Feedback-controlled photolytic gas phase nitric oxide delivery from S-nitrosothiol-doped silicone rubber films

Abstract

Constant therapeutic gas phase nitric oxide (NO) delivery is achieved from S-nitrosothiol (RSNO) type NO donor doped silicone rubber films using feedback-controlled photolysis. For photo-release of the NO gas, the intensity of the LED light source is controlled via a PID (proportional–integral–derivative) controller implemented on a microcontroller. The NO concentration within the emitted gas phase is monitored continuously with a commercial amperometric NO gas sensor. NO release was accurately adjustable up to 10ppm across a broad range of setpoints with response times of roughly 1 min or less. When NO is generated into an air recipient stream, lower NO yields and a comparable level of toxic nitrogen dioxide (NO₂) formation is observed. However, NO gas generated into an N₂ recipient gas stream can be blended into pure O₂ with very low NO₂ formation. Following scale-up, this technology could be used for point-of-care gas phase NO generation as an alternative for currently used gas cylinder technology for treatment of health conditions where inhaled NO is beneficial, such as pulmonary hypertension, hypoxemia, and cystic fibrosis.

1. Introduction

Therapeutic use of gas phase nitric oxide (NO) has several important applications in the medicine. In addition to its well-known vasodilator action, NO is a potent and natural antimicrobial/antiviral agent normally present at moderate levels (0.2–1.0ppm) in the upper airways/sinuses of healthy individuals to help prevent chronic upper airway and lung infections, and to control ciliary beat frequency [1–3]. Since its first medical application > 20 years ago, inhaled nitric oxide (iNO) has become a mainstay of intensive care for lung failure patients [4]. iNO is essential in neonatology, lung transplantation, and pulmonary hypertension [5–12]. It is also used in treatment of pneumonia, acute respiratory distress syndrome (ARDS), and in other medical applications [7–9,13–17]. There is also great interest in using NO within the sweep gas of oxygenators employed in extracorporeal (EC) procedures to prevent activation of platelets on the blood side and to mitigate the occurrence of systemic inflammatory response syndrome (SIRS) [18]. Further, inhaled NO may provide a new strategy to improve penumbral blood flow and neuronal survival in stroke or other ischemic conditions [19]. A recent study also showed the benefits of using low-dose (10 ppm) inhaled NO as adjunct therapy for enhanced effect of antibiotics to treat acute Pseudomonas aeruginosa infection in cystic fibrosis (CF) [20].

Although inhaled NO has proven to be safe for inhalation therapy, to date it can be administered only in the hospital setting and it is not available as an in-home therapy for patients with chronic pulmonary diseases and pulmonary hypertension. The biggest barrier for more widespread application of inhaled NO therapy is the very high cost associated with its use [21,22]. This relates to the reactivity and low-storage stability [23] of NO gas in gas tanks due to the disproportionation of NO at high pressures into toxic NO₂ and dinitrogen oxide (N₂O) [23], and other safety concerns associated with the high concentration (800 ppm) NO tanks currently employed. Thus, gas cylinder-based technologies have very limited portability and generally air travel is not allowed on commercial airlines with pressurized gas tanks. In-situ generation of NO gas from stable, solid phase NO donors would address these stability and safety issues.

There have been several efforts to develop alternative technologies to the currently used gas cylinder-based NO delivery system. Yu et al. reported a system in which NO can be generated from atmospheric nitrogen using a pulsed electric discharge [24]. Lovich et al. published a technology which generated NO₂ from liquid N₂O₄ and then converts the NO₂ to NO at room temperature with an ascorbic acid cartridge [25,26]. This system requires careful monitoring and scrubbing of the very high levels of toxic NO₂ gas that could potentially be present in the output gas stream (from liquid N₂O₄).

The Meyerhoff group recently developed a technology for in-situ electrochemical generation of gas phase NO via copper(II)-tri(2-pyridylmethyl)amine mediated reduction of nitrite ions within a liquid phase. The liquid phase is continuously circulated through a hollow silicone fiber-based gas separator to create the desired levels of NO in a recipient air or nitrogen stream for final medical use [27,28].

The NO generation method presented here offers a very simple and cost-effective way of generating NO gas at therapeutic levels by photolysis of a solid RSNO donor embedded in silicone rubber films. Proof-of-principle experiments using this approach were recently reported by our team [29]. The purpose of this paper is to report on the development and capability of a continuous monitoring and feedback control system that can precisely control the output gas phase levels of NO delivered by this new photolysis method that will greatly enhance its potential biomedical applications.

