Keywords

The clinical diagnosis of concussion largely relies on external indicators of neurological injury for diagnosis. The ever-evolving definition of concussion has been described as “a clinical syndrome of biomechanically induced alteration of brain function typically affecting memory and orientation, which may involve loss of consciousness” [1, 2]. Despite the vague clinical description, advances in imaging technologies, the development of clinical biomarkers, and preclinical animal studies have been used to identify subtle but significant microstructural abnormalities and identified molecular cascades occurring in the absence of macroscopic injury to the brain, which collectively contribute to changes in cognitive status. These pathophysiological events highlight the contribution of concussions as an important contributor to the spectrum of traumatic brain injury (TBI) despite being classified as a mild traumatic event. As technologies evolve, the ability to reliably and accurately detect subtle cellular changes in cerebral structure and function gives hope for the more accurate diagnosis of injuries, with potential prognostic and therapeutic applications. Understanding the link between biophysical transduction of force from tissue to cellular and even ultrastructural scales for an injury type that is notoriously heterogeneous is challenging from a preclinical modeling perspective. In turn, this challenge is amplified when it comes to developing therapeutic strategies [3].

This chapter focuses on our current understanding of the pathophysiology associated with these processes and their contribution to the evolving pathophysiology of concussion. Although many of the mechanisms described below are also applicable to moderate and severe TBIs, it can be assumed that these cellular processes are also shared with mild TBI albeit at a lesser degree. There are, however, notable differences between injury severities which further suggest that not all mechanisms of secondary injury are applicable or relevant across the injury spectrum of TBI. Preclinical models have helped to bridge the gaps in knowledge from the limitations of noninvasive methods in the clinic and provide a better understanding of the discrete molecular and cellular events unfolding within the brain. Although developing clinically relevant, accurate, and reliable models of clinical concussion remains a challenge, animal models, computer simulations, and in vitro systems remain an important component of understanding the pathophysiology of concussion [4, 5].

Normal and Abnormal Physiological States

It is interesting to consider that even under normal conditions there exists considerable pliability and tolerance within an otherwise static central nervous system (CNS). For example, the very act of bending the spine or rapid rotation of the head involves some substantial biomechanical force transduction to axons within the CNS. Body movement lends itself to some degree of nerve fiber stretching and bending or inertial motion of otherwise static tissues but ultimately returning to a normal state [6]. In vitro studies have demonstrated that axons have a high tolerance to stretch and exhibit a delayed elasticity phenomenon, whereby axons subject to up to 65% tensile strain generally recover their original shape afterward [7].

Similarly, the brain is largely protected from inertial acceleration by the fluidity of the brain encased in a rigid skull surrounded by cerebral spinal fluid (CSF). Dramatic and instantaneous acceleration or deceleration (<50 ms) beyond the cushioning capacity of the CSF generates substantial linear and rotational shear forces that result in axonal injury [8]. In particular, the anisotropic properties of white matter tissue in the brain lend itself to higher levels of strain and shear depending on the direction and magnitude of the force being applied [9,10,11]. Microcompartment in vitro studies further support the argument that axons or neurites are more susceptible to injury than the cell soma [12]. In the same vein, computational models of axons that incorporate cytoskeletal dynamics between actin and spectrin, as well as the anchoring protein, ankyrin, point to greater intrinsic stiffness in axons relative to the cell bodies which further imply increased susceptibility to injury relative to other brain structures [13]. While imaging studies of mTBI patients support the vulnerability of white matter in concussion, the underlying pathophysiology remains a complex issue to resolve (Fig. 2.1).

Fig. 2.1
figure 1

(a) A schematic representation of the coronal section of a human brain. At the macroscopic level, rotational forces resulting from head trauma can translate into inertial movement of the brain resulting in deformation of tissues and brain structures. These forces include shear strain on white matter tracts that are particularly vulnerable to excessive inertial loading. (b) At a microscopic scale, the biomechanical transduction of forces on white matter tracts can result in rapid stretching of axons regardless of the orientation of fibers. Studies suggest that shear strain rarely results in primary axotomy. The majority of axons return to their normal state. However, a subset of axons develops morphological characteristics associated with diffuse axonal injury. These include abnormal morphology, such as axonal swelling and varicosities associated with activation of molecular pathways and disruption of cytoskeletal networks

