Bioanalytical strategies for the quantification of xenobiotics in biological fluids and tissues 2

Chapter 2 Determination of estrogens in ovarian cyst fluid samples by porous membrane protected micro-solid-phase- extraction combined with gas chromatography-mass spectrometry... 2.1 Pr

Chapter 2 Determination of estrogens in ovarian cyst fluid samples by porous membrane protected micro-solid-phase- extraction combined with gas chromatography-mass

spectrometry

2.1 Preface to Chapter 2

To compare the levels of estrogens in benign and malignant ovarian tumor cyst fluids, a cost effective and environmentally friendly extraction technique using porous membrane protected µ-SPE is described A sorbent (ethylsilane (C2) modified silica) (20 mg) was packed in a porous polypropylene envelope (2 cm × 1.5 cm) whose edges were heat sealed to secure the contents The µ-SPE device was conditioned with acetone and placed in a stirred (1:5) diluted cyst fluid sample solution (10 mL) to extract estrogens for 60 min After extraction, the analytes were desorbed and simultaneously derivatized with a 5:1 mixture of acetone and N,O-bis(trimethylsilyl)-trifluoroacetamide The extract (2 µL) was analyzed by gas chromatography–mass spectrometry Various extraction, desorption and derivatization conditions were optimized for µ-SPE With this simple technique, low limits of detection of between 9 and 22 ng L−1 and linear range from the detection limits up to

50 µg L−1 were achieved The optimized method was used to extract estrogens from cyst fluid samples obtained from patients with malignant and benign ovarian tumors The results showed a pattern of higher levels of estrogen accumulation in benign as compared to malignant samples in the samples tested This implies that estrogens might play a role in the malignancy associated with epithelial ovarian cancer along

with other compounding factors

2.2 Introduction

Estrogens are a group of steroid hormones which primarily function is to regulate the reproductive systems of both female and male animals and humans Over the past several decades, estrogens have received much attention due to their association with many types of human gynaecological cancer [1] In 2002, estrogens were first listed as known human carcinogens by the U.S Department of Health and Human Services in its Report on Carcinogens (10th edition), based on sufficient evidence from human epidemiology studies [2] These studies showed that use of estrogen replacement therapy by postmenopausal women is associated with a consistent increase in the risk of uterine endometrial cancer and a less consistent increase in the risk of breast and ovarian cancer Some evidence suggests that use of oral contraceptives may also increase the risk of breast cancer [2] The exposure to estrogens comes from both natural hormones that are secreted by the ovaries (e.g 17β-estradiol and its metabolite estrone) and synthetic forms (e.g 17α-ethynylestradiol and diethylstilbestrol) that are widely found as the ingredient of medication for estrogen replacement therapy, oral contraceptives, and many cosmetics [3, 4] The presence of estrogens in human body fluids such as follicular cyst fluid and nipple aspirate fluid has been demonstrated by many studies [5-8] Therefore, determining the level of estrogens in these body fluids would be very important for the study of their roles in the carcinogenesis of ovarian and breast cancer

The role of estrogen in the progression of gynaecological cancers such as ovarian cancer is well documented [9] A high correlation was reported between the presence of certain types of estrogen receptors (ER) and the prevalence of

gynaecological cancers [10] Determination of estrogens in tumor specimens and accumulating fluids in the cyst (cyst fluid) could reveal information on the cancer Based on this information (estrogen-positive or -negative) the nature of therapy to be administered to patients, and the prognosis of the cancer may be determined following assessment of genes responsible or ER positive and negative status [11,12] The challenges in determining the quantity of estrogen arises from the fact that (i) the amount of cyst fluid sample available is very small; and these (ii) samples are characterized by their complexity Therefore, high preconcentration with efficient sample clean up are required for cyst fluid sample analysis Techniques for extraction

of estrogens in aqueous samples include the established SPE [13], SPME [14] and more recently, polymer-coated hollow fibre microextraction [15] and stir-bar sorptive extraction [16] All these techniques normally require extensive sample clean up from complex samples such as cyst fluid Therefore the aim of this study is to develop a better and alternative procedure for extracting estrogens from cyst fluids, that involves

no or little additional clean up

A novel, low cost and environmentally friendly extraction technique, called porous membrane-protected µ-SPE, was used for the extraction of various target analytes from complex samples without additional sample clean up [17-20] The µ-SPE device consists of sorbent enclosed within a ca 2 cm × 1.5 cm membrane envelope and is ideally suited to the extraction from a limited amount of sample The judicious choice of sorbent materials, and therefore to some extent, the selectivity of µ-SPE can be fine-tuned With the protection afforded by the porous polypropylene membrane, the elimination of substances such as particulates, proteins and humic substances, which can interfere with the extraction, is easily accomplished without

