Characterization of Voltage-Gated Potassium Channels in Human Neural Progenitor Cells

Background Voltage-gated potassium (Kv) channels are among the earliest ion channels to appear during brain development, suggesting a functional requirement for progenitor cell proliferation and/or differentiation. We tested this hypothesis, using human neural progenitor cells (hNPCs) as a model system. Methodology/Principal Findings In proliferating hNPCs a broad spectrum of Kv channel subtypes was identified using quantitative real-time PCR with a predominant expression of the A-type channel Kv4.2. In whole-cell patch-clamp recordings Kv currents were separated into a large transient component characteristic for fast-inactivating A-type potassium channels (IA) and a small, sustained component produced by delayed-rectifying channels (IK). During differentiation the expression of IA as well as A-type channel transcripts dramatically decreased, while IK producing delayed-rectifiers were upregulated. Both Kv currents were differentially inhibited by selective neurotoxins like phrixotoxin-1 and α-dendrotoxin as well as by antagonists like 4-aminopyridine, ammoniumchloride, tetraethylammonium chloride and quinidine. In viability and proliferation assays chronic inhibition of the A-type currents severely disturbed the cell cycle and precluded proper hNPC proliferation, while the blockade of delayed-rectifiers by α-dendrotoxin increased proliferation. Conclusions/Significance These findings suggest that A-type potassium currents are essential for proper proliferation of immature multipotent hNPCs.


Introduction
Human neural progenitor cells (hNPCs) isolated from fetal brain tissue are considered a promising source for cell replacement therapies in neurodegenerative disorders [1]. They bear an immense potential to proliferate and represent an appropriate in vitro model for investigating mechanisms of early human brain development [2] including ion channel function. The expression of ion channels and their physiological properties are modulated during cell differentiation [3,4]. Vice versa, ion channels are involved in the regulation of cell differentiation [5]. Proliferation may also be modulated by ion channel activity, whereas the expression of functional voltage-gated potassium (K v ) channel subtypes seems to be particularly important. For example, proliferation of activated immune cells is repressed by K v 1.3 blockade [6], and tumor cell divisions are reduced by selective inhibition of Ca 2+ -activated potassium channel subtypes [7]. In contrast, the selective blockade of K v 1.3 and 3.1 in rat neural progenitor cells increased proliferation [8].
While immature progenitor cells rarely exhibit sodium currents and cannot generate action potentials [9,10], functional K v channels are expressed early during brain maturation with developmentally regulated and highly cell type specific patterns [11][12][13]. In Drosophila CNS precursors, the expression of K v currents seemed to be cell autonomous, while other currents changed, when cell-cell contacts occurred [14]. Therefore, potassium channel function is assumed to be a key requirement for proper progenitor cell proliferation and also may pave the way for neuronal differentiation [15][16][17].
After identification of the four K v channel genes Shaker, Shab, ShaI and Shaw in Drosophila [18,19], 8 related gene families were discovered in mammals [20]. Among these, K v 1, K v 2, K v 3 and K v 4 can form homomeric and heteromeric channels, while K v 5, K v 6, K v 8 and K v 9 are 'electrically silent' and become conducting only after building heteromers with subtypes of K v 1-4 [21]. Potassium channels regulate neuronal excitability by setting resting membrane potentials as well as firing thresholds and by repolarizing action potentials [22,23]. In most cells, voltageactivated potassium (K v ) outward currents exhibit a transient component, which is characterized as the fast-inactivating A-type current (I A ), and a non-inactivating or slowly inactivating sustained component that comprises delayed-rectifying currents with slow (I DR ) or fast (I D ) activation kinetics [24,25]. Early functional investigations pointed out that I A is typically involved in setting the interspike interval [22], while I DR is essential for fast repolarization of action potentials and consequently contributes to repetitive firing [22,26]. Biophysical separation of these two currents can be obtained by the design of appropriate voltage protocols [14,27]. However, due to the diversity of K v channels additional pharmacological isolation of current components is often required [25]. Classical agents to block neuronal K v channels are tetraethylammonium chloride (TEA), which was described to be more effective at blocking I DR [28], and 4-aminopyridine (4-AP), which was commonly used to inhibit I A [29]. Other potent inhibitors of neuronal K + currents are quinidine (QND) a structural isomer of the antidysrhythmic drug quinine, that has been used as a Na + channel blocker [30], and the TEA analogon NH 4 Cl. Naturally occuring toxins like a-dendrotoxin (aDTX), margatoxin (MTX) and phrixotoxin (PTX) are highly selective for single K v channel subtypes and very potent, because of their strong binding affinity [31][32][33][34].
