SEARCH

SEARCH BY CITATION

Keywords:

  • cannabinoid;
  • endocannabinoid;
  • lipid raft;
  • membrane raft;
  • microglia;
  • Caveolin-1;
  • flotillin-1;
  • cannabidiol;
  • CB1 receptor;
  • mass spectrometry

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

BACKGROUND AND PURPOSE N-acyl ethanolamines (NAEs) and 2-arachidonoyl glycerol (2-AG) are endogenous cannabinoids and along with related lipids are synthesized on demand from membrane phospholipids. Here, we have studied the compartmentalization of NAEs and 2-AG into lipid raft fractions isolated from the caveolin-1-lacking microglial cell line BV-2, following vehicle or cannabidiol (CBD) treatment. Results were compared with those from the caveolin-1-positive F-11 cell line.

EXPERIMENTAL APPROACH BV-2 cells were incubated with CBD or vehicle. Cells were fractionated using a detergent-free continuous OptiPrep density gradient. Lipids in fractions were quantified using HPLC/MS/MS. Proteins were measured using Western blot.

KEY RESULTS BV-2 cells were devoid of caveolin-1. Lipid rafts were isolated from BV-2 cells as confirmed by co-localization with flotillin-1 and sphingomyelin. Small amounts of cannabinoid CB1 receptors were found in lipid raft fractions. After incubation with CBD, levels and distribution in lipid rafts of 2-AG, N-arachidonoyl ethanolamine (AEA), and N-oleoyl ethanolamine (OEA) were not changed. Conversely, the levels of the saturated N-stearoyl ethanolamine (SEA) and N-palmitoyl ethanolamine (PEA) were elevated in lipid raft fractions. In whole cells with growth medium, CBD treatment increased AEA and OEA time-dependently, while levels of 2-AG, PEA and SEA did not change.

CONCLUSIONS AND IMPLICATIONS Whereas levels of 2-AG were not affected by CBD treatment, the distribution and levels of NAEs showed significant changes. Among the NAEs, the degree of acyl chain saturation predicted the compartmentalization after CBD treatment suggesting a shift in cell signalling activity.

LINKED ARTICLES This article is part of a themed section on Cannabinoids in Biology and Medicine. To view the other articles in this section visit http://dx.doi.org/10.1111/bph.2012.165.issue-8. To view Part I of Cannabinoids in Biology and Medicine visit http://dx.doi.org/10.1111/bph.2011.163.issue-7


Abbreviations
2-AG

2-arachidonoyl glycerol

ABHD6

α-β-hydrolase domain 6

AEA

N-arachidonoyl ethanolamine

CBD

cannabidiol

CB1

cannabinoid receptor 1

CB2

cannabinoid receptor 2

NAPE-PLD

N-acyl phosphatidylethanolamine phospholipase D

OEA

N-oleoyl ethanolamine

PEA

N-palmitoyl ethanolamine

PPARα

peroxisome proliferator-activated receptor α

SEA

N-stearoyl ethanolamine

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

Membrane rafts are defined as small (10–200 nm), heterogeneous, dynamic, cholesterol and sphingolipid-rich domains. The term membrane raft (interchangeable with lipid raft) is used to convey the importance of both lipids and proteins in membrane raft formation and function (Pike, 2006). Lipid rafts are involved in many functions including intracellular signalling, cellular polarity, molecular sorting, membrane transport and endocytosis (Moffett et al., 2000; Gaus et al., 2003; Chini and Parenti, 2004; Wilson et al., 2004; Lajoie and Nabi, 2010). Caveolae are a subtype of membrane/lipid rafts with electron microscopically visible plasma membrane invaginations. The protein caveolin-1 is a necessary (but not sufficient) component for the formation of caveolae in non-muscle tissues (see Lajoie and Nabi, 2010). Caveolar membrane rafts are involved in lipid molecule endocytosis, transcytosis, transport, trafficking and efflux (Fielding and Fielding, 1995; 1997; Czarny et al., 1999; Sharma et al., 2003; Luo et al., 2010). Other membrane/lipid raft subtypes are less well characterized and include non-caveolar lipid rafts of different types (Lajoie and Nabi, 2010).

Lipid raft/caveolae have been proposed to compartmentalize the endocannabinoid signalling machinery in several cellular systems (Keren and Sarne, 2003; McFarland et al., 2004; 2008Bari et al., 2005a,b; 2006; 2008; McFarland and Barker, 2005; Sarnataro et al., 2005, 2006; Oddi et al., 2007; Placzek et al., 2008; Rimmerman et al., 2008; Maccarrone et al., 2009). Specifically, several lines of evidence support an association of the cannabinoid CB1 receptor (nomenclature follows Alexander et al., 2011) with lipid raft/caveolae including: (i) CB1 receptor C-terminal acylation domain is required for proper interactions with lipid raft-associated G proteins (Mukhopadhyay et al., 1999; Barnett-Norris et al., 2005; Fay et al., 2005; Xie and Chen, 2005); (ii) CB1 receptor internalization in human embryonic kidney 293 cells-CB1 transfected cells that occurs via both caveolae and clathrin-coated pits (Keren and Sarne, 2003); (iii) increased CB1 receptor binding and signalling following cholesterol depletion in C6 glioma cells (Bari et al., 2005a,b; 2006); (iv) CB1 receptor association with lipid raft fractions/ non-lipid raft fractions in MDA-MB-231 breast cancer cells which depends on receptor activation/antagonism (Sarnataro et al., 2005; 2006); and (v) CB1 receptor localization within lipid rafts in human endothelial cells, and CB1 receptor co-localization with caveolin-1 in C6 glioma cells (Bari et al., 2006; 2008).

