Hydrogen peroxide suppresses TRPM4 trafficking to the apical membrane in mouse cortical collecting duct principal cells

Abstract

A Ca²⁺-activated nonselective cation channel (NSC_Ca) is found in principal cells of the mouse cortical collecting duct (CCD). However, the molecular identity of this channel remains unclear. We used mpkCCD_c14 cells, a mouse CCD principal cell line, to determine whether NSC_Ca represents the transient receptor potential (TRP) channel, the melastatin subfamily 4 (TRPM4). A Ca²⁺-sensitive single-channel current was observed in inside-out patches excised from the apical membrane of mpkCCD_c14 cells. Like TRPM4 channels found in other cell types, this channel has an equal permeability for Na⁺ and K⁺ and has a linear current-voltage relationship with a slope conductance of ~23 pS. The channel was inhibited by a specific TRPM4 inhibitor, 9-phenanthrol. Moreover, the frequency of observing this channel was dramatically decreased in TRPM4 knockdown mpkCCD_c14 cells. Unlike those previously reported in other cell types, the TRPM4 in mpkCCD_c14 cells was unable to be activated by hydrogen peroxide (H₂O₂). Conversely, after treatment with H₂O₂, TRPM4 density in the apical membrane of mpkCCD_c14 cells was significantly decreased. The channel in intact cell-attached patches was activated by ionomycin (a Ca²⁺ ionophore), but not by ATP (a purinergic P₂ receptor agonist). These data suggest that the NSC_Ca current previously described in CCD principal cells is actually carried through TRPM4 channels. However, the physiological role of this channel in the CCD remains to be further determined.

the collecting duct is responsible for absorption of up to 5% of the fluid filtered by the glomerulus, which is critically important for final adjustment of extracellular fluid volume and blood pressure. The cortical collecting duct (CCD) is composed of both principal and intercalated cells. The principal cells reabsorb Na⁺ transcellularly through epithelial sodium channels (ENaC) and secrete K⁺ via potassium channels. However, the renal epithelial cells, including CCD principal cells, also express a Ca²⁺-activated, nonselective cation channel (NSC_Ca), which is permeable to both Na⁺ and K⁺ (9, 12). Since it is activated by intracellular Ca²⁺, this channel may be regulated by receptor-mediated elevation of intracellular Ca²⁺. We and others have shown that the purinergic P2Y₂ receptor downregulates ENaC in principal cells via a phospholipase C-dependent reduction of membrane phosphatidylinositol-4, 5-bisphosphate (PIP₂) (17, 18, 24, 25). However, P2Y₂ receptor-induced PIP₂ hydrolysis also stimulates Ca²⁺ release from intracellular stores via its product inositol triphosphate (15, 36, 41). Therefore, P2Y₂ receptors may regulate the NSC_Ca channels by elevating intracellular Ca²⁺ in CCD cells.

The two melastatin subfamily members of transient receptor potential channels (TRPM), TRPM4 and TRPM5, are Ca²⁺-activated, nonselective cation channels with unitary conductance of ~20-25-pS (10, 13, 27). Therefore, it is likely that either TRPM4 or TRPM5 is the molecular identity of NSC_Ca. Although both TRPM4 and TRPM5 can be activated by intracellular Ca²⁺, their activation patterns are different. TRPM4 is constitutively activated by intracellular Ca²⁺; in contrast, TRPM5 is transiently activated and shows decreased activity when intracellular Ca²⁺ levels are higher than 1 μM (27). Furthermore, TRPM4 is inhibited by intracellular ATP, while TRPM5 is insensitive to intracellular ATP (35). Therefore, besides the gene-silencing method, these different characteristics can be used to determine which type of TRPM represents the previously described NSC_Ca channel.

