Screening for putative epimutagens in neural stem cells derived from human induced pluripotent stem cells

Abstract

The prevalence of pediatric neurological disorders has increased in recent years and has emerged as a major social concern. While genetic factors play a role, environmental factors may also contribute to these disorders, particularly exposure to harmful chemicals in the fetal environment during pregnancy. However, the concentrations of these chemicals in the fetal environment are extremely low compared with those in experimental conditions, and their influence on the development of neurological disorders after birth remains unclear. Epigenetic regulation plays a crucial role in gene expression and/or chromatin formation, ensuring normal cellular development and differentiation. The relationship between impaired epigenetic regulation and neurological disorders has been previously reported. Therefore, this study aimed to identify chemicals that disrupt epigenetic systems at detectable concentrations in pregnant mothers' serum using a human neural differentiation model. Among the 11 chemicals, including pesticides and heavy metals, exposure to octachlorodipropyl ether (S-421) altered heterochromatin formation in neural stem cells at concentrations found in maternal and umbilical cord blood serum. Additionally, exposure to several chemicals, including S-421, affected neural differentiation, leading to excessive neural fiber growth. Furthermore, S-421 exposure induced DNA hypermethylation of the neural-related genes Cadherin 2 (CDH2) and Sry-related HMG-Box gene 10 (SOX10) in neural stem cells. These findings suggest that S-421, which is present in the fetal environment, functions as an epimutagen that disrupts the epigenetic system within biologically relevant exposure levels. The established epimutagenic screening is expected to provide new insights into the relationship between prenatal chemical exposure and postnatal childhood neurological disorders.

INTRODUCTION

The global prevalence of childhood neurological disorders has risen significantly, posing a major social concern. These disorders are believed to result from a combination of genetic and environmental factors. Concerns of fetal exposure to certain environmental chemicals exist, which may increase the risk of neurological disorders. Epidemiological studies indicate an association between prenatal exposure to chemicals, including pesticides and a higher risk of neurological and behavioral disorders in children (Llop et al., 2013; Rauh et al., 2011; Harari et al., 2010; Eskenazi et al., 2007). However, the concentrations of chemicals detected in the maternal-fetal environment are extremely low (approximately 0.1–100 ppb levels) compared to the toxic levels reported in pharmacological studies. Consequently, chemicals in the fetal environment may contribute to neurological disorders, although the precise molecular mechanisms underlying their role in the pathogenesis remain unknown. The early diagnosis and prompt initiation of treatment are critical for the management of pediatric neurological disorders. Therefore, evaluation of the effects of chemicals on neural development and differentiation could facilitate the early detection of neurological disorders, taking into account the actual chemical exposure in pregnant mothers.

Epigenetic systems, which consist of DNA methylation and histone modifications, play a key role in regulating cell type-specific gene expression (Ikegami et al., 2009; Golob et al., 2008; Lieb et al., 2006; Shiota, 2004). During cell differentiation, epigenetic state undergoes dynamic changes, influencing the direction of differentiation. Furthermore, epigenetic modifications affect chromatin structure, where chromatin condensation in heterochromatin regions is associated with DNA methylation and repressive histone modifications (Cedar and Bergman, 2009). Disruptions in epigenetic regulation are associated with alterations in gene expression and chromatin structure, contributing to various diseases, including cancer and neurological disorders (Fujita and Yamashita, 2018; Kramer and van Bokhoven, 2009; Ushijima and Okochi-Takada, 2005). Thus, epigenetic system is considered a crucial regulator of normal cell differentiation and development.

Chemicals that disrupt epigenetic systems are referred to as epimutagens (Holliday and Ho, 2002), and certain environmental chemicals act as epimutagens. Diethylphosphate (DEP), cotinine, octachlorodipropyl ether (S-421), mercury, and selenium detected in maternal and umbilical cord blood samples during pregnancy function as epimutagens in mouse and human pluripotent stem cells at biologically relevant exposure concentrations (Arai et al., 2015; Arai et al., 2011). This suggests that certain chemicals may disrupt the epigenetic status within the biologically exposure range, adversely affecting cell differentiation. Thus, the detection of epimutagens may provide new insights into the pathogenesis of potential chemically induced diseases, including neurological diseases.

Chemical toxicity has primarily been evaluated using mice, rats, or cancer cell lines. However, sensitivity to chemicals varies widely across animal species and cell types (Boncler et al., 2019; Wilson et al., 2014; Chen et al., 2020). This highlights the advantage of using a human cell-derived neural differentiation system to evaluate chemicals that may be relevant to human neurological disorders. Therefore, this study aims to investigate the effects of chemicals detected in the fetal environment, within biologically relevant exposure ranges, on neuronal epigenetic systems using a developed epimutagen screening system based on induced pluripotent stem cell (iPSC)-derived human neuronal differentiation.

MATERIALS AND METHODS

Human induced pluripotent stem cell culture and neural stem cell differentiation

The human induced pluripotent stem cells (RNA-MRC-iPS-6) used in this study have been previously described (Sekiya et al., 2022). Human iPSCs were maintained on iMatrix-511 (Nippi, Tokyo, Japan) in StemFit AK02 (Ajinomoto, Kawasaki, Japan). Handling of human cells adhered to the ethical standards of the Helsinki declaration.

Human iPSC-derived neural stem cells (NSCs) were cultured in PSC Neural Induction Medium (Thermo Fisher Scientific, Waltham, MA, USA) according to the protocol of the manufacturer. Briefly, iPSCs were dissociated into single cells using TrypLE Select (Thermo Fisher Scientific) and plated on iMatrix-511 in StemFit AK02. The next day, the culture medium was replaced with Neurobasal Medium (Thermo Fisher Scientific) supplemented with a neural induction supplement (Thermo Fisher Scientific) under 5% oxygen tension. Following 7 days of neural induction, neural expansion was performed. Cells were again dissociated into single cells using TrypLE Select and cultured on iMatrix-511 for 5 days in equal volumes of Neurobasal Medium and Advanced DMEM/F12 (Thermo Fisher Scientific) supplemented with a neural induction supplement.

