Transient receptor potential V4 channel stimulation induces reversible epithelial cell permeability in MDCK cell monolayers

Abstract

The transient receptor potential V4 channel (TRPV4) is responsive to a variety of physical and chemical stimuli, including a synthetic agonist GSK1016790A (GSK). Here, we show that TRPV4 is functionally expressed in, and that GSK induces the reversible opening of tight junctions (TJs) in epithelial Madin–Darby canine kidney II monolayers. Stimulation of TRPV4 by GSK induces an increase in fluorescein isothiocyanate‐conjugated dextran (4 kDa) permeability and a reduction in transepithelial resistance, and these responses are blocked by pretreatment with the specific TRPV4 antagonist. Small conductance, but not large conductance Ca²⁺‐activated K⁺ channels, TRPA1 channel, and cofilin activation are involved in TRPV4‐mediated reversible opening of TJs. These results suggest that a novel mechanism underlies TRPV4‐mediated regulation of the tightness of epithelial barriers.

Abbreviations

4α‐PDD, 4α‐phorbol 12,13‐didecanoate

A96, A‐967079

AITC, allyl isothiocyanate

BK channels, large conductance Ca²⁺‐activated K⁺ channels

FD4, fluorescein isothiocyanate‐conjugated dextran (4 kDa)

GSK, GSK1016790A

HBSS, Hank's Balanced Salt Solution

HC, HC‐067047

HRP, horseradish peroxidase

LatA, latrunculin A

MDCK, Madin–Darby canine kidney

RuRed, ruthenium red

SK channels, small conductance Ca²⁺‐activated K⁺ channels

TER, transepithelial electrical resistance

TJ, tight junction

TRP, transient receptor potential

Transient receptor potential V4 channel (TRPV4) is a nonselective cation channel with a moderate preference for Ca²⁺. TRPV4 exhibits a polymodal gating and is responsive to a large variety of physical and chemical stimuli, including moderate heat, hypotonicity‐induced cell swelling, flow, endogenous ligands, and synthetic agonists, such as GSK1016790A (GSK) [1]. TRPV4 has been found in both excitable and nonexcitable cells in many different tissues, where it regulates diverse cellular functions. For instance, TRPV4 has been shown to regulate the barrier function of endothelial [2, 3], and epithelial cells [4-10]. In skin keratinocytes, TRPV4 activation is important for tight junction (TJ) establishment and also strengthens TJ barrier function [4, 5, 8]. Interestingly, it has also been reported that TRPV4 activation leads to increased paracellular permeability or decreased transepithelial electrical resistance (TER) in mammary and intestinal epithelial cell lines [7, 10], although our study here shows the different patterns of epithelial permeability increase and the mode of actions.

We have been interested in the regulation of TJ permeability and have analyzed the capsaicin‐induced reversible increase in paracellular permeability [11-16]. Recently, we reported that capsaicin and natural compounds containing an α,β‐unsaturated moiety open TJs reversibly via TRPA1 activation, but not by TRPV1 activation [17] in monolayers of the canine kidney epithelial cell line Madin–Darby canine kidney (MDCK) II, although capsaicin is an agonist of the TRPV1 cation channel.

Here, we functionally characterize TRPV4 in MDCK II epithelial cells. TRPV4 mRNA expression was confirmed, and its activation a specific agonist, GSK, resulted in rapid Ca²⁺ influx. Treatment of the MDCK II monolayer with GSK induced a reversible increase in fluorescein isothiocyanate‐conjugated dextran (4 kDa; FD4) permeability and a reversible decrease of TER. These responses were blocked by pretreatment with a specific TRPV4 antagonist, HC‐067047 (HC) [18], suggesting that TRPV4 activation is involved in TJ opening. Ca²⁺ entry through TRPA1 and cofilin activation are involved in capsaicin‐induced reversible opening of TJs. Although Ca²⁺ entry via GSK triggers an immediate decrease in TER and an increase in FD4 permeability, cofilin activation did not occur at all. Our data show that TRPV4 stimulation induces an increase in reversible paracellular permeability in MDCK II monolayers by a different mechanism from TRPA1 activation.

