TiO2 nanoparticles induce omphalocele in chicken embryo by disrupting Wnt signaling pathway

Titanium dioxide nanoparticles (TiO2 NPs) are among abundantly used metal oxide NPs but their interactions with biomolecules and subsequent embryonic toxicity in higher vertebrates is not extensively reported. Physicochemical interactions of TiO2 NPs with egg albumen reveals that lower doses of TiO2 NPs (10 and 25 µg/ml) accounted for higher friccohesity and activation energy but an increment in molecular radii was recorded at higher doses (50 and 100 µg/ml). FTIR analysis revealed conformational changes in secondary structure of egg albumen as a result of electrostratic interactions between egg albumen and TiO2 NPs. The morphometric data of chicken embryo recorded a reduction at all the doses of TiO2 NPs, but toxicity and developmental deformity (omphalocele and flexed limbs) were recorded at lower doses only. Inductively coupled plasma optical emission spectrometry (ICP-OES) confirmed presence of Ti in chicken embryos. mRNA levels of genes involved in canonical and non-canonical Wnt signaling were lowered following TiO2 NPs treatment resulting in free radical mediated disruption of lateral plate mesoderm and somite myogenesis. Conformational changes in egg albumen and subsequent developmental deformity in chicken embryo following TiO2 NPs treatment warrants detailed studies of NP toxicity at lower doses prior to their biomedical applications.

Nanotechnology is a rapidly expanding field, with a wide range of applications in communications, robotics, medicine, clothes, sporting goods, etc 1,2 . According to a recent survey, the number of nanotechnology-based consumer products available in the world market is more than 1800 3 . The increased use of nanomaterials is also under scrutiny due to their adverse effects on the environment, physiology and overall survival of organisms. Titanium dioxide nanoparticles (TiO 2 NPs) are the most abundantly used nano metal oxides with their documented industrial uses in pigments and additives for paints, paper, ceramics, plastics, foods, and other products. The estimated worldwide production of TiO 2 NPs is 10000 tons/year for 2011-2014 and 2.5 million metric tons/ year by 2025 4 . Therefore, risk assessment studies have predicted that TiO 2 NPs will be the most prevalent nanomaterials in environment 5 .
Cytotoxic potential of TiO 2 NPs is well documented in a variety of cell lines. Oxidative DNA damage and apoptosis in HepG2 cells and in human epidermal cells 6 , apoptosis and/or necrosis in human astrocytoma (astrocytes-like) U87 cells 7 and mitochondrial dysfunction in BRL 3A cells 8 are some of the recent reports on cytotoxicity of TiO 2 NPs. Toxicity of TiO 2 NPs based on difference in their size has been documented in nematodes 9 and earthworm 10 . The ability of TiO 2 NPs to produce reactive oxygen species and surface charge are the reasons accredited for their toxicity 6,11 . Several engineered nanometals including TiO 2 NPs have been known to persist in the food chain and move across trophic levels resulting in various forms of toxic manifestations 11 . Hence, their effect on reproductive performance and embryonic development cannot be ignored. Accelerated hatching of larvae and deformed embryos in zebrafish 12 and histopathological changes in juvenile carp 13 are few evidences on TiO 2 NPs induced toxicity on embryonic and post-hatch development. Hatching inhibition and malformation of embryos of Abalone have been reported following TiO 2 NPs exposure 14 . Also, prenatal exposure of TiO 2 NPs in female rats impacts genes controlling brain development in offspring 15 providing compelling evidences on systemic and developmental toxicity.

Results
In DLS analysis, TiO 2 NPs presented a single distribution with peak centered at 88.6 nm. The plot showed that the nanoparticles have a narrow size distribution with an average diameter of about 88.6 nm (Supplementary Figure S1). Physicochemical analysis. Results shown herein are quantification of interaction of peptide bonds with TiO 2 NPs and alteration in the Lennard Jone potential that varies spontaneity and strength of interactive force. There was a decrement in density (1.031064 kg.m −3 ) of TiO 2 NPs + albumen at 10 µg/ml, whereas, 25, 50 and 100 µg/ml recorded steadily ascending values (1.031567, 1.031979 and 1.032092 kg.m −3 respectively) ( Fig. 1a Table S2). An increase in molecular radii (5-11.53 nm) was observed at 1-10 µg/ml TiO 2 NPs. Further, a steady increase in molecular radii (15.69, 19.66 and 24.45 nm) was observed at 25, 50 and 100 µg/ml doses respectively ( Fig. 1f and Supplementary Table S2).
