G Model
ARTICLE IN PRESS
CHROMB-18846; No. of Pages 10
Journal of Chromatography B, xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb
Review
Analysis of eicosanoids by LC-MS/MS and GC-MS/MS: A historical retrospect and a discussion夽 Dimitrios Tsikas ∗ , Alexander A. Zoerner Institute of Clinical Pharmacology, Hannover Medical School, Carl-Neuberg-Str. 1, 30625 Hannover, Germany
a r t i c l e
i n f o
Article history: Received 17 December 2013 Received in revised form 10 March 2014 Accepted 13 March 2014 Available online xxx We dedicate this article to Prof. Jürgen C. Frölich’ 75th anniversary. Jürgen C. Frölich is a pioneer in analytical, biological and clinical pharmacological eicosanoids research, an enthusiastic advocate of the gas chromatography-tandem mass spectrometry technology, and our mentor. Keywords: Artefactual formation Blood Plasma Tandem mass spectrometry Urine Validation
a b s t r a c t Eicosanoids are a large family that derives from arachidonic acid, i.e., eicosatetraenoic acid. Prominent members include prostaglandins, thromboxane and leukotrienes. They are biologically highly active lipid mediators and play multiple physiological roles. GC-MS/MS has played a pivotal role in the identification and quantification of eicosanoids in biological samples. This technology generated a solid knowledge of their analytical chemistry, biochemistry, physiology and pharmacology. Since about a decade, GC-MS and GC-MS/MS are increasingly displaced by the seemingly more simple, rapid and powerful LC-MS/MS in the area of instrumental analysis of physiological substances, drugs and their metabolites. In this article, we review and discuss LC-MS/MS methods published over the last decade from the perspective of the GC-MS/MS user. Our analysis revealed that the shift from the adult GC-MS/MS to the youthful emerging LC-MS/MS technology in eicosanoid analysis is associated with several important challenges. Known pitfalls and problematic issues discovered by eicosanoid pioneers by using GC-MS/MS are often ignored by LC-MS/MS users. Established reference values and intervals provided by GC-MS-based methods are not considered properly in developing and validating LC-MS/MS methods. Virtually, there is a belief in the unlimited capability of the LC-MS/MS technique in eicosanoid analysis, a thought that simulates analytical certainty. LC-MS/MS users should profit from the plethora of solid knowledge acquired from the use of GC-MS/MS in eicosanoid analysis in basic and clinical research. © 2014 Elsevier B.V. All rights reserved.
Contents 1. 2. 3.
4. 5. 6.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis of eicosanoids by mass spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Recall past knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. A brief historical retrospection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Artefactual ex vivo formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Ex vivo stimulation of eicosanoids synthesis in blood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Sensitivity—basal concentrations and limits of detection and quantitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Method validation and comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biological and pharmacological applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
夽 This paper is part of the special issues ACIDS edited by A.A. Zoerner and D. Tsikas. ∗ Corresponding author. Tel.: +49 511 532 3984; fax: +49 511 532 2750. E-mail address: [email protected] (D. Tsikas). http://dx.doi.org/10.1016/j.jchromb.2014.03.017 1570-0232/© 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: D. Tsikas, A.A. Zoerner, J. Chromatogr. B (2014), http://dx.doi.org/10.1016/j.jchromb.2014.03.017
00 00 00 00 00 00 00 00 00 00 00 00
G Model CHROMB-18846; No. of Pages 10
ARTICLE IN PRESS D. Tsikas, A.A. Zoerner / J. Chromatogr. B xxx (2014) xxx–xxx
2
Nomenclature ADMA AEA BAL CCS CID COX CSF CYP EBC (EC)NICI
Asymmetric dimethylarginine Arachidonic acid ethanol amide (anandamide) broncho alveolar liquid Cell culture supernatant Collision-induced dissociation Cyclooxygenase Cerebrospinal fluid Cytochrome P450 Exhaled breath condensate (Electron-capture) negative-ion chemical ionization EETs Epoxyeicosatrienoic acids European Medicines Agency EMA ESI Electrospray ionization FDA Federal Drug Administration GC-MS Gas chromatogram-mass spectrometry GC-MS/MS Gas chromatogram-tandem mass spectrometry Hydroxyeicosanoic acid HEA HETE(s) Hydroxyeicosatetraenoic acid(s) IAC Immunoaffinity chromatography LC-MS Liquid chromatogram-mass spectrometry LC-MS/MS Liquid chromatogram-tandem mass spectrometry LLOD Lower limit of detection LLOQ Lower limit of quantitation Lipoxygenase LO LT Leukotriene MD Microdialysate MOX Methoxyamine Nordihydroguaiaretic acid NDGA PAR Peak area ratio Pentafluorobenzyl PFB PFB-Br Pentafluorobenzyl bromide Prostaglandin PG PGE-MUM Prostaglandin E major urinary metabolite Prostaglandin H synthase PGH SRM Selected-reaction monitoring TMS Trimethylsilyl TMSOH Trimethylsilanol Thromboxane Tx
Scheme 1. A part of the arachidonic acid cascade illustrating four major pathways. (1) The cyclooxygenase (COX) or prostaglandin H synthase (PGHS) pathway includes thromboxane (TxA2 ) and prostacyclin (PGI2 ) which are spontaneously converted to their stable analogs TxB2 and 6-keto-PGF1␣ , respectively. (2) The lipoxygenase (LO) pathway includes several hydroxyeicosatetraenoic acids (HETEs), leukotriene B4 (LTB4 ) and the cysteinyl leukotrienes C4 (LTC4 ), LTD4 and LTE4 . For simplicity, the 20-HETE is not shown. (3) The cytochrome P450 (CYP) family generates various epoxyeicosatrienoic acids (EETs). (4) Many eicosanoids including F2 -isoprostanes are generated both from free arachidonic acid and from arachidonic acid esterified to lipids. For the sake of simplicity, the pathways are shown for the free form of arachidonic acid.
