Planar nitric oxide (NO)-selective ultramicroelectrode sensor for measuring localized NO surface concentrations at xerogel microarrays

Abstract

A planar ultramicroelectrode nitric oxide (NO) sensor was fabricated to measure the local NO surface concentrations from NO-releasing microarrays of varying geometries. The sensor consisted of platinized Pt (25 µm) working electrode and a silver paint reference electrode coated with a thin silicone rubber gas permeable membrane. An internal hydrogel layer separated the Pt working electrode and gas permeable membrane. The total diameter of the sensor was ≤50 µm, and demonstrated negligible analyte trapping effects. The sensitivity and response time of the ultramicroelectrode sensor to NO were 0.19 ± 0.07 pA nM⁻¹ and 1–4 s, respectively, with a 5 nM limit of detection. The sensor was employed to correlate the steady-state NO surface concentration and observed platelet adhesion resistance. Results indicate that the required steady-state NO concentration necessary to inhibit platelet adhesion to the micropatterned xerogels depends on the xerogel geometry.

---

Introduction

Although tremendous efforts have focused on the development of more biocompatible polymers, the utility of such materials for preparing in vivo devices such as sensors remains limited.^1,2 Thrombus formation and/or implant-associated infection continue to hinder the utility of polymeric materials for in vivo applications. Polymers designed to slowly release low levels of nitric oxide (NO) have emerged as a new class of biomaterials.^3–6 Indeed, the discovery that endogenously produced NO regulates a number of physiological processes including the regulation of platelet adhesion and activation,⁷ angiogenesis,⁸ phagocytosis,⁹ and wound healing,¹⁰ has helped researchers harness its potential for therapeutic applications. For NO-releasing polymers in particular, NO's therapeutic effects are localized at the implant surface due to its extremely short half-life (i.e., <1 s in oxygenated biological milieu),¹¹ and thus the potential problems related to other common systemic anticoagulant and antibiotic therapies are avoided.

We have recently reported the synthesis of xerogel materials capable of releasing therapeutic levels of NO for extended periods.^5,12,13 Sol–gel chemistry allows for the synthesis of materials with a broad range of NO release properties by simply altering the identity and/or relative ratios of the NO donor precursor (aminosilane) in the starting sol. We have also reported the use of micropatterning methods to selectively modify substrates with precisely positioned NO-releasing xerogel microstructures such that regions of the underlying surface remain unmodified (i.e., between the xerogel features).^13–15 Methyltrimethoxysilane (MTMOS) xerogels doped with (aminoethylaminomethyl) phenethyltrimethoxysilane (AEMP3) were initially used in these studies due to the ability to readily form micropatterns using standard soft lithography methods.^13–15 As studied via chemiluminescence detection, the NO release properties of the AEMP3/MTMOS micropatterned xerogels were governed by a variety of factors including xerogel precursor structures and concentrations, microarray dimensions, and xerogel solution immersion time.^13–15In vitro platelet adhesion experiments were used to assess the blood compatibility of substrates modified with such NO-releasing xerogel micropatterns. The results of these studies indicated that at a NO surface flux of 2.2 pmol cm⁻² s⁻¹, arrays of xerogel lines separated by up to 50 µm were equally as thromboresistant as uniform xerogel coatings (i.e., NO-releasing films).^13–15 When the microstructure separation was reduced to 10 µm, a NO surface flux of only 0.42 pmol cm⁻² s⁻¹ effectively prevented platelet adhesion.¹⁵ Similarly, Dobmeier and Schoenfisch¹⁶ reported a 50% reduction in bacterial adhesion (Pseudomonas aeruginosa) to NO-releasing (1.0 pmol cm⁻² s⁻¹) arrays for microstructure separations up to 200 µm.¹⁶ Collectively, these findings indicate that substrates modified with NO-releasing xerogel arrays may be useful for improving the biocompatibility of an interface while preserving the properties (e.g., functionality) of the underlying surface. Nitric oxide-releasing xerogel microarrays have proven particularly useful as sensor coatings as such interfaces allow for adequate analyte mass transport to the sensor surface while maintaining overall thromboresistivity. Indeed, a miniaturized enzymatic glucose biosensor modified with a NO-releasing xerogel microarray was characterized by significantly greater sensitivity compared to a sensor modified with a xerogel film due to enhanced mass transport of analyte to the sensor surface through unmodified regions between xerogel lines.^14,17

To date, NO-release properties of xerogel-based films and microarrays have been determined using chemiluminescence.^13–15 Such instrumentation allows for the measurement of NO in bulk solution in real time. A surface flux can be determined by approximating the geometry of the NO-releasing film or microarray. Although surface flux is a useful parameter for comparing the NO release from different polymeric films and/or microarrays, a steady-state NO surface concentration (i.e., at the interface) may provide an improved understanding on the minimum NO release necessary to influence thromboresistivity. Developing methods for determining the localized NO surface concentration is important for tuning NO-releasing interfaces for specific applications such as designing more blood-compatible implantable sensors.

