molecular species of the alcohol biomarker ... · molecular species of the alcohol biomarker...

Molecular Species of the Alcohol BiomarkerPhosphatidylethanol in Human Blood

Measured by LC-MSAnders Helander1* and Yufang Zheng1

BACKGROUND: The alcohol biomarker phosphati-dylethanol (PEth) comprises a group of ethanol-derived phospholipids formed from phosphatidylcho-line by phospholipase D. The PEth molecular specieshave a common phosphoethanol head group ontowhich 2 fatty acid moieties are attached. We developedan electrospray ionization (ESI) LC-MS method forqualitative and quantitative measurement of differentPEth species in human blood.

METHODS: We subjected a total lipid extract of wholeblood to HPLC gradient separation on a C4 columnand performed LC-ESI-MS analysis using selected ionmonitoring of deprotonated molecules for the PEthspecies and phosphatidylpropanol (internal standard).Identification of individual PEth species was based onESI–tandem mass spectrometry (MS/MS) analysis ofproduct ions.

RESULTS: The fatty acid moieties were the major prod-uct ions of PEth, based on comparison with PEth-16:0/16:0, 18:1/18:1, and 16:0/18:1 reference material. ForLC-MS analysis of different PEth species in blood, weused a calibration curve covering 0.2–7.0 �mol/LPEth-16:0/18:1. The lower limit of quantitation of themethod was �0.1 �mol/L, and intra- and interassayCVs were �9% and �11%. In blood samples collectedfrom 38 alcohol patients, the total PEth concentrationranged between 0.1 and 21.7 �mol/L (mean 8.9). PEth-16:0/18:1 and 16:0/18:2 were the predominant molec-ular species, accounting for approximately 37% and25%, respectively, of total PEth. PEth-16:0/20:4 andmixtures of 18:1/18:1 plus 18:0/18:2 (not separated us-ing selected ion monitoring because of identical molec-ular masses) and 16:0/20:3 plus 18:1/18.2 made up ap-proximately 13%, 12%, and 8%.

CONCLUSIONS: This LC-MS method allows simulta-neous qualitative and quantitative measurement of

several PEth molecular species in whole bloodsamples.© 2009 American Association for Clinical Chemistry

To aid in the early identification of people engaged inharmful alcohol consumption, and for monitoringalcohol-dependent patients during treatment, researchefforts have focused on developing laboratory tests foralcohol biomarkers with a longer detection windowthan that offered by ethanol testing (1 ). The conju-gated minor ethanol metabolites ethyl glucuronide (2 )and ethyl sulfate (3 ) and fatty acid ethyl esters (4 ) areexamples of such tests. Ethyl glucuronide and ethyl sul-fate are excreted in urine for a much longer time thanethanol and hence have gained popularity as diagnos-tically sensitive and specific tests to spot recent alcoholconsumption (5–7 ). For detection and follow-up oflong-term risky or heavy drinking, measurement of thealcohol-related change in the serum transferrin glyco-form profile known as carbohydrate-deficient trans-ferrin (CDT)2 (8, 9 ) has become a standard method,often used in combination with liver function tests(e.g., �-glutamyltransferase) (1 ).

Phosphatidylethanol (PEth) is another indicatorof high alcohol consumption identified some time ago(10, 11 ) that so far has received limited clinical interest.PEth is an ethanol-derived phospholipid formed fromphosphatidylcholine (PC) in cell membranes by atransphosphatidylation reaction catalyzed by phos-pholipase D (12 ). Phospholipase D normally hydro-lyzes PC into phosphatidic acid and choline, but be-cause the affinity for ethanol is �1000-fold higher thanfor water, PEth is formed at the expense of phospha-tidic acid when ethanol is present (13, 14 ). In clinicalstudies, PEth was not detected after a single high alco-hol intake but after sustained drinking of more than

1 Department of Clinical Neuroscience, Karolinska Institute, Stockholm, Sweden.* Address correspondence to this author at: Alcohol Laboratory, L7:03, Karolinska

University Hospital Solna, SE-171 76 Stockholm, Sweden. Fax �46-8-51771532; e-mail [email protected].

Received November 19, 2008; accepted April 23, 2009.Previously published online at DOI: 10.1373/clinchem.2008.120923

2 Nonstandard abbreviations: CDT, carbohydrate-deficient transferrin; PEth, phos-phatidylethanol; PC, phosphatidylcholine; CE, capillary electrophoresis; PProp,phosphatidylpropanol; ESI, electrospray ionization; SIM, selected ion monitor-ing; MS/MS, tandem mass spectrometry; SRM, selected reaction monitoring;LOD, limit of detection; LOQ, lower limit of quantitation.

Clinical Chemistry 55:71395–1405 (2009)

Lipids, Lipoproteins, and Cardiovascular Risk Factors

1395

approximately 50 g/day for 3 weeks (15 ). The PEthconcentration in blood was reported to correlate withthe amount of alcohol consumed during the previous 2weeks (16 ), and PEth may be detected for up to 2– 4weeks after cessation of heavy drinking (17–19 ).