2. Experimental section

S-nitroso-N-acetylpenicillamine (SNAP) was purchased from Pharmablock (USA). The purity of SNAP was determined by absorbance at 340 nm in phosphate buffered saline with 10 μM EDTA (PBSE) (ε_SNAP,340 = 1024M⁻¹ cm⁻¹ [30]) and was considered when calculating the actual loadings of the film and determining the NO and NO₂ yields.

Thirteen w/w% loaded SNAP doped silicone films were prepared as described in our previous work [29]. Briefly, base and curing agent of Sylgard 184 polydimethylsiloxane (PDMS) were combined in 10:1 w/w ratio and mixed thoroughly. Then, the solid NO donor was blended into it, degassed and cured at room temperature for at least 24 h. Cured films were placed into vacuum for another 24 h in order to remove volatiles from the film. RSNO-doped PDMS films were stored at room temperature in air, protected from light.

Sixmm and 12.5 mm diameter test pieces were cut from the RSNO loaded silicone rubber films using a biopsy punch and cork borer, respectively, for feedback controlled NO release experiments. The thickness of the films was measured using a micrometer (Mitutoyo). The NO donor doped films were placed onto a micro structured surface (Buehler MicroCloth) to prevent gas accumulation between the glass wall of the photolysis cell and the silicone film.

As a model system the same experimental setup was used as described in our earlier work [29], with the difference that in this case the light intensity of the Thorlabs M385LP1, M470 L3 or M565 L3 LED light sources was feedback-controlled. The light intensity was modulated through a LEDD1B T-Cube driver (Thorlabs) in trigger mode using TTL pulse width modulated (PWM) signal of the microcontroller. The driver current was maximized at 1000 mA and the maximal LED optical power density was set to 51 mW/cm² at 100% LED PWM duty cycle using a Thorlabs PM-16–401 calibrated optical power meter below the empty photolysis chamber.

For feedback control, a PID controller was implemented on an Arduino Uno compatible development board microcontroller (Ruggeduino-SE, Rugged Circuits), which continuously adjusted the intensity of the light source based on the error signal derived from the setpoint NO concentration and the measured NO concentration in the gas stream.

The NO concentration of the delivered gas stream was measured with an amperometric NO sensor (Alphasense, NO-B4) equipped with an individual sensor board (ISB) for polarizing the working electrode to +200 mV and converting current output to a voltage signal measured by the microcontroller. The NO₂ level was monitored with an amperometric NO₂ sensor (Alphasense, NO2-B43F) equipped with ISB and gas hood assembly (Alphasense). NO and NO₂ sensors with a wider working range (Alphasense, NO-A4 and NO2-A43F, respectively) were used with analog-front-end (AFE) boards and gas hood assembly (Alphasense) (see below Fig. 7). The voltage signal of the ISBs or AFE board was digitalized with a 16-bit analog-to-digital (A/D) converter (Adafruit ADS1115). The ISB or AFE board was powered from stabilized 5 V voltage directly from the Arduino board. 10 nF and 100 nF capacitors were used to decouple the voltage supply and A/D converter from ISB or AFE board in order to decrease emitted noise onto the sensor.

Fig. 7.

Schematic shows the experimental setup for oxygen enriching of NO gas emitted into nitrogen recipient gas (top) and feedback controlled NO release from 12.5 mm diameter (weight: 270mg, thickness: 2.5 mm) PDMS-SNAP film with target NO level of 10.0 ppm using 385 nm LED light source (bottom). Recipient gas was 40 SCCM N₂ and it was blended with 160 SCCM O₂. NO concentration was measured after mixing the two gas streams. PID parameters: K_p = 0.05, K_i = 0.00125, K_d = 0.0275.

A 2000mA 3.7V lithium-ion battery (Adafruit) was used on a PowerBoost 500 Shield (Adafruit) as an uninterruptible power supply to keep the sensors always biased. Experimental data were logged onto a microSD card using a data logging shield for Arduino (Adafruit).

A 24″ long 0.060″ diameter Nafion tubing (Perma Pure) was used to adjust the humidity of the gas stream to an ambient level and to prevent drying of the electrochemical sensors [31]. The sample line was a 1 m-long black polytetrafluorethylene tubing with 0.787 mm inner and 3.175 mm outer diameter.

A photolysis cell was made from borosilicate glass with a quartz window [29]. In order to assess the temperature change of the PDMS-SNAP films during photo-release type J thermocouple (using Fluke 52 dual input thermometer with 0.05%+ 0.3°C accuracy) was inserted into the film and the ambient temperature was measured at the inlet of the photolysis cell.

