Progress toward the development of a microchip electrophoresis separation-based sensor with electrochemical detection for on-line in vivo monitoring of catecholamines

Abstract

The development of a separation-based sensor for catecholamines based on microdialysis (MD) coupled to microchip electrophoresis (ME) with electrochemical (EC) detection is described. The device consists of a pyrolyzed photoresist film working electrode and a poly(dimethylsiloxane) microchip with a flow-gated sample injection interface. The chip was partially reversibly sealed to the glass substrate by selectively exposing only the top section of the chip to plasma. This partially reversible chip/electrode integration process not only allows the reuse of the working electrode but also greatly enhanced the reproducibility of electrode alignment with the separation channel. The developed MD-ME-EC system was then tested using L-DOPA, 3-O-MD, HVA, DOPAC, and dopamine standards, which were separated in less than 100 seconds using a background electrolyte consisting of 15 mM sodium phosphate (pH 7.4), 15 mM sodium dodecyl sulphate, and 2.5 mM boric acid. A potential of +1.0 V vs. Ag/AgCl was used for amperometric detection of the analytes. The device was evaluated for on-line monitoring of the conversion of L-DOPA to dopamine in vitro and for monitoring dopamine release in an anesthetized rat in vivo following high K⁺ stimulation. The system was able to detect stimulated dopamine release in vivo but not endogenous levels of dopamine.

Graphical Abstract

Development of an easily fabricated MD-ME-EC set up for continuous on-line in vivo monitoring of neuroactive compounds in rat brain including dopamine.

Introduction

The continuous monitoring of neuroactive compounds in awake, freely roaming animals is critical for understanding neurological and behavioral disorders as well as for developing and evaluating the efficacy of drugs that are used to treat these conditions. When elucidating the neurochemistry of behavior-related disorders, the ability to correlate changes in concentration of neuroactive compounds with behavior is advantageous for a broader understanding of the disease. Most current methods employed for correlating behavior with concentrations of neuroactive compounds involve the collection of off-line microdialysis samples from freely moving or tethered animals in constrained environments such as a Raturn^®.^1–4 Off-line sample analysis provides some benefits for monitoring neurotransmitter release, including the freedom to use any relevant analytical instrumentation for the analysis. However, in some cases, temporal resolution can be compromised due to the high volume injection requirements (1–10 μL) of most conventional separation techniques. Therefore, with the aim of preserving the temporal information, both liquid chromatography and capillary electrophoresis techniques have been integrated with microdialysis sampling for on-line monitoring of neuroactive compounds in the brain.^4–9 However, due to the bulky hardware and complex mechanisms of operation, employing these techniques for on-animal behavior-related studies is not very feasible. On the other hand, a more compact field-portable microchip-based system can be employed on-animal for continuous in vivo monitoring while the animal is freely roaming and behaving in a less constrained environment.¹⁰ In this regard, methods that are capable of on-animal sensing to monitor concentration changes of neuroactive compounds would be more beneficial to understanding the neurochemistry of behavioral disorders. The ideal on-animal sensor must be one that can be operated wirelessly, is inexpensive, causes minimal discomfort to the animal, and can be employed for continuous real-time on-line measurements of multiple neurotransmitters simultaneously.

In comparison to other currently available separation techniques, microchip electrophoresis fulfills most of the aforementioned requirements for an on-animal separation-based sensor. Microchips are small and planar, making them easy to integrate into a portable analysis system. As a result of the planar separation platform and the very small sample (pL-nL)¹¹ and reservoir volume (μL) requirements, microchip electrophoresis devices can be directly integrated with microdialysis sampling.^{12, 13} In addition, microchip electrophoresis separations are generally faster than other separation methods such as liquid chromatography and capillary electrophoresis. Therefore, once coupled with continuous microdialysis sampling, microchip electrophoresis separation-based sensors can be used to obtain near real-time dynamic information regarding the chemical composition of the sample.