Natality and Morphometry of Chick embryos. Lower doses (10 and 25 µg/ml) of TiO 2 NPs treatment accounted for 12.5% and 25% viable embryos respectively on 19 th day of incubation. Also, 56.25% and 43.75% embryos were found to be malformed at 10 and 25 µg/ml doses. However, higher doses (50 and 100 µg/ml) recorded viable embryos ranging between 75-87.5% (Fig. 3a). Morphometry of the embryos (whole weight and length) recorded significant decrement at all the doses (10-100 µg/ml) (Fig. 3b). Whole weights of liver, brain and heart showed non-significant decrement at all the said doses (Supplementary Table S4).

ICP-OES analysis of embryos.
After 4-days of TiO 2 NPs treatment, the contents of Ti in the chick embryo were measured by ICP-OES. Significantly high levels of Ti was detected (3 times increase) in embryos of eggs treated with 10 µg/ml TiO 2 . But, higher dose (100 µg/ml) accounted for a moderate non-significant content of Ti in embryos (Fig. 3c).
Deformity. Control and TiO 2 NPs treated chick embryos were examined as per Hamburger-Hamilton standard that revealed presence of flexed limbs at 10 and 25 µg/ml doses of TiO 2 NPs (Fig. 3d). Also, omphalocele (ventral body wall defect) was observed at 10 µg/ml dose. These deformities were not seen at any of the higher doses (50 and 100 µg/ml). Further confirmation of flexed limbs of 10 and 25 µg/ml TiO 2 NPs treated embryos was obtained by alcian blue-alizarine red staining (Fig. 3e).
Expression of Wnt signaling genes. RT-PCR analysis was performed to assess the effect of TiO 2 NPs on expression of key genes of canonical (CTNNB1, PITX2 and LEF1), non-canonical Wnt/Ca 2+ (WNT11, PRKCA and CAMK2D) and Planar Cell Polarity (ROCK1 and ROCK2) pathways associated with Wnt signaling. Expression levels of genes of canonical pathway (CTNNB1, PITX2 and LEF1) were downregulated significantly in embryos treated with TiO 2 NPs (10 µg/ml). A Similar trend of significant decrement was also observed in cadmium treated embryos, whereas, TiO 2 bulk treatment could not manifest any significant change ( Fig. 4a-c). Expression levels of key genes of non-canonical Wnt/Ca 2+ Wnt signaling (WNT11, PRKCA and CAMK2D) showed significantly lowered expression levels following TiO 2 NPs or cadmium treatment. However, the TiO 2 bulk treatment showed non-significant changes in the expression levels of the said genes ( Fig. 4d-f). mRNA expression of key genes of Planar Cell Polarity pathway (ROCK1 and ROCK2) accounted for non-significant decrement following TiO 2 NPs or cadmium treatment. TiO 2 bulk treatment accounted for moderately significant increment in ROCK1 expression and non-significant increment in expression of ROCK2 ( Fig. 4g and h). Expression levels of HOXD13 showed significantly lowered expression levels following TiO 2 NPs or cadmium treatment, whereas, the TiO 2 bulk treatment showed non-significant changes (Fig. 4i).

Somite development.
It was observed that 10 µg/ml dose of TiO 2 NPs accounted for 20% decrement in the number of somites after 24 h which was comparable to that of the cadmium treated group, whereas, TiO 2 bulk treatment could not manifest any significant change (Fig. 4j).

