Nutrient deficiency develops frequently in nasopharyngeal carcinoma cell (CNE-2Z) due to the characteristics of aggregation and uncontrolled proliferation. Therefore, starvation can induce autophagy in these cells. Chloride channel 3 (ClC-3), a member of the chloride channel family, is involved in various biological processes. However, whether ClC-3 plays an important role in starvation-induced autophagy is unclear. In this study, Earle's balanced salt solution (EBSS) was used to induce autophagy in CNE-2Z cells. We found that autophagy and the chloride current induced by EBSS were inhibited by chloride channel blockers. ClC-3 knockdown inhibited the degradation of LC3-II and P62. Furthermore, when reactive oxygen species (ROS) generation was suppressed by antioxidant N-acetyl-l-cysteine (L-NAC) pretreatment, EBSS-induced autophagy was inhibited, and the chloride current was unable to be activated. Nevertheless, ClC-3 knockdown had little effect on ROS levels, indicating that ROS acted upstream of ClC-3 and that both ROS and ClC-3 participated in EBSS-induced autophagy regulation in CNE-2Z.
Nasopharyngeal carcinoma (NPC) is a common malignant cancer of the head and neck arising from the epithelium of the nasopharyngeal mucosa . It shows a characteristic geographical distribution with a high prevalence in Asia, particularly in the Guangdong Province in Southern China where the estimated annual incidence varies from 15 to 50 cases per 100 000 individuals [2,3]. Although advances in diagnostic imaging, radiation therapy and concurrent chemoradiotherapy have achieved better locoregional control, the final treatment outcomes and the prognosis are not satisfactory.
As an evolutionarily conserved process, autophagy is a mechanism that cells use to adapt to different microenvironments, such as hypoxia, extreme pH and nutrition deficiency . Due to cell aggregation and uncontrolled proliferation in tumors , cells are incapable of assimilating enough nutrients, resulting in starvation-induced autophagy .
Chloride channels participate in many biological processes . ClC-3, as a significant member of the voltage-gated chloride channel family, participates in cell proliferation, the cell cycle and apoptosis, especially in tumor cells . A recent study has shown that ClC-3 was rapidly activated by bufalin and subsequent apoptosis was induced through inhibition of the PI3K/Akt/mTOR pathway , which was significantly relevant with autophagy.
Reactive oxygen species (ROS) play an important role in many physiological processes, such as apoptosis and autophagy [10–13]. Under the pressure of starvation, excessive ROS is generated, which damages cells by making functional changes in DNA, lipid and protein [12,14]. Our previous study demonstrated that ClC-3-mediated chloride current induced by zoledronic acid was suppressed by inhibition of ROS generation . However, in Earle's balanced salt solution (EBSS)-induced autophagy in CNE-2Z cells, the relationship between ROS and ClC-3 is unknown.
In this study, we examined the hypothesis that autophagy induced by starvation is mediated through chloride channels, especially the relationship between CIC-3 and autophagy. EBSS was used to activate autophagy, providing a microenvironment that lacks nutrients for the lumped human NPC cell CNE-2Z. The autophagy-related proteins LC3 and P62 were used as markers to detect autophagy. The relationships between ClC-3 and ROS in starvation-induced autophagy were also examined.
Materials and methods
The poorly differentiated human NPC cell CNE-2Z was grown in 25 cm2 plastic culture flasks in RPMI 1640 medium (Gibco, U.S.A.), supplemented with 10% FBS, 100 IU/ml penicillin, and 100 g/ml streptomycin (37°C, 5% CO2). The cells were subcultured every 2 days with 0.25% trypsin. For chloride current measurement, the cells were relocated on 22 mm round coverslips and incubated for 1 h. For starvation incubation, RPMI 1640 medium was replaced with EBSS after growing cells at a cell density of 70–80% and carried on cell culture in 6 h for the following experiments.
