In vitro toxicity evaluation of heavy metals in urban air particulate matter on human lung epithelial cells

Science of the Total Environment 678 (2019) 301–308

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

In vitro toxicity evaluation of heavy metals in urban air particulate matter on human lung epithelial cells Yue Yuan, Yun Wu ⁎, Xinlei Ge ⁎, Dongyang Nie, Mei Wang, Haitao Zhou, Mindong Chen Jiangsu Key Laboratory of Atmospheric Environment Monitoring and Pollution Control (AEMPC), Collaborative Innovation Center of Atmospheric Environment and Equipment Technology (AEET), School of Environmental Science and Engineering, Nanjing University of Information Science & Technology, Nanjing 210044, China

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• Metal fractions made a great contribution (22.9% ± 11.5%) to PM-related cytotoxicity. • Zn, Cr, Mn, Fe, Cu, and Pb might account for PM toxicity in A549 cells. • Mixture toxicity was increased when co-exposed with Mn, but decreased in the presence of Fe. • The IAI model exhibited better stability in metal mixture toxicity prediction.

a r t i c l e

i n f o

Article history: Received 3 December 2018 Received in revised form 28 April 2019 Accepted 29 April 2019 Available online 30 April 2019 Editor: Pingqing Fu Keywords: Atmospheric particulate matter Heavy metals A549 cells Combined toxicity Toxicity prediction model

a b s t r a c t Heavy metals are widely recognized as toxic components in urban air particulate matter (PM). However, the major toxic metals and their interactions are poorly understood. In this study, we attempted to explore the toxicity contribution and combined effects of PM-bounded metals in human lung epithelial cells (A549). Real-time cell analysis indicated that the critical toxic concentration (EC50) of PM detected in this study was 107.90 mg/L (r2 = 1.00, p b 0.01). The cell viability of A549 increased significantly (12.3%) after metal removal in PM, demonstrating an important contribution of metal components to PM toxicity. Among eleven elements examined (Zn, Cr, Mn, Fe, Ni, Cu, As, Se, Sr, Cd, and Pb), six heavy metals (Zn, Cr, Mn, Fe, Cu, and Pb) might account for PM toxicity in A549 cells, and their co-exposure led to a high mortality of A549 cells (36.5 ± 7.3%). For combination treatments, cell mortality caused by single or multiple metal mixtures was usually alleviated by Fe addition, while it was often aggravated in the presence of Mn. The varying effects of other metals (Zn, Cu, Pb and Cr) on different metal mixtures might be explained by their interactions (e.g., similar or dissimilar membrane transporters and intracellular targets). Furthermore, the concentration addition model (CA), independent action model (IA), integrated addition model (IAM) and integrated addition and interaction model (IAI) were used to predict mixture toxicity, and the IAI model exhibited the least variation between observed and predicted toxic effects (r2 = 0.87, p b 0.01). Our results highlight the potential contribution from heavy metals and their interactions to PM toxicity, and promote the application of toxicity prediction models on metal components in PM. © 2019 Elsevier B.V. All rights reserved.

1. Introduction ⁎ Corresponding authors. E-mail addresses: [email protected] (Y. Wu), [email protected] (X. Ge).

https://doi.org/10.1016/j.scitotenv.2019.04.431 0048-9697/© 2019 Elsevier B.V. All rights reserved.

Particulate matter (PM) pollution has been reported to be a crucial risk factor for adverse health outcomes, e.g., cardiovascular disease

302

Y. Yuan et al. / Science of the Total Environment 678 (2019) 301–308

and respiratory disease (Nie et al., 2018). PM is a complicated mixture of elemental carbon, sulfate, nitrate, ammonium, metals, and many other organic compounds (Ge et al., 2011; Badaloni et al., 2017). Among these PM constituents, it has been suggested that some metals are associated with the PM-induced biological damages observed in laboratory cells (Uski et al., 2015), animals (Lippmann and Chen, 2009) and cohort communities (Suh et al., 2011). Therefore, although heavy metals account for only a small proportion of PM mass, their potential toxicity should not be ignored or underrated (Lai et al., 2018; Wang et al., 2018). Both essential and non-essential metals have been detected in ambient PM. Even for those that have bio-functions, daily intake or exposure would be harmful to human beings if exceeding the toxicity thresholds. Generally, several mechanisms are involved in heavy metal toxicity: 1) combining with the sulphydryl groups of cysteine or sulfur atoms of methionine; 2) replacing the metal ion in some functional enzymes; 3) catalyzing the formation of reactive oxygen species (Gurer and Ercal, 2000; Wu et al., 2016). Different metals might have different target sites, target organs or modes of action, thus leading to different toxicity profiles (Wu et al., 2016). It should be noted that multiple heavy metals often coexist in atmospheric particulates, metal-metal interactions are critical to the final observed effect. Because a metal element can affect the accumulation and distribution of another, and synergism or antagonism would occur after multi-metal co-exposure (Mwamba et al., 2016). Wu et al. (2012) found that a mixture of Cu and Zn showed an antagonistic effect, while the mixture of Cu and Ag had a synergistic effect on primary human endometrial epithelial cells. Using sea urchin embryo-larval bioassay, Xu et al. (2011) discovered that the interactions among Cu, Pb, Zn and Cd were mainly synergistic in binary combinations, but they were generally concentration-additive in ternary or quaternary mixtures. Numerous studies have been conducted to explore the toxicity of single metal or their combined toxicity in the aquatic and terrestrial environment. However, very little is known about the metal-metal interactions in atmospheric particles, especially in human subject experimentation. The content and composition of heavy metals vary greatly among atmospheric particles of different origins and sizes; it is virtually impossible to mimic the toxic effects of all possible combinations experimentally. Several toxicity prediction models have been proposed to study the joint effect of pollutants, some of them can be used for estimating toxicity of heavy metals (Kim et al., 2012; Meyer et al., 2015). The concentration addition (CA) model assumes that different substances have identical molecular target site and show the same effect (Loewe and Muischnek, 1926; Plackett and Hewlett, 1952; Pöch, 1993). Correspondingly, the independent action (IA) model hypothesizes that the components in a mixture have different modes of action but with the same toxicological endpoint (Bliss, 1939; Plackett and Hewlett, 1952; Pöch, 1993). In some cases these two processes cannot be treated separately, the integrated addition model (IAM) is then developed by merging them together (Olmstead and Leblanc, 2004). These models, although simple in application, have ignored any interactions that might likely exist among metals. Therefore, the integrated addition and interaction model (IAI) is extended by introducing the parameter K-function to express the influence of one chemical on another, which inevitably leads to increases in complexity (Rider and Leblanc, 2005). According to our knowledge, relatively few studies have examined the feasibility of these toxicity prediction models for metals contained in PM. Therefore, the present study attempted to investigate the toxic contribution, combined effects and model prediction of metals contained in PM on human lung epithelial cells (A549). We selected an urban particulate standard reference material (SRM 1648a) and estimated the contribution of heavy metals to particle toxicity. We further studied the interactions among the toxic heavy metals we identified in a permutation. Afterwards, we compared the differences between observed and predicted effects with four toxicity prediction models (i.e., the CA, IA, IAM, and IAI models), to explore the most reliable mixture model for heavy metals.

