Portable Nitric Oxide (NO) Generator Based on Electrochemical Reduction of Nitrite for Potential Applications in Inhaled NO Therapy and Cardiopulmonary Bypass Surgery

Abstract

A new portable gas phase nitric oxide (NO) generator is described for potential applications in inhaled NO (INO) therapy and during cardiopulmonary bypass (CPB) surgery. In this system, NO is produced at the surface of a large-area mesh working electrode by electrochemical reduction of nitrite ions in the presence of a soluble copper(II)-ligand electron transfer mediator complex. The NO generated is then transported into gas phase by either direct purging with nitrogen/air or via circulating the electrolyte/nitrite solution through a gas extraction silicone fiber-based membrane-dialyzer assembly. Gas phase NO concentrations can be tuned in the range of 5–1000 ppm (parts per million by volume for gaseous species), in proportion to a constant cathodic current applied between the working and counter electrodes. This new NO generation process has the advantages of rapid production times (5 min to steady-state), high Faraday NO production efficiency (ca. 93%), excellent stability, and very low cost when using air as the carrier gas for NO (in the membrane dialyzer configuration), enabling the development of potentially portable INO devices. In this initial work, the new system is examined for the effectiveness of gaseous NO to reduce the systemic inflammatory response (SIR) during CPB, where 500 ppm of NO added to the sweep gas of the oxygenator or to the cardiotomy suction air in a CPB system is shown to prevent activation of white blood cells (granulocytes and monocytes) during extracorporeal circulation with cardiotomy suction conducted with five pigs.

Graphical Abstract

INTRODUCTION

Nitric oxide (NO) is endogenously produced and has important physiological functions that include: increasing vasodilation, preventing platelet adhesion/activation, promoting wound healing and angiogenesis, and serving as a potent antimicrobial agent released by macrophages and nasal epithelial cells to fight infection.^1–8 Direct inhalation of nitric oxide (INO) therapy is a treatment approved by the US Food and Drug Administration (FDA) for persistent pulmonary hypertension of newborn babies (PPHN)^9,10 and has been demonstrated to improve oxygenation and reduce the need for higher-risk extracorporeal membrane oxygenation (ECMO) therapy.¹¹ INO not only induces preferential pulmonary vasodilation and lowers pulmonary vascular resistance, but also has a beneficial effect on treatment of other illnesses including pneumonia,¹² stroke,¹³ and acute respiratory distress syndrome (ARDS).¹⁴ Recent studies have reported the use of INO as an inhaled antiseptic agent in the treatment of cystic fibrosis¹⁵ and tuberculosis,¹⁶ and as an anti-inflammatory agent to modulate immune response and promote survival in patients with malaria.¹⁷ INO has also been demonstrated to provide neuroprotection and reduce brain damage.¹⁸

One other significant area for potential clinical use of gas phase NO is in the sweep gas of oxygenators and within the cardiotomy suction air used in cardiopulmonary bypass (CPB) surgery extracorporeal circuits. CPB causes in some patients a severe systemic inflammatory response (SIR) that is associated with multiple organ failure (MOF), where the severity is related to the length of CPB.¹⁹ It is known that SIR is stimulated by blood-surface interactions and include leukocyte and platelet activation with release of cytokines.²⁰ Further, in vitro studies have shown that the inflammatory response and associated hemolysis during cardiotomy suction are related to the air exposure in that portion of the circuit and leukocytes are preconditioned by air exposure, priming them for adhesion and activation on the circuit’s surfaces.²¹ The anti-inflammatory properties of NO may also prove beneficial for reducing such complications with these procedures.

At present, INO and other therapeutic biomedical applications of NO require the use of a NO gas cylinder and a complex delivery device to regulate and monitor NO concentrations. Hence, the use of NO is considered one of the most expensive drugs in neonatal medicine since INO has a cost of ca. $3000 per day per patient,^22,23 but is still considered cost-effective in terms of decreasing the need for ECMO and being essential to prevent death of some neonates.²⁴ However, NO pressurized in a conventional gas cylinder can undergo a disproportionation reaction to form N₂O and highly toxic NO₂, limiting the use-life of a gas cylinders for medical applications that contain NO at levels >800 ppm.²⁵ Therefore, NO gas cylinders at ≤800 ppm levels are required for current INO distribution systems and these cylinders are heavy, cumber-some, and very expensive. Thus, there is a great need to develop an inexpensive yet portable source of relatively pure NO for use in INO and other biomedical applications to make use of gas phase NO more available to a greater number of patients, facilitating clinical trials of different gas phase NO therapies on more diseases, and ultimately making INO available for home use (e.g., for cystic fibrosis patients) and in remote areas.

Alternate gas phase NO generation techniques have already been proposed. These include catalytic conversion of liquid NO₂/N₂O₄ into NO²⁶ and generation of NO from air via pulsed electrical discharges.²⁷ However, in these techniques a large amount of highly toxic NO₂ needs to be used as the starting material or is produced as byproducts, which induces significant safety concerns.²⁸

In earlier work reported by our group, a controllable and inexpensive method was developed to generate NO by electrochemical reduction of nitrite using a copper(II)-tri(2-pyridylmethyl)amine (Cu(II)TPMA) complex as a mediator, and this approach was applied to develop intravascular (IV) catheters and sensors that emit low levels of NO to prevent clotting and infection.^29,30 We also demonstrated the temporal profile of NO generation in the gas phase in the 0–400 ppb range using a nitrogen purge of 0.2 L/min in a beaker of nitrite ions with added Cu(II)TPMA mediator, and the gas phase level could be modulated readily by applying different cathodic potentials to the working electrode. In this present work, the generation of gas phase NO at concentrations that are much more relevant for INO therapy and the oxygenator/cardiotomy suction air application, up to 500 ppm and at higher flow rates from 0.2–5 L/min, is explored using two different designs of the new electrochemical NO generation system in combination with a new Cu(II)-ligand complex that helps produce NO more efficiently than Cu(II)TPMA.

