Determination of cytotoxicity following oxidative treatment of pharmaceutical residues in wastewater

effectiveness of a wastewater treatment plant (WWTP). In certain cases, discharged metabolites are transformed back into their pristine structure and become bioactive again. Other compounds are persistent and can withstand conventional wastewater treatment. When WWTP effluents are released in surface waters, pristine and persistent chemicals can affect the aquatic environment. To complement WWTPs and circumvent incomplete removal of unwanted chemicals or pharmaceuticals, on-site wastewater treatment can contribute to their removal. Advanced oxidation processes (AOPs) are very powerful techniques for the abatement of pharmaceuticals, however, under certain circumstances reactive toxic by-products can be produced. We studied the application of on-site AOPs in a laboratory setting. It is expected that treatment at the contamination source can eliminate the worst polluters. Thermal plasma and UV/H 2 O 2 oxidation were applied on simulation matrices, Milli-Q and synthetic sewage water spiked with 10 different pharmaceuticals in a range of 0.1 up to 2400 μ g/L. In addition, untreated end-of-pipe hospital effluent was also subjected to oxidative treatment. The matrices were activated for 180 min and added to cultured HeLa cells. The cells were 24 h and 48 h exposed at 37 ◦ C and subsequently markers for oxidative stress and viability were measured. During the UV/H 2 O 2 treatment periods no toxicity was observed. After thermal plasma activation of Milli-Q water (150 and 180 min) toxicity was observed. Direct application of

• Thermal plasma and UV-C/H 2 O 2 treatment were applied in contaminated matrices.• Cytotoxicity from oxidative treated matrices was evaluated using an oxidative stress probe and viability marker.With the presence of multiple chemicals in sewage waters, contaminants may adversely affect the effectiveness of a wastewater treatment plant (WWTP).In certain cases, discharged metabolites are transformed back into their pristine structure and become bioactive again.Other compounds are persistent and can withstand conventional wastewater treatment.When WWTP effluents are released in surface waters, pristine and persistent chemicals can affect the aquatic environment.To complement WWTPs and circumvent incomplete removal of unwanted chemicals or pharmaceuticals, on-site wastewater treatment can contribute to their removal.Advanced oxidation processes (AOPs) are very powerful techniques for the abatement of pharmaceuticals, however, under certain circumstances reactive toxic by-products can be produced.We studied the application of on-site AOPs in a laboratory setting.It is expected that treatment at the contamination source can eliminate the worst polluters.Thermal plasma and UV/H 2 O 2 oxidation were applied on simulation matrices, Milli-Q and synthetic sewage water spiked with 10 different pharmaceuticals in a range of 0.1 up to 2400 μg/L.In addition, untreated end-ofpipe hospital effluent was also subjected to oxidative treatment.The matrices were activated for 180 min and added to cultured HeLa cells.The cells were 24 h and 48 h exposed at 37 • C and subsequently markers for oxidative stress and viability were measured.During the UV/H 2 O 2 treatment periods no toxicity was observed.
After thermal plasma activation of Milli-Q water (150 and 180 min) toxicity was observed.Direct application of thermal plasma treatment in hospital sewage water caused elimination of toxic substances.The low cytotoxicity of treated pharmaceutical residues is likely to become negligible if plasma pre-treated on-site wastewater is further diluted with other sewage water streams, before reaching the WWTP.Our study suggests that AOPs may be promising technologies to remove a substantial portion of pharmaceutical components by degradation at the source.Further studies will have to be performed to verify the feasibility of upscaling this technology from the benchtop to practice.

