Eag1 and Bestrophin 1 Are Up-regulated in Fast-growing Colonic Cancer Cells*

Ion channels like voltage-gated ether-á-go-go (Eag1) K+ channels or Ca2+-activated Cl– channels have been shown to support cell proliferation. Bestrophin 1 (Best1) has been proposed to form Ca2+-activated Cl– channels in epithelial cells. Here we show that original T84 colonic carcinoma cells grow slowly (T84-slow) and express low amounts of Eag1 and Best1, whereas spontaneously transformed T84 cells grow fast (T84-fast) and express high levels of both proteins. Both Eag1 and Best1 currents are up-regulated in T84-fast cells. Eag1 currents were cell cycle-dependent with up-regulation during G1/S transition. T84-slow, but not T84-fast, cells formed tight monolayers when grown on permeable supports. RNA interference inhibition of Eag1 and Best1 reduced proliferation of T84-fast cells, whereas overexpression of Best1 turned T84-slow into fast-growing cells. Eag1 and Best1 improve intracellular Ca2+ signaling and cell volume regulation. These results establish a novel role for bestrophins in cell proliferation.

Ion channels like voltage-gated ether-á-go-go (Eag1) K ؉ channels or Ca 2؉ -activated Cl ؊ channels have been shown to support cell proliferation. Bestrophin 1 (Best1) has been proposed to form Ca 2؉ -activated Cl ؊ channels in epithelial cells.

Here we show that original T 84 colonic carcinoma cells grow slowly (T 84 -slow) and express low amounts of Eag1 and Best1, whereas spontaneously transformed T 84 cells grow fast (T 84fast) and express high levels of both proteins. Both Eag1 and Best1 currents are up-regulated in T 84 -fast cells. Eag1 currents were cell cycle-dependent with up-regulation during G 1 /S transition. T 84 -slow, but not T 84 -fast, cells formed tight monolayers when grown on permeable supports. RNA interference inhibition of Eag1 and Best1 reduced proliferation of T 84 -fast cells, whereas overexpression of Best1 turned T 84 -slow into fastgrowing cells. Eag1 and Best1 improve intracellular Ca 2؉ signaling and cell volume regulation. These results establish a novel role for bestrophins in cell proliferation.
An increasing number of studies demonstrate proliferative effects of membrane ion channels (1). Voltage-gated K ϩ channels and other types of K ϩ channels are expressed in numerous types of tumors, where they may serve as diagnostic and prognostic markers and potential drug targets (2)(3)(4). Eag1 channels are probably necessary for progression through the G 1 phase and G 0 /G 1 transition of the cell cycle (5). A recent study demonstrates the hyperpolarizing effects of Eag1 and other Kv channels on the membrane voltage of T 84 cells, which supports intracellular pH regulation and Ca 2ϩ increase necessary for proliferation (6).
Much less is known about the role of Ca 2ϩ -activated Cl Ϫ channels in cell proliferation (7). This may be due to the ongoing controversy regarding the molecular nature of Ca 2ϩactivated Cl Ϫ channels (8). A family of putative Ca 2ϩ -activated Cl Ϫ channels has been identified that also controls cell-cell adhesion, apoptosis, and the cell cycle. However, the structure and biophysical properties of these channels are poorly understood (9,10). Recent studies defined bestrophin proteins as bona fide Ca 2ϩ -activated Cl Ϫ channels. The Cl Ϫ currents gen-erated upon expression of bestrophin show many of the properties found for native Ca 2ϩ -activated Cl Ϫ currents (11)(12)(13). However, bestrophins have also been proposed to function as regulators of voltage-gated L-type Ca 2ϩ channels (14). Our own ongoing work in epithelial cells supports both concepts, in that bestrophins may form part of a Cl Ϫ channel complex or may couple intracellular Ca 2ϩ signals to Cl Ϫ channels of unknown molecular identity (15).
In the present report we demonstrate that both voltage-gated Eag1 K ϩ channels and bestrophin 1 (Best1) Cl Ϫ channels support proliferation of fast-growing T 84 colonic carcinoma cells. The fast-growing T 84 cell clone was obtained through spontaneous transformation of slow-growing T 84 cells. In contrast to slow-growing T 84 cells, transformed cells do not form polarized monolayers and show a remarkable up-regulation of Eag1 and Best1 expression. We demonstrate that both currents are in charge of enhanced cell proliferation.