The physiological role of NSC_Ca channels in CCD has remained unclear for more than two decades. Determining the molecular identity of NSC_Ca channels would provide important information to uncover its physiological role. If the NSC_Ca channel is TRPM4, it should control CCD principal cell viability, because TRPM4 channel activity is closely associated with cell death (31). Inhibition of TRPM4 prevents lipopolysaccharide-induced endothelial cell death (3). Hydrogen peroxide (H₂O₂) stimulates TRPM4 channels in a variety of cell types including endothelial cells (28) and cardiomyocytes (23). H₂O₂ is elevated in the rat kidney after high salt challenge (33). Therefore, activation of TRPM4 by H₂O₂ may mediate the kidney damage caused by oxidative stress.

The present study shows that the NSC_Ca current previously described in CCD is carried through TRPM4 and that H₂O₂ does not stimulate TRPM4 in CCD cells, but conversely reduces TRPM4 density in the apical membrane. The physiological role of TRPM4 in CCD remains to be further determined.

METHODS

Cell culture.

The mpkCCDc₁₄ line is an immortalized mouse collecting duct principal cell line, which was cultured as described previously (4). These cells were cultured in a 1:1 mixture of DMEM and Ham’s F-12 medium (GIBCO) supplemented with 20 mM HEPES, 2 mM l-glutamine, 50 nM dexamethasone, 1 nM triiodothyronine, 2% heat-inactivated FBS, and 0.1% penicillin-streptomycin. The mpkCCD_c14 cells were plated at a density of 75,000 cells/cm and grown on permeable supports to maintain cell polarization (Costar Transwells; 0.4-µm pore, 24-mm diameter) and cultured for at least 7 days before the experiments. To knock down the TRPM4 expression, the mpkCCD_c14 cells were transiently transfected with TRPM4 silencing short hairpin RNA (shRNA) carried by a lentiviral vector [TRCN0000068683, Sigma-Aldrich (5′-CCGGCCTGGGTAATGTGGTCAGTTACTCGAGTAACTGACCACATTACCCAGGTTTTTG-3′)] and scramble shRNA (5′-CCGGCAACAAGATGAAGAGCACCAACTCGAGTTGGTGCT CTTCATCTTGTTGTTTTT-3′) carried by a lentiviral vector (SHC002V, Sigma-Aldrich). The efficiency of TRPM4 knockdown in mpkCCD_c14 cells was confirmed by Western blot analysis before the designated experiments.

Patch-clamp recording.

Single-channel currents were recorded from mpkCCDc₁₄ cells under the voltage-clamp mode with an Axopatch-200B amplifier (Molecular Devices, Sunnyvale, CA), using the cell-attached and excised inside-out patch-clamp configurations, as described previously (37, 43). Data were acquired and sampled with a low-pass, 1-kHz, eight-pole Bessel filter using a Digidata 1440A analog-digital interface (Axon Instruments). MpkCCDc₁₄ cells were thoroughly washed with NaCl solution containing (in mM) 145 NaCl, 1 MgCl₂, 1 CaCl₂, 10 glucose, and 10 HEPES, pH adjusted 7.4 with NaOH. This NaCl solution was used for filling the bath in the patch chamber and filling the patch pipette. The patch pipette was pulled with borosilicate glass, giving a tip resistance of 5–8 MΩ when filled with NaCl solution. Only the patches with a seal resistance >2 GΩ were used. Experiments were conducted at room temperature (22–25°C). Before analysis, the single-channel traces were further filtered at 100 Hz. The single-channel amplitude was constructed by all-point amplitude histogram, and the histograms were fit using multiple Gaussians and optimized using a simplex algorithm. The open probability (P_O) was calculated as P_O = NP_O/N, where N (N was estimated by the current-amplitude histogram during at least a 5-min recording period) is the apparent number of active channels in the patch. The current-voltage (I–V) relationships were constructed using the single-channel amplitude (I) at the indicated voltages as a function of voltages, and the slope conductance was fit with linear regression using SigmaPlot software (Jandel Scientific).