Introduction
Many different anthropogenic pollutants are discarded via industrial, domestic and hospital sewage water for removal in wastewater treatment plants (WWTPs).At WWTPs multiple steps are applied to reduce the content of suspended solids and overall organic carbon by 80-90% (Peake et al., 2016).Due to the complexity of sewage water and the presence of multiple chemicals such as personal care products, industrial compounds, pesticides, detergents and pharmaceuticals, individual pollutants or mixtures may adversely affect the WWTP removal efficiency causing inhibitory or toxic effects by affecting the bioactivity of activated sludge (Carucci et al., 2006;Henriques et al., 2007;Vasiliadou et al., 2018).This leads to incomplete WWTP removal or enzymatic conversion of metabolites into the original bioactive structure.Emissions of insufficiently treated WWTP may have adverse effects in the aquatic environment (Ajo et al., 2018;Abbas et al., 2019;Moermond et al., 2020;van Bergen et al., 2021).The occurrence of pharmaceutical residues in the environment is a known risk for aquatic ecosystems, antimicrobial resistance and drinking water quality (Isidori et al., 2016;Peake et al., 2016;Vasiliadou et al., 2018).Despite the difficulty to reveal the complete ecotoxicological effect of pharmaceuticals in water, the potential risk for water quality of various compounds is well demonstrated by bioassays or in silico tools (Fent et al., 2006;De Garcia et al., 2014;Abbas et al., 2019).For the removal of residual pharmaceutical compounds, different techniques could be applied either at WWTPs or at drinking water production facilities by treatment with membrane filtration, reversed osmosis, powdered activated carbon in activated sludge (PACAS) or advanced oxidation processes (AOPs) (Beier et al., 2010;Streicher et al., 2016;Banaschik et al., 2018;Ajo et al., 2018).Where certain techniques rely on the removal by transferring pollutants to another phase or an adsorbent, AOPs in contrast, are based on oxidation reactions with reactive species for the degradation of micropollutants (Magureanu et al., 2018;Ma et al., 2019).For example, with UV-C irradiation of hydrogen peroxide (H 2 O 2 ), hydroxyl radicals ( • OH) are formed which react with organic molecules.Another AOP is plasma technology, where electric energy is discharged into gas phase over a liquid containing residual pharmaceuticals.The energetic electrons effectively induce excitation, dissociation and ionization of the gas discharge medium, yielding amongst others nitrogen and oxygen-based excited states, radical species and ions.Primarily produced reactive nitrogen (RNS) and reactive oxygen species (ROS) will react with pathogens and organic molecules by disinfecting and degrading pollutants in wastewater (Shen et al., 2016;Gerrity et al., 2010;Back et al., 2018;Thirumdas et al., 2018).It is demonstrated that AOPs including plasma oxidation are promising techniques for on-site wastewater treatment prior to conventional WWTP (Ajo et al., 2018;Graumans et al., 2021).Despite the rapid degradation of organic molecules and disinfection properties, certain AOPs may also produce harmful oxidation by-products, which can be temporal upon reaction progress and degree of mineralization (Parkinson et al., 2001;Heringa et al., 2011;Hofman-Caris et al., 2015;Jaeger et al., 2015Dong et al., 2017;Du et al., 2017;Sillanpaa et al., 2018;Gilca et al., 2020).Formation of toxic oxidative intermediates is especially of concern for drinking water production (Parkinson et al., 2001;Heringa et al., 2011;Hofman-Caris et al., 2015).Photocatalytic removal of natural organic matter (NOM), using techniques like UV/H 2 O 2 and UV/O 3 in combination with medium pressure lamps (MP) can convert micropollutants in even more harmful components by the reaction between NOM, nitrate and photolytically produced radicals (Parkinson et al., 2001;Heringa et al., 2011;Hofman-Caris et al., 2015;Gilca et al., 2020).Next to that, in presence of bromide, the AOP ozonation (O 3 ) can generate bromated toxic by-products (Wu et al., 2019;Yang et al., 2019).Additional filtration steps with activated carbon are needed for the removal of these toxic intermediates formed by oxidation (Heringa et al., 2011;Back et al., 2018).To circumvent and minimize the formation of oxidative by-products, it is suggested to apply AOPs early in the wastewater chain.By reducing pharmaceutical residues and pathogens at the contamination source, conventional wastewater treatment plants are relieved and complemented.To investigate the suitability of plasma technology as a pretreatment technique for hospital sewage water, we tested oxidatively treated matrices on HeLa-cells.The cellular system was directly exposed to the oxidatively treated matrices and overall toxicity was evaluated.HeLa cells are uterus tumour cells of human origin, which can proliferate rapidly, making them a suitable cell line for cytotoxicity assays (Symons et al., 2001).

Cell culture
HeLa cells (American Type Culture Collection, CCL-2) were maintained at 37 • C in a humidified atmosphere of 5% (v/v) CO 2 in DMEM medium containing 25 mM glucose, GlutaMAX, 1 mM pyruvate supplemented with 10% (v/v) FCS.Medium was renewed 2-3 times a week and cells were passed (1:10) every week after trypsinisation with 0.05% trypsin-EDTA in serum free medium when cells reached approximately 90% confluence.

Simulation matrices and hospital sewage water
To evaluate toxicity before and after UV-C/H 2 O 2 and thermal plasma induced oxidation, Milli-Q water and synthetic sewage water matrices were spiked with a mixture of 10 pharmaceuticals IOM, DIA, CIP, FLU, DF, MET, CP, CB, APAP and MF, see table S2.Pharmaceuticals were individually spiked from freshly prepared 1 mg/mL stock solution.From the 1 mg/mL stock solution consecutive concentrations of 24000,100,150,20,40,20,40,20,850,370 μg/L were prepared in 500 mL Milli-Q or SSW.The prepared amounts in the simulation matrices were a tenfold higher than measured in raw hospital sewage water, this to simulate the detected concentration (1.0-2400 μg/L) in the cell system (Graumans et al., 2021).

Synthetic sewage
Synthetic sewage water (SSW) was freshly prepared prior to each experiment.The composition of SSW was prepared according a standardized protocol from the Organization for Economic Co-operation and Development (OECD). 100 mL of freshly tapped Milli-Q was used to prepare all analytes separately.The final composition of SSW consists of the salts: 28.0 mg K 2 HPO 4 , 7.0 mg NaCl, 3.0 mg CaCl 2 ⋅2H 2 O and 0.5 mg MgSO 4 including organic constituents 160.0 mg Peptone, 110.0 mg Meat Extract, 30.0 mg Urea dissolved in 1000 mL Milli-Q water (OECD, 2001).

Hospital sewage water
End-of-pipe hospital sewage water (HSW) was used during plasma and UV-C/H 2 O 2 treatment.HSW was derived from a composite sample, which was periodically sampled between Monday -Thursday 16-19 September 2019 (Graumans et al., 2021).The collected HSW was directly filtered upon arrival in the laboratory using vacuum filtration with a Whatman Filter grade 1 (Maidstone, United Kingdom).Different aliquots were prepared and stored in the freezer (− 20 • C) till analysis.Prior to oxidative treatment, 500 mL frozen HSW samples were thawed and subsequently poured in a glass beaker.

