Abstract
All-trans retinoic acid (ATRA) plays key roles in neurogenesis mediated by retinoic acid receptors (RARs). RARs are important targets for the therapeutic regulation of neurogenesis but effective drug development depends on modelling-based strategies to design high-specificity ligands in combination with good biological assays to discriminate between target-specificity and off-target effects. Using neuronal differentiation as a model, the aim of this study was to test the hypothesis that responses across different temporal scales and assay platforms can be used as comparable measures of retinoid activity. In biological assays based on cell phenotype or behaviour, two structurally similar synthetic retinoids, differing in RAR affinity and specificity, retained their relative activities across different temporal scales. In contrast, assays based on the transcriptional activation of specific genes in their normal genomic context were less concordant with biological assays. Gene-induction assays for retinoid activity as modulators of neurogenesis require careful interpretation in the light of variation in ligand-receptor affinity, receptor expression and gene function. A better characterization of neuronal phenotypes and their regulation by retinoids is badly needed as a framework for understanding how to regulate neuronal development.