Materials and methods

Cell culture and reagents

Madin–Darby canine kidney II cells were cultured in Dulbecco's Modified Eagle's Medium (Nacalai Tesque, Kyoto, Japan) supplemented with 10% fetal bovine serum (HyClone, GE Healthcare, Buckinghamshire, UK) and 1% penicillin–streptomycin (Nacalai Tesque) in a humidified atmosphere containing 5% CO₂.

GSK, HC, paxilline, and A‐967079 were from Cayman (Ann Arbor, MI, USA). Capsaicin and FD4 were purchased from Sigma (St. Louis, MO, USA). Latrunculin A (LatA) and ruthenium red (RuRed) were from Wako Pure Chemical Industries (Osaka, Japan). Ionomycin and allyl isothiocyanate (AITC) were purchased from LKT Laboratories (St. Paul, MN, USA) and Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan) respectively.

Anti‐cofilin antibody (#3312) was purchased from Cell Signaling Technology (Beverly, MA, USA). Anti‐β‐actin (#125K4769) and anti‐claudin‐2 (#SAB4503544), anti‐phospho‐cofilin (Ser 3, #sc‐21867‐R), anti‐claudin‐1 (#13050‐1‐AP), and anti‐TRPV4 (#ab94868) antibodies were from Sigma, Santa Cruz Biotechnology (Santa Cruz, CA, USA), Proteintech Group, Inc. (Rosemont, IL, USA), and Abcam (Cambrigde, UK) respectively. Anti‐occludin (#71‐1500) and anti‐claudin‐4 (#329400) antibodies were from Invitrogen (Grand Island, NY, USA). Horseradish peroxidase (HRP)‐conjugated anti‐mouse and anti‐rabbit IgGs (#5220‐0341 and #5220‐0336, respectively) were from Kirkegaard & Perry Laboratories (Gaithersburg, MD, USA).

All other reagents were of reagent grade and purchased from Nacalai Tesque unless otherwise noted.

RNA preparation and PCR

RNA was isolated from a MDCK II monolayer established as described below using a NucleoSpin RNA II kit (Macherey‐Nagel, Düren, Germany) according to the manufacturer's protocol. First‐strand cDNA was synthesized using a PrimeScript II 1st strand cDNA synthesis kit (TAKARA, Kyoto, Japan). PCR was performed using KOD Dash (TOYOBO, Osaka, Japan) in a TAKARA PCR Thermal Cycler Dice^® Gradient with specific primer sets as follows: GAPDH, 5′‐GAAAGCTGCCAAATATGACGACATC‐3′ (forward) and 5′‐TCCAGGAGGCTCTTACTCCTTG‐3′ (reverse); TRPV4, 5′‐GGACGAGGTGAACTGGTCTCACT‐3′ (forward) and 5′‐GGAGCATCGTCAGTCCTCCACT‐3′ (reverse); and TRPA1, 5′‐CTACATGCATAATGTAGAGG‐3′ (forward) and 5′‐GGAGCAGATATTAATAGCAC‐3′ (reverse). PCR condition was as follows: 1 cycle at 98 °C for 30 s, 30 cycles at 94 °C for 30 s, 55 °C for 20 s, and 74 °C for 30 s followed by 1 cycle at 74 °C for 4 min.

Cytotoxicity assay with MDCK II monolayer

Madin–Darby canine kidney II cells were seeded into a 96‐well plate (Iwaki, Tokyo, Japan) at a density of 3.4 × 10⁴ cells per well. The cells were cultured for 3 days to establish monolayer integrity and then treated with various concentrations of GSK. After 5 h, 10 μL of Cell Counting Kit‐8 (Dojindo Laboratories, Kumamoto, Japan) reagent was added to each well. The plate was incubated at 37 °C, and absorbance at 450 nm was then measured using an iMark Microplate Reader (Bio‐Rad Laboratories, Inc., Hercules, CA, USA) prior to calculation of IC₅₀ values.

