Cellular mechanisms of traumatic brain injury

Abstract

Mild traumatic brain injury (mTBI) is an acute injury with immediate and medium-term symptom presentation. However, our mechanistic understanding of mTBI and how mechanical loading of soft cellular tissues leads to injury is limited. The aim of this review is to introduce this interdisciplinary field to non-experts and provide an overview of our current understanding of how mechanical trauma contributes to cellular injury. Here, we compare the significance of various measures of mechanical loading including strain magnitude, strain rate, loading mode, and frequency, and their relative significance for cell and tissue injury in in vitro and ex vivo experimental models reported in the literature. Interestingly, while it is difficult to define a precise injury threshold value based on strain magnitude alone, cellular injury is commonly observed at strain rates of >0.1 s^-1, higher than rates observed in many normal cell functions (< 0.01 s^-1). We explore the role of the plasma membrane, cytoskeleton, and specialized structures in maintaining cell integrity during traumatic injury.

Introduction

As part of their normal physiology, cells and tissues undergo changes in shape in response to internally generated and external mechanical forces. For example, adult connective tissues are subjected to mechanical forces as part of everyday movement (such as tendons, skin, muscle etc.), and embryonic epithelial layers are stretched by internally generated forces during formation of the internal organs. In addition to passively sustaining mechanical loads, cells actively sense and respond to mechanical stimuli in their environment through a process referred to as mechano-transduction^1,2,3. Cellular responses to load depend greatly on cell type and loading conditions, which can impact cell proliferation^4,5,6, viability^7,8,9, and gene expression^10,11,12,13. In addition, mechanical loading of cells and tissues plays an important role in tissue homeostasis. In many connective tissues the absence of mechanical load can lead to tissue degeneration (i.e., muscle atrophy, loss in bone density)^14,15,16. Given that sustaining and generating mechanical load is critical for tissue function and homeostasis, many tissues are specialized for this purpose and are composed of extracellular matrix proteins (e.g., collagen, elastin) that are organized into load-bearing structures. In contrast, approximately 80% of the volume of human brain tissue are cells, with a ratio of 5:3:1 for neurons, glia and endothelial cells¹⁷.

Not all mechanical loading of cells and tissues is physiological. Beyond a certain threshold, the amount of deformation or the speed at which a tissue is deformed leads to failure, which can present as symptoms such as tissue rupture¹⁸, haemorrhaging¹⁹, blistering, and cracking²⁰. At the cellular level, pathological mechanical loading leads to cell death^{21,22,23,24,25,26} changes in cell morphology^27,28,29, and subsequently alteration in cell function. What defines the difference between physiological and pathological mechanical loading conditions for a particular tissue, and at what point does this transition occur? This question is significant for both the design of personal protective equipment that aims to attenuate traumatic tissue deformation, and our understanding of the mechanisms of disease. Traumatic brain injury (TBI) provides a relevant example, where impact- or blast-based injuries are caused by excessive brain tissue deformation. Instances of TBI are common, affecting 1153 in every 100,000 individuals per year^30,31 and ranging in severity. Mild traumatic brain injury (mTBI) includes conditions such as concussion, whereas severe traumatic brain injury is characterized by complications such as post-traumatic seizures³². mTBI is particularly challenging, as it often goes undiagnosed, but can still lead to long-term medical implications, including Persistent Post-Concussive Symptoms (PPCS). These symptoms, which are developed in approximately 15-25% of mTBI patients^33,34, include chronic headaches, increased likelihood of depression, and dementia^32,35.

While prior studies have primarily explored strain-induced damage at the whole cell and tissue levels, our understanding of the resulting mechanical failure of specific subcellular structures is limited. In this work, we aim to provide an analysis of existing studies to assess our current understanding of injury at the cellular and subcellular levels. Understanding the link between mechanical loading of tissues and the subsequent effect on cellular function requires an interdisciplinary approach that combines concepts from engineering and life sciences. First, we introduce basic measures from mechanical engineering that are commonly used to describe soft tissue deformation. Next, we explore different loading conditions for cells and tissues reported in the literature that are considered physiological or pathological. Finally, we focus on traumatic brain injury and explore the effect of different loading conditions has on brain cells by evaluating how different cellular structures respond to mechanical load and under what conditions they are likely to mechanically fail and lead to injury.

