PAPER IN FOREFRONT

Use of LC–MS–MS as an alternative to currently availableimmunoassay methods to quantitate corticosterone in egg yolkand albumen

Siegrid De Baere & Tom Rosendahl Larsen & Mathias Devreese &

Patrick De Backer & Liesbeth De Neve & Graham Fairhurst &Luc Lens & Siska Croubels

Received: 13 August 2014 /Revised: 12 October 2014 /Accepted: 13 October 2014# Springer-Verlag Berlin Heidelberg 2014

Abstract Corticosterone (CORT) is the dominant plasmaglucocorticoid in birds. There has been increasing interest inthe function of CORT in avian egg yolk and in the potential touse CORT concentrations in eggs to quantify stress and toassess the effect of maternal stress on offspring. The concen-tration of CORT in egg yolk is most frequently assessed usingenzyme or radioimmunoassays, alone or in combination withhigh-performance liquid chromatography. However, the quan-tification of CORT is frequently hampered by the presence ofhigh concentrations of other steroid hormones which cross-react with the CORT antibody. As an alternative, we devel-oped a sensitive and specific LC–MS–MS method. Thesample-preparation procedure consisted of a protein–lipidprecipitation step, followed by defatting and clean-up usinga C18 SPE column. Chromatography was performed on anAcquity C18 BEH column (50 mm×2.1 mm i.d., dp: 1.7 μm,run-time: 6 min), using 0.1 % formic acid in both water (A)and acetonitrile (B) as mobile phases. TheMS–MS instrument

was operated in the positive-electrospray-ionization mode.The method was validated in-house according to EuropeanGuidelines (linearity, accuracy and precision, limits of quan-tification and detection, specificity, stability) and the resultsfell within the accepted ranges. The method was successfullyused for the analysis of CORT in yolk and albumen of eggscollected from eight breeding lesser black-backed gulls at aFlemish coastal colony. CORT concentrations were in therange 42.4–166.3 pg g−1 in albumen and<LOQ (75 pg g−1)–762.5 pg g−1 in yolk.

Keywords Corticosterone . LC–MS–MS . Eggs . Yolk .

Albumen . Steroid hormones

Abbreviations11-deoxy-CRTL 11-Deoxycortisol17-α-OH-PROG 17-α-Hydroxyprogesteronea Slope calibration curveALD AldosteroneANDR AndrostenedioneAPCI Atmospheric-pressure chemical

ionizationCORT CorticosteroneCORT-d8 Deuterated corticosteroneCRTL CortisolDHEA DehydroepiandrosteroneEIA Enzyme immunoassayESI Electrospray ionizationESTR EstradiolEU European UnionFDA Food and Drug Administrationg Goodness-of-fit coefficientHPLC High-performance liquid

chromatographyIS Internal standard

Published in the topical collection on Hormone and Veterinary DrugResidue Analysiswith guest editors Siska Croubels, Els Daeseleire, SarahDe Saeger, Peter Van Eenoo, and Lynn Vanhaecke.

Electronic supplementary material The online version of this article(doi:10.1007/s00216-014-8269-7) contains supplementary material,which is available to authorized users.

S. De Baere (*) :M. Devreese : P. De Backer : S. CroubelsDepartment of Pharmacology, Toxicology and Biochemistry,Laboratory of Pharmacology and Toxicology, Faculty of VeterinaryMedicine, Ghent University, 9820 Merelbeke, Belgiume-mail: siegrid.debaere@ugent.be

T. Rosendahl Larsen : L. De Neve : L. LensDepartment of Biology, Terrestrial Ecology Unit, Faculty of Science,Ghent University, 9000 Ghent, Belgium

G. FairhurstDepartment of Biology, University of Saskatchewan, Saskatoon,SK S7N 5E2, Canada

Anal Bioanal ChemDOI 10.1007/s00216-014-8269-7

LC–MS–MS Liquid chromatography tandemmass spectrometry

LOD Limit of detectionLOQ Limit of quantificationL–PPT Lipid–protein precipitationMRM Multiple reaction monitoringm/z Mass-to-charge ratioN2 NitrogenPBS Phosphate-buffered salinePREG PregnenolonePROG ProgesteroneQC Quality controlr Correlation coefficientRA Apparent recoveryRE Extraction recoveryRIA Radioimmunoassayrpm Rotations per minuteRSD Relative standard deviationRRT Relative retention timeS/N Signal-to-noise ratioSPE Solid-phase extractionSSE Signal suppression and/or

enhancementTEST TestosteroneTr Retention time

Introduction

Maternal steroid hormones are present in the yolks of avianeggs [1, 2]. Levels of these hormones vary both within andbetween clutches in response to a range of factors, includingposition in the laying order, food availability, season, qualityof the male, and population density and social interactions,and have both short and long-lasting effects on offspringdevelopment [3]. Most studies investigating these maternaleffects have focused on the measurement of androgens andgestagens in eggs of different species [2]. In recent years, therehas been an increasing interest in the function of corti-costerone (CORT)—the major glucocorticoid in birds—in avian egg yolk and albumen, and in the potential touse CORT concentrations in eggs to quantify maternalstress and its effect on early offspring development andfitness [4–8]. However, mechanisms and effects are notyet clearly known.

The concentration of CORT in avian egg yolk and albumenis most frequently assessed using enzyme (EIA) [6] or radio-immunoassays (RIA) [8, 9]. However, analysis of CORT inegg yolk is hampered by the presence of high concentrationsof gestagens, including progesterone (PROG), pregnenolone(PREG), and cortisol (CRTL), which cross-react with theCORT antibodies used [2, 10]. PROG levels in egg yolk were

reported to be approximately 1000 ng g−1, whereas PREGconcentrations were up to 10-fold higher [1, 4, 11, 12]. Incontrast, reported CORT concentrations in chicken andjunglefowl yolk were in the range 0.2–0.9 ng g−1 [4, 13]. Asa consequence, a small cross-reactivity of the mentionedgestagens with the CORT antibody (e.g. 1.4 % for PROGand 1.3 % for 17-α-hydroxyprogesterone (17-α-PROG)) [2]may result in overestimation of the true CORT concentrationin yolk.

