Improved Serotonin Measurement with Fast-Scan Cyclic Voltammetry: Mitigating Fouling by SSRIs

Abstract

Selective serotonin reuptake inhibitors (SSRIs) have been used for decades to treat disorders linked to serotonin dysregulation in the brain. Moreover, SSRIs are often used in studies aimed at measuring serotonin with fast-scan cyclic voltammetry (FSCV) in living tissues. Here, we show that three different SSRIs – fluoxetine, escitalopram, and sertraline – significantly diminish the faradaic oxidation current of serotonin when employing the commonly used Jackson waveform. Coating carbon-fiber microelectrodes (CFMs) with Nafion resulted in further degradation of peak current, increased response times, and decreased background charging currents compared to bare CFMs. To decrease fouling, we employed a recently published extended serotonin waveform, which scans to a maximum positive potential of +1.3 V, rather than +1.0 V used in the Jackson waveform. Use of this waveform with bare CFMs alleviated the decrease in faradaic current, indicating decreased electrode fouling. Collectively, our results suggest that fouling considerations are important when designing FSCV experiments that employ SSRIs and that they can be overcome by using the appropriate waveform.

Introduction

Serotonin (5-hydroxytryptamine) is a biogenic indoleamine that plays crucial roles in numerous physiological functions, including regulation of appetite and mood in the brain^1–3 and enhancing nutrient absorption and energy storage in the gut.^4–6 Despite ~90% of the body’s serotonin being distributed in enterochromaffin cells in the gut,⁷ serotonin’s roles as a neuromodulator in the brain, especially in mood and emotion regulation, have been a prominent area of research since the mid-20^th century.^{8, 9} Given that the dysfunction of serotonin dynamics in the brain correlates with anxiety and depression disorders,^9–14 the current first-line medication of choice for these disorders is oral administration of selective serotonin reuptake inhibitors (SSRIs), more commonly known as anti-depressants.^{15, 16} Although the long-term mechanisms of SSRIs and their therapeutic pathophysiology are still under investigation,^17–19 the primary target of interest of these drugs is the inhibition of the serotonin transporter (SERT), thereby elevating extracellular serotonin concentrations. Therefore, analytical methods that measure the release and reuptake dynamics of serotonin, as well as the effects of SSRIs on those dynamics, are useful for understanding the underlying mechanisms through which anxiety and depression disorders manifest.

Fast-scan cyclic voltammetry (FSCV) at carbon-fiber microelectrodes (CFMs) is one method that can measure the rapid dynamics of electroactive molecules in the brain, including serotonin. Wightman’s group made progress in measuring serotonin with FSCV by developing the “Jackson” waveform which scans from +0.2 V to +1.0 V to −0.1 V to +0.2 V at 1000 V/s with a +0.2 V holding potential.²⁰ This waveform was developed for several methodological reasons. First, serotonin produces nearly indistinguishable electrochemical signals from dopamine, another prominent neurotransmitter. Both molecules oxidize at approximately +0.6 V, while serotonin reduces at 0 V and dopamine reduces at −0.2 V when employing conventional fast-scan triangular waveforms.^20–22 However, the Jackson waveform demonstrated an enhanced selectivity of serotonin over dopamine at physiological concentrations. This is likely due to the Jackson waveform staying above the reduction potential of dopamine-o-quinone which prevents its recycling back to dopamine. Second, the oxidation of serotonin and its more prominent metabolite, 5-hydroxyindoleacetic acid (5-HIAA), produces radicals that rapidly polymerize on the electrode surface, thereby passivating the electrode surface.^{23, 24} The +0.2 V holding potential prevents adsorption of serotonin, and the 1000 V/s scan rate outruns the radical polymerization, therefore decreasing fouling. Lastly, electrodes were modified with Nafion, a perfluorinated cation-exchange polymer, to prevent the negatively charged 5-HIAA from reaching the electrode surface while allowing the positively charged serotonin to pass freely. It also increases sensitivity which is advantageous for measuring the low concentrations of serotonin present in brain tissue.^25–27

Several studies have been conducted to measure serotonin using these techniques, and some of these studies utilized SSRIs, such as citalopram and its isolated S-isomer escitalopram, to characterize their effects on serotonin release and reuptake dynamics.^{25, 28–32} As anticipated, the bulk of this research shows that administration of SSRIs increases the synaptic release of serotonin. However, recent findings suggest that the Jackson waveform is especially susceptible to fouling in comparison to other waveforms.²⁴ Additionally, to our knowledge, no investigation of the effects of SSRIs on electrochemical detection of serotonin using the established methods has been published in the primary literature. Understanding these effects is crucial for optimizing the accuracy of serotonin measurements, especially in tissues that contain concentrations at or below the nanomolar range.