Discussion
Nanomaterials have been reported to interact with protein molecules in unique ways and form a 'protein corona' that alters its physicochemical identity and affect its bio-distribution, kinetics and subsequent toxicity 22 . A previous study in our lab had shown that TiO 2 NPs interact with protein components of RPMI-1640 and result in higher indices of intermolecular interaction 23 . Egg albumen is reservoir of protein in an avian egg that meets the nutritional requirements of an embryo. Besides egg shell, shell membrane and chorio-allantoic membrane; egg albumen also regulates the trafficking of exogenous elements by acting as a natural biological barrier 24 . In the present study, a dose dependent increase in density of albumen was observed following addition of TiO 2 NPs. But, relatively lowest density observed at 10 µg/ml hints at effective dispersion of TiO 2 NPs in egg albumen. Higher intermolecular forces and cohesion are the key factors that determine viscosity and surface tension of liquids. In our study, a dose dependent decrement in viscosity and a reciprocal increment in surface tension are in support of our claim that higher intermolecular forces are as a result of higher concentration of TiO 2 NPs. Friccohesity is a product of frictional and cohesive forces within similar (protein-protein) and dissimilar (protein-nanoparticle) molecules 25 . A dose dependent decrement in friccohesity suggests weaker inter conversion between cohesive and . Expression of Wnt signaling pathway-related genes in TiO 2 NPs-treated chicken embryos. The expression of Wnt signaling pathway-related genes (a-h) including CTNNB1, PITX2, LEF1, WNT11, PRKCA, CAMK2D, ROCK1 and ROCK2 and (i) limb development gene HOXD13 was analyzed using reverse transcription polymerase chain reaction (RT-PCR) in control and TiO 2 NPs-treated embryos (n = 3), 4 h after treatment in shell-less culture at 60 h. All the Wnt signaling pathway related genes and limb development gene were downregulated in TiO 2 NPs-treated embryos compared to those of control embryos. *p < 0.05, **p < 0.01, ***p < 0.001, ns = not significant. (j) Somite numbers in control and TiO 2 NPs-treated embryos 24 h after treatment in shell-less culture (HH-23). There is a reduction in the number of somites in TiO 2 NPs-treated embryos as compared to the control embryos. The data are expressed as Mean ± SD. Statistical analysis was done by one way ANOVA followed by Dunnett's test. *p < 0.05, **p < 0.01, ***p < 0.001, ns = not significant. frictional forces implying towards a stronger nanoparticles-egg albumen interaction. Also, a decrement in activation energy is an indicator of increased proportion of collision/chemical reaction between the test compounds 23 . Relatively higher indices of activation energy recorded in TiO 2 NPs + albumen (10 µg/ml) is suggestive of more quantum of interaction between TiO 2 NPs and egg albumen. Molecular radii play an important role in dispersion of nanometals and its impact on biological systems. A dose dependent increment in molecular radii resulting due to TiO 2 NPs-egg albumen interactions implies towards formation of nanoparticle agglomerates at higher doses.
TiO 2 NPs + albumen recorded a shift in H-O-H stretching peak to 3478 cm −1 confirming interaction between TiO 2 NPs and albumen. Further a shift in peak from 2071 cm −1 (in albumen) to 2083 cm −1 (TiO 2 NPs + albumen) is attributable to interaction between C-O and amide groups of amino acids present in albumen. Peaks observed in the amide I and II secondary fingerprint region (in albumen) at 1651 cm −1 and 1642 cm −1 are attributable to C=O stretching and H-O-H bending respectively. However, a minor shift in C=O stretching and depletion of H-O-H bending (at 1641 cm −1 ) was possibly on account of electrostatic interaction due to Vander Waal forces taking place between albumen and TiO 2 NPs. Peak in amide III region (at 1243 cm −1 ) in albumen occurs due to N-H bending and C-N stretching of amino groups but, TiO 2 NPs + albumen recorded a shift (at 1551 cm −1 ) from amide III to amide II region. This shift also portrays major conformational changes in secondary components (α helices and β sheet) of proteins possibly due to their interaction with TiO 2 NPs. Fourier-self deconvolution approach was employed to assess secondary conformational changes in amide I and II region. Venyaminov and Kalnin 26 had reported that amide peak at 1610 ± 4 corresponds to NH bending of CO-NH 2 bond in glutamine. In our study, amide peak at 1610 cm −1 in TiO 2 NPs + albumen indicates possible deformity of glutamate in egg albumen. Role of glutamate in nutrition and metabolism is well reported 27 and hence impact of structurally altered glutamate on developing chicken embryo is postulated herein. Further, an increase in β sheet and α helices in TiO 2 NPs + albumen are possibly due to TiO 2 NPs mediated conformational changes, formation of aggregates or amyloids with protein moieties in egg albumen. These findings are the first to showcase interaction of TiO 2 NPs with egg albumen and the said physicochemical alterations.
