate 11.02 ± 0.05 min, subsynovial fluid 11.03 ± 0.06 min, lymph 11.09 ± 0.02 min; pooled results from all experiments, P= 0.80, one-way ANOVA).
, http://www.100md.com
The tret results thus provided limited evidence for selective synovial permeability to chains of differing size within a heterodisperse HA sample of physiological average molecular size.
Discussion
The principal new finding was that the permeation of HA through synovial interstitial matrix is a graded function of HA molecular size. We discuss below the earlier work in this area, the mechanisms involved, the relevance of the findings to the concentration polarization hypothesis, and their pathophysiological significance.
, 百拇医药
Previous evidence indicating synovial molecular sieving
The quantitative results here are in broad agreement with a histological assessment of HA permeation by Asari et al. (1998). They found that intra-articular fluorescein-labelled HA of Mw 2300 kDa hardly penetrated the synovial lining of the dog knee, whereas 840 kDa HA penetrated it more readily. Size-selective molecular sieving by synovium is also indicated by the observation that the clearance of large proteoglycans and radio-colloid from the joint cavity is slower than that of albumin (Page-Thomas et al. 1987; Reimann et al. 1989).
, http://www.100md.com
Coleman et al. (2000) studied the relation between intra-articular fluid pressure and trans-synovial volume flow in the presence of 2000 kDa, 530 kDa, 300 kDa and 90 kDa HA. Although they did not collect subsynovial fluid or lymph, they noted that HA accumulated in the joint cavity over the course of the study in proportion to its molecular size, as found here (Fig. 3). In this earlier work, however, filtration rate varied both during each study and between classes of HA. The Rasp values reported by Coleman et al. (2000) were 0.79 for HA2000 at a mean filtration rate of 3 μl min–1, 0.25 for HA 530 at 10 μl min–1 and 0.12 for HA90 at 15 μl min–1. Since it was later shown that the HA reflected fraction falls when filtration rate is increased (Sabaratnam et al. 2004), the findings of Coleman et al. (2000) were potentially open to an alternative interpretation. In the present study, however, the filtration rate was held constant throughout the experiment and was matched for each HA preparation. The new results confirm that the increase in Rasp with increasing Mw is independent of the effect of filtration rate on Rasp.
, 百拇医药
Quantitative relation between reflected fraction and molecular domain size; equivalent pore size of extracellular matrix
The relation between HA reflected fraction and molecular domain radius is plotted in Fig. 6, with additional results for HA, 2000 kDa dextran and 67 kDa albumin from previous work (see figure legend). The simplest explanation for the relation is that large solutes experience partial steric exclusion from the water space in the extracellular matrix. The greater the solute domain size relative to the effective pore size in the matrix, the greater the solute exclusion and hence the greater the reflection of the solute during filtration. This hypothesis can be assessed quantitatively as follows.
, 百拇医药
Molecular radius is expressed as the Flory Rg value except in the case of albumin (Stokes–Einstein radius 3.6 nm). with S.E.M. bars are mean of HA reflected fractions from current study. without S.E.M. are Rasp values from Coleman et al. (2000), measured at lower filtration rates and a higher HA concentration (3.6 mg ml–1). is the reflected fraction for plasma albumin (Coleman et al. 2000). is reflected fraction for 2000 kDa dextran (Scott et al. 2000b). is reflection coefficient for HA2000 at 0.2 mg ml–1 extrapolated from the relation between reflected fraction and filtration rate; bars are 95% confidence limits (Sabaratnam et al. 2004). Dashed line is the theoretical relation between reflection coefficient and solid sphere radius rs (lower, offset x scale) based on steric exclusion in cylindrical pores of radius 33 nm (Anderson & Malone, 1974). To plot the polymers it is assumed that, due to their chain flexibility, effective polymer rs= 0.26Rg (see text). Dotted line is theoretical based on steric exclusion in a randomly orientated molecular fibre matrix (Ogston, 1970) of polymer concentration 11.5 mg ml–1 (Scott et al. 2003).