Laser-induced fluorescence (LIF) detection is the most common detection method for microchip electrophoresis-based analysis.^{11, 14} However, very few small neuroactive compounds exhibit natural fluorescence. As a result, most assays for neurotransmitters include derivatization with a fluorophore prior to LIF detection.¹⁵ Integration of a sample-derivatization step into an on-line separation system can increase the lag time and adversely affect the temporal resolution. Electrochemical detection has several advantages for the detection of neuroactive compounds separated by microchip electrophoresis. First, many important neuroactive compounds, including L-DOPA (l-3,4-dihydroxyphenylalanine), dopamine, adenosine, and serotonin, are naturally electroactive and can be directly detected using amperometric detection.¹⁶ In addition, electrochemical detection can be directly integrated with microchip electrophoresis separation devices without the use of expensive and bulky optics or electronics.^17–20 Lastly, microdialysis sampling (MD) combined with microchip electrophoresis (ME) and electrochemical detection (EC) opens the possibility of developing field-portable sensors that can be employed for continuous on-animal monitoring of neuroactive compounds in the brain.¹⁰

Previously, our group developed an on-animal MD-ME-EC-based separation-sensor that can be remotely operated to monitor drug metabolism in freely roaming sheep.¹⁰ In a proof-of-principle study, the authors successfully demonstrated the ability of the sensor to monitor nitrite production in sheep following subcutaneous perfusion of nitroglycerin through a linear microdialysis probe.¹⁰ An all-glass microchip device integrated with flow-gated sample injection and a platinum working electrode was employed for amperometric detection in these studies. For the detection of catecholamines, carbon-based electrodes provide better electrochemical performance compared to metal electrodes.²¹ However, carbon electrodes have limited stability under the high temperature and pressure conditions used for bonding of all-glass devices. The fabrication of a poly(dimethylsiloxane) (PDMS)/glass MD-ME-EC hybrid device containing a pyrolyzed photoresist film (PPF) working electrode was recently described by Saylor et al.²² This device was fabricated by bonding a plasma-activated PDMS layer containing the separation channel with the electrode containing glass plate to obtain a permanent MD-ME interface.

Accurate positioning of the detector electrode with the separation channel plays a critical role for obtaining a reproducible signal-to-noise ratio in ME-EC as well as maintaining good separation efficiency. One of the major limitations of the plasma-activated chip-substrate bonding protocol described above is that it requires quick (in less than 2 min)and precise alignment of the electrode with the channel prior to bonding. This can lead to a high rate of chip fabrication failure and irreproducible electrode-channel alignment. In this paper, we describe the development and evaluation of a simplified and robust PDMS/glass device fabrication method to integrate a PPF electrode for electrochemical detection of catecholamines combined with microdialysis sampling. The new approach makes it possible to readjust the position of the electrode in the microchip multiple times to obtain the optimum channel-electrode alignment. The new system was first demonstrated for near real-time on-line monitoring of L-DOPA metabolism in vitro. The system was then used for the first time to monitor endogenous dopamine release in vivo following high K⁺ stimulation.

Materials and methods

Materials

The following chemicals were purchased from the designated sources and used as received: 3,4-dihydroxy-L-phenylalanine (L-DOPA), homovanillic acid (HVA), 3,4-dihydroxyphenylacetic acid (DOPAC), dopamine hydrochloride (DA), 3-methoxytyramine hydrochloride (3-MT), 3-O-methyldopa (3-O-MD), boric acid, sodium phosphate monobasic, sodium phosphate dibasic, potassium phosphate mono basic and potassium phosphate dibasic (Sigma Aldrich, St. Louis, MO, USA) and NaOH and 2-propanol (IPA) (Fisher Scientific, Fairlawn, NJ, USA); sodium dodecyl sulfate (SDS) (Thermo Scientific, Waltham, MA, USA); AZ 1518 positive photoresist and AZ 300 MIF developer (AZ Electronic Materials, Somerville, NJ, USA); SU-8 10 and SU-8 developer (Micro-Chem, Newton, MA, USA); and poly(dimethyl siloxane) and curing agent (Sylgard 184 silicon elastomer base and curing agent, Dow Corning Corp., Midland, MI, USA). Additionally, the following materials were used for electrode preparation: quartz glass plates (4 in × 2.5 in × 0.085 in, Glass Fab, Rochester, NY, USA); copper wire (22 gauge, Westlake Hardware, Lawrence, KS, USA); epoxy (J-B Weld, Sulphur Springs, TX, USA), and colloidal silver liquid (Ted Pella, Inc., Redding, CA, USA). All the solutions were prepared in 18.2 (MΩcm) water (Millipore, Kansas City, MO, USA).

