ELSEVIER
Journal of Chromatography B
journal homepage: www.elsevier.com/locate/chromb
JOURNAL OF CHROMATOGRAPHY B
ScienceDirect
Development of a solid phase extraction method for the simultaneous determination of steroid hormones in H295R cell line using liquid chromatography-tandem mass spectrometry
CrossMark
Jonas Abdel-Khalik*, Erland Björklund 1, Martin Hansen **
Toxicology Laboratory, Analytical Biosciences, Department of Pharmacy, Faculty of Health and Medical Sciences, University of Copenhagen, Universitetsparken 2, DK-2100 Copenhagen, Denmark
ARTICLE INFO
Article history: Received 21 December 2012
Accepted 17 July 2013 Available online 25 July 2013
Keywords: Solid phase extraction LC-MS/MS Cholesterol Steroid hormones H295R incubation medium Steroidogenesis
ABSTRACT
The H295R in vitro cell line produces the majority of the steroidogenesis, for which reason it is commonly used as a screening tool for endocrine disrupting chemicals. Simultaneous determination of the precursor cholesterol and key steroid hormones could give a broad insight into the mechanistic disruption of the steroidogenesis. Steroid hormones have primarily been extracted from H295R incubation medium by means of liquid-liquid extraction (LLE) and the obtained recoveries and matrix effects have typically not been stated or assessed. In the present study a solid-phase extraction (SPE) method was developed and validated for the simultaneous extraction of cholesterol and five key steroid hormones pregnenolone, 17-hydroxyprogesterone, testosterone, cortisol and aldosterone from H295R incubation medium, and finally detected by LC-MS/MS. Cholesterol was recovered at a level of 55.7%, while steroid hormone recoveries ranged from 98.2 to 109.4%. Matrix effects varied between -0.6% and 62.8%. Intra-day precision was deemed acceptable, but the inter-day precision for pregnenolone and aldosterone exceeded the precision limit of 15% RSD. Although LLE has been the most frequently used extraction method in H295R studies, however, our investigation has shown that SPE may relatively easily extract and recover steroid hormones, potentially replacing LLE.
@ 2013 Elsevier B.V. All rights reserved.
1. Introduction
The adrenal cortex of the adrenal gland has the capability of pro- ducing steroid hormones from the precursor cholesterol [1], and the enzyme catalyzed reactions in the formation of these steroid
Abbreviations: 17OHProg, 17-hydroxyprogesterone; ALD, aldosterone; Chol, cholesterol; Chol IS, d6-cholesterol; F, cortisol; F IS, d4-cortisol; GC, gas chromatog- raphy; H295R, human adrenocortical carcinoma cell line; HPLC, high performance liquid chromatography; IS, internal standard; LC, liquid chromatography; LODinstr, instrumental limit of detection; LOQinstr, instrumental limit of quantification; LLE, liquid-liquid extraction; MS/MS, tandem mass spectrometry; OECD, orga- nization for economic cooperation and development; Preg, pregnenolone; RSD, relative standard deviation; SPE, solid phase extraction; T, testosterone; T IS, d3- testosterone.
* Corresponding author. Present address: Institute of Mass Spectrometry, College of Medicine, Swansea University, Singleton Park, Swansea SA2 8PP, UK. Tel .: 00447598742565.
** Corresponding author. Present addresses: Department of Civil & Environmen- tal Engineering, Stanford University, CA, United States; Department of Growth and Reproduction, Copenhagen University Hospital, Denmark. Tel .: +45 51 76 70 58. E-mail addresses: J.A.F.A.ABDEL-KHALIK.744116@swansea.ac.uk
(J. Abdel-Khalik), martin.hansen@stanford.edu (M. Hansen).
1 Present address: School of Education and Environment, Division of Natural Sciences, Kristianstad University, SE-291 88 Kristianstad, Sweden.
hormones are collectively called the steroidogenesis [2]. Human adrenocortical carcinoma cells (H295R cells) have the same physi- ological characteristics as fetal cells of the adrenal cortex and thus can produce steroid hormones from all zones in the adult, healthy adrenal cortex [3-5]. The H295R human adrenocortical carcinoma cell line has therefore been presented as a rapid in vitro tool to determine the effects of endocrine disruption chemicals on the steroidogenesis [6]. By exposing these cells to possible endocrine disrupting chemicals, and thereafter determine the level of secreted steroid hormones relative to control incubations, allows for a more mechanistic understanding of steroidogenesis disruption [7].
Recently, the Organization for Economic Cooperation and Devel- opment (OECD) guideline for testing of various chemicals ability to disrupting the steroidogenesis was released [8], focusing on the expression of estradiol and testosterone as endpoints. Endocrine disrupting chemicals may however affect multiple parts of the steroidogenesis, and an effect on one part of the steroidogenesis may potentially lead to disruption of another part of the pathway. Consequently development of analytical methods capable of simul- taneously analyzing multiple steroid hormones is crucial [9,10].
Extraction of organics in water samples has classically been per- formed by means of liquid-liquid extraction (LLE) and is also the most common sample preparation method for steroid hormones in H295R incubation medium [7,11-25]. However, LLE methods
are often labor-intensive, and time-consuming, and require strict control of experimental conditions and generally involve large volumes of solvents (typically ether in the case of steroid hor- mones) [26,27]. Solid-phase extraction (SPE) often offers a more sophisticated approach, with faster extraction time, lowering of hazardous solvent consumption, and no emulsion formation [28]. Yet the number of publications utilizing SPE for the extraction of steroid hormones from H295R incubation medium has been lim- ited [9,29-31]. Furthermore these studies do not clearly state the SPE recoveries of steroid hormones of the developed method. Con- sequently there is a lack of knowledge concerning the performance of SPE during sample preparation and extraction of steroid hor- mones from H295R media. Liquid chromatography (LC-MS/MS) [11-15,19-22,29-31] and gas chromatography-tandem mass spectrometry (GC-MS/MS) [9,10,29] have been utilized to simulta- neously determine several extracted steroid hormones secreted by H295R cells. LC-MS/MS is generally preferred over GC-MS/MS due to time-consuming derivatization processes often being required prior to analysis by GC-MS/MS combined with longer times of analysis in GC-MS/MS [10,32].
In this study six steroid hormones were selected as endpoints. Structures and physicochemical properties of these can be seen in Table 1. Cholesterol, pregnenolone, 17-hydroxyprogesterone, testosterone, cortisol and aldosterone were chosen, as they cover the majority of the adrenal steroidogenesis. They vary in phy- siochemical properties as shown in Table 1, and each endpoint represents a different steroid class: pregnenolone for progestagens, 17-hydroxyprogesterone for hydroxylated progestagens, testos- terone for androgens, cortisol for glucocorticoids and aldosterone for mineralocorticoids. Another endpoint selection criterion was the physiological significance of the steroid. Cholesterol and pre- gnenolone are suitable endpoints, as they are implied in the rate limiting steps of the steroidogenesis [1,33]. The level of 17- hydroxyprogesterone is a frequently determined endpoint in the clinical settings, as its blood level can be used to assess the function- ality of the adrenal gland, which is typically done for babies [34]. Furthermore 17-hydroxyprogesterone is an important endpoint in the screening for congenital adrenal hyperplasia in infants [34]. Testosterone is essential for the development and maintenance of the male phenotype [35]. Cortisol was selected, as it regulates the metabolism, stress and immune response [36], and is commonly analyzed in clinical settings to diagnose adrenal hyperfunction or insufficiency [37]. Finally aldosterone was selected, as it is essen- tial for the regulation of the extracellular fluid volume by increasing sodium reabsorption and stimulating potassium excretion by the kidneys [38]. Although included in the OECD guideline [8] estradiol, an estrogen, was not analyzed as it would complicate LC-MS/MS analysis which was desired to be kept as simple as possible. Estro- gens if nonderivatized are usually analyzed in negative mode while the remaining steroid hormones if nonderivatized typically are ana- lyzed in positive mode [27].
