Inhibition of human N‐ and T‐type calcium channels by an ortho‐phenoxyanilide derivative, MONIRO‐1

Abstract

Background and Purpose

Voltage‐gated calcium channels are involved in nociception in the CNS and in the periphery. N‐type (Ca_v2.2) and T‐type (Ca_v3.1, Ca_v3.2 and Ca_v3.3) voltage‐gated calcium channels are particularly important in studying and treating pain and epilepsy.

Experimental Approach

In this study, whole‐cell patch clamp electrophysiology was used to assess the potency and mechanism of action of a novel ortho‐phenoxylanilide derivative, MONIRO‐1, against a panel of voltage‐gated calcium channels including Ca_v1.2, Ca_v1.3, Ca_v2.1, Ca_v2.2, Ca_v2.3, Ca_v3.1, Ca_v3.2 and Ca_v3.3.

Key Results

MONIRO‐1 was 5‐ to 20‐fold more potent at inhibiting human T‐type calcium channels, hCa_v3.1, hCa_v3.2 and hCa_v3.3 (IC₅₀: 3.3 ± 0.3, 1.7 ± 0.1 and 7.2 ± 0.3 μM, respectively) than N‐type calcium channel, hCa_v2.2 (IC₅₀: 34.0 ± 3.6 μM). It interacted with L‐type calcium channels Ca_v1.2 and Ca_v1.3 with significantly lower potency (IC₅₀ > 100 μM) and did not inhibit hCa_v2.1 or hCa_v2.3 channels at concentrations as high as 100 μM. State‐ and use‐dependent inhibition of hCa_v2.2 channels was observed, whereas stronger inhibition occurred at high stimulation frequencies for hCa_v3.1 channels suggesting a different mode of action between these two channels.

Conclusions and Implications

Selectivity, potency, reversibility and multi‐modal effects distinguish MONIRO‐1 from other low MW inhibitors acting on Ca_v channels involved in pain and/or epilepsy pathways. High‐frequency firing increased the affinity for MONIRO‐1 for both hCa_v2.2 and hCa_v3.1 channels. Such Ca_v channel modulators have potential clinical use in the treatment of epilepsies, neuropathic pain and other nociceptive pathophysiologies.

Linked Articles

This article is part of a themed section on Recent Advances in Targeting Ion Channels to Treat Chronic Pain. To view the other articles in this section visit http://onlinelibrary.wiley.com/doi/10.1111/bph.v175.12/issuetoc

Abbreviations

hERG

human ether à‐go‐go related gene

SSI

steady‐state inactivation

VGCCs

voltage‐gated calcium channels

Introduction

Sensory nerve conduction relies on the coordinated activation of a suite of ion channels to generate fast and accurate responses to mechanical, thermal and chemical stimuli. In the periphery, harmful signals are perceived by ‘nociceptive’ neurons and relayed centrally via voltage‐activated ion channels (http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=81, http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=82, http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=80) and ligand‐gated ion channels (http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=78, http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=79, http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=77) (Waxman and Zamponi, 2014) as part of the ‘pain‐signalling cascade’ (Gold and Gebhart, 2010). In chronic pain, these pathways are characterized by hyperactivity (Waxman and Zamponi, 2014), and the effects of noxious stimuli are enhanced (hyperalgesia) to the point where non‐noxious stimuli are painful (allodynia) (see Guo and Hu, 2014). Voltage‐gated calcium channels (VGCCs) involved in pain pathways are localized to dendrites of primary sensory neurons and contribute to integrate nociceptive stimuli at the spine by regulating the release of pro‐nociceptive neurotransmitters, including substance P, calcitonin gene‐related peptide and glutamate, at the presynaptic nerve terminus (Simms and Zamponi, 2014). Therefore, modulating specific Ca_v subtypes could be a promising way to treat chronic pain.

Five distinct voltage‐gated calcium current types have been described in mammalian cells: L‐ (http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=528&familyId=80&familyType=IC, http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=529&familyId=80&familyType=IC, http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=530&familyId=80&familyType=IC and http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=531&familyId=80&familyType=IC), P/Q‐ (http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=532&familyId=80&familyType=IC), N‐ (http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=533&familyId=80&familyType=IC), R‐ (http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=534&familyId=80&familyType=IC) and T‐ (http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=535&familyId=80&familyType=IC, http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=536&familyId=80&familyType=IC, http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=537&familyId=80&familyType=IC) types (Catterall et al., 2005). Importantly, each Ca_v subtype has distinct electrophysiological properties, tissue‐specific localization and pharmacological profiles (Catterall et al., 2005). N‐type calcium channels have been a major focus of research because they are localized at the presynaptic terminals of sensory nerves in the dorsal horn of the spinal cord, where they play a critical role in pain signalling (Saegusa et al., 2001; Lee, 2014). As a result, Ca_v2.2 channels are strongly implicated in a range of chronic pain conditions including neuropathic pain (pain caused by injury to the nervous system). Specific Ca_v2.2 channel inhibitors have clinical applications (Altier and Zamponi, 2004; Perret and Luo, 2009; Bourinet et al., 2014), such as http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=2536 (Ziconotide). Ziconotide selectively targets Ca_v2.2 channels and is used to effectively treat intractable chronic pain conditions (Pope and Deer, 2013). However, due to its chemical nature, Ziconotide must be administered into the cerebrospinal fluid via the intrathecal route and has a relatively narrow therapeutic window (Pope and Deer, 2013).

T‐type calcium channels modulate neuronal excitability in the peripheral and central nervous systems, where they are linked to hyperexcitable states such as epilepsy, pain and tremor disorders like Parkinson's disease (Bourinet et al., 2005; Choi et al., 2007; Kim et al., 2001). Previously characterized T‐type calcium channels inhibitors, such as mibefradil and ethoximide, have analgesic properties, suggesting that T‐type channels may be involved in pain signalling (Matthews and Dickenson, 2001; Todorovic et al., 2002; Flatters and Bennett, 2004; Wen et al., 2010). Unfortunately, http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=2522 and ethoximide target a wide range of voltage‐gated ion channels, including sodium, potassium and chloride channels (Chouabe et al., 1998; Nilius et al., 1998; Eller et al., 2000).

