Contents lists available at ScienceDirect

Journal of Steroid Biochemistry and Molecular Biology

journal homepage: www.elsevier.com/locate/jsbmb

The art of measuring steroidsPrinciples and practice of current hormonal steroid analysis

S.A. Wudya,⁎, G. Schulerb, A. Sánchez-Guijoa, M.F. Hartmanna

a Steroid Research &Mass Spectrometry Unit, Laboratory for Translational Hormone Analytics in Paediatric Endocrinology, Division of PaediatricEndocrinology & Diabetology, Center of Child and Adolescent Medicine, Justus Liebig University, Giessen, Germanyb Veterinary Clinic for Obstetrics, Gynecology and Andrology of Large and Small Animals, Faculty of Veterinary Medicine, Justus Liebig University, Giessen, Germany

A R T I C L E I N F O

Keywords:SteroidImmunoassayGas chromatographyLiquid chromatographyMass spectrometryMetabolomics

A B S T R A C T

Steroids are small and highly important structural or signalling molecules in living organisms and their meta-bolism is complex. Due to the multiplicity of enzymes involved there are many different steroid related disorders.E.g., an individual enzyme defect is rather rare but can share various clinical symptoms and can thus be hardlydiagnosed clinically. Therefore, reliable hormonal determination still presents the most reasonable initial di-agnostic approach and helps to avoid uncritical and expensive attempts at molecular diagnostic testing. It alsopresents a backbone of monitoring these complex patients. In science, reliable hormone measurement is indis-pensable for the elucidation of new mechanisms of steroid hormone actions.

Steroid analytics is highly challenging and should never be considered trivial. Most common methods forsteroid determination comprise traditionally immunoassay, or more recently, mass spectrometry based methods.It is absolutely necessary that clinicians and scientists know the methods they are applying by heart. With theintroduction of automated direct assays, a loss of quality could be observed over the last two decades in the fieldof steroid immunoassays.

This review wants to meet the need for profound information and orientation in the field of steroid analysis.The pros and cons of the most important methods, such as immunoassays and mass spectrometry based methodswill be discussed. The focus of the latter will lie on gas chromatography-mass spectrometry (GC–MS) as well asliquid chromatography-mass spectrometry (LC–MS). Selected analytical applications from our DeutscheForschungsgemeinschaft Research Group FOR 1369 “Sulfated Steroids in Reproduction” will illustrate thecontents.

In brief, immunoassays have for long presented the traditional technique for steroid analysis. They are easy toset up. Only one analyte can be measured per immunoassay. Specificity problems can arise and caution has to beexerted especially regarding direct assays lacking purification steps. Mass spectrometry based methods providestructural information on the analyte and thus higher specificity. In combination with chromatographic tech-niques, they permit the simultaneous determination of a multitude of analytes. Highest specificity can be ob-tained using GC–MS, a sophisticated but most powerful tool for characterizing steroid metabolomes. LC–MS is atrue high throughput technique and highly suited for detecting complex steroids. GC–MS and LC–MS are notcompeting but complementary techniques.

Since reliable steroid determination requires extremely high expertise in the field of analytics as well assteroid biochemistry, it is recommended that collaborations and networking with highly specialized centers ofexpertise are developed.

1. Introduction

1.1. Measuring steroids, an art?

It is not easy to reliably measure steroids. When one looks up the

meaning of art, e.g. in Merriam-Webster’s Dictionary [1] one of thedefinitions of art is “an occupation requiring knowledge or skill”. Thesetwo elements, in depth knowledge as well as skill are indispensableprerequisites for successful steroid determination. Both, clinicians andscientists, ordering tests for steroid hormones, have to know the

http://dx.doi.org/10.1016/j.jsbmb.2017.09.003Received 31 March 2017; Received in revised form 18 July 2017; Accepted 5 September 2017

⁎ Corresponding author at: Head, Division of Paediatric Endocrinology & Diabetology, Center of Child and Adolescent Medicine, Justus Liebig University, Feulgenstrasse 12, 35392Giessen, Germany.

E-mail address: stefan.wudy@paediat.med.uni-giessen.de (S.A. Wudy).

Journal of Steroid Biochemistry and Molecular Biology 179 (2018) 88–103

Available online 28 September 20170960-0760/ © 2017 Elsevier Ltd. All rights reserved.

T

methods applied by heart. Conversely, laboratory scientists involved insteroid measurements have not only to be skilled analysts; they alsoneed to have in depth knowledge of steroid metabolism and its dys-functions. Furthermore, clients and analysts need to communicate andcooperate as closely as possible to select the right assay and to interpretthe results correctly [2].

1.2. Measuring steroids in times of political deregulation

As with all societal events, political circumstances also influencemedical and scientific actions. The determination of steroid hormones,their precursors and their metabolites has once been a mainstay ofendocrinology, the science devoted to the action of hormones. It is in-conceivable but true that during the last two to three decades, primarilyausterity measures have led to a dissociation of hormone laboratoriesfrom endocrinology and finally to their absorption by often remotehuge centralized, supposedly more economic laboratory platforms,where steroids represent just one class of parameters among a multitudeof others. Little wonder that this development went hand in hand with aloss of expertise in assay development, selection, interpretation, andinadequate economic valuation, finally leading to a crisis in the qualityof steroid analysis in the clinic and research.

This development is more worrisome than one tends to think and itis typical against the background of the current political situation. It isembedded in a world of politically intended increased deregulation,leading to augmented and fierce competition and finally to the survivalof the economically strongest. Furthermore, current reasoning tends toequalize frequent occurrence with importance. These changes are fatalif transferred to medicine and basic science. Disorders of steroid me-tabolism are rare and belong to the most complex disorders. It goeswithout saying that these latter qualities will neither put patients withsteroid disorders nor scientists trying to elucidate the underlying me-chanisms in a position to present a democratically or an economicallystrong community. It is therefore high time for politicians to eventuallyrealize that the current socio-economic system highly discriminatesagainst and threatens both patients with rare and complex disorders aswell as the activity of scientists devoted to investigate these diseases.

1.3. Why measuring steroids in times of molecular genetics?

Truly, we have seen many impressive advances in the field of mo-lecular genetics, but admittedly not all phenomena in a patient can beexplained by characterizing his genome. We have learnt that a finger-print of his metabolic profile is often much more informative andconcordant with the clinical phenotype. In addition, comprehensivemetabolic assessment is not only a powerful diagnostic tool but alsoallows for monitoring of the disease.

This article is devoted to all of those who wish to be informed on theprinciples and practice of current analysis of natural steroid hormones,their precursors and metabolites. The article addresses clinicians, aswell as hormone analysts, and researchers. To keep its scope at a rea-sonable limit, steroids such as bile acids or vitamin D and its metabo-lites have not been enclosed in this review.

While immunoassay has for long been the predominant assaytechnique for measuring steroids, the introduction of mass spectro-metry (MS) based techniques has increased the variety of analyticalmethods. Especially one of the recent developments, liquid chromato-graphy-mass spectrometry (LC–MS) is about to “flood” the market. Theneed for profound information and orientation in the field of steroidanalysis is further reflected in an increasing number of conflictingpublications showing that obviously competition and confusion hasarisen in the field of steroid analysis particularly between immunoassayand MS based techniques [3]. But also in the field of MS based steroidanalytics there seems uncertainty about the roles of gas chromato-graphy-mass spectrometry (GC–MS), the older, and LC–MS, the youngertechnique. Thus it seems to be an opportune time to provide a differ-entiated view and perspective reconsidering the pros and cons of eachanalytical approach.

1.4. Role of DFG research group 1369 “Sulfated Steroids in Reproduction”

To exemplify the contents of this article, the inclined reader willfind examples of typical applications of the analytical techniquesmentioned and discussed. These examples were taken from collabora-tions of the authors within Research Group 1369 “Sulfated Steroids inReproduction”, a worldwide unique research consortium investigatinghitherto unknown biological functions of sulfated steroids, which hasbeen funded by the Deutsche Forschungsgemeinschaft (DFG). To ensuremost reliable steroid analysis within this research project, one sub-project was devoted to MS based steroid analytics (“LC–MS and GC–MSbased Steroidomics”, Stefan A. Wudy, principal investigator). As thisspecial edition of the Journal of Steroid Biochemistry and MolecularBiology will entirely be devoted to essential outcomes of this researchgroup, the reader will find more applications in further publications ofthis volume.

1.5. Steroids, structure and nomenclature

Steroids are small molecules. However, they are essential for prac-tically all forms of life. On the one hand, they are indispensable for theformation of cell structures. On the other hand, they can act as systemicor local signaling molecules (hormones, paracrine or intracrine reg-ulatory factors), constituting an elaborate and highly important in-formation transfer system. Already tiny structural changes can result indramatic functional changes.

The structural common feature of steroids is the 4-membered hy-drocarbon ring system (sterane, cyclopentanoperhydrophenanthrene)consisting of three six-membered carbon rings A, B, C and one 5-membered carbon ring D (Fig. 1). All thousands of natural and syntheticsteroids are derivatives of that core. Most steroid compounds are de-rived from the following six basic hydrocarbons: the C18 steroids es-tranes, their name rings in trivial names like estradiol, estrone; the C19

steroids androstanes, their name reminds us of androgens; the C21

steroids resulting from pregnane, its echo – pregnancy – points to ge-stagens; the C24 steroids cholanes, to be found in cholic alcohols andacids; the C27 steroids cholestanes, who form the foundation of sterols;

Fig. 1. Sterane (cyclopentanoperhydrophenan-threne) and cholesterol.The structural common feature of steroids is the 4-membered hydrocarbon ring system consisting ofthree six-membered carbon rings A, B, C and one 5-membered carbon ring D. The numbering of thecarbon sceleton is shown for cholesterol.

S.A. Wudy et al. Journal of Steroid Biochemistry and Molecular Biology 179 (2018) 88–103

89

and the C27 secosteroid 9,10-secocholestane, which leads to vitamin Dand its metabolites [4,5].

The procedure for naming a steroid according to the InternationalUnion of Pure and Applied Chemistry (IUPAC) nomenclature [6] takesinto consideration the above mentioned hydrocarbon class, unsatura-tion, substituent atoms or groups, esters, ethers, and other derivativesof alcohols, derivatives of carboxylic acids, geometric isomerism andskeletal modifications. However in practice it is rather cumbersome touse the systematic names, e.g. talking about 11β,17,21-trihydrox-ypregn-4-ene-3,20-dione. It is much easier, to use the trivial names andto talk about the same compound as cortisol or hydrocortisone.Sometimes, capital letters are used as abbreviations for steroids, e.g. F isused for cortisol. This phenomenon dates back to the early days of thefield, when the steroids isolated by the groups of Kendall, Reichstein orWintersteiner were – unfortunately not coherently – consecutivelynamed with capital letters from the alphabet. Admittedly, dealing withtrivial names or letters might be confusing for beginners. For thosehaving worked longer in the field, their shortness has undeniably itsadvantages. Chemical structure databases, such as ChemSpider [7],provide quick orientation on nomenclature and structure of steroids.

1.6. Historical aspects of early steroid analysis

Measurements of steroids have been performed already in the pre-immunoassay era by various methods such as bioassays (animal ex-periments, in vitro bioassays), physico-chemical methods and bio-chemical methods. First quantitative data on steroid hormone con-centrations relied on colorimetric staining [8]. However, theapplicability of these methods was limited due to their inherent highexpenditure of work, low sensitivity and/or specificity.

2. Steroid analysis by immunoassay

2.1. History and analytical principle of the radioimmunoassay

It was a milestone in endocrinology, when Yalow and Berson pub-lished in 1959 the establishment of a radioimmunological method forthe measurement of a hormone (insulin) in human plasma. In precedingexperiments they had observed that adding increasing amounts of un-labelled insulin to a known constant amount of antibody bound radi-olabelled insulin progressively displaced the latter from the antibody.By measuring how much labelled insulin (tracer) was released, theycould work out the concentration of the unlabelled hormone (analyte)in a sample by comparing its inhibitory effect on the binding of thetracer to specific antibody with the inhibitory effect of known stan-dards.

This basic principle is still followed in many of the steroid im-munoassays until today (Fig. 2). It took some years until the potentialsof the radioimmunoassay (RIA) had been generally realized, whichwere above all the specificity by the use of specific antisera, sensitivity,low sample volumes and technical simplicity which enabled a highthroughput of samples. However, by the late 1960s, RIA had become animportant tool in endocrinology and has subsequently expanded be-yond endocrinology to many other fields of biomedical research and toclinical laboratories [9–11].

