Effects of cannabinoids on the development of chick embryos in ovo

We have examined the effects of the synthetic cannabinoids HU 210 and HU 211, the plant-derived cannabidiol and the endogenous cannabinoid anandamide on the viability and development of chick embryos. Fertilized White Leghorn chicken eggs were injected with the test compounds or carrier vehicle, via a drilled small hole in the egg, directly into the egg yolk. After nine days of exposure, the embryonal viability, length and wet weight of embryos, and wet weight of brains were measured, and the development stages were assessed according to the Hamburger and Hamilton (HH) scale. The potent synthetic cannabinoid receptor agonist HU 210 and the non-psychotropic cannabidiol were embryotoxic at the highest concentrations examined (10 µM and 50 µM, respectively), with no viable embryos after the HU 210 injection, and 20% viability after the cannabidiol injections. The effects of HU 210 on the chick embryo were attenuated by α-tocopherol and the cannabinoid receptor antagonist AM251, whereas only α-tocopherol gave a statistically significant protection against the embryotoxic effects of cannabidiol. This study shows that exposure to plant-derived or synthetic cannabinoids during early embryonal development decreases embryonal viability. Extrapolation of data across species is of course difficult, but the data would argue against the use of cannabinoids, be it recreationally or therapeutically, during pregnancy.

"Spice" products, herbal mixtures added with an increasing number of different synthetic CBs that are smoked to obtain cannabis-like effects 19 , of which many lack toxicological data 20 .
Data concerning the effects of cannabinoids on the developing chick embryos are rare and represented by some studies on the role of CB 1 receptor signaling during neurodevelopment 21 , and the effects of THC and analogues on nucleic acid synthesis 22 , neural development 23 , and the function of embryonic heart cells 24 .
In the present study, we have investigated the effects of the synthetic cannabinoids HU 210 and HU 211, the plant-derived cannabidiol (CBD) and the endogenous cannabinoid anandamide (AEA) on the viability and development of chick embryos.
The avian CB 1 receptors have a very close homology to human CB 1 receptors and the chicken CB 1 receptor pharmacology, assessed in binding experiments, resembles very closely its rodent equivalent 25 . We have previously shown that the CB 1 receptor recognition sites in the chicken brain have similar pharmacological properties to those in the rat brain 26 . The CB 1 receptors retain functional coupling to their G-proteins, since the synthetic cannabinoid agonist CP 55,940, at nanomolar concentrations produced a robust decrease in the cAMP response to forskolin in a manner that was blocked by the selective CB 1 receptor antagonist AM251 27 .
HU 210 is one of the most potent synthetic cannabinoids yet discovered, being 100-800 times more potent than Δ 9 -THC 28,29 , with a long duration of action (1-2 days in rhesus monkeys 30 ).
HU 210 is an analogue of the tricyclic benzopyran Δ 9 -THC with reported K i values of 0.06-1.6 nM at the CB 1 receptor and 0.52 nM at the CB 2 receptor [31][32][33] . We have previously showed that synthetic CB receptor agonists found in Spice branded products 34,35 , such as HU 210 and CP 55,940, are cytotoxic in various in vitro cell models [36][37][38] . It is therefore possible that, beside the well-known side effects of cannabis or cannabis-like drugs, prolonged and/or excessive recreational use of HU 210 may be cytotoxic to a wide range of cells in the body, especially cells undergoing rapid mitotic division such as immune cells and cells in the developing embryo. HU 211 is a nonpsychotropic enantiomer of HU 210 lacking activity at CB receptors but acts as a glutamate NMDA receptor antagonist 39 . CBD is a natural constituent of Cannabis sativa that lacks psychotomimetic properties. CBD is a weak, noncompetitive, and negative allosteric modulator of CB 1 receptors [40][41][42] , however, even if the pharmacological action of CBD seems to not involve cannabinoid receptors, CBD exerts several important effects on the central nervous system, including anxiolytic, antipsychotic 43 analgesic, or antiepileptic effects 44,45 . AEA, a N-acylethanolamine fatty acid amide, was the first endogenous ligand (endocannabinoid) to the CB receptors to be discovered 46 . The affinity of AEA to the human CB receptors have been reported to be (K i values) 239 nM at the CB 1 receptor and 440 nM at the CB 2 receptor 47 . The pharmacological properties of AEA, that is rapidly degraded by the intracellular enzyme fatty acid amide hydrolase (FAAH) 48 , is characterized by actions at a number of non-cannabinoid receptor targets such as the transient receptor potential vanilloid 1 (TRPV1) 7 and the peroxisome proliferation-activated receptor 49 .
The effects of the specific CB 1 receptor antagonist AM251 and the antioxidant α-tocopherol were also investigated since these compounds have blocked or attenuated the antiproliferative/cytotoxic effects of CBs in various in vitro cell models [36][37][38]50 .