Chemical exposure to neural stem cells

To assess the effects of chemicals on neuronal cell development and differentiation, 11 chemicals were selected, including pesticides, tobacco, and heavy metals, which have been reported to be associated with neurological disorders such as autism spectrum disorder in previous epidemiological studies (Carter and Blizard, 2016; Harari et al., 2010; Eskenazi et al., 2007). NSC exposure to the chemicals began the day after neural expansion. NSCs were cultured with chemicals (diethylphosphate (DEP), 3,5,6-trichloro-2-pyridinol (TCP), octachlorodipropyl ether (S-421), nicotine, cotinine, di(2-ethylhexyl)phthalate (DEHP), mono(2-ethylhexyl) phthalate (MEHP), mercury (Hg), selenium (Se), cadmium (Cd), and lead (Pb)) at concentrations equivalent to the serum levels or 10-fold higher levels determined in a previous study (Arai et al., 2011). This was based on the concentrations in cord blood and/or the serum of pregnant mothers (Table 1). Hg (Kanto Chemical, Tokyo, Japan), Se (Kanto Chemical), Cd (Fujifilm Wako Pure Chemical, Osaka, Japan), and Pb (Fujifilm Wako Pure Chemical) were diluted with nitric acid (HNO₃, Fujifilm Wako Pure Chemical). DEP (Fujifilm Wako Pure Chemical), TCP (Fujifilm Wako Pure Chemical), S-421 (Sigma-Aldrich), nicotine (Sigma-Aldrich), cotinine (Sigma-Aldrich), DEHP (Kanto Chemical), and MEHP (Kanto Chemical) were diluted with ethanol (EtOH, Fujifilm Wako Pure Chemical). The final concentrations of HNO₃ and EtOH were 0.0063% and 0.1%, respectively. Furthermore, 5-aza-2-deoxycytidine (5-aza-dC, Sigma-Aldrich), a well-known epimutagen, was used as a positive control in the epimutagen screening.

Table 1. Chemicals used in this study.

Chemical	a Exposure concentration (ppb) 1×	a Exposure concentration (ppb) 10×	Diluting solvent
DEP	0.1	1.0	EtOH
TCP	0.1	1.0	EtOH
S-421	0.01	0.1	EtOH
Nicotine	100	1000	EtOH
Cotinine	100	1000	EtOH
DEHP	1.2	12.0	EtOH
MEHP	5.2	52.0	EtOH
Hg	1.0	10.0	HNO3
Se	100	1000	HNO3
Cd	0.1	1.0	HNO3
Pb	1.0	10.0	HNO3

^a, determined as previous study (Arai et al., 2011)

1×, serum level concentration detected in pregnant mothers/cord blood serum

10×, ten-fold higher level than pregnant mothers/cord blood serum concentration

Neural differentiation

For neural differentiation, NSCs were dissociated into single cells using TrypLE Select and cultured on Matrigel (Corning, Corning, NY, USA) in neurobasal medium supplemented with B27 supplement (Thermo Fisher Scientific), 10 ng/mL brain-derived neurotrophic factor (Fujifilm Wako Pure Chemical), 10 ng/mL glial cell-derived neurotrophic factor (Fujifilm Wako Pure Chemical), 1% GlutaMax (Thermo Fisher Scientific), 1% MEM Non-essential amino acids (Thermo Fisher Scientific), and 200 µM of ascorbic acid (Fujifilm Wako Pure Chemical) for 14 days under 5% oxygen tension. The culture medium was replaced every 3 days. All experiments were conducted independently at least twice.

Quantitative gene expression analysis

Total RNA was extracted from NSCs employing the RNeasy Plus Mini Kit (Qiagen, Hilden, Germany). First-strand cDNA synthesis was performed utilizing the ReverTra Ace (TOYOBO, Osaka, Japan) with random hexamers. Quantitative real-time polymerase chain reaction (qPCR) was conducted using AriaMx (Agilent Technologies, Santa Clara, CA, USA) with SYBR Green PCR master mix (Applied Biosystems, Woburn, MA, USA). Data were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression levels, and relative expression was determined using the 2(-ΔΔ CT) method (Livak and Schmittgen, 2001). Table S1 lists the PCR primer sequences used in this study. All experiments were performed in triplicate, and the results are presented as mean values with standard deviations.

Slide preparation and image acquisition for heterochromatin analysis

Slide preparation was performed as previously described (Arai et al., 2011). Briefly, NSCs exposed to chemicals or their solvents were fixed with a fixative solution consisting of EtOH (A) and acetic acid (B) (A/B:3/1, v/v). Following centrifugation to remove the supernatant, the cells were resuspended in the fixative solution, and interphase nuclei were then spread on a glass slide. For heterochromatin analysis, nuclei were stained with 0.5 μg/mL of 4′,6-diamino-2-phenylindole (DAPI) (Dojindo, Kumamoto, Japan) for 5 min. The samples were mounted using an antifade medium, PermaFluor Aqueous Mounting Medium (Thermo Fisher Scientific). All reactions were performed at a temperature of 20−25°C. Fluorescent images of DAPI staining were acquired using a fluorescent microscope (BZ-9000; KEYENCE, Osaka, Japan). Following deconvolution with BZ-9000 software, the number of DAPI signals in the nuclus was determined using ImageJ software by the National Institute of Health (http://rsb.info.nif.gov/ij/). Acquired RGB images were converted to 8-bit grayscale using linear scaling, setting intensity values from 0 to 255. The intensity thresholds for DAPI images were automatically determined within the range of 28 to 30 using the ImageJ program and were shown in red pseudocolor. Finally, the number of signals per nucleus (n = 195–465) was determined using ImageJ software.

Immunofluorescence

NSCs and neurons differentiated from NSCs were fixed with 4% paraformaldehyde for 10 min. Following permeabilization with 0.2% Triton X-100 for 5 min, the samples were incubated with a blocking buffer containing 1% bovine serum albumin and 0.01% Tween-20 in phosphate-buffered saline for 60 min. The samples were then incubated for 60 min with anti-NESTIN (NES) rabbit polyclonal antibodies (19483-1-AP; Proteintech, Rosemont, IL, USA) for NSCs or anti-TUBB3 mouse monoclonal antibodies (MMS-435P; BioLegend, Princeton, NJ, USA) for neurons, both diluted 1:200 in blocking buffer. The samples were then washed thrice with phosphate-buffered saline containing 0.05% Tween-20. After incubation for 60 min with fluorescent secondary antibodies (Alexa Fluor 594 goat anti-rabbit immunoglobulin G for anti-NES antibody or Alexa Fluor 488 goat anti-mouse immunoglobulin G for anti-TUBB3 antibody, Thermo Fisher Scientific) diluted 1:200 in blocking buffer, the samples were washed again and counterstained with 0.2 μg/mL DAPI. All reactions were performed at a temperature of 20−25°C. Immunofluorescence images were acquired using BZ-9000, and images of TUBB3-positive neuronal fiber areas in nine visual fields were quantified using ImageJ software. Briefly, RGB images were converted to 8-bit grayscale (0−255 gradation). The intensity thresholds for TUBB3 images were set within 20 to 42 using the automatic threshold setting of the ImageJ program. The neuronal fiber area was then measured using ImageJ software. All experiments were independently performed at least twice.