MATERIALS AND METHODS
Cell Culture and Proliferation Studies-Human colorectal carcinoma epithelial T 84 cells (ATCC, Manassas, VA) were grown in Dulbecco's modified Eagle's medium/Ham's F-12 medium (1:1) supplemented with 10% fetal bovine serum, 2 mM glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin (Invitrogen, Karlsruhe, Germany) at 5% CO 2 and 37°C. Cells were seeded on fibronectin (Invitrogen)/collagen (Cellon)-coated glass coverslips or permeable supports (Snapwell, Costar). Typically these cells grow slowly as polarized monolayers (T 84 -slow). Because of spontaneous transformation, a T 84 cell line was selected that grew remarkably faster (T 84 -fast). For proliferation assays, cells were plated at a density of 2000 cells/ 0.35 cm 2 and incubated 2 days later with either niflumic acid (0.01-100 mM) or astemizole (0.5-5000 nM). Cell proliferation was assessed by 5-bromo-2Ј-deoxyuridine (BrdUrd) 3 incorporation using an enzyme-linked immunosorbent assay kit (Roche Diagnostics, Penzberg, Germany) and cell counting. The cell number was assessed after fixation in 3.7% formaldehyde and 0.5% Triton X-100 for 30 min at room temperature and after staining with Mayer's hemalaun (Merck, Darmstadt, Germany) for 5 min. Digitized microscopic images were taken (Fluovert FS, Leitz), and nuclei were counted using imaging software (TINA version 2.09g). Toxicity of the blockers was * This work was supported by Deutsche Forschungsgemeinschaft Grant SFB699 A7 and the Else-Krö ner-Fresenius-Stiftung. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 Both authors contributed equally to this work. 2 To whom correspondence should be addressed. assessed using trypan blue (Sigma). Each experiment was performed at least in triplicate. Cell Cycle, FACS Analysis, Caspase Assay, and RT-PCR-Cells were synchronized into early G 1 by 24-h serum starvation. Incubation in thymidine (2 mM) (Sigma) halted the cells at G 1 /S transition. 36-h treatment with demecolcine (0.05 g/ml) (Sigma) synchronized cells into M phase. Synchronization was verified by FACS (COULTER EPICS XL-MCL, Beckman) using propidium iodide staining of the DNA (Sigma). Apoptosis was analyzed after 8-and 24-h incubation with the protein kinase C inhibitor staurosporine (1 M) (Sigma) and with detection of cleaved caspase-3 in Western blots using rabbit antihuman caspase-3 antibody (1:1000) (Cell Signaling Technology, Inc., Danvers, MA). For RT-PCR, total RNA was isolated using NucleoSpin RNA II columns (Macherey-Nagel, Düren, Germany). 1 g of total RNA was reverse-transcribed for 1 h at 37°C using random primer and RT (Moloney murine leukemia virus reverse transcriptase, Promega, Mannheim, Germany). For PCR, the following primers were used: hEag1 (KCNH1, GenBank TM accession number NM_002238), 5Ј-CGCATGA-ACTACCTGAAGACG-3Ј (sense) and 5Ј-TCTGTGGATGG-GGCGATGTTC-3Ј (antisense), 560 bp; and hBest1 (VMD2, GenBank TM accession number NM_004183), 5Ј-CTGCTCTG-CTACTACATCATC-3Ј (sense) and 5Ј-GTGTCCACACTGA-GTACGC-3Ј (antisense), 552 bp. The conditions were 94°C for 2 min and 35 cycles of 94°C for 30 s, 58°C for 30 s, and 72°C for 1 min. PCR products were visualized by loading on 2% agarose gels and verified by sequencing.
Down-regulation of Best1 and Eag1 Expression by RNAi-Three different batches (A, B, and C) of duplexes of 21-nucleotide RNAi with a 3Ј-overhanging TT were purchased from Invitrogen. The sense strands of the RNAi used to silence the Best1 gene were 5Ј-AAUUCCUGUCGACAAUCCAGUU-GGU-3Ј (A), 5Ј-AUCUCAUCCACAGCCAACAGGGACA-3Ј (B), and 5Ј-UAAAUAAAGCGGAUGAUGUAGUAGC-3Ј (C). RNAi sequences for Eag1 are shown in Ref. 6. Fluorophore-labeled RNAi and exposure to the transfection reagent Lipofectamine 2000 (Invitrogen) served as controls. After 48 h, cells were further processed in proliferation assays for Western blotting.
