Blog

By David Aguinaga, Mireia Medrano, Ignacio Vega-Quiroga, Katia Gysling, Enric I. Canela, Gemma Navarro, and Rafael Franco

Excerpt from the article published in Frontiers in Molecular Neuroscience, 12 February 2018, Sec. Molecular Signalling and Pathways, Volume 11 – 2018, DOI: https://doi.org/10.3389/fnmol.2018.00017

Editor’s Highlitghts

• In acute cocaine exposure, σ₁R modulation of D₁R-mediated signaling prevails, but in longer exposures, there is a shift to regulation by σ₂R. 
• σ₂R becomes the main player in conditions of chronic exposure to the drug.
• Some of the addictive and motor actions of cocaine are the result of a balance between cocaine-σ₁R versus cocaine-σ₂R impact on activation of D₁R and D₂R.

Abstract

Sigma σ₁ and σ₂ receptors are targets of cocaine. Despite sharing a similar name, the two receptors are structurally unrelated and their physiological role is unknown. Cocaine increases the level of dopamine, a key neurotransmitter in CNS motor control and reward areas. While the drug also affects dopaminergic signaling by allosteric modulations exerted by σ₁R interacting with dopamine D₁and D₂ receptors, the potential regulation of dopaminergic transmission by σ₂R is also unknown. We here demonstrate that σ₂R may form heteroreceptor complexes with D₁ but not with D₂ receptors. Remarkably σ₁, σ₂, and D₁ receptors may form heterotrimers with particular signaling properties. Determination of cAMP levels, MAP kinase activation and label-free assays demonstrate allosteric interactions within the trimer. Importantly, the presence of σ₂R induces bias in signal transduction as σ₂R ligands increase cAMP signaling whereas reduce MAP kinase activation. These effects, which are opposite to those exerted via σ₁R, suggest that the D₁ receptor-mediated signaling depends on the degree of trimer formation and the differential balance of sigma receptor and heteroreceptor expression in acute versus chronic cocaine consumption. Although the physiological role is unknown, the heteroreceptor complex formed by σ₁, σ₂, and D₁ receptors arise as relevant to convey the cocaine actions on motor control and reward circuits and as a key factor in acquisition of the addictive habit.

Introduction

In advanced societies cocaine addiction is an important health and socio-economic problem. Cocaine use begins recreationally and the seeking behavior is based on a feeling of general well-being. Drug addiction is the result of plastic changes in areas of the brain that have dopamine as the main neurotransmitter, particularly in the ventral tegmental area(VTA) (see Lüscher, 2013 and references therein). The main consequence of cocaine consumption in the central nervous system (CNS) is an increase in interneuronal dopamine levels, which is not limited to VTA but extends to other structures, such as the basal ganglia (Wise, 1984; Bradberry, 2008). It was thought that the inhibition of dopamine transporters was at the root of all the effects caused by this drug of abuse. However, there is strong evidence showing that cocaine exerts effects by a direct interaction with sigma receptors. Two different sigma receptors have been identified that are functionally and structurally unrelated. Although endogenous ligands are not known and the physiological function of sigma receptors is unclear, these receptors share the ability to bind cocaine. On the one hand, sigma-1 receptor (σ₁R) is a chaperone that spans once the membrane bilayer and whose recently reported structure consists of a homotrimer (Schmidt et al., 2016). On the other hand, sigma-2 receptor (σ₂R) was identified as a member of the family of membrane-associated progesterone receptors; apart from σ₂R (PGRMC1), three other human members are identified: PGRMC2, neuferricin, and neudesin. They are haem proteins displaying a cytochrome b₅-fold domain. While σ₂R dimerization affects proliferation and chemoresistance in tumor/metastasis in vitro models and xenograft- based tumor/metastasis models, the mode of action in the periphery and the CNS are virtually unknown (Kabe et al., 2016; reviewed in Cahill, 2017). The interaction of σ₁R with dopamine receptors and the relevant role that σ₁R exerts on the modulation of dopaminergic signaling by cocaine has been reported. In contrast, no study has been undertaken to know whether the binding of cocaine to σ₂R results in dopaminergic regulation.

