Activation of human AML14.3D10 eosinophils by nanoparticles: Modulatory activity on apoptosis and cytokine production

Abstract

Eosinophilic inflammation is frequently observed in response to nanoparticle (NP) exposure in airway rodent models of allergies where the number of eosinophils is increased in lungs. Despite this, it is surprising that the potential cytotoxic effect of NP, as well as their direct role on eosinophils is poorly documented. The present study investigated how different NP can alter the biology of the human eosinophilic cell line AML14.3D10. It was found that among NP forms of CeO₂, ZnO, TiO₂, and nanosilver of 20 nm (AgNP₂₀) or 70 nm (AgNP₇₀) diameters, only ZnO and AgNP₂₀ induced apoptosis. Caspases-7 and -9 were not activated by the tested NP while caspase-3 was activated by AgNP₂₀ only. However, both ZnO and AgNP₂₀ induced cytoskeletal breakdown as evidenced by the cleavage of lamin B₁. Using an ELISArray approach for the simultaneous detection of several analytes (cytokines/chemokines), it was found that only ZnO and AgNP₂₀ increased the production of different analytes including the potent pro-inflammatory CXCL8 (IL-8) chemokine. From the data here, we conclude that toxic effects of some NP could be observed in human eosinophil-like cells and that this could be related, at least partially, by induction of apoptosis and production of cytokines and chemokines involved in inflammation. The results of this study also indicate that distinct NP do not activate similarly human eosinophils, since ZnO and AgNP₂₀ induce apoptosis and cytokine production while others such as TiO₂, CeO₂, and AgNP₇₀ do not.

Keywords:

• Apoptosis
• cytokines
• cytoskeleton
• eosinophils
• nanoparticles

Introduction

The use of engineered nanoparticles (NP) in a variety of sectors, including electronics, biomedical sciences, pharmaceutical industry, cosmetics, etc., is in rapid expansion. Consequently, the probability of human exposure to NP has risen (Oberdorster et al. 2005). Because NP could enter human body via inhalation, ingestion, and dermal exposure, they can reach the blood circulation and come in close contact with immune cells. Therefore, it is crucial to evaluate the risk that NP may represent to humans.

One of the major toxic effects observed following NP exposure in a variety of in vitro and in vivo models is inflammation, based especially on two criteria: (i) an increased production of cytokines in in vitro and in vivo models of inflammation; and (ii) an increased number of granulocytes observed in lungs and/or bronchoalveolar lavages following administration of NP via inhalation, intratracheal instillation, nebulization, etc. (Chen et al. 2006, 2011; Cho et al. 2010; Rossi et al. 2010; Srinivas et al. 2011; Sahu et al. 2013). Of note, an increase number of granulocyte eosinophils are especially observed in response to NP exposure in airway rodent models of allergy (Cho et al. 2010; Larsen et al. 2010; Brandenberger et al. 2013; Roy et al. 2013).

Despite the fact that high number of neutrophils and/or eosinophils is observed in several studies following NP exposure, the role of NP in the biology of human neutrophils and eosinophil is poorly understood. However, in the past few years, direct interactions between NP and human neutrophils have been investigated, leading to the observation that several neutrophil responses can be altered in response to NP exposure. These include respiratory burst, degranulation, cytokine production, and apoptosis (Goncalves et al. 2010; Abrikossova et al. 2012; Babin et al. 2013; Poirier et al. 2014). Unlike neutrophils, there is presently no study reporting the direct roles of NP on human eosinophils. Accordingly, the present study utilized the human eosinophilic cell line AML14.3D10, a good model for studying eosinophil cell biology or at least to obtain information on cancerous cells of eosinophilic origin (Baumann & Paul 1998), to investigate effects of a panel of well-characterized NP. Among the NP agents analyzed were titanium dioxide (TiO₂), cerium oxide (CeO₂), zinc oxide (ZnO), as well as nanosilver of initial diameter of 20 nm (AgNP₂₀) or 70 nm (AgNP₇₀) for their effects on apoptosis and cytokine production by the AML14.3D10 cells.