Oxidative treatment and cellular exposure
500 mL of simulation matrices were activated in triplicate (n = 3) and exposed to HeLa cells after treatment using plasma and UV-C/H 2 O 2 treatment.Both AOPs, were also used for the duplicate (n = 2) treatment of 500 mL of HSW samples, prior to cellular incubation.Cells were seeded in 96-wells plates (BD Falcon 96 well imaging plate black/clear, Fisher Scientific) (24 h) prior to exposure.During HeLa cell exposure the untreated and treated matrices were added in triplicate (n = 3 wells), dissolved in a 1:10 concentration ratio (20 μL sample: 180 μL medium (v/v)) serum-free DMEM medium.For each experiment three controls were taken in the form of a medium control, vehicle control and a cytotoxic positive control using 500 μM H 2 O 2 .The cell plates were incubated for 24 h and 48 h at 37 • C in a humidified atmosphere of 5% (v/v) CO 2 .

Thermal plasma treatment
A laboratory-scale plasma activation unit from VitalFluid (Eindhoven, the Netherlands) was used for plasma discharge in the matrices.The plasma unit comprises a 150 Watt (W), 1 MHz dual resonant power modulator (Pemen et al., 2017;Hoeben et al., 2019).The test samples were placed in an 800 mL glass beaker and thermal plasma discharges were applied with maximum power output of 150 Watt (W) in ambient air over water.During this process electrical discharges with typically 8 kV peak-to-peak voltage were generated.To maintain the temperature of the treated solution below 35 • C during 180 min plasma activation, a Peltier cooling was applied.

UV-C/H 2 O 2
For the UV-C irradiation a modified Aetaire UV air disinfection unit was used.UV-C irradiation is 180 min stationary applied through an opening of 28 cm × 7 cm.UV-C is irradiated at 253.7 nm (Philips PL-L 60W/4P HO UV-C lamp, Eindhoven, the Netherlands) with a distance of 20 cm between the lamp and aqueous matrix surface.At the start of the experiment 0.22 mM (~10 mg/L) hydrogen peroxide (H 2 O 2 ) from J.T. Baker, Avantor Performance Materials (Deventer, the Netherlands) is added to the matrix.Minimizing reflection and UV scattering, a container with an aperture (12 cm, wide) was placed over the 800 mL glass beaker.

Oxidative stress and viability assay 2.5.1. Oxidative stress probe CM-H 2 DCFDA
After 24 h and 48 h incubation cells were washed with 200 μL PBS.Next, cells were stained with 5 μM CM-H 2 DCFDA probe diluted in serum-free DMEM.Cells were incubated for 30 min at 37 • C before the access medium containing the probe was aspirated and were washed with PBS.To induce fluorescent signal, 100 μL of 500 μM H 2 O 2 was added to the culture medium.Cells were incubated for 5 min at 37 • C and the fluorescence was measured at 488/520 nm using a microplate spectrophotometer from Molecular Devices, Multimode Microplate Reader Spectramax iD5 (San Jose, USA).

Cell viability marker crystal violet
After cellular exposure, medium was removed and washed with PBS.
Subsequently, 50 μL 0.5% (v/v) crystal violet solution was added to the cells and placed for 20 min at room temperature.The crystal violet marker will stain both proteins and nuclei, but, when cells undergo apoptosis, a lower amount of crystal violet staining will occur since death cells are detached and thus not visualized (Feoktistova et al., 2016).After incubation at room temperature, the staining solution is removed and cells were washed thoroughly 5 times with PBS.Plates were air-dried for 4 h before the addition of 100 μL methanol, which will dissolve the crystal violet marker.Crystal violet absorbance is measured at 570 nm using a microplate spectrophotometer from Molecular Devices, Multimode Microplate Reader Spectramax iD5 (San Jose, USA).

Data analysis
Fluorescence and viability levels were plotted according to eq. (1) by using GraphPad Prism 5.03 and 9.3.1 statistical software (GraphPad Inc., CA, USA).The amount of fluorescence produced compared to the medium control, was used as a measure for oxidative stress.Fluorescence levels above >100.0%do indicate oxidative stress levels deviating from to the medium control.Eq. ( 1) is also applicable to calculate cell viability levels.Apparent cell viability (%) were determined by taking the crystal violet absorbance of a sample divided by the absorbance of the medium control multiplied by 100%.Cell viability levels lower than <50% were classified as severe cytotoxic.

Oxidative stress (%) =
Fluorescence (sample) Fluorescence (Medium control) • 100% (1) For each untreated (0 min) or oxidatively treated sample (5-180 min), three independent wells were 24 h and 48 h exposed to the HeLa cells.The mean (x) value was calculated for each treatment period, including the standard deviation (σ) and the standard error of the mean (SEM).To confirm a significant difference in cell viability levels after an oxidation experiment, one-way analysis of variance (ANOVA) was applied.For the determination of statistically significant effects per treated time-period compared to the 100% medium control, the T-statistic was calculated, using Bonferroni's Multiple comparison test in GraphPad Prism 5.03 and 9.3.1 statistical software.Statistically significant effects are labelled with an asterisk (*): *; p < 0.05, **; p < 0.01 M.H.F.Graumans et al. and ***; p < 0.001 (Miller and Miller, 2010).

