The application of Reporter Gene Assays for the detection ofendocrine disruptors in sport supplements.

Plotan, M., Scippo, M-L., Muller, M., Antignac, J-P., Malone, E., Bovee, T., Mitchell, S., Elliott, C., & Connolly, L.(2011). The application of Reporter Gene Assays for the detection of endocrine disruptors in sport supplements.Analytica Chimica Acta, 700(1-2), 34-40. https://doi.org/10.1016/j.aca.2010.12.014

Published in:Analytica Chimica Acta

Queen's University Belfast - Research Portal:Link to publication record in Queen's University Belfast Research Portal

General rightsCopyright for the publications made accessible via the Queen's University Belfast Research Portal is retained by the author(s) and / or othercopyright owners and it is a condition of accessing these publications that users recognise and abide by the legal requirements associatedwith these rights.

Take down policyThe Research Portal is Queen's institutional repository that provides access to Queen's research output. Every effort has been made toensure that content in the Research Portal does not infringe any person's rights, or applicable UK laws. If you discover content in theResearch Portal that you believe breaches copyright or violates any law, please contact openaccess@qub.ac.uk.

Download date:05. Oct. 2020

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Analytica Chimica Acta 700 (2011) 34– 40

Contents lists available at ScienceDirect

Analytica Chimica Acta

jou rn al hom epa ge: www.elsev ier .com/ locate /aca

The application of reporter gene assays for the detection of endocrine disruptorsin sport supplements

Monika Plotana, Christopher T. Elliotta, Marie Louise Scippob, Marc Mullerc,Jean-Philippe Antignacd, Edward Malonee, Toine F.H. Boveef, Samuel Mitchell g, Lisa Connollya,∗

a Institute of Agri-Food and Land Use, School of Biological Sciences, Queen’s University Belfast, Belfast BT95AG, Northern Ireland, United Kingdomb Department of Food Sciences, University of Liege, 4000 Liege, Belgiumc Molecular Biology and Genetic Engineering GIGA-R, University of Liege, 4000 Liege, Belgiumd LABERCA, ENVN, USC INRA 2013, BP 50707, 44 307, Nantes, Francee The State Laboratory, Young’s Cross, Celbridge, Co. Kildare, Irelandf RIKILT Institute of Food Safety, P.O. Box 230, AE Wageningen 6700, The Netherlandsg Agri-Food and Biosciences Institute, Belfast BT9 5PX, United Kingdom

a r t i c l e i n f o

Article history:Received 2 September 2010Received in revised form 6 December 2010Accepted 8 December 2010Available online 16 December 2010

Keywords:Endocrine disruptorEstrogenAndrogenBioassayHealth food supplementFood safety

a b s t r a c t

The increasing availability and use of sports supplements is of concern as highlighted by a number ofstudies reporting endocrine disruptor contamination in such products. The health food supplement mar-ket, including sport supplements, is growing across the Developed World. Therefore, the need to ensurethe quality and safety of sport supplements for the consumer is essential.

The development and validation of two reporter gene assays coupled with solid phase sample prepara-tion enabling the detection of estrogenic and androgenic constituents in sport supplements is reported.Both assays were shown to be of high sensitivity with the estrogen and androgen reporter gene assayshaving an EC50 of 0.01 ng mL−1 and 0.16 ng mL−1 respectively.

The developed assays were applied in a survey of 63 sport supplements samples obtained across theIsland of Ireland with an additional seven reference samples previously investigated using LC–MS/MS.Androgen and estrogen bio-activity was found in 71% of the investigated samples. Bio-activity profil-ing was further broken down into agonists, partial agonists and antagonists. Supplements (13) withthe strongest estrogenic bio-activity were chosen for further investigation. LC–MS/MS analysis of thesesamples determined the presence of phytoestrogens in seven of them. Supplements (38) with androgenbio-activity were also selected for further investigation. Androgen agonist bio-activity was detected in12 supplements, antagonistic bio-activity was detected in 16 and partial antagonistic bio-activity wasdetected in 10. A further group of supplements (7) did not present androgenic bio-activity when testedalone but enhanced the androgenic agonist bio-activity of dihydrotestosterone when combined.

The developed assays offer advantages in detection of known, unknown and low-level mixtures ofendocrine disruptors over existing analytical screening techniques. For the detection and identification ofconstituent hormonally active compounds the combination of biological and physio-chemical techniquesis optimal.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

The endocrine system controls important physiological eventswithin the body through hormone signalling and cellular receptors.Estrogens, the female sex hormones, are responsible for growth,development and function of the female reproductive system [1,2].Androgens, the male sex hormones, regulate spermatogenesis, areresponsible for male sex characteristics, influence behaviour andred blood cell production [3].

∗ Corresponding author. Tel.: +44 28 9097 6668; fax: +44 28 909706513.E-mail address: l.connolly@qub.ac.uk (L. Connolly).

