1371 PERMEABILITY OF HYALURONIC ACID SOLUTIONS The Effects of Matrix Concentration, Calcium, and pH MARCUS E. CARR, JR. and NORTIN M. HADLER The permeability of hyaluronic acid solutions was measured in a chamber closed at each end by semipermeable membranes. permeability was strongly concentration dependent below 1.0 gm% hyaluronic acid. Small changes in the ionic environment of 1.0 gm% hyaluronic acid solutions also altered permeability. The permeability of such solutions decreased by a factor of 4.7 in the presence of as little a s 5 mM calcium ion and showed optimum permeability at pH 7.0. The pH optimum was not noted in hyaluronic acid preparations known to contain only small amounts of protein. Hyaluronic acid is found in connective tissue throughout the body. In the early 1950s, Day (1,2) demonstrated the important role of hyaluronic acid in determining the rate of solvent flow through the extracellular space of connective tissue. Subsequently, only the study of Ogston and Sherman in 1961 (3) has provided some insight into the mechanisms of bulk solvent flow through a hyaluronate matrix. In the current study we used a simple apparatus to directly measure the flow of solvent through solutions of hyaluronic acid under vari- From the Departments of Medicine and Bacteriology (Immunology), University of North Carolina, School of Medicine, Chapel Hill, North Carolina. Supported by Research Grant AM-I9992 from the National Institutes of Health. Marcus E. Carr, Jr., PhD, MD: Post-Doctoral Trainee; Nortin M. Hadler, MI): Associate Professor of Medicine and Bacteriologv (Immunology) and recipient of an Established Investigatorship from the American ileart Association. Address reprint requests to Dr. Hadler, Dept. of Medicine, 938 Faculty Laboratory Office Building 231H, University of North Carolina, Chapel Hill, NC 27514. Submitted for publication January 28, 1980; accepted in revised form July 30, 1980. Arthritis and Rheumatism, Vol. 23, No. 12 (December 1980) ous conditions of pH, ionic strength, and calcium concentration. The flow of a solution through a well defined volume of sample material is governed by Darcy’s equation. This is an empirical relationship based on a classic experiment of flow through filter beds (4). D = Q v v:/Ftp D is the permeability or Darcy constant of the medium. All of the other parameters-viscosity of the permeating solution (v), length of the sample column (q),cross sectional area (F) in the direction of flow (Q/t), and the pressure applied ( p e a r e known or are easily measured. MATERIALS AND METHODS Hyaluronic acid from human umbilical cord (H-175 I, Grade 1) was purchased as a potassium salt from Sigma Chemical Company (St. Louis, MO). It was dissolved by placing the hyaluronic acid in test tubes, layering the solvent on top, and allowing the tube to stand in a water bath at 37°C for 24 hours. This was sufficient to allow all bubbles to rise to the top. The tubes were then placed in a refrigerator at 2°C. Just prior to use, the tubes were removed from the refrigerator and brought to 35°C by heating in a water bath for 30 minutes. In addition to the Sigma samples, two other umbilical cord hyaluronic acid preparations were examined. Hyaluronic acid from human umbilical cords (batch #7017) was obtained from Miles Laboratories (Goodwood, South Africa). The sample was received as a powder of a potassium salt, and was stated to contain 1.4% protein and to have an intrinsic viscosity of 69 at 25°C. A third sample in the form of Sodium hyaluronate was purchased from the laboratory of Professor E. A. Balazs. The preparation had a stated specific viscosity of 2700 centipoise and was reported to contain 0.9% protein. Ionic strength of all solutions was set by adding NaCl from a stock I.OM solution. All solutions contained 0. IOM sodium chloride unless otherwise indicated. Phosphate buffers CARR AND HADLER 1372 were prepared as 0.50M stock solutions and the pH was set by the addition of 8.OM NaOH or concentrated H,PO,. The buffer was always diluted 1 to 10 in all hyaluronic acid solutions. The pH of all hyaluronic acid solutions was 7.0 unless otherwise indicated. All chemicals were of reagent grade and all solutions were prepared in doubly