Introduction

There are numerous reports regarding the potential of inducing skin sensitization using chemicals [1]; however, whether skin can be sensitized using nanoparticles (NPs) remains unclear.

It is well-known that NPs released into the environment can enter the human body and potentially cause damage to organs [2, 3]. Metal oxide NPs have been used in various applications including the industrial, electrical, pharmaceutical, and biomedical fields because of their unique physicochemical properties compared to bulk chemicals [4]. These physicochemical properties include a small size, easy dissolution, and a surface charge. While desirable, these properties can also lead to various adverse effects in humans. Their very small size, which is characteristic of NPs, increases the surface area to volume ratio, showing higher toxicity than larger particles of the same composition with a lower surface area to volume ratio [5]. Thus, it is important to analyze the physicochemical properties of NPs because they can persist in vivo [6,7,8].

The current knowledge on the chemical and biological mechanisms associated with skin sensitization has been summarized in the form of an adverse outcome pathway, starting with the molecular initiating event through intermediate events to the adverse effect, namely allergic contact dermatitis [9]. The first key event is that electrophilic chemicals in the irritant initially covalently react with nucleophilic thiol and primary amines in skin proteins. The second key event occurs within the keratinocytes and includes inflammatory responses, as well as changes in gene expression associated with specific cell signaling pathways such as the antioxidant/electrophile response element (ARE)-dependent pathways [10]. The ARE-Nrf2 Luciferase KeratinoSens™ test, representing the second key event, can be used to discriminate between skin sensitizers and non-sensitizers in accordance with the United Nations Globally Harmonized System of Classification and Labelling of Chemicals [9].

Park et al. [11] reported that titanium oxide NPs do not induce skin sensitization in mice according to a local lymph node BrdU-enzyme-linked immunosorbent assay. In contrast, it has been shown that gold NPs can bind non-covalently to proteins and affect the immune system [12]. Moreover NPs may release free chemicals, which may have skin sensitization properties [13, 14]. However, there is little information available describing the skin sensitization potential of NPs.

Therefore, this study was performed to evaluate the skin sensitization potential of metal oxide NPs using the ARE-Nrf2 Luciferase KeratinoSens™ assay.

Materials and methods

NPs

CeO2, Co3O4, CuO, and TiO2 NPs were purchased from Nanostructured & Amorphous Materials (Houston, TX, USA). Fe2O3 NPs were purchased from American Elements (Los Angeles, CA, USA). NiO NPs were purchased from US-Nano Co. (Houston, TX, USA). ZnO NPs were purchased from Sumitomo (Osaka, Japan). Their primary size was confirmed by transmission electron microscopy (JEM-1200EX II, JEOL, Tokyo, Japan). The hydrodynamic size, polydispersity, and zeta potential of the NPs were measured using a Zetasizer-Nano ZS instrument (Malvern Instruments, Malvern, UK) in different vehicles including distilled water (DW), dimethyl sulfoxide (DMSO), and Dulbecco’s Modified Eagle’s Medium (DMEM; GIBCO, Grand Island, NY, USA) containing 1% heat-inactivated fetal bovine serum (FBS; GIBCO). The levels of endotoxin were evaluated using an Endpoint Chromogenic Limulus Amoebocyte Lysate assay (Cambrex, Walkersville, MD, USA).

Vehicle selection process

According to OECD TG 442D, the test materials were prepared in one of three different vehicles: DMSO (CAS: 67-68-5), DW, and culture medium (DMEM containing 1% FBS). Initially, these three types of vehicle were evaluated to select one for further study. Because NPs often initially agglomerate in vehicles, dispersion of the different metal oxide NPs was performed by using an ultra-sonicator [15]. Unlike the existing guideline method in which the solvent providing the best solubility is selected, because the nanomaterials are not soluble in solvents, the vehicle decision was based on the satisfactory uniform dispersion of the nanomaterials so that the NPs could be used to treat cells. CuO NPs were directly dispersed for 10 min in each solvent (DMSO, DW, and DMEM) by ultra-sonication, and then the hydration size of the NPs was measured by dynamic light scattering (Zetasizer-Nano ZS).

Preparation of NP suspensions

The suspensions of NPs in media were prepared by modifying the previously described method [16]. Briefly, the NP stock solution was dispersed in DW at a concentration of 200 mM and sonicated at 40 kHz with 100 W output power for 10 min in a bath type sonicator (Saehan-Sonic, Seoul, Korea). Next, DMEM supplemented with 1% FBS was added to different working concentrations (0.98–2000 µM).

