Chemical and toxicological assessment of leachates from UV-degraded plastic materials using in-vitro bioassays

Plastic polymers, artificial weathering and solid phase extraction

Selected plastic polymers , artificial weathering and solid phase extraction (SPE) were identical as published by Klein et al. (2021). Table 1 shows the selected polymer types for this study: polypropylene (PP), polyethylene terephthalate (PET), polystyrene (PS), low-density polyethylene (LDPE) and polyvinylchloride (PVC). In order to also consider supposedly more sustainable products, recyclates (R) of PET, LDPE and PVC, polybutylene succinate (Bio-PBS) as bio-based and a starch blend (SB) as biodegradable plastics were analyzed additionally. The SB is a mixture of petroleum-based (polybutylene adipate terephthalate—PBAT) and bio-based material (thermoplastic starch, glycerin and polylactic acid—PLA) as described in Klein et al. (2021) already. For each plastic category two materials, obtained as pellets with a diameter <5 mm were selected, except for the two PVC products. These PVC samples (transparent plate and sheet pile) were cut into pellet-sized pieces to increase the surface area and thus to achieve similar leaching conditions. The leachable surface area was measured for an aliquot of 100 pellets per sample to the nearest 0.001 mm on each dimension under a stereo microscope (Olympus SZ40) and extrapolated to the tested mass (Klein et al., 2021).

LDPE-R is a post-industrial recyclate from the packaging industry. It is mechanically recycled plastics and contains additives in contrast to the additive-free LDPE.

Artificial weathering and SPE were performed in duplicates by project partners from Goethe University Frankfurt (Germany). The starch blend (SB) was processed later at IWW in the same way (Wegener, 2020). Pellet number/100 g and surface area (cm²/L⁻¹) of the decomposed plastic material are given in Table S1. Also, pH and conductivity of the water samples after the treatment processes (T1 = dark control, T2 = UV-C, T3 = UV-A/B, T4 = UVA/B_aq.) were analyzed and are presented in Table S2.

In short and relevant for the present study: for artificial weathering 100 g of each plastic polymer was treated under four different conditions. The simulated scenarios included ‘atmospheric’ as well as ‘aquatic’ weathering. For ‘atmospheric’ weathering conditions, the test materials were irradiated under laboratory conditions for 24 h by either UV-C at 250 nm wavelength (treatment 2, T2) or UV-A/B at 280–400 nm wavelength (treatment 3, T3). To represent an ‘aquatic’ exposure scenario and to cover volatile compounds eventually lost under T2 and T3, test materials were leached for 24 h in 1.1 L of ultrapure water during UV-A/B irradiation at 280–400 nm (T4). A corresponding dark control (T1) was treated as T4 without irradiation. Further weathering conditions e.g., temperature, UV-intensities, lamp characterization and calculation of the simulated irradiance time during weathering are described in Klein et al. (2021).

SPE was performed with 1 L leachates using Telos C18/ENV cartridges (Kinesis GmbH, Wertheim, Germany) to retain hydrophobic compounds (Abbas et al., 2019). Loaded and dried cartridges from one duplicate were sent to IWW (Mülheim, Germany). After elution using acetone and methanol (both, VWR International GmbH, Hannover, Germany), the extracts were resolved in dimethyl sulfoxide (DMSO, VWR International GmbH, Hannover, Germany) to an enrichment factor of 5,000. Extracts were stored at −20 ° C until testing. Previously, 200 µL of the corresponding eluate were stored at −20 °C for chemical analysis at the Federal Institute of Hydrology (Koblenz, Germany).

Cell culture

p53- and ERα-transfected human osteosarcoma cell line U2-OS (BioDetection Systems BV, Amsterdam, Netherlands) were cultured in Dulbecco’s modified eagle medium with nutrient mixture F-12 (DMEM/F12; Thermo Fisher Scientific, Waltham, MA, USA) and phenol red (c.c.pro GmbH, Oberdorla, Germany). The medium was supplemented with 8.0% fetal calf serum (FCS), 0.6% Gentamicin and 1.0% non-essential amino acids (MEM-NEAA, all from c.c.pro GmbH). Both cell types were cultured at conditions of 37 °C, 97% humidity and 5% CO₂ atmosphere (CO₂-Incubator INC 153, Memmert, Nussloch, Germany).

