Marijuana and other derivatives of the plant Cannabis
spp. have been used for thousands of years for their therapeutic and mood-altering/recreational properties. Cannabis
-derived psychoactive compounds, mainly ∆9-tetrahydrocannabinol (∆9-THC) and cannabidiol (CBD), can pass the placental barrier and be transferred to the embryo [1
]. It is now recognized that Cannabis
consumption during pregnancy can exert adverse consequences on the progeny, including anxiety, cognitive and attention deficits, as well as depression [2
In the central nervous system (CNS), ∆9-THC, CBD and various other synthetic cannabinoids bind the G protein-coupled receptors of the cannabinoid family, CB1R and CB2R [3
]. The cognate endogenous ligands, known as endocannabinoids (eCBs), are small signaling lipids synthesized “on demand” from membrane lipids. The most abundant eCB in the CNS is 2-arachidonoylglycerol (2-AG), derived from the cleavage of diacylglycerol by diacyl-glycerol lipase (DAGL) and degraded by the action of monoacylglycerol lipase (MAGL) [4
]. eCBs are mainly recognized by the CB1 and CB2 receptors, although they can also bind to other receptor types (TRP, PPARs, etc.). During CNS development, CB1R expression levels gradually increase and are localized on developing axonal projections, while CB2Rs are rather expressed by uncommitted precursor cells [7
]. The localization of the eCB-synthesizing enzyme DAGL usually overlaps with that of CB1Rs in preterminal axons and growth cones, while MAGL is actively excluded from growth cones until synaptogenesis is concluded [8
]. Upon release, eCBs have been shown to act in both autocrine and paracrine fashion and can cross cell membranes. In contrast to the primarily autocrine mechanism of action in the immature CNS, in the adult synapses it has been shown that eCBs act as retrograde signals (released post-synaptically, and acts on pre-synaptic termini to modulate neurotransmitter release) through a paracrine mechanism.
Following binding of eCBs, phyto-CBs and synthetic CBs to CB1R, a variety of transduction pathways are activated in a time- and region-dependent way, depending on the cell type and receptor repertoires. In particular, eCBs have been proposed to activate Rho-class of guanosine tri-phosphate phosphatases (Rho-GTPases) and to control the neuronal actomyosin cytoskeleton [9
]. It has been demonstrated that [10
] direct interaction of CB1 with the WAVE1/RAC1 complex, which acts on actin-binding proteins, results in actin remodeling and growth cone dynamics. Activated CB1R can also modulate microtubule dynamics at the growth cone, via the c-Jun NH(2)-terminal kinases (JNKs) pathway [11
]. By modulating the actomyosin network and microtubule dynamics at the growth cone [9
], in addition to axon growth and guidance, CB1 receptors have been shown to participate in the control of synaptogenesis and synaptic plasticity in various experimental models [14
Data from the medical literature support the role of CB1R in brain wiring during early development. Fetuses exposed to cannabis during embryonic life have increased risk of cognitive deficit [17
], attention deficit [19
] and anxiety and depression [20
], as well as defects in neuronal migration and axonal pathfinding in several brain districts [2
]. Most experimental studies have focused mainly on the cerebral cortex, the hippocampus and the visual system in mice, where CB1R regulates axon guidance and map formation [22
]. Importantly, exposure of pregnant mice to ∆9-THC results in life-long circuit modification and altered wiring in the offspring [29
]. In zebrafish embryos, interfering with CB1R signaling has been shown to result in abnormal axonal growth and fasciculation in the anterior and posterior commissures of the forebrain. Moreover, exuberant axons of reticulospinal neurons in the hindbrain were found to cross the midline or to deviate from their trajectory and turn backward [30
]. Similarly, the selective pharmacological inhibition of CB1R and CB2R in early embryos resulted in altered development of the locomotor system at later stages [31
], and treatment with ∆9-THC or CBD resulted in altered motor neurons morphology, synaptic activity at the neuromuscular junction and locomotor responses to sound [32
]. Reducing the level of 2-AG via knockdown of Daglα
resulted in abnormal behaviors, characterized by stereotyped movement and altered motion perception [33