For ion selectivity experiments, 145 mM NaCl in the bath was replaced by 145 mM KCl, or 145 mM NaCl in the pipettes was replaced by 145 mM NMDG-Cl, and the data were corrected for junction potentials at the ground bridge (3 M KCl in 3% agar). The free Ca²⁺ concentration after chelation of CaCl₂ with EGTA was determined using the free Web software Winamac (Stanford University, Stanford, CA), as previously described (40).

Biotinylation and Western blot assay.

Biotinylation assays of the plasma membrane were performed as described before (14, 16). Briefly, after each treatment the mpkCCDc₁₄ cells were incubated with a freshly prepared solution of 1.0 mg/mL EZ-Link sulfo-N-hydroxysuccinimide disulfide-biotin (Pierce, 21331) in borate buffer for 30 min at 4°C. The biotin reaction was quenched for 5 min with 0.1 mM lysine. An equal amount of lysate protein (either 0.5 or 1 mg, depending on which experiment) from each sample was respectively incubated with 25 µl of immobilized streptavidin-agarose beads (20349, Pierce) at 4°C for overnight with gentle shaking. The beads were washed four times with RIPA buffer. Cell lysates from either whole-cell (50 μg) or biotinylated plasma membranes (30 μg) were loaded and separated by a 10% SDS-polyacrylamide gel and transferred to polyvinylidene fluoride membranes. The membranes were then blocked in 5% nonfat dry milk for 1 h, followed by incubation with rabbit polyclonal anti-TRPM4 antibody (1:200 dilution; ACC-044; Alomone Laboratories) at 4°C overnight. GAPDH and β-actin were used as internal controls. Antibodies to GAPDH and β-actin were purchased from Santa Cruz Biotechnology (sc-25778 and sc-1615) and were used in 1:1,000 dilutions.

Chemicals.

All chemicals for electrophysiological recordings were purchased from Sigma-Aldrich (St Louis, MO) except when specified.

Data analysis.

Data are reported as means ± SE. Statistical analysis was performed with SigmaPlot and SigmaStat software (Jandel Scientific). Student's t-test was used between two groups. Analysis of variance was used for multiple comparisons. Results were considered significant if P < 0.05.

RESULTS

NSC_Ca is present in mpkCCD_c14 cells.

Previous studies have shown that an NSC_Ca channel was observed in M1 mouse CCD cells (12). However, it is unknown whether the cultured mpkCCDc₁₄ cells also contain such a channel. Therefore, we used the excised inside-out patch-clamp technique to determine whether an NSC_Ca single-channel current can be detected in mpkCCDc₁₄. In cell-attached patches, there were no channel openings even when +40 mV was applied to the patch pipette (−V_pipette = −40 mV), presumably due to low intracellular Ca²⁺ concentration ([Ca²⁺]_i). However, after formation of the inside-out patches (exposure of the patch membrane to 1 mM Ca²⁺ in the bath), single-channel currents were immediately detected (top trace; Fig. 1A). No current was observed when no voltage existed and Na⁺ concentration is symmetrical across the patch membrane (middle trace; Fig. 1A). Switching the holding potential to +40 mV (−V_pipette) reversed the inward current to an outward current (bottom trace; Fig. 1A). To determine whether the current is carried by cations or anions, the pipettes were filled with a solution containing 145 mM NMDG-Cl to replace Na⁺ in the pipettes. Under this condition, formation of the inside-out patch at the holding potential of +40 mV led to an outward current, which is caused by Na⁺ efflux into the patch pipette due to the Na⁺ concentration gradient (top trace; Fig. 1B); however, no current was observed at a holding potential of +80 mV, which is the calculated reversal potential of Na⁺, suggesting that the current was carried by cations (bottom trace; Fig. 1B).

Fig. 1.

