Skip to main content
Download PDF
Download ePub

Understanding the toxicological effects of TiO2 nanoparticles extracted from sunscreens on human keratinocytes and skin explants

Particle and Fibre Toxicology volume 21, Article number: 49 (2024) Cite this article

Abstract

Background

Inorganic ultraviolet filters such as titanium dioxide nanoparticles are frequently used in sunscreens. Numerous toxicological studies in vitro and in vivo have been conducted using pristine standard reference nanomaterials of these inorganic filters. While convenient, this approach is not realistic because the complex environment of sunscreen formulations could change the physicochemical properties of the nanoparticles and lead to vastly different toxicological outcomes. Therefore, this study focused on characterizing nanoparticles extracted from commercial sunscreen and evaluating the associated toxicological impacts upon exposure to human keratinocytes and human skin explants.

Results

Titanium dioxide nanoparticles were extracted from commercial sunscreens and thoroughly characterized. The identity of the associated molecular corona on the extracted nanoparticles was also evaluated. Cell metabolic and proliferation profiles, mitochondrial superoxide activity, reactive oxygen species levels, and genotoxicity induced through exposure to the nanoparticles were studied in vitro using a human keratinocyte cell line. The cell response was significantly different after treatment with pristine nanoparticles compared to corresponding sunscreen-extracted nanoparticles. Pristine spherical nanoparticles resulted in more pronounced toxicity in 2D cultured keratinocytes compared to extracted nanoparticles but did not impact wound-edge migration significantly in 3D ex vivo human skin explant models. Additionally, extracted rod-shaped nanoparticles had greater toxic impacts in keratinocytes in vitro and retarded wound-edge migration in the ex vivo model compared to the extracted spherical nanoparticles. Nevertheless, these heightened cell responses were not associated with any increase in phosphorylated γH2AX (which is indicative of DNA damage) both in vitro and ex vivo.

Conclusions

This study shows the feasibility of extracting nanoparticles from personal care products such as sunscreens to obtain relevant forms to model real-life exposure scenarios. Overall, sunscreen-extracted nanoparticles were found to be less toxic compared to pristine equivalents but retarded wound-edge migration more significantly. Skin explant cultures provide a more realistic alternative to monolayer cell cultures, although the differential outcomes between the models need more in-depth evaluation.

Background

Titanium dioxide (TiO2) is commonly used in sunscreens as an inorganic ultraviolet (UV) filter together with organic UV filters such as benzophenones and salicylates. The global production of TiO2 between 2003 and 2004 was estimated at 1000 tons, and the market size has been projected to reach USD 28 billion by 2026 [1, 2]. Sunscreens contain the highest quantity of metal oxide nanoparticles among commercial products.

Recent evidence has shown that the detrimental effects on human health and the environment arise from organic UV filters used in sunscreens, such as avobenzone, oxybenzone, and octinoxate [3,4,5]. Therefore, increased use of inorganic UV filters in sunscreens is expected. The United States Food and Drug Administration (FDA) classifies TiO2 as “generally regarded as safe” (GRAS), and a maximum permissible limit of 25 wt% is allowed in sunscreens [6]. Nevertheless, many TiO2-based sunscreens have shifted to using nano-sized TiO2 for improved efficacy and aesthetics. This change results in uncertainties due to possible undesirable outcomes from nanoparticle (NP) interactions with biological systems, which require in-depth investigation to ascertain safety [7].

The general consensus is that particles could become more reactive as their size decreases. There has been a pervasive introduction of nanomaterials in products for human, which warrants the consideration of health and safety impacts resulting from their interactions with the human body. We hypothesize that the molecular compounds in sunscreen are adsorbed onto nanoparticulate UV filters in a similar way to the protein corona phenomenon [8]. This adsorption could change the NPs’ physiochemical properties and possibly affect the outcomes of toxicity tests [9]. Furthermore, human skin is another unique environment as well as biological system that contains a plethora of endogenous and exogenous molecules that could further change the NPs [10].

In conventional safety assessments, pristine NPs are administered directly to cultured cells throughout a cocktail of cell culture medium. However, the corona formed on the NPs could produce contrasting effects that depend on the type of cell being tested [11,12,13,14]. A considerable number of corona studies have been reported, but there is still limited understanding of the effects of NP coronas in a dermal environment. This raises the question of whether real-life sunscreen-extracted NPs with relevant coronas could post any detrimental effect on keratinocytes and the skin.

We have previously shown that pristine TiO2 and zinc oxide (ZnO) NPs lead to different responses in primary human epidermal keratinocytes, including direct cell death and more subtle impacts such as autophagy induction [15, 16]. However, transformations of the NPs arising from interactions with the environment were not considered in those studies. This knowledge gap has not been well addressed in the literature thus far and is recognized by the FDA as well [17]. Recently, 3D in vitro skin models such as EpiSkin [18] and EpiDerm [19] have been used in addition to monolayer cell cultures to reflect real skin conditions as closely as possible. Nonetheless, ex vivo skin explants from human patients or volunteers provide the closest representation of human skin. Martins et al. used this approach and found that their solid lipid NPs were successful in protecting keratinocytes from UVB radiation [20].

A more recent study used full-thickness human skin explants and demonstrated that TiO2 NPs reduced the production of inflammatory cytokines IL-6 and IL-8 upon exposure of human skin explants to UVB [21]. Limited studies have used skin explants to conduct nanotoxicology tests with relevant NPs. Hence, the objective of this study was to evaluate the biological impact of TiO2 NPs extracted from sunscreen products when exposed to monolayer keratinocyte cultures in vitro in comparison with ex vivo skin explant cultures.

Results

Physicochemical characteristics of NPs

Five different TiO2 NPs were used in this study. TiO2 P25 Aeroxide (P25) was chosen because it represents the standard reference material (SRM). It is primarily used as a benchmark and tool for evaluation of the potential environmental, health, and safety risks that might be associated with manufactured nanomaterials during their product life cycle. Additionally, experiments were also done with TiO2 NPs that were extracted from commercially available sunscreens, as well as their size- and shape-matched pristine TiO2 NPs. These sample groups are referred to as extracted sphere (ExS), pristine sphere (PrS), extracted rod (ExR), and pristine rod (PrR).

Primary sizes of the five different NPs studied were analyzed with a transmission electron microscope (TEM). The representative TEM images in Fig. 1A–E show that P25, PrS, and ExS were spherical, whereas PrR and ExR were rod-like particles. At higher magnifications (Fig. 1F–J), the formation of a corona was detected for PrS, ExS, and ExR, as evidenced by a thin coating layer surrounding the NPs, which is indicated by white arrows.

Fig. 1
figure 1

Representative images of nanoparticles analyzed in this study; P25, PrS, ExS, PrR, and ExR. A-E) High Resolution TEM images indicating the size and shape of each nanoparticle; F-J) TEM images showing the corona formation on each nanoparticle, white arrows indicating the corona formation around the nanoparticles; K-O) Histogram of primary particle size and/or aspect ratio, n = 100. Legend: P25 – Pristine P25; PrS – Pristine Sphere; ExS – Extracted Sphere; PrR – Pristine Rod; ExR – Extracted Rod

Full size image

The size distributions and dimensions determined from TEM images are shown in Fig. 1K–O. P25 had an average primary size of 25 nm, which is consistent with the manufacturer’s datasheet [22]. The average primary sizes of PrS and ExS were 88% and 24% larger than that of P25. The aspect ratios of the non-spherical PrR and ExR were 9.48 and 10.03, respectively.

The results indicated that a wide variety of different sizes and shapes of NPs were used in the commercial sunscreens. When NPs were dispersed in deionized (DI) water, their hydrodynamic sizes were significantly different (Supplementary Table 1). Only ExS had a larger hydrodynamic size than P25, while PrS, PrR, and ExR had a smaller hydrodynamic size than P25. The zeta potential was negative for all the NPs except for P25, which suggests that the corona on the NPs could have played a role in changing the zeta potential of the pristine TiO2 NPs.

Interestingly, when the NPs were dispersed in EpiGRO cell culture media, the hydrodynamic size of ExS, PrR, and ExR increased by approximately 100% compared to that in DI water, but those of P25 and PrS increased drastically by 1,500% and 1,160%, respectively. The dramatic increase in the case of P25 could be attributed to the high amount of interaction of the pristine particles with components in the cell culture media, while this effect was mediated by the corona of other NPs, which was not seen on ExS, PrR, and ExR [23].

After 24 h of incubation in EpiGRO media, there was no significant increase in the hydrodynamic sizes of all NP groups. The zeta potential, however, equilibrated between − 7.1 and − 13.6 for all NP groups. This indicated that negatively charged components of the media contributed to the corona on the NPs, which is consistent with other reports [23].

