M3 Receptor Pathway Stimulates Rapid Transcription of the CB1 Receptor Activation through Calcium Signalling and the CNR1 Gene Promoter

In this study, we have investigated a possible mechanism that enables CB1/M3 receptor cross-talk, using SH-SY5Y cells as a model system. Our results show that M3 receptor activation initiates signaling that rapidly upregulates the CNR1 gene, resulting in a greatly potentiated CB1 receptor response to agonists. Calcium homeostasis plays an essential intermediary role in this functional CB1/M3 receptor cross-talk. We show that M3 receptor-triggered calcium release greatly increases CB1 receptor expression via both transcriptional and translational activity, by enhancing CNR1 promoter activity. The co-expression of M3 and CB1 receptors in brain areas such as the nucleus accumbens and amygdala support the hypothesis that the altered synaptic plasticity observed after exposure to cannabinoids involves cross-talk with the M3 receptor subtype. In this context, M3 receptors and their interaction with the cannabinoid system at the transcriptional level represent a potential pharmacogenomic target not only for the develop of new drugs for addressing addiction and tolerance. but also to understand the mechanisms underpinning response stratification to cannabinoids.


Introduction
Understanding how the cascade of molecular events that follows receptor activation alters gene expression at the level of transcription, is the central pillar of the emerging field of pharmacogenomics. This is particularly important in understanding why a proportion of cannabinoid drug users suffer side effects such as addiction and withdrawal. For example, changes in intracellular Ca 2+ concentration are essential in synaptic plasticity, such as the learning and memory processes of long-term potentiation and depression [1]. Similarly, they may have a role in the development of dependence and withdrawal associated with the abuse of drugs like cannabinoids [2]. In this context, intracellular Ca 2+ homeostasis is crucial for the functions that neuronal systems such as the dopaminergic and glutamatergic systems display during the adaptive responses (tolerance, receptor sensitization, dependence, and withdrawal) involved in the on-set of addiction [3][4][5]. Importantly in this context, intracellular Ca 2+ is the major intracellular messenger that links synaptic activity in neurons to gene expression, and it is involved in the regulation of signal transduction networks of G protein-coupled receptors (GPCRs), including cannabinoid receptors [6][7][8][9][10]. Recently, a role for the cholinergic system in drug abuse disorder has been emerging, with several lines of evidence demonstrating a close functional coupling between acetylcholine (ACh) and opioid transmission [11]. Similarly, the cholinergic system has been found to play a

M 3 Receptor Activation Primes CB 1 Receptors in Human SH-SY5Y Neuroblastoma Cells
As detailed above, calcium imaging studies in human SH-SY5Y neuroblastoma cells show that CB 1 receptor activation (via Gα i/o ) stimulates Ca 2+ release from intracellular stores only after concomitant M 3 receptor activation (via Gα q/11 ) [10]. Here, we addressed the question of whether this priming by the M 3 receptor also affects the CB 1 receptor-driven electrophysiological response. To elucidate this, whole cell patch-clamp recordings were obtained of currents evoked by CB 1 receptor activation alone, or after pre-exposure to ACh, from SH-SY5Y cells grown on glass coverslips. The activation of CB 1 receptors by the localized application of the CB 1 receptor full agonist, WIN 55,212-2 (WIN, 100 nM) ejected by pressure from the tip of a nearby pipette, produced only a small inward current when applied to naïve cells (Figure 1Aa,Ba). Conversely, the pressure application of ACh to activate M 3 receptors (1 µM; Figure 1Ab) induced a substantial inward current. However, strikingly, if applied within 5 min after ACh application, WIN 55,212-2-induced inward currents were greatly enhanced (p < 0.001) (Figure 1Ab,Bb). This enhanced response was due to ACh-priming, since it was abolished either in presence of the selective M 3 receptor antagonist 4-DAMP (10 nM, Figure 1Ac,Bc) or the selective CB 1 receptor antagonist AM 251 (1 µM, Figure 1Ad,Bd). Thus, ACh priming, via muscarinic M 3 receptors stimulation, induced a greatly enhanced electrophysiological response to CB 1 receptor activation. ejected by pressure from the tip of a nearby pipette, produced only a small inward current when applied to naïve cells (Figure 1Aa, Ba). Conversely, the pressure application of ACh to activate M3 receptors (1 µM; Figure 1Ab) induced a substantial inward current. However, strikingly, if applied within 5 min after ACh application, WIN 55,212-2-induced inward currents were greatly enhanced (p < 0.001) (Figure 1Ab, Bb). This enhanced response was due to ACh-priming, since it was abolished either in presence of the selective M3 receptor antagonist 4-DAMP (10 nM, Figure 1Ac, Bc) or the selective CB1 receptor antagonist AM 251 (1 µM, Figure 1Ad, Bd). Thus, ACh priming, via muscarinic M3 receptors stimulation, induced a greatly enhanced electrophysiological response to CB1 receptor activation.

