Nanotechnology can be defined as the techniques aimed at characterizing and producing materials on the nanometer scale (< 100 nm) and exhibiting specific physical/chemical properties and functions. Nanoparticles (NPs) are currently commercially manufactured and can be carbon- or metal-based materials (quantum-dots, nanogold, metal oxides), dendrimers and composites. Carbon nanoparticles will be produced in tons and consequently will lead to increased human and environmental exposure due to normal use, fugitive emissions, accidental spills and disposal of materials after use. Despite their wide application, little is known about their human health and environmental implications.

Numerous epidemiological studies have associated exposure to small particles such as combustion-generated fine particles with lung cancer, heart disease, asthma and/or increased mortality. Both Donalson et al., 12 and Oberdörster 3 concluded in reviews that ultrafine particles of low-solubility and low toxicity materials are more inflammogenic in the rat lung than larger particles of the same material. Additionally, NPs are able to penetrate deeply into the respiratory tract. Once deposited in the alveolar region, they may translocate to blood and to sites distant from their portal of entry such as the liver, spleen, kidney and brain 345678. Their migration to distant sites is an important issue with regard to their toxicity. The kidney is particularly susceptible to xenobiotics owing to its high blood supply and ability to concentrate toxins. Few studies have examined the impact of NPs in kidney, while both glomerular structures during plasma ultrafiltration and tubular epithelial cells may be exposed to NPs. Chen et al. 9 clearly observed damage to proximal tubular cells in mice exposed to copper NPs. Wang et al. 10 also observed signs of glomerulonephritis and pathological degeneration after oral titanium dioxide administration, within the renal proximal convoluted tubules. Additionally, recent bio-distribution studies confirmed NPs in kidneys and the influence of size or surface treatment on in vivo tissue distribution 111213.

In the present studies, the effects of NP exposure on renal cells and their potential toxicity were investigated. The kidney is composed of different types of cells with varying sensitivities to toxic substances. Assays were conducted on two different cell lines (mesangial cell line, IP15 and proximal epithelial tubular cell line, LLC-PK₁). These cell models were used by considering two important levels in nephrotoxicity. Mesangial cells are perivascular pericytes located within the central portion of the glomerular tuft between the capillary loops and are involved in the control of glomerular hemodynamics 14. IP15 cells represent a human stable immortalized mesangial cell line and a suitable model to study in vitro cytotoxicity 1516. LLC-PK₁ cells constitute an established cell line derived from normal pig kidney displaying several characteristics of the proximal tubule. This cell type is characterized by well-developed basal infolding and an apical brush-border, intense pinocytotic activity and variable transport or co-transport. They are also involved in intensive toxic accumulation 17.

Particle size, size distribution and dispersion in relevant biological media were first defined to assess toxicity accurately. NPs characterization in terms on the chemical properties was quantified in a Knudsen flow reactor using different probe gases that heterogeneously interact with the functional groups present on the NPs surface (Setyan, Sauvain and Rossi, personal publication submitted). Cellular morphology and mitochondrial function (WST-1 assay) for toxicity evaluations were assessed under controlled conditions. Cells were stained with phalloidin-FITC to detect of the cytoskeletal component, F-actin. To model the cellular uptake of particles, we used electron microscopy and modified fluorescent carboxylate and sulphate-modified polystyrene latex beads of various diameters. An additional study on oxidative stress was conducted using 2'–7'dichlorodihydrofluorescein diacetate (DCFH-DA) ROS assay. The mesangial and proximal tubular cell types were exposed to carbon black (CB) and to titanium dioxide (TiO₂) particles with different average diameters. The data presented here constitute the first step towards the determination of a dose-effect correlation and a risk assessment of CB and TiO₂ NPs on renal cells.