N-arachidonoyl glycerol (2-AG) and N-arachidonoyl ethanolamine (AEA), two major endocannabinoids that interact with CB1 and CB2 receptors have been characterized in lipid rafts (Rimmerman et al., 2008). We previously investigated their compartmentalization in caveolin-1-expressing F-11 cells (a dorsal root ganglion-like cell line). We showed that membrane/lipid rafts (rich in caveolin-1, flotillin-1 and lipid markers) compartmentalize the biochemical machinery for the production of 2-AG that includes the 2-AG precursor arachidonoyl-containing diacyl glycerol, and the synthetic enzyme diacyl glycerol lipase α (DGLα; Rimmerman et al., 2008). Additionally, endogenous AEA and one of its biosynthetic enzymes, N-acyl phosphatidylethanolamine phospholipase D (NAPE-PLD) were found in lipid raft as well as in non-raft fractions (Rimmerman et al., 2008). A similar trend was observed following incubation with exogenous deuterium labelled AEA [(2H8)-AEA]. On the other hand, deuterium labelled arachidonic acid, a direct metabolite of [2H8]-AEA, was localized mostly to non-lipid raft fractions consistent with the theory of caveolar-mediated endocytosis of AEA (McFarland et al., 2004; McFarland and Barker, 2005).

AEA belongs to a larger family of N-acyl ethanolamines (NAE) such as the monounsaturated [(18:1)-containing acyl chain]N-oleoyl ethanolamine (OEA), and the saturated [(18:0) and (16:0)-containing acyl chain], N-stearoyl ethanolamine (SEA) and N-palmitoyl ethanolamine (PEA) respectively. NAEs were shown to increase during brain disease, and to produce anti-inflammatory effects through different mechanisms (see Franklin et al., 2003; Hansen, 2010). OEA was shown to interact with the nuclear receptor peroxisome proliferator-activated receptor α (PPARα) and the orphan G protein coupled receptor, GPR119. PEA was shown to interact with PPARα, GPR119 and GPR55 (Fu et al., 2005; Lo Verme et al., 2005a,b; Overton et al., 2006, 2008; O'Sullivan et al., 2007; Godlewski et al., 2009). SEA interacts with a yet unidentified target to modulate cellular signalling (Maccarrone et al., 2002; Hansen, 2010). The compartmentalization of OEA and SEA in lipid rafts has not yet been investigated.

The mechanisms leading to 2-AG and NAE metabolism differ between cell types, reflecting differential expression of gene products, enzymes, lipid membrane composition and trafficking mechanisms. In this study, we characterize the membrane compartmentalization of 2-AG and NAEs into lipid rafts in the BV-2 microglial cell line, which we found to be devoid of caveolin-1. BV-2 microglial cells express CB1 and CB2 receptors, the lysophosphatidylinositol/cannabinoid receptor GPR55 and the putative abnormal-cannabidiol receptor GPR18 (Pietr et al., 2009; McHugh et al., 2010; Stella, 2010). In addition, they express the enzyme α-β-hydrolase domain 6 (ABHD6) that controls the accumulation of 2-AG and the efficacy of this compound at cannabinoid receptors (Marrs et al., 2010). Furthermore, PEA was shown to be metabolized via a URB602-sensitive enzyme of yet unknown identity, and AEA through fatty acid amide hydrolase (FAAH) in BV-2 cells (Muccioli et al., 2007; Muccioli and Stella, 2008).

Our group recently reported that the non-psychoactive plant cannabinoid, cannabidiol (CBD) inhibits pro-inflammatory pathways in lipopolysaccharide-activated BV-2 cells (Kozela et al., 2010). In addition, our gene array analysis studies indicate that CBD exerts immunosuppressive effects by regulating stress response genes (Juknat et al., 2012). Here, we compared the compartmentalization of 2-AG and NAEs in lipid raft fractions, as well as their levels in whole cells with growth medium, between vehicle and CBD-treated cells. We show that BV-2 cells do not express caveolin-1 mRNA or protein, and thus, are devoid of the caveolar-membrane/lipid raft subtype. The compartmentalization of 2-AG in these cells was generally similar to F-11 cells (which express caveolin-1); however, while the membrane/lipid rafts in F-11 cells were well separated from the non-raft fractions by their density profile (Rimmerman et al., 2008), BV-2 lipid rafts and non-lipid raft fractions showed a much closer density. The CB1 receptor is localized mostly to non- lipid raft fractions with small amounts present in lipid raft fractions. In addition, we show that following CBD treatment, the levels of 2-AG were not affected in membrane/lipid raft fractions or whole cells with growth medium. The levels of AEA and OEA increased significantly in whole cells with growth medium, while the levels of PEA and SEA increased significantly in membrane/lipid rafts fractions.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

Cell culture

The BV-2 murine microglial cell line, originally generated by E. Blasi (University of Perugia, Perugia, Italy; Blasi et al., 1990), was kindly provided by Prof E.J. Choi from the Korea University (Seoul, Korea). Cells were grown under 5% CO2 at 37°C in Dulbecco's modified Eagle's medium containing high glucose (Gibco-BRL, Gaithersburg, MD, USA) supplemented with 5% heat-inactivated foetal bovine serum, streptomycin (100 µg·mL−1) and penicillin (100 U·mL−1) (Biological Industries Ltd, Kibbutz Beit Haemek, Israel). Cells were used up to passage 25.