2. Materials and methods 2.1. Chemicals A standard ambient particulate matter sample (SRM 1648a, with an average diameter of 5.85 μm) from the National Institute of Standards and Technology (NIST) was used to evaluate metal-related toxicity. The concentrations of eleven metals (Zn, Cr, Mn, Fe, Ni, Cu, As, Se, Sr, Cd and Pb) in the PM samples range from 0.66 to 4800 mg/kg, which are reported in the Certificate of Analysis. Analytical grade metal salts of ZnCl2, CrCl2, NiCl2·6H2O, CuCl2·2H2O, SrCl2·6H2O, CdCl2 and PbCl2 were purchased from Aladdin (Shanghai, China). MnCl2, FeCl3·6H2O, AsNaO2 and Na2O3Se were obtained from Xiya Reagent (Shandong, China). Chelex 100 chelating Resin was purchased from Bio-Rad Laboratories (USA) to remove heavy metals from particles. NaOH (AR grade) was obtained from Sinopharm Chemical Reagent Co., Ltd. (China), and HCl (GR grade) was purchased from Shanghai Hushi Laboratorial Equipment Co., Ltd. 2.2. Cell culture The A549 cells (human lung epithelial cells), obtained from Nanjing University of Chinese Medicine, were cultured in Minimum Essential Medium (MEM, KeyGEN BioTECH, China) with 10% heat-inactivated fetal bovine serum (FBS, Gibco). The cells were routinely grown in a humidified incubator at 37 °C in a 5% CO2 atmosphere. All experiments were performed at 80% cell confluence. Phosphate Buffer Saline (PBS) and trypsin (0.25%) were purchased from KeyGEN BioTECH, China. 2.3. Real-time cell analysis The level of exposure to PM in the following experiments was determined by the xCELLigence real time cell analysis system (RTCA S16, ACEA Biosciences). The A549 cells were seeded in 16-well E-plates at a density of 5000 cells per well. The E-plate was first kept at room temperature for 30 min, then it was transferred into the humidified incubator for 16 h an until stable baseline was reached (Murphy et al., 2016). After washing twice with PBS, the adhered cells were subjected to PM exposure for 24 h. An SRM 1648a stock solution was prepared with sterile PBS (0.01 M, pH = 7.4), sterilized by ultraviolet irradiation and then diluted with FBS-free MEM to form the working solutions. There were 4 PM-concentration treatments with two replicates each, and the nominal concentrations were 0, 50, 100 and 200 mg/L. These exposure concentrations of PM were used based on the results of our preliminary experiments for PM concentrations ranging from 0 to1000 mg/L. Cell adhesion and proliferation were continuously monitored by detecting the electrical impedance. The transformed cell index (CI) was analyzed with RTCA Data Analysis Software 1.0 and the calculated EC50 was used for the following experiments. 2.4. Cell exposure At the PM concentration of EC50, the concentrations of metals contained in PM solution were 480 μg/L of Zn, 40.2 μg/L of Cr, 79 μg/L of Mn, 3920 μg/L of Fe, 8.11 μg/L of Ni, 61 μg/L of Cu, 11.55 μg/L of As, 2.84 μg/L of Se, 21.5 μg/L of Sr, 7.27 μg/L of Cd, and 655 μg/L of Pb. Such metal concentrations were prepared to mimic the concentrations of metals in PM, individually and jointly. In single-metal exposure experiments, A549 cells were exposed to eleven metals at the designated concentrations, and six toxic elements were identified in the preexperiment. Then eight concentrations of each toxic metal were prepared in sextuplicate and their concentrations were: 0.01, 0.1, 1, 10, 100, 1000, 1500, 2000 mg/L for Fe, 0.01, 0.1, 0.5, 1, 5, 10, 50, 100 mg/L for Mn, 0.01, 0.1, 0.5, 0.6, 0.8, 1, 5, 10 mg/L for Pb, 0.01, 0.1, 0.5, 1, 2, 5, 10, 50 mg/L for Cu, 0.01, 0.1, 1, 10, 50, 100, 500, 1000 mg/L for Cr, and 0.01, 0.1, 1, 3, 5, 10, 20, 30 mg/L for Zn. In mixed-metal toxicity tests,