In order to eliminate the effect of the flow rate on the gas sensors (Fig. S3) the flow rate of the nitrogen (Cryogenic Gases) or air recipient gas was controlled (at 200 SCCM) with a mass flow controller (Alicat MCS series). A photograph of the basic system with its main components are shown in the Supporting info (Fig. S1). For evacuating the photolysis chamber a micro diaphragm pump was used (TCS D250).

The NO-B4 electrochemical NO gas sensor was calibrated using a 4.9 ppm NO (balance N₂, primary standard; from Cryogenic Gases) or 5.1 ppm NO (balance nitrogen, NIST traceable certified standard; from Grainger) calibration gases. The NO-A4 electrochemical NO gas sensor was calibrated with 44.3 ppm NO (balance N₂, primary standard; from Cryogenic Gases). The NO-B43F NO₂ sensor was calibrated using a 5.2 ppm NO₂ tank (balance air, NIST traceable certified standard; from Grainger). The NO-A43F NO₂ sensor was calibrated with 20.4 ppm NO₂ (balance N₂, certified standard; from Cryogenic Gases). Calibrations were performed at atmospheric pressure, at 200 SCCM flow rate. The main sources of uncertainty and estimated errors were as follows for the output NO levels: sensor (±5%), certified standard calibration gas (±2%). Thus, the estimated uncertainty of the measured output NO concentration was about ±7%.

Yields were calculated from the measured NO and NO₂ data, flow rates and actual RSNO loading of the films as described in our previous work [29]. Total yield represents the sum of NO and NO₂ yields. The main sources of uncertainty and estimated errors affecting the NO release duration and total yield were as follows: purity of NO donor (±1%), actual loading of film (±2%), film weight (±1%), mass flow (±1%) and measured NO level (±7%). Thus, the estimated uncertainty of the NO release duration and total yield was about ±12%.

3. Results and discussion

We demonstrated previously that NO can be photo-released from S-nitroso-N-acetyl-penicillamine (SNAP) particles embedded within silicone rubber film into the gas phase and the release kinetics is wavelength and optical power density dependent [29]. However, gas phase NO delivery at stable concentrations from these RSNO doped silicone rubber films is not feasible with the photolysis of embedded NO donors using constant light intensity [29]. In order to achieve very stable NO levels in the delivered gas stream, the active feedback control of the photolyzing light intensity is required. For this purpose, we employed an inexpensive commercial amperometric NO sensor (see Experimental section for details) to continuously monitor the mole fraction of NO in the delivered gas stream. Based on measured NO concentration with the sensor, a PID feedback loop was developed to continuously adjust the light intensity to generate the desired gas phase NO level. Fig. 1 shows a block diagram of the feedback-control electronics and the experimental setup.

Fig. 1.

Block diagram of control electronics (top) and the feedback-controlled photo release setup (bottom).

The microcontroller converts the digitalized analog voltage signal from the Individual Sensor Board (ISB) to NO concentration using calibration constants stored in the non-volatile memory (EEPROM) of the microcontroller. The PID constants were manually determined to provide tight control of the NO concentration in the output gas stream as follows: K_p =0.2, K_i =0.005, K_d =0.11 constants were used for all the experiments reported here, except where otherwise stated. Although the transient of photolysis is not a monotonic linear function of the light intensity [29], it is still possible to achieve rather stable NO release from the RSNO-doped films using the PID controller with the set-up shown in Fig. 1.

For testing the feasibility of the feedback-control we used LED light sources with different nominal wavelengths (385 nm, 470 nm, 565 nm), which were selected in our previous study [29]. SNAP-doped polydimethylsiloxane (PDMS) films were used as NO source and the target NO level was set to 1 ppm. With feedback control the films emitted stable 1 ppm NO into a 200 SCCM N₂ recipient gas stream for several hours using all the different LED light sources examined (Fig. 2A and B). To reach the target NO level (1 ppm) it took ca. 5 min with the 385 nm deep UV light source, with only a slight initial overshoot. By contrast, it took significantly longer with exposure to a longer wavelength light source, i.e., ca. 40 min with the 470 nm blue and 7.5 h with the 565 nm green lights. Fig. 2 displays the necessary LED PWM duty cycle (%) transients in order to achieve the target gas phase NO levels. Using the 385 nm deep UV light, most of the NO payload was instantly available, and increasing light intensity was necessary to maintain the target 1 ppm NO level. After reaching a given depletion of the films, the flux of NO released from the film (i.e., NO release rate/film surface area) was limited by the maximum light intensity, and the target NO level could not be maintained. After this point, the NO level of the output gas stream started to decrease. This occured at 84%, 86% and 52% cumulative NO release (% of theoretical based on moles of RSNO in the film) with the 385nm, 470 nm and 565 nm light sources, respectively. However, when continuing the photolysis for 48 h, most of the NO loading was released but with lower generated levels of NO in the gas stream. That is, 99% and 95% of the total possible payload emission was possible with the 385 nm and 470 nm light sources, respectively. The 565 nm light source triggered release of only 79% of the loading over the 48 h test period.

