ε‐Viniferin is a dehydrodimer of resveratrol, a polyphenol synthesized in many plants, including grapevine. The present study investigated the effects of ε‐viniferin and resveratrol on epithelial secretory and barrier functions in isolated rat small and large intestinal mucosa. Mucosa–submucosa tissue preparations of various segments of the rat large and small intestines were mounted on Ussing chambers, and short‐circuit current (Isc) and tissue conductance (Gt) were continuously measured. The mucosal addition of ε‐viniferin (>10−5 mol/L) and resveratrol (>10−4 mol/L) to the cecal mucosa, which was the most sensitive region, induced an increase in Isc and a rapid phase decrease (P‐1) followed by rapid (P‐2) and broad (P‐3) peak increases in Gt in concentration‐dependent manners. Mucosal ε‐viniferin (10−4 mol/L), but not resveratrol (10−4 mol/L), increased the permeability of FITC‐conjugated dextran (4 kDa). The mucosal ε‐viniferin–evoked changes in Isc (Cl− secretion), but not in Gt, were attenuated by a selective cyclooxygenase (COX)‐1 inhibitor and a selective EP4 prostaglandin receptor. The mucosal ε‐viniferin–evoked increase in Isc was partially attenuated, and P‐2, but not P‐1 or P‐3, change in Gt was abolished by a transient receptor potential cation channel, subfamily A, member 1 (TRPA1) inhibitor. Moreover, the mucosal ε‐viniferin concentration‐dependently attenuated the mucosal propionate (1 mmol/L)‐evoked increases in Isc and Gt. Immunohistochemical studies revealed COX‐1–immunoreactive epithelial cells in the cecal crypt. The present study showed that mucosal ε‐viniferin modulated transepithelial ion transport and permeability, possibly by activating sensory epithelial cells expressing COX‐1 and TRPA1. Moreover, mucosal ε‐viniferin decreased mucosal sensitivity to other luminal molecules such as short‐chain fatty acids. In conclusion, these results suggest that ε‐viniferin modifies intestinal mucosal transport and barrier functions.
ε‐Viniferin is a dehydrodimer of resveratrol (3,5,4′‐trihydroxy‐trans‐stilbene), a polyphenol expressed in many plants, including grapevine. Resveratrol has a variety of beneficial effects, including antioxidant (Farghali et al. 2013), anti‐inflammatory (Csiszar 2011), and anticancer effects (Jang et al. 1997; Aggarwal et al. 2004). Additionally, it has been reported that resveratrol oligomers (oligostilbenoids), like ε‐viniferin, have a variety of beneficial effects (Xue et al. 2014). However, it is thought that resveratrol oligomers are poorly absorbed in the intestine, unlike resveratrol (Willenberg et al. 2015).
The intestinal epithelia are sensory tissues for luminal molecules (Kaji et al. 2011b; Breer et al. 2012; Furness et al. 2013). Luminal nutrient sensing in the gastrointestinal (GI) tract induces a variety of physiological responses mediated by diverse gut hormones. For example, nutrients stimulate gastrin‐containing enteroendocrine G cells to release gastrin and induce gastric acid secretion. In the large intestine, the most predominant (~100 mmol/L) anions are short‐chain fatty acids (SCFAs), which are enterobacterial fermented products, that induce secretory (Yajima 1988; Karaki and Kuwahara 2011) and contractile responses (Yajima 1985; Mitsui et al. 2005) by activating the luminal side, but not the serosal side. Moreover, we previously reported that a bitter taste (Kaji et al. 2009) and an olfactory compound (Kaji et al. 2011a) evoke secretory responses by activating the luminal side of the GI tract. Therefore, it is possible that some polyphenols, including resveratrol, ε‐viniferin, and other oligomers activate the luminal chemosensory system in the GI tract.
It was reported that apical administration of resveratrol stimulates cAMP‐dependent chloride secretion (Cl−) in a colonic epithelium‐like monolayer‐cultured cell line, T84 cells, and in the mouse jejunum (Blumenstein et al. 2005). Conversely, it was reported that resveratrol oligomers, including ε‐viniferin, inhibited cAMP‐dependent (forskolin‐activated) Cl− secretion (Zhang et al. 2014). Unfortunately, these earlier reports did not investigate the effects of resveratrol or its oligomers on epithelial barrier functions.
Therefore, the aims of the present study were to investigate the mucosal effects of resveratrol and ε‐viniferin on epithelial secretory functions and ion permeability in isolated rat small and large intestines.
Materials and Methods
Animals and tissue preparation
Male Sprague Dawley (SD) rats (mean ± standard error of the mean [SEM], 335.2 ± 10.3 g, n = 62) were purchased from Japan SLC, Inc. (Hamamatsu, Japan), and were allowed food and water ad libitum before all experiments. The animals were handled and euthanized in accordance with the Guidelines for the Care and Use of Laboratory Animals of the University of Shizuoka, and the study was approved by the University of Shizuoka Animal Usage Ethics Committee. SD rats were anesthetized with isoflurane and decapitated with a guillotine. The terminal ileum, cecum, proximal colon, middle colon, distal colon, and rectum were removed and placed in ice‐cold Krebs–Ringer solution saturated with O2 and CO2 (see Solutions). Tissues were cut along the mesenteric border, and pinned flat to silicon rubber‐lined Petri dishes with the Krebs–Ringer solution. The smooth muscle layer of the tissue was removed using microscissors under a stereomicroscope, and the mucosa–submucosa preparations were mounted between the halves of Ussing flux chambers (CHM2; World Precision Instruments, Inc., Sarasota, FL) to measure transepithelial ion transport and permeability.
