Introduction

Nanomaterials (NMs) are used in a variety of applications, including cosmetics and personal care products, electronics, drug delivery systems, medical diagnostics, manufacturing technologies, and paints. In particular, in medicine, nanoparticles (NPs) offer a unique opportunity to overcome physiological barriers and deliver bioactive agents to specific cells and targets (Guarnieri et al. 2011; 2013; Kumar et al. 2012). However, the use of NMs is a threat because of their potential toxic effects on living organisms (Kumar et al. 2012). Therefore, an increasing effort of the research community has recently been devoted to manage the sustainability of NPs employment in biology. Experimental evidence shows that NPs may enter the environment and be associated with the increased risk of certain diseases in living organisms (Bacchetta et al. 2012; Casado et al. 2013; Kumar 2006), including humans (Kumar et al. 2012). In particular, it is well known that aquatic environments are at risk of exposure to pollutants, as they represent a sink for most environmental contaminants (Krysanov et al. 2010; Scown et al. 2010; Casado et al. 2013). This new form of pollution, known as nanotoxicity, introduces a serious problem to the scientific community. The relative novelty of the subject, the complexity of the interaction occurring between NMs and living systems, and the lack of history or methods (Kahru and Dubourguier 2010; Kumar et al. 2012), make it difficult to provide standardized measurements of the toxicity level of the new nanotechnology industry products. NPs have the same chemical composition of the bulk, yet have different toxicological properties according to size, shape, and surface covering. Dose, route of administration, and exposure as well are critical factors that affect the degree of toxicity produced by any particular type of NP. It is for this reason that a careful and rigorous toxicity testing is necessary before any NP is declared to be safe for broad use (Sanfins et al. 2014; Kumar et al. 2012). In particular, polystyrene is one of the most heavily used polymers in routine daily activity and is a potential strong pollutant of aquatic environment. Among its many applications, it is commonly used as a food container and carrier. In nanomedicine and nanotechnology, understanding the potential consequences of employing nanoscale materials in various applications, especially in embryology and biomedical field, is critically an important procedure for the protection of the general public.

Liu et al. (2011) showed that uncoated PSNPs are generally inert and nontoxic, when negatively charged. However, Hardy et al. (2013), investigating comparatively the immunological imprints of uncoated polystyrene nanoparticles (PSNPs) 50 nm (PS50G) and 500 nm (PS500G), found that these particles imprint a differentially modulated induction of acute allergic airways inflammation, with PS50G but not PS500G significantly inhibiting adaptive allergen-specific immunity. These effects are not limited to the respiratory system, as increases in acute cardiovascular events were also observed (Enright et al. 2013). During mouse embryogenesis, fluorescent polystyrene particles were employed, as an efficient and safe tracking method to examine the roles of particles size and surface modification in particles translocation, by Tian et al. (2009). Yet, under this condition, growth inhibition was observed in embryos containing those nanoparticles. Mouse embryos derived from mother exposed to mixed-size PSNPs did not show inhibition of embryo development at the blastocyst stage, although they were internalized, suggesting a lack of embryotoxicity (Bosnam et al. 2005). However, the observation that fluorescent PSNPs up to 240 nm were able to cross the placental barrier (Wick et al. 2010) requires further studies for screening embryotoxicity, although they do not affect the intrinsic characteristics of this organ. Although PSNPs deleterious effects were observed in aquatic species (Casado et al. 2013), such as zebrafish (Cedervall et al. 2012), the effects of PSNPs on aquatic embryos have not been investigated yet.

The aim of our work is to understand the effects that the accumulation of uncoated fluorescent 50 nm PSNPs has during the embryogenesis of the well-known and robust model Xenopus laevis, for the assessment of PSNPs effects in aquatic larvae, thereby contributing to the understanding of their possible toxicity and subsequent harm for general community. In Xenopus, reports dealing with the potential effects of these nanoparticles are not available. X. laevis is an aquatic organism development of which occurs outside the mothers body, therefore allowing monitoring anlagen morphogenesis in detail. Although the use of X. laevis has been hampered by its relatively long reproduction time compared with D. melanogaster and C. elegans, its large embryonic cells and the ease of manipulation in early embryogenesis continue to have remarkable research potential. Xenopus facilitates biophysical and physiological approaches to understand developmental signals that can be transferred to higher vertebrates, including humans (Tomlinson et al. 2005; Takagi et al. 2013). The PSNPs size range (40–60 nm) used in our work is known to represent the most efficient size for cellular uptake (Nam et al. 2013) and were chosen as model to investigate the suitability of their possible employments in future studies such as carrier or tracking methods (Chang 2010). We have employed the range of concentration used to test these NPs in other organisms (Casado et al. 2013) and applied the Frog Embryo Teratogenesis Assay-Xenopus (FETAX) protocol. We assayed two procedures of administration: the dilution in culture medium (contact embryos) and microinjection in the early stage of development (one of the cells at the two cells stage). Interestingly, in microinjected embryos, the uninjected side can be considered as the natural control of each test to check if particles penetrate into the sister cell, while, in contact embryos, the penetration in adjacent cells cannot be easily tested. In contact embryos, nanoparticles, differently from the injected embryos, encounter a complex mixture of extracellular proteins in the intercellular space, which cover them and can determine their interaction with the cellular environment (Fleischer and Payne 2014; Xu et al. 2014). The three endpoints of this test, i.e., mortality, malformations, and growth inhibition, render it a powerful and flexible bioassay to evaluate pollutants in water (Dumont et al. 1983).