Cell culture

A transgenic cell line with a stable insertion of the Luciferase reporter gene under control of the ARE-element KeratinoSens™ was obtained from Givaudan Suisse SA (Vernier, Switzerland). The cells were cultured in DMEM supplemented with 10% FBS, 0.5 mg/mL Geneticin (G418; Sigma-Aldrich, St Louis, MO, USA). KeratinoSens™ cells were sub-cultured every 3–4 days at 80–90% confluence for a maximum of 25 passages. For the experiments, KeratinoSens™ cells were seeded into 96-well plates at a density of 1 × 104 cells/well and the medium replaced with fresh medium (DMEM supplemented with 1% FBS) and then incubated in a humidified atmosphere of 5% CO2 at 37 °C.

Treatment with NPs and evaluation of their cytotoxicity

To evaluate cell viability, KeratinoSens™ cells were seeded into 96-well plates at a density of 1 × 104 cells/well and incubated overnight to reach approximately 80% confluency. The cells were washed once with pre-warmed DPBS (Gibco) followed by addition of fresh medium containing test materials (0.98–2000 µM) and incubation for 48 h. Cell viability was measured in a thiazolyl blue tetrazolium bromide (3-(4,5-dimethylthiazo-2-yl)-2,5-diphenyl-tetrazolium bromide assay reduction test (Promega, Madison, WI, USA). To exclude colorimetric interference from NPs present in the cells, the supernatant was transferred into clear 96-well plates and the absorbance was measured at 570 nm with a microplate reader (Tecan, Männedorf, Switzerland). The cell viability (%) was calculated based on the optical density of the vehicle control and blank.

Evaluation of luciferase expression in KeratinoSens™ cells treated with metal oxide NPs

To assess the induction of luciferase activity in KeratinoSens™ cells, the cells were seeded into 96-well plates at a density of 1 × 104 cells/well and incubated overnight to approximately 80% confluency. The cells were washed three times with pre-warmed DPBS (Gibco) followed by addition of fresh medium containing test materials (0.98–2000 µM), and incubated for 48 h. Luciferase activity was measured using the One-Glo™ Luciferase assay kit (Promega). The luminescence intensity of each sample was measured using a luminometer (Promega) and multi-microplate reader (Synergy 2, BioTek, Winooski, VT, USA). Luciferase induction was calculated based on the luminescence values of the vehicle control and blank.

Results

Selection of test vehicle for the NPs

Suspension of NPs in both DMSO and DMEM culture medium as test vehicles showed larger degrees of agglomeration than suspensions in DW (Fig. 1). Particularly, DMSO showed the severest level of aggregate formation, with particles up to 1 µm in size observed. As the test solvent, DW produced NPs with the smallest hydration size and most evenly dispersed particles, and therefore was selected as the test vehicle.

Fig. 1
figure 1

Selection of the test vehicle. a Hydrodynamic size and b polydispersity index of CuO NPs in DMSO, DW, and culture media (DMEM contained 1% FBS). Data are expressed as mean ± standard deviation values from six independent experiments

Full size image

Physicochemical properties of metal oxide NPs

Transmission electron microscopy images of the seven metal oxide NPs used in this study are shown in Fig. 2. The average sizes of the different NPs were < 100 nm in DW vehicles, showing a similar size to that provided by NP manufacturers. The physicochemical properties of the seven metal oxide NPs are summarized in Table 1. Hydrodynamic size analysis of the seven metal oxide NPs in working solution revealed that all types of NPs were agglomerated compared with their primary size (nm). Measurement of the zeta potential showed that all NPs were positively charged in DW, whereas in DMEM they were negatively charged, with zeta potentials of − 23 to − 32 mV. The Limulus Amoebocyte Lysate test showed that all metal oxide NPs had endotoxin levels that were lower than the limit of detection (0.1 U/mL).

Fig. 2
figure 2

Transmission electron microscopy images of the seven metal oxide NPs in distilled water. a CeO2 (15–30 nm), b Co3O4 (50–80 nm), c CuO (30–50 nm), d Fe2O3 (100 nm), e NiO (10–20 nm), f TiO2 (15 nm), and g ZnO (20 nm)

Full size image
Table 1 Physicochemical properties of the seven different metal oxide NPs
Full size table

Evaluation of NPs in the KeratinoSens™ assay

The seven metal NPs were assessed for their skin sensitization potential using the KeratinoSens™ assay; the data are shown in Table 2 and Fig. 3. CuO NPs induced the activity of the luciferase reporter by more than 1.5-fold, suggesting that they cause skin sensitization. The six other metal oxide NPs did not increase the luciferase activity in the KeratinoSens™ assay. The EC 1.5 value for the CuO NPs was 1.38 µM, classifying it as a sensitizer, whereas the EC 1.5 values were > 1000 µM for the remaining six metal oxide NPs, classifying them as non-sensitizers.