X-ray diffraction (XRD) confirmed the presence of TiO2 in all NP samples. P25 showed unique peaks for TiO2 at 27.5º, 36.1º, and 54.3º for the anatase phase (01-083-5914) and 25.3º, 37.8º, and 48.2º for the rutile phase (01-079-6029). PrS, ExS, PrR, and ExR contained only the rutile phase of TiO2 (04-008-7850, 04-008-7850, 01-086-4330, and 04-006-1930, respectively; Fig. 2A). This is consistent with previous results, which indicate that rutile-phase TiO2 is more commonly used in sunscreen formulations due to its lower toxicity and photocatalytic activity under UV light compared to anatase-phase TiO2 [24].

Fig. 2
figure 2

(A) XRD spectra of nanoparticles, suggesting the presence of Anatase (A) phase and Rutile (R) phase TiO2. Only P25 contains a combination of Anatase and Rutile phase whereas the remaining groups only contain Rutile phase; (B) TEM-EDX spectra of nanoparticles, indicating the presence of Ti and O peaks, suggesting the presence of TiO2. ExS contains an additional Si peak and PrR contains an additional Al peak, indicating the presence of an additional passivation layer around the TiO2 nanoparticle. Cu signals originate from the copper TEM grids used. Legend: P25 – Pristine P25; PrS – Pristine Sphere; ExS – Extracted Sphere; PrR – Pristine Rod; ExR – Extracted Rod

Full size image

High-resolution TEM images were also obtained to observe the interplanar distances and diffraction rings of individual NPs (Supplementary Fig. 1B–D). P25 had an interplanar d-spacing of 3.76 Å which coincided with the unit cell value of a = 3.7845 Å (01-083-5914) (Supplementary Fig. 2). PrS, ExS, PrR, and ExR had interplanar d-spacing ranging from 2.62 to 3.00 Å, which coincided with the c value of the unit cell (04-008-7850, 04-008-7850, 01-086-433, and 04-006-1930, respectively; Supplementary Figs. 3–5). This further supports that the NPs in TEM images were TiO2.

Energy dispersive X-ray spectrometry (EDX) analysis also confirmed the presence of the elements titanium (Ti Kα) and oxygen (O Kα) in all 5 NP samples. Both PrR and ExR contained an aluminum (Al) peak, indicating the possibility of an alumina (Al2O3) passivation layer on the NPs. ExS exhibited a silicon (Si) peak that was absent from the PrS spectrum (Fig. 2B). Scanning transmission electron microscopy (STEM) was combined with EDX to obtain a 3D elemental map of the distribution of titanium and oxygen (Supplementary Fig. 1A).

Spectroscopy measurements of the absorbance in the UV region of 100–400 nm showed good agreement with published data regarding the use of these NPs as UV filters [25] (Supplementary Fig. 1E). Only PrS showed higher absorbance than the SRM P25. The remaining samples (ExS, PrR, and ExR) showed lower absorbance than P25. The differences in UV absorbance between the NPs were intriguing. In particular, PrR and PrS showed a marked difference despite both being anatase TiO2. The differences in UV absorbance of these different forms support the rationale of incorporating both organic and inorganic UV filters in commercial sunscreens.

Corona analysis of NPs

Fourier-transform infrared (FTIR) spectra of the NPs were recorded (Fig. 3A), and the characteristic peaks are summarized in Supplementary Table 2 [26]. No distinguishable peaks were detected for both P25 and PrS. The identifiable peaks for PrR were narrowed down to the possible compounds aluminum stearate, alumina, and simethicone, as stated in the technical datasheet from the manufacturer [27]. The first two peaks at ∼ 2920 cm− 1 and ∼ 1467 cm− 1 were due to C-H stretching and bending, respectively, and were assigned to aluminum stearate [28]. The next two bands at ∼ 1423 cm− 1 and ∼ 1350 cm− 1 were from O-H bending and were assigned to alumina.

Fig. 3
figure 3

(A) FTIR analysis of NPs in DI Water. Areas of interest are highlighted as Areas A and B where individual peaks were identified. (B) TGA of nanoparticles in EpiGRO media. (C) LC-MS/MS analysis of 5 most abundant proteins found in the corona on the nanoparticles suspended in EpiGRO media. Legend: P25 – Pristine P25; PrS – Pristine Sphere; ExS – Extracted Sphere; PrR – Pristine Rod; ExR – Extracted Rod

Full size image

Adsorbed H2O molecules could have reacted with CO2 at room temperature alongside aluminum and formed aluminum bicarbonate [29]. The final two bands at ∼ 1260 cm− 1 and ∼ 1100 cm− 1 were due to Si-CH3 stretching and Si-O stretching, respectively, and were assigned to simethicone [30,31,32]. The extracted NPs showed intriguing results due to the presence of a corona composed of compounds used in the sunscreen formulation. For ExS, the bands at ∼ 2920 cm− 1, ∼ 1663 cm− 1, ∼ 1467 cm− 1, and ∼ 1423/1350 cm− 1 represented chemical bonds in ingredients that are commonly found in sunscreens and suggest that these components are present in the corona. The peak at ∼ 1727 cm− 1 represented C = O stretching in aliphatic ketones, which are found in some chemical UV blockers in sunscreens, while the peak at ∼ 1100 cm− 1 was indicative of a silicon passivation layer on the TiO2 NPs.

The only uncertainty was the band at ∼ 1410 cm− 1, which represented S = O stretching. The reason for this uncertainty was that the ingredient list did not contain any compounds that have S = O bonds. For ExR, the peak at ∼ 1727 cm− 1 similarly suggested the presence of chemical UV blockers that are used in sunscreens. There were also peaks at ∼ 1260 cm− 1 and ∼ 1100 cm− 1, which corresponded to silicon-based materials.

The composition of the coronas on the NPs were determined with thermogravimetric analysis (TGA) following 24 h of incubation in EpiGRO media. The final weight percentages of the samples stabilized at 650 °C and were recorded as 97.5% (P25), 95.8% (PrS), 84.2% (ExS), 74.0% (PrR) and 69.9% (ExR) (Fig. 3B). The TGA results strongly indicated significance differences in the corona mass percentages on each NP.

To distinguish the corona profiles of each of the NPs, we employed liquid chromatography with tandem mass spectrometry (LC-MS/MS). The detected peptides were compared to the UniProt database to determine the identities and accession codes of the proteins present. The most abundant protein identified was epididymis secretory sperm-binding protein Li 71p (A0A0K0K1H8), which is also known as serotransferrin (Fig. 3C). This iron-binding protein helps in stimulating cell proliferation.

The second most abundant protein was gamma-enolase (P09104), which binds to calcium and promotes cell survival. The third was alpha-enolase (P06733), which is involved in growth control and glycolysis. The fourth was fructose-bisphosphate aldolase A (P04075), which also participates in glycolysis and acts as a scaffolding protein. The fifth was an actin isoform (Q8WVW5). Compared to their pristine counterparts (PrR and PrS), ExR and ExS had 85% and 57% lower amounts of gamma-enolase, as well as 40% and 10% lower amounts of alpha-enolase, respectively.

In vitro dosimetry model, sterility, and endotoxin tests for Ker-CT cell studies

Since the five different NP samples had varying hydrodynamic sizes, primary sizes, shapes, and corona profiles, it is important to establish a standardized delivered dose for experiments involving the hTERT-immortalized keratinocyte cell line (Ker-CT). This was done using the in vitro sedimentation, diffusion, and dosimetry (ISDD) model in combination with the volumetric centrifugation method (VCM) [33].

The necessary critical sonication energy (DSEcr) to produce stable suspensions of P25, PrS, ExS, PrR, and ExR were determined to be 529 J/ml, 353 J/ml, 1,764 J/ml, 353 J/ml, and 0 J/ml, respectively (Supplementary Fig. 6A). ISDD simulation was conducted to determine the deposition rate of the NPs over a period of 24 h (Supplementary Fig. 6B). Lastly, the data were correlated with real-life exposure calculations performed in a previous study [15, 34] to determine the required administered dose to simulate a realistic range of exposures (0.01 µg/cm2 to 5 µg/cm2) (Supplementary Table 3).

Sterility and endotoxin tests were conducted to ensure that toxicity measurements were directly resulting from NP exposure. All samples showed no bacteria colony formation up to 21 days of culture (Supplementary Fig. 7). Endotoxin test results were corrected based on the administered dosage. PrR and ExR at a dose of 5 µg/cm2 had higher endotoxin levels than the recommended limit of 0.5 EU/ml determined by the United States FDA for pharmaceutical applications (Supplementary Table 4) [35]. Regardless, the higher endotoxin levels in these 2 NP groups did not adversely impact cell responses upon high dose exposure.

