Quantifying steroid hormones in amniotic fluid by ultra-performance liquid chromatography and tandem mass spectrometry

Simultaneous assessment of steroid hormone Background: concentrations in amniotic fluid is of importance for elucidating long-term consequences of intrauterine processes, and of broad scientific and clinical relevance. The objective of the study was to develop sensitive and specific analytical ultra-performance liquid chromatography tandem mass spectrometry (UPLC-MS/MS) methods for determination of steroid hormone concentrations in human amniotic fluid, and to provide proof-of-concept of their applicability. Methods were validated according to linearity, limit-of-detection, Methods: limit-of-quantification, recovery, intraand inter-assay precision, and applied to 275 amniotic fluid samples. Limits-of-quantification (S/N 10:1) were 0.05 ng/mL for cortisol, Results: = dehydroepiandrosterone sulfate (DHEAS), estradiol, estriol, and testosterone, and 0.01 ng/mL and 1.0 ng/mL for cortisone and dehydroepiandrosterone (DHEA), respectively. Good interand intra-assay precision were observed. Cortisol, cortisone, DHEAS, estradiol and estriol concentration were quantified in all samples. By lack of sensitivity of the analytical method, DHEA was quantified in 11 samples only. Testosterone was quantified in 119 of 275 samples analyzed. 116 of the quantified testosterone samples were from male offspring (out of 138 male offspring). These specific and sensitive methods offer a simple and Conclusion: 1 1,2 2,3 1 4 4 4 4 5 5 6-9 10 1 2 3 4 5 6 7 8 9 10

Amniotic fluid, with its nutritive and protective functions, is essential for fetal well-being and development 1 . It is contained by the amniotic cavity enclosed by the amnion and chorion 2 . Amniotic fluid originally comes from maternal plasma and passes through the fetal membranes, being fueled by secretions from the fetal respiratory and gastrointestinal tract, from the oral cavity, and from fetal skin before keratinization occurs 1,3-5 . Furthermore, cells from the amnion layers secrete proteins into the amniotic fluid 2 . At the beginning of pregnancy, the biochemical composition of human amniotic fluid resembles fetal extracellular fluid, probably due to passage across the still unkeratinized fetal skin 6 . Given the interplay between amniotic fluid and fetal physiology, amniotic fluid has been used as an important source of potential biomarkers for fetal pathologies 7 .
Therefore, there is a strong need for powerful and easily applicable analytical procedures for research as well as diagnostic purposes, to assess fetal steroids in amniotic fluid, overcoming previous limitations of sensitivity and specificity. However, as yet, respective methods are scarce. Steroid hormones can be quantified by traditional immunoassays but with several limitations as a lack of sensitivity making the quantification of small amounts of steroids difficult or a lack of specificity due to antibody cross-reactivity resulting in results higher than true concentrations 25 . The presence of interfering substances such as autoantibodies, exogenous substances, hemolysis or lipaemia in patient samples can alter the measurable concentration of the analyte 26 . The matrix effect problems also affect immunoassays. Furthermore, a relatively large sample volume is required as immunoanalysis provides only single-analyte assays and this is particularly a problem for precious samples such as amniotic fluid. Mass-spectrometric techniques are an alternative to address immunoassays issues. It allows for highly sensitive and specific simultaneous assessment of multiple biomarkers, for example in the context of metabolic and proteomic profiling in amniotic fluid 7, 25 . Recently, first studies reported on the successful application of different mass-spectrometric approaches, including gas-chromatography mass spectrometry (GC-MS) and liquid-chromatography tandem mass spectrometry (LC-MS/ MS), for concomitant assessment of multiple selected steroids in amniotic fluid [26][27][28] .
The main objective of this study was to develop highly sensitive and specific analytical ultra-performance liquid chromatography tandem mass spectrometry (UPLC-MS/MS) methods for determination of cortisol, cortisone, DHEA, DHEAS, estradiol, estriol, and testosterone concentrations in human amniotic fluid, and provide proof-of-principle for the method's applicability in amniotic fluid samples of pregnant women, collected within obstetric care.