The role of σ₁R as relevant target of cocaine was suspected due to the moderate affinity of drug binding to the receptor (Matsumoto et al., 2003; Hayashi and Su, 2005). Therefore, it seems that the “physiologically” relevant concentrations of cocaine can both inhibit the uptake of dopamine and activate σ₁R. σ₁R-cocaine interaction intervenes in the triggering of locomotor and convulsive actions of the drug (Menkel et al., 1991; Matsumoto et al., 2001a,b, 2002; Barr et al., 2015). In addition, synthetic drugs that act as σ₁R agonists and antagonists, respectively, potentiate (Matsumoto et al., 2002, 2003) and reduce (Matsumoto et al., 2004) cocaine actions. More recent studies have identified in both heterologous expression systems and natural sources an interaction between σ₁R and dopamine receptors (Navarro et al., 2010; Moreno et al., 2014; Borroto-Escuela et al., 2017). Accordingly, it has been suggested that dual antagonism of σ₁R and inhibition of the dopamine DAT transporter can effectively block cocaine self-administration (Katz et al., 2016). Matsumoto et al. (2007) reported that treatment with synthetic drugs that act on σ₂R attenuates cocaine-derived behavior in mice. Although the selectivity of the compounds was poor, 1 year later, Mésangeau et al. (2008) designed an approach for converting selective σ₁R ligands into σ₂R selective ligands that, importantly, showed anti-cocaine activity. Furthermore, it has been observed that treatment with σ₂R antagonists counteract locomotor stimulation induced by cocaine in mice (Lever et al., 2014; Guo and Zhen, 2015).

An important physiological consequence of cocaine ingestion is an increase in motor activity, which is controlled by basal-ganglia brain circuits. Motor control is exerted by the direct and indirect pathways of the basal ganglia and associated nuclei. Of the five types of dopamine receptors, the D₁ (D₁R) is enriched in the direct pathway, while the D₂ (D₂R) is enriched in the indirect route. The balance of the dopaminergic input in the two circuits results in fine-tuning motor control. The locomotor hyperactivity resulting from cocaine use probably reflects a lack of balance in these two routes. The objective of this work was to investigate how the binding of cocaine to σ₂R affects dopaminergic signaling mediated by D₁R and/or D₂R. We first investigated whether σ₂R interacts with D₁R or with D₂R and, subsequently, we observed how cocaine could affect in a σ₁R-independent but σ₂R-dependent fashion the signal transduction triggered by agonist activation of D₁R but not of D₂R.

