Solution-processed graphene films for electrochemical monitoring of extracellular nitric oxide released by breast cancer cells

Nitric oxide (NO) plays an important role in cardiovascular function, immune response, and intercellular signaling. However, due to its short lifetime, real-time detection of NO is challenging. Herein, an electrochemical sensor based on fibronectin-modified, solution-processed graphene ink for NO detection is developed using a facile fabrication method involving spin-coating and hot-plate annealing. The sensor is first electrochemically characterized with a NO donor, spermine NONOate, exhibiting a dynamic range of 10–1000 μM. The fibronectin-functionalized graphene supports the attachment and growth of MDA-MB-231 breast cancer cells, as confirmed by optical microscopy. Extracellular NO production is stimulated using the amino acid L-arginine. NO production results in morphological changes to the adhered cells, which are reversible upon the addition of the NO synthase antagonist Nω-nitro-L-arginine methyl ester. The production of NO is also confirmed using real-time amperometric measurements with the fibronectin-functionalized graphene sensors. While this work focuses on NO detection, this potentially scalable platform could be extended to other cell types with envisioned applications including the high-throughput evaluation of therapeutics and biocompatible coatings.


Introduction
Nitric oxide (NO) is an important small molecule produced from amino acid L-arginine by the nitric oxide synthase (NOS) enzyme family (Michel and Feron 1997).NO is involved in numerous physiological functions, including vasodilation as an endothelium-derived relaxation factor, immune system function, and neurological processes in the brain (Heales et al 1999, Habib and Ali 2011, Maiorana et al 2012, Džoljić et al 2015).Moreover, NO plays an important role in both promoting and inhibiting tumor growth, depending upon the concentration, timing, and location (Xu et al 2002, Korde Choudhari et al 2013).Relatively high levels of NO, particularly associated with macrophages (Macmicking et al 1997), can have tumoricidal effects (Garbán and Bonavida 1999).On the other hand, NO can cause oxidative damage to DNA (Derojas- Walker et al 1995), which could lead to deleterious mutations and tumor promotion (Korde Choudhari et al 2013).Furthermore, NOS expression has been reported in various cancer cell types, including breast cancer (Reveneau et al 1999, Vakkala et al 2000, Loibl et al 2002).For MDA-MB-231 triple negative breast cancer cells, NO induces the upregulation of NOS and cancer biomarkers, alters cell migration, and impacts the response of the MDA-MB-231 cells to chemotherapeutic agents (Heinecke et al 2014).Additionally, NOS expression in breast cancer cells promotes metastasis and has been correlated to poor prognosis and poor patient outcomes (Walsh et al 2016, Chung et al 2021).Given the importance of NO in health and life science, and especially its role in cancer, monitoring and quantifying NO concentration and dynamics is important for understanding disease progression and for the development of novel therapeutics.
Numerous methods have been reported for the detection of NO, including fluorescence measurement (Lim et al 2006), colorimetric assays (Sun et al 2003), electron paramagnetic resonance spectroscopy (Weaver et al 2005), and chemiluminescence measurement (Pinder et al 2008).Fluorescent probes, for example, are being used for in vitro and in vivo applications due to their high sensitivity and their ability to be used for spatial monitoring of NO, especially in biological media (Lim et al 2006).Although effective, many of these methods are endpoint measurements, which are not conducive to detecting NO in real-time.In contrast, electrochemical sensors have the potential to enable monitoring NO levels