TMEM16A/ANO1 Channels Contribute to the Myogenic Response in Cerebral Arteries

Animal protocols were reviewed and approved by the Animal Care and Use Committee of the University of Tennessee Health Science Center. Male Sprague-Dawley rats (age, 6–8 weeks) were euthanized by intraperitoneal injection of sodium pentobarbital (150 mg/kg). The brain was removed and placed into physiological saline solution of composition (in mmol/L): 112 NaCl, 4.8 KCl, 24 NaHCO₃, 1.8 CaCl₂, 1.2 MgSO₄, 1.2 KH₂PO₄, and 10 glucose, gassed with 21% O₂–5% CO₂–74% N₂ to pH 7.4. Resistance-size (≈200-μm diameter) posterior cerebral, cerebellar, and middle cerebral arteries were dissected from the brain, pooled, and used for experimentation, unless specified.

Smooth muscle cells were isolated from cerebral arteries as previously described.¹⁶ Patch-clamp electrophysiology was performed using isolated cerebral artery smooth muscle cells or human embryonic kidney 293 (HEK293) cells expressing recombinant TMEM16A channels. For whole-cell current measurements, the pipette solution contained (in mmol/L): 126 CsCl, 10 HEPES, 10 D-glucose, 1 EGTA or 1 BAPTA, 1 Mg ATP, 0.2 GTP·Na, and 40 sucrose with pH adjusted to 7.2 with CsOH. Total MgCl₂ was adjusted to give final free Mg²⁺ of 1 mmol/L. For whole-cell current experiments on arterial smooth muscle cells, pipette free Ca²⁺ was 200 nmol/L. For whole-cell current measurements in HEK293 cells expressing recombinant TMEM16A channels, pipette free Ca²⁺ was 1 μmol/L. Free Mg²⁺ and Ca²⁺ were calculated using WebmaxC Standard (http://www.stanford.edu/~cpatton/webmaxcS.htm) and confirmed using a Ca²⁺-sensitive and reference electrode (Corning; Acton, MA). Bath solutions used for whole-cell recordings are described in Online Table I. For cell-attached patch measurements, the pipette solution contained (in mmol/L): 134 NaCl, 6 KCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES, 10 glucose, 1 TEA⁺ and 5 4-aminopyridine (pH 7.4, NaOH). For cell-attached patch measurements, the bath solution contained (in mmol/L): 140 KCl, 10 glucose, 10 HEPES, 2 CaCl₂, 1 MgCl₂ (pH 7.4, KOH). To study anion permeability, Cl⁻ was replaced with equimolar aspartate or I⁻. The osmolarity of solutions was measured using a Wescor 5500 Vapor Pressure Osmometer (Logan, UT). To minimize junction potential, the reference Ag/AgCl electrode was immersed in a solution of 3 mmol/L KCl continuous with an agar bridge (4% agar in 3 mmol/L KCl). Whole-cell Cl⁻ currents were measured by applying 1.5-second voltage steps between −80 mV and +80 mV in 20 mV increments, using an interpulse holding potential of −40 mV. Currents were normalized to membrane capacitance. Cell-attached currents were measured at a steady membrane potential of −80 mV. Pharmacological agents and rabbit monoclonal anti-TMEM16A antibody (Abcam) were introduced directly into the experimental chamber. Boiled (15 minutes at 98°C), denatured TMEM16A antibody served as a control for active antibody. Pressure-induced stretch was applied to the plasma membrane contained within the patch pipette, using an ez-gSEAL 100B controller (Neo Biosystems). Membrane currents were recorded using an Axopatch 200B amplifier equipped with a CV 203BU headstage, Digidata 1332A, and Clampex 8 or 9 (Molecular Devices). Whole-cell currents were filtered at 1 kHz, using a low pass Bessel filter and digitized at 4 kHz. Single-channel currents were filtered at 2 kHz and digitized at 8 kHz. The relative anion permeability ratio of I⁻ to Cl⁻ (P_I/P_Cl) or aspartate (Asp) to Cl⁻ (P_Asp/P_Cl) was calculated using the shift in reversal potential (E_rev) and the constant field equation:

where X is I⁻ or Asp and zF/RT is −0.039 at 25°C.