Results

σ₂R May Form Complexes with Dopamine D₁ But Not with Dopamine D₂Receptors

Two different sigma receptors have been described, the non-opioid receptor, σ₁R, and the PGRMC-1 protein, also known as σ₂. Despite the endogenous ligands are not known, the two sigma receptors may bind cocaine. While recent studies have demonstrated that σ₁R is involved in cocaine modulation of dopamine receptor function, a similar study on σ₂R-mediated modulation of dopaminergic signaling is lacking. We first evaluated in a heterologous expression system whether σ₂R may colocalize with dopamine receptors at the plasma membrane. Immunocytochemistry assays were undertaken in HEK-293T cells expressing σ₂R fused to Rluc and either dopamine D₁R fused to YFP or dopamine D₂R fused to YFP. The σ₂R expression was identified by a specific antibody against Rluc protein and a secondary Cy3 antibody, while dopamine receptor-YFP expression was identified by its own fluorescence. D₁R (green) was detectable at the plasma membrane level while σ₂R (red) was expressed both in intracellular structures and at the plasma membrane, where it colocalized (yellow) with D₁R (Figure 1A, left images). When a similar experiment was developed with D₂R, similar results were obtained indicating that D₂R and σ₂R colocalize at the cell surface (Figure 1A, right images). When the immunocytochemical assays were performed in cells pretreated with 30 μM cocaine for 30 min, the level of colocalization between σ₂R and D₁R or D₂R was similar, indicating that cocaine pretreatment did affect neither cell surface expression of σ₂R, D₁R or D₂R nor receptor colocalization. Next, we determined whether σ₂R may form heteromer complexes with dopamine D₁ or D₂ receptors. For this purpose, we took advantage of energy transfer assays and in situ proximity ligation assay (PLA), which allows the identification of close proximity between two proteins (<17 nm) (Borroto-Escuela et al., 2011; Trifilieff et al., 2011). For PLA, HEK-293T cells expressing σ₂R and either D₁R or D₂R were treated with specific primary antibodies against σ₂R and against each of the dopamine receptors. Interestingly, the red punctuated signal around Hoechst-stained nuclei was much higher for D₁R and σ₂R than for D₂R and σ₂R (82 versus 27% of labeled cells) (Figure 1B). Finally, we developed bioluminescence energy transfer assays in HEK-293T cells transfected with cDNAs for D₁R-Rluc or D₂R-Rluc and increasing amounts of cDNA for σ₂R-YFP. Interestingly, a saturable BRET curve was obtained (BRET_max 50 ± 3, BRET₅₀ 190 ± 40) (Figure 2A) indicating a specific interaction between D₁R-σ₂R; in contrast, a linear signal was obtained between D₂R-σ₂R (Figure 2B) suggesting a lack of interaction between them. When the same experiments were undertaken in cells treated with cocaine, similar results were obtained for the D₁R-Rluc/σ₂R-YFP donor/acceptor pair (BRET_max 82 ± 10, BRET₅₀680 ± 200), indicating that cocaine did not significantly affect the interaction (Figure 2B).

Dopamine D₁R, σ₁R, and σ₂R May Form Heterotrimeric Complexes

Dopamine D₁ and σ₁ receptors may form heteromeric complexes in HEK-293T cells (Navarro et al., 2010). To confirm whether in our experimental conditions D₁R-Rluc may act as a donor of σ₁R-YFP, BRET experiments undertaken in cotransfected HEK-293T cells provided a saturable curve thus indicating the interaction between σ₁R and D₁R (Figure 2C). We then hypothesized that σ₁R and σ₂R could be interacting together. Accordingly, BRET assays were performed in HEK-293T cells expressing a constant amount of σ₁R-Rluc and increasing amounts of σ₂R-YFP. The unspecific linear signal obtained (Figure 2D) suggested that no interaction was occurring between the two sigma receptors. We then performed assays to investigate whether σ₁R and σ₂R competed for the binding to D₁R. BRET experiments were then developed in HEK-293T cells expressing a constant amount of σ₁R-Rluc and D₁-YFP and increasing amounts of non-fused σ₂R. The results indicated that σ₂R was not able to compete with σ₁R for heteromer formation since the energy transfer between donor and acceptor was not altered (Figure 2E). When a similar experiment was performed expressing a constant amount of D₁-Rluc and of σ₂R-YFP and increasing amounts of non-fused σ₁R, the results indicated that low expression levels of σ₁R increased BRET signals; however, higher expression levels σ₁R were able to displace σ₂R out of the heteromer, as reflected by a significant decrease in BRET signal (Figure 2F). This result could reflect the formation of D₁R-σ₁R-σ₂R heterotrimer complexes, where the interaction of σ₁R to the σ₂R-D₁R complex could create a structural change in turn leading to increasing the energy transfer between Rluc and YFP. To confirm this possibility, sequential resonance energy transfer (SRET) assays, which permits detection of trimers (Carriba et al., 2008), were developed in HEK-293T cells expressing a constant amount of σ₁R-Rluc and of D₁R-YFP and increasing amounts of σ₂R-Cherry. The saturable SRET curve indicates that formation of σ₁R-D₁R-σ₂R heteromer complexes was occurring (Figures 2G,H). The negative control was performed by substituting σ₂R-RFP by the cannabinoid CB₁ receptor fused to RFP thus confirming the specificity of the triple σ₁R-D₁R-σ₂R interaction (Figure 2G).

