Journal of Neurvchemistry Lippincott Williams & Wilkins, Inc., Philadelphia 0 1999 International Society for Neurochemistry

GCMS Analysis of Anandamide and Quantification of N-Arachidonoylphosphatidylethanolamides in Various Brain

Regions, Spinal Cord, Testis, and Spleen of the Rat

Hsiu-Ying T. Yang, Farouk Karoum, *Christian Felder, ?Henry Badger, 'fTao-Chin Lin Wang, and ?Sanford P. Markey

National Institute of Mental Health, Neuroscience Center at St. Elizabeths, Washington, D. C.; *Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, Indiana; and ?Laboratory of Neurotoxicology,

National Institute of Mental Health, Bethesda, Maryland, U.S.A.

Abstract: Anandamide [N-arachidonoylethanolamide (NAE)] was initially isolated from porcine brain and proposed as an endogenous ligand for cannabinoid receptors in 1992. Accumulating evidence has now suggested that, in the tissue, NAE is generated from N-arachidonoylphosphatidylethanolarnides (N-ArPEs) by phosphodiesterase. In this study a sensitive and specific procedure was developed to quantify NAE and N-ArPE, including organic solvent extraction, reverse- phase C-18 cartridge separation, derivatization, and gas chromatography/mass spectrometry (GUMS) analysis. NAE is converted by a two-step derivatization procedure to a pentafluorobenzoyl ester followed by pentafluoropropionyl acylation. Quantification was per- formed by isotope dilution GC/MS using deuterium- labeled NAE (NAE-*H,) as an internal standard. The same chemical derivatization was applicable to N-ArPE quantification. The separated N-ArPE fractions were converted by a two-step cleavage/derivatization pro- cedure into the pentafluorobenzoyl ester of NAE and then to its pentafluoropropionyl amide. The derivative was quantified by GC/MS using deuterium-labeled 1,2- [2H,]dioleoyl-sn-glycero-3-phospho(arachidonoyl)eth- anolamide as an internal standard. Using these meth- ods, we have found that endogenous NAE levels in rat brain, spleen, testis, liver, lung, and heart were below the level of quantification achievable (0.1 pmol/mg of protein) but that N-ArPE is readily quantifiable and is widely distributed in the rat CNS with the highest level in the spinal cord. The striaturn, hippocampus, and accumbens contain intermediate concentrations of N-ArPE, whereas the value is lowest in the cerebel- lum. Key Words: Anandamide-N-Arachidonoylphos- phatidylethanolamides-Endogenous cannabinoid- Gas chromatography/mass spectrometry-Deuterium labeling-Reverse-phase C-18 cartridge. J. Neurochem. 72, 1959-1 968 (1 999).

N-Arachidonoylethanolamide [5,8,11, 14-eicosatetra- eneamide, N-(2-hydroxyethyl-(d 2); anandamide; NAE]

was initially isolated from porcine brain and proposed as one of the endogenous ligands for cannabinoid receptors (Devane et al., 1992). NAE displayed similar in vivo and in vitro effects as a cannabinoid receptor agonist acting through the CB1 cannabinoid receptor (for review, see Felder and Glass, 1998). NAE induces antinociception, hypomotility, hypothermia, and catalepsy when injected into mice, and, in vitro, it is capable of inhibiting forskolin- induced accumulation of intracellular cyclic AMP and N- type Ca2+ channels in several types of cells (Felder et al., 1993; Vogel et al., 1993).

The biosynthesis of NAE has been investigated by several groups, and two mechanisms have been pro- posed. One possible biosynthetic mechanism is energy- independent condensation of ethanolamine and arachi- donate catalyzed by either an NAE synthetase or an amidohydrolase working in reverse (Deutsch and Chin, 1993; Devane and Axelrod, 1994; Kruszka and Gross, 1994; Ueda et al., 1995). A second NAE-forming process involves cleavage of NAE from N-arachidonoylphos- phatidylethanolamide (N-ArPE) catalyzed by phosphodi- esterase (Schmid et al., 1983; Di Marzo et al., 1994; Cadas et al., 1996~ ; Sugiura et al., 1996~). The identifi- cation of endogenous N-ArPE (Di Marzo et al., 1994; Cadas et al., 19966, 1997; Sugiura et al., 1996u,b) and of an enzyme capable of forming NAE from N-ArPE strongly supports the latter biosynthetic pathway. Recent

Received October 9, 1998; revised manuscript received December 7, 1998; accepted December 16, 1998.

Address correspondence and reprint requests to Dr. S. P. Markey at Laboratory of Neurotoxicology, Building 10, Room 3D42, National Institute of Mental Health, Bethesda, MD 20892-1262, U.S.A.

The first two authors contributed equally to this research. Abbreviations used: N-ArPE, N-arachidonoylphosphatidylethanol-

amide; GCMS, gas chromatography/mass spectrometry; NAE, N-arachidonylethanolamide; PFBz, pentafluorobenzoyl; PFBzC1, pen- tafluorobenzoylchloride; PFP, pentafluoropropionate; PFPA, pentaflu- oropropionic anhydride; TFA, trifluoroacetic acid.

1959

1960 H.-Y. T. YANG ET AL,

studies have confirmed the existence of N-acyltrans- ferase, an enzyme catalyzing the transference of an arachidonoyl group from the sn-1 carbon of phospholip- ids to the amino group of phosphatidylethanolaine, as originally described by Schmid and co-workers for sat- urated N-acylphosphatidylethanolamide (for review, see Schmid et al., 1990, 1996), and phospholipid substrates necessary for the biosynthesis of N-acylphosphatidyleth- anolamide (Berdyshev et al., 1996; Sugiura et al., 1996b; Cadas et al., 1997), giving further support to the role of N-ArPE as an NAE precursor. However, the information on this NAE biosynthetic pathway is limited; for exam- ple, levels of N-ArPE in brain regions are currently unavailable. To study the regulation of NAE biosynthe- sis, especially in vivo, it is desirable to have a sensitive and relatively simple method to measure the very small amounts of endogenous N-ArPE, in addition to NAE, in mammalian tissues.