To overcome this lack of specificity, small molecule T‐type calcium channel inhibitors were designed and have been shown to provide analgesic effects in mouse models of inflammatory and neuropathic pain (Kraus et al., 2010; Reger et al., 2011; Bladen et al., 2015). Of particular note is the analgesic, low MW compound http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=7724, which is more potent against inactivated than closed or resting channels (IC₅₀ values of 100 and 10 μM at Ca_v3.1 respectively), indicating its effects on targets are state‐dependent (Francois et al., 2013). Based on these findings, N‐ and T‐type calcium channels are of particular interest for the study and treatment of chronic pain (see recent reviews by Cheong and Shin, 2013; Lee, 2013; Sekiguchi and Kawabata, 2013; François et al., 2014; Lee, 2014; Simms and Zamponi, 2014).

During the last 5 years, numerous reports have described small molecule Ca_v2.2 channel inhibitors as alternatives to Ziconotide (see review by Lee, 2014). Regrettably, clinical development of the most promising of these small‐molecules, Z160, a reformulated form of http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=7765 (Zamponi et al., 2009) (compound 1, Figure 1), was discontinued after it failed to meet end points in phase II clinical trials (Eller et al., 2000). Due to this and the need for effective and safe therapeutics for chronic pain, the development of new classes of small‐molecule Ca_v channel antagonists is of significant interest.

Figure 1.

Chemical structures of compounds 1 (Z160/NP118809), 2 (MONIRO‐1) and 3 (NNC 55‐0396) relevant to this study.

The pharmacophore of Ca_v2.2‐selective http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=2535 (McCleskey et al., 1987; Baell et al., 2004; Baell et al., 2006; Andersson et al., 2009; Duggan et al., 2009) has been carefully studied and has led to the identification of a series of compounds that block Ca_v2.2 channels in a concentration‐dependent manner (Tranberg et al., 2012).

We focused on developing low MW calcium channel inhibitors with enhanced biophysical properties including reversibility, voltage‐ and/or state‐dependence. Compounds acting via these mechanisms can strongly inhibit channels upon membrane depolarization (Hondeghem and Katzung, 1984; Butterworth and Strichartz, 1990) by preferentially targeting the open and/or inactivated states over the resting state. Such compounds would facilitate inhibition of high‐frequency firing or depolarised nerves that can predominate in chronic pain states and could act within a wider therapeutic window.

To develop these calcium channel inhibitors, we screened a small library of related derivatives, using FLIPR Tetra® high‐throughput cellular screening system, for their ability to inhibit Ca_v2.2‐mediated calcium responses in SH‐SY5Y cells using the fluorescent indicator Calcium 4 dye (Molecular Devices, Sunnyvale, CA, USA) (Gleeson et al., 2015). In this assay, MONIRO‐1 (compound 2, Figure 1) inhibited Ca²⁺ responses elicited by KCl‐mediated depolarization with an IC₅₀ of 124 μM. Subsequently, we characterized the ortho‐phenoxyanilide derivative MONIRO‐1 electrophysiologically. To directly assess the biophysical properties of MONIRO‐1 inhibition of VGCCs, we used the whole‐cell patch clamp recording technique to study a comprehensive array of Ca_v channels heterologously or stably expressed in HEK‐293 cells. MONIRO‐1 was 5–100 times more potent against T‐type calcium channels (Ca_v3.1, Ca_v3.2 and Ca_v3.3) than N‐type (Ca_v2.2) and L‐type (Ca_v1.2 and Ca_v1.3) channels. No inhibition of P/Q (Ca_v2.1) or R‐type currents (Ca_v2.3) was detected in the presence of up to 100 μM MONIRO‐1. Furthermore, MONIRO‐1 modulated Ca_v2.2 differently at various stimulation rates and in a state‐dependent fashion, preferentially inhibiting hyperactive channels such as those likely to predominate in pathologically augmented pain states.

Methods

Cell culture and transfection

HEK293 cells containing the SV40 Large T‐antigen (HEK293T) were cultured at 37°C, 5% CO₂ in Dulbecco's Modified Eagle's Medium (Invitrogen Life Technologies, Australia) supplemented with 10% fetal bovine serum (Invitrogen), 1% penicillin and streptomycin (Invitrogen). HEK293T cells stably expressing the human Ca_v2.1 (P/Q‐type) calcium channel (α_1A‐2 splice variant; GenBank accession no. AF004883.1), Ca_v2.2 (N‐type) calcium channel (transcript variant 1, NM_000718) or Ca_v2.3 (R‐type) calcium channel (neuronal α_1E‐3 splice variant, also called α_1E‐c, L29385.1) were obtained from Merck and cultured according to procedures described by Dai et al. (2008). These cell lines co‐express recombinant human Ca_v α_2bδ‐1 (M76559.1) and Ca_v β_3a (NM_000725) auxiliary subunits (Dai et al., 2008). Cells are plated onto 12 mm cover glass and incubated at 30°C, 5% CO₂ for 24–48 h prior to recording.

In a separate series of experiments, HEK293T cells were transiently co‐transfected with plasmid cDNAs encoding mouse Ca_v1.2 (provided by Dr D. Lipscombe, Addgene plasmid # 26572; Helton et al., 2005) or human Ca_v1.3 (provided by Dr G. Zamponi, University of Calgary) with α₂δ‐1 and β₃, Ca_v3.1 (provided by Dr G. Zamponi), human Ca_v3.2 (CACNA1HB; Origene Technologies, Rockville, MD), human Ca_v3.3 a1Ic‐HE3‐pcDNA3 was a gift from Dr E. Perez‐Reyes (Gomora et al., 2002) (Addgene plasmid # 45810) and K_v11.1 (human ether à‐go‐go related gene [hERG]) (provided by Dr J. Vandenberg, Victor Chang Cardiac Research Institute, Sydney). After transfection, cells were plated on glass coverslips and incubated at 37°C, 5% CO₂ overnight. The transfection medium was then replaced with culture medium, and cells were incubated at 30°C, 5% CO₂.