Ussing chamber experiments
The cross‐sectional area of the mucosa–submucosa preparations in the Ussing flux chamber was 0.64 cm2, and both sides were bathed with 10 mL of Krebs–Ringer solution, which was recirculated from a glass circulation reservoir (#5210; World Precision Instruments, Inc.) maintained at 37°C and gassed with 95% O2/5% CO2. The transepithelial potential difference (PD; the serosal electrode served as the reference) was measured using paired Ag–AgCl electrodes (EKV; World Precision Instruments, Inc.) through Krebs–agar bridges, and was clamped at 0 mV by applying a short‐circuit current (Isc) with another pair of Ag–AgCl electrodes (EKC; World Precision Instruments, Inc.) connected to a voltage‐clamp apparatus (CEZ‐9100; Nihon‐Koden, Tokyo, Japan). The lumen negative electrogenic current represents a positive Isc per unit area of tissue (μA/cm2). To record tissue conductance (Gt [mS/cm2]), voltage command pulses (10 mV, 3 sec duration) were applied at 1 min intervals, and Gt was calculated using Ohm's law as the current needed to alter the clamped voltage. The current output was continuously recorded on a data acquisition and analog‐to‐digital conversion system (PowerLab 4/26; ADInstruments, Cattle Hill, Australia). Before the experiments, the tissues were stabilized for 1 h and tissue viability was checked by electrical field stimulation (EFS; 25 V, 5 Hz, 0.5 msec duration for 2 min) using a pair of aluminum foil ribbon electrodes.
The permeability of fluorescein isothiocyanate (FITC)‐conjugated dextran (average molecular weight 4000; FD4) in the mucosa–submucosal tissue preparation was measured under short‐circuit conditions in the Ussing chamber. In this experiment, FD4 (10−4 mol/L) was added to the mucosal bathing solution 30 min before the addition of EtOH (10 μL; vehicle control), ε‐viniferin (10−4 mol/L), and resveratrol (10−4 mol/L). Samples (100 μL) were taken from the serosal bathing solution, and fresh Krebs–Ringer (100 μL) was added at −15, 0 15, 30, 45, and 60 min (6 samples). The concentrations of FD4 were determined by measuring the fluorescence intensity of samples and standard solutions using a microplate reader (Varioskan Flash; Thermo Fisher Scientific, Inc., Yokohama, Japan) at emission and excitation wavelengths of 520 nm and 490 nm, respectively. Based on a volume of the serosal bathing solution (Vm) of 10 mL and sample volume (Vs) of 100 μL, the total FD4 mass that crossed the tissue at each time (Mn) was: where C1, C2, … Cn are the sample concentrations at sampling times −15 … 60 min. The FD4 flux (Jm→s) per unit area (0.64 cm2) was calculated as:where J1m→s, J2m→s, …Jn‐1m→s are the FD4 fluxes in each period (−15–0 min, 0–15 min, … 45–60 min). Because the decrease in the mucosal concentration of FD4 during experiments is considered to be negligible, the permeability (Pnm→s) of FD4 in each period was calculated by Fick's law as follows:where Cm and Cs (Cm ≫ Cs) are the concentrations of FD4 in the mucosal and serosal bathing solutions, respectively. This method was built in reference to a chapter in the textbook, Molecular Biopharmaceutics (Brodin et al. 2009).
The Krebs–Ringer solution contained (in mmol/L) 117 NaCl, 4.7 KCl, 1.2 MgCl2, 1.2 NaH2PO4, 25 NaHCO3, and 2.5 CaCl2, and was saturated with 95% O2/5% CO2. The Cl−‐free modified Krebs–Ringer solution was prepared by replacing Cl− with an equal concentration of gluconate−, and the concentration of Ca‐gluconate2 was increased to 8 mmol/L to compensate for the Ca2+‐buffering effects of gluconate. The mucosal bathing solutions were supplemented with 5 mmol/L glucose.
ε‐Viniferin, resveratrol, and sodium propionate were purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). Carbachol (CCh), FD4, bumetanide, 5‐nitro‐2‐(3‐phenylpropylamino) benzoic acid (NPPB), atropine, hexamethonium, SC‐560, and NS‐398 were purchased from Sigma (St. Louis, MO). TTX and HC030031 were purchased from Tocris (Ellisville, MO). Piroxicam was purchased from Biomol Research Laboratories (Plymouth Meeting, PA). 4,4′‐Diisothiocyanatostilbene‐2,2′‐disulfonic acid (DIDS) was purchased from ANA SPEC (Fremont, CA). AH‐6809 was purchased from Cayman Chemical (Ann Arbor, MI). ONO‐8713, ONO‐AE3‐208, and ONO‐AE3‐240 were kind gifts from ONO Pharmaceutical Co., Ltd. (Osaka, Japan).
ε‐Viniferin and resveratrol were dissolved in ethanol (EtOH) just before use. TTX was dissolved in citrate buffer (pH 4.8) and stored at −20°C until use. CCh, FD4, propionate, atropine, and HEX were dissolved in distilled water, and stored at 4°C until use. All of the other chemicals were dissolved in DMSO and stored at −20°C until use.
The isolated cecal tissues were immediately frozen with Tissue‐Tek optimal cutting temperature compound (Sakura Finetek Japan Co., Ltd., Tokyo, Japan) in a −75°C acetone bath (Histo‐Tek Pino; Sakura Finetek Japan Co., Ltd.). The tissues were cut into 4‐μm‐thick sections on a cryostat (CM1100; Leica Microsystems, Weltzlar, Germany). The sections were placed on glass slides, fixed in 100% methanol at −20°C for 10 min, and air‐dried under a cold blower for 30 min. The sections were incubated with 10% normal donkey serum and 1% Triton X‐100 in phosphate‐buffered saline (PBS) at room temperature for 1 h to block nonspecific antibody binding. The sections were incubated with goat anticyclooxygenase (COX)‐1 antibody (×500 dilution; sc‐1754; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) in PBS containing 0.3% Triton X‐100 at 4°C overnight. After washing in PBS (3 × 10 min), the sections were incubated with donkey anti‐goat IgG conjugated with Alexa594 (×500 dilution; A11058; Life Technologies, Carlsbad, CA) at room temperature for 1 h. The sections were washed in PBS (3 × 10 min), and coverslips were mounted on the glass slides with mounting medium (DakoCytomation, Glostrup, Denmark). Immunoreactivity for COX‐1 was visualized using a fluorescence microscope (Axio Observer Z1; Carl Zeiss, Oberkochen, Germany), and the images were captured using a cooled charge‐coupled device digital camera and digital imaging software (AxioCam and AxioVision; Carl Zeiss).