The aim of this study was to develop and validate a SPE method capable of simultaneously extracting cholesterol and five key steroid hormones from H295R incubation medium; viz. cholesterol, pregnenolone, 17-hydroxyprogesterone, testosterone, cortisol and aldosterone. To achieve this goal a simple LC-MS/MS method capable of simultaneously analyzing the six steroid hormones also had to be established as part of the investigation, which is described in some detail in the first part of this paper.
2. Materials and methods
2.1. Materials and reagents
Cholesterol, pregnenolone, 17-hydroxyprogesterone, testos- terone, cortisol and aldosterone were all purchased from
Sigma-Aldrich (Glostrup, Denmark). The level of purity was above 99% for cholesterol, while it was above 96% for the remaining steroid hormones. Deuterated analogs were used as internal standards (IS); d6-cholesterol, d4-cortisol and d3-testosterone were obtained from Toronto Research Chemicals (North York, ON, Canada), all with a isotopic purity above 98%. Ammonium acetate had a purity level above 98% and was purchased from Sigma-Aldrich (Glostrup, Denmark). All utilized solvents were of analytical grade and obtained from Lab-scan analytical sciences (Fisher Scientific Biotech Line, Slangerup, Denmark). Stock solu- tions (in the range 100-500 µg/mL) of each internal standard were prepared in methanol. From the stock solutions a methanol mixture was prepared containing the internal standards at a concentration of 37.5, 4.5 and 15.0 µg/mL for d6-cholesterol, d3-testosterone and d4-cortisol, respectively. Each sample of H295R incubation medium was spiked with 20.0 L IS mixture, i.e. 750, 90, and 300 ng of d6-cholesterol, d3-testosterone and d4-cortisol, respectively. The concentration of each non-labeled steroid hormone in a mixture dissolved in methanol was 37.6, 37.7, 3.2, 4.5, 15.0 and 15.0 µg/mL for cholesterol, pregnenolone, 17-hydroxyprogesterone, testosterone, cortisol and aldosterone, respectively. Each sample was either pre- or post-spiked with 20.0 µL mixture, i.e. an absolute amount of 752, 754, 64, 90, 300 and 300 ng of cholesterol, pregnenolone, 17-hydroxyprogesterone, testosterone, cortisol and aldosterone, respectively. H295R incuba- tion medium composed of Dulbecco’s Modified Eagle’s Medium and Ham’s F-12 Nutrient mixture medium (GibcoBRL Life Technologies, Paisar, UK) supplemented with 10 mL/L of ITS+ premix (BD Bio- science, Brøndby, Denmark) and 25 mL/L Nu-serum (BD Bioscience, Brøndby, Denmark).
2.2. Chromatographic and mass spectrometric conditions
The modular HPLC (High Performance Liquid Chromatography) system (Agilent 1100 Series; Agilent Technologies, Palo Alto, CA, USA) comprised of an autosampler held at 4℃ (model G1367A), a column compartment oven kept at 30℃ (G1316A), a quater- nary pump (G1311A) and a degasser (G1322A). Chromatographic separation was carried out on a XTerra MS C18 analytical column (100 mm × 2.1 mm, 3.5 um particles with 125 Å pore size). Mobile phases A and B composed of methanol: water at a v/v-ratio of 50:50 and 99:1, respectively, both with 2.5 mM ammonium acetate. Gra- dient elution of the steroid hormones was performed by pumping 100% mobile phase A isocratically from minutes 0 to 14 which over the course of 1 min was changed to 100% mobile phase B that then was pumped isocratically from minutes 15 to 24. After the steroid hormone elution mobile phase A was pumped again iso- cratically from minutes 25 to 35 for system re-equilibration. Flow rate was set to 250 pL/min and an injection volume of 10 uL was used. For detection an API-2000 triple-quadrupole mass spectrom- eter (Applied Biosystems, Foster City, CA, USA) equipped with an electrospray source was operated in positive mode under multiple reaction monitoring conditions during analysis. The ion transitions for the MS/MS analysis of each analyte and internal standard are listed in Table 2, which also includes the tandem mass spectrom- eter working parameters. Obtained chromatographic peak areas were acquired by the Analyst 1.4 software package (MDS Sciex) and processed in the same software, along with Microsoft Office Excel 2007 and GraphPad Prism v. 5.0 (GraphPad Software, San Diego, CA, USA).
2.3. Sample preparation
The final SPE method contained the following procedural steps. In order to stabilize the steroid hormones 1.5 mL H295R incuba- tion medium was initially pH adjusted to pH 3.0 ± 0.1 with diluted
Table 1 Physicochemical properties, structures and proposed LC-MS/MS fragmentation patterns of cholesterol and the five steroid hormones investigated in the present study. The observed fragments obtained during multiple reaction monitoring were confirmed by previous studies identified in the literature.
| Steroid | Physicochemical properties | Precursor | Quantifier | Qualifier | Ref. |
|---|---|---|---|---|---|
| Chol | C27H46O M: 386.7 | H | H | [10,58-60] | |
| CAS: 57-88-5 | "I | ||||
| Sw: 0.1 | HO NH4 | + | |||
| log P: 8.74 | , 369 | ||||
| 404.5 m/z | 369.4 m/z | ||||
| 0 - 281 | 0 | ||||
| H | H | H | [61-63] | ||
| Preg | C21 H32 O2 | III | |||
| M: 316.5 g/mol | H | + | |||
| CAS: 145-13-1 | H2O | + | |||
| Sw: 7.1 | 299 ¥ | ||||
| log P: 4.22 | (281) | 299.4 m/z | 281.3 m/z | ||
| 317.2 m/z | |||||
| O HO H | 0 + | HỎ | |||
| 17OH- Prog | C21 H30O3 | I 2 H | [64-66] | ||
| M: 330.5 g/mol CAS: 68-96-2 | HỒ 97 | ||||
| Sw: 6.5 | 4 109 | 97.2 m/z | 109.0 m/z | ||
| log P: 3.17 | 331.3 m/z OH | ||||
| T | C19 H 28 O2 | H | 0 + | HỒ | [67-70] |
| M: 288.4 g/mol | A | ||||
| CAS: 58-22-0 | HỒ 97- - | ||||
| Sw: 23.4 | 109 | 97.2 m/z | 109.0 m/z | ||
| log P: 3.32 | 289.2 m/z | ||||
| 0 HO OH HO | 0 HO OH HO | ||||
| F | C21H30O5 | 1- H - | o + | H I | [65,71,72] |
| M: 362.5 g/mol | H | H | |||
| CAS: 50-23-7 | HỒ | ||||
| Sw: 320 | 267- 2 121- | 121.1 m/z | + | ||
| log P: 1.61 | 363.2 m/z | 267.3 m/z | |||
| 315 O 343 | 0 O | HOH2C + 0 | |||
| O A OH L . HO | HO | H | |||
| ALD | C21 H28 O5 | H | I H H | [73-75] | |
| M: 360.4 g/mol | H | ||||
| CAS: 52-39-1 | + | 0 | HO | ||
| Sw: 51.2 | HỒ | ||||
| log P: 1.08 | 361.4 m/z | 343.2 m/z | 315.2 m/z |
CxHyOz: molecular formula; M: molar mass (g/mol); P: octanol-water coefficient; Sw: water solubility at 37 ℃ (µg/mL).