Solutions and chemicals
EBSS, without calcium, magnesium and phenol red, was purchased from Gibco (EBSS, 14155-063, Thermo Fisher Scientific, Waltham, MA). The formulation of EBSS (in mM) was 5.33 KCl, 117.24 NaCl, 26.19 NaHCO3, 1.01 NaH2PO4·H2O, and 5.56 d-Glucose. The antioxidant N-acetyl-l-cysteine (L-NAC, A7250, 100 mM in PBS) and the autophagy inhibitor chloroquine phosphate (C6628, 40 mM in DMSO) were purchased from Sigma–Aldrich (St Louis, MO, U.S.A). The components of the pipette solution and isotonic solution (in mM) were 70 N-methyl-d-glucamine chloride (NMDG-Cl), 1.2 MgCl2, 10 HEPES, 1 EGTA, 140 d-mannitol, and 2 ATP. Tris-base was used to achieve a pH of 7.4, and d-mannitol was used to adjust the osmolarity to 300 mOsmol/L. The chloride channel blockers 4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid disodium salt hydrate (DIDS, D3514) and 5-nitro-2-(3-phenylpropylamino) benzoic acid (NPPB, N4497) was purchased from Sigma–Aldrich (St. Louis, U.S.A.), dissolved in DMSO and diluted to 100 mM. All of the stock solutions were stored at −20°C and diluted to the final concentration before application.
Whole-cell current recordings
The whole-cell currents of CNE-2Z cells were measured by the patch-clamp technique with an EPC-7 patch-clamp amplifier (List Electronic, Darmstadt, Germany). The CNE-2Z cells were put into 22 mm round coverslips and incubated at 37°C for 1 h in order to make cells adherent in RPMI 1640 medium before chloride current was recorded. Before current was recorded, the base current was set to zero potential. When the background chloride currents were stable for 5 min, the cells were perfused with the required solutions. The patch-clamp pipettes, which were made from glass microtubules on a two-stage vertical pipette puller (PC-10, Sutter Instrument Co., U.S.A.), displayed a 4–6 MΩ resistance when filled with the pipette solution. After establishing a whole-cell mode, 200-ms pulses of ±0, ±40, and ±80 mV were stepped repeatedly with 4-s intervals between steps. A computer was utilized to record command voltages and whole-cell currents through a laboratory interface (CED 1401, Cambridge, U.K.). EPC software was used to perform the current analysis (CED, Cambridge, U.K.). All experiments were processed at room temperature (20–25°C). When current reached the value of 1.5 times of based current, the correspondent time of inactivated currents was defined. The equilibrium potential was calculated following Nernst equation: , where R is gas constant, T is absolute temperature, n is number of electrons transferred, F is Faraday constant, [Cl−]out and [Cl−]in are chloride concentration in bath solution and pipette solution.
The ClC-3 siRNA sequence was as follows: sense: 5′-CAA UGG AUU UCC UGU CAU ATT-3′ and antisense: 5′-UAU GAC AGG AAA UCC AUU GTA-3′. The sense strand of the negative control siRNA (NC siRNA) was 5′-UUCUCCGAACGUGUCACGUTT-3′, and the antisense strand was 5′-ACGUGACACGUUCGGAGAATT-3′. CNE-2Z cells were cultured in RPMI 1640 medium without antibiotics in 6-well culture plates until reaching a 40–60% cell density. The transfections were performed with siRNAs by Lipofectamine 2000 (1 µl in 500 µl serum-free medium, Invitrogen) for 6 h, and subsequently, the cells were cultured in PMI 1640 medium containing serum and antibiotics for 48 h.
Western blot analysis
CNE-2Z cells were lysed by RIPA lysis buffer (FUDE, FD008), and 30 µg of protein was loaded onto SDS–PAGE gels for each sample and subsequently transferred to PVDF membranes (0.45 µm, Millipore, Bedford, MA). The membranes were blocked at room temperature for 2 h in 5% evaporated milk dissolved in TBST (0.1% Tween 20). The membranes were incubated overnight at 4°C with primary antibodies, including rabbit anti-ClC-3 antibody (1:1000, ab28736, Abcam); rabbit anti-GAPDH antibody (1:750, D16H11, Cell Signaling); rabbit anti-P62 antibody (1:1000, PM045, MBL) and rabbit anti-LC3 antibody (1:1000, L8918, Sigma). The secondary HRP-conjugated antibody was goat anti-rabbit IgG (H + L) (SA00001-2, Proteintech). The band intensity was detected by Western Blot Luminol Reagent (Santa Cruz, CA, U.S.A.).
