• No se han encontrado resultados

Introducció

In document CIÈNCIES SOCIALS I LA SEUA DIDÀCTICA (página 153-156)

Modulation of cAMP production was the rst second messenger system convincingly demonstrated to be mediated by cannabinoids via the CB1 receptor. Both exo- and endocannabinoids inhibit adenylyl cyclase via G proteins. Numerous studies have shown the involvement of cannabinoids in the modulation of cAMP levels in cells of various types in culture, as well as in homogenates of brain regions.

In the neuroblastoma (N18TG2) or neuroblastoma X glioma (NG108-15) cell lines, cannabinoid-induced decreases in cAMP formation were shown not to be due to interaction with prostanoid, opioid, muscarinic, or adrenergic systems (reviewed by Howlett, 1995). Studies in NG108-15 cells showed that cannabinoid-induced inhibition of cyclic-AMP formation is rapid and reversible, occurs at low concentrations of the cannabinoids, and follows a structure–activity relationship, and its stereoselec-tivity is similar to that observed for behavioral measures (with the exception of the anticonvulsant effects of the cannabinoids). Furthermore, the effect on cAMP was not blocked by antagonists of other classical neurotransmitters (binding is not displaced by classical neurotransmitters) and is mediated via coupling to the Gi protein, because pertussis toxin attenuates activity of the commonly used synthetic cannabinoid agonist, CP55940. These studies, along with supporting work evaluating the stereoselectivity and the Hill coefcient for binding of the synthetic bicyclic cannabinoid, CP55940 (Devane et al., 1988), established the CB1 cannabinoid receptor linkage through a Gi protein to the modulation of cyclic AMP.

The potency of numerous cannabinoids to inhibit cAMP formation in the neuroblastoma cells was found to correlate to the antinociceptive effects of the drugs in vivo (Howlett et al., 1988). The cannabinoid-induced antinociception was proposed to be due to the modulation of adenylyl cyclase.

The potency of various cannabinoids to displace CP55940 binding and to inhibit adenylyl cyclase has been shown to be similar in rank order to the production of not only antinociception but also hypothermia, spontaneous activity, and catalepsy by the cannabinoids (Compton et al., 1993; Little et al., 1988). The aminoalkylindole analogs—WIN55212-2 is the prototype—have also been shown to inhibit adenylyl cyclase activity in rat brain membranes and compete for cannabinoid-binding sites (Pacheco et al., 1991). These compounds are interesting in that they were initially developed as nonsteroidal anti-inammatory agents, and some analogs have demonstrable in vitro cannabinoid antagonist effects.

Soon after its discovery, AEA was shown to inhibit cAMP accumulation in CB1-transfected CHO cells and N18TG2 cells (Felder et al., 1993; Vogel et al., 1993). Similarly, in the initial discovery of 2-AG, Mechoulam and coauthors reported its ability to inhibit adenylyl cyclase activity in CB1- and CB2-transfected COS cells (Mechoulam et al., 1995). Both AEA and 2-AG are full (or nearly full) agonists in the inhibition of adenylyl cyclase via the CB1 receptor (Stella et al., 1997).

Early studies of forskolin-stimulated cAMP in mouse brain synaptosomes showed biphasic responses to !9-THC and !8-THC (Little and Martin, 1991). Some cannabinoids did not alter cAMP levels. These early studies suggesting biphasic effects of cannabinoids have been conrmed, both in brain and in CB1-receptor-transfected cells (Glass and Felder, 1997; Bonhaus et al., 1998).

Furthermore, a mutational study points to selective G-protein activation as a possible mechanism for the biphasic effects of cannabinoids, because a point mutation produced a constitutively active CB1 receptor that coupled to Gs instead of Gi (Abadji et al., 1999). AEA is a partial agonist in activation of this stimulatory pathway (Bonhaus et al., 1998). 2-arachidonylglycerol has not been examined for stimulatory effects.

The determinants of which cAMP signaling pathway is activated by CB1 may include agonist-specic G-protein interactions, as well as the isoform of adenylyl cyclase present in the cell type examined. There is mounting evidence that CB1 and CB2 receptors interact with different sets of G proteins (Mackie et al., 1995; Glass and Northup, 1999; McAllister et al., 1999). In addition, it has recently been shown in studies from several laboratories that different ligands promote interactions with different G proteins (Selley et al., 1996; Bonhaus et al., 1998; Grifn et al., 1998; Glass and Northup, 1999; Kearn et al., 1999; Tao et al., 1999; Mukhopadhyay et al., 2000). Furthermore, one ligand can activate several G proteins (Glass and Northup, 1999; Mukhopadhyay et al., 2000; Prather et al., 2000). In addition, coexpression of adenylyl cyclase isoforms 1,3,5,6, or 8 with CB1 or CB2 receptors resulted in inhibition of cAMP, whereas isoforms 2,4 or 7 resulted in increases in cAMP formation (Rhee et al., 1998).