Assay optimization
The effectiveness and sensitivity of the HeLa cytotoxicity assay was determined by using a fixed volume of 20 μL hydrogen peroxide (H 2 O 2 ) at different concentrations.H 2 O 2 is a known cytotoxic compound, since it is a source of hydroxyl radicals ( • OH) when exposed to UV irradiation.
• OH radicals in a cell system will initiate lipid peroxidation and DNA hydroxylation (Day and Suzuki, 2006;Ayuso et al., 2020).The inactivation applicability of aqueous H 2 O 2 was demonstrated in vitro on human and murine tumour cell lines.To observe cytotoxic effects after 24 h H 2 O 2 exposure, CM-H 2 DCFDA fluorescent probe and crystal violet were used as markers for oxidative stress and cell viability, respectively.Ascending H 2 O 2 concentrations cause oxidative stress and reduce viability levels significantly, see Fig. 1 and Table S1.As demonstrated in this figure H 2 O 2 concentrations higher than >100 μM do cause a decline in viability.Elevated fluorescence signal, compared to apparent viability of control, is a measure for cellular oxidative stress.
Increased fluorescence levels relative to the medium control, taken as 100%, indicate cellular oxidative stress (expressed as % relative to the medium control).When CM-H 2 DCFDA probe is oxidized by intracellular reactive oxygen species (ROS), a fluorescent adduct is produced within the living cell (Fig S1).In case of apoptosis the probe will leak out and is washed away during the colouring process.Crystal violet staining is a useful marker for the verification of viability (Sliwka et al., 2016;Feoktistova et al., 2016).Compared to enzymatic colorants the crystal violet indicator tags proteins and the external surface of the DNA double helix.By measuring the absorbance of the coloured product, viable adherent cells are quantified.With a plotted dose-response curve, the half maximal effect concentration (EC 50 ) was calculated for the different H 2 O 2 concentrations Fig S2 .EC 50 values of 99 μM and 16 μM were found with CM-H 2 DCFDA and crystal violet indicators, respectively.The difference between the EC 50 values is explained by the different cellular targets of both markers (Feoktistova et al., 2016;Ayuso et al., 2020).The observed fluorescence signal, even in a case of cellular decline, might overestimate the viability when CM-H 2 DCFDA probe is used alone.It was demonstrated in intestinal (IPEC-J2) cells that cellular efflux of the CM-H 2 DCFDA probe can occur.Other interfering factors were attributed to the washing buffer, polystyrene well plate or from non-oxidized probe (Ayuso et al., 2020).To minimize autofluorescence we also measured empty polystyrene wells.Despite differences between the EC 50 values for CM-H 2 DCFDA and crystal violet, 500 μM H 2 O 2 can be used as a positive control for confirmation of detected cell death.With 500 μM H 2 O 2 viability levels less than <25% for both CM-H 2 DCFDA (14 ± 25%) and crystal violet (21 ± 6%) markers were found.