Data analysis and statistics
All physiologic data in the present study are expressed as the mean ± SEM. Multiple comparisons were made using Dunnett's test and Tukey's test for each group relative to the control group and among all combinations of groups, respectively. Paired t tests were used to determine the statistical significance between two groups because some tissue preparations from the same animals were used in both experimental groups. In all analyses, Ps < 0.05 were considered statistically significant.
Effects of mucosal and serosal ε‐viniferin and resveratrol on Isc and Gt in the mucosa–submucosa preparations of rat cecum
In the rat cecal mucosa–submucosal preparations, mucosal, but not serosal, addition of ε‐viniferin (bath concentration, 10−4 mol/L) evoked a monophasic increase in Isc (Fig. 1A; ∆Isc: 73.93 ± 16.75 μA/cm2, ∆time: 5.68 ± 0.52 min) and triphasic changes in Gt (Fig. 1B; P‐1: ∆Gt −0.61 ± 10.07 mS/cm2, ∆time 1.8 ± 0.2 min; P‐2: ∆Gt 3.91 ± 1.13 mS/cm2, ∆time: 5.6 ± 0.2 min; P‐3: ∆Gt 4.48 ± 0.76 mS/cm2, ∆time 20.6 ± 2.3 min; n = 5). In contrast, mucosal addition of 10−4 mol/L resveratrol evoked very small changes in Isc and Gt (Fig. 1E and C), while mucosal, but not serosal, addition of 3 × 10−4 mol/L resveratrol evoked biphasic changes in Isc (Fig. 1C; P‐1: ∆Isc −9.84 ± 2.76 μA/cm2, ∆time 3.55 ± 0.42 min; P‐2: ∆Isc 54.98 ± 5.76 μA/cm2, ∆time 24.62 ± 4.77 min; n = 4) and Gt, likely via ε‐viniferin–evoked P‐1 (∆Gt −0.25 ± 0.09 mS/cm2, ∆time 2.3 ± 0.5 min) and P‐3 (∆Gt 3.83 ± 0.71 mS/cm2, ∆time 16.8 ± 1.7 min) (n = 4; Fig. 1D). At a lower concentration (10−5 mol/L), ε‐viniferin evoked a small, but long‐lasting decrease in Gt, as shown in Figure 1E. Mucosal or serosal addition of the vehicle, EtOH (10 μL), did not affect Isc or Gt.
Various concentrations of ε‐viniferin (3 × 10−6 to 3 × 10−4 mol/L) and resveratrol (10−5 to 3 × 10−4 mol/L) were added to the mucosal bathing solution, and the mean changes in Isc (Fig. 1E) and Gt (Fig. 1F) were plotted and fitted to the Hill equation. The half maximal (50%) effective concentration (EC50) of the mucosal ε‐viniferin–evoked ∆Isc was 4.81 × 10−5 mol/L, the maximum effect (Emax) was 87.19 μA/cm2, and the Hill coefficient value (nH) was 2.37 (coefficient of determination [R2] = 1.000). At 10−5 mol/L or lower concentrations, mucosal ε‐viniferin only evoked a decrease in Gt, which was long‐lasting. The EC50 of the mucosal ε‐viniferin–evoked P‐1 ∆Gt was 2.53 × 10−6 mol/L, Emax was −0.69 mS/cm2, and nH was 1.22 (R2 = 0.887). Although it is unclear whether the P‐2 ∆Gt evoked by 3 × 10−4 mol/L ε‐viniferin was at the maximum effective concentration, the EC50 of ∆Gt was calculated to be 9.44 × 10−5 mol/L, when we assumed that the Emax was the mean value for ∆Gt at 3 × 10−4 mol/L ε‐viniferin (6.92 mS/cm2) and that the nH was same as the that for P‐3 described below (R2 = 1.000). Therefore, the EC50 for mucosal ε‐viniferin–evoked P‐3 ∆Gt was calculated to be 2.35 × 10−5 mol/L, the Emax of mucosal ε‐viniferin–evoked P‐3 ∆Gt was 4.49 mS/cm2, and nH was 4.53 (R2 = 0.999).
The resveratrol‐evoked P‐1 and P‐2 ∆Isc did not seem to reach the maximum concentrations, but the EC50 of P‐2 ∆Isc was 2.52 × 10−4 mol/L (R2 = 0.975) when we assumed that Emax and nH were the same as those for ε‐viniferin–evoked ∆Isc. Likewise, the resveratrol‐evoked increase in Gt did seem to reach the maximum concentration, but the EC50 was 1.82 × 10−4 mol/L (R2 = 0.912) when we assumed the Emax and nH were the same as those for ε‐viniferin–evoked P‐3 ∆Gt.
Although ε‐viniferin may evoke more potent responses at concentrations of >3 × 10−4 mol/L, ε‐viniferin at these concentrations evoked unlimited continuous increases in Gt. This indicates that ε‐viniferin at concentrations >3 × 10−4 mol/L may have cytotoxic effects because the epithelial barrier function was thought to have failed. Thus, ε‐viniferin was used at 10−4 mol/L in the following experiments.
Segmental differences in the effects of mucosal ε‐viniferin in the rat small and large intestines
The effects of mucosal ε‐viniferin (10−4 mol/L) on Isc and Gt were measured and compared among the rat terminal ileum, cecum, proximal colon, middle colon, and distal colon. Mucosal addition of ε‐viniferin (10−4 mol/L) had weak effects on Isc in the intestinal tissues, except for the cecum (Fig. 2A). Regarding Gt, the ε‐viniferin (10−4 mol/L)–evoked increase was most potent in the cecum and the magnitude of the increase gradually decreased from the proximal to distal segments of the large intestine (Fig. 2B). Thus, the other experiments in the present study were performed using the rat cecum.
