Contents lists available at SciVerse ScienceDirect
Journal of Chromatography B
j ourna l h om epage: www.elsev ier .com/ locate /chromb
ast liquid chromatography–quadrupole linear ion trap-masspectrometry analysis of polyunsaturated fatty acids andicosanoids in human plasma�
inda Kortza,b,∗, Juliane Dorowa,b, Susen Beckera,b, Joachim Thierya,b, Uta Ceglareka,b
Institute of Laboratory Medicine, Clinical Chemistry and Molecular Diagnostics, University Hospital Leipzig, Liebigstr. 27, 04103 Leipzig, GermanyLIFE – Leipzig Research Center for Civilization Diseases, Universität Leipzig, Germany
a r t i c l e i n f o
rticle history:eceived 8 October 2012ccepted 12 March 2013vailable online xxx
a b s t r a c t
Profiling of polyunsaturated fatty acids (PUFAs) and their oxidized metabolites, mainly eicosanoids,in human plasma by fast liquid chromatography–mass spectrometry is described. Sample preparationinvolved protein precipitation of 200 �L plasma followed by on-line solid-phase extraction. 7 PUFAs and94 oxidized metabolites were separated utilizing a C-18 column packed with 2.6 �m core–shell particlesin 7 min. The analytes and deuterium-labeled standards were detected via scheduled multiple reaction
eywords:ass spectrometry
ast liquid chromatographyinear ion trapn-line solid-phase extractionicosanoidsolyunsaturated fatty acid metabolism
monitoring transitions (123 sMRM). Simultaneously, linear ion trap fragment spectra were acquired forconfirmation, if necessary. The lower limit of quantitation ranged between 200 and 1000 ng/mL for thePUFAs and 10–1000 pg/mL for the metabolites. The method was applied to a study on plasma samplesfrom 50 healthy subjects.
Eicosanoids are a class of over 100 lipid mediators derived fromhe twenty-carbon polyunsaturated fatty acids (PUFAs) arachidonic
Please cite this article in press as: L. Kortz, et al., J. Chromatogr. B (2013), h
cid (AA), eicosapentaenoic acid (EPA) and dihomo-�-linoleniccid (DHGLA) [1,2]. They are generated via the cyclooxygen-se (COX), lipoxygenase (LOX) and cytochrome P450 (CYP450)
Abbreviations: AA, arachidonic acid; CE, collision energy; COX, cyclooxygenase;ytochrome P450, CYP450; DHET, dihydroxyeicosatrienoic acid; DHA, docosa-exaenoic acid; DHGLA, dihomo-�-linolenic acid; EET, epoxyeicosatrienoic acid;PA, eicosapentaenoic acid; EPI, enhanced product ion; ESI, electrospray ioniza-ion; HETE, hydroxyeicosatetraenoic acid; HODE, hydroxyoctadecadienoic acid; LA,inoleic acid; LC–MS, liquid chromatography–mass spectrometry; LLOQ, lower limitf quantitation; LOD, limit of detection; LOX, lipoxygenase; LC-QqLIT-MS, liquidhromatography–quadrupole linear ion trap-mass spectrometry; MRM, multipleeaction monitoring; PUFA, polyunsaturated fatty acid; sMRM, scheduled multipleeaction monitoring; S/N, signal-to-noise ratio; SPE, solid-phase extraction; UHPLC,ltra high performance liquid chromatography; ULOQ, upper limit of quantitation.� This paper belongs to the Special Issue, Fast Liquid Chromatography, edited byaraskevas D. Tzanavaras and Constantinos Zacharis (Guest Editors).∗ Corresponding author at: Institute of Laboratory Medicine, Clinical Chemistry
enzymes. Prostaglandins and thromboxanes are COX products,hydroxyeicosatetraenoic acids (HETEs) and leukotrienes derivemainly from LOX, whereas epoxyeicosatrienoic acids (EETs) andseveral HETEs are CYP450 products [1]. Isoprostanes are formednon-enzymatically from AA [3]. Beside eicosanoids there existLOX-derived metabolites from docosahexaenoic acid (DHA) calledprotectins and resolvins [4] and LOX-derived metabolites fromlinoleic acid (LA) called hydroxyoctadecadienoic acids (HODEs) [5].
