Sage Journals: Discover world-class research

Abstract

The epithelial sodium channel (ENaC) plays a crucial role in salt and water homeostasis and is primarily involved in sodium reabsorption in the kidney and lung. Modulators of ENaC function, particularly within lung epithelia, could offer potential treatments for a number of diseases. As a constitutively active sodium channel, ENaC expression at the cell membrane is highly regulated through rapid turnover. This short half-life of the channel at the membrane and cytotoxicity from overexpression pose a problem for reagent generation and assay development in drug discovery. We have generated an HEK293 stable cell line expressing ENaC β and γ subunits containing the PY motif trafficking mutations found in Liddle’s syndrome to overcome rapid channel turnover at the membrane. A BacMam virus was used to transiently express the ENaC α subunit to reconstitute channel function to reduce the toxicity associated with long-term overexpression. We have configured a 384-well FLIPR membrane potential antagonist assay for high-throughput screening and an IonWorks Quattro electrophysiology antagonist assay that is predictive of potency values derived from primary lung epithelial cell short-circuit measurements. The triage strategy for compound screening and profiling against this target using these assays has resulted in the discovery of novel chemotypes.

Keywords

epithelial sodium channel FLIPR IonWorks

Introduction

The epithelial Na⁺ channel (ENaC) is a member of the ENaC/degenerin ion channel family.¹ ENaC is located on the apical membrane of polarized epithelial cells in the lung, distal nephron, distal colon, sweat ducts, salivary glands and taste buds, where it plays a key role in the reabsorption of Na⁺.^1–4 It is critical in maintaining Na⁺ and K⁺ haemostasis¹ as well as the ionic content, volume, and osmolarity of extraepithelial fluids in the lung.⁵

ENaC is composed of three homologous subunits, termed α, β, and γ, that are encoded by distinct genes.⁶ Although functional channels can be formed by assembly of α subunits alone, β and γ are required for maximal expression and relevant channel activity on the cell surface.⁶ The precise structure of the native channel remains unresolved, but based on crystallographic studies of the homologous ASIC1 channel, ENaC is believed to have 1:1:1 α, β, γ subunit stoichiometry.^7,8 Importantly, ENaC is constitutively active, so wherever the channel is functionally expressed, it continuously conducts Na⁺. The surface density of ENaC, and hence the transmembrane Na⁺ flux, is under a complex hormonal control involving transcriptional and post-transcriptional regulation of its synthesis and insertion as well as retrieval of active channels from the apical membrane.⁹ Under normal conditions, ENaC has a membrane half-life of less than 1 h, because of binding of ubiquitin ligase Nedd4-2 to C-terminal domains of the channel, which target the protein for lysosomal degradation.^10,11

Mutations in the C-terminus PY motif of the β and γ subunits disrupt this interaction and lead to accumulation of active channels on the cell surface and a subsequent increase in Na⁺ reabsorption and K⁺ loss.¹² These mutations are clinically manifested as autosomal dominant Liddle’s syndrome, characterized by elevated blood pressure, hyperkalemia, and pseudohyperaldosteronism.¹³ Such gain-of-function mutations in ENaC may also contribute to the pathophysiology of cystic fibrosis.¹⁴ Conversely, loss of function mutations in ENaC cause autosomal recessive severe systematic pseudohypoaldosteronism (type 1: PHA-1), which is characterized by hypotension with renal insensitivity to aldosterone.¹³ Accordingly, modulators of ENaC channels may have useful therapeutic benefit. Cystic fibrosis is most widely pursued through mucus hydration and clearance by aerosol delivery of channel inhibitors delivered directly to the lung.^15–19

To date, efforts to identify novel channel modulators have been hampered by the availability of suitable in vitro screening methods. Although both conventional patch clamp electrophysiology and short-circuit current measurements in epithelial cells using Ussing chambers can yield highly relevant functional measures of ENaC, each lack the throughput and ease of use to support wider-scale screening and compound profiling. Here, we report the development, optimization, and validation of novel fluorescence (FLIPR membrane potential [V_m]) and automated patch clamp electrophysiology assays for human ENaC. In combination, these assays enable high-throughput and relevant screening for novel ENaC modulators.

Materials and Methods

Compounds and Materials

Three standard ENaC small-molecule inhibitors, amiloride, benzamil, and phenamil, were purchased from Sigma Aldrich (St. Louis, MO). A series of novel ENaC blockers first described by Collingwood and Smith²⁰ and Hirsh et al.^15,16 were synthesized by GSK Medicinal Chemistry or obtained from the GSK compound collection. All compounds were initially solubilized in 100% DMSO to a concentration of 10 mM, with the exception of amiloride (30 mM DMSO stock for FLIPR assays). Compounds were further diluted in buffer before addition to the assay. Membrane potential indicator dyes and other FLIPR reagents, as well as all cell culture materials, were purchased from Invitrogen (Carlsbad, CA), unless stated otherwise.

Recombinant Expression of ENaC in HEK293 Cells

ENaC subunit cDNAs were amplified from human kidney (α and β) and human brain (γ) marathon libraries. The sequences are identical to GenBank references NM_001038, NM_000336, and NM_001039, respectively. The α subunit cDNA fragment was placed under the control of the CMV promoter in the pBacMire shuttle vector and transfected into SF9 insect cells, and BacMam viruses were generated as previously described by Condreay et al.²¹ The β and γ subunits were subcloned into pBacMire (neomycin/geneticin resistance) and pBacMireHyg (hygromycin resistance) respectively, under the control of the CMV promoter. HEK293 cells were co-transfected with β and γ subunit constructs, and stable cell lines were established under the selection of 0.5 mg/mL geneticin and 0.2 mg/mL hygromycin. Cells were maintained in DMEM Ham’s F12 medium supplemented with glutamax and 10% fetal bovine serum (FBS), at 37 °C with 5% CO₂, dissociated with TrypLE, and split 1/5 to 1/10 for propagation. Mutations in the β (from CCC-AAC-TAT to GCC-AAC-TTA, corresponding to P618A Y620L) and γ (from CCC to TAG, corresponding to P624Stop) subunits (ENaC β [P618A Y620L; β*] and γ*: ENaC γ [P624Stop; γ*], respectively) were created using the Stratagene QuikChange XL Site-Directed Mutagenesis Kit (cat. No. 200516) according to the manufacturer’s instructions. HEK293 ENaC βγ and β*γ* stable cell lines were maintained under a constant selection of 0.5 mg/mL geneticin and 0.2 mg/mL hygromycin. For functional experiments, the α subunit was introduced to form the full ENaC channel via a BacMam virus added 18 to 24 h prior to the start of the experiment.