Materials and methods

Nanoparticles, chemicals, agonists, and antibodies

TiO₂, CeO₂, and ZnO NP were obtained from Vivenano (Toronto, ONT, Canada). TiO₂ anatase crystals, in aqueous suspension stabilized by polyacrylate sodium (≤ 10-nm particle size, 90%); metal content: 97% TiO₂ determined by inductively coupled plasma (ICP); cerium oxide (CeO₂) water-dispersible and non-aggregating (crystalline CeO₂) in aqueous suspension; carboxy-functionalized and stabilized by sodium polyacrylate, ≤ 10 nm particle size, 90%) with a metal content of 90% Ce/10% Pr as determined by (ICP) and ZnO water-dispersible and non-aggregating ZnO (crystalline ZnO) carboxy-functionalized in aqueous suspension stabilized by sodium polyacrylate (≤ 20 nm particle size, 90%) with metal content of 90% Zn as determined by (ICP).

RPMI 1640, HEPES, and penicillin/streptomycin (P/S) were purchased from Gibco BRL (Grand Island, NY). Bovine serum albumin (BSA), Viscum album agglutinin-1 (VAA-I), and monoclonal antibodies to human cytoskeletal proteins (anti-gelsolin antibody, clone GS-2C4 and anti-vinculin, clone Vin 11-5) were purchased from Sigma (St. Louis, MO). The silver NP of 20 nm or 70 nm of diameter, i.e., AgNP₂₀ and AgNP₇₀, (Pelco^® citrate Biopure™ silver) were purchased from Ted Pella (Redding, CA). Anti-vimentin (clone V9), polyclonal anti-lamin B₁, and rabbit polyclonal anti-GAPDH antibody were from Santa Cruz Biotechnologies (Santa Cruz, CA). Horseradish peroxidase (HRP)-conjugated goat anti-mouse and goat anti-rabbit antibodies were purchased from Jackson Immunoresearch (West Grove, PA). Mouse mAb anti-caspase-3 (3G2), anti-caspase-7 (C7), and anti-capase-9 (C9) were purchased from Cell Signaling (Danvers, MA).

Size distribution and zeta potential measurements

The size distribution and the surface charge (zeta potential) of NP were determined by dynamic light scattering using a Malvern Zetasizer Nano-ZS (model ZEN3600 from Malvern Instruments Inc. (Westborough, MA). Measurements were performed in the same experimental conditions that NP were used in vitro: at 37 °C and at a concentration of 100 μg NP/ml in RPMI-1640 medium supplemented with 10% fetal calf serum and 1% penicillin– streptomycin (Gibco) with 1 mM sodium pyruvate. Before performing the analysis, the sterility of all NP used was checked.

Since the classical Limulus amebocyte lysate assay used to determine endotoxin level is a method in which NP could interfere (Neun & Dobrovolskaia 2011), sterility was instead tested as previously published (Goncalves & Girard 2013). In brief, NP suspensions were incubated in Lysogeny broth agar plates for 48 h and the plates then observed for the presence or absence of colonies. As a positive control, a suspension of intentionally contaminated TiO₂ NP (strictly reserved to perform this kind of analysis) was used. In contrast to the positive control, all the NP used in this study appears to be negative for endotoxin in the limit of this approach, since no colonies were observed (data not shown).

Cell line

Human eosinophilic AML14.3D10 cells (referred to as 3D10 cells hereafter) were generated and kindly provided by Drs. C. C. Paul and M. Baumann (Wright State University, Dayton, OH) (Baumann & Paul 1998). The non-adherent 3D10 cells were cultured at 37 °C in a 5% CO₂ atmosphere in the same buffer as above. Cells were cultivated at a density between 2 × 10⁵ and 1 × 10⁶/ml, and the medium was changed three times per week. Before performing the different experiments, the cell viability was systematically verified by trypan blue exclusion and was found to routinely be >98%.