The advantages of using an NO-selective electrochemical sensor to measure NO include the ability to accurately reflect the spatial and temporal distribution of NO in biological environments. Electrochemical sensors provide direct, real-time measurement of NO, and are also relatively inexpensive. A miniaturized planar NO sensor (∼150 µm sensing tip size) was recently reported for making NO surface concentration measurements.¹⁸ Notably, the placement of the sensor adjacent to the NO-releasing substrate led to the accumulation of NO in the region between the sensor and substrate, which the authors referred to as “analyte trapping effects”, resulting in the measurement of artificially elevated NO concentration levels.¹⁸ Overcoming this problem requires further miniaturization of the sensor. Herein, we report the development of an ultramicroelectrode (sensing tip size ≤50 µm) NO sensor for accurately measuring steady-state surface NO concentration levels above NO-releasing xerogel microarrays. The electrochemical performance of the ultramicroelectrode sensor, including sensitivity, selectivity, response time, and stability, are reported. Furthermore, surface concentration measurements for a range of NO-releasing xerogel microarrays are compared with in vitro platelet adhesion results. We attempt to identify the surface NO concentrations necessary to inhibit platelet adhesion.

Experimental

Materials

(Heptadecafluoro-1,1,2,2-tetra-hydrodecyl)trichlorosilane, methyl-trimethoxysilane (MTMOS) and (aminoethylaminomethyl)phenethyltrimethoxysilane (AEMP3) were purchased from Gelest (Tullytown, PA). Silicon master templates, prepared using standard photolithography methods, were obtained from MCNC (Research Triangle Park, NC). Sylgard 184 (poly(dimethylsiloxane), PDMS) was purchased from Fisher Scientific (Philadelphia, PA). Whole blood from healthy pigs was received from the Francis Owen Blood Research Laboratory (UNC-CH; Chapel Hill, NC). Platinum (Pt) wire (25 µm outer diameter, od) and silver (Ag) wire (76 µm od) were purchased from Alfa Aesar (Ward Hill, MA) and Aldrich (St. Louis, MO), respectively. Borosilicate glass capillaries (1 mm od) and room temperature-vulcanizing silicone rubber (3140 RTV) were purchased from World Precision Instruments Inc. (Sarasota, FL) and Dow Corning (Midland, MA), respectively. Nitric oxide (NO), argon (Ar), and nitrogen (N₂) were purchased from National Welders Supply (Raleigh, NC). Distilled water was purified to 18.2 MΩ cm⁻¹ with a Millipore Milli-Q Gradient A-10 system (Bedford, MA). All other solvents and chemicals were analytical-reagent grade and used as received.

Synthesis of nitric oxide-releasing xerogels

Sol solution was prepared by combining 200 µL absolute ethanol (EtOH), 11 µL water, 120 µL MTMOS, 80 µL AEMP3, and sonicating for 5 min at room temperature. Glass substrates (25 mm × 8 mm) were sonicated in EtOH for 20 min, rinsed with EtOH, dried under a stream of N₂, and ozone-cleaned for 20 min in a BioForce TipCleaner (Ames, IA). Xerogel films (thickness ∼ 10 µm) were prepared by casting 10 µl of the sol solution onto the surface of the clean glass substrate. The xerogel-modified surfaces were cured for 24 h under ambient conditions. Diamine groups in the cured xerogel were then converted to diazeniumdiolate NO donors in an in-house reactor.¹⁹ The reaction chamber was flushed thoroughly with Ar to remove air and moisture and pressurized with 5 atm NO. After 3 d, the reaction chamber was flushed thoroughly with Ar again to remove unreacted NO. The diazeniumdiolate-modified xerogels were removed from the reactor and stored in a sealed container at −20 °C until use.

Synthesis of nitric oxide-releasing xerogel micropatterns

Silicon (Si) wafers (1 cm²) etched with arrays of straight lines (length = 8 mm, width = 50 µm, depth = 10 µm, and individual line separations ranging from 10–200 µm) were cleaned of residual organic contaminants by immersion in Piranha solution consisting of a 3 ∶ 7 mixture (v/v) of H₂O₂ (30%): sulfuric acid (H₂SO₄, 18 M). (Care should be taken when using Piranha solution since it is a very strong oxidant and can spontaneously detonate upon contact with organic material.) The wafers were then rinsed with water and EtOH, and dried under a stream of N₂. To prevent bonding between PDMS and the Si surface, the wafers were incubated in (heptadecafluoro-1,1,2,2-tetra-hydrodecyl)-trichlorosilane vapor for 1 h under a N₂ environment.²⁰ The PDMS precursor and cross-linking solutions were combined in a 10 ∶ 1 ratio, mixed thoroughly, and poured over the Si wafers to yield a film with thickness ∼5 mm. The mixture was degassed under vacuum for 30 min, and cured at 70 °C for 1 h. The elastomeric templates were then detached from the Si wafers, rinsed with EtOH, and dried in a stream of N₂.

