Involvement of AMP-activated Protein Kinase (AMPK) in Regulation of Cell Membrane Potential in a Gastric Cancer Cell Line

Membrane potential (Vmem) is a key bioelectric property of non-excitable cells that plays important roles in regulating cell proliferation. However, the regulation of Vmem itself remains largely unexplored. We found that, under nutrient starvation, during which cell division is inhibited, MKN45 gastric cancer cells were in a hyperpolarized state associated with a high intracellular chloride concentration. AMP-activated protein kinase (AMPK) activity increased, and expression of cystic fibrosis transmembrane conductance regulator (CFTR) decreased, in nutrient-starved cells. Furthermore, the increase in intracellular chloride concentration level and Vmem hyperpolarization in nutrient-starved cells was suppressed by inhibition of AMPK activity. Intracellular chloride concentrations and hyperpolarization increased after over-activation of AMPK using the specific activator AICAR or suppression of CFTR activity using specific inhibitor GlyH-101. Under these conditions, proliferation of MKN45 cells was inhibited. These results reveal that AMPK controls the dynamic change in Vmem by regulating CFTR and influencing the intracellular chloride concentration, which in turn influences cell-cycle progression. These findings offer new insights into the mechanisms underlying cell-cycle arrest regulated by AMPK and CFTR.


Nutrient starvation induces reversible hyperpolarization and suppression of cell division in MKN45 cells.
To assess the effect of nutrient starvation on V mem , we cultured MKN45 cells in Earle's balanced salt solution (EBSS, lacking glucose and free amino acids) without FBS to induce nutrient starvation. V mem was measured using DiBAC 4 (3), an anionic and membrane potential-sensitive dye. As shown in Fig. 1A, starvation treatment of MKN45 cells led to an immediate and significant decrease in the fluorescence intensity of DiBAC 4 (3) in a time-dependent manner, indicating that cells were hyperpolarized during nutrient starvation. Furthermore, nutrient-starved cells rapidly depolarized to basal level within 2 hours after the EBSS was replaced with growth medium (Fig. 1B). These results indicated that cells up-regulate V mem in response to nutrient starvation.
A large body of data indicates that nutrient starvation induces cell-cycle arrest 49,50 . We evaluated the effect of nutrient starvation on cell division by Western blot to detect phosphorylation of histone H3 at serine 10 (H3S10). H3S10 levels decreased, accompanied by time-dependent hyperpolarization, with the extension of processing time, suggesting that cell division was suppressed after starvation treatment (Fig. 1C). CCK-8 assay of nutrient-starved cells revealed that cell proliferation was suppressed as expected (Fig. 1D). Flow cytometry analysis revealed that the cell cycle was arrested at S phase after starvation treatment (Fig. S1). After replacement of EBSS, phosphorylation of H3 was increased to the basal level within 2 hours, indicating that cell division resumed after starvation terminated (Fig. 1E). A flow cytometric analysis confirmed that the cell cycle progressed from S Scientific REPORTS | (2018) 8:6028 | DOI: 10.1038/s41598-018-24460-6 phase to G2/M phase after termination of starvation treatment (Fig. S2). These results revealed that non-dividing cells were hyperpolarized, and that cells in division were relatively depolarized, consistent with the results of previous studies.

Nutrient starvation induces an increase in intracellular chloride concentration by suppressing CFTR in MKN45 cells.
To determine whether the hyperpolarization of MKN45 cells involved chloride ion, we measured intracellular chloride concentration under nutrient starvation using a chloride-selective fluorescence dye, MQAE. We found that fluorescence intensity decreased after starvation, indicating that this treatment increased the intracellular chloride concentration in these cells ( Fig. 2A).
Next, we focused on the chloride ion channel CFTR, which directly controls intracellular chloride concentration. Expression of CFTR was down-regulated after starvation treatment (Fig. 2B), but returned to the basal level after starvation was terminated (Fig. 2C), indicating that expression of CFTR was suppressed during starvation. Immunofluorescence analysis revealed that the distribution of CFTR on the cell membranes decreased after starvation treatment, but this was not accompanied by an obvious decrease in the cytoplasmic level of CFTR (Fig. 2D), that which might or might not relate to the reported fragmentation of CFTR in cells into differently sized N and C terminal fragments 51 . These results suggested that nutrient-starved cells were hyperpolarized by an increase in the intracellular chloride concentration due to suppression of CFTR expression.