For CALUX-assays, cells were seeded in DMEM low glucose (Sigma-Aldrich, Germany) medium (free of phenol red supplemented with 4.8% stripped FCS (BDS), 1.6% gentamicin (c.c.pro), and 1.0% MEM-NEAA (c.c.pro), at a density of 1 × 10⁴ cells in 100 µL medium per well of 96-well-plates. ERα-cells were incubated under the above conditions for 24 h and p53-cells for 48 h to allow adherence and confluent growth before exposure to the respective sample.

Cytotoxicity

The MTT assay is used to determine the metabolic activity of eukaryotic cells and is based on the reduction of the soluble, yellow tetrazolium salt 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) to a blue-purple, insoluble formazan by dehydrogenases localized in the mitochondria of the cells. After cell lysis, formazan can be quantified photometrically. The amount of formazan is proportional to the number of metabolically active eukaryotic cells.

In detail, cells were exposed to different concentrations of the samples for 24 h. 10% DMSO served as a positive control. Subsequently, cells were washed and stained with MTT-solution (5 mg MTT/ml in PBS, Merck KGaA, Darmstadt, Germany). After 2 h of incubation, cell lysis (lysis solution: 99.4 ml DMSO, 0.6 ml acetic acid, 10 g sodium dodecyl sulphate, Merck KGaA) was performed and absorbance at 595 nm was measured using the Spark 10M plate reader (Tecan Deutschland GmbH, Crailsheim, Germany).

The MTT-assay was applied as pre-test in order to avoid false negative results in the CALUX-assays using eukaryotic cell lines. Only the concentration of the respective polymer extract or substance classified as non-cytotoxic with the MTT assay (cell viability >70% according to ISO standard (ISO 10993-5, 2009)) was applied in the CALUX assays to determine estrogenic (ERα) and genotoxic (p53) activity (Mosmann, 1983).

Estrogenicity

The ERα-CALUX-assay, used for the detection of estrogenic effects was performed according to ISO 19040-3 (2018). The assay is used to detect receptor-mediated estrogenic activities. “CALUX” stands for chemical activated luciferase gene expression. The human osteosarcoma cell line U2-OS (BioDetection Systems BV, Amsterdam, NL) is genetically modified. The reporter gene luciferase has been coupled to the estrogen receptor gene (ERα). In case an estrogen-active substance binds to the ERα receptor in the cell, the corresponding gene is read and consequently also the reporter gene. Thereupon luciferase is produced, which reduces the added luciferin to oxyluciferin under the emission of light. The luminescence intensity correlates directly with the amount of substance bound to the receptor. To determine estrogenicity equivalent concentrations, a 17β-estradiol (E2) standard series (0–91 ng E2 equivalents per litre, EEQ/L) is included in each assay, which allows the expression of results as ng EEQ/L (BioDetection Systems, 2014; Besselink, 2015).

24 h after seeding in assay medium, adherend and confluent cells were exposed to the respective samples for another 24 h. This was followed by 15 min cell lysis (lysis-buffer, 500 ml cell culture water, 3 g Tris base, 2 mL DTT, 0.73 g CDTA, 100 mL Glycerol, 10 mL Triton X-100; Sigma-Aldrich) and subsequent luciferin (Glow Mix, BDS) addition. Immediately afterwards, luminescence was measured using a Spark 10M plate reader.

A statistical analysis using One-way ANOVA und Tukey’s post-hoc test (Prism; GraphPad, San Diego, CA, USA) was performed in order to proof if estrogenic effects measured in the plastic leachates were statistically different from the process control.

