Hormone signaling plays an essential role during fetal life and is vital for brain development. Endocrine-disrupting chemicals can interfere with the hormonal milieu during this critical time-period, disrupting key neurodevelopmental processes. Hence, there is a need for the development of assays that evaluate developmental neurotoxicity (DNT) induced by an endocrine mode of action. Herein, we evaluated the neural progenitor C17.2 cell line as an in vitro test system to aid in the detection of endocrine disruption-induced DNT. For this, C17.2 cells were exposed during 10 days of differentiation to agonists and antagonists of the thyroid hormone (THR), glucocorticoid (GR), retinoic acid (RAR), retinoic x (RXR), oxysterol (LXR), estrogen (ER), androgen (AR), and peroxisome proliferator activated delta (PPARβ/δ) receptors, as well as to the agonist of the vitamin D (VDR) receptor. Upon exposure and differentiation, neuronal morphology (neurite outgrowth and branching) and the percentage of neurons in culture were assessed by immunofluorescence. For this, the cells were stained for βIII-tubulin (neuronal marker). C17.2 cells decreased neurite outgrowth and branching in response to RAR, RXR and PPARβ/δ agonists. Exposure to the GR agonist increased the number of cells differentiating into neurons, while exposure to the RXR agonist had the opposite effect. With this approach, we demonstrate that C17.2 cells are responsive to GR, RAR, RXR, and PPARβ/δ agonists and hence could be useful to develop a test system for hazard assessment of endocrine disruption-induced DNT.

Background/Objectives: Many pregnant women globally suffer from depression and are routinely prescribed selective serotonin reuptake inhibitors (SSRIs). These drugs function by blocking the re-uptake of serotonin by the serotonin transporter (SERT) into neurons, resulting in its accumulation in the presynaptic cleft. Despite a large amount of research suggesting a potential link to neurodevelopmental disorders in children whose mothers took these drugs during pregnancy, their possible adverse effects are still debated, and results are contradictory. On the other hand, there is an immediate need for improved cell-based models for developmental neurotoxicity studies (DNT) to minimize the use of animals in research. Methods: In this study, we aimed to assess the effects of clinically relevant concentrations of paroxetine (PAR), fluoxetine (FLX), and citalopram (CIT)—on maturing neurons derived from human neural stem cells using multiple endpoints. Results: Although none of the tested concentrations of FLX, CIT, or PAR significantly affected cell viability, FLX (10 µM) exhibited the highest reduction in viability compared to the other drugs. Regarding neurite outgrowth, CIT did not have a significant effect. However, FLX (10 µM) significantly reduced both mean neurite outgrowth and mean processes, PAR significantly reduced mean processes, and showed a trend of dysregulation of multiple genes associated with neuronal development at therapeutic-relevant serum concentrations. Conclusions: Transcriptomic data and uptake experiments found no SERT activity in the system, suggesting that the adverse effects of FLX and PAR are independent of SERT.

Acrylamide (ACR) is a known neurotoxicant and developmental neurotoxicant. As a soft electrophile, ACR reacts with thiol groups in cysteine. One hypothesis of ACR induced neurotoxicity and developmental neurotoxicity (DNT) is conjugation with reduced glutathione (GSH) leading to GSH depletion, increased reactive oxygen species (ROS) production and further oxidative stress and cellular damage. In this regard, we have investigated the effect of ACR on neuronal differentiation, glutathione levels and ROS production in the human neuroblastoma SH-SY5Y cell model. After 9 days of differentiation and exposure, ACR significantly impaired area neurites per cell at non-cytotoxic concentrations (0.33 μM and 10 μM). Furthermore, 10 μM ACR dysregulated 9 mRNA markers important for neuronal development, 5 of them being associated with cytoskeleton organization and axonal guidance. At the non-cytotoxic concentrations that significantly attenuate neuronal differentiation, ACR did neither decrease the level of GSH or total glutathione levels, nor increased ROS production. In addition, the expression of 5 mRNA markers for cellular stress was assessed with no significant altered regulation after ACR exposure up to 320 μM. Thus, ACR-induced DNT is not due to GSH depletion and increased ROS production, neither at non-cytotoxic nor cytotoxic concentrations, in the SH-SH5Y model during differentiation.

Acrylamide (ACR) is a known neurotoxicant that can pass the placenta and has been detected in breast milk. Some in vivo and in vitro studies indicate that ACR exposure might lead to developmental neurotoxicity (DNT). Here, we have developed a physiologically-based toxicokinetic model for a pregnant human population using PK-Sim. We performed an in vitro to in vivo extrapolation (IVIVE) of data collected from human neuroblastoma SH-SY5Y cells exposed during differentiation to ACR. The developed PBTK model was successfully evaluated and predicted fetal plasma concentrations in the low nM range after exposing the model to an estimated average daily intake for pregnant women. The IVIVE showed that low concentrations of ACR (fM-nM) that induced attenuated differentiation of the SH-SY5Y neuronal cell model, were relevant for human exposure to ACR from oral intake. However, doses estimated in the IVIVE from concentrations in the µM range, were found to be unrealistic by exposure through food intake for an average daily intake. However, in case of exposure due to environmental pollution or occupational exposure, these concentrations may be reached in fetal plasma. The findings in this study raise the concern regarding ACR exposure during pregnancy as well as the relevance of testing concentrations in vitro that are several orders of magnitude higher than the predicted fetal plasma concentrations.