, 百拇医药
A simple approach to characterizing the irregular aqueous channels through an extracellular biopolymer matrix in terms of a single parameter is to work out the radius of a cylindrical, water-filled pore that would have the same excluding properties – the ‘equivalent cylindrical pore’ model. For a pore of radius rp and a solid, neutral spherical solute of radius rs, the solute reflection coefficient can be related to rs/rp through the solute partition coefficient , which is the complement of the steric exclusion fraction for solute in a narrow pore (Anderson & Malone, 1974; Curry, 1984) -
, 百拇医药
where = (1 –rs/rp)2. The term 1 –rs/rp describes the radial space available to the solute centre of mass relative to water. Second-order exclusion through charge interactions could increase the degree of exclusion but are not considered here. Eqn (1) was applied to dextran reflection data by Scott et al. (2000b), who estimated the upper limit of synovial rp to be = 87 nm. A further estimate can be made from the reflection coefficient = 0.91 for HA2000 (Sabaratnam et al. 2004), though this entails estimating the equivalent solid sphere radius rs of HA2000. This is much smaller than Rg (101–177 nm, Table 1), because the HA chain is flexible and can deform to access pores smaller than Rg. Munch et al. (1979) showed that linear polyelectrolytes behave hydrodynamically as an equivalent solid sphere whose radius is less than half Rg. Their data for the molecular sieving of 3000 kDa hydrolysed polyacrylamide across a Nucleopore membrane with 100 nm radius pores indicates that rs is 0.26Rg at = 0.91. Substitution of rs= 0.26Rg and = 0.91 for HA2000 into eqn (1) provides a new estimate of synovial matrix rp, namely 33–59 nm. From this we can assess whether steric exclusion can explain the experimental results, as follows.
, http://www.100md.com
The dashed line in Fig. 6 shows theoretical synovial reflection coefficients for rp= 33 nm (eqn (1)). For HA140 (Rg= 21–34 nm; rs= 0.26Rg), the predicted is 0.08–0.09, which is close to the average observed reflected fraction, 0.088 ± 0.033 (Table 1). For HA500 (Rg= 39–72 nm; rs= 0.26Rg), the predicted is 0.27–0.29, which is close to the average observed reflected fraction, 0.26 ± 0.06. Steric exclusion can thus account reasonably well for the size dependence of HA reflection. It should be noted that the value adopted for solid sphere scaling (here rs/Rg= 0.26) does not materially affect the above conclusion, because a bigger scaling factor increases not only the value of rs for a given Rg, but also the estimate of rp. Concentration polarization theory predicts that the HA2000 reflected fraction in the present study should fall well below the theoretical at raised filtration rates, which it does (Fig. 6) (Sabaratnam et al. 2004; Lu et al. 2004). This aspect is discussed further below.
, 百拇医药
An alternative approach to steric exclusion is to relate it to a more realistic structural model (Ogston, 1970). If the molecular chains (glycosaminoglycans, proteoglycan core proteins, glycoproteins) of the extracellular, extrafibrillar matrix are randomly distributed, the fraction of the water space Kav available to a solid spherical solute of radius rs is given by:
where C is polymer concentration, v is effective specific volume (0.65 ml g–1) and rf is chain radius. Rabbit synovial extrafibrillar space contains 3.9 mg ml–1 glycosaminoglycan (Price et al. 1996) and an estimated 7.6 mg ml–1 of glycoprotein and proteoglycan core protein, giving a total C= 11.5 mg ml–1 (Scott et al. 2003). Fibre radius rf is approximated here as 0.87 nm, which is a weighted average for glycosaminoglycan rf (0.6 nm) and protein rf (1.0 nm). Theoretical for these values, based on eqns (1) and (2), are shown as a dotted line in Fig. 6. The fit appears less good than with the simple equivalent pore model, but both models show that steric exclusion has the potential to explain the experimental findings.