Few reports have linked the eCB/CB1R signaling functions to the maturation and physiology of the hypothalamic neuroendocrine systems controlling reproduction and appetite [34
]. The gonadotropin releasing hormone (GnRH) neurons, which are central for the control of sexual maturation, fertility and reproduction, have been extensively studied. In this context, eCBs have been shown to modulate the input and the firing activity of GnRH neurons [36
], and CB1R activation inhibits presynaptic GABA release [37
]. Accordingly, the literature reports a link between perturbations of CB1R (either due to consumption of phyto-cannabinoids or exposure to synthetic cannabinoids) and altered functionality of the neuroendocrine gonadotropic system, puberty and fertility deficits [35
Another important prominent effect of the eCB system is the regulation of appetite and energy balance. The hypothalamic network of peptidergic neurons controlling food intake/energy metabolism, enhancing feeding behavior, includes mainly neurons expressing pro-opiomelanocortin (POMC), neuropeptide Y (NPY) and agouti-related peptide (AgRP) [39
]. This system is highly conserved among vertebrates. Pharmacological blockade of CB1R suppresses hunger and induces hypophagia, and the CB1R-antagonist Rimonabant in the past has been proposed for the treatment of obesity [40
]. Importantly, eCB/CB1R signaling has been shown to have an impact on AgRP/NPY neurons [43
] and indeed its perturbation has been associated with altered appetite control, energy storage/consumption and metabolic conditions [44
While most previous works have focused on postnatal and adult hypothalamic functions, little is known on the role of CB1R and its misactivation for the correct development of hypothalamic neuroendocrine neurons and their connectivity needed to attain control of sexual maturation and/or reproduction and appetite/energy storage. This question is highly relevant, to fully assess the impact of exposure to these (and other molecules) during embryonic and early infancy on human health, and the possible link with adult pathologies, such as reduced fertility and altered food intake. Furthermore, these studies aimed to elucidate the role of eCB ligands and the mechanisms underlying axonal miswiring upon altered CB1R activation [7
We have investigated the possible role of CB1R altered activity on the extension, fasciculation and pathfinding of GnRH3 and AgRP axons during forebrain development in zebrafish. Using transgenic lines in which GnRH3 and AgRP neurons are fluorescently labeled, we show that pharmacological and genetic manipulation of CB1R resulted in abnormal guidance, fasciculation and routing of GnRH3 and AgRP1 axons during embryonic development, indicating that CB1R is a possible regulator of the development of reproduction- and food intake-related neurons.
In the present paper we show that pharmacological manipulation and gene knockdown of CB1R in early zebrafish embryos cause profound alterations in the ability of GnRH3 and AgRP1 axons to organize in discrete and well-fasciculated commissures, accompanied by misrouting and an altered neuron number. We also provide protein coexpression analyses that confirm, in early embryos, the localization of CB1 receptors in commissural structures and longitudinal tracts associated with GnRH3 fibers. Thus, we propose that CB1R is involved in the fasciculation and guidance of neuroendocrine GnRH3 and AgRP1 axons during forebrain development, implying high sensitivity of these neuroendocrine systems to neurodevelopmental exposure to exogenous cannabinoids.
During brain development, axonal projections elongate, fasciculate and defasciculate in distinctive pathways, often crossing the midline and reaching the appropriate targets guided by the orchestrated interactions between axon tracts and the environment at distinct domains, as well as homo- and hetero-philic interactions among axonal fibers. Several signaling and adhesion molecules involved in axon guidance and fasciculation have been identified [64
]. Recent reports suggest that CB1R signaling is likely to modulate one or more of these signaling cascades during neural circuit formation, however most studies have focused on a few brain regions and neuronal types, in particular corticofugal and retinofugal neurons [22
]. Here we show a novel role of eCBs signaling via CB1R on the development of two well-characterized neuroendocrine systems in zebrafish, the GnRH3 hypophysiotropic system and the AgRP1 system, involved in the central control of reproduction and food intake, respectively.
