2020 Volume 45 Issue 6 Pages 305-317
The aim of the present study was to evaluate the underlying mechanism of multi-walled carbon nanotubes (MWCNT) induced cellular response and their potential cross-talk, specifically, between endoplasmic reticulum (ER) stress, MAPK activation and apoptosis and how these nano-bio interactions depend on the physico-chemical properties of MWCNT. For this purpose, human bronchial epithelial (Beas2B) and human hepatoma (HepG2) cell lines, were exposed to five kinds of MWCNTs which differ in functionalization and aspect ratios. Tissue-specific sensitivity was evident for calcium homeostasis, ER-stress response, MAPK activation and apoptosis, which further depended on surface functionalization as well as aspect ratios of MWCNT. By applying specific pharmaceutical inhibitors, relevant biomarkers gene and proteins expressions, we found that possibly MWCNT induce activation of IRE1α-XPB1 pathway-mediated ER-stress response, which in turn trigger apoptosis through JNK activation in both type of cells but with variable intensity. The information presented here would have relevance in better understanding of MWCNT toxicity and their safer applications.
The unique physical, thermal and electrical properties of multi-walled carbon nanotubes (MWCNT), concentrically stacked cylindrical tubes of graphitic carbon, have led to wide range of applications not only in the fields of electronics, emission devices, and energy storage devices but also as drug delivery carriers, biosensor, therapeutics, and diagnostic materials in the biomedical and nanomedicine fields (Castranova et al., 2013; Heister et al., 2013; Lanone et al., 2013). Surface functionalization of MWCNT by introducing hydrophilic chemical groups (such as, hydroxyl -OH or carboxyl groups-COOH) is the principle way to overcome low solubility and to improve the efficacy of MWCNTs’ applications, specifically in the biomedical field (Prato et al., 2008; Vardharajula et al., 2012).
Considerable health concerning environmental, occupational as well as in consumer domain has immensely increased with the wide production and applications of MWCNT and their functionalized derivatives (Castranova et al., 2013; Li et al., 2013; Albini et al., 2015). The foregoing original research as well as recent reviews highlighted that length, diameter, surface functionalization, aspect ratio as well as the target cell types are the principle determinants of the nano-bio interactions of MWCNT (Sweeney et al., 2014; Hamilton et al., 2013; Han et al., 2012; Kim et al., 2011; Liu et al., 2012; Nagai et al., 2011; Bussy et al., 2012). A noted apprehension is that the modification of any one or a combination of these properties could modulate the mode of nano-bio interactions of MWCNT (Lanone et al., 2013). The well-known mechanisms of MWCNT-induced toxicity include oxidative stress, inflammatory responses, lung granulomas formation, interstitial fibrosis, DNA damage, autophagy, apoptosis, malignant transformations, DNA methylation etc. (Liu et al., 2013; Wang et al., 2011a; Ghosh et al., 2020; Toyokuni, 2013; Sargent et al., 2014; Chatterjee et al., 2017; Shvedova et al., 2012; Poulsen et al., 2015; Wang et al., 2011b; Hamilton et al., 2013). However, only a few studies reported on endoplasmic reticulum (ER) stress (Chang et al., 2018; Long et al., 2019; Sun et al., 2019) and MAPK activation (Lee et al., 2012; Sweeney et al., 2014) in mammalian cell model due to MWCNT exposure.
The endoplasmic reticulum (ER) is an essential organelle which plays critical role in cell survival and functions, such as, calcium homeostasis, lipid biosynthesis, etc. The ER maintains a surveillance mechanism for protein folding and works as a sensor for misfolded protein to guide them in the degradatory pathway (Malhotra and Kaufman, 2007). The misfolded or unfolded proteins caused ER stress, also known as unfolded protein response (UPR), which is an evolutionarily conserved cell signaling pathway (Harding et al., 2002; Garg et al., 2012; Shen et al., 2004). The activation of transmembrane stress-sensing proteins in the ER, including inositol-requiring protein 1 (IRE1), PKR-like endoplasmic reticulum kinase (PERK), and activating transcription factor-6 (ATF-6), either gives rise to signals for cellular self-recovery or, alternatively, promotes cell death (Garg et al., 2012; Harding et al., 2002; Tabas and Ron, 2011). Calcium homeostasis also plays crucial role in ER/mitochondria cross-talk and hence induces various modes of cell death, autophagy, apoptosis, necrosis (Pinton et al., 2008; Romagnoli et al., 2007; Kaufman and Malhotra, 2014; Malhotra and Kaufman, 2011).
