Brodifacoum does not modulate human cannabinoid receptor-mediated hyperpolarization of AtT20 cells or inhibition of adenylyl cyclase in HEK 293 cells

Background Synthetic cannabinoids are a commonly used class of recreational drugs that can have significant adverse effects. There have been sporadic reports of co-consumption of illicit drugs with rodenticides such as warfarin and brodifacoum (BFC) over the past 20 years but recently, hundreds of people have been reported to have been poisoned with a mixture of synthetic cannabinoids and BFC. We have sought to establish whether BFC directly affects cannabinoid receptors, or their activation by the synthetic cannabinoid CP55940 or the phytocannabinoid Δ9-tetrahydrocannabinol (Δ9-THC). Methods The effects of BFC on the hyperpolarization of wild type AtT20 cells, or AtT20 cells stably expressing human CB1- or CB2- receptors, were studied using a fluorescent assay of membrane potential. The effect of BFC on CB1- and CB2-mediated inhibition of forskolin-stimulated adenylyl cyclase (AC) activation was measured using a BRET assay of cAMP levels in HEK 293 cells stably expressing human CB1 or CB2. Results BFC did not activate CB1 or CB2 receptors, or affect the hyperpolarization of wild type AtT20 cells produced by somatostatin. BFC (1 µM) did not affect the hyperpolarization of AtT20-CB1 or AtT20-CB2 cells produced by CP55940 or Δ9-THC. BFC (1 µM) did not affect the inhibition of forskolin-stimulated AC activity by CP55940 in HEK 293 cells expressing CB1 or CB2. BFC (1 µM) also failed to affect the desensitization of CB1 and CB2 signaling produced by prolonged (30 min) application of CP55940 or Δ9-THC to AtT20 cells. Discussion BFC is not a cannabinoid receptor agonist, and appeared not to affect cannabinoid receptor activation. Our data suggests there is no pharmacodynamic rationale for mixing BFC with synthetic cannabinoids; however, it does not speak to whether BFC may affect synthetic cannabinoid metabolism or biodistribution. The reasons underlying the mixing of BFC with synthetic cannabinoids are unknown, and it remains to be established whether the “contamination” was deliberate or accidental. However, the consequences for people who ingested the mixture were often serious, and sometimes fatal, but this seems unlikely to be due to BFC action at cannabinoid receptors.


INTRODUCTION
is an inhibitor of vitamin K epoxide reductase and active ingredient of rodenticides (King & Tran, 2015). There have been sporadic reports of brodifacoum consumption with drugs such as cocaine and cannabis (La Rosa, Clarke & Lefkowitz, 1997;Waien, Hayes Jr & Leonardo, 2001;Spahr, Maul & Rodgers, 2007), however, a large number of people were recently hospitalized with poisoning by brodifacoum and related compounds following ingestion of what are believed to be synthetic cannabinoid receptor agonists (SCRAs) (Kelkar et al., 2018;Riley et al., 2019;Moritz et al., 2018;Panigrahi, Jones & Rowe, 2018). There is limited evidence to suggest that people have on occasions deliberately combined brodifacoum with cannabis (La Rosa, Clarke & Lefkowitz, 1997;Spahr, Maul & Rodgers, 2007), and the apparent mixing of brodifacoum with a variety of different SCRA could be a deliberate attempt to enhance the effects of the drugs through either a pharmacokinetic or pharmacodynamic mechanism. In this study, we have examined the effects of brodifacoum on the acute signalling of human CB 1 and CB 2 receptors in AtT20 and HEK 293 cells. In AtT20 cells, activation of heterologously expressed CB 1 or CB 2 produces a hyperpolarization, mediated by activation of G protein-gated inwardly rectifying K channels (Mackie et al., 1995;Banister et al., 2016). In CB1-or CB2-expressing HEK 293 cells, we measured the real time modulation of forskolin-stimulated cAMP accumulation (Cawston et al., 2013). We found that cannabinoid-induced signaling was not affected by brodifacoum, indicating that combining SCRA with brodifacoum is not likely to enhance user experience through interactions with cannabinoid receptors.