Y. Yuan et al. / Science of the Total Environment 678 (2019) 301–308

the previously identified toxic metals were mixed combinationally to investigate their joint action on cytotoxicity. Chelex resin was demonstrated to effectively remove most of the metal ion contaminants without affecting the nonmetallic components (Shafer et al., 2010; Heo et al., 2015). A stock solution of SRM 1648a was chelated with Chelex 100 resin, filtered through nylon net (75 μm), and then diluted to the EC50 concentration to explore the contribution of heavy metals to PMrelated cytotoxicity. 2.5. Cell viability assay

2.6. Toxicity modeling The joint effects of heavy metals on A549 cells were modeled and predicted by using four mathematical models: concentration addition model (CA), independent action model (IA), integrated addition model (IAM), and integrated addition and interaction model (IAI), respectively. Their parameters were derived from the concentration response curves (CRCs) of individual metals or mixtures on A549 cells. Based on the CA model, the total concentration of components in a mixture provoking x effect (ECxmix) can be calculated using the following formula: ECxmix ¼


n ∑ i¼1

pi ECxi

−1 ;

ð1Þ

where ECxi is the concentration of the ith component provoking x effect solely; pi is the percentage of the ith component with respect to the total concentration of components in the mixture. In accordance with the conception of the IA model, joint effects can be obtained by the product of fractional effects of individual mixture components: EðCmix Þ ¼ 1−

n Y ½1−EðCi Þ;

ð2Þ

i¼1

where E(Ci) is the individual effect of the ith component if present in the concentration C; E(Cmix) is the total effect of the mixture with the total concentration Cmix. Besides, Olmstead and Leblanc (2005) merged the CA and IA models into one model (IAM), and calculated the joint toxicity according to Eq. (3):

Rmix

the concentration of the ith component (i = 1…n); p' is the average power or average slope of the individual components within a group. As a comparison, an extended IAI model was used in this study to deal with components interacted with each other. Here the combined effect was calculated based on the K-function for the interactions and the model was written as following (Rider and Leblanc, 2005):

Rmix

The cytotoxicity of PM and associated heavy metals was measured by WST-8 assay (Cell Counting Kit-8, Beyotime, China) according to the manufacturer instructions. Briefly, A549 cells were seeded in 96-well microplates at a density of 5 × 103 cells per well, and incubated overnight at 37 °C, 5% CO2. Afterward, culture medium was replaced by the prepared samples in sextuplicate. The control treatment was identical to the test treatment with the exception of sample placement. After 24 h of exposure, 10 μL of CCK-8 solution was added into each well and incubated at 37 °C for 1 h. The absorbance of samples was quantified using a microplate reader (Thermo Scientific, USA) at 450 nm. The viability of cells was calculated according to the following formula: cell viability (%) = (A − B) / (C − B) × 100%, where A represents the absorbance of samples; B is the absorbance of blank controls; and C is the absorbance of negative controls. All experiments were replicated at least three times.

8 > > > > < n > Y ¼ 1− 1− > I¼1 > > > > 1þ :

1 1 C n ∑i¼1 EC i

9 > > > > > = !p0 >; > > > > ;

303

9 8 > > > > > > > > > > > > = n < Y 1 ; ¼ 1− 1− 1 > > > I¼1 > > 1 þ 0> > >   > > > n k ðC ÞC p > : ∑i¼1 a;iECa50 i ;

ð4Þ

i

where ka,i is a function describing the extent to which component a presents in the mixture as concentration Ca affects the effective concentration of component i. 2.7. Statistical analysis The data were analyzed with MATLAB software (MathWorks, USA) and were expressed as mean ± standard deviation (SD). The EC50 in WST-8 assay was estimated by applying a linear interpolation method (ICPIN software, version 2.0, USEPA). Statistical analysis was performed using SPSS version 22.0 (IBM company, USA). Differences among treatments were assessed by one-way analysis of variance followed by Tukey's post hoc test and statistical significance was accepted at p b 0.05. 3. Results 3.1. Optical exposure concentration and single toxicity The toxic effects of heavy metals in PM vary greatly and depend on many factors, including atmospheric PM composition and particle size (Zou et al., 2016; Jia et al., 2017). To reduce the number of possible combinations, we used an urban particulate matter standard reference material, and investigated the “effective dose” or the EC50 by using the xCELLigence RTCA S16 system in vitro (Fig. 1). The EC50 value of PM measured in this study was 107.90 mg/L (r2 = 1.00, p b 0.01). Therefore, 100 mg/L of PM suspension was selected as the nominal exposure concentration for the next experiments. Among the heavy metals in SRM 1648a (100 mg/L), we first explored which heavy metal(s) contributed to the observed toxicity of PM under the hypothesis that all metal ions could be released into the liquid suspension. We discovered that, among all tested elements (Zn, Cr, Mn, Fe, Ni, Cu, As, Se, Sr, Cd, Pb), exposure to Zn (480 μg/L), Cr (40.2 μg/L), Mn (79 μg/L), Fe (3920 μg/L), Cu (61 μg/L) or Pb (655 μg/L) significantly decreased the cell viability of A549 (Fig. 2). In contrast, exposure to Ni (8.11 μg/L), As (11.55 μg/L), Se (2.84 μg/L), Sr (21.5 μg/L) or Cd (7.37 μg/L) did not significantly suppress A549 cell viability. To further investigate the characteristics of the six toxic metals, their dose-response curves were tested under identical conditions (Fig. 3), and the calculated EC50 values of cell index were as follows: 4.95 mg/L of Zn, 471.19 mg/L of Cr, 34.73 mg/L of Mn, 1192.05 mg/L of Fe, 3.81 mg/L of Cu or 0.93 mg/L of Pb. 3.2. Mixture toxicity of heavy metals in PM

ð3Þ

50i

where Rmix is the combined toxicity of components in the mixture; EC50i is the concentration of the ith component that causes a 50% effect; Ci is