Cerebral arteries were homogenized using Laemmli sample buffer (2.5% SDS, 10% glycerol, 0.01% bromphenol blue, and 5% β-mercaptoethanol in 100 mmol/L Tris·HCl, pH 6.8) and centrifuged at 6000g for 10 minutes to remove cellular debris. Proteins (40 μg/lane) were separated on a 7.5% SDS-polyacrylamide gel and transferred onto nitrocellulose membranes. Blots were physically cut at 75 kDa to permit probing for TMEM16A, TRPC6, TRPM4, or TRPP2 at the higher molecular weight and for actin at the lower molecular weight. Membranes were incubated with rabbit monoclonal anti-TMEM16A (1:100, Abcam), rabbit anti-TRPC6 (1:250, Sigma), rabbit anti-TRPM4 (1:500, Thermo Scientific), rabbit anti-TRPP2 (1:1000, Johns Hopkins Polycystic Kidney Disease Research and Clinical Core Center, or 1:100, Santa Cruz) and mouse monoclonal antiactin (1:5000 dilution, Chemicon International) primary antibodies overnight at 4°C in Tris-buffered saline (TBS) with 0.1% Tween 20 (TBS-T) and 5% nonfat dry milk. Proteins were visualized using horseradish peroxidase–conjugated secondary antibody (1:10 000 dilution; Pierce) and a chemiluminescent detection kit (Pierce). Band intensity was quantified by digital densitometry using Quantity One software (Bio-Rad). Protein band intensity was normalized to actin.

Three small interference RNA (siRNA) sequences targeting TMEM16A or negative control siRNA (Invitrogen), as used previously,¹² were inserted intracellularly into cerebral arteries using either reverse permeabilization, as described,^17–21 or a Bex CUY21Vivo-SQ electroporator. Arteries were then maintained in serum-free DMEM F12 media supplemented with 1% penicillin-streptomycin (Sigma) for 4 days after reverse permeabilization or 3 days after electroporation at 37°C in a sterile incubator (21% O₂, 5% CO₂). Western blotting was used to compare the effect of TMEM16A siRNA with control siRNA on protein expression. Band intensity of proteins from arteries treated with either TMEM16A siRNA or control siRNA were compared on the same membranes. Reverse permeabilization and electroporation similarly reduced TMEM16A protein (reverse permeabilization, 62 ± 5% of control siRNA, n=7; electroporator, 56 ± 1% of control siRNA, n=3) in arteries (P>0.05).

HEK293 cells were maintained in DMEM supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin under standard tissue culture conditions (21% O₂–5% CO₂; 37°C). HEK293 cells were transiently transfected with pcDNA3 encoding full-length recombinant TMEM16A (2 μg), a kind gift from Dr Luis Galietta, Istituto Giannina Gaslini, Italy. Transfection was done using Effectene (Qiagen). Cells were used between 36 and 72 hours after transfection.

Experiments were performed using endothelium-denuded middle cerebral arteries. Arteries were maintained in physiological saline solution containing (in mmol/L): 112 NaCl, 4.8 KCl, 26 NaHCO₃, 1.8 CaCl₂, 1.2 MgSO4, 1.2 KH₂PO4, and 10 glucose, gassed with 74% N₂, 21% O₂, 5% CO₂ (pH 7.4). Artery segments 1 to 2 mm in length were cannulated at each end in a temperature-controlled perfusion chamber (Living Systems Instrumentation; Burlington, VT). Intravascular pressure was altered by means of a reservoir and monitored using a pressure transducer.

Arterial wall diameter was measured at 1 Hz using a CCD camera attached to a Nikon TS100-F microscope and the automatic edge detection function of IonWizard software (Ionoptix, Milton, MA). Luminal flow was absent during experiments. Myogenic tone (%) was calculated as 100×(1-active diameter/passive diameter). Endothelial denudation was confirmed using methods previously described.¹⁸

Membrane potential measurements were obtained in arteries at 10 or 60 mm Hg that had developed steady-state myogenic tone. This was done by maintaining arteries at steady pressure for at least 2hours and confirmed using edge detection. Membrane potential was measured by inserting glass microelectrodes filled with 3 mol/L KCl (50–90 mΩ) into the adventitial side of pressurized arteries. Membrane potential was recorded using a WPI FD223 amplifier and digitized using pClamp 9.2 software (Axon Instruments) and a personal computer. Criteria for successful intracellular impalements were (1) a sharp negative change in potential on insertion; (2) stable voltage for at least 1 minute after entry; (3) a sharp positive voltage deflection on exit from the recorded cell; and (4) a <10% change in tip resistance after the impalement.