, http://www.100md.com
Relation of findings to concentration polarization and attenuation of synovial fluid loss
The volume of synovial fluid in a normal joint is very small (rabbit knee 50 μl, human knee 500–1000 μl), and a major role of HA is to prevent this vital fluid from being squeezed out of the joint cavity during a sustained flexion (outflow buffering, see Introduction). The effectiveness of outflow buffering is a graded function of HA molecular size (Coleman et al. 2000), and buffering is lost as Mw is reduced, i.e. fluid drains away more rapidly. According to the current concentration polarization hypothesis, the outflow buffering depends on HA reflection (Coleman et al. 1999; Lu et al. 2004). Reflection raises the HA concentration at the membrane surface and the resulting osmotic pressure attenuates the water outflow. Loss of outflow buffering with reduced HA Mw was attributed to a fall in the HA reflected fraction. The new findings in Figs 3 and 4 support this explanation, because they show that the HA reflected fraction does indeed fall as HA chain length is reduced.
, 百拇医药
Both concentration polarization theory and experimental observation indicate a negative relation between HA reflected fraction and trans-synovial filtration rate (Sabaratnam et al. 2004: Lu et al. 2004). This accounts for the finding that the HA2000 reflected fraction (0.63 ± 0.04 at filtration rate 35 ml min–1) is smaller than the measured or theoretical HA2000 reflection coefficient (0.91, in Fig. 6). The difference between measured reflected fraction and theoretical is less marked for the smaller HA preparations, because a fall in Mw increases the HA diffusivity and synovial permeability, reducing the degree of concentration polarization and thus allowing the reflected fraction to approach closer to the reflection coefficient at a given filtration rate.
, 百拇医药
Patho-physiological significance for arthritis
HA chain length is reduced in arthritic effusions and can be as little as 28% of normal size (Balazs et al. 1967; Bjelle et al. 1982; Dahl et al. 1985; Yingsung et al. 2003). The reduced chain length is probably due mainly to degradation by free oxygen radicals in the inflamed joint (McCord, 1974; Greenwall & Moak, 1986; Baker et al. 1989; Halliwell, 1995; Haubeck et al. 1995; Schenck et al. 1995). It is also possible that pro-inflammatory cytokines alter the expression of hyaluronan synthase (HAS) isoforms (Castor & Dorstewitz, 1966; Vuorio et al. 1982). Of the three mammalian isoforms, HAS2 is most commonly expressed and synthesizes large-Mw HA, whereas HAS3 synthesizes chains of only 200–300 kDa. HAS3 is up-regulated by cytokines found in inflamed joints, such as IL-1and TNF (Spicer & Nguyen, 1999; Ohno et al. 2001). Dahl & Husby (1985) reported HA chains as small as 50 kDa in the supernatant of cultured rheumatoid cells. The present findings show that such chains would quickly leak out of the joint cavity, leaving behind the longer chains. In rheumatoid synovial fluid the hyaluronan loss may be offset to some degree by the presence of SHAP–hyaluronan complexes, but their quantitative significance is not yet clear (Yingsung et al. 2003).
, 百拇医药
The reduced HA chainlength in arthritis has a number of functional consequences. First, the fall in intrinsic viscosity (Table 1) reduces the hydrodynamic lubricating ability of the synovial fluid, as long recognized (Balazs et al. 1967). Second, in arthritis joint movements are limited and effusion pressures are always positive, as in the present study. Our results indicate that a fall in HA reflected fraction will reduce the build-up of a concentration polarization layer at the synovial surface during fluid drainage. There will therefore be less osmotic buffering of outflow from the arthritic joint. This can be viewed as a useful biological response in that it facilitates the resolution of the intra-articular effusion that inevitably accompanies arthritis. The reduced reflection also has a downside, however, because the increased loss of HA will contribute to the known fall in its intra-articular concentration, which will reduce further the lubricating power of the intra-articular fluid. Increased loss of HA also calls for a faster synthesis rate to replace it. The latter is a known effect of some pro-inflammatory cytokines, but it increases the metabolic burden on the synovial lining cells, which are in a relatively hostile, acidotic and hypoxic environment in arthritis.
