Anal. Chem. 1996, 68, 1439-1443
Effect of Surface-Attached Heparin on the Response of Potassium-Selective Electrodes Kelly A. Brooks, John R. Allen,† Pamela W. Feldhoff,‡ and Leonidas G. Bachas*
Department of Chemistry and Center of Membrane Sciences, University of Kentucky, Lexington, Kentucky 40506-0055
Heparin (or hydrolyzed heparin) was covalently attached on the surface of derivatized cellulose triacetate membranes, which were subsequently impregnated with the potassium-selective ionophore valinomycin. The resulting ion-selective electrodes presented near-Nernstian response to potassium and had selectivity coefficients of the same order of magnitude as those of conventional poly(vinyl chloride)-based electrodes. It was found that the heparin layer does not alter significantly the response characteristics of the electrodes. The biological activity of the immobilized heparin (or hydrolyzed heparin) was measured in terms of its inactivation of blood coagulation factor Xa. It was found that the covalently anchored hydrolyzed heparin was not biologically active, but the immobilized heparin was able to inactivate factor Xa. Therefore, by covalently attaching heparin on the surface of ion-selective electrodes, electrodes with improved blood compatibility characteristics may be prepared. The development of ion-selective electrodes (ISEs) capable of monitoring physiologically important ions in biological samples has been the focus of increased investigation.1-6 Among these ISEs, those based on poly(vinyl chloride) (PVC) membranes impregnated with the ionophore valinomycin5,7 have shown excellent response and selectivity for potassium and are the standards against which all other potassium ISEs are compared. Many clinical analyzers based on ISEs have been developed and are commercially available for monitoring analytes in clinical laboratories and at the bedside of critically ill patients.1,8-10 However, when potassium-selective, or other ion-selective, electrodes based on PVC membranes are used in undiluted whole blood for extended periods of time, one encounters the problem of fibrin formation and clotting.11-13 In the case of PVC, the time of maximum thrombus deposition is on the order of 15-20 † Current address: Department of Chemistry, West Virginia State College, P.O. Box 1000, Institute, WV 25112. ‡ J. G. Brown Cancer Center, University of Louisville School of Medicine, 529 South Jackson Street, Louisville, KY 40292. (1) Yim, H.; Kibbey, C. E.; Ma, S.; Kliza, D. M.; Liu, D.; Park, S.; Torre, C. E.; Meyerhoff, M. E. Biosens. Bioelectron. 1993, 8, 1-38. (2) Collison, M. E.; Meyerhoff, M. E. Anal. Chem. 1990, 62, 425A-437A. (3) Rouilly, M.; Rusterholz, B.; Spichiger, U. E.; Simon, W. Clin. Chem. 1990, 36, 466-469. (4) Thompson, M.; Vandenberg, E. T. Clin. Biochem. 1986, 19, 255-261. (5) Oesch, U.; Ammann, D.; Simon, W. Clin. Chem. 1986, 32, 1448-1459. (6) Sibbald, A.; Covington, A. K.; Carter, R. F. Med. Biol. Eng. Comput. 1985, 23, 329-338. (7) Fiedler, U.; Ruzicka, J. Anal. Chim. Acta 1973, 67, 179-193. (8) Burritt, M. F. Clin. Chem. 1990, 36B, 1562-1566. (9) Meyerhoff, M. E. Clin. Chem. 1990, 36B, 1567-1572. (10) Bowers, G.; Brassard, C.; Sena, S. Clin. Chem. 1986, 32, 1437-1447. (11) Meyerhoff, M. E. Trends Anal. Chem. 1993, 12, 257-266.