FLIPR Membrane Potential Assay

Cells expressing ENaC β and γ subunits were plated onto 384-well, black, clear-bottomed assay plates (Greiner, Monroe, NC) for 18 to 48 h at 37 °C in culture media supplemented with ENaC α subunit BacMam virus (multiple of infection [MOI] = 2.4) at a density of 15,000 cells per well. The medium was then aspirated using a Tecan Plate Washer, leaving 10 µL residual volume in each well. Twenty microliters of FLIPR membrane potential dyes from Molecular Probes (Eugene, OR) (Blue: FMP [cat. No. R8042] or FMP2 [cat. No. F8181]; Red [cat. No. R8126] or DiBac2 [cat. No. B438]), diluted at the manufacturer’s recommended concentrations in Tyrode’s buffer (Sigma cat. No. T2145 supplemented with sodium bicarbonate: in mM, NaCl 136.9, KCl 2.7, H₂NaO₄P 0.4, CaCl₂ 1.8, MgCl₂ 1.1, D-glucose 5.6, NaHCO₃ 11.9; pH adjusted to 7.35 with NaOH), was then added to each well using a Multidrop and incubated at room temperature for 1 h. Compound plates with 384 wells were prepared by adding 50 µL of Tyrode’s buffer containing 0.04% pluronic acid into each well preloaded with 1 µL of 1/3 serial dilution of compounds in DMSO. 10 μL of the diluted compounds was transferred to each well in the assay plate using the dispensing head of a FLIPR (Molecular Devices) while simultaneously monitoring whole-well fluorescence (excitation/emission λ 510–545/565–625 nm). Amiloride (150 µM) and DMSO were used as high (100% channel block) and low (vehicle) controls, respectively. The kinetic changes in fluorescence intensity following compound addition were recorded using SoftMax Pro software. The average steady-state fluorescence intensity values were measured before and after drug addition and exported for data analysis. Curve fitting and –lg(IC₅₀ at [M] concentration) (pIC₅₀) determination were carried out using a four-parameter logistic model (Excel XC50 Module). For Na⁺ add-back experiments, Na⁺-free buffer contained (mM) N-methyl-D-glucamine 132, KCl 2, CaCl₂ 2, MgCl₂ 1, HEPES 10, and D-glucose 5, pH 7.3 with HCl.

IonWorks Planar Array Electrophysiology

HEK293 ENaC β*γ* stably expressing cells were seeded at ~3 × 10⁶ cells in T175 cm² flasks and grown for 18 to 24 h. ENaC α subunit BacMam was added at an MOI of 25, and cells were maintained for a further 18 to 24 h in medium containing 70 mM NaCl to minimize channel-mediated cellular toxicity (50:50 mix of standard growth medium and DMEM/F12 without NaCl, supplemented with 10% FBS).

Planar array electrophysiology recordings were made on an IonWorks Quattro instrument using both single-cell (HT) and population patch-clamp (PPC) configurations.^22–26 Cells were prepared immediately prior to experimentation as previously described²⁶ and resuspended in warm external buffer (in mM, Na-gluconate 120, KCl 5, NaCl 20, CaCl₂ 1, MgCl₂ 2, HEPES 10, pH adjusted to 7.35 with NaOH, 0.2 µm filter sterilized) to yield a final cell density of ~2 × 10⁶ cells per mL. Internal buffer was composed of (mM) K-gluconate 120, NaCl 5, KCl 20, HEPES 10, CaCl₂ 2, MgCl₂ 1, pH adjusted to 7.2 with KOH (0.2 µm filter sterilized). Cell perforation was achieved using internal buffer supplemented with 0.2 mg/mL amphotericin B.

Individual cells (or wells for PPC) were voltage clamped and the test protocols applied in groups of 48 using the electronics (E-) head.²⁶ A holding potential (V_m) of −40 mV was applied for 5 s followed by a ramp (250 ms, −100 mV to +100 mV with a 20 ms hold at each extremity) and step (50 ms, 0 mV; 50 ms, +30 mV) protocol. The corresponding currents were recorded first in the absence (pre-) and then the presence (post-) of amiloride, test compound, or vehicle (compound incubation time of 3–5 min). To reduce any potential compound carryover effects due to compound sticking to the fluidics (F-) head pins, the 50% DMSO solvent wash facility was used in these experiments. Leak subtraction was disabled, and currents were measured either at 0 mV (where the contribution of nonbiological leak is negligible²⁴) or −100 mV (and referenced to 100% block amiloride controls in PPC). Curve fitting and pIC₅₀ determinations were carried out by fitting to a four-parameter logistic equation (Microsoft Excel XC50 Module).

Ussing Chamber Short-Circuit Current Recording

Human airway epithelium cultures (HEAC) grown in air-liquid interface (AIR-100-SNP, 12 tissues in Snapwell inserts) were purchased from MatTek Corp (Ashland, MA). On the day of use, Snapwell inserts containing EpiAirway tissue (1 cm² diameter) were placed in Ussing chambers containing Krebs bicarbonate buffer (KBS; composed of [mM] NaCl, 118; KCl, 5.4; NaH₂PO₄, 1.2; MgSO₄, 1.2; CaCl₂, 2.5; NaHCO₃, 25.0; dextrose, 11.1, and gassed with 95% O₂ and 5% CO₂, pH 7.4) for bioelectric measurements. Throughout the experiment, each half chamber contained 5 mL of KBS maintained at 37 °C and bubbled with 95% O₂/5% CO₂. Tissues were equilibrated under open-circuit conditions until transepithelial potential was stable (typically within 45 min) with washing every 15 min.

