*Article* **Role of Autophagy in Zinc Oxide Nanoparticles-Induced Apoptosis of Mouse LEYDIG Cells**

**Jingcao Shen 1,**†**, Dan Yang 1,**†**, Xingfan Zhou 2, Yuqian Wang 2, Shichuan Tang 2, Hong Yin 3, Jinglei Wang 1, Rui Chen 2,\* and Jiaxiang Chen 1,4,\***


Received: 28 July 2019; Accepted: 16 August 2019; Published: 19 August 2019

**Abstract:** Zinc oxide nanoparticles (ZnO NPs) have shown adverse health impact on the human male reproductive system, with evidence of inducing apoptosis. However, whether or not ZnO NPs could promote autophagy, and the possible role of autophagy in the progress of apoptosis, remain unclear. In the current study, in vitro and in vivo toxicological responses of ZnO NPs were explored by using a mouse model and mouse Leydig cell line. It was found that intragastrical exposure of ZnO NPs to mice for 28 days at the concentrations of 100, 200, and 400 mg/kg/day disrupted the seminiferous epithelium of the testis and decreased the sperm density in the epididymis. Furthermore, serum testosterone levels were markedly reduced. The induction of apoptosis and autophagy in the testis tissues was disclosed by up-regulating the protein levels of cleaved Caspase-8, cleaved Caspase-3, Bax, LC3-II, Atg 5, and Beclin 1, accompanied by down-regulation of Bcl 2. In vitro tests showed that ZnO NPs could induce apoptosis and autophagy with the generation of oxidative stress. Specific inhibition of autophagy pathway significantly decreased the cell viability and up-regulated the apoptosis level in mouse Leydig TM3 cells. In summary, ZnO NPs can induce apoptosis and autophagy via oxidative stress, and autophagy might play a protective role in ZnO NPs-induced apoptosis of mouse Leydig cells.

**Keywords:** ZnO NPs; Leydig cells; apoptosis; autophagy; oxidative stress

#### **1. Introduction**

Nanotechnology manipulates matters at the atomic, molecular, and supramolecular scales and has grown rapidly worldwide in the past decades. With the development of nanotechnology, environmental exposure to nanoparticles (NPs) is increasing dramatically [1,2]. Metal oxide nanoparticles are the most abundantly produced types of engineered nanomaterials in industry [3]. Among them, zinc oxide nanoparticles (ZnO NPs) are used in various applications, such as cosmetics, rubber manufacture, pigments, food additives, biosensors, chemical fibers, bioimaging, and antibacterial agents, due to their low production cost and unique physicochemical properties [4]. ZnO NPs may enter human bodies by various routes, including inhalation, dermal penetration, injection, and ingestion [5]. These NPs can then accumulate in various organs, such as the liver, spleen, lungs, kidney, and heart via circulation, and may produce adverse consequences, such as edema and degeneration of hepatocytes, inflammation of the pancreas, or damage to the stomach and spleen [6,7]. Consequently, toxicity research and health risk assessments of ZnO NPs have attracted tremendous attention recently [8].

Tissue damage due to NPs exposure arises from direct cell–NPs interaction and is associated with local concentrations of exogenous substances in the tissues, i.e., the nanoparticle itself or solubilized ions [9,10]. Previous researches show that zinc ion dissolved from the surface of ZnO NPs is a primary reason for its cytotoxicity [11,12]. Reducing the ion release from the surface, such as by pre-coating with a protein corona, could greatly decrease their cytotoxicity [13,14]. Furthermore, our previous research illustrates that ZnO NPs and their soluble ions can induce significant cellular endoplasmic reticulum (ER) stress responses before triggering ER-related apoptosis [12]. Generation of reactive oxygen species (ROS) is generally involved in cellular damage from the exposure to ZnO NPs [12,13,15]. ROS are chemically-reactive molecules containing oxygen, which are generated as by-products of biological oxidation during mitochondrial respiration under physiological conditions. ROS include both free radicals, such as nitric oxide (NO), superoxide (O2 •−), and hydroxyl radical (•OH), and peroxides [16]. Reduced glutathione (GSH) and antioxidant enzymes such as glutathione peroxidase (GSH-PX) and superoxide dismutase (SOD) are normally used to scavenge ROS. Oxidative stress occurs when there is an imbalance between ROS production and the cellular antioxidant defence system [17,18].

Spermatogenesis consists of highly organized and sequential steps of undifferentiated spermatogonial stem cell proliferation and differentiation, which generates functional sperms in the testis [19–21]. ZnO NPs can cause vacuolization of germinal epithelium and sloughing of germ, and even decrease the sperm number and motility in the epididymis [22]. In addition to Sertoli cells, Leydig cells play an important role in maintaining spermatogenesis and are prone to being affected by various chemicals [23,24]. ZnO NPs have been reported to exert cytotoxic effects on mouse Leydig cells [25,26]. Similar toxic effects were revealed in the testis of six-month-old common carp Cyprinus carpio after exposure to 10, 50, and 100 μg/L ZnO NPs for 21 days [27]. Furthermore, increasing evidence suggests that the toxicity of ZnO NPs may result from ROS production [9,28,29].

Autophagy is an evolutionarily-conserved, highly-regulated lysosomal degradative pathway involving the delivery of cytoplasmic cargo to the lysosome, which occurs at low basal levels to perform protein and organelle turnover in normal situations [30,31]. Autophagy can be induced during starvation or growth factor withdrawal in order to generate more intracellular nutrients and energy [32]. Autophagy can also be induced under stressful conditions such as neurodegenerative diseases, pathogen infections, chemotherapy, and chemical exposure [33–37]. Increasing evidence has shown that ZnO NPs can induce autophagy in immune cells, normal skin cells, gastrointestinal tract cells, and kidney tissue [7,38–40]. Until now, there has been no evidence that ZnO NPs exposure could induce autophagy in testis tissue.

The aims of the present study were to investigate whether oxidative stress was involved in ZnO NPs-induced apoptosis and autophagy of mouse Leydig cells, and to determine the role of autophagy in ZnO NPs-induced apoptosis. These results will provide fundamental understanding of ZnO NPs-induced spermatogenesis failure.

#### **2. Results**

#### *2.1. Characteristics and Morphology of ZnO NPs*

Transmission electron microscopy (TEM) test shows the primary size of ZnO NPs is about 30 nm with a propensity to agglomerate (Figure S2). These characteristics are comparable to previous publications using the same nanoparticles [41,42]. The hydrodynamic sizes and zeta potentials of ZnO NPs suspended in water are 66.36 ± 0.39 nm (PDI = 0.167, *n* = 3) and 38.25 ± 1.06 mV (*n* = 3),respectively.

#### *2.2. ZnO NPs Cause Testis Damage to Male Mice*

As shown in Figure 1A, the testes of vehicle-treated mice showed normal seminiferous tubules lined with both spermatogenic cells and Sertoli cells. No detached germ cells were found in the tubular lumen. In the 100 mg/kg/day ZnO NPs exposure group, no significant morphologic changes were observed at the seminiferous epithelium. However, the seminiferous tubule demonstrated mildly disorganized histo-architecture in the 200 mg/kg/day group. In the 400 mg/kg/day group, seminiferous tubules exhibited disintegration of the germinal epithelium, germ cell depletion, and a reduction in round sperm. There was a significant decrease in sperm density of the epididymis after exposure to 100, 200, or 400 mg ZnO NPs/kg/day compared to the vehicle control group (Figure 1B), indicating that ZnO NPs exposure significantly inhibited spermatogenesis.

**Figure 1.** Intragastrical exposure of zinc oxide nanoparticles (ZnO NPs) cause toxic damage to the mouse male reproductive system. (**A**) Testes were obtained from male mice treated with 0 (**a**), 100 (**b**), 200 (**c**), or 400 (**d**) mg ZnO NPs/kg/day for 28 days. The testes were stained with hematoxylin and eosin (HE) and then were visualized under an IX51 Olympus microscope. The disruption of the seminiferous epithelium in the testis is indicated by arrows. Magnification: 100×. (**B**) Epididymides were obtained from male mice treated with 0 (**a**), 100 (**b**), 200 (**c**), or 400 (**d**) mg ZnO NPs/kg/day for 28 days, and stained with HE. The sperm in the epididymis are indicated by an asterisk. Magnification: 200×. (**C**) The protein levels of cleaved Caspase-3, cleaved Caspase-8, Bax, and Bcl 2 and (**E**) the levels of LC3, Beclin 1, and Atg 5 were detected by Western blot; Actin was used as an internal control. (**D,F**) The relative protein levels were quantified by densitometry. (**G**) The serum testosterone concentration. The experiment was done in triplicate and repeated three times (*n* = 9). Data were analyzed by one-way ANOVA. \* *p* < 0.05.

