Mitochondria has two membranes: The outer membrane, which separates the organelle from the cytosol, and the inner membrane, surrounding the mitochondrial matrix, thus defining three different specialized regions [17]. The outer membrane separates the mitochondria from the cytosol, allowing the passage of metabolites through porins (POR) of the Voltage-Dependent Anion Channel (VDAC) and proteins that make up the family of Translocases of outer membrane (TOM) [18]. Further, outer membrane proteins regulate mitochondrial dynamics, mainly mitofusins 1 and 2 (Mfn1 and Mfn2) and mitochondrial fission protein 1 (FIS1) and proteins of the B-cell lymphoma-2 family (Bcl-2) such as Bcl-2 homologous antagonist killer (BAK). Also, outer membrane proteins interact with inner membrane proteins: the G alpha-q (G$\alpha$q/11) protein interacts with the optic atrophy protein 1 (OPA1) and the Sorting Assembly Machinery (SAM) protein interacts with the Mitochondrial contact site and cristae organizing system (MICOS) [17]. Additionally, mitochondrial protein mitoNEET containing Asn–Glu–Glu–Thr (NEET) sequence (also referred as mitoNEET), an iron-containing outer mitochondrial membrane protein that regulates oxidative capacity and calcium buffering by mitochondria, can also participate in energy metabolism transferring electrons to oxygen or ubiquinone [19]. In addition, the outer membrane also has specific regions for interaction with other organelles: Mitochondria-associated membranes (MAM), which represent a region of the endoplasmic reticulum (ER) reversibly tether to mitochondria [20]; and the inter mitochondrial junctions (IMJ), which interact with other mitochondria [21].

The inner membrane delimits the intermembrane space (IM), structuring the mitochondrial cristae that separate the IM from the mitochondrial matrix [17]. Importantly, cristae are fundamental structures for mitochondria and not simple invaginations. In fact, tight tubular junctions regulated by different proteins separate cristae from the intermembrane space, forming specialized compartments to limit the diffusion of molecules important for the electron transport system and guarantee maximum functionality [22]. Namely, the inner membrane contains the proteins of the Translocases of the Inner Membrane (TIM) family, which transport proteins to the matrix, OPA1, MICOS, and cardiolipin (CL) interacting for the correct assembly and localization of inner membrane proteins. Specifically, G$\alpha$q/11 (from the outer membrane), Presenilins-associated rhomboid-like protein (PARL), reactive oxygen species modulator 1 (ROMO1), regulate OPA1 to keep the cristae junctions closed [17]. Also, the interaction of ATP proteases of the matrix (m-AAA) with the complex prohibitin–stomatin-like protein 2 (PHB-SLP2) anchored in cardiolipin, influences OPA1 activity [17]. Thus, the interaction of these proteins guarantees the functionality of cristae, their structure, and the disposition of complexes of the electron transport system [23].

Conceptually, mitochondrial oxidation of metabolic substrates in the tricarboxylic acid cycle (TCA) or the cytosol via malate-aspartate shuttles directs the flow of electrons from reduced nicotinamide adenine dinucleotide (NADH) and Reduced adenine flavin dinucleotide (FADH₂) through the respiratory complexes to oxygen and concomitantly generates a hydrogen proton gradient and the synthesis of ATP. The inner membrane houses the components of the electron transport system: Complex I (NADH ubiquinone oxidoreductase or NADH dehydrogenase); Complex II (ubiquinone succinate oxidoreductase or succinate dehydrogenase); Complex III (ubiquinol-cytochrome C oxidoreductase); Complex IV (cytochrome C oxidase) where O₂ consumption occurs and finally; Complex V, F1Fo ATP synthase, where ATP synthesis occurs. Super complexes organize the components of the electron transport system along the inner membrane (Fig. 1) and may present different sets of complexes (for example I, III, IV and V; II, III, IV and V; in addition to from the classic described from I to V), always notoriously converging to F1Fo ATP Synthase dimers located at the edge of cristae [17].