Interaction of TiO 2 NPs with egg albumen prompted us to assess its impact on embryonic development using chicken egg as a model. Significant reduction in morphometric indices (body weight and length) and higher percentage mortality was recorded in developing chicken embryos at lower doses of TiO 2 NPs (10 µg/ml). But, the higher doses of TiO 2 NPs (50 or 100 µg/ml) failed to elicit a dose-dependent toxicological response possibly because TiO 2 NPs underwent physicochemical alterations as evidenced by relatively higher indices of density, viscosity and friccohesity coupled with lower activation energy hinting at formation of NP agglomerates. Whereas, lower extent of NP-egg albumen interactions observed at 10 µg/ml of TiO 2 NPs was instrumental in its effective bio-distribution and manifested said toxicity. Percentage mortality of bulk TiO 2 treated chicken embryos was comparable to that of higher doses (50 or 100 µg/ml) of TiO 2 NPs thus providing conclusive evidence that an altered physicochemical identity of NPs failed to induce a dose dependent toxicity in chicken embryos. Nanoparticles are known to cross biological barriers like the blood brain barrier and blood placenta barrier 28 . The results obtained herein indicate that the TiO 2 NPs could cross biological barriers within an avian egg and reach the embryo. The same was confirmed by ICP-OES studies that revealed presence of higher levels of TiO 2 NPs in the embryonic tissue at the lower dose (10 µg/ml).
Omphalocele is a ventral body wall defect and is accompanied by herniation of midgut into the abdominal cavity, failure in fusion of the anterior abdominal wall with 1/3000 frequency of occurrence in human population 29 . Teratogenic agents such as cadmium 30 , specific radiations 31 , fungal toxins 32 , etc. are known to induce omphalocele in various animal models. However, no known nanomaterials have been reported to induce omphalocele. Wnt signaling pathway has been implicated in various events of embryonic development such as cell differentiation, survival, migration, proliferation, adhesion and somite formation 33 . Canonical Wnts relay their signal via ß-Catenin pathway that control cell fate determination 33 . Whereas, the non-canonical Wnt signaling either through Wnt/Ca 2+ pathway or planar cell polarity pathway that controls cell adhesion and movement 33 . Results obtained herein were compared with cadmium induced omphalocele chicken embryo model 30 . PITX2, a bicoid-type homeodomain transcription factor, has known to be regulated by ß-Catenin dependent Wnt pathway 34 . In the Wnt/ß-Catenin pathway, the accumulation of ß-Catenin in the nucleus converts DNA-binding factor, lymphoid enhancing factor-1 (LEF1), to a transcriptional activator and is regulated through direct physical interaction with PITX2 and ß-Catenin 35 . In this study, downregulation in expression levels of CTNNB1, PITX2 and LEF1 following TiO 2 NPs treatment (10 µg/ml) could be a key factor in the disruption of somite myogenesis by inhibiting Wnt/ß-Catenin pathway. It has been postulated that cells from somites migrate into the parietal layer of lateral plate mesoderm (LPM) to assist in forming the lateral body folds 36 . PITX2 is known to regulate cell survival 37 and its downregulation may induce abnormal apoptosis in the somite and LPM that could further interfere with the movement of the lateral body wall folds 29 . These results justify the decrement in somite count obtained in our study following TiO 2 NPs treatment (10 µg/ml). WNT11, a member of the noncanonical Wnts, is an important epithelialization factor acting on the dermomyotome whereas, PRKCA and CaMK2D control actin-cytoskeleton organization and cell contractility 33,38 . Previous studies had implicated PRKCA and CaMK2D (activated by WNT11) in the regulation of cell-cell adhesion molecules (CAMs) such as cadherins. The resultant linkages between E-cadherin and actin filaments reinforce the cell-cell junctional