Small-molecule activators of TMEM16A, a calcium-activated chloride channel, stimulate epithelial chloride secretion and intestinal contraction

Abstract

TMEM16A (ANO1) is a calcium-activated chloride channel (CaCC) expressed in secretory epithelia, smooth muscle, and other tissues. Cell-based functional screening of ∼110,000 compounds revealed compounds that activated TMEM16A CaCC conductance without increasing cytoplasmic Ca²⁺. By patch-clamp, N-aroylaminothiazole “activators” (E_act) strongly increased Cl⁻ current at 0 Ca²⁺, whereas tetrazolylbenzamide “potentiators” (F_act) were not active at 0 Ca²⁺ but reduced the EC₅₀ for Ca²⁺-dependent TMEM16A activation. Of 682 analogs tested, the most potent activator (E_act) and potentiator (F_act) produced large and more sustained CaCC Cl⁻ currents than general agonists of Ca²⁺ signaling, with EC₅₀ 3–6 μM and Cl⁻ conductance comparable to that induced transiently by Ca²⁺-elevating purinergic agonists. Analogs of activators were identified that fully inhibited TMEM16A Cl⁻ conductance, providing further evidence for direct TMEM16A binding. The TMEM16A activators increased CaCC conductance in human salivary and airway submucosal gland epithelial cells, and IL-4 treated bronchial cells, and stimulated submucosal gland secretion in human bronchi and smooth muscle contraction in mouse intestine. Small-molecule, TMEM16A-targeted activators may be useful for drug therapy of cystic fibrosis, dry mouth, and gastrointestinal hypomotility disorders, and for pharmacological dissection of TMEM16A function.—Namkung, W., Yao, Z., Finkbeiner, W. E., Verkman, A. S. Small-molecule activators of TMEM16A, a calcium-activated chloride channel, stimulate epithelial chloride secretion and intestinal contraction.

Calcium-activated Cl⁻ channels (CaCCs) are widely expressed in epithelial and nonepithelial cell types, where they facilitate epithelial fluid secretion, smooth muscle contraction, neurosensory signaling, and other functions (1–3). TMEM16A (alternative name, anoctamin-1, ANO1) was identified as a CaCC, as its heterologous expression in oocytes and mammalian cells produced outwardly rectifying, Ca²⁺-sensitive Cl⁻ currents (4–6). TMEM16A is expressed in epithelial cells in airways, salivary gland, intestine, and other tissues, as well as in arterial smooth muscle, intestinal pacemaker cells, sensory neurons, and various tumors (4, 7–9). Though TMEM16A-knockout mice die just after birth because of tracheomalacia (10), electrophysiological measurements in the neonatal knockout mice suggested TMEM16A involvement in chloride secretion in salivary gland (11) and airway (12) epithelia. Evidence has also been reported for TMEM16A involvement in intestinal and vascular smooth muscle contraction, nociception, and bile formation (9, 13–15).

We recently identified, by high-throughput screening, small-molecule inhibitors of TMEM16A chloride conductance. Some compounds, including tannic acid and related gallotannins (16) and the arylaminothiophene CaCC_inh-A01 (17), function as nonselective CaCC inhibitors that inhibit TMEM16A and other, as yet unidentified, CaCCs in multiple cell types. We proposed that CaCC inhibition by gallotannins in red wines and green teas may account, in part, for their health benefits, including reduced risk of cardiovascular disease. TMEM16A-selective inhibitors were also identified, including the aminophenylthiazole, T16A_inh-A01 (18). T16A_inh-A01 inhibited CaCC Cl⁻ current in TMEM16A-transfected cells and in cultures of human salivary gland and IL-4-treated bronchial epithelia, but not in intestine, providing pharmacological data on TMEM16A involvement in CaCC function in various tissues.

Here, we report the identification of small-molecule TMEM16A activators that target TMEM16A itself rather than upstream Ca²⁺ signaling. TMEM16A activators are useful research tools for pharmacological dissection of TMEM16A function, and potential drug candidates for treatment of salivary gland dysfunction, as in Sjogren's syndrome and following radiation injury, as well as for cystic fibrosis, dry eye syndrome, intestinal hypomotility, and other disorders associated with Cl⁻ channel dysfunction (2, 19). In cystic fibrosis, the rationale for CaCC activator therapy is the activation of alternative, non-CFTR chloride channels in airway epithelium where CFTR is dysfunctional. Two CaCC activator therapies for cystic fibrosis have been in clinical trials, including a P2Y₂ receptor antagonist (denufosol) (20), which acts through Ca²⁺ elevation, and a bacterial polycyclic peptide (duramycin) (21). A P2Y₂ receptor agonist is also in clinical trials for dry eye disease (22). CaCC activators that target CaCCs directly without cytoplasmic Ca²⁺ elevation might offer more targeted therapy than general agonists of Ca²⁺ signaling and, unlike receptor agonist therapy, could produce more sustained CaCC activation and hence, offer greater efficacy. Here, we report the identification and activation mechanism of TMEM16A-targeted activators and evidence for their potential therapeutic utility in cystic fibrosis, dry mouth, and intestinal hypomotility.

MATERIALS AND METHODS

Chemicals and solutions

Amiloride, ATP, UTP, and other chemicals, unless otherwise indicated, were purchased from Sigma (St. Louis, MO, USA). 1-(2-Methoxyethyl)-2-thiourea was purchased from Oakwood Products (West Columbia, SC, USA). T16A_inh-A01 and CFTR_inh-172 were synthesized as described previously (23). The compound collections used for screening included ∼100,000 synthetic small molecules from ChemDiv (San Diego, CA, USA) and Asinex (Winston Salem, NC, USA), and ∼7500 purified natural products from Analyticon (Potsdam, Germany), Timtek (Newark, NJ, USA), and Biomol (Plymouth Meeting, PA, USA). Compounds were maintained as DMSO stock solutions. Structure-activity analysis was done on analogs purchased from ChemDiv and Asinex. The HCO₃⁻-buffered solution contained (in mM): 120 NaCl, 5 KCl, 1 MgCl₂, 1 CaCl₂, 10 d-glucose, 5 HEPES, and 25 NaHCO₃ (pH 7.4). In the half-Cl⁻ solution, 65 mM NaCl in the HCO₃⁻-buffered solution was replaced by Na gluconate.