Methods
Study participants and amniotic fluid collection 282 pregnant women were recruited via Prenatal Medicine and Genetics, Duesseldorf, Germany, who were undergoing regular amniocenteses for diagnostic reasons with maternal age as the only risk factor. Samples from seven women were not included in the analyses because of problems with either the consent forms or the amniotic fluid samples. Sociodemographic and obstetric characteristics of the 275 women included in the analyses are presented in Table 1. Amniotic fluid samples were collected in the years 2010 to 2012, between week 14 and week 18 of gestation. Samples were collected, using GREINER BIO-ONE (Frickenhausen, GERMANY) tubes CRYO'S 124 and stored at -80°C for later analyses. Biochemical analyses were conducted in 2013 at the laboratory of the Institute of Legal Medicine at the University of Strasbourg, France.
All participants provided written informed consent that their samples were used in research after genetic examinations. The study protocol was approved by the Ethical Committee of Faculty of Science and Mathematics, Heinrich-Heine-University, Duesseldorf, Germany.

Solution preparation.
Cortisol, testosterone and estradiol solutions were prepared in MeOH at a final concentration of 2.5, 25 and 250 ng/mL. Cortisone, DHEAS, estriol solutions were prepared in MeOH at a final concentration of 2.5, 25, 250 and 2500 ng/mL. DHEA solutions were prepared in MeOH at a final concentration of 0.025, 0.25, and 2.5 and 25 µg/mL. Deuterated cortisol was prepared in MeOH at a final concentration of 250 ng/mL. These solutions were stable for at least 6 months at 4°C. Sorensen buffer was prepared by adding 38.8 mL KH 2 PO 4 (9.07 g/L) to 6.12 mL Na 2 HPO 4 (11.87 g/L); pH value was adjusted to pH 7.6. NaHCO 3 buffer was prepared by adding 1 L of distilled water to 8,4 g of NaHCO 3 ; pH value was adjusted to pH 10.5. Dansyl chloride solution was prepared by adding 20 mL of acetone to 20 mg of dansyl chloride.
Calibration standards and quality control. Distilled water was used for calibration standards. All molecules of interest were thus not detectable in the blank samples. Calibration standards were prepared at concentrations ranging from limit-of-quantification (LOQ) to high concentration for each analyte. Intra-and inter-assay precision were determined by enriching 250 µL of distilled water with the analytes of interest at three concentrations (low, medium and high) for each hormone.
Sample preparation for cortisol, cortisone, DHEA, DHEAS and testosterone determination. 0.25 mL of amniotic fluid were mixed in 0.75 mL Sorensen buffer (pH 7.6) in the presence of 20 ng/mL cortisol-d4 as internal standard. For further purification, SPME Oasis® HLB extraction plates were used. Activation was operated with 0.2 mL MeOH, followed by 0.2 mL deionized water. The incubation medium was centrifuged and the supernatant was removed and deposited on the activated plate, then rinsed with 0.2 mL deionized water/MeOH (95:5, v/v). The plate was allowed to dry for 5 minutes at room temperature. Analytes were eluted with 35 µL ACN/isopropanol (40:60, v/v) with 2% of concentrated NH 4 OH, followed by 35 µL deionized water. Ten µL of this extract were directly injected into the UPLC-MS/MS system.

Sample preparation for estradiol and estriol determination.
3 mL of diethylether was added to 0.25 mL of amniotic fluid in the presence of 20 ng/mL cortisol-d4 as internal standard. Samples were agitated 15 minutes. Samples were centrifuged 15 minutes at 3000g and the organic phase was transferred into a new glass tube. The supernatant was evaporated at 40°C under a constant stream of nitrogen until the samples were completely dried. Finally, 50 µL of NaHCO 3 buffer (pH 10.5) and 50µL of dansyl-chloride solution were added and the tube was vortexed for 30 second. The mix was heated 3 minutes at 60°C. Ten µL of this extract were directly injected into the UPLS-MS/MS system.

Chromatographic and mass spectrometric conditions.
A Waters (Milford, USA) Acquity UPLC system with a column heater, autosampler, and a 10µL injection loop was used. Analytes were separated at 30°C on a Waters Acquity UPLC BEH C18 column (1.7 µm, 100 x, 2.1-mm). Separation was achieved by gradient elution with 0.1% HCOOH (pH 2.6) and ACN at a flow rate of 0.4 mL/min (0 to 6 min, 90% HCOOH and 10% ACN; 6 to 9 min, 40% HCOOH and 60% ACN; 9 to 11 min, 90% HCOOH and 10% ACN). The total run time was 11 minutes, including periods required for injection and equilibrating the column before the next injection.
Detection was carried out by a Quattro Premier XE tandem mass spectrometer (MS/MS) (Waters Micromass, Manchester, UK). This mass spectrometer was equipped with an electrospray ionization probe and operated switching between positive and negative ionization mode.