Materials and methods

Animals

Adult X. laevis were obtained from Nasco (Fort Atkinsons, Wisconsin, USA). They were kept and used at the Department of Biology of the University of Naples, Federico II, according to the guidelines and policies dictated by the University Animal Welfare Office in agreement with international rules and in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health of the Italian Ministry of Health. The protocol was approved by the Committee on the Ethics of Animal Experiments (Centro Servizi Veterinari) of the University of Naples Federico II (Permit Number: 2014/0017970). All procedures were performed according to Italian ministerial authorization (DL 116/92) and European regulations on the protection of animals employed for experimental and other scientific purposes. All surgical procedures were performed under ethyl 3-aminobenzoate methanesulfonate (Sigma). All trials were adopted to minimize suffering. To obtain eggs, X. laevis females were injected in the dorsal lymphatic sac with 500 units of Gonase (AMSA) in amphibian Ringer solution (111 mM NaCl, 1.3 mM CaCl2, 2 mM KCl, 0.8 mM MgSO4, in 25 mM Hepes, pH 7.8). Fertilized eggs and embryos were obtained by standard insemination methods and staged according to Nieuwkoop and Faber (1956).

PSNPs characterization

Red fluorescent, unmodified polystyrene NPs with diameters of about 50 nm were purchased from Duke Scientific Corporation. Composition: Polystyrene. Dyes: Firefly fluorescent red (542/612 nm). Density: 1.05 g/cm3 (1.5 × 1014NPs/ml). Index of Refraction: 1.59 at 589 nm (25 °C). Dynamic light scattering (DLS), made with a Zetasizer Nano-ZS (Malvern Instruments, Worcestershire, UK), was performed to measure PSNPs size and z-potential. Measurements were conducted at 25 and 18 °C, with or without sonication. We performed the DLS on amphibian solutions, 10 % Ringer pH 6.5 or FETAX pH 7.4 (106 mM NaCl, 11 mM NaHCO3, 4 mM KCl, 1 mM CaCl2, 4 mM CaSO4, 3 mM MgSO4) or in H2O, containing the PSNPs at three different concentrations: 4.5, 9, and 18 mg/L. We made these tests both in the absence and in the presence of embryos (18° C, the breeding temperature). Each measurement was repeated five times, and all the measurements were done in triplicate. The measurements were carried out in agreement with the embryonic stages evaluated for mortality data. i.e., stages: 10, 14, 28, 46, corresponding to days: 1, 2, 4, and 7 post-fertilization, respectively.

PSNPs exposition and microinjections

Stages 2 embryos were placed and grown in 10 % Ringer or FETAX solution containing PSNPs at the following concentrations: 4.5, 9, or 18 mg/L. FETAX can be used to test single compounds (Bantle and Dawson 1988). All embryos were harvested until stage 45/46. At stage 40/41, X. laevis embryos opened their mouth (Nieuwkoop and Faber 1956), and ingestion became the main route of PSNPs intake. Since in the following days the grazing behavior of the larvae became very active, the number of ingested particles increased. The embryos survival and phenotype were checked daily. All experiments were carried out at about 18 °C in triplicate. As control, sibling embryos were used. The mortality percentage was calculated on the number of dead embryos versus their total number at the beginning of the test. Dose-dependent correlation between PSNPs concentration and embryos mortality rates was also analyzed. The relationship between the control and the treated groups, along with the percentage of dead embryos and the observed malformations, was investigated with Chi square test, using the Yate’s correction for continuity or Fisher’s Exact test. To evaluate differences in growth retardation among groups, the ANOVA with nonparametric Kruskal-Wallis test and Dunns test for post hoc analysis was used to compare all pairs of columns. The survival distributions in control and experimental groups were also assessed in terms of significance using nonparametric Mantel-Cox test. For all statistical tests, we employed the Graph Pad Prism 5 software (San Diego, California USA). The methods utilized here are comparable to those of other authors who have investigated the NPs on aquatic organism (Kahru and Dubourguier 2010; Krysanov et al. 2010; Bacchetta et al. 2012; Pompa et al. 2010). Alternatively, the PSNPs were microinjected into embryos using a Drummond ‘Nanoject II’. The nanoparticle solutions were injected into the animal hemisphere of a single cell of two-cell stage embryos. After injection, embryos were kept in 3 % Ficoll (Fisher Bioreagent)/10 % Ringer at about 18 °C. The maximum volume we could inject was 4.6 nL. The concentrations of the solutions were the same as those utilized for contact experiments. The solutions of microinjections contain 1.38 × 106 (0.08 ng) or 0.69× 106 (0.04 ng) or 0.345× 106 (0.02 ng) PSNPs diluted in phenol red (vital dye) and 300 pg of pCS2MTGFP mRNA coding for GFP (Green Fluorescent Protein). GFP was always co-injected to label the injected side. Capped synthetic RNAs were generated by in vitro transcription using the Sp6 Message Machine kit (Ambion, Austin, TX, USA). As control, GFP-injected and -uninjected sibling embryos were utilized. The phenotypes of the injected embryos were scored when embryos reached stage 11, 11.5, 13, 19, 25, 28, 35, or 45/46, corresponding to the known stages of anlagen morphogenesis. All the samples were photographed using a Leica MZ16F UV stereomicroscope, equipped with a Leica DFC 300Fx camera and IM50 Image Manager Software.

