Anal. Chem. 1995,67,522-527
Optical Detection of Macromolecular Heparin via Selective Coextraction into Thin Polymeric Films Enju Wang* Department of Chemistty, St. John's University, 8000 Utopia Parkway, Jamaica, New Yo& 11439 Mark E. Meyerhoff
Department of Chemistty, The Universiiy of Michigan, Ann Arbor, Michigan 48109
Victor C. Yang College of Pharmacy, The Univemty of Michigan, Ann Abor, Michigan 48109
Thin plasticized polymer films,poly(vinyl chloride) doped with a speci6c ion pairing quaternary ammonium compound, tridodecylmethylammonium chloride, and a lipophilic pH indicator, 3-hydr0~-4-(4-nitrophenyhm)phenyl octadeconate, are shown to exhibit signiscantand analytically useful optical response toward macromolecular heparin. The response mechanism is based on favorable extraction of heparin into the bulk organic film, owing to the specific ion-pairingcomplexation reaction between the quaternary ammonium species and the pobanion. A simultaneous coextraction of hydrogen ions results in protonation of the pH chromophore and hence a change in the optical absorbance of the polymeric film. When used in a limited volume/hed exposure (10 min) detection mode, film absorbances change as a function of the initial heparin concentration in the range of 0.2-3.0 units/mL (1.2-18 pg/mL). The practical measurement response time is controlled by heparin diffusion through the stagnant diffusion layer adjacent to the surface of the film as well as within the bulk of the polymer film and is shown to increase with the molecular weight of the heparin species tested. No optical response to heparin is observed when a strong heparin complexingagent (e.g., protamine) is present in the test solution, suggestingthat the polymer film can be used to conveniently monitor heparin-protamine (or other antagonist) titrations. The theory relating to the operation of the sensingfilm in either the equilibrium or the kinetic mode and the SeledVh' Of the optimized film to heparin relative to small anions are presented. Over the past few years, a wide range of selective chemistries used previously to devise potentiometric polymer membranetype ion-selective electrodes (ISEs) have been adapted successfully to develop new optical ion-selective sensor^.^-^ In general, these (1) Morf, W. E.; Seiiler, K; Lehmann, B.; Behringer, Ch.; Hartman K; Simon, W. Pure Ajfil. Chem. 1989, 61, 1613.
(2) Simon, W Mod, W. E.; Seiler, K.; Spichiger-Keller, U. E. Frensenius]. Anal. Chem. 1990,337, 26. (3) Seiler, IC;Simon, W. Anal. Chim. Acta 1992,266, 73. (4) Gantzer, M. L.; Hemmes, P. R ; Wong, D. Eur. Pat Appl. EP 153.641 (Cl. G01N33/00), Sept 4, 1985.
522 Analytical Chemistry, Vol. 67, No. 3, February 7, 7995
optical sensors are prepared by incorporating appropriate lipophilic ionophores and pH indicators into thin (2-5 pm) plasticized poly (vinyl chloride) (WC)films cast on glass plates. In the case of cation sensors, selective extraction of the target cation by a neutral ionophore (e.g., valinomycin for K+) is coupled to a simultaneous deprotonation of the pH indicator in the bulk of the organic film, resulting a change in the film absorbance at a given The same basic strategy has also been applied for detection of specific anions,+12 although in these cases, simultaneous protonation of the indicator within the lilm occurs regardless of whether the ionophore serves as a neutral, charged, or general dissociated anion exchangetype carrier. Unlike the response of polymer membrane electrodes, which is governed by electrochemical equilibria at phase boundaries, the response of ion-selective optical sensors is dependent on a mass/charge balance reaction within the bulk of the polymeric film, and thus, a stable optical signal can be achieved only when the entire film is in chemical (analyte) equilibrium with the sample s ~ l u t i o n . ~ J ~ Recently we demonstrated that specially formulated polymer membranes doped with tridodecylmethylammonium ions (TDMA) exhibit siflcant potentiometric response to submicromolar levels of heparin,14-17 a highly sulfated polysaccharide used routinely as an anticoagulant in various medical procedures. The (5) Ng, R H.; Sparks, M. S.; Statland, B. E. Clin. Chem. 1992,38, 1371. (6) Spichiger, U. E.; Seiler, K; Wang,K; Suter, G.; Morf, W. E.; Simon, W. Proceedings of EC04, The Hague, Netherlands, 12-13 March, 1991. Proc. SPIE-Inf. Sot. Opt. Eng. 1991, 1510,118. (7) Lerchi, M.; Bakker, E.; Rusterholtz, B.; Simon, w. Anal. Chem. 1 9 9 2 , 6 4 , 1534. (8)Lerchi,M.; h i t t e r , E.; Simon,W.; Pretsch, E. Anal. Chem. 1994,66,1713. (9) Tan,S. S. S.;Hauser. P. C.; Chaniotakis. N. A; Suter, G.;Simon, W. Chimia 1 9 8 9 , 4 3 . 