Functional up-regulation of HERG K+ channels in neoplastic hematopoietic cells.

Kv1.3 channels regulate proliferation of normal lymphocytes, but the role of voltage-gated potassium channels in transformed hematopoietic cells is not known. We examined transcripts for Kv1.3, h-erg, h-eag, and BEC1 genes in primary lymphocytes and leukemias and in several hematopoietic cell lines. Surprisingly, BEC1, formerly thought to be brain-specific, was present in all the primary leukemias examined, in resting peripheral blood lymphocytes, and in proliferating activated tonsillar cells, lymphocytes from Sjögren's patients, and Epstein-Barr virus-transformed B-cells. Only h-erg mRNA was up-regulated in the cancer cells, but this was not due to proliferation per se, because it was not elevated in any of the proliferating noncancerous lymphocyte types examined. Nor did h-erg transcript levels correlate with the B-cell subset, because it was elevated in immature neoplastic B-CLL cells (CD5(+)) and in a CD5(-) Burkitt's lymphoma cell line (Raji) but not in Sjögren's syndrome cells (enriched in CD5(+) B-cells) or Epstein-Barr virus-transformed B-cells, which are mature CD5(-) B-cells. The protein and whole cell current levels roughly corresponded with the amount of mRNA expressed in three hematopoietic cell lines: CEM (an acute lymphoblastic leukemic line), K562 (a chronic myelogenous leukemic line), and U937 (an acute promyelocytic leukemic line). The selective HERG channel blocker, E-4031, reduced proliferation of CEM, U937, and K562 cells, and this appears to be the first direct evidence of a functional role for the HERG current in cancer cells. Selective up-regulation of h-erg appears to occur in neoplastic hematopoietic cells, thus providing a marker and potential therapeutic target.

properties of Kv channels and even the types expressed have been linked to several cardiac and neurological diseases (4,5). In immune cells, Kv channels play roles in proliferation (6,7), cytotoxicity (8), volume regulation (9,10), and the respiratory burst (11); however, with a few notable exceptions, the particular Kv channel has not been identified. There are many channels in the Kv superfamily, which is divided into Shaker (Kv1-Kv4) and ether-a-go-go (eag) families. In normal T lymphocytes, Kv1.3 channels have been clearly demonstrated as important for proliferation and volume regulation (for reviews, see Refs. 7, 9, and 12). Because pharmacological blockade of Kv1.3 dramatically inhibits activation of naive T-cells, this channel has become a therapeutic target for controlling T-cell-mediated inflammation (13,14). The commonly proposed role of Kv channels in proliferation is to maintain the driving force for calcium influx, thereby affecting calcium-dependent cell cycle control proteins.
In contrast to normal lymphocytes, the role of Kv channels in the proliferation of transformed immune cells (i.e. leukemias and lymphomas) is not known. For instance, in the human myeloid leukemia cell line, ML-1, the nonselective Kv channel blocker 4-aminopyridine reduced proliferation and induced differentiation (15)(16)(17). The specific Kv channel or mechanism of involvement was not identified. In direct contrast to these findings, it was proposed that the absence of a delayed rectifier (Kv) current contributed to the malignant state of megakaryocytes from myelogenous leukemia patients and in a human erythroleukemia cell line (18).
Aberrant expression of Kv channels has been detected in some cancers of nonhematopoietic origin. The most compelling evidence for a role in cell transformation, and thus a therapeutic potential, exists for the eag family. eag transcripts and protein are present in numerous primary human cancers, such as endometroid adenocarcinomas, and in several human and murine cell lines (19 -21). Evidence for their oncogenic potential is that overexpression of eag in CHO or NIH3T3 cells induced several features characteristic of malignant transformation (faster growth, loss of growth factor and substrate dependence, and loss of contact inhibition), and antisense reduction of eag mRNA expression decreased the proliferation of several tumor cell lines (20).