Physiological And Pathological Deformations Of Cells And Tissues

Quantitative metrics for characterizing changes in cell and tissue shape

Descriptive, quantitative, and normalized metrics are useful tools for comparing changes in cell and tissue shape and for distinguishing between physiological and pathological loading conditions. Firstly, changes in shape can be characterized by the mode or ‘type’ of deformation. Cells and tissues experience three primary types of mechanical deformations, namely tension, compression, and shear (see further reading and Box 1 for detailed discussion on modes of deformation). Secondly, changes in cell and tissue shape can be quantified by the parameter strain, which measures the amount of deformation. Strain is a second-order tensor that can be represented by a square matrix (e.g., a 3×3 matrix for 3D strain). For example, brain tissue is structurally anisotropic (i.e., direction-dependent) due to structures such as white matter tracts, or nerves. This is observed in mechanical testing data, where ex vivo tissue samples from the corpus callosum were measured to be relatively stiffer when tension was applied along the fiber tracts as opposed to in the transverse direction, while they were stiffer in compression applied transverse to the fiber tract directions³⁶. It has been suggested that these directional differences are marginal and secondary to anatomical differences seen across various regions in the brain and may therefore be considered mechanically isotropic³⁶ (i.e., uniform in all directions) in its strain response³⁶. However, computational models of head injury have increasingly considered integrating representative tractography to be essential for accurate injury prediction^37,38,39,40. While different physical definitions of strain exist (see Box 1), focusing on a particular mode of strain can provide convenient comparative metrics for studying complex objects such as cells and tissues and are widely used^{41,42,43,44,45,46}. The most commonly used definitions of strain are engineering strain and Green strain. The engineering strain (ε) is used for small deformation (ε < 0.01), whereas Green strain is applicable to a general large deformation. When a deformation becomes too large, cells and tissues can rupture or delaminate leading to cell death. The term ‘ultimate strain’ defines the strain magnitude value at which failure occurs. Thirdly, as most biological tissues and cells are viscoelastic (a time-dependent material property), strain rate,$\dot{\varepsilon }$, is used to describe the rate at which a given strain magnitude is applied. Finally, many cell types and tissues experience dynamic and repetitive loading conditions, such as a beating heart. Repetitive deformation can be characterized in terms of frequency, which is the number of loading cycles per second (or Hertz). For an example of these values, see Box 2.

Box 1 Strain and Modes of Deformation

Strain is a second order tensor that is usually represented by a square matrix (e.g., a 3×3 matrix for 3D strain). Tensile and compressive strains are normal strains that act along the surface normal, whereas shear strain involves deformation parallel to the surface plane. Normal and shear strains are dimensionless and form the component of a strain matrix. The diagonal components ${\varepsilon }_{{xx}}$, ${\varepsilon }_{{yy}}$ and ${\varepsilon }_{{zz}}$ of the 3D strain tensor shown below ($\textbf{$\varepsilon$}$) describe the normal strain in the x-, y- and z-direction respectively, whereas the off-diagonal components $\frac{{\gamma }_{{xy}}}{2}$, $\frac{{\gamma }_{{xz}}}{2}$ and $\frac{{\gamma }_{{yz}}}{2}$ represent shear strain.

$\textbf{$\varepsilon$} =\left[\begin{array}{ccc}{\varepsilon }_{{xx}} & \frac{{\gamma }_{{xy}}}{2} & \frac{{\gamma }_{{xz}}}{2}\\ \frac{{\gamma }_{{yx}}}{2} & {\varepsilon }_{{yy}} & \frac{{\gamma }_{{yz}}}{2}\\ \frac{{\gamma }_{{zx}}}{2} & \frac{{\gamma }_{{zy}}}{2} & {\varepsilon }_{{zz}}\end{array}\right]$

Physical Definitions of Strain

Engineering strain (ε) The normal strain is defined as the ratio of the change in length ($\varDelta L={l-L}_{0}$) to the original length of a material $\left({\varepsilon }_{{yy}}=\frac{\varDelta L}{{L}_{0}}\right)$. The shear component of the engineering strain describes the shape distortion and is defined as ${\gamma }_{{xy}}=\frac{\triangle x}{{L}_{0}}\approx$ α where ∆x indicates the relative displacement between two transverse planes. The engineering strain is suitable for linear elastic materials undergoing small deformation (i.e., length changes that are less than two orders of magnitude of the characteristic length) with negligible rigid body motion.

