OUP user menu

Neuroinflammation and the dynamic lesion in traumatic brain injury

Erin D. Bigler
DOI: http://dx.doi.org/10.1093/brain/aws342 9-11 First published online: 30 January 2013

Until the latter half of the 20th century, moderate to severe traumatic brain injury (TBI) was rarely a survivable event (Masel and DeWitt, 2010). Fortunately, improvements in neuroimaging, and emergency medical and neurosurgical care have dramatically increased survival along with improved understanding and characterization of the acute effects of TBI. With improved survival, it was also assumed that understanding the acute effects of TBI would lead to strategies for better outcome, helpful in rehabilitation and more predictive of recovery or disability.

During this era, standardizations emerged to categorize acute TBI by severity, most commonly based on the Glasgow Coma Scale, whether loss of consciousness occurred, duration of post-traumatic amnesia, presence/absence of skull fracture and neuroimaging findings. Because TBI has an obvious aetiology and onset, classic and well-defined neuropathologies from initial injury through chronic phases have been well characterized, including the macroscopic pathologies identified using neuroimaging. In fact, considerable optimism emerged that, as neuroimaging methods improved, detection of TBI-related pathologies such as cerebral oedema, contusions, intraparenchymal haemorrhages, epidural and subdural haematomas, that better predicted long-term outcome would occur. However, modelling long-term TBI outcome, in particular cognitive outcome, has proved elusive, even taking into account all of the acute TBI variables. Understanding outcome amounts to more than the severity of injury or even the size, type and location of traumatic lesions. What may be missing in TBI outcome models reveals an incomplete understanding of how these lesions influence neural networks of the functioning brain (Caeyenberghs et al., 2012). In other words, how well a patient with TBI returns to pre-injury level of function may depend most on how well functional neural networks survive and/or adapt to the injury. Networks depend on white matter integrity and therefore understanding white matter pathology is crucial in understanding TBI outcome.

As elegantly described many years ago, white matter is particularly vulnerable to injury in TBI. Strich (1956) observed ‘diffuse degeneration of the white matter of the cerebral hemispheres’ (p.163) in patients with severe TBI who survived 5–15 months, but did not have any other gross pathology (i.e. no haematomas, contusions, etc.). A decade later, Peerless and Rewcastle (1967) discussed the vulnerability of white matter damage from TBI as ‘shear injuries’, followed by Adams et al. (1982) introducing the term ‘diffuse axonal injury’. Research on both human and animal diffuse axonal injury led to refinements not only in understanding the time course and distribution of diffuse axonal injury but also the broader concept of axonal damage referred to as traumatic axonal injury (Maxwell et al., 1997; Povlishock and Christman, 1995). White matter disruption and disconnection from traumatically initiated axonal damage led Buki and Povlishock (2006) to state that in TBI ‘all roads lead to disconnection’ (p. 181). It should be emphasized that this disconnection need not be a literal physical disconnection in the sense of axon degeneration, but may also include physiological and synaptic disconnection or decoupling due to compromised axonal function, potentially related to membrane, cytoskeletal abnormalities or other cellular aberrations.

Because TBI begins with an acute event, the neuropathological end-point is presumed to follow a time course of a few months followed by static abnormalities. However, longitudinal studies of neurocognitive and neurobehavioural outcome from TBI demonstrate a highly variable course long after TBI lesions would be assumed to have reached a steady state. For example, McMillan et al. (2012) followed a prospective cohort of >200 TBI survivors to 12–14 years post-injury, observing a dynamic process of changing disability over time. Of particular importance was the observation that level of disability worsened in a substantial number of survivors, with fewer improving once disabled. If the lesion in TBI becomes static, why is there variability in neurobehavioural outcome over time?

Animal studies have shown progressive brain changes well beyond the early chronic stage (Bramlett and Dietrich, 2007), as have neuroimaging studies carried out longitudinally in humans. Ng et al. (2008) used quantitative MRI to compare brain volume loss in patients with TBI with moderate to severe injuries at 4.5 months, a time frame when major brain volumetric changes have already occurred (Blatter et al., 1997), to 2.5 years post-injury. Significant progressive volume loss was observed. In a separate study, Farbota et al. (2012) examined patients with TBI at 2 months, 1 year and 4 years post-injury. White matter volume loss continued up to 4 years post-injury, implicating active neuropathological changes well beyond an assumed static phase. Additionally, Tomaiuolo et al. (2012), in an extended series of single cases followed for >8 years, demonstrated progressive volume loss involving the corpus callosum in severe TBI.