One reason for the limited clinical use of PEth asalcohol biomarker is probably that the analyticalmethod employed, lipid extraction of whole blood fol-lowed by HPLC analysis using an evaporative light-scattering detector (18, 20 ), is not well suited to rou-tine use in clinical laboratories. An alternative capillaryelectrophoresis (CE) method with UV detection wasrecently introduced (21 ), and development of an im-munoassay based on a monoclonal PEth antibody hasbeen initiated (22 ).

PEth is not a single molecular species but a groupof phospholipids with a common nonpolar phospho-ethanol head group onto which 2 fatty acid moieties,typically with a chain length of 16, 18, or 20 carbons(18, 23 ), are attached at positions sn-1 and sn-2. Giventhat PEth originates from PC, the fatty acid composi-tion of PEth is likely to mirror that of PC. The numer-ous combinations of chain lengths and double bondsenable formation of a large number of PC species (24 ),with 16:0/18:1 (nomenclature for fatty acids � [num-ber of carbons]:[number of double bonds]) and 16:0/18:2 being the major fatty acid combinations in PCextracted from human erythrocyte membranes (25–27).

Current analytical methods for PEth based onHPLC (19, 20 ) and CE (21 ), and any immunoassaytargeting the head group (22 ), measure the sum of allPEth species. Measurement of individual species is fea-sible by LC-MS (18, 28 ), depending on the variablelength and number of unsaturated bonds of the fattyacid chains. We aimed to develop an LC-MS methodfor quantitative and qualitative measurement of differ-ent PEth species in human blood and determine theprofile of molecular species in samples from heavydrinkers.

Materials and Methods

CHEMICALS

We obtained PEth reference materials containing2 palmitic acids (1,2-dipalmitoyl-sn-glycero-3-phosphoethanol; PEth-16:0/16:0) and 2 oleic acids(1,2-dioleoyl-sn-glycero-3-phosphoethanol; PEth-18:1/18:1) and phosphatidylpropanol (PProp-18:1/18:1;internal standard) from Avanti Polar Lipids; anotherPEth reference material containing 1 palmitic and1 oleic acid (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanol; PEth-16:0/18:1) from Biomol Re-search Laboratories; and acetonitrile (HPLC grade),ammonium acetate, and isopropanol (analyticalgrade) from Merck. All other chemicals were of an-

alytical or HPLC grade, and the water was of HPLCgrade.

We prepared stock solutions of PEth referencematerials (14.0 �mol/L) and PProp (1.35 �mol/L) in a1:5 solution of 2 mmol/L ammonium acetate and ace-tonitrile and stored them at �20 °C until use. The stocksolutions were stable for at least 3 months.

BLOOD SAMPLES

The blood specimens used for method developmentwere surplus volumes from the clinical samples poolsent to the Alcohol Laboratory (Karolinska UniversityHospital, Stockholm, Sweden) for testing of alcoholuse and abuse. We obtained additional samples fromhealthy volunteers reporting consumption of �100 gethanol/week. For comparison with the HPLC methodfor total PEth (20 ), we obtained blood samples fromLund University Hospital (Sweden). Blood was col-lected into EDTA tubes and stored at 4 °C for �3 daysbefore analysis. PEth is stable for at least 3 weeks inrefrigerated blood samples (20 ). The ethics committeeat the Karolinska University Hospital approved thestudy.

SAMPLE PREPARATION

We prepared total lipid extracts of whole blood (mainlyerythrocyte membranes (19 )) by stepwise addition of100 �L blood to 600 �L isopropanol and 50 �L 1.35�mol/L PProp internal standard under constantvortex-mixing (20, 29 ). Thereafter, samples weregently mixed for 10 min. We then added hexane (2�450 �L) with mixing after each addition, and mixedagain for another 10 min. The samples were finally cen-trifuged for 10 min at 2000g at 4 °C. The clear superna-tants were transferred to new glass tubes and evapo-rated to dryness under a stream of nitrogen gas at 30 °Cusing a metal block. The final dried extract was dis-solved in 50 �L hexane, followed by 50 �L acetonitrileand 75 �L isopropanol, and transferred to 0.3-mL glassautosampler vials.

LC-ESI-MS QUANTIFICATION OF PEth SPECIES

We used an Agilent 1100 series liquid chromatographicsystem connected to an LC/MSD SL mass spectromet-ric detector with the electrospray ionization (ESI) in-terface operated in negative ion mode together withChemStation software. The conditions used were dry-ing gas flow 10.0 �, nebulizer gas 20 �, drying gastemperature 350 °C, and capillary voltage 3000 V.

