From Pan-Reactive KV7 Channel Opener to Subtype Selective Opener/Inhibitor by Addition of a Methyl Group

The voltage-gated potassium channels of the KV7 family (KV7.1–5) play important roles in controlling neuronal excitability and are therefore attractive targets for treatment of CNS disorders linked to hyperexcitability. One of the main challenges in developing KV7 channel active drugs has been to identify compounds capable of discriminating between the neuronally expressed subtypes (KV7.2–5), aiding the identification of the subunit composition of KV7 currents in various tissues, and possessing better therapeutic potential for particular indications. By taking advantage of the structure-activity relationship of acrylamide KV7 channel openers and the effects of these compounds on mutant KV7 channels, we have designed and synthesized a novel KV7 channel modulator with a unique profile. The compound, named SMB-1, is an inhibitor of KV7.2 and an activator of KV7.4. SMB-1 inhibits KV7.2 by reducing the current amplitude and increasing the time constant for the slow component of the activation kinetics. The activation of KV7.4 is seen as an increase in the current amplitude and a slowing of the deactivation kinetics. Experiments studying mutant channels with a compromised binding site for the KV7.2–5 opener retigabine indicate that SMB-1 binds within the same pocket as retigabine for both inhibition of KV7.2 and activation of KV7.4. SMB-1 may serve as a valuable tool for KV7 channel research and may be used as a template for further design of better subtype selective KV7 channel modulators. A compound with this profile could hold novel therapeutic potential such as the treatment of both positive and cognitive symptoms in schizophrenia.


Introduction
The neuronally expressed members of the K V 7 family (K V 7. [2][3][4][5] are the molecular correlates of the M-current, a slowly activating and slowly deactivating potassium current [1,2,3]. K V 7 channels open at voltages below the threshold required for generation of an action potential and play a fundamental role in the control of neuronal excitability. Accordingly, mutations in the genes encoding K V 7.2 and K V 7.3 result in a form of neonatal epilepsy [4,5,6]. Hence, drugs that increase the activity of K V 7 channels have a great therapeutic potential for the treatment of diseases characterized by hyperexcitability, such as epilepsy and neuropathic pain. Retigabine (Trobalt/Potiga), which activates K V 7.2-5, was approved as an add-on treatment for partial onset seizures in 2011 and has proven effective in preclinical models for a wide variety of diseases [7,8,9,10].
Based on mutation studies, retigabine has been shown to bind to a hydrophobic pocket between transmembrane segments 5 and 6 of the channel proteins. Specifically, channel activation by retigabine is critically dependent on a tryptophan residue (W236 in K V 7.2) in the cytoplasmic part of S5 [11,12]. The cardiac K V 7.1 channel has a leucine at this position explaining its resistance to retigabine-induced enhancement. It appears that the tryptophan residue constitutes a structural element of a promiscuous binding site in the channels, since compounds which are structurally different from retigabine, like BMS-204352, (S)-1 and (S)-2 also lose their activating effects when the tryptophan is substituted for a leucine [13,14]. Yet, compounds like ZnPy [15], ICA-27243 [16,17] and fenamic acids [18] are not dependent on its integrity, suggesting that other activator binding sites exist. A leucine within the pore loop and a leucine extending from S6 of the adjacent subunit have also been found to be important residues for retigabine sensitivity [19]. These residues are conserved between the neuronal subtypes whereas K V 7.1 carries valines in both positions. Introduction of the three critical residues in K V 7.1 confers full retigabine sensitivity [19].
We have previously shown that the acrylamide (S)-2 ( fig. 1A) activates K V 7.2-5 [14]. For K V 7.4 and K V 7.5 the effect of (S)-2 is purely positive, while the compound has a bimodal effect on homomeric K V 7.2 and heteromeric K V 7.2/3 channels. For K V 7.2, the positive effects of (S)-2 are a hyperpolarizing shift in the voltage-dependence of activation, a slowing of the deactivation kinetics (t deact ) and an acceleration of the fast component of the activation kinetics (t act_fast ). At voltages below 210 mV the compound also increases the current amplitude and accelerates the slow component of the activation kinetics (t act_slow ). However, at voltages above 210 mV the compound has a secondary inhibitory effect. At these depolarized voltages the effect on the current amplitude and t act_slow crosses over and becomes inhibitory. When we tested (S)-2 on a mutated K V 7.2 channel where the tryptophan residue in S5 (which is essential for the effect of retigabine) was substituted for a leucine (K V 7.2-W236L) all activating effects of (S)-2 were lost and the compound became purely inhibitory. Based on these observations we proposed that the mechanism behind the inhibitory effect on K V 7.2-W236L and the secondary inhibitory action at the WT K V 7.2 may be the same and that the effect of (S)-2 on K V 7.2 can be divided into two entities: 1) an activating part visible as a shift in the voltagedependence of activation and 2) an inhibitory part visible as a decrease in the current amplitude and an increase in t act_slow . According to this proposal, the interpretation of the experimental data is that the W236L mutation disables only the (S)-2 induced activation and thereby unmasks the inhibitory component of the drug effect. Consistent with this proposal is the observation that for K V 7.4, where the effect of (S)-2 is purely positive, the retigabine insensitive mutant (K V 7.4-W242L) is fully insensitive to (S)-2.
This raised the question of whether it would be possible to design a compound that mimics the effect of the W-L mutation, resulting in an inhibitor of K V 7.2 that would not affect K V 7.4. Interestingly, it is described in the structure activity-relationship of the acrylamide class of compounds that methylation of the amide nitrogen of a related acrylamide K V 7.2 opener (S)-1 results in a compound that inhibits K V 7.2 [20]. In our hands (S)-1 is also an inhibitor of K V 7.2-W236L (unpublished observation) and has been reported to be an inhibitor of K V 7.4-W242L [13]. We hypothesized that this chemical modification of the acrylamide may have the same effect as the W-L mutation where the methylated version of (S)-1 would be an inhibitor of both K V 7.2 and K V 7.4, while the methylated version of (S)-2 would inhibit K V 7.2 but not K V 7.4. Here we describe the synthesis and characterization of an analogue of (S)-2, termed SMB-1, where the amide nitrogen is methylated ( fig. 1B). We show that SMB-1 is indeed an inhibitor of K V 7.2. In contrast, K V 7.4 is activated by SMB-1. However, SMB-1 does not have sufficient penetration into the rodent brain to allow examination of its in vivo CNS profile.

