Anandamide levels in human female reproductive tissues: Solid-phase extraction and measurement by ultraperformance liquid chromatography tandem mass spectrometry

Analytical Biochemistry 400 (2010) 155–162

Contents lists available at ScienceDirect

Analytical Biochemistry journal homepage: www.elsevier.com/locate/yabio

Anandamide levels in human female reproductive tissues: Solid-phase extraction and measurement by ultraperformance liquid chromatography tandem mass spectrometry Timothy H. Marczylo, Patricia M.W. Lam, Akwasi A. Amoako, Justin C. Konje * Endocannabinoid Research Group, Reproductive Sciences Section, Department of Cancer Studies and Molecular Medicine, University of Leicester, Robert Kilpatrick Clinical Sciences Building, Leicester Royal Infirmary, P.O. Box 65, Leicester, Leicestershire LE2 7LX, UK

a r t i c l e

i n f o

Article history: Received 25 August 2009 Received in revised form 13 December 2009 Accepted 16 December 2009 Available online 22 December 2009 Keywords: Endocannabinoid Anandamide Solid-phase extraction

a b s t r a c t Anandamide (N-arachidonoylethanolamide), a bioactive lipid, is reported to play a role in pregnancy maintenance and parturition. Our aims were to (1) evaluate AEA levels at the human maternal:fetal interface and (2) validate the use of solid-phase extraction of AEA from tissues. AEA was analyzed in cord and maternal blood, amniotic fluid, placenta, and fetal membranes collected during Caesarean section (n = 14). Extraction efficiencies were 42 and 36% for the placenta and the fetal membranes, respectively. Tissue AEA was quantified using an isotope-dilution method and UPLC-ESI-MS/MS giving intra- and inter-day variability for tissues spiked with 0.2, 1, and 5 pmol/g AEA of less than 12%. Accuracy for these spiked samples was between 95% and 103% for fetal membranes and between 99% and 114% for placenta. Mean AEA concentrations were 2.72 ± 1.04 pmol/g for placenta and 1.19 ± 0.68 pmol/g for fetal membranes, and 0.93 ± 0.28, 0.88 ± 0.33, 0.77 ± 0.30, and 0.06 ± 0.04 nM for maternal, umbilical vein, and umbilical artery plasma and amniotic fluid. Higher AEA concentrations were found in placenta compared to fetal membranes (P < 0.0001), in umbilical vein compared with umbilical artery (P = 0.0015), and in plasma from maternal circulation compared with umbilical artery (P = 0.0152). The relevance of these changes in AEA concentrations at the maternal:fetal interface requires further investigation. Ó 2010 Elsevier Inc. All rights reserved.

Cannabis (active component D9-tetrahydrocannabinol) use is associated with pregnancy complications such as low birth weight, fetal growth restriction, placental abruption, preterm birth, stillbirths, and spontaneous miscarriage [1–4]. These complications are induced via manipulation of signaling pathways involving the cannabinoid receptors (CB1 and CB2). Endocannabinoids are bioactive lipids representing the endogenous ligands for these cannabinoid receptors. The first of this family of eicosanoid-related compounds to be identified was N-arachidonoylethanolamide (anandamide, AEA)1 (Fig. 1) which was extracted from porcine brain [5]. Other endocannabinoids that have since been identified include 2-arachidonoylglycerol (2AG), noladin ether, virodhamine, and Narachidonoyl dopamine [6–9]. The endocannabinoids, their receptors, and the enzymes responsible for their synthesis such as N-acylphosphatidylethan* Corresponding author. Fax: +44 116 252 5846. E-mail address: [email protected] (J.C. Konje). 1 Abbreviations used: AEA, N-arachidonoylethanolamide (anandamide); AEA-d8, deuterated anandamide; ECS, endocannabinoid system; FAAH, fatty acid amide hydrolase; LOD, limit of detection; LOQ, limit of quantification; LPE, liquid-phase extraction; NAPE-PLD, N-acylphosphatidylethanolamide phospholipase D; SPE, solidphase extraction; UPLC-MS/MS, ultraperformance liquid chromatography tandem mass spectrometry. 0003-2697/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2009.12.025

olamide phospholipase D (NAPE-PLD) and their degradation such as fatty acid amide hydrolase (FAAH) and monoacyl glycerol lipase (MAGL) compose the endocannabinoid system (ECS). Recently numerous roles have been proposed for the involvement of the ECS in fertilization, blastocyst development, implantation, pregnancy maintenance, and parturition [10–13]. AEA levels measured at the non-implantation site of the uterus in pregnant mice are the highest identified in mammalian tissue to date and down-regulation of AEA levels at the implantation site is associated with increased uterine receptivity [14]. Maternal plasma AEA levels decline progressively during pregnancy but become significantly elevated in labor [10,15]. Additionally, increased plasma AEA concentrations have been shown to be predictive of miscarriage [13,16]. While AEA levels fall, rat placental FAAH activity increases throughout pregnancy [17]. Together these observations support the hypothesis that low systemic levels of AEA are required for normal pregnancy progression [16]. The ECS has been identified in the placenta [18,19], which suggests an important role for the ECS in this organ and in labor. Despite these observations of the apparent importance of the ECS at the maternal:fetal interface, AEA concentrations have not been extensively investigated in human reproductive tissues. Measurement of AEA in neurological tissues has a number of pitfalls