To further investigate the potential mechanism of ZnO NPs-induced spermatogenesis failure, the apoptosis level in the mouse testis tissues was assessed. As can be seen from Figure 1C,D, ZnO NPs significantly increased the levels of apoptosis-related proteins, including cleaved Caspase-8, cleaved Caspase-3 and Bax, along with a decreased protein level of Bcl 2 in the testis tissue, which indicates that ZnO NPs induced apoptosis of the testis tissue. Additionally, ZnO NPs markedly increased the ratio of LC3-II/LC3-I, as well as the levels of autophagy proteins Atg 5 and Beclin 1, indicating that ZnO NPs induced autophagy of the testis tissue (Figure 1E,F). Furthermore, ZnO NPs decreased the serum testosterone concentration in a dose-dependent manner (*p* < 0.05), which implies that ZnO NPs disrupted the physiological function of the male reproductive system by targeting the Leydig cells (Figure 1G).

#### *2.3. ZnO NPs Induce Apoptosis of Mouse Leydig TM3 Cells*

The content of testosterone dramatically decreased in the ZnO NPs-treated groups, which implies that ZnO NPs might cause damage to Leydig cells. To further verify the hypothesis, mouse Leydig TM3 cell line was utilized as an in vitro model. As shown in Figure 2A, ZnO NPs at concentrations of 3, 4, and 8 μg/mL significantly inhibited cell viability. Further tests showed that the cell viability was further suppressed at time points of 24, 48, and 72 h post-exposure to 4 μg/mL ZnO NPs (Figure 2B). To determine whether the anti-proliferative effect of ZnO NPs resulted from apoptosis, the apoptosis-related proteins were investigated, including cleaved Caspase-8, cleaved Caspase-3, Bcl 2 and Bax, after the cells were incubated with 0, 2, 3, and 4 μg/mL ZnO NPs for 24 h. It was shown that ZnO NPs dramatically increased the protein levels of cleaved Caspase-8, cleaved Caspase-3, and Bax, as well as decreased Bcl 2 protein level (Figure 2C,D). Furthermore, ZnO NPs increased the numbers of AnnexinV-FITC positive staining cells (Figure 2E). These results indicate that ZnO NPs induced apoptosis of mouse Leydig TM3 cells.

**Figure 2.** ZnO NPs induce apoptosis in mouse Leydig TM3 cells. 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyltetrazolium bromide (MTT) assay results of mouse Leydig TM3 cells treated with 0–8 μg/mL ZnO NPs for 24 h (**A**) or treated with 4 μg/mL ZnO NPs for 24~72 h (**B**). (**C**) The cells were treated with 0–4 μg/mL ZnO NPs for 24 h; then, the protein levels of cleaved Caspase-3, cleaved Caspase-8, Bcl 2, and Bax were investigated by Western blot; Actin was used as an internal control. (**D**) The relative protein levels were quantified by densitometry. (**E**) The cells were treated with 0 (**a**), 2 (**b**), 3 (**c**), 4 (**d**) μg/mL ZnO NPs for 24 h, then the AnnexinV-FITC positive staining cells were counted by flow cytometry. The experiment was done in triplicate and repeated three times. Data were analyzed by one-way ANOVA. \* *p* < 0.05.

#### *2.4. ZnO NPs Induce Apoptosis through Activation of Oxidative Stress*

In order to investigate whether oxidative stress was involved in ZnO NPs-induced apoptosis of mouse Leydig TM3 cells, the contents of malondialdehyde (MDA) and GSH and the enzyme activities of SOD and GSH-PX were determined after the cells were treated with ZnO NPs for 24 h. As shown in Figure 3, ZnO NPs significantly increased MDA level in the cells in a dose-dependent manner, whereas the content of GSH and the activities of the antioxidant enzymes SOD and GSH-PX were decreased in the ZnO NPs-treated cells, which implies that ZnO NPs induced oxidative stress in mouse Leydig TM3 cells. The same markers were detected after the cells were treated with H2O2 and the cell viability was significantly inhibited with the induction of apoptosis, suggesting that oxidative stress could induce apoptosis of mouse Leydig TM3 cells (Figure S3). The mouse Leydig TM3 cells were treated with 4 μg/mL ZnO NPs for 24 h in the presence or absence of 5 mM NAC, an inhibitor of ROS, to further confirm the role of oxidative stress in ZnO NPs-induced apoptosis. As shown in Figure 4, inhibition of viability and induction of apoptosis by ZnO NPs was significantly rescued by NAC. These results illustrate that oxidative stress was involved in ZnO NPs-induced apoptosis of mouse Leydig TM3 cells.

**Figure 3.** ZnO NPs induce oxidative stress in mouse Leydig TM3 cells. Mouse Leydig TM3 cells were treated with 0–4 μg/mL ZnO NPs for 24 h; then the contents of MDA (**A**) and GSH (**B**) and the enzyme activities of SOD (**C**) and GSH-PX (**D**) were determined. The experiment was done in triplicate and repeated three times. Data were analyzed by one-way ANOVA. \* *p* < 0.05.

**Figure 4.** Oxidative stress is involved in ZnO NPs-induced apoptosis of mouse Leydig TM3 cells. Mouse Leydig TM3 cells were treated with 4 μg/mL ZnO NPs for 24 h in the absence or presence of 5 mM NAC, then cell viability (**A**), the protein levels of cleaved Caspase-8, cleaved Caspase-3, Bcl 2, and Bax (**B**) and the AnnexinV-FITC positive staining cells (**D**) were detected by MTT assay, Western blot, and flow cytometry, respectively. (**C**) The relative protein levels of cleaved Caspase-8, cleaved Caspase-3, Bcl 2, and Bax were quantified by densitometry. The experiment was done in triplicate and repeated three times. Data were analyzed by one-way ANOVA. \* *p* < 0.05.

#### *2.5. Oxidative Stress is Involved in ZnO NPs-Induced Autophagy*

As shown in Figure 5A,B, ZnO NPs increased the ratio of LC3-II to LC3-I, as well as the protein levels of Atg 5 and Beclin 1. Similarly, H2O2 markedly increased the ratio of LC3-II to LC3-I and the contents of Atg 5 and Beclin 1, indicating that oxidative stress could induce autophagy of mouse Leydig TM3 cells (Figure S4). Furthermore, inhibition of oxidative stress could rescue the induction of autophagy by ZnO NPs (Figure 5C,D). ZnO NPs-induced autophagy was further investigated by TEM. As shown in Figure 5E, there were relatively few autophagosomes in the cytoplasm of the control cells, while autophagic vacuoles containing extensively-degraded organelles (such as mitochondria and endoplasmic reticulum) significantly increased in both ZnO NPs-treated cells and starvation-treated cells. Interestingly, inhibition of oxidative stress decreased the number of autophagosomes. These results suggest that oxidative stress played an important role in ZnO NPs-induced autophagy. It is worthy to note that ZnO NPs-induced autophagy and apoptosis of mouse Leydig TM3 cells might be closely related to the soluble zinc ions, as similar bio-effects were observed after the cells were treated with 0–1 μg/mL ZnCl2 (Figure S5).

#### *2.6. Inhibition of Autophagy Increases ZnO NP-Induced Apoptosis*

As apoptosis and autophagy were both induced by ZnO NPs, the effects of autophagy on ZnO NPs-induced apoptosis were studied. Cell viability was measured after the cells were treated with 4 μg/mL ZnO NPs for 24 h in the absence or presence of an autophagy inhibitor, either 10 mM 3-Methyladenine (3-MA) or 1 μM Wortmannin (Wort). Compared with the ZnO NPs-treated cells, inhibition of autophagy further decreased viability of mouse Leydig TM3 cells (Figure 6A) and up-regulated the protein levels of cleaved Caspase-8, cleaved Caspase-3, and Bax, accompanied by the down-regulation of Bcl 2 protein (Figure 6B,C). The number of AnnexinV-FITC positive staining cells were also markedly increased when autophagy was inhibited (Figure 6D). These results indicate that autophagy might play a protective role in ZnO NPs-induced apoptosis of mouse Leydig TM3 cells.

**Figure 5.** Oxidative stress is involved in ZnO NPs-induced autophagy of mouse Leydig TM3 cells. Mouse Leydig TM3 cells were treated with 0–4 μg/mL ZnO NPs for 24 h (**A**) or treated with 4 μg/mL ZnO NPs for 24 h in absence or presence of 5 mM NAC (**C**); then, the protein levels of LC 3, Atg 5, and Beclin 1 were quantified by Western blot. (**B,D**) The relative protein levels of LC 3, Atg 5, and Beclin 1 were quantified by densitometry. (**E**) The cells were treated with ddH2O, bars: 1 μm, (**a**), 4 μg/mL ZnO NPs (**b**), or 5 mM N-acetyl-L-cysteine (NAC) plus 4 μg/mL ZnO NPs for 24 h (**d**). Then, autophagic vacuoles in the cells were visualized by transmission electron microscopy (TEM), with starvation-treated cells as a positive control (**c**). The autophagic vacuoles are indicated by white arrows. The experiment was done in triplicate and repeated three times. Data were analyzed by one-way ANOVA. \* *p* < 0.05.

