Comparison of Cell Expression Formats for the Characterization of GABA_A Channels Using a Microfluidic Patch Clamp System

Abstract

Ensemble recording and microfluidic perfusion are recently introduced techniques aimed at removing the laborious nature and low recording success rates of manual patch clamp. Here, we present assay characteristics for these features integrated into one automated electrophysiology platform as applied to the study of GABA_A channels. A variety of cell types and methods of GABA_A channel expression were successfully studied (defined as I_GABA>500 pA), including stably transfected human embryonic kidney (HEK) cells expressing α₁β₃γ₂ GABA_A channels, frozen ready-to-assay (RTA) HEK cells expressing α₁β₃γ₂ or α₃β₃γ₂ GABA_A channels, transiently transfected HEK293T cells expressing α₁β₃γ₂ GABA_A channels, and immortalized cultures of human airway smooth muscle cells endogenously expressing GABA_A channels. Current measurements were successfully studied in multiple cell types with multiple modes of channel expression in response to several classic GABA_A channel agonists, antagonists, and allosteric modulators. We obtained success rates above 95% for transiently or stably transfected HEK cells and frozen RTA HEK cells expressing GABA_A channels. Tissue-derived immortalized cultures of airway smooth muscle cells exhibited a slightly lower recording success rate of 75% using automated patch, which was much higher than the 5% success rate using manual patch clamp technique by the same research group. Responses to agonists, antagonists, and allosteric modulators compared well to previously reported manual patch results. The data demonstrate that both the biophysics and pharmacologic characterization of GABA_A channels in a wide variety of cell formats can be performed using this automated patch clamp system.

Introduction

γ-Amino butyric acid (GABA) is the primary inhibitory neurotransmitter in the central nervous system. GABA_A receptors belong to a family of ligand-gated ion channels mediating fast synaptic transmission. Ionotropic GABA_A receptors consist of 5 protein subunits arranged around a central pore that constitutes the actual ion channel. There are 16 different subunits comprising the GABA_A receptor family, including α_1–6, β_1–3, γ_1–3, δ, ɛ, π, and θ. GABA_A receptor subtype combinations, α₁β_2/3γ₂, α₂β₃γ₂, and α₃β₃γ₂, are the most abundant, and are mainly present in the brain. These channels conduct a Cl⁻ current when exposed to agonists. The increased Cl⁻ conductance drives the membrane potential toward the reversal potential of the Cl⁻ ion (∼75 mV in neurons) inhibiting the firing of new action potentials.^1,2 GABA_A receptors are of major importance as pharmacological targets for anxiolytics, anti-schizophrenics, anesthetics, and sleep aids.

Patch clamp recording is an information-rich assay that has become the gold standard in the study of ion channels, as well as for the selection and development of ion channel drug candidates. Patch clamp recordings are real-time measurements of the ionic currents passing through the cellular membrane of a single cell, providing a real-time recording of ion channel activity. Patch clamp studies define a compound's effect on channel desensitization, kinetics, and gating, which reveals many aspects of drug and ion channel interaction. While cell-based fluorescence assays for screening ion channels has proven to be powerful for assessing a large number of compounds quickly in ion channel targeted drug discovery efforts,^3,4 hit confirmation and mechanism of action studies require the use of patch clamp recordings.

Automated patch clamp systems have been developed over the last decade to achieve higher experimental throughput as compared to conventional manual patch clamp, which is slow and laborious. Automated patch clamp systems integrate robotics for liquid handling, and enables multiple recordings in parallel.^5–8 It is expected that increasing capacities of such systems will allow for a greater capacity in the conduct of ion channel studies and ion channel targeted compound screening.

The overwhelming majority of automated electrophysiology experiments employ recombinant cell lines stably overly expressing a particular ion channel of interest. Such recombinant cell lines are expensive to establish or acquire from vendors and may lack the normal channel kinetics of endogenously expressed channels. Currently, only a small number of stable recombinant cell lines expressing different GABA_A channels are commercially available, which limits essential investigations to determine ligand effects based on different GABA_A receptor subunit compositions. Furthermore, the gradual loss of expression levels of stably transfected plasmids at higher passages can influence the integrity of results. Recently, a new assay format, frozen ready-to-assay (RTA) cells, has been introduced to the market.⁹ These cryopreserved cells are thawed on demand and offer convenience and improved consistency by avoiding variations between different cell passages. Furthermore, using frozen RTA cells decouples the experimental schedule from cell culture, which leads to important time savings. For smaller scale studies, RTA cells are also more affordable as compared to acquisition of the immortalized stable cell lines.

If the study of different GABA_A subunit compositions is desired, transiently transfected cells allow maximum flexibility. The transient transfection of a specific combination of GABA_A subunits can be performed on demand as long as the cDNA is available, while only maintaining one cell type in culture. The limitations of this approach include the expression heterogeneity of transfected plasmids, uncertainty in the true subunit combination achieved for a given ion channel and damage to the cell membrane during the transfection, which can lead to low success rates using manual patch or single-cell configured automated patch clamp systems.