Xerogel arrays were prepared according to a previously described method for forming micropatterns.^13,14 Briefly, a micromolded template was placed in conformal contact with a clean glass substrate previously embedded in a PDMS support. Sol solution deposited on the PDMS support was then drawn into the microchannels by capillary action.^13,14 After curing for 24 h under ambient conditions, the template was removed to reveal the xerogel microarray. Excess xerogel was separated from the micropattern by removing the glass substrate from the elastomeric support. Diamine functional groups in the cured xerogel were then converted to diazeniumdiolate NO donors in an in-house reactor as described above.¹⁹

The rate and duration of NO release from diazeniumdiolate-modified xerogels were determined using a Sievers NOA™ 280i Chemiluminescence Nitric Oxide Analyzer (Boulder, CO). Xerogels were placed in a reaction cell containing 15 mL of Tyrode's buffer (137 mM sodium chloride, 5.6 mM D-(+)-glucose, 3.3 mM potassium phosphate monobasic, 2.7 mM potassium chloride; pH 7.4) at 37 °C. Nitric oxide released from the xerogel was transported to the analyzer by a stream of N₂ passed through the reaction cell. The level of NO (ppb) was measured in real-time and used to calculate the NO surface flux (pmol cm⁻² s⁻¹) using the dimensions of the xerogel film or pattern.

Blood compatibility evaluation

Nitric oxide-releasing xerogel microarrays were prepared on glass substrates as described above. In vitro platelet adhesion to NO releasing micropatterns was evaluated after first immersing the microarrays in Tyrode's buffer at physiological temperature and pH for 1–24 h prior to incubation in platelet rich plasma. Micropattern biocompatibility was evaluated relative to bare glass slides, and control xerogel micropatterns (i.e., not charged with NO).

Platelet rich plasma (PRP) was obtained from acid citrate dextrose (ACD)-anticoagulated porcine blood (1 part ACD to 9 parts whole blood) by centrifugation at 200 g for 30 min at room temperature.²¹ To ensure normal platelet activity, the concentration of calcium in the PRP was adjusted to 0.25–0.50 mM using calcium chloride (CaCl₂).²¹ Substrates were then incubated with PRP for 1 h at 37 °C, rinsed thoroughly with Tyrode's buffer (pH 7.4) to remove loosely adhered platelets, and immersed in 1% glutaraldehyde solution (v/v, Tyrode's buffer) for 30 min to preserve cell morphology. The surfaces were then rinsed with both Tyrode's buffer and water, and dehydrated by sequential immersion for 5 min in 50%, 75% and 95% ethanol (v/v, water), followed by 100% ethanol for 10 min and hexamethyldisilazane (overnight). Phase contrast optical micrographs were obtained using a Zeiss Axiovert 200 inverted microscope (Chester, VA). Images from 10 distinct regions (200 × 200 µm) on at least 3 different samples were digitally processed (% opaqueness) to obtain semi-quantitative platelet surface coverage values. Analysis of variance (ANOVA) statistical analysis was performed to determine the significance of differences between platelet adhesion to control and NO-releasing xerogels, with p ≤ 0.001 indicating a significant difference between data sets.

Fabrication of the planar ultramicroelectrode nitric oxide sensor

A schematic of the miniaturized planar amperometric NO sensor, prepared according to previously described methods,^18,22 is shown in Fig. 1. Glass-sealed Pt (25 µm od) and coiled Ag (76 µm od) wires served as the working and reference electrodes, respectively. The reference electrode was immobilized in close proximity to the working electrode using silver paint. The diameter of the sensing tip was minimized by mechanically polishing the insulating glass sheath to achieve a conical shape with a final tip diameter of ∼50 µm. The bare Pt disk electrode was then platinized in 3% chloroplatinic acid (v/v; water) solution by cycling the potential from +0.6 V to −0.35 V at a scan rate 20 mV s⁻¹ using a CHI900 scanning electrochemical microscope (CH Instruments; Austin, TX).¹⁸ The sensing tip was then coated with an internal hydrogel layer containing 1 wt% Methocel, 30 mM sodium chloride (NaCl), and 0.3 mM HCl, pH 3.5, serving to ensure electrical contact between the platinized Pt working and silver reference electrodes. The internal hydrogel layer was allowed to dry for 1 h before further modifying the sensor with an outer silicone rubber gas permeable membrane formed by dip coating the sensor in a 1% (w/v) solution of RTV-3140 in tetrahydrofuran (THF) and drying overnight under ambient conditions.

