New aryl and acylsulfonamides as state-dependent inhibitors of Na v 1.3 voltage-gated sodium channel

Voltage-gated sodium channels (Na v s) play an essential role in neurotransmission, and their dysfunction is often a cause of various neurological disorders. The Na v 1.3 isoform is found in the CNS and upregulated after injury in the periphery, but its role in human physiology has not yet been fully elucidated. Reports suggest that selective Na v 1.3 inhibitors could be used as novel therapeutics to treat pain or neurodevelopmental disorders. Few selective inhibitors of this channel are known in the literature. In this work, we report the discovery of a new series of aryl and acylsulfonamides as state-dependent inhibitors of Na v 1.3 channels. Using a ligand-based 3D similarity search and subsequent hit optimization, we identi昀椀ed and prepared a series of 47 novel compounds and tested them on Na v 1.3, Na v 1.5, and a selected subset also on Na v 1.7 channels in a QPatch patch-clamp electrophysiology assay. Eight compounds had an IC 50 value of less than 1 μ M against the Na v 1.3 channel inactivated state, with one compound displaying an IC 50 value of 20 nM, whereas activity against the inactivated state of the Na v 1.5 channel and Na v 1.7 channel was approximately 20-fold weaker. None of the compounds showed use-dependent inhibition of the cardiac isoform Na v 1.5 at a concentration of 30 μ M. Further selectivity testing of the most promising hits was measured using the two-electrode voltage-clamp method against the closed state of the Na v 1.1-Na v 1.8 channels, and compound 15b displayed small, yet selective, effects against the Na v 1.3 channel, with no activity against the other isoforms. Additional selectivity testing of promising hits against the inactivated state of the Na v 1.3, Na v 1.7, and Na v 1.8 channels revealed several compounds with robust and selective activity against the inactivated state of the Na v 1.3 channel among the three isoforms tested. Moreover, the compounds were not cytotoxic at a concentration of 50 μ M, as demonstrated by the assay in human HepG2 cells (hepatocellular carcinoma cells). The novel state-dependent inhibitors of Na v 1.3 discovered in this work provide a valuable tool to better evaluate this channel as a potential drug target.


Introduction
Voltage-gated sodium channels (Na v s) are transmembrane proteins that open and close in response to membrane potential, controlling the 昀氀ux of sodium ions from the extracellular to the intracellular side [1][2][3].They are expressed in various electrically excitable cells where they are responsible for electrical signalling.The main part of the channels consists of a large α-subunit with four homologous domains, DI-DIV, arranged to form a central pore.Each of the domains contains a voltage-sensor domain (VSD) consisting of four transmembrane helices, S1-S4, which responds to changes in membrane potential and affects the functional state of the channel [4].Although the α-subunit alone is suf昀椀cient for ion conduction, it is usually associated with one or more β-subunits that regulate the kinetics of channel gating, cell surface expression or act as adhesion molecules [5].There are nine known members of the Na v 1 channel family, Na v 1.1-Na v 1.9, with different functional properties and expression patterns in cells, such as peripheral and central neurons (Na v 1.1-1.3 and 1.6-1.9),skeletal muscles (Na v 1.4), and cardiac muscle (Na v 1.5) [2].
Na v channels exist in three main functional states: closed (resting), open, and inactivated [1].The af昀椀nities of Na v channel inhibitors for each of these states are often different.Many known inhibitors block preferentially the open or inactivated state, which is referred to as a state-dependent action.Additionally, many compounds bind with higher af昀椀nity to channels in cells with higher 昀椀ring frequencies, known as use dependence.State and use dependence are preferred properties of Na v channel inhibitors, as this enhances their binding to damaged nerves with pathological 昀椀ring patterns [3].The abnormal activities of Na v isoforms have been associated with various diseases, such as epilepsy (Na v 1.1, 1.2, 1.3, 1.6), periodic paralysis (Na v 1.4), cardiac arrhythmias (Na v 1.5), CNS tremor, ataxia and dystonia (Na v 1.6), and hyper-or hyposensitivity to pain (Na v 1.3, 1.7, 1.8, 1.9).
Early Na v channel modulators, such as local anaesthetics (lidocaine, bupivacaine), anticonvulsants (carbamazepine), antiepileptics (phenytoin) and antiarrhythmics (mexiletine) were largely subtype nonselective.To avoid possible CNS (dizziness, sedation, convulsions) or cardiovascular (arrhythmias, cardiotoxicity) side effects, subtypeselective inhibitors are being investigated.However, due to the high degree of sequence similarity between Na v subtypes, it is dif昀椀cult to achieve subtype selectivity [3,4].In recent years, the discovery of Na v channel inhibitors has been facilitated by an advanced understanding of the biology and genetics of Na v channels [6,7], new automated screening technologies [8], new heterologous ion channel expression systems [9], structural data of bacterial [10][11][12][13][14] and eukaryotic [5,15,16] Na v channels in various functional states, and co-crystal structures of Na v channels with small molecules [17].
Na v 1.7, Na v 1.8, and Na v 1.9 channels have been identi昀椀ed as the major contributors to nociceptive signalling [4].Of these channels, Na v 1.7 is the best studied and is considered a key player in pain reception.The role of Na v 1.3 is less well understood, but some studies suggest that it is also involved in pain pathways [18][19][20][21].Na v 1.3 channels are mainly expressed in the CNS of the embryonic brain and their expression decreases signi昀椀cantly after birth [22].However, it has been shown that nerve injury can lead to upregulation of Na v 1.3, resulting in neuronal hyperexcitability and pain [20,23].In addition, Na v 1.3 is expressed in neutrophils and its inhibition may have anti-in昀氀ammatory effects [24].Mutations in the SCN3A gene, which encodes the α-subunit of Na v 1.3, have been associated with neurodevelopmental disorders, such as epilepsy and brain malformations [25][26][27][28].
One of the most important binding sites for the development of subtype-selective small molecule Na v channel inhibitors that has emerged in recent years is located on the extracellular side of the fourth VSD [17,29,30].The binding site lies between helices S2, S3, and S4 and is partially immersed in the membrane bilayer, as shown by mutational studies [30] and the crystal structure of a small molecule inhibitor bound to this site [17].Inhibitors targeting this site belong to the structural class of aryl-and acylsulfonamides.The anionic arylsulfonamide warhead of the inhibitors makes contacts with the fourth arginine of the S4 helix, while contacts with residues of the S2 and S3 helix are important to achieve isoform selectivity.Aryl-and acylsulfonamides bind to the inactivated state of the channel, stabilizing the voltage sensor in the activated (or "up") conformation and thus trapping the channel in the nonconducting inactivated state.They exhibit state-and use-dependent blockade.
Most aryl-and acylsulfonamides have been investigated as Na v 1.7 channel blockers for the treatment of pain and many of them show very good selectivity pro昀椀les over Na v 1.5 and other isoforms [31][32][33][34][35][36].Several Na v 1.7 inhibitors have advanced through various stages of preclinical and clinical development [2,29,37].However, despite their very high potency and selectivity, it has proven dif昀椀cult to translate these compounds into successful therapeutics [38,39].Many of the known inhibitors have poor drug metabolism and pharmacokinetic (DMPK) properties, poor blood-brain barrier permeability, and high plasma protein binding, so it remains dif昀椀cult to achieve potent analgesic effects in vivo [32].In addition to Na v 1.7, arylsulfonamides have also been investigated as Na v 1.3 inhibitors, and some of the reported compounds achieved good potency and isoform selectivity [16,30,40].
Given the recent discoveries demonstrating the important role of Na v 1.3 in human physiology, new potent and selective Na v 1.3 inhibitors are needed to better evaluate this channel as a potential drug target.