Formation of inside-out patches activates a cation channel in mpkCCD_c14 cells. A: representative currents were recorded using a symmetrical 145 mM NaCl solution. Excision of a cell-attached patch into a 145 mM NaCl solution containing 1 mM Ca²⁺ resulted in activation of inward single-channel current when the patch was voltage clamped at a holding potential of −40 mV (−V_pipette = −40 mV; top trace), but no current was observed at 0 mV (middle trace). Switching the holding potential to +40 mV reversed the inward current to an outward current (−V_pipette = +40 mV; bottom trace). Such channel activity was repeatedly observed in 10 of 17 inside-out patches; no channels were observed in the other 7 patches (empty patches). B: after replacement of 145 mM Na⁺ in the patch pipette with 145 mM NMDG⁺, formation of the inside-out patch mode induced an outward current at a holding potential of −40 mV (−V_pipette = −40 mV; top trace), and no current was observed at −80 mV (−V_pipette = −80 mV; bottom trace). Such channel activity was repeatedly observed in 8 of 13 inside-out patches; no channels were observed in the other 5 patches (empty patches).

We further examined the Ca²⁺ sensitivity of this channel by exposing the patch membrane to the bath containing different concentrations of free Ca²⁺. Consistently, this current was observed in the inside-out patch mode in the presence of 10⁻³ M free Ca²⁺ at a holding potential of +40 mV; however, the channel activity in the same patch almost disappeared when the bath solution was replaced with a solution containing 10⁻⁷ M free Ca²⁺ (Fig. 2A; top trace). The channel P_O was tremendously decreased from 0.81 ± 0.04 to almost undetectable levels (Fig. 2B; P < 0.001; n = 7). Conversely, the channel activity was significantly increased after Ca²⁺ concentration in the bath was increased from 10⁻⁵ to 10⁻³ M. (Fig. 2A; bottom trace). P_O was dramatically increased, from 0.14 ± 0.02 to 0.81 ± 0.05 (Fig. 2B; P < 0.001; n = 6). The P_Os in response to different concentrations of bath free Ca²⁺ were plotted and fitted with the Hill equation (Fig. 2C; n = 5), showing that Ca²⁺ activates this channel with an EC₅₀ of 32.6 μM. These data suggest that this cation-permeable channel is activated by intracellular Ca²⁺, but the sensitivity is relatively low.

Fig. 2.

Ca²⁺ activates this cation channel in a dose-dependent manner. A: representative single-channel currents recorded either from 2 inside-out patches; one was recorded before and after replacing the bath solution containing 10⁻³ M Ca²⁺ with a similar solution but containing 10⁻⁷ M Ca²⁺ (top trace), whereas the other was recorded before and after replacing the bath solution containing 10⁻⁵ M Ca²⁺ with a similar solution but containing 10⁻³ M Ca²⁺ (bottom trace). B: summary plots of mean open probability (P_O) under each condition. C: dose-dependent elevation of channel activity by Ca²⁺. Channel P_O was plotted as a function of Ca²⁺ concentration ([Ca²⁺]) in the bath. Data were fitted with the Hill equation.

To determine single-channel conductance and ion selectivity, the Ca²⁺-activated single-channel currents were examined at different holding potentials using different solutions (Fig. 3). The inside-out patch was exposed first to a bath solution containing 145 mM NaCl and 10⁻³ M free Ca²⁺ (Fig. 3A), and then the bath solution was replaced with another solution containing 145 mM KCl and 10⁻³ M free Ca²⁺. The data show that the substitution of Na⁺ with K⁺ did not alter the reversal potential of the channel (Fig. 3B). The channel exhibited a linear I–V with the reversal potential near 0 mV and a slope conductance of either 23.4 pS when the bath contained 145 mM Na⁺ or 21.8 pS (n = 6) when the bath contained 145 mM K⁺ (n = 5). However, the substitution of Na⁺ in the patch pipette with NMDG⁺, a big cation, shifted the reversal potential from near zero to −78 mV (n = 4). These results suggest that the channel exhibits equal permeability for Na⁺ and K⁺ and is impermeable to NMDG⁺. Taking the above data together, we conclude that NSC_Ca is present in mpkCCD_c14 cells.

