Characterization of Next-Generation Inhibitors for the Inward-Rectifier Potassium Channel K_ir2.1: Discovery of VU6080824

Abstract

ML133 is a selective inhibitor of the inward-rectifier potassium channel K_ir2.1 and has found extensive use as a tool with which to probe K_ir biology. Despite its utility as a tool compound, ML133 has only modest on-target potency (manual patch clamp (MPC) K_ir2.1 IC₅₀ = 1.5 μM, pH 7.4), and its in vivo pharmacokinetics (PK) were previously uncharacterized. In the present study, we report a next-generation series of K_ir2.1 inhibitors based on the ML133 scaffold, along with the rat PK of ML133 and selected analogs. Compound 5s (VU6080824) was ultimately identified as having superior potency to ML133 in both the thallium flux and MPC functional assays and has excellent PK properties suitable for use as an improved K_ir2.1 tool compound in rodents.

The inward-rectifier potassium channels (K_ir) play critical roles in diverse biological functions and are characterized by broad tissue distribution. At least 7 subfamilies of K_ir channels have been identified (K_ir1-K_ir7), and each class is further subdivided into specific isoforms. Most K_ir channels exist as either homomeric or heteromeric tetramers arranged around a central pore, through which potassium ions (K⁺) move along an electrochemical gradient. “Inward-rectifier” refers to the tendency of the K_ir channels to pass this ionic current predominantly in the inward direction under voltage-clamp conditions.

It is well-established that the K_ir channels are valuable therapeutic targets for a wide range of indications including cardiovascular, metabolic, renal, and neurological. − Specifically, K_ir2.1, the second member of the K_ir class to be cloned, is broadly expressed in bodily tissues including cardiac cells, macrophages, neutrophils, astrocytes, and glial cells. Moreover, K_ir2.1 mutations have been linked to disorders including periodic paralysis, Anderson-Tawil syndrome (ATS), and short-QT syndrome. − Selective modulation of K_ir2.1, therefore, would be of immense value both for therapeutic development and as a tool to further understand K_ir2.1 biology.

The high sequence homology of the K_ir families within the pore domain, including the K_ir2 family, has challenged the development of selective tool molecules. , In general, the known pharmacological toolbox for the K_ir channels contains only weak and/or nonselective isoform inhibitors (although selective inhibitors for certain isoforms, e.g., K_ir1.1, have been described). Such tool molecules are often further limited by poor physicochemical properties and lack of “drug-likeness”. Although ion flux-based high-throughput screening (HTS) approaches have found some success in the identification of new chemical matter for various members of the K_ir family, many of the available chemotypes require extensive optimization. In the case of K_ir2.1 specifically, the available tools are limited to weakly active molecules repurposed from other indications (e.g., tamoxifen, chloroquine, and thiopental), − nonspecific divalent cation inhibition (barium, Ba²⁺), or the HTS-derived semiselective inhibitor ML133.

ML133 (N-(4-methoxybenzyl)-1-(naphthalen-1-yl)­methanamine) is a pH-dependent K_ir2.1 probe identified from a thallium flux HTS campaign in association with the Molecular Libraries Probe Production Centers Network (MLPCN) (Figure ). Although ML133 represents an encouraging step toward the identification of a potent small molecule K_ir probe (human K_ir2.1 manual patch clamp (MPC) IC₅₀ = 1.8 μM, pH = 7.4 in the original report), the compound is relatively nonselective against the other channels in the K_ir2.x family. Encouragingly, however, ML133 has appreciable selectivity against other K_ir channels including K_ir1.1, K_ir4.1 and K_ir7.1. The on-target potency and selectivity of ML133 has facilitated its use as a probe across many disease models including sour taste transduction, regulation of neuronal processes, modulation of autophagy in endothelial progenitor cells, and traumatic brain injury.

1.

Chemical structure, manual patch clamp (MPC), and selectivity data for ML133.