EXPERIMENTAL SECTION

Sodium nitrite (99.99%), N-(2-hydroxyethyl)piperazine-N′-(2-ethanesulfonic acid) (HEPES) acid (99.5%), HEPES sodium salt (99.5%), copper(II) sulfate pentahydrate (99.999%), and 1,4,7-trimethyl-1,4,7-triazacyclononane (Me₃TACN) (97%) were purchased from Sigma-Aldrich (St. Louis, MO) and used as received. All aqueous solutions were prepared with deionized water (18 MΩ cm⁻¹) from a Milli-Q system (Millipore Corp., Billerica, MA). Gold (Au) mesh electrodes (99.99%, 52 mesh, 5 cm × 10 cm) and Platinum (Pt) mesh electrodes (99.99%, 52 mesh, 5 cm × 5 cm) were purchased from Alfa Aesar (Ward Hill, MA). Teflon PFA-coated platinum wires (0.125 mm OD) from A–M Systems (Sequim, WA) were used as lead wires for the mesh electrodes. Nitric oxide standard gas cylinders (45 ppm) were products from Cryogenic Gas Inc. (Detroit, MI).

A continuous micro liquid pump TCS M200 S and a micro air pump TCS D3K were purchased from Servoflo Corp. (Lexington, MA). The gas permeable membrane modules PermSelect PDMSXA-2500 were products of Medarray (Ann Arbor, MI).

All the electrochemical experiments were performed on a Gamry 600 potentiostat (Warminster, PA). The generation of NO was measured in real time via chemiluminescence using a Sievers Nitric Oxide Analyzer (NOA 280i) from GE Analytics (Boulder, CO), a CLD 822 CM NO/NOx analyzer from ECO Physics, Inc. (Ann Arbor, MI) or a lab made gas phase amperometric NO sensor.³¹

Preparation of the Copper Catalyst.

Copper(II) sulfate pentahydrate (7 mM) and Me₃TACN (7 mM) were added to a solution containing 0.5 M HEPES buffer (pH 7.3) with 1 M NaNO₂, the solution was used directly for NO generation and testing.

For characterization purposes, Cu(II)Me₃TACN (structure is shown in Figure S1) was isolated as a solid by the following procedure. Copper(II) sulfate pentahydrate (113.71 mg, 0.455 mmol) and Me₃TACN (78.0 mg, 0.455 mmol) were dissolved in 10 mL methanol and stirred for 1 h, resulting in an intense blue colored solution. All solvent was then removed via reduced pressure. The resulting blue solid was recrystallized by dissolving in a minimal amount of methanol, layering with diethyl ether and placing in a freezer overnight. The resulting blue solid was collected via vacuum filtration, washed with additional diethyl ether and dried under vacuum. FT-IR (ATR): 2904, 1647, 1503, 1454, 1302, 1217, 1133, 1079, 1059, 1007, 983, 941, 924, 895, 786, 741, 656 cm⁻¹ (Figure S2). UV–vis (H₂O) λ_max: 270, 680 nm (Figure S3). EPR (H₂O, 30% glycerol; 110 K): g_⊥ = 2.055, g_∥ = 2.292, A_∥ = 463 MHz (Figure S5).

Cu(II)Me₃TACN ${\text{NO}}_2^{-}$ complex: The isolated Cu(II)-Me₃TACN solid was redissolved in H₂O resulting in a bright blue solution and mixed with solid NaNO₂ or NaNO₂ dissolved in H₂O to give the stated concentration, resulting in a green solution that was directly characterized. UV–vis (H₂O) λ_max: 276, 354 (NaNO₂), 652 nm (Figure S4). EPR (H₂O, 30% glycerol; 110 K): g_⊥ = 2.045, g_∥ = 2.25, A_∥ = 503 MHz (Figure S6).

Electrochemical Cell Configuration.

A homemade glass cell with ports for a bubbler and gas outlet was filled with a solution (80 mL) of 7 mM Cu(II)Me₃TACN, 1.0 M NaNO₂ and 0.5 M HEPES buffer (pH 7.3). A two-electrode system employing a constant current mode was used for electrochemical reduction of nitrite. A 5 cm × 10 cm Au mesh was used as the working electrode, and a 5 cm × 5 cm Pt mesh served as the counter/reference electrode. Both electrodes were fully submerged in the solution and separated to prevent shorting the circuit.

Gas NO Generation and Measurement.

In one mode of operation, a fritted glass bubbler was placed in the glass cell containing the solution for supplying a carrier gas (N₂ or air) to purge the NO generated on the electrode surface into the gas phase. The flow rate of nitrogen or air was controlled by a digital mass flow controller. In second design, the nitrite/Cu(II)-complex solution in the glass cell was continuously pumped through a gas separation silicone fiber dialyzer (PDMSXA-2500) via a micro liquid pump, and circulated back to the glass cell. A nitrogen or air recipient stream was introduced into the dialyzer module through the gas inlet port. The carrier gas sweeps through the inside of silicone fibers while the circulating solution from the electrochemical cell flows over the outside of the silicone fibers. The NO in the solution phase was separated and received into the gas phase because of its high permeability through the walls of the silicone fibers.

