JBC Papers in Press. Published on November 30, 2011 as Manuscript M111.311688 The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.M111.311688 Glioma Specific Cation Conductance Regulates Migration and Cell Cycle Progression* Arun K Rooj1, Carmel M McNicholas1, Rafal Bartoszewski2, Zsuzsanna Bebok2, Dale J Benos1,2#, and Catherine M Fuller1,2. From the Departments of Physiology and Biophysics1 and Cell Biology2; University of Alabama at Birmingham, Birmingham AL 35294. *Running Title: Sodium-dependent migration and proliferation in glioma cells To whom correspondence should be addressed: Catherine M. Fuller Ph.D., Dept. of Physiology and Biophysics, MCLM 830, University of Alabama at Birmingham, 1530 3rd Avenue South, Birmingham, AL-35294, Tel: 205-934-6227, Fax: 205-975-7679, Email: [email protected] Keywords:, ASIC-1, ENaC, sodium, psalmotoxin-1, MAP-kinase # Deceased ASIC1. In contrast, knocking down δENaC, D o Background: Cation transport contributes to which is not a component of the glioma cation w n migration and proliferation of tumor cells. channel complex, had no effect on CKI lo a d Results: Sodium current block decreased ERK expression. Phosphorylation of ERK1/2 was ed phosphorylation and increased expression of cell also inhibited by PcTX-1, benzamil, and knock- fro m cycle inhibitors. down of ASIC1 but not δENaC in D54MG cells. h ttp Conclusion: Activity of an ENaC/ASIC cation Our data suggest that a specific cation ://w channel is required to maintain the glioma cell conductance composed of ASIC and ENaC w w phenotype. subunits, regulates migration and cell cycle .jb c Significance: Activity of a membrane cation progression in gliomas. .o rg channel influences signaling pathways to effect Glioblastoma multiforme, (GBM), a grade IV b/ y changes in migration and proliferation of glioma g primary brain tumor is the most common and u e cells. aggressive of all brain tumors, exhibiting high st o rates of proliferation and migration, often setting n J a n SUMMARY up secondary foci some distance from the primary ua In the present study, we have investigated the site. Ion channel activity is closely linked to ry 2 4 role of a glioma specific cation channel changes in proliferation and migration of tumor , 2 0 assembled from subunits of the Deg/ENaC cells (1-3), although the role of ion channels in the 19 superfamily, in the regulation of migration and mechanisms underlying these processes is less cell cycle progression in glioma cells. Channel well understood. There is increasing evidence that inhibition by Psalmotoxin-1 (PcTX-1), expression of Ca2+, K+ and Cl- (4-6) channels is significantly inhibited migration and altered in many cancer cells and that the cellular proliferation of D54-MG glioma cells. Both pathophysiology is influenced by abnormal PcTX-1 and benzamil, an amiloride analog, activities of these channels. Other basic caused cell cycle arrest of D54-MG cells in phenotypic characteristics of tumor cells, such as G0/G1 phases (by 30% and 40%, respectively) cell adhesion (7), and interaction with the and reduced cell accumulation in S and G2/M extracellular matrix (8), are also influenced by ion phases after 24h of incubation. Both PcTX-1 channel activity. Regulation of these parameters and benzamil up-regulated expression of cyclin- in both normal and tumor cells is also under the dependent kinase inhibitor proteins (CKI) control of a diverse array of signaling pathways. p21Cip1 and p27Kip1. Similar results were Chief among these is the mitogen activated protein obtained in U87MG and primary GBM cells kinase (MAPK) pathway, which regulates glioma maintained in primary culture and following cell migration, proliferation, and differentiation knockdown of one of the component subunits, (9,10). A key downstream signaling element of 1 Copyright 2011 by The American Society for Biochemistry and Molecular Biology, Inc. this pathway is the highly pluiripotent signaling Dulbecco’s modified Eagle’s medium/F-12 intermediate ERK1/2. Activation of this pathway medium (Invitrogen, Carlsbad, CA) supplemented is generally attributed to ligand binding to EGFR with 10% fetal bovine serum (Thermo-Fisher, and related receptors, initiating a classical tyrosine Rockford, IL) in the absence of antibiotics. To kinase signaling cascade. In epithelial cells, ERK generate stable cell lines, D54-MG cells were activation is associated with decreased surface transfected with 4µg of a truncated-eGFP-ASIC1 expression of ENaC sodium channels, (11,12), or eGFP-δENaC cDNA as previously described exerting effects on mechanisms of channel (15). Following transfection cells were cultured retrieval from the membrane. for 72h, transferred to a T75 flask and selected We have previously shown that malignant glioma with G418 (500 µg/ml; Mediatech, Manassas, cells express an inward Na+ current which is not VA). After initial selection, GFP positive cells detected in normal astrocytes or low grade gliomas were sterile sorted by FACS. Stable transfectants and which is blocked by the diuretic amiloride were maintained in DMEM-F12 media with 10% (13). This cation current is due to expression of a FBS. All constructs used were the human cross-clade channel composed of members of the variants, subcloned into pcDNA3.1 for