Cell culture

Fisher rat thyroid (FRT) cells were stably transfected with human TMEM16A [TMEM16A(abc), cDNA provided by Dr. Luis Galietta, Gaslini Institute, Genoa, Italy] and the halide sensor YFP-H148Q/I152L/F46L. Cells were plated in 96-well black-walled microplates (Corning, Corning, NY, USA) at a density of 20,000 cells/well in Coon's modified F12 medium supplemented with 5% FCS, 2 mM l-glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin. The human submandibular cell line, A253 (HTB 41; American Type Culture Collection, Manassas, VA, USA) was cultured in complete McCoy's 5A medium supplemented with 10% FBS, 100 U/ml penicillin and 100 μg/ml streptomycin.

Tracheal and bronchial tissues were obtained from non-CF and CF patients following lung transplantation or from post- mortem examinations performed within 24 h after death. Non-CF tissues were from individuals without significant pulmonary airway disease. The Committee on Human Research at the University of California, San Francisco, approved the use of human tissues for these studies. Primary cultures of non-CF and CF human bronchial epithelial (HBE) cells were grown at an air-liquid interface as described previously (24). Cells were plated at a density of 5 × 10⁵/cm² onto 12-mm diameter, 0.4-μm pore polycarbonate cell culture inserts (Snapwell; Corning, Lowell, MA, USA) precoated with human placental collagen (15 μg/cm²; Sigma). Cultures were grown at an air-liquid interface in ALI medium at 37°C in 5% CO₂/95% air (25). Medium was changed every 2–3 days. Cultures were used 21–30 days after plating, at which time transepithelial resistance (R_te) was 400–1000 Ω/cm², and an airway surface liquid film was seen.

Primary cultures of non-CF human tracheal gland (HTG) serous cells were generated from the trachea and mainstem bronchi under conditions that induced serous cell differentiation (26). Briefly, after removal of surface epithelium, the gland-rich submucosal tissues were dissected from between the cartilaginous rings. Small segments of gland tubules and acinar structures were isolated by enzymatic digestion, as described previously. Gland fragments were plated in T-25 flasks in DMEM/F12 supplemented with 20% FBS, penicillin (10⁵ mU/ml), streptomycin (100 μg/ml), gentamicin (100 μg/ml), and amphotericin B (2.5 μg/ml). The next day, cultures were rinsed with PBS, and plating medium was replaced with bronchial epithelial growth medium (BEGM; Lonza, Basel, Switzerland). Medium was changed every 24 h for 3 days, and every 2 days thereafter. When the outgrowths of cells from attached acini reached ∼80% confluence, they were removed by trypsinization (0.05% trypsin, 0.02% EDTA) and plated (3×10⁵ cells) onto 12-mm cell culture inserts coated with human placental collagen (15 μg/cm²). Serous gland cells were grown at an air-liquid interface on 0.4-μm pore polyester cell culture inserts (Snapwell) in DMEM/F12 supplemented with insulin (10 μg/ml), transferrin (5 μg/ml), retinoic acid (5×10⁻⁸ M), hydrocortisone (0.5 μg/ml), triidothyronine (20 ng/ml), BSA (2 mg/ml), 0.1% Ultroser G serum substitute (Pall Corp., Port Washington, NY, USA), and gentamicin (50 μg/ml). Cells were studied after 10–14 days, with R_te >100 Ω · cm².

Screening procedures

High-throughput screening was done using an automated screening platform (Beckman, Brea, CA, USA) equipped with FluoStar fluorescence plate readers (BMG Lab Technologies, Durham, NC, USA) as described previously (23). Each well of a 96-well plate was washed 3 times with PBS (200 μl/wash), leaving 50 μl PBS. Test compounds (0.5 μl) were added to each well at 25 μM final concentration. After 10 min, 96-well plates were transferred to a plate reader for fluorescence assay. Each well was assayed individually for TMEM16A-mediated I⁻ influx by recording fluorescence continuously (400 ms/point) for 2 s (baseline), then 50 μl of a 140 mM I⁻ solution was added. The initial rate of I⁻ influx was computed from fluorescence data by nonlinear regression.

Short-circuit current

Snapwell inserts containing TMEM16A-expressing FRT, HBE, CF HBE, or HTG cells were mounted in Ussing chambers (Physiological Instruments, San Diego, CA, USA). Amiloride, CFTR_inh-172, UTP, ATP, T16A_inh-A01, and TMEM16A activators were added to the apical solution, and an equal volume of vehicle was added at the same time to the basolateral solution. Symmetrical HCO₃⁻-buffered solutions were used for HBE, CF HBE, and HTG cells. For FRT cells, the hemichambers were filled with a half-Cl⁻ solution (apical) and the HCO₃⁻-buffered solution (basolateral), and the basolateral membrane was permeabilized with 250 μg/ml amphotericin B, as described previously (16). Cells were bathed for a 10-min stabilization period and aerated with 95% O₂/5% CO₂ at 37°C or room temperature. Apical membrane current (for FRT cells) and short-circuit current were measured using an EVC4000 multi-channel V/I clamp (World Precision Instruments, Sarasota, FL, USA) and recorded using PowerLab/8sp (AD Instruments, Castle Hill, NSW, Australia).