**Figure 6.** Inhibition of autophagy increases the ZnO NPs-induced apoptosis level in mouse Leydig TM3 cells. Mouse Leydig TM3 cells were treated with 4 μg/mL ZnO NPs for 24 h in the absence or presence of 10 mM 3-MA or 1 μM Wortmannin (Wort), then cell viability (**A**), the protein levels of cleaved Caspase-8, cleaved Caspase-3, Bcl 2, and Bax (**B**) and the AnnexinV-FITC positive staining cells (**D**) were tested by MTT assay, Western blot, and flow cytometry, respectively. (**C**) The relative protein levels of cleaved Caspase-8, cleaved Caspase-3, Bcl 2, and Bax were quantified by densitometry. The experiment was done in triplicate and repeated three times. Data were analyzed by one-way ANOVA. \* *p* < 0.05.

#### **3. Discussion**

Exposure to ZnO NPs for humans is inevitable due to their wide applications in commercial and industrial products. Thus, the adverse effects from the exposure of ZnO NPs need clear definition. Nanoparticles can pass through the blood–brain barrier (BBB), blood–testis barrier (BTB), and blood–air barrier (BAB), with the ability to accumulate in the brain, the testis, or peripheral organs [43–46]. Recently, Qian et al. showed that ZnO NPs could cause adverse effects throughout the male reproductive system by impairing the BTB [47]. Oral dose toxicity has been reported in SD mice after repetitive exposure to positively-charged 100 nm ZnO NPs over 14 or 90 days, and the target organs were found to be the spleen, stomach, and pancreas, with a no-observed-adverse-effect dose level of about 125 mg/kg (b.w.) [48,49]. In our research, it was shown that 28-day gavage exposure of ZnO NPs (30 nm positively charged) at the concentrations of 100, 200, and 400 mg/kg/day caused disruption and atrophy of the seminiferous epithelium in the testis of mice. Furthermore, the sperm density in the epididymis significantly decreased in the ZnO NPs-treated groups, which was in good agreement with some previous work [22,27]. This toxic dosage range is also similar to the research of Hong et al., in which they tested the toxicity on embryo-fetal development in rats from 15 days of repeated oral doses of 20 nm negatively-charged ZnO NPs [50]. Therefore, ZnO NPs, with the high chance of daily contact and exposure, may pose a high risk of reproductive toxicity after long-term accumulation in the human body.

ZnO NPs have been shown to induce apoptosis in many cells such as human epidermal keratinocytes, human aortic endothelial cells, human liver and kidney podocytes [51–54]. Han et al. showed that ZnO NPs took cytotoxic effects on mouse testicular cells and induced apoptosis in Leydig cells [25]. In this research, we confirmed that ZnO NPs up-regulated the protein levels of Bax, cleaved Caspase-3, and cleaved Caspase-8 in the testis tissue, as well as decreased the protein level of Bcl 2, which indicates that ZnO NPs could induce apoptosis in the testis.

Autophagy protein LC3, a widely used marker of mammalian autophagy, has two forms, i.e., a cytosolic form (LC3-I) and an autophagic vesicle-associated form (LC3-II). During induction of autophagy, LC3-I covalently conjugates with phosphatidylethanolamine and develops LC3-II, which is recruited and bound to the autophagosome membrane [33,55]. The conversion of LC3-I to LC3-II is considered to be a crucial step in initiating autophagy [56], with the amount of LC3-II related to the extent of autophagosome formation [55]. In the present study, ZnO NPs exposure significantly increased the ratio of LC3-II to LC3-I in the testis tissue, along with similar up-regulation of autophagy proteins Atg 5 and Beclin 1. These results implied that ZnO NPs could induce autophagy in the testis tissue.

The primary function of Leydig cells is the synthesis and secretion of androgen, which plays an important role in spermatogenesis [57]. In our study, ZnO NPs could decrease serum testosterone level, indicating that Leydig cells might be the target for ZnO NPs-induced spermatogenesis failure. To further verify this hypothesis, the mouse Leydig TM3 cell line was utilized as an in vitro research model. In agreement with the in vivo findings, ZnO NPs exposure inhibited viability and induced apoptosis of mouse Leydig TM3 cells. Thus, it is reasonable to speculate that the inhibition of cell viability upon ZnO NPs exposure might result from the induction of apoptosis.

Oxidative stress has been identified as a critical pathophysiological mechanism of reproductive toxicity from environmental chemicals or organophosphorus compounds [58]. Asani et al. showed that ZnO NPs could induce oxidative stress in pancreatic β-cells [59]. Similar to the toxicity effects of H2O2, exposure to ZnO NPs significantly increased the MDA in the cells, along with a marked decrease in both the GSH levels and the enzyme activities of SOD and GSH-PX. Further evidence demonstrates that apoptosis could be distinctly reduced when oxidative stress was inhibited, which confirmed that oxidative stress was involved in ZnO NPs-induced apoptosis of mouse Leydig TM3 cells. Oxidative stress has been shown to induce autophagy and plays an important role in chemical-induced autophagy [21,60]. In the current study, ZnO NPs exposure induced autophagy of mouse Leydig TM3

cells, which could be inhibited by NAC, a scavenger of ROS. Collectively, these results provide clear evidence that oxidative stress was critical in ZnO NPs-induced autophagy in mouse Leydig TM3 cells.

Both cell survival and death can be related to autophagy when the cells are subject to stressful conditions. In most circumstances, autophagy will promote cell survival [35,61]. However, autophagy is also considered to be a form of non-apoptotic programmed cell death—"type II" or "autophagic" cell death [62,63]. To investigate the role of autophagy in ZnO NPs-induced apoptosis, apoptosis was measured after the treatment of ZnO NPs in the absence or presence of autophagy inhibitor. Surprisingly, inhibition of autophagy could further induce apoptosis of mouse Leydig TM3 cells (Figure 7). These results illustrate that autophagy plays a cytoprotective role in ZnO NPs-induced apoptosis of mouse Leydig TM3 cells.

**Figure 7.** Schematic representation of the activation mechanism of apoptosis and cytoprotective autophagy in mouse Leydig cells after ZnO NP exposure. The up-regulation expression of protein is indicated by up arrow (↑), and down-regulation expression is indicated by down arrow (↓) in schematic illustration.

#### **4. Materials and Methods**

#### *4.1. Reagents*

ZnO NPs (No. 721077), N-acetyl-L-cysteine (A7250), 3-Methyladenine (M9281), and Wortmannin (12-338) were obtained from Sigma (St. Louis, MO, USA). Mouse Leydig cell line (TM3) was obtained from the Cell Culture Center of the Institute of Basic Medical Science, Chinese Academy of Medical Sciences (Beijing, China). Anti-Caspase-3 (sc-7148), anti-Caspase-8 (sc-7890), anti-Bax (sc-493), anti-Bcl-2 (sc-492), and anti-β-actin (sc-69879) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Anti-LC3 (PD014), anti-Atg5 (PM050), and anti-Beclin-1 (PD017) were gained from MBL Co. Ltd. (Nagoya, Japan). The AnnexinV-FITC/PI Apoptosis Kit (V13242) was purchased from Invitrogen Life Technologies (Waltham, MA, USA). Oxidation-antioxidation assay kits of malondialdehyde (MDA) (A003-1), glutathione (GSH) (A006-1), superoxide dismutase (SOD) (A001-1-1) and glutathione

peroxidase (GSH-PX) (A005), testosterone Assay Kit (H090), and protease inhibitor cocktail (W060) were bought from Nanjing Jiancheng Bioengineering Institute (Nanjing, China).

#### *4.2. Nanoparticles and Characterization*

In characterization tests, ZnO NPs dispersed in sterile Milli-Q water (final concentration 1 mg/mL, Milford, MA, USA) were put into an ultrasound bath (100 W, Shanghai, China) to break up the aggregates before transmission electron microscopy (TEM, JEM-200CX, JEOL, Japan) and dynamic light scattering (DLS) analyses (Malvern Zeta sizer Nano ZS, Malvern Instruments, U.K.).

#### *4.3. Animal Administration*

Adult male Kunming mice (8 weeks old, 20–25 g) were obtained from the Shanghai Laboratory Animal Center, Chinese Academy of Sciences (CAS, Shanghai, China). The mice were housed in an isolated and air-conditioned animal room with water and rodent food supplement. All animals were acclimated to this environment for at least one week prior to the experiment. Experiments were approved by the Animal Ethics Committee of Nanchang University, China, SYKX2015-0001, 12 October 2015. The mice were intragastrically (i.g.) administered with ZnO NPs (0, 100, 200, 400 mg/kg/day, diluted in water) for 28 days and were then anesthetized with carbon dioxide inhalation, followed by cervical dislocation. The serum samples were collected following standard operation procedures. Then, the testes and the epididymis were quickly dissected free of fat, decapsulated, and frozen in liquid nitrogen.

#### *4.4. Histology*

Male mouse testis and epididymis tissues were stained with hematoxylin and eosin (HE) according to the method described by Chen et al. [43].

#### *4.5. Western Blotting Analysis*

The homogenized testis tissue and mouse Leydig TM3 cells were harvested in lysis buffer (50 mM Tris pH 7.5, 0.3 M NaCl, 5 mM EGTA, 1 mM EDTA, 0.5% Triton X-100, 0.5% NP40) containing protease inhibitor cocktail. Then, the supernatants were collected for Western blot after centrifuge for 10 min at 12 000× *g*. All primary antibodies and their recommended secondary antibodies were diluted 1:1000 and 1:5000, respectively.