The detailed mechanisms of ER-stress induction by MWCNT and how it is related to cell death remains to be elucidated. Here, we addressed the potentiality of MWCNT to induce ER stress response-mediated apoptosis in in vitro systems and how it differs with the cell types as well as MWCNTs’ physicochemical properties. To this end, we used five types of MWCNT which differ in length, diameter as well as surface functionalization and two different cultured cell systems, human bronchial epithelial cells (Beas2B cell) and human hepatoma cell line (HepG2 cell). The mechanisms were evaluated with/without relevant inhibitor, gene and protein expressions of ER-calcium depletion and mitochondrial calcium loading, ER-stress response, MAPK activation and apoptosis as endpoints.
The commercial MWCNTs [pristine, functionalized derivatives – carboxylated (-COOH), hydroxylated (-OH) and hydroxylated with different tube length and diameter] were purchased from CheapTubes.com (Cambridgeport, VT, USA) and were additionally characterized, further, with X-ray photoelectron spectroscopy (XPS), transmission electron microscopy (TEM) and a dynamic light scattering spectrometer (DLS).
XPS (Sigma Probe, ThermoVG, UK) was used to examine the surface chemical modification of the MWCNTs. The data were obtained by a hemispherical analyzer equipped by a monochromatic Al X-ray source (15 kV, 100W, 400 micrometer) operating at a vacuum (2 × 109 mb). The XPS peaks were analyzed by using Gaussian components after a Shirley background subtraction. The O/C atomic ratios of the samples were obtained by using peak area ratios of the XPS core levels and the sensitivity factor (SF) of each element in XPS.
Structure and shape were investigated by TEM (Carl Zeiss LIBRA 120). The samples for TEM were prepared by drop casting a diluted suspension (50 mg/L) onto a carbon film with 300 square mesh copper grids and dried at room temperature for 24 hr.
The size distribution and ζ-potential of the MWCNTs (30 mg/L in DMEM culture media) were evaluated by using a photal dynamic light scattering spectrometer (DLS) (ELSZ-1000, Otsuka Electronics Co., Inc., Osaka, Japan).
Beas2B cells (human bronchial epithelial cells) were cultured in DMEM/F12 (GIBCO, Grand Island, NY, USA) and HepG2 (human liver carcinoma cells) were cultured in MEM (GIBCO) and both were supplemented with 10% (v/v) fetal bovine serum and 1% (v/v) antibiotics, at 37°C in a 5% CO2 atmosphere.
The MWCNTs were freshly prepared in cell culture medium (DMEM/F12 or MEM) at the desired concentrations (respective EC20 of each MWCNT for each cell lines) with appropriate amount from the stock (1000 mg/L in distilled water) and were sonicated for 15 min before biological exposure.
Approximately, 5 x 103 cells/ well were seeded in 96-well plates 24 hr prior to treatment. Then the cells were pre-incubated with 1,2-bis(o-aminophenoxy)ethane-N,N,N’,N’-tetraacetetic acid) (BAPTA) (5 µM), 4-Phenylbutyrate (4-PBA) (5 mM) and solubrinal (25 µM) pretreatment for 1 hr followed by addition of MWCNTs (respective EC50 doses), and after incubation for 24 hr, cytotoxicity was measured with the EZ-Cytox method as described previously (Chatterjee et al., 2017). The dose of each inhibitor was selected by testing less than 10% cytotoxicity with the inhibitors exposure alone (data not shown).
Total RNA from MWCNT-treated samples (respective EC20 for 24 hr) were extracted by using RNA extraction kit (NucleoSpin, MachereyeNagel, Düren, Germany) and the quantity and quality of RNAs were detected in spectrophotometer as well as with agarose gel separation.
