Cannabidiol changes P-gp and BCRP expression in trophoblast cell lines

Objectives. Marijuana is the most commonly used illicit drug during pregnancy. Due to high lipophilicity, cannabinoids can easily penetrate physiological barriers like the human placenta and jeopardize the developing fetus. We evaluated the impact of cannabidiol (CBD), a major non-psychoactive cannabinoid, on P-glycoprotein (P-gp) and Breast Cancer Resistance Protein (BCRP) expression, and P-gp function in a placental model, BeWo and Jar choriocarcinoma cell lines (using P-gp induced MCF7 cells (MCF7/P-gp) for comparison). Study design. Following the establishment of the basal expression of these transporters in the membrane fraction of all three cell lines, P-gp and BCRP protein and mRNA levels were determined following chronic (24–72 h) exposure to CBD, by Western Blot and qPCR. CBD impact on P-gp efflux function was examined by uptake of specific P-gp fluorescent substrates (calcein-AM, DiOC2(3) and rhodamine123(rh123)). Cyclosporine A (CsA) served as a positive control. Results. Chronic exposure to CBD resulted in significant changes in the protein and mRNA levels of both transporters. While P-gp was down-regulated, BCRP levels were up-regulated in the choriocarcinoma cell lines. CBD had a remarkably different influence on P-gp and BCRP expression in MCF7/P-gp cells, demonstrating that these are cell type specific effects. P-gp dependent efflux (of calcein, DiOC2(3) and rh123) was inhibited upon short-term exposure to CBD. Conclusions. Our study shows that CBD might alter P-gp and BCRP expression in the human placenta, and inhibit P-gp efflux function. We conclude that marijuana use during pregnancy may reduce placental protective functions and change its morphological and physiological characteristics.


INTRODUCTION
Until recently it was assumed that prenatal exposure to marijuana is frequently combined with other drugs (e.g., tobacco and alcohol), making cannabis effects difficult to isolate and assess (Kozer & Koren, 2001;Moore et al., 2010). However, the studies of Dekker et al. and Hayatbakhsh et al. suggested that cannabis use prior to, during the first trimester, or throughout gestation is associated with a higher risk for a low birth weight and neonatal length, as well as preterm labor (Dekker et al., 2012;Hayatbakhsh et al., 2012). Moreover, the presence of the endocannabinoid receptors CB1 and CB2 on placental syncytiotrophoblast (Habayeb et al., 2008a), along with marijuana being the most popular drug of abuse among pregnant population (Brown & Graves, 2013), raised the need to understand the effect of cannabinoids on the placenta.
Preliminary data from in vitro models indicate that cannabinoids may interact with human P-gp (ABCB1) and BCRP (ABCG2). Acute and long-term exposures to cannabinoids were shown to alter P-gp expression and function (Holland et al., 2006;Zhu et al., 2006). However, BCRP exposure to cannabinoids demonstrated functional inhibition only (Holland et al., 2007).
We have already assessed the interaction of CBD with placental BCRP, finding that in both, in vitro and ex vivo systems, CBD inhibited its efflux function (Feinshtein et al., 2013). These results together with other recent findings regarding the effect of cannabinoid on the ABC transporter led us to investigate whether CBD affects placental P-gp on functional and placental P-gp and BCRP on expressional levels (Holland et al., 2006;Zhu et al., 2006;Holland et al., 2007;Holland, Allen & Arnold, 2008).
In the present work the implications of CBD exposure on P-gp and BCRP expression and P-gp function is tested in a human trophoblast-like cell lines BeWo and Jar, as placental model.

Subcellular fractionation
Fractionation procedure was carried out as previously described (Golan, Schreiber & Avissar, 2009), for the isolation of the membrane fraction. The membrane fraction was diluted in sample buffer 1:3 (10% v/v glycerol, 20% v/v SDS 20%, 5% v/v β-mercaptoethanol, 0.05% w/v bromophenol blue, pH 6.8), boiled for 5 min at 95 • C (for BCRP determination) or incubated for 30 min at 37 • C (for P-gp determination) and frozen at −80 • C until assayed. Aliquots were taken for protein determination using the Lowry assay. Fractions purity was verified using specific markers: Na + /K + ATPase for membranes and NF/kB p65 for cytosol.

