LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 193

9

LC-MS in doping control

Detlef Thieme

Introduction adjacent fields with similar analytical prospects, like veterinary residue control (predominantly dealing with identification of growth promoters Definition of doping in various matrices), forensic sciences (the major- ity of doping-relevant substances are scheduled Doping analysis comprises a diversity of sub- as controlled substances in most countries), stance classes with different pharmaceutical and environmental analysis (e.g. steroids in waste chemical properties. Therefore, the discussion of water) or clinical chemistry (e.g. due to the the suitability of liquid chromatography-mass increasing relevance of steroid hormone replace- spectrometry (LC-MS) in doping analysis needs ment therapy). to distinguish various categories. This chapter describes the key fields of appli- According to its formal definition, a doping cation of LC-MS in routine doping control (i.e. violation in sports can be caused by various screening analysis, confirmation and quantifica- events, e.g.: tion of positive results) extra to particular • the detection of a prohibited substance or research activities. The latter are focused on the metabolites or markers of that substance (as intended technical improvements (e.g. extension defined by the recent document [1] of the of detection time windows, reduction of turn- World Anti-Doping Agency [WADA]) in the around times and costs) of conventional analyti- athlete’s specimen cal procedures and, in particular, on the detection • the use of prohibited substances or methods of prohibited substances that cannot be ade- • possession or trafficking prohibited substances quately identified so far (e.g. growth hormone). • refusing without compelling justification to The arrangement follows mainly historical submit a sample. and technical considerations, and does not necessarily represent the frequency or relevance This definition is clearly legally motivated and of the application of LC-MS. does not support the discussion of technical issues. The number and classes of prohibited sub- LC-MS in doping control – historical and stances is very complex in human sports, where β technical aspects selected stimulants, narcotics, hormones, 2- agonists, anti-oestrogenic agents and diuretics are Some peculiar legal and technical principles in covered. The situation becomes even more con- doping control influence analytical strategies in fusing if the term ‘doping’ is extended to animal athletes drug testing: (e.g. equestrian) sports, where any application of pharmaceutical drugs is totally prohibited and • The substance-based doping definition priori- even substance classes like muscle relaxants or tises target analyses compared with general mild stimulants are included. unknown screening procedures. Sensitive and Moreover, doping analysis is closely related to specific selected ion monitoring (SIM) or

193 LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 194

194 Chapter 9 • LC-MS in doping control

selected reaction monitoring (SRM) experi- nique in anti-doping research and routine [3] ments are much more frequent than scanning analyses. The issue of peptide hormones was experiments. already tackled in the mid-1990s, because gas • Urine, which is the preferred specimen for chromatography (GC)-MS analysis could not doping control due to the ease of sample solve the problem of identification of macromol- collection and relatively high concentrations ecular compounds. The potential of LC-MS to dif- of xenobiotics, requires a careful consideration ferentiate intact growth hormone obtained from of substance metabolism, including conjuga- different manufacturers, quantify the insulin-like tion. Minor biochemical pathways leading to growth factor (IGF-1) and characterise human long-term metabolites are often more import- chorionic gonadotrophin (hCG) after tryptic ant than active parent compounds. The digestion had already been reported in 1994 [4, relevance of quantitative analyses is reduced 5]. However, these ‘proofs of principles’ demon- to a few ‘threshold substances’. This group strate the general usefulness of LC-MS for the comprises compounds that are accepted below identification of peptide hormone doping, but certain threshold concentrations, because low are not used routinely to date, mainly due to amounts may be due to a permitted adminis- sensitivity limitations. tration (e.g. inhalation of salbutamol) or an The subsequent developments were mainly endogenous origin of the substance (e.g. characterised by practical improvements. Sub- natural levels of testosterone). In general, stances (e.g. mesocarb [6]) and substance groups qualitative substance identification is a (e.g. diuretics [7]) causing severe analytical sufficient proof of a doping offence. problems (stability) or inconvenience (time- • The differentiation between substance prohi- consuming derivatisation reactions) were bition ‘in competition’ and ‘out of competi- covered by efficient LC-MS assays, while other tion’ requires modified analytical procedures screenings remained unchanged, due to the with respect to numbers of included sub- availability of well-established and validated stances and threshold concentrations. GC-MS procedures. • A major analytical challenge consists in the In contrast, there is an obvious preference to verification of the prohibited administration use LC-MS in cases of upcoming new substances of endogenous substances like testosterone, (like the ‘designer steroid’ tetrahydrogestrinone human growth hormone (hGH) or erythro- [THG] and the stimulant modafinil) or substance poietin (EPO). In such cases, minor quantita- groups (e.g. corticosteroids, included in the list of tive (e.g. amount of steroids compared with prohibited substances in 2003). endogenous precursors or biochemical There is no preferred default LC-MS instru- byproducts) or qualitative (e.g. glycosylation mentation in doping control analyses. Almost of proteins) deviations need to be identified. any technical variant of ionisation – electrospray • MS plays an outstanding role in doping ionisation (ESI), atmospheric pressure chemical analysis and was originally considered as a ionisation (APCI) or photo-ionisation (APPI) – mandatory analytical technique for confirma- has been applied in combination with quadru- tion of substance identity. Exceptions were pole (Q), ion trap (IT) or time-of-flight (TOF) mass later acknowledged in the field of peptide analysers, whether as single or tandem mass hormones. spectrometers. Certain reports of related technical develop- Approaches to the application of an LC-MS ments of LC-MS seem to be just coincidentally coupling in the framework of doping control associated with doping, e.g.: were already reported in 1981 [2], when a combination of LC-MS equipped with a moving • The introduction of isotope ratio MS linked to belt was used for MS confirmation of corticos- LC enables the identification of the origin of teroids. The relevance of LC-MS application in substances. In particular, a differentiation of doping control was ruled by practical demands, endogenous production from synthetic mat- resulting in an early implementation of the tech- erial becomes possible in principle. However, LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 195

Small molecules 195

applications presented so far [8] are rather complex too. It summarises any substance that insensitive (requiring 400 ng substance on- may interfere with doping analysis by: column) and therefore of no practical value in • diluting the urine and accelerating excretion routine cases. (diuretics) • The introduction of Fourier-transform ion • suppression of reabsorption of xenobiotics cyclotron resonance (FTICR) MS to identify (uricosurics, e.g. probenicid) corticosteroids [9] is probably technically • manipulation of endogenous steroid profiles motivated. The high expense of this technique (administration of epitestosterone, used as an does not permit routine applications. Never- endogenous reference of urinary steroid con- theless, it is clear that high-resolution (HR)-MS centrations, is able to conceal elevated levels is essential for identification of multiply of testosterone). charged intact peptide hormones (hGH) [10]. Affordable routine instruments could greatly improve the detection and characterisation of Diuretics proteins. Additionally, progress in chromatographic separ- The main motivation to prohibit the use of ation (e.g. column switching [11], use of graphi- diuretics in sports is the intended reduction of tised carbon [12] or chiral [13] columns) needs to body mass in weight-classified sports. The second be achieved, particularly in the field of peptide reason is a masking effect. Due to forced diuresis, hormones. The improvement of ionisation the clearance of prohibited substances (e.g. ana- efficiency appears to be a crucial aspect of steroid bolic steroids) may be accelerated and the urinary analysis. Derivatisation (dansylation [14]) of concentration may drop below the detection steroids and attempts to improve APPI ionisation limit or threshold. Different types of diuretics (e.g. using anisol as dopant gas) are both specifi- (thiazide-like, loop and potassium-sparing cally focused on a sensitivity enhancement [15]. diuretics) may be distinguished on the basis of their pharmacological properties (Table 9.1). However, these classes are not differentiated with regard to their doping relevance. The first two Small molecules groups are supposed to be most efficient in doping because of their high potency. They are The majority of doping-relevant substances characterised by acidic groups (acid amides, typi- belong to the category of small molecules. The cally sulphonamides) and are therefore suitable most common definition of substance groups is for negative ionisation. In contrast to these com- based on pharmaceutical activity, distinguishing pounds, the group of potassium-sparing diuretics stimulants, narcotics, cannabinoids, anabolic is characterised by steroid structures (aldosterone agents, β-agonists, anti-oestrogens, masking antagonists, e.g. canrenone) or cycloamidine agents and glucocorticoids. However, this sched- structures (triamterene), and is more appropriate ule is not suitable for analytical consideration as to protonation and subsequent detection in pos- structurally and analytically unrelated species itive-ionisation mode. may be arbitrarily grouped together (e.g. clen- Other compounds with diuretic (side) effects, buterol and testosterone as anabolic agents) like osmodiuretics (mannitol) or xanthine deriv- because of their equal intended activity. atives (caffeine), are no longer mentioned in the However, one substance may be scheduled in list of doping substances, because prohibited use β different groups. The 2-agonist salbutamol, for could not be certainly discriminated from their instance, may be considered as a permitted anti- permitted applications as food ingredients. asthmatic, as a stimulant or as an anabolic agent, The identification of diuretics in doping depending on the occasion of the doping control control was originally carried out by GC-MS and (in/out of competition) and on its urinary con- LC-UV diode array detection (LC-DAD) [7, 16]. centration. The group of masking agents is rather The reduced polarity of the methyl derivatives LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 196