Bioelectric measurements were made with an epithelial voltage clamp (model VCC MC6) and a self-contained Ussing chamber system (both purchased from Warner Instruments, Inc., Hamden, CT), digitized by a DataQ DI-720 series real-time Data Acquisition System (Physiologic Instruments, San Diego, CA) and displayed and recorded using Acquire and Analyze 2.3 Software (Physiologic Instruments, version 2.3.177). Once transepithelial potential was stable, transepithelial voltage was clamped at zero, and the resulting short-circuit current (I_sc ) was recorded prior to and following the addition of ENaC blocker compounds that were added in a series of cumulative increasing concentrations (threefold increments). ENaC blockers were prepared as a 10 mM stock solution in DMSO and stored at −20 °C. Working solutions were prepared daily as a series of 10-fold dilutions in KBS. Data analysis was carried out using GraphPad Prism (version 4.03).

Results

Pharmacology of ENaC in Human Airway Epithelia

Prior to assembling high-throughput assays, the pharmacology of ENaC was determined by short-circuit current (I_SC) measurements in HEAC grown at an air-liquid interface.²⁷ We considered this HEAC assay the gold standard because human ENaC channels, presumably composed of relevant subunits and auxiliary regulatory proteins, are endogenously expressed under conditions that most closely resemble the environment in native polarized epithelia. The absolute and relative potencies of 17 known inhibitors (amiloride, benzamil, phenamil, published and patent literature compounds) were quantified and used as the target pharmacology profile to which we attempted to align (correlate) higher-throughput assays. Under the experimental conditions, HEAC exhibited large, stable I_SC of 31.4 ± 12.7 µA/cm² (n = 45). Addition of amiloride (100 µM) reduced the I_SC by >95% to 1.4 ± 1.6 µA/cm² (n = 6). In cumulative concentration-response curve assays, all 17 inhibitors produced a substantial, concentration-dependent reduction in I_SC ( Fig. 1 ). The pIC₅₀ value for amiloride was 6.47 ± 0.03 (n = 6, Hill slope = 0.97) with a maximum percentage inhibition of 96.0 ± 5.0 (n = 6) that was shown to be reversible on washout (86.3% ± 6.3%, n = 6). These data were consistent with previously published values.²⁸ The ENaC inhibitor Novartis-5 (mean pIC₅₀ values of 8.7 ± 0.04, n = 2, Hill slope = 0.72, maximum inhibition = 93.6% ± 2.7%, reversibility = 14.5% ± 0.3%) was >100,000-fold more potent than the least potent compound trimethoprim (pIC₅₀ = 3.46, n = 1, Hill slope = 0.75, maximum inhibition = 68.4%, reversibility = 80.6%).

Figure 1.

Ussing chamber recordings of human airway epithelium. (A) Representative trace of apical membrane sodium short-circuit current (I_sc) measured in Ussing chambers on human airway epithelium in the presence of increasing concentrations of the epithelial sodium channel (ENaC) blocker amiloride. Each concentration was applied for approximately 1 min. Recordings were performed at 37 °C. (B) Concentration-response curve analysis, where the x-axis is the log of amiloride concentration (molar) and the y-axis the response (expressed as percentage maximum inhibition). Each point represents a single test occasion (n = 6). The pIC₅₀ value for amiloride was 6.47 ± 0.03 with a Hill slope of 0.97 observed. (C) Summary table of some of the exemplar ENaC inhibitors tested in the Ussing chamber.

Analysis of ENaC Activity in Lung- and Kidney-Derived Epithelial Cell Lines

As ENaC is known to be present in lung and kidney epithelia, we first examined a range of lung- and kidney-derived epithelial cell lines for the possibility of developing high-throughput assays using endogenously expressed ENaC. These cell lines included human bronchial epithelia (16HBE 14o- [kindly provided by Dr. D. C. Gruenert, University of California, San Francisco, California, USA] and Beas 2B [ATCC CRL-9609]), human lung carcinoma A549 (ATCC CRL-185), human lung adenocarcinoma NCI-H441 (ATCC HTB-174), human epithelial kidney HEK293 (ATCC CRL-1573), and dog kidney MDCK (ATCC CCL-34). Amiloride-sensitive whole-cell currents were measured using single-hole planar array electrophysiology but were found to be too small to conduct detailed pharmacology studies upon: the largest signal obtained with H441 cells was <150 pA at 0 mV ( Table 1 ). Efforts to increase the functional expression by pretreating cells with dexamethasone, forskolin, or 1-methyl-3-(2-methylpropyl)-7H-purine-2,6-dione were unsuccessful (data not shown).

Table 1.

Characterization of lung- and kidney-derived epithelial cell lines for ENaC activity measured by high-throughput automated electrophysiology.

Cells	Mean Current at 0 mV (pA)	Number of Cells Tested	Success Rate (% Cells)	Mean Seal Resistance Pre (MΩ)	Mean Seal Resistance Post (MΩ)
16HBE14o-	45.6 ± 5.1	1536	48.6 ± 8.9	118.9 ± 7.2	116.2 ± 11.3
Beas2B	52.7 ± 5.7	768	33.9 ± 8.9	120.2 ± 3.2	113.6 ± 13.9
A549	37.4 ± 3.9	4224	35.2 ± 3.3	155.8 ± 6.9	151.4 ± 7.7
H441	147.5 ± 6.5	768	49.5 ± 3.1	70.3 ± 2.7	80.2 ± 0.6
HEK293	48.5 ± 2.6	256	96.1 ± 3.1	123.8 ± 2.2	112.7 ± 4.7
MDCK	60.4 ± 81.0	256	96.9 ± 3.1	69.0 ± 1.1	66.0 ± 0.4

The number of cells tested equals the total number of wells. Mean seal resistances pre- and post-30 µM amiloride addition were recorded. Success rate was calculated as the percentage of cells sealing at values greater than 30 MΩ. Current responses were recorded at 0 mV holding potential and calculated by subtracting post- from pre-amiloride addition.