To circumvent this problem, some authors subjected yolksamples to chromatographic purification and fraction collec-tion before determination of the CORTconcentration using animmunoassay [4, 6, 10]. Another alternative is the use ofliquid chromatography coupled with tandem mass-spectrometry detection (LC–MS–MS): this technique avoidsthe challenge of antibody cross-reactivity because of its highspecificity. Until now, only Sas et al. [13] have used LC–MS–MS for the quantitation of CORT in yolk of laying hens.Sample preparation was performed using accelerated solventextraction and the chromatographic run-time was rather long(30 min per sample). Dexamethasone was used as internalstandard. Furthermore, LC–MS–MS has been used by severalauthors for the quantification of steroids in other biologicalmatrices, including plasma [14, 15], serum [16], and tissues[15, 17].

The objective of this study was to develop a highly specificLC–MS–MS method that can be used as an alternative toexisting immunoassays for the quantitative determination ofCORT in egg yolk and albumen. Special attention was paid tothe optimization of the sample-preparation procedure to selec-tively remove interfering substances, for example lipids andproteins, and to obtain a sensitivity that was as high as possi-ble, given that CORT levels in egg yolk and albumen areusually reported to be in the sub-nanogram range [4, 13].

The method was validated in-house according to EuropeanGuidelines [18, 19], and the following variables were evalu-ated: linearity, within-run and between-run accuracy and pre-cision, limits of quantification and detection, recovery, matrixeffect, stability in solvent, extract, and matrix during storage,and stability during three freeze–thaw cycles. Thereafter, themethod was successfully used for the analysis of CORT inyolk and albumen of eggs collected from eight breeding lesserblack-backed gulls at a Flemish coastal colony.

Materials and methods

Chemicals, products, and reagents

The analytical standards of aldosterone (ALD), androstenedi-one (ANDR), corticosterone (CORT), cortisol (CRTL), 11-deoxycortisol (11-deoxy-CRTL), dehydroepiandrosterone(DHEA), estradiol (ESTR), 17-α-hydroxyprogesterone

S. De Baere et al.

(17-α-OH-PROG), pregnenolone (PREG), progesterone(PROG), and testosterone (TEST), were obtained fromSigma–Aldrich (Bornem, Belgium) and stored according tothe manufacturers’ instructions. Corticosterone-d8 (CORT-d8), which was used as internal standard (IS), was purchasedfrom Toronto Research Chemicals Inc. (North York, ON,Canada). The solvents (water and acetonitrile) and additive(formic acid, 99 %) used for the preparation of the UPLCmobile phase were of ULC–MSgrade and were obtained fromBiosolve (Valkenswaard, The Netherlands). All other solventsand reagents were of HPLC grade (water, methanol, andacetonitrile) or analytical grade (formic acid, n-hexane, ethylacetate, and ethanol) andwere purchased fromVWR (Leuven,Belgium). Phosphate buffered saline (PBS) was obtained fromSigma–Aldrich (Bornem, Belgium).

Bond Elut C18 (100 mg mL−1) solid-phase extraction(SPE) columns were from Agilent (Diegem, Belgium).

Millex-GN nylon filter units (0.22 μm) were obtained fromMerck–Millipore (Overijse, Belgium).

Preparation of standard solutions

Stock solutions of all steroids were prepared in methanol at aconcentration of 1.0 mg mL−1, except for ANDR, which waspurchased as a 1.0 mg mL−1 solution in acetonitrile.

Working solutions of 10.0 μg mL−1, 100.0 ng mL−1,10.0 ng mL−1, 1.0 ng mL−1, and 0.1 ng mL−1 CORT wereprepared in methanol by appropriately diluting the stock so-lution. For CORT-d8, a working solution of 10.0 ng mL−1 wasprepared in methanol.

A combined standard solution of 10μg mL−1 of all steroidswas prepared in methanol. To test the specificity of theLC–MS–MS method, the solution was further diluted withmethanol to reach a final concentration of 1 ng mL−1 of allsteroids.

All stock and working solutions were stored at≤−15 °C forthree months. Thereafter, they were replaced by freshly pre-pared solutions.

Egg samples

Chicken egg samples

For the preparation of matrix-matched calibrator and qualitycontrol (QC) samples, commercially bought chicken eggs(n =6) were used. The albumen and yolk were separatedmanually for each egg. Thereafter, the individual albumenand yolk samples were combined by vortex mixing and storedat≤−15 °C. Before use, the combined albumen and combinedyolk were checked for the basal CORT levels, which were41.5 and 108.3 pg g−1, respectively.

Gull egg samples

To evaluate the applicability of the developed method, fullclutches from eight lesser black-backed gull breeding pairswere collected during spring 2013 at a restricted area of theouter port of Zeebrugge (Belgium, 51°21’N, 03°11’E) thathosts ca. 4500 breeding pairs annually. After collection, theeggs were frozen at≤−75 °C until separation of albumen andyolk. For separation, the total eggs were weighed and thawedat room temperature for 10 min, followed by the removal ofthe egg shell. The eggs were transferred to a glass beaker andthawed for an additional hour, followed by the separation ofthe liquid albumen from the frozen yolk. From each fraction,two 1 g aliquots were weighed in a 15 mL falcon tube andstored at≤−75 °C until analysis.

Sample pretreatment

Sample processing was performed in plastic tubes to prevent theadsorption of steroids to glassware [15]. One hundred μL ISworking solution (CORT-d8, 10 ngmL−1) was added to each 1 gegg yolk or albumen, followed by a vortex mixing (15 s) andequilibration (5 min, room temperature) step. After the additionof 5 mL 1% formic acid in acetonitrile, the samples were vortexmixed for 30 s and centrifuged for 10 min at 4750 rpm and 4 °C.A 2.5 mL aliquot of the supernatant was transferred to another15 mL falcon tube and defatted by vortex mixing for 30 s with2.5 mL n-hexane. After centrifugation (10 min, 4750 rpm, 4 °C)and removal of the n-hexane phase, the lower phase was trans-ferred to another tube and evaporated using a gentle stream ofnitrogen (N2; ~50 °C). The dry residue was reconstituted in100 μL methanol, vortexed for 15 s, and diluted with 900 μLwater. The sample was further purified using a C18 SPE column,which was previously conditioned with 1 mL methanol and1 mL water and was washed—after sample loading—with1 mL water and 1 mL n-hexane. The analytes were eluted using2 mL ethyl acetate, and the eluate was evaporated until dry. Theresidue was redissolved in 125 μL methanol and—after vortexmixing (15 s)—diluted with an equal volume of water. Thesample was passed through a 0.22 μm Millex nylon syringefilter and transferred to an autosampler vial. A 10 μL aliquotwas injected onto the LC–MS–MS instrument.