257. (10) Tan,S. S. S.; Hauser, P. C.; Wang, K; Fluri, IC; Seiler, K.; Rusterholtz, B.; Suter, G.; Kruettli, M.; Spichiger, U. E.; Simon, W. Anal. Chim. Acta 1991, 225, 35. (11) Wang, E.; Meyerhoff, M. E. Anal. Chim Acta 1993,283,673. (12) Kuratli, M.; Badertscher, M.; Rusterholz, B.; Simon,W. Anal. Chem. 1993, 65,3473. (13) Janata, J. Anal. Chem. 1990,62, 33R (14) Ma, S.-C.; Meyerhoff, M. E.; Yang, V. C. Anal. Chem. 1992, 64,394. (15) Ma, S.-C.; Yang, V. C.; Fu, B.; Meyerhoff, M. E. Anal. Chem. 1993, 65, 2078. (16) Yun, J.-H.; Ma, S.;Fu,B.;Yang, V. C.; Meyerhoff, M. E.]. Electroanal. 1993, 5,719. (17) Fu,B.; Bakker, E.; Yun, J. H.; Yang, V. C.; Meyerhoff, M. E. Anal. Chem. 1994, 66,2250. 0003-2700/95/0367-0522$9.00/0 Q 1995 American Chemical Society
membrane electrode displays high selectivityfor heparin and has been applied successfully for monitoring heparin levels in whole blood as well as for determining the binding constants for the interaction of heparin with a variety of macromolecules (e.g., protamine, polybrene, etc.) . 1 5 3 The response mechanism has most recently been ascribed to a nonequilibrium steady-state change in the phase boundary potential at the sample/membrane interface.17 This steady-stateresponse results, surprisingly,from the rather favorable extraction of hydrophilic heparin into the polymer membrane due to formation of strong ion pairs with the TDMA species. Herein, we report that this same extraction chemistry can also be used to develop a thin optical sensing polymer film that responds to polyanionic heparin levels in the physiologically relevant concentration range. PRINCIPLE OF OPERATION
To take advantage of the strong ion-pairing reactions of heparin with TDMA (or other &CH3N+ species) to devise an optical heparin sensing film, it is necessary to further dope the polymer film with a lipophilic pH indicator to serve as the chromophore. Coextraction of protons and heparin can then occur simultaneously, heparin complexing with the TDMA &+) and the hydrogen ions protonating the indicator and-), thereby changing the absorbance of the film. For such a system, the overall coextraction equilibrium reaction can be written as follows: zL&
+ zInd, + zHlq + Hep& = Lpep,, + zIndH,,
0
where Ind, refers to the deprotonated pH indicator and Lzrgto the quaternary ammonium species, both in the organic polymer film, while Hrq and Hepi; refer to the protons and heparin in the aqueous sample phase. The equilibrium constant for this reaction can be expressed as
in the sample. The theory governing the response of the optical heparin sensing film in both of these two modes is discussed below. Concentration Mode. To operate optical sensing film in the concentration detection mode, the equilibrium reaction (reaction I) should be reversible. At equilibrium, the relationship between the concentration of heparin in the sample [Hep] and the relative absorbance can be expressed as33
(3)
The sensitivity,s (slope), of the response function, a vs log[Hepl, is then given by
Since the charge on the heparin polyanion, 2, is very high (z 70) ,19the slope of the optical response of the sensing film should be extremely small over a wide range of heparin concentrations. Thus, such a concentration mode measurement method would not provide useful analytical signals. Limited Volume Mode. If the following conditions are met, (a) the equilibrium constant for reaction I is very large, (b) the sample is maintained at an appropriate pH (via a buffer) so that sample heparin concentration controls the equilibrium position of the coextraction process (reaction I), (c) the volume of sample is limited, and (d) [L+lVr >> z[HeplinitVs(where 6 and V, represent the volumes of the polymer film and aqueous test sample, respectively), then the equilibrium heparin concentration ([Hep]) in aqueous test solution is negligible compared to its initial concentration ([Heplinit), Le., [HeplfiJ[Hepl > 100. That is, before all the L+ is effectively titrated with heparin, the following relationship holds: x
z[HeplhitVs = [IndHl V, The optical signal, i.e., absorbance, is dependent on the amount of indicator in a given form within the film. The relative absorbance (a)of the polymer film, defined as the fraction of the indicator in the deprotonated form, can be expressed as