Drugs
(-) CP 55940 was from Cayman Chemical (#90084; Ann Arbor MI, USA), 9tetrahydrocannabinol (THC) was from THCPharm (Frankfurt, Germany) and was a kind gift from the Lambert Initiative for Cannabis Therapeutics (University of Sydney). Brodifacoum was from Sigma-Aldrich (#46036), and forskolin was from Ascent Scientific Ltd.
Proprietary FLIPR membrane potential dye (blue, #R8034, Molecular Devices, Sunnyvale CA) was dissolved in Hank's Balanced Salt Solution (HBSS) of composition (mM) NaCl 145, HEPES 22, Na 2 HPO 4 0.338, NaHCO 3 4.17, KH 2 PO 4 0.441, MgSO 4 0.407, MgCl 2 0.493, CaCl 2 1.26, glucose 5.56 (pH 7.4, osmolarity 315 ± 15) and added to the cells an hour before fluorescence reading began. Dye was used at 50% of the manufacturers recommended concentration, and cells were incubated at 37 • C in humidified room air for loading. Plates were read using a Flexstation 3 (Molecular Devices) plate reader at 37 • C. Plates were excited at a wavelength of 530 nm, emission was measured at 565 nm, with cut-off filter at 550 nm. Drugs were added using the pipetting function of the Flexstation in a volume of 20 µl after recording 60-120 s of baseline fluorescence. Readings were made every 2 s. Drug stocks were made up in DMSO (#D8418, Sigma-Aldrich) and diluted on the day of experiment, the final concentration of DMSO in the assay was 0.1%.
Data were expressed as the percentage change in baseline fluorescence produced by drug addition. The change in fluorescence produced by vehicle (0.1% DMSO) addition was subtracted from the traces before this calculation. Data is expressed as the mean ± SEM of at least 5 independent determinations performed in duplicate, unless otherwise noted. Pooled data was fit to a four-parameter logistic equation in Graphpad PRISM 7 (GraphPad Software, San Diego CA, USA).

Assay of cAMP levels
Human embryonic kidney (HEK) 293 FlpIn cells stably transfected with human CB 1 or CB 2 receptors tagged with three haemagglutinin epitopes at the amino terminus and human G protein gated inwardly rectifying potassium channel 4 (GIRK4) were used (the construction of these cells will be described in another place, and we did not assay CB receptor coupling to GIRK4 in this study). Cells were grown in DMEM containing 10% FBS and 100 units/ml/penicillin, 100 µg/ml streptomycin and were maintained under selection with hygromycin (80 µg ml −1 ) and G418 (400 µg ml −1 ). HEK 293 FlpIn cells were originally obtained from Life Technologies (now Thermofisher, #75007).
Cellular cAMP levels were measured using the pcDNA3L-His-CAMYEL plasmid, which encodes the cAMP sensor YFP-Epac-RLuc (CAMYEL), (Cawston et al., 2013). The pcDNA3L-His-CAMYEL was a kind gift from Dr. Angela Finch (The University of New South Wales, NSW, Australia), and originally obtained from American Type Culture Collection (Manassas, VI, USA). Cells were seeded in 10 cm dishes at a density of 6,000,0000 such that they would be 60-70% confluent the next day. The day after seeding, pcDNA3L-His-CAMYEL plasmid was transiently transfected into cells using linear polyethyleneimine (PEI, m.w. 25 kDa) (#23966, Polysciences, Warrington, PA, USA). The DNA-PEI complex mixture was added to the cells at the ratio of 1:6, and incubated for 24 h in 5% CO 2 at 37 • C. After the incubation, cells were detached from the dish using trypsin/EDTA and the pellet was resuspended in 10 ml Leibovitz's L-15, no phenol red (#21083027; Gibco) media supplemented with 1% FBS, 100 units/ml/penicillin, 100 µg/ml streptomycin and 15 mM glucose. The cells were seeded at a density of 100,000 cells per well in poly D-lysine (Sigma-Aldrich) coated, white wall, clear bottom 96 well microplates. Cells were incubated overnight at 37 • C in ambient CO 2 .
