Cyto- and genotoxic profile of groundwater used as drinking water supply before and after disinfection

A directive of the European Union (98/83/EC) regulates the quality of drinking water, and defines the microbiological and physical–chemical parameters to allow its use for humans. The promulgation of the directive became very useful, as both groundwater and surface water have been polluted in recent decades, mainly because of human activities.

The quality of surface water has been analyzed in several studies (Alves et al. 2014; Mendoza et al. 2014; Wilbers et al. 2014), while groundwater has received less attention as it has been considered pollution-free. However, several recent studies have reported that contaminants, including pharmacologically active compounds (e.g. pharmaceutical drugs, hormonal substances, and drugs of abuse), industrial compounds, various products used for personal hygiene, and pesticides, have been detected in groundwater (Petrovic et al. 2003; Radjenovic et al. 2007; Dieter 2010; Martínez Bueno et al. 2012). These contaminants and their degradation products can be very persistent, thereby becoming a risk for human health (Postigo et al. 2008; Soares et al. 2008; Reungoat et al. 2010). In particular, pollutants with genotoxic properties may lead to serious adverse effects, including cancer, even when present at low concentrations (Jurado et al. 2012; Lapworth et al. 2012; Pal et al. 2014). The contemporary use of several complementary in vitro tests able to identify different types of genotoxic damage is a widespread monitoring approach to assess the presence of genotoxins in water environments (Lah et al. 2005; Marabini et al. 2007). Among the available tests, the most used are the Ames test, which allows the identification of mutagens with a different mechanism of action (Maron & Ames 1983), the Alkaline version of the Comet assay that detects primary DNA damage (Olive & Banath 2006) and the micronucleus assay, which mainly measures clastogenic and aneugenic effects (Fenech et al. 1999). The assessment of raw water (RW) is interesting to predict the type and quality of tap water. In fact, several studies have shown that the formation of disinfection by-products (DBPs) is connected to the chemical and physical properties of RW (Fulladosa et al. 2004; Roy & Bickerton 2011; Huang et al. 2014). In particular, Teksoy et al. (2008) have shown a relationship between the formation of DBPs and the pH, the temperature and the concentration of organic matter, and of bromine dissolved in the water supply. Several studies have shown how contaminants in groundwater (e.g. pentavalent arsenic and bromine) can affect the disinfection process, resulting in the formation of substances that are cytotoxic, genotoxic and teratogenic (Hunter et al. 1996; Plewa et al. 2002; Richardson et al. 2008; Richardson 2009; Huang et al. 2014). Though the levels of these compounds in water are low, and the exposure of humans through water consumption is rather limited, the severity of the health risk may be high because such exposure is maintained throughout life (Carraro et al. 2000; Hrudey 2009). Several studies have indeed reported a link between the consumption of chlorinated drinking water and cancer of the urinary, gastrointestinal, and nervous systems, as well as reproductive disorders (Ijsselmuiden et al. 1992; Koivusalo et al. 1994, 1995; Nieuwenhuijsen et al. 2000; Woo et al. 2002; Villanueva et al. 2004). For this reason, the Environmental Protection Agency of the United States (US EPA) recommends a level of zero in drinking water for compounds that are known or probable carcinogens.

The goal of this study was to identify a methodological approach to better evaluate the cyto- and genotoxic effects induced by groundwater. In addition to the compounds found through chemical and physical analysis, many unknown compounds are present in groundwater. Known and unknown toxic substances become the components of complex environmental mixtures that may induce side effects on human health. In fact, though the European directive (98/83/EC) establishes maximum levels for contaminants in RW (introducing threshold levels for acute exposure to toxic compounds), this does not take into account chronic low dose exposure to pollutants, generally present in complex mixtures. Consequently, water pollution may be a serious problem for public health and the aquatic ecosystem.

Reagents for the biological assays and general laboratory chemicals were from Sigma-Aldrich (St Louis, MO, USA). C18 cartridges (Sep-Pak Plus tC18 Environmental Cartridges) were from Waters Chromatography (Milford, MA, USA).