Given the widespread success of ML133 as a K_ir2.1-specific molecular probe, we became interested in the identification of next-generation K_ir2.1 inhibitors based on the ML133 scaffold. The previously reported structure–activity relationship (SAR) data for ML133 were quite steep, and minor modifications to the substituted dibenzylamine scaffold were generally observed to ablate all K_ir2.1 activity (for example, all examined replacements to the 4-methoxybenzyl motif resulted in inactive compounds). Additionally, the rodent pharmacokinetics (PK) for ML133 were not known, further limiting its utility as an in vivo tool (i.e., understanding appropriate doses for achieving sufficient plasma exposure, correlation of PK with any observed behavioral phenotypes in animal studies). Toward the identification of next-generation, ML133-based K_ir2.1 inhibitors, our goals for the present study were therefore 2-fold: (1) an expanded and comprehensive SAR campaign, examining all facets of the ML133 scaffold, to identify molecules with improved K_ir2.1 inhibition, and (2) PK characterization for ML133 and related analogs in rats to further enable in vivo work, as the MLPCN did not allow for in-depth DMPK profiling at the time.

We initially sought to further explore replacements to the 4-methoxybenzyl group, and a series of molecules were synthesized in which the 1-naphthyl pendant was held constant in the context of 4-methoxybenzyl replacements. Analogs in this series were prepared via a simple reductive amination reaction between 1-naphthaldehyde (2) and the corresponding amine (Scheme ).

1. Synthesis of Compounds 3a-l .

^a Reagents and conditions: (a) 4-substituted benzylamine, NaBH­(OAc)₃, DCM, r.t., 4–80% (3a–i). (b) 2-Methoxyethanol, Cs₂CO₃, RockPhos-Pd-G3, toluene, 100 °C, 30% (3j). (c) Isothiazolidine 1,1-dioxide, K₃PO₄, [Pd­(allyl)­(t-BuBrettPhos)]­OTf, t-BuOH, 110 °C, 11% (3k). (d) Dimethylphosphine oxide, K₃PO₄, Pd₂(dba)₃, XantPhos, 1,4-dioxane, 110 °C, 36% (3l).

In the original SAR report, all modifications to the 4-methoxybenzyl moiety were found to ablate K_ir2.1 activity, and selected examples were reported (e.g., 4-Cl and 4-OCF₃). With an understanding of this steep SAR, we therefore decided to begin by profiling specific hydrogen bond acceptors as methoxy replacements, hoping to more closely mimic the polarity and hydrogen-bonding characteristics of the 4-OMe group. Initially, a diverse group of hydrogen bond acceptor motifs were surveyed at the 4-phenyl position (exemplified by 3a–e, Table ). Unfortunately, as was observed in the original SAR campaign, all attempted modifications/replacements to the methoxy substituent were unfruitful. In some cases, weaker activity (> 10 μM, with diminished % efficacy) was observed (as in the case of difluoromethoxy analog 3a, nitrile 3b, and pyrazole 3e). Cyclization of the left-hand pendant (3g and 3h, prepared via analogous reductive amination chemistry) was also detrimental to K_ir2.1 potency, as was the introduction of a hydrogen bond donor (primary alcohol 3f). To access a greater diversity of 4-substitutions, additional O-, N-, and P-linked analogs were prepared via transition metal cross-coupling chemistry starting from 4-bromo intermediate 3i (exemplified by 3j–l). − For cross-couplings involving challenging nucleophiles (e.g., sultam 3k), a π-allylpalladium precatalyst derived from tBuBrettPhos ([Pd­(allyl)­(tBuBrettPhos)]­OTf, prepared as previously described) proved particularly useful (Scheme ). Unfortunately, all examined replacements to the 4-methoxy group continued to dramatically diminish K_ir2.1 activity. These results further corroborate the necessity of the 4-methoxy substitution for maintaining K_ir2.1 inhibition within this chemotype.