NPs uptake in 2D cell culture

Potential NP uptake, translocation, and preferential localization were investigated using Ker-CT cells. Generally, all types of NPs were taken up by the cells (Fig. 4A and B, Supplementary Fig. 8A–12A). The NPs were found within the cytoplasm but not in the nucleus. The NPs were mainly confined within the cells’ vacuoles and endosomal compartments. The NPs in the cells were confirmed to be TiO2 using TEM-EDX mapping (Supplementary Fig. 8B–12B).

Fig. 4
figure 4

(A) Representative TEM images of Ker-CT cells indicating ExS NP uptake. (B) Representative TEM images of Ker-CT cells indicating ExR NP uptake. (C) Alamar blue assay of Ker-CT after 24 h of exposure to NPs. (D) CellROX assay of Ker-CT after 24 h of exposure to NPs. Data represent means ± SD, n = 3; #p < 0.05 compared to control (ctrl), Student’s t-test; * p < 0.05; one-way ANOVA, Tukey’s HSD. Legend: P25 – Pristine P25; ExS – Extracted Sphere; ExR – Extracted Rod

Full size image

2D in vitro cytotoxicity and genotoxicity studies

Cytotoxicity and genotoxicity studies were conducted using Ker-CT. Light microscopy images showed that cells exposed to the NPs for 24 h were healthy after receiving an administered dose of 0.01 µg/cm2 (Supplementary Fig. 13). However, significant differences were observed at the highest administered dose of 5 µg/cm2. At this dose, only cells exposed to ExR appeared healthy, while those in the other NP groups presented unhealthy morphologies. P25 induced a 20% reduction in cell metabolic activity after 24 h of exposure at doses of 0.1 µg/cm2 or more. ExS and ExR did not exhibit any concentration-dependent impact on cell metabolism (Fig. 4C). Similar observations were seen in the proliferation of Ker-CT cells exposed to the different NPs (Supplementary Fig. 14A).

Immunofluorescence analysis was performed on the Ker-CT cells to observe any phosphorylated H2AX (p-γH2AX) signals, which are a marker of double-stranded DNA damage. Cells exposed to UV for 30 min were used as positive controls. Additionally, ZnO NPs were used as a positive NP control to induce p-γH2AX expression. When Ker-CT cells were exposed to any of the NPs, there were no observable p-γH2AX signals at all exposure doses of 0.01 to 5 µg/cm2. This indicated that the NPs did not induce significant DNA damage even though some of them could potentially be cytotoxic at high concentrations (Supplementary Fig. 15).

Mitochondria superoxide and reactive oxygen species (ROS) production in Ker-CT cells

Elevated ROS levels were detected in Ker-CT cells that had been exposed to P25 and ExR at 1 µg/cm2 and 0.1 µg/cm2, respectively (Fig. 4D). Significantly elevated levels of mitochondria superoxide activity were also observed in the P25 and ExR treatment groups at 1 µg/cm2 and above (Supplementary Fig. 14B). ROS elevation and mitochondria superoxide induction were not significantly changed when the cells were treated with ExS. P25 elicited a significantly higher level of ROS production in Ker-CT cells at a dose of 5 µg/cm2 in comparison to ExS and ExR. A similar trend was observed with mitochondrial superoxide expression.

Toxicity studies using 3D ex vivo skin explants

Hematoxylin and eosin (H&E) staining and wound-closure rates of 3D ex vivo skin explants were used to elucidate the potential impact of the NPs on skin. Although the application of sunscreens directly onto wounds may be rare, post-operative wound care involving intended application of ZnO-containing sunscreen has been reported [36]. More realistically, unintended NP exposure and penetration into the skin through a wound could occur through superficial epidermal wounds, such as minor cuts and abrasions sustained during sports while having sunscreens applied on the skin. This is especially plausible given the general advice that sunscreens should be re-applied regularly during prolonged periods of sun exposure.

This wounded skin model allows us to reflect the highest possible exposure of NPs to the viable layers of the skin, to obtain information about the interaction of these NPs with keratinocytes, and to investigate the most extreme effects possible on the wound-closure process. All treated groups showed a similar trend to the control group, which received no treatment (Fig. 5A, Supplementary Fig. 16). H&E staining showed the appearance of a migrating tongue shape from Day 1 after wound creation. On day 3 clearly, the wound showed a migrating tongue shape extending from the wound edge, indicating a normal wound-healing process.

Fig. 5
figure 5

(A) Representative H&E images of 3D skin explants subjected to P25 treatment at 100 µg/ml compared to untreated control. (B) Representative TEM images of untreated control and P25 treated 3D skin explants. Solid white arrows indicate melanosomes and dashed white arrows indicate uptaken NPs. (C) Plot of migrating tongue length of wounded 3D skin explants after exposure to P25, ExS, and ExR, compared to untreated controls, at Days 1 and 3; N = 3, n = 3, lower and upper fences mark 25th and 75th percentile values; Medians are labelled with black/blue lines; means are indicated with light blue lines; whiskers represent 10th and 90th percentile; outliers are marked in red. #p < 0.05, one-way ANOVA, Tukey’s HSD. Legend: P25 – Pristine P25; ExS – Extracted Sphere; ExR – Extracted Rod

Full size image

TEM images of the skin explants (Fig. 5B) showed that NP uptake by the keratinocytes occurred, and the NPs were typically localized in the perinuclear regions of the cytoplasm (dotted arrows) [37]. These NPs were found in vacuoles and endosomes of the keratinocytes, which likely arrived there through endocytosis [38, 39]. No NPs were found within cell nuclei. NPs were commonly localized near melanosomes, although the significance of this is unknown.

In terms of wound-edge migration, the data suggest that rod-shaped NPs negatively affected keratinocyte migration in the 3D skin explants at day 3 of treatment (Fig. 5C). No significant differences were observed across all treatment groups on day 1 in comparison to untreated controls. Additionally, no NPs resulted in any observable p-γH2AX signals, indicating no DNA damage in the keratinocytes at all exposure doses (Supplementary Fig. 17).

Discussion

The aim of this study was to assess the validity of sunscreen-extracted NPs to provide a more realistic model to understand any potential toxicity effects using 2D in vitro keratinocytes and 3D ex vivo skin models. Most reports of TiO2 NP-induced dermal toxicity have used P25 and did not consider the contributions of the microenvironments within complex sunscreen formulations and cell culture media [15, 40]. These microenvironments result in corona formation due to the interactions of the NPs with the organic fractions, which could change the particles’ physiochemical properties and possibly affect toxicity assessments [9].

TEM images (Fig. 1A–E) showed that TiO2 NPs of different shapes and sizes were used in different sunscreen formulations. This provided early indications that P25 alone should not be used as a reference material for such toxicity studies [15]. Additionally, the size and shape of a single NP are known to affect its effects in biological systems, so using just P25 would not provide an accurate representation [41].

Higher-magnification TEM images (Fig. 1F–J) also showed coronas on ExS and ExR NPs, which likely originated from the ingredients within the sunscreen formulation [42, 43]. P25 and PrR showed no corona formation around the TiO2 NPs, while PrS did, presumably resulted from the ingredients used to manufacture the particles [27]. As a result of corona differences, in both DI water and EpiGRO media, the hydrodynamic sizes of all TiO2 NP samples differed significantly from their primary sizes, as well as between the different NPs [44]. Additionally, the positively charged P25, negatively charged PrS, ExS, and ExR, and neutral PrR in DI water all became negatively charged NPs when incubated in EpiGRO media (Supplementary Table 1), indicating that these NPs were sterically stable in the media [45]. ExS and ExR constituted less than 5 wt% of the sunscreen formulation, which is within the typical range of 2–15 wt% used in sunscreens according to the United States Environmental Protection Agency and the maximum allowable limit of 25 wt% stipulated by the United States FDA [6, 46].

We successfully characterized NP coronas arising from interactions with both sunscreen and cell culture media. Both FTIR and LC-MS/MS indicated unique compositions of the coronas on each NP type and provided evidence that physicochemical properties between NPs will vary due to corona differences. This is important information that should be considered when evaluating NP toxicity. There is no information about whether the sunscreen manufacturer used NPs with silica/aluminum coatings, but the FTIR analysis revealed strong Si-O bonds on ExS and ExR, which could indicate the presence of such coatings or coronas containing silicone-based compounds found in other sunscreen ingredients.

TEM images of Ker-CT cells showed that all types of NPs were taken up by the cells (Fig. 4A). The NPs were mainly confined to vacuoles in the cytoplasm but were absent from the nuclei, which is consistent with current literature [47, 48]. Sterility tests showed no viable microorganisms on the NPs, and all NP groups except PrR and ExR had endotoxin levels that were below the maximum limit in the United States FDA guidelines for pharmaceutical applications (0.5 EU/ml). The endotoxin levels in PrR and ExR were consistent between 3 batches of extraction. Nonetheless, the measured levels of ∼ 0.5–0.7 EU/ml (Supplementary Table 4) were still lower than the FDA limit.