expression. Deg/ENaC channel superfamily; acid-sensitive ion Transwell Migration Assay - Transwell channel 1, (ASIC1), and two members of the migration assays were conducted as previously D epithelial sodium channel family, α-, and ENaC described (16). Cells were split one day prior to o w (13-16). The amiloride-sensitive cation the experiment and treated with PcTX-1 (100 nM), nlo a conductance found in glioma cells is also blocked control peptide (100 nM) (Peptides International, de d by psalmotoxin-1 (PcTX-1), a peptide isolated Louisville, KY) or benzamil (100 µM) (Sigma- fro m from a West Indies tarantula Psalmopoeus Aldrich, St. Louis, MO). These drug h cambridgei (17,18), and which is a highly specific concentrations were used for all subsequent ttp blocker of ASIC1. Knocking down expression of experiments. The control peptide had an identical ://w w any of the three Deg/ENaC subunits abolishes the amino acid sequence as PcTX-1 except all the w.jb current and slows migration (15). Furthermore, cysteines were replaced with alanine, thus c.o the regulatory volume increase of glioma cells breaking the disulfide bonds of this inhibitor brg/ following cell shrinkage by hyperosmolar cysteine knot fold toxin. On the day of the y g solutions was completely abolished by PcTX-1 experiment, 1 × 104 cells were added to the insert, ues and amiloride (19). Based on these previous and allowed to migrate for 5h (37°C, 95% air/5% t on J findings, we wanted to determine if PcTX-1 could CO ), in the continued presence of the drugs, an 2 u a affect glioma cell migration and proliferation. We which were added to both sides of the filter. ry 2 report here that exposure to PcTX1, an amiloride Following incubation, cells were fixed (4% 4 , 2 analog benzamil, low extracellular [Na+], or paraformaldehyde) for 10 min and stained with 01 9 knockdown of the ASIC1 subunit, reduced 1% crystal violet solution (5 min). Images from migration, caused accumulation of p21Cip1 and five random fields per insert were taken at 20X p27Kip1 slowing proliferation, and decreased magnification. All experiments were done in phosphorylation of ERK1/2 in two tumor cell triplicate and repeated a minimum of three times lines, D54-MG and U-87MG, and in GBM cells per condition. maintained in primary culture. These observations Scratch Wound Migration Assay - D54-MG reveal that activity of this channel is essential for cells (either wild type or stably transfected with the maintenance of the glioma cell phenotype. truncated eGFP-ASIC1) were seeded on a 6 well plate at a density of 7 × 105 cells/well. The EXPERIMENTAL PROCEDURES monolayer was scarred with a 200µL sterile Cell Culture - The cell lines, D54-MG, and pipette tip and washed two times with serum free primary cultures of human glioblastoma cells were DMEM/F-12 media. Immediately after scarring, kind gifts of Dr. D Bigner, (Duke University, cells were treated with PcTX-1, control peptide, or Durham, NC) and Dr. G.Y Gillespie, (University benzamil for 24h (DMEM/F12 with 2% FBS) at of Alabama at Birmingham, Neurosurgery Brain 37oC. Images were taken at 0h and after 24h of Tissue Bank). U87MG cells were purchased from treatment and the cell-free area measured. The ATCC. The cells were cultured and maintained in 2 experiments were performed at least three times IN), for 30 min at 4oC with shaking. Cell lysates per condition. Images were captured on a Nikon were homogenized by passing 10 x through a 22g 200 TEV inverted microscope. IPLab software needle and centrifuged (13,200 rpm 30 min 4oC). was used to measure the region of interest (ROI) Protein concentration of the supernatant was in the scratch area. measured using BCA protein assay (Thermo- Isolation of Cellular RNA –Total cellular RNA Scientific) and 25-50µg of protein lysates were was isolated from D54-MG cells using RNeasy used per lane for SDS-PAGE and immunoblotting. (Qiagen, Valencia, CA), according to the Lysates were heated at 95oC for 6-7 min in 1X manufacturer's recommendations. RNA was Laemmli sample buffer (25% glycerol, 2% SDS, isolated from D54-MG cells treated with PcTX-1 0.01% bromophenol blue, 10% β- (100nM) or control peptide (100nM) or benzamil mercaptoethanol, 62.5 mM Tris HCl, pH 6.8) and (100µM) or incubated in DMEM-F12 (with 2% or subjected to SDS-PAGE over 8% or 12% 10% FBS) or Kreb’s buffer or NMDG+ buffer for separating gels. Proteins were transferred to 24h or from D54MG cells stably transfected with Immobilon-P transfer membranes (Millipore, either DN-eGFP-ASIC1 cDNA or DN-eGFP- Hayward, CA). Following transfer, membranes δENaC cDNA. RNA concentration was calculated were blocked for 1h with 5% nonfat dry milk or based on the absorbance at 260 nm. RNA samples 10% BSA (for phosphor-specific antibody) in were stored at −20 °C. Tris-buffered saline (100mM Tris (pH-7.5), 150 Do w mM NaCl), with Tween-20 (0.1%: Bio-Rad, n Measurement of Relative mRNA Levels Using lo Hercules, CA; TBS-T) for 1h at room temperature ad Real Time PCR - Real time PCR to measure p21 e d (Hs00355782_m1) and p27 (Hs00153277_m1) a1n0d% p BroSbAed i nw iTthB Spr-iTm oarvye rannigtihbto dati e4s oiCn. 5B%lo mts iwlke orer from mRNA was performed using TaqMan® One-Step h RT-PCR Master Mix Reagents (Applied wseacsohnedda rwyi tha nTtiBboSd-Tie s( 3xco, n5j umgiante),d a ntod phroorbseedra wdiisthh ttp://w Biosystems Carlsbad, CA) according to the w peroxidase (HRP; Thermo-Fisher) in 5% milk in w manufacturer’s directions (Relative TBS-T. The blots were developed in SuperSignal .jbc Quantification; Applied Biosystems 7300/7500 .o West Pico substrate (Thermo-Fisher) and exposed rg Real Time PCR system; 2004). 