Inhibition of CFTR suppresses cell division by inducing hyperpolarization in MKN45 cells.
To explore the function of CFTR in modulation of V mem and cell division, we used the CFTR-specific inhibitor GlyH-101 to inhibit CFTR activity. The fluorescence intensity of MQAE decreased in GlyH-101-treated cells, indicating that inhibition of CFTR induced an increase in intracellular chloride concentration (Fig. 3A). In addition, we monitored the effects of CFTR inhibition on V mem . GlyH-101 treatment led to an immediate and significant Membrane potential of MKN45 cells after starvation treatment. Membrane potential was determined using a fluorescent bioelectricity reporter, DiBAC 4 (3) (green), as described in "Methods". BF = bright field. Scale bars, 50 µm. Relative DiBAC 4 (3) fluorescence intensity changes were quantified (bottom). All data are expressed as means ± SD. (B) MKN45 cells were pretreated with EBSS for 6 hours, and the membrane potential of the cells following termination of starvation treatment was detected using DiBAC 4 (3) (green). BF = bright field. Scale bars, 50 µm. Relative DiBAC 4 (3) fluorescence intensity changes were quantified (bottom). All data are expressed as means ± SD. (C) Western blot analysis of phosphorylation of H3 after starvation treatment. Quantitative data of optical band densitometry are shown. All data are expressed as means ± SD. **P < 0.05. (D) Proliferation rates of MKN45 cells after starvation treatment were assessed by CCK-8 assay. All data are expressed as means ± SD. (E) Western blot analysis of expression and phosphorylation of H3 after termination of starvation. Means ± SD. **P < 0.05. decrease in DiBAC 4 (3) fluorescence intensity, suggesting that cells were hyperpolarized after CFTR inhibition (Fig. 3B).
To study the influence of hyperpolarization on cell division, we measured phosphorylation of H3 and BrdU incorporation and performed a CCK-8 assay on CFTR-inhibited cells. Western blot analysis revealed that phosphorylation of H3 was down-regulated in GlyH-101-treated cells (Fig. 3C). BrdU incorporation was observed in the control group, but no BrdU signal was detectable in GlyH-101-treated cells (Fig. 3D). The CCK-8 assay confirmed that GlyH-101 treatment inhibited proliferation (Fig. 3E). Together, these results demonstrate that inhibition of CFTR activity suppresses cell division.
Suppression of the increase in AMPK activity prevents hyperpolarization and suppression of cell division during nutrient starvation. AMPK, an energy sensor, is over-activated in response to metabolic stress 37 . To explore the regulation of V mem during nutrient starvation, we monitored the phosphorylation of AMPK during and after starvation. AMPK activity significantly increased in cells hyperpolarized in response to starvation (Fig. 4A), and then decreased to the basal level after starvation was terminated (Fig. 4B). These results suggested that AMPK activity is involved in the up-regulation of V mem during nutrient starvation.
To verify the involvement of AMPK, we used compound C to inhibit the phosphorylation of AMPK in response to nutrient starvation. Western blot showed that phosphorylation of AMPK increased, whereas the H3S10 decreased, in nutrient-starved cells (Fig. 4C, lane 2 and lane 6). However, in comparison with that in cells that were only starved, phosphorylation of AMPK decreased in cells subjected to combined starvation and compound C treatment (Fig. 4C, lanes 3-5 and 7-9), suggesting that the increase in AMPK activity was suppressed to a certain extent by compound C during starvation. Meanwhile, the H3S10 level was increased nearly to the normal level in cells co-treated with compound C and starvation (Fig. 4C, lanes 3-5 and 7-9), indicating that suppression of the increase of AMPK activity prevented the suppression of cell division after starvation.