An endocrine disruptor (ED) is defined by the European Com-mission (1996) [4] as “an exogenous substance or a mixture, thatalters function(s) of the endocrine system and consequently causesadverse health effects in an intact organism, or its progeny, or (sub)populations”. EDs may function through various modes of actionsuch as mimicking or antagonising the effects of hormones, disrupt-ing natural hormone production pathways, disrupting hormonesecretion or metabolism, or disrupting expression of their nuclearreceptors [5]. Disruption of the endocrine system can lead to serioushealth effects. Estrogenic EDs may result in reproductive disordersand lead to the development of breast and endometrial cancer inwomen, or testis cancer and infertility in men [5,6]. AndrogenicEDs may be responsible for many other serious health problems;

0003-2670/$ – see front matter © 2010 Elsevier B.V. All rights reserved.doi:10.1016/j.aca.2010.12.014

Author's personal copy

M. Plotan et al. / Analytica Chimica Acta 700 (2011) 34– 40 35

including prostate carcinoma, reproduction system abnormalities,hepatotoxicity, heart disease and psychiatric symptoms [7–9].

Exposure to EDs via our diet is a major route of concern [10].Recently, there has been a significant rise in the use of healthfood products in the Developed World, including dietary supple-ments, by the general public. The Island of Ireland is typical of thistrend. Food supplements as defined by the European CommunitiesRegulation, 2007 [11]; are products containing nutrients or othersubstances that elicit nutritional and physiological effects. Thereferred Regulation specifies nutrients as vitamins and mineralsand provides an exact list of allowed compounds for food sup-plements production. These health food products including sportsupplements can be obtained via a widening range of outlets suchas the internet, pharmacies, supermarkets, health food shops andoutlets linked to sport centres. Many consumers view such supple-ments as a means of improving their health status. They believethat these products positively influence overall health, enhancesports performance and in some cases result in muscle gain withincreased strength. Similar effects can be obtained through anabolicandrogenic steroid action. However, both exo- and endogenousanabolic steroid compounds are banned in any type of supplemen-tation [11]. The ‘General Food Law Regulation’ clearly specifies thatmedicinal products should not be included in any type of food.However, concerns over the accidental, environmental or deliber-ate contamination of sport supplements during the manufacturingprocess or from the raw ingredients have been highlighted in anincreasing number of studies reporting the contamination of sportsupplements with hormonally active substances [12,13]. Delbekeet al. reported the significant presence of pro-hormones and hor-mones in two sport supplements which did not have hormonalsubstances listed on the product label [14]. Consequently, inges-tion of doses much lower then the manufacturers’ recommendeddose resulted in positive urine doping test results in athletes forup to 144 h. Anabolic androgenic contamination of food supple-ments was reported in another study by Kafrouni et al. [15]. In thiscase two young men had been taking contaminated supplementsfor 4 weeks and had developed serious cholestatic liver injury. Themen did not require a liver transplant but recovery by hospitalisa-tion took 8–15 weeks. Therefore, it is important that an effectiveapproach for monitoring such foods for anabolic compounds isdeveloped.

The most reliable way of detecting hormonally active con-stituents is by monitoring biological activity through their naturaltarget molecules, the steroid hormone receptors. Reporter GeneAssays (RGAs), incorporating relevant receptors and a reportergene such as luciferase have previously been used to detecthormonally active compounds through biological activity [16].This type of bioassay enables detection of known, unknown andlow-level cocktails of EDs and offers advantages over traditionalassays which will only detect a limited range of known com-pounds due to recognition based on structure rather than biologicalactivity.

The present study focused on two groups of steroid hormones,estrogens and androgens due to the importance of their physi-ological effects. The overall aim was to develop estrogenic andandrogenic RGA screening assays for the detection of EDs in sportsupplements available to consumers on the Island of Ireland marketand apply these methods in an all Island survey.

2. Experimental

2.1. Reagents and chemicals

Cell culture reagents including Dulbecco’s Modified EagleMedium (DMEM) medium, general and hormone depleted serum,

penicillin–streptomycin, trypsin, phosphate-buffered saline (PBS)and lysis reagent were supplied by Invitrogen Ltd (Paisley, UK). Ref-erence standards including: 17�-estradiol, 17�-estradiol, estrone,dihydrotestosterone (DHT), testosterone, dehydroepiandrosterone(DHEA), bisphenol A, trenbolone, fulvestrant (ICI 182,780), flu-tamide, dimethyl sulfoxide (DMSO) and thiazolyl blue tetrazoliumbromide were supplied by Sigma (Poole, Dorset, UK). Falcon tis-sue culture flasks were obtained from BD Biosciences (Oxford, UK).Specialised plates used in the reporter gene assays were suppliedby Greiner Bio-One (Stonehouse, UK). A luciferase kit (PromegaE1500) was supplied by MSc Ltd, Ireland. Dispersive SPE CitrateExtraction Tubes and PSA/ENVI-Carb SPE cleanup Tubes 2 as wellas HPLC grade water, tert-Butyl-methyl-ether, glacial acetic acidand sodium acetate were supplied by Sigma (Poole, Dorset, UK).Oasis HLB glass and plastic cartridges (5 cc/200 mg and 6 cc/200 mg,respectively) were supplied by Waters Chromatography Ireland(Dublin, Ireland). Methanol and acetone were obtained by BDH(Poole, Dorset, UK). Acetonitrile was supplied by Analab (Lisburn,UK).