Nitric oxide in the carrier gas (N₂ or air) was measured by a chemiluminescence NO analyzer or via an amperometric NO sensor developed in our laboratory.³¹ Both the CLD NO/NOx and NOA 280i are calibrated with zero NO (N₂) and standard 45 ppm of NO or 10 ppm of NO₂ in N₂ cylinders. For the Sievers Nitric Oxide Analyzer (NOA 280i), a splitter was used if gas flow rate was higher than 0.2 L/min. To measure NO level with the electrochemical NO sensor, a gas stream was also split from the main stream at a flow rate of 0.05 L/min and flowed over the surface of the sensor, with excess NO at the outlet of the sensor being scavenged by an activated carbon cartridge.

Amperometric NO Sensor.

The amperometric Pt-Nafion gas phase nitric oxide sensor was fabricated according to our previous report,³¹ with minor changes. Briefly, Nafion 117 films (DuPont, Wilmington, DE) were cut into ca. 1.6 cm diameter circles and cleaned of impurities by boiling in 3 M nitric acid for 1 h, and then by boiling in deionized water for 1 h. Platinum was chemically deposited into/onto the solid-polymer electrolyte using the impregnation-reduction method. The Nafion 117 membrane was mounted between two glass cells with 0.92 cm diameter openings (0.66 cm² apparent geometric area), and one side was exposed to 2 mM Pt(NH₃)₄Cl₂ and incubated for 20 h at 37 °C. After that time, impregnation solution was removed and cell was washed with deionized water. Next, 50 mM NaBH₄ in 0.1 M NaOH was placed at the same side of the glass cell, and the chemical reduction was allowed to proceed for 1 h at 37 °C. Subsequently, the Pt-Nafion membrane was boiled in deionized water for 1 h to remove any remaining Pt complexes and reducing agents. The Pt-Nafion membrane was mounted in glass sensor assembly with the metallic side of the membrane electrode facing the gas phase. A 10 mm × 2 mm piece of 50 μm thick Au foil was used as the working electrode lead, and secured between the SPE membrane electrode and the gas inlet/outlet section of the sensor. A single junction Ag/AgCl (saturated KCl) reference electrode and bare Pt auxiliary electrode were placed in the liquid chamber filled with 0.5 M H₂SO₄ internal electrolyte. A CHI 206B potentiostat (CH Instruments, Austin, TX) was used to apply a potential to the working electrode (1 V vs Ag/AgCl) and record the sensor’s output currents. A MC-200SCCM mass flow controller (Alicat Scientific, Tucson, AZ) was used to deliver calibration gas and the gaseous sample to the gas phase sensor at a constant flow rate.

Porcine Model of CPB Studying the Effect of Nitric Oxide.

The experimental procedure was performed in an ovine model following protocol approval by the University of Michigan Institutional Animal Care and Use Committee (IACUC). All pigs used for the experiment were treated in compliance with the Guide for Care and Use of Laboratory Animals, eighth edition.³² Pigs were anesthetized, surgically instrumented, and connected to a CPB circuit. All animals were triaged prior to surgery to have normal preoperative white blood cell count between 14 and 20 K/uL, if an animal had white blood cell (WBC) out of this range; the animal was not used in the study. In addition, per our laboratory protocols, all animals received a prophylactic dose of IV antibiotics (1 g nafcillin and 80 mg of gentamycin) 1 h prior surgical incision. The blood in the control group (n = 5) was not exposed to air since no cardiotomy suction was applied, while the air-blood interface (ABI) group (n = 5) was placed on a three-pump circuit, which exposed the blood to air with cardiotomy suction. The treating group followed the same procedure as the ABI group but with 500 ppm of NO present in the sweep gas of the oxygenator or added to cardiotomy suction air. The schematics of the in vivo venoarterial (VA)-CPB circuit porcine models are shown in Figure 7, below.

Figure 7.

Schematics of the in vivo VA-CPB circuits using porcine model. (A) VA control model with minimal air-blood interface; (B) VA air-blood interface (ABI) model exposed the blood to air with cardiotomy suction; (C) INO group with gaseous NO delivery in the cardiotomy suction loop; (D) INO group with gaseous NO delivery in the oxygenator sweep gas inlet. RA: right arm; Ao: aorta (via carotid artery); IVC: inferior vena cava; Oxy: oxygenator.

Under general anesthesia, a venous cannula was placed to drain blood from the right atrium (RA) and inferior vena cava (IVC) into an open venous reservoir of the heart-lung machine. The venous blood was then passed through a pump and oxygenator (blood flow 1 L/min), and was reinfused through a cannula placed in the ascending aorta via the carotid artery to complete the venoarterial (VA) circuit. In the treatment group, a dose of 500 ppm of NO from the proposed E-chem NO generator was added to the gas inlet of the oxygenator within the CPB system or directly purged into the bloodstream at the cardiotomy suction/air interface. After 2 h on the circuit, the pigs were recovered and monitored for 96 h in order to assess the long-term effects of the blood’s exposure to air and negative pressure as compared to the baseline measurements. After 96 h, the pigs were euthanized and necropsies were performed. Systemic hemodynamics and blood serum samples were collected during the CPB procedure and during the recovery period to assess levels of blood activation and results were analyzed using an unpaired two-tailed Student t test with a p-value <0.05 considered significant. Systemic hemodynamics were stable for all of the animals throughout the 96 h postoperative period.