There are many variants of strain measures developed for different types of loading set-ups. Some of them are defined below based on the uniaxial loading scenario (i.e., normal strain component):

True strain / logarithmic strain (${\varepsilon }_{t}$) The natural logarithm of the ratio of the final length to the original length $\left({\varepsilon }_{t}=\mathrm{ln}\left(\frac{\varDelta l}{{L}_{0}}\right)\right)$.

This definition accounts for large deformations and is suitable for plastic deformations.

Green strain (${\varepsilon }_{G}$) A strain measure defined with respect to the undeformed (reference) configuration $\left({\varepsilon }_{G}=\frac{1}{2}\left(\frac{{l}^{2}-{{L}_{0}}^{2}}{{{L}_{0}}^{2}}\right)\right)$.

This definition is used in finite strain theory and large deformation analysis, along with Almansi strain.

Almansi strain (${\varepsilon }_{A}$) A strain measure defined with respect to the deformed (current) configuration $\left({\varepsilon }_{A}=\frac{1}{2}\left(\frac{{l}^{2}-{{L}_{0}}^{2}}{{l}^{2}}\right)\right)$.

Box 2 Example values of engineering strain

Example 1: Consider a 90-µm-long section of the axon of a neuron⁴⁶ stretched along its long axis and therefore subjected to tensile strain. If the section of the axon undergoes a 4.5 µm extension in 2 seconds, it would experience an engineering strain magnitude of 0.05 µm/µm, sometimes reported as 5%, at a strain rate of 0.025 s^-1. If this cell underwent cyclic loading, being stretched and then relaxed back to its original shape 10 times in 5 seconds, it undergoes a loading frequency of 2 Hz.

Example 2:

Epithelial cells lining blood vessel walls experience shear strains due to blood flow. Consider a cell with dimensions of 40 μm in length and 5 μm in thickness. If a small shear stress is applied, displacing the top surface of the cell laterally by 0.1 μm relative to the bottom, the cell experiences a shear strain of approximately 0.02 radians, indicating a change of angle by approximately 1.1°.

Cell and tissue strains

Different strain magnitudes, strain rates, and loading frequencies are observed in diverse cell types and tissues as part of their physiological function. During embryonic development and in processes such as dorsal closure^47,48 and gastrulation^49,50, epithelial sheets can endure large strain magnitudes, up to 0.5–1.0, which typically occur at low strain rates, 10^-4 to 10^-3s^-151. In contrast, in adult tissues, strain magnitudes range from 0.1 to 1.0 can occur at high strain rates exceeding 0.1 s^{-152,53,54,55}. For example, the strain magnitudes associated with cyclic stretching of lungs range between 0.04–0.12 which occurs at a frequency of 0.2–1 Hz during breathing⁵⁶. Heart valves experience strain magnitudes as high as 1.47^52,57, while diastolic motion of myocardium involves strain magnitudes of 0.1–0.2 and loading frequency of ~1 Hz⁵⁸. Biological tissues often experience a complex combination of deformation modes, strain magnitudes, strain rates, and loading frequencies. A summary of different strain magnitudes and strain rates sustained by selected tissue types is show in Fig. 1.

Fig. 1: Graphical representation of the range of strain loading conditions associated with physiological and pathological processes of the body.

a Simplified graphical representation and b scatter plot of strain rate versus strain magnitude plot of various physiological and pathological processes. Processes are grouped by the relevant organs and tissues. Values for strain obtained for different tissues including bones^{191,192,193,194,195}, lungs^56,196,197, ligaments^198,199, and cartilage²⁰⁰; brain^{61,62,63,64,65,66,71,72,201}, heart^52,57,58,196, skin⁵⁵ and intestines²⁰². Each point plotted represents the average value from an individual study. The bounding lines are qualitative and are for visual reference only.