connection 39 . In our study, downregulation of WNT11, PRKCA and CaMK2D genes after TiO 2 NPs treatment (10 µg/ml) possibly interfered with actin-cytoskeleton organization, cell movement and cell adhesion, thus disrupting noncanonical Wnt/Ca 2+ signaling that resulted in omphalocele. Rho kinases (ROCK) are involved in the regulation of various cellular functions (contraction, adhesion, migration, proliferation and apoptosis) including tissue closure during embryonic development. ROCK1 and ROCK2 mediate signaling from Rho to the actin cytoskeleton in the Wnt non-canonical pathway 40 . ROCK1 knockout (KO), ROCK2 KO, and ROCK1/2 double heterozygous mice has been reported to exhibit omphalocele phenotype due to disorganization of actin filament in the epithelial cells of umbilical ring 41  assembly, resulting in the failure of ventral body wall closure resulting in omphalocele. Defects in ventral body wall closure and omphalocele has also been reported with accompanying limb deformities in genetically modified experimental models 42 . Hox genes are important regulators of limb pattern in vertebrate development, HOXD13 misexpression in the hindlimb results in shortening of the long bones, including the femur, the tibia, the fibula and the tarsometatarsals 43 . In our study, significantly lowered expression of HOXD13 in TiO 2 NPs treated embryos corroborate with the observed omphalocele. Cadmium is known to use Ca 2+ ion channels and membrane transporters to enter in to the cells of a developing embryo. Further, it disrupts lateral plate mesodermal cells and induces omphalocele 30 . Therefore, cadmium treated chicken embryos were used as a disease control in our study wherein; expression levels of key genes of the Wnt signaling pathways were comparable to TiO 2 NP treated embryos. TiO 2 NPs are also known to cause free radicals induced cellular damage 6 . Free radical induced disruption of lateral plate mesoderm and somite myogenesis culminating in omphalocele in TiO 2 NPs treated chicken embryos is hypothesized in our study.
Besides their widespread industrial use, TiO 2 NPs have gained prominence in biomedical applications due to their long term photostability, superior biocompatibility, catalytic efficiency and a strong oxidizing power 44,45 . Photodynamic therapy for cancer, cell imaging, genetic engineering, drug delivery and biosensors are some of the reported biomedical applications of TiO 2 NPs 44,45 . Also, their use in diagnosis of cardiovascular diseases, diabetes mellitus, cancer and orthopaedic disorders underlines their prominence. But, omphalocele formation only at sub lethal (lower) concentrations reported herein raises concerns of toxicity benchmarks impacting foetal development. Hence, it raises an urge to study interactions of nanoparticles with biomolecules vis-à-vis particle size or surface modifications prior to their use in diagnostics or biomedical applications.

Conclusion
Nanometal oxides witness a wide range of biomolecules in a physiological environment that can alter their behavior and responses. In the present study, TiO 2 NPs were found to interact with egg albumen as evidenced by changes in their proteinic secondary structure. These interactions could possibly allow TiO 2 NPs to traverse the biological barriers (shell membrane and CAM) within chicken egg and affect the growth and development of embryos and cause malformations like omphalocele and flexed limbs. Also, the observed mortality and significant decrement in morphometry (whole weight and length) are attributable to TiO 2 NPs-albumen interactions. Omphalocele formation in TiO 2 NPs treated groups is possibly due to the disruption of somite myogenesis as evidenced by alterations in expression of key genes of Wnt signaling pathway. Hence, use of TiO 2 NPs in diagnostics and therapy warrants a detailed research in embryos by taking into account its particle size, surface modifications and interaction with biomolecules.