Despite recent advances in transfection technology, disparities persist between experiments using recombinant cell lines as compared to primary cells with endogenous levels of ion channel expression that are extracted from tissue samples. In nontransfected cells, the assay targets are expressed in a more physiologically relevant environment, and the complex interplay of the endogenously expressed ion channels, second messengers, and other cell signaling proteins can be better recapitulated than in recombinant cell lines. Therefore, there is a growing interest in studying endogenously expressed ion channels to achieve a higher level of physiological relevance. Limitations of transiently transfected cells and natively expressing cells are overcome by the use of the IonFlux™ instrument, which records the average current and resistance of 20 cell ensembles and utilizes continuous microfluidic perfusion.

Previous publications have described the principles of operation for the IonFlux system, and shown electrophysiology and pharmacology characterization for both ligand-gated and voltage-gated ion channel targets on the IonFlux platform, which applies whole cell voltage clamp to an ensemble of 20 cells in parallel.⁷ The IonFlux is the first microfluidic patch clamp system featuring continuous recording coupled with fast solution exchange—an ideal combination for the study of ligand gated channels.⁷ In the present study, we focused on determining the robustness of GABA assays in a number of different cell expression formats and report the success rate, throughput, and accuracy of pharmacology results as compared to available manual patch clamp references. The different assay types explored were as follows: recombinant human embryonic kidney (HEK293) cells stably expressing the α₁β₃γ₂ GABA_A channel; frozen RTA recombinant HEK293 cells expressing either the α₁β₃γ₂ or α₃β₃γ₂ GABA_A channel; transiently transfected HEK293T cells expressing the α₁β₃γ₂ GABA_A channel; immortalized ex vivo human airway smooth muscle cells endogenously expressing GABA_A channels composed of α₄, α₅, β₃, γ₁, γ₂, δ, or θ subunits.^10–12 The side by side comparison of different cell expression formats provides an overview for the selection of the appropriate protocol and expression systems, outlining tradeoffs between biological relevance, assay flexibility, and raw experimental capacity. Ensemble recording coupled with continuous fluid perfusion yields high success rates for the study of GABA_A receptors in transfected HEK293 cells or endogenously expressed in smooth muscle cells.

Materials and Methods

Reagents

GABA, bicuculline, picrotoxin, muscimol, and diazepam were purchased from Sigma-Aldrich, and used without further purification. HZ166 was provided by J.M.C.

Cell Culture and Preparation of Recombinant GABA_A Receptor-Expressing HEK293 Cells

Cells expressing hGABA_A (Millipore PrecisION™ hGABA_A α₁β₃γ₂-HEK Recombinant Cell Line; Cat# CYL3053) were cultured in 175-cm² filter-top flasks containing DMEM/F12 glutamax (Invitrogen #10569-010), 10% fetal bovine serum (FBS) (Thermo #SH30070.02), 1% nonessential amino acids (Invitrogen #11140), 400 μg/mL geneticin (Invitrogen #10131), 100 μg/mL hygromycin B (Invitrogen #10687), and 0.625 μg/mL puromycin (Clontech #631306) at 37°C and 5% CO₂. The cells were kept below 80% confluency and split once a week. For the cell preparation, flasks were first washed with 2 mL of Ca²⁺ and Mg²⁺-free phosphate-buffered saline (PBS) followed by the addition of 5 mL of Detachin™ solution for cell detachment. After release, the cell suspension was spun for 90 s (180 g) prior to being resuspended in extracellular solution (5×10⁶ cell/mL).

Cell Culture and Preparation of Frozen RTA Recombinant GABA_A Receptor-Expressing HEK293 Cells

Two RTA cell lines—one expressing hGABA_A α₁β₃γ₂ (Millipore #CYL3053RTA) and the other one expressing hGABA_A α₃β₃γ₂ (Millipore #CYL3068RTA)—were tested for this study. For storage, cells were kept in liquid nitrogen before use. For the assay, cells were thawed by immediately immersing the vial in a 37°C water bath. As soon as the ice was thawed, 1 mL of prewarmed HEK293 serum-free media (Invitrogen #11686) was added to the vial. The cell solution was transferred dropwise into a 15-mL conical tube holding 10 mL of prewarmed HEK 293 serum-free media. The cell suspension solution was centrifuged at 180 g for 90 s to separate cells from freezing media. The cell pellet was resuspended in 10 mL warm HEK 293 serum-free media, and the cells were incubated for 30 min at 37°C, washed, and resuspended in extracellular solution at a density of 3–5×10⁶ cells/mL before the electrophysiology experiment.