σ₂R Activation Blocks Dopamine D₁R Signaling

Our next aim was to characterize the functionality of the σ₁R-D₁R-σ₂R heterotrimer structure in HEK-293T cells treated with cocaine. It should be noted that σ₁R (Navarro et al., 2010) and σ₂R (Johannessen et al., 2011) are endogenously expressed in HEK-293T cells; consequently, we used a siRNA approach to silence σ₁R or σ₂R expression thus impeding heterotrimer formation. When HEK-293T cells were transfected with D₁R and siRNA for σ₁R, SKF-81297-induced a significant increase in cAMP levels, that was inhibited by pretreatment with cocaine or with the σ₂R agonist, PB-28, indicating that cocaine decreases D₁R-mediated cAMP signaling function through its binding to σ₂R (Figure 3A). When HEK-293T cells were transfected with D₁R and siRNA for σ₂R, the results indicated that cocaine pretreatment potentiated agonist-induced cAMP levels, which was evidence of cocaine action upon binding to the σ₁R (Figure 3B). The next set of results is consistent with a reciprocal modulation of signaling mediated by cocaine binding to σ₂R and σ₁R; while cocaine via σ₁R positively modulates cAMP levels, it inhibits cAMP signaling via σ₂R. Accordingly, no effect of cocaine was observed in HEK-293T expressing D₁R and the two endogenous sigma receptors (Figure 3C). The lack of modulation exerted by cocaine upon simultaneous binding to both σ₁R and σ₂R likely reflects a balance which would, in a physiological set-up, depend on the relative expression of the two sigma receptors. In fact, when HEK-293T cells were transfected with D₁R and both siRNA for σ₁R and σ₂R, cocaine or the specific σ₂R agonist, PB-28, had no effect, indicating that cocaine modulation over D₁R depends on σ₁R and σ₂R expression (Figure 3D).

We next investigated whether cocaine binding to σ₂R receptors could still modulate D₂R-mediated signaling. HEK-293T cells transfected with cDNAs for D₂R and siRNA for σ₁R, responded to the selective-D₂R agonist, sumanirole. In these cells the G_i-mediated decrease of forskolin-induced cAMP accumulation due to G_i coupling was not affected by cocaine pretreatment (Figure 3E). These results agree with the lack of interaction between σ₂R and dopamine D₂R (see Figure 2B). As a control, we confirmed that when the σ₁R-cocaine modulation over D₂R was assayed, i.e., silencing σ₂R expression, cocaine was able to block the sumanirole-induced effect (Figure 3F). These results agree with those in Navarro et al. (2013) in the sense that they reflect the consequence of a physical interaction between σ₁R and D₂R receptors. In agreement with this hypothesis, HEK-293T cells expressing D₂R and endogenous sigma receptors behaved as cells in which the σ₂R was silenced (Figure 3G). As a further control, HEK-293T cells treated with siRNAs to silence both sigma receptors showed no modulation by cocaine over D₂R-mediated signaling (Figure 3H), thus reinforcing the idea that cocaine effect over D₂R depends on σ₁R expression.