Full size image

Imaging White Matter Injury in Concussion

Diffuse axonal injury (DAI) is a common component across all TBI severities [9, 14,15,16]. However, the dispersed and subtle structural alterations associated with concussion are inherently difficult to verify clinically, despite being the most likely contributor to morbidity after injury. Computed tomography (CT) and magnetic resonance imaging (MRI) are largely limited to the detection of macroscopic abnormalities and lack the sensitivity to detect the subtle white matter changes that may occur with concussions [17, 18]. Biomarker studies, improved imaging technologies, and animal models of TBI have demonstrated that the absence of overt changes in brain following mTBI and concussion does not exclude the occurrence of pathophysiological changes which contribute to morbidity.

Diffuse tensor imaging (DTI) is an advanced MRI sequence that provides a more sensitive technique for identifying injuries to white matter tracts [18]. The underlying principle is that DTI estimates rates of water diffusion in different spatial directions, with white matter tracts exhibiting highly directional, anisotropic diffusion parallel to fiber bundles [18, 19]. Its application in TBI involves the detection of changes in diffusion properties within axonal fiber bundles, which presumably indicate some form of physical anomaly in the white matter tracts [20]. While there is no doubt that DTI provides some information on structural disturbances to the brain after TBI , to date, the most effective application of this relatively new imaging modality has been in moderate and severe TBIs. Its potential application in concussion is still in proof-of-concept stage [21, 22]. Although noninvasive imaging of mTBI pathology has not been fully translated into clinical utility [23], there is well-documented postmortem evidence of axonal pathology associated with TBI that continues to be bolstered by complementary preclinical animal studies [5, 14]. Bridging the uncertainty of pathophysiological mechanisms in human cases of concussion with what is known from more severe forms of TBI and animal models will help further validate the use of imaging modalities in characterizing concussion pathophysiology.

Axonal Injury at the Microscopic Level

The collective contribution of subcellular structural alterations, metabolic changes, and ionic shifts in the brain results in impairment of neurotransmission that contributes to the morbidity associated with concussion [24]. Computational modeling conducted at the tissue, cellular [25, 26], and ultrastructural levels [27] has demonstrated the physical transfer of strain, shear, and stresses from macroscopic to microscopic scales in greater detail [8]. These studies are useful in conjunction with histological studies which highlight the complex physical relationship between individual axons within white matter bundles [28] and other constituents of the brain including glia, myelin sheathing, and vascular cells [6]. Differing viscoelastic properties among varying cell types and structures create a complex environment when transduction of biomechanical forces is taken into consideration.

The axonal membrane demonstrates a spectrum of injury responses dependent on the force of injury. For example, in fluid percussion injured cats, mTBI was shown not to result in disrupted axonal membranes but was still associated with neurofilament disruption and compaction. However, with increasing injury severity, axonal membrane permeability becomes a pathological feature of injury [29,30,31]. Similarly, in vitro studies have demonstrated that mammalian axonal membranes display a remarkable tolerance to stretch injury, with uptake of low molecular weight fluorescent dyes only in instances of higher strain rates resulting in primary axotomy [7]. A direct translation of critical injury thresholds between in vitro and in vivo injury models is difficult to implement, due to the simplicity of in vitro systems compared to the heterogeneity of cells and tissue composition of in vivo systems. However, both injury models agree that primary axotomy resulting from initial strain or impact is in fact a rare occurrence. Experimental evidence suggests that axonal injury or disconnection is an evolving and progressive condition [10, 28]. Human postmortem studies have confirmed that the development of neurofilament misalignment and the formation of retraction bulbs is a delayed process even in severe cases of TBI which suggests that primary axotomy is a rare occurrence at the time of impact [32]. Detailed in vitro investigation has demonstrated that stretch injury conditions that mimic in vivo injury forces result in a heterogeneous response among cytoskeletal elements. For example, early periodic breaks in the microtubule structure were reported [33], while fast transported beta-amyloid precursor protein (β-APP) and slow transported heavy neurofilament (NF200) demonstrated different morphological deposition patterns along injured axons [34]. The described process of events involves the disruption of microtubule networks which are generally regarded as rigid structure within axons and pathways for axonal transport [10, 14]. The misalignment and disruption of the microtubule network set into motion a molecular sequence of failure in axonal transport, resulting in the accumulation of proteins within axons and the formation of axonal varicosities. Collectively, these studies suggest that concussions are likely not a cause of primary axotomy, but rather a sequence of molecular events that unfolds after trauma leads to axonal dysfunction or disconnection. Understanding these processes and their course of evolution is paramount to the development of therapeutic strategies.