Results and discussion
Design.The design of the present series of compounds was based on the arylsulfonamide class of Na v inhibitors that bind to the extracellular small molecule binding site on VSD4.Compounds I and II (Fig. 1) with IC 50 values for human Na v 1.3 in the low nanomolar range served as starting points [41].The extracellular VSD4 binding site has greater sequence diversity among different Na v subtypes than the central pore region, making this site attractive for the development of subtype-selective Na v inhibitors.The general design protocol is shown in Fig. 1.First, a computational approach was used with a ligand-based 3D similarity search.Compound II was used as a query, and the similarity search was performed on a drug-like subset of the ZINC library of compounds using a combination of ROCS and EON from OpenEye Sci-enti昀椀c Software, Inc [42,43].The obtained results were evaluated, and sixteen compounds (Table 3) were selected based on their calculated ET_Combo similarities (Supplementary information, Table 1S) and structural diversity, and purchased.In addition, new potential Na v 1.3 modulators were designed by systematic structural modi昀椀cations of inhibitors I and II (type I compounds, Fig. 2).Moreover, an Icagen compound III [44] (Fig. 3), recently used by our research group to characterize endogenous sodium channels in the ND7-23 cell line [45], was used as a starting compound for the preparation of new analogues due to its high state-dependent activity against human Na v 1.3 (IC 50 of 56 nM for the inactivated state and 18 μM for the resting state), and no effects on Na v 1.4, Na v 1.5, Na v 1.6, and Na v 1.7 [45].The design strategy for new Na v 1.3 inhibitors based on the compound III is shown in Fig.