The analysis of these bioactive metabolites in cells, tissuesand body fluids is of growing interest but challenging due tolow endogenous concentrations (ng/L-range), isomeric and isobaricstructures, and in vitro generation. In the past, eicosanoids weremainly analyzed by gas chromatography–mass spectrometry [6–8]or by immunoassays [3]. Recently, liquid chromatography–massspectrometry (LC–MS) became more important for eicosanoidanalysis due to its less laborious sample pretreatment (no derivati-zation) and the potential for multi-analyte testing. LC–MS methodsare either focused on selected eicosanoids [9,10] or on profiling upto 50 and more metabolites with typical analysis times of 20 min[11–15].
ttp://dx.doi.org/10.1016/j.jchromb.2013.03.012
For chromatographic separation of eicosanoid isobars and regio-isomers high chromatographic resolution power is required, whichusually results in long analysis times. For example, 8-iso-PGF2�
and isomers were separated in 12 min [16], whereas PGE2 isomers
Please cite this article in press as: L. Kortz, et al., J. Chromatogr. B (2013), h
ARTICLE ING Model
CHROMB-18332; No. of Pages 5
2 L. Kortz et al. / J. ChromatogrTa
ble
1Li
mit
s
of
det
ecti
on
(LO
Ds)
and
lin
ear
ran
ges
of
the
PUFA
/PU
FA-m
etab
olit
es
sele
cted
for
vali
dat
ion
as
wel
l as
intr
a-as
say
(n
=
10)
and
inte
r-as
say
vari
abil
ity
(n
=
15
over
5
day
s).
Com
pou
nd
LOD
Lin
ear
ran
geIn
tra-
assa
y
pre
cisi
onIn
ter-
assa
y
pre
cisi
on
pg/
mL
(ng/
mL)
ap
g/m
L
(ng/
mL)
aM
ean
, pg/
mL
(ng/
mL)
aC
V, %
Mea
n, p
g/m
L
(ng/
mL)
aC
V, %
Tetr
anor
-PG
EM/T
etra
nor
-PG
DM
10
25
–
10,0
00
542
7.0
574
10.0
iPF 2
�-I
V50
100
–
10,0
00
267
14.1
259
30.7
TxB
250
100
–
10,0
00
527
7.2
565
9.0
PGF 2
�50
100
–
10,0
00
395
11.8
361
24.6
15-k
eto-
PGE 2
/8-i
so-1
5-ke
to-P
GE 2
50
100
–
10,0
00
176.
8
12.1
168
17.8
5(S)
,6(R
)-LX
A4
50
100
–
10,0
00
207.
79.
719
3
18.2
LTE 4
100
250
–
10,0
00
1394
12.4
1439
13.9
PGB
250
100
–
10,0
00
440
12.7
467
33.4
7(S)
-Mar
esin
150
100
–
10,0
00
1157
10.7
1105
12.3
10(S
),17
(S)-
DiH
DoH
e
50
100
–
10,0
00
1483
9.3
1448
14.7
LTB
450
250
–
5000
370
4.4
361
6.3
5,6-
DH
ET
50
250
–
10,0
00
555
3.3
508
10.3
5-H
EPE
50
250
–
10,0
00
816
7.7
701
15.4
13-H
OD
E
50
100
–
10,0
00
11,0
11
6.5
12,3
15
9.7
15-o
xo-E
TE
100
500
–
10,0
00
343
6.9
347
12.7
11-H
ETE
50
100
–
10,0
00
1915
5.6
1921
12.4
12-H
ETE
100
500
–
10,0
00
3249
11.1
3631
20.1
5-H
ETE
100
250
–
5000
4777
10.8
5022
15.7
11,1
2-EE
T
100
500
–
10,0
00
383
5.5
361
12.7
AA
a10
0
500
–
10,0
00
1006
3.5
922
9.6
PRESS. B xxx (2013) xxx– xxx
needed 65 min [17] using 4 and 3 �m particles, respectively.Recent developments in chromatographic supports and instru-mentation allow rapid and efficient separations based on columnsfilled with sub-2 �m particles (ultra high performance liquidchromatography, UHPLC) or core–shell material [18]. The result-ing peak widths <10 s are challenging for MS-detection if multiplereaction monitoring (MRM) experiments with 5–20 ms/MRM areapplied. The simultaneous generation of a characteristic fragmentspectrum for analyte confirmation using a hybrid quadrupole linearion trap instrument (QqLIT) [19] is also demanding in combinationwith fast chromatography due to prolonged scan times.