#### *4.6. Detection of Testosterone Content*

The testosterone level in the serum was determined by ELISA kit according to the manufacturer's instructions (Nanjing Jiancheng Bioengineering Institute, Nanjing, China).

#### *4.7. Cell Culture and ZnO NP Treatment*

Mouse Leydig TM3 cells were cultured at 37 ◦C in a 5% CO2 atmosphere in Dulbecco's modified Eagle's medium (DMEM, Gibco, Langley, OK, USA), supplemented with 5% horse serum (Gibco, Langley, OK, USA) and 2.5% fetal bovine serum (FBS, Gibco, Langley, OK, USA). The cultured cells were seeded and incubated for 24 h before exposure to varying concentrations of ZnO NPs or ZnCl2 as per experimental designs. The summary of the research design was illustrated in Figure S1.

#### *4.8. Cell Viability Assay*

The cells were seeded at a density of 1 × 104 cells per well in medium in 96-well plates and incubated for 24 h. The medium was then replaced with ZnO NPs of indicated concentrations in the presence or absence of 5 mM NAC, 10 mM 3-MA, or 1 μM Wortmannin for 24 h. Cell viability was determined by measuring the absorbance at 570 nm after the cells were incubated with 0.5 mg/mL MTT in medium for 4 h.

#### *4.9. AnnexinV-FITC*/*PI Apoptosis Assay*

Apoptosis was determined by using an AnnexinV-FITC/PI Apoptosis Kit from Invitrogen Life Technologies (Waltham, MA, USA) as described previously [64].

#### *4.10. Oxidative Stress Measurement*

The resultant supernatants of homogenized mouse Leydig TM3 cells were utilized to determine the activities of GSH-PX and SOD and the levels of GSH and MDA by using the commercial kits following the manufacturer's instructions. The protein concentration was detected by the Bradford assay.

#### *4.11. Transmission Electron Microscopy (TEM) Analysis*

Mouse Leydig TM3 cells were treated with ddH2O, 4 μg/mL ZnO NPs or 5 mM NAC, plus 4 μg/mL ZnO NPs for 24 h. Then, the autophagic vacuoles were observed by TEM as previously described [65]. The cells treated for 2 h by starvation media (140 mM NaCl, 1 mM CaCl2, 1 mM MgCl2, 5 mM glucose, and 20 mM HEPES at pH = 7.4 supplemented with 1 % BSA) were used as the positive control of autophagy.

#### *4.12. Statistical Analysis*

The data were represented as means ± SE. Statistical analyses were performed using a one-way ANOVA with Newman–Keuls multiple range test. *p* < 0.05 was considered statistically significant.

#### **5. Conclusions**

ZnO NPs could cause disruption and atrophy of seminiferous epithelium, and even damage to spermatogenesis in male mice. Apoptosis and autophagy were induced by ZnO NPs in the testis tissue with a decreased level of serum testosterone. In vitro studies demonstrated that ZnO NPs markedly inhibited the viability of mouse Leydig TM3 cells and induced apoptosis and autophagy. Oxidative stress was also induced after the cells were treated with ZnO NPs, while inhibition of oxidative stress could rescue the induction of apoptosis and autophagy, indicating that oxidative stress was involved in ZnO NPs-induced apoptosis and autophagy. However, suppression of autophagy further inhibited cell viability with increase of the apoptosis levels. Taken together, we have provided detailed evidence that oxidative stress is involved in ZnO NPs-induced apoptosis and autophagy of mouse Leydig TM3 cells, while autophagy contributes to counteract the reproductive toxicity of ZnO NPs in the testis.

**Supplementary Materials:** Supplementary materials can be found at http://www.mdpi.com/1422-0067/20/16/ 4042/s1.

**Author Contributions:** J.S., D.Y., R.C., and J.C. conceived and designed the study. J.S., D.Y., X.Z., Y.W., and S.T. carried out all the experiments. H.Y. and J.W. performed statistical analysis. J.S., H.Y., R.C., and J.C. drafted the paper and amended the paper. All authors read and approved the final manuscript.

**Funding:** We thank the financial support from the National Natural Science Foundation of China (No. 81660255, No. 81360098 and No. 21777036), the Young Scientist Training Project of Jiangxi Province, China (No. 20153BCB23032), the Research fund for postgraduate program of Nanchang University (No. cx2015179), and the Youth Plan of Beijing Academy of Science and Technology (YC201809).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Fate Determination of ZnO in Commercial Foods and Human Intestinal Cells**

#### **Ye-Rin Jeon, Jin Yu and Soo-Jin Choi \***

Division of Applied Food System, Major of Food Science & Technology, Seoul Women's University, Seoul 01797, Korea; yrjeon0715@swu.ac.kr (Y.-R.J.); ky5031@swu.ac.kr (J.Y.)

**\*** Correspondence: sjchoi@swu.ac.kr; Tel.: +82-2-970-5634; Fax: +82-2-970-5977

Received: 13 December 2019; Accepted: 7 January 2020; Published: 9 January 2020

**Abstract:** (1) Background: Zinc oxide (ZnO) particles are widely used as zinc (Zn) fortifiers, because Zn is essential for various cellular functions. Nanotechnology developments may lead to production of nano-sized ZnO, although nanoparticles (NPs) are not intended to be used as food additives. Current regulations do not specify the size distribution of NPs. Moreover, ZnO is easily dissolved into Zn ions under acidic conditions. However, the fate of ZnO in commercial foods or during intestinal transit is still poorly understood. (2) Methods: We established surfactant-based cloud point extraction (CPE) for ZnO NP detection as intact particle forms using pristine ZnO-NP-spiked powdered or liquid foods. The fate determination and dissolution characterization of ZnO were carried out in commercial foods and human intestinal cells using in vitro intestinal transport and ex vivo small intestine absorption models. (3) Results: The results demonstrated that the CPE can effectively separate ZnO particles and Zn ions in food matrices and cells. The major fate of ZnO in powdered foods was in particle form, in contrast to its ionic fate in liquid beverages. The fate of ZnO was closely related to the extent of its dissolution in food or biomatrices. ZnO NPs were internalized into cells in both particle and ion form, but dissolved into ions with time, probably forming a Zn–ligand complex. ZnO was transported through intestinal barriers and absorbed in the small intestine primarily as Zn ions, but a small amount of ZnO was absorbed as particles. (4) Conclusion: The fate of ZnO is highly dependent on food matrix type, showing particle and ionic fates in powdered foods and liquid beverages, respectively. The major intracellular and intestinal absorption fates of ZnO NPs were Zn ions, but a small portion of ZnO particle fate was also observed after intestinal transit. These findings suggest that the toxicity of ZnO is mainly related to the Zn ion, but potential toxicity resulting from ZnO particles cannot be completely excluded.

**Keywords:** zinc oxide; fate; cloud point extraction; dissolution; commercial food; intestinal absorption

#### **1. Introduction**

Nanomaterials have been widely applied to diverse industries, such as electronics, medicine, pharmaceutics, cosmetics, and foods. In particular, food additive particles, including silicon dioxide (SiO2), titanium dioxide (TiO2), and zinc oxide (ZnO) are widely used as anticaking agents, coloring agents, and nutritional fortifiers, respectively [1–3]. Among them, ZnO can be added to milk, dairy products, cereals, and beverages. This is due to the fact that zinc (Zn) plays an essential role as a trace element in the immune system, cell function, enzyme activity, and signaling in human body [4–6]. ZnO is used as a food additive in the United States (US) and Europe, and classified as "Generally Recognized as Safe" by the US Food and Drug Administration [7]. The acceptable daily intake (ADI) for Zn in Zn-fortified functional foods is in the range of 2.55 to 12 mg in the Republic of Korea [8]. Nanotechnology developments have led to the production of nano-sized ZnO particles, although nanomaterials that range in size from 1 to 100 nm are not intended to be used as food additives. Indeed, the particle size distribution for ZnO as a food additive is unspecified [9].

The fate determination of ZnO is of importance for interpreting and understanding its potential toxicity, because the toxic effects resulting from ZnO nanoparticles (NPs), NP aggregates/agglomerates, micro-sized particles, or Zn ions are completely different [10,11]. A few reports have demonstrated the fates of ZnO in biological systems or environmental water samples [12–14]. ZnO NPs were reported to be primarily present as ionic forms in tissues after oral administration in rats [15–17]. Zn ion release from ZnO NPs in simulated gastric fluid increased to 14% has also been reported [18,19], suggesting their partial dissolution property. Studies have demonstrated that ZnO can be easily decomposed to Zn ions in acidic solutions or biological fluids [20–23]. Since most foods have a slightly acidic pH and Zn-fortified foods are taken up by the oral route, ZnO can be dissolved into Zn ions in food matrices to some extent, and does not present in its particle form in the body. However, there is no clear information about the fate of ZnO in commercial foods and human intestinal cells during intestinal transit.

