De rol van bio-energetica bij letsel-geïnduceerde axonale hergroei: glycolyse als sleutelmechanisme

The regenerative capacity of the adult mammalian central nervous system (CNS) is limited. Following injury, most neurons die, and the few survivors are unable to regenerate their axons, leading to irreversible damage. Regrowth is limited first by neuron-extrinsic mechanisms, such as glial scarring and the release of growth-inhibiting molecules following injury. Second, even if these are cleared, the growth potential of adult CNS neurons is severely restricted by neuron-intrinsic mechanisms, such as the progressive decrease in expression of regeneration-associated genes during development. Decades of research have culminated in the discovery of genetic interventions able to partially counteract such inhibition. One of these is the deletion of phosphatase and tensin homolog (Pten) and suppressor of cytokine signalling 3 (Socs3), which leads to robust axonal regrowth in both the optic nerve and spinal cord. Nonetheless, full circuit restoration and functional recovery remain unachievable. According to the established paradigm of axonal regeneration, mitochondria accumulate at the growth cone and provide energy to fuel and support axonal elongation. Unfortunately, mitochondrial transport is progressively decreased during neuronal maturation, with increased docking at the site of synapses. For this reason, strategies to induce axonal regrowth so far aimed at restoring mitochondrial transport to fuel the regeneration process. Nonetheless, as injury has been demonstrated to depolarise the local population of mitochondria, leading to their dysfunction and subsequent depletion of axonal ATP, an important knowledge gap remains: how can axons withstand the initial energy crisis caused by injury and overcome it to prevent degeneration and initiate regrowth? To answer this question, we aimed to expand our understanding of the metabolic adaptations that are necessary within CNS axons to facilitate a regrowth response. This entailed an in-depth characterization of mitochondrial dynamics and intracellular bioenergetics, which in turn requires longitudinal assessment of the neurons, live imaging, and the ability to chemically interfere with the metabolic network in a precise and temporally controlled manner. All of these would be extremely challenging in vivo, especially in the optic nerve. Therefore, we established a culture of primary retinal cells from postnatal Ptenfl/fl;Socs3fl/fl mice in microfluidic devices. We demonstrated that matured postnatal retinal ganglion cells (RGCs) exhibit a phenotype comparable to adult ones in terms of activation of regeneration-associated pathways, or the lack thereof. Moreover, multicompartment microfluidics allowed us to separate the neuronal cell bodies in the somatic channel from their axons, which can grow through the microgrooves and extend into the axonal channel. This fluidic isolation enabled clean axonal injuries and independent treatment of the axonal and somato-dendritic compartments. Using this microfluidic setup, we observed increased regrowth and enhanced mitochondrial trafficking in the axons of retinal ganglion cells with the deletion of both Pten and Socs3. While wild-type axons relied on mitochondrial metabolism, after injury, in the absence of Pten and Socs3, energy production was demonstrated to be supported by local glycolysis. Limiting the glycolytic flux with galactose medium impaired energy production in Pten and Socs3 co-deleted axons and completely reversed their regeneration phenotype, both in terms of the number of regenerating axons and the length of regrowth. Together, these observations reveal that glycolytic ATP, combined with sustained mitochondrial transport, is essential for injury-induced axonal regrowth. To validate and narrow down our findings to a set of critical molecular players, we proceeded to identify a transcriptomic basis in vivo for the metabolic adaptations we identified with our functional data in vitro. We did so by leveraging a previously published single-cell RNA sequencing dataset of injured RGCs, which included the deletion of Pten and Socs3. We re-analysed this dataset with a focus on metabolic pathways. We found that RGCs co-deleted for Pten and Socs3 promptly undergo a switch in the expression of metabolism-related genes after optic nerve crush injury. This includes the increased expression of antioxidant genes and, most importantly, the upregulation of genes involved in glycolysis, lactate production and NAD salvage. These findings align with our in vitro data, showing that the glycolytic pathway plays a pivotal role in counteracting the energetic deficiency induced by axonal damage and in providing ATP to induce regeneration. As lactate production links glycolysis with the maintenance of the NAD redox state, we inhibited it after injury in our microfluidic setup, specifically within the axons. This reverted the phenotype induced by Pten and Socs3 co-deletion in terms of the percentage of axons surviving injury and initiating regrowth. On the other hand, axon elongation was not affected. As such, while our previous findings show that the entire glycolytic pathway plays an important role in all phases of regeneration, lactate production is required only to support the initial injury response and induction of regrowth. In addition to the experimental work, we also developed image analysis tools that enabled the study of RGC phenotypes at a single-cell level. This first coincided with the development of RGCode, a deep learning pipeline that fully automates the counting of RGCs on retinal RBPMS-stained whole-mounts, measures the retinal area and calculates RGC density. The generation of this deep learning model allowed us to later develop new quantification tools with minimal training data via transfer learning. A prominent use within this thesis was the generation of a model to segment the somata of cultured RGCs, allowing us to fully automate the measurement of immunolabelled regeneration-associated markers or quantify FRET signals at a single-cell level. Collectively, this PhD work provides a set of tools aimed at the study of neuronal bioenergetics and presents data that constitute a significant advancement in the understanding of the metabolic underpinnings of axonal regrowth. We elect glycolysis as a crucial pathway for successful neuroprotection and injury-induced axonal regeneration in mammals, and as a potential avenue for novel strategies to induce circuit repair. Nonetheless, this does not negate the critical role of mitochondria in this process. The shift in perspective provided by our findings can be the starting point to reevaluate their role in axonal regrowth. This might lie beyond the simple generation of energy and instead among the secondary contributions of mitochondria to neuronal metabolism, such as the synthesis of macromolecule precursors.