2.2. Reporter gene assay

2.2.1. Estrogen reporter gene assayAn estrogen responsive, reporter gene cell line (MMV-Luc) was

previously produced by stable transfection of MCF-7 cell line withthe MMTV-Luc reporter plasmid [17]. Prior to running the assay, thecells were passaged at least twice in hormone free assay medium(DMEM, 10% hormone depleted serum) to remove endogenous hor-mones. Additionally, the assay was performed using phenol redfree DMEM media due to the estrogenicity of phenol red. A stan-dard curve was generated using 17�-estradiol. Cells were seeded inhormone free assay media at a concentration of 4 × 105 cells mL−1

in 96 well plates and incubated overnight at 37 ◦C in 8% CO2.The next day, assay media spiked with various concentrations ofstandard was added to the cells and incubated at 37 ◦C for 24 h.The supernatant was discarded and cells washed with PBS (pH7) prior to lysis by the addition of lysis buffer and shaking for10 min at 37 ◦C. Luciferase substrate was injected and the assayread using a luminometer (Mithras LB 940, Berthold Technologies).The assay was developed for agonistic and antagonistic activitydetection.

2.2.2. Androgen reporter gene assayAs with the estrogen reporter gene assay, androgen responsive

cell line (TARM-Luc) was previously produced by stable trans-fection of T47D cell line with the MMTV-Luc reporter plasmid[17]. However, the MMV-Luc cell line intrinsically expresses theestrogen receptor while in the case of the TARM-Luc cell line, thepSV-AR0 expression vector, encoding the androgen receptor gene,was additionally transfected into the T47D cell line due to theabsence of the expression of substantial amounts of the androgenreceptor. The assay was performed in the same way as in the caseof the estrogen RGA, with a few exceptions. Phenol red free DMEMmedia was not required as phenol red does not present androgenic-ity. Moreover, for standard curve development DHT was used as thecalibrant and cells treated with the standard were incubated 48 h,prior to reading the results.

2.3. Cross-reactivity study

Cross-reactivity profiles for estrogen and androgen assays wereperformed using 17�-estradiol, estrone and testosterone, tren-bolone; respectively. The experiments were performed in the sameway as described for 17�-estradiol and DHT (Section 2.2) and eachEC50 was calculated.

Author's personal copy

36 M. Plotan et al. / Analytica Chimica Acta 700 (2011) 34– 40

Table 1Comparison of LC–MS/MS and RGA methods based on seven different sport supplements analysis performed by Horseracing Forensic Laboratory (HFL) and Queen’s UniversityBelfast (QUB). ‘–’ no response observed; ‘+’ positive response observed, ‘+++’ very strong positive response observed.

Sample No. HFL results (LC–MS/MS) QUB results (RGA)

Estrogenic compounds Androgenic compounds Estrogen assay response Androgen assay response

1 − − − −2 − − +++ −3 − − + −4 − Andosteronedione + +5 − DHEA − −6 Estrenedione DHEA + −7 Unknown Unknown + −

2.4. MTT assay

Cytotoxicity testing of all samples was performed on both celllines, MMV-Luc and TARM-Luc. The assay was performed with afew modifications to the method as described by Mossman [18].Cells were seeded in 96 well plates at a concentration of 4 × 105

cells mL−1 and incubated overnight at 37 ◦C. The following day,standards and extracts were added for a further 24 h. Cells werewashed with PBS and MTT reagent (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide 5 mg mL−1 was added (20 �L per50 �L media) and incubated for 4 h at 37 ◦C. The liquid was carefullyremoved and 200 �L DMSO added. Plates were shaken at 37 ◦C for10 min prior to reading absorbance on a Tecan, Safire 2 plate readerat 570 nm.

2.5. Extraction and clean-up of sport supplements

2.5.1. Oasis HLB cartridge SPEAn extraction procedure was developed for protein and carbo-

hydrate based products. The procedure involved pre-treatment,liquid/liquid extraction [19] of 1 g of sample homogenised with10 mL of HPLC grade water, (un)spiked with relevant hormone,followed by Oasis HLB cartridge extraction. Columns 5 cc/200 mgwere used for estrogen assay (glass) and 6 cc/200 mg for andro-gen assay (plastic), and conditioned with 3 mL of methyl tert-butylether, methanol and water. After sample loading, cartridges werewashed with 3 mL of 5% methanol in water and eluted by 6 mLof 10% methanol in methyl tert-butyl ether. Eluent was evapo-rated under a nitrogen stream at 40 ◦C and re-suspended in 250 �Lmethanol.

2.5.2. ‘QuEChERS’ methodAn extraction procedure was developed for protein, carbo-

hydrate, plant and herbal extract, mineral, vitamin, amino acidbased products. Dispersive solid phase extraction (dSPE), i.e. ‘theQuEChERS’ method, involved mixing 1 g of sample with 10 mLof acetonitrile, (un)spiked with relevant hormone. dSPE CitrateExtraction Tube was used for pre-treatment, and further cleanupwas performed using the PSA/ENVI-Carb SPE CleanUp Tube 2. Thecollected aliquot was evaporated under a nitrogen stream at 40 ◦Cand re-suspended in 250 �L methanol.