156

Anandamide levels in human reproductive tissues / T.H. Marczylo et al. / Anal. Biochem. 400 (2010) 155–162

ally analyzed within 1 week of sample collection. Random placenta and fetal membrane biopsy samples were taken and dissected into approximately 1-g aliquots before washing repeatedly with volumes of ice-cold PBS until the PBS ran clear to remove blood contamination. The tissues were then either analyzed for AEA immediately or flash-frozen in liquid nitrogen and stored at 80 °C prior to analysis. Anandamide extraction and measurement *

*

*

*

*

*

*

*

Fig. 1. Structure of AEA showing sites of deuteration in AEA-d8 (), proposed site of fragmentation (- - -) and protonated ethanolamine daughter ion.

that have made the published data highly variable [20]. For example, AEA concentrations increase postmortem in brain tissues [21,22] and ex vivo in whole blood [23] and silica-based extraction techniques give falsely high AEA concentrations due to deuterium transfer from the internal standard [24]. Our aims here were to develop a reliable and reproducible method of AEA extraction and measurement in reproductive tissues and to use these techniques to determine AEA concentrations at the human maternal:fetal interface at term. Materials and methods Chemicals Anandamide and octadeuterated anandamide (AEA-d8, purity >95% AEA and >99% deuterium incorporation; Fig. 1) were obtained from Cayman Chemicals (Ann Arbor, MI). HPLC grade acetonitrile, chloroform, formic acid, phosphoric acid, and methanol were purchased from Fisher Scientific (Loughborough, UK) and HPLC grade water was produced using a water purification system (Maxima ELGA, ELGA, High Wycombe, UK). Subjects and sample collection Women (n = 14) due to undergo elective Caesarean section, including two with twins, were recruited and gave informed consent to take part in this study which had been approved by the Leicester and Rutland Research Ethics Committee. As soon as the baby was separated from the cord, a segment was isolated between clamps and blood collected separately from the artery and the vein into 9-ml Sarstedt-monovette EDTA collection tubes (Sartedt, Leicester, UK). The pH and gases (pCO2 and pO2) in the blood were immediately measured using a blood gas analyzer (Roche Diagnostics, Basle, Switzerland) to ensure that the samples were from the appropriate vessels. The umbilical samples were accepted only when there was a difference in the pH and gas pressures of the two samples. Maternal blood was collected at the same time as the umbilical blood samples. The collected blood samples were placed on ice and transferred to the laboratory where plasma was obtained by centrifugation of whole blood at 1200g for 30 min at 4 °C. Prior to incision of the amniotic sac, amniotic fluid was collected into sterile universal bottles for transport to the laboratory. This was also centrifuged (1200g for 20 min at 4 °C) and the supernatant recovered. The plasma and amniotic fluid were then transferred into 7-ml Kimble scintillation vials (Kinesis, St. Neots, Cambridge, UK) and either analyzed immediately for AEA or stored as 1-ml aliquots at 80 °C. The stored samples were usu-

Frozen tissue was ground to a fine powder with a pestle and mortar in liquid nitrogen. Ground tissue (100 mg) was suspended in 2 ml aqueous o-phosphoric acid (2.5% v/v), spiked with internal standard (12.5 pmol/g tissue), vortexed thoroughly, and homogenized with a glass Potter-Elvehjem homogenizer. Homogenate was centrifuged at 1500g for 30 min at 4 °C and the supernatant decanted. AEA was extracted from tissue supernatant, plasma, and amniotic fluid as described previously [25] using solid-phase extraction, dried under a constant stream of nitrogen, and then resuspended in acetonitrile (80 ll) prior to analysis by ultraperformance liquid chromatography-ESI-tandem mass spectrometry (UPLC-ESI-MS/MS) in positive-ion mode using multiple reaction monitoring as described previously by Lam et al. [15] with slight modifications to the transitions employed and the capillary voltage. Transitions employed for the detection of AEA and AEA-d8 were m/z 348.25 > 61.9 and m/z 356.25 > 62.9, respectively. Dwell time was 20 ms, capillary voltage 0.70 kV, cone voltage 16 V, and collision energy 17 eV. The daughter ion of AEA-d8 at m/z 62.9 is consistent with previous observations using this standard [15,27] and implies that there is an intermolecular transfer of a deuterium from the fatty acid to the ethanolamine side chain. AEA was quantified by an isotope-dilution method with reference to a 7-point standard curve (0.23–19 nM) using relative response = peak area/ (AEA-d8 area/concentration AEA-d8) and Quanlynx software (Waters Inc.) as described previously [15]. For comparison, AEA was also extracted using the chloroform:methanol extraction method described by Koga et al. [26]. Validation of tissue extraction method To determine the reproducibility and reliability of the SPE extraction method compared with solvent extraction, placenta and fetal membrane samples (n = 3) from the same patient were first extracted on three separate days using both the SPE method described above and a previously described solvent extraction method [26]. Accuracy and precision for the SPE extraction were determined using aliquots of placenta and fetal membrane (100 mg) spiked with exogenous AEA at low (0.2 pmol/g), median (1 pmol/g), and high (5 pmol/g) physiological concentrations. Three extractions of each concentration were undertaken on three separate days concurrently with three unspiked samples. Accuracy was determined as (mean observed level)/(mean endogenous level + spiked AEA concentration)  100%. Precision was determined as the relative standard deviation (RSD = [standard deviation/ mean]  100%). The peak area observed for the AEA-d8 internal standard was used to calculate the efficiency of each extraction technique in both tissues by comparison with an equal concentration of nonextracted AEA-d8 in acetonitrile (n = 9). LOD and LOQ as determined by spectrogram peaks yielding mean signal-to-noise ratios of 3 and 10, respectively, were calculated using AEA-d8 spiked into placenta (100 mg, n = 3) and fetal membranes (100 mg, n = 3). The presence of endogenous levels of AEA necessitates the employment of AEAd8 as a surrogate for the determination of LOD and LOQ. Ion suppression by components of the extracted tissue matrix may contribute to the observation of false negative results.