Initially, the establishment of RIA methods for measurements ofsteroids was mainly hampered by difficulties to produce suitable anti-sera. Due to their low molecular weight, steroids are not readily im-munogenic, the more as their structure is independent from the species.Other more or less specific binding molecules with a sufficiently highaffinity for the analyte can also be applied instead of a specific antibody(ligand-binding assays). Thus, various naturally occurring steroidbinding proteins have been used such as corticosteroid binding protein(CBG; for glucocorticoids plus progestagens), sex hormone bindingglobulin (SHBG; for estrogens plus androgens) or steroid receptorsprepared from cytosolic fractions (radioreceptor assays) [12–15].

However, due to the lack of specificity (CBG, SHBG) or instability(steroid receptors), their use was widely abandoned after suitable an-tibodies became available. Nevertheless, radioreceptor assays may stillbe helpful as a screening method for the detection of unknown steroidalor non-steroidal substances interacting with the rather promiscuoussteroid receptors allowing for the binding of other endogenous orexogenous molecules in addition to the steroid considered as the gen-uine ligand, e.g. phytoestrogens [16].

Eventually, the problem with the production of steroid antisera wasovercome by the use of immunogens consisting of the steroid of interestcoupled to a large immunogenic carrier molecule such as serum al-bumin or keyhole limpet hemocyanin, giving the steroid the role of anincomplete antigen (hapten). A subpopulation of antibodies obtainedafter immunization recognizes the free steroid alone and thus may actas specific binding sites for the analyte in immunoassays [17–21]. Inaddition to free steroids, assays for conjugated (sulfated or glucur-onidated) forms have been developed, too. Conjugated steroids mayalso be measured as the corresponding free form after chemical (sol-volysis) or enzymatic hydrolysis and extraction [22,23].

2.2. Hook effect

A rare, but important pitfall in immunoassays is the Hook effect. Inthis condition low antigen determination occurs due to excessivequantities of antigen impairing antigen-antibody binding. The Hookeffect leads to falsely low results and may cause misdiagnosis [24].

2.3. Specificity of immunoassays

A major concern related to the application of immunological

Fig. 2. Basic principle of a competitive radioimmunoassay.The assay system is set up by defined amounts of a specific antibody and of the radi-olabelled steroid (tracer). The analyte present in a sample competes with the tracer for thebinding sites of the antibodies. With increasing concentrations of the analyte, a higherproportion of the tracer is displaced from the antibodies. Measurement of the radio-activity either in the fraction of the free or in the fraction of the antibody bound steroidsafter separation allows to work out the concentration of the analyte in a sample bycomparing its inhibitory effect on the binding of the tracer to specific antibody with theinhibitory effect of known standards.

S.A. Wudy et al. Journal of Steroid Biochemistry and Molecular Biology 179 (2018) 88–103

90

methods for steroid measurements is specificity. Specificity of im-munoassays depends on the specific interaction of the binding site ofantibodies with their target molecule. However, in the case of steroids,the target is exceptionally small in relation to the binding fragment ofthe immunoglobulin molecule, and already subtle modifications at thecarbon backbone can cause enormous diversity in functional activity.E.g. estradiol-17β is a highly potent estrogen, whereas estrogens such asestrone or estradiol-17α are far less active.

2.4. Cross reactivity

Thus, depending on the structure of a certain target steroid and thecharacteristics of the antiserum applied, immunoassays for steroidmeasurements may not be capable of discriminating efficiently betweenthe analyte and other structurally closely related steroids. This mightnot be a problem in the absence of cross-reactive steroids or in casetheir concentrations in a sample are only low. However, a roughoverestimation of analyte concentrations will occur even with weaklycross-reacting steroids when their concentrations are high in relation tothe analyte [25,26].

To avoid pitfalls from cross-reactivity, the applicant has to know thecross-reactivity of the antiserum used. The extent of cross-reactions maybe dependent on various factors such as species, sex, age, type of sample(e.g. blood, urine, feces, saliva, milk, tissue, cell culture supernatant),treatments with steroidal or steroid-like drugs and the presence of othersteroidogenic sources, e.g. placenta, fetal adrenal cortex or tumors[25,27–29].

2.5. Specificity of steroid antisera

The specificity of steroid antisera is significantly influenced by thetype of the immunogen, namely by the position of the chemical bridgebetween the steroid and the carrier protein. Due to the relatively smallsize of steroid molecules, antibodies usable in immunoassays are pre-ferentially those directed against the part of the steroid molecule far-thermost from the chemical bridge to the carrier protein. However, anantiserum obtained after immunization against e.g. estradiol-17ß cou-pled via the C-3 atom of the A-ring, will not discriminate well betweenphenolic and neutral steroids exhibiting a D-ring identical to that ofestradiol-17ß (e.g. testosterone, androst‐5‐ene‐3β,17β-diol or 17ß-es-tradiol-3-conjugates). On the other hand, immunization against an es-tradiol-17-conjugate will generate an antiserum only poorly dis-criminating between various free estrogens. Therefore, it has beenrecommended that for the production of highly specific antisera, ster-oids should be conjugated to proteins at sites on the B or C ring of thesterane backbone. Thus, it is basically possible to tailor the specificity ofsteroid antisera by selecting the optimal site of the steroid molecule forlinkage to the carrier protein [30,31].

Nevertheless, specificity may still be a problem in certain circum-stances, as for example the accurate measurement of 5α-dihy-drotestosterone or testosterone in the presence of both steroids [18,32].In this case, they need to be physically separated prior to im-munological measurement, e.g. by chromatographic methods. On theother hand, rather non-specific antisera may be very useful as “group-specific” reagents in certain applications, e.g. for the assessment ofstructurally closely related metabolites (sum signal) for non-invasivesteroid analysis in faeces from wild or zoo animals to monitor adrenalor ovarian function, or for non-invasive pregnancy diagnosis [33–35].

2.6. Matrix effects in immunoassays

Another important issue are matrix effects. A RIA method estab-lished and validated in a certain species may not be readily applicableto another one due to species-specific components of the sample in-terfering with the assay system. Matrix effects may be especially aproblem when immunoassay is performed with unprocessed samples

(direct assay). Common reasons for matrix effects may be the presenceof binding proteins, cross-reacting steroidal or nonsteroidal moleculesand other mainly lipophilic compounds [19,36,37] present in thesample. Often, their nature remains unidentified. Matrix effects mayalso be a problem within a certain species, when measurements arecarried out in different body fluids or tissues, e.g. tissue homogenates,saliva, milk or feces. Several techniques help to overcome matrix ef-fects, such as liquid or solid phase extraction of the samples (extractionassay), chromatographic prepurification, dilution, heating, precipita-tion or addition of displacing reagents [19,20,37].

2.7. Nonradioactive immunoassays

The performance of radioimmunological measurements is proble-matic due to the safety of the laboratory personnel, disposal of radio-active waste and the high costs for special laboratory facility andcounting equipment. Thus, starting already in the late 1960s effortswere made to replace the radioactive reporter label by nonradioactivealternatives such as enzymes (for review see Lequin 2005 [38]). Later,besides enzyme activities other reporter labels were used, e.g. fluor-escent dyes or ruthenium complexes with electroluminescent properties[39]. The first papers on the measurement of sex steroids and cortisolby enzyme immunoassays (EIA) were published in the mid to late 1970s(for review see van Weemen et al. [40]). Owing to the more challengingsteric situation, the establishment of a reliable EIA is generally clearlymore elaborate than the development of a corresponding RIA usingtritiated tracer. In competitive EIAs (as in RIAs applying iodinatedtracers) a crucial factor is the type and position of the bridge usedduring the production of the immunogen in relation to the bridgeduring tracer synthesis [41,42]. Another specific source of problems inEIAs are factors altering the activity of marker enzymes. Neverthelessmany reliable and useful nonradioactive steroid immunoassays havebeen developed with a similar or even higher accuracy and precision incomparison to corresponding tritium-based RIAs [40,43].

2.8. Automated, direct immunoassays

A revolution especially in clinical steroid measurement came in the1980s with the emergence of commercially available kits used in au-tomated instruments enabling steroid measurements with a hithertounknown very high pace at relatively low costs. However, as thesetechniques are usually performed without preceding sample processing(direct assays), erroneous results can easily happen [44,45].

With the dominance of commercial automated immunoassays theoverall situation of steroid analysis has significantly changed. In the1970s and 1980s steroid immunoassays were essentially “home-brewed” methods, developed and thoroughly validated by researcherswho in many cases were also involved in clinical or experimental an-imal work. Nowadays in commercial or clinical laboratories steroids arejust a few parameters among many others, the personnel operating theautomated platform has generally no specific knowledge in methodo-logical aspects of steroid measurement, fully dependent on informationprovided by the package insert of the reagent kit and has no extensivecontact with patients or the scientific environment that generated thesamples. Because of this situation, there is usually no profound ex-pertise concerning the interpretation of the hormone concentrationsmeasured. Results are frequently issued to the sender together with areference range adopted from the literature and may have been estab-lished using a different method. For economic reasons determinationsare generally performed as single estimations. Thus, technical outlierscannot be identified. Double estimations are recommended.

2.9. Reliability of immunoassays

Recently a discussion came up about the value and reliability ofimmunoassays in todaýs measurement of steroid hormones [3]. Albeit

S.A. Wudy et al. Journal of Steroid Biochemistry and Molecular Biology 179 (2018) 88–103

91

basically important, this discussion is problematic as it results mainlyfrom the fact that immunoassays have been frequently pushed beyondtheir limits during the last years [44]. Applicants must be aware of theadvantages and pitfalls of their analytical methods. If properly appliedand validated, immunoassay can still be valuable tools in clinical andresearch laboratories. Checking their application with mass spectro-metric based methods is highly recommended (Fig. 3).

3. Steroid analysis by mass spectrometry

3.1. Introduction

3.1.1. General principle of mass spectrometryMS has developed into the perhaps most versatile and exact of all

analytical techniques permitting generation of qualitative as well asquantitative data on the atomic and molecular structure of inorganicand organic materials [46].

The principle of analysis by MS consists in generating multiple ionsfrom a sample, separating them in an electromagnetic field according totheir mass to charge (m/z) ratios, and finally, recording their relativeabundances [47].

Thus, a mass spectrometer consists of three elementary components,an ion source, a mass analyser, and a detector. The substances of in-terest are converted into ions, either positive or negative. MS separationis achieved by determining the mass-to-charge ratio (m/z) of the io-nized analyte. This process takes place in the mass analyzer. There areseveral types of analyzers, such as quadrupoles, ion trap analyzers,orbitrap analyzers or time-of-flight analyzers. As techniques other thanthe quadrupole are barely used [48–50] the focus of this review will lieon the latter. Quadrupoles are composed of four cylindrical rods be-tween which both electrostatic and radio-frequency fields can be ad-justed, allowing for the selection of ions with a particular m/z value.Other ions cannot reach the detector in the MS as their trajectorieswould not be stable in the quadrupole.

3.1.2. ChromatographyChromatography is the most relevant and most widely used se-

paration technique in the study of biological samples. Various chro-matographic separation methods are applied to research and clinicalstudies. The ultimate goal of chromatography is to achieve the bestpossible separation of the individual compounds of interest in a solu-tion. A chromatographic experiment is always based on the sameprinciple: the components of the mixture are separated over time ac-cording to their affinity for two immiscible phases. The sample is firstinjected or loaded onto a stationary phase. Later, a continuous flow of a

mobile phase is applied to the stationary phase and the analytes areseparated according to their chemical and physical properties and in-teraction with the two phases. Those molecules with a stronger inter-action with the stationary phase would be the last to elute, whereascompounds with weakest interactions would leave the stationary phaseearlier.

3.1.3. MetabolomicsChromatography provides only partial information about the nature

of the compounds one intends to analyze (polarity of the compound,specific retention time). Therefore, an additional identification tech-nique is required to achieve proper characterization of the constituentsin a mixture. Due to its versatility and specificity, MS is the leadingchoice for detection and quantification purposes. It allows for analysisof a wide range of compounds with excellent specificity and selectivity.