blood sampling. This is important because artefactual formation may be abundant and exceed several-fold the concentrations that prevail in vivo (see below). Therefore, primary circulating eicosanoids such as leukotriene B4 (LTB4 ), thromboxane B2 (TxB2 ), i.e., the stable hydrolysis product of TxA2 , and 6-keto-prostaglandin F1␣ (6kPGF1␣ ), i.e., the stable hydrolysis product of prostacyclin (PGI2 ), are not useful biomarkers of eicosanoid synthesis in health and disease or in pharmacotherapy. The problem of ex vivo artefactual formation of eicosanoids has been successfully overcome by measuring in the circulation certain metabolites, commonly their dehydro- and dinor-metabolites such as 11-dehydro-TxB2 , (11dh-TxB2 ), 2,3-dinor-TxB2 and 2,3-dinor-6kPGF1␣ (Scheme 2). Such eicosanoid metabolites cannot be formed ex vivo, at least not to a degree that could compromise analysis. Nevertheless, eicosanoid synthesis is preferentially assessed by measuring eicosanoids in urine rather than in blood (i.e., plasma or serum) [1]. 2. Analysis of eicosanoids by mass spectrometry
1. Introduction Arachidonic acid (5Z,8Z,11Z,14Z-eicosatetraenoic acid) is the common precursor of a variety of biologically highly active lipid mediators. Enzymatically formed intermediate products, such as leukotriene A4 (LTA4 ) and prostaglandin H2 (PGH2 ), are further converted by the action of different enzymes in a cascadelike manner to generate numerous, in part structurally closely related carboxyl-groups containing substances, collectively named eicosanoids (Scheme 1). Arachidonic acid, free and esterified to lipids, is also oxidized by non-enzymatic reactions to form numerous isomeric compounds, the iso-eicosanoids, such as isoleukotrienes, F2 -isoprostanes, and iso-thromboxane [1,2]. Primary eicosanoids and iso-eicosanoids circulate in blood and are extensively metabolized by - and -oxidation. Eicosanoids, i.e., both primary eicosanoids and their metabolites, are present in blood and urine at concentrations in the pM to nM range. Wide structural variety and extremely low quantity render unequivocal identification and reliable quantification of these substances in biological samples incomparably challenging. In addition, both eicosanoids and iso-eicosanoids can be formed artefactually ex vivo, notably during
Historically, GC-MS and GC-MS/MS have played a central role both in the identification and in the quantification of eicosanoids and iso-eicosanoids in biological samples [1]. Our current knowledge of the biology, pharmacology and analytical chemistry of eicosanoids is primarily based on the application of the GC-MS and GC-MS/MS technologies over the past three to four decades. GC-MS and GC-MS/MS approaches have also been proved useful to validate immunology-based assays for eicosanoids [1]. However, RIA and ELISA methods for eicosanoids are generally artefact-prone, lack specificity and may produce hardly interpretable results. Nevertheless, they have been and are still frequently used in experimental and clinical research. Due to great advances in the LC-MS/MS technology, especially over the past two decades, this relatively new approach is increasingly used both in biomedical and life sciences [1–4], and in clinical chemistry [5]. The main reason for the invasion is its putative simplicity, basically due to the omission of multiple derivatization steps. Derivatization in the GC-MS and GC-MS/MS analysis of eicosanoids is absolutely required because eicosanoids have two or three thermally labile chemical functionalities, i.e.,
Please cite this article in press as: D. Tsikas, A.A. Zoerner, J. Chromatogr. B (2014), http://dx.doi.org/10.1016/j.jchromb.2014.03.017
G Model CHROMB-18846; No. of Pages 10
ARTICLE IN PRESS D. Tsikas, A.A. Zoerner / J. Chromatogr. B xxx (2014) xxx–xxx
3
3. Recall past knowledge The chemical structures of selected primary eicosanoids and their index metabolites discussed below are shown in Scheme 2. 3.1. A brief historical retrospection
Scheme 2. Selected stable primary eicosanoids (left panel) and their corresponding major metabolites resulting from - and/or -oxidation from their precursors (right panel). PGE-MUM, major urinary metabolite of E prostaglandins. See also Scheme 1.