sulfuric acid. Subsequently each sample was spiked with 20.0 µL of the IS mixture (described in Section 2.1). SPE was performed using C18 cartridges (500 mg, 3 mL reservoir, Varian Inc., Palo Alto, CA, USA) preconditioned with 2x 3 mL heptane, 3 mL acetone, 2x 3 mL methanol and lastly with 2x 3 mL tap water adjusted to pH 3.0. Samples were quantitatively transferred to the cartridges by flush- ing the sample test tube twice with 1 mL of pH-adjusted tap water and extracted at a rate of 1-2 mL/min using a vacuum manifold (IST
Vacmaster, Biotage, Uppsala, Sweden). Immediately after enrich- ment, the SPE cartridges were washed with 2x 3 mL tap water (pH 3.0) followed by air-drying using the vacuum manifold for 30 min. Finally, analytes were eluted from the SPE cartridges with 7 mL mobile phase B into a test tube and evaporated to dryness under a gentle stream of nitrogen. Evaporation was followed by reconsti- tution in 1.5 mL of a 50:50 mixture of mobile phases A and B, aided by 3 min of whirl mixing at 2600 rpm.
| Ion source specific | Compound specific | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ISV (V) | Temp. (°C) | GS1 (Psi) | GS2 (Psi) | CUR (Psi) | CAD (Psi) | Compounds | Precursor | Products | DP (V) | FP (V) | EP (V) | CE (V) | CXP (V) |
| 20 (138 kPa) | 3 (21 kPa) | Cholesterol | 404.5(M+NH4) | q1: 369.4 | 130 | 330 | 3 | 15 | 8 | ||||
| 15 (103 kPa) | d6-Cholesterol | 410.6(M+NH4) | q1: 375.2 | 130 | 330 | 3 | 15 | 8 | |||||
| 5500 | Pregnenolone | 317.2 (M+H) | q1: 299.4q2: 281.3 | 80 | 400 | 10 | 20 | 6 | |||||
| 200 | 25 (172 kPa) | 17OHProg | 331.3 (M+H) | q1: 97.2 q2: 109.0 | 80 | 400 | 12 | 35 | 2 | ||||
| Testosterone | 289.2 (M+H) | q1: 97.2 q2: 109.0 | 75 | 400 | 12 | 33 | 2 | ||||||
| d3-Testosterone | 292.2 (M+H) | q1: 97.0 q2: 109.0 | 52 | 340 | 10 | 38 | 2 | ||||||
| Cortisol | 363.2 (M+H) | q1: 121.1q2: 267.3 | 80 | 400 | 10 | 32 | 4 | ||||||
| d4-Cortisol | 367.2 (M+H) | q1: 121.1q2: 331.3 | 55 | 325 | 10 | 37 | 3 | ||||||
| Aldosterone | 361.4 (M+H) | q1: 343.2q2: 315.2 | 50 | 350 | 9 | 28 | 7 | ||||||
CAD: collision-activated dissociation; CE: collision energy; CUR: curtain gas; CXP: collision cell exit potential; DP: declustering potential; EP: entrance potential; FP: focusing potential; GS1: nebulizer; GS2: gas heater; ISV: ion spray voltage; M: molecular ion; Temp .: temperature; q1: quantifier; q2: qualifier. The transition between precursor and quantifier (fragment) was used to assess the recovered level of steroid hormones, while the transition between precursor and qualifier (fragment) was used as an additional parameter of peak identification.
2.4. Validation of LC-MS/MS method
Calibration curves using five concentration levels for each steroid were constructed by diluting the standard mixture with water containing ammonium acetate, resulting in five solutions of the six steroid hormones in 50:50 (v/v) mixtures of mobile phases A and B. Each solution was prepared in triplicate. The HPLC vial concentration ranges for the calibration curves were 301-704 ng/mL for cholesterol and pregnenolone, 26-60 ng/ml for 17-hydroxyprogesterone, 37-86 ng/ml for testosterone and lastly 119-280 ng/mL for cortisol and aldosterone. Peak areas of the anal- yses were subtracted with the peak areas from analyses of blank 50:50 (v/v) mixture of mobile phases A and B, which followed each analysis of a standard solution to ensure no interference from any carry-over although none was observed. Calibration curves were obtained by correlating peak areas in counts per second as a function of the analyte concentration in ng/ml. Deuterium labeled internal standards were not included when determin- ing the calibration curves, as the curves were based on analysis of neat standards. Furthermore the calibration curves were only determined to assess a range for each analyte, within which the peak area and concentration correlated linearly. Thereby it was ensured that the recovery of steroid hormones during SPE method development and validation were within their respective linear ranges. For each steroid hormone, only the ion transition between precursor and quantifier was used during calibration curve deter- mination. It was however set as a criterion that the peak area ratios between the qualifier and the quantifier ion transitions should not deviate more than 20% from a fixed ratio [10,39]. Intra- (n=3) and inter-day (n=6) instrument precisions were determined from relative standard deviations (RSDs) of the peak areas obtained for a triplicate injection during calibration curve determination. Analyses were performed on a solution containing the steroid hor- mones at concentration levels near the middle of their respective standard curves. For the LC-MS/MS instrument the limits of detec- tion (LODinstr) and quantification (LOQinstr) were calculated for each analyte as recommended by the ICH guideline from 2005 [40] fol- lowing Eqs. (1) and (2). Here o is the standard deviation of the peak areas obtained, when analyzing in triplicate the least concentrated mixture of each analyte during determination of the calibration curves. S is the slope of the calibration curve for each steroid hor- mone.