The combination of MS with separation methods such as GC or LCleads to so called “hyphenated techniques”. In these combinations, themass spectrometer serves as the detector. Whereas one immunoassayallows for the determination of a single steroid only, these hyphenatedanalytical techniques provide a further unique advantage: the si-multaneous determination of a multitude of analytes in a single run.This feature of multicomponent analysis permitted a renaissance inmetabolism research. It paved the way for the field of metabolomics,which is the systematic study of small molecule metabolites char-acterizing a biological sample. In case the instrument records all ions ofa particular mass range in a non-discriminatory, unbiased way, its modeof operation is called scanning or “untargeted”. In case the instrumentrecords only preselected ions, usually ions typical of particular analytes,the approach is a selected or “targeted” one.

3.1.4. Internal standardsFor quantitative analysis, suitable internal standards are a pre-

requisite. A big advantage of MS based analytical methods is thatradioactive labels are not needed. In general, an internal standardshould be chemically as close to the analyte as possible. Non-biologicalsteroids, e.g. epimers can serve as internal standards. However, stableisotope-labeled analogs of the analytes present nearly ideal internalstandards, since they have the advantage of showing practically thesame chemical and chromatographic properties. Furthermore, theyallow procedural losses to be disregarded and can easily be dis-tinguished from their unlabeled counterparts in the mass spectrometerby monitoring different ions. Labeling with deuterium is much easier toperform than labeling the carbon skeleton with 13C. However, deu-terium has to be incorporated at chemically stable positions that are notsubject to biologic attack. 13C-labeld steroids have the advantage of thestability of the label and the avoidance of isotope effects. With eithertechnique, high isotopic enrichment is important [51].

While stable isotope labeled internal standards had to be synthe-sized in the early days of mass spectrometric steroid analysis [52–54],most of them have become commercially available nowadays.

3.2. Steroid analysis by gas chromatography–mass spectrometry

3.2.1. Historical milestonesThe first separation of steroids by GC was achieved in 1960 [55]. 4

years later, the development of a robust interface permitted the com-bination of GC with MS [56]. In 1968, Horning published on the firstcomprehensive urinary steroid analysis by GC–MS [57]. Major ad-vances since Horning’s seminal work have been availability of strongderivatization reagents, C18 cartridges and introduction of capillarycolumns and robust mass spectrometers.

3.2.2. Steroid profile, targeted and untargeted analysesIt was Horning, who used the expression “steroid profile” for the first

time in his publication in 1968. It had the meaning of encompassing alldetectable urinary steroids in a non-selective way and somehow

Fig. 3. Long-term testosterone profiles in an adult boar comparatively obtained from anin house RIA [131] and LC–MS/MS [109]. Consistency between the two methods wasexcellent regarding the lower testosterone concentrations. Regarding high testosteroneconcentrations, the RIA showed a problem with linearity in the upper part of the mea-suring range. Consecutively, the upper limit of the measuring range of the RIA was re-duced to 5 ng/ml.

S.A. Wudy et al. Journal of Steroid Biochemistry and Molecular Biology 179 (2018) 88–103

92

Horning was anticipating what we nowadays term an untargeted me-tabolomics approach. Nowadays, the term steroid profile is used in asloppy, confusing, and inconsistent way. It should particularly not beused in case only a few steroids are being measured. In such a case, theword “panel” would be more appropriate. In our view, the term steroidprofile should be abandoned and replaced by more exact terms takinginto consideration the respective analytical approaches. Thus, a char-acterization of the steroid metabolome of a biological unit, e.g. an or-ganism or a cell culture, could either be targeted or untargeted (tar-geted or untargeted steroid metabolome analysis). Furthermore, each ofthese approaches could be either qualitative or quantitative, respec-tively.

3.2.3. Analytical principle of gas chromatography–mass spectrometryFor the analysis of steroids, GC still provides the best resolution of

all separation techniques. Usually, helium or hydrogen is used as mo-bile phase (carrier gas). Since the 1980s, capillary columns coated witha liquid stationary phase have been in use for analyzing steroids.Capillary columns are nowadays flexible, consisting of fused silica glasscovered with a protecting material (e.g. polyamide). The chemicalcharacteristics of the stationary phase and its film thickness, as well aslength and inner diameter of the column are responsible for the col-umn’s separation power. Most commonly, stationary phases are poly-siloxanes which can be modulated in their polarity by adding differentfunctional groups. As GC usually requires high temperatures, a pre-requisite is that analytes are vaporized without decomposition.Therefore, it is often necessary to chemically modify, i.e. derivatize theanalytes before injecting them in the GC. Various derivatization pro-cedures have been described [58]. Most common are silylation oracylation of the functional groups of the steroid molecules. Nowadays,mass spectrometric detection [59] is standard in GC based steroidanalysis (Fig. 4). Ionization is achieved by electron impact ionization(EI), a high energetic “hard ionization technique” resulting in numerousfragments of the molecular ion. When recording these fragments ac-cording to m/z ratios, highly specific mass spectra can be obtained.

Their fragmental pattern is characterized by an extremely high intra-and inter-instrumental reproducibility.

The MS can be operated in two different modes. In the full massrange “scanning mode”, the MS detects all ions generated in the ionsource from the molecules of the sample. The resulting chromatogramreflects the “total ion current” (TIC). At each time point of the chro-matogram, all recorded m/z values are available and thus full massspectra are obtainable (Fig. 5). In this non-selective (untargeted) steroidmetabolomics approach identification of unknowns is possible bycomparing obtained mass spectra with spectral libraries. In the “se-lected ion monitoring” (SIM) mode, the MS is focused on detecting onlyone or multiple specific fragment ions representative of a certain ana-lyte. This results in a higher degree of sensitivity. This targeted steroidmetabolomics approach is ideally suited for quantification.

3.2.4. Essentials of method validationIn analytical method development validation is a crucial step to

demonstrate that reliable results are generated for the performed ana-lysis. Therefore, method validation should include at least determina-tion of accuracy, precision, specificity, sensitivity, reproducibility andstability. Several guidelines exist for method validation [60,61].

3.2.5. GC–MS in steroid analysisVarious groups of steroids such as glucocorticoids, gestagens, an-

drogens, estrogens and sterols can be measured with GC–MS methods.Abnormalities in steroid biosynthesis and excretion can be found byanalyzing these steroid hormones in different biological materials likeurine and blood with the findings of either enzyme deficiencies or al-tered steroid hormone excretions in patients with complex diseasessuch as obesity or cancer. A comprehensive overview dealing with thedifferent aspects of sample preparation as well as the clinical applica-tions for GC–MS steroid analysis can be found in Choi et al. [62]. In thefollowing two sections, typical applications of GC–MS to steroid ana-lysis in the Giessen Steroid Research &Mass Spectrometry laboratorywill be described.

Fig. 4. Schematic figure for the principle of GC–MS.After sample preparation, an aliquot of the derivatized sample is injected in the GC. Using Helium as carrier gas the sample mixture is separated into single compounds by passing throughthe GC column. Afterwards all compounds are ionized in the ion source of the MS, separated and detected. The resulting chromatogram shows the peaks with the mass spectra for eachcompound.

S.A. Wudy et al. Journal of Steroid Biochemistry and Molecular Biology 179 (2018) 88–103

93

3.2.6. Application: urinary steroid metabolome analysisThe mainstay of GC–MS steroid analysis has been urinary steroid

metabolome analysis. Urinary steroid analysis was the original form ofsteroid disorder diagnosis as analytical sensitivity in the 1950s wasinsufficient for serum analysis. Most disorders in steroid metabolismhave been first characterized by analysis of urine. The technique hasbeen further developed and refined especially by Shackleton [63].Multiple steroid metabolites (Table 1) can be simultaneously de-termined allowing for the non invasive diagnosis of practically allsteroid related disorders [64–66].

The non-selective measurement of urinary metabolites excretedover 24-h provides the most integrated picture of the steroid hormonalstatus of a patient. When the steroid concentrations are analyzed in a24-h urine sample the resulting steroid excretion rates allow to estimatethe integrated output of adrenocortical and gonadal steroid production[67]. Thus, hormonal production rates or compliance of hydrocortisonetherapy can be assessed [65]. GC–MS data is usually interpreted bycomparing concentrations or excretion rates of metabolites with re-ference values. Precursor to product ratios of metabolites can be used toassess enzyme activities. Since quantitative datasets of multiple com-ponents are very difficult to present visually, unbiased systems biologyapproaches are applied nowadays. For instance, the metabolic patternof patients with 17α-hydroxylase/17,20-lyase deficiency was char-acterized using a novel application of a mathematical approach adaptedfrom gene expression data analysis [68]. Concerning non-Mendeliancomplex diseases, such as non syndromic obesity, clustering of similarsteroidal fingerprints were found to present specific steroid metabo-lomic disease signatures and lead to reclassification of patients to sev-eral groups [69].

The method used at the Giessen steroid &mass spectrometry unit isoriginally based on the one used by Shackleton [63,67]. It not onlyserves for clinical analyses but is also used for research purposes tocharacterize disease specific steroid metabolomes and identify steroidsin biological fluids, e.g. cell culture supernatants or tissue samples[70,71].

In brief, urine (typically 5 ml) is extracted by solid phase extraction,dried, reconstituted and hydrolyzed with Helix pomatia enzyme. Theresulting free steroids are extracted once more (solid phase extraction).After addition of internal standard, the sample is derivatized to formmethyloxime-trimethylsilyl ethers. Samples are analyzed on an AgilentTechnologies 6890 series GC that is directly interfaced to an Agilent

Technologies 5975 inert XL mass selective detector. The derivatizedsamples are analyzed during a temperature programmed run between210 and 270 °C on an Optima-1 MS fused silica column (length, 25 m;film thickness, 0.1 μm; inner diameter, 0.2 mm; Macherey-Nagel).Helium is used as carrier gas. For all urinary steroids measured, intra-assay precision varies between 1,7% (for 17b-Adiol) and 9,5% (for 20α-DHF) and interassay precision ranges between 1,1% (for a-cortol) and9,5% (for 11-OH-An) [72].

For getting all the information on the steroid metabolome of asample, the scanning mode of the mass spectrometer is used [73–75].On the other hand, the selected ion monitoring mode provides a moreselective approach with much higher sensitivity for the analytes and istherefore used for quantification. Two characteristic ions (quantifierand qualifier ions) are measured per analyte. The choice of these ionsdepends on no other compound being present in the sample whichwould give a response for these ions at the same retention time. Fig. 2shows the chromatogram (TIC, total ion current) as well as an examplefor a mass spectrum of the methyloxime-trimethylsilylether derivativesof a standard solution of urinary steroids.

3.2.7. Application: stable isotope dilution GC–MS analysis of serum steroidsIsotope dilution GC–MS (ID/GC–MS) is used for quantitative ana-

lysis of steroid hormones in plasma whenever highest specificity isneeded. A cocktail of stable isotope labeled analogs of the analytes isadded to the sample. Since the early nineties the team of the GiessenSteroid Research &Mass Spectrometry laboratory has been developinga constantly expanding ID/GC–MS method for steroid analysis in var-ious biological fluids and tissues. We are currently able to simulta-neously analyze 11 steroids including cortisol, 11-deoxycortisol, 17-hydroxyprogesterone, 17‐hydroxypregnenolone, corticosterone, pro-gesterone, testosterone, 4‐androstenedione, dehydroepiandrosterone,dihydrotestosterone and 5α-androstane-3α,17β-diol.

Our method for plasma steroid analysis consists of incubation andequilibration of plasma (typically 0.5 ml) with a mixture of the deut-erated internal standards. This step is then followed by solvent ex-traction, clean up by Sephadex LH-20 gel chromatography, and deri-vatization (perfluoroacylation). An aliquot of the derivatised extract issubjected to analysis by GC–MS using SIM [59]. Intra-assay precisionvaried between 0,8% for progesterone and 6,8% for 4-androstenedione.Interassay precision ranged between 0,5% for progesterone and 9,1%for 4-androstenedione. In Fig. 6, typical selected ion chromatograms of

Fig. 5. TIC (upper panel) and example of mass spectrum (lower panel, THE) of the methyloxime-trimethylsilylether derivatives of a standard solution of urinary steroids.