The aim of the current study was the development of a rapidprofiling method combining on-line solid-phase extraction forsample-clean up and preconcentration with fast LC and QqLIT-MSfor PUFAs and their oxidized metabolites.
2. Experimental
2.1. Standards and solutions
Unlabeled and deuterium-labeled standards were obtainedfrom Cayman Chemical (Ann Arbor, MI, USA) (see SupplementTables 1 and 2). Standards were diluted in methanol and storedat −80 ◦C. Acetonitrile, 2-propanol, methanol and formic acid,all ULC–MS grade, were purchased from Biosolve (Valkenswaard,Netherlands). Water was obtained in-house from a nanopure waterpurification system (Thermo Scientific, Germany). Precipitationsolution consisting of methanol–zinc sulfate 4:1 v/v was pre-pared with a solution of 89 g/L zinc sulfate heptahydrate (p.A.Merck Darmstadt, Germany) in water. Calibrators were preparedby diluting the stock solutions with isotonic saline solution to finalconcentrations of 5, 10, 25, 50, 100, 250, 500, 1000, 2500, 5000,10,000 ng/mL for PUFAs and 5, 10, 25, 50, 100, 250, 500, 1000,2500, 5000, 10,000 pg/mL for the metabolites. Calibration curveswere performed with 1/x weighted linear regression to account forthe smaller concentration values. Quality controls were preparedusing pooled plasma spiked with appropriate volumes of PUFA andmetabolite stock solutions. Labeled standard solutions were dilutedfrom stock to a final concentration of 5 ng/mL in methanol–water,50:50 (v/v) (50 ng/mL for AA-d8).
2.2. Collection of samples
EDTA-K3-blood samples were collected from 50 healthysubjects (ethical approval 082-10-190-42010) and stored approx-imately 0.5 h at room temperature after blood taking. Aftercentrifugation at 3220 × g for 10 min the plasma was stored at−80 ◦C in safe-lock tubes (Eppendorf, Hamburg, Germany) untilanalysis.
2.3. Sample preparation
200 �L EDTA-plasma were transferred in 1.5 mL polypropylenetubes and mixed with 50 �L of labeled standard solution and 400 �Lprecipitation solution for 2 min. After 5 min of centrifugation at10,000 × g, 250 �L of clear supernatant were transferred into anautosampler vial and stored at 10 ◦C in a temperature controlledautosampler until injection of 200 �L to the online-SPE-LC–MS/MSsystem.
2.4. On-line SPE-LC–MS/MS analysis
ttp://dx.doi.org/10.1016/j.jchromb.2013.03.012
For MS/MS analysis a 5500 QTrap mass spectrometer (AB Sciex,Darmstadt) with electrospray ionization (ESI) in negative ion modewas applied (for mass spectrometric parameters see SupplementTable 1). Quantitative analysis was performed using sMRM scans
cquiring MRM data for each analyte in a time window of thexpected retention time ±30 s. Ion source temperature was set to00 ◦C, nebulizer gas to 55 psi, auxilliary gas to 70 psi and ionizationoltage to −4500 V.