Whole mount in situ hybridization

Wild-type X. laevis embryos, or embryos injected in one blastomere at stage 2 with PSNPs diluted in phenol red, were fixed in MEMFA (MOPS (4-morpholinepropanesulfonic acid) EGTA MgSO4 FormAldehyde) (100 mM MOPS pH 7.4, 2 mM EGTA, 1 mM MgSO4 3.7 % v/v Formaldehyde) (see Gont et al. 1993) at stage 11, 11.5, 13, 19, 25, and 45/46 and stored in 100 % ethanol at -20 °C. In situ hybridizations (see Vaccaro et al. 2006) were performed with antisense digoxigenin-labeled RNA of bra, or myod1, or sox9 synthesized with SP6 polymerase (Roche, Mannheim, Germany). In situ hybridizations with antisense sox9 RNA were also performed on embryos exposed to PSNPs diluted 18 mg/L and fixed in MEMFA at stage 46. All hybridized embryos were photographed using a Leica MZ16F equipped with a Leica DFC 300Fx camera and IM50 Image Manager Software. myod1, and sox9 probes are a kind gift from Dr M. Ori, University of Pisa, Italy.

Cloning of fragment bra for in situ hybridization

For total RNA extraction, embryos at stage 19 were homogenized using Tri-Reagent (Sigma) according to the supplier instructions. cDNA synthesis was performed using SuperScriptVILO cDNA Syntesis Kit (Invitrogen). The partial coding DNA of bra (880 bp) was obtained by PCR, using the primers forward: 5′-tcacccagactcacccaactt-3′ and reverse: 5′-gtgccgtgacatcatactgg-3′, on the basis of the X. laevis sequence of the brachyury (T) gene (Xbra), (GenBank: M77243.1). The PCR products were cloned into the pCR2 vector (TA Cloning Dual Promoter, Life Technologies, Carlsbad, CA, USA). The product was purified using the miniprep High Pure Plasmid Isolation Kit (Roche, Mannheim; Germany). pCR2-bra was digested with Not1 (Roche) and purified using PureLink PCR Purification Kit. Riboprobes were transcribed with Sp6 polymerase using RNA labeling (SP6/T7) kit (Roche) and purified with mini Quick Spin Columns (Roche).

Histology

Embryos were fixed at stage 46 in 4 % formaldehyde at 4 °C and stored in 100 % methanol at −20 °C. Frozen sections of 7 µm thickness were obtained after embedding and freezing in Killik (Bio Optica, Milan, Italy). Nuclei were counter-stained with DAPI (4′,6-Diamidino-2-Phenylindole,Dihydrochloride, SIGMA) (1:1,000 in 100 % ethanol). Sections were observed and photographed using a Leica CTR 6500 UV microscope equipped with a Leica Application Suite.

Confocal microscopy

Embryos were fixed at stage 46 in MEMFA and then 60-μm-thick sections were obtained after embedding and freezing in Killik (Bio Optica Milan, Italy). Intestinal sections prepared for observation under the confocal microscope were incubated with WGA-FITC conjugates (Wheat Germ Agglutinin-Fluorescein IsoThioCyanate, Molecular probes), a kind gift from Dr D. Guarnieri (CRIB, Naples, Italy), diluted 1:200 in PBS containing 0.5 % BSA and 0.1 % Triton X100 for 1 h at RT. The WGA excess was eliminated by several washes in PBS.

Sections were observed and photographed using a Leica SP5 confocal laser scanning microscope. An Ar laser was used to produce the excitation laser line at 488 nm, and a HeNe laser was used to produce the excitation laser line at 543 nm. Fluorescence emission wavelength-bands were 500–530 nm for green fluorescence (488 nm excitation) and 560–650 nm for red fluorescence (543 nm excitation). Image acquisition was standardized maintaining laser power, photomultiplier, pinhole aperture, and confocal scanner settings constant for all experiments. Sections were spaced at 0.7 µm intervals, and 25 × 0.95 a HCX IRAPO L 25 × 0.95 W 0.17 water immersion objective was utilized. For each sample, a total number of 80 optical sections (1.4 µm thick) were analyzed. Images were processed by LAS AF confocal software.