RESULTS

Cu(II)Me₃TACN as Mediator for Electrochemical Reduction of Nitrite.

The new Cu(II)Me₃TACN mediator complex (ligand structure shown in Figure S1), inspired by the active site of copper nitrite reductase,³⁴ was initially studied by cyclic voltammetry (CV) and bulk electrolysis in the presence of nitrite. This Cu(II)-ligand complex was characterized for its formal redox potential as well as the maximum NO production rate. For these experiments, a 1 mM catalyst solution (in electrolyte) with 0.1 M nitrite was used. As shown in Figure S1 (supplemental file), the Cu(II)Me₃TACN complex exhibits an E_1/2 of −0.24 V vs a Ag/AgCl reference electrode that was determined from the midpoint of both the cathodic and anodic peak potentials in the CV at a scan rate of 50 mV/s under deoxygenated conditions (solution purged with nitrogen). The maximum NO generation rate under these conditions is 5.7 μmol⁻¹cm⁻² (based on the geometric surface area of the Pt working electrode) as determined by a chemiluminescence nitric oxide analyzer (NOA) during bulk electrolysis of the same solution in the presence of 0.1 M NaNO₂ at −0.8 V for reduction of nitrite. Based on the ratio of the moles of NO measured by the NOA over a given time period (integration of the NOA signal over time) divided by the total charge passed in the electrochemical cell (current × time) divided by Faraday’s constant, the Faradaic efficiency of the Cu(II)Me₃TACN mediated reaction to produce NO was determined to be 93%. This far exceeds the efficiency provided by the Cu(II)TPMA species (6%) when examined under the same experimental conditions. The cyclic voltammogram of Cu(II)Me₃TACN in the presence of different concentrations of nitrite exhibits a catalytic cathodic wave that increases as the nitrite concentration increases (see Figure S1).

Electrochemical Gas Phase NO Generator with Fritted Gas Bubbler.

In the new generators, NO is produced on the surface of a large area Au mesh electrode surface that is submerged in the solution of sodium nitrite and Cu(II)-Me₃TACN. Generation of gas phase NO to the delivery system can be accomplished with either of the two configurations shown in Figure 1. The initial design (Figure 1A) uses a glass bubbler placed in the solution to purge it with N₂ or air that sweep the NO gas generated at the mesh working electrode into the gas phase. A membrane filter was added to the emitted NO/N₂ stream to remove any possible aerosol droplets that may contain nitrite ions and the Cu(II) ligand complex. The two-electrode system employing a 5 cm × 10 cm Au mesh electrode as working electrode and a 5 cm × 5 cm Pt mesh as the counter electrode was used to control the reaction and resulting gas phase NO concentration (we used an Au working electrode and a Pt counter electrode for most of the studies reported here). By applying different constant currents between the two electrodes, a wide range of NO levels (from 5 ppm to >1000 ppm) can be generated in the gas phase when N₂ is used as the carrier gas. The time for the NO concentration to change from one level to another when changing the applied current is <5 min, demonstrating a good temporal control of the system (Figure 2A). The fluctuation of the purging gas flow rate may lead to the observed spikes in the time trace, which can be eliminated by using a precise mass flow controller. Under a given purge gas flow rate, the gas phase NO concentration in N₂ is proportional to the current applied as shown in Figure 2B, and the associated faradaic efficiency is relatively stable (fluctuation <10%) for different currents. The electrochemical generation of NO is stable for continuous use for at least 1 week (Figure 2C).

Figure 1.

Gas phase NO generator designs investigated in this work. Design A: with direct bubbling of the solution. Design B: with circulation of the nitrite/Cu(II)-Me₃TACN solution through silicone fiber-based gas separation module.

Figure 2.

NO data from the gas phase NO generator with direct bubbling of the solution (Design A in Figure 1). (A) Time trace of NO level (ppm) vs applied current; (B) calibration of gas phase NO (ppm) vs applied current; and (C) example of gas phase NO levels generated over 7 d of continuous generation/measurement of NO from the same solution of nitrite/Cu(II)Me₃TACN. Working electrode: 5 cm × 10 cm Au mesh; counter electrode: 5 cm × 5 cm Pt mesh; N₂ bubbling rate: 0.2 L/min.

The NO/N₂ stream can be mixed with another stream of air or O₂ to provide desired NO levels for biomedical applications. As demonstrated in Figure 3, by mixing NO/N₂ from the generator with different ratio of N₂, air (21% O₂), or pure O₂, various gas phase NO concentrations can be obtained. The final NO concentration is proportional to the dilution ratio. Using air or pure O₂ as diluting gas does not change the NO concentration because of fast mixing and short contact time of NO and O₂.

Figure 3.

Production of different gas phase NO concentrations (ppm) using different mixing ratios with N₂ (green open triangle), air (black open circle), and 100% O₂ (red open square). (A) The flow rate of NO/N₂ from the generator is 0.1 L/min for set A; (B) flow rate set at 0.2 L/min; and (C) 0.5 L/min. Applied current is 10 mA; working electrode: 5 cm × 10 cm Au mesh; counter electrode: 5 cm × 5 cm Pt mesh. The NO generator system with direct purging shown in Figure 1A was used to obtain these data.