carboxylic, hydroxylic and keto groups (Scheme 2). This lends the LC-MS/MS approach the potential for online and high-throughput analysis. These particular attributes attract the interest of analysts who increasingly apply LC-MS/MS both in experimental and in clinical research. With regard to ionization and more so to collisioninduced dissociation (CID), LC-MS/MS and GC-MS/MS seem to produce structurally closely related ions (Scheme 3). At present, LC-MS/MS seems to rapidly and massively displace GC-MS and GC-MS/MS in various analytical disciplines especially including eicosanoids analysis in biological samples. Yet, “blind” trust in the attractive LC-MS/MS technology and non-consideration of several decades back-dating acquirement of basic analytical knowledge also hold the potential for fallacious analytical certainty. Neglecting previous knowledge, including recognition and avoidance of pre-analytical pitfalls and generation of references values and intervals for eicosanoids established by GC-MS and GC-MS/MS pioneers in this field, is a risky progress in the LC-MS/MS-based analysis of circulating eicosanoids. Development of LC-MS/MS-based methods for the quantitative analysis of eicosanoids and other biomarkers in biological samples and validation of these methods by established GC-MS- and GC-MS/MS methods is highly advisable, if not even an obligation. But, consideration of findings revealed by GC-MS and GC-MS/MS over the past decades seems to be neglected by some LC-MS/MS users in the area of eicosanoids analysis. We wish to discuss this issue in more detail below.
Prostacyclin (prostaglandin I2 , PGI2 ) was discovered in the vascular endothelium in 1976 [6]. At physiological pH, prostacyclin is highly unstable and hydrolyses spontaneously to the stable 6-keto-PGF1␣ [7]. The major urinary metabolites of prostacyclin and 6-keto-PGF1␣ infused into healthy volunteers have been identified as 2,3-dinor-6-keto-PGF1␣ and 13,14-dihydro-2,3-dinor6-keto-PGF1␣ [8]. From the two dinor metabolites of prostacyclin, 2,3-dinor-6-keto-PGF1␣ was found to be its major metabolite in the urine of healthy humans [9] and became generally accepted as the index parameter of prostacyclin [1] (Scheme 2). With regard to circulating prostanoids, the sampling technique has been found to influence the basal level of 6-keto-PGF1␣ in plasma due to artefactual formation. Instead of 6-keto-PGF1␣ , measurement of circulating 6,15-diketo-13,14-dihydro-PGF1␣ , a major circulating metabolite of prostacyclin in man [10,11], is preferred because it cannot be formed in vitro. Thromboxane A2 was discovered in 1975 in platelets [12]. At physiological pH, TxA2 is extremely labile and is non-enzymatically converted to TxB2 . 2,3-dinor-TxB2 has been identified as the major urinary metabolite of systemically administered TxB2 in man [13]. In total, twenty urinary metabolites of TxB2 have been identified, among them 2,3-dinor-TxB2 and 11-dehydro-TxB2 were the most abundant [14]. Quantification of urinary 2,3-dinor-TxB2 and 11dehydro-TxB2 is a useful measure of in vivo production of TxB2 in man [1] (Scheme 2). 11-dehydro-TxB2 has been shown to be an appropriate quantitative index metabolite of TxA2 both in plasma [15,16] and in urine [17]. Leukotriene B4 was discovered in 1979 as a dihydroxy metabolite (5,12-dihydroxy-eicosatetraenoic acid) of arachidonic acid by rabbit peritoneal polymorphonuclear leucocytes [18–20]. A biologically active B4 -isoleukotriene has been found to be formed by free radical-catalyzed peroxidation of glycerophospholipids independent of 5-LO [21]. Cysteinyl leukotrienes are known since the 1930s as ‘slow reacting substances of anaphylaxis’ (SRS-A). Structural elucidation of the first member of cysteinyl leukotrienes, i.e., LTC4 , was reported in 1979 [18,22,23]. Catalytical saturation of the double bonds and desulphurisation of the cysteinyl moiety of cysteinyl leukotrienes by noble catalysts and hydrogen gas yields quantitatively 5-hydroxy-eicosanoic acid (5-HEA), which can be utilized to measure cysteinyl leukotrienes by GC-MS [24–26]. LTE4 has been shown to be the major metabolite of LTC4 in urine in humans and in the monkey [27–30] (Scheme 2). Urinary LTE4 has been turned out to be a useful index metabolite for whole body production of cysteinyl leukotrienes in humans [1]. 3.2. Artefactual ex vivo formation Electron-capture negative-ion chemical ionization (ECNICI) permits the most sensitive GC-MS and GC-MS/MS analysis of eicosanoids after proper derivatization with fluorine-rich reagents such as pentafluorobenzyl (PFB) bromide (Scheme 3). Hughes et al. quantitated LTB4 in serum of healthy humans by GC-MS as a PFB ester trimethylsilyl ether derivative [31]. When blood was collected into the 5-LO inhibitor nordihydroguaiaretic acid (NDGA), the LTB4 concentration was determined to be 10 ± 4 pg/mL [31]. By contrast, in blood that was not collected into NDGA, the LTB4 concentration in serum was ten-fold higher, indicating artefactual ex vivo LTB4 formation [31]. This finding was also observed by Blair et al. [32]. In
Please cite this article in press as: D. Tsikas, A.A. Zoerner, J. Chromatogr. B (2014), http://dx.doi.org/10.1016/j.jchromb.2014.03.017
G Model CHROMB-18846; No. of Pages 10 4
ARTICLE IN PRESS D. Tsikas, A.A. Zoerner / J. Chromatogr. B xxx (2014) xxx–xxx
Scheme 3. Analysis of prostaglandin E2 (PGE2 ) by LC-MS/MS (left panel) and GC-MS/MS (right panel). For sensitive GC-MS/MS analysis of PGE2 in the negative-ion chemical ionization mode (NICI), PGE2 must be derivatized in a three-step process. By means of pentafluorobenzyl bromide (PFB-Br) the pentafluorobenzyl (PFB) ester of PGE2 is prepared. By means of methoxyamine (MOX) hydrochloride the keto group of PGE2 is methoximated and yields the syn- and anti-isomers (not shown). Finally, the remaining two hydroxyl groups are etherified by means of N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA). Electrospray ionization in the negative mode (ESI-) of native PGE2 free acid yields the carboxylate of PGE2 , i.e., to generate the anion with the mass-to-charge (m/z) ratio of 351. Collision-induced dissociation (CID) of this precursor ion yields several product ions. The anion m/z 271 [M − PFB−2 × H2 O − CO2 ]− is frequently used for the quantitative analysis of PGE2 in the selected-reaction monitoring (SRM) mode. In GC-MS/MS, NICI of the fully derivatized PGE2 generates the most intense precursor anion at m/z 351 [M − PFB]− . CID of this precursor ion yields several product ions. The anion m/z 268 [M − PFB − 2 × TMSOH − CH3 OH–CO2 ]− is frequently used for the quantitative GC-MS/MS analysis of PGE2 in the SRM mode. TMSOH, trimethylsilanol.