In the present study we show that proliferating hNPCs express functional K v channels, while they do neither exhibit sodium currents nor action potential firing. An overview of the investigated K v 1-4 channels and their published functional characteristics is given in Table S1. The expression pattern of K v channel subtypes was investigated in immature hNPCs predominantly expressing the A-type channel transcript K v 4.2 and in differentiating cells, which showed decreased A-type channel formation. On the other hand, delayed-rectifying channels were upregulated during differentiation. Whole-cell K v currents were separated biophysically into a transient, I A -like current and a sustained component denoted as I K . Both current components exhibited different sensitivities towards individual K v antagonists, which we utilized to unravel their specific contributions to cell viability and progenitor cell proliferation. The inhibition of I A significantly reduced the proliferation capacity and cell viability, indicating an important role of Atype potassium channels for proliferation and survival of hNPCs.
Voltage-gated currents were activated from a holding potential of 2100 mV by depolarizing steps to 100 mV in 10 mV increments (300 ms). Steady-state inactivation of K v currents was determined via hyperpolarizing prepulses increasing in 10 mV increments from 2130 mV to 50 mV (500 ms) followed by a test pulse to 50 mV (300 ms). Current amplitudes were measured between 0 and 20 (transient component, t.c.) and between 280 and 300 ms (sustained component, s.c.) of each depolarizing voltage pulse. Biophysical separation of a delayedrectifier current (I K ) was obtained in activation protocols by a depolarizing prepulse to 240 mV (500 ms), which inactivated the transient A-type current (I A ). I A could be isolated in inactivation protocols by a test pulse to 0 mV, because it activated at slightly more negative potentials than I K . Both current components were additionally separated pharmacologically by application of 10 mM 4-AP to proliferating hNPCs, with I K being identified as the 4-AP-insensitive component measured in activation protocols and I A was isolated by subtracting the 4-AP-insensitive component of steady-state inactivation currents from control currents (Fig. 1). K v currents evoked in activation protocols were converted to chord conductances assuming a reversal potential (V rev ) of 282 mV (calculated according to 130 mM K + inside/5.4 mM K + outside). Values were normalized to the peak amplitudes and fitted to the Boltzmann distribution using Origin 6.1 (OriginLab Corporation, Northampton MA, USA): where V 1/2 is the half maximal activation/inactivation, and dV the slope of the voltage dependency.
For dose-response relationships the inhibition of biophysically separated peak currents was determined during a single depolarizing voltage step from 2100 mV to 100 mV (240 mV prepulse, for I K ) or to 0 mV (2130 mV prepulse, for I A ). At the same time antagonists were applied starting 30 s prior to the test pulses. Values were normalized to peak amplitudes recorded in the absence of antagonists and fitted with the Hill equation using Origin 6.1: In inactivation protocols I A was revealed by a test pulse to 0 mV only since it activated at slightly more negative potentials than I K . During each voltage step peak values of the transient component were measured between 0 and 20 ms and sustained currents were determined between 280 and 300 ms. Chord conductances and current values respectively were normalized to their peak amplitudes and fitted to a Boltzmann distribution and current-voltage-relationships of control currents (A), pharmacologically (B) as well as biophysically (C) separated currents were calculated (iii, see Tab. 1). Note the similar I-V relations for both separation procedures. doi:10.1371/journal.pone.0006168.g001 with IC 50 being the half maximal inhibitory concentration, and dc the Hill coefficient determining the slope of the concentration dependency.