It is desirable to demonstrate that instead of N₂, air can also serve as a carrier gas. However, as shown in Figure 4A, when purging the nitrite/Cu(II)Me₃TACN solution with air, the NO measured in the emitted gas phase is decreased more than 80% at an applied current of 5 mA. This is because the electrolyte solution is saturated with O₂ and the high concentration of NO generated locally at the surface of the Au electrode can react with the oxygen at an appreciable rate, producing oxidized products. It might also be possible that superoxide can be generated in this case by direct reduction of O₂. Superoxide reacts extremely fast with NO, at a diffusion controlled rate (k ~ 10¹⁰ mol⁻¹s⁻¹).³⁹ Although ppm levels of NO can still be produced with this configuration while the air is used as a purge gas, the levels >50 ppm cannot be reached. Interestingly, simultaneous measurement of NO and NOx species with the CLD 822 CM NO/NOx analyzer demonstrated that in both conditions NO is still the predominant species in the gas phase. Other nitrogen oxides such as N₂O and NO₂ are <1 ppm for NO concentrations from 5 to 500 ppm.

Figure 4.

NO concentration comparison between air and N₂ as purge gas. A: NO generator using the configuration shown in Figure 1A; B: NO generator using the configuration in Figure 1B. Working electrode: 5 cm × 10 cm Au mesh; counter electrode: 5 cm × 5 cm Pt mesh; applied current: 5 mA.

Electrochemical NO Generator with Gas separation Silicone Fibers.

In order to generate higher concentrations of pure NO and using air as the carrier gas, the solution in the electrochemical cell was pumped rapidly and continuously into a gas extraction device that contains thousands of silicone hollow fibers (see Figure 1B). By passing a recipient gas (air) on the other side of the silicone fibers, NO is extracted into the gas phase due to its high permeability through the silicone materials,^30,35 while the solution and all nongaseous species are circulated back to the glass electrochemical cell. This design allows for a fast removal of NO from the electrode surface and the solution so that the reaction of NO with oxygen (coming from the air) is minimized. Figure 4 shows that with design A in Figure 1, when N₂ is replaced with air as the purge gas, this causes a dramatic decrease of NO production from 350 to 30 ppm (Figure 4A). However, when design B with the silicone fiber-based gas separation module is employed, the NO level decrease in going from nitrogen to air as the recipient gas through the fibers is only from 380 to 190 ppm under identical experimental conditions (Figure 4B). The solution enters the fiber module through the inlet port to the tube side and flows through the outside of the hollow silicone rubber fibers. In the fiber inner side, a sweep/carrier gas flows therein to carry away the permeating NO. The gaseous NO in the solution phase with higher permeability will efficiently transfer at a significant rate across the walls of silicone hollow fibers into the carrier gas. The solution circulation NO generator allows easy and reasonably rapid temporal control of the resulting gas phase NO concentration by adjusting the applied current or the recipient gas flow rate. At a fixed gas flow rate, the NO level can be easily tuned by changing the applied current and the resulting NO concentration is proportional to the current as shown in Figure 5B. The time for the NO level to change while applying a different current is less than 5 min (Figure 5A). A long-term stability experiment showed that over 24 h a stable NO with high concentrations up to 400 ppm can be produced continuously using the design B.

Figure 5.

NO data from the NO generator with solution circulation (Figure 1B). Different currents were applied for the stepwise change; 0.1 L/min and 0.2 L/min air flow rate through fibers were used. (A) Time trace of NO (ppm) vs applied current with different air flow rates; (B) calibration of NO (ppm) vs applied current at 0.1 L/min air flow rate. Working electrode: 5 cm × 10 cm Au mesh; counter electrode: 5 cm × 5 cm Pt mesh.

The purity of NO in the gas phase generated using the design B (Figure 1B) was investigated with the ECO-Physics CLD 822 CM NO/NOx analyzer by measuring the NO and NO₂ levels simultaneously. This analyzer has a highly efficient NOx converter that converts all NOx into NO, and the total NO is measured and compared to the NO level detected in the unmodified stream. The difference is identified with the NO₂ level. The ECO-Physics NO₂ detection system that we employed to quantitate the gas phase NO₂ levels is a widely employed commercial system designed to detect NO₂ levels with reasonably good accuracy. Hence, we are confident that the levels we have measured are accurate. The results shown in Figure 6 demonstrate a relatively low level of NO₂ (<1.5 ppm) during continuous generation of NO at ca. 60 ppm level with the air as the carrier gas. The NO₂ level did not increase when the current was changed to increase the gas phase NO concentration. To validate the experimental data, a bag of 43 ppm of NO from an NO cylinder mixing with air for 15 min, was measured with the analyzer at 16 min. Under these conditions 8 ppm of NO₂ was produced which corresponds well to the reaction rate of NO with O₂ in the air.³³

Figure 6.

NOx, NO, and NO₂ measured by the CLD 822 CM NO/NOx analyzer for the design B NO generator. Blue line: total NO (NOx) concentration in ppm after conversion; red line: NO concentration; green line: NO₂ concentration by subtracting the NO concentration from the NOx concentration. At 16 min, the sample was switched to a bag of 43 ppm of NO from an NO cylinder mixed with air for 15 min. Air flow rate from the gas pump: 1 L/min; the NO generating solution was circulated through a micro liquid pump and a PDMSXA-2500 silicone fiber module was used as gas separation device; applied current: 50 mA and 40 mA; working electrode: 5 cm × 10 cm Au mesh; counter electrode: 5 cm × 5 cm Pt mesh.

Monitoring NO from the Generator with Amperometric NO Sensor and Application for in Vivo Animal Studies.