human plasma, LTB4 concentration was reported to be about half the serum concentration in the same volunteers [31], suggesting LTB4 formation during clotting. By GC-MS/MS in the ECNICI mode, we found that LTB4 is present in plasma of healthy humans from blood that was not collected into NDGA at concentrations lower than 37 pg/mL [33], confirming the observations by Hughes et al. [31] and Blair et al. [32]. Thus, ex vivo artefactual formation of LTB4 is a major obstacle in the quantitative determination of circulating LTB4 , if not useful precautions are taken. Abundant ex vivo formation is not specific to LTB4 but also occurs to other eicosanoids in blood, notably TxA2 [34]. Upon activation, for instance during blood sampling and processing, large amounts of TxA2 may be produced artefactually by platelets. This particular circumstance can be utilized to measure TxA2 synthesis in platelets and the effects of drugs such as acetylsalicylic acid [35] (see also next section). By means of a sophisticated twodimensional LC-MS/MS and by collecting blood into indomethacin, a cyclooxygenase (COX) inhibitor, prostaglandin E2 (PGE2 ) and prostaglandin F2␣ (PGF2␣ ) were measured in human plasma at basal concentrations of 0.4–1.6 pg/mL and 2.9–11.4 pg/mL, respectively [36] (Table 1). These concentrations are very close to those measured by GC-MS/MS [34,40]. In contrast, Ferreiro-Vera et al. measured by LC-MS/MS PGF2␣ concentrations in human serum being almost 10,000 times higher, suggesting massive artefactual ex vivo formation due to the absence of any added COX inhibitor during blood sampling [41] (see also Ref. [17]). In conclusion, measurement of primary eicosanoids in the circulation should be generally regarded as an unreliable method to assess their synthesis in vivo due to potential and abundant ex vivo formation [34]. Another potential problem associated with eicosanoids analysis may be formation of numerous isomers which must be separated chromatographically. LTB4 and its isomers can easily be separated
each from other by capillary GC and normal phase LC [31], and this must be ensured in LC coupled to MS too [53]. 3.3. Ex vivo stimulation of eicosanoids synthesis in blood Stimulation of LTB4 synthesis in human blood, for instance by using calcium ionophore A23187, has been frequently investigated by GC-MS (see for instance Refs. [31,32]). Although there is considerable linear dependency of LTB4 synthesis in human whole blood on calcium ionophore A23187 concentration, remarkable stimulation of LTB4 synthesis can be achieved by only 1 M calcium ionophore A23187 [32], and this potent stimulation is unlikely to be achieved by disease or drugs. Therefore, utilization of extraordinary high calcium ionophore A23187 concentrations such as 14 M [53] is not very meaningful. Finally, because not all blood cells produce eicosanoids, their synthesis is best assessed by investigating blood cells that are primarily responsible for particular eicosanoids, e.g., platelets for TxA2 , and monocytes for LTB4 and cysteinyl leukotrienes [54]. 3.4. Sensitivity—basal concentrations and limits of detection and quantitation Table 1 summarizes reported concentrations of selected (mainly primary) eicosanoids in plasma, serum, and other biological fluids of healthy humans as measured by reported validated GC-MS, GC-MS/MS and LC-MS/MS methods. As can be clearly seen from Table 1, the concentration of eicosanoids in the circulation and other biological fluids of healthy humans lies in the lowest pg/mLrange. Most of the reported LC-MS/MS methods have several-fold higher lower limits of quantitation (LLOQ) values than reported GCMS/MS methods. For example, the LLOQ value of a recently reported
Please cite this article in press as: D. Tsikas, A.A. Zoerner, J. Chromatogr. B (2014), http://dx.doi.org/10.1016/j.jchromb.2014.03.017
G Model CHROMB-18846; No. of Pages 10
ARTICLE IN PRESS D. Tsikas, A.A. Zoerner / J. Chromatogr. B xxx (2014) xxx–xxx
5
Table 1 Reported concentrations for selected leukotrienes, prostaglandins and thromboxane in biological fluids of healthy humans as measured by approaches based on GC-MS, GC-MS/MS and LC-MS/MS. Eicosanoid
Basal concentration
Reference
Approach
Leukotriene B4 LTB4 LTB4 LTB4 LTB4 LTB4
ARTICLE IN PRESS
CHROMB-18846; No. of Pages 10
Journal of Chromatography B, xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb
Review
Analysis of eicosanoids by LC-MS/MS and GC-MS/MS: A historical retrospect and a discussion夽 Dimitrios Tsikas ∗ , Alexander A. Zoerner Institute of Clinical Pharmacology, Hannover Medical School, Carl-Neuberg-Str. 1, 30625 Hannover, Germany
a r t i c l e
i n f o
Article history: Received 17 December 2013 Received in revised form 10 March 2014 Accepted 13 March 2014 Available online xxx We dedicate this article to Prof. Jürgen C. Frölich’ 75th anniversary. Jürgen C. Frölich is a pioneer in analytical, biological and clinical pharmacological eicosanoids research, an enthusiastic advocate of the gas chromatography-tandem mass spectrometry technology, and our mentor. Keywords: Artefactual formation Blood Plasma Tandem mass spectrometry Urine Validation