Intracellular Ca²⁺ measurements

Madin–Darby canine kidney II cells were seeded into 96‐well plates with clear bottoms and black walls (Greiner Bio‐One, Frickenhausen, Germany) at a density of 3.4 × 10⁴ cells per well. The cells were cultured for 3 days, with the medium being changed every day to establish monolayer integrity. After experimental treatment, monolayers were washed with Hank's Balanced Salt Solution (HBSS; 136.9 mm NaCl, 5.4 mm KCl, 1.3 mm CaCl₂, 0.8 mm MgSO₄, 0.34 mm Na₂HPO₄, 0.44 mm KH₂PO₄, 1.0 g·mL⁻¹ glucose, and 4.2 mm NaHCO₃) and treated with Fluo‐8 loading buffer [5 μm Fluo‐8 (AAT Bioquest, Sunnyvale, CA, USA), 2.5 mm probenecid (Cosmo Bio, Tokyo, Japan), and 1% F‐127 (Biotium, Hayward, CA, USA) in HBSS] at 37 °C for 1 h. Cells were washed with HBSS twice, and recording buffer (2.5 mm probenecid in HBSS) and various compounds were then added. Fluo‐8 fluorescence was measured every 3 min for 60 min using a PowerscanHT fluorescence microplate reader (Dainippon Sumitomo Pharma, Osaka, Japan) at excitation and emission wavelengths of 485 and 528 nm respectively.

Transport studies

Madin–Darby canine kidney II cells were seeded into 6.5‐mm‐diameter transwells (pore size 0.4 mm; Corning Inc., Corning, NY, USA) coated with collagen at a density of 3.4 × 10⁴ cells per well. The cells were cultured for 3 days to establish monolayer integrity. Transwell plates were washed three times, incubated with HBSS, and equilibrated for 1 h at 37 °C. HBSS containing 1.0% w/v of FD4 was placed on the apical side, and transport enhancer was also added to the apical side. The basolateral side was exposed to HBSS, which was refreshed at predetermined intervals. Samples collected from the basolateral compartments were analyzed for FD4 using a PowerscanHT fluorescence microplate reader at an excitation wavelength of 485 nm and an emission wavelength of 530 nm.

Hank's Balanced Salt Solution lacking CaCl₂ was used in experiments with Ca²⁺‐depleted conditions.

TER measurements

Madin–Darby canine kidney II cells were seeded into 12‐mm‐diameter transwells (pore size 0.4 mm) coated with collagen at a density of 1.1 × 10⁵ cells per well. The cells were cultured for 3 days to establish monolayer integrity. TER was measured as described previously [17].

Hank's Balanced Salt Solution lacking CaCl₂ was used in experiments with Ca²⁺‐depleted conditions, and TER was measured before and after adding drugs.

Localization of FD4 in cell layer

Madin–Darby canine kidney II monolayers grown on eight well coverglass chamber (Iwaki) were washed with HBSS, and incubated with HBSS for 1 h at 37 °C. The monolayers were treated with HBSS (0.2 mL) containing 1.0% w/v of FD4 and drugs for 30 min, with 2 μg·mL⁻¹ Hoechst 33258 (Sigma) for the last 10 min, then washed three times with HBSS, and fluorescence images were acquired with a Leica DMI6000B microscope (Leica Microsystems, Heidelberg, Germany) within 20 min. The z‐stacks were collected at 0.24 μm intervals. Image stacks were deconvoluted using leica application suite x software (Leica Microsystems).

Immunoblotting

Madin–Darby canine kidney II cells were seeded into 24‐well plates (Iwaki) at a density of 2.1 × 10⁵ cells per well. The cells were cultured for 3 days, and the medium was changed every day to establish monolayer integrity. After specific treatments, monolayers were washed once with phosphate‐buffered saline and lysed with SDS sample buffer (62.5 mm Tris‐HCl, pH 6.8, 2% SDS, 10% glycerol, and 5% β‐mercaptoethanol). After sonication, the cell extracts were boiled at 65 °C for 10 min, separated by sodium dodecyl sulfate‐polyacrylamide gel electrophoresis, transferred to a polyvinylidene fluoride microporous membrane (Wako), blocked with 5% skimmed milk (Megmilk Snowbrand, Sapporo, Japan), probed with the appropriate primary and HRP‐conjugated anti‐IgG secondary antibodies, and visualized by enhanced chemiluminescence (Nacalai Tesque). Images were produced using a Sayaka Imager (DRC, Tokyo, Japan). Band intensities of phospho‐cofilin and actin were measured with imagej software (NIH, Bethesda, MD, USA) and normalized to actin.