NPs have been shown to reach the systemic circulation after inhalation, ingestion or intravenous injection, with further distribution to several organs such as liver, spleen, kidneys, brain or heart 4192021. However, particle size clearly affects their distribution, as suggested by De Jong et al. 22. Indeed, after intravenous injection in the rat tail vein, systemic distribution was evidenced for gold NPs (10 nm diameter), with further detection in various organs including liver, spleen, kidney, testis, thymus, heart, lung and brain. Larger particles were also systemically distributed, but detected in the liver and spleen only. This is in accordance with recent bio-distribution experiments showing NPs in the kidneys 101123. Wang et al., observed damage of kidneys after TiO₂ single oral gavage due to the small size and particle clearance 10. Moreover, exposition to copper NPs clearly led to renal proximal tubular cell damage in mice, with swollen glomeruli reflecting glomerulonephritis, altered blood biochemical indexes, in particular noticeable changes in urea nitrogen, creatinine and alkaline phosphatase levels, all suggesting altered renal functions 9. Others studies suggested that nano-drugs could be targeted and incorporated by renal glomerular mesangial cells 2425. Meanwhile, short- or long-term impacts on renal functions after NPs accumulation in the kidneys have not been assessed yet, and, to our knowledge, few in vitro data are currently available on NPs toxicity in renal cells.

In this study, we evaluated the potential in vitro effects of different commercially available NPs (carbon black -FW2, P60 and LB101- or TiO₂ nanoparticles) on glomerular mesangial cells (IP15) or epithelial proximal tubular cells (LLC-PK₁). We have previously evaluated and validated the use of such cell lines to assess renal toxicity 162627. Therefore, developing such in vitro models to assess NPs renal toxicity would be of particular interest regarding development of routine screening tests and investigation of NPs precise mechanisms of action.

First, we analyzed NPs by electronic microscopy and turbidimetry under conditions close to toxicity experiments in the RPMI 1640- free serum medium. We showed that these particles do not necessarily retain their nano-size in solution. Indeed, carbon black as well as TiO₂ particles tended to aggregate into larger complexes, with for example an 8-, 13- or 6 fold-increased size for FW2, P60 or LB101 aggregates respectively, compared to the initial size of isolated particles.

Measuring the viability of cells, by neutral red or LDH assay after chemical exposition is a routinely made method to assess toxicity information. Another standard assay used for analysis is the measurement of the activity of the mitochondrial dehydrogenases by formation of a formazan product which is water soluble in the WST-1 assay in contrast to the formazan product generated in the MTT assay. Carbon black particles can adsorbed substances such as neutral red or MTT 282930. We confirmed this result in some of our experiments, showing interferences caused by these particles in the neutral red cytotoxicity assay. Keeping in mind these interactions, all of our microscopy characterization and cytotoxicity assays were therefore conducted in a red-phenol free medium without FBS supplementation. We assessed NPs potential cytotoxicity by use of the WST-1 assay; the WST-1 assay has been validated in various in vitro toxicity assays using such particles 3132.

In vitro cytotoxicity assays conducted in this study highlighted different responses in function of the nanoparticle tested and/or the cell type considered. On both IP15 and LLC-PK₁ cells, the smallest FW2 NP was found to be the most cytotoxic with classic dose-behavior. For the other NPs tested (LB101 and TiO₂), different cytotoxic profiles were found, with LLC-PK₁ cells being more sensitive than IP15 cells. On IP15 cells, CB LB101, P60 and TiO₂ displayed less or no toxicity at the doses tested. On LLC-PK₁ cells, the cytotoxic response of LB101 was different and size-dependent for TiO₂. Previous results showed the ability of NPs to induce various effects in function of the cell type considered. For example, Patra et al. reported a gold nanoparticle-induced death response in a human carcinoma lung cell line (A549 cells) whereas no effect was observed in BHK21 (hamster kidney) or HepG2 (human hepatocellular liver carcinoma) cells 33. Accordingly, such differences were observed in our experiments for example with TiO₂ particles, inducing cell death in LLC-PK₁ but not in IP15 cells. The degree of cytotoxicity was correlated to the particle size, in accordance with numerous studies showing that smaller TiO₂ and CB particles are more toxic than larger ones 1343536. However, apart from size, other structural characteristics may also participate in the various cytotoxic effects observed when studying different cell types. For example, Warheit et al. demonstrated that crystal structure and surface chemical reactivity may also be involved 13. In particular, chemical reactive groups present at the surface of the particles may differ from one particle type to another as recently suggested by our group (Setyan et al., personal publication, submitted), therefore leading to various cytotoxic profiles when interacting with different cell types,