Functional Expression of Recombinant ENaC

We next explored expression of human recombinant ENaC in mammalian host cells. As the channel is constitutively open, long-term overexpression is likely to be associated with cellular toxicity due to excess Na⁺ influx. Thus, an HEK293 stable cell line co-expressing the β and γ subunits was cultivated, and the fully functional channel constituted by transient introduction of the α subunit via BacMam viral transduction (for 18–24 h prior to the functional assay). To validate this expression strategy, ENaC activity was initially assessed by determining the magnitude of the hyperpolarization response to amiloride in a fluorescence-based membrane potential assay on the FLIPR ( Fig. 2A – D ). As anticipated, amiloride did not hyperpolarize HEK293 cells stably expressing either βγ or the α subunit alone (via BacMam), beyond the signal observed in untransfected cells. Introduction of the α subunit to βγ-expressing cells produced only a small increase in the amiloride response, which we ascribe to rapid degradation of the wild-type proteins.¹¹

Figure 2.

(A–D) Comparison of the activity of subunit combinations of epithelial sodium channel (ENaC) activity in HEK293 cells by FLIPR using FMP Blue dye. The FMP Membrane Potential Assay Kit can detect changes in potential across the cellular membrane. (A) ENaC is constitutively open, causing an influx of sodium ions into the cells, leading to membrane potential depolarization and an increase in dye fluorescence. (B) Following the addition of ENaC blockers, the membrane potential repolarizes (going down) and dye fluorescence decreases. (C) Time course of concentration-dependent amiloride-evoked fluorescent intensity changes (Δ relative fluorescent units [RFUs]) in the FMP Blue FLIPR assay with HEK293 β*γ* stable cell line transduced with α subunit BacMam for 24 h. (D) Fluorescence intensity (RFUs) of ENaC subunit combinations in response to amiloride (31 µM) and trypsin (12.5 µg/mL). Host HEK293 alone, ENaC βγ or β*γ*-expressing stable cell lines were transduced with ENaC α subunit BacMam for 24 h. Trypsin or buffer alone was added to the wells 5 min prior to online amiloride addition. Basal values (open bars) were taken before amiloride addition. The signal window (filled bars) was calculated by subtracting the basal value of the same well from averaged plateau phase values following amiloride addition. (E–H) Optimization of FLIPR assay. FLIPR assays were performed using FMP2 Blue with an HEK293 β*γ* stable cell line plus α subunit BacMam. (E) Effect of viral titration on signal window (bars, calculated by subtracting basal values of the same well from steady-state values following amiloride addition) and Z′ (line). (F) Effect of viral titration on amiloride concentration-response curves. Note the right shift of the curves in response to a higher viral load. (G) Effect of viral titration on the pIC₅₀ values of a set of known ENaC inhibitors (represented by the different shapes). A multiple of infection of 2.4 was selected for all further experiments. (H) Robustness single short tests of two replicates of >1000 randomly selected compounds (blue) using the FLIPR assay, with DMSO low control (yellow) and 150 µM amiloride high control (cyan).

To overcome this, mutations were introduced to the PY motif of β (P618A Y620L) and γ (P624Stop) subunits (referred to as β* and γ*, respectively) to mimic those found in Liddle’s syndrome, which have been shown to significantly increase the cell surface expression and stability of ENaC.¹¹ Expression of the α subunit into a β*γ* stably expressing cell line yielded significantly higher basal fluorescence than observed with wild-type subunits, indicating an ENaC-mediated cellular depolarization. Consistent with this, the amiloride-evoked hyperpolarization was >2-fold higher in αβ*γ* compared with αβγ-expressing cells ( Fig. 2D ).

Interestingly, despite the apparent low functional activity of ENaC αβγ, the basal fluorescence and amiloride-induced hyperpolarization were dramatically enhanced when cells were treated with trypsin, a serine protease enzyme. Similar observations were made with α β*γ* cells ( Fig. 2D ). Proteolytic treatment of ENaC increases channel open-state probability by releasing Na⁺-“self-inhibition,” and PY motif mutations further enhance this effect.^29,30 We did not pursue this approach further because preliminary studies indicated that trypsin treatment attenuated the effects of ENaC blockers (data not shown) and would likely yield discrepant pharmacology values.

To verify functional activity and basic biophysics of ENaC, α β*γ* automated planar array electrophysiology methods were applied prior to subsequent FLIPR assay optimization and screening (discussed later in this article).

Development and Optimization of ENaC FLIPR Membrane Potential Assays

A range of membrane-potential sensitive fluorescent dyes were evaluated as suitable reporters for high-throughput (384-well) FLIPR assays. Pilot studies with DiBAC and FMP Red were unsuccessful because of a high background fluorescence and low amiloride-sensitive signal window in ENaC α β*γ* HEK293 cells. With the FMP Blue kit, clear hyperpolarizing responses to amiloride were observed, but the potency (mean pIC₅₀ value 5.22) was 24-fold lower than observed in the HEAC Ussing chamber assay (mean pIC₅₀ value of 6.58). Strikingly, the highly potent inhibitor Novartis-5 (HEAC pIC₅₀ 8.7) was >1000-fold less active in this FMP Blue membrane potential assay (pIC₅₀ <5). Switching to a second-generation FMP2 Blue Kit produced some improvement. Although the measured potencies of amiloride and phenamil remained unchanged, we observed a large increase in the effect of Novartis-5 (pIC₅₀ 6.6 FMP2) and some other ENaC blockers ( Fig. 3 ). Interestingly, the FMP Blue and FMP2 Blue kits differed in the results obtained in Na⁺ substitution and add-back experiments. When switching from Na⁺-containing Tyrodes to an NMDG-based Na⁺-free solution, an amiloride-evoked reduction in fluorescence (hyperpolarization) was observed with FMP ( Fig. 3A , B ) but not FMP2 ( Fig. 3C , D ). Conversely, when going from Na⁺-free to Na⁺-containing (Na⁺ add-back protocol), a far greater depolarizing response was observed with FMP2 compared with FMP (~15,000 vs ~5000 fluorescence units, corresponding to ~100% vs ~30% increase over basal reading obtained prior to Na⁺ addition). The explanation of this is presently unclear as the complete kit formulations are undisclosed by the vendor. Inasmuch as the FMP2 kit membrane potential data displayed the anticipated profile of ENaC regarding Na⁺ modulation and blocker effects (albeit with potency underestimates; Fig. 3E – H ), we used this kit for all subsequent experiments.