LC–MS–MS method for quantification

The LC system consisted of an Acquity UPLC H-ClassQuaternary Solvent Manager and Flow-Through-NeedleSample Manager with temperature-controlled tray and columnoven from Waters (Zellik, Belgium). Chromatographic separa-tion was achieved on an Acquity UPLC BEH C18 column(50 mm×2.1 mm i.d., dp: 1.7 μm) in combination with anAcquity BEH C18 1.7 μm Vanguard pre-column, both fromWaters.

LC-MS-MS Analysis of cortocosterone in egg yolk and albumen

Mobile phase A consisted of 0.1 % formic acid in water,and mobile phase B was 0.1 % formic acid in acetonitrile. Agradient elution was performed: 0–0.2 min (80 % A, 20 % B),2.0 min (linear gradient to 80 % B), 2.0–3.8 min (20 % A,80 % B), 4.0 min (linear gradient to 80 % A), 4.0–6.0 min(80 % A, 20 % B). The flow was 400 μL min−1.

The temperatures of the column oven and autosampler traywere set at 60 °C and 8 °C, respectively.

The UPLC column effluent was interfaced to a Xevo TQ-SMS–MS system, equipped with an electrospray-ionization(ESI) probe operating in the positive mode (all fromWaters). A divert valve was used, and the UPLC effluentwas directed to the mass spectrometer from 1.5 to 2.5 min.

Instrument settings were optimized by direct infusion ofworking solutions of 0.1 μg mL−1 each compound (flow:10 μL min−1) in combination with the mobile phase (50 %A, 50 % B, flow: 200 μL min−1).

The following settings were used: capillary voltage: 2.0 kV,cone: 20 V, source offset: 50 V, desolvation temperature:600 °C, desolvation gas: 1000 L h−1, cone gas: 150 L h−1,nebulizer pressure: 7.0 bar, LM resolution 1 and 2: 2.72 and2.86, respectively, HM resolution 1 and 2: 14.90 and 15.15,respectively, ion energy 1 and 2: 0.2 and 0.7, respectively,collision gas flow: 0.15 mL min−1.

MS–MS acquisition was performed in multiple-reaction-monitoring (MRM) mode. The MRM transitions that weremonitored for CORT and the internal standard are shown inTable 1.

LC–MS–MS method for specificity testing

The UPLC inlet method and the MS tune settings were the sameas those used for CORT quantification (see section above). TheMS–MS method contained the MRM transitions for CORT andCORT-d8, and those for all steroid hormones used for specificitytesting (see Table 1). No divert valve was used.

In-house method validation

In-house validation was performed on chicken yolk and albu-men, because this matrix was commercially available. A set ofvariables that were in compliance with reference guidelinesdefined by the EU [18, 19] and in the literature [20–22] wereevaluated.

Linearity

Calibration curves were generated to evaluate the linear rela-tionship between instrument response (i.e. peak-area ratioCORT/IS) and known concentration of CORT (y=ax+b).Matrix-matched calibration-curve samples were prepared byspiking 1 g aliquots of combined chicken yolk or albumenwith CORT levels of 10, 25, 50, 100, 250, 500, 1000, and

2500 pg g−1. The standard working solutions were directlyapplied onto the homogenized samples, together with the ISsolution, followed by a vortex-mixing step. After 5 min equil-ibration, the sample-preparation procedure was performed asdescribed above. The correlation coefficient (r) and goodness-of-fit coefficient (g) were determined, and had to be ≥0.99 and≤20 %, respectively [18, 20]. The basal concentration ofCORT in the combined chicken yolk or albumen and in theunknown samples was calculated by dividing the peak arearatio CORT/IS by the slope of the calibration curve (a). Inaddition, a calibration curve based on extracted standard so-lutions was prepared by spiking and extracting 1 mL aliquotswater in the same way as the matrix-matched calibratorsamples.

Accuracy and precision

Within-run precision (repeatability) and accuracy were deter-mined by analyzing six chicken yolk or albumen samples thatwere spiked with CORT at a low (25 pg g−1), medium(100 pg g−1), and high (1000 pg g−1) concentration on thesame day. The between-run precision and accuracy weredetermined by analyzing quality-control samples spiked withCORT at the same concentration levels together with eachanalytical batch of samples, run on different days (n ≥2 perconcentration and per day). Accuracy was evaluated by cal-culating the percentage of deviation of the measured valuecompared with the theoretical spiked value. The acceptancecriteria for accuracy were: −50 % to +20 % and −30 % to +10 % for concentrations <1 ng g−1 and 1–10 ng g−1, respec-tively. The precision was evaluated by calculating the relativestandard deviation (RSD), which had to be below the RSDmax

values of 30 % (for concentrations <1 ng g−1) and 25 %(concentrations of 1–10 ng g−1) [19]. Because RSDmax

values for between-run precision would be very high(45.3–90.5 %) when calculated by the Horwitz equation(RSDmax =2(1-0.5logConc)), it was decided to use the sameRSDmax values as for the within-run precision.

Limit of quantification

Because CORT is a naturally occurring steroid hormone, itwas difficult to obtain blank yolk and albumen. The limit ofquantification (LOQ) was therefore determined as follows: sixchicken yolk or albumen samples were spiked with a CORTconcentration of 10 pg g−1, which corresponded to a realCORT concentration of 136.2 pg g−1 and 48.0 pg g−1 in yolkand albumen, respectively, if the basal CORT concentrationwas taken into account. The accuracy, within-run precision,and signal-to-noise ratio (S/N) were evaluated. The LOQ wasdetermined by calculating the CORT concentration thatcorresponded to an S/N of 9, i.e. 25 pg g−1 for albumen and75 pg g−1 for yolk. To confirm this LOQ, six aliquots of a gull

S. De Baere et al.

egg sample containing CORT concentrations of ~75.0 pg g−1

(yolk) and ~25.0 pg g−1 (albumen) were extracted using thesample-preparation procedure described above, and analyzedusing LC–MS–MS. The accuracy, precision, and S/N ratiowere evaluated.

Limit of detection

The limit of detection (LOD) was defined as the lowestconcentration of the analyte that could be recognized by thedetector with an S/N ≥3. The LOD values were calculatedusing the samples spiked at a CORT level of 10 pg g−1, andconfirmed by the analysis of six 1 mL aliquots of a standardsolution containing 25.0 pg mL−1 (yolk) and 10.0 pg mL−1

(albumen) CORT.