Patch clamp

Whole-cell recordings were made at room temperature on TMEM16A-expressing FRT cells and human submandibular A253 cells. The bath solution contained (in mM): 140 NMDG-Cl, 1 CaCl₂, 1 MgCl₂, 10 glucose, and 10 HEPES (pH 7.4). The pipette solution contained (in mM): 130 CsCl, 0.5 EGTA, 1 MgCl₂, 1 Tris-ATP, and 10 HEPES (pH 7.2). Different concentrations of free calcium in pipette solution were obtained by replacing 0.5 mM EGTA with 5 mM EGTA and using different amounts of CaCl₂ in the pipette solution. Pipettes were pulled from borosilicate glass and had resistances of 3–5 MΩ after fire polishing. Seal resistances were between 3 and 10 GΩ. After establishing the whole-cell configuration, TMEM16A was activated by 100 μM ATP, TMEM16A activators, or different concentrations of free calcium in the pipette solution. Whole-cell currents were elicited by applying hyperpolarizing and depolarizing voltage pulses from a holding potential of 0 mV to potentials between −100 and +100 mV in steps of 20 mV. Recordings were made at room temperature using an Axopatch-200B (Axon Instruments). Currents were digitized with a Digidata 1440A converter (Axon Instruments), filtered at 5 kHz, and sampled at 1 kHz.

Cytoplasmic calcium measurements

FRT cells in 96-well black-walled microplates were loaded with Fluo-4 NW per the manufacturer's protocol (Invitrogen, Carlsbad, CA, USA). Fluo-4 fluorescence was measured with a FLUOstar Optima fluorescence plate reader equipped with syringe pumps and custom Fluo-4 excitation/emission filters (485/538 nm).

Immunoblot

CF HBE and HTG cells were lysed with cell lysis buffer [50 mM Tris-HCl, pH 7.4, 1% Nonidet P-40, 0.25% sodium deoxycholate, 150 mM NaCl, 1 mM EDTA, 1 mM Na₃VO₄, and protease inhibitor mixture (Roche Applied Science, Indianapolis, IN, USA)]. Cell debris was removed by centrifugation, and proteins in the supernatant were resolved by SDS-PAGE and immunoblotted using standard procedures [transfer to polyvinylidene difluoride membrane, 1 h blocking in 5% nonfat dry milk, primary TMEM16A antibody (1:1000 dilution, ab16293, Abcam, Cambridge, MA, USA) and secondary antibody incubations, and enhanced chemiluminescence detection].

Immunohistochemistry

Paraffin tissue sections (5 μm) were dewaxed in two changes of Clear-Rite (Thermo Scientific, Waltham, MA, USA) then rehydrated through a series of graded alcohols. Slides were submerged in 3% hydrogen peroxide for 10 min to quench endogenous peroxidase activity. Heat-induced antigen retrieval was performed by boiling slides in Borg antigen retrieval buffer (Biocare Medical, Concord, CA, USA) for 10 min at 125°C. Slides were blocked with protein block (Dako, Carpinteria, CA, USA) for 10 min and incubated for 60 min with TMEM16A primary antibody (NBP1–49559, Novus Biologicals, Littleton, CO, USA). Antibody detection was done using the SuperPicture Polymer Kit (Invitrogen). 3,3′-Diaminobenzidine) was used to develop the stains. Slides were counterstained with hematoxylin and photographed.

Gland fluid secretion

Human airways were obtained from human subjects following lung transplantation and the California Lung Transplantation Donation Network. For optical recording of mucus (fluid) secretion in airway glands, a fragment of human trachea or bronchus of ∼1 cm² with underlying glands was dissected from the cartilage and mounted in a 37°C chamber allowing serosal solution exchange. The mucosal surface was rinsed and blotted dry with a cotton swab and further dried with an air stream, after which ∼100 μl of water-saturated mineral oil was placed on the surface. Agonists and inhibitors were added to the serosal side by complete bath replacement. Mucus bubbles in the oil layer were imaged using a Nikon SMZ stereozoom epifluorescence microscope (Nikon, Tokyo, Japan) equipped with P-HR Plan Apo ×1.6 objective lens (working distance 24 mm) and Hamamatsu ORCA-ER CCD camera (Hamamatsu Photonics, Hamamatsu, Japan). Mucus bubble volume was deduced from bubble size, as described previously (27).

Intestinal smooth muscle contraction

Wild-type CD1 mice (age 8–10 wk) were killed by avertin overdose (200 mg/kg). The ileum was removed and washed with ice-cold HCO₃⁻-buffered solution. The ends of the ileal segments were tied with silk thread and connected to a force transducer. Ileal segments were equilibrated for 60 min with a resting force of ∼1 mN, with changes of the bathing solution every 15 min. Tension was monitored continuously with a fixed-range precision force transducer (TSD, 125 C; Biopac, Goleta, CA) connected to a differential amplifier (DA 100B; Biopac). Data were recorded using MP100, Biopac digital acquisition system, and analyzed using Acknowledge 3.5.7 software (Biopac).

Synthesis procedures

Flash chromatography was performed on a CombiFlash Companion chromatography system (Teledyne Isco, Lincoln, NE). ¹H and ¹³C NMR images were obtained on a Bruker 300 MHz instrument (Bruker Corp., Billerica, MA, USA). High-resolution mass spectrometry was done at a core facility at the University of California (Riverside, CA, USA). Elemental analyses were done at the Micro-Mass Facility (University of California, Berkeley, CA, USA).