(type II compounds).
To facilitate the design of new compounds, inhibitors I-III were visually divided into three structural parts: (i) the substituted pyrazol-5amine (I and II) or 2-phenylcyclopropane-1-carboxamide (III) on the left hand side (LHS, coloured red), (ii) the central benzene ring (coloured black), and (iii) the N-(thiazol-2-yl)sulfonamide moiety on the right hand side (RHS, coloured blue) (Fig. 2).
Subtype selectivity of inhibitors can be achieved primarily by varying the LHS, which usually contains hydrophobic substituents.In the 昀椀rst series of analogues, the substituted pyrazol-5-amine moiety on the LHS was replaced by various substituted pyrrole-2-carboxamide moieties (types Ia-b, Fig. 2).Phenyl, benzyl, or substituted benzyl groups were attached to the pyrrole nitrogen to mimic the structure of the substituents of compounds I and II.In some analogues, positions 4 and of the pyrrole ring were substituted by bromine atoms (types Ia-b, Fig. 2).The substitution pattern on the disubstituted central benzene ring was either 1,4 or 1,3 (type Ia, Fig. 2).The charged warhead at the RHS is critical for establishing interactions with the gating arginine at the VSD.Therefore, we retained the N-(thiazol-2-yl)sulfonamide moiety at this position for most analogues.In some compounds, an acetate group was introduced on the sulfonamide nitrogen in place of the thiazole group to determine whether the thiazole ring was essential for high binding af昀椀nity (type Ia, Fig. 2).In addition to the arylsulfonamides (types Ia and IIa, Figs. 2 and 3), we prepared a series of so-called acylsulfonamides in which N-(methylsulfonyl)carbamoyl or N-(N,N-dimethylsulfonyl)carbamoyl substituents were introduced at the RHS (types Ib-c and IIb-c, Figs. 2 and 3).In the acylsulfonamide series, the predicted pKa of the sulfonamide nitrogen is lower (pKa ~4.2, Marvin-Sketch from Chemaxon Ltd) than for N-(thiazol-2-yl)sulphonamides (pKa ~5.6, MarvinSketch from Chemaxon Ltd), and the compounds also have lower lipophilicity and volume.For the type Ic compounds, benzofuran-2-yl and 4-chlorophenoxyeth-1-yl substituents were introduced into the LHS, inspired by some of the most promising hits obtained in the similarity search (TSS-34, TSS-42, Table 3).
The compound III was recently shown by our group to be a potent inhibitor of human Na v 1.3 with an IC 50 of 56 nM for the inactivated state of the channel and with good selectivity for other Na v 1.x isoforms [45].
We prepared a series of its analogues (Fig. 3) to further investigate the importance of the 2-phenylcyclopropane-1-carboxamide moiety (coloured red, Fig. 3) for biological activity.Both aryl-(type IIa, Fig. 3) and acylsulfonamide series (types IIb-c, Fig. 3) were prepared.For the type IIc compounds, the 2-phenylcyclopropane group was attached to the central benzene ring via a methyleneoxy group to increase the 昀氀exibility of the molecules.
For the synthesis of 4-amino-N-(sulfamoyl)benzamides 23a and 23b (Scheme 6), methanesulfonamide (21a) and N,N-dimethylsulfamide (21b) were 昀椀rst prepared by the reaction of methanesulfonyl chloride (20a) or dimethylsulfamoyl chloride (20b) with ammonia.In the next step, a proton from the NH 2 group of 21a or 21b was removed with sodium hydride, the obtained N-nucleophile was reacted with 4-nitrobenzoyl chloride (16).Finally, the nitro groups of 22a-b were reduced by catalytic hydrogenation.
The carbonyl group of 11a was 昀椀rst reduced with lithium aluminium hydride, and then the resulting alcohol 33 was reacted with methyl 3-or 4-hydroxybenzoate in a Mitsunobu reaction.Compounds 34a and 34b were subjected to alkaline hydrolysis, and the resulting carboxylic acids 35a and 35b were activated with oxalyl chloride in the form of acid chlorides and then reacted with methanesulfonamide (21a), which was pretreated with sodium hydride to remove a proton from the amino group.
Inhibitory activity on Na v 1.3 and Na v 1.5 channels.All 31 synthesized compounds and 16 compounds identi昀椀ed by a 3D ligand-based similarity search were evaluated for their inhibitory effect on the human Na v 1.3 channel and for their selectivity toward the cardiac Na v 1.5 channel.Channels were expressed in Chinese hamster ovary cells (CHO), and screening was performed with an automated patch clamp electrophysiology assay on the Sophion QPatch HT system (Sophion Bioscience A/S), as described in the Experimental section.IC 50 values are reported in Tables 1-3 and represent the concentrations of the compound that inhibit a sodium channel current by 50%.Tetrodotoxin and amitriptyline were used as positive controls.
The electrophysiology experiments on Na v 1.3 channels were designed to re昀氀ect both the resting (closed) state and the inactivated state of the channel.The inactivated state is often the preferred state for pharmacological intervention because under pathophysiological conditions the 昀椀ring frequency is often higher, resulting in a higher percentage of channels in this functional state.To evaluate the effect of the compounds on the resting state of the channel, a 20-ms activating step to −20 mV was applied starting from a holding potential of −100 mV (Peak 1, Tables 1-3).To evaluate the block of the inactivated state of the channel, the second activating pulse was applied after a 5-s prepulse to half inactivation potential (Peak 2, Tables 1-3).Based on the block of the resting state and the inactivated state, the state selectivity was calculated by comparing the IC 50 values at Peak1 and Peak 2.
The block of the Na v 1.5 channel isoform is associated with an increased risk of cardiac adverse events.Voltage protocols re昀氀ecting its physiological cardiac sinus rhythm frequency state were developed for this channel.A pulse train consisting of 10 repetitive activating test pulses was applied at a frequency of 1 Hz.Peak inward currents were determined from the 昀椀rst (Pulse 1) and the tenth (Pulse 10) pulse of each recorded pulse train (Tables 1-3).
Of the 47 compounds tested, eight had IC 50 values for the Na v 1.3 channel in the submicromolar concentration range, and one (15b) showed IC 50 values in the nanomolar range (20 nM).All active compounds acted as state-dependent modulators of the Na v 1.3 channel, preferentially blocking the inactivated state of the channels and showing lower activity in the closed state of the channels.The selectivity index between the closed and inactivated states was about one hundred for the most promising inhibitors and about three hundred for the most potent compound 15b.None of the compounds inhibited the Na v 1.5 channel at a concentration of 30 μM.The structure-activity relationships of the compounds for inhibition of Na v 1.3 are described below.
Table 2 shows the inhibitory activity of compounds 12a-e, 13, 14a-i, 15a-b, and 19.Compounds 14a-i and 15a-b, with a thiazol-2-yl substituent attached to the sulfonamide nitrogen at the RHS, were about 20to 60-fold more effective than compounds 12a-e and 13 with an acetate group at that position (Scheme 4).This is evident when comparing
Compounds with the 1,4-disubstitution pattern on the central benzene ring were about 10 times more potent than their analogues with the 1,3-disubstitution pattern.This is evident when comparing 14a (IC   The most active compound in the series was 15b (IC 50 ; 0.020 μM) with a 2-(3,4-di昀氀uorophenyl)cyclopropane-1-carboxamide group at the LHS.The two 昀氀uorine substituents on the LHS phenyl ring slightly increased the activity compared with compound III (IC 50 ; 0.056 μM).