, 百拇医药
To summarize, a study of HA transport in vivo showed that the retention of HA in joints depends critically on its chain length. Reduction of hyaluronan chain length, as in arthritis, increases the rate of loss of not only HA but also synovial fluid itself, because there is reduced concentration polarization and hence reduced outflow buffering.
References
Anderson JL & Malone DM (1974). Mechanism of osmotic flow in porous membranes. Biophys J 14, 957–982.
, 百拇医药
Asari A, Miyauchi S, Matsuzaka S, Ito T, Kominami E & Uchiyama Y (1998). Molecular weight-dependent effects of hyaluronate on the arthritic synovium. Arch Histol Cytol 61, 125–135.
Baker MS, Green SP & Lowther DA (1989). Changes in the viscosity of hyaluronic acid after exposure to a myeloperoxidase-derived oxidant. Arthritis Rheum 32, 461–469.
Balazs EA, Watson D, Duff IF & Roseman S (1967). Hyaluronic acid in synovial fluid. I. Molecular parameters of hyaluronic acid in normal and arthritic human fluids. Arthritis Rheum 10, 357–376.
, 百拇医药
Bjelle A, Anderson T & Granath K (1982). Molecular weight distribution of hyaluronic acid of human synovial fluid in rheumatic diseases. Scand J Rheumatol 12, 133–138.
Brown TJ, Alcorn D & Fraser JRE (1999). Absorption of hyaluronan applied to the surface of intact skin. J Invest Dermatol 113, 740–746.
Brown TJ, Laurent UBG & Fraser JRE (1991). Turnover of hyaluronan in synovial joints: elimination of labelled hyaluronan from the knee joint of the rabbit. Exp Physiol 76, 125–134.
, 百拇医药
Castor CW & Dorstewitz EL (1966). Abnormalities of connective tissue cells cultured from patients with rheumatoid arthritis. I. Relative unresponsiveness of rheumatoid synovial cells to hydrocortisone. J Lab Clin Med 68, 306–313.
Coleman PJ, Scott D, Mason RM & Levick JR (1999). Characterization of the effect of high molecular weight hyaluronan on trans-synovial flow in rabbit knees. J Physiol 514, 265–282.
Coleman PJ, Scott D, Mason RM & Levick JR (2000). Role of hyaluronan chain length in buffering interstitial flow across synovium in rabbits. J Physiol 526, 425–434.
, 百拇医药
Coleman PJ, Scott D, Ray J, Mason RM & Levick JR (1997). Hyaluronan secretion into the synovial cavity of rabbit knees and comparison with albumin turnover. J Physiol 503, 645–656.
Curry FE (1984). Mechanics and thermodynamics of transcapillary exchange. In Handbook of Physiology, section 2, The Cardiovascular System, vol IV The Microcirculation, ed Renkin EM & Michel CC, pp. 309–374, American Physiological Society, Bethesda.
, 百拇医药
Dahl LB, Dahl IM, Engstrm-Laurent A & Granath K (1985). Concentration and molecular weight of sodium hyaluronate in synovial fluid from patients with rheumatoid arthritis and other arthropathies. Ann Rheumatic Dis 44, 817–822.
Dahl IMS & Husby G (1985). Hyaluronic acid production in vitro by synovial lining cells from normal and rheumatoid joints. Ann Rheumatic Dis 44, 647–657.
Davies DV (1946). The lymphatics of the synovial membrane. J Anat 80, 21–23.
, http://www.100md.com
Day AJ & Sheehan JK (2001). Hyaluronan: polysaccharide chaos to protein organisation. Curr Opin Struct Biol 11, 617–622.
de Gennes P-G (1979). Scaling Concepts in Polymer Physics. Cornell University Press, Ithaca and London.