Chromatographic separation of the lipid extractswas achieved on a 50 by 3 mm, 5-�m HyPurity C4column (Thermo Scientific) maintained at 25 °C. TheLC system was operated in gradient mode with solventA being 20% 2 mmol/L ammonium acetate and 80%

1396 Clinical Chemistry 55:7 (2009)

acetonitrile and solvent B 100% isopropanol. Fromsample injection until 2.0 min, isocratic elution with90% A and 10% B was used; from 2.0 –3.0 min, a lineargradient to 50% B; from 3.0 – 6.0 min, a linear gradientto 100% B; from 6.0 –7.0 min, isocratic elution with100% B; and from 7.0 – 8.0 min, a linear gradient backto 90% A and 10% B. The flow rate was 200 �L/min,and the sample injection volume, 10 �L. Recondition-ing the column with 90% A and 10% B for 10 min aftereach injection improved the reproducibility of the re-tention times.

We performed LC-ESI-MS analysis using selectedion monitoring (SIM) of the deprotonated moleculesfor the different PEth species and PProp (Table 1). Wedetermined the PEth concentration in unknown wholeblood samples from the peak area ratio between analyteand internal standard by reference to a calibrationcurve. The calibration curve was produced by spikingPEth-negative blood with 0.2–7.0 �mol/L PEth-16:0/18:1 or 18:1/18:1 prepared from the standard stock so-lutions by dilution with solvent A.

LC-ESI–TANDEM MASS SPECTROMETRY IDENTIFICATION OF

PEth SPECIES

The LC-ESI–tandem mass spectrometry (MS/MS) sys-tem was a Perkin-Elmer series 200 LC system con-nected to Sciex API 2000MS, with the ESI interface op-erated in negative ion mode, and Analyst 1.1 software(Applied Biosystems). The LC conditions were identi-cal to those employed for single MS analysis.

We performed LC-ESI-MS/MS analysis using se-lected reaction monitoring (SRM) to detect the majordeprotonated fragments from each PEth species (Table1). The conditions used were curtain gas 10 �, collisiongas 3 �, ion spray voltage �3500 V, temperature450 °C, nebulizer gas 30 �, auxiliary gas 45 �, declus-tering potential �61 V, focusing potential �350 V, en-trance potential �10 V, and collision energy �40 V.

VALIDATION OF THE LC-MS METHOD

The analytical imprecision (CV%) of the LC-ESI-MSmethod was determined for 4 blood samples contain-ing 0.3–7.0 �mol/L PEth, using triplicate measure-ments over 5 days (Clinical and Laboratory StandardsInstitute EP 15-A2 guideline). We calculated linearityequations using 6 PEth-negative blood samples spikedwith 0.2–7.0 �mol/L PEth-16:0/18:1 or 18:1/18:1.

To study the possible impact by matrix effects(30 ), PEth-16:0/16:0, 16:0/18:1, and 18:1/18:1 stan-dards at 1.0 �mol/L were infused postcolumn at aconstant rate of 10 �L/min, whereas 6 extractedPEth-negative blood samples were injected via theautosampler. Pre- and postextraction addition experi-ments (n � 6) compared the detector responses forstandard and internal standard in blanks and bloodsamples.

Results

MONITORING OF PEth SPECIES BY LC-MS AND LC-MS/MS

Chemical structures for the PEth species included inthis study and for PProp used as the internal standardare shown in Fig. 1. In phospholipids isolated fromhuman erythrocytes, saturated fatty acids (e.g., 16:0,palmitic acid) are mainly located in the sn-1 positionand unsaturated fatty acids (e.g., 18:2, linoleic acid)mainly in the sn-2 position (26, 31 ). The correspond-ing deprotonated molecules used in SIM of differentPEth species and PProp are listed in Table 1. Referencematerial was available for PEth-16:0/16:0, 16:0/18:1,and 18:1/18:1, whereas target masses for the other mo-lecular species were calculated.

The major product ions of the different PEth spe-cies (Table 1), as identified by ESI-MS/MS, corre-sponded to the fatty acid chains (Fig. 2A). The identityof the products was confirmed by analysis of PEth ref-erence material, where PEth-16:0/16:0 and 18:1/18:1

Table 1. Mass spectrometric parameters for theLC-ESI-MS and LC-ESI-MS/MS methods for

quantitative and qualitative measurement ofdifferent PEth molecular species in total lipid

extracts of human whole blood.

Molecular species

LC-MS LC-MS/MS

SIM, m/zaMajor product ions,

m/za (fatty acid)

PEth-16:0/16:0 675.7 255.5 (16:0)

PEth-16:0/18:2 699.7 255.5 (16:0)

279.5 (18:2)

PEth-16:0/18:1 701.7 255.5 (16:0)

281.5 (18:1)

PEth-16:0/20:4 723.7 255.5 (16:0)

303.5 (20:4)

PEth-16:0/20.3,PEth-18:1/18:2

725.7 255.5 (16:0)279.5 (18:2)

281.5 (18:1)

305.5 (20:3)

PEth-18:1/18:1,PEth-18:0/18:2

727.7 283.5 (18:0)281.5 (18:1)

279.5 (18:2)

PEth-18:0/18:1 729.7 283.5 (18:0)

281.5 (18:1)

PProp-18:1/18:1 741.7 281.5 (18:1)

a Analysis was performed using ESI in negative ion mode ([M � H]�).