N-(2-methoxyethyl)-4-phenyl-2-thiazolamine (compound 1)

A solution of 2-bromoacetylphenone (1.59 g, 7.99 mmol) and 1-(2-methoxyethyl)-2-thiourea (1.02 g, 7.61 mmol) in ethanol (30 ml) was refluxed under argon for 4 h. After the reaction mixture was cooled to room temperature, saturated aqueous NaHCO₃ was slowly added. Ethanol was removed under reduced pressure. The resulting suspension was extracted twice with CH₂Cl₂. The combined organic phase was washed with water, dried (Na₂SO₄), and concentrated. Chromatography [silica, hexanes:ethyl acetate (9:1 to 4:1)] yielded a white crystalline solid (1.62 g, 91%), melting point 57–60°C. ¹H NMR (300 MHz, CDCl₃) δ 3.39 (s, 3H), 3.53 (m, 2H), 3.64 (m, 2H), 5.46 (br, 1H), 6.70 (s, 1H), 7.27 (m, 1H), 7.37 (m, 2H), and 7.80 (m, 2H). ¹³C NMR (75 mHz, CDCl₃) δ 45.2, 58.8, 70.7, 101.0, 126.0, 127.6, and 128.5. ESI-MS calculated for C₁₂H₁₅N₂OS 235.0900; found 235.0902.

N-(2-methoxyethyl)-N-(4-phenyl-2-thiazolyl)-2,3,4-trimethoxybenzeneacetamide (E_act)

A solution of compound 1 (1.00 g, 4.27 mmol) and anhydrous pyridine (690 ml, 8.54 mmol) in anhydrous toluene (40 ml) was stirred for 5 min. To the reaction mixture was added a solution of 3,4,5-trimethoxybenzoyl chloride (1.47 g, 6.41 mmol) in anhydrous toluene (20 ml). The mixture was refluxed under argon for 4.5 h, cooled to room temperature, and poured into water and ethyl acetate. The organic phase was collected, dried (Na₂SO₄) and concentrated. Chromatography [silica, hexanes:ethyl acetate (4:1 to 3:1)] yielded a crystalline solid (1.33 g, 73%), melting point 97–100°C. ¹H NMR (300 MHz, CDCl₃) δ 3.28 (s, 3H), 3.86 (t, J = 5.4, 2H), 3.88 (s, 6H), 3.90 (s, 6H), 4.48 (t, J = 5.4, 2H), 6.92 (s, 2H), 7.26 (s, 1H), 7.33 (m, 1H), 7.43 (m, 2H), and 7.90 (m, 2H). ¹³C NMR (75 mHz, CDCl₃) δ 49.6, 56.2, 58.9, 61.0, 69.7, 105.4, 109.4, 126.0, 127.9, and 128.7. ESI-MS is calculated for C₂₂H₂₅N₂O₅S 429.1479; found 429.1477.

Synthesis of F_act analogs, exemplified with N-(4-bromophenyl)-3-(1H-tetrazol-1-yl)benzamide (F_act)

Thionyl chloride (477 ml, 6.58 mmol) and 3-(1H-tetrazol-1-yl)benzoic acid (50 mg, 0.263 mmol) was heated to 80°C in a screw-cap vial. After a clear solution was observed, the residue solid on the wall was washed by gentle shaking. After 1.5 h, the reaction was cooled to room temperature. The reaction mixture was concentrated to dryness under reduced pressure. The resulting white solid was suspended in CH₂Cl₂ (1.5 ml) and treated with p-bromoaniline (90 mg, 0.526 mmol) and triethylamine (100 ml, 0.719 mmol). The reaction was stirred at room temperature for 14 h, mixed with silica gel, and concentrated to dryness. Chromatography [silica, CH₂Cl₂:methanol (98:2 to 1:3)] yielded an off-white solid (48 mg, 53%), decomposition point 203°C. ¹H NMR (300 MHz, DMSO-d₆) d 7.58 (d, J = 4.9, 2H), 7.79 (d, J = 4.9, 2H), 7.85 (δ, J = 8.0, 1H), 8.10–8.18 (m, 2H), 8.47–8.51 (m, 1H), 10.22 (s, 1H), and 10.62 (br, 1H, NH). ¹³C NMR (75 mHz, DMSO-d₆) δ 120.18, 122.08, 123.92, 128.50, 130.11, 131.30, and 142.25. ESI-MS is calculated for C₁₄H₁₀BrN₅O 344.0141; found 344.0152.

RESULTS

Discovery and characterization of TMEM16A activators

TMEM16A activators were identified from screening of ∼110,000 synthetic drug-like compounds, purified natural products, and approved/investigational drugs. Our cell-based screen utilized FRT cells coexpressing human TMEM16A and the I⁻-sensing yellow fluorescent protein YFP-H148Q/I152L/F46L. FRT cells were chosen because of their low basal I⁻ and Cl⁻ transport, rapid growth on uncoated plastic, strong stable expression of transfected proteins, and formation of tight junctions for measurements of transepithelial short-circuit current (23). As diagrammed in Fig. 1A, test compounds at 25 μM final concentration were added 10 min prior to I⁻ addition. The 10-min incubation was chosen to allow for compound transport into cytoplasm and to minimize false positives from compounds that elevate cytoplasmic Ca²⁺ transiently. Fluorescence from individual wells of 96-well plates was measured just prior to and for 6 s after I⁻ addition for computation of initial I⁻ influx rate. Figure 1B shows representative fluorescence data from single wells showing positive (ionomycin) and negative (vehicle only) controls, and examples of inactive and active compounds.

Figure 1.

Identification of small-molecule TMEM16A activators by high-throughput screening. A) Screening protocol. FRT cells stably expressing TMEM16A and the halide-sensitive cytoplasmic fluorescent sensor YFP-H148Q/I152L/F46L were incubated for 10 min with test compound. Fluorescence was monitored in response to addition of iodide. B) Fluorescence measured in single wells of 96-well plates, showing vehicle and positive (ionomycin) controls and examples of inactive and active compounds. C) Structures of TMEM16A activators of six different chemical classes.