REGULATION OF ION CHANNELS

The recent discovery that endocannabinoids serve as retrograde messengers in multiple brain regions has allowed for the integration of past data concerning endocannabinoid inuence on ion currents into a specic unit in a complete functional neurotransmitter circuit (Ohno-Shosaku et al., 2001;

Kreitzer and Regehr, 2001; Wilson and Nicoll, 2001). Many important initial studies laid the foundation for these pathways, as detailed in the following paragraphs.

Inhibition of Voltage-Dependent Calcium Channels

The rst experiments utilizing an endocannabinoid to affect ion channel activity demonstrated the inhibition of N- and Q-type voltage-dependent calcium channels (VDCC) in cell lines transfected with the cDNA for CB1 receptors or containing the native protein (Mackie et al., 1993; Mackie et al., 1995). The inhibition produced by AEA was blocked by pertussis toxin and not because of the activation of the cAMP pathway(Mackie et al., 1993). The effect observed was likely the result of a direct interaction of G subunits with calcium channels (Ikeda, 1996).

Numerous studies using synthetic cannabinoids have shown that activation of CB1 receptors on axonal terminals leads to inhibition of calcium channels and a subsequent decrease in the release of neurotransmitter (reviewed in Pertwee, 1997). As CB1 receptors have been localized to terminals containing acetylcholine, GABA, glutamate, or norepinephrine, the physiological outcome pro-duced by receptor activation clearly depends upon which cell populations are being activated. The presynaptic inhibition of calcium channels on axon terminals that synapse on pyramidal neurons of the hippocampus correlates with the ability of endocannabinoids to inhibit long-term potentiation (Terranova et al., 1995; Stella et al., 1997; Misner and Sullivan, 1999). Cannabinoids may alter learning and memory through this mechanism. This point has been reiterated in recent investigations focusing on retrograde signaling in the hippocampus and cerebellum and is discussed in the following paragraphs.

The distribution of presynaptic cannabinoid receptors, the postsynaptic localization of both precursors to endocannabinoid synthesis, and the enzyme involved in their metabolism suggested that endocannabinoids might act as retrograde signaling molecules; these molecules would be released from postsynaptic cells to inhibit presynaptic transmitter release via CB1 receptor activation (Di Marzo et al., 1994; Stella et al., 1997; Egertova et al., 1998; Elphick and Egertova, 2001). An intact system demonstrating the interplay between all these components was discovered by groups that were not focused on endocannabinoids but on the phenomenon termed depolarization-induced suppression of inhibition or excitation (DSI or DSE).

Retrograde signaling by endocannabinoids results in DSI and DSE and was rst demonstrated in the hippocampus and cerebellum (Ohno-Shosaku et al., 2001; Kreitzer and Regehr, 2001; Wilson and Nicoll, 2001). The investigations showed that activation of postsynaptic neurons resulted in the release of endocannabinoids from these neurons. The endogenous ligands then acted as retro-grade signaling molecules to inhibit presynaptic calcium inux in axonal terminals and subsequently reduced the release of neurotransmitter.

Using CB1 knockout animals, a study in hippocampal slices denitively showed that activation of the CB1 receptor was responsible for the downstream inhibition of presynaptic VDCC in cells in which DSI was produced (Wilson et al., 2001). The investigators also showed that a specic subtype of GABAergic interneuron is targeted by endocannabinoids. In this case, activation of CB1 receptors led to the specic inhibition of N- but not P/Q- type calcium channels, and this resulted in decreased release of neurotransmitter.

In contrast, in GABAergic interneurons which synapse with Purkinje cells in the cerebellum, an interaction between CB1 receptors and VDCC does not lead to decreased transmitter release (Kreitzer et al., 2002). The authors demonstrated that activation of a K+ current by CB1 receptors

in GABAergic interneurons led to the inhibition of ring in these cells. Therefore, in presynaptic terminals, the interaction between CB1 receptors and ion channels in the endocannabinoid retrograde signaling pathway can change substantially depending on the cell type involved.