Despite a relative resilience of the axonal membrane to shear forces, at the subcellular level the complex organization of the cytoskeleton involves dynamic reorganization in response to structural disruption [6]. The response to cytoskeletal disruption can manifest in the form of axonal swellings. Similar to a motor accident on a highway creating congestion as incoming traffic continues to backlog, transported proteins accumulate within disrupted networks of microtubules and neurofilaments that no longer support continuous paths to traverse. This characteristic histopathology finding is a hallmark for diffuse axonal injury resulting in axonal swellings and eventually disconnection bulbs [35]. β-APP has been used as a general histopathology marker of axonal injury in this regard [36]. Mild TBI has been shown to result in deposition of β-APP in human postmortem studies [37]. These findings have also been replicated in a swine model of rotational acceleration injury [38]. Furthermore, these animal studies provide evidence that axonal injury can occur in the absence of loss of consciousness [38], leading to a greater appreciation of the sensitivity of the brain to mild traumatic forces.

Mechanisms of Ionic Dysregulation and Calcium Pathways

Based on the disruption of cytoskeletal networks and the relative impermeability of cell membranes to primary axotomy, the question arises as to how cytoskeletal disruption is initiated. Although primary physical disruption likely plays some role, it is the delayed evolution of axonal pathology that suggests active cellular mechanisms are involved in the development of axonal pathology. One major event in concussion at the cellular level is shifted in ionic equilibrium within neurons and glia which lead to dysregulation of calcium signaling pathways with detrimental consequences. Large early increases in extracellular glutamate and potassium posttrauma have been demonstrated in rat models of mTBI [39, 40]. The altered neurotransmitter levels and ion flux coincide with spreading waves of depolarization observed experimentally [41] as well as clinically [42, 43]. The increased concentration of extracellular glutamate is presumably responsible for large influxes of extracellular calcium mediated by ionotropic N-methyl D-aspartate (NMDA) receptors [44], which further contribute to cellular depolarization through voltage-gated sodium channels. Interestingly, sublethal stretch injury in vitro in primary neuronal cultures has been shown to increase susceptibility to subsequent NMDA challenge [45]. Although performed in vitro, these findings suggest a state of increased vulnerability of the brain to repeat concussion.

Alterations in ion concentrations within intact axons have also been shown in vitro to be due to the activation of mechano-sensitive voltage-gated sodium channels [46]. The proposed sequence of events involves influx of sodium resulting in subsequent activation of voltage-gated calcium channels as well as reversal of sodium–calcium (Na+-Ca2+) exchangers [47]. The net effect is an increase in intra-axonal calcium, which results in the activation of calcium-sensitive proteases, such as calpain, or further release of intracellular calcium stores from the mitochondria or endoplasmic reticulum leading to a feedforward process of calcium overload [48]. While these mechanisms hold true for more severe forms of axonal injury, other data indicate that intracellular calcium stores are likely involved in the initial calcium spike observed after mild trauma [49]. Moreover, there are biphasic responses to calcium dynamics in the cell, suggesting that the initial phase is a result of intracellular calcium, while subsequent calcium waves are propagated through primarily extracellular sources [49, 50]. These data highlight the complexity and ever-changing physiology of cell receptors and signaling in response to injury and further highlight the difficulties in treating these conditions.

Calcium plays a critical role in numerous cell functions and acts as a linchpin in the activation of pathophysiological processes linked to axonal and neuronal degeneration [51]. The numerous processes initiated by calcium include excessive protease activation, phosphatase activation, initiation of apoptosis, mitochondrial failure, and the reversal of sodium–calcium exchangers leading to membrane depolarization [48, 52]. Importantly, the route of calcium entry, also referred to as the “source specificity hypothesis,” [44] determines subsequent activation of discrete downstream signaling pathways adding another layer of complexity to the calcium signaling process [48, 53].