The cyanotoxin cylindrospermopsin (CYN) has been postulated to cause neurotoxicity, although the studies in this concern are very few. In addition, some studies in vitro indicate its possible effects on development. Furthermore, pesticides can be present in the same environmental samples as cyanotoxins. Therefore, chlor-pyrifos (CPF) has been one of the most common pesticides used worldwide. The aim of this report was to study the effects of CYN, isolated and in combination with CPF, in a developmental neurotoxicity in vitro model. The human neuroblastoma SH-SY5Y cell line was exposed during 6 days of differentiation to both toxics to study their effects on cell viability and neurite outgrowth. To further evaluate effects of both toxicants on cholinergic signaling, their agonistic and antagonistic activities on the alpha 7 homomeric nicotinic acetylcholine receptor (nAChR) were studied upon acute exposure. Moreover, a transcriptomic analysis by qPCR was performed after 6 days of CYN-exposure during differentiation. The results showed a concentration-dependent decrease on both cell viability and neurite outgrowth for both toxics isolated, leading to effective concentration 20 (EC20) values of 0.35 mu M and 0.097 mu M for CYN on cell viability and neurite outgrowth, respectively, and 100 mu M and 58 mu M for CPF, while the combination demonstrated no significant variations. In addition, 95 mu M and 285 mu M CPF demonstrated to act as an antagonist to nicotine on the nAChR, although CYN up to 2.4 mu M had no effect on the efficacy of these receptors. Additionally, the EC20 for CYN (0.097 mu M) on neurite outgrowth downregulated expression of the 5 genes NTNG2 (netrin G2), KCNJ11 (potassium channel), SLC18A3 (vesicular acetylcholine transporter), APOE (apolipoprotein E), and SEMA6B (semaphorin 6B), that are all important for neuronal development. Thus, this study points out the importance of studying the effects of CYN in terms of neurotoxicity and developmental neurotoxicity.

Current guidelines for developmental neurotoxicity (DNT) evaluation are based on animal models. These have limitations so more relevant, efficient and robust approaches for DNT assessment are needed. We have used the human SH-SY5Y neuroblastoma cell model to evaluate a panel of 93 mRNA markers that are frequent in Neuronal diseases and functional annotations and also differentially expressed during retinoic acid-induced differentiation in the cell model. Rotenone, valproic acid (VPA), acrylamide (ACR) and methylmercury chloride (MeHg) were used as DNT positive compounds. Tolbutamide, D-mannitol and clofibrate were used as DNT negative compounds. To determine concentrations for exposure for gene expression analysis, we developed a pipeline for neurite outgrowth assessment by live-cell imaging. In addition, cell viability was measured by the resazurin assay. Gene expression was analyzed by RT-qPCR after 6 days of exposure during differentiation to concentrations of the DNT positive compounds that affected neurite outgrowth, but with no or minimal effect on cell viability. Methylmercury affected cell viability at lower concentrations than neurite outgrowth, hence the cells were exposed with the highest non-cytotoxic concentration. Rotenone (7.3 nM) induced 32 differentially expressed genes (DEGs), ACR (70 µM) 8 DEGs, and VPA (75 µM) 16 DEGs. No individual genes were significantly dysregulated by all 3 DNT positive compounds (p < 0.05), but 9 genes were differentially expressed by 2 of them. Methylmercury (0.8 nM) was used to validate the 9 DEGs. The expression of SEMA5A (encoding semaphorin 5A) and CHRNA7 (encoding nicotinic acetylcholine receptor subunit α7) was downregulated by all 4 DNT positive compounds. None of the DNT negative compounds dysregulated any of the 9 DEGs in common for the DNT positive compounds. We suggest that SEMA5A or CHRNA7 should be further evaluated as biomarkers for DNT studies in vitro since they also are involved in neurodevelopmental adverse outcomes in humans.

Analysis of the transcriptomic alterations upon chemical challenge, provides in depth mechanistic information on the compound’s toxic mode of action, by revealing specific pathway activation and other transcriptional modulations. Mapping changes in cellular behaviour to chemical insult, facilitates the characterisation of chemical hazard. In this study, we assessed the transcriptional landscape of mitochondrial impairment through the inhibition of the electron transport chain (ETC) in a human renal proximal tubular cell line (RPTEC/TERT1). We identified the unfolded protein response pathway (UPR), particularly the PERK/ATF4 branch as a common cellular response across ETC I, II and III inhibitions. This finding and the specific genes elaborated may aid the identification of mitochondrial liabilities of chemicals in both legacy data and prospective transcriptomic studies.