where A is the measured absorbance at a given wavelength, A0 and A1 are the absorbances of the film when the indicator in its fully protonated and deprotonated forms, respectively, and [Ind-I and [Ind~]are the concentrations of the deprotonated indicator and the total amount of indicator in the film. Based on the required mass balance and electroneutrality conditions within the polymer film, efficient ion coextraction requires that the total cation and anionic sites in the membrane be nearly equal, i.e., [Ind~]= [bl.Depending on how sample is introduced to the optical sensing film, the film can be operated in either the concentration mode, where the concentration of the analyte remains unchanged, or in the limited volume mode, where the extraction chemistry alters the final concentration of analyte
(5)
The relative absorbance can then be derived as
Thus when the sensing film is operated in the limited volume mode, the amount of heparin that can be detected is controlled by the total quaternary ammonium and pH indicator sites within the organic film. From eq 6, the sensitivity,s (aa/a[Hep]iniJ, can be expressed as
Hence, the sensitivity in this mode of operation is inversely proportional to the total concentration of the indicator (or the ion pairing agent) in the film, but proportional to the phase volume ratio of sample solution and the organic polymer film. Since the Analytical Chemistry, Vol. 67, No. 3, February 1, 1995
523
charge of heparin is very high, the mole sensitivity for heparin detection is also quite high, and the range of heparin concentrations detectable within a limited sample volume is very narrow. The key remaining issue is the expected response time of the polymer films toward heparin. If heparin complexation with the quaternary ammonium ion sites in the film is relatively rapid, then optical response time will be controlled by mass transport of heparin from the bulk sample solution to the membrane/polymer film interface and diffusion of the complex within the bulk of the organic film. The diffusion coefficient for the latter process has recently been estimated by potentiometric measurements to be in the range of 3 x lo-* -1 x 10-9 cm2/s depending on the plasticizer content of the W C films.17 The rate of mass transport to the surface of the film will be controlled by both convection within the bulk of the sample solution (i.e., via stirring) and diffusion through the stagnant Nernst diffusion layer adjacent to the surface of the polymer film.For the limited volume measurement mode, the flux of heparin into the film will continually decrease, owing to the diminution in the heparin concentration in the bulk solution and the decrease in available quaternary ammonium species at the surface of the film. Thus, in practice, equilibrium response, even using a limited volume of sample, would take several hours to achieve for samples containing relatively low concentrations of heparin. Consequently, practical application of the optical heparin sensing films will require use of the limited volume measurement mode in conjunction with a fixed exposure time of the film with the sample (Le., nonequilibrium measurements). EXPERIMENTAL SECTION
Reagents. Tridodecylmethylammoniumchloride (TIMAC) was obtained from Polyscience Inc. (Wanington, PA), and Aliquat 336s (trioctylmethylammoniumchloride) and toluene were from Aldrich Chemical Co. (Milwaukee, wr) . Tetradodecylammonium bromide (TIDAB), bis(2-ethylhexyl) sebacate (DOS), %hydroxy4(4nitrophenylazo)phenyl octadecanoate (ETH 2412, pH dye), tetrahydrofuran (THF),and high molecular weight poly(viny1 chloride) (WC) were obtained from Fluka (Ronkonkoma, NY). Tris (hydroxymethyl)aminomethane eris)and sodium salicylate were purchased from Sigma (St. Louis, MO). Solid heparin (sodium salt from porcine intestine mucosa, 169 USP units/mg) was from Hepar Industries, Inc. (Franklin, OH). Heparin fragments of different molecular weights, obtained by nitrous acid hydrolysis of heparin, were graceously provided by the Institute of Choay (Paris, France) and Pharmuka Laboratories (Gennevitlies, France). Protamine sulfate, poly-L-lysine, and POlybrene were products of S i a Chemical CO. AU other chemicals were commercially available products. Standard solutions and buffers were prepared with distilled deionized water. Polymer F i b Preparation. WC films were prepared as described for ion-selective electrodes that contain 30 wt % of polymeric matrix. THF was used to dissolve the film components to prepare a homogeneous cocktail required for spin casting films. The preferred optical sensing film contained the following: 1.7 wt % of chromophore (ETH 2412), 2.0 wt % of anion carrier (e.g., TDMAC), and 66.3 wt % of bis(2ethylhexyl) sebacate (DOS). Blank reference films had the same composition but without the (18) Bakker, E.; Simon, W. Anal. Chem. 1992,64, 1805. (19) Casu, B. In Heparin and Related Polysaccharides, Structure and Activities;