S.A. Wudy et al. Journal of Steroid Biochemistry and Molecular Biology 179 (2018) 88–103

94

the steroids present in a real plasma sample are shown.This method is also used as a reference technique for quality control

of immunoassays. In an earlier investigation we were among the first toshow that various immunoassay techniques can severely overestimateconcentrations of 17-hydroxyprogesterone [25]. The method has fur-ther been applied to the determination of steroids in various speciessuch as mice. Thus we established first reference values for variousmarker steroids of adrenal and gonadal origin in mice [76].

3.2.8. Outlook: potential role of gas chromatography–tandem massspectrometry

In contrast to GC–MS, GC–tandem mass spectrometry (GC–MS/MS)using a triple quadrupole is a relatively new technique that has not beenwidely used yet. A future role of GC–MS/MS might lie in its potential tostill discover and characterize novel steroids and steroid biomarkers, ashas for instance been shown in conditions such as 21-hydroxyalasedeficiency [77].

3.3. Steroid analysis by liquid chromatography–mass spectrometry

3.3.1. Historical milestonesThe development of so-called “soft” ionization techniques and the

successful coupling of LC with MS presented further breakthroughs inthe field of mass spectrometric analysis. Thermospray LC–MS was thefirst LC–MS technique to be applied to the analysis of steroids and itproved particularly suitable for the analysis of intact complex steroids,such as sulfated steroids [26]. Substantial further progress was madewith the introduction of MS/MS, especially the triple quadrupole. Thistechnique compensates for the rather low chromatographic capacity ofLC. All these developments have contributed to LC–MS having reallycome of age [78] in the field of mass spectrometric steroid analysis.

3.3.2. Liquid chromatographyLC includes all those techniques in which the mobile phase is a li-

quid. The majority of the liquid chromatography methods make use ofpacked columns. They are easy to work with, they do not need pre-paration of the stationary phase as they are commercially available, andthey allow good separations. Among these, the most common columnsin research laboratories are HPLC columns (High Performance LiquidChromatography). This technique applies high flows to perform quickanalysis. Nowadays faster approaches are available, including UltrahighPressure Liquid Chromatography (UPLC), which requires a specific in-strument. Fused-core technologies work with the same apparatus asHPLC [79].

Reversed-phase chemistry prevails among the different availablestationary phases [58]. In the reversed phase material, a silica gel actsas support for the structures that will interact with the analysis. Thosestructures are covalently bonded alkyl chains or phenyl structures. Thealkyl chains typically range from 4 to 30 carbons, although 18 carbonsare most common (C18). The term “reversed-phase” was adopted his-torically in contraposition to the chemical nature of the first materialsapplied to chromatographic separations (normal phase or polar phase).In case of reversed-phase stationary phases, lipophilic analytes interactlonger with the phase. More, polar compounds do not interact thatintensely and leave the column earlier. In order to elute lipophiliccompounds, an adequate mobile phase has to be used, capable of sur-passing the affinity of the analyte for the stationary phase.

More recently, the use of normal phased HILIC columns(Hydrophilic Interaction Liquid Chromatography) has gained re-levance, especially in the field of metabolome research [80]. Thischromatography permits the identification of very polar compoundswhich otherwise would elute too early from reversed-phased columns[81].

3.3.3. Liquid chromatography–mass spectrometry, “soft ionization”Unlike GC–MS, in which the compounds usually require ionization,

most compounds analyzed by LC–MS are already charged ions in theliquid phase. However, the ions in the liquid phase must be isolatedfrom the solvent to form gas-phase ions. Only these are amenable to beanalyzed by MS. The coupling between LC and the MS is the mostchallenging restriction of the technique, as not every solvent or everyanalyte are optimal for the formation of gas-phase ions. This processtakes place in the ion source. Typical ion sources include electrosprayionization (ESI), atmospheric pressure chemical ionization (APCI), andatmospheric pressure photoionization (APPI). Among these, the mostcommon are ESI and APCI. ESI is widely applied to the study of polarcompounds and APCI is preferred to analyze compounds which are noteasily ionized. Both ESI and APCI are non-disintegrating, “soft” ioni-zation techniques, producing only a low degree of fragmentation(Fig. 7). Commonly used solvents in LC–MS are methanol and acet-onitrile due to their good miscibility with water, the good solubility of awide range of compounds, and their volatility.

Table 1Origin of major urinary steroid metabolites excreted in children and adults, simulta-neously determined by GC–MS urinary steroid metabolome analysis. Steroids are ar-ranged according to the number of C atoms.

Abbreviation Urinary steroid metabolite Origin of urinary steroid

C18 steroids estrogensE1 estrone estrogensE2 estradiolE3 estriol

C19 steroids androgensT testosterone testosteroneAn androsterone DHEA, androstenedione,

testosteroneEt etiocholanaloneDHEA dehydroepiandrosterone DHEA-Sulfate16OH-DHEA 16α-hydroxy-DHEAA5T-16α 5-androstene-3ß,16α,17ß-triolA5-3ß,17ß 5-androstene-3ß,17ß-diol DHEAA5-3α,17ß 5-androstene-3α,17ß-diol11-O-An 11-oxo-androsterone cortisol, 11-hydroxy-

androstenedione11-OH-An 11-hydroxy-androsterone11-OH-Et 11-hydroxy-etiocholanolone

C21 steroids progestagensPD pregnanediol progesteronePT pregnanetriol 17-hydroxyprogesterone17-OH-Po 17α-hydroxypregnanolonePdiol 17α-hydroxyallopregnanoloneP5D pregnenediol pregnenoloneP5T 5-pregnene-3ß,17α,20α-triol 17-hydroxypregnenolone11-O-PT 11-oxo-pregnanetriol 21-deoxycortisolTHS tetrahydro-11-deoxycortisol 11-deoxycortisol

C21 steroids glucocorticoidsF cortisol cortisolTHF tetrahydrocortisol5α-THF 5α-tetrahydrocortisolα-C α-cortolß-C ß-cortol6ß-OH-F 6ß-hydroxycortisol20α-DHF 20α-dihydrocortisolTHE tetrahydrocortisone cortisoneα-Cl α-cortoloneß-Cl ß-cortolone

C21 steroids mineralocorticoidsTHA tetrahydro-11-dehydro-

corticosteronecorticosterone

THB tetrahydrocorticosterone5α-THB 5α-tetrahydrocorticosteroneTH-DOC tetrahydro-11-

deoxycorticosterone11-deoxycorticosterone

S.A. Wudy et al. Journal of Steroid Biochemistry and Molecular Biology 179 (2018) 88–103

95

3.3.4. Liquid chromatography–tandem mass spectrometryFor MS/MS the most usual analyzer configuration is the triple

quadrupole (3Q) (Fig. 8). In the first quadrupole (Q1), the ion or ionswhich are going to be measured are selected. All other ions are

excluded from the system. A second quadrupole (Q2) acts as a collisioncell. A collision gas, generally argon, is responsible for fragmentation ofthe molecules. In the third quadrupole (Q3) a specific fragment orfragments from the compound of interest obtained in Q2 are selected

Fig. 6. Chromatograms of the steroid analysis in aplasma sample.Corresponding ion traces of analyte and deuterated in-ternal standards (deuterium = d) were superimposed.The following steroids were measured: testosterone (T)m/z 680.4/683.4 for T/d3-T, 4‐androstenedione (4A) m/z 482.3/484.3 for 4A/d2-4A, 5α-androstane-3α,17β-diol(AD) m/z 470.3/473.3 for AD/d3-AD, m/z 270.2/272.2for DHEA/d2-DHEA, dihydrotestosterone (DHT) m/z414.3/417.3 for DHT/d3-DHT, 17-hydroxyprogesterone(17OHP) m/z 465.4/469.4 for 17OHP/d4-17OHP,17‐hydroxypregnenolone (17PE) m/z 467.4/471.4 for17PE/d4-17PE, 11-deoxycortisol (S) m/z 465.2/467.2 forS/d2-S, cortisol (F) m/z 489.3/491.3 for F/d2-F, proges-terone (Prog) m/z 510.3/518.3 for Prog/d8-Prog andcorticosterone (B) 720.4/726.4 for B/d6-B.

S.A. Wudy et al. Journal of Steroid Biochemistry and Molecular Biology 179 (2018) 88–103

96

and measured. This operating mode is known as selected or single re-action monitoring (SRM). In case of multiple analytes it is common touse the term multiple reaction monitoring (MRM). The line of frag-mentation, “transitions”, is usually very specific for each compound.

3Q configuration provides better performance as each step of thefragmentation experiment takes place in a different analyzer. In con-trast, in an ion trap, the whole process takes place in a single analyzer.

Although MRM is the normal operation mode in tandem massspectrometry (TMS), triple quadrupoles are versatile and several otherworking modes are possible. Targeted-mode quantifications can becarried out when the compound of interest is known and its chemicalstandard is available. In TIC scan mode, all compounds within a certainm/z range are allowed to reach the detector, e.g. between m/z100–1000 for steroids. In this experiment only one quadrupole is re-quired. The signal for the compounds of interest can be studied ex-tracting the information for specific m/z values (extracted ion

chromatograms). In the SIM mode only several typical m/z valuescorresponding to the analytes are recorded. 3Q configuration can pro-vide a lot of useful information in experiments devoted to find newmetabolites. In precursor ion mode only the m/z from compounds witha specific fragmentation product is selected. Product ion mode providesthe fragmentation pattern from a specific m/z. In neutral loss scan, them/z of compounds with a specific neutral loss is selected.

3.3.5. Essentials of method validationAs with all other bioanalytical methods, validation is a pre-requisite

prior to its application to laboratory routine use. This is especially truein case of clinical analyses. The Food and Drug Administration (FDA,USA) and many other agencies provide guidelines for bioanalyticalmethod validation [82]. The validation process includes the assessmentof several parameters. Most important are sensitivity, precision, re-producibility, accuracy, limits of quantification and detection,

Fig. 7. Mass spectrum of DHEAS as detected in Q1.The base peak of the spectrum at m/z 367 ressemblesthe molecular ion after abstraction of a proton[M−H]−. The LC–MS/MS was run in electrospraynegative ionization mode.

Fig. 8. Schematic principle of the LC–MS/MS.After sample preparation, an aliquot of the sample is injected in the LC. After chromatographic separation the analytes are ionized in the ion source of the MS, separated and detected. Theresulting chromatogram shows the peak with the MS/MS spectrum for the compound.

S.A. Wudy et al. Journal of Steroid Biochemistry and Molecular Biology 179 (2018) 88–103

97

Fig. 9. Merged chromatograms of sulfated steroids in higher concentrations in human serum [110]. Upper panel depicts the merged chromatograms of a healthy male. Panel on thebottom shows the chromatograms for a mix of standards (100 ng/ml).

S.A. Wudy et al. Journal of Steroid Biochemistry and Molecular Biology 179 (2018) 88–103

98

recovery, stability, carryover and matrix effects [83–85]. InternationalOrganization for Standardization (ISO) standard 17511 [86] and Clin-ical and Laboratory Standards Institute (CLSI) EP32-R [87] containrecommendations on how to implement calibrator traceability.

3.3.6. Matrix effects in LC–MSOf special importance is to study matrix effects in LC–MS experi-

ments [88]. This is due to the particular nature of the LC–MS hy-phenation. As explained above, an interface between the LC system andthe MS is needed, because the compounds are dissolved in a liquid andneed to be transformed into the gas phase to be detected by MS. Whenthis process takes place, other compounds stemming from the biologicalmatrix will come along with the compound(s) of interest. The presenceof these compounds can affect the ionization process in some cases. Thisbecomes evident, when the results are compared with those obtainedfrom the quantification of the analyte dissolved in a pure solvent so-lution. Matrix effects can increase the ionization of the analyte (ionenhancement) or, more commonly, decrease its ionization efficiency(ion suppression). Post column infusion experiments allow for assess-ment of matrix effects. The use of stable isotopic labelled internalstandards helps to overcome matrix effects. Proper sample preparationis a critical step in method development, because it might provide asignificant reduction of matrix effects too. Matrix effects may be un-avoidable in some cases after sample cleanup and/or the use of internalstandard. In such a case, calibration curves must be done using the samematrix or a proper matrix surrogate (i.e. charcoal-treated matrix) [89].