Extraction of neural stem cell-specific hypomethylated sites from genome-wide DNA methylation data

To extract NSC-specific hypomethylated gene sites associated with neuronal cell development and differentiation, genome-wide DNA methylation data were analyzed using the Illumina Infinium Human MethylationEPIC BeadChip (Illumina Inc., CA, USA). DNA methylation data for two human somatic cell lines (Edom22 and IMR90) were obtained from a previous study (Nishino et al., 2021) and they are available in the Gene Expression Omnibus database (https://www.ncbi.nim.gov/geo/) under the accession number: GSE141521. In this study, genome-wide DNA methylation data for two NSC lines (NSC_Edom and NSC_IMR) derived from iPSCs established from Edom22 and IMR90 were additionally obtained using the Human MethylationEPIC BeadChip (GSE328809). These two NSC lines have been confirmed to be positive for Paired box protein-6 (PAX6), a neural stem cell marker gene in previous study (Sekiya et al., 2022). In the BeadChip data, methylated and unmethylated signals were quantified as β-value, representing DNA methylation levels on a scale from 0 (completely unmethylated) to 1 (completely methylated). Comparing the DNA methylation data of these four samples, NSC-specific hypomethylated sites were determined.

DNA methylation analysis by COBRA assay

DNA methylation analysis was conducted using a bisulfite reaction of purified DNA with the EZ DNA Methylation-Gold Kit (Zymo Research, Irvine, CA, USA). The COBRA assay was conducted as previously described (Xiong and Laird, 1997). Briefly, bisulfite-treated DNA was amplified with BioTaq HS DNA polymerase (BIOLINE, London, UK) using specific PCR primers (Table S2). Following amplification, the PCR fragments were digested with HpyCH4IV (New England Biolabs, Ipswich, MA, USA) or TaqI-v2 (New England Biolabs) for 3 hr and then electrophoresed. Digested PCR fragments represent methylated DNA, while undigested fragments indicate unmethylated DNA. The luminescence intensity of DNA fragments was measured using ImageJ software, and DNA methylation levels were calculated using the following formula: DNA methylation levels = luminescence intensity of digested DNA fragments/total luminescence intensity of undigested and digested DNA fragments.

Statistical analysis

Statistical comparisons of the number of DAPI signals and neuronal fiber area detected via anti-TUBB3 antibodies were performed using the Wilcoxon test, while those of gene expression and COBRA assay data were analyzed using the Student’s t-test.

RESULTS

Effects of chemicals on heterochromatin formation in human neural stem cells

To investigate the effects of chemical exposure on epigenetic systems during human neural differentiation, we first assessed whether chemicals at serum-detectable concentrations influenced the cytotoxicity and differentiation ability into neural stem cells. Human iPSC-derived NSCs were exposed to 11 chemicals, including pesticides and heavy metals (DEP, TCP, S-421, nicotine, cotinine, DEHP, MEHP, Hg, Se, Cd, and Pb), based on previously identified maternal and/or cord blood serum concentrations (Fig. 1A). Following 4 days of exposure, the number of NSCs in the chemical-exposed groups were counted, and no significant variations were observed compared to the control groups treated with EtOH- or HNO₃ (Fig. 1B). These findings suggest that the serum concentrations of the tested chemicals were not cytotoxic. Subsequent immunostaining with an antibody against the neural stem cell marker NES demonstrated that all NSCs exposed to the 11 chemicals exhibited NES positive (Fig. 1C). Furthermore, no variations were observed in the gene expression levels of neural stem cell markers NES, PAX6, and Sry-related HMG-Box gene 2 (SOX2) between the chemical-treated and control EtOH- or HNO₃-exposed NSCs (Fig. 1D). These findings indicate that exposure to maternal serum concentrations of the 11 chemicals neither induces cell death nor inhibits differentiation into NSCs.

Fig. 1

Chemical exposure to human iPSC-derived NSCs. (A) Schematic diagram of NSC culture exposed to 11 chemicals (DEP, TCP, S-421, nicotine, cotinine, DEHP, MEHP, Hg, Se, Cd, and Pb) at serum-level concentration (1×). (B) Verification of chemical-induced cytotoxicity. NSCs were exposed to each chemical for 4 days, and cell numbers were measured. Data are presented as mean ± SE based on ratios relative to NSCs treated with solvent control, EtOH-, or HNO₃. (C) Expression of stem cell marker NES in chemical-exposed NSCs. After 4 days of exposure to chemicals or their solvents, NSCs were immunostained with an anti-NES antibody. Red and blue indicate anti-NES antibody and nuclear staining with DAPI, respectively. Scale bar = 250 µm. (D) Expression of neural stem cell marker genes, NES, PAX6, and SOX2, in chemical-exposed NSCs, as analyzed via qPCR. Relative expression levels of each gene are presented as mean ± SD based on GAPDH expression.

We then investigated the effects of these chemicals on the epigenetic systems of NSCs. The structure of heterochromatin, formed by DNA methylation and histone modifications, changes during cellular development and differentiation (Allshire and Madhani, 2018). Thus, heterochromatin structure is regarded as an indicator of the cellular state. Furthermore, because heterochromatin is easily visualized through DAPI nuclear staining, quantifying heterochromatin signals enables us to identify chemicals that affect the epigenetic state of mouse/human pluripotent stem cells (Arai et al., 2015; Arai et al., 2011). In this study, we employed a previously established epimutagen screening system for human iPSC-derived NSCs.

Human NSCs were exposed to maternal/cord blood serum concentrations (1×) or ten times these serum concentrations (10×) of the 11 chemicals for 4 days (Fig. 2A). First, we exposed NSCs to 5-aza-dC, a DNA methyltransferase inhibitor known as an epimutagen (Holliday and Ho, 2002). Analysis of DAPI-stained images revealed that exposure to 1 µM 5-aza-dC significantly increased the number of heterochromatin signals compared to that of the control (Fig. 2B). In contrast, 0.1% EtOH and 0.0063% HNO₃, used as solvents to dilute the chemicals, did not affect the number of heterochromatin signals compared to the untreated group (Fig. 2C). These findings suggest that epimutagen screening using heterochromatin as an indicator is applicable to human iPSC-derived NSCs.