The objective of the study is to develop a µ-SPE technique for the determination of estrogens in benign and malignant human ovarian cyst fluid This is the first instance where the µ-SPE technique is applied to human cyst fluid samples The information regarding the levels of estrogen in tumor ovarian cyst fluids might play an important role in disease diagnostics This work also investigates the feasibility of applying the simple µ-SPE technique to a complex biological matrix

Figure 2.1 Chemical structures of the estrogens studied: (a) diethylstilbestrol, (b)

estrone, (c) 17β-estradiol, (d) 17α-ethynylestradiol

2.3.2 Human cyst fluid samples

Cyst fluid obtained from benign and malignant ovarian tumor samples were collected following approval from the Domain Specific Review Board, National Health Group, Singapore Twenty cyst fluid samples were collected from patients who were diagnosed to have benign and malignant cysts Small volumes of cyst fluid were collected from patients and in initial studies raw cyst fluid samples without dilution were used for µ-SPE, but this resulted in poor precision and significant matrix interference However, sample dilution with ultrapure water to a 1:1 ratio improved the extraction precision and extraction efficiency It is probable that the dilution reduced the extent of interferences by the protein (clogging on the membrane) and the low viscosity of the matrix that allowed more efficient extraction

Standard safety precautions were put in place during the handling of body fluids All body fluids and solvents used in this project were decontaminated according to standard biohazard disposal protocols

2.3.3 GC-MS

Analyses were carried out using a Shimadzu (Kyoto, Japan) QP2010 GC–MS system equipped with a Shimadzu AOC-20i autosampler and a DB-5 (J & W Scientific, Folsom, CA, USA) fused silica capillary column (30 m × 0.32 mm internal diameter, 0.25 µm film thickness) Helium (purity 99.9999%) was used as the carrier gas at a flow rate of 2.0 mL min-1 Samples (2 µL) were injected in splitless mode The injection temperature was set at 300◦C and the interface temperature kept at

280◦C The GC temperature program used was as follows: initial temperature 90◦C held for 2 min, then increased by 30◦Cmin-1 to 280◦C, and held for 2 min The standard mixtures and extracts were analyzed in selected ion monitoring mode with a detector voltage of 1.5 kV

2.3.4 Preparation of µ-SPE device

The preparation of the µ-SPE device has been described previously [19] The µ-SPE device consists of sorbent materials enclosed within a polypropylene sheet membrane envelope To prepare the device, the longer edge of a polypropylene sheet was folded over to a width of ~2 cm The edge of the fold-over flap was then heat sealed using an electrical sealer to the main sheet The fold-over section was then trimmed off from the main membrane sheet The former was then cut (at ~1.5 cm intervals) into individual (2 cm × 1.5 cm) pieces One of the two open ends of each piece was then heat-sealed A glass Pasteur pipet and a glass funnel were used to introduce sorbent (20 mg) into the resulting membrane envelope via the remaining open end that was then heat-sealed to secure the contents Before use, each µ-SPE device was conditioned (ultrasonication for 10 min with 5mL of acetone) and stored

in the same solvent

2.3.5 µ-SPE procedure

For extraction, the µ-SPE device after drying in air for few minutes was placed

in 10 mL of sample solution The sample solution was agitated at 105 rad s-1 for 60 min to facilitate extraction The device tumbled freely within the sample during extraction After extraction, the device was taken out of the sample solution, dried thoroughly with lint free tissue and placed in a 500 µL autosampler vial for desorption 100 µL of acetone and BSTFA mixture (5:1 ratio) was added and ultrasonicated for 8 min After desorption, the µ-SPE was removed from the desorption vial and the extract (~ 80 µL) was kept in a water bath at 60◦C for 20 min Keeping the extract in warm condition before analysis will facilitate the derivatization process especially for biological matrices Finally, 2 µL of derivatized extract was injected into the GC-MS for analysis

