Molecular Mechanisms of Zinc Oxide Nanoparticle-Induced Genotoxicity Short Running Title: Genotoxicity of ZnO NPs

Background: Zinc oxide nanoparticles (ZnO NPs) are among the most frequently applied nanomaterials in consumer products. Evidence exists regarding the cytotoxic effects of ZnO NPs in mammalian cells; however, knowledge about the potential genotoxicity of ZnO NPs is rare, and results presented in the current literature are inconsistent. Objectives: The aim of this review is to summarize the existing data regarding the DNA damage that ZnO NPs induce, and focus on the possible molecular mechanisms underlying genotoxic events. Methods: Electronic literature databases were systematically searched for studies that report on the genotoxicity of ZnO NPs. Results: Several methods and different endpoints demonstrate the genotoxic potential of ZnO NPs. Most publications describe in vitro assessments of the oxidative DNA damage triggered by dissoluted Zn2+ ions. Most genotoxicological investigations of ZnO NPs address acute exposure situations. Conclusion: Existing evidence indicates that ZnO NPs possibly have the potential to damage DNA. However, there is a lack of long-term exposure experiments that clarify the intracellular bioaccumulation of ZnO NPs and the possible mechanisms of DNA repair and cell survival.


Introduction
Over the past 15 years, nanotechnology has increasingly gained in importance in industry, biomedicine, and research. According to the current definition of the European Union (EU), nanomaterials are natural, incidental, or manufactured materials that contain particles in an unbound state, either as aggregates or as agglomerates. At least 50% of these particles must exhibit one or more external dimension within the size range of 1-100 nm [1]. Surface properties become more important as a function of the size reduction of a material. Thus, nanoparticles (NPs) have completely different mechanical, optical, electrical, magnetic, and catalytic properties compared with larger particles of the same composition. Hence, the bioactivity of NPs significantly differs from that of their fine-size analogues [2]. Zinc oxide (ZnO) NPs are among the most commonly used nanomaterials in industrial applications. Despite their increasing usage in consumer products, the safety aspects of ZnO NPs remain uncertain. In particular, information regarding the possible genotoxicity of ZnO NPs is rare, and partially contradictory. The aim of this review is to summarize the literature published between 2009 and 2017 that covers the genotoxicity of ZnO NPs in mammalian and non-mammalian in vitro and in vivo systems, and to estimate the current risk of using ZnO NPs in consumer products. Furthermore, information on the molecular mechanisms of ZnO NP-induced DNA damage will also be outlined and discussed.

Exposure Routes
For the toxicological evaluation of NPs, knowledge regarding the routes of intake is essential. Knowledge regarding its bioavailability and resorption is also important. Possible intake routes of NPs in humans are the gastrointestinal tract, the skin, and the airways. For consumers, dermal exposure is the most likely way to come into contact with ZnO NPs due to the high number of cosmetic products containing ZnO NPs. The stratum corneum, known as the upper layer of the skin, seems to be a sufficient barrier against ZnO NP penetration into the epidermis, as shown by several authors [6,7]. It was clearly demonstrated that ZnO NPs were not able to penetrate healthy and intact human or porcine skin. Although NPs may be retained in the hair follicle ostium or skin folds, they are usually sufficiently eliminated by sebum flow [8]. However, skin damages, for example after excessive sun bathing, may harm this protection layer, and lead to possible toxicological effects from NPs. Cytotoxic or genotoxic effects only seem to be relevant in proliferating cells, which can be found in the basal layers of the epidermis. This is why the application of ZnO NPs to injured or defective skin is discussed as being potentially dangerous. The ingestion of ZnO NPs and contact with intestinal mucosa must be evaluated equally. In particular, chronic intestinal illness may lead to a defect in the mucosa barrier, which consequently may lead to an enhanced toxicity. Further studies are needed to evaluate the correlation between the grade of skin damage and the hazard of ZnO NPs.
