Biolubrication: Hyaluronic Acid and the Influence on Its Interfacial Viscosity of an Antiinflammatory Drug

Introduction. How to understand the lubrication of animal joints? In seeking to dissect this complex problem into manageable components, it is interesting to consider hyaluronic acid (HA), more currently called hyaluron, a linear polysaccharide with the repeat unit of the disaccharide D-glucoronic acid and N-acetyl-Dglucosamine (see Figure 1a), which exists ubiquitously in most organisms with highest concentration in soft connective tissues, such as cartilage, synovial joints, and the vitreous body of the eye.1,2 The most accepted structure is a 2-fold, tapelike helix motif with five hydrogen bonds per tetrasaccharide unit of HA, which helps explain the ubiquitous interactions of HA with lipids, membranes, and also itself.1,3,4 In pharmacological applications, it is widely studied for use as a therapeutic agent in joint disease and related issues.5 In parallel, it has caught the attention of engineers, and various controversial lubrication schemes have been proposed to explain its function in the lubrication of animal joints.6-9 However, it is interesting to note that the pharmacological and the engineering communities have developed to be largely independent of one another. In the pharmacological community, the relation of chemical structure to biomechanical functions such as lubrication and shock absorbency is seldom considered.1,5 Conversely, engineers recognize lubrication functions but seldom consider the possible influence on lubricating properties of physiological conditions, such as ionic strength and drug additives.10-12 The present study is a modest attempt to begin to rectify this split by examining how the interfacial rheology of HA is modulated by the presence of an antiinflammatory drug, D-penicillamine (D-pen),13 which is capable of hydrogen bonding to HA. We selected D-pen as a model system; as a drug, it is rarely prescribed these days. In the mammalian body, the mechanical function of HA is surely complex because it coexists with plasma proteins. The viscosity properties of protein-free HA in bulk dilute solution, in the presence and in the absence of D-pen, were studied recently by Colby and coworkers.14 The drug was found to cause a moderate reduction of the viscosity, and the reasonable hypothesis was advanced that the reason was disruption of intramolecular hydrogen bonding.14 In the study below, HA was compressed between two surfaces such that the local concentration of HA much exceeded dilute. As will be seen below, this situation generates an effect of the opposite sign, an enhanced effective viscosity, and we hypothesize that this reflects the introduction of transient intermolecular cross-links via transient intermolecular hydrogen bonding. To the best of our knowledge, no previous rheological study of HA was conducted in these conditions where HA was controllably confined to spacings as small as a few nanometers, thus potentially mimicking the condition the flow of synovial fluid between gaps in cartilage surfaces. As this comparison must ultimately fail because cartilage is softer than the solid surfaces that we studied, we restricted the comparison to low compressive loads, such that neither cartilage nor the model surfaces that we employed deformed appreciably. To put this in perspective, this study focuses on the liquid side of the solid-liquid interface. Experimental Section. The experimental platform was a home-built surface forces apparatus (SFA) modified to study the viscoelastic shear response of confined fluids as a function of frequency and shear rate.15,16 However, as mica is negatively charged in water and the negatively charged HA does not adsorb to it owing to charge-charge repulsion,17 it was necessary to find a method by which to connect HA to the solid surface underneath. The strategy to modify the solid surface via a polymer connector was made with the following criteria in mind. First, the charge should be positive to promote adsorption onto it of the negatively charged polysaccharide. Second, surface coverage of the connector should be controllable in order to control the quantity of HA adsorbed. Finally, the thin-film viscoelastic responses of the connector itself should be known as a base of comparison. Given these considerations, supported soft connective layers were prepared by allowing the adsorption of quaternized poly(4-vinylpyridine) (QPVP), purchased as poly(4-vinylpyridine) from Polymer Source Inc., Quebec, Canada, and quaternized in our laboratory to a degree of 98%.18 The weight-average molecular weight of the narrow-distribution samples after quaternization was 69 000 g mol-1. Control experiments without added HA showed the final incompressible thickness of adsorbed QPVP-QPVP was 2 nm, corresponding to the adsorbed amount of <1 mg m-2. From previous studies in this laboratory, we know the force-distance relations and thin-film viscoelastic shear responses of QPVP, which were qualitatively different19,20 from the new data reported below. The adsorption concentration was kept low, 5 ppm, such that the polymer connector is known to adsorb in the thin “pancake” configuration.19 The sample preparation was as follows. Freshly cleaved mica sheets were glued onto cylindrical silica disks using Epon 1004, immersed into 0.005 mg mL-1 Figure 1. Structures of (a) hyaluronic acid (HA) and (b) D-penicillamine, whose three zones of potential hydrogen bonding are circled. 973 Macromolecules 2003, 36, 973-976