Design and Characterization of Novel Small Molecule Activators of Excitatory Amino Acid Transporter 2

Abstract

Excitotoxicity in the brain is a causal factor in several neurological and neurodegenerative disorders. Excitatory amino acid transporter 2 (EAAT2), an astrocytic glutamate transporter involved in the clearance of >80% of synaptic glutamate, is considered a therapeutically relevant target for excitotoxicity. We have previously designed GT951, an activator of EAAT2 with nanomolar efficacy but limited in vivo bioavailability. In this study, a pharmacophore-based screening and optimization resulted in the design of GTS467 and GTS511. GTS467 and GTS511 have low nanomolar efficacy in the glutamate uptake assay. Pharmacokinetic profiles (PK) of GTS511 show a >6 h half-life and higher bioavailability in plasma and the brain under all three routes of administration in rats. Similarly, GTS467 has high oral bioavailability (80–85%) in the brain and plasma with a >1 h half-life under all three dosing routes. These encouraging efficacy and PK profiles suggest that GTS511 and GTS467 can be further developed to treat neurological disorders caused by excitotoxicity.

Glutamate is a major excitatory neurotransmitter in the central nervous system that plays a key role in neuroplasticity, cognition, learning, memory, and development.¹⁻³ The neurotransmission of glutamate is tightly regulated by a network of metabotropic and ionotropic receptors and transporters that are distributed on neurons, glia, and other cell types in the brain.^4,5 Under healthy conditions, depolarization of the glutamate neurons leads to the release of glutamate by the presynaptic vesicles in a calcium-dependent manner, which then activates the post synaptic ionotropic NMDA (N-methyl-d-aspartate) and AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) receptors facilitating the activation of calcium dependent signaling cascades in the post synaptic neuron.⁶⁻⁸ Excess glutamate is quickly removed from the synapse by the astrocytic glutamate reuptake transporters, which is an active transport involving cotransport of glutamate along with 3 Na⁺ followed by counter transport of K⁺ ion.^6,9,10 This removal of excess glutamate from the synapse prevents hyper stimulation of the post synaptic and extra synaptic NMDA (N-methyl-d-aspartate) receptors to prevent excitotoxicity.^6,11

The glutamate transporters are classified into five subtypes in rodents and humans.¹² The three rodent transporters are called GLAST (glutamate–aspartate transporter), GLT1 (glutamate transporter 1), and EAAC1 (excitatory amino acid carrier 1), while the human homologues are called the excitatory amino acid transporters (EAAT) EAAT1, EAAT2, and EAAT3, respectively.^13,14,11 The remaining two EAAT4 and EAAT5 share common nomenclature in rodents and humans.⁶ These transporters are differentially expressed in various brain regions or cell types and accordingly have different functional significance.^9,11,12 EAAT2 is the most common subtype and accounts for 85–90% of glutamate reuptake from the terminals.^6,15,16 Dysfunction or downregulation of EAAT2 can lead to excess accumulation of glutamate causing calcium mediated excitotoxicity and death.¹⁷⁻¹⁹ This form of excitotoxicity has been implicated in several neurological disorders including amyotrophic lateral sclerosis (ALS), Alzheimer’s disease (AD), Huntington’s disease (HD), and Parkinson’s disease (PD), and dysfunction of EAAT2 has been documented in these disorders, suggesting EAAT2 could be a novel drug target to treat excitotoxicity in these disorders.²⁰⁻²³ In fact, ceftriaxone (Table-1), a β-lactam antibiotic, which has nonspecific EAAT2 activator function, was advanced through Phase 3 clinical trials for treatment of ALS.^24,25 In an effort to develop novel EAAT2 activators, a cell-based enzyme linked immunosorbent assay (ELISA) was developed where the primary astrocyte line was stably transfected with the designed vector for EAAT2 modulation.²⁶ This assay was used for high throughput screening of a library of compounds for glutamate reuptake. Several hits were identified with micromolar efficacy in a glutamate reuptake assay in a modified PA-EAAT2 cell line (rat primary astrocyte cell line lacking endogenous EAAT2 expression but stably express hEAAT2 or human EAAT2 promoter).^26,27 More recently, a small molecule activator of EAAT2, LDN/OSU-0212320 (Table 1), was designed and tested for efficacy in preclinical models of ALS and epilepsy.^28,29 However, none of these molecules have progressed to become FDA-approved drugs due to either off-target toxicity or lack of translatable physicochemical properties.^30,31 In this study, we developed novel activators of EAAT2 using hybrid structure-based screening and optimization. The molecules were tested for glutamate uptake using a transflex plate seeded with the Madin Darby Canine Kidney (MDCK) cell line overexpressing hEAAT2, and their PK profile was further characterized in rats.

