Harmful effects of metal(loid) oxide nanoparticles.

The incorporation of nanomaterials (NMs), including metal(loid) oxide (MOx) nanoparticles (NPs), in the most diversified consumer products, has grown enormously in recent decades. Consequently, the contact between humans and these materials increased, as well as their presence in the environment. This fact has raised concerns and uncertainties about the possible risks of NMs to human health and the adverse effects on the environment. These concerns underline the need and importance of assessing its nanosecurity. The present review focuses on the main mechanisms underlying the MOx NPs toxicity, illustrated with different biological models: release of toxic ions, cellular uptake of NPs, oxidative stress, shading effect on photosynthetic microorganisms, physical restrain and damage of cell wall. Additionally, the biological models used to evaluate the potential hazardous of nanomaterials are briefly presented, with particular emphasis on the yeast Saccharomyces cerevisiae, as an alternative model in nanotoxicology. An overview containing recent scientific advances on cellular responses (toxic symptoms exhibited by yeasts) resulting from the interaction with MOx NPs (inhibition of cell proliferation, cell wall damage, alteration of function and morphology of organelles, presence of oxidative stress bio-indicators, gene expression changes, genotoxicity and cell dead) is critically presented. The elucidation of the toxic modes of action of MOx NPs in yeast cells can be very useful in providing additional clues about the impact of NPs on the physiology and metabolism of the eukaryotic cell. Current and future trends of MOx NPs toxicity, regarding their possible impacts on the environment and human health, are discussed. KEY POINTS: • The potential hazardous effects of MOx NPs are critically reviewed. • An overview of the main mechanisms associated with MOx NPs toxicity is presented. • Scientific advances about yeast cell responses to MOx NPs are updated and discussed.

Introduction

Nanomaterials (NMs) are defined as “chemical substances or materials with particle sizes between 1 to 100 nm in at least one dimension” (ECHA 2020). Due to their nanometer size, they present huge surface-to-volume ratios, exhibiting unique physical and chemical properties (such as catalytic, optical, magnetic, electronic and mechanical) that are different from those of materials on a larger or “bulk” scale (Klaine et al. 2013). The exceptional properties exhibited by NMs have led to their incorporation in many products in various sectors such as agriculture, automotive, construction, cosmetics, electronics, environment, food, home appliance, medicine, petroleum and printing (NPD 2020).

The rapid expansion of production and use of NMs inevitably raised concerns about their safety for human health and the environment. The physical form and the chemical reactivity that makes NMs distinctive also provide them the potential to interfere with biological processes and produce hazardous effects. Humans can be intentional (through nanomedicine or personal healthcare products) or unintentionally exposed to NMs (released from food packaging); additionally, occupational exposure (as consequence of industrial processes) should also be considered (Lombi et al. 2019; Klaper 2020). Examples of intentional application of NMs in the environment include their use in environmental remediation (Guerra et al. 2018; Qian et al. 2020) or in agricultural practices (Servin et al. 2015; Usman et al. 2020). Unintentionally release of NMs in the environment includes the following: (i) the release due to the life cycle of products incorporating NPs, such as paints, cosmetics, sunscreens (Sun et al. 2016; Wu et al. 2020); and, (ii) accidental spills or industrial liquid effluents, such as those emitted by textile industries during the washing of nanotextiles (Yetisen et al. 2016). It was estimated that of the global NMs produced, 63–91% reach landfills, 8–28% are released into soils, 0.4–0.7% in natural water bodies, and 0.1–1.5% are released into the atmosphere (Keller et al. 2013).

NMs can be divided into five categories: carbon-based (single and multi-walled carbon nanotubes, graphene and fullerenes); metal-based (metal(loid) oxides; zerovalent metals such as iron, silver and gold); dendrimers (hyperbranched polymers, dendrigraft polymers and dendrons); semiconductor nanocrystals, known as quantum dots; and composites (constituted by two different NMs or NMs combined with larger, bulk-type materials; and NMs combined with synthetic polymers or resins) (EPA 2017).

Among NMs, metal(loid) oxide (MOx) nanoparticles (NPs) have received considerable attention largely due to their variety of uses namely in optics and electronics, healthcare, construction, automotive and personal care products (Laurent et al. 2018), and it will be the subject of the present review. The market for MOx NPs is expected to growth at a compound annual growth over 7% globally during the period of 2020–2025. However the current framework of uncertainty arising from the COVID-19 pandemic it may hinder the growth of this market sector (Mordor Intelegence 2020).

In the last five years, several review papers have been published about NMs, NPs or more specifically on MOx NPs. Some reviews refer more broadly to NMs, namely about their behaviour, fate, bioavailability and effects on the environment (Pulido-Reyes et al. 2017; Lead et al. 2018; Spurgeon et al. 2020; Zhao et al. 2020) or the toxicity mechanisms associated with NMs over algal cells (Chen et al. 2019). Within NMs, reviews on NPs, namely about the influence of their physico-chemical properties on ecotoxicology, in terrestrial and aquatic systems (Bundschuh et al. 2018; Nguyen et al. 2020; Roma et al. 2020), the effects on freshwater organisms (Deniel et al. 2019), genotoxicity (Mortezaee et al. 2019) or the mechanisms associated with cell dead by necrosis, apoptosis and autophagy (Mohammadinejad et al. 2019; Paunovic et al. 2020) have been published. A more specific review about the toxic effects of NiO in aquatic organisms (Meyer et al. 2020) was recently published.

The present work summarises the main mechanisms underlying MOx NPs toxicity. The biological models used to assess nanotoxicity are briefly presented, with particular emphasis on the yeast Saccharomyces cerevisiae as a valuable and alternative model in nanotoxicology. An updated overview of yeast cell responses to stress induced by MOx NPs is critically reviewed. Finally, current and future trends in the assessment of MOx NPs toxicity, regarding their possible impact on the environment and human health, are discussed.