Another calcium channel subtype inhibited by endocannabinoids via CB1 receptors includes the native L-type Ca2+ channels in cat cerebral vascular smooth muscle cells (VSMC) (Gebremedhin et al., 1999). The inhibition of the L-type calcium current by AEA in VSMC may relate to the ability of AEA to relax preconstricted cerebral vessels in similar tissue.

Synthetic cannabinoid agonists were also shown to produce inhibition of L-type calcium channels of bipolar cells in retinal slices from larval tiger salamander (Straiker et al., 1999). The presence of 2-AG, PEA, and oleylethanolamide, but not AEA, was reported in the tissue used in this study but the direct application of these ligands was not tested in regard to L-type calcium channel activity.

Interestingly, a follow-up study by Straiker and Sullivan found that cannabinoids enhanced L-type calcium channel activity in rods but inhibited channel activity in cones (Straiker and Sullivan, 2003).

In both rods and cones, a potassium current was also inhibited (Straiker and Sullivan, 2003). The enhancement of L-type calcium currents in rods was due to the modulation of the cAMP pathway by CB1, however, the signaling pathways leading to the other effects were not examined. Future studies focusing on the endocannabinoid signaling pathway in this tissue will help explain the functional relevance of differential ionic effects produced by CB1 receptors in the eye.

Modulation of Potassium Currents

As discussed earlier, endocannabinoids can enhance K+ currents in cerebellar slices leading to decreased neurotransmitter release in the retrograde signaling pathway (Kreitzer et al., 2002). The

rst indication that AEA could modulate potassium channels was shown in AtT20 cells transfected with the CB1 receptor (Mackie et al., 1995). In this investigation, application of AEA caused an enhancement of a G-protein-coupled inwardly rectifying potassium current (Kir /GIRK). Cannab-inoid ligands have also been shown to enhance A-type potassium currents in hippocampal cells, an effect that was indirectly mediated by inhibition of cAMP accumulation (Deadwyler et al., 1995). In the cerebellar slices in which endocannabinoid retrograde signaling was investigated, it was determined that CB1 receptors most likely modulated a Kir-type potassium current vs. KA; the exact type of current affected still needs to be elucidated (Kreitzer et al., 2002). In the nucleus accumbens, CB1 receptors located on glutametergic afferents can also potentially modulate KA and Kir currents leading to decreased neurotransmitter release, but endocannabinoids have not been studied in this pathway (Robbe et al., 2001).

AEA can enhance potassium currents in Xenopus oocytes transiently expressing CB1 receptors and GIRK channels (Henry and Chavkin, 1995; McAllister et al., 1999). The efcient coupling between the receptors and channels has allowed for a convenient system to study structure–activity relationships of the CB1 receptor (Jin et al., 1999; McAllister et al., 2002).

In the brain, almost all investigations point to a presynaptic locus for CB1 receptor modulation of ionic channels (Elphick and Egertova, 2001). However, one study carried out in hippocampal slices suggests that a postsynaptic interaction between CB1 receptors and specic potassium chan-nels can occur. In CA1 hippocampal neurons, methanandamide acting through CB1 receptors decreased postsynaptic K+ M-current (IM) (Schweitzer, 2000). The author did not suggest a direct interaction of CB1 receptor-activated G proteins with the channels but rather hypothesized that CB1 -mediated stimulation of intracellular calcium stores, leading to increase in intracellular calcium, could be one of the mechanisms behind the inhibition of IM. Regardless, this study suggests an interesting potential addition to the endocannabinoid retrograde signaling pathway. If a postsynaptic CA1 pyramidal cell is activated, releasing endocannabinoids, it may potentially have two ways to modulate its activity: the indirect decreased release of neurotransmitter from presynaptic terminals via CB1 and a direct change in excitability through activation of postsynaptic CB1 receptors. If IM

is inhibited in a cell, it makes it harder for the cell to repolarize after an action potential (Marrion, 1997). In general, CB1 receptors have a presynaptic localization in the hippocampus and cerebellum (Elphick and Egertova, 2001; Tsou et al., 1998). Future studies will be needed to determine if an intact in vivo system with these characteristics exists.

MODULATIONOF INTRACELLULAR CALCIUM

Early studies of endocannabinoid inuence over intracellular calcium levels reported non-receptor-mediated effects in brain tissue and cells transfected with CB1 receptors (Felder et al., 1992; Felder et al., 1995; Mombouli et al., 1999). However, more recent studies in cell lines and tissue with native CB1 receptors have demonstrated CB1-receptor-mediated activation of intracellular calcium stores. The key determinant between the production of receptor and nonreceptor effects appears to be the cellular background used.

