Anal. Chem. 2002, 74, 5937-5941
Nitric Oxide-Releasing Fluorescence-Based Oxygen Sensing Polymeric Films Mark H. Schoenfisch,† Huiping Zhang, Megan C. Frost, and Mark E. Meyerhoff*
Department of Chemistry, University of Michigan, Ann Arbor, Michigan 48109-1055
The in vitro analytical performance of fluorescence-based oxygen sensing polymeric films prepared with silicone rubbers that spontaneously release nitric oxide (NO) is examined. The use of NO-release polymers for fabricating functional optical sensors is proposed as a potential solution to lingering biocompatibility and concomitant performance problems encountered with prototype intravascular optical oxygen sensors. Plasticized silicone rubber films formulated with two distinct types of diazeniumdiolate NO donors release NO for more than 24 h. The optical oxygen sensing films prepared by doping these NO release polymeric materials with oxygen indicators (pyrene/ perylene donor/acceptor pair) display different analytical responses, as compared to controls without NO release capability. Nonlinear Stern-Volmer behavior is observed for single-layer NO release oxygen sensors owing to heterogeneous environments for the pyrene/perylene pair and a concomitant quenching of the fluorescence by excess amine sites in such films. In contrast, a dual-layer configuration using an underlying NO-release silicone rubber layer covered with a second polymeric layer containing the fluorescent indicators is shown to yield identical sensitivity and linearity toward oxygen as conventional non-NO-releasing oxygen sensing films, while still providing surface NO fluxes necessary to yield more thromboresistive devices. Recent studies in this laboratory1,2 and elsewhere3,4 have demonstrated the potential utility of nitric oxide (NO)-releasing polymers in creating more thromboresistant material/blood interfaces. Indeed, NO is well-known to be a potent inhibitor of both platelet adhesion and aggregation and is, in fact, continuously released at low levels (1 × 10-10 mol cm-2 min-1) by the vascular endothelium.5-7 This contributes significantly to the nonthrom* To whom correspondence should be addressed. Email: mmeyerho@ umich.edu. † Current address: Department of Chemistry, University of North Carolina, Chapel Hill, NC, 27599. (1) Mowery, K. A.; Schoenfisch, M. H.; Saavedra, J. E.; Keefer, L. K.; Meyerhoff, M. E. Biomaterials 2000, 21, 9-21. (2) Zhang, H.; Annich, G. M.; Miskulin, J.; Osterholzer, K.; Merz, S. I.; Bartlett, R. H.; Meyerhoff, M. E. Biomaterials 2002, 23, 1485-1494. (3) Smith, D. J.; Chakravarth, D.; Pulfey, S.; Simmons, M.; Hrabie, J. A.; Citro, M. L.; Saavedra, J. E.; Davies, K. M.; Hutshell, T. C.; Mooradian, D. L.; Hanson, S. K.; Keefer, L. K. J. Med. Chem. 1996, 39, 1148-1156. (4) Duan, X.; Lewis, R. S. Biomaterials 2002, 23, 1197-1203. 10.1021/ac026075y CCC: $22.00 Published on Web 11/01/2002
© 2002 American Chemical Society
bogenic properties of this innermost layer of healthy blood vessel walls.8 One potential analytical application of such NO-releasing polymers is the fabrication of in vivo chemical sensors that function more reliably once implanted in blood vessels owing to their enhanced thromboresistivity derived from the continuous flux of NO from the surface of the sensor. It has been shown that a variety of electrochemical ion and gas sensors can be fabricated with outer polymeric membranes/coatings that release low levels of NO without negatively impacting the analytical response characteristics of these devices.9,10 Furthermore, both Schoenfisch et al.11 and Frost et al.12 have clearly demonstrated the improved in vivo thromboresistivity and concomitant analytical performance that can be achieved when amperometric oxygen sensing catheters are modified with outer NO-releasing silicone rubber coatings prior to intravascular implantation in animal models. Optical sensors have received increased attention as intravascular sensing devices because of certain inherent advantages, including the ease of miniaturization (resulting in the ability to prepare multiple