2.4 Results and discussion

µ-SPE is an equilibrium based extraction procedure involving the dynamic partitioning of analytes between the sorbent material and the sample solution [19] The analytical factors that influence extraction efficiency such as the type of sorbent, amount of sorbent, extraction time and desorption time, sample pH and ionic strength were evaluated by a stepwise univariate approach

Figure 2.2 Suitability of various sorbents for µ-SPE from spiked samples Samples

were spiked at levels of 10 µg L-1 of each analyte µ-SPE conditions: samples were extracted for 30 min with 5-min desorption by ultrasonication using 150 µL of acetone, and 20 min derivatization at 60◦C; 15 mg of sorbent was used

Initially, the selection of a suitable sorbent was considered Various sorbents including ethylsilane (C2) modified silica, octylsilane (C8) modified silica and octadecylsilane (C18) modified silica, activated carbon, Carbograph, Haye-Sep A and Haye-Sep B were evaluated (Figure 2.2) Estrogens are polar compounds and appeared to have greater interaction with the relatively polar C2 sorbent compared with the others under acidic (pH 2) condition After selecting C2 as a suitable sorbent, the amount of sorbent material was varied from 6 to 20 mg It was found that with increasing sorbent amount, the extraction efficiency increased, as denoted by higher peak areas during GC-MS analysis (Figure 2.3) Placing >20 mg of sorbent in to an envelope made the device too large to fit into the autosampler vial As a result, desorption was not efficient since the device could not be immersed completely in the solvent Thus, 20 mg of sorbent was the maximum amount used in all experiments

Figure 2.3 Effect of sorbent mass on µ-SPE from spiked samples Samples were

spiked at levels of 10 µg L-1 of each analyte µ-SPE conditions: samples were

extracted for 30 min with 5-min desorption by ultrasonication using 150 µL of acetone and 20 min derivatization at 60◦C

The effect of extraction time was investigated since mass transfer is a dependent process Extractions of between 10 and 60 min were studied in order to determine the adsorption profile of the estrogens (Figure 2.4) To facilitate mass transfer and to decrease equilibration time, the sample was stirred at 105 rad s-1continuously at room temperature During extraction, the mass transfer of analyte from the sample solution to the sorbent determines the extraction efficiency [18] A longer extraction time (60 min) gave better analyte enrichment; probably more time was required for the analyte to diffuse through the porous membrane, and onto the sorbent material Since the total time was considerable (88 min comprising of 60 min for extraction and 28 min for desorption and derivatization), we did not further extend the extraction time, and 60 min was selected

Figure 2.4 Extraction time profiles of estrogens Samples were spiked at levels of 10

µg L-1 of each analyte µ-SPE conditions: samples were desorbed by ultrasonication using 150 µL of acetone for 5 min, 20 min derivatization at 60◦C; 15 mg of sorbent was used

The salting-out effect has been widely used to enhance the extraction efficiency of polar compounds in extraction and microextraction techniques [13-16] Addition of salt decreases the solubility of polar analytes in aqueous samples [24, 25] and thus, in this case, favours extraction by the sorbent The effect of salt on extraction efficiency was determined by adding sodium chloride (NaCl) (from 5 to 30% (w/v)) to the sample The highest peak areas were obtained when 5% NaCl was used

0.00E+00 5.00E+06 1.00E+07 1.50E+07

DES Estrone Estradiol Ethynylestradiol

Figure 2.5 Ionic strength profile of estrogens for different salt concentrattion

Samples were spiked at level of 5 µgL-1 of each analyte µ-SPE conditions: samples were extracted for 60 min with 100 µL of acetone as desorption solvent, 20 min derivatization at 60◦C; 15 mg of sorbent was used

Estrogens are ionisable compounds and their extraction behaviour at different sample pH (from 2 to 12) was investigated Sample pH was adjusted by the addition

of 1M hydrochloric acid and 1M sodium hydroxide respectively At a sample pH of 2, better extraction efficiency was achieved when compared to neutral or basic conditions Acidic sample pH had previously been used for extracting these compounds [26] Based on this, a sample pH of 2 was used for further experiments