Fig. 2

Epimutagen screening in human NSCs using heterochromatin formation. (A) Schematic diagram of NSC culture exposed to chemicals at serum-level concentration or 10-fold higher concentration. (B) Effects of 5-aza-dC treatment on heterochromatin formation. After treatment with 1 µM 5-aza-dC, nuclear staining with DAPI was performed, and deconvolution was carried out with the microscope software. Subsequently, strongly stained heterochromatin with DAPI was visualized as pseudocolor (red) based on a threshold using ImageJ software (left panel). The number of heterochromatin signals was quantified using ImageJ software and displayed as a boxplot (right panel). Scale bar = 250 µm. * p < 0.01. (C) Number of heterochromatin signals in NSCs exposed to the chemical diluting solvents EtOH or HNO₃. Non-treat., untreated group. (D) Effects of chemicals on heterochromatin formation. The number of heterochromatin signals was quantified using ImageJ software and was presented as a boxplot. −, solvent-treated control; 1×, serum-level concentration; 10×, 10-fold higher concentration than serum level; * p < 0.01.

Next, we analyzed the effects of the 11 chemicals on heterochromatin formation (Fig. 2D). Exposure to chemicals at serum concentrations resulted in a similar number of heterochromatin signals to that of control solvent exposure, except for S-421, suggesting that chemical exposure did not affect heterochromatin formation. Conversely, S-421 exposure significantly increased the number of heterochromatin signals compared to that of the control. This trend was also observed when NSCs were exposed to concentrations ten times higher than serum levels (Fig. 2D), indicating that S-421 functions as an epimutagen that alters heterochromatin formation in human neural stem cells within biologically relevant exposure ranges. Furthermore, four other chemicals (nicotine, cotinine, DEHP, and Hg) significantly increased the number of heterochromatin signals when NSCs were exposed to concentrations 10 times higher than serum concentrations (Fig. 2D). These findings suggest that effects of epimutagen increases in a chemical concentration-dependent manner. Heterochromatin analysis could not be performed in Se-exposed NSCs owing to cell death occurring at 10-fold higher exposure than serum concentration. A similar trend is observed in mouse embryonic stem cells (Arai et al., 2011).

Sustained effects of S-421 on heterochromatin formation in neural stem cells

Exposure to serum concentrations of S-421 altered heterochromatin formation in NSCs. To investigate the sustained effect of S-421 on heterochromatin formation, we analyzed the heterochromatin state in NSCs 4 days after removing S-421 at serum concentration (Fig. 3A). The number of heterochromatin signals in NSCs after S-421 removal was significantly higher than in NSCs after EtOH removal. This change in heterochromatin signal counts is consistent with the S-421 exposure results in Fig. 2D, suggesting that the effects of S-421 on heterochromatin formation is not transient but may persist even after removal of the chemical.

Fig. 3

Verification of recovery of heterochromatin changes by chemical removal. (A) Schematic diagram of NSC culture. NSCs were exposed to serum concentrations of S-421 or the solvent control EtOH for 4 days, after which NSCs were passaged without chemical or EtOH and cultured for another 4 days. (B) Heterochromatin formation in NSCs after S-421 removal. The number of heterochromatin signals was quantified based on the DAPI images (left panel), and the results are presented as a boxplot (right panel). Scale bar = 250 µm. * p < 0.01.

Chemical exposure promotes neuronal fiber overgrowth

Several chemicals altered heterochromatin formation in human NSCs at concentrations up to 10-fold higher than serum concentrations. Furthermore, heterochromatin changes induced by exposure to serum concentrations of S-421 tended to persist even after removal of the chemical. We then investigated whether chemical exposure to NSCs affected their subsequent neural differentiation. NSCs were exposed to 11 chemicals at serum concentrations or their solvents, after which the chemicals/solvents were removed to induce neural differentiation (Fig. 4A). Fourteen days after induction, neuronal fibers were visualized via immunostaining with an antibody against TUBB3, a neuron-specific marker (Fig. 4B). Measurements using ImageJ software demonstrated that exposure to DEP, TCP, S-421, nicotine, cotinine, and DEHP significantly increased neuronal fiber area (Fig. 4C). In contrast, no significant difference in neuronal fiber area was observed between the solvent (EtOH or HNO₃)-treated and untreated groups (Fig. S1). These findings suggest that NSC exposure to these chemicals affects subsequent neural differentiation and promotes fiber overgrowth. Fig. 4D provides a summary of the effects of chemical exposure on heterochromatin formation and neural differentiation. Six chemicals (DEP, TCP, S-421, nicotine, cotinine, and DEHP) affected neural differentiation at serum-level concentrations, with S-421 also affecting heterochromatin formation. Furthermore, among the six chemicals that affected neural differentiation, nicotine, cotinine, and DEHP induced changes in heterochromatin formation upon exposure to concentrations 10-fold higher than serum levels.

Fig. 4

Effects of chemicals on neural differentiation. (A) Schematic diagram illustrating the neural differentiation process of NSCs. NSCs were exposed to serum concentrations of 11 chemicals for 4 days, after which the chemicals were removed, and neural differentiation was induced. (B) Images of neuronal cells after 14 days of differentiation. Neuronal fibers were stained with an anti-TUBB3 antibody. Scale bar = 250 µm. (C) Measurement of neuronal fiber area of NSC-derived neurons. The neuronal fiber area was measured using ImageJ software based on images acquired with the anti-TUBB3 antibody and are presented as mean ± SE. * p < 0.01. (D) Summary of the adverse effects of chemicals on heterochromatin formation and neural differentiation. *, Heterochromatin analysis could not be performed in Se-exposed NSCs owing to cell death occurring at 10-fold higher exposure than serum concentration. 1×, serum-level concentration; 10×, 10-fold higher concentration than serum level.

S-421-induced DNA methylation changes in neural-related genes

Exposure to serum concentrations of several chemicals influenced neural differentiation, with S-421 also altering heterochromatin formation. Following that, we examined the effect of chemical exposure on DNA methylation in genic regions. Typically, DNA methylation influences cellular properties and functions. Genes critical for maintaining the properties of iPSCs, including OCT3/4 and NANOG, are specifically hypomethylated in human iPSCs (Nishino et al., 2011). Similarly, neural-related genes that are hypomethylated in NSCs are believed to play a key role in their properties and subsequent neural differentiation. Therefore, we first attempted to extract neural-related genes that are specifically hypomethylated in neural stem cells using genome-wide DNA methylation data from the Human Methylation EPIC BeadChip.