Method validation
The following parameters were tested to validate the method: limit-of-detection (LOD), LOQ, linearity, recovery, intra-and inter-assay precision.
The LOD and LOQ were determined by analysis of replicate blank samples (n=6) spiked with hormones at various concentrations. The LOD and LOQ were estimated as giving a signalto-noise ratio greater than 3 and 10 respectively, for each of the quantitative ions transitions monitored.
Linearity was tested by the preparation of calibration curves ranging from the LOQ to 20 ng/mL for cortisol, to 75 ng/mL for cortisone, to 500 ng/mL for DHEA, to 50 ng/mL for DHEAS, to 10 ng/mL for testosterone, to 100 ng/mL for estriol and to 25 ng/mL for estradiol. Three curves were developed for 9 concentration levels of each hormone. Linearity of the method was expressed by the correlation coefficient (r 2 ).
Analyte recovery of the extraction procedure was determined by analyzing replicate blank samples (n=3) spiked with hormone (LOQ, medium and high concentration) against replicate blank extracts (n=3) spiked at the same levels after extraction.
Precision was evaluated using three solutions with all hormones (LOQ, medium and high concentration) and expressed as coefficient of variation (CV). The intra-assay precision was assessed by determining these samples on one day (n=10 for each sample), while the inter-assay precision was assessed over 8 days (n=8 for each sample).

Cortisol, cortisone, DHEA, DHEAS, testosterone assessment LOD, LOQ and linearity
The LOD and LOQ, as well as the CVs are presented in Table 2, for both cortisol, cortisone, DHEA, DHEAS, testosterone.
The calibration curve obtained showed good linear responses with a r 2 of 0.9974, 0.9907, 0.9956, 0.9856 and 0.9940 for cortisol, cortisone, DHEA, DHEAS and testosterone respectively from the range of 0.05 to 20 ng/mL for cortisol, 0.01 to 75 ng/mL for cortisone, 1 to 500 ng/mL for DHEA, 0.05 to 50 ng/mL for DHEAS and 0.05 to 10 ng/mL for testosterone. Table 3 shows the recovery tested for each analyte at three different concentrations (LOQ, medium and high concentrations). The extraction recoveries were > 71%, > 86%, > 73%, > 80% and > 52% for cortisol, cortisone, DHEA, DHEAS and testosterone respectively.

Precision
Intra-(n=10) and inter-assay (n=8) precision were evaluated for each hormone at 3 concentrations (LOQ, medium, and high). The intra-assay (n=10) precision values were less than 20% for all hormones at LOQ and less than 15% for all hormones at medium and high concentration. Good inter-assay (n=8) values (less than 20% for both analytes at LOQ and less than 15% at other medium and high concentrations) were obtained, as outlined in Table 4.

Estradiol and estriol assessment LOD, LOQ and linearity
The LOD and LOQ, as well as the CVs are presented in Table 5, for both estradiol and estriol.
The calibration curves showed good linear responses with a r 2 of 0.9967 and 0.9979 for estradiol and estriol respectively, within the range of 0.01 to 25 ng/mL for estradiol and 0.05 to 100 ng/mL for estriol. Table 6 shows the recovery tested for each analyte at three different concentrations (LOQ, medium and high). The extraction recoveries were > 77% and > 92% for estradiol and estriol respectively.

Precision
Intra-(n=10) and inter-assay (n=8) precision were evaluated for estradiol and estriol at 3 concentrations (LOQ, medium and high). The intra-assay (n=10) values were less than 16% for both estradiol and estriol and good inter-assay (n=8) values (less than 17% for both analytes) were obtained, as demonstrated in Table 7.
Application on amniotic fluid samples. UPLC-MS/MS analyses were performed on 275 amniotic fluid samples (see Table 8, Figure 3 and Figure 4). Cortisol, cortisone, DHEAS, estradiol and estriol concentrations were quantified in all samples. DHEA was quantified in 11 of the 275 samples analyzed. Testosterone was quantified in 119 of the 275 samples analyzed (samples from pregnancies with male offspring -all twin or higher order pregnancies resulted in male offspring only: quantified in 116 out of 138 samples analyzed; samples from pregnancies with female offspring: quantified in 3 out of 126 samples analyzed). For DHEA, in 35 of the analyzed samples, concentrations were below the LOQ of 1 ng/mL but above the limit-of-detection of 0.75 ng/mL. For testosterone, in 39 of the analyzed samples, concentrations were below the LOQ of 0.05 ng/mL but above the limit-of-detection of 0.02 ng/mL.