Measurement of the Intracellular Ca 2ϩ Concentration and Cell Volume-T 84 cells were loaded with 2 M Fura-2/AM (Molecular Probes, Eugene, OR) in Opti-MEM I medium (Invitrogen) with 2.5 mM probenecid for 1 h at room temperature. Fluorescence was detected in cells perfused with Ringer's solution containing 2.5 mM probenecid (Sigma) at 37°C using an inverted microscope (IMT-2, Olympus, Nürnberg, Germany) and a high speed polychromator system (VisiChrome, Puchheim, Germany). Fura-2 was excited at 340/380 nm, and emission was recorded between 470 and 550 nm using a chargecoupled device camera (CoolSnap HQ, Visitron). [Ca 2ϩ ] i was calculated from the 340/380 nm fluorescence ratio after background subtraction. The formula used to calculate [Ca 2ϩ where R is the observed fluorescence ratio. The values R max and R min (maximum and minimum ratios) and the constant S f2 /S b2 (fluorescence of free and Ca 2ϩ -bound Fura-2 at 380 nm) were calculated using 2 mol/liter ionomycin (Calbiochem), 5 mol/liter nigericin, 10 mol/liter monensin (Sigma), and 5 mmol/liter EGTA to equilibrate intracellular and extracellular Ca 2ϩ in intact Fura-2-loaded cells. The dissociation constant for the Fura-2⅐Ca 2ϩ complex was taken as 224 nmol/liter.
Patch Clamping-Cell culture dishes were mounted on the stage of an inverted microscope (IM35, Zeiss) and perfused continuously (37°C) with Ringer's solution. Patch clamp experiments were performed in the fast and slow whole cell configuration. Patch pipettes had an input resistance of 2-4 megaohms when filled with a solution containing 30 mM KCl, 95 mM potassium gluconate, 1.2 mM NaH 2 PO 4 , 4.8 mM Na 2 HPO 4 , 1 mM EGTA, 0.758 mM calcium gluconate, 1.034 mM MgCl 2 , 5 mM D-glucose, and 3 mM ATP. pH was adjusted to 7.2; Ca 2ϩ activity was 0.1 M. The access conductance was monitored continuously and was 60 -120 nanosiemens. Currents (voltage clamp) and voltages (current clamp) were recorded using an EPC7 amplifier (HEKA, Darmstadt, Germany), an LH1600 interface and PULSE software (HEKA), and Chart software (ADInstruments, Spechbach, Germany). In intervals, membrane voltages were clamped in steps of 10 mV from Ϫ50 to ϩ50 mV relative to resting potential. Membrane conductance G m was calculated from the measured current (I) and clamped membrane voltage values according to Ohm's law.
Cell volume was measured directly by a Zeiss Axiovert 200M/ApoTome microscope using Axiovision software or was assessed by fluorescence measurements in calcein (2 M; Molecular Probes)-loaded cells at an excitation of 500 nm and emission of 520 -550 nm. The experiments were done in the presence of 2.5 mM probenecid. The control isotonic solution (290 mosM) was prepared by adding 120 mM mannitol. The hypotonic (170 mosM) and control isotonic solutions contained 85 mM NaCl.
Materials and Statistical Analysis-All compounds used were of the highest purity grade available. Astemizole, niflumic acid (NFA), DIDS, tetrapentylammonium, carbachol, and ATP were all from Sigma. Student's t test (for paired or unpaired samples as appropriate) and analysis of variance (ANOVA) were used for statistical analysis. p Ͻ 0.05 was accepted as significant.

Fast-growing Colonic Carcinoma Cells (T 84 -fast) Express
Eag1 and Best1-T 84 cells typically grow in slowly expanding patches (T 84 -slow). At passage number 73, the cell line spontaneously changed its growth pattern, i.e. the cells grew remark-

Eag1 and Best1 Enhance Proliferation in T 84 Cells
ably faster (T 84 -fast) as single cells, non-polarized and above each other (Fig. 1, A and B). T 84 -slow cells formed tight monolayers with a transepithelial resistance of 2.2 Ϯ 0.3 kiloohms ϫ cm 2 (n ϭ 25) when grown on permeable supports, whereas T 84 -fast cells did not (0.12 Ϯ 0.07 kiloohms ϫ cm 2 ; n ϭ 23). Thus, growth pattern or phenotype was not changed by the way the cells were grown. We examined apoptosis in both cell lines by Western blotting and analyzed uncleaved and cleaved caspase-3. Small amounts of cleaved caspase-3 were detected in T 84 -slow, but not in T 84 -fast, cells after treatment with the protein kinase C inhibitor staurosporine (1 M) (Fig. 1C).