Transitions from physiological to pathological loading

When strain amplitudes and rates are large, especially for soft tissues that are not specialized for load-bearing, or diseased tissues with attenuated mechanical properties, the applied mechanical loading becomes pathological. A tissue-specific threshold that determines whether a given loading is beneficial or detrimental may exist. We sought to determine the approximate threshold values for strain magnitude and strain rate from the values reported in the literature. When plotting as a function of strain magnitude alone, it is difficult to define a precise injury threshold value although a positive correlation between increasing magnitude and instances of cell injury is both intuitive and seen in the data (Fig. 2a). When plotting as a function of strain rate (Fig. 2b), physiological cellular loading often appears to occur at strain rates <0.01 s^-1 with many reported pathological loading conditions commonly observed at strain rates of >0.1 s^-1. One hypothesis that can be made from this observation is that over this range, some of the cellular structures responsible for maintaining cell shape cannot reorganize to adapt to the deformation occurring on this timescale (for example a strain magnitude of 1.0, at a rate of 0.1 s^-1, takes ~10 seconds). To explore this hypothesis, we discuss the different cellular structures important for maintaining brain cell shape in the context of the timescale for their turnover (binding and assembly, un-binding and dis-assembly) and repair, and how strain rates that exceed these limits lead to cellular injury and TBI. While discussing these structures it is important to note that tissue loading in vivo is a complex combination of magnitudes and rates in three-dimensional space. Making an estimate of timescale and the turnover of different cellular structures provides a simple starting point for the complex problem of brain tissue injury.

Fig. 2: Physiological and pathological responses of cells under strain loading.

a Physiological^203,204,205 and pathological^{21,24,62,126,202,206,207,208} cellular responses at various strain magnitudes. b Physiological^51,204,205 and pathological^{21,126,202,207,209} cellular responses at various strain rates. Physiological responses are shown in green text, whereas the pathological responses are shown in red.

Mechanisms Of Cellular Injury In TBI

Due to the clinical significance of traumatic brain injury, there has been an ongoing effort to estimate the strain magnitudes and strain rates that occur in vivo for brain tissue during impact-based (e.g. collisions during sports) and blast-based (e.g. exposure to explosions) TBI events. There have also been concerted efforts to study TBI in the context of age-related neurodegenerative diseases, as neurodegenerative diseases compromise the brain’s structural integrity and resilience, making it more susceptible to traumatic brain injury (TBI) (see, for example, Abdi et al. ⁵⁹). In addition, TBI itself can accelerate neurodegenerative processes, creating a cycle of increased vulnerability and progressive cognitive decline (e.g., Brett et al.⁶⁰). Existing studies suggest that the average strain rate experienced by brain tissue during impact TBI could reach as high as 52 s^-1. In vitro models and computational simulations of inertia-driven impact TBI have demonstrated that brain tissues deform by peak strain magnitudes of 0.2–0.5 at rates of 10 to 50 s^{−1 61,62,63,64,65,66,67,68,69,70} (Fig. 3a). In blast-driven TBI, computational simulations of the head subjected to shock wave loading predict brain tissue strain magnitude of ≤0.1, but at high rates ranging from 12 to 960 s^-171,72. Together, these studies indicate that TBI can occur over a broad range of loading conditions depending on the context of the injury. For the basis of our discussion, we consider mild Traumatic Brain Injury (mTBI) events to occur at a representative average strain magnitude of ~0.3 and strain rate of ~10 s^-1.