Total RNA isolation and PCR analysis
Total RNA was isolated from proliferating hNPCs as well as from differentiated cells (4 tissue preparations each) grown in 75 cm 2 PLO/FN-precoated culture flasks using the RNeasy mini kit (QIAGEN Sciences, Germantown MD, USA) according to the manufacturer's protocol. First-strand cDNA was prepared from total RNA using the RevertAid first strand cDNA synthesis kit (Fermentas International Inc., Burlington, Canada). 30 ml samples of total RNA were transcribed to cDNA with 600 U of reverse transcriptase. The reaction mixture of 60 ml further contained 5 mM oligo(dT) 18 primer, 0.5 mM nucleotide triphosphates (dNTPs), 50 mM KCl, 4 mM MgCl 2 , 10 mM dithiothreitol (DTT) and 50 mM Tris-HCl (pH 8.3). Oligonucleotide primers for subtypes of the K v channel families 1-4 (see Table S2; MWG Biotech AG, Ebersberg, Germany) were designed to flank intron sequences, if feasible, using Primer 3 software (http://frodo.wi.mit. edu) and tested by means of conventional PCR analysis. PCR samples contained: 100 ng cDNA, 0.625 U Taq DNA polymerase (Fermentas International Inc., Burlington, Canada), 2 mM forward and reverse primers, 1 mM dNTPs, 50 mM KCl, 2.5 mM MgCl 2 and 10 mM Tris-HCl (pH 8.8) in a final volume of 25 ml. The amplifications were performed in a Peltier thermal cycler (MJ Research Inc., Bio-Rad, Watertown MA, USA) using the following protocol: 95uC for 4 min to activate the Taq polymerase, followed by 30 cycles of 95uC for 45 s, 55uC for 40 s and 72uC for 1 min, amplification was stopped at 72uC for 10 min. Aliquots of the PCR reactions were analysed by 2% agarose gel electrophoresis and visualized by ethidium bromide fluorescence using a MultiImage light cabinet and the analysis software AlphaImager 120 v. 5.1 (Alpha Innotech Corporation, San Leandro CA, USA).
Quantitative real-time PCR was performed using 300 ng cDNA from total RNA, 600 nM forward and reverse primers, Platinum-SYBR Green qPCR SupermixH (SYBR Green I, 0.375 U Platinum Taq DNA polymerase, 20 mM Tris-HCl (pH 8.4), 50 mM KCl, 3 mM MgCl2, dNTPs 200 mM each, 0.25 U UDG) and 100 nM 6carboxy-X-rhodamine (both from Invitrogen) using the following protocol in an MX 3000P instrument (Stratagene, La Jolla, CA, USA): 2 min 50uC, 2 min 95uC and 50 cycles of 15 s 95uC, 30 s 60uC. To confirm a single amplicon a product melting curve was recorded. Threshold cycle (Ct) values were placed within the exponential phase of the PCR as described previously by Engemaier et al. (2006). Ct values of 4-12 independent experiments, each performed in duplicate, were normalized to ribosomal protein L22 (Ct2Ct RPL22 = DCt) [46]. DCt values were converted to 2 2 D Ct to calculate the relative expression levels [35].

Cell viability
Evaluation of cell viability was performed by a tetrazolium salt assay using the reagent 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, Sigma-Aldrich GmbH). In viable cells MTT is converted by the mitochondrial dehydrogenase to a blue formazan product [47,48,49]. HNPCs were seeded into 96-well PLO/FN-precoated culture plates (10,000 cells/well, 3 tissue preparations) and incubated for 24 h at 37uC before K v antagonists were added. Cells were treated for 72 h with different concentrations of 4-AP (0.1-2 mM), PTX (1-1000 nM), NH 4 Cl (1-100 mM), QND (5-100 mM), DTX (0.01-10 mM), MTX (1-500 nM) and TEA (1-100 mM). Additionally, electrophysiologically determined inhibitory doses (IC 50 /IC 80 ) were used to compare the effects on cell viability (for each concentration n$4 well). Untreated cells were used as control. After the culture period, 10 ml of 5 mg/ml MTT stock solution were added to each well. Following additional 4 h of incubation at 37uC, culture medium was rejected. To solubilize the formazan crystals 100 ml of 5% acid isopropyl alcohol was applied to the adherent cells and plates were placed on a shaker for at least 30 s. Cell viability was determined colorimetrically at 570 nm using the automated Synergy HT multi-mode microplate reader equipped with the analysis software Gen 5 (BioTek Instruments Inc., Winooski, VT, USA). Absorbance values were normalized to control values of untreated cells.