196 Chapter 9 • LC-MS in doping control

Table 9.1 Typical classes and examples of prohibited diuretics

Class Typical modifications Examples Chemical structure

OO Thiazide R1 = H, alkyl, subst. phenyl benzthiazide R = Cl, CF hydrochlorothiazide H2NO2S S 2 3 NH

R2 N R1

H

Thiazide analogues R1 = subst. amide mefruside H2NO2SR1 R2 = H, OH xipamide

Cl R2

R Loop diuretics R1 = Cl, phenoxy furosemide 2 (furosemide type) R2 = amine bumetanide R1

H2NO2S COOH

Potassium-sparing diuretics, canrenone O aldosterone antagonists spironolactone O

O

Cycloamidine derivatives triamterene O NH amiloride Cl N N NH2 H

H2N N NH2

permitted a GC separation and subsequent of sensitivity [3, 18]. As stated earlier, there are identification of the majority of diuretics in SIM strongly acidic as well as basic compounds mode. This approach includes analytical limi- among the class of diuretics, requiring either tations, in addition to the requirement for a negative- or positive-ionisation modes, respec- time-consuming derivatisation step. Certain sub- tively. Basic diuretics may well be combined with stances do not form stable, reproducible and other screening procedures (e.g. anabolic uniform methyl derivatives. Chromatographic steroids) to avoid a re-injection of samples in artefacts (Figure 9.1) need to be factored into the different ionisation modes. Alternatively, the screening [17] and there remained at least one option of a scan-to-scan polarity switching was diuretic (benzthiazide) undetectable in GC-MS. utilised to detect both groups of diuretics simul- Therefore, LC-DAD was additionally applied as a taneously [19, 20]. complementary analytical technique. The The relatively high urinary concentrations obvious benefits of LC-ESI-MS are revocation of combined with the sensitivity of the technique derivatisation, improvement of comprehensive- reduces chromatographic separation to a ness, reduction of turn-around times and increase minimum and diminishes turn-around times to a LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 197

Small molecules 197

field of horse testing [22] because of the general β prohibition of all 2-agonists, regardless of their intended action, concentration or adminis- β tration route. Typically, 2-agonists (Table 9.2) are

Dealkylation/ identified in positive ESI mode using the proton- Sulphoxidation ring opening ated molecule as precursor ion. Characteristic fragmentation reactions are losses of the terminal isobutene group, resulting in [M – 56]+ fragments, whether or not combined with losses of water [23]. Approaches to differentiate between inhala- Figure 9.1 Diuretics with thiazide structure are converted tional and prohibited systemic (e.g. oral) appli- by hydrolysis or sulphoxidation during sample transporta- cation of salbutamol were based on quantitative tion, storage or preparation, which may cause analytical examinations (values greater than 1 mg/L were, problems in GC-MS. Respective artefacts need to be fac- according to the WADA regulations, considered tored into LC-MS screening. as an adverse finding) or investigations of salbu- tamol enantiomers. A discrimination function was derived from the higher amounts of the S(+) few minutes. A combination with automated relative to R(–) isomer after oral administration. solid-phase extraction (SPE) may be applied to This approach requires the combination of chiral establish high-throughput LC-MS screening [21]. LC separation with MS detection, whether on- The assay is sufficiently selective to differentiate line or off-line [13]. concentrations above a minimum required per- formance level (MRPL [17]) of 250 µg/L from a blank matrix (Figure 9.2). The potential influence β-Blockers of ion suppression is not critical, because there are no threshold values and the quantification of β-Adrenergic blocking agents (β-blockers) are diuretics is not required in screening analyses. prohibited due to their reduction of heart rate, blood pressure and hand tremor. Doping controls are consequently restricted to competition con- β 2-Agonists trols in particular sports where steadiness is important (archery, shooting, etc.). The classification of these sympathomimetic β-Blockers are characterised by a very similar agents in doping control has been frequently chemical structure. With a few exceptions (e.g. modified. At present, there is a differentiation sotalol or carvedilol), they represent derivatives between substances that are permitted by of oxypropanolamine terminated by t-butyl or inhalation to treat asthma (requiring a thera- isopropyl groups and aromatic substituents peutic use exemption), while others are pro- (Table 9.3). hibited due to their potential stimulating or GC-MS was the conventional screening anabolic effect. Clenbuterol, which is supposed technique for β-blockers in doping control. to be the most potent anabolic β-agonist, consti- Derivatisation with N-methyl-N-trimethylsilyltri- tuted a particular analytical challenge as typical fluoroacetamide (MSTFA), if necessary combined urinary concentrations are of the order of magni- with N-methyl-bis-trifluoroacetamide (MBTFA), tude of 1 µg/L. Due to its two chlorine atoms, it was applied after hydrolysis of the conjugates and is well suited for HR-MS and its trimethylsilyl isolation from urine [24]. The application of (TMS) derivative was preferably identified by LC-MS represents a useful alternative to avoid either GC-HR-MS or GC-MS-MS. Typical LC-MS this time-consuming derivatisation step, and the screening procedures are slightly less sensitive formation of unstable derivatives and artefacts than GC-MS procedures, but more comprehen- in some cases (e.g. acebutolol). All β-blockers sive, which appears to be more important in the contain a secondary amino group accounting for LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 198

198 Chapter 9 • LC-MS in doping control

7.0e4 (a) Intensity (cps)

0.0

2.7e4 (b) Intensity (cps)

0.0 1.0 2.0 Time (min)

Figure 9.2 Multiple-reaction monitoring (MRM) experiments permit the screening of low amounts of prohibited diuretics in negative-ionisation mode. The turnaround time and sample preparation may be reduced to a minimum due to the out- standing selectivity. A mixture of 20 diuretics extracted from a urine matrix (a) is compared with a blank urine containing mefruside as internal standard (IS) (b).

their protonation and detection in positive (M – 77) in addition to the formation of an mode. According to their high structural simi- N-isopropyl-propanolamine fragment (mass-to- larity, there are group-specific fragmentation charge [m/z] ratio 116) [25]. reactions for both classes of β-blockers (Table 9.3). The proposal of a combined screening of Substances with a terminal t-butyl group are diuretics and β-blockers agents, utilising a characterised by loss of isobutene (M – 56), scan-to-scan polarity-switching technique [19], is usually in combination with loss of water, while very promising for clinical purposes, because most of isopropyl terminated compounds both substance groups are frequently combined undergo a loss of an isopropylamino group in the treatment of hypertension. However, this LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 199

Small molecules 199

Table 9.2 Chemical structures of typical β-agonists

Class Typical modifications Examples Chemical structure

Aniline R1, R2 = Cl, Br, CN, OH clenbuterol R1 R = C(CH ) , subst. phenyl brombuterol 3 3 3 OH fenoterol H2N NH