Data analysis

graphpad prizm 6, (GraphPad Software Inc., San Diego, CA, USA) was used to perform the analysis of variance for multiple comparisons. P < 0.05 was considered to be statistically significant.

Results

TRPV4 is functionally expressed in MDCK II monolayer

We previously demonstrated that TRPA1 is involved in the reversible opening of TJs in MDCK II epithelial monolayers [17]. Because TRPA1 is a TRP channel protein and TRPV4, the other TRP channel protein, is abundantly expressed in the kidney [19], we examined the mRNA expression of TRPV4 in MDCK II monolayers and detected TRPV4 at similar levels as TRPA1 (Fig. 1A). We also examined the expression of TRPV4 on the protein level by western blot analyses using an anti‐TRPV4 antibody. Figure 1B shows the expression of a protein of ~ 90 kDa, which is in good agreement with the others reported [20, 21].

Figure 1

Functional TRPV4 is expressed by MDCK II monolayers. (A) Expressions of TRPV4 and TRPA1 were confirmed by RT‐PCR. Electrophoretic separation of PCR products produced with specific primers is shown. GAPDH was used as a control. Expected product sizes are as follows: GAPDH, 271 bp; TRPV4, 270 bp; and TRPA1, 272 bp. (B) TRPV4 protein expression was analyzed by immuoblotting using anti‐TRPV4 antibody in MDCK II monolayer. (C) Several doses of GSK were employed to analyze Ca²⁺ influx in MDCK II monolayers. (○) DMSO control, (×) 10 μm ionomycin, (▲) 0.3 nm, (●) 1 nm, (■) 3 nm, and (◆) 10 nm GSK. Each value represents the mean ± standard deviation in triplicate experiments. (D) Cells were pretreated with TRPV4 antagonists (10 μm RuRed and 1 μm HC) and TRPA1 antagonist (1 μm A96) for 30 min prior to 1 nm GSK addition. (○) DMSO control, (×) 10 μm ionomycin, (Δ) 10 μm RuRed, (◊)1 μm HC, (□) 1 μm A96, (●) 1 nm GSK, (▲) 10 μm RuRed + 1 nm GSK, and (◆) 1 μm HC + 1 nm GSK, and (■) 1 μm A96 + 1 nm GSK. Each value represents the mean ± standard deviation in triplicate experiments.

To confirm functional TRPV4 expression in MDCK II monolayers, we examined cellular responses to the reported TRPV4 agonist GSK using Fluo‐8, a fluorescent Ca²⁺‐imaging probe (Fig. 1C). The positive control, ionomycin, immediately increased cellular Ca²⁺ concentration and was then attenuated after around 30 min. Ca²⁺ influx initiated by GSK could be detected from 1 nm and increased dose‐dependently. To analyze the contribution of TRPV4 channels to the Ca²⁺ influx, we used ruthenium red (RuRed), which acts as a broad inhibitor of a wide range of cation channels, including TRPV family members, and a TRPV4‐specific antagonist, HC‐067047 (HC). One nanomolar GSK induced rapid and sustained Ca²⁺ influx that was completely blocked by 10 μm RuRed and 1 μm HC (Fig. 1D). These data demonstrate that MDCK II monolayers express functional TRPV4 ion channels.

The TRPV4 agonist GSK induces reversible TJ opening in MDCK II monolayers

Before investigating the effects of GSK on TJ function, we examined the cytotoxicity of GSK to MDCK II monolayers. The IC₅₀ of GSK was determined to be 5.8 ± 1.2 nm, and GSK did not exhibit notable cytotoxicity at 1 nm after 5 h of exposure to the monolayers (Fig. S1).

