Cytotoxicity of Nanoparticles

International Journal of Molecular Sciences

Esther E. Fröhlich 1 and Eleonore Fröhlich 2,*
1 Institute of Experimental and Clinical Pharmacology, Medical University of Graz, Universitätsplatz 4,
Graz A-8010, Austria; esther.froehlich@medunigraz.at
2 Center for Medical Research, Medical University of Graz, Stiftingtalstr. 24, Graz A-8010, Austria
* Correspondence: eleonore.froehlich@medunigraz.at; Tel.: +43-316-3857-3011; Fax: +43-316-3857-3009
Academic Editors: Michael Routledge and Bing Yan
Received: 23 February 2016; Accepted: 31 March 2016; Published: 6 April 2016
Abstract: Toxicity of nanoparticles (NPs) upon oral exposure has been studied in animals using physiological changes, behavior, histology, and blood analysis for evaluation. The effects recorded include the combination of the action on cells of the exposed animal and the reaction of the microorganisms that populate the external and internal surfaces of the body. The importance of these microorganisms, collectively termed as microbiota, for the health of the host has been widely recognized. They may also influence toxicity of NPs but these effects are difficult to differentiate from toxicity on cells of the gastrointestinal tract. To estimate the likelihood of preferential damage of the microbiota by NPs the relative sensitivity of enterocytes and bacteria was compared.
For this comparison NPs with antimicrobial action present in consumer products were chosen. The comparison of cytotoxicity with Escherichia coli as representative for intestinal bacteria and on gastrointestinal cells revealed that silver NPs damaged bacteria at lower concentrations than enterocytes, while the opposite was true for zinc oxide NPs. These results indicate that silver NPs
may cause adverse effects by selectively affecting the gut microbiota. Fecal transplantation from NP-exposed animals to unexposed ones offers the possibility to verify this hypothesis.
Keywords: silver; zinc oxide; nanotoxicology; cytotoxicity; antimicrobial effects

1. Introduction
Contact of nanoparticles (NPs) with the gastrointestinal microbiota occurs mainly via ingestion of solid food, water, cosmetics, and personal care products. In the latter NPs are included as active ingredients to inhibit, for instance, biofilm formation on teeth. However, an action of the NPs in these products on the human body is not intended. The broad use of NPs and its accumulation in
the environment warrants studies not solely focusing on acute toxicity in human cells. On the other hand, it is known that microbiota have a pronounced influence on human health [1–3]. Due to the close interaction between microbiota and the human organism, it might be hypothesized that effects of NPs on inherent or intestinal bacteria could have an effect on human health. The human body offers multiple sites for microbiota, like the oral cavity or the gastrointestinal tract. The latter harbors the largest community of bacterial members and has attracted interest from diverse areas of research. The intestinal microbiota comprises bacteria, fungi, viruses and arachaea, which form a complex ecosystem and live in a close relationship with the host. The environment, genetics, diet, and antibiotics shape and alter the microbiota and further influence the interaction
between the microbiota and the host. These influences can lead to microbial imbalance (dysbiosis) and may promote susceptibility to diseases. The intestinal microbiota fulfills multiple functions for the host including the processing of otherwise indigestible food compounds, synthesis of vitamins,
colonization resistance, regulation of the metabolism, and development of the immune system [4,5].