One important consequence of calcium influx is the activation of calpains, a family of cysteine proteases with numerous cell functions including cytoskeletal remodeling, cell signaling, differentiation, and vesicular trafficking under physiological conditions [53, 54]. The two calpains relevant to diseases in the CNS are activated at micro- and millimolar concentrations of calcium [55], denoted as either μ-calpain, m-calpain, or calpain-1 and calpain-2, respectively [56]. While studies suggest that TBI results in influxes of calcium in the millimolar range [51], it has been difficult to parse out the individual contributions of each isoform due to overlapping cleavage targets and a lack of isoform-specific inhibitors [56]. However, several studies point toward μ-calpain as the mediator of neurodegeneration after insult [57]. Regardless, calpain activation is associated with the cleavage of alpha II spectrin, an important structural component of the axonal cytoskeleton [58]. The breakdown of intact 280 kDa alpha II spectrin to a lower molecular weight product of 145/150 kDa is a hallmark indicator of calpain-mediated axonal injury [59]. The specificity of calpain’s cleavage site makes it an attractive target as a clinical biomarker of axonal injury as it has been shown to be detectable in CSF after injury [59]. Moreover, alpha II spectrin is further cleaved into a 120 kDa product associated with cleaved caspase-3 activity [60]. Although concussion typically does not result in significant neuronal loss, there is evidence that apoptosis can occur in some instances [61]. Thus, alpha II spectrin may be useful to delineate the underlying injury pathways of axonal injury or delayed cell death occurring after mTBI (Fig. 2.2).

Fig. 2.2
figure 2

Cartoon depicts some of the molecular mechanisms described in in vitro studies believed to contribute to secondary axonal injury after trauma. Mechano-sensitive voltage-gated sodium channels are activated by physical insult resulting in sodium ion influx into axons. The increased sodium ion concentration in axons causes reversal of sodium–calcium exchangers. Increased calcium ion influx can mobilize calpains resulting in proteolysis of cytoskeletal proteins such as alpha II spectrin. Shedding of spectrin into the interstitial spaces finds its way into peripheral circulation and can be detected in some instances as a serum biomarker. Physical damage to microtubule networks results in the disruption of axonal transport. Beta-amyloid precursor protein accumulates within axons due to transport failure resulting in axonal swelling. This has been shown to result in axonal disconnection

Full size image

Military Concussion

Military operations over the last two decades have highlighted a shift in warfare injuries, with a larger percentage attributable to survivable blast exposure and head trauma [62]. Blast exposures and mTBI contribute to a large segment of the brain-injured cohort and have become increasingly recognized as a potentially unique form of mTBI [63]. Evidence suggests that non-fatal blast exposure in the absence of contusion also results in axonal injury but with patterns distinct from conventional head trauma [64]. Data from animal models of subclinical blast trauma point toward significant and delayed alterations in heavy neurofilament expression, which is in turn associated with impaired electrophysiological function in white matter and behavioral impairments [65, 66]. In these studies of mild non-blunt force trauma, calpain-mediated breakdown of alpha II spectrin has been observed in the absence of neuronal cell loss [66, 67]. Mild blast studies have also demonstrated neurovascular changes including disruption of the blood–brain barrier (BBB) involving degeneration of astrocytic endfeet following blast exposure [68, 69]. There is also evidence to suggest that soldiers exposed to blast injuries develop deposits of phosphorylated tau, a hallmark indicator of chronic traumatic encephalopathy [70]. Although tau deposition is generally attributed to repetitive head trauma, there is well-documented postmortem and in vivo imaging studies describing tau deposition years following a single TBI event highlighting the complexity of the response to various modalities of TBI [71, 72]. It should also be noted that blast modeling in animals, particularly the defining criteria for mild or low-level blast and its correlation to human low-level blast exposure, is not well-established or standardized [73]. Moreover, blast injuries in human cases are often confounded with comorbidities including post-traumatic stress disorder (PTSD) and limited to imaging studies which provide little insight into the underlying molecular mechanisms of injury.