Denlinger JL (1982). Metabolism of Sodium Hyaluronate in Articular and Ocular Tissues. Doctoral Dissertation, Université de Sciences et Techniques de Lille, Lille.
Flory PJ (1971). Principles of Polymer Chemistry. Cornell University Press, Ithaca, New York.
, 百拇医药
Fraser JRE & Laurent TC (1996). Hyaluronan. In Extracellular Matrix, Vol 2. Molecular Components and Interactions, ed. Comper WD, pp. 141–199, Harwood Academic Publishers, Amsterdam.
Greenwall RA & Moak SA (1986). Degradation of hyaluronic acid by polymorphonuclear leucocytes. Inflammation 10, 15–30.
Halliwell B (1995). Oxygen radicals, nitric oxide and human inflammatory joint disease. Ann Rheumatic Dis 54, 505–510.
, 百拇医药
Hardingham T (2004). Solution properties of hyaluronan. In Chemistry and Biology of Hyalurona, ed. Garg HG & Hales CA, pp. 1–19, Elsevier, Amsterdam.
Haubeck HD, Kock R, Fischer D-C, Leur EV, Hoffmeister K & Greiling H (1995). Transforming growth factor B1, a major stimulator of hyaluronan synthesis in human synovial lining cells. Arthritis Rheum 38, 669–677.
Johnson M, Kamm R, Ethier CR & Pedley T (1987). Scaling laws and the effects of concentration polarization on the permeability of hyaluronic acid. Physicochem Hydrodynamics 9, 427–444.
, 百拇医药
Laurent TC, Ryan M & Pietrustkiewicz A (1960). Fractionation of hyaluronic acid. The polydispersity of hyaluronic acid from the bovine vitreous body. Biochemica et Biophysica acta 42, 476–485.
Levick JR (1987). Flow through interstitium and other fibrous matrices. Q J Exp Physiol 72, 409–438.
Levick JR, Price FM & Mason RM (1996). Synovial matrix-synovial fluid system of joints. In Extracellular Matrix, Vol. 1. Tissue Function ed. Comper WD, pp. 328–377. Harwood Academic Publishers, Amsterdam.
, 百拇医药
Lu Y, Levick JR & Wang W (2004). Concentration polarization of hyaluronan on the surface of the synovial lining of infused joints. J Physiol 561, 559–573.
Lu Y, Levick JR & Wang W (2005). Synovial fluid retention in pressurised joint cavities is achieved by hyaluronan concentration polarisation. Microcirculation (in press).
McCord JM (1974). Free radicals and inflammation: protection of synovial fluid by superoxide dismutase. Science 185, 529–531.
, 百拇医药
McDonald JN & Levick JR (1995). Effect of intra-articular hyaluronan on pressure–flow relation across synovium in anaesthetised rabbits. J Physiol 485, 179–193.
Munch WD, Zestar LP & Anderson JL (1979). Rejection of polyelectrolytes from microporous membranes. J Memb Sci 5, 77–102.
Nagai S (1987). The basic study of the synovial membrane of the normal rabbit knee by using the induction system of carbon particles. J Tokyo Med Coll 45, 488–498.
, 百拇医药
Ogston AG (1970). On the interaction of solute molecules with porous networks. J Phys Chem 74, 668–669.
Ohno S, Tanimoto K, Fujimoto K, Ijuin C, Honda K, Tanaka N, Doi T, Nakahara M & Tanne K (2001). Molecular cloning of rabbit hyaluronan synthases and their expression patterns in synovial membrane and articular cartilage. Biochim Biophys Acta 1520, 71–78.
Page-Thomas DP, Bard D, King B & Dingle JT (1987). Clearance of proteoglycan from joint cavities. Ann Rheum Dis 46, 934–937.
, 百拇医药
Price FM, Levick JR & Mason RM (1996). Glycosaminoglycan concentration in synovium & other tissues of rabbit knee in relation to hydraulic resistance. J Physiol 495, 803–820.