Fig. 3.

Replacement of Na⁺ with membrane-impermeable NMDG⁺, but not K⁺, shifted the reversal potential. A: representative single-channel currents recorded from an inside-out patch exposed to a 145 mM NaCl solution containing 1 mM free Ca²⁺ at different holding potentials. B: representative single-channel currents recorded from another inside-out patch exposed to a 145 mM KCl solution containing 1 mM free Ca²⁺ at different holding potentials. C: current-voltage relationships were plotted with single-channel amplitudes as a function of potentials applied to patch pipettes, when patch pipettes contained a 145 mM NaCl solution while the bath contained either the same NaCl solution (circles) or a 145 mM KCl solution (squares), or when patch pipettes contained 145 mM NMDG-Cl while the bath contained a 145 mM NaCl solution (triangles). Fitting the data with the linear regression (solid line through the data) shows a similar slope conductance under each condition.

NSC_Ca is not affected by extracellular ATP, but inhibited by intracellular ATP.

Previous studies have shown that CCD principal cells express a purinergic P2Y₂ receptor (26). Since activation of P2Y₂ receptors causes Ca²⁺ release from its intracellular stores (38), we hypothesized that ATP might stimulate the NSC_Ca channel by inducing Ca²⁺ release in mpkCCD_c14 cells. Surprisingly, the data show that in cell-attached patches application of 100 μM ATP to the bath did not induce any channel activity (Fig. 4A; top trace) in a total of five cell-attached patches (Fig. 4B). In contrast, application of 5 μM ionomycin (a Ca²⁺ ionophore) to the bath activated the channel (Fig. 4A; bottom trace). The P_O was increased, from undetectable to 0.34 ± 0.06 (P < 0.01; n = 5). Interestingly, in inside-out patches the channel activity of NSC_Ca was almost completely inhibited after the inner side of the patch membrane was exposed to 1 mM ATP in the bath; the inhibition was immediately reversed after ATP was washed out of the bath (Fig. 4C). Mean P_O was decreased, from 0.78 ± 0.06 to 0.02 ± 0.01 (P < 0.01; n = 6), and recovered to 0.77 ± 0.06 after the wash (Fig. 4D). These data suggest that the NSC_Ca channel is not affected by Ca²⁺ release from its intracellular stores induced by extracellular ATP, but activated by Ca²⁺ influx from the extracellular side of the cells induced by ionomycin.

Fig. 4.

ATP inside the cell inhibits channel activity. A: cell-attached single-channel recordings before and after application of either 100 μM ATP (top trace) or 5 μM ionomycin (bottom trace) to the bath. B: summary plots of P_O calculated from recordings under each condition (ATP_e indicates that ATP was applied to extracellular side of the cell). C: inside-out single-channel recordings before and after application of 1 mM ATP to the bath, and after washout of ATP. D: summary plots of P_O calculated from recordings under each condition (ATP_i indicates that ATP was applied to intracellular side of the cell).

NSC_Ca is inhibited by a TRPM4 inhibitor and diminished by TRPM4 knockdown.