LODinstr = 3.3 x~
₹
LOQinstr = 10 x~
₾
2.5. Validation of SPE method
H295R incubation medium samples adjusted to pH 3.0 were spiked with 20.0 uL of the standard mixture and 20.0 pL of the IS mixture. Responses from the analysis of these samples were used to determine matrix effects, absolute recovery, relative recovery and process efficiency [41-44]. As for the neat standards during calibration curve determination (Section 2.4) only the transition between precursor and quantifier was used to assess the recov- eries and process efficiencies for each steroid hormone. An upper limit of 20% deviation from a fixed ratio for the peak area ratios between the qualifier and the quantifier ion transitions was again set. For each steroid hormone the deviation from a fixed ratio was collectively determined for Sets 1 and 2, which are described in the following. Three sets of samples were prepared and spiked with identical concentrations of steroid hormones: Set 1 was H295R medium spiked prior to SPE extraction (pre-spiked); Set 2 was H295R medium extracted through the SPE and thereafter spiked
(2) (1)
(post-spiked); Set 3 were pure solutions of the analytes in a 50:50 (v/v) mixture of mobile phases A and B. From these sets of samples the matrix effect (Eq. (3)) was determined as described by Stahnke et al. [43], process efficiency (Eq. (4)) as well as absolute recover- ies (Eq. (5)) were determined according to Matuszewski et al. [44], while relative recoveries were determined as described in Hansen et al. [45]:
Matrix effect (%) = (( Set 3 response )-1) x 100% (3)
Set 2 response
Process efficiency (%) = Set 1 reponse Set 3 reponse
× 100% (4)
Absolute recovery (%) = Set 1 reponse Set 2 reponse
× 100% (5)
Relative recovery (%) = Set 1 reponseanalyte/Set 1 reponseIs Set 2 reponseanalyte/Set 2 reponseIs × 100% (6)
response analyte and responseis were the peak areas obtained for an analyte or IS, respectively. d6-Cholesterol was used as an IS for cholesterol, d3-testosterone was used as an IS for pregnenolone, 17-hydroxyprogesterone and testosterone, while d4-cortisol was used as an IS for cortisol and aldosterone.
3. Results and discussion
3.1. LC-MS/MS method
A prerequisite for the development and validation of an SPE method suitable for a broad range of steroid hormones in H295R medium was the establishment of a detection technique com- bining an efficient liquid chromatographic separation with an appropriate ionization method prior to MS/MS. This method was then utilized in all SPE method optimizations. The six calibra- tion curves were all linear, as the coefficient of determination (R2) was in the range of 0.9600-0.9924, with R2 being >0.9900 for 17-hydroxyprogesterone, testosterone and cortisol. Calibra- tion ranges for 17-hydroxyprogesterone, testosterone and cortisol might therefore have exceeded 26-60, 37-86 and 119-280 ng/mL, respectively. In contrast a tendency for non-linearity and impre- cision was observed outside the ranges of 301-704ng/ml for cholesterol and 119-280 ng/mL for aldosterone (data not shown). Precision, quantifier/qualifier ratio, LODinstr and LOQinstr for the six endpoints analyzed by the developed LC-MS/MS method are displayed in Table 3. According to the FDA guideline from 2001, the precision measured as % RSD should not exceed 15% [46]. In the present study, the intra-day precision was in the range of 0.9-15.9% and therefore deemed acceptable; analysis of choles- terol was the most imprecise. For most analytes, inter-day precision
was good and below 8%, though analysis of aldosterone was some- what outside the accepted range with a value of 18.8%. Finally analysis of pregnenolone showed the poorest inter-day precision with a value close to 40%, which might partly be explained by a slightly inconsistent quantifier/qualifier ion-ratio, as % RSD of the ratio for pregnenolone was 14.0% during calibration curve deter- mination. However, overall variation of the quantifier/qualifier ion-ratio was acceptable for all six steroid hormones, as they fell in the range 2.1-14.0% (Table 3). The MS/MS-ion fragmentation was also investigated carefully by analyzing constantly infused neat standards (approximately 5 µg/mL in 50:50 A:B). Proposed fragmentation patterns of the steroid hormones and agreements between other studies are summarized in Table 1. Depending on the analyte, the LODinstr and LOQinstr varied from the lower to the higher ng/ml range, with LOQinstr being as low as 1.0 ng/ml for testosterone and as high as 350.1 ng/mL for pregnenolone (Table 3). Based on the obtained LOQinstr levels, the developed LC-MS/MS method may be sufficiently sensitive to quantify cholesterol, 17- hydroxyprogesterone, testosterone and possibly cortisol during a regular H295R study while pregnenolone and especially aldos- terone are questionable [7,9,14,18,47,48]. This does however seem to depend on the experimental conditions [7-9,14,18,47-49]. First and foremost the LC-MS/MS method was sufficiently sensitive to serve its purpose of aiding the development and validation of the SPE methodology.
3.2. Optimization of SPE conditions
The method developed here was based on a previously devel- oped method for the determination of three groups of steroid hormones (progestagens, androgens and estrogens) in blood using a C18 SPE column, followed by additional clean-up steps (silica and aminopropyl), derivatization and analysis by GC-MS/MS [10]. Simi- larities to that method include stabilizing the H295R medium to pH 3.0 by adding diluted sulfuric acid. Furthermore the C18 SPE column conditioning was the same applying heptane, acetone, methanol and pH 3 adjusted water (see Section 2.3). However, the previous method did not apply a rinse step, and elution was performed with acetone, which was evaporated off prior to reconstitution in chlo- roform and external clean-up on packed silica gel and aminopropyl columns. Therefore a number of critical steps had to be evaluated including rinsing, elution and reconstitution prior to analysis on LC-MS/MS, as outlined below.
3.2.1. Rinsing step of samples attached to SPE cartridges
In the present study on H295R medium a rinsing step was applied to replace external clean-up (from Ref. [10]) in order reduce overall sample preparation time. During the development of the SPE method it was experienced that inclusion of column washing with pH 3 adjusted tap water significantly reduced the amount of insoluble residues in the eluate after solvent evaporation. Conse- quently, washing of the cartridge eluted matrix compounds initially
| Chol | Preg | 17OHProg | T | F | ALD | |
|---|---|---|---|---|---|---|
| Precision (% RSD) | ||||||
| Intra-day (n=3) | 15.9 | 12.2 | 1.0 | 0.9 | 1.3 | 8.1 |
| Inter-day (n=6) | 7.4 | 37.6 | 6.2 | 3.4 | 4.3 | 18.7 |
| Quantifier/qualifier ratio (n=3) Calibration curve | – | 1.9±0.3 | 1.0±0.1 | 1.0±0.0 | 3.7±0.2 | 2.9±0.2 |
| LOD and LOQ (ng/ml) | ||||||
| LODinstr | 31.1 | 115.5 | 0.5 | 0.3 | 4.3 | 45.6 |
| LOQinstr | 94.4 | 350.1 | 1.4 | 1.0 | 13.1 | 138.3 |
withheld by the SPE column, giving cleaner extracts. It should be noted though that column washing might only be significant, when the steroid hormones are analyzed by LC-MS/MS. Other stud- ies which have analyzed the steroid hormones by LC-MS/MS also included column washing with fairly aqueous solutions in their SPE procedures [29-31]. In contrast studies using SPE and GC-MS/MS for analysis of steroid hormones in H295R incubation medium did not include column washing in their SPE procedure [9,29]. The most significant reason for the lack of rinsing may be that these SPE procedures were followed by for example LLE, which contribute to sample cleanup [9,29]. Another reason for the SPE wash differ- ence between GC and LC based methods may be that the latter technology is more sensitive to matrix components, affecting MS- ionization efficiency [44,50]. As a final note pH adjusted tap water may be used in the developed SPE method as no steroids were detected in procedural blanks during development of the original SPE method (data not shown) [10]. Furthermore any ions in the applied tap water in the present study did not seem to significantly interact with the analytes as no significant complexes between steroid hormones and ions were detected during LC-MS/MS anal- ysis (Table 2).