NAE and other N-acylethanolamides in biological tis- sues are usually extracted into neutral organic solvents such as chloroform (Schmid et al., 1995u, 1996) or chlorofodmethanol mixtures (Kempe et al., 1996; Ca- das et al., 1997) using procedures modified after that of Folch (Shaikh, 1994). After purification by silica gel column, TLC, or HPLC, NAE has been detected and quantified by liquid or gas chromatographylmass spec- trometry (GUMS) (Devane et al., 1992; Schmid et al., 1995b; Felder et al., 1996; Kempe et al., 1996; Schmid et al., 1996). There are a range of values reported in the literature with regard to endogenous levels of NAE in various tissues. In this communication we report simpli- fied methods to separate and quantify endogenous NAE, demonstrate that a solid-phase cartridge separates NAE from its precursor N-ArPE (Schmid et al., 1983; Di Marzo et al., 1994; Cadas et al., 1996b, 1997; Sugiura et al., 1996u,b), and confirm that endogenous NAE levels in the rat brain are below our limit of quantification (-0.1 pmoVmg of protein), consistent with the findings of others (Kempe et al., 1996).

N-ArPE is not a single molecular species, but rather the name used to denote a family of 1,2-diacyl-sn-glyc- ero-3-phospho-(N-arachidonoyl)ethanolamides and the plasmalogen family of 1 -( 1 '-alkenyl)-2-acyl-sn2-glycero- 3-phospho-(N-arachidonoyl)ethanolamides, which both contain the same NAE substructure and are substrates for phospholipase D (Schmid et al., 1990). The plasmalo- gens are present in equal or greater amounts in bovine brain as the diacyl-N-ArPE and are likely present in other extracts of N-acylphosphatidylethanolamide (Schmid et al., 1997~). In this study, both of these families are designated as N-ArPE and were separated from other lipids first by using a reverse-phase C-18 cartridge and then cleaved in a derivatization reaction to form the same derivative as used in the quantification of NAE. Thus, a sensitive and specific method for NAE was successfully applied to assay N-ArPE in various brain regions, the spinal cord, the testis, and the spleen of the rat.

MATERIALS AND METHODS

Materials Male Sprague-Dawley rats (Charles River, Raleigh, NC,

U.S.A.) weighing 200-250 g (49-56 days old) were housed with a 12-h light period with free access to food and water. All rats were acclimatized for 1 week before use. Procedures for all animals used were in accordance with the NIH Guidelines for the Care and Use of Laboratory Animals and were approved by the NIMH Animal Care and Use Committee. Sep-Pak C-18 cartridges were purchased from Waters Chromatography Divi- sion, Millipore Corp. (Milford, MA, U.S.A.). NAE was pur- chased from Research Biochemicals International (Natick, MA, U.S.A.) and from Biomol (Plymouth Meeting, PA, U.S.A.). Deuterium-labeled (5,6,8,9,11,12,14,15 NAE-2H, 98%, 99 at- om% 'H,) was purchased after custom synthesis from Biomol. 1,2-Dioleoyl-sn-glycero-3-phospho-(N-arachidonoyl)ethanol- amide (dioleoyl-N-ArPE) was purchased from Avanti Polar Lipids (Alabaster, AL, U.S.A.), and 1,2-dioleoyl-sn-glycero-3- pho~pho-(N-5,6,8,9,11,12,14,15-~H, arachidonoy1)ethanolamide (dioleoyl-N-ArPE-2H,) was purchased after custom synthesis from Avanti. Pentafluorobenzoylchloride (PFBzC1) and pen- tafluoropropionic anhydride (PFPA) were purchased from Pierce Chemical Co. (Rockford, IL, U.S.A.). Phospholipase D (from Streptornyces chrornofiscus) was obtained from Sigma Chemical Co. (St. Louis, MO, U.S.A.).

Protein content was determined according to the method of Lowry et al. (1951) using bovine serum albumin as the stan- dard.

Lipid extraction and fractionation All rats were killed by decapitation, and their brains were

rapidly removed, frozen in dry ice, and kept frozen (at -30 or -70°C) until processed. For comparison of different extraction methods, the frozen whole brains (minus cerebellum and me- dulla) were divided into two halves on a cold plate, and tissues were immediately homogenized (5 ml per half brain) in chlo- rofondmethanol mixture (2: l), ethyl acetate, or chloroform using an Ultra-Turax Tissuemizer (Tekmar, Cincinnati, OH, U.S.A.). For experiments to determine the quantities of NAE or N-ArPE in tissue, NAE-2H, (see figure legends) or dioleoyl- N ~ u P E - ~ H , was added to the homogenization solvent. For experiments to compare the composition of fractions eluting from solid-phase cartridges, 150 pmol of NAE-2H, was added to fractions 1-8, and 150 pmol of dioleoyl-N-ArPE-2H, was added to fractions 9-16. Fifty-microliter aliquots of the chlo- roform and chlorofordmethanol homogenate were removed for protein content determinations. The homogenates were then centrifuged at 4,900 g for 20 min, and the clear supernatants were removed and dried with a gentle stream of nitrogen. Pellets from the ethyl acetate homogenates were used for protein content determinations, because it was not possible to remove a uniform sample from the ethyl acetate homogenates. The dried extracts were redissolved in 1 ml of ethanol, soni- cated if necessary, and centrifuged at 12,000 g for 5 min. Then aliquots (200 pl) of the supernatants were removed for frac- tionation with reverse-phase C- 18 cartridges.

For the determination of regional N-ArPE distribution, rats were killed by decapitation. Brains were rapidly removed and dissected on ice into various regions. The tissues were then frozen immediately in dry ice and kept at -70°C until pro- cessed. Frozen tissue was homogenized in 1 ml of ethyl acetate containing 30 pmol of dioleoyl-N-ArPE-2H, using a 40-W Vibracell sonic homogenizer (Sonics and Materials, Danbury, CT, U.S.A.). The homogenates were ,.centrifuged, the clear

J. Neurochem., Vol. 72, No. 5, 1999

G U M S ANALYSIS OF NAE AND N-ArPE 1961

FIG. 1. Derivatization of NAE and N-APE. Formation of NAE-PFBz- PFP is shown from NAE directly or from N-ArPE by a two-step cleav- age/derivatization procedure.

c!

supernatants were dried as described above, and the pellets were used for protein content determination. The dried extracts were dissolved in 200 p1 of ethanol and centrifuged, and then the clear supematants were fractionated by reverse-phase C-18 cartridges.