IMMUNOCYTOCHEMISTRY AND CONFOCAL MICROSCOPY
MCF7/P-gp, BeWo and Jar cells were seeded on cover slips and grown to 70% subconfluency. Cells were then fixed in PFA for 15 min and stained with mouse anti-human CD 243 (MDR-1) antibodies, diluted in PBS, containing 3% BSA, for 1 h at room temperature, followed by incubation with Alexa Fluor 488 goat anti-mouse antibody in PBS containing 3% BSA, for 1/2 h at room temperature. Cover-slips were washed 3 times and mounted onto glass slides with DAPI (4 ,6-diamidino-2-phenylindole)-containing fluorescent mounting medium (DAPIFluoromount-G; SouthernBiotech). Immunofluorescence was detected by an Olympus FV-1000 Spectral confocal laserscanning microscope with excitation at 488 nm and emission at 520 nm. Image analysis was performed using ImageJ v. 1.40C software.

FACS analysis of BCRP levels
Cells were grown to confluency in 6-well culture dishes, trypsinized, counted, and 0.5 × 10 6 cells were placed in light resistant vials. Cells were washed three times with cold PBS (and each time centrifuged at 1200 g for 5 min), and incubated for 30 min with BSA 3% in PBS. Following additional PBS wash, 50 µL of BSA 1% in PBS and antibodies were added (1:10 anti-BCRP FITC conjugated antibody or 1:10 CBL602F FTIC conjugated isotype negative control). Cells were incubated in the dark at 4 • C for 1 h, washed with ice cold PBS, re-suspended (in 0.5 ml PBS) and analyzed by FACS (BD FACS Vantage).

Immunoblotting
Membrane fractions were thawed on the day of assay. Protein aliquots (60-100 µg/lane) were taken for protein separation by SDS-PAGE as previously described (Golan, Schreiber & Avissar, 2009). Semi-quantitative analysis was carried out using a computerized image analysis system (EZQuant-Gel 2.11; EZQuant Biology Software Solutions Ltd., Israel). Equal protein loading was ensured by normalization to Na + /K + ATPase (for P-gp) or actin (for BCRP).

RNA extraction, reverse transcription, real-time polymerase chain reaction (qPCR)
Isolation and purification of total RNA from BeWo and Jar cells was carried out using EZ-RNA Kit (Biological Industries, Israel) according to manufacturer's instructions. 1 µg of total RNA was used for reverse transcription using High Capacity cDNA RT kit (Applied Biosystems, Foster City, CA), in 20 µl reaction volume. BCRP and P-gp mRNA was measured by qRT-PCR, as indicated in the manufacturer protocol (Applied Biosystems, Foster City, CA), and performed by the Applied Biosystems Real Time PCR system (7500 system), using TaqMan probes and primers for human BCRP, P-gp and actin (Applied Biosystems, Foster City, CA). The cycling conditions for all primers were as follows: hold for 10 min at 95 • C, followed by 40 cycles consisting of two steps, 15 s at 95 • C (denaturing), and 1 min at 60 • C (annealing-extension). The threshold cycle, which correlates inversely with the mRNA levels of target, was measured as the cycle number at which the reporter fluorescent emission increases above a threshold level. P-gp and BCRP mRNA levels were normalized to actin mRNA in the same samples. Results were analyzed by the 2 − − C T method, demonstrating the relative changes in gene expression from real-time quantitative PCR experiments, using 7500 Software v2.0.4 (Applied Biosystems, Foster City, CA).

Statistical analysis
All statistics and graphs were carried out using GraphPad Prism5 software. Student's t-test or one-way ANOVA followed by appropriate Bonferroni corrections were used.