Another member of the eag channel family, the human eagrelated gene (h-erg), appears to be aberrantly expressed in some cancer cells of epithelial, muscle, and neuronal origin (19,21,22) and in the human preosteoclastic cell line, FLG 29.1 (19,23,24), but its role in oncogenesis is unknown. There is broad interest in HERG in normal cells because its naturally occurring mutations in cardiac muscle cause type II heart arrhythmias (25)(26)(27). More recently, h-erg has been identified in some neuronal cells and cell lines, where it may contribute to the resting membrane potential (28,29), spike frequency adaptation (30), seizure-like activity (31), and hormone secretion (32)(33)(34)(35).
In the present study we examined expression of Kv1.3 and two eag family members, h-erg and the recently cloned BEC1, in a broad range of human hematopoietic cell lines, in cells from several chronic lymphocytic leukemia patients, and in peripheral blood lymphocytes from several normal individuals. h-erg was expressed in all of the hematopoietic cell types examined, except Epstein-Barr virus (EBV)-transformed B-cells. h-erg expression (but not BEC1 or Kv1.3) was consistently higher in all the cancer cells than in the noncancer cells. We detected macroscopic HERG currents in three human leukemic cell lines: the T-cell lymphoblastic leukemia, CEM, the chronic myelogenous leukemia, K562, and the histiocytic leukemia, U937, and the average current amplitude correlated with the amount of protein detected by Western blot analysis. Using the specific HERG blocker E-4031, we show that the HERG current plays a role in proliferation of these leukemic cell lines.

EXPERIMENTAL PROCEDURES
Cells-Normal peripheral blood lymphocytes (PBL) and EBV-transformed B-lymphocytes were generated in the F. Tsui laboratory. Primary cells from patients with Sjögren's syndrome and B-cell chronic lymphocytic leukemia (B-CLL) were obtained from Dr. A. Bookman (Toronto Western Hospital) and Dr. J. Scott (Toronto General Hospital). Permanent human cell lines representing Burkitt's lymphoma (Raji), chronic myelogenous leukemia (K562), acute lymphoblastic leukemia (CEM), and acute promyelocytic leukemia (U937 and HL-60) were obtained from ATCC (Manassas, VA). Primary tonsillar lymphocytes were activated by a 72-h incubation in 1% (w/v) of the mitogen, phytohemagglutinin (Sigma). All cells were cultured in endotoxin-free ␣-minimal essential medium that was supplemented with 26 mM NaHCO 3 , 100 mg/l streptomycin, and 10% fetal bovine serum. The cultures were grown in a humidified incubator at 37°C, 5% CO 2 . The cells were counted each day with a hemacytometer using 0.4% trypan blue and then passaged every 2 days to maintain them in the exponential growth phase (i.e. maintained at 1-2 ϫ 10 5 cells/ml). All of the cell culture reagents were purchased from Invitrogen.
Reverse Transcriptase-Polymerase Chain Reaction-For all primary cells and cell lines the total mRNA was extracted using the TRIzol method (Invitrogen). Total mRNA (0.5 g) was reverse transcribed at 37°C for 30 min in 1ϫ RT buffer (50 mM KCl, 10 mM Tris, pH 8.3, 1.5 mM MgCl 2 , 0.01% gelatin) with 1 mM dNTP, 5 mM MgCl 2 , 1.5 units/l RNasin, 100 g/ml oligo(dT), and 50 units SuperScript II reverse transcriptase. After first-strand cDNA synthesis, the PCR amplifications were conducted using the GeneAmp PCR 9600 system (PerkinElmer Life Sciences) with the following primers. For h-erg the forward primer was 5Ј-CAGCGGCTGTACTCGGGCACAG-3Ј, and the reverse primer was 5Ј-CAGAAGTGGTCGGAGAACTC-3Ј (product size, 567 bp). For Kv1.3, the forward primer was 5Ј-TCGAGACGCAGCTGAAGAC-3Ј, and the reverse primer was 5Ј-GGTACTCGAAGAGCAGCCAC-3Ј (product size, ϳ350 bp). For h-eag, the forward primer was 5Ј-CGCATGAAC-TACCTGAAGACG-3Ј, and the reverse primer was 5Ј-TCTGTGGAT-GGGGCGATGTTC-3Ј (data not shown). For BEC1, the forward primer was 5Ј-CACTGTCTCTGGCCCTGCG-3Ј, and the reverse primer was 5Ј-TGTCCACCTCTGCAGAGCCT-3Ј (product size, 366 bp). The PCR mixture was incubated for 1 min at 95°C, and then 0.5 units of TaqDNA polymerase (PerkinElmer Life Sciences) was added. The mixture was subjected to 30 cycles of 95°C for 1 min, a 15-s denaturing phase at 94°C, a 15-s annealing phase (at 62°C for h-erg, 60°C for Kv1.3, and 64°C for BEC1), and a 15-s extension phase at 72°C, with a final extension phase of 6 min at 72°C. The resulting DNA products were resolved in 2% agarose gels labeled with 0.5 mg/ml ethidium bromide, and their fluorescence intensities were measured with a fluorescence reader (Bio-Rad Fluor-S MultiImager). The identities of the RT-PCR-amplified products were confirmed by sequencing (ACGT, Toronto, Canada).