Results

PSNPs characterization

To elucidate the colloidal stability of nanoparticles, we carried out measurements of the z-potential. It was reported that the colloidal stability of NPs in suspension is strongly pH-dependent, mainly due to electrostatic repulsion. The PSNPs z-potentials measured on Ringer, FETAX and in water, within 1 h from dilution at the concentrations of 9 and 18 mg/L are about 30 mV, indicating that PSNPs do not form aggregates in the solutions we utilized (Table S1). This is in agreement with data in the literature reporting that, at z-potential values of ±30 mV, electrostatic interactions between particles are strong enough for electrostatic stability, while at intermediate values, near their isoelectric point, particles can flocculate (Guarnieri et al. 2013). The measures at 4.5 mg/L are not reported herein because the particles were too diluted to give a good quality analysis. In addition, our DLS measurements indicated that there was no need to sonicate NPs suspensions before use. Sonication is expected to damage the particles to some extent producing fragments that could increase the polydispersity value of the NP sample, making the obtained results less accurate. The pH values of the PSNP solutions, in the absence of embryos, do not vary in time. In the presence of embryos, starting from the second day, the pH of the Ringer solution changes from 6.5 to 7, while the pH of the FETAX remains unchanged. The z-average of PSNPs, in FETAX solution, in the absence of embryos and at the concentrations of both 9 mg/L and at 18 mg/L reaches about 80 nm from the second day. In the presence of the embryos, the PSNPs showed strong aggregation up to about 300 nm (Tables 1, S2; Figs. 1, S1).

Table 1 Dynamic light scattering PSNPs in FETAX solution
Full size table
Fig. 1
figure 1figure 1

Graphs of the PSNP solutions measurements performed at the DLS. Each curve represents the average of five measurements, each performed in triplicate without embryos (a, b, left) or with embryos (a, b, right). In both cases, it is possible to observe a time dependent appearance of the peaks related to the increasing size of the particles. In the solution with embryos the PSNPs increase begins earlier and is more evident than in the solution without embryos

Full size image

Concentration-dependent mortality

The embryos harvested in the presence of 18 mg/L PSNPs (n = 225) showed a percentage of mortality (about 20 %) similar to that found for control sibling embryos. In contrast, the embryos harvested in the presence of 9 mg/L PSNPs (n = 225) or 4.5 mg/L (n = 225) showed, respectively, 40 % and over 50 % of mortality rates (Table 2). These mortality values are statistically very relevant compared to those of the control sibling embryos (n = 325) (Table 2; Fig. 2a). Therefore, our data provide evidence that more pronounced PSNP dilutions produce mortality increase in these embryos (Fig. 2a). This correlation has been verified by Chi square test for trend (P < 0.0001; Table 2) and by the inverse linear relationship between the percentages of dead embryos and concentration (P = 0.0345, r 2 = 0.9971; Fig. 3a). To further analyze the effect of PSNPs on the embryos, we performed injections of PSNPs and/or GFP in one blastomere of X. laevis two-cell embryos for a total of 550 embryos. While the embryos injected with GFP alone (n = 100) showed mortality similar to that of sibling embryos (Table 2), the injection of PSNPs produced mortality values of 66 % in embryos injected with 4.5 mg/L, 70 % in embryos injected with 9 mg/L, and of 80 % in embryos injected with 18 mg/L (Table 2; Fig. 2b). According to the Chi square test for trend, the observed mortality increase is dose dependent (P = 0.0035; Table 2). The graph in Fig. 3b shows direct correlation between mortality and concentration of nanoparticles injected (P = 0.0334, r 2 = 0.9973).

Table 2 Embryotoxic effects in X. laevis
Full size table
Fig. 2
figure 2

Mortality evaluation in PSNP-treated embryos. The mortality distributions were evaluated by non-parametric Mantel-Cox test (a contact embryos) giving P > 0.05 (PSNPs treatment with 18 mg/L compared with the control; P = 0.85) and P < 0.0001 (PSNPs treatment with 9 or 4.5 mg/L compared with control). b Injected embryos: PSNPs treatment with 18 or 9 or 4.5 mg/L in respect to control show P < 0.0001. The experimental points represent the average from nine independent experiments for contact embryos and five experiments for injected embryos. The error bars indicate the standard error

Full size image
Fig. 3
figure 3

Correlation between mortality and PSNPs concentrations upon contact and injection. a Correlation between mortality percentage and PSNPs concentrations in contact embryos (P = 0.0345, r 2 = 0.9971) and b injected embryos (P = 0.0334, r 2 = 0.9973)

Full size image

PSNPs affect embryo phenotype

PSNPs microinjected or contacted embryos produced the same type of abnormalities in developing embryos (Figs. 4, 5). These malformations concern derivatives of all three germ layers: ectoderm, endoderm, and mesoderm. Deformations of the tail were observed with regard to length, shape, and number (4–50 %) (Figs. 4 a–c, 5b, d, e, h; Table 3) (Gont et al. 1993). Embryos showed various abnormalities of the head (13–37 %) (Fig. 5a; Table 3) as small eyes (Fig. 4g), the absence of or variously deformed eyes (Fig. 5b, c, e, f, h), and cone-shaped eyes (Fig. 5b, e, f).