Ker-CT cells had a dose-dependent response to P25, which reduced metabolic activity of the cultures with increasing dosage. Neither ExS nor ExR significantly reduced cell metabolic activity (Fig. 4C), indicating that cell responses to NP interaction changed with respect to differences in the corona on the extracted version of both NPs compared to that on P25. One reason could be the presence of gamma-enolase and alpha-enolase (ENO1) proteins found in the corona. A high amount of ENO1 suppresses cell proliferation and glycolysis, which would reduce overall cell metabolic activity upon exposure to P25 [49]. ENO1 levels were 1.3-times higher on P25 than ExS and 5.5-times higher on P25 than ExR for exposures of more than 0.1 µg/cm2 (Supplementary Table 5).

The strong Si-O stretching peak in the FTIR analysis indicated that silicon-based compounds were present in the corona of ExS and ExR (Supplementary Table 2), but not P25 [50, 51]. The resulting passivation layer would reduce the reactivity of the original ExS and ExR NPs. There was a significant increase in intracellular ROS and mitochondria superoxide levels when Ker-CT cells were exposed to P25 and ExR at 1 µg/cm2 and above. A contributing factor could be the significant variations in NP corona compositions (Supplementary Table 5). Differences in corona composition could disguise the identity of the particle differently and consequently vary particle uptake and toxicological outcomes, a phenomenon akin to the “Trojan horse” effect [36].

Additionally, several authors have also shown that passivation of TiO2 NPs by silica or aluminum during the manufacturing process could prevent the formation of ROS, which would mitigate any potential ROS-induced toxicological outcomes [52]. This is supported by PrS having led to 20% lower cell metabolic activity than ExS (Supplementary Fig. 18A) and eliciting significantly higher formation of mitochondria superoxide (Supplementary Fig. 18B). Overall, ExS presented the lowest toxicity impact on Ker-CT cells compared to all other NPs, suggesting that a composite corona made up of the sunscreen matrix and components of the cell culture media helped to mitigate NP toxicity [53].

Fetal bovine serum (FBS) has had the most well-studied drastic effects on NP–cell interactions [54, 55]. No previous studies have systematically considered the sunscreen corona on inorganic NP-based UV filters and correlated it to toxicological outcomes. In contrast to spherical NPs, PrR was less toxic than ExR (Supplementary Fig. 19). The rod shape of PrR could have led to an increase in uptake and thus an increase in ROS generation [56]. The results suggest that using extracted NPs instead of pristine ones provides more reliable and representative toxicological information, which has been suggested previously [57].

NPs are reported to be able to break the double strands of DNA [58], but there was lack of elevated p-γH2AX signals in our in vitro and ex vivo studies. This could be attributed to the corona formation on the NPs. The corona could have acted as a protective barrier, which would be coupled with the fact that TiO2 does not dissociate into ions, unlike zinc oxide (ZnO) NPs, which are known to cause elevated p-γH2AX signals [59].

Co-exposure to both NPs and UV would be a more realistic scenario than the conditions examined in this study. However, there are significant complications when including UV exposure, especially in terms of uncertainties in the methodology. For example, we have to ensure that the same amount of UV energy reaches the NPs in different environments (sunscreen vs. culture media), which can then be translated into similar photoactivation or ROS generation in TiO2 NPs. We would also have to be able to measure and correlate the produced ROS, which would be short-lived. Hence, we decided to focus only on the NPs with the corona and to leave the examination of UV exposure for future studies after these technical challenges are resolved.

Data from 3D skin explants showed similar results to the 2D cell cultures. Rod-shaped NPs had more adverse effects on keratinocytes than spherical NPs in both the 2D and 3D models. There were no significant differences between all treatment groups in comparison to the controls on day 1 of treatment. On day 3, only ExR (at both concentrations of 100 µg/ml and 500 µg/ml) led to a significant reduction in the migrating tongue length compared to the control (Fig. 5C). NPs with higher aspect ratios are generally expected to elicit more toxic response [60].

When comparing ExR data against its pristine counterpart (PrR) (Supplementary Fig. 20), it was evident that ExR significantly reduced the migrating tongue length, which strongly suggests that the corona had a major impact on the cell response to the NPs. The same observations were also made when ExS was compared to its pristine counterpart (PrS) and resulted in a significantly lower migrating tongue length compared to PrS on day 3 (Supplementary Fig. 21). This result corresponded to the results from our 2D cell culture study.

We evaluated the differences between relative and absolute measurements for the migrating tongue length and concluded that absolute measurements would provide the most accurate data for comparison. The reason is that the speed of the epithelial tongue migration is the same between each NP group regardless of the original wound size. Comparing the relative measurement (% wound closure) would mask differences due to the variations in the original wound size. In future studies, we will attempt to develop an ex vivo model with consistent wound sizes, which will offer greater flexibility in data interpretation.

p-γH2AX expression in the 3D explants supported the conclusion from 2D cultures, where no observable double-stranded DNA breaks were recorded across all exposure scenarios. While ROS levels could be directly measured in 2D model, the same measurement could not be employed for tissue sections that heavily rely on the indirect detection of ROS oxidative byproducts, such as peroxidized lipid. The ROS level increased around 20–30% in the 2D model, and normal keratinocytes have robust anti-oxidative defense mechanisms to modulate oxidative homeostasis, so we expect the oxidative damage in the 3D model to be minimal or nonexistent. This could also be indirectly validated by our DNA damage analysis, which showed consistent results in both 2D and 3D models.

DNA damage is one of the likely outcomes of highly elevated intracellular ROS if left unchecked, and the absence of DNA damage in the 3D model suggested that there was minimal perturbation of ROS homeostasis in the skin explant. Further studies with a larger sample size for 3D ex vivo skin explants will help to confirm the results and give information about subtle mechanistic differences between the different treatments. Overall, our results demonstrated that the changes of NPs in relevant environments must be considered to produce relevant toxicological assessment outcomes. In the present case, the corona derived from sunscreen matrix and cell culture media influenced the results. A more physiologically representative experimental model such as the skin explant cultures used in this study can be used to validate results from 2D cell culture studies.

Conclusions

TiO2 NPs were successfully extracted from commercial sunscreens while preserving their corona signature. They were characterized alongside their pristine counterparts and compared against the widely reported reference TiO2 material P25. The extracted NPs were confirmed to be TiO2, and their unique coronas were characterized. Based on cell metabolic activity, proliferation rate, and the production of ROS and mitochondria superoxide, ExS had the highest cytotoxicity impact in vitro, followed by ExR and then P25. However, ExR had the highest capacity to retard wound-edge migration in the ex vivo model, followed by ExS and P25.

This study validated the hypothesis that TiO2 NPs extracted from sunscreens have differing physicochemical properties compared to pristine reference materials. These extracted TiO2 NPs are more relevant for use in models to study nanotoxicological impacts based on real-life exposure scenarios. In summary, extracted TiO2 NPs induced lower cytotoxic pressures on keratinocytes and retarded wound-edge migration in vitro and ex vivo.

Methods

Nanoparticle sources and extraction protocol

SRMs are intended primarily for use as a benchmark and tool for evaluation of the potential environmental, health, and safety risks that might be associated with manufactured nanomaterials during their product life cycle [61]. In this study, pristine P25 (AEROXIDE® TiO2 P25, Evonik Corporation, Germany) was designated as the SRM for TiO2. Four other NPs were used in this study and divided into two categories primarily according to their shape. Each of the five NPs was given an abbreviation, as shown in Supplementary Table 6.

Pristine TiO2 (Natpure Screen TWG and Solaveil™ CT-12 W) was obtained from Sensient Cosmetic Technologies, SEA and Oceania, and Croda Singapore [27, 62, 63]. Commercial sunscreens were purchased over the counter in Singapore (Cetaphil UVA/UVB Defense, 50 ml, and Cetaphil SPF 50 + Light Gel, 50 ml) [42, 43] (Supplementary Table 6). The sunscreens’ lists of ingredients can be found in Supplementary Table 7.

Among the samples studied, only PrR NP was found to be coated as received. We did not perform any additional coating or modification on any of the NPs used in the study. The extraction protocol for ExS and ExR was modified from previous studies [64,65,66]. Briefly, a 100-mg sample of sunscreen was weighed, 10 ml of hexane was added, and sonication was performed to create a 1% w/v suspension. Next, the suspension was centrifuged at 2,350 rcf for 60 min, and the resulting pellet was washed twice with DI water. The pellet was then resuspended in DI water with sonication and used for all subsequent experiments.

Primary size, morphology, and elemental composition characterization

The TiO2 NPs were imaged using TEM (JEOL 2010 HR and JEOL 2100 F) with an accelerating voltage of 200 kV. The particles were prepared as a stock solution in DI water and sonicated using a probe sonicator (Qsonica Q125) at their respective critical sonication energies. They were then drop-cast onto pure carbon film and left to dry fully overnight at room temperature.

