SRF regulates the generation of neuroprotective astrocytes in the brain

Abstract

In response to injuries, infections or in neurodegenerative disorders, astrocytes get activated to become reactive. This phenomenon is called astrogliosis and is marked by a spectrum of changes which encompasses structural, functional and genetic changes in astrocytes. Until now, the molecular mechanisms regulating astrogliosis remain poorly understood. Traditionally, upregulation of GFAP and cellular hypertrophy were considered to be the markers of astrogliosis. But these markers vary with respect to the degree of astrogliosis and studies are ongoing to fully characterize and understand astrogliosis. On the basis of severity, astrogliosis is broadly divided into three categories – mild to moderate, severe diffuse and severe astrogliosis with compact glial scar formation. Mild to moderate astrogliosis is marked by upregulation of GFAP and several other genes. Here, reactive astrocytes become hypertrophic while maintaining their domains. Once the neural insult is resolved, these astrocytes revert back to their non-reactive states. In severe astrogliosis, there is more prominent upregulation of GFAP along with other genes and more hypertrophy of astrocytes as compared to mild to moderate astrogliosis. These reactive astrocytes lose their territories and their processes merge with each other. This may lead to permanent changes in the architecture of the brain. Severe astrogliosis with scar formation includes similar changes as severe astrogliosis including enhanced GFAP expression and hypertrophy of astrocytes, but these changes become much more pronounced. In addition, these reactive astrocytes proliferate and their processes overlap extensively to form a compact, dense and narrow glial scar. Here, the tissue undergoes prominent structural changes, which are long-standing and are likely to persist even after tissue repair is complete. Our lab had previously shown that conditional deletion of the stimulus-dependent response factor, serum response factor (SRF) in neural precursor cells (Srf-Nestin-cKO mutant mice) leads to a significant loss in astrocyte and oligodendrocyte numbers both in vivo and in vitro. This suggests a critical role for SRF in glial specification. However, whether astrocyte differentiation is decreased or delayed in the Srf-Nestin-cKO mice was not clear since these mutant mice exhibited neonatal lethality and astrocyte differentiation peaks around postnatal days 2 to 7. To further address this, SRF was deleted specifically in astrocytes using a GFAP-Cre transgenic mouse line. In the resultant conditional mutant mice, Srff/f;GFAPCre (SRF-GFAP-cKO), astrocytes numbers were significantly reduced at birth similar to that observed in the Srf-Nestin-cKO mutant mice. However, at 3 weeks of age, SRF-GFAP-cKO mutant mice exhibited astrogliosis in the brain, with hypertrophic astrocytes exhibiting increased expression of GFAP, vimentin and nestin. The reactive astrocytes observed in all regions of the brain in SRF-GFAP-cKO mice persisted throughout adulthood and were accompanied by reactive microglia. Previous studies have shown that depletion of microglia prevented the generation of neurotoxic A1 astrocytes. We found that depletion of microglia following treatment with pexidart 5562, PLX5562 (a drug that depletes microglia by inhibiting colony stimulating factor 1 receptor signalling), did not affect astrogliosis in SRF-GFAP-cKO mice suggesting that the gliosis state of SRF-deficient reactive astrocytes is not dependent on microglia. SRF-GFAP-cKO mice exhibited reduced astrocyte numbers at birth and these astrocytes later became reactive. A likely cause for gliosis in these mutant mice is that the astrocytes became reactive in response to this change in astrocyte numbers and any accompanying change in tissue homeostasis. Or it could reflect a genetic requirement of SRF for maintenance of non-reactive astrocytes. To address this, we generated a tamoxifen-inducible conditional mutant mouse, Srff/f;GFAPCreERT2 (SRF-GFAPERT2-cKO). In the SRF-GFAPERT2-cKO mutant mice, Cre recombinase is expressed in astrocytes starting embryonic day 16.5 but Cre-mediated recombination is temporally regulated by tamoxifen. Tamoxifen administration to 4-8 wk old SRF-GFAPERT2-cKO mice resulted in astrogliosis at 2 months post-injection and the reactive astrocytes also persisted throughout adulthood. No cell death or cell proliferation was observed in either SRF-GFAP-cKO or SRF-GFAPERT2-cKO mice suggesting that cell death is not the cause for reactive astrocytes and that these astrocytes are not proliferative. We then asked whether normal functions of astrocytes such as maintenance of blood brain barrier is affected in these SRF mutant mice, We found that the blood brain barrier was intact and did not exhibit any leakiness in both the lines of SRF mutant mice. Neurotoxic A1 astrocytes have been shown to kill neurons and oligodendrocytes in vitro and were severely compromised in their normal functions such as secretion of pro-synaptogenic factors, maintenance of synapses and synaptic transmission. In contrast, we did not observe any cell death in these SRF mutant mice and expression of astrocyte secreted pro-synaptogenic factors were unaffected. Furthermore, we observed no deficits in basal synaptic transmission and LTP in SRF-GFAPERT2-cKO mice in both young and aged adults. Gene expression analyses by quantitative RT-PCR of mRNA obtained from the neocortex of SRF-GFAP-cKO mutant mice showed a predominant expression of neuroprotective A2 reactive astrocyte markers as compared to neurotoxic A1 reactive astrocyte markers. These results together indicated that SRF-deficient astrocytes are unlikely to be neurotoxic and may be neuroprotective. We next asked whether SRF-deficient reactive astrocytes can provide neuroprotection using two models of neural injury: stab-wound injury and kainic acid-induced neurotoxicity. We found that control mice exhibited significant cell death in all regions of hippocampus at 7-days post-kainic acid administration. In contrast, SRF-deficient mice showed significantly decreased cell death. Also, compared to control littermates, SRF mutant mice exhibited smaller wound cavity at 7-10 days post-stab wound injury that suggested faster wound-healing. Together, our findings suggest that SRF regulates the generation of neuroprotective astrocytes in the mouse brain. Identification of SRF-target genes may provide novel therapeutic targets for promoting efficient neuronal recovery following traumatic brain injury and epilepsy. Previous studies have shown that astrocytes play a role in promoting the differentiation of oligodendrocyte precursor cells (OPCs) into oligodendrocytes during development. Astrocytes influence both differentiation of OPCs and myelination of the damaged neurons in the injured and/or diseased adult brain negatively as well as positively. While characterizing the SRF-GFAP-cKO mutant mice, we observed severe deficits in myelination in 3 week old mutant mice. The myelination deficits became more prominent with time and ultimately resulted in loss of white matter as was evident from the absence of corpus callosum and poorly connected hippocampus in older mutant mice (3 months and older). Subsequent analysis revealed a significant reduction in the number of oligodendrocyte precursor cells (OPCs) in the brains of 2-wk old SRF-GFAP-cKO mice as compared to control mice. However, there was no evidence of Cre expression and SRF deletion in OPCs in these mice. In contrast to SRF-GFAP-cKO mutant mice, postnatal ablation of SRF in tamoxifen injected SRF-GFAPERT2-cKO mutant mice did not show any discernible deficits in myelination. Together, our findings suggest that SRF-deficient astrocytes may regulate OPC differentiation and/or maturation, and myelination in a paracrine manner during early development in the mouse brain.