Here, we analyze the effects of three common SSRIs – fluoxetine, escitalopram, and sertraline – on serotonin detection with FSCV at bare and Nafion CFMs using the Jackson waveform. These measurements were accomplished by developing an in vitro flow cell assay in which serotonin oxidation current was measured before, during, and after SSRI exposure in solution. Exposure of Nafion-coated CFMs to these SSRIs significantly decreased the peak serotonin oxidation current, while bare CFMs were only affected by fluoxetine and escitalopram. Washout of the SSRI did not fully restore serotonin signal. The use of the recently published extended serotonin waveform²⁴ helped mitigate fouling effects observed with all three SSRIs, indicating that it may be a useful alternative for measuring serotonin in these applications.

Experimental

Chemicals and reagents

Serotonin hydrochloride, 5-hydroxy-3-indoleacetic acid, Nafion perfluorinated resin (5 wt.%), fluoxetine hydrochloride, and sertraline hydrochloride were purchased from Sigma-Aldrich (St. Louis, MO). Escitalopram oxalate was purchased from Tocris Bioscience (Bristol, UK). Artificial cerebrospinal fluid (aCSF) was made up of 131 mM NaCl, 2 mM KCl, 1.25 mM KH₂PO₄, 20 mM NaHCO₃, 2 mM MgSO₄, 2.5 mM CaCl₂, and 10 mM HEPES in purified (18.2 MΩ) H₂O and pH was adjusted to 7.4. A 1.0 mM stock solution of serotonin and 5-HIAA was prepared in 0.2 M HClO₄, shielded from light, and remade after 3 days. A 10 mM stock solution of each SSRI was prepared in DMSO and stored at 4°C and shielded from light. Fresh solutions of 100 nM serotonin, 1 μm 5-HIAA and 10 μm SSRI were prepared the day of each experiment.

Carbon-fiber microelectrode preparation

CFMs were fabricated as previously described. Briefly, 7 μm diameter carbon fibers (Goodfellow Cambridge LTD, Huntingdon, UK) were aspirated into glass capillaries (1.2 mm D.D. and 0.68 mm I.D., 4 in long; A-M Systems, Inc., Carlsborg, WA) using a vacuum pump. The capillaries were then pulled using a PE-22 heated coil puller (Narishige Int. USA, East Meadow, NY) to form two electrodes. Exposed carbon fibers were trimmed to a length of 40–60 μm. CFMs were sealed by dipping in epoxy resin (EPON resin 815C and EPIKURE 3234 curing agent, Miller-Stephenson, Danbury, CT) for 45 seconds and cured at 100°C for 1 hr. Prior to experiments, electrodes were soaked in isopropanol for at least 10 min.

Nafion-coating procedure

Nafion polymer was electrodeposited onto the CFM surface as previously described. Briefly, CFMs were soaked in isopropanol for at least 30 min to thoroughly clean the surface. Immediately after soaking, they were dipped in Nafion resin and a constant +1.0 V potential against a Ag/AgCl reference electrode was applied for 60 seconds using a DC power supply (Shenzhen Korad Technology, Shenzhen, China). The coated electrodes were air-dried for 30 seconds before curing at 70°C for 10 minutes.

Electrochemical Instrumentation

All flow cell experiments were conducted in a grounded Faraday cage. The electrochemical system was comprised of a CFM working electrode backfilled with 0.5 M potassium acetate and a chlorided Ag/AgCl wire as a reference electrode. Potentials were applied to the working electrode using a ChemClamp potentiostat and headstage (Dagan, Minneapolis, MN). A syringe pump (New Era Pump Systems, Farmingdale, NY) was used to push aCSF through a six-port loop injector (Valco Instruments, Houston, TX) at a flow rate of 2 mL/min. The CFM tip was submerged in solution at the flow cell output. Data were collected and analyzed using TarHeel (Department of Chemistry, University of North Carolina at Chapel Hill) and Knowmad (Knowmad Technologies, LLC, Tuscon, AZ) software. All data were collected at a frequency of 10 Hz and low pass filtered at 3 kHz.