Fig. 1 A diagram of the planar amperometric NO-selective ultramicroelectrode sensor.

Nitric oxide surface concentration measurements

A CHI900 scanning electrochemical microscope (CH Instruments; Austin, TX) was used to apply a potential of +0.75 V (vs. Ag/AgCl) to the working electrode. To ensure sensor stability, calibrations were performed before and after NO surface measurements using standard NO solutions. A NO stock solution was prepared daily by bubbling deoxygenated PBS with Ar for 20 min followed by NO for 20 min to achieve a final NO concentration of 1.9 mM.²³ Surface concentration measurements were obtained by positioning the sensing probe directly above the NO-releasing xerogel film or microarray using a computer-controlled inchworm motor (CH Instruments, Austin, TX). Current was measured for 10 min to achieve a steady-state level, and then converted to NO concentration using the calibration curve.

Results and discussion

Microarrays consisting of 40% AEMP3 (balance MTMOS) xerogel lines (50 µm wide, 10 µm high) separated by 10–200 µm of exposed glass were employed to systematically evaluate the effects of microstructure separation and NO surface concentration on platelet adhesion to patterned surfaces. An ultramicroelectrode sensor capable of monitoring the localized NO concentration in close proximity (i.e., 10 µm) to the xerogel microstructure source was fabricated, characterized, and used to investigate the effects of NO surface concentration on platelet adhesion.

Evaluation of micropattern blood compatibility

An established procedure using platelet-rich porcine plasma was employed to assess in vitro platelet adhesion to xerogel micropatterns since such studies are useful in predicting thromboresistivity in humans.²⁴ Representative phase contrast optical micrographs of platelet adhesion to 40% AEMP3 (balance MTMOS) arrays after different pre-incubation (buffer immersion prior to platelet incubation) times are shown in Fig. 2. As anticipated based on previous studies,^13,14,17 considerable platelet adhesion and aggregation were observed for all control surfaces regardless of microstructure separation. Previous studies have also demonstrated that at extended soak periods, the NO surface flux, as determined by chemiluminescence, subsides with time as the finite NO donor reservoir within the polymer becomes depleted.^13–15 Thus, the blood compatibility of xerogel microarrays was evaluated at NO surface fluxes of 6.7 ± 0.3, 1.7 ± 0.4, and 0.42 ± 0.03 pmol cm⁻² s⁻¹, corresponding to solution pre-incubation times of 1, 6, and 24 h, respectively. As shown in Fig. 2B, minimal platelet adhesion was observed after 1 h pre-incubation for microstructure separations up to 50 µm. These results are consistent with previous in vitro and in vivo reports in which polymers doped with diazeniumdiolate NO donors exhibited enhanced thromboresistivity due to NO release.⁴ After 6 h pre-incubation, a slight increase in platelet adhesion was observed, presumably due to partial depletion of the NO donor reservoir within the xerogel microstructures.^13–15 After 24 h, a significant increase in platelet adhesion was observed (Fig. 2D), suggesting the concentration of NO generated from the depleted xerogel was no longer sufficient to impart blood compatibility to the substrate. Semi-quantitative platelet surface coverage values for micropatterns with line separations ranging from 10–200 µm are provided in Table 1. Nitric oxide-releasing arrays with 10 µm separations between xerogel lines were effectively resistant to platelet adhesion (relative to control micropatterns) for the duration of the experiments as evidenced by ANOVA calculations (p < 0.001). Increasing the microstructure separation to 25 µm resulted in substrates that were effectively thromboresistant for 6 h, however a slight increase in platelet adhesion was observed at 24 h. Despite this increase, ANOVA statistics confirmed that after 24 h the NO-releasing arrays were significantly more blood compatible than control arrays (p = 3.8 × 10⁻⁶). Substrates prepared with 50 µm xerogel line separations were characterized by good blood compatibility initially, however increased platelet adhesion was observed after only 6 h. Despite a slight compromise in blood compatibility relative to surfaces at t = 1 h, ANOVA statistical analysis confirmed that after 6 h solution pre-incubation, the surfaces were significantly more resistant to platelet adhesion than controls that did not release NO (p = 1.1 × 10⁻¹³). After 24 h pre-incubation, however, the NO-releasing and control arrays were indistinguishable (p = 0.39). Micropatterns with xerogel line separations of 100 µm and 200 µm exhibited poor blood compatibility relative to controls regardless of solution pre-incubation time, suggesting that the concentration of NO generated by these surfaces is below the minimum threshold necessary to prevent platelet adhesion.

