Uptake and cytotoxic effects of multi-walled carbon nanotubes in human bronchial epithelial cells

Abstract

Carbon nanotubes (CNT) are cytotoxic to several cell types. However, the mechanism of CNT toxicity has not been fully studied, and dosimetric analyses of CNT in the cell culture system are lacking. Here, we describe a novel, high throughput method to measure cellular uptake of CNT using turbimetry. BEAS-2B, a human bronchial epithelial cell line, was used to investigate cellular uptake, cytotoxicity, and inflammatory effects of multi-walled CNT (MWCNT). The cytotoxicity of MWCNT was higher than that of crocidolite asbestos in BEAS-2B cells. The IC₅₀ of MWCNT was 12 μg/ml, whereas that of asbestos (crocidolite) was 678 μg/ml. Over the course of 5 to 8 h, BEAS-2B cells took up 17–18% of the MWCNT when they were added to the culture medium at a concentration of 10 μg/ml. BEAS-2B cells were exposed to 2, 5, or 10 μg/ml of MWCNT, and total RNA was extracted for cytokine cDNA primer array assays. The culture supernatant was collected for cytokine antibody array assays. Cytokines IL-6 and IL-8 increased in a dose dependent manner at both the mRNA and protein levels. Migration inhibitory factor (MIF) also increased in the culture supernatant in response to MWCNT. A phosphokinase array study using lysates from BEAS-2B cells exposed to MWCNT indicated that phosphorylation of p38, ERK1, and HSP27 increased significantly in response to MWCNT. Results from a reporter gene assays using the NF-κB or AP-1 promoter linked to the luciferase gene in transiently transfected CHO-KI cells revealed that NF-κB was activated following MWCNT exposure, while AP-1 was not changed. Collectively, MWCNT activated NF-κB, enhanced phosphorylation of MAP kinase pathway components, and increased production of proinflammatory cytokines in human bronchial epithelial cells.

Introduction

A group of carbon nanotubes (CNT) is one of the most remarkable nanomaterials because of its mechanical and electric characteristics and because of its fibrous feature like asbestos. Both CNT and asbestos are biopersistent and able to cause persistent stress in pulmonary cells when deposited in the lung, although a recent study suggested that single-walled CNT were degraded by myeloperoxidase in the presence of hydrogen peroxide and chloride ions (Kagan et al., 2010). The inflammatory and fibrotic potency of CNT in the lung has been well documented in mice (Lam et al., 2004, Shvedova et al., 2005) and rats (Warheit et al., 2004, Muller et al., 2008) following intratracheal instillation CNT. However, a bolus intratracheal instillation of biopersistent particulates may cause much more serious effects in the lung than a moderate inhalational exposure (Mitchell et al., 2007, Muller et al., 2009). Acute exposure to a high concentration of CNT (30 mg/m³) caused subpleural fibrosis in mice, while a low concentration (1 mg/m³) did not cause such effects (Ryman-Rasmussen et al., 2009). Inhalation toxicity of CNT clearly depends on the burden on the lung (Ma-Hock et al., 2009).

The length and the aspect ratio, defined as a/√bc where a–c are the three dimensions of a particle, are important determinants for the toxicity of fiber both in vitro and in vivo. Fenoglio et al. (2000) reported that the cytotoxicity of pure silica zeolites was correlated with both the external surface area and aspect ratio in J774 murine macrophages. Stanton et al. (1981) clearly demonstrated that the incidence of pleural sarcoma correlated best with a number of fibers that measured 0.25 μm or less in diameter and more than 8 μm in length in rats when implanted in the pleurae. Poland et al. (2008) recently reported that long amosite and CNT are more potent than corresponding short fibers in causing inflammation and granulomatous lesions in the diaphragm of mice following intraperitoneal injection.

In vitro toxicity of CNT has been reported in several types of cells including macrophages (Jia et al., 2005, Kagan et al., 2006, Hirano et al., 2008), epithelial cells (Cui et al., 2005, Ye et al., 2009), mesothelial cells (Wick et al., 2007, Pacurari et al., 2008) and keratinocytes (Shvedova et al., 2003, Monteiro-Riviere et al., 2005). It is obvious that CNT are cytotoxic and that CNT-exposed cells exhibit inflammatory responses. One of the most serious difficulties nanotoxicologists face in pursuing in vitro tests is a poor dispersion of CNT in the culture medium. It has been shown that single-walled CNT (SWCNT) readily agglomerate in the culture medium and no internalization of SWCNT was observed in human alveolar carcinoma cells (A549) at 24 h post exposure, although the number of lamellar bodies was increased by SWCNT (Davoren et al., 2007). Dipalmitoylphophatidylcholine (DPPC) improved the degree of SWCNT dispersion in medium and in turn, increased the in vitro toxicity of SWCNT in A549 human lung carcinoma cells (Herzog et al., 2009). In contrast multi-walled CNT (MWCNT) were found as agglomerates in the culture medium even in the presence of 1% DPPC or ethanol and caused neither cell membrane injury nor oxidative stress in A549 cells (Tabet et al., 2009). The A549 cells did not associate with MWCNT whereas chrysotile and crocidolite asbestos fibers were internalized in the cells in this study. Contrary to this Wick et al. (2007) reported that agglomerated SWCNT caused more adverse effects than well-dispersed ones in mesothelial cells (MSTO-211H); however, most cells appeared to have associated well with SWCNT even though SWCNT were agglomerated. CNT are hydrophobic because of their graphene structure and easily agglomerate in the aqueous solution. The agglomeration of particles affects the effective dose in the cellular systems. Thus, a quantitative measurement of cellular uptake of CNT is essential for assessing the actual effective dose in in vitro systems.

The mode of cytotoxic effects of CNT may differ between cell types. We proposed that macrophages are killed by MWCNT by direct membrane injury or necrosis rather than apoptosis (Hirano et al., 2008). It has been shown that CNT generated reactive oxygen species (ROS), activated NF-κB, and increased IL-8 mRNA levels in A549 human lung epithelial cells (Ye et al., 2009). Peroxide accumulation and lipid peroxidation were observed in SWCNT-exposed human epidermal keratinocytes (HaCaT) (Shvedova et al., 2003). Macrophages are motile cells and the cellular responses of macrophages to CNT may be different from those of epithelial cells.

Although it is clear that well-dispersed CNT are cytotoxic once the cell membrane reacts with CNT, the mechanism of CNT toxicity has not been fully elucidated. We herein report a quantitative and high-throughput method to measure cellular uptake of CNT in vitro. We used a reporter gene assay, antibody arrays for cytokines and protein phosphorylation, and cDNA primer arrays to assess the comprehensive cellular responses to CNT in human bronchial epithelial cells.