Semi-quantitative RT-PCR analysis was performed for voltage-gated K ϩ channels (Eag1) and putative Ca 2ϩ -activated Cl Ϫ channels (Best1) using ␤-actin as an internal standard. Densitometric analysis suggested transcript numbers for Eag1 and Best1 that were at least 10 times higher in T 84 -fast cells. Thus much higher protein expression was found for Eag1 and Best1 in T 84 -fast cells in Western blots (Fig. 1D). The membrane con-ductance properties were analyzed in T 84 -slow and T 84 -fast cells in whole cell patch clamp experiments, and membrane voltages were measured in the current clamp mode. T 84 -slow cells (n ϭ 8) were more hyperpolarized (Ϫ51.7 Ϯ 5.6 mV) and had a lower base-line conductance (3.8 Ϯ 0.7 nanosiemens) than T 84 -fast cells (Ϫ36.6 Ϯ 2.9 mV and 23.8 Ϯ 2.9 nanosiemens, n ϭ 8). The increased base-line conductance and depolarized membrane voltages of T 84 -fast cells were due to higher activity of Cl Ϫ channels because replacement of extracellular Cl Ϫ by gluconate (5Cl) shifted the I/V curve to more positive clamp voltages. This was not observed for T 84slow cells (Fig. 1E).
Eag1 Controls Proliferation of Fast-but Not Slow-growing T 84 Cells-Eag1 has been demonstrated to support proliferation of several different cell types (6,16). We therefore examined if high expression levels of Eag1 correlate with increased proliferation of T 84 -fast cells. To that end, T 84 -slow and T 84 -fast cells were treated with three different batches (A-C) of siRNA for Eag1. Incubation of the cells with either fluorescently labeled scrambled oligos or lipid (transfection reagent) and nontreated cells served as controls. Measurement of BrdUrd incorporation clearly indicates inhibition of proliferation of T 84 -fast cells after treatment with siRNA for Eag1 ( Fig. 2A). siRNA treatment of T 84 -slow cells did not affect cell proliferation, suggesting a proliferative function of Eag1 only in T 84 -fast cells. Notably, Eag1 expression was significantly up-regulated by the readdition of 10% fetal calf serum in serum-starved cells (data not shown). Using the inhibitor astemizole (5 M), we examined the contribution of Eag1 to whole cell currents measured in T 84 -slow and T 84 -fast cells. Astemizole inhibited whole cell currents in both cell types; however, the effect was more pronounced in T 84 -fast cells, and thus astemizole-sensitive whole cell currents were significantly larger in T 84 -fast cells (Fig. 2, B and C). The normalized conductance/voltage relationship for Eag1 (astemizole-sensitive whole cell conductances) was not different for T 84 -fast (solid line) and T 84 -slow (dashed line) cells (Fig. 2D).
A previous report (6) supplied evidence that hyperpolarized Eag1 currents assist in the increase of intracellular Ca 2ϩ ([Ca 2ϩ ] i ) when cells are stimulated with secretagogues (such as ATP and carbachol) or mitogens and may thereby support proliferation. Both Ca 2ϩ release from endoplasmic reticulum stores and Ca 2ϩ influx through store-operated Ca 2ϩ channels depend on Kv channel function as reported earlier (6). We compared the increase in [Ca 2ϩ ] i in Fura-2-loaded T 84 -slow and T 84 -fast cells upon stimulation with ATP (100 M). ATP binds to purinergic P2Y 2 receptors and thereby induces a peak and plateau [Ca 2ϩ ] i increase in both cell lines (Fig. 2E). The peak [Ca 2ϩ ] i increase was doubled in T 84 -fast cells when compared with T 84 -slow cells, which suggests a role of Eag1 in Ca 2ϩ signaling in T 84 -fast cells (Fig. 2, E and F). Enhanced Ca 2ϩ signaling is not due to an increased expression of the abundant purinergic receptor P2Y 2 in T 84 -fast cells. In contrast, P2Y 2 expression was slightly higher in T 84 -slow cells (Fig. 2G).
