Determination of Hypoglycemic Agents in Surface Water Samples Using SPE-LC-MS/MS Method

Antidiabetic compounds are a class of emerging contaminants in environment, for which there are no regulations in the world environmental legislation. These compounds are among the most widely used drugs in the world due to the large number of patients with diabetic conditions. The presence of these pollutants in the environment is insufficiently studied, so efficient analytical methods are needed to allow their detection at trace levels (ng/L). For the simultaneously quantification of the five antidiabetics (glyburide, metformin, glipizide, gliclazide, glimepiride) and one bio-degradation product (guanyl urea) in surface water samples a SPE-LC-MS/MS (solid phase extraction -liquid chromatography coupled with mass spectrometry detection) method was validated using real river water samples. The compounds were separated on C18 LC column in 9 minutes at 30C using a gradient of mobile phase of 0.1% formic acid and acetonitrile. Good performance parameters were obtained using the method: low limits of quantification (LOQs 0.1-2.4 ng/L), precision (repeatability 3.5-7.2% and reproducibility 6.5-12.7%) and determination coefficients (higher than 0.99). The most contaminated river was represented by Ialomita, which had a total concentration of antidiabetics of 112.1 ng/L in the downstream point, followed by the Siret and Dambovita rivers, which had a total concentration of antidiabetics of 66.3 ng/L and 57.3 ng/L, respectively, also in the downstream points.