Enhanced CB1 Mediated Response in the Presence of ACh Is Due to the Release of Intracellular Ca 2+
To determine if the enhanced WIN 55,212-2 response due to ACh priming required the release of Ca 2+ from internal stores, we tested if it was abolished when recording in calcium-free extracellular buffer (0 Ca 2+ /EGTA buffer), which depletes intracellular Ca 2+ A small inward current after CB 1 receptor activation (WIN 55,212−2, 100 nM), was observed in SH−SY5Y cells (a), 5 min pre-application of acetylcholine (ACh, 1 µM) greatly enhances the inward current induced by WIN 55,212−2 stimulation (b). Superfusion with either M 3 (4−DAMP, 10 nM) or CB 1 (AM251, 1 µM) receptor antagonists (c,d respectively) were able to completely abolish the ACh priming effect on CB 1 receptors. During superfusion with Ca 2+ -free/EGTA (100 µM) buffer, the ACh−dependent priming effect on CB 1 receptors was again abolished (e). (B) Histogram representation of the M 3 receptor priming on the subsequent cellular response induced by CB 1 receptor stimulation in presence of 4−DAMP (c), AM251 (d) and EGTA (e). *** p < 0.001 vs. (a) WIN only, and ### p < 0.001 vs. (b) ACh-WIN. One-way analysis of variance followed by Multiple Comparison Test, with Bonferroni's correction. n values for each experiment are shown in brackets below each bar.

Enhanced CB 1 Mediated Response in the Presence of ACh Is Due to the Release of Intracellular Ca 2+
To determine if the enhanced WIN 55,212-2 response due to ACh priming required the release of Ca 2+ from internal stores, we tested if it was abolished when recording in calcium-free extracellular buffer (0 Ca 2+ /EGTA buffer), which depletes intracellular Ca 2+ stores. This was indeed the case, since the ACh-primed enhancement of the WIN 55,212-2 response was greatly reduced or abolished in the absence of Ca 2+ (Figure 1Ae,Be). This indicates that the enhanced electrophysiological response to CB 1 agonist after the ACh priming, was Ca 2+ -dependent.

ACh-Primed Enhancement of the CB 1 -Mediated Response Is Modulated at the Transcriptional and Translational Level
We next investigated the mechanism by which intracellular Ca 2+ enhances the electrophysiological response to WIN 55,212-2. Because it has been shown that CB 1 receptors undergo continuous and rapid replacement and degradation [22], we explored the possibility that intracellular Ca 2+ release accelerates de novo transcription of CB 1 receptors. We therefore tested this hypothesis by the inhibition of the cellular machinery (transcription and translation) involved in protein production, to see if it impaired the enhancing effects of ACh on CB 1 activity. Consistent with this hypothesis, the 30 min pre-incubation of SH-SY5Y cells with actinomycin D (Act D, a RNA polymerase inhibitor, 1 mM), blocked the AChpriming-induced response to WIN 55,212-2 (Figure 2Ac,Bc). Furthermore, when cells were pre-incubated (30 min) with cycloheximide (CHE, a protein synthesis inhibitor, 10 nM), a similar effect upon the ACh-mediated priming-induced response to a CB 1 receptor agonist was also observed (Figure 2Ad,Bd).
We next investigated the mechanism by which intracellular Ca 2+ enhances the electrophysiological response to WIN 55, Because it has been shown that CB1 receptors undergo continuous and rapid replacement and degradation [22], we explored the possibility that intracellular Ca 2+ release accelerates de novo transcription of CB1 receptors. We therefore tested this hypothesis by the inhibition of the cellular machinery (transcription and translation) involved in protein production, to see if it impaired the enhancing effects of ACh on CB1 activity. Consistent with this hypothesis, the 30 min pre-incubation of SH-SY5Y cells with actinomycin D (Act D, a RNA polymerase inhibitor, 1 mM), blocked the ACh-priming-induced response to WIN 55,212-2 (Figure 2Ac,Bc). Furthermore, when cells were pre-incubated (30 min) with cycloheximide (CHE, a protein synthesis inhibitor, 10 nM), a similar effect upon the ACh-mediated priming-induced response to a CB1 receptor agonist was also observed (Figure 2Ad,Bd).  These experiments suggest that the enhancement of the CB 1 -mediated response by ACh involved both transcriptional and translational mechanisms.
If translational mechanisms are involved in the enhancement, Western blot analysis of SH-SY5Y cells should reveal changes in expression levels of CB 1 receptor proteins. Indeed, 5 min of ACh stimulation (1 µM) did significantly increase expressed CB 1 receptor protein levels ( Figure 3a). To test the hypothesis that ACh application stimulated CB 1 expression at the transcriptional level, Western blot analysis was performed on SHSY-5Y cells after treatment with the RNA polymerase inhibitor actinomycin D during ACh stimulation ( Figure 3b). This prevented the increase in CB 1 protein expression, in keeping with the hypothesis that M 3 stimulated expression of CB 1 receptor acted at the transcriptional level. The increased expression was also blocked when cells were incubated (30 min) with the protein synthesis inhibitor cycloheximide (10 nM) (Figure 3c). These findings both support the contention that the M 3 -induced increase of CB 1 receptor expression is controlled at both transcriptional and translational levels. ulation ( Figure 3b). This prevented the increase in CB1 protein expression, in keepi the hypothesis that M3 stimulated expression of CB1 receptor acted at the transcr level. The increased expression was also blocked when cells were incubated (30 m the protein synthesis inhibitor cycloheximide (10 nM) (Figure 3c). These findin support the contention that the M3-induced increase of CB1 receptor expression trolled at both transcriptional and translational levels.