According to this, a flow cytometric analysis was performed to substantiate the effects on cell cycle (see Methods S1).

Cell proliferation
Progenitor cell proliferation was quantified by a colorimetric immunoassay based on the measurement of 5-bromo-2-deoxyuridine (BrdU) incorporation during DNA synthesis [50] (cell proliferation ELISA, Roche Diagnostics GmbH, Mannheim, Germany). HNPCs were seeded into 96-well PLO/FN-precoated culture plates (10,000 cells/well, 3 tissue preparations) and incubated for 24 h at 37uC before K v antagonists were added. Cells were treated for 72 h with electrophysiologically determined inhibitory doses (IC 50 /IC 80 ) of 4-AP, PTX, NH 4 Cl, QND, DTX, MTX and TEA (for each concentration n$4 well). Untreated cells were used as control. After the culture period, 100 mM BrdU was added to each well. During a labeling period of 4 h at 37uC the pyrimidine analogue BrdU was incorporated in place of thymidine into the DNA of proliferating cells. After rejecting the labeling medium 200 ml FixDenat solution were added to each well to fix the cells and denature DNA during an incubation period of 30 min at 20-22uC. The fixing solution was rejected and 100 ml/ well anti-BrdU-POD antibody solution was added to bind the incorporated BrdU. The cells were incubated for 90 min at 20-22uC and subsequently washed 3 times with phosphate-buffered saline. By adding 100 ml/well tetramethylbenzidine solution the substrate reaction was started and immune complexes were detected within 5-10 min. The reaction product was quantified by measuring the absorbance at 370 nm (reference wavelength 492 nm) using a scanning multiwell spectrophotometer equipped with the analysis software Gen 5 (Synergy HT multi-mode microplate reader, BioTek Instruments Inc., Winooski, VT, USA). The absorbance values directly correlate to the amount of DNA synthesis and hereby to the number of proliferating cells and were normalized to absorbance values of untreated cells.

Statistical analysis
Data were expressed as mean6standard error (SEM). Statistical differences were calculated with Students's t-test (two-tailed, unpaired) using Origin 6.1 (OriginLab Corporation, Northampton MA, USA) or one-way ANOVA, followed by Tukey's post-hoc test using GraphPad Prim 3 (GraphPad Software Inc., La Jolla, USA); p values#0.05 were considered significant.

Voltage-gated potassium currents in proliferating hNPCs
To characterize the voltage-dependency of voltage-gated potassium (K v ) currents in proliferating human neural progenitor cells (hNPCs) outward currents were elicited in whole-cell voltageclamp recordings either in activation protocols or steady-state inactivation protocols (Fig. 1A). We found that outward currents were composed of a transient (t.c.) and a sustained component (s.c.). Typically, the inactivating current component is considered as I A and the sustained component as I DR or I D [22,24]. We did not distinguish between I DR and I D and denoted the sustained component as I K . The transient component of whole-cell potassium outward currents reached maximal capacitance-corrected current densities of 329642 pA/pF in activation protocols, while the sustained component measured only 5667 pA/pF (n = [35][36][37][38]. In inactivation protocols lower current densities were obtained (t.c. 227623 pA/pF, s.c. 2164 pA/pF, n = 36-38; Fig. 1A) due to the reduced driving force. Inactivation data of the sustained current (s.c.) were best fit with a sum of two Boltzmann equations (Fig. 1A iii). Because the first component showed values similar to the transient component, this likely reflects the contribution of I A to the sustained component. I A and I K were pharmacologically separated by application of 10 mM 4-aminopyridine (4-AP). I K was classified as 4-APinsensitive current in activation protocols (3065 pA/pF, n = 10-13) and contributed 10% to the transient and 47% to the sustained whole-cell current. I A was isolated as 4-AP-sensitive component during steady-state inactivation (207666 pA/pF, n = 6-7) and constituted 90% of the transient and 53% of the sustained component of K v outward currents (Fig. 1B). In addition, biophysical separation of I K was performed in activation protocols by a depolarizing prepulse to 240 mV, which caused inactivation of I A . The biophysically measured I K amplitudes (3263 pA/pF, n = 36-38) were comparable to pharmacologically determined values. Also the voltage dependency was similar, while halfmaximal activation values differed between the two separation methods. Because I A was activated at slightly more negative potentials than I K , it was isolated in inactivation protocols by a test pulse to 0 mV and had amplitudes of 9668 pA/pF (n = 33-36)smaller than the pharmacologically separated I A , which we attribute to the smaller driving force during depolarization to 0 mV instead of 50 mV. The current-voltage relationships of I A were similar with pharmacological and biophysical separation (Fig. 1C, Tab. 1), indicating that the same current was separated. In the following experiments biophysical separation was used, since we determined the sensibility of I A and I K against different K v antagonists in dose-response curves.
