Adaptive changes of the ground squirrel retina during hibernation

National Eye Institute

Investigators

Linked publications & trials

Abstract

Axonal injury and subsequent neuronal death underpin the pathology of many neurological disorders from acute neural injuries to chronic neurological diseases. It remains controversial how axonal damage leads to neuronal death, but likely involves interactions between neurons, glia, and local immune cells. Promoting neuronal survival after axonal injury will be of tremendous clinical value, not only for neuroprotection per se, but also to improve the starting point from which attempts aimed at axon regeneration and functional recovery can be initiated. In an effort to develop innovative neuroprotective strategies we turned to the thirteen-lined ground squirrel (TLGS), a mammalian hibernator. Previously, TLGS neurons have been shown to have a unique mitochondrial reaction to cold stress that enable them to preserve neural structure and function. Central to this cold-adaptive response is the suppression of mitochondrial reactive oxygen species (mtROS) production, which has also been implicated in many pathological conditions such as stroke and spinal cord injury. Additionally, mtROS is also closely associated with innate immune responses, which have a substantial impact on neural injury and survival through the process of neuroinflammation. However, it is unknown whether such connection between metabolic adaption and the innate immune response occurs in hibernation and whether it could subsequently affect neural survival against injury. To explore whether the unique metabolic adaptations in hibernation render the neural tissues resilient to axonal injury and to uncover the potential underlying mechanisms of neuroprotection, we selected optic nerve crush (ONC), an axonal injury model without the complications of vascular disruption. More importantly, we developed procedures to perform ONC on hibernating TLGSs without disturbing torpor. In active TLGS, retinal ganglion cells (RGCs) underwent profound cell death following ONC, reaching approximately 80% cell loss 14d post-ONC, whereas torpid TLGS remarkably displayed virtually no cell loss. To probe the differential dynamics and mechanisms of degeneration following ONC, we developed a partial ONC (pONC) model, in which the linear structure of the TLGS optic nerve was crushed only on the nasal side thereby allowing the temporal half to function as an internal control (a nasal-temporal ratio of 0.8 reflects the natural gradient of RGCs). In active TLGS, the time course of cell death after ONC mirrors that seen in rat(25) and mouse(26) whereas torpid TLGS exhibited little RGC loss. While it is possible that lower body temperature (Tb) may slow RGC death and clearance (i.e. Q10 effect), it is unlikely to account for the torpor-associated survival completely, as RGC death was largely absent, rather than delayed, at 21 days and even 6 weeks post-injury in torpid TLGS, while torpid TLGS after ONC occasionally awoke from hibernation still suffered RGC death despite remaining in the cold room. Moreover, torpid TLGS went through multiple rounds of interbout arousal after ONC, during which their Tb rose to 37 C yet did not cumulatively elicit RGC death. Taken together, these observations indicate that torpor bestows TLGS with certain temperature-independent mechanisms that protect RGCs from axonal injury-induced neuronal death. To understand why RGCs in active and torpid TLGSs responded so differently to the same axonal injury, we performed RNA sequencing analyses on injured optic nerve samples (from nerve-head to chiasm). We then categorized genes enriched (not cell-type-specific) with major cell types in the optic nerve (neurons, astrocytes, microglia/macrophage, oligodendrocytes, and oligodendrocyte precursor cells), to provide clues for identifying the cell types that responded most differently to ONC in active vs. torpid conditions. Among differentially expressed genes (DEGs) between crushed active and torpid TLGS samples, we found that 58.8% (80/136) expressed at significantly higher levels in active samples were enriched in microglia/macrophage, whereas only 6.8% (10/148) of DEGs that were higher in torpid samples were microglia/macrophage-associated. Notably, such DEGs included those associated with microglia/macrophage activation such as CD68, CD74, C1QB, and TREM2, thereby suggesting a differential innate immune response to ONC in active vs. torpid TLGS. We then directly examined microglia/macrophages at the ONC injury site. In active TLGS samples, we found massive aggregation of IBA1+ cells, or myeloid cells without differentiating between resident microglia or infiltrating monocytes. Such myeloid cell accumulation started as early as day 1 post-ONC and persisted throughout the observation period (day 21). Moreover, these cells were positive for CD68 labeling as well as for other macrophage/phagocytotic markers, such as F4/80 and MFGE8, and adopted a round, ameboid shape, indicating that they are activated myeloid cells. In stark contrast, such IBA1+ cell aggregation was absent in samples from torpid TLGSs. Instead, a cell-sparse region at the crush site was apparent. These findings corroborate the transcriptomic analysis indicating differential macrophage/microglial responses at the ONC site between active and torpid TLGS. The apparent correlation between microglia responses and RGC death following ONC in active TLGS and the lack of both in torpid TLGS raised the question of why microglia in hibernating TLGS did not react to axonal injury? We are focusing on possible metabolic regulation of microglial activity.