CB1 Receptor Expression Modulated by M3 Receptor Stimulation Is Controlled by th Promoter Locus
If the enhanced response to WIN 55,212-2 via M3 receptor pre-stimulation i an increase in CB1 receptor mRNA expression, this should occur within the tim observed in the electrophysiology experiments. To examine the kinetics of these with more precision, we performed quantitative PCR analysis on total RNA deriv treated and untreated SH-SY5Y cells with CNR1 gene mRNA specific primers ( Fig  Consistent with the proposed hypothesis, the expression of the CNR1 mRNA wa icantly increased by ACh stimulation with the maximum increase reached after 5 the cholinergic agonist administration (Figure 4b). Conversely, Act D, which inhib scription, abolished the ACh-induced up-regulation of CNR1 transcription (Fig  c(b)). On the other hand, pre-incubation with the protein synthesis inhibito heximide, did not cause any detectable effect upon ACh-induced up-regulation o transcription (Figure 4a, c(c)). Taken together, these results support the hypothes

CB 1 Receptor Expression Modulated by M 3 Receptor Stimulation Is Controlled by the CNR1 Promoter Locus
If the enhanced response to WIN 55,212-2 via M 3 receptor pre-stimulation is due to an increase in CB 1 receptor mRNA expression, this should occur within the timeframe observed in the electrophysiology experiments. To examine the kinetics of these changes with more precision, we performed quantitative PCR analysis on total RNA derived from treated and untreated SH-SY5Y cells with CNR1 gene mRNA specific primers (Figure 4a). Consistent with the proposed hypothesis, the expression of the CNR1 mRNA was significantly increased by ACh stimulation with the maximum increase reached after 5 min of the cholinergic agonist administration (Figure 4b). Conversely, Act D, which inhibits transcription, abolished the ACh-induced up-regulation of CNR1 transcription ( Figure 4A,Cb). On the other hand, pre-incubation with the protein synthesis inhibitor cycloheximide, did not cause any detectable effect upon ACh-induced up-regulation of CNR1 transcription ( Figure 4A,Cc). Taken together, these results support the hypothesis that a transcriptional regulatory relationship between ACh stimulation and CNR1 transcription exists. Moreover, since in SH-SY5Y cells, ACh stimulation causes intracellular Ca 2+ mobilization mainly through M 3 receptors [23], these results corroborate the electrophysiological observations above showing that M 3 receptor-mediated priming could modulate CB 1 receptor protein levels.
lization mainly through M3 receptors [23], these results corroborate the electrophysiological observations above showing that M3 receptor-mediated priming could modulate CB1 receptor protein levels.

Genomic Annotation of the CNR1 Promoter Indicates That It Is Active in Neuroblastoma Cells, Binds Core Transcriptional Proteins and Contains Putative cAMP Response Element-Binding Protein (CREB) Binding Sites
These data so far suggested that in SH-SY5Y neuroblastoma cells, ACh activation of M3 receptors transcriptionally regulates the CNR1 locus and its protein product, the CB1 receptor. The stimulation of muscarinic AChRs coupled to Gαq/11 (i.e., M3 receptors) has been shown to lead to the activation of CREB in neuronal cells, which could be how these effects occur in our experimental system [24,25]. Bioinformatics analysis was therefore One-way analysis of variance followed by Multiple Comparison test with Bonferroni's correction (n = 6).

Genomic Annotation of the CNR1 Promoter Indicates That It Is Active in Neuroblastoma Cells, Binds Core Transcriptional Proteins and Contains Putative cAMP Response Element-Binding Protein (CREB) Binding Sites
These data so far suggested that in SH-SY5Y neuroblastoma cells, ACh activation of M 3 receptors transcriptionally regulates the CNR1 locus and its protein product, the CB 1 receptor. The stimulation of muscarinic AChRs coupled to Gα q/11 (i.e., M 3 receptors) has been shown to lead to the activation of CREB in neuronal cells, which could be how these effects occur in our experimental system [24,25]. Bioinformatics analysis was therefore conducted for the CNR1 promoter region in these cells for characteristic features of a promoter element with CREB binding sites ( Figure 5). This analysis revealed that the CNR1 promoter is in an active/open conformation, consistent with its sensitivity to DNAseI digestion in a number of different cell types, including SK-N-SH cells from which SH-SY5Y cells derive ( Figure 5B). Binding studies also suggest that the 5 end of the CNR1 locus is a promoter element, as it binds RNA polymerase II (RNApol), TBP and TAF1, which are members of the transcriptional pre-initiation complex ( Figure 5C). Further, the promoter region shows marked signals for H3K4me3 histone modifications, a characteristic of active promoter regions ( Figure 5D). Taken together, the current genomic annotation of the 5 region of the CNR1 locus is consistent with previous work indicating the presence of a promoter at this locus [24], i.e., functioning as a CNR1 promoter. moter element with CREB binding sites ( Figure 5). This analysis revealed that the CNR1 promoter is in an active/open conformation, consistent with its sensitivity to DNAseI digestion in a number of different cell types, including SK-N-SH cells from which SH-SY5Y cells derive ( Figure 5B). Binding studies also suggest that the 5′ end of the CNR1 locus is a promoter element, as it binds RNA polymerase II (RNApol), TBP and TAF1, which are members of the transcriptional pre-initiation complex ( Figure 5C). Further, the promoter region shows marked signals for H3K4me3 histone modifications, a characteristic of active promoter regions ( Figure 5D). Taken together, the current genomic annotation of the 5′ region of the CNR1 locus is consistent with previous work indicating the presence of a promoter at this locus [24], i.e., functioning as a CNR1 promoter. Therefore, overall, these data support the hypothesis that the CNR1 promoter contains a response element that responds to signal transduction pathways triggered by activation of mAChRs, that coordinately regulate the expression of the CB1 receptor. Further examination of the CNR1 promoter was therefore conducted to establish if it could bind Therefore, overall, these data support the hypothesis that the CNR1 promoter contains a response element that responds to signal transduction pathways triggered by activation of mAChRs, that coordinately regulate the expression of the CB 1 receptor. Further examination of the CNR1 promoter was therefore conducted to establish if it could bind to, and regulate transcription in response to, ACh-induced CREB activation. Consistent with this, JASPAR analysis ( Figure 5E) revealed numerous highly conserved putative CREB binding sites. This analysis indicates, therefore, the presence of a highly conserved mechanism by which AChinduced transcription regulation can alter the levels of CB 1 receptor protein by activating transcription factors such as CREB to modulate the activity of key regulatory domains of the CNR1 locus. However, it is possible that the interaction of CREB with the CNR1 promoter is indirect rather than through the direct binding of CREB to the CNR1 promoter. Further analysis of this possible interaction is ongoing.