a b s t r a c t Eicosanoids are a large family that derives from arachidonic acid, i.e., eicosatetraenoic acid. Prominent members include prostaglandins, thromboxane and leukotrienes. They are biologically highly active lipid mediators and play multiple physiological roles. GC-MS/MS has played a pivotal role in the identification and quantification of eicosanoids in biological samples. This technology generated a solid knowledge of their analytical chemistry, biochemistry, physiology and pharmacology. Since about a decade, GC-MS and GC-MS/MS are increasingly displaced by the seemingly more simple, rapid and powerful LC-MS/MS in the area of instrumental analysis of physiological substances, drugs and their metabolites. In this article, we review and discuss LC-MS/MS methods published over the last decade from the perspective of the GC-MS/MS user. Our analysis revealed that the shift from the adult GC-MS/MS to the youthful emerging LC-MS/MS technology in eicosanoid analysis is associated with several important challenges. Known pitfalls and problematic issues discovered by eicosanoid pioneers by using GC-MS/MS are often ignored by LC-MS/MS users. Established reference values and intervals provided by GC-MS-based methods are not considered properly in developing and validating LC-MS/MS methods. Virtually, there is a belief in the unlimited capability of the LC-MS/MS technique in eicosanoid analysis, a thought that simulates analytical certainty. LC-MS/MS users should profit from the plethora of solid knowledge acquired from the use of GC-MS/MS in eicosanoid analysis in basic and clinical research. © 2014 Elsevier B.V. All rights reserved.
Contents 1. 2. 3.
4. 5. 6.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis of eicosanoids by mass spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Recall past knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. A brief historical retrospection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Artefactual ex vivo formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Ex vivo stimulation of eicosanoids synthesis in blood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Sensitivity—basal concentrations and limits of detection and quantitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Method validation and comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biological and pharmacological applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
夽 This paper is part of the special issues ACIDS edited by A.A. Zoerner and D. Tsikas. ∗ Corresponding author. Tel.: +49 511 532 3984; fax: +49 511 532 2750. E-mail address: [email protected] (D. Tsikas). http://dx.doi.org/10.1016/j.jchromb.2014.03.017 1570-0232/© 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: D. Tsikas, A.A. Zoerner, J. Chromatogr. B (2014), http://dx.doi.org/10.1016/j.jchromb.2014.03.017
00 00 00 00 00 00 00 00 00 00 00 00
G Model CHROMB-18846; No. of Pages 10
ARTICLE IN PRESS D. Tsikas, A.A. Zoerner / J. Chromatogr. B xxx (2014) xxx–xxx
2
Nomenclature ADMA AEA BAL CCS CID COX CSF CYP EBC (EC)NICI
Asymmetric dimethylarginine Arachidonic acid ethanol amide (anandamide) broncho alveolar liquid Cell culture supernatant Collision-induced dissociation Cyclooxygenase Cerebrospinal fluid Cytochrome P450 Exhaled breath condensate (Electron-capture) negative-ion chemical ionization EETs Epoxyeicosatrienoic acids European Medicines Agency EMA ESI Electrospray ionization FDA Federal Drug Administration GC-MS Gas chromatogram-mass spectrometry GC-MS/MS Gas chromatogram-tandem mass spectrometry Hydroxyeicosanoic acid HEA HETE(s) Hydroxyeicosatetraenoic acid(s) IAC Immunoaffinity chromatography LC-MS Liquid chromatogram-mass spectrometry LC-MS/MS Liquid chromatogram-tandem mass spectrometry LLOD Lower limit of detection LLOQ Lower limit of quantitation Lipoxygenase LO LT Leukotriene MD Microdialysate MOX Methoxyamine Nordihydroguaiaretic acid NDGA PAR Peak area ratio Pentafluorobenzyl PFB PFB-Br Pentafluorobenzyl bromide Prostaglandin PG PGE-MUM Prostaglandin E major urinary metabolite Prostaglandin H synthase PGH SRM Selected-reaction monitoring TMS Trimethylsilyl TMSOH Trimethylsilanol Thromboxane Tx
Scheme 1. A part of the arachidonic acid cascade illustrating four major pathways. (1) The cyclooxygenase (COX) or prostaglandin H synthase (PGHS) pathway includes thromboxane (TxA2 ) and prostacyclin (PGI2 ) which are spontaneously converted to their stable analogs TxB2 and 6-keto-PGF1␣ , respectively. (2) The lipoxygenase (LO) pathway includes several hydroxyeicosatetraenoic acids (HETEs), leukotriene B4 (LTB4 ) and the cysteinyl leukotrienes C4 (LTC4 ), LTD4 and LTE4 . For simplicity, the 20-HETE is not shown. (3) The cytochrome P450 (CYP) family generates various epoxyeicosatrienoic acids (EETs). (4) Many eicosanoids including F2 -isoprostanes are generated both from free arachidonic acid and from arachidonic acid esterified to lipids. For the sake of simplicity, the pathways are shown for the free form of arachidonic acid.