Genotoxicity

Chemical analysis

The eluates of Solid Phase Extraxtion (SPE) were measured via HPLC-HR-MS as described previously in detail by Klein et al. (2021). Analysis was performed using a TripleTOF 6600 (SCIEX) coupled with an ESI source to a binary HPLC instrument (1260 Infinity, Agilent) equipped with a reversed phase C18 column (Zorbax Eclipse Plus, 2.1 mm ×150 mm, 3.5 µm, Agilent). As eluent a water-acetonitril gradient was used, acidified with 0.1vol% formic acid at a flow of 300 µL min⁻¹. The method used both ESI(+) and ESI(-) ionization mode (scan mode 100–1200 Da). Samples were diluted 1:1 with ultrapure water and injection volume was 50 µL (Klein et al., 2021). The data analysis was processed according to a non-target approach as described by Jewell et al. (2020) and Köppe et al. (2020). After data acquisition, peak picking, componentization, alignment of the components and blank correction, the components and their intensities were summed for each sample and ionization mode. The data are summarized in Table S3 of Supplementary Information. The total count and intensity of all detected components were normalized according to the highest and lowest values and displayed in the heat map (see Fig. 1).

Cytotoxicity

In order to determine cytotoxic effects of the different plastic leachates on eukaryotic cells and to determine the non-cytotoxic concentrations to perform the CALUX assays on estrogenicity and genotoxicity, the MTT assay was performed using ERα- and p53-cells. Test concentrations in the CALUX-assays were chosen based on the results of the MTT-assay.

Two different leachate concentrations were tested in the MTT assay with ERα-cells: 178 and 445 g/L P-EQ (starch blend pellets (SB) 98 and 196 g/L P-EQ). At the highest concentration tested, cytotoxic effects (<70% cell viability) were measurable in some plastic leachates (see Fig. S1). At a concentration of 178 g/L P-EQ (SB: 196 g/L P-EQ) leachates from all treatments (T1–T4) did not show cytotoxic effects. Therefore, the ERα-CALUX was performed with leachate concentrations of 178 g/L P-EQ (SB: 196 g/L P-EQ), in order to be able to compare the results of the different plastic polymers as well as the different treatments.

In the MTT-assay with p53-cells, three different concentrations of the leachates were tested: 445, 890 and 1780 g/L P-EQ (SB: 392, 979 and 1958 g/L P-EQ). The highest two concentrations showed cytotoxic effects in most of the leachates (see Fig. S2). At the lowest tested concentration leachates from all treatments (T1–T4) did not show cytotoxic effects. Therefore, the p53-CALUX was performed with leachate concentrations of 445 g/L P-EQ (SB 1958 g/L P-EQ) in order to be able to compare the results of the different plastic polymers as well as the different treatments. Using the same P-EQ concentrations allowed a ranking of the investigated polymers with respect to the potential to release estrogenic and genotoxic compounds as a consequence of weathering.

Estrogenicity

The ERα-CALUX was performed with exposure concentrations of the plastic leachates of 178 g/L P-EQ (196 g/L P-EQ for SB). As the procedural blanks (pb) of the SPE using Telos C18/ENV cartridges (Kinesis GmbH) showed slight estrogenic effects (0.29 ng EEQ/L), this background signal was subtracted from results of the leachates, except for SB (no background signal was found here).

The results of the different treatments are shown in Table 2. All values in bold exceed the recommended limit value of 0.4 ng EEQ/L for surface waters (Kienle et al., 2015). Compared to the procedural blank significantly different estrogenic effects of 0.63, 0.58 and 0.94 ng EEQ/L were detected for the PVC-A leachate after UV-C treatment (T2), the PVC-R leachate of the dark control (T1) and the SB leachate after aqueous UV-A/B exposure, respectively. For PVC-A leachate after UV-A/B treatment (T3), an estrogenicity level of 0.44 ng EEQ/L was detected. However, the difference to the procedural blank was not statistically significant. In general, EEQ concentrations were higher in leachates of plastic polymers treated with UV-radiation (except for PVC-R). For SB, estrogenic effects were solely found after the “aquatic” weathering scenario (T4). Estrogenic effects of the different leachate samples at different treatment procedures (T1–T4) are shown in Figure S3.