Part of the explanation for these progressive changes, particularly in white matter, far beyond the stable lesion phase may be the result of TBI-initiated chronic neuroinflammatory processes. In this issue of Brain, Johnson et al. (2013) show that chronic neuroinflammatory changes are present in the white matter of the brain and relate to progressive changes in white matter integrity following TBI. In the descriptions of diffuse axonal injury and traumatic axonal injury, acute and subacute inflammatory reactions have been well characterized for some time but active long-term neuroinflammation associated with white matter degeneration in the chronic phase in the human is something new. The authors sampled different regions of the corpus callosum and adjacent parasagittal cortex, in cases with TBI surviving from 10 h to 47 years post-injury compared with age-matched uninjured controls. The corpus callosum may be the most vulnerable white matter region of the brain and potentially the ideal brain region in which to explore these relationships. Furthermore, the corpus callosum is vital to interhemispheric connectivity and overall functional connectivity of the brain. With survival of 3 months or more from injury, some cases with TBI displayed extensive, densely packed, reactive microglia (CR3/43- and/or CD68-immunoreactive), a pathology not seen in controls or the acutely injured cases. Cases that exhibited inflammatory pathology were also the ones with ongoing white matter degradation and callosal atrophy. All of this suggests that neuroinflammatory processes may play an ongoing role within white matter of the injured brain disrupting functional connectivity years after a single moderate to severe TBI.

Neuroinflammation in brain injury has always been viewed as a double-edged sword mediating both short-term beneficial effects for injured parenchyma and neuronal survival but also adverse influences that contribute to secondary brain damage and neuronal loss (Morganti-Kossmann et al., 2002). Johnson et al. (2013) raise many interesting speculations about both acute and long-term outcome in TBI, its treatment and follow-up. Could chronic neuroinflammatory changes in some TBI survivors be responsible for the variability in outcome? It is intriguing to see that only a subset (28% of those who survived >1 year) of the patients with TBI had identifiable indicators of chronic neuroinflammation. Interestingly, as observed by McMillan et al. (2012), only a subset of individuals with TBI change from non-disabled status to disabled over time. Could the presence of neuroinflammatory reactions be a factor related to resilience versus disability following TBI? Is neuroinflammation a marker associated with disability? Similarly, interesting speculations can be made about the role of neuroinflammation in the evolution of post-traumatic epilepsy and the onset of neuropsychiatric disorder post-injury, which may often be delayed for years post-injury. Most importantly, as raised by Johnson et al. (2013), could this be a factor in the association of TBI occurring early in life and linked later with dementia? The heuristic value of these findings provides a foundation for the full spectrum of basic and applied studies on the role of neuroinflammation in TBI outcome.

Returning to the discussion of brain white matter and neural connectivity, how does neuroinflammation disrupt functioning of the network? Is this an active process that affects axonal functioning? Before the moment that shear-tensile forces mechanically deform and damage brain parenchyma, the brain has been a passive partner in whatever is about to happen. Similarly, post-injury the assumption has been that active pathological processes are time-limited. However, as discussed by Amor et al. (2010), there are numerous ways for the damaged neuron and its glial cell environment actively to influence neuroinflammation.

The study by Johnson et al. (2013) suggests that some injured neurons may not be just a passive by-product of the trauma, but active in the neuroinflammatory process well into the chronic phase. If that is the case, the potential for in vivo biomarkers of neuroinflammation may be particularly important, including sequential neuroimaging that quantifies white matter damage. Current MRI techniques including spectroscopy and diffusion tensor imaging may provide in vivo information about neuroinflammation (Harris et al., 2012; Hunter et al., 2012). New neuroimaging techniques involving functional connectivity mapping are already providing important insights into disrupted connectivity from TBI, both acutely as well as chronically (Caeyenberghs et al., 2012; Shumskaya et al., 2012). The combination of neuroinflammation biomarkers with neuroimaging indicators of white matter integrity may provide wonderful insights into TBI outcome and provide better predictive models.

The lesions in TBI may be far more dynamic in their influence over the lifespan than previously assumed. As Masel and DeWitt (2010) suggest, TBI should be viewed as a disease process not just a solitary event. New insights into the chronicity of neuroinflammation in TBI add another feature of complexity to understanding TBI in the context of what Marklund and Hillered (2011) label as ‘the most complicated disease of the most complex organ of the body’ (p. 1207).