Microchip fabrication

A silicon master with a double ‘t’ design used for PDMS microchip fabrication was constructed using a previously published photolithography protocol.²² Briefly, SU-8 2010 negative photoresist was spin-coated onto a silicon wafer at 1650 rpm for 30 s, resulting in 15-μm tall features. The film was then soft baked for 3 min at 95 °C. To transfer the microchannel design onto the photoresist layer, the silicon wafer with the negative photoresist film was covered with a transparency mask (Infinite Graphics, Minneapolis, MN, USA) with the desired microchannel design and then exposed to UV light for 10 s (15 mW/cm²) using a UV flood source (ABM Inc., Scotts Valley, CA, USA). After the photoresist exposure, the silicon master was baked at 95 °C for 3 min on a programmable hotplate (Thermo Scientific, Waltham, MA, USA). SU-8 developer was used to develop the silicon master, and it was rinsed with isopropyl alcohol and dried with N₂ gas prior to the hard bake at 200 °C for 2 h. The separation and buffer flow channels were 40 μm wide and the microdialysis sampling channel was 600 μm wide. All other microchip dimensions are given in Figure 1. To create the PDMS microchip, 20 g of a 10:1 mixture of PDMS:curing agent was poured onto the silicon wafer and cured overnight. A 4-mm Harris Uni-core biopsy punch was used to create buffer, buffer waste, and sample waste reservoirs.

figure 1.

flow gated sample injection profile of a “double t” microchip (a) and an electropherogram (b) obtained for a standard mixture (50 μm) on such a chip. standard mixture was perfused at 1 μl/min flow rate with the absence of a md probe. bge: 15 mm phosphate at ph 7.4 with 2.5 mm boric acid and 15 mm sds. 1.0 v detection voltage at end-channel detection.

Electrode fabrication

The PPF electrodes used for these studies were fabricated using the previously described procedure with a few modifications.²³ Briefly, AZ 1518 positive photoresist was spin-coated onto a clean quartz glass plate at 100 rpm for 10 s and then at 2000 rpm for 20 s. The coated plates were prebaked at 100 °C for 3 min, covered with the desired positive photomask, and exposed to UV light for 7 s (15 mW/cm²). After exposure, the plate was developed in AZ 300 MIF developer for 13 s and quickly rinsed with NANOpure water. After drying under nitrogen, the plates were post-baked at 100 °C for 10 min. The features on the quartz plate were then pyrolyzed using a Lindberg/Blue M Three-Zone Tube Furnace (Cole-Parmer) under N₂ gas. The temperature of the furnace was increased from ambient temperature to 925 °C at the rate of 5.5 °C/min, held at 925 °C for 1 h, and then allowed to cool to room temperature. Amperometric detection was conducted at 35 μm wide electrodes. All electrodes fabricated using this method were 440–480 nm in height.

Assembly of hybrid PDMS/glass microchips for on-line MD-ME analysis

To couple microdialysis sampling with ME-EC, a protocol for generating a partially irreversible bond between the PDMS chip and glass substrate was developed. The PDMS chip was first reversibly aligned on the glass platform with a PPF electrode, leaving the desired gap between the microchannel and the electrode for end-channel detection (Figure 2a). Once aligned, the top section of the chip was then delaminated and folded back to create a gap between the glass and the channel side of the microchip (Figure 2b). A handheld corona discharger (Model BD 20: Electro-Technic Products, Inc., Chicago, IL, USA) was slowly moved across this gap for about 1 min to simultaneously expose the glass and the channel side of the chip to plasma (Figure 2c). Once exposed, the tape holding the top of the chip up was removed and the chip was allowed to reseal with the glass surface (Figure 2d). Slight pressure can be applied to the chip to remove air pockets between the surfaces. To facilitate the bonding, the device was heated for up to 5 min using a commercial hair drier and left to bond for at least 10 min. To couple the microchip to the microdialysis pump, a 20-gauge stainless steel blunt needle was used to make a hole in the PDMS for the sample inlet prior to alignment of the electrode. A stainless steel connector pin with a similar diameter was used to direct the microdialysate into the sampling channel of the chip.

figure 2.

improved chip bonding procedure: a) reversibly align the microchannel with the ppf electrode, b) fold back the top section of the chip from the glass and temporarily tape to the opposite end of the glass plate, c) simultaneously expose the top sections of the pdms chip and the glass substrate, d) re-seal the chip onto substrate by removing the tape and insert a stainless-steel tube to connect the chip to md flow.