The biophysical characteristics of the NSC_Ca channel we found in in mpkCCD_c14 cells are similar to those of the TRPM4 channel (13). Moreover, our previous studies have shown that TRPM4 mRNA is present in mouse CCD principal cells (43). To determine whether TRPM4 is the molecular identity of NSC_Ca, we first used a specific TRPM4 inhibitor, 9-phenanthrol. The data show that NSC_Ca activity was dramatically reduced after application of 100 µM 9-phenanthrol to the bath (Fig. 5A). Mean P_O was significantly decreased, from 0.57 ± 0.08 to 0.15 ± 0.04 in the first 30 s and to 0.02 ± 0.01 in the second 30 s (n = 7; P < 0.01) (Fig. 5B). Second, we used TRPM4 silencing shRNA carried by a lentiviral vector to knock down TRPM4 in mpkCCD_c14 cells. Scramble shRNA (the sequences shown in methods) served as a control. After knockdown of TRPM4, most patches did not contain NSC_Ca channels (Fig. 5C). The frequency of observing NSC_Ca channels in inside-out patches was significantly decreased, from 63.6% (14 of 22 patches) to 31.8% (7 of 24 patches) (Fig. 5D). Western blotting confirmed that TRPM4 shRNA significantly reduced TRPM4 expression. Consistent with previous studies (6, 20), only one specific band of TRPM4 was shown in the blot. These data together suggest that the molecular identity of the NSC_Ca in mpkCCDc14 cells is TRPM4.

Fig. 5.

Transient receptor potential channel, the melastatin subfamily 4 (TRPM4) blockade and knockdown reduce channel activity. A: representative single-channel current recorded from an inside-out patch before and after replacing the bath solution with a similar solution but containing 100 μM 9-phenanthrol (a specific TRPM4 blocker). B: summary plots of mean P_O. C: representative single-channel currents recorded from either a control cell (top trace) or a TRPM4 knockdown cell (bottom trace). D: summary plots of percentage of patches containing channels in control and TRPM4 knockdown cells. E: Western blot of TRPM4 from cells under control conditions or transfected with either control RNA or TRPM4 short hairpin (sh) RNA (left). Data represent 3 separate experiments, showing that TRPM4 shRNA significantly reduced TRPM4 expression (right).

H₂O₂ reduces TRPM4 density in the apical membrane of mpkCCD_c14 cells.

Recent studies suggest that elevation of H₂O₂ in the kidney participates in salt-sensitive hypertension (19) and that H₂O₂ stimulates TRPM4 in other type of cells (23, 28). Surprisingly, we found that acute application of H₂O₂ did not affect TRPM4 channel activity in mpkCCDc14 cells, no matter whether the bath contained either 10⁻³ or 10⁻⁶ M free Ca²⁺ (Fig. 6A). As shown in Fig. 6B, the mean P_Os were 0.86 ± 0.04 (before) vs. 0.83 ± 0.04 (after 100 μM H₂O₂; P = 0.8; n = 7) and 0.81 ± 0.05 (before) vs. 0.80 ± 0.05 (after 500 μM H₂O₂; P = 0.8; n = 7); even when TRPM4 was only slightly activated by 10⁻⁶ M free Ca²⁺ in the bath, 500 μM H₂O₂ still failed to activate TRPM4; mean Po remained at low levels (0.03 ± 0.01 vs. 0.03 ± 0.01; P = 1.0; n = 6). These data suggest that intracellular H₂O₂, at least, does not acutely regulate TRPM4 activity in mpkCCD_c14 cells.

Fig. 6.

H₂O₂ does not affect TRPM4 channel activity in mpkCCD_c14 cells. A: representative single-channel currents recorded from 3 inside-out patches. Addition of 100 or 500 μM H₂O₂ to the bath did not alter channel activity, no matter whether the channel was strongly activated by 10⁻³ M Ca²⁺ (top and middle traces) or slightly activated by 10⁻⁶ M Ca²⁺ (bottom trace). B: summary plots of P_O calculated from recordings before and after application of H₂O₂ under each condition.