Reimann I, Vittas D, Nielsen SL & Svalastoga E (1989). Lymphatic transport from normal and synovitic knees in rabbits. Acta Orthop Scand 60, 185–187.
Sabaratnam S, Coleman PJ, Badrick E, Mason RM & Levick JR (2002a). Interactive effect of chondroitin sulphate C and hyaluronan on fluid movement across rabbit synovium. J Physiol 540, 271–284.
, 百拇医药
Sabaratnam S, Mason RM & Levick JR (2002b). Inside-out cannulation of fine lymphatic trunks used to quantify coupling between trans-synovial flow and lymphatic drainage from rabbit knees. Microvasc Res 64, 1–13.
Sabaratnam S, Mason RM & Levick JR (2003). Molecular sieving of hyaluronan by synovial interstitial matrix and lymphatic capillary endothelium evaluated by lymph analysis in rabbits. Microvasc Res 66, 227–236. Erratum to all figures. Microvasc Res (2004) 67, 245–246.
, http://www.100md.com
Sabaratnam S, Mason RM & Levick JR (2004). Filtration rate-dependence of hyaluronan reflection by joint-to-lymph barrier in rabbits; evidence for concentration polarisation. J Physiol 557, 909–922.
Schenck P, Schneider S, Miehlke R & Prehm P (1995). Synthesis and degradation of hyaluronate by synovia from patients with rheumatoid arthritis. J Rheumatol 22, 400–405.
Scott D, Coleman PJ, Abiona A, Ashhurst DE, Mason RM & Levick JR (1998b). Effect of depletion of glycosaminoglycans and non-collagenous proteins on interstitial hydraulic permeability in rabbit synovium. J Physiol 511, 629–643.
, 百拇医药
Scott D, Coleman PJ, Mason RM & Levick JR (1998a). Direct evidence for the partial reflection of hyaluronan molecules by the lining of joints during trans-synovial flow. J Physiol 508, 610–623.
Scott D, Coleman PJ, Mason RM & Levick JR (2000a). Concentration-dependence of interstitial flow buffering by hyaluronan in synovial joints. Microvasc Res 59, 345–353.
Scott D, Coleman PJ, Mason RM & Levick JR (2000b). Action of polysaccharides of similar average mass but differing molecular Volume and charge on fluid drainage through synovial interstitium in rabbit knees. J Physiol 528, 609–618.
, 百拇医药
Scott D, Levick JR & Miserocchi G (2003). Non-linear dependence of interstitial fluid pressure on joint cavity pressure and implications for interstitial resistance in rabbit knee. Acta Physiol Scand 179, 93–101.
Spicer AP & Nguyen TK (1999). Mammalian hyaluronan synthase: investigation of functional relationships in vivo. Biochem Soc Trans 27, 109–115.
Vuorio E, Einola S, Hakkarainen S & Penttinen S (1982). Synthesis of underpolymerized hyaluronic acid by fibroblasts cultured from rheumatoid and non-rheumatoid synovitis. Rheumatol Int 2, 97–102.
, 百拇医药
Wik HB & Wik O (1998). Rheology of hyaluronan. In The Chemistry, Biology and Medical Applications of Hyaluronan and its Derivatives, ed. Laurent TC, pp. 25–32. Portland Press, London.
Xu H, Edwards JC, Banerji S, Prevo R, Jackson DG & Athanasou NA (2003). Distribution of lymphatic vessels in normal and arthritic human synovial tissue. Ann Rheum Dis 62, 1277–1279.
Yamashita S & Ohkubo M (1993). Distribution and three-dimensional reconstruction of lymphatic vessels of the elbow joint capsule of rabbits. Acta Anat Nipponica 68, 513–521.
Yingsung W, Zhuo L, Morgelin M, Yoneda M, Kida D, Watanabe H, Ishiguro N, Iwata H & Kimata K (2003). Molecular heterogeneity of the SHAP-hyaluronan complex. J Biol Chem 278, 32710–32718., http://www.100md.com(S. Sabaratnam, V. Arunan,)