Cell viability and apoptosis determination

3D10 cells were incubated for up to 24 h in the presence or absence of 100 μg/ml of TiO₂, CeO₂, ZnO, AgNP₂₀, or AgNP₇₀ and cell viability was then determined by trypan blue exclusion at 6, 12, 18, and 24 h. The concentration of 100 μg/ml was selected based on previous studies performed with human granulocytes (Goncalves et al. 2010; Babin et al. 2013; Goncalves & Girard 2014; Poirier et al. 2014, 2016). For apoptosis, cells were incubated for 24 h and apoptosis then evaluated by flow cytometry by counting the numbers of FITC-Annexin-V-positive cells. In some experiments, apoptosis was confirmed by cytology from cyto-centrifuged sample preparations of cells using a Cytotek^® centrifuge (Miles Scientific, Naperville, IL) and processed as documented previously (Lavastre et al. 2005). Cells were examined by light microscopy at 400× final magnification.

Caspase activation et cytoskeletal protein expression

Cells (10⁶ cells/ml) were incubated in the presence or absence of buffer, VAA-I, or the test NP for 24 h and the expression of caspase and cytoskeletal proteins were then evaluated by Western blot as published previously (Lavastre et al. 2002). In brief, cells were harvested for the preparation of cell lysates in 1× Laemmli sample buffer. Aliquots corresponding to 10⁶ cells were loaded onto 10% sodium dodecyl sulfate-polyacrylamide gels, electrophoresed, and then their proteins electrotransferred to polyvinylidene difluoride membranes. Non-specific binding sites on the membranes were blocked with blocking solution (5% BSA in Tris-buffered saline-Tween [25 mM Tris-HCl (pH 7.8), 190 mM NaCl, 0.15% Tween-20]) for 1 h at room temperature. Membranes were then washed and incubated with anti-human pro-caspase or cyto-skeletal antibodies (all at 1:1000) overnight at 4 °C. After several washes with Tris-buffered saline-0.1% Tween 20, the membranes were incubated with HRP-labeled goat anti-mouse IgG antibodies (1:20 000) or HRP-labeled goat anti-rabbit IgG antibody (1:20 000) for 1 h at room temperature in blocking solution. At the end of the experiments, the membranes were stripped and re-probed with anti-GAPDH antibody (1:500) to verify equal protein loading. In all cases, bands were revealed with an enhanced chemiluminescence-Western blotting detection system and assessed in a Chemidoc^TM MP imaging system (both Bio-Rad, Hercules, CA).

Cytokine and chemokine detection

Supernatants from four different batches of 3D10 cells treated separately were harvested and pooled for the simultaneous detection of several cytokines and chemokines using two Multi-Analyte ELISArrays (Qiagen Inc.-Canada, Toronto, Canada). The first kit allowed for simultaneous detection of cytokines and chemokines including interleukin (IL)-1α, IL-1β, IL-2, IL-4, IL-6, IL-8, IL-10, IL-12, IL-17A, interferon (IFN)-γ, tumor necrosis factor (TNF)-α, and GM-CSF. The second kit allowed for detection of chemokines IL-8, MCP-1, RANTES, MIP-1α, MIP-1β, IP-10, I-TAC, MIG, Eotaxin, TARC, MDC, and GR α. Only IL-8 was common to both kits. The measurement of IL-8 production was also determined using a commercially available ELISA kit (Invitrogen Canada Inc., Toronto, Canada) as per the instructions of the manufacturer. The level of sensitivity of the kit was < 5 pg IL-8/ml.

Statistical analysis

A one-way analyses of variance (ANOVA) followed by Dunnett's test was used to evaluate for significance differences between controls and each sample. Statistical analyses were performed using Prism for Windows (v5.0, GraphPad Software, San Diego, CA). Statistical significance was set at p < 0.05.

Results

Characterization of NP

As indicated in Table 1, the size distribution of AgNP₂₀ and AgNP₇₀ was higher than the starting material with a proportion of ≈100% of the NP suspensions with a mean diameter of 158.9 ± 6.2 nm and 113.2 ± 2.6% nm, respectively. The presence of large aggregates was noted for CeO₂, ZnO, and TiO₂ NP suspensions. The zeta potential varies among the NP where the values (mean ± SEM, n = 3) were −10.1 ± 0.5, −9.2 ± 1.2, −11.0 ± 0.5, 0.01 ± 1.20, and −12.1 ± 0.8 mV for AgNP₂₀, AgNP₇₀, CeO₂, ZnO, and TiO₂, respectively.