SSRI flow cell assay

A simple flow cell assay was developed to investigate the confounding effects of SSRIs on serotonin detection. CFMs were cycled at 60 Hz and 10 Hz, both for 10 min, to stabilize the background charging. A “pre-treatment” baseline signal was obtained by injecting a solution of 100 nM serotonin into the flow cell three separate times (5 seconds each). Next, the flow solution was changed so that it contained either 10 μm SSRI solution or fresh aCSF as a control, and the CFMs were cycled for 15 min. After cycling, the electrodes treated with SSRI were evaluated by injecting 100 nM serotonin and 10 μm SSRI (“post-treatment”) to assess serotonin signal. The control electrodes underwent the same assay steps, except that no SSRI was present in any solutions. Finally, CFMs were cycled for 5 min in aCSF, and a final triplicate injection of 100 nM serotonin (“washout”) was conducted. This assay was also conducted in combination with 5-HIAA to assess compounding fouling effects. In a separate set of measurements, we also assessed the effects of short-term SSRI exposure by injecting solutions containing 10 μm SSRI for 60 s (see Supplementary Figure 3). The post-treatment and washout signals were normalized to pre-treatment signals for each CFM.

We applied either the Jackson waveform or the Extended Serotonin Waveform to the electrodes. The Jackson waveform scans from +0.2 V to +1.0 V to −0.1 V to +0.2 V at a scan rate of 1000 V/s. The potential is held at +0.2 V between scans and is applied at a frequency of 10 Hz. The Extended Serotonin Waveform differs from the Jackson waveform in that the switching potential is extended to +1.3 V instead of +1.0 V; all other parameters are consistent with the Jackson waveform.

Data analysis

For each electrode, the post-treatment and washout oxidation currents obtained from the injection of serotonin were normalized against pre-treatment oxidation currents. These values are expressed as a percentage of the initial serotonin signal and were averaged across electrodes. The presentation and statistical analysis of data were carried out using Graphpad Prism version 7.03 (San Diego, CA). For statistical analyses, we used two-way ANOVA with Dunnett or Tukey’s post hoc analyses to determine overall main effects between electrode treatment types and significant differences between individual groups. For all data, an N-value of three to five electrodes was used and p < 0.05 was taken to indicate significance. Plotted values and error bars on all bar graphs indicate mean ± SEM.

The limit of detection (LOD) for serotonin was estimated using a ratio method as described in Dunham et al.²⁴ The LOD was calculated as 3 times the tested concentration (100 nM) divided by the signal-to-noise ratio of that tested concentration. To calculate noise, the standard deviation of baseline current was taken from 0–2 s (N = 20) in i vs t traces.

Results

SSRIs decrease serotonin oxidation current measured with Jackson waveform

The effects of treatment with SSRIs on electrode performance when using the Jackson waveform is shown in Fig. 1. The injection of serotonin against a naïve electrode yields a CV with an oxidation peak occurring between +0.55 to +0.65 V and a reduction peak occurring between −0.1 to 0 V vs Ag/AgCl (Fig. 1A). However, the currents noticeably decreased after 15 min of cycling in aCSF solution with 10 μM fluoxetine (Fig. 1B), indicating that this SSRI interferes with the electrochemical detection of serotonin.

Figure 1. SSRIs decrease redox current of serotonin in a flow cell when we use the Jackson waveform.

Representative color plots, i vs time and CVs show the change in serotonin redox current before and after addition of SSRI to solution. Data shown here were collected using Nafion CFMs and 10 μM fluoxetine. Plots of i vs time were extracted at E_p,a (horizontal dotted line) and CVs at average peak current (vertical dotted line). (A) A 5 second injection of 100 nM serotonin was conducted in triplicate to determine pre-treatment serotonin signal. (B) Addition of 10 μM SSRI to flow and injection solutions (post-treatment) significantly reduces serotonin signal.

To investigate this phenomenon in more depth, we employed a flow cell assay to evaluate the effects of SSRI exposure on electrode performance. For each electrode, we first injected serotonin to obtain reference CVs (‘pre-treatment’). We then treated electrodes with fluoxetine (FXT; Fig. 2A), escitalopram (ESP; Fig. 2B), or sertraline (STR; Fig. 2C) by allowing these SSRIs to flow over the surface for 15 min while the waveform was being applied. Next, we injected serotonin in the presence of the SSRI and obtained a series of CVs (‘post-treatment’). We then washed out the SSRIs and collected additional CVs of serotonin (‘washout’). Control electrodes were treated identically, except that no SSRI was used.

Figure 2. The effects of SSRIs on serotonin signal between bare and Nafion CFMs.