Additionally, a concept of utmost importance is the proton-motive force, the combination of the mitochondrial membrane potential ($\mathrm{\Delta }$$\mathrm{\Psi }$m; generated by pumping H⁺ protons) from the matrix into the intramembrane space via complexes I, III, and IV) and the differential chemistry of the H⁺ gradient ($\mathrm{\Delta }$pH) allowing the synthesis of ATP to occur in complex V and the return of protons to the matrix. Thus, proton leakage, either inducibly via the action of uncoupling proteins (UCP), or non-inducible (pathological) can influence mitochondrial processes. Succinate oxidation through Complex II donates electrons to ubiquinone and consequently to Complex III, independently of Complex I, a phenomenon called ‘Convergent flow’ comprising an important aspect of ATP synthesis via the transport system [24].

Likewise, the transport system also produces ROS, the physiological byproducts of normal oxygen metabolism. Categorically, ROS include free radicals, mainly superoxide (O₂^•-), hydroxyl radical (OH^•-), and singlet oxygen (¹O₂), as well as non-radical species such as hydrogen peroxide (H₂O₂). Under physiological conditions, several endogenous antioxidants prevent oxidative damage, such as the superoxide dismutase (SOD), glutathione peroxidase (GPx), and catalase (CAT) enzymes. Strikingly, excess oxidative damage impairs the integrity and permeability of cell membranes, simultaneously causing dysfunctions in the electron transport system, a hallmark of many diseases [25]. Classically, ROS overproduction causes $\mathrm{\Delta }$$\mathrm{\Psi }$m hyperpolarization, pro-apoptotic stimuli such as cytochrome c and caspase activation, and specific oxidation of mitochondrial DNA (mtDNA), contributing to apoptosis [26].

Noteworthy, the complex interaction of substrate concentration, oxidized and reduced forms of Nicotinamide adenine dinucleotide (NAD⁺/NADH) ratio, ATP/ADP ratio, and, more recently, calcium concentrations regulate mitochondrial oxidative metabolism. Conversely, mitochondrial calcium concentrations are low, ranging from 0,1 to 10 µM, and rapidly respond to increases in cytosolic calcium [27]. Then, the mitochondrial calcium uniporter (MCU) effectively senses extra mitochondrial calcium levels and mediates Ca²⁺ entry into the mitochondrial matrix, partially dissipating the mitochondrial membrane potential ($\mathrm{\Delta }$$\mathrm{\Psi }$m). This parallels with the activation of three enzymes of the mitochondrial matrix (Pyruvate dehydrogenase, Isocitrate dehydrogenase, and alpha-ketoglutarate dehydrogenase) in addition to regulating the malate-aspartate shuttle, the glycerol-3-phosphate shuttle and directly the complex III and V of the transport system of electrons [28, 29].

Additionally, Ca²⁺ regulates respiration independently of ATP demand [30]. To maintain calcium homeostasis, the mitochondrial sodium/calcium/lithium exchanger (NCLX) balances influx by extruding one Ca²⁺ in exchange with three sodium ions [31]. Notably, mitochondrial Ca²⁺ effects on different processes may differ, ranging from activation to shutdown. For example, moderate increases are necessary and sufficient to adjust ATP production to cellular demand, and Ca²⁺ overload leads to disruption of mitochondrial membrane integrity and irreversible oxidative damage decreasing ATP production capacity [32]. This phenomenon essentially results from three mechanisms: (I) increased mitochondrial Ca²⁺ uptake, after efflux from the endoplasmic reticulum and influx of Ca²⁺ from the extracellular space; (II) reduction of Ca²⁺ extrusion through NCLX; and (III) changes in mitochondrial Ca²⁺ buffering (Fig. 2). Therefore, impairment of mitochondrial calcium homeostasis is a recognized feature of the early stages of neurodegeneration, with direct implications in CNS diseases, with synergistic effects occurring between disruption of mitochondrial membrane potential, calcium overload, and ROS in mediating cell damage [27, 33].

Owing to the importance and complex relationship between $\mathrm{\Delta }$$\mathrm{\Psi }$m, ROS production, calcium homeostasis, and ATP synthesis, another important point is how mitochondria react to changes in these components: mitochondria is a dynamic organelle, with a life cycle also involving two processes called fusion and fission. Both fusion and fission processes are fundamental for maintaining the mitochondrial function, adapting the morphology, and forming a dynamic interconnected network, maintaining mitochondrial integrity and quality [34]. Moreover, the life cycle of mitochondria comprises distinct periods: the solitary period, pre-fusion, network formation, and post-fusion period (Fig. 3). Furthermore, mammalian mitochondria characteristically exhibit brief fusion that may trigger selective fission, where segregated (solitary) mitochondria can sustain normal membrane potential or present depolarization, ultimately being eliminated by mitophagy [34]. For instance, decreases in the mitochondrial membrane potential upregulates expression of Phosphatase and tensin homolog-induced putative kinase 1 (PINK1), which aggregates on the mitochondrial membrane, causing activation of parkin to promote mitophagy.