OriginLab and GraphPad InStat software were used for statistical analyses. Values are expressed as mean±SEM. Student t test was used for comparing paired and unpaired data from 2 populations, and ANOVA with Student-Newman-Keuls post hoc test was used for multiple group comparisons. P<0.05 was considered significant. Power analysis was performed on all data in which P>0.05 to verify that sample size was sufficient to give a power value of >0.8. All-points histograms were fit with a multipeak gaussian function, using Microcal Origin.

Expanded materials and methods are provided in the Online Data Supplement.

To study the mechanosensitivity of Cl⁻ currents in arterial smooth muscle cells, we measured responses to hyposmotic bath solution, a commonly used method to induce cell swelling.^8,22–24 Experiments were performed using solutions that abolish Na⁺ and K⁺ currents (Online Table I). In isosmotic solution (300 mOsm/L), smooth muscle cells generated outwardly rectifying currents (Figure 1A and 1B). Switching to hyposmotic bath solution (250 mOsm/L) increased currents from 4.0 ± 0.5 to 11.3 ± 2.0 pA/pF or 2.8-fold (at +80 mV) and linearized the current-voltage (I-V) relationship, reducing the rectification index (I₈₀/I₋₈₀) from ≈3.1 to 2.3 (Figure 1A and 1B). The swelling-activated current peaked in ≈90 seconds (Online Figure I). Equimolar replacement of 126 mmol/L external Cl⁻ with I⁻ caused a −26.1 ± 1.7-mV hyperpolarizing shift in E_rev, indicating a P_I/P_Cl of 2.9 (Figure 1B). In contrast, equimolar replacement of 126 mmol/L external Cl⁻ with aspartate shifted E_rev by +13.2 ± 4.2 mV, indicating a P_Asp/P_Cl of 0.6 (Figure 1B). These data indicate that cell swelling activates a Cl⁻ current (I_Cl) in arterial smooth muscle cells.

Next, we examined the molecular identity of channels that generate swelling-activated Cl⁻ currents in arterial smooth muscle cells. A TMEM16A antibody that inhibits currents generated by recombinant TMEM16A channels¹² reduced mean swelling-activated Cl⁻ currents in smooth muscle cells from 12.5 ± 1.9 to 4.0 ± 0.9 pA/pF, or by ≈71% (at +80 mV; Figure 1C and 1D). In contrast, boiled TMEM16A antibody did not alter swelling-activated Cl⁻ currents (Figure 1D). RNAi was used as a complementary approach. Western blotting indicated that TMEM16A siRNA reduced arterial TMEM16A protein by 43 ± 3% when compared with control siRNA (Figure 2A). In contrast, TMEM16A siRNA did not alter TRPC6, TRPM4, or TRPP2 expression (Figure 2B and Online Figure II). Hyposmotic (250 mOsm/L) bath solution increased Cl⁻ currents from 4.2 ± 0.8 to 14.1 ± 2.7 pA/pF (at +80 mV), or ≈3.3-fold in control siRNA-treated smooth muscle cells (Figure 2C and 2D). In contrast, hyposmotic bath solution increased Cl⁻ current density only 1.9-fold (at +80 mV) in smooth muscle cells in which TMEM16A was knocked down. In hyposmotic bath solution, mean Cl⁻ current density in TMEM siRNA-treated cells was ≈30% of that in control siRNA-treated cells (Figure 2D). These results indicate that cell swelling activates TMEM16A currents in arterial smooth muscle cells.

Mechanisms by which cell swelling activates TMEM16A currents in arterial smooth muscle cells were investigated. Extracellular Ca²⁺ removal abolished swelling-activated TMEM16A currents, whereas nimodipine, a voltage-dependent Ca²⁺ channel inhibitor, and thapsigargin, a SR Ca²⁺-ATPase inhibitor that depletes SR Ca²⁺ load, did not alter swelling-activated TMEM16A currents (Figure 3A and 3B). These data suggest that swelling activates TMEM16A currents by inducing Ca²⁺ influx through a pathway that is independent of voltage-dependent Ca²⁺ channels. Therefore, we tested the hypothesis that swelling activates 1 or more nonselective cation channels, leading to Ca²⁺ influx that activates TMEM16A currents. SKF-96365 and Gd³⁺, nonselective cation channel blockers, inhibited swelling-induced TMEM16A currents in smooth muscle cells. In contrast, SKF-96365 and Gd³⁺ did not alter currents generated by recombinant TMEM16A channels expressed in HEK293 cells (Online Figure III). Next, we studied whether Ca²⁺ influx activates TMEM16A channels via a local signaling mechanism. This hypothesis has merit because the Ca²⁺-sensitivity of recombinant TMEM16A channels is lower than the physiological global [Ca²⁺]_i range in arterial smooth muscle cells.^5,10 Consistent with our hypothesis, equimolar replacement of pipette solution EGTA for BAPTA, a fast Ca²⁺ chelator, abolished swelling-activated TMEM16A currents. These data indicate that swelling activates nonselective cation channels, which generate a local intracellular Ca²⁺ signal that stimulates TMEM16A currents, in arterial smooth muscle cells.