Fig. 3: Responses of tissue and cellular models of the brain to different strain magnitudes and rates.

a A representative strain field observed within a head surrogate intended to simulate the brain under mTBI conditions. The strain field was captured at the time point corresponding to the peak maximum principal strain of 0.20 magnitude at rate of 7 s^-170 (© Rovt et al. 2023, reprinted with permission from Elsevier). b Axons show swellings (bulbs) two hours after high strain magnitude, high strain magnitude (0.58–0.77) and high strain rate (26–35 s^-1)¹⁰⁷ (© Smith et al. 1999, Society for Neuroscience). c Degree of axonal undulation is shown to be proportional to the strain amplitude applied at a rate of 44 s^-142 (© Tang-Schomer et al. 2009, reprinted with permission from Wiley & Sons). d Model of the axonal cortical cytoskeleton; The axonal core contains bundled microtubules cross-linked by microtubule-associated proteins. Surrounding this core are neurofilaments and an outer scaffold of periodic F-actin rings connected by α/β-spectrin tetramers aligned along the axon¹⁸⁰ (cross-sections of the tetramers shown as dimers) (Adapted from Dubey et al. 2020, used under CC-BY license). e Super-resolution images of axons labelled with anti-βII spectrin (primary) and Alexa Fluor 488 (secondary). Repeated patterns along the axon length indicating ring structures are observed. Scale bar: 1 μm¹⁸⁰ (© Dubey et al. 2020, used under CC-BY license).

Cellular responses to high strain rate loading

What happens to the cellular constituents of brain tissue when loaded by a strain magnitude of ~0.3 at a strain rate of ~10 s^-1 associated with mTBI? Several different cellular structures are well documented as having an important role in sustaining cell shape, which include the plasma membrane, cytoskeleton, cytoplasm as well as a secondary role for some organelles. Furthermore, cells in the brain, such as neurons, often contain specialized structures such as axons, dendrites, and spines that give cells complex morphologies and features that could be damaged during mechanical loading. In general, cellular structures can accommodate changes in cell shape by firstly ‘unravelling’ or aligning in the direction of mechanical loading (an entropic change), and secondly by ‘stretching’ which depends on the mechanical properties of the proteins, lipids, and molecules that constitute that structure (an enthalpic change). Cellular structures that can accommodate high strain magnitudes therefore typically either have a lot of ‘extra material’ that can unravel under stretch or consist of a material that is durable and can stretch without mechanical failure. As part of a living system, many cellular structures undergo active remodelling and turnover and are repaired as a normal part of cell function. The timescale for critical processes such as protein diffusion, binding, synthesis, and repair, and the timescale over which deformation occurs therefore have complex roles in cellular responses to strain rate.

To study the mechanical response of cells and tissues to external mechanical loading, a range of different tools have been developed. In vitro methods for exposing cell cultures and tissues to mechanical stretch are now widely used to evaluate injury in response to deformation^{73,74,75,76,77}. In addition to whole cell and tissue loading, biophysical tools such as atomic force microscopy^78,79, micropipette aspiration⁸⁰, and optical tweezers⁸¹ can be used to characterize the mechanical properties of subcellular structures, including the properties of cytoskeletal polymers^82,83,84 and the response of the plasma membrane^85,86. Similarly, a range of optical microscopy and spectroscopy techniques have been developed to measure the dynamics of proteins and molecules within subcellular structures including single molecule imaging^87,88, Fluorescence Correlation Spectroscopy (FCS)^89,90, and Fluorescence Recovery after Photobleaching (FRAP)^91,92,93,94. In the following sections, we use data reported in the literature that were obtained from whole cell testing, biophysical tools, and microscopy/spectroscopy to make comparisons between the role of strain magnitude, strain rate, and loading frequency on brain cell injury, in the context of a transition from physiological to pathological loading of brain tissue.