Table 1. Characterization of the NP used in this study^a.

	AgNP20	AgNP70	CeO2	ZnO	TiO2
Size (nm)^b (% intensity)^c	158.9 ± 6.2 (99.2 ± 0.7)	113.7 ± 2.6 (100)	29.9 ± 18.6 (70.0 ± 27.5)	53.4 ± 4.6 (50.1 ± 3.1)	95.2 ± 13.1 (66.9 ± 1.3)
	1482 ± 2557 (0.7 ± 0.7)		1416 ± 2400 (33.1 ± 27.5)	11.9 ± 0.3 (42.0 ± 0.8)	11.8 ± 0.9 (31.6 ± 2.2)
			3247 ± 2814 (1.0 ± 0.9)	2907 ± 658 (7.9 ± 2.4)	2483 ± 2164 (1.6 ± 1.7)
Zeta potential^a	−10.1 ± 0.5	−9.2 ± 1.2	−11.0 ± 0.5	0.01 ± 1.2	−12.1 ± 0.8

a Measurements were performed in RPMI 1640 medium supplemented with 10% fetal calf serum, 1% penicillin–streptomycin and with 1 mM sodium pyruvate, at 37 °C.

b Data are from three different lectures and results shown are means ± SEM.

c Intensity represents the proportion (%) of each peak.

Evaluation of cell viability and apoptosis in human 3D10 eosinophil cells in response to NP

Figure 1(A) illustrates that none of the tested NP induced necrosis in > ≈17% of cells after 6, 12, 18, or 24 h as well as the potent pro-apoptotic plant lectin VAA-I (VAA). Thus, the tested NP did not appear more cytotoxic than VAA-I. Because apoptotic cells exclude trypan blue and are, therefore, considered viable, apoptosis was next determined. As illustrated in Figure 1(B), ZnO and AgNP₂₀ were found to significantly increase apoptosis from 5.7 ± 0.7% (mean ± SEM, n = 4 different cell batches) to 11.5 ± 0.5% and to 15.9 ± 0.6%, respectively, according to the determination of FITC-Annexin-V⁺ cells. Of note, VAA-I also significantly increased apoptosis to a level of 54.9 ± 1.3%. These results were confirmed by cytology in two separate experiments where, among the tested NP, only ZnO and AgNP₂₀ were found to induce apoptosis (Figure 1(C)).

Figure 1. Evaluation of cell viability and apoptosis in human 3D10 eosinophil cells in response to nanoparticles. 3D10 eosinophil cells (1 × 10⁶ cells/ml, 2 ml/well) were incubated in the presence of buffer (Ctrl), 1 μg/ml of the pro-apoptotic plant lectin VAA-I (VAA) or 100 μg/ml of the indicated NP for 6-24 h (A) or 24 h (B and C) and cell viability (A) was assessed by trypan blue exclusion assay and apoptosis was evaluated by flow cytometry (B) or cytology (C). (A) Two independent kinetic experiments are illustrated. (B) Results are means ± SEM (n = 4). (C) Results are from two separate experiments (Exp. 1 and Exp. 2). Inset, typical images illustrating that most of the cells are normal in Ctrl when compared with incubation with VAA where the vast majority of cells are apoptotic (arrows). *p < 0.05 versus control (Ctrl).

Caspase activation in NP-treated 3D10 cells

The study next determined whether or not both ZnO and AgNP₂₀ could induce processing of pro-caspases-3, -7, and -9 in comparison with VAA-I known to induce apoptosis in these cells by a caspase-dependent mechanism (Lavastre et al. 2005). In contrast to VAA-I that markedly induced pro-caspase-3 processing, AgNP₂₀ but not ZnO slightly-to-moderately activated caspase-3 (Figure 2). The other NP tested did not alter the basal level of pro-caspase-3. Similar results were observed for casp-7, a closely caspase-3-related caspase. Caspase-9, highly expressed in 3D10 cells, was found to be activated by VAA-I as evidenced by its cleavage while ZnO and AgNP₂₀ did not, or minimally activate it, respectively.