The Biology of Biomarkers for Concussion

Biomarkers have the potential to noninvasively diagnose the presence of concussion. However, their clinical application has been met with limited success. The shortcomings of biomarker utility in concussion are in part due to the broad clinical definition based on external symptoms, while the underlying physiological criteria have not been specifically defined [74]. This disconnect is also complicated by evidence pointing to the pathophysiology of concussion extending beyond the window of clinical symptoms [75,76,77]. These imply a disconnect between the current gold standard of clinical diagnosis for concussion and the molecular tools meant to supplant this standard. Despite these challenges, the principles of biomarker development are based on our understanding of cellular and subcellular changes occurring in neural tissues in response to trauma which are presumably important contributors to clinical outcome.

Cleavage substrates of calpain are presumably released from injured neurons or axotomized axons and find their way into the CSF and blood. These biomarkers provide insight into proteolytic cleavage targets and an opportunity for the development of a noninvasive diagnostic tool. Neurofilaments have also been evaluated clinically as biomarkers of axonal injury and provide some insight into the pathophysiology of axonal injury after concussion. Neurofilaments subtypes consisting of light, medium, and heavy chains are the largest contributors to the intermediate filament family that make up the neuronal cytoskeleton along with actin microfilaments and the larger caliber microtubules [78]. In animal models, there is a demonstrated correlation between injury severity and the amount of detectable serum and CSF phosphorylated heavy neurofilament [79]. Serum presence of neurofilament light (NFL) chain after repetitive head injury has been shown to correlate with injury severity in repetitively concussed athletes [80]. Serum presence of NFL chain also correlates with CSF levels. Interestingly, elevated NFL was detected up to 3 months post-injury in boxers [81], suggesting a prolonged shedding or injury to axons, which further supports that notion of concussion as an ongoing neurodegenerative process.

A common theme in studies of white matter injury is that heterogeneity exists among axons, not only in composition (e.g., myelinated vs unmyelinated), but also in their response to injury. In vitro studies have demonstrated that the neurofilament response to stretch varies depending on the degree of mechanical insult applied [82]. The varied response to different degrees of mechanical injury in vitro is consistent with in vivo observations which indicate differing immunoreactivity subtypes following TBI [83]. Similarly, neurofilament compaction, believed to be a result of phosphatase activity on the sidearm structures [84], does not occur in the same axons as those exhibiting microtubule destabilization and impaired axonal transport [85,86,87]. The varying axonal response to injury across these studies were reported under conditions of moderate to severe modeled TBI and whether these mechanisms are involved in concussion remains to be further elucidated. However, these studies are valuable for demonstrating the heterogenic axonal response to TBI.

Mild stretch injury induces increased neurofilament immunoreactivity in axons [82] which also correlates with increases in heavy neurofilament expression observed in fluid percussion models and mild blast in vivo models [66, 88]. Similarly, increases in phosphorylated heavy neurofilament have been detected in serum samples from boxers [89]. Phosphorylated heavy neurofilaments are predominantly localized in long axons [90], which make them a potential marker for axonal injury. Numerous clinical and animal studies have examined the prognostic and diagnostic value of serum and CSF biomarkers. However, their meaningful application in human concussion and mTBI remains to be further clarified due to technical hurdles in terms of thresholds of detection and whether the select measures are truly indicative of underlying pathophysiology [91, 92] (Fig. 2.3).

Fig. 2.3
figure 3

In vitro studies suggest that axon membrane permeability is not a common occurrence in mTBI. Thus, calcium entry either occurs from reversal of sodium–calcium exchangers or secondary calcium release from the axonal endoplasmic reticulum. Other observed mechanisms of calcium-induced injury to axons include activation of phosphatases resulting in loss of sidearms projections in neurofilaments. Sidearm projections are believed to be responsible for maintaining axon caliber and their loss results in neurofilament compaction contributing to changes to axon diameter. Phosphorylated neurofilaments have been detected in serum samples from concussed patients. While it is assumed that membrane disruption is responsible for the shedding of proteins such as neurofilaments and GFAP, there remains the issue as to how these proteins end up in peripheral circulation given that concussed patients and mild TBI animal models demonstrated little evidence of axotomy