Subsequently, the dynamics of these processes allow mitochondria to retain a more stable membrane potential after fission, although it can vary by up to 5 mV from the solitary period. Therefore, fission results in functionally different mitochondria. For fission to occur, the protein Dynamin-related protein 1 (DRP1) interacts with two proteins anchored in the outer mitochondrial membrane: the Mitochondrial Fusion Factor (Mff) and Mitochondrial fission 1 protein (FIS1), to a lesser extent [35]. In contrast, fusion combines the contents of the original mitochondria and requires an intact mitochondrial membrane potential, regenerating damaged mitochondria. Further, the fusion allows for rapid diffusion of mitochondrial matrix proteins combined with slower migration of inner and outer membrane components. In mammals, different GTPases of the dynamin family regulate mitochondrial fusion: in the inner mitochondrial membrane, OPA1 and mitofusins in the outer mitochondrial membrane (Mnf1 and Mnf2). A greater detail of these mechanisms is beyond the scope of this review; however, the reader may benefit from resorting to other sources [34, 36].

Importantly, the mitochondrial signaling that results in apoptosis (programmed cell death) results from the interaction of Ca²⁺ concentration with proteins of the Bcl-2 family. Anti-apoptotic proteins B-cell lymphoma 2 (Bcl-2) and B-cell lymphoma-extra large (Bcl-XL) reduce endoplasmic reticulum Ca²⁺ concentrations and upregulate mitochondria to maintain cellular homeostasis. This is possible thanks to its location on the membrane of both organelles. Conversely, the Bcl-2-associated X protein (BAX) and Bcl-2 homologous antagonist/killer protein (BAK), belonging to the same family, have a pro-apoptotic effect [27, 33]. Likewise, the regulation of this pathway is directly proportional to the concentrations of calcium released by the endoplasmic reticulum and consequent mitochondrial uptake, since an exacerbated increase results in the opening of the mitochondrial permeability transition pore (mPTP), consequently changing the morphology, and inducing the release of cytochrome c, and activation of caspases, determining pro-apoptotic stimuli [27, 33]. However, the cellular and molecular mechanisms involved in these progressive changes (Fig. 4) can simultaneously interact to exacerbate injury and initiate neuronal repair [34, 36]. The pathophysiological mechanisms associated with TBI involve primary injury resulting from mechanical or inertial damage to white and gray matter, leading to cellular membrane disruption, content release, and diffuse axonal injury [37, 38]. Secondary injury refers to the progression of changes associated with primary brain injury, resulting in various neurological alterations that can have devastating lifelong consequences, such as persistent activation of a series of neurotoxic event cascades triggers the progression of structural damage, loss of neuronal function and connectivity, culminating in the death of neural cells adjacent to the injury focus [37]. Therefore, the extent and severity of secondary damage are proportional to the location and intensity of the primary insult. Mechanisms underlying secondary damage include excessive increases in extracellular glutamate (excitotoxicity), impairments in calcium (Ca²⁺) homeostasis, persistent inflammatory response, neuroenergetic deficiency, redox imbalance, vascular system dysfunction, ischemia, axonal protein accumulation, and mitochondrial dysfunction [39, 40].

Experimental studies have revealed that the phospholipid cardiolipin translocates from the inner to the outer mitochondrial membrane following TBI, thereby tagging damaged mitochondria for selective removal via mitophagy [39]. These mitochondrial distress signals elicit local and systemic responses by interacting with specific receptors on immune cells. Likewise, activation of the mPTP is crucial for cell survival post-TBI [41]. In this context, mitochondrial Ca²⁺ is regulated by two transporters: the mitochondrial calcium uniporter (MCU), facilitating Ca²⁺ influx regulated by changes in the membrane potential, and the mitochondrial Na⁺/Ca²⁺/Li⁺ exchanger (NCLX), mediating Ca²⁺ efflux which is potentially inhibited following TBI [42]. Excessive Ca²⁺ within mitochondria can trigger the opening of the mPTP, resulting in mitochondrial edema and reduced oxidative metabolism, releasing mitochondrial Ca²⁺ and activating detrimental calcium-dependent proteases such as calpain [43]. Additionally, the loss of cytochrome c through the mPTP activates cellular apoptosis via caspases [41].

