Enhanced Brain Disposition and Effects of Δ9-Tetrahydrocannabinol in P-Glycoprotein and Breast Cancer Resistance Protein Knockout Mice

The ABC transporters P-glycoprotein (P-gp, Abcb1) and breast cancer resistance protein (Bcrp, Abcg2) regulate the CNS disposition of many drugs. The main psychoactive constituent of cannabis Δ9-tetrahydrocannabinol (THC) has affinity for P-gp and Bcrp, however it is unknown whether these transporters modulate the brain accumulation of THC and its functional effects on the CNS. Here we aim to show that mice devoid of Abcb1 and Abcg2 retain higher brain THC levels and are more sensitive to cannabinoid-induced hypothermia than wild-type (WT) mice. Abcb1a/b (−/−), Abcg2 (−/−) and wild-type (WT) mice were injected with THC before brain and blood were collected and THC concentrations determined. Another cohort of mice was examined for THC-induced hypothermia by measuring rectal body temperature. Brain THC concentrations were higher in both Abcb1a/b (−/−) and Abcg2 (−/−) mice than WT mice. ABC transporter knockout mice exhibited delayed elimination of THC from the brain with the effect being more prominent in Abcg2 (−/−) mice. ABC transporter knockout mice were more sensitive to THC-induced hypothermia compared to WT mice. These results show P-gp and Bcrp prolong the brain disposition and hypothermic effects of THC and offer a novel mechanism for both genetic vulnerability to the psychoactive effects of cannabis and drug interactions between CNS therapies and cannabis.


Introduction
Cannabis is the most widely used illicit drug in the world and may trigger psychiatric disorders such as psychosis and addiction in genetically vulnerable individuals [1]. Research identifying the exact genes that comprise genetic vulnerability to cannabisinduced brain disorders is in its infancy. A catechol-o-methyltransferase gene variant (COMT) increases the risk of developing schizophrenia in adolescent cannabis users [2,3]. Our own preclinical research has shown that the schizophrenia susceptibility gene, neuregulin 1, modulates the neurobehavioural effects of the main psychoactive constituent of cannabis, D 9 -tetrahydrocannabinol (THC) [4,5,6,7]. While studies have focused on pharmacodynamic aspects of gene-cannabinoid interactions, here we aim to explore a mechanism involving adenosine triphosphate (ATP)binding cassette (ABC) transporters that might enhance disposition of THC in the brain.
ABC transporters are a family of drug efflux pumps that utilize ATP hydrolysis to transport substrates across biological membranes and were originally discovered due to their role in mediating multidrug resistance (MDR) to anti-cancer drugs [8]. They regulate the disposition of many drugs in tissues because they are localised in excretory organs such as the liver, intestine and the blood-brain barrier (BBB) [9,10,11,12]. Two of the most characterized ABC transporters, P-glycoprotein (P-gp, encoded by MDR1/ABCB1) and breast cancer resistance protein (Bcrp, encoded by BCRP/ABCG2), are both found at the apical surface of capillary endothelial cells that comprise the BBB. P-gp has been shown to regulate the brain accumulation of various central nervous system (CNS) drugs including antipsychotics, selective serotonin reuptake inhibitors, opioids and antiepileptic agents [13,14,15]. Similarly Bcrp affects brain disposition of drugs [16] and often cooperates with P-gp to extrude compounds from the brain because these transporters share overlapping substrate specificities [17,18,19].
A recent study suggests P-gp regulates the brain disposition of THC, as an ABCB1 polymorphism was associated with an increased risk of developing cannabis dependence [20]. In vitro studies show cannabinoids bind ABC transporters and that THC stimulates P-gp ATPase activity and inhibits Bcrp [21,22,23,24,25]. Further, CFI mice naturally deficient in P-gp expression accumulate higher plasma levels of THC following oral dosing than wild-type (WT) mice implying that THC is a substrate of P-gp [26]. However no prior studies have directly shown that Pgp and Bcrp regulate the brain disposition of THC.
One strategy to examine whether THC is a substrate of P-gp and Bcrp with implications for its CNS-disposition is to utilize mice devoid of ABC transporter genes [14,27,28]. Prior studies have demonstrated that Abcb1a/b (2/2) and Abcg2 (2/2) mice attain higher brain levels of P-gp and Bcrp substrate drugs [14,19].