Biological models used in nanotoxicology

Brief overview of the models used in nanotoxicity assessment

An array of biological models have been used in ecotoxicity studies, which include (in parentheses it can be found typical examples employed): bacteria (Escherichia coli and Vibrio fischeri), yeasts (S. cerevisiae), microalgae (Pseudokirchneriella subcapitata and Chlorella vulgaris), protozoa (Tetrahymena thermophila), rotifers (Brachionus plicatilis), crustaceans (Daphnia magna), annelids (Eisenia fetida), nematodes (Caenorhabditis elegans), cnidarians (Hydra attenuata), molluscs (Potamopyrgus antipodarum), echinoderms (Lytechinus pictus), amphibians (Xenopus laevis) and fishes (Danio rerio) (Juganson et al. 2015; Minetto et al. 2016; Libralato et al. 2017).

Although animal testing is still the predominant model use for the risk assessment of chemicals (Hartung and Rovida 2009), due to the pressure from public opinion and legal demand, supported by ethical reasons, the replacement of animals for cheaper and more human-relevant alternatives have been proposed based on the use of cell lines. Thus, different mammalian cell lines have been used in toxicity assays with MOx NPs (Al2O3, CuO, NiO, TiO2 and ZnO), comprising models of different human systems, such as respiratory, digestive, renal, immune and skin (Ivask et al. 2014; Naseer et al. 2018; Czyzowska and Barbasz 2020).

The yeast S. cerevisiae as an important tool in nanotoxicology

S. cerevisiae is the most commonly used yeast in industrial applications, receiving the status of Generally Recognized As Safe (GRAS) microorganism by the United States Food and Drug Administration (FDA 2018). This yeast is easy to manipulate and cultivate, does not require expensive ingredients in the formulation of the culture media and presents a short generation time. It was the first eukaryotic organism with the genome completely sequenced (Goffeau et al. 1996).

The yeast S. cerevisiae presents a cellular structure and organization related to animal cells. About 30% of genes associated with human diseases have a yeast orthologue (Foury 1997), which makes this yeast an attractive model organism to study diseases in humans. Mitochondrial respiration can be manipulated by the loss of mitochondrial DNA or by changing the growth conditions, making this yeast an appropriate model for elucidating the role of mitochondria in ROS generation, as well as mitochondrial diseases associated with oxidative phosphorylation (Malina et al. 2018); this information can be readily transported to higher eukaryotes via the Gene Ontology (Howe et al. 2018).

This yeast features a set of important tools that include the complete gene deletion collection (Giaever and Nislow 2014) and the possibility of achieving high-throughput data, such as obtained from transcriptomics, proteomics and metabolomics analysis (Braconi et al. 2016). The use of yeasts in the assessment of toxicity of environmental pollutants (including NMs) does not raise ethical issues and is well suited in a first toxicity screening, because reduces costs and toxic wastes and replaces/limits the use of animal models (dos Santos and Sa-Correia 2015).

However, this model also has limitations. The unicellular nature of this organism does not make possible to provide specific toxicological data about tissues or organs. In addition, it presents a cell wall (in contrast to animal cells), which can act as a barrier to toxicants, many efflux pumps and detoxification mechanisms, which can be the cause of the greater tolerance of yeasts to toxics, compared with eukaryotic cells of higher organisms (dos Santos et al. 2012; Braconi et al. 2016).

Global mechanisms underlying to MOx NPs toxicity

MOx NPs can present a toxic effect by several mechanisms, in some cases even by more than one. The main mechanisms are summarised below and depicted in Fig. 1.

Fig. 1

Outline of the main toxic mechanisms associated with metal(loid) oxide nanoparticles. Please see text for details. a NPs solubilisation: release of toxic ions. b NPs passage through wall pores versus release of metal ions on NPs-cell surface interface. c Direct effect on the cell surface. d Cellular uptake of NPs. e Oxidative stress. f Shading effect (on photosynthetic microorganisms): homoagglomeration of NPs. g Physical restraint: heteroagglomeration. CAT, catalase; Cl, chloroplast; CW, cell wall; ETC, electron transport chain; Mt, mitochondrion; NPs, nanoparticles; PM, plasma membrane; ROS, reactive oxygen species; SOD, superoxide dismutase

NPs solubilisation: release of toxic ions

MOx NPs dissolution, to a greater or lesser extent, is a common transformation process, which is dependent on their physico-chemical properties (chemical composition and size), presence of stabilizing agents and chemical composition of the medium, namely, pH, ionic strength (IS), presence of anions (phosphate and sulphate) and natural organic matter (Quigg et al. 2013; Amde et al. 2017).

Once in solution, the ions diffuse in the medium and reach the cells, where they produce a deleterious effect after intracellular accumulation (Fig. 1A). Metal ions, as charged chemical species, do not diffuse freely across the plasma membrane. Thus, different membrane transport proteins (pumps and channels) are involved in their influx (Argueello et al. 2012).