3.3.7. LC–MS in steroid analysisMeanwhile, LC–MS methods have been developed for nearly all

classes of steroids. Methods are now available for all clinically relevantclassical steroid hormones in adults comprising glucocorticoids, pro-gestines, androgens, estrogens and mineralocorticoids [90–96], as wellas in children [97–99]. Furthermore, methods have been developed forvitamin D, a seco steroid, and its metabolites [100].

Oxysterols are increasingly recognised as important players inatherosclerosis, cancer or neurodegenerative disorders [101]. LC–MScan assist in detection and characterization of new steroidal biomarkers[102]. Furthermore, the technique is of invaluable help in dopinganalysis [103] and suitable tool for analyzing bile acids [104,105].While most common matrices are biological fluids such as serum/plasma or urine, LC–MS is also suitable in determining steroids in ratherunusual matrices such as hair [106], saliva [107] or follicular fluid[108].

As has already been pointed out before, LC–MS is the technique ofchoice for determining conjugated steroids [26], since the techniqueallows for the analysis of the intact molecule [109–113].

3.3.8. Application: LC–MS analysis of sulfated steroidsThe analysis of sulfated steroids has been the focus of the DFG re-

search group 1369, “Sulfated Steroids in Reproduction” in which allauthors participated. Since LC–MS is the method of choice for mea-suring the intact conjugated steroid, a method for measuring the bio-logically most important steroid sulfates has been developed. We couldachieve the development of the most comprehensive LC–MS profile forsteroid sulfates which can be characterized as follows:

We developed a new method for the quantification of up to 11sulfated steroids in human serum [110]. This method was preceded byanother which allowed for the quantification of sulfated steroids andunconjugated compounds in several biological fluids [109].

300 μl of serum (or plasma) and the calibrators are incubated with amix of several internal standards. Then, protein is precipitated with1 ml of ACN-ZnSO4 [89 g/l, 4:1 (v/v)] and the supernatant isolatedafter centrifugation. The supernatant is mixed with 3 ml of water in aglass tube, and transferred onto a conditioned SepPak C18 cartridge.The cartridge is washed first with 3 ml of water, followed by 3 ml ofn‐hexane. Next wash is with 4 ml of chloroform, and the final step wash

is with 4 ml of methanol. Unconjugated steroids can be collected fromthe chloroform fraction, and sulfated steroids are eluted with methanol.The methanolic fraction is then evaporated with nitrogen at 40 °C andthen reconstituted in 250 μl of a solution containing 79.75% water,10% MeOH, 10% ACN, and 0.25% ammonium hydroxide. After cen-trifugation, 10 μl are injected in the LC–MS/MS. Examples of chroma-tograms of sulfated steroids in human serum are given in Fig. 9.

These methods have been pivotal for several projects of the DFGgroup “Sulfated steroids in reproduction”. Recently, we studied thevariation of sulfated steroids with age in both control group and inpatients with steroid sulfatase deficiency (STSD) [114]. Most sulfatedsteroids were increased in STSD, but levels of androsterone sulfate,were decreased. Cholesterol sulfate was found to increase with agingunlike the rest of sulfated steroids, in both groups. Concentrations ofsulfated and unconjugated steroids were calculated in several otherprojects. These included the characterization of the transport of sulfatedsteroids by SOAT (Sodium-dependent Organic Anion Transporter)transporter [115], the profile of steroids in post-pubertal boars [116] orthe study of the transport of the placental estriol precursor 16α-hy-droxy-dehydroepiandrosterone-3-sulfate [117].

4. Conclusions and recommendations

4.1. Challenges of steroid analysis

The reliable determination of steroids is anything but trivial. First, ithas to be realized that steroid metabolism is complex per se and there isa multitude of different steroid precursors, steroid hormones or steroidmetabolites. Some of them have been attributed diagnostic potential,some not, and not all relevant steroids have been identified yet.

For instance, sulfated steroids have hitherto been regarded as water-soluble end products to be excreted via the kidneys. However, with thediscovery of specific transport channels for steroid sulfates, such asSOAT (sodium dependent anion transporter), this view has changed andit has been the intention of this DFG research Group 1369 “SulfatedSteroids in Reproduction” to characterize and unravel new biologicalfunctions of this class of steroids.

Steroid metabolism in a living creature is not static at all. For in-stance, in the human being, various steroidogenic phases and milieuscan be discerned, e.g. fetal and neonatal period, puberty, menopause,etc. Consecutively, reference ranges vary according to sex, age anddevelopmental stage. Some steroids can be present in very high con-centrations, whereas others are to be found only in trace amounts.

Moreover, steroids are not easy to measure. Steroids are small mo-lecules. They differ only in subtle structural changes, a circumstancewhich makes them not easily amenable to analysis but, regarding theirbiological activity, can lead to enormous functional differences.

4.2. Which technique to choose for steroid analysis?

While hitherto steroid analysis by immunoassays has been theprevailing analytical method, the advent of MS based analytical tech-niques has given rise to an ongoing debate on the correct choice of theanalytical method. Especially LC–MS seems to have swamped themarket and field of steroid analysis. For non-specialists these develop-ments might be highly confusing, thus, a differentiated view on whichanalytical technique to properly use for the analysis of steroids is des-perately needed. It needs to be envisioned that reliable steroid analysisin medicine as well as in research is a “sine qua non” regarding lifelongimplications for the patient or validity of scientific data.

4.3. Pros and cons of immunoassay

For long, immunoassays have been the mainstay of steroid analysis.The reliability of immunoassays depends primarily on the specificity ofthe antibody and the presence of potentially cross reactive interferants.

S.A. Wudy et al. Journal of Steroid Biochemistry and Molecular Biology 179 (2018) 88–103

99

Especially with the techniques in the eighties and nineties of last cen-tury, relative reliable results could be obtained. This was due to thefacts that extraction or chromatographic steps were used to purifysamples and researchers were fully acquainted with the principle andperformance of these methods. Of course, only one analyte can bemeasured per immunoassay. However, determination of unknownsteroids by immunoassay is not possible and no assays are available forunusual or exotic steroids.

With the advent of new commercially available automated directimmunoassays and the centralisation of hormone assays in remotecentral laboratory platforms, severe deterioration in steroid assayquality resulted. Due to their non-specificity the new assays are proneto inaccuracy. Furthermore, a loss in expertise regarding steroid assaydevelopment, selection and interpretation can be noted.

4.4. Pros and cons of mass spectrometry based steroid determination

The fact that MS based steroid analysis currently provides thehighest specificity consists in the fact that MS provides exact structuralinformation on the analyte by recording its typical ions. If combinedwith a chromatographic technique, the retention time of a particularanalyte presents a further specific criterion of identification.Furthermore, these approaches allow determination of a multitude ofanalytes in a single run (“steroid profile”). Stable isotope labelled in-ternal standards present practically ideal internal standards and allowfor exactquantification.

Though progress has been made in the development of robust benchtop instruments and companies started to market kits for steroid de-termination, access to mass spectrometers is limited, their operationneeds skilled personnel and methods are not easy to set up.

4.5. Immunoassays versus mass spectrometry

There is no need to condemn immunoassays in general. It needs tobe respected that not all clinicians and researchers have access to MSbased technologies. It might be advantageous that immunoassays donot need highly trained operators and that they are relatively easy to setup and to perform. Immunoassays should only be used for purposes andanalytical situations they have been developed for. They must not bepushed beyond their limits. The immunoassay cannot use internalstandards. In case of uncertain results or analytical situations, it ishighly recommended to have the performance of immunoassayschecked by MS based techniques. In difficult analytical situations, e.g.the neonatal period, we could show that immunoassays can severelyoverestimate plasma steroid concentrations, e.g. of 17-hydro-xyprogesterone [25].

4.6. Pros and cons of steroid analysis by gas chromatography-massspectrometry

GC has the greatest potential for separation of steroids and is thusexcellently suited for the simultaneous detection of a multitude ofsteroid metabolites. Its combination with MS gave rise to the extremelyrobust hyphenated technique of GC–MS. The higher degree of frag-mentation with electron impact ionisation and the extremely high sta-bility of fragmentation patterns are best prerequisites for providinghighest specificity in steroid analysis, allowing for obtaining qualitativeand quantitative data, and making the technique highly suited as re-ference technology (gold standard) and quality control for im-munoassays and LC–MS. Furthermore, the database of steroid frag-mentations in GC–MS is unexcelled and will be an indispensableresource for steroid identification. Thus, GC–MS will also present thegold standard in the characterization of steroid metabolomes: its per-sistent significance consists in its property being the most powerful toolfor diagnosing and defining new disorders of steroid biosynthesis aswell as for characterizing metabolome changes, e.g. in the evaluation of

new drug therapies [5,73].However, due to the prerequisite of a derivatization step, sample

workup is more elaborate and time consuming. Compared with LC–MS,GC–MS run times are longer, not allowing for being used as a highthroughput method. Complex, e.g. conjugated steroids cannot be ana-lyzed as intact molecules but need to be hydrolyzed. In general, GC–MSsteroid analysis needs extremely high analytical and biochemical spe-cialist knowledge available in just a few highly specialized suprar-egional laboratories.

4.7. Pros and cons of steroid analysis by LC–MS/MS

The importance of LC–MS/MS in steroid analysis has rapidly in-creased during the last decade [118]. The first reason for this are theshort instrumental run times. Secondly, sample preparation is easy andshort. Little wonder that the resulting fast turnaround times make themethod perfectly suited as a high throughput method. And thirdly, thetechnique is capable of analyzing intact complex molecules such assulfated or glucuronidated steroids [109,119] which makes it themethod of choice for determining such compounds, e.g. as has been thecase in the DFG Research Group 1369 “Sulfated Steroids in Reproduc-tion”. Last but not least, the system permits a high degree of automationwith the use of online sample preparation approaches [120,121].

Despite all this, the technique requires solid knowledge of both li-quid chromatography and MS, and proper method validation is man-datory [122]. In some cases it is still necessary to derivatize the ster-oids, e.g. in cases certain matrices lead to poor ionization or steroids arepresent in very low concentrations, e.g. oxysterols [123] or estrogens[95]. A further disadvantage of LC–MS/MS lies in the limited avail-ability of stable isotope labelled internal standards for complex steroids.

4.8. GC–MS and LC–MS are complementary techniques in steroid analysis

The fact that LC–MS can be used as a high throughput method hasled to a constant move towards LC–MS for steroid analysis and a certain“hype” during the last few years. This development goes hand in handand fits to the change in temporal structures of modernity, which isacceleration [124]. Consecutively commercial companies spotted amarket for selling instruments and even methods (kits) for steroidhormone determination and an increasing number of laboratories haveswitched and will switch to this technique. As has already been statedabove, the technique requires solid knowledge in analytics and bio-chemistry and it may be severely doubted, whether it is reasonable toset up this technique at every corner.

The almost exponential increase in publications concerning steroidanalysis by LC–MS and the tenor of many of these articles conveys theimpression that LC–MS might replace GC–MS in steroid analysis.However, this perception is not applicable! LC–MS has clearly its ad-vantages in the rapid determination of unconjugated and conjugatedsteroids by targeted approaches, but it must simultaneously be realizedthat its specificity is severely hampered by the following factors: firstly,chromatographic resolution of LC is much lower than that of GC,especially for relatively similar analytes such as steroid isomers andepimers. Chromatographic separation will even be further hampered incase chromatographic run times are tried to be kept short. Secondly,soft ionization leads to a low rate of fragmentation. As many steroidshave identical or rather similar molecular weights and structure, theirfragmentation patterns share many similarities and their determinationis not possible unless clear chromatographic separation of isobars isachieved. Thirdly, some classes of steroids, e.g. ring saturated steroidsor steroids bearing a double bond between C5 and C6 are difficult toionize either by ESI or APCI. Fourthly, soft ionization – in contrast toelectron impact ionization – is highly susceptible to matrix effects. Allthese factors can lead to a severe reduction in specificity of LC–MS andmakes the technique far less suited for untargeted steroid determina-tion. In contrast, though a more elaborate and time consuming

S.A. Wudy et al. Journal of Steroid Biochemistry and Molecular Biology 179 (2018) 88–103

100

procedure – GC–MS allows for determining steroids by both targetedand untargeted approaches with highest specificity and will be un-rivalled in the characterization of steroid metabolomes. Therefore,GC–MS and LC–MS are in fact complementary techniques.