The nitric oxide from the electrochemical NO generator was used in extracorporeal circuit procedures involving a porcine model. A constant current of 80 mA was applied in the generator of design configuration A (Figure 1A) and nitrogen with flow rate of 0.55 L/min was used as the purge gas in the NO generator. This provides 1000 ppm gaseous NO in N₂. NO from the generator (1000 ppm) was then split into two streams, one with 0.05 L/min NO in N₂ was delivered to the amperometric NO sensor for continuous monitoring of NO levels; and the rest of the NO/N₂ (0.5 L/min) stream was mixed with 100% O₂ in 1:1 ratio immediately before being delivered to the animal to provide the 500 ppm level in a 50% oxygen background (so high oxygen is going into the oxygenator or air portion of the cardiotomy suction unit employed in the animal experiments), so there is very little time for the NO to react with the oxygen. The NO sensor exhibits rapid response to gas phase NO, with linear response from 5 ppb–1000 ppm of NO and a response time of ~ <1 min. The NO sensor results were compared to measurements made with the conventional chemiluminescence NO analyzer, and showed an excellent correlation (Figure S7).

The effectiveness of gaseous nitric oxide (NO) produced by our new electrochemical NO generator to reduce SIR during CPB was studied by using the VA-CPB models (schematics shown in Figure 7) in pigs. The control group, without exposure to air by cardiotomy suction, had low CD11b expression, while ABI group with cardiotomy suction showed large increase in CD11b expression by granulocytes and monocytes. This indicates that WBC activation and inflammatory response is likely caused by the cardiotomy suction portion of the CPB system. When a dose of 500 ppm of NO was added to the CPB system in the sweep gas of the oxygenator or directly to the bloodstream via the air used in the cardiotomy suction portion of the system, the expression of CD11b on granulocytes (Figure 8A) and monocytes (Figure 8B) was preserved within normal values and followed the trend of the control group during CPB. These results demonstrate the protective effects of NO on white blood cell activation, using NO produced by the electrochemical NO generator.

Figure 8.

CD11b expression by granulocytes and monocytes normalized to baseline at different time points. (A) CD11b by granulocytes; (B) CD11b by monocytes. The addition of NO in the CPB system prevented activation of WBC (granulocytes and monocytes). Control: without exposure to air; ABI: the blood exposed to air with cardiotomy suction. ABI + 500 ppm of NO in sweep (n = 3) group: animals sustained exposure to the blood-air interface plus the addition of NO (500 ppm after mixing with O₂) in the sweep gas of the oxygenator; ABI+500 ppm of NO cs: animals sustained exposure to the blood-air interface plus the addition of 500 ppm of NO in cardiotomy suction.

DISCUSSION

NO gas cylinders (typically 800 ppm of NO in N₂) and a complex delivery unit is the only FDA approved device for INO and other therapeutic biomedical applications of NO. The dose of NO gas administered varies based on the illness, but falls within the range of 0.1 to 100 ppm, for current medical use (where direct inhalation is the mode of delivery to the patient). However, even higher levels may be needed when used in other applications (e.g., bypass surgery), since only a fraction of the NO present in the sweep gas of the oxygenator or the air of the cardiotomy suction line will actually get into the patient. Hence, the goal of this study was to develop a portable and controllable NO generating/delivery device that can provide NO levels of 0.1–500 ppm in the final gas stream for at least 1 d or longer, so that the same system can be used for all possible clinical applications in a hospital setting or potentially at home.

Several Cu(II) model systems have been shown to mediate nitrite reduction to NO.³⁴ Similar to the previous catalyst, Cu(II)TPMA,³⁰ the cyclic voltammogram of Cu(II)Me₃TACN in the presence of different concentrations of nitrite exhibits a catalytic cathodic wave that increases as the nitrite concentration increases (see Figure S1). The anodic peak corresponds to the oxidation of Cu(I) that is generated from Cu(II) complex in the solution. This oxidation reaction is suppressed in the presence of higher nitrite concentrations, because now Cu(I) as a mediator quickly reacts with nitrite to make NO, and hence, on the anodic scan, there is less Cu(I) to be oxidized at the electrode surface, and the signal is decreased. Upon adding more nitrite to the solution, more Cu(I) will react, and the anodic signal is decreased even further. The anodic signal will disappear completely, when all of the Cu(I) created in the cathodic scan is consumed by nitrite immediately. The conditions for Cu(II)Me₃TACN can be tuned for this species to electrochemically generate predominately NO. The excess nitrite (1 M) used in our experiments can competitively bind to the Cu(I/II) center of the Me₃TACN complex (after reduction of nitrite to NO), and prevent NO binding to the Cu(I/II) center, thereby suppressing N₂O generation, which has been confirmed by high NO purity and a Faradaic efficiency of 93% measured for the Cu(II)Me₃TACN mediated NO generation. Structurally, the methyl groups in Cu(II)Me₃TACN provide more steric hindrance and therefore slow down the disproportionation reaction, that requires two Cu(I) complexes to approach each other in close proximity. Indeed when using a Cu(II)TACN, ligand without the methyl groups, no nitrite reduction to NO can be observed due to the much faster disproportionation reaction with k^dis = 2.50 × 10³ M⁻¹ s⁻¹. The new copper(II) complex also exhibits better chemical stability. During storage in air for 60 d, the NO generating solution produced a very similar amount of NO at a given (applied) current on different days.