MDC staining was used to detect the formation of autophagosomes and acidic organelles. CNE-2Z cells were cultured in 6-well culture plates for 24 h until reaching a 40–60% cell density on the plate. The supernatant was removed, and the cells were resuspended with trypsin (2.5 mg/ml) and then incubated with the MDC (1 µmol/l) at 37°C for 30 min. The samples were analyzed by fluorescence microscopy (excitation 388 nm and emission 450 nm, 400× magnification).
The cells were trypsinized and incubated with DCFH-DA for 1 h and resuspended three times in RPMI 1640 medium (without serum and antibiotics). Then, flow cytometry (488 nm excitation and 525 nm emission) was used to detect the ROS level. The relative fluorescence intensity was calculated as follows: (fluorescence intensity in the experiment group/fluorescence intensity in the control group) × 100%.
The values in this study are presented as the mean ± SE (n, number of frequencies), which were analyzed with one-way ANOVA, as appropriate. P < 0.05 was considered to show a statistically significant difference. All statistical analyses were performed using SPSS 13.0 software (Chicago, IL, U.S.A.).
EBSS activates chloride currents in CNE-2Z cells
A basal current density was recorded with the values of 5.68 ± 1.41 pA/pF at +80 mV and −4.62 ± 0.99 pA/pF at −80 mV. Under EBSS-induced starvation, a current was induced in 107.48 ± 21.32 min, which reached a peak within 174.16 ± 18.88 min (n = 10, Figure 1A,B,E,F). The results showed an outward rectification with a current density of 48.09 ± 6.41 pA/pF at +80 mV and −35.36 ± 4.53 pA/pF at −80 mV (n = 10). A significant difference had revealed between the current density at +80 mV and −80 mV (**P < 0.01). The I–V curve demonstrated that the current reversed at −9.98 ± 2.64 mV (n = 10, Figure 1D,H), closed to the with the value of −13.5 mV.
EBSS induces chloride current in CNE-2Z cells.
To identify the characteristics of the EBSS-induced chloride current, the chloride channel blockers, NPPB and DIDS, were used to inhibit the chloride current. When the currents were stable within 3–5 min, the cells were exposed to the chloride inhibitor NPPB or DIDS (100 µM), showing a mean current density of 14.19 ± 3.25 pA/pF at +80 mV and −10.79 ± 3.08 pA/pF at −80 mV (Figure 1C,G). The difference between the current activated under EBSS condition and that under NPPB and DIDS condition was significant (**P < 0.01). These results initially demonstrated that EBSS induced the activation of the chloride current in CNE-2Z cells.
Chloride channel blocker inhibits the autophagy induced by EBSS
To verify whether chloride channels regulate autophagy in CNE-2Z cells, autophagic relevant protein expression was analyzed. After treatment with EBSS for 6 h, the expression of the autophagy substrate P62 was decreased, and the ratio of LC3-II/LC3-I was increased (**P < 0.01), exhibiting the progression of autophagy. Expression of LC3-II was increased by pretreatment with the chloride channel blockers NPPB and DIDS for 20 min. However, P62 suppression was prevented under EBSS conditions with NPPB or DIDS pretreatment (Figure 2A–D). The suppression of LC3-II and P62 degradation implied that the degradation of autophagosomes was inhibited.
Chloride channel blockers NPPB and DIDS suppressed EBSS-induced autophagy in CNE-2Z cells.