As an alternative approach to investigate I_Cl activation by osmolarity, negative pressure was applied to cell-attached patches to induce membrane stretch. These experiments were performed using a pipette solution containing physiological 140 mmol/L NaCl and 2 mmol/L Ca²⁺. Experiments were performed at −80 mV with the pipette solution containing 1 mmol/L TEA⁺ and 5 mmol/L 4-aminopyridine to block large-conductance Ca²⁺-activated K⁺ (BK_Ca) and voltage-dependent K⁺ (Kv) channels, respectively. The bath solution contained 140 mmol/L KCl to depolarize smooth muscle cells and permit efficient voltage-clamp of the plasma membrane patch within the pipette. Application of −40 mm Hg pipette pressure reversibly activated channels in membrane patches (Figure 4A through 4E). In 29 of 37 patches, −40 mm Hg pressure stimulated multiple simultaneously gating ion channels (Figure 4A). Averaging these recordings indicated that pressure stimulated a mean current of ≈2.0 ± 0.3 pA (Figure 4A). To determine the contribution of TMEM16A channels to these currents, experiments were repeated in the presence of the TMEM16A inhibitory antibody. The TMEM16A antibody reduced pressure-induced membrane currents by ≈50% (P<0.05, Figure 4A). In 8 of 37 patches, −40 mm Hg pressure activated single channels that when analyzed using all-points histograms had a mean amplitude of 0.36 ± 0.01 pA (Figure 4B and 4C). Inclusion of the TMEM16A antibody in the pipette solution inhibited these single channels (Figure 4B and 4C). Next, we compared the properties of these stretch-activated single channels to recombinant TMEM16A channels expressed in HEK293 cells. At −80 mV, the amplitude of single recombinant TMEM16A channels was 0.37 ± 0.02 pA (Figure 4D and 4E). This amplitude is almost identical to that of the stretch-activated, TMEM16A antibody–inhibited channels in smooth muscle cells (Figure 4B and 4C). The TMEM16A antibody also blocked recombinant TMEM16A channels (Figure 4D and 4E). These data indicate that pressure-induced membrane stretch activates TMEM16A channels in arterial smooth muscle cells.

To examine physiological functions of smooth muscle cell TMEM16A channels, membrane potential regulation by intravascular pressure was measured in endothelium-denuded arteries. At 10 mm Hg, TMEM16A knockdown did not alter mean arterial smooth muscle cell membrane potential (Figure 5A and 5B). An intravascular pressure elevation from 10 to 60 mm Hg depolarized control siRNA-treated arteries from −66.8 ± 2.3 to −35.5 ± 1.5 mV, or by ≈31.3 mV (Figure 5A and 5B). In contrast, the same pressure elevation depolarized TMEM16A siRNA-treated arteries by ≈15.6 mV, or by ≈50% of that in control arteries (Figure 5A and 5B).

Myogenic tone was measured at intravascular pressures between 20 and 100 mm Hg in endothelium-denuded arteries. Elevating intravascular pressure induced a graded elevation in myogenic tone in control siRNA-treated arteries (Figure 6A and 6B). TMEM16A knockdown reduced myogenic tone at pressures between 40 and 100 mm Hg by between 43% and 49% (Figure 6A and 6B). In contrast, TMEM16A knockdown did not alter membrane depolarization-induced (60 mmol/L K⁺) vasoconstriction (Figure 6C). TMEM16A knockdown also did not alter passive arterial diameter (control siRNA, 243.1 ± 11.6 μm, n=7; TMEM16A siRNA, 238.4 ± 11.3 μm, n=7; P>0.05). These data indicate that intravascular pressure–induced smooth muscle cell TMEM16A channel activation contributes to arterial depolarization and thus, vasoconstriction.