We next examined whether or not GSK can increase TJ permeability. An FD4 solution was applied to the apical sides of MDCK II monolayers grown in transwells, which were then treated with different concentrations of GSK. Every 30 min, the basolateral solution was collected to quantify the amount of permeated FD4 (Fig. 2A,B). Latrunculin A (LatA), an irreversible TJ opener, increased FD4 permeability continuously [13]. As with LatA, 3 nm GSK continuously increased FD4 permeability from 30 min to 5 h, indicating an irreversible TJ opening. Note that in this figure, we use a logarithmic scale for the Y‐axis (relative cumulative fluorescence intensity), because 3 nm GSK increased permeability much more than LatA. In contrast, 1 nm GSK induced reversible opening of TJs, which opened in as little as 15 min (Figs 2C,E and 3B,F) and then closed in around 2 h. A lower dose of GSK (0.3 nm) did not induce FD4 permeability. We therefore used 1 nm GSK for further experiments.

Figure 2

GSK reversibly increases the paracellular permeability of MDCK II monolayers. (A) GSK modulated the epithelial permeability of MDCK II monolayer. Three nanomolar GSK induced an irreversible FD4 permeability increase, 1 nm GSK reversibly increased permeability, and 0.3 nm GSK did not. (○) DMSO control, (□) 0.1 μm LatA, and (▲) 0.3 nm, (●) 1 nm, and (◆) 3 nm GSK. Typical data of three independent experiments is shown. (B) The efficiencies of TJ permeability enhancement induced by 0.1 μm LatA and 1 nm GSK were evaluated by the cumulative FD4 amounts transported until 2 h relative to the vehicle control. Each value represents the mean ± standard deviation of three independent experiments. (C) The specific TRPV4 antagonist 1 μm HC abolished the FD4 permeability increase induced by 1 nm GSK. (○) DMSO control, (□) 0.1 μm LatA, (◊) 1 μm HC, (●) 1 nm GSK, (◆) 1 μm HC + 1 nm GSK, and (■) 1 μm HC + 0.1 μm LatA. Typical data of three independent experiments is shown. (D) The inhibitory effect of HC on the TJ permeability enhancement induced by GSK was evaluated by comparing the cumulative FD4 amounts transported by 1 μm HC, 1 nm GSK, and 1 μm HC + 1 nm GSK until 2 h relative to the vehicle control. Each value represents the mean ± standard deviation of four independent experiments. *** indicate that the value is significantly (P < 0.001) different from others. (E) The SK channel blocker apamin (1 μm), but not the BK channel blocker paxilline (30 μm), inhibit the FD4 permeability increase induced by 1 nm GSK. (○) DMSO control, (□) 0.1 μm LatA, (◊) 30 μm paxilline, (Δ) 1 μm apamin, (●) 1 nm GSK, (◆) 30 μm paxilline + 1 nm GSK, and (▲) 1 μm apamin + 1 nm GSK. Each value represents the mean ± standard deviation in triplicate experiments. (F) Deconvoluted images of MDCK II monolayers after 30 min incubation with (left) 1% FD4 and DMSO, (right) 1% FD4 and 1 nm GSK. Top: XY sections of merged images. Bottom: XZ sections of merged images. Top is apical, bottom is basolateral. Scale bar: 10 μm.