Lipid analysis from whole BV-2 cells and media

BV-2 cells were grown for 24 h on 10 cm plates in growth medium. Media was replaced with 5 mL fresh growth medium 1 h before treatment. BV-2 cells were then treated with either vehicle (ethanol 0.1%) or CBD (10 µM). Cells together with media were scraped into 7 mL 100% HPLC-grade methanol. Whole cells with growth medium were collected from independent plates at four time points (10, 30, 60 and 210 min; n= 3 per condition; except for vehicle at 210 min, n= 2). The solution was then vortexed and centrifuged at 200× g for 5 min. Internal standards were added to supernatant samples and HPLC-grade water was added to make a 75% aqueous solution. Lipids were then extracted as previously described (Bradshaw et al., 2006). Briefly, 500 mg C8 Bond Elut solid phase extraction columns (Varian, Harbor city, CA, USA) were conditioned with 5 mL HPLC-grade methanol followed by 3.0 mL HPLC-grade water. The 75% aqueous solutions were loaded onto separate columns, which were then washed with 3 mL water. Two sequential elutions (1.5 mL 50%, and 2 mL 100% methanol) were collected for mass spectrometric analysis. Samples were dried down by speed vac and reconstituted in 400 µL 100% HPLC-grade methanol for lipid analysis.

Membrane fractionation

Procedures for fractionation were adapted from the method of Macdonald and Pike (2005). Cells were plated 24 h before the experiment. On the day of the experiment, cells from four 150 mm plates for each experimental condition were washed once with PBS and scraped into a total of 25 mL of 4°C base buffer (20 mM Tris-HCl, 250 mM sucrose, pH 7.8) containing 1 mM calcium chloride and 1 mM magnesium chloride. Cells were centrifuged in a polystyrene tube at 200× g for 7 min at 4°C. The pellet was re-suspended in 1.0 mL of base buffer containing calcium and magnesium, and protease inhibitors were added (1 mM phenylmethylsulphonyl fluoride, 20 µg·mL−1 leupeptin, 1 µg·mL−1 aprotinin and 2 µM pepstatin). Cells were then lysed by passage through a 27 g × 3″ needle, 20 times, and centrifuged at 1000× g for 10 min at 4°C. The post-nuclear supernatant was transferred to a new tube, the pellet was re-suspended in base buffer containing calcium chloride, magnesium chloride and protease inhibitors, and the procedure was repeated. The 2 mL post-nuclear supernatants were combined and 50% OptiPrep (Axis Shield, Dundee, UK) in base buffer was added to give a 4 mL solution of 25% OptiPrep. An 8 mL continuous density gradient of 20%–0% OptiPrep in base buffer was formed on top of the 4 mL 25% OptiPrep solution in Ultra-ClearTM (14 × 89 mm) centrifuge tubes. The gradient was centrifuged using an SW-41 swinging bucket rotor (Beckman Coulter, Fullerton, CA, USA) for 90 min at 52 000× g at 4°C. 16 fractions of 0.75 mL were collected from each gradient.

To confirm our results, we used a second fractionation method according to Ostrom et al. (2001) which was modified from the original method of Song et al. (1996). Briefly, cells were scraped into 500 mM sodium bicarbonate, pH 11 (including protease inhibitors). Cells were then homogenized 10 times with a Dounce homogenizer, then were polytroned 10 s × 3 times, and sonicated 20 s × 3 times. The 2 mL homogenate was mixed with 2 mL of 90% sucrose (containing MES 25 mM, and NaCl 150 mM) to a final concentration of 45% sucrose. The 4 mL homogenate was placed in the bottom of an Ultra-ClearTM (14 × 89 mm) centrifuge tube. On top of the homogenate were placed 4 mL 35% sucrose (containing MES 25 mM, NaCl 150 mM, Na2CO3 250 mM), and 4 mL of 5% sucrose in the same buffer to form a discontinuous gradient. The gradient was centrifuged using an SW-41 swinging bucket rotor for 20 h at 160 000× g at 4°C. Eleven fractions of 1 mL were collected from each gradient.

Mass spectrometric analysis of lipids in fractions

For lipid analysis, 0.6 mL of each fraction was removed and 2 mL of HPLC-grade methanol were added. [2H8]-AEA (200 pmol) was added to each sample and diluted with HPLC-grade water to make a 75% aqueous solution. Lipids were extracted as previously described (Bradshaw et al., 2006). Briefly, 500 mg C8 Bond Elut solid phase extraction columns (Varian) were conditioned with 5 mL HPLC-grade methanol, followed by 3.0 mL HPLC water. The 75% aqueous solutions containing the fractions were loaded onto separate columns, which were then washed with 20 mL water. Five sequential elutions (1.5 mL each of 30, 50, 85 and 100% methanol) were collected for mass spectrometric analysis. As described previously (Bradshaw et al., 2006), sample analysis of lipids was carried out as follows. An aliquot of each of the eluates was loaded using a Shimadzu SCL10Avp (Wilmington, DE, USA) or an LC Packings Ultimate-3000 (Sunnyvale, CA, USA) autosampler onto a reversed phase Zorbax 2.1 × 50 mm C8 column maintained at 40°C. HPLC gradient formation at a flow rate of 200 µL·min−1 was achieved by a system comprised of a Shimadzu controller and two Shimadzu LC10ADvp pumps or an LC Packings controller and an LPG-3000 loading pump. Lipid levels in the samples were analysed in multiple reaction monitoring (MRM) mode on a triple quadrupole mass spectrometer, using either the API 3000 or the API 4000 (Applied Biosystems/MDS SCIEX, Foster city, CA, USA), with electrospray ionization. Methods for lipid analysis were created and optimized by flow injection of lipid standards. All calculations for quantitation experiments were based on calibration curves using synthetic standards. The following molecular ion and fragment ion pairs were used to quantify lipids in MRM mode: (18:0)-sphingomyelin 731.7/184.1; 2-AG 379.3/287.3; AEA 348.2/62.1; PEA 300.2/62.1; SEA 328.2/62.1; OEA 326.2/62.1; CBD 315.2/193.1.