0003-2700/96/0368-1439$12.00/0
© 1996 American Chemical Society
min.12,14 This interferes with both the potentiometric response and the selectivity patterns of the ISEs.11,13 Therefore, there exists a need to develop biocompatible ion-selective electrodes in order to avoid clotting on the sensor membrane, thereby making longterm in vivo use of ISEs more practical. Several approaches have been proposed for the development of electrodes that are compatible with blood samples. A popular direction has been the replacement of conventional PVC membranes with more biocompatible polymers such as silicone rubber12,15-18 and polyurethane (Tecoflex).12,19,20 In addition, membranes based on copolymers of vinyl chloride, vinyl acetate, and vinyl alcohol gave ISEs that exhibited reduced nonspecific adsorption of proteins,12,21-23 which is a requirement for the development of biocompatible devices.12,14,24,25 Membranes based on polyurethane and silicone rubber, however, demonstrated an overall better performance in terms of lack of adsorption of fibrinogen, which implies a reduced probability for thrombogenesis.12 Aminated PVC and carboxylated PVC (PVC-COOH) have also been proposed as membranes for ISEs because they are resistant to protein adsorption.21,22,26-30 Buck and co-workers found these modified PVC polymers to be useful for applications in blood samples.26-29 Further, they demonstrated that by reducing the amount of plasticizer in electrodes based on PVCCOOH and polyurethane, the biocompatibility of the devices (12) Cha, G. S.; Liu, D.; Meyerhoff, M. E.; Cantor, H. C.; Midgley, A. R.; Goldberg, H. D.; Brown, R. B. Anal. Chem. 1991, 63, 1666-1672. (13) Fogt, E. J. Clin. Chem. 1990, 36B, 1573-1580. (14) Salzman, E. W., Ed. Interaction of the Blood with Natural and Artificial Surfaces; Marcel Dekker: New York, 1981. (15) Marrazza, G.; Mascini, M. Electroanalysis 1992, 4, 41-43. (16) Mascini, M.; Marrazza, G. Anal. Chim. Acta 1990, 231, 125-128. (17) Jenny, H.-B.; Riess, C.; Ammann, D.; Magyar, B.; Asper, R.; Simon, W. Mikrochim. Acta 1980, II, 309-315. (18) Crowe, W.; Mayevsky, A.; Mela, L. Am. J. Physiol. 1977, 233, C56-C60. (19) Lindner, E.; Cosofret, V. V.; Ufer, S.; Buck, R. P.; Kao, W. J.; Neuman, M. R.; Anderson, J. M. J. Biomed. Mater. Res. 1994, 28, 591-601. (20) Espadas-Torre, C.; Meyerhoff, M. E. Anal. Chem. 1995, 67, 3108-3114. (21) Bricker, J.; Daunert, S.; Bachas, L. G.; Valiente, M. Anal. Chem. 1991, 63, 1585-1589. (22) Florido, A.; Daunert, S.; Bachas, L. G. Electroanalysis 1991, 3, 177-182. (23) Durselen, L. F.; Wegman, D.; May, K.; Oesch, U.; Simon, W. Anal. Chem. 1988, 60, 1455-1458. (24) Mannhalter, C. Sens. Actuators B 1993, 11, 273-279. (25) Ratner, B. D. J. Biomed. Mater. Res. 1993, 27, 283-287. (26) Cosofret, V. V.; Lindner, E.; Johnson, T. A.; Neuman, M. R. Talanta 1994, 41, 931-938. (27) Cosofret, V. V.; Erdo¨sy, M.; Lindner, E.; Johnson, T. A.; Buck, R. P. Anal. Lett. 1994, 27, 3039-3063. (28) Lindner, E.; Cosofret, V. V.; Ufer, S.; Johnson, T. A.; Ash, R. B.; Nagle, H. T.; Neuman, M. R.; Buck, R. P. Fresenius J. Anal. Chem. 1993, 346, 584588. (29) Lindner, E.; Cosofret, V. V.; Ufer, S.; Buck, R. P.; Kusy, R. P.; Ash, R. B.; Nagle, H. T. J. Chem. Soc., Faraday Trans. 1993, 89, 361-367. (30) Ma, S.-C.; Meyerhoff, M. E. Mikrochim. Acta 1990, I, 197-208.