In certain MOx NPs, the ions dissolved seem to be the major factor in their ecotoxicity. This is the case of the toxic impact of ZnO NPs over bacteria (Heinlaan et al. 2008; Li et al. 2011; Wang et al. 2016), yeasts (Kasemets et al. 2009; Bayat et al. 2014), crustaceans (Heinlaan et al. 2008; Wiench et al. 2009; Vimercati et al. 2020), microalgae (Franklin et al. 2007; Miller et al. 2010; Lee and An 2013; Aravantinou et al. 2015; Schiavo et al. 2016) or mammalian cell lines (Brunner et al. 2006; Zhang et al. 2012), which is totally or mainly caused by solubilized Zn ions. A Multi-omics approach (transcriptomics, metabolomics and lipidomics) confirmed that metal ions mediated the main toxicological pathways of ZnO NPs in lung epithelial A549 cells (Dekkers et al. 2018). In the same line, CuO NPs toxicity over bacteria (Bondarenko et al. 2012), yeasts (Kasemets et al. 2009; Kasemets et al. 2013; Bayat et al. 2014; Bao et al. 2015), green algae (Aruoja et al. 2009; von Moos et al. 2015), crustaceans (Heinlaan et al. 2008; Jo et al. 2012) or human cell lines (Cohen et al. 2013; He et al. 2020) was partially or completely explained by dissolved Cu ions. The toxicity of NiO and SnO2 over P. subcapitata (Sousa et al. 2018b; Sousa et al. 2019b) or Mn3O4 and Y3O3 over S. cerevisiae (Moriyama et al. 2019; Sousa et al. 2019a) can also be mainly attributed to the respective ions leached from the NPs.

NPs passage through wall pores versus release of metal ions on NPs-cell surface interface

The cell wall, present in plants, in most bacteria, yeasts and in many microalgae, is the primary site of interaction and the first barrier to the passage of NPs from the extracellular medium to the cytoplasm. This cellular structure is absent in animal cells and protozoa. The second barrier is the plasma membrane, common to all cells. Chemical composition and architecture of the cell wall vary according to the organism, being their thickness an important factor in the determination of the NPs internalisation (Chen et al. 2019). This cell structure is generally seen as a porous matrix. For instance, the yeast S. cerevisiae presents cell wall pores of 200 nm, which can be enlarged up to 400 nm under stress conditions (Pereira and Geibel 1999). Cell wall pore diameter of 7–8 nm in marine macro algae (Zemke-White et al. 2000) and pore channels with an average diameter of 20–200 nm in unicellular green microalgae were also described (Anissimova and Staer 2018). It is conceivable that, in organisms with cell wall, NPs and NPs homoagglomerates smaller than wall pores can pass through this cellular structure and reach the plasma membrane (Fig. 1B, top). Conversely, NPs and NPs homoagglomerates larger than pore size are very unlikely to pass through cell wall (sieve effect).

It was described that CeO2 NPs, coated with polyvinylpyrrolidone and presenting a size of 4–5 nm, could cross the cell wall of the microalga Chlamydomonas reinhardtii being internalised into intracellular vesicles (Taylor et al. 2016). Using different microscopy techniques, the internalisation of CuO NPs in S. cerevisiae (Vasco et al. 2017) and in the C. reinhardtii (Yin et al. 2020) was also described. In the case of C. reinhardtii, CuO NPs were largely accumulated in the vacuoles (Yin et al. 2020). Conversely, it has been suggested that the internalisation of CeO2 NPs (with a nominal size of 25 nm and presenting in solution agglomerates with an average size of 146 nm) in C. reinhardtii was rather unlikely (nee Rohder et al. 2018). A similar conclusion was achieved in other studies with algae (Rodea-Palomares et al. 2011; Pulido-Reyes et al. 2015; Angel et al. 2015) and cyanobacterium (Rodea-Palomares et al. 2011). In this line, the examination by transmission electron microscopy (TEM)-energy-dispersive X-ray spectroscopy of yeast cells treated with CuO (Bao et al. 2015), NiO (Sousa et al. 2018a), or ZnO NPs (Zhang et al. 2016) did not detect MOx NPs inside cells, suggesting that these NPs could not be taken by yeast cells. In resume, although examples of NPs internalisation in organisms with cell wall can be found in the literature, this is a debatable issue, especially the passage through the cell wall of agglomerated NPs.

However, different studies have attributed the toxic effects of MOx NPs to the NPs themselves rather than to the ions coming from them. This is the case of NiO (Sousa et al. 2018a), SiO2 (Sousa et al. 2019a) and TiO2 (Bayat et al. 2014) over yeasts as well as Al2O3, CeO2, Fe3O4, Mn3O4 and WO3 NPs on microalgae (Aruoja et al. 2015; Angel et al. 2015; Sousa et al. 2019b). In this context, MOx NPs can present a toxic effect by an indirect mechanism: particles adhere tightly to the cell wall of the microorganisms and enhance the release of metals at the NP-cell wall interface, leading to the activation of toxic responses (Fig. 1B, bottom). This indirect mechanism was suggested for to explain the antibacterial activity of Fe2O3, Co3O4 and NiO NPs (Wang et al. 2016) and the toxicity of CuO and NiO on yeasts (Kasemets et al. 2013; Bao et al. 2015; Sousa et al. 2018a) and CeO2 NPs on microalgae (Angel et al. 2015).

Direct effect on the cell surface

After contact and adsorption to the cell wall, NPs can clog the pores of the wall, limiting the exchange of chemical species (including nutrients) between the surrounding medium and the cell (Fig. 1C, top) or induce physical damage to the cell wall (Fig. 1C, bottom).

In agreement with the first possibility, it was described that C. reinhardtii incubated with TiO2 presented the cell surface coated with NPs, which can hinder the exchange of substances between the cell and the surrounding milieu (Chen et al. 2012). The functionalization and the type of functionalization of CuO and ZnO can influence the level of NPs adsorption to S. cerevisiae and C. reinhardtii cell wall and the respective anti-fungi and anti-algal activity (Halbus et al. 2019; Halbus et al. 2020).