Introduction
A large variety of pharmaceuticals such as antidiabetics, β-blockers, analgesics, antibiotics, antidepressants, lipid regulators, hormones have been monitored and detected in the environment, particularly in surface waters and wastewaters [1,2]. A high number of administered pharmaceuticals passes the human body unchanged by excretion and enters into wastewater. The excreted and unchanged pharmaceuticals pass the sewage treatment plant (STP) and the incomplete removal contributes to environmental presence [3]. The presence of pharmaceuticals residues in the aquatic environment represent one of the most urgent emerging environmental issues [4]. The active substances used in obtaining pharmaceuticals from antidiabetic class were used in the treatment of diabetes mellitus or prediabetes treatment. Antidiabetics prescribed include the next classes: meglitinide (repaglinide), sulfonylurea derivatives (gliclazide, glibenclamide, glimepiride), biguanidine (metformin) [5]. These drugs were frequently detected in WWTP influents at ng/L concentration level, whereas in some cases comparable concentrations in the treated effluent were noticed. In 2019, the International Diabetes Federation reported the number of diabetic patients worldwide (20-79 years) as 463 million [6]. In Romania it was estimated in 2019 that the number of diabetes patients reached 900,000 [7]. Metformin (N, N-dimethyl-biguanide) is the most consumed antidiabetic for treat type 2 diabetes, but also it is prescribed as a cytostatic product [8,9]. Because MET is consumed intensive by a large number of diabetic patients, it has a high polarity (low octanol water partition coefficient, log Kow -2.6), is not metabolized by the human body and is eliminated unchanged by urine (90%) in 12 h https://doi.org/10.37358/RC. 20.7.8252 and the rest by feces, is expected to be present in the influent treatment plants from where it is eliminated by effluent in the receiving rivers [10][11][12]. MET was determined with high concentrations of the order ng/L in the surface waters analyzed [13,14].
Scheurer et al., reported in 2009 the occurrence of metformin in Germany, in three WWTPs with a median of 110 μg/L in the influent and 11.4 μg/L in the effluent, respectively [15]. In German river waters, MET was detected in a range from 102 ng/L in the Lake Constance, 349 ng/L in the Weser river, 100 ng/L in the Rhine river, up to 1700 ng/L in the Elbe river [10,16]. In Belgium, Metformin was detected in all WWTPs influent samples ranging from 20 μg/L to 94 μg/L with a medium concentration of 46 μg/L [17]. Kolpin et al. reported on the occurrence of metformin in United States surface waters and metformin was detected in 4.8% of 84 samples investigated with a maximum concentration of 0.15 μg/L and a medium concentration of 0.11 μg/L [18]. In China, metformin was detected in eleven Wastewater Treatment Plants (WWTPs) ranged from 1.7 μg/L to 239.0 μg/L, with an average value of 68.3 μg/L [19]. Glibenclamide (GLB), also known as glyburide is extensively metabolized, mainly by hydroxylation of the cyclohexyl moiety of the molecule, whereas its excretion rate as a parent compound are rather low, 35 and 42% in urine and feces, respectively. GLB was determined in surface waters at ng/L to µg/L levels [20]. Glimepiride (GMP) and gliclazide (GLZ) are metabolized in the human body, generating metabolites that exhibit pharmacological activity. Gliclazide was detected in river water at low concentrations at ng/L levels [20].
Ecotoxicological information for selected compounds is still too limited, particularly regarding chronic and behavioral data. For metformin a LC50 value of >982 mg/L for Lepomis macrochirus and an EC50 value of 130 mg/L for Daphnia magna were reported. Guanyl urea showed no toxic effects on the bacterial community in a manometric respiratory test at a concentration of 11.9 mg/L [21]. The physical/chemical properties of selected compounds are presented in In Romania, data on the presence of antidiabetics in surface waters, intended for the production of drinking water, are not available. In general, environmental studies on pharmaceutical contaminants have focused on the following chemical classes: antibiotics (macrolides, sulfonamides, quinolones, penicillin's, tetracyclines), non-steroidal anti-inflammatory drugs (NSAIDs), antiepileptics, lipid regulators [23][24][25][26][27][28][29][30]. Also, multiple studies have been carried out for the emerging contaminants of the type of metallic elements present in the environmental components which are not regulated in the national environmental legislation [31,32].
Thus, it is necessary to carry out analytical studies for the determination of the antidiabetic's concentrations (metformin, gliclazide, glimepiride, glyburide, glipizide and one degradation product: guanyl urea), from surface waters and for the evaluation of the potential impact of the effluents that they have on the quality of the receiving rivers. The main aim of the paper was to validate an SPE-LC-MS/MS method that would allow the quantification of hypoglycemic agents at traces (ng/L) levels in surface waters. Then, these concentration values were used for quantitative estimation of the potential impact of wastewater treatment plants that discharge waste water on the receiving inland waters. This was achieved by comparing the levels of antidiabetics from river samples taken downstream from those collected upstream from the treatment plants. Antidiabetic pollutants from WWTP effluents are continuously introduced into receiving rivers where they can irreversible affects the aquatic https://doi.org/10.37358/RC.20.7.8252 microorganisms. Then, the surface water potentially contaminated with antidiabetics represents the source for obtaining drinking water for the resident population. Thus, rigorous analytical control is required regarding the occurrence of these compounds in surface water and in the drinking water.

Instrument/Equipment and Operating Parameters
The analytical determinations of antidiabetic contaminants were conducted on a 1260 UHPLC system (Agilent Technologies, Germany) which was equipped with a triple quadrupole mass spectrometer (QQQ) Model 6410 Agilent (Agilent Technologies, Waldron, Germany). The compounds were separated on C18 LC column, in 9 minutes, at 30 0 C using a gradient of mobile phase of 0.1% formic acid and acetonitrile (0.2mL/min flow rate). The injection volume for calibration standard and for sample extract was 10 µL. The ionization of compounds was realized by positive electrospray ionization in MS source using the optimal parameters shown in Table 2, 3. The system was controled by Mass Hunter software from Agilent Technologies. Formic acid was used in mobile phase to obtain good peak shape and for the production of the precursor ion [M-H] + . Ionization of compounds was performed using the next optimized settings: gas temperature 300 0 C, capillary voltage, 3000 V, nitrogen nebulizer gas flow rate (10 L/min), nebulizer pressure 50 psi, the cell acceleration voltage (CAV) 4 V, collision energy 10-25V, fragmentor voltage 80-120V.  The two most intense product ions were selected for the analysis. Table 4 presents the optimized mass spectrometer /QQQ parameters for the determination of antidiabetic drugs in the environment samples. The adduct [M-H] + was used as the precursor ion for MS determinations in the positive ionization mode. The first product ion as abundance was used for quantification and the second ion as abundance for confirmation.