Fig. 4
figure 4

PSNPs injection effects on embryos development. a–g Embryos injected with PSNPs and GFP mRNA into one blastomere at two-cell stage and harvested until stage 45/46. The injected side is marked by GFP. a–c The embryos show a reduced length and a forked tail (dashed lines). In c the embryo is caught at about stage 25 when it stopped growing. d Embryo showing abnormal distribution of pigment (dorsal view, arrowheads). e–f Embryos (ventral view) showing an immature gut (circle) compared with sibling embryos at the same stage of growth (see h). g Embryo with small eye (arrow), edema (spot) and primitive gut (circle). h Control sibling embryo injected only with GFP mRNA

Full size image
Fig. 5
figure 5

PSNPs contact effects on embryos development. a Set of stage 45/46 embryos showing anomalies of different types. b–d, h Embryos with edema in ventral anterior zone (spot), b, c, e, h embryos with different anomalies of eyes (arrow), b, d, e, h embryos with anomalies of the tail (asterisk), c, d, h embryos with abnormal intestine (circle), e embryos with anomalous pigment distribution (arrowheads). DAPI staining of cryostat sections of exposed embryos, showing irregular stratification of retina in f (arrow, white lines define retina layers: ONL outer nuclear layer, INL inner layer, and GCL ganglion cell layer) and in g malformation of neural tube (empty arrowhead), compared with wild types. h Right and left views of a st. 41 embryo are shown. The embryo misses one eye (arrow) and displays anomalies of the tail (dashed lines) and of the intestine (circle) as well as an edema (spot). i–j Wild-type st.45 embryo in dorsal i and ventral views j

Full size image
Table 3 Pattern malformations in X. laevis
Full size table

Sections carried out through the head showed abnormalities to the retina layers and malformations of the corresponding encephalon in contact-embryos (Fig. 5f–g). Abnormalities were observed also in epidermal melanocytes, which were randomly distributed and often clustered (13–28 %) (Figs. 4d, 5e; Table 3) and in the winding of the intestine that appeared immature with respect to control sibling tadpoles (13–36 %) (Figs. 4 e–g, 5c, d, h; Table 3) (Chalmers and Slack 1998). Moreover, the contacted and microinjected embryos showed variously extended edema in the ventral anterior zone (6–24 %) (Figs. 4g, 5b, c, d, h; Table 3). Head malformations and edema were not observed in 4.5 mg/L contact embryos at stages 41–45 when the counts were performed (Table 3). The total absence in these embryos of head malformation or edema correlates with the high percentage of mortality presented early in development which starts from the neurula stage. At that stage, these embryos showed malformations of the prospective head so severe and an edema so extensive, which were incompatible with embryo survival (data not shown). Instead, by comparing mortality results in Tables 2, 3, it is evident that the survived embryos are the ones that showed comparable malformations. In both treatments, embryos often displayed more than one of the mentioned malformations: we define ‘severe’ malformed embryo when the phenotypes display two malformations, and ‘monster’ when the phenotypes carry at least three malformations, thus rendering it difficult to recognize their original pattern of development. In particular, contact embryos treated with 18 mg/L grew more slowly than sibling embryos and show 27 % monster phenotypes (Table 3). Moreover, the 9 mg/L and the 18 mg/L contacted embryos were significantly shorter than the control embryos (Fig. 6). Embryos microinjected in one blastomere of two-cell stage showed the amended phenotype in both the injected and the noninjected side (Table 3). In Xenopus, the first two cells give rise to the left and right sides of the embryo. While microinjection of GFP had no effect on the embryos development as shown in Fig. 4h, the PSNPs microinjected embryos showed modification on both the left and right sides, thus indicating that the injected PSNPs pass from one cell to another affecting the development of the whole embryo. The red fluorophore conjugated to nanoparticles allowed us to use the confocal microscopy to observe their localization in the various embryonic districts. PSNPs exposed embryos fixed at stage 45–46 show aggregates of PSNPs in the digestive tract (Fig. 7a–a3), in the eyes (Fig. 7b–b3) and in the pharynx (Fig. 7c–c4), but not in the brain (Fig. 7c–c2). In the intestine, the aggregates were present both in its lumen and in its wall (Fig. 7a–a2), as shown in particular in Fig. 7a3. Pharynx presents PSNPs aggregates of various sizes (Fig. 7c–c4). In the eyes, PSNPs are found around the lens and between the pigmented layer and the outer layer of retina (ONL) (Fig. 7b–b3), as well as in the optic nerve evenly distributed throughout the structure (Fig. 7b–b3). The images in Fig. 7 show that the NPs assayed can penetrate in all embryonic tissues if in direct contact. Magnification of digestive tract cells showed the presence of the PSNPs in the cytoplasm, in the nucleus, and in the cellular periphery (Fig. 7d). The aggregates are more abundant in the gut tissue with respect to the pharynx and the eyes (Fig. 7e).

