Kaempferol, CFTR inhibitor-172 (CFTRinh-172), glibenclamide, 5-nitro-2-(3-phenylpropylamino) benzoic acid (NPPB), calcium-activated chloride channel inhibitor (CaCCinh-A01), 4,4′-Diisothiocyanatostilbene-2,2′-disulfonate (DIDS), forskolin, 1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA-AM), amphotericin B, H89, AG490, and tyrphostin A23 were dissolved in dimethyl sulfoxide (DMSO). The final concentration of DMSO was controlled at 0.01% in all tests. Amiloride, carbachol (CCh), 8-(4-chlorophenylthio) adenosine-3′,5′-cyclic monophosphate (8cpt-cAMP), and sodium orthovanadate (vanadate) were dissolved in distilled water. Bumetanide was dissolved in ethanol. The vehicles had no effect on I _sc . All these chemicals were purchased from Sigma-Aldrich (MO, United States). Mouse anti-CFTR antibody, mouse anti-β-actin antibody, HRP-conjugated goat anti-mouse IgG antibody, and chemiluminescence ECL were obtained from Santa Cruz Biotechnology (CA, United States). Bicinchoninic acid assay (BCA assay) was purchased from Visual Protein (TW, ROC). Laemmli loading buffer and mercaptoethanol were purchased from Bio-rad (CA, United States).

Human colonic adenocarcinoma cell line (T84 cells) purchased from the American Type Culture Collection (VA, United States), were plated on a 100-mm cell culture dish. The T84 cells (passage number 61–86) were cultured in standard media composed of DMEM-Ham’s F-12 (DMEM-F12; 1:1) with 10% fetal bovine serum, and 1% penicillin-streptomycin at 37°C in 5% CO₂ incubator. Cell culture reagents and supplies were purchased from Gibco (NY, United States). Culture media were refreshed every 2 days. When the cells reached 80% confluence, they were detached using trypsin-EDTA (0.25%) and transferred into new suitable cell culture vessels.

Cytotoxicity effects of kaempferol on T84 cells were determined using the MTT colorimetric assay. The cells were seeded in a 48-well plate at 1 × 10⁵ cells/well and incubated for 7 days. Afterward, the cells were treated with various concentrations of kaempferol (1, 5, 10, 50, and 100 µM) in standard media for 24 and 48 h. An equivalent quantity of DMSO was used as vehicle control. 10% MTT solution in media (125 µL) was added to each well and incubated in a CO₂ incubator for 3 h. The MTT-derived formazan crystals were dissolved in 100 µL of DMSO, and the absorbance was measured in optical density (OD) unit at wavelength of 570 nm and 620 nm using a microplate reader (Epoch, Biotek, VM, United States). OD unit of kaempferol at 570 nm subtraction with OD unit at 620 nm was calculated for cytotoxicity analysis by comparing with those of the vehicle DMSO. The experiments were conducted in duplicate and repeated at least three times for each concentration of kaempferol.

T84 cells (3 × 10⁵ cells/well) were cultured into the Snapwell insert with a 0.4 µm pore polycarbonate membrane (Costar, MA, United States) for 10–14 days to form the complete monolayer detected by assessing Transepithelial electrical resistance (TEER) using a Millicell ERS-2 volt-ohm meter coupled to Ag/AgCl electrodes (World Precision Instrument, FL, United States). The monolayers with high resistance (1,500–3,000 Ω cm²) were selected for Ussing experiments. The monolayers were mounted in Ussing chambers bathed with standard Ringer solution pH 7.4 (in mM: 118 NaCl, 4.5 KCl, 2.5 CaCl₂, 0.54 MgCl₂, 25 NaHCO₃, 1.5 NaH₂PO₄). The solution was bubbled with 95% O₂ and 5% CO₂ (carbogen) and maintained at 37°C. Short-circuit currents (I _sc ) and transepithelial potential difference (PD) were measured by an EVC-4,000 voltage/current clamp (World Precision Instrument, United States) with Ag/AgCl electrodes connected to bathing solution via agar bridges. Transepithelial conductance (G) was calculated using Ohm’s law (G = I _sc /PD). The monolayers were short-circuited, except for measuring the PD value before and after adding test substances. The voltage clamp data was sent through a PowerLab 4/35 A/D converter and recorded on an Intel® Core™ i5-3,570, 3.40 GHz, 3,401 MHz Processors. The monolayers were equilibrated for at least 30 min. Using an apical side as a reference, a positive I _sc indicated a net flux of positive charges from the apical side to the basolateral side (absorptive direction) or net flux of negative charges from the basolateral side to the apical side (secretory direction). In anion substitution experiments, I _sc was measured from monolayers bathed with Cl⁻ free, HCO₃ ⁻-free or Cl⁻ and HCO₃ ⁻-free Ringer solution. In Cl⁻ free solution, gluconate salts were used to replace Cl⁻, and bubbled with carbogen, while in HCO₃ ⁻ free solution, HEPES buffer was used to replace HCO₃ ⁻. Both HCO₃ ⁻ free and Cl⁻-HCO₃ ⁻ free Ringer solutions were bubbled with 100% O₂.