Fig. 2 Representative phase contrast optical micrographs (200 µm × 200 µm) of porcine platelet adhesion to (A) control and NO-releasing xerogel microarrays after (B) 1 h, (C) 6 h, and (D) 24 h immersion (pre-incubation) in buffer at physiological temperature and pH prior to incubation in platelet rich plasma. Platelets are shown in white.

Table 1 Effect of microarray dimensions on porcine platelet adhesion^a

Microarray/µm^c	Platelet coverage (%)^b
	Control	1 h	6 h	24 h
a For each measurement, n ≥ 3. Experiments performed at physiological temperature. b Relative to glass. c Microstructure dimensions reported as width∶separation.
50 ∶ 10	117 ± 68	11 ± 9	19 ± 16	24 ± 18
50 ∶ 25	135 ± 51	11 ± 10	12 ± 11	70 ± 43
50 ∶ 50	142 ± 57	9 ± 5	43 ± 28	129 ± 70
50 ∶ 100	165 ± 36	102 ± 69	206 ± 60	259 ± 71
50 ∶ 200	164 ± 52	139 ± 78	255 ± 50	220 ± 49

---

Characterization of the planar nitric oxide-selective ultramicroelectrode

To identify the minimum localized surface concentration of NO necessary to enhance the blood compatibility of a variety of micropatterned substrates, a NO-selective ultramicroelectrode sensor was developed. Obtaining accurate NO concentration measurements requires the use of a probe that can be positioned in close proximity (i.e., 10 µm) to the xerogel surface to ensure a relatively short diffusion path length since NO is rapidly converted to nitrite in oxygenated media.¹⁸

The planar ultramicroelectrode NO sensor designed for these experiments is shown in Fig. 1 and consisted of a platinized Pt working electrode and a Ag paint reference electrode positioned behind a thin internal hydrogel layer and coated with a silicone rubber gas permeable membrane. Previous reports on the use of platinized Pt electrodes demonstrated significant enhancements in sensitivity relative to unmodified Pt due to increased surface roughness.^25–27 In addition, the potential necessary for oxidizing NO to nitrite is reduced from +0.9 V to +0.75 V (vs. Ag/AgCl), thereby increasing sensor stability for longer periods.²⁸ A representative dynamic response curve for the platinized Pt ultramicroelectrode NO sensor is shown in Fig. 3. The sensitivity of the ultramicroelectrode sensor to changes in NO concentration was determined based on the measured current in response to discrete injections of a NO standard solution. The sensor was characterized by a sensitivity of 0.19 ± 0.07 pA/nM (n = 15) and a detection limit of 5 nM. Although the detection limit of this sensor was slightly higher than previously reported poly(tetrafluoroethylene)-based NO microsensors,¹⁸ problems associated with analyte trapping effects due to electrode size were avoided.¹⁸ Furthermore, the thin silicone rubber membrane did not pose a significant diffusion barrier, as evidenced by a rapid sensor response time to changes in NO concentration. Specifically, the sensor response time, determined from the dynamic response curves as the time required to reach 90% of steady-state current when the NO concentration was varied from 10 to 100 nM, was between 1–4 s. Silicone rubber was used as the gas permeable membrane since it has been shown to effectively exclude common electrochemical interferences (i.e., species oxidized at the same potential as NO), including nitrite, ascorbate, and other charged species that could potentially interfere with NO detection in aqueous solution or biological milieu.^29–31 Indeed, the ultramicroelectrode sensor did not respond to nitrite, the primary oxidation product of NO, even at concentrations up to 1 mM.

Fig. 3 Dynamic response of the ultramicroelectrode NO sensor to changes in NO concentration. Measurements were obtained in deoxygenated PBS at physiological pH with constant stirring.

Sensor calibrations were performed before and after surface NO measurements to ensure sensor stability. Representative calibration plots of a sensor before and immediately after NO surface measurements are shown in Fig. 4. Significant changes in sensitivity were not observed, suggesting that the sensor was not adversely affected by the experimental conditions necessary for determining surface concentration values. In addition, the stability of the ultramicroelectrode sensor was maintained after storage in PBS under ambient conditions for 24 h.

Fig. 4 Calibration plots for the ultramicroelectrode NO sensor immediately (■) before and (●) after surface concentration measurements, and (▲) after storing in PBS for 24 h. NO surface measurements were conducted for 3 h in aerated PBS (pH 7.4) without stirring.