The Promoter Region of CNR1 Is Sensitive to M 3 Receptor Stimulation
A number of physiological and pathological conditions involving CB 1 receptor expression level and functionality have been related to the genetic regulation of the CNR1 promoter region [24]. The next experiment therefore tested if the previously observed M 3 receptor-stimulated increase in CNR1 transcriptional activity was via the CNR1 promoter. A luciferase reporter plasmid was produced, containing 1 kb of the CNR1 promoter region, based on observations published by Zhang et al. [23] and information produced from our datamining ( Figure 5). The reporter construct was then transfected into SH-SY5Y cells that were subsequently stimulated with ACh and the luciferase reporter gene expression was then quantified as a marker of CNR1 promoter activity ( Figure 6). Consistent with both our studies of Ca 2+ imaging [10] and electrophysiology (this study), a familiar pattern of responses was observed. Thus, the stimulation of the CB 1 receptors in naïve cells with the agonist WIN 55,212-2 (100 nM, 5 min) only marginally increased luciferase expression driven by the CNR1 promoter (Figure 6a). However, stimulation with ACh (1 µM, 15 min) produced a significant increase in luciferase expression, indicative of increased CNR1 promoter activity (Figure 6b). Strikingly, co-stimulation with ACh and WIN 55,212-2 stimulated CNR1 promoter activity to an even greater magnitude than ACh alone (Figure 6c). These data are therefore consistent with a model that the CNR1 promoter region is a component of the regulatory systems driving the interaction between the cannabinoid and ACh systems, by up-regulating the expression of the CB 1 encoding CNR1 gene. key regulatory domains of the CNR1 locus. However, it is possible that the interaction o CREB with the CNR1 promoter is indirect rather than through the direct binding of CREB to the CNR1 promoter. Further analysis of this possible interaction is ongoing.

The Promoter Region of CNR1 Is Sensitive to M3 Receptor Stimulation
A number of physiological and pathological conditions involving CB1 receptor ex pression level and functionality have been related to the genetic regulation of the CNR1 promoter region [24]. The next experiment therefore tested if the previously observed M receptor-stimulated increase in CNR1 transcriptional activity was via the CNR1 promoter A luciferase reporter plasmid was produced, containing 1 kb of the CNR1 promoter re gion, based on observations published by Zhang et al. [23] and information produced from our datamining ( Figure 5). The reporter construct was then transfected into SH-SY5Y cells that were subsequently stimulated with ACh and the luciferase reporter gene expres sion was then quantified as a marker of CNR1 promoter activity ( Figure 6). Consisten with both our studies of Ca 2+ imaging [10] and electrophysiology (this study), a familiar pattern of responses was observed. Thus, the stimulation of the CB1 receptors in naïve cells with the agonist WIN 55,212-2 (100 nM, 5 min) only marginally increased luciferase expression driven by the CNR1 promoter (Figure 6a). However, stimulation with ACh (1 µM, 15 min) produced a significant increase in luciferase expression, indicative of in creased CNR1 promoter activity (Figure 6b). Strikingly, co-stimulation with ACh and WIN 55,212-2 stimulated CNR1 promoter activity to an even greater magnitude than ACh alone ( Figure 6c). These data are therefore consistent with a model that the CNR1 pro moter region is a component of the regulatory systems driving the interaction between the cannabinoid and ACh systems, by up-regulating the expression of the CB1 encoding CNR1 gene.   , 1 µM, (b)), or with a combination of both (c). Following stimulation of transfected SH−SY5Y cells with WIN 55,212-2, no significant change was seen. However, after ACh treatment, a significant increase in luciferase expression was observed. A greater magnitude of CNR1 promoter activity was observed when the transfected cells were stimulated with both CB 1 (WIN) and M 3 (ACh) receptor agonists. * p < 0.05 vs. Ctrl, ** p < 0.01 vs. Ctrl. One-way analysis of variance followed by Multiple Comparison test, with Bonferroni's correction (n = 6).