σ₂R Activation Potentiates Dopamine D₁R MAP Kinase Phosphorylation

To further understand the cocaine effect over D₁R function, MAP kinase signaling was evaluated in HEK-293T cells transfected with cDNAs for D₁R and siRNA for either σ₁R or σ₂R. In cells expressing D₁R with silenced σ₁R, i.e., expressing D₁R and σ₂R (Figure 4A), cocaine pretreatment increased agonist (SFK-81297)-induced ERK1/2 phosphorylation, while in cells with silenced σ₂R, i.e., expressing D₁R and σ₁R (Figure 4B), cocaine decreased agonist-induced ERK1/2 phosphorylation. These results are evidence of potentiation by cocaine-σ₂R of MAP kinase signaling, and potentiation by cocaine-σ₁R of G_s-protein dependent signaling. In cells expressing D₁R and the two sigma receptors, no effect of cocaine pretreatment on pERK1/2 levels was observed, in agreement with the above-described balance resulting from reciprocal sigma-receptor-mediated cocaine effects (Figure 4C). As a further control, cocaine did not alter the SKF-81297-induced ERK1/2 phosphorylation in HEK-293T cells expressing D₁R and with silenced sigma receptors (Figure 4D). A similar experimental design was used to undertake dynamic mass redistribution (DMR) assays. DMR is a label-free technique useful to investigate the activation of G-protein coupled receptors (Grundmann and Kostenis, 2015; Medrano et al., 2017). On the one hand, in cells expressing D₁R and σ₂R, cocaine blocked SKF-81297-induced increase in the DMR signal in a similar way as the selective σ₂R ligand, PB-28, did (Figure 4E). On the other hand, the SKF-81297 effect was potentiated by cocaine pretreatment in cells expressing D₁R and σ₁R (Figure 4F). Once more, cocaine modulation on D₁R-agonist-induced effects was not found in cells expressing D₁R and both sigma receptors (Figure 4G). As DMR in cells expressing D₁R mainly reflects G_s-coupling (Kebig et al., 2009; Schröder et al., 2009; Hamamoto et al., 2015), these results are similar to those obtained in cAMP read-outs. Another control was performed to show that pretreatment with the σ₂R selective agonist, PB-28, did not result in any signal modulation in cells expressing D₁R but silenced σ₁R and σ₂R expression (Figure 4H).

σ₂R Activation Blocks Dopamine D₁R-Mediated Signaling in Primary Cultures of Striatal Neurons

A proximity ligation assay (PLA) was used to determine in primary cultures of striatal neurons whether D₁R-σ₂R complex expression was affected by cocaine pretreatment. Consequently, specific antibodies against D₁R and σ₂R were used in neurons treated or not with cocaine for 30 min (Figure 5A). 32% of cells showed punctuated staining (with 2.2 red spots/cell containing spots) surrounding Hoechst-stained nuclei (Figure 5B). These results indicate the occurrence of D₁-σ₂ heteroreceptor complexes in striatal primary cultures of neurons. A control done in the absence of primary antibodies led to 18% of labeled cells (with 1.2 red spots/cell containing spots). The percentage of positive cells after a 30-min treatment with cocaine was around 30 (with 2 red spots/cell containing spots) (Figure 5B). Thus, cocaine pretreatment did not significantly alter D₁R-σ₂R complex formation. When PLA was developed to detect D₂R and σ₂R complexes, the results (19% with 1.3 red spots/cell containing spots) were similar to those in the negative control (20% with 1.4 red spots/ cell containing spots), i.e., no evidence of heteroreceptor formation was obtained. Pretreatment with cocaine did not lead to the appearance of heteromer complexes formed by D₂R and σ₂R (Figure 5B). These results agree with the BRET assays that did not find sign of interaction between the D₂R-Rluc and σ₂R-YFP but between the D₁R-Rluc and σ₂R-YFP pair.

To demonstrate the effect of cocaine over D₁R-mediated signaling in a more physiological environment, we analyzed cAMP and MAP kinase signaling pathways in primary cultures of striatal neurons. As striatal neurons express the two sigma receptors, the siRNA approach was used to silence sigma receptor expression. On the one hand, in neurons transfected with siRNA for σ₁R, and consequently expressing D₁R and σ₂R, cocaine and PB-28 led to a decrease in agonist-induced cAMP levels and to an enhancement in MAP kinase signaling (Figures 5C,G). On the other hand, in neurons transfected with the siRNA for σ₂R, and consequently expressing D₁R and σ₁R, cocaine but not PB-28 induced an increase in the cAMP signal and a decrease in the ERK1/2 phosphorylation signal (Figures 5D,H). Most of these results agree with those obtained in the heterologous system. However, in striatal neurons expressing D₁R and both sigma receptors, cocaine treatment led to a net effect that showed predominance of σ₁R- versus σ₂R-mediated modulation (Figures 5E,I). These findings could be due to a higher expression of σ₁R-D₁R complexes versus σ₂R-D₁R but they may also result from the lower affinity of the cocaine/σ₂R binding (Lever et al., 2016). Finally, another control was performed to show that pretreatment with the σ₂R selective agonist, PB-28, or with cocaine, did not result in any signal modulation in cells expressing D₁R but silenced σ₁R and σ₂R expression (Figures 5F,J).