The latter, in particular, is thought to play an important role when it comes to various diseases including, but not limited to, autoimmune diseases, allergies, inflammatory bowel disease, intestinal infections, obesity, and even neurodevelopmental disorders [4,6,7]. The presence of an established
microbiota is of particular importance for normal immune responses, as mice raised under germ-free conditions have been shown to possess an underdeveloped immune system [7,8]. Furthermore, several studies have demonstrated that the microbiota shaped in the first years has an immense impact on an individual’s health status later on in life. Children born by caesarian section have been shown to be at greater risk of developing diseases like asthma, celiac disease or obesity [6]. Due to the growing appreciation of the importance of the microbiota with regard to diseases and the great achievements in sequencing methods and databases, the microbiota has even been suggested as a
diagnostic and prognostic biomarker [1,6]. As each individual harbors its own microbiota profile, much like a (personal) fingerprint, the use of microbiota as biomarker would be a big achievement in the field of personalized medicine [1]. The application of the microbiota as therapeutic has already started. Various studies have shown that in patients with recurrent Clostridium difficile (C. difficile)
infection fecal microbiota transplantation—a procedure in which fecal matter, or stool, is collected from a tested donor, mixed with a saline or other solution, strained, and placed in the patient by colonoscopy, endoscopy, sigmoidoscopy, or enema—was effective and has emerged as a promising new treatment for C. difficile infection. However, the choice of a donor microbiota needs to be made
wisely and further development and research in this field are still needed [4,9].
In vitro and in vivo studies have usually focused on direct toxic effects of NPs on the exposed cells and organisms, while few studies have investigated potential effects of NPs on the oro-gastrointestinal microbiota of the host. The effects of zinc oxide (ZnO) and cerium oxide (CeO2) NPs at 0.01 g/L and of 3 mg/L titanium oxide (TiO2) on microbiota isolates from one healthy donor cultured in a
custom colon reactor indicated that NPs affected short fatty acid production, hydrophobicity, sugar content of the extracellular matrix and electrophoretic mobility [10]. Other studies have focused on the composition of the microbiota and have found that silver (Ag) NPs of 14 nm did not alter the ratio of Bacteroides to Firmicutes after oral exposure of rats [11]. The lack of obvious changes in the
microbiota composition has also been reported after exposure of mice to 20 and 110 nm Ag NPs [12]. Williams et al., on the other hand, have detected size- and dose-dependent changes in ileal-mucosal microbial populations after oral gavage of rats with 10, 75 and 110 nm Ag NPs [13]. After treatment with 10 nm Ag NPs greater proportions of Firmicutes phyla, along with a decrease in the Lactobacillus
genus were observed. In the absence of morphological damage to enterocytes, the population of lactic acid bacteria was increased in the guts of Japanese quail that received colloidal 25 mg/kg Ag NPs in their drinking water [14]. Exposure to 110 nm Ag NPs caused a decrease in Firmicutes at the highest concentration of 36 mg/kg. 60–100 nm Ag NPs also reduced coliforms in the gut microbiota of weaning pigs [15]. When using in vitro exposures of the porcine microbiota samples effects were even more pronounced; coliforms were markedly and lactobacilli slightly reduced. In synthetic stool mixtures of 33 different isolates from a healthy human donor polyvinylpyrrolidone-capped 10 nm Ag NPsincreased the abundance of Escherichia coli (E. coli) [16]. These changes were observed at concentrations of ¥100 mg/L. Differences in particle concentrations (9–40 mg/kg) as well as in sizes (10–110 nm) could serve as explanation for the (above-mentioned) contradictory findings. Effects of larger Ag NPs could be explained by closer contact with the bacteria, while smaller NPs could be absorbed by the
intestinal tract. Furthermore, adult rats, mice, humans, quails, and weaning pigs differ in composition and stability of the gut microbiota. Different study results can be explained by the fact that the samples used for microbiome analysis may originate from luminal content or from gut tissue. Moreover, the site of specimen collection influences the results, as the number and composition of gut microbiota
changes depending on the location in the intestinal tract. In addition to that, the choice of method for the analysis of the microbiota generates method-specific results and may lead to additional bias [17]. The studies mentioned above do not give any indication of changes in organ histology, blood count, and clinical  chemistry of the exposed animals. Effects on microbiota, on the other hand, could explain some effects of NPs observed in in vivo studies. Oral exposure of rats to ZnO NPs induced not only liver damage but also behavioral changes in the treated animals [18]. Since behavioral changes due to alterations of the gut microbiota have been reported [19–21], it might be hypothesized that ZnO
caused the behavioral effects by affecting the gut microbiota.
Due to the conflicting results regarding the effects of Ag NPs on the microbiota in animal experiments, the limitations of the methodologies employed so far to assess effects on microbiota, and the limitations of rodent studies for the human microbiota, another approach will be used in this paper in order to evaluate the possibility that NPs act selectively on the gut microbiota. As most research on the microbiota has focused on the role of the bacteria, this review will also concentrate on the bacterial fraction in the gut, and compare the sensitivity of bacteria and intestinal cells to NPs. The selection of the NPs in this review is based on the hypothesis that selective damage of microbiota would be most likely for NPs that are taken up by the oral route and possess antimicrobial activity.