R2 R3

Phenol R1 = OH, subst. alkyl salbutamol R1 R2 = alkyl, subst. phenyl orciprenaline OH

HO NH

R2

Benzazepinone zilpaterol

HN

HO

N

O N H

Table 9.3 Chemical structures of typical β-blocking agents prohibited in particular sports

Class Modifications Examples Chemical structure

Isopropylamine R1 = substituted aromatic rings atenolol acebutolol N propranolol R1 O H OH

t-Butylamine R1 = substituted aromatic rings bupranolol carteolol N timolol R1 O H OH

approach seems to be less rational in doping Steroids control, because the intention of an abuse of both substance classes and the scope of their This substance class is characterised by a uniform prohibition (concerning sports, in competition structure consisting of a modified sterane skele- versus out of competition) are different. ton (Table 9.4). Due to this apolar structure, the LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 200

200 Chapter 9 • LC-MS in doping control

efficiency of ionisation and suitability of LC-MS Synthetic anabolic steroids mainly depend on molecular substitutions, The group of anabolic steroids still includes a resulting in tremendous variations of sensitivity diversity of similar structures, which cannot be among different steroids. Saturated steroid mol- systematically separated. The subgroups listed in ecules (e.g. by saturation of rings or reduction of Table 9.4 (biologically active anabolic steroids, keto groups) may hardly be ionised by protona- precursors and metabolites) overlap each other tion or formation of ammonium adducts. The from legal as well as biochemical perspectives. presence of conjugated double bonds (3-keto-4- These compounds (e.g. androstenedione) may be ene, steroids), oxidation of the sterane moiety or considered as precursors and metabolites in the other polar ring substitutions (e.g. the pyrazol biosynthesis of steroids, but they also originate ring in stanozolol) improves the ionisation rate from prohibited application of synthetic ana- significantly. Therefore, this apparently uniform logues (pro-hormones) of endogenous steroids or substance class needs to be divided into the even from synthetic hormones. The analytical following groups accounting for their LC-MS result of a urine analysis does not necessarily properties. allow the differentiation of an approved medication, the abuse of pro-hormones (legally

Table 9.4 Chemical structures of selected endogenous and synthetic steroids

Class Typical modifications Example Chemical structure OH Endogenous oxidation/reduction intestosterone 12 steroids (including positions 3/17, saturation 17 their precursors of 4 double bond, 5-α/β 19 C D 16 and metabolites) isomers 1 2 A B 3 5 O 4 6

OH Synthetic 17α alkylation, double oxandrolone CH steroids bonds at position 1–2, 4–5 3 or 5–6, A-ring condensation, 4-chlorination, 19-demethylation O

O

Esters acetate, propionate, testosterone O decanoate esters O

(CH2)nCH3

O LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 201

Small molecules 201

tolerated in certain countries) or the application comprehensive and sensitive LC-MS methods for of scheduled prohibited steroids. the identification of anabolic steroids. Appli- The attempts to develop LC-MS methods for cation in routine steroid analysis is mainly the identification of anabolic steroids of the so- focused on selected polar steroid molecules called ‘free fraction’ (i.e. slightly polar steroids (Figure 9.4), e.g. stanozolol [26, 30], boldenone which are excreted unconjugated in urine) [31], trenbolone [32, 33] or THG [29]. These polar demonstrated significant diversity of ionisation steroids often encounter analytical difficulties in principles [26]. All anabolic steroids are detected GC-MS, due to the formation of instable deriva- in positive mode. However, the appearance of the tives and artefacts after silylation. The appli- most abundant precursor ion depends signifi- cation of an additional derivatisation (e.g. cantly on structural modifications of the sterane methoxime derivatives of THG) followed by an skeleton and is almost unpredictable. Protonated extra GC-MS procedure would be required to + + ions [M + H] , ammonium adducts [M + NH4] identify these substance groups. Therefore, LC- and fragments resulting from loss of up to three MS is a beneficial alternative to the extra effort + molecules of water [M – nH2O] were reported as of additional derivatisations or modified GC base peaks. The balance between these ions is techniques. determined by proton affinity and therefore conjugated double bonds (3 keto-4-ene steroids, Endogenous steroids pyrazol ring condensation, aromatic rings) are the most obvious structural indicators for an The problem of endogenous steroids in drug improved protonation. Physical conditions in testing human athletes is mainly reduced to the the ion source appear to represent another sensi- quantitative balance between testosterone and its tive factor of ionisation efficiency. Comparing biochemical byproduct epitestosterone (see ESI, APCI and APPI for various anabolic steroids, ‘Steroid conjugates’). In addition, there is an it turned out that the choice of a precursor upcoming interest in the quantification of depends significantly on the ionisation tech- endogenous steroids related to the therapeutic nique. Oxandrolone (Table 9.4), for example, administration of steroids (treatment of testo- formed predominantly the [M + H]+ ion in ESI, sterone deficiency [34]) and to their increasing + + the adduct [M + NH4] in APCI and the [M – H2O] relevance as lifestyle drugs. Therefore, the quan- fragment in APPI [27]. titative evaluation of endogenous steroid profiles Moreover, there is no general consensus about using LC-MS-MS is of increasing importance in the preferences of various ionisation techniques clinical chemistry. for steroid analysis (Figure 9.3). Controversial The growing number of precursors of en- evaluations of APPI [27, 28] suggest that techni- dogenous anabolic steroids (so-called pro- cal differences between manufacturers’ concepts hormones) that are widely available as nutrition are more significant than physical constraints. supplements and abused in sports and body- The typical acquisition mode of these methods is building constitutes another analytical chal- MRM, including two to three fragmentation lenge. Biochemical precursors of testosterone reactions per analyte. LC-MS has contributed to (e.g. androstenedione, dehydroepiandrosterone) the structural elucidation in the case of the new were the first products on the market, but they upcoming anabolic steroid tetrahydrogestrinone have been recently replaced by synthetic steroids, (THG) (Figure 9.3) [29] and provided an efficient representing slight structural modifications of complementary screening method, directed to the endogenous compounds (e.g. 1-testosterone, the identification of polar steroids (e.g. tren- 1,5α-androstenedione, where the location of the bolone, THG, stanozolol). double bond is shifted from the 4 to the 1 posi- However, unpredictable formation of pre- tion). Analytical problems to identify these cursor ions, relatively low ion abundances, and compounds by LC-MS are comparable to those unspecific fragmentation reactions in combin- encountered for their endogenous counterparts ation with the existence of numerous isomeric [35]. The technique is probably not sufficient for steroids and metabolites complicate the design of the unambiguous identification of all potentially LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 202

202 Chapter 9 • LC-MS in doping control

(a)

XIC of +MRM: 313.2/159.1 amu 5.31 8.5e4 Max. 8.7e4 cps THG

OH

O M=312.1 Intensity (cps)

0.0 1.0 2.0 3.0 4.0 5.0 6.0 Time (min) (b) XIC of +MRM: 313.2/159.1 amu 1.8e4

Epitestosterone-d3 IS Intensity (cps)

Max. 2250.0 cps

5.52

0.0 1.0 2.0 3.0 4.0 5.0 6.0 Time (min)

Figure 9.3 Comparison of LC-MS-MS detection of epitestosterone and THG in (a) APPI and (b) ESI. Both chromatograms were run under identical LC conditions. APPI shows an outstanding sensitivity for detection of THG, but is limited to steroids with chromophoric molecular structures, while ESI appears to be the most versatile ionisation technique. LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 203

Small molecules 203

269→161, CH3 3e4 269→105 OH

HO H Epimetendiol 6.76 Intensity (cps)

0.0

6.01 2e5 345→345, 345→97 OH CH OH 3

N HN

5.65 3′-OH-Stanozolol Intensity (cps)

5.15

0.0 5.0 6.0 7.0 8.0 Time (min)