Annals of the New York Academy of Science 556 Ofosu, F.A, e t al., Eds.; The New York Academy of Sciences: New York, 1989.
524 Analytical Chemistry, Vol. 67, No. 3, February 1, 1995
pH indicator. The film casting solutions were prepared by dissolving a total amount of 200 mg of components (in appropriate proportions) in 2 mL of THF. Films of about 2-5 pm thickness were cast from these solutions onto glass slides (0.9 x 5 cm, 1 mm in thickness) by spin coating. Optical Measurements. Absorbance measurements were made on a UV/vis double beam spectrophotometer (Lambda 6B, Perkin Elmer) with the polymer film coated glass slides placed in a commercial quartz cell (1 x 1 x 4 cm3) containing the test solution. All quantitative measurements were made at 540 nm. The films were first soaked for more than 20 min in a Tris-SO1 buffer (PH 7.4) until the membrane changed color and had a stable absorbance value before the k s t heparin measurement was made. Sample solutions with various heparin concentrations were prepared by adding aliquots of a standard concentrated heparin solution to a fixed volume of buffer (0.05 M Tris-SO1 or phosphate, pH 7.4). For most experiments, calibration was performed by first soaking the films in 10 mL of the heparin sample solution with constant stirring and then transferring the films to the cuvette containing the same sample solution for immediate absorbance measurements. The spectrophotometer was zeroed either by zeroing the absorbance of fully hydrated reference lilm at 540 nm or by zeroing the absorbance of a heparin sensing film at 750 nm (where the dye does not absorb). Optical selectivity coefficients of the films for heparin over other anions were determined using the separate solution m e t h ~ d . ~ .All '~ measurements and preequilibrations were carried out under ambient conditions (23 "C). Iiquid-Liquid Coextraction Experiments. Preliminary liquid-liquid coextraction experiments were performed in a separatory funnel. The organic phase (10 mL toluene or DOS as solvent) contained 9.0 x M of the pH indicator, ETH 2412, and 1.0 x M TDMAC or other quaternary ammonium salts. A 0.2 M NaHC03 solution was used as the background electrolyte of the aqueous phase. After a preliminary equilibration period, in which the organic phase changed color from yellow to purple (from protonated to deprotonated form of indicator), standard amounts of heparin from a stock solution of 1000 units/mL were added to the aqueous phase (10 mL) and the solutions mixed thoroughly for 1 min to achieve a extraction equilibrium. The absorbance (at 540 nm) of the organic phase was then measured after each addition of heparin. Titration of Protamine. The ability of the heparin sensing films to follow the titration of given levels of protamine with heparin was demonstrated by first dissolving a known amount of protamine in 10 mL of Tris-SO1 buffer (PH 7.4). Aliquots of a heparin standard solution were then added stepwise. After each addition of heparin, the film was exposed to the test solution for exactly 10 min and then rapidly transferred to the cuvette (containing buffer only) for measurement of the absorbance at 540 nm. RESULTS AND DISCUSSION
Iiquid-liquid Extraction. Preliminary liquid-liquid extraction experiments were carried out to determine the feasibility of detecting heparin optically by coextraction with protons into an organic phase. Using appropriate amounts of quaternary ammonium salts and the lipophilic pH indicator ETH 2412 dissolved in either toluene or DOS, the pH indicator was found to completely deprotonate at aqueous phase pH values > 8.0; therefore, a 0.2 M NaHC03 solution was used as the background electrolyte/
Relative absorbance (a)
Absorbance
1
0.221
'
e
l
-.2.0
0.18 0.14
0.10 0.06
IHeplinit (U/mL)
Flgure I.Plots of relative aborbance at 540 nm for liquid-liquid extraction of heparin from 0.2 M sodium bicarbonatesolution (10 mL) M into toluene solution (10 mL) containing ETH 2412 at 9.0 x and either TDDACI (0) or TDMACI (0)at 1.O x 10-4 M.