In rat brain slices and in cerebeller granule neurons in culture, endocannabinoids were shown to modulate calcium ux through NMDA channels (Hampson et al,1998; Netzeband et al., 1999).

In rat brain slices, AEA inhibited calcium inux brought about by addition of NMDA (Hampson et al., 1998). In this model, addition of NMDA alone caused an increase in intracellular calcium.

The proposed mechanism responsible for the endocannabinoid effect was an opposing inhibition of calcium entry into the cells brought about by CB1-receptor inhibition of voltage-dependent P/Q-type calcium channels.

In primary cerebellar cultures and acutely isolated cerebellar granule neurons, methanandamide has been demonstrated to enhance NMDA-evoked calcium release (Netzeband et al., 1999). This effect was due to activation of the CB1 receptor. In a study using inhibitors of VDCC, a variety of compounds known to modulate intracellular calcium release, and PKC and PKA inhibitors, it was determined that the enhancement of NMDA-evoked calcium release was due to CB1-receptor activation of the phospholipase C pathway; this led to the downstream release of calcium from intracellular stores. In this study, it was also noted that blockade of the phospholipase C pathway unmasked a CB1-mediated inhibition of the NMDA-evoked calcium release (Netzeband et al., 1999). This is consistent with the study reported by Hampson et al. (1998) in rat brain slices, referred to earlier.

To relate the ndings reported by both groups, Netzeband et al. (1999) hypothesized the existence of an activity-dependent pathway that is differentially modulated by endocannabinoids.

They proposed that, under conditions of strong cellular stimulation, an inhibitory effect on voltage-gated calcium channels by endocannabinoids would be apparent, whereas under periods of lower stimulation, the PLC-mediated pathway would dominate.

After the discovery of the endocannabinoid 2-AG (Sugiura et al., 1995; Mechoulam et al., 1995), Sugiura and colleagues demonstrated a rapid and transient calcium increase that was produced by 2-AG but not other structural analogs such as 2-palmityol-glycerol, 2-oleoyl-glycerol, and 2-linoleoyl-glycerol (Sugiura et al., 1995; Mechoulam et al., 1995). AEA was a partial agonist in this system, as were the potent synthetic analogs HU210 and CP55940. The cell line used was a neuroblastoma–glioma hybrid (NG108-15) that contained endogenous CB1 receptors. The endocannabinoid-mediated increase in transient Ca2+ release was both CB1 and Gi/Go mediatedand could be produced by the synthetic agonist WIN55212-2 but not the inactive isomer WIN55212-3.

The investigators suggested CB1 receptors could activate PLC through G') subunits as these subunits have previously been demonstrated to stimulate PLC' in NG108-15 cells (Jin et al., 1994).

It would be interesting to determine if 2-AG produces similar calcium transients in neuronal populations in the CNS where CB1 receptors exist. Perhaps, future studies will be able to relate this effect in NG108-15 cells to endogenous signaling in native neuronal tissue.

Recently, the CB1 receptor was shown to couple the activated FGF receptor to an axonal growth response via increased calcium inux into cerebellar granule neurons (Williams et al., 2003). CB1 receptor antagonists inhibited the neurite outgrowth response stimulated by N-cadherin and FGF2.

Synthetic AEA and 2-AG analogs, arachidonyl-2-chloroethylamide and 2-arachidonylglycerol ether, were shown to enhance neurite outgrowth. Furthermore, activation of the FGF receptor resulted in activation of phospholipase C with the subsequent hydrolysis of DAG (Williams et al., 2003). The initial hydrolysis of DAG at the sn-1 position will generate 2-AG (Stella et al., 1997), the proposed endogenous neuromodulator of neurite outgrowth (Williams et al., 2003). In this case, activation of the CB1 receptor required calcium inux into the neurons; the agonist-induced neurite outgrowth response was inhibited by N- and L-type calcium channel antagonists (Williams et al., 2003).

NITRIC OXIDE

Anandamide and 2-arachidonylglycerol can directly stimulate nitric oxide release via CB1 (Stefano et al., 1996; Stefano et al., 2000). This direct stimulation of constitutive nitric oxide synthase has been shown in immunocytes, microglia, and monocytes, as well as in endothelial cells from blood vessels (Stefano et al., 2000). As constitutive nitric oxide synthase isoforms are Ca2+-calmodulin dependent, an intermediate step in this pathway may be increased intracellular calcium, as described earlier.