optical fiber-based analyte sensors), no requirement for a reference electrode, and the potential for greater signal stability by employing multiple wavelength measurements.13-15 A number of commercial products based on optical oxygen, carbon dioxide, and pH sensors were introduced in the mid-1990s,16,17 but almost all have been removed from the market because of lingering analytical performance issues resulting from blood compatibility issues. (5) Mellion, B. T.; Ignarro, L. J.; Ohlstein, E. H.; Pontecorvo, E. G.; Hyman, A. L.; Kadowitz, P. H. Blood 1981, 57, 946-955. (6) Radomski, M. W.; Salas, E. Artherosclerosis 1995, 118 (Suppl.), S69-S80. (7) Vaughn, M. W.; Kuo, L.; Liao, J. C. Am. J. Physiol. (Heart Circ. Physiol.), 1998, 274, H2163-H2176. (8) Ware, J. A.; Heistad, D. D. New Eng. J. Med. 1993, 328, 628-635. (9) Mowery, K. A.; Schoenfisch, M. H.; Baliga, N.; Wahr, J. A.; Meyerhoff, M. E. Electroanalysis 1999, 11, 681-686. (10) Espadas-Torre, C.; Oklejas, V.; Mowery, K.; Meyerhoff, M. E. J. Am. Chem. Soc. 1997, 119, 2321-2322. (11) Schoenfisch, M. H.; Mowery, K. A.; Radar, M. V.; Baliga, N.; Wahr, J. A.; Meyerhoff, M. E. Anal. Chem. 2000, 72, 1119-1126. (12) Frost, M. C.; Rudich, S. M.; Zhang, H.; Maraschio, M. A.; Meyerhoff, M. E. Anal. Chem. 2002, 074, 5942-5947. (13) Meyerhoff, M. E. Trends Anal. Chem. 1993, 12, 257-266. (14) Wahr, J. A.; Tremper, K. K. J. Card. Vasc. Anesth. 1994, 8, 342-353. (15) Ferguson, J. A.; Healey, B. G.; Bronk, K. S.; Barnard, S. M.; Walt, D. R. Anal. Chim. Acta 1997, 340, 123-131. (16) Mahutte, C. K.; Sassoon, C. S. H.; Muro, J. R.; Hansmann, D. R.; Maxwell, T. P.; Miller, W. W.; Yafuso, M. J. Clin. Monit. 1990, 6, 147-157. (17) Uchida, T.; Makita, K.; Tsunoda, Y.; Toyooka, H.; Amaha, K. Can. J. Anaesth. 1994, 41, 64-70.
Analytical Chemistry, Vol. 74, No. 23, December 1, 2002 5937
The chemistry used to prepare optical sensors for blood gas measurements varies considerably, depending on the analyte of interest. Optical measurement of the partial pressure of oxygen (PO2) is commonly based on the fluorescence quenching of an indicator (i.e., fluorophore) doped in a polymer matrix.18,19 Typically, silicone rubber (SR) is the polymeric material of choice because of its high permeability to oxygen,20 ease of handling, optical transparency, and exemplary adhesion to glass fibers. The most commonly used indicators include polycyclic aromatic hydrocarbons (PAHs) (e.g., fluoranthene, pyrene, etc.)21-24 and transition metal-organic complexes (e.g., ruthenium 4,7-diphenyl1,10-phenanthroline, ruthenium(II)tris(dipyridine)).25-28 The PO2 in solution is determined by measuring changes in the intensity of the fluorescence signal (i.e., as O2 level increases, observed fluorescence intensity decreases). The theoretical relationship between fluorescence intensity and PO2 is given by the SternVolmer equation:19,29
I0/I ) 1 + KSVPO2
(1)
where I0 is the fluorescence intensity in the absence of oxygen, I is the intensity of the signal at a given PO2 level (in Torr), and KSV (in Torr-1) is the Stern-Volmer quenching constant for the particular indicator. The goal of this research effort was to explore the feasibility of fabricating fluorescent oxygen sensors using novel NO-releasing SRs, and thereby assess whether the chemistry of NO release is truly compatible with the chemistry required for fluorescence oxygen sensing. PAH-based optical oxygen sensors were initially chosen for study over those based on transition-metal complexes, since it was believed that NO would be less reactive toward PAHs. Furthermore, a two-fluorophore system consisting of pyrene as a donor molecule and perylene as an energy acceptor molecule were chosen for this work because of their high sensitivity toward oxygen and large Stokes’ shift which facilitate the separation of analytical signal from the excitation radiation.30 Prototype optical sensors were prepared on quartz slides using either a small molecule diazeniumdiolate NO donor, (Z)-1-[N-methyl-N-[6-(Nmethylammoniohexyl)amino]]-diazen-1-ium-1,2-diolate (MAHMA/ N2O2) doped within SR1 or a newly developed modified silicone (18) Wise, D. L.; Wingard, L. B., Jr., Eds.; Biosensors with Fiberoptics; Humana: Clifton, NJ, 1991. (19) Wolfbeis, O. S. Ed.; Fiber Optic Chemical Sensors and Biosensors; CRC Press: Boca Raton, FL, 1991; Vols. I and II. (20) Brandrup, J., Immergut, E. H., Eds.; The Polymer Handbook, 3rd ed.; Wiley: New York, 1989; pp 435-449. (21) Wolfbeis, O. S.; Offenbacher, H.; Kroneis, H.; Marsoner, H. Mikrochim. Acta 1984, 1, 153-158. (22) Cox, M. E.; Dunn, B. Appl. Opt. 1985, 24, 2114-2120. (23) Gehrich, J. L.; Lubbers, D. W.; Opitz, N.; Hansmann, D. R.; Miller, W. W.; Tusa, J. K.; Yafuso, M. IEEE Trans. Biomed. Eng. 1986, 33, 117-132. (24) Sharma, A.; Wolfbeis, O. S. Appl. Spectrosc. 1988, 42, 1009-1011. (25) Wolfbeis, O. S.; Leiner, M. J. P.; Posch, H. E. Mikrochim. Acta 1986, 3, 359-366. (26) Bacon, J. R.; Demas, J. N. Anal. Chem. 1987, 59, 2780-2785. (27) Carraway, E. R.; Demas, J. N.; DeGraff, B. A.; Bacon, J. R. Anal. Chem. 1991, 63, 337-342. (28) MacGraith, B. D.; McDonagh, C. M.; O’Keefe, G.; Keyes, E. T.; Vos, J. G.; O’Kelly, B.; McGilp, J. F. Analyst 1993, 118, 385-388. (29) Murov, S. L. In Handbook of Photochemistry; Hug, G. L., Ed.; Marcel Dekker: New York, 1993. (30) Sharma, A.; Wolfbeis, O. S. Anal. Chim. Acta 1988, 212, 261-265.
5938
Analytical Chemistry, Vol. 74, No. 23, December 1, 2002
Figure 1. Configuration of (a) single-layer and (b) dual-layer fluorescent oxygen-sensing films.
rubber polymer that possesses anchored diazeniumdiolate functional groups on the cross-links of the polymer matrix (DACA6/N2O2-SR).2 Both of these NO-releasing materials have been previously shown to improve the in vivo thromboresistivity and performance of electrochemical oxygen sensors. Herein, the in vitro analytical performance of the corresponding fluorescencebased O2-sensing SR films that continuously release NO is reported. EXPERIMENTAL SECTION Materials and Reagents. Silanol-terminated poly(dimethylsiloxane) (PDMS), methyltrimethoxysilane, dibutyltin dilaurate, pyrene, perylene, dioctyl sebacate (DOS), and phosphate-buffered saline (PBS, pH 7.4, 0.01 M phosphate, and 0.14 M saline) were obtained from Sigma-Aldrich (Milwaukee, WI). N-(6-Aminohexyl)3-aminopropyl-trimethoxysilane (DACA-6) and hexamethyldisilazane-treated fumed silica were products of Gelest (Tullytown, PA). Nitric oxide (NO) and argon (Ar) gases were obtained from Cryogenic Gases (Detroit, MI). Toluene (certified A.C.S. grade) was purchased from Fisher Scientific (Pittsburgh, PA). MAHMA/ N2O2 was a gift from the Laboratory of Comparative Carcinogenesis at the National Cancer Institute (Frederick, MD). All chemicals were used as received. Preparation of Silicone Rubber (SR) Sensing Films. Control SR sensing films were prepared by mixing PDMS (200 mg), methyltrimethoxysilane (6.5 mg), dibutyltin dilaurate (1 mg), fumed silica (40 mg), and DOS (25 mg) in a 2.0-mL toluene solution consisting of 1.0 mL of 7 mM pyrene (toluene) and 1.0 mL of 4.5 mM perylene (toluene). Five drops of the cocktail solution were cast onto a quartz slide (width 1.2 cm, length 3.8 cm) as illustrated in Figure 1a. Cross-linked SR films were then formed on quartz slides by curing under ambient moisture in air for 48 h. The final concentrations of pyrene and perylene in the film were ∼35 and 22 mM, respectively. The thickness of sensing films was estimated to be ∼80 µm, given that the density of SR films is 1.2 g/cm2. The MAHMA/N2O2-doped SR sensing films were prepared by adding 10 mg of MAHMA/N2O2 powder to the above SR cocktail prior to casting films. Preparation of DACA-SR Sensing Films. DACA-6-SR sensing films (see Figure 1a) were prepared by mixing PDMS (200 mg), DACA-6 (23.6 mg), dibutyltin dilaurate (0.2 mg), and fumed silica (40 mg) in a 2.0-mL toluene solution consisting of 1.0 mL of 7 mM pyrene (toluene) and 1.0 mL of 4.5 mM perylene (toluene). Five drops of the cocktail solution were cast onto the quartz slide. Cross-linked DACA-6-SR films were then formed on quartz slides by curing under ambient moisture for 48 h.