Sampling of groundwater was performed at two water treatment plants from the same aquifer system of the same river basin in Northern Italy, which together serve 127,000 inhabitants. At these plants, the only treatment performed on waters is disinfection. Chlorine dioxide is used as a disinfectant for the first plant (Plant 1), while sodium hypochlorite is used in the second one (Plant 2). Fifty liters of water was sampled in spring and autumn of the same year in each plant before (RW) and after disinfection (disinfected water (DW)). According to the US EPA 525.2 method (US EPA 1994), the water samples were immediately taken to the laboratory, acidified with hydrochloric acid (pH 2–2.5), and filtered through tri-functional C18 cartridges. Cartridge activation was performed with 40 mL of the following solvents: ethyl acetate, dichloromethane and methanol (Guzzella et al. 2006). The water was sucked through the cartridge (20 L/cartridge, at a constant flow of about 20 mL/min) with a pump in a multi-sample concentration system (VAC ELUT SPS 24, Varian, Leini, Italy). The elution of each cartridge was made using ethyl acetate, dichloromethane and methanol (40 mL/each). The eluates were then reduced to a small volume by means of a rotating vacuum evaporator, dried under nitrogen flow, and re-dissolved in dimethyl sulfoxide (DMSO) at an extract concentration equivalent to 25 L of water per 1 mL of DMSO (Pellacani et al. 2006). RW and DW (∼1 L) were also sampled for chemical analysis.

Physical and chemical analyses were performed at the treatment plants, using the analytical methods for the determination of physical/chemical parameters in water intended for human consumption according to the Italian Legislative Decree 31/2001 (transposition of European Directive 98/83/EC) and its integrations (Ottaviani & Bonadonna 2007).

The mutagenic effect of water samples was studied using S. typhimurium (strains TA 98 and TA 100) with or without microsomal activation (S9 liver extract of rats treated with Aroclor 1260) to detect direct and indirect mutagenic compounds (Maron & Ames 1983). Positive controls were 2-aminofluorene (5 μg/plate) and hycanthone (25 μg/plate), while DMSO (80 μL/plate) was tested as a negative control. The results of the Salmonella assay are reported as the mean of three replicates with their relative standard deviation. The results of the assay were considered positive if two consecutive dose levels or the highest non-toxic dose level produced a response at least twice that of the solvent control and at least two of these consecutive doses showed a dose-response relationship. For each sample tested, three concentrations of extract corresponding to 0.5, 1.0 and 2.0 L of water and for each concentration three independent replicates were performed (three repeated measurements of the sample that represent independent measures of the error associated with equipment). The revertants were counted after 48 hours of incubation at 37 °C. Cytotoxicity was detected by a clearing or diminution of the background lawn.

The micronucleus assay was performed using blood samples from healthy donors who provided written, informed consent for use of their blood sample in the study. Whole blood cultures were set up adding 0.5 mL (about 1× 0⁶ lymphocytes/5 mL) of blood to 4.5 mL of culture medium containing RPMI 1640, 15% heat-inactivated fetal calf serum, 1% phytohemoagglutinin, 1 mM L-glutamine, 100 IU penicillin, and 100 mg/mL streptomycin. The cultures were incubated at 37 °C, 5% CO₂ for 72 h. Treatments were carried out at 24 h of incubation by adding water extracts dissolved in DMSO (0.15, 0.30, 0.60 and 1.2 Leq/5 mL; Leq = liters equivalent). Mitomycin C (0.3 μM) and vinblastine (40 ng/mL) were used as positive controls, and DMSO (12 μ/mL) as negative control. Each treatment was performed in duplicate. Cytochalasin-B (6 μg/mL) was added to all culture tubes at 44 h. The cultures were harvested at 72 h and treated with a hypotonic solution (KCl 0.075 M) for 2–3 minutes at room temperature; the cells were then fixed twice in ice-cold acetic acid:methanol (10:1) solution. Pre-coded slides were cytocentrifuged, air dried and stained with May Grunwald-Giemsa. Micronuclei were scored using a light microscope at 400× magnification (Surrallés et al. 1995). In each slide, 1,000 binucleated lymphocytes were examined to discover the presence of micronuclei, and the micronuclei frequency was then calculated for each treatment. Cell-cycle parameters were evaluated by classifying the 500 cells according to the number of nuclei. The cytokinesis block proliferation index (CBPI) was calculated using the formula: CBPI = [M₁ + 2M₂ + 3(M₃ + M₄)]/N, where M₁ through M₄ indicate the number of cells with 1–4 nuclei, and N indicates the total number of cells scored (Eastmond & Tucker 1989; OECD 2014). The cytotoxicity of compounds is assessed as a decrease in CBPI as a consequence of suppressed cell proliferation (Surrallés et al. 1995). Necrotic and apoptotic cells (discriminated and recognized by morphological criteria) were scored according to the guidelines of Fenech et al. (1999).