Each grid was imaged using TEM (JEOL 2100 F), and EDX mapping was conducted to obtain the elemental composition and mapping of each sample. Additionally, each TEM grid was imaged using the JEOL 2200 F to obtain high-resolution images to recreate the line profile of the NPs, and a diffraction pattern was obtained. Primary particle diameters were determined using ImageJ based on 100 randomly selected NPs [67]. The diameter was used to calculate the primary size of NPs with a circular morphology, while the longitudinal and transverse lengths were measured for NPs with rod-like structure to determine the aspect ratio.

Hydrodynamic size and zeta potential characterization

Each type of NP was suspended in a stock solution containing DI water (10 mg/ml) and sonicated using a probe sonicator (Qsonica Q125) at their respective critical sonication energies (Supplementary Table 8). They were then diluted to a final working concentration of 500 µg/ml with various media (DI water and EpiGRO). Dynamic light scattering (DLS, Malvern Zetasizer) analysis was then performed to quantify the hydrodynamic size and zeta potential.

Crystallographic analysis

Each pristine NP sample was suspended in a stock solution with a concentration of 10 mg/ml, while the extracted NP samples were used immediately after extraction with no modification. Each sample was drop-cast onto a low-background silicon disc and left to dry overnight. The discs were then analyzed using XRD (Shimadzu XRD-6000) at 10°< 2θ < 90° with a step size of 0.02° and 1 scan speed.

Corona analysis

Each NP was incubated in EpiGRO media for 24 h before being freeze-dried, while control NP samples were incubated in DI water before being freeze-dried. All samples were then analyzed with FTIR spectroscopy (Perkin Elmer Frontier) in attenuated total reflectance (ATR) mode. Another batch of freeze-dried NPs incubated in EpiGRO media was sent for analysis using LC-MS/MS. In parallel, NPs incubated in EpiGRO media were also analyzed by TGA (TA Instruments Q500), for which samples were heated to 600 °C at a rate of 1 °C/minute.

UV absorbance

Each NP type was analyzed with a UV-Vis spectrometer (UV-2700) at a concentration of 100 µg/ml.

Dosimetry, endotoxin, and sterility studies

Dosimetry evaluation was done using an established protocol based on an extension of the ISDD model [33]. To simulate real-life dermal exposures, the highest sunscreen application frequency reported by Lorenzo et al. [34] was used to calculate the corresponding NP exposure per unit skin area and consequent penetrated NP doses as described previously [15]. Consideration was also given to the scenario in which a compromised skin barrier could allow increased NP penetration [68]. Based on this approach, realistic in vitro exposure dosages were estimated to be between 0.026 and 0.531 µg/cm² [15].

Based on the result, we selected in vitro exposure dosages of 0.01, 0.1, 1, and 5 µg/cm² for this study (Supplementary Table 3). These values represent the full spectrum of exposure, which covers both physiologically relevant doses and high doses that are expected to elicit a toxic response in cells according to current literature [69]. In the ex vivo explant study, a damaged skin scenario was examined to simulate the unintended application of sunscreen on an open wound, which would hypothetically result in higher NP exposure due to the compromised skin barrier. Such a situation is likely to occur during outdoor activities where minor skin injuries are sustained and sunscreen is reapplied [68]. The damaged skin model also serves as an extreme exposure scenario to understand the potential cellular effects of NPs in sunscreens, which have been shown to penetrate the intact skin barrier, albeit in small amounts [70, 71].

Each suspension of NPs was first determined to be stable over a period of 24 h by applying the required sonication energy. The suspension was determined to be stable if the hydrodynamic size did not deviate by more than 5%. The dispersed and stable suspensions were then incubated in complete cell culture media to determine the effective densities of the agglomerated NPs. Lastly, MATLAB was used to simulate the deposition rate of each NP sample.

The endotoxin test was conducted using the Pierce™ LAL Chromogenic Endotoxin Quantitation Kit (Thermo Fisher). A stock solution was prepared for each NP at 1 mg/ml, and the pH was adjusted to the range of 6.0 to 8.0 before dilution to a final concentration of 10 µg/ml in endotoxin-free water. A standard curve (0.1–1.0 EU/ml) was separately generated using the positive control provided in the kit. Two separate triplicate sample groups were prepared. One group was analyzed without the addition of lyophilized amoebocyte lysate (LAL) as a negative control against signals due to the NPs.

A sterility test was conducted based on the WHO’s protocol [72, 73]. A stock solution was prepared for each NP at 1 mg/ml and diluted to a final concentration of 500 µg/ml in thioglycolate broth (USP Alternative). Each solution was incubated at 30 °C in a stationary oven for 21 days. 100 µg of each solution was taken on days 1, 3, 7, and 21, spread onto a Tryptic soy agar plate, and incubated in a stationary oven at 30 °C for 16 h. The agar plate was then visually inspected for any signs of bacteria colony formation. Data were collected in triplicate for analysis.

Cell culture

Ker-CTs from adult human skin were purchased from the American Type Culture Collection (ATCC CRL-4048™, USA). The Ker-CTs were cultured in EpiGRO™ medium (Merck) with 1% penicillin-streptomycin at 37 °C in 5% CO2. The cells were maintained for 2–4 days until 80% confluence was reached, harvested with 0.25% Trypsin-EDTA, and centrifuged at 290 g for 5 min. Next, the cell pellet was re-suspended in EpiGRO medium, and a hemacytometer was used to count the number of cells. Lastly, the cells were seeded at a density of 10,000 cells/cm2 and incubated at 37 °C in 5% CO2 for 24 h to allow complete attachment before further treatments were conducted.

Cell metabolism and proliferation assays

AlamarBlue Cell Viability Reagent (Invitrogen) was used to quantify cell viability. Cells were treated with the different NP groups for 24 h before washing them twice with PBS. 10% alamarBlue was diluted from stock solution in EpiGRO medium, and 120 µl were added into each well. The well plate was then incubated for 4 h, and 100 µl was aliquoted into a black 96-well plate. Fluorescence was measured at an excitation wavelength of 540 nm and emission wavelength of 590 nm using a microplate reader (Infinite M200, TECAN Inc).

A Quant-iT™ PicoGreen™ dsDNA Assay (Invitrogen) was used to quantify cell proliferation immediately after the alamarBlue assay. The remaining alamarBlue solution was removed, and each well was washed with PBS twice. 150 µl of DI water were added to each well, and the well plate was subjected to 2 freeze-thaw cycles. 100 µl of DI water were aliquoted into a new 96-black well plate, and 100 µl of PicoGreen solution were then added. Fluorescence was measured at an excitation wavelength 485 nm and emission wavelength of 538 nm using a microplate reader (Infinite M200, TECAN Inc).

ROS and mitochondria superoxide assays

CellROX™ was used to quantify the amount of ROS generated by the cells. Ker-CTs were seeded on Costar® 96-well clear-bottom black well plates and treated with the NPs for 8 h. The cells were then washed with PBS twice before administering CELLROX™ reagent and incubated for 30 min according to the manufacturer’s protocol. Hoechst 33342 was administered concurrently with the CELLROX™ reagent to stain cell nuclei. The cells were washed again with PBS twice and replaced in 100 µl of fresh PBS before analysis on a microplate reader.

MitoSOX™ was used to quantify the amount of mitochondria superoxide generated by the cells. Ker-CTs were seeded on Costar® 96-well clear-bottom black well plates and treated with the NPs for 8 h. Subsequently, the cells were washed with PBS twice before administering MitoSOX™ reagent and incubated for 30 min according to the manufacturer’s protocol. Hoechst 33342 was administered concurrently with the MitoSOX™ reagent to stain cell nuclei. The cells were washed again with PBS twice and replaced in 100 µl of fresh PBS before analysis on a microplate reader.

Immunofluorescence analysis

Immunofluorescence staining was used to detect any p-γH2AX due to double-stranded DNA breaks in the cells. Ker-CTs were seeded on glass chamber wells and treated with the different NPs for 24 h. Next, the medium was removed, and cells were washed with PBS twice. The cells were fixed with 4% paraformaldehyde and permeabilized with 0.2% Triton X-100.

Samples were then incubated with histone H2A.X polyclonal antibody (Invitrogen) overnight after blocking with a buffer of 2% bovine serum albumin with 0.1% Triton X-100. The secondary antibody and Hoescht 33342 were then incubated with the samples before mounting with a coverslip. The samples were imaged with a confocal microscope (Leica Stellaris 5).