Table 1. Chemical Structures of Known EAAT2 Activators and Their EC50 Values in Glutamate Uptake Assays.

Our previous efforts to design and develop small molecule activators of EAAT2 resulted in a series of tetrazole core containing compounds exemplified by GT951 and GT949.³² GT951 and GT949 demonstrated high potency to EAAT2 in the in vitro glutamate uptake assay in COS cells overexpressing EAAT2, with EC50 values of 0.8 nM and 0.26 nM, respectively (Table 1) and in primary astrocytes, the EC50 values were ∼0.3 nM and 1 nM, respectively. Interestingly, the molecules showed high selectivity to EAAT2 in comparison to other EAATs and did not have any functional effects on NMDA-mediated calcium response.³² Further, using site-directed mutagenesis studies, the likely binding site of these molecules was confirmed to bind to an allosteric site at the interface between the trimerization and transport domain but proximal to the substrate binding domain.³² Despite the high efficacy of these molecules, they have very poor drug-like properties including high lipophilicity (logP > 6) and poor aqueous solubility which limit their bioavailability in vivo. To evaluate the bioavailability of GT951 in plasma and brain, we evaluated the pharmacokinetics of GT951 in CD-1 mice. GT951 was formulated as a suspension in 0.5% w/v methylcellulose and injected intraperitoneally to CD-1 mice at 10 mg/kg dose. Results from the study showed that GT951 had half-lives of 3 and 3.5 h and C_max values of 717 and 30 ng/mL in plasma and brain tissue, respectively (Table 2).

Table 2. Pharmacokinetic Profile of GT951 Evaluated in CD-1 Mice Dosed with 10 mg/kg Intraperitoneally in Plasma and Brain Samples.

parameter	GT951 brain	GT951 plasma
dose(mg/kg)	10	10
AUC	86	1339
T1/2α (h)	3.5	3
Cmax (ng/mL)	30	717
Tmax (h)	0.5	0.5

To develop EAAT2 activators with better pharmacokinetic profile and drug-like properties, we designed a pharmacophore using the interaction profile of GT951 with EAAT2 and screened a drug-like library of small molecules to identify 3822 hit molecules. These molecules were then docked to the binding pocket of EAAT2 lined by residues such as Val191, Lys299, Ala362, Asp471, Trp472, Asp475, Arg476, and Thr479. Ten best ranking molecules that had high docking scores were then synthesized and purified to >95% purity as determined by ¹H NMR.

Ten test molecules along with GT951, which was used as a positive control, were tested at 1 and 3 μM concentration in a glutamate uptake assay using BioIVT’s kit with their customized Transflex plate. Glutamate levels in cell lysates were quantified by high-resolution liquid chromatography mass spectrometry (LCMS). To improve separation and detection, we precolumn derivatized samples with benzoyl chloride, which has been utilized for glutamate detection in complex samples including microdialysates.³³⁻³⁵ After derivatization, molecules were separated by UPLC and analyzed by high-resolution mass spectrometry. We observed a signal for the [M + H]+ ion and a much stronger signal for the sodium adduct [M + Na]+ of benzylated glutamate. Using a calibration curve of an authentic glutamate standard, glutamate levels in each of the samples were quantified. Screening was performed at 1 and 3 μM and compounds that increased glutamate levels in cells beyond that of vehicle control (normalized to 100%) were considered as hits (Figure 1). Compounds GTS467, GTS477, GTS511, GTS619, and the positive control GT951 had increased levels of glutamate at 1 and 3 μM and were then tested for dose dependency to derive EC50 values. It must be noted that some of the compounds had a trend toward a decrease in glutamate levels at higher concentration but were not statistically different from 1 μM, suggesting saturating effects at the transporters. Using the same assay, the EC₅₀ values for these compounds were evaluated with GTS467 as 35.1 ± 1.0 nM; GTS511 as 3.8 ± 2.2 nM; GTS619 as 78.5 ± 1.0 nM; and GT951 as 11.4 ± 3 nM. GTS477 was unstable at room temperature and was not further tested. Docking studies show that all these molecules have strong interactions at the binding site facilitated by hydrogen-bonded interactions, cation-π and hydrophobic interactions likely contributing to their high docking scores and EC₅₀ values (Figure 2). In addition to the racemate, the two enantiomers of GTS467 were also docked independently and the results showed similar interaction profiles and docking scores suggesting they may have similar activities. Future studies will include enantiomeric separation and testing of the individual isomers.

Figure 1.

Molecules were screened at 1 and 3 μM concentrations in a glutamate uptake assay in MDCK cells transfected with EAAT2. Results were normalized to the glutamate levels from vehicle control (dotted line). Compounds GTS467, GTS477, GTS511, GTS619, and GT951 were identified as positive hits from this assay.