Electrophysiology

Electrophysiological experiments were carried out 1–3 days after transfection using the whole‐cell patch clamp recording technique. Cells were superfused with a solution containing 110 mM NaCl, 10 mM CaCl₂, 1 mM MgCl₂, 5 mM CsCl, 30 mM TEA‐Cl, 10 mM D‐glucose and 10 mM HEPES, pH 7.4 with TEA‐OH (~310 mOsmol kg⁻¹). Fire‐polished borosilicate patch pipettes with tip‐resistance values of 1–3 MΩ were filled with an intracellular solution containing 125 mM K⁺gluconate, 5 mM NaCl, 2 mM MgCl₂, 5 mM EGTA, 4 mM MgATP and 10 mM HEPES, pH 7.2 with CsOH (~300 mOsmol kg⁻¹). Depolarization‐activated calcium currents were recorded at room temperature (21–23°C) using a Multiclamp 700B patch clamp amplifier (Molecular Devices, Sunnyvale, CA, USA) controlled by Digidata 1322A/Clampex 9.2 acquisition systems. MONIRO‐1 stock solution was dissolved in 5% DMSO to a concentration of 5 mM. MONIRO‐1 was further diluted in recording solution to appropriate concentrations. Previous experiments using the maximal concentration of DMSO used, 0.1% at 100 μM, did not modulate VGCC currents in recording solution.

Depolarization‐activated calcium currents (I_Ca) were evoked from a holding potential of −80 mV, with a 135 ms (50 ms for T‐type) voltage step to a test potential (determined from peak of I‐V curve) applied every 10 s. When evaluating I_Ca amplitude in the presence and absence of various compounds, test depolarizations of varying durations were applied at a frequency of 0.03–2 Hz, from a holding potential of −60, −80 or −100 mV to a test potential determined from an I–V relationship to elicit peak inward current. To examine the effects of pulse duration on MONIRO‐1 block and recovery from block, pulses of either 45 or 135 ms for Ca_v2.2 (30 μM) or 15 or 50 ms for Ca_v3.1 (3 μM) were elicited at a frequency of 0.1 Hz (V_h = −80 mV). Similarly, to study the frequency dependence of MONIRO‐1 inhibition, test pulses (45 ms for Ca_v2.2 or 15 ms for Ca_v3.1) were elicited at rates of 0.03, 0.1 and 0.2 Hz for Ca_v2.2 (10 μM MONIRO‐1) and 0.03, 0.1, 0.2, 1 and 2 Hz for Ca_v3.1 (3 μM MONIRO‐1). To determine the time constants for activation and inactivation of macroscopic currents, we used a single exponential equation to fit the traces in control and after MONIRO‐1 application.

Activation curves were measured from test pulses from −80 to +80 mV every 5 mV for Ca_v2.2 and Ca_v3.1 channels (V_h = −80 mV) before and after MONIRO‐1 application. Steady‐state inactivation (SSI) curves were generated from various pre‐pulse potentials (−120 to 0 mV for Ca_v2.2; −85 to −20 mV for Ca_v3.1) of 10 s (Ca_v2.2) or 1 s (Ca_v3.1). Recovery from inactivation was examined using a double‐pulse protocol, varying the time between a fully inactivating pre‐pulse and the test pulse repeated every 15 s, using a concentration of MONIRO‐1, which inhibits ~50% of the current. Recovery data were best fit by a double exponential. Currents were filtered at 3 kHz and sampled at 10 kHz. Leak and capacitive currents were subtracted using a −P/4 pulse protocol. Data were stored digitally on a computer for further analysis.

Data analysis

The data and statistical analysis comply with the recommendations on experimental design and analysis in pharmacology (Curtis et al., 2015). All data reported as mean ± SEM. Data analysis and graphs were built using OriginPro (Origin Lab Corporation, North Hampton, MA, USA). Concentration–response curves were generated by plotting current in the presence of MONIRO‐1 (I_compound) over current prior to application (I_control) versus a range of MONIRO‐1 concentrations tested. Typically, four to six concentrations per channel were screened with five to nine cells per concentration. Concentration–response data were plotted using a rectangular hyperbola assuming a Hill coefficient of 1, using the following expression,

$$\left({\mathrm{I}}_{\text{compound}}/{\mathrm{I}}_{\text{control}}\right)=1/\left(1+{\mathrm{IC}}_{50}/\left[\text{drug}\right]\right),$$

where IC₅₀ is the concentration at which half‐maximal inhibition was observed.

Steady‐state activation and inactivation curves were generated using a modified Boltzmann equation,

$$\mathrm{G}=1/\left(1+exp\left(\left(\mathrm{Vm}-{\mathrm{V}}_{0.5}\right)/\mathrm{\kappa a}\right)\right),$$

where V_m is the test potential, V_0.5 is the half‐maximal activation potential and κa is the slope factor. SSA and SSI data were each recorded from five to seven individual experiments per channel examined. Recovery from inactivation curves were the results from five to six independent experiments where the complete set of recovery from inactivation time points were acquired. The data were fit using a double exponential of the following equation,

$$\mathrm{I}=1+{{\mathrm{A}}_{\text{fast}}}^{*}exp\left(-\mathrm{t}/{\mathrm{\tau }}_{\text{fast}}\right)+{{\mathrm{A}}_{\text{slow}}}^{*}exp\left(-\mathrm{t}/{\mathrm{\tau }}_{\text{slow}}\right),$$

where τ is the fast and slow time constant, t is the recovery time and A is the amplitude of the fast and slow time constants. Statistical significance (P < 0.05) was determined using unpaired Student's t‐tests or one‐way ANOVA with Tukey's post hoc test.