Nanoparticle uptake study

Ker-CTs were seeded in 6-well plates and treated with each NP for 24 h. Cells were washed with PBS twice, and then trypsin was added. The resulting solution was centrifuged to obtain a cell pellet, which was fixed with 2% formaldehyde and 2.5% glutaraldehyde in PBS. Fixed samples were then dehydrated, embedded in araldite resin (Araldite 502, Electron Microscopy Studies), and sectioned with an ultramicrotome (Leica EM UC7)). Ultrathin sections were deposited onto a TEM copper grid and stained with UranyLess (UranyLess EM Stain 22409, Electron Microscopy Studies) and lead citrate (Lead Citrate 22410-01, Electron Microscopy Studies) before TEM imaging (JEOL 2100 F). Elemental mapping was conducted using STEM-EDX mode.

3D ex vivo skin explants

All experimental procedures involving donated patient samples were reviewed and approved by the National Healthcare Group Domain Specific Review Board (NHG DSRB ref: 2016/00370). A total of 8 female patients who underwent routine abdominoplasty were recruited. All patients were of Chinese ethnicity with ages of 50–60 years. Donated skin samples were collected immediately after surgery and brought to the laboratory for processing. Samples were preliminarily washed with PBS and 70% ethanol before thorough cleaning with PBS containing 2% antibiotic-antimycotic. Subcutaneous fat was removed with regular surgical tools, and then 6-mm skin biopsies were taken.

The dermis of each biopsy was trimmed down manually with scissors, and a wound was manually created on the epidermis. The epidermis was first pinched using fine tweezers to lift the epidermis, and the skin fold directly below the tweezers was cut with fine scissors to create an elliptical partial-thickness excisional wound. The skin explants were kept in culture at the air-liquid interface in a 6-well deep well plate containing tissue culture inserts. The cell culture medium consisted of Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 4 mM L-glutamine, 1% antibiotic-antimycotic, and 10% FBS.

A stock suspension of each NP group was prepared in DI water. The stock was diluted to the administered concentration in the same DMEM medium used for the skin explants. Next, 10 µL of the NP suspension were introduced to the wound bed, and complete DMEM of the same volume was used as vehicle control.

The cell culture medium has no direct relevance to the real-life scenario of TiO2 application, but it was necessary to prevent desiccation and maintain a cell-compatible environment to allow for wound-edge migration. Such an environment is also closer to wound or interstitial fluids compared to water or the sunscreen matrix. Additionally, using the cell culture media allows us to maintain the NPs at their most stable hydrodynamic size during the exposure. At 1 and 3 days after treatment, the skin biopsies were cut in half with a razor blade. Each half of the skin explant was placed in two separate fixing solutions of either 4% formaldehyde or 2% formaldehyde + 2.5% glutaraldehyde for 24 h.

Biopsy samples fixed in 4% formaldehyde was processed for histological analysis. Each biopsy sample was first dehydrated in a series of ascending ethanol concentrations of 70%, 80%, 90%, and 100% for 1 h each, followed by a second change to 100% ethanol overnight. The samples were then soaked in xylene for 7 h, followed by a second change overnight. Next, the samples were impregnated in paraffin wax at 60 °C for 8 h, followed by a second change overnight.

On the fourth day, each sample was embedded in paraffin blocks in cassettes. These blocks were then sectioned into 5-µm-thick samples using a microtome and collected on glass slides. The other half of the biopsy sample fixed in 2% formaldehyde + 2.5% glutaraldehyde was processed for TEM sectioning and imaging. The samples were post-fixed in a solution containing 2% osmium tetroxide (Merck) for 2 h and washed with DI water twice, followed by the protocol described above for TEM imaging.

H&E staining

Each glass slide was heated up on a slide warmer at 60 °C, placed in glass slide holders, and dewaxed using 3 rounds of xylene for 3 min each. Next, they were rehydrated in a series of descending ethanol concentrations of 100%, 90%, 80%, and 70% for 3 min each. Following this, the samples were stained with hematoxylin for 5 min, acid alcohol for 15 s, Scott’s tap water for 2 min, and eosin for 2 min, followed by a 30-second wash under running tap water. Lastly, the samples were dehydrated in a series of ascending ethanol concentrations of 70%, 80%, and 90% for 30 s each, followed by 100% ethanol for 3 min each. The glass slides were then soaked in 3 separate xylene washes for 3 min each, followed by mounting with a glass cover slip with Cytoseal™ 60 (Electron Microscopy Studies).

Statistical analysis

All data are presented as the mean ± standard deviation. Experiments were carried out in triplicate (n = 3). One-way analysis of variance (ANOVA) with Tukey’s post-hoc test was performed to compare means when there were more than two sample groups. All statistical analyzes were conducted using the software Origin 2023, and p < 0.05 was used to determine statistical differences in all analyzes.

Data availability

Data is provided within the manuscript or supplementary information files.

References

  1. Borm PJ, Robbins D, Haubold S, Kuhlbusch T, Fissan H, Donaldson K, et al. The potential risks of nanomaterials: a review carried out for ECETOC. Part Fibre Toxicol. 2006;3(1):11.

    Article  PubMed  PubMed Central  Google Scholar 

  2. Global Market Insights Inc. [Internet]. [cited 2023 Sep 11]. Titanium Dioxide Market Size and Share | Industry Trends – 2026. https://www.gminsights.com/industry-analysis/titanium-dioxide-market

  3. Calafat AM, Wong LY, Ye X, Reidy JA, Needham LL. Concentrations of the sunscreen agent benzophenone-3 in residents of the United States: National Health and Nutrition Examination Survey 2003–2004. Environ Health Perspect. 2008;116(7):893–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. Downs CA, Kramarsky-Winter E, Segal R, Fauth J, Knutson S, Bronstein O, et al. Toxicopathological effects of the sunscreen UV filter, oxybenzone (benzophenone-3), on coral planulae and cultured primary cells and its environmental contamination in Hawaii and the US Virgin Islands. Arch Environ Contam Toxicol. 2016;70(2):265–88.

    Article  CAS  PubMed  Google Scholar 

  5. Dermatology MA. of. Hawaii Sunscreen Ban Actions [Internet]. 2018. https://www.massacademyofdermatology.org/news/403915/Hawaii-Sunscreen-Ban-Actions.htm

  6. Administration F. and D. Sunscreen Drug Products for Over-the-Counter Human Use [Internet]. 2019. https://www.federalregister.gov/documents/2019/02/26/2019-03019/sunscreen-drug-products-for-over-the-counter-human-use

  7. Nel A, Xia T, Mädler L, Li N. Toxic potential of materials at the nanolevel. Science. 2006;311(5761):622–7.

    Article  CAS  PubMed  Google Scholar 

  8. Hadjidemetriou M, Kostarelos K. Evolution of the nanoparticle corona. Nat Nanotechnol. 2017;12(4):288–90.

    Article  CAS  PubMed  Google Scholar 

  9. Lu Q, Yu H, Zhao T, Zhu G, Li X. Nanoparticles with transformable physicochemical properties for overcoming biological barriers. Nanoscale. 2023.

  10. Zhang J, Guo W, Li Q, Wang Z, Liu S. The effects and the potential mechanism of environmental transformation of metal nanoparticles on their toxicity in organisms. Environ Sci Nano. 2018;5(11):2482–99.

    Article  CAS  Google Scholar 

  11. Carnovale C, Bryant G, Shukla R, Bansal V. Identifying trends in Gold Nanoparticle toxicity and uptake: size, shape, capping Ligand, and Biological Corona. ACS Omega. 2019;4(1):242–56.

    Article  CAS  Google Scholar 

  12. Ding L, Yao C, Yin X, Li C, Huang Y, Wu M, et al. Size, shape, and Protein Corona Determine Cellular Uptake and removal mechanisms of gold nanoparticles. Small. 2018;14(42):1801451.

    Article  Google Scholar 

  13. Liu N, Tang M, Ding J. The interaction between nanoparticles-protein corona complex and cells and its toxic effect on cells. Chemosphere. 2020;245:125624.

    Article  CAS  PubMed  Google Scholar 

  14. Rabel M, Warncke P, Thürmer M, Grüttner C, Bergemann C, Kurland HD, et al. The differences of the impact of a lipid and protein corona on the colloidal stability, toxicity, and degradation behavior of iron oxide nanoparticles. Nanoscale. 2021;13(20):9415–35.

    Article  CAS  PubMed  Google Scholar 

  15. Gautam A, Rakshit M, Nguyen KT, Kathawala MH, Nguyen LTH, Tay CY, et al. Understanding the implications of engineered nanoparticle induced autophagy in human epidermal keratinocytes in vitro. NanoImpact. 2019;15:100177.

    Article  Google Scholar 

  16. Zhao Y, Howe JL, Yu Z, Leong DT, Chu JJH, Loo JSC, et al. Exposure to titanium dioxide nanoparticles induces autophagy in primary human keratinocytes. Small. 2013;9(3):387–92.