buffer for these preliminary experiments. Figure 1 shows the relative absorbance at 540 nm of two organic phases (toluene) equilibrated with increasing amounts of porcine heparin using two different quaternary species O M A and TDDA) as selective ionpairing agents. As shown, when TDDA is employed, essentially no optical response to heparin is observed over the heparin concentration range tested (note 1unit/mL heparin = 6 pg/mL). On the other hand, when TDMA is present in the organic phase, a very large decrease in the organic phase absorbance at 540 nm is observed, suggesting a significant degree of heparin/proton coextraction into the organic phase. Similar coextraction was found when Aliquat 336 (trioctylmethylammonium)was used as the ion-pairing agent (data not shown). This supports the notion that tight ion-pairing in the organic phase is predicated on accessibility of the positively charged nitrogen to interact electrostaticallywith the negatively charged sulfonate and carboxylate groups present within the heparin structure. The reaction stoichiometry of TDMA with heparin, calculated on the basis of the change in the absorbance of the organic phase (assuming that protonation of one pH indicator molecule corresponds to ionpairing of one quaternary ammonium site with sulfonate/carboxylate group of heparin), was determined to be 12 f 0.5 pg/unit (n = 3). This value agrees well with the reported literature value of 11pg/unit.lg Heparin Response of Polymeric Films. The preliminary liquid-liquid extraction experiments clearly indicated that it is possible to detect heparin optically by coextraction with protons into an appropriately formulated organic phase. It was anticipated, however, that adaptation of this extraction chemistry to thin polymer films would require signilicantly longer equilibration times, owing to the relatively slow diffusion of macromolecular heparin across stagnant layer and within the polymer films. Figure 2 shows the change in the absorbance spectrum with time of a thin (4 pm) PVC (30 wt %)/DOS (66.3 wt %) film containing 2 wt %TDMACand 1.7 wt % ETH 2412 after exposure to a 2 units/ mL solution of heparin in 10 mL of Tris-SO4 buffer, pH 7.4. The polarity of the organic polymer film differs from that of the bulk liquid phases used above, and this changes the effective pKa of the indicator, necessitating the use of a lower pH for the bathing sample solution. As shown in the inset in Figure 2, it takes approximately 20 min for the film to equilibrate fully when the
0.0
0
1
2
3
4
IHeplrnit WmL)
Figure 3. Relative optical response to varying levels of heparin in 10 mL of buffer solution (pH 7.4) after 10 min (0) and 5 h (0) of exposure to the sample solution. The film formulation is the same as used in Figure 2. Data presented are mean values of five sets of measurements, all with fresh membranes.