Next, we monitored the effects of AMPK activity on V mem . The intensity of DiBAC 4 (3) signal in co-treated cells increased significantly relative to that in cells that were only starved, indicating that co-treated cells were depolarized relative to starvation-only cells (Fig. 4D). However, compound C alone did not maintain the V mem of co-treated cells at the normal level ( Fig. 4D), likely because compound C could not completely suppress the increase in AMPK activity during starvation (Fig. 4C). These results indicated that the hyperpolarization induced by starvation is mediated by an increase in AMPK activity, and also revealed the regulatory function of AMPK activity and V mem in cell division during nutrient starvation.

Increased AMPK activity induces hyperpolarization by suppressing CFTR, and suppresses cell division in MKN45 cells.
To investigate the function of AMPK in regulation of V mem during nutrient starvation, we used AICAR, an AMPK-specific activator, to artificially increase AMPK activity. Phosphorylation of AMPK was elevated after treatment with 1 mM AICAR, and AMPK activity increased in a time-dependent manner (Fig. 5A). To determine whether the suppression of CFTR expression was regulated by AMPK during starvation, we analyzed the expression of CFTR in AICAR-treated cells. The expression level of CFTR in MKN45 cells was down-regulated after AICAR treatment for 18 hours (Fig. 5A). Immunofluorescence analysis revealed that the distribution of CFTR on cell membranes was reduced in AICAR-treated cells (Fig. 5B). In addition, elevated AMPK activation caused an increase in intracellular chloride concentration (Fig. 5C). The fluorescent signal of DiBAC 4 (3) significantly decreased after AICAR treatment (Fig. 5D), indicating that the cells became hyperpolarized after AMPK activity increased. These findings were similar to those obtained in nutrient-starved cells, and indicated that the down-regulation of CFTR was mediated by an increase in AMPK activity during nutrient starvation.
To study the effect of AMPK-induced hyperpolarization on cell division, we performed BrdU incorporation and CCK-8 assays on AICAR-treated cells. As shown in Fig. 6A, in contrast to the control group, no BrdU signal was observed in AICAR-treated cells, indicating that division was suppressed by the drug treatment. Phosphorylation of H3 was also diminished in AICAR-treated cells (Fig. 6B). The CCK-8 assay confirmed that cell proliferation was suppressed after AICAR treatment (Fig. 6C). Furthermore, flow cytometry analysis revealed that the cell cycle arrested at S phase after treatment with AICAR (Fig. 6D). Together, these results indicated that elevated AMPK activity suppressed the cell division.
Hyperpolarization induced by over-activation of AMPK was reversible in MKN45 cells. We next investigated whether over-activation of AMPK induced by AICAR treatment was reversible. Phosphorylation of AMPK decreased to the basal level within 12 hours after removal of AICAR (Fig. 7A). Accompanied by the decrease in AMPK activity, V mem of the AICAR-pretreated cells became relatively depolarized after AICAR removal for 12 hours, when compared with that of the AICAR-treated cells, and the V mem eventually returned to a nearly normal level (Fig. 7B). Meanwhile, flow cytometry analysis showed that the cells were released from S phase to G2/M phase after removal of AICAR, indicating that cell division resumed after AMPK activity decreased to its basal level (Fig. 7C). These results further confirmed that hyperpolarization was induced by the increase in AMPK activity.
Based on these results, we propose a tentative model of nutrient starvation-mediated hyperpolarization of V mem and regulation of the AMPK signaling pathway in MKN45 cells (Fig. 8). According to this model, AMPK activity increases in response to nutrient starvation, and inhibits expression and activity of CFTR. In turn, suppression of CFTR induces hyperpolarization of V mem , and subsequently inhibits cell division.

Discussion
In this study, we found that nutrient starvation induced reversible hyperpolarization and cell-cycle arrest in the gastric cancer cell line MKN45. This hyperpolarization was induced by an increase in AMPK activation, which inhibited chloride efflux by suppressing CFTR expression. As a cellular energy sensor, AMPK is a therapeutic target in cancer. Acting downstream of tumor suppressor LKB1, AMPK regulates cell growth and proliferation through modulation of multiple signaling pathways, including inhibition of the mTOR pathway and stimulation of p53 52,53 . However, our findings reveal the first mechanistic details of AMPK as a powerful regulator of V mem , which also influences cell division. Elevation of AMPK activity induced hyperpolarization, whereas cells depolarized following reduction of AMPK activity. The changes in V mem induced by AMPK activation were closely related to cell division.