Synthesis of cDNAs was performed by a reverse transcriptase (RT) reaction and the PCR amplification was carried out with a thermal cycler (Bio-Rad, Hercules, CA, USA). Real-time RT-PCR analysis was accomplished with CFX manager (Bio-Rad) using the IQTM SYBR Green SuperMix (Bio-Rad). The primers were constructed (by Primer3plus) based on sequences available in NCBI and the qRT-PCR conditions were optimized (efficiency and sensitivity tests) for each primer prior to the experiment (Supplementary Table S2). Three biological replicates each in triplicate were used for each qRT-PCR analysis. Analysis of negative control reactions (without RT and all reagents except template) confirmed no DNA contamination. The gene expressions were normalized by using GAPDH and ACTB as a housekeeping genes.
Cell samples were harvested from each treatment and cell extracts were prepared in NP40 protein extraction buffer and the protein concentration was measured by Bradford method. Equal amounts (50 µg) of proteins were separated on 15% SDS-PAGE gels and transferred to nitrocellulose membranes. The membranes were blocked with 3% BSA in TBST at room temperature. The primary antibody was used at 1:1000 dilutions and the secondary antibody was used at 1:2000 dilutions. Blots were developed using enhanced chemiluminescence western blotting detection kit (Amersham, Little Chalfont, UK). The tested proteins were Bax, Bcl-2, γ-H2AX and β-Actin (antibodies purchased from Cell Signaling, Beverly, MA, USA).
The cells were plated on poly-L-lysine-coated cover glass (at 1 x 105 cells/mL) and exposed to MWCNTs respective EC20 doses for 24 hr. After incubation the respective fluorescence dyes were added to the cells, and the cells were further incubated for 30 min under the same growth conditions and were then washed with PBS after removing the loading dye. Microscopic images were collected using the fluorescence microscope (Leica DM IL) equipped with Leica DCF 420C camera. The endoplasmic reticulum (ER) stress was detected with ER-Tracker Blue-White DPX (1 µM, Molecular Probes, Life Technologies), mitochondrial Ca2+ level was determined with Rhod2 AM (Molecular Probes, Life Technologies) and the apoptotic bodies or condensed nuclei were detected after staining with Hoechst 33342 (10 µg/mL, Sigma-Aldrich, St. Louis, MO, USA).
Significance of differences among/between treatments was determined using one-way analysis of variance (ANOVA) and followed by a post-hoc test (Tukey, p < 0.05) in SPSS 12.0KO (SPSS Inc., Chicago, IL, USA) and graphs were prepared in Sigma Plot (Version 12.0).
The characterizations of all MWCNTs (including in-house analysis as well as from manufacturer) were previously published (Chatterjee et al., 2016, 2017) and summarized in Table 1. In total, five kinds of MWCNTs were used in this study and divided in two groups, one group differing in surface functionalization [pristine (Prist-SS), hydroxylation (OH-SS) and carboxylation (COOH-SS)] but possessing similar length and diameter (length: 0.5-2.0 μm, diameter: 8-15 nm) and the other group with same surface functionalization (hydroxylation) but differing in length and diameter (OH-SS with length: 0.5-2.0 μm, diameter: 8-15 nm, aspect ratio: 33.33-250; OH-LS with length: 10-30 μm, diameter: 8-15 nm, aspect ratio: 666.67-3750; OH-LB with length: 10-30 μm, diameter 20-30 nm, aspect ratio: 333.33-1500) (Table 1). The order of carbon content was exhibited as Prist-SS > OH-LB > COOH-SS > OH-LS > OH-SS in XPS measurement and oxygen content was in a reverse order (Table 1). The aggregation behavior and colloidal stability of MWCNTs under different time points (0 and 24 hr) were evaluated with simple visual evidence of settling in cell culture media as well as with DLS for ζ-potential (Table 1). The order of visual dispersion stability in culture media after 24 hr was OH-SS > COOH-SS > OH-LS > Prist-SS > OH-LB, which was also in agreement with measured ζ-potential (Table 1).
The EC20 of 24-hr doses (Table S1) of each MWCNT to each type of cells (Chatterjee et al., 2016, 2017), were chosen for further mechanistic endpoint, ER stress, MAPK-activation, DNA damage and apoptosis. For viability rescue due to inhibitor exposures, the respective EC50 of 24-hr doses (Table S1) were selected.
