Cyanobacterial anatoxin-a does not induce in vitro developmental neurotoxicity, but changes gene expression patterns in co-exposure with all-trans retinoic acid.

Cyanobacterial blooms are increasing in frequency and intensity globally, and impacting recreational waters as well as waters used for drinking water provisioning. They are sources of bioactive metabolites including retinoids and the neurotoxin anatoxin-a. Here, we investigated the effects of anatoxin-a on a differentiating in vitro human neural stem cell model previously characterised with retinoic acids. Effects on protein and gene expression upon exposure for 9 or 18 days to anatoxin-a alone or in co-exposure with all-trans retinoic acid were evaluated using a panel of neural and glial differentiation biomarkers. Anatoxin-a did not cause distinct developmental neurotoxicity alone, or in co-exposure with retinoic acid. However, in line with its excitotoxicity, in co-exposure with 200 nM all-trans retinoic acid it reduced the differentiation of acetylcholinergic neuron subtypes in the culture at 1000 nM (highest tested concentration). While this could have substantial functional implications for the developing nervous system, there is no indication for developmental neurotoxicity beyond its (excito-)toxicity to acetylcholinergic neurons, which only occurred in co-exposure to all-trans retinoic acid.

Effects of all-trans and 9-cis retinoic acid on differentiating human neural stem cells in vitro.

Cyanobacterial blooms are known sources of environmentally-occurring retinoid compounds, including all-trans and 9-cis retinoic acids (RAs). The developmental hazard for aquatic organisms has been described, while the implications for human health hazard assessment are not yet sufficiently characterized. Here, we employ a human neural stem cell model that can differentiate in vitro into a mixed culture of neurons and glia. Cells were exposed to non-cytotoxic 8-1000 nM all-trans or 9-cis RA for 9-18 days (DIV13 and DIV22, respectively). Impact on biomarkers was analyzed on gene expression (RT-qPCR) and protein level (western blot and proteomics) at both time points; network patterning (immunofluorescence) on DIV22. RA exposure significantly concentration-dependently increased gene expression of retinoic acid receptors and the metabolizing enzyme CYP26A1, confirming the chemical-specific response of the model. Expression of thyroid hormone signaling-related genes remained mostly unchanged. Markers of neural progenitors/stem cells (PAX6, SOX1, SOX2, NESTIN) were decreased with increasing RA concentrations, though a basal population remained. Neural markers (DCX, TUJ1, MAP2, NeuN, SYP) remained unchanged or were decreased at high concentrations (200-1000 nM). Conversely, (astro-)glial marker S100ß was increased concentration-dependently on DIV22. Together, the biomarker analysis indicates an RA-dependent promotion of glial cell fates over neural differentiation, despite the increased abundance of neural protein biomarkers during differentiation. Interestingly, RA exposure induced substantial changes to the cell culture morphology: while low concentrations resulted in a network-like differentiation pattern, high concentrations (200-1000 nM RA) almost completely prevented such network patterning. After functional confirmation for implications in network function, such morphological features could present a proxy for network formation assessment, an apical key event in (neuro-)developmental Adverse Outcome Pathways. The described application of a human in vitro model for (developmental) neurotoxicity to emerging environmentally-relevant retinoids contributes to the evidence-base for the use of differentiating human in vitro models for human health hazard and risk assessment.