Using these mice we aim to show that Abcb1a/b (2/2) and Abcg2 (2/2) retain higher brain levels of THC and that this has functional consequences for CNS-mediated actions of THC.
Augmented THC-induced hypothermia in Abcb1a/b (2/2) and Abcg2 (2/2) mice Given that both Abcb1a/b (2/2) and Abcg2 (2/2) mice retained higher concentrations of THC in the brain we proceeded to test whether this had functional consequences by examining whether Abcb1a/b (2/2) and Abcg2 (2/2) mice were more sensitive than WT mice to the hypothermic actions of THC -an effect mediated by the CNS (see Figure 2). THC exposure in WT mice significantly reduced body temperature from baseline at 30 min post-injection (Wilcoxon test, P = 0.0180) and this hypothermia disappeared 2-3 h following THC exposure. Abcg2 (2/2) mice had significantly greater THC-induced hypothermia than WT mice at the 2 and 3 h time-points post-injection (Mann-Whitney U test, P = 0.0452 and P = 0.0319 respectively). Abcb1a/b (2/2) mice had significantly greater hypothermia than WT mice at 2 h post-injection only (Mann-Whitney U test, P = 0.0266).
Our results accord with prior in vitro and in vivo evidence indicating THC is a substrate of P-gp. THC was first shown to have affinity for P-gp as it weakly stimulated human P-gpdependent ATPase activity [25]. Bonhomme-Faivre et al. (2008) then demonstrated THC is a P-gp substrate by showing that CFI mice, which are naturally deficient in Abcb1a, attain higher plasma levels of THC than WT mice following a 25 mg/kg oral THC dose [26]. Our results extend this research by showing P-gp regulates brain disposition of THC. Further, we provide first evidence that THC is a substrate of Bcrp and that this transporter regulates brain and blood levels of THC. This finding is consistent with our prior in vitro evidence showing THC inhibits Bcrp [22] and suggests THC-induced enhancement of mitoxantrone accumulation in Bcrp-overexpressing cells occurs via competitive modulation of Bcrp transport.
Our results highlight that THC is a dual substrate for P-gp and Bcrp which is not surprising given the considerable overlap in the substrate specificities of P-gp and Bcrp and that many drugs are transported by both these efflux proteins [17,18,19,32]. THC appears a better substrate for Bcrp than P-gp because Abcg2 (2/2) mice retained higher levels of THC than Abcb1a/b (2/2) mice in the brain at 3 h post-administration. This is at odds with findings that P-gp has approximately 5-fold higher expression in murine brain vascular endothelial cells than Bcrp [33,34]. However, our laboratory has shown that THC inhibits Bcrp but not P-gp because THC increases accumulation of the Bcrp substrate mitoxantrone but not the P-gp substrate rhodamine 123 in cells selectively expressing these respective transporters [22,23]. Taken together, our results imply THC may have significantly lower affinity for P-gp than Bcrp.
Our results suggest that P-gp and Bcrp expressed on endothelial cells that comprise the BBB assist in the active transport of THC from inside the brain back into the peripheral blood supply. Abcb1a/b (2/2) and Abcg2 (2/2) mice, without these transporters, display delayed brain elimination of THC leading to the accumulation of this psychoactive constituent of cannabis in the CNS. This implies that Abcb1a/b (2/2) and Abcg2 (2/2) mice could also be more sensitive to CNS-mediated functional effects of THC. It is well established that THC promotes hypothermia through its action on the CNS largely by affecting thermoregulatory neurons in the hypothalamus [35]. For example, a classic study by Fitton and Pertwee (1982) demonstrated that intracerebroventricular and intrahypothalamic injections of THC promoted reductions in rectal body temperature in mice [36]. Here we show that mice devoid of P-gp or Brcp are more sensitive to hypothermia promoted by THC -a likely functional consequence of enhanced THC concentrations in the brain affecting thermoregulatory neurons in these mice.