Materials and Methods
Availability of Data and Materials. The datasets supporting the conclusions of this article are included within the article.

Physicochemical analysis.
The physicochemical properties such as density, viscosity, surface tension, activation energy, friccohesity and molecular radii were assessed in absence or presence of TiO 2 NPs in egg albumen (freshly collected). TiO 2 NPs were suspended in egg albumen at 1, 5, 10, 25, 50 and 100 μg/ml concentrations. Density of TiO 2 NPs in water and egg albumen were determined with Anton Paar Density and Sound velocity Meter (DSA 5000 M). Density was calculated using equation 1: 2 (ρ° at m → 0 is limiting density, Sρ is the 1 st slope) Viscosity was measured as viscous flow times (VFT) using Borosil Mansingh Survismeter 23 at physiological temperature of 37 °C (LAUDA ALPHA RA 8 thermostat) and calculated by equation 2: (η° is viscosity of water and t°, t are flow times of solvent and mixtures respectively) The η data were regressed with following equation 3: (η° atm → 0 is limiting viscosity; S η is the 1 st degree slope). Surface tension was measured by counting pendent drop numbers (PDN) using Borosil Mansingh Survismeter and calculated by equation 4: (γ° is surface tension of water, η° and η are pendent drop numbers of medium and solutions respectively) The γ data were regressed for limiting values γ° at m → 0 with following equation 5: γ = γ°+ γ S m (5) (γ° is limiting surface tension, and S γ is the 1 st degree slope) Friccohesity was calculated using Mansingh equation 6 46 : (σ is friccohesity, t and t° are the sample and solution viscous flow times respectively, η° and η are the pendant drop numbers of medium and solutions respectively) Reference friccohesity was calculated by equation 7 where, η° and γ° are the viscosity and surface tension of references respectively. For activation energy, the partial molar volume V 2 was calculated with following equation 8: 2 (M is molar mass, ρ° is density of water and ρ is density of solution) The V 1 for water or albumen at 37 °C is calculated with equation 9: 1 V 1 and V 2 are used for calculating activation energy by using equation 10: (Δµ 1 * is activation energy of water or albumen, R is gas constant, h is Planck constant and N is Avogadro number (6.023 × 10 23 ). Activation energy (Δμ 2 * J/mol) was calculated by using equation 11:  Shell-less culture and dosing. In a separate set of experiment, procured eggs were incubated for 60 h in standard conditions and later were explanted into shell-less culture as per Dugan, et al. 50 . The embryos were divided into four groups of six eggs each viz control (50 µl PBS), positive control (50 µl of 50 µM CdCl 2 ), TiO 2 NPs treated (50 µl of 10 µg/ml) and bulk TiO 2 (50 µl of 10 µg/ml). Dosing was done directly on blastodisc using a micropipette and embryos were incubated for 4 h or 24 h.
Autopsy, RNA isolation and qPCR study. Developing embryonal discs (three per group, HH 17; whole embryo) were transferred in RNA later solution (Invitrogen, California, USA). Total RNA was isolated using TRIzol reagent (Invitrogen, California, USA) and cDNA was synthesized by reverse transcription of 1 μg of total RNA using iScript cDNA Synthesis kit (BIORAD, California, USA). For HOXD13, total RNA was isolated from limb bud of 4 day old control and treated embryos. Quantitative RT-PCR was performed using SYBR Select Master Mix (Applied Biosystems) in QuantStudio12K (Life Technologies) real-time PCR machine with primers (Table 1) to detect selected messenger RNA (mRNA) targets. The relative mRNA expression levels were normalized against expression levels of GAPDH for each sample and analyzed using 2 −∆∆CT method 51 .
Somite development. The embryos (three per group) were dissected from their membranes 24 hours after treatment (HH 23) and inspected using the dissecting microscope to count somite numbers.
Statistical analysis. Data analysis was carried out by unpaired Student's t-test or one way analysis of variance (ANOVA) using Graph Pad Prism 6.0 (CA, USA). Differences between control and treatment groups were deemed to be significant when P < 0.05.