Effects of mucosal ε‐viniferin and resveratrol on FD4 permeability in the rat cecum
To investigate the effects of mucosal ε‐viniferin and resveratrol on the barrier function against a nonionic macromolecule, the effects of mucosal ε‐viniferin and resveratrol on the transepithelial permeability of FD4 from the mucosal to serosal side (FD4 Pm→s) were measured in terms of Gt. Basal Isc was 18.07 ± 1.62 μA/cm2, basal Gt was 11.15 ± 0.41 mS/cm2, and basal FD4 Pm→s was 1.26 ± 0.20 × 10−7 cm/s (n = 20).
Mucosal addition of ε‐viniferin evoked an increase in Isc, and the mean Isc reached its maximum at 7 min (ΔIsc: 83.56 ± 17.12 μA/cm2, n = 8). The Gt first decreased at 3 min (P‐1 ΔGt: −0.45 ± 0.05 mS/cm2, n = 8), the P‐2 peak occurred at 8 min (P‐2 ΔGt: 4.00 ± 0.82 mS/cm2, n = 8), and the P‐3 peak occurred at 26 min (P‐3 ΔGt: 3.70 ± 0.84 mS/cm2, n = 8). FD4 Pm→s increased, and reached a maximum (3.98 ± 0.95 × 10−7 cm/s, n = 8) at 30–45 min after mucosal addition of ε‐viniferin. This value was significantly greater (P < 0.05 by Dunnett's test) than the FD4 Pm→s after the addition of the vehicle control (EtOH; 0.138 ± 0.025 × 10−6 cm/s, n = 7) at 30–45 min.
After the mucosal addition of resveratrol (10−4 mol/L), Isc decreased initially (P‐1 ΔIsc: −4.24 ± 2.06 μA/cm2, n = 6, at 5 min), and then slowly increased (P‐2 ΔIsc: 7.40 ± 4.31 μA/cm2, n = 6, at 37 min). Gt slowly increased (ΔGt: 1.06 ± 0.22 mS/cm2, n = 6, at 17 min). However, FD4 permeability did not significantly change.
Identification of ion species involved in the ε‐viniferin–evoked increase in Isc and Gt in the rat cecum
To identify the ion species involved in the ε‐viniferin–induced increase in Isc, the bathing solutions in the serosal, mucosal, and both chambers were replaced with Cl−‐free/gluconate solution. Before mounting the tissues, the fluid resistances of the normal solution and Cl−‐free solution in each chamber were measured (normal: 69.5 ± 1.0 Ω; Cl−‐free: 114.7 ± 0.9 Ω; n = 3). All preparations were bathed with normal solution at first. After the bathing solutions of the serosal, mucosal, or both sides were replaced with Cl−‐free solution, and the compensations of fluid resistance were reset after changing the solutions. Basal Isc significantly increased in the serosal Cl−‐free solution, but decreased in the mucosal Cl−‐free solution (Table 1). In contrast, the basal Gts were not significantly different between the Cl−‐free and sham control tests for each side (Table 1). The negative basal Isc in the mucosal Cl−‐free solution and the positive basal Isc in the serosal Cl−‐free solution are considered to be due to the liquid junction potentials between agar bridges and bathing solution.
The ε‐viniferin (10−4 mol/L)–evoked increases in Isc were almost completely abolished for Cl−‐free solution on the serosal and both sides, but not the mucosal side (P < 0.01 by Dunnett's test, n = 4; Fig. 4A). These results indicate that the mucosal ε‐viniferin–evoked increase in Isc is due to electrogenic Cl− secretion. Moreover, with serosal Cl−‐free solution, mucosal ε‐viniferin evoked a decrease in Isc (−5.55 ± 1.66 μA/cm2; Fig. 4A). Serosal and both side Cl−‐free solutions led to unlimited increases in Gt, which made it difficult or impossible to measure the mucosal ε‐viniferin–evoked changes in Gt.
To confirm that mucosal ε‐viniferin evoked Cl− secretion, a Na+‐K+‐2Cl− (NKCC) transporter inhibitor, bumetanide (10−4 mol/L), was added to the serosal bathing solution. Alternatively, a Ca2+‐dependent channel blocker, DIDS (10−4 mol/L), or a cAMP‐dependent Cl− channel (especially cystic fibrosis transmembrane regulator [CFTR]) blocker, NPPB (10−5 to 10−4 mol/L), were added to the mucosal bathing solution before the mucosal addition of ε‐viniferin. Bumetanide (10−4 mol/L) and NPPB (≥5 × 10−5 mol/L) significantly attenuated the ε‐viniferin–evoked increase in Isc, but not DIDS (Fig. 5A). Although bumetanide attenuated the ε‐viniferin–evoked increase in Isc, it did not attenuate the Gt responses (Fig. 5B). In contrast, NPPB (5 × 10−5 mol/L) enhanced the ε‐viniferin–evoked P‐1 Gt reduction and nearly abolished the P‐3 Gt change (Fig. 5B). The ε‐viniferin–evoked changes in Gt in the presence of 10−4 mol/L NPPB were not measured because Gt gradually increased after the addition of 10−4 mol/L NPPB.