157

Anandamide levels in human reproductive tissues / T.H. Marczylo et al. / Anal. Biochem. 400 (2010) 155–162

Although the use of deuterated internal standard should adjust for ion suppression, this phenomenon was investigated in more detail. Aliquots of placenta and fetal membranes (100 mg) were homogenized as described above but not spiked with AEA-d8 until after elution from the SPE cartridge (n = 5) to determine the contribution of endogenous AEA. At the same time identical aliquots were spiked with AEA-d8 and AEA (1 pmol/g) for quantification of the sum of endogenous and spiked AEA. These samples were evaporated to dryness and resuspended in acetonitrile. For comparison the same concentrations of AEA-d8 and AEA as spiked into the samples were made in acetonitrile and the effects of the extracted matrix on measured AEA determined using the equation [total AEA]  [endogenous AEA]/[AEA standard]  100%.

Table 2 Intra- and inter-day accuracy and precision for placenta and fetal membrane tissues spiked with AEA. AEA added to tissues (pmol/g) Fetal membranes

Placenta

Effects of sample handling on quantified tissue AEA concentration The effects of ex vivo tissue processing time on tissue AEA concentrations were assessed for both placenta and fetal membranes. Fresh tissues were processed as described under Materials and methods and then AEA was extracted immediately or tissues were kept on ice for 30 min, 2 h, and 4 h prior to extraction. For each time point, three tissue samples were extracted and the AEA concentration was measured in triplicate. To investigate the effects of poststorage tissue handling on AEA levels, placental and fetal membrane samples taken from different patients (n = 3) were processed in triplicate either following direct homogenization of frozen tissue or following thawing on ice for 30 min and 2 h. In addition, AEA concentrations were measured in tissues taken from the same women either immediately from frozen or following 1, 2, or 3 further freeze–thaw cycles. Placental samples were allowed to thaw completely on ice before refreezing in liquid nitrogen and storage at 80 °C for at least 12 h for each cycle. AEA was extracted as described above and measured by UPLC-MS/MS. To assess the effect of long-term storage on AEA levels in tissues placenta (n = 5) and fetal membrane (n = 5) biopsy samples were reassessed for AEA concentration following at least 4 months storage at 80 °C. Statistics Paired analysis of AEA concentrations in matched samples was undertaken using the Student t test (Prism 5, GraphPad Software Inc.). Results Validation of AEA extraction from tissues using solid phase Validation parameters are presented in Tables 1 and 2. Extraction of AEA from tissues using SPE or LPE did not give rise to any significant changes in measured AEA concentrations. Furthermore, there was little difference in the extraction efficiency of each method for both tissues, though in placenta the extraction efficiency

Intra-day (n = 3) Mean SD RSD (%) Accuracy (%) Inter-day (n = 9) Mean SD RSD (%) Accuracy (%) Intra-day (n = 3) Mean SD RSD (%) Accuracy (%) Inter-day (n = 9) Mean SD RSD (%) Accuracy (%)

0

0.2

1.0

5.0

0.85 0.07 7.70 n.a.

1.03 0.08 8.10 98.0

1.90 0.03 1.32 103

5.66 0.06 1.02 96.8

0.89 0.08 9.11 n.a.

1.04 0.10 9.94 95.8

1.80 0.17 9.67 95.6

5.60 0.23 4.16 95.1

2.07 0.10 4.94 n.a.

2.25 0.03 1.29 99.1

3.13 0.36 11.57 102

7.46 0.36 4.81 106

2.02 0.08 4.04 n.a.

2.54 0.31 12.23 114

3.21 0.38 11.95 106

7.45 0.32 4.30 106

n.a., not applicable.

was slightly but nonsignificantly better using SPE (42.3% compared with 34.1%). Limits of detection and quantification were determined using AEA-d8 and found to be comparable between the two extraction methods (Table 1). Detection limits for the placenta were slightly higher (187.5 fmol/g) than for fetal membranes (125 fmol/g). LOD and LOQ for the two tissues using both LPE and SPE extraction methods were 187.5 and 312.5 fmol/g for the placenta and 125 and 187.5 fmol/g for fetal membranes, respectively (Table 1 and Fig. 2). The measurements of AEA in fetal membranes and placenta following SPE were further validated using samples spiked with three concentrations of AEA (0.2, 1.0, and 5.0 pmol/g) representing the lower, median, and upper concentrations observed for these human tissues in our studies. The results are presented in Table 2 and demonstrate excellent degrees of accuracy and precision both within day and inter-day. RSD for all concentrations were <10% for fetal membranes and <12.5% for placenta and accuracy was between 95% and 103% and between 99% and 114% for fetal membranes and placenta, respectively. No ion suppression was observed with either placenta or fetal membranes. Relative concentrations of AEA in these tissues compared with comparable samples in acetonitrile alone gave values of 100.2% and 99.5%, respectively, for fetal membranes and placenta.