Materials

Compound 2, N‐(2‐(4‐fluorophenoxy)phenyl)‐4‐(3‐guanidinopropoxy)benzamide hydrochloride, MONIRO‐1, was synthesized from 2‐(4‐fluorophenoxy)aniline in four steps, as reported previously (Gleeson et al., 2015). ¹H NMR (400 MHz, CD³OD): δ 7.88–7.85 (m, 1H), 7.78–7.76 (m, 2H), 7.21–7.18 (m, 2H), 7.07–6.98 (m, 6H), 6.97–6.94 (m, 1H), 4.13 (t, J = 5.9 Hz, 2H), 3.41 (t, J = 6.8 Hz, 2H), 2.11–2.05 (m, 2H). ¹³C NMR (100 MHz, CD³OD): δ 168.2, 163.2, 160.2 (d, J = 239 Hz), 158.7, 154.4, 151.0, 130.5, 130.4, 127.9, 127.6, 126.9, 125.0, 120.8 (d, J = 8.3 Hz), 120.1, 117.2 (d, J = 23.4 Hz), 115.4, 66.3, 39.5, 29.5. ¹⁹F NMR (376 MHz, CD³OD): δ −120.6, purity >92%, determined by HPLC analysis (λ = 254 nm), all tests were performed using the same batch of compound. The positive control, compound 1 (Z160/NP118809) (Zikolova and Ninov, 1972; Zamponi et al., 2009), was prepared by N‐(3‐dimethylaminopropyl)‐N′‐ethylcarbodiimide‐assisted acylation of N‐benzhydryl piperazine with 3,3‐diphenylpropanoic acid. Control compound NNC 55–0396 dihydrochloride was provided by Alomone Labs (Jerusalem, Israel).

Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Southan et al., 2016), and are permanently archived in the Concise Guide to PHARMACOLOGY 2015/16 (Alexander et al., 2015).

Results

Selective inhibition of human N‐ and T‐type calcium channel subtypes

Whole‐cell calcium currents were elicited by depolarizing pulses at a frequency of 0.1 Hz from a holding potential of −80 mV. MONIRO‐1 reversibly inhibited channel currents mediated by Ca_v3.2 (IC₅₀ 1.7 ± 0.1 μM, n = 5–9 cells per concentration), Ca_v3.1 (IC₅₀ 3.3 ± 0.3 μM, n = 5–6 cells per concentration), Ca_v3.3 (IC₅₀ 7.2 ± 0.3 μM, n = 5 cells per concentration) and Ca_v2.2 (IC₅₀ 34.0 ± 3.6 μM, n = 5–7 cells per concentration) (Figures 2 and 3A). The effects of MONIRO‐1 were further investigated on a panel of VGCCs. At 100 μM, MONIRO‐1 had no apparent effect on Ca_v2.1 (n = 6) or Ca_v2.3 (n = 5) (Figure 3B) but inhibited Ca_v1.2 and Ca_v1.3‐mediated currents by 46 ± 1% (n = 5) and 27 ± 1% (n = 5) respectively.

Figure 2.

MONIRO‐1 inhibition of human Ca_v2.2‐ and Ca_v3.1‐mediated currents in HEK293 cells. (A) Superimposed whole‐cell Ca²⁺ currents recorded from HEK293 cells stably expressing human N‐type calcium (hCa_v2.2) channels (α_1B + β³ + α²δ1) in response to depolarizing pulses (135 ms duration) to +20 mV from a holding potential of −80 mV. Representative current traces obtained in the absence (a) and presence of 50 (b) and 100 μM (d) MONIRO‐1 and after respective washouts (c, e). (B) Peak inward Ca²⁺ current amplitude plotted as a function of time. Time course of inhibition and current amplitude recovery in the presence and following washout of 50 and 100 μM MONIRO‐1. Representative currents shown in (A) were obtained at the times indicated by the lower case letters. (C) Superimposed whole‐cell Ca²⁺ currents recorded from HEK293 cells transiently expressing human T‐type calcium channel, hCa_v3.1. These Ca²⁺ currents are in response to depolarizing pulses (50 ms duration) to −25 mV from a holding potential of −80 mV. Representative current traces obtained in the absence (a) and presence of 10 μM (b) and 100 μM (c) MONIRO‐1. (D) Peak inward Ca²⁺ current amplitude plotted as a function of time. Time course of inhibition and recovery of current amplitude in the absence (a) and presence of 10 (b) and 100 μM (c) MONIRO‐1.

Figure 3.

MONIRO‐1 inhibition of voltage‐gated calcium channels. (A) Concentration–response of MONIRO‐1 at hCa_v2.2, hCa_v3.1, hCa_v3.2 and hCa_v3.3 channels expressed in HEK293 cells. Fitted curves gave IC₅₀ values as shown in the Results. (B) Fraction of current inhibited by 100 (solid) or 10 μM (striped) MONIRO‐1 at Ca_v1.3, Ca_v2.1, Ca_v2.2, Ca_v2.3, Ca_v3.1 and Ca_v3.2 channels. Cells were held at −80 mV and pulsed at 0.1 Hz to the voltage determined to elicit the peak inward current previously determined from an I–V protocol. The number of cells is indicated in parentheses.

In comparison, Z160/NP118809 (Zamponi et al., 2009) inhibited Ca_v2.2‐mediated currents with an IC₅₀ of 0.28 ± 0.02 μM (n = 5–6 cells per concentration), without affecting Ca_v3.1 currents (maximal concentration tested 1 μM, Table 1). Application of 10 μM NNC 55‐0396, a T‐type calcium channel inhibitor, inhibited hCa_v3.2 currents by 96 ± 3% (n = 5). Furthermore, Z160/NP118809 inhibited http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=572 channels with an IC₅₀ of 7.4 μM (Zamponi et al., 2009) whereas 100 μM MONIRO‐1 did not inhibit hERG channel‐mediated K⁺ currents (Table 1). Table 1 compares the IC₅₀ values obtained for MONIRO‐1 inhibition of L‐, P/Q‐, N‐, R‐ and T‐type calcium channels and the hERG potassium channel, with that of the control compounds, Z160/NP118809 and http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=4269, and other known N‐ and T‐type calcium channel inhibitors where electrophysiology data were obtained across a large portion of the VGCC subtypes.

