The synapse in traumatic brain injury

Abstract

Traumatic brain injury (TBI) is a leading cause of death and disability worldwide and is a risk factor for dementia later in life. Research into the pathophysiology of TBI has focused on the impact of injury on the neuron. However, recent advances have shown that TBI has a major impact on synapse structure and function through a combination of the immediate mechanical insult and the ensuing secondary injury processes, leading to synapse loss. In this review, we highlight the role of the synapse in TBI pathophysiology with a focus on the confluence of multiple secondary injury processes including excitotoxicity, inflammation and oxidative stress. The primary insult triggers a cascade of events in each of these secondary processes and we discuss the complex interplay that occurs at the synapse. We also examine how the synapse is impacted by traumatic axonal injury and the role it may play in the spread of tau after TBI. We propose that astrocytes play a crucial role by mediating both synapse loss and recovery. Finally, we highlight recent developments in the field including synapse molecular imaging, fluid biomarkers and therapeutics. In particular, we discuss advances in our understanding of synapse diversity and suggest that the new technology of synaptome mapping may prove useful in identifying synapses that are vulnerable or resistant to TBI.

Keywords: astrocyte; inflammation; microglia; synaptome; synaptopathy.

Figure 1

Summary of pathological changes following traumatic brain injury. After TBI, synapse density is reduced for up to 7 days, followed by a recovery which plateaus around 30 days (Scheff et al., 2005). Abnormal accumulation of APP peaks 1 day after TBI but can persist for months to years after injury (Johnson et al., 2013). Tau spreads through the brain after TBI and neurofibrillary tangles are a hallmark of chronic traumatic encephalopathy (Edwards et al., 2020). There is a rapid rise in extracellular glutamate after TBI, and levels may remain abnormally elevated up to a week from the injury (Vespa et al., 1998; Hong et al., 2001). Microglia show bimodal peaks after TBI at 7 days and then 30 days (Simon et al., 2017). Markers of oxidative stress appear rapidly after injury and may persist up to 6 months after TBI (Licastro et al., 2016).

Figure 2

TBI secondary injury and the postsynaptic density. There is a surge of glutamate following TBI (a) (Vespa et al., 1998; Hong et al., 2001), which leads to influx of Ca²⁺ into the cell at the NMDA and AMPA receptors (b) (Nilsson et al., 1996). PSD95, which is bound to PKCα, promotes GluR2-deficient AMPA receptors and exacerbates TBI excitotoxicity (c) (Spaethling et al., 2008). The release of TNFα also acts upon the AMPA receptors and reduces expression of GluR2, which worsens excitotoxicity (d) (Beattie et al., 2002; Stellwagen et al., 2005). Triggered by CR3, microglia lead to long-term depression by activation of NADPH oxidase and GluR2-mediated AMPA receptor internalization (e) (Zhang et al., 2014). PSD95 is bound to neuronal nitric oxide synthase which plays a role in promoting oxidative stress (f) (Brenman et al., 1996).

Figure 3

Role of astrocytes in post-traumatic synapse loss and repair. The upregulation of astrocytic ephrin B1 and STAT3 phosphorylation following TBI have been found to coincide with a decrease in VGlut1 (a) (Nikolakopoulou et al., 2016). The tonic release of astrocytic d-serine has been associated with synapse damage after TBI (b) (Perez et al., 2017). Astrocytic calcineurin/NFAT signalling pathway induces synaptic strength loss in CA1 and a reduction in synaptic protein loss (PSD95 and GluR1) (c) (Furman et al., 2016). MMP3 has been implicated in maladaptive synaptic recovery after TBI (e) (Falo et al., 2006). Hevin has been shown to promote post-traumatic synaptogenesis which was amplified by the blockade of the calcineurin/NFAT pathway (d) (Furman et al., 2016). Thrombospondin 1 (which is regulated by STAT3) has been found to promote synapse recovery after motor neuron injury (f) (Tyzack et al., 2014).