Effects of sample handling on quantified AEA levels in placenta and fetal membranes Delayed tissue processing following collection had a significant effect on AEA concentration (Fig. 3). AEA levels in the placenta were modestly elevated after 30 min (P = 0.0232) and 2 h (P = 0.0482) and increased more significantly (P = 0.0033) 4 h after

Table 1 Validation of two techniques for the extraction of AEA from placenta and fetal membranes. Parameter

AEA content (pmol/g, mean ± SD) Concentration variability (RSD) Extraction efficiency (%) LOD (AEA-d8 fmol/g tissue) LOQ (AEA-d8 fmol/g tissue)

Placenta

Fetal membrane

SPE

LPE

SPE

LPE

0.97 ± 0.17 17.7 42.3 ± 8.9 187.5 312.5

1.05 ± 0.20 19.4 34.1 ± 7.0 187.5 312.5

0.46 ± 0.06 13.5 35.9 ± 8.9 125 187.5

0.59 ± 0.14 23.3 35.6 ± 20.1 125 187.5

158

Anandamide levels in human reproductive tissues / T.H. Marczylo et al. / Anal. Biochem. 400 (2010) 155–162

100

A

0 100

1.50

1.60

1.70

1.80

1.90

2.00

2.10

2.20

2.30

1.50

1.60

1.70

1.80

1.90

2.00

2.10

2.20

2.30

1.50

1.60

1.70

1.80

1.90

2.00

2.10

2.20

2.30

1.50

1.60

1.70

1.80

1.90

2.00

2.10

2.20

2.30

1.50

1.60

1.70

1.80

1.90

2.00

2.10

2.20

2.30

1.70

1.80

1.90

2.00

2.10

2.20

2.30

1.80

1.90

2.00

2.10

2.20

2.30

B

0 100

C

0 100

D 0 100

E 0 100

F

0 1.50

100

1.60

G 0 100

1.50

1.60

1.70

1.50

1.60

1.70

H 0

1.80

1.90

2.00

2.10

2.20

2.30

Retention time (min) Fig. 2. UPLC-MS/MS spectrograms for AEA (A and E) and AEA-d8 (B–D and F–H) following extraction from placenta (A–D) and fetal membranes (E–H). Spectrograms representing extractions at the LOQ and LOD for AEA-d8 are presented for placenta (312.5 fmol/g, C, 187.5 fmol/g, D) and fetal membranes (187.5 fmol/g, G, 125 fmol/g, H). Analyses were conducted as described under Materials and methods and transitions used for the detection of AEA and AEA-d8 were m/z 348.25 > 61.9 and m/z 356.25 > 62.9, respectively.

collection (Fig. 3). In contrast fetal membrane AEA concentrations were only elevated significantly after 4 h (P = 0.0417). Delayed extraction following removal of tissues from 80 °C storage was responsible for a modest but significant (P = 0.01 after 2 h) increase in AEA concentration in both placenta and fetal membranes (Fig. 4). When placenta samples were kept on ice for 4 h, AEA concentration continued to increase with concentrations approximately threefold higher than that of immediately processed tissue (data not shown). If tissues underwent two freeze– thaw cycles there was a significant (P < 0.005) rise in AEA levels in both placenta and fetal membranes (Fig. 5). Following three cycles, AEA was further increased (P < 0.001) in placenta but not fetal membranes (Fig. 5). AEA levels in both placenta and fetal membranes were modestly but not significantly (P > 0.05) decreased following long-term (>4 month) storage at 80 °C (Fig. 6A).

tively, for the placenta and fetal membranes. Analysis of the matched pairs of tissue samples showed a significantly (P < 0.0001) higher AEA content in the placenta compared with fetal membranes. The mean concentrations of AEA in maternal plasma and plasma from the umbilical artery and vein were 0.93 ± 0.28, 0.77 ± 0.30, and 0.88±0.33 pmol/ml, respectively. AEA concentrations in amniotic fluid (0.06 ± 0.04 pmol/ml) were measureable but significantly (P < 0.0001) lower than plasma AEA concentrations. Individual values for the 14 volunteers are presented in Table 3 (maternal blood could not be obtained from two subjects). Paired statistical analyses found plasma AEA levels to be significantly higher in the umbilical vein compared with the umbilical artery (Fig. 7, P = 0.0015). Maternal plasma AEA levels were also significantly higher (P = 0.0152) than those in the umbilical artery (Fig. 7). There was no statistically significant difference between levels of AEA in maternal plasma and plasma from umbilical vein.

AEA concentrations at the maternal:fetal interface Discussion The quantified AEA concentrations in the placenta (n = 11) and fetal membranes (n = 10) are presented in Fig. 6B. The mean AEA concentrations were 2.72 ± 1.04 and 1.19 ± 0.68 pmol/g, respec-

To the best of our knowledge, this is the first time that AEA has been quantified in human placenta and fetal membranes. Tissue

159

Anandamide levels in human reproductive tissues / T.H. Marczylo et al. / Anal. Biochem. 400 (2010) 155–162

4.0

**

** 3.0

2.0

** 1.0

0.0 1

2

3

4

Ex-vivo processing time (h)

**

4

2

**

*

9.0

*

A

6.0

3.0

0.0 0.0 2.5

1.0

2.0

*

B

2.0 1.5 1.0 0.5 0.0 0.0

1.0

1

2

3

Freeze-Thaw cycles

Fig. 3. Ex vivo changes in AEA concentration in placenta and fetal membrane. Placenta (}) and fetal membrane (j) were processed in triplicate at different times following collection. Results are presented as mean ± standard deviation of experiments conducted in triplicate.  represents values significantly (P < 0.005) different from all other time points and  represent values significantly (P < 0.05) different from time 0 only.