Cancer cells, which exhibit sustained proliferative signaling and aberrant changes in the cell cycle, are effective models in which to study V mem and the ionic regulation, or mis-regulation, of cell division. Alterations and dysfunction of V mem and ion channels have been observed in a variety of cancers 9,54,55 . In general, depolarization serves as a signal that initiates mitosis and DNA synthesis. The mean V mem levels of cancer cells are consistently depolarized relative to those of most normal somatic cells 3 . For example, MCF-7, a breast cancer cell line, has a V mem of −9 mV at G1 phase and hyperpolarizes to about −30 mV in S phase, whereas MCF10A, a normal breast cell line, has a mean V mem value between −40 and −58 mV 3,56 . Selective depolarization of cells expressing the glycine-gated chloride channel in Xenopus embryos results in over-proliferation of melanocytes, but this cancer-related phenotype can be reversed by expression of a hyperpolarizing channel 57 . Also in the Xenopus model, monitoring of a distinctly depolarized resting potential was used for early detection of tumors. Moreover, hyperpolarization itself, regardless of the types of ion channels, can prevent the formation of tumors 58,59 . Here, we found that hyperpolarized MKN45 cells were non-dividing. Further, prevention of hyperpolarization rescued the cell division, revealing that V mem is a bioelectric regulator, and not merely a marker. These results confirmed the general regulation of cell division by V mem .
Fluctuations in chloride concentration may contribute to changes in V mem during the cell cycle, and chloride currents play functional roles in the proliferation of a variety of cell types 3,9,20 . The chloride channel blocker 5-N-2-(3-phenylpropylamino) benzoic acid (NPPB) inhibits entry into S phase, and also suppresses the transition from quiescence into the cell cycle in NIH3T3 cells 60 . Moreover, several chloride channels have been implicated in cell division. For example, absence of CLC-3 chloride channel expression inhibits proliferation of vascular smooth muscle cells and glioma cells 61,62 . Fluctuations in chloride concentration also contribute to changes in cell volume for the survival of cells, and animal cells show a regulatory volume decrease by releasing intracellular Clafter osmotic swelling 63 .
Our investigation of the role of AMPK in regulation of V mem revealed that chloride concentration was influenced by AMPK activity in MKN45 cells: specifically, intracellular chloride concentration significantly increased with AMPK activation, and decreased upon inactivation of AMPK, indicating that AMPK regulates chloride concentration and efflux. Specific inhibition of CFTR, a bidirectional anion channel, increased the intracellular chloride concentration, resulting in hyperpolarization. Because fluctuation in the chloride concentration is necessary for maintenance of V mem and cell division, mis-regulation of chloride concentration inhibited cell division. CFTR is an interaction partner of AMPK-α1 subunit, and its activation by the PKA pathway is inhibited by direct phosphorylation of AMPK 10,12 . In this study, we found that the level and membrane localization of CFTR significantly decreased after starvation-induced AMPK activation and AICAR treatment, suggesting the existence of a novel mechanism whereby AMPK down-regulates CFTR. AMPK is an upstream kinase of Nedd4-2, an ubiquitin ligase that targets a variety of ion channel proteins. Stimulation of Nedd4-2 by AMPK promotes endocytosis and degradation of targeted ion channel proteins such as ENaC, Kir2.1, and Orai1 46,47,64 . Otherwise, AMPK inhibits phosphatase and tensin homolog (PTEN) activity via glycogen synthase kinase 3β to suppress KATP channel trafficking 65 . A. Prince and colleagues recently proposed that PTEN binds CFTR which adds another level of complexity 66 . In addition, AMPK can inhibit activation of nuclear factor kappa B (NF-kB), thereby down-regulating NF-kB-mediated transcription of genes such as Orai1/STIM1 [67][68][69] . Nucleoside diphosphate kinase A (NDPK-A) was found interacts with both AMPK and CFTR in airway epithelia, and NDPK-A catalytic function is required for the AMPK-dependent inhibition of CFTR activity 70 . All of these down-regulatory mechanisms provide clues regarding the reduction of CFTR expression by AMPK activation. However, not only does AMPK regulate CFTR function, but that CFTR, via metabolic-dependent interactions with AMPK, was found regulate AMPK function and subcellular localization in epithelia 71 . The results suggest that CFTR may reciprocally regulate AMPK function in our system.