2. Oral Ingestion of Nanoparticles 

Ag, silica (SiO2), TiO2, and ZnO NPs are most relevant for oral ingestion because they are added as ingredients to food and contained in health care products. Since 2007, the use of NPs in food and
beverages has increased from 64 to 72 products. TiO2, ZnO, SiO2 are produced in the highest amounts, while Ag NPs are used in the highest number of products [22]. 2.1. Estimated Amounts of Daily Intake Ag NPs are used in food packaging, added as antimicrobial agent (E174) to beef, and serve in alginate gel coatings of carrots and asparagus to prevent water loss [23]. Ag NPs can also leach and
migrate from plastic bags and reusable food containers. Leaching of Ag from reusable food containers to food simulants in water was 5 ng/cm2 over 10 days, polyethylene bags released 10 ng/cm2 after 10 days, and 34 ng/cm2 of Ag were released after 3 use cycles from food storage containers [24–26].
Migration of Ag NPs from plastic food containers amounted to 1.66–31.46 ng/cm2 [27]. In addition, bioaccumulation in plants and fungi by Ag content of wastewater, incorporation into sewage sludge and spreading on agricultural fields as well as accumulation within food fish results in human exposure to Ag NPs [28,29]. Based only on food intake, daily Ag consumption is estimated to amount
to 20–80 g/day [30]. SiO2 NPs of different composition are labelled as E551, E554, E556, or E559, and used for instance as an anti-caking agent. The amount ingested daily is estimated to be 1.8 mg/kg (around 126 mg/day for a 70 kg person) [31]. The Scientific Committee on Food of the European Food Safety Authority has estimated the daily intake of SiO2 at 20–50 mg for a 60 kg person [32]. TiO2, gold (Au), platinum (Pt) and ZnO NPs are ingredients of sunscreens and toothpastes [33]. Highest concentrations of TiO2 NPs (E171), however, have been found in sweets [34]. Chewing gums and cookies contain around 1–5 g/mg of E171 and these authors estimated the daily intake to reach 0.45 mg/kg for an adult (around 31.5 mg/day for a 70 kg person) and 1 mg/kg for children.
Lomer et al. indicated daily ingestions of 2.5 mg for a 70 kg person in one study and 5.9 mg in another one [35,36]. Daily intake amounts estimated by Powell et al. [37] were at 5 mg and by Shi et al. at 300–400 g [38]. Based on survey data on daily food intake and fluid consumption rough estimates can be made of approximate concentrations in the gastrointestinal tract. The National Diet & Nutrition Survey reported 449.7 g of solid food (protein, carbohydrates, and fat) for men and 328.1 g for women [39]. Average fluid intake was at 1.98 L/day with considerable variations across countries
(lowest: 1.5 L, Japan; highest: 2.47 L, Germany; [40]). Under the assumption that a volume of 2.5 L of food (solid and fluids) is ingested, and based on the highest and lowest daily ingestions of the respective
NPs that have been published, the following concentrations can be estimated: 0.008–0.032 g/mL Ag NP, 9.3–50.4 g/mL SiO2 and 0.12–12.6 g/mL TiO2 NPs. ZnO NPs are included in nutritional supplements, such as multivitamins, and may be released from food packaging [41]. ZnO NPs were also detected in freshwater snails, showing that these animals accumulated ZnO particles present
in water [42]. Estimation of oral ingestion of ZnO NPs is complicated because ZnO ingestion may occur in addition to accidental uptake of health care products and oral uptake of food also through nutritional supplements. Dietary zinc deficiency is a global health problem and a dietary intake of 5–20 mg/person/day of zinc is recommended by the European Commission [32]. Zinc is essential
for cells and contained in a variety of proteins (transcription factors, enzymes, etc.). It also plays an important role in bacterial defense because secretion of zinc by mucosal surfaces makes bacteria more sensitive to immune cell killing [43]. Accidental uptake of ZnO NPs might cause zinc levels to increase into the toxic range of >50 mg/person/day [32]. To estimate the effects of NPs in food on gut microbiota and intestinal cells, the fact that the particles have more direct contact with bacteria in the lumen than with epithelial cells of the gastrointestinal tract also needs to be taken into account. The epithelial cells of the oro-gastrointestinal
tract are covered by a mucus layer that consists of a firmly and a loosely adherent layer and can reach a total thickness of up to 1000 m, which produces a strong barrier preventing both bacteria and NPs from penetrating cells [44]. Mucus restricts cellular access of NPs both by bonding to mucus fibers through ionic and hydrophobic interactions and by size filtering (for more details see for instance [45]). Residence time of food is shortest in the stomach (3–5 h) and longest in the large intestine (20–30 h) [46].
Residence time in the gastrointestinal tract and thickness of the mucus layer explain why the absorption
of 500 nm TiO2 particles in the stomach is lowest (0.06%) and highest in the large intestine (4%) [47]. 