Membrane permeability studies were performed using amphotericin B-permeabilized monolayer. To determine apical Cl⁻ current (I _Cl ), the apical side was filled with a high concentrated KCl Ringer solution pH 7.4 (in mM: 111.5 KCl, 25 NaHCO₃, 12 D-glucose, 1.8 Na₂HPO₄, 1.25 CaCl₂, 1 MgSO₄, 0.2 NaH₂PO₄) and the basolateral side filled with a KMeSO₄ Ringer solution pH 7.4 (in mM: 120 KMeSO₄, 20 KHCO₃, 15 mannitol, 5 NaCl, 2 calcium gluconate, 1.3 K₂HPO₄, 1 MgSO₄, 0.3 KH₂PO₄). Amphotericin B (50 µM) was added into basolateral solution for basolateral membrane permeabilization. Basolateral K⁺ current (I _Kb ) was measured when the apical side was filled with KMeSO₄ Ringer solution in the presence of amphotericin B, while the basolateral side was filled with a NaMeSO₄ Ringer solution pH 7.4 (in mM: 120 NaMeSO₄, 30 mannitol, 5 NaCl, 3 calcium gluconate, 1 MgSO₄, 20 KHCO₃, 0.3 KH₂PO₄, 1.3 K₂HPO₄). All solutions were maintained at 37°C and bubbled with carbogen.

The effect of kaempferol on the CFTR protein expression was tested by Western blot analysis. T84 cells (3 × 10⁵ cells/well) were cultured in 60-mm culture dishes and then treated with either 50 µM or 100 µM of kaempferol in standard media or DMSO alone for 24 h. Total proteins were extracted from the cells by a lysis buffer containing (in mM) 46.5 Tris (adjust pH 7.4), 150 mM NaCl, 1 EDTA, 1 NaF, 1 PMSF, 1% NP-40, and 1% protease inhibitor mixture (Sigma, United States) and measured concentrations by the Bicinchoninic acid assay (BCA assay). Protein samples (60 μg/mL) were loaded into 7.5% Sodium Dodecyl Sulfate polyacrylamide gel electrophoresis (SDS-PAGE), run at 100 V for 120 min, and subsequently transferred to a polyvinylidene difluoride (PVDF) membrane (Millipore, United States). The membrane was blocked with 5% non-fat milk for 1 h before incubation at 4°C overnight with primary antibodies, i.e., mouse monoclonal anti-CFTR (1:100 dilution), and anti-β-actin (1:1,000 dilution) and further incubated at room temperature for 1 h with horseradish peroxidase-conjugated goat anti-mouse IgG secondary antibody (1:2,000 dilution). Immunoreactive signals were visualized by the ECL substrate and imaged by the ChemiDoc Imaging System (Bio-Rad, United States). The protein expression of CFTR was analyzed by ImageJ (NIH, United States) and normalized to β-actin. All experiments were replicated at least three times to verify reproducibility.

Data were expressed as mean ± standard deviation; n referred to the number of independent experiments from different passages in each group. Statistical differences between mean values of the control and one treated group were analyzed by Student’s t-test. To compare cytotoxic effect of different treatments on different times, two-way analysis of variance (two-way ANOVA) was performed followed by Dunnett’s post-hoc test for comparison with control. Analysis of variance (ANOVA) followed by Tukey’s post-hoc test were used to compare the various concentration effect of drugs. All statistics were analyzed using GraphPad Prism software version 8.0.2, Inc., San Diego, CA. A value of P < 0.05 was considered statistically significant.