Food additive ZnO is directly added to complex food matrices in solution or powdered form. Hence, the interactions between ZnO and food components should also be considered. Indeed, ZnO NPs have been demonstrated to interact with food components, such as proteins and saccharides, forming particle–matrix corona (nanocorona covered by matrices) [23–25]. These interactions were also found to affect cytotoxicity, intestinal transport efficiency, and oral absorption [24–27]. Thus, separating ZnO from complex food matrices in intact particle form is challenging. Currently, the most widely applied method for detecting ZnO is inductively coupled plasma–atomic emission spectroscopy (ICP-AES), ICP–mass spectroscopy, or atomic absorption spectroscopy. However, these methods require pre-digestion procedures with acids to digest organic matrices, and therefore are used for analyzing total Zn levels only and cannot be used to distinguish ZnO particles from Zn ions.

On the other hand, cloud point extraction (CPE) using Triton X-114 (TX-114) was first described and developed for the analysis of trace NPs, such as gold, silver, and copper as particle forms [28–30]. TX-114 is a surfactant, is cost-effective, and has the advantage of forming micelles at room temperature (20–25 ◦C), contributing to easy manipulation. To date, TX-114-based CPE studies have focused on the determination of silver NPs in water or cells [31,32]. ZnO separation by CPE in aqueous phase or environmental waters where little organic matrix is present has also been reported [33]. A CPE-based approach was used for the detection of aluminum (Al) or Zn in foodstuffs. However, these studies used ethylenediaminetetraacetic acid (EDTA) or 8-hydroxyquinoline as chelating agents for precipitation of Al or Zn ions in precipitated surfactant-rich phase by CPE [34,35]. Moreover, the CPE method was applied after pre-treatments such as filtering, nitric acid treatment, and dry-ashing were applied in order to digest organic matrices in foods [34,35]. No attempt has been made to detect ZnO in its particle form in complex systems, such as in foods or biomatrices.

In the present study, an analytical method for determining ZnO with intact particle size and shape in foods and human intestinal cells was established using surfactant-based CPE. The fates of ZnO in commercial processed foods, including liquid and powdered types, were determined based on the optimized method established with pristine ZnO-NP-spiked food matrices. Finally, fate determination was assessed in human intestinal cells, in vitro models of human intestinal barriers, and an ex vivo intestinal absorption model.

#### **2. Results**

#### *2.1. Optimization and Characterization of CPE for ZnO NPs*

TX-114-based CPE was first optimized utilizing pristine ZnO NPs. The optimization of ZnO dispersant was carried out with humic acid (HA) and glucose (Glc), and compared with NPs in distilled and deionized water (DDW) as a control at 37 and 45 ◦C, respectively. HA was chosen as a dispersing agent based on the research of Majedi et al. [33], and Glc is known to play a role in NP dispersion as well [25,36]. The physicochemical properties of ZnO NPs after dispersion (Step 1), during CPE (Step 2: surfactant-rich and aqueous phases without centrifugation), and after CPE (Step 3: surfactant-rich and

aqueous phases separated by centrifugation) were characterized (Figure 1). When particle size, shape, and morphology were examined by field emission-scanning electron microscopy (FE-SEM), pristine ZnO NPs dispersed in DDW, HA, and Glc (Step 1) had an average primary particle size of ~60 nm with irregular shapes (Figure 2A). The precipitated, TX-114-rich phase of ZnO NPs after CPE (Step 3) had similar primary particle size distributions compared to pristine NPs in different dispersants, showing no significant differences in particle size between Step 1 and Step 3 (Figure 2B, *p* > 0.05). FE-SEM images during CPE process (Step 2) could not be clearly obtained due to the presence of high level of organic material TX-114.

**Figure 1.** Schematic illustration of Triton X-114 (TX-114)-based cloud point extraction (CPE) procedure.

The hydrodynamic radii of ZnO NPs consequently increased when ZnO NPs were covered by TX-114 and after micelle formations (Step 2). Thus, the formation of ZnO NPs-TX-114 micelles was confirmed through dynamic light scattering (DLS) analysis. The hydrodynamic diameters of pristine ZnO NPs in DDW or HA were significantly smaller than those in Glc at Step 1 (Table 1). The hydrodynamic radii of ZnO NPs in different dispersants dramatically increased during the CPE process (Step 2) in all cases, but decreased to the same levels similar to that in pristine NPs after CPE (Step 3), only when ZnO NPs were dispersed in HA (Table 1). No statistical differences in temperatures (37 and 45 ◦C) were found (*p* > 0.05), except in ZnO NPs in HA during CPE (Step 2). On the other hand, the zeta potential values changed to slightly negative charges in all cases after CPE (Step 3, Table 1).


**Table 1.** Hydrodynamic radii and zeta potentials of pristine ZnO nanoparticles (NPs) in different CPE conditions.

Different upper-case letters (A,B,C) indicate significant differences among different CPE conditions (*p* < 0.05). Different lower-case letters (a,b,c) indicate significant differences among different CPE steps (*p* < 0.05). Abbreviation: DDW, distilled and deionized water; HA, humic acid; Glc, glucose.

#### *2.2. Recovery of Pristine ZnO NPs by CPE*

The recovery (%) of ZnO NPs in different dispersants by CPE was checked by quantifying Zn amounts in both aqueous and TX-114-rich phases (supernatant and precipitate after CPE, respectively) in order to confirm the efficiency of the CPE procedure. As shown in Figure 2C, more than 82% of the

ZnO NPs were recovered from the precipitates as particles after CPE, regardless of CPE conditions. Only small portions (less than 2%) of ZnO NPs were detected in supernatants as ions after CPE. The highest total recovery (94.9%) of ZnO NPs in both supernatant and precipitate after CPE was found when ZnO NPs were dispersed in HA at 45 ◦C. It is worth noting that 93.4% of pristine ZnO NPs in HA were obtained in the precipitates as particles after CPE. Hence, this condition was further used for fate determination of ZnO in commercial foods and human intestinal cells.

**Figure 2.** Field emission-scanning electron microscopy (FE-SEM) images and size distribution of pristine ZnO NPs in different dispersants (**A**) before CPE and (**B**) after CPE. (**C**) Recovery (%) of Zn ions, ZnO particles, and total Zn levels obtained from pristine ZnO NPs by CPE. Size distributions were obtained by randomly selecting 100 particles from FE-SEM images. Different lower-case letters (a,b,c) indicate significant differences among different CPE conditions (*p* < 0.05).

#### *2.3. Dissolution Property of ZnO NPs in Food*/*Bio Matrices*

The solubility of ZnO NPs in food matrices and cell culture medium was first checked prior to fate determination. Two different food types which were representative of powdered or liquid foods were used. Powdered foods (coffee mix and skim milk) and liquid beverages (milk and sports drink) were selected based on the potential utilization of ZnO in Zn-fortified foods. Figure 3A showed that the solubilities of ZnO NPs in different dispersants ranged from 0.8% to 1.7%. The pHs of ZnO dispersed in DDW, HA, and Glc were 8.6, 8.1, and 8.7, respectively.

On the other hand, the solubilities of ZnO in food matrices increased, reaching 39.4%, 30.0%, and 49.2% in coffee mix, skim milk, and milk, respectively. There was a dramatic increase in solubility observed in sports drink, which reached up to 90.9% (Figure 3B). The pHs of ZnO-spiked food matrices were 6.2, 6.9, 6.9, and 3.3 in coffee mix, skim milk, milk, and sports drink, respectively. The solubility of ZnO in cell culture minimum essential medium (MEM) (pH 7.0) was ~18% after 0.5 h and increased up to 24.8% after 6–24 h (Figure 3C).

**Figure 3.** Dissolution properties of ZnO NPs in (**A**) dispersants, (**B**) food matrices, and (**C**) cell culture medium. Different upper-case letters (A,B) indicate significant differences among different dispersants (*p* < 0.05). Different lower-case letters (a,b) indicate significant differences among incubation times (*p* < 0.05).

#### *2.4. Characterization and Fate of ZnO-NP-Spiked Foods*

The reliability and accuracy of the CPE method for fate determination of ZnO in commercial foods were checked using ZnO-NP-spiked foods. DLS results demonstrated that all ZnO NPs recovered from the precipitates of ZnO-NP-spiked coffee mix, skim milk, and milk after CPE had statistically significant similarities in hydrodynamic radii compared to pristine ZnO NPs (*p* > 0.05, Figure S1). On the other hand, no particle forms were detected in the precipitates of ZnO-NP-spiked sports drink

after CPE (Figure S1). SEM and energy dispersive X-ray spectroscopy (EDS) analysis revealed the presence of Zn elements recovered from the precipitated TX-114-rich phases of all ZnO-NP-spiked foods after CPE, except sports drink (Figure 4A,B).

**Figure 4.** (**A**) FE-SEM images and (**B**) energy dispersive X-ray spectroscopy (EDS) analysis of the precipitates of ZnO-NP-spiked foods after CPE. (**C**) Fate (%) of ZnO NPs in ZnO-NP-spiked foods by CPE, followed by inductively coupled plasma–atomic emission spectroscopy (ICP-AES) quantification.

Quantitative ICP-AES analysis results on the supernatants and precipitates after CPE are presented in Figure 4C. About 59.5%, 59.5%, and 51.3% of ZnO were present as particle forms in coffee mix, skim milk, and milk, respectively, and 93.7% of ZnO NPs were dissolved and detected as Zn ions in sports drink. The total Zn recoveries (%) from both Zn ions and ZnO particles ranged from 92.6% to 97.5%.

