animals
Article HPLC MS-MS Analysis Shows Measurement of Corticosterone in Egg Albumen Is Not a Valid Indicator of Chicken Welfare
Malcolm P. Caulfield * and Matthew P. Padula
School of Life Sciences and Proteomics Core Facility, Faculty of Science, The University of Technology Sydney, Ultimo 2007, Australia; [email protected] * Correspondence: malcolm.caulfi[email protected]
Received: 16 April 2020; Accepted: 6 May 2020; Published: 9 May 2020
Simple Summary: Growing interest in the welfare of farmed animals, particularly those in restrictive housing, has stimulated attempts to define simple measures of welfare. One such claimed measure involves the analysis of so-called stress hormones. In chickens, the main ‘stress hormone’ is corticosterone. In Australia, reviews of legislation relating to the welfare of chickens housed in cages have relied heavily on non-invasive measures of corticosterone, in particular those using egg white (albumen). All of those measures have used antibodies to quantify the corticosterone. Recently, doubts have been raised about the specificity of these measurement techniques. In this study, we demonstrate that high-resolution chromatographic separation of extracted egg albumen, followed by mass spectrometry, reveals that corticosterone is barely detectable in chicken egg albumen. Previous work using immunoassays reported levels of 0.5 to 20 ng/g. We have found egg albumen corticosterone concentrations of about 50 pg/g. We conclude there is so little corticosterone in egg albumen that it is not routinely usable as an indicator of chicken welfare. We have also found significant amounts of other steroids (progesterone, cortisol) in chicken egg white, which may have contributed to the levels reported in the antibody studies.
Abstract: Assessment of animal welfare can include analysis of physiological parameters, as well as behavior and health. Levels of adrenocortical hormones such as cortisol (and corticosterone in chickens) have been relied on as indicators of stress. Elevations in those hormones have been said to be correlated with poor welfare, while levels in the normal range have been interpreted to mean that animals are in a good state of welfare. Procuring blood samples from animals for hormone measures can in itself be stressful and cause increases in the target hormones. To overcome this problem, indirect measures of cortisol and corticosterone have been developed. In chickens, corticosterone levels in egg albumen are said to be a useful indirect measure, and have been used in several recent studies as indicators of chicken welfare. All of the measures of chicken egg albumen corticosterone in welfare studies have used immunoassays, and have reported values ranging from about 0.5 to over 20 ng/g. Using these measures, egg albumen from chickens housed in conventional cages or free ranging has been said to have indistinguishable corticosterone levels. This has been used to support the conclusion that chickens kept in conventional cages are not experiencing stress and are in a good state of welfare. In this study, we have used high-pressure liquid chromatography (HPLC) coupled with mass spectrometry (MS) to measure corticosterone in egg albumen. We found levels of corticosterone (median level about 50 pg/g) in egg albumen which were just above the limit of detection. By contrast, we found significant levels of progesterone and cortisol, hormones which would be expected to cross react with anti-corticosterone antibodies, and which therefore might explain the high reported levels of corticosterone using immunoassay. We conclude that because corticosterone levels in egg albumen are negligible, they cannot be used as an indicator of chicken welfare.
Animals 2020, 10, 821; doi:10.3390/ani10050821 www.mdpi.com/journal/animals Animals 2020, 10, 821 2 of 10
Keywords: animal welfare; chickens; layer hens; stress hormones; corticosterone; corticosteroids; mass spectrometry; high-pressure liquid chromatography
1. Introduction There is increasing public concern about close confinement of animals in intensive farming systems. One such system is the conventional cage for egg-laying chickens. Nicol and colleagues recently provided a useful description of conventional cages and free range systems for layer hens, and a summary of studies of factors affecting animal welfare in those systems [1]. It is accepted that assessment of animal welfare should involve a range of techniques, including behavioural, clinical (i.e., health) and physiological. Thus, there are said to be three ‘conceptual frameworks’ which can be applied: affective state, natural living and biological functioning [2]. An extensively used physiological measure (under the ‘biological functioning’ framework) is the assay of steroid effector hormones released upon activation of the hypothalamic–pituitary–adrenal (HPA) axis [3]. It is said that stressors cause elevation of adrenal corticosteroids (principally corticosterone in chickens [4]), and that because stress results in poor welfare, raised levels of corticosterone indicate a poor welfare state. The uncritical use of corticosteroid levels as an indicator of animal welfare has long been the subject of criticism [5], particularly for different egg-laying chicken housing systems. Different systems have been reported as being associated with increases, decreases or no difference in corticosterone levels [1,5]. The frequent use of corticosteroid measurements in animal welfare studies in recent decades has been facilitated by the availability of antibodies for the target molecules, and the resultant wide range of commercially available immunoassay kits. Most of the early studies of corticosteroids used plasma. However, it is apparent that the restraint of an animal to take a blood sample, and the act of taking the sample, will itself cause increases in the measured steroid. This has prompted efforts to develop non-invasive techniques for steroid measurement, including corticosterone, in chickens [6,7]. In particular, measures of faecal metabolites of corticosterone, using hplc and immunoassay, may have promise as an indirect measure [7]. One widely used measure in egg-laying chickens is the direct (i.e., without chromatographic separation) immunoassay of corticosterone in egg albumen [8]. Equivalent levels of egg albumen corticosterone in chickens that are housed free-range or in battery cages [9] have been used to support the idea that free-range hens are not necessarily less stressed than cage hens [10]. Further studies have applied egg corticosterone as an indicator of chicken welfare in studies of free-range chickens [11], ‘furnished cages’ [12] and floor space allowance and nest box access [13,14]. The use of direct immunoassays to measure steroids in human medicine, and in areas such as sports doping, has been heavily criticized, primarily because anti-steroid antibodies lack specificity, with significant and confounding cross-reactivity with related molecules. It is now accepted that mass spectrometry methods represent the ‘gold standard’ for steroid measurements [15]. Regarding measures in eggs, problems with the cross-reactivity of anti-corticosterone antibodies were highlighted in a study of bird egg yolk, where high-performance liquid chromatography (HPLC) separation, followed by immunoassay, indicated that the dominant immunoreactivity occurred at a peak corresponding to progesterone and similar molecules. There was very little corticosterone [16]. Consistent with this, a mass spectrometric analysis of chicken egg albumen indicated that corticosterone was barely detectable (at just under 50 pg/g) [17]. The aim of the present study was to develop a method for measuring corticosterone and similar steroid molecules in egg albumen using HPLC coupled with mass spectrometry.