Genotoxicity

The p53-CALUX was performed with exposure concentrations of the plastic leachates of 445 g/L P-EQ (1958 g/L P-EQ for SB). Plastic leachates were judged to be genotoxic if the induction factor (IF) was greater than or equal than 1.5 in at least two of three test replicates and on average (Van der Linden et al., 2014). A genotoxic potential could be determined in leachates of three plastic species and solely in assays performed without metabolization of substances (Table 3). PS-HI was genotoxic after UV-C treatment (T2, IF = 1.55), LDPE-R after “aquatic” weathering with UV-A/B (T4, IF = 1.52) and SB after all weathering scenarios, except UV-C treatment (T1, IF = 1.50; T3, IF = 1.61; T4, IF = 1.89). Hence for SB the threefold concentration (1958 g/L P-EQ) compared to other leachates was tested. The remaining leachates were not genotoxic as the mean IF was either <1.5 or 2 of 3 replicates were negative. The comparison of test results with and without S9 (metabolic activation) is shown in Figure S4. Significant effects were always detectable without S9 supplementation.

As second genotoxicity assay, the Umu-test was applied. Here, plastic leachates were tested in 4 concentrations ranging from 0.67 to 5.34 g P-EQ/mL. (SB 0.37 to 2.94 g/mL P-EQ). Samples were judged to be clearly genotoxic if the induction rate was greater than 1.5 for both test replicates. The results, indicating the lowest genotoxic concentration with or without S9, are presented in Table 4. All obtained results with and without S9 can be found in Newe (2019) and Wegener (2020) and in Figure S5.

The LDPE-R leachates were genotoxic after every type of weathering scenario and after “aquatic” weathering with UV-A/B (T4) already at the lowest tested concentration of 0.67 g/mL P-EQ. The leachates of the weathered SB also exhibited genotoxic potential after T1, T2, and T4. A genotoxic effect after metabolization of substances present in the leachates occurred only for LDPE-R in the dark control (T2).

Comparison of the toxicological potential of different plastic species after weathering

For comparison of the biological activity of the different investigated plastic leachates, a heat map was created based on the bioassay results of this study (Fig. 1). The heat map shows clearly that leachates of LDPE-R have the highest genotoxic potential (Umu-assay) after T4 treatment. PVC-A and PVC-R show the highest estrogenic effects at T1 and T2 treatment, respectively. SB is genotoxic as shown in the p53-CALUX without S9-Mix (T1: IF = 1.50; T3: IF = 1.61; T4: IF = 1.89) as well as in the Umu-assay without S9 (T1, T3, T4). Leachates of PS-HI were also genotoxic in p53-CALUX without S9 (T2 UV-C, IF = 1.55) (see Fig. 1 and Fig. S4/S5).

Chemical analysis

As already discussed in Klein et al. (2021), LC-QTOF results can be used to get an overview about the complexity of the compounds present in the leachates and for a rough estimation of the released amounts. Informative value of intensities and comparability of different plastic samples is very limited, due to high differences in detector response for different molecules in LC-MS analysis. However, the data revealed obvious differences in leachates from the plastic samples and UV induced changes in the leaching behavior. Fig. 1 illustrates counts and intensities of components found in the plastic eluates.

LDPE-R, both biodegradable plastics and both PVC samples released the highest number of compounds (between 1012 (PVC-A, T1) and 4873 (SB, T4), see Tab. S3). PP-H released the fewest number of compounds (between 37 (T2) and 78 (T4); Tab. S3). A clear increase of counts caused by UV irradiation could be observed for LDPE-R and PVC-A. For Bio-PBS and SB only a small effect and for the other plastics no effect could be observed.