In our previous study, nutrient starvation induced autophagy in MKN45 cells, and the autophagy was observed after 6 hours of starvation treatment 72 . In this study, the cells hyperpolarized rapidly within 30 minutes of starvation. The response of V mem to starvation was much earlier than that of autophagy. Thus, we supposed that the cells hyperpolarized rapidly to suppress cells division and decrease metabolic level, thus resisting the starvation environment. As the starvation time extending, the autophagy level increased, which further responded to the environmental stress.
In conclusion, we found that elevation of AMPK activity down-regulates expression of CFTR in MKN45 cells, resulting in accumulation of intracellular chloride and subsequent hyperpolarization, ultimately blocking cell division. Both the AMPK signaling system and the V mem regulation network are fairly complex. Our findings provide important insight into the coupling between AMPK, V mem regulation, and ion channel activity. Hence, modulation of AMPK activity represents a potential therapeutic strategy for using in disorders caused by AMPK-sensitive dysregulation of V mem .

Methods
Cell Culture. The MKN45 human gastric cancer cell line was purchased from Bioleaf Biotech. MKN45 Cells were cultured in RPMI-1640 (Corning) supplemented with 10% fetal bovine serum (GIBCO), 100 IU/ml penicillin and 100 μg/ml streptomycin. All cells were cultured in a 37 °C humidified atmosphere containing 5% CO 2 and 95% air.  Immunofluorescence Assay. Cells were fixed in 4% PFA at least 2 hours at room temperature or 4 °C overnight. Then washed the fixed cells in PBS. Cells were blocked in blocking buffer (1% BSA in PBS with 0.25% Triton X-100) for 30-60 minutes at room temperature. Subsequently, cells were incubated overnight at 4 °C with primary antibodies at a dilution of 1:100. The slides were washed three times with PBS and incubated with FITC-conjugated secondary antibody (Invitrogen) at room temperature. Cell nuclei were incubated with DAPI for 15 minutes.
Western Blot Analysis. We performed the Western blot analysis as described previously 74 . Whole-cell lysates were separated by denaturing 10% SDS-PAGE gels and transferred to polyvinylidene fluoride membranes (Bio-Rad). The membranes were incubated at 4 °C overnight with specific primary antibodies after blocking in 1% blocking buffer for 1 hour, primary antibodies used were: anti-phospho-AMPK anti-phospho-H3, anti-H3 (Cell Signaling Technology), and anti-CFTR (Abcam). An anti-tubulin antibody (Sigma-Aldrich) was used as a loading control. Then the membranes were incubated in the appropriate secondary antibodies for 2 hours at room temperature. Use the BM Chemiluminescence Western Blotting kit (Roche) to detect the immunoreactive bands.
Cell Proliferation Assay. Cell viability and proliferation were assessed using CCK-8 kit (Beyotime Biotechnology). Cells were treated with EBSS, after that, Cells were planted in 96-well plates (2000 cells/well) in triplicate for 12 hours. Use the Multiskan EX plate reader (Thermo Fisher Scientific) to quantify the viable cells by measuring absorbance at 450 nM.
Cell-Cycle Analysis. Fixed cells were treated with RNase A (100 μg/ml; Sigma) and stained for 30 minutes with propidium iodide (50 μg/ml; Sigma) at 4 °C. Cell-cycle analysis was performed on a Beckman Coulter Flow Cytometer (FC500MCL).

Statistical analyses.
All quantitative data are expressed as means ± SD at least of three independent experiments. Comparisons between test and control values were analyzed via t test and one way ANOVA, and P values less than 0.05 were considered to be significant.