We then examined whether chronic application of H₂O₂ can affect TRPM4 density in the apical membrane of mpkCCD_c14 cells. After mpkCCD_c14 cells were treated with 100 μM H₂O₂ for 24 h, both control cells and H₂O₂-treated cells were fixed and stained with an anti-TRPM4 antibody. To set confocal microscopy optical sections across or near the apical membrane, the tight junction protein zonula occludens (ZO)-1 was also labeled with its antibody. Confocal microscopy data from four separate experiments consistently show that the fluorescent intensity of TRPM4 (shown in red) in or near the apical membrane (as marked by ZO-1, shown in green) was dramatically reduced (Fig. 7A). Biotinylation of the apical membrane shows that the apical TRPM4 protein was significantly reduced by 65% after treatment of the cells with H₂O₂ for 24 h, whereas total TRPM4 was not altered (Fig. 7B; n = 6). Patch-clamp data also show that TRMP4 channels were scarcely seen in inside-out patches, even when the bath contained 1 mM free Ca²⁺ (Fig. 7C). The frequency of observing TRPM4 channels was significantly reduced, from 69% (11 of 16 patches in control cells) to 40% (6 of 15 patches in H₂O₂-treated cells) (Fig. 7D). These data together suggest that H₂O₂ reduces TRPM4 density in the apical membrane of mpkCCD_c14 cells.

Fig. 7.

H₂O₂ reduces apical TRPM4 density in mpkCCD_c14 cells. A: representative confocal microscopy images taken from either control mpkCCD_c14 cells (top) or the cells treated with 100 μM H₂O₂ for 24 h (bottom). TRPM4 levels in mpkCCD_c14 cell monolayer were examined through immunostaining the cells with its antibody (shown in red). Optical sections were set at or near the apical membrane according to where the tight junction protein ZO-1 was localized (shown in green). Scale bar = 20 μm. B: representative Western blots of either total or biotinylated TRPM4 protein in control cells or the cells treated with 100 μM H₂O₂ for 24 h; the data represent 5 experiments, showing that biotinylated (apical) TRPM4 was significantly reduced by H₂O₂. C: representative TRPM4 single-channel currents recorded from control cells or the cells treated with 100 μM H₂O₂ for 24 h. D: summary plots of patches containing TRPM4 channels under control condition or after treatment of the cells with H₂O₂.

DISCUSSION

It has long been noticed that NSC_Ca is expressed in CCD cells (12). However, the role of this channel in renal physiology remains unclear, mainly because its molecular identity is unknown. We demonstrate, for the first time, that TRPM4 represents the NSC_Ca previously described in the principal cells of mouse CCD. We conclude that the NSC_Ca current detected in mpkCCD_c14 cells is carried through TRPM4, but not through TRPM5 because 1) ATP at the concentration of 1 mM completely inhibits channel activity; however, the same concentration of ATP does not inhibit TRPM5 (35); 2) TRPM4 and TRPM5 have distinct Ca²⁺ sensitivity: TRPM4 is constitutively activated by high levels of [Ca²⁺]_i, whereas TRPM5 is transiently activated by lower [Ca²⁺]_i and shows reduced activity when the levels of [Ca²⁺]_i is higher than 1 μM (27); and 3) TRPM4 knockdown significantly reduces the channel density in the cells; in addition, NSC_Ca current can be inhibited by 9-phenanthrol (a specific TRPM4 inhibitor). Although the NSC_Ca detected in mpkCCD_c14 cells has a similar single-channel conductance with a channel complex formed by TRPP2 and TRPV4 in mouse CCD (43), even without the gene-silencing data, it is unlikely that the NSC_Ca we showed here represents the TRPP2/TRPV4 channel complex because the TRPP2/TRPM4 channel is insensitive to intracellular ATP (43).