In the new NO generation systems, the constant current method was used to produce NO in contrast to the constant potential method introduced in the earlier catheter studies.³⁰ This is advantageous because the gas phase NO concentration is proportional to the NO flux from the electrode surface at steady-state, which, in turn, is directly determined by the level of current passed through the working mesh electrode. Therefore, the constant current method provides a more stable NO concentration/generation rate and is less sensitive to fluctuations in temperature or solution composition compared to using the constant potential mode. In addition, the constant current method consists of only two electrodes, making a conventional reference electrode unnecessary. Further, different inert metals or alloys can be used as working and counter electrodes. Beyond Au, we have studied Pt mesh and stainless steel mesh electrodes as well, and have obtained similar NO generation trends. With 5 cm ×10 cm Au or Pt mesh electrodes, NO concentrations from 5 to 1000 ppm can be generated when changing the current from 0 to 80 mA (the potential difference is shown in Table S1 in SI), whereas under the same conditions the stainless steel mesh electrodes produce only 20% less NO compared to the Au/Pt mesh electrodes. These results suggest that we can create low cost and robust high surface area working electrodes with stainless steel mesh or very thin layers of Pt or Au deposited on the surface of the stainless steel mesh electrodes allowing for the preparation of very inexpensive NO generation systems.

At a fixed purge gas flow rate, highly stable gas phase NO concentrations ranging from 5 ppm to over 1000 ppm can be generated with Design A by applying different constant currents between the two electrodes when N₂ is used as purge gas through the bubbler. The current efficiency increases as the bubbling/purge flow rate increases. Higher flow rate induces more efficient mass transfer, for both the catalyst to reach the surface of the working electrode, as well as the substrate (nitrite) mixing with the reduced catalyst very near the electrode surface after the Cu(I)-ligand complex is created electrochemically. Also, a higher flow rate captures/purges more NO over a given time period, decreasing the concentration of NO in solution more effectively, preventing the cross reaction of produced NO at the counter electrode and disproportionation of NO to N₂O by the catalyst.

Nitric oxide generated from the electrochemical device on demand at ambient pressure (when needed), avoids contamination of the NO gas with N₂O and toxic NO₂, resulting from the disproportionation of NO when stored over extended periods of time. This remains a critical problem for NO in conventional high pressure cylinders. The new NO generator can produce relatively pure NO with negligible byproducts of N₂O and NO₂ based on the NO purity test and calculated Faraday efficiency. Further, any possible aerosol droplets that contains nitrite ions and the Cu(II) complex can be removed by adding a membrane filter to the outlet NO/N₂ stream that can then be merged with another stream of air or oxygen for NO delivery at a desired concentration for an INO application. Based on the amount of nitrite (1.0 M) in the solution, the amount of NO (in moles) that can be produced from one 80 mL volume of solution using the electrochemical device is equivalent to more than 10 commercially available cylinders of INO (D-size, 800 ppm, 2000 psig). The electrodes and electronics are very stable and can be reused many times; the consumables are only the solution, which is very inexpensive compared to the current INO cylinder.

Using nitrogen as the carrier gas can eliminate the reaction between O₂ and NO, and prevent the copper catalyst scavenging by the O₂ that produces superoxide, which is very efficient in destroying NO. On the other hand, it is desirable to demonstrate that instead of N₂, air can also serve as a carrier gas so that more portable and low cost NO generators can be fabricated with an air pump to provide the carrier gas. When using air as the bubbling gas in Design A, the system cannot produce more than 50 ppm of NO. This is likely due to the rapid reaction of NO and O₂ with water to form nitrite in the solution close to the working electrode surface, as well as the scavenging of Cu(I) by O₂. The reaction between NO and O₂ is second order with respect to NO both in the gas phase and in the solution phase, resulting in an increased reaction rate at higher NO concentrations. In the gas phase, it takes 12 s to produce 1 ppm of NO₂ from the reaction of 200 ppm of NO and air,³³ while in the solution phase, the product of NO oxidation is nitrite. This explains why no significant amount of NO₂ (<1 ppm) was produced even when air is used as a carrier gas in the experiment since most of the reaction between NO and O₂ occurs in the solution phase.

To overcome this problem, we fabricated the second design (B) to obtain gaseous NO from the nitrite/electrolyte solution (Figure 1B). Here, the solution in the electrochemical cell is pumped/circulated rapidly into a gas extraction device by a micro liquid pump with a flow rate up to 700 mL/min. When applying a constant current to the mesh electrodes, high levels of NO are produced on the electrode surface and are carried away immediately when the solution is rapidly and continuously pumped into a silicone fiber module. We used a PermSelect silicone rubber hollow fiber membrane oxygenator from Medarray (Ann Arbor, MI) to separate NO from the solution. The silicone membrane modules use silicone hollow fibers with exceptional gas transfer properties. Because silicone is dense (nonporous), liquids cannot grossly transfer through the fiber walls, while NO has high solubility and permeability in the walls of the silicone rubber fibers. A sweep gas was introduced to receive NO from the solution on the other side of the silicone fibers, while the solution and components within it are retained and pumped back to the electrochemical cell. This new design allows for a fast removal of NO from the electrode surface and the solution and results in nearly a 6-fold enhancement in the gas phase levels of NO compared to direct bubbling of the nitrite solution (when air is used instead of N₂ as the purge gas). Indeed, device configuration B is more suitable for INO therapy. With this design, 500 ppm of NO can be readily be obtained even with air as the carrier gas through a micro gas exchanger. The example shown in Figure 4 does not illustrate the highest levels of current that can be applied to create the highest possible levels of NO. Further, a more efficient hollow fiber membrane oxygenator unit and pumping system can reduce the loss of NO and provide even higher concentrations of gas phase NO. A low level of NO₂, around 1 ppm, can be further eliminated with an NO₂ scrubber (e.g., zinc oxide particles) when air is used as the carrier gas in design B.