We can show that the potent synthetic CB receptor agonist HU 210 as well as the plant-derived CBD, with low CB receptor activity, are embryotoxic in a manner that could be attenuated by 100 µM α-tocopherol or by 1 µM AM251, when injected in ovo.

Results
To examine the concentration-dependent effects of CBs on the development and survival of chick embryos, the test compounds were injected directly into the egg yolk. Of 305 eggs used in the study, 40 eggs were injected with saline, 41 eggs with DMSO, 12 eggs with ethanol, and 212 eggs with the test compounds at day 1, 4 and 7 after incubation. At incubation day 10 (corresponds approximately to stage 36 according to the HH scale), 290 eggs were examined for weight and vascularization of the egg, embryonal mean crown-rump (C-R) length, mean wet embryonal weight, and mean wet brain weight ( Table 1). The mean weight loss of fertilized eggs injected with saline and incubated for 10 days was −2.53 ± 0.07 g (mean ± SEM). None of the injected compounds significantly affected the egg weight loss.
Embryonal viability and the Hamilton and Hamburger development stage were assessed based on the embryonal length and weight measurements, and overall 88% viable embryos were found in the saline control group with an average HH stage of 35 ± 0.83 (means ± SEM), 83% viability when exposed to DMSO (HH stage 34 ± 1.1), and ethanol treatment resulted in 89% viability (HH stage 35 ± 1.6). Upon visual examination, no gross morphological malformations of e.g. limbs, heart, beak, brain or eyes were observed in CB-treated chick embryos or in the controls.
The synthetic CB HU 210 (Fig. 1A), at a concentration of 10 µM but not at 1 µM or 3 µM, killed all chick embryos (n = 12) and the average HH stage that had been reached was only 19 ± 3.5 (P < 0.0001 vs. control DMSO treatment). This stage corresponds to approximately day 2-3 according to the HH scale 51 . Thus at least 50% of the embryos died after the first injection of 10 µM HU 210, and after the second injection (day 4) all embryos were dead. HU 211 did not produce any statistically significant effects on the chick embryos (Fig. 1B). On the other hand, CBD showed dose-dependent effects on both the viability and the development ( Fig. 1C; P = 0.0003). The viability dropped 50% at a concentration of 5 µM and the HH stage decreased to 27 ± 4.0 (corresponding to approximately day 5 according to the HH scale), and at 50 µM, the viability was 20% and the HH stage only 11 ± 4.9 (n = 10; P < 0.001 when compared to DMSO controls). AEA showed a weak, but statistically insignificant, dose-dependent effect. At 50 µM, the viability was 80% and the HH stage 30 ± 3.3 (Fig. 1D).
In a subsequent series of experiments, we tested the effects of the CB 1 receptor antagonist AM251 and the antioxidant α-tocopherol on the embryotoxicity produced by HU 210 and CBD. Neither α-tocopherol (100 µM), nor AM 251 (1 µM) had any statistically significant effects per se on the viability and development of chick embryos (Table 1 and Fig. 1E). However, both compounds protected against the embryotoxicity produced by 10 µM HU 210, since the embryo viability increased from 33% in eggs treated with HU 210 to 67% in eggs injected with HU 210 + or α-tocopherol, and 78% when injected with the combination of HU 210 and AM251. The HH stage also increased from 24 ± 2.1 (HU 210 per se), to 32 ± 2.8 (P < 0.05) and 33 ± 2.1 (P < 0.01), respectively. α-Tocopherol also protected against the toxicity produced by 50 µM CBD (22% viability; HH stage 24 ± 2.6; P < 0.01 vs. DMSO controls) by increasing the viability to 78% and restoring the HH stage to 33 ± 2.2 (P < 0.05 vs. CBD per se). AM 251 did not produce a statistically significant protection against the embryotoxic effect of CBD. The vascularization score of the chicken eggs followed the pattern of the HH stages, with an average of 1.97 ± 0.03 in eggs injected with saline that were significantly reduced to 1.16 ± 0.16 in eggs injected with 10 µM HU 210 and 0.95 ± 0.18 in eggs injected with 50 µM CBD (Table 1).
Since drug-induced changes in maternal and placental physiology are eliminated, the model is especially useful for studies focusing on the molecular mechanisms by which drugs alter embryonal development. We have used in ovo yolk injections to administer the compounds. This method has been shown to be superior in order to minimize perturbations to the embryos and to decrease the occurrence of gross morphological abnormalities, induced by the treatment mode per se 60 .