In proliferating hNPCs half-maximal activation of I K was determined at 10 to 30 mV by fitting activation curves of normalized chord conductances to the Boltzmann distribution. Fitted inactivation curves of current values showed half-maximal inactivation of I A at 260 to 270 mV (Fig. 1iii, Table 1). Wholecell K v currents were constituted to 90% by I A and to 10% by I K . Thus, A-type currents are the predominant potassium outward currents in immature, proliferating hNPCs.

Comparison of K v currents in hNPCs and differentiated cells
To investigate the development of K v currents during differentiation, hNPCs were exposed to a differentiation medium (DM) for 14 days prior to the recording ( Fig. 2A, B). Differentiated hNPCs represent a heterogenous population of cells composed of neurons (,50% Tuj1 positive), astrocytes (,30% GFAP-positive), oligodendrocytes and cells that do not differentiate [10]. After 14 days of differentiation they exhibited no remarkable expression of sodium inward currents, which is consistent with Schaarschmidt et al. (2009) [10].
The biophysically separated I K showed similar half-maximal activation (6 mV in DM vs. 91 mV in PM), but lower voltage dependency (11 mV/e-fold in DM vs. 22 mV/efold in PM). Current-voltage relationships of the transient I A were comparable -half-maximal inactivation at 272 mV in DM vs. 277 mV in PM, voltage dependency 8 mV/e-fold in DM vs. 7 mV/e-fold in PM (Fig. 2iii, Tab. 2).
Furthermore, in differentiated cells the mean current density of I K was significantly increased (4566 pA/pF in DM vs. 2963 pA/ pF in PM, n = 23-36), while I A amplitudes decreased (54612 pA/ pF vs. 9668 pA/pF in PM, n = 22-36; Fig. 2C). Thus, during differentiation I K seems to be upregulated, while I A is smaller than in proliferating hNPCs.

Biological equivalents of voltage-gated potassium currents (K v )
K v channel subtypes in proliferating hNPCs as well as in differentiated cells were identified using reverse transcription   polymerase chain reaction (RT-PCR) analysis based on mRNA expression. Specific primers for several K v channel subtypes were designed and tested by means of conventional PCR (see Table S2, Fig. 3A). A comprehensive expression pattern of almost all tested subtypes of the K v channel families 1-4 was detected in hNPCs except K v 1.4, 3.2 and 3.3). This broad spectrum of K v channels was maintained during differentiation. The expression of several K v channel transcripts was quantified by real-time PCR analysis (Fig. 3B). The A-type channel transcript K v 4.2 showed the highest expression level and, thus, seemed to contribute predominantly to the generation of K v currents in proliferating hNPCs. During differentiation the expression of the A-type channel K v 4.2 was significantly downregulated. Also the expression of other A-type channels decreased, while the delayedrectifier transcripts K v 1.1, 1.7, 2.1, 2.2 and 3.1 considerably increased. This is in line with the electrophysiologically observed increase in I K and decrease in I A in differentiated cells compared to immature hNPCs.