3.2.2. Elution solvent of samples attached to SPE cartridges
The SPE method developed by Hansen et al. [10] utilized acetone as elution solvent, and was initially attempted here, since acetone generally has a high eluting strength [51]. However acetone was found inappropriate, since a vast amount of solidified eluate was obtained after evaporation. The vast amount made it difficult to avoid particles from appearing in the reconstituted solution, which may lead to peak distortion and carryover in the obtained chro- matograms [51,52]. As a consequence other elution solvents were tested. Here methanol based solvents were utilized, since methanol has a lower eluting strength than acetone, and thus implied a lower risk of eluting undesirable matrix components [51]. Furthermore pure methanol was used to elute steroid hormones from a C18 cartridge in Rosenmai et al. [31] and a Strata X cartridge in Rijk et al. [29]. In Table 4 absolute and relative recoveries, process effi- ciencies and matrix effects following elution with either 7 mL pure mobile phase B or 7 mL of a 50:50 (v/v)% mixture of mobile phases A and B are shown. As can be seen from the table, cholesterol is not recovered from the column when eluting with a mixture of mobile phases A and B. When eluting with mobile phase B only, the absolute recovery and process efficiency of cholesterol is how- ever approximately 60 and 85%, respectively. In contrast, no matter whether mobile phase B alone or a mixture of mobile phases A and B were utilized, the remaining steroid hormones were recovered at a level of at least 51.7% with regard to absolute recovery and pro- cess efficiency. Relative recoveries were at least 66.6%. These data
indicate that pure methanol (mobile phase B) is necessary for suitable recovery of cholesterol (60.4%). The remaining steroid hormones may however be recovered by using lower levels of methanol the 50:50 (v/v)% mixture of mobile phases A and B (which contains approximately a total of 75% methanol) and is an option for anyone not interested in analyzing cholesterol. In either case it should be noted in Table 4 that the matrix effect was 26.6% at most, which typically is considered to be within a reasonable range [44,53,54].
3.2.3. Reconstitution solvent after SPE elution and solvent evaporation
During the development and validation of the SPE method, the entire solidified residue in the eluate after solvent evapora- tion could not be dissolved in mixtures of methanol and water or methanol alone. Different approaches were therefore attempted prior to method validation in order to reconstitute cholesterol and the five steroid hormones without particles appearing in the HPLC vial. First and foremost it was experienced that centrifugation did not precipitate the particles in solution, while the analytes were observed to adsorb to a 0.2 um sterile filter (data not shown). Ultra- sonication for 15 min and heating the samples at 50℃ for 5 h did however lead to dissolution or degradation of the particles. This latter approach was however time-consuming. Instead reconstitu- tion was attempted aided by 3 min of whirl mixing at 2600 rpm. Although the entire eluate was not reconstituted, initial studies for this approach indicated higher process efficiencies (for the whole method) when reconstituting in a 50:50 (v/v)% mixture of mobile phases A and B rather than mobile phase A alone (data not shown). The 50:50 (v/v)% mixture of both mobile phases contained approx- imately 75 (v/v)% methanol compared to the content of 50 (v/v)% methanol in mobile phase A alone. Correlation between the content of methanol in the reconstitution solvent and recovery was hence thought to exist. A comparative study of reconstituting in either 75 or 90 (v/v)% methanol was therefore performed as shown in Table 5. Although the recoveries, process efficiencies and matrix effect did not seem to significantly differ between the latter reconstitution solvents, peak broadening and slight peak distortion in LC-MS/MS chromatograms were observed for aldosterone, cortisol and testos- terone when reconstituting in 90 (v/v)% methanol (Fig. 1, lower chromatogram). The peaks were fairly sharp when reconstituting in 75 (v/v)% methanol (Fig. 1, middle chromatogram). Reconstitu- tion in 75 (v/v)% methanol (50:50 (v/v)% mixture of mobile phases A and B), aided by 3-5 min of whirl mixing at 2600 rpm, was therefore selected as the final reconstitution solvent. For compar- ison the reconstitution solvent in the study by Rijk et al. [29] was mostly aqueous based, 80% water, in order to make the reconsti- tution solvent compatible with the initial LC gradient conditions.
| Solvent | Chol | Preg | 17OHProg | T | F | ALD | Chol IS | TIS | F IS |
|---|---|---|---|---|---|---|---|---|---|
| Absolute recovery (%) | |||||||||
| 50:50 A:B | 11.2 | 72.7 | 93.1 | 102.0 | 176.2 | 122.9 | 0.0 | 109.3 | 176.5 |
| B | 60.4 | 118.2 | 120.3 | 129.7 | 185.6 | 179.2 | 52.3 | 136.5 | 195.8 |
| Relative recovery (%) | |||||||||
| 50:50 A:B | 0.0 | 66.6 | 85.2 | 93.3 | 99.9 | 69.6 | |||
| B | 115.5 | 86.6 | 88.1 | 95.0 | 94.8 | 91.5 | |||
| Process efficiency (%) | |||||||||
| 50:50 A:B | 12.9 | 95.6 | 83.4 | 86.7 | 51.7 | 62.2 | 0.0 | 69.8 | 46.1 |
| B | 84.8 | 115.5 | 108.3 | 104.0 | 105.3 | 90.8 | 55.4 | 81.7 | 88.1 |
| Matrix effect (%) | |||||||||
| 50:50 A:B | 8.6 | 15.5 | 1.7 | -4.8 | 3.6 | -21.1 | -17.3 | -31.4 | -12.9 |
| B | 26.6 | 19.7 | 17.0 | 8.2 | 15.1 | -14.5 | 25.4 | -20.5 | -5.4 |
| Solvent | Chol | Preg | 17OHProg | T | F | ALD | Chol IS | TIS | F IS |
|---|---|---|---|---|---|---|---|---|---|
| Absolute recovery (%) | |||||||||
| 50:50 A:B | 60.4 | 118.2 | 120.3 | 129.7 | 185.6 | 179.2 | 52.3 | 136.5 | 195.8 |
| 90 B | 63.9 | 117.0 | 96.3 | 101.4 | 103.8 | 113.0 | 42.9 | 104.7 | 102.5 |
| Relative recovery (%) | |||||||||
| 50:50 A:B | 115.5 | 86.6 | 88.1 | 95.0 | 94.8 | 91.5 | |||
| 90 B | 149.1 | 111.7 | 92.0 | 96.9 | 101.2 | 110.2 | |||
| Process efficiency (%) | |||||||||
| 50:50 A:B | 84.8 | 115.5 | 108.3 | 104.0 | 105.3 | 90.8 | 55.4 | 81.7 | 88.1 |
| 90 B | 73.2 | 94.9 | 96.7 | 89.2 | 79.3 | 74.3 | 40.1 | 68.0 | 66.3 |
| Matrix effect (%) | |||||||||
| 50:50 A:B | 26.6 | 19.7 | 17.0 | 8.2 | 15.1 | -14.5 | 25.4 | -20.5 | -5.4 |
| 90 B | 23.0 | 1.2 | 6.6 | -1.5 | 6.7 | -12.5 | -8.5 | -21.0 | -10.5 |
Reconstitution was however firstly performed in pure methanol before water was added [29], seemingly confirming the need for a high methanol content when reconstituting steroid hormones.