Full size image

Biomarker studies have been useful in identifying some generalized cellular changes in addition to neurofilament shedding and disruption. One such example is the detection of glial fibrillary acidic protein (GFAP). GFAP is an astrocytic scaffolding protein whose presence in serum has recently been shown to correlate with MRI abnormalities [93] and has been detected in serum samples of mild and moderate TBI patients [94]. As with all biomarkers of TBI, the reliability of GFAP as an indicator of mTBI is not firmly established. For example, no reported changes in GFAP were found in serum samples from Olympic boxers, while phosphorylated tau, a suspected indicator of chronic traumatic encephalopathy, was detected [95]. There are numerous reasons for the discrepancies between these findings including timing of sampling, type of injuries sustained, and methods used for detection. Similar to the issues surrounding the clinical usefulness of DTI in the detection of concussion, further studies are required to better understand the physiology and temporal sequence of protein shedding into the blood from the CNS for the application of biomarkers to be both reliable and accurate [91, 96].

Repetitive Concussion and Implications for Neurodegenerative Diseases

Animal studies examining GFAP immunoreactivity after mTBI have demonstrated an increase in GFAP expression after injury and increased the presence of microglial activation [97]. Not surprisingly, repetitive head injury in rodents also demonstrates increased the presence of GFAP expression and immune reactivity in brain tissue [98]. In addition to increase in GFAP expression in reactive astrocytes, increased presence of phosphorylated tau (p-tau) protein in rodent models [98] is consistent with the neuropathology of chronic traumatic encephalopathy (CTE). CTE is the term used to describe the neurodegenerative condition characterized by perivascular neurofibrillary tangles of p-tau which frequently occurs in tandem with tau expressing astrocytic tangles seen in a significant number of repetitively head-injured cases [99, 100]. Although there is no clear consensus on whether CTE constitutes a distinct clinical condition [101] or is part of a larger spectrum of tauopathies, the diagnosis for CTE can only be confirmed through postmortem histological analysis and is graded in four stages based on the distribution and density of p-tau [100, 102]. Given that stages of CTE pathology can be categorized, the molecular course of events suggests an evolution of injury but also the potential for intervention at limiting the progression of CTE.

In addition to the association with neurodegenerative tauopathies, concussion and mTBI are also associated with an increased risk of developing dementias [103]. The link between concussion and dementia demonstrates an overlap in pathology and molecular pathways with some evidence pointing toward concussion as an accelerator for those at risk or increasing risk of development of neurodegenerative disorders such as Alzheimer’s disease [104]. At a basic research level, understanding the progression of tauopathies has been difficult despite the recent development of numerous models of mild and repetitive brain trauma [105,106,107]. Although histological examination in these models demonstrates tau deposition [105], the time course of detection is relatively acute compared to the life-long development in human cases of concussion. The overlapping pathology of other neurodegenerative disorders poses a challenge in trying to parse the effects of tau deposition in the sole context of mTBI, considering the natural progression of dementias and AD. This necessitates longitudinal studies when attempting to establish risk of disease development. This further highlights the challenges in modeling human concussion in animals [4, 108] and extrapolating information between species and reconciling what are shared mechanisms of injury progression, regardless of temporal discrepancies.

Emerging Research

An area of important clinical significance in the acute management of brain trauma is the integrity of the BBB and its effects on cerebrovascular dynamics. While the majority of the work to date has focused on the effects on concussion on white matter injury, there is evidence from animal models that BBB and neurovascular disruption are also components of mTBI [106, 109, 110]. A recent mTBI case study indicates potential detection of neurovascular compromise suggesting that the neurovascular unit may also be at risk in addition to neurons and glia [111]. A critical limitation in the clinical evaluation of BBB is the lack of direct assessment on its integrity. Serum biomarkers are a surrogate measure of a presumed leakiness of the BBB, allowing passage of otherwise impermeable proteins into peripheral circulation and vice versa. However, as indicated, the reliability and the interpretation of serum biomarkers, particularly in light of the discovery of the perivascular glymphatic system [112] and a sinus-associated lymphatic system [113] in the brain poses new questions about the mechanism of protein leakage into circulation following TBI [112, 114].

These recent anatomical findings bring to light the complexity of the brain and our limited understanding, even in its native state. There is an obvious need for further study in understanding the interactions between anatomical structures at the cellular and molecular levels in the context of concussion. Advances in imaging technologies in parallel with advanced molecular techniques are on the verge of providing an integrated understanding of preclinical and clinical pathophysiological mechanisms underlying concussion.