not-yet-known not-yet-known not-yet-known unknown Discussion and Conclusions The major finding in the present work is that MRS5474 has selective actions in hippocampal epileptic tissue. Another significant discovery involves the identification of adenosine A3R as the targets for MRS5474 actions within the hippocampus. The evidence that extracellular levels of adenosine increase during seizures (Dale and Frenguelli, 2009), raised attention for the antiseizure role of adenosine and of its inhibitory A1R (Rombo et al, 2018; Beamer et al., 2021). However, the druggability of A1R is hampered by their ubiquitous distribution, a major cause for the on-target side effects that have been identified in pre-clinical studies focusing on A1R in epilepsy. Among unwanted effects are bradycardia, atrioventricular block, a reduction in atrial contractility, and sedation (Jacobson and Gao, 2006; Nguyen et al., 2023). The search for structurally modified A1R ligands that could achieve selectivity for some of the therapeutically relevant actions of these receptors, led to the discovery of the distinguishable properties of MRS5474 (Tosh et al., 2012a). MRS5474 is a small molecule with favourable pharmacokinetics. In the mouse, 1mg/kg (p.o) has a bioavailability 97.7%, a long half-life (4.96 hours), an exposure (AUC0-inf) of 3700 ng·h/mL, a plasma concentration maintained at >530 nM for at least 8 hours, a clearance of 4.52 mL/min/kg, and distribution volume of 1.94 L/kg (Tosh et al., 2019). No lethality was detected at concentrations up to 60 mg/kg (i.p.) (Tosh et al., 2019), which is remarkably higher than the ED50 value for its antiseizure action (Tosh et al., 2012a). Using the StarDrop software (Segall, 2012), MRS5474 is predicted to cross the blood-brain barrier more readily than other adenosine receptor ligands, as CPA, with a predicted brain:blood ratio of 11.7%. It was tested in gpcrMAX and Kinome screens (Eurofins DiscoverX, Fremont, CA, USA) and found to have no major off-target interactions. Altogether, these pharmacokinetic parameters highlight the suitability of MRS5474 for drug development. MRS5474 has A1R selectivity and agonist-like properties (Tosh et al., 2012a; Carlin et al., 2017). Therefore, it was highly surprising when we observed the lack of effect of MRS5474 upon synaptic potentials and excitatory synaptic currents at the hippocampus, since the inhibitory actions of A1R in the hippocampus have been known for a long time (Sebastião et al., 1990). Searching for an action upon GABAergic transmission (Sebastião and Ribeiro, 2023), we detected that MRS5474 inhibits GAT-1 mediated GABA uptake, but through an A3R rather than an A1R mediated mechanism. Indeed, the inhibitory effect of MRS5474 upon GAT-1 mediated GABA was insensitive to A1R selective antagonism, was mimicked by a selective A3R agonist, and was fully inhibited by MRS1523, an A3R antagonist. MRS1523 is selective for A3R but at the concentrations used it may also antagonize A2AR (Ki 2µM for rat A2AR) (Li et al., 1998). However, the involvement of A2AR in the inhibitory action of MRS5474 upon GAT-1 is highly unlikely since A2AR facilitate, rather than inhibit, GAT-1 in nerve endings (Cristóvão-Ferreira et al., 2009) and astrocytes (Cristóvão-Ferreira et al., 2013). Clinical trials with A3R agonists proved them to be well tolerated and with few side effects (Jacobson et al., 2019; Coppi et al., 2022). The A3R has a very low expression in cardiomyocytes and was even shown to be cardioprotective (Wan et al., 2019), as well as neuroprotective (Von Lubitz et al., 1999; Cheng et al., 2022). Of high importance is the neuroprotective potential of a mixed A1R/A3R agonist in both rodent and non-human primate models of stroke (Liston et al., 2020, 2022), which led to two human Phase I clinical safety trials and a likely approaching Phase II clinical trial for acute stroke and traumatic brain injury (https://www.astrocytepharma.com/portfolio/, accessed March 22, 2024). Interestingly, stroke and traumatic brain injury are leading causes of epilepsy. Neuroinflammation is also a hallmark of epilepsy (Palumbo et al., 2023; Vezzani et al., 2023), and A3R agonists have anti-inflammatory actions (Jacobson et al., 2019). The role of A3R in epilepsy has, however, been controversial (Świąder et al., 2014; Rombo et al., 2018; Beamer et al., 2021; Tescarollo et al., 2020). This may result from the initial lack of selective ligands for A3R (Gao et al., 2023), as well as a lack of knowledge of the species differences in ligand selectivity. Importantly, we now proved effectiveness of MRS5474 in human tissue, and showed that it acts through A3R. Another issue to consider when addressing the action of A3R in epilepsy, is the involvement of the