3.3. Method validation
3.3.1. Recoveries and process efficiency
Once the rinsing, elution and reconstitution steps were opti- mized, absolute and relative recoveries, process efficiencies and matrix effect obtained for the final SPE method was evaluated as outlined in Table 6 and summarized graphically in Fig. 2. First and foremost during the SPE method validation, the quantifier/qualifier ion-ratio was acceptable, as it fell within 1.9 for testosterone to 16.9% for pregnenolone. Except for cholesterol and its inter- nal standard, the absolute and relative recoveries along with the process efficiencies were in the range of 82.0-111.2%, which are considered as acceptable levels [44,45,53,55]. Complete valida- tions of SPE methods should cover three concentration levels (low, medium and high) and 6 replicates at each concentration level [46]. Yet, the % RSDs of the various recoveries and process efficiencies in Table 6 show good precision of the SPE method.
For cholesterol the relative recovery was 101.9%, but the abso- lute recovery was limited to 55.7%. Although the process efficiency was determined to be 97.7% for cholesterol, the process efficiency for its internal standard was only 50.5%. The high process efficiency for cholesterol is arguably due to ion enhancement, since the matrix effect for cholesterol was determined to be 62.8% (Table 6). Hansen et al. [10] recovered 119 ±8.4% of cholesterol, after pH 3 adjusted plasma samples were spiked first with 100 ng of cholesterol and then SPE extracted and eluted with acetone, and finally analyzed by GC-MS/MS (personal communication). However, acetone caused vast particle formation as discussed above (Section 3.2) when preparing for LC-MS/MS analysis. In the study by Hansen et al. [10], no precipitation was observed as the extract was redissolved
in n-heptane. Despite the recoveries of cholesterol being some- what suboptimal by eluting with mobile phase B in the present study, the extraction was consistent and precise, as the % RSD for the absolute and relative recoveries and the process efficiency of cholesterol were in the range of 9.2-14.5% (Table 6) [39,46]. Fur- thermore it proved difficult to increase the recovery of cholesterol without simultaneously increasing the presence of water-insoluble matrix compounds in the eluate. Taking these two statements into account, the obtained recovery of cholesterol was considered acceptable.
3.3.2. Matrix effect
The matrix effect for each steroid was determined to be in the range of -0.6 (ion suppression) to 62.8% (ion enhancement) (see Table 6 and Fig. 2). Of the steroids, only cholesterol and pregneno- lone experienced a fairly significant matrix effect, 62.8 and 32.8% respectively [44,53,54]. In Fig. 1 the upper and middle LC-MS/MS chromatograms were obtained for, respectively, a standard mix- ture of the six analytes and an injection of a sample from Set 2 treated by the validated procedure described in Section 2.3. When comparing the two chromatograms, the baseline in the chro- matogram for a Set 2 sample is unaffected by matrix effects. Peak shapes are however not as sharp in the chromatogram for Set 2 as in the standard mixture. This is especially seen for cholesterol and its internal standard. In general the level of matrix effects were thus observable, yet not significantly limiting the analyti- cal method. Neither Nielsen et al. [9], Rijk et al. [29], Iwaoka et al. [30] nor Rosenmai et al. [31] state the level of matrix effects in their H295R medium studies. Depending on the selectivity of their sample preparation methods toward the extraction of steroid hor- mones, the matrix effects may be higher or lower than obtained in the present study [44,53,54]. It should however be noted that bioanalysis of cholesterol by LC-MS/MS often may be compli- cated by matrix effects due to ion enhancement [56]. The benefits
| Chol Preg | 17OHProg | T | F | ALD | Chol IS | TIS | F IS |
|---|---|---|---|---|---|---|---|
| Absolute recovery (%) (n=3) 55.7 ± 13.6 104.4 ± 25.6 | 98.2 ± 24.5 | 98.2 ± 23.6 | 109.4 ± 19.7 | 98.5 ± 23.0 | 55.1 ± 16.2 | 100.6 ± 15.4 | 111.2 ± 17.7 |
| Relative recovery (%) (n=3) 101.9 ± 9.2 104.0 ± 16.3 | 98.8 ± 14.7 | 98.6 ± 10.5 | 98.5 ± 5.0 | 88.8 ± 10.7 | |||
| Process efficiency (%) (n=3) 97.7 ± 14.5 103.1 ±26.8 | 100.1 ± 10.8 | 103.3 ± 3.5 | 96.1 ± 6.8 | 96.6 ± 6.8 | 50.5 ± 16.4 | 90.4 ± 6.7 | 82.0 ± 5.8 |
| Matrix effect (%) (n=3) 62.8 ± 10.5 32.8 ± 23.7 | 6.9 ± 9.3 | 13.0 ± 3.5 | 15.0 ± 4.5 | 19.8 ± 7.0 | 4.4 ± 10.8 | -0.6 ± 1.7 | -0.2 ± 4.0 |
5.303
(Offsets: +3 min
17.423
+500 cps) FIS
(Offsets:
-3 min
+500 cps)
2000
5.454
TIS
20.150
Intensity (cps)
F
17.726 T
17OHProg
3.636 ALD
29.089
1000
(Offsets:
+1 min
+500 cps)
21.817 Preg
Chol IS
Chs
29.241
0
5.303
17.272 (Offsets: -3 min +500 cps)
2000
Intensity (cps)
(Offsets: +3 min +500 cps) FIS
20.150
17OHProg
5.303
TIS
F
17.575
T
29.089
1000
3.485
(Offsets:
ALD
+1 min
+500 cps)
21.817 Preg
Chol IS
Chol
29.241
0
2000
5.303
17.423
20.150
Intensity (cps)
(Offsets: +3 min
(Offsets:
8
+500 cps)
-3 min
+500 cps)
2
5.303
FIS
TIS
17.575
=
29.241
1000
F
T
(Offsets:
+1 min
21.817
+500 cps)
3.485
Preg
Chol IS
ALD
Chol
29.392
0
0
10
20
30
Time (min)
of using appropriate internal standards of steroid hormones to correct for matrix effects have been shown in the literature [42,57]. In the present study the use of internal standards was omitted from the final methodology and calibration, as matrix effect and process efficiency were observed to be different for analytes and internal standards (Table 6). Furthermore for this reason in addition to being costly usage of an equivalent deuterium labeled internal standard for each of the six analytes was not performed.
150
Absolute recovery
Relative recovery
100
Process efficiency
Matrix effect
0°
50
0
3
5
5
Chol
Preg
17OHProg
1
TI
ALD
4. Conclusion
A SPE methodology was developed and validated, which rel- atively easily extracted cholesterol and the five selected steroid hormones from H295R incubation medium. Compared to two other studies in the literature using SPE and LC-MS or LC-MS/MS for the analysis of steroid hormones in H295R incubation medium [29,30], the developed method was not necessarily easier to per- form, but was the first study to focus and give data on the extraction of steroids from H295R incubation medium by SPE. The steroid hormones were recovered with absolute recoveries above 98.2% and process efficiencies above 96.1%. Cholesterol was however only recovered 55.7% in absolute terms. An attempt was made to increase the recovery of cholesterol, but it was found diffi- cult to increase the recovery of cholesterol without eluting highly hydrophobic compounds, which could not easily be dissolved in aqueous solution. Although LLE has been the most frequently used extraction method in a H295R study, the present study has shown that a SPE method may relatively easily extract and recover steroid hormones.