CB1 receptors also appear to mediate inhibition of inducible nitric oxide synthase. Studies in microglia showed a dose-dependent inhibition of lipopolysaccharide-induced and/or )-interferon-induced nitric oxide release that was antagonized by the CB1 receptor antagonist SR141716A (Waksman et al., 1999). Recently, AEA has been demonstrated to inhibit )-interferon- and HIV-1 Tat protein–induced release of nitric oxide in rat C6 glioma cells (Esposito et al., 2002). AEA also inhibited lipopolysaccharide-stimulated release of nitric oxide from glia in primary culture; how-ever, in this case, the CB2 receptor also appears to be involved (Molina-Holgado et al., 2002). The pathway activated in glia involved the subsequent release of interleukin-1, indicating a role of AEA in the modulation of brain injury (Esposito et al., 2002; Molina-Holgado et al., 2003).

REGULATION OF INTRACELLULAR KINASES

Early studies using synthetic cannabinoid agonists showed a CB1-receptor-dependent stimulation of the two MAP kinases (MAPK), p42 and p44 kDa (Bouaboula, Poinot-Chazel, et al., 1995). This effect on MAPK was independent of the cAMP pathway and led to downstream changes in the immediate-early gene Krox-24 (Bouaboula, Bourrie, et al., 1995). Futhermore, phosphatidylinositol 3-kinase (PI3K) was an intermediate in the pathway leading to CB1 receptor upregulation of MAPK (Bouaboula et al., 1997). More recent studies have expanded this pathway in the presence of endocannabinoids. AEA activation of p42 and p44 kDa MAPK, now more commonly termed extracellular signal-regulated kinase (ERK) 1 and 2 or ERK, led to the induction of chemotaxis and chemokinesis in HEK cells transfected with CB1 receptors (Song and Zhong, 2000).

AEA can reduce progenitor cell differentiation by inhibiting the Rap1/B-Raf-ERK pathway through CB1 receptors in developing neurons (Rueda et al., 2002). The inhibition observed in vitro was also correlated with in vivo activity of the compounds in the hippocampus. Interestingly, although intraperitoneal administration of methanandamide did not decrease the total number of dividing cells in the subgranular zone of the dentate gyrus of the adult rat, it did inhibit the ability of neuronal progenitors to reach a mature phenotype in this region (Rueda et al., 2002).

Endocannabinoids, working through the CB1 receptor, also activate ERK in hippocampal slices through a pathway that involves G"-cAMP but not G')-PI3K (Derkinderen, Ledent, et al., 2001;

Derkinderen et al., 2003). The same group also showed that AEA and 2-AG increase tyrosine phosphorylation of focal adhesion kinase (FAK) in rat hippocampal slices (Derkinderen, Toutant, et al., 2001). This effect was mediated downstream of CB1 receptor inhibition of cAMP (Derkinderen et al., 1996). The tyrosine kinase Fyn, which associated with FAK during endocannabinoid treatments, appeared to be critical to the effects produced, because in Fyn knockout mice, stimulation of FAK by 2-AG was lost (Derkinderen, Toutant, et al., 2001). Another intriguing nding by Derkinderen

(2003) (Derkinderen et al., 2003) was that in the Fyn knockout mice, 2-AG stimulation of ERK was lost. This suggests that Fyn is a common link between ERK and FAK activation in hippocampal slices. Pathways utilizing ERK and FAK have been associated with changes in long-term potentiation (LTP) in the hippocampus (Grant et al., 1995; Kojima et al., 1997). These proteins may be an important part of endocannabinoid signaling, associated with effects produced downstream of acute actions such as DSI.

ANTITUMOR ACTIVITY

Endocannabinoids have been shown to reduce nerve growth factor–induced and basal cell prolif-eration of human breast cancer cells (De Petrocellis et al., 1998; Melck et al., 1999; Melck et al., 2000). Early studies using EFM-19 and MCF-7 cells conrmed a CB1-receptor-dependent effect that resulted in the downregulation of both prolactin receptors (PRLr) and a downstream breast-cancer-susceptibility gene, brca1 (De Petrocellis et al., 1998). Prolactin has been shown to act as

Endocannabinoids have been shown to reduce nerve growth factor–induced and basal cell prolif-eration of human breast cancer cells (De Petrocellis et al., 1998; Melck et al., 1999; Melck et al., 2000). Early studies using EFM-19 and MCF-7 cells conrmed a CB1-receptor-dependent effect that resulted in the downregulation of both prolactin receptors (PRLr) and a downstream breast-cancer-susceptibility gene, brca1 (De Petrocellis et al., 1998). Prolactin has been shown to act as

In document CIÈNCIES SOCIALS I LA SEUA DIDÀCTICA (página 153-156)