The 1,4-disubstitution pattern on the central benzene ring in compound III was signi昀椀cantly more optimal than the 1,3-disubstitution pattern in compound 15a (IC 50 ; 5.1 μM).Finally, replacement of the RHS sulfonamide bond in compound III with the amide bond in compound 19 (IC 50 ; 0.99 μM) resulted in twofold lower activity.Fig. 4 shows the example current traces of the resting-state (P1) and inactivated-state (P2) current for Na v 1.3 in the presence of 15b, DMSO and TTX (Fig. 4a), an overview of the voltage protocol diagram (Fig. 4b), and a concentration-response curve of the effects of 15b on the amplitude of the sodium currents observed during P1 and P2 (Fig. 4c).
Table 2 shows the inhibitory activities of the prepared acyl sulfonamides 29a-d, 30, 31a-c, 32a-c, and 36a-b in which N-(methylsulfonyl) carbamoyl or N-(N,N-dimethylsulfonyl)carbamoyl substituents were introduced at the RHS.Interestingly, this modi昀椀cation completely abolished the activity of the compounds at Na v 1.3 channels, as none of the prepared derivatives showed inhibitory activity at concentrations lower than 30 μM.
Table 3 shows the inhibitory activity of the compounds selected by a 3D ligand-based similarity search using Na v 1.3 inhibitor II (Fig. 1) as query.ZINC database was 昀椀rst screened using ROCS similarity searching software.Compounds in the hitlist were ranked according to the Tani-motoCombo score, which is a sum of ShapeTanimoto and Color-Tanimoto scores.In general, the highest ranked compounds had higher ShapeTanimoto than ColorTanimoto similarities.Therefore, ROCS hitlist was used further in EON similarity searching, which, in addition to the shape similarity, calculated the electrostatic similarities between the pre-aligned query molecule and compounds from the ROCS hitlist.This ligand-based virtual screening methodology resulted in a library of potential Na v 1.3 inhibitors similar in shape and electrostatic properties to inhibitor II.Sixteen of virtual hits were purchased and tested for Na v 1.3 inhibition (Table 3).The strongest inhibitors in the series were TSS-34 (IC 50 ; 0.32 μM) and TSS-42 (IC 50 ; 0.24 μM), both containing the N-(thiazol-2-yl)sulfonamide moiety at the RHS.These two compounds showed a state selectivity index of about one hundred and did not inhibit Na v 1.5 at 30 μM.Both TSS-34 and TSS-42 have a 1,4-disubstitution pattern on the central benzene ring and a hydrophobic group on the LHS, containing a chloro substituent attached to the aromatic ring.The third most potent compound from 3D similarity searching, TSS-39 (IC 50 ; 3.8 μM), also contains an N-(thiazol-2-yl)sulfonamide moiety at the RHS, suggesting that this structural feature is optimised for binding to VSD4 and for making contacts with the fourth arginine of the S4 helix.Replacement of the thiazol-2-yl substituent with other aromatic groups, e.g., 5-methylisoxazol-3-yl (TSS-37, IC 50 ; >30 μM), phenyl (TSS-41, IC 50 ; >30 μM), 2,6-dimethylphenyl (TSS-42, IC 50 ; >30 μM) and 3,4dimethylisoxazol-5-yl (TSS-48, IC 50 ; >30 μM), or aliphatic groups, e. g., acyl (TSS-35, TSS-36, and TSS-46, IC 50 ; >30, 21, and 15 μM, respectively) and pyrrolidin-1-yl (TSS-40, IC 50 ; 20 μM), resulted in lower Na v 1.3 inhibitory activity.
The binding modes of the most potent Na v 1.3 inhibitors in our series, 14i, 15b and TSS-42, were investigated by molecular docking (Fig. 5).The recently published cryo-EM structure of Na v 1.3 in complex with the inhibitor ICA121431 (PDB entry 7W7F) [16] bound to the VSD of DIV  was used for docking calculations using the Glide XP protocol (Schrödinger LLC).The N-(thiazol-2-yl)sulfonamide moiety was placed in the same binding conformation and orientation as in the case of inhibitor ICA121431.In the binding pocket, it formed ionic interactions with the Arg1627 and Arg1630 side chains, while additional hydrogen bonds were formed between the thiazole nitrogen atom and the Arg1630 guanidinium group, and the sulfonamide oxygen atom and the Asn1562 side chain.In addition, there were also hydrophobic contacts with Ala1626.This interaction network is the key determinant for the potency of the inhibitors.The 1,4-substituted phenyl ring formed hydrophobic contacts with Met1604, while the additional lipophilic substituent of 14i, 15b and TSS-42 stabilized the binding conformation of the inhibitor by interacting with Leu1563 and Phe1605.Inhibitory activity on other Na v isoforms.For the selected most promising inhibitors of Na v 1.3 (14f, 14g, 14i, 15b and 19), activities on Na v 1.7 channels were assessed using an automated patch clamp assay.Channels were expressed in CHO cells, and the assay was performed as described in Experimental section.The results are shown in Table 4.The compounds showed state-dependent activity as there was an enhanced level of block for the inactivated state of the channels and reduced potency against the closed state of the channel.Compounds 14f, 14g, 14i and 15b had IC 50 values between 7 and 9 μM, while compound 19 was inactive up to 30 μM.Overall, activity on Na v 1.7 was about 20 times lower for compounds 14f, 14g, and 14i than on Na v 1.3 and about 350 times lower for the strongest compound 15b.To investigate the statedependent effects of selected compounds on Na v 1.5 channels, we performed a similar experimental protocol as for Na v 1.3 and Na v 1.7 channels.The results are shown in Table 4.The compounds showed state-dependent activity on Na v 1.5 by inhibiting only the inactivated state of the channel and exerting no activity on the resting state up to 30 μM.For compounds 14f, 14g, 14i and 15b, the IC 50 values measured in the inactivated state ranged from 3 to 4 μM.For compounds 14f, 14g and 14i, the activity on Na v 1.5 was about 10 times lower than on Na v 1.3, and for the most active compound 15b, the activity was 150 times lower.
To further investigate the Na v 1.x selectivity, for the most promising inhibitors of the Na v 1.3 channels with IC 50 values below 10 μM (12c, 12e, 13, 14b-i, 15a-b, 19, TSS-34, TSS-39, TSS-42), and for some other selected compounds (29a-d, 30, 31a-c, 32a-c, 36a-b, TSS-33), the inhibitory activities on the closed state of the Na v 1.1, Na v 1.2, Na v 1.3, Na v 1.4, Na v 1.5, Na v 1.6, Na v 1.7, and Na v 1.8 channel isoforms were determined (Supplementary information, Table 2S).For these experiments, Na v channels were expressed in Xenopus laevis oocytes, and the two-electrode voltage-clamp method was used for electrophysiology experiments.All compounds were tested at a concentration of 1 μM against the closed state of the sodium channels, and at this concentration, none of the compounds displayed any activity against the Na v 1.1-Na v 1.2 or Na v 1.4-Na v 1.8 channel isoforms.The same two-electrode voltage-clamp method was used to determine the activity of selected compounds (14d, 14f, 14g, 14i, 15b, TSS-34, TSS-42) also against the closed state of the Na v 1.3 channel (Supplementary information, Table 2S).Representative whole-cell current traces in the presence of compound 15b, the voltage protocol used for the experiments, and the bar graph showing the percent inhibition of Na v 1.3, Na v 1.6, and Na v 1.7 current by 1 μM 15b are shown in Fig. 1S (Supplementary information).
The most potent compound in the series was compound 15b with a moderate 9% blockade of the resting-state current at 1 μM.However, since we have already shown that the compounds act by blocking the inactivated state of the channels, we decided to also evaluate the statedependent effects.To determine state-dependent inhibition, the    activities of selected compounds (14d, 14f, 14g, 14i, 15b, TSS-34, TSS-42) at Na v 1.3, Na v 1.7, and Na v 1.8 channels expressed in Xenopus laevis oocytes were determined by the two-electrode voltage-clamp method using a standard two-pulse protocol as described for the automated patch-clamp electrophysiology assay (Supplementary information, Table 3S).The compounds showed no activity at Na v 1.None of the compounds showed signi昀椀cant cytotoxic activity at a concentration of 50 μM (Supplementary information, Table 4S).