Extracted samples were tested immediately or stored at −20 ◦C.Prior to application to the RGAs and MTT assay, the extracts werediluted in assay medium 1:200, v/v. When samples exhibited a toxiceffect, additional serial dilutions in assay medium were performed,until tolerable conditions for the RGAs were achieved.

2.6. Extraction recoveries and matrix effect

Pure water or acetonitrile was spiked with 17�-estradiol orDHT at 0.5 ng mL−1 and 3 ng mL−1 concentrations, respectively andextracted to determine the procedure’s recoveries. A matrix effectstudy was performed for the extraction method based on Oasis

HLB SPE. Pure water and sample homogenate spiked with 17�-estradiol at concentrations between 0.0001 and 2.7 ng mL−1 or DHTin range 0.001–29 ng mL−1 were extracted and applied to RGA. Thedeveloped standard curves were compared to 17�-estradiol or DHTstandard curves generated as described in Section 2.2.

2.7. Sport supplements

A total of 63 sport supplement samples were obtained from var-ious retail outlets across the Island of Ireland, via the internet andsome were provided by the Irish Medicine Board. Additionally, 7certified negative and positive control sport supplements (for estro-gens and androgens) were provided by the Horseracing ForensicLaboratory, UK (HFL) as shown in Table 1. The sport supplementswere composed of different ingredients but generally included pro-teins, carbohydrates, amino acids, vitamins, minerals, plant andherbal extracts. The supplements were present in various physicalforms including powders, tablets, capsules or liquids.

2.8. Data analysis

Standard curves were calculated based on the sigmoidal dose-response curve equation using the Slide Write Plus software, whereY is the response, X logarithm of concentration, while bottomis 0 and top 100 percentage of maximum achieved response;Y = bottom + (top − bottom)/(1 + 10(log EC

50−X)).

The EC50 is the standard concentration that produces a 50%fold-induction of the maximal response achieved. Cross-reactivitywas determined for the two most potent natural androgen andestrogen hormones, DHT and 17�-estradiol, respectively. The per-centage value for the cross-reactivity of each tested hormone wascalculated by the division of the EC50 of DHT or 17�-estradiol byEC50 of androgen or estrogen and multiplied by 100. Fold inductionwas estimated by division of sample response by negative controlresponse.

Extraction recoveries were calculated based on standard curveequations, where the X value was calculated for a given Y value,response achieved from the assay. Calculated in this way concen-tration was divided by the value of spiking concentration beforeextraction and multiplied by 100 to achieve a % extraction recovery.

Toxic effect was determined based on the absorbance (Abs) val-ues from the formula:

% toxicity = (negative control Abs − sample Abs)/

(negative control Abs) ∗ 100

3. Results and discussion

3.1. Standard curves

Dose response curves generated in assay media using 17�-estradiol in the estrogen assay and DHT in the androgen assay

Author's personal copy

M. Plotan et al. / Analytica Chimica Acta 700 (2011) 34– 40 37

Fig. 1. Dose response standard curves. Mean induction ± standard deviation in relation to untreated cells is presented. (a) MMV-Luc cell line treated with increasingconcentrations of 17�-estradiol (n = 7). Luciferase activity measured after 24 h. (b) TARM-Luc cell line treated with increasing concentrations of DHT (n = 8). Luciferaseactivity measured after 48 h.

are presented in Fig. 1. The calculated EC50 for the estrogen andandrogen assay were 0.01 ng mL−1 and 0.16 ng mL−1 respectively.

3.2. Cross-reactivity study

Assay cross-reactivity (% ACR) profiles were obtained for boththe estrogen and androgen RGAs. ACR in the MMV-Luc assay was100% for 17�-estradiol (reference hormone), while for less potentestrogenic hormones it was 5% and 4.5% for 17�-estradiol andestrone, respectively. ACR in the TARM-Luc cell line was 100% fordihydrotestosterone (reference hormone), 62% for testosteroneand 28% for trenbolone. Testosterone is a weaker androgen thenDHT and has a binding affinity for androgen receptor in theregion of 25% [20]. Trenbolone has been shown [20] to displaycomparably similar androgenic potency to DHT. However, in thepresent study trenbolone was shown to be less potent than DHT asdemonstrated by their EC50 values of 0.58 and 0.16 respectively.This unexpected lower cross-reactivity may be explained by a veryhigh fold induction.

3.3. Extraction recoveries and comparison of both extractionprocedures

The efficiency of the HLB based extraction and ‘QuEChERS’extraction was determined for both the estrogen and androgenassay.

The percentage recoveries presented in Table 2 indicate thatboth solid phase extractions show high recoveries. In determin-ing the most appropriate extraction procedure a second factor wastaken into consideration, i.e. the applicability of the method to dif-ferent types of samples. Eleven sport supplements were extractedby both methods and applied to the RGAs and MTT assay in order tocheck the toxicities of the different extracts generated. All extractsfrom samples which contained mainly proteins and/or carbohy-drates presented similar RGA results with no toxic effects beingobserved. Conversely, products containing a high amount of plantand herbal extracts or mixtures of minerals, amino acids and vita-

Fig. 2. Matrix effect in estrogen RGA. Standard curves in medium (cell cultureassay media), extracted HPLC water and extracted sport supplements, spiked beforeextraction with different concentrations of 17�-estradiol.

mins with other synthetic compounds, were more problematic assummarised in Table 2. HLB extracts were found to be highly toxicin a number of cases and the determined toxicity was equivalentto the intensity of each extract pigmentation. However, all extractsproduced by the ‘QuEChERS’ method were non toxic and suitablefor the RGAs and this method was observed to remove most of thepigments present.