Western blot analysis

Membrane components present in gradient fractions were lysed by the addition of sodium dodecyl sulphate (SDS) (Sigma-Aldrich), NP-40 (EMD Biosciences Inc., La Jolla, CA, USA) and Triton X-100 (Sigma-Aldrich) to a final concentration of 0.5, 1 and 1% respectively. Samples were incubated on ice for 30 min. Laemmli Loading buffer was added to the samples and they were denatured for 3 min at 70°C using a hot water bath. Aliquots of 20 µL from each sample were separated by an 8% or 10% SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes. The membranes were blocked in blocking buffer consisting of TBST [10 mM Tris-HCl (pH 7.5), 150 mM NaCl and 0.1% Tween-20] with 5% w/v skim milk. The membranes were washed five times with TBST and incubated overnight at 4°C in the presence of primary antibody or antibody + blocking peptide in 3% BSA. The following affinity-purified antibodies were used for the Western blot analyses: anti-flotillin-1 (1:500 dilution), anti-caveolin-1 (1:500 dilution), anti-transient receptor potential channel V2 (TRPV2; 1:500 dilution), anti-CB2 receptor (1:500 dilution), anti-CB1 receptor (1: 500 dilution), anti- β-actin (1:1000–2000 dilution) and anti-lysosomal-associated membrane protein 1 (LAMP1) (1: 1000 dilution). The membranes were washed five times with TBST, and then incubated for 45–60 min with either HRP-goat anti-rabbit IgG, HRP-goat anti-mouse IgG (1:10 000 dilution) or HRP-rat anti-mouse IgG (1:10 000) in 5% milk (or 3% BSA for the CB1 receptor primary antibody). The blots were washed five times and processed using chemiluminescent detection. The detection of CB1 receptors needed a much longer film exposure compared with CB2 receptors, due to lower levels of the CB1 receptor protein.

Isolation of total RNA, reverse transcription and real-time quantitative PCR (qPCR)

Primers for mouse caveolin-1 (NM_007616.3) and for mouse caveolin-2 (NM_016900.3) were designed as assay on demand by Qiagen (Hilden, Germany). The following primers were used for mouse caveolin-1, forward AGCTCACATTACAGCTCTGCCCTT, reverse AGTGTCGGCAAGACTGAAGGAGAA; for caveolin-2, forward TTGCGGGTATCCTGTTTGCT, reverse AGTTGCATGCTGACCGATGA; and for mouse β-microglobulin, forward ATGGGAAGCCGAACATACTG, reverse CAGTCTCAGTGGGGGTGAAT. RNA was extracted using the Versagene RNA purification kit (Gentra, Minneapolis, MN, USA), and RNA samples (2 µg) were reverse transcribed using the QuantiTect Reverse Transcription Kit from Qiagen including DNase treatment of contaminating genomic DNA. Expression of caveolin-1 and caveolin-2 mRNAs were determined by qPCR, using β-microglobulin as a normalizing gene, as previously described (Butovsky et al., 2006). Normal and mock reversed transcribed samples (in the absence of reverse transcriptase), as well as no template controls (total mix without cDNA), were run for each of the examined mRNA's. qPCR reactions were subjected to an initial step of 15 min at 95°C to activate the HotStar Taq DNA polymerase, followed by 40 cycles consisting of 15 s at 94°C, 30 s at 60°C and 30 s at 72°C. Fluorescence was measured at the end of each elongation step. Data were analysed using the Rotor-Gene software (Corbett Life Sciences, Mortlake, Australia) and a threshold cycle value Ct was calculated from the exponential phase of each PCR sample. Amounts of mRNAs were calculated and expressed in relative units of SYBR Green fluorescence.

Data analysis

Data are shown as means ± SEM. Differences in lipid levels were assessed by one-way analysis of variance (ANOVA) with Fisher's least significant difference (LSD) post hoc test or Bonferroni post hoc test (SPSS, Chicago, IL, USA), as specified in the Figure legends. Differences were considered significant for P < 0.05.

Materials

[2H8]-N-arachidonoyl ethanolamine [(2H8)-AEA], AEA, PEA, SEA and 2-AG were purchased from Cayman Chemical (Ann Arbor, MI, USA). CBD was received from NIDA. Sphingomyelin standards were purchased from Avanti Polar Lipids, Inc (Alabaster, AL, USA). High-performance liquid chromatography (HPLC) grade methanol, acetonitrile and isopropyl alcohol were purchased from VWR international (Plainview, NY, USA) and HPLC-grade water was obtained using a MilliQ Gradient apparatus from Millipore (Milford, MA, USA). HPLC-grade acetic acid and ammonium acetate were purchased from Sigma-Aldrich (St. Louis, MO, USA).