Western Analysis-The cells were lysed in RIPA buffer containing 50 mM Tris, pH 8.0, 1% Nonidet P-40, 150 mM NaCl, 1 mM EDTA, 2 mM Na 3 V0 4 , 1 mM NaF, and the protease inhibitors phenylmethylsulfonyl fluoride (1 mM), aprotinin (1 g/ml), and leupeptin (1 g/ml), and one protease inhibitor mixture tablet (Roche Molecular Biochemicals). The protein content of the supernatant was measured with the Bradford protein assay (Bio-Rad). Aliquots of 20 -50 g of each lysate were boiled for 3 min in 3ϫ sample buffer (Bio-Rad) and then separated by 10% SDS-PAGE. The proteins were transferred to a nitrocellulose membrane, blocked with 5% nonfat milk in TBS containing 0.1% Tween 20 (TBST), and then incubated with anti-HERG antibody (Alomone Laboratories, Jerusalem, Israel; APC-062, residues 1106 -1159 of HERG) overnight at 4°C. After two washes with TBST and two washes with TBS, the membranes were incubated (for 1 h at room temperature) with horseradish peroxidase-conjugated secondary antibody (1:3000; Cedarlane Labs, Hornby, Canada). Following another two washes with TBST and two washes with TBS, labeled proteins were visualized using ECL on high performance chemiluminescence film (Hyperfilm TM ; Amersham Biosciences).
Electrophysiology-Whole cell patch-clamp recordings were obtained from CEM, K562, U937, Raji, and HL-60 cells using an Axopatch 200 or 200A amplifier (Axon Instruments, Union City, CA), using 5-kHz filtering and on-line compensation for series resistance and capacitance. All of the signals were analyzed using pCLAMP 6.0 or 8.0 software (Axon Instruments). The data were stored on a computer and displayed using the Origin program (version 6.1; Origin Labs, Northampton, MA). Pipettes with resistances of 3-5 M⍀ were made from thin walled borosilicate glass capillaries (WPI, Sarasota). While HERG currents were being recorded, the cells were superfused with an extracellular solution containing 130 mM potassium aspartate, 1 mM CaCl 2 , 1 mM MgCl 2 , 5 mM D-glucose, 10 mM HEPES, pH 7.4, adjusted to 300 mOsm with sucrose. The pipette solution contained 130 mM potassium aspartate, 2 mM CaCl 2 , 1 mM MgCl 2 , 10 mM EGTA, 2 mM K 2 ATP, and 10 mM HEPES, titrated with KOH to pH 7.2, and adjusted to 290 mOsm with sucrose. Aspartate was used as the major anion to reduce potential contamination by Cl Ϫ currents. All of the recordings were made at 20 -23°C. At normal [K ϩ ] o outward HERG K ϩ currents are difficult to detect because of rapid inactivation of the channels at depolarized potentials (36 -40). However, the currents can be amplified and inward currents can be measured at hyperpolarized potentials by using high extracellular K ϩ concentrations; we used 130 mM (Nernst potential, ϳ0 mV), and hyperpolarizing steps from a holding potential of ϩ20 mV. Stock solutions of CsCl (50 mM) and BaCl 2 (1 mM) were made in distilled water and kept at room temperature. The selective HERG channel blocker, E-4031 (Alomone Laboratories) (39, 73) was prepared as a 1 mM stock solution in distilled water, stored at Ϫ20°C, and then diluted in bath solution to the final concentration.