Based on data from two NSCs derived from different human iPSC lines (NSC_Edom and NSC_IMR) and their corresponding donor somatic cells (Edom22 and IMR90), we attempted to extract NSC-specific hypomethylated sites (NSC-hypo-sites) with methylation levels > 0.5 lower than those of somatic cells (Fig. 5A). From approximately 800,000 sites, we extracted NSC-hypo-sites located near genes, identifying 4,493 and 3,202 sites in NSC_Edom and NSC_IMR, respectively (Fig. 5B). Among these, 1,912 NSC-hypo-sites were obtained, showing hypomethylation in both NSCs, and 436 sites corresponding to 285 genes were located upstream of gene regions. These DNA methylation levels of 436 sites are shown in a heatmap (Fig. 5C). Furthermore, genes associated with neural development and differentiation were identified through gene ontology analysis using DAVID, a functional annotation tool provided by the National Institutes of Health (https://davidbioinformatics.nih.gov). This analysis yielded 81 sites corresponding to 48 genes, which were obtained as neural-related NSC-hypo-sites (Fig. 5C). Of these sites, 14 analyzable genes were selected for DNA methylation analysis using COBRA assay in NSCs exposed to DEP, TCP, S-421, nicotine, cotinine, and DEHP, as these chemicals were associated with neuronal fiber overgrowth. The COBRA assay was selected for DNA methylation analysis owing to its simplicity, speed, and suitability for multiple sample analysis in this study. DNA methylation occurs at the cytosine of CpG sequences, and the COBRA assay evaluates DNA methylation level by cleaving PCR-amplified DNA at CpG sites recognized by restriction enzymes (Xiong and Laird, 1997). Therefore, CpG sites at or near the NSC hypo-sites that could be recognized by restriction enzymes were selected for analysis.

Fig. 5

Effects of S-421 exposure on DNA methylation of genic regions in NSCs. (A) Schematic diagram of neural stem cell-specific hypomethylated sites (NSC-hypo-sites) selection. (B) After excluding the data of detection p-value (> 0.05) and missing values (NaN) using Human MethylationEPIC BeadChip data, NSC-hypo-sites were identified with DNA methylation levels > 0.5 lower in two neural stem cell lines derived from different human iPSC lines (NSC_Edom and NSC_IMR) compared to somatic cells (Edom22 and IMR90). Furthermore, focusing on the upstream gene regions, 436 sites corresponding to 285 genes were extracted. (C) DNA methylation data of 436 NSC-hypo-sites. A heatmap displays methylation data from NSCs and somatic cells. Gene ontology analysis identified 81 sites (48 genes) associated with neural-related NSC-hypo-sites, of which 14 sites were selected for DNA methylation analysis using COBRA assay. (D) DNA methylation analysis in NSCs exposed to S-421. DNA methylation status of CDH2 and SOX10 in NSCs exposed to serum concentrations of S-421 or diluting solvent (Control) for 4 days was analyzed using COBRA assay and confirmed by electrophoresis. White and black arrows indicate unmethylated and methylated bands, respectively. (E) DNA methylation levels were calculated based on the intensity of methylated and unmethylated bands and are presented as mean ± SE. * p < 0.05

In NSCs exposed to serum concentrations of five chemicals (DEP, TCP, nicotine, cotinine, and DEHP) other than S-421, DNA methylation levels of the analyzed genes were comparable to those in control NSCs exposed to EtOH (Table 2). However, DNA methylation status of Cadherin 2 (CDH2) and Sry-related HMG-Box gene 10 (SOX10) differed between S-421-exposed and control NSCs (Fig. 5D). The DNA methylation levels of these gene sites were significantly increased in S-421-exposed NSCs (Fig. 5E). These findings suggest that S-421, which alters heterochromatin formation, also influences the DNA methylation status of neural-related gene when exposed to serum-level concentrations.

Table 2. DNA methylation level analyzed by COBRA assay in chemical (1×) exposed NSCs.

Gene	DEP	TCP	S-421	Nicotine	Cotinine	DEHP	Solvent a	Enzyme b
APBB1	0	0	0	0	0	0	0	H
CDH2	0.63	0.62	0.72*	0.63	0.63	0.62	0.63	H
CNTNAP1	0	0	0	0	0	0	0	T
FGF20	0	0	0	0	0	0	0	H
SLC4A8	0	0	0.04	0.01	0	0	0.02	H
NODAL	0	0	0	0	0	0	0	H
PAX6	0	0	0	0	0	0	0	H
PRDM8	0	0	0	0	0	0	0	H
RHOH	0.35	0.32	0.34	0.32	0.30	0.31	0.32	H
SALL4	0	0	0	0	0	0	0	H
SATB2	0	0	0	0	0	0	0	H
SOX6	0.04	0.04	0.06	0.07	0.04	0.06	0.05	H
SOX10	0.25	0.24	0.31*	0.23	0.25	0.22	0.25	T
WNT8B	0.16	0.16	0.17	0.14	0.15	0.14	0.14	H

^a, 0.1% EtOH

^b, Restriction enzymes used in the COBRA assay. H, HpyCH4IV; T, Taq-v2.

1×, serum level concentration detected in pregnant mothers/cord blood serum

*p < 0.05 (comparing chemical- and solvent-exposed NSCs)

DISCUSSION

To investigate the effects of chemicals detected in the fetal environment on the epigenetic systems of neuronal cells, we utilized a neural cell differentiation system derived from human iPSCs. Based on previously determined maternal and umbilical cord blood serum concentrations, S-421 altered heterochromatin formation and affected the DNA methylation of neural-related genes in neural stem cells. These findings suggest that certain chemicals detected in the fetal environment act as epimutagens on neuronal cells within biological exposure levels. Previous toxicity assessments of chemicals were primarily focused on cell death and required analysis at concentrations hundreds of times higher than the biological exposure level. In contrast, exposure to the chemicals at serum-level concentrations in this study did not significantly induce cell death. This highlights the importance of epigenetic analysis in evaluating the effects of extremely small amounts of chemicals within biologically relevant exposure ranges on human neural differentiation. Furthermore, changes in chromatin structure are implicated in various diseases (Spielmann et al., 2018; Vogel-Ciernia and Wood, 2014), and abnormal neural development is associated with chromatin dysplasia (Fujita and Yamashita, 2018; Kramer and van Bokhoven, 2009). Thus, detecting epimutagen based on heterochromatin formation is expected to contribute to a new method for assessing normal/abnormal human neural cell development and differentiation.