The affinity purified rabbit polyclonal antibodies L15-CB1, and CTt-CB2 were provided by Dr Ken Mackie, Bloomington, IN, USA and used with their respective blocking peptides. The TRPV2 rabbit polyclonal antibody (ACC-039) targeting an extracellular epitope was purchased from Alomone (Jerusalem, Israel) and was used with the supplied blocking peptide. Mouse monoclonal antibody for caveolin-1 (C37120), and mouse flotillin-1 (312–428) were purchased from BD Transduction Laboratories (San Diego, CA, USA). Mouse monoclonal antibody against β-actin was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA), and rat monoclonal anti-LAMP-1 (1D4B) was purchased from Developmental studies hybridoma bank (NIH).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

Localization of lipid raft markers

BV-2 cells were fractionated using a continuous (0–20% OptiPrep), detergent-free, density gradient (Macdonald and Pike, 2005). The distribution of the lipid raft marker [18:0]-sphingomyelin in BV-2 cells (Figure 1A) was equivalent to the distribution of the membrane/lipid rafts protein marker flotillin-1 (Figure 1B). Note that the lipid rafts isolated from BV-2 cells (fractions 7–10) are distributed differently on the continuous OptiPrep density gradient when compared with lipid rafts isolated from F-11 cells (fractions 10–13; Rimmerman et al., 2008). In the BV-2 cells fraction 6 represents an interface fraction where some mixing of lipid rafts and other cellular components occurs. The phenomena that membranes from different cell lines have different average densities, resulting in lipid rafts and other membranes to be recovered at slightly shifted positions on the density gradient, was previously reported (Macdonald and Pike, 2005). Figure 1C shows a comparison of the distribution of the lipid raft marker [18:0]-sphingomyelin in both cell lines. β-actin, a cytoskeletal protein associated with both raft and non-raft fractions was observed in high concentrations in non-lipid raft fractions, and found in low levels in lipid raft fractions (Figure 1B). The lysosomal marker LAMP-1 was found in the lipid raft fractions as previously reported (Figure 1B; Zhai et al., 2009). The membrane raft protein caveolin-1, which was shown to be associated with the caveolae lipid raft subtype and necessary for their formation (Drab et al., 2001) was absent from BV-2 cells (Figure 1D). Caveolin-1 was present in CHO cells (Figure 1D), F11 cells (Rimmerman et al., 2008) and human endothelial cells (data not shown). qPCR analysis demonstrated that BV-2 cells were negative for caveolin-1 mRNA (no reaction product was obtained after 40 PCR cycles; mouse primary astrocytes were used as positive control). BV-2 cells were found to express caveolin-2 mRNA (reaction product appeared after 22 cycles).

image

Figure 1. Distribution of membrane raft markers across BV-2 fractions. (A) [18:0]-sphingomyelin is highest in fraction 8 (significantly different from fractions 1–5, and 11–16, P < 0.05, n= 3; one-way anova, Fisher's LSD post hoc). Data are presented as the total quantity in picomoles (pmol) recovered from each fraction. Error bars represent standard error of the mean. (B) Western blot showing the distribution of flotillin-1, LAMP-1, and β-actin. The membrane raft marker flotillin-1 is concentrated in fractions 6–10. β-actin is localized mostly to fractions 1–7, with small amounts in the lighter fractions 8–9. LAMP-1, a lysosomal marker is spread through the lipid rafts and lighter fractions 6–13. (C) Comparison of [18:0]-sphingomyelin distribution between BV-2 (left y-axis) and F-11 (right y-axis) cells. Data for F-11 cells (from Rimmerman et al. 2008) was re-plotted in picomoles. (D) Western blot showing that BV-2 cells are devoid of the protein caveolin-1, CHO cells serve as a positive control.

Download figure to PowerPoint

Compartmentalization of endocannabinoids in BV-2 cells

2-AG levels in control cells were highest in the lipid raft fractions (6–9; Figure 2A) as previously shown by us for the F-11 cells, albeit with a general leftward shift of lipid raft fractions on the density gradient when compared with F-11 cells (Figure 2B; Rimmerman et al., 2008). The other endocannabinoid, AEA, was present in both lipid raft and non-lipid raft fractions (Figure 2C). Owing to the low levels of AEA, it was not possible to detect it in all of the fractions. Western blot analysis revealed localization of CB2 receptors and of the vanilloid receptor TRPV2 to non-raft fractions (Figure 2D). High levels of CB1 receptors were localized to the high-density non-lipid raft fractions, while smaller amounts were found through the lipid raft fractions (Figure 2D). The localization of CB1 receptors (and other proteins) between non-lipid rafts and lipid rafts was verified by a second fractionation method (Song et al., 1996; Ostrom et al., 2001) which showed similar results (Figure 2E), i.e., the presence of flotillin-1, LAMP-1, small amounts of CB1 receptor and lack of caveolin-1 in lipid rafts isolated from BV-2 cells.

image

Figure 2. Distribution of endocannabinoids and CB receptors across BV-2 fractions. (A) 2-AG (in pmol) is highest in interface fraction 6, and significantly different from fractions 1–5 and 10–16 (P < 0.05, n= 3, one-way anova, Fisher's LSD post hoc). (B) Comparison of 2-AG distribution between BV-2 (left y-axis) and F-11 (right y-axis) cells. Data for F-11 cells was re-plotted from Rimmerman et al. (2008). (C) AEA is found in non-lipid raft fraction 5, 6 and lipid raft fraction 9. (D) CB2 receptors and the vanilloid receptor TRPV2 are localized to non-lipid raft fractions 1–5 and interface fraction 6. CB1 receptors are present mostly in non-lipid raft fractions 1–5 and 6. Small amounts are present in lipid raft fractions. (E) Western blot analysis of fractions obtained using a second fractionation method confirmed the compartmentalization of β-actin, TRPV2 and most of CB1 receptors in the non-lipid raft fractions. Flotillin-1, LAMP-1 and part of the CB1 receptors are present in lipid raft fractions. Caveolin-1 is not found in any of the fractions. These results are consistent with the OptiPrep fractionation method.