Samples in ethanol were applied to C-18 cartridges pre- equilibrated with 50% CHJN in water containing 0.1% tri- fluoroacetic acid (TFA). The C-18 cartridges were then eluted with increasingly nonpolar solvents (all containing 0.1% TFA), using 3 ml of 50% CH,CN/water, then 6 ml of CH,CN, 3 ml of methanol, and finally 4 ml of 90% chloroform in methanol; 1-ml fractions were collected. The fractions in Reactivials were dried with gentle streams of nitrogen and used for derivatiza- tion and GC/MS quantification.

Derivatization and GC/MS analysis The fractions from C-18 cartridges were converted by a

two-step cleavage/derivatization procedure suitable for the quantification of either NAE or N-ArPE shown in the reaction scheme in Fig. 1. NAE reacts with PFBzCl using mild condi- tions, without amine catalysis, to form a monopentafluoroben- zoyl [mono-(PFBz)] ester. N-ArPE is cleaved under the same conditions with PFBzCl to form the same PFBz ester of NAE. In a second step, the 3FBz-NAE derivatives are reacted with

> NAE-PFBz

0

NAE-PFBz-PFP

PFPA using amine catalysis to form an NAE-PFBz-pentaflu- oropropionate (PFP) derivative. In a 1-ml Reactivial, 50 p1 of anhydrous ethyl acetate (stored over CaH,) is added to the dry residue from either evaporated solvent extracts or standards. PFBzCl in CH,Cl, (10 pl of a 20% solution, stored over CaH,) is added. The vials are tightly capped, and the mixture is heated at 65°C for 30 min. After cooling, the mixture is evaporated under a gentle stream of dry nitrogen. The dried residue is reacted to form the combined PFPlPFBz derivative by addition of 10 p1 of 1% triethylamine in anhydrous ethyl acetate and 100 pl of PFP (care is exercised to assure anhydrous condi- tions). The vials are tightly capped and heated at 65°C for 30 min. After cooling to room temperature, the products are con- centrated to dryness under dry nitrogen and reconstituted in aliquots of 10 pl of anhydrous ethyl acetate. From 1 to 2 p1 of this solution is injected into the gas chromatograph. Gas chro- matographic separation of NAE-PFBz-PFP was accomplished using an SPBl (Supelco, Bellafonte, PA, U.S.A.) fused silica capillary column (15 m X 0.32 mm id. , 0.5 p m film thickness). Split injection was used at a ratio of -l:lO, with the column oven temperature maintained isothermally at 260°C and the injection port temperature set at 220°C. Mass spectrometric analysis and quantification were performed with a Finnigan

J. Neurochem., Vol. 72, No. 5, 1999

1962 H.-Y. T. YANG ET AL.

- 1 47

473 275

hl- 667 667

0 100 200 300 400 500 600 700

MR

100 200 300 400

(Sunnyvale, CA, U.S.A.) 4600 quadrupole mass spectrometer. Electron capture negative ionization was generated with meth- ane at an ion source pressure of 0.15 torr and an ionizer temperature of 160°C.

Phospholipase D digestion of N-ArPE fraction followed with analysis of NAE by GC/MS

One whole brain (minus cerebellum and medulla portions) was homogenized in 10 ml of ethyl acetate using an Ultra- Turax Tissuemizer, and the homogenate was centrifuged. The clear supernatant was removed and dried by a gentle stream of nitrogen. The dried extract was fractionated using C-18 car- tridges as described above. The N-ArPE fractions were com- bined, divided into two equal aliquots, dried, and used for phospholipase D digestion according to published methods (Imamura and Horiuti, 1979). One N-ArPE aliquot was incu- bated with 10 units of phospholipase D (from Streptomyces chromofuscus) in 250 p l of 0.04 M, pH 8.0 Tris buffer con- taining 4 pM CaCl, and 2.5 mg of Triton X-100 at 37°C for 4 h. The other N-ArPE aliquot was incubated with heat-inac- tivated enzyme and used as control. The enzyme reaction mixture was heated at 100°C for 10 min and extracted with 1 ml of ethyl acetate after adding 100 pmol of 2H,-NAE. The ethyl acetate extract was dried under nitrogen, redissolved in ethanol, and fractionated again using the C-18 cartridge, as described above. The fraction eluting with NAE characteristics was applied to a reverse-phase HPLC column (Beckman ODS-5,4 X 250 mm). The HPLC column was then eluted with a linear gradient of 50-80% CH,CN containing 0.05% TFA over 30 rnin at a flow rate of 1 mumin, conditions in which elution of NAE peaks occurs at fraction 29 when 1-ml fractions were collected. Fractions 28-30 were collected in a Reactivial dried in a vacuum concentrator (Savant Instruments, Hicks-

500 Scan

FIG. 2. A Electron capture negative ionization mass spectrum of NAE-PFBz-PFP derivative formed from NAE or dioleoyl-N-ArPE. B Selected ion record of NAE-PFBz-PFP. The selected ion monitoring trace re- sulted from a mixture of 50 pmol of NAE and NAE-2H, as their PFBz-PFP derivatives: -1/20th of the total sample was injected at a split ratio of -1:15.

ville, NY, U.S.A.) and analyzed by GC/MS. This experiment was repeated with a second rat brain.

RESULTS

Derivatization Figure 1 shows the reaction sequence used for deriva-

tization of NAE and N-ArPE. Dioleoyl-N-ArPE is rep- resentative of the N-ArPE families and was used to optimize the reaction conditions. Fluorinated acyl deriv- atives of NAE were investigated because of their en- hanced electron capturing properties and superior detec- tion by electron capture negative ionization. Several der- ivatization schemes were compared, including formation of the NAE-PFBz or NAE-diPFBz (Kempe et al., 1996) and the NAE-PFBz-PFP derivative, with the NAE-PFBz- PFP providing the best results. The NAE-PFBz-PFP derivative formed reproducibly and exhibited prominent structure-specific ions for facile selected ion monitoring. Excluding triethylamine as a catalyst for the formation of the PFP derivative produces a second reaction product, possibly the PFBz-0-PFP derivative versus the PFBz-N- PFP derivative. The PFBz-N-PFP derivative of either NAE or N-ArPE exhibits the electron capture negative ionization spectrum in Fig. 2A with a base peak at m/z 540 resulting from the loss of COC,F, (M-147); the spectrum of the minor derivative (data not shown) ex- hibits a prominent ion at m/z 523, representing the loss of C,F,CO,H. The volumes of reagents and heating periods used for derivatization were optimized for the production