Samples Treatment and Analysis of the Antidiabetics
The SPE-LC-MS / MS method previously developed for wastewater samples (effluent and influent treatment plants) was validated using surface water. Thus, the method used a volume of 500 mL samples of river water that was subjected to the entire procedure without addition of standard and with known addition of standard. It was used an enrichment factor of 500 for each water sample. First the sample was filtered on a 0.45 µm glass filter (4.7 mm diameter) to remove the suspended matter that can block the SPE extraction material. Then, the pH of the sample was adjusted to 10 with 0.24% ammonium hydroxide, after which the entire volume of water was passed through SPE cartridge, preconditioned with 2x4 mL methanol and 2x4 mL ultrapure water with pH 10. The extraction was performed on an automatic SPE equipment type Dionex Autotrace 280 (Thermo Scientific) using cartridges Strata X (500mg / 6mL, Phenomenex). To remove the interfering matrix, the adsorbent material was washed with ultrapure water, after that the cartridge was air dried (20 min.) in order to remove traces of water. The compounds of interest were eluted from the SPE cartridge with 2x3 mL methanol. To change the extraction solvent, the obtained extracts were evaporated in a water bath (50 0 C) under a dry nitrogen stream, after which the organic residue was taken up with 1 mL of the mobile phase (0.1 formic acid: acetonitrile, 50/50).

Validation Study
The method has been validated in terms of linearity, limit of quantification, efficiency of recovery and precision (repeatability, reproducibility). Linear regressions (1-100 ng/mL) were obtained for each antidiabetic compound by injecting of 5 calibration solutions with increasing concentrations. Regressions were accepted if the coefficients of determination were over 0.99. The limit of quantification was calculated as the minimum analyte concentration that can be determined from a spiked surface water sample with a concentration of compound for which the signal to noise (S/N) ratio is 10, following the entire extraction and analysis process. The recovery was experimentally determined from a river water sample with the addition of 50 ng/L calibration mix solution. Also, the water sample was analyzed without addition, and the determined antidiabetic drugs were subtracted from the sample with addition. A recovery of 70-120% was considered good for accuracy experiments. To calculate the precision, to 4 sub-samples of surface water (500 mL) were added 1 mL calibration solution, 50 ng/L mixture of antidiabetics in the mobile phase. The samples were extracted and analyzed in the same day, determining the repeatability, expressed as RSD (residual standard deviation), and on different four days, calculating the reproducibility. Precision was accepted if the values of repeatability and internal reproducibility were below 15%.

Sample Collections
Surface water samples were collected in November 2019 in a single day, from 5 rivers (Siret, Bahlui, Ialomita, Dambovita, Somes) (Table 5). Thus, samples were collected from downstream and upstream of the municipal wastewater treatment plant (WWTP). Sampling points were located 100m before the station (upstream) and 50m after the station (downstream). The sample was collected in a 1L volume glass falcon, then stored at 4 0 C during transport to the laboratory and extracted within 48 h.

Validation of method
The mass spectrometer detector response was linear in the range of 1-100 ng/mL for all compounds except Gua (5-100ng/mL), yielding coefficients of determination between 0.99-0.998 (Table 6, Figure 2). The limit of quantification had low values (0.10-2.45 ng/L) allowing the simultaneous determination of antidiabetics from surface water samples using the LC-MS/MS method. The method presented corresponding performance characteristics as repeatability, generating intraday precision within the range of 3.5-7.2% and internal reproducibility (inter-days precision) in the range of 6.5-12.7%. The validation parameter proves that the method is sensitive, accurate and precise. The validation parameters of the method and data obtained by the external standard calibration methodology for all antidiabetic compounds in surface water are presented in Table 3.