Eag1 Activity in T 84 -fast Cells Is Cell Cycle-dependent-In other cell types Eag1 activity varies during the cell cycle (16). Therefore we examined cell cycle dependence of Eag1 in T 84fast cells. Cells were synchronized in early G 1 (eG 1 ), G 1 /S, or M phase (see "Materials and Methods"), and synchronization was verified by FACS analysis (Fig. 3A). We found that astemizolesensitive whole cell currents were augmented in cells synchronized in G 1 /S when compared with eG 1 or M phase (Fig. 3B). T 84 cells also express other voltage-gated K ϩ channels such as Kv1.5 and Kv3.4, which are blocked by the inhibitor tetrapen-tylammonium (6). However, in contrast to Eag1, tetrapentylammonium-sensitive whole cell currents were not cell cycledependent (data not shown). We further compared the increase in [Ca 2ϩ ] i in T 84 -fast cells synchronized in eG 1 and G 1 /S phase. Both peak and plateau [Ca 2ϩ ] i increases were significantly enhanced in G 1 /S cells. Moreover, the Eag1 blocker astemizole inhibited peak and plateau [Ca 2ϩ ] i increases in both cell cycle phases, but the effect of astemizole on plateau [Ca 2ϩ ] i was augmented in the G 1 /S phase (Fig. 3C). These results supply evidence for cell cycle-regulated Eag1 currents in T 84 -fast cells and cell cycle-dependent effects of Eag1 on Ca 2ϩ signaling.
Best1 Controls Proliferation of Fast-but Not Slow-growing T 84 Cells-Because T 84 -fast cells show much higher expression of Best1 when compared with T 84 -slow cells (Fig. 4A), we examined the effect of this putative Ca 2ϩ -activated Cl Ϫ channel on cell proliferation. T 84 -slow and T 84 -fast cells were treated with three different batches (A-C) of siRNA for Best1, which reduced Best1 levels in T 84 -fast cells. Expression levels in T 84 -   n ϭ 14), G 1 /S (n ϭ 9), and M (n ϭ 9) phase. G Aste was significant during all cell cycle phases but was enhanced during G 1 /S (ANOVA). nS, nanosiemens. C, upper panels, summary of the peak and plateau [Ca 2ϩ ] i increases induced by carbachol in T 84 -fast cells synchronized into eG 1 and the effects of 0.5 M astemizole (Aste; n ϭ 73); lower panels, summary of the peak and plateau [Ca 2ϩ ] i increases induced by carbachol in T 84 -fast cells synchronized into G 1 /S and the effects of 0.5 M astemizole (n ϭ 64). The effect of astemizole was significant in all series (paired t test) and was enhanced for the plateau [Ca 2ϩ ] i increase in G 1 /S when compared with eG 1 (unpaired t test). con, control. *, significant difference when compared with control; #, significant difference when compared with eG 1 . slow cells were already very low, so more protein was loaded on the gel as indicated by the ␤-actin signal (Fig. 4, A and B). Incubation of the cells with either fluorescently labeled scrambled oligos or transfection reagent (lipid) and non-treated cells served as controls. Measurement of BrdUrd incorporation clearly indicates inhibition of proliferation of T 84 -fast cells after treatment with siRNA for Best1 (Fig. 4C). No effects of siRNA were seen in T 84 -slow cells, suggesting a proliferative function of Best1 only in T 84 -fast cells.
In whole cell patch clamp experiments, the high base-line conductance found in T 84 -fast cells could be partially inhibited by the blockers of Ca 2ϩ -activated Cl Ϫ channels, NFA and DIDS (both 100 M). Neither inhibitor had an effect in T 84 -slow cells (Fig. 5A). Stimulation with ATP (100 M) to increase intracellular Ca 2ϩ (Fig. 2D) activated a whole cell current in T 84 -slow, but not in T 84 -fast, cells (Fig. 5B). In T 84 -slow cells, ATP activated primarily a K ϩ conductance as indicated by the hyperpolarizing effect of ATP on the membrane voltage (Fig. 5C). Few effects of ATP were seen in T 84 -fast cells. Correspondingly, replacement of extracellular Cl Ϫ by impermeable gluconate (5Cl) showed no effects in T 84 -slow cells but reduced the baseline conductance in T 84 -fast cells (Fig. 5D). Taken together, these results suggest active Best1 Cl Ϫ channels in non-stimulated T 84 -fast cells, which cause enhanced proliferation.