Electrophoresis procedure

All the ME-EC analyses were carried out using devices with either a simple ‘t’ or double ‘t’ configuration. Spellman CZE 1000R (Hauppauge, NY, USA) high voltage power supplies were used for all separations and LabView (National Instruments, Austin, TX, USA) software written in-house was used to control the voltage output. The separation was accomplished by applying +1900 V across the separation channel. For the simple ‘t’ devices, the electrokinetic gate between the buffer and the sample flow was established by applying +1600 V as the sample voltage. Sample was injected into the separation channel by floating the separation voltage for 1.5 s. The injection flow profile is shown in Figure 1.

In vitro and in vivo sample preparation

All animal experiments were performed in compliance with all applicable federal statutes and regulations related to animals and experiments involving animals and adhere to the principles stated in the Guide for the Care and Use of Laboratory Animals (8^th edition) published by the National Academy of Sciences in 2011. All protocols were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Kansas and meet the standards set by the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC).

Homogenized brain samples

Brains isolated from male Sprague-Dawley rats weighing 250–280 g were used for the in vitro L-DOPA metabolism studies. Brains were extracted postmortem and immediately submerged in 3 mL of either 15 mM pH 7.4 phosphate for immediate use or artificial cerebral spinal fluid (aCSF) for later use. Brains in aCSF were stored at –80 °C until the time of the experiment. The brains were homogenized using a mechanical tissue homogenizer (VirTis Company, Gardiner, NY, USA). The homogenate was then centrifuged at 10 000g for 15 min; 1.5 mL of the supernatant was used for on-line analysis. Sampling was performed using a 3-cm long linear microdialysis probe (PAN membrane, 20 kDa MWCO, and 220/320 μm ID/OD) fabricated in-house using the following procedure. A 3.4 cm-long piece of PAN membrane was carefully cut using a sharp razor blade. Then two 15 cm long polyimide tubes (122/163 μm ID/OD) were inserted into the membrane (about 2 mm from each side). UV curable glue was used to fix the membrane with the polyimide tubes. Care must be taken not to clog the polyimide tube inlets and membrane with the glue. Similarly, two 1 cm long PTFE tubes were glued to the free ends of the polyimide tubes to facilitate the connection between the linear probe with the syringe pump and the microchip inlet. All on-line MD-ME-EC experiments were performed at a 1 μL/min sampling flow rate.

In vivo sampling from anesthetized rats

Sprague-Dawley rats weighing 250–300 g were anesthetized via inhalation of isoflurane (2%) followed by an intraperitoneal injection of 100 mg/kg ketamine and 10 mg/kg xylazine diluted in saline. Rats were connected to a constant isoflurane supply (0.5–2%) to maintain the anesthesia throughout the brain surgery and during the sampling. The rat was placed in a stereotaxic instrument for placing the microdialysis guide cannula into the striatum of the brain, following the coordinates A/P +0.7, M/L −2.7 and V/D −3.4 (from bregma).²² The guide cannula was held in place using dental acrylic and metal screws. The microdialysis probe was inserted through the guide cannula once the guide cannula was in place. Prior to the on-line experiments, the rat was allowed to recover from surgery for at least 45 min, during which time 15 mM sodium phosphate solution (NaH₂PO₄/Na₂HPO₄) at pH 7.4 was continually perfused through the microdialysis probe at a flow rate of 1 μL/min.

For the on-line in vivo experiments, the 20 kDa polyarylethersulfone (PAES) microdialysis probe (2 mm or 4 mm membrane, Harvard apporatus, Hilliston, MA, USA) was connected to the syringe pump and the ME-EC microchip using 15-cm fluorinated ethylene propylene (FEP, 0.65 mm O.D and 0.12 mm I.D) tubes with microdialysis connectors on each side of the probe. Unless otherwise mentioned, a flow rate of 1.0 μL/min was employed for microdialysis sampling. For DA release experiments, 15 mM sodium phosphate buffer was replaced with 100 mM K⁺ solution in the form of potassium phosphate buffer (~56 mM, K₂HPO₄/KH₂PO₄) at pH 7.4.