D₁R-Mediated Signaling Is Modulated by σ₁R in Acute and by σ₂R in Chronic Conditions

In situ PLAs were performed to identify D₁R-σ₁R and D₁R-σ₂R heteroreceptor complexes in striatal sections from Sprague–Dawley rats receiving cocaine under acute or chronic regimes (see section “Materials and Methods”) (Figure 6A). When striatal sections of vehicle-treated animals were analyzed, it was observed that 38.5% of cells showed D₁R-σ₁R complexes with 2.5 red spots/cell containing spots, while only 25% of cells showed D₁R-σ₂R complexes with 2.1 dots/cell (Figure 6B). When Sprague–Dawley rats were acutely treated with cocaine, it was observed that both D₁R-σ₁R and D₁R-σ₂R complex expression increased. However, the D₁R-σ₁R complexes doubled its expression while D₁R-σ₂R complex expression suffered a slight increase (respectively, 54% of cells showed red spots with 4.5 spots/cell and 33% with 2.3 spots/cell). Interestingly, in the case of rats chronically treated with cocaine, the D₁R-σ₁R heteromer complex expression was not affected (34% of cells showed red spots with 2.4 spots/cell containing spots) compared to control animals, while the D₁R-σ₂R heteromer expression significantly increased (35% of cells containing spots with 3.4 spots/cell containing spots) (Figure 6B). These results indicate that acute cocaine treatment strongly increases D₁R-σ₁R complexes formation in striatal rat sections but chronic cocaine treatment only drives D₁R-σ₂R complex expression. Then, we questioned if the cocaine-induced alterations in D₁R-σ₁R and D₁R-σ₂R complex expression had signaling consequences. To do so, we analyzed SKF-81297-induced cAMP production in primary cultures of striatal neurons pretreated with vehicle or cocaine for different times (from 0.5 h to 7 days). Interestingly, we observed that at short times SKF-81297-induced cAMP levels were further increased. In agreement with results in HEK-293 cells, cocaine binding to the σ₁R induced a positive modulation over dopamine D₁R-mediated signaling. When primary cultures of neurons were longer exposed to cocaine (1–7 days), SKF-81297-induced increase in cAMP levels was inhibited (Figure 6C). Taking into account the results in HEK-293 cells such effect seems associated to D₁R-σ₂R complex formation and to the ability of σ₂R to counteract the SKF-81297-induced increases of cAMP. To check whether these interpretations were correct, i.e., if sigma receptors were responsible of cocaine-induced modulations over D₁R-mediated signaling, primary striatal neurons were transfected with siRNA specific for σ₁R or σ₂R. On the one hand, cocaine pretreatment (0.5 h to 7 days) blocked SKF-81297-induced accumulation of cAMP levels in primary cultures of neurons transfected with siRNA for σ₁R, i.e., expressing D₁R and σ₂R (Figure 6D). On the other hand, 0.5 h and 2 h pretreatment of cocaine potentiated the SKF-81297-induced increases in cAMP levels in primary neurons transfected with siRNA for σ₂R, i.e., expressing D₁R and σ₁R. However, longer periods of cocaine exposure (1–7 days) produced no effect (Figure 6E). These results suggest that in acute cocaine treatment D₁R form heteromers mainly with σ₁R, prevailing the D₁R-σ₁R-mediated signaling. In contrast, in the chronic situation, the increase of σ₁R-D₁ heteromer complex expression observed in acute conditions disappear but the increase in the D₁R-σ₂R complex expression is maintained, being the σ₂R responsible of the cocaine modulation over D₁R, hence prevailing the D₁R-σ₂R-mediated signaling.