AEA concentration (pmol/g)

***

0

0

AEA concentration (pmol/g)

6

* AEA concentration (pmol/g)

AEA concentration (pmol/g)

5.0

2.0

Tissue Processing Time (h) Fig. 4. Effect of poststorage handling of placenta (A) and fetal membranes (B). Each line represents tissue from a different individual and each data point represents the mean ± standard deviation of three observations.  represents data significantly (P = 0.01) different from time 0.

extraction efficiencies using SPE, based on the retrieval of spiked deuterated AEA, were modest (Table 1) but were comparable to those observed here using an established LPE technique [26] (Table 1). The lower extraction efficiencies observed here compared with previously published LPE methods are likely as a consequence of the lower, more physiological, concentrations of AEA-d8 (1.25 pmol/sample, 12.5 pmol/g) utilized for these calculations compared with previous reports. For example, Koga et al. [26] used

Fig. 5. Effect of freeze–thaw cycles on tissue AEA concentrations in placenta (}) and fetal membranes (j). Results are presented as mean ± standard deviation of experiments conducted in triplicate. Significant differences to 1 freeze–thaw cycle are represented by  (P < 0.05),  (P < 0.005), and  (P < 0.001).

25, 125, and 250 pmol AEA/sample to achieve recoveries of 68%, 67%, and 73%. Despite these lower recoveries, we observed a marked improvement in LOD and LOQ compared with those published previously. The LODs reported here of 187.5 and 125 fmol/ g for placenta and fetal membrane are equivalent to 1.64 and 1.09 fmol on column. Richardson et al. [27] reported LOQ for AEA extracted from rat brain to be 10 pmol/g and an LOD of 25 fmol on column; Williams et al. [28] reported similar LOD values also using rat brain, whereas Kingsley and Marnett employed silver ion chromatography to improve the LOD to 14 fmol on column [29]. The improved sensitivity afforded by the use of UPLC also shortened the run time compared with previously published methods [25]. The use of deuterated AEA, in addition to acting as an internal standard for the quantification of AEA, also helped confirm the identity of the observed peak as AEA by coelution. Virodhamine (O-arachidonoylethanolamine) demonstrates an identical fragmentation to AEA and consequently may contribute to false positives. However, virodhamine and AEA have distinctly different retention times of 1.52 and 1.65 min, respectively [15]. Furthermore, the specific fragmentation of AEA to yield a protonated ethanolamine daughter ion, and coelution of sample peaks with both deuterated and nondeuterated standards, suggests strongly that the observed results are not the consequence of false positives. Although occasional samples yielded very small peaks with retention times approximating that of virodhamine (Fig. 2A), coelution studies demonstrated that virodhamine did not contribute to these minor peaks (data not shown). Accuracy and precision in tissue samples spiked with 3 physiologically relevant concentrations of AEA (Table 2) were comparable or better than previous published methods using brain [28,29]. No comparable data are available for reproductive tissues. Whether using LPE and SPE, AEA concentrations were comparable. SPE, however, is expected to produce cleaner samples which are likely to decrease any matrix effects and increase the column lifetime, reduce background noise, and increase sensitivity. Previously LPE extracts were cleaned using silica SPE; however, evidence suggests that this methodology is compromised by deuterium transfer from the internal standard [24]. This flaw may be subverted by the use of 15N-, 17O-, or 13C-labeled standards. Here we observe modest, but significant, increases in AEA levels in both the placenta and the fetal membranes. However, significant rises in AEA were only observed in fetal membranes after 4 h

160

Anandamide levels in human reproductive tissues / T.H. Marczylo et al. / Anal. Biochem. 400 (2010) 155–162 Table 3 AEA concentrations in maternal and umbilical plasma and in amniotic fluid. Patient ID

AEA concentration (nM) Maternal

Umbilical artery

Umbilical vein

Amniotic fluid

1 (twins)

1.71 1.71 1.38 0.99 1.00 0.59 0.60 1.01 0.62 0.87 0.99 0.92 0.92   1.18  

1.13 0.75 0.82 1.23 0.90 0.72 0.96 0.69 0.48 0.32 0.46 0.79 0.80 0.57 0.58 0.74

0.91 1.00 1.08 1.42 1.02 0.83 1.23 0.72 0.62 0.38 0.51 0.76 0.94 0.73 0.81 0.84

0.04 0.06 0.02 0.03 0.06 0.05 0.13 0.18 0.06 0.04 0.04 0.06 0.02 0.05 0.02 0.08

2 3 4 5 6 7 8 9 10 11 (twins) 12 13 14  

No sample collected.