CREB Activation in SH-SY5Y Cells Following Pre-Stimulation of M 3 Receptors
If the promoter region of CNR1 is indeed activated by M 3 receptor stimulation, it should also be possible to detect an increase in the activated form of CREB (phosphoCREB). Consistent with this, ACh administration produced a significant increase of phosphoCREB in SH-SY5Y cells (Figure 7a). This indicates that CREB is a downstream target of M 3 receptors stimulation.

CREB Activation in SH-SY5Y Cells Following Pre-Stimulation of M3 Receptors
If the promoter region of CNR1 is indeed activated by M3 receptor stimulation, it should also be possible to detect an increase in the activated form of CREB (phospho-CREB). Consistent with this, ACh administration produced a significant increase of phos-phoCREB in SH-SY5Y cells (Figure 7a). This indicates that CREB is a downstream target of M3 receptors stimulation.

Discussion
The disturbing symptoms of cannabis withdrawal have been shown to involve ACh signaling [26]. The current study aimed to determine whether there is cross-talk between the cellular signaling cascades governing cannabinoid receptor activation and ACh signaling, to explore a possible mechanism contributing to withdrawal symptoms. Collectively, our present and previous studies confirmed the presence of cross-talk between acetylcholine M3 receptors and cannabinoid CB1 receptors, at a cellular level through calcium signaling [6,7,10,27], and a direct interaction with the promoter region of the CNR1 gene. Overall, these findings support a model whereby M3 pre-activation evokes an increase in the cannabinoid-induced intracellular Ca 2+ concentration, triggering CREB phosphorylation. There is also a concomitant, or possibly a directly consequential, rapid increase in both CB1 promoter activation and expression of the CB1 gene.
Both cannabinoid CB1 and acetylcholine M3 receptors are highly expressed in SH-SY5Y cells used in the present study, which is why they serve as good model for areas within the central nervous system, such as nucleus accumbens and amygdala, known to be linked to addiction [28][29][30][31]. Here, these receptors also play an important role in processes such as sensory and motor processing, sleep, nociception, mood, stress response, attention, arousal, memory, motivation, and reward [32,33]. There is also evidence for cholinergic involvement in the initiation of addictive processes. However, the mechanism by which M3 receptors affect the addiction induced by substance of abuse remains unclear [34,35]. Using a combination of electrophysiology, pharmacology, molecular biology, and bioinformatics, the current study supports a novel hypothesis whereby M3 receptors, via Ca 2+ mediated regulation of the cross-talk, stimulate CB1 promoter activity, which may be involved in the complex events leading to the development of addictive behaviors or withdrawal syndromes, or both.