Plasma membrane

The plasma membrane separates the interior of mammalian cells from the external environment and consists of a lipid bilayer packed with membrane proteins^95,96,97. While lipid bilayers themselves cannot accommodate large planar strains (up to 0.02–0.04^98,99), the cellular plasma membrane is not flat, and membrane material contained within caveolae (small 50–100 nm membrane invaginations) and other structures can be unravelled under stretch to buffer rapid changes in cellular strain. For example, cells subjected to hypo-osmotic shock, which increases their volume by ~35%, results in a reduction in the number of caveolae by ~30%¹⁰⁰. In migration of primordial germ cells, membrane protrusions (blebs) form through the flattening of membrane tubes^99,101. In addition to changes in membrane topography, lipids and proteins in the plasma membrane can diffuse in the plane of the membrane and are mobile, potentially allowing them to reorganize under strain. For example, in supported lipid bilayers recovery of fluorescence occurs within ~20–30 seconds after photobleaching¹⁰², with similar values reported for the recovery of membrane-bound proteins by lateral diffusion and exchange with the cytoplasm in ~20 seconds¹⁰³. It has also been shown that changes in plasma membrane tension following changes in cellular strain coordinates the activation of exocytosis and contraction¹⁰⁴. In addition to passive effects, mechanical strain has also been suggested to directly affect lipid order and plasma membrane mobility, slowing recovery rates¹⁰⁵. The plasma membrane is therefore a dynamic cellular structure that can respond to the strains associated with mTBI injury both directly through changes in shape and dynamically through planar diffusion and turnover.

Indeed, changes in membrane morphology are often a hallmark of cellular injury in TBI. For example, Diffuse axonal injury (DAI) is a form of TBI-induced injury characterized by extensive stretch- and shear-induced damage to axonal fibres due to rotational forces experienced during sudden head acceleration or deceleration¹⁰⁶. Early-stage axonal injuries in DAI are marked by bulging along the long axis of the axons (axonal swelling). Uniaxial stretch experiments on neuronal cells¹⁰⁷ showed that while axons could withstand strain magnitudes of up to 0.65 without severing, they exhibited cytoplasmic and membrane bulging upon returning to their original length (Fig. 3b). Other experiments have reported axonal swellings and the formation of axonal bulbs at strain magnitudes exceeding 0.22 and strain rates over 27 s^-1, with these changes becoming more pronounced as strain amplitude and rate increased^108,109. Pathological axonal membrane swellings where microtubule bundles have disintegrated into loops or waves have been observed in both TBI and neurodegenerative diseases, and highlighting the significance of the interaction between the plasma membrane and underlying cytoskeletal structures^{42,107,108,109,110,111,112,113,114}. Interestingly, recent work has also suggested that axonal bulbs may have a protective effect, preventing Ca²⁺ influx into non-stressed regions of the axon¹¹⁵. Additionally, axonal injury disrupts fast axonal transport, leading to the abnormal accumulation of amyloid precursor protein (APP) at sites of axonal swelling. This APP buildup, a hallmark of DAI, results from impaired anterograde transport and has been implicated in secondary neurodegenerative processes following TBI^116,117,118.

In addition to changes in membrane morphology, rupturing of the plasma membrane has also been suggested as a mechanism for axonal injury. Traumatic injury has been shown to increase membrane permeability, which correlates with the rate and magnitude of loading^{119,120,121,122,123,124,125}. In neuronal injuries, membrane rupture, or the formation of membrane pores under strain (mechanoporation), frequently occurs before neuronal degeneration in tension and shear traumatic brain injury experiments, at strain magnitude of 0.3 and strain rate of 10 s^-1126. Furthermore, membrane mechanoporation can trigger calcium influx, reactive oxygen species (ROS) generation, and mitochondrial dysfunction, leading to further cellular damage and death⁴⁵. Mechanical stimulation of the membrane is also closely linked with the regulation of mechanosensitive ion channels. Mechanosensitive ion channels, such as Piezo 1 and Piezo2, play a critical role in maintaining neuronal homeostasis by regulating processes like cell migration and differentiation through controlled Ca²⁺ influx^{127,128,129,130}. Piezo 2 channel in particular helps modulate sensory processes under normal conditions^129,131. However, TBI conditions can dysregulate Piezo2, leading to an excessive calcium influx. This disruption not only increases cellular stress but also contributes to neuronal death and the production of inflammatory cytokines like TNF-α and IL-1β¹³². For example, cells stretched at strain magnitudes of 0.73–0.9, at a strain rate of 33 s^-1 exhibited prolonged plasma membrane damage and an increase in cell death (up to 45%), particularly after 12-24 hours post-injury¹³³.

