Neuron-Glia Signaling Regulates the Onset of the Antidepressant Response

Background:

Over the years, the number of individuals experiencing depressive episodes have been on the rise. With around 350 million individuals affected worldwide, it remains one of the major disabling health conditions, creating psychosocial and economic burdens. Several brain regions are affected by depression, including the hippocampus, prefrontal cortex, anterior cingulate cortex, amygdala, and nucleus accumbens. Consequently, depression causes a global deficit in information processing in the brain, manifesting as behavioural and cognitive changes in affected individuals. The complexity of depressive disorders lies in their multimodal associations with genetic mutations, neurochemical imbalance, neuroendocrine system dysfunction, cytokine imbalance, and structural alterations at a neural circuit and synaptic level (1). Pharmacological intervention via selective serotonin reuptake inhibitors (SSRIs) have been shown to restrict the diminishing level of serotonin (5-HT) in the synaptic clefts, causing an improvement in mood. Treatment response involves ameliorating the structural and functional abnormalities associated with depression, which is ineffective in almost one-third of the patients. However, in most responding patients, the improvements associated with SSRIs take several weeks to appear (2). In this preprint, Vicky Yao, Revathy Chottekalapanda and colleagues address the cellular and molecular changes happening as a response to antidepressant (fluoxetine- Prozac) treatment at a temporal resolution.

Key results:
To induce depression, FVB-Tg mice (expressing EGFP under the S100a10 promoter) on C57BL/6 background were socially isolated from weaning (P21) until 10-weeks of age. Using single-nuclei RNAseq, the authors first assessed the changes in gene expression at a whole cerebral cortex (involved in mood regulation) level at different time points of vehicle or fluoxetine treatment (mixed with water)- 3, 7, and 10 days. 36 clusters of neuronal, inter-neuronal, and non-neuronal cell types were identified that expressed genes differentially. Upon fluoxetine treatment, several genes that were upregulated in response to stress were brought back to their normalized level in a time-dependent manner: 10 days of treatment had the highest number of genes that were normalized by fluoxetine. These changes were observed in non-neuronal cell populations as well as neuronal and inter-neuronal cells. The observed responses were brought about as early as 3 days post-treatment in neurons of cortical layers 2/3 and 6, and astrocytes. Interestingly, among all cell types, astrocytes showed the strongest transcriptional response at all time points, and was the first glial cell type to respond at 3 days. Oligodendrocyte progenitor cells (OPCs) and mature oligodendrocytes underwent gene expression changes on days 7 and 10 respectively, suggesting the likelihood of OPC maturation. Microglia did not respond to fluoxetine treatment, which remains an unanswered question from this study. Quite interestingly, these early glial alterations were followed by transcriptional changes in multiple neuronal populations on day 10. These results indicate that glial cells respond first to chronic antidepressant treatment followed by neurons.

Gene ontology and pathway analyses of the differentially expressed genes were used to identify which of the glial functions were involved in mediating neuronal properties. Overall, astrocyte-specific gene changes were attributed to regulating cellular morphology, neurotrophin signaling and regulation of OPCs. However, gene alterations in neuronal populations regulated neuronal remodelling and plasticity-inducing pathways that included neurotransmission, lipid and cholesterol metabolism, morphology regulation, synaptogenesis and synapse maintenance. Genes involved in myelination and axonal regulation had responded in mature oligodendrocytes. The authors also identified several upstream signaling modulators, involved in both neurogenic (TCF7L2, RxRB, FGF2) and neurotrophic (VEGF, BDNF) functions, responsible for the observed transcriptomic changes after fluoxetine treatment. Collectively, these results suggest that the early glial functions progressively contribute and lead to neuronal adaptations, as these adaptations are observed closer to the timeline of 10-14 days when behavioural improvement is observed in antidepressant mouse models.