The damage of the cell wall of algae and yeast cells incubated with ZnO NPs was described (Suman et al. 2015; Zhang et al. 2016; Babele et al. 2018). In the same way, it has been reported that TiO2 NPs or TiO2 NPs surface-bound humic acid adhered to algal cells (Chlorella spp., Karenia brevis, Nitzschia closterium and Skeletonema costatum) and could destroy the cell wall and enter the cells, inducing plasmolysis (Lin et al. 2012; Xia et al. 2015; Li et al. 2015). In the same mode, it has been proposed that the toxic effect of CeO2 NPs on P. subcapitata can be mediated, mainly, by a physical effect due to a close adsorption of the NPs on the cell surface (Manier et al. 2013). This possibility is in line with other observations that describe that CeO2 NPs are strongly adsorbed to Anabaena sp. and completely disrupt cyanobacterium cell wall and membrane (Rodea-Palomares et al. 2011). It is important to note that the adhesion of MOx NPs to the cell wall does not necessarily imply the triggering of a toxic effect. For instance, the attachment of La2O3 NPs to the cell wall did not produce morphological changes on Chlorella spp. (Balusamy et al. 2015).

Cellular uptake of NPs

After crossing the cell wall (in organisms with a cell wall and when such passage is possible), MOx NPs meet the cell membrane and two types of events can occur: damage of the membrane (due to physical disruption) (Chen et al. 2019) or passage through the membrane by endocytosis (Fig. 1D), a process well known in mammalian cells (Oh and Park 2014). Endocytosis was also suggested in microalgae and bacteria but their predominance and mechanisms are unknown (von Moos et al. 2014).

Once inside the cells, MOx NPs can undergo several alterations, such as redox transformations and complexation (Chen et al. 2019) or can be solubilized inside of acidic lysosomes (Fig. 1D) and exerts their toxicity by a Trojan horse-type mechanism (Oh and Park 2014). The intracellular dissolution mechanism affords the trafficking of toxic metal ions into the cells. Thus, CuO and ZnO NPs intracellular dissolution and release of Cu and Zn ions, respectively, were described in different mammalian (including human) cell lines (Xia et al. 2008; Cronholm et al. 2013; Condello et al. 2016; He et al. 2020).

Oxidative stress

Oxidative stress (OS) occurs when it is observed an imbalance between the generation of reactive species (RS) and the level of antioxidant defences, either enzymatic or non-enzymatic (Halliwell and Gutteridge 2015). MOx NPs themselves, or the metals released from them, can present a pro-oxidant potential, i.e. the capacity to generate the production of RS or hindering/consuming antioxidant defences (Nel et al. 2006). Reactive oxygen species (ROS) production (such as superoxide radical (O2·−), hydroxyl radical (HO) and hydrogen peroxide (H2O2)) can occur under a cell free milieu, i.e. extracellularly (abiotic ROS) or via interaction of NPs (or the released ions) with biologicals systems (biotic ROS) (Fig. 1E).

ROS can be generated at the NP surface (Fig. 1E, top). The bioavailability and valence state of redox-active elements influences strongly the level of ROS generation by NPs. In a general way, the capacity of MOx NPs to generate ROS is dependent on their chemical composition, purity, (particle) size, shape and surface reactivity (von Moos and Slaveykova 2014). In certain NPs, such as TiO2, the ability to produce ROS may also require light or an ultraviolet source to excite the NPs surface (Xia et al. 2008; Guo et al. 2011) (Fig. 1E, top). ROS generation, in abiotic conditions, by CeO2, Co3O4, CuO and Sb2O3 NPs were described (Xia et al. 2008; Bayat et al. 2014; Aruoja et al. 2015).

Intracellular (biotic) ROS can have origin in the endoplasmic reticulum, peroxisomes and in electron transport processes in mitochondria and chloroplasts (in eukaryotic photosynthetic organisms) (Lesser 2006; del Rio and Lopez-Huertas 2016) (Fig. 1E, bottom). MOx NPs, even containing redox-inactive metals, such as NiO (Siddiqui et al. 2012; Ahamed et al. 2013; Oukarroum et al. 2017; Sousa et al. 2018c; Sousa et al. 2018b) and ZnO NPs (De Berardis et al. 2010; Kumar et al. 2011; Alarifi et al. 2013; Lu et al. 2015; Ng et al. 2017) can induce intracellular ROS and oxidative stress. It was shown that P. subcapitata algal cells exposed to NiO NPs presented a reduced activity of the photosystem II (ɸPSII) and a decreased electron flow in the electron transport chain (ETC). The electrons deflected from photosynthetic ETC probably are used to generate ROS (Sousa et al. 2018b) (Fig. 1E, bottom). Compatible with this possibility, intracellular ROS accumulation and decrease of ɸPSII were also observed in microalgae (P. subcapitata or C. reinhardtii) exposed to Al2O3, CeO2, Cr2O3, Mn3O4, SiO2 or SnO2 NPs (Rodea-Palomares et al. 2012; da Costa et al. 2016; Sousa et al. 2019b) and in the cyanobacterium Anabaena CPB4337 exposed to CeO2 NPs (Rodea-Palomares et al. 2012).

It was proposed that in P. subcapitata cells exposed to NiO NPs, the disturbance of photosynthetic performance and the increase of intracellular ROS, combined with a reduction of metabolic (esterasic) activity may cause the arrest of algal cell cycle which, in turn, may origin the increase of cell volume and the appearance of aberrant morphology and, ultimately, the arrest of algal growth (Sousa et al. 2018b). Similarly, it was observed that Al2O3, Mn3O4 and SiO2 NPs induced the growth inhibition of P. subcapitata as a consequence of the cumulative effect of adverse outcomes, such as intracellular ROS accumulation, loss of metabolic activity and reduction of ɸPSII (Sousa et al. 2019b). A reduction of chlorophyll a content and increase of ROS was also observed in C. minutissima exposed to Co3O4 NPs (Sharan and Nara 2020).