Microtubules and Axons

Microtubules are cylindrical filaments formed from protofilaments of tubulin dimers and play a critical role in processes such as cell division¹³⁴. Microtubules have a high bending rigidity (i.e. a persistence length of several millimetres^135,136) and have been reported to have a non-negligible role in determining the mechanical properties of single cells in compression^137,138,139. Microtubules are highly dynamic and exhibit dynamic instability, characterized by continuous phases of polymerization and depolymerization (turnover time of 1–10 min)^140,141. The dynamic behavior of microtubules is regulated by interactions with microtubule-associated proteins (MAPs), which bind to microtubules and influence their stability, dynamics, and cellular interactions. For example, MAP2 serves as a marker of dendritic injury and synaptic plasticity following TBI and has a role in stabilizing microtubules and supporting neuronal structure^142,143. Similarly, MAP6 (also known as STOP) has been shown to contribute to microtubule stabilization post-TBI¹⁴⁴. The MAP tau is essential for stabilizing and organizing microtubules in parallel alignment within axons, and the cross-linking of tau helps maintain uniform spacing between adjacent microtubules and reduces depolymerization^{145,146,147,148,149}. Axons in neurons are supported by tau-stabilized microtubules, which exhibit a long turnover time of >30 minutes^150,151. The interplay between turnover dynamics and the mechanical response is particularly significant for microtubules. For example, mechanical stresses have been reported to cause damage to the microtubule lattice, which can repair itself through the addition of new tubulin dimers¹⁵².

In axons, continuous bundles of cross-linked microtubules are thought to provide structural support against different forms of mechanical loading, such as tension and compression¹¹¹. Indeed, mechanical loading can lead to microtubule bending which can lead to internal tension, compression, and shear. This will in turn promote breaking and subsequent depolymerization of microtubules^153,154. Following loading at a large strain magnitude of 0.3-0.75 and a fixed strain rate of 44 s^-1, undulating distortions have been observed to form along axons (Fig. 3c) which coincides with buckling, breakage, and progressive loss of microtubules⁴². Mechanically induced damage of microtubules and subsequent axonal failure are strain rate dependent^{71,155,156,157}, consistent with the notion that bending, potential breakage and disruption of the microtubule network^153,154 occur particularly in regions of the axon where the microtubules are relatively stable with long turnover times.

The strain rate sensitivity of microtubule-based injury may also be attributed to the viscoelastic properties of the neuronal cytoskeleton^158,159. It has been hypothesised¹⁶⁰ that the tau proteins are viscoelastic: being flexible at low strain rates ( ~ 0.01 s^-1) allowing for reversible sliding of microtubules^126,161,162 but stiffen up under strain applied at high rate, transferring significant load onto microtubules and potentially causing damage¹⁶⁰. This microtubular failure likely disrupts neurite transport, leading to an accumulation of vesicles and organelles, which manifests morphologically as axonal beading¹⁶³, and causing electrophysiological dysfunction of the neurons⁶⁶.

Actin, growth cone, dendrites, and spines

The actin cytoskeleton plays a critical role in determining cell shape and mechanical properties^2,164,165. Actin filaments are semi-flexible polymers (i.e. a persistence length ~10–20 µm¹⁶⁶) that assemble into higher order structures such as meshes, bundles, networks and fibres, to enable cells to sustain external load, generate forces, and change shape. The assembly of different actin structures is facilitated by the interaction between actin filaments and an array of regulatory proteins that include actin filament nucleating proteins, crosslinking proteins, and motor proteins^167,168. All these elements, including actin filaments themselves, are dynamic, allowing cells to dynamically change shape and generate mechanical forces. The turnover timescale of actin filaments and their regulatory proteins is structure dependent and have been measured using fluorescent fusion proteins to actin FRAP. For example, filaments in the actin cortex are highly dynamic, turning over in ~11 seconds^169,170. By comparison actin filaments and crosslinking proteins in structures such as stress fibres are more stable, taking hundreds of seconds to turn over^171,172,173. Interestingly, disease-causing mutations to actin regulatory proteins can lead to excessive polymerization of actin filaments^174,175 or over-crosslinking of actin structures subsequently changing the turnover dynamics of different actin structures and cellular mechanical properties^176,177.