Discussion
The disturbing symptoms of cannabis withdrawal have been shown to involve ACh signaling [26]. The current study aimed to determine whether there is cross-talk between the cellular signaling cascades governing cannabinoid receptor activation and ACh signaling, to explore a possible mechanism contributing to withdrawal symptoms. Collectively, our present and previous studies confirmed the presence of cross-talk between acetylcholine M 3 receptors and cannabinoid CB 1 receptors, at a cellular level through calcium signaling [6,7,10,27], and a direct interaction with the promoter region of the CNR1 gene. Overall, these findings support a model whereby M 3 pre-activation evokes an increase in the cannabinoid-induced intracellular Ca 2+ concentration, triggering CREB phosphorylation. There is also a concomitant, or possibly a directly consequential, rapid increase in both CB 1 promoter activation and expression of the CB 1 gene.
Both cannabinoid CB 1 and acetylcholine M 3 receptors are highly expressed in SH-SY5Y cells used in the present study, which is why they serve as good model for areas within the central nervous system, such as nucleus accumbens and amygdala, known to be linked to addiction [28][29][30][31]. Here, these receptors also play an important role in processes such as sensory and motor processing, sleep, nociception, mood, stress response, attention, arousal, memory, motivation, and reward [32,33]. There is also evidence for cholinergic involvement in the initiation of addictive processes. However, the mechanism by which M 3 receptors affect the addiction induced by substance of abuse remains unclear [34,35]. Using a combination of electrophysiology, pharmacology, molecular biology, and bioinformatics, the current study supports a novel hypothesis whereby M 3 receptors, via Ca 2+ mediated regulation of the cross-talk, stimulate CB 1 promoter activity, which may be involved in the complex events leading to the development of addictive behaviors or withdrawal syndromes, or both.
To further elucidate the regulation of the cross-talk between receptors we first performed electrophysiological studies. Our results suggest that the ACh-primed increase in CB 1 receptor-mediated inward current was a consequence of a series of coordinated events, ultimately leading to the activation of transcriptional mechanisms at the CNR1 locus that encodes the CB 1 receptor. The abolition by 4-DAMP showed that M 3 pre-stimulation was essential in the initiation of CB 1 regulated depolarization. Our previous studies showed calcium mobilization acts as a major mediator of this cross-talk. Endoplasmic reticulum Ca 2+ homeostasis and the stable interaction between SERCA2b and GPCRs, a receptor class which includes CB 1 receptors, takes place during or soon after co-translation of the latter. This is a crucial event that enhances expression of GPCR receptors in a variety of cells [36]. Likewise, the stimulation of receptors that increase levels of the second messenger inositol trisphosphate (IP 3 ), such as M 3 receptors, cause the opening of the store-operated Ca 2+ release-activated Ca 2+ (CRAC) channels, activating components of the Ca 2+ signaling pathway (STIM proteins, Orai channels, calmodulin). This is known to trigger transient transcriptional responses [37][38][39]. We therefore performed Western blotting, qPCR experiments and bioinformatics analysis, which indicated the pre-stimulation of M 3 receptors in SH-SY5Y trigger the activation of a transcriptional mechanisms leading to new CB 1 receptor expression, thereby suggesting the involvement of a CB 1 promoter-centred autocrine loop.
Our results indicate ACh pre-stimulation up-regulates CB 1 receptors, by increasing CB 1 gene promoter activity, mRNA transcript and protein.
The role of transcription in the M 3 receptor mediated regulation of CB 1 receptors was particularly supported by the fact that CB 1 receptor up-regulation was blocked by incubation with Act D, an antagonist of RNA polymerase II activity. The active involvement of the CNR1 promoter region was further supported by analysis of the CNR1 promoter on-line data repositories (UCSC genome database), which revealed active promoter markers such as histone modification marks (H3K4me3), binding of the core transcriptional apparatus (RNA polymerase II) and DNAaseI hypersensitivity binding sites are all present within the CNR1 promoter region in cell lines from which neuroblastomas such as SH-SY5Y cells are derived [40][41][42][43].
The active involvement of the CNRI promoter in the interaction of ACh and CB 1 signaling with the expression of the CB 1 gene was supported by the use of luciferase reporter plasmid containing the proximal promoter region sequence of CNR1 [24,44]. Individually, M 3 receptor and CB 1 agonisms both increase the activity of the CNR1 proximal promoter. However, combined M 3 and CB 1 agonism made the promoter much more highly active compared to their individual effects. These observations support the model that at least part of the molecular mechanism by which M 3 stimulation enhances the electrophysiological response of CB 1 activation is by activation of the CNR1 promoter. Since it is well established that release of intracellular Ca 2+ can directly activate transcriptional factors, such as NFAT, NFkB, DREAM, CREB and c-fos [45][46][47][48][49], our previous observations suggest that intracellular Ca 2+ , mobilized due to M 3 receptor stimulation, is likely to be the main second messenger involved in the CNR1 gene regulation. Indeed, the UCSC genome browser provides evidence that several of these transcription factors (i.e., NFkB, c-fos) bind to the CNR1 promoter region in neuroblastoma cell lines [24]. We also identified a number of potential transcriptional factor binding sites in the CNR1 promoter region.
We focused our particular attention on CREB, since this transcription factor can be regulated by intracellular calcium concentration, which we have shown follows stimulation of M 3 receptors in SH-SY5Y cells [10]. Our Western blotting showed a small but significant increase of the phosphorylated active form of CREB, indicating it may be activated by the stimulation of M 3 receptors. This is consistent with studies in neurons showing intracellular Ca 2+ signals can lead to CREB phosphorylation via activation of the CaM-kinase family of proteins [50]. In this context, in SH-SY5Y cells, muscarinic M 3 receptor has been reported to regulate type 1 and type 8 adenylyl cyclase isoforms sensitive to Ca 2+ /calmodulin (CaM) or protein kinase C (PKC) [51]. In particular, it has been proposed that intracellular Ca 2+ oscillations may be decoded by AC8 into parallel cAMP oscillations that in turn, could differentially be decoded by downstream targets (i.e., CREB) and potentially initiate events such as gene expression and cell differentiation [52].
Further, nuclear Ca 2+ can be critical for activating regulatory pathways that control the timing and the nature of the CREB response. The hyper-excitable state of neurons in the nucleus accumbens after chronic exposure to drugs of abuse has been associated with changes in both Ca 2+ homeostasis and CREB activation [10,25,[53][54][55].
These observations, therefore, support a model whereby Ca 2+ stimulation increases CREB phosphorylation. However, further studies are needed to determine if the effect of CREB phosphorylation is relayed through direct CNR1 promoter binding or result from an indirect interaction.
This current study is not the first to show a rapid modulation of CB 1 receptor expression. A rapid increase of CB 1 receptor expression has also been observed in human [56], other mammalian cell [57], and animal models [58], supporting the structure-specific plasticity induced by cannabinoid drugs indicated here. Interestingly, the rapid activation of the CNR1 gene observed in our experiments follows a very similar timeline to early gene activation in other studies [59][60][61].
Altogether, our observations provide a mechanistic model to explain the functional cross-talk between M 3 and CB 1 receptors (Gα q/11 -Gα i/o cross-talk) that is responsible for the increased pharmacological response by ACh-priming ( Figure 8). In this model, at a basal level, the stimulation of CB 1 receptors produces a modest response, but the CNR1 promoter region is maintained in a state poised to respond if appropriately activated. The presence of an EZH2 binding site supports this situation. EZH2 is the functional enzymatic component of the Polycomb Repressive Complex (PRC2) that maintains gene promoter regions in a "poised" state, whereby release of PRC2 results in rapid transcriptional initiation [62,63]. Following M 3 receptor activation, the increase in intracellular Ca 2+ concentration primes the CNR1 promoter via the genetic machinery responsible for higher CB 1 receptor protein expression. The increased insertion of these CB 1 receptors into the cell membrane subsequently results in greater pharmacological responsiveness. Such increased responsiveness could underlie the symptoms of withdrawal.
In support of this model, the cross-talk mechanism we indicated by observations in the current proposal has been reported in the regulation of other genes. For example, a regulatory region that controls the BDNF gene in the hippocampus requires PKA and PKC pathways to interact appropriately [64]. Our model, therefore, could provide a significant step in the understanding of the cellular processes involving the neuronal responses to drugs of abuse such as cannabinoids. While most research to date has focused on synaptic plasticity to understand these processes, our current findings indicate a further mechanism. That is, changes in intrinsic neuronal excitability that are not synapse-specific, may play substantial roles in strongly modulating responses. Since a well-known feature of drug addiction is enhanced drug sensitivity, the observed Ca 2+ mediated "homeostatic plasticity", might be an important contributor to these cellular and behavioral adaptations involved in drugs of abuse phenomena [55].
Within the wider context of the burgeoning field of pharmacogenomics, our demonstration of the involvement of the CB 1 promoter region in the observed crosstalk between the M 3 and CB 1 provides an opportunity to further explore the specific protein-DNA interactions involved. More importantly still, the identification of the role of the CB 1 promoter in this mechanism will lead to an understanding of how genetic variation within the CB 1 gene promoter, or even environmentally driven epigenetic modifications such as DNA-methylation, could lead to greater or lesser susceptibility to cannabinoid addiction and withdrawal. moter regions in a "poised" state, whereby release of PRC2 results in rapid transcriptional initiation [62,63]. Following M3 receptor activation, the increase in intracellular Ca 2+ concentration primes the CNR1 promoter via the genetic machinery responsible for higher CB1 receptor protein expression. The increased insertion of these CB1 receptors into the cell membrane subsequently results in greater pharmacological responsiveness. Such increased responsiveness could underlie the symptoms of withdrawal.

