Abstract by Kristine Egested Sloth Wilhelmsen

γ-Aminobutyric acid (GABA) is the main inhibitory neurotransmitter in the central nervous system (CNS) acting via GABA type A and GABA type B receptors (GABA_ARs, GABA_BRs). GABA_ARs are pentameric ligand-gated ion channels with the subunit composition determining cell-type and synaptic localization and pharmacological properties. The δ-subunit predominantly assembles into functional receptors as α_4/6βδ that are exclusively extrasynaptically located and mediate so-called tonic inhibition. Well-controlled tonic inhibition is essential for proper development and functioning of the CNS. Hence, dysregulated tonic inhibition is implicated in several neurological disorders including mood disorders, various types of epilepsy, and stroke. In 2023, the drug Zurzuvae^TM was the first δ-preferring drug to be registered for clinical use in post-partum depression by the U.S. Food and Drug Administration (FDA), marking a landmark in the field. The clinical relevance in potentiating or inhibiting tonic inhibition necessitates δ-selective ligands, of which very few have been reported. A specific challenge in the field is to fill the gap relating to the lack of δ-selective inhibitors. Such δ-selective ligands could provide valuable insight into the structure, pharmacology and physiology of δ-containing GABA_ARs. A relevant starting point is the delta-selective compound 2 (DS2), a positive allosteric modulator (PAM) of δ-containing GABA_ARs, yet with unclear molecular interaction sites and limited brain permeability. Thus, the overall aim of this PhD thesis was to develop and characterize novel tool compounds selective for δ-containing GABA_ARs to guide future drug development.

With DS2 as the structural scaffold, a small library of compounds (23 in total) was designed in-house, and initially profiled for allosteric modulatory activity at the α₄β₁δ receptor expressed in HEK293 cells using the FLIPR membrane potential (FMP) assay and whole-cell patch-clamp electrophysiology. This led to the identification of several novel PAMs (8 ligands), e.g., 1c displaying increased apparent efficacy over DS2 in the FMP assay. Most importantly, in this process, we identified a negative allosteric modulator (NAM) (1e), having only minor structural modifications compared to the identified PAMs. In follow-up experiments, 1e was found to be inactive at synaptic γ₂-containing and binary α₄β₃ GABA_ARs. However, at α₄β_2/3δ GABA_ARs, 1e acted as a PAM. An analogue IP-27 was later identified. Further studies are needed to clarify the β-subunit dependence and further selectivity profile of these compounds.

To discern PAM and NAM activity on the molecular level, potential α₄⁽⁺⁾δ^(-) and α₄⁽⁺⁾β₁^(-) binding interfaces in the transmembrane domain (TMD) were investigated in α₄β₁δ receptors. This was based on an unpublished cryogenic electron microscopy (cryo-EM) structure of α₄β₃δ with the DS2 analogue, Compound 30, as well as a previously suggested binding site for DS2. Site-directed mutagenesis and cell-based functional assays revealed the importance of the α₄Ser303 residue for PAM activity but not for NAM activity of DS2-related analogues. Conversely, the δGln253 residue located in the α₄⁽⁺⁾δ^(-) TMD interface, appeared as a key molecular determinant for discerning positive and negative modulation.

To directly investigate the binding sites for DS2-type compounds, a novel radioligand, [³H]-Compound 30, was developed and characterized in radioligand membrane binding assays and in vitro autoradiography. Whereas binding experiments in native cortical membranes indicated shared or overlapping binding sites of Compound 30, DS2 and 1e, and the expected rank order, binding to recombinant GABA_ARs was unsuccessful, precluding validation of the binding sites indicated from the mutational studies. Furthermore, autoradiography did not yield sufficient binding levels. Possibly due to a lack of sensitivity of [³H]-Compound 30.

Finally, first attempts of a biological proof-of-concept for a clinical potential of our novel NAMs were initiated through a collaborative setup. To this end, the ability of 1e to dampen gain-of-function (GOF) traits of α₄(T300I) and δ(L260V) mutations identified in patients with developmental and epileptic encephalopathies (DEE) was examined. In patch-clamp recordings of HEK cells expressing α₄β₁δ receptors carrying the respective mutations, the expected increase in GABA currents was confirmed. Furthermore, it was demonstrated that 1e could inhibit GABA-induced currents at recombinant receptors. In mouse slice recordings from thalamic relay neurons, both NAMs (10 μM) were able to increase agonist-induced tonic currents, but, interestingly, a marked inhibition of 1e was observed in pyramidal neurons of the somatosensory cortex. These differences may relate to different regional expression of α₄βδ receptors and warrants further studies. In parallel, both NAMs were found to have good brain permeability in mice, permitting future in vivo studies.

Overall, this PhD thesis has provided novel insights into tonic inhibition and identified novel tool compounds that may lay the groundwork for the understanding and further study of δ-containing GABA_AR pharmacology. Particularly, the work has led to the identification of the very first δ-subunit preferring NAMs, which signifies a shift in the field and may hold clinical potential in conditions of undesirably high tonic inhibition, such as GOF patient mutations or absence epilepsies where tonic inhibition is known to be unproductively high.