In neurons, actin plays a central role in the formation of distinct structures that are significant for their response to strain loading¹⁷⁸. In the axon, actin filament rings form along the length of the axon interconnected by spectrin tetramers, with a periodicity of 180-190 nm¹⁷⁹ (Fig. 3d). Actin-spectrin networks have been proposed to have a role in the mechanical response of the axon in TBI. For example, Dubey et. al. ¹⁸⁰. used a custom-built stretching apparatus to investigate the mechanical response of axons to strain loading. They observed that the actin-spectrin cytoskeleton (Fig. 3e) is a prominent contributor to axon mechanics by functioning as a shock absorber. Spectrin proteins are proposed to reversibly unravel¹⁸¹ by a strain magnitude of up to 0.1, softening the axon and buffering changes in length that could potentially cause damage. In other work, an increase in actin-spectrin spacing from 183 nm to 202 nm has been observed adjacent to neuronal swellings following injury by a strain magnitude of ~0.1¹⁸². Furthermore, it has been proposed that softening of the spectrin scaffold under repeated mechanical loading exposes microtubules in axons to increased stress during repeated TBI events¹⁸³. In addition to the formation of actin rings in axons, actin plays a critical role in dynamic processes such as the formation of growth cones, dendritic spines, and protrusions. In many of these structures, actin dynamics and remodelling are critical for plasticity and changes in actin dynamics can lead to neurodegenerative disease^184,185,186. For example, it has been shown that actin-binding protein cofilin interferes with the actin polymerization process when the cell is subjected to stress, leading to the formation of persistent rod-like structures which are linked to Alzheimer’s and Huntington’s disease^{187,188,189,190}. Changes in the turnover dynamics of these disease-associated structures and the corresponding effects on the strain response of these cells is an interesting area of future research.

Conclusion

Traumatic brain injuries (TBI) are a common and significant health concern, with both short- and long-term detrimental effects. Cellular injury in TBI depends on strain magnitude and strain rate, presenting a complex combination of loading conditions. Interestingly, many cell types are able to tolerate high strain magnitudes if loaded at a quasi-static strain rate, but sustain damage from loading applied at high strain rate. This suggests that the turnover of cellular structures, such as the membrane and cytoskeletal polymers, allows cells to adapt to slow physiological loads, but not fast loading. This highlights the significance of considering various mechanical loading parameters (e.g. strain magnitude, strain rate, loading mode, and frequency) in determining cellular and tissue injury. Among these, high strain rate emerges as a critical factor contributing to cellular damage, which can be linked to the mechanical properties and dynamic responses of subcellular structures. In this work, we have introduced some of the key structures that contribute to cellular responses to strain. The plasma membrane, with its rapid turnover time ( < 10 s), experiences in-plane tension and serves as a barrier against mechanical deformation. The cytoskeletal components exhibit a range of mechanical responses: actin in the cortex (turnover ~11 s) and stress fibres ( > 100 s) primarily counteract tension, while microtubules (1–10 min) and MAP-stabilized microtubules ( > 30 min) resist tension, compression, and shear forces. These turnover times highlight the varying capacities of cellular structures to respond to mechanical strain, influencing the extent of injury and potential recovery following traumatic loading. Understanding these behaviours and characteristics is crucial for developing targeted interventions to mitigate cellular damage.

Future research in traumatic brain injury (TBI) should focus on elucidating the interplay between these structural components under different loading conditions, with particular attention to how the dynamics of individual cell components may influence injury thresholds. A possible area of focus could include the study of membrane-cytoskeletal linker proteins and cytoskeletal crosslinking proteins, combined with measurements of their dynamics and mechanical properties, to develop a holistic understanding of cellular injury mechanisms in TBI. Additionally, exploring potential therapeutic strategies to modulate cytoskeletal dynamics and membrane resilience may provide new avenues for mitigating TBI-induced cellular damage. Ultimately, a deeper understanding of these mechanical interactions will enhance our ability to develop effective protective and regenerative strategies in neurotrauma research.