Programming of neural cells by (endo)cannabinoids: from physiological rules to emerging therapies

Abstract

Among the many signalling lipids, endocannabinoids are increasingly recognized for their important roles in neuronal and glial development. Recent experimental evidence suggests that, during neuronal differentiation, endocannabinoid signalling undergoes a fundamental switch from the prenatal determination of cell fate to the homeostatic regulation of synaptic neurotransmission and bioenergetics in the mature nervous system. These studies also offer novel insights into neuropsychiatric disease mechanisms and contribute to the public debate about the benefits and the risks of cannabis use during pregnancy and in adolescence.

Fig. 1. Molecular architecture of the endocannabinoid system during synaptogenesis and at mature synapses

Neuronal and glial components of developing and mature synapses are shown. It is important to note that the molecular architecture shown here is typical. There are likely to be differences in neurotransmitter system-specific and developmentally-regulated enzyme and/or receptor expression and function at different types of synapse and at different stages of development. Most prominently, monoacylglycerol lipase (MAGL) is excluded from motile growth cones until synaptogenesis commences^,. Strike-through symbols indicate catalytic enzyme activity throughout. (a) In developing synapse, AEA (green circles) and 2-AG (brown circles) orchestrate signaling by binding to their target receptors: CB₁R, CB₂R, GPR55, and TRPV1. Their availability is determined by biosynthesis enzymes (NAPE-PLD and DAGLα/β) and degrading enzymes (FAAH and MAGL). Note that information on alternative enzymes (e.g., ABHD6 and ABHD12) is at present not available. Unlike other endocannabinoid-binding receptors, CB₁Rs are preferentially recruited to and signal within cholesterol-enriched membrane microdomains termed lipid rafts. (b) At the mature synapse, the availability of AEA and 2-AG is determined by ABHD6 and ABHD12 in addition to FAAH and MAGL, and also by transmembrane (EMT) and intracellular (AIT) transport mechanisms (e.g., fatty acid binding proteins, heat shock protein 70, and FAAH-like AEA transporter), and storage organelles (adiposomes or lipid droplets). There is compelling evidence that key receptor and enzyme components of the endocannabinoid system partition distinctly, both intracellularly and amongst pre- and post-synaptic neurons, microglia and astrocytes. CB₂Rs are expressed mainly upon brain injury. Abbreviations: ABHD6/12, α/β-hydrolase domain-containing 6/12 hydrolases; AIT, AEA intracellular transporter; CB₁R/CB₂R, G protein-coupled type-1 and type-2 cannabinoid receptors; DAGLα/β, sn-1-diacylglycerol lipase α/β; EMT, putative endocannabinoid transmembrane transporter; ER, endoplasmic reticulum; FAAH, fatty-acid amide hydrolase; GPR55, G protein-coupled receptor 55; MAGL, monoacylglycerol lipase; NAPE-PLD, N-acylphosphatidylethanolamine-specific phospholipase D; TRPV1, transient receptor potential vanilloid 1 channel.

Fig. 2. Molecular architecture of endocannabinoid signaling during corticogenesis, with emphasis on neurogenesis and neuronal migration

(a) During mid- or late-gestation in rodents, 2-AG-rich cortical microdomains are thought to repulse post-mitotic neurons that express CB₁R⁺, including radially migrating pyramidal cells and tangentially migrating GABA interneurons, in the cerebral cortex. DAGL expression in the fetal ventricular proliferative zone (1) and in the cortical plate (2) can produce physiologically-relevant extracellular 2-AG concentrations (pink shading). This molecular arrangement, producing “corridors” sparse in 2-AG (white areas, 3), could explain some features of spatially segregated radial (pyramidal cell) and tangential (interneuron) neuronal migration. Radial glia (green, 4), acting as scaffolds for migrating neurons, can synthesize and subsequently degrade endocannabinoids, thus promoting the endocannabinoid-mediated radial detachment of neurons for final positioning (5). Abbreviations: CP, cortical plate; dms/sms, deep/superficial migratory streams; IZ, intermediate zone; MZ; marginal zone; SVZ, subventricular zone; VZ, ventricular zone. (b) At E14.5, CB₁R mRNA is predominantly expressed by neurons in the cortical plate (cp) and hippocampal primordium (hc). Arrows denote CB₁R mRNA expression in likely interneurons that leave the ganglionic eminence (ge) and migrate towards the cerebral cortex. Abbreviations: iz, intermediate zone; lv, lateral ventricle; mz; marginal zone; spt, septum. (c) In mid-gestational mouse brain, DAGLα is found expressed at high levels by pyramidal cells and targeted towards their axons. Immunohistochemistry was performed in GAD67-GFP knock-in mice, marking some GABAergic interneurons migrating in the superficial (sms) and deep migratory streams (dms). (d) CB₁R mRNA expression concentrates in the cortical plate and proliferative germinal layers (ne)^, in the human fetal brain (second trimester). Data in (b) were modified from refs., while images in (c) and (d) are courtesy of Dr. Erik Keimpema and Dr. Yasmin L. Hurd, respectively. Scale bar = 110 μm (b), 30 μm (c), 200 μm (d).