1. K_ir2.1 Thallium Flux Potency and Efficacy Data for ML133, Compounds 3a-h, 3j-l .

^aLive cells expressing human K_ir2.1 were stimulated with thallium solution and measured in real time for changes in fluorescent signal (482ex, 536em). The slopes of the change in fluorescence were fit to a four-parameter logistic equation to estimate the reported IC₅₀ (μM) and efficacy (% inhibition). Test compounds that did not fit the 4PL model within the concentration range tested (> 10 μM) have the 30 μM measured value shown expressed relative to 30 μM ML133 (% inhibition). Values represent a single experiment tested in triplicate. ML133 estimated IC₅₀ 10.1 μM (pIC₅₀ 4.98 ± 0.05 SEM) is the mean reported over 9 experimental plates in triplicate representing dates of the test compounds included in the table.

Next, we turned our attention to replacements to the 1-naphthyl pendant (Scheme ). Starting from (4-methoxyphenyl)­methanamine (4), analogs 5a–i were prepared via reductive amination chemistry, and analogs 5j and 5k were accessed via substitution chemistry from 4-methoxybenzyl chloride (6). Analogously to Scheme , a greater diversity of substitutions could be accessed utilizing cross-coupling chemistry. Specifically, Suzuki-Miyaura chemistry starting from aryl bromide intermediate 5i afforded analogs 5m–u; to avoid a historically challenging coupling with a 2-pyridyl boron reagent, aryl bromide 5i was converted to the pinacol boronate prior to coupling with 2-bromopyridine to give 5l. Buchwald-Hartwig cross-couplings smoothy furnished analogs 5v-z. The SAR results for 5a–h, 5j–z are outlined in Table . In short, this exercise was found to be much more fruitful; after synthesizing > 50 novel ML133 analogs in this context, several were found to have comparable or improved thallium flux potency relative to ML133 (Table ).

2. Synthesis of Compounds 5a-z .

^a Reagents and conditions: (a) substituted aldehyde, NaBH­(OAc)₃, DCM, r.t., 28–70% (5a–i). (b) Substituted amine, K₂CO₃, MeCN, r.t., 7–25% (5j–k). (c) Substituted boronic acid or ester, Cs₂CO₃, Pd­(dppf)­Cl₂, 1,4-dioxane, H₂O, 100 °C, 41–66% (5m–u). (d) Substituted amine, Cs₂CO₃, Pd₂(dba)₃, RuPhos, 1,4-dioxane, 100 °C, 66–83% (5v–y). (e) 1,2-Thiazinane 1,1-dioxide, K₃PO₄, [Pd­(allyl)­(t-BuBrettPhos)]­OTf, 1,4-dioxane, 100 °C, 42% (5z). (f) (i) Bis­(pinacolato)­diboron, potassium acetate, Pd­(dppf)­Cl₂, 1,4-dioxane, 100 °C; (ii) 2-bromopyridine, Cs₂CO₃, Pd­(dppf)­Cl₂, 1,4-dioxane, H₂O, 100 °C, 16% (5l).

2. K_ir2.1 Thallium Flux Potency and Efficacy Data for ML133, Compounds 5a-h, 5j-z, 7 .

^aLive cells expressing human K_ir2.1 were stimulated with thallium solution and measured in real time for changes in fluorescent signal (482ex, 536em). The slopes of the change in fluorescence were fit to a four-parameter logistic equation to estimate the reported IC₅₀ (μM) and efficacy (% inhibition). Test compounds that did not fit the 4PL model within the concentration range tested (> 10 μM) have the 30 μM measured value shown expressed relative to 30 μM ML133 (% inhibition). Values represent a single experiment tested in triplicate. ML133 estimated IC₅₀ 10.1 μM (pIC₅₀ 4.98 ± 0.05 SEM) is the mean reported over 9 experimental plates in triplicate representing dates of the test compounds included in the table.