Effect of probe–sample separation on analyte detection

Nitric oxide is a highly reactive radical gas that is readily oxidized to nitrite through reaction with oxygen in aqueous solution.³² Therefore, the detection of NO generated from the surface of a xerogel prior to its conversion to nitrite requires that the probe be positioned in close proximity to the source of NO. Previous studies have demonstrated that the measurement of NO at the surface of a synthetic film is significantly affected by the sensing probe geometry.¹⁸ Indeed, the presence of the sensor near the surface of an NO-releasing material has been shown to impair free NO diffusion, resulting in the accumulation of NO between the sample and probe and the measurement of artificially elevated NO levels.¹⁸ Lee et al. reported a significant reduction in such trapping of NO upon reducing the sensing tip diameter from 1 mm to 150 µm at a probe–sample separation of 10 µm.¹⁸ The development of an ultramicroelectrode (≤50 µm od) sensor should allow for more accurate determination of localized NO concentrations by further minimizing trapping effects.

The effect of sample–probe separation on the measured NO concentration above the surface of a xerogel film was evaluated using the ultramicroelectrode sensor to identify if analyte trapping was occurring. The NO-releasing xerogel films employed in these experiments were composed of 40% AEMP3 (balance MTMOS), and the substrates were incubated in PBS (pH 7.4) for 1 h prior to NO surface concentration measurements.¹⁴ Discrete measurements were made for ca. 10 min to ensure a stable steady-state response. In the absence of analyte trapping effects, similar measured NO concentrations at different sensor-sample separations would be expected since the diffusion coefficient and half-life of NO are 3300 µm² s⁻¹³² and 4 s,^33,34 respectively at physiological temperature in aerated aqueous solution. As shown in Fig. 5, the measured NO concentrations at sensor–sample separations from 10–50 µm were similar, suggesting that the sensor geometry did not trap NO diffusing from the xerogel source. Indeed, steady-state NO concentrations of 280 nM ± 10 nM and 270 nM ± 8 nM were measured at probe–sample separations of 10 µm and 100 µm, respectively. For subsequent measurements, the probe was positioned at a distance of 10 µm from the xerogel surface since minimal NO trapping was observed in this configuration, and this distance could be achieved reproducibly.

Fig. 5 Effect of sensor–sample separation distance on the measured NO concentration above the surface of a NO-releasing xerogel films. Distance between sensor and substrate: A (10 µm), B (30 µm), C (50 µm) and D (100 µm).

Measurement of NO concentration at the surface of xerogel microarrays

To assess the relationship between the localized surface concentration of NO and platelet adhesion to micropatterned substrates, a variety of NO-releasing microarrays were prepared as described above. The microarrays employed in this study consisted of 50 µm wide xerogel lines (height = 10 µm, length = 8 mm) separated by distances ranging from 10 to 200 µm. The micropatterned substrates were immersed in PBS at physiological temperature and pH for 1 h prior to surface concentration measurements to ensure a constant rate of NO release. The sensing probe was then positioned at a distance of 10 µm directly above the surface of the xerogel microarray. Previous studies by Lee et al. demonstrated that a steady-state current was measured with a microelectrode NO sensor (∼150 µm od) at a distance of 10 µm from an NO-releasing substrate within 150 s.¹⁸ Thus, measurements with the ultramicroelectrode sensor described above were obtained for 5 min to ensure that the reported concentrations represented steady-state values. Concentration measurements were obtained after 1, 6, and 24 h immersion to correlate NO concentration with the observed platelet adhesion trends.

The localized concentrations of NO generated from 40% AEMP3 (balance MTMOS) xerogel microarrays of different geometry are provided in Table 2. As expected, the largest NO concentration for each type of micropattern was observed after 1 h immersion. The highest NO concentration, 237 ± 11 nM, was measured above arrays with xerogel line separations of 10 µm. At increased microstructure separations, lower NO surface concentrations were measured (Table 2). For microstructure separation distances up to 50 µm, NO concentrations were measurable for the duration of the study (i.e., 24 h) with NO levels of 13 ± 4 and 8 ± 5 nM above the microarrays with 25 and 50 µm separations, respectively. A decrease in NO concentration would be expected at larger line separations for a constant xerogel line width, since the fraction of unmodified glass (unable to release NO) directly beneath the sensing tip increases. Microarrays with line separations of 100 and 200 µm were characterized by initial NO surface concentrations of 50 ± 17 and 35 ± 21 nM, respectively. While NO release was measurable after 6 h solution immersion, the level of NO generation at longer times (>24 h) was below the detection limit of the sensor (∼5 nM). These observations are consistent with previous bulk chemiluminescence measurements used to characterize the NO release properties of AEMP3/MTMOS xerogels and are attributed to depletion of the finite NO donor reservoir within the xerogel lines with time.^13,14,17

Table 2 Localized NO surface concentration measurements measured 10 µm above the surface of xerogel arrays^a