Table 1.

Comparison of IC₅₀ values obtained for MONIRO‐1 inhibition of L‐, P/Q‐, N‐, R‐ and T‐type calcium channels and the hERG potassium channel, with control compounds Z160/NP118809 and NNC 55‐0396 and other known N‐ and T‐type calcium channel inhibitors

	L‐type (μM)	P/Q‐type (μM)	N‐type (μM)	R‐type (μM)	T‐type (μM)	hERG (μM)
Z160/NP118809 (Zamponi et al., 2009)	12b (Cav1.2)	–	0.28a	–	–	7.4b
aMONIRO‐1	27% @ 100 (Cav1.3) 46% @ 100 (Cav1.2)	>100	34	>100	3.3 (Cav3.1) 1.7 (Cav3.2) 7.2 (Cav3.3)	>100
A‐1048400 (Scott et al., 2012)	28	1.3	0.8	–	0.9 (Cav3.2)	–
Z944 (Tringham et al., 2012)	32 (Cav1.2)	–	11	–	0.05 (Cav3.1) 0.16 (Cav3.2)	7.8
NNC 55–0396 (Chen et al., 2014)	–	–	–	–	6.8b (Cav3.1) 96% @10a (Cav3.2)	–
Ziconotide (Wang et al., 2016)	–	>10	0.18b	–	–	–
TTA‐A2 (Kraus et al., 2010)	>30b (Cav1.2)	>30b	>30bRecovery from block	>30b	0.09 (Cav3.1) 0.09 (Cav3.2)	–
TROX‐1 (Swensen et al., 2012)	–	0.29	0.19	0.28	–	–

^aThis work.

^bStudies were performed with HEK293 cells stably expressing the desired voltage‐gated calcium channel. –, not tested.

The current inhibition properties varied between the different MONIRO‐1‐sensitive Ca_v subtypes, so we analysed the interaction between MONIRO‐1 and T‐type (Ca_v3.1) and N‐type (Ca_v2.2) calcium channels.

Differential modulation of N‐ and T‐type calcium channels: dependence on frequency and holding potential

We examined the inhibitory characteristics of MONIRO‐1 at different test pulse durations and frequencies on Ca_v2.2 (Figure 4A, B) and Ca_v3.1 channels (Figure 5A, B), as representatives of the VGCC types targeted by MONIRO‐1. Ca_v2.2 currents in response to shorter (45 ms) depolarizations were 1.5‐times more sensitive to MONIRO‐1 inhibition than those elicited by a longer (135 ms) pulse. Channels held open by the longer pulse also recovered more completely from block (82 ± 9%, n = 5) than those briefly opened (45 ms; 47 ± 3%, n = 5) (Figure 4A). In agreement with these findings, higher frequency stimulation of Ca_v2.2 channels rendered a population of channels that was more potently inhibited by MONIRO‐1 (0.03, 0.1 and 0.2 Hz; Figure 4B). At a frequency of 0.2 Hz, MONIRO‐1 inhibited Ca_v2.2 current by 29 ± 1% (n = 5), whereas at 0.1 and 0.03 Hz, only 21 ± 1% (n = 5) and 16 ± 1% (n = 5) of the current amplitude was inhibited respectively. Interestingly, the effects of MONIRO‐1 on Ca_v2.2 channels are in marked contrast to its effects on Ca_v3.1 channels (Figure 5A, B). In the case of Ca_v3.1 channels, pulse duration did not affect Ca_v3.1‐mediated current inhibition (% block 15 ms: 43 ± 3%, n = 5; and 50 ms: 51 ± 7%, n = 5), recovery from inhibition (% recovery 15 ms: 95 ± 1.3%, n = 5; and 50 ms: 99 ± 0.8%, n = 5) or frequency dependence of block over the same range as those tested for Ca_v2.2 channels (% block at 0.2 Hz = 43 ± 1.5%, n = 5; 0.1 Hz = 42 ± 2.3%, n = 5; and 0.03 Hz = 44 ± 1.6%, n = 5). However, at faster rates of 1 and 2 Hz, there was a significant increase in MONIRO‐1 block of Ca_v3.1 channels (% block at 1 Hz = 61.0 ± 1.5%, n = 6; and 2 Hz = 71.0 ± 2.0%, n = 5). This would increase threefold the affinity for channels firing at faster frequencies of 2 Hz compared with those firing at rates <0.2 Hz.

Figure 4.

MONIRO‐1 block of human Ca_v2.2 channels. (A) (i) Fraction blocked and (ii) recovery from block by 30 μM MONIRO‐1, with a depolarizing pulse duration of 45 ms or 135 ms, from a holding potential of −80 mV to a test potential that elicits maximal inward current, repeated every 10 s. The number of cells is indicated in parentheses. (B) Fraction of current inhibited by 10 μM MONIRO‐1 at three different frequencies (0.03, 0.1 and 0.2 Hz; holding potential −80 mV). The number of cells is indicated in parentheses. (C) Effect of holding potential (−100, −80 and −60 mV) on the fraction of current blocked by 30 μM MONIRO‐1 (pulsed at 0.1 Hz). The number of cells is indicated in parentheses. (D) Activation and SSI curves in the absence (control) and presence of 30 μM MONIRO‐1. Activation curves were generated through indicated depolarizing test pulses (135 ms) from a holding potential of −80 mV, repeated every 10 s. For SSI curves, selected pre‐pulse potentials (−120 to 0 mV in 5 mV increments for 10 s) were applied prior to a test pulse (135 ms) determined from the peak current, repeated every 15 s from a holding potential of −80 mV. (E) Recovery from inactivation in the absence (control) and presence of 30 μM MONIRO‐1. A two‐pulse protocol (Inset) was repeated every 15 s, where the time was varied between the first fully inactivating pulse and the test pulse (0–10 s). *P < 0.05, significantly different as indicated; n.s. = not significant; Student's t‐test.