LIT-fragment spectra were acquired from injections of sin-le eicosanoid standards using enhanced product ion (EPI) scans,riggered by the sMRM scans. LIT fill time was set to ‘dynamic’ andPI data was acquired at a scan speed of 10,000 Da/s. For all ana-ytes a library of EPI spectra was generated with a collision energyCE) of 30 ± 15 eV.
A Shimadzu UFLC LC-20A Prominence liquid chromatographyystem (Shimadzu Deutschland GmbH, Duisburg, Germany) con-isting of a high pressure gradient system (2 LC-AD pumps), ansocratic pump (LC-20 ADSP) for on-line SPE and a CTO-20AColumn oven with a sixport switching valve was used. Sample injec-ion was carried out with a CTC-PAL autosampler (Axel-Semrau,prockhövel, Germany). For on-line SPE a Strata-X extractionolumn (20 × 2 mm i.d., 25 �m particle size, Phenomenex, Aschaf-enburg, Germany) was applied. LC separation was performed on
core–shell LC column (Kinetex C18, 100 × 2.1 mm i.d., 2.6 �marticle size, Phenomenex, Aschaffenburg, Germany) with ancetonitrile–water gradient described in detail in Supplementable 3.
.5. Method validation
Limit of detection (LOD) was determined from standard solu-ions using a signal-to-noise (S/N) ratio of 3. Lower limit ofuantitation (LLOQ, S/N ratio ≥ 10) and the upper limit of quan-itation (ULOQ) were defined by the linear calibration curve.
Precision and recovery were approved for 20 selected PUFAsnd metabolites (see Supplement Table 1). Intra-assay and inter-
Please cite this article in press as: L. Kortz, et al., J. Chromatogr. B (2013), h
ssay precision were determined by analyzing 10 replicates of theuality controls on 1 day and 5 consecutive days (3 replicates).ecovery was determined by standard addition to pooled plasmac = 300 ng/mL for PUFAs, 300 pg/mL for metabolites).
3. Results and discussion
3.1. Mass spectrometry
Electrospray ionization in negative mode was chosen as thePUFAs and their metabolites build [M−H]− quasimolecular ionsdue to their carboxylic acid moiety. MRMs were selected based onintensity and specificity of fragments. Several eicosanoid isomerslike the AA derived group of diol regio-isomers (DHET) showedcompound specific fragments and could therefore be separated bya characteristic MRM. In contrast, some prostanoids as the regio-isomers PGE2 and PGD2 or the stereo-isomers 8-iso-PGF2� andPGF2� share the same prominent fragment ions and therefore needto be chromatographically resolved. The 123 mass transitions (101analytes, 22 labeled standards) were acquired in scheduled MRMmode with a fixed target scan time of 0.25 s and a detection win-dow of ±30 s of the expected retention times. A maximum of 42MRMs overlap, resulting in scan times of ≥6 ms for each MRM (seeSupplement Figure 1 for time distribution). MRM and EPI scans forfragment spectra generation resulted in a total cycle time of 0.56 s.This is a precondition to assure approx. 10 data points/peak whencombined with the core–shell column for fast chromatography.