Increased Proliferation and Cl Ϫ Conductance in Best1-transfected T 84 -slow Cells-To further demonstrate that Best1 contributes to proliferation of T 84 cells, we expressed human Best1 in T 84 -slow cells. As shown in Fig. 6A, transfection of 100 ng of exogenous Best1 increased Best1 expression in T 84 -slow cells almost to the level found in T 84 -slow cells. No change was seen in mock-transfected cells. Notably, overexpression of Best1 changed the growth pattern of T 84 -slow cells toward that found for T 84 -fast cells (Fig. 6B). Moreover, after expression of Best1, a DIDS-sensitive whole cell current appeared in T 84 -slow cells   8 -11). *, significant effects of NFA and DIDS (paired t test). S, microsiemens. B, whole cell currents measured in T 84 -slow and T 84 -fast cells. Stimulation of the cells with ATP (100 M) activated a whole cell conductance only in T 84 -slow, but not T 84 -fast, cells. con, control. C, summaries for the whole cell conductances (upper panels) and membrane voltages (lower panels) measured in T 84 -slow (n ϭ 13) and T 84 -fast (n ϭ 14) cells and the effects of ATP (100 M). *, significant effects on whole cell conductance and membrane voltages (paired t test). D, summaries for the whole cell conductances measured in T 84 -slow (n ϭ 9) and T 84 -fast (n ϭ 8) cells and the effect of replacement of extracellular Cl Ϫ by gluconate (5Cl) and ATP (100 M). *, significant activation of conductance by ATP in T 84 -slow cells and significant inhibition by gluconate (5Cl) (paired t test). nS, nanosiemens.
that was not found in mock-transfected or parental cells (Fig.  6C). Measurement of the BrdUrd incorporation indicated a significant increase in proliferation of Best1-transfected T 84 -slow cells that was not observed for mock-transfected cells (Fig. 6D). In summary, both Eag1 K ϩ channels and Best1 Cl Ϫ channels are up-regulated in spontaneously transformed T 84 cells, where they augment proliferation.
Eag1 and Best1 Support Intracellular Ca 2ϩ Signaling and Volume Regulation, Respectively-Eag1 K ϩ channels (and to some degree Best1) have a clear impact on intracellular Ca 2ϩ signaling. This was demonstrated by stimulation with ATP (100 M), which increased intracellular peak (endoplasmic reticulum Ca 2ϩ release) and plateau (Ca 2ϩ influx through store-operated Ca 2ϩ channels) increases in T 84 -fast cells. Treatment with siRNA for Eag1 significantly reduced peak and plateau Ca 2ϩ increases (Fig. 7A). Volume measurements indicate stronger cell swelling and potent regulatory volume decrease in T 84 -fast when compared with T 84 -slow cells, as identified by direct volume measurements and indirect measurements of calcein fluorescence (Fig. 7, B-E). T 84 -fast cells treated with Best1 siRNA show reduced regulatory volume decrease (Fig.  7B). As both intracellular Ca 2ϩ and cell volume control are essential for the mitotic cell cycle, these results may provide a mechanism for the proliferative role of Eag1 and Best1.

Transformation of Colonic Epithelial Cells Increases Proliferation and Ion
Conductances-T 84 colonic carcinoma cells are well established (17,18). They grow slowly in patches and form tight and polarized monolayers when cultured on permeable supports. T 84 cells were used for numerous electrophysiological studies and resemble a model for electrolyte transport in the colonic epithelium (17). They show many aspects of native epithelial cells such as relatively hyperpolarized membrane voltage and Cl Ϫ secretion by cAMP-dependent cystic fibrosis transmembrane conductance regulator channels. When stimulated by secretagogues such as ATP and carbachol that increase the intracellular Ca 2ϩ concentration, predominantly K ϩ channels are activated, hyperpolarizing the membrane voltage. This is comparable with colonic crypt cells, which also activate Ca 2ϩactivated K ϩ channels upon stimulation of basolateral muscarinic M3 and apical purinergic P2Y 2 receptors (18,19). In contrast, T 84 cells that had undergone spontaneous transformation were remarkably different: they proliferated much faster, were unable to polarize or form tight monolayers when grown on permeable supports, showed typical features of malignancy, and had different membrane conductances.