Microarray/µm^b	NO surface concentration/nM^c
	1 h	6 h	24 h
a For each measurement, n ≥ 6. b Microstructure dimensions reported as width ∶ separation. c Measured at a probe–xerogel separation of 10 µm. d Below sensor detection limit.
50 ∶ 10	237 ± 11	131 ± 9	20 ± 5
50 ∶ 25	175 ± 18	94 ± 7	13 ± 4
50 ∶ 50	102 ± 22	63 ± 8	8 ± 5
50 ∶ 100	50 ± 17	23 ± 6	n/a^d
50 ∶ 200	35 ± 21	10 ± 9	n/a^d

---

Results from the NO surface concentration measurements indicate that the minimum NO concentration necessary for improving micropattern thromboresistivity is highly dependent on the microarray geometry. Indeed, patterns consisting of xerogel lines separated by ≤25 µm exhibited enhanced biocompatibility relative to control arrays at a minimum NO surface concentration of 13 ± 4 nM. Increasing the microstructure separation to 50 µm resulted in substrates that required a significantly higher NO surface concentration (63 ± 8 nM) to effectively resist platelet adhesion. These results demonstrate the potential for NO-releasing xerogel arrays with microstructures positioned ≤50 µm apart to resist platelet adhesion at a minimum sustained NO surface concentration of ∼65 nM. Micropatterns prepared with microstructure separations ≥100 µm did not resist platelet adhesion, indicating that a NO surface concentration >63 ± 8 nM would be required to render such interfaces more resistant to platelet adhesion. Unfortunately, the maximum NO concentrations that can be achieved with this xerogel composition are 50 ± 17 and 35 ± 21 nM for microstructure separations of 100 and 200 µm, respectively. Robbins et al. recently reported on the NO release properties of ethyltrimethoxysilane (ETMOS) xerogels doped with other types of aminosilane NO donor precursors, specifically N-(2-aminoethyl)-3-aminopropyl-trimethoxysilane (AEAP3) and N-(6-aminohexyl)aminopropyltrimethoxysilane (AHAP3).¹⁵ Xerogel microarrays prepared with AEAP3 and AHAP3 were characterized by 4- and 7-fold increases in NO release capability, respectively, relative to comparable AEMP3 arrays.¹⁵ Thus, by employing such arrays it may be feasible to achieve a sufficiently high NO surface concentration to prevent platelet adhesion to micropatterns with larger (i.e., ≥100 µm) microstructure separations.

While these results are promising, it is important to note that in vitro platelet adhesion experiments do not account for the numerous NO scavengers present in the body (e.g., proteins, thiols, transition metals, etc.) that would be expected to decrease the local concentration of NO.³⁵ Under these conditions, higher NO surface concentrations may be required to achieve enhanced blood compatibility.

Conclusion

Results presented herein demonstrated that the NO-selective ultramicroelectrode is an appropriate tool for determining accurate steady-state surface NO concentrations. Trapping effects were negligible since the miniaturized sensing probe (≤50 µm diameter) did not block (i.e., trap) NO from NO-releasing micropatterned xerogels. This particular NO-selective ultramicroelectrode sensor may prove useful for NO measurements in biological milieu due to its high sensitivity and fast response time.

In vitro platelet adhesion experiments at NO-releasing xerogel microarrays indicated that a microstructure separation of 50 µm and a 63 ± 8 nM NO surface concentration effectively resisted platelet adhesion, while microstructure separations ≥100 µm were incapable of reducing platelet adhesion with the existing NO fluxes. This limit may be circumvented by employing other NO-releasing xerogels that release greater NO levels than AEMP3/MTMOS xerogels.

Acknowledgements

This research was supported by the National Institutes of Health (NIH EB000708). The authors would like to thank the University of North Carolina at Chapel Hill Francis Owen Blood Research Laboratory for supplying porcine blood samples. MER gratefully acknowledges a Linda Dykstra Science Dissertation Fellowship from The University of North Carolina at Chapel Hill Graduate School (MER).