Effects of mucosal resveratrol on Isc, Gt, and ε‐viniferin–evoked responses in the rat cecum
Although mucosal addition of 10–4 mol/L resveratrol evoked only a weak decrease in P‐1 ΔIsc (−5.66 ± 1.61 μA/cm2) and an increase in ΔGt (0.63 ± 0.30 mS/cm2) (Fig. 6; n = 4), resveratrol at 3 × 10−4 mol/L elicited Isc and Gt responses, which were similar to those induced by 10−4 mol/L ε‐viniferin (Fig. 1E and F). We hypothesized that the lower concentration of resveratrol might have an antagonistic function to ε‐viniferin. To investigate this hypothesis, ε‐viniferin (10−4 mol/L) was added to the mucosal bathing solution 30 min after mucosal addition of resveratrol (at 0 to 3 × 10−4 mol/L), and the changes in Isc and Gt were measured. In this experiment, resveratrol significantly attenuated the ε‐viniferin–evoked changes in Isc and Gt in concentration‐dependent manners (Fig. 6). Mucosal resveratrol at 10−4 mol/L significantly attenuated the mucosal ε‐viniferin (10−4 mol/L)–evoked increase in Isc to 26.07 ± 4.12% of the control value (P < 0.05, n = 4) and P‐3 ∆Gt to 29.31 ± 8.20% of the control value (P < 0.01, n = 5). When the mucosal bathing solution contained 3 × 10−4 mol/L resveratrol, the administration of ε‐viniferin (10−4 mol/L) decreased Isc (P‐1 ΔIsc: −8.60 ± 0.71 μA/cm2, P < 0.001, n = 4) and abolished the P‐2 response. However, there were no statistically significant differences because the P‐2 response in the control conditions was weak and varied considerably between tests. Moreover, in the presence of 3 × 10−4 mol/L resveratrol, ε‐viniferin (10−4 mol/L) caused a continuous increase in Gt, as shown in Figure 6C, so P‐3 ∆Gt could not be determined.
Effects of a transient receptor potential cation channel, subfamily A, member 1 (TRPA1) inhibitor on mucosal ε‐viniferin–evoked responses in the rat cecum
In our previous study, we reported that mucosal administration of a pungent principle, allyl isothiocyanate (AITC), evoked anion secretion by activating TRPA1 channels (Kaji et al. 2012). Therefore, we investigated whether the ε‐viniferin–evoked changes in Isc and Gt were mediated via TRPA1 channels, like those evoked by AITC. In this setting, a TRPA1 inhibitor, HC030031 (10−4 mol/L), significantly reduced the ε‐viniferin–evoked increase in Isc from 60.95 ± 10.70 μA/cm2 to 38.04 ± 8.02 μA/cm2 (P < 0.05, paired t test, n = 4; Fig. 7B), and the peak time after the addition of ε‐viniferin was significantly delayed from 6.71 ± 0.93 min to 14.68 ± 2.05 min (P < 0.05, paired t test, n = 4; Fig. 7A). For Gt, the ε‐viniferin–evoked P‐1 and P‐3 responses were not affected, but the P‐2 response was almost completely abolished by HC030031, as shown in Figure 7A and B (P < 0.01, n = 4).
Lack of effects of neural blockade and cholinergic antagonists, and the involvement of COX activity in the mucosal ε‐viniferin–evoked Isc and Gt responses in the rat cecum
The neural reflex pathways in the submucosal plexus in the gut wall play important roles in luminal stimuli‐evoked secretory responses in the intestinal mucosa (Cooke 1998; Karaki and Kuwahara 2004). Thus, we investigated whether neural and cholinergic pathways are involved in the ε‐viniferin–induced Isc and Gt responses by adding TTX (a neural blocker; 10−6 mol/L), HEX (a nicotinic ACh receptor antagonist; 10−4 mol/L), or atropine (a muscarinic ACh receptor antagonist) to the serosal bathing solution 30 min before the addition of ε‐viniferin (10−4 mol/L). However, TTX, HEX, and atropine did not affect the ε‐viniferin–evoked changes in Isc and Gt (n = 4, Fig. 8A, B).
In our previous studies, some mucosal stimulants including a bitter tastant (6‐n‐propyl‐2‐thiouracil) (Kaji et al. 2009), an odorant (thymol) (Kaji et al. 2011a), and a pungent principle (AITC) (Kaji et al. 2012) increased Isc, and these effects were mediated by COX because piroxicam attenuated these effects. Therefore, we hypothesized that the ε‐viniferin–evoked Isc and Gt responses were dependent on COX activity. To test this hypothesis, a nonselective COX inhibitor (piroxicam; 10−5 mol/L), a selective COX‐1 inhibitor (SC‐560; 10−5 mol/L), a selective COX‐2 inhibitor (NS‐398; 10−5 mol/L), or a combination of SC‐560 and NS‐398 were added to the serosal bathing solution 30 min before the mucosal addition of ε‐viniferin (10−4 mol/L). Piroxicam, SC‐560, and the combination of SC‐560 and NS‐398, but not NS‐398 alone, significantly attenuated the ε‐viniferin–evoked increase in Isc from the control value of 59.40 ± 10.53 μA/cm2 (n = 6) to 13.33 ± 2.79 μA/cm2 with piroxicam (n = 4; P < 0.01, Dunnett's test), to 23.10 ± 6.36 μA/cm2 with SC‐560 (n = 6; P < 005, Dunnett's test), and 22.92 ± 7.70 μA/cm2 with the combination of SC‐560 and NS‐398 (Fig. 8B). These results indicate that the ε‐viniferin–evoked increase in Isc is mediated by COX‐1 production of prostaglandin (P)G. In contrast, piroxicam, SC‐560, and NS‐398 did not significantly affect the ε‐viniferin–evoked changes in Gt (Fig. 8B).
Effects of selective prostaglandin E2 receptor antagonists on mucosal ε‐viniferin–evoked responses in the rat cecum
To determine which PG receptor subtype mediates the ε‐viniferin–evoked increase in Isc, a selective EP1 receptor antagonist (ONO‐8713; 10−5 mol/L), an EP1 and EP2 receptor antagonist (AH‐6809; 10−5 mol/L), a selective EP3 receptor antagonist (EP3 > EP4; ONO‐AE3‐240; 10−6 to 10−5 mol/L), or a selective EP4 receptor antagonist (EP4 > EP3; ONO‐AE3‐208; 10−7 to 10−5 mol/L) were added to the serosal bathing solution 30 min before the mucosal addition of ε‐viniferin (10−4 mol/L). In this experiment, neither ONO‐8713 nor AH‐6809 affected the ε‐viniferin–evoked increase in Isc, but ONO‐AE3‐208 at concentrations ≥10−6 mol/L significantly attenuated the ε‐viniferin–evoked increase in Isc (from 84.47 ± 18.14 μA/cm2 to 17.81 ± 10.21 μA/cm2 at 10−6 mol/L; P < 0.05, n = 4; Fig. 9A). These results indicate that the ε‐viniferin–evoked increase in Isc is mediated by EP4 receptors.