18 S rRNA b/ to X-ray film. The X-ray films were scanned y (Hs99999901_m1) was amplified as an internal g using a Syngene G-Box and images were analyzed ue control and used as a reference. for densitometry by GeneTools software st o n Flow Cytometry - Cells were seeded in a six (Syngene). Ja well plate at a density of 2 × 105/well. After 24h nu a Antibodies and Drugs - The following ry the cells were washed three times with PBS and 2 antibodies were used - mouse anti-GFP 4 then incubated with serum free DMEM/F12 for , 2 monoclonal antibody (Abgent, San Diego, CA) at 0 48h to synchronize the cells in G0/G1 phase. The 1:2000; mouse anti-p21 monoclonal antibody 19 cells were incubated with either PcTX-1, control (Abcam, Cambridge, MA) at 1:2000; mouse anti- peptide, or benzamil for 24h in reduced serum p27 monoclonal antibody (Cell Signaling, media (DMEM/F12 with 2% FBS) at 37oC. After Danvers, MA) at 1:1000; rabbit anti-ERK1/2 fixing with 95% ethanol, the cells were treated polyclonal antibody (Millipore) at 1:10,000; with RNAase solution (1mg/ml in PBS; Thermo- mouse anti-phospho-ERK1/2 monoclonal antibody Fisher), stained with propidium iodide (40µg/ml) (Millipore) at 1:10,000; rabbit anti-ASIC1 and (Sigma-Aldrich) and sorted at the FACS core anti-δENaC (Santa-Cruz Biotechnology, Santa facility in the Center for AIDS Research (UAB). Cruz, CA) at 1:500; mouse anti-actin monoclonal Cell cycle phases were analyzed by FACS-Diva antibody (Millipore) at 1: 200,000. Because software (Beckton-Dickinson, Franklin Lakes, glioma cells release proteases, we used PcTX-1 NJ). (and control peptide) at 100 nM for long-term Cell Lysates, SDS-PAGE and Immunoblotting - (24h) inhibition of the glioma specific cation Cells were washed twice with cold PBS and lysed conductance. Similarly, we conducted in buffer (150 mM NaCl, 5 mM EDTA, 50 mM experiments at reduced (2%) serum concentrations Tris pH 7.4, 1% Triton-X-100, Complete® protease inhibitor cocktail (Roche, Indianapolis, 3 to limit binding of the toxin to plasma proteins. (24h) inhibition of the cation conductance by Benzamil and amiloride were used at 100 µM. PcTX-1 affected the migration of D54-MG cells. Electrophysiology - Whole-cell current If the channel is crucial for the migratory recordings were obtained using the amphotericin B phenotype of glioma cells, PcTX-1 treated D54- perforated patch technique as previously described MG cells should show a lower rate of migration as (20). Cells were mounted in a flow-through compared to untreated cells or cells treated with chamber on the stage of a Leica DM IRB inverted the control peptide. As shown in Fig. 1A/B, we microscope (Leica Microsystems, Heidelberg, found that PcTX-1 and benzamil significantly Germany). Bath solution exchange was achieved inhibited migration (by 48 ± 3% and 62 ± 4% using a pinch valve control system converging on respectively, n ≥ 3) as compared to untreated cells an 8-1 manifold. Tips of borosilicate recording or cells treated with control peptide. Similarly, in pipettes (5-7 mΩ) were back filled with pipette the scratch wound/healing assay, (Fig. 1C/D), solution (150 mM KCl, 10mM HEPES, pH 7.2 cells, treated with PcTX-1 and benzamil had (Tris/HCl)) then with the same solution containing significantly lower rates of migration (by 37 ± ~0.2mg/ml Amphotericin B (Sigma-Aldrich). 13% and 42 ±18% respectively, n≥4), than Currents were obtained using an Axopatch 200B untreated cells or cells treated with control patch clamp amplifier (Axon Instruments/ peptide. D Molecular Devices, Sunnyvale, CA) with voltage PcTX-1 causes cell cycle arrest at G /G phase o 0 1 w commands and data acquisition controlled by and inhibition of cation current in D54-MG nlo a Clampex software (pClamp 8, Axon Instruments) glioma cells - Because closure of the scratch de d and digitized (Digidata 1321A interface, Axon reflects both cell migration and proliferation, we fro m Instruments) at a sampling frequency of 2 kHZ. examined the effect of PcTX-1 on cell cycle h Current-voltage relationships were obtained using progression in D54-MG cells. Using FACS ttp a pulse protocol in which cells were stepped from analysis, we found cell cycle arrest in the G /G ://w 0 1 w a -40 mV holding potential from -100 to +80 mV phase in PcTX-1and benzamil treated cells by 30 ± w .jb in 20 mV increments for 250 ms. Mean currents 7.4% and 40 ± 1%, respectively, as compared to c.o were obtained from the average of 3 sweeps untreated cells or cells treated with control peptide brg/ during the 200-250 ms period of each sweep using (Fig. 2A; n≥5). The number of cells in S-, and y g u Csollaumtipofnist c8o nstoaifntweda r(ei n( AmxMon) 1I4n0s trNuamCeln, t4s).0. KBaCtlh, Gim2/pMly ingp hatsheast twhee re inhciobnitcioomn itaonft lyt her edcuactieodn, est on J 1.8 CaCl , 1.0 MgCl , 10 glucose, 10 HEPES, pH conductance directly influenced cell cycle an 2 2 u a 7.4 (NaOH/HCl). Appropriate