Airway exposure via inhalation is the predominant means of contact for workers in the chemical, cosmetic, or paint industries [4]. Nanosized particles are able to reach the peripheral airway sites, such as the bronchiolar and alveolar regions. If not carried away by mucociliary transport mechanisms, NPs may affect alveolar cells and cause toxic, genotoxic, or inflammatory effects [4]. Inhalation exposure to ZnO NPs seems to be an important hazard, and risk assessment is urgently needed within this context [9]. Indeed, the airway exposure of NPs seems to be very important in the toxicological circumstances. According to Vermylen et al., the intake of superfine structures via inhalation has profound negative local and systemic side effects, such as an enhanced risk of cardiovascular diseases [10]. These very small particles are able to penetrate the tracheobronchial tree. In particular, ultrafine particles, which have a diameter less than 100 nm, are able to pass directly into the blood stream [10,11]. Some studies hypothesize that NPs might be able to reach the brain along peripheral nerves [12,13]. This may offer a therapeutic option as well. However, toxicological evaluations are warranted.

Genotoxicity of ZnO NPs
The difference between the volume and surface of NPs enables their variety of chemical, physical, and biological properties [14]. Due to their small size, large surface area, and physicochemical characteristics, NPs may exhibit unpredictable genotoxic properties. The biological properties depend on the manufacturing procedure, agglomeration and aggregation tendencies, and surface coating. During the manufacturing processes, the particle diameters are not homogeneous. Due to their surface, NPs tend to aggregate, which implicates the need for dispersions. Surface coating is a suitable method for preventing the aggregation of NPs. These above-mentioned circumstances significantly influence the toxicity of NPs. Kwon et al. showed that small NPs cross the cellular membranes more easily, which leads to an increased potency of DNA damage. Accumulated NPs might be internalized into the cell mainly during the mitosis process. According to Liu et al., a crucial determinant of toxicity is the solubility of ZnO NPs, which is influenced by various factors, including the pH of the environment in tissues, cells, and organelles [15]. ZnO NPs and other particles such as silver are soluble, and may release ions. Unlike silver, Zn is an important component of several enzymes and transcription factors in the human body. After incorporation, ZnO NPs may dissolve into Zn 2+ and trigger several signaling pathways and cascades, which might lead to an enhanced influx of calcium, gene upregulation, or the release of pro-inflammatory markers [16]. The solubility of NPs such as silver (Ag), copper (Cu), or ZnO is one of the main contributors to their toxicity. Ag, Cu, and ZnO NPs have some commonalities. Their elemental composition is metallic, they fight the growth of microorganisms, they have a negative surface charge, and most importantly, all of them are soluble [17]. Nevertheless, there are also differences between these metallic particles. According to Bondarenko, although their particle size is similar, their toxicity is likely different. Cu ions may be involved in electron-transfer processes, in contrast to Ag and Zn [17].
According to Golbamaki, the genotoxic effects of NPs may be classified as either "primary genotoxicity" or "secondary genotoxicity". Reactive oxygen species (ROS) generation during particle-induced inflammation is the cause of secondary genotoxicity, whereas primary genotoxicity refers to genotoxic effects without inflammation [18]. There are studies that point to the correlation between particle size and toxicity [9]. However, information concerning the size dependency of NP-induced toxicity is contradictory. Warheit et al. did not observe any variation in the toxicity levels of large and small TiO2 NPs [19]. However, Golbamaki and Karlsson detected significantly increased DNA damage after cell exposure to larger micrometer-sized particles compared with smaller NPs [18,20]. Due to these inhomogeneous statements, the size dependency of nanotoxicity and nanogenotoxicity needs to be clarified. NP size must always be characterized exactly in order to provide comparable data in the context of the current literature.
Over the past 10 years, studies focusing on the nanotoxicity of ZnO have been continuously published. However, most of these studies primarily address the cytotoxic aspects of ZnO NPs. Dose-response correlations between ZnO NP concentration and cellular viability are investigated in most studies. However, DNA damage occurs at significantly lower concentrations compared with cytotoxic effects. Hence, genotoxicological evaluations of NPs must be performed at non-cytotoxic doses. Although ZnO NPs are frequently applied in industry and research, data on the genotoxic potential of this material is quite limited [21].