Representative cyclic voltammograms show the change in serotonin signal between pre-treatment, post-treatment, and washout steps in the assay. Bar graphs on the bottom show the normalized serotonin oxidation currents for post-treatment and washout assay steps with standard error of the mean (SEM) error bars for each group. All groups were normalized to their respective pre-treatment signals for each electrode. Cyclic voltammograms for FXT (A), ESP (B), and STR (C) all demonstrate a decrease in serotonin signal after the addition of SSRI with Nafion-coated CFMs being preferentially fouled compared to bare CFMs. (D) For bare CFMs, post-treatment measurements of serotonin caused a significant decrease in signal for FXT (N = 5, p < 0.001) and ESP (N = 5, p < 0.05), but not STR (N = 5, p = 0.69) compared to the control group (N = 4). No difference in the signal was observed after SSRI washout. (E) For Nafion-coated CFMs, all three SSRIs significantly decreased serotonin signal (N = 5 for all SSRIs; p <0.0001) compared to the control group (N = 4). Washout of SSRIs still resulted in a significantly decreased serotonin signal for FXT (p < 0.05), ESP (p < 0.05), and STR: (p < 0.001) compared to the control group (statistics: Two-way ANOVA, Dunnett’s post hoc).

For each electrode, the average post-treatment current was normalized against the average pre-treatment current. Thus, the values on the bar graph are expressed as average percentages ± SEM obtained from N = 4 or 5 electrodes. Treatment of bare electrodes (Fig. 2D) with FXT and ESP significantly decreased peak current responses versus control electrodes (FXT, 49.5 ± 2.7%, p < 0.001 and ESP, 69.8 ± 4.7%, p < 0.05 versus Control, 95.0 ± 3.0%). STR did not provide a significant effect. After washout, signals from SSRI-treated electrodes somewhat increased, but there was no significant main effect (p = 0.15). Additionally, no signals from SSRI-treated electrodes differed from those of the respective control electrodes. Treatment of Nafion-coated electrodes (Fig. 2E) with SSRIs resulted in more profound decreases in oxidation current (FXT, 16.7 ± 4.0%, ESP, 28.4 ± 4.2%, and STR, 14.8 ± 2.1%, p < 0.0001 versus Control, 92.0 ± 1.0%). There was a significant main effect of treatment washout, suggesting that the washout step partially restored the serotonin oxidation currents (p < 0.0001).

Fouling from repeated measurements of serotonin when using the Jackson waveform has been previously observed.²⁴ This decrease is caused, in part, by film formation on the electrode surface due to the polymerization of radical oxidation products of serotonin, as previously described. Fouling by serotonin polymerization may have also contributed to the decreases observed in the SSRI groups. However, SSRIs are likely responsible for the majority of the signal loss since there were significant differences between measurements from control and SSRI treated electrodes. The increase in signal arising from washout further supports this notion. However, the fact that, even after washout, the currents obtained with Nafion coated electrodes are significantly diminished suggests that this coating enhances electrode fouling. It is possible that these SSRIs are selectively partitioning into the Nafion coating and continuing to impede access of the serotonin to the electrode surface.

Effects of SSRI treatment on electrode response time

FSCV offers good temporal resolution, with an update rate of 10 CVs/s commonly used. This rate allows the measurement of sub-second changes in analyte concentration in real time in many biological systems. However, electrodes must have a sufficient response time, which we define generally as the time required for the oxidation current to approach its maximum, to accurately measure changes in extracellular neurotransmitter levels.

With this issue in mind, we analyzed data from the SSRI assay to examine the effects of SSRIs on CFM response time (t_res). Here, we defined t_res specifically as the time measured between 10 to 80% of the maximum oxidation current sampled at the anodic potential that provides the maximum currents on the CVs (E_p,a). Fig. 3A shows raw data with illustrative t_res values obtained from pre-treatment to post-treatment signals. Bare CFMs (Fig 3B) exhibited no significant post-treatment differences in t_res between control and SSRI groups, but significant decreases did occur after washout of FXT and ESP (FXT, 70.7 ± 8.2%, ESP, 74.9 ± 7.1%, p < 0.05 versus Control, 134.2 ± 31.3%). Post-treatment t_res values were generally lower than the control, and there was a significant main effect of treatment type (p < 0.01). For Nafion-coated CFMs, no significant differences in t_res were observed between control and SSRI groups for any step in the assay (Fig 3C). However, in this case, post-treatment t_res values were generally higher than the control, and there was a significant main effect of treatment type (p < 0.05).

Figure 3. The effects of SSRI exposure on electrode response time of bare and Nafion-coated CFMs.

Electrode response time (t_res) was determined by the time between 10% and 80% of i_max of the anodic peak current vs time curve. Bar graphs on the bottom show the normalized t_res values for post-treatment and washout assay steps with standard error of the mean (SEM) error bars for each group. All groups were normalized to their respective pre-treatment values. (A) Representative i vs t curves taken at anodic peak current for serotonin show minimal changes in t_res between pre-treatment (left) and post-treatment (right) measurements. (B) No significant differences in t_res were observed for bare CFMs between any SSRIs and the control group for post-treatment, but washout was significant for FXT (p < 0.05) and ESP (p < 0.05) compared to the control. (C) No significant differences were observed in Nafion CFMs for any SSRI for any category (statistics: Two-way ANOVA, Dunnett’s post hoc, N = 4 or 5 electrodes).