Figure 3

Ca²⁺ influx, but not cofilin and TRPA1, is involved in GSK‐induced modulation of MDCK II epithelia. (A, B) Ca²⁺ depletion blocks the GSK‐induced TER decrease (A) and FD4 permeability increase (B). MDCK II monolayers were treated with 1 nm GSK, with or without 1.3 mm Ca²⁺. Changes in TER and amounts of permeated FD4 were analyzed at each time point. Each value represents the mean ± standard deviation of three independent experiments. (○) DMSO control, (●) 1 nm GSK, (□) DMSO control under Ca²⁺‐depleted conditions, (■) 1 nm GSK under Ca²⁺‐depleted conditions. Dunnett's test has been performed comparing with GSK treatment at each time. *P < 0.05; **P < 0.01; ***P < 0.001. (C, D) GSK did not induce a notable change in the phosphorylation of cofilin protein. MDCK II monolayers were treated with 1 nm GSK or 300 μm capsaicin for the indicated durations. Representative data from three independent western blot analyses of phosphorylated cofilin, cofilin, and actin are shown. The bar shows a densitometric analysis of phosphorylated cofilin. The band intensity at each time point is relative to that of time 0 and is shown as the mean ± standard deviation of three independent experiments. Dunnett's test has been performed comparing each treatment with time 0 control. *P < 0.05; **P < 0.01; ***P < 0.001. (E) GSK did not induce notable changes in the expression of occludin and claudin‐1, ‐2, ‐4 proteins. MDCK II monolayers were treated with 1 nm GSK for the indicated durations. Representative data from three independent western blot analyses are shown. The bar shows the densitometric analysis of each protein expression. The band intensity at each time point is relative to that of time 0 and is shown as the mean ± standard deviation of three independent experiments. Dunnett's test has been performed comparing each treatment with time 0 control, and found to be not significant. (F, H) TRPV4 antagonist (1 μm HC) pretreatment for 30 min inhibited the GSK‐induced FD4 permeability increase (F) and TER decrease (H), but a TRPA1 antagonist (1 μm A96) did not. (○) DMSO control, (◊) 1 μm HC, (Δ) 1 μm A96, (●) 1 nm GSK, (◆) 1 μm HC + 1 nm GSK, and (▲) 1 μm A96 + 1 nm GSK. Typical data of three independent experiments is shown (F), or each value represents the mean ± standard deviation of three independent experiments (H). In (H), Dunnett's test has been performed comparing with vehicle control at each time. ***P < 0.001. (G) The inhibitory effect of HC and no influence of A96 on the TJ permeability enhancement induced by GSK was evaluated by comparing the cumulative FD4 amounts transported by 1 μm HC, 1 μm A96, 1 nm GSK, 1 μm HC + 1 nm GSK and 1 μm A96 + 1 nm GSK until 2 h relative to the vehicle control. Each value represents the mean ± standard deviation of three independent experiments. Dunnett's test has been performed comparing with the control. **P < 0.01.

We previously reported that capsaicin and other natural α,β‐unsaturated carbonyl compounds induce reversible opening of TJs via Ca²⁺ influx [11-13]. We therefore investigated TRPV4 involvement in GSK‐induced reversible TJ opening and found that HC inhibited GSK‐induced FD4 permeability completely (Fig. 2C,D), suggesting that the TJ permeability increase induced by GSK is mediated by TRPV4. HC did not affect LatA‐induced permeability increase at all under the same conditions, because LatA opens TJs irreversibly by depolymerizing actin filaments.

It has been reported that Ca²⁺ entry via TRPV4 upon 4α‐phorbol 12,13‐didecanoate (4α‐PDD) stimulation activates large conductance Ca²⁺‐activated K⁺ channels (BK channels) in a mouse mammary cell line, resulting in an increase in transcellular/paracellular permeability for at least until 24 h [7]. Because BK channels are expressed in kidney epithelial cells [22], there is a possibility that TJ opening is correlated with the activation of BK channels. We therefore investigated the involvement of BK channels in GSK‐induced TJ opening. The specific BK channel blocker, paxilline, did not block both GSK‐induced FD4 permeability (Fig. 2E) and TER decrease (Fig. S2), suggesting that BK channels are not involved in GSK‐induced TJ opening.

On the other hand, there is a report that TRPV4‐mediated Ca²⁺ influx activates small conductance Ca²⁺‐activated K⁺ channels (SK channels) in renal collecting duct cells [23]. When we employed SK channel inhibitor apamin to the GSK‐induced FD4 permeability experiment, it blunted the permeability increase (Fig. 2E). Taken together, not BK but SK channels are suggested to be involved in GSK‐induced TJ opening.

Finally, transport of FD4 across the MDCK II monolayer was demonstrated using fluorescent microscopy as described in Materials and methods. Images were taken after incubation of the monolayers with FD4 with or without GSK, as shown in Fig. 2F. In control monolayers, the uptake of fluorescent was not generally observed. Addition of GSK resulted in increased paracellular uptake of fluorescence.