OS is associated with the damage of biological molecules such as cellular lipids (via lipid peroxidation, LPO), carbohydrates, proteins and DNA (Halliwell and Gutteridge 2015) being considered a major mechanism of NPs toxicity (Nel et al. 2006; Xia et al. 2006). Accordingly, ROS production by various MOx NPs, namely, Al2O3, Ce2O, CuO, Mn3O4, NiO, SiO2, SnO2, TiO2 and ZnO with the consequent cell oxidative disturbances, which include, LPO and cell membrane damage (loss of integrity), overwhelmed antioxidant defence system, reduced mitochondrial function, chromatin condensation, DNA damage and cell death via apoptotic pathway, over different biological models have been described; examples are the following: bacteria (Kumar et al. 2011; Rodea-Palomares et al. 2012), yeasts (Zhang et al. 2016; Babele et al. 2018; Sousa et al. 2018c; Sousa et al. 2019a), freshwater and marine microalgae (Rodea-Palomares et al. 2012; von Moos et al. 2015; Xia et al. 2015; Suman et al. 2015; von Moos et al. 2016; Oukarroum et al. 2017; Dauda et al. 2017; Sendra et al. 2018; Sousa et al. 2018b; Sousa et al. 2019b), carp (Cyprinus carpio) larva (Naeemi et al. 2020) and human cell lines (Karlsson et al. 2008; Park et al. 2008; Lu et al. 2015; Duan et al. 2015; Rajiv et al. 2016; Subramaniam et al. 2020).

Shading effect (on photosynthetic microorganisms): homoagglomeration of NPs

MOx NPs, in aqueous medium, can interact with each other and form clusters of NPs (homoagglomerates), which can have an important impact on their bioavailability, fate in the environment and toxicity (Vale et al. 2016). The homoagglomeration process (as well as heteroagglomeration—please see below) depends on the concentration and characteristics of the NPs (chemical composition, morphology and charge), the NPs surface functionalisation, the physico-chemical properties of the medium where they are suspended, namely, IS (influenced by water salinity and hardness), pH of the solution (which affects the NPs surface charge) and the existence of natural organic matter (NOM). NOM can adsorb to the NPs surface and origin their steric stabilisation or act as bridges promoting the agglomeration of NPs (Amde et al. 2017; Yu et al. 2018). The adsorption of NOM to NPs, with the formation of coated NPs (corona-coated NPs), can be associated with the transfer of these NMs from algae to higher trophic levels consumers (crustaceans and fish) presenting unknown environmental and human health risks (Xu et al. 2020).

Aqueous suspensions of dispersed or homoagglomerated MOx NPs are, sometimes, opaque. Due to light absorption or scattering by NPs or NPs homoagglomerates, a reduction of the light availability can occur (shading effect, Fig. 1F). This effect can influence the photosynthetic efficiency of organisms like cyanobacteria and algae (Navarro et al. 2008). In agreement with this possibility, a significant decrease in light absorbance of C. reinhardtii algal suspensions, in comparison with control (NPs free), due to CuO NPs at concentrations higher than 1 mg L−1 was described (Cheloni et al. 2016). Sadiq et al. (2011) suggested that the growth inhibition and chlorophyll content reduction observed in algal cells incubated with Al2O3 could be attributed to the decrease of light availability owing to the attachment of the NPs onto cell wall of Chlorella spp. A shading effect was also attributed to Co NPs in the inhibition of S. costatum growth (Chen et al. 2018).

However, other authors did not observe any significant effect on the 72 h growth of the alga P. subcapitata, regardless of the concentrations of CeO2, CuO and ZnO NPs tested; in the case of TiO2, even for the relative opaque suspensions, containing easily settled NPs homoagglomerates, a growth reduction was not observed (Aruoja et al. 2009; Rogers et al. 2010; Hartmann et al. 2010). A shading effect was also excluded, as the main mechanism of ZnO nanotoxicity on Chlorella spp. (Ji et al. 2011) and TiO2 on K. brevis and S. costatum, although the algae were almost entirely covered by TiO2 NPs agglomerates (Li et al. 2015). This means that, at least, for the organisms and NPs reported above, the shading effect does not appear to be the main mechanism of toxicity.

Physical restraint: heteroagglomeration

Another possibility of NPs inducing a toxic effect is through the co-agglomeration of NPs (or NPs homoagglomerates) with cells—formation of heteroagglomerates. These micro or even macroscopic heteroagglomerates can lead to the reduction of light, nutrients, or oxygen availability, due to the trapping of cells inside the agglomerates (Fig. 1G). In this context, Aruoja et al. (2009) described the co-agglomeration of TiO2 NPs with algal cells of P. subcapitata; the formation of large clusters entrapped almost all algae. Thus, it was suggested that the observed growth inhibition could be attributed to the reduced availability of light in entrapped cells. According to the authors, the shading effect may contribute (or play a major role) to the toxicity of TiO2 NPs on algae. A similar mechanism (limitation of essential nutrients due to physical restriction caused by heteroagglomeration) has been proposed for cyanobacterium cells exposed to CeO2 NPs; bacteria incubated with CeO2 NPs were found completely entrapped inside the heteroagglomerates, which leads the authors not to exclude the possibility that the nutrients transport into the cells may have been severely impaired (Rodea-Palomares et al. 2011). However, phosphate or micronutrients depletion, due to the adsorption on NPs surface, alone, did not allow to explain CeO2 toxicity to P. subcapitata (Rogers et al. 2010).