References
[1] P.A. Fail, C.S. Sloan, J.D. Johnson, V.J. Brown, Steroidogenesis Screening Assays and Endocrine Disruptors, U.S. Environmental Protection Agency Endocrine Disruptor Screening Program, EPA Contract Number 68-W- 01-023, Work Assignments 2-6 and 5-5, Task 3, 2005 (document online; cited 04.08.12). Available from http://www.epa.gov/endo/pubs/edmvs/ steroidogenesis_drp_final.3.29.05.pdf
[2] C. Furuta, S. Noda, C. Li, A.K. Suzuki, S. Taneda, G. Watanabe, K. Taya, Toxicol. Appl. Pharmacol. 229 (2008) 109.
[3] J. Sanderson, W. Seinen, J. Giesy, M. Van Den Berg, Toxicol. Sci. 54 (2000) 121.
[4] W.E. Rainey, K. Saner, B.P. Schimmer, Mol. Cell. Endocrinol. 228 (2004) 23.
[5] W.E. Rainey, I.M. Bird, C. Sawetawan, N.A. Hanley, J.L. McCarthy, E.A. McGee, R. Wester, J.I. Mason, J. Clin. Endocrinol. Metab. 77 (1993) 731.
[6] B. Staels, D.W. Hum, W.L. Miller, Mol. Endocrinol. 7 (1993) 423.
[7] M. Hecker, J.L. Newsted, M.B. Murphy, E.B. Higley, P.D. Jones, R. Wu, J.P. Giesy, Toxicol. Appl. Pharmacol. 217 (2006) 114.
[8] OECD, Test No. 456: H295R Steroidogenesis Assay, OECD Guidelines for the Testing of Chemicals, Section 4: Health Effects, OECD Publishing, July 2011 (doc- ument online; cited 04.08.12). Available from http://browse.oecdbookshop. org/oecd/pdfs/free/9745601e.pdf
[9] F.K. Nielsen, C.H. Hansen, J.A. Fey, M. Hansen, N.W. Jacobsen, B. Halling- Sørensen, E. Björklund, B. Styrishave, Toxicol. In Vitro 26 (2012) 343.
[10] M. Hansen, N.W. Jacobsen, F.K. Nielsen, E. Björklund, B. Styrishave, B. Halling- Sørensen, Anal. Bioanal. Chem. 400 (2011) 3409.
[11] F. Zhang, D.L. Rick, L.H. Kan, A.W. Perala, D.R. Geter, M.J. Lebaron, M.J. Bartels, Rapid Commun. Mass Spectrom. 25 (2011) 3123.
[12] X. Zhang, H. Chang, S. Wiseman, Y. He, E. Higley, P. Jones, C.K. Wong, A. Al- Khedhairy, J.P. Giesy, M. Hecker, Toxicol. Sci. 121 (2011) 320.
[13] Y. Xing, M.A. Edwards, C. Ahlem, M. Kennedy, A. Cohen, C.E. Gomez-Sanchez, W.E. Rainey, J. Endocrinol. 209 (2011) 327.
[14] C. Liu, X. Zhang, H. Chang, P. Jones, S. Wiseman, J. Naile, M. Hecker, J.P. Giesy, B. Zhou, Toxicol. Appl. Pharmacol. 247 (2010) 222.
[15] U. Muller-Vieira, M. Angotti, R.W. Hartmann, J. Steroid Biochem. Mol. Biol. 96 (2005) 259.
[16] T. Gracia, P. Jones, E. Higley, K. Hilscherova, J. Newsted, M. Murphy, A. Chan, X. Zhang, M. Hecker, P. Lam, R. Wu, J. Giesy, Environ. Sci. Pollut. Res. 15 (2008) 332.
[17] T. Gracia, K. Hilscherova, P.D. Jones, J.L. Newsted, E.B. Higley, X. Zhang, M. Hecker, M.B. Murphy, R.M. Yu, P.K. Lam, R.S. Wu, J.P. Giesy, Toxicol. Appl. Pharmacol. 225 (2007) 142.
[18] T. Gracia, K. Hilscherova, P.D. Jones, J.L. Newsted, X. Zhang, M. Hecker, E.B. Higley, J.T. Sanderson, R.M. Yu, R.S. Wu, J.P. Giesy, Ecotoxicol. Environ. Saf. 65 (2006) 293.
[19] P. Liakos, D. Lenz, R. Bernhardt, J.J. Feige, G. Defaye, J. Endocrinol. 176 (2003) 69.
[20] L. Schloms, K.H. Storbeck, P. Swart, W.C. Gelderblom, A.C. Swart, J. Steroid Biochem. Mol. Biol. 128 (2012) 128.
[21] S. Komarnytsky, D. Esposito, A. Poulev, I. Raskin, J. Cell. Physiol. 228 (2013) 1120.
[22] R. van der Pas, L.J. Hofland, J. Hofland, A.E. Taylor, W. Arlt, J. Steenbergen, P.M. van Koetsveld, W.W. de Herder, F.H. de Jong, R.A. Feelders, J. Endocrinol. 215 (2012) 403.
[23] E.B. Higley, J.L. Newsted, X. Zhang, J.P. Giesy, M. Hecker, Environ. Sci. Pollut. Res. Int. 17 (2010) 1137.
[24] K. Ji, K. Choi, S. Lee, S. Park, J.S. Khim, E.H. Jo, X. Zhang, J.P. Giesy, J. Hazard. Mater. 182 (2010) 494.
[25] Y. Ma, J. Han, Y. Guo, P.K. Lam, R.S. Wu, J.P. Giesy, X. Zhang, B. Zhou, Aquat. Toxicol. 106/107 (2012) 173.
[26] V. Camel, Spectrochim. Acta Part B: At. Spectrosc. 58 (2003) 1177.
[27] J. Abdel-Khalik, E. Björklund, M. Hansen, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 928 (2013) 58.
[28] D. Hagen, C. Markell, G. Schmitt, Anal. Chim. Acta 236 (1990) 157.
[29] J.C. Rijk, A.A. Peijnenburg, M.H. Blokland, A. Lommen, L.R. Hoogenboom, T. Bovee, Chem. Res. Toxicol. 25 (2012) 1720.
[30] Y. Iwaoka, R. Hashimoto, H. Koizumi, J. Yu, T. Okabe, Life Sci. 86 (2010) 894.
[31] A.K. Rosenmai, F.K. Nielsen, M. Pedersen, N. Hadrup, X. Trier, J.H. Christensen, A.M. Vinggaard, Toxicol. Appl. Pharmacol. 266 (2013) 132.
[32] N. Krone, B.A. Hughes, G.G. Lavery, P.M. Stewart, W. Arlt, C.H. Shackleton, J. Steroid Biochem. Mol. Biol. 121 (2010) 496.
[33] A.H. Payne, D.B. Hales, Endocr. Rev. 25 (2004) 947.
[34] K. Sarafoglou, J.H. Himes, J.M. Lacey, B.C. Netzel, R.J. Singh, D. Matern, Horm. Res. Paediatr. 75 (2011) 19.
[35] D.T. Harwood, D.J. Handelsman, Clin. Chim. Acta 409 (2009) 78.