Overall, slight decreases in t_res were observed for bare CFMs after treatment with SSRIs. Although counterintuitive, this improvement in electrode response is fairly proportional to the respective decreases in serotonin oxidation current observed in Fig. 2D, which may indicate that electron transfer kinetics of bare CFMs are moderately preserved after treatment with SSRIs. However, this proportionality was not observed for Nafion CFMs, which generally showed slight increases in t_res after treatment despite substantial decreases in serotonin oxidation current. Direct comparison of Nafion CFMs to bare CFMs shows a significant increase in post-treatment (p < 0.05) and washout (p < 0.01) t_res with FXT (Supplementary Figure 1). Additionally, there was a significant main effect of electrode coating after washout, which indicates that Nafion CFMs generally underwent greater increases in t_res relative to bare CFMs (p < 0.05).

With a Nafion coating, analytes must diffuse through the polymer layer, and this can increase electrode response times.^{25, 33, 34} Indeed, the raw pre-treatment t_res values for Nafion CFMs (0.95 ± 0.07 s) were more sluggish compared to bare CFMs (0.81 ± 0.08 s). The post-treatment increase in t_res for Nafion CFMs may be explained by the aggregation of the SSRIs into the Nafion polymer matrix due to their high lipophilicity. This would effectively “clog” the Nafion which decreases diffusion and induces relative increases in response times as a result. Additional studies are needed to probe the mechanism for this phenomenon.

Effects of SSRI treatment on background charging current

The rapidly changing potentials associated with FSCV produce large, non-faradaic charging currents that are caused by the formation of electrical double layers (EDLs) at electrode surfaces. This background charging current is digitally subtracted to reveal the smaller faradaic currents that arise from analyte oxidation and reduction.^{35, 36} The magnitude of the background charging current is directly affected by the surface integrity of the electrode and its environment, thereby providing a proxy for assessing the fouling of the CFM surface by SSRIs. To investigate fouling by SSRIs, the background charging currents were recorded, CFMs were cycled in 10 μM SSRI solution for 15 min, and then the background charging current was recorded again. This procedure was used to compare bare and Nafion CFMs to their respective control groups.

Fig. 4A shows representative cyclic voltammograms of background current before and after SSRI exposure. Exposure of electrodes to the SSRIs for 15 minutes had a minimal effect on background current when bare CFMs were used. However, background currents decreased substantially when exposing Nafion CFMs to FXT (Fig. 4A) and the other SSRIs (Supplementary Figure 2). To quantify the change in background current, CVs taken after cycling in SSRI solutions were normalized against pre-treatment CVs and the median value was obtained. Fig. 4B shows the combined results for the change in background current. The background current from bare CFMs did not decrease after cycling in any of the SSRIs compared to the bare control. In contrast, treatment of Nafion-coated CFMs with these SSRIs resulted in significant decreases in background current (FXT, 64.5 ± 14.8%, ESP, 83.8 ± 3.0%, and STR, 80.9 ± 4.5%, p < 0.01 versus Control, 98.3 ± 1.8%).

Figure 4. Relative change in background current after exposure to SSRIs.

Background current was allowed to stabilize by cycling at 60 Hz and 10 Hz consecutively for 10 min each for each CFM. Background current was recorded before and after the CFM was cycled in 10 μM SSRI solution for 15 minutes. Control groups were cycled in aCSF for 15 minutes between recordings of the background current. (A) Representative CVs exhibit a decrease in the background current after cycling in SSRI solution for 15 minutes for both bare and Nafion CFMs. (B) Bar graph of the normalized background current with standard error of the mean (SEM) error bars. The values were obtained by normalizing the current from the post-treatment CV to the pre-treatment CV at every point in the waveform and taking the median value for each CFM. A significant decrease in background current was observed for Nafion CFMs for all three SSRIs (FXT: p < 0.0001; ESP: p < 0.01; STR: p < 0.001; N = 5 for all SSRIs) compared to the control group (N = 4). No significant differences were observed for bare CFMs for any SSRI compared to the control (statistics: Two-way ANOVA, Dunnett’s post hoc).