Our results offer an additional pharmacokinetic mechanism for interactions between the approximately 80 different cannabinoid constituents found in the cannabis plant. Interaction between THC and another plant-derived cannabinoid constituent cannabidiol (CBD) is of particular interest as CBD modulates the actions of THC [31,37,38,39,40,41]. Indeed this was part of the rationale behind the development of SativexH, a cannabinoid formulation containing a 1.07:1 THC : CBD dose, approved for use in various countries for the treatment of spasticity associated with multiple sclerosis. CBD's modulation of THC effects might be partly explained by pharmacokinetic mechanisms because CBD increases the brain levels of THC and accentuates some of its neurobehavioural effects [31,37,42]. Our present findings showing THC is a substrate of P-gp and Bcrp, combined with prior research indicating CBD inhibits both P-gp and Bcrp [23,25], highlights a novel ABC transporter-mediated mechanism whereby CBD might inhibit the transport of THC from the brain parenchyma back into the blood.
Cannabis-drug interactions should also be considered in future research as substrates for P-gp and Bcrp include several CNS drugs that might be co-administered with cannabis such as anticonvulsants, antidepressants and antipsychotics [14,43,44,45,46,47,48,49]. This may be particularly problematic for antipsychotic treatment as there are higher rates of cannabis use in schizophrenia patients than the general population and cannabis use is associated with poor antipsychotic treatment outcomes [50,51,52]. THC behaving as a competitive substrate for P-gp might increase the brain and blood concentrations of P-gp substrate antipsychotic drugs such as risperidone and olanzapine [14,48,53] and increase dose-related side-effects of these drugs. Alternatively, antipsychotic drugs by inhibiting P-gp and Bcrp [54] might increase CNS accumulation of THC and increase psychotic reactions and relapse in these patients.
The present findings raise the question of whether ABCB1 and ABCG2 polymorphisms might explain genetic susceptibility to cannabis-induced psychosis or the adverse psychoactive effects of this drug (e.g. addiction, anxiety and cognitive impairment) by altering the brain disposition of THC. This would provide a pharmacokinetic source of interindividual variability in response to cannabis use in addition to other genes that modulate the pharmacodynamic actions of the drug [2,4,5,6]. Thus far, to the best of our knowledge, only one human study has addressed whether single nucleotide polymorphisms (SNPs) in ABC transporter genes confers greater vulnerability to CNS disorders. This study showed that the ABCB1 variant C3435T (CC genotype) Figure 2. Greater THC-induced hypothermia in Abcb1a/b (2/2) and Abcg2 (2/2) mice compared to WT mice. THC-induced hypothermia in WT, Abcb1a/b (2/2) and Abcg2 (2/2) mice at 0.5, 2 and 3 h post-THC injection. Data expressed as a change from baseline (pre-THC injection) rectal body temperature reading in these mice at the various time-points (mean 6 SEM). Abcb1a/b (2/2) or Abcg2 (2/2) compared to WT group at specific time-points * P,0.05, baseline versus post-THC injection body temperature in WT mice # P,0.05. For WT mice n = 7, Abcb1a/b (2/2) n = 6 and Abcg2 (2/2) n = 6. doi:10.1371/journal.pone.0035937.g002 increased the risk of developing cannabis dependence, as this SNP was more prevalent in individuals dependent on cannabis than controls [20]. The CC genotype has been associated with increased expression of P-gp, whereas the TT genotype decreased P-gp expression [55]. Therefore, dependent on genotype the SNP C3435T might decrease or increase THC brain levels and on one hand increase the propensity of individuals for cannabis dependence (CC genotype) or increase the risk of developing cannabis-induced psychosis (TT genotype). Our findings reinforce the need for future studies to more closely examine the relationship between SNPs in ABCB1 and ABCG2 and cannabis-related psychiatric problems such as addiction, schizophrenia and panic disorder.
In conclusion our results show that the main psychoactive constituent of cannabis THC is a substrate for the ABC transporters P-gp and Bcrp. Abcb1a/b (2/2) and Abcg2 (2/2) mice attained higher brain THC concentrations and were more sensitive to the CNS-mediated hypothermic effects of THC than WT mice. These results provide a novel mechanism for genetic vulnerability to the psychoactive effects of cannabis, the most widely used illicit drug in the world. Further they highlight a novel mechanism for drug interactions between cannabis use and CNS drugs that are substrates for the ABC transporters P-gp and Bcrp.