in realtime, especially if they are interfaced closely with the source of NO (Guo et al 2022).The ability to monitor changes in NO level continuously compared to periodically (e.g.cyclic voltammetry (CV)) can be advantageous given the short lifetime of NO and its rapid dynamics.Recent examples of electrochemical sensors for NO detection are based on noble metals (e.g.Au, Pt, Ag) and their nanostructures (Zhang and Oyama 2005, Quinton et al 2011, Thi Kim et al 2014, Govindhan and Chen 2016), two-dimensional (2D) materials (e.g.graphene, transition-metal dichalcogenides, MXenes) (Liu et al 2013, 2015, Yoon et al 2017), transition metal compounds/nanostructures (Shahid et al 2015, Wang et al 2021) and proteinfunctionalized electrodes (Wu et al 2010, Chen andZhao 2012).The 2D materials are especially interesting for electrochemical applications due to the large surface-to-volume ratio and in some cases, high electrical conductivity and biocompatibility (Bolotsky et al 2019, Nguyen et al 2020, Rohaizad et al 2021).In particular, graphene, a 2D allotrope of carbon, has shown remarkable properties (such as high conductivity and biocompatibility) and as a result, has seen widespread use in electrochemical sensing (Suvarnaphaet and Pechprasarn 2017).Numerous reports have shown graphene modified with gold (Ting et al 2013, Li et al 2015, Geetha Bai et al 2017), platinum (Mathew et al 2021), or bimetallic nanoparticles (Govindhan andChen 2016, Liu et al 2016).In an effort to move away from noble metal usage, other examples have demonstrated graphenebased NO sensors modified with peptides (Guo et al 2012), for example.Even sensors for NO based on graphene with varying degrees of oxidation have been reported (Liu et al 2013), although no measurement of extracellular NO has been reported in this particular example.Another report shows remarkably low detection limits and high sensitivities with hemin-functionalized graphene, along with the capability for real-time extracellular NO detection (Jiang et al 2013).Despite the exceptional performance, the mechanical exfoliation technique used to prepare the graphene limits the scalability of this technology.Alternatively, the ability to prepare graphene directly in solution has enabled its production on an increasingly larger scale, which is crucial for enabling future high-throughput and low-cost applications.
Herein, we explore solution-processed graphene ink for real-time electrochemical detection of extracellular NO secreted by human breast cancer cells that are grown directly on the graphene surface.The graphene films are prepared on silicon substrates using a simple two-step spin-coating and annealing process.By functionalizing the graphene surface with the extracellular matrix protein fibronectin, cells can attach and proliferate directly on the graphene film, which is in good agreement with previous reports (Kenry et al 2016).Close contact between the graphene and cell layer is crucial because of the short lifetime (one to tens of seconds, Conner and Grisham 1995, Thomas et al 2001, Habib and Ali 2011) of NO.The generation of extracellular NO is confirmed electrochemically using chronoamperometry measurements with L-arginine as an NOS protagonist.The generation of NO can be halted by the addition of Nω-nitro-L-arginine methyl ester (L-NAME), a NOS antagonist.Furthermore, optical microscopy is utilized to visually inspect the effect of NO production on cell morphology, which changes drastically after the addition of L-arginine and then reverts upon addition of L-NAME.While this work demonstrates the detection of NO generation, the biocompatibility and electrochemical properties of the solutionprocessed graphene films make the developed sensing chips a well-suited platform to investigate other systems, such as neurotransmitter release from neural cells and drug screening.