Cytotoxic effect of kaempferol treatment at concentrations of 1, 5, 10, 50, and 100 µM for 24 or 48 h on T84 cells was determined by the MTT assay. The results showed that all treatments with kaempferol for 24 or 48 h had no toxic effects to T84 cells as compared with the corresponding DMSO control (Figure 1). Conversely, the cell viability was proportionally increased with kaempferol concentrations from 1 to 50 µM. Treatment with kaempferol 10 and 50 µM for 24 and 48 h significantly increased T84 cell viability compared with the corresponding DMSO (n = 3–5, P < 0.05, Figure 1). However, the kaempferol treatments at the same concentration did not significantly change cell viability between 24 and 48 h.

Under basal conditions, the average I _sc , PD, and G values of T84 cell monolayers mounted with standard Ringer solution in Ussing chambers were 1.18 ± 0.66 μA/cm², −1.96 ± 1.05 mV, and 0.62 ± 0.18 mS/cm² (n = 102), respectively. We first investigated the effect and route of kaempferol at the effective concentration of 50 µM from previous study on basal I _sc . We found that kaempferol 50 µM increased I _sc when added to the apical side and slightly increased I _sc after a subsequent addition to the basolateral side (n = 4, Figure 2A). Meanwhile, a basolateral addition of kaempferol led to a small increase in I _sc , and a subsequent apical addition significantly increased I _sc (n = 4, Figure 2B). This indicates the apical membrane is a major site of kaempferol action. To ensure the maximum effect on I _sc , kaempferol at each concentration was applied to both apical and basolateral solution.

Bilateral addition of kaempferol (50 µM) induced an increase in I _sc within 7–8 min to a maximal response of 19.54 ± 4.46 μA/cm² (n = 8, Figure 2C), then decreased slightly and maintained continuously up to 90 min. A single dose of kaempferol varying from 1 to 100 µM was additionally tested for dose-dependent response. As shown in Figure 2D, kaempferol 10, 50, and 100 µM significantly increased the I _sc as compared with DMSO (n = 4, P < 0.0001) with a maximum activation at 50 and 100 µM. Thus, kaempferol 50 μM, an observed minimum concentration that produced the highest I _sc activation, was selected for the remaining experiments.

As kaempferol increased I _sc response in T84 cell monolayers, we further investigated the ionic basis of the kaempferol-induced I _sc by applying various pharmacological channel blockers in apical solution before (pretreatment) and after (posttreatment) kaempferol addition. As shown in Figure 3, kaempferol-induced increase in I _sc was not inhibited by amiloride (10 µM), the Na⁺ channel blocker, whether added before or after kaempferol (Figures 3A, 5A,B). Pretreatment of CFTRinh-172 (50 µM), the selective CFTR inhibitor, failed to inhibit the kaempferol-increased I _sc (Figures 3B, 5A) whereas posttreatment of CFTRinh-172 decreased the I _sc by 53% from the maximum effect of kaempferol (−10.29 ± 2.48 μA/cm², n = 8, Figures 3B, 5B). Conversely, the kaempferol response was significantly inhibited by pretreatment with 5-Nitro-2-(3-phenylpropylamino) benzoic acid (NPPB) (100 µM) or glibenclamide (200 µM), CFTR inhibitors by 84% (3.09 ± 0.23 μA/cm², n = 4, P < 0.05, Figures 3C, 5A) or 64% (7.06 ± 1.15 μA/cm², n = 4, P < 0.05, Figures 3D, 5A), respectively, as compared with kaempferol alone. The posttreatment with NPPB or glibenclamide decreased the kaempferol response by 60% (−11.79 ± 1.82 μA/cm², n = 4, P < 0.05, Figures 3C, 5B) or 58% (−11.33 ± 0.96 μA/cm², n = 4, P < 0.05, Figures 3D, 5B), respectively.

Our study also showed that pretreatment with CaCCinh-A01 (30 µM) or DIDS (100 µM), CaCC inhibitors, failed to inhibit the kaempferol-induced increase in I _sc (n = 4, Figures 4A,B, 5A). Nevertheless, posttreatment with CaCCinh-A01 or DIDS decreased the I _sc from the maximum effect of kaempferol by 10% (−1.92 ± 1.35 μA/cm², n = 5, Figures 4A, 5B) or 16% (−3.10 ± 1.08 μA/cm², n = 5, Figures 4B, 5B), respectively. Besides, pretreatment with bumetanide (200 µM) in basolateral solution to block the Cl⁻ uptake through NKCC cotransporter reduced the kaempferol-induced I _sc by 61% (7.53 ± 0.73 μA/cm², n = 4, P < 0.05, Figures 4C, 5A), and posttreatment decreased the I _sc from maximum kaempferol response by 58% (−11.05 ± 1.34 μA/cm², n = 4, Figures 4C, 5B). Taken together, our findings indicate kaempferol stimulation of anion secretion, albeit with little active CFTR.