Analytical Chemistry, Vol. 68, No. 8, April 15, 1996 1439
improved,19,31 and in most cases the response characteristics of the corresponding ISEs did not change substantially.19,31 In the case of a Ca2+-selective electrode based on ETH 1001 and Tecoflex polyurethane, even membranes with no plasticizer gave functional electrodes.19 Another popular approach when using electrodes in vivo is to supply heparin to the blood stream. Heparin is a linear polysaccharide, and its anticoagulant activity is ascribed to its effects on several steps of the coagulation cascade process that leads to the production of fibrin.32 In one example, an ISE catheter was inserted into a dog that had been treated with heparin.33 Absence of heparin caused deviations in response as compared to the response of the ISE when heparin was present. This method has the advantage of simplicity, in that no special modification of the electrode is needed, but it suffers from the disadvantage of subjecting the patient to heparin treatment, which may cause hemorrhagic complications. Because of its demonstrated anticoagulation properties (heparin can catalyze the binding between antithrombin and thrombin to prevent conversion of fibrinogen to fibrin), heparin has been attached to the surface of various biomaterials, such as polyurethane urea, polystyrene, polypropylene, and poly(dimethylsiloxane)-poly(ethylene oxide) block copolymers in order to improve their blood compatibility during implantation.34-37 The surfaceimmobilized heparin was bioactive and resulted in materials with improved hemocompatibility properties. However, there is no report on the use of such material in sensors. In the present study, heparin was covalently attached to ionselective electrode membranes in order to improve their compatibility with blood samples. Because PVC is not easily derivatized for heparin attachment, cellulose triacetate (CTA) was employed as the polymer matrix. In that respect, asymmetric CTA membranes were used that were prepared by a process similar to that described by Cha and Meyerhoff.38 By using asymmetric membranes, it was possible to attach heparin to only one of the two sides of the membrane. The feasibility of coating ion-selective membranes with heparin and hydrolyzed heparin in order to improve the compatibility of potassium-selective electrodes with biological samples is demonstrated. EXPERIMENTAL SECTION Reagents. Valinomycin, porcine intestinal mucosal heparin, and carbonyldiimidazole (CDI) were obtained from Sigma (St. Louis, MO). Cellulose triacetate and chromatographic grade poly(vinyl chloride) were from Eastman Kodak (Rochester, NY) and Polysciences (Warrington, PA), respectively. Tetrahydrofuran (THF) was purchased from Fisher (Fair Lawn, NJ) and tris(hydroxymethyl)aminomethane (Tris) from Research Organics (31) Cosofret, V. V.; Erdo ¨sy, M.; Buck, R. P.; Kao, W. J.; Anderson, J. M.; Lindner, E.; Neuman, M. R. Analyst 1994, 119, 2283-2292. (32) Danishefsky, I.; Ahrens, M.; Klein, S. Biochim. Biophys. Acta 1977, 498, 215-222. (33) Collison, M. E.; Aebli, G. V.; Petty, J.; Meyerhoff, M. E. Anal. Chem. 1989, 61, 2365-2372. (34) Nilsson, U. R.; Larm, O.; Nilsson, B.; Storm, K. E.; Elwing, H.; Nilsson Ekdahl, K. Scand. J. Immunol. 1993, 37, 349-354. (35) Nojiri, C.; Park, K. D.; Grainger, D. W.; Jacobs, H. A.; Okano, T.; Koyanagi, H.; Kim, S. W. ASAIO Trans. 1990, 36, M168-M172. (36) Grainger, D. W.; Okano, T.; Kim, S. W.; Castner, D. G.; Ratner, B. D.; Briggs, D.; Sung, Y. K. J. Biomed. Mater. Res. 1990, 24, 547-571. (37) Piao, A. Z.; Nojiri, C.; Park, K. D.; Jacobs, H.; Feijen, J.; Kim, S. W. J. Biomater. Sci. Polym. Ed. 1990, 1, 299-313. (38) Cha, G. S.; Meyerhoff, M. E. Talanta 1989, 36, 271-278.