Mechanistic comparison with capsaicin TJ opening via TRPA1

We next analyzed the mechanisms underlying the GSK‐induced reversible opening of TJs. We have previously shown that Ca²⁺ entry triggers capsaicin‐induced TJ opening [11, 12]; therefore, we first confirmed whether this is also the case with GSK‐induced TJ opening. To obtain direct evidence for a Ca²⁺ influx requirement for GSK‐induced reversible TJ opening, we removed Ca²⁺ from the assay buffer (Materials and methods). Because Ca²⁺ depletion itself gradually opens TJs [11, 24-27], we measured TER and FD4 permeability for 30 min. GSK induced a considerable decrease in TER within 5 min in the presence of Ca²⁺ (Fig. 3A), whereas the DMSO control did not. In contrast, GSK did not lead to any decrease in TER in the absence of Ca²⁺, which was the same as DMSO control until 10 min. After 15 min, GSK decreased TER gradually, but so did the DMSO control, suggesting that this TER decrease was caused by Ca²⁺ depletion. Consistent with this result, the increase in FD4 permeability induced by GSK was also abolished in the absence of Ca²⁺ (Fig. 3B), although Ca²⁺ depletion itself was relatively less affective. Taken together, these results suggest that Ca²⁺ entry is necessary to induce reversible TJ opening by GSK, similarly to what is observed with capsaicin.

Next, we focused on cofilin dephosphoryaltion (activation), which is a critical signal for capsaicin‐induced TJ opening [11, 12]. We have shown that almost all cofilin is dephosphorylated as quickly as 10 min and then rephosphorylated at around 60 min upon capsaicin treatment. Because GSK induced a rapid increase in TJ permeability within 10 min (Fig. 3A,B), we analyzed changes in cofilin phosphorylation within the first 20 min of treatment (Fig. 3C, left). Capsaicin induced a significant decrease in cofilin phosphorylation as quickly as 5 min, and this persisted over 20 min. In contrast, GSK did not produce such a decrease even though it should have already induced TJ opening by this time. These observations were confirmed by densitometric analysis (Fig. 3C, right). The same results were obtained in longer incubations up to 120 min. Consistent with our previous report, cofilin phosphorylation recovered after 60 min of capsaicin treatment, although GSK did not induce any notable change in its phosphorylation level (Fig. 3D). These results suggest that although both GSK and capsaicin require a Ca²⁺ influx to trigger TJ opening, the resulting downstream signaling appears to be different. In addition, we also analyzed the expression of occludin, claudin‐1, ‐2, ‐4, as they are the main regulator of TJ, however, there were no significant change (Fig. 3E). Furthermore, we also examined the distributions of occludin and claudin‐1 by immunofluorescence, but could not observe marked difference between control and GSK treatment (Fig. S3).

It is possible that TRPV4 activation by GSK might induce TJ opening via TRPA1 sensitization/activation [28-31]. We therefore examined the involvement of TRPA1 in GSK‐induced TJ opening pathways to further elucidate the difference in mechanisms between GSK and capsaicin. The Ca²⁺ influx induced by GSK was unaffected by the TRPA1 antagonist A‐967079 (A96; Fig. 1D). We confirmed that the TRPA1 agonist AITC also induced Ca²⁺ influx, which was blocked by A96, but not by HC under the same conditions (Fig. S4). The ineffectiveness of the TRPA1 antagonist at inhibiting the GSK‐induced epithelial permeability increase was confirmed by both the FD4 permeability (Fig. 3F,G) and TER experiments (Fig. 3H). These results suggest that, although sharing necessity of a Ca²⁺ influx, GSK opens TJs via a different channel and downstream mechanism than capsaicin.

Discussion

In the present study, we evaluated TRPV4 expression and involvement in the reversible opening of TJs in MDCK II epithelial monolayers. The TRPV4 agonist GSK increased intracellular Ca²⁺ in MDCK II monolayer, and the responses were significantly inhibited by the specific TRPV4 agonist HC, but not by the TRPA1 antagonist A96. These results indicate that TRPV4 is functionally expressed in MDCK II monolayers.