On the following day, drugs were prepared in HBSS containing 0.1 mg ml −1 BSA. For measurement of cAMP inhibition, all the drugs were made in 3 µM of forskolin. Coelenterazine-h substrate (2.5 µM) (#S2011; Promega, Madison, WI, USA) was added to the cells, and incubated for 5 mins prior to the addition of drugs or vehicle. Luminescence was measured using a Flexstation 3 (Molecular Devices) microplate reader at 37 • C at an emission wavelength of 461 nm and 542 nm simultaneously, with an integration time of 1 s. Drugs were added in a volume of 10 µl (10×) to each well to give the desired concentration. The final concentration of DMSO in each well was always 0.1%. Raw data are presented as inverse BRET ratio of emission at 461/542. Background reading (no substrate) was subtracted from raw values before calculating ratios. For convenience, values are expressed such that an increase in ratio correlates with increase in cAMP production. Area under the curve (AUC) analysis was performed in GraphPad prism (Graph Pad Software Inc., San Diego, CA, USA), and data were expressed as percentage of the difference between basal (vehicle, 0%) and forskolin (100%) values over a 5-minute period after forskolin addition.
For experiments examining the potential interaction between brodifacoum and cannabinoids, the cells were pre-treated with 1 µM of brodifacoum (or vehicle) and the response to a subsequent addition of SCRAs was measured. The concentration of DMSO (0.1%) was kept constant for the brodifacoum-treated and control cells. Data was normally distributed (D'Agostino and Pearson normality test, PRISM), differences between groups were tested using unpaired Student's t -Test (PRISM). Statistical significance was defined as P < 0.05.

RESULTS
Acute application of brodifacoum for 5 min at concentrations up to 30 µM did not significantly affect the fluorescence of AtT20 cells expressing CB 1 or CB 2 receptors (Fig. 1). Prolonged exposure to brodifacoum at concentrations greater than 10 µM produced decreases in fluorescence in AtT20 cells expressing CB receptors as well as wild type cells, and so for experiments examining the potential interaction between brodifacoum and cannabinoids we used a concentration of 1 µM.

DISCUSSION
The principal finding of this study is that brodifacoum does not affect CB 1 or CB 2 signaling, either to K channels in AtT20 cells or adenylyl cyclase in HEK 293 cells. In the assay of K channel activation, there was no effect on the concentration response relationship for CP55940 or THC, and brodifacoum did not affect the desensitization of signaling produced by prolonged application of either drug. Brodifacoum had no effect on the potency, maximum effect or time-dependence of the actions of the high efficacy synthetic cannabinoid CP55940 or the lower efficacy phytocannabinoid THC, indicating that it is unlikely to act as modulator of the pharmacodynamic effects of cannabinoids.
Activation of GIRK is mediated by the Gβγ subunits of G protein heterotrimers, and many Gi/Go coupled receptors effectively signal through this pathway in AtT20 cells (e.g., Mackie et al., 1995;Günther, Culler & Schulz, 2016;Knapman et al., 2013;Heblinski, Bladen  & Connor, 2019). We have previously used the fluorescent measurement of membrane potential to study CB 1 and CB 2 agonists, antagonists, and allosteric modulators of CB 1 (Cawston et al., 2013). Inhibition of adenylyl cyclase activity by CB receptors is mediated via the Gα subunits of G protein heterotrimers, and brodifacoum also failed to affect this signal transduction pathway. The precise cellular signaling mechanisms responsible for the subjective effects of Cannabis and synthetic cannabinoid agonists are not established, although the signal transduction of cannabinoid receptors has been extensively studied (Howlett & Abood, 2017;Ibsen, Connor & Glass, 2017) and it is unlikely that any one pathway is responsible. It remains formally possible that brodifacoum could selectively modulate pathways other than Gβγ -mediated activation of GIRK or Gα-mediated inhibition of cAMP accumulation, but the lack of any effect whatsoever on the effects of CP55940 or THC suggests that ligand interactions with cannabinoid receptors are unaffected by brodifacoum. The concentration of brodifacoum in blood or brain after co-ingestion with synthetic cannabinoids is unknown. However, concentrations of up to 3 µM have been reported in the serum of people who have deliberately ingested large quantities of rat poison (Weitzel et al., 1990;Hollinger & Pastoor, 1993), and inhalation of BFC via smoked synthetic cannabinoids may produce higher serum concentrations of BFC than ingestion. Brodifacoum at 1 µM failed to affect CB 1 or CB 2 receptor signaling when measured continuously over a period of 30 min, and 10 µM brodifacoum failed to mimic or affect the acute response to a maximally effective concentration of CP 55940, although at this concentration prolonged application of brodifacoum produced a decrease in the fluorescence of wild type AtT20 cells, as well as those expressing CB 1 and CB 2 receptors. This effect at higher concentrations may reflect direct interactions of brodifacoum with cell membranes (Marangoni et al., 2016). Concentrations of brodifacoum in the upper range of what we tested are achieved only after ingestion of large amounts of rat bait, it is possible that they could be achieved while ingesting contaminated synthetic cannabinoids, but this remains unreported.