Anion substitution tests were additionally carried out to confirm the ionic basis of the kaempferol effect on anion secretion. As shown in Figures 6A,B, adding kaempferol to normal Ringer solution increased the I _sc by 19.54 ± 4.46 μA/cm² (n = 8). The kaempferol-induced I _sc was unaffected when HCO₃ ⁻ was substituted (18.49 ± 2.07 μA/cm², n = 4), but it was significantly decreased by 96% in Cl⁻ free (0.73 ± 0.45 μA/cm², n = 4, P < 0.05) or Cl⁻ and HCO₃ ⁻ free solution (0.78 ± 0.35 μA/cm², n = 4, P < 0.05). This indicates that the kaempferol-stimulated I _sc is Cl⁻ dependent.

The involvement of intracellular cAMP or Ca²⁺ in kaempferol-sensitive I _sc was further investigated under Cl⁻ secretion activated by forskolin or carbachol. The kaempferol-sensitive I _sc was determined before and after addition of forskolin or carbachol as shown in Figure 7. Addition of forskolin (10 μM, bilateral) induced the maximal I _sc response to 172.91 ± 12.89 μA/cm² (n = 4, Figures 7A,D). A subsequent addition of kaempferol decreased the forskolin-stimulated I _sc by 55% (−94.42 ± 15.57 μA/cm², n = 4, Figures 7A,C). Pretreatment with kaempferol increased I _sc and reduced the forskolin-activated I _sc to 25.81 ± 4.52 μA/cm² (n = 4), which was less than the I _sc induced by forskolin alone by 85% (Figures 7A,D). To determine the kaempferol effect on Ca²⁺-activated Cl⁻ secretion, it was shown in Figure 7B that carbachol (100 μM, basolateral) increased I _sc by 5.87 ± 0.55 μA/cm² (n = 4). Subsequent addition of kaempferol further increased the I _sc response to 58.34 ± 8.05 μA/cm² (n = 4), which was three times greater than the kaempferol alone (Figures 7B,C). However, pretreatment with kaempferol greatly increased the maximal I _sc response induced by carbachol to 129.92 ± 2.74 μA/cm² (n = 4, Figures 7B,D).

The basolateral membrane permeabilized monolayers were used to examine the effect of kaempferol on apical Cl⁻ permeability. Bilateral addition of kaempferol maximally increased I _Cl by −59.83 ± 9.20 μA/cm² (n = 6, Figures 8A,D), as shown by the negative deflection of the current. Like the I _sc , either pretreatment or posttreatment with amiloride (10 μM, apical) did not inhibit the stimulatory effect of kaempferol on the I _Cl . The kaempferol-stimulated I _Cl after amiloride was −51.56 ± 4.62 μA/cm², n = 4, Figures 8A,D). In contrast, apical pretreatment of CFTRinh-172 (50 µM) or CaCCinh-A01 (30 µM) inhibited the kaempferol-induced increase in I _Cl by 98% (−1.13 ± 0.95 μA/cm², n = 4, P < 0.05, Figures 8B–D) and 58% (−25.29 ± 11.74 μA/cm², n = 4, P < 0.05, Figures 8C,D), respectively. Posttreatment with CFTRinh-172 completely inhibited the kaempferol response and basal I _Cl which accounted for 117% (69.89 ± 16.36 μA/cm², n = 4, P < 0.05, Figures 8B,E) and CaCCinh-A01 markedly inhibited 80% from the maximum kaempferol response (47.84 ± 4.85 μA/cm², n = 4, P < 0.05, Figures 8C,E).

In the apical membrane-permeabilized monolayer, kaempferol at 50 µM increased the maximum I _Kb by 5.89 ± 1.05 μA/cm² within 7–8 min, as shown by the positive deflection of the current, and then declined slightly (n = 12, Figure 9A). Basolateral pretreatment with a non-specific K⁺ channel blocker, BaCl₂ (5 mM), and a Ca²⁺ activated K⁺ channel blocker, clotrimazole (50 µM) significantly inhibited the kaempferol-activated I _Kb by 55% (2.66 ± 0.67 μA/cm², n = 5, P < 0.05, Figures 9B,D) and by 72% (1.67 ± 1.14 μA/cm², n = 5, P < 0.05, Figures 9C,D), respectively. Posttreatment with BaCl₂ and clotrimazole inhibited the maximum kaempferol response by 62% (−3.63 ± 0.34 μA/cm², n = 4, P < 0.05, Figures 9B,E) and by 117% (−6.88 ± 1.16 μA/cm², n = 4, P < 0.05, Figures 9C,E), respectively. These findings demonstrate that kaempferol activates Cl⁻ secretion mainly through CFTR and partially through CaCC and basolateral K⁺ channels.