Fig. 6
figure 6

Growth retardation analysis. The growth retardations were evaluated with ANOVA statistical test giving P < 0.0001. 18 mg/L versus control: P < 0.001; 9 mg/L versus control: 0.001 < P < 0.01 and 4.5 mg/L versus control: P > 0.05. The error bars indicate the standard error

Full size image
Fig. 7
figure 7

PSNPs confocal localization. The distributions of PSNPs in the intestine a–a 3 , in the eye b–b 3 , and in the brain/pharynx c–c 4 are here shown by confocal microscopy. Figures a–c shows fluorescent optical sections; figures a 1 –c 1 shows the same sections in brightfield and figures; a 2 –c 2 shows the merge between them. In a 3 is shown an enlargement of the area indicated by the empty arrowhead in a–a 2 ; the nanoparticles are in the tissue (a 3 , empty arrowhead). In the eyes b–b 2 ; nanoparticles are distributed in the ONL of the retina (short arrow), in lens (thin arrow), and in the optic nerve (arrow); white lines define retina ONL, INL, and GCL. In b 3 is shown the control optical section. In c–c 2 , the brain does not show aggregates of PSNPs (see the area over the white line), while in the pharyngeal area c–c 4 (see the area under the line that was used to highlight the zone), there are interspersed aggregates of nanoparticles (arrowhead). A contrast-enhanced image of merge c 3 together with a magnification c 4 of pharynx area (arrowhead) confirms the presence of fluorescent aggregates. In d magnification of the gut cells, stained with WGA-FITC conjugated, shows the PSNPs. The red-yellow PSNPs are seen in the cytoplasm, in the nucleus, and at the cell periphery (circle). In e the histogram shows the aggregates size distribution in intestine (red), eye (blue), and pharynx (green). a–a 3 , b–b 3 , and c–c 4 scale bar = 50 µm, d scale bar = 5 µm. (Color figure online)

Full size image

Expressions of bra, myod1, and sox9 mRNAs are amended in embryos treated with PSNPs

The malformations observed in embryos treated with PSNPs led us to investigate whether also an altered expression of genes is involved in the early embryonic development. Nanoparticles can induce changes in gene expression (Fan et al. 2012). For this purpose, we used bra (early mesoderm marker), myod1 (paraxial mesoderm marker), and sox9 (neural crests migration marker). We performed the hybridizations of bra and myod1 mRNAs, on PSNPs injected (Fig. S2 a’–c’ for riboprobes bra and d’–f’ for myod1) and control, uninjected sibling embryos (Fig. S2 a–c for bra and d–f for myod1). The latter expressed bra in the marginal zone shortly before the beginning of gastrulation in a crown-shaped pattern (Fig. S2 a). During gastrulation, transcription of bra was maintained in the prospective notochord along the A/P axis (Fig. S2 b) and, at the end of gastrulation, bra could be detected mostly in the notochord and in the posterior mesoderm (Fig. S2 c) (Smith et al. 1991; Kwan and Kirschner 2003). In PSNPs injected embryos (Fig. S2 a’–c’) at the beginning of gastrulation, a partial expression of bra was observed that did not cover the entire marginal zone (Fig. S 2a’) suggesting changes in the migration of the prospective mesoderm that did not fully migrate to form a complete notochord (Fig. S2 b’). At the neurula stage, the notochord appeared displaced sideways (Fig. S2 c’). In similarly microinjected embryos, the myod1 expression was displaced as well (Fig. S2 d’–f’). During gastrulation, myod1 localization did not have the typical horseshoe shape (compare Fig. S2 d with Fig. S2d’). The expression of myod1 was altered also later at the early stage of neurula, where the expression was rather asymmetric in contrast to wild-type embryos (Fig. S2 e–e’) and at stage 25, where the riboprobe localization extended up to the head (Fig. S2 f–f’) (Hopwood et al. 1989, 1992; Scales et al. 1990).

Abnormalities in the migration of the neural crest, which is currently regarded as a germ layer, were assessed employing sox9 riboprobe. Untreated embryos (Fig. S3 a–b) expressed this messenger at the levels of the branchial arches (Fig. S3 a–b,), facial cartilage (Fig. S3 b), olfactory bulbs (Fig. S3 b), and epidermis (Fig. S3 a–b) (El Jamil et al. 2008; Spokony et al. 2002). Stage 46 contact-embryos exhibited altered expression of sox9 mRNA. In these embryos, sox9 showed anomalous and variable expressions in the epidermis and in the malformed gut (Fig. S3 d–e). In the eyes, a strong hybridization is detected in the retina (Figs. S3 e). Little sox9 expression was found in the underdeveloped branchial arches region (Fig. S3 c–d) and in the misplaced olfactory bulbs (Fig. S3 c).