Materials and reagents
The graphene ink is purchased from Millipore-Sigma.The Spermine NONOate (SN) is purchased from Cayman Chemical and stored at −80 • C until use.Dulbecco's phosphate buffered saline (PBS, pH 7.4) and Dulbecco's Modified Eagle's Medium (DMEM) are purchased from Corning Inc. Larginine, L-NAME, (3-mercaptopropyl) trimethoxysilane (MPTS), and toluene (anhydrous) are purchased from Sigma-Aldrich.Fibronectin is purchased from Gibco.Fetal bovine serum (FBS) is purchased from Atlanta Biologicals and gentamicin is purchased from Gibco.Si/SiO 2 wafers are purchased from NOVA Wafers (285 nm, wet thermal oxide, p-type).Acetone and isopropanol (IPA) are purchased from VWR. Polydimethylsiloxane (PDMS, Sylgard 184) is prepared in a 10:1 ratio of elastomer to curing agent, degassed under vacuum, and cured at ∼60 • C in an oven.

Preparation of graphene ink films
The graphene films are prepared using a similar strategy as reported previously (Butler et al 2021).The 1 × 1 cm 2 Si/SiO 2 chips are ultrasonicated sequentially in acetone and IPA for 15 min each followed by rinsing in deionized water (Milli-Q, 18.2 MΩ cm) and drying under N 2 flow.The chips are then immersed overnight in MPTS (5% by vol. in toluene) at room temperature.Then, the chips are rinsed with toluene and dried.The 60 µl of graphene ink is spin-coated on each chip at 5000 rpm for 30 s.After spin-coating, the samples are annealed on a hot plate at 300 • C for 30 min under ambient conditions.

Device assembly
The spin-coated and processed graphene films are contacted using copper tape.To passivate the copper tape and ensure a consistent active sensor area, a piece of polyimide tape with a 4 mm diameter hole punched out is attached to the Si/SiO 2 chip.For measurements with cells, a PDMS well (8 mm diameter) is attached to the sensor using a thin layer of Ecoflex (Smooth-On Inc.) and cured at room temperature for at least 1 h.

Material characterization
Scanning electron microscopy images are acquired in a ThermoFisher Verios instrument at a working distance of 2.6 mm.The graphene films are not coated with any conductive layer for imaging and are grounded to the sample chuck using copper tape.X-ray photoelectron spectroscopy (XPS) measurements are carried out using a Physical Electronics VersaProbe II instrument (Chanhassen, MN).An Al Kα x-ray source is used at 45 • takeoff angle.Charge neutralization is carried out using low energy (<5 eV) electrons and Ar ions.Raman measurements are performed with a Horiba LabRAM instrument using a 532 nm laser and 300 gr mm −1 grating.The spectra are measured in 10 × 10 s acquisitions.Spectra of three distinct areas are measured and the mean and standard deviation values are calculated.

Electrochemical characterization
Electrochemical measurements are performed with either a PalmSens4 single-channel or MultiPalmSens4 multi-channel potentiostat.Ag/AgCl (3 M NaCl) is used as the reference electrode and Pt wire is used as the counter electrode.CV measurements are performed at a scan rate of 50 mV s −1 unless otherwise specified.Electrochemical impedance spectroscopy (EIS) measurements are performed at a DC potential of 0 V vs. the open-circuit potential and a small AC signal of 5 mV.Chronoamperometry measurements with SN in N 2 -purged PBS are performed in a C3 Cell Stand (Bioanalytical Systems Inc.) at a potential of 0.8 V vs. Ag/AgCl and a magnetic stirrer at 400 rpm.N 2 blankets the solution during the measurement.The current is allowed time to reach an approximately steady-state before adding SN.The SN solutions are held in a 37 • C water bath when not in use.

Cell culture
The human breast cancer cell line MDA-MB-231 (HTB-26), purchased from the American Type Culture Collection (ATCC), is used for this study and the cells are grown according to the recommendations of ATCC.MDA-MB-231 cells are maintained in DMEM supplemented with 10% FBS and 50 µg ml −1 gentamicin.Media is changed every 3 d, and the cells are grown in a 37 • C incubator in a humidified 5% CO 2 environment.

Fibronectin coating
The PDMS wells mounted on graphene ink are sterilized with 70% ethanol for 5 min and are washed with 1× PBS twice.Fibronectin solution in 1× PBS is pipetted onto the wells and incubated for 2 h at room temperature.Fibronectin concentrations of 10, 25, and 50 µg ml −1 are used to examine cell attachment to the graphene ink substrate.A fibronectin concentration of 25 µg ml −1 is used for coating the graphene for electrochemical assays and is kept constant for all the samples across different measurements.Fibronectin concentration was used as recommended by the manufacturer and is consistent with literature studies that have used fibronectin for cell adherence of MDA-MB-231 breast cancer cells (Maity et al 2011, Taherian et al 2011, Frahs et al 2019).

Seeding cells onto the graphene film
The graphene ink samples coated with fibronectin are washed with 1× PBS twice to remove excess fibronectin and 200 µl of cell culture growth medium is added into the wells.MDA-MB-231 cells are seeded at a density of 10 000 cells per well (50 000 cells ml −1 ).The devices are then placed in the incubator at 37 • C to allow for the cells to attach to the fibronectincoated graphene surface.