The formation of heteroagglomerates and the respective algal entrapment, may, by itself, not induce a toxic effect. In fact, it was reported that Al2O3, Mn3O4 and SnO2 NPs form heteroagglomerates with algal cells. The observation of cells inside, at the periphery of the structures and in the surrounding medium, together with the easy dispersibility of the agglomerates makes it hardly plausible that the toxicity induced by these NPs may be due to nutritional limitations induced by hetero agglomeration (Sousa et al. 2019b).

Yeast responses to MOx NPs stress

Although different yeasts have been used in the assessment of antifungal properties of MOx NPs, the main workhorse in ecotoxicity studies with these NMs is the yeast S. cerevisiae. Thus, unless stated otherwise, when in the text below it is mentioned the word “yeast”, it means S. cerevisiae.

The knowledge of the cellular responses (toxic symptoms exhibited by yeasts) to MOx NPs (described below and depicted in Fig. 2) is important in the identification of potential targets and biomarkers of the toxic action of NPs. Additionally, these information can be useful in the elucidation of the specific modes of action by which MOx NPs interact with the eukaryotic cells and affect their physiology and metabolism.

Fig. 2

Schematic representation of the principal molecular and physiological yeast responses to metal(loid) oxide nanoparticles. Please see text for details. CW, cell wall; ΔΨm, mitochondrial membrane potential; FUN-1, 2-chloro-4-(2,3-dihydro-3-methyl-(benzo-1,3-thiazol-2-yl)-methylidene)-1-phenylquinolinium iodide; GSH, reduced glutathione; MT, mitochondria; mtDNA, mitochondrial DNA; N, nucleus; NPs, nanoparticles; PBN, N-tertbutyl-α-phenylnitrone; PM, plasma membrane; ROS, reactive oxygen species; V, vacuole

Inhibition of cell proliferation

The impact of MOx NPs on the ability of a cell to divide (yeast proliferation) has been evaluated either by a clonogenic assay (viability assay) or in a liquid culture medium (growth inhibition assay). A reduction of the % of viability, in a dose-dependent way, was observed when yeast cells were exposed to different NPs: Al2O3, NiO, Mn3O4, SiO2 and SnO2 (Sousa et al. 2018a; Sousa et al. 2019a). 24 h-IC50 values of 4.8 mg/L CuO (Kasemets et al. 2013) and 5–20 mg/L ZnO (Babele et al. 2018) were described. Higher NPs concentrations were required to inhibit growth. Thus, yeast growth inhibition, in rich medium, was described for CuO (8 h-EC50 of 20.7 mg/L) and for ZnO NPs (8 h-EC50 121–134 mg/L) in malt extract (ME) medium (Kasemets et al. 2009). However, no growth inhibition in yeast peptone dextrose (YPD) broth was observed when yeast cells were exposed to 100 mg/L Al2O3, NiO, Mn3O4, SiO2, and SnO2 NPs (Sousa et al. 2018a; Sousa et al. 2019a).

The incubation of yeast cells in water or a buffer solution revealed to be a more sensitive method for the assessment of MOx NPs toxicity rather than protein-rich liquid culture medium (ME or YPD) (Kasemets et al. 2013; Suppi et al. 2015; Sousa et al. 2018a; Sousa et al. 2019a). This difference of sensitivity can be partially explained by the presence of proteins in the culture media which can be adsorbed to NPs, and form a protein layer, which is called protein corona (Kharazian et al. 2016); the “coating” of MOx NPs with proteins can reduce their toxicity (Nguyen and Lee 2017). Additionally, the organic ligands, in rich medium, complex the toxic ions making them less bioavailable and thus less toxic (Hughes and Poole 1991).

Cell wall damage

The exposure of S. cerevisiae to ZnO NPs induced cell wall damage (Babele et al. 2018). Yeasts with morphology changed from elliptical to irregular shape and with cell wall deformed, with sunken areas or deficiencies or even broken or partially broken were described after being exposed to ZnO NPs (Zhang et al. 2016). Yeast cells treated with TiO2 and CuO NPs presented the wall with an undulating appearance. In cells exposed to CuO NPs, a cell wall with a thicker and folded appearance was also described (Bayat et al. 2014).

Mutant strains with deficient genes associated with cell wall organization and biogenesis (such as KRE6, HOC1 and BCK1) were more sensitive to ZnO NPs than the respective wild-type strain; the increased susceptibility of ZnO treated cells to sonication confirmed that ZnO NPs affected cell wall function and integrity (Márquez et al. 2018).

An increase in the chitin content, a marker of cell wall stress and an upregulation in the expression of chitin synthesis (CHS1, CHS3 and CHS5) genes were described in S. cerevisiae treated with ZnO NPs (Babele et al. 2018). Similarly, the yeast Pichia pastoris incubated with TiO2 NPs presented an increased chitin content in the cell wall (Liu et al. 2016).

Modification of metabolic activity

Another option used to assess the toxicity of MOx NPs in yeasts is to evaluate its impact on the general metabolic status of the cells. For this purpose, different fluorescent probes have been used such as fluorescein diacetate (FDA), 2-chloro-4-(2,3-dihydro-3-methyl-(benzo-1,3-thiazol-2-yl)-methylidene)-1-phenylquinolinium iodide (FUN-1), and resazurin (Alamar Blue reagent). Yeast cells exposed to CuO NPs presented a reduction of metabolic (reductase) activity (Mashock et al. 2016). In the same way, yeasts treated with Al2O3, Mn3O4, NiO, SiO2 and SnO2 NPs presented a reduced metabolic capability traduced by a decreased ability to process the probe FUN-1 and by a diminished esterase activity (Sousa et al. 2018a; Sousa et al. 2019a). It was suggested that the reduction of esterasic activity could be a consequence of OS (Sousa et al. 2019a), since intracellular ROS accumulation could lead to the oxidation of sensitive amino acid residues of the enzymes, such as those containing aromatic side chain or sulfhydryl groups (Cecarini et al. 2007).