Cell Proliferation-The CyQUANT cell proliferation assay kit (Molecular Probes, Eugene, OR) (41,42) was used in conjunction with an HTS7000 BioAssay Reader (PerkinElmer Life Sciences) to assess the effect of the HERG channel blocker, E-4031, on the proliferation of the leukemic cell lines, HL-60, CEM, K562, and U937. This nucleic acidbased fluorescent assay for determining the number of cells exhibits a linear response between about 50 and 50,000 cells/well (42), which was confirmed in our hands. For K562 and U937 cells, 5,000 cells were loaded into each well of 96-well plates (Sarstedt, Newton, NC). For the nonadherent HL-60 and CEM cells, the plates were coated with 5 g/ml poly-L-lysine (Sigma), and 10,000 cells/well were used. After each cell type was incubated with or without 1 M E-4031 for 48 h, the fluorescence intensity measured with the CyQUANT assay was compared with intensities from 5000 or 10000 cells (the starting number of cells) of the same type. An increase in fluorescence signal in any well (compared with the starting number of cells) indicates an increase in cell numbers; i.e. proliferation. The fluorescence intensities, with or without E-4031, were compared to assess the effect of blocking HERG on the normal proliferation of each cell line. Our data from the CyQUANT assay are expressed as means Ϯ S.E., and a paired Student's t test was used to assess the ability of E-4031 to suppress proliferation, with p Ͻ 0.05 taken as statistically significant.

RESULTS
Expression of h-erg mRNA and Protein-RT-PCR was used to detect transcripts for HERG, Kv1.3, and BEC1 channels (h-erg, Kv1.3, BEC1), and ␤-actin in cells from leukemia patients, several leukemic cell lines, and normal human lymphocytes (Fig. 1). To compare expression between different cell types (Table I), the fluorescence intensity of each ethidium bromide-labeled product was first normalized to the ␤-actin signal from the same sample. Then the mRNA expression in the primary leukemic cells and leukemic cell lines was compared with expression in normal (i.e. unstimulated) mature PBL. All nine patients with B-cell chronic lymphocytic leukemia had substantially higher h-erg transcript levels (9 -15-fold higher) than normal PBL. All the hematopoietic cells lines (except HL-60 and Ramos) had elevated h-erg transcript levels: 6-fold higher in the two Burkitt's lymphoma cell lines (Raji, BL2), 10-fold higher in the pro-myelocytic leukemia, U937, 18-fold higher in the pro-B-cell acute lymphoblastic leukemia, CEM, and 19-fold higher in the human chronic myelogenous leukemia, K562.
Three cell types were used to assess whether h-erg up-regulation simply corresponds with an activated, proliferative state. h-erg mRNA was not up-regulated in activated primary tonsillar cells, which are mostly proliferating T lymphocytes, or in highly proliferating EBV-transformed B-cells (CD5 Ϫ mature B-cells) where h-erg transcripts were nearly undetectable. Finally, the h-erg level was not increased in PBL from Sjögren's syndrome patients (n ϭ 3) compared with normal PBL. Sjögren's syndrome is classified as a disorder that lies between hyperplasia and neoplasia, but the highly proliferative mature lymphocytes are not cancerous. h-erg mRNA was 6 -11-fold higher in the immature neoplastic B-CLL cells (95% CD5 ϩ B-cells) than in Sjögren's syndrome cells, which are enriched in CD5 ϩ B-cells (and CD4 ϩ T-cells). These comparisons also rule out the possibility that h-erg up-regulation depends on the B-cell type (CD5 ϩ versus CD5 Ϫ ).
In contrast with up-regulated h-erg expression, Kv1.3 transcripts were not significantly increased in any of the cell types. The primary leukemia cells and activated tonsillar cells had normal levels of Kv1.3 mRNA expression, the levels were lower in the cell lines than in PBL (Table I), and Kv1.3 was not detected in K562 cells (Fig. 1). BEC1 was present in all of the primary leukemias and lymphocytes examined but was not detected in the cell lines, except for CEM and MOLT-4. As a further indication that h-erg was selectively up-regulated among the Kv channels expressed, transcripts for the related channel, h-eag, were not detected in any primary leukemic cells or nontransformed lymphocytes examined and only at low levels in K562 and the acute lymphoblastic leukemia cell line, MOLT-4 (data not shown).