LC-MS Measurement of PEth in Human Blood

Clinical Chemistry 55:7 (2009) 1397

produced 1 major product ion each (m/z 255.5 for 16:0fatty acid and m/z 281.5 for 18:1 fatty acid) and PEth-16:0/18:1 both of them. Besides the major fatty acidproduct ions, 2 low-intensity ions of PEth-16:0/18:1were detected. These corresponded to the phospho-ethanol head group (m/z 125) (23 ) and to a loss ofthe 18:1 fatty acid from position sn-2 (m/z 437) (28 )(Fig. 2A).

The identification of PEth species in blood sam-ples was based on SRM analysis of the unique massratios for the fatty acid chains to the parent compound,producing similar product ion ratios as with the refer-ence material (the m/z 281.5/255.5 ratios for PEth-16:0/18:1 were 3.0 and 3.1, respectively) (Fig. 2B). Some ofthe examined molecular species have an identical mo-lecular mass (e.g., m/z 727.7 for PEth-18:1/18:1 and18:0/18:2) and also very similar retention times, andthus were not separated in the LC-MS chromatograms(Fig. 3). However, for mixtures of 2 or more PEth spe-cies with identical mass, SRM analysis enabled individ-ual identification owing to the unique combinations offatty acid chains.

QUANTIFICATION OF PEth SPECIES BY LC-MS

The PEth species and PProp internal standard elutedover a narrow time range with retention times ofaround 3 min (Fig. 3). In routine use (typically �40samples/run), the intraassay imprecision for absoluteand relative (PEth/PProp) retention times were typi-cally �4% and �1%, respectively. No interfering peakswere observed on routine analysis of �100 PEth-negative clinical blood samples (data not shown).

The limit of detection (LOD) (signal-to-noise ra-tio �3) of the LC-MS method was �0.02 �mol/L forall PEth species examined, and the correspondinglower limit of quantitation (LOQ) (signal-to-noise ra-tio �10) was �0.1 �mol/L. The pre- and postextrac-tion addition experiments showed an absolute extrac-tion efficiency of approximately 80% for both PEth andPProp. The postcolumn infusion experiments showedno ion suppression or enhancement at the retentiontimes of PEth (tested with 16:0/16:0, 16:0/18:1, and 18:1/18:1 reference materials) and the internal standard(data not shown). Still, in the postextraction additionexperiments carried out with blood, the peak intensi-

Fig. 1. Structures of different PEth species and phosphatidylpropanol.

Tentative structures of the molecular species of PEth and PProp [internal standard (IS)] measured by the LC-MS method.Nomenclature for fatty acids � (number of carbons):(number of double bonds). *Commercially available material.


Fig. 2. LC-MS/MS analysis of PEth-16:0/18:1.

(A), LC-ESI-MS/MS product ion scan of PEth-16:0/18:1 reference material producing 2 major fatty acid chain fragments (m/z255.5 for 16:0 and m/z 281.5 for 18:1) and 2 minor fragments (m/z 125 for the phosphoethanol head group [Holbrook et al.(23 )] and m/z 437 for a loss of the 18:1 fatty acid chain from position sn-2 [Tolonen et al. (28 )]). (B), LC-ESI-MS/MSchromatograms for the deprotonated molecules of PEth-16:0/18:1 (m/z 701.7/255.5 for 16:0; m/z 701.7/281.5 for 18:1) for aPEth-negative blood sample spiked with PEth-16:0/18:1 (top), the blood sample without addition of PEth (middle), and a poolof whole blood samples collected from heavy drinkers (total PEth concentration 7.5 �mol/L).



ties for both PEth and PProp were approximately 90%compared with the blank, demonstrating approxi-mately 10% absolute matrix effect. However, as allchanges for PEth and PProp were of the same magni-tude, the internal standard compensated for losses dur-ing sample extraction and analysis. Accordingly, when5 blood samples containing 0 –2.6 �mol/L total PEthwere spiked with 1.0 �mol/L and 10.0 �mol/L PEth-16:0/18:1 before the extraction procedure, mean ana-lytical recoveries were 101% and 103%, respectively.