Results and discussion

Evaluation of the partially reversible PDMS/Glass MD-ME-EC interface

In the previous work, Saylor et al. used a 5 cm long double ‘t’ shaped microchip with flow-gated injection to introduce the microdialysis sample into the separation channel.²² In that design, the PDMS/glass hybrid chip had to be able to withstand the pressure generated from the hydrodynamic flow used for microdialysis sampling. This was achieved by simultaneously treating the PDMS and glass surfaces using a plasma cleaner, under vacuum.²² During the plasma treatment process, the PPF electrode portion of the glass was covered with a piece of sacrificial PDMS to protect the electrode, and to achieve reversible bonding around the electrode, so that it could be reused. Electrode alignment was then performed after the surface plasma treatment process. However, due to fast surface deactivation, the two oxidized surfaces had to be brought together very quickly, and in one attempt, to obtain optimal sealing. This process often resulted in poor microchannel/electrode alignments due to inadequate time available to use a microscope to view the microchannel/electrode interface. Therefore, this method would often result in several wasted microchips to get one working setup.

In the improved fabrication process presented in this paper, the chip-bonding method has been greatly simplified by replacing the vacuum-sealed plasma cleaner with a handheld corona discharger. In contrast to the plasma cleaner, the handheld corona discharger can be used to expose only a selected portion of the substrate without having to insulate the rest of the surface. In the present method, a the PDMS layer containing the separation channel was first reversibly aligned on the glass platform at the PPF electrode, leaving a 5 μm gap between the microchannel and the electrode for end-channel detection. The top section of the chip was folded back and temporarily taped to the end of the glass substrate. The two surfaces were then oxidized by slowly moving the handheld corona discharger across the gap between the chip and the glass. During this treatment process only the top section of the PMDS chip and the glass surface directly under the chip were exposed to the plasma, while the bottom section of the PDMS chip remained reversibly aligned with the electrode (Figure 2). In this manner, the desired microchannel/electrode alignment can be reproducibly obtained in less time without having to sacrifice additional microchips. More importantly, this method avoids any damage to the electrode due to the plasma exposure.

Evaluation of system for monitoring of dopamine generated in vitro

Figure 1b shows an electropherogram recorded using the developed MD-ME-EC device for mixtures of standard solutions of 3-O-MD, L-DOPA, HVA, DOPAC, DA and 3-MT. The optimization of the microchip electrophoresis separation and electrochemical detection of these compounds has been described previously.^{22, 24} For both the in vitro and in vivo studies, 15 mM phosphate buffer at pH 7.4, 15 mM SDS, and 2.5 mM boric acid was used as the background electrolyte. For the in vitro MD-ME-EC study, the formation of metabolites of L-DOPA was monitored by adding L-DOPA (0.5 mM) to a brain homogenate. 3-nitrotyrosine (0.3 mM) was also added and used as the internal standard. Microdialysis sampling was conducted by using a linear MD probe and using the phosphate buffer as a perfusate. A representative electropherogram recorded for in vitro on-line monitoring of L-DOPA metabolism in brain homogenate is shown in Figure 3. Under the separation conditions tested, the PDMS microchip lasted more than 4 hours as long as the buffer was replenished frequently (once every 6½ min). However, as seen in Figure 3, this periodic sample replenishing can cause random fluctuation in the background current. On the other hand, these frequent buffer replenishments act to preserve the microchip separation and prevent the injection to injection migration time variability caused by the fluctuations in separation current, Joule heating, and pH changes in the buffer reservoirs.

figure 3.

on-line monitoring of l-dopa metabolism in vitro by homogenized rat brain. bge: 15 mm phosphate at ph 7.4 with 2.5 mm boric acid and 15 mm sds. 1.0 v detection voltage at end-channel detection. 15 mm phosphate was perfused through the linear md probe at 1 μl/min sampling flow rate. insets: a. baseline response for brain homogenate b. first appearance of l-dopa and nitrotyrosine (internal standard) signals and c. electropherogram after 40 minutes from l-dopa spiking respectively. ‘*’ indicates the beginning of each sample injection.