Effects of S-421 exposure on heterochromatin formation persisted even after the chemical was removed, suggesting that epigenetic changes induced by S-421 may be long-lasting. S-421 is utilized as a potentiator of pyrethroids and organophosphate pesticides and as an insecticide (Krieger et al., 2003) and has been associated with mucosal symptoms in sick building syndrome (Kanazawa et al., 2010). Since S-421 is a stable compound, it persists in the environment and it is found at high concentrations in breast milk (Arai et al., 2011; Miyazaki, 1982). Furthermore, in this study, exposure of NSCs to S-421 induced abnormal neuronal fiber overgrowth during subsequent neural differentiation, even after the chemical was removed. A previous study has reported that combined exposure of human iPSCs to multiple chemicals, including S-421, led to decreased expression of neural marker genes during subsequent neural differentiation (Arai et al., 2015). Furthermore, in mouse embryonic stem cells, a single exposure to S-421 resulted in abnormal hypoplasia of embryoid bodies derived from the three germ layers, including neuronal cells (Arai et al., 2011). These findings suggest that S-421 has a particularly strong adverse effect on stem cells, which are fragile before differentiation, and further disrupt subsequent neural differentiation. This implies that fetal exposure to S-421 may have strong and long-term effects on the development of children even after birth. Therefore, further investigation of S-421 is necessary, particularly regarding the neurodevelopment of children, including studies on actual exposure conditions and long-term development follow-ups.

In contrast, our previous study has shown that a single exposure to S-421 did not affect heterochromatin formation in human iPSCs (Arai et al., 2015). This indicated that the epimutagenic effects of S-421 differed between iPSCs and NSCs. It has been reported that sensitivity to chemicals varies with the differentiation stage of iPSC-derived neuronal cells and that different chemicals affect different stages of neural differentiation (Pei et al., 2016). Because iPSCs can differentiate into various cell types, they are useful for evaluating the cytotoxicity and impaired cell differentiation induced by chemicals in a wide range of tissues and organs. However, because the differentiation fate of NSCs has already been determined in the neural direction, it is possible to evaluate the effects of chemicals specifically on neurotoxicity in brain development or neural network formation. Therefore, selecting an appropriate cell type is crucial for evaluating the effects of chemicals on cellular development and differentiation. The chemical screening system developed in this study using the neural differentiation process is considered an effective method for evaluating epimutagens in neuronal cells.

As an in vitro model of neurological diseases, impaired neural differentiation has been reported in iPSCs derived from patients with schizophrenia (Toyoshima et al., 2016). In contrast, all adverse effects of chemical exposure observed in this study resulted in excessive neuronal fiber expansion. Patients with autism spectrum disorder exhibit increased brain weight and greater number and size of neurons compared to those of healthy individuals (Courchesne et al., 2011). These findings suggest that the expansion of neuronal fibers observed after chemical exposure in this study resembles the phenotype of autism spectrum disorder. Fetal exposure to chemicals increases the incidence of autism spectrum disorder (Ornoy et al., 2015), and cotinine and DEHP, which induced abnormal neural differentiation in this study, have already been linked to the disorder (Carter and Blizard, 2016). This implies that the chemical-induced neural differentiation abnormalities may provide insights into the onset of autism spectrum disorders. Neurological disorders are diagnosed based on behavioral criteria or academic achievement tests; however, diagnosing the disease in children is considered difficult. Therefore, the epimutagen screening system in human neuronal cells established in the present study, in combination with actual data regarding maternal exposure to chemicals during pregnancy, may facilitate the development of novel diagnostic methods for neurological disorders.

Although no statistically significant difference was observed, neurons derived from NSCs exposed to HNO₃ tended to have a slightly larger neuronal fiber area than untreated neurons without the solvents. Similarly, neurons derived from NSCs exposed to Se using HNO₃ as the diluting solvent tended to have a slightly larger neuronal fiber area than those exposed to other chemicals using HNO₃ as the diluting solvent. These results imply that combined exposure to Se and HNO₃ may exacerbate the adverse effects on neural differentiation owing to a synergistic effect. In our previous study, we found that even when a single chemical exposure did not affect heterochromatin formation, combined exposure to multiple chemicals altered heterochromatin formation and neural differentiation (Arai et al., 2015). Therefore, to evaluate whether a chemical acts as an epimutagen, it is necessary to carefully consider the concentration and type of the diluting solvent.

Exposure to serum concentrations of DEP, TCP, nicotine, cotinine, and DEHP induced abnormal expansion of neuronal fibers without affecting heterochromatin formation and DNA methylation of neural-related genes. This suggests that, although the exact mechanisms remain unclear, these chemicals may influence neural differentiation through pathways independent of heterochromatin formation, such as altering histone modifications in euchromatin. However, at concentrations ten times higher than serum levels, nicotine, cotinine, and DEHP altered heterochromatin formation. Previous analyses of chemicals detected in maternal and umbilical cord blood serum show that the exposure range varies between 2−30 times the minimum detected value (Arai et al., 2011), indicating that the 10-fold serum concentrations analyzed in this study are within the range of actual biological exposure. Additionally, inducing neural differentiation in vitro is time-consuming and costly, requiring several weeks to months. The epimutagen screening system developed in this study offers a strategy to predict in advance which chemicals will affect neural differentiation, potentially reducing time and costs. Therefore, detecting epimutagens in neuronal cells at multiple concentrations within the biological exposure range is thought to be effective in thoroughly detecting chemicals that may affect neural differentiation.

S-421 exposure also affected DNA methylation of neural-related genes CDH2 and SOX10 in NSCs. CDH2 plays a role in synaptic function, and its deficiency is associated with neurological disorders (Accogli et al., 2019). SOX10 is a key transcription factor involved in the development and differentiation of neural crest and peripheral neurons (Pingault et al., 2022). DNA methylation is tightly regulated, and its status, especially in genic regions, changes throughout the developmental stage, thereby forming cell type-specific DNA methylation patterns (Ikegami et al., 2009; Shiota, 2004). Therefore, aberrant DNA methylation at a specific stage may impair normal cell differentiation based on the formation of DNA methylation patterns. In other words, abnormal DNA methylation during fetal development may hinder future cell development and differentiation even after birth. The relationship between DNA methylation alterations and differentiation abnormalities induced by chemical exposure remains unclear. However, in NSCs, S-421 exposure adversely affected subsequent neural differentiation, along with the disruption of the DNA methylation status of neural-related genes. In chemical-related neurological diseases, the time of exposure to chemicals and the time of disease onset are often different; this makes it difficult to elucidate the molecular mechanisms underlying the onset of neurological diseases. If such relationships between DNA methylation alterations and differentiation abnormalities are clarified in the future, epigenetics may help explain the delay between fetal exposure to chemicals during pregnancy and the onset of neurological diseases in children. Recently, epigenome editing, a technique for artificially modifying epigenetics, has been actively developed. Selective DNA methylation modifications of neural-related genes, including CDH2 and SOX10, through epigenome editing, may provide new insights into the mechanisms underlying chemical-related neurological diseases.