J. Neurochem., Vol. 72, No. 5, 1999

GUMS ANALYSIS OF NAE AND N-ArPE 1963

TABLE 1. N-ArPE and NAE levels in ethyl acetate, chloroform, and chloroform/methanol ( 2 : l ) extracts

of rat brain

Extraction

Level (pmol/mg of protein)

N-ArPE NAE

Ethyl acetate (7) 1.8 t 0.1 0.2-3.7 Chloroform (5) 1.5 C 0.3 <o. 1 Chlorofordmethanol (5) 1.8 IT 0.4 <o. 1

Data are mean ? SEM values, with the number of animals in parentheses.

of a single derivative, nearly eliminating the appearance of the minor component. We compared the use of ethyl acetate, 1% pyridine in ethyl acetate, and 1% triethyl- amine in dichloromethane with that of 1 % triethylamine in ethyl acetate in the reaction of PFPA with the NAE- PFBz derivative. The use of 1% triethylamine in ethyl acetate produces a consistently high yield of a single NAE-PFBz-PFP derivative.

Figure 2B shows single ion monitoring of the PFBz- PFP derivatives of a mixture of 50 pmol of NAE (natural isotope abundance) and 50 pmol of NAE-2H,. In the absence of tissue extracts, the mass spectrometric limit of detection (signal-to-noise ratio, -3) of NAE-PFBz-PFP standards is 0.2 pmol per tube (2 p l of a total volume of 15 p1 of reconstituted derivative corresponding to 10 pg in ethyl acetate injected into the injection port main- tained at a split ratio of 1:lO).

Extraction To find the most suitable solvent for NAE and N-ArPE

extraction from tissues, brains were homogenized in three different solvents containing NAE-2H, or dioleoyl- N-ArPE-2H, as internal standard. As shown in Table 1, undetectable levels of NAE were observed when chlo- roform or chlorofodmethanol mixtures were used as the homogenizatiodextraction solvent. Irreproducible and variable levels of NAE were observed when ethyl acetate was used. Extracts appeared to contain less chem- ical background when chloroform was used, and conse- quently a determination of sample recovery and assay linearity was conducted using standard additions of 0-3 pmol quantities of NAE and 5 pmol of NAE-2H, to chloroform homogenates of rat brain tissue (Fig. 3A). In contrast, approximately the same levels of N-ArPE were measured in all three solvents (Table 1); subsequently, ethyl acetate was used for tissue extraction in the re- gional distribution study. With dioleoyl-N-ArPE-2Hx as an internal standard and monitoring ions with m/z ratio of 540 (M-147 for NAE) and 548 (M-147 for 2H,-NAE), a linear standard curve was obtained with 0.6-220 pmol per sample of dioleoyl-N-ArPE (Fig. 3B). However, the practical limit of detection in brain, spinal cord, and peripheral tissues (spleen, testis, liver, lung, and heart) is -1 pmo1/100 mg of tissue sample (corresponding to -0.1 pmol/mg of protein). The presence of 1 pmol of NAE in an extract of 100 mg of brain tissue would be

readily measurable. Homogenizing spinal cord and the above-mentioned peripheral tissues (2 100 mg) in chlo- roform failed to reveal detectable quantities of NAE, indicating levels (0.1 pmol/mg of protein.

Separation Figure 4 shows chromatograms constructed by plot-

ting the appearance of NAE-PFBz-PFP ions relative to a constant quantity of NAE-2H, or dioleoyl-N-ArPE-2Hx added to each sample. Authentic NAE elutes from the C-18 cartridge with acetonitrile containing 0.1% TFA in fraction 4; dioleoyl-N-ArPE elutes using 90% chloro- form in methanol containing 0.1% TFA at fractions 13 and 14 (Fig. 4A). When chloroform extracts of rat brain are assayed, there is no detectable NAE in fraction 4, but N-ArPE is readily detected in fractions 13 and 14 (Fig. 4B). There are substances eluted in fractions 1-3 that can generate NAE after treatment with acid or base. Figure 4C shows a chromatogram of a chloroform extract as in Fig. 4B, but after each fraction was evaporated to dry- ness, it was dissolved in 3 M HCl, then extracted with ethyl acetate, evaporated to dryness, and derivatized. Abundant NAE-PFBz-PFP ions are observed in fractions 1-3 in Fig. 4C, but these derive from compounds other

y=0.301x - 0.004

0.75 .

0.5 .

0.25 .

1 1.5 2 2.5 3 c o o d 0 0.5

e 0 w NAE (pmol)

e E

v)

16

14

0 5 10 15 20 25

- 0 20 40 So 80 I00 120 140 160 IS0 200 ; 0

DIOLEOYL-N-ARPE (pmol)

FIG. 3. Upper panel: Standard additions of NAE to rat brain extract. Recovery is shown of 0-3 pmol of authentic NAE added to the same amount of chloroform brain extract (1 ml of organic phase from a total of 5 ml of whole brain homogenate). NAE-2H, (5 pmol) was used as the internal standard. Lower panel: Stan- dard curve of N-ArPE. Quantities of dioleoyl-N-ArPE (0.6-220 pmol) were derivatized in the presence of dioleoyl-N-ArPE-'H, (50 pmol for the 20-200-pmol samples and 10 pmol for the 0.6-24.3-pmol samples). NAE-PFBz-PFP derivatives were ana- lyzed by GC/MS.

J. Neurochem., Vol. 72, No. 5, 1999

I964 H.-Y. T. YANG ET AL.