GSK induced a reversible paracellular permeability increase and TER decrease at nontoxic concentrations, suggesting that the TJ opening was reversible. The GSK‐induced TJ opening was potently inhibited by the TRPV4 antagonist (HC), but not by paxilline. GSK‐induced TER decreases eventually returned to basal levels, suggesting that cell viability was normal. Because it has been reported that TRPV4 activation is associated with a rapid downregulation of TRPV4 at the plasma membrane, leading to rapid desensitization [32, 33], GSK stimulation on TRPV4 might be temporary, resulting in reversible effects.

TRPV4 activation increases epithelial permeability in the mouse mammary cell line HC11, and this is prevented by the BK channel blocker paxilline in the early phase [7]. We have not demonstrated the expression of BK channels in MDCK II cells, but they are known to be expressed in epithelial cells of the kidney and in C11‐MDCK cells [22, 34]. However, we observed no inhibition of GSK‐induced TJ opening by paxilline. This observation, together with a report that the permeability increase is not reversed until at least 24 h in 4α‐PDD‐stimulated HC11 cells [7], suggests that the processes and the resulting effects including reversibility of the epithelial permeability increase, and the underlying mechanisms are different from what we revealed here. Instead, here we found that SK channels are involved in GSK‐induced TJ opening. We are now analyzing the interplay between TRPV4 and SK channels and the downstream mechanism. On the other hand, several reports have shown that TRPV4 activation strengthens epithelial TJs [4-10], and we are currently investigating the basis of these differences, including TJ protein modulation.

We have been analyzing the reagents that induce the reversible opening of TJs, for example, capsaicin and natural compounds containing an α,β‐unsaturated moiety, and have discovered that a specific cofilin dephosphorylation and F‐actin modulation are important for TJ opening to be reversible [13, 17]. Here, we observed the highly reproducible and complete reversal in TRPV4‐mediated TJ opening. We primarily expected that the TJ opening mechanism of GSK would be the same as capsaicin's, that is, that GSK action would also involve Ca²⁺ influx and cofilin‐actin modulation, because these two compounds induced a similar reversible pattern of TJ opening triggered by Ca²⁺ influx in MDCK II monolayers. However, GSK did not activate cofilin even though Ca²⁺ influx is involved early in the opening of TJ by both capsaicin and GSK. Because capsaicin and GSK activate different channels, i.e., TRPA1 for capsaicin and TRPV4 for GSK, these results suggest that the Ca²⁺ increase induced by the stimulation of different TRP channels recruits different molecules and/or transmits different signals for barrier regulation.

The MDCK cell line was isolated from the kidney cortex [35], and high expression of TRPV4 has been shown in the kidney [19]. There are several reports of TRPV4 expression in the nephron segment in kidney. One showed that water‐impermeant tubular segments, such as the thick ascending limb and the distal convoluted tubule, express abundant TRPV4, suggesting an osmoregulatory role for TRPV4 in these nephron segments where an osmotic gradient is generated [36]. Subsequent studies have reported that TRPV4 is expressed in several tubule segments, and collecting ducts, where it appears to function as a flow sensor inducing a rapid influx of Ca²⁺ [37, 38]. Consistently, it has also been reported that TRPV4 is an essential component of the ciliary flow sensor and is associated with transient Ca²⁺ influx in MDCK cells [39]. Since TRPV4 responds to osmotic cell swelling and flow‐induced mechanical stress, leading to an influx of Ca²⁺ [23, 40], and since we observed that Ca²⁺ influx is responsible for the reversible opening of TJs (Fig. 3A), TRPV4‐mediated TJ regulation may function to adjust to changes in extracellular osmolarity and/or flow rate under physiological conditions.

In conclusion, we have shown that TRPV4 activation induces Ca²⁺ influx, resulting in the reversible opening of TJs. An analysis of the mechanism underlying this opening, and its physiological role, is underway.

Acknowledgments

This work was supported by grants from Japan Society for Bioscience, Biotechnology, and Agrochemistry to TU (JSBBA Innovative Research Program Award), and the Center of Innovation Program funded by the Ministry of Education Culture and Sports (MEXT), Japan and Japan Science and Technology (JST) to YN. The authors thank Enago (www.enago.jp) for the English language review.

Author contributions

MM and YY performed the experiments, analysis, and interpretation of data. TU and YN contributed to the conception and design of the study, analysis and interpretation of data, and writing of the manuscript.