Figure 9.4 Threshold concentrations of 2 µg/L of 3-OH-stanozolol and epimetendiol in extracts from a urinary matrix. The polar stanozolol metabolite is exceptionally well ionised and permits a sensitive identification, while detection of the medium polar epimetendiol is hampered by a limited intensity and specificity.

relevant steroids, but represents a helpful supple- urine as such and, therefore, not included in ment to GC-MS. routine doping analysis. The general benefit of the identification of steroid esters is an unequivocal proof of illegal Steroid esters administration of a synthetic compound. Any Steroids are typically administered as fatty acid trace amount of the exogenous esters provides a esters by intramuscular injection. These com- clear indication for a doping offence, while the pounds are stored in adipose tissues, rapidly corresponding free steroids need to be distin- hydrolysed in blood, not markedly excreted in guished from natural endogenous levels. LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 204

204 Chapter 9 • LC-MS in doping control

LC-APCI-MS-MS was reported to facilitate a with glucuronidase from Helix pomatia, which sensitive identification of testosterone [36] or does not cleave the corresponding sulphates. 19-nortestosterone esters [37] in equine plasma. Logically, the testosterone/epitestosterone ratios The formation of dominant [M + H]+ precursor are by consensus referred to the total amount of ions was reported for all steroid esters. free and glucuronidated steroids. It was suggested that elevated relative amounts of epitestosterone sulphate may cause increased testosterone/ Steroid conjugates epitestosterone ratios when measured by conven- The detection of the abuse of endogenous tional GC-MS. A potential racial bias of phase 2 steroids (testosterone or its precursors) is based biotransformation was discussed. on a quantitative evaluation of the urinary The quantification of intact conjugates, which concentrations of testosterone and epitestos- is made possible by LC-MS, seems to be a con- terone. The latter steroid is a byproduct of the clusive option to examine the individual influence biosynthesis of steroids and supposed to be of glucurono- and sulpho-conjugation on testos- suppressed after the administration of steroids. terone/epitestosterone ratios directly and circum- According to WADA criteria, further investi- vent the uncertainty of hydrolysis recovery [38, gations to exclude physiological deviations are 39]. Steroid glucuronides were reported to form mandatory, if the ratio of testosterone/epitestos- different adducts (protonated, ammoniated and terone exceeds a value of 6. This definition is sodiated ions) in positive ESI. Depending on the mainly empirical and derived from the conven- declustering potentials, the [M + H]+ was found to tional analytical GC-MS procedure, which was be most suitable due to its high signal-to-noise traditionally based on the quantification of bis- (S/N) ratios. Reasonably specific fragments TMS derivatives of both epimers. Hydrolysis of (aglycone and its singly or doubly dehydrated the conjugates (mainly glucuronides and sul- fragment ions) were chosen for sensitive MRM phates, Table 9.5) was carried out after hydrolysis experiments. The use of a deuterated IS for each

Table 9.5 Chemical structures of steroid conjugates

Conjugation Examples Chemical structure

Glucuronides androsterone O glucuronide COOH

HO O

HO O

OH

OSO H Sulphates testosterone 3 sulphate

O LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 205

Small molecules 205

conjugate was found to be mandatory due to the produce considerable higher amounts of the potential-ion suppression in the urinary matrix. dehydrated aglycone fragments (relative to the The assumed influence of sulphation on elevated deconjugated steroid) than 5α isomers [41]. testosterone/epitestosterone ratios was not sup- ported by LC-MS-MS examinations [39]. Corticosteroids Quantitative uncertainties of conjugate hydrolysis are less crucial in the case of exogenous Glucocorticosteroids (Table 9.6) are prohibited steroids, because anti-doping legislation does not when administered systemically (orally or by require a quantitative threshold. Approaches to injection), whereas all other administration identify glucuronides of synthetic steroids by LC- routes require medical notification. Due to their MS were directed to structural investigations of influence on protein and carbohydrate metab- steroid glucuronides and automation of the ana- olism, they are known to be abused as growth lytical process by application of on-line micro- promoters in food-producing animals and may extraction [40]. Steroid glucuronides proved to reasonably be abused in sports. At the moment, exhibit similar ionisation principles to their free there is no analytical solution for a reliable differ- analogues. The presence of conjugated double entiation of the administration pathway and ana- bonds increases the proton affinity, resulting in an lytically positive cases may be rejected by the elevated abundance of [M + H]+ pseudo-molecular availability of a therapeutic use exemption. ESI in ions, while saturated molecules tend to form negative mode was consistently reported to be ammonium adducts rather than protonated ions. the most efficient ionisation for a corticosteroid The steroid conformation (5α versus 5β linkage of screening method [42, 43]. Typical fragmentation

the A and B rings of the sterane skeleton) was reactions are loss of the CH2OH moiety, water or reported to influence the fragmentation of A- hydrofluoric acid. The LC-MS assays are carried ring-saturated steroids; 5β isomers were found to out either in single-MS mode using diagnostic

Table 9.6 Chemical structures of synthetic corticosteroids

Class Typical modifications Examples Chemical structure

Cortisone 1–2 double bond prednisone OH O OH

O

Cortisol 1–2 double bond prednisolone O R R1 = H, OH betamethasone HO 5 R = H, OH, CH triamcinolone R 2 3 1 R , R = C(CH ) R 1 2 3 2 4 R 2 R3, R4 = H, F

O

R 3 LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 206

206 Chapter 9 • LC-MS in doping control

fragments or in MS-MS mode, where [M + ecgonine) are comparable candidates for an – CH3COO] adducts serve as precursor ions due to insertion into LC-MS screening procedures. the absence of [M – H]– pseudo-molecular ions. The identification of the major urinary Alternatively, the identification of intact metabolite of ∆9-tetrahydrocannabinol (THC) corticosteroid conjugates in bovine urine was (the glucuronide of 11-nor-9-carboxy-THC) in evaluated [44]. Base peaks detected in positive ESI doping control does not represent any particular mode are sodium adducts, while a predominant [M exception compared with forensic urine analysis –H]– ion is observed in negative mode. Both pre- (see Chapter 8). A threshold value of 15 µg/L cursors exhibit a low fragmentation rate. MRM specifies a potential doping violation; however, experiments in negative-ionisation mode using a there are no uniform sanctions and cannabinoids [M–H]– → [M–H]– pseudo-transition (monitored were downgraded to the group of ‘specified sub- at elevated collision energy) were found to produce stances, which are particularly susceptible to specific signals due to the high ion stability. unintentional anti-doping rule violations . . . and which are less likely to be successfully abused as doping agents’ [1]. Any eligible LC-MS assay for Other prohibited substances (stimulants, cannabinoids [46] may be applied in sports drug narcotics, cannabinoids) testing without modifications. The identification of alkylated xanthine deriv- There are numerous other prohibited substances atives is no longer essential in human doping in doping control eligible for application of LC- analysis after its removal from the list of pro- MS, like stimulants or narcotics. Respective hibited substances in 2004, although recent analytical procedures are available in clinical, publications have dealt with its identification veterinary or forensic toxicology and may cer- and quantification by LC-MS in horses [47, 48]. tainly be adopted in doping analysis. However, According to the required performance specifica- conventional screening procedures based on GC tion of the Association of Official Racing Chem- with nitrogen phosphorous-selective detection ists (AORC), concentrations as low as 100 µg/L (NPD) or GC-MS are still state of the art, because need to be detected in equine urine. the concentration and stability of respective Among numerous pharmaceutical substances compounds are sufficiently high. Moreover, the that are controlled in equine doping control, list of prohibited stimulants and narcotics has there are quaternary ammonium anti-cholinergic been revised and condensed recently, simplifying agents (e.g. isopropamide, glycopyrrolate) that the analytical challenges. appear to be well suited for LC-MS. According to Analytical development in these areas is their ionic structure, quaternary ammonium focused on upcoming (e.g. modafinil) or critical drugs require special pre-treatment (ion pair for- substances (e.g. mesocarb). The latter stimulant is mation) to enable conventional liquid–liquid thermally instable, forms irreproducible artefacts extraction (LLE) or LC separation. Examining the and is therefore difficult to identify by conven- identification of eight quaternary ammonium tional GC-MS. The long history of attempts to use drugs in horse urine, the application of capillary various ionisation techniques (particle beam, electrophoresis was found to be more appropriate thermospray, ESI, APCI) to detect mesocarb and (enhanced sensitivity and separation power) elucidate its biotransformation [45] demonstrates than LC [49]. the analytical difficulties when tackling this sub- stance. Mesocarb does not fit analytically into its native pharmacological group of stimulants and Large molecules is typically integrated into other LC-MS screening procedures (e.g. diuretics). Other isolated substances causing GC prob- Introduction lems due to their high polarity and/or low thermal stability (e.g. the anti-oestrogenic Several proteins and peptide hormones are clomiphene or the cocaine metabolite benzoyl- included in the list of prohibited substances, LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 207