The presence of HERG protein was assessed in CEM, K562, U937, and HL-60 cells using a polyclonal anti-HERG antibody and Western blot analysis. We detected immunoreactive doublets at ϳ130 and 140 kDa, with the highest level in K562 cells, followed by CEM and then U937 cells, and with very low levels in HL-60 cells (Fig. 2). A doublet with similarly sized bands has been previously reported for both endogenous and heterologously expressed HERG (24,(43)(44)(45)(46)(47).
Patch-Clamp Analysis of HERG Current-An important feature of the HERG current is its unusual gating kinetics, that is, the channels are opened by depolarization but inactivate extremely quickly (36 -40, 48). Inactivation is rapidly removed by hyperpolarization, allowing the inward current to flow, and then the channels close (deactivate) slowly. Thus, instead of producing outward potassium currents like other Kv channels, the HERG current is mainly inward. HERG is most easily studied as an inward "tail" current during steps to very negative potentials after a depolarization to open the channels and when using a high external potassium concentration to increase the driving force and current amplitude. We monitored whole cell HERG currents in CEM, K562, and U937 cells by using an extracellular K ϩ concentration of 130 mM (Nernst potential, ϳ0 mV). The channels were activated at the holding potential of ϩ20 mV, and the inward tail current was measured during steps to a variety of negative potentials (Fig. 3). U937 cells also exhibited an outward Kv current when extracellular K ϩ was low and the holding potential was very negative. This current was blocked by the potent Kv1.3 blocker, agitoxin-2, or inactivated by raising the holding potential. The presence of a Kv1.3-like current is consistent with the robust expression of Kv1.3 transcripts (Fig. 1) and an earlier report of a Kv1.3-like current in U937 cells (8). CEM cells also had a small Kv1.3-like current. In all subsequent experiments, we eliminated Kv1.3like currents by holding the membrane potential at ϩ20 mV, a potential at which steady-state inactivation is complete. E-4031 was used to assess the amplitude of the HERG component, i.e. the current remaining in the presence of 1 M E-4031 was subtracted from the total current to yield the E-4031sensitive HERG current. For CEM cells, 10 of 14 cells had measurable HERG currents, and the average peak inward current was Ϫ82 Ϯ 11 pA (at Ϫ140 mV). For the 8 of 13 K562 cells with detectable HERG current, the average value was Ϫ97 Ϯ 18 pA (at Ϫ120 mV). For U937 cells, 6 of 10 cells had measurable HERG currents, with an average amplitude of Ϫ52 Ϯ 14 pA (at Ϫ120 mV). No HERG currents were detected in Raji or HL-60 cells (data not shown).
Because the cell cycle regulates expression of rat ether-àgo-go Kv channels heterologously expressed in Xenopus oocytes FIG. 1. Expression of transcripts for the Kv channels, HERG,  BEC1, and Kv1.3, and for ␤-actin. RT-PCR was conducted on several hematopoietic cell lines, primary cells from leukemia patients, and other sources of peripheral blood lymphocytes, using the primers and conditions given under "Experimental Procedures." For abbreviations see Table I.

FIG. 2. Expression of HERG protein in four human hematopoietic cell lines.
A typical HERG doublet was seen at about 130 and 140 kDa after a polyclonal anti-HERG antibody (Alomone Laboratories) was incubated with protein from cell lysates (50 g protein loaded/lane) and separated by 6.5% SDS-PAGE.
(49), we assessed whether the HERG current amplitude was simply a function of the cell size (constant current density) or whether there were correlations that might indicate cell cycledependent expression, e.g. more channels in enlarged cells preceding cell division. The relative cell size was determined by measuring the cell capacitance, and the current density (peak current/unit capacitance) for each cell was plotted as a function of the cell's capacitance (Fig. 4). For all cells exhibiting HERG current, regardless of the cell type, the total current did not vary with capacitance; thus there were similar absolute numbers of active channels in each cell. In addition, cells with no HERG current were of varied size. Taken together, there was no indication that expression of active HERG channels was dependent on the cell cycle.