To further examine the effects of chloride channels, MDC staining was used to detect the formation of autophagic vacuoles and acidic organelles. The results showed that when the chloride channel was blocked by NPPB or DIDS, the green fluorescence signal was increased in the cytoplasm (Figure 2E), exhibiting the augmentation of autophagosomes and acidic vesicles. These results showed that chloride channel suppression inhibited autophagy by preventing the fusion of autophagosomes and lysosomes.
ClC-3 knockdown suppresses EBSS-induced current and autophagy
To determine whether ClC-3 regulates EBSS-induced chloride currents, we knocked down ClC-3 expression in CNE-2Z cells with ClC-3 siRNA conjugated to 5′-FAM, resulting in 50% ClC-3 knockdown (n = 3, Figure 3A,B, **P < 0.01). The density of the background current was recorded as 1.32 ± 0.21 pA/pF at +80 mV and −1.16 ± 0.28 pA/pF at −80 mV (n = 5, Figure 3C). The density of the current was 10.69 ± 0.91 pA/pF at +80 mV and −9.05 ± 0.52 pA/pF at −80 mV (n = 5, Figure 3D) when the cells were exposed to EBSS, in contrast with the control group which recorded a current density of 52.14 ± 5.75 pA/pF at + 80 mV and −38.01 ± 4.52 pA/pF at −80 mV (n = 5, Figure 3C). The I–V curve demonstrated that the current reversed at −8.63 ± 2.15 mV (n = 5, Figure 3E), closed to the with the value of −13.5 mV. The difference between currents with ClC-3 siRNA treatment and the control at ±80 mV was significant (**P < 0.01). The results indicated that ClC-3 involved in EBSS-induced chloride current.
EBSS-activated chloride current and autophagy were suppressed by ClC-3 siRNA treatment.
In line with down-regulation of ClC-3 knockdown on EBSS-induced chloride current, the EBSS-induced autophagy was also notably suppressed. Compared with the control group, LC3-II/LC3-I was increased and P62 degradation was suppressed in the ClC-3 knockdown group (Figure 3F,G), similar to that in the group pretreated with the typical autophagy inhibitor chloroquine (Figure 3H,I) for 1 h, which inhibits the fusion of autophagosomes and lysosomes, inhibiting the degradation of P62 and LC3-II. MDC staining showed that after treatment with ClC-3 siRNA, there were significantly more green fluorescent dots in the treated cells than in the control cells (Figure 3J). Contrast with chloroquine pretreatment, these results indicated that ClC-3 suppression inhibited autophagy by preventing autophagosome degradation in CNE-2Z cells. These provide evidence that ClC-3 up-regulates EBSS-induced autophagy.
EBSS-generated ROS involving chloride currents and autophagy
To investigate the role of ROS in EBSS-induced autophagy in CNE-2Z cells, L-NAC, an antioxidant, was used to scavenge the generation of ROS. Here, we recorded the chloride current in CNE-2Z cells after pretreatment with 1 mM L-NAC within 1 h. Then, the substrate was changed to EBSS containing 1 mM L-NAC until the chloride current peaked. The current density was 44.00 ± 2.75 pA/pF at +80 mV and −32.72 ± 1.77 pA/pF at −80 mV (n = 5, Figure 4A) when exposing to EBSS. When the cells were pretreated with the antioxidant L-NAC, the current density was 4.54 ± 0.88 pA/pF at +80 mV and −3.01 ± 0.37 pA/pF at −80 mV (n = 5, Figure 4B). There was a significant difference between the current recorded in EBSS and that in EBSS with 1 mM L-NAC (Figure 4C, **P < 0.01).
The antioxidant L-NAC inhibited EBSS-induced chloride current and autophagy.
The results of MDC staining revealed that enhanced green fluorescent dots were recorded after treated with EBSS for 6 h (Figure 4D). When ROS were inhibited by L-NAC, the green fluorescence signal was decreased in the cytoplasm, demonstrating that declining ROS levels inhibited the generation of autophagosomes or acidic vesicles. This result indicated that reduced ROS levels inhibited EBSS-activated autophagy. When CNE-2Z cells were pretreated with the antioxidant L-NAC, ClC-3 expression was suppressed, P62 degradation was prevented, and the transformation of LC3-I to LC3-II was inhibited (Figure 4E,F). L-NAC decreased ROS levels and inhibited EBSS-induced autophagy at the early stage but not the late stage.