P=0.0015 P=0.0152

AEA concentration (nM)

1.8 1.5 1.2 0.9 0.6 0.3 0.0

Maternal plasma

Umbilical vein plasma

Unbilical artery plasma

Amniotic fluid

Fig. 7. Umbilical and maternal plasma and amniotic fluid AEA concentrations. Data are presented as individual mean plasma concentrations from 16 births including two sets of twins. P values following paired analyses (t test) are presented. Fig. 6. AEA concentrations present in placenta and fetal membranes. (A) The effects of prolonged storage at 80 °C (A) in placenta and fetal membranes (n = 5); data are presented as mean ± standard deviation. (B) Concentrations present in a cohort of women undergoing Caesarean section. Data are presented as mean ± standard deviation of 13 placenta and 12 fetal membrane samples representing 11 matched pairs. *Tissue levels in placenta were significantly (P = 0.0003) higher than paired fetal membrane.

(Fig. 3). In addition, delays in processing tissues, poststorage at 80 °C, also caused a significant elevation in AEA concentrations (Fig. 4). AEA concentrations were also increased significantly (P = 0.0006) in the placenta following repeated freeze–thaw cycles (Fig. 5). No similar increase was observed with fetal membranes. The increases in tissue AEA concentrations following delayed ex vivo processing are consistent with previous observations in human and rodent brain [21,22,31] and human plasma [23] each of which demonstrated a modest increase in AEA over the first few hours. In the brain, this increase has been shown to continue with time [21,22,31]. Our data therefore add to that already available for brain and indicate that for stabilization of AEA levels in tissues, handling time should be kept to a minimum after collection and if samples are to be stored they should be processed directly from

frozen. We observed no degradation of AEA levels so we suggest that FAAH inhibitors are not required. The absence of reliable NAPE-PLD inhibitors restricts investigations into how these agents may improve stabilization. When AEA levels were investigated following storage at 80 °C for at least 4 months modest but nonsignificant (P > 0.05) decreases in concentrations were observed. To the best of our knowledge no comparable, previous studies on AEA stability in tissues have been undertaken; however, three recently conducted studies have investigated AEA stability in human plasma up to 2 months [34–36]. One study [34] found AEA levels to be rather variable after 1 day, 1 week, and 2 months after storage at 80 °C, though increases in AEA concentration were not time dependent and after 2 months AEA values increased by 21.6% compared with freshly analyzed plasma. In the other two studies no significant changes in AEA concentrations were observed after 1 month at 80 °C [35] or 2 months at 70 °C [36]. AEA levels in freshly processed placenta were significantly higher than in fetal membranes (Fig. 6B). Although AEA levels have not been measured in these tissues before, the higher concentration in placenta is in keeping with previous observations that FAAH, the

Anandamide levels in human reproductive tissues / T.H. Marczylo et al. / Anal. Biochem. 400 (2010) 155–162

enzyme which degrades AEA, is more strongly expressed in fetal membranes [18]. Tissue levels of AEA reported here are low compared with levels reported from several tissues in experimental animals including the mouse uterus which has been reported to have the highest concentration (1–7 lg/g or 2.87–20.09 nmol/g) of AEA yet identified [14,26,27,32]. Whether this represents a tissue or a species-specific effect or is a consequence of different sample handling processes requires further examination. Increased FAAH expression in the early stages of pregnancy, both in maternal mononuclear cells [16] and in placenta [19], is associated with a good prognosis for pregnancy maintenance. These observations suggest that the placenta acts as a barrier to AEA transport to the fetus during the early stages of pregnancy. Given the rise in circulating AEA levels at term and the absence of placental FAAH reported by Acone et al. [30], we wanted to investigate the levels of AEA to which the fetus is exposed. The overall significantly higher AEA levels in the umbilical vein compared to the artery suggest that there is either transport across the placenta or placental production of AEA. Either way, term placenta may not be a barrier to AEA transport as previously suggested. However, in 7 women AEA levels were lower in umbilical vein while 3 women had higher levels and 3 showed no difference (Table 1). This variability suggests that the overall picture is far more complicated and requires further investigations. In agreement with our previously published data [25], AEA levels were slightly higher in plasma from the umbilical vein compared with that from the umbilical artery (P = 0.0015). When twin births were excluded, this significance increased to P < 0.0001 and maternal AEA was less significantly higher than that of the umbilical artery (P = 0.0486). Consequently, these initial observations from this small study suggest that the fetus utilizes some of the AEA passed from mother or placenta to fetus. AEA levels in amniotic fluid were very low (0.02–0.18 nM) in contrast to previous observations [33] (8 nM). There are several possible explanations for this discrepancy. First, the samples were collected at different gestational ages; our samples were collected at term while in the earlier study, the samples were collected at around 16 weeks gestation. At term amniotic fluid largely comprises fetal urine which may have a dilution effect on AEA concentrations. We have previously shown that urine contains no measurable AEA [25]. Furthermore, the previous study utilized silica clean up which has been shown to give falsely high values because of deuterium transfer from the internal standard [24]. Conclusion AEA can be reproducibly quantified in human placenta and fetal membranes using SPE and UPLC-MS/MS. Levels in both tissues are in the low the picomole per gram range which is significantly lower than previously observed in other animal and human genital tissues. Delays in processing and freeze–thaw cycles can affect the reproducibility of AEA determination especially in the placenta where concentrations rise significantly. AEA is also detectable in cord blood, suggesting that the placenta does not restrict fetal exposure to AEA. Plasma AEA in the umbilical artery was modestly but significantly lower when compared with matched umbilical vein levels. How these observations impact on physiological changes at term and what effects these changes in AEA concentrations have on the fetus require further detailed examination. Acknowledgments This work was funded in part by generous grants from PerkinElmer and the British United Provident Association (BUPA) Foundation.

