HPLC-MS/MS multiclass determination of steroid hormones in environmental waters after preconcentration on the carbonaceous sorbent HA-C@silica

2.1 Chemicals and materials

All chemicals were reagent grade or higher in quality. Silica microparticles (40–63 μm, surface area 550 m² g⁻¹, pore volume 0.8 cm³ g⁻¹), humic acids sodium salt (technical grade), formic acid (FA), ammonium hydroxide solution (NH₄OH), ammonium fluoride (NH₄F), polypropylene tubes and polyethylene frits, high purity steroid hormones standards (E1, E2, EE2, TST, PROG, M-PROG, cortisone (CORT), hydrocortisone (H-CORT), prednisone (PRED), prednisolone (PREDLO), betamethasone (BETA), dexamethasone (DEXA) were supplied by Sigma-Aldrich (Milan, Italy). Analytical grade H-PROG standard was purchased from Steroids (Cologno Monzese, Italy). Analytical grade triamcinolone (TRIAM) and fluocinolone (FLUO) standards were purchased from Farmabios (Groppello Cairoli, Italy). HPLC gradient grade methanol (MeOH), acetonitrile (ACN) and ultrapure water were supplied by VWR (Milan, Italy).

Stock solutions of 1000 μg mL⁻¹ of each steroid hormone were prepared in MeOH and stored in the dark (4 °C). Steroid hormones working solutions of 0.15–1 μg mL⁻¹ were prepared by dilution from a 10 μg mL⁻¹ solution, and they were renewed weekly.

2.2 SPE procedure using HA-C@silica

The SPE procedure was carried out placing in a 3 mL SPE cartridge 200 mg HA-C@silica between two polyethylene frits. HA-C@silica is obtained by a thermal treatment of HAs immobilized onto micrometric silica, as reported in previous work (Speltini et al., 2018). Briefly, HAs (100 mg) were dissolved in 50 mL water in a round-bottom flask, and the suspension obtained by adding 1 g silica was magnetically stirred for 2 min. After water evaporation under vacuum, the resulting brownish material was thermally treated at 600 °C for 1 h under N₂ flow into a cylindrical oven (heating ramp 10 °C min⁻¹, cooling ramp 10 °C min⁻¹).

Water samples (500–1000 mL for river water and 250 mL for UWWTP effluent were loaded onto the SPE cartridge at a flow rate of 5 mL min⁻¹, by a multi-channel peristaltic pump (Gilson, Italy) and then the cartridge was dried under vacuum for 5 min. The analytes were eluted (1 mL min⁻¹) with 4 mL MeOH and 4 mL ACN, sequentially. For sample spiking ≤ 25 ng L⁻¹, the extracts were evaporated under N₂ flow and reconstituted with 0.25 mL MeOH, prior HPLC-ESI-MS/MS analysis.

2.3 Water samples

Tap water was from the Pavia municipal waterworks. Northern Italy surface waters were collected in Spring and Summer 2018 at 30–50 cm depth, directly in amber glass bottles, from various rivers (Tanaro, Ticino, Staffora), and at the outlet of Pavia UWWTP. Samples were stored in the dark (4 °C) and analyzed within 24 h of collection to check the absence/presence of the target analytes. Staffora river and UWWTP effluent were selected as blank environmental samples due to the absence of all analytes (<MDLs) as verified from recovery tests and quantification by the standard addition method. The physical-chemical parameters of the water samples are reported in Table S1, Electronic Supplementary Material.

2.4 HPLC-ESI-MS/MS analysis

The target analytes were analysed with a HPLC apparatus Agilent 1260 Infinity coupled with an Agilent 6460C MS spectrometer ESI-MS/MS system (Cernusco sul Naviglio, Italy).

The MS spectrometer was tuned up by direct injection of individual analyte solutions (1 mg L⁻¹ in MeOH), and different experimental conditions were investigated to maximize the signal response and increase detection sensitivity. First, the HPLC-ESI-MS/MS conditions proposed in the previous work (Speltini et al., 2018) were applied, but a poor ionization of the target analytes, in particular of oestrogens, was observed. Based on the literature (Shimko et al., 2019; Agilent Application Note), NH₄F was selected as additive for the aqueous phase to increase sensitivity.

The MS operating parameters, optimized by Agilent Mass Hunter Source Optimizer Software (Agilent, USA) were: drying gas (N₂) temperature 350 °C; drying gas flow 12 L min⁻¹; nebulizer 50 psi; sheath gas temperature 400 °C; sheath gas flow 12 L min⁻¹; capillary voltage 4000 V positive, 3000 V negative; nozzle voltage 0 V positive, 1500 V negative; electron multiplier voltage (EMV) 200 V positive, 0 V negative; cell accelerated voltage 4 V positive, 1 V negative.

The quantitative analysis of the target compounds was performed in multiple-reaction monitoring (MRM) mode, using the two most intense transitions of each compound, as reported in Table 1.

The chromatographic separation was achieved on a Zorbax Eclipse Plus C18 column (4.6 × 100 mm, 3.5 μm), maintained at 25 °C (±0.8 °C), and the injection volume was 5.0 μL. Elution was performed by 1 mM NH₄F in the aqueous mobile phase (A) and ACN (B) according to the following program: 30% B for 0.5 min, linear gradient to 85% B in 12 min, maintenance of 85% B for 5 min, washing step with 100% B for 2.5 min. The initial conditions were re-established by 7-min equilibration time (0.5 mL min⁻¹ flow rate, column temperature 25 ± 0.8 °C). A typical MRM chromatogram of a standard solution (150 μg L⁻¹) is reported in Fig. 2.

Close

Fig. 2

A typical MRM chromatogram of a standard solution of sex hormones (150 μg L⁻¹).