Figure 3.

Comparison of dyes for the FLIPR assay using HEK293 β*γ* stable cell line with α subunit BacMam. Fluorescence (relative fluorescent units) was recorded online by FLIPR. Cells were loaded with FMP Blue (A,B) and FMP2 Blue (C,D). Fluorescence values were subtracted by the average basal reads of the same well before the online addition of Tyrode’s buffer containing 31 µM amiloride (amiloride) or an equal amount of DMSO to cells preincubated in Tyrode’s buffer; Tyrode’s buffer containing 31 µM amiloride to epithelial sodium channel β*γ* cells without α subunit BacMam (Ctrl cells); or Na⁺-free buffer in the presence (Na⁺ add-back) or absence (Na⁺ free) of 31 µM amiloride to cells preincubated with Na⁺-free buffer. (E, F, and G) Exemplar concentration-response curves to amiloride, phenamil, and Novartis-5 for FMP Blue (circles) and FMP2 Blue (triangles), with pIC₅₀ values labeled next to the curves. (H) Correlation of pIC₅₀ values obtained with FMP2 Blue (filled circles) and FMP Blue (open circles) with HEAC I_SC.

In an attempt to address the underestimation of ENaC-blocking activity, we considered the nonlinearity of channel block with the change in membrane potential. Theoretically, with high levels of functional activity of ENaC, a substantial proportion of channels will need to be blocked before the membrane will begin to hyperpolarize. To explore this, we titrated the expression levels of ENaC by modifying the multiplicity of infection of the BacMam for the α-subunit. Across the range of 7.12 to 1.2 MOI, we observed a clear increase in compound potency as the MOI was reduced (i.e., fewer channels were expressed; Fig. 2E – H ). There was little change in overall assay signal window (relative fluorescent unit [RFU]) or assay performance (Z′) between MOI 7.12 and 2.4. Below an MOI of 2.4 the assay quality parameters deteriorated. From this, an MOI of 2.4 was selected as the optimal condition for sensitive pharmacology and assay performance. For the standard blockers, the final format of the FLIPR membrane potential assay underestimated the true ENaC-blocking potency by an average of 17-fold ( Fig. 3H ).

As a final test of robustness, four plates of test compounds (352 samples per plate at a final assay concentration of 10 µM, 16 high [150 uM amiloride], 16 low [0.5% DMSO] controls) were tested in duplicate. The Z′ value was 0.80 ± 0.05 (SD, n = 8), and the activity of the test samples was centered on 0% inhibition for both replicates ( Fig. 2H ).

Development and Optimization of ENaC IonWorks Patch Clamp Electrophysiology Assay

Taking learnings from assays for other non–voltage-gated channels,^24,26 we next developed a planar array patch clamp electrophysiological assay for ENaC using IonWorks Quattro. In single-hole experiments on control HEK293 β*γ*-expressing cells, grown in standard media without the α-subunit, the measured seal resistance was 197.1 ± 63.6 MΩ (n = 366). A voltage step to −100 mV (20 ms holding prior to ramp) produced a small inward current of 320 ± 140 pA (n = 356) that was relatively unchanged in the presence of amiloride (data not shown). Figure 4A shows that the voltage ramp (200 ms, −100 to +100 mV) produced a small linear current (615 pA at −100 mV in this example), reversing close to 0 mV, which was insensitive to amiloride (10 µM; <10% block at −100 mV). This current was similar to that observed in wild-type (untransfected) cells and is likely a (contaminating) non-ENaC leak current ( Table 1 ). In the presence of the co-expressed α-subunit (BacMam), functional ENaC channels can be formed. Figure 4B shows that the voltage ramp (200 ms, −100 to +100 mV) produced a much larger linear current (2712 pA or 2.71 nA at −100 mV), reversing between +20 and +30 mV, which was highly sensitive to amiloride (10 µM; 76% block at −100 mV). The ENaC current, measured by subtracting the amiloride-resistant current from the control (no-amiloride) signal, displayed inward rectification and reversed toward the Na⁺ equilibrium potential (+58 mV under the ionic conditions). When compared in HT mode, large, inward amiloride-sensitive currents (mean value of −0.91 ± 1.26 nA; n = 153, values ranging from −6.62 to +1.06 nA) were observed in some but not all cells ( Fig. 4C ). In these cells, the initial seal resistances were highly variable but overall lower ( Fig. 4D ; mean HT value of 58.4 ± 50.8 MΩ, n = 191). Application of amiloride markedly increased the (membrane) resistance in cells that had low initial values (see Fig. 4D ; mean HT value 123.7 ± 57.5 MΩ, n = 190), indicating constitutive ENaC activity.

Figure 4.

Population patch-clamp planar array electrophysiology recordings of epithelial sodium channel (ENaC)–derived currents in human recombinant ENaC αβ*γ*-expressing HEK cells. (A,B) Representative current-voltage relationship (200 ms, –80 to +80 mV) obtained before (black) and after (gray) amiloride (10 µM). The amiloride-sensitive current is indicated by a dashed line. In the absence of the alpha (β*γ* only) subunit, no ENaC-derived current is observed (A). When the alpha subunit is expressed (αβ*γ*), an inwardly rectifying relationship detected is consistent with an ENaC-mediated current. The E_rev was calculated to be ~+33 mV. (C) Population histograms for the amiloride-sensitive currents (I_pre-post) recorded at −100 mV in HEK cells expressing ENaC αβ*γ* cultivated in normal growth medium by single hole (left) and population patch-clamp (PPC; middle) or in low-sodium medium by PPC (right). Histograms were generated by measuring responses in 153 to 191 wells, which were then binned at 0.1 nA and counted. Mean current values are indicated on the graphs. Note the tighter current distribution in PPC when compared with single hole. When the cells were cultured in low-sodium medium, the variance in current distribution improved. (D) Seal resistance determined from the pre- and post-amiloride measurements in HEK cells expressing ENaC αβ*γ* cells cultivated in normal medium for single hole (left) and PPC (middle) and in low sodium medium for PPC (right). Data were generated by measuring responses in 153 to 191 wells (same as above). Data points of individual wells before and after amiloride addition were connected by lines. (E) Concentration-response curve analysis, where the x-axis is the log of amiloride concentration (molar) and the y-axis the response (expressed as percentage of maximum inhibition). Each point represents an average of up to four wells (quadruplicate, tested in pseudo-96 on a 384-well plate) in PPC mode, low-sodium growth medium. A pIC₅₀ value of 7.11 ± 0.06 and Hill slope of 0.87 were observed for the averaged data. (F) Summary table of exemplar ENaC inhibitors tested in IonWorks Quattro using final assay conditions (PPC, low-sodium growth medium).