Fig. 3. Design logic of endocannabinoid signaling during neurite outgrowth and synaptogenesis

(a) Signal transduction mechanisms implicated in the CB₁R-mediated control of cortical neuron specification and morphological differentiation. Activation of tyrosine kinase receptors (particularly the fibroblast growth factor receptor (FGFR) and the high-affinity nerve growth factor (NGF) receptor, TrkA) and their activity-dependent phosphorylation are thought to induce 2-AG production via sequential activation of phospholipase Cγ (producing diacylglycerol; DAG) and sn-1-diacylglycerol lipase α (DAGLα). The dashed arrow in the plasma membrane indicates lateral 2-AG diffusion that can activate CB₁Rs in an autocrine fashion. Signaling via G proteins recruited to CB₁Rs upon agonist binding regulates neuronal morphology by e.g., the phosphorylation of C-Jun N-terminal kinases (e.g., JNK1), which trigger the rapid degradation of SCG10/stathmin-2 and alter to cytoskeletal instability. Alternatively, CB₁R activation can modulate the activity of Rho-family GTPases, particularly RhoA, to induce growth cone repulsion and collapse^,. Neurotrophin (blue circles) and 2-AG signaling can coincidently activate PI3K/Akt signaling. This, in turn, influences the activity of the transcriptional regulators Pax6 and CREB and their control of neural progenitor proliferation and fate decisions (for review see ref.). Cytoplasmic BRCA1 is referred to as one of the candidate E3 ubiquitin ligases controlling MAGL degradation^,. (b) Spatial segregation of molecular determinants of 2-AG signaling during the corticothalamic-thalamocortical axonal “handshake”. Left panel: corticofugal axons are CB₁R⁺ (red circles) whereas thalamocortical axons are CB₁R-but MAGL⁺ (green circles). Scale bar = 400 μm. Right panel: corticofugal axons harbor DAGLs (ref.) and can use paracrine 2-AG signaling for fasciculation (“1”). In turn, autocrine 2-AG signaling in corticofugal axons might be sufficient to promote their elongation (“2”). This molecular layout is compatible with MAGL⁺ thalamocortical axons limiting the spatial spread of 2-AG (“3”; dashed line indicates 2-AG inactivation), thus controlling the distribution of corticofugal fascicles and confining their growth trajectories to a subpallial corridor (“4”). Accordingly, pharmacological inhibition of MAGL activity during corticogenesis disrupts the formation of the corticofugal projection system. (c) The subcellular switch of DAGL and MAGL during neuronal polarization and synaptogenesis. In immature neurons, DAGLα is localized to the primary neurite (quiescent axon) including the growth cone for autocrine signaling. However, DAGLα is excluded from more proximal parts of the axon and redistributed to the somatodendritic axis of neurons once synapses formed. In contrast, MAGL becomes enriched in the presynapse, where it likely assumes a role as “stop” signal to limit 2-AG-mediated neurite elongation. The precisely timed molecular reconfiguration of 2-AG signaling supports a continuum of endocannabinoid actions during neuronal differentiation, leading up to the retrograde control of synaptic neurotransmission (inset). Here, DAGLα is selectively enriched in the perisynaptic annulus of dendritic spines (pink shading) apposing glutamatergic afferents^,. Abbreviations: ctx, cerebral cortex; cfa, corticofugal axon; f, fimbria; hc, hippocampus; lv, lateral ventricle; tca, thalamocortical axon; th, thalamus.

Fig. 4. Defective development of the corticofugal system following genetic manipulation of CB₁Rs

Conditional deletion of CB₁Rs from cortical pyramidal cells results in errant corticofugal axon fasciculation (cfa; arrows) in mice . Note that enlarged axon fascicles were also seen in extracortical areas, such as the striatum (cpu), where local CB₁R expression was unaffected. Abbreviation: G5, golga 5, a coiled-coil membrane protein that likely plays a role in vesicle tethering and docking; L1-NCAM, L1 neural cell adhesion molecule. Scale bars = 25 μm. Data were modified from ref..

Fig. 5. Dysregulation of cannabinoid receptor signaling in glioma cells

(a) In neural progenitors, agonist binding to cannabinoid receptors couples to the PI3K/Akt/mTORC1 pathway via G_i proteins. Trans-activated (by phosphorylation) receptor tyrosine kinases (Trks) might produce signal amplification by also using mTORC1 as a molecular effector. Akt can elicit cell growth and survival effects (thus being mitogenic) either by inhibiting glycogen synthase kinase 3β (GSK-3β) and activating β-catenin or by activating mTORC1, which leads to p27 inhibition and Pax6 phosphorylation and expressional up-regulation. (b) In glioma cells, cannabinoids trigger ER stress by engaging (at least) two mechanisms: cannabinoid receptors stimulate de novo synthesis of ceramide in the ER via G_i-dependent and perhaps also G_i-independent mechanisms^,; and TRPV1 receptors on the ER, which mediate Ca²⁺ release from this organelle to the cytoplasm and, conceivably, deplete ER Ca²⁺ stores. Ceramide accumulation and Ca²⁺ depletion in the ER converge at the phosphorylation (that is, inhibition) of eIF2α and the induction of ATF-4, which, in turn, triggers cell death by two signaling cascades: up-regulation of TRIB-3 expression, leading to the inhibition of the AKT-mTORC1 axis; and up-regulation of ATF-3 expression.