2.2. Changes of Nanoparticle Properties in the Gastrointestinal Tract

Physicochemical parameters, size and surface properties are strongly influenced by contact with biological fluids [48]. For oral ingestion, mechanical forces and the prominent pH changes along the oro-gastrointestinal tract need to be considered as well. In the stomach, contractions of up to 150 mm Hg have been measured, but effects on NP agglomeration and aggregation are largely unknown [49].
Changes of pH along the oro-gastrointestinal tract are prominent in the fasted state, but usually buffered to a range of pH 2–6 in the presence of food. A low pH can increase dissolution of particles and enzymes in the digestive fluids can induce denudation of particles. Ag NPs show agglomeration in synthetic gastric fluid by partial dissolution and release of Ag+ [50,51]. Based on these findings particle growth was described as partial dissolution of Ag particles in the acidic environment and formation of AgCl on the particles’ surface by Ag+ in combination with Cl released from the environment. The influence of dissolution appeared to be more pronounced for smaller particles as they agglomerated
to a higher degree [52]. Less conclusive data has been obtained concerning the role of particle coating because polyvinylpyrrolidone-coated Ag particles from various sources behaved differently. Gastric fluid induced rapid dissolution of ZnO NPs [53], while NPs with a lower solubility in acid solutions, such as SiO2 NPs, agglomerated in gastric fluid [54]. Stabilization of NPs in food products reduced the extent of changes by oro-gastrointestinal fluids and SiO2 NPs integrated in food products deaggregated again after sequential treatment with saliva, gastric juice, and intestinal fluid. Several reports have shown the effect of the food matrix on dispersion and stability; SiO2 NPs were better dispersed in low fat coffee creamer than in water, agglomerated in saliva and deaggregated in gastric fluid containing digestive enzymes [52]. SiO2 NPs (E551) in coffee, soup, and pancake were nanosized to 30%, 13%, and 5% before and to 80%, 15%, and 15% after subsequent incubation with artificial saliva, gastric juice,
and duodenal juice + bile [55]. Furthermore, binding of macromolecules affects the biological action of NPs. The composition of the particle coating (commonly termed as “protein corona”) differs according to the composition of the surrounding media [48]. In the gastrointestinal tract, the protein corona consists of bile salts and proteins. The effects of fluids of the digestive tract on the biological effects of these NPs in intestinal cells have been reported differently. Digestion of Ag NPs with food compounds did not change uptake by Caco-2 cells, while digestion in the absence of food decreased cellular uptake to 60% [56]. Treatment with digestive solutions reduced the potential to generate reactive oxygen
species of SiO2 NPs without affecting cytotoxicity [57]. Adhesion to enterocytes by the presence of a protein corona was influenced in such a way that coating with bovine serum albumin and casein reduced adhesion of the particles to Caco-2 cells, while coating with meat extract had no effect on cell adhesion of 20, 100, and 200 nm polystyrene particles [58]. Incubation in murine intestinal fluid,
however, increased adherence of 20 and 200 nm particles to Caco-2 cells.