Conclusion
The Na v 1.3 channel is mainly expressed in the CNS of the embryonic brain and has been shown to be upregulated after nerve injury, and mutations in Na v 1.3 are associated with childhood epilepsies and developmental encephalopathies [27,28].Few selective Na v 1.3 inhibitors are known in the literature, so the discovery of new and potent inhibitors of this channel would allow us to study its role in more detail.Selective Na v 1.3 inhibitors could potentially be used as novel therapeutics to treat pain or neurodevelopmental disorders.In this work, we developed a series of novel Na v 1.3 inhibitors by using known Na v 1.3 inhibitors I-III as starting compounds.A total of 31 compounds were synthesized and 16 compounds were selected based on 3D ligand-based similarity search.All active compounds acted as state-dependent inhibitors of Na v 1.3 by inhibiting the inactivated state of the channel and exerting signi昀椀cantly lower activity on the closed form of the channel.
Eight compounds had an IC 50 value of less than 1 μM, and one of the compounds had an IC 50 value of 20 nM.None of the compounds showed use-dependent inhibition of the cardiac isoform Na v 1.5 at a concentration of 30 μM.Selectivity was also determined for the closed state of Na v channel isoforms (Na v 1.1-Na v 1.8).Whereas compounds displayed no effect on the closed state of Na v 1.1-Na v 1.2 or Na v 1.4-Na v 1.8 channel isoforms, compound 15b displayed small, yet selective, effects on the  The most important part of the molecule for the good inhibitory effect on Na v 1.3 proved to be the N-(thiazol-2-yl)sulfonamide moiety at the RHS.At the central benzene ring, the 1,4-disubstitution pattern was most optimal.Various lipophilic groups were tolerated at the LHS, usually containing a terminal aromatic group that could contain halogen substituents.The most potent inhibitor, 15b, contained a 2-(3,4-di昀氀uorophenyl)cyclopropane-1-carboxamide group at the LHS.The activity was also good for the molecules with a benzylated pyrrole ring at this position.Thus, the LHS of the inhibitors seems to be the most suitable part for further optimization.Overall, the Na v 1.3 inhibitors we discovered in this work could be used as tools for studying the biological role of this Na v isoform or for their further development into more potent and selective inhibitors.