To verify whether cAMP plays a role in the kaempferol response, the kaempferol-activated I _Cl was determined before and after adding forskolin or 8cpt-cAMP which stimulated Cl⁻ secretion through CFTR. As shown in Figure 10, addition of forskolin (10 μM, apical and basolateral) or 8cpt-cAMP (100 μM, basolateral) activated the maximal I _Cl response by −223.85 ± 1.08 μA/cm² (n = 4, Figures 10A,C) or −222.30 ± 35.90 μA/cm² (n = 4, Figures 10B,D), respectively. The subsequent addition of kaempferol did not affect the forskolin or 8cpt-cAMP-stimulated I _Cl response pattern (Figures 10A,B). Conversely, pretreatment with kaempferol increased the I _Cl and decreased the forskolin or 8cpt-cAMP-activated I _Cl to −21.19 ± 9.97 μA/cm² (n = 4, Figures 10A,C) or −3.01 ± 0.74 μA/cm² (n = 4, Figures 10B–D), respectively, which was less than forskolin alone by 91% or 8cpt-cAMP alone by 99%.

To determine whether the Cl⁻ secretion induced by kaempferol was dependent of intracellular Ca²⁺, the kaempferol response was examined in the presence of BAPTA-AM, a membrane-permeable Ca²⁺ chelator. Pretreatment with BAPTA-AM (50 μM, apical and basolateral) for 30 min did not change the baseline. A subsequent addition of kaempferol increased the I _sc response to 18.05 ± 2.25 μA/cm² (n = 4, Figures 11A,B). The kaempferol-activated I _sc in the presence of BAPTA-AM was not significantly different compared with kaempferol alone (21.55 ± 5.82 μA/cm²; Figure 11B). This evidence suggests that the effect of kaempferol on activating Cl⁻ secretion may not involve the Ca²⁺ mobilization but possibly relate to the cAMP signaling pathway.

Cl⁻ secretion in colonic epithelium occurring mainly through CFTR has been reported to be regulated by PKA or protein tyrosine kinase (PTK) phosphorylation and protein tyrosine phosphatase (PTT) dephosphorylation. To further investigate the mechanisms of kaempferol on regulating Cl⁻ secretion via protein kinases and phosphatases, the kaempferol-activated apical I _Cl was determined in the presence of kinase or phosphatase inhibitors for 30 min in apical and basolateral solutions as shown in Figure 11C. Pretreatment with H89 (10 µM), a PKA inhibitor, significantly decreased the kaempferol-activated I _Cl by 88% (−7.15 ± 4.25 μA/cm², n = 4, P < 0.05, Figures 11C,D). AG490 (10 µM) or tyrphostin A23 (100 µM), PTK inhibitors, inhibited the kaempferol-activated I _Cl by 16% (−50.04 ± 6.03 μA/cm², n = 4) or by 16% (−50.00 ± 1.32 μA/cm², n = 4), respectively. Vanadate (100 µM), a PTT inhibitor, increased the kaempferol-activated I _Cl by 3% (−61.66 ± 7.16 μA/cm², n = 4). Nonetheless, the kaempferol response in the presence or absence of AG490, tyrphostin A23, or vanadate was not statistically significant (Figure 11D). Our results reveal the mechanism of kaempferol action via a protein kinase A activity.

To examine the effect of kaempferol on regulating Cl⁻ transport protein, the expression of CFTR was analyzed by Western blot analysis. T84 cells were cultured in standard media for 7 days before treatment with kaempferol (50 or 100 µM) or DMSO for 24 h. The protein bands demonstrated CFTR and β-actin with a molecular weight of 165 and 43 kDa, respectively (Figure 12A). The protein band density measured as the expression ratio of CFTR to β-actin revealed a slight but not statistically significant increase in the CFTR expression in the DMSO group compared with the control group that received no treatment (Figure 12B). Kaempferol 50 and 100 µM treatment significantly increased CFTR expression by 23% (1.20 ± 0.13, n = 5, P < 0.05, Figure 12B) and 25% (1.22 ± 0.12, n = 5, P < 0.05, Figure 12B) compared with DMSO, respectively. The kaempferol effect upregulates CFTR expression according to the increased concentration.