161

References [1] C.S. Conner, Marijuana and alcohol use in pregnancy, Drug Intell. Clin. Pharm. 18 (1984) 233–234. [2] C.C. Felder, M. Glass, Cannabinoid receptors and their endogenous agonists, Annu. Rev. Pharmacol. Toxicol. 38 (1998) 179–200. [3] L.L. Hatch, M.B. Bracken, Effects of marijuana use in pregnancy on fetal growth, Am. J. Epidemiol. 124 (1986) 986–993. [4] B. Zuckerman, D.A. Frank, R. Hingson, H. Amaro, S.M. Levenson, H. Kayne, S. Parker, R. Vinci, K. Aboagye, L.E. Fried, H. Cabral, R. Timperi, H. Bauchner, Effects of maternal marijuana and cocaine use on fetal growth, N. Engl. J. Med. 320 (1989) 762–768. [5] W.A. Devane, L. Hanus, A. Breuer, R.G. Pertwee, L.A. Stevenson, G. Griffin, D. Gibson, A. Mandelbaum, A. Etinger, R. Mechoulam, Isolation and structure of a brain constituent that binds to the cannabinoid receptor, Science 258 (1992) 1946–1949. [6] R. Mechoulam, S. Ben-Shabat, L. Hanuš, M. Ligumsky, N.E. Kaminski, A.R. Schatz, A. Gopher, S. Almog, B.R. Martin, D.R. Compton, R.G. Pertwee, G. Griffin, M. Bayewitch, J. Marg, Z. Vogel, Identification of an endogenous 2monoglyceride, present in canine gut, that binds to cannabinoid receptors, Biochem. Pharmacol. 50 (1995) 83–90. [7] A.C. Porter, J.M. Sauer, M.D. Knierman, G.W. Becker, M.J. Berna, J. Bao, G.G. Nomikos, P. Carter, F.P. Bymaster, A. Baker Leese, C.C. Felder, Characterization of a novel endocannabinoid, virodhamine, with antagonist activity at the CB1 receptor, J. Pharmacol. Exp. Ther. 301 (2002) 1020–1024. [8] T. Bisogno, D. Melck, M.Y. Bodrov, N.M. Gretskaya, W. Bezuglov, L. De Petrocellis, V. Di Marzo, N-Acyl-dopamines: novel synthetic CB1 cannabinoid-receptor ligands and inhibitors of anandamide inactivation with cannabimimetic activity in vitro and in vivo, Biochem. J. 351 (2000) 817–824. [9] L. Hanus, S. Abu-Lafi, E. Fride, A. Breuer, Z. Vogel, D.E. Shalev, I. Kustanovich, R. Mechoulam, 2-Arachidonyl glyceryl ether, an endogenous agonist of the cannabinoid CB1 receptor, Proc. Natl. Acad. Sci. USA 98 (2001) 3662–3665. [10] O.M.H. Habayeb, A.H. Taylor, M.D. Evans, M.S. Cooke, D.J. Taylor, S.C. Bell, J.C. Konje, Plasma levels of the endocannabinoid anandamide in women-A potential role in pregnancy maintenance and labor?, J Clin. Endocrinol. Metab. 89 (2004) 5482–5487. [11] H. Wang, S.K. Dey, M. Maccarrone, Jekyll and Hyde: two faces of cannabinoid aboring in female fertility, Endocr. Rev. 27 (2006) 427–448. [12] A.H. Taylor, C. Ang, S.C. Bell, J.C. Konje, The role of the endocannabinoid system in gametogenesis, implantation and early pregnancy, Hum. Reprod. Update 13 (2007) 501–513. [13] O.M.H. Habayeb, A.H. Taylor, M. Finney, M.D. Evans, J.C. Konje, Plasma anandamide concentration and pregnancy outcome in women with threatened miscarriage, J. Am. Med. Assoc. 299 (2008) 1135–1136. [14] P.C. Schmid, B.C. Paria, R.J. Krebsbach, H.H.O. Schmid, S.K. Dey, Changes in anandamide levels in mouse uterus are associated with uterine receptivity for embryo implantation, Proc. Natl. Acad. Sci. USA 94 (1997) 4188–4192. [15] P.M. W Lam, T.H. Marczylo, M. El-Talatini, M. Finney, V. Nallendran, A.H. Taylor, J.C. Konje, Ultra performance liquid chromatography tandem mass spectrometry method for the measurement of anandamide in human plasma, Anal. Biochem. 380 (2008) 195–201. [16] M. Maccarrone, T. Bisogno, H. Valensise, N. Lazzarin, F. Fezza, C. Manna, V. Di Marzo, A. Finazzi-Agro, Low fatty acid hydrolase and high anandamide levels are associated with failure to achieve an ongoing pregnancy after IVF and embryo transfer, Mol. Hum. Reprod. 8 (2002) 188–195. [17] M. Cella, G.F. Leguizamón, M.S. Sordelli, M. Cervini, T. Guadagnoli, M.L. Ribeiro, A.M. Franchi, M.G. Farina, Dual effects of anandamide on rat placenta nitric oxide synthesis, Placenta 29 (2008) 699–707. [18] B. Park, H.M. Gibbons, M.D. Mitchell, M. Glass, Identification of the CB1 cannabinoid receptor and fatty acid amide hydrolase (FAAH) in the human placenta, Placenta 24 (2003) 990–995. [19] O.M.H. Habayeb, A.H. Taylor, S.C. Bell, D.J. Taylor, J.C. Konje, Expression of the endocannabinoid system in human first trimester placenta and its role in trophoblast proliferation, Endocrinology 149 (2008) 5052–5060. [20] T.H. Marczylo, J.C. Konje, Quantification of endocannabinoids in clinical samples: pitfalls and future directions, in: J.K. Lang (Ed.), Handbook on Mass Spectrometry: Instrumentation, Data and Analysis, and Applications, Nova Science, New York. [21] P.C. Schmid, R.J. Krebsbach, S.R. Perry, T.M. Dettmer, J.L. Maasson, H.H.O. Schmid, Occurrence and post-mortem generation of anandamide and other long-chain N-acylethanolamines in mammalian brain, FEBS Lett. 375 (1995) 117–120. [22] K. Kempe, F.F. Hsu, A. Bohrer, J. Turk, Isotope dilution mass spectrometric measurements indicate that arachidonoylethanolamide, the proposed endogenous ligand of the cannabinoid receptor, accumulates in rat brain tissue post mortem but is contained at low levels in or is absent from fresh tissue, J. Biol. Chem. 271 (1996) 17287–17295. [23] M. Vogeser, D. Hauer, S.C. Azad, E. Huber, M. Storr, G. Schelling, Release of anandamide from blood cells, Clin. Chem. Lab. Med. 44 (2006) 488–491. [24] S. Hardison, S.T. Weintraub, A. Giuffrida, Quantification of endocannabinoids in rat biological samples by GC/MS: technical and theoretical considerations, Prostaglandins Other Lipid Mediat. 81 (2006) 106–112. [25] T.H. Marczylo, P.M.W. Lam, V. Nallendran, A.H. Taylor, J.C. Konje, A solid-phase method for the extraction and measurement of anandamide from multiple human biomatrices, Anal. Biochem. 384 (2009) 106–113.