Download figure to PowerPoint

Three NAEs: OEA, PEA and SEA showed a distribution similar to that of 2-AG. Specifically, OEA and PEA levels were highest in the interface fraction 6 and in the lipid raft fractions 7–9 (their levels were lower in fractions 1–5 and 10–16 (Figures 3A, B, P < 0.05, n= 3, one-way anova, Fisher's LSD post hoc). SEA was highest in interface fraction 6 which was significantly different from fractions 1–4 and 10–16 (Figures 3C, P < 0.05, n= 3, one-way anova, Fisher's LSD post hoc).

image

Figure 3. Distribution of the NAEs, OEA, PEA and SEA, across BV-2 fractions. (A) OEA levels are highest in interface fraction 6 which is significantly different from fractions 1–5 and 10–16 (P < 0.05, n= 3, one-way anova, Fisher's LSD post hoc). (B) PEA levels are highest in interface fraction 6 which is significantly different from fractions 1–5 and 10–16 (P < 0.05, n= 3, one-way anova, Fisher's LSD post hoc).(C) SEA is highest in interface fraction 6 which is significantly different from fractions 1–4 and 10–16 (P < 0.05, n= 3, one-way anova, Fisher's LSD post hoc).

Download figure to PowerPoint

Compartmentalization of endocannabinoids following treatment with CBD

Treatment with CBD (10 µM) for 3 h did not change the distribution pattern of the lipid raft marker [18:0]-sphingomyelin, or protein marker flotillin-1 (Figure 4A, C). CBD was localized to interface fraction 6 and to a lesser degree to lipid raft fraction 7 (Figure 4B). The distribution of β-actin, LAMP-1, CB1, CB2 and, TRPV2 was not affected by CBD (Figure 4C). Similarly, levels and distribution of 2-AG and AEA were equivalent to vehicle-treated cells (Figure 5A, B). However, following CBD treatment, the level of OEA in lipid rafts was not increased (Figure 6A), whereas the saturated NAEs, PEA and SEA were significantly increased in lipid raft fractions, demonstrating a shift towards the lighter fractions (Figure 6B, C).

image

Figure 4. Distribution of membrane raft markers and CBD across BV-2 fractions following CBD treatment. (A) [18:0]-sphingomyelin is highest in fraction 9 (significantly different from fractions 1–6, and 12–16, P < 0.05, n= 3; one-way anova, Fisher's LSD post hoc). (B) CBD is highest in fractions 6 and 7 (CBD levels found in fraction 6 are significantly different from fractions 1–5 and 8–16; P < 0.02; n= 3, one-way anova, Fisher's LSD post hoc). (C) Lipid raft marker flotillin-1 is concentrated in fractions 6–11. β-actin is localized mostly to fractions 1–6. LAMP-1 is spread through lipid rafts and the lighter fractions 6–14. CB2 receptors and the vanilloid receptor TRPV2 are localized to non-membrane raft fractions 1–5, and interface fraction 6. CB1 receptors are compartmentalized mainly to non-lipid raft fractions with small amounts appearing in lipid raft fractions.

Download figure to PowerPoint

image

Figure 5. Distribution of the endocannabinoids AEA and 2-AG across BV-2 fractions following CBD treatment. (A) 2-AG levels and distribution are not affected by CBD treatment. (B) Following CBD treatment, AEA was found in the heavier fraction 5, interface fraction 6, and lipid raft fractions (7 and 9).

Download figure to PowerPoint

image

Figure 6. Distribution of the NAEs, PEA, SEA, and OEA across BV-2 fractions following CBD treatment. (A) OEA levels are highest in fraction 7 in the CBD-treated cells and are significantly different from fractions 1–5 and 10–16 (P < 0.05; n= 3, one-way anova, Fisher's LSD post hoc. The levels of OEA are not significantly increased in CBD-treated cells compared with vehicle-treated cells. (B) In CBD-treated cells, PEA peaked in fraction 7 and was increased in fractions 6–10 (fraction 7 is significantly different from fractions 1–5 and 11–16; P < 0.02; n= 3, one-way anova, Fisher's LSD post hoc). Lipid raft fractions 7–10 in CBD-treated cells have significantly increased levels of PEA when compared with vehicle-treated cells (#, fraction 7, P < 0.03; fraction 8, P < 0.06; fraction 9, P < 0.03; fraction 10, P < 0.05; n= 3, one-way anova, Fisher's LSD post hoc). (C) SEA levels are highest in fraction 7 in the CBD-treated cells and are significantly different from fractions 1–5 and 11–16, P < 0.01; n= 3, one-way anova, Fisher's LSD post hoc). Lipid raft fractions (7–10) in the CBD-treated cells have significantly increased levels of SEA (fraction 7, P < 0.005; fraction 8, P < 0.02; fraction 9, P < 0.05; fraction 10, P < 0.02; n= 3, one-way anova, Fisher's LSD post hoc).