2. Materials and Methods All solvents (methanol, acetonitrile, isopropanol, hexane and ethyl acetate) were of HPLC grade (B&J) and were obtained from ChemSupply. MS (mass spectrometry) grade formic acid (Fluka) was obtained from ChemSupply. Water was purified to MS standard by carbon adsorber and Animals 2020, 10, x 3 of 10 Animals 2020, 10, 821 3 of 10 were carried out in plasticware. Autosampler vials were polypropylene microvials (Shimadzu). Solid mixed-bedphase extraction ion exchange (SPE) was resins done (Sartorius). using C-18To 50 avoidmg columns problems (Discovery with steroids DSC-18) binding from Merck. to glass The [18 ], following compounds were obtained from Merck: corticosterone, cortisol, 11-deoxycortisol, 11- all procedures were carried out in plasticware. Autosampler vials were polypropylene microvials deoxycorticosterone and progesterone. Deuterated corticosterone (d8-corticosterone) for use as (Shimadzu). Solid phase extraction (SPE) was done using C-18 50 mg columns (Discovery DSC-18) from internal standard (2,2,4,6,6,17A,21,21-D8, 97%–98%) was obtained from Cambridge Isotope Merck. The following compounds were obtained from Merck: corticosterone, cortisol, 11-deoxycortisol, Laboratories. 11-deoxycorticosterone and progesterone. Deuterated corticosterone (d8-corticosterone) for use as internal Figure 1 shows some of the major synthetic and metabolic pathways for the steroids studied. standard (2,2,4,6,6,17A,21,21-D8, 97%–98%) was obtained from Cambridge Isotope Laboratories. Figure1 shows some of the major synthetic and metabolic pathways for the steroids studied.
17 alpha hydroxypregnenolone
progesterone
17 alpha hydroxyprogesterone
11-deoxycorticosterone
11-deoxycortisol corticosterone
cortisol
Figure 1. Major corticosteroid metabolic pathways (see [4] for a more detailed pathway). Large arrows show pathways to other metabolites. Figure 1. Major corticosteroid metabolic pathways (see [4] for a more detailed pathway). Large arrows Standardsshow pathways were preparedto other metabolites. as a 1 mg /mL solution in methanol. A mix of corticosterone, cortisol, 11-deoxy cortisol, 11-deoxy corticosterone, progesterone and d8-corticosterone was prepared in 50% Standards were prepared as a 1 mg/ml solution in methanol. A mix of corticosterone, cortisol, methanol/water in concentrations ranging from 0.02 to 4 ng/mL. 11-deoxy cortisol, 11-deoxy corticosterone, progesterone and d8-corticosterone was prepared in 50% Twelve conventional cage eggs from hens of the Isa Brown strain were kindly supplied by methanol/water in concentrations ranging from 0.02 to 4 ng/mL. Dr Tamsyn Crowley, School of Environmental and Rural Science, University of New England, Twelve conventional cage eggs from hens of the Isa Brown strain were kindly supplied by Dr Armidale,Tamsyn NewCrowley, South School Wales, of Environmental Australia. They and were Rural shipped Science, overnightUniversity toof New the laboratory England, Armidale, in Sydney, where albumen was separated and stored at 80 C until use. All eggs were collected within a few New South Wales, Australia. They were shipped− ◦ overnight to the laboratory in Sydney, where hoursalbumen of lay. was Mouse separated serum and was stored kindly at supplied−80 oC until by use. Dr FionaAll eggs Ryan, were University collected within of Technology, a few hours Sydney. of Albumen samples were thawed overnight at 4 C, then homogenized on ice using a Heidolph lay. Mouse serum was kindly supplied by Dr Fiona− Ryan,◦ University of Technology, Sydney. Diax 600 homogenizer, with four bursts of 30 s each. The criterion for adequate homogenization was that the samples could be easily pipetted with a 1 mL pipette tip. Samples of albumen (1 mL = 1 g) were pipetted into a plastic 15 mL Falcon tube, and 40 µL d8-corticosterone 0.125 µg/mL was added to give a final concentration of internal standard of 0.5 ng/g. Animals 2020, 10, x 4 of 10