162

Anandamide levels in human reproductive tissues / T.H. Marczylo et al. / Anal. Biochem. 400 (2010) 155–162

[26] D. Koga, T. Santa, T. Fukushima, H. Homma, K. Imai, Liquid chromatographicatmospheric pressure chemical ionization mass spectrometric determination of anandamide and its analogs in rat brain and peripheral tissues, J. Chromatogr. B 690 (1997) 7–13. [27] D. Richardson, C.A. Ortori, V. Chapman, D.A. Kendall, D.A. Barrett, Quantitative profiling of endocannabinoids and related compounds in rat brain using liquid chromatography-tandem electrospray ionization mass spectrometry, Anal. Biochem. 360 (2007) 216–226. [28] J. Williams, J. Wood, L. Pandarinathan, D.A. Karanian, B.A. Bahr, P. Vouros, A. Makriyannis, Quantitative method for the profiling of the endocannabinoid metabolome by LC-atmospheric pressure chemical ionization-MS, Anal. Chem. 79 (2007) 5582–5593. [29] P.J. Kingsley, L.J. Marnett, Analysis of endocannabinoids by Ag+ coordination tandem mass spectrometry, Anal. Biochem. 314 (2003) 8–15. [30] G. Acone, E. Trabucco, N. Colacurci, L. Cobeliis, Low type I cannabinoid receptor levels characterize placental villous in labouring delivery, Placenta 30 (2009) 203–205. [31] M. Palkovits, J. Harvey-White, J. Liu, Z.S. Kovacs, M. Bobest, G. Lovas, A.G. Bago, G. Kunos, Regional distribution and effects of postmortem delay on endocannabinoid content of the human brain, Neuroscience 152 (2008) 1032–1039.

[32] H. Wang, H. Xie, X. Sun, P.J. Kingsley, L.J. Marnett, B.F. Cravatt, S.K. Dey, Differential regulation of endocannabinoid synthesis and degradation in the uterus during embryo implantation, Prostaglandins Other Lipid Mediat. 83 (2007) 62–74. [33] H. Schuel, L.J. Burkman, J. Lippes, K. Crickard, E. Forester, D. Piomelli, A. Giuffrida, N-Acylethanolamines in human reproductive fluids, Chem. Phys. Lipids 121 (2002) 211–227. [34] M.G.J. Balvers, K.C.M. Verhoeckx, R.F. Witkamp, Development and validation of a quantitative method for the determination of 12 endocannabinoids and related compounds in human plasma using liquid chromatography-tandem mass spectrometry, J. Chromatogr. B 877 (2009) 1583–1590. [35] A. Thomas, G. Hopfgartner, C. Giroud, C. Staub, Quantitative and qualitative profiling of endocannabinoids in human plasma using a triple quadrupole linear ion trap mass spectrometer with liquid chromatography, Rapid Commun. Mass Spectrom. 23 (2009) 629–638. [36] A.A. Zoerner, F.M. Gutzki, M.T. Suchy, B. Beckmann, S. Engeli, J. Jordan, D. Tsikas, Targeted stable-isotope dilution GC-MS/MS analysis of the endocannabinoid anandamide and other fatty acid ethanol amides in human plasma, J. Chromatogr. B 877 (2009) 2909–2923.