Due to the high extraction efficiency and low toxicity of the‘QuEChERS’ method this was selected for all further sample prepa-ration work.

3.4. Matrix effect

A study of the matrix effects was performed in both RGAs andthe results are shown in Figs. 2 and 3. In both assays, only a veryweak matrix effect was observed.

Table 2Comparison of HLB and ‘QuEChERS’ solid phase extraction procedures. Recoveries with standard deviations (n = 10) for both estrogen and androgen RGAs. Toxicity study(MTT assay) of 11 different sport supplements: ‘+’ successful extraction, no toxic effect; ‘−’unsuccessful extraction or/and toxic effect.

Extraction method Recoveries Main supplement ingredient

Estrogen RGA Androgen RGA Proteins Carbohydrates Plant and herbal extracts Minerals, amino acids and vitamins

SPE:HLB 94 ± 23% 104 ± 12% + + − −dSPE:‘QuEChERS’ 99 ± 16% 94 ± 14% + + + +

Author's personal copy

38 M. Plotan et al. / Analytica Chimica Acta 700 (2011) 34– 40

Fig. 3. Matrix effect in androgen RGA. Standard curves in medium (cell cultureassay media), extracted HPLC water and extracted sport supplements, spiked beforeextraction with different concentrations of DHT.

3.5. Reference samples

Sport supplement samples kindly provided by Dr Phil Teale,Horseracing Forensic Laboratory and previously investigated usingan LC–MS/MS method [21] were examined by the RGAs. This studyaimed to confirm the suitability of RGA analysis for the detectionof hormonally active constituents in sport supplements. Table 1describes the comparison of data produced by both RGA and LC-MS/MS analysis. The RGA method detected an estrogenic responsein sample 6, a sample in which LC–MS/MS identified the presenceof the estrogenic compound estrenedione. In samples 2, 3 and 4 theRGA also detected the presence of estrogenic compounds howeverno matching evidence was found by LC–MS/MS. As the LC–MS/MSprocedure was specifically designed to detect the presence of illegalanabolic compounds, a second LC–MS/MS procedure was appliedwhich was designed to detect phytoestrogens. The profiling for 13compounds was performed using the procedure of Antignac et al.[19]. The samples profiled as estrogenic positive by RGA containedhigh concentrations of six plant estrogens, most notably in sam-ple 2 which showed the strongest response in the estrogen RGA(Table 3).

In relation to the androgen RGA analysis; agonist activity wasdetected only in sample 4. LC–MS/MS investigation had revealedthat sample 4 contained androstenedione. Samples 5 and 6 didnot present any activity in the androgen RGA. However, LC–MS/MSanalysis had confirmed the presence of DHEA, which is a naturalprecursor of steroid hormones such as testosterone, DHT, estroneand estradiol in humans [3]. This weak androgen has been shownto have a very low binding affinity for the androgen receptor [20].

3.6. Sport supplement study

Sixty three sport supplement samples were collected and inves-tigated by estrogenic and androgenic RGAs. Experiments weredesigned to detect agonistic and antagonistic activities. However,when the resulting data were analysed more groupings (n = 5) of

results for both assays were required to explain the effects observed(Fig. 4a and b).

3.6.1. Estrogen RGAEighteen negatives were found from the 63 samples tested for

the presence of estrogenic compounds, including 5 products thatshowed a very low level of activity close to the negative responselevel. These 5 samples were a common type of supplement withwhey protein and carbohydrates as their main constituents andmay require alternative sample treatment to remove a potentialmatrix effect. Of the 45 supplements showing a response, three sub-groups were apparent. As presented in Fig. 4a; 24 were classifiedas partial agonists, 17 were classified as agonists showing addi-tive responses, and 4 classified as agonists displayed synergisticresponses. Samples showing a similar or higher response than 17�-estradiol at a concentration of 0.5 ng mL−1 in the RGA were selectedand tested for a range of estrogenic compounds by LC–MS/MSmethod [22], including natural and synthetic estrogens, estrogenicmycotoxins and their metabolites and esters forms. None of thepotential hormonal substances (� and �-estradiol, estrone, ethyny-loestradiol, dienestrol, diethylstibbestrol, hexestrol, delmadiononeacetate, zeranol, taleranol, zearalenone, � and � zearalenol) werefound in any of the samples investigated. Additionally, confirma-tory analysis was also performed to look for estrogenic ester formsof some of the drugs, none were detected. However, in 7 samples(all soy based products), from selected supplements, the presenceof isoflavone-genistein (a known estrogen agonists) was confirmedby LC–MS/MS, explaining the positive response observed in theestrogen RGA.

In summary, the study detected the presence of estrogenic com-pounds in 45 of the 63 samples examined. Mass spectrometryanalysis confirmed the presence of estrogenic substances in onlya few examples, with mainly phytoestrogens being found in soybased products. The activity detected in the other samples may beexplained by the presence of natural and synthetic estrogens notincluded in the mass spectrometric profiles, estrogenic environ-mental contaminants, such as pesticides, fungicides, plasticisers,etc. These may contaminate final product during the manufactur-ing processes or originate in raw materials from with which theywere produced.