3.2. Fast chromatography
For chromatographic separation of the 101 PUFAs and oxidizedmetabolites a C-18 column with 2.6 �m core–shell particles waschosen. A main advantage of core–shell compared to UHPLCcolumns with particle sizes <2 �m is the lower back pressurewhich averaged 350 bar. Peak widths of 4–6 s were found incontrast to 12–18 s using a 4 �m particle column [18]. In Fig. 1a LC–MS/MS chromatogram of a standard mixture containing
ttp://dx.doi.org/10.1016/j.jchromb.2013.03.012
the 101 analytes is shown. Chromatographic separation couldbe achieved in 7 min after 1 min on-line sample clean-up andpreconcentration. Total analysis time including on-line SPE andre-equilibration of the columns was 13 min. Prostaglandins and
4 L. Kortz et al. / J. Chromatogr. B xxx (2013) xxx– xxx
Fm
tb5hbcostT1�
3
TepfiPaP2
dpPCmiScao2(p
3
oaIo
Fig. 2. Chromatograms of plasma samples including chromatograms of assignedlabeled standards: (A) TxB2, (1, 2 peaks of the TxB2 anomers, see [24]), c = 396 pg/mL,TxB2-d4; (B) 14,15-DHET c = 526 pg/mL, 8,9-DHET-d11; (C) 12-HHT c = 918 pg/mL,15-HETE-d8; (D) 5-HEPE c = 633 pg/mL, 15-HETE-d8; (E) 9-HODE: 1, EZ-diasteromer;2, presumably EE-diastereomer (see LIT investigation) c = 6060 pg/mL, 13-HODE-d4;
ig. 1. LC-MS/MS chromatogram of a standard mixture of PUFAs and oxidizedetaboites (c = 1000 ng/mL for PUFAs and 1000 pg/mL for metabolites).
hromboxanes elute in the range of 1.7–3.7 min, leukotrienesetween 2.6 and 4.2 min, diols (DiHETEs, DHETs) between 4.0 and.4 min, monohydroxyl compounds (HETEs, HEPEs, HODEs) andydroperoxides (HpETEs) between 5.1 and 6.2 min, epoxides (EETs)etween 6.5 and 6.6 min and PUFAs between 7.2 and 8.0 min. Forhromatograms of each extracted MRM see Supplement Fig. 2. Theptimization of the chromatographic separation was focused on ahort analysis time. As consequence, it was not possible to separatehe following isomers which are quantified as sum parameters:etranor-PGEM/Tetranor-PGDM, 15-keto-PGF2�/8-iso-PGE2/PGE2,5-keto-PGE2/8-iso-15-keto-PGE2, 8-iso-PGA2/PGA2/PGJ2, and-linolenic acid/�-linolenic acid.
.3. Method validation
LODs and linear calibration ranges (Table 1, Supplementable 4) were found for PUFAs and most metabolites in thexpected plasma concentration range, with the exception ofrostaglandins and leukotrienes which is a limitation of the pro-ling method. LLOQ ranged between 200–1000 ng/mL for theUFAs and 10–1000 pg/mL for the metabolites. Chromatogramst LLOQ level are provided in Supplement Fig. 3. ULOQ forUFAs and metabolites ranged between 5000–10,000 ng/mL and500–10,000 pg/mL, respectively.
20 PUFAs/PUFA-metabolites were selected for method vali-ation, representing the different structures and enzymaticathways. Intra-assay coefficients of variation (CV) of these 20UFA/PUFA-metabolites ranged from 3.3 to 14.1% and inter-assayV from 6.3 to 33.4% (see Table 1). The high CVs of some metabolitesight be explained by the fact that the assigned labeled standard
s not best suited. Additionally, poor stability may be a reason.tandard addition experiments for determination of recovery werehallenging due to poor stability of PUFAs/PUFA-metabolites andnalyte losses resulting from adsorption to vials which was previ-usly reported [15]. Reasonable recovery among the investigated0 PUFAs and metabolites was found for LTB4 (79%), 5,6-DHET79%), 13-HODE (95%) and PGB2 (119%). For the other selected com-ounds most recovery rates were below 35%.