To overcome the large cell-to-cell variability in ENaC current amplitude, further studies were conducted in PPC mode in which the ensemble (average) current from up to 64 cells per well was measured. We also explored whether reducing the Na⁺ concentration in the growth/transfection medium for the HEK293 αβ*γ* cells would improve assay quality, hypothesizing that a high basal ENaC-mediated Na⁺ flux may be detrimental to cell health for electrophysiology recordings. When cultivated in standard sodium-containing culture media, the mean seal resistance of the cells in PPC was 29.5 ± 4.0 MΩ, n = 192, and was increased to only 49.5 ± 11.0 MΩ (n = 192) in the presence of amiloride ( Fig. 4D ). After culturing in low-sodium-containing media (half that of standard culture media), improved seal resistances (46.2 ± 6.0 MΩ and amiloride = 77.8 ± 16.4 MΩ [n = 191] for basal and amiloride recordings, respectively) were observed, likely owing to improved cell health. The amplitude of the ENaC-specific currents was not increased under low-sodium conditions; however, the consistency was greatly improved across the test plate (see Fig. 4C , Gaussian curve fitting).

Figure 4E shows the pharmacology of amiloride recorded under final assay conditions (PPC, low sodium) and plotted as percentage of maximum control (10 µM amiloride). The pIC₅₀ value for amiloride was 7.24 ± 0.08 (n = 11) with a Hill slope of 1.08. The pharmacology of amiloride was unchanged in the presence of low-sodium versus normal media conditions (data not shown). The pharmacology of other known ENaC blockers is summarized in Figure 4F . Because of the variability of the assay, each data point was tested in quadruplicate on the plate (pseudo 96-well plate format). Average robust Z′ values of >0.1 were observed for this assay, with a cutoff of >0 applied to data for screening. All compounds were tested with a final assay concentration of 1% DMSO. This assay has previously been shown to be tolerable of ≤2% DMSO final concentration (data not shown).

Comparative Pharmacology, Hit Identification, and Triage

Using this IonWorks assay format, potency values for a set of known ENaC blockers were obtained and correlated well with that from the Ussing chamber (R² = 0.81), without a noticeable general shift in potency ( Fig. 5A ). Correlation with FLIPR was also good (R² = 0.83; Fig. 5C ). The FLIPR assay understated the pIC₅₀ values for the more potent compounds, by about 1 log unit ( Fig. 5B , C ).

Figure 5.

Correlation plots for assay comparison. (A) IonWorks pIC₅₀ values showed excellent correlation with that of the Ussing chamber with a slope of 1.26 and an R² of 0.81. (B) Correlation of the FLIPR pIC₅₀ data with that of the Ussing chamber generated a slope of 0.90 and R² of 0.45. The lower R² value is due to the poor spread of pIC₅₀ data for correlation. (C) Correlation of the FLIPR pIC₅₀ data with IonWorks yielded a slope of 0.57 and an R² value of 0.83. Note the gradual reduction of pIC₅₀ values for higher-potency compounds determined by FLIPR. (D) Results of focused screening of >2000 compounds (including some known ENaC inhibitors) tested at a single concentration of 10 µM in duplicate by FLIPR and IonWorks Quattro and the values compared. Compounds with >30% inhibition in the IonWorks assay are labeled in gray, and some were tested in the full curve. The pIC₅₀ values from IonWorks are indicated by circle size (larger = more potent).

We next compiled a focused set of ~2000 compounds based on homologies to several known ENaC blockers and profiled with the FLIPR and IonWorks assays in parallel initially in single-well format at 10 µM. pIC₅₀ values for the hits were determined in subsequent experiments. The results are illustrated in Figure 5D . When 30% cutoff of response was applied, a hit rate 1.8% were obtained using IonWorks. Most of these hits were also positive in the FLIPR assay using the 30% cutoff, representing 1.4% of the total compounds, and this was increased to 1.7% when the 20% cutoff was applied. This result shows that the FLIPR V_m assay is capable of detecting most of the genuine hits but also reported a high rate of hits unconfirmed by IonWorks (10.9% and 14.4% with 30% and 20% cutoff, respectively), perhaps as a result of lower assay sensitivity and specificity.

Discussion

Owing to its rate-limiting role in regulating Na⁺ absorption and Na⁺/K⁺ balance, ENaC is clearly a potential point for therapeutic intervention for a number of diseases and has long attracted interest from academics and pharmaceuticals alike. It is therefore rather surprising that, to our knowledge, no high-throughput FLIPR or electrophysiological assay for this channel has been documented. An obvious answer might be the rather low level of channel activity of native epithelial cell lines that express wild-type ENaC ( Table 1 ). A high level of and uniform expression of ion channels is critical for the development of robust high-throughput assays, particularly planar array electrophysiology such as IonWorks.³¹ Our approach has addressed these difficulties in several ways. First, the β and γ subunits were expressed in a stable cell line using IRES, a technology that enables uniform high-level expression in stable cell lines.³¹ Second, mutations in the β and γ subunits that mimic those found in Liddle’s syndrome were introduced, which are known to prevent fast internalization and degradation as well as to increase the open probability of the channel.³⁰ Third, functional channels were reconstituted only transiently following the addition of α subunit BacMam, to avoid long-term functional overexpression, which may cause cell toxicity due to excessive cell swelling.