The series of stacked electropherograms shown in Figure 4a demonstrate the changes occurring over time with the homogenate/L-DOPA reaction mixture. Prior to the addition of L-DOPA, four distinct peaks were apparent from the brain homogenate. Although the exact identities of the first three peaks remain uncertain, the largest peak (4) appearing at around 80 s was later confirmed to be ascorbic acid (AA) by comparing migration time of standard AA solution with that obtained for microdialysis sample. The lag time of the device was determined to be approximately 8 minutes as indicated in Figure 3. Since the lag time is largely governed by the microdialysis flow rate, the total length of the flow tubing, and the dead volume of the connectors, it could be reduced in the future by reducing the length of tubing between the animal and the chip, employing a higher microdialysis flow rate, or using shorter, smaller bore tubing.

figure 4.

a) enzymatic formation of da from l-dopa in rat brain homogenate b) continuous production of da as a function of time. peak identity, 1), 2) and 3) are unknown, 4) ascorbic acid, 5) l-dopa, 6) 3-nitrotyrosine (is), 7) dopamine, and 8) hva (shoulder on the left side of the ascorbic acid peak).

Dopamine, 3-OMD, DOPAC, 3-MT, and HVA have been identified as the major brain metabolites of L-DOPA in vivo.²⁴ There are mainly three enzymes that contribute to L-DOPA metabolism in the central nervous system (CNS). These are aromatic L-amino acid decarboxylase (AADC), monoamine oxidase (MAO-A/B), and catechol-O-methyl transferase (COMT).²⁵ Initially, L-DOPA is metabolized to DA by AADC. Then, DA is transformed into DOPAC by MAO. COMT is responsible for the conversion of DOPAC to HVA. However, as can be seen in Figure 4a, the only metabolite that could be identified in the homogenized rat brain was DA. This can be explained by the sample preparation procedure used. In this study, only the supernatant of the centrifuged brain homogenate was used for analysis. Therefore, the extent of the L-DOPA metabolism depends on the availability of the enzymes and the cofactors present in the supernatant. AADC exists both in the CNS (dopamine and serotonin neurons) and in capillary endothelial cells,²⁶ whereas the other two enzymes are present either inside cell organelles (MAO, in the outer mitochondrial wall and glia) or in the postsynaptic spaces of dopaminergic neurons (COMT) in the CNS.²⁷ Therefore, it is possible to extract more AADC than the other two enzymes into the supernatant.

Since the goal of this study was to test the feasibility of using the developed MD-ME-EC sensor for continuous on-line monitoring, no attempt was made to improve the in vitro experimental protocol to obtain complete L-DOPA metabolism. Figure 4b shows a graph of the continuous formation of DA as a result of the reaction between added L-DOPA and aromatic L-amino acid decarboxylase present in rat brain homogenate. With these preliminary data, we confirmed the feasibility of employing the developed ME-EC separation system for continuous on-line monitoring of microdialysis samples. After confirming the suitability of the device for on-line in vitro measurements, the system was then employed for in vivo monitoring of stimulated release of dopamine in an anesthetized rat.

On-line MD-ME-EC monitoring of stimulated release of dopamine in vivo

It is well known that dopamine release can be stimulated by an electrical pulse or drugs or by introducing depolarizing agents to DA-rich regions in the rat brain.^{7, 28–30} In previous studies, high concentrations of depolarizing agents such as high K⁺ solutions have been used to evoke DA release in the striatum.^{7, 29} It is believed that the amount of DA that is released is primarily dependent on the K⁺ ion concentration and the duration of the infusion. K⁺ concentrations ranging from 50 mM to 100 mM have been shown to induce release of DA into the extracellular space. An off-line microdialysis-based study conducted by Ripley et al. reported a 2900% increase of extracellular DA over the basal level after the addition of 100 mM K⁺ to the perfusate.²⁹ Additionally, an on-line microdialysis-liquid chromatography study conducted by Weber et al. reported concentrations of dopamine over 1 μM in microdialysates obtained following 60 and 100 mM K⁺ stimulations.⁷

In the present study, we employed our on-line MD-ME-EC system to monitor near real-time release of DA in an anaesthetized rat in response to high K⁺ concentration pulses. Following the protocol described in the methods section, the rat was subjected to two separate short-term perfusions of 100 mM K⁺ with 65 min recovery time between stimulations. As has been reported by Ripley et al., a 1-hr recovery time between 10 min long, 100 mM K⁺ perfusions is sufficient to avoid any effect from the prior stimulation on the next evoked release.²⁹