1440
Analytical Chemistry, Vol. 68, No. 8, April 15, 1996
(Cleveland, OH). The plasticizers o-nitrophenyl octyl ether (NPOE) and bis(2-ethylhexyl) sebacate (DOS), as well as the lipophilic salt potassium tetrakis(chlorophenyl)borate (KTClPB), were purchased from Fluka (Ronkonkoma, NY). Di-n-amyl phthalate (DAP) was a product of Chem Service (Media, PA). Chloroform, methylene chloride, 1,1,2,2-tetrachloroethane, and all chloride salts were purchased from Aldrich (Milwaukee, WI). All standard solutions and buffers were prepared with deionized (MilliQ, Millipore, Bedford, MA) distilled water. Hydrolyzed heparin was prepared by controlled nitrous acid hydrolysis. The heparin preparations were characterized by gel filtration utilizing a Spherogel TSK 3000 SW (7.5 mm × 300 mm) column that was calibrated with sodium styrene sulfonate standards (Pressure Chemical, Pittsburgh, PA) and a Hitachi liquid chromatograph. Heparin (or hydrolyzed heparin) was eluted with 0.50 M Na2SO4. The commercially obtained heparin had Mw 23 450 and Mn 12 680, while the hydrolyzed heparin had Mw 14 930 and Mn 4254. Cellulose Triacetate Control Membranes. A base layer of the membrane was prepared by dissolving 74 mg of CTA in 1.1 mL of methylene chloride, 0.40 mL of chloroform, and 0.40 mL of 1,1,2,2-tetrachloroethane and casting the mixture in a 31-mmi.d. glass ring placed on a glass plate. After 2 days, the membrane was removed from the glass mold, and an ionophore-containing cocktail was poured into a 22-mm-i.d. ring that was placed on top of the base layer. Unless otherwise stated, the ionophorecontaining cocktail was composed of 1 mg of valinomycin, 100 µL of NPOE, 35 mg of CTA, and 0.42 mg of KTClPB, all dissolved in a mixture of 0.80 mL of methylene chloride and 0.80 mL of chloroform. The solvent was allowed to evaporate for 2 days as the two layers fused into a single asymmetric membrane. Asymmetric Hydrolyzed CTA Control Membranes. The base layer of the membrane was prepared in the same manner as above. After the solvent was allowed to evaporate for 2 days, the membrane, which has raised edges at the area of contact with the glass ring, was removed from the mold and floated for 4.5 h in 1.0 M NaOH. Because of the raised edges, only one of the surfaces of the membrane was in contact with the NaOH solution. After the hydrolysis step was completed, the membrane was rinsed with deionized water. The second layer, containing the active membrane components, was then added as described above. Asymmetric Heparin-Coated CTA Membranes. The base layer was prepared and hydrolyzed as described above. However, after treatment with NaOH and rinsing with water, the bottom side of the membranes (i.e., the one with the free OH groups) was immersed in cold deionized water, and 324 mg of CDI was added in five increments over a 15-min period, giving a final concentration of 0.10 M CDI. Following this activation period, the membranes were immediately incubated overnight in a 1% (w/v) heparin solution in 1.0 M sodium carbonate, pH 10, to promote covalent heparin attachment. Both hydrolyzed and nonhydrolyzed heparin were used. The membranes were removed from the heparin solution and rinsed sequentially with 0.10 M NaHCO3 (pH 8.5), distilled water, acetate buffer (pH 4.0), and distilled water. Finally, the second layer was added to the unmodified side of the membranes as described above. Figure 1 outlines this membrane preparation scheme. ISE Construction and Setup. Once the asymmetric membranes were prepared, 6-mm-i.d. disks were cut from them and mounted in Philips IS-561 electrode bodies (Glasblaserei Mo¨ller,
0.10 M NaCl in 0.10 M Tris-HCl (pH 7.5), to which additions of 0.010 M KCl (containing 0.10 M NaCl) were made. Determination of Amount of Heparin Coated on the CTA Surface. After the heparin or hydrolyzed heparin was immobilized on the surface of the CTA membrane, the amount of bound heparin was measured by a colorimetric assay using toluidine blue, a dye that binds to heparin. This assay can determine the amount of heparin attached to the surface of the CTA membrane.40 Biological (Anticoagulant) Activity of CTA Membranes. The biological activity for the heparin-bound and control membranes was determined utilizing an antifactor Xa activity assay (Helena, Beaumont, TX). The active surface of the membranes was placed in contact with a solution containing an excess of factor Xa and antithrombin and was allowed to incubate for a set time. This incubation allows heparin on the membrane surface to activate antithrombin, which in turn neutralizes (i.e., inhibits) factor Xa in proportion to the amount of heparin present. A chromogenic substrate for factor Xa was then added. This substrate contains a chromophore attached to a polypeptide. Factor Xa is a peptidase and is able to cleave this chromophore from the polypeptide, which results in a shift in the absorption peak of the chromophore. The remaining amount of active factor Xa could, thus, be determined by monitoring the absorbance at 405 nm. Figure 1. Schematic representation of the protocol used to prepare ion-selective cellulose triacetate membranes with surface-attached heparin or hydrolyzed heparin.