The dual-layer DACA-6-SR sensing films (see Figure 1b) were prepared using the above DACA-6-SR cocktail but excluding pyrene and perylene from the innermost layer (layer was ∼500 µm thick). The DACA-6-SR layer (without indicators) was then covered with a SR layer (with indicators) by adding 5 drops of the SR cocktail on top of the DACA-6-SR layer (sensing layer was ∼80 µm thick). The NO-releasing DACA-6/N2O2-SR-based oxygen-sensing films (both single-layer and dual-layer) were prepared by directly reacting the above DACA-6-SR films with NO at 80 psi in a high-pressure NO-reactor system for ∼55 h. NO-Release Measurements. Nitric oxide release from cured sensing films at room temperature was monitored via chemiluminescence with a Sievers Nitric Oxide Analyzer 280 (Boulder, CO). Fluorescence Measurements. A Shimadzu model RF-1501 spectrofluorometer equipped with a special quartz cuvette for mounting quartz slides was used to evaluate the optical sensing films. Ultraviolet excitation (320 nm) from a 150 W xenon lamp was focused onto the sensing film at an optimized angle of 35° to minimize direct reflection of the excitation beam into the detector. In addition, a 450-nm cutoff filter was placed between the quartz sample and the detector to further reduce stray light. Fluorescence emission was monitored at 470 nm as a function of PO2. Oxygen levels were systematically varied in the cuvette by removing the PBS solution with a glass pipet and adding a tonometered solution with a different PO2. Care was taken to minimize the formation of air bubbles on the membrane layer and to allow adequate equilibration time (8 min) for oxygen diffusion into/out of the film prior to acquiring fluorescence spectra. RESULTS AND DISCUSSION Optical sensors made on quartz slides were used to study whether oxygen sensors can be prepared with NO-releasing SRs and if their analytical performance is equivalent to sensing films that do not generate NO (i.e., controls). The analytical performance of these films was evaluated by varying PO2 in the cuvette and measuring the corresponding fluorescence intensity at the maximum emission wavelength (λemission ) 470 nm) for the donoracceptor complex. Typical fluorescence spectra obtained from a control oxygensensing film as a function of PO2 showed the expected intensity decrease with increasing PO2, resulting in a linear Stern-Volmer plot, shown in Figure 2. Optical sensors prepared with either MAHMA/N2O2-doped SR or DACA-6/N2O2-SR films were also evaluated. Sensors Prepared with MAHMA/N2O2-Doped SR Films. NO-releasing optical sensors were prepared by doping MAHMA/ N2O2 particles within SR sensing films. The insoluble MAHMA/ N2O2 particles with varied sizes were heterogeneously dispersed into the SR matrix. The analytical performance of the resulting NO-releasing fluorescent sensors was strongly influenced by the dispersion of the MAHMA/N2O2 particles, as indicated by their lower sensitivity relative to controls (KSV ) 0.00141 ( 0.00018 versus 0.00598 ( 0.00002, respectively) and a nonlinear response (Figure 2). The observed decrease in sensitivity and variations in quenching behavior are attributed to poor optical film quality owing to the MAHMA/N2O2 particles and inherent polymer heterogeneity, respectively. Nonlinear oxygen quenching and the existence of
Figure 2. Stern-Volmer plots for the quenching of the pyrene/ perylene system in single-layer SR (9) and MAHMA/N2O2-doped SR (b) films by oxygen. Each value is an average of four measurements.
