The intracellular signaling pathways related to GUO effects were the first targets to be identified. It was already shown that GUO effect on cell proliferation is dependent on cyclic AMP (cAMP) level increase (Gysbers and Rathbone 1996; Su et al., 2009). Also, some protective effects are directly related to Phosphoinositide 3-kinase/Protein kinase B (PI3K/Akt) pathway (Dal-Cim et al., 2013; Dal-Cim et al., 2012; Molz et al., 2011; D'Alimonte et al., 2007; Giuliani et al., 2015) and the mitogen-activated protein kinase/extracellular signal-regulated kinase (MAPK/ERK) pathway (Di Iorio et al., 2004; Dal-Cim et al., 2011). Recently, the involvement of protein kinase C (PKC) was also identified, as GUO effect of increasing the glial glutamate transporter-1 (Glt-1) membrane expression after an oxygen/glucose deprivation (OGD) protocol in astrocytes is abolished by PKC or MAPK/ERK inhibition (Dal-Cim et al., 2019).

Since GUO effects evoke intracellular signaling pathways, the need for a membrane receptor target is claimed. The effects of GUO are not altered by nucleosides transporter blockers (Oleskovicz et al., 2008; Giuliani et al., 2015; Decker et al., 2019), indicating that its effects are mediated through interaction with some molecular target at the cellular membrane not yet identified. Some studies pointed to a putative selective GUO receptor in rat brain membranes through [³H]GUO binding analysis (Traversa et al., 2002; Traversa et al., 2003). Additionally, a study from Volpini and collaborators suggested GUO interaction with a G-protein coupled receptor (GPCR) (Volpini et al., 2011). In this line of evidence, the pharmacological blockade of GUO neuroprotective effect against ischemia-like in vitro protocol with Pertussis toxin also indicates a GPCR interaction (Dal-Cim et al., 2013). While the extracellular effects of purinergic adenine-based nucleotides and nucleoside are better characterized and their respective P2 and P1 receptors are recognized (Palmer and Stiles 1995), guanine-based purines are still orphan molecules.

Apart from the purinergic system, it was shown that GUO effects may depend on a potassium channel interaction. GUO effects of increasing cellular viability in hippocampal slices subjected to OGD and in SH-SY5Y neuroblastoma cells subjected to mitochondrial damage are blocked by large (big) conductance calcium-activated potassium channels (BK) inhibitors (Oleskovicz et al., 2008; Dal-Cim et al., 2012; Dal-Cim et al., 2013). A couple of studies also suggested the relation of GUO effects with GPR23 (Di Liberto et al., 2012) or CD40 receptors (D'Alimonte et al., 2007). Although there is a suggestion for GUO receptors (a selective one, or other putative receptors), a great number of results support that GUO effects are mediated by adenosine receptors (Lanznaster, et al., 2016).

Adenosine plays a pivotal role as a neuromodulator and presents neurotrophic effects acting through its P1 receptors, which are composed of four different GPCRs (A₁R, A_2AR, A_2BR, and A₃R). A₁R and A₃R are typically coupled to Gi proteins and thus inhibit adenylyl cyclase activity, whereas A_2AR and A_2BR are coupled to Gs proteins and increase the production of cAMP (Zimmermann, 2011). P1 receptors are expressed in neurons, astrocytes, oligodendrocytes and microglia and their stimulation activates multiple functions, such as synaptic plasticity and presynaptic neuromodulation (Daré et al., 2007; Burnstock, Fredholm, and Verkhratsky 2011; Burnstock and Ulrich 2011). In addition, A₁R and A_2AR are the main responsible for adenosine actions on the CNS, while A₁R is well expressed in the whole brain, A_2AR is enriched in some particular areas such as the striatum, hippocampus, raphe nuclei and locus coeruleus (Palmer and Stiles 1995). And, to our knowledge, besides one study showing that a preferential A_2BR antagonist partially decreased the mitogenic activity of GUO in astrocytes (Ciccarelli et al., 2000), only A₁R and A_2AR have been related to GUO effects.

Some results directly imply GUO effects with the A₁R. Most of GUO known effects are abolished by previous incubation or treatment with the selective A₁R antagonist DPCPX. In vitro protocols of brain ischemia in hippocampus slices and cortical astrocyte cultures demonstrated that DPCPX abolished the protective effects of GUO in ROS production, glutamate uptake, and cell viability (Dal-Cim et al., 2013; Dal-Cim et al., 2019). Similarly, DPCPX prevented GUO protective effect against mitochondrial oxidative stress in human neuroblastoma SH-SY5Y cells (Dal-Cim et al., 2012). Also, it was shown that in primary culture of both neurons and astrocytes, GUO increases global Small Ubiquitin-like MOdifier (SUMO)2/3-ylation at neuroprotective concentrations, an effect abolished by DPCPX preincubation (Zanella et al., 2020). In vivo protocols also display the same pattern related to A₁R. In a reserpine-treated mice protocol, where animals develop a parkinsonian tremor and striatal damage, GUO reverses the motor impairment and decreases ROS level in the striatum, but GUO efficacy is lost when animals are pretreated with DPCPX (Massari et al., 2020). In a traumatic brain injury model in rats, it was seen that mitochondrial dysfunction in the cerebral cortex is reversed by GUO treatment. However, this effect is no longer seen if the animals are pretreated with DPCPX (Gerbatin et al., 2019). In the same way, DPCPX reversed the anxiolytic-like effect induced by GUO, as well as the GUO capacity of decreasing the synaptosomal K⁺-stimulated glutamate release (Almeida et al., 2017). It is important to mention that DPCPX is also considered to be an inverse agonist of A₁R (Weyler et al., 2006). Additionally, reports are now revealing some molecules that display a biased agonism (a ligand-dependent differential intracellular signaling) on A₁R, an issue that still needs additional studies (Vecchio et al., 2018). Taken together, these data could suggest that GUO effects are mediated by A₁R activation. However, in heterologous transfection of A₁R in HEK293 cells, GUO does not induce calcium mobilization as observed with an A₁R agonist (R-PIA) treatment (as a positive control). Moreover, GUO has no effect upon R-PIA-inducing calcium mobilization through A₁R (Lanznaster, et al., 2019).