An aza-scan of the 1-naphthyl group largely resulted in compounds with attenuated activity relative to ML133, although isoquinoline 5b, while approximately 2-fold less potent than ML133 in the thallium flux assay, was found to be highly efficacious (140% inhibition). Alternative 6/5 bicyclic replacements (e.g., imidazopyridine 5f and benzo­[d]­[1,3]­dioxole 5g) were inactive, although, interestingly, regioisomeric benzo­[d]­[1,3]­dioxole 5h was found to be roughly equipotent to ML133. 1-([1,1’-Biphenyl]-3-yl) and 1-([1,1’-biphenyl]-4-yl) replacements (5j and 5k) were also tolerated, particularly 1-([1,1’-biphenyl]-4-yl) analog 5k (thallium flux IC₅₀ = 8.2 μM). Accordingly, we next held the 1-([1,1’-biphenyl]-4-yl) motif constant, and surveyed systematic substitutions at the ortho, meta, and para positions (aza: 5l–5n; and fluoro: 5o–5q). In both series, substitution at the meta position afforded analogs with the greatest potencies (3-pyridine 5m and 3-fluorophenyl 5p). Additional meta substitutions (3-methyl 5r, 3-methoxy 5s and 3-cyano 5t) were also well tolerated and indeed afforded the most potent K_ir2.1 inhibitors within this chemotype to date (thallium flux IC₅₀ s < 6 μM for 5s and 5t). Substitution with alternative heteroaromatics at the 4-phenyl position were not tolerated (e.g., pyrazole 5u). Within the N-linked series (5v–5z), prepared via Buchwald-Hartwig chemistry, all examined compounds were either inactive or had attenuated activity compared to ML133, apart from pyrrolidine 5v (see Table ).

O–Demethylation of aryl methoxy groups is a common metabolic degradation pathway, , and, in parallel, we were interested in the design of an ML133 analog that might circumvent this potential biotransformation. Accordingly, we also synthesized the direct OCD₃ analog of ML133 (7, Scheme ). In the thallium flux assay, 7 was found to be approximately equipotent to ML133 (see Table ).

3. Synthesis of Compound 7 .

^a Reagents and conditions: (a) methanol-d4, Cs₂CO₃, RockPhos-Pd-G3, toluene, 100 °C, 42%.

Selected novel analogs were further profiled relative to ML133 in the manual patch clamp (MPC) assay. In the original ML133 report, the compound was found to have 1.8 μM potency at pH 7.4 (ML133 is a pH-dependent probe due to the presence of the basic amine). Upon retesting, ML133 was found to have a potency value of 1.51 μM, in good agreement with the initial publication. Selected next-generation analogs with comparable or improved potency in the thallium flux assay were chosen for MPC profiling (Table ). Interestingly, OCD₃ analog 7 was approximately 3–4-fold less potent than ML133 in this context (as was benzo­[d]­[1,3]­dioxole 5h). 1-([1,1’-Biphenyl]-4-yl) analog 5k was equipotent relative to ML133. Encouragingly, meta-substituted 1-([1,1’-biphenyl]-4-yl) analogs 5j, 5p, 5s and 5t were all ∼ 3–5-fold more potent than ML133 in MPC; these analogs are therefore the most potent K_ir2.1 analogs yet described for the dibenzylamine chemotype.

3. Manual Patch Clamp Data for Selected K_ir2.1 Inhibitors .

Compound	IC50, μM	Emax,%
ML133	1.51	99
5h	3.03	92
5j	0.39	100
5k	1.39	99
5p	0.30	98
5s	0.35	100
5t	0.48	98
7	3.94	82

^aElectrophysiological characterization of K_ir2.1 current inhibition. Responses were measured at −120 mV and normalized to the 2 mM Ba²⁺ full block. IC₅₀ values were derived from a four-parameter logistic fit. E _max indicates the % inhibition of Ba²⁺-sensitive current at the highest tested concentration (3–30 μM). MPC experiments were run at pH 7.4 with a top compound concentration of 10 μM, unless noted.

^b30 μM top concentration.

^c3 μM top concentration.