vehicle controls progression. The cell cycle is regulated by ry 2 were performed. different classes of cyclin dependent kinases 4 , 2 Data Analysis and Statistics - Statistical (CDKs) whose activities are controlled by CDK 01 9 analysis was done with Microsoft Excel or inhibitors (CKIs) (21). We therefore explored GraphPad Prism 5. All experiments were whether expression of two CKIs, p21Cip1 and performed at least three times. Statistical p27Kip1, were also affected by the toxin. As shown significance is displayed on the figure according to in Fig. 2B, expression of both CKIs was the level of significance (*P<0.05; **P<0.01; significantly increased when D54-MG cells were ***P<0.001). Data are presented as mean ± SD treated with PcTX-1 or benzamil for 24h. and two-tailed t-tests (unpaired or paired) were Reduction of FBS in the media to 2% to reduce performed for two group comparisons and one binding of toxin to plasma proteins was without way ANOVA followed by post-hoc tests were effect (n≥6). used to compare three or multiple groups. We also verified that the amiloride-sensitive cation current which we had previously recorded RESULTS from glioma cells was inhibited by long-term PcTX-1 inhibits migration and proliferation of exposure to PcTX-1 and benzamil. We found that D54-MG glioma cells - As ASIC-1 is one subunit approximately 32% (n≥6) of the total basal current of the glioma specific heteromeric ion channel at -80 mV was amiloride-sensitive in untreated complex, we wanted to determine if prolonged cells (Fig. 2C). In contrast, in PcTX-1 and 4 benzamil pre-incubated cells, only 3.5% and 6.3% as compared to wild type cells. The percentage of respectively of the remaining basal current was cells in S and G /M phases was significantly 2 amiloride-sensitive, consistent with a nearly decreased in D54MG-A1DN cells (by 29 ± 7% complete abrogation of current by the drugs. and 66 ± 18% respectively, n≥3). Consistent with Under these conditions, the control peptide had no our FACS data, we also found that ASIC-1 effect. The I/V relationships for each condition knockdown significantly increased the expression described above are shown in Fig. 2D. of p21Cip1 (by 3 fold, n≥6), and p27Kip1 (by 2.5 Representative individual whole-cell current fold, n≥5), as compared to the wild type cells (Fig. records under each experimental condition are 4D). As we have previously shown that δENaC is shown in Supplemental Fig. 1. not a part of the glioma cation channel complex as knocking down this subunit did not affect the Knock-down of ASIC-1 expression inhibits amiloride-sensitive glioma cation current (15), we migration and cell-cycle progression in D54-MG cells - We generated a stable D54-MG cell line established a stable D54-MG glioma cell line expressing an eGFP-tagged ASIC-1 construct expressing a GFP-tagged truncated (S35X) δENaC (DN-eGFP-ASIC1), containing a premature stop construct. Western blot analysis (Fig. 5A) of mutation at Y67 (Y67X). This construct whole cell lysates showed that the expression of effectively knocks down expression of both glycosylated (upper band ~100 kDa) and non- endogenous ASIC1, but does not affect expression glycosylated (lower band ~75 kDa) forms of Do w of other related subunits (15). We analyzed whole δENaC (22) were significantly reduced (by 50 ± nlo cell lysates obtained from both wild type D54-MG 15%, n≥3) in D54MG-δENaC-DN cells as ade d (trDa5n4sfMecGte-dW Tw) ithc elltsh e anDd ND-e5G4F-MP-GA SIcCel1ls csDtaNblAy cpo2m1Cpipa1r eadn dt op 2D7K5i4p1M wGe-rWe Tsi mciellalrsl.y Euxnparfefesscitoend boyf h from (D54MG-A1DN) for ASIC-1 protein expression. knocking down δENaC (Fig. 5B). ttp Quantitative densitometric analysis showed that We also repeated these experiments in wild-type ://w w the expression of ASIC-1 was significantly D54-MG cells under conditions of low (25 mM) w .jb decreased by 58 ± 13% (n≥4) in D54MG-A1DN external Na+. Incubation with low [Na+] media c.o cells (Fig. 3A). Whole-cell patch-clamp recording increased the proportion of cells in G0/G1 phase as brg/ demonstrated that only 9% of basal current compared to those in normal buffer. The number y g u eaxphpirboixtiemd atealmy il3o2r%ide i-nse nthseit iwviitlyd -tyinp e ccoelnltsr a(snt≥ t6o; owfe rcee lclso rirne sspuobnsdeiqnugelyn t loSw pehr aisne caenldls Gin2c/Mub apthedas eins est on J Figs. 3B, 3C). Representative individual whole- NMDG buffer, indicating that the extracellular an u a cell current records of wild type and stably sodium concentration is a critical regulator of the ry 2 transfected cells under each experimental cell cycle. As shown in Figs 6A and 6B, lowering 4 , 2 condition are shown in Supplemental Fig. 2. the external sodium concentrations evoked 0 1 9 We hypothesized that knocking down ASIC-1 in essentially identical results to those for the glioma cells would have similar inhibitory effects D54MG-A1DN cells. The expression of p21Cip1 as long term inhibition of ASIC-1 by PcTX-1 or (n≥5) and p27Kip1 (n≥4) were significantly benzamil. In both transwell migration (Fig. 4A), upregulated when the cells were exposed to low and scratch/wound healing assays (Fig. 4B), [Na+] buffer as compared to control buffer. These D54MG-A1DN cells showed a significant findings suggest that disrupting