In conclusion, epimutagens that disrupt epigenetic systems in human neuronal cells were successfully detected based on heterochromatin formation. Disrupting epigenetic regulation at specific differentiation stages can have detrimental effects on subsequent cellular development and differentiation. Thus, an epimutagen screening system for human neuronal cells could enable the identification of chemicals with the potential to cause abnormalities in neural differentiation within biologically relevant exposure ranges, helping to elucidate the causes of neurological disorders in children resulting from fetal exposure to harmful chemicals.

ACKNOWLEDGMENTS

YA thanks Ms. Shiori Nakamura for her kind cooperation with this project.

Funding

This work was supported by the CERI Competitive Research Grant from the Chemicals Evaluation and Research Institute, Japan (CERI) (to YA) and Grant-in-Aid for Scientific Research (21K12277) (to YA).

Conflict of interest

The authors declare no competing interests.

Data availability

The data in this study are included in the article/supplementary materials. Contact the corresponding authors directly to request the underlying data.

Author contributions

Conceptualization: Yoshikazu Arai

Investigation: Yoshikazu Arai and Koichiro Nishino

Writing – original draft: Yoshikazu Arai

Ethical approval and consent to participate

Not applicable.

Patient consent for publication

Not applicable.

REFERENCES

•  Accogli,  A.,  Calabretta,  S.,  St-Onge,  J.,  Boudrahem-Addour,  N.,  Dionne-Laporte,  A.,  Joset,  P.,  Azzarello-Burri,  S.,  Rauch,  A.,  Krier,  J.,  Fieg,  E.,  Pallais,  J.C.,  McConkie-Rosell,  A.,  McDonald,  M.,  Freedman,  S.F.,  Rivière,  J.B.,  Lafond-Lapalme,  J.,  Simpson,  B.N.,  Hopkin,  R.J.,  Trimouille,  A.,  Van-Gils,  J.,  Begtrup,  A.,  McWalter,  K.,  Delphine,  H.,  Keren,  B.,  Genevieve,  D.,  Argilli,  E.,  Sherr,  E.H.,  Severino,  M.,  Rouleau,  G.A.,  Yam,  P.T.,  Charron,  F. and  Srour,  M.; Undiagnosed Diseases Network. (2019): De Novo Pathogenic Variants in N-cadherin Cause a Syndromic Neurodevelopmental Disorder with Corpus Collosum, Axon, Cardiac, Ocular, and Genital Defects. Am. J. Hum. Genet., 105, 854-868.
•  Allshire,  R.C. and  Madhani,  H.D. (2018): Ten principles of heterochromatin formation and function. Nat. Rev. Mol. Cell Biol., 19, 229-244.
•  Arai,  Y.,  Hayakawa,  K.,  Arai,  D.,  Ito,  R.,  Iwasaki,  Y.,  Saito,  K.,  Akutsu,  K.,  Takatori,  S.,  Ishii,  R.,  Hayashi,  R.,  Izumi,  S.,  Sugino,  N.,  Kondo,  F.,  Horie,  M.,  Nakazawa,  H.,  Makino,  T.,  Hirosawa,  M.,  Shiota,  K. and  Ohgane,  J. (2015): Putative Epimutagens in Maternal Peripheral and Cord Blood Samples Identified Using Human Induced Pluripotent Stem Cells. BioMed Res. Int., 2015, 876047.
•  Arai,  Y.,  Ohgane,  J.,  Yagi,  S.,  Ito,  R.,  Iwasaki,  Y.,  Saito,  K.,  Akutsu,  K.,  Takatori,  S.,  Ishii,  R.,  Hayashi,  R.,  Izumi,  S.,  Sugino,  N.,  Kondo,  F.,  Horie,  M.,  Nakazawa,  H.,  Makino,  T. and  Shiota,  K. (2011): Epigenetic assessment of environmental chemicals detected in maternal peripheral and cord blood samples. J. Reprod. Dev., 57, 507-517.
•  Boncler,  M.,  Lukasiak,  M.,  Dastych,  J.,  Golanski,  J. and  Watala,  C. (2019): Differentiated mitochondrial function in mouse 3T3 fibroblasts and human epithelial or endothelial cells in response to chemical exposure. Basic Clin. Pharmacol. Toxicol., 124, 199-210.
• Carter, C.J, Blizard, R.A. (2016): Autism genes are selectively targeted by environmental pollutants including pesticides, heavy metals, bisphenol A, phthalates and many others in food, cosmetics or household products. Neurochem Int., S0197–0186, 30197-30198.
•  Cedar,  H. and  Bergman,  Y. (2009): Linking DNA methylation and histone modification: patterns and paradigms. Nat. Rev. Genet., 10, 295-304.
•  Chen,  Z.,  Liu,  Y.,  Wright,  F.A.,  Chiu,  W.A. and  Rusyn,  I. (2020): Rapid hazard characterization of environmental chemicals using a compendium of human cell lines from different organs. Altern. Anim. Exp., 37, 623-638.
•  Courchesne,  E.,  Mouton,  P.R.,  Calhoun,  M.E.,  Semendeferi,  K.,  Ahrens-Barbeau,  C.,  Hallet,  M.J.,  Barnes,  C.C. and  Pierce,  K. (2011): Neuron number and size in prefrontal cortex of children with autism. JAMA, 306, 2001-2010.
•  Eskenazi,  B.,  Marks,  A.R.,  Bradman,  A.,  Harley,  K.,  Barr,  D.B.,  Johnson,  C.,  Morga,  N. and  Jewell,  N.P. (2007): Organophosphate pesticide exposure and neurodevelopment in young Mexican-American children. Environ. Health Perspect., 115, 792-798.
•  Fujita,  Y. and  Yamashita,  T. (2018): Spatial organization of genome architecture in neuronal development and disease. Neurochem. Int., 119, 49-56.
•  Golob,  J.L.,  Paige,  S.L.,  Muskheli,  V.,  Pabon,  L. and  Murry,  C.E. (2008): Chromatin remodeling during mouse and human embryonic stem cell differentiation. Dev. Dyn., 237, 1389-1398.