CBD impact on BCRP and P-gp protein expression
The changes in P-gp and BCRP protein levels in the membrane fraction of BeWo, Jar and MCF7/Pgp cells were studied. Following determination of proper cell fractionation (Fig. 1A), the basal expression of these two transporters in non-treated cells was determined by Western Blot analysis (Fig. 1B). Due to unexpected P-gp expression profile in all three cell lines we verified its basal expression by immunocytochemical fluorescent staining (Fig. 1C). It can be seen that our BeWo, Jar and MCF7/P-gp cells express detectable levels of P-gp. For P-gp expression, the results of the Western Blot were confirmed by immunocytochemistry showing that P-gp expression was the highest in MCF7/P-gp and the lowest in Jar cells. For BCRP (ABCG2), results were confirmed by FACS analysis, showing that BCRP expression was much higher in JAR cells compared to BeWo cells (in full accordance with Western Blot analysis) (Fig. 1D).
The changes in membrane BCRP and P-gp levels in BeWo cells following long-term exposure to CBD are displayed in Fig. 2. Indeed, BCRP levels were significantly increased in a concentration-dependent manner (Fig. 2B) that was not time-dependent ( Fig. 2A). Following long-term exposure to CBD, P-gp protein levels significantly decreased, in a concentration-dependent manner (Fig. 2D), with no time-dependent effect (Fig. 2C). Similarly, in Jar cells (Fig. 3), BCRP protein concentrations was significantly elevated following long term exposure to CBD, in a concentration but not time-dependent effect (Figs. 3A and 3B). Due to very low P-gp initial (baseline) expression in the Jar cell line (Figs. 1B and 1C), the expected down-regulation in P-gp expression following long-term exposure to CBD is demonstrated in experiments of 72 h treatment only. P-gp is vaguely visualized in the control group while it is almost undetectable in the CBD treated group (Fig. 3C). 48 h exposure yielded the same results (P-gp levels were almost undetectabledata not shown), making time and concentration dependent comparisons impractical.
To examine whether CBD effects are cell specific, the same experimental routine was applied to MCF/P-gp cells. In these cells, BCRP and P-gp behavior under CBD  Data is displayed in means ± s.d. of at least three (n = 9) independent experiments of each concentration or time point. One representative blot is presented for each experimental group. One-Way ANOVA, followed by Bonferroni's multiple comparison test, * * p < 0.01, * * * p < 0.0001 compared to control or comparison between groups as indicated.
influence was profoundly different from that seen in choriocarcinoma cell lines. P-gp levels dramatically increased following long-term treatment with CBD, in a concentration-but not time-dependent manner (Fig. 4A). At the same time, BCRP expression was not affected by CBD (neither time-, nor concentration-dependent effect was observed) (Fig. 4B). Thus, we conclude that CBD has a cell-type specific influence upon long-term cellular exposure. CBD impact on MCF7/P-gp cells was not investigated further.

Figure 3 Long-term exposure of Jar cells to CBD: changes in P-gp and BCRP protein levels. Long-term exposure of Jar cells to CBD: changes in membrane BCRP and P-gp expressed levels. Changes in BCRP-CBD concentration-dependent (B) and time-dependent (A). Changes in P-gp (C).
Data is displayed in means ± s.d. of at least three (n = 12) independent experiments of each concentration or time point. One representative blot is presented for each experimental group. One-Way ANOVA, followed by Bonferroni's multiple comparison test (Student's t test for (C)), * p < 0.05, * * * p < 0.0001 compared to control or comparison between groups as indicated.

CBD impact on BCRP and P-gp mRNA
To have a glimpse into the mechanism underlying the changes seen on the protein level in BCRP and P-gp, we further focused to track the changes that occur on mRNA level of these two transporters. In both, BeWo and Jar cells, BCRP mRNA quantification showed elevation following long-term exposure to CBD (Figs. 5A and 5C), and matched the findings on protein level, supporting transcriptional up-regulation. Likewise, P-gp mRNA expression significantly dropped following long-term treatment with CBD, matching the outcomes seen on the protein level (Figs. 5B and 5D), indicating that the changes in P-gp levels following long-term CBD exposure results from transcriptional down-regulation rather than post-transcriptional changes.

CBD impact on P-gp function
Only BeWo and MCF7/P-gp cell lines were tested for P-gp inhibition by CBD due to the fact that Jar cells expressed very low P-gp levels. In a preliminary study we observed a non-cell type specific P-gp inhibition by 25 µM CBD in both cell lines, as intracellular calcein and rh123 fluorescence was significantly higher (elevation of 101 ± 50% and 70 ± 30%, respectively) in the presence of CBD (Figs. 6A and 6B). We further examined whether lower concentration of CBD 10 µM also inhibits P-gp, using two different P-gp specific substrates, calcein-AM and DiOC2(3). P-gp was inhibited following short-term (1 h) exposure to CBD, as significantly more DiOC2 (3) and calcein (elevation of 20 ± 8% and 24 ± 9%, respectively) accumulated in the cells (Figs. 7A and 7B).

Figure 4 Long-term exposure of MCF/P-gp cells to CBD: changes in P-gp and BCRP protein levels.
Long-term exposure of MCF/P-gp cells to CBD: changes in membrane P-gp (A) and BCRP (B) expressed levels. Protein levels are given as percent of control levels. Data is displayed in means ± s.d. of at least three (n = 9) independent experiments of each concentration or time point. One representative blot is presented for each experimental group. One-Way ANOVA, followed by Bonferroni's multiple comparison test. * * p < 0.01, * * * p < 0.0001 compared to control or comparison between groups as indicated.