To make certain the relationship between ClC-3 expression and ROS levels, ClC-3 was knocked down by ClC-3 siRNA, and ROS levels were detected in CNE-2Z cells. After treatment with EBSS in CNE-2Z cells for 6 h, the ROS level in CNE-2Z cells was significantly increased (**P < 0.01) but had little change in cells which pretreated with L-NAC (Figure 4G,H). Meanwhile, cells treated with ClC-3 siRNA was not significantly different from that in cells without ClC-3 knockdown (Figure 4I,J). The difference between the ClC-3 siRNA group and the negative control was not significant (P > 0.05). These results showed that ROS functioned upstream of ClC-3 in EBSS-induced autophagy.
Autophagy is an important physiology process to resist the pressure from nutrition deficiency . In our study, EBSS was used to activate autophagy in CNE-2Z cells . EBSS has certain advantages compared with other conventional inducers. As an autophagy inducer, EBSS can be considered with respect to the induction time rather than the drug concentration such as rapamycin . Another advantage is that EBSS is not cytotoxic .
Autophagy is closely associated with chloride channels. Cystic fibrosis transmembrane conductance regulator (CFTR) positively regulates autophagy. ΔF508 mutation of CFTR results in crossing bonding of BECN1 and accumulation of SQSTM1/P62 . Besides, ClC-7, as a member of chloride voltage-gated channel, has been reported to contribute to lysosomal storage disease and an increase in LC3-II levels . In our study, activated chloride current and enhanced autophagy were recorded under the pressure of EBSS-induced starvation. The current was weakly outward rectified and was inhibited when exposed to the chloride channel blockers NPPB and DIDS, respectively. It has been suggested that chloride channels are regarded as potential regulators of autophagy. For example, CLIC4 has been reported to regulate autophagy in human glioma U251 cells .
ClC-3 is generally expressed in cell membranes and acidic intracellular compartments, such as late endosomes, lysosomes, and mature autophagosomes . ClC-3 is significantly involved in migration, apoptosis and cell cycle [24–26]. Our results have shown that NPPB or DIDS significantly inhibited EBSS-induced chloride current. However, they are non-specific and low selective chloride channel blockers . Therefore, in order to investigate the function of ClC-3, the use of non-functional mutants of ClC-3 is of particular importance. It has been reported that mutation of Ser109 at the N-terminal of ClC-3 channels significantly inhibited the calcium-calmodulin kinase II (CaMKII)-dependent chloride current . TNF-α could induce chloride current in wild-type vascular smooth muscle cells (VSMC), but this current was inactivated in ClC-3 null VSMC . ClC-3 T532A (mutation of Thr532 to Ala) significantly decreased angiotensin II (AngII)-induced Cl− current and migration of VSMC . In this study, ClC-3 siRNA was used to knockdown of ClC-3 expression specifically. We found that ClC-3 was a key chloride channel in EBSS-induced current, and its expression was increased after EBSS-induced autophagy. A new type of chloride current was found, which was activated slowly, representing the slowly stimulating chloride channel in starvation-activated autophagy. However, the mechanism of ClC-3 in starvation-induced autophagy in CNE-2Z cells is not yet known. As a Cl−/H+ exchanger, ClC-3 could mediate coupled exchange of anions and protons across biological membranes in order to ensure proton supplementation [31,32]. ClC-3 localizes to the cell membrane and subcellular organelles [33,34], and autophagy activation is closely associated with these cell compartments. In addition, it has been demonstrated that different ClC isoforms expression varies in cellular compartments and with separate intracellular targeting . ClC-3a and ClC-3b can only be found in intracellular compartments and they localize to the late endosomal/lysosomal system, whereas ClC-3c is part of the recycling endosome with a large proportion of transporters expressed in the surface membrane . Jonathan W. Wojtkowiak et al.  found that the number of acidic organelles increased when autophagy was activated, and proton transporters contributed to this process. In particular, ClC-3 localizes preferentially in endosomal compartments . The generation of this slow-type chloride current that we found in CNE-2Z cells may be associated with continuous and endogenous proton-activated anion conductance. In this study, we showed that ClC-3 knockdown worked similar to chloroquine, which was able to increase the pH of the intracellular environment, hence tentatively suggesting that CIC-3 may inhibit the fusion of autophagosomes and lysosomes.