•  Harari,  R.,  Julvez,  J.,  Murata,  K.,  Barr,  D.,  Bellinger,  D.C.,  Debes,  F. and  Grandjean,  P. (2010): Neurobehavioral deficits and increased blood pressure in school-age children prenatally exposed to pesticides. Environ. Health Perspect., 118, 890-896.
•  Holliday,  R. and  Ho,  T. (2002): DNA methylation and epigenetic inheritance. Methods, 27, 179-183.
•  Ikegami,  K.,  Ohgane,  J.,  Tanaka,  S.,  Yagi,  S. and  Shiota,  K. (2009): Interplay between DNA methylation, histone modification and chromatin remodeling in stem cells and during development. Int. J. Dev. Biol., 53, 203-214.
•  Kanazawa,  A.,  Saito,  I.,  Araki,  A.,  Takeda,  M.,  Ma,  M.,  Saijo,  Y. and  Kishi,  R. (2010): Association between indoor exposure to semi-volatile organic compounds and building-related symptoms among the occupants of residential dwellings. Indoor Air, 20, 72-84.
•  Kramer,  J.M. and  van Bokhoven,  H. (2009): Genetic and epigenetic defects in mental retardation. Int. J. Biochem. Cell Biol., 41, 96-107.
•  Krieger,  R.I.,  Dinoff,  T.M. and  Zhang,  X. (2003): Octachlorodipropyl ether (s-2) mosquito coils are inadequately studied for residential use in Asia and illegal in the United States. Environ. Health Perspect., 111, 1439-1442.
•  Lieb,  J.D.,  Beck,  S.,  Bulyk,  M.L.,  Farnham,  P.,  Hattori,  N.,  Henikoff,  S.,  Liu,  X.S.,  Okumura,  K.,  Shiota,  K.,  Ushijima,  T. and  Greally,  J.M. (2006): Applying whole-genome studies of epigenetic regulation to study human disease. Cytogenet. Genome Res., 114, 1-15.
•  Livak,  K.J. and  Schmittgen,  T.D. (2001): Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)). Method. Methods, 25, 402-408.
•  Llop,  S.,  Julvez,  J.,  Fernandez-Somoano,  A.,  Santa Marina,  L.,  Vizcaino,  E.,  Iñiguez,  C.,  Lertxundi,  N.,  Gascón,  M.,  Rebagliato,  M. and  Ballester,  F. (2013): Prenatal and postnatal insecticide use and infant neuropsychological development in a multicenter birth cohort study. Environ. Int., 59, 175-182.
•  Miyazaki,  T. (1982): Residues of the synergist S-421 in human milk collected from the Tokyo metropolitan area. Bull. Environ. Contam. Toxicol., 29, 566-569.
•  Nishino,  K.,  Takasawa,  K.,  Okamura,  K.,  Arai,  Y.,  Sekiya,  A.,  Akutsu,  H. and  Umezawa,  A. (2021): Identification of an epigenetic signature in human induced pluripotent stem cells using a linear machine learning model. Hum. Cell, 34, 99-110.
•  Nishino,  K.,  Toyoda,  M.,  Yamazaki-Inoue,  M.,  Fukawatase,  Y.,  Chikazawa,  E.,  Sakaguchi,  H.,  Akutsu,  H. and  Umezawa,  A. (2011): DNA methylation dynamics in human induced pluripotent stem cells over time. PLoS Genet., 7, e1002085.
•  Ornoy,  A.,  Weinstein-Fudim,  L. and  Ergaz,  Z. (2015): Prenatal factors associated with autism spectrum disorder (ASD). Reprod. Toxicol., 56, 155-169.
• Pei Y, Peng J, Behl M, Sipes NS, Shockley KR, Rao MS, Tice RR, Zeng X. (2016): Comparative neurotoxicity screening in human iPSC-derived neural stem cells, neurons and astrocytes. Brain Res., 1638(Pt A), 57-73.
•  Pingault,  V.,  Zerad,  L.,  Bertani-Torres,  W. and  Bondurand,  N. (2022): SOX10: 20 years of phenotypic plurality and current understanding of its developmental function. J. Med. Genet., 59, 105-114.
•  Rauh,  V.,  Arunajadai,  S.,  Horton,  M.,  Perera,  F.,  Hoepner,  L.,  Barr,  D.B. and  Whyatt,  R. (2011): Seven-year neurodevelopmental scores and prenatal exposure to chlorpyrifos, a common agricultural pesticide. Environ. Health Perspect., 119, 1196-1201.
•  Sekiya,  A.,  Takasawa,  K.,  Arai,  Y.,  Horike,  S.I.,  Akutsu,  H.,  Umezawa,  A. and  Nishino,  K. (2022): Variation of DNA methylation on the IRX1/2 genes is responsible for the neural differentiation propensity in human induced pluripotent stem cells. Regen. Ther., 21, 620-630.
•  Shiota,  K. (2004): DNA methylation profiles of CpG islands for cellular differentiation and development in mammals. Cytogenet. Genome Res., 105, 325-334.
•  Spielmann,  M.,  Lupiáñez,  D.G. and  Mundlos,  S. (2018): Structural variation in the 3D genome. Nat. Rev. Genet., 19, 453-467.
•  Toyoshima,  M.,  Akamatsu,  W.,  Okada,  Y.,  Ohnishi,  T.,  Balan,  S.,  Hisano,  Y.,  Iwayama,  Y.,  Toyota,  T.,  Matsumoto,  T.,  Itasaka,  N.,  Sugiyama,  S.,  Tanaka,  M.,  Yano,  M.,  Dean,  B.,  Okano,  H. and  Yoshikawa,  T. (2016): Analysis of induced pluripotent stem cells carrying 22q11.2 deletion. Transl. Psychiatry, 6, e934.
•  Ushijima,  T. and  Okochi-Takada,  E. (2005): Aberrant methylations in cancer cells: where do they come from? Cancer Sci., 96, 206-211.
•  Vogel-Ciernia,  A. and  Wood,  M.A. (2014): Neuron-specific chromatin remodeling: a missing link in epigenetic mechanisms underlying synaptic plasticity, memory, and intellectual disability disorders. Neuropharmacology, 80, 18-27.
•  Wilson,  M.S.,  Graham,  J.R. and  Ball,  A.J. (2014): Multiparametric High Content Analysis for assessment of neurotoxicity in differentiated neuronal cell lines and human embryonic stem cell-derived neurons. Neurotoxicology, 42, 33-48.
•  Xiong,  Z. and  Laird,  P.W. (1997): COBRA: a sensitive and quantitative DNA methylation assay. Nucleic Acids Res., 25, 2532-2534.