A Standards 0.6 3 I

1 3 5 7 9 1 1 1 3 1 5

B CHa, Brain Extract 0 . 5 , I

&J 1 3 5 7 9 1 1 1 3 1 5

E C Aclditied CHC13 Brain Extract

0.25 1 I

1 3 5 7 9 1 1 1 3 1 5

Fraction

FIG. 4. Reverse-phase C-18 cartridge fractionation of NAE and N-ArPE. A A mixture of NAE and dioleoyl-N-ArPE (150 pmol each) in 200 pI of ethanol was applied to a C-18 cartridge. The cartridge was then eluted with 3 ml of 50% CH,CN, 6 ml of 100% CH,CN, 3 ml of CH,OH, and 4 ml of 90% CHCI, in CH,OH (0.1 % TFA was included in every eluent), and 1 -ml fractions were collected. The eluate fractions were dried after adding NAE-'H, (100 pmol to fractions 1-8) and dioleoyl-N-ArPE-2H, (47.71 prnol to fractions 9-16) and analyzed. B Rat brain (minus cerebellum and medulla) was homogenized in chloroform as described in Materials and Methods. The homogenate was subjected to C-18 cartridge fractionation and analyzed as described above. C Separation of components from a rat brain chloroform extract and subsequent 3 M HCI exposure. Each chromatographic frac- tion was evaporated to dryness, reconstituted in 3 M HCI, ex- tracted into ethyl acetate, evaporated, and derivatized.

than NAE or N-ArPE as defined by their chromato- graphic properties.

N-ArPE analyses Figure 5 shows a typical selected ion recording ob-

tained from an ethyl acetate extract of rat brain for the N-ArPE fraction (combined fractions 13 and 14 from Fig. 4B) analyzed by G C M S by monitoring the d z ratio 540 (for NAE) and 548 (for 2H,-NAE). To confirm further the identity of N-ArPE, a brain extract was frac- tionated by the C-18 cartridge, and the N-ArPE fraction was digested with phospholipase D. The NAE produced was isolated by HPLC and then analyzed by GCMS. The amounts of NAE generated by the phospholipase D

1173410

A 1:40 2:m 2:m 255 1:40

Elution lime

FIG. 5. Selected ion recording of NAE-PFBz-PFP derivative formed from N-ArPE from rat brain tissue extract. N-ArPE was extracted from the frozen striatum of rat brain with ethyl acetate containing dioleoyl-N-ArPE-*H,, partially purified by a C-18 car- tridge, and then analyzed as described in Materials and Meth- ods. The upper recording represents the detection of N-ArPE as the NAE-PFBz-PFP derivative (m/z 540); the lower recording represents that of dioleoyl-N-ArPE-2H, as the corresponding 2H,-NAE-PFBz-PFP derivative (mk 548).

digestion were found to be 158.8 and 118.8 pmol/g of brain in two experiments.

N-ArPE was found to be widely distributed in the rat CNS (Table 2), with the highest concentration in the spinal cord. The striatum, hippocampus, and accumbens contain intermediate concentrations of N-ArPE, whereas the value is lowest in cerebellum.

DISCUSSION

Initial efforts to derivatize directly organic solvent extracts indicated that the complex mixture of lipids in rat brain extracts precluded successful G C M S analyses without additional separation, as previously noted (Fon- tana et al., 1995). We investigated solid-phase cartridge separations as a means of separating N-ArPE from NAE. Direct derivatization and mass spectrometric assay of each fraction collected from the solid-phase cartridge indicated that NAE-PFBz-PFP was formed from frac- tions eluting not only coincident with NAE standards, but also coincident with dioleoyl-N-ArPE elution. These studies demonstrated that the derivatization chemistry developed for NAE produced the identical reaction prod-

TABLE 2. N-ArPE in brain, spinal cord, spleen, and testis

Region (n) N-ArPE (pmoVmg of protein)

Hypothalamus ( 5 ) 1.0 ? 0.1 Frontal cortex (5) 0.9 ? 0.1 Accumbens (5 ) 1.5 ? 0.1 Striaturn ( 5 ) 1.9 2 0.1 Hippocampus ( 5 ) 1.5 2 0.1 Cerebellum (5) 0.50 ? 0.1 Spinal cord ( 5 ) 3.2 ? 0.4

0.5 2 0.1 Spleen (3) Testis (5) 4.7 ? 0.8

Data are mean ? SEM values, with the number of animals in parentheses.

J. Neurochem., Vol. 72, No. 5, 1999

GC/MS ANALYSIS OF NAE AND N-ArPE 1965

TABLE 3. Comparison of reported concentrations of NAE

Species and tissue NAE amount pmoVmg of protein" Reference

Pig brain 0.6 mg/4.5 kg 3.8 (est) Devane et al. (1992) Pig brain 6 +- 1 ng/g 0.17 ? .03 Schmid P. C. et al. (19956), Sheep brain Not detectable Not detectable Schmid H. H. 0. et al. (1996) Cow brain Wistar rat brain 4.3 +- 1.1 pmol/g 0.04 ? 0.001 (est) Sugiura et al. (1996~) Sprague-Dawley rat brain Not detectable <0.17 Kempe et al. (1996) Sprague-Dawley rat brain 20 pmol/g of thalamus 0.2-0.29 (est) Felder et al. (1996)

Human brain 30 pmoVg of cerebellum 0.3-1.5 (est)

Wistar rat brain Cadas et al. ( 1 997) P. lividus sea urchin Bisogno et al. (1997) Sprague-Dawley rat brain Age and tissue-dependent Koga et al. (1997)

3 4 ng/g 0.09-0.11

29 pmoVg of hippocampus

110 pmoVg of hippocampus 11 +- 7 pmoVg of tissue 12 +- 4 pmoVg of tissue

0.11 ? 0.07 (est) 0.12 5 0.04 (est)

Cerebellum (6 weeks) 50 2 27 pmol/g 0.5 (est) Cerebellum (12 weeks) 73 ? 19 pmol/g 0.7 (est) Hippocampus (6 weeks) 33 ? 13 pmol/g 0.3 (est) Hippocampus (12 weeks) 201 +- 110 pmoVg 2.0 (est)

6 weeks 2.1 2 3.6 pmol/g 0.02 (est) 12 weeks 325 ? 50 pmoUg 3.3 (est)

Rat spleen

Rat blood 5.2 pmoVml Est not possible Giuffrida and Piomelli (1998) Bivalve molluscs

Mouse peritoneal 17-30 pmoVpmo1 of lipid Est not possible Schmid et al. (1997~)

Mouse uterus 1-7 pglg wet weight 2,900-20,000 Schmid et al. (1997b) Rat kidney 2.79 +- 1.1 ng/g 0.08 +- 0.03 Deutsch et al. (1997) Sprague-Dawley rat brain Not detectable <0.1 Present study

Not detectable to 2.0 +- 0.9 ng/g Not detectable to 0.06 L 0.03

Sepe et al. (1998)

macrophages phosphorus

Est, estimate. a Calculations assume tissue is 10% protein per wet weight.

uct for N-ArPE through a combined cleavage and deri- vatization process (Fig. 1).