References

1. Biomaterials Engineering and Devices: Human Applications, vol. 1: Fundamentals and Vascular and Carrier Applications, ed. D. L. Wise, D. J. Trantolo, K.-U. Lewandrowski, J. D. Gresser, M. V. Cattaneo, and M. J. Yazemiski, Humana Press, New Jersey, 2000.
2. J. H. Braybrook, Biocompatibility Assessment of Medical Devices and Materials (Biomaterials Science and Engineering Series), John Wiley & Sons, New York, 1997.
3. M. C. Frost, M. M. Reynolds and M. E. Meyerhoff, Biomaterials, 2005, 26, 1685.
4. L. K. Keefer, Annu. Rev. Pharmacol. Toxicol., 2003, 43, 585.
5. B. J. Nablo, T.-Y. Chen and M. H. Schoenfisch, J. Am. Chem. Soc., 2001, 123, 9712.
6. D. J. Smith, D. Chakravarthy, S. Pulfer, M. L. Simmons, J. A. Harbie, M. L. Citro, J. E. Saavedra, K. M. Davis, T. C. Hutsell, D. L. Mooradian, S. R. Hanson and L. K. Keefer, J. Med. Chem., 1996, 39, 1148.
7. M. W. Radomski, D. D. Rees, A. Dutra and S. Moncada, Br. J. Pharmacol., 1992, 107, 745.
8. R. D. Rudic, E. G. Shelsey, N. Maeda, O. Smithies, S. S. Segal and W. C. Sessa, J. Clin. Invest., 1998, 101, 731.
9. J. E. Albina and J. S. Reichner, Cancer Metastasis Rev., 1998, 17, 19.
10. D. Efron, D. Most and A. Barbul, Curr. Opin. Clin. Nutr. Metab. Care, 2000, 3, 197.
11. Methods in Nitric Oxide Research, ed. M. Feelisch and J. Stamler, John Wiley & Sons, New York, 1996.
12. S. M. Marxer, A. R. Rothrock, B. J. Nablo, M. E. Robbins and M. H. Schoenfisch, Chem. Mater., 2003, 15, 4193.
13. M. E. Robbins and M. H. Schoenfisch, J. Am. Chem. Soc., 2003, 125, 6068.
14. M. E. Robbins, E. D. Hopper and M. H. Schoenfisch, Langmuir, 2004, 20, 10296.
15. M. E. Robbins, B. K. Oh and M. H. Schoenfisch, Chem. Mater., 2005, 17, 3288–3296.
16. K. P. Dobmeier and M. H. Schoenfisch, Biomacromolecules, 2004, 5, 2493.
17. B. K. Oh, M. E. Robbins, B. J. Nablo and M. H. Schoenfisch, Biosens. Bioelectron., 2005, 21, 749–757.
18. Y. Lee, B. K. Oh and M. E. Meyerhoff, Anal. Chem., 2004, 76, 536.
19. J. A. Hrabie, J. R. Klose, D. A. Wink and L. K. Keefer, J. Org. Chem., 1993, 58, 1472.
20. C. D. James, R. C. Davis, L. Kam, H. G. Craighead and M. Isaacson, Langmuir, 1998, 14, 741.
21. J. Cazenave and J. Mulvihill, in The Role of Platelets in Blood-Biomaterial Interactions, ed. Y. Missirlis and J.-L. Wautier, Kluwer, Dordrecht, 1993, pp. 69––80.
22. Scanning Electrochemical Microscopy, ed. A. J. Bard and M. V. Mirkin, Marcel Dekker Inc., New York, 2001.
23. Handbook of Chemistry and Physics, ed. D. R. Lide, CRC Press, Inc., Boca Raton, FL, 1990–1991.
24. E. F. Grabowski, P. Didisheim, J. C. Lewis, J. T. Franta and J. Q. Stropp, Trans.–Am. Soc. Artif. Intern. Organs, 1977, 23, 141.
25. J. C. Vidal, E. Garcia-Ruiz and J. R. Castillo, J. Pharm. Biomed. Anal., 2000, 24, 51.
26. J. C. Vidal, E. Garcia-Ruiz and J. R. Castillo, Anal. Sci., 2002, 18, 537.
27. Q. S. Li, B. C. Ye, B. X. Liu and J. J. Zhong, Biosens. Bioelectron., 1999, 14, 327.
28. Electrochemical Methods: Fundamentals and Applications, ed. A. J. Bard and L. R. Faulkner, John Wiley & Sons, New York, 2001.
29. D. Kato, M. Sakata, C. Hirayama, Y. Hirata, F. Mizutani and M. Kunitake, Chem. Lett., 2002, 1190.
30. F. Mizutani, Y. Hirata, S. Yabuki and S. Iijima, Chem. Lett., 2000, 802.
31. F. Mizutani, S. Yabuki, T. Sawaguchi, Y. Hirata, Y. Sato and S. Iijima, Sens. Actuators, B, 2001, 76, 489.
32. V. G. Kharitonov, A. R. Sundquist and V. S. Sharma, J. Biol. Chem., 1994, 269, 5581.
33. J. R. Lancaster, Jr., Proc. Natl. Acad. Sci. USA, 1994, 91, 8137.
34. T. Malinski, Z. Taha, S. Grunfeld, S. Patton, M. Kapturczak and P. Tomboulian, Biochem. Biophys. Res. Commun., 1993, 193, 1076.
35. M. Kelm, Biochim. Biophys. Acta, 1999, 1411, 273.

---