Pharmacological inhibition of K v currents in hNPCs
There is a broad spectrum of specific and less specific K v antagonists [25]. To selectively inhibit either I A or I K we tested some of the most frequently used blockers on hNPCs and monitored the concentration-dependency of their inhibitory effects on the biophysically separated current components (Fig. 4,  Table 2). We started with 4-aminopyridine (4-AP) typically considered as K v blocker preferentially affecting I A , but with moderate specificity [24,29]. 4-AP inhibited I A with IC 50 = 1.7 mM and a Hill slope of 1.4 (Fig. 4i). I K was not completely blocked. To selectively inhibit I A the spider toxin phrixotoxin-1 was used, which acts as an antagonist on K v 4.2 and 4.3 channels [34]. Because in hNPCs I A is predominantly carried by K v 4.2 (Fig. 3B, Tab. S1), this current component was sufficiently blocked with IC 50 = 1.8 mM and slope 0.5, while I K was not affected (Fig. 4ii). The quaternary ammonium salt NH 4 Cl was actually used to inhibit I K . Because of the higher polarity compared to its analogon tetraethylammonium chloride (TEA), it is considered to act on the outer quaternary ammonium ion receptor of K v channels [51]. Surprisingly, it stronger inhibited I A (IC 50 = 35.5 mM, slope 0.9) than I K (IC 50 = 255.6 mM, slope 1.2; Fig. 4iii). But compared to 4-AP and PTX high doses were required.
Furthermore, the classical potassium channel antagonist TEA, which is typically used to block I K , but with moderate specificity, was applied to the cells [24,28]. TEA blocked I K with an IC 50 value of 18 mM and a Hill slope of 0.5 marginally stronger than I A with IC 50 = 49 mM and slope 1.1 (Fig. 4iv). As a fifth antagonist quinidine (QND) -a classical Na + channel blocker, which is reported to non-specifically block I K and I A -was tested [30,52]. We found that in hNPCs the IC 50 value for I K inhibition (IC 50 = 3.4 mM, slope 0.8) was significantly lower than for blocking I A (IC 50 = 42.0 mM, slope 0.8; Fig. 4v). Additionally, the  neurotoxins a-dendrotoxin (DTX) and margatoxin respectively (MTX; see Results S1) were applied, which are considered to specifically affect K v 1 subtypes [31]. We found that both selectively blocked I K (DTX with IC 50 = 163.9 nM and Hill slope 0.7, MTX with IC 50 = 180.7 nM and slope 0.5), while they were ineffective in blocking I A (Fig. 4v, Fig. S1). Because the channel transcripts K v 1.2 and 1.3 showed low expression levels, MTX predominantly inhibited K v 1.1, while DTX additionally blocked K v 1.6 (Fig. 3B, Tab. S1).
Taken together, 4-AP and NH 4 Cl preferentially and PTX specifically blocked I A , while QND stronger and DTX selectively inhibited I K . TEA acted as a non-specific K v channel blocker in hNPCs.

Biological effects of K v channel inhibition in hNPCs
We further investigated whether K v channels play a role in cell survival. Towards this end, we applied various concentrations of the K v antagonists for 3 days prior to analysis by MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay, which colorimetrically measured the production of MTT formazan in viable cells (Fig. 5A). According to the above findings, 4-AP, PTX and NH 4 Cl were utilized to inhibit I A , while QND (low doses) and DTX were applied as specific I K blockers. TEA and higher doses of QND blocked all K v currents. For QND treatment the extracellular KCl concentration had to be raised to 10 mM in order to achieve sufficient inhibitory effects, probably due to its action as an open channel blocker [53,54].
A significant reduction of cell viability was observed after blocking I A with electrophysiologically determined inhibitory doses (IC 50 /IC 80 ) of 4-AP, PTX and NH 4 Cl as well as after treatment with TEA and QND, which blocked all K v currents. On the contrary, low doses of QND and DTX, which specifically inhibited I K , did not affect cell viability (Fig. 5B). To substantiate these results, cell cycle analysis was performed, that yielded a similar increase in apoptosis after inhibition of I A and all K v currents respectively, while specific I K antagonists did not induce cell death (see Results S2, Fig. S2).