Figure 2. Effect of NP on caspase activation. (A) 3D10 eosinophil cells (1 × 10⁶ cells/ml; 2 ml/well) were incubated in the presence of buffer (ctrl), 1 μg/ml VAA-I (VAA) or with 100 μg/ml of the indicated NP for 24 h and expression of pro-caspase-3, 7 or 9 was then assessed by Western blot (see Materials and methods section). Membranes were then stripped and re-probed with an anti-GAPDH antibody to verify protein loading. Results shown are from one representative experiment out of at least three. (B) Corresponding densitometry.

ZnO and AgNP₂₀ induce degradation of cytoskeletal lamin B₁ protein in 3D10 cells

Knowing that ZnO and AgNP₂₀ induced apoptosis but only minimally caspases, the study attempted to verify if these NP would alter the cytoskeleton, expecting they would not. Unexpectedly, both ZnO and AgNP₂₀, but not the other NP, were found to cleave lamin B₁ onto the characteristic p46 fragment (Figure 3). As expected, VAA-I was found to markedly cleave lamin B₁ to the p46 fragment. However, unlike VAA-I, none of the tested NP were found to induce degradation of vimentin. Gelsolin was only minimally cleaved by ZnO when compared to VAA-I. Vinculin was also not cleaved by any of the test NP.

Figure 3. Expression of cytoskeletal proteins in response to NP. (A) 3D10 eosinophil cells (1 × 10⁶ cells/ml; 2 ml/well) were incubated for 24 h in the presence of buffer (Ctrl), 1 μg/ml VAA-I (VAA), or with 100 μg/ml of the indicated NP and the expression of vimentin, lamin B₁, gelsolin, and vinculin was then determined by Western blot analyses. Membranes were then stripped and re-probed with an anti-GAPDH antibody to verify protein loading. Results shown are from one representative experiment out of at least two. (B) Corresponding densitometry.

Production of cytokines and chemokines in NP-treated 3D10 cells

Figure 4(A) illustrates the results obtained (with first kit) that showed that production of different cytokines was increased when compared with controls only with ZnO or AgNP₂₀. More precisely, ZnO increased 3D10 cell production of IL-8, MCP-4, RANTES, MIP-1α, MIP-1β, IP-10, and ITAC while AgNP₂₀ also increased, but to a lesser extent, production of IL-8, RANTES, MIP-1α, MIP-1β, and IP-10, but not of MCP-4, IP-10, or ITAC. The second kit allowed detection of other analytes, but IL-8 was also included in this kit and the results testified to reproducibility of the technique (see IL-8 in Figure 4(A) and (B)). Of note, only the production of TNFα was increased by treatment with ZnO or AgNP₂₀. Production of all other analytes evaluated was not affected by any of the test NP.

Figure 4. Cytokine and chemokine production in NP-induced eosinophils. 3D10 eosinophil cells (1 × 10⁶ cells/ml; 2 ml/well) were incubated for 24 h in the presence of buffer or with 100 μg/ml of the indicated NP and the supernatants from different cell batches were then pooled and used for the antibody array assay to detect production of indicated cytokines/chemokines. CTRL^-and CTRL⁺ are, respectively, technical negative and positive controls included in the kit.

ZnO and AgNP₂₀ increase the production of IL-8

To confirm the results obtained using the ELISarrays, and since IL-8 is an important pro-inflammatory molecule, amounts of IL-8 released in the supernatants of NP-treated 3D10 cells were evaluated by ELISA. As shown in Figure 5, only ZnO and AgNP₂₀ significantly increased the production of IL-8 from 18.5 ± 7.5 (mean ± SEM, n = 4) for control cells to 2084 ± 245 or 252 ± 47 pg/ml, respectively, for cells treated with ZnO or AgNP₂₀. Treatment with the other NP never gave rise to levels that exceeded 27 pg/ml.