Molecular biology
Point mutations were introduced using mutated oligonucleotide extension (PfuTurbo Polymerase, Stratagene, La Jolla, CA, USA) from a plasmid template harboring the cDNA of interest, digested with DpnI (Fermentas, St. Leon-Roth, Germany) and transformed into E.coli XL1 Blue cells. The construct was verified by complete DNA sequencing of the cDNA insert. cRNA was prepared from linearized human wild-type (WT) and mutant K V 7 channels in the pGEM-HE vector using the T7 m-Message Machine kit (Ambion, Austin, TX, USA) according to the manufacturer's instructions.

Electrophysiology
Extraction of Xenopus laevis oocytes and injection of cRNA was performed as described previously [14]. Oocytes were kept in Modified Barth's Saline (in mM: 88 NaCl, 1 KCl, 2.4 NaHCO 3 , 0.41 CaCl 2 , 0.82 MgSO 4 , 0.3 Ca(NO 3 ) 2 , 15 HEPES, pH 7.4 supplemented with 100 U/mL penicillin and 100 mg/mL streptomycin) at 18uC, and currents were recorded after 2-7 days. The care of Xenopus laevis and the oocyte extraction procedure were performed according to national guidelines and approved by the Danish Animal Experiments Inspectorate. K V 7 currents in Xenopus laevis oocytes were recorded using twoelectrode voltage-clamp at room temperature in Ringer buffer (in mM: 115 NaCl, 2.5 KCl, 1.8 CaCl 2 , 0.1 MgCl 2 , 10 HEPES, pH 7.4) as described previously [14]. Data was acquired using pCLAMP 10.2 software (Molecular Devices, CA, USA) and analyzed using pCLAMP 10.2 and GraphPad Prism 4.0 (GraphPad Software Inc., CA, USA). Currents were elicited from a holding potential of 280 mV by 5 s steps to potentials between 2100 and +50 mV in 10 mV increments, followed by a 2 s step to 2120 mV.
SMB-1 was dissolved in dimethyl sulfoxide (DMSO) to obtain a concentrated stock solution. On the day of experiments the stock solution was thawed and diluted in Ringer buffer to the final concentrations. A concentration of 10 mM SMB-1 was used unless indicated otherwise. The final DMSO concentration never exceeded 0.1%.