distilled and deionized water. Perfusion measurements were made by using the apparatus illustrated in Figure 1. The perfusing fluid was forced through the hyaluronic acid matrix by a pressure head generated by a nitrogen gas tank. The gauges on the tank were used as a first approximation of the applied pressure, but precise pressure measurements were made with a mercury rnanometer. Centimeters of mercury were converted to dynes/crnz by multiplying by the factor 1.33322 X 104. Flow was measured by observing the movement of the meniscus in a pipette which served as a flowmeter. The attachment of graph paper to the shaft of the pipette allowed for h e r gradations of rneasurement. By switching between 1.0 and 0.1 ml pipettes, we were able to measure both fast and slow flow. All experiments were performed at ambient temperature. In all cases, the perfusing fluid was identical in composition to the hyaluronic solution being perfused minus the hyaluronic acid. This apparatus is very similar to the one described previously (5). The hyaluronic acid solution was contained within a plexiglass chamber (Figure 2) closed at both ends by membranes. The chamber itself was threaded at both ends in order to accept the female portion of a Millipore filter holder (cat. #SXO001300). The Unipore membranes were purchased from Biorad Laboratories and had a pore size of 1000 Angstroms. The membranes were supported by grids which were cut from the male end of a Millipore filter holder. Both grids and rnembranes were sealed into place by silicon rings. The inside diameter of the chamber was 0.476 crn and the length 1.0 cm. The chambers were filled with hyaluronic acid solutions by two techniques: for the more dilute solutions, pasteur pipettes were used; for the more concentrated solutions, the gel-like flow in silicon ring flow out membrane support grid .silicon ring perfusion chamber silicon ring mernbrane grid 'support Figure 2. Chamber used to contain hyaluronic acid solutions during perfusion measurements. material was spooned into the chamber. In all cases, loading was accomplished by removing one end of the chamber. The membranes themselves offer some resistance to flow. Blank measurements were made prior to each hyaluronic acid solution measurement with the chamber filled with water and at the same pressure as was used to measure flow through the hyaluronic acid solutions. The inverse of the flow value for the filters plus water was subtracted from the inverse of the flow value for the filters plus hyaluronic acid. Ultraviolet (UV) spectra of the three hyaluronic solutions were obtained by utilizing a Cary 118 spectrophotometer. The solutions were scanned from 300 to 260 nm in 1.0 cm pathlength quartz cells. The hyaluronic acid samples were all prepared in 0.1M KCl and were unbuffered. A solution of Nacetylglucosamine and glucuronic acid in O. 1M KCl was scanned as a control. A solution of 0.1M KCl served as a blank for all scans. RESULTS U Figure 1. Schematic diagram of the apparatus used to measure the flow through hyaluronic acid solutions. 1, Nitrogen tank used to apply pressure head. 2, Mercury manometer: a, mercury pot; b, vertical mercury column; e, meter stick; d, spill pot. 3, Three-way valve to control pressure head. 4, Connection to allow loading of perfusion fluid. 5, Chamber containing test sample. 6, Three-way stopcock to allow switching between flowmeters. 7, e, 1.0 ml pipette for fast flow; f, 0.10 ml pipette for slow flow. The ability of the described instrument to measure accurately the very low flows observed in concentrated hyaluronic acid is illustrated in Figure 3. A major point of concern was whether the hyaluronic acid would be forced through the solvent and onto the exiting membrane. If this occurred, one would observe nonlinear flow; the flow would gradually slow with time. The fact that the volume flow is linear with time therefore indicates that coating of the membrane is not occurring during the period of measurement. A second method used to check for membrane coating was to reverse the direction of solvent flow after a given period of flow in one direction. The results obtained with reversed flow were essentially the same as the original results. PERMEABILITY