The NAE data obtained in this study are comparable with those of an earlier GCMS report (Kempe et al., 1996), in terms of both the detection limit and the ab- sence of measurable quantities of NAE. Table 3 com- pares published quantitative determinations of NAE. One group has found levels in rat brain below our limit of detection (Sugiura et al., 1996~) apparently using more sensitive analytical methods. Others have reported greater tissue concentrations of NAE than observed in this study, and as has been previously observed, there are several reasons for the detection of higher levels of NAE. Two groups (Kempe et al., 1996; Schmid et al., 1996) have reported the postmortem generation of NAE. Others have noted the generation of NAE-like artifacts (Fontana et al., 1995) that could be separated by HPLC and have recommended the use of chlorofordmethanol in prefer- ence to ethyl acetate or diethyl ether and the use of silica gel minicolumns for cleanup before GCMS analyses.

The use of a reverse-phase cartridge as a separation device in this study clearly separated potentially inter- fering substances in the analysis of NAE. Figure 4A shows the separation of standards of NAE and N-ArPE by the reverse-phase cartridge. Figure 4B shows that there is no detectable NAE in a chloroform extract of rat brain when eluted fractions are evaporated to dryness

and reacted with PFBzCl and PFPA. Kempe et al. (1996) observed that if brain extracts were treated with base, large quantities of NAE appeared. Likewise, we ob- served that exposure of brain extracts to either base or acid significantly enhanced the appearance of NAE. Large quantities of NAE can be generated chemically (nonenzymatically) during sample preparation. To dem- onstrate that the compounds generating NAE chemically were separated by the reverse-phase cartridge procedure, isolated fractions were further analyzed (Fig. 4C). There are substances that are more polar than NAE (fractions 1-3) and that can form NAE on exposure to acid, pos- sibly by amidation or hydrolysis and rearrangement.

Studies in brain (Sugiura et al., 1996~; Cadas et al., 1997), cultured brain neurons (Di Marzo et al., 1994; Cadas et al., 1996b), and testis (Sugiura et al., 1996b) have strongly suggested that NAE biosynthesis involves Ca2+-dependent formation of N-ArPE as an intermediate molecule and subsequent enzymatic cleavage of NAE from the preexisting N-ArPE located in plasma mem- brane by a phosphodiesterase (Cadas et al., 1996b). In supporting this NAE biosynthetic pathway, the existence of N-ArPE in brain and testis has now been demonstrated in several studies (Sugiura et al., 1996~; Cadas et al., 1997; Schmid et al., 1998). The methods used for the N-ArPE identification include isolation of N-ArPE from tissue extracts with column chromatography or TLC,

J. Neurochem., Vol. 72, No. 5, 1999

1966 H.-Y. T. YANG ET AL.

TABLE 4. Comparison of reported concentrations of N-ArPE

N-ArPE ( p m o h g of Tissue N-ArPE amount protein)" Reference

Wistar rat brain Sea urchin P. lividus

Wistar rat brain Sprague-Dawley rat kidney

Sprague-Dawley rat testis

Mouse uterus

ovaries

Peritoneal macrophages

Bivalve molluscs

50.2 ? 27.8 p m d g 6 ? 5 pmol/g of

22 2 16 pmoVg 7.77 pmollpmol of

phosphorus 90.87 pmol/pmol of

phosphorus 1.5-11.9 pmol/pmol

of phosphorus 3.44.1 pmollpmol

of phosphorus Not detectable to 1.9

? 0.9 pmol/g

tissue

0.50 % 0.3 (est) Estimate not possible

0.22 2 .16 (est) 0.65 (est)

Estimate not possible

Estimate not possible

Estimate not possible

0.02 2 0.1 (est)

Sugiura et al. (1996~) Bisogno et al. (1997)

Cadas et al. (1997) Deutsch et al. (1997)

Schmid et al. (1998)

Schmid et al. (1997b)

Schmid et al. (1997~)

Sepe et al. (1998)

Assuming 10% protein per wet weight

digestion of N-ArPE fraction with bacterial phospho- lipase D, purification of NAE generated by column chro- matography or TLC, and then subsequent analysis of NAE by G C N S or HPLC after derivatization. These methods require large amounts of tissue and would be difficult to apply to measurements of small amounts of N-ArPE in discrete tissue regions but have been applied to a range of animal tissue organs as summarized in Table 4. The methods used in this study permit measure- ment of N-ArPE levels in discrete brain regions.

A chemical method that cleaves N-ArPE in a manner analogous to phospholipase D and then derivatizes the NAE portion of N-ArPE appears to be an ideal way to quantify integratively the family of endogenous N-ArPEs after reverse-phase C- 18 cartridge fractionation (Fig. 4). The identities of the R, acyl or alkenyl group at C1 and R, acyl group at C2 of glycero-3-phospho-(N-arachido- noy1)ethanolamide (Fig. 1) remain unknown. The par- tially purified N-ArPE fraction is free from NAE and other lipid material that may interfere with the GCMS analysis. The N-ArPE fraction from the C-18 cartridge fractionation was further confirmed by phospholipase D digestion followed with HPLC isolation and GCMS analysis of the released NAE. Furthermore, the amounts of N-ArPE estimated after phospholipase D digestion (- 1.59 and 1.19 pmol/mg of protein, assuming that 1 g of tissue contains - 100 mg of protein) are similar to that determined by the direct GCMS method (Table 1).

In this study, the levels of N-ArPE determined were not found to be influenced by the three different extrac- tion solvents used (Table 1). Addition of dioleoy1-N- ArPE-,H, as an internal standard during homogenization is largely responsible for these consistent measurements of N-ArPE levels. The level of N-ArPE detected in rat brain in this study (Tables 1 and 2) is -3.5 times higher than that reported by one group (Sugiura et al., 19968) and 8.5 times higher than that reported in another study (Cadas et al., 1997). The values in Table 1 are for the rat brain having no cerebellum and medulla, whereas the

levels reported in Table 4 are for whole rat brains. This difference may partially account for the higher N-ArPE levels in Table 1, as indicated by the regional differences reported in Table 2. The different separation and chem- ical derivatization procedures may further contribute to the higher levels that we detected. N-Acylethanolamide phospholipids were found to accumulate during postde- capitation ischemia; this effect was more pronounced in developing rat brain than in adult rat brain (Natarajan et al., 1986). To minimize this postdecapitation effect, rat brains from adult rats were removed, frozen in dry ice immediately after decapitation, and kept frozen until processed for N-ArPE analysis. Thus, the accumulation of N-ArPE due to the postdecapitation effect should be very minimal; consequently, it is not likely to be a serious contributing factor for the higher N-ArPE values found in this study.