Noteworthy, in most experimental models of CNS diseases, mitochondrial dysfunction is a common source of cellular metabolic crisis [8, 9], especially following TBI [44, 45]. This partially reflects the pathological increase in intracellular calcium concentrations and mitochondrial calcium buffering, consequently decreasing mitochondrial ATP production, impairing other cellular functions, and concomitantly increasing ROS generations [46, 47]. Thus, with increased calcium exceeding the capacity for mitochondrial sequestration, selective destruction of the cytoskeleton can lead to cell death by apoptotic neurodegeneration [48]. Excess calcium induces the fragmentation of structural proteins of the neuronal cytoskeleton, such as alpha-spectrin, in addition to stimulating the hyperphosphorylation of Tau, classic molecular biomarkers of neurodegeneration, via calpain (mainly calpain-2) and caspases activities [49].

In this context, therapeutic interventions targeting mitochondria may delay or prevent secondary cascades that lead to cell death and neurobehavioral disability. Thus, the emerging interest in mitochondrial function in TBI reinforces the historically established importance of this organelle in the scientific literature. Although substantial conceptual evidence currently considers brain mitochondria as a potential therapeutic target against neurodegeneration, no pharmacological intervention with proven efficacy is available to date [11, 12, 13].

Nevertheless, research on rodents demonstrated promising outcomes in enhancing mitochondrial function following TBI. Currently, several drugs targeting mitochondrial metabolism are possible candidates for further repurposing for TBI patients, albeit most of the proposed interventions fail to provide benefits in clinical trials [11, 13]. In this section, we focus on mitochondrial-targetting drugs that showed promising results in pre-clinical studies that also were tested in human clinical trials, regardless of TBI severity. For instance, N-acetyl cysteine (NAC) has demonstrated significant neuroprotective effects in various animal models, particularly by mitigating secondary neuronal injury resulting from TBI, via the upregulation of glutathione levels (GSH), which functions as a free radical scavenger and an antioxidant, improving outcomes in TBI [50, 51]. Recently, a double-blind, placebo-controlled human trial evaluated the efficacy of NAC in patients with blast-induced mild traumatic brain injury. The treatment group received 2 g of NAC twice daily for the first four days, followed by 1.5 g of NAC twice daily for three days, and showed significant improvement (p 0.01) in different neurological symptoms, and enhanced chances of recovery [52].

Similarly, Peroxisome proliferator-activated receptor (PPAR) agonists, including rosiglitazone and pioglitazone, also presented neuroprotective properties in a preclinical study, potentially reducing axonal injury modulating apoptosis, autophagy, and microglial activation [53]. Interestingly, the pioglitazone therapeutic window is delayed, with better results obtained following treatment initiated 18 hours following injury [54]. The effects also depend on dosage and mitoNEET presence, which modulates calcium-related mitochondrial mechanisms [55]. In addition, two studies showed significant effects on mitochondrial dysfunction after severe and mild preclinical TBI, which extended to sustained cognitive performance and reduced cortical damage, with modest antioxidant and tissue-sparing effects [19, 54]. Deng et al. [56] combined an animal model of TBI with brain tissue, blood, and cerebrospinal fluid from forty-five patients with moderate to severe TBI [57]. Strikingly, the study reported brain edema and high expression of the inflammatory factors Interleukin-6 (IL-6), nitric oxide (NO), and the apoptotic factor caspase-3 in TBI patients. Treatment with pioglitazone in TBI rats showed beneficial effects, and PPAR type gamma (PPAR$\gamma$) activation may partly explain the reported results [56]. As PPAR$\gamma$ promotes oxidative phosphorylation, antioxidant defense, and mitochondrial biogenesis, pioglitazone emerges as a potential drug for TBI [58].

Importantly, mild mitochondrial uncoupling with exogenous stimuli to supplement the endogenous response to TBI has shown beneficial effects, implementing low doses of drugs that promote mitochondrial uncoupling, such as 2,4-dinitrophenol (DNP), carbonyl cyanide-4-phenylhydrazone (FCCP), and mitochondrial uncoupler prodrug of 2,4-dinitrophenol (MP201), which demonstrated positive effects in pre-clinical studies, albeit still under scrutiny regarding human clinical trials [59, 60]. Moreover, a few small-scale clinical trials evaluated cyclosporin A, which delays mPTP opening and regulates calcium buffering, showing no beneficial effects [57, 61].