We studied the joint toxicity of these six heavy metals (Zn, Cr, Mn, Fe, Cu and Pb) in a permutation. For the binary mixtures, the cell mortality increased more significantly in A549 cells exposed to Fe + Mn than those exposed to Fe or Mn alone (Fig. 4A). The same tendency was also obtained in the treatments of Fe + Cu, Mn + Zn, Mn + Cu, Mn + Cr, Mn + Pb or Pb + Cu. Differently, the metal toxicity decreased when cells were exposed to Fe + Pb (85.1 ± 7.4%), Fe + Cr (85.1 ± 5.8%), Zn

304

Y. Yuan et al. / Science of the Total Environment 678 (2019) 301–308

Fig. 1. Real time cell index of A549 exposed to gradient concentrations of atmospheric particulate matter. The Formula of EC50: Y = Bottom − (Top − Bottom) / (1–10^ ((Log EC50 − X) ∗ Hillslope)). Bottom: −3.166e-001, Top: −4.510e-002, Slope: −6.037, EC50 = 107.90 mg/L, r2 = 1.00, p b 0.01.

+ Pb (81.1 ± 6.0%), Zn + Cr (81.1 ± 6.2%), Pb + Cr (67.3 ± 10.6%) or Cu + Cr (90.8 ± 9.2%), comparing with those exposed to only one of the two heavy metals. With respect to the ternary mixtures, we observed that some heavy metals could lower the toxicity of other binary mixtures (Fig. 5). For example, the obtained cell viability of the combination of Fe + Mn + Pb was 69.0 ± 7.6%, higher than the cell viability in the Mn + Pb treatments (45.0 ± 4.8%). Similar effects were observed when Fe was combined with Zn + Pb, Zn + Cu, Zn + Cr, Mn + Pb, Mn + Cu, Mn + Cr, Pb + Cu, or Pb + Cr, and when Zn was combined with Mn + Pb, Mn + Cu, Pb + Cu, Pb + Cr, Fe + Cr, Fe + Cu, or Fe + Pb. Interestingly, Cr and Pb could also reduce the cell mortality caused by other binary mixtures. Distinct effects were often obtained for Cu when co-exposed with Fe + Zn, Fe + Pb, Fe + Cr, Zn + Mn, or Zn + Cr. Likewise, Mn was found to induce greater toxicity when coexposed with all other binary mixtures, except for Pb + Cu. Afterward, cytotoxicity of quaternary mixtures was conducted to explore the influences of two heavy metals on the other four among all the six heavy metals (Fe, Zn, Mn, Pb, Cu and Cr). Exposure to the combination of Fe + Pb and Cu + Cr + Mn + Zn (65.2 ± 7.3%) significantly increased A549 cell viability compared to that of Cu + Cr + Mn + Zn (49.7 ± 6.9%) (Fig. 4B). Such enhancement was also observed when Fe + Cr, Fe + Cu, Fe + Pb or Cu + Cr was combined with the other four metals. In contrary, co-exposure to Pb + Cr, Pb + Cu, Mn + Cr, Mn + Cu, Mn + Pb, Zn + Cr, Zn + Cu, Zn + Pb, Zn + Mn or Fe + Mn

Fig. 2. Single toxicity of heavy metals to A549 cells in vitro. The concentrations of these tested elements were 480 μg/L of Zn, 40.2 μg/L of Cr, 79 μg/L of Mn, 3920 μg/L of Fe, 8.11 μg/L of Ni, 61 μg/L of Cu, 11.55 μg/L of As, 2.84 μg/L of Se, 21.5 μg/L of Sr, 7.27 μg/L of Cd, and 655 μg/L of Pb. The concentration of SRM 1648a was 100 mg/L. Bars represent means ± SD. ***p b 0.001 versus control. *p b 0.05 versus control.

increased the combined toxicity of the other four metals, correspondingly. In addition, the effects of each metal on the remaining five mixtures were further investigated using quinary systems. An obvious decrease in the cell viability was observed in the absence of Pb (\\Pb, 56.5 ± 9.0%) or Fe (\\Fe, 56.1 ± 4.1%), as compared to that of the hexanary mixture (Fig. 6). However, Mn, Cu, Cr or Zn had opposite effects on the cell mortality caused by the other five metals.

3.3. Contribution of heavy metals to PM toxicity The viability of A549 cells after exposure to 100 mg/L PM was 47.7 ± 5.2%, indicating that the toxicity data were acceptable. We then investigated the relative contributions of all the heavy metals in PM (SRM 1648a). The cell viability increased to 60.1% after the removal of Chelex-labile metals (Fig. 6). Thus, the contribution of metals contained in SRM 1648a was estimated to be approximately 22.9 ± 11.5% (calculated as the decreased mortality after metal removal divided by the mortality after PM exposure). Moreover, co-exposure to eleven selected metals resulted in a dramatic decrease in cell viability (56.8 ± 7.1%), indicating that these eleven metals might play a major role in the cell damage and death. Additionally, co-exposure to six toxic metals led to a high mortality of A549 cells (36.5 ± 7.3%), which contributed the majority of the metal-related toxicity (83.8 ± 19.1%, calculated as the mortality after exposure to six toxic metals divided by the mortality after exposure to eleven metals).

Fig. 3. Concentration response curves (CRCs): single effect of six toxic heavy metals on A549.

Y. Yuan et al. / Science of the Total Environment 678 (2019) 301–308

305

Fig. 4. In vitro toxicity of binary heavy metals in A549 lung cells. (A) In vitro toxicity of quaternary and hexanary heavy metals in A549 lung cells. (B) The used concentrations for the six heavy metals were 480 μg/L of Zn, 40.2 μg/L of Cr, 79 μg/L of Mn, 3920 μg/L of Fe, 61 μg/L of Cu and 655 μg/L of Pb. Bars represent means ± SD. ***p b 0.001 versus control. **p b 0.01 versus control. *p b 0.05 versus control.