4.9. Reliability of stable isotope labeled internal standards

This important aspect has not been given enough attention in thecurrent analytical literature! Recently attention has been drawn to thephenomenon that the same LC–MS assay for serum testosterone yieldeddifferent results when various stable isotope labelled internal standardswere applied [54]. As a consequence stable isotope labelled internalstandards have to be chosen carefully. An important criterion is stabi-lity. Most standards used are labelled with deuterium because this is themost easy and inexpensive way to synthesise them allowing for highisotopic enrichment [51]. However, a huge drawback in the use ofdeuterium labelled internal standards is their susceptibility to loss oflabel due to chemical reactions during the analytical procedure (e.g.changes in pH). The easier the label is introduced the easier it is lost!E.g. deuterium labels adjacent to carbonyl groups are easily introducedby exchange reactions but are susceptible to back exchange unless thecarbonyl group is modified or removed. 13C-labeled steroids are verystable, however difficult to synthesize and yet available for only asmaller number of steroids. To avoid isotopic overlap with naturallyoccurring steroids there is a consensus that the number of deuteriumatoms to be introduced in the molecule should at least be two, pre-ferably three. Too many labels may have disadvantages, e.g. alteredchromatographic behaviour [125].

4.10. Harmonisation

Immunoassay as well as mass spectrometric methods are nowadayscommon for measuring steroid hormones in clinical and research la-boratories. In the future it can be expected that for the sake of higherspecificity, immunoassays might be more and more replaced by MSbased methods. However, substantial interlaboratory variability [126]has been described even among LC–MS steroid assays. Most of the ex-isting LC–MS methods are in house methods and it is a matter of con-cern that up to now no comprehensive programs for harmonisation andstandardization have been available. In this context, the authors re-commend caution on automatically putting any LC–MS based methodon the level of a gold standard.

Harmonisation of the different analytical methods is required toachieve the best outcomes for steroid analysis in clinic and research.The pillars of harmonisation are 1) external quality assurance, 2)availability of certified reference materials, 3) reference measurementprocedures, and 4) reference laboratories, reference intervals and de-cision points [53]. The term commutability refers to the ability ofcertified reference material to show interassay accuracies comparableto those obtained in the measurement of the same analyte in a biolo-gical sample [127]. The Asia Pacific Federation for Clinical Biochem-istry and Laboratory Medicine (APFCB) mass spectrometry harmoni-sation project on testosterone measured by LC–MS/MS presented a firstworldwide attempt and showed an impressive reduction in method bias[53]. Concerning the first pillar, a great step forward has been the in-troduction of the first peer comparison of 5α-dihydrotestosterone(DHT) measured by MS and immunoassay laboratories [128]. An iso-tope dilution GC–MS/MS method was used for target setting of the DHTvalues. DHT is one of the most important analytes in diagnosing dis-orders of sex development and it is difficult to measure both by im-munoassay and LC–MS. The Joint Committee for Traceability in La-boratory Medicine (JCTLM) promotes worldwide standardization ofclinical laboratory tests and provides information on reference mate-rials and reference methods [129].

4.11. Final recommendations

As steroid analysis is complex, it is essential that clinicians and re-searchers know the pitfalls and limitations of the methods, which theyare going to use by heart. In case of doubt, it is always good to contactthe laboratory in advance for consulting and planning which approachto choose. Sophisticated analytical questions should be handled incollaboration with highly specialized supraregional centers of expertiseproviding the required knowledge in analytics and steroid biochem-istry. Only highly specialized and experienced analysts will have “tacitknowledge” at their disposal, knowledge, which might be essential butcannot be passed on by written or verbal communication [130] makingsteroid analysis a real art.

It is likely that mass spectrometric based steroid determination willprevail in future. Good and reliable steroid analytics is worth its priceand especially in our times becoming increasingly dominated bydumping prices, it has to be kept in mind that quality is not equivalentto luxury. It should remain a matter of course, that both clinics andscience are based on reliable analytics.

Acknowledgements

This work was supported by the German Research Foundation(DFG) within DFG Research Group 1369 “Sulfated Steroids inReproduction”, subprojects 4 (G. Schuler, principal investigator, SCHU1195/4-2) and 7 (S. A. Wudy, principal investigator, WU 148/6-2).

References

[1] Merriam-Webster’s Dictionary, https://www.merriam-webster.com/.[2] M.W. Duncan, Good mass spectrometry and its place in good science, J. Mass

Spectrom. 47 (2012) 795–809.[3] D.J. Handelsman, L. Wartofsky, Requirement for mass spectrometry sex steroid

assays in the Journal of Clinical Endocrinology and Metabolism, J. Clin.Endocrinol. Metab. 98 (2013) 3971–3973.

[4] R.F. Greaves, G. Jevalikar, J.K. Hewitt, M.R. Zacharin, A guide to understandingthe steroid pathway: new insights and diagnostic implications, Clin. Biochem. 47(2014) 5–15.

[5] C. Kamrath, S.A. Wudy, N. Krone, Steroid biochemistry, Endocr. Dev. 27 (2014)41–52.

[6] G. Moss, Nomenclature of steroids (Recommendations 1989), Pure Appl. Chem. 61(1989) 1783–1822.

[7] ChemSpider, http://www.chemspider.com/.[8] I.E. Bush, The Chromatography of Steroids, Pergamon Press, Oxford (UK), 1961.[9] R.S. Yalow, S.A. Berson, Assay of plasma insulin in human subjects by im-

munological methods, Nature 184 (Suppl. 21) (1959) 1648–1649.[10] R.S. Yalow, Radioimmunoassay: a probe for the fine structure of biologic systems,

Science 200 (1978) 1236–1245.[11] S. Glick, Rosalyn sussman yalow (1921–2011), Nature 474 (2011) 580.[12] C.S. Corker, D. Exley, F. Naftolin, Assay of 17beta-oestradiol by competitive pro-

tein binding methods, Acta Endocrinol. Suppl. (Copenh.) 147 (1970) 305–319.[13] M.L. De Souza, H.O. Williamson, L.O. Moody, E. Diczfalusy, Further assessment of

the reliability of progesterone assays by competitive protein binding, ActaEndocrinol. Suppl. (Copenh.) 147 (1970) 171–187.

[14] C.A. Nugent, D. Mayes, Reliabiltiy of plasma testosterone assays by competitiveprotein binding methods, Acta Endocrinol. Suppl. (Copenh.) 147 (1970) 257–274.

[15] W.R. Slaunwhite Jr., A.A. Sandberg, Analysis of corticosteroids, Acta Endocrinol.Suppl. (Copenh.) 147 (1970) 144–154.

[16] H. Khodabandehlou, B. Hoffmann, J. Pallauf, Investigations in central Hessia onthe occurrence of estrogenic activity content of cattle feed, Dtsch. Tierarztl.Wochenschr. 104 (1997) 291–294.

[17] Steroid Immunoassay, Proceedings of the Fifth Tenovus Workshop Alpha OmegaPublishing, Cardiff, UK, 1975.

[18] S.L. Jeffcoate, Recent developments in steroid radioimmunoassays, J. Reprod.Fertil. 51 (1977) 267–272.

[19] S.L. Jeffcoate, Progress towards the wider use of better steroid immunoassays, J.Steroid Biochem. 11 (1979) 1051–1055.

[20] E. Nieschlag, E.J. Wickings, A review of radioimmunoassay for steroids, Z. Klin.Chem. Klin. Biochem. 13 (1975) 261–271.

[21] J.J. Pratt, Steroid immunoassay in clinical chemistry, Clin. Chem. 24 (1978)1869–1890.

[22] B. Hoffmann, F. Gentz, K. Failing, Investigations into the course of progesterone-,oestrogen- and eCG-concentrations during normal and impaired pregnancy in themare*, Reprod. Domest. Anim. 31 (1996) 717–723.

[23] F. Schwarzenberger, G.S. Toole, H.L. Christie, J.I. Raeside, Plasma levels of severalandrogens and estrogens from birth to puberty in male domestic pigs, ActaEndocrinol. (Copenh.) 128 (1993) 173–177.

S.A. Wudy et al. Journal of Steroid Biochemistry and Molecular Biology 179 (2018) 88–103

101

[24] L. Akin, S. Kurtoglu, M. Kendirci, M.A. Akin, M.F. Hartmann, S.A. Wudy, Hookeffect: a pitfall leading to misdiagnosis of hypoaldosteronism in an infant withpseudohypoaldosteronism, Horm. Res. Paediatr. 74 (2010) 72–75.

[25] S.A. Wudy, U.A. Wachter, J. Homoki, W.M. Teller, 17 Alpha-hydroxyprogesterone,4-androstenedione, and testosterone profiled by routine stable isotope dilution/gas chromatography-mass spectrometry in plasma of children, Pediatr. Res. 38(1995) 76–80.

[26] C.H. Shackleton, C. Kletke, S. Wudy, J.H. Pratt, Dehydroepiandrosterone sulfatequantification in serum using high-performance liquid chromatography/massspectrometry and a deuterated internal standard: a technique suitable for routineuse or as a reference method, Steroids 55 (1990) 472–478.

[27] W. Rosner, R.J. Auchus, R. Azziz, P.M. Sluss, H. Raff, Position statement: utility,limitations, and pitfalls in measuring testosterone: an endocrine society positionstatement, J. Clin. Endocrinol. Metab. 92 (2007) 405–413.

[28] D.J. Haisenleder, A.H. Schoenfelder, E.S. Marcinko, L.M. Geddis, J.C. Marshall,Estimation of estradiol in mouse serum samples: evaluation of commercial estra-diol immunoassays, Endocrinology 152 (2011) 4443–4447.

[29] F.Z. Stanczyk, N.J. Clarke, Measurement of estradiol–challenges ahead, J. Clin.Endocrinol. Metab. 99 (2014) 56–58.

[30] G.D. Niswender, Influence of the site of conjugation on the specificity of antibodiesto progesterone, Steroids 22 (1973) 413–424.

[31] D.Y. Yoon, M.J. Choi, I.S. Choe, T.W. Chung, S.M. Byun, Influence of the con-jugation site on the specificity of monoclonal antibodies to progesterone and onthe performance of direct enzyme immunoassay, Biochem. Mol. Biol. Int. 31(1993) 553–563.

[32] J.F. Yarrow, D.T. Beck, C.F. Conover, L.A. Beggs, B.A. Goldberger, S.E. Borst,Invalidation of a commercially available human 5alpha-dihydrotestosterone im-munoassay, Steroids 78 (2013) 1220–1225.

[33] E. Mostl, R. Palme, Hormones as indicators of stress, Domest. Anim. Endocrinol. 23(2002) 67–74.

[34] F. Schwarzenberger, G. Speckbacher, E. Bamberg, Plasma and fecal progestagenevaluations during and after the breeding season of the female vicuna (Vicunavicuna), Theriogenology 43 (1995) 625–634.

[35] F. Schwarzenberger, C. Walzer, K. Tomasova, J. Vahala, J. Meister,K.L. Goodrowe, J. Zima, G. Strauss, M. Lynch, Faecal progesterone metaboliteanalysis for non-invasive monitoring of reproductive function in the white rhi-noceros (Ceratotherium simum), Anim. Reprod. Sci. 53 (1998) 173–190.

[36] J.M. Rash, I. Jerkunica, D.S. Sgoutas, Lipid interference in steroid radio-immunoassay, Clin. Chem. 26 (1980) 84–88.

[37] T.G. Shrivastav, Matrix interference in direct total testosterone enzyme im-munoassay and its eliminiation with the use of non-crossreactive steroids in serumbased standards, Health Popul.—Perspect. Issues 25 (2002) 55–64.

[38] R.M. Lequin, Enzyme immunoassay (EIA)/enzyme-linked immunosorbent assay(ELISA), Clin. Chem. 51 (2005) 2415–2418.

[39] G.F. Blackburn, H.P. Shah, J.H. Kenten, J. Leland, R.A. Kamin, J. Link,J. Peterman, M.J. Powell, A. Shah, D.B. Talley, et al., Electrochemiluminescencedetection for development of immunoassays and DNA probe assays for clinicaldiagnostics, Clin. Chem. 37 (1991) 1534–1539.

[40] B.K. van Weemen, A.M. Bosch, E.C. Dawson, A.H. Schuurs, Enzyme-immunoassayof steroids: possibilities and pitfalls, J. Steroid Biochem. 11 (1979) 147–151.