Figure 5 shows a representative electropherogram recorded using the developed MD-ME-EC system for K⁺-induced DA release. In this experiment, 10-min and 20-min long stimulations were initiated by the delivery of 100 mM K⁺ via retrodialysis through the probe implanted in the striatum, while monitoring of the changes in the microdialysate composition was carried out continuously. Prior to stimulation and during the recovery time, 15 mM sodium phosphate buffer was perfused through the rat brain. Due to the significant difference between the conductivities of the separation BGE (15 mM sodium phosphate at pH 7.4, 15 mM SDS, and 2.5 mM boric acid) and the stimulation solution (100 mM potassium phosphate), noticeable background current fluctuations were observed when the stimulation solution entered the separation channel (Figure 5, points A and B). The appearance and disappearance of these fluctuations can be used to estimate the lag time of K⁺ solution reaching the microchip as well as the time window in which to expect the DA efflux signal in the electropherogram.

figure 5.

on-line md-me-ec monitoring of k⁺-stimulated da release in the striatum of an anesthetized rat. ‘▲’ indicates the background current fluctuations caused by buffer replenishments. a and b represent the beginning of background current fluctuations due to the arrival of the k⁺-containing microdialysate plug into the separation channel. ‘*’ represent sample injections.

In addition to the noticeable background fluctuations (highlighted area on Figure 5), two additional peaks (X and Y) were observed in the electropherograms during both K⁺ perfusion time ranges. During the first K⁺ stimulation period, the presence of both X and Y was noticeable for about 11 min (Figure 6a). Similarly, during the second stimulation, peak X was observed for about 19 min, which is comparable to the K⁺ stimulation time (Figure 6b). In contrast to the first stimulation, peak Y was not prominently visible during the second stimulation. By comparing the DA signal obtained for the standard DA solution analyzed using the same MD-ME-EC system on the same day, peak X was identified as DA. The time course of the appearance of DA resulting from high potassium stimulation can be seen in Figure 7.

figure 6.

detection of k⁺ induced da release (peak x) during two discrete stimulations of the same animal. a) first k⁺ induced da release, the stimulation was conducted for 10 min. b) show the production of da during the second k⁺ stimulation event, the second stimulation was conducted for 20 min. traces a and b denote the first appearance of k⁺-related system peak on the electropherogram after the first and the second k⁺ stimulation respectively. all the other traces show the time from the appearance of trace a and b for the two stimulations.

figure 7.

k⁺ induced da release in anesthetized rat brain during the first (10 min) and second (20 min) stimulations.

The limit of detection for dopamine in the on-line system is approximately 1 μM. The concentrations of dopamine released due to high K⁺ in the rat experiment described here were high enough to detect by the on-line system. However, the limits of detection/quantitation are not yet sufficient to consistently detect dopamine release in all animals or monitor basal levels of dopamine. Therefore, further improvement is needed to use this system in quantitative, on-line in vivo monitoring of extracellular catecholamine concentrations such as dopamine, HVA, and DOPAC in the rat brain.

Concluding remarks

An improved method was developed for fabricating a MD-ME-EC sensor with PPF carbon electrodes. The modified method significantly reduced the device fabrication time and greatly minimized the waste of microchips due to poor microchannel/electrode alignment. Under the optimal separation conditions, the ME-EC system has the ability to separate five major metabolites in L-DOPA metabolic pathway, namely 3-O-MD, HVA, DOPAC, dopamine, and 3-MT in less than 100 s with near-baseline resolution. Both in vitro and in vivo studies demonstrated in this research proved the ability of the MD-ME-EC for use in long-time, on-line monitoring of microdialysis samples. This is also the first report of using a MD-ME-EC system to monitor the release of endogenous dopamine following high potassium stimulation.

Although the work presented in this paper is primarily focused on demonstrating the ability of the device for monitoring DA in vivo, microchip electrophoresis separation-based sensors can be used to monitor multiple dopamine metabolites at the same time. In that regard, future work will concentrate on improving the limits of detection for the on-line system and its use for monitoring L-DOPA transport and metabolism across the blood-brain barrier. In the future this same device can be used to study the influence of drugs and other chemical species on the concentration of catecholamines and other endogenous electroactive species in the brain. Current research is utilizing the device to look at the effects of L- and D-glutamate and oxidative stress on the concentrations of ascorbate and dopamine in the brain.