3.6.2. Androgen RGAThe androgenic RGA enabled investigation and analyses of

target products in relation to naturally occurring androgens, dihy-drotestosterone (DHT) at concentration of 2.9 ng g−1.

Of the 63 supplements tested, 18 were negative and showedno androgenic activity. The largest grouping was represented byandrogenic antagonists (26), of which 16 antagonised the recep-tor by blocking and 10 which partially antagonised the receptor bymimicking hormone action with lower efficiency. In 12 samples,agonistic substances, both with additive and synergistic responsesin the androgen RGA were found. An additional group of 7 sam-ples showed no agonistic activity when analysed alone. However,a significant response was observed when these samples wereanalysed in the presence of DHT and was greatly enhanced in com-parison to the control containing the same concentration of DHT

Table 3Phytoestrogen profiling performed by LC–MS/MS in LABBERCA (Nantes, France) on three sport supplement samples which showed estrogenic responses when analysed byRGA. Phytoestrogen concentrations are shown in ng g−1.

Sample No. Daidzeine Genisteine Secoisolariciresinol Enterolactone Equol Apigenine

2 946 1700 215 1290 290 136403 202 120 178 830 40 44904 226 190 405 920 20 5290

Author's personal copy

M. Plotan et al. / Analytica Chimica Acta 700 (2011) 34– 40 39

Fig. 4. ED profile for 63 sport supplements investigated by RGA: (a) samples having estrogenic activity equivalent to 17�-estradiol control detected using MMV-Luc cell linebased RGA are shown, (b) androgen active samples being equivalent to DHT control concentration were detected using the TARM-Luc cell line based RGA.

only. Yeh et al. [23] and Tyagi et al. [24] report that potent estro-gens can enhance the response in androgen RGAs. This may bedue to translocation of the androgen receptor and trans-activationof target gene promoters by high concentrations of estrogens like17�-estradiol. In 4 of these 7 samples, strong estrogenic agonistswere observed by the estrogen RGA. Moreover, flavonoids causea stabilisation of the luciferase marker enzyme [25]. Genistein forinstance is not able to give a response in an androgen RGA, but sta-bilises the luciferase that is induced by DHT. In combination it mayseem like there is a compound that strengthens the signal inducedby DHT.

However, 3 of these samples did not show estrogenic activity.Increased androgenic activity may also be a result of enhancingthe androgen-mediated transcription through increasing androgenreceptor expression as has recently been shown by the chemicaltriclocarban [26].

A comparative study using a yeast based reporter gene assay[27] was performed on 3 samples presenting the strongest andro-genic agonistic activity by the mammalian RGA. The supplementsshowed no response in comparison to a testosterone control addedat 1.45 �g g−1. The observed differences between the two RGAsystems may be due to the differential sensitivity of the two proce-dures. Alternatively, the differences may be due to the presence offlavonoids that stabilise the luciferase marker enzyme, this effectis not observed in the yeast with yEGFP as a marker protein.

Four samples, showing the strongest antagonistic effect, wereanalysed by mass spectrometry. A profile of 75 pesticides havingandrogenic antagonistic effect was examined by GC/MS, LC–MS TOFand GC/MS/MS method [28]. No pesticide residues were found tobe present in any sample at concentrations higher than the limit ofdetection (1 ng g−1).

In summary, the androgen RGA was able to detect activity in 45of the 63 tested sport supplements. The exact nature of the causeof the activity has yet to be determined.

4. Conclusions

The developed RGAs have been shown to be very sensitiveanalytical tools when combined with the ‘QuEChERS’ extractionprocedure for the detection of androgenic and estrogenic bio-activity in sport supplements. Comparison of the estrogen andandrogen RGA results with LC–MS/MS analysis show that theassays are capable of detecting EDs including steroid hormonesand phytoestrogens. The RGAs are also capable of detecting lowconcentrations of EDs and weak steroids. However, these assays areunable to detect steroid precursors having low affinity for nuclearreceptors.

The identification of the constituents responsible for thedetected ED bio-activity requires a confirmatory method. The use

of a confirmatory method alone may present false negatives due tostructural specificities. An RGA in parallel with a LC–MS/MS methodwas successfully used in a previous study [27] in detecting anabolicsteroids in a number of dietary supplements, previously declarednegative by mass spectrometry analysis. Therefore, a combinedapproach of screening and confirmatory method is most suitablefor the detection and identification of ED constituents.