Large molecules 207

due to their anabolic effect, as in the case of linear ITs) are used for research investigations, hGH or hCG, or due to an increase in the oxygen but are not available for routine applications. transportation capacity of blood, like EPO and The identification of macromolecular com- haemoglobin-based oxygen carriers (HBOCs). pounds by LC-MS in routine doping analysis is so These substances are synthetic or recombinant far restricted to HBOCs and synthetic insulin. analogues of endogenous hormones. The possi- bility to discriminate between an abuse of pro- hibited substances and endogenous production Human chorionic gonadotrophin depends on the structural modifications of the respective exogenous compound. Possible ana- This gonadotrophic hormone stimulates the lytical approaches are: endogenous production of testosterone and is therefore prohibited in men. Quantification of • quantitative evaluations of compounds with a hCG by two different immunological assays is significant concentration difference between recognised by WADA as sufficient analytical tech- basal levels and exogenous administrations nique. Different conventional cut-off values of (hCG, IGF-1) 10 or 25 IU/L were suggested (but not officially • identification of variations of the primary adopted) to discriminate normal values and structure (insulin, HBOCs) pathological situations or misuse. An LC-MS pro- • discrimination of mass variants of proteins cedure using an IT MS in positive ESI mode was (20- and 22-kDa isomers of hGH) found to be suitable for a sensitive quantification • investigation of charge variants resulting from procedure of hCG [50]. Immunoaffinity extrac- post-translational modifications (e.g. glyco- tion was used to enrich the peptide from the sylation of EPO). urinary matrix prior to the tryptic digestion of In fact, the analytical approach depends on the the glycoprotein. A residue of the β-subunit con- current availability of pharmaceutical substances taining 17 amino acids was chosen as a signifi- which are potentially abused. EPO, for instance, cant marker for hCG; in particular, a distinct was originally available as recombinant protein structural specificity compared with the similar with a primary structure identical to the en- subunit of the luteinising hormone was con- dogenous hormone. The introduction of syn- firmed. MRM experiments of the doubly charged thetic variants with modified amino acid peptide allowed a quantification down to thresh- sequences required an immediate adaptation of old concentrations of 5 IU/L. Examination of the the analytical approach. glycosylation of hCG revealed a considerable The practical application of LC-MS techniques micro-heterogeneity [51], which does not affect to identify large molecules in routine doping the evaluation of doping cases because a reliable control is still restricted by sensitivity limitations differentiation is possible based on a wide con- and the requirement of extensive and selective centration gap between endogenous and abnor- sample clean-up. A shotgun approach based on mal hCG levels in men. digestion of proteins followed by HR separation of the digests is most frequently applied. Selective clean-up procedures (e.g. immunoaffinity enrich- Human growth hormone ment) of the resulting complex peptide mixtures need to be applied to reduce ion suppression and Human GH stimulates the production of IGF-I, to be able to identify relevant proteins at low leading to a promotion of protein synthesis, amounts. increase of muscle mass, reduction of the amount The characterisation of structural particulari- of stored fat and an induction of the growth of ties based on the identification of intact proteins long bones. The secretion from pituitary glands requires relatively high amounts of sample occurs in three to five daily pulses, during which material and HR-MS for identification of the the basal serum concentrations (around 3 µg/L) multiply charged precursor ions. Adequate ana- are temporarily greatly exceeded and return lytical techniques (e.g. FTICR MS combined with rapidly (half-life around 15 min) to normal. LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 208

208 Chapter 9 • LC-MS in doping control

Therefore, hGH serum concentrations are not raphy [10]. The detection was carried out by HR- suitable for the detection of its abuse. Instead, a LC-MS (FTICR), which was coupled to a linear IT. combined evaluation of concentrations of IGF-I, In another study, qualitative investigations on its binding protein (IGF-BP3) and bone markers hGH were carried out using LC-MS-MS (FTICR). was proposed to reveal abuse of growth hormone Based on the isolation and accurate mass measur- in human athletes [52]. ing of the [M + 17H]17+, a deviation of the In horses, a cut-off serum concentration of molecular mass of 4 Da (compared with non- 700 µg/L of IGF-I was anticipated as the criterion post-translationally modified hGH) could be for growth hormone administration [53]. A sen- detected and attributed to the formation of two sitive LC-ESI-MS quantification procedure (limit disulphide bonds [56]. of quantification = 30 µg/L IGF-I) was reported to The identification of structural modifications be eligible for routine screening. The use of Arg3- like disulphide linkage or variations of glycosyl- IGF-I, which is derived from IGF-I by replacement ation, influencing proper folding of proteins, of Glu3, as IS, appears to be essential, due to its may be of diagnostic value in doping analysis. similar properties in the affinity chromatography Present technical constraints (relatively high cut- clean-up. Full-scan MS was applied to detect the off values and/or high amounts of sample mater- intact pseudo-molecular ion ([M + 7H]7+, m/z ial, requirement of high mass ranges and MS 1093.4) of IGF-I, which is used for quantification resolution, and high MS accumulation times that in equine serum samples [53]. are incompatible with typical LC peak widths and Another analytical approach is based on the prohibit on-line coupling) seem to impede the evaluation of qualitative variations of growth application of LC-MS to the identification of hormone. There are several mass variants of the hGH abuse in the near future. predominant protein containing 191 amino acids corresponding to a molecular mass of 22 kDa. The most interesting one is a 20-kDa Erythropoietin molecule, which is derived by deletion of residues 32–46 and exhibits comparable physiological EPO is a 34-kDa glycoprotein hormone control- activity. The concentration ratio of both variants ling red blood cell production and is therefore a does not depend significantly on gender, age, potent doping agent to enhance the oxygen body height and weight [54]. Administration of transport capacity of blood. Recombinant human any hGH isoform suppresses the endogenous EPO (rhEPO) has been available since 1989 and secretion of both variants. Accordingly, abuse of cannot be directly distinguished from en- commercially available hGH may be detected by dogenous EPO due to its identical amino acid an elevated 22/20 kDa ratio, because it contains sequence. Indirect methods (e.g. haematological exclusively the 22-kDa variant. Studies to prove parameters) served as an indicator for a potential the appropriateness of this approach based on abuse. The variable glycosylation is determined enzyme-linked immunoassays are documented by post-translational modification, which [55]. Current alternative quantification pro- depends on the availability of enzymes in the cedures of low-level proteins by LC-MS, based on respective cells, resulting in a wide diversity of sophisticated analytical equipment, e.g. FTICR carbohydrates, typically terminated by one or MS, require extensive sample preparation and do two sialic acid residues. not yet achieve the required sensitivity [10, 56]. The number and location of sialic acid residues Identification of two tryptic peptides of hGH per molecule (Figure 9.5) determine the formation (22 kDa), which was obtained from digestion of of quaternary isoforms and influence the plasma LC fractionated plasma, provided limits of detec- half-life of EPO. Removal of sialic acid groups tion 10-fold higher than relevant clinical plasma leads to a total inactivity, due to a rapid clearance concentrations [57]. Another approach was from blood circulation, while new EPO variants capable of identifying clinically relevant levels (i.e. darbepoetin) provide a prolonged activity, (5 µg/L) from unfractionated plasma, applying triggered by the inclusion of additional oligo- tagged peptides, isolated by affinity chromatog- saccharides, which are terminated by sialic acid. LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 209

Large molecules 209

1 165 F L-Fucose G D-Galactose F Asn-24 M N G S M D-Mannose N N M M N G S N N-Acetylglucosamine N G S N-Acetylneuraminic acid (sialic acid) Asn-38 N F N Ser-126 G S M S M M N N N N G G G G S SSN G S G N S F S G N M Asn-83 M N N S G N M G N

Figure 9.5 Chemical structure of EPO.