Download figure to PowerPoint

In order to investigate the kinetics of the CBD-induced changes in 2-AG and NAEs levels, cells were incubated with vehicle or CBD and analysed for lipid content (including growth medium) at different time points (10, 30, 60, and 210 min). 2-AG levels in whole cells + growth medium did not change following CBD treatment (Figure 7A). Conversely, AEA levels were significantly increased in whole cells with growth medium at 60 and 210 min compared with vehicle-treated cells (Figure 7B). OEA levels were also significantly increased in whole cells with growth medium at 60 and 210 min compared with vehicle-treated cells (Figure 7C), whereas PEA and SEA levels were not markedly increased (Figure 7D, E).

image

Figure 7. 2-AG and NAE levels in whole cells + media, following CBD treatment. (A) 2-AG levels in whole cells + growth medium are not significantly different between CBD-treated and vehicle-treated cells. (B) AEA levels are increased in whole cells + growth medium from CBD-treated cells at 60 min and 210 min, #, P < 0.001; n= 3, at CBD 210 min, n= 2, Bonferroni post hoc. (C) OEA levels are increased in whole cells + growth medium from CBD-treated cells compared with control at 60 min #, P < 0.02; at 210 min, *P < 0.001; n= 3 (at CBD 210 min, n= 2), one-way anova, Bonferroni post hoc. (D) PEA levels in whole cells + growth medium are not significantly different between CBD-treated and vehicle-treated cells. (E) SEA levels in whole cells + growth medium are not significantly different between CBD-treated and vehicle-treated cells.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

Microglia are resident macrophages that serve as early host defence against pathogens in the CNS. Chronic activation of microglial cells seems to be a major factor in the development of neuroinflammation and contributes to neuronal damage and pathology associated with neurodegeneration (Farooqui et al., 2007). The murine BV-2 cell line is a useful model of microglial cells (Blasi et al., 1990; Bocchini et al., 1992), and we recently showed that CBD inhibits the anti-inflammatory response and alters gene expression of stress genes in these cells (Kozela et al., 2010; Juknat et al., 2012). Additionally, BV-2 cells express a functional cannabinoid system including related receptors: CB1, CB2, GPR55, GPR18, the metabolic enzymes: ABHD6 and FAAH, the endocannabinoids 2-AG and AEA, and other NAEs (Pietr et al., 2009; Stella, 2009; Marrs et al., 2010; McHugh et al., 2010).

Compartmentalization of 2-AG and NAEs in membrane domains is not well characterized, and lipid raft/caveolae are emerging as important domains involved in their signal transduction (McFarland et al., 2004; 2008; Bari et al., 2005a,b; 2006; 2008; McFarland and Barker, 2005; Sarnataro et al., 2005, 2006; Oddi et al., 2007; Placzek et al., 2008; Rimmerman et al., 2008; Yates and Barker, 2009; Maccarrone et al., 2010). We previously investigated the compartmentalization of endocannabinoid signalling in F-11 cells which served as a model of peripheral neurons. Here, we report that the lipid rafts of F-11 cells and of BV-2 cells show different densities, and appear in slightly different locations on the gradient (using identical density gradient methodology). Lipid rafts in the F-11 cells are well separated from non-lipid raft fractions (forming two separate peaks). In contrast, BV-2 lipid raft fractions exhibit a higher density as evident by the localization of lipid raft markers, flotillin-1 and sphingomyelin. The fact that membranes from different cells exhibit altered average densities has been previously reported (Macdonald and Pike, 2005). Differences in protein expression and lipid composition may explain these differences. Indeed, we find that BV-2 cells, in contrast to F-11 cells, are devoid of caveolin-1. Caveolin-1, which is necessary for the formation of caveolae, was reported to modulate cholesterol transport (Murata et al., 1995; Ikonen and Parton, 2000). Pike et al. (2002) reported that human epidermal carcinoma cells devoid of caveolin-1 had significantly reduced cholesterol levels compared with human epidermal carcinoma cells which had been transfected with mouse caveolin-1. Such differences in protein lipid content (e.g. caveolin-1 and cholesterol) may contribute to the density differences observed between F-11 and BV-2 lipid rafts.

Compartmentalization of 2-AG in BV-2 lipid rafts

In F-11 cells, endogenous 2-AG was concentrated in lipid raft fractions, where its levels were significantly higher than in any other fraction. The non-lipid raft fractions contained much lower levels of 2-AG. A similar phenomenon was observed in the BV-2 cells. Incubation of BV-2 cells with CBD did not change the distribution or levels of 2-AG between fractions. Whole cells with growth medium levels were also not affected by the CBD treatment, and the synthetic enzyme DGLα was below our detection limit using Western blot analysis. To conclude, 2-AG has a high affinity for lipid rafts in two cell lines, and CBD did not affect 2-AG metabolism or synthesis in BV-2 cells.

Compartmentalization of NAEs in BV-2 lipid rafts

We previously investigated the compartmentalization of AEA in F-11 fractions and found that AEA was present in significant levels in both lipid raft and non- lipid raft fractions. Using deuterium-labelled AEA, we determined that deuterium-labeled free arachidonic acid, a direct metabolic product of AEA hydrolysis, was found in non-lipid raft fractions. We postulated that AEA was endocytosed and trafficked to be hydrolysed in non-lipid raft FAAH-containing membranes. This is consistent with the proposed theory of lipid raft/caveolae-mediated AEA endocytosis (McFarland and Barker, 2005). AEA levels in several of the BV-2 membrane fractions were below our detection limit. Wherever AEA was detectable, significant levels were measured in both lipid rafts and non-lipid raft fractions (similar to AEA distribution in F-11 cells). Remarkably, AEA levels in whole cells with growth medium were significantly increased following CBD treatment in a time-dependent manner. Because the levels of AEA in membrane fractions cannot account for these changes, we hypothesize that excess AEA production following CBD treatment is released into the growth medium.