In order to gain relevant information on the developmental effects of the synthetic CB1R ligands employed in this study (WIN55, Rimonabant and AM251) we tested the possible toxic effects of their chronic exposure on zebrafish embryos from 0 to 72 or 96 hpf using a wide range of concentrations. Embryo survival was affected only by the highest agonist concentration (1 μM WIN55), and therefore this concentration was excluded from subsequent experiments. The selected ligands concentrations did not interfere with the overall embryo development and the hatching rate. These results partly agree with a previous study [70
], in which AM251 treatment (10 and 20 nM) on zebrafish embryos for 72–96 hpf did not lead to morphological changes but decreased the hatching rate.
We analyzed the impact of chronic CB1R ligand exposure on the developing GnRH3 and AgRP1 systems, taking advantage of the fluorescently labelled neurons present in the gnrh3::EGFP
and the agrp1::mCherry
]. Pharmacological interference with CB1R in gnrh3::EGFP
embryos clearly showed abnormal fasciculation of GnRH3 axons in the anterior commissure, with some ectopic axons detaching from the commissure itself. Defective phenotypes resulted from exposures to all three tested drugs (CB1R agonist, inverse agonist and antagonist) even at the lowest concentrations, indicating that interfering with the endogenous eCB system, either through over-activation or inhibition, equally disrupts the modulatory control exerted by CB1R on axonal navigation and fasciculation. We obtained similar results by knocking down the CB1R with a morpholino strategy, confirming previous observations made on the anterior commissure [30
]. Furthermore, by immunocytochemistry we show co-expression of CB1R and GnRH3::EGFP in several fiber systems, although evidence of co-localization of the two labeling in exactly the same axons was found only in the anterior commissure and in longitudinal fiber tracts. Together these data suggest that at least part of the axons crossing the anterior commissure are affected by CB1R ligand exposure, including CB1R-expressing GnRH3 fibers.
In addition to the defects observed in fiber tracts, we found a significant decrease in the number of GnRH3 cell bodies (terminal nerve-associated GnRH3 neurons) following both antagonist (AM251) exposure and CB1R knockdown. These results are consistent with the previously reported role of CB1R as a regulator of progenitor niches [7
]. Alternatively, CB1R could be involved in the timing of GnRH3 neuronal migration. A role of CB1R in neuronal migration has been previously reported during radial migration of immature pyramidal neurons in the cortex [23
], and eCB signaling has been shown to promote migration of newborn neurons along the rostral migratory stream in the postnatal mouse brain [14
]. Further experiments are needed to define a possible role of CB1R in GnRH3 neuronal migration.
Pharmacological interference of CB1R in agrp1::mCherry zebrafish embryos resulted in anterior commissure defects similar to those observed in gnrh3::EGFP embryos, suggesting that a common mechanism affected both GnRH3- and AgRP1-containing fibers. We found an increase in the number of AgRP1+ cell bodies in the hypothalamus in AM251-treated embryos, however this result could be affected by the methodology we chose for counting cells. It is possible that the apparent augmented number of AgRP1 cells depends on a decrease in fiber number, which would be in line with an axonal defect.
In our experiments, we observed similar phenotypes when applying CB1R agonists or antagonist/inverse agonists. This is consistent with the fact that different studies reported opposite effects of CB1R activation on axonal growth. Indeed, in some cases CB1R-activation has been reported to induce collapse of the growth cone and axon retraction, while its inhibition resulted in axon elongation [9
]; other studies reported the opposite effects [23
]. It should be kept in mind that while eCBs can reach very high local concentrations due to spatially and temporally coordinated activity of their synthetic and degrading enzymes, exogenous CB1R agonists and antagonists engage their targets indiscriminately, possibly impacting the spatial- and chrono-dynamics of eCB signaling in developing axons in a similar fashion.