Treatments with L-arginine and L-NAME
For electrochemical assays, 72 h post-seeding, cells are taken out of the incubator, and media is aspirated from the wells.The cells are washed with 1× PBS twice before adding fresh 1× PBS.The counter and reference electrodes are inserted into each well and connections are made to the potentiostat.Stock concentrations of 50 mg ml −1 of L-arginine and L-NAME are made in water.After achieving the baseline reading, cells are treated with 4 mM L-arginine or a combination of 4 mM L-arginine and 8 mM L-NAME.The control samples without the cells containing only 1× PBS are also treated with the same concentrations.The electrochemical assay is then performed for 1 h.

Morphological assays
For morphological studies, cells are seeded on fibronectin-coated glass substrates and fibronectincoated graphene ink-coated glass substrates.The 24or 72 h post-seeding, cells are taken out of the incubator, and media is aspirated.The cells are washed with 1× PBS twice before adding fresh 1× PBS.Cells are then treated with 4 mM L-arginine or a combination of 4 mM L-arginine and 8 mM L-NAME and incubated for 2 h.Cells are imaged with a 20× objective on a Nikon Eclipse Ti-E inverted fluorescence microscope equipped with a Photometrics CoolSNAP HQ2 camera.Imaging conditions and settings were kept constant for all sample treatments.Cell spread area and cell aspect ratio are determined from phase contrast images using the 'shape descriptors' measurement tool in ImageJ software.Data are reported for at least 170 cells for each experimental condition.An analysis of variance followed by Tukey's post hoc test was performed using Minitab software for sample comparison.Differences between the means of the cell spread area and the aspect ratio for various treatment conditions were considered significant for p-values less than 0.05.

Live/dead assay
Assessment of cell viability following treatments with L-arginine and the combination of L-arginine and L-NAME was performed using a ThermoFisher LIVE/DEAD viability/cytotoxicity kit as per the manufacturer's recommended protocol.Briefly, following treatment of cells with the stimulants for 2 h, the cells were washed with 1× PBS and treated with 2 µM of calcein AM and 4 µM ethidium homodimer-1 for 30 min at room temperature.The treated samples were then visualized using fluorescence microscopy.Cells are imaged with a 10× or 20× objective on a Nikon Eclipse Ti-E inverted fluorescence microscope equipped with a Photometrics CoolSNAP HQ2 camera.