Principal findings of the study
There is a dual effect of CBD on the expression of BCRP and P-gp transporters in trophoblast-like and in MCF7/P-gp cell lines. Under long-term CBD exposure BCRP and P-gp present cell-type specific changes in protein and mRNA levels. Occurring already at the transcriptional level, P-gp protein expression is down-regulated, while that of BCRP is up-regulated. Upon short-term exposure, P-gp efflux function is inhibited by CBD.

Cannabinoid effects on pregnancy
Cannabis is extensively used in Western society as a recreational drug. The effect of this drug on pregnancy outcome was under constant debate; however, recently cannabis consumption was reported to be associated with adverse pregnancy outcomes, including preterm birth and fetal growth restriction (Dekker et al., 2012;Hayatbakhsh et al., 2012). Although the mechanisms in which phytocannabinoids exert their effects are not well understood, there is accumulating evidence that endocannabinoids (like anandamide) can influence reproduction, including fertilization, implantation, angiogenesis, embryo development and placental growth (Taylor et al., 2007;Habayeb et al., 2008a;Lewis et al., 2012;Solinas et al., 2012;Sun & Dey, 2012). One possible route to convey the effect of endocannabinoids is through the CB1 and CB2 receptors that are expressed by the human Data is displayed as means ± s.d. of at least three (n = 6) independent experiments for each time point. One-Way ANOVA, followed by Bonferroni's multiple comparison test. * p < 0.05, * * p < 0.01, * * * p < 0.0001 compared to control.  Data presented from at least 2 independent experiments (n = 9, n = 6, respectively), as means ± s.d. Statistical significance determined by One-Way ANOVA, followed by Bonferroni's multiple comparison test for (A) and student's t-test for (B). * * * p < 0.0001, * p < 0.05.
trophoblast and provide a direct target for cannabinoids (Kenney et al., 1999;Habayeb et al., 2008a;Habayeb et al., 2008b). However, not all endocannabinoids exert their effect through the classic cannabinoids receptors. For example, CBD lacks the central effects of cannabis and works through CB1 and CB2 independent model of action (Scuderi et al., 2009). Since our goal was to test the direct cannabinoid effect on trophoblast transporters BCRP and P-gp, and CBD showed the most potent inhibitory effect (among major cannabinoids) on these transporters (Holland et al., 2006;Zhu et al., 2006;Holland et al., 2007) it made CBD the ideal cannabinoid for the present study.

BCRP and P-gp significance in pregnancy
BCRP and P-gp play a key role in the transport of drugs and endogenous compounds in the human placenta, affecting the outcome of pregnancy (Robey et al., 2009;Myllynen, Kummu & Sieppi, 2010). They transport a broad variety of structurally diverse compounds, some of which are congruent (Frohlich et al., 2004;Mathias, Hitti & Unadkat, 2005;Zhou, 2008). In addition, BCRP transports a wide range of substrates, including fetal hormonal precursors such as estrone-3-sulfate, naturally occurring carcinogens, porphyries and ceramides (Imai et al., 2003;Krishnamurthy & Schuetz, 2005;Evseenko et al., 2007;Evseenko, Paxton & Keelan, 2007;Mao, 2008;Dietrich et al., 2011). P-gp is expressed in the apical membrane of syncytiotrophoblast and probably is the main placental protective transporter during the first trimester. Its expression (both mRNA and protein) is the highest during the first trimester and decreases with advancing gestation (Gil et al., 2005;Mathias, Hitti & Unadkat, 2005;Sun et al., 2006). BCRP plays an important role as a survival factor in BeWo cells as well as in the human placenta. It is thought to have a protective antiapoptotic role in the trophoblast, regulating their survival under low oxygen conditions (Krishnamurthy & Schuetz, 2006;Yeboah et al., 2006;Evseenko et al., 2007;Evseenko, Paxton & Keelan, 2007).
The trophoblast expression of BCRP may change in different pregnancy complications. Indeed, placentas of women with preterm labor and intra-amniotic inflammation had higher expressions of this transporter than those of women with preterm labor without inflammation. In addition, the mRNA expression of BCRP correlated with that of IL-8, which also increased significantly in placentas of women with preterm labor and inflammation, suggesting that the transfer of drugs across the placenta may be altered in cases of preterm labor with inflammation (Mason et al., 2011).