4. Discussion 4.1. Heavy metals in PM and their toxic effects Numerous studies have analyzed the pollution levels, specific sources and chemical compositions of heavy metals in ambient PM, due to their adverse health effects. It was reported that entire heavy metal fractions could account for approximately 0.3% to ~5.0% of PM mass concentration in the off-line analysis (Qi et al., 2016; Ge et al., 2017; Elhadi et al., 2018; Lai et al., 2018; Liu et al., 2018; Tadros et al., 2018; Vaio et al., 2018; Chen et al., 2019). In this study, the NIST SRM 1648a sample, which contained a relatively high proportion of heavy metals (5.32%), was used to better understand the toxicity of heavy metals in PM and their interactions. Consistent with previous results (Kouassi et al., 2010), the critical toxic concentration of PM (EC50) was approximately 100 mg/L (Fig. 1). At this exposure concentration, the cell viability of A549 increased from 47.7 ± 5.2% (before chelating resin addition) to 60.1 ± 5.8% (after chelating resin addition). Chelex

Fig. 5. In vitro toxicity of ternary heavy metals in A549 lung cells. The used concentrations for the six heavy metals were 480 μg/L of Zn, 40.2 μg/L of Cr, 79 μg/L of Mn, 3920 μg/L of Fe, 61 μg/L of Cu and 655 μg/L of Pb. Bars represent means ± SD.

resin was demonstrated to effectively remove most of the metal ion contaminants without affecting the nonmetallic components (Shafer et al., 2010; Heo et al., 2015). Therefore, the contribution of heavy metals to the toxicity of SRM 1648a was estimated to 22.9 ± 11.5% (excluding the interactions between metals and non-metal components in PM). It should be noted that this contribution might be underrated due to incomplete removal (e.g., Cr, Fe, As). Although relatively low metal contents were detected in PM, our results indicated that the corresponding effects of heavy metals in PM cannot be ignored and underrated. Heo et al. (2015) found that after the removal of metal ions, the production of ROS and TNF-α was reduced to 36% and 23% versus control, respectively. A previous study also reported a good correlation between PM metal components and health endpoints in vivo/vitro, again implying the important contribution of heavy metals to the toxicity of atmospheric particulate matter (Lippmann, 2014). Eleven heavy metals, Zn, Cr, Mn, Fe, Ni, Cu, As, Se, Sr, Cd and Pb, were chosen for further experiments because of their high toxicities or relatively high levels in the NIST SRM 1648a sample. The contents of these eleven elements account for N5.28% of the total PM mass. During 24-h

Fig. 6. Combinational effects of multiple heavy metals on A549 cell viability. –M (i.e., Fe, Mn, Cu, Pb, Cr, Zn): hexanary mixture in the absence of M. The concentrations of these tested elements were 480 μg/L of Zn, 40.2 μg/L of Cr, 79 μg/L of Mn, 3920 μg/L of Fe, 8.11 μg/L of Ni, 61 μg/L of Cu, 11.55 μg/L of As, 2.84 μg/L of Se, 21.5 μg/L of Sr, 7.27 μg/L of Cd, and 655 μg/L of Pb. Bars represent means ± SD. ***p b 0.001 versus control.

306

Y. Yuan et al. / Science of the Total Environment 678 (2019) 301–308

exposure to single-metal, Fe, Mn, Cu, Pb, Cr and Zn significantly inhibited A549 cell viability. Generally, Fe, Mn, Cu and Zn are essential for normal development and body function below their respective thresholds, but they are toxic to humans when overexposure happens. Fe and Cu are thought to be effective catalysts in Fenton reactions, and the formed free radicals are harmful to biomolecules such as protein, lipid and DNA (Jomova and Valko, 2011; Hsu et al., 2018). Although the precise mechanism is not fully understood, toxic effects of Zn were often attributed to Cu deficiency (Plum et al., 2010), while Mn toxicity might be associated with its effects on other essential metals, including Fe, Zn and Cu (Crossgrove and Zheng, 2004). Cr and Pb are typical toxic metals that not only compete for the transport proteins or cofactor binding sites with the essential metals, but also lead to elevated intracellular oxidative stress (Jomova and Valko, 2011). Sun et al. (2016) examined eight heavy metals in coal combustion related PM2.5, and found that Pb, Cr and Cu contributed to PM2.5 toxicity in Caenorhabditis elegans. However, to the best of our knowledge, no attempt has been made to identify the major toxic metals in PM in human cells. It should be noted that other five metals (Ni, As, Se, Sr, and Cd) co-exposure might have a synergistic effect, although there was no adverse effect after exposure to single metal at their proportional doses. Therefore, biological responses induced by single or multiple exposures to these metals need to be further studied in the future. 4.2. Interactions among toxic heavy metals Traditional toxicity studies are commonly focused on testing the effects of single toxicant. However, it does not reflect real-world scenarios in which humans are exposed to multiple chemicals, especially to the atmospheric particles with multi-components. Thus, we ran our experiments on signed permutations of six identified toxic heavy metals, in order to mimic the combination exposure under controlled laboratory conditions. With very few exceptions, manganese (Mn), the ubiquitous element essential for mammal growth and development, aggravated the toxic effects of all other metal combinations. For example, A549 cells treated with Mn + Fe exerted significantly higher cell mortality (34.2 ± 6.7%) than those treated with Fe alone (15.3 ± 5.8%). Identical results were obtained for ternary, quaternary, quinary and hexanary mixtures. This is in line with Lu et al. (2018) who found that Mn + Pb induced more-than-additive (synergistic) effects on the reproduction and lifespan of C. elegans. Chandra et al. (1981) also found that coexposure to Mn and Pb could lead to a lighter brain weight of rat than the exposure to individual metals alone. It has been reported that manganese cation could perturb the bilayer structures of cell membrane and increase its permeability, resulting in increased bioaccumulation of other toxic metals (Suarez et al., 1995; Suwalsky et al., 2010). Significantly increased Fe uptake in rat cells has been reported during coexposure with Mn (Zheng and Zhao, 2001). Moreover, toxic exposure to Mn also disrupts the dynamics of the endoplasmic reticulum and Golgi apparatus, which play vital roles in regulating protein synthesis along the secretory pathway (Towler et al., 2000). Therefore, Mn might increase the toxicity of other metals by affecting the synthesis of proteins (metallothionein) for heavy metal detoxification (Erikson and Aschner, 2002). However, the effects of PM contained Mn on other components and the underlying mechanisms involved must be further examined carefully. Conversely, according to our data observed in vitro, Fe generally played an important role in alleviating cell mortality caused by other metals. Among all multi-exposure treatments containing Fe (31 combinations), only three treatments (Fe + Zn + Mn, Fe + Cu + Cr, Fe + Zn + Mn + Pb) had a significantly higher cell mortalities than the corresponding treatments without Fe addition. Tallkvist et al. (2000) discovered that increased Fe exposure could not only decrease cellular uptake of Fe, Mn and Zn, but also reduce the expression of divalent metal transporter 1 (DMT1), which was considered as a widespread and nonspecific transporter for divalent metals (e.g., Fe, Pb, Zn, Cu, and Mn).