The main objective of this study was to develop methods forthe detection of estrogenic and androgenic EDs in sport supple-ments. This objective was fulfilled. However, the data obtained wasmore complex than expected. Results indicating the presence ofagonistic, partial agonistic and antagonistic responses were alsoobtained. Additionally, a few supplements contained extracts thatenhanced the hormone response but did not have hormonal activ-ity themselves. It may be that these potential EDs are not actingon the steroid receptor directly, but can influence the activation ofthe receptor, strengthening the hormone and hormone receptorinteraction or enhancing nuclear receptor synthesis or produc-tion. Another possibility is their interaction and positive influenceon the cofactors that are involved in the induction of the steroidendocrine pathway within the cell [29] or the previously men-tioned stabilisation of the luciferase marker enzyme by flavonoids[25]. These new-classes of EDs are being extensively investigatedby scientists [6]. The complexity of results in this study impliesthat different types of hormonally active constituents are presentin health food products and the physiological effects they may haveon consumers is unknown. Besides the favourable effects that sportsupplements may have on body composition, the potential ED sub-stances present in these products may also have undesirable effects.The list of detrimental effects they may cause in humans and theirprogeny is described in many reviews and clinical studies providinglinks with infertility, heart problems and different types of carci-noma [8,30–32]. Furthermore, of concern is that this study obtaineda positive hormonal response in more then 70% of the tested prod-ucts. This reinforces increasing publicised data about the presenceof EDs in healthfood supplements [12,13,27].

Confirmatory analysis was carried out to search for syntheticEDs which could be responsible for the bio-activity observed insupplements. However this approach did not confirm ED iden-tity in most cases. Many samples also contained plant or herbalingredients. A range of 7 phytoestrogens which are able to elicitan estrogenic response were present in approximately 20% of theinvestigated samples. These samples elicited double the responsefound by 0.5 ng mL−1 of 17�-estradiol. This study highlights thepotential source of natural EDs in the diet. It may be advisable tocarry out a risk assessment on such products having similar ingre-dients, including soy and plant extracts.

In conclusion, bio-activity screening in combination with con-firmatory analysis is optimal for the detection of EDs in sport

Author's personal copy

40 M. Plotan et al. / Analytica Chimica Acta 700 (2011) 34– 40

supplements. It is also clear that potential EDs are constituents ofa major proportion of food supplements purchased on the Islandof Ireland and elsewhere. Therefore, legislation is essential for thecontrol of these types of foodstuff. However, further research onthe detection, characterisation and potential adverse health effectcaused by EDs of plant origin is also necessary.

Acknowledgements

The authors would like to thank Mr Hugo Bonar from The IrishMedicine Board and Dr Phil Teale from Horseracing Forencing Lab-oratory for the control samples. We also thank Dr Martin Danaherfor project support and Miss Caroline Frizzell for technical sup-port. This research was carried out under the “Health Food” projectfunded by Irish Department of Agriculture, Fisheries and Food; andThe Health Research Board (HRB), in collaboration with the Depart-ment of Health and Children, under the Food for Health ResearchInitiative (FHRI).

References

[1] S.S. Nussey and S.A. Whitehead (2001) Endocrinology. An IntegratedApproach. Taylor & Francis, http://www.ncbi.nlm.nih.gov/bookshelf/br.fcgi?book=endocrin.[19th September 2008].

[2] D.M. Oh, T.J. Philips, Sex hormones and wound healing, Wounds 18 (1) (2006)p.8–18.

[3] D.G. Gardner, D. Shoback, Glucocorticoids and Adrenal Androgens, TestesGreenspans, in: Basic and Clinical Endocrinology, 8th ed., The McGraw-Hill,New York, 2007, pp. 405, 470–474, 865.

[4] European Commission (1996), European Workshop on the Impact of EndocrineDisrupters on Human Health and Wildlife, 2–4 December 1996, Weybridge, UK.

[5] C. Sonnenschein, A.M. Soto, An updated review of environmental estrogen andandrogen mimics and antagonists, Journal of Steroid Biochemistry and Molec-ular Biology 65 (1998) 143–150.

[6] R.M. Sharpe, N.E. Skakkebaek, Male reproductive disorders and the role ofendocrine disruption: Advances in understanding and identification of areasfor future research, Pure and Applied Chemistry 75 (2003) 2023–2038.

[7] M.S. Juhn, Popular sports supplements and ergogenic aids, Sports Medicine 33(2003) 921–939.

[8] C. Maravelias, A. Dona, M. Stefanidou, C. Spiliopoulou, Adverse effects ofanabolic steroids in athletes A constant threat, Toxicology Letters 158 (2005)167–175.

[9] H.G. Pope, D.L. Katz, Affective and physiologic symptoms associated withanabolic steroid use, The American Journal of Psychiatry 145 (1988) 487–490.

[10] L. Connolly, Endocrine disrupting chemicals: origins, fates and transmissioninto the food chain, in: I. Shaw (Ed.), Endocrine Disrupting Chemicals In Food,Woodhead Publishing Ltd, 2009, pp. 103–125 (5).

[11] European Communities (Food Supplements) Regulations 2007, S>1. No.506 of 2007, Directive 2006/37/EC [20th October 2008] http://www.fsai.ie/legislation/food/eu docs/food supplements/SI 506 2007.pdf.

[12] R.J. Maughan, Contamination of dietary supplements and positive drug tests insport, Journal of Sports Sciences 23 (2005) 883–889.

[13] H. Geyer, M.K. Parr, U. Mareck, U. Reinhart, Y. Schrander, W. Schanzer, Analysisof non-hormonal nutritional supplements for anabolic – androgenic steroids –results of an International Study, Journal of Sports Medicine 25 (2004) 124–129.