#### *2.5. Fate of ZnO in Commercial Foods*

The presence of ZnO particles or Zn ions in Zn-fortified commercial foods was determined based on the established CPE method. Commercial foods that indicated ZnO as an additive as seen in product labeling were chosen, and another product containing Zn gluconate as a Zn fortifier was also used for comparative study. When the supernatants and precipitates after CPE were quantitatively analyzed by ICP-AES, the fate of ZnO differed depending on food type (Figure 5). The highest concentration of ZnO was detected as particle forms (~77–99%) in powdered or dried foods (chocolate powder, powdered probiotics, and cereals). ZnO was determined to be mainly present as Zn ions (~92–98%) in liquid beverages (functional peptide beverage and fruit juice). It is worth noting that the Zn ionic fate of Zn gluconate after CPE was also found in another functional mineral beverage (Figure 5F). The pHs of peptide, fruit juice, and mineral beverages were 4.7, 3.4, and 2.7, respectively, whereas the pHs of powdered chocolate, probiotics, and cereals were 7.3, 7.5, and 6.5, respectively. The total Zn recoveries (%) from both ZnO particles and Zn ions ranged from 96.9% to 102.8%, which was calculated based on composition labeling in commercial foods.

**Figure 5.** Fate (%) of ZnO in ZnO-added commercial foods: (**A**) chocolate powder, (**B**) powdered probiotics, (**C**) cereals, (**D**) functional peptide beverage, (**E**) fruit juice. (**F**) Fate of Zn gluconate in a functional mineral beverage by CPE.

#### *2.6. Intracellular Fate of ZnO NPs*

The fate of ZnO NPs was evaluated by CPE in human intestinal cells, and intracellular Zn levels in both supernatants and precipitates after CPE were quantified after cell lysis. Figure 6A shows that the uptake of ZnO NPs, as measured by total Zn levels, increased in concert with incubation time and reached a plateau at 1 h post-incubation. Intracellular ZnO NPs separated by CPE increased with time and reached a maximum level at 6 h, and then returned to normal control level after 24 h. On the other hand, intracellular Zn ion concentrations gradually increased and the highest Zn ion level was detected at 24 h, which was statistically similar to total Zn level at 24 h. When we compared ZnO NP/Zn ion ratios, most ZnO NPs were present as Zn ions (~71%) at 0.5 h, and both ZnO NPs and Zn ions were present at almost similar levels at 6 h (~53 vs. 47%). Zn ion forms (~87%) were primarily found at 24 h.

The intracellular fate of ZnO NPs was also checked with a Zn-selective fluorescent probe during incubation and examined by confocal microscopy. Figure 6B shows that the fluorescence intensity increased at 0.5–1 h and decreased thereafter. Magnified confocal images clearly demonstrated that the fluorescence intensity inside cells resulting from Zn ions was higher at 0.5–1 h than that at 6–24 h. Slightly increased fluorescence was also observed in control cells without ZnO NPs, attributed to basal intracellular Zn ion levels.

**Figure 6.** (**A**) Intracellular fate of ZnO NPs in human intestinal Caco-2 cells, as determined by CPE followed by ICP-AES analysis. (**B**) Confocal microscopic images of intracellular Zn ions, as determined by a Zn-selective fluorescent probe, TSQ in Caco-2 cells. Different upper-case letters (A,B) indicate significant differences among Zn ion, ZnO, and total Zn (*p* < 0.05). Different lower-case letters (a,b,c) indicate significant differences among different incubation times (*p* < 0.05). Abbreviation: ND, not detectable.

#### *2.7. Intestinal Transport and Absorption Fate of ZnO NPs*

The fate of ZnO NPs after intestinal transport was examined using in vitro Caco-2 monolayer and follicle-associated epithelium (FAE) models, representing the intestinal tight junction and microfold (M) cells in Peyer's patches, respectively. Figure 7A,B show that ZnO NPs were transported through both Caco-2 monolayer and M cells primarily in Zn ion form. No significant differences in transport amount were found between the Caco-2 monolayer and M cells (*p* > 0.05).

We also evaluated the intestinal absorption fate of ZnO NPs using ex vivo everted small intestine sacs, which reflect small intestinal absorption in vivo. These results demonstrated that ZnO NPs were primarily absorbed into the body across the mucosa of the small intestine as Zn ionic forms (Figure 7C). It is also worth noting that ZnO can be slightly but significantly transported by M cells and absorbed in the small intestine in particle form after 0.5–6 h (Figure 7B,C).

**Figure 7.** Intestinal transport fate of ZnO NPs obtained from in vitro models of (**A**) Caco-2 monolayer and (**B**) follicle-associated epithelium (FAE) by CPE. (**C**) Intestinal absorption fate of ZnO NPs using ex vivo everted small intestine sacs. Different lower-case letters (a,b,c) indicate significant differences among different incubation times (*p* < 0.05). Abbreviation: ND, not detectable.

#### **3. Discussion**

ZnO NPs have a wide range of applications in products intended for human consumption, such as foods and cosmetics, which raises increasing concerns about their potential toxicity. The fate determination of NPs is important, and can answer the fundamental question of whether NPs are present as intact particles or dissolved ionized forms. This type of research is also useful for understanding whether the toxicity of ZnO results from nano-sized particles or from Zn ions. In particular, fate determination is crucial for partially soluble NPs in commercial products or biological environments, such as ZnO NPs and silver NPs [37,38]. However, separating NPs and ionized forms from complex food and biological matrices without affecting intact particle size and morphology is challenging.

In this study, TX-114-based CPE was first optimized using pristine ZnO NPs. The results obtained revealed that ZnO NPs can be effectively separated into ZnO particles and Zn ions: (1) ZnO NPs were recovered from the precipitated TX-114-rich phase after CPE as intact particles (Figure 2); (2) the hydrodynamic radii and primary particle sizes of pristine ZnO NPs obtained by CPE were statistically similar to those of pristine (Table 1, Figure 2), suggesting that particle size and size distribution were not affected during CPE procedure; (3) more than 82% of pristine ZnO NPs under all variable CPE conditions and ~93% of ZnO NPs in HA at 45 ◦C were recovered in the precipitates after CPE (Figure 2C), suggesting the efficacy of CPE for separating ZnO NPs as particle form; and (4) small amounts (less than 2%) of ZnO NPs were detected in supernatants as Zn ions (Figure 2C), indicating that only a minimum amount of Zn ions was released from ZnO NPs during CPE. This was also confirmed by low dissolution property (less than 2%) of ZnO NPs in different dispersants (Figure 3A). On the other hand, zeta potential values were affected and changed to slightly negative charges, approaching zero zeta potentials in all cases after CPE (Table 1). It is known that zeta potentials close to zero are optimal for the formation of NPs-TX-114 micelles [30,31]. The results obtained with pristine ZnO NPs suggest that TX-114-based CPE, especially in HA as a dispersant at 45 ◦C, can effectively separate ZnO particles and Zn ions without affecting primary particle size, size distribution, or morphology.

The dissolution property of ZnO NPs was highly affected by food matrices and cell culture medium (Figure 3B,C). The solubilities of ZnO NPs in coffee mix, skim milk, and milk ranged from ~30–49% (Figure 3B), although the pHs of ZnO-NP-spiked food matrices were close to neutral, except for sports drink (pH 3.3). It is known that ZnO can be dissolved into Zn ions under acidic conditions [39–41]. Thus, the high solubility (~91%) of ZnO in sports drink can be explained by the low pH of the sports drink. Increased solubilities of ZnO in other food matrices, such as coffee mix, skim milk, and milk, seem to be related to its interaction with food components. Meanwhile, the solubilities of ZnO NPs in MEM cell culture medium (pH 7.0) were ~18% to 25% over the incubation time (Figure 3C), which likely resulted from their interactions with various components found in the MEM [38,42]. The solubility of ZnO could, therefore, be highly affected by the presence of food or biomatrices.

The efficacy of CPE was also confirmed using ZnO-NP-spiked foods, showing the reliability and accuracy of the CPE for fate determination of ZnO NPs in foods: (1) DLS results recovered from the precipitates of ZnO-NP-spiked coffee mix, skim milk, and milk after CPE had similar hydrodynamic radii compared to pristine ZnO NPs (Supplementary Figure S1), indicating that ZnO NPs were well recovered as intact particle forms after CPE; (2) no particle forms were detected in the precipitates of ZnO-NP-spiked sports drink after CPE (Supplementary Figure S1), suggesting the complete dissolution of ZnO into Zn ions; (3) SEM-EDS analysis revealed the presence of Zn elements recovered from the precipitates of all ZnO-NP-spiked foods after CPE, except in sports drink (Figure 4A,B), which was in good agreement with DLS results; and (4) ~51–60% of ZnO were present as particles in coffee mix, skim milk, and milk, whereas almost all ZnO NPs were detected as Zn ions in sports drink after CPE, which is highly consistent with the dissolution properties of ZnO (Figure 3B). Indeed, no significant differences were found between solubilized Zn ions in food matrices (Figure 3B) and Zn ions recovered after CPE (Figure 4C, *p* > 0.05). Hence, the fate of ZnO NPs in food matrices can be determined as both intact particle and Zn ion forms by applying CPE, without affecting particle size and solubility. The total Zn recoveries (%) from both ZnO particles and Zn ions ranged from 92.6% to 97.5%, suggesting the accuracy of the analytical procedure.