Animals
All Experiments were approved by the University of Sydney Animal Ethics Committee in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. Agematched female Abcb1a/b (2/2), Abcg2 (2/2) and wild-type (WT, FVB background strain) mice were used in this study. Abcb1a/b (2/2) and Abcg2 (2/2) mice were purchased from Taconic farms (New York, USA) and were developed by Professor Alfred Schinkel and colleagues at the Netherlands Cancer Institute, Amsterdam (see [27,28] for the initial characterisation of these animals). Female mice were used to allow comparison of our data with a prior study that assessed the disposition of numerous CNS drugs in female WT and Abcb1a/b (2/2) mice [14]. Following shipment of founder mice to Australia from the US all mice genotypes were bred in-house and the offspring were used for experimentation following rearing under identical conditions. Mice were 2-3 months old for our first study assessing the brain and blood concentrations of THC and 4.5 months old for the second study examining THC-induced hypothermia. All mice were group-housed with 4-6 mice per cage in a temperature and humidity controlled room and maintained on a reverse 12-h light/12-h dark cycle. Experimental sessions took place during the dark cycle. Mice had ad libitum access to water and rodent lab chow.
Drug treatment D 9 -Tetrahydrocannabinol (THC Pharm GmbH, Frankfurt) was dissolved in absolute ethanol before being added to an equal amount of Tween 80 and diluted in 0.9% saline to give a final stock of ethanol/Tween 80/saline (1:1:18). All drugs were freshly prepared before use at an injection volume of 1 ml/kg. THC was administered intraperitoneally (i.p.) at a dose of 3 mg/kg for the measurement of THC concentration in brain and blood. At 0.5, 1, 2 and 3 h post-THC injections all mice (Abcb1a/b (2/2), Abcg2 (2/2) and WT animals, n = 5-6 per group) were anaesthetised with isoflurane and blood was collected via cardiac puncture before brains were extracted and snap frozen in liquid nitrogen. To avoid coagulation blood was stored in EDTA tubes. Both brain and blood were stored at 220uC prior to gas chromatography mass spectrometry (GC-MS) analysis. For the body temperature study mice (n = 6-7 per group) were first tested for baseline rectal body temperature as previously described (Boucher et al, 2011b). All mice were then administered an i.p. injection of 10 mg/kg THC before rectal body temperature measurements at 30 min, 2 h and 3 h post-injection were taken. The higher 10 mg/kg dose of THC was administered in this study as, unlike 3 mg/kg THC, this dose significantly reduced body temperature in WT mice, allowing observation of any potentiation of this effect in Abcb1a/b (2/2) and Abcg2 (2/2) mice.

GC-MS analysis of THC in biological matrices
The methods used for determining brain and blood THC concentrations were as described previously [29,30,31]. Briefly for blood THC analysis, a total of 50 ml of (deuterated) D3-THC (0.25 mg/L) internal standard solution was added to each 0.5 ml sample of whole blood. Acetate buffer was added (pH 4.0) to THC and was extracted with 1-chlorobutane solvent. Calibration standards were prepared by spiking drug-free sheep blood at concentrations of 1.25-100 ng/ml for THC. The standards were vortexed and treated identically to other blood samples. Following extraction and complete drying under a nitrogen gas stream, the samples underwent derivatization of the polar functional groups (COOH, OH) with bis(trimethylsilyl)trifluroacetamide (BSTFA). The limit of quantification (LOQ) for the blood analysis was 1.25 ng/ml. For brain THC analysis, brains were dissected in half in the sagittal plane and each hemisphere was weighed. A single hemisphere was then homogenized in 0.9% saline at a 1:2 ratio (w/v). A total of 50 ml of (deuterated) D3-THC (0.25 mg/L) internal standard solution, 2 ml of 0.2 M sodium hydroxide (pH 13) and 3 ml of acetonitrile were added to each 0.5 ml sample of homogenized brain samples. Calibration standards were prepared by spiking drug-free mouse brains at concentrations of 5-200 ng/g for THC. They were then completely dried under nitrogen gas stream and underwent identical extraction and derivatization as blood samples, as described above (LOQ = 5 ng/ g for the brain analysis). All quantification was performed by GC-MS (Shimadzu 2010Plus system) using electron impact ionization in selective ion mode.