Specificity

The specificity of the method was evaluated with respect to othersteroid hormones which can naturally occur in yolk or albumenor which may have cross-reactivity in the RIA and EIA assaystraditionally used for the analysis of CORT. Therefore, a standardmixture containing CORT, CORT-d8, and ten other steroids(ALD, ANDR, CRTL, 11-deoxy-CRTL, DHEA, ESTR, 17-α-OH-PROG, PREG, PROG, and TEST) at a concentration of1 ng mL−1 was injected using the LC–MS–MS method forspecificity testing. The (relative) retention time of each analyte

was evaluated. In addition, a standard mixture containing allsteroids, except CORT and CORT-d8, was injected using theLC–MS–MS method for quantification. Possible interference ofthe analytes in the MRM trace of CORTwas evaluated.

Carry-over

The carry-over was evaluated by analyzing a solvent sample(water–methanol, 50:50, v/v) just after the highest calibratorsample. The S/N of the eventual CORT peak in the solventsample had to be <3.

Storage stability

The storage stability of CORT was evaluated in stock(1 mg mL−1) and working (10 ng mL−1) solutions duringstorage at≤−15 °C. A solution was regarded as stable if thepeak area of CORT fell within ±10 % of the correspondingpeak area for a freshly prepared solution. Furthermore, thestability of CORT was evaluated in extracted samples duringstorage at 2–8 °C, in matrix during storage at≤−75 °C, andduring three freeze–thaw cycles (freezing at≤−75 °C andthawing to room temperature). For each stability experiment,three chicken yolk and albumen samples were spiked withCORT at concentrations of 100 and 1000 pg g−1. CORTconcentrations were determined using freshly prepared

Table 1 Multiple-reaction-monitoring (MRM) transitionsa, MS–MS conditions, and retention times (Tr) for CORT, CORT-d8, and 10 steroid hormonesused to investigate the specificity of the method

Analyte MMb

(g mol−1)Precursor ion(m/z)c

Product ions(m/z)

Cone(V)

Collisionenergy (eV)

Dwelltime (s)

Tr(min)

ALD 360.4 361.4 343.2 30 22 0.010 1.69

CTRL 362.5 363.2 121.0 30 28 0.010 1.85

CORT-d8 354.2 355.2 125.1d

337.3e4545

1515

0.0250.025

2.10

CORT 346.3 347.3 97.1e

121.0d4545

2422

0.0250.025

2.11

TEST 288.4 289.2 97.0 30 22 0.01 2.36

ANDR 286.4 287.1 97.0 30 22 0.01 2.48

DHEA 288.4 271.1 213.3 30 25 0.01 2.49

PROG 314.5 315.2 97.0 30 22 0.01 2.85

PREG 316.5 317.2 281.3 30 25 0.01 2.85

17-α-OH-PROG 330.5 331.2 97.0 30 24 0.01 2.50

ESTR 272.4 255.0 159.0 30 25 0.01 2.30

11-deoxy-CTRL 346.5 347.2347.3

97.0121.0

3045

2522

0.010.025

2.16

aMRM transitions were based on the TargetLynx Quanpedia database of the Xevo TQ-S system and on Refs. [14, 16, 24]bMolecular masscMass-to-charge ratiod Quantifier ione Qualifier ion

LC-MS-MS Analysis of cortocosterone in egg yolk and albumen

calibration curves and had to fall within −30% to +10% of thetheoretical concentrations.

Recovery and matrix effects

Experiments evaluating the extraction recovery (RE), apparentrecovery (RA), and matrix effects (signal suppression and/orenhancement, SSE) were performed as follows: blank matrixwas spiked before extraction (= spiked) and after extraction(= spiked extract) at a concentration of 10 ng g−1. In addition,standard solutions containing CORT and CORT-d8 at thesame concentration were prepared [17, 22].

The extraction recovery and apparent recovery were deter-mined by comparing the peak areas of the analytes in thespiked samples with those in the spiked extracts (RE) and thestandard solutions (RA).

The signal suppression and/or enhancement was evaluatedby comparing the peak areas of the analytes in the spikedextracts with the related peak areas in the standard solutions.All experiments were performed in triplicate.

Results and discussion

During the optimization of a method for the analysis of CORTin avian eggs, two points of interest were taken into account:

1. the sensitivity had to be as high as possible, because eggCORT levels reported in the literature were in the sub-nanogram concentration range; and

2. the method had to be specific to avoid cross-reaction withother steroids that are present in high concentrations inavian eggs.

Sample extraction and clean-up

Egg yolk and albumen contain high amounts of lipids andproteins, respectively, which can interfere with steroid analysis.Sample preparation is therefore especially important for LC–MS–MS analysis, to minimize matrix effects and to maximizesensitivity. Traditionally, proteins or nonpolar lipids are precipi-tated using ethanol [9] or methanol [6], or steroids are extractedusing organic solvents (liquid–liquid extraction using petroleumether–diethyl ether (30:70, v/v) in combination with Celite col-umn chromatography [7]). Other authors combine precipitationof lipids with further purification of the extract using solid-phaseextraction (SPE), which gives superior results in terms of recov-ery and removal of interfering compounds (protein precipitationwith 100 % methanol in combination with Sep-Pak C18cartridge [2, 6]). Sas et al. used accelerated solvent extraction(ASE-200) [13] before LC–MS–MS.

For this study, several extraction procedures of increasingcomplexity were tested:

1. lipid–protein precipitation (L–PPT) using 1% formic acidin acetonitrile;

2. L–PPT in combination with a defatting step using n-hexane; and

3. L–PPT+defatting in combination with solid-phase extrac-tion (SPE) using a C18 extraction column.

SPE was based on the procedure of Koren et al. [14] for theanalysis of steroids in animal plasma. The SPE columns werewashed using two solvents with different polarities, i.e. waterand n-hexane, to remove polar interferences and lipids, re-spectively, without losing CORT [14]. Ethyl acetate was usedas elution solvent, rather than methanol, because it can beevaporated much more quickly. During the elution step,the substances causing the yellow color of the yolkextract (xanthophylls) were completely eluted from theC18 column, but they were retained by the 0.22 μm nylonsyringe filter during the final filtration step, resulting in clearextracts.

Recovery and matrix effects

LC–MS–MS is known for its specificity and selectivity, but ithas been revealed that co-eluting matrix components mayaffect the ionization efficiency. This phenomenon can bereduced by performing an extensive sample clean-up, andtherefore we tested several extraction procedures. Each time,the extraction recovery, matrix effect, and apparent recoverywere evaluated. As shown in Fig. 1, the RE reduced withincreasing complexity of the sample preparation procedure,whereas the opposite was observed for matrix effects. Tominimize contamination of the LC–MS–MS instrument, theprocedure consisting of lipid–protein precipitation, defatting,and C18 SPE was chosen as the final procedure, resulting inan RE of 82.6 % and 71.3 %, an SSE of 90.5 % and 97.4 %,and an RA of 74.7 % and 69.5 % for CORT in egg albumenand yolk, respectively.