[14] T.F. Delbeke, P.V. Eenoo, W.V. Thuyne, N. Desmet, Prohormones and sport, TheJournal of Steroid Biochemistry and Molecular Biology 83 (2003) 245–251.

[15] M.I. Kafrouni, R.A. Anders, S. Verma, Hepatotoxicity associated with dietarysupplements containing anabolic steroids, Clinical Gastroenterology and Hep-atology 5 (7) (2007) 809–812.

[16] L. Connolly, K. Cai, E. Van der Heiden, M.L. Scippo, M. Muller, J. Torbin, C. Elliott,Detection of glucocorticoid bioactivity in bovine urine samples using a reportergene assay, Analytica Chimica Acta 637 (1–2) (2009) 321–327.

[17] P. Willemsen, M.L. Scippo, G. Kausel, J. Figueroa, G. Maghuin-Rogister, J.A. Mar-tial, M. Muller, Use of reporter cell lines for detection of endocrine-disrupteractivity, Analytical and Bioanalytical Chemistry 378 (2004) 655–663.

[18] T. Mossman, Rapid colorimetric assay for cellular growth and survival: applica-tion to proliferation and cytotoxicity assays, Journal of Immunological Methods65 (1983) 55–63.

[19] J.P. Antignac, I. Gaudin-Hirret, H. Naegeli, R. Cariou, C. Elliott, B. Le Bizec,Multi-functional sample preparation procedure for measuring phytoestrogensin milk, cereals, and baby-food by liquid-chromatography tandem mass spec-trometry with subsequent determination of their estrogenic activity usingtranscriptomic assay, Analytica Chimica Acta 637 (1–2) (2008) 55–63.

[20] M.L. Scippo, C.V.D. Weerdt, P. Willemsen, J.M. Francois, F. Rentier-Delrue, M.Muller, J.A. Martial, G. Maghuin-Rogister, Detection of illegal growth promotersin biological samples using receptor binding assays, Analytica Chimica Acta 473(2002) 135–141.

[21] J. Scarth, A. Clarke, J. Hands, P. Teale, R. Macarthur, J. Kay, Validation of aquantitative multi-residue urinary assay for the detection of androgen, oestro-gen and progestagen abuse in the bovine, Chromatographia 71 (3–4) (2009)241–252.

[22] E. Malone, C. Elliott, D. Kennedy, L. Regan, Rapid confirmatory method for thedetection of sixteen synthetic growth promoters and bisphenol A in bovine milkusing dispersive solid-phase extraction and liquid chromatography–tandemmass spectrometry, Journal of Chromatography B 878 (2010)1077–1084.

[23] S. Yeh, H. Miyamoto, H. Shima, C. Chang, From estrogen to androgen receptor:a new pathway for sex hormones in prostate, in: Proceedings of the NationalAcademy of Sciences 95, 1998, pp. 5527–5532.

[24] R.K. Tyagi, Y. Lavrovsky, S.C. Ahn, C.S. Song, B. Chatterje, A.K. Roy, Dynamicsof intracellular movement and nucleocytoplasmic recycling of the ligand-activated androgen receptor in living cells, Molecular Endocrinology 14 (8)(2000) 1162–1174.

[25] A.M. Sotoca, T.F.H. Bovee, W. Brand, N. Velikova, S. Boeren, A.J. Murk, J. Ver-voort, I.M.C.M. Rietjens, Superinduction of estrogen receptor mediated geneexpression in luciferace based reporter gene assays is mediated by a post-transcriptional mechanism, Journal of Steroid Biochemistry and MolecularBiology 122 (4) (2010) 204–211.

[26] J. Chen, K.C. Ahn, N.A. Gee, M.I. Ahmend, A.J. Duleba, L. Zhao, S.J. Gee, D.B. Ham-mock, B.L. Lasley, Triclocarban enhances testosterone action: a new type ofendocrine disruptor? Endocrinology 149 (2008) 1173–1179.

[27] J.C.W. Rijk, T.F.H. Bovee, S. Wang, C. Van Poucke, C. Van Peteghem, M.W.F. Nie-len, Detection of anabolic steroids in dietary supplements: the added value ofan androgen yeast bioassay in parallel with a liquid chromatography–tandemmass spectrometry screening method, Analytica Chimica Acta 637 (2009)305–314.

[28] R. Hill, S. Head, S. Baker, M. Gregg, D. Shealy, S. Bailey, C. Williams, E. Sampson,L. Needham, Pesticide residues in urine of adults living in the United States:reference range concentrations, Environmental Research 71 (2) (1995) 99–108.

[29] M.M. Tabb, B. Blumberg, New modes of action for endocrine-disrupting chem-icals, Molecular Endocrinology 20 (2006) 475–482.

[30] J.A. McLachlan, E. Simpson, M. Martin, Endocrine disrupters and femalereproductive health, Best Practice and Research Clinical Endocrinology andMetabolism 20 (2006) 63–75.

[31] S.M. Snedeker, Pesticides and breast cancer risk: a review of DDT, DDE andDieldrin’ Environmental Health Perspectives 109 (2001) 35–47.

[32] M. Yang, M.S. Park, H.S. Lee, Endocrine disrupting chemicals: human exposureand health risks, Journal of Environmental Science and Health Part C 24 (2006)183–224.