Experimental section
Electrophysiology.Automated patch clamp.The synthesized compounds were evaluated for their inhibitory effects on human voltagegated sodium channels Na v 1.3 and Na v 1.5 using the automated patch clamp electrophysiology technique on the Sophion QPatch HT system (Sophion Bioscience A/S).The IC 50 values of the compounds were calculated from concentration-response curves measured at four relevant concentrations between 0.3 and 10 μM.Cells were detached from T175 cell culture 昀氀asks with trypsin-EDTA (0.05%) and stored in serumfree media on board the QPatch HT system.Cells were sampled, washed, and resuspended in extracellular recording solution by the QPatch HT before being added to the wells of the chip.0.1% DMSO v/v solution was applied to the cells to achieve stable control recording (4 min total), which was completed by applying test sample concentrations (4 min incubation per test concentration).Samples were prepared in extracellular solution with serial dilutions ranging from 10 to 0.3 μM concentration.Measurements on Na v 1.3 were performed using a standard twopulse voltage protocol.Starting from a holding potential of −100 mV, a 20 ms activating step was applied to −20 mV to measure the effect of the compounds on resting-state block.The second activating pulse was applied to the half-inactivation potential after a 5-s prepulse to assess block on the inactivated state of the channel.This protocol was applied with an interval of 0.067 Hz.To measure the Na v 1.5 isoform, 10 pulses were applied from −20 mV to a holding potential of −100 mV at 1 Hz.This protocol was applied with an interval of 0.016 Hz for the entire duration of the experiment.Peak inward current measurement of Na v 1.3 was performed for both the closed and inactivated test pulses for each sweep and for Na v 1.5 from the 10th pulse.Dimethyl sulfoxide (DMSO) was used as a control, and its concentration was kept constant under all conditions.Data were recorded using QPatch assay software (v5.0).Percent peak current inhibition was calculated as the mean peak current value for the last three sweeps measured in each concentration test period relative to the last three sweeps recorded during the vehicle control period.Sigmoidal concentration-response curves were 昀椀tted to the inhibition data using Xl昀椀t (IDBS).Data are presented as mean ± standard deviation for at least 3 independent observations.Two-electrode Voltage Clamp.Expression of voltage-gated ion channels in Xenopus laevis oocytes.For the expression of Nav channels, including hNa v 1.1, rNa v 1.2, rNa v 1.3, rNa v 1.4, hNa v 1.5, mNa v 1.6, hNa v 1.7, hNa v 1.8, together with auxiliary subunits rβ1 and hβ1, in Xenopus oocytes, the linearized plasmids were transcribed using the T7 or SP6 mMESSAGE mMACHINE transcription kit (Ambion®, Carlsbad, California, USA).Xenopus laevis oocytes at stage V-VI were isolated by partial ovariectomy as previously described [48].Animals were anesthetized by immersion in 0.1% tricaine methanesulfonate solution (Sigma®) (pH 7.0) for 15 min.The isolated oocytes were defolliculated with 1.5 mg/mL collagenase.Into the defolliculated oocytes, 50 nL of cRNA was injected at a concentration of 1 ng/nL using a microinjector (Drummond Scienti昀椀c®, Broomall, Pennsylvania, USA).Oocytes were incubated in a solution containing (in mM): NaCl, 96; KCl, 2; CaCl 2 , 1.8; MgCl 2 , 2; and HEPES, 5, at a pH of 7.4 supplemented with 50 mg/L gentamycin sulfate.Frogs were used in accordance with license number LA1210239 of the Laboratory of Toxicology and Pharmacology, University of Leuven.All animal care and experimental procedures were in accordance with the guidelines of the "European Convention for the Protection of Vertebrate Animals used for Experimental and other Sci-enti昀椀c Purposes" (Strasbourg, 18. III.1986).
Electrophysiological recordings.Two-electrode voltage-clamp recordings were performed at room temperature (18-22 • C) using a Geneclamp 500 ampli昀椀er (Molecular Devices®, Downingtown, Pennsylvania, USA) controlled by a pClamp data acquisition system (Axon Instruments®, Union City, California, USA).Whole-cell currents of oocytes were recorded 1-4 days after mRNA injection.The composition of the bath solution was (in mM): NaCl, 96; KCl, 2; CaCl 2 , 1.8; MgCl 2 , 2; and HEPES, 5, at pH 7.4.The voltage and current electrodes were 昀椀lled with 3 M KCl.The resistances of the two electrodes were kept between 0.8 and 1.5 MΩ.The elicited currents were sampled at 20 kHz and 昀椀ltered at 2 kHz using a four-pole low-pass Bessel 昀椀lter.Leak subtraction was performed with a -P/4 protocol.For electrophysiological analysis of compounds, a series of protocols were performed at a holding potential of −90 mV.Na + current traces were elicited by 100 ms depolarizations to Vmax (the voltage corresponding to the maximum Na + current under control conditions).The effects of the compounds on steady-state inactivation were examined using a standard 2-step protocol.In this protocol, 100-ms conditioning 5-mV step prepulses ranging from −90 to 60 mV were followed by a 50-ms test pulse to −10 mV.Compounds 14d, 14f, 14g, 14i, 15b, TSS -34 and TSS -42 were also measured on Na v 1.3, Na v 1.7 and Na v 1.8 using a similar standard two-pulse voltage protocol as described for the automated patch clamp experiments.All data are presented as mean ± standard deviation (SD) of at least 5 independent experiments (n ≥ 5).All data were tested for normality with a D'Agustino Pearson omnibus normality test.All data were tested for statistical signi昀椀cance with the Bonferroni test or the Dunn test.Data were analyzed using pClamp Clamp昀椀t 10.0 (Molecular Devices®, Downingtown, Pennsylvania, USA) and Origin 7.5 software (Originlab®, Northampton, Massachusetts, USA).
3D similarity searching.A library of drug-like molecules was downloaded from the ZINC database [49].For these compounds, a library of conformers was generated using the OMEGA software (OMEGA 2.5.1.4:OpenEye Scienti昀椀c Software, Santa Fe, NM. http://www.eyesopen.com)[50].Default settings were used to generate the conformers, resulting in a maximum of 200 conformers per ligand.
3D similarity search was 昀椀rst implemented in ROCS (ROCS 3.3.1.2:OpenEye Scienti昀椀c Software, Santa Fe, NM. http://www.eyesopen.com)[51], using the compound II (Fig. 1) as a query.ROCS represents atoms as 3D Gaussian functions [52,53] and calculates similarity as a function of the volume overlaps between alignments of the pre-generated molecular conformers.Chemical ("color") similarity is measured using overlaps between dummy atoms that mark chemical functionalities of interest: Hydrogen bond donors and acceptors, charged functional groups, rings, and hydrophobic groups.The similarity scores for shape (molecular geometry) and color (presence of relevant pharmacophores) are usually combined into a single score (TanimotoCombo) that can be used to rank screening molecules against a query molecule [54].Ten thousand highest ranked compounds from the ROCS similarity search were used for virtual screening using EON (EON 2.3.1.2:OpenEye Sci-enti昀椀c Software, Santa Fe, NM. http://www.eyesopen.com),which calculates the electrostatic similarity between two compounds in the form of an Electrostatic Tanimoto (ET) score.
Visualization of the hit list of 1000 compounds from EON virtual screening was performed using VIDA software (VIDA 4.3.0.4:OpenEye Scienti昀椀c Software, Santa Fe, NM. http://www.eyesopen.com).Hits were ranked according to the ET_combo score, which is the sum of the EON ShapeTanimoto and ET_pb scores (Supplementary information, Table 1S).Based on these results, 16 compounds were purchased (Table 3) and tested for Na v 1.3 channel inhibition.
Molecular docking.Ligand structures in SMILES format were opened in Maestro (Schrödinger, LLC, New York, NY, USA, 2020).Energy-minimized conformations of compounds 14i, 15b, and TSS-42 were generated with the LigPrep wizard using the OPLS3 force 昀椀eld.The protonation states of the ligands were calculated at pH 7.4 using Epik.Stereoisomers were generated for racemic compounds 15b and TSS-42.Molecular docking calculations were performed using Schrödinger Release 2020-1 (Schrödinger, LLC, New York, NY, USA, 2020).The cryo-EM structure of Na v 1.3 in complex with the inhibitor ICA121431 (PDB entry: 7W7F) [16] was retrieved from the Protein Data Bank.The protein was then prepared using the Protein Preparation Wizard with default settings.The receptor grid was calculated for the ligand-binding site, and compounds 14i, 15b and TSS-42 were docked using the Glide XP protocol, as implemented in Schrödinger Release 2020-1 (Glide, Schrödinger, LLC, New York, NY, USA, 2020).The highest scored docking pose of each compound was used for presentation.The 昀椀gures were created in PyMOL.
In vitro cytotoxicity measurements.Cytotoxicity of selected compounds at a concentration of 50 μM was determined using the MTS assay Madison, WI, USA) was added to the wells, and the plates were incubated for an additional 3 h.Absorbance at 490 nm was measured using a Synergy H4 microplate reader (BioTek, Winooski, VT, USA).To determine cell viability, the results of wells containing cells treated with the test compound were normalized with the results of cells incubated in 0.5% DMSO.Statistical signi昀椀cance (p < 0.05) was calculated using a two-tailed Student's t-test between the treated groups and 0.5% DMSO.Independent experiments were performed in triplicate and repeated twice.Results are expressed as the mean values of the independent measurements.Chemistry.Chemicals were purchased from Apollo Scienti昀椀c (Stockport, UK), TCI (Tokyo, Japan), Sigma-Aldrich (St. Louis, USA), and Acros Organics (Geel, Belgium).Thin-layer chromatography (TLC) was performed on Merck 60 F254 silica gel plates (0.25 mm) under visualization with UV light and spray reagents.Column chromatography was performed on silica gel 60 (particle size 240-400 mesh).IR spectra were recorded on a Thermo Nicolet Nexus 470 ESP FT-IR spectrometer (Thermo Fisher Scienti昀椀c, Waltham, USA).HPLC analyses were performed on an Agilent Technologies 1100 instrument (Agilent Technologies, Santa Clara, USA) with a G1316A thermostat, a G1313A autosampler, and a G1365B UV-Vis detector using a Phenomenex Luna 5-μm C18 column (4.6 × 150 mm or 4.6 × 250 mm, Phenomenex, Torrance, USA) and a 昀氀ow rate of 1.0 mL/min.The eluent consisted of tri昀氀uoroacetic acid (0.1% in water) as solvent A and acetonitrile as solvent B. Melting points were determined using a Reichert hot stage microscope and are uncorrected. 1H and 13 C NMR spectra were recorded at 400 and 100 MHz, respectively, using a Bruker AVANCE III spectrometer (Bruker Corporation, Billerica, USA) in CDCl 3 or DMSO-d solutions with TMS as the internal standard.Mass spectra were recorded using a VG Analytical Autospec Q mass spectrometer (Fisons, VG Analytical, Manchester, UK).The purity of the tested compounds was determined to be ≥ 95%.