Molecular Mechanisms of Genotoxicity and Evaluation of Oxidative DNA Damage
It is crucial to understand the molecular mechanisms of genotoxicity caused by ZnO NPs in order to provide a valid risk assessment. Although several groups have contributed data towards elucidating these pathways, the associated mechanisms and correlations still remain unclear. The role of Zn ions cannot definitely be ruled out at this stage. Auffan et al. showed that chemically stable metallic nanoparticles have no significant cellular toxicity, whereas nanoparticles that are able to be oxidized, reduced, or dissolved are cytotoxic and genotoxic for cellular organisms [22]. Results from the Wuerzburg group suggest a correlation between ion concentration and genotoxic effects [23], but other groups could not confirm these results in several test systems (micronucleus test, comet assay, and γ H2AX) in a human neural cell line [24].
Autophagy is a lysosome-dependent degradation process that is usually activated in stress situations. Roy et al. identified autophagy activation as a major modulator of ZnO NP-induced cellular toxicity [25]. The detection of increased autophagosome formation and several autophagy marker proteins was reported. ROS generation was identified to be a major trigger for the induction of autophagy. Antioxidant enzymes inhibited cell death and reduced autophagy marker protein expression. The important role of autophagy in ZnO NP-induced toxicity was demonstrated by our group as well. Similar to the results reported by Roy et al., cellular damage could be reduced by counteracting oxidative stress and autophagy [26]. The correlation between autophagosome formation and apoptosis is controversially discussed in the literature. According to Vessoni et al., autophagy is a reaction to DNA damage, and plays an ambiguous role in regulating cell fate [27]. On the one hand, autophagy may promote cell protection, e.g., by degrading pro-apoptotic proteins or by supporting DNA repair. On the other hand, autophagy may also lead to cytotoxic events through the degradation of anti-apoptotic and DNA repair-related proteins [28]. In fact, ZnO NP-induced oxidative DNA damage stimulates autophagy pathways, and thus may influence the balance between cell survival and cytotoxicity. Pati et al. demonstrated an inhibition of DNA repair mechanisms. The reduction in the macrophage cell viability was due to the arrest of the cell cycle at the G0/G1 phase, the inhibition of superoxide dismutase, catalase, and eventually ROS [29].
Kononenko et al. demonstrated a concentration-dependent genotoxicity and cytotoxicity. DNA and chromosomal damage was accompanied by a reduction of glutathione S-transferase and catalase activity [30].
The amount of DNA damage does not only depend on the tested NP itself, but also on the exposed target cell, and the cell's genetic and proteomic properties in particular. ZnO NPs were shown to activate the p53 pathway by several groups [25,31,32]. ZnO NP-induced DNA damage should usually be forced by p53-associated apoptosis. Ng et al. examined a p53 knockdown fibroblast cell line exposed to ZnO NPs, and found a resistance to ZnO NP-mediated apoptosis, as well as a progressive cellular proliferation, indicating a possible first step to carcinogenesis.
The photogenotoxicity of ZnO NPs is a very important topic. UV irradiation was shown to enhance the cytotoxic properties of ZnO NPs in the A549 cell line by Yang and Ma [33]. Wang et al. reported on the oxidative DNA damage induced by ZnO NPs during UVA (ultraviolet) and visible light irradiation in a dose-dependent manner in HaCaT human skin keratinocytes [34]. The authors proclaimed a photogenotoxic potential of ZnO NPs in combination with UV light. These findings must be discussed critically, especially with respect to the use of ZnO NPs in sunscreen products. Contrary results were published by Demir et al., who demonstrated ZnO NP-related DNA damage in human and mouse cell lines using the micronucleus test and comet assay [35]. Furthermore, they observed anchorage-independent cell growth after NP exposure, which can be interpreted as an initial step towards malignant cell transformation. However, UVB exposure antagonized these effects. Future research projects can be expected to illuminate the interactions between UV light and ZnO NPs regarding DNA damage or DNA protection.