The treatment with MWCNTs at 24 hr caused depletion of ER calcium stores and significant increase in mitochondrial Ca2+ levels in both HepG2 (Fig. 1A) and Beas2B (Fig. 1B) cell lines. Almost all types of MWCNTs displayed clear mitochondrial Ca2+ level overloading except the OH-LB one, and the order in HepG2 cells was OH-SS > OH-LS ≥ Prist-SS > COOH-SS > OH-LB (Fig. 1A) while the order in Beas2B cells was COOH-SS ≥ Prist-SS > OH-SS ≥ OH-LS > OH-LB (Fig. 1B). Moreover, marked time-dependent increases in mitochondrial Ca2+ levels from 4 hr to 24 hr were evident in both types of cell lines (Fig. S1A and S1B). Likewise, depletion of ER Ca2+ stores and mitochondrial Ca2+ overloading, all types of MWCNTs displayed sharp endoplasmic reticulum (ER) stress. The order of ER stress in HepG2 cells was OH-SS > Prist-SS > OH-LS > COOH-SS > OH-LB (Fig. 1C), whereas, the order of COOH-SS > Prist-SS ≥ OH-SS > OH-LS > OH-LB was noticed in Beas2B Cells (Fig. 1D). In addition, the time-dependent increases in ER stress were evident between 4-hr to 24-hr exposures to MWCNTs (Fig. S1C and S1D). The pretreatment with known calcium chelator, BAPTA, showed rescue in viability, mainly in OH-SS, in both type of cells and COOH-SS in Beas2B (Fig. 1E) which was actually in agreement with mitochondrial Ca2+ level overloading (Fig. 1A and Fig. 1B). It is a well-known fact that calcium homeostasis is tightly linked with endoplasmic reticulum function with mitochondrial bioenergetics (Kaufman and Malhotra, 2014). In addition, the ER stress inhibitors, 4 BPA, exhibited viability rescue in almost all MWCNT treatments except OH-LB (Fig. 1E) which further support that the MWCNTs caused ER stress in both types of cells (Fig. 1C and Fig. 1D). To shed light on the effects of MWCNT-mediated Ca2+ depletion from ER, we have chosen gene expressions of the three luminal ER Ca2+-regulating proteins (Clapham, 2007; Prins and Michalak, 2011). The IP3R1 receptor-mediated calcium release from ER were evident in MWCNT-treated HepG2 cells but not in Beas2B cells, while calcineurin (PPP3CA) up-regulations were observed in both types of cells, except OH-LB-treated Beas2B cells (Fig. 1F). The deregulation of calreticulin (CALR), the Ca2+ binding protein in ER lumen, was observed both in HepG2 cells and Beas2B cells, specifically in the OH-LS- and COOH-SS-exposed ones (Fig. 1F). Furthermore, the expressions of ER stress sensor genes, GRP94, GRP78, PERK, ATF4, eIF2a, CHOP, IRE1a, ATF6, and spliced XBP1 were measured to elucidate the underlying mechanism of MWCNT-mediated ER stress response. The highly significant increases in IRE1a gene expressions were evident in both types of MWCNT-exposed cells, particularly in COOH-SS-treated cells (Fig. 1G). Besides, GRP78 and XBP1 (spliced) gene expressions were also found as significant, specifically in COOH-SS-exposed cells. Conversely, PERK and eIF2a genes did not exhibit significant alterations in their expression, except in OH-SS-exposed cells (Fig. 1G). Thus, MWCNTs treatment, in general, causes ER stress through IRE1a-XBP1 pathway but does not evoke the PERK-eIF2a pathway, except OH-SS, which might possibly elicit both pathways in both cells (Fig. 1G). This hypothesis was further supported by the pretreatments with solubrinal, the ER stress inhibitor specific for eIF2a inhibition. The solubrinal pretreatment did not show significant rescue in viability of MWCNT-exposed cells, except both OH-SS-treated cells (Fig. 1E). In summary, tissue-specific ER-stress response varies among surface functionalization as well as length and diameter. Our results showed that in respect of ER-stress and mitochondrial Ca2+ overloading, COOH-SS is the most sensitive in Beas2B cells whereas OH-SS is the most sensitive in HepG2 cells and shorter diameter and shorter length possess higher response than their longer counterparts. In agreement with our study, a previous study reported that significant modulation of autophagy-ER stress mediated cytotoxicity in shorter diameter MWCNT in comparison to longer ones (Zhao et al., 2019). Unlike another previous report, we found shorter length MWCNT caused more profound ER-stress response (Long et al., 2017). The reason of differential response might lie in the choice of cell model (Sweeney et al., 2014).