For this assay, a whole blood sample from healthy volunteers (who provided written, informed consent for its use in the study) was centrifuged twice in an erylyse solution (155 mM NH₄Cl, 5 mM KHCO₃, 0.005 mM Na₂EDTA, pH 7.4) for leukocyte isolation, and was then washed and resuspended at 10⁶ cell/mL in phosphate buffered saline. The cell suspension was added to a microcentrifuge tube together with water extract for 1 h at 37 °C, to produce the following concentrations: 0.25, 0.50, 0.75, 1.00, and 1.25 Leq/mL (Pellacani et al. 2006). DMSO (50 μL/mL) was used as a negative control, while ethyl methanesulfonate (2 mM) served as a positive control. The Comet assay was performed with cell viability >70% (Tice et al. 2000), and cell viability was determined using the fluorescein diacetate/ethidium bromide assay (Merk & Speit 1999). Following this passage, the Comet assay was performed as described by Singh et al. (1988), with minor modifications as described by Buschini et al. (2001). Alkaline DNA unwinding (pH > 13) was carried out for 20 minutes, and electrophoresis was conducted for 20 minutes at 0.78 V cm⁻¹ and 300 mA. After staining with 100 μL ethidium bromide (10 μL/mL), the slides were examined with a fluorescence microscope (Leica DMLS, 400×) equipped with an excitation filter, BP 515–560 nm, and a barrier filter, LP 580 nm, using an image analysis system (Cometa Release^® 2.1 Sarin, Florence, Italy). Fifty randomly-selected cells per slide (two slides per sample) were analyzed. The samples were coded and evaluated blind. The Comet parameter total migration length (TL) was used to measure DNA damage, and cells with complete disintegration of the head region (ghost cells (GC)/hedgehog) were also recorded to evaluate the necrosis/apoptosis phenomena.

An SPSS 11 package (SPSS Inc., Chicago, IL, USA) was used for statistical analysis. The median values of DNA migration (TL) of each experiment (Comet assay) were used for analysis of variance (ANOVA) (Dunnett's post hoc test). The statistical analysis of micronuclei frequency was performed using the χ²-test. CBPI data were analyzed by ANOVA (Tukey post hoc test).

All chemical and physical parameters of RW from both plants had results within the limits required by the Italian legislative decree 31/2001 (D.Lgs 31/2001 and subsequent amendments and additions) transposing the European Directive 98/83/EC (Table 1). Nevertheless, some molecules of toxicological interest were present above the detection limit but under the maximum allowable concentration (MAC) described by the Italian law. In particular, the following contaminants were identified: chloroform, bromoform, carbon tetrachloride, 1,1,1-trichloroethane, and tetrachlorethylene (Table 1). In most cases, levels were similar in the two plants at both sampling periods. The concentration of total trihalomethanes (THMs) in the RW collected in June and October presents similar values in both plants. THMs were also measured in water in both plants after disinfection (Table 1). As shown in Table 2, disinfection of water with chlorine dioxide (Plant 1) did not modify the levels of THMs; in contrast, in Plant 2, disinfection of water with sodium hypochlorite caused a four- to five-fold increase in THM levels (Table 2).

Table 3 shows the results of the mutagenicity studies. In general, samples of raw and DW, in both plant and at both sampling points, were negative in both Salmonella strains.

Although no significant result has been achieved, it should be noted that both aqueous extracts (RW and DW) taken in Plant 1 in June are able to induce a doubling of the spontaneous frequency of revertants in strain TA98 (without metabolic activation) (Table 3). The same samples, in the presence of metabolic activation (S9), showed no mutagenic activity, highlighting detoxifying activity of S9.

In Plant 1, all water samples induced a decrease of CBPI, with the exception of the DW sample collected in June, which had no significant effects on the CBPI. None of the water samples caused any induction of micronuclei, it is conceivable that the reduced proliferation could have masked possible genotoxic effects. No effects on proliferation was induced by the RWs from Plant 2, while waters sampled immediately after disinfection with sodium hypochlorite induced anti-proliferative effects at the highest concentrations tested (Figure 1). However, the RW sampled in June (at the concentration equivalent to 1.2 L) induced breakage and /or loss of chromosomes, detected from micronuclei counts (Figure 1).