An unsolved issue is whether nanoparticles are internalized into cells, and if so, which internalization mechanisms are involved. To assess this question, we first conducted experiments to evaluate whether fluorescent latex beads may be internalized in both cell types. We therefore showed that small particles (30 nm) were internalized into the cell cytoplasm, whereas larger particles (500 nm) remained outside the cells. This result therefore demonstrates that particle internalization is possible in these cells, but is size-dependent. However, the size of the latex beads internalized is consistent with the size of our NPs of interest, suggesting a possible internalization mechanism for carbon black -FW2, P60 and LB101- or TiO₂ NPs. Apart from particle size, other features such as anionic or cationic groups on surface, in particular by use of carboxylate or sulphate polystyrene latex beads, may also affect particles internalization. This will be investigated in the future. F-actin fluorescent microscopy images revealed an increase of cell vacuoles and electronic microscopy images show NPs trapped in these vacuoles. Single particles were no or rarely observed inside the cytoplasmic cell, perhaps due to their small size. Internalization seemed to be more important in LLC-PK₁ cells. These epithelial cells are polarized with a typical apical brush border membrane and also display an intense endocytic activity; these characteristics may therefore participate in enhanced NPs internalization in these cells compared to IP15 cells, and may therefore take part in the enhanced cytotoxicity observed in LLC-PK₁ cells.

ROS generation was investigated thereafter by use of the DCFH-DA assay. For this study, the oxidation of DCFH was chosen on the basis of the reactivity of this molecule with different ROS including H₂O₂ and superoxide. In the literature, various NPs have been demonstrated to generate ROS and to a greater extent more than larger particles, and this mechanism is thought to play a role in their cytotoxic effects 3738. In our experiments, no generation of ROS was evidenced in any cell type after treatment with P60, LB101 or TiO₂ particles. By contrast, exposure to FW2 NP, evidenced in our previous experiments as the most cytotoxic particle type, significantly enhanced production of ROS in both IP15 and LLC-PK₁ cells. The precise mechanisms activated by NPs to induce ROS generation and induce cytotoxicity are not known yet. NPs may activate intracellular pathways involving in particular increase in calcium intracellular concentrations 3940, involvement of MAPK 4142, and activation of transcription factors leading in particular to synthesis of pro-inflammatory cytokines and chemokines 136. In addition, when particles enter the cell, they could induce oxidative stress by disturbing the balance between oxidant and anti-oxidant processes, as for example the glutathione system. In epithelial cell line, Stone et al. 43, demonstrated that ultrafine carbon can induce glutathione depletion. The signaling mechanisms activated by such nanoparticles to induce cytotoxicity, in particular through generation of reactive oxygen species, therefore merit further attention.

In this work, we described for the first time the effects of CB and TiO₂nanoparticles in renal cells in vitro. These effects may have relevance in vivo considering the ability of nanoparticles to reach the systemic circulation and to be further distributed to numerous organs, in particular to the kidneys. We described the ability of these nanoparticles to exert a cytotoxic effect on renal cells, and suggested involvement of particle internalization as well as activation of intracellular mechanisms that might include generation of reactive oxygen species. However, we showed that cytotoxicity was particle type-, particle size- and cell type-dependent. Moreover, the doses of nanoparticles that were cytotoxic in vitro correspond to a concentration of 20 μg/cm², which may not be physiologically relevant. However, these results raised numerous questions regarding NPs long term secondary effects, bioavailability and precise toxicological effects, in particular in renal cells. The further development of in vitro models, in particular cell culture models, their validation regarding translation of the results obtained in vitro to in vivo models and their standardization will be of great help to assess these questions and investigate the nanoparticle precise mechanisms of action to induce cell toxicity.

The authors declare that they have no competing interests.

BL planned the study and contributed to all the sections; JJ carried out the cytotoxicity, ROS and immunofluorescent labeling studies. ES performed SEM and TEM nanoparticles analyses. DO and FM helped to prepare all cell specimens for electron microscopy. JF performed TEM observations. JC and PB provided advice and infrastructure. The study design and preparation of the manuscript were by COC.