Therefore, competitive inhibition in the presence of Fe may be able to explain the detoxification effects of Fe against other metals in A549 cells. Besides, the essential elements (Zn, Cu and Cr) and the nonessential elements (Pb and Cr) had variable effects on the toxicity of other metals. These phenomena may be interpreted by the competitive or accelerative effects of metals on the accumulation, action, detoxification and efflux processes of other metals. For instance, Zn, an essential element required for many enzymes and proteins, is generally thought to play protective roles against oxidative stress caused by other metals (Kozlowski et al., 2009). However, the induction of metallothionein by Zn was also reported to be associated with increased accumulation of toxic metals (Torra et al., 1995). As an example of non-essential elements, Pb was found experimentally to be able to increase Zn excretion in rats, because of replacing Zn in Zn-containing enzymes or competing with Zn for metallothionein-like proteins (Victery et al., 1982). In contrary, Pb exposure was demonstrated to increase intracellular Cu accumulation, which resulted from an increased Cu uptake and a decreased Cu efflux (Qian et al., 1999). 4.3. Metal mixture toxicity prediction It is obvious that we can hardly measure the toxicity of metal mixtures for each combination. Complex pollution situation, various interactions among pollutants, as well as limited experimental resources together conspired to increase demand for the toxicity prediction of mixtures. The CA and IA models were first used to analyze the multimetal toxicity based on the logistic-fitted concentration response curve (CRCs) of each individual metal in the mixture (Fig. 7A, B). According to our observations (N = 57), we found that both models often overestimated the observed toxicity of metal mixtures. For the CA model, only several predicted values closely matched the observed results, while for the IA model almost all data points located above the 1:1 fitted line. Obviously, one should be careful when using these two models to predict the toxicity of metal components in PM. The CA model assumes that the mixture components act in a similar way, while the IA model proposes that the components act through different mechanisms (Pöch, 1993; Kim et al., 2012). However, neither of these models considers the interacting effect of metal mixture as demonstrated above, and thus the prediction accuracy of the CA and IA models is currently debated (Xu et al., 2011; Meyer et al., 2015). Besides, based on the notion of these two models, the mode of effect should be told apart easily though it is complicated in human cell lines (Rudzok et al., 2010). Hadrup et al. (2013) also discovered that only part of the predicted values (40%–60%) were acceptable by using the CA and IA models in human kidney H295R cells. Moreover, the merged CA and IA model, the integrated addition model (IAM), were also used to predict mixture toxicity as described in previous studies (Altenburger et al., 2005; Olmstead and Leblanc, 2005). Although both similar and dissimilar action modes were considered, this model is limited to components that have no interactions, and most of the predicted effects were dispersed from the observed value (r2 = 0.711, Fig. 7C). To deal with such limitation, Rider and Leblanc (2005) developed an integrated model, the integrated addition and interaction model (IAI), to incorporate mixture interactions using the K-function. Because the K-functions need to be experimental established, which is a drawback of model prediction, we estimated the K value by the ratio of EC50i to the empirical EC50 of the ith component when co-exposure with other components. After calculation, 57 combinational treatments were well fitted by the IAI model, with a high r2 value (r2 = 0.87, p b 0.01) after a linear regression analysis (Fig. 7D). A comparison between the IAM and the IAI model indicated that the latter might be more reasonable to estimate the modes of action of metal mixtures. Moreover, these results demonstrate again that there are considerable interactions among the six heavy metals as discussed earlier. The IAI model was initially utilized in pesticide toxicity prediction and obtained a good fit with the r2 value of 0.72 in D. magna (Rider and Leblanc, 2005).

Y. Yuan et al. / Science of the Total Environment 678 (2019) 301–308

307

Fig. 7. Evaluation of prediction models for A549 cell sensitivity. (A) The concentration addition model, CA. (B) the independent action model, IA. (C) the integrated addition model, IAI. (D) the integrated addition and interaction model, IAM. Mix (n) indicates the multi-exposure treatments (n = 2–6). The used concentrations for the six heavy metals were 480 μg/L of Zn, 40.2 μg/L of Cr, 79 μg/L of Mn, 3920 μg/L of Fe, 61 μg/L of Cu and 655 μg/L of Pb. Total points: N = 57.