In Italy, the main source (ISTAT 2012) of drinking tap water is groundwater. Generally, this water is not subjected to purification treatments, due to its good quality resulting from filtration by permeable layers and by protection from overlying impermeable layers (De Giglio et al. 2015). However, in the last decades the quality of groundwater has decreased, principally because of the considerable exploitation of water resources that induces a progressive deterioration of their quality (Widory et al. 2004; Burkart & Stoner 2007; Huang et al. 2012). The quality of groundwater also depends on the features of the soil and its filtration ability, which may reduce the degree of water contamination (Goeppert & Goldscheider 2011; Atherholt et al. 2013). Moreover, seasonal variation and climatic factors may change the natural replenishment of aquifers, and may worsen the quantity of groundwater (Makwe & Chup 2013; Rea et al. 2014; Sun et al. 2014; Tornevi et al. 2014).

The safety of groundwater resources represents a primary issue, and complex activities are needed to preserve their integrity. Biological monitoring allows detection of toxic substances undetectable through chemical and physical analysis, and allows identification of antagonistic/synergistic effects caused by toxic substances present in complex mixtures (Pellacani et al. 2005, 2006). Environmental risks are often evaluated analyzing the single pollutants, and do not take into consideration the effect of chemical mixtures, thus underestimating the risk to human health (Backhaus & Faust 2012). A more complex approach, involving a battery of bio-toxicological tests, is an important tool to properly assess the quality of groundwater and drinking water. To evaluate the drinking water's mutagenic and/or genotoxic risks and to get information in the short term, several biological assays, such as the Ames test, the micronucleus test and Comet assay, may be used, as in the present study.

Though the RW of both plants at both sampling periods complied with D.Lgs 31/2001 (and subsequent amendments and additions), a number of contaminants were detected (carbon tetrachloride, 1,1,1-trichloroethane, tetrachlorethylene, bromoform and chloroform). Though the exposure level to any single pollutant is low, chronic consumption of the complex mixture may represent a risk factor for human health due its ability to generate synergic effects between toxic compounds.

Carbon tetrachloride is considered a possible carcinogen to humans (IARC 1999). This molecule results in toxicity to the liver and kidney (WHO 2004a). It does not appear mutagenic to Salmonella typhimurium, however it induces DNA damage and mutations in single studies with Escherichia coli (IPCS 1999). In vitro, carbon tetrachloride produces a low induction of DNA damage, unscheduled DNA synthesis, sister chromatid exchange or chromosomal aberrations (IPCS 1999). This chemical is also able to induce damage in mammalian cells during cytokinesis (McGregor & Lang 1996).

1,1,1-trichloroethane is a substance that can cause liver effects after repeated or long-term exposure (IARC 1999). Several studies evaluated genotoxic effects of 1,1,1-trichloroethane (US EPA 2005). The results in the Ames Salmonella assay are contradictory: in fact some test results were negative, while others, performed to minimize evaporation of the compound, were positive (ATSDR 2006). 1,1,1-trichloroethane induces chromosomal aberrations in Chinese hamster ovary cells in vitro and is positive in most mammalian cell transformation assays (ATSDR 2006).

Tetrachloroethylene has been found in drinking water and in groundwater worldwide (WHO 2003). It persists in water, where volatilization does not occur, but does not cause bioaccumulation in food chains (ATDSR 1993). It is able to generate developmental toxic effects (WHO 2003) and it is considered a probable human carcinogen (IARC 1995). A short-term assay demonstrated that tetrachloroethylene causes single-strand DNA breaks, but it could not induce chromosomal aberrations. Moreover, this compound showed negative effects in Salmonella typhimurium, Escherichia coli, and Saccharomyces cerevisiae, both with and without microsomal activation (WHO 1984, 2003).

Bromoform is a substance that can cause liver toxicity (IARC 1999). It is considered a mutagen (IARC 1991; GlobalTox 2002). Bromoform was positive in the Ames test with or without metabolic activation (ATSDR 2005). The National Toxicology Program (ATSDR 2005) indicates that bromoform is able to generate micronuclei formation and it is positive to SCE assay. However, bromoform presents negative results for chromosomal aberrations in mouse bone marrow (ATSDR 2005). Potter et al. (1996) found that bromoform is not able to generate DNA strand breaks in male F344 rats.