General procedure B synthesis of compounds 2b-c and 2e (with 2b as an example)
A solution of compound 1b (1.00 g, 4.71 mmol) in dry DMF (5 mL) was cooled on an ice bath, sodium hydride (60% dispersion in mineral oil, 207 mg, 5.18 mmol) was added portion wise and the obtained mixture was stirred for 0.5 h.A solution of benzyl bromide (0.560 mL, 4.71 mmol) in DMF (1 mL) was added dropwise and the mixture was stirred at rt for 2 h.Ethyl acetate (50 mL) was added to the solution, the organic phase was washed with water (2 × 20 mL), 10% citric acid (2 × 20 mL) and brine (2 × 15 mL), dried over Na 2 SO 4 , 昀椀ltered and evaporated under reduced pressure.The crude product was puri昀椀ed with 昀氀ash column chromatography using ethyl acetate/petroleum ether (1:10) as solvent, to obtain 2b (0.973 g) as a brown oil.

General procedure C. Synthesis of compounds 3a and 3d-e (with 3a as an example)
Compound 2a (0.630 g, 3.13 mmol) was dissolved in tetrahydrofuran (10 mL), 2 M NaOH (6.26 mL, 12.5 mmol) was added and the mixture was heated at 50 • C for 15 h.The mixture was neutralized with 1 M HCl and concentrated under reduced pressure, the residual aqueous solution was acidi昀椀ed to pH 2 with 1 M HCl and the product extracted with ethyl acetate (2 × 10 mL).The combined organic phases were washed with 0.1 M HCl (2 × 10 mL) and brine (2 × 10 mL), dried over Na 2 SO 4 , 昀椀ltered and the solvent evaporated under reduced pressure to afford 3a as white solid (533 mg).

General procedure D. Synthesis of compounds 3b-c (with 3b as an example)
Compound 2b (0.913 g, 3.02 mmol) was dissolved in tetrahydrofuran (10 mL), 2 M NaOH (4.53 mL, 9.05 mmol) was added and the mixture was stirred at rt for 5 h.The mixture was neutralized with 1 M HCl and concentrated under reduced pressure, the residual aqueous solution was acidi昀椀ed to pH 2 with 1 M HCl and the product extracted with ethyl acetate (2 × 10 mL).The combined organic phases were washed with 0.1 M HCl (2 × 10 mL) and brine (2 × 10 mL), dried over Na 2 SO 4 , 昀椀ltered and the solvent evaporated under reduced pressure to afford 3a as white crystals (420 mg).

N-(N,N-Dimethylsulfamoyl)-4-nitrobenzamide (22b)
To a solution of compound 21b (1.20 g, 9.67 mmol) in dry tetrahydrofuran (10 mL) cooled on an ice bath sodium hydride (387 mg, 9.67 mmol, 60% dispersion in mineral oil) was added portionwise.After 30 min a solution of 4-nitrobenzoyl chloride (1.20 g, 6.44 mmol) in tetrahydrofurane 10 mL was added dropwise and the mixture was stirred for 1 h at rt and for 15 h at 50 • C. The solvent was removed under reduced pressure and to the residue ethyl acetate (40 mL) and 0.5 M HCl (20 mL) were added.The layers were separated, the organic phase was washed with 0.5 M HCl (2 × 20 mL) and brine (20 mL), dried over Na 2 SO 4 , 昀椀ltered and evaporated under reduced pressure.To the solid residue ether (20 mL) was added, the suspension was sonicated, the solid was 昀椀ltered off, washed with ether and dried.The crude product was puri昀椀ed with 昀氀ash column chromatography using dichloromethane/methanol (20:1) as solvent, to obtain 22b (0.700 g) as white crystals.Yield 40% (700 mg); mp 159-164

4-Amino-N-(methylsulfonyl)benzamide (23a)
To a solution of compound 22a (200 mg, 0.934 mmol) in a mixture of ethanol (5 mL) and glacial acetic acid (5 mL) under argon Pd-C (50 mg) was added and the reaction mixture was stirred under hydrogen atmosphere for 15 h.The catalyst was 昀椀ltered off and the solvent was removed under reduced pressure.To the residue ethyl acetate (100 mL) and 0.1 M HCl (10 mL) were added, the phases were separated and organic phase was washed with brine (5 mL), dried over Na 2 SO 4 , 昀椀ltered and the solvent evaporated under reduced pressure to obtain product 23a (456 mg) as an off-white solid.Yield 90% (158 mg); mp 160-164

4-Amino-N-(N,N-dimethylsulfamoyl)benzamide (23b)
To a solution of compound 22a (0.692 mg, 2.531 mmol) in methanol (50 mL) under argon Pd-C (200 mg) was added and the reaction mixture was stirred under hydrogen atmosphere for 15 h.The catalyst was 昀椀ltered off and the solvent was removed under reduced pressure.To the crude product ether (10 mL) was added, the suspension was sonicated and the solid was 昀椀ltered off, washed with ether (5 mL) and dried to obtain product 23b (585 mg) as a white solid.Yield 95% (585 mg); mp 130-134

General procedure G. Synthesis of compounds 27a and 27b (with 27a as an example)
To a solution of compound 26a (1.52 g, 7.08 mmol) in a mixture of methanol (10 mL) and water (2 mL) 2 M LiOH (5.31 mL, 10.6 mmol) was added dropwise and the mixture was stirred at rt for 15 h.The solvent was removed under reduced pressure, to the residue ethyl acetate (20 mL) and 0.5 M HCl (20 mL) were added, the phases were separated, organic phase was washed with 0.5 M HCl (20 mL) and brine (10 mL), dried over Na 2 SO 4 , 昀椀ltered and the solvent evaporated under reduced pressure to obtain 27a (1.20 g) as white crystals.[67,68] Synthesized according to General procedure G. White crystals; yield 84% (1.

General procedure H. Synthesis of compounds 34a and 34b (with 34a as an example)
To a solution of methyl 3-hydroxybenzoate (312 mg, 2.05 mmol) and triphenylphosphine (0.808 g, 3.08 mmol) in dry THF (20 mL) compound 33 (0.500 g, 3.08) and DEAD (1.69 mL, 3.70 mmol) were added consecutively.The mixture was stirred at rt for 3 h and then heated at 50 • C for 15 h.The solvent was removed under reduced pressure and the residue was puri昀椀ed with 昀氀esh column chromatography using ethyl acetate/petroleum ether (1:10) as solvent, to obtain 34a (290 mg) as a pink oil.