The LC-MS method for quantification of PEth spe-cies in whole blood samples produced linear results forPEth-16:0/18:1 in the concentration range 0.2–20.0�mol/L (see Supplemental Fig. 1, inset, in the Data Sup-plement that accompanies the online version of this arti-

cle at www.clinchem.org/content/vol55/issue7). Calibra-tion curves prepared by spiking PEth-negative bloodsamples with PEth-16:0/18:1 or 18:1/18:1 reference mate-rial showed nearly identical responses (online Supple-mental Fig. 1), indicating that the MS response for differ-ent molecular species was essentially the same. Hence, forroutine quantification of PEth species in whole blood,PEth-16:0/18:1, being the major species in clinical sam-ples (see data below), was chosen for preparation of thecalibration curve and 0.2–7.0 �mol/L was employed asthe measuring range for each molecular species. It shouldbe noted that this corresponded to a routine measuringrange for total PEth from 0.2 �mol/L up to approximately20 �mol/L, given that PEth-16:0/18:1 accounted for ap-proximately 35% of total PEth in human blood.

Fig. 3. LC-MS measurement of different PEth molecular species in blood.

Example of LC-ESI-MS chromatograms for the deprotonated molecules of PEth molecular species and PProp (internal standard)in a whole blood sample obtained from a heavy drinker (total PEth concentration 7.5 �mol/L). IS, internal standard.


The intraassay CV% for total PEth concentrationin the measuring range was 4.5%– 8.6%, and the valuesfor interassay CV% were �11% (online SupplementalTable 1).

PEth STABILITY AND FORMATION FROM ETHANOL ON STORAGE

In agreement with previous observations (20 ), all PEthspecies were demonstrated to be stable for at least 5days in whole blood samples stored at 4 °C (data notshown). Furthermore, when 38 patient blood samplescontaining 0 –20 �mol/L PEth were reanalyzed afterstorage for 14 months at �80 °C, there was no indica-tion of declining total values (Passing–Bablok regres-sion equation: y2007 � 1.0308x2008 � 0.3731) orchanges in the species profile.

To study further the risk for postsampling synthe-sis of PEth on storage of blood samples containing eth-anol (32, 33 ), PEth-negative samples from 4 controlsubjects collected in EDTA, heparin, and citrate tubeswere spiked with ethanol to a final concentration of 2g/L. Aliquots were taken at the start and after differenttimes of storage at 20 °C, 4 °C, �20 °C, and �80 °C foranalysis of PEth species by the LC-MS method. No for-mation of PEth was observed after storage for 4 h and24 h, whereas concentrations ranging up to 0.75�mol/L were found after 72 h storage at 20 °C or�20 °C (online Supplemental Fig. 2). In the bloodfrom 1 donor, total PEth concentrations at or above 0.2�mol/L (i.e., within the measuring range) were gener-ated in all samples, whereas concentrations �0.1�mol/L were found in 1 blood sample and undetect-able amounts in the remaining 2.

DISTRIBUTION OF PEth SPECIES IN BLOOD FROM HEAVY

DRINKERS

The LC-MS method was applied for determination ofthe PEth species profile in blood samples collectedfrom 38 heavy drinkers recently admitted for alcoholdetoxification. When the patients tested negative forethanol (breath test), they were transferred to an inpa-tient treatment ward where blood sampling took placewithin 2 days after admission. PEth was detected in allsamples, the total concentration range being 0.1–21.7�mol/L, with a mean of 8.9 �mol/L (median 9.6).The results demonstrated that PEth-16:0/18:1 and16:0/18:2 were the predominant species (Fig. 4A), ac-counting for on average 37% and 26%, respectively, oftotal PEth by this method. Owing to interindividualvariations, PEth-16:0/18:2 was sometimes the majorform. PEth-16:0/20:4 (m/z 723.7) and mixtures of 18:1/18:1 plus 18:0/18:2 (m/z 727.7; not separated in theSIM method because of identical mass) and 16:0/20:3plus 18:1/18.2 (m/z 725.7) made up approximately13%, 12%, and 8%, respectively (Fig. 4A). When 1 out-patient was followed with serial testing over an 11-week

period, including a relapse into heavy drinking, thetime course for individual PEth species were similarbut not identical (Fig. 4B). In this case, PEth-16:0/18:2became the quantitatively most important species dur-ing the relapse.

COMPARISON OF PEth RESULTS BY LC-MS AND HPLC

In 21 blood samples where the total PEth concentra-tion had already been determined by an HPLC method(20 ), LC-ESI-MS analysis was carried out for a com-parison. The 2 methods produced similar total PEthvalues in the low concentration range (�3 �mol/L),whereas the LC-MS method generally gave higher val-ues above this threshold (Fig. 5).

Discussion

The 9 PEth species covered in our LC-MS method witheither SIM or SRM detection were selected from thecomposition of PC molecular species in human eryth-rocytes (27 ) and the results of previous MS studiesidentifying 16:0 and 18:1 as the major fatty acid chainsin PEth (18, 23 ). In agreement with the species distri-bution for PC (26 ), the present study confirmed 16:0/18:1 and 16:0/18:2 as the predominant fatty acid com-binations in PEth extracted from human whole bloodsamples (i.e., mainly erythrocyte membranes (19 )), to-gether making up approximately 60% of total PEthspecies measured by this method. However, individualvariations in the molecular composition of PEth spe-cies were demonstrated both between individuals andwithin the same patient during a relapse. It should bepointed out that the fatty acid profile is, to some extent,also influenced by dietary factors (34 ).