Figure 5.

MONIRO‐1 block of human Ca_v3.1 channels. (A) (i) Bar graph showing fraction blocked and (ii) recovery from block by 3 μM MONIRO‐1, with a depolarizing pulse duration of 15 ms or 50 ms, from a holding potential of −80 mV repeated every 10 s. The number of cells is indicated in parentheses. (B) Bar graph of the fractional current inhibited by 3 μM MONIRO‐1 at five different frequencies (0.03, 0.1, 0.2, 1 and 2 Hz; holding potential −80 mV). The number of cells is indicated in parentheses. (C) Effect of holding potential (−100, −80 and −60 mV) on the fraction of current inhibited by 3 μM MONIRO‐1 (pulsed at 0.1 Hz). The number of cells is indicated in parentheses. (D) Activation and SSI curves in the absence (control) and presence of 3 μM MONIRO‐1. Activation curves were generated through indicated depolarizing test pulses (50 ms) from a holding potential of −80 mV, repeated every 10 s. For SSI curves, selected pre‐pulse potential values (−85 to −20 mV in 5 mV increments for 1 s) were applied prior to a test pulse (50 ms) determined from the peak current, repeated every 10 s from a holding potential of −80 mV. (E) Recovery from inactivation in the absence (control) and presence of 3 μM MONIRO‐1. A two‐pulse protocol (Inset) was repeated every 15 s, where the time was varied between the first fully inactivating pulse and the test pulse. *P < 0.05, significantly different as indicated; n.s. = not significant; Student's t‐test.

To further explore the differential effects of MONIRO‐1 on Ca_v2.2 and Ca_v3.1 channels, we investigated the effect of membrane holding potential on block of Ca_v2.2 and Ca_v3.1 channels. MONIRO‐1 could more effectively block Ca_v2.2 channels at more depolarized potentials (block at −100 mV = 34 ± 1.8%; −80 mV = 43 ± 3.2%; and −60 mV = 77 ± 5.5%, n = 5). In contrast, block at all of the holding potentials tested did not affect the degree by which MONIRO‐1 inhibited Ca_v3.1 channels (block at −100 mV = 41 ± 0.5%; −80 mV = 46 ± 5.1%; and −60 mV = 43 ± 0.6%, n = 5) (Figures 4C and 5C). This result highlights the state‐dependent differences between MONIRO‐1 inhibition of Ca_v2.2 and Ca_v3.1 channels.

Effects of MONIRO‐1 on N‐ and T‐type calcium channel activation, inactivation and recovery from inactivation

To identify any effects that MONIRO‐1 may have on calcium current kinetics and to understand the mechanisms behind the enhanced inhibition at depolarised membrane potentials, we examined its effects at half‐maximal concentrations on Ca_v2.2 and Ca_v3.1 activation and SSI (Figures 4D and 5D). Exposure to MONIRO‐1 did not affect the voltage‐dependence of activation of Ca_v2.2 or Ca_v3.1 channels (Table 2). However, there was a significant ~10 mV hyperpolarizing shift in Ca_v2.2 SSI. An apparent decrease in the voltage‐dependence of SSI was also seen. However, no significant changes were observed in Ca_v3.1 SSI in the presence of MONIRO‐1 or voltage‐dependence of SSI. Channel activation and inactivation time constants were examined, with and without MONIRO‐1, for Ca_v2.2 (30 μM) and Ca_v3.1 channels (3 μM). MONIRO‐1 had no effect on either the activation or inactivation time constants of Ca_v2.2 channels (Table 2). However, MONIRO‐1 application increased both activation and inactivation time constants for Ca_v3.1 channels (Table 2).

Table 2.

Comparison of activation and inactivation parameters between Ca_v2.2 and Ca_v3.1 channels before and after application of MONIRO‐1

	Cav2.2		Cav3.1
	Control	MONIRO‐1 (30 μM)a	Control	MONIRO‐1 (3 μM)a
τ activation	1.81 ± 0.15 ms (5)	1.51 ± 0.17 ms (5)	1.90 ± 0.08 ms (5)	3.45 ± 0.51 ms (5)*
τ inactivation	61.75 ± 13.69 ms (5)	44.45 ± 7.61 ms (5)	10.78 ± 0.52 ms (5)	17.91 ± 0.74 ms (5)*
V0.5 activation	12.5 ± 0.5 mV (5)	12.4 ± 0.4 mV (5)	−34.0 ± 0.4 mV (7)	−34.4 ± 0.6 mV (7)
V0.5 inactivation	−41.3 ± 0.7 mV (5)	−52.9 ± 1.3 mV (5)*	−53.5 ± 0.6 mV (6)	−55.1 ± 0.6 mV (6)
Slope factor	5.3 ± 0.6 (5)	10.2 ± 1.2 (5)*	4.4 ± 0.6 (6)	6.1 ± 0.5 (6)

^* P < 0.05 compared with control, Students t‐test. Data presented as mean ± SEM and n is given in brackets.

^a Concentration chosen based on IC₅₀ of individual channel.

Finally, we investigated the effect of MONIRO‐1 on Ca_v2.2 and Ca_v3.1 channel recovery from inactivation. MONIRO‐1 had no observed effect on Ca_v2.2 recovery from inactivation time constants or their relative contribution to either time constants (control τ_fast = 0.22 ± 0.1 s, A_fast = 57 ± 8.5%; τ_slow = 2.8 ± 0.7 s, A_slow = 45 ± 8.6%; MONIRO‐1 τ_fast = 0.14± 0.1 s, A_fast = 43 ± 7.0%; τ_slow = 2.4 ± 0.4 s, A_slow = 58 ± 6.2%, n = 5 respectively). However, MONIRO‐1 slowed recovery of Ca_v3.1 channels from inactivation by decreasing the relative contribution of the fast recovery time constant while increasing the contribution of the slow time constant (control τ_fast = 0.051 ± 0.008 s, A_fast = 80.1 ± 5.1%; τ_slow = 0.20 ± 0.11 s, A_slow = 21.8 ± 5.4%; MONIRO‐1 τ_fast = 0.026 ± 0.007 s, A_fast = 29.6 ± 4.2%; τ_slow = 0.33 ± 0.03 s, A_slow = 70 ± 4.1%, n = 6; P < 0.05). Therefore, MONIRO‐1 can inhibit Ca_v3.1 channels more efficiently at firing frequencies >1 Hz (Figure 5B), as indicated by the slowed recovery from inactivation and frequency dependence, such as those prevailing in hyperexcitable neurons (Cain and Snutch, 2010).