General procedure I. Synthesis of compounds 36a and 36b (with 36a as an example)
To a solution of compound 35a (100 mg, 0.372 mmol) in dry dichloromethane (5 mL) oxalyl chloride (2 M solution in dichloromethane, 0.410 mL, 0.818 mmol) was added and the mixture was stirred at rt for 15 h.The solvent was removed under reduced pressure and to the residue dry tetrahydrofuran (5 mL) was added to obtain solution A. In a separate 昀氀ask, compound 21a (71 mg, 0.745 mmol) was dissolved in dry tetrahydrofuran (15 mL) and to the obtained solution sodium hydride (403 mg, 10.5 mmol, 60% dispersion in mineral oil) was added portionwise at 0 • C.After 30 min solution A was added dropwise and the mixture was stirred for 1 h at rt and for 15 h at 50 • C. The solvent was removed under reduced pressure and to the residue ethyl acetate (30 mL) and 0.5 M HCl (10 mL) were added.The layers were separated, the organic phase was washed with 0.5 M HCl (2 × 10 mL) and brine (10 mL), dried over Na 2 SO 4 , 昀椀ltered and evaporated under reduced pressure.To the solid residue ether (10 mL) was added, the suspension was sonicated, the solid was 昀椀ltered off, washed with ether and dried to obtain 35a (35 mg) as a white solid.

Declaration of competing interest
The authors declare that they have no known competing 昀椀nancial interests or personal relationships that could have appeared to in昀氀uence the work reported in this paper.

Fig. 2 .
Fig. 2. Proposed structural modi昀椀cations of Na v 1.3 inhibitors I and II for the design of improved aryl-and acylsulfonamide Na v 1.3 inhibitors (type I compounds).

N
.Zidar et al.
7 and Na v 1.8 channels up to a concentration of 1 μM, con昀椀rming their isoform selectivity.However, the compounds blocked Na v 1.3 current in the inactivated state (Peak 2) with 21-63% inhibition at 1 μM.The most active compound was 15b with a 63% blockade of the inactivated-state current at 1 μM.Since this compound blocked the resting-state current (Peak 1) to a lesser extent (9% block at 1 μM), we con昀椀rmed its promising state-dependent activity.Therefore, compounds with low IC 50 values at Na v 1.3 channels (e.g., 14d, 14f, 14g, 14i, 15b, TSS-34, TSS-

Fig. 4 .
Fig. 4. Characterization of compound 15b as an inhibitor of the Na v 1.3 channel.Compound 15b was applied to CHO cells stably expressing the Na v 1.3 channel to determine its potency and state dependence.a) Example current traces of resting-state (P1) and inactivated-state current (P2, after a 5-s prepulse to half of the inactivation potential) for a typical recording showing concentration-dependent inhibition of inward currents (black traces = vehicle, blue traces = 300 nM 15b, red traces = 300 nM TTX).The scale bar shows 2 nA on the y-axis and 20 ms on the xaxis.b) An overview of the voltage protocol diagram.c) A concentration-response curve of the effects of 15b on the amplitude of sodium currents observed during P1 and P2.

Fig. 5 .
Fig. 5. Binding modes of a) 14i (in magenta sticks), b) 15b (in green sticks) and c) TSS-42 (in yellow sticks) in the Na v 1.3 VSD of DIV (in grey cartoon, PDB entry: 7W7F).For clarity, only amino acids forming hydrogen bonds (black dashed lines), ionic interactions and hydrophobic interactions with inhibitors are shown as sticks.

Table 1 Inhibitory activities of compounds 12a-e, 13, 14a-i, 15a-b and 19 on
human Na v 1.3 channels (state-dependent inhibition) and Na v 1.5 channels (use-dependent inhibition) expressed in CHO cells determined using the QPatch.

Peak 1 a Peak 2 b State Selectivity (Peak 1/Peak 2) IC 50 ratio d N e Pulse 1 f Pulse 10 g N IC 50 (μM) c
(continued on next page) N.Zidar et al.

Table 2
Inhibitory activities of compounds 29a-d, 30, 31a-c, 32a-c and 36a-b on human Na v 1.3 channels (state-dependent inhibition) and Na v 1.5 channels (use-dependent inhibition) expressed in CHO cells determined using the QPatch.

Table 3
Inhibitory activities of compounds identi昀椀ed with similarity search on human Na v 1.3 channels (state-dependent inhibition) and Na v 1.5 channels (use-dependent inhibition) expressed in CHO cells determined using the QPatch.
(continued on next page) N.Zidar et al.

Table 3
(continued ) a Resting-state Na v 1.3 current.b Inactivated-state Na v 1.3 current.c Concentration of compound that inhibits the channel current by 50%.d Ratio between resting-and inactivated-state IC 50 valuesa measure of state-dependent inhibition of Na v 1.3.e Number of independent experiments.f Na v 1.5 tonic 1 Hz pulse 1 QPatch potency.g Na v 1.5 phasic pulse 10 QPatch potency.h Not determined.i Tetrodotoxin.j Amitriptyline.

Table 4
Inhibitory activities of compounds 14f

-g, 14i, 15b and 19 on
human Na v 1.7 and Na v 1.5 channels (state-dependent inhibition) expressed in CHO cells determined using the QPatch.Na v 1.3 channel isoform.In addition, the activities of seven selected compounds on Na v 1.3, Na v 1.7, and Na v 1.8 channels expressed in Xenopus laevis oocytes were determined by the two-electrode voltage clamp method with the same two-pulse protocol used to assess Nav1.3 channel activity in the automated patch-clamp experiments.The compounds showed robust effects on the inactivated state of the Na v 1.3 channel at 1 μM, but no effect on the inactivated state of the Na v 1.7 or Na v 1.8 channel, demonstrating that the compounds are selective for the inactivated state of Na v 1.3 channel when tested among the Na v 1.3, Na v 1.7, and Na v 1.8 channel isoforms.On the other hand, compounds showed activity on Na v 1.3 channels expressed in oocytes, albeit with slightly lower potency compared with the activity observed in patch clamp experiments.
a Resting-state current.b Inactivated-state current.c Concentration of compound that inhibits the channel current by 50%.d Ratio between resting-and inactivated-state IC 50 valuesa measure of state-dependent inhibition.e Number of independent experiments.f Not determined.g Amitriptyline.