Cell Culture and Preparation of Transiently Transfected HEK293T Cells Expressing the α₁β₃γ₂ GABA_A Channel

HEK293T cells (ATCC) were cultured at 37°C and 5% CO₂ in 75 cm² flasks using growth media composed of MEM/EBSS without phenol red but with L-glutamine (2 mM) and glucose (1 mM) (Hyclone #SH30024FS), 1% nonessential amino acids (Hyclone #SH3023801), 1 mM sodium pyruvate (Hyclone SH3023901), penicillin (100 units/mL) and streptomycin (100 μg/mL) (Hyclone #SV30010), and 10% heat inactivated FBS (Hyclone #SH3008803HI). All flasks were treated for 10 min at 37°C with a 1% Matrigel (BD #CB40230A) solution in MEM/EBSS and split three times a week. The transfections were performed at 50%–70% confluency. Therefore, growth media were aspirated and replaced with assay media containing MEM/EBSS without phenol red but with L-glutamine (2 mM), glucose (1 mM) (Hyclone #SH30024FS), 1% nonessential amino acids (Hyclone #SH3023801), 1 mM sodium pyruvate (Hyclone SH3023901), penicillin (100 units/mL) and streptomycin (100 μg/mL) (Hyclone #SV30010), and 10% dialyzed FBS (Invitrogen #26400-036). A transfection mixture was then added to the flask. The mixture consisted of 1.5 mL of serum-free MEM/EBSS media, 5 μg of each of the GABA_A receptor subunit DNA (α₁, β₃, and γ₂), Lipofectamine™ LTX (75 μL), and PLUSTM reagent (25 μL) (Invitrogen, #15338-100). After 24 h, without changing media, the cells were washed with 10 mL of Ca²⁺ and Mg²⁺-free PBS, followed by the addition of 3 mL of Detachin solution, after which the cells were incubated for 2 to 5 min depending on their confluency. The cell suspension was added to 7 mL of growth media to inhibit Detachin, followed by centrifugation (2 min at 180 g) and resuspension in serum-free MEM/EBSS media (5 mL). The centrifugation and resuspension were repeated two more times. Cells were then resuspended in HEK293 serum-free media (Invitrogen #11686) (5 mL) and placed on a shaker for 30 min. The cells were centrifuged again and then resuspended in extracellular solution (5 mL). The centrifugation and resuspension in extracellular solution was repeated two more times, resulting in cell suspension of 5×10⁶ cells/mL, which was then used for electrophysiology experiments.

Cell Culture and Preparation of Immortalized Human Airway Smooth Muscle Cells Endogenously Expressing GABA_A Channels

Human bronchial smooth muscle cells were a kind gift from Dr. William Gerthoffer (University of South Alabama) and were immortalized by the stable expression of human telomerase reverse transcriptase (hTERT).¹⁰ Expression of hTERT extends the life span of smooth muscle cells, but does not affect the endogenous level of expressed channels. Cells were originally isolated from human lung surgical specimens under approved institutional review board (IRB) protocols. Cells were grown to confluence in 15-cm-diameter collagen-coated (BD Biosciences [354236] rat tail collagen type 1, using 5 μg/cm²) plastic dishes. Cells were grown in M199 media containing 10% fetal bovine serum, 0.25 ng/mL epidermal growth factor, 1 ng/mL fibroblast growth factor, ITS supplement (1 mg/mL insulin, 0.55 mg/mL transferrin, and 0.67 μg/mL sodium selenium), and antibiotics (100 units/mL penicillin G sodium, 100 μg/mL streptomycin sulfate, and 0.25 μg/mL amphotericin B) in a humidified atmosphere of 5% CO₂ at 37°C. Media were aspirated from the dishes and collagenase type IV (Sigma C5138) (500 units/mL in 5 mL basal M199 media) was added, and the dishes were incubated 15 min at 37°C in a cell culture incubator. Cells were released from the dish by gentle trituration using a pipette and transferred to a 15-mL conical tube for centrifugation (300 g, 3 min). The cell pellet was resuspended in the desired external buffer solution and washed twice. Cell solution (3–5×10⁶ cells/mL) was then loaded to the inlet wells in a 96-well IonFlux microfluidic plate.

Automated Patch Clamp and Ensemble Recording

The IonFlux platform utilizes integrated microfluidics and the wells on the 96-well plate are interconnected through microfluidic channels. Buffers, compound solutions, cells, and waste are deposited and handled in a single plate within a standard 96-well-plate format. The solutions are moved by pneumatic pumps, eliminating automated pipetting and possible cross contamination. The IonFlux plate layout consists of units of 12 wells: two wells contain intracellular solutions, 1 contains extracellular solution plus cells, 8 contain compounds of interest, and 1 well is used for waste collection. Table 1 shows an example of an assay protocol with the 12-well plate layout, loading parameters and assay steps. A 96-well plate contains 8 patterns, and can provide 16 recordings (384-well plates provide 64 recording channels). Cells are captured from suspension by applying suction to microscopic channels in the ensemble of the recording array. Once the array is fully occupied by 20 individual cells, the applied suction breaks the cell membrane of captured cells, establishing whole cell voltage clamp. The calculated current over time is dictated by the compound affinity, ion channel kinetics, and solution exchange speed. IonFlux utilizes microfluidics with the fluid flow regulated by valves and regulators that control the duration and timing of compound application, which are synchronous across the plate. Compound solutions can be introduced or replaced within ∼100 ms (0%–90% washout),⁷ which can well accommodate GABA_A channel recording.