Matrix effects were further minimized by the use of anisotope-labelled internal standard (CORT-d8).

LC–MS–MS analysis

Chromatography

Chromatography was performed on an Acquity H-ClassUPLC instrument using a reversed-phase BEH C18 column(50 mm×2.1 mm i.d.) with a small particle size (1.7μm). Thishad the advantage that the chromatographic analysis of onealbumen or yolk sample could be accomplished within 6 min,which is much shorter than the time for the LC–MS–MS

S. De Baere et al.

method reported by Sas et al. (30 min) [13]. In addition,CORT (Tr =2.11 min, relative retention time (RRT) =1.006)could be distinguished from the isobaric compound 11-deoxycortisol (Tr =2.16 min, RRT =1.029) (see Fig. 2aand Table 1), which was important for the specificity ofthe method [16] and usually not investigated by otherauthors [13, 15].

Mass spectrometry

CORT and CORT-d8 were tuned in both the positive atmo-spheric-pressure-chemical- ionization (APCI) andelectrospray-ionization (ESI) mode, with comparable sensi-tivity. Although several reported methods for the analysis ofsteroid hormones in biological matrices use APCI [13, 14, 16],the ESI mode was chosen for this method because this wasmost frequently used at our laboratory.

Detection was performed in the multiple-reaction-monitoring (MRM) mode. After fragmentation of the proton-ated molecular ion [M−H]+ of CORT and CORT-d8, the twomost intense product ions were selected for quantitative andqualitative purposes, respectively. Quantitation was per-formed using the MRM transitions with m/z 347.3>121.0(CORT) and 355.2>125.1 (CORT-d8), and the product ionswith m/z 97.2 and 337.3 were used as the respective qualifierions. In a real sample, a peak was identified as CORT on thebasis of the calculation of the relative retention time (RRT=TrCORT/TrCORT-d8) and the ion ratio (= peak areaqualifier ion/peakareaquantifier ion). The identity of a peak was confirmed if theion ratio fell within ±25 % of the mean ion ratio of the spikedsamples [18] and if the RRT fell within ±1% of the highest andlowest RRT of CORT in the spiked samples (calibrator andquality control (QC) samples) of the same analytical batch.This latter criterion was more stringent than the ±2.5 % rangefor RRT that is allowed for LC–MS–MS applications in the2002/657/EC document, but could be justified by the use of ahighly reproducible UPLC–MS–MS instrument (see below).

Internal standard

Internal standardization was used to compensate for analytelosses during sample preparation and for matrix effects duringLC–MS–MS analysis. An isotope-labelled internal standard(CORT-d8) was chosen, because the structural, physical, andchemical properties of such a compound are very similar tothose of the analyte of interest. In addition, the internal stan-dard was used for analyte identification by calculation ofrelative retention times (TrCORT/TrCORT-d8), and forquantification.

Method validation

Linearity Two methods of quantification were tested: usingspiked yolk and albumen calibration samples originating fromcombined chicken eggs (= matrix-matched calibration) andspiked water samples (= standard calibration) that had beensubjected to the sample-preparation procedure.

Using the matrix-matched calibration method, nine-pointcalibration curves—including a zero-calibrator sample—wereconstructed by plotting the peak-area ratio of CORT/IS versusthe spiked CORTconcentrations and applying weighted linear

0.00

20.00

40.00

60.00

80.00

100.00

120.00a

b

c

L/PPT L/PPT-Hex L/PPT-Hex-C18

Ext

ract

ion

rec

ove

ry (

%)

Albumen

Yolk

0.00

20.00

40.00

60.00

80.00

100.00

L/PPT L/PPT-Hex L/PPT-Hex-C18

Sig

nal

su

pp

ress

ion

/en

han

cem

ent

(%)

Albumen

Yolk

0.00

20.00

40.00

60.00

80.00

100.00

L/PPT L/PPT-Hex L/PPT-Hex-C18

Ap

par

ent

reco

very

(%

)

Albumen

Yolk

Fig. 1 (a) Extraction recovery, (b) matrix effect (signal suppression orenhancement), and (c) apparent recovery for CORT (10 ng g−1) inalbumen and yolk with different sample-clean-up procedures: lipid–pro-tein precipitation (L–PPT); L–PPT+defatting with n-hexane (L–PPT-Hex); L–PPT+defatting with n-hexane+C18 solid-phase extraction(L–PPT-Hex-C18)

LC-MS-MS Analysis of cortocosterone in egg yolk and albumen

regression (factor 1/x). The calibration curves covered aspiked concentration range of 10 to 2500 pg g−1, whichencompassed the concentrations observed in real samples(see section “Analysis of biological samples”). As can be seenfrom Fig. 3a, positive intercepts were obtained because of thepresence of endogenous CORT levels in chicken yolk (108.3±17.0 pg g−1) and albumen (41.5±5.6 pg g−1). CORT concen-trations in the gull samples were determined by applying theformula x=y/a (x=CORT concentration (pg g−1); y=area ratioCORT/CORT-d8, and a=slope). As can be seen from Table S1in the Electronic Supplementary Material (ESM), correlationcoefficients (r) were ≥0.99 and goodness-of-fit coefficients (g)≤20% for both albumen and yolk, indicating the good linearityof the method.

In comparison, CORT concentrations were also determinedusing a nine-point standard calibration curve by applying theformula x=(y−b)/a (b=intercept). As can be seen from Fig. 3b,a good correlation (r2>0.99) could be obtained between theresults obtained with both types of calibration curves, whichmeans that the use of a matrix-matched calibrationmodel is notreally necessary. This indicates that the deuterated internal

standard fully compensates for matrix effects on the LC–MS–MS instrument, which is in accordance with [23].

Within-run and between-run accuracy and precision The ac-ceptability ranges for within-run and between-run precisionand accuracy were met for all compounds at the specifiedlevels (see ESM Table S2).