Further, guanosine, an endogenous nucleoside, also showed neuroprotective effects following mild TBI, increasing oxidative phosphorylation and ETS function, resulting in preserved motor performance and cognitive functions [62], an effect mediated by the A_2⁢a Adenosine receptor [63]. Likewise, testosterone can be considered, since it potentially modulates several neurodegenerative processes in mice subjected to severe TBI. A particular study [42] showed that TBI increased calcium uptake and impaired calcium extrusion via the NCLX, effects that were linked to a deterioration in mitochondrial membrane potential dynamics and a decreased efficiency in mitochondrial ATP synthesis coupling, and these effects were mitigated by testosterone. Also, there was an observed increase in hydrogen peroxide production post-TBI, without a corresponding change in the immunocontent of the mitochondrial Superoxide dismutase 2 (SOD2) protein, and testosterone prevented the increase in tau fragmentation and its phosphorylation at serine 396, and increased in alpha-spectrin cleavage via calpain-2 activation. Additionally, testosterone decreased both caspase-3 activation and the Bax/Bcl-2 ratio, indicating downregulation of mitochondrial apoptotic signaling pathways. Hence, testosterone administration after severe TBI enhanced calcium extrusion through the NCLX exchanger and improved mitochondrial function, reducing the overexpression of molecular factors associated with neurodegeneration. Thus, since TBI often induces hypopituitarism and testosterone reduction, testosterone may help improve metabolic, neurological and functional outcomes, albeit additional investigations are necessary [64]. Notably, a small clinical trial reported functional improvements in men receiving testosterone following severe TBI when compared to rehabilitation alone [65]. Another promising approach is mitochondrial transplantation, in which mitochondria from healthy cells are transferred to stressed cells in damaged tissue, to achieve neuroprotective effects and mitigate cognitive dysfunctions [66, 67].

Remarkably, the failure of translation from bench to bedside in the field of TBI is due to many challenges, including failure the lack of translation of effects from preclinical models to humans, and the fact that current clinical trials can be full of shortcomings and design flaws [68, 69]. Much of the literature has focused on severe TBI, drawing criticism due to the challenges coming from the large variability in TBI pathology (contusions, diffuse axonal injury, subdural hematomas), patient demographics (age, sex), and secondary complications [70]. Furthermore, many often highlight other problems: provision of emergency care and critical care is uneven across regions; surgical management varies; inconsistent rehabilitation protocols during hospitalization and post-discharge; no assessment of pharmacokinetics/biology for new therapies; and poor sensitivity of outcomes (e.g., Glasgow Coma Scale) [70, 71]. These limitations may hinder the potential of mitochondrial regulation in TBI, although numerous initiatives are being taken to overcome them [71, 72].

Given the complexity of mitochondrial functions, current literature indicates that researchers can combine tools to optimize studies exploring mitochondrial mechanisms in different physiological contexts, including TBI. Hence, considering the appropriate sample and striving to combine different analyses encompassing the main functions of mitochondria is of utmost importance. This review reinforces a mitochondrial diagnostic model in preclinical rodent studies based on various guidelines for mitochondrial function analysis [73, 74]. Firstly, Percoll density gradient centrifugation [75, 76] can provide mitochondria-enriched samples from isolated nerve terminals and extra-synaptic mitochondria from the whole brain or single hemisphere [77], or tissue homogenates from specific brain regions [42, 78], albeit density gradient centrifugation may limit small brain region assessments due to low mitochondrial yield. Recently, the fractionated mitochondrial magnetic separation (FMMS), employing magnetic anti-Tom22 antibodies can provide better yield in obtaining isolated functional synaptic and non-synaptic mitochondria populations from mouse cortex and hippocampus [79]. Subsequently, following the combination with a respiration buffer [42, 78], oxygen consumption, and ATP synthesis capacity are evaluated through high-resolution respirometry. Further, spectrophotometric assays for other important aspects (i.e., hydrogen peroxide synthesis, membrane potential dynamics, and calcium handling capacity) complement this analysis. Additionally, techniques to assess glucose uptake and cellular signaling, such as immunoblotting or immunohistochemistry, can add mechanistic insights to explain the outcomes (Fig. 5).