.4. Eicosanoid profiling in human plasma samples
Plasma concentrations of the 6 PUFAs, 14 eicosanoids and 3
Please cite this article in press as: L. Kortz, et al., J. Chromatogr. B (2013), h
ther oxidized PUFA metabolites found in our 50 study subjectsre provided as medians in Table 2 (for chromatograms see Fig. 2).n all analyzed plasma samples a second peak at 5.85 min was foundn the sMRM of 9-HODE in contrast to standard injections showing
(F) 5-HETE c = 775 pg/mL, 5-HETE-d8; (G) 14,15-EET c = 236 pg/mL, 8,9-DHET-d11;(H) AA c = 1380 ng/mL: 1, AA-d8; 2, unknown peak.
only one peak at 5.70 min (Fig. 3A and B). To investigate this sig-nal the LIT spectra were compared to our fragment spectra library.The EPI spectra of 9-HODE (Z,E-diastereomer in the standard) andthe unknown peak share the same prominent fragment of m/z 171(Fig. 3C and D respectively). As HODE make characteristic frag-ments depending on the position of the hydroxy-group, it is mostlikely the E,E-diastereomer of 9-HODE described by Yoshida et al.[20].
TxB2 was detected only in parts of the plasma sam-ples, which is explained by the fact that TxB2 and otherprostaglandins/leukotrienes have endogenous concentrations aslow as 1 pg/mL [21] and are thus below the LOD. Generally, con-centrations measured with our method were in a comparable rangeas the concentration values determined by the profiling methodsof Quehenberger et al. [22] and Psychogios et al. [23]. Differences
ttp://dx.doi.org/10.1016/j.jchromb.2013.03.012
may be explained by the use of different standards and calibratorsets. However, the influence of inter-individual variation and pre-analytical standardization is known to be high for these analytes.
L. Kortz et al. / J. Chromatogr. B xxx (2013) xxx– xxx 5
(B)
5.5 5.7 5.9 6.1Time, min
0.0
1.4
2.8
4.0
Inte
nsity
×10
4 , cp
s 21(A)
5.5 5.7 5.9 6.10.0
2.5
4.5
6.3
1
Inte
nsity
×10
5 , cp
s
Time, min
(C)
60 100 140 180 220 260 300m/z, Da
0.0
1.4
2.8
4.0 295.3
171.2 277.2
Inte
nsity
×10
6 , cp
s (D)
60 100 140 180 220 260 3000.0
3.0
6.0
8.8 295.2
277.4171.1
m/z, Da
Inte
nsity
×10
5 , cp
s
sMRM295 / 171
sMRM295 / 171
Product ion s ofm/z 295 at 5.70 min
Product ion s ofm/z 295 at 5.85 min
F 00 pgs
Tt
4
ao
A
tfS
A
i2
R
[
[
[[[
[
[
[[[
[
[
[
ig. 3. Application of the LIT function: chromatogram of 9-HODE standard, c = 10,0pectrum of 9-HODE standard (C), and unknown peak (D).
herefore, in a next step preanalytical protocols from blood takingill sample storage have to be validated.
. Conclusion
A fast LC–MS/MS method was developed including on-line SPEnd LIT fragment confirmation for the profiling of 7 PUFAs and 94xidized metabolites within a total analysis time of 13 min.
cknowledgements
This publication is supported by LIFE – Leipzig Research Cen-er for Civilization Diseases, Universität Leipzig. This project wasunded by means of the European Social Fund and the Free State ofaxony.
ppendix A. Supplementary data
Supplementary data associated with this article can be found,n the online version, at http://dx.doi.org/10.1016/j.jchromb.013.03.012.
Please cite this article in press as: L. Kortz, et al., J. Chromatogr. B (2013), h
[2] M. Bruegel, U. Ceglarek, J. Thiery, LaboratoriumsMedizin 33 (2009) 333.[3] S. Basu, Antioxid. Redox Signal. 10 (2008) 1405.[4] C.N. Serhan, N.A. Petasis, Chem. Rev. 111 (2011) 5922.[5] N. Youhnovski, D. Schulz, C. Schwarz, G. Spiteller, K. Schubert, Z. Naturforsch C
58 (2003) 268.
[
[
/mL, (A) and plasma sample c = 5921 pg/mL, (B, 1: 9-HODE; 2: unknown peak); EPI
[6] D. Tsikas, J. Chromatogr. B: Biomed. Sci. Appl. 717 (1998) 201.[7] H. Tsukamoto, T. Hishinuma, T. Mikkaichi, H. Nakamura, T. Yamazaki, Y.