We show that extracellular ATP, which can cause Ca²⁺ release from its intracellular stores, is unable to activate TRPM4. However, ionomycin, which can induce Ca²⁺ influx into the cells, does activate the channel. We argue that the P₂ receptor-caused Ca²⁺ release from intracellular stores is probably unable to reach the threshold concentration which is required to activate TRPM4. In contrast, the ionomycin-induced Ca²⁺ influx appears to strongly elevate local Ca²⁺ to activate TRPM4. Previous studies have shown that a strong elevation of intracellular Ca²⁺ (> 1 μM) is required to activate TRPM4 (13, 21). Our data indicate that the Ca²⁺ sensitivity of TRPM4 channels in the cultured mpkCCD_c14 cells is even lower than those previously reported in other cell types, which has an EC₅₀ of 32.6 μM. Our recent studies suggest that mitochondria can form a band below the apical membrane of CCD principal cells to function as a Ca²⁺ barrier (34). This mitochondria band may take up the released Ca²⁺ which is induced by ATP before it diffuses to the apical membrane where the channel is located. Therefore, we propose that receptor-mediated Ca²⁺ release may not be an important regulator of TRPM4 in CCD principal cells, at least not under normal physiological conditions. Since the Ca²⁺-permeable canonical family of TRP channels (TRPC) is localized in the apical membrane of the CCD (8), Ca²⁺ influx through TRPC channels should provide an efficient pathway for a rapid and strong elevation of intracellular Ca²⁺, especially when the mitochondria band is intact. TRPM4 may be activated in the patients with the focal segmental glomerulosclerosis, because in these patients TRPC6 is constitutively activated due to gain-of-function mutations (39).

Although TRPM4 is strongly inhibited by intracellular ATP in excised inside-out patches, in intact cells intracellular ATP may not be able to completely turn the channel off, but only modulate channel activity. This assumption is supported by the fact that ATP-sensitive potassium channels are also strongly inhibited by intracellular ATP in excised inside-out patches (5, 22), but the channel has spontaneous activity in intact cells (11). Reducing phosphatidylinositol 4,5-bisphosphate is considered to be the reason that causes ATP-sensitive potassium channels to be strongly inhibited by ATP in excised inside-out patches, but not in intact cells (1, 2, 30). This mechanism may be applied to the TRPM4 channel as well, because TRPM4 channels are also regulated by both phosphatidylinositol 4,5-bisphosphate and ATP (7, 42). In other words, in intact cells TRPM4 can be activated by elevation of intracellular Ca²⁺ without reducing ATP in the cells. However, when cell death signals are initiated, both elevation of intracellular Ca²⁺ and reduction of intracellular ATP can occur simultaneously; TRPM4 should be strongly activated under these conditions to facilitate cell death.

In addition to intracellular Ca²⁺ and ATP, TRPM4 is also regulated by oxidative stress. In human umbilical vein endothelial cells, TRPM4 mediates H₂O₂-induced cell migration and adhesion (28). In TRPM4-overexpressing HEK-293 cells, TRPM4 participates in H₂O₂-induced necrosis (32). Therefore, we originally hypothesized that TRPM4 may mediate H₂O₂-induced kidney damage. However, in CCD cells, we show that H₂O₂ does not acutely regulate TRPM4, but chronically reduces apical surface expression of TRPM4. This paradox may serve as an interesting topic for us to investigate: a possible unique regulation of TRPM4 by H₂O₂ in CCD principal cells.

GRANTS

The work was supported by grants from the Key Project of Chinese National Program for Fundamental Research and Development (973 Program 2012CB517803 and 2014CB542401 to Z.-R. Z.), National Nature Science Foundation of China (81320108002 and 81270340 to Z.-R. Z.), and Department of Health and Human Services (National Institutes of Health Grant R01-DK100582 to H. -P. M.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.-M.W., Y.-J.Z., Y.-X.L., Q.-Q.H., Z.-R.W., S.-P.W., and L.Z. performed experiments; M.-M.W., Y.-J.Z., S.-P.W., and H.-P.M. analyzed data; M.-M.W., Y.-J.Z., Z.-R.Z., and H.-P.M. prepared figures; M.-M.W., Y.-J.Z., T.L.T., Z.-R.Z., and H.-P.M. drafted manuscript; M.-M.W., Z.-R.Z., and H.-P.M. approved final version of manuscript; Y.-J.Z., A.A.A., T.L.T., Z.-R.Z., and H.-P.M. interpreted results of experiments; A.A.A., T.L.T., Z.-R.Z., and H.-P.M. edited and revised manuscript.