Several case reports suggest an interaction between therapeutic warfarin and cannabis or cannabidiol (Grayson et al., 2018;Yamreudeewong et al., 2009;Damkier et al., 2019). It has been suggested that cannabinoid inhibition of enzymes responsible for the metabolism of warfarin can increase blood levels of the drug, and while these studies have focussed on potentially dangerous changes in warfarin concentration, levels of cannabinoids could also be reciprocally elevated. Such interactions may inform a decision to deliberately combine ''superwarfarin'' with SCRA, as has been previously suggested for cannabis (La Rosa, Clarke & Lefkowitz, 1997;Spahr, Maul & Rodgers, 2007), although whether brodifacoum is metabolized by pathways shared with SCRA in humans is unknown. Information about how or even whether BFC is metabolized in humans is very sparse, although available evidence suggests metabolism is very limited or absent (Hauck, Feinstein & Van Breeman, 2016). Apart from the obvious danger of ingesting brodifacoum, altering the metabolism of SCRA is likely to have unpredictable consequences, as some metabolites of SCRA retain cannabinoid receptor activity (e.g., Brents et al., 2011;Chimalakonda et al., 2012;Longworth et al., 2017;Cannaert et al., 2016), and may contribute to the overall SCRA experience.
Ingestion of brodifacoum is relatively common, while death from exposure is rare, owing to ready treatment with vitamin K (King & Tran, 2015;Gummin et al., 2018). The high number of deaths associated with the combination of SCRA and anticoagulants in 2018 (at least eight; Connors, 2018) may point to an interaction between the drugs. It may also reflect the identity and dose of the synthetic cannabinoid(s) consumed, as well as the general health status of the drug users. Deaths from synthetic cannabinoid exposure are uncommon, but well documented (e.g., Kasper et al., 2015;Trecki, Gerona & Schwartz, 2015). While there is a general acceptance that brodifacoum or a similar agent is responsible for the coagulopathies associated with synthetic cannabinoid ingestion, identification of the synthetic cannabinoid has not been reported in most cases, but a recent report identified a metabolite of AB-FUBINACA in one patient following ingestion of ''King Kong'', a brodifacoum laced SCRA (Riley et al., 2019). It seems unlikely, though, that brodifacoum would interact with higher efficacy or potency SCRAs at cannabinoid receptors when it clearly does not interact with CP55940 or THC signaling (Noble et al., 2019;Sachdev et al., in press). Intriguingly, several groups have reported cannabinoid receptor ligands based on a coumarin scaffold (Behrenswerth et al., 2009;Han et al., 2015). While these drugs have been reported to be either antagonists/inverse agonists (Behrenswerth et al., 2009) or CB 2 -selective agonists (Han et al., 2015), they remain largely uncharacterized. Given the propensity of chemists producing and, in some cases, designing cannabinoids for the recreational market, it cannot be ruled out that coagulopathy may be an unanticipated adverse effect of a synthetic cannabinoid, which may have arisen from a novel, coumarinbased cannabinoid that retains some of the vitamin K epoxide reductase inhibitory of warfarin and brodifacoum.
In conclusion, we report that brodifacoum does not appear to be an agonist or antagonist of human cannabinoid receptors, and it also does not appear to be an allosteric modulator of CB 1 or CB 2 activation of K channels or inhibition of adenylyl cyclase. Why brodifacoum has been mixed with synthetic cannabinoid receptor agonists remains a matter for speculation, although an intended effect on synthetic cannabinoid drug pharmacokinetics cannot entirely be ruled out.