It has been shown that the cannabinoid CB 1 receptor expression spatially and temporally follows neuronal differentiation in the early chick embryo 61 , and we have previously shown that the pharmacological properties of the chick CB 1 receptor are broadly similar to the rodent counterpart 26 . By using the CB 1 receptor-selective antagonist radioligand [ 3 H]SR 141716 A we found that chicken brain CB 1 receptors are pharmacologically similar to rat brain CB 1 receptors, and that rates of association and dissociation, K D values and inhibition by a variety of cannabinoid ligands were very similar in membranes from chicken and rat brain 26 . In addition, the shift to the right of the inhibition of [ 3 H]SR 141716A binding by CP 55,940 in the presence of nonhydrolysable GTP analogues 26,27 is in line with a previous studies in the rat brain, and would suggest that in the chicken membranes, the CB 1 receptors retain functional coupling to their G-proteins.
In present study, the synthetic CB HU 210 and the plant-derived CBD showed embryotoxic, rather than teratotoxic, effects in the chick embryo model system. Psychoyos et al. 23 showed that the water-soluble analogue of Δ 9 -THC, O-2545, caused impaired formation of brain, heart, somite, and spinal cord primordial, during early embryogenesis in the chick embryo 23 . The ability of a CB 1 receptor antagonist to block these effects was, however, not investigated. This is of importance, given that cannabinoids can affect cell viability by both CB receptor-dependent and -independent mechanisms 62 .
We could not observe any gross structural abnormalities of embryonal skeletal system such as beak deformities, absent, supernumerary or deformed limbs, nor malformations of the heart, brain or eyes in the CB-treated embryos. This is supported by a study from 1976, where THC reduced the embryonal body weight without causing any teratological effects, when injection into yolk sac of chicken eggs at various times of incubation 22 . One explanation to this discrepancy is that the CBs induce early (after the first injection at day 1) teratological damage that we have been unable to observe at day 10 due to an early embryonic death.
HU 210 produced a pronounced embryotoxicity at the added concentration of 10 µM. In contrast, the enantiomer HU 211, which lacks CB receptor-mediated effects, was without embryotoxic effects. This implicates the involvement of CB 1 receptor in the effects of HU 210 and the conclusion is further supported by the finding that the CB 1 receptor antagonist AM251 reduces the toxicity produced by HU 210.
The endogenous CB ligand AEA, on the other hand, was without effect upon embryonic development at the concentrations tested. AEA is, however, avidly metabolised by the enzyme fatty acid amide hydrolase, which is present in the chicken 63 , and this presumably limits its embryotoxicity. In future studies, it would be instructive to determine whether or not embryonic toxicity is seen when this enzyme is inhibited, although in general the enzyme is less sensitive to pharmacological inhibition than its rodent counterpart 63 , which may be a complicating factor. CBD, with reported weak CB receptor activity but approved by the Food and Drug Administration for treatment of children with refractory epilepsy 64 , also induced embryotoxicity, albeit at a concentration that was five times higher than HU 210.
The teratogenic potential of CBD has been evaluated in zebrafish, and it was found that CBD did not cause any malformations of the embryos but delayed the hatching compared to control embryos 65 . In the present study, the embryotoxicity by CBD was not significantly blocked by AM251 but was sensitive to the antioxidant α-tocopherol (as was that of HU 210). CBD has been reported both to act as a protective antioxidant 66 and as a toxic agent producing an increase in reactive oxygen species 67 in vitro. Our data would suggest that the latter mechanism may be operative in the chick embryo, although as a caveat it should be pointed out that α-tocopherol is by no means a specific antioxidant since it demonstrates effects that are unrelated to the antioxidant activity and possibly reflect specific interactions with enzymes, structural proteins, lipids and transcription factors 68,69 .
The most obvious mechanistical explanation to the anti-embryotoxic effects of α-tocopherol would be its antioxidative properties. However, we have published data showing that α-tocopherol has the ability to protect against the antiproliferative or cytotoxic effects of cannabinoids in human colorectal carcinoma cancer cells 37 , rat C6 glioma cells 36,50 , and mouse embryonal P19 teratocarcinoma cells 38 , whereas other non-specific antioxidants as N-acetyl-L-cysteine, N-nitroarginine methyl ester or ascorbic acid have failed to protect the cells. α-Tocopherol has even attenuated CB-induced toxicity in C6 cells in the absence of large-scale lipid peroxidation. There are numerous mechanistical explanations for the protective effects of α-tocopherol, since this compound can interact as a ligand at certain receptors and transcription factors and has been shown to inhibit protein kinase C, 5-lipoxygenase and phospholipase A 2 and activate protein phosphatase 2A and diacylglycerol kinase 70 . Some genes (e. g. scavenger receptors, α-TTP, α-tropomyosin, matrix metalloproteinase-19 and collagenase) are modulated by α-tocopherol at the transcriptional level 70 . α-Tocopherol also inhibits cell proliferation 71 , platelet aggregation 72 and monocyte adhesion 73 .