    Article  PubMed  Google Scholar 

  17. Nutrition C, for FS. and A. Guidance for Industry: Safety of Nanomaterials in Cosmetic Products [Internet]. FDA; 2020 [cited 2024 Feb 9]. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/guidance-industry-safety-nanomaterials-cosmetic-products

  18. Liang Y, Simaiti A, Xu M, Lv S, Jiang H, He X, et al. Antagonistic skin toxicity of Co-exposure to Physical Sunscreen ingredients Zinc Oxide and Titanium Dioxide nanoparticles. Nanomaterials. 2022;12(16):2769.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Wills JW, Hondow N, Thomas AD, Chapman KE, Fish D, Maffeis TG, et al. Genetic toxicity assessment of engineered nanoparticles using a 3D in vitro skin model (EpiDermTM). Part Fibre Toxicol. 2015;13(1):1–21.

    Article  Google Scholar 

  20. Martins RM, de Siqueira Martins S, Barbosa GLF, Fonseca MJV, Rochette PJ, Moulin VJ, et al. Photoprotective effect of solid lipid nanoparticles of rutin against UVB radiation damage on skin biopsies and tissue-engineered skin. J Microencapsul. 2022;39(7–8):668–79.

    Article  CAS  PubMed  Google Scholar 

  21. Ailawadi S, Talreja R, Panstingel N, Sulentic C. Titanium Dioxide and Zinc Oxide nanoparticles in Sunscreen: potential impact on Cytokine expression in human skin pre. -and Post-UVB Exposure; 2022.

  22. Silica Specialist. - Evonik Industries [Internet]. [cited 2023 Sep 20]. https://www.silica-specialist.com/en

  23. Sabuncu AC, Grubbs J, Qian S, Abdel-Fattah TM, Stacey MW, Beskok A. Probing nanoparticle interactions in cell culture media. Colloids Surf B Biointerfaces. 2012;95:96–102.

    Article  CAS  PubMed  Google Scholar 

  24. Chen CW, Huang JH, Lai TC, Jan YH, Hsiao M, Chen CH, et al. Evaluation of the intracellular uptake and cytotoxicity effect of TiO 2 nanostructures for various human oral and lung cells under dark conditions. Toxicol Res. 2016;5(1):303–11.

    Article  CAS  Google Scholar 

  25. Popov AP, Priezzhev A, Lademann J, Myllylä R. TiO2 nanoparticles as an effective UV-B radiation skin-protective compound in sunscreens. J Phys Appl Phys. 2005;38(15):2564.

    Article  CAS  Google Scholar 

  26. Merck. IR Spectrum Table & Chart [Internet]. https://www.sigmaaldrich.com/SG/en/technical-documents/technical-article/analytical-chemistry/photometry-and-reflectometry/ir-spectrum-table

  27. SolaveilTM CT-12 W by Croda. - Personal Care & Cosmetics [Internet]. [cited 2023 Sep 20]. https://www.ulprospector.com/en/eu/PersonalCare/Detail/1409/6611765/Solaveil-CT-12W

  28. Egerton TA, Everall NJ, Tooley IR. Characterization of TiO2 nanoparticles surface modified with aluminum stearate. Langmuir. 2005;21(7):3172–8.

    Article  CAS  PubMed  Google Scholar 

  29. Lee D, Condrate Sr R. An FTIR spectral investigation of the structural species found on alumina surfaces. Mater Lett. 1995;23(4–6):241–6.

    Article  CAS  Google Scholar 

  30. Ahmed MH, Byrne JA, McLaughlin J, Elhissi A, Ahmed W. Comparison between FTIR and XPS characterization of amino acid glycine adsorption onto diamond-like carbon (DLC) and silicon doped DLC. Appl Surf Sci. 2013;273:507–14.

    Article  CAS  Google Scholar 

  31. Luna-López J, Carrillo-López J, Aceves-Mijares M, Morales-Sánchez A y, Falcony C. FTIR and photoluminescence of annealed silicon rich oxide films. Superf Vacío. 2009;22(1):11–4.

    Google Scholar 

  32. Flinn ACT, BD. INFRARED SPECTROSCOPY: A POTENTIAL QUALITY ASSURANCE METHOD FOR COMPOSITE BONDING SURFACE PREPARATION [Internet]. https://depts.washington.edu/amtas/events/jams_12/papers/paper-flinn.pdf

  33. DeLoid GM, Cohen JM, Pyrgiotakis G, Demokritou P. Preparation, characterization, and in vitro dosimetry of dispersed, engineered nanomaterials. Nat Protoc. 2017;12(2):355–71.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Lorenz C, Von Goetz N, Scheringer M, Wormuth M, Hungerbühler K. Potential exposure of German consumers to engineered nanoparticles in cosmetics and personal care products. Nanotoxicology. 2011;5(1):12–29.

    Article  CAS  PubMed  Google Scholar 

  35. Administration USF. D. Bacterial Endotoxins/Pyrogens [Internet]. https://www.fda.gov/inspections-compliance-enforcement-and-criminal-investigations/inspection-technical-guides/bacterial-endotoxinspyrogens

  36. Di Giosia M, Zerbetto F, Calvaresi M. Incorporation of Molecular nanoparticles Inside proteins: the Trojan Horse Approach in Theranostics. Acc Mater Res. 2021;2(8):594–605.

    Article  Google Scholar 

  37. Skin Penetration and Cellular Uptake of Amorphous Silica Nanoparticles with Variable Size. Surface Functionalization, and Colloidal Stability | ACS Nano [Internet]. [cited 2023 Oct 13]. https://pubs.acs.org/doi/abs/https://doiorg.publicaciones.saludcastillayleon.es/10.1021/nn301622h

  38. Harush-Frenkel O, Debotton N, Benita S, Altschuler Y. Targeting of nanoparticles to the clathrin-mediated endocytic pathway. Biochem Biophys Res Commun. 2007;353(1):26–32.

    Article  CAS  PubMed  Google Scholar 

  39. de Almeida MS, Susnik E, Drasler B, Taladriz-Blanco P, Petri-Fink A, Rothen-Rutishauser B. Understanding nanoparticle endocytosis to improve targeting strategies in nanomedicine. Chem Soc Rev. 2021;50(9):5397–434.

    Article  Google Scholar 

  40. Horie M, Sugino S, Kato H, Tabei Y, Nakamura A, Yoshida Y. Does photocatalytic activity of TiO2 nanoparticles correspond to photo-cytotoxicity? Cellular uptake of TiO2 nanoparticles is important in their photo-cytotoxicity. Toxicol Mech Methods. 2016;26(4):284–94.

    Article  CAS  PubMed  Google Scholar 

  41. Albanese A, Tang PS, Chan WC. The effect of nanoparticle size, shape, and surface chemistry on biological systems. Annu Rev Biomed Eng. 2012;14(1):1–16.

    Article  CAS  PubMed  Google Scholar 

  42. Cetaphil Sunscreen with SPF. 50+ | Cetaphil Singapore [Internet]. [cited 2023 Sep 20]. https://www.cetaphil.com.sg/sunscreens/cetaphil-sun-spf-50%2B-light-gel/SG_7612076600645.html

  43. Cetaphil Sun SPF. 50 + Light Gel | Cetaphil Singapore [Internet]. [cited 2023 Sep 20]. https://www.cetaphil.com.sg/sunscreens/cetaphil-sun-spf-50%2B-light-gel/SG_7612076600645.html

  44. Lundqvist M, Stigler J, Elia G, Lynch I, Cedervall T, Dawson KA. Nanoparticle size and surface properties determine the protein corona with possible implications for biological impacts. Proc Natl Acad Sci. 2008;105(38):14265–70.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. Clogston JD, Patri AK. Zeta potential measurement. Charact Nanopart Intend Drug Deliv. 2011;63–70.

  46. Agency UEP. Nanomaterial Case Studies: Nanoscale Titanium Dioxide in Water Treatment and in Topical Sunscreen [Internet]. https://ofmpub.epa.gov/eims/eimscomm.getfile?p_download_id=499980

  47. Zhang LW, Monteiro-Riviere NA. Toxicity assessment of six titanium dioxide nanoparticles in human epidermal keratinocytes. Cutan Ocul Toxicol. 2019;38(1):66–80.

    Article  CAS  PubMed  Google Scholar 

  48. Vales G, Rubio L, Marcos R. Long-term exposures to low doses of titanium dioxide nanoparticles induce cell transformation, but not genotoxic damage in BEAS-2B cells. Nanotoxicology. 2015;9(5):568–78.

    Article  CAS  PubMed  Google Scholar 

  49. Tohgasaki T, Ozawa N, Yoshino T, Ishiwatari S, Matsukuma S, Yanagi S, et al. Enolase-1 expression in the stratum corneum is elevated with parakeratosis of atopic dermatitis and disrupts the cellular tight junction barrier in keratinocytes. Int J Cosmet Sci. 2018;40(2):178–86.