LOQ and LOD The results for the LOQ and LOD evaluationare shown in ESM Table S2. As can be seen, the calculatedLOQ for egg yolk (~75 pg g−1) was threefold higher than thatfor egg albumen (~25 pg g−1). This could be attributed to thelower apparent recovery for CORT in yolk, caused by thehigher complexity of this matrix compared with albumen. Thecalculated LOQ values were confirmed by analyzing the S/Nof a gull yolk and a gull albumen sample that contained mean(n =6) CORT levels of 79.3 pg g−1 and 29.6 pg g−1, respec-tively. To confirm the calculated LOD values, standard solu-tions with CORT concentrations of 10 and 25 pg mL−1 weresubjected to the sample-preparation procedure and the S/N ofthe CORT peak was evaluated.

Analytea b c

CTRL

Time (min) Time (min)

ALD

CORT

11-deoxy-CRL

17-α-OH-PROG

PREG

PROG

TEST

ANDR

DHEA

ESTR1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.8528391

1.695066

2.1234783

2.1576835

2.12;9357

2.5199572

2.891829

2.852209

2.37174104

2.482644

2.48188531

2.491453

2.3024746

95 2.00 2.05 2.10 2.15 2.20 2.25 2

2.1340602

95 2.00 2.05 2.10 2.15 2.20 2.25 2

2.1686234

2.1312148

m/z 121 m/z 97

Rel

ativ

e ab

un

dan

ce

main product ions of precursor m/z 347.3

CORT

11-deoxy-CRL

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

1.50 2.00 2.50 3.00

2.51112622.47

107331.862741

2.933352

1.379452

1.968089 2.35

3753

2.193333

1.622047

2.193172

2.51306032.16

25811

2.87849475

2.095785

3.0957359

2.856215901

2.73378781.92

214102.445817

3.0328603 3.4

50

2.37231166

1.7612545

2.49;144178

2.7211322

2.844044

2.495404732

2.714472

2.8192222.76

66672.042948

2.86286212.09

62871.993358 2.62;8621

3.0910257

m/z 347.4 > 121.1

m/z 347.4 > 97.0

Abu

ndan

ce (

%)

Abu

ndan

ce (

%)

Tr = 2.12 min

Mixed standard solution (5 ng mL-1) Gull yolk sample (CORT: 213 pg g-1)

0

0.2

0.4

0.6

0.8

1

Fig. 2 Specificity of the method: (a) LC–MS–MS chromatogram of amixed standard solution containing CRTL, ALD, CORT, 11-deoxy-CRTL, 17-α-OH-PROG, PREG, PROG, TEST, ANDR, DHEA, andESTR at a concentration of 5 ng mL−1, revealing that other steroidhormones do not interfere with the trace of CORT; (b) extracted ion

chromatogram showing the MRM traces m/z 347.4>121.1 and m/z347.4>97.0 for CORT and 11-deoxy-CRTL, respectively, and the differ-ence in relative abundance of the product ions with m/z =121.1 and 97.0for both compounds; (c) LC–MS–MS chromatogram of a gull yolksample with a CORT concentration of 213 pg g−1

S. De Baere et al.

The obtained LOQ values were low enough to enable thequantification of endogenous CORT levels in egg yolk andalbumen of avian species. The LOD values were in the samerange as or lower than those reported by other authors for thequantification of CORT in yolk using EIA or RIA: LOD,10 pg g−1 yolk [4], 0.3 ng g−1 yolk [2, 6, 10].

Specificity The results of the specificity experiments are shownin Fig. 2. As can be seen from Fig. 2a, a good chromatographicseparation between CORT and the investigated steroids wasachieved, except for the isobaric compound 11-deoxy-CRTL,which was not completely baseline resolved (Fig. 2b). However,interference from 11-deoxy-CRTL was limited because the rel-ative abundance of the ion with m/z =121.1 compared with theionwithm/z =97.0was completely different for CORT (ion ratio~2.10) versus 11-deoxy-CRTL (ion ratio ~0.05) (Fig. 2b). Thismeans that identification of an unknown peak as CORTcould beachieved on the basis of the ion ratio (m/z 121.1 or 97.0).

Identification was also performed on the basis of the RRT,which had to fall within a ±1 % range of the highest and lowestRRT of CORT in spiked samples of the same analytical batch.From the validation experiments it was observed that the RRTof

CORT in spiked samples was highly reproducible, with a meanvalue of 1.006±0.0017 (n =72, RSD =0.17 %) in albumen and1.006±0.0012 (n =96, RSD =0.12 %) in yolk. On the basis ofthese results, an RRT range of 0.996–1.016 was determined,which did not overlap with the RRT range for 11-deoxy-CTRL(1.019–1.039). ThemeanRRTof CORT in incurred albumen andyolk samples was 1.006±0.0015 (n =89, RSD =0.15 %) and1.005±0.0018 (n =93, RSD =0.18 %), respectively, and fellwithin the range of 0.996–1.016.

These results indicate that the LC–MS–MS method wepresent here is highly specific for the determination ofCORT, which is of utmost importance for correct quantifica-tion of CORT in yolk and albumen. This technique is superiorto the EIA and RIA methods reported in the literature, whichsuffer from cross-reactivity of other steroid hormones (pro-gesterone, cortisol, pregnenolone, 17-α-OH-PROG, andro-stenedione, and testosterone) with the CORT antibody [2, 6,10, 12]. In Fig. 2c, a chromatogram is shown from the analysisof a yolk extract using the LC–MS–MSmethod for specificitytesting. As can be seen, the following steroid hormones couldbe detected: CRTL (Tr =1.86 min), 17-α-OH-PROG(Tr =2.51 min), PREG (Tr =2.87 min), PROG (Tr =2.85 min,

0

0.5

1

1.5

2

2.5

3a

b

0 200 400 600

Are

a ra

tio

CO

RT

/IS

CORT concentration (pg g-1)

stand

yolk

albumen

Linear (stand)

Linear (yolk)

Linear (albumen)

y = 0.915x - 7.0064 R² = 0.9981

0.0

100.0

200.0

300.0

400.0

500.0

600.0

700.0

800.0

0.0 200.0 400.0 600.0 800.0 1000.0

CO

RT

yo

lk c

on

cen

trat

ion

(p

g g

-1),

stan

dar

d c

alib

rati

on

CORT yolk concentration (pg g-1), matrix-matched calibration

Fig. 3 (a) Calibration curves forCORT using standard or matrix-matched (yolk, albumen) calibra-tion samples, calibration model:y=ax+b. (b) Line-fit plot of theresults of the analysis of CORT ingull yolk samples using the stan-dard and matrix-matched calibra-tion models; r2=correlationcoefficient

LC-MS-MS Analysis of cortocosterone in egg yolk and albumen

signal off the scale, indicating the high concentrations ofPROG in the egg yolk), TEST (Tr =2.37 min), and andro-stenedione (Tr =2.49 min). In egg albumen the followingsteroid hormones could be determined, but to a lesser extentthan in egg yolk: CRTL, 17-α-OH-PROG, PREG, PROG,TEST, ANDR, and DHEA (chromatogram not shown).