Multiple mechanisms downstream of CB1R activation in developing neurons have been suggested. It has been reported that CB1R acts via the WAVE complex and Rac1-GTPases, and via Rho-GTPase and the Rho-dependent ROCK activation [9
]; these pathways appear to mainly control the dynamics of the actin cytoskeleton. Notably, these intracellular signaling pathways are essential for neuronal “motility” during embryo development, including migration, neuritogenesis, spinogenesis and synaptogenesis [74
Another reported mechanism involves the stathmin family of proteins and the fine modulation of microtubule dynamics at the growth cone. Stathmin-2, also known as SCG10 (superior cervical ganglia neural-specific 10) is a neuron-specific protein highly expressed in the developing forebrain, able to stabilize the plus ends both at steady state and early during polymerization by increasing the rate and extent of growth and facilitating microtubule extension into filopodia [13
]. Overexpression of stathmin-2 strongly enhances neurite outgrowth, while its phosphorylation by the JNKs negatively regulates its microtubule destabilizing activity, suggesting that the JNK-stathmin-2 pathway may link extracellular signals to the rearrangement of the neuronal cytoskeleton [11
]. By applying a bioinformatic analysis based on conserved coexpression and transcriptomic datasets from the developing hypothalamus, we report the identification of stathmin-2 as most likely to be involved in the effect of CB1R on axon elongation/fasciculation/guidance. Recently, stathmin-2 has been shown to be a specific molecular target for a CB1R-mediated effect of ∆9-THC in mouse and possibly in human fetal nervous systems [29
]. By restoring stathmin-2 expression in the context of ∆9-THC exposure and consequent CB1R activation, the authors were able to acutely correct the altered cytoskeleton dynamics [29
]. Highly relevant for neuroendocrine neurons, stathmin expression has been shown to modulate the migratory properties of mammalian GnRH neurons in vitro [60
]. Thus, several evidence including ours indicate that stathmin-2/SCG10 participates in a regulatory cascade that is influenced by the activity and/or misactivity of the CB1R and controls cytoskeleton dynamics during neurite elongation and possibly neuronal migration.
4. Materials and Methods
4.1. Zebrafish Strains and Treatments
All procedures using zebrafish (Danio rerio
) were authorized by the Ethical Committee of the University of Torino and the Italian Ministry of Health (approval code: N. 425/2016-PR; 27 April 2016). The wild-type fish strain TL-AB was used. The fish strain Tg(gnrh3::EGFP
] was obtained from Prof. Y. Zohar (Univ. Maryland Biotechnology Institute, Baltimore, MD, USA). The fish strain named, for simplicity, agrp1::mCherry
was obtained by breeding the TgBAC(agrp:GAL4-VP16)
with the UAS::nfsB-mCherry
strains, followed by selection of the reporter fishes by quick fluorescent examination at low magnification [47
]. Adult fish were routinely maintained under a 14 h light and 10 h dark photoperiod at approximately 28 °C, bred and genotyped according to standard procedures. Allelic transmission followed the expected mendelian ratios. Eggs were generated by natural mating, and following fertilization were collected, treated and maintained under a 12 h light and 12 h dark photoperiod, incubated at 28 °C.
The Tg(gnrh3::EGFP) embryos were grown in the presence of 0.003% 1-phenyl-2-thiourea (PTU) to prevent the formation of melanin pigment, which could interfere with the visualization of fluorescence in neurons and fibers; the agrp1::mCherry embryos were grown in fish water without PTU, because the pigmentation in this case did not affect the analysis under microscope.
4.2. Evaluation of Survival Rate and Hatching Rate
The endpoints used to assess developmental toxicity comprised embryos survival (percentage of live embryos/total) and hatching rate (percentage of hatching embryos/living embryos), observed every 24 h during the whole exposure period (72 or 96 hpf). Dead embryos were removed without solution renewal.