Figure 5. ZnO and AgNP₂₀ induce IL-8 production in eosinophils. 3D10 eosinophil cells (1 × 10⁶ cells/ml; 2 ml/well) were incubated in the presence of buffer (Ctrl), 1 μg/ml VAA-I (VAA) or with 100 μg/ml of the indicated NP and supernatants were tested by ELISA to quantify the production of IL-8. *p < 0.05 versus control (Ctrl).

Discussion

The results of the present study indicated that the biology of human eosinophil-like 3D10 cells could be altered by NP. To the best of our knowledge, this is the first study investigating interactions between NP and human cells of eosinophilic origin. Accumulation of eosinophils is a common feature of allergic airway inflammation and there are evidences for a role of these cells in pulmonary diseases, including asthma. In addition, some NP were found to exert some pro-inflammatory activity in airway diseases in vivo (Larsen et al., 2010; Brandenberger et al., 2013). Here, among the tested NP, it was interesting to note that only ZnO and AgNP₂₀ were found to activate these cells. Therefore, NP does not necessarily act similarly on eosinophils and, according to the present data, induction of apoptosis is a potential mechanism of toxicity of ZnO and AgNP₂₀. The presence of large aggregates in AgNP₂₀ is not without any precedent since, as previously documented for human neutrophils (Poirier et al. 2014), and as for these cells, we cannot rule out here the possibility that such aggregates can induce the observed effect in 3D10 cells. Interestingly, we recently documented that AgNP₇₀ and AgNP₂₀ alter differently the biology of human neutrophils (Poirier et al. 2016). The fact that AgNP₂₀ were more potent than AgNP₇₀ could be related to their ability to penetrate more easily inside cells. Therefore, it is tempting to speculate here that because AgNP₂₀ possess more biological activity in 3D10 cells than AgNP₇₀, this could also be linked to differences in cellular nanoparticle uptake. This remains to be determined.

Of note, in the present study, 3D10 cells were incubated with NP for only 24 h. This was not like several studies where cells or cell lines were incubated with NP for up to 72 h and even longer in some cases (Zhang et al. 2011; Al-Qubaisi et al. 2013; Chuang et al. 2013; Hou et al. 2013; Kimura et al. 2013; Tsai et al. 2013). Nevertheless, in experimental similar conditions to those used in the present study (100 μg/ml for 24 h), TiO₂ NP were found to be anti-apoptotic in human neutrophils (Goncalves et al. 2010), whereas AgNP₂₀ were rather pro-apoptotic (Poirier et al. 2014). Therefore, a given NP can act similarly on different cell types (e.g., AgNP₂₀) or not (e.g., TiO₂). Several pro-apoptotic NP, including those investigated here, are known to activate caspases in different cell types (Kang et al. 2009; Lanone & Boczkowski 2011; Guo et al. 2013).

In contrast, in the present study, the tested caspases were not, or only minimally, activated by ZnO and AgNP₂₀. Despite this, lamin B₁ was cleaved by both ZnO and AgNP₂₀. These results are in agreement with our previous observation that the pro-apoptotic molecule VAA-I cleaved several cytoskeletal proteins, including lamin B₁ in 3D10 cells (Lavastre et al. 2005). However, the ability of VAA-I to cleave lamin B₁ was found to be a caspase-dependent process since pre-treatment of cells with the pan-caspase inhibitor z-VAD, completely reversed the degradation/cleavage of lamin B₁ by VAA-I. Therefore, it is tempting to speculate that ZnO and AgNP₂₀ induce lamin B₁ degradation by proteases other than caspases. For example, calpains were recently shown to be involved in eosinophil apoptosis. In primary human eosinophils, lamin B degradation was found to be a caspase-3- and 6-dependent process (Ilmarinen-Salo et al. 2010). We did not investigate here the potential role of caspase-6 in ZnO- or AgNP₂₀-induced lamin B₁ degradation in leukemia eosinophils, since, unlike primary human eosinophils (Ilmarinen-Salo et al. 2010), there is no evidence yet that 3D10 cells express this caspase.