Differences in compound potency between the thallium flux and MPC assays are routinely observed, with MPC considered the “gold standard” measurement (in MPC, voltage clamp electrophysiology directly measures the potassium current at specific voltages and channel open states, rather than a nonphysiological cation (e.g., thallium). − The potency difference observed for deuterated compound 7 relative to ML133 in this context is noteworthy and in line with the steep SAR observed for the attempted replacements to the 4-methoxy group. Further study, however, will be necessary to refine the true extent to which this hydrogen/deuterium exchange affects the MPC potency, and to understand any mechanism(s) behind this observation.

While ML133 has found use as a K_ir2.1 tool molecule across many disease models, the in vivo PK of this and related compounds within the dibenzylamine chemotype have not previously been characterized. As a highly efficient method for profiling the PK of a compound set, we therefore examined ML133 (alongside 9 additional next-generation K_ir2.1 inhibitors) in our rat PK PBL cassette dosing platform (Table ). − From this experimental setup, administration of a 4-compound bolus dose, plus a standard control, can be used to determine a variety of PK parameters in a high-throughput fashion. Specifically, Sprague–Dawley (SD) rats were given 0.2 mg/kg of each K_ir2.1 inhibitor (i.v. dosing, 4 compounds +1 control, total dose 1 mg/kg) with PK measured across set time points up to 24 h. Additionally, a separate cohort of animals was utilized to assess total plasma-brain levels (PBL, ratio reported as K_p) of compound at a 0.25 time point. In this fashion, standard plasma-based PK parameters (elimination t _1/2, mean resonance time (MRT), plasma clearance (CL_p), and volume of distribution (V_ss)) were measured alongside a parallel assessment of total brain exposure.

4. Rat PK PBL Cassette Data for Selected K_ir2.1 Inhibitors .

Compound	Elim. t 1/2 (h)	MRT (h)	CLp (mL/min/kg)	Vss (L/kg)	AUC (h*ng/mL)	Kp
ML133	0.75	0.78	31.9	1.50	104	3.36
7	1.39	1.24	22.8	1.70	146	3.30
5b	0.29	0.36	18.2	0.39	183	1.05
5h	1.33	0.87	30.4	1.58	110	2.13
5j	1.18	1.00	31.7	1.90	105	4.23
5k	1.68	1.08	22.6	1.46	148	1.82
5p	0.69	0.76	30.6	1.39	109	2.27
5r	1.73	1.31	20.6	1.62	162	2.29
5s	1.56	0.99	17.7	1.06	189	1.81
5t	1.66	0.99	20.2	1.20	165	1.25

^aCompounds are administered to male SD rats (n = 1) as a 0.2 mg/kg i.v. dose (8% EtOH, 33% PEG400, 58% DMSO (0.5 mL/kg)).

In general, compounds within this chemotype were found to have short to moderate t _1/2s (0.29 – 1.73 h) and MRTs (0.36 – 1.31 h), with CL_p values in the low to moderate range (17.7 – 31.9 mL/min/kg). All profiled compounds were also found to have a moderate V_ss (0.39 – 1.90 L/kg), and all were brain penetrant (K_p s ≥ 1.05). Interestingly, OCD₃ analog 7 (VU6079685) was found to have a ∼ 2-fold longer t _1/2 compared to ML133, and a ∼ 1.4-fold lower overall plasma clearance. These data indicate that O-demethylation may contribute to the metabolic clearance of this chemotype, an effect that can be readily attenuated through deuterium incorporation (kinetic isotope effect). ,

Although several compounds in this series proved attractive with respect to overall PK, analog 5s (VU6080824) was particularly encouraging. This compound displayed low plasma clearance (CL_p = 17.7 mL/min/kg), with a t _1/2 of 1.56 h. Additionally, VU6080824 was found to be robustly brain penetrant (K_p = 1.81). These PK data, alongside the improved MPC potency of this compound relative to ML133 (IC₅₀ = 0.35 μM, pH 7.4), highlight VU6080824 as a next-generation K_ir2.1 inhibitor with superior potency and PK properties relative to ML133.