Primary screening yielded 40 compounds that increased I⁻ influx by >2 mM/s at 25 μM (>50% of maximal I⁻ influx produced by 100 μM ATP) and had EC₅₀ of <10 μM. Figure 1C shows structures of active compounds from six chemical classes. Secondary screens were done to identify compounds that increased I⁻ influx by targeting TMEM16A. Measurements on TMEM16A null cells (expressing YFP alone) showed that none of the active compounds increased I⁻ influx. However, C_act and D_act increased TMEM16A-mediated I⁻ efflux by producing sustained elevation of cytoplasmic Ca²⁺ (Fig. 2A). Though such Ca²⁺ agonists are of potential interest, they were not studied further. Of the remaining compounds that did not elevate Ca²⁺, we chose the N-aroylaminothiazole E_act and the tetrazolylbenzamide F_act for further characterization because they produced maximal TMEM16A activation (compared to ATP effect) in apical membrane current measurements, whereas A_act and B_act produced only ∼ 50% maximal activation.

Figure 2.

Characterization of TMEM16A activators. A) Cytoplasmic calcium measured by Fluo-4 fluorescence. Arrow indicates addition of 100 μM ATP (gray line) or 10 μM of indicated TMEM16A activators. B) Apical membrane current measured in TMEM16A-expressing FRT cells in the presence of a transepithelial chloride gradient and after basolateral membrane permeabilization. Left and center panels: representative current traces showing ATP (100 μM), E_act- or F_act-stimulated TMEM16A Cl⁻ current. T16A_inh-A01 (10 μM) was added where indicated. Inset shows long-time E_act effect. Right panel: concentration-activation data summary (mean±se, n=4–6). C) Synergistic effect of E_act and F_act. Left and center panels: representative current traces showing synergy. Right panel: data summary of low doses of TMEM16A activation (mean±se, n=5). *P < 0.05. D) Apical membrane current measured in FRT cells transfected with mouse TMEM16A or TMEM16B. E) Effect of E_act and F_act on CFTR and ENaC. Left panel: FRT cells expressing wild-type CFTR and YFP indicator were pretreated for 5 min with 10 μM E_act and F_act. Forskolin (10 μM) was added as indicated. Right panel: HBE cells were pretreated for 5 min with 10 μM E_act and F_act, with amiloride (10 μM) added as indicated.

Figure 2B (left and center panels) shows measurements of apical membrane Cl⁻ current in TMEM16A-expressing FRT cells in which the cell basolateral membrane was permeabilized with amphotericin B and a mucosal-to-serosal Cl⁻ gradient was applied. The purinergic agonist ATP, which transiently elevates cytoplasmic Ca²⁺, produced a large but transient elevation in Cl⁻ current. E_act and F_act produced large, concentration-dependent increases in Cl⁻ current, which were inhibited by the TMEM16A-selective inhibitor T16A_inh-A01. The current increase was sustained for >10 min (Fig. 2B, inset). The concentration-activation data gave EC₅₀ of ∼3 μM for E_act and ∼6 μM for F_act (Fig. 2B, right panel). Figure 2C shows synergy between E_act and F_act for TMEM16A activation, suggesting distinct mechanisms of action. Whereas 1–3 μM F_act produced little TMEM16A activation alone, it greatly increased TMEM16A current following 1 μM E_act.

Figure 2D (left panel) shows that E_act was effective in producing Cl⁻ current in mouse TMEM16A, which supports its testing in mouse tissues. Figure 2D (right panel) shows activation by E_act of TMEM16B, the other TMEM16 isoform having CaCC activity. Neither E_act nor F_act affected CFTR Cl⁻ conductance or ENaC Na⁺ conductance (Fig. 2E), which are often found in epithelial cell mucosal membranes where TMEM16A is expressed.

More than 1000 analogs of the B, E, and F classes were tested for TMEM16A inhibition activity, reasoning that small structural changes can convert an agonist into an antagonist. As shown in Fig. 3A, we found compounds of the B and E classes that fully inhibited TMEM16A Cl⁻ conductance. In each case relatively minor chemical structural changes (common scaffolds shown in red in Fig. 3A) converted an activator to an inhibitor, supporting the conclusion that these compounds target TMEM16A directly.

Figure 3.

Structure-activity analysis and synthesis of TMEM16A activators. A) Structural similarities between TMEM16A inhibitors and activators (common scaffolds shown in red). Apical membrane current measurements show activation of TMEM16A by B_act (top panels), inhibition of ionomycin (1 μM)-induced TMEM16A currents by pretreatment B and E class analogs (each 10 μM; bottom panels). B) Summary of structural determinants for TMEM16A activation. Left panel: E_act class. Right panel: F_act class. C) Synthesis of E_act and F_act analogs (see Materials and Methods).

E_act and F_act analogs were assayed to establish structure-activity relationships and to select the best compound(s) for resynthesis in highly pure form for biological studies. Of 673 commercially available class E analogs screened, 18 compounds increased TMEM16A Cl⁻ conductance. Nine of the 10 most potent compounds had a 2,3,4-trimethoxyphenyl (TMOP) group at the R¹ position (Fig. 4). Similar compounds but with 4-methoxyphenyl or 2,5-dimethoxyphenyl at R¹ were inactive, as summarized in Fig. 3B (left panel). Of the active compound with TMOP at R¹, the most potent compounds had benzyl, (tetrahydrofuran-2-yl)methyl, or methoxyethyl groups at R². Compounds containing an additional carbon on the benzyl group were inactive, as were compounds with methylene replacing the oxygen in methoxyethyl group. Methoxyethyl at R² and phenyl group at R³ gave one of the most potent analogs. Limited SAR on the F class was done on 9 synthesized analogs, as analogs were not available from commercial sources. We found that while 4-bromo at R¹ increased TMEM16A Cl⁻ conductance, the 2-bromo or 3-bromo analogs did not, and that compounds containing 4-chloro, 4-nitro, 4-ethoxycarbonyl, or 4-dimethylamino were inactive.

Figure 4.

Structure-activity analysis of E_act and F_act analogs. EC₅₀ values were determined from fluorescence plate reader assay.