Discussion
We developed an analytical method to simultaneously measure cortisol, cortisone, DHEAS, and testosterone, and an analytical method to simultaneously measure estradiol and estriol in human amniotic fluid, using an LC-MS/MS assay, with good linearity, sensitivity, specificity and accuracy without interferences between the molecules and a low LOQ. Five out of the seven hormones were quantifiable in all amniotic fluid samples, while DHEA was quantifiable only in a few samples and testosterone was primarily quantifiable in samples from women carrying a male fetus.
The results add to previous findings that indicated that different mass-spectrometric approaches, including gas-chromatography mass spectrometry (GC-MS) and LC-MS/MS, can be applied for assessment of steroids in amniotic fluid 27,28 .
Comparisons of our results on steroid concentrations in human amniotic fluid with data from previous studies are hampered by substantial heterogeneity across previous reports. For example, even when restricting comparisons to findings also based on LC-MS/MS, cortisol concentrations in human amniotic fluid have previously been reported as higher 28 as well as Abbreviations: CV, coefficient of variation; DHEA, dehydroepiandrosterone; DHEAS, dehydroepiandrosterone sulfate; LOD, limit-of-detection; LOQ, limit-of-quantification assessed via mass spectrometric approaches 27,28 , and with previous findings on estradiol and estriol concentrations in human amniotic fluid, assessed via immunoassay-based approaches 32,33 (we are not aware of any respective information from previous studies using mass spectrometric approaches).
Our study has several strengths. We developed analytical methods to quantify a range of steroid hormones in human amniotic fluid with low minimal amount of sample and a total analytical run time of 11 minutes, allowing simultaneous assessment of several hormones. Using state-of-the-art approaches, we confirmed    Abbreviations: CV, coefficient of variation; LOD, limit-of-detection; LOQ, limit-of-quantification       35 . In view of these studies, the LOQ of our method was not sensitive enough to detect a low concentration of testosterone in amniotic fluid particularly for female fetuses. However, identifying a concentration as being below LOQ may also be of diagnostic value. Second, timing of amniotic fluid collection was not evenly distributed across pregnancy, so we cannot draw conclusions regarding changes in amniotic fluid steroid concentrations throughout gestation. Third, amniotic fluid samples were not immediately assessed after collection, so we cannot exclude degradation of analytes between sample collection and analysis, even though storage conditions would have made substantial degradation rather unlikely. Finally, as amniotic fluid is not routinely collected during every pregnancy, we cannot assume that our sample of study participants is representative with regard to the population of pregnant women. However, we drew our participants from women, undergoing amniotic fluid collection during obstetric care just for the reason that the pregnant women were above 30 years of age.
Hence we expect our sample of study participants being largely comparable to a typical sample of women undergoing amniotic fluid testing.
Future studies should further address the role of storing conditions and time interval between collection and analyses. Moreover, future studies should aim at further increasing the sensitivity of the methods. Finally, further studies should test the functional significance of the methods and provide data on their validity and applicability in clinical and research settings.
The methods may provide new opportunities for future applications in a range of fields, such as endocrinology and beyond, providing a tool to determine steroid concentrations in pathological conditions characterized by disturbances of steroid hormones and related enzymatic activities.
We here demonstrated that cortisol, cortisone, DHEAS, estradiol, estriol, and testosterone concentrations are easily detectable and quantifiable in human amniotic fluid, using state-of-the-art and reproducible LC-MS/MS techniques with good linearity, sensitivity, specificity, and accuracy without interferences for the molecules and with a low LOQ. The methods may provide an appropriate approach for a wide range of clinical and research applications, including the field of endocrinology and others.

Data availability
Individual (non-aggregated) data cannot be made publicly available, due to ethical restrictions. In order to access this data, data must be requested from the corresponding author. Data requestors will have to provide: i) written description and legally binding confirmation that their data use is within the scope of the study, as outlined in the ethical board request and written informed consent (respective information will be provided by the corresponding author on request); ii) detailed written description and legally binding confirmation of their actions to be taken to protect the data (e.g., with regard to transfer, storage, back-up, destruction, misuse, and use by other parties), as legally required and to current national and international standards (data protection concept); and iii) legally binding and written confirmation and description that their use of these data is in line with all applicable national and international laws (e.g., the General Data Protection Regulation of the EU). In case of requests of hormone data only, less restrictive regulations may apply. The funders had no role in study design, in the collection, analysis and interpretation of data, in the writing of the report, and in the decision to submit the paper for publication.