The identity of the PEth species was partly basedon comparison with reference material that was avail-able for 3 molecular species, whereas theoretical targetmasses for the other species were calculated. The chro-matographic peaks were finally assigned to distinctPEth species based on MS/MS monitoring of the cor-responding fatty acid fragment ions, albeit not identi-fying the exact positioning of double bonds. In contrastto the stable polar head group of PC that is usually usedas the target product in SRM analysis, producing acharacteristic fragment of the phosphocholine moietyat m/z 184 in positive ion mode, the nonpolar phos-phoethanol head group of PEth (m/z 125 in negativeion mode) (23 ) is unstable. Accordingly, the main frag-ment ions of PEth were the fatty acid chains that werereadily detected in negative ion mode. For chromato-graphic peaks that could represent mixtures of 2 ormore PEth species with identical mass, SRM analysisenabled individual quantification due to the uniquecombinations of fatty acid chains. For routine mea-surement of the 9 PEth species covered in this study,



Fig. 4. Profile of PEth species in blood samples from alcohol patients.

(A), Distribution of molecular species of PEth in whole blood samples collected from 38 patients undergoing treatment foralcohol-related problems. Some species were not distinguishable in the ESI-MS method because of identical masses. Data for individualsubjects are connected with lines. Inset: Mean (SD) values for relative amounts of each PEth species (% of total PEth). (B), Time coursefor individual PEth species in blood samples collected from 1 outpatient over 11 weeks, including a relapse into heavy drinking.


LC-MS analysis in SIM mode was used. It should benoted that monitoring at least 4 PEth species in SIMmode, or 1 precursor ion and 2 product ions in SRMmode, could both meet proposed requirements forconfirmatory LC-MS analysis (i.e., a minimum of 4identification points) (35 ) and thereby produce legallydefendable results.

The LC-MS method allowed simultaneous mea-surement of the major and some minor PEth species inhuman whole blood samples, as well as determinationof the total concentration (i.e., sum of all species cov-ered) in a clinically relevant concentration range. Thisrepresents a methodological development comparedwith HPLC and CE methods that can only determine atotal amount (20, 21 ). To reduce the analysis time ofthe LC-MS method in routine use, a 2-column switch-ing setup can be applied. For clinical application ofPEth as alcohol biomarker, previous studies have usedcutoffs for total PEth in the range of approximately0.2– 0.7 �mol/L, depending on the LOQ of the HPLCmethod at the time. In Sweden, 0.7 �mol/L is currentlyused as the routine clinical threshold. Based on previ-ous studies, the lower limit of the measuring range forthe LC-MS method was set to 0.2 �mol/L for each PEthspecies. However, the method can detect at least 20

times lower concentrations, thereby possibly also al-lowing for detection of lower drinking levels. This hy-pothesis is supported by observations that the apparentalcohol consumption cutoff (based on self-report) de-tectable by PEth was approximately 50 g/day at an LOQof 0.7 �mol/L total PEth by HPLC (15 ), whereasamounts below 40 g/day were detectable when theLOQ was reduced to approximately 0.2 �mol/L (16 ).

For use as an alcohol biomarker, an advantage ofPEth over some conventional analytes (e.g., liver func-tion tests) is the theoretical high specificity for alcohol,being a direct ethanol metabolite. Still, observations ofindividual PEth formation rates (32 ) indicated that itmight not be possible to link the PEth concentration inblood to a precise drinking level. A main drawback withPEth is the risk for postsampling production on storageof ethanol-containing samples, also when frozen at�20 °C as demonstrated in this and previous publica-tions (32, 33 ), which could generate false-positive re-sults. This risk is especially serious in postmortem ex-aminations, because even if the deceased subject hadnot ingested alcohol before death, artifactual ethanolformation between time of death and autopsy due tomicrobial action is a common problem (36, 37 ). Forthat reason, special precaution related to handling and

Fig. 5. Comparison of PEth concentrations in patient blood by LC-MS and HPLC.

Comparison of total PEth concentrations for 21 blood samples determined by the LC-ESI-MS method and an HPLC method (20 )that is routinely employed for PEth testing. Inset: The same data set presented as a Bland–Altman plot.



storage of samples before analysis is required if PEthresults are to be used for medicolegal matters.

In conclusion, the LC-MS method allowed for si-multaneous qualitative and quantitative analysis of sev-eral PEth species in human whole blood. The fatty acidcomposition observed for the different molecular speciesagreed with that reported for PC, indicating that PEthformation from PC by action of phospholipase D is a gen-eral process and not limited to certain molecular species.To simplify future analysis of PEth as alcohol biomarkerand allow for standardization of measurement, it shouldbe advantageous to focus on distinct molecular species,which is possible by LC-MS, instead of the total concen-tration. Based on the patient data, the 2 predominantPEth species to monitor clinically would be 16:0/18:1 and16:0/18:2, as these together accounted for approximately60% of the total amount in blood from heavy drinkers.