Table 1.

Example of Assay Protocol

For compound applications, pressure was applied to the appropriate compound wells, introducing a compound dissolved in extracellular solution rapidly to the patch clamped cells. For recording of GABA_A channel currents, cells were clamped at a holding potential of −80 mV. The extracellular solution was (in mM):138 NaCl, 4 KCl, 1 MgCl₂, 1.8 CaCl₂, 10 HEPES, 5.6 glucose, pH 7.4 with NaOH. The intracellular solution was in mM: 60 KCl, 70 KF, 15 NaCl, 5 EGTA, 5 HEPES, pH 7.2 with KOH.

Each GABA application lasted for 3 s, followed by a 30 to 60 s of wash to allow the channels to recover in stably transfected cultured or RTA HEK293 cells. Bicuculline or picrotoxin at each concentration was preincubated for 3 min before GABA (10 μM) and bicuculline or picrotoxin co-application. Allosteric ligands diazepam or HZ166 were co-applied for stably transfected and transiently transfected cell with 1 or 0.3 μM GABA, respectively.

Data Analysis

Data analysis and graphical presentation were performed using a combination of IonFlux software, GraphPad Prism 5, and Microsoft Excel. Dose–response curves were analyzed using nonlinear regression (GraphPad Prism 5, variable slope, sigmoidal dose response). All data are shown with arithmetic mean±standard error of mean (SEM).

Results

The Current Response to GABA or Muscimol in Cultured Recombinant HEK293 Cells Stably Expressing the α₁β₃γ₂ GABA_A Channel Was Reliably Determined with the Automated Patch Clamp Instrument IonFlux

The inward Cl⁻ current ensemble of 20 HEK293 cells, stably expressing the α₁β₃γ₂ GABA_A channel in the presence of 10 μM GABA, is shown in Figure 1a as a screen capture of IonFlux software. The GABA-evoked Cl⁻ current increased to peak (10%–90%) within 400±50 ms. GABA concentrations between 0.1 and 30 μM were applied to determine the dose response of HEK293 cells that stably express the α₁β₃γ₂ GABA_A channel (Fig. 1b). At higher GABA concentrations, the channel showed faster desensitization (Fig. 1b, inset). The wash out of GABA using extracellular solution took ∼500 ms. The calculated EC₅₀ value from dose response was 2.87±0.24 μM, which is similar to the values reported in the literature, which ranged from 2.00 to 3.29 μM.^13–15 A second classic GABA_A agonist, muscimol, was used to demonstrate correct pharmacology and desensitization kinetics using this microfluidic platform in these HEK293 cells stably transfected with the α₁β₃γ₂ GABA_A channel (Fig. 1c). Muscimol showed faster desensitization than GABA stimulation as expected. Muscimol (0.1–100 μM) induced a dose-dependent current with an EC₅₀ of 1.43±0.38 μM consistent with values published with this GABA_A channel composition in HEK cells.¹⁶ Additionally, the consistency of measurements from plate to plate was determined for eight plates comparing the GABA (10 μM)-induced Cl⁻ currents of 128 cell ensembles across 8 plates compared to a vehicle control (0.1% dimethyl sulfoxide [DMSO]) in the extracellular solution (Fig. 1d). Average current from 10 μM GABA was 14.95±3.21 nA, whereas vehicle control (0.1% DMSO) yielded 0.16±0.03 nA, which was about 2 orders of magnitude lower than the measured GABA current.

Fig. 1.

(a) Representative current sweeps of GABA-induced current from HEK cells stably transfected with α₁β₃γ₂ GABA_A receptors. (b) GABA dose–response curve (0.1–30 μM) in recombinant HEK cells expressing the α₁β₃γ₂ GABA_A receptor (n=40). Current data of independent measurements in duplicate were normalized to highest response for each dose response used for fitting data to a four-parameter (variable slope) nonlinear regression equation. Inset: Inward Cl⁻ currents are shown from one ensemble of cells exposed to GABA with increasing concentrations (0.1–30 μM). (c) Muscimol dose–response curve (0.1–100 μM) in stably transfected HEK cells expressing the α₁β₃γ₂ GABA_A receptor (n=6). Current data of independent measurements in duplicate were normalized to highest response for each dose response used for fitting data to a four-parameter (variable slope) nonlinear regression equation. Inset: Inward Cl⁻ currents are shown from one ensemble of cells exposed to muscimol with increasing concentrations (0.1–100 μM). (d) Consistency of GABA-induced current responses in HEK cell stably expressing the α₁β₃γ₂ GABA_A receptor. Eight plates were tested, and the average current at 10 μM GABA was ∼10 nA (shown as white bars), whereas the vehicle control (0.1% dimethyl sulfoxide) yielded an ∼100 pA current, which was 2 orders of magnitude lower than GABA current, and shown as black bars (P=0.003). GABA, γ-amino butyric acid; HEK, human embryonic kidney.