Fig. 2.

Feedback-controlled NO release from 6mm diameter 13% loaded PDMS-SNAP films (weight: 37.3(±4.9) mg, thickness: 1.5(±0.2) mm) into N₂. NO level in the delivered gas stream (solid line) and the cumulative NO release (dashed line) using (A) LED light sources with different wavelengths (385 nm, 470 nm, 565 nm) with (B) corresponding PWM duty cycles of the LED, and (C) a 385 nm LED light source with different target NO levels (5.0 ppm, 2.5 ppm, 1.0 ppm) with (D) corresponding PWM duty cycles.

The SNAP loaded film did not exhibit any significant loss in NO loading during a 6 month storage period, with > 90% of the loading still available for gas phase NO generation.

With an increase in the NO target level, the duration of stable NO release period decreased (Fig. 2C and D). For example, using SNAP doped films, the duration of stable NO release was 33.8 h at 1.0ppm, 10.3 h at 2.5 ppm and only 3.1 h at 5.0 ppm target NO level. At the 5.0 ppm target NO level, only 50% of the NO payload was released. After reaching this cumulative NO release, the system was not able to compensate by increasing the light intensity and the delivered NO concentration of the gas phase started to decrease. At 2.5 ppm and 1.0 ppm NO target levels, 75% and 84% of the NO loading was available at the target NO level, respectively. The effect of the film thickness on the NO release was assessed by measuring the NO release from PDMS-SNAP films with 1.0 mm, 1.5mm and 1.9mm thickness using the 385 nm LED light source. The NO release decreased to 90% of the setpoint when 89%, 86% and 67% of the total loading was reached, respectively (Fig. S2). Thus, thinner films are more favorable to achieve the desired zero-order release profile.

For testing the stability of the feedback control loop, we perturbated the system both by changing the target NO level setpoint (Fig. 3A) and the flow rate of recipient gas (Fig. 3B) during the photolysis NO release process. After these perturbations of the system, the NO generation stabilized at the target NO level in a reasonable time period (i.e., <2 min to reach 90% of setpoint) and the NO release remained stable. As expected, delivery of carrier gas with a higher NO level required higher LED intensity. When the flow rate of recipient gas stream was decreased, the PID controller decreased the light intensity accordingly, and less NO release was necessary to maintain the same level of NO at the lower flow rate of the gas stream. The sensitivity of the amperometric sensor is not completely independent of the sample flow rate. Changes in the flow rate from 200 SCCM (standard cubic centimeters per minute) to 300 SCCM or to 100 SCCM yielded a ca. ±4% error in the sensitivity (Fig. S3) in the delivered gas concentration with the current setup. In the future, a parallel sampling line for NO measurement with a steady flow rate may be necessary to incorporate in the system in order to avoid the variation in accuracy of the NO measurement at different flow rates of the delivered gas. When illuminated with constant optical power density (385 nm LED, 51 mW/cm²) the temperature of the PDMS-SNAP films increased by 8 °C, while the temperature of blank PDMS film increased only by 6 °C. Melvin et al. recently found that the decomposition temperatures of SNAP is 132.8 ± 0.9°C [32]. Thus, the heat generated by the LED light does not appear to play a major role in NO release in the demonstrated setup.

Fig. 3.

(A) Feedback-controlled NO release from 6mm diameter PDMS-SNAP film (weight: 28.0 mg, thickness: 1.1 mm) with target NO level changed stepwise to 0.5, 1.0, 1.5, 2.0, 2.5 and 5.0 ppm and then back in the reverse direction. Recipient gas was N₂. (B) Feedback-controlled NO release from 6mm diameter PDMS-SNAP film (weight: 30.5 mg, thickness: 1.2 mm) with target NO level of 2.5 ppm using 385 nm LED light source and system response to perturbation of the flow rate of N₂ recipient gas. Middle plots show the changes in flow rate of N₂ recipient gas during the experiment. Bottom plots show the associated PWM duty cycle of the 385nm LED light.