These findings reveal that this method could be used for thequantitative determination of other steroids in albumen andyolk of oviparous species, which could be an interesting futureapplication and a valuable alternative to the currently usedimmunoassays.

Carry-over No carry-over of analytes from one sample toanother was observed.

Stability The results of the stability testing are shown in ESMTable S3. As can be seen, CORTwas stable for at least three daysin extracted samples (stored at 2–8 °C), for at least three monthsin stock and working solutions during storage at≤−15 °C and inmatrix at≤−75 °C, and during three freeze–thaw cycles.

Analysis of biological samples

The applicability of the developed LC–MS–MS method wasrevealed by the analysis of CORT in yolk and albumen oflesser black-backed gull eggs (n =23) collected at a naturalbreeding colony.

Analysis of egg yolk

The distribution of CORTcan vary within the egg layers [6], andtherefore the yolk was mixed using a vortex mixer before thesampling of 1 g aliquots for analysis. CORTconcentrations werein the range<LOQ (75 pg g−1)–762.5 pg g−1, and varied betweendifferent eggs of one clutch. These concentrations were in thesame range as those reported by Sas et al. (200–900 pg g−1 inyolk of caged and free-range hens) [13], Rettenbacher et al. [4](800 pg g−1 in junglefowl yolk), and Navara et al. [7] (770–1610 pg g−1), but lower than those reported by Almasi et al. [6](13–27 ng g−1 in yolk of control American barn owl).

Analysis of egg albumen

The concentrations of CORT in egg albumen ranged from42.4 pg g−1–166.3 pg g−1 and were lower than those reportedby Navara et al. [7] (0.44–0.52 ng mL−1) and Saino et al. [8](~1–1.5 ng mL−1). However, the higher concentrations in thelatter study can be attributed to the fact that no chromatographicpurification of the albumen extract was performed beforeradioimmunoassay.

These results indicate that this method can be used toinvestigate relationships between natural or human-inducedmaternal stress and the production of the stress hormoneCORT in avian species. A paper from a study that used thismethod to relate effects of timing of breeding on early off-spring development in lesser black-backed gulls is currently inpreparation (Larsen et al., in prep). Detailed quantitative re-sults for CORT in albumen and yolk of the investigated eggswill be given in this article.

Conclusions

We successfully developed a highly sensitive and specific LC–MS–MS method for the quantitative determination of CORT inavian yolk and albumen. The method was validated according toEU regulations, and detailed results for linearity, within-run andbetween-run precision and accuracy, LOQ, LOD, and stabilitywere presented. In addition, the specificity was obtained withrespect to ten other steroid hormones which can be present inyolk and albumen at high concentrations, and this is an advan-tage over currently used EIA or RIA methods.

Yolk and albumen samples taken from lesser black-backedgull clutches were quantitatively analyzed using the describedLC–MS–MS method. The method was sensitive enough todetermine CORT levels in both matrices at concentrationsdown to 25 pg g−1 in albumen and 75 pg g−1 in yolk.

In conclusion, this is—as far as we are aware—the firsttime a validated LC–MS–MS method has been reported forthe quantitative analysis of CORT at low concentrations inboth avian yolk and albumen. This method can be used as analternative to currently used EIA and RIA methods, which areknown to suffer from cross-reactivity.

Acknowledgments The authors wish to thankMrs J. Lambrecht for herkind assistance during sample analysis, Hans Matheve for his help withdata collection, and Marjut Paljakka and Cátia Santos for their help withsample preparation. This studywas financially supported by FWOprojectG0E1614N to LL and LDN.

Conflicts of interest The authors report no conflicts of interest.

References

1. Lipar JL, Ketterson ED, Nolan V, Casto JM (1999) Egg Yolk LayersVary in the Concentration of Steroid Hormones in Two AvianSpecies. Gen Comp Endocrinol 115:220–227

2. Rettenbacher S, Möstl E, Groothuis TGG (2009) Gestagensand glucocorticoids in chicken eggs. Gen Comp Endocrinol164:125–129

S. De Baere et al.

3. von Engelhardt N, Henriksen R, Groothuis TGG (2009) Steroids inchicken egg yolk: Metabolism and uptake during early embryonicdevelopment. Gen Comp Endocrinol 163:175–183

4. Rettenbacher S, Groothuis TG, Henriksen R, Möstl E (2013)Corticosterone in bird eggs: The importance of analytical validation.Wien Tierärztl Monat 100:283–290

5. Groothuis TGG, Schwabl H (2008) Hormone-mediated maternaleffects in birds: mechanisms matter but what do we know of them.Phil Trans R Soc B 363:1647–1661

6. Almasi B, Rettenbacher S, Müller C, Brill S, Wagner H, Jenni L(2012) Maternal corticosterone is transferred into the egg yolk. GenComp Endocrinol 178:139–144

7. Navara KJ, Pinson SE (2010) Yolk and albumen corticosteroneconcentrations in eggs laid by white versus brown caged laying hens.Poult Sci 89:1509–1513

8. Saino N, Romano M, Ferrari RP, Martinelli R, Møller AP (2005) JExp Zool 303A:998–1006

9. Kozlowski CP, Bauman JE, Hahn DC (2009) A SimplifiedMethod forExtracting Androgens From Avian Egg Yolks. Zool Biol 28:137–143

10. Quillfeldt P, Poisbleau M, Parenteau C, Trouvé C, Demongin L, vanNoordwijk HJ, Möstl E (2011) Measuring corticosterone in seabirdegg yolk and the presence of high yolk gestagen concentrations. GenComp Endocrinol 173:11–14

11. Möstl E, Spendier H, Kotrschal K (2001) Concentration of immuno-reactive progesterone and androgens in the yolk of hens’ eggs (Gallusdomesticus). Wien Tierärztl Monat 88(3):62–65

12. Bertin A, Richard-Yris M-A, Houdelier C, Lumineau S, Möstl E,Kuchar A, Hirschenhauser K, Kotrschal K (2008) Habituation tohumans affects yolk steroid levels and offspring phenotype in quail.Horm Behav 54:396–402