Effects on Proliferation of Blocking HERG Channels-The HERG channel blocker, E-4031, was used to assess the role of HERG in proliferation of HL-60, CEM, K562, and U937 cell lines (Fig. 5). Proliferation was measured with a nucleic acidbased assay under two conditions: with the normal extracellular K ϩ concentration ([K ϩ ] o ) of 5 mM and with elevated [K ϩ ] o (30 mM), which was used to approximate conditions in many tumors (see "Discussion"). For each cell line, a standard number of cells per well (either 5,000 or 10,000) were incubated for 48 h with or without 1 M E-4031, and then the average fluorescence level in three replicate wells was assessed, for n separate experiments. In the absence of E-4031, each cell type proliferated well, as seen from the higher fluorescence signal compared with the 5,000 or 10,000 freshly plated cells (Fig. 5A). E-4031, at a concentration that fully blocked the HERG current (Fig. 3) 0.02, n ϭ 4). The number of K562 cells was significantly reduced by E-4031 at normal [K ϩ ] o (p Ͻ 0.007, n ϭ 3) but not at elevated [K ϩ ] o . As a control for the increased osmolarity of the elevated [K ϩ ] o medium, each cell type was grown in medium with 60 mM sucrose added; there was no effect on proliferation (data not shown). The following control experiment shows that E-4031 did not reduce the cell numbers by causing significant cell death. Be-cause the CyQUANT assay requires that the cells be lysed, we measured the fluorescence in wells without adding lysis buffer and determined that the background fluorescence (which would arise from dead and lysed cells) was not significantly elevated after the 48-h growth period. As a further indication that E-4031 was not toxic, proliferation of HL-60 cells was not significantly affected at normal [K ϩ ] o (p Ͼ 0.2, n ϭ 12; Fig. 5) or at elevated [K ϩ ] o (data not shown). The effect of HERG blockade was specific because several other K ϩ channel blockers that do not block HERG channels did not affect proliferation at normal [K ϩ ] o : the Kv1.3 blocker, agitoxin-2 (5 nM); the SK4 blocker, clotrimazole (250 nM); or the inward-rectifier blocker, barium (1 mM) (data not shown). The effects of these blockers were not assessed at elevated [K ϩ ] o . (50), Src (51), Rous sarcoma virus (52), v-Fms (53), or SV40 (54) increased K ϩ channel activity; K ϩ channel up-regulation might be a common feature of cancer cells. However, based on our results, the particular K ϩ channel that is up-regulated may be characteristic of the cancer type. We identified transcripts for four Kv channel types in leukemia cells and hematopoietic cell lines: Kv1.3, BEC1, h-eag, and h-erg. Kv1.3 mRNA was expressed in all of the cells at similar levels to unstimulated lymphocytes except in K562 cells, where it was undetectable. We previously reported that K562 cells do not express Kv1.3 currents (8). The lack of up-regulation of Kv1.3 mRNA (41). In normal lymphocytes Kv1.3 is a target for developing new anti-inflammatory drugs because it plays important roles in T-cell proliferation; thus it is interesting that it was not up-regulated in the cancerous cells. A small Kv1.3-like current was present in U937 and CEM cells, which had low levels of Kv1.3 mRNA, but not in K562 cells, in which transcripts were not detected.

Kv1.3 in stimulated tonsillar cells, EBV-transformed B-cells, or cells from Sjögren's patients is not surprising, because activation of normal peripheral T-cells is only accompanied by a very small increase in
BEC1 is a recently cloned channel gene that is related to eag, and Northern analysis only detected it in human brain, where it is enriched in the telencephalon (55). Surprisingly, we found BEC1 transcripts in all of the primary leukemias examined, as well as in resting PBL, activated tonsillar cells, lymphocytes from Sjögren's patients, and EBV-transformed B-cells. This BEC1 expression in immune cells is a novel finding. The only paper we could find on BEC1 (55) failed to detect transcripts in thymus, spleen, or peripheral blood leukocytes, perhaps because of the sensitivity of Northern analysis, which is lower than that of RT-PCR analysis. For the hematopoietic cell lines, we detected BEC1 only in CEM cells and at very low levels in MOLT-4 cells. We conclude that BEC1 is not a marker of the cancerous state because it was not elevated in cancer cells compared with normal lymphocytes.