Figure 2.

Schematic representation of compounds (A) GTS467 and (B) GTS511 docked to the binding pocket of EAAT2 generated using ligand interaction module of MOE shows several conserved interactions between ligand molecules in the binding pocket that resulted in high docking scores.

Based on their in vitro profiles, GTS467 and GTS511 were then tested for their pharmacokinetic profiles in male Wistar rats at under all three routes of administration (intraperitoneal, IP, at 10 mg/kg; postoral, PO, at 10 mg/kg; and intravenous, IV, at 5 mg/kg), and their profiles in plasma and brain are shown in Figure 3. GTS511 had C_max values of 567, 450, and 42.86 ng/mL and a half-life of 6.08, 6.22, and 5.34 h in plasma under IV, IP, and PO dosing, respectively, with a bioavailability of 58.36 and 17.51% for IP and PO dosing. In brain samples, GTS511 had a C_max of 370.6, 204.1, and 24.98 ng/mg of brain tissue with a half-life of 1.52, 2.26, and undetectable levels under IV, IP, and PO dosing, respectively (Table 3). GTS467 showed C_max values of 307, 262, and 204 and half-lives of 0.6, 1, and 4.76 h under IV, IP, and PO dosing, respectively, in plasma. GTS467 also had high bioavailability of 85 and 80% under IP and PO dosing, respectively, in plasma. In brain lysates, GTS467 had C_max values of 1221, 2461, and 1173 ng/mg of brain tissue and half-lives of 1.03, 1.62, and 1.81 h under IV, IP, and PO dosing, respectively (Table 4).

Figure 3.

Pharmacokinetic profile of GTS511 in (A) plasma and (B) brain; GTS467 in (C) plasma and (D) brain under IV, IP, and PO administration is shown. The profiles show GTS511 has low bioavailability under oral dosing while GTS467 had ∼10 fold higher concentration in brain tissue compared to its plasma distribution (note the change in the Y-axis scale).

Table 3. Pharmacokinetic Profile of GTS511 Evaluated in Wistar Rats Dosed with 10 mg/kg IP and PO or 5 mg/kg IV in Plasma and Brain Samples.

	brain route (dose)			plasma route (dose)
parameter	IV (5 mg/kg)	IP (10 mg/kg)	PO (10 mg/kg)	IV (5 mg/kg)	IP (10 mg/kg)	PO (10 mg/kg)
Cmax(ng/g)	370.60	204.10	24.98	567.22	450.27	42.86
Tmax (h)	0.50	0.50	1.38	0.50	0.63	3.25
AUC (h ng/g)	443.23	429.85	183.00	919.43	1073.21	321.93
t1/2 (h)	1.52	2.26	UD	6.08	6.22	5.34
F (%)	NA	ND	ND	NA	58.36	17.51

Table 4. Pharmacokinetic Profile of GTS467 Evaluated in Wistar Rats Dosed with 10 mg/kg IP and PO or 5 mg/kg IV in Plasma and Brain Samples.

	brain route (dose)			plasma route (dose)
parameter	IV (5 mg/kg)	IP (10 mg/kg)	PO (10 mg/kg)	IV (5 mg/kg)	IP (10 mg/kg)	PO (10 mg/kg)
Cmax (ng/g)	1221.05	2461.17	1173.13	307.055	262.370	204.40
Tmax (h)	0.50	0.50	0.88	0.500	0.500	0.875
AUC (h ng/g)	1959.70	4414.35	2816.80	315.635	536.768	507.172
t1/2 (h)	1.03	1.62	1.81	0.651	1.016	4.764
F (%)	NA	ND	ND	NA	85.03	80.34

In conclusion, results from our study show that GTS467 and GTS511 are novel EAAT2 activators that have high affinity to EAAT2 and favorable drug-like properties. EAAT2 is a therapeutically viable target and plays a significant role in glutamate homeostasis in the brain. Dysregulation of glutamate transport leads to excitotoxicity—a condition that is known to be either causal or can exacerbate conditions such as PD, HD, ALS, stroke, and epilepsy. Currently, there are no known drugs on the market that can treat excitotoxicity without inducing fatal side effects. Thus, development of novel molecules such as GTS467 and GTS511 is essential. Current and future studies in our laboratory will include testing these compounds in rodent models of HD, PD, and epilepsy.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.2c00304.

• All the experimental details relevant to the compound synthetic protocols, synthesis validations, compound structural characterizations and purification, compound testing, and pharmacokinetic evaluations (PDF)

The authors declare the following competing financial interest(s): S.K. is the named inventor on the U.S. Patent Application 17/847,490, filed June 23, 2022, which covers all the molecules described in the manuscript.