Antidiabetics Occurrence in Surface Water
A total of 10 surface water samples (from 5 receiving rivers) taken in the downstream and in the upstream points of the wastewater treatment plants of some municipalities were analyzed in order to determine the concentrations of antidiabetics. At the same time, the aim of this study was to evaluate the chemical quality of surface water used in the production of drinking water. Compounds that were ubiquitous (100% detection frequency) in all samples analyzed were metformin and guanyl urea being detected in all 5 rivers both upstream and downstream of the treatment plants. The compound detected with the highest frequency (90%) was gliclazide followed by glipizide which was determined only in 50% of the samples. The glibenclamide and glimepiride were never detected in the analyzed surface waters.  In order to assess the potential impact of WWTP effluents discharged in river, we calculated an increase factor (IF) for each antidiabetic compounds with equation (1): were, cdw is the antidiabetic concentration in surface water sampled from downstream of the WWTP and cup is the antidiabetic concentration in surface water taken from river in the upstream of WWTP.
Regarding the potential impact of the WWTPs on the surface water quality, it was observed that the Bahlui river presented the highest increase factor (22) for the concentration of glipizide, probably due to Galati WWTP, followed by the Dambovita river which had an increase factor of 5.7 in the case of gliclazide, generated probably by Glina (Bucharest) WWTP (Figure 4). The next degree of potential impact corresponded to GUA (3.4) and MET (3.1) in the Ialomita river probably due to the effluent discharged by the Targoviste WWTP. In the case of Somes river, for the gliclazide, an increase factor of 3 it was obtained, probably due to the discharge of the Cluj-Napoca WWTP effluent in the Ialomita river in the downstream area. The potential impact of the 5 WWTPs was strong for four rivers (Bahlui, Dambovita, Ialomita, Somes), but the most pronounced was the WWTP Iasi effluent on Bahlui river, followed by WWTP Bucharest for Dambovita river. The antidiabetic concentration increased, after the effluent discharge, with factors of 22 from <LOQ to 3.5 ng/L for glipizide (Bahlui), 5.7 from 3.1 to 20.9ng/L for gliclazide (Dambovita) and 4.9 from 2.4 to 14.2ng/L for metformin in Bahlui.
These values are similar or lower than the concentrations reported by other paper in Germany rivers (Lake Constance MET 102 ng/L, Rhine river MET 100ng/l), or in USA rivers (MET 150ng/L) [10,14,18]. The most contaminated river was represented by Ialomita, which had a total concentration of antidiabetics of 112.1 ng/L in the downstream point, followed by the Siret and Dambovita rivers, which had a total concentration of antidiabetics of 66.3 ng/L and 57.3 ng/L, respectively, also in the downstream points.

Conclusions
A SPE-LC/MS/MS method was validated in order to quantify 5 antidiabetic compounds (metformin, glimepiride, glyburide, gliclazide, glipizide) and one degradation (guanyl ure) in surface water samples. The limit of quantification (LOQ) ranged in the interval of 0.1-2.45 ng/L. The recovery rates obtained for spiked samples were between 57.4 and 105.2%, proving that the method is accurate. The linear regressions (1-100ng/L) used to calibrate the LC-MS/MS system had determination coefficients higher than 0.99. The method was precise having good intra-day precision (3.5-7.2%) and inter-day precision (6.5-12.7%). The most contaminated river was represented by Ialomita, which had a total concentration of antidiabetics of 112.1 ng/L in the downstream point, followed by the Siret and Dambovita rivers, which had a total concentration of antidiabetics of 66.3 ng/L and 57.3 ng/L, respectively, also in the downstream points.