Within-Generational Disruption of the Stress Response by Fluoxetine and Other Environmental Contaminants in Zebrafish

Abstract

Selective serotonin reuptake inhibitors (SSRIs), like fluoxetine, are widely used to treat depressive disorders during pregnancy. These antidepressants reach water reservoirs through sewage treatment facilities and expose the aquatic vertebrates, including fish. It has been shown that early-life exposure to fluoxetine could disrupt the normal function of the stress axis by decreasing the level of circulating glucocorticoids in humans, rodents, and teleosts. Our lab recently showed that early life exposure to fluoxetine resulted in transgenerational hyporcortisolism and altered exploratory behaviour in adult male zebrafish and their descendant male adults for at least three generations. In the current study, we used a stress-responsive transgenic zebrafish line (SR4G) that expresses green fluorescence protein (eGFP) under the control of six consecutive glucocorticoid response elements. The effects of developmental exposure to fluoxetine on the transcriptional profiles of genes in the larval head and male adult telencephalon and hypothalamus were analysed using high throughput RNA sequencing. We also assessed the potential of eGFP mRNA to evaluate blunted stress response as an alternative to cortisol immunoassay measurements. The effects of bisphenol A, vinclozolin and fluoxetine were ytested in the SR4G line. Developmental exposure to fluoxetine resulted in a life-long dysregulation of pathways involved in nervous system development, stress response, and lipid metabolism in both larvae and adult zebrafish. Numerous differentially expressed genes in zebrafish are orthologous to genes in Homo sapiens linked the development of the major depressive disorder and epigenetics regulation and include bdnf, trkb, npas4, per1, per2, dnmt3a, adarb1, adaeb2, hdac4, hdac5, hdac8, and atf2. It is suggested that the dysregulation of the primary transcription regulators of circadian rhythm (clocka) and stress response (nr3c1), amongst others, were the potential drivers of the observed life-long effects. Furthermore, we report on a significant positive linear correlation between cortisol levels and eGFP mRNA levels in SR4G transgenic zebrafish larvae (R2> 0.9). Random forest and logistic regression models trained by eGFP mRNA levels both correctly predicted the blunted stress response. The negative predictive value (NPV) for both models was 100%. Models based on the mRNA levels of 11 genes associated with neurogenesis, stress response and depression resulted in a similar 100% NPV. These findings provide evidence for a life-long effect of developmental exposure to fluoxetine. This study also provides a proof-of-concept for an in vivo biomonitoring assay to screen chemicals for their stress-disrupting potential.