GABAergic system. An A3R-mediated pro-epileptiform action has been detected under GABAA receptor blockade conditions (Laudadio & Psarropoulou, 2004), therefore precluding the observation of any modification of GABAergic transmission. As we show here, A3R may predominantly affect GABAergic rather than glutamatergic transmission. It may also be relevant to distinguish between phasic inhibition (mediated by synaptic GABAA receptors) from tonic inhibition (mediated by extrasynaptic GABAA receptors). Drugs with A3R antagonistic characteristics have been shown to decrease GABAA receptor desensitization in oocytes injected with tissue samples from a variety of human epileptic tissues (Roseti et al., 2008; 2009), which is compatible with a pro-convulsive action of A3R. In contrast, we observed that MRS5474, by activating A3R, selectively potentiates GABAergic currents in oocytes injected with human epileptic hippocampal tissue, thus an action compatible with an antiseizure activity. Whole GABAergic currents include those mediated by synaptic (fast desensitizing) and extrasynaptic GABAA (slow desensitizing) receptors. One may thus suggest that MRS5474, by activating A3R, predominantly affects slow-desensitizing extrasynaptic GABAergic currents. In intact tissues, inhibition of GAT-1 by MRS5474 will even favour an extrasynaptic action, since it would facilitate GABA spillover to extrasynaptic sites. Extrasynaptic GABAA receptor activation is of high value for seizure control (Richerson, 2004; Naylor, 2023). The use of human epileptic fresh slices would allow to address the relative activity of MRS5474 and of A3R in tonic (extrasynaptic) vs phasic (synaptic) inhibition, as well as directly evaluating the action of A3R upon GAT-1 activity in humans. However, the availability of fresh human tissues is scarce, and even more for control tissues. Nevertheless, we obtained evidence for overexpression of A3R in human epileptic tissue, which likely acts as an amplification mechanism and explains the lack of effect of MRS5474 in non-epileptic tissue. We could not detect the well-known A1R-mediated inhibition of excitatory inputs to CA1 pyramidal cells in non-epileptic rodents, even though MRS5474 binds to both A1R and A3R, with even higher affinity for A1R (Tosh et al., 2012a; Carlin et al., 2017). The reasons for this lack of A1R-mediated effects of MRS5474 may result, at least in part, from the ability of A3R to desensitize A1R in healthy tissue (Dunwiddie et al., 1997). Alternatively, one may speculate on biased agonism (McNeill et al., 2021) by MRS5474. Context-specific biased agonism (McNeill et al., 2021) could also help to explain the epilepsy-selective actions of this compound detected in the present work. Further studies are worthwhile to explore this possibility and its putative therapeutic implications. The present results, together with previous evidence that MRS5474 has antiseizure action without appreciable negative side effects in cardiac function (Tosh et al., 2012a), strongly support the idea that targeting A3R in epilepsy may lead to fewer side effects than targeting A1R. Targeting A3R may also have fewer side effects than targeting GAT-1. MRS5474 could inhibit GAT-1 even in non-epileptic tissue but the change in extracellular GABA concentrations is likely small since it did not affect excitatory transmission in healthy tissue. This may result from either overexpression of A3R in epileptic tissue (present work), amplifying A3R mediated actions, and/or alteration in GAT-1 in epilepsy (Medina-Ceja et al., 2012, Su et al., 2015). Inhibitors of GAT-1, such as tiagabine currently used in clinics, have sedative effects (Kälviäinen, 2001; Bauer & Cooper-Mahkorn, 2008, Masocha & Parvathy, 2016), though clearly of lower magnitude than benzodiazepines. In contrast, no sedative effects have been detected in animals administered with MRS5474 (Tosh et al., 2012a), which may be related to our finding that MRS5474 selectively affects hyperexcitable tissue, mostly sparing non-epileptic tissue. In conclusion , this work provided, for the first time, evidence of an A3R-mediated action upon GABA transporters and GABAergic currents. It also provides evidence that a drug with A3R agonist properties, MRS5474, predominantly affects epileptic tissue. Therefore, we disclosed a novel mechanism of action of a drug that proved to be effective in an animal model of epilepsy and highlighted A3R as a target for antiseizure medications. These findings pave the way for further preclinical studies to explore the therapeutic potential of MRS5474 or other A3R agonists against epilepsy, either as adds-on drugs with other antiseizure drugs or even as antiepileptogenic drugs, due to their combined anti-inflammatory action.