Determination of calcium homeostasis and endoplasmic reticulum (ER) stress in HepG2 and BEAS-2B cells due to MWCNTs exposures for 24 hr at their respective EC20 doses. A. Measurement of mitochondrial Ca2+ levels by Rhod2 AM (1 µM) probing in HepG2 cells. B. Measurement of mitochondrial Ca2+ levels by Rhod2 AM (1 µM) probing in BEAS-2B cells. C. Evaluation of ER staining with ER tracker blue-white DPX probing in HepG2 cells. D. Evaluation of ER staining with ER tracker blue-white DPX probing in BEAS-2B cells. E. Determination of mechanism of mitochondrial Ca2+ levels and ER-stress with BAPTA, 4-PBA and solubrinal pretreatment and the rescue of viability due to MWCNTs exposure at their respective EC50 doses. F. Alteration of calcium channel-related genes (CAL, IP3R1 and PPP3CA) expressions. G. Variation of ER-stress-related genes (GPR94, GPR78, CHOP, ATF4, ATF6, PERK, eIE2a, IRE1a, XBP1) expressions. Data presented as mean ± SEM, *p < 0.05, **p < 0.001, *** p < 0.0001.
Very interesting tissue-specific and functionalization-specific effects were evident in MAPKs activations (Fig. 2). The phospho ERK1/2 (p-ERK1/2) were only differentially expressed in Beas2B cells but not in HepG2 cells. The suppression of p-ERK1/2 was evident due to MWCNTs exposure, specifically in Prist-SS-exposed Beas2B cells. By the same token, phospho p38 (p-p38) was only activated in Beas2B cells and showed no changes in HepG2 cells. In particular, OH-LB- and OH-SS-treated Beas2B cells showed significant higher activation of p-p38 than control (Fig. 2). By the way of contrast, phospho JNK (p-JNK) was commonly activated in both types of cells, and the order of activation in Beas2B cells was Prist-SS > COOH-SS > OH-SS ≥ OH-LS > OH-LB while the order of activation in HepG2 cells was OH-LS > OH-SS > Prist-SS > OH-LB > COOH-SS (Fig. 2). In agreement with our study, Sweeney et al. (2014) reported markedly altered MAPK (ERK, JNK and p38) signaling proteins responses that vary according to the target cell type, as well as the aspect ratio of the MWCNT. Similar MAPK activation, i.e., activation of JNK but not ERK1/2, mode was evidenced in MWCNT-exposed IL-1 cell model (Arnoldussen et al., 2015). Besides, MAPK (JNK, 38MAPK, ERK) activation were reported in COOH-MWCNT-treated marcophages (Zhang et al., 2015), conversely, no significant MAPK (ERK, JNK and p38) activation was reported in MWCNT-treated mouse macrophages (Hirano et al., 2008).
Effects of MWCNTs on MAPK (ERK, p-ERK, JNK, p-JNK, p38 and p-p38) protein expressions due to MWCNTs exposure for 24 hr at their respective EC20 doses.
The order of double strand break potency, the expressions of γ-H2AX, in MWCNT-exposed HepG2 cells was OH-SS > COOH-SS ≥ Prist-SS > OH-LS ≥ OH-LB, conversely, the order in MWCNT-treated Beas2B cells was COOH-SS > OH-LS > OH-SS > OH-LB ≥ Prist-SS (Fig. 3). Previous studies also support that MWCNT cause DNA strand break and higher γ-H2AX protein expressions (Guo et al., 2011; Cveticanin et al., 2010). The genotoxicity, specifically the DNA double strand break, shed light on the fact of carcinogenic potentiality of MWCNTs (Wang et al., 2011a; Sargent et al., 2014; Nagai et al., 2011; Lindberg et al., 2013; Toyokuni, 2013).
Double-stranded DNA damage measured with γ-H2AX protein expression after 24 hr of exposure of MWCNTs at their respective EC20 doses.