This alteration is based on a modification of five columns [60]. An LC-ESI-MS investigation of amino acids of the polypeptide and respective various commercially available EPO forms substances may be distinguished from natural (epoetin α, epoetin β, darbepoetin) resulted in EPO by analysis of the primary structure. Another astonishingly good LC separation. The full-scan variant of EPO (SEPO) consists of a synthetic MS of the three variants shows characteristic par- peptide that is conjugated with a polymer, ticularities in ESI (on-line micro LC coupling) as aiming for a reduction of the biological diversity. well as in matrix-assisted laser desorption ionisa- The identification of EPO variants in doping tion (MALDI) mode (examination of the LC frac- control urine samples is to date based on a chemi- tions) [61]. The injected quantity of protein was luminescence detection of its isoforms after about 100 ng and, hence, far beyond the amount immunological enrichment and electrophoretic available in routine analyses. The limited MS separation [58]. The logical MS approaches to an sensitivity is due to the heterogeneity of glyco- identification in combination with LC or capil- sylation, because the ion abundance is spread lary electrophoresis are hampered by the limited over a large number of individual molecules. A sensitivity. deglycosylation of the EPO molecule leads to a Application of capillary electrophoresis- and reduction of diversity and increase of MS sensitiv- LC-IT MS enabled the quantitative characterisa- ity. Analyses of deglycosylated EPO (rhEPO and tion of EPO reference material after cleavage darbepoetin) by MALDI showed a good match of and derivatisation. The relative amount of sialic its molecular mass with the assumed structure acid per molecule of rhEPO was found to be [62]. LC-ESI-MS experiments allowed a sensitive 17.6 mol/mol. This was supposed to provide a identification of [M + nH]n+ pseudo-molecular potential parameter to differentiate recombinant ions (n = 5–8) of rhEPO after application of from endogenous EPO [59]. Alternatively, a 250 fmol on-column. The additional identifi- characteristic sulphation of N-linked oligo- cation of peptide, obtained from endoprotease saccharides in EPO molecules from different cell Glu-C digestion, provides complementary struc- lines was detected, based on MS examinations tural information and is supposed to permit a (negative-ionisation mode) after LC separation of sufficient identification of rhEPO and darbe- the protein digest using graphitised carbon poetin in race horses and greyhounds [62]. LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 210

210 Chapter 9 • LC-MS in doping control

Insulin separation. Alternatively, the identification of the B chains can be applied after reduction of the The presumptive abuse of insulin among athletes disulphide bonds. and body-builders is due to its stimulation of endogenous protein synthesis. Several cases of hypoglycaemia (i.e. significantly reduced levels Haemoglobin-based oxygen carriers of blood sugar) were observed in fatal incidents in body-building, indicating its high popularity in Similar to the administration of EPO or synthetic this field. Insulin (Figure 9.6) belongs to the perfluorocarbons, the intention of an abuse of group of prohibited peptide hormones according HBOCs in sports is an elevation of the oxygen to the recent WADA definition [1]. There are transportation capacity of blood. Various various medications available, based on struc- cross-linked bovine or human haemoglobin tural alterations of insulin. The main purpose of preparations (e.g. Hemopure, Hemolink, Oxy- these structural variations consists in the regu- globin) are approved for the treatment of animals lation of its bioavailability. Insulin tends to self- or humans. The intact oxyglobin was reported association to biologically inactive hexamers, to be inadequate for an efficient confirmation of which can be suppressed by structural modifica- haemoglobin by LC-MS in blood samples, tions like switching of the positions of lysine and because of the formation of one single dominant proline (B28 versus B29, Humalog) or replace- fragment that was interfered with by impurities ment of proline with an aspartic acid residue from haemolysed plasma. Sufficient variations (B29, Novolog). Alternatively, long-acting between the primary structures of human and insulin was synthesised by elevation of its iso- bovine haemoglobin permit the differentiation electric point (modification in A and B chains, of both proteins, based on the identification of Lantus). diagnostic tryptic peptides. Specific peptides The identification of intact insulin derivatives, resulting from α and β chains of haemoglobin carried out on a QTOF instrument equipped with were found to be suitable for a differentiation of a nano-ESI ion source, revealed the predominant both species [64, 65]. Subunits of the α (residue formation of corresponding multiply charged 69–90, 2367 Da) and the β chain (residue 40–58, molecules (z = 5–7). This enables a mass-specific 2090 Da) were commonly used as markers for differentiation of the synthetic derivatives from bovine haemoglobin. Human haemoglobin was human insulin [63] (Figure 9.7) after LC identified by recording of a diagnostic peptide

Tyr Gln A-chain Leu Leu Ser Glu 11 Asn Cys Ile Tyr 21 Cys Asn − Ser COO Cys Gln Cys Thr Glu Val Gly Leu Cys Cys 21 Glu Ile His Val 1 Gly COO− Gly Arg Gln Ser Leu Thr NH + Gly 3 Asn His Tyr Lys 11 Phe Val Leu Leu Pro Phe 1 Phe Val Ala B-chain Thr + Glu Tyr NH3

Figure 9.6 Chemical structure of insulin. LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 211

Large molecules 211

[M + 6H]6+ 968.86 (a)

20.0%

18.0%

16.0%

14.0%

12.0%

10.0%

8.0% Percentage abundance Percentage + 5+ 6.0% [M 5H] 1162.42 4.0%

7+ [M + 7H] 966.03 2.0% 830.60 1167.01 0.0% 800850 900 950 10001050 1100 1150 12001250 1300 m/z 100% (b) 6.14 Humalog

Lantus

6.05

6.14 Novolog Relative abundance Relative

5.5 5.6 5.7 5.8 5.9 6.0 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7.0 Retention time (min)

Figure 9.7 ESI full-scan spectrum of human insulin (a) generating the multiply charged molecules [M + 5H]5+, [M + 6H]6+ and [M + 7H]7+ at m/z 1162.4, 968.9 and 830.6, respectively. Extracted ion chromatogram (b) of a plasma sample fortified with Humalog (m/z 1162), Novolog (m/z 1166) and Lantus (m/z 1213) at 10 pmol/mL. Insulins were analysed as intact proteins. (Reproduced from Thevis et al. [63] with permission.) LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 212