To unravel the contribution of K v channels to progenitor cell proliferation we performed a BrdU assay after 3 days of potassium channel blockade according to MTT assay. The inhibition of Atype K v channels by 4-AP, PTX and NH 4 Cl significantly impaired cell proliferation. Blocking all K v currents by TEA and QND had similar effects. In contrast, specific inhibition of delayed-rectifier channels by low doses of QND did not affect proliferation, while the application of DTX even increased proliferation of hNPCs (Fig. 6).
Taken together, these results demonstrate a substantial effect of K v currents on cell survival and proliferation, mainly mediated by I A .

Discussion
We tested the hypothesis that voltage-gated potassium (K v ) channels play a functional role in the development of human neural progenitor cells (hNPCs).
In whole-cell patch clamp recordings the biophysical separation of two K v currents, I A and I K , was obtained by different voltage protocols. The transient current I A was determined in steady-state inactivation protocols by a test pulse to 0 mV, because it is activated at slightly more negative potentials than I K . I K was measured as the sustained outward current in activation protocols following a prepulse to 240 mV, which inactivated I A . Two types of delayed rectifier currents have been described previously: I DR and I D . While I DR is slowly activated with a time to peak of 50-100 ms and does not show pronounced steady-state inactivation, the delay current I D is rapidly activated and slowly inactivated [24,25]. We could not distinguish between I DR and I D and denoted this current as I K . Similar currents for I A and I K were obtained by pharmacological separation. Whole-cell K v currents were constituted to 90% by I A and to 10% by I K .
During differentiation I K amplitudes increased, while I A decreased without considerable changes in current-voltage dependencies. An increase in voltage-activated K v currents during development was observed before in several other cell types, for example in rat retinal ganglion cells [55] or in rat cerebellar granule cells [56]. However, downregulation of I A has not been described so far.
In hNPCs a broad pattern of K v channel subtypes was detected with almost all K v 1-4 channels being expressed except K v 1.4, 3.2 and 3.3. The A-type channel transcript K v 4.2 showed predominant expression levels and, thus, seems to have a critical impact on the physiological characteristics of immature progenitor cells. However, expression of K v channel mRNA and electrophysiological or pharmacological K v properties are quite distinct [25]. Although the in vitro expression of individual a subunits lead to generation of either classical I A or I DR currents [57,58], the physiological properties may be dramatically changed by formation of heteromultimers [59], b subunit association [60,61], the degree of phosphorylation [62,63] as well as the oxidative state [64,65]. Therefore, we combined molecular expression studies with the physiological and pharmacological characterization of K v channels. Whereas the high expression of K v 4.2 mRNA is in line with the 90 percent contribution of I A to whole-cell K v currents, I K -producing delayed-rectifier channels are less prominent. Recently, in rat NPCs derived from the subventricular zone I A was found to be mediated by K v 4.3 and I K by K v 2.1 [9], while in rat midbrain-derived NPCs high levels of the DR channels K v 1.3 and K v 3.1 as well as the A-type channel K v 1.4 were expressed [8]. Thus, K v channel expression seems to be not only region, but also species specific. During differentiation of hNPCs the formation of A-type channels significantly decreased, while delayed-rectifying channels are upregulated analogous to a reduction in I A and an increase in the generation of I K currents.