Electrophysiology
The whole-cell configuration of the patch clamp technique [65] was used to record from cultured SH-SY5Y neuroblastoma cells at room temperature (20-23 • C). The pipettes had resistances of 3-7 MΩ and were filled with 140 mM CsCl, 1.1 mM EGTA, 2 mM MgCl 2 , 0.1 mM CaCl 2 , 2 mM ATP, 10 mM HEPES, and the pH of the internal patch pipette solution was adjusted to 7.2 by addition of Tris buffer. The patch pipettes were made from Pyrex borosilicate glass capillaries (Plowden and Thompson Ltd., Dial Glass Works, Stourbridge, UK) using a two-stage vertical microelectrode puller (Model 730, David Kopf Instruments, Tujunca, CA, USA). During experiments, the cells were maintained in a 35 mm diameter culture dish with extracellular recording buffer medium: 152 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 2 mM CaCl 2 , 10 mM HEPES and 10 mM Glucose. The pH was adjusted to 7.4 by addition of NaOH. The osmolarity of the external recording medium and internal patch pipette solutions were adjusted to 320 and 310 mOsm, respectively, with sucrose. Frozen external medium solutions were freshly defrosted and CaCl 2 was added as required. In some cases, CaCl 2 was substituted with BaCl 2 (2 mM) in the extracellular recording or internal patch pipette solution, as the charge carrier for calcium channel currents. In some experiments CaCl 2 was omitted and the concentration of the calcium chelator EGTA was raised to 20 mM to test for calcium-dependent conductance. The cells were viewed using a Diaphot-TMD (Nikon Europe B.V. Amstelveen, The Netherlands) inverted microscope at x200 magnification. A ground electrode in a KCl bath was connected to the adjacent recording bath by a KCl-agar bridge. Patched cells were voltage clamped with an Axoclamp-2A amplifier (Axon Instruments, Molecular Devices, Wokingham, UK) at a sampling rate of 15-25 kHz in the discontinuous single electrode voltage clamp mode, and initially held at −70 mV. Voltage steps were triggered, and the step duration controlled, by a D4030 pulse generator (Digitimiter, Welwyn Garden City, UK). Data were acquired with Scope 3.6.11 software (PowerLab, AD Instruments Ltd, Oxford, UK). Antagonists were bath-applied (M 3 antagonist, 4-DAMP; CB 1 antagonist/inverse agonist, AM251) while agonists were applied focally to the surface of the cells using a low-pressure (~7K Pa) ejection (ACh, WIN; PDES-2DX-LA, npi electronic GmbH, Tamm, Germany) from micropipettes with a diameter of 10 µm, with the pipette tip placed approximately 100 µm from the cell from which responses were being recorded.

Quantitative Reverse-Transcription PCR
A two-step method for the quantification of CNR1 transcript levels was used. SH-SY5Y neuroblastoma cells total RNA was extracted using ReliaPrep RNA cell miniprep system (Promega, Southampton, UK) in accordance with manufacturer's recommendations. A DNaseI incubation step was included in the protocol. Purified RNA concentrations were measured on a spectrophotometer (Nanodrop, ThermoFisher Scientific, Loughborough, UK) and aliquots of purified RNA were made to a final concentration of 60 ng/µL with RNase-free H 2 O. RNA was reverse transcribed to cDNA using the High-Capacity cDNA reverse transcription kit (Applied Biosciences, ThermoFisher Scientific, Loughborough, UK) using random primers, 600 ng of RNA and the following thermocycling conditions: 10 min at 25 • C then 120 min at 37 • C and finally 5 min at 85 • C. cDNA samples were diluted to a final concentration of 10 ng/µL with pure water after the completion of reverse transcription. Quantitative PCR was conducted with 2× SYBR green master mix (Roche, Welwyn Garden City, UK) in a Light Cycler 480 (Roche), with all data being analyzed on Light Cycler 480 software (Roche, Welwyn Garden City, UK). Master Mix reaction solutions were prepared containing 1× SYBR green, 500 nM final concentration CNR1 primers and TBP primers as a reference gene [66]. Master mixes were divided 18 µL/well in 384-well plates and cDNA samples were loaded at 2 µL/well in triplicate. Standard curves were produced for both primer pairs using a 5-fold serial dilution of one of the cDNA samples to provide a range from undiluted to 125 times diluted. cDNA samples for the production of a standard curve were loaded at 2 µL/well in duplicate. Standard curves from CNR1 primers and TBP primers were used to calculate primer efficiency values specific to both the primers being utilized and the sample being measured. Quantitative PCR thermocycling conditions were as follows: 5 min at 95 • C, 45 cycles of: 10 s at 95 • C, 10 s at 56 • C, 10 s at 72 • C, followed by a melting curve analysis using: 5 s at 95 • C, 1 min at 65 • C, 97 • C continuous.