OF HA SOLUTIONS 1373 x 3.0 t -1 t 10” 2.0t- i 0 1.0 0 0 0 [HA] in gm0lo TIME Figure 3. Volume flow of water through hyaluronic acid (Sigma preparation) solutions versus time in seconds. The concentrations of hyaluronic acid are indicated at the right in gm%. All solutions contained 0.10M sodium chloride and were buffered at pH 7.0 with 0.05M sodium phosphate. Permeability displays a large concentration dependence below 1.5 gm% (Figure 4). The effect of Ca++ ion concentration on the permeability of 1% hyaluronic acid solutions (Sigma preparation) is depicted in Figure 5. Even solutions with concentrations of Ca++as low as 5 mM are nearly five times less permeable than those with no added Ca++.The effect of pH on the permeability of 1% hyaluronic acid solutions (Sigma preparations) is depicted in Figure 6. At neutral pH there is an increase in permeability which falls off at either higher or lower pH. The results of UV scans of the three hyaluronic acid preparations are presented in Figure 7. As can be seen, the Sigma hyaluronic acid Preparation has an absorbance at 280 nm that is 2.2 times that of the Miles product. The control solution of nearly equimolar amounts of N-acetylglucosamine and glucuronic acid has a negligible absorbance between 300 and 260 nm. Figure 4. The effect of the concentration of hyaluronic acid (HA) on the permeability or Darcy constant (D) of the solution. These solutions are all buffered at pH 7.0 in 0.05 sodium phosphate, 0.10 sodium chloride. If the sodium chloride is omitted from the 1.0% solution, D measures 49.6 X lo-”. The study illustrated employed the Sigma preparation of hyaluronic acid. 6.3 X ml/second through a 1.0 gm% solution as compared to 5.4 X for a 0.84 gm solution reported by Ogston (3). The rather marked effects of small changes in the ionic environment of the hyaluronic acid matrix on the permeability of the solution have not been detailed previously, but they were not totally unexpected in view of the reported effects of such changes on the physical properties and conformation of hyaluronic acid. The structure of hyaluronic acid as revealed by x-ray diffraction is a left hand helix. The aggregation of these :I lo” 2 0 I 0 DISCUSSION The permeability of a hyaluronate matrix to a simple solvent at neutral pH reported here is in agreement with that reported previously. We found a flow of 50 I 100 I I 150 200 [Ca++]in mM Figure 5. The effect of calcium ions on the permeability of 1.0 gm% hyaluronic acid solutions (Sigma preparation). All solutions were buffered at pH 7.0 with 0.05M phosphate. 1374 CARR AND HADLER chains in the putty-like matrices subjected to x-ray diffraction is dependent on pH and counter ions. Sodium ions tend to stabilize the hyaluronic acid matrix by forming intermolecular salt bridges (6,7). Under certain conditions the chains are arranged anti-parallel to one another (8,9) and can be further stabilized through calcium bridges (9). The rheologic properties of solutions of hyaluronic acid are also sensitive to ionic strength and pH (10). Light- scattering (11,12) and optical rotary dispersion (ORD) studies (13) have provided additional evidence of effect of ionic strength and pH on the conformation of hyaluronate in solution. Analysis by I3C nuclear magnetic resonance (NMR) has demonstrated the importance of calcium in determining the order present in a thick matrix of hyaluronic acid (14). The permeability pH optimum of 7.0 (Figure 6) cannot be explained by any known property of the hyaluronate itself in the matrix (15). A contaminant, most likely a protein, probably serves as a modulator of matrix permeability. In 'recent experiments we have demonstrated that the Miles preparation, which appears to have the least protein content (Figure 7), lacks the pH optimum of the Sigma preparation. Future studies will aim at the identification and isolation of the constituent present in some hyaluronic acid preparations that imparts the pH sensitivity to matrix permeability. The current study illustrates the importance of comparing the physical properties of preparations of matrix of varying I I I NAG+ GLU.10M KCI 0 260 I 270 280 290 3 0 0 nm h Figure 7. Ultraviolet (UV) spectra of three different