not-yet-known not-yet-known not-yet-known unknown Introduction Epilepsy is a common neurological disorder that affects over 1% of the global population and has substantial societal consequences, with about 30% of the patients being refractory or becoming refractory to commonly available antiseizure medications (Kwan et al., 2011; Kalilani et al., 2018, Löscher & Klein, 2021). Furthermore, most of the available drugs have considerable cognitive side effects since they act directly through key molecules in brain function. An ideal antiseizure medication would be one affecting only diseased tissue, to modulate, rather than directly block, the activity of molecules relevant to the control of brain activity. Such a drug would be expected to display fewer cognitive side effects. Adenosine is a homeostatic/allostatic modulator of synaptic activity (Diógenes et al., 2014; Xiao et al., 2019; Jacobson et al., 2022) due to the balanced activity of enzymes and transporters that control its intracellular and extracellular concentration (Boison & Jarvis, 2021). Once in the extracellular space adenosine activates membrane embedded G protein-coupled receptors, four subtypes being known, A1R, A2AR, A2BR and A3R (Sebastião & Ribeiro, 2009; Gao et al., 2023). Adenosine has long been considered an endogenous anticonvulsant (Dunwiddie, 1980). Dysregulation of the homeostatic/allostatic control of synaptic activity by adenosine may lead to epilepsy (Sandau et al., 2016). The antiseizure action of adenosine is mediated by inhibitory A1R present in brain tissues, namely, cerebral cortex and hippocampus (Dunwiddie & Masino, 2001; Sebastião & Ribeiro, 2009). However, the widespread distribution of the A1R throughout the body has hampered the development of A1R modulators as antiseizure medications. Indeed, the A1R is also widely expressed in the nerve endings of the autonomic nervous system and in effector organs, including the heart. Therefore, a major concern is the bradycardic and negative inotropic actions of A1R agonists (Baltos et al., 2023). Numerous adenosine derivatives have been synthesized over the last two decades, to identify molecules that could serve as prototypes of novel therapeutics. This was the case of MRS5474, a truncated N-methanocarba nucleoside that has selectivity for A1R (Ki: 48 nM for human; 3.2 nM for mouse) over the A3R (Ki: 470 nM for human, 1056 nM for mouse), and low affinity for the A2AR (Ki 4 µM for human, > 10 µM for mouse), as assessed by standard radioligand binding approaches in CHO or HEK cells expressing the human or mouse adenosine receptor subtypes (Tosh et al., 2012a; Carlin et al., 2017). Functional data showed that MRS5474 can act as a full agonist of the A1R and partial agonist of the A3R, having negligible activity on the A2BR (Tosh et al., 2012a). A striking characteristic of MRS5474, reported by Tosh et al (2012a), was its antiseizure activity in the 6 Hz minimal clonic seizure mouse model of epilepsy. However, despite acting as a full A1R agonist, as assessed by inhibition of adenylate cyclase activity in CHO cells transfected with the human (h)A1R, it did not cause alterations in the RotaRod test in mice, indicating the absence of cardiac inhibitory actions. This prompted us to hypothesize that MRS5474 could act through another receptor than A1R and/or through a brain/disease specific mechanism. To explore these possibilities, we tested the effects of MRS5474 in the hippocampus in a variety of models, including animal and human hippocampal tissue, addressing both excitatory and inhibitory transmission endpoints. Taken together, data shows that MRS5474, under the experimental conditions, rather than activating an A1R, activates A3R that are likely overexpressed in epileptic tissue. Notably, the actions of MRS5474 in excitatory synaptic transmission as well as in GABAergic currents could only be detected in epileptic tissue but not in control tissue, underscoring A3R as promising targets for the development of antiseizure medications.