Curve fitting and statistical analysis
Current -voltage (I-V) relationship curves were generated by normalizing the steady-state peak current measured at potentials between 2100 and +50 mV to the current measured at +50 mV in control recordings and plotting the values against the test potential.
The voltage-dependence of activation was determined from tail current analysis using the current measured immediately after the transient capacitive current after stepping to 2120 mV from potentials between 2100 mV and +50 mV. Data was normalized to extend from 0-1 and the tail current-voltage relationship was fitted to the Boltzmann equation: where I max is the maximum tail current, I min is the minimum tail current, V 0.5 is the potential for half maximal activation and k is the slope factor. A V 0.5 was calculated for each individual experiment and statistical significance was estimated by paired two-tailed Student's t-test. Activation kinetics for K V 7.2 and K V 7.2-W236L was determined by fitting the traces recorded at potentials between 240 and +40 mV to a double exponential function: where I(t) is the current at time t, I fast and I slow are the current amplitudes at infinite times, and t fast and t slow are the time constants of the fast and slow components, respectively. Activation kinetics for K V 7.4 and K V 7.4-W242L was determined by fitting the traces recorded at potentials between 230 and +50 mV to a single exponential function: where I(t) is the current at time t, I 0 is the peak current and t is the activation time constant. For investigation of deactivation kinetics the tail current traces recorded at 2120 mV when stepping from the channels' respective ,V 0.5 were fitted to a single exponential function (Eq. 3).
Dose-response curves from electrophysiological experiments were made by plotting the increase in steady state current at + 50 mV expressed in percentages as a function of drug concentration. The data were then analyzed by non-linear regression and fitted to the equation for sigmoidal dose-response with variable slope: where R 1 is the initial response, R 2 is the maximum response, X is the logarithm of the drug concentration and nH is the slope (Hill coefficient) of the curve. For data concerning activation kinetics statistical significance was determined by two-way ANOVA followed by Bonferroni posttests (if the overall P-value of the drug factor was less than 0.05). Before the statistical analyses of the fast component of the activation for K V 7.2 and K V 7.2-W236L the measured time constants were log-transformed to meet the assumption of normality. For traces fitted to double exponential functions only those where the sum of I fast and I slow equaled the peak current were included in the analysis.
For analysis of the remaining data statistical significance was determined by two-tailed Student's t-test if single comparisons were made and by two-way ANOVA followed by Bonferroni posttests if multiple comparisons were made. Statistical analyses were carried out using GraphPad Prism 4.0. p,0.05 was accepted for statistical difference. All values are shown as mean 6 S.E.M.

Rodent exposure studies
Ethical permission for the in vivo rat procedures used in these studies was granted by the animal welfare committee, appointed by the Danish Ministry of Justice and all animal procedures were carried out in compliance with EC Directive 86/609/EEC and with the Danish law regulating experiments on animals.
Brain and plasma exposure of SMB-1 was evaluated in male Sprague-Dawley rats (225-250 g, Charles River Lab., UK). SMB-1 was administered subcutaneously in solution at a dose of 20 mg/ kg in 25% solutol (dose volume 5 ml/kg) and exposure was assessed after 30 and 120 min (n = 3). Under isoflurane anesthesia, cardiac blood was obtained in EDTA-coated tubes and centrifuged for 10 min at 4uC, after which plasma was harvested. While still under anesthesia, the animal was decapitated, the brain was removed and brain homogenate was prepared by homogenizing the whole brain with 70% acetonitrile (1:4 v/v) followed by centrifugation and collection of the supernatant. Concentrations of SMB-1 were determined using turboflow chromatography (dual column, focus mode; Thermo Fisher Scientific, Waltham, MA) followed by tandem mass spectrometry detection (Sciex API-3000 MS; Applied Biosystems, Foster City, CA). The limit of detection was 1 ng/ml in plasma and 5 ng/g in brain.