Author Contributions: All authors confirmed they have contributed tothe intellectual content of this paper and have met the following 3 re-

quirements: (a) significant contributions to the conception and design,acquisition of data, or analysis and interpretation of data; (b) draftingor revising the article for intellectual content; and (c) final approval ofthe published article.

Authors’ Disclosures of Potential Conflicts of Interest: Uponmanuscript submission, all authors completed the Disclosures of Poten-tial Conflict of Interest form. Potential conflicts of interest:

Employment or Leadership: None declared.Consultant or Advisory Role: None declared.Stock Ownership: None declared.Honoraria: None declared.Research Funding: A. Helander, financial support through the re-gional agreement on medical training and clinical research (ALF)between Stockholm County Council and the Karolinska Institute.Expert Testimony: None declared.

Role of Sponsor: The funding organizations played no role in thedesign of study, choice of enrolled patients, review and interpretationof data, or preparation or approval of manuscript.

Acknowledgments: The authors thank Dr Therese Hansson for pro-viding blood samples used for comparison with the routine HPLCmethod, and Professor Olof Beck for valuable comments on themanuscript.

References

1. Helander A. Biological markers in alcoholism.J Neural Transm Suppl 2003;(66):15–32.

2. Schmitt G, Aderjan R, Keller T, Wu M. Ethylglucuronide: an unusual ethanol metabolite inhumans. Synthesis, analytical data, and determi-nation in serum and urine. J Anal Toxicol 1995;19:91–4.

3. Helander A, Beck O. Mass spectrometric identifi-cation of ethyl sulfate as an ethanol metabolite inhumans. Clin Chem 2004;50:936–7.

4. Best CA, Laposata M. Fatty acid ethyl esters: toxicnon-oxidative metabolites of ethanol and markersof ethanol intake. Front Biosci 2003;8:e202–17.

5. Erim Y, Bottcher M, Dahmen U, Beck O, BroelschCE, Helander A. Urinary ethyl glucuronide testingdetects alcohol consumption in alcoholic liverdisease patients awaiting liver transplantation.Liver Transpl 2007;13:757–61.

6. Politi L, Leone F, Morini L, Polettini A. Bioanalyti-cal procedures for determination of conjugates orfatty acid esters of ethanol as markers of ethanolconsumption: a review. Anal Biochem 2007;368:1–16.

7. Helander A, Bottcher M, Fehr C, Dahmen N, BeckO. Detection times for urinary ethyl glucuronideand ethyl sulfate in heavy drinkers during alcoholdetoxification. Alcohol Alcohol 2009;44:55–61.

8. Jeppsson JO, Arndt T, Schellenberg F, Wielders JP,Anton RF, Whitfield JB, Helander A. Toward stan-dardization of carbohydrate-deficient transferrin(CDT) measurements: I. Analyte definition andproposal of a candidate reference method. ClinChem Lab Med 2007;45:558–62.

9. Bergstrom JP, Helander A. Influence of alcoholuse, ethnicity, age, gender, BMI and smoking onthe serum transferrin glycoform pattern: implica-tions for use of carbohydrate-deficient transferrin(CDT) as alcohol biomarker. Clin Chim Acta 2008;388:59–67.

10. Alling C, Gustavsson L, Anggård E. An abnormal

phospholipid in rat organs after ethanol treat-ment. FEBS Lett 1983;152:24–8.

11. Yang SF, Freer S, Benson AA. Transphosphatidy-lation by phospholipase D. J Biol Chem 1967;242:477–84.

12. Gustavsson L. Phosphatidylethanol formation:specific effects of ethanol mediated via phospho-lipase D. Alcohol Alcohol 1995;30:391–406.

13. Kobayashi M, Kanfer JN. Phosphatidylethanol for-mation via transphosphatidylation by rat brainsynaptosomal phospholipase D. J Neurochem1987;48:1597–603.

14. Chalifa-Caspi V, Eli Y, Liscovitch M. Kinetic anal-ysis in mixed micelles of partially purified ratbrain phospholipase D activity and its activationby phosphatidylinositol 4,5-bisphosphate. Neuro-chem Res 1998;23:589–99.

15. Varga A, Hansson P, Lundqvist C, Alling C. Phos-phatidylethanol in blood as a marker of ethanolconsumption in healthy volunteers: comparisonwith other markers. Alcohol Clin Exp Res 1998;22:1832–7.

16. Aradottir S, Asanovska G, Gjerss S, Hansson P,Alling C. Phosphatidylethanol (PEth) concentra-tions in blood are correlated to reported alcoholintake in alcohol-dependent patients. Alcohol Al-cohol 2006;41:431–7.