ROS have been described as a significant factor in autophagy . The high ROS levels were generated by curcumin, which has been reported to stimulate the conversion of LC3 and the degradation of the autophagic label protein SQSTM1/P62 . In our study, ROS levels were suppressed by the antioxidant L-NAC, and the conversion of LC3-I to LC3-II was also inhibited, demonstrating that ROS stimulated the activation of autophagy. It has been reported that elevated ROS production and subsequent activation of AKT facilitated TP53 degradation in hepatocarcinoma, thus promoting autophagy and survival . Our results showed that ROS induced autophagy via chloride channel activation. However, the mechanism by which ROS regulates the chloride current is unknown. One suggested mechanism is that the volume-sensitive outwardly rectifying (VSOR) chloride channel and the Ca2+-permeable chloride channel cluster together with the Nox family and act as a part of ‘signaling nanodomains’ in the specialized plasma membrane [15,38]. Interactions of cytoskeleton-related proteins are also regulated in the same way by ROS generation from Nox-accelerated VSOR chloride channel activation. Another possible mechanism is that concomitant utilization of H+ and O2− dismutation provides a potential across the membrane which activates a Cl− channel . Consistent with our findings, the chloride current was notably suppressed by antioxidant L-NAC. Meanwhile, ClC-3 expression was increased when ROS generation was improved, but ClC-3 knockdown had a little effect on ROS levels. This implied that ROS functioned upstream of ClC-3. In contrast with the report by Su et al. , ClC-3 knockdown decreases the ROS levels generated by cisplatin in human glioma U251 cells, supporting the view that cisplatin is able to regulate autophagy or apoptosis via other pathways. Here, we assumed that ROS are essential to the function and expression of ClC-3 because of their putative roles in the formation of internal disulfide bonds in ClC-3 and the modification of thiolate anions of the cysteines .
In conclusion, we demonstrated that EBSS treatment of CNE-2Z activated autophagy via ROS accumulation, which was found to subsequently activate ClC-3. All of these observations implied that ClC-3 influences the progression of autophagy. Further understanding of the mechanisms of chloride channel induced autophagy may provide insights into the treatment and therapies for NPC.
cystic fibrosis transmembrane conductance regulator
voltage-gated chloride channel 3
4,4′-diisothiothiocyanatostilbene-2,2′-disulfonic acid disodium salt hydrate
Earle's balanced salt solution
reactive oxygen species
vascular smooth muscle cells
volume-sensitive outwardly rectifying
Z.C., S.P., L.C. and L.W conceived of and designed the research. S.P., L.C. and L.W supervised the study. Y.Z., Z.C., S.P., L.C. and L.W. designed the experiments. Y.Z., Z.C., Z.G., X.Y., M.Y., C.Z., J.L., P.X. and L.Z. performed experiments and analyzed data. C.Z., S.P., L.C. and L.W. interpreted the data and wrote the manuscript. T.J.C.J. revised the manuscript. All authors read and approved the final manuscript.
The study was supported by the National Natural Science Foundation of China (81272223, 81273539), the Science and Technology Programs of Guangdong Province (2017A050501021, 2013B051000059) and the Natural Science Foundation of Guangdong Province (2016A030313495).
Fluorescence microscope images and flow cytometry data were collected at Jiabing Tian's Medical Experimental Center, Jinan University.
The Authors declare that there are no competing interests associated with the manuscript.
These authors contributed equally to this work.