Albumen samples were thawed overnight at −4 oC, then homogenized on ice using a Heidolph Diax 600 homogenizer, with four bursts of 30 seconds each. The criterion for adequate homogenization was that the samples could be easily pipetted with a 1 mL pipette tip. AnimalsSamples2020, 10 ,of 821 albumen (1 mL = 1 g) were pipetted into a plastic 15 mL Falcon tube, and 40 µL4 d8- of 10 corticosterone 0.125 µg/mL was added to give a final concentration of internal standard of 0.5 ng/g. The extraction method was based on that of De Baere et al. [15], incorporating liquid–liquid extraction withThe extractionprotein precipitation, method was baseddefatting on that with of De hexa Baerene etand al. [15further], incorporating purification liquid–liquid on SPE columns. extraction Acetonitrilewith protein (5 precipitation, mL) containing defatting 1% formic with hexaneacid wa ands added further to purificationeach spiked on 1mL SPE sample columns. and Acetonitrile vortexed for(5 30s. mL) The containing samples 1%were formic centrifuged acid was for 10 added minutes to each at 5000 spiked rpm at 1mL 4 oC. sample A total andof 2.5 vortexed mL supernatant for 30 s. wasThe taken, samples and were 2.5 centrifugedmL hexane forwas 10 added, min at 5000and the rpm tubes at 4 ◦ wereC. A totalvortexed of 2.5 for mL 15 supernatant seconds. The was upper taken, hexaneand 2.5 layer mL hexanewas discarded. was added, Acetonitrile and the tubeswas evaporated were vortexed with fora stream 15 s. Theof nitrogen upper hexane at 40 oC layer and wasthe residuediscarded. resuspended Acetonitrile in was100 µL evaporated methanol, with which a stream was vortexed of nitrogen for at 30 40 seconds,◦C and the and residue 900 µL resuspended water was µ µ added.in 100 SPEL methanol, columns were which conditioned was vortexed with for the 30 additi s, andon 900 of 1 LmL water methanol, was added. followed SPE by columns l mL water, were afterconditioned which the with sample the addition was added. of 1 After mL methanol, the samp followedle had passed by l mLthrough water, the after column, which the the column sample was was washedadded. with After 1 themL sample water, hadfollowed passed by through1 mL hexane the column,. Samples the were column eluted was wi washedth two 1 with mL volumes 1 mL water, of ethylfollowed acetate. by 1Elution mL hexane. of all solvents Samples was were by eluted gravity. with The two eluate 1 mL was volumes dried under of ethyl nitrogen acetate. gas Elution at 40oC of andall solventsresuspended was byin 125 gravity. µL methanol, The eluate vortexed was dried for under30 seconds, nitrogen followed gas at by 40 ◦theC andaddition resuspended of 125 µL in µ µ water.125 L The methanol, resulting vortexed sample for was 30 s,centrifuged followed by at the 10,000 addition rpm of(4 125oC) forL water. 10 minutes The resulting and 100 sample µL of the was µ supernatantcentrifuged atwas 10,000 added rpm to (4 autosampler◦C) for 10 min vials and for 100 HPLC-MS-MSL of the supernatant analysis. was The added autosampler to autosampler was maintainedvials for HPLC-MS-MS at 5oC. Studies analysis. using The mouse autosampler serum as was a positive maintained control at 5 ◦forC. Studiescorticosterone using mouse extracted serum a µ serumas a positive volume control of 100 forµL. corticosterone All other extraction extracted anda analysis serum volume steps were of 100 as forL. egg All otheralbumen. extraction A total and of µ 10µlanalysis volumes steps of were samples as for were egg injected albumen. onto A totalthe HPLC of 10 column.L volumes of samples were injected onto the HPLCThe column. HPLC–MS system was based on a Shimadzu LCMS 8060 triple quadrupole mass spectrometer,The HPLC–MS with systemtwo Nexera was based LC30AD on a Shimadzupumps, a LCMS CTO-20AC 8060 triple column quadrupole oven massand a spectrometer, SIL-30AC autosampler.with two Nexera The LC30ADHPLC column pumps, was a CTO-20AC an Agilent columnZorbax Eclipse oven and Plus a SIL-30AC C18 column, autosampler. having dimensions The HPLC column was an Agilent Zorbax Eclipse Plus C18 column, having dimensions 2.1 100 mm packed 2.1 x 100 mm packed with 1.8 µm particles, maintained in the column oven at 40 oC.× Mobile phase A µ waswith 0.1% 1.8 formicm particles, acid in maintainedwater, and mobile in the columnphase B ovenwas acetonitrile at 40 ◦C. Mobile plus 0.1% phase formic A was acid. 0.1% Flow formic rate wasacid 0.4 in mL/minute. water, and mobile Two gradients phase B waswere acetonitrile used (Figure plus 2). 0.1% formic acid. Flow rate was 0.4 mL/min. Two gradients were used (Figure2).