Na+ transport reduction in migration (by 40 ± 5%, n≥3, and by across the membrane by whatever mechanism 39± 16%, n≥4), respectively, as compared to wild inhibits cell migration and proliferation. type D54-MG cells. Inhibition of the glioma cation conductance Similarly, ASIC-1 knock down inhibited decreases phosphorylation of ERK1/2 - The progression through the cell cycle consistent with mitogen activated protein kinase (MAPK) the results from our earlier experiments (Fig. 4C). signaling cascade is a crucial intracellular After 48h of serum starvation, exposing D54MG- regulatory pathway which controls diverse cellular A1DN cells to DMEM-F12/10% FBS for 24h, processes including cell proliferation, cell cycle significantly increased the percentage of the cell progression, cell survival and angiogenesis (9,23). population in the G /G phase (by 19 ± 5%, n≥3), 0 1 Our findings that glioma cell migration, 5 proliferation and the cell cycle were down- PCR analysis demonstrated that in all cases the regulated by PcTX-1, benzamil, low extracellular increases in p21Cip1 expression were due to an [Na+] and by knocking down the expression of increase in mRNA; in contrast, message levels of o ASIC-1, prompted us to determine the p27Kip1 were not affected by any experimental phosphorylation status of ERK1/2 in all of our maneuver (Supplementary Fig. S5). previous experimental conditions. We found that phosphorylation of ERK1/2 was significantly DISCUSSION inhibited when we treated the cells with PcTX-1 Grade IV gliomas are highly invasive tumors, (by 82 ± 22%, n≥3 or benzamil (by 82 ± 24%, distinguished by their ability to migrate through n≥3), as compared to control cells (in 2% or 10% the brain parenchyma to establish secondary foci FBS) or cells treated with control peptide (Fig. distant from the primary tumor. Multiple 7A). Stably knocking down ASIC-1 in D54-MG laboratories have reported that ion transport plays cells also significantly decreased ERK1/2 a critical role in both migration and proliferation phosphorylation, by 47 ± 20%; n≥6; Fig. 7B). In of glioma cells, and potassium, chloride, calcium, contrast knocking down δENaC in this glioma cell and voltage-gated sodium channels have all been line had no effect on phosphorylation of ERK1/2 implicated (1,28,29). Earlier studies proposed that (Fig. 7C). Low [Na+] also down-regulated Na+ influx was itself a mitogenic signal, initiating phosphorylation of ERK1/2 in D54-MG glioma cell cycle progression in neuroblastoma-glioma D o w cells following 24h of incubation (Fig. 7D, hybrid cells, as inhibitors of sodium transport n lo n≥4).These findings suggest that normal channel reduced the DNA synthesis required for cell ad e d a c t i v Eitxyp irse sresiqouni roefd pf2o1r CEipR1 Kan1d/2 p p2h7oKsipp1h ios riynlcartieoans.e d, pgrroadliifeenrat tioesnt ab(3li0sh,3e1d) . bFyu rtthhee rmNoar+e/,K +thAeT Psaosdei umis h from and ERK1/2 phosphorylation decreased, by PcTX1 importantly used to drive uptake of metabolic ttp and benzamil in primary GBM cells - To substrates such as glucose and amino acids, as ://w w determine if our findings were more widely well as to remove harmful metabolites such as H+ w applicable, we evaluated the effect of current via turnover of the Na+/H+ exchanger (NHE). .jbc.o ipnhhoisbpithioonry laotnio nc eilnl pcryimclea ryp rGogBrMes scioenll s.a n Fdi gE. R8KA Creeglulsl atcianng tarlasnos pcoornt torof lN tah+e, irK +i,n taenrdn aCl l-v aocluromses tbhye by grg/ shows that there was a significantly higher plasma membrane, swelling or shrinking as ues percentage of cells in the G /G phase following required. Although much less sensitive to t on PcTX-1 and benzamil treatm0en1t (26% n≥6, and amiloride than the prototypical renal sodium Jan u 25% n≥6), respectively, as compared to the channel αβ ENaC, (Kiamil < 0.1 µM; (32)), the ary controls. The percentage of cells in S and G /M glioma cation channel that we have identified is 24 phases were correspondingly reduced. As sho2wn critically involved in migration and proliferation , 20 1 in Fig. 8B, expression of p21Cip1 and p27Kip1 in (13-16). This channel is inhibited by amiloride and 9 primary GBM cells was significantly increased by its analogs and is blocked by PcTX-1, to which it treatment with PcTX-1 and benzamil; n≥5). is exquisitely sensitive (Kiamil 10-100 µM; KiPcTX-1 Similarly, the phosphorylation of ERK1/2 was ~ 40 pM; (17)). This peptide exhibits high significantly reduced in PcTX-1 (by 85 ± 5%, n≥4) specificity for ASIC1, with no effect on ASIC2, or benzamil (by 79 ± 24%, n≥4) treated primary heteromers of ASIC subunits or αβ ENaC (18). GBM cells. Essentially identical results were However, PcTX-1 prevented regulatory volume obtained in a second cell line U87MG, with the increase in glioma cells following a hyperosmotic exception of the changes in p21, which were either challenge (19). It has previously been suggested not affected or slightly decreased by the blockers. that in addition ENaC plays an important role in It is possible that p21 is not an essential regulator wound healing (33,34) and proliferation (35). of the cell cycle in U87 cells, as other studies have Both ASIC and ENaC channels contribute to shown that p21 is much slower to respond to a migration of vascular smooth muscle cells (36,37), variety of experimental maneuvers in this cell type although in neither case are the underlying as compared to other glioma cells (24- mechanisms well understood. 