Discussion

Sigma receptors are relevant in cocaine addiction, because binding of cocaine to these receptors modulates dopaminergic transmission. Although cocaine can bind to both σ₁R and σ₂R, they are not closely related and no common structural properties have been identified. With respect to the modulation of receptor-mediated signaling, a relevant difference is revealed by the formation of heteroreceptor complexes. On the one hand, σ₁R interacts with D₁ and D₂ dopamine receptors (Navarro et al., 2010; Moreno et al., 2014). Interestingly, we here report that σ₂R may form heteromeric complexes with D₁R but not with D₂R.

Information on PGRMC1/σ₂R expression in brain is partial. Intlekofer and Petersen (2011)confirmed data by Krebs et al. (2000) showing enrichment of the receptor in nuclei of the hypothalamus that are important for female reproduction. Petersen et al. (2013) in 2013, reviewed neuroanatomical data on the expression PGRMC1 and related proteins in CNS neuroendocrine nuclei. To our knowledge, information on expression in other neural regions is either absent or preliminary. Interestingly, a recently developed fluorescent probe tested in rat brain indicates that the receptor is more present in neurons than in glial cells (Zeng et al., 2016). Despite good in vitro properties, some of the radiolabeled probes that were developed for in vivo σ2 receptor imaging have not reached the final objective (Abate et al., 2013; Selivanova et al., 2015). In contrast, recently reported ¹⁸F-labeled PET probes, with enhanced brain uptake and σ₂R selectivity (in mice), show promise for in vivo imaging of the receptor in the human brain (Wang et al., 2017). Surely these novel tools will be instrumental to achieve a more detailed mapping of the receptor in the CNS, specially in those areas in which dopamine receptors are expressed. There is, however, strong evidence of expression in the striatum; for instance, a recent report shows that receptor agonist regulate dopaminergic input into the striatum and the receptor is presynaptically expressed the nucleus accumbens (Klawonn et al., 2017). Furthermore, pioneering studies by Werling and colleagues showed σ₂R involvement in control of dopamine transporter activity in striatum (Derbez et al., 2002) and that the striatal receptor was a target of cocaine (Nuwayhid and Werling, 2006).

The results presented here and those already reported (Navarro et al., 2010, 2013) show that in equivalent experimental configurations, cocaine binding to σ₁R improves the accumulation of cAMP mediated by D₁R and inhibits MAP kinase signaling. Cocaine, via σ₂R, blocks D₁-mediated cAMP accumulation and enhances MAP kinase activation. Importantly, similar results were obtained in HEK-293T cells and primary neuronal cultures.

D₁R can form complexes and high- order heteromers by interacting simultaneously with the σ₁ and σ₂ receptors, σ₁R being able to displace σ₂R, but not vice versa. Navarro et al. (2010)reported an increase in the plasma membrane expression of σ₁R after acute exposure to cocaine. When increase in σ₁R levels in the plasma membrane occurs, σ₂R is displaced from the D₁R-σ₂R or D₁R-σ₂R-σ₁R heteroreceptor complexes. Such phenomenon results in increasing the amount of D₁-σ₁ heteroreceptors and D₁R signaling whereas, as reported by Navarro et al. (2013), reducing D₂R-mediated actions. However, in a longer exposure to cocaine, the signaling mediated by the dopamine D₁R fits more with that occurring via a D₁R-σ₂R functional unit. These data suggest that the initial cocaine-induced overexpression in the plasma membrane of the σ₁R is transient; once these levels decrease, due to internalization or other still unknown mechanisms, σ₂ is the predominant receptor forming heteromers with dopamine D₁R.