The authors then wanted to study how the glia influenced the neuronal functions. The authors wanted to mimic the physiological events in vivo in a minimal neuron-glia culture system. Hence, they developed a model system by treating primary mixed cortical cultures with the neurotransmitter, serotonin (5-HT), the levels of which are increased in the synaptic cleft upon chronic SSRI treatment. The authors tested whether increasing 5-HT acutely (measured after 8h of induction) or chronically (treated daily and measured on days 2, 5, 7, 10, 14 days after 8h of induction) in cortical mixed cultures (neuron-glia) versus neuronal cultures (only neurons) would show a transcriptional response same as in vivo upon fluoxetine treatment. As a proxy to analyse the molecular response to fluoxetine, they chose Fgf2 and Bdnf which were identified as upstream regulators in their snRNAseq study. Both of these factors were found to be essential for generating an antidepressant response, and have previously been shown to remain downregulated in stressed and depressed animals. Another gene, S100a10 (p11 protein), previously shown to be activated downstream of FGF2 and BDNF signalling, and also regulated by stress and induced in the late phase of fluoxetine treatment (between 9 and 14 days) also acted as a molecular read-out. Analysis of the transcriptional profiles of Bdnf, Fgf2, and S100a10 when analysed by qRT-PCR revealed a gene-specific temporal response (~3-folds for Fgf2 between day 5 and day7; ~3.5-folds for Bdnf between day 7 and day 10; ~2-fold for S100a10 between day 7 and day 10) upon chronic treatment. This temporal response was observed only in mixed cultures and not neuronal cultures substantiating the significant of role of glia-neuron interaction. Using fluorescent in-situ hybridization (FISH), the authors found that Fgf2 stimulation occurred in astrocytes, Bdnf and S100a10 in neurons. Together these results suggest a combinatorial, temporally-regulated response of glia and neurons towards serotonin treatment, wherein astrocyte FGF2 synthesis initiates the serotonin response.

In order to determine if such neuron-glia coupling depended on cell-cell contact or was secretion-mediated,  conditioned media from 5-HT treated primary mixed cortical cultures (5, 7, 10, and 14 days of treatment) was transferred into primary cortical neurons and the expression profile of S100a10 was checked. Conditioned media from prolonged treatment (10 and 14 days), but not from 3 and 7-days of treatment, was sufficient to induce neuronal S100a10 expression, suggesting the presence of neuron-glia derived secreted factors in mediating neuronal adaptations to serotonin/fluoxetine treatment. Since Bdnf expression also peaked by 10 days of 5-HT treatment, the authors hypothesized that BDNF could be one of the molecules secreted by the 5-HT-treated mixed cultures. Hence, they depleted BDNF from the conditioned media before transferring to neuronal cultures, or blocked the binding of BDNF to its receptor by inhibiting TrkB phosphorylation. They found that S100a10 was not stimulated during the two stated conditions, thereby demonstrating that BDNF was one of the secreted components from the mixed cortical cultures at 10 days of 5-HT treatment. The authors further addressed the potential involvement of FGF2 signalling in this network and found a reciprocal cross-regulation between FGF2 and BDNF signaling. FGF2 signalling directly increased Bdnf levels in mixed cortical cultures, but failed to do so in primary neurons. Although Fgf2 was found to be located primarily in astrocytes, FGF2 is known to be secreted and could potentially bind to FGF receptors in any of the glial cells. Hence which glial cell type regulates Bdnf synthesis is yet to be determined. In summary, the authors have identified the central role of neuron-glia communication in response to chronic serotonin availability, initiated by glial FGF2 synthesis and signalling, followed by secretion of glia-derived factors (one of which is Bdnf), these factors then act on neurons to stimulate neuronal remodelling genes and their functions (one of which is S100a10) to produce an antidepressant response.

Figure 1: Mechanism of temporally-regulated mechanism of antidepressant response. Glial cells are the first molecular responders to increased level of serotonin extracellularly. Astrocytes increase the production of FGF2, that in turn allows for elevation in BDNF production by neurons. BDNF signaling increases the neuronal expression of S100a10, thereby bringing about an antidepressant response. Figure created in BioRender.com (Based on figure 5D from preprint).

The reason behind choosing this preprint:
The first line of pharmacological treatment (a stage reached when counselling is rendered ineffective) involving SSRIs has a delayed impact in alleviating symptoms in responding patients. Because the symptoms assessed in cases of depressive disorders are at a behavioural and cognitive level, understanding the cellular and molecular basis of antidepressant response would eventually help in improving the quality of treatment right from the very beginning. Findings of this preprint are also of great importance when it comes to modulating the molecular behaviour of glial cells in treating the disorder.

References:

1. Maletic V, Robinson M, Oakes T, Iyengar S, Ball S, Russell J. Neurobiology of depression: an integrated view of key findings. International journal of clinical practice. 2007;61(12):2030-40.
2. Machado-Vieira R, Baumann J, Wheeler-Castillo C, Latov D, Henter ID, Salvadore G, et al. The timing of antidepressant effects: a comparison of diverse pharmacological and somatic treatments. Pharmaceuticals. 2010;3(1):19-41.

 

Tags: mental health

doi: https://doi.org/10.1242/prelights.30727

Read preprint (1 votes)