The mutations introduced to the PY motif of ENaC β and γ subunits did not appear to affect compound potency values generated by IonWorks. The values were close to that by measuring I_SC of HEAC ( Fig. 5A ), despite the large differences in the level of channel expression, cell background, and the technology between the two platforms. This is likely due to the fact that the PY motif is close to the extreme C-termini of the proteins, which are intracellular and far away from the channel pore,⁸ and are mainly involved in the regulation of trafficking rather than the biophysical properties of the channel. Existing tool compounds such as amiloride interact directly with the pore without the need to enter the cell.^8,32

Despite the difference in functional FLIPR responses under normal assay conditions, trypsinization significantly elevated the activity of not only αβ*γ* mutant ENaC but also αβγ wild-type channel, generating equivalent maximal fluorescent responses in the FLIPR assay ( Fig. 2 ). Proteolysis is a well-documented mechanism of controlling the open probability of the channel.^33,34 The cause for the equivalent maximal responses of the wild-type and mutant ENaC requires further investigation. If dye saturation is not the cause, alternative explanations could include the differences in proteolytic regulation between wild-type and mutant ENaC carriers and the implications for health and disease in these populations.

A reliable FLIPR assay was possible only with the second-generation FMP2 Blue dye, which overcomes several limitations of the first-generation FMP formulation by the ability to detect Na⁺-dependent depolarization and respond to all of the known ENaC blockers tested. The potency values were about 1 log unit lower than obtained with the Ussing chamber. The reason for this is unclear, but this level of potency shift is often true for membrane potential FLIPR assays. It is unlikely due to the buffering of compounds by excess amounts of nonfunctional ENaC, as potency values from IonWorks patch clamp recordings using the same cells were in much better agreement with the Ussing chamber. An interesting observation is that the potency of these compounds in FLIPR was quite sensitive to the amount of viral load to the cells, which may reflect the level of functional ENaC expressed on the cell surface.

The viral load of the IonWorks assay was about 10 times that of FLIPR, which would lead to significantly higher Na⁺ influx during cultivation, causing cellular swelling and compromising membrane integrity. This was resolved by growing the cells in low-Na⁺ medium. The change of growth conditions did not lead to alterations in the I-V relationship or the potency of amiloride, suggesting that the channel properties per se are unlikely altered.

The IonWorks Quattro assay using the ENaC recombinant cell line is highly predictable of compound potencies of HEAC I_SC determined by the Ussing chamber. The FLIPR assay, on the other hand, was not only less sensitive but also reported a high degree of false-positives. Nevertheless, it detected the majority of the high-potency IonWorks positives in the focused screen. The significantly lower cost and simplicity of the FLIPR assay merit its consideration as a frontline assay to screen or profile large numbers of compounds. The IonWorks assay is best suited for postscreen activities, such as confirming target engagement and driving SAR campaigns. The HEAC I_SC assay would be conducted prior to in vivo tests. A scheme of lead discovery that incorporates these assays is outlined in Supplemental Figure S1.

Technology is perhaps now ripe for the systematic screening and profiling of diversity compound libraries for novel modulators for diseases either directly or indirectly involving ENaC function or dysfunction.

Footnotes

Acknowledgements

The authors would like to acknowledge Sergio Senar-Sancho, Oyee Chiu, Sandra Arpino, Simon Dowell, and Jeff Clare for general and technical support.

Supplementary material for this article is available on the Journal of Biomolecular Screening Web site at .

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

References

Garty

Palmer

L. G.

Epithelial Sodium Channels: Function, Structure, and Regulation. Physiol. Rev. 1997, 77, 359–396.

Rossier

B. C.

Pradervand

Schild

. Epithelial Sodium Channel and the Control of Sodium Balance: Interaction between Genetic and Environmental Factors. Annu. Rev. Physiol. 2002, 64, 877–897.

Schild

The Epithelial Sodium Channel: From Molecule to Disease. Rev. Physiol. Biochem. Pharmacol. 2004, 151, 93–107.

Kellenberger

Schild

Epithelial Sodium Channel/Degenerin Family of Ion Channels: A Variety of Functions for a Shared Structure. Physiol. Rev. 2002, 82, 735–767.

Matalon

O’Brodovich

Sodium Channels in Alveolar Epithelial Cells: Molecular Characterization, Biophysical Properties, and Physiological Significance. Annu. Rev. Physiol. 1999, 61, 627–661.

Canessa

C. M.

Schild

Buell

. Amiloride-Sensitive Epithelial Na⁺ Channel Is Made of Three Homologous Subunits. Nature 1994, 367, 463–467.

Jasti

Furukawa

Gonzales

E. B.

. Structure of Acid-Sensing Ion Channel 1 at 1.9 A Resolution and Low pH. Nature 2007, 449, 316–323.

Stockand

J. D.

Staruschenko

Pochynyuk

. Insight toward Epithelial Na⁺ Channel Mechanism Revealed by the Acid-Sensing Ion Channel 1 Structure. IUBMB Life 2008, 60, 620–628.

Eaton

D. C.

Malik

Bao

H. F.

. Regulation of Epithelial Sodium Channel Trafficking by Ubiquitination. Proc. Am. Thorac. Soc. 2010, 7, 54–64.

10.

Hanwell

Ishikawa

Saleki

. Trafficking and Cell Surface Stability of the Epithelial Na⁺ Channel Expressed in Epithelial Madin-Darby Canine Kidney Cells. J. Biol. Chem. 2002, 277, 9772–9779.

11.

Pribanic

Debonneville

. The PY Motif of ENaC, Mutated in Liddle Syndrome, Regulates Channel Internalization, Sorting and Mobilization from Subapical Pool. Traffic 2007, 8, 1246–1264.

12.

Abriel

H. J.

Loffing

J. F.

Rebhun

J. H.

. Defective Regulation of the Epithelial Na⁺ Channel by Nedd4 in Liddle’s Syndrome. J. Clin. Invest. 1999, 103, 667–673.

13.

Schild

The Epithelia Sodium Channel and the Control of Sodium Balance. Biochim. Biophys. Acta 2010, 1802, 1159–1165.

14.

Rauh

Diakov

Tzschoppe

. A Mutation of the Epithelial Sodium Channel Associated with Atypical Cystic Fibrosis Increases Channel Open Probability and Reduces Na⁺ Self Inhibition. J. Physiol. 2010, 588, 1211–1225.

15.