For configuration B, a stable NO can be continuously generated for at least 2 days. We observed that adsorption of the copper complex mediator on the PVC tubing during solution circulation will eventually create a situation where the loss of the mediator decreases the levels of NO that can be produced. Further, the continuous exposure to air decreases the activity of the copper catalyst. New solution/components (including more Cu(II) complex) can be added to the electrochemical cell without stopping the NO generation process, to further enhance the lifetime of achieving the desired level of NO gas. The stability and duration can also be improved by using more stable mediators and immobilizing the copper complex onto the electrode surface. These approaches are now being examined in our laboratory.

The new E-chem NO generator integrated with an amperometric NO sensor was used to study the cause of SIR during CPB with ABI and prevention by NO. Previous research suggests that air exposure due to cardiotomy suction is the major cause of white blood cell activation and inflammatory response in CPB.³⁶ Deleterious effects of cardiotomy suction are well-known and surgeons try to limit the amount of suction, and process the cardiotomy suction blood, but it is a necessary component of open heart operations, and blood exposure to air is currently inevitable during cardiac surgery. Blood which has been extravasated into the pericardial or pleural cavities and which is subsequently aspirated by cardiotomy suction markedly differs from the intravascular blood or blood within a closed CPB circuit. The concurrent suction of air results in highly turbulent flow with high shear stresses at the air-fluid interface. This causes cellular damage and activates all of the humoral cascades involved in the systemic inflammatory response. CD11b is expressed on the surface of many leukocytes including monocytes, neutrophils, natural killer cells, granulocytes and macrophages. Granulocyte CD11b expression is used as a pharmacodynamic biomarker of the degree of inflammation. The high CD11b expression by granulocytes and monocytes from the ABI group with cardiotomy suction is consistent with this conclusion. Nitric oxide is endogenously produced and known to prevent platelet adhesion/activation and inflammatory response.^3,6 Hence, if NO can serve as anti-inflammatory agent, the levels of CD11b expression should remain low. Initial tests with 50 ppm of NO did not induce significant changes in the CPB experiments. When using cardiotomy suction, adding 500 ppm of NO into the sweep gas of the oxygenator or directly into the blood can prevent WBC activation. The CD11b expressions by granulocytes and monocytes are similar to the control group without ABI. In addition, it was confirmed that 500 ppm of NO does not have a negative effect in the control by comparing the platelet count and methemoglobin (MetHb) levels of the control and ABI groups with the NO group.

The total moles of NO delivered during the 2 h swine studies can be estimated from the gas phase NO concentration delivered to the animal (either to oxygenator or cardiotomy suction air line), the flow rate, and time. The moles NO are calculated to be 0.005 mol (NO: 500 ppm; flow rate: 2 L/min; time: 2 h) if we assume there is no loss of NO. Of course, this is the maximum level if all the NO delivered actually gets into the blood, which is not possible, since the transport of NO through the fibers of the oxygenator used in the animal studies is likely <50%. NO that does get into the blood reacts with Fe(II)Hb very quickly. Therefore, the total amount of NO is much less than the total moles of Fe(II)Hb in blood (about 0.12 mol of Fe(II)Hb in 4 L blood) and this explains why there is no significant measurable change in MetHb levels observed in our studies.

MetHb can be measured to assess the potential toxicity of NO as cyanosis and hypoxia induced by high levels of NO have been reported, leading to MetHb levels >15% and >30%, respectively.^37,38 The maximum MetHb value measured for all animals treated with 500 ppm of NO during this study was <2%, which lacks clinical significance. As shown in Figure 8, nitric oxide has a very beneficial effect in reducing WBC activation with no negative effect in cardiopulmonary bypass with cardiotomy suction. It also has been reported that the delivery of nitric oxide to the oxygenator gas flow during pediatric cardiopulmonary bypass reduced the incidence of low cardiac output syndrome by varying degrees.⁴⁰ However, the exact mechanism of how NO inhibits the cause of SIR during CPB and how the prevention by NO works is still unknown. This will require more fundamental studies that demands a robust and cost-effective NO delivery system. Hence, the new portable NO generator described herein will be used for future studies in our laboratory to answer this question.

CONCLUSIONS

Two different designs for a novel electrochemical NO generation system are described in this article and tested subsequently for generating relatively pure gaseous NO from inorganic nitrite ions (by a copper catalyst) for potential biomedical applications. The new NO generators are robust and stable, and can produce a broad range of gas phase NO concentrations that can be used for INO or other applications (cardiopulmonary bypass surgery). By applying different constant currents, NO concentrations in the gas phase can be easily tuned in the 5–200 ppm range at flow rates up to 5 L/min, which is relevant for INO therapy and can potentially be a substitute for pressurized NO gas cylinders currently used in hospitals. For pulmonary delivery, the NO concentration can be monitored continuously with an amperometric NO sensor to control the NO in the therapeutic range of 5–100 ppm. The system using air as a carrier gas for NO suggested the possibility of developing a more portable and cost-effective INO device for use in the field or developing countries. The robust NO generators described herein can also serve as tools for fundamental biological/physiological studies that require a strictly controlled NO concentration in the gas phase, e.g., the effect of NO on airway related diseases, both in vitro and in vivo, could be investigated with our new generators. In this work, this new system has been adapted for the CPB application, wherein NO was added to the oxygenator sweep gas and/or the air used for cardiotomy suction. In this application, NO can alleviate significant complications arising from the activation of platelets, blood clot formation, and systemic inflammatory response during the open heart surgery and ECMO.