The CB1 receptor was previously compartmentalized in lipid raft/caveolae membranes in several cellular systems (Keren and Sarne, 2003; Bari et al., 2005b; 2006; 2008; Sarnataro et al., 2005; 2006). In BV-2 cell fractions, we observed compartmentalization of CB1 receptors mostly to non-lipid raft membranes. Using two different fractionation methods we show that only small amounts of CB1 receptors are localized to lipid rafts. It is yet to be determined whether association of CB1 receptors with BV-2 lipid rafts is affected by the lack of caveolin-1, which may well affect cholesterol balance in the cells. This association may explain apparent differences in lipid raft localization of CB1 receptors between cell types. Consistent with previous reports, CB2 receptors did not compartmentalize in lipid rafts (Bari et al., 2006; Rimmerman et al., 2008).

Levels of the monounsaturated N-acyl ethanolamine derivative OEA showed a similar trend to AEA. OEA levels in the membrane fractions showed a slight, non-significant increase following CBD treatment. However, OEA levels in whole cells with growth medium were significantly increased following CBD incubation in a time-dependent manner. The accumulation of AEA and OEA in whole cells with growth medium following CBD treatment, suggests that CBD is inhibiting their uptake and/or hydrolysis. Although AEA uptake has been extensively investigated, not much is known about the uptake of OEA (Di Marzo et al., 1994; Beltramo et al., 1997; Bisogno et al., 1997; 2001b; Jacobsson and Fowler, 2001; Fowler and Jacobsson, 2002; McFarland and Barker, 2005; Hermann et al., 2006; Thors et al., 2007; Placzek et al., 2008; Kaczocha et al., 2006; 2009; Oddi et al., 2008; 2009; Maccarrone et al., 2010). In contrast, the mechanisms of NAE hydrolysis have been determined (Cravatt et al., 1996). In this context, CBD was previously shown to inhibit AEA uptake and hydrolysis in several systems. For example, CBD inhibited AEA hydrolysis in mouse brain microsomes (Watanabe et al., 1996). In addition, CBD inhibited AEA uptake by RBL-2H3 with Ki∼ 11 µM (Rakhshan et al., 2000). Furthermore, Bisogno et al. (2001a) reported that CBD inhibited [14C]-AEA uptake in RBL-2H3 cells (IC50 of 22 µM), and [14C]-AEA hydrolysis in N18TG2 cell membranes (IC50 of 27.5 µM). Different results were reported by Massi et al. (2008) where CBD stimulated the activity of FAAH, and decreased AEA content in tumors from CBD-treated nude mice, and in U87 human glioma cells. Our results in BV-2 cells support the reported inhibition by CBD of AEA uptake and/or hydrolysis, and extend it to OEA. Additionally, we showed that following CBD treatment, the levels of the saturated NAEs, PEA and SEA significantly increased in lipid raft fractions. This is also consistent with an inhibitory effect of CBD on NAE metabolism and the compartmentalization of this accumulation to lipid rafts. We also found that CBD treatment did not affect 2-AG levels, consistent with the findings of Marrs et al. (2010) showing that 2-AG hydrolysis in intact cells proceeds through the enzyme ABHD6 rather than FAAH. Aside from the effects of CBD on FAAH activity, there may be other mechanisms leading to the increased accumulation of NAEs. For example, we have found that CBD increased, by 90%, the mRNA level of the enzyme protein tyrosine phosphatase, non receptor 22 (PTPN22) (Rimmerman, unpublished experiments), which was shown to be involved in AEA synthesis from phospho-AEA in a macrophage cell line (Liu et al., 2006).

CBD was previously shown to exert immunosuppressive effects on immune cells in vitro and in vivo (Ignatowska-Jankowska et al., 2009; Kozela et al., 2010; 2011; Ruiz-Valdepenas et al., 2011). It is not clear whether the immunosuppressive effects of CBD are mediated via the CBD-induced increases in NAEs, and the NAEs differential compartmentalization. At least for AEA, immunosuppressive effects through the CB2 receptor have been described (Cencioni et al., 2010; Correa et al., 2011). Additional research will be needed to determine the contribution of NAEs compartmentalization to immune function.

In conclusion, CBD treatment generally increased NAEs accumulation in BV-2 cells. The increase in NAE accumulation may be explained by inhibition of FAAH and/or NAE uptake, induced by CBD. We report that following CBD treatment, NAEs undergo different compartmentalization, depending on the saturation of the fatty acid. The saturated NAEs (PEA and SEA) markedly increased in lipid raft fractions, while the unsaturated NAEs (AEA and OEA) markedly increased in whole cells with growth medium, probably via release to the medium. Our results support the idea that although NAE levels are controlled by similar enzymatic mechanisms, they compartmentalize and signal via divergent pathways. In addition, we hypothesize that the anti-inflammatory effects of CBD may be mediated by the increased accumulation of anti-inflammatory NAEs in microglial cells.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest
  9. References

This work was supported by the Dr Miriam and Sheldon G. Adelson Medical Research Foundation. A.J and N.R. were supported by the Israeli Center for Absorption in Science.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflicts of interest
  9. References