Endocrine disrupting chemicals in Italian drinking water systems: Insights from a three-year investigation combining chemical and effect-based tools

Abstract

Drinking water quality can be compromised by endocrine-disrupting chemicals (EDCs). Three phenolic compounds [bisphenol A (BPA), nonylphenol (NP), and 4-octylphenol (OP)] and three hormones [17β-estradiol (E2), estrone (E1), and 17α-ethinylestradiol (EE2)] were analyzed as EDCs potentially occurring in source and drinking water from three full-scale drinking water treatment plants (DWTPs) in the Romagna area (Italy) by a combined approach of HPLC-MS/MS target analysis and effect-based tests for estrogenicity and genotoxicity. The EDC removal efficiency was evaluated at different steps along the treatment process in the most advanced DWTP. NP prevailed in all samples, followed by BPA. Sporadic contamination by OP and E1/E2 appeared only in the source waters; EE2 was never detected. No estrogenic or genotoxic activity was found, except for two samples showing estrogenicity well below the effect-based trigger value suggested for drinking water safety (0.9 ng/L EEQ). BPA and NP levels were largely below the threshold value; however, increases were observed after the intermediate steps of the treatment chain. The good quality of the water relied on the last step, i.e. the activated carbon filtration. DWTPs may represent an extra source of EDCs and monitoring chemical occurrence at all steps of the process is advisable to improve efficiency.

Keywords: Drinking water; Effect-based methods; Endocrine-disrupting chemicals (EDCs); HPLC/MS-MS; Water treatment efficiency.

Graphical abstract

Fig. 1

Targeted EDCs in the present study.

Fig. 2

Geographical position and schematic representation of treatment processes at the three drinking water treatment plants (DWTPs) investigated: NIP; Capaccio, CAP; and Standiana, STN. CER: Canale Emiliano-Romagnolo. The sampling points are indicated with dots and triangles. For each sampling point chemical or biological analyses further applied are indicated in the boxes at the bottom.

Fig. 3

BPA and NP removal efficiency of the three DWTPs investigated (NIP; Capaccio, CAP; and Standiana, STN) during the summer campaigns (July and September 2020, 2021, and 2022) (n = 6).

Fig. 4

BPA and NP removal efficiency evaluated after PRECL (pre-chlorination), flocculation (FLOCC), ultrafiltration (ULTR), and after biological activated carbon (BAC) filtration treatment steps of Standiana (STN) DWTP. [A] Removal efficiency during the summer campaigns (July and September 2020, 2021, and 2022) (n = 6). [B] Removal efficiency during the winter campaigns (January, February and/or March 2020, 2021, and 2022) (n = 5).

Fig. 5

[A] Estrogenic activity observed at REF 20 after exposure of MCF-7 cells to drinking water samples from three DWTPs (NIP; Standiana, STN; Capaccio, CAP) collected during six campaigns in 2020, 2021 and 2022. Data are expressed as the mean of proliferative effect (PE) ± SE of different experiments, each conducted in quadruplicate (n = 4). Activity is normalized to negative control, which was set at 1.0 (dotted line). [B] Estrogenic activity in drinking water samples from NIP and STN DWTPs in July 2021 (violet bars) compared in the presence of 10⁻⁷ M Tamoxifen (TMX; red bars). NC: negative control (1.0 ± 0.06). Data are expressed as the mean of PE ± SE of different experiments, each conducted in quadruplicate (n = 3). *P < 0.05 vs negative control.

Fig. 6

Genotoxic response of drinking water samples from three DWTPs (NIP; Standiana, STN; and Capaccio, CAP) collected during summer campaigns in 2020, 2021, and 2022. Data are expressed as the mean of micronuclei (n°/1000 binucleated cells) ± SD (n = 3). Genotoxicity is compared to negative control (0.1 % methanol), which was set at 11.4 ± 2.8 (dotted line). Positive control using Mitomycin C showed a MN frequency of 29.4 ± 3.3.