Tomioka, M. Mizugaki, J. Chromatogr. B: Analyt. Technol. Biomed. Life Sci. 774(2002) 205.
[8] A. Ferretti, V.P. Flanagan, J. Chromatogr. Biomed. Appl. 622 (1993) 109.[9] P. Montuschi, J. Chromatogr. B: Analyt. Technol. Biomed. Life Sci. 877 (2009)
1272.10] M. Enzler, S. Schipp, L.B. Nicolas, J. Dingemanse, C. Siethoff, J. Chromatogr. B:
Analyt. Technol. Biomed. Life Sci. 901 (2012) 67.11] L. Kiss, Y. Roder, J. Bier, N. Weissmann, W. Seeger, F. Grimminger, Anal. Bioanal.
Chem. 390 (2008) 697.12] M. Masoodi, A. Nicolaou, Rapid Commun. Mass Spectrom. 20 (2006) 3023.13] J. Yang, K. Schmelzer, K. Georgi, B.D. Hammock, Anal. Chem. 81 (2009) 8085.14] D.S. Dumlao, M.W. Buczynski, P.C. Norris, R. Harkewicz, E.A. Dennis, Biochim.
Biophys. Acta 1811 (2011) 724.15] K. Strassburg, A.M. Huijbrechts, K.A. Kortekaas, J.H. Lindeman, T.L. Pedersen, A.
Dane, R. Berger, A. Brenkman, T. Hankemeier, J. van Duynhoven, E. Kalkhoven,J.W. Newman, R.J. Vreeken, Anal. Bioanal. Chem. 404 (2012) 1413.
16] M. Haschke, Y.L. Zhang, C. Kahle, J. Klawitter, M. Korecka, L.M. Shaw, U. Chris-tians, Clin. Chem. 53 (2007) 489.
17] S.A. Brose, B.T. Thuen, M.Y. Golovko, J. Lipid Res. 52 (2011) 850.18] L. Kortz, C. Helmschrodt, U. Ceglarek, Anal. Bioanal. Chem. 399 (2011) 2635.19] L. Kortz, R. Geyer, U. Ludwig, M. Planert, M. Bruegel, A. Leichtle, G.M. Fiedler, J.
Thiery, U. Ceglarek, LaboratoriumsMedizin 33 (2009) 341.20] Y. Yoshida, S. Kodai, S. Takemura, Y. Minamiyama, E. Niki, Anal. Biochem. 379
(2008) 105.21] H. Schweer, C.O. Meese, B. Watzer, H.W. Seyberth, Biol. Mass Spectrom. 23
rill, S. Bandyopadhyay, K.N. Jones, S. Kelly, R.L. Shaner, C.M. Sullards, E. Wang,R.C. Murphy, R.M. Barkley, T.J. Leiker, C.R.H. Raetz, Z. Guan, G.M. Laird, D.A. Six,D.W. Russell, J.G. McDonald, S. Subramaniam, E. Fahy, E.A. Dennis, J. Lipid Res.51 (2010) 3299.
ttp://dx.doi.org/10.1016/j.jchromb.2013.03.012
23] N. Psychogios, D.D. Hau, J. Peng, A.C. Guo, R. Mandal, S. Bouatra, I. Sinelnikov, R.Krishnamurthy, R. Eisner, B. Gautam, N. Young, J. Xia, C. Knox, E. Dong, P. Huang,Z. Hollander, T.L. Pedersen, S.R. Smith, F. Bamforth, R. Greiner, B. McManus, J.W.Newman, T. Goodfriend, D.S. Wishart, PLoS ONE 6 (2011) e16957.
24] H. John, W. Schlegel, J. Chromatogr. B: Biomed. Sci. Appl. 698 (1997) 9.