Discussion

In this study, we examined the effects of an ortho‐phenoxyanilide on the inhibition and biophysical properties of VGCCs. This ortho‐phenoxyanilide, MONIRO‐1, is a low MW compound and was selected for its ability to inhibit VGCCs in a high‐throughput calcium influx assay (Gleeson et al., 2015). To ascertain its selectivity and mode of action, we examined its potency of block and biophysical properties on a variety of VGCCs. MONIRO‐1 was shown to be an N‐ and T‐type calcium channel inhibitor with advantageous biophysical properties, including state‐ and use‐dependence. It specifically targets Ca_v3.2 and Ca_v2.2 channels, which have been implicated in various pain pathways, and Ca_v3.1 channels, which are involved in cell signalling and epilepsy (Choi, 2013; Waxman and Zamponi, 2014).

Given that MONIRO‐1 inhibits depolarization‐activated calcium responses from SH‐SY5Y cells (Gleeson et al., 2015), we determined its efficacy and selectivity against several VGCCs expressed in HEK293 cells. After screening representative members of each VGCC family (L‐, N‐, P/Q‐, R‐ and T‐types), we established that MONIRO‐1 preferentially inhibits T‐type channels (Ca_v3.1, Ca_v3.2 and Ca_v3.3 channels), modestly interacts with N‐type Ca_v2.2 channels (with 5–20 times less affinity compared to T‐type calcium channels) and has poor efficacy for L‐type (Ca_v1.2 and Ca_v1.3 channels IC₅₀ > 100 μM). No other VGCCs investigated (P/Q‐type Ca_v2.1; R‐type Ca_v2.3) responded to high concentrations of MONIRO‐1. hERG, a typical off‐target channel for low MW compounds, was not modulated by MONIRO‐1 at concentrations up to 100 μM. Subsequently, we characterized the mechanism of MONIRO‐1 inhibition on its primary targets, N‐ and T‐type calcium channels.

In the present study, MONIRO‐1 inhibited Ca_v3.1 channels with an IC₅₀ of 3 μM, so we examined the mechanistic details of this inhibition. Many T‐type calcium channel inhibitors have been identified in assays looking for inactivation state‐dependent inhibitors. These studies have led to compounds that predominantly bind and alter SSI, including TTA‐A2, A‐1048400 and NNC 55‐0396, unlike MONIRO‐1, which does not affect SSI of Ca_v3.1 channels. MONIRO‐1 (3 μM) significantly slowed both channel activation and inactivation kinetics, without discernible effects on either V_0.5act or V_0.5inact. MONIRO‐1 slowed recovery of Ca_v3.1 from inactivation at rates >1 Hz, which was reflected as an increase in current inhibition at 1 and 2 Hz (Figure 5B, E), suggesting that MONIRO‐1 would preferentially target epileptic firing neurons over normal firing neurons. Taken together, the lack of effect of MONIRO‐1 on other biophysical properties examined, including the lack of effect of holding potential, suggests that it acts by binding to prevent Ca_v3.1 channels from transitioning from a non‐conducting state to the open state. Therefore, the prolonged recovery from inactivation caused by MONIRO‐1 inhibition of Ca_v3.1‐mediated currents is likely to be due to the slower dissociation of the compound from the open‐blocked channels or due to its accelerating channel transitions into a slow inactivated channel conformation. A similar phenomenon has been observed for mibefradil inhibition of Ca_v2.1 channels (Timin et al., 2004).

Ca_v3.1 channels are key players in the circuitry involved in sleep including the reticular thalamic nucleus, thalamic relay neurons and neocortical pyramidal cells (Chen et al., 2014). Ca_v3.1 channels are highly expressed in thalamocortical neurons (Kim et al., 2001; Song et al., 2004), where they initiate entry of these neurons into bursting mode – a typical activity mode in absence seizures (Cain and Snutch, 2013). Consequently, Ca_v3.1 channels have been directly linked to absence of epilepsies (Kim et al., 2001). Disease mutations in Ca_v3.2 channels have also been linked to both rat models and human idiopathic generalized epilepsy (Khosravani et al., 2005; Powell et al., 2009), suggesting that inhibitors of both Ca_v3.1 and Ca_v3.2 channels could decrease neuronal firing induced by hyperactivity of these channels during epilepsy. Compound Z944, which has higher affinity for the inactivated state of T‐type calcium channels over the resting state, reduced seizure activity in terms of time spent in seizure and the total number of seizures per hour (Tringham et al., 2012). During burst firing, VGCCs repetitively open and close at frequencies of 5–10 Hz (Cain and Snutch, 2010). Ideally, an anti‐epileptogenic drug targeting Ca_v3.1 or Ca_v3.2 channels in epileptogenic foci, that is, on neurons in bursting mode where the channels are in the open state. Our results suggest that MONIRO‐1 modulates Ca_v3.1 channels by holding the channel in a non‐conducting state, thus prolonging recovery from inactivation, which makes the compound well suited for the study and/or potentially for the amelioration of Ca_v3.1 or Ca_v3.2‐mediated epileptogenesis in neurons entering bursting mode.