Anderson and Welsh (20) as well as Morris and Frizzell (21) reported earlier that expression of ion channels in epithelial cells is significantly affected by the underlying substrate. Thus Ca 2ϩ -activated Cl Ϫ channels were prominent in non-differentiated HT 29 colonic carcinoma cells but disappeared with polarization (22). Although T 84 -fast cells did not polarize when grown on permeable supports, a change in Best1 expression upon differentiation of filter-grown HT 29 cells may explain earlier results (22).
A change in membrane ion conductance has also been reported for spontaneously transformed Madin-Darby canine kidney cells. These fast-growing cells express Ca 2ϩ -activated K ϩ channels that largely enhance cell migration (23). The fastgrowing T 84 cells described in the present study showed a strong increase in the expression of the Best1 Cl Ϫ channel along with Eag1 potassium channels. Eag1 currents were detected despite a relatively high (100 nM) Ca 2ϩ concentration in resting cells. However, up to 1 M [Ca 2ϩ ] i is required for complete inhibition of Eag1, and thus a fraction (ϳ40%) of the Eag1 conductance is still detectable in T 84 -fast cells (24). Notably, the Cl Ϫ channels were already active in non-stimulated cells, i.e. without purinergic stimulation. Additional increase of intracellular Ca 2ϩ by ATP or carbachol did not further increase Cl Ϫ conductance. A similar result was obtained when Best1 was overexpressed in T 84 -slow cells and has also been observed when Best1 was expressed in HEK293 cells (8). This suggests that additional components may be required for Ca 2ϩ regulation of Best1 currents, which may not be present in cancer cells. The presence of these DIDS-sensitive Cl Ϫ currents along with astemizole-inhibited Eag1 K ϩ currents clearly augmented cell proliferation.
Ion Channels Induce Malignancy and Metastatic Cell Growth-Clonal selection of fast proliferating T 84 cells (T84SF) has been reported previously (25). These cells demonstrate invasive and metastatic cell growth when transferred into nude mice. Basal tyrosine phosphorylation and expression of Src kinase were enhanced in these T84SF cells (25). Along this line, the Src inhibitor PP2 reduced cell growth, invasion, and cell adhesion of T84SF cells (26). In Jurkat T lymphocytes, Src kinase controls voltage-dependent Kv1.3 K ϩ channels, which also affect cell proliferation (27). Src kinase may also be up-regulated in fastgrowing T 84 cells and may be responsible for changes in membrane conductance. It will be interesting to examine in subsequent studies the impact of Eag1 and Best1 on cell migration and tissue invasion since we found previously that genomic amplification of Eag1 in human colorectal carcinoma is an independent marker of adverse prognosis (28).
How Do Ion Channels Determine Malignancy?-K ϩ and Cl Ϫ channels are essential for cell migration and metastasis of cancer (29). It has been shown that enhanced intracellular Ca 2ϩ activates Kv channels during intestinal wound healing (30). Cell migration and formation of tumor metastasis are due to fluctuations in the activity of membrane transporters and ion channels because they cause localized cell swelling and shrinkage (29). These changes in cell volume appear to be a prerequisite for cell migration and malignant invasion. There are only a few studies that have investigated the role of Cl Ϫ channels in cell migration. Cl Ϫ channels are probably necessary for cell movement because K ϩ transport needs to be accompanied by a counterion (23). Notably, bestrophin has been shown to operate as a volume-sensitive Cl Ϫ channel (31).
A Novel Function of Bestrophin for Cell Proliferation?-Cl Ϫ currents induced by expression of bestrophins share many of the properties attributed to Ca 2ϩ -activated Cl Ϫ channels, such as anion selectivity of I Ϫ Ͼ Cl Ϫ and inhibition by NFA and DIDS (11)(12)(13)15). However, there is an ongoing controversy regarding whether bestrophins are actually channel-forming proteins or, rather, regulators of ion channels. Moreover, other proteins have also been proposed as molecular candidates for the Ca 2ϩ -activated Cl Ϫ channel (reviewed in Ref. 8). The present results would support the role of Best1 as a Cl Ϫ channel. As known from previous studies, Ca 2ϩ and volume-regulated Cl Ϫ channels support cell proliferation, and bestrophin is activated by an increase in intracellular Ca 2ϩ as well as cell swelling (7,32). In an ongoing study with M1-collecting duct cells expressing high levels of Best1, we found inhibition of proliferation by the Cl Ϫ channel blockers DIDS and NFA. 4 Thus bestrophins may provide the molecular basis for an understanding of the role of Ca 2ϩ and volume-regulated Cl Ϫ channels in cell proliferation.