Membrane permeability and free fraction determination
The bidirectional permeability (P app ) of SMB-1 was examined in an MDCK cell system expressing human MDR1 [ABCB1, Pglycoprotein (P-gp)] as described previously [21]. Permeation of the test compound from apical (A) to basal (B) direction or B to A direction was determined in triplicate over a 150-min incubation at 37uC. The efflux ratio was calculated as the ratio between P app (B to A) and P app (A to B). Free fraction of SMB-1 was determined using a standard equilibrium dialysis methodology with freshly isolated rat brain homogenate or plasma [22]. Equilibrium dialysis was performed by incubating at 37uC for 5 h in triplicates.

SMB-1 is an inhibitor of K V 7.2
We heterologously expressed K V 7 channels in Xenopus laevis oocytes and assessed their biophysical properties by two-electrode voltage-clamp. Upon activating voltage-steps, K V 7.2 channels displayed slowly activating and deactivating currents as expected [1]. Representative current traces are shown in Figure 1C. Application of 10 mM SMB-1 led to a reduction of current amplitudes at all activating potentials ( fig. 1C and D). At +50 mV the current amplitude was 81.763.1% (n = 9) of control. A small, but significant, shift to more positive potentials was observed in the voltage-dependence of activation (V 0.5 : control: 222.562.2 mV; SMB-1: -18.062.7 mV, n = 8, p,0.05, fig. 1E). We also assessed kinetic parameters by determining the time constants of deactivation (t deact ) and activation (t act_fast and t act_slow ). Deactivation kinetics were significantly slowed in the presence of SMB-1 (control: 20.261.4 ms, SMB-1: 37.363.3 ms, n = 7, p,0.001, fig. 1F). SMB-1 did not affect t act_fast ( fig. 1G) but increased t act_slow at 220 mV and above (fig. 1H). These effects of SMB-1 are remarkably similar to the effect of (S)-2 on K V 7.2-W236L [14].
To determine the potency of SMB-1 on K V 7.2 the compound was tested at four concentrations (0.4, 2, 10 and 50 mM). Visual inspection of the solution indicated that the drug precipitated at concentrations above 50 mM. The current amplitude in percentage of control at +50 mV was plotted against the logarithm of the concentration. Fitting the data to the equation for sigmoidal doseresponse with variable slope determined the IC 50 value of SMB-1 to ,7.4 mM (pIC 50 : 5.160.2 M, n = 5, fig. 1I). The compound was not able to fully inhibit the current within the concentration range tested, reaching a maximal inhibition of 67.665.3% of control (n = 5) at +50 mV.

SMB-1 is an activator of K V 7.4
Next, we addressed the effect of SMB-1 on K V 7.4. Submitting Xenopus laevis oocytes expressing K V 7.4 channels to a voltage-step protocol elicited slowly activating currents. A representative experiment is shown in Figure 2A. Application of SMB-1 resulted in increased current amplitudes at 240 mV and above ( fig. 2A and B). The current amplitude was 184.5612.4% of control (n = 8) at +50 mV in the presence of 10 mM SMB-1, which is low compared to the efficacy of (S)-2 on K V 7.4 (1927% of control at + 40 mV) [14]. SMB-1 did not affect the voltage-dependence of activation of K V 7.

Dependency on the critical tryptophan residue for inhibition/activation by SMB-1
To determine whether SMB-1 was dependent on the critical tryptophan residue in the 'retigabine binding-site' for its inhibitory/activating effects, the compound was applied to mutant channel proteins K V 7.2-W236L and K V 7.4-W242L ( fig. 3 and 4). K V 7.2-W236L was inhibited by SMB-1 similar to wild-type (WT) K V 7.2 ( fig. 3A and B). The current amplitude at +50 mV was 68.863.9% of control (n = 6). We observed a trend towards an increase in the half-maximal value for voltage-dependence of activation V 0.5 (control: 227.162.4 mV; SMB-1: 2 15.0610.1 mV, n = 4, p = 0.2, fig. 3C). There was a small but significant increase in t deact (control: 23.362.9 ms; SMB-1: 27.063.0 ms, n = 6, p,0.05, fig. 3D). Like for WT K V 7.2, there was no significant effect of SMB-1 on t act_fast ( fig. 4E), while t act_slow was significantly increased at potentials above 220 mV ( fig. 3F).
Hence, the activating effect of SMB-1 on K V 7.4 is critically dependent on the tryptophan residue while the inhibitory effect on K V 7.2 is not.