Pharmacological Responses of Cultured Recombinant HEK293 Cells Stably Expressing the α₁β₃γ₂ GABA_A Channel in the Presence of the Antagonists Bicuculline or Picrotoxin or the Allosteric Modulator Diazepam as Determined in the IonFlux Microfluidic Platform

Figure 2a depicts the dose–response inhibition of 10 μM GABA by the competitive GABA_A antagonist bicuculline at the α₁β₃γ₂ GABA_A channel stably expressed in HEK293 cells. The calculated IC₅₀ value of 13.9±0.5 μM is higher than the reported IC₅₀ value of 1.9±0.3 μM.^17–19 Similarly, 10 μM GABA was dose dependently inhibited by the GABA_A channel antagonist picrotoxin (0.3–100 μM) with an IC₅₀ of 13.6±0.3 μM in HEK293 cells stably expressing the α₁β₃γ₂ GABA_A channel (Fig. 2b).^13,20 GABA_A receptors containing α₁, α₂, α₃, or α₅, any β subunit, and γ₂ subunits are sensitive to benzodiazepine modulation.^1,21 Benzodiazepines function as positive allosteric modulators of GABA at GABA_A channels and have been the focus of intense research due to the extensive therapeutic use of benzodiazepines. The dose-dependent potentiation of diazepam (0.01–10 μM) in the presence of 1 μM GABA in HEK293 cells stably transfected with the α₁β₃γ₂ GABA_A channel is shown in Figure 2c. The calculated EC₅₀ value for diazepam was 420±10 nM with a current potentiation of ∼300%, which is similar to values reported in the literature using manual patch clamp techniques.²²

Fig. 2.

(a) Dose-dependent inhibition of 10 μM GABA (an EC₈₀) by co-application of the GABA_A receptor antagonist bicuculline (0.03–100 μM) in HEK cells stably expressing α₁β₃γ₂ GABA_A receptors. Current data of n=5 independent measurements in duplicate were normalized to highest response for each dose response and used for fitting data to a four-parameter (variable slope) nonlinear regression equation. Inset: Representative dose–response sweeps of bicuculline (0.03–100 μM) inhibiting 10 μM GABA-induced current from one ensemble of HEK cells stably transfected with α₁β₃γ₂ GABA_A receptors. (b) Dose-dependent inhibition of 10 μM GABA (an EC₈₀) by co-application of the GABA_A receptor antagonist picrotoxin (0.5–100 μM) in HEK cells expressing α₁β₃γ₂ GABA_A receptors. Current data of n=6 independent measurements in duplicate were normalized and used for fitting data to a four-parameter (variable slope) nonlinear regression equation. (c) Dose-dependent potentiation of 1 μM GABA (an EC₂₀) by co-application of the GABA_A receptor allosteric enhancer diazepam (0.01–10 μM) in HEK cells expressing α₁β₃γ₂ GABA_A receptors. Current data of n=4 independent measurements in duplicate were normalized and used for fitting data to a four-parameter (variable slope) nonlinear regression equation. Inset: Inward Cl⁻ currents are shown from one ensemble of cells exposed to diazepam (0.01–10 μM) in the presence of 1 μM GABA (EC₂₀).

Frozen RTA HEK293 Cells, Stably Expressing GABA_A Channels α₁β₃γ₂ or α₃β₃γ₂, Yielded Similar Results as Cultured HEK293 Cells Stably Expressing the Same Channels

Two RTA HEK293 cell lines stably expressing α₃β₃γ₂ or α₁β₃γ₂ GABA_A channels were compared with cultured recombinant cells expressing the α₁β₃γ₂ GABA_A channel using an IonFlux-HT plate, which can perform 64 recordings in parallel. Figure 3 depicts representative dose–response sweeps of GABA dose responses (0.3–30 μM) in HEK293 RTA cells expressing the α₁β₃γ₂ (Fig. 3a) or α₃β₃γ₂ (Fig. 3b) GABA_A channels. GABA was more potent at α₃β₃γ₂ GABA_A channel (EC₅₀ value=0.9±0.04 μM) as compared to cells expressing α₁β₃γ₂ GABA_A channels (EC₅₀ value=1.8±0.1 μM) (Fig. 3c). The normalized response at 10 μM GABA showed that the α₃β₃γ₂ GABA_A channel had a slower on time and desensitization as compared to the α₁β₃γ₂ GABA_A channel (Fig. 3d). Cultured recombinant HEK293 cells and RTA HEK293 cells with the same GABA_A channel, α₁β₃γ₂, showed very similar kinetics profiles. Previous studies substantiate that the EC₅₀ value and affinity of GABA agonists are highly dependent on subunit composition of GABA_A channels.^17,23–25 The EC₅₀ values are in good agreement with previous manual patch clamp studies. The average seal resistance from cultured HEK cells on the same plate was ∼700 MΩ, as compared to ∼500 MΩ from RTA HEK cells.