212 Chapter 9 • LC-MS in doping control

from the β chain (residue 42–60, 2059 Da). 2. Houghton E, Dumasia MC, Wellby JK. The use of Doubly or triply charged ions of these peptides combined high performance liquid chromatog- are suitable for LC-ESI-MS-MS screening. The raphy negative ion chemical ionization mass assay is able to identify, confirm and quantify spectrometry to confirm the administration of HBOCs in human and equine plasma samples. synthetic corticosteroids to horses. Biomed Mass Spectrom 1981; 8: 558–564. Respective assays were established using either QTOF or triple-quadrupole (QqQ) MS [64, 65]. 3. Garbis SD, Hanley L, Kalita S. Detection of thiazide- Sample preparation requires a laborious combi- based diuretics in equine urine by liquid chroma- nation of SPE, filtration of macromolecules and tography/mass spectrometry. J AOAC Int 1998; 81: enzymatic digestion. 948–957. 4. Bowers LD. Analytical advances in detection of per- formance-enhancing compounds. Clin Chem 1997; 43: 1299–1304. Summary 5. Bowers LD, Fregien K. HPLC/MS confirmation of peptide hormones in urine: an evaluation of detec- LC-MS has been introduced in different branches tion limits. In: Donike M, Geyer H, Gotzmann A, of doping analysis. et al., eds, Recent Advances in Doping Analysis 6. Qualitative progress in the analysis of highly Cologne: Sport und Buch Strauß, 1994: 175–184. polar substances (diuretics) could be achieved in 6. Thieme D, Große J, Lang R, et al. Detection of combination with a significant reduction of mesocarb metabolite by LC-TS/MS. In: Donike M, sample preparation. Other low-mass pharma- Geyer H, Gotzmann A, et al., eds, Recent Advances in ceutical substances (stimulants, narcotics, β- Doping Analysis 2. Cologne: Sport und Buch Strauß, blockers) may certainly be analysed by LC-MS, 1995: 275–284. but well-established and validated GC assays are 7. Ventura R, Fraisse D, Becchi M, et al. Approach to still dominant. the analysis of diuretics and masking agents by Progress in steroid analysis is mainly focused high-performance liquid chromatography-mass on substances with high proton affinity (polar spectrometry in doping control. J Chromatogr 1991; substituents, conjugated double bonds, corti- 562: 723–736. costeroids, steroid conjugates). The identification 8. Krummen M, Hilkert AW, Juchelka D, et al. A new of relevant steroids by LC-MS has considerably concept for isotope ratio monitoring liquid advanced the classic GC-MS screening for chromatography/mass spectrometry. Rapid anabolic steroids, but an extensive replacement Commun Mass Spectrom 2004; 18: 2260–2266. of the conventional surveys is not likely. Finally, there is increasing application of 9. Greig MJ, Bolanos B, Quenzer T, et al. Fourier trans- form ion cyclotron resonance mass spectrometry LC-MS to the analysis of doping-relevant pro- using atmospheric pressure photoionization for teins. Synthetic (e.g. insulin) or non-human (e.g. high-resolution analyses of corticosteroids. Rapid haemoglobin) peptides may be positively dis- Commun Mass Spectrom 2003; 17: 2763–2768. criminated from endogenous analogues, whereas recent approaches to an unequivocal identifi- 10. Wu SL, Choudhary G, Ramstrom M, et al. Evalu- ation of shotgun sequencing for proteomic analysis cation of abuse of recombinant EPO or growth of human plasma using HPLC coupled with either hormone are promising, but so far insufficiently ion trap or Fourier transform mass spectrometry. sensitive and reliable for routine application. J Proteome Res 2003; 2: 383–393. 11. Magnusson MO, Sandstrom R. Quantitative analy- sis of eight testosterone metabolites using column References switching and liquid chromatography/tandem mass spectrometry. Rapid Commun Mass Spectrom 2004; 18: 1089–1094. 1. World Anti-Doping Agency. The Prohibited List 2004. International Standard. Montreal: WADA, 12. Kawasaki N, Ohta M, Itoh S, et al. Usefulness of 2004. sugar mapping by liquid chromatography/mass LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 213

References 213

spectrometry in comparability assessments of glyco- methods for beta2-agonists in human or equine protein products. Biologicals 2002; 30: 113–123. urine. J Mass Spectrom 2003; 38: 1197–206. 13. Berges R, Segura J, Ventura R, et al. Discrimination 24. Segura J, Ventura R, Jurado C. Derivatization pro- of prohibited oral use of salbutamol from cedures for gas chromatographic-mass spectro- authorized inhaled asthma treatment. Clin Chem metric determination of xenobiotics in biological 2000; 46: 1365–1375. samples, with special attention to drugs of abuse and doping agents. J Chromatogr B 1998; 713: 14. Nelson RE, Grebe SK, DJ OK, et al. Liquid chroma- 61–90. tography-tandem mass spectrometry assay for simultaneous measurement of estradiol and 25. Thevis M, Opfermann G, Schanzer W. High speed estrone in human plasma. Clin Chem 2004; 50: determination of beta-receptor blocking agents in 373–384. human urine by liquid chromatography/tandem 15. Kauppila TJ, Kostiainen R, Bruins AP. Anisole, a new mass spectrometry. Biomed Chromatogr 2001; 15: dopant for atmospheric pressure photoionization 393–402. mass spectrometry of low proton affinity, low 26. Leinonen A, Kuuranne T, Kotiaho T, et al. Screening ionization energy compounds. Rapid Commun Mass of free 17-alkyl-substituted anabolic steroids in Spectrom 2004; 18: 808–815. human urine by liquid chromatography-electro- 16. Ventura R, Segura J. Detection of diuretic agents in spray ionization tandem mass spectrometry. doping control. J Chromatogr B 1996; 687: 127–144. Steroids 2004; 69: 101–109. 17. World Anti-Doping Agency. Technical Document 27. Leinonen A, Kuuranne T, Kostiainen R. Liquid TD2004 MRPL. Montreal: WADA, 2004: 1–2. chromatography/mass spectrometry in anabolic steroid analysis – optimization and comparison of 18. Thieme D, Grosse J, Lang R, et al. Screening, confir- three ionization techniques: electrospray ioniza- mation and quantification of diuretics in urine for tion, atmospheric pressure chemical ionization doping control analysis by high-performance and atmospheric pressure photoionization. J Mass liquid chromatography-atmospheric pressure Spectrom 2002; 37: 693–698. ionisation tandem mass spectrometry. J Chromatogr B 2001; 757: 49–57. 28. Guo T, Chan M, Soldin SJ. Steroid profiles using liquid chromatography-tandem mass spectrometry 19. Deventer K, Van Eenoo P, Delbeke FT. Simultaneous with atmospheric pressure photoionization source. determination of beta-blocking agents and diuret- Arch Pathol Lab Med 2004; 128: 469–475. ics in doping analysis by liquid chromatography/ mass spectrometry with scan-to-scan polarity 29. Catlin DH, Sekera MH, Ahrens BD, et al. Tetra- switching. Rapid Commun Mass Spectrom 2004; 19: hydrogestrinone: discovery, synthesis, and detec- 90–98. tion in urine. Rapid Commun Mass Spectrom 2004; 18: 1245–1249. 20. Deventer K, Delbeke FT, Roels K, et al. Screening for 18 diuretics and probenecid in doping analysis by 30. Draisci R, Palleschi L, Marchiafava C, et al. Confir- liquid chromatography-tandem mass spectrom- matory analysis of residues of stanozolol and its etry. Biomed Chromatogr 2002; 16: 529–535. major metabolite in bovine urine by liquid chromatography-tandem mass spectrometry. J 21. Goebel C, Trout G, Kazlauskas R. Rapid screening Chromatogr A 2001; 926: 69–77. method for diuretics in doping control using automated solid phase extraction and liquid 31. Draisci R, Palleschi L, Ferretti E, et al. Confirmatory chromatography-electrospray tandem mass spec- analysis of 17beta-boldenone, 17alpha-boldenone trometry. Anal Chim Acta 2003; 502: 65–74. and androsta-1,4-diene-3,17-dione in bovine urine by liquid chromatography-tandem mass spectrom- 22. Lehner AF, Harkins JD, Karpiesiuk W, et al. Clenbu- etry. J Chromatogr B 2003; 789: 219–226. terol in the horse: confirmation and quantitation of serum clenbuterol by LC-MS-MS after oral and 32. Buiarelli F, Cartoni GP, Coccioli F, et al. Determina- intratracheal administration. J Anal Toxicol 2001; tion of trenbolone and its metabolite in bovine 25: 280–227. fluids by liquid chromatography-tandem mass spectrometry. J Chromatogr B 2003; 784: 1–15. 23. Thevis M, Opfermann G, Schanzer W. Liquid chromatography/electrospray ionization tandem 33. Van Poucke C, Van Peteghem C. Development mass spectrometric screening and confirmation and validation of a multi-analyte method for the LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 214