Considerable research into pain physiology has focused on voltage‐gated N‐ and T‐type calcium channels. Ca_v3.2 channels are expressed in peripheral sensory neurons, where they modulate pain processing by controlling dorsal root ganglion neuroexcitability (Carbone and Lux, 1984). Ca_v2.2 channels are located at presynaptic terminals in the CNS, where they modulate neurotransmitter release and fast synaptic transmission (Wheeler et al., 1994). MONIRO‐1 inhibits Ca_v2.2 channels under resting state conditions with an IC₅₀ of ~30 μM, which is approximately 10‐fold higher than that observed the T‐type channels. However, the effect of MONIRO‐1 on Ca_v2.2 channels is enhanced at higher frequencies, where faster pacing of the channel increased inhibition. At frequencies as high as 0.2 Hz (the maximum tested), the apparent affinity increased to ~20 μM, with no observable change to the time course of activation or inactivation of the channel (Figure 4B and Table 2).

Following this observation, we examined the potential state dependence of the activity of MONIRO‐1. Short (45 ms) depolarizing test pulses produced greater inhibition of Ca_v2.2 channels as well as less recovery after washout compared with a longer (135 ms) test pulse, suggesting a possible decrease in affinity for the open and/or inactivated state (Figure 4A). MONIRO‐1 had a pronounced effect on Ca_v2.2 channel SSI that was not paralleled in Ca_v3.1 channels (Figure 4D), along with strong dependence on holding potential. At −60 mV, ~80% of the whole‐cell current was inhibited compared with −100 mV where only ~35% of the current was blocked (Figure 4C). Such a hyperpolarizing shift in SSI and the dependence on holding potential suggests that MONIRO‐1 preferentially binds to the inactivated state of N‐type calcium channels (Bean et al., 1983; Bean, 1984; Sanguinetti and Kass, 1984). Interestingly, unlike its effect on T‐type calcium channels, MONIRO‐1 did not influence N‐type calcium channel recovery from inactivation (Figure 5E), suggesting a distinct mode of action against the two channel families.

Nociceptive stimuli cause depolarization, calcium channel opening, intracellular [Ca²⁺] increase and neuronal hyper‐excitability, as shown by faster firing rates (Winquist et al., 2005). Nociceptive stimuli reduce the effective open time and increase the frequency of channel opening (Winquist et al., 2005). Sustained neuronal firing in pain syndromes increases cytosolic calcium levels in the dorsal horn and shifts calcium channels into the inactivated state. Therefore, drugs that target Ca_v2.2 channels are best suited to ‘prefer’ those channels in the inactivated state (predominant in pain syndromes), allowing normal neurotransmitter release signalled by these channels in neurons not in a pain state.

A few compounds have recently reported to exhibit use‐ and/or state‐dependent inhibition of calcium channels involved in pain signalling, including A‐1048400 (Scott et al., 2012) (compound 3, Figure 1), Z160/NP118809 1 (Zamponi et al., 2009) and http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=7766 (Abbadie et al., 2010). To explore MONIRO‐1 inhibitory mechanism of Ca_v2.2, we examined stimulation conditions comparable with those occurring in the physiological setting. Given that Ca_v2.2 channels can potentially accumulate in the inactivated state in hyperactive pain state conditions (Winquist et al., 2005), the selective effect of MONIRO‐1 on channels in this state could be a valuable pharmacological property that would enable therapeutic targeting during neuropathic pain treatment. Frequency dependence and preference of MONIRO‐1 for the inactivated Ca_v2.2 channels suggest that it or related compounds may prove useful in the prevention of high‐frequency firing episodes, an important signalling mode occurring during pain transmission. Therefore, MONIRO‐1 may preferentially inhibit channels in hyperexcitable states in pain‐specific loci, leaving normally firing neurons unaffected (Liao et al., 2011).

In conclusion, we have identified MONIRO‐1 as a low MW antagonist of voltage‐gated calcium channels with 10‐ to 100‐fold higher affinity for Ca_v3.1, Ca_v3.2, Ca_v3.3 and Ca_v2.2 channels than other classes of VGCCs examined. MONIRO‐1 is a valuable tool for the study of VGCCs because it inhibits N‐ and T‐type calcium channels through distinct mechanisms. Given its state‐ and use‐dependent inhibition of Ca_v2.2 channels (preferentially inhibits the inactivated state) and Ca_v3.1 channels (prohibiting channel transitions to the open state), while exhibiting frequency‐dependent block of both channels, MONIRO‐1 is a novel lead compound that could be used to develop Ca_v channel blockers to treat VGCC pathologies, including neuropathic pain and epilepsy. MONIRO‐1 shows a novel mechanism of block of hCav3.1 channels compared with previously described T‐type calcium channel blockers, in that it slows channel activation and inactivation kinetics and traps channels at high‐frequency firing (>1 Hz) from transitioning to the open state while not altering SSI. Prior to in vivo assessment of their analgesic potential, MONIRO‐1 and analogues need to be characterized at other voltage‐gated ion channels in pain pathways, including Na_v1.7 and Na_v1.8 channels.

Author contributions

J.R.M., P.J.D., K.L.T. and D.J.A. designed the experiments; J.R.M., L.M., E.C.G. and S.S. performed the experiments; J.R.M. analysed the electrophysiological data; J.R.M., P.J.D., K.L.T. and D.J.A. prepared the figures; J.R.M. wrote the manuscript; J.R.M., R.J.L., P.J.D., K.L.T. and D.J.A. revised the manuscript. All authors reviewed the manuscript.

Conflict of interest

The authors declare no conflicts of interest.

Declaration of transparency and scientific rigour

This http://onlinelibrary.wiley.com/doi/10.1111/bph.13405/abstract acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research recommended by funding agencies, publishers and other organisations engaged with supporting research.

McArthur, J. R. , Motin, L. , Gleeson, E. C. , Spiller, S. , Lewis, R. J. , Duggan, P. J. , Tuck, K. L. , and Adams, D. J. (2018) Inhibition of human N‐ and T‐type calcium channels by an ortho‐phenoxyanilide derivative, MONIRO‐1. British Journal of Pharmacology, 175: 2284–2295. doi: 10.1111/bph.13910.