However, absence of evidence is not evidence of absence, since the results do not rule out a reasonable conclusion to make based on our findings when taken together that the CBs produce their embryotoxic effects, at least in part, as a result of oxidative stress.
However, this conclusion must be validated in further experiments. A final note concerns whether the concentrations used are pharmacologically relevant. In the study investigating the embryotoxic effects of O-2545, Psychoyos et al. 23 reported that their concentration range "correspond to levels approximately 60-to 116-fold higher than levels of Δ 9 -THC detected in humans following inhalation of a (2019) 9:13486 | https://doi.org/10.1038/s41598-019-50004-7 www.nature.com/scientificreports www.nature.com/scientificreports/ single marijuana cigarette with 3.55-20% Δ 9 -THC content 74 , and thus reflects the concentrations of cannabinoid found in chronic cannabis users". We have used much lower concentrations of CBs in our study, but it is difficult to discuss about clinically relevant concentrations since there are, to our best knowledge, no data available on concentrations found in human beings of e.g. HU 210 after consumption of Spice. It is of course important to be careful when attempting to interpret our findings in the chicken embryo to the human conditions, but since basic molecular and cellular control of early organogenesis is highly evolutionary conserved among species, the results obtained suggest a risk of cannabinoid consumption during pregnancy that may cause early miscarriage or developmental disturbances. This is important when considering the current discussion about legalisation of medical marijuana and therapeutic application of cannabinoid-based drugs.
Handling of chick embryos and drug administration. Fertilized White Leghorn chicken eggs (Gallus gallus), obtained from Agrisera AB (Vännäs, Sweden), were weighted and incubated with the blunt end up in an humified egg incubator (Agroswede, Malmö, Sweden) at 38 °C. The eggs were turned automatically every two hours. Before incubation, the eggs were weighed, cleaned with 70% ethanol to decrease external contamination and marked with a graphite pencil. After 24 hours of incubation, a six mm hole was made in the eggs using a Co/ Tech mini grinder drill (Clas Ohlson, Sweden), and test compounds were injected with a sterile 27G 1½ needle into the yolk, carefully without harming the embryo. The hole was sealed with PVC-tape (Clas Ohlson, Sweden) between the injections. The injections were repeated day 4 and 7 and before each injection the eggs were cleaned with 70% ethanol. CBs are highly lipophilic and thus requires solubilization with DMSO, ethanol, acetonitril or surfactant agents, all of which are potentially embryotoxic. To overcome this problem, we have kept the vehicle concentration to approximately 0.05% (v/v) by serial dilution of the cannabinoids in anhydrous DMSO (≥99.9%) and α-tocopherol in ethanol (≥95%). Each stock solution was then diluted in saline to final concentrations of which 100 µl was injected into the yolk. All statistical comparisons are made against vehicle controls, i.e. eggs injected with 100 µl of 0.0498% (v/v) DMSO or 0.0475% (v/v) EtOH. Fertilized chick eggs continuously loose about 0.6-0.7% in weight per day due to evaporation of water from the eggs 75 . This is essential for the formation of the air cell and to facilitate optimised water and mineral balances in the different embryonic compartments formed during embryonic development. All eggs examined were weighted before incubation and on the day of analysis (day 10). After 10 days of incubation, the eggs were cracked and the vascularization (i.e. amount of blood vessels and capillaries) of the yolk sac and the chorioallantoic membrane (CAM) was visually examined. The vascularization was scored 0-2, where 0 is no vascularization, 1 small blood vessels with fewer capillaries, and 2 normal large blood vessels with extensive capillary network evenly distributed over the CAM surface (see Supplementary Fig. S1).
The embryos were removed from the eggshell and dissected from the extra-embryonic membranes, and viability, gross morphological abnormalities, wet embryo weights, crown-rump lengths and wet brain (optic lobes and cerebral hemispheres) weights were examined. Based on these measurements, the Hamilton and Hamburger stage (which go from 1 to 46) were determined. The Hamilton and Hamburger stages describe the structural development of the chick embryo and thereby separate morphological from temporal development, since developmental status for a given chronological age can vary widely 51 . This study was reviewed by the Regional Animal Ethics Review Board in Umeå (ref. no. A 15-08), who returned it with an advisory opinion: because the chick embryos were taken before day 12, ethical approval was not required according to Swedish law (SFS 1988:539; SJVFS 2015:24).
Statistics. Statistical differences with respect to the appropriate controls were assessed by one-way ANOVA with post-hoc Bonferroni's multiple comparisons test for the C-R length, embryo and brain weight data, post-hoc Dunnett´s multiple comparison test for the Hamilton-Hamburger scores, and Kruskal-Wallis nonparametric test with Dunn's test for multiple comparisons for the vascularization scores using the statistical package built into the GraphPad Prism computer programme (GraphPad Software Inc., San Diego, CA, USA).