To understand the nature of these changes further, 2 mL injections of SSRI solution were tracked for 2 min (Supplementary Figure 3). For bare CFMs, injections of each SSRI caused an initial decrease in background current of ~0.5% followed by a slower decease until completion of injection. Clearance of the SSRIs led to a gradual increase toward baseline current until a new, decreased baseline was established. A similar trend, albeit more pronounced, occurred when using Nafion CFMs with two exceptions. Injection of FXT caused a rapid and consistent rate of decrease over the course of the injection until 30% reduction in background current was reached; this decrease was followed by an equally rapid, consistent increase toward baseline after clearance. The other exception occurred upon injection of STR, in which no return to baseline was observed after clearance. These results indicate that SSRIs induce gradual changes to the background current that can be partially restored, which correlates with the partial recovery of signal after the washout step of the assay.

Decreasing of the background current, both immediate and long-term, indicates a change in the electrode surface and its environment.^{37, 38} This is likely due to adsorption of the SSRIs to the CFM surface. Non-electroactive organic molecules that contain amine functionalities, such as the SSRIs tested here, are able to adsorb to carbon-fiber electrodes.³⁷ Additionally, all three SSRIs contain aromatic rings which could interact with the aromatic rings of the graphitic carbon by pi stacking, thereby leading to surface adsorption. The substantial decrease in background charging current with Nafion CFMs may be attributed to the aggregation of SSRIs in the polymer which could be enhancing adsorption as well as disrupting the EDL formation.

Extended Serotonin Waveform Mitigates SSRI effects

Collectively, the presented data indicate that SSRIs significantly inhibit the detection of serotonin when the Jackson waveform is applied. Recent work by Dunham et al.²⁴ investigated alternatives to the Jackson waveform that could improve resistance to fouling while maintaining selectivity and low limits of detection. One such waveform they tested was the extended serotonin waveform (ESW); a modified version of the Jackson waveform which has an increased switching potential of +1.3 V. This waveform was hypothesized to prevent fouling through regeneration of the carbon surface by breaking carbon-carbon bonds, which occurs above 1.0 V.³⁹ Indeed, the ESW exhibited moderate fouling resistance compared to the Jackson waveform while maintaining selectivity over dopamine. Given our hypothesis that SSRI adsorption is the root cause of the phenomena observed in this work, we chose the ESW as a candidate to prevent fouling by SSRIs.

We applied our SSRI assay, in which we employed the Jackson waveform, to bare CFMs using the ESW. Fig. 5A shows representative CVs of serotonin signal for each SSRI using the ESW. Although treatment with SSRIs decreased serotonin oxidation currents when using the Jackson waveform for bare and Nafion CFMs, we found no significant difference in oxidation current compared to when the ESW was used (Fig. 5B). Furthermore, washout of the SSRIs yielded no significant changes in signal compared to the post-treatment results (Fig. 5B). Additionally, no significant difference was observed in t_res (Fig. 5C) or background current (Fig. 5D). Thus, the ESW appeared to improve the detection of serotonin compared to the Jackson waveform when carrying out flow cell analysis (see Supplementary Fig. 4). Nevertheless, it will be important to validate its use when using SSRIs in model organisms.

Figure 5. The Extended Serotonin Waveform prevents the fouling effects of SSRIs.

Representative cyclic voltammograms show the change in serotonin signal between pre-treatment, post-treatment, and washout steps in the assay. Bar graphs on the bottom show the normalized serotonin oxidation currents for the post-treatment assay step for each SSRI and waveform/CFM surface with standard error of the mean (SEM) error bars for each group. All groups were normalized to their respective pre-treatment signals. (A) The fouling effects observed using the Jackson waveform (bare and Nafion CFMs) are not observed using the ESW. (B) No significant decrease in serotonin signal was observed post-treatment for all three SSRIs, and washout signals stayed consistent as well. (C). Post-treatment electrode response times increased relative to pre-treatment response for all SSRIs except for FXT using the ESW, but no SSRI differed from controls. (D) The background current generated by the ESW did not change after cycling in SSRIs for 15 minutes.

Interestingly, negative current peaks were consistently generated at the switching potential using the ESW (Fig. 5A). It is possible that these peaks are an artifact of increased adsorption of serotonin. Increasing the switching potential to 1.3 V was found to increase adsorption of catecholamines to CFMs³⁹. Nevertheless, these peaks did not overlap with and confound the oxidation peaks used for analysis.

Extended Serotonin Waveform Mitigates Fouling by 5-Hydroxyindoleacetic Acid

As previously stated, fouling by 5-HIAA, which exists at much higher concentrations than serotonin, is a major concern when applying these methods in brain tissues, and coating electrodes with Nafion is commonly used to prevent long-term fouling by 5-HIAA. The work in Dunham et al. demonstrated the propensity of long-term fouling by 5-HIAA using the Jackson waveform with bare CFMs. They also showed that the ESW was more resistant to this fouling, as has been demonstrated in this work with SSRIs.