Nor was expression of the ether a-go-go gene, h-eag, upregulated; it was detectable at low levels in only two cell lines (K562, . This is particularly notable because the EAG channel has been proposed as a link between oncogenic transformation and K ϩ channel expression. Although high h-eag expression is normally restricted to the adult brain, it has been found in several somatic cancer cell lines (e.g. breast and cervical carcinomas, neuroblastoma). Moreover, proliferation of these cells is inhibited by reducing eag expression (with specific antisense oligonucleotides) (20), and conversely, transfecting eag into mammalian cells conferred a transformed phenotype characterized by faster growth, loss of contact inhibition and growth factor and substrate dependence, and tumor progression when injected into SCID mice. In contrast with Kv1.3, BEC1, and h-eag, we observed dramatic up-regulation of h-erg in all the hematopoietic cell lines examined and, even more importantly, in all the primary leukemias tested. h-erg has also been detected in the human preosteoclastic cell line, FLG 29.1 (19,23,24). Our observation that h-erg was not up-regulated in proliferating noncancerous lymphocytes (activated tonsillar cells, EBV-transformed cells, cells from Sjögren's Syndrome patients), indicates that it is not a marker of proliferation but is selectively up-regulated in leukemic cells. We are not suggesting that HERG is restricted to cancers of hematopoietic origin. It is also up-regulated in a cell line derived from the macrophage-like brain microglia (39,47), is expressed more frequently in human endometrial cancer tissue compared with normal and hyperplastic endometrium (21) and is present in some cancer cells of epithelial, muscle, and neuronal origin (19,21,22). Nor is h-erg expression restricted to cancerous tissue; it is expressed in astrocytes (56) and, most notably, in cardiac muscle, where its mutations cause type II heart arrhythmias (25)(26)(27).
Roles of K ϩ Channels in Proliferation-Specific K ϩ channels are involved in activation and proliferation of normal lymphocytes. Kv1.3 activity is important during the early activation phase of naive T lymphocytes (6,57), whereas in activated T-cells the SK4 Ca 2ϩ /calmodulin-activated K ϩ channel is dramatically up-regulated, and not surprisingly, it then plays a greater role in the restimulation that underlies the secondary immune response (41). Different K ϩ channels may underlie the function of transformed lymphocytes. Activity of an unidenti-fied K ϩ channel regulates some functions of human myeloblastic leukemia ML-1 cells and rat basophilic leukemia cells: the early responses to ␤-integrin-mediated adhesion that lead to cell spreading and the EGF-mediated signal-transduction that is required for initiating cell proliferation (17,24,58). Our findings may be of clinical interest. The widespread expression of h-erg transcripts in primary leukemic cells but not in proliferating nontransformed lymphocytes and the significant reduction in proliferation by blocking HERG channels in several hematopoietic cell lines may indicate a role for HERG in the progression of leukemias. The sensitivity of RT-PCR analysis, along with the relative ease of acquiring peripheral blood lymphocytes, may offer a further means of differentiating leukemias from other lymphoproliferative disorders. Because the HERG blocker did not fully inhibit proliferation, it may be that Kv currents modulate, rather than control, tumor cell proliferation. However, there may also be differences in their roles in primary cancer cells, e.g. even in the compelling study linking aberrant Kv channel expression to oncogenesis, reduction of h-eag expression with an antisense oligonucleotide decreased proliferation by only 17-40% in four tumor cell lines (20).