In the thallium flux assay, VU6080824 was found to be selective relative to K_ir1.1 and K_ir4.1 up to 10 μM (K_ir1.1 and K_ir4.1 IC₅₀ s > 10 μM), although additional refinement will be necessary to understand the true selectivity ratios for each given the relatively modest potency of VU6080824 in the K_ir2.1 thallium flux assay (IC₅₀ = 5.9 μM). With respect to selectivity against other members of the K_ir2.x family, VU6080824 was found to be approximately 2-fold selective relative to K_ir2.2 and K_ir2.3, as measured by MPC (IC₅₀s = 0.75 μM and 0.86 μM, respectively). These data are in line with the overall selectivity profile observed for ML133, and indicate that further structural refinements will be necessary to improve selectivity within the Kir2.x family.

Given the favorable PK results observed for OCD₃ analog 7 relative to ML133, the three possible OCD₃ analogs of VU6080824 (5s) were also prepared in an analogous fashion (left-hand pendant OCD₃, right-hand pendant OCD₃, left–right dual OCD₃; see the Supporting Information for further details). In this context, however, none of the deuterated analogs were superior to 5s with respect to PK, and all showed comparable (if slightly inferior) clearance profiles. Additionally, all three OCD₃ analogs displayed comparable K_ir2.1 potencies in both the thallium flux (4.3 – 5.4 μM) and MPC assays (0.21 – 0.33 μM) to 5s, and similar selectivity profiles relative to K_ir2.2 and K_ir2.3 (generally 1–3 fold selectivity; see the Supporting Information for full potency and PK details). The effects of deuterium incorporation within this chemotype thus appear to be highly context dependent.

In conclusion, we have profiled a novel set of K_ir2.1 inhibitors, using the substituted dibenzylamine ML133 chemotype as a scaffold for chemical diversification. From this exercise, substitutions on the 1-naphthyl pendant identified next-generation analogs with improved potency relative to ML133 in both the thallium flux and MPC functional assays. Selected compounds, along with ML133, were further profiled in rat PK PBL cassettes to understand the in vivo properties of this series. VU6080824 (1-(3′-Methoxy-[1,1’-biphenyl]-4-yl)-N-(4-methoxybenzyl)­methanamine) ultimately emerged as a superior K_ir2.1 tool molecule, which we hope will find widespread use as a probe for K_ir biology.

Safety Statement. No unexpected or unusually high safety hazards were encountered.

Glossary

Abbreviations

ATS

Anderson-Tawil Syndrome

AUC

area under the curve

CL_p

plasma clearance (mL/min/kg)

DCM

dichloromethane

GPCR

G protein-coupled receptor

HTS

high-throughput screen

K_ir

inward-rectifier potassium channel

K_p

total brain to plasma ratio

MeCN

acetonitrile

ML133

(N-(4-methoxybenzyl)-1-(naphthalen-1-yl)­methanamine)

MLPCN

Molecular Libraries Probe Production Centers Network

MPC

manual patch clamp

MRT

mean resonance time (h)

PBL

plasma-brain levels

PK

pharmacokinetics

SAR

structure–activity relationship

SD

Sprague–Dawley rats

t _1/2

half-life (h)

t-BuOH

tert-butanol

V_ss

volume of distribution (steady state, L/kg)

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.5c00297.

• Experimental methods for compound synthesis and characterization, and experimental protocols for molecular pharmacology (thallium flux and manual patch clamp functional assays). Experimental methods for PK cassette studies (PDF)

∇.

R.A.D., D.H.H., and A.M.B. performed synthetic chemistry and compound characterization. R.M.L. performed manual patch clamp experiments and data analysis. L.L., Y.B., and E.L.D. performed thallium flux experiments and data analysis. S.K. and A.T.G. performed PK experiments and bioanalysis. O.B. analyzed PK experimental data. C.W.L., J.S.D., D.W.E., and A.M.B. oversaw experimental design and conceived the study, and A.M.B. wrote the manuscript with final approval from all authors. R.A.D. and D.H.H. contributed equally.

The authors declare no competing financial interest.