The most potent compounds of the E and F classes were synthesized in highly pure form for further characterization and biological studies (Fig. 3C). E_act was synthesized in two steps. 2-Aminothiazole was obtained by reflux of 2-bromoacetyl-phenone and 1-(2-methoxyethyl)-2-thiourea in ethanol (Fig. 3C). E_act was obtained by reaction of 2-aminothiazole and 2,3,4-trimethoxybenzoyl chloride using anhydrous pyridine in anhydrous toluene. The yield (73%) was comparable to that reported for similar reactions between N-alkyl-2-aminothiazoles and benzoyl chloride (28). F_act analogs were synthesized from 3-(1H-tetrazol-1-yl)benzoic acid and the corresponding anilines. Because amide formation with 1,1′-carbonyldiimidazole as the coupling agent did not drive the reaction to completion, we adopted a two-step, one-pot procedure (29). The benzoic acid was first treated with neat SOCl₂ at 80°C for 1.5 h. After removal of excess SOCl₂ by rotary evaporation, the resulting acid chloride was suspended in CH₂Cl₂ and treated with anilines and TEA to yield F_act compounds in 48–65% yield.

The Ca²⁺ dependence of TMEM16A activation by E_act and F_act was investigated. Apical membrane current measurements done with 0 Ca²⁺ apical and basolateral solutions in the presence of cycloplazonic acid (to deplete intracellular Ca²⁺ stores) showed TMEM16A Cl⁻ currents induced by E_act, but not by ATP or F_act (Fig. 5A). Whole-cell currents were then recorded by patch clamp at different cytoplasmic (pipette) [Ca²⁺]. In the absence of activators, Fig. 5B (top panel) shows increasing TMEM16A Cl⁻ current with outward rectification at relatively low [Ca²⁺] and near linear currents at high [Ca²⁺], in agreement with prior patch-clamp studies of TMEM16A (4). E_act strongly activated TMEM16A at 0 Ca²⁺, producing outwardly rectifying currents, with more linear currents at higher [Ca²⁺] (Fig. 5B, bottom and right panels). In contrast, F_act did not produce Cl⁻ current at 0 Ca²⁺, but increased Cl⁻ current (compared to no compound) at submaximal Ca²⁺ (Fig. 5C). Maximum TMEM16A Cl⁻ current at high [Ca²⁺] was not further increased by E_act or F_act. Thus, E_act and F_act activate TMEM16A by different mechanisms: E_act as a largely Ca²⁺-independent activator, and F_act as a potentiator of Ca²⁺ activation. T16A_inh-A01 completely blocked Cl⁻ currents produced by E_act or F_act (Fig. 5D), as expected.

Figure 5.

Patch-clamp analysis of Ca²⁺ requirements for TMEM16A activation by E_act and F_act. A) Apical membrane current measured in TMEM16A-expressing FRT cells. ER calcium stores were depleted by CPA (50 μM, 30 min) and 0 CaCl₂ in bath. ATP (100 μM), F_act (10 μM), E_act (10 μM), and T16A_inh-A01 (10 μM) were added as indicated. B, C) Whole-cell TMEM16A currents were recorded at a holding potential at 0 mV, and pulsing to voltages between ± 80 mV (in steps of 20 mV) in the absence and presence of 3 μM E_act (B) or 10 μM F_act (C). Left panels: free calcium concentration of pipette solutions were clamped at 0 μM (black), 0.07 μM (green), 0.15 μM (red), and 1 μM (blue). E_act (3 μM) or F_act (10 μM) was added as indicated. Right panels: current-voltage (I/V) plots of mean currents at the middle of each voltage pulse. D) TMEM16A inhibited by 10 μM T16A_inh-A01 after stimulation by E_act or F_act.

Epithelial fluid secretion and intestinal smooth muscle contraction

Prior studies using TMEM16A inhibitors and RNAi knockdown indicated that TMEM16A is a minor component of total CaCC conductance in human bronchial surface epithelial cell cultures under basal conditions, but that TMEM16A is strongly up-regulated after IL-4 treatment for 24 h (18). Supporting this conclusion, we found here that E_act did not induce Cl⁻ current in untreated human CF bronchial epithelial cells, whereas UTP, which elevates cytoplasmic Ca²⁺ and hence non-TMEM16A CaCC(s), produced a large Cl⁻ current (Fig. 6A, left panel). Remarkably, E_act induced Cl⁻ current in IL-4-treated CF bronchial cells, which was blocked by the TMEM16A-selective inhibitor T16A_inh-A01 (Fig. 6A, center and right panels).

Figure 6.

Airway epithelial chloride secretion. A) Short-circuit current in CF HBE cells. Left and center panels: E_act and UTP (100 μM) were added in control (left) and IL-4 (10 ng/ml, 24 h, center) treated CF HBE cells. Right panels: summary of E_act-induced, T16A_inh-A01-sensitive peak current (mean±se, n=6–8). *P < 0.05. ENaC was inhibited by 10 μM amiloride. B) E_act (10 μM) and UTP (100 μM) induced CaCC Cl⁻ current measured in primary cultures of non-CF human tracheal gland (HTG) serous cells. Left and center panels: TMEM16A, CFTR, and ENaC were inhibited by pretreatment with T16A_inh-A01, CFTR_inh-172, and amiloride, respectively. Inset: TMEM16A immunoblot in whole cell homogenates of CF HBE and HTG cells. Right panel: summary of UTP and E_act-induced peak current in the presence and absence of T16A_inh-A01 (mean±se, n=3). *P < 0.05.