CBD impact on BCRP and P-gp upon long-term exposure
In the current study we have demonstrated for the first time that the exposure of trophoblast-like cell lines BeWo and Jar to CBD is associated with two distinct patterns of effects on the expression of the BCRP and P-gp transporters. The use of these cell lines, instead of primary trophoblast cultures, results from the fact that primary trophoblasts may rapidly differentiate in culture, continuously changing their gene and protein expression (Evseenko, Paxton & Keelan, 2006). Comparison of these parameters with regard to CBD influences in such a dynamic in vitro environment would be almost impossible. Moreover, BeWo and Jar cell lines, derived from human gestational choriocarcinoma, commonly used as an in vitro model for trophoblast toxicology studies, and they offer a suitable model to study certain aspects of human trophoblast physiology, without the aspect of inter-patient variability (Sullivan, 2004;Khare et al., 2006;Myren et al., 2007).
The first effect that was observed following long-term exposure is that CBD transcriptionally inhibited P-gp membrane expression (as mRNA levels of P-gp were also dramatically reduced following long-term exposure to CBD). The long-term inhibitory effect of CBD on P-gp protein was previously reported in drug-selected human T lymphoblastoid leukaemia cell line (CEM/VLB(100)) (Holland et al., 2006). The cell specificity of CBD was previously reported (Khare et al., 2006;Hu, Ren & Shi, 2011), yet our study is the first to detect this phenomenon in trophoblast-like cell lines. This observation may have clinical implications, in light of the role of P-gp as one of the key transport mechanisms for numerous drugs in the human placenta, especially during the venerable period of the first trimester in which all the fetal organs are formed. Thus, our results might be especially important to women who use cannabis on a regular basis during the first trimester and are treated with other drugs that are P-gp substrates.
The expressional up-regulation of BCRP mRNA and protein in trophoblast-like cell lines following long-term exposure to CBD is the second novel effect reported in our study. This observation is similar to that reported in trophoblasts of women with preterm labor and inflammation, suggesting that (I) long-term exposure to CBD may elicit an inflammatory response in trophoblast-like cell lines (Mason et al., 2011); and (II) that this phenomenon could be a direct compensational consequence of P-gp down-regulation, providing a working defense line to the developing embryo. Interestingly, compensation of such nature was already described in murine placentas (Hutson, Koren & Matthews, 2010). Nonetheless, the signal transduction of CBD action in human placenta/trophoblast/choriocarcinoma cells has not been clearly reported yet and needs to be further elucidated.

CBD impact on P-gp upon acute exposure
In agreement with previously published data, we found that CBD holds inhibiting properties over P-gp efflux function (Holland et al., 2006;Nieri et al., 2006;Zhu et al., 2006). Moreover, our results show inhibition of P-gp efflux function in cell lines that naturally express this transporter. It can be seen that CBD 10 µM (Fig. 7) yielded inhibition effects lower than CBD 25 µM (Fig. 6). Similar to results we recently presented (Feinshtein et al., 2013), this observation could imply that CBD inhibits P-gp in a concentration dependent fashion. However, due to different quantification methods used in the present study, this should be further elucidated.
Of note, many of the drugs that are prescribed and considered safe to use during pregnancy are in fact P-gp or BCRP substrates, like Loratadine and H 2 blockers (i.e., Ranitidine and Cimetidine) (Collett et al., 1999;Chen et al., 2003;Li et al., 2008;Schwarz et al., 2008;Dahan & Amidon, 2009;Gill, O'Brien & Koren, 2009;Matok et al., 2010). Our finding may have clinical implications, suggesting that the use of cannabis during gestation may alter drug transport through the trophoblast and lead to the absence of a functional placental barrier during the first trimester, leaving the developing embryo unprotected at this vulnerable period of pregnancy. Moreover, the trophoblasts of pregnant women exposed to marijuana may exhibit some resistance to apoptotic and inflammatory processes due to the effect of CBD on BCRP expression.

CONCLUSIONS
Following cannabis consumption, all the drugs that are P-gp substrates can potentially penetrate the human placental barrier at higher rates when combined with CBD, and therefore their safety under these conditions is to be questioned. Additionally, changes in placental BCRP expression profile might lead to altered transplacental transport of BCRP substrates, such as medications, naturally occurring carcinogens, hormonal precursors and apoptotic molecules, and influence pregnancy outcomes.