Plasma treatment of spiked Milli-Q water
Xenobiotics such as pharmaceuticals can generate reactive oxygen species (ROS) in living organisms during metabolism (Lutterbeck et al., 2015).Insufficient inactivation of ROS can cause damage to cellular macromolecules and too high concentrations of endogenously produced ROS (H 2 O 2 and O 2 − ) might initiate apoptosis (Ayuso et al., 2020).To simulate a contaminated water matrix, 10 different pharmaceuticals were spiked in Milli-Q water (see supplemental information paragraph S2).Milli-Q water is a clean matrix that can be used to show the possible effects of produced oxidative degradation intermediates as a reference without the influence of other constituents (Banaschik et al., 2018;Graumans et al., 2021).During a previous degradation study we spiked Milli-Q with an initial concentration of 5 μg/L for the tested pharmaceuticals.After 120 min plasma activation, different conversion levels were found for the parent compounds, having abatement levels for IOM (67%), DIA (41%), CIP (99%), FLU (83%), DF (100%), MET (72%), CP (100%), CB (91%), APAP (100%) and MF (18%) (Graumans et al., 2021).Milli-Q used as vehicle control did not show any significant abnormalities compared to the medium control, see Fig. 2A.No decline in viability was observed for the untreated Milli-Q sample containing pharmaceuticals after 24 h exposure.Expected is that the initial concentrations are too low to cause acute toxicity or cell death.Only moderately cellular oxidative stress is observed, mainly for the plasma activated periods 5 and 90-180 min.In contrast, after 48 h exposure, a clear elevation in fluorescence signal was detected for the treatment periods 5 and 15 min in combination with a severe decrease in viability levels for treatment duration 90, 120 and 180 min (significantly different p < 0.05 from the medium control).Here also the untreated matrix (0 min) resulted to some extent in increased fluorescence (~156%), indicating that prolonged exposure to a pharmaceutical mixture influences cellular homeostasis, see Fig. 2B.During thermal plasma treatment, reactive nitrogen (RNS) and oxygen species (ROS) are produced.These reactive species cause pharmaceutical degradation.In our previous study, we have demonstrated in Milli-Q water that 100% degradation for DF and APAP occurs within min and complete CP elimination in 30 min (Graumans et al., 2021).The elevated fluorescence levels observed for plasma treatment periods of and 15 min indicated that rapidly produced degradation intermediates may cause cellular oxidative stress.According to other oxidation studies it is well explained that during oxidative treatment multiple transformation intermediates are formed (Lutterbeck et al., 2015;Banaschik et al., 2018;Graumans et al., 2020).Lutterbeck demonstrated for the CP alone 5 different intermediates after UV/H 2 O 2 and UV/TiO 2 treatment.In a similar study with non-thermal plasma treatment a four-step degradation pathway was anticipated for DF with the possibility for the formation of 21 different intermediates (Banaschik et al., 2018).During a qualitative pharmaceutical degradation study we have also demonstrated multiple reaction intermediates for the compounds CB, DF and APAP (see supplemental information paragraph S3).Identification of these degradation intermediates could have provided more information regarding their toxicity.However, based on the high levels of oxidative stress, seen in the cell system after 48 h incubation, it is most likely that intermediates generated in Milli-Q water cause this unwanted cellular effect.Next to increased oxidative stress levels, a decrease in cell viability was seen.Statistically significant differences among the medium control for treatment periods 90, 120 and 180 min were observed  S4.Substantial decrease in viability levels were primarily demonstrated for the prolonged plasma treatment periods longer than 90 min.A decrease in viability may not be solely attributed to the formed degradation intermediates, but could have also been caused by the reactivity of plasma activated water itself (Hoeben et al., 2019;Graumans et al., 2021).Prolonged plasma activation results in an acidic environment (Table S17).Under less neutral circumstances (pH < 6) nitration reactions may complement the hydroxyl radical interaction during plasma oxidation (Magureanu et al., 2018;Hoeben et al., 2019;Graumans et al., 2020Graumans et al., 2021).Although RNS are useful for chemical degradation, incorporation of nitrogen containing compounds into aromatic structures can result in mutagenic by-products (Hofman-Caris et al., 2015).On the other hand, reactive species cannot react exhaustively in Milli-Q water, due to the absence of larger organic molecules or buffering minerals (Giannakis et al., 2018).The methanol fraction (0.64 mM) derived from all diluted stock standards was 1.3% (v/v) in 500 mL test matrix.It is expected that the methanol content is also rapidly broken down by oxidation products.This effect was demonstrated by Popov et al. (2010), who showed a gradual decline of 43 mM CH 3 OH using UV/H 2 O 2 oxidation.As reported by Hoeben et al. (2019), prolonged plasma activation will increase the concentration of reactive species such as H 2 O 2 , HNO 2 , HNO 3 and transient reactive nitrogen species.They used a similar thermal plasma generator to produce 4900 μM (NO 2 − ), 1800 μM (NO 3 − ) and 370 μM (H 2 O 2 ) in deionised water during 120 Watt plasma activation for 25 min (Hoeben et al., 2019).
Compared to our EC 50 values it was found that <100 μM H 2 O 2 is already a sufficient concentration to induce cytotoxicity resulting in substantial viability loss.Because of the continuous addition of multiple reactive oxygen and nitrogen species, it is explicable why HeLa cell viability decreases after 48 h exposure.

Plasma treatment synthetic sewage water
Synthetic sewage water (SSW) is a suitable surrogate matrix for studying pharmaceutical degradation in actual wastewater (OECD, 2001;Ajo et al., 2018).In our previous study, 5 μg/L was selected as initial pharmaceutical concentration in SSW.The conversion levels by thermal plasma activation were; IOM (56%), DIA (15%), CIP (70%), FLU (83%), DF (99%), MET (65%), CP (100%), CB (79%), APAP (100%) and MF (73%) (Graumans et al., 2021).To detect possible cytotoxicity after thermal plasma activation in a more realistic environment than Milli-Q water, a surrogate for sewage water was prepared.Comparing synthetic sewage water cytotoxicity results with plasma-treated Milli-Q water, a direct difference in oxidative stress and viability was detected, see Fig. 3. Measured elevated levels in viability >120% indicate that cell growth occurred.Increased viability levels were also seen for the vehicle control samples, representing that SSW matrix might have contributed beneficially to cell medium quality.Certain constituents from SSW matrix, such as mineral salts, can maintain osmotic balance and organic residues, such as meat extract and peptone, act as building blocks (Arora, 2013).Regarding the oxidative stress levels, a less pronounced increase in fluorescence was seen after 24 h and 48 h cellular incubation.Lower variance between the datapoints may be explained by a less hostile reactive environment in synthetic sewage matrix compared to Milli-Q water.It is expected that with the presence of minerals and organic constituents, plasma-produced reactive species are inhibited or transformed into less reactive radicals (Giannakis et al., 2018;Graumans et al., 2021).Using the viability confirmation assay by staining biomass, it is seen that the majority of the exposed cells do not die or detach in the cell system, see Fig. 3A.For the 5-120 min plasma-treatment periods no statistically significant differences (p < 0.05) were observed, only a significant reduction in viability was found.The gradual decline in viability was 35, 46% and 34 and 42% for prolonged treatment periods of 150 and 180 min, respectively, following 24 h and 48 h of cellular exposure (see Table S8 and S9 in the supplemental information).For the treatment periods of 5-120 min, it is expected that minerals and organic constituents in SSW do have sufficient buffer capacity to inhibit RNS and ROS.On the other hand, in AOP systems where • OH radicals are continuously generated, organic matter can act as an additional source for oxidative species by forming transient organic matter radical species (DOM • ) (Westerhoff et al., 2007;Dong and Rosario-Ortiz, 2012;Peake et al., 2016).
Therefore, after plasma degradation it is expected that in the absence of high concentrations of pharmaceutical residues, organic matter will be an additional source for reactive species.As for Milli-Q water, the gradual decrease in viability after prolonged plasma activation is attributed to additionally produced reactive species, affecting the ambient HeLa cell environment (Hoeben et al., 2019;Trebinska-Stryjewska et al., 2020).The very minimal cytotoxicity observed in our laboratory set-up is likely to become negligible when plasma pre-treated on-site wastewater is further diluted with other sewage water streams before reaching conventional sewage treatment.By using AOPs in healthcare facilities, the pharmaceutical residues can be degraded when concentrations are still high, eliminating a substantial portion of components by degradation at the source.