4.3. Morpholino Injections
To down-modulate CB1R we utilized a conventional antisense morpholino oligonucleotide (MO)-mediated strategy (GeneTools, LLC, Philomat, OR, USA). The z-cb1r
MO was designed as previously published [30
]. The MOs used for injection were
|z-cb1r MO:||5′ CTAGAGGAAACCTGTCGGAGGAAAT 3′|
|Mismatch_PPFIA scrambled MO (control):||5′ TCGTGGCCATCAACTCGAACA 3′.|
Zygotes were collected at the 1-cell stage and injected under stereological examination with 50 µM, 200 µM or 400 µM of z-cb1r MO (or the scrambled MO as control), in the presence of Phenol Red for subsequent selection of the injected embryos. Uninjected embryos were also analyzed.
4.4. Western Blot
Zebrafish embryonic heads at 72 hpf were lysed manually in modified RIPA buffer composed as follows: 150 mM NaCl, 50 mM Tris pH 7.5 with 1% Triton X-100, 0.5% Na deoxycholate, 0.1% SDS, 2 mM EDTA, 1 mM sodium orthovanadate, 1 mM sodium fluoride and protease inhibitors. Homogenates were centrifuged at 3000 rpm for 5 min at 4 °C, extracts were quantified with BCA method (Bradford). A total of 30 µg of proteins/lane were separated in 10% acrylamide gel and transferred to a PVDF membrane (Millipore, Burlington, MA, USA). Membranes were saturated with 5% BSA in 0.3% TBS-Tween-20, washed in TBS-Tween, incubated overnight at 4 °C with an anti-CB1R primary antibody (diluted 1:800), then washed in TBS-Tween, incubated with an anti-rabbit IgG HRP-conjugated secondary antibody (diluted 1:3000; BioRad, Hercules, CA, USA), washed and developed with chemo-luminescence reagent Luminata TM Forte Western HRP Substrate (Millipore). Images were acquired with the ChemiDoc system (Bio-Rad), exported and analyzed.
gnrh3::EGFP fish embryos were collected at 72 hpf and processed as described above. Embryos were sacrificed with a tricaine overdose, fixed in 4% PFA overnight, then washed in PBS. Solution was replaced with 0.1M PB and 7% sucrose at 4 °C for 12 h, then with 0.1M PB and 30% sucrose at 4 °C overnight. Samples were then embedded in OCT (Bio Optica, Mila, Italy) and sectioned at 14 μm thickness. Sections were washed in 0.01M PBS pH 7.4 for 10 min and incubated in normal goat serum (diluted 1:100 in PBS-Triton-X 100), then incubated with the anti-CB1R antibody (diluted 1:600 in PBS-Triton-X 100-normal goat serum) O/N. Slices were then washed in 0.01M PBS pH 7.4 for 10 min and incubated with anti-rabbit IgG CY3-conjugated secondary antibody (diluted 1:800 in PBS-Triton-X 100; Jackson ImmunoResearch Laboratories, Philadelphia, PA, USA) for 45 min. For cellular nuclei visualization, slices were incubated with DAPI (4′,6-diamidine-2′-phenylindole dihydrochloride), then washed in PBS and mounted in DABCO. The fluorescent signal was visualized using a Leica TCS SP5 (Leica Microsystem, Wetzlar, Germany).
4.6. Bioinformatic Searches and Statistical Analyses
Known and putative protein::protein interactions were searched for using the Human Protein Reference Database (http://www.hprd.org
). Only those which appear in the list of at least two entry genes were further considered. Conserved coexpression in transcriptomic datasets was searched for using FuncPred (https://www.mbc.unito.it/en/notizie/funcpred
]. Conversion of the “gene ID” from human to mouse was done using the BioMart tool (http://www.biomart.org
As a reference dataset of differentially expressed genes in the hypothalamus, we used the dataset GSE21278 present in Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/
]. This data set provides extensive gene expression data on developing mouse hypothalamic tissues, done using the Affymetrix MOE430 microarray.