Certainly, a detailed characterization of the physicochemical properties of ZnO NPs is crucial in order to understand the partially divergent statements in the literature. Bhattacharya et al. underscored the important role of the physical properties of NPs. They showed that rod-shaped ZnO NPs induced significantly more DNA damage in peripheral blood mononuclear cells compared with spherical NPs [36]. Coatings may also influence the genotoxic potential of ZnO NPs, as shown by Yin et al., who demonstrated the extended DNA damage of NPs coated with poly (methacrylic acid) (PMAA) compared with uncoated particles [37]. The surface activity and large surface area of NPs lead to a high sorption capability, and thus induce further toxic effects. NPs can function as carriers of absorbed toxic substances, and thus enhance their bioavailability [38]. The majority of the current data regarding the genotoxic effects of nanoparticulate ZnO are based on in vitro investigations. In cells, NPs induce inflammation, genotoxic effects, and damages via the generation of reactive oxygen species (ROS). Sharma et al. published several studies on the genotoxicity of ZnO NPs in a variety of cell systems. They observed DNA damage using the single cell microgel electrophoresis (comet) assay in the HepG2 human liver cell line and the A-431 human epidermal cell line. Cells were exposed to ZnO NPs for 6 h [39,40]. The generation of ROS was demonstrated and discussed as a possible trigger of in vitro genotoxicity in both studies. Patel et al. found the generation of ROS in the A-431 cell line following the application of ZnO NPs. In this publication, ZnO NPs induced cell death, as well as a cell cycle arrest in the S and G2/M phase [41]. Tyrosine phosphorylation was shown to be another promoter of DNA damage in HepG2 cells [42]. Transmission and scanning electron microscopy are the usual tools for the investigation of cellular NP uptake, although these methods are quite time-consuming and technically challenging. Condello et al. demonstrated the entrance of ZnO NPs into human colon carcinoma cells, either by passive diffusion, endocytosis, or both. The entrance mode was dependent on the agglomeration state of the nanomaterial [43]. Toduka et al. used side-scattered light in flow cytometry as an indicator of NP uptake into mammalian cells [44]. Several nanomaterials were tested, including ZnO NPs in Chinese hamster ovary (CHO)-K1 cells using this method. Particles were internalized into the cells, and thus induced a high ROS production, which was directly correlated with the genotoxic events shown by the generation of the phosphorylated histone γ-H2AX. The co-cultivation with the antioxidant N-acetylcysteine (NAC) counteracted DNA damage. Kermanizadeh et al. also showed the important role of oxidative stress through demonstrating a suppression of the toxic potential of ZnO NPs by the antioxidant Trolox in a hepatocyte cell line [45]. DNA damage and pro-inflammatory IL-8 production were induced by oxidative stress and ROS generation. Other groups also published similar results demonstrating the positive correlation between oxidative stress and DNA damage [46,47]. The generation of ROS was mainly assessed by the dichloro-dihydro-fluorescein diacetate (DCFH-DA) assay. Various markers for oxidative stress were evaluated, e.g., glutathione (GSH) reduction, elevated gluthatione, malondialdehyde, superoxide dismutase, and catalase. The photogenotoxicity of ZnO NPs, including a high cellular uptake, was shown in Allium cepa [48]. Other groups also demonstrated the connection between DNA damage and ROS production [43,49].
Most of the studies on nanogenotoxicity were performed using cell lines instead of primary cells. Due to high interindividual variation and the difficulty of standardizing cellular harvest, repetitive experiments with large numbers of patients are necessary in order to assess representative data on primary cells. However, primary cells are neither immortalized nor transformed. Thus, the similarity to cells within the origin tissue is usually higher compared with transformed cell lines. This is why studies with primary cells are supplementary to those using standardized cell lines, and can contribute to common knowledge on nanogenotoxicology. Sharma et al. presented a study using primary human epidermal keratinocytes, a relevant target cell for ZnO NPs, which are mainly used in cosmetics applied to the human skin. ZnO NPs were internalized by the cells, as shown by transmission electron microscopy, where they induced a DNA fragmentation after 6 h of exposure at a concentration of 8 µg/mL [50]. Our own group used primary human nasal mucosa to evaluate the genotoxicity of ZnO NPs. Nasal mucosa belongs to the most important primary contact regions of humans with volatile xenobiotics. Cells of the nasal mucosa are representative of the entire human upper aerodigestive tract. Distinct three-dimensional cell culture systems serve to imitate the in vivo situation quite closely [51]. The genotoxic potential of ZnO NPs was proven in human nasal mucosa cells in an air-liquid interface cell culture, as well as by the extended secretion of IL-816. Primary human adipose tissue-derived mesenchymal stem cells showed DNA damage and pro-inflammatory cytokine production after ZnO NP exposure as well. The stem cell migration capacity was impaired significantly after NP exposure. Interestingly, ZnO NPs were internalized into the cells in high amounts after 24 h, and remained in the cytoplasm for over three weeks, indicating bioaccumulation of the particles. Future studies should illuminate cellular uptake dynamics and exclusion mechanisms. The intracellular persistence of NPs could be a severe problem, since even low exposure doses can lead to critical intracellular concentrations after repetitive contact [52]. The repetitive exposure of nasal mucosa mini organ cultures induced an enhanced genotoxic effect, and 24 h after exposure the DNA damage even increased, probably due to persisting NPs in the cells and the ongoing production of ROS [53]. Ghosh et al. investigated the genotoxic effects of ZnO NPs on human peripheral blood mononuclear cells. The in vitro tests revealed weak genotoxic effects. A significant decrease of mitochondrial membrane potential was also detected [54]. Branica et al. demonstrated a significant increase of DNA damage in human lymphocytes after exposure to ZnO NPs [55].