Bioinformatics
In silico bioinformatic analysis of the CNR1 locus and CNR1 proximal promoter region were conducted using the ECR browser, rVista and the Transfac professional (v 10.2) transcription factor binding site algorithm hosted by the ECR browser [67]. CREB-binding sites within the human CNR1 promoter sequence were searched for using a matrix similarity value of 0.85 or greater. Further bioinformatics analysis of the CNR1 locus and CNR1 proximal promoter region was conducted using data hosted by the UCSC browser [68]. CNR1 locus annotation derived from RefSeq and Aceview. Polyadenylation sites by Gencode version 19 and consensus coding DNA sequences (CDS) came from the collaborative project between NCBI, EBI, HGNC, MGI and WTS. DNaseI hypersensitivity plots, ChIP-seq data on transcription factor binding, histone modifications and chromatin state segmentation data were collected from ENCODE [69] project data hosted by the UCSC browser.

Cloning
The CNR1 proximal promoter fragment was amplified from human placental DNA using the following primers: forward 5 CGCAGCCAGGTAGCGAACG 3 and reverse 5 TTTCGTTCTAGCGGACAAC 3 . Primers were designed to amplify the negative genomic DNA strand since CNR1 is transcribed from this strand. Following PCR amplification, the specific CNR1 proximal promoter amplicon was cloned into a linearized pGEM-T Easy vector using T4 DNA ligase (Promega, Southampton, UK). The CNR1 proximal promoter fragment was then cloned into pGL4.23 (Promega, Southampton, UK) by digesting both vector and insert plasmids with EcoRI and SacI to form pCNR1prom-Luc. High quality pCNR1prom reporter plasmid was purified from transformed E. coli using an endotoxin free maxi prep kit (Qiagen, Manchester, UK).

Luciferase Reporter Assays
Cells (≈1 × 10 4 ) were plated onto 24-well plates 24 h before transfection with Transfast (Promega, Southampton, UK). Transfections were optimized according to the manufacturer's instructions. In brief, reporter plasmid (the CNR1prom-Luc and pGL4.74 (renilla luciferase Promega), 0.5 µg per well) was mixed with serum-free cell medium (200 µL per well), and Transfast was added at a ratio of 2:1 (3 µL per well). Cells were then washed twice with PBS, and DNA medium mixture was added for 1 h, after which 1 mL of relevant cell medium containing 10% FBS was added. Following transfection and treatment of SH-SY5Y neuroblastoma cells, culture plates were placed on ice and washed with cold PBS twice followed by 20 min incubation at room temperature with 150 µL of 1× passive lysis buffer (Promega, Southampton, UK). Cell lysates were collected in pre-cooled micro tubes and centrifuged at 4000× g for 5 min at 4 • C. Cleared lysates were then loaded (20 µL/well) into a 96-well luminometer plate (Nunc, Roskilde, Denmark) in duplicate. Firefly luciferase and Renilla luciferase intensities were measured on GloMax ® 96 Microplate Luminometer (Promega, Southampton, UK) using the Dual-Luciferase reporter assay system in accordance with manufacturer's instructions. Reporter assay data were analyzed by calculating the ratio between firefly luciferase and Renilla luciferase intensities to control for differences in transfection efficiency. Subsequently, percentage change was calculated for treated samples as compared with control samples.

Statistical Analysis
The data and statistical analysis comply with the recommendations on experimental design and analysis in pharmacology [70]. Specifically, all averaged data sets are expressed as the mean ± S.E.M of n independent experiments. Statistical analysis was performed by either Student's unpaired t-test or one-way analysis of variance (ANOVA) followed by Bonferroni's Multiple Comparison post hoc test as appropriate (GraphPad Prism, San Diego, CA, USA).

Conclusions
This study revealed for the first time a functional interaction between CB 1 and M 3 receptors, that through the intracellular Ca 2+ mobilization pathway, is responsible for a rapid transcription of the CNR1 gene. Our findings clearly demonstrate how the physiological responses due to CB 1 receptor stimulation could be controlled and modulated by the cholinergic system. This newly described mechanism suggests the possibility of targeting the cannabinoid system from a different angle, with the cholinergic system playing a central role. In this contest, considering the cholinergic system the primary target, its activation could modulate cannabinoid system pharmacological responses, as well as limiting the manifestation of the unwanted side effects (withdrawal and addiction).