Discussion

The impact of nanostructured products on the ecosystem is not completely clear, albeit some preliminary reports led to reasons of full concern (Kumar et al. 2012; Scown et al. 2010). In particular, the potential embryotoxic effects of NPs on both aquatic organisms (Bacchetta et al. 2012; Casado et al. 2013; Krysanov et al. 2010) and laboratory mammals (Bosman et al. 2005; Tian et al. 2009) deserve full attention. NPs may in fact constitute a serious danger not only for these organisms, but also for their consequent domino effects, in particular along the aquatic food chain (Cedervall et al. 2012). To improve the understanding of this topic, in the present study, we investigate the potential effects of the PSNPs accumulation in X. laevis embryos for the assessment of their effects on aquatic larvae in view of their possible employments in future studies (Chang 2010). In particular, we have tested uncoated fluorescent of about 50 nm PSNPs because it is well known that maximal cellular uptake of NPs can take place in the intermediate size range of 40–60 nm and that fluorescence offers an efficient tracking method (Nam et al. 2013). We have employed a range of concentrations that have been used to test these NPs in other organisms (Casado et al. 2013) applying the FETAX protocol (Bacchetta et al. 2012; Bantle and Dawson 1988).

DLS data show that at the beginning of the experiment, the PSNPs utilized have a diameter of about 54 nm in amphibian Ringer or FETAX. Analyses of the mortality in contacted-embryos suggest that they die depending on the nanoparticles concentration; the mortality is inversely dependent on NPs concentration. Apparently, these data are in contrast with the literature (Casado et al. 2013). Mortality increases observed in contact-embryos suggest that the NPs penetrate more efficiently at a lower concentration. Higher concentrations of PSNPs produce aggregates before entering the embryos more readily than at lower concentrations. Our data show that the PSNPs at the concentrations of both 9 and 18 mg/L, starting from the second day, aggregate in the presence of the embryos. The PSNPs at the concentration of 4.5 mg/L cannot be measured using the DLS. The fact that we have observed large aggregates of PSNPs by confocal microscopy at the tissue borders argues in favor of this hypothesis. It should also be mentioned that, in contact-embryos, the aggregates penetration may be slowed down by envelopes covering the embryos. In fact, when PSNPs were added at two-cell stage, they have to pass through the vitelline envelope, an extracellular barrier that accompanies the embryos until stage 32, and in later stages, they have to pass the tegument. In the injected embryos, the mortality rate increases according to the increment of PSNPs amount. This increase is statistically significant. Based on our data, microinjections appear a particularly appealing method for investigating PSNPs effects, as the amount of the administered PSNPs is more likely to reach the embryo cytoplasm in unmodified conditions, and we are sure that nanoparticles are inside the embryos. Comparing the effects obtained with both techniques, we could understand that the effects we observed in contact embryos are due to PSNPs, as they are the same in both kind of treatments. Interestingly, although embryos were microinjected only in one side, the effects on the phenotype are present in both the injected and the noninjected sides, indicating that the PSNPs can pass from one cell to another. Lai et al. (2007) showed that 24-nm polymer nanoparticles can exploit a non-clathrin, non-caveolae, and cholesterol-independent pathway for nondegradative trafficking. Moreover, polystyrene microspheres ranging from 50 nm to 3 μm fed to female rats are absorbed into the gastrointestinal tract, but particles larger than 100 nm do not reach the bone marrow, and those larger than 300 nm are absent from blood and do not penetrate the tissues (Jani et al. 1990). However, it should be mentioned that the observations made on contact embryos where the PSNPs might have been modified through transporting process from extracellular to intracellular environment may represent more valid information for environmental studies than observations made through directly injected PSNPs (Xu et al. 2014). We showed that PSNPs treated groups, contacted or injected, display the same modifications statistically relevant and higher compared with controls in pigment distribution, malformations of the head, tail, and edema in the anterior ventral region and of the gut winding. Moreover, these malformations can be variously associated. They create embryos with phenotypes dramatically different from the wild type (monsters) or induce their mortality. In addition, the embryos grew more slowly than the sibling embryos. PSNPs aggregates were observed particularly in the intestine, in accordance with the fact that at stage 40/41 X. laevis embryos opened their mouth and ingestion becomes the main route of PSNPs intake. In particular, we have showed that PSNPs entered the gut mucosa, the eyes, and the pharynx and were distributed in the cytoplasm and in the nucleus indicating that they can penetrate easily into the cells. Moreover, the diencephalon anomalies we observed might have been caused by small amount of particles, in a quantity not detectable by our methods, or indirectly by the surrounding tissues. Interestingly, Symens et al. (2011), using the Xenopus nuclear envelope re-assembly assay, found that the nuclear enclosure of nanoparticles they observed were dependent on size and charge of the polystyrene beads. The delayed distribution or non placement of specific mRNAs we showed in PSNPs injected embryos during early development may cause the abnormal development observed in treated embryos, including the tail formation that develops as a direct continuation of events initiated during gastrulation (Gont et al. 1993). Brachyury expression characterizes the migratory mesodermic cells of vertebrates and is required for mesoderm formation in Xenopus, an indispensable component of all the organs (Yanagisawa et al. 1981; Smith et al. 1991). myod expression is involved in the activation of muscle genes in the somites of embryos (Taylor et al. 1991), sox9 incorrect localization could produce a defect in the migratory activity of the neural crest multipotent cells first located at the lateral edges of the neural plate. As they reach their targets, neural crests differentiate in various cell types, including epidermal melanocytes (Le Douarin and Creuzet 2011) and several malformations we observed in the treated embryos, such as defects in pigmentation are in agreement with neural crests incorrect migration. These data suggest that, at the experimental concentrations we used, the 50 nm PSNPs have a toxic and a possible additional teratogenic potential, in contrast to the fact that the polystyrene is an inert material. Thus, PSNPs cannot be considered as a simple neutral vehicle when in living cells. This conclusion should be taken in account before using PSNPs for medical treatments, as suggested by recent literature on airway cells (McCarthy et al. 2011), intestinal epithelium (Mahler et al. 2012) and blood coagulation (Oslakovic et al. 2012) all depending in different ways on PSNPs size and surface chemistry (Cedervall et al. 2007; Lundqvist et al. 2008). In particular, modified PSNPs induce a growth inhibition of the mouse embryos (Tian et al. 2009) similar to the effect we remarked in Xenopus. Several NPs (i.e., silver, gold and silica NPs) are able to overpass the embryonic barriers and to penetrate zebrafish embryos even at early stages, thus contributing to generate adverse developmental effects (Lee et al. 2007; Browning et al. 2009; Fent et al. 2010).