13. Sas B, Domány G, Gyimóthy I, Gaál Kovácsné K, Süth M (2006)Influence of the type of management system on corticosterone trans-fer into eggs in laying hens. Acta Vet Hung 54(3):343–352

14. Koren L, Ng ESM, Soma KK, Wynne-Edwards KE (2012) SamplePreparation and Liquid Chromatography-Tandem MassSpectrometry for Multiple Steroids in Mammalian and AvianCirculation. PLoS ONE 7:e32496

15. Samtani MN, JuskoWJ (2007) Quantification of dexamethasone andcorticosterone in rat biofluids and fetal tissue using highly sensitiveanalytical methods: assay validation and application to a pharmaco-kinetic study. Biomed Chromatogr 21:585–597

16. Fanelli F, Belluomo I, Di Lallo VD, Cuomo G, De Iasio R, BacciniM, Casadio E, Casetta B, Vicennati V, Gambineri A, Grossi G,Pasquali R, Pagotto U (2011) Serum steroid profiling by isotopicdilution-liquid chromatography-mass spectrometry: Comparisonwith current immunoassays and reference intervals in healthy adults.Steroids 76:244–253

17. Rönquist-Nii Y, Edlund PO (2005) Determination of corticosteroidsin tissue samples by liquid chromatography-tandem mass spectrom-etry. J Pharm Biomed 37:341–350

18. Commision Decision 2002/657/EC implementing Council Directive 96/23/EC concerning the performances of analytical methods and the inter-pretation of results, Official Journal of the European Communities, N°L221, 17/08/2002, Decision of 12 August 2002, European Commision,Directorate General for Public Health and Consumers Protection

19. VICHGL49 (2012-Final), Guidance for Industry, Studies to evaluatethe metabolism and residue kinetics of veterinary drugs in foodproducing animals: validation of analytical methods used in residuedepletion studies, February 2011

20. Knecht J, Stork G (1974) Prozentuales und Logarithmisches Verfahrenzur Berechnung von Eichkurven. Fresen’ Z Anal Chem 270:97–99

21. R.J. Heitzman, Ed. Veterinary Drug Residues, Report Eur. 14126-EN, Commision of the EC, Brussels-Luxembourg, 1994

22. Matuszewski BK, Constanzer ML, Chavez-Eng CM (2003) Strategiesfor the Assessment of Matrix Effect in Quantitative BioanalyticalMethods Based on HPLC-MS/MS. Anal Chem 75:3019–3030

23. Hewavitharana AK (2011) Matrix matching in liquidchromatography-mass spectrometry with stable isotope labelled in-ternal standards – is it necessary? J Chromatogr A 1218:359–361

24. Abdel-Khalik J, Björklund E, Hansen M (2013) Development of asolid phase extraction method for the simultaneous determination ofsteroid hormones in H295R cell line using liquid chromatography-tandem mass spectrometry. J Chromatogr B 935:61–69

Siegrid De Baere is a post-doctoral researcher at the Labora-tory of Pharmacology and Toxi-cology (Faculty of VeterinaryMedicine, Ghent University, Bel-gium). Her main research focuseson the development of LC–MS–MS methods to quantify xenobi-otics (veterinary medicinal prod-ucts and mycotoxins) and endog-enous compounds in biologicalmatrices of different origin.

TomRosendahl Larsen is a doc-toral fellow at the Terrestrial Ecol-ogy Unit (TEREC), Ghent Uni-versity, Belgium. He is nowworking in the field of bird ecol-ogy with focus on the interactionbetween breeding strategies andstress hormones.

Mathias Devreese is a veterinar-ian working as post-doctoral fel-low at the Laboratory of Pharma-cology and Toxicology (Facultyof Veterinary Medicine, GhentUniversity, Belgium). He hasbeen working for several yearson the development of LC–MS–MS methods to quantify xenobi-otics (veterinary medicinal prod-ucts and mycotoxins) in animalmatrices. He is also resident of theEuropean College of VeterinaryPharmacology and Toxicology.

LC-MS-MS Analysis of cortocosterone in egg yolk and albumen

Patrick De Backer is full profes-sor andManager of theDepartmentof Pharmacology, Toxicology andBiochemistry of the Faculty of Vet-erinary Medicine of Ghent Univer-sity in Belgium. He is diplomate ofthe European College of VeterinaryPharmacology and Toxicology. Hisresearch in veterinary pharmacolo-gy focuses on the pharmacokinet-ics, pharmacodynamics, and resi-dues of veterinary drugs and con-taminants in several animal species;and on the development of suitableanimal models for drug research.

Liesbeth De Neve is a postdoc-toral researcher at the TerrestrialEcology Unit at the Biology De-partment of Ghent University. Hermain research focuses on the ef-fect of life-history trade-offs andenvironmental stressors on avianreproductive behavior. To investi-gate underlying mechanisms sheoften makes use of different phys-iological factors including innateand cell-meditated immune re-sponses, hormones, and contami-nation markers.

Graham D. Fairhurst is a post-doctoral fellow at the Universityof Saskatchewan in Saskatoon,Canada. His research seeks to un-derstand the mechanisms thathelp birds cope with change andhow they contribute to fitnessprospects in a changing world.Since 2006, he has been workingwith corticosterone in feathers in avariety of ecophysiological con-texts with the objective of devel-oping practical biomarkers forecology and conservation.

Luc Lens is full professor of Ter-restrial Ecology and director ofthe Terrestrial Ecology Unit atthe Biology Department of GhentUniversity. He has been workingfor several years on the behavioralecology, population and spatialecology, and stress physiology oftemperate and tropical birds in ur-ban, forested, and coastal land-scapes. Methods comprise aviaryand field experiments, demo-graphic, genetic, eco-physiologi-cal, toxicological, and biometricsampling, behavioral observa-

tions, and the use of state-of-the-art tracking devices.

Siska Croubels is full professorand head of the Laboratory ofPharmacology and Toxicology atthe Faculty of Veterinary Medi-cine of Ghent University in Bel-gium. She is associate member ofthe European College of Veteri-nary Pharmacology and Toxicol-ogy. Her research group in veter-inary toxicology is a member oft h e M Y T O X p l a t f o r m(www.mytox.be), dealing withthe prevention and control ofmolds in the food chain; myco-toxin analysis and method devel-

opment; identification, prediction and characterization of fungi; andeffects of mycotoxins on human and animal health. Her research inveterinary pharmacology focuses on the pharmacokinetics, pharmacody-namics, and residues of veterinary drugs and contaminants in severalanimal species, and on the development of suitable animal models fordrug research.

S. De Baere et al.