Table 2 Seven metal oxide NPs evaluated in KeratinoSens™ assay
Full size table
Fig. 3
figure 3

The induction of luciferase activity (closed diamonds) and cell viability (open squares) in the KeratinoSens™ assay. KeratinoSens™ cells were treated with seven metal oxide NPs: a CeO2, b Co3O4, c CuO, d Fe2O3, e NiO, f TiO2, and g ZnO. h Positive control (cinnamic aldehyde, 4–64 µM) was tested in parallel. Data are expressed as mean ± standard deviation values (n = 6)

Full size image

Discussion

This study was performed to evaluate the skin sensitization potential of metal oxide NPs using the ARE-Nrf2 Luciferase KeratinoSens™ assay. This assay is based on the second key event of skin sensitization, i.e., it evaluates whether a test substance sensitizes skin at the keratinocyte level [9]. The level of reproducibility in the predictions expected from the KeratinoSens™ assay is approximately 85% within and between laboratories, The accuracy of identifying a skin sensitizer by this test method has been demonstrated to be 77% (155/201) with a sensitivity of 78% (71/91) [17, 18].

Collectively, all available data indicate that the KeratinoSens™ assay is useful for identifying the risk of skin sensitization by compounds containing a variety of organic functional groups, pre-reactor, skin sensitization, and physicochemical properties [17,18,19,20]. Furthermore, although limited, data assessing various mixtures of compounds have been generated using this test method [21]. After 145 chemical tests, the KeratinoSens™ assay has been recognized as an OECD guideline and is currently being evaluated as an alternative test guideline [9, 17]. In general, skin sensitization tests are performed using purified chemicals, but evaluation of materials that are not completely dissolved has also been reported [21, 22].

The dispersion stability of nanomaterials has been identified as an important parameter affecting the environmental behavior of nanomaterials. This parameter depends on the physicochemical characteristics of the nanomaterial itself, physicochemical characteristics of the suspension medium, method by which the suspension is prepared, concentration of the nanomaterial, and concentration of other substances and particles in the suspension [23]. It is very important to produce a stable and uniform dispersion to accurately identify the toxicity of nanomaterials, as nanomaterial aggregates may exert different biological effects compared to well-dispersed nanomaterials [15]. Therefore, in our study, DW, which induces the smallest aggregate size, was used as a vehicle to fabricate the initial NPs stock solution. This was an essential choice for accurate quantification during dilution with a working solution.

The surface charge is suggested as one of the factors which control various biological responses to NPs. Depending on the surface charge of the particles, intracellular internalization may be affected [24, 25]. The surface charge, which showed a large variation for each NPs, showed a similar charge when diluted with the final working solution. This seems to be due to the formation of protein-corona due to a number of proteins contained in the culture media [15]. As a result, there was no association between zeta-potential and skin sensitization in our study results.

In this study, the KeratinoSens™ assay was performed to assess the potential skin sensitization effects of seven metal oxide NPs. One metal oxide NP, CuO, showed positive results, whereas the remaining six metal oxide NPs showed negative results. Both TiO2 and ZnO NPs are used in the cosmetics and other industries, and it has been reported that they do not show skin sensitization when evaluated using an LLNA local lymph node assay skin sensitization test. The same result was observed in our KeratinoSens™ assay [11, 26]. It has been reported that iron oxide does not induce sensitization, and cerium is not reported of the case of sensitization [27, 28]. Nickel and cobalt are shown the difference between the present study and previous studies which suggest them as sensitizers [29, 30]. Additional studies will be needed for the difference. In line with our study, copper oxide has been reported to induce sensitization in the LLNA test [31].

According to Cho et al. [32] and Jeong et al. [33], CuO, ZnO NPs can rapidly dissolve in lysosomes and yield their ions. In addition, free Cu-ions induced cytotoxicity in human skin organ culture [34]. In our study, the particularly high cytotoxicity of CuO and ZnO NPs could be explained by the toxicity of the released ions due to dissolution of the nanoparticles. In this study, since the degree of ionization of the NPs was not determined, the exact concentrations of metal ions released from metal oxide NPs could not be confirmed. Further studies are required to assess the sensitization effects of ionized copper ions in KeratinoSens™ cells.

We used the KeratinoSens™ assay to assess the skin sensitization potential of nanomaterials and found that CuO NPs induce skin sensitization. These results suggest that the ARE-Nrf2 Luciferase KeratinoSens™ assay may be useful for evaluating the potential of some NPs to induce skin sensitization. More studies are needed to more accurately evaluate the sensitization of NPs, and it seems necessary to revise skin sensitization guidelines for NPs based on various studies. Finally, we hope that these results may allow for the development of alternative test methods for testing the toxicity testing of nanomaterials.