Preparation and characterization of fibronectin-modified graphene ink films
Preparation of the graphene films begins by spincoating a commercially available graphene ink onto a SiO 2 /Si wafer.The samples are then annealed at 300 • C for 30 min on a hot-plate to remove any remaining solvent and decompose the stabilizing polymer in the ink, hence improving the electrical conductivity (Secor et al 2015).The resulting graphene film is comprised of individual graphene flakes approximately 100-400 nm in lateral size as determined by atomic force microscopy (figure 1(a)).Raman spectroscopy (figure 1(b)) confirms the presence of the characteristic Raman bands for graphitic carbon species.A comparison of the D peak intensity, which is attributable to defects present in the graphene lattice, to the G peak can provide an estimate of the defect density in the graphene film (Ferrari andRobertson 2004, Cançado et al 2011).A ratio (I D /I G ) of 0.39 ± 0.01 is indicative of a moderate defect density, noting that some amount of defects can be advantageous for electrochemical applications due to significantly faster electron transfer at defect sites (Zhong et al 2014).
In addition to Raman spectroscopy, XPS measurements are performed to elucidate the elemental composition of the samples.Figure 1(c) shows a representative survey spectrum with carbon, oxygen, and silicon peaks present with atomic percentages of 89.3 ± 1.3, 8.1 ± 0.8, and 2.6 ± 0.5 (n = 3), respectively.While carbon and oxygen are to be expected for graphene, the presence of silicon is attributed to the underlying substrate.Figure 1(d) provides a high-resolution spectrum of the C 1s region.The main sp 2 carbon feature at ∼284.4 eV is fitted using an asymmetric peak (Kovtun et al 2019) with a line shape determined by measuring freshly cleaved highly-ordered pyrolytic graphite (pure sp 2 carbon).The tail of the asymmetric peak has been shown to provide a better quantification of the oxygen functional groups present on the graphene surface (Kovtun et al 2019).After fitting the sp 2 feature, smaller symmetric peaks corresponding to sp 3 defects and various oxygen functional groups are fitted: sp 3 (285.0eV), C-OH (285.6 eV), C-O-C (286.8 eV), C=O (288.0 eV), and O-C=O (289.0 eV).Lastly, two broad peaks at binding energies 6.4 and 10.1 eV higher than the main sp 2 feature are attributed to plasmon/shake-up contributions (Leiro et al 2003, Kovtun et al 2019).
In order to promote cell attachment, the graphene film is functionalized with fibronectin (Quinton et al 2011), which is an extracellular matrix protein.
Electrochemical characterization before and after the fibronectin treatment is undertaken to confirm the successful functionalization of the graphene film (figure 2) without completely hindering its electrochemical activity.CV with 5 mM [Fe(CN) 6 ] 4− shows the characteristic oxidation and reduction peaks for exemplary samples with and without fibronectin (figure 2(a)).After fibronectin modification, the electron transfer is more sluggish, as indicated by the suppression and increased separation of the redox peaks, with ∆E peak increasing from 209.5 ± 39.3 mV to 346.8 ± 23.8 mV (n = 4).Moreover, the peak ratio (I ox /I red ) increases from 1.07 ± 0.07 to 1.66 ± 0.31, noting an ideal value of 1. EIS measurements are also carried out to complement the CV analysis.EIS probes the interfacial charge transfer at the electrode surface by applying a small (5 mV) AC signal of varying frequency centered at the open circuit potential (Bard et al 2022).The exemplary Nyquist plots shown in figure 2(c) are modeled using a modified Randles circuit (SI figure S1), whereby the doublelayer capacitance term (C dl ) is replaced by a constant phase element (Q CPE ) to account for heterogeneities at the electrode surface (Cördoba-Torres et al 2015).The solution resistance (R s ) is consistent for samples with and without fibronectin modification (figure 2(d)), which is expected as the electrolyte is the same.The charge transfer resistance (R ct ), an indication of the facility of the redox reaction, is found to increase from 9.19 ± 4.31 kΩ to 36.99 ± 10.86 kΩ after fibronectin treatment, which is consistent with the CV data.Surface modification with a protein layer commonly leads to an increased charge transfer resistance as the proteins effectively slow or 'block' the electron transfer (Tolba et al 2012).Overall, the more sluggish electron transfer kinetics seen after fibronectin modification allude to the successful functionalization of the graphene, which is important for efficient cell attachment and growth on the electrode surface.

Electrochemical characterization with synthetic NO
After preparing the graphene films, the electrochemical response is benchmarked with synthetic NO.To do so, we employ a NO donor, SN, to prepare serial aliquots of NO in a N 2 -purged PBS solution.SN is chosen due to its previously characterized decomposition kinetics (Ramamurthi and Lewis 1997), where the authors reported a maximum NO concentration that was approximately ten-fold less than the initial SN concentration and occurred after ∼10 min.This timeframe corresponds well to experimental timeframes, whereas other NONOates, such as diethylamine NONOate and diethylenetriamine NONOate, have very rapid or lengthy decomposition kinetics, respectively.First, we perform CV measurements at a scan rate of 50 mV s −1 with a range of SN concentrations to characterize the oxidation of NO (figure 3(a)).From the CV data, a concentrationdependent oxidation wave can be seen at ∼0.9 V vs. Ag/AgCl, which is in agreement with previous reports of electrochemical NO  In addition to CV measurements, chronoamperometry at a constant potential of 0.8 V vs. Ag/AgCl is performed to characterize the sensor response in real time.Figure 3(c) shows the current-time data after a 10 min stabilization period for SN concentrations from 67 nm to 36 µM.The lower panel of figure 3(c) provides a closer look at early time points where small undulations in current can be seen for SN concentrations as low as 0.13 µM.Well-defined steps in current can be seen at a SN concentration of 0.73 µM, noting the actual NO concentration will be lower (Ramamurthi andLewis 1997, Majumder et al 2014).To determine the response time of the graphene sensor, we focus on the well-defined step after adding 6.4 µM of SN.Based on these results, we estimate a response time at 95% of the resulting current to be 17.7 s.While the response time is not as short as other reports (Xu et al 2014), perhaps due to the reduced electron transfer rate caused by the presence of fibronectin, the amperometric measurements indicate that the sensor responds reasonably well in real-time at nM-level concentrations of NO.Previous reports have implicated in vitro and in vivo levels of NO to be in the range of pM to µM depending on the cell type, tissue type, expression of NOS isoforms, experimental design, and measurement approach (Nagano 1999, He and Liu 2001, Mocellin et al 2007, Hall and Garthwaite 2009, Liu et al 2017, Dou et al 2019).A comparison of this work to related graphene-based NO sensors is provided in table 1.