several decay profiles as a function of film heterogeneity have been previously reported for SR-based optical sensors.26-28,31-33 For example, Xu et al.32 observed reduced quenching as the amount of polar cross-linker and, thus, the size and number of polar domains within the SR matrix were increased. As a result, the tendency of pyrene to dissolve in or associate with more polar components leads to inefficient quenching. Klimant and Wolfbeis33 also reported the nonlinear quenching behavior of ruthenium diimine complexes in SR films. They attributed the nonlinear quenching to differences in indicator environments, specifically a homogeneous distribution and an aggregated state within the polymer matrix. Clearly, multiple interfacial interactions between the indicator and the polymer support result in a nonlinear response to oxygen. Our findings indicate that different quenching mechanisms are operative when pyrene/perylene is combined with heterogeneously dispersed MAHMA/N2O2 particles. Optical Sensors Prepared with DACA-6-SR Films (SingleLayer). To eliminate film heterogeneity caused by the incorporation of insoluble MAHMA/N2O2 particles, the use of DACA-6/ N2O2-SR films was investigated. This polymer was recently prepared and characterized by Zhang et al.2 and shown to yield significantly enhanced thromboresistivity in two different animal models.2,12 The polymer consists of conventional PDMS that is cross-linked with a diamine-alkyl silane cross-linker. The secondary amine sites within the cross-linker can be converted to NO donors in situ after curing. The analytical performance of the DACA-6-SR-based sensors (without NO-release) was first examined. A nonlinear response to oxygen was observed for DACA6-SR films doped with the perylene/pyrene pair (see Figure 3). Notably, enhanced quenching was observed with the DACA-6SR films relative to the perylene/pyrene-doped SR controls. This behavior is likely the result of additional quenching of pyrene/ perylene by amines. Goodpaster et al.34 recently reported that amines effectively quench the fluorescence of various PAHs (including pyrene). Clearly, a multiple quenching system that does not respond linearly to oxygen is nonideal for preparing functional optical oxygen sensors. (31) Demas, J. N.; DeGraff, B. A. Makromol. Chem. 1992, 59, 35-51. (32) Xu, W.; Schmidt, R.; Whaley, M.; Demas, J. N.; DeGraff, B. A.; Karikari, E. K.; Farmer, B. L. Anal. Chem. 1995, 67, 3172-3180. (33) Klimant, I.; Wolfbeis, O. S. Anal. Chem. 1995, 67, 3160-3166. (34) Goodpaster, J. V.; McGuffin, V. L. Anal. Chem. 2000, 72, 1072-1077.
Analytical Chemistry, Vol. 74, No. 23, December 1, 2002
5939
Figure 3. Stern-Volmer plots for the quenching of the pyrene/ perylene system in single-layer SR (9) and DACA-6-SR (b) films by oxygen. Each value is an average of four measurements.