Chloroform is a probable human carcinogen (IARC 1999). Its principal removal process is volatilization (US EPA 1984). However, in groundwater chloroform is persistent, due to restricted volatilization and slow biodegradation (Environment Canada & Health Canada 2001). Chloroform presents contradictory mutagenic activity in Salmonella typhimurium and Escherichia coli bacteria assays (WHO 1984, 2003). Several studies did not indicate genotoxic potential (chromosome aberrations, sister chromatid exchange and micronuclei formation) in vitro and in vivo (IPCS 2004). However, other studies reported that chloroform is able to induce micronuclei formation and chromosome damage in rats (Robbiano et al. 1998; Fujie et al. 1990). Moreover, Morimoto & Koizumi (1983) indicate a positive SCE result in mouse bone marrow. In conclusion, chloroform is slightly mutagenic and it does not appear to be genotoxic (WHO 2004b).

Taking into account the pollutants detected in the RWs analyzed, our data seem to confirm literature evidence that stated that the Ames test and Comet assay, in general, are not able to detect the genotoxic effects of these molecules (Hartmann & Speit 1995; IPCS 1999; WHO 2003; ATSDR 2006). As an example, tetrachloroethylene induces a clear cytotoxic effect on human blood cells treated in vitro (Perocco et al. 1983; Hartmann & Speit 1995), but does not alter DNA migration in the Comet assay when tested in the absence or presence of metabolic activation (Hartmann & Speit 1995).

The main indication obtained from the three assays carried out in this study is that the prevalent effect induced by water samples was cell toxicity, as indicated by the reduction of viability, the increase of GC, and by the reduction of CBPI. Furthermore, the results show that cytotoxicity is mainly observed with RW samples rather than samples after disinfection. As reported in the literature, the presence of toxic chemicals in RW may influence the toxic properties of the same water after disinfection (Nikolaou et al. 2004; Teksoy et al. 2008; Feretti et al. 2012; Huang et al. 2014). The results indeed show that the disinfection process (in both seasons) alters the characteristics of RW, reducing its cytotoxicity and improving its quality, independent of the biocidal chemical utilized (chlorine dioxide or sodium hypochlorite). Peculiar results were detected through MN assay when the RW of Plant 2, sampled in June, was analyzed. In the absence of anti-proliferative effects, the RW induced a significant increase in micronucleated cells. After disinfection, DW induced anti-proliferative effects and no genotoxic effect. These data support the hypothesis that genotoxic data obtained may be underestimated/masked by the strong cytotoxic effect. Nevertheless, the water toxicity decrease after disinfection, detected by the reduction of leukocyte viability and GC increase when a Comet assay was performed, could be related to a possible detoxication of compounds present in the RW. It is unclear how the chlorination processes can break down the toxic load of raw groundwater, as detected in our study. Chlorine is able to transform several water micropollutants; oxidation, addition and electrophilic substitution reactions with organic compounds are possible pathways (Deborde & von Gunten 2008). Chlorine is known to interact with dissolved natural organic matter (DNOM). DBPs, such as THMs, can result from the reaction of chlorine and chlorine dioxide with natural organic molecules. In our samples, we found a mild induction of total THMs after disinfection far below the MAC, evidencing a low concentration of DNOM.

The toxicological activity detected in our study could also be related to the occurrence in Italian water resources of emerging water pollutants such as endocrine disruptors, some pesticides, some pharmaceuticals and some azo-dyes and/or their metabolites (Meffe & de Bustamante 2014). We can hypothesize that these pollutants could be detoxified by the disinfection processes. Some of these molecules, such as endocrine disrupting chemicals and pharmaceuticals and personal care products, seem to be degraded through advanced oxidation processes in water effluents (Esplugas et al. 2007).

The methodological approach used, which involves a combination of chemical and physical analysis and a battery of bio-toxicological tests (with different molecular targets), has allowed the identification of a strong cytotoxic activity of concentrates of raw groundwater that result within the limits required by law. The disinfection processes, in particular the use of hypochlorite, induced an unexpected reduction of toxicity. More extensive studies are needed to identify groundwater micropollutants, single or in mixture, which can produce the toxicological effect detected in this study and that could represent a risk for human health.

The author(s) declare no potential conflicts of interest in the research, authorship, and/or publication of this article.

This work was supported, in part, by the Spinner Consortium (DEPACQUE; protocol: 100/03), funded by the European Union.

We would like to thank Dr Silvia Cagliari for the English revision of the manuscript and Dr Mirca Lazzaretti for technical assistance.