Mycotoxin Fumonisin B1 Interferes Sphingolipid Metabolisms and Neural Tube Closure during Early Embryogenesis in Brown Tsaiya Ducks

Fumonisin B1 (FB1) is among the most common contaminants produced by Fusarium spp. fungus from corns and animal feeds. Although FB1 has been known to cause physical or functional defects of embryos in humans and several animal species such as Syrian hamsters, rabbits, and rodents, little is known about the precise toxicity to the embryos and the underlying mechanisms have not been fully addressed. The present study aimed to investigate its developmental toxicity and potential mechanisms of action on sphingolipid metabolism in Brown Tsaiya Ducks (BTDs) embryos. We examined the effect of various FB1 dosages (0, 10, 20 and 40 µg/embryo) on BTD embryogenesis 72 h post-incubation. The sphingomyelin content of duck embryos decreased (p < 0.05) in the highest FB1-treated group (40 µg). Failure of neural tube closure was observed in treated embryos and the expression levels of a neurulation-related gene, sonic hedgehog (Shh) was abnormally decreased. The sphingolipid metabolism-related genes including N-acylsphingosine amidohydrolase 1 (ASAH1), and ceramide synthase 6 (CERS6) expressions were altered in the treated embryos compared to those in the control embryos. Apparently, FB1 have interfered sphingolipid metabolisms by inhibiting the functions of ceramide synthase and folate transporters. In conclusion, FB1-caused developmental retardation and abnormalities, such as neural tube defects in Brown Tsaiya Duck embryos, as well as are partly mediated by the disruption of sphingolipid metabolisms.


Introduction
Fumonisins (FBs), discovered in 1988, are mycotoxins produced by fungi of the Fusarium species including Fusarium verticillioides and Fusarium proliferatum [1,2]. The most abundant FBs are the B-series, fumonisin B 1 (FB 1 ), B 2 (FB 2 ) and B 3 (FB 3 ), which can be found in natural conditions. The FB 1 and FB 2 limitations of the Food and Drug Administration (FDA) in the US [3] and European Union (EU Regulation 1126/2007) [4] were 800-4000 and 2000-4000 µg/kg for cereal-based products, respectively. The average FB 1 concentration of food exceeds these mycotoxin limits in some countries such as Lebanon and Brazil [5,6]. Among them, FB 1 is the most toxic and over 70% of food products are contaminated by this toxin [7][8][9], some of which might reach the maximal concentration limit and can be harmful to animal and human health [10].
Previous studies have shown that the consumption of FB 1 contaminated corn has been associated with esophageal cancer in humans [11][12][13] and with various animal diseases including equine leucoencephalomalacia and hepatotoxicity, as well as porcine pulmonary edema [14][15][16]. FB 1 has been indicated being associated with the toxicity in different organs, such as livers, kidneys, lungs, and intestine [17][18][19] as well as in nervous and cardiovascular systems of animals [20][21][22].
It has been found that the structural similarity of FBs with the sphingoid bases sphinganine (Sa) and sphingosine (So) is the main cause to the disruption of sphingolipid metabolisms. FB 1 can inhibit ceramide synthesis by an unsubstituted primary amino group at C2 that is structurally similar to Sa and So. This leads to the accumulation of Sa and So during the process of sphingolipid biosynthesis [23][24][25]. Moreover, the ratio of free sphingoid bases (Sa/So) is found increase in blood, tissues (e.g., lung, liver, and intestine) and in cultured cells [26,27] by the FB 1 administration. Therefore, inhibition of ceramide synthase by FB 1 causes cell damage due to the disruption of the membrane integrity [28].
It has been known that FB 1 also causes the malfunction of folate transporter and neural tube defects (NTDs), a common congenital abnormality occurs when the embryonic neural tube fails to properly close during the first few weeks of mouse development [29]. The mechanism of FB 1 action in mammals is due to the inhibition of ceramide synthesis. FB 1 causes the interruption or aberrant acylation of Sa in the endoplasmic reticulum, the major site of membrane lipid biosynthesis in mammalian cells; in turn, the reduced synthesis of sphingomyelin compromises the proper function of glycosylphosphatidylinositol (GPI)-anchored proteins, such as the folate transporter [30,31] on cell membranes. However, the molecular and cellular mechanisms of FB 1 action on, e.g., NTDs, in avian species remain to be determined. Therefore, the present study was aimed to use an in ovo model system to assess the related gene expression profiles and development toxicity of duck embryos after the treatments of various FB 1 doses. The possible molecular mechanism of FB 1 during the development of the nervous system in duck embryos was investigated.

Developmental Toxicity of FB 1 in Brown Tsaiya Duck (BTD) Embryos
Treatments with FB 1 did not cause embryo death at 72 h post-incubation (Table 1). Results showed that the viability of embryos was unaffected by the treatment of FB 1 . However, FB 1 caused a growth retardation and delay of the developmental stage, evaluated by embryonic crown-to-tail length (ECTL) and numbers of somites. The percentages of malformation were higher in the FB 1 treated groups compared to the untreated control (0% vs. 73.7-88.9%; p < 0.0001).  Table 2 (p = 0.0417) and Figure 1. The percentages of abnormal neural tube formation and somitogenesis were also significantly increased in FB 1 -treated embryos (p < 0.0001) when compared with that in the non-FB 1 treated control. Morphological examination of the neural tube and the somite by transverse histological sections showed disclosure or defective structures; somitomeres appeared more edematous or misaligned in all groups treated with FB 1 compared to those in the control (Figure 1).  The expression of a neurulation-related gene, sonic hedgehog (Shh), was down-regulated in the FB1-treated group (p = 0.0227). However, the marker genes for somite development and neural tube related-genes, Paired Box 3 (Pax3) and Paired Box 7 (Pax7), were unaffected among different groups ( Figure 2).  The expression of a neurulation-related gene, sonic hedgehog (Shh), was downregulated in the FB 1 -treated group (p = 0.0227). However, the marker genes for somite development and neural tube related-genes, Paired Box 3 (Pax3) and Paired Box 7 (Pax7), were unaffected among different groups ( Figure 2).
show abnormal images of the neural tube defect after injection with (B) (10), (C) (20), or (D) (40) µg/embryo of FB1, respectively. The black dash-line indicates the position of the transverse section. The lower (A1-D1) panel shows the transverse sections corresponding to (A-D) in the upper panel, respectively. (A1) A normal neural tube closure; (B1-D1) defective neural tube closure (blue dot box), as well as the cells of the somitomeres appear more edematous or misaligned in comparison with the non-FB1 injected control. The red and black arrows indicate the neural tube and somite, respectively. NC: notochord. Scale bar = 0.5 mm.
The expression of a neurulation-related gene, sonic hedgehog (Shh), was down-regulated in the FB1-treated group (p = 0.0227). However, the marker genes for somite development and neural tube related-genes, Paired Box 3 (Pax3) and Paired Box 7 (Pax7), were unaffected among different groups ( Figure 2).

Exposure to FB 1 Alters the Sphingolipid Metabolism and the Related-Gene Expression
The sphingomyelin content was examined for its dose-response relationship with the inhibition of ceramide synthase in sphingolipid metabolism pathway after FB 1 treatment. Results showed that the sphingomyelin content decreased after the highest dose (40 µg) of FB 1 treatment in BTD embryos ( Figure 3, p = 0.0232), when compared with the non-injected control group.

Exposure to FB1 Alters the Sphingolipid Metabolism and the Related-Gene Expression
The sphingomyelin content was examined for its dose-response relationship with the inhibition of ceramide synthase in sphingolipid metabolism pathway after FB1 treatment. Results showed that the sphingomyelin content decreased after the highest dose (40 µg) of FB1 treatment in BTD embryos ( Figure 3, p = 0.0232), when compared with the noninjected control group. FB1-induced NTDs in BTD embryos could be due to the disturbance of sphingolipid metabolisms. Therefore, we investigated the expression of sphingolipid metabolism-related genes in duck embryonic tissues. Results showed that the sphingolipid metabolism related-genes (CERS3, CERS5, DEGS1, SGPL1, SGPP1, SPHK1, and PLPP1) (Figure 4), genes coding for sphingomyelin synthase (SMPD3 and SGMS1) and glucosylceramide synthase-related gene; UGCG ( Figure 5) did not differ among the treatment groups, but N-acylsphingosine amindohydrolase 1 (ASAH1) expression was up-regulated in the FB1treated embryos relative to that in the control embryos (p = 0.0326). Embryos treated with the highest dose of FB1 down-regulated ceramide synthase 6 (CERS6) and 5,10-methylenetetrahydrofolate reductase (MTHFR) gene expressions compared to those of the con- FB 1 -induced NTDs in BTD embryos could be due to the disturbance of sphingolipid metabolisms. Therefore, we investigated the expression of sphingolipid metabolism-related genes in duck embryonic tissues. Results showed that the sphingolipid metabolism relatedgenes (CERS3, CERS5, DEGS1, SGPL1, SGPP1, SPHK1, and PLPP1) (Figure 4), genes coding for sphingomyelin synthase (SMPD3 and SGMS1) and glucosylceramide synthase-related gene; UGCG ( Figure 5) did not differ among the treatment groups, but N-acylsphingosine amindohydrolase 1 (ASAH1) expression was up-regulated in the FB 1 -treated embryos relative to that in the control embryos (p = 0.0326). Embryos treated with the highest dose of FB 1 down-regulated ceramide synthase 6 (CERS6) and 5,10-methylenetetrahydrofolate reductase (MTHFR) gene expressions compared to those of the control embryos.
FB1-induced NTDs in BTD embryos could be due to the disturbance of sphingolipid metabolisms. Therefore, we investigated the expression of sphingolipid metabolism-related genes in duck embryonic tissues. Results showed that the sphingolipid metabolism related-genes (CERS3, CERS5, DEGS1, SGPL1, SGPP1, SPHK1, and PLPP1) (Figure 4), genes coding for sphingomyelin synthase (SMPD3 and SGMS1) and glucosylceramide synthase-related gene; UGCG ( Figure 5) did not differ among the treatment groups, but N-acylsphingosine amindohydrolase 1 (ASAH1) expression was up-regulated in the FB1treated embryos relative to that in the control embryos (p = 0.0326). Embryos treated with the highest dose of FB1 down-regulated ceramide synthase 6 (CERS6) and 5,10-methylenetetrahydrofolate reductase (MTHFR) gene expressions compared to those of the control embryos.

Discussion
It has been known that FB1 toxicity involves in the disturbance of the sphingolipid metabolisms. In the present study, we found that FB1 was a strong inhibitor of ceramide synthase. It disturbed the metabolism of the sphingolipids ( Figure 6) which is essential for stabilizing the structure and function of the developing embryos.

Discussion
It has been known that FB 1 toxicity involves in the disturbance of the sphingolipid metabolisms. In the present study, we found that FB 1 was a strong inhibitor of ceramide synthase. It disturbed the metabolism of the sphingolipids ( Figure 6) which is essential for stabilizing the structure and function of the developing embryos. all other treatment groups was detected (p > 0.05). a, b Bars without the same superscripts differ (p < 0.05); three replicates.

Discussion
It has been known that FB1 toxicity involves in the disturbance of the sphingolipid metabolisms. In the present study, we found that FB1 was a strong inhibitor of ceramide synthase. It disturbed the metabolism of the sphingolipids ( Figure 6) which is essential for stabilizing the structure and function of the developing embryos. The avian embryo is an excellent model system to study the developmental toxicity and mechanisms involving the patterning during early embryogenesis [34]. In previous Figure 6. The hypothesis of de novo sphingolipid pathways shows the synthesis of ceramide and sphingomyelin reported by Lumsangkul et al. [33]. The ceramide synthesis pathway could be blocked by FB 1 indicated by red line.
The avian embryo is an excellent model system to study the developmental toxicity and mechanisms involving the patterning during early embryogenesis [34]. In previous studies, avian embryos were used to study neurulation, somitogenesis and the effects of various chemicals, such as choline, caffeine, and glucose, on early development of the neural tube and somites [35][36][37][38][39]. In the present study, BTD embryos were used to investigate the toxicity of FB 1 on the induction of NTDs. The possible molecular mechanism of FB 1 in relation to the developmental toxicity of BTD embryos was also studied.
Due to little evidence available on the toxic effect of FB 1 to avian embryogenesis, to our best knowledge, the present study is the first report on the FB 1 toxicity during early embryogenesis and later development. The FB 1 intoxicated the BTD embryos by affecting their developmental stages manifested by the growth retardation and consequently the NTDs. The retardation of the embryonic development measured by the ECTL, somite numbers (Table 1), as well as the FB 1 -induced NTDs was confirmed. Our finding is consistent with the observation reported by Liao et al. [40] who demonstrated that FB 1induced NTDs of fetuses when a pregnant mouse was fed with FB 1 -contaminated feeds.
Although the NTDs were not discernible in general morphology of the whole embryo ( Figure 1A-D), unambiguous NTD structures were observed through the transverse histological sections at the lambo-sacral region or lower trunk ( Figure 1B1-D1). Morphological analysis by H&E staining revealed clearly the presence of NTDs in the FB1-treated embryos ( Figure 1B1-D1) but not in the control group ( Figure 1A,A1). Normally, neural tube closure is a dynamic process that starts from the head region of the neural plate and progresses toward the caudal region, where the two opposing neural folds elevate and fuse together at the midline by epithelial cell adhesion molecule (EpCAM) [41,42]. The fused edge at the dorsal neural tube is smooth and curved; however, embryos exposed to FB 1 could result in forming a discontinuous dorsal neural tube edge ( Figure 1B1-D1). Our observation clearly showed that FB 1 could disrupt the cellular activity during neurulation and consequently lead to NTDs.
Segmentation is initiated at very early stage in development through the formation of embryonic somites. Somitogenesis plays a very important role in establishing the bone and skeletal muscles in the body and limbs [43], which is closely associated with the neurulation process. It has been shown that the neural tube is required for proper somitogenesis and differentiation [44]. The development and differentiation of somites are dependent on signals emitted from the ipsilateral neural tube [45]. In this study, we Toxins 2021, 13, 743 7 of 12 also found that numbers of somites decreased (Table 1) and somitomeres appeared more irregularly scattered in FB 1 -treated embryos compared to those in the control embryos ( Figure 1B1-D1).
Sonic hedgehog (Shh) protein, an important morphogen, is generated from the notochord and the floor plate to regulate many morphogenetic events during early embryo development, such as somitogenesis [46]. It is also involved in myogenesis by inducing myogenic factor (Myf5) expression directly and myoblast determination protein (MyoD) indirectly [47]. Paired Box 7 (Pax7) and Paired Box 3 (Pax3) are two key genes expressing in dorsal neural tube and dorsal pre-migratory neural crest cells (NCCs). These genes are expressed in a population of muscle precursor cells to maintain their uncommitted state throughout embryonic development as well as to play a key role in embryonic muscle development [48][49][50]. In the present study, these somitogenesis and neurulation related-genes were interfered along with a decreased expression of Shh after FB 1 treatment (Figure 2).
Due to the structural similarity to sphingosine (So) and sphinganine (Sa), the primary precursors of sphingolipids, FB 1 could compete with Sa and So for their integration into the sphingolipid metabolism pathway. However, it is unclear why the expressions of sphingolipid metabolism related-genes (CERS5, SGPL1, SGPP1, SPHK1, and PLPP1), sphingomyelin synthase (SMPD3 and SGMS1) and glucosylceramide synthase-related genes (UGCG) were not affected by FB 1 , which requires more investigation.
To our knowledge, we are the first report to analyze different isoforms of CERS gene in avian species. We found that the isoform CERS6 was related to FB 1 toxicity, likely, via inhibition of ceramide synthase activity and disturbance of sphingolipids metabolisms in BTDs. As abovementioned, we also showed that FB 1 increased incidence of NTDs by impairing the functions of folate transporters. Sphingomyelin is a major component of the plasma membrane and is required for the proper function of GPI-anchored proteins of the folate transporter [29]. In this study, we found that the sphingomyelin content in BTD embryo was reduced by FB 1 treatments (Figure 3) due to the inhibition of sphingolipid biosynthesis. We also found that MTHFR gene in folate metabolism pathway was altered in FB 1 -treated BTD embryos. Fumonisin B 1 might decrease the glycosphingolipids synthesis that in turn increased the incidence of incomplete neurulation or NTDs during embryogenesis.
Ceramidases (N-acylsphingosine amidohydrolase, ASAH1) are a family of hydrolases that directly regulates the intracellular balance of ceramide by catalyzing the degradation of ceramide into sphingosine. Because ceramide degradation is the only source of sphingosine, these enzymes are not only essential for modulating ceramide-mediated signaling but also for the functions of sphingosine and sphingosine-1-phosphate. In the mouse, ASAH1 expresses during early embryogenesis and disruption of ASAH1 gene results in embryonic lethality [51]. In addition, ASAH1 overexpression has been reported in various human cancers [52]. The sphingosine is a sensitive biomarker for FB 1 exposure in animals, as well as being proposed for monitoring FB 1 exposure in humans. It correlates with liver and kidney toxicity and often a precede sign of intoxication [53,54]. In the present study, the highest dose (40 µg) of FB1 exposure up-regulated ASAH1 expression compared to the control group. Such alterations might be resulted in changes of sphingosine levels and imbalance of ceramide through modulating the expression of ASAH1 gene. Therefore, embryos showing a disturbance ceramide synthesis profile may be involved in neurodegeneration and reduction of neural cell numbers in the early developing brain [55][56][57], and finally led to NTDs of fetuses.
The metabolism and the mode of action of the FB 1 in avian species are largely unknown. The present study provided the evidence about the effect of FB 1 exposure on the levels of ceramide synthase and sphingomyelin. Moreover, FB 1 might also affect the function of folate transporter and cause NTDs. As abovementioned, folate transporter is a major component of the plasma membrane and is required for the proper function of GPI-anchored proteins. Our observations echo with the findings of Marasas et al. (2004) and Liao et al. (2004) [29,40], who have reported that FB 1 caused both neural tube and craniofacial defects in mouse embryos by inhibiting sphingolipid biosynthesis, as well as folate transport. Moreover, sphingolipids and cholesterol are typically embedded in the lipids rafts of extracellular leaflet of the cell membrane; interactions between cholesterol, sphingomyelin and FB 1 warrant further investigation [58,59].
In conclusion, FB 1 induces growth retardation, developmental abnormalities and neural tube defects of BTD embryos. The underlying mechanism is, at least, partially mediated by disrupting ceramide synthesis in sphingolipid metabolisms.

Eggs Used and Conditions of Incubation
A total of 80 fertilized BTD eggs obtained from Yilan Livestock Research Institute, Council of Agriculture were used in this study. All the eggs were preselected based on weight (61 ± 0.09 g) and randomly assigned into four treatment groups; twenty BTD eggs in each group were incubated in an incubator (YC-M10, Yongcheng Incubation Machine Technology Co., Ltd., Changhua, Taiwan) at 37.5 • C with 65% relative humidity. The whole study was carried out in strict accordance with the guideline recommended and approved by the Institutional Animal Care and Use Committee 365 (IACUC) of the National Chung Hsing University (Permit number: 100-02; date of approval: 24 January 2011).

Treatment Concentrations of Fumonisin B 1
The FB 1 was purchased from (Cat No. 116355-83-0, Sigma-Aldrich, St. Louis, MO, USA). The FB 1 solutions were prepared with deuterium-depleted water (DDW) to a stock concentration of 1 mg/mL (1000 ppm) and diluted into 10, 20 and 40 µg before use.
Various FB 1 injection dosages (0, 10, 20 and 40 µg/embryo) were directly applied at 0 h post-incubation to embryos in ovo. Briefly, and approximately 100 µL FB 1 solution was injected into the air chamber of the egg via a small hole made at the blunt-end of the egg. After that, the hole of the egg was sealed with the adhesive tape. The treated embryos were then incubated for 72 h for further observation.

Embryo Viability and Development
Twenty BTD eggs per treatment group were respectively dissected at 72 h postincubation to determine embryo viability. Embryonic development and viability were defined and distinguished based on the standard speed of developmental progression established previously [32]; parameters including heartbeats, formation of yolk-sac and blood vessels, brain development, as well as organogenesis were examined.

Histological Assessments
After 72 h post-incubation, ten embryos were sampled from each treatment group and were fixed in 10% buffered formalin for 24 h. The tissues of embryos were pre-embedded in standard agarose gel prior to embedding in paraffin wax. Three-micrometer thick sections were cut consecutively from the paraffin blocks, mounted onto glass slides, de-paraffinized, and then dehydrated. Cross sections of the neural tube at the lumbosacral region were stained with hematoxylin and eosin (H&E) to assess morphologies of the neural tube and somite structures. The sections were examined and recorded using a DinoCapture 2.0 digital microscope (Dino-Lite, Los Angeles, CA, USA) attached to a Nikon LABOPHOT-2 binocular microscope (Nikon, Tokyo, Japan).

Sphingomyelin Quantification by Colorimetric Assay
Ten embryos from each group were collected and crushed by ultrasonic cell crusher (40 Hz) followed by the extraction with chloroform and methanol (2:1, v/v). Liquids were dried in a vacuum dryer and re-dissolved by using tert-Butyl alcohol (TBA). Sphingomyelin levels were analyzed by using sphingomyelin colorimetric Assay Kit (Cayman, MI, USA, Item no: 10009928). Sphingomyelin content was quantified with a Molecular Devices ELISA reader (Sunnyvale, CA, USA) at a wavelength 595 nm based on a sphingomyelin standard curve.

Revers Transcription and Quantitative Real-Time PCR (qPCR) Analyses
Total RNAs were extracted with a total RNA extraction kit (Invitrogen, PureLink™ RNA Mini Kit, Carlsbad, CA, USA). Cytoplasmic RNAs from embryos were reversetranscribed to generate first-strand cDNA by iScript™ cDNA Synthesis Kit (Bio-Rad, Hercules, CA, USA). Candidate genes (Table 3) were chosen for qPCR analysis. The qPCR was performed in triplicate using 100 ng of cDNA, 0.8 (0.25) mM of primers and iTaq Universal SYBR Green supermix (2X) on a CFX Connect™ Real-Time PCR System (Bio-Rad, Hercules, CA, USA). Thermal cycling conditions were 95 • C for 30 s (holding stage), 40 cycles of 95 • C for 15 s and 60 • C for 30 s (cycling stage) and followed by 95 • C for 15 s, 60 • C for 60 s, 95 • C for 15 s (melt curve stage). The relative expression of genes was analyzed according to the 2 −∆∆Ct method. Each sample of each group was measured in triplicate and the assay was repeated three times. The quantification was standardized to an endogenous control GAPDH.

Statistical Analyses
All experimental data were analyzed using analysis of variance (ANOVA) procedure of SAS Enterprise Guide Software V.9.4 (SAS Institute, Cary, NC, USA). Least square means (LSM) were compared by using Tukey's test. A probability level at p < 0.05 was considered as statistically significant.