Prior reports suggested the involvement of TMEM16A in CaCC activity in airway submucosal glands (30, 31). To verify TMEM16A function, short-circuit current measurements were done in primary cultures of human tracheal submucosal gland epithelial cells that were grown under conditions that preserve serous-type phenotype (26). Figure 6B (left panel) shows increased Cl⁻ current in response to UTP and E_act, which was largely abolished by T16A_inh-A01 pretreatment. Figure 6B (middle panel) shows increased Cl⁻ conductance with E_act in the absence of UTP pretreatment. Immunoblot (Fig. 6B, inset) confirmed TMEM16A protein in the glandular epithelial cell cultures. Averaged peak Cl⁻ currents are summarized in Fig. 6B (right panel).

Immunostaining of human non-CF and CF bronchi showed TMEM16A expression at the luminal membrane of submucosal gland serous cells, but not gland mucous cells or surface epithelial cells (Fig. 7A). Similar staining was found in non-CF and CF human bronchi. Fluid secretion was measured in individual submucosal glands from the increasing size of mucus (fluid) bubbles in which airway fragments were mounted in a 37°C perfusion chamber and the mucosal solution was covered with oil. Figure 7B (top panel) shows mucus bubbles following addition of carbachol and E_act in human bronchi. Figure 7B (bottom panel) summarizes the secretion rate from many mucus droplets, showing nonzero basal secretion, and significantly increased secretion following serosal application of submaximal carbachol and E_act. T16A_inh-A01 pretreatment abolished basal and agonist-stimulated gland fluid secretion. The data summary in Fig. 7C shows E_act-induced gland fluid secretion in non-CF and CF bronchi of comparable magnitude to that induced by maximal cAMP stimulation by forskolin.

Figure 7.

Airway submucosal gland fluid secretion in human bronchi. A) TMEM16A immunohistochemistry in CF (left panel) and non-CF (right panel) human bronchi showing apical membrane expression in serous gland epithelial cells (arrows). Scale bar = 20 μm. B) Mucus (fluid) secretion in human bronchi. Top panels: images of mucus bubbles formed under oil in response to basolateral application of 300 nM carbachol (CCh) and 20 μM E_act. TMEM16A was inhibited by 30 μM T16A_inh-A01. Individual fluid bubbles marked with arrowheads. Scale bar = 0.5 mm. Bottom panels: CCh and E_act-induced secretion rates. Where indicated, tissues were pretreated with T16A_inh-A01 (30 μM). Each point is the average of measurements made from 20 glands (mean±se). C) Summary of human gland fluid secretion rates measured at 20 min after addition of 20 μM E_act, and 30 min after application of 300 nM carbachol (CCh) and 10 μM forskolin (20–66 glands from 3 tracheas and 4 bronchi). In CF-bronchi, 6 glands from one donor were stimulated by E_act. *P < 0.05.

Prior studies implicated TMEM16A as the principle CaCC in salivary gland epithelium (4, 11, 18). TMEM16A immunostaining showed that TMEM16A is expressed on the apical surface of acinar epithelial cells in human parotid gland (Fig. 8A, left panel). A253 cells, a human salivary gland epithelial cell line, express TMEM16A (Fig. 8A, right panel). By whole-cell patch-clamp E_act strongly increased Cl⁻ current in A253 cells (Fig. 8B).

Figure 8.

Salivary gland epithelial cell Cl⁻ secretion. A) Expression of TMEM16A in human salivary gland. Left panel: TMEM16A immunostaining in human parotid gland. Scale bar = 20 μm. Right panel: immunoblot of TMEM16A in FRT-TMEM16A and A253 cells. B) Whole-cell patch-clamp recordings in A253 cells. Left panel: CaCC and TMEM16A chloride current induced by 100 μM ATP and 10 μM E_act, respectively. Right panel: current/voltage (I/V) plot of mean currents at the middle of each voltage pulse (voltages between −80 and +80 mV in steps of 20 mV).

TMEM16A in expressed in the interstitial cells of Cajal, the pacemaker cells that control smooth muscle contraction dynamics in stomach and intestine (8, 32). We investigated the effect of E_act on mouse ileal smooth muscle contraction in an ex vivo intestinal preparation. Figure 9A shows considerable constitutive activity of mouse ileal muscle segments at baseline, with large, spontaneous intestinal contractions that were inhibited by T16A_inh-A01 (left panel). Resting and maximum tone were increased by carbachol, without change in contraction frequency (Fig. 9, middle panel). E_act produced a very small increase in resting but not maximum tone (Fig. 9, right panel). We reasoned that the constitutive TMEM16A activity in this ex vivo preparation could obscure E_act activity. To reveal E_act effects, atropine was first added to inhibit basal contractions. Figure 9B shows that atropine prevented the carbachol effect, which is Ca²⁺ dependent, and revealed a large increase in contraction amplitude following E_act, whose action is Ca²⁺ independent. Figure 9C summarizes data on contraction frequency, resting tone, and maximal tone.

Figure 9.

Intestinal smooth muscle contraction. A) Representative traces from mouse ileal segments showing effects of T16A_inh-A01 (10 μM), carbachol (CCh, 1 μM) and E_act (10 μM). B) Effect of CCh (1 μM) and E_act (10 μM) following atropine (1 μM). C) Summary of contraction frequency (left panel) and resting and maximum tone (right panel; mean±se, n=4–7). *P < 0.05.

DISCUSSION

A functional, cell-based screen of small-molecule collections revealed several chemical classes of TMEM16A activators that produced strong and more sustained Cl⁻ currents than Ca²⁺-elevating purinergic agonists in multiple cell types without elevating cytoplasmic Ca²⁺. Two classes of compounds with distinct activating mechanisms, activators and potentiators, were identified. Though various agonists of cytoplasmic Ca²⁺ have been available and studied in clinical trials, direct-acting CaCC modulators have not been reported previously. The more sustained CaCC activation produced by the compounds identified here could translate to improved efficacy compared to Ca²⁺ agonists, such as P2Y₂ agonists, which generally produce only transient elevation in cytoplasmic Ca²⁺ and, consequently, in Cl⁻ secretion. The recently reported phase 3 trial of the P2Y₂ agonist denufosol (20), which failed to show clinical efficacy, may be related to its limited duration of action. In addition to producing more sustained activation of Cl⁻ conductance, direct-acting CaCC activators also have the theoretical advantage over Ca²⁺ agonists of greater target specificity.