27)(Supplemental Figs S3 and S4). Real-time 6 In the present study we have begun to tumors (45), and ERK1/2 phosphorylation is a key dissect out these mechanisms by examining the signaling event controlling migration and effect of current inhibition on cell cycle proliferation of multiple cancer cells including intermediates and the MAP kinase pathway. The gliomas (10,23,46). However, activation of cell cycle is a tightly regulated process which is ERK1/2 in the absence of EGFR amplification has involved in growth, division and differentiation. It been reported, suggesting that non-classical is regulated by both cyclin dependent kinases and pathways are also involved in the regulation of this inhibitor proteins including p21Cip1 and p27Kip1 important transcriptional regulator (45). Down- (38,39). We found that expression of these CKI regulation of ERK1/2 is intimately associated with proteins was up-regulated when the glioma current cell cycle inhibition and accumulation of p27Kip1 was inhibited under all conditions tested. This was (47,48) and p21Cip1 (49). ERK is also a potent specific to the glioma conductance, as knocking regulator of glioma cell migration; inhibition of down ENaC had no effect on CKI expression. ROCK reduced phosphorylation of ERK1/2 and Expression of p21Cip1 and p27Kip1 are primarily decreased cell migration (50). Several transporter regulated by transcriptional and post-translational proteins are also targets of ERK1/2 mechanisms respectively (40), and this was phosphorylation. ERK phosphorylates β- and confirmed by our real time PCR results, which ENaC, increasing retrieval of αβ ENaC from the showed an increased level of mRNA for p21Cip1 membrane (12,51), thus decreasing current. Do w but not p27Kip1 in every condition where we found However, we have now demonstrated the n lo up-regulation of protein expression (Supp. Fig. 5). reciprocal relationship in the glioma cell, as down- ad e It should be noted that consistent with expression regulating the Na+ current inhibited d fro of a non-selective cation channel, we found that phosphorylation of ERK1/2. ERK activation also m h the outward K+ currents, were also blocked by stimulates turnover of the Na+/H+ exchanger or ttp PcTX-1 as well as by benzamil, suggesting that K+ NHE, which is widely implicated in tumor growth ://w w efflux from cells was also inhibited (17). Similar and proliferation (52). This exchanger normally w .jb accumulation of these two CKI proteins was found operates to maintain pH homeostasis, although in c i .o in oligodendrocyte precursors under conditions of gliomas this protein is up-regulated and the pH is rg K+ channel block, and it was suggested that the quite alkaline (53). The NHE is effectiveily by/ g underlying mechanism involved a change in inhibited by 100 µM amiloride, and although we ue s membrane potential (41). However, we observed used 100 µM benzamil, a less potent amiloride t o n J little change in reversal potential in the presence of analog as a positive control (54), it is likely that a n u benzamil or PcTX-1 in our studies, (Fig.2), the effects of benzamil are attributable to ary suggesting that membrane potential is unlikely to inhibition of both the glioma cation conductance 24 be a key trigger. Expression of p21Cip1 was and NHE. However, inhibition of NHE does not , 20 1 associated with increased mRNA, consistent with account for the effects of ASIC1 knock-down on 9 the expression of wild-type p53 in D54-MG cells CKI accumulation and ERK phosphorylation, (42,43). The dissociation between p27Kip1 suggesting a less prominent role for NHE in this accumulation and mRNA may reflect rapid system. turnover of message, increased translation, or In summary therefore, we have shown that increased protein stability (44). blocking of plasma membrane cation channel Our finding that ASIC-1 mediated glioma complex inhibits migration and proliferation of cell cycle arrest was associated with increased glioma cells. At the least, this most likely expression of CKIs, prompted us to examine the involves inhibition of ERK1/2 phosphorylation involvement of the MAP-kinase pathway by and subsequent accumulation of the CKI proteins evaluating the phosphorylation status of ERK1/2. p21Cip1 and p27Kip1. How changes in channel Activation of ERK1/2 is classically downstream of activity are transduced to the MAP kinase pathway the EGFR, which is over-expressed in most GBM remains to be determined. 7 REFERENCES 1. Kunzelmann, K. (2005) J Membr Biol 205, 159-173 2. Roderick, H. L., and Cook, S. J. (2008) Nat Rev Cancer 8, 361-375 3. Soroceanu, L., Manning, T. J., Jr., and Sontheimer, H. (1999) J. Neurosci. 19, 5942-5954 4. Prevarskaya, N., Zhang, L., and Barritt, G. (2007) Biochim Biophys Acta 1772, 937-946 5. Ullrich, N., Gillespie, G. Y., and Sontheimer, H. (1996) Neuroreport 7, 1020-1024 6. Wang, Z. (2004) Pflug. Arch. 448, 274-286 7. Chioni, A. M., Brackenbury, W. J., Calhoun, J. D., Isom, L. L., and Djamgoz, M. B. (2009) Intl. J. Biochem. Cell Biol. 41, 1216-1227 8. Ransom, C. B., O'Neal, J. T., and Sontheimer, H. (2001) J. Neurosci. 