3. Antimicrobial Activity of Nanoparticles
Due to their antimicrobial activity, aluminum oxide (Al2O3), Ag, copper oxide (CuO), and ZnO NPs are the particles most likely to affect the gut microbiota [59]. Antimicrobial activity of TiO2 NPs was linked to photoactivation and effects were recorded only after illumination [60–62]. Naked SiO2 NPs did not possess prominent antimicrobial action, while SiO2 NPs grafted with antibacterial
agents, such as antibacterial polymers, quaternary ammonium compounds, and antimicrobial tricosan displayed antibacterial properties [63]. Antimicrobial effects caused by Al2O3 and CuO NPs are not relevant for human oral exposure because these particles are only contained in products that do not have a high probability of being ingested, such as abrasives and scratch-proof car paints (Al2O3) or antimicrobial coatings of pillowcases and socks (CuO) [64].
Therefore, the combination of exposure by the oral route and antimicrobial action restricts the candidates for potential damage of the microbiota to Ag and ZnO NPs.

4. Effects of Nanoparticles on Prokaryotic and Eukaryotic Cells

Toxicity of NPs to bacteria and mammalian cells is linked to the increased reactivity of these particles due to their large surface. However, cellular action of NPs differs between prokaryotic and eukaryotic cells due to their different composition and morphology. One important point is the absence of active uptake mechanisms in bacteria, except planctomycete Gemmata obscuriglobus [65].
The plasma membrane of mammalian cells measures 7.5 nm [66]. Uptake into the cells can occur either by diffusion or by active (endosomal) uptake mechanisms. These uptake routes are globally classified as clathrin-dependent (clathrin-mediated) and clathrin-independent. The latter consists of caveolin, clathrin- and caveolin-independent routes, and macropinocytotsis (Figure 1A). Clathrin- and caveolin-independent routes include Arf6-, flotillin-, Cdc42- and RhoA-dependent uptake [67]. For a more detailed summary of the uptake routes for NPs, the reader is referred to reviews dedicated to this topic, for example [67–69]. The cell wall of bacteria differs between gram-positive and gram-negative
bacteria. Gram-positive bacteria possess one cytoplasmic membrane and one thick peptidoglycan layer of an entire thickness between 20 and 80 nm, while the 5–10 nm thick cell wall of gram-negative bacteria consists of two cell membranes and one thin peptidoglycan layer [70] (Figure 1B). Additional differences between bacterial and mammalian cells include the around 50 times larger cell size of mammalian cells, the presence of membranes around the nucleus and of membrane-enclosed organelles (endosomes, lysosomes, autophagosomes, mitochondria, peroxisomes, etc.), and the cytoskeleton (Figure 2A). Bacteria, on the other hand, have a cell wall instead of a plasma membrane and a circuit chromosome devoid of histones [71] (Figure 2B). The majority of NPs that enter
mammalian cells by active mechanisms are transported to lysosomes. There, low-biodegradable NPs can accumulate, metal ions can be released and increase cytotoxicity [72]. When NPs reach the cytoplasm by diffusion across the plasma membrane, the release of metal ions and cytotoxicity are lower [73]. The lack of endocytosis in bacteria has the important consequence that NPs enter bacteria
only by destroying the bacterial wall and cell membrane [74] (Figure 2B). The NPs anchor to the bacterial wall and penetrate it causing structural changes. Ag NPs probably bind to thiol groups of membrane proteins. Electrostatic attraction to the cell membrane, however, is less likely because both membrane and Ag NPs are negatively charged. Another option is the formation of irregular pits at the bacterial surface leading to NP accumulation followed by progressive release of lipopolysaccharides and membrane proteins facilitating uptake by bacteria [75].

(This study continues, for a copy please contact the GPH office)