Pharmacological investigations revealed different sensitivities of I A and I K to the applied K v antagonists. PTX selectively blocked K v 4.2 and 4.3 [34], which contribute largely to I A , and, thus was sufficient in blocking A-type currents in hNPCs. 4-AP is traditionally used as a blocker of A-type potassium channels [24,29]. In hNPCs 4-AP preferentially inhibited I A , but with less specificity. Since I K was not completely blocked, IC 80 values were used to block I A , but an inhibition of delayed-rectifying channels could not be excluded. Selective inhibition of K v 1 delayed rectifier channels was obtained by DTX or MTX [31,32]. Especially DTX sufficiently blocked K v 1.1 and 1.6, which showed the highest expression levels among delayed-rectifying K v channels in hNPCs. In hNPCs low doses of the classical Na + channel blocker QND preferentially affected I K (IC 50 = 3 mM), while higher concentrations also inhibited I A (IC 50 = 43 mM). To obtain appropriate effects it was necessary to add 10 mM KCl to QND-treated cells due to its action as an open channel blocker [53,54]. TEA, traditionally used as an inhibitor of DR potassium channels [24,28], non-specifically blocked both current components and showed at best a slight preference in blocking I K . Hence, the biological effects of A-type channel inhibition were investigated in cell viability and proliferation assays using 4-AP and NH 4 Cl to preferentially and PTX to selectively block I A , while low doses of QND and DTX specifically inhibited I K . TEA acted as an unspecific K v channel blocker in hNPCs.
Potassium channel function is assumed to be a key requirement for proper progenitor cell proliferation and also essential for functional neuronal differentiation [15][16][17]66,67]. In mature neurons K v currents regulate neuronal excitability, while in undifferentiated neural progenitors they are speculated to be involved in cell proliferation [9]. By using the spider toxin PTX we were able to selectively block A-type channels and, thus, to investigate their specific contribution to cell viability and proliferation. In hNPCs a concentration-dependent reduction in cell viability and proliferation was observed after specific I A inhibition with PTX. Less specific (4-AP, NH 4 Cl) as well as nonspecific K v antagonists (TEA, QND) showed similar toxicity. These results indicate that voltage-activated A-type currents generated predominantly by K v 4.2 channels are likely to play a key role for proliferation and survival of hNPCs. This hypothesis is underlined by a downregulation of functional A-type channels with disrupting proliferation and inducing cell differentiation. Similar findings were obtained in adult neural progenitor cells, which showed an injury-induced increase in proliferation mediated by A-type K v 4 channels [17]. Because of its fast activation and inactivation properties, I A prevents mature neurons from responding to fast depolarizations [24], whereas in immature progenitor cells neuronal excitability is absent, but the occurrence of Ca 2+ transients and their regulation by K + channels has been described [67]. In this respect, the hyperpolarizing effect of K + channels on the plasma membrane was thought to provide a driving force for the influx of Ca 2+ , which was believed to trigger cell proliferation [68,69]. However, the exact mechanisms and tasks of I A in proliferating neural progenitor cells remain to be fully elucidated. In contrast, the proliferation of oligodendrocyte progenitor cells is supposed to be controlled by the activity of several DR channels of the K v 1 family [70,71] suggesting different functions of K v channels in neural and glial progenitors.
Furthermore, by using the snake toxin DTX we were able to selectively block I K . DTX did not cause accelerated cell death, but slightly increased proliferation of hNPCs. If we vice versa disrupted proliferation and induced differentiation, functional delayed-rectifier channels were upregulated. An increase in proliferation was also described in rat midbrain-derived NPCs after selective blockade of the DR channels K v 1.3 and 3.1. Two explanations were described: First, a Ca 2+ independent regulation via cell cycle mechanisms. Second, the mediation by a higher open probability of voltage-gated Ca 2+ channels in response to the depolarizing effect caused by the K v channel block and an increase of intracellular Ca 2+ [8]. However, the fact that our data on differentiated hNPCs were obtained from a heterogeneous population of about 50% neuronal and 30% glial cells [10] allows no definitive conclusion about the role of delayed-rectifying potassium channels in the development of mature functional properties.
In summary, hNPCs generated K v currents that consist to 90% of A-type currents predominantly produced by K v 4.2 channels. Whereas delayed-rectifying currents mainly generated by K v 1.1 and 1.6 were small. Inhibiting I A function caused a dramatic decrease in proliferation and extensive cell death and, vice versa, disrupting proliferation reduced A-type current formation. These findings emphasize that even A-type potassium channels may play a key role in proliferation and survival of immature progenitor cells. On the other hand, the inhibition of I K was less toxic and in case of DTX even increased progenitor cell proliferation. This is in line with the finding that non-proliferating, differentiating cells upregulated these channels.