Inhibitory effects of mucosal ε‐viniferin on the mucosal propionate‐evoked Isc response
Short‐chain fatty acids, including acetate (two carbons), propionate (three carbons), and butyrate (four carbons), are the predominant anions in the large intestine and exist at concentrations of ≥100 mmol/L. They are produced by bacterial fermentation of indigestible dietary fibers and oligosaccharides. Mucosal SCFAs (with a potency order of propionate ≥ butyrate ≫ acetate) are known to stimulate the large intestinal mucosa to secrete transepithelial anions (Yajima 1988; Karaki and Kuwahara 2011). In our previous studies, luminal thymol and AITC attenuated propionate‐evoked anion secretion in the rat colon (Kaji et al. 2011a, 2012). Therefore, we hypothesized that ε‐viniferin would attenuate the mucosal propionate‐evoked secretory responses. To confirm this hypothesis, mucosal propionate (10−3 mol/L)‐evoked Isc responses were measured 30 min after the addition of ε‐viniferin at a variety of concentrations, as described earlier.
In the absence of ε‐viniferin, propionate (10−3 mol/L) transiently increased Isc (ΔIsc: 222.33 ± 17.45 μA/cm2; Fig. 11A) and Gt (ΔGt: 6.39 ± 0.67 mS/cm2; Fig. 11B) (n = 6). However, the mucosal propionate (10−3 mol/L)‐evoked increases in Isc and Gt were attenuated by mucosal ε‐viniferin in a concentration‐dependent manner at ε‐viniferin concentrations ranging from 3 × 10−6 to 3 × 10−4 mol/L (Fig. 11). Pretreatment with 3 × 10−4 mol/L ε‐viniferin nearly abolished the propionate response. The concentration–response curves of ε‐viniferin for inhibiting the propionate‐evoked Isc and Gt responses were drawn by fitting the data to the inhibitory Hill equation, as described in Figure 11. In this analysis, the half‐maximal inhibitory concentration (IC50) of ε‐viniferin for ΔIsc was 4.12 × 10−5 mol/L and the nH was 2.11 (R2 = 0.998). For Gt, the IC50 was 2.68 × 10−5 mol/L and nH was 1.68 (R2 = 0.985), where 119.4%, which was the mean value of the propionate‐evoked Gt in 3 × 10−6 mol/L ε‐viniferin, was used as the value for the maximal effect.
Immunohistochemistry for COX‐1 in the rat cecum
It was hypothesized that luminal ε‐viniferin might stimulate some epithelial sensory cells expressing COX‐1, because luminal ε‐viniferin evoked Isc and Gt responses, but serosal ε‐viniferin did not. An earlier study revealed that COX‐1–expressing epithelial cells are scattered throughout the crypts in the rat large intestine (Shao et al. 1999). However, there are no previous reports describing COX‐1 expression in the rat cecum. Thus, we performed immunohistochemical analysis of COX‐1 expression in the rat cecum. Figure 10 shows the images of COX‐1 immunoreactivity in the rat cecal crypt. COX‐1 immunoreactivity was detected on scattered cells in the crypt epithelial cells (Fig. 10).
The results of the present study show that ε‐viniferin, a dehydrodimer of resveratrol, stimulated the rat intestinal mucosa, especially the cecal mucosa, from its luminal side, whereas a 10 times higher concentration of resveratrol was needed to evoke the same effect. Mucosal ε‐viniferin had three major effects: (1) induce transepithelial electrogenic Cl− secretion by activating EP4 receptors via PGs produced by COX‐1; (2) elicit a rapid decrease and a sustained increase in transepithelial ion permeability; and (3) increase transepithelial permeability of a nonionic macromolecule, FD4. In addition, mucosal ε‐viniferin inhibited mucosal propionate‐evoked Cl− secretion. These results indicate that ε‐viniferin stimulates the intestinal epithelium from the luminal side and effects mucosal barrier functions.
Mucosal ε‐viniferin– and resveratrol‐evoked changes in Isc and Gt
Mucosal ε‐viniferin at concentrations ≥10−5 mol/L elicited a concentration‐dependent monophasic positive change in Isc and triphasic changes in Gt, which comprised an abrupt decrease (P‐1) followed by fast (P‐2) and sustained (P‐3) increases in the rat cecal mucosa, whereas serosal administration of 10−4 mol/L ε‐viniferin did not (Fig. 1A and B). On the other hand, 10−4 mol/L resveratrol evoked little, but 3 × 10−4 mol/L resveratrol evoked biphasic changes in Isc, which included a weak negative ∆Isc (P‐1) and a sustained positive ∆Isc (P‐2), and a single sustained increase in Gt. In contrast, serosal administration of 3 × 10−4 mol/L resveratrol did not affect Isc and a small decrease (P‐1) and increase (P‐2) in Gt (Fig. 1C and D) were observed. These results suggest that ε‐viniferin stimulates apical targets on the epithelium with a potency 10 times greater than that of resveratrol. Moreover, the first negative ∆Isc evoked by ε‐viniferin was apparently masked by the positive phase ∆Isc. In fact, the ε‐viniferin–evoked negative ∆Isc was observed when the positive ∆Isc was inhibited, as in the serosal Cl−‐free condition described below.
At lower concentrations (≤10−5 mol/L), ε‐viniferin evoked the long‐lasting decreases in P‐1 Gt, as shown in Figure 1E. This suggested that, at higher concentration (>3 × 10−5 mol/L), the ε‐viniferin–evoked changes in Gt consisted of an abrupt and long‐lasting decrease in ion permeability and subsequent increases in ion permeability. At lower concentration, ε‐viniferin may enhance epithelial barrier function, and at higher concentrations ε‐viniferin may further enhance ionic and nonionic permeability (as mentioned below) to induce secretory and inflammatory functions as part of a host defense mechanism.