Figure 2. Chromatographic gradients used in the present study. Figure 2. chromatographic gradients used in the present study. Gradient 1 was used in preliminary experiments and was adequate for the separation of all the studiedGradient steroids, 1 was used apart in from preliminary corticosterone. experiments Better and separation was adequate of corticosterone for the separation from other of all peaks the studiedwith similar steroids, retention apart from times corticosterone. was achieved Better with Gradientseparation 2, of and corticosterone this was used from for other quantification peaks with of similarcorticosterone retention in times samples. was Because achieved of with the length Gradient of the 2, gradientsand this (15was min), usedretention for quantification times shifted of corticosteroneby about 0.2min in samples. during theBecause 8-h analysis of the length time, of which the gradients was necessary (15 minutes), to process retention all the times samples shifted and bystandards. about 0.2 Tominutes allow during for this, the standards 8-hour analysis at 0.1 ng time,/mL werewhich interposed was necessary at every to process 5th sample. all the Retention samples andtime standards. data from To those allow standards for this, were standards used to at adjust 0.1 ng/mL the retention were interposed times for corticosterone at every 5th sample. peaks in Retentionthe samples. time data from those standards were used to adjust the retention times for corticosterone peaksTable in the1 samples.shows the MS parameters for precursor ions and fragment ions, together with observed retention times, for the steroidsused in the present study. Animals 2020, 10, x 5 of 10
Table 1 shows the MS parameters for precursor ions and fragment ions, together with observed retentionAnimals times,2020, 10 for, 821 the steroids used in the present study. 5 of 10
Table 1. Multiple reaction monitoring (MRM) transitions and retention times for the studied steroids. Table 1. Multiple reaction monitoring (MRM) transitions and retention times for the studied steroids. Precursor Quantifier Qualifier Retention Time Steroid Precursor Quantifier Qualifier Retention Time Steroid Ion Ion Ion (min) Ion Ion Ion (min) corticosterone 347.2 121.1 329.3 8.60 5.011 1 d8 corticosteronecorticosterone 355.2 347.2 125.1 121.1 337.3 329.3 8.60 5.01 8.49 d8 corticosterone 355.2 125.1 337.3 8.49 cortisol 363.2 121.1 91.1 6.54 cortisol 363.2 121.1 91.1 6.54 11-deoxy cc11-deoxy cc 331.1 331.1 97.1 97.1 331.1 331.1 11.40 11.40 11-deoxycortisol11-deoxycortisol 347.1 347.1 97.1 97.1 109.1 109.1 9.08 9.08 progesteroneprogesterone 315.1 315.1 97.1 97.1 109.1 109.1 12.10 12.10 1 Retention1 Retention times times are are for for gradient gradient 1,1,except except for for the the italicized italicized time fortime corticosterone, for corticosterone, which was which obtained was with obtainedgradient with 2. 11-deoxygradient cc2. is11-deo 11-deoxycorticosterone.xy cc is 11-deoxycorticosterone. All other parameter settings for the mass spectrometer were as previously described [19]. Dwell All other parameter settings for the mass spectrometer were as previously described [19]. time for transitions was 50ms. Data were acquired and analysed with Shimadzu LabSolutions v5.91 Dwell time for transitions was 50 ms. Data were acquired and analysed with Shimadzu LabSolutions software. The values reported for each steroid are integrated peak counts for the quantifier ion and v5.91 software. The values reported for each steroid are integrated peak counts for the quantifier ion are shown in Table 1. GraphPad Prism 8.3.1 was used for statistical analysis and graphing. and are shown in Table1. GraphPad Prism 8.3.1 was used for statistical analysis and graphing.
3. Results3. Results TracesTraces obtained obtained for forthe thecorticosterone corticosterone standard standard curve curve and anda plot a plot of those of those traces traces are areshown shown in in FigureFigure 3. The3. The limit limit of quantification of quantification was was determined determined to be to 0.02 be 0.02 ng/mL ng / mL(Figure (Figure 3a inset).3a inset).
Figure 3. (a) Traces for increasing concentrations of corticosterone ranging from 0.02 to 4 ng/mL, using gradient 2. The major peak represents the quantifier ion at m/z 121.1. The inset in (a) shows the Figureresponse 3. (a) Traces to 0.02 for ng increasing/mL corticosterone concentrations on a compressedof corticosterone time ranging scale. (b from) A standard 0.02 to 4ng/ml, line plotting using the gradientintegrated 2. The peak major areas peak (filled represents circles) fromthe quanti (a). Rfier2 for ion the at fitted m/z line 121.1. was The 0.994. inset in (a) shows the response to 0.02ng/ml corticosterone on a compressed time scale. (b) A standard line plotting the integratedFigure peak4 shows areas overlaid (filled circles) MS traces from for(a). albumenR2 for the fitted extracts, line withwas 0.994. corticosterone peaks (filled arrows), and other peaks at around the retention time for corticosterone, obtained with Gradient 1 (a) and Gradient 2 (b), using albumen samples from 12 cage eggs. Animals 2020, 10, x 6 of 10
Figure 4 shows overlaid MS traces for albumen extracts, with corticosterone peaks (filled Animals 2020, 10, x 6 of 10 arrows), and other peaks at around the retention time for corticosterone, obtained with Gradient 1 (a) and FigureGradient 4 2shows (b), using overlaid albumen MS samplestraces for from albume 12 cagen extracts, eggs. with corticosterone peaks (filled arrows), and other peaks at around the retention time for corticosterone, obtained with Gradient 1 Animals 2020, 10, 821 6 of 10 (a) and Gradient 2 (b), using albumen samples from 12 cage eggs.