Motor control in the basal ganglia is achieved through a complex circuit composed of GABAergic neurons that contain mainly D₁R (direct pathway) and GABAergic neurons containing mostly D₂R (indirect pathway) (Grillner and Robertson, 2016). Fine motor control is achieved by a balance of dopaminergic signals, one via D₁ receptors, which are G_scoupled, and another via D₂ receptors, which are G_i coupled (Jenner, 1995; Gerfen, 2000). The deterioration of motor control by cocaine depends on the imbalance of the direct/indirect pathway, but the underlying mechanism remains unclear. Although the scenario is complex, cocaine is known to increase cAMP levels in cells expressing D₁R-σ₁R (Navarro et al., 2010). Therefore, cocaine seems to be increasing in the direct pathway the cAMP-dependent dopaminergic output, namely activation of protein kinase A and cAMP-regulated DARPP-32 phosphoprotein (Svenningsson et al., 2004). Through the same receptor (σ₁R), cocaine leads to a deterioration of the dopaminergic performance of the indirect route (Navarro et al., 2013). In addition to the imbalance resulting from these σ₁R-dependent effects, our results demonstrate that trimers of D₁, σ_1, and σ₂ receptors may be formed and that cocaine acting on these heteromers reduces the negative modulation exerted by the D₁R-σ₁R complexes.

The results here presented also show that the MAP kinase signaling pathway is particularly affected by the action of cocaine upon dopamine-sigma heteroreceptors. While in cells expressing the D₁R-σ₁R heteromer, cocaine decreased ERK1/2 phosphorylation, cocaine did the opposite in cells expressing the D₁R-σ₂R heteromer. It is known that ERKs are involved in the plastic changes induced by the consumption of drugs of abuse (Radwanska et al., 2005). In addition, the inhibition of ERK phosphorylation alters learned place-preference in a paradigm of drug-of-abuse consumption, whereas activation of ERK1/2 is necessary to establish the association between place preference and drug consumption (Valjent et al., 2006; Du et al., 2017). In this context, knocking down ERK1 has shown that enhanced ERK2 signaling and repeated exposure to the drug facilitate the plastic changes leading to drug addiction (Ferguson et al., 2006). It should be noted that the temporal pattern of MAP kinase activation in the mouse brain is differently induced by addictive or non-addictive drugs (Valjent et al., 2004). Interestingly, Zhang et al. (2017) have described that D₁ receptor antagonists alter in cocaine-treated mice the length of nucleus accumbens postsynaptic densities, i.e., cocaine-induced long-term plasticity; however, the mechanism underlying this phenomenon has not been described. According to the previous reports and to our results, it may be suggested that potentiation of MAP kinase pathway mediated by the D₁R-σ₂R heteromer may be the mechanism by which the σ₂R would induce long term neuronal plasticity. The predominant role of the σ₁R in acute cocaine use shifts to a more relevant role of σ₂R in the chronic condition leading to the establishment of addiction. In any case, the relative expression of the two receptors in a given neuron seems important in determining the fate of the cell when the drug of abuse is consumed.

In acute cocaine exposure, σ₁R modulation of D₁R-mediated signaling prevails, but in longer exposures, there is a shift to regulation by σ₂R. Recently, Singer et al. (2017) have determined that neuronal plasticity initiate 2 h after cocaine exposure. The mechanism of action described in this paper cannot explain some of the results reported by Matsumoto et al. (2007) and Lever et al. (2014) who report that σ₂R receptor antagonists block the effects of cocaine-induced hyperlocomotion. It should be, however, noted that a recent report show benefits of a σ₂R -selective agonist, siramesine, for decreasing cocaine effects via reduction of dopaminergic and glutamatergic input to the striatum (Klawonn et al., 2017). On the one hand, it is a reasonable assumption that σ₁R is more involved in the regulation of D₁R signaling at acute exposure. However, it remains to be determined whether some of the results reported on the impact of cocaine on locomotion are due to the use of non-selective ligands, that is, ligands that can bind to both sigma receptors and alter their function. Alternatively, it may happen that σ₂R is also affecting the direct route in acute conditions. What our results undoubtedly indicate is that σ₂R becomes the main player in conditions of chronic exposure to the drug. In summary, some of the addictive and motor actions of cocaine are the result of a balance between cocaine-σ₁R versus cocaine-σ₂R impact on activation of D₁R and D₂R (and D₁R-D₂R, see Perreault et al., 2016) in ad hoc CNS circuits.