All data were statistically evaluated using commercially available software (SPSS version 18.0 for Windows, SPSS Inc., Chicago, IL, USA; and Microsoft Excel 2011, Microsoft Corporation, Albuquerque, NM, USA). Significance was calculated using a t-test and one-way ANOVA followed by Bonferroni’s post hoc test for multiple comparisons.
4.7. Real-Time qPCR Analysis
Total RNA was extracted from 72 hpf embryo heads with the TRIzol reagent (Life Technologies, Carlsbad, CA, USA), following the manufacturer’s instructions. Genomic DNA was destroyed by passing each sample into a 22G needle connected to a 1 mL syringe. Chloroform was added and, after 15–20 min of centrifugation at 12,000× g
at 4 °C, RNA was precipitated in isopropanol. Samples were centrifuged at 12,000× g
at 4 °C for 15 min and the RNA was then washed with 75% ethanol. The RNA pellet was briefly air-dried, resuspended in 20 µL sterile water and stored at −20 °C. Samples were quantified using a NanoDrop1000 spectrophotometer (Nanodrop Technologies, Inc., Wilmington, DE, USA). Real-time qPCR was performed using Superscript III Platinum One-step qRT-PCR system (Invitrogen, Carlsbad, CA, USA) and the thermal cycler Rotor Gene Q (Qiagen, Hilden, Germany). Each RNA sample was analyzed in three technical replicates containing 50 ng of total RNA. Primers sequences were
Quantification of mRNA abundance in each sample was done using a standard curve, built with several dilution of the samples. The abundance of the z-β-actin housekeeping mRNA was used for normalization. The mRNA expression was calculated relative to the one of control embryos (=1).
4.8. CB1 Ligands Exposure and Confocal Microscope Analysis
Upon fertilization, at the 1-cell stage, embryos were treated with the following ligands (Tocris Bioscience, Bristol, UK): CB1R-agonist WIN 55,212-2 (renamed WIN55 for simplicity, 1 nM, 10 nM and 100 nM); CB1R-inverse agonist Rimonabant (SR141716A) (1 nM, 10 nM, 100 nM and 1 μM); CB1R-antagonist AM251 (100 nM, 1 μM and 5 μM). All these compounds were dissolved in 0.1% dimethyl-sulfoxide (DMSO) and added to the fish water for 72 h (gnrh3::EGFP) or 96 h (agrp1::mCherry) without solution replacement for the entire exposure period. At 24 hpf 0.003% PTU was added to the water for gnrh3::EGFP fish embryos. Embryos were grown in 6-well Petri dishes, 30 embryos/well. Negative control embryos were treated with 0.1% DMSO alone.
For whole mount visualization, the gnrh3::EGFP and agrp1::mCherry embryos were collected at 72 hpf or 96 hpf, sacrificed with a tricaine overdose, fixed with 4% paraformaldehyde (PFA) at 4 °C O/N, washed in PBS and embedded in 4% low melting agarose. The apical portion of the head was manually dissected from the rest of the embryo. Confocal microscopy examination was carried out at 20× and 40× magnification with a Leica TCS SP5 or Leica TCS SP8 (Leica Microsystems). Images were acquired as Z-stacks of 5 μm-thick optical sections. For counting, stacks of Z-slices at 40× were used, images were processed and analyzed with the assistance of ImageJ software (NIH, Bethesda, MD, USA) after adjustment of contrast and brightness.
4.9. Statistical Analyses
Data are expressed as mean ± standard error of the mean (SEM) from three independent experiments. The results of pharmacological treatments/morpholino injections were compared to control (DMSO-treated/uninjected embryos). Statistical analyses were performed using commercially available software (SPSS version 18.0 for Windows and Microsoft Excel 2011). Significance was calculated using Fisher’s exact test for the analysis of altered phenotypes and one-way analysis of variance (ANOVA) for all other experiments, followed by Bonferroni’s and Tuckey’s post hoc tests for multiple comparisons. p ≤ 0.05 was taken as minimum level of significance.