In contrast to the series of publications stating the possible genotoxicity of ZnO NPs, there are other studies showing no evidence of DNA damage. Nam et al. classified ZnO NPs as well as Zn ions as non-genotoxic in the so-called SOS chromotest [56]. In addition, Kwon et al. did not find any genotoxic effects in several in vitro and in vivo assays that used differently sized and differently charged particles [57]. In a study conducted by Alaraby et al., no toxicity or oxidative stress induction was observed in vivo. Furthermore, no significant changes in the frequency of mutant clones or percentage of DNA in tail (comet assay) were measured, although significant changes in Hsp70 and p53 gene expression were detected [58].
Sahu et al. demonstrated the cytotoxic effects and inflammatory potential of ZnO NPs in a human monocyte cell line, but did not observe any DNA damage [59]. Bayat et al. critically discussed the test systems that are routinely used for genotoxicity assessments. They stated that in vitro genotoxicity testing is probably unreliable because different test systems produce inconsistent results [60].

In Vivo Studies
Only a few studies can be found that evaluate the genotoxicity of ZnO NPs in vivo. Pati et al. investigated the toxicity of ZnO NPs in mice. In this publication, ZnO NPs were dispersed in water by vortex mixing. Afterwards, the animals were fed with water containing NPs in order to demonstrate oral exposure. ZnO NP-treated animals showed signs of toxicity, which was associated with severe DNA damage in peripheral blood and bone marrow cells. Moreover, DNA repair mechanisms were inhibited and enhanced organ inflammation was detected, as well as a disturbance of wound healing [29]. Sharma et al. used a mouse model for subacute oral exposure to ZnO NPs for two weeks. NPs accumulated in the liver and induced DNA damage in liver cells. The authors used an Fpg-modified comet assay to prove that oxidative stress induced DNA damage [32]. Ali et al. found a reduction in glutathione, glutathione-S-transferase, and glutathione peroxidase, as well as an increase in malondialdehyde and catalase in Lymnaea luteola freshwater snails after ZnO NP exposure for 24 and 96 h. Genotoxic effects were found in the digestive gland cells treated with ZnO NPs [61]. Li et al. used a mouse model to prove the biodistribution and genotoxicity of orally administered and intraperitoneally injected ZnO NPs [62]. Baky et al. examined the cardiotoxic effects of ZnO NPs in rats [63], and found that alpha-lipoic acid and vitamin E reduced the DNA damage in cardiac cells. Zhao et al. found DNA damage in embryo-larval zebrafish [64]. The authors compared the toxic effects of Zn ions and ZnO NPs, and demonstrated that ions only partially contribute to the toxic effects. In contrast, triethoxycaprylylsilane-coated ZnO NPs did not induce DNA damage in lung cells from rats after inhalation exposure [65]. Ghosh et al. showed a reduced mitochondrial membrane potential in bone marrow cells in vivo. Furthermore, an enhanced oxidative stress, a G0/G1 cell cycle arrest, and chromosomal aberration with micronuclei formation were measured [54]. In the study conducted by Ng et al., a significant toxicity was observed in melanogaster F1 progenies upon ingestion of ZnO NPs. The egg-to-adult viability of the flies was significantly reduced, which was closely associated with ROS induction by ZnO NPs. Nuclear factor E2-related factor 2 was identified to play a role in ZnO NP-mediated ROS production [49]. Anand  urchins. ZnO NPs provoked damages to immune cells in adult echinoids and transmissible effects to offspring [67].