blood sampling. This is important because artefactual formation may be abundant and exceed several-fold the concentrations that prevail in vivo (see below). Therefore, primary circulating eicosanoids such as leukotriene B4 (LTB4 ), thromboxane B2 (TxB2 ), i.e., the stable hydrolysis product of TxA2 , and 6-keto-prostaglandin F1␣ (6kPGF1␣ ), i.e., the stable hydrolysis product of prostacyclin (PGI2 ), are not useful biomarkers of eicosanoid synthesis in health and disease or in pharmacotherapy. The problem of ex vivo artefactual formation of eicosanoids has been successfully overcome by measuring in the circulation certain metabolites, commonly their dehydro- and dinor-metabolites such as 11-dehydro-TxB2 , (11dh-TxB2 ), 2,3-dinor-TxB2 and 2,3-dinor-6kPGF1␣ (Scheme 2). Such eicosanoid metabolites cannot be formed ex vivo, at least not to a degree that could compromise analysis. Nevertheless, eicosanoid synthesis is preferentially assessed by measuring eicosanoids in urine rather than in blood (i.e., plasma or serum) [1]. 2. Analysis of eicosanoids by mass spectrometry
1. Introduction Arachidonic acid (5Z,8Z,11Z,14Z-eicosatetraenoic acid) is the common precursor of a variety of biologically highly active lipid mediators. Enzymatically formed intermediate products, such as leukotriene A4 (LTA4 ) and prostaglandin H2 (PGH2 ), are further converted by the action of different enzymes in a cascadelike manner to generate numerous, in part structurally closely related carboxyl-groups containing substances, collectively named eicosanoids (Scheme 1). Arachidonic acid, free and esterified to lipids, is also oxidized by non-enzymatic reactions to form numerous isomeric compounds, the iso-eicosanoids, such as isoleukotrienes, F2 -isoprostanes, and iso-thromboxane [1,2]. Primary eicosanoids and iso-eicosanoids circulate in blood and are extensively metabolized by - and -oxidation. Eicosanoids, i.e., both primary eicosanoids and their metabolites, are present in blood and urine at concentrations in the pM to nM range. Wide structural variety and extremely low quantity render unequivocal identification and reliable quantification of these substances in biological samples incomparably challenging. In addition, both eicosanoids and iso-eicosanoids can be formed artefactually ex vivo, notably during
Historically, GC-MS and GC-MS/MS have played a central role both in the identification and in the quantification of eicosanoids and iso-eicosanoids in biological samples [1]. Our current knowledge of the biology, pharmacology and analytical chemistry of eicosanoids is primarily based on the application of the GC-MS and GC-MS/MS technologies over the past three to four decades. GC-MS and GC-MS/MS approaches have also been proved useful to validate immunology-based assays for eicosanoids [1]. However, RIA and ELISA methods for eicosanoids are generally artefact-prone, lack specificity and may produce hardly interpretable results. Nevertheless, they have been and are still frequently used in experimental and clinical research. Due to great advances in the LC-MS/MS technology, especially over the past two decades, this relatively new approach is increasingly used both in biomedical and life sciences [1–4], and in clinical chemistry [5]. The main reason for the invasion is its putative simplicity, basically due to the omission of multiple derivatization steps. Derivatization in the GC-MS and GC-MS/MS analysis of eicosanoids is absolutely required because eicosanoids have two or three thermally labile chemical functionalities, i.e.,
Please cite this article in press as: D. Tsikas, A.A. Zoerner, J. Chromatogr. B (2014), http://dx.doi.org/10.1016/j.jchromb.2014.03.017
G Model CHROMB-18846; No. of Pages 10
ARTICLE IN PRESS D. Tsikas, A.A. Zoerner / J. Chromatogr. B xxx (2014) xxx–xxx
3
3. Recall past knowledge The chemical structures of selected primary eicosanoids and their index metabolites discussed below are shown in Scheme 2. 3.1. A brief historical retrospection
Scheme 2. Selected stable primary eicosanoids (left panel) and their corresponding major metabolites resulting from - and/or -oxidation from their precursors (right panel). PGE-MUM, major urinary metabolite of E prostaglandins. See also Scheme 1.
carboxylic, hydroxylic and keto groups (Scheme 2). This lends the LC-MS/MS approach the potential for online and high-throughput analysis. These particular attributes attract the interest of analysts who increasingly apply LC-MS/MS both in experimental and in clinical research. With regard to ionization and more so to collisioninduced dissociation (CID), LC-MS/MS and GC-MS/MS seem to produce structurally closely related ions (Scheme 3). At present, LC-MS/MS seems to rapidly and massively displace GC-MS and GC-MS/MS in various analytical disciplines especially including eicosanoids analysis in biological samples. Yet, “blind” trust in the attractive LC-MS/MS technology and non-consideration of several decades back-dating acquirement of basic analytical knowledge also hold the potential for fallacious analytical certainty. Neglecting previous knowledge, including recognition and avoidance of pre-analytical pitfalls and generation of references values and intervals for eicosanoids established by GC-MS and GC-MS/MS pioneers in this field, is a risky progress in the LC-MS/MS-based analysis of circulating eicosanoids. Development of LC-MS/MS-based methods for the quantitative analysis of eicosanoids and other biomarkers in biological samples and validation of these methods by established GC-MS- and GC-MS/MS methods is highly advisable, if not even an obligation. But, consideration of findings revealed by GC-MS and GC-MS/MS over the past decades seems to be neglected by some LC-MS/MS users in the area of eicosanoids analysis. We wish to discuss this issue in more detail below.