Zurich), with the heparin-modified surface facing the sample solution. The two types of control asymmetric membranes (i.e., with hydrolyzed and nonhydrolyzed CTA surfaces) were also mounted in the same manner. The internal filling solution was 0.010 M KCl, except when testing the ISE for its response to NH4+, at which point a buffered internal filling solution was used, 0.100 M Tris-HCl, pH 7.5. A double-junction Ag/AgCl electrode (Orion Model 90-02-00) with an Orion (90-02-02) internal filling solution was used as the reference electrode. The outer compartment of this reference was filled with 0.100 M Tris-HCl, pH 7.5. Potentiometric responses were measured with a Fisher Accumet digital pH/mV meter (Model 910) and registered on a Linear (Model 1200; Reno, NV) strip-chart recorder. Prior to initial use, the electrodes were conditioned in 0.010 M KCl solution overnight. Between each use, the ISEs were conditioned in 0.010 M KCl for at least 1 h. Evaluation of Membrane Response and Selectivity. Calibration of the electrodes was accomplished in a beaker that was thermostated at 25 °C with a Fisher Isotemp circulator bath. Additions of incrementally sized aliquots of cation standard solutions were made to a stirred initial buffer solution of 5.00 mL of 0.100 M Tris-HCl (pH 7.5). The response of the electrodes was monitored and recorded. Calibration curves were constructed by plotting the potential versus the logarithm of the concentration of the cation present in the buffer solution. The detection limits were determined by using the IUPAC recommendations.39 Selectivity coefficients for potassium over sodium were determined by using the fixed-interference method39 and sample solutions of (39) Buck, R. P.; Lindner, E. Pure Appl. Chem. 1994, 66, 2527-2536.
RESULTS AND DISCUSSION Asymmetric cellulose triacetate membranes have been successfully prepared using other neutral ion carriers.38 The goal of the present study is to investigate whether asymmetric CTA membranes can also be prepared using valinomycin as the ionophore to yield electrodes that are selective for potassium. More importantly, this study is aimed at demonstrating that heparin can be attached covalently to the surface of the membrane without a significant loss in the response characteristics of the ISE, thus leading to ISEs with increased blood compatibility. The potentiometric response characteristics of the control membranes were evaluated initially. In particular, the influence of the membrane composition on the response of the ISEs was investigated by using various plasticizers in an attempt to produce electrodes with optimal detection limits and Nernstian responses. A comparison of different plasticizers indicated the inferiority of DOS and DAP in comparison to NPOE. Therefore, NPOE was used in all subsequent membrane preparations. It was also found that addition of the lipophilic salt, KTClPB, was effective in helping produce calibration plots with slopes of 52 ( 3 mV/decade (n ) 3) and in increasing the linear response range from 1 × 10-4 to at least 1 × 10-2 M K+. The CTA-based control membrane electrodes had a detection limit of (4.1 ( 1.0) × 10-5 M (n ) 3) for potassium and selectivity coefficient for potassium over sodium of (1.1 ( 0.4) × 10-3 (n ) 5). Although there is some reduction in the detection limit and selectivity coefficient for potassium over sodium when compared to those of conventional PVC-based ISEs that employ the same ionophore,41 these response characteristics are sufficient to allow the use of the CTA-based ISE in the determination of potassium in blood with