Effect of SMB-1 on channels with mutations in the refined retigabine binding site
To test if SMB-1 depends on other residues important for the action of retigabine, we generated channels where the two leucine residues (L275 and L299 in K V 7.2, L281 and L305 in K V 7.4) shown to be part of the binding site in K V 7.3 [19] were substituted with valines, the corresponding residues in K V 7.1. The resulting channels K V 7. The L281V mutation in K V 7.4 resulted in a non-functional channel. A mutation affecting this residue (L281S) has been found in patients with DFNA2 [24], indicating that it may be very critical for K V 7.4 channel function.
The L275V mutation in K V 7.2 affected the efficacy of SMB-1 ( fig. 5A). The current amplitude was 90.464.1% of control at + 50 mV (n = 4). K V 7.2-L299V was inhibited to a similar degree as the WT channel (current amplitude 69.365.6% of control at + 50 mV, n = 4, fig. 5B). The L305V mutation in K V 7.4 abolished the effect of the compound (current amplitude 100.8610.3% of control at +50 mV, n = 6, fig. 5C).

Exposure of SMB-1 in rats
Following subcutaneous administration of 20 mg/kg, SMB-1 was found to be rapidly absorbed systemically, and high plasma concentrations were observed (Table 1). In contrast, we found the brain exposure to be very low relative to plasma with corresponding brain/plasma partition coefficients in the range of 0.04-0.07. Considering the contribution of residual blood in the brain tissue to the brain partition coefficient, these results suggest that SMB-1 possesses limited brain penetration capabilities. In line with these findings, in vitro transport studies in MDCK cells showed that SMB-1 exhibited low membrane permeability of 2.560.3610 26 cm/s with an efflux ratio of 0.7. In addition, SMB-1 was found to be highly bound in plasma (free fraction 0.660.4%) and in brain homogenate (free fraction 0.160.1%) (mean 6 S.E.M., n = 3).