Mitochondrial perturbation is a key event in chemical-induced organ toxicities that is incompletely understood. Here, we studied how electron transport chain (ETC) complex I, II, or III (CI, CII and CIII) inhibitors affect mitochondrial functionality, stress response activation, and cell viability using a combination of high-content imaging and TempO-Seq in HepG2 hepatocyte cells. CI and CIII inhibitors perturbed mitochondrial membrane potential (MMP) and mitochondrial and cellular ATP levels in a concentration- and time-dependent fashion and, under conditions preventing a switch to glycolysis attenuated cell viability, whereas CII inhibitors had no effect. TempO-Seq analysis of changes in mRNA expression pointed to a shared cellular response to CI and CIII inhibition. First, to define specific ETC inhibition responses, a gene set responsive toward ETC inhibition (and not to genotoxic, oxidative, or endoplasmic reticulum stress) was identified using targeted TempO-Seq in HepG2. Silencing of one of these genes, NOS3, exacerbated the impact of CI and CIII inhibitors on cell viability, indicating its functional implication in cellular responses to mitochondrial stress. Then by monitoring dynamic responses to ETC inhibition using a HepG2 GFP reporter panel for different classes of stress response pathways and applying pathway and gene network analysis to TempO-Seq data, we looked for downstream cellular events of ETC inhibition and identified the amino acid response (AAR) as being triggered in HepG2 by ETC inhibition. Through in silico approaches we provide evidence indicating that a similar AAR is associated with exposure to mitochondrial toxicants in primary human hepatocytes. Altogether, we (i) unravel quantitative, time- and concentration-resolved cellular responses to mitochondrial perturbation, (ii) identify a gene set associated with adaptation to exposure to active ETC inhibitors, and (iii) show that ER stress and an AAR accompany ETC inhibition in HepG2 and primary hepatocytes. 

Methylmercury (MeHg) is a developmental neurotoxicant, and one potential mechanism of MeHg toxicity is epigenetic dysregulation. In a recent meta-analysis of epigenome-wide association studies (EWAS), associations between prenatal MeHg exposure and DNA methylation at several genomic sites were identified in blood from newborns and children. While EWASs reveal human-relevant associations, experimental studies are required to validate the relationship between exposure and DNA methylation changes, and to assess if such changes have implications for gene expression. Herein, we studied DNA methylation and gene expression of five of the top genes identified in the EWAS meta-analysis, MED31, MRPL19, GGH, GRK1, and LYSMD3, upon MeHg exposure in human SH-SY5Y cells exposed to 8 or 40 nM of MeHg during differentiation, using bisulfite-pyrosequencing and qPCR, respectively. The concentrations were selected to cover the range of MeHg concentrations in cord blood (2–8.5 μg/L) observed in the cohorts included in the EWAS. Exposure to MeHg increased DNA methylation at MED31, a transcriptional regulator essential for fetal development. The results were in concordance with the epidemiological findings where more MED31 methylation was associated with higher concentrations of MeHg. Additionally, we found a non-significant decrease in DNA methylation at GGH, which corresponds to the direction of change observed in the EWAS, and a significant correlation of GGH methylation with its expression. In conclusion, this study corroborates some of the EWAS findings and puts forward candidate genes involved in MeHg’s effects on the developing brain, thus highlighting the value of experimental validation of epidemiological association studies.

The need for reliable, sensitive (developmental) neurotoxicity testing of chemicals has steadily increased. Given the limited capacities for routine testing according to accepted regulatory guidelines, there is potential risk to human health and the environment. Most toxicity studies are based on mammalian test systems, which have been questioned for low sensitivity, limited relevance for humans, and animal welfare considerations. This increased the need for alternative models, one of which is the zebrafish (Danio rerio) embryo. This study assessed selected neonicotinoids at sub-lethal concentrations for their effects on embryonic development and behavior. The fish embryo acute toxicity test (OECD TG 236) determined the lowest observable effective concentrations, which were used as the highest test concentrations in subsequent behavioral assays. In the FET test, no severe compound-induced sublethal effects were seen at < 100 µM. In the coiling assay, exposure to ≥ 1.25 µM nicotine (positive control) affected both the burst duration and burst count per minute, whereas ≥ 50 µM thiacloprid affected the mean burst duration. Exposure to ≥ 50 µM acetamiprid and imidacloprid induced significant alterations in both mean burst duration and burst count per minute. In the swimming assay, 100 µM acetamiprid induced alterations in the frequency and extent of movements, whilst nicotine exposure only induced non-significant changes. All behavioral changes could be correlated to findings in mammalian studies. Given the quest for alternative test methods of (developmental) neurotoxicity, zebrafish embryo behavior testing could be integrated into a future tiered testing scheme.