Fig. 3.

(a) Representative dose–response sweeps of GABA (0.3–30 μM)-induced inward Cl⁻ current from one ensemble of ready-to-assay (RTA) HEK cells transfected with α₁β₃γ₂ GABA_A receptors. (b) Representative dose–response sweeps of GABA (0.3–30 μM)-induced inward Cl⁻ current from one ensemble of RTA HEK cells transfected with α₃β₃γ₂ GABA_A receptors. (c) Dose–response curves of GABA (0.1–30 μM)-induced currents in cultured HEK cells stably expressing the α₁β₃γ₂ subtype, (●) (n=16), RTA HEK cells expressing the α₁β₃γ₂ subtype (■) (n=14), and RTA HEK cells expressing the α₃β₃γ₂ subtype (▴) (n=8). (d) Overlay of representative current sweeps with 10 μM GABA in RTA HEK cells expressing either the α₁β₃γ₂ or α₃β₃γ₂ GABA_A receptor. ECS, extracellular solution.

Transiently Transfected HEK293T Cells Expressing GABA_A Channel α₁β₃γ₂ Yield Robust Currents in Response to GABA, Bicuculline, and the Benzodiazepine-Site Modulator HZ166

The GABA dose response (0.3–30 μM) for transiently transfected HEK293 cells expressing the α₁β₃γ₂ GABA_A channel is shown in Figure 4a along with an inset of current sweeps. The GABA response kinetic profile in these transiently transfected cells was similar to the kinetic profile in response to GABA seen in stably transfected HEK293 cells expressing the same subtype (α₁β₃γ₂) of GABA_A channels (Fig. 1b). The calculated EC₅₀ value was 1.3±0.29 μM, which is well in the range with EC₅₀ values of 1.8–2.9 μM from the cultured or RTA HEK293 cells expressing this same GABA_A channel subtype (Fig. 3c). Dose-dependent inhibition of GABA-induced currents was demonstrated by bicuculline (0.03–100 μM). After a 3-min pretreatment with varying concentrations of bicuculline, a 4-s co-application of bicuculline with 1 μM GABA (EC₈₀) (Fig. 4b). The calculated IC₅₀ value was 15.9±5.45 μM, a value similar to the IC₅₀ of bicuculline in HEK293 cells stably expressing the α₁β₃γ₂ GABA_A channel (Fig. 2a). HZ166 is benzodiazepine-based compound that exhibits antiseizure activities in mice and rats.²⁶ To determine its ability to modulate the α₁β₃γ₂ GABA_A receptor, transiently transfected HEK293T cells were first exposed to 0.3 μM GABA (EC₂₀) followed by an application of HZ166 and a co-application of HZ166 and 0.3 μM GABA. At a concentration of 30 μM, HZ166 showed a current potentiation of ∼200%. The modulation and the determined IC₅₀ value of 670±290 nM was similar to previously reported results using transfected oocytes.^26,27

Fig. 4.

(a) GABA dose–response curve (0.03–30 μM) using transiently transfected HEK293T cells expressing the α₁β₃γ₂ channel. Inset: Inward Cl⁻ currents are shown from one ensemble of cells exposed to GABA with increasing concentrations (0.03–30 μM). Current data of n=26 independent measurements in duplicate were normalized to highest response for each dose response used for fitting data to a four-parameter (variable slope) nonlinear regression equation. (b) Bicuculline dose–response curve (0.03–100 μM) using transiently transfected HEK293T cells expressing the α₁β₃γ₂ channel. Current data of n=8 independent measurements in duplicate were normalized to highest response for each dose response and used for fitting data to a four-parameter (variable slope) nonlinear regression equation. Inset: Inward Cl⁻ currents are shown from one ensemble of cells exposed to GABA (EC₈₀ concentration=10 μM) and increasing concentrations of bicuculline (0.03–100 μM). (c) HZ166 dose–response curve (0.03–60 μM) using transiently transfected HEK293T cells expressing the α₁β₃γ₂ channel. Current data of n = 8 independent measurements in duplicate were normalized to highest response for each dose response used for fitting data to a four-parameter (variable slope) nonlinear regression equation. The GABA concentration of 0.13 μM was potentiated 200% in the presence of HZ166. Inset: Inward Cl⁻ currents are shown from one ensemble of cells exposed to GABA (EC₂₀ concentration = 0.3 μM) with increasing concentrations of HZ166 (0.03–60 μM).