Segmental differences in the effects of mucosal ε‐viniferin
The most potent effects of ε‐viniferin in terms of the Isc and Gt responses occurred in the cecal mucosa (Fig. 2). The physiologic meaning of these findings might be related to the primary function of the cecum as a fermentation tank. The gut microbiota in adult humans, especially in the colon, consists of more than 100 trillion microbes of at least 400 species (Bourlioux et al. 2003). The cecum in rodents, including rats, has not atrophied like that in humans, and is the major reservoir of microbiota in the gut. Thus, a variety of compounds produced by fermentation, including organic acids, alcohols, aldehydes, and phenols (Garner et al. 2007), are thought to accumulate in the cecal lumen owing to the activity of cecal microbiota. We hypothesize that the cecal epithelial membrane monitors the luminal fermentation conditions by sensing the luminal content. This is because maintaining adequate luminal conditions for microbiota within the major fermentation site is likely to be extremely important for gut homeostasis. Accordingly, the cecal mucosa might be the most sensitive region for sensing luminal chemicals, including ε‐viniferin.
Mucosal ε‐viniferin–evoked increase in FD4 Pm→s
Mucosal ε‐viniferin (10−4 mol/L), but not resveratrol (10−4 mol/L), elicited a transient (15–45 min after addition) increase in FD4 Pm→s (Fig. 3). In our previous study, mucosal addition of thymol also increased Isc, Gt, and FD4 Pm→s in the rat colon (Kaji et al. 2011a). In the present study, ε‐viniferin (10−4 mol/L) also increased FD4 Pm→s in the rat cecum. These findings suggest that luminal ε‐viniferin directly and/or indirectly affects epithelial tight junctions, and increases paracellular permeability of nonionic macromolecules. However, the mechanism underlying the ε‐viniferin–evoked increase in FD4 Pm→s needs to be examined in future studies.
Identification of ion species involved in the mucosal ε‐viniferin–induced changes in Isc and Gt
The ε‐viniferin–induced increase in Isc was due to transepithelial Cl− secretion, because these responses were attenuated by Cl−‐free solution on the serosal and both sides, but not on the mucosal side (Fig. 4A and B). It has been proposed that intestinal fluid secretion is induced by transepithelial electrogenic anion, especially Cl−, secretion elicited by the collaborative functions of basolateral and apical transporters and channels (Karaki and Kuwahara 2004). In particular, the basolateral NKCC1 cotransporter promotes Cl− uptake from the basolateral membrane, whereas anion channels, especially CFTR channels in the apical membrane, release anions, especially Cl−, into the luminal side (Karaki and Kuwahara 2004). The present results indicate that the mucosal ε‐viniferin–evoked positive ∆Isc was attenuated by serosal bumetanide, a NKCC inhibitor, and by mucosal NPPB, a cAMP‐dependent Cl− channel (including CFTR) blocker (Fig. 5). Accordingly, the ε‐viniferin–evoked positive ∆Isc may be driven by Cl− secretion. In the serosal Cl−‐free condition, mucosal ε‐viniferin evoked a negative ∆Isc, as did resveratrol (Fig. 5). This suggests that the initial ε‐viniferin–evoked negative ∆Isc was masked in normal conditions. When both sides contained Cl−‐free solutions, ε‐viniferin did not evoke significant negative ∆Isc (Fig. 5). This suggests that the initial ε‐viniferin–evoked negative ∆Isc, which was masked in normal conditions, is due to the presence of mucosal Cl− and might be based on electrogenic Cl− absorption. However, little is known about electrogenic Cl− absorption in the mammalian intestine. Therefore, further studies are necessary to clarify the mechanism involved in the initial ε‐viniferin– or resveratrol‐evoked negative ∆Isc.
Although the ε‐viniferin–evoked P‐1, P‐2, and P‐3 ∆Gt were not significantly affected by serosal bumetanide (10−4 mol/L) and apical DIDS (10−4 mol/L), the apical administration of NPPB at concentrations ≥5 × 10−5 mol/L significantly enhanced P‐1 and significantly attenuated P‐3 ∆Gt (Fig. 5B). The enhancement of the ε‐viniferin–evoked P‐1 ∆Gt by NPPB might be shown by the NPPB‐mediated abolishment of the ε‐viniferin–evoked increase in Gt. The ε‐viniferin–evoked change in Gt was likely due to the change in the condition of tight junctions. However, the mechanism by which mucosal NPPB inhibited the ε‐viniferin–evoked P‐2 and P‐3 ∆Gt was unclear, so further studies are need to clarify the mechanism underlying the ε‐viniferin–evoked changes in Gt.
Antagonistic effects of resveratrol on ε‐viniferin–evoked changes in Isc and Gt
Although mucosal resveratrol at 10−4 mol/L hardly affected Isc or Gt, it inhibited the mucosal ε‐viniferin–evoked increases in Isc and the increase in Gt (Fig. 6). This suggests that 10−4 mol/L resveratrol has antagonistic effects against the receptors for ε‐viniferin on the apical membrane of the epithelium without causing a tissue response. However, mucosal resveratrol did not inhibit the initial ε‐viniferin–evoked decrease in Gt (Fig. 6C and D). These results suggest that the ε‐viniferin–evoked increase and decrease in Gt are mediated via different mechanisms.
Moreover, we hypothesize that the apical membrane of cecal epithelial cells express some receptors, on which ε‐viniferin acts as an agonist, and resveratrol acts as an antagonist at ≤10−4 mol/L or as an agonist at >10−4 mol/L. However, the identity of these receptors and the molecular mechanisms involved in the binding of ε‐viniferin and resveratrol to these receptors are still unknown. Future studies should address these questions.