For NO inhalation therapy, the NO delivery is necessary in air or in oxygen enriched air. We observed, however, a significantly lower NO yield (24%) when air was used in place of N₂ as the carrier gas for NO delivery (Fig. 4). The corresponding duration of NO delivery at 1 ppm setpoint was also significantly shorter using an air stream. We hypothesize that the reduced NO delivery is related to the reaction of oxygen and nitric oxide to form nitrogen dioxide (NO₂).

Fig. 4.

Feedback-controlled NO release from 6mm diameter PDMS-SNAP films (weight: 37.2(±1.2) mg, thickness: 1.5(±0.1)mm) into N₂ and air with target NO level of 1.0 ppm using 385nm LED light source. Dashed lines show the cumulative yield. Recipient gas was 200 SCCM N₂ (black) or 200 SCCM air (red). Bottom plot shows the corresponding LED PWM duty cycles (%).

NO₂ is a toxic, unavoidable contaminant associated with NO therapy. For iNO, the maximum level of inhaled NO₂ should be always < 3 ppm [33] and future iNO technologies should minimize NO₂ exposure. Measuring the NO₂ levels in the gas stream indicated significant NO₂ formation when photo release was performed into air recipient gas stream, but no NO₂ release was observed when an N₂ recipient gas stream was employed (Fig. 5). The level of NO₂ formation cannot be explained by the oxidation of generated NO gas in the headspace of the photolysis chamber and in the tubing path. The total yield (NO + NO₂), however, was 49(±6)% in air vs. 89(±6)% in N₂. We also observed yellow discoloration of the films photobleached in air (Fig. 6). Therefore, we postulate that the generated NO₂ amount was actually higher, but it was absorbed on the silica particle filler of the Sylgard 184 PDMS material, which then caused discoloration of the films.

Fig. 5.

Feedback controlled NO release from 6mm diameter PDMS-SNAP films (weight: 20.8(±0.9) mg, thickness: 0.9(±0.1) mm) into N₂ (left) and air (right) with target NO level of 1.0ppm using 385nm LED light source. Shaded areas show standard deviation of three measurements.

Fig. 6.

PDMS-SNAP samples after photo-release with 1ppm NO target level in N₂ and air.

Although the NO₂ formation occurs in the silicone rubber, and NO gas generation from PDMS-SNAP films into the air/oxygen stream is suboptimal, it is also possible to generate the NO gas into a N₂ stream (Fig. 7) followed by blending the gas stream into O₂ to reduce NO₂ formation. With this method therapeutically relevant NO level (10 ppm) can be generated into hyperoxic gas mixture (80% O₂ in N₂) stream for 8 h, without significant NO₂ formation. This could be useful for low-dose NO delivery to treat pulmonary infections of CF patients [20] following scale-up to match necessary flow rates for low-flow inhaled NO therapy through nasal cannula.

We can also conclude that the main source of the NO₂ formation is the NO releasing film and a negligible amount of NO₂ is formed at the blending point of the generated NO and the O₂ gases or in the tubing between the mixing point and the sensor. At higher NO levels, however, these sources of NO₂ will have to be considered. Although we did not see any direct impact of the humidity on NO photo-generation (data not shown), humidification of the gas stream with hot saturated water vapor may decrease the NO₂ formation in the delivery line [34].

Another possible approach is using a vacuum to facilitate the NO diffusion from the silicone film into the gas phase, thus decreasing residence time of NO in the silicone and the NO₂ formation (Fig. S4). However, a sweep gas is necessary at very low flow rate (1 SCCM, air) for efficient gas transport from the evacuated photolysis chamber into the O₂ stream (Fig. S5). The advantage of this approach is that, it does not require a N₂ source (e.g., N₂ cylinder) for the NO generation; however, it requires an additional vacuum pump.

4. Conclusion

Generating steady and adjustable therapeutic levels of NO into a N₂ recipient gas stream by feedback-controlled photolysis of RSNO type NO donors embedded into silicone rubber films is possible and is potentially applicable for in-situ generation of NO gas for iNO therapy. Although O₂ present in the recipient gas increases the levels of emitted toxic NO₂ gas in the delivered gas stream, blending of NO generated in N₂ stream into O₂ is possible to provide considerably lower NO₂ generation. Thus, after scale-up the present technology offers a possible alternative to currently used cumbersome and expensive cylinder-based NO delivery technologies and provides an attractive new approach for creating a portable, in-home, low-dose NO generation system for therapeutic applications.