stirred solution contains 2.0 units/mL of heparin ([HepluJ. When fresh films of the same chemical composition and thickness are exposed to a solution containing 0.5 units/mL of heparin in buffer, the absorbance at 540 nm continues to decrease over an extended 3 h time period but eventually reaches the same value as the 2.0 units/mL solution. This suggests that, when exposed to a limited volume of test solution (10 mL) containing as little as 0.5 unit/ mL of heparin, the total amount of heparin present is sufficient to completely saturate the quaternary ammonium sites within the film (based on reaction I), albeit over an impractically long equilibration period. Indeed, as shown in Figure 3, full equilibration (5 h) of films in contact with a limited volume (10 mL) of sample solution exhibit signiticant optical response at a very low levels of heparin (0-0.5 unit/mL). However, since such long equilibration is not desirable, the films can also be used in a limited volume/fixed exposure time mode, such that the optical response after immersion for only 10 min in the heparin solution can be used for analytical purposes. As expected and as shown in Figure 3, operation in such a nonequilibrium sensing mode results in a significant shift in the detection range for monitoring polyanion Analytical Chemistry, Vol. 67, No. 3, February 1, 1995
525
heparin levels with the polymer films. Of course, a further reduction in the exposure time and/or a change in the limited sample volume exposed to film (or phase volume ratio) would also change the practical measurement range of this sensing system. The response to heparin in the presence of 0.12 M of NaCl (added to Tris-SO4 buffer) in the fixed exposure mode (10 min) exhibits a response curve similar to that shown in Figure 3, when a is calculated using A1 as the absorbance in the buffer/ NaCl background, since the absorbance in this background decreases by about 20%compared to its maximum value in the Tris-HzS04 buffer. This experiment clearly indicates that large optical signals for heparin can be observed even in the presence of greater than physiological levels of chloride ion. It is important to note that the optical response of the optimized polymer film can in fact be reversed by soaking it in large amount of buffer solution or by first soaking in a high concentration of NaCl (3 M) for 5 min to back-extract the heparin from the film (i.e., regenerate TDMACl in the organic phase) and then soaking it in buffer to regenerate a functional film (i.e., TDMA-Ind). However, since the pH indicator incorporated into the optical sensing film does not have very high lipophilicity, repeated regeneration of the film by such treatment results in a continuous decrease in the absolute optical absorption of the film at 540 nm. Use of a more lipophilic form (derivative) of the indicator could potentially solve this problem. In addition, use of much thinner organic films (Le., < 1pm, perhaps in conjunction with evanescent wave-type absorbance measurements to offset loss of transmission path length, could potentially yield a true heparin sensor, with enhanced reversibility and continuous monitoring capabilities. Response of the Polymeric Film to Other Anions. Polymer membranes/films doped with TDMACl have been employed previously to prepare both potentiometric and optical sensors that exhibit response to small anions.10,20In general, such chemistry yields an anion selectivity pattern that depends on the lipophilicity of the anions present in the sample solution. However, as suggested previously during studies with the electrochemical heparin sensor,15the favorable extraction of heparin by TDMA into polymer films can yield a detection method with surprisingly high selectivity for heparin, even relative to very lipophilic smaller anionic species (e.g., perchlorate, salicylate, etc.). To examine the selectivity of the optical heparin sensing film, responses to various anions were determined in the Tris-SO4 buffer, pH 7.4 (using sodium salts). The selectivity was calculated in accordance with the following expression:
a*
X
where Kee is the equilibrium constant for heparin/proton coextraction, and CH,C H ~and ~ , CX are the sample concentrations of protons, heparin, and interferent anions, respectively. For simplicity, anion concentrations rather than activities are used to avoid calculating activity coefficients for the polyanions. Further, the heparin was treated as a singly charged species since measurements were made under nonequilibrium fixed exposure time conditions. The apparent selectivity coefficients were calculated by taking the ratio of heparin and interferent anion concentrations (20) Wegmann, D.; Weiss, H.; Ammann, D.; Mod, W. E.; Retsch, E.; Sugahara, K Simon,W. Amal. Chim. Acta 1986,183, 83.
526 Analytical Chemistry, Vol. 67, No. 3, February 1, 1995
Table I.Apparent Selectlvity Coefficients of the Optlcal Films to Anions**
anion SalINO3-
BrSCNc104-
PPo4"-
log e ; x
anion
-2.5 -2.5
NOz-
-3.3 -3.7 -2.8
HzP04'-
c1-
OAc-
HC03-
-1.5
sod2-
-0.9
citrate
log
e;$
-4.6 -5.8