Optical Sensors Prepared with DACA-6 /N2O2-SR Films (Dual-Layer). To circumvent the problem of enhanced quenching by amine groups within the DACA-6-SR matrix, a dual-layer configuration was employed to fabricate oxygen sensors. As illustrated in Figure 1b, a DACA-6-SR layer without indicator is covered with a SR layer doped with indicator. The top SR layer serves as the oxygen-sensing layer, while the bottom DACA-6SR layer, once reacted with NO to form DACA-6/N2O2-SR, serves as the NO-releasing layer. It is also possible to reverse the order of the layers, although all subsequent data presented is for the case where the outer layer is sensitive to oxygen. The analytical performance of DACA-6-SR-based sensors was found to be comparable to controls (KSV ) 0.00642 ( 0.00021 versus 0.00598 ( 0.00002, respectively with R2 values of 0.99). The dual-layer configuration effectively eliminates interference from the amine groups within the DACA-6-SR matrix. This duallayer configuration would probably not be useful for the MAHMA/ N2O2-doped sensing films because, unlike DACA-6-SR, the amine groups of which are covalently attached to the polymer, the parent diamine that is present after NO-release could potentially diffuse into the sensing layer. This would result in enhanced quenching of the fluorophores, leading to the same nonlinear behavior observed with the DACA-6-SR single-layer film. To examine the effect of localized NO generation on the analytical performance of the sensors, NO-releasing fluorescent sensors (i.e., DACA-6/N2O2-SR-based sensors) were prepared by in situ formation of diazeniumdiolates in the DACA-SR matrix (see Experimental Section). At ambient temperature, the NO flux reached 6.5 × 10-10 mole/cm2‚min initially, as shown in Figure 4. It then decreased and remained at a level of 4 × 10-10 mole/ cm2/min for at least 24 h. This flux is greater than that observed from healthy unstimulated endothelial cells.7 The quenching constant for the NO-releasing sensing film, KSV, was determined to be 0.00589 ( 0.00034 during the initial calibration after 2 h of NO-release, 0.00668 ( 0.00053 after 12 h, and 0.00658 ( 0.00017 after 22 h (see Figure 4 inset; all R2 values were 0.99). These values are again nearly identical to those obtained for the duallayer DACA-6-SR and control sensors and indicate that continuous release of NO for at least 24 h does not compromise the sensitivity or linearity of the sensing films. The dynamic response time of the sensors was affected only by the thickness of the SR layers and not by the presence of NO-release (data not shown). 5940 Analytical Chemistry, Vol. 74, No. 23, December 1, 2002
Figure 4. Plot of NO flux from a dual-layer DACA-6/N2O2-SR sensing film at ambient temperature for 24 h. Inset shows the SternVolmer plots for the quenching of the pyrene/perylene system at 2 (9), 12 (b) and 22 h (2). The dashed line shows the level of NO flux from unstimulated endothelial cells.
An obvious concern with the dual-layer sensor configuration is the stability of the sensor. As described above, the sensing films exhibit nonlinear behavior when the pyrene/perylene pair is incorporated directly into the DACA-6-SR layer. It is possible that the fluorophores may, over time, diffuse from the SR sensing layer into the DACA-6-SR layer and negatively influence sensor response to oxygen. To simulate the effect of this type of diffusion, a sensing film was prepared by incorporating pyrene/perylene into both the DACA-6-SR and SR layers. The KSV for this sensing film was 0.00612 ( 0.00010 and exhibited linear behavior (R2)0.99), indicating that even if the pyrene/perylene pair diffuses into the underlying DACA-6-SR layer, the sensitivity and linearity of the resulting film are not compromised, since the majority of the analytical signal originates from the layer that does not contain amines. CONCLUSIONS The in vitro analytical performance of pyrene/perylene based oxygen sensors (on quartz slides) prepared with two different NOreleasing SR films has been evaluated. Sensors doped with MAHMA/N2O2 particles were found to exhibit decreased sensitivity and a nonlinear optical response, as compared to control sensors. These properties are attributed to the film heterogeneity caused by the incorporation of insoluble MAHMA/N2O2 particles into the films. To circumvent the problem of film heterogeneity, DACA-6-SR sensing films (single-layer) were prepared and examined. Amine groups (necessary for NO release) within the film were found to quench the fluorescence of pyrene/perylene, resulting in a nonlinear response to oxygen. To reduce amine interference, the indicator and NO release layers were sequentially cast, separating the components in highly cross-linked layers. Identical optical responses between the DACA-6-SR dual-layer sensors and the blank sensors were obtained using this approach. These dual-layer films were further reacted with NO to prepare NO-releasing sensors. It was found that the low level of NO release
from the sensor did not influence the O2 response characteristics (i.e., sensitivity and linearity). Hence, the dual-layer DACA-6/ N2O2-SR films represent good candidates for preparing miniaturized NO-releasing fluorescence-based oxygen sensors at the distal end of optical fibers for fabrication of more thromboresistive intravasular optical oxygen sensors.
ACKNOWLEDGMENT The authors are grateful for the support of this research by the National Institutes of Health (NIH EB00783-05). Received for review August 24, 2002. Accepted October 1, 2002. AC026075Y
Analytical Chemistry, Vol. 74, No. 23, December 1, 2002
5941