[36] A. Ohlsson, N. Cedergreen, A. Oskarsson, E. Ullerås, Toxicology 275 (2010) 21.
[37] M.M. Kushnir, A.L. Rockwood, W.L. Roberts, B. Yue, J. Bergquist, A.W. Meikle, Clin. Biochem. 44 (2011) 77.
[38] T.J. Rosol, J.T. Yarrington, J. Latendresse, C.C. Capen, Toxicol. Pathol. 29 (2001) 41.
[39] M.P. Schlüsener, K. Bester, Rapid Commun. Mass Spectrom. 19 (2005) 3269.
[40] ICH, International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use, Validation of Analytical Pro- cedures: Text and Methodology Q2(R1), November 2005 (document online; cited 07.09.12). Available from http://www.ich.org/fileadmin/Public_Web _Site/ICH_Products/Guidelines/Quality/Q2_R1/Step4/Q2_R1 __ Guideline.pdf
[41] E. Rozet, R.D. Marini, E. Ziemons, B. Boulanger, P. Hubert, J. Pharm. Biomed. Anal. 55 (2011) 848.
[42] M. Hansen, K.A. Krogh, B. Halling-Sørensen, E. Björklund, Anal. Methods 3 (2011) 1087.
[43] H. Stahnke, S. Kittlaus, G. Kempe, L. Alder, Anal. Chem. 84 (2012) 1474.
[44] B.K. Matuszewski, M.L. Constanzer, C.M. Chavez-Eng, Anal. Chem. 75 (2003) 3019.
[45] M. Hansen, E. Björklund, K.A. Krogh, B. Halling-Sørensen, TrAC, Trends Anal. Chem. 28 (2009) 521.
[46] U.S. FDA, Guidance for Industry - Bioanalytical Method Validation, May 2001 (document online; cited 07.09.12). Available from http://www.fda.gov/ downloads/Drugs/ … /Guidances/ucm070107.pdf
[47] M.B. Kjærstad, C. Taxvig, C. Nellemann, A.M. Vinggaard, H.R. Andersen, Reprod. Toxicol. 30 (2010) 573.
[48] A. Ohlsson, E. Ullerås, A. Oskarsson, Toxicol. Lett. 191 (2009) 174.
[49] M. Hecker, H. Hollert, R. Cooper, A.M. Vinggaard, Y. Akahori, M. Murphy, C. Nellemann, E. Higley, J. Newsted, J. Laskey, A. Buckalew, S. Grund, S. Maletz, J. Giesy, G. Timm, Environ. Sci. Pollut. Res. Int. 18 (2011) 503.
[50] J. Hajslová, J. Zrostliková, J. Chromatogr. A 1000 (2003) 181.
[51] D. Harris, Quantitative Chemical Analysis, 7th ed., W.H. Freeman and Company, New York, 2007.
[52] R. Venn (Ed.), Principles and Practice of Bioanalysis, 2nd ed., CRC Press, Florida, 2008.
[53] A. Van Eeckhaut, K. Lanckmans, S. Sarre, I. Smolders, Y. Michotte, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 877 (2009) 2198.
[54] F.T. Peters, D. Remane, Anal. Bioanal. Chem. 403 (2012) 2155.
[55] H. Runnqvist, S.A. Bak, M. Hansen, B. Styrishave, B. Halling-Sørensen, E. Björk- lund, J. Chromatogr. A 1217 (2010) 2447.
[56] R. MacNeill, A. Farside, V. Acharya, R. Stromeyer, B. Urbanowicz, M. Moussallie, Bioanalysis 4 (2012) 1163.
[57] T. Koal, D. Schmiederer, H. Pham-Tuan, C. Rohring, M. Rauh, J. Steroid Biochem. Mol. Biol. 129 (2012) 129.
[58] Genome Alberta, Genome Canada, Human Metabolome Database Version 3.0, Cholesterol (HMDB00067), 2012 (homepage online; cited 27.10.12). Available from http://www.hmdb.ca/metabolites/HMDB00067
[59] J.G. McDonald, B.M. Thompson, E.C. McCrum, D.W. Russell, Methods Enzymol. 432 (2007) 145.
[60] I. Surowiec, M. Koc, H. Antti, P. Wikstrom, T. Moritz, J. Sep. Sci. 34 (2011) 2650.
[61] Genome Alberta, Genome Canada, Human Metabolome Database Version 3.0, Pregnenolone (HMDB00253), 2012 (homepage online; cited 27.10.12). Avail- able from http://www.hmdb.ca/metabolites/HMDB00253
[62] I. Miksik, K. Mikuliková, J. Pácha, M. Kucka, Z. Deyl, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 800 (2004) 145.
[63] W.J. Chia, A.M. Jenner, A.A. Farooqui, W.Y. Ong, Exp. Brain Res. 186 (2008) 143.
[64] Genome Alberta, Genome Canada, Human Metabolome Database Version 3.0, 17-Hydroxyprogesterone (HMDB00374), 2012 (homepage online; cited 27.10.12). Available from http://www.hmdb.ca/metabolites/HMDB00374
[65] V.M. Carvalho, O.H. Nakamura, J.G. Vieira, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 872 (2008) 154.
[66] Y. Liu, C.E. Uboh, L.R. Soma, X. Li, F. Guan, Y. You, J.A. Rudy, J.W. Chen, Drug Test. Anal. 3 (2011) 54.
[67] Genome Alberta, Genome Canada, Human Metabolome Database Version 3.0, Testosterone (HMDB00234), 2012 (homepage online; cited 27.10.12). Available from http://www.hmdb.ca/metabolites/HMDB00234
[68] T.M. Williams, A.J. Kind, E. Houghton, D.W. Hill, J. Mass Spectrom. 34 (1999) 206.
[69] B.J. Vanderford, R.A. Pearson, D.J. Rexing, S.A. Snyder, Anal. Chem. 75 (2003) 6265.
[70] I.P. Baskova, Z. Ferner, A.S. Balkina, S.A. Kozin, O.V. Kharitonova, L.L. Zavalova, V.G. Zgoda, Biochem. (Moscow) Suppl. Ser. B: Biomed. Chem. 2 (2008) 215.
[71] Genome Alberta, Genome Canada, Human Metabolome Database Version 3.0, Cortisol (HMDB00063), 2012 (homepage online; cited 27.10.12). Available from http://www.hmdb.ca/metabolites/HMDB00063
[72] F. Fanelli, I. Belluomo, V.D. Di Lallo, G. Cuomo, R. De lasio, M. Baccini, E. Casadio, B. Casetta, V. Vicennati, A. Gambineri, G. Grossi, R. Pasquali, U. Pagotto, Steroids 76 (2011) 244.
[73] Genome Alberta, Genome Canada, Human Metabolome Database Version 2.5, Aldosterone (HMDB00037), 2012 (homepage online; cited 11.08.12). Available from http://www.hmdb.ca/metabolites/HMDB00037
[74] U. Ceglarek, L. Kortz, A. Leichtle, G.M. Fiedler, J. Kratzsch, J. Thiery, Clin. Chim. Acta 401 (2009) 114.
[75] H.J. Cho, J.D. Kim, W.Y. Lee, B.C. Chung, M.H. Choi, Anal. Chim. Acta 632 (2009) 101.