To understand the utility of the ESW compared to the Jackson waveform in biological applications that use SSRIs, we applied our assay again and incorporated 5-HIAA into the treatment. In these experiments, CFMs were treated with either 1 μM 5-HIAA or a combination of 1 μM 5-HIAA and 10 μM FXT (5-HIAA + FXT). Fig. 6 shows normalized oxidation current of serotonin. Control groups for each waveform, previously shown in Fig. 2 and Fig. 5, were included for comparison. For bare CFMs (Fig. 6A), treatment with 5-HIAA slightly decreased the signal, and 5-HIAA + FXT led to a further decrease, but neither treatment significantly differed from controls; no significant differences were observed after washout. For Nafion CFMs (Fig. 6B), treatment with 5-HIAA led to a significant decrease in signal and the combination with FXT decimated the signal (5-HIAA, 55.9 ± 11.2%, p < 0.001; 5-HIAA + FXT, 5.9 ± 3.0%, p < 0.0001 versus Control, 92.0 ± 1.0%). Washout partially restored signal for both groups, although signals were still significantly decreased compared to controls. For ESW CFMs (Fig. 6C), treatment with 5-HIAA did not decrease the signal, which remained consistent with the control signal both post-treatment and after washout. Treatment with 5-HIAA + FXT slightly decreased the signal, but it was not significant, and washout fully restored the signal.

Figure 6. The Extended Serotonin Waveform mitigates fouling from 5-HIAA and FXT.

Bare graphs show the normalized serotonin oxidation currents for post-treatment and washout assay steps for 1 μM 5-HIAA and 1 μM 5-HIAA with 10 μM FXT for each waveform/CFM surface with standard error of the mean (SEM) error bars for each group. All groups were normalized to their respective pre-treatment signals. The control groups were incorporated using the relevant data shown in previous figures. (A) For bare CFMs, no significant differences in post-treatment or washout steps were observed between any treatment group and controls. (B) For Nafion CFMs, significant post-treatment differences were observed for 5-HIAA (N = 3, p < 0.001) and 5-HIAA + FXT (N = 3, p < 0.0001) compared to controls (N = 4). Washout signals remained significantly decreased for 5-HIAA (p < 0.05) and 5-HIAA + FXT (p < 0.001) compared to controls. There was a significant main effect of treatment type (p < 0.0001) and assay step (p < 0.01) as well. (C) For ESW CFMs, no significant differences in post-treatment or washout steps were observed between any treatment group and controls (statistics: Two-way ANOVA, Tukey’s post-hoc, N = 3 or 4 electrodes).

Effects of Fouling on Detection Limits

Due to low concentrations of serotonin in brain tissues, electrodes must possess a sufficiently low limit of detection (LOD). As the fouling in this work has shown to significantly decrease serotonin oxidation current, it is important to quantify the changes in the LOD due to treatment with SSRIs and 5-HIAA. Table 1 shows the LOD of serotonin for each waveform before and after treatment with each SSRI. Bare CFMs that used the Jackson waveform had the highest LOD (3.0 ± 1.1 nM) which increased less than 2-fold after treatment with SSRIs. Nafion CFM that used the Jackson waveform had the lowest LOD (2.0 ± 0.8 nM), which agrees with other studies that show enhanced sensitivity of Nafion-coated electrodes.^25–27 However, treatment with SSRIs led to a 4- to 10-fold increase in the LOD, which corresponds to the significant decreases in signal observed in Fig. 2E. The LOD using the ESW CFMs was in the middle of the other waveforms (2.6 ± 0.8 nM) and did not experience any changes after treatment with any of the SSRIs and produced the lowest post-treatment LODs out of all three waveforms.

Table 1:

LODs of serotonin before and after treatment with each SSRI and waveform.

Waveform	Average Pre-treatment LOD (nM) (N = 15)	Average Post-FXT LOD (nM) (N = 5)	Average Post-ESP LOD (nM) (N = 5)	Average Post-STR LOD (nM) (N = 5)
Jackson (bare)	3.0 ± 1.1	4.8 ± 1.0	5.1 ± 1.5	4.3 ± 1.6
Jackson (Nafion)	2.0 ± 0.8	16.4 ± 4.9	8.3 ± 4.4	19.1 ± 8.2
ESW (bare)	2.6 ± 0.8	2.4 ± 1.8	2.2 ± 0.7	2.3 ± 1.4

Table 2 shows the LODs after treatment of 5-HIAA and 5-HIAA + FXT. Treatment with 5-HIAA marginally increased the LODs for each waveform. The addition of FXT led to slight additional increases in LODs for bare and ESW CFMs, but Nafion CFMs experienced a 40-fold increase relative to the pre-treatment LOD. Overall, these results demonstrate the robustness of the ESW which produced the lowest LODs in the presence of SSRIs, 5-HIAA, and the combination of the two.