Figure 4. Peaks from MS traces using Gradient 1 (a) and Gradient 2 (b) for albumen extracts (12) from cageFigure eggs 4.showingPeaks fromthe corticosterone MS traces using peak Gradient (filled arrows) 1 (a) and and Gradientother peaks 2 ( bat) forsimilar albumen retention extracts times, (12) includingfromFigure cage 4.the Peaks eggs peaks showingfrom for 11-deoxyMS thetraces corticosterone cortisol using Gradient (open peak arrows 1 (a) (filled and). Major arrows)Gradient tick andmarks 2 (b) other for repr albumen peaksesent at0.5 extracts similar minutes. retention(12) Bars from representtimes,cage eggs including 1000 showing counts. the the peaks corticosterone for 11-deoxy peak cortisol (filled (open arrows) arrows). and other Major peaks tick at marks similar represent retention 0.5 times, min. Barsincluding represent the peaks 1000 counts.for 11-deoxy cortisol (open arrows). Major tick marks represent 0.5 minutes. Bars Whilerepresent Gradient 1000 counts.1 clearly separated peaks for corticosterone and the isobaric molecule 11-deoxy cortisol,While it was Gradient apparent 1 clearlythere was separated at least peaks one forearlier corticosterone peak at almost and thethe isobaricsame retention molecule time 11-deoxy for corticosterone,cortisol,While it wasGradient which apparent would 1 clearly there interfere separated was atwith least peaks the one inte for earlier grationcorticosterone peak of peaks at almostand at thethe the isobaric apparent same molecule retention corticosterone 11-deoxy time for retentioncorticosterone,cortisol, time. it was Samples which apparent would run there with interfere wasGradient withat least the 2 integrationshowedone earlier clearer of peak peaks corticosterone at at almost the apparent the peaks same corticosterone (Figureretention 4b), retentiontime and for Gradienttime.corticosterone, Samples 2 was used run which with for wouldmeasurement Gradient interfere 2 showed of withcorticosterone clearer the inte corticosteronegration peaks. of It peaks peaksis important (Figureat the 4 apparenttob) note, and that, Gradient corticosterone with 2the was exceptionusedretention for of measurement time. one reading, Samples of even run corticosterone usingwith GradientGradient peaks. 22, showed al Itmost is important allclearer corticosterone tocorticosterone note that, peaks with peaks were the exceptionjust(Figure above 4b), of the oneand levelreading,Gradient of detection. even 2 was using used Gradient for measurement 2, almost allof corticosterone peakspeaks. were It is important just above theto note level that, of detection. with the exceptionGradientGradient of 1 one gave 1 gave reading, good good separation even separation using of Gradientofall all the the other other2, al studiedmost studied all steroids corticosterone steroids (Figure (Figure peaks5)5 and) and were was was usedjust used above for for the the the quantificationquantificationlevel of detection. of ofsteroids steroids other other than than corticosterone corticosterone in in samples samples (Figure (Figure 6).6). Gradient 1 gave good separation of all the other studied steroids (Figure 5) and was used for the quantification of steroids other than corticosterone in samples (Figure 6).
Figure 5. MS trace from a mix of standards of the studied steroids (1 ng/mL) showing peaks for cortisol, d8-corticosterone (d8), corticosterone (cc), 11-deoxycortisol (11-dc), 11-deoxycorticosterone (11-dcc) and progesterone, using Gradient 1. Traces from both quantifier (the larger trace for each peak) and Figure 5. MS trace from a mix of standards of the studied steroids (1ng/ml) showing peaks for cortisol, d8-corticosteronequalifier ions are (d8), shown. corticosterone (cc), 11-deoxycortisol (11-dc), 11-deoxycorticosterone (11-dcc) andFigure progesterone, 5. MS trace using from Gradient a mix of 1.standards Traces from of the both studie quantifierd steroids (the (1ng/ml) larger traceshowing for eachpeaks peak) for cortisol, and qualifierd8-corticosterone ions are shown. (d8), corticosterone (cc), 11-deoxycortisol (11-dc), 11-deoxycorticosterone (11-dcc) and progesterone, using Gradient 1. Traces from both quantifier (the larger trace for each peak) and qualifier ions are shown. Animals 2020, 10, x 7 of 10
Animals 2020, 10, x 7 of 10 Animals 2020, 10, 821 7 of 10