Action of E_act and F_act probably involves direct interaction with the TMEM16A protein, as they did not elevate cytoplasmic Ca²⁺ and were effective in patch-clamp studies in which [Ca²⁺] was clamped by the pipette solution. Further, the generation of TMEM16A inhibitors by minor structural modification of activators strongly support a direct binding mechanism. The molecular site of activator interaction on the TMEM16A protein is not known and will require mutagenesis or purification/binding studies. On the basis of the assumed topography and domain structure of TMEM16A (4), we speculate that E_act interacts at or very near the Ca²⁺ binding site, perhaps at the cluster of four contiguous glutamic acid residues localized in the first intracellular loop (7). F_act, which increased the Ca²⁺ sensitivity of TMEM16A activation, might act by an allosteric mechanism at a site distinct from the Ca²⁺ binding site.

Lung disease pathogenesis in CF is thought to involve distinct defects in airway submucosal gland fluid secretion, causing secretion of hyperviscous mucus (33), and in airway surface transport, causing reduced airway surface liquid volume (34). Both related defects are likely attributed to defective CFTR-mediated Cl⁻ secretion, and perhaps ENaC hyperactivity, though the evidence remains controversial (35). Small-molecule therapies under development to correct the underlying CFTR defect include correctors, potentiators, and read-though enhancers to restore Cl⁻ conductance in cells expressing CF-causing mutant CFTRs (2, 36), as well as gene replacement (37). Therapies targeting the activation of alternative Cl⁻ channels, the CaCCs, have received considerable attention as well, as CaCCs are robustly expressed in non-CF and CF airways where CFTR is normally expressed. Our results here provide evidence that TMEM16A activators increase secretion by submucosal glands, as well as by airway surface cells exposed to a proinflammatory milieu. The long-acting TMEM16A activators identified here may thus be useful for CF therapy.

Dry mouth due to salivary gland dysfunction is seen following radiation therapy for head and neck cancers, in Sjogren's syndrome, and most commonly, with unknown (idiopathic) etiology. Dry eye (keratoconjunctivis sicca) is a related disorder that is very common in the elderly, which results from lacrimal or meibomian gland dysfunction. TMEM16A is the major ion channel regulating saliva secretion by salivary gland acinar epithelial cells (11). As such, TMEM16A activators are predicted to potentiate, or amplify, endogenous cholinergic signals to maximize saliva secretion from functional salivary gland acini. Because Cl⁻ and saliva fluid secretion require activity of basolateral membrane K⁺ channels, TMEM16A-targeted activators are potentially advantageous over nonselective Ca²⁺-elevating agonists, as they would primarily amplify the effect of physiological stimuli rather than cause fluid secretion on their own. Follow-on studies are indicated of TMEM16A activator therapy in rodent models of dry mouth involving radiation- or immune-mediated salivary gland injury.

Interstitial cells of Cajal (ICC) generate slow-wave pacemaker activity in smooth muscle in the gastrointestinal (GI) tract. TMEM16A is expressed in ICC but not in the GI smooth muscle cells (8, 9, 38). A recent report showed that pharmacological inhibition or genetic deletion of TMEM16A abolished slow waves in murine small intestine (9). We found that the TMEM16A-selective inhibitor T16A_inh-01 greatly reduced smooth muscle contraction in mouse intestine ex vivo, and that the TMEM16A activator E_act increased contraction, restoring contraction following atropine inhibition. These data suggest that GI smooth muscle motility can be modulated by pharmacological activation or inhibition of TMEM16A. TMEM16A activators may be of therapeutic utility for GI motility disorders such as slow transit constipation, and inhibitors for disorders associated with hypermotility. Of interest, TMEM16A modulators changed the strength but not the frequency of intestinal smooth muscle contraction. Pharmacological inhibition of CaCCs by nonspecific inhibitors of CaCCs (niflumic acid, DIDS) reduce the frequency and blocked slow waves in murine intestine (9). Thus, TMEM16A may be involved in the regulation of intestinal smooth muscle contractility and other CaCCs in the regulation of contraction frequency.

There is prior literature on the chemistry and biological properties of analogs of the class E and F compounds identified here as TMEM16A activators. 4-Aryl-1,3-thiazol-2-yl-N-methylbenzamide analogs have been reported as inhibitors of mGluR1 (28) malonyl-CoA decarboxylase (39), and HCV RNA polymerase (40). Analysis of 2-aminothiazole analogs having a secondary amide structure (rather than the tertiary amide in E_act), have revealed analogs as thrombopoietin agonists (41), VEGFR-2 kinase inhibitors (42), adenosine receptor antagonists (43), RNA polymerase inhibitors (40), and mGluR1 antagonists (28). These compounds also have antimicrobial (44), hypolipidemic (45), and antiatherogenic (46) activities. The published data suggest favorable pharmacological properties and low toxicity of compounds related to class E and F activators, though pharmacological testing of the specific TMEM16A activators identified here will be needed for their further preclinical development.

In summary, the TMEM16A activators identified here by high-throughput screening likely act by binding to the TMEM16A protein. They produced large and more sustained Cl⁻ currents than Ca²⁺-elevating purinergic agonists in salivary gland and airway submucosal gland epithelial cells and IL-4-treated bronchial surface epithelial cells, and stimulated submucosal gland secretion in human bronchi and smooth muscle contraction in mouse intestine. Small-molecule, TMEM16A-targeted activators may have broad utility for treatment of cystic fibrosis, dry mouth, and dry eye syndromes, and motility disorders of the gastrointestinal tract.