The Activity of Endogenous GABA_A Channel in Human Airway Smooth Muscle Cells Can Be Determined with the IonFlux

The expression and function of endogenous ionotropic GABA_A channels on airway smooth muscle cells was reported recently.^{11,12,28–30} Figure 5a demonstrates a GABA dose response (1–1,000 μM) eliciting current in human immortalized airway smooth muscle cells using the IonFlux platform. The determined EC₅₀ value was 34±10 μM. The average seal in these airway smooth muscle cells on IonFlux was ∼100 MΩ. About 12/16 human airway smooth muscle cell ensembles demonstrated a detectable current (defined as current >500 pA). The success rate of measuring these GABA-induced currents in an ensemble of cells was ∼75% as compared to manual whole cell patch clamping of single cells that had a success rate of 5% (data not shown).

Fig. 5.

GABA dose–response curve (1–1,000 μM) in cultured human airway smooth muscle cells. Current data of n=6 independent measurements in duplicate were normalized to highest response for each dose response used for fitting data to a four-parameter (variable slope) nonlinear regression equation. Inset: Inward Cl⁻ currents are shown from one ensemble of cells exposed to GABA with increasing concentrations (1–1,000 μM).

High Success Rates for GABA Responses in Stably Transfected (in Active Culture or RTA), Transiently Transfected, and Native Cells with Endogenous GABA_A Channel Expression Using the IonFlux Were Observed

A high average seal resistance for patch clamp assays is essential in order to measure changes in ion flux. Although seal resistances of higher than 1 GΩ were reported for manual patch clamp, similar values were observed for the automated patch clamp. Recombinant HEK293 cells, expressing GABA_A channels, exhibited seal resistances between 500 and 700 MΩ in whole cell (Fig. 6a). Transient transfected cells and native airway smooth muscle cells showed seal resistances between 60 and 100 MΩ. Nevertheless, high success rates were observed, defined as a measurable current of >500 pA (Fig. 6b). The difference in success rates of GABA-induced currents was most pronounced in native human airway smooth muscle cells endogenously expressing GABA_A channels. This automated microfluidic platform method of measuring current from an ensemble of cells yielded detectable currents in 75% of experiments compared with only a 5% success rate using traditional single whole cell manual recordings, representing a significant advance and throughput for current studies in these cells.

Fig. 6.

(a) Average seal resistance (MΩ) for different cell types expressing GABA_A receptors as measured on IonFlux^™. (b) The success rate of recording GABA-induced currents (defined as measurable current >500 pA) on IonFlux^™ from ensembles of 20 cells from different cell types expressing GABA_A receptors.

Discussion

Electrophysiology studies of GABA_A channels were performed on the IonFlux automated patch clamp system, in various cell expression formats including recombinant cell lines stably expressing a specific GABA_A channel (Millipore, α₁β₃γ₂), frozen RTA recombinant cells expressing two different subtypes of GABA_A channels (Millipore RTA, α₁β₃γ₂ and α₃β₃γ₂), transiently transfected HEK293T cells expressing the α₁β₃γ₂ GABA channel, and immortalized human airway smooth muscle cells endogenously expressing GABA_A channels. One GABA_A channel (α₁β₃γ₂) was tested in three different assay formats using cultured recombinant HEK293 cells, RTA-frozen recombinant stably transfected HEK293 cells, and transiently transfected HEK293T cells (Figs. 3a, 4c). The results for this particular subtype showed a high level of consistency for GABA-induced current kinetics profile with EC₅₀ values of 2.9±0.1 μM (stably transfected), 1.8±0.1 μM (RTA), and 1.3±0.29 μM (transiently transfected), respectively. Different GABA_A channels showed distinctive current desensitization, and potency profiles in response to GABA stimulation as expected given the different biologic function of the receptor subtypes. Dose–response studies of the GABA_A receptor antagonists bicuculline and picrotoxin and the allosteric modulators diazepam and HZ166 confirmed the correlation of responses between automated and manual patch clamp. Overall, the improved success rates observed for the IonFlux in comparison with manual patch clamp demonstrates that automation does not compromise quality.

HEK293 cells are useful for the performance for the automated patch clamp assay, because HEK293 cells can be easily prepared as a monodisperse suspension. Nevertheless, other cell lines such as primary cells have been successfully used. The IonFlux utilizes the ensemble recording with 20 cells patched per ensemble. Both current and seal resistance are averaged during the recording, significantly improving the success rate for assays where cells have heterogeneous expression levels, such as transiently transfected and primary cells.

Endogenous GABA_A channels expressed by primary cells provide the most physiological relevant information, but also present the most challenging assay format due to the low expression level and heterogeneity of channel expression. The results from endogenous GABA_A channels from immortalized human airway smooth muscle cells demonstrated that IonFlux system can measure GABA-dependent currents from endogenously expressed ion channels at a much higher success rate (75% for immortalized smooth muscle cells in a 20-cell ensemble, as compared to <5% success using manual patch clamp in single cells).