Figure 6. Illustrative MS trace for an albumen extract separated using Gradient 1. Major peaks are indicated,Figure 6. Illustrativeand traces from MS trace both forquantifier an albumen (larger extract trace for separated each peak) using and Gradient qualifier 1. ions Major are peaksshown. are indicated,Figure 6. and Illustrative traces from MS both trace quantifier for an albumen (larger traceextract for separated each peak) using and qualifierGradient ions 1. Major are shown. peaks are Standardindicated, lines and tracesobtained from forboth steroids quantifier other (larger than trace corticosterone for each peak) and were qualifier linear ions in arethe shown. range of Standard lines obtained for steroids other than corticosterone were linear in the range of concentrations studied, and were fitted with a straight line. R2 values for all regressions were greater concentrationsStandard studied, lines obtained and were for fittedsteroids with other a straight than corticosterone line. R2 values were for linear all regressions in the range were of than 0.998 (data not shown). greaterconcentrations than 0.998 studied, (data not and shown). were fitted with a straight line. R2 values for all regressions were greater Mean recovery, calculated from levels of d8 corticosterone in samples, was 48.1%. The 95% thanMean 0.998 recovery, (data not calculated shown). from levels of d8 corticosterone in samples, was 48.1%. The 95% confidence interval for recoveries was from 44% to 52.2%. Figure 7 shows individual datapoints for confidenceMean interval recovery, for recoveriescalculated wasfrom from levels 44% of tod8 52.2%. corticosterone Figure7 shows in samples, individual was datapoints48.1%. The for 95% corticosterone and the other studied steroids. In our opinion, the scatter of datapoints was corticosteroneconfidence interval and the for other recoveries studied steroids.was from In 44% our to opinion, 52.2%. the Figure scatter 7 shows of datapoints individual was datapoints inconsistent for inconsistent with a normal distribution. It seemed appropriate, therefore, to use the median of all withcorticosterone a normal distribution. and the other It seemed studied appropriate, steroids. therefore, In our toopinion, use the medianthe scatter of all of data datapoints to represent was data to represent central tendency, and to calculate P values for comparison of cage and free-range centralinconsistent tendency, with and a tonormal calculate distribup valuestion. forIt seemed comparison appropriate, of cage andtherefore, free-range to use albumen the median samples of all albumen samples using a non-parametric test (Mann–Whitney). usingdata a to non-parametric represent central test tendency, (Mann–Whitney). and to calculate P values for comparison of cage and free-range albumen samples using a non-parametric test (Mann–Whitney). (a) (b) (a) (b)
Figure 7. Steroid levels in egg albumen. Values are shown in (a) for corticosterone and in (b) for 11-deoxy cortisol (11.dc), cortisol, progesterone and 11-deoxy corticosterone (11-dcc). Figure 7. Steroid levels in egg albumen. Values are shown in (a) for corticosterone and in (b) for 11- deoxySteroid cortisol concentrations (11.dc), cortisol, in egg progeste albumenrone samples and 11-deoxy are shown corticosterone in Table (11-dcc).2. Corticosterone levels were Figure 7. Steroid levels in egg albumen. Values are shown in (a) for corticosterone and in (b) for 11- very low, as was 11-deoxycorticosterone. There were significant quantities of 11-deoxycortisol, cortisol deoxy cortisol (11.dc), cortisol, progesterone and 11-deoxy corticosterone (11-dcc). and progesterone. Mean (nTable= 6) 2. corticosterone Median steroid levels concentrations in mouse in serum egg albumen. were 71 ng/mL, ranging from 49 to 80 ng/mL, which is similarTable to earlier 2. Median reported steroid values concentrations using HPLC in egg and albumen. MS [20 ]. Animals 2020, 10, 821 8 of 10
Table 2. Median steroid concentrations in egg albumen.
Steroid 1 Concentration (ng/g) corticosterone 0.043 (0.029–0.091) 11-deoxycortisol 0.335 (0.27–0.79) cortisol 0.374 (0.24–1.60) 11-deoxycorticosterone 0.057 (0.033–0.08) progesterone 1.251 (0.81–0.84) 1 Values are from 12 eggs. Values in brackets are 95% confidence limits of the median; corticosterone values are corrected for recovery.
4. Discussion It is a problem besetting animal welfare science that the measurement of ‘stress hormones’ is ‘dangerously easy’ [21]. This issue has been acknowledged and addressed in human clinical chemistry [15], but has yet to have any impact on the routine measurement of cortisol and corticosterone in animal welfare studies. The present study, using HPLC and mass spectrometry, has demonstrated that corticosterone levels in chicken egg albumen are just above the threshold of detectability. Median values of corticosterone were less than 50 pg/g. These values are consistent with those found in commercially bought eggs in an earlier HPLC–mass spectrometry study [17]. In albumen samples, we found unidentified peaks with indistinguishable precursor and fragment masses, and similar retention times to corticosterone. These peaks were better separated from the corticosterone peak by using an extended chromatographic gradient. We also found that chicken egg albumen contains significant quantities of progesterone (about 1.2 ng/g) and cortisol (about 0.4 ng/g), with detectable, but lower, levels of 11-deoxycortisol and 11-deoxycorticosterone. These data show that earlier studies of egg albumen using immunoassay probably greatly overestimated corticosterone content, by up to 20-fold, given the cross-reactivity of anti-corticosterone antibodies for closely related steroids such as cortisol and progesterone [16,22]. This will have been