Data regarding GUO dependence on the A_2AR signaling are conflicting. While most data show that antagonism of A_2AR has no impact on GUO promoting effects (Almeida et al., 2016; Almeida et al., 2017; Gerbatin et al., 2019; Massari et al., 2020; Zanella et al., 2020) some reports are showing otherwise (Dal-Cim et al., 2012; Decker et al., 2019). Surprisingly, the pharmacological use of the A_2AR agonist CGS21680 shows a clear counteraction of GUO-mediating effects. Like the pretreatment with an A₁R antagonist, A_2AR agonist also abolishes the protective effects of GUO over ROS production, glutamate uptake, and cell viability on those in vitro protocols of brain ischemia (Dal-Cim et al., 2013; Dal-Cim et al., 2019). Recently, it was seen that in mice that do not express A_2AR (i.e. A_2AR Knock-out mice, A2AR^−/-) the preventive GUO effect on ROS production and on cell viability is impaired (Lanznaster et al., 2019). Important to notice, GUO per se does not induce cAMP increase in HEK293 cells transfected with A_2AR, neither interfere with cAMP level increase induced by the A_2AR agonist CGS 26180 (Lanznaster et al., 2019).

The dubious effect of GUO on adenosine receptors can also be interpreted through the oligomeric interaction of these receptors.

The understanding of GPCRs physiology and pharmacology has changed in the last 2 decades. This is due to the growing evidence that they can form homomers (homo-oligomerization, from the same GPCRs) and heteromers (hetero-oligomerization of different GPCRs). This oligomerization induces changes in biochemical properties of GPCRs. It is well established that adenosine receptors can form oligomers among themselves and with receptors for other neurotransmitters, such as dopamine receptors (Ginés et al., 2000; Fuxe et al., 2005; Ciruela et al., 2006; Navarro, et al., 2018; Borroto-Escuela et al., 2018; Ferré and Ciruela 2019; Cortés et al., 2019). It is known that A₁ and A_2A receptors form functional oligomers with each other and that the A₁R-A_2AR heteromer plays an important role in modulating the control of cortico-striatal function (Ciruela et al., 2006). This control takes place through the activation of the presynaptic A₁R or A_2AR, which depends on the concentration of adenosine, as a low concentration would activate A₁R while a high concentration would activate A_2AR, resulting in a lesser or greater release of glutamate, respectively (Ciruela et al., 2006). Moreover, the A₁R-A_2AR heteromer seems to have a role in glutamate clearance by modulating the expression of the excitatory amino acid transporter 2 (EAAT2) in astrocytes (Hou et al., 2020). Also, adenosine interaction with A₁R-A_2AR heteromer in astrocytes has been shown to control extracellular gamma-aminobutyric acid (GABA) uptake via modulation of GABA transporters (Cristóvão-Ferreira et al., 2013). In this way, it is proposed that A₁R-A_2AR heteromer works as an adenosine concentration-sensing device that implies a cross-communication between Gi and Gs proteins guided by the C-terminal tail of the A_2AR (Navarro, Cordomí, Brugarolas, et al., 2018).

The structure of A₁R and A_2AR heterodimerization was recently proposed through in silico molecular modeling (Navarro et al., 2016; Navarro, Cordomí, Brugarolas, et al., 2018). A₁R-A_2AR heteromer may be organized as a tetramer structure composed of two homodimers of A₁R and two homodimers of A_2AR. The homodimerization of A₁R and A_2AR occurs through the transmembrane (TM) 4/5 interface while the heterodimerization takes place through the TM 5/6 interface of these receptors (Navarro, Cordomí, Brugarolas, et al., 2018).

In this line, using an in vitro approach with transfected HEK293 cells, we recently showed that GUO-induced effects require both A₁R and A_2AR co-expression. GUO was able to decrease A_2AR binding affinity and cAMP response evoked by a selective A_2AR ligand but only in cells expressing both A₁R and A_2AR. Also, GUO had no effect on A₁R signaling in the presence or absence of A_2AR co-expression (Lanznaster et al., 2019). Thereby, we interpret that GUO interacts with the adenosine receptors as a heteromeric entity, thus the most adjusted hypothesis is that GUO could be acting as a negative modulator of A_2AR, but only in the presence of A₁R. It is feasible to speculate that the physical interaction between A₁R and A_2AR could lead to an increase of A_2AR affinity for GUO. CGS21680 could be directly interfering in the GUO signaling on A_2AR, whereas DPCPX interacting with A₁R may be responsible for allosteric modulation of GUO A_2AR affinity upon the A₁R-A_2AR heteromers (Figure 1). Indeed, GUO modulation over other adenosine-forming heteromers could not be discarded and might be different among brain structures, once that it may depend on differential receptors expression. Intriguingly, the GUO protective effect is lost in hippocampal but not in striatal slices from A2AR^−/- mice (Lanznaster et al., 2019; Massari et al., 2020). Therefore, more studies regarding GUO interactions with adenosine oligomers are necessary.