Identification and development of TRPM4 antagonists to counteract neuronal excitotoxicity

Abstract

Neurodegeneration in central nervous system disorders is linked to dysregulated neuronal calcium. Direct inhibition of glutamate-induced neuronal calcium influx, particularly via N-methyl-D-aspartate receptors (NMDAR), has led to adverse effects and clinical trial failures. A more feasible approach is to modulate NMDAR activity or calcium signaling indirectly. In this respect, the calcium-activated non-selective cation channel transient receptor potential melastatin 4 (TRPM4) has been identified as a promising target. However, high affinity and specific antagonists are lacking. Here, we conducted high-throughput screening of a compound library to identify high affinity TRPM4 antagonists. This yielded five lead compound series with nanomolar half-maximal inhibitory concentration values. Through medicinal chemistry optimization of two series, we established detailed structure-activity relationships and inhibition of excitotoxicity in neurons. Moreover, we identified their potential binding site supported by electrophysiological measurements. These potent TRPM4 antagonists are promising drugs for treating neurodegenerative disorders and TRPM4-related pathologies, potentially overcoming previous therapeutic challenges.

Keywords: Molecular biology; Molecular neuroscience; Neuroscience.

Graphical abstract

Figure 1

Identification of five lead TRPM4 antagonist series by high-throughput screening (A) Schematic illustration of the high-throughput screening (HTS) workflow for the identification of TRPM4 antagonists. The HTS involved screening a library comprising 256,286 small molecules. Active compounds were identified using a kinetic FLIPR assay based on membrane-potential sensitive dyes followed by subsequent validation of their activity and determination of IC50 values with an automated patch clamp system (QPatch). (B) Multidimensional scaling (MDS) analysis presenting the distribution of 357 compounds identified during the high-throughput screening. The respective IC50 values determined by QPatch measurements are represented, with compounds exhibiting IC50 values below ≤1 μM highlighted. (C) Assignment of lead compounds and active orthologs to five distinct compound series summarizing chemical features of the identified TRPM4 antagonists. Also see Figures S1–S3.

Figure 2

Medicinal chemistry of lead compounds and structure-activity relationship (A) Multidimensional scaling (MDS) analysis of 152 derivatives from Series 1 with IC50 values color coded for 78 active compounds. (B) Illustration of the correlation between lead similarity and IC50. (C) Series 1 compounds identified through high-throughput screening (HTS) and medicinal chemistry (MedChem), with IC50 values below 1 μM. The focus lies on modifications of the phenyl residue that maintain IC50s in the nanomolar range. (D) Exploration of structure-activity relationship (SAR) elements, determined through MedChem optimization in Series 1. (E) Comparison of Series 3 compounds with known TRPM4 antagonists of the anthranilic acid ortholog family. (F) Exploration of SAR elements determined through MedChem optimization in Series 3. Also see Figures S2–S5.

Figure 3

Neuroprotective in vitro evaluation and in vivo pharmacokinetics (A) In vitro evaluation of compound activity on glutamate-induced excitotoxicity in mature primary neuronal cultures. Neuronal cultures were treated with 5 μM of the specified compound or vehicle control 5 h prior to glutamate stimulation, assessing compound rescue activity and toxicity relative to minimal (Vehicle +0 μM Glu) and maximal (Vehicle +50 μM Glu) excitotoxicity. Time course of a representative compound with virtual endpoint at 15 h post stimulation and assay window definition (left). Toxicity of compound treated cultures with and without glutamate stimulation (right). Data are represented as mean and 95% confidence interval; n ≥ 4; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; One-sample Wilcoxon signed-rank test. (B) Assessment of CMP312 on mitochondrial integrity after glutamate-induced excitotoxicity. Mitochondrial membrane potential was measured by the mean fluorescence intensity of TMRE normalized to MitoTracker. Data are represented as mean ± SEM; n = 5; paired t-test. (C) In vivo pharmacokinetics of four selected compounds. Time course of compound concentration in the brain, heart, and plasma after a single dose via intraperitoneal (i.p.) or intravenous (i.v.) application. Data are represented as mean ± SEM. Half-life estimates by non-compartmental analysis are shown as vertical line and IC50 value as dashed horizontal line. (D) Compound concentration in the brain relative to respective IC50 concentrations after a single dose via intra-peritoneal or intra-venous routes. Also see Figure S6.

Figure 4

Electrophysiological analysis of proposed compound binding site (A) Proposed compound binding site. (left) Aligned superimposition of drTRPM5 structure (blue) in complex with its antagonist NDNA (green) and hsTRPM4 (magenta). Transmembrane helices (S) and the TRP domain are labeled. drTrpm5 binding pocket of NDNA (middle) and the corresponding pocket in hsTRPM4 (right). Divergent amino acids are labeled. (B and C) Whole cell voltage-clamp analysis of basic electrophysiological properties of HEK293T cells expressing wild-type (WT) or mutant hsTRPM4. (B) Representative time course of current levels at −80 and +80 mV recorded in hsTRPM4-expressing HEK293T cells. TRPM4 currents were elicited upon patch rupture by 100 μM CaCl₂, loaded in the pipette solution. Extracellular Na⁺ was replaced by equimolar N-methyl-D-glucamine (NMDG⁺) to identify TRPM4 currents. Individual current-to-voltage relationship measured in the same hsTRPM4-expressing HEK293T cells are shown next to the representative time course. IV trace numbers indicate the respective time points of the measurement to calculate peak current, maximal inhibition and plateau current. (C) Quantification of the peak current, maximal inactivated current and the steady-state plateau current between WT and hsTRPM4-L907A and hsTRPM4-S924A mutants. Data are represented as mean ± SEM. (D and E) Assessment of Ca²⁺ sensitivity of WT and mutant hsTRPM4. (D) Representative time course of current levels at −80 mV and +80 mV recorded in hsTRPM4-expressing HEK293T cells using the inside-out patch-clamp technique. TRPM4 currents were activated by 500 μM Ca²⁺ and deactivated by 10 mM EGTA. Individual current-to-voltage relationship measured in the same hsTRPM4-expressing HEK293T cell are shown next to the representative time course. (E) The Ca²⁺ concentration-to-current relationship of WT and hsTRPM4-L907A and hsTRPM4-S924A mutants. Likelihood maximization was used to estimate the parameters for the sigmoidal fitting curves with y-maximum, slope and midpoint of 0.94, 8.05 and 0.55 (WT); 0.90, 18.00 and 0.16 (L907A); 0.91 and 17.57 (S924A). (F) Quantification of relative inhibition in whole cell voltage-clamp configuration of WT and mutant hsTRPM4 by Series 1 compounds (CMP312, CMP343), Series 3 compound (CMP233) and CBA. Data are represented as mean ± SEM. (G) Comparison of relative inhibition in whole cell voltage-clamp configuration between mouse and human TRPM4. C: n ≥ 10; F: n ≥ 5; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; Wilcoxon rank-sum test. Also see Figures S7–S10.