compounded by the presence of other unidentified molecules, which may have similar structures to corticosterone and which may interact with the antibodies used. Our findings go some way to explaining the greater than 50-fold range of egg albumen corticosterone levels reported using immunoassays, even in separate studies from the same laboratory, which employed both radioimmunoassay and enzyme-linked immunosorbent assay [13,14]. The main conclusion from this work is that immunoassays of corticosterone in chicken egg albumen cannot be used as measures of chicken welfare. The bulk of what is measured is probably not corticosterone. This is perhaps not surprising, given that corticosterone must move from the plasma to albumen during the few hours in which it is formed [23], and is present in chicken plasma at relatively low levels (about 0.5 ng/mL: [24]). Similarities between egg corticosterone levels in chickens housed in different housing systems, particularly conventional cages and free ranging, continue to be used to support the contention that chickens housed in cages have good welfare [10]. Given that it is likely that corticosterone is not what is being measured, such conclusions are arguably questionable. Our finding that there is almost undetectable corticosterone in chicken egg albumen adds to the substantial list of criticisms of the assay of ‘stress hormones’ in animal welfare, even where those measures do not involve blood sampling. Thus, adrenocortical hormones undergo significant natural diurnal variation, with pulsatile rises and falls superimposed on that rhythm [25]. Given this, measures at one timepoint are unlikely to reflect relevant levels. Moreover, given the substantial variations in levels observed, there is the obvious question as to the significance of those variations [5]. They are unlikely to represent diurnal variations in welfare. Exposure to sustained stress results in a down-regulation of the adrenocortical response, making measures of these hormones unsuitable for anything other than acute stress situations [3]. This has particular importance where the question is the Animals 2020, 10, 821 9 of 10 effect of long-term housing on welfare. A further serious criticism of reliance on cortisol or corticosterone measures is that many situations which are demonstrably stressful (and associated with marked changes in other welfare measures) are not associated with increases in those hormones [5,26–28]. Furthermore, it is clear that corticosteroids can be elevated in response to arousal, as well as stress [1], requiring the ‘valence’ of the response to be defined before anything can be concluded about stress or animal welfare. The conclusion is that, unless corticosterone (or other ‘stress hormone’ measures) can be validated and verified as measures of welfare, they should be discounted in favour of other measures, including behavioural, ethological and health measures, all of which have been applied extensively in the area of chicken welfare [29]. This is particularly the case where the assessment of welfare is relied on as part of the legislative process relating to the implementation of animal welfare law [30].
Author Contributions: Conceptualization, M.P.C.; methodology, M.P.C. and M.P.P.; formal analysis, M.P.C.; investigation, M.P.C.; resources, M.P.P.; data curation, M.P.C.; writing—original draft preparation, M.P.C.; writing—review and editing, M.P.P.; supervision, M.P.P.; project administration, M.P.P. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Conflicts of Interest: The authors declare no conflict of interest.
References
1. Nicol, C.J.; Bouwsema, J.; Caplen, G.; Davies, A.C.; Hockenhull, J.; Lambton, S.L.; Lines, J.A.; Mullan, S.; Weeks,C.A. Farmed Bird Welfare Science Review. Availableonline: http://agriculture.vic.gov.au/__data/assets/ pdf_file/0019/370126/Farmed-Bird-Welfare-Science-Review-Oct-2017.pdf (accessed on 14 February 2020). 2. Hemsworth, P.H.; Mellor, D.J.; Cronin, G.M.; Tilbrook, A.J. Scientific assessment of animal welfare. N. Z. Vet. J. 2015, 63, 24–30. [CrossRef][PubMed] 3. Spencer, R.L.; Deak, T. A users guide to HPA axis research. Physiol. Behav. 2017, 178, 43–65. [CrossRef] [PubMed] 4. Carsia, R.V. Adrenals. In Sturkie’s Avian Physiology, 6th ed.; Scanes, C.G., Ed.; Academic Press: Cambridge, MA, USA, 2014; pp. 577–611. 5. Rushen, J.A. Problems associated with the interpretation of physiological data in the assessment of animal welfare. Appl. Anim. Behav. Sci. 1991, 28, 381–386. [CrossRef] 6. Scanes, C.G. Biology of stress in poultry with emphasis on glucocorticoids and the heterophil to lymphocyte ratio. Poult. Sci. 2016, 95, 2208–2215. [CrossRef][PubMed] 7. Palme, R. Non-invasive measurement of glucocorticoids: Advances and problems. Physiol. Behav. 2019, 199, 229–243. [CrossRef][PubMed] 8. Downing, J.A.; Bryden, W.L. Determination of corticosterone concentrations in egg albumen: A non-invasive indicator of stress in laying hens. Physiol. Behav. 