N-ArPE was found to be widely distributed in the brain and spinal cord. Significantly ( p < 0.05), but not much, higher concentrations of N-ArPE were detected in extracts from striatum, hippocampus, and accumbens than in those from hypothalamus, frontal cortex, and cerebellum (Table 2). It is interesting that high densities of cannabinoid receptors were also reported in hip- pocampus and striatal areas (Herkenham, 1991), al- though an exact correlation was not found between the distribution of receptors and that of N-ArPE. A compar- ison of the brain regional levels of N-ArPE with those of NAE may be useful in understanding the metabolic re- lationship between these two lipids. Because NAE levels were below detection limits in this study, no direct cor- relation is possible. The NAE concentration differences in various brain regions that have been reported by others are not marked (Felder et al., 1996; Koga et al., 1997), although both N-ArPE and NAE were found to be rela- tively abundant in hippocampus. Several cannabinoids have been found to exert analgesic effects through ac- tions at the level of the spinal cord (Gilbert, 1981 ; Yaksh, 198 1). These observations and the high level of N-ArPE

J. Neurochem., Vol. 72, No. 5 , 1999

GC/MS ANALYSIS OF NAE AND N-ArPE 1967

found in the rat spinal cord suggest that the spinal cord may be an important site for exploring the functional role of endogenous NAE in antinociception.

In the two peripheral tissues we analyzed, a much higher concentration of N-ArPE was found in the testis in comparison with the spleen. Others have also reported high concentrations of N-ArPE in testis (Table 4). These observations correlate well with the distribution of N- acyltransferase, the enzyme that catalyzes the biosynthe- sis of N-ArPE by transferring an arachidonate group from the sn-1 carbon of phospholipids to the amino group of phosphatidylethanolamine. N-Acyltransferase enzyme was found to be much more abundant in the rat testis than in spleen (Cadas et al., 1997).

The N-ArPE analysis developed in this study is highly sensitive and relatively simple and can be readily applied to analyze the small amounts of N-ArPE in various regions of a single rat brain. The method should provide a useful tool to study the regulation of N-ArPE and, in turn, NAE biosynthesis in vivo.

REFERENCES

Berdyshev E. V., Boichot E., and Lagente V. (1996) Anandamide-a new look on fatty acid ethanolamides. J. Lipid Mediaf. Cell Signal. 15, 49-67.

Bisogno T., Ventriglia M., Milone A., Mosca M., Cimino G., and Di Marzo V. (1997) Occurrence and metabolism of anandamide and related acyl-ethanolamides in ovaries of the sea urchin Paracen- trotus lividus. Biochim. Biophys. Acta 1345, 338-348.

Cadas H., Gaillet S., Beltramo M., Venance L., and Piomelli D. (19964 Biosynthesis of an endogenous cannabinoid precursor in neurons and its control by calcium and CAMP. J. Neurosci. 16, 3934-3942.

Cadas H., Schinelli S., and Piomelli D. (19966) Membrane localization of N-acylphosphatidylethanolamine in central neurons: studies with exogenous phospholipases. J. Lipid Mediat. Cell Signal. 14, 63-70.

Cadas H., di Tomaso E., and Piomelli D. (1997) Occurrence and biosynthesis of endogenous cannabinoid precursor, N-arachido- noyl phosphatidylethanolamine, in rat brain. J. Neurosci. 17,

Deutsch D. G. and Chin S. A. (1993) Enzymatic synthesis and degra- dation of anandamide, a cannabinoid receptor agonist. Biochem. Pharmacol. 46,79 1-796.

Deutsch D. G., Goligorsky M. S., Schmid P. C., Krebsbach R. J., Schmid H. H., Das S. K., Dey S. K., Arreaza G., Thorup C., Stefan0 G., and Moore L. C. (1997) Production and physiological actions of anandamide in the vasculature of the rat kidney. J. Clin. Invest. 100, 1538-1546.

Devane W. A. and Axelrod J. (1994) Enzymatic synthesis of anand- amide, an endogenous ligand for the cannabinoid receptor, by brain membranes. Proc. Natl. Acad. Sci. USA 91, 6698-6701.

Devane W. A,, Hanus L., Breuer A,, Pertwee R. G., Stevenson L. A,, Griffin G., Gibson D., Mandelbaum A,, Etinger A,, and Mechou- lam R. (1992) Isolation and structure of a brain constituent that binds to the cannabinoid receptor. Science 258, 1946-1949.

Di Marzo V., Fontana A,, Cadas H., Schinelli S., Cimino G., Schwartz J. C., and Piomelli D. (1994) Formation and inactivation of endogenous cannabinoid anandamide in central neurons. Narure 372, 686-691.

Felder C. C. and Glass M. (1998) Cannabinoid receptors and their endogenous agonists. Annu. Rev. Pharmacol. Toxicol. 38, 179- 200.

Felder C. C., Briley E. M., Axelrod J., Simpson J. T., Mackie K., and Devane W. A. (1993) Anandamide, an endogenous cannabimi- metic eicosanoid, binds to the cloned human cannabinoid receptor

1226-1242.

and stimulates receptor-mediated signal transduction. Proc. Natl. Acad. Sci. USA 90,7656-7660.

Felder C. C., Nielsen A,, Briley E. M., Palkovits M., Priller J., Axelrod J., Nguyen D. N., Richardson J. M., Riggin R. M., Koppel G. A., Paul S. M., and Becker G . W. (1996) Isolation and measurement of the endogenous cannabinoid receptor agonist, anandamide, in brain and peripheral tissues of human and rat. FEBS Lett. 393, 23 1-235.

Fontana A,, Di Marzo V., Cadas H., and Piomelli D. (1995) Analysis of anandamide, an endogenous cannabinoid substance, and of other natural N-acylethanolamines. Prostaglandins Leukot. Es- sent. Fatty Acids 53, 301-308.