A disturbance in lipids biosynthesis was described in yeasts treated with MOx NPs. The modification in the cellular distribution of lipid biosynthetic enzymes (Fas1 and Fas2) and the induction and accumulation of lipids droplets (LDs) in yeast cells treated with ZnO (Babele et al. 2018) or CuO NPs (Bayat et al. 2014) was observed. In the same way, an enhancement in LDs, a decrease of the relative content of saturated fatty acids, an increase of the content of unsaturated fatty acids (UFA), and an upregulation of the genes involved in UFA synthesis (FAD9A, FAD9B, FAD12 and FAD15) was described in the yeast P. pastoris exposed to TiO2 NPs (Yu et al. 2015).

Using powerful techniques such as proteomics, metabolomics and system biology-based pathway analysis it was found that in S. cerevisiae cells exposed to ZnO NPs, almost 40% of proteins are down regulated and the metabolome deregulated. More specifically, it was found that a wide range of key metabolites involved in central carbon metabolism, cofactors synthesis, amino acid and fatty acid biosynthesis, purines and pyrimidines, nucleoside and nucleotide biosynthetic pathways were repressed (Babele 2019). By a similar approach (transcriptomic and proteome profile analysis), it was found that ZnO and ZnFe2O4 NPs induced dysfunction of cholesterol biosynthesis in an alveolar rat macrophage cell line (Doumandji et al. 2020).

Intracellular ROS generation

S. cerevisiae cells exposed to Al2O3, Mn3O4, NiO, SiO2, SnO2 and ZnO NPs accumulated significantly more intracellular ROS than control (Zhang et al. 2016; Babele et al. 2018; Sousa et al. 2018c; Sousa et al. 2019a). The co-incubation of yeast cells with Al2O3, Mn3O4, NiO, SiO2 and SnO2 NPs and the antioxidants ascorbic acid or N-tertbutyl-α-phenylnitrone quenched intracellular ROS and significantly restored cell viability and metabolic activity, suggesting a ROS-mediated mechanism in cytotoxicity induced by these NPs over yeast cells (Sousa et al. 2018c; Sousa et al. 2019a). TiO2 NPs also induced a dose-dependent accumulation of intracellular ROS in the yeast P. pastoris (Chen et al. 2019).

Mitochondrial respiratory chain seems to be an important source of ROS since wild-type yeast cells under nitrogen atmosphere as well as mutants lacking respiratory chain (without mitochondrial DNA, ρ0) presented decreased intracellular levels of ROS and augmented resistance to NiO NPs (Sousa et al. 2018c). It was hypothesised that the Ni ions released from NPs can disturb the electron transport at mitochondria by displacing iron from the ETC of the inner mitochondrial membrane. Probably, the leakage of electrons from the ETC to oxygen lead to the production of O2·− which, in turn, is most likely the main origin of H2O2 (Sousa et al. 2018c). In fact, both ROS types (O2·− and H2O2) were detected in S. cerevisiae cells treated with NiO NPs (Sousa et al. 2018c).

Transcriptomic analysis revealed that Y3O3 NPs, at high concentration (1–4 g L−1), induced the upregulation of oxidative stress genes in S. cerevisiae (Moriyama et al. 2019).

Reduction of non-enzymatic and enzymatic antioxidant defences

Yeast cells present non-enzymatic (which the most abundant is reduced glutathione, GSH) and enzymatic defence mechanisms such as superoxide dismutases (SOD1 and SOD2), catalases (CatT and CatA), glutathione peroxidases (such as Gpx3 and Grx1) and thioredoxin peroxidases (like, Tsa1 and Prx1) to preserve intracellular redox equilibrium and survive (Jamieson 1998; Herrero et al. 2008).

Reduced glutathione seems to be involved in the fight against OS induced by NiO as revealed by the decrease of cellular GSH level in yeasts incubated with these NPs. Supporting this observation, it was shown that mutant strains without (gsh1Δ) or with a reduced level of GSH (gsh2Δ) presented augmented levels of ROS and susceptibility to NiO NPs (Sousa et al. 2018c). The depletion of the GSH levels can be a consequence of the increased consumption in the scavenging of free oxygen radicals induced by NPs or due to the affinity of metal ions (such as Ni2+) to cysteine residue of GSH, leading to a reduction of cellular antioxidant defences (Sousa et al. 2018c). Similarly, TiO2 NPs induced an accentuated reduction in GSH concentration in the yeast P. pastoris (Liu et al. 2016).

Single-gene mutant strains devoid of the main antioxidant enzymatic defences (Sod1p, Sod2p, Ctt1p, Cta1p, Gpx3p, Grx1p, Tsa1p and Pprx1p) did not present augmented vulnerability to NiO NPs comparatively to wild-type strain (Sousa et al. 2018c); the absence of a sensitive phenotype, in these deleted strains, can be attributed to gene redundancy or the presence of compensatory parallel pathways (Dawes 2004). A similar effect was observed with sod1Δ and sod2Δ mutant strains exposed to CuO NPs (Kasemets et al. 2013). However, the yeast P. pastoris treated with TiO2 NPs presented a downregulation of the genes (cSOD, GLR1, GPX1 and TRR1) encoding to enzymes associated with ROS scavenging system (Liu et al. 2016).