2008, 95, 381–387. [CrossRef][PubMed] 9. Downing, J. Non-Invasive Assessment of Stress in Commercial Housing Systems. A Report for the Australian Egg Corporation Limited. 2012. Available online: https://www.australianeggs.org.au/dmsdocument/529- non-invasive-assessment-of-stress-in-commercial-housing-systems (accessed on 14 February 2020). 10. University of Sydney News. Cage or Free Range Don’t Put Too Much Stress on Housing Systems. Available online: https://www.sydney.edu.au/news/84.html?newsstoryid=8876 (accessed on 25 March 2020). 11. Campbell, D.L.M.; Hinch, G.N.; Downing, J.A.; Lee, C. Early enrichment in free-range laying hens: Effects on ranging behavior, welfare and response to stressors. Animal 2017, 12, 575–584. [CrossRef][PubMed] 12. Barnett, J.L.; Tauson, R.; Downing, J.A.; Janardhana, V.; Lowenthal, J.W.; Butler, K.L.; Cronin, G.M. The effects of a perch, dust bath, and nest box, either alone or in combination as used in furnished cages, on the welfare of laying hens. Poult. Sci. 2009, 88, 456–470. [CrossRef][PubMed] 13. Engel, J.M. The Effects of Floor Space Allowance and Nest Box Access on the Welfare of Caged Laying Hens (Gallus gallus domesticus). Ph.D. Thesis, University of Melbourne, Melbourne, Australia, 2016. 14. Engel, J.M.; Widowski, T.M.; Tilbrook, A.J.; Butler, K.L.; Hemsworth, P.H. The effects of floor space and nest box access on the physiology and behavior of caged laying hens. Poult. Sci. 2019, 98, 533–547. [CrossRef] [PubMed] Animals 2020, 10, 821 10 of 10
15. Handelsman, D.J. Mass spectrometry, immunoassay and valid steroid measurements in reproductive medicine and science. Hum. Reprod. 2017, 32, 1147–1150. [CrossRef][PubMed] 16. Quillfeldt, P.; Poisbleau, M.; Parenteau, C.; Trouve, C.; Demongin, L.; Van Noordwijk, H.; Mostl, E. Measuring corticosterone in seabird egg yolk and the presence of high yolk gestagen concentrations. Gen. Comp. Endocr. 2011, 173, 11–14. [CrossRef][PubMed] 17. De Baere, S.; Rosendahl Larsen, T.; Devreese, M.; De Backer, P.; De Neve, L.; Fairhurst, G.; Lens, L.; Croubels, S. Use of LC-MS-MS as an alternative to currently available immunoassay methods to quantitate corticosterone in egg yolk and albumen. Anal. Bioanal. Chem. 2015, 407, 4351–4362. [CrossRef][PubMed] 18. Makin, H.L.J.; Honour, J.W.; Shackleton, C.H.L.; Griffiths, W.J. General methods for the extraction, purification, and measurement of steroids by chromatography and mass spectrometry. In Steroid Analysis; Makin, H.L.J., Gower, D.B., Eds.; Springer Science+Business Media: London, UK, 2010. [CrossRef] 19. Karvaly, G.; Kovacs, K.; Meszaros, K.; Kocsis, I.; Patocs, A.; Vasarhelyi, B. The comprehensive characterization of adrenocortical steroidogenesis using two-dimensional ultra-performance liquid chromatography—electrospray ionization tandem mass spectrometry. J. Pharm. Biomed. 2018, 153, 274–283. [CrossRef][PubMed] 20. Li, H.; Sheng, L.-P.; Wang, B.; Yang, Z.-L.; Liu, S.-Y. An optimized method for corticosterone analysis in mouse plasma by ultra-performance liquid chromatography-full-scan high-resolution accurate mass spectrometry. J. Chromatogr. Sci. 2015, 53, 285–294. [CrossRef][PubMed] 21. Webster, J.A. Animal Welfare: Limping Towards Eden; Universities Federation for Animal Welfare: Wallingford, UK, 2005; p. 30. 22. Aerts, J. Quantification of a glucocorticoid profile in non-pooled samples is pivotal in stress research across vertebrates. Front. Endocrinol. 2018, 9, 635. [CrossRef][PubMed] 23. Johnson, A.L. Reproduction in the female. In Sturkie’s Avian Physiology, 6th ed.; Scanes, C.G., Ed.; Academic Press: Cambridge, MA, USA, 2014; pp. 635–665. 24. Fallahsharoudi, A.; De Kock, N.; Johnsson, M.; Ubhayasekera, S.J.K.A.; Bergquist, J.; Wright, D.; Jensen, P. Domestication effects on stress induced steroid secretion and adrenal gene expression in chickens. Sci. Rep. 2015, 5, 15345. [CrossRef][PubMed] 25. Sarabdjitsingh, R.A.; Joels, M.; De Kloet, E.R. Glucocorticoid pulsatility and rapid corticosteroid actions in the central stress response. Physiol. Behav. 2012, 106, 73–80. [CrossRef][PubMed] 26. McFarlane, J.M.; Curtis, S.E.; Shanks, R.D.; Carmer, G.G. Multiple concurrent stressors in chick. 1. Effects on weight gain, feed intake and behavior. Poult. Sci. 1989, 68, 501–509. [CrossRef][PubMed] 27. McFarlane, J.M.; Curtis, S.E.; Simon, J.; Izquierdo, O.A. Multiple concurrent stressors in chicks. 2. Effects on haematologic, body composition and pathologic traits. Poult. Sci. 1989, 68, 510–521. [CrossRef][PubMed] 28. McFarlane, J.M.; Curtis, S.E. Multiple concurrent stressors in chicks. 3. Effects on plasma corticosterone and the heterophil/lymphocyte ratio. Poult. Sci. 1989, 68, 522–527. [CrossRef][PubMed] 29. Nicol, C.J. The Behavioural Biology of Chickens; CABI: Wallingford, UK, 2015. 30. Animal Health Australia. Proposed Australian Animal Welfare Standards and Guidelines. Poultry. Consultation Regulatory Impact Statement. 2017. Available online: http://www.animalwelfarestandards.net. au/files/2015/07/Public-Cons-Version-Poultry-RIS-Nov-2017.pdf (accessed on 14 February 2020).
© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).