The MWCNT-induced apoptosis was measured in reference to nuclear morphological criteria by employing Hoechst 33342 staining. The MWCNT-treated cells showed DNA condensation and fragmentation at 24 hr in HepG2 cells (Fig. 4A) as well as Beas2B cells (Fig. 4B) at their respective EC20 doses. In general, Prist-SS showed the highest fluorescence intensity in both types of cells. But the time dependency of apoptosis induction was not clear from 4 hr to 24 hr with Hoechst 33342 staining (Fig. S2A and S2B). The prominent apoptotic features, an increase in BAX and a corresponding decrease in Bcl-2 expressions, were evident in MWCNT-treated cells at respective EC20 doses (Fig. 4C). The imbalance of BAX/BCl2 ratio was most significant in COOH-SS-followed by Prist-SS-treated Beas2B cells; conversely, it was highest in Prist-SS-exposed HepG2 cells followed by OH-SS ones; but the highest BAX, the pro-apoptotic marker, expression was found in OH-SS-exposed HepG2 cells (Fig. 4C). To find out the ER-stress mediated apoptosis, Bax/Bcl-2 expressions were also measured in 4-PBA ± MWCNT-treated conditions. The alterations in Bax (increased) and Bcl-2 (decreased) expressions due to MWCNT exposures were rescued in 4-PBA pre-treatment (Fig. 4C). In general, MWCNT activates apoptosis pathway through ER-stress, however, the intensity of this cellular response depends on aspect ratio, surface functionalization (Sun et al., 2019). In addition, other executioners of apoptosis, the caspases (CASP9, CASP8, CASP3) and FAS gene expressions displayed further tissue type as well as MWCNTs functionalization specificities (Fig. 4D). The FAS gene was noticeably expressed only in OH-LB and Prist-SS exposed HepG2 cells, whereas it was significantly expressed in all MWCNT-treated Beas2B cells, except OH-SS ones. CASP9 was the most significant gene expressed in HepG2 cells and the order of gene expressions was OH-SS > COOH-SS > Prist-SS > OH-LS > OH-LB. However, CASP8 was the most ascertainably expressed gene in Beas2B cells and the order of expressions was COOH-SS > OH-SS > Prist-SS > OH-LS > OH-LB (Fig. 4D). Therefore, the MWCNT-exposed HepG2 cells might use mitochondrial apoptotic pathway and Beas2B cells use extrinsic apoptotic pathway. The extrinsic apoptotic pathway was most common in Prist-SS and OH-LB MWCNTs treatment, irrespective of cell types. In agreement with our study, it was demonstrated that MWCNT induced apoptosis via mitochondrial pathway in macrophage cells (Wang et al., 2012), in rat lung epithelial cells (Ravichandran et al., 2009). In contrast to our study, it is reported that hydroxylation (Liu et al., 2014b) and carboxylation (Liu et al., 2014a) of MWCNT decreased the mitochondrial apoptotic pathway and thereby functionalization of MWCNT reduced cytotoxicity. This might be because of variations in cell types as well as dimension and aspect ratio of MWCNT.
Evaluation of MWCNTs induced apoptosis in HepG2 cells and BEAS-2B cells at their respective EC20 doses for 24 hr. A. Measurement of DNA condensation and fragmentation measured by Hoechst 33342 staining in HepG2 cells. B. Measurement of DNA condensation and fragmentation measured by Hoechst 33342 staining in BEAS-2B cells. C. Changes in Bcl-2 (antiapoptotic) and Bax (pro-apoptotic) proteins expression. D. Alteration of FAS and caspases genes (CASP9, CASP8, CASP3) expressions. Data presented as mean ± SEM, *p < 0.05, **p < 0.001.
In summary, we found that cellular responses vary according to the target cell type, aspect ratio as well as the surface functionalization of the MWCNT. In general, HepG2 cells were more sensitive to hydroxylated MWCNT (short diameter, short length) while Beas2B cells were more sensitive to carboxylated (short diameter, short length) MWCNT. Significant increase in ER Ca2+ depletion and mitochondrial Ca2+ overloading, ER-stress, DNA double strand damage, MAPK (JNK) activation and increased apoptosis in MWCNT exposed cells. Most probably, the IRE1α-XPB1 pathway mediated ER-stress response trigger apoptosis through JNK in MWCNT-treated cells. The results of the present study could be useful for regulatory consideration as well as safe-by-design concept of MWCNT applications, in particular, for biomedical applications.
This work was supported by the 2017 Research Fund of the University of Seoul.
The authors declare that there is no conflict of interest.