Hirsh

A. J.

Molino

B. F.

Zhang

. Design, Synthesis, and Structure-Activity Relationships of Novel 2-Substituted Pyrazinoylguanidine Epithelial Sodium Channel Blockers: Drugs for Cystic Fibrosis and Chronic Bronchitis. J. Med. Chem. 2006, 49, 4098–4115.

16.

Hirsh

A. J.

Zhang

Zamurs

. Pharmacological Properties of N-(3,5-diamino-6-chloropyrazine-2-carbonyl)-N′-4-[4-(2,3-dihydroxypropoxy)phenyl]butyl-guanidine methanesulfonate (552-02), a Novel Epithelial Sodium Channel Blocker with Potential Clinical Efficacy for Cystic Fibrosis Lung Disease. J. Pharmacol. Exp. Ther. 2008, 325, 77–88.

17.

Hunt

Atherton-Watson

H. C.

Axford

. Discovery of a Novel Chemotype of Potent Human ENaC Blockers Using a Bioisostere Approach. Part 1: Quaternary Amines. Bioorg. Med. Chem. Lett. 2012, 22, 929–932.

18.

Hunt

Atherton-Watson

H. C.

Collingwood

S. P.

. Discovery of a Novel Chemotype of Potent Human ENaC Blockers Using a Bioisostere Approach. Part 2: α-Branched Quaternary Amines. Bioorg. Med. Chem. Lett. 2012, 22, 2877–2879.

19.

Clark

K. L.

Hughes

S. A.

Bulsara

. Pharmacological Characterization of a Novel ENaCα siRNA (GSK2225745) with Potential for the Treatment of Cystic Fibrosis. Mol. Ther. Nucl. Acids 2013, 15, e65.

20.

Collingwood

S. P.

Smith

Pyrazine Derivatives as Epithelial Sodium Channel Blocker. U.S. Patents WO2007/071396A2 and WO2007/071400A1. Jun 28, 2007.

21.

Condreay

Witherspoon

Clay

. Transient and Stable Gene Expression in Mammalian Cells Transduced with a Recombinant Baculovirus Vector. Proc. Natl. Acad. Sci. U. S. A. 1999, 96, 127–132.

22.

Shroeder

Neagle

Trezise

D. J.

. Ionworks HT: A New High-Throughput Electrophysiology Measurement Platform. J. Biomol. Screen, 2003, 8, 1–14.

23.

Finkel

Wittel

Yang

. Population Patch Clamp Improves Data Consistency and Success Rates in the Measurement of Ionic Currents. J. Biomol. Screen. 2006, 11, 488–496.

24.

John

V. H.

Dale

T. J

Hollands

E. C.

. Novel 384-Well Population Patch Clamp Electrophysiology Assays for Ca²⁺-Activated K⁺ Channels. J. Biomol. Screen. 2007, 12, 50–60.

25.

Dale

T. J.

Townsend

Hollands

E. C.

. Population Patch Clamp Electrophysiology: A Breakthrough Technology for Ion Channel Screening. Mol. Biosyst. 2007, 3, 714–722.

26.

Hollands

E. C.

Dale

T. J.

Baxter

A. W.

. Population Patch-Clamp Electrophysiology Analysis of Recombinant GABAA alpha1beta3gamma2 Channels Expressed in HEK-293 Cells. J. Biomol. Screen. 2009, 14, 769–780.

27.

Sheppard

D. N.

Hug

M. J.

Transepithelial Electrical Measurements with the Ussing Chamber. J. Cyst. Fibros. 2004(suppl 2), 123–126.

28.

Jiang

Ingbar

D. H.

O’Grady

S. M.

Adrenergic Regulation of Ion Transport across Adult Alveolar Epithelial Cells: Effects on Cl- Channel Activation and Transport Function in Cultures with an Apical Air Interface. J. Membr. Biol. 2001, 181, 195–204.

29.

Chraïbi

Horisberger

J. D.

Na Self Inhibition of Human Epithelial Na Channel: Temperature Dependence and Effect of Extracellular Proteases. J. Gen. Physiol. 2002, 120, 133–145.

30.

Knight

K. K.

Olson

D. R.

Zhou

. Liddle’s Syndrome Mutations Increase Na⁺ Transport through Dual Effects on Epithelial Na⁺ Channel Surface Expression and Proteolytic Cleavage. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 2805–2808.

31.

Clare

J. J.

Chen

M. X.

Downie

D. L.

. Use of Planar Array Electrophysiology for the Development of Robust Ion Channel Cell Lines. Comb. Chem. High Throughput Screen 2009, 12, 96–106.

32.

Kashlan

O. B.

Sheng

Kleyman

T. R.

On the Interaction between Amiloride and Its Putative Alpha-Subunit Epithelial Na⁺ Channel Binding Site. J. Biol. Chem. 2005, 280, 26206–26215.

33.

Chraïbi

Vallet

Firsov

. Protease Modulation of the Activity of the Epithelial Sodium Channel Expressed in Xenopus Oocytes. Gen. Physiol. 1998, 111, 127–138.

34.

Jovov

Berdiev

B. K.

Fuller

C. M.

. The Serine Protease Trypsin Cleaves C Termini of Beta- and Gamma-Subunits of Epithelial Na⁺ Channels. J. Biol. Chem. 2002, 277, 4134–4140.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.09 MB

0.00 MB

Validation and Optimization of Novel High-Throughput Assays for Human Epithelial Sodium Channels

Abstract

Keywords

Introduction

Materials and Methods

Compounds and Materials

Recombinant Expression of ENaC in HEK293 Cells

FLIPR Membrane Potential Assay

IonWorks Planar Array Electrophysiology

Ussing Chamber Short-Circuit Current Recording

Results

Pharmacology of ENaC in Human Airway Epithelia

Analysis of ENaC Activity in Lung- and Kidney-Derived Epithelial Cell Lines

Functional Expression of Recombinant ENaC

Development and Optimization of ENaC FLIPR Membrane Potential Assays

Development and Optimization of ENaC IonWorks Patch Clamp Electrophysiology Assay

Comparative Pharmacology, Hit Identification, and Triage

Discussion

Footnotes

Acknowledgements

Declaration of Conflicting Interests

Funding

References

Supplementary Material