Morphological characterization of cells grown on fibronectin-modified graphene sensors
In vitro monitoring and detection of extracellular metabolites and byproducts is crucial for enabling high-throughput screening of therapeutics and a better understanding of various biophysical phenomena.To this end, we characterize the generation of NO in the MDA-MB-231 human breast cancer cell line, which can be stimulated with L-arginine to release NO (Marullo et al 2021).Cell attachment on graphene ink-coated substrates was monitored as a function of fibronectin concentration (10, 25, 50 µg ml −1 ) used for coating the graphene surface.The number of attached cells per mm 2 of the substrate surface increases with increasing fibronectin concentration (SI figure S2).For electrochemical assays, graphene substrates were coated with 25 µg ml −1 fibronectin and cell morphologies were observed using phase contrast microscopy at 24 and 72 h post-seeding.While cells attach to the surfaces by 24 h post-seeding, the cells require more time to spread onto the surface (SI figure S3).Indeed, cells cultured on fibronectin-coated glass and graphene surfaces for 72 h exhibit increased spreading on the substrate surface as shown in microscopy images in figure 4(a) and as quantified in terms of a projected cell area in figure 4(b).Furthermore, the cells have an elongated morphology, which is characteristic of mesenchymal-like migratory cells, and is quantified by cell aspect ratio (figure 4(c)).We also analyzed the viability of cells in response to treatment with L-arginine or a combination of Larginine and L-NAME by performing a Live/Dead Assay.The presence of intracellular esterase activity in live cells converts calcein AM to calcein indicated by green fluorescence for live cells.On the other hand, ethidium homodimer-1 is absorbed by cells with a damaged cell membrane and it then binds to nucleic acid, producing a red fluorescence stain for dead cells.As shown in fluorescence microscopy images in figures 4(d), a large majority of cells treated with L-arginine or a combination of L-arginine and L-NAME for 2 h show green calcein staining and are thus viable.The percentage of live cells is higher than 90% for cells seeded on fibronectin-coated glass.Furthermore, approximately 80% of cells seeded on fibronectin-coated graphene substrates are viable both with and without treatment with the stimulants.

Electrochemical detection of extracellular NO with fibronectin-modified graphene sensors
After confirming successful attachment of cells to the fibronectin-modified graphene, amperometric measurements are carried out to detect extracellular NO from cells stimulated with L-arginine.A schematic of the test setup is shown in figure 5(a), in addition to an image of the test setup shown in figure S4 of the supplementary information.A PDMS well is attached to the sensor to house the cells and media during culturing.Before testing, the media is replaced with PBS to serve as the supporting electrolyte.Figure 5(b) shows the current-time profiles for four different experimental conditions (replicate experiments shown in supplementary figure S5).Like before, the cells are seeded on the sensor 72 h before testing to ensure there is enough time for cells to attach to the sensor surface.Allowing the cells to attach to the sensor surface helps minimize the cell-sensor distance, which is critical for detecting short-lived species, such as NO.All four experimental conditions show a sharp spike upon the addition of the L-arginine or L-arginine/L-NAME, which could be due to acute changes in the double-layer capacitance.For the cells with both L-arginine and L-NAME (red curve), the current quickly falls back to the baseline.However, for the cells with only L-arginine (grey curve), a broad domelike feature and elevated current is seen after the initial spike.A similar feature has been observed in previous amperometric measurements with cells (Burmeister et al 2002, Walker et al 2007) and has been attributed to the release of multiple species with different kinetic behavior.While a similar feature is seen upon the addition of L-arginine in the absence of cells (blue curve), indicating the graphene electrode responds to L-arginine alone, the magnitude of the current increase is not as large as when cells are present.It is possible that the current increase seen after adding Larginine to the cells contains contributions from both the L-arginine and the generated extracellular NO.