214 Chapter 9 • LC-MS in doping control

detection of anabolic steroids in bovine urine with confirmation of synthetic corticosteroids in doping liquid chromatography-tandem mass spectrom- urine samples by liquid chromatography-electro- etry. J Chromatogr B 2002; 772: 211–217. spray ionisation mass spectrometry. J Chromatogr A 2001; 926: 87–95. 34. Wang C, Catlin DH, Demers LM, et al. Measure- ment of total serum testosterone in adult men: 44. Antignac JP, Le Bizec B, Monteau F, et al. Study of comparison of current laboratory methods versus natural and artificial corticosteroid phase II liquid chromatography-tandem mass spectrom- metabolites in bovine urine using HPLC-MS/MS. etry. J Clin Endocrinol Metab 2004; 89: 534–543. Steroids 2002; 67: 873–882. 35. Reilly CA, Crouch DJ. Analysis of the nutritional 45. Appolonova SA, Shpak AV, Semenov VA. Liquid supplement 1AD, its metabolites, and related chromatography-electrospray ionization ion trap endogenous hormones in biological matrices using mass spectrometry for analysis of mesocarb and its liquid chromatography-tandem mass spectrom- metabolites in human urine. J Chromatogr B 2004; etry. J Anal Toxicol 2004; 28: 1–10. 800: 281–289. 36. Shackleton CH, Chuang H, Kim J, et al. Electrospray 46. Maralikova B, Weinmann W. Simultaneous mass spectrometry of testosterone esters: potential determination of delta9-tetrahydrocannabinol, 11- for use in doping control. Steroids 1997; 62: hydroxy-delta9-tetrahydrocannabinol and 11-nor- 523–529. 9-carboxy-delta9-tetrahydrocannabinol in human plasma by high-performance liquid chromatog- 37. Kim JY, Choi MH, Kim SJ, et al. Measurement of raphy/tandem mass spectrometry. J Mass Spectrom 19-nortestosterone and its esters in equine plasma 2004; 39: 526–531. by high-performance liquid chromatography with tandem mass spectrometry. Rapid Commun Mass 47. Thevis M, Opfermann G, Krug O, et al. Electrospray Spectrom 2000; 14: 1835–1840. ionization mass spectrometric characterization and quantitation of xanthine derivatives using 38. Borts DJ, Bowers LD. Direct measurement of isotopically labelled analogues: an application for urinary testosterone and epitestosterone conju- equine doping control analysis. Rapid Commun gates using high-performance liquid chromatog- Mass Spectrom 2004; 18: 1553–1160. raphy/tandem mass spectrometry. J Mass Spectrom 2000; 35: 50–61. 48. Todi F, Mendonca M, Ryan M, et al. The confirma- tion and control of metabolic caffeine in standard- 39. Bowers LD, Sanaullah. Direct measurement of bred horses after administration of theophylline. steroid sulfate and glucuronide conjugates with J Vet Pharmacol Ther 1999; 22: 333–342. high-performance liquid chromatography-mass spectrometry. J Chromatogr B 1996; 687: 61–68. 49. Tang FP, Leung GN, Wan TS. Analyses of quaternary ammonium drugs in horse urine by capillary 40. Kuuranne T, Kotiaho T, Pedersen-Bjergaard S, et electrophoresis-mass spectrometry. Electrophoresis al. Feasibility of a liquid-phase microextraction 2001; 22: 2201–2209. sample clean-up and liquid chromatographic/mass spectrometric screening method for selected 50. Liu C, Bowers LD. Mass spectrometric characteriza- anabolic steroid glucuronides in biological tion of the beta-subunit of human chorionic samples. J Mass Spectrom 2003; 38: 16–26. gonadotropin. J Mass Spectrom 1997; 32: 33–42. 41. Kuuranne T, Aitio O, Vahermo M, et al. Enzyme- 51. Jacoby ES, Kicman AT, Laidler P, et al. Determination assisted synthesis and structure characterization of of the glycoforms of human chorionic gonado- glucuronide conjugates of methyltestosterone (17 tropin beta-core fragment by matrix-assisted laser alpha-methylandrost-4-en-17 beta-ol-3-one) and desorption/ionization time-of-flight mass spec- nandrolone (estr-4-en-17 beta-ol-3-one) metabo- trometry. Clin Chem 2000; 46: 1796–1803. lites. Bioconjug Chem 2002; 13: 194–199. 52. Sonksen PH. Insulin, growth hormone and sport. 42. Deventer K, Delbeke FT. Validation of a screening J Endocrinol 2001; 170: 13–25. method for corticosteroids in doping analysis by 53. de Kock SS, Rodgers JP, Swanepoel BC. Growth liquid chromatography/tandem mass spectrom- hormone abuse in the horse: preliminary assess- etry. Rapid Commun Mass Spectrom 2003; 17: ment of a mass spectrometric procedure for IGF-1 2107–2114. identification and quantitation. Rapid Commun 43. Fluri K, Rivier L, Dienes-Nagy A, et al. Method for Mass Spectrom 2001; 15: 1191–1197. LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 215

References 215

54. Tsushima T, Katoh Y, Miyachi Y, et al. Serum con- 60. Kawasaki N, Haishima Y, Ohta M, et al. Structural centrations of 20K human growth hormone in analysis of sulfated N-linked oligosaccharides in normal adults and patients with various endocrine erythropoietin. Glycobiology 2001; 11: 1043–1049. disorders. Study Group of 20K hGH. Endocr J 2000; 61. Caldini A, Moneti G, Fanelli A, et al. Epoetin alpha, 47(Suppl): S17–S21. epoetin beta and darbepoetin alfa: two-dimensional 55. Wu Z, Bidlingmaier M, Dall R, et al. Detection of gel electrophoresis isoforms characterization and doping with human growth hormone. Lancet 1999; mass spectrometry analysis. Proteomics 2003; 3: 353: 895. 937–941. 56. Wu SL, Jardine I, Hancock WS, et al. A new and 62. Stanley SM, Poljak A. Matrix-assisted laser- sensitive on-line liquid chromatography/mass desorption time-of flight ionisation and high- spectrometric approach for top-down protein performance liquid chromatography-electrospray analysis: the comprehensive analysis of human ionisation mass spectral analyses of two glycosy- growth hormone in an E. coli lysate using a hybrid lated recombinant epoetins. J Chromatogr B 2003; linear ion trap/Fourier transform ion cyclotron 785: 205–218. resonance mass spectrometer. Rapid Commun Mass 63. Thevis M, Ogorzalek Loo RR, et al. Mass spectro- Spectrom 2004; 18: 2201–2207. metric identification of synthetic insulin. In: 57. Wu SL, Amato H, Biringer R, et al. Targeted pro- Schänzer W, Geyer H, Gotzmann A, et al., eds, teomics of low-level proteins in human plasma by Recent Advances in Doping Analysis 11. Cologne: LC/MSn: using human growth hormone as a model Sport und Buch Strauß, 2003: 227–237. system. J Proteome Res 2002; 1: 459–465. 64. Guan F, Uboh C, Soma L, et al. Unique tryptic pep- 58. Lasne F, Martin L, Crepin N, et al. Detection of tides specific for bovine and human hemoglobin in isoelectric profiles of erythropoietin in urine: the detection and confirmation of hemoglobin- differentiation of natural and administered recom- based oxygen carriers. Anal Chem 2004; 76: binant hormones. Anal Biochem 2002; 311: 5118–5126. 119–126. 65. Thevis M, Ogorzalek Loo RR, Loo JA, et al. Doping 59. Che FY, Shao XX, Wang KY, et al. Characterization control analysis of bovine hemoglobin-based of derivatization of sialic acid with 2- oxygen therapeutics in human plasma by LC- aminoacridone and determination of sialic acid electrospray ionization-MS/MS. Anal Chem 2003; content in glycoproteins by capillary electro- 75: 3287–3293. phoresis and high performance liquid chromatog- raphy-ion trap mass spectrometry. Electrophoresis 1999; 20: 2930–2937. LCMS_Chap09 (JB-D) 8/5/06 3:14 pm Page 216