21, 7674-7683 9. Fanton, C. P., McMahon, M., and Pieper, R. O. (2001) J. Biol. Chem. 276, 18871-18877 10. McLendon, R. E., Turner, K., Perkinson, K., and Rich, J. (2007) Arch Pathol Lab Med 131, 1585- 1590 11. Falin, R. A., and Cotton, C. U. (2007) J Gen Physiol 130, 313-328 12. Soundararajan, R., Melters, D., Shih, I. C., Wang, J., and Pearce, D. (2009) Proc. Natl. Acad. Sci. USA 106, 7804-7809 13. Berdiev, B. K., Xia, J., McLean, L. A., Markert, J. M., Gillespie, G. Y., Mapstone, T. B., Naren, D o w A. P., Jovov, B., Bubien, J. K., Ji, H. L., Fuller, C. M., Kirk, K. L., and Benos, D. J. (2003) J. n lo Biol. Chem. 278, 15023-15034 ad e 14. Bubien, J. K., Keeton, D. A., Fuller, C. M., Gillespie, G. Y., Reddy, A. T., Mapstone, T. B., and d fro Benos, D. J. (1999) Am J Physiol 276, C1405-1410 m h 15. Kapoor, N., Bartoszewski, R., Qadri, Y. J., Bebok, Z., Bubien, J. K., Fuller, C. M., and Benos, D. ttp J. (2009) J. Biol. Chem. 284, 24526-24541 ://w w 16. Vila-Carriles, W. H., Kovacs, G. G., Jovov, B., Zhou, Z. H., Pahwa, A. K., Colby, G., Esimai, O., w Gillespie, G. Y., Mapstone, T. B., Markert, J. M., Fuller, C. M., Bubien, J. K., and Benos, D. J. .jbc .o (2006) J. Biol. Chem. 281, 19220-19232 rg 17. Bubien, J. K., Ji, H.-L., Gillespie, G. Y., Fuller, C. M., Markert, J. M., Mapstone, T. B., and by/ g Benos, D. J. (2004) Am. J. Physiol.Cell Physiol. 287, C1282-C1291 ue s 18. Escoubas, P., De Weille, J. R., Lecoq, A., Diochot, S., Waldmann, R., Champigny, G., Moinier, t o n D., Menez, A., and Lazdunski, M. (2000) J. Biol. Chem. 275, 25116-25121 Ja n u 19. Ross, S. B., Fuller, C. M., Bubien, J. K., and Benos, D. J. (2007) Am J Physiol Cell Physiol 293, a ry C1181-1185 2 4 20. Akaike, N., and Harata, N. (1994) Jap.J Physiol. 44, 433-473 , 2 0 1 21. Sherr, C. J., and Roberts, J. M. (1999) Genes Dev 13, 1501-1512 9 22. Haerteis, S., Krueger, B., Korbmacher, C., and Rauh, R. (2009) J. Biol. Chem. 284, 29024-29040 23. Nakabayashi, H., and Shimizu, K. Cancer Sci 102, 393-399 24. Jia, L., Soengas, M. S., and Sun, Y. (2009) Cancer Res 69, 4974-4982 25. Komata, T., Kanzawa, T., Takeuchi, H., Germano, I. M., Schreiber, M., Kondo, Y., and Kondo, S. (2003) Br J Cancer 88, 1277-1280 26. Wang, K., Pan, L., Che, X., Cui, D., and Li, C. (2010) J Neurooncol 98, 319-327 27. Xu, J., Sampath, D., Lang, F. F., Prabhu, S., Rao, G., Fuller, G. N., Liu, Y., and Puduvalli, V. K. (2011) J Neurooncol 105, 241-251 28. Cuddapah, V. A., and Sontheimer, H. (2011) Am J Physiol Cell Physiol 29. Joshi, A. D., Parsons, D. W., Velculescu, V. E., and Riggins, G. J. (2011) Mol Cancer 10, 17 30. O'Donnell, M. E., Cragoe, E., Jr., and Villereal, M. L. (1983) J Pharmacol Exp Ther 226, 368- 372 31. O'Donnell, M. E., and Villereal, M. L. (1982) J. Cell Physiol. 113, 405-412 32. Kellenberger, S., and Schild, L. (2002) Physiol Rev 82, 735-767 33. Chifflet, S., Hernandez, J. A., and Grasso, S. (2005) Am J Physiol Cell Physiol 288, C1420-1430 8 34. Del Monaco, S. M., Marino, G. I., Assef, Y. A., Damiano, A. E., and Kotsias, B. A. (2009) J Membr Biol 232, 1-13 35. Bondarava, M., Li, T., Endl, E., and Wehner, F. (2009) Pflug. Arch. 458, 675-687 36. Grifoni, S. C., Gannon, K. P., Stec, D. E., and Drummond, H. A. (2006) Am J Physiol Heart Circ Physiol 291, H3076-3086 37. Grifoni, S. C., Jernigan, N. L., Hamilton, G., and Drummond, H. A. (2008) Microvasc Res 75, 202-210 38. George, J., Banik, N. L., and Ray, S. K. Intl J Biochem Cell Biol 42, 1164-1173 39. Herrero-Gonzalez, S., Gangoso, E., Giaume, C., Naus, C. C., Medina, J. M., and Tabernero, A. Oncogene 29, 5712-5723 40. Slingerland, J., and Pagano, M. (2000) J. Cell Physiol. 183, 10-17 41. Ghiani, C. A., Yuan, X., Eisen, A. M., Knutson, P. L., DePinho, R. A., McBain, C. J., and Gallo, V. (1999) J. Neurosci. 19, 5380-5392 42. Gomez-Manzano, C., Fueyo, J., Kyritsis, A. P., Steck, P. A., Roth, J. A., McDonnell, T. J., Steck, K. D., Levin, V. A., and Yung, W. K. (1996) Cancer Res 56, 694-699 43. Jung, Y. S., Qian, Y., and Chen, X. (2010) Cell Signal 22, 1003-1012 44. Vervoorts, J., and Luscher, B. (2008) Cell Mol Life Sci 65, 3255-3264 45. Lopez-Gines, C., Gil-Benso, R., Benito, R., Mata, M., Pereda, J., Sastre, J., Roldan, P., Gonzalez- D o w Darder, J., and Cerda-Nicolas, M. (2008) Neuropathology 28, 507-515 n lo 46. Glassmann, A., Reichmann, K., Scheffler, B., Glas, M., Veit, N., and Probstmeier, R. (2011) Int J ad e d Oncol fro 47. Gysin, S., Lee, S. H., Dean, N. M., and McMahon, M. (2005) Cancer Res 65, 4870-4880 m h 48. Kortylewski, M., Heinrich, P. C., Kauffmann, M. E., Bohm, M., MacKiewicz, A., and Behrmann, ttp I. (2001) Biochem J 357, 297-303 ://w w 49. Hwang, C. Y., Lee, C., and Kwon, K. S. (2009) Mol Cell Biol 29, 3379-3389 w 50. Zohrabian, V. M., Forzani, B., Chau, Z., Murali, R., and Jhanwar-Uniyal, M. (2009) Anticancer .jbc .o Res 29, 119-123 rg b/ 51. Yang, L. M., Rinke, R., and Korbmacher, C. (2006) J. Biol. Chem. 281, 9859-9868 y g 52. Slepkov, E. R., Rainey, J. K., Sykes, B. D., and Fliegel, L. (2007) Biochem J 401, 623-633 ue s 53. McLean, L. A., Roscoe, J., Jorgensen, N. K., Gorin, F. A., and Cala, P. M. (2000) Am J Physiol t o n Cell Physiol 278, C676-688 Ja n u 54. Nakamura, K., Kamouchi, M., Kitazono, T., Kuroda, J., Matsuo, R., Hagiwara, N., Ishikawa, E., a ry