Toxicity of hospital sewage water
To understand how thermal plasma oxidation performs in a real sewage water sample, untreated and oxidatively treated hospital wastewater (HSW) was exposed to cultured HeLa cells for 24 h and 48 h.In our a previous degradation study 10 pharmaceuticals were detected in a concentration range of 0.1-2400 μg/L.Direct application of 120 min plasma activation resulted in the abatement for IOM (33%), DIA (24%), CIP (70%), FLU (12%), DF (94%), MET (43%), CP (100%), CB (36%), APAP (100%) and MF (78%) (Graumans et al., 2021).The HSW is a much more complex matrix than spiked Milli-Q and synthetic sewage water.This may explain the viability drop of ~65% after 24 h incubation (see Fig. 4A).By application of 150 Watt thermal plasma, this effect is, however, rapidly intervened after 30 min treatment, since similar viability levels were produced for the medium control.A completely different response was observed after 48 h of cellular exposure, in which no decline was found in viability for untreated HSW.The observed crystal violet signal (111% ± 47) in combination with low fluorescence levels indicated that organic contaminants reacted with crystal violet.
That the viability level of the untreated HSW sample increased from 35% (24 h) to 111.0% (48 h) after prolonged cellular incubation, indicates the growth of bacteria.From other studies it is known that many different strains of bacteria co-exist in sewage water (Chahal et al., 2016).Based on the very low CM-H 2 DCFDA fluorescence signal at 0 min (17% ± 3), which is almost equal to that of the H 2 O 2 control (16% ± 4), it is suggested that crystal violet colorant visualized bacterial biomass.Crystal violet is a known colorant for Gram-positive bacteria (Davies et al., 1983).According to the cellular interaction of CM-H 2 DCFDA probe, it is further endorsed that primarily Gram-positive bacteria are visualized in untreated HSW, since certain prokaryotic cells do not produce glutathione (Smirnova and Oktyabrsky, 2005;Fahey, 2013).CM-H 2 DCFDA probe binds to the thiol-group of glutathione, in case of oxidative stress reactive oxygen species do produce a fluorescent adduct.Abatement of viability levels for timepoints 5, 15 and 30 min further endorse the suggestion that bacteria are degraded by disinfectant properties of plasma oxidation (Thirumdas et al., 2018).Comparison of the HSW data with the simulation matrices, makes it difficult to determine the exact contaminants that cause direct toxicity of the untreated HSW samples.It is expected that the cocktail of untreated HSW contains pharmaceutical residues, metabolites, detergents, hormones, pathogens and bacteria who all might contribute to viability reduction (Ajo et al., 2018;Abbas et al., 2019).That cell viability levels decrease after the addition of untreated wastewater is in line with other studies (Zegura  2009); Jaeger et al., 2015).During our experimental set up we added 20 μL matrix in 180 μL medium (10% (v/v)) untreated and oxidatively treated matrix to HeLa cells.In the study of Zegura et al. (2009), human liver cells (HepG2) were exposed for 20 h to different volumes of 0, 5, 10, 20, 30 and 40% industrial effluent, hospital effluent, WWTP effluent and river water.In their study cell viability reduction levels of 30% were considered as cytotoxic.Different volumes of industrial effluent and river water showed a dose-dependent decrease in cell viability.For the same 10% volume, only river water indicated a significant reduction of viable HepG2 cells.The direct toxicity we found for 10% v/v untreated HSW might be explained by the water composition and that the concentration of pharmaceutical residues is highest at the contamination source.In the study of Zegura et al. (2009), no direct cytotoxicity of HepG2 cells exposed to 40% v/v for 20 h incubation was seen, but 2 out of their 5 HSW samples were found to be genotoxic.Incubation of untreated HSW for 20 h did not immediately result in a major decrease of HepG2 cells, probably because of filtration through a 0.22 μm filter, prior to cellular exposure (Zegura et al., 2009).In our preliminary HSW experiment, we also applied a filtration step using Costar® Spin-X® centrifuge filtration tubes (0.22 μm).During this exploratory experiment only oxidative stress levels of ~177% ± 46 for filtered HSW were determined compared to 11% ± 7 for unfiltered HSW, see supplementary paragraph S5.Based on this result we suggest that filtration removes bacterial toxins.To demonstrate the most realistic effect of plasma treatment on HSW, we decided to not use a filtration step prior to cellular incubation.To determine potential toxicity of real wastewater samples it must be noted that too much sample preparation can lead to miss interpretation (Abbas et al., 2019).Despite that the HSW matrix can cause difficulty in visualising the viability levels of HeLa cells using crystal violet, it is clear from the overall results that plasma treatment will have an overall beneficial effect on wastewater quality.Plasma oxidative treatment of 45 min will result in a better cellular environment, since HeLa cell viability loss is reversed.

Cytotoxicity measurements before and after UV-C/H 2 O 2 treatment
The use of UV-C/H 2 O 2 is also an effective technology for the degradation of pharmaceutical residues.Photolysis of a hydrogen peroxide solution will produce reactive hydroxyl radicals ( • OH), which interact with nearly all organic compounds.To investigate whether UV-C/H 2 O 2 treatment might generate harmful oxidative degradation products, cytotoxicity of HeLa cells was measured for untreated and treated Milli-Q, SSW and HSW, respectively (Fig. 5A-F).During an earlier oxidative degradation study, pharmaceuticals (5 μg/L) were spiked in Milli-Q and SSW.It was found that the parent compounds were rapidly converted between 9.6 and 100% using 120 min UV-C/H 2 O 2 treatment.The detected pharmaceutical concentrations in HSW showed conversion levels of 4 up to 100% (Graumans et al., 2021).No direct decrease in viability levels (>50%) was observed compared to the 500 μM H 2 O 2 positive control condition.More specifically, for the oxidative treatment of spiked Milli-Q water, minimal effect on the HeLa cells was found after 24 h exposure.Elevated fluorescence levels were only seen after 48 h cellular incubation, indicating that during prolonged UV-C/H 2 O 2 oxidative treatment intermediates are formed which initiate oxidative stress levels.Since Milli-Q is a clean matrix, it is plausible that degradation intermediates were derived from the initial spiked pharmaceuticals (paragraph S3).As compared to plasma oxidation, UV-C/H 2 O 2 treatment resulted in a less hostile cellular environment.A minimal decrease in cell viability levels were found after UV-C/H 2 O 2 oxidation, which might be caused by the addition of 10 mg/L (220 μM) H 2 O 2 at the start of the treatment period.In contrast to the continuously generated RNS and ROS species with plasma oxidation, we expect that the photolysis reaction between UV-C and H 2 O 2 is more exhaustive in nature.With UV-C irradiation the highest emitted energy photons (254 nm, 4.9 eV) cannot photochemically split atmospheric oxygen, nor water, thus ozone and direct hydroxyl radical formation do not occur.Next to that, no RNS such as HNO 2 , HNO 3 and ONOOH are introduced in the matrix during UV-C/H 2 O 2 treatment, leading to a less acidic and reactive water environment (Table S19).On the other hand, photolysis of the residual pharmaceuticals by UV-C photons is likely, including the rapid chemical degradation by • OH radicals derived from the single added 10 mg/L (220 μM) H 2 O 2 solution (Graumans et al., 2021).
Important to mention is that in SSW and HSW, UV-C irradiation is absorbed by organic constituents, and that the formed • OH radicals are likely to be quenched or inhibited by minerals such as phosphate or chlorine (Giannakis et al., 2018;Graumans et al., 2021).Despite the inhibitory effects of organic constituents during oxidative degradation, other studies found that application of UV irradiation in drinking water with high natural organic matter (NOM) content can generate mutagenic degradation by-products (Parkinson et al., 2001;Heringa et al., 2011;Hofman-Caris et al., 2015).
With the application of AOPs at the wastewater source, a different NOM content is expected compared to surface or ground water used for drinking water preparation.From our oxidatively treated SSW and HSW samples, minimal direct cytotoxic effects were found, since a decrease in viability levels was not detected, see Fig. 5C-F.This can be partly explained by the use of a low-pressure UV lamp at 254 nm in contrast to a medium pressure UV lamp, which was applied in the studies of Heringa et al. (2011) andHofman-Caris et al. (2015).Medium pressure lamps emit multiple UV wavelengths between 200 and 300 nm, causing changes in the molecular NOM structure (Heringa et al., 2011).In the presence of NO 3 , MP UV irradiation can produce mutagenic by-products, between photolytic nitrate products and NOM (Hofman-Caris., 2015).Furthermore, by alteration of the NOM structure, bound heavy metals could be released, increasing the toxicity (Parkinson et al., 2001).Next to the UV-C source, also the concentration and composition of NOM is of importance since this mixture of constituents will affect the chemical degradation efficiency greatly.Westerhoff et al. (2007) found that their reference NOM from Suwannee River fulvic acid water (US) had much higher UV 254 nm absorbance than their WWTP reference sample.Based on this observation, it is expected that during our UV-C/H 2 O 2 experiment lower photolysis of NOM in HSW occurred.This supports our observations of the rapid degradation of UV-C absorbing x-ray contrast agents IOM and DIA (Graumans et al., 2021).NOM material with unsaturated double bonds (C = C) are prone to absorb UV within the range of 250-280 nm.It is expected for our oxidatively treated matrices SSW and HSW that the NOM content had a different composition, which is less prone to direct change in molecular structure after 254 nm irradiation.The minimal cytotoxic effects in HeLa cells exposed to UV-C/H 2 O 2 -treated SSW and HSW suggests that this could be a useful on-site sewage water treatment technique.