Summary
Although evaluations of the genotoxicity of ZnO NPs are not consistent, there seems to be reliable evidence supporting the potential for them to damage the DNA in human cells. Genotoxic events were demonstrated using several methods and different endpoints. Besides the comet assay, the micronucleus test, the chromosomal aberration assay, and the γ H2AX method were used. The correlation between oxidative stress and DNA damage can be easily proved by the Fpg-modified comet assay and by the interaction with antioxidants such as N-acetylcysteine. Research has shown the internalization of ZnO NPs into the cells via endocytosis or several other mechanisms such as macropinocytosis. Intracellular distribution was observed by transmission electron microscopy as well as by alternative methods such as side scatter flow cytometry. While there is still some controversy surrounding the possible transfer of ZnO NPs into the nucleus, a distribution into cell organelles can definitely be observed. The inclusion into lysosomes seems to be of major importance, since due to the low pH milieu of lysosomes, ZnO dissolves and Zn 2+ ions are released. Ion release from ZnO NPs may already occur in the cellular expansion medium. Research studies also discuss both intracellular and extracellular Zn 2+ release as main triggers for DNA damage. Even if ZnO NPs are not able to enter the nucleus, Zn 2+ ions affect DNA integrity in a dose-dependent manner. Lysosomes release Zn 2+ ions into the cytoplasm, which is then a trigger for ROS generation. Several research groups have proven this phenomenon by using the DCFH-DA assay. Markers for oxidative stress such as GSH reduction, elevated gluthatione, malondialdehyde, superoxide dismutase, and catalase were analyzed after ZnO NP exposure. As a reaction to disrupted DNA integrity, lysosomes develop into autophagosomes, which can be detected by transmission electron microscopy or indirectly by several protein markers such as LC3 II or beclin-1. The role of autophagy on apoptosis or cell survival is still unclear, and only a few studies address the topic of DNA repair capacity after NP exposure. There is evidence indicating the insufficient repair of DNA disintegrity after ZnO NP exposure, which can be explained by trapped NPs in intracellular departments, and an ongoing trigger for ROS-induced DNA damage. Figure 1 shows a hypothetical model of ZnO NP-induced genotoxicity.
develop into autophagosomes, which can be detected by transmission electron microscopy or indirectly by several protein markers such as LC3 II or beclin-1. The role of autophagy on apoptosis or cell survival is still unclear, and only a few studies address the topic of DNA repair capacity after NP exposure. There is evidence indicating the insufficient repair of DNA disintegrity after ZnO NP exposure, which can be explained by trapped NPs in intracellular departments, and an ongoing trigger for ROS-induced DNA damage. Figure 1 shows a hypothetical model of ZnO NP-induced genotoxicity.

Conclusions and Recommendations for Future Research
At present, there is still limited information regarding the genotoxic potential of ZnO NPs. Due to inconsistencies in the data available, it is nearly impossible to give recommendations or properly assess the risk of ZnO NP application. Most studies on the hazardous effects of ZnO NPs focus on cytotoxicity. However, ZnO NPs seem to belong to a group of nanomaterials that are able to cause DNA damage. Thus, further genotoxicological evaluation is needed. A strictly detailed and standardized physicochemical characterization of the tested NPs is obligatory in order to produce comparable and informative genotoxicological data. The authors refer to the recommendations of Landsiedel et al. (2010) [65] regarding nanotoxicological study design. Most genotoxicological investigations on ZnO NPs address acute exposure situations. That is why our knowledge of bioaccumulation and long-term exposure effects is only fragmentary. Hence, test systems need to be established in order to clarify these questions, and the biological mechanisms responsible for DNA damage must be analyzed continuously. ZnO NPs are very promising and highly effective materials, and a proper characterization of the genotoxic issues is mandatory in order to apply them reasonably and safely. Table 1 summarizes relevant publications on ZnO NP-associated genotoxicity mechanisms. The order of listed NPs in Table 1 was sorted according to the particle size, beginning with the smallest. We did not observe any tendency that the results regarding genotoxic potency varied within the two groups of particles smaller or larger than 100 nm. Although the group of larger particles did not exactly fit the definition of NPs, they seem to be still small enough to exhibit comparable toxic properties as compared with NPs <100 nm.