preparations of hyaluronic acid. All solutions contained 0. IOM potassium chloride. HAS = Sigma preparation (4.08 mg/ml); HAB = Balazs preparation (4.58 mg/ml); HAM = Miles preparation (4.70 mg/ml); NAG + GLU = N-acetyl glucosamine (5.0 mg/ml) + glucuronic acid (4.89 mg/ml). All measurements were made at 22.1'C. The UV absorption at this range of wavelength reflects contamination of the hyaluronic acid preparation particularly by proteins. degrees of purity. We may at times purify away the most interesting properties of the matrix. The sensitivity of hyaluronate matrix permeability to slight alterations in solvent suggests a scheme by which a cell communicates with and thereby alters its extracellular environment (16). The cell may control its juxtaposed extracellular matrix by secreting various ions that act on proteins which in turn modulate changes in the matrix. Given this model, the matrix no longer appears to be a space filling substance, but rather takes on the role of a dynamic extension of the cell that is directly under cellular control. ' I a 16 1 D 4tn I "5.5 6 l a a I 7 T 8 PH Figure 6. The effect of pH on the permeability (D X lo-") of 1.0 gm% hyaluronic acid solutions (Sigma preparation). All solutions contain 0.10Msodium chloride. The Miles preparation has no demonstrable pH sensitivity of permeability. REFERENCES 1. Day TD: Connective tissue permeability and the mode of action of hyaluronidase. Nature 166:785-786, 1950 2. Day TD: The permeability of interstitial connective tissue and the nature of the interfibrillary substance. J Physiol 17:l-8, 1952 3. Ogston AG, Sherman TF: Effects of hyaluronic acid upon diffusion of solutes and flow of solvent. J Physiol 156:6774, 1961 4. Darcy H: Les Fontaines Publiques de la Ville de Dijon. Pans, 1856 PERMEABILITY OF HA SOLUTIONS 5. Carr ME, Shen LL, Hermans J: Mass-length ratio of fi- 6. 7. 8. 9. 10. brin fibers from gel permeation and light scattering. Biopolymers 16:l-15, 1977 Guss JM, Hukins DWL, Smith PJC, Winter WT, Arnott S, Moorhouse R, Rees DA: Hyaluronic acid molecular conformation and interactions in two sodium salts. J Mol Biol 95:359-384, 1975 Winter WT, Smith PJC, Arnott S: Hyaluronic acid: structure of a fully extended 3-fold helical sodium salt and comparison with the less extended 4-fold helical forms. J Mol Biol 99:219-235, 1975 Sheehan JK, Gardner KH, Atkins EDT: Hyaluronic acid: a double helical structure in the presence of potassium at low pH and found also with the cations ammonium, rubidium and caesium. J Mol Biol 117:113-135, 1977 Winter WT, Arnott S: Hyaluronic acid: the role of divalent cations in conformation and packing. J Mol Biol 117:761-784, 1977 Gibbs DA, Merrill EW, Smith KA, Balazs EA: Rheology of hyaluronic acid. Biopolymers 6:777-79 1, 1968 1375 11. Hallett FR, Gray AL: Quasi-elastic light scattering studies of hyaluronic acid solutions. Biochim Biophys Acta 343:648-655, 1974 12. Mathews MB, Decker L: Conformation of hyaluronate in neutral and alkaline solutions. Biochim Biophys Acta 498:259-263, 1977 13. Chakrabarti B: Effect of counterions on the conformation of hyaluronic acid. Arch Biochem Biophys 180:146-150, 1977 14. Napier MA, Hadler NM: Effect of calcium on structure and function of a hyaluronic acid matrix: carbon-13 nuclear magnetic resonance analysis and the diffusional behavior of small solutes. Proc Natl Acad Sci USA 75:22612265, 1978 15. Hadler NM, Napier MA: Structure of hyaluronic acid in synovial fluids and its influence on the movement of solutes. Semin Arthritis Rheum 7:141-152, 1977 16. Hadler NM: The biology of the extracellular space. Clinics Rheum Dis (in press) New Approaches to Internal Medicine The course “New Approaches to Internal Medicine” will be held February 8-14, 1981 at the Wildwood Inn, Snowmass, Colorado. The objective will be to present the definition, pathogenesis, appropriate diagnostic approach and best current management to problems commonly seen in general medical practice. Registration is $250. Physicians practicing in Colorado may attend for $190. For further information, contact the Office of Postgraduate Medical Education, The University of Colorado School of Medicine, 4200 East Ninth Avenue, Denver, CO, 80262 or call 303-3945241.