However, this model has not been widely used due to the limited data of K-function, especially in human cell models. 5. Conclusion In vitro cytotoxicity of metals contained in PM was assessed and predicted in human A549 cells. With the aid of chelating resin, A549 cell viability increased by 12.3% at the critical toxic concentration of PM (EC50), indicating an important role of metals in PM-induced adverse health effects. Among 11 tested elements (Zn, Cr, Mn, Fe, Ni, Cu, As, Se, Sr, Cd, Pb), Zn, Cr, Mn, Fe, Cu and Pb significantly suppressed the cell viability of A549, and their calculated EC50 values followed the order of Fe N Cr N Mn N Zn ≥ Cu N Pb. Generally, Mn increased the toxic effects of other metal mixtures, Fe could lighten the toxicity caused by other metals, while Zn, Cu, Pb and Cr exerted opposite effects on different metal combinations. These metal-metal interactions may be attributed to their similar or dissimilar membrane transporters and intracellular targets. Moreover, a good linear regression was observed between observed and predicted toxic effects by the IAI model. Although other models performed below expectation, their applications need to be further evaluated for more complex metal combinations. Acknowledgments This work was supported by the Natural Science Foundation of China [grant number 91543115, 91544220 and 21577065]. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.scitotenv.2019.04.431.

References Altenburger, R., et al., 2005. Algal toxicity of nitrobenzenes: combined effect analysis as a pharmacological probe for similar modes of interaction. Environ. Toxicol. Chem. 24 (2), 324–333. Badaloni, C., et al., 2017. Effects of long-term exposure to particulate matter and metal components on mortality in the Rome longitudinal study. Environ. Int. 109, 146. Bliss, C.I., 1939. The toxicity of poisons applied jointly. Ann Appl Biol 26, 586–615. Chandra, A.V., et al., 1981. Behavioral and neurochemical changes in rats simultaneously exposed to manganese and lead. Arch. Toxicol. 49 (1), 49. Chen, Y., et al., 2019. Chemical characteristics of PM2.5 and water-soluble organic nitrogen in Yangzhou, China. Atmosphere 10, 178. Crossgrove, J., Zheng, W., 2004. Manganese toxicity upon overexposure. NMR Biomed. 17 (8), 544–553. Elhadi, R.E., et al., 2018. Seasonal variations of atmospheric particulate matter and its content of heavy metals in Klang Valley, Malaysia. Aerosol Air Qual. Res. 18 (5), 1148–1161. Erikson, K., Aschner, M., 2002. Manganese causes differential regulation of glutamate transporter (GLAST) taurine transporter and metallothionein in cultured rat astrocytes. Neurotoxicology 23 (4), 595–602. Ge, X., et al., 2011. Atmospheric amines – part I. A review. Atmos. Environ. 45 (3), 524–546. Ge, X., et al., 2017. Aerosol characteristics and sources in Yangzhou, China resolved by offline aerosol mass spectrometry and other techniques. Environ. Pollut. 225, 74–85. Gurer, H., Ercal, N., 2000. Can antioxidants be beneficial in the treatment of lead poisoning? Free. Radical. Bio. Med. 29 (10), 927–945. Hadrup, N., et al., 2013. Concentration addition, independent action and generalized concentration addition models for mixture effect prediction of sex hormone synthesis in vitro. PLoS One 8 (8), e70490. Heo, J., et al., 2015. Assessing the role of chemical components in cellular responses to atmospheric particle matter (PM) through chemical fractionation of PM extracts. Anal. Bioanal. Chem. 407 (20), 5953–5963. Hsu, H.W., et al., 2018. Environmental and dietary exposure to copper and its cellular mechanisms linking to Alzheimer's disease. Toxicol. Sci. 163 (2), 338–345. Jia, Y.Y., et al., 2017. Toxicity research of PM2.5 compositions in vitro. Int. J. Env. Res. Pub. He. 14 (3), 232. Jomova, K., Valko, M., 2011. Advances in metal-induced oxidative stress and human disease. Toxicology 283 (2), 65–87. Kim, J., et al., 2012. Reliable predictive computational toxicology methods for mixture toxicity: toward the development of innovative integrated models for environmental risk assessment. Rev. Environ. Sci. Bio. 12 (3), 235–256.