The same trends were observed in commercial foods in which ZnO addition was indicated on product labeling. Most ZnO particles were present as Zn ions in liquid foods (functional peptide beverage and fruit juice), while the fate of ZnO was found to be mainly the particle forms in powdered foods (chocolate powder, powdered probiotics, and cereals, Figure 5). It is worth noting that another functional mineral beverage fortified with Zn gluconate was found to contain Zn ions after CPE (Figure 5F), supporting the reliability of the results. The pHs of liquid beverages were 2.7–4.7, which affected ZnO dissolution property due to the fact that ZnO dissolves more rapidly in acidic solutions. The slight presence of ZnO as particles in liquid beverages seems to be attributable to a Zn ion complex formed with other food components. The total Zn recoveries (%) from both ZnO and Zn ions ranged from 96.9% to 102.8%, and was calculated based on composition labeling in commercial foods and ICP-AES analysis, supporting the accuracy of the CPE procedure. Zn ion ratios in ZnO-NP-spiked powdered foods (Figure 4) were higher than those in commercial powdered or dried foods (Figure 5). The various processing steps used for commercial food products, such as formulation, mixing, and thermal or drying treatment, may increase the stability of ZnO in processed, powdered foods, contributing to its low dissolution in food matrices. Taken together, it is probable that ZnO as a food additive is primarily present as a particle in powdered or dried foods, but can be easily decomposed into Zn ions in liquid foods.

The intracellular fate of ZnO NPs, determined by CPE followed by ICP-AES analysis, revealed that ZnO NPs were taken up in particle forms, but slowly dissolved into Zn ions after a certain time inside cells (Figure 6A). A portion of dissolved Zn ions from ZnO in cell culture medium can be also rapidly taken up by cells, considering that the dissolution property of ZnO NPs in MEM is ~18–25% (Figure 3C). Vandebriel et al. also demonstrated that ZnO NPs can be taken up by cells by in both particle and ionic forms, which is consistent with our findings [43]. Paek et al. reported that Zn ions can be more rapidly and massively absorbed into the bloodstream after oral administration in rats [18], which may support the rapid cellular uptake of Zn ions compared to ZnO (Figure 6A). Taken together, both ZnO NPs and Zn ions can be internalized into cells, but the major fate of ZnO NPs is to become ionized inside cells over time. Gilbert et al. demonstrated that ZnO NPs were completely dissolved into Zn ions after cellular internalization [12]. Wang et al. also reported that ZnO NPs were internalized into cells by endocytosis and localized within acidic lysosomes, releasing Zn ions from internalized ZnO NPs [44–46]. ZnO NPs were reported to cause cytotoxicity associated with an increase in the Zn ions released inside cells [47,48]. However, contradictory results were obtained by confocal microscopy using a Zn–selective fluorescent probe, showing elevated Zn ion levels at 0.5–1 h and decreased Zn ions thereafter (Figure 6B). The discrepancy might be explained by complex formation between Zn ions and other molecular ligands inside cells, which was evidenced by Zn–S bond formation in tissues after oral administration of ZnO NPs in rats [15]. It is, therefore, strongly likely that ZnO NPs are dissolved into Zn ions and form Zn–molecular ligand complexes after internalization into cells, which is in good agreement with the results obtained by Gilbert et al. [12].

The intestinal transport and absorption fate of ZnO NPs, evaluated using in vitro models of human intestinal barriers and ex vivo everted small intestine sacs, was determined to be primarily Zn ion forms. This result suggests that ZnO NPs can be taken up by cells in both particle and ionized forms (Figure 6A), but most ZnO particles are dissolved into Zn ions during intestinal transit and absorption (Figure 7). Our previous report demonstrated that the ex vivo solubility of ZnO NPs in rat-extracted intestinal fluid was ~9% [23], supporting their high dissolution during intestinal transit. Thus, the major fate of absorbed ZnO NPs in the small intestine is likely to be the ionized forms. However, a small amount of ZnO can be also transported by M cells and absorbed as particle form (Figure 7B,C), suggesting different toxicokinetic behaviors of ZnO compared to those of Zn ions, as reported by previous research [18]. Intestinal transport of NPs by M cells was also demonstrated [23,49,50].

#### **4. Materials and Methods**

#### *4.1. Materials*

ZnO NPs (<100 nm), D-(+)-glucose, humic acid (sodium salt), TX-114, EDTA, formalin, ammonium chloride (NH4Cl), and monosodium phosphate (NaH2PO4) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Sodium hydroxide (NaOH), sodium chloride (NaCl), potassium chloride (KCl), calcium chloride (CaCl2), magnesium chloride (MgCl2), sodium bicarbonate (NaHCO3), nitric acid (HNO3), and hydrogen peroxide (H2O2) were supplied by Samchun Pure Chemical Co., Ltd. (Pyeongtaek, Gyeonggi-do, Korea). Conical-bottom glass centrifuge tubes (15 mL) were obtained from Daeyoung Science (Seoul, Korea). MEM, Roswell Park Memorial Institute (RPMI) 1640 medium, Dulbecco's modified eagle's medium (DMEM), heat-inactivated fetal bovine serum (FBS), penicillin, streptomycin, Dulbecco's phosphate-buffered saline (DPBS), and phosphate-buffered saline (PBS) were purchased from Welgene Inc. (Gyeongsan, Gyeongsangbuk-do, Korea). N-(6-Methoxy-8-quinolyl)-p-toluenesulfonamide (TSQ) was obtained from Enzo Life Science Inc. (Farmingdale, NY, USA). Matrigel® was from Corning Inc. (Corning, NY, USA). Transwell® polycarbonate inserts were purchased from SPL Life Science Co., Ltd. (Pocheon, Gyeonggi-do, Korea).

Commercial foods used for the spiking experiment were as follows: coffee mix, skim milk, milk, and sports drink. Commercial Zn-fortified products analyzed were as follows: chocolate powder (ZnO added), powdered probiotics (ZnO added), cereals (ZnO added), functional peptide beverage (ZnO added), fruit juice (ZnO added), and functional mineral beverage (Zn gluconate added), all from international brands found in markets in Seoul, Republic of Korea, in 2019.

#### *4.2. Characterization*

Primary particle size, shape, and chemical characterization were determined by FE-SEM (JSM-7100F, JEOL, Tokyo, Japan), equipped with EDS (Aztec, Oxford Instruments, Abingdon, UK). Zeta potentials and hydrodynamic radii of particles (1 mg/mL) were measured with Zetasizer Nano System (Malvern Instruments, Worcestershire, UK) after stirring for 30 min, sonication for 15 min (Bransonic 5800, Branson Ultrasonics, Danbury, CT, USA), and dilution (0.1 mg/mL).

#### *4.3. Optimization of CPE for Pristine ZnO NPs*

ZnO NPs (100 μg/mL) were dispersed in 7 mL of DDW, HA (final concentration of 10 μg/mL), or Glc (final concentration of 1% (*w*/*v*)) solution by stirring for 30 min, followed by sonication for 15 min. CPE was processed as described by Majedi et al. [33]. The suspensions (7 mL) were transferred to conical-bottom glass centrifuge tubes and the pH was adjusted to 10 by adding an NaOH solution. Next, 5% (*w*/*v*) TX-114 (0.5 mL) and 0.2 M NaCl solution (0.75 mL) were added to the suspensions. After dilution to 10 mL with DDW, the mixtures were incubated for 30 min at 37 or 45 ◦C and centrifuged for 5 min at 2500× *g* at 25 ◦C. The precipitates and supernatants were analyzed by ICP-AES (JY2000 Ultrace, HORIBA Jobin Yvon, Longjumeau, France) after digestion with HNO3 as described in Section 4.12.

#### *4.4. Fate Determination of ZnO NPs in Food Matrices*

ZnO NPs (10 mg) were spiked into 10 g of powdered foods, such as coffee mix and skim milk powder, and mixed well. Next, 0.1 g of the mixed powders were dispersed in 7 mL of HA solution, and the dispersions were stirred for 30 min followed by sonication for 15 min. In parallel, 10 mg of ZnO NPs were spiked with 100 mL of HA-added liquid foods, such as milk and sports drink. Spiked samples (1 mL) were then diluted to 7 mL by adding HA solution and stirring for 30 min, followed by sonication for 15 min. The Zn concentrations were chosen based on the ADI for Zn in Zn-fortified functional foods (2.55–12 mg) in the Republic of Korea [8]. The same procedure was applied as described in Section 4.3.