Discussion
Although an increasing number of compounds has been found to activate or inhibit K V 7 channels there is still a lack of truly subtype specific drugs that can discriminate between K V 7.2-5. This has hampered the determination of the individual contribution of each subtype to the in vivo effects of e.g. retigabine and made it difficult to establish which subtype to target for each of the many indications where K V 7 channels have been suggested as attractive therapeutic targets.
In an attempt to develop subtype-specific K V 7 active drugs, we took advantage of previous findings that had established the so- called retigabine binding site and the SAR of acrylamide K V 7 channel compounds [11,12,19,20]. Trying to mimic the effect of a tryptophan substitution, we designed and synthesized a methylated version of (S)-2, termed SMB-1. This compound oppositely affects K V 7.2 and K V 7.4 channels. SMB-1 inhibits K V 7.2 by reducing the current amplitude and increasing the time constant for the slow component of the activation kinetics remarkably similar to the effects of (S)-2 on K V 7.2-W236L [14]. The potency of (S)-2 on K V 7.2-W236L (IC 50 = 8.1 mM) and SMB-1 on K V 7.2 (IC 50 = 7.4 mM) is similar [14]. Hence, our data suggests that methylation of the amide nitrogen disables the activating component of (S)-2's effect on K V 7.2 while preserving the inhibitory component, and thereby recapitulates the pharmacological consequence of the W236L mutation.
The increase in K V 7.4 current levels by SMB-1 was weak compared to the high efficacy of (S)-2 [14]. As K V 7.4-W242L was inhibited by SMB-1, we speculate that the methyl-group in SMB-1 promotes an inhibitory binding mode and thereby tips the activation/inhibition equilibrium towards reduced efficacy relative to the effects of (S)-2 on K V 7.4. The positive effect of SMB-1 on K V 7.4 was critically dependent on the tryptophan residue in S5. Furthermore, K V 7.4-L305V was SMB-1 insensitive, leading us to conclude that SMB-1 binds to the promiscuous retigabine binding site for activation of K V 7.4.
In contrast, the inhibitory action of SMB-1 on K V 7.2 was not critically dependent on the tryptophan residue. However, the L275V mutation in K V 7.2 reduced the inhibitory efficacy of 10 mM SMB-1, indicating that it's binding to K V 7.2 also occurs within the border of this binding pocket. Since this site accommodates pan-reactive compounds, it has been suggested that it would not allow for discrimination between neuronal K V 7 channel subtypes [25]. These findings demonstrate that this binding site can indeed accommodate compounds that oppositely affect neuronal K V 7 channels thereby giving rise to subtype selectivity. To further examine the basis of the opposite effect of SMB-1 on K V 7.2 and K V 7.4 in more detail, a cross-mutational approach may be applied where non-conserved residues within the binding pocket are swapped between the two subtypes.
What may this compound be used for? Foremost, determining the contribution of different K V 7 channel subtypes to currents in various regions may assign particular physiological functions to particular subtypes.
Compounds modulating K V 7 channels may have a therapeutic potential in relation to schizophrenia. This disease is characterized by psychotic or positive symptoms such as hallucinations, delusions, thought disorder, disorganized speech and paranoia; negative symptoms, which include lack of mental activities such as thoughts and motivation; and cognitive symptoms comprising failure of working memory, impairments of learning and attentional dysfunction [26]. According to the dopamine hypothesis of schizophrenia the positive symptoms result from over-active neurotransmission in the mesolimbic dopamine pathway [27,28]. This hypothesis is based on the fact that antipsychotic drugs work by blocking dopamine D 2 receptors, while compounds that increase dopamine levels can induce psychosis or exacerbate schizophrenia [27,29]. On the other hand, a hypodopaminergic state in the frontal cortical terminal fields of the mesocortical dopamine neurons is thought to underlie the negative and cognitive symptoms [28]. The antipsychotic drugs available on the market today are more or less efficient in ameliorating the positive symptoms but have little or no effect on the negative and cognitive symptoms. Compounds with differential modulatory action across the K V 7 channel family could aid in determining the most optimal subtype profile. K V 7 channel inhibitors were initially developed as cognition enhancers based on their ability to increase acetylcholine release [30,31,32]. This effect probably stems from inhibition of K V 7.2 and K V 7.2/3 channels as the reference inhibitor XE-991 has been shown to more potently inhibit K V 7.2 and K V 7.2/3 than K V 7.4 channels albeit the data are reported from different laboratories [1,33]. K V 7 channels also play a role in the dopaminergic system and retigabine has anti-dopaminergic effects in the basal ganglia [9,34,35,36]. K V 7.4 has been suggested as the subtype mediating the efficacy of pan-reactive K V 7 channel openers in modulation of the dopaminergic system [9,35,36]. Hence, we speculate that a  balanced inhibition of K V 7.2 and K V 7.2/3 in combination with enhancement of K V 7.4 current may be efficacious with respect to targeting of both positive and cognitive symptoms in patients suffering from schizophrenia. As this is hypothetical, proof of this concept necessitates an evaluation of SMB-1 in animal models of psychosis and cognition. Unfortunately, even though SMB-1 was not identified as a substrate for the efflux transporter P-gp judged from the efflux ratio in MDR1-MDCK cells, we found it to have low membrane permeability with poor CNS distribution in rats limiting its in vivo pharmacological potential after systemic administration in the CNS. However, the recent years have provided evidence for a role of the so-called neuronal K V 7 channels in the vasculature [37,38,39]. K V 7.4 channels in particular have been involved in the regulation of vascular tone in e.g. renal, coronary, and cerebral arteries [40,41,42]. Hence, further studies are warranted to investigate whether SMB-1 might have vasorelaxant effects which could in turn have an impact on cerebral circulation.
In conclusion, we have identified a novel K V 7 channel active compound with a unique characteristic in that it inhibits K V 7.2 but activates K V 7.4. SMB-1 provides valuable insight into the pharmacology of K V 7 channels and may be used as a template for the further design of better subtype selective K V 7 channel modulators.
Author Contributions Table 1. Plasma and brain exposure of SMB-1 in rats following subcutaneous administration of 20 mg/kg.