[41] G.D. Nordblom, R. Webb, R.E. Counsell, B.G. England, A chemical approach tosolving bridging phenomena in steroid radioimmunoassays, Steroids 38 (1981)161–173.

[42] L.X. Tiefenauer, D.M. Bodmer, W. Frei, R.Y. Andres, Prevention of bridge bindingin immunoassays: a general estradiol tracer structure, J. Steroid Biochem. 32(1989) 251–257.

[43] H. Meyer, Enzymimmunologische Meßverfahren zur Hormonanalytik, EnkeStuttgart, 1989.

[44] R.J. Auchus, Steroid assays and endocrinology: best practices for basic scientists,Endocrinology 155 (2014) 2049–2051.

[45] J. Taieb, C. Benattar, A.S. Birr, A. Lindenbaum, Limitations of steroid determi-nation by direct immunoassay, Clin. Chem. 48 (2002) 583–585.

[46] D.A. Skoog, F.J. Holler, S.R. Crouch, Principles of Instrumental Analysis, ThomsonBrooks/Cole, 2007.

[47] S. Maher, F.P.M. Jjunju, S. Taylor, Colloquium: 100 years of mass spectrometry:perspectives and future trends, Rev. Mod. Phys. 87 (2015) 113–135.

[48] A. Gaudl, J. Kratzsch, Y.J. Bae, W. Kiess, J. Thiery, U. Ceglarek, Liquid chroma-tography quadrupole linear ion trap mass spectrometry for quantitative steroidhormone analysis in plasma urine, saliva and hair, J. Chromatogr. A 1464 (2016)64–71.

[49] A. Römpp, S. Guenther, Y. Schober, O. Schulz, Z. Takats, W. Kummer, B. Spengler,Histology by mass spectrometry: label-free tissue characterization obtained fromhigh-accuracy bioanalytical imaging, Angew. Chem. Int. Ed. 49 (2010)3834–3838.

[50] D.F. Cobice, C.L. Mackay, R.J. Goodwin, A. McBride, P.R. Langridge-Smith,S.P. Webster, B.R. Walker, R. Andrew, Mass spectrometry imaging for dissectingsteroid intracrinology within target tissues, Anal. Chem. 85 (2013) 11576–11584.

[51] S.A. Wudy, Synthetic procedures for the preparation of deuterium-labeled analogsof naturally occurring steroids, Steroids 55 (1990) 463–471.

[52] S.A. Wudy, U.A. Wachter, J. Homoki, W.M. Teller, C.H. Shackleton, Androgenmetabolism assessment by routine gas chromatography/mass spectrometry pro-filing of plasma steroids: part 1 unconjugated steroids, Steroids 57 (1992)319–324.

[53] R.F. Greaves, C.S. Ho, K.E. Hoad, J. Joseph, B. McWhinney, J.P. Gill, T. Koal,C. Fouracre, H.P. Iu, B.R. Cooke, C. Boyder, H.T. Pham, L.M. Jolly, Achievementsand future directions of the APFCB mass spectrometry harmonisation project on

serum testosterone, Clin. Biochem. Rev. 37 (2016) 63–84.[54] L.J. Owen, B.G. Keevil, Testosterone measurement by liquid chromatography

tandem mass spectrometry: the importance of internal standard choice, Ann. Clin.Biochem. 49 (2012) 600–602.

[55] W.J. Vandenheuvel, E.C. Horning, Gas chromatography of adrenal cortical steroidhormones, Biochem. Biophys. Res. Commun. 3 (1960) 356–360.

[56] P. Eneroth, B. Gordon, R. Ryhage, J. Sjovall, Identification of mono- and dihydroxybile acids in human feces by gas-liquid chromatography and mass spectrometry, J.Lipid Res. 7 (1966) 511–523.

[57] E. Horning, Gas phase analytical methods for the study of steroid hormones andtheir metabolites, Gas Phase Chromatography of Steroids, Springer, 1968, pp.1–71.

[58] Steroid Analysis, in: H.L. Makin, D.B. Gower (Eds.), 2nd ed., Springer, Heidelberg,London, New York, 2010.

[59] A. Sanchez-Guijo, M.F. Hartmann, S.A. Wudy, Introduction to gas chromato-graphy-mass spectrometry, Methods Mol. Biol. 1065 (2013) 27–44.

[60] E.M. Agency, Guideline on bioanalytical method validation, http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2011/08/WC500109686.pdf.

[61] F.a.D.A. U.S. Department of Health and Human Services, Center for DrugEvaluation and Research, Guidance for industry: bioanalytical method validation,http://www.fda.gov/downloads/Drugs/Guidances/ucm070107.pdf, 2001.

[62] M.H. Choi, B.C. Chung, Bringing GC–MS profiling of steroids into clinical appli-cations, Mass Spectrom. Rev. 34 (2015) 219–236.

[63] C.H. Shackleton, Mass spectrometry in the diagnosis of steroid-related disordersand in hypertension research, J. Steroid Biochem. Mol. Biol. 45 (1993) 127–140.

[64] C.H. Shackleton, Profiling steroid hormones and urinary steroids, J. Chromatogr.379 (1986) 91–156.

[65] S.A. Wudy, M.F. Hartmann, Gas chromatography-mass spectrometry profiling ofsteroids in times of molecular biology, Horm. Metab. Res. 36 (2004) 415–422.

[66] A. Kulle, N. Krone, P.M. Holterhus, G. Schuler, R. Greaves, A. Juul, Y. de Rijke,M. Hartmann, A. Saba, O. Hiort, S. Wudy, Diagnosis of endocrine disease: steroidhormone analysis in diagnosis and treatment of DSD position paper of EU COSTaction BM 1303 DSDnet, Eur. J. Endocrinol. 176 (2017) P1–P9.

[67] S.A. Wudy, M.F. Hartmann, T. Remer, Sexual dimorphism in cortisol secretionstarts after age 10 in healthy children: urinary cortisol metabolite excretion ratesduring growth, Am. J. Physiol. Endocrinol. Metab. 293 (2007) E970–976.

[68] D. Tiosano, R. Navon, O. Flor, C. Knopf, M. Hartmann, S. Wudy, Z. Yakhini,Z. Hochberg, A steroid metabolomic approach to 17α-hydroxylase/17, 20 lyasedeficiency, Metabolomics 6 (2010) 417–426.

[69] A. Gawlik, M. Shmoish, M.F. Hartmann, E. Malecka-Tendera, S.A. Wudy,Z. Hochberg, Steroid metabolomic disease signature of nonsyndromic childhoodobesity, J. Clin. Endocrinol. Metab. 101 (2016) 4329–4337.

[70] A. Hobler, N. Kagawa, M.C. Hutter, M.F. Hartmann, S.A. Wudy, F. Hannemann,R. Bernhardt, Human aldosterone synthase: recombinant expression in E. coli andpurification enables a detailed biochemical analysis of the protein on the mole-cular level, J. Steroid Biochem. Mol. Biol. 132 (2012) 57–65.

[71] J. Neunzig, M. Milhim, L. Schiffer, Y. Khatri, J. Zapp, A. Sanchez-Guijo,M.F. Hartmann, S.A. Wudy, R. Bernhardt, The steroid metabolite 16(beta)-OH-androstenedione generated by CYP21A2 serves as a substrate for CYP19A1, J.Steroid Biochem. Mol. Biol. 167 (2017) 182–191.

[72] T. Remer, K.R. Boye, M. Hartmann, C.M. Neu, E. Schoenau, F. Manz, S.A. Wudy,Adrenarche and bone modeling and remodeling at the proximal radius: weakandrogens make stronger cortical bone in healthy children, J. Bone Miner. Res. 18(2003) 1539–1546.

[73] N. Krone, B.A. Hughes, G.G. Lavery, P.M. Stewart, W. Arlt, C.H. Shackleton, Gaschromatography/mass spectrometry (GC/MS) remains a pre-eminent discoverytool in clinical steroid investigations even in the era of fast liquid chromatographytandem mass spectrometry (LC/MS/MS), J. Steroid Biochem. Mol. Biol. 121(2010) 496–504.

[74] J.W. Honour, Steroid profiling, Ann. Clin. Biochem. 34 (Pt. 1) (1997) 32–44.[75] J.W. Honour, C.G. Brook, Clinical indications for the use of urinary steroid profiles

in neonates and children, Ann. Clin. Biochem. 34 (Pt. 1) (1997) 45–54.[76] M. Langhammer, M. Michaelis, M.F. Hartmann, S.A. Wudy, A. Sobczak,

G. Nurnberg, N. Reinsch, J. Schon, J.M. Weitzel, Reproductive performance pri-marily depends on the female genotype in a two-factorial breeding experimentusing high-fertility mouse lines, Reproduction 153 (2017) 361–368.

[77] S. Christakoudi, D.A. Cowan, N.F. Taylor, Steroids excreted in urine by neonateswith 21-hydroxylase deficiency: characterization, using GC–MS and GC–MS/MS,of the D-ring and side chain structure of pregnanes and pregnenes, Steroids 75(2010) 34–52.

[78] S.A. Wudy, M.H. Choi, Steroid LC–MS has come of age, J. Steroid Biochem. Mol.Biol. 162 (2016) 1–3.

[79] O. Nunez, H. Gallart-Ayala, C.P. Martins, P. Lucci, R. Busquets, State-of-the-art infast liquid chromatography-mass spectrometry for bio-analytical applications, J.Chromatogr. B Analyt. Technol. Biomed. Life Sci. 927 (2013) 3–21.

[80] D.P. Kloos, H. Lingeman, W.M. Niessen, A.M. Deelder, M. Giera, O.A. Mayboroda,Evaluation of different column chemistries for fast urinary metabolic profiling, J.Chromatogr. B Analyt. Technol. Biomed. Life Sci. 927 (2013) 90–96.

[81] F. Qin, Y.Y. Zhao, M.B. Sawyer, X.F. Li, Hydrophilic interaction liquid chroma-tography-tandem mass spectrometry determination of estrogen conjugates inhuman urine, Anal. Chem. 80 (2008) 3404–3411.

[82] M.J. Ruiz-Angel, M.C. Garcia-Alvarez-Coque, A. Berthod, S. Carda-Broch, Areanalysts doing method validation in liquid chromatography? J. Chromatogr. A1353 (2014) 2–9.

[83] J.W. Honour, Development and validation of a quantitative assay based on tandem

S.A. Wudy et al. Journal of Steroid Biochemistry and Molecular Biology 179 (2018) 88–103

102

mass spectrometry, Ann. Clin. Biochem. 48 (2011) 97–111.[84] C.a.L.S.I. (CLSI), C62 – Liquid Chromatography-Mass Spectrometry Methods,

https://clsi.org/standards/products/clinical-chemistry-and-toxicology/c62/,2014.

[85] C.a.L.S.I. (CLSI), C50 – Mass Spectrometry in the Clinical Laboratory: GeneralPrinciples and Guidance, https://clsi.org/standards/products/clinical-chemistry-and-toxicology/c50/, 2007.

[86] I.O.f.S. (ISO), ISO 17511:2003 – In vitro diagnostic medical devices –Measurement of quantities in biological samples – Metrological traceability ofvalues assigned to calibrators and control materials, https://www.iso.org/standard/30716.html, 2003.

[87] C.a.L.S.I. (CLSI), EP32 – Metrological Traceability and Its Implementation, AReport, https://clsi.org/standards/products/method-evaluation/ep32/, 2006.

[88] P.J. Taylor, Matrix effects: the Achilles heel of quantitative high-performance li-quid chromatography-electrospray-tandem mass spectrometry, Clin. Biochem. 38(2005) 328–334.

[89] O. Gonzalez, M.E. Blanco, G. Iriarte, L. Bartolome, M.I. Maguregui, R.M. Alonso,Bioanalytical chromatographic method validation according to current regulationswith a special focus on the non-well defined parameters limit of quantification,robustness and matrix effect, J. Chromatogr. A 1353 (2014) 10–27.

[90] S.J. Soldin, O.P. Soldin, Steroid hormone analysis by tandem mass spectrometry,Clin. Chem. 55 (2009) 1061–1066.

[91] B.R. Stolze, V. Gounden, J. Gu, E.A. Elliott, L.S. Masika, B.S. Abel, D.P. Merke,M.C. Skarulis, S.J. Soldin, An improved micro-method for the measurement ofsteroid profiles by APPI-LC–MS/MS and its use in assessing diurnal effects onsteroid concentrations and optimizing the diagnosis and treatment of adrenal in-sufficiency and CAH, J. Steroid Biochem. Mol. Biol. 162 (2016) 110–116.

