Molecular Mechanisms Regulating the Pro-Regenerative Glial Cell Response to Spinal Cord Injury in Axolotl

Abstract

Axolotl salamanders have the remarkable ability to functionally regenerate after spinal cord injury. In response to injury, glial cells adjacent to the lesion undergo a pro-regenerative response, proliferate and migrate to reconnect the spinal cord and create a permissive environment for axon regeneration. This is in stark contrast to the mammalian response to spinal cord injury. Damaged astrocytes undergo reactive gliosis and contribute to a glial and fibrotic scar by secreting axon growth inhibitory molecules like chondroitin sulfate proteoglycans and collagens. This ultimately results in failed axon regeneration and a loss of sensory and motor function below the lesion. Why the pro-regenerative glial cell response in axolotl is so different from mammalian astrocytes and the identities of pro-regenerative downstream molecular pathways were not well known. To this end, we identified a dynamic change in glial cell membrane potential that was necessary for the pro-regenerative glial cell response to injury. Disruption of glial cell depolarization by either genetic or pharmacologic approaches inhibited the pro-regenerative glial cell response to injury and blocked spinal cord regeneration. Transcriptional profiling and biochemical approaches identified the ERK/c-Fos signaling pathway as key effector molecules downstream of glial cell depolarization. Investigations into the identity of the c-Fos binding partner revealed that JunB, not the canonical c-Jun, is the c-Fos binding partner in axolotl glial cells. While reactive astrocytes in mammals express AP-1cFos/cJun which functions to promote reactive gliosis, glial scar formation, and inhibit spinal cord regeneration, AP-1cFos/JunB represses expression of reactive gliosis associated genes. Therefore, we hypothesized that differential composition of AP-1 could regulate the different cellular responses to injury. Consistent with our hypothesis, the ectopic overexpression of AP-1cFos/cJun in axolotl glial cells leads to defects in axon regeneration, similar to mammals. To determine how glial cells repress c-Jun expression, we identified a miR-200a binding site in the 3’ untranslated region of axolotl c-Jun transcript. Using in vivo and in vitro approaches, we showed that axolotl c-Jun is a direct target of miR-200a. Additionally, inhibition of miR-200a leads to axon regeneration defects reminiscent of the AP-1cFos/cJun overexpression phenotype. Finally, transcriptomic profiling of miR-200a inhibitor-electroporated spinal cords revealed differential expression of a subset of genes involved with reactive gliosis, the glial scar, extracellular matrix remodeling, inflammation, migration, and axon guidance compared to control spinal cords. Collectively these results reveal that miR-200a inhibits signaling networks involved with reactive gliosis, the glial scar, and other processes necessary for spinal cord regeneration. Further examination of the RNA sequencing data revealed that miR-200a inhibition led to the expression of the mesoderm transcription factor Brachyury and down-regulation of a host of neural genes, including Sox2. This expression profile is reminiscent of a more developmentally primitive spinal cord progenitor population called neuromesodermal progenitors. This suggests that miR-200a inhibition led to the loss of the neural identity and acquisition of a more neuromesodermal progenitor-like state. Subsequent analysis revealed miR-200a inhibition indirectly promotes Brachyury expression, specifically in axolotl glial cells, perhaps via modulation of FGF and Wnt signaling molecules. Whether modulation of Brachyury expression is sufficient to induce glial cells to fully dedifferentiate into NMPs and contribute to mesoderm-derived tissues (muscle/cartilage) during regeneration is not clear. In summary, my thesis research identified an injury-induced change in glial cell membrane potential that was necessary for the pro-regenerative glial cell response to spinal cord injury. The injury-induced change in glial cell membrane potential is up-stream of ERK signaling and c-Fos expression. In damaged mammalian astrocytes, c-Fos heterodimerizes with c-Jun to promote reactive gliosis and glial scar formation. However, a majority of axolotl glial cells do not express c-Jun and instead c-Fos heterodimerizes with JunB to form AP-1cFos/JunB. AP-1cFos/JunB functions to inhibit reactive gliosis, glial scar formation, and promote the pro-regenerative glial cell response. Ectopic overexpression of AP-1cFos/cJun in axolotl glial cells inhibits spinal cord regeneration. Axolotl glial cells express the microRNA miR-200a, which functions to repress c-Jun expression. Inhibition of miR-200a blocks spinal cord regeneration, leading to differential expression of genes involved with reactive gliosis and glial scar formation. Finally, miR-200a may play an additional role in stabilizing the neural identity of neural progenitor cells during axolotl spinal cord regeneration. Inhibition of miR-200a could result in dedifferentiation of glial cells to a more neuromesodermal progenitor-like identity, perhaps by modulating Wnt and FGF signaling.