Conclusion
This work demonstrates a solution-processed graphene ink film as a sensor for electrochemical detection of NO.By functionalizing the graphene film with a fibronectin extracellular matrix protein, human breast cancer cells are able to attach directly to the electrode surface, as confirmed by optical microscopy.Minimizing the cell-sensor spacing is important for NO detection due to its short lifetime in solution.With cells attached to the electrode surface, NO is detected in real-time using L-arginine to stimulate its generation.Profound changes in cell morphology are also observed upon adding L-arginine to the cell environment.We believe the acute elevation of NO levels in the surrounding environment has a deleterious effect on the cells and results in reduced spreading on the graphene surface.These effects are reversed after adding L-NAME, which quenches the generation of NO, further supporting this speculation.
Adapting this work to a future cardiovascular point-of-care device would require a number of changes and improvements.For instance, a pointof-care device that analyzes physiological samples would require improvements to sensor selectivity.This can be accomplished for example by coating the graphene with a NO-selective membrane (Thangavel andRamaraj 2008, Shim et al 2010) or reducing the NO oxidation potential to minimize the contribution from interfering species, such as L-arginine.Moreover, a point-of-care device would require a more compact electrochemical cell, ideally with the reference, counter, and working electrodes all on-chip with a small footprint.While we did not demonstrate such a setup in this work, we have shown in a previous report an all-ink, on-chip electrochemical sensor for dopamine (Butler et al 2021).Furthermore, a point-of-care device based on the mechanism reported here would require specific handling and storage when being used in the field and maintaining the sterility/cleanliness of the sensor would also be critical.
Our target in this work was ex vivo application in order to study how cells respond to external biochemical stimulation.Ex vivo cell experiments have great importance for drug development, disease modeling, and mechanistic characterization.Having the ability to monitor NO release in real-time, as opposed to periodically or using an endpoint assay, is crucial for these applications.NO levels can change quickly, and these dynamics may not be captured by periodic or endpoint measurements.However, even excessive NO levels for a short time can have adverse effects on cells and having the ability to detect the entirety of these fluctuations could provide better insight into underlying phenomena.Although this work focused on detecting NO, the demonstrated ability of these graphene films to support cell attachment makes them a potential platform to study other cell systems.For example, the affinity of graphene for dopamine could be useful for quantifying its concentrations in cultures of neural cells.Furthermore, the ease with which solution-processed graphene films can be deposited on various substrates, especially flexible and stretchable materials, could make them suitable for studying cells under mechanical stress.

Figure 1 .
Figure 1.Material characterization of graphene films.(a) An exemplary atomic force micrograph of the spin-coated graphene film.(b) Raman spectrum of the graphene film (λ = 532 nm).Peak ratios are determined from spectra recorded at 3 distinct locations on the same sample.(c) An x-ray photoelectron spectroscopy (XPS) survey scan demonstrating the elemental composition of the graphene film.(d) A high-resolution XPS scan of the carbon 1s region and corresponding peak fitting of the various components.