[92] J.M. Hawley, B.G. Keevil, Endogenous glucocorticoid analysis by liquid chroma-tography-tandem mass spectrometry in routine clinical laboratories, J. SteroidBiochem. Mol. Biol. 162 (2016) 27–40.

[93] M.M. Kushnir, A.L. Rockwood, W.L. Roberts, B. Yue, J. Bergquist, A.W. Meikle,Liquid chromatography tandem mass spectrometry for analysis of steroids inclinical laboratories, Clin. Biochem. 44 (2011) 77–88.

[94] P.J. Taylor, D.P. Cooper, R.D. Gordon, M. Stowasser, Measurement of aldosteronein human plasma by semiautomated HPLC-tandem mass spectrometry, Clin.Chem. 55 (2009) 1155–1162.

[95] Q. Wang, C. Mesaros, I.A. Blair, Ultra-high sensitivity analysis of estrogens forspecial populations in serum and plasma by liquid chromatography–mass spec-trometry: assay considerations and suggested practices, J. Steroid Biochem. Mol.Biol. 162 (2016) 70–79.

[96] M. Vogeser, K.G. Parhofer, Liquid chromatography tandem-mass spectrometry(LC-MS/MS)–technique and applications in endocrinology, Exp. Clin. Endocrinol.Diabetes 115 (2007) 559–570.

[97] A.E. Kulle, M. Welzel, P.M. Holterhus, F.G. Riepe, Implementation of a liquidchromatography tandem mass spectrometry assay for eight adrenal C-21 steroidsand pediatric reference data, Horm. Res. Paediatr. 79 (2013) 22–31.

[98] O.P. Soldin, H. Sharma, L. Husted, S.J. Soldin, Pediatric reference intervals foraldosterone 17alpha-hydroxyprogesterone, dehydroepiandrosterone, testosteroneand 25-hydroxy vitamin D3 using tandem mass spectrometry, Clin. Biochem. 42(2009) 823–827.

[99] M. Peitzsch, T. Dekkers, M. Haase, F.C. Sweep, I. Quack, G. Antoch, G. Siegert,J.W. Lenders, J. Deinum, H.S. Willenberg, G. Eisenhofer, An LC–MS/MS methodfor steroid profiling during adrenal venous sampling for investigation of primaryaldosteronism, J. Steroid Biochem. Mol. Biol. 145 (2015) 75–84.

[100] B.C. Netzel, K.W. Cradic, E.T. Bro, A.B. Girtman, R.C. Cyr, R.J. Singh, S.K. Grebe,Increasing liquid chromatography-tandem mass spectrometry throughput by masstagging: a sample-multiplexed high-throughput assay for 25-hydroxyvitamin D2and D3, Clin. Chem. 57 (2011) 431–440.

[101] W.J. Griffiths, J. Abdel-Khalik, P.J. Crick, E. Yutuc, Y. Wang, New methods foranalysis of oxysterols and related compounds by LC-MS, J. Steroid Biochem. Mol.Biol. 162 (2016) 4–26.

[102] J. Marcos, O.J. Pozo, Current LC-MS methods and procedures applied to theidentification of new steroid metabolites, J. Steroid Biochem. Mol. Biol. 162(2016) 41–56.

[103] T. Piper, W. Schanzer, M. Thevis, Revisiting the metabolism of 19-nortestosteroneusing isotope ratio and high resolution/high accuracy mass spectrometry, J.Steroid Biochem. Mol. Biol. 162 (2016) 80–91.

[104] K. Wegner, S. Just, L. Gau, H. Mueller, P. Gerard, P. Lepage, T. Clavel, S. Rohn,Rapid analysis of bile acids in different biological matrices using LC-ESI–MS/MSfor the investigation of bile acid transformation by mammalian gut bacteria, Anal.Bioanal. Chem. 409 (2017) 1231–1245.

[105] A. Sanchez-Guijo, C. Blaschka, M.F. Hartmann, C. Wrenzycki, S.A. Wudy, Profilingof bile acids in bovine follicular fluid by fused-core-LC-MS/MS, J. SteroidBiochem. Mol. Biol. 162 (2016) 117–125.

[106] W. Gao, C. Kirschbaum, J. Grass, T. Stalder, LC-MS based analysis of endogenoussteroid hormones in human hair, J. Steroid Biochem. Mol. Biol. 162 (2016) 92–99.

[107] M. Mezzullo, A. Fazzini, A. Gambineri, G. Di Dalmazi, R. Mazza, C. Pelusi,V. Vicennati, R. Pasquali, U. Pagotto, F. Fanelli, Parallel diurnal fluctuation oftestosterone, androstenedione, dehydroepiandrosterone and 17OHprogesterone asassessed in serum and saliva: validation of a novel liquid chromatography-tandemmass spectrometry method for salivary steroid profiling, Clin. Chem. Lab. Med.(2017).

[108] M.M. Kushnir, T. Naessen, K. Wanggren, J. Hreinsson, A.L. Rockwood,A.W. Meikle, J. Bergquist, Exploratory study of the association of steroid profilesin stimulated ovarian follicular fluid with outcomes of IVF treatment, J. SteroidBiochem. Mol. Biol. 162 (2016) 126–133.

[109] C.E. Galuska, M.F. Hartmann, A. Sanchez-Guijo, K. Bakhaus, J. Geyer, G. Schuler,K.P. Zimmer, S.A. Wudy, Profiling intact steroid sulfates and unconjugated ster-oids in biological fluids by liquid chromatography-tandem mass spectrometry(LC–MS-MS), Analyst 138 (2013) 3792–3801.

[110] A. Sanchez-Guijo, V. Oji, M.F. Hartmann, H. Traupe, S.A. Wudy, Simultaneousquantification of cholesterol sulfate, androgen sulfates, and progestagen sulfates inhuman serum by LC–MS/MS, J. Lipid Res. 56 (2015) 1843–1851.

[111] M. Doue, G. Dervilly-Pinel, K. Pouponneau, F. Monteau, B. Le Bizec, Analysis ofglucuronide and sulfate steroids in urine by ultra-high-performance supercritical-fluid chromatography hyphenated tandem mass spectrometry, Anal. Bioanal.Chem. 407 (2015) 4473–4484.

[112] M.D. McLeod, C.C. Waller, A. Esquivel, G. Balcells, R. Ventura, J. Segura,O.J. Pozo, A constant ion loss method for the untargeted detection of bissulfatemetabolites, Anal. Chem. (2016).

[113] S.H. Lee, N. Lee, Y. Hong, B.C. Chung, M.H. Choi, Simultaneous analysis of freeand sulfated steroids by liquid chromatography/mass spectrometry with selectivemass spectrometric scan modes and polarity switching, Anal. Chem. 88 (2016)11624–11630.

[114] A. Sanchez-Guijo, J. Neunzig, A. Gerber, V. Oji, M.F. Hartmann, H.C. Schuppe,H. Traupe, R. Bernhardt, S.A. Wudy, Role of steroid sulfatase in steroid home-ostasis and characterization of the sulfated steroid pathway: evidence from steroidsulfatase deficiency, Mol. Cell. Endocrinol. 437 (2016) 142–153.

[115] D. Fietz, K. Bakhaus, B. Wapelhorst, G. Grosser, S. Gunther, J. Alber, B. Doring,S. Kliesch, W. Weidner, C.E. Galuska, M.F. Hartmann, S.A. Wudy, M. Bergmann,J. Geyer, Membrane transporters for sulfated steroids in the human testis–cellularlocalization, expression pattern and functional analysis, PLoS One 8 (2013)e62638.

[116] G. Schuler, Y. Dezhkam, L. Bingsohn, B. Hoffmann, K. Failing, C.E. Galuska,M.F. Hartmann, A. Sanchez-Guijo, S.A. Wudy, Free and sulfated steroids secretionin postpubertal boars (Sus scrofa domestica), Reproduction 148 (2014) 303–314.

[117] H. Schweigmann, A. Sanchez-Guijo, B. Ugele, K. Hartmann, M.F. Hartmann,M. Bergmann, C. Pfarrer, B. Doring, S.A. Wudy, E. Petzinger, J. Geyer, G. Grosser,Transport of the placental estriol precursor 16alpha-hydroxy-dehydroepian-drosterone sulfate (16alpha-OH-DHEAS) by stably transfected OAT4-, SOAT-, andNTCP-HEK293 cells, J. Steroid Biochem. Mol. Biol. 143 (2014) 259–265.

[118] C. Shackleton, Clinical steroid mass spectrometry: a 45-year history culminating inHPLC-MS/MS becoming an essential tool for patient diagnosis, J. Steroid Biochem.Mol. Biol. 121 (2010) 481–490.

[119] F. Badoud, E. Grata, J. Boccard, D. Guillarme, J.L. Veuthey, S. Rudaz, M. Saugy,Quantification of glucuronidated and sulfated steroids in human urine by ultra-high pressure liquid chromatography quadrupole time-of-flight mass spectro-metry, Anal. Bioanal. Chem. 400 (2011) 503–516.

[120] A. Sanchez-Guijo, M.F. Hartmann, L. Shi, T. Remer, S.A. Wudy, Determination offree cortisol and free cortisone in human urine by on-line turbulent flow chro-matography coupled to fused-core chromatography-tandem mass spectrometry(TFC-HPLC–MS/MS), Anal. Bioanal. Chem. 406 (2014) 793–801.

[121] T. Soeborg, H. Frederiksen, P. Fruekilde, T.H. Johannsen, A. Juul, A.M. Andersson,Serum concentrations of DHEA DHEAS, 17alpha-hydroxyprogesterone, Delta4-androstenedione and testosterone in children determined by TurboFlow-LC–MS/MS, Clin. Chim. Acta 419 (2013) 95–101.

[122] V.M. Carvalho, The coming of age of liquid chromatography coupled to tandemmass spectrometry in the endocrinology laboratory, J. Chromatogr. B Analyt.Technol. Biomed. Life Sci. 883-884 (2012) 50–58.

[123] W.J. Griffiths, M. Hornshaw, G. Woffendin, S.F. Baker, A. Lockhart,S. Heidelberger, M. Gustafsson, J. Sjovall, Y. Wang, Discovering oxysterols inplasma: a window on the metabolome, J. Proteome Res. 7 (2008) 3602–3612.

[124] H. Rosa, Social Acceleration: A New Theory of Modernity, Columbia UniversityPress, 2013.

[125] R. Knuppen, O. Haupt, H.O. Hoppen, Synthesis of [2H8] estradiol [2H7] estrone,[2H6] 2-hydroxyestrone and [2H6] 4-hydroxyestrone as internal standards forselected ion monitoring, Steroids 39 (1982) 667–673.

[126] H.W. Vesper, S. Bhasin, C. Wang, S.S. Tai, L.A. Dodge, R.J. Singh, J. Nelson,S. Ohorodnik, N.J. Clarke, W.A. Salameh, C.R. Parker Jr., R. Razdan, E.A. Monsell,G.L. Myers, Interlaboratory comparison study of serum total testosterone [cor-rected] measurements performed by mass spectrometry methods, Steroids 74(2009) 498–503.

[127] H.W. Vesper, W.G. Miller, G.L. Myers, Reference materials and commutability,Clin. Biochem. Rev. 28 (2007) 139–147.

[128] R.F. Greaves, L. Jolly, M.F. Hartmann, C.S. Ho, R.K. Kam, J. Joseph, C. Boyder,S.A. Wudy, Harmonisation of serum dihydrotestosterone analysis: establishment ofan external quality assurance program, Clin. Chem. Lab. Med. (2016).

[129] JCTLM database, http://www.bipm.org/jctlm/.[130] A little knowledge, Nature 472 (2011) 135–135.[131] C. Ludwig, P.O. Desmoulins, M.A. Driancourt, S. Goericke-Pesch, B. Hoffmann,

Reversible downregulation of endocrine and germinative testicular function(hormonal castration) in the dog with the GnRH-agonist azagly-nafarelin as aremovable implant Gonazon; a preclinical trial, Theriogenology 71 (2009)1037–1045.

S.A. Wudy et al. Journal of Steroid Biochemistry and Molecular Biology 179 (2018) 88–103

103