K ϩ channel expression may also be linked to specific phases of the cell cycle. During progression from the G 1 to S phase, many cells undergo changes in membrane potential, cell volume, cytoplasmic pH, and ion content (59,60), that, in principle, could arise from differing K ϩ channel expression. A requirement for Kv1.3 channel activity was rigorously demonstrated for progression of T lymphocytes through the G 1 FIG. 4. HERG current density as a function of cell size, which was monitored by the whole cell capacitance. Each data point represents an individual cell, with different symbols representing each leukemic cell line. The current density, i.e. peak current (in pA) per unit cell capacitance (in picofarad), was calculated from the E-4031-sensitive current during a voltage-clamp step to Ϫ120 mV (K562, n ϭ 13; U937, n ϭ 10) or to Ϫ140 mV (CEM, n ϭ 14). phase (13,61). Other studies show that K ϩ channel expression correlates with the cell cycle and may play a role in progression through the G 1 phase (59,(62)(63)(64), but the specific channels were not identified. The currents did not resemble HERG and, based on their drug sensitivity, were most likely Kv1 family members. Expression of EAG current may also be cell cycle-dependent because activity of heterologously expressed EAG channels was suppressed during maturation of Xenopus oocytes (49,65). We did not directly assess K ϩ channel expression during the cell cycle, but as an initial indication we examined current density as a function of cell size for three of the hematopoietic cell lines. Presumably the largest cells in each culture were just before cell division. There was no compelling relationship. That is, cells with no detectable HERG current were a wide range of sizes, and in cells expressing current there was a generally inverse correlation between cell size and current density. This would occur if the total number of active channels per cell were similar but distributed over a larger surface area in large cells.
We monitored cell numbers following proliferation, which is one key property of leukemic cells. In the future, it will be important to examine the role of specific K ϩ channels, including HERG, in leukemia cell death because leukemias result from aberrant survival, a balance of proliferation and apoptosis, of hematopoietic cells at different differentiation states. In North America, chronic lymphocytic leukemia is the most common form of adult leukemia, and the cells are about 95% of the B-cell phenotype: mostly auto-reactive, antibody-producing CD5 ϩ mature B-cells (66). Most B-CLL cells overexpress a crucial family of apoptosis regulators, Bcl-2 (67). Bcl-2 family members control B-cell receptor-induced apoptosis (68), and high Bcl-2 expression prevents apoptosis (69). In contrast, Tcell apoptosis is not affected in B-CLL patients, possibly because Bcl-2 proteins do not regulate Fas-induced apoptosis, the major means of reducing T-cells. Other factors must control the transformation of non-B-cell hematopoietic cells. In the myeloblastic ML-1 cell line, serum factors stimulate K ϩ channel activity and proliferation (70); thus, one possibility is that aberrant blood factors affect K ϩ channel activity, reduce apoptosis, and cause other hematopoietic cells to become leukemias or lymphomas. The role of HERG channels in cell death has not yet been investigated; however, there is a precedent for Kv channel involvement in apoptosis. Fas receptor-mediated inhibition of Kv1.3 channels may influence apoptosis in the human leukemic T-cell line Jurkat (71), and two apoptotic proteins, Reaper and Grim, are thought to facilitate apoptosis by stably blocking Kv channels (72). In both studies, the interpretation was that inhibiting K ϩ channels depolarizes the cells and initiates a caspase-dependent apoptotic pathway.
In nontransformed human lymphocytes, K ϩ channels maintain a hyperpolarized membrane potential, which promotes a rise in intracellular Ca 2ϩ and triggers a cascade of events that culminates in interleukin-2 production and proliferation (6,13,61). The membrane potential appears to be correlated with the mitotic activity of many cell types. Terminally differentiated cells in the G 0 phase are usually reported to be very hyperpolarized, quiescent cells that require mitogen stimulation to enter the cell cycle (e.g. lymphocytes) are moderately hyperpolarized, and continuously cycling cells (e.g. tumor cells) that do not enter G 0 phase are very depolarized (59). By moderately depolarizing the resting potential of cancerous cells, the HERG current is expected to affect such cellular processes as calcium entry and volume regulation, which in turn may affect proliferation, differentiation, or apoptosis. An intriguing possibility is that a moderate depolarization in leukemic cells creates a negative feedback by reducing the driving force for calcium entry, and if this reduces Ca 2ϩ -induced cell death or differentiation, it may promote the leukemic state.