In a biological fluid, nanoparticles are not naked but covered with a protein “corona” which mediates the biological effects of NPs and changes over time as it moves from one fluid compartment to another (Dell’Orco et al. 2010; Xu et al. 2014). A remarkable specificity in the composition of the “corona” depending on NPs size and surface chemistry was shown (Lundqvist et al. 2008, 2011). Proteins in the “corona” may perturb aggregate structures (Cabaleiro-Lago et al. 2010), produce loss or gain of function (Oslakovic et al. 2012) or an increased inflammatory response (Chang 2010). These effects are comparable to those that PSNPs produce in our experiments, suggesting that the “corona” effect may occur in Xenopus blastomeres, as shown in other studies (Cedervall et al. 2012; Xu et al. 2014).

In addition, the effects we observed may depend on the “mechanical stress” induced by the large quantity of PSNPs that accumulate in the embryonic anlagen (Bacchetta et al. 2012). Significant malformations in the intestine of X. laevis treated with TiO2 NPs, that do not release toxic metal ions similarly to PSNPs (Yamamoto et al. 2004), may be caused by “mechanical stress” due to the large quantities of TiO2 released in the digestive tract (Bacchetta et al. 2012). PSNPs microinjections at the beginning of the development appear to displace mRNAs which is important in the early stages of the development. The effects produced may depend on the change of gene expression induced by PSNPs (Fan et al. 2012) or “mechanical stress” resulting from the large quantity of PSNPs or to threshold concentrations able to affect tissues. If the latter hypothesis were correct, this would imply that PSNPs might interfere with mechanisms of cellular movements, thus contributing to cause the observed potential teratogenic effect. Further studies are necessary to minimize the effects of these NPs.

Conclusions

In conclusion, our data suggest that PSNPs aggregation in solution can depend not only on pH but also on the presence of embryos. In embryos, in spite of the different methods of PSNPs administration, we detected, in both cases, anomalous distribution of pigmentation; malformations of the head, gut, and tail; edema in the anterior ventral zone; and a shorter body length compared with the wild type; these anomalies were variously distributed in treated embryos. Moreover, the treated embryos showed an increase in mortality and grew more slowly than control sibling embryos. Confocal microscopy of contact-embryos showed PSNPs in the intestinal cells, while in other organs, PSNPs were not observed. We found an anomalous distribution of the genes involved in the early embryonic development and neural crest migration. The results of our research show that the PSNPs we used produce malformations despite the presumed inert chemical properties of the polystyrene bulk and indicate that the lethal effect of PSNPs depends both on exposure concentration level and on the ability they have to penetrate the tissue. Our data also demonstrate that the tests we used in Xenopus may be powerful and flexible bioassays for evaluation of pollutants in aquatic embryos.