Natural mineral particles are cytotoxic to rainbow trout gill epithelial cells in vitro

Abstract

Worldwide increases in fluvial fine sediment are a threat to aquatic animal health. Fluvial fine sediment is always a mixture of particles whose mineralogical composition differs depending on the sediment source and catchment area geology. Nonetheless, whether particle impact in aquatic organisms differs between mineral species remains to be investigated. This study applied an in vitro approach to evaluate cytotoxicity and uptake of four common fluvial mineral particles (quartz, feldspar, mica, and kaolin; concentrations: 10, 50, 250 mg L(-1)) in the rainbow trout epithelial gill cell line RTgill-W1. Cells were exposed for 24, 48, 72, and 96 h. Cytotoxicity assays for cell membrane integrity (propidium iodide assay), oxidative stress (H2DCF-DA assay), and metabolic activity (MTT assay) were applied. These assays were complemented with cell counts and transmission electron microscopy. Regardless of mineral species, particles ≤ 2 µm in diameter were taken up by the cells, suggesting that particles of all mineral species came into contact and interacted with the cells. Not all particles, however, caused strong cytotoxicity: Among all assays the tectosilicates quartz and feldspar caused sporadic maximum changes of 0.8-1.2-fold compared to controls. In contrast, cytotoxicity of the clay particles was distinctly stronger and even differed between the two particle types: mica induced concentration-dependent increases in free radicals, with consistent 1.6-1.8-fold-changes at the 250 mg L(-1) concentration, and a dilated endoplasmic reticulum. Kaolin caused concentration-dependent increases in cell membrane damage, with consistent 1.3-1.6-fold increases at the 250 mg L(-1) concentration. All effects occurred in the presence or absence of 10% fetal bovine serum. Cell numbers per se were marginally affected. Results indicate that (i.) natural mineral particles can be cytotoxic to gill epithelial cells, (ii.) their cytotoxic potential differs between mineral species, with clay particles being more cytotoxic, and (iii.) some clays might induce effects comparable to engineered nanoparticles.

Figure 1. Control cells as well as particle uptake and effects.

Pictures were taken after 72(control) or with 250 mg L⁻¹ of the respective particle. Shown are examples for A.) Control cells, with marked nucleus (Nu) and arrows pointing to rudimentary microridges. B.) Control cell detail, with nucleus (Nu), mitochondria (Mi) and a myelin body (MB). C.) Quartz particle, phagocytized. D.) Feldspar particle, phagocytized. E.) Mica particle, phagocytized. F.) Mica particle, dilated endoplasmic reticulum. G.) Kaolin particle, phagocytized. H.) Kaolin particle phagocytosis (arrows, with inset showing detail of particle marked on the upper left). I.) Kaolin particle in cytoplasm without surrounding membrane. In all insets bar denotes 0.5 µm. Note: white areas around particles represent artifacts due to sectioning.

Figure 2. Cytotoxic effects and particle interference in the applied cytotoxicity assays.

Shown are cytotoxic effects of the particles in cells (black symbols) and particle interference with the respective assay (grey symbols), all data after 72 h exposure. Symbols near the x-axis denote significant differences (p<0.05) to respective control for cytotoxicity (asterisks) and particle interference (circles) data. Labels on x-axis denote control (C, no particles), as well as low (L, 10 mg L⁻¹), medium (M, 50 mg L⁻¹) and high (H, 250 mg L⁻¹) treatment level, grouped according to mineral species. Data points are mean ± SE. Note: Shown are summary statistics calculated from raw-data, while significance was tested with linear mixed-effect models to adjust for clustering of the wells on plates.

Figure 3. Effects in cell numbers and baseline metabolic activity in control wells.

A.) Cell counts after 72 h exposure to 250 mg L⁻¹ mineral particles. Shown are exposure with (left group) and without (right group) FBS. Open symbols are controls (C, no particles), and filled symbols denote mineral particle (quartz, Q; feldspar, F; mica, M; kaolin, K). Asterisks near the x-axis denote significant (p<0.05) differences to respective control; B.) Changes in metabolic activity of controls during the experiment, shown are controls with (circles) and without (diamonds) FBS. Dashed lines are best fit lines with respective Pearson product-moment correlation coefficients (σ) given near the x-axis. In both graphs data points are mean ± SE. Note: Summary statistics shown in A.) were calculated from raw-data, while significance was tested with a generalized linear mixed-effect model to adjust for clustering of the wells on plates.