    Article  CAS  PubMed  Google Scholar 

  50. Warheit DB, Webb TR, Reed KL, Frerichs S, Sayes CM. Pulmonary toxicity study in rats with three forms of ultrafine-TiO2 particles: Differential responses related to surface properties. Toxicology. 2007;230(1):90–104.

    Article  CAS  PubMed  Google Scholar 

  51. Serpone N, Salinaro A, Horikoshi S, Hidaka H. Beneficial effects of photo-inactive titanium dioxide specimens on plasmid DNA, human cells and yeast cells exposed to UVA/UVB simulated sunlight. J Photochem Photobiol Chem. 2006;179(1):200–12.

    Article  CAS  Google Scholar 

  52. Xia T, Kovochich M, Brant J, Hotze M, Sempf J, Oberley T, et al. Comparison of the abilities of ambient and manufactured nanoparticles to induce cellular toxicity according to an oxidative stress paradigm. Nano Lett. 2006;6(8):1794–807.

    Article  CAS  PubMed  Google Scholar 

  53. Ehrenberg MS, Friedman AE, Finkelstein JN, Oberdörster G, McGrath JL. The influence of protein adsorption on nanoparticle association with cultured endothelial cells. Biomaterials. 2009;30(4):603–10.

    Article  CAS  PubMed  Google Scholar 

  54. Francia V, Yang K, Deville S, Reker-Smit C, Nelissen I, Salvati A. Corona composition can affect the mechanisms cells use to internalize nanoparticles. ACS Nano. 2019;13(10):11107–21.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  55. Fleischer CC, Payne CK. Secondary structure of corona proteins determines the cell surface receptors used by nanoparticles. J Phys Chem B. 2014;118(49):14017–26.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  56. Dunnill C, Patton T, Brennan J, Barrett J, Dryden M, Cooke J, et al. Reactive oxygen species (ROS) and wound healing: the functional role of ROS and emerging ROS-modulating technologies for augmentation of the healing process. Int Wound J. 2017;14(1):89–96.

    Article  PubMed  Google Scholar 

  57. Aggarwal P, Hall JB, McLeland CB, Dobrovolskaia MA, McNeil SE. Nanoparticle interaction with plasma proteins as it relates to particle biodistribution, biocompatibility and therapeutic efficacy. Adv Drug Deliv Rev. 2009;61(6):428–37.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  58. Toyooka T, Amano T, Ibuki Y. Titanium dioxide particles phosphorylate histone H2AX independent of ROS production. Mutat Res Toxicol Environ Mutagen. 2012;742(1–2):84–91.

    Article  CAS  Google Scholar 

  59. Ng KW, Khoo SPK, Heng BC, Setyawati MI, Tan EC, Zhao X, et al. The role of the tumor suppressor p53 pathway in the cellular DNA damage response to zinc oxide nanoparticles. Biomaterials. 2011;32(32):8218–25.

    Article  CAS  PubMed  Google Scholar 

  60. Fubini B, Fenoglio I, Tomatis M, Turci F. Effect of chemical composition and state of the surface on the toxic response to high aspect ratio nanomaterials. Nanomed. 2011;6(5):899–920.

    Article  CAS  Google Scholar 

  61. Technology NI. of S and. Standard Reference Material® 1898 Titanium Dioxide Nanomaterial [Internet]. 2020. https://www-s.nist.gov/srmors/certificates/1898.pdf

  62. Evonik. AEROXIDE® TiO2 P 25 [Internet]. https://www.silica-specialist.com/en/product/PR_52000356

  63. Natpure® Screen TWG by SENSIENT BEAUTY - Personal Care &. Cosmetics [Internet]. [cited 2023 Sep 20]. https://www.ulprospector.com/en/na/PersonalCare/Detail/817/991552/Natpure-Screen-TWG

  64. Benítez-Martínez S, López-Lorente ÁI, Valcárcel M. Determination of TiO 2 nanoparticles in sunscreen using N-doped graphene quantum dots as a fluorescent probe. Microchim Acta. 2016;183(2):781–9.

    Article  Google Scholar 

  65. Nischwitz V, Goenaga-Infante H. Improved sample preparation and quality control for the characterisation of titanium dioxide nanoparticles in sunscreens using flow field flow fractionation on-line with inductively coupled plasma mass spectrometry. J Anal Spectrom. 2012;27(7):1084–92.

    Article  CAS  Google Scholar 

  66. Philippe A, Košík J, Welle A, Guigner JM, Clemens O, Schaumann GE. Extraction and characterization methods for titanium dioxide nanoparticles from commercialized sunscreens. Environ Sci Nano. 2018;5(1):191–202.

    Article  CAS  Google Scholar 

  67. Schneider CA, Rasband WS, Eliceiri KW. NIH Image to ImageJ: 25 years of image analysis. Nat Methods. 2012;9(7):671–5.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  68. Miquel-Jeanjean C, Crépel F, Raufast V, Payre B, Datas L, Bessou-Touya S, et al. Penetration study of formulated Nanosized Titanium Dioxide in models of Damaged and Sun-irradiated skins. Photochem Photobiol. 2012;88(6):1513–21.

    Article  CAS  PubMed  Google Scholar 

  69. Park SB, Jung WH, Kim KY, Koh B. Toxicity assessment of SiO2 and TiO2 in normal colon cells, in vivo and in human colon organoids. Molecules. 2020;25(16):3594.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  70. Cao W, Liu M, Wang C, Jing G, Cao Y. Influences of repeated exposure to low levels of TiO2 nanoparticles on Kruppel-like factors in 3D Caco‐2 spheroids and mouse intestines. J Appl Toxicol. 2023;43(5):706–18.

    Article  CAS  PubMed  Google Scholar 

  71. Filipe P, Silva J, Silva R, Cirne de Castro J, Marques Gomes M, Alves L, et al. Stratum corneum is an effective barrier to TiO2 and ZnO nanoparticle percutaneous absorption. Skin Pharmacol Physiol. 2009;22(5):266–75.

    Article  CAS  PubMed  Google Scholar 

  72. Organization WH. 3.2 TEST FOR STERILITY [Internet]. 2012. https://www.who.int/medicines/publications/pharmacopoeia/TestForSterility-RevGenMethod_QAS11-413FINALMarch2012.pdf

  73. Beltran-Huarac J, Zhang Z, Pyrgiotakis G, DeLoid G, Vaze N, Demokritou P. Development of reference metal and metal oxide engineered nanomaterials for nanotoxicology research using high throughput and precision flame spray synthesis approaches. NanoImpact. 2018;10:26–37.

    Article  PubMed  Google Scholar 

Download references

Acknowledgements

The authors acknowledge the Facility for Analysis, Characterization, Testing and Simulation (FACTS), NTU, for the support in electron microscopy analysis.

Funding

This research was supported by the Ministry of Education, Singapore (AcRF Tier 1: RG10/20, RG7/22). DYDK is a recipient of the Interdisciplinary Graduate Programme – Health Technologies Research Scholarship from Nanyang Technological University.

Author information

Authors and Affiliations

  1. School of Materials Science and Engineering, Nanyang Technological University, Singapore, Singapore

    Darien Yu De Kwek, Magdiel Inggrid Setyawati, Archana Gautam, Sunil S. Adav & Kee Woei Ng

  2. Interdisciplinary Graduate Programme (IGP), Graduate College, Singapore, Singapore

    Darien Yu De Kwek

  3. Tan Tock Seng Hospital, Singapore, Singapore

    Ee Cherk Cheong

  4. Nanyang Environment and Water Research Institute (NEWRI), Singapore, Singapore

    Kee Woei Ng

  5. Skin Research Institute of Singapore (SRIS), Singapore, Singapore

    Kee Woei Ng

Authors
  1. Darien Yu De Kwek
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Magdiel Inggrid Setyawati
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Archana Gautam
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Sunil S. Adav
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Ee Cherk Cheong
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Kee Woei Ng
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

DYDK designed, carried out the study, and wrote the manuscript. MIS and AG provided technical support and assisted in the 3D skin explant experiments. SSA provided technical support and helped to analyse the LCMS/MS data. ECC provided technical support in the 3D skin explant experiments, analysed the data and reviewed the manuscript. KWN supervised the study, planned the experiments, analysed the data and reviewed the manuscript.

Corresponding author

Correspondence to Kee Woei Ng.

Ethics declarations

Ethics approval and consent to participate

All experimental procedures involving donated patient samples were reviewed and approved by the National Healthcare Group Domain Specific Review Board (NHG DSRB ref: 2016/00370).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

De Kwek, D.Y., Setyawati, M.I., Gautam, A. et al. Understanding the toxicological effects of TiO2 nanoparticles extracted from sunscreens on human keratinocytes and skin explants. Part Fibre Toxicol 21, 49 (2024). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12989-024-00610-9

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12989-024-00610-9

Keywords

Particle and Fibre Toxicology

ISSN: 1743-8977

Contact us