308

Y. Yuan et al. / Science of the Total Environment 678 (2019) 301–308

Kouassi, K.S., et al., 2010. Oxidative damage induced in A549 cells by physically and chemically characterized air particulate matter (PM2.5) collected in Abidjan, Côte d'ivoire. J. Appl. Toxicol. 30 (4), 310–320. Kozlowski, H., et al., 2009. Copper, iron, and zinc ions homeostasis and their role in neurodegenerative disorders (metal uptake, transport, distribution and regulation). Coordin. Chem. Rev. 253 (21), 2665–2685. Lai, C.H., et al., 2018. Effects of heavy metals on health risk and characteristic in surrounding atmosphere of tire manufacturing plant, Taiwan. RSC Adv. 8 (6), 3041–3050. Lippmann, M., 2014. Toxicological and epidemiological studies of cardiovascular effects of ambient air fine particulate matter (PM2.5) and its chemical components: coherence and public health implications. Crit. Rev. Toxicol. 44 (4), 299–347. Lippmann, M., Chen, L.C., 2009. Health effects of concentrated ambient air particulate matter (CAPs) and its components. Crit. Rev. Toxicol. 39 (10), 865–913. Liu, K., et al., 2018. Sources and health risks of heavy metals in PM2.5 in a campus in a typical suburb area of Taiyuan, north China. Atmosphere (Basel) 9 (2), 46. Loewe, S., Muischnek, H., 1926. Über Kombinationswirkungen I. Mitteilung: Hilfsmittel der Fragestellung. Naunyn-Schmiedeberg's Arch. Exp. Pathol. Pharmakol. 114, 313–326. Lu, C., et al., 2018. Toxicity interactions between manganese (Mn) and lead (Pb) or cadmium (Cd) in a model organism the nematode C. elegans. Environ. Sci. Pollut. R. 25 (16), 15378–15389. Meyer, J.S., et al., 2015. Metal mixtures modeling evaluation project: 1. Background. Environ. Toxicol. Chem. 34 (4), 726–740. Murphy, S.M., et al., 2016. Optimization of an in vitro bioassay to monitor growth and formation of myotubes in real time. Biosci. Rep. 36 (6), e00330. Mwamba, T.M., et al., 2016. Interactive effects of cadmium and copper on metal accumulation, oxidative stress, and mineral composition in Brassica napus. Int. J. Environ. Sci. Te. 13 (9), 2163–2174. Nie, D., et al., 2018. Characterization of fine particulate matter and associated health burden in Nanjing. Int. J. Env. Res. Pub. He. 15 (4), 602. Olmstead, A.W., Leblanc, G.A., 2004. Toxicity assessment of environmentally relevant pollutant mixtures using a heuristic model. Integr. Environ. Asses. 1 (2), 114–122. Olmstead, A.W., Leblanc, G.A., 2005. Joint action of polycyclic aromatic hydrocarbons: predictive modeling of sublethal toxicity. Aquat. Toxicol. 75 (3), 253–262. Plackett, R.L., Hewlett, P.S., 1952. Quantal responses to mixtures of poisons. J. R. Stat. Soc. Ser. B Methodol. 14, 141–163. Plum, L.M., et al., 2010. The essential toxin: impact of zinc on human health. Int. J. Env. Res. Pub. He. 7 (4), 1342–1365. Pöch, G., 1993. Combined Effects of Drugs and Toxic Agents: Modern Evaluation in Theories and Practice. Springer-Verlag, Wien and New York. Qi, L., et al., 2016. Source identification of trace elements in the atmosphere during the second Asian Youth Games in Nanjing, China: influence of control measures on air quality. Atmos. Pollut. Res. 7, 547–556. Qian, Y., et al., 1999. Effect of lead exposure and accumulation on copper homeostasis in cultured C6 rat glioma cells. Toxicol. Appl. Pharm. 158 (1), 41. Rider, C.V., Leblanc, G.A., 2005. An integrated addition and interaction model for assessing toxicity of chemical mixtures. Toxicol. Sci. 87 (2), 520–528.

Rudzok, S., et al., 2010. Measuring and modeling of binary mixture effects of pharmaceuticals and nickel on cell viability/cytotoxicity in the human hepatoma derived cell line HepG2. Toxicol. Appl. Pharm. 244 (3), 336–343. Shafer, M.M., et al., 2010. Reactive oxygen species activity and chemical speciation of sizefractionated atmospheric particulate matter from Lahore, Pakistan: an important role for transition metals. J. Environ. Monitor. 12 (3), 704–715. Suarez, N., et al., 1995. Cellular neurotoxicity of trivalent manganese bound to transferrin or pyrophosphate studied in human neuroblastoma (SH-SY5Y) cell cultures. Toxicol. in Vitro 9 (5), 717–721. Suh, H.H., et al., 2011. Chemical properties of air pollutants and cause-specific hospital admissions among the elderly in Atlanta, Georgia. Environ. Health. Persp. 119 (10), 1421–1428. Sun, L., et al., 2016. Contribution of heavy metals to toxicity of coal combustion related fine particulate matter (PM2.5) in Caenorhabditis elegans with wild-type or susceptible genetic background. Chemosphere 144, 2392–2400. Suwalsky, M., et al., 2010. Mn2+ exerts stronger structural effects than the Mn-citrate complex on the human erythrocyte membrane and molecular models. J. Inorg. Biochem. 104(1), 55–61. Tadros, C.V., et al., 2018. Chemical characterisation and source identification of atmospheric aerosols in the Snowy Mountains, south-eastern Australia. Sci. Total Environ. 630, 432–443. Tallkvist, J., et al., 2000. Functional and molecular responses of human intestinal Caco-2 cells to iron treatment. Am. J. Clin. Nutr. 72 (3), 770–775. Torra, M., et al., 1995. Cadmium and zinc relationships in the liver and kidney of humans exposed to environmental cadmium. Sci. Total Environ. 170 (1–2), 53–57. Towler, M.C., et al., 2000. The manganese cation disrupts membrane dynamics along the secretory pathway. Exp. Cell Res. 259 (1), 167–179. Uski, O., et al., 2015. Effect of fuel zinc content on toxicological responses of particulate matter from pellet combustion in vitro. Sci. Total Environ. 511, 331–340. Vaio, P.D., et al., 2018. Heavy metals size distribution in PM10 and environmental-sanitary risk analysis in Acerra (Italy). Atmosphere (Basel) 9 (2), 58. Victery, W., et al., 1982. Lead increases urinary zinc excretion in rats. Biol. Trace Elem. Res. 4 (2–3), 211. Wang, Y., et al., 2018. Assessing the cytotoxicity of ambient particulate matter (PM) using Chinese hamster ovary (CHO) cells and its relationship with the PM chemical composition and oxidative potential. Atmos. Environ. 179, 132–141. Wu, J., et al., 2012. In vitro cytotoxicity of Cu2+, Zn2+, Ag+ and their mixtures on primary human endometrial epithelial cells. Contraception. 85(5), 509–518. Wu, X., et al., 2016. A review of toxicity and mechanisms of individual and mixtures of heavy metals in the environment. Environ. Sci. Pollut. R. 23 (9), 8244–8259. Xu, X., et al., 2011. Assessment of toxic interactions of heavy metals in multi-component mixtures using sea urchin embryo-larval bioassay. Toxicol. in Vitro 25 (1), 294–300. Zheng, W., Zhao, Q., 2001. Iron overload following manganese exposure in cultured neuronal, but not neuroglial cells. Brain Res. 897 (1), 175–179. Zou, Y., et al., 2016. Water soluble and insoluble components of urban PM2.5 and their cytotoxic effects on epithelial cells (A549) in vitro. Environ. Pollut. 212, 627–635.