The EC₅₀ values for GABA are very different in transfected HEK cells overexpressing large numbers of pure populations of a given GABA_A receptor subunit combination compared to the GABA_A receptors endogenously expressed in native human airway smooth muscle cells. We believe that this difference is accounted for by the subunit composition of the GABA_A channel in human airway smooth muscle cells whose native stoichiometry and composition are not yet fully elucidated.¹¹ We were the first to describe GABA_A receptors on human airway smooth muscle cells and a complete mRNA survey of all known subunits demonstrated a limited repertoire of subunits including the α₄, α₅, β₃, γ₁, γ₂, γ₃, δ, π, and θ subunits.¹¹ However, the precise combinations of subunits that comprise the native GABA_A receptor in native human airway smooth muscle cells remain unknown. It is possible that compositions that include the atypical subunits (e.g., π and θ) may account for a decreased affinity for GABA in smooth muscle as has been demonstrated by the co-expression of the θ subunit with the α₂β₃γ₁ subunits in HEK cells, which demonstrated an increased GABA EC₅₀ value of 62 μM.³¹ Further support of this altered agonist affinity of GABA_A receptors in airway smooth muscle is supported by a recent publication¹² using the same cells in this same microfluidic platform demonstrating an EC₅₀ value of 15 μM for THIP (gaboxadol), which is 10-fold higher than EC₅₀ values reported for classic heterologously expressed GABA_A channels not containing an α₄ subunit and nearly 100-fold higher than EC₅₀'s for heterologously expressed GABA_A channels containing α₄ subunits.

The GABA_A receptors belong to an important family of ligand-gated ion channels, which are widely distributed in human central nervous system and serve as the predominant inhibitory channel. However, it is challenging to characterize the large number of subunit combinations expressed in GABA_A channels with traditional recombinant cell lines because only a few types of stably transfected cell lines expressing different GABA_A channels are available. Furthermore, it is very costly and time consuming to maintain these cells in culture. Commercially available frozen RTA cells are an excellent alternative, allowing researchers to determine ion channel activity without being equipped to maintain a large number of mammalian cells in culture, and provide larger, well-characterized production lots.

Transiently transfected cells, on the other hand, provide maximum flexibility in studying different ion channel subtypes, even those that are not available as stably transfected cells. The IonFlux system records from a 20-cell ensemble and features continuous microfluidics compound perfusion, which overcomes the challenges from expression heterogeneity, and achieves a better success rate. These features compensate for heterogeneous expression levels and reduce the total assay time as compared to other systems that require robotic pipetting for compound applications. The average seals across the ensemble varied between 100 and 600 MΩ, depending on the cell type used. This is due to both the substrate used, and the fact that resistances are in parallel and the overall resistance will be dominated by the lowest resistance in the 20-cell ensemble.

In conclusion, we have evaluated the performance of the IonFlux-automated patch clamp systems for use with a variety of GABA_A channel cell expression systems. Manual patch clamp is a recognized standard for obtaining electrophysiology data, but also a labor-intensive and low throughput approach, which usually requires years of user training to master; the presented approach increases throughput by over an order of magnitude and reduces the training period to weeks. For the cell lines tested, we also found the success rates to be increased considerably through the use of ensemble recording as compared to single cell recording.

We anticipate that automated patch clamp instruments will accelerate current electrophysiology research and provide us with the opportunity to better understand the distribution and function of ion channels. Additionally, these instruments will facilitate the screening of small molecules to identify new drug candidates. Finally, it makes ion channel research accessible to a broader spectrum of researchers, who may otherwise be discouraged by the challenges of traditional patch clamping techniques.

Abbreviations

DMSO

dimethyl sulfoxide

EC₅₀

compound concentration at which ion channel current is 50% of the maximum response

FBS

fetal bovine serum

GABA

γ-amino butyric acid

HEK293

human embryonic kidney

hTERT

human telomerase reverse transcriptase

IC₅₀

compound concentration at which ion channel current is 50% of the current in the absence of compound

MEM/EBSS

minimum essential medium with Earle's balanced salts

PBS

phosphate-buffered saline

RTA

ready-to-assay

Acknowledgments

The α₁, β₃, and γ₂ plasmids were kindly provided by Prof. Dr. Werner Sieghart (Department of Biochemistry and Molecular Biology of the Nervous System, Medical University of Vienna, Austria).

Disclosure Statement

J.M.C., N.Y., and L.A.A. are employees of the University of Wisconsin–Milwaukee. Their work was supported by the National Institutes of Health NIMH MIL100371. P.D.Y. was supported by NIH GM008464 and C.W.E. was supported by GM065281. Q.C., J.J., S.S., and C.I.Z. are all employees of Fluxion Biosciences.