Prostacyclin (prostaglandin I2 , PGI2 ) was discovered in the vascular endothelium in 1976 [6]. At physiological pH, prostacyclin is highly unstable and hydrolyses spontaneously to the stable 6-keto-PGF1␣ [7]. The major urinary metabolites of prostacyclin and 6-keto-PGF1␣ infused into healthy volunteers have been identified as 2,3-dinor-6-keto-PGF1␣ and 13,14-dihydro-2,3-dinor6-keto-PGF1␣ [8]. From the two dinor metabolites of prostacyclin, 2,3-dinor-6-keto-PGF1␣ was found to be its major metabolite in the urine of healthy humans [9] and became generally accepted as the index parameter of prostacyclin [1] (Scheme 2). With regard to circulating prostanoids, the sampling technique has been found to influence the basal level of 6-keto-PGF1␣ in plasma due to artefactual formation. Instead of 6-keto-PGF1␣ , measurement of circulating 6,15-diketo-13,14-dihydro-PGF1␣ , a major circulating metabolite of prostacyclin in man [10,11], is preferred because it cannot be formed in vitro. Thromboxane A2 was discovered in 1975 in platelets [12]. At physiological pH, TxA2 is extremely labile and is non-enzymatically converted to TxB2 . 2,3-dinor-TxB2 has been identified as the major urinary metabolite of systemically administered TxB2 in man [13]. In total, twenty urinary metabolites of TxB2 have been identified, among them 2,3-dinor-TxB2 and 11-dehydro-TxB2 were the most abundant [14]. Quantification of urinary 2,3-dinor-TxB2 and 11dehydro-TxB2 is a useful measure of in vivo production of TxB2 in man [1] (Scheme 2). 11-dehydro-TxB2 has been shown to be an appropriate quantitative index metabolite of TxA2 both in plasma [15,16] and in urine [17]. Leukotriene B4 was discovered in 1979 as a dihydroxy metabolite (5,12-dihydroxy-eicosatetraenoic acid) of arachidonic acid by rabbit peritoneal polymorphonuclear leucocytes [18–20]. A biologically active B4 -isoleukotriene has been found to be formed by free radical-catalyzed peroxidation of glycerophospholipids independent of 5-LO [21]. Cysteinyl leukotrienes are known since the 1930s as ‘slow reacting substances of anaphylaxis’ (SRS-A). Structural elucidation of the first member of cysteinyl leukotrienes, i.e., LTC4 , was reported in 1979 [18,22,23]. Catalytical saturation of the double bonds and desulphurisation of the cysteinyl moiety of cysteinyl leukotrienes by noble catalysts and hydrogen gas yields quantitatively 5-hydroxy-eicosanoic acid (5-HEA), which can be utilized to measure cysteinyl leukotrienes by GC-MS [24–26]. LTE4 has been shown to be the major metabolite of LTC4 in urine in humans and in the monkey [27–30] (Scheme 2). Urinary LTE4 has been turned out to be a useful index metabolite for whole body production of cysteinyl leukotrienes in humans [1]. 3.2. Artefactual ex vivo formation Electron-capture negative-ion chemical ionization (ECNICI) permits the most sensitive GC-MS and GC-MS/MS analysis of eicosanoids after proper derivatization with fluorine-rich reagents such as pentafluorobenzyl (PFB) bromide (Scheme 3). Hughes et al. quantitated LTB4 in serum of healthy humans by GC-MS as a PFB ester trimethylsilyl ether derivative [31]. When blood was collected into the 5-LO inhibitor nordihydroguaiaretic acid (NDGA), the LTB4 concentration was determined to be 10 ± 4 pg/mL [31]. By contrast, in blood that was not collected into NDGA, the LTB4 concentration in serum was ten-fold higher, indicating artefactual ex vivo LTB4 formation [31]. This finding was also observed by Blair et al. [32]. In
Please cite this article in press as: D. Tsikas, A.A. Zoerner, J. Chromatogr. B (2014), http://dx.doi.org/10.1016/j.jchromb.2014.03.017
G Model CHROMB-18846; No. of Pages 10 4
ARTICLE IN PRESS D. Tsikas, A.A. Zoerner / J. Chromatogr. B xxx (2014) xxx–xxx
Scheme 3. Analysis of prostaglandin E2 (PGE2 ) by LC-MS/MS (left panel) and GC-MS/MS (right panel). For sensitive GC-MS/MS analysis of PGE2 in the negative-ion chemical ionization mode (NICI), PGE2 must be derivatized in a three-step process. By means of pentafluorobenzyl bromide (PFB-Br) the pentafluorobenzyl (PFB) ester of PGE2 is prepared. By means of methoxyamine (MOX) hydrochloride the keto group of PGE2 is methoximated and yields the syn- and anti-isomers (not shown). Finally, the remaining two hydroxyl groups are etherified by means of N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA). Electrospray ionization in the negative mode (ESI-) of native PGE2 free acid yields the carboxylate of PGE2 , i.e., to generate the anion with the mass-to-charge (m/z) ratio of 351. Collision-induced dissociation (CID) of this precursor ion yields several product ions. The anion m/z 271 [M − PFB−2 × H2 O − CO2 ]− is frequently used for the quantitative analysis of PGE2 in the selected-reaction monitoring (SRM) mode. In GC-MS/MS, NICI of the fully derivatized PGE2 generates the most intense precursor anion at m/z 351 [M − PFB]− . CID of this precursor ion yields several product ions. The anion m/z 268 [M − PFB − 2 × TMSOH − CH3 OH–CO2 ]− is frequently used for the quantitative GC-MS/MS analysis of PGE2 in the SRM mode. TMSOH, trimethylsilanol.