Gilbert P. E. (1981) A comparison of THC, nantradol, nabilone, and morphine in the chronic spinal dog. J. Clin. Pharmacol. 21

Giuffrida A. and Piomelli D. (1998) Isotope dilution GCMS determi- nation of anandamide and other fatty acylethanolamides in rat blood plasma. FEBS Lett. 422, 373-376.

Herkenham M. (1991) Characterization and localization of cannabinoid receptors in brain: an in vitro technique using slide-mounted tissue sections. NIDA Res. Monogr. 112, 129-145.

Imamura S. and Horiuti Y. (1979) Purification of Streptomyces chro- mofuscus phospholipase D by hydrophobic affinity chromatogra- phy on palmitoyl cellulose. J . Biochem. (Tokyo) 85, 79-95.

Kempe K., Hsu F.-F., Bohrer A,, and Turk J. (1996) Isotope dilution mass spectrometric measurements indicate that arachidonyleth- anolamide, the proposed endogenous ligand of the cannabinoid receptor, accumulates in rat brain tissue post mortem but is con- tained at low levels in or is absent from fresh tissue. J. Biol. Chem. 271, 17287-17295.

Koga D., Santa T., Fukushima T., Homma H., and Imai K. (1997) Liquid chromatographic-atmospheric pressure chemical ionization mass spectrometric determination of anandamide and its analogs in rat brain and peripheral tissues. J. Chromatogr. B Biomed. Sci.

Kruszka K. K. and Gross R. W. (1994) The ATP- and CoA-indepen- dent synthesis of arachidonoylethanolamide. J . Biol. Chem. 269, 14345-14348.

Lowry 0. H., Rosebrough N. J., Farr A. L., and Randall R. J. (1951) Protein measurement with the Folin phenol reagent. J . Biol. Chem.

Natarajan V., Schmid P. C., and Schrnid H. H. (1986) N-Acylethanol- amine phospholipid metabolism in normal and ischemic rat brain. Biochim. Biophys. Acta 878, 32-41.

Schmid H. H. O., Schmid P. C., and Natarajan V. (1990) N-Acylated glycerophospholipids and their derivatives. Prog. Lipid Res. 29,

Schmid H. H. O., Schmid P. C., and Natarajan V. (1996) The N- acylation-phosphodiesterase pathway and cell signalling. Chem. Phys. Lipids 80, 133-142.

Schmid P. C., Reddy P. V., Natarajan V., and Schmid H. H. 0. (1983) Metabolism of N-acylethanolamine phospholipids by a mamma- lian phosphodiesterase of the phospholipase D type. J. Biol. Chem.

Schmid P. C., Deli E., and Schmid H. H. (1995~) Generation and remodeling of phospholipid molecular species in rat hepatocytes. Arch. Biochem. Biophys. 319, 168-176.

Schmid P. C., Krebsbach R. J., Perry S. R., Dettmer T. M., Maasson J. L., and Schmid H. H. 0. (1995b) Occurrence and postmortem generation of anandamide and other long-chain N-acylethanol- amines in mammalian brain. FEBS Lett. 375, 117-120.

Schmid P. C., Kuwae T., Krebsbach R. .I., Schmid H. H., Matsuda S., Kanemitsu N., Nakamura A,, Mirnura Y., Ueda N., Kurahashi Y., and Yamamoto S. (1997~) Anandamide and other N-acylethanol- amines in mouse peritoneal macrophages. Metabolism of anand- amide, an endogenous cannabinoid receptor ligand, in porcine ocular tissues. Chem. Phys. Lipids 87, 103-1 10.

Schmid P. C., Paria B. C., Krebsbach R. J., Schmid H. H. O., and Dey S. K. (1997b) Changes in anandamide levels in mouse uterus are associated with uterine receptivity for embryo implantation. Proc. Natl. Acad. Sci. USA 94,4188-4192.

(SUPPI.), 311s-319s.

A&. 690,7-13.

193, 265-273.

1-43.

258,9302-9306.

J. Neurochem.. Vol. 72, No. 5, 1999

I968 H.-Y. T. YANG ET AL.

Schmid P. C., Schwindenhammer D., Krebshach R. J., and Schmid H. H. (1998) Alternative pathways of anandamide biosynthesis in rat testes. Chem. Phys. Lipids 92, 21-35.

Sepe N., De Petrocellis L., Montanaro F., Cimino G., and Di Marzo V. (1998) Bioactive long chain N-acylethanolamines in five species of edible bivalve molluscs. Possible implications for mollusc physiology and sea food industry. Biochim. Biophys. Acta 1389, 101-111.

Shaikh N. A. (1994) Assessment of various techniques for the quanti- tative extraction of lysophospholipids from myocardial tissues. Anal. Biochem. 216, 313-321.

Sugiura T., Kondo S., Sukagawa A,, Tonegawa T., Nakane S., Ya- mashita A,, Ishima y., and Waku K. (1996a) Transacylase-medi- ated and phosphodiesterase-mediated synthesis of N-arachido- noylethanolamine, an endogenous cannabinoid-receptor ligand, in rat brain microsomes. Comparison with synthesis from free ara- chidonic acid and ethanolamine. Eur. J. Biochem. 240, 53-62.

Sugiura T., Kondo S., Sukagawa A,, Tonegawa T., Nakane S., Ya- mashita A., and Waku K. (19966) Enzymatic synthesis of anan- damide, an endogenous cannabinoid receptor ligand, through N- acylphosphatidylethanolamine pathway in testis: involvement of Ca(2 +)-dependent transacylase and phosphodiesterase activities. Biochem. Biophys. Res. Comrnun. 218, 113-1 11.

Ueda N., Kurahashi Y., Yamamoto S., and Tokunaga T. (1995) Partial purification and characterization of the porcine brain enzyme hydrolyzing and synthesizing anandamide. J. Biol. Chem. 270, 23823-23827.

Vogel Z . , Barg J., Levy R., Saya D., Heldman E., and Mechoulam R. (1993) Anandamide, a brain endogenous compound, interacts specifically with cannabinoid receptors and inhibits adenylate cyclase. J. Neurochem. 61, 352-355.

Yaksh T. L. (1981) The antinociceptive effects of intrathecally admin- istered levonantradol and desacetyllevonantradol in the rat. J. Clin. Pharmacol. 21 (Suppl.), 334s-340s.

J. Neurochem., Val. 72, No. 5, 1999