Strengths and limitations of the cytotoxicity experimentation
One of the strengths of this experimental study is that the cytotoxicity levels were measured for multiple oxidatively treated wastewater matrices and hospital sewage water, see Fig. 6.Cytotoxicity was determined before and after thermal plasma and UV-C/H 2 O 2 treatment, demonstrating the usefulness of advanced oxidation at a contamination source.Our laboratory-scale setup may serve as an indicator for further experimental design for the application of advanced oxidation technology.On the other hand, it is important to mention that a limitation of our work is that we have obtained cytotoxicity results based on a mixture of pharmaceuticals in a wide concentration range.It is therefore difficult to establish whether an individual compound or a mixture effect has caused the harmful cellular effects.On top of that, with the application of thermal plasma activation and UV-C/H 2 O 2 treatment, distinct chemical degradation kinetics occur, producing different reactive species.For translation of the cytotoxic results, an additional mutagenic or more specific ecotoxicity assay could elucidate or underpin specific observations.The usage of crystal violet for assessing cell viability is not the most appropriate control for our complex HSW matrix, but, in combination with the CM-H 2 DCFDA probe informative results on oxidative stress and viability were obtained.With the current experimental setup we have demonstrated the possibilities of advanced oxidation of end-of-pipe contaminated water matrices, the performance of which could have been explained even better with the oxidative degradation of multiple source oriented sewage water compositions.This laboratory-scale experiment is therefore the preferred step to develop a pilot-testing system for the treatment of specialized hospital wastewater streams.

Conclusion
From the results it is clear that the initial concentration of pharmaceuticals, the composition of the matrix, oxidative treatment duration, the amount of formed oxidative species and the ability of the cellular system to recover, are all determinants of residual toxicity.Our experimental setup showed that both thermal plasma and UV-C/H 2 O 2 are capable of substantially degrading pharmaceuticals in hospital sewage water.Our results may therefore contribute to the adoption of both oxidative technologies for the pre-treatment of wastewater with a high load of pharmaceuticals at the source.Plasma activation provides a continuous production of ROS and RNS, which may directly affect cell viability in Milli-Q water if not diluted or inhibited.Further studies are required, including bioassays, which inform the potential impact on the aqueous ecosystem.Direct application of AOPs at the contamination source is considered a promising technology to reduce pharmaceuticals in sewage water at the source.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgement
The MEDUWA project and this work were supported by the INTER-REG Deutschland-Nederland program [Grant number 142118].We would like thank VitalFluid for their help in providing the thermal plasma unit and UV-C source.The authors also would like to thank Charlotte Hoogstraten and Tim Somers for their technical advice and

•
Distinct cellular effects were found in simulation matrices oxidatively treated either by thermal plasma or UV-C/H 2 O 2 .• Application of oxidative treatment at the contamination source can reduce cytotoxicity.• Laboratory scale oxidative treatment is the first step in validation of plasma deployment in wastewater treatment.released in the aquatic environment due to incomplete removal from wastewater.

Fig. 1 .
Fig. 1.Different concentrations of 20 μL H 2 O 2 were added to cultivated HeLa cells, compared to a medium control and positive control (500 μM H 2 O 2 ).CM-H 2 DCFDA fluorescence probe was used to indicate oxidative stress ( ).Crystal violet staining was used to confirm cell viability levels ( ) ***; p < 0.001 (t-test with Bonferroni correction).(For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 2 .
Fig. 2. Milli-Q water matrix spiked with pharmaceutical residues was 0-180 min plasma treated and subsequently added to cultured HeLa cells for 24 h (A) and 48 h (B).CM-H 2 DCFDA ( ) fluorescent tag is used to demonstrate oxidative stress, the crystal violet colorant ( ) is applied for the viability confirmation.The black line (− ) is used as reference to indicate the unaffected HeLa cell condition (100%, medium control).The red line ( ) demonstrates 50% cell viability.**; p < 0.01 and ***; p < 0.001 (t-test with Bonferroni correction).(For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3 .
Fig. 3. Synthetic sewage water matrix containing 10 pharmaceutical residues was plasma treated and subsequently incubated for 24 h (A) and 48 h (B).CM-H 2 DCFDA ( ) fluorescent tag is used to demonstrate oxidative stress and the crystal violet colorant ( ) is applied for confirming viability.Unaffected HeLa cell condition (100%, medium control) is represented by (− ).The red line ( ) demonstrates the 50%, viability.**; p < 0.01 and ***; p < 0.001 (t-test with Bonferroni correction).(For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 5 .
Fig. 5. UV-C/H 2 O 2 oxidative treatment applied to different spiked matrices.Oxidative stress was measured with CM-H 2 DCFDA ( ) and cell viability is confirmed with crystal violet colorant ( ).The black line (− ) is the 100% reference and the red line ( ) demonstrates 50% viability.The oxidative treatment effects of Milli-Q water were measured for 24 h (A) and 48 h (B) on HeLa cells.These effects were also examined after treatment of synthetic sewage 24 h (C) and 48 h (D) and 24 h (E) and 48 h (F) incubated hospital sewage water.***; p < 0.001 (t-test with Bonferroni correction).(For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 6 .
Fig. 6.Combined cytotoxicity results obtained in different oxidatively treated matrices Milli-Q, SSW, HSW.A heatmap visualization is used to compare the oxidative stress results (A and C) and viability levels (B and D).