17. Hansson P, Caron M, Johnson G, Gustavsson L,Alling C. Blood phosphatidylethanol as a marker ofalcohol abuse: levels in alcoholic males during with-drawal. Alcohol Clin Exp Res 1997;21:108–10.

18. Gunnarsson T, Karlsson A, Hansson P, Johnson G,Alling C, Odham G. Determination of phosphati-dylethanol in blood from alcoholic males usinghigh-performance liquid chromatography andevaporative light scattering or electrospray massspectrometric detection. J Chromatogr B BiomedSci Appl 1998;705:243–9.

19. Varga A, Hansson P, Johnson G, Alling C. Nor-malization rate and cellular localization of phos-

phatidylethanol in whole blood from chronic al-coholics. Clin Chim Acta 2000;299:141–50.

20. Aradottir S, Olsson BL. Methodological modifica-tions on quantification of phosphatidylethanol inblood from humans abusing alcohol, using high-performance liquid chromatography and evapo-rative light scattering detection. BMC Biochem2005;6:18.

21. Varga A, Nilsson S. Nonaqueous capillary elec-trophoresis for analysis of the ethanol consump-tion biomarker phosphatidylethanol. Electro-phoresis 2008;29:1667–71.

22. Nissinen AE, Makela SM, Vuoristo JT, LiisananttiMK, Hannuksela ML, Horkko S, SavolainenMJ. Immunological detection of in vitro formedphosphatidylethanol—an alcohol biomarker—with monoclonal antibodies. Alcohol Clin Exp Res2008;32:921–8.

23. Holbrook PG, Pannell LK, Murata Y, Daly JW.Molecular species analysis of a product of phos-pholipase D activation: phosphatidylethanol isformed from phosphatidylcholine in phorbolester- and bradykinin-stimulated PC12 cells.J Biol Chem 1992;267:16834–40.

24. Houjou T, Yamatani K, Imagawa M, Shimizu T,Taguchi R. A shotgun tandem mass spectrometricanalysis of phospholipids with normal-phaseand/or reverse-phase liquid chromatography/elec-trospray ionization mass spectrometry. RapidCommun Mass Spectrom 2005;19:654–66.

25. Connor WE, Lin DS, Thomas G, Ey F, DeLougheryT, Zhu N. Abnormal phospholipid molecular spe-cies of erythrocytes in sickle cell anemia. J LipidRes 1997;38:2516–28.

26. Ekroos K, Ejsing CS, Bahr U, Karas M, Simons K,Shevchenko A. Charting molecular compositionof phosphatidylcholines by fatty acid scanningand ion trap MS3 fragmentation. J Lipid Res2003;44:2181–92.

27. Leidl K, Liebisch G, Richter D, Schmitz G. Mass


spectrometric analysis of lipid species of humancirculating blood cells. Biochim Biophys Acta2008;1781:655–64.

28. Tolonen A, Lehto TM, Hannuksela ML, SavolainenMJ. A method for determination of phosphati-dylethanol from high density lipoproteins byreversed-phase HPLC with TOF-MS detection.Anal Biochem 2005;341:83–8.

29. Radin NS. Extraction of tissue lipids with a solventof low toxicity. Methods Enzymol 1981;72:5–7.

30. Annesley TM. Ion suppression in mass spectrom-etry. Clin Chem 2003;49:1041–4.

31. Beermann C, Mobius M, Winterling N, SchmittJJ, Boehm G. sn-Position determination of

phospholipid-linked fatty acids derived fromerythrocytes by liquid chromatography electros-pray ionization ion-trap mass spectrometry. Lipids2005;40:211–8.

32. Aradottir S, Moller K, Alling C. Phosphatidyletha-nol formation and degradation in human and ratblood. Alcohol Alcohol 2004;39:8–13.

33. Aradottir S, Seidl S, Wurst FM, Jonsson BA, AllingC. Phosphatidylethanol in human organs andblood: a study on autopsy material and influencesby storage conditions. Alcohol Clin Exp Res 2004;28:1718–23.

34. Hodson L, Skeaff CM, Fielding BA. Fatty acidcomposition of adipose tissue and blood in hu-

mans and its use as a biomarker of dietary intake.Prog Lipid Res 2008;47:348–80.

35. Rivier L. Criteria for the identification of compoundsby liquid chromatography-mass spectrometry andliquid chromatography-multiple mass spectrometryin forensic toxicology and doping analysis. AnalChim Acta 2003;492:69–82.

36. Kugelberg FC, Jones AW. Interpreting results ofethanol analysis in postmortem specimens: a re-view of the literature. Forensic Sci Int 2007;165:10–29.

37. O’Neal CL, Poklis A. Postmortem production ofethanol and factors that influence interpretation.Am J Forens Med Pathol 1996;17:8–20.



molecular species of the alcohol biomarker ... · molecular species of the alcohol biomarker...

Documents