#### *4.5. Dissolution Property of ZnO NPs in Food*/*Bio Matrices*

ZnO NPs in DDW, different dispersants, or food matrices were prepared and stirred for 30 min. The same concentration of ZnO NPs for optimization of CPE and fate determination in food matrices were used. ZnO NPs (50 μg/mL) in cell culture medium MEM were incubated for 0.5, 1, 6, and 24 h with gentle shaking (180 rpm) at 37 ◦C. The ZnO suspensions were then centrifuged (16,000× *g*) for

15 min, and the supernatants were subjected to ICP-AES analysis after pre-digestion as described in Section 4.12.

#### *4.6. Fate Determination of ZnO in Zn-Fortified Commercial Foods*

Zn-fortified powdered foods (10 g) were homogenized in an agate mortar. Samples thus homogenized (0.1 g) were dispersed in 7 mL of HA solution, and the dispersions were stirred for 30 min and sonicated for 15 min. Commercial Zn-fortified beverages (1 mL) were diluted to 7 mL by adding HA solution, and the solutions were stirred for 30 min followed by sonication for 15 min. The same procedure was applied as described in Section 4.3.

#### *4.7. Cell Culture*

Human intestinal epithelial Caco-2 cells were purchased from the Korean Cell Line Bank (KCLB; Seoul, Korea). The cells were cultured in MEM containing 10% FBS, 100 units/mL of penicillin, and 100 μg/mL of streptomycin in a 5% CO2 incubator at 37 ◦C.

#### *4.8. Cellular Uptake and Intracellular Fate of ZnO NPs*

The cells were plated at a density of 1 <sup>×</sup> 106 cells/well and incubated with ZnO NPs (50 <sup>μ</sup>g/mL) for 0.5, 1, 6, and 24 h. After washing three times with DPBS, 5 mM EDTA in DPBS was used to treated the cells for 40 s in order to remove adsorbed NPs on the surface of cell membrane. After washing with PBS three times, the cells were harvested with a scraper, centrifuged, and re-suspended in 1 mL of DDW to determine the intracellular fate of ZnO NPs by CPE and cellular uptake quantification by ICP-AES. Cells in the absence of ZnO NPs were used as controls.

The suspended cells (1 mL) were transferred to conical-bottom glass centrifuge tubes and sonicated four times for 10 s on ice with a 150 W ultrasonic processor (Sonics & Materials Inc., Newtown, CT, USA). After dilution to 7 mL by adding HA solution, the same procedure was applied as described in Section 4.3.

#### *4.9. Intracellular Fate of ZnO NPs by Confocal Microscopy*

The cells were plated at a density of 2 <sup>×</sup> 104 cells on a glass coverslip, and ZnO NPs (50 <sup>μ</sup>g/mL) were treated for 0.5, 1, 6, and 24 h. After washing with DPBS, the cells were fixed with 500 μL of freshly made 3.7% formalin (containing 1.5% methanol) in DPBS on ice for 20 min. Next, 50 mM NH4Cl in DPBS was added and incubation was continued on ice for 30 min. After washing twice with DPBS, 50 μL of 30 μM fluorescent probe for Zn ions, TSQ, was added and incubated for 30 min in the dark at room temperature. Finally, the cells were rinsed three times with DPBS and visualized using a D-Eclipse C1 confocal microscope (Nikon Instech. Co., Kawasaki, Japan), equipped with Ar (488 nm) and HeNe (543 nm) lasers. Image acquisition and analysis were performed with EZ-C1 2.3 software (Nikon Instech. Co., Kawasaki, Japan). Each experiment was repeated twice on separate days.

#### *4.10. Intestinal Transport Fate of ZnO NPs*

Human Burkitt's lymphoma Raji B cells, supplied from the KCLB, were cultured in RPMI 1640 medium containing FBS (10%), non-essential amino acids (1%), L-glutamine (1%), penicillin (100 units/mL), and streptomycin (100 μg/mL) in a 5% CO2 incubator at 37 ◦C. The FAE model, mimicking M cells, was established as described previously [23,51]. After coating Transwell® polycarbonate inserts with Matrigel® matrix prepared in serum-free DMEM for 2 h, supernatants were removed, and inserts were then washed with serum-free DMEM. Caco-2 cells (1 <sup>×</sup> 106 cells/well) were seeded on the apical sides and grown for 14 days. Lymphoma Raji B cells (1 <sup>×</sup> 10<sup>6</sup> cells/well) were added to the basolateral sides, and these co-cultures were maintained for 5 days until trans epithelial electrical resistance (TEER) values reached 150–200 Ω cm2. The apical medium of the monolayers was then replaced by medium containing ZnO NPs (50 μg/mL), and incubation continued for 0.5 and 6 h.

Caco-2 monoculture was also used to evaluate the transported fate of ZnO NPs through intestinal epithelial tight junction barrier. Caco-2 cells (4.5 <sup>×</sup> 10<sup>5</sup> cells/well) were seeded on upper inserts and further cultured for 21 days (TEER <sup>≥</sup> <sup>300</sup> <sup>Ω</sup> cm2). Apical medium of the monolayers was then replaced by medium containing ZnO NPs (50 μg/mL), and incubation continued for 0.5 and 6 h.

Basolateral solutions (1 mL) were collected in a conical-bottom glass centrifuge tube, and diluted to 7 mL with HA solution. The same procedure was applied as described in Section 4.3.

#### *4.11. Intestinal Absorption Fate of ZnO NPs*

Eight-week-old male Sprague Dawley (SD; 200–250 g) rats were purchased from Koatech Co. (Pyeongtaek, Gyeonggi-do, Korea). Animals were housed in plastic laboratory animal cages in a ventilated room, maintained at 20 ± 2 ◦C and 60% ± 10% relative humidity with a 12 h light/dark cycle. Water and commercial complete laboratory food for rats were available ad libitum. Animals were environmentally acclimated for 5 days before treatment. All animal experiments were performed in compliance with the guideline issued by the Animal and Ethics Review Committee of Seoul Women's University (SWU IACUC-2019A-1), Republic of Korea.

Everted small intestine sacs were prepared as previously described [52]. Briefly, two male rats were fasted overnight (water available) and sacrificed by CO2 euthanasia. The small intestines were collected, washed three times with Tyrode's solution (containing 0.8 g of NaCl, 0.02 g of KCl, 0.02 g of CaCl2, 0.01 g of MgCl2, 0.1 g of NaHCO3, 0.005 g of NaH2PO4, and 0.1 g of glucose in 100 mL of distilled water), cut into sections (5 cm in length), and everted on a puncture needle (0.8 mm in diameter). After one end was clamped, the everted sacs were filled with 200 μL of Tyrode's solution and then tied using silk braided sutures. Each sac was placed in a six well plate containing 3 mL of ZnO NPs (50 μg/mL) for 0.5 and 6 h in a humidified 5% CO2 atmosphere at 37 ◦C. The solution in the interior sac was collected and the same procedure was applied after dilution to 7 mL with HA solution, as described in Section 4.3.

#### *4.12. ICP-AES Analysis*

All samples were pre-digested with 10 mL of ultrapure HNO3 at ~160 ◦C, and 1 mL of H2O2 solution was added and heated until the samples were colorless and until the solution was completely evaporated. The digested samples were diluted with 3 mL of DDW, and total Zn concentrations were determined by ICP-AES (JY2000 Ultrace, HORIBA Jobin Yvon).

#### *4.13. Statistical Analysis*

Results are expressed as means ± standard deviations. One-way analysis of variance with Tukey's test in SAS Ver.9.4 (SAS Institute Inc., Cary, NC, USA) was used to determine the significances of intergroup differences. Statistical significance was accepted for *p* values < 0.05.

#### **5. Conclusions**

Surfactant TX-114-based CPE was optimized and established for fate determination of ZnO in commercial foods and human intestinal cells. The solubility of ZnO was not affected by dispersants used for CPE, but was highly affected by food matrices or cell culture medium, showing dissolved fate to some extent. ZnO was found to be mainly present as particle forms in powdered or dried foods, whereas its major fate in liquid beverages was in Zn ionic form. On the other hand, ZnO NPs were internalized into cells as both particles and Zn ions, but slowly dissolved into Zn ions upon time, probably forming Zn–ligand complexes inside the cells. ZnO NPs were found to be transported through intestinal barriers and absorbed in the small intestine primarily as Zn ions. However, a portion of ZnO NPs could be absorbed into the body as particles. The toxicity of ZnO NPs is, therefore, likely to be mainly associated with Zn ion toxicity, but long-term potential toxicity resulting from particle forms cannot be completely excluded. These findings will be useful for understanding the potential toxicity of ZnO NPs and for their wide application to commercial foods at safe levels.

*Int. J. Mol. Sci.* **2020**, *21*, 433

**Supplementary Materials:** Supplementary materials can be found at http://www.mdpi.com/1422-0067/21/2/433/s1. Figure S1: Hydrodynamic radii of pristine ZnO and ZnO obtained by CPE in ZnO-NP-spiked foods.

**Author Contributions:** S.-J.C. conceived and designed the study, and wrote the manuscript; Y.-R.J. performed all experiments; J.Y. contributed to the discussion. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (2018R1A2B6001238), and by the Nano Material Technology Development Program (NRF-2018M3A7B6051668) of the National Research Foundation (NRF) funded by the Ministry of Science and ICT.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

#### *Article*