Table 2:

LODs of serotonin before and after treatment with 5-HIAA and 5-HIAA + FXT

Waveform	Average Pre-treatment LOD (nM) (N = 6)	Average Post-5-HIAA LOD (nM) (N = 3)	Average Post-5-HIAA + FXT LOD (nM) (N = 3)
Jackson (bare)	3.7 ± 1.1	4.3 ± 1.3	5.7 ± 1.5
Jackson (Nafion)	1.2 ± 0.7	1.9 ± 0.1	46 ± 43
ESW (bare)	1.6 ± 0.8	1.8 ± 0.7	2.2 ± 1.4

Discussion

This work has demonstrated the propensity for SSRIs to foul CFMs when using the Jackson waveform. Under certain conditions, this fouling may prevent accurate measurements of serotonin by decreasing signal intensity and electrode response over time of exposure. Nafion CFMs were especially susceptible to fouling with substantial decreases in serotonin oxidation current and electrode response depending on the SSRI used. Although Nafion is required to prevent fouling due to 5-HIAA²⁵, this work suggests that these benefits may be offset when SSRIs are used in combination with the Jackson waveform. However, the fouling observed by SSRIs and 5-HIAA was ultimately prevented by use of the ESW. The ESW also maintained consistently low LODs after treatment with SSRIs and 5-HIAA compared to the Jackson waveform. The extended +1.3 V switching potential is the most likely mechanism that decreases fouling as it would break carbon-carbon bonds and regenerate the electrode surface.

The mechanism of fouling with SSRIs is unclear. One possible explanation is adsorption of the SSRIs to the electrode surface. All three SSRIs used in this study contain aromatic rings which may facilitate pi-stacking interactions on the graphitic carbon surface. Yet, while FXT and ESP significantly decreased serotonin signal for bare CFMs, STR did not, which suggests other factors may contribute. The profound decrease in Nafion CFM performance suggests that additional mechanisms of fouling are associated with Nafion-coating. Indeed, despite the fouling resistance of the ESW CFMs, coating these electrodes with Nafion resulted in significant decreases in background current after treatment with FXT (Supplementary Figure 5). All three SSRIs are highly lipophilic, with volume of distribution (V_d) values of 20–45 L/kg,⁴⁰ 14–20 L/kg,^{40, 41} and 20 L/kg⁴⁰ for FXT, ESP, and STR, respectively. The perfluorinated backbone of Nafion polymer may facilitate accumulation of lipophilic small molecules like SSRIs in the polymer. This may “clog” the Nafion polymer matrix and prevent adequate mass transport to the electrode surface and disruption of the EDL, which would explain the poor performance of Nafion CFMs in this work. Future investigations could probe the mechanisms of these fouling phenomena using surface analysis techniques, such as AFM or SEM with EDX, to test our hypotheses.

Our results are particularly important for FSCV experiments that measure serotonin in combination with high concentrations of select SSRIs, including ex vivo applications such as brain slice experiments or other applications that may perfuse or repeatedly spritz concentrated SSRI solutions for extended periods of time. Regarding in vivo applications, SSRI concentrations in the extracellular brain matrix are likely at lower concentrations than those used in this work. Indeed, many in vivo studies using Nafion CFMs with the Jackson waveform have successfully measured increases in serotonin after administration of SSRIs.^{25, 28–32} It is important to note, however, that long-term administration of SSRIs has led to accumulation of SSRIs in brain tissues at concentrations greater than 10 μmol/dm³ in rats⁴² and humans^{43, 44} due to their high lipophilicity. It is unclear to what extent SSRIs could leach out and interact with implanted CFMs, so fouling by SSRIs may still be a practical consideration for in vivo work. Based on our results, the ESW may be more suitable than the Jackson waveform for serotonin applications that use SSRIs and require selectivity over dopamine as well as fouling resistance.

Conclusions

Here, we have demonstrated that SSRIs foul CFMs using the Jackson waveform, which prevents accurate and high-temporal resolution measurements of serotonin with FSCV. Although coating with Nafion is an important methodological requirement for measuring serotonin in brain tissue, it is especially prone to fouling by selected SSRIs. However, the ESW appeared to decrease the fouling effects of these SSRIs as well as 5-HIAA, suggesting that it may be a useful alternative to the Jackson waveform. Nevertheless, more work will be required to validate its use in model organisms treated with the SSRIs used in this work. Additionally, our work demonstrates the importance of examining the effects of other pharmacological agents on electrode response.