The mucosal ε‐viniferin–evoked increase in Isc and P‐2 increase in Gt was mediated via TRPA1
We previously reported that a luminal stimulant, AITC, a TRPA1 agonist, induced anion secretion via EP4 receptors (Kaji et al. 2012). Therefore, the ε‐viniferin–evoked Cl− secretion following the release of PGs may occur via the same signaling pathway to the AITC‐evoked response. The peak ∆Isc of the ε‐viniferin–evoked increase in Isc was significantly, but not completely, inhibited by a TRPA1 blocker, HC030031, and the peak time after the addition of ε‐viniferin was significantly delayed in the presence of HC030031 (Fig. 7A and B). Moreover, the ε‐viniferin–evoked P‐2 ∆Gt was completely abolished by HC030031 (Fig. 7C and D). Therefore, the ε‐viniferin–evoked increase in Isc was partially due to TRPA1 and the increase in P‐2 ∆Gt was completely due to TRPA1. There are currently no reports showing that ε‐viniferin activates TRPA1, but mucosal ε‐viniferin may activate apical TRPA1 expressed on epithelial cells (Kaji et al. 2012).
Mechanisms involved in mucosal ε‐viniferin–evoked Cl− secretion and changes in ion permeability
The present results suggest that ε‐viniferin–evoked Cl− secretion is due to the production of PG by COX‐1, and that the Gt response is independent of PG (Fig. 8C). ONO‐AE3‐208, a selective EP4 (EP4 > EP3) antagonist, significantly attenuated ε‐viniferin–evoked Cl− secretion (Fig. 9A). Therefore, it appears that luminal ε‐viniferin stimulates COX‐1 activity and PG production in sensory epithelial cells in the cecum. An earlier study revealed that COX‐1–expressing epithelial cells are scattered throughout the crypts in the rat colon (Shao et al. 1999), and we confirmed the existence of these cells in the rat cecal epithelium by immunohistochemistry (Fig. 10). Accordingly, it appears that PGs activate EP4 receptors on the secretory epithelial cells, ultimately inducing Cl− secretion. It has been reported that EP2 and EP4 receptors, but not EP1 and EP3 receptors, mediate PGE2‐evoked Cl− secretion via the cAMP pathway (Mosa et al. 2008).
In addition, the ε‐viniferin–evoked changes in ion permeability did not involve neural pathways or PG‐related pathways. Therefore, it seems likely that the ε‐viniferin–evoked changes in ion permeability occurred through direct effects of ε‐viniferin on the epithelial cells.
Inhibitory effects of mucosal ε‐viniferin on mucosal propionate‐evoked Cl− secretion
Bacterial fermentation in the lumen of the large intestine produces numerous metabolites. The predominant molecules are SCFAs, particularly acetate, propionate, and butyrate. Propionate and butyrate, but not acetate, were reported to induce anion secretion in the rat (Yajima 1988) and guinea pig (Karaki and Kuwahara 2011) colon. Therefore, some molecules, which modulate the effects of SCFAs in the colon, critically affect the physiologic and/or pathophysiologic conditions of the large intestine. The present study showed that mucosal ε‐viniferin attenuated the luminal propionate (1 mmol/L)‐evoked increase in Isc and Gt in a concentration‐dependent manner (Fig. 11). This indicates that ε‐viniferin does not inhibit the secretory functions of epithelial cells, but instead suggests that ε‐viniferin might affect the mechanism for sensing propionate. Although the mechanism by which ε‐viniferin may inhibit the propionate‐evoked Isc response is unclear, the data suggest that the inhibitory effects of ε‐viniferin on the propionate‐evoked responses are mediated by positive cooperative binding because nH was >1. The propionate‐evoked response is thought to be mediated by its receptors, namely free fatty acid receptor 2 (FFA2 or GPR43) and/or FFA3 (GPR41) (Karaki et al. 2006, 2008; Tazoe et al. 2009; Karaki and Kuwahara 2011). Thus, ε‐viniferin may allosterically bind to these receptors, with a possible stoichiometry of 2:1 because the nH was nearly 2. Nevertheless, further studies are necessary to confirm this hypothesis. Moreover, at the lower concentration of 3 × 10−6 mol/L, ε‐viniferin very weakly enhanced the propionate‐evoked increase in Gt (Fig. 11C). Therefore, the effects of lower concentrations of ε‐viniferin on transepithelial ion permeability need to be investigated in future studies.
Physiologic relevance of sensing and responding to mucosal ε‐viniferin in the intestinal mucosa
In the gastrointestinal lumen, especially in the large intestine, the luminal microbiota can synthesize a variety of compounds and some of these compounds may have cytotoxic effects. Thus, the intestinal mucosa has protective roles, including enhancing the integrity of the epithelial barrier and fluid secretion when the mucosa senses potentially cytotoxic chemicals. We hypothesized that the tissue concentrations of PGs may constitute an alarm‐response system, in which an increase in the PG concentration might enhance the protective functions of the mucosa in host defense (Karaki and Kuwahara 2004). Therefore, ε‐viniferin may activate COX‐expressing cecal crypt cells, increase the tissue PG level as part of a tissue alarm‐response system, and enhance the tissue's host defense functions.
The present study shows that administration of ε‐viniferin to the mucosal side of the rat intestine modulates transepithelial ion transport, ion permeability, and the permeability of nonionic macromolecule as well as the effects of the other luminal molecules, such as SCFAs. These results also imply that ε‐viniferin has beneficial effects on intestinal functions by enhancing the mucosal host defense mechanism.
Conflict of Interest
A. Kuwahara have received research fund from FANCL Corporation. I. Ishikawa is an employee of FANCL Corporation.
This study was partially supported by JSPS Grant‐in‐Aid for Young Scientists (B) to S.‐i. Karaki (No. 00363903).
- Manuscript Received: June 3, 2015.
- Manuscript Revised: April 6, 2016.
- Manuscript Accepted: April 18, 2016.
- © 2016 The Authors. Physiological Reports published by Wiley Periodicals, Inc. on behalf of the American Physiological Society and The Physiological Society.
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.