Export to PPT

2.5 Analytical evaluation of the SPE/HPLC-ESI-MS/MS procedure

In term of a consistent application of the proposed procedure to multiclass enrichment of steroid hormones from real-world water samples, the entire analytical procedure was within-laboratory evaluated based on the main figures of merit (Speltini et al., 2018; Rui Zhang et al., 2014; Tölgyesi et al., 2010), viz. trueness, precision, linearity, detection and quantification limits.

3.1 Solid-phase extraction on HA-C@silica

As described in Section 2.2, this work focuses on multiclass enrichment of three classes of steroid hormones, using HA-C@silica, a mixed-mode sorbent material recently successfully applied by this research group for preconcentration of GCs (Speltini et al., 2018) and benzotriazoles and benzothiazoles (Speltini et al., 2019) from environmental waters. Due to its peculiar characteristics – viz. the presence of sp² hybridized carbon and polar groups- HA-C@silica was here tested for the extraction of oestrogens, progestins and androgens. It was investigated firstly in tap water samples (500 mL) enriched with 1 μg L⁻¹ of each compound. Based on literature data about the commercial SPE cartridges (Pailler et al., 2009; Moreira et al., 2011; Jauković et al., 2017; Kumar et al., 2009; Kuster et al., 2009; Chen et al., 2012), the extraction was carried out at the sample native pH and under acidic conditions (pH 3). Generally (data shown in Table S2, Supplementary material), the pH influence was not so pronounced considering the RSDs, although higher extraction efficiencies were achieved for all analytes at native pH values. Therefore, the extraction was done with no pH adjustment.

In these preliminary tests, elution was performed with MeOH (4 mL), which provided quantitative recovery for all analytes except for PROG (50%). Hence, further tests were done in the same experimental conditions (500 mL tap water spiked with 1 μg L⁻¹ of each hormone) to assess the performance of other eluents: MeOH + 2% v/v NH₄OH (alkaline solvent), MeOH + 0.2% v/v FA (acidic solvent) and ACN (aprotic solvent). Alkaline and acidic MeOH did not lead to improvements in the elution of PROG, while the aprotic solvent provided quantitative recovery for this compound, which indeed does not present —OH groups in the skeleton. On the other hand, ACN provided lower recovery than ones obtained with MeOH for the other compounds, as reported in Table S3, Supplementary material. By these evidences, it seemed clear that the SPE procedure has to be performed with elution in two sequential steps, using 4 mL MeOH followed by 4 mL ACN.

HA-C@silica was then tested first in a river water sample not containing the analytes (blank sample from Staffora river) and on the same sample after spiking (500 mL enriched at 25–600 ng L⁻¹) to verify the recovery. As it is shown in Table 2, good performance was obtained at the different concentrations, with recovery slightly lower for PROG. The highest EF (4000) was achieved by preconcentrating 1 L of river water enriched with 5 ng L⁻¹, obtaining satisfying recoveries also at the lowest concentration level for at least three consecutive extractions (71–124%).

The SPE procedure was then evaluated on a more complex aqueous matrix, that is UWWTP effluent (physical-chemical parameters are reported in Table S1, Electronic Supplementary Material). In this matrix, accurate recoveries were achieved processing a sample volume of 250 mL at the three concentration levels investigated (25, 300 and 1000 ng L⁻¹), for all contaminants (see Table 2).

These results collected on different water matrices demonstrate the real applicability of HA-C@silica for preconcentration of different classes of sex hormones in environmental waters at low nanograms per litre levels, with results comparable to those reported in the most recent literature (Zhang and Fent, 2018).

3.2 HA-C@silica SPE followed by HPLC-ESI-MS/MS: Figures of merit

3.3 Determination of steroid hormones in real-world samples

Based on the good results obtained by recovery tests, the proposed analytical method was applied to enrichment of real-world surface waters (Tanaro and Ticino rivers) collected in high-density populated areas of Northern Italy, and of UWWTP effluent sampled at the Pavia WWTP upstream the UV treatment. As expected and as reported in Table 3 and Fig. S1, steroid hormones were detected at few nanograms per litre, attesting their environmental diffusion in surface waters. In particular, high levels of oestrogens were found, probably due to their extensive usage and poor metabolization (Huysman et al., 2017; Barreiros et al., 2016). The results highlight the applicability of the proposed SPE coupled to HPLC-ESI-MS/MS method for environmental monitoring of steroid hormones at the low nanograms per litre levels. However, deeper monitoring investigations are required to be more aware about the actual contamination levels of these xenobiotic compounds in the water ecosystems considered.

3.4 Application of method extended to GCs in river water

The good preconcentration results obtained on HA-C@silica in this work and in the previous one for GCs (Speltini et al., 2018) prompted us to verify the applicability of the HPLC-ESI-MS/MS method for the simultaneous and quantitative determination of the four classes of steroid hormones (oestrogens, progestins, androgens and GCs). For this reason, eight GCs (Speltini et al., 2018) were separated and fragmented in the new chromatographic conditions, proposed in this work (see Electronic Table S4, Supplementary Material).

Recovery tests were then performed in triplicate on the blank river water matrix (1 L) enriched with 5 ng L⁻¹ of the fifteen analytes, obtaining quantitative recoveries (see Table 4).

At the same time, an aliquot of Ticino river sample was analysed; oestrogens, progestins, androgens were confirmed (see Table 3) and among GCs, CORT, H-CORT and DEXA were quantified at concentrations of 2–3 ng L⁻¹, (see Fig. S2, Supplementary Material).

These findings show that all compounds are quantitatively pre-concentrated on the same cartridge and quantified in a single chromatographic run, further widening the method applicability compared to the most recent methods proposed in the literature (Zhang and Fent, 2018).