Ooboshi, H., Ibayashi, S., and Iida, M. (2008) Am J Physiol Heart Circ Physiol 294, H1700-1707 2 4 , 2 0 1 9 ACKNOWLEDGEMENTS We would like to thank Melissa McCarthy for excellent cell culture assistance and Drs. Yancey Gillespie, Edlira Clark, Niren Kapoor and Yawar Qadri for helpful discussions. This study was supported by NIH Grant # DK37206. FIGURE LEGENDS FIGURE 1 . Inhibition of ASIC-1 by PcTX-1 reduces migration and proliferation of D54-MG glioma cells. A. Representative microscopic images of transwell filters, showing migration of D54-MG cells following 24h exposure to PcTX-1, control peptide, or benzamil. Drugs were present on both sides of the filter throughout the 5h migration period. In each experiment, the number of cells on the underside of the filter was counted from five randomly selected fields of each image (n≥3). B. Bar graph showing the average number of migrated cells per field under each condition (normalized to control). C. Scratch/wound assay. Phase contrast image of D54-MG cells untreated or treated with PcTX-1, control 9 peptide, or benzamil at 0h and 24h after scarring. D. Summary bar graph showing results expressed as % wound area covered at 24 h, normalized to control conditions (n≥4). FIGURE 2. Effect of inhibition of ASIC-1 on cell cycle progression and whole-cell current of D54- MG glioma cells. A. Flow cytometric analysis of D54-MG cells treated with PcTX-1, control peptide, and benzamil for 24h. DNA content is shown as 2n and 4n in the X-axis where 2n = cells in G /G phase 0 1 and 4n = cells in G /M phase. The bar graph shows the percentage of cells in each cell cycle phase for the 2 different experimental conditions, n≥4. B. Expression of p21Cip1 and p27Kip1 proteins in D54-MG cells treated with PcTX-1, control peptide or benzamil for 24h. β-actin was used as a loading control. Accompanying bar graph shows the normalized density as compared to 2% FBS control n≥6. C. Bar graph showing average amiloride-sensitive currents at -80mV for D54-MG cells pre-treated with PcTX-1, control peptide, or benzamil for 24h. Currents were recorded in the continued presence of drug in the bath solution n≥6. D. Summary I/V curves showing basal current and that following addition of amiloride to the bath solution. FIGURE 3. Knockdown of ASIC1 expression inhibits amiloride-sensitive current in D54-MG glioma cells. A. Lysates from untransfected wild type D54-MG (D54MG-WT) and D54-MG cells stably transfected with DN-eGFP-ASIC1 cDNA (D54MG-A1DN) were analyzed. The expression of ASIC-1 D o w was significantly decreased in D54MG-A1DN cells compared to wild type cells, n≥6. B. Average n lo amiloride sensitive currents at -80mV, n≥6. C. I/V curves show significant attenuation of amiloride- ad e d sensitive whole-cell current after knocking down ASIC1. fro m h FIGURE 4. Knock-down of ASIC1 inhibits migration and proliferation of D54-MG cells. Summary ttp data showing A migration of untransfected D54MG-WT, D54MG-A1DN and D54-MG-WT cells treated ://w w with benzamil for 24h evaluated by transwell migration as described for Fig. 1 and B, scratch/wound w healing assay at 0h and 24h after scarring n≥4. C. Flow cytometric analysis of D54MG-WT, D54MG- .jbc .o A1DN and D54-MG cells treated with benzamil for 24h; the percentage of cells in each cell cycle phase rg for different experimental conditions is shown n≥3. D. Expression of p21Cip1 (n≥6), and p27kip1 (n≥5), by/ g was determined in D54MG-A1DN cells following re-feeding with 10% FBS after 48h of serum ue s starvation. Actin was used as a loading control. Each bar represents normalized densitometry compared t o n to untransfected D54MG-WT cells. Ja n u a ry FIGURE 5. Knocking down δENaC has no effect CKI expression. A. Lysates from untransfected wild 2 4 type D54-MG and D54MG-δENaC-DN cells were separated by SDS-PAGE and immunoblotting. Each , 20 1 bar represents the normalized density of both bands compared to untransfected D54-MG cells, n≥3. B. 9 No difference in expression of p21Cip1 and p27kip1 proteins was detected between wild type D54-MG and D54MG-δENaC-DN cells. Wild type cells treated with were used as a positive control (n≥6). FIGURE 6. Effect of low sodium on cell cycle progression of D54-MG glioma cells. A. FACS analysis of D54-MG cells incubated in either control Krebs buffer (in mM: 118 NaCl, 4.7 KCl, 1.2 MgSO , 1.2 4 KH PO , 1.2 CaCl , 10 glucose, 25 NaHCO , pH 7.4) or low Na+ Krebs buffer with equimolar NMDG-Cl 2 4 2 3 substituted for NaCl for 24h. The bar graph represents the percentage of number of cells in each cell cycle phase for different experimental conditions. Data are representative of six independent experiments. B. Expression of p21Cip1 (n≥5), and p27kip1 (n≥4), was determined in wild type D54MG cells incubated in either control Krebs buffer or low Na+ Krebs buffer for 24h after 48h of serum starvation. Actin was used as a loading control. Each bar represents normalized densitometry compared to untransfected D54MG- WT cells. FIGURE 7. Inhibition of ERK1/2 phosphorylation. A. Immunoblot analysis of lysates from D54-MG cells treated with PcTX-1, control peptide, or benzamil for 24h. The blots were probed for phospho- ERK1/2 and then stripped and reprobed for total ERK1/2, n≥3. B. Immunoblot analysis of lysates from 10
Description: