Tributyltin Inhibits Neural Induction of Human Induced Pluripotent Stem Cells

Tributyltin (TBT), one of the organotin compounds, is a well-known environmental pollutant. In our recent study, we reported that TBT induces mitochondrial dysfunction, in human-induced pluripotent stem cells (iPSCs) through the degradation of mitofusin1 (Mfn1), which is a mitochondrial fusion factor. However, the effect of TBT toxicity on the developmental process of iPSCs was not clear. The present study examined the effect of TBT on the differentiation of iPSCs into the ectodermal, mesodermal, and endodermal germ layers. We found that exposure to nanomolar concentration of TBT (50 nM) selectively inhibited the induction of iPSCs into the ectoderm, which is the first step in neurogenesis. We further assessed the effect of TBT on neural differentiation and found that it reduced the expression of several neural differentiation marker genes, which were also downregulated by Mfn1 knockdown in iPSCs. Taken together, these results indicate that TBT induces developmental neurotoxicity via Mfn1-mediated mitochondrial dysfunction in iPSCs.

which is responsible for cellular energy supply, by allowing impaired mitochondria to be recycled 14,17 . Therefore, mitochondrial dynamics are necessary for cell survival as well as adaptation to changing conditions needed for cell growth 18 . Several studies have shown the relationship between mitochondrial fragmentation and cellular and neurodevelopmental defects. For example, Mfn1 knockout mice show developmental delay at the midgestational embryonic stage and ultimately die 19 . In addition, embryonic fibroblasts from these knockout mice display distinct types of fragmented mitochondria, which is a phenotypical characteristic of a severe reduction in mitochondrial fusion 19 . Based on these findings, we hypothesized that nanomolar TBT could also affect the developmental process of iPSCs, which can differentiate into somatic cells from three developmental germ layers (ectoderm, mesoderm, endoderm) 20 .
In the present study, we investigated the effect of TBT on the differentiation of iPSCs into germ layers as a model of human organ development 21 . Our results showed that 50 nM TBT selectively suppressed the induction of iPSCs into the ectoderm during neurogenesis. Moreover, TBT reduced the expression of several neural differentiation marker genes, which were also downregulated by Mfn1 knockdown. These data suggest that TBT-induced neurodevelopmental toxicity involves Mfn1-mediated mitochondrial dysfunction in human iPSCs, without affecting mesodermal and endodermal inductions.

Effect of TBT on Differentiation of iPSCs into Three Germ Layers.
To examine the effects of TBT on fetal development, we studied its effects on the differentiation of iPSCs into ectodermal, mesodermal, and endodermal germ layers. In ectoderm induction, we found that treatment with 50 nM TBT significantly reduced the gene expression of the OTX2 marker that regulates neurogenesis 22 (Fig. 1a). We also found that TBT reduced another ectodermal marker IRX1 expression 23 (Fig. 1b). In contrast, TBT had little effect on the inductions of mesodermal (BRACHYURY, MIXL1) and endodermal (SOX17, FOXA2) markers [24][25][26][27] (Fig. 1c-f). TBT at 50 nM has the ability to bind to PPARγ with a higher affinity than that of the intrinsic ligands, and these genomic transcriptional activations have been reported to mediate neurodevelopmental defects 5 . To investigate the molecular mechanisms by which TBT inhibits ectodermal induction, we examined the effect of the PPARγ agonist rosiglitazone (RGZ), which had been confirmed in our previous report, as having agonistic effects on PPARγ 10 . We found that RGZ did not reduce OTX2 expression (Fig. S1). Taken together, these data suggest that TBT inhibits ectodermal induction in iPSCs regardless of PPARγ activation.

Effect of TBT on Neural Differentiation of iPSCs.
The production of ectodermal germ layer is the first step in neurogenesis 28 . Next, to examine the effects of TBT on neural differentiation, we performed the differentiation into neural progenitor cells (NPCs) from iPSCs (Fig. 2a). There has been reported for neural differentiation methods from iPSCs based on different protocols, such as neural induction through embryoid body formation 29 , inhibition of TGF-β and BMP signaling pathways (dual SMAD inhibition) 28 , or forced expression of neurogenin-2 with puromycin selection 30 . We chose dual SMAD inhibition protocol, because it is simple, non-viral, low-priced, non-time consuming, highly efficient, reproducible and similar to in vivo neurogenic processes among these methods. We examined the expression of several neural differentiation markers. Immunocytochemical analysis showed strong expression levels of PAX6, a marker of neuroectoderm 28 , by day 4 compared to that on day 0 while TBT exposure significantly decreased the percentage of PAX6 positive cells at day 4 ( Fig. 2b). In addition, real-time PCR analysis revealed that TBT significantly downregulated the expression of Nestin (day 8), which is a marker of NPCs 31 (Fig. 2c). These data suggest that TBT has an inhibitory effect on the neural induction of iPSCs.

Effect of Mfn1
Knockdown on Neural Differentiation of iPSCs. We previously found that TBT induced mitochondrial dysfunction by degrading Mfn1 11 . In addition, as described above, Mfn1 is reported to be involved in neural development 19 . To investigate the involvement of Mfn1 in the effects of TBT on neural induction, we performed a knockdown (KD) of Mfn1 using lentivirus-delivered shRNAs. Our previous study showed that the KD was selective for Mfn1 and not Mfn2, with an efficiency of approximately 70% 32 . The Mfn1 KD cells were used to perform neural induction. We found that Mfn1 KD significantly reduced the gene expression of OTX2 at day 2 after neural induction (Fig. 3a). Immunocytochemical analysis revealed a strong expression of PAX6 by day 4 after neural induction in the control cells while Mfn1 KD significantly reduced the percentage of PAX6 positive cells at day 4 ( Fig. 3b). We further found that Mfn1 KD significantly decreased the gene expression of Nestin at day 8 (Fig. 3c). These data suggest that Mfn1 is involved in the TBT-mediated negative effects on neural induction of iPSCs.

Negative Regulation of Neural Induction by TBT Exposure. A previous report indicates that Mfn1
directly binds Ras and Raf, thereby inhibiting Ras-Raf-ERK signaling, as determined using biochemical analysis 33,34 . Thus, ERK has been reported to be activated after depletion of Mfn1 35 . Moreover, ERK signaling is known to inhibit neural induction via OTX2 silencing in human embryonic stem cells 36 . Therefore, we focused on investigating the involvement of ERK signaling in the effect of TBT on neural induction. We found that TBT exposure significantly increased basal ERK phosphorylation levels, while this effect was neglected by treatment with the ERK inhibitor U0126 ( Fig. 4a and b). To further investigate whether OTX2 downregulation in TBT-exposed cells was mediated by ERK signaling, we examined the effect of U0126 on OTX2 expression. Only U0126 treatment increased OTX2 expression during ectodermal induction (Fig. 4c), suggesting that ERK signaling negatively regulated neural induction. In contrast, U0126 suppressed the expressions of BRACHYURY and SOX17 marker genes in mesodermal and endodermal induction respectively ( Fig. 4d and e), supporting previous reports that ERK signaling contributes to mesodermal and endodermal differentiation 37,38 . U0126 quenched the negative Figure 1. TBT inhibits ectodermal induction of iPSCs. (a,b) Ectodermal induction was initiated after exposure to 50 nM TBT for 24 h. Cells were continuously exposed to TBT throughout the induction. At day 4 after the induction, the expression of ectodermal markers (OTX2, IRX1) was examined using real-time PCR analysis. (c,d) Mesodermal induction was initiated after exposure to 50 nM TBT for 24 h. Cells were continuously exposed to TBT throughout the induction. At day 1 after the induction, the expression of mesoderm markers (BRACHYURY, MIXL1) was examined using real-time PCR analysis. (e,f) Endodermal induction was initiated after exposure to 50 nM TBT for 24 h. Cells were continuously exposed to TBT throughout the induction. At day 4 after the induction, the expression of endoderm markers (SOX17, FOXA2) was examined using real-time PCR analysis. Each bar represents the mean ± SD from three independent experiments. *P < 0.05.

Discussion
In the present study, we demonstrated that exposure to 50 nM TBT inhibited ectoderm induction by suppression of OTX2 in iPSCs. In addition, the negative effect of TBT on neurogenesis was likely mediated by Mfn1 degradation, followed by ERK phosphorylation. Based on the data obtained in our study, we propose a mechanism by which the TBT-induced developmental neurotoxicity is mediated by mitochondrial dysfunction (Fig. 5).
We showed that TBT selectively inhibited ectodermal induction during the differentiation of iPSCs into the three germ layers (Fig. 1). A previous report suggests the presence of organotin compounds such as TBT, at concentrations between 50 and 400 nM in human blood 3 . Thus, the concentration (50 nM) of TBT used in our studies NPC induction from iPSCs using dual SMAD inhibition. Neural induction was initiated after exposure to 50 nM TBT for 24 h. Cells were continuously exposed to TBT throughout neural differentiation. (b) At day 4 after neural induction, the expression of neuroectodermal marker, PAX6, was observed by immunocytochemistry using anti-PAX6 antibodies. Nuclei were counterstained with DAPI. PAX6 positive nuclei were counted using the ImageJ software. Bar = 100 μm. (c) At day 8 after neural induction, the expression of NPC marker, Nestin, was examined using real-time PCR analysis. Each bar represents the mean ± SD from three independent experiments. *P < 0.05.  (a) Cells were exposed to TBT (50 nM) or TBT + U0126 (5 μM) for 24 h. ERK phosphorylation was analyzed by western blotting using anti-phospho-ERK antibodies. Cropped blots were shown and the full-length blots were indicated in Supplementary Fig. 3. (b) Relative densities of bands were quantified using ImageJ software. Relative changes in expression were determined by normalization to total ERK protein level. (c) At day 2 after ectodermal (neural) induction with TBT or TBT + U0126, the expression of OTX2 gene was analyzed using real-time PCR. (d) At day 1 after mesodermal induction with TBT or TBT + U0126, the expression of BRACHYURY gene was analyzed using real-time PCR. (e) At day 4 after endoderm induction with TBT or TBT + U0126, the expression of SOX17 gene was analyzed using real-time PCR. Each bar represents the mean ± SD from three independent experiments. *P < 0.05. are relevant to human exposure levels. Moreover, the inhibitory effect of TBT on ectodermal induction was also supported by the knockdown of Mfn1, which mimics the effect of TBT on Mfn1 11 . In contrast, TBT did not affect mesodermal or endodermal induction of iPSCs. A previous report has shown that the ablation of Mfn1 in the embryonic mouse heart, which is derived from embryonic mesodermal germ layer, arrests its development 39 . Therefore, TBT may affect the differentiation processes after the mesodermal germ layer, which in turn may potentially lead to cardiomyocyte. Further studies are needed to elucidate whether TBT actions contain the stage selectivity in the differentiation into mesodermal-or endodermal-derived cells.
We demonstrated that TBT negatively affected the levels of OTX2 in the neural differentiation (Fig. 1), which acts as a transcriptional regulator during forebrain development in vertebrates 40 . In addition, the targeted disruption of PAX6 in rodents led to the loss of anterior neural tissues 41 , suggesting a central role for PAX6 in forebrain development. TBT exposure causes decreases in forebrain weight with a reduction of synaptogenic markers in the developing rat brain 42 . These TBT-induced defects of the forebrain may be caused by transcriptional silencing of anterior neural markers, such as OTX2 and PAX6, during early neurogenesis. As a neural progenitor marker, Nestin was downregulated during the neural differentiation of iPSCs exposed to TBT, and further studies in NPCs are required to determine whether TBT affects their differentiation into specific neuronal subtypes, e.g., glutamatergic, GABAergic, or cholinergic.
We further demonstrated that the negative effect of TBT on neurogenesis was probably mediated by Mfn1 degradation, followed by ERK phosphorylation (Fig. 4). A previous report indicates that Mfn1 directly binds Ras and Raf, thereby inhibiting Ras-Raf-ERK signaling, as determined using biochemical analysis 33,34 . Thus, Mfn1 reduction by TBT or shRNA could reverse the inhibition of ERK signaling. The mobilization of Ca 2+ from intracellular stores including the mitochondria was reported to result in the phosphorylation of MAPKs, the authors stating that the process was suppressed by the chelation of intracellular Ca 2+ in human T lymphoblastoid cells 43 . The established mitochondrial uptake of any Ca 2+ that accumulated in the cytosol was shown to be dependent on MMP 44 with the mitochondrial dysfunction induced by TBT exposure probably causing an overload of Ca 2+ , which resulted in ERK activation. Moreover, the ERK signaling was reported to inhibit neural induction by silencing PAX6 through upregulation of the stemness factors NANOG/OCT4 and downregulation of OTX2 36 . Thus, TBT-induced ERK phosphorylation could downregulate PAX6 expression by the suppression of OTX2 in mediating its developmental neurotoxicity in iPSCs. Moreover, ERK signaling evoked by TBT may affect the expression of other PAX6 regulatory factors such as NANOG/OCT4. In future studies, it would be expedient for us to further investigate the mechanisms underlying the TBT-induced negative regulation of neural induction via ERK.
In summary, our results demonstrate a novel mechanism underlying the cytotoxicity and neurodevelopmental toxicity of TBT in iPSCs. There seems to be line-to-line differences in hiPSCs 45 . Although we used 253G1 cells in the present studies, line-to-line differences of hiPSCs in the neural differentiation should be investigated in future. Recently, significant progress has been made in the induction of differentiation of pluripotent stem cells into a variety of cell types 46 . Further studies are needed to evaluate the developmental effects of TBT on various types of iPSC-derived cells. Moreover, we showed that the TBT-induced cytotoxicity was likely caused by Mfn1-mediated mitochondrial dysfunction, which is also involved in the toxicities of other endocrine disruptive chemicals 47,48 , such as chlorpyrifos 32 and silver nanoparticles 49 . Thus, further investigation of mitochondrial functions influenced by Mfn1 would be an important next step to examine the mechanisms underlying the toxicities induced by chemicals.
Cell culture. We used the human iPSC line 253G1 (Riken BRC Cell Bank, Tsukuba, Ibaraki, Japan), which was established through the retroviral transduction of Oct3/4, Sox2, and Klf4 into adult human dermal fibroblasts 50 . The cell culture was performed as previously reported 11 . Briefly, the cells were cultured under feeder-free conditions using human embryonic stem cell (ESC)-qualified Matrigel (BD Biosciences, San Jose, CA, USA) and TeSR-E8 medium (Stemcell Technologies, Vancouver, BC, Canada) at 37 °C in an atmosphere containing 5% CO 2 . For passaging, the iPSC colonies were dissociated into single cells using Accumax (Innovative Cell Technologies, San Diego, CA, USA) and cultured in TeSR-E8 medium supplemented with the ROCK-inhibitor Y-27632 (10 μM) for the first two days.
Neural differentiation. The neuronal lineages derived from the ectodermal germ layer were induced using the dual SMAD inhibition protocol as previously described with modifications 28 . Briefly, iPSC colonies were dissociated into single cells with Accumax. The cells were seeded at a density of 7 × 10 4 cells/cm 2 in the TeSR-E8 medium on Matrigel-coated plates to reach a near confluent level within 2 days after seeding. The initial ectoderm differentiation was performed using a knockout serum replacement (KSR) medium [Knockout DMEM (Thermo Fisher Scientific) containing KSR (Thermo Fisher Scientific), L-glutamine (Thermo Fisher Scientific), non-essential amino acids (NEAA; Thermo Fisher Scientific), 2-ME, PS] containing SB431542 (TGFβ inhibitor, 10 μM) and LDN193189 (BMP inhibitor, 1 μg/ml). After 4 days, N2 medium [Neurobasal containing N2 (Thermo Fisher Scientific), B27 (minus vitamin A, Thermo Fisher Scientific), GlutaMAX (Thermo Fisher Scientific), PS] was added to the KSR medium with LDN193189 every 2 days.

Mesoderm Induction.
For the induction of the mesodermal germ layer, a cardiomyocyte differentiation protocol was used as previously described with modifications 24 . Briefly, iPSC colonies were dissociated into single cells using Accumax. The cells were seeded at a density of 6 × 10 4 cells/cm 2 in the TeSR-E8 medium on Matrigel-coated plates to achieve an approximately 80-90% confluence within 2 days after seeding. Then, the medium was replaced with RPMI1640 (Nacalai Tesque, Kyoto, Japan) containing B27 (minus insulin, Thermo Fisher Scientific), CHIR99021 (a GSK3 inhibitor, 10 μM), and PS for 1 day.
Endoderm Induction. For the induction of endodermal germ layer, hepatic differentiation protocol was used as previously described with modifications 25 . Briefly, iPSC colonies were dissociated into single cells with Accumax. The cells were seeded at a density of 5 × 10 4 cells/cm 2 in the TeSR-E8 medium on Matrigel-coated plates to reach an approximately 70% confluence level within 2 days after seeding. Then, the medium was replaced with RPMI1640 containing B27 (Thermo Fisher Scientific), activinA (100 ng/ml; R&D Systems, Minneapolis, MN, USA), Wnt3a (50 ng/ml; R&D Systems), and PS. The next day, NaB (0.5 mM) was added to the culture medium, followed by a 1-day incubation period and then the medium was replaced with RPMI1640 containing B27, activinA, Wnt3a, and PS for an additional 2 days.
Immunocytochemistry. Cell staining was performed as previously described 51 . Briefly, cells were cultured on glass coverslips, fixed in 4% paraformaldehyde in phosphate-buffered (PBS, pH 7.4) for 15 min at room temperature, and then incubated with anti-PAX6 polyclonal antibodies (1:100, Biolegend, San Diego, CA, USA) overnight at 4 °C. Then, the cells were incubated with Alexa 555-conjugated secondary antibodies (1:200, Thermo Fisher Scientific) for 1 h at room temperature. Nuclei were counterstained with DAPI (Nacalai Tesque). Fluorescence images were obtained using a BIOREVO BZ-9000 fluorescent microscope (Keyence, Osaka, Japan). PAX6 positive nuclei were counted using the ImageJ software (NIH, Bethesda, MD, USA). Gene Knockdown Using Short Hairpin RNA (shRNA). Knockdown experiments were performed using Mfn1 shRNA lentiviruses from Sigma-Aldrich (MISSION shRNA), as previously reported 8 . A scrambled hairpin sequence was used as the negative control. Briefly, the cells were infected with the viruses at a multiplicity of infection (moi) of 1 in the presence of 8 μg/mL hexadimethrine bromide (Sigma-Aldrich) for 24 h. After the medium exchange, the cells were selected with 1 μg/mL puromycin for 24 h and cultured for an additional 72 h prior to the functional analyses.

Statistical analysis.
All the data are presented as means ± standard deviation (SD) from three independent experiments and the Student's t-test was used to analyze the data in Figs 1, 2c, 3a,c and S1. An analysis of variance (ANOVA) followed by the Bonferroni posthoc test was used to analyze the data in Figs 2b, 3b and 4. P-values < 0.05 were considered statistically significant.