Interestingly, among the tested NP, ZnO and AgNP₂₀ were not only the two NP found to induce apoptosis but also the only ones that increased cytokine production in 3D10 cells. However, it is difficult to make a link between the pro-apoptotic activity of ZnO and AgNP₂₀ with their ability to increase cytokine production. For example, the fact that among the tested analytes, only TNFα is known to mediate apoptosis in a variety of cells. In addition, the TNFα production is increased only by ZnO and not by AgNP₂₀ that are both pro-apoptotic in 3D10 cells. Also, it was previously demonstrated that addition of TNFα to IL-15 in human eosinophil cultures did not promote or accelerate their rate of apoptosis (Hoontrakoon et al. 2002) and, moreover, TNFα was found to be anti-apoptotic by itself in these cells (Tsukahara et al. 1999). Although ZnO and AgNP₂₀ increased the production of other analytes in the present study, these latter, including IL-8, MIP-1α, and MIP-1β are known to be potent chemoattractants and not as pro-apoptotic agents. This concords with potential pro-inflammatory activity of ZnO and/or AgNP₂₀. In this sense, ZnO NP was recently found to increase the production of IL-8 in human bronchial epithelial cells (Wu et al. 2010). Interestingly, ZnO NP are known to provoke metal fume fever (Kuschner et al. 1995), where an increased number of certain analytes, especially TNFα, IL-6, and IL-8, are observed in bronchoalveolar fluids (Kuschner et al. 1997). The current results are also in agreement with this observation since ZnO was found to increase the production of both TNFα and IL-8 in human 3D10 cells, but not of IL-6.

The pro-inflammatory activity of nanosilver NP is less documented than ZnO NP and the results are not clear. In a subacute murine inhalation model, nanosilver was found to induce minimal lung toxicity or inflammation (Stebounova et al. 2011). However, the average nano-silver NP size used in this study was ≈10 nm rather than 20 nm. In another study, nanosilver particles less than 100 nm and microsilver particles (250 μm) were implanted into the back muscles of rats and inflammation response levels measured based on patho- and morphological changes (Chen et al. 2007). After 90 and 180 d, local inflammation reaction induced by the implanted nanosilver particles was more serious than the microsilver particles. More recently, a higher propensity for citrate 20 nm NP to generate acute neutrophilic inflammation in the lung and to produce some cytokines was reported (Wang et al. 2013). In particular, the production of LIX and MCP-1 was increased in response to citrate 20 nm NP. Unfortunately, LIX, a chemokine usually associated with cell migration and activation of neutrophils, was not an analyte among the ones tested in the present study. Also, although we observed an increase production of MCP-1 in response to ZnO, we did not observe this for AgNP₂₀. However, as measured by ELISA, it was clear that human eosinophils could be directly activated by both NP to produce large amounts of IL-8, a potent pro-inflammatory molecule know to attract leukocytes, especially neutrophils (Kobayashi, 2008).

Conclusions

The results of this study indicate that NP do not activate similarly human eosinophils, since some (ZnO and AgNP₂₀) can induce apoptosis and cytokine production while others (TiO₂, CeO₂, and AgNP₇₀) do not have such effects. Moreover, NP belonging to a common group, namely metallic oxides such as TiO₂, CeO₂, and ZnO also impact differently on eosinophil cell biology. This is also true for a given NP with a different starting diameter; AgNP₂₀ induce apoptosis and cytokine production, but not AgNP₇₀. The fact that some NP can directly activate eosinophils to release chemokines that possess the capacity to attract leukocytes, including eosinophils themselves, support a role of these cells in cytokine networking leading to allergic airway inflammation.

Funding information

This study was partially supported by grants from Natural Sciences and Engineering Research Council of Canada (NSERC) and the Institut de Recherche Robert-Sauvé en Santé et en Sécurité du Travail (IRSST).

Disclosure statement

The authors declare no conflicts of interest. The authors alone are responsible for the content of this manuscript.