Figure 2 .
Figure 2. Modification of the graphene film with fibronectin.(a) Cyclic voltammograms from −0.6 V to 0.8 V with 5 mM [Fe(CN)6] 4− in PBS before and after fibronectin treatment (scan rate = 50 mV s −1 ).(b) The resulting peak separation (∆E peak ) and peak ratio (Iox/I red ) as determined from CV measurements (n = 4).(c) Electrochemical impedance spectroscopy measurements before and after fibronectin modification (V dc = open circuit potential, Vac = 5 mV).(d) The extracted charge transfer and solution resistance from EIS measurements (n = 4).
sensors (Shahid et al 2015, Liu et al 2016).The resulting calibration curve, ∆I I0 = I−I0 I0 × 100 where I is the current at a particular SN concentration and I 0 is the baseline current in the absence of SN, measured at 0.96 V vs. Ag/AgCl is shown in figure 3(b).For the region from 10 to 1000 µM SN, the fitting equation (equation (1), R 2 = 0.9998) is,

Figure 3 .
Figure 3. Electrochemical characterization with NO donor, spermine NONOate (SN).(a) Exemplary CV scans from 0 V to 1 V with increasing concentrations of SN (scan rate = 50 mV s −1 ) and (b) the resulting calibration curve (n = 4).The inset highlights the dynamic range of our sensor along with the linear fit.(c) Chronoamperometric measurements (E = 0.8 V vs Ag/AgCl, stir rate = 400 rpm) across a range of SN concentrations with the bottom panel focusing on the low concentration region.

FET
= field-effect transistor, DEA = diethylamine, LPS = lipopolysaccharide, PB = phosphate buffer, PBS = phosphate buffered saline, NP = nanoparticle, IL = ionic liquid, rGO = reduced graphene oxide, erGO = electrochemically-reduced graphene oxide, RGD = Arginylglycylaspartic acid, SN = spermine NONOate.To stimulate NO generation, L-arginine is added to the cells cultured on fibronectin-coated graphene, which promotes NO production enzymatically via NOS(Privett et al 2010).A previous study demonstrated that 4 mM L-arginine promotes an increase in NO generation in MCF7 breast cancer cells(Dou et al 2019).For our studies, 2 h after stimulation with 4 mM L-arginine, the cell morphology appears much more spherical, possibly due to the cells contracting in the presence of excess NO in the surrounding environment.Treatment with a NOS inhibitor can block L-arginine induced NO generation in MDA-MB-231 cells (Marullo et al 2021).To quench the NO generation, cells were treated with both L-arginine and L-NAME.After 2 h of stimulation, the inhibition of NO appears to promote a more well-attached and spread morphology.Figures4(b) and (c) show the cell spread area and cell aspect ratio under different experimental conditions.Cells seeded on graphene and treated with L-arginine exhibited significantly lower cell spread areas and aspect ratios as compared to when cells were cultured on glass, graphene, and graphene with a combination of L-arginine and L-NAME treatments.

Figure 4 .
Figure 4. Characterization of cell attachment on graphene films coated with 25 µg ml −1 fibronectin.(a) Phase contrast microscopy images of MDA-MB-231 cells cultured for 72 h on glass, graphene, graphene after addition of 4 mM L-arginine, and after addition of 4 mM L-arginine and 8 mM L-NAME.Scale bars: 50 µm.Quantification of (b) cell spread area and (c) cell aspect ratio (elongation) for different treatment conditions.At least 170 cells were quantified for each treatment condition and the mean value is shown by the bold line.* * * p < 0.001.(d) Fluorescence microscopy images from a Live (green)/Dead (red) assay for MDA-MB-231 cells seeded for 72 h on fibronectin-coated glass and graphene substrates then treated with L-arginine or L-arginine/L-NAME for 2 h.The percentage of viable cells for each treatment condition is indicated on the images.Scale bars: 50 µm.

Figure 5 .
Figure 5. Real-time electrochemical monitoring of extracellular NO.(a) A schematic of the sensor with cells cultured directly on-chip.(b) Chronoamperometric measurements at E = 0.8 V vs. Ag/AgCl with cells (MDA-MB-231) stimulated by L-arginine (L-Arg) or a mixture of L-Arg and L-NAME.

Table 1 .
Comparison of graphene-based electrochemical NO sensors.