Loss of cell membrane integrity

One of the outcomes of high ROS levels is the lipid peroxidation. Large-scale lipid peroxidation leads to increased membrane fluidity, efflux of cytosolic components and, ultimately, loss of plasma membrane integrity and cell death (Avery 2011). Consistent with this scenario, it was shown that the exposure of yeasts to NiO NPs leads to a progressive depolarization (reduction of the membrane potential) and an increase of permeability of the yeast plasma membrane, in cells under oxidative stress (Sousa et al. 2019c). Similarly, yeasts incubated with ZnO NPs displayed intracellular ROS and an augmented cell membrane permeability (Babele et al. 2018). It was observed that strains with deletion of genes involved in the biosynthesis of ergosterol (ERG2 and ERG28), a sterol that affects membrane fluidity, and in transmembrane transport (PKR1), displayed enhanced susceptibility to ZnO NPs, which suggested that these NPs disrupt cell membrane integrity and impair their proper function (transport) (Márquez et al. 2018).

The impact of MOx NPs on yeast cell membrane can be dose-dependent. Thus, the incubation of yeasts for 24 h with 100 mg L−1 Al2O3, In2O3, Mn3O4, SiO2 and SnO2 did not induce the permeabilization of the cell membrane (Sousa et al. 2019a). However, the exposure of the same yeasts to some of these NPs (Al2O3, Mn2O3 and SiO2), but at higher concentration (1000 mg L−1), during 10 h, caused a significant loss of membrane integrity (Garcia-Saucedo et al. 2011; Otero-Gonzalez et al. 2013).

Alteration of function and morphology of mitochondria and endoplasmic reticulum

Mitochondrial membrane potential (ΔΨm) is an essential component for energy-producing and non-producing mitochondrial functions (Zorova et al. 2018). The depolarization of the mitochondrial membrane, i.e. the dissipation of mitochondrial membrane potential in yeast cells treated with NiO NPs (Sousa et al. 2019c) and the alteration of the architecture of mitochondria in cells incubated with ZnO NPs were described (Babele et al. 2018).

The exposure to ZnO NPs also severely affected the architecture and function of the endoplasmic reticulum, in yeasts, through modulation of unfolded protein response (Babele et al. 2018).

Modification of vacuole architecture and induction of autophagy

Yeast cells treated with ZnO NPs presented a drastic modification of vacuoles morphology (Bayat et al. 2014; Babele et al. 2018) and a redistribution of Atg8-GFP to vacuoles, indicating the induction of autophagy (Babele et al. 2018). It was observed (by TEM) the modification of vacuole shape and its disruption, in yeasts incubated with TiO2 and CuO NPs, respectively. Dark deposits in vacuoles (TiO2 treated cells) or in vesicles (ZnO or CuO treated cells) were also described (Bayat et al. 2014).

Mitochondrial and genomic DNA damage

DNA damage is commonly found during OS, being mitochondrial DNA (mtDNA) a very sensitive target (Richter et al. 1988). Exposure of S. cerevisiae cells to NiO NPs led to the mtDNA damage with the consequent abolition of respiration (incapacity to grow on non-fermentable carbon sources) and the formation of typical respiratory-deficient colonies, commonly known as petite mutants (Sousa et al. 2019c).

Using the canavanine assay, it was shown the damage of nuclear DNA in yeasts incubated with NiO NPs (Sousa et al. 2019c). Nuclear DNA damage, measured using the comet assay, in yeast cells treated with CuO, TiO2 and ZnO NPs was also described (Bayat et al. 2014).

Apoptotic cell death

The exposure of yeast cells to NiO NPs induced regulated cell death, with typical apoptotic hallmarks such as damage of cell membrane, loss of cell viability, phosphatidylserine exposure at the outer cytoplasmic membrane leaflet, and nuclear chromatin condensation, in a process dependent on de novo protein synthesis and apoptotic regulators/executors (Yca1p and Aif1p) (Sousa et al. 2019c). The sequence of events associated with the induction of cell death in S. cerevisiae by NiO NPs was described (Sousa et al. 2019c). Other studies also indicated a cell dead apoptotic pathway in human cell lines exposed to CuO (Siddiqui et al. 2013), NiO (Siddiqui et al. 2012), ZnO (Keerthana and Kumar 2020), or binary mixtures of Al2O3 and ZnO NPs (Koerich et al. 2020).

Concluding remarks

Products containing NMs grown enormously in the last decades. Nanotechnology Consumer Products Inventory, updated in 2013, listed 1814 consumer products containing NMs, from 622 companies, in which products containing metals and metal oxides correspond to the largest group, constituting 37% of products (Vance et al. 2015). Concomitant, and understandably, concerns about environmental, health and safety implications of NMs have also raised.

Over the last decade, and thanks to a substantial research effort, important progress concerning the impact of NPs in terrestrial and aquatic systems as well as about their mechanisms of toxicity has been observed. However, substantial gaps still exist that require further attention, namely regarding to MOx NPs concentrations used in the assays, the time and the type of exposure.

Although it is difficult to accurately detect NPs in aquatic environments, it is estimated that their concentration in surface waters varies from ng L−1 to μg L−1 (Gottschalk et al. 2013). However, it is common to find studies that use NPs concentrations greater than 100 mg L−1, reaching 1000 mg L−1 or even more. Another challenge is to study the impact of sub-lethal concentrations of NPs during a long-term exposure (covering multiple generations of the organism), in a repeated way, to get more information on chronic exposure to MOx NPs, in order to adopt the necessary protective measures regarding the use of products containing MOx NPs.

A more systematic approach is needed in nanotoxicology research. Thus, future studies should combine high-throughput molecular profiling technologies (transcriptomics, proteomics and metabolomics) with more traditional approaches (physiological studies), to give a holistic understanding of cellular responses to MOx NPs (and NMs in a general way) and allow the elucidation of the mechanisms associated with its toxicity.