human plasma, LTB4 concentration was reported to be about half the serum concentration in the same volunteers [31], suggesting LTB4 formation during clotting. By GC-MS/MS in the ECNICI mode, we found that LTB4 is present in plasma of healthy humans from blood that was not collected into NDGA at concentrations lower than 37 pg/mL [33], confirming the observations by Hughes et al. [31] and Blair et al. [32]. Thus, ex vivo artefactual formation of LTB4 is a major obstacle in the quantitative determination of circulating LTB4 , if not useful precautions are taken. Abundant ex vivo formation is not specific to LTB4 but also occurs to other eicosanoids in blood, notably TxA2 [34]. Upon activation, for instance during blood sampling and processing, large amounts of TxA2 may be produced artefactually by platelets. This particular circumstance can be utilized to measure TxA2 synthesis in platelets and the effects of drugs such as acetylsalicylic acid [35] (see also next section). By means of a sophisticated twodimensional LC-MS/MS and by collecting blood into indomethacin, a cyclooxygenase (COX) inhibitor, prostaglandin E2 (PGE2 ) and prostaglandin F2␣ (PGF2␣ ) were measured in human plasma at basal concentrations of 0.4–1.6 pg/mL and 2.9–11.4 pg/mL, respectively [36] (Table 1). These concentrations are very close to those measured by GC-MS/MS [34,40]. In contrast, Ferreiro-Vera et al. measured by LC-MS/MS PGF2␣ concentrations in human serum being almost 10,000 times higher, suggesting massive artefactual ex vivo formation due to the absence of any added COX inhibitor during blood sampling [41] (see also Ref. [17]). In conclusion, measurement of primary eicosanoids in the circulation should be generally regarded as an unreliable method to assess their synthesis in vivo due to potential and abundant ex vivo formation [34]. Another potential problem associated with eicosanoids analysis may be formation of numerous isomers which must be separated chromatographically. LTB4 and its isomers can easily be separated
each from other by capillary GC and normal phase LC [31], and this must be ensured in LC coupled to MS too [53]. 3.3. Ex vivo stimulation of eicosanoids synthesis in blood Stimulation of LTB4 synthesis in human blood, for instance by using calcium ionophore A23187, has been frequently investigated by GC-MS (see for instance Refs. [31,32]). Although there is considerable linear dependency of LTB4 synthesis in human whole blood on calcium ionophore A23187 concentration, remarkable stimulation of LTB4 synthesis can be achieved by only 1 M calcium ionophore A23187 [32], and this potent stimulation is unlikely to be achieved by disease or drugs. Therefore, utilization of extraordinary high calcium ionophore A23187 concentrations such as 14 M [53] is not very meaningful. Finally, because not all blood cells produce eicosanoids, their synthesis is best assessed by investigating blood cells that are primarily responsible for particular eicosanoids, e.g., platelets for TxA2 , and monocytes for LTB4 and cysteinyl leukotrienes [54]. 3.4. Sensitivity—basal concentrations and limits of detection and quantitation Table 1 summarizes reported concentrations of selected (mainly primary) eicosanoids in plasma, serum, and other biological fluids of healthy humans as measured by reported validated GC-MS, GC-MS/MS and LC-MS/MS methods. As can be clearly seen from Table 1, the concentration of eicosanoids in the circulation and other biological fluids of healthy humans lies in the lowest pg/mLrange. Most of the reported LC-MS/MS methods have several-fold higher lower limits of quantitation (LLOQ) values than reported GCMS/MS methods. For example, the LLOQ value of a recently reported
Please cite this article in press as: D. Tsikas, A.A. Zoerner, J. Chromatogr. B (2014), http://dx.doi.org/10.1016/j.jchromb.2014.03.017
G Model CHROMB-18846; No. of Pages 10
ARTICLE IN PRESS D. Tsikas, A.A. Zoerner / J. Chromatogr. B xxx (2014) xxx–xxx
5
Table 1 Reported concentrations for selected leukotrienes, prostaglandins and thromboxane in biological fluids of healthy humans as measured by approaches based on GC-MS, GC-MS/MS and LC-MS/MS. Eicosanoid
Basal concentration
Reference
Approach
Leukotriene B4 LTB4 LTB4 LTB4 LTB4 LTB4
