Inflammation in Traumatic Brain Injury

Abstract

There is an increasing evidence that inflammation contributes to clinical and functional outcomes in traumatic brain injury (TBI). Many successful target-engaging, lesion-reducing, symptom-alleviating, and function-improving interventions in animal models of TBI have failed to show efficacy in clinical trials. Timing and immunological context are paramount for the direction, quality, and intensity of immune responses to TBI and the resulting neuroanatomical, clinical, and functional course. We present components of the immune system implicated in TBI, potential immune targets, and target-engaging interventions. The main objective of our article is to point toward modifiable molecular and cellular mechanisms that may modify the outcomes in TBI, and contribute to increasing the translational value of interventions that have been identified in animal models of TBI.

Traumatic brain injury (TBI) is an all-encompassing term to describe an intracranial injury after an external mechanical force results in damage to the brain structure or alteration in its functioning [1]. Both the causes and courses of TBI are heterogenous. Mechanisms of injury resulting in TBI are varied, including blunt forces (from bumps to blows), inertial loading, penetrating wounds, and blast injuries [1]. Just as the causes of TBI vary, the extent of damage also varies. Immediate injury after TBI can result in complications ranging from an unnoticed bruise with transient deficits to bleeding, or brain dysfunction resulting in permanent disability, or even death [2]. In evaluation of potential TBI patients, it is important to gain an accurate history of injury and complications during injury (including loss of consciousness), evaluate and treat symptoms [2], and to perform serial assessments in the post-injury period [3].

Etiological and pathophysiological heterogeneity of TBI are two of the major factors that produce inconsistent results between animal research and randomized clinical trials. It is essential to identify precise mechanisms for the primary [1] and secondary injuries associated with TBI to adequately target specific subgroups that are characterized by shared pathophysiology. After TBI, the cellular cascade of inflammation could potentially have beneficial effects [4–7]. However, if inflammation is too intense, prolonged, and unremitting, it may be harmful. Secondary injury has a slower progression ranging from minutes to years following the mechanical insult and is driven by several pathways, including: 1) free radical generation causing damage to molecular structures of cellular membranes [8], 2) excitotoxicity secondary to excess of glutamate [9, 10], and 3) systemic and central nervous system (CNS) immune activation [11]. It has been hypothesized that immune activation and modulation can significantly change the clinical and functional course in TBI [12]. Within minutes of impact, glia and recruited immune cells generate signals that initiate an inflammatory cascade [13]. The activation of the immune system has a vital role in cleaning the areas of the primary injury and preventing subsequent infections. However, to this day, targeting immune activation has been unsuccessful in improving clinical outcome in patients with TBI [14, 15], perhaps because of the greater heterogeneity of the causes of inflammation and their interactions in clinical TBI as compared to animal models, as well as genetic, environmental, pharmacological, and cultural differences between participants. The complex immunological mechanisms, including the influence of interacting immune-modifying factors involved in human TBI, presents challenges in developing an adequate animal model.

Interventional paradigms may be designed to:

1. Target early inflammatory processes that stimulate beneficial effects and inhibit detrimental ones.

2. Modulate overly prolonged or intense immune activation.

3. Clarify how immune challenges before or following TBI (e.g., reactivating infections, allergen exposure in atopic individuals, or exacerbation of autoimmune disorders) may affect its course and outcomes.

A better understanding of the complex molecular cascade of inflammation and immune function post-TBI is vital for both patient-specific diagnostic interventions and development of novel treatments, as well as in developing coupling between animal models and clinical interventional studies.

SEVERITY OF TBI

The methods of assessing the severity of TBIs in a variety of clinical settings include the mechanism [closed versus (vs.) penetrating], the clinical severity, and the type and degree of structural damage [16]. The most common tool used to evaluate TBI severity in the clinical setting is the Glasgow Coma Scale (GCS). The GCS includes assessment of best verbal response, best eye-opening response, and best motor response, with an overall score range of 3–15, with 3 being the most severe and 15 the least severe [1].

The three major categories of TBI severity are mild, moderate, and severe, often dichotomized into mild vs. moderate/severe.

EPIDEMIOLOGY OF TRAUMATIC BRAIN INJURY

TBI is a significant public health problem. It has been referred to as the “silent epidemic”. Yet, even if underestimated, the epidemiological characteristics of TBI are staggering. The lifetime prevalence of TBI is up to 40% of adults [17]. The global estimate of TBI indicates an incidence of 64 million and 74 million new cases yearly worldwide, with the highest incidence in USA, Canada, and Europe [18]. Overall, the most common causes of TBI are falls and motor vehicle accidents [19]. TBI incidence peaks three times during the lifespan: childhood, adolescence, and older adulthood [17]. TBI is the leading cause of death in children over 1 year old worldwide [20]. The estimated incidence of TBI in children aged 0–14 years old is about 475,000 in the United States [21]. The geriatric population incurs the most TBI-related hospital visits and deaths. Given the aging population, the number of older adults afflicted by TBI is likely to continue to increase [17]. Military personnel are also at particularly high risk due to combat, accidents, and activities during training [22].

ANIMAL MODELS OF TRAUMATIC BRAIN INJURY

Presenting the immune changes induced by TBI, common for all models and specific to each animal model in detail, is beyond the scope of the current article. Yet, we will briefly describe specific models with some of the inflammation findings uncovered using those models. For a presentation in depth, the reader is directed to a recent review of the topic [23].

Given the complexity of TBI, it is not surprising to see a variety of approaches that aim at modeling it in animals [24–26]. Animal models of TBI are commonly designed and classified by how the head injury is induced in the animal subject, rather than by its severity [25, 26]. Severity of injury in a particular animal model depends on the specific mechanical impact parameters that are used in the experiment. Therefore, in most cases, the same animal model can be used to represent both severe and milder forms of TBI, depending on the particular parameters used in the study [27]. At minimum, animal models of TBI can be classified by a single distinguishing factor, namely, closed vs. open head models. In the closed-head injury models, the animal’s skull is left intact, while in open-head models, craniotomy or craniectomy is performed to open a window on the skull to reach the brain for induction of TBI. Neuroinflammation is a key feature of all animal TBI models. Proliferation of reactive astrocytes and microglia, which is a hallmark cellular response to TBI and is often evidenced by upregulation of certain proteins, such as glial fibrillary acidic protein (GFAP) and ionized calcium-binding adaptor molecule 1 (Iba1), along with increased levels of proinflammatory cytokines and chemokines, has been consistently observed across different types of animal models of TBI [23, 28, 29]. Another common finding is that repeated injuries exacerbate the neuroinflammatory response across different animal models [1, 30].

Closed-head models

Most closed-head animal TBI models require a blunt impact to the head for the induction of TBI. Some models only mimic the rotational forces that the brain is subjected to during TBI but not the mechanical impact [31–35]. Blast-injury models can be considered impactless since brain injury is achieved by the shock wave from the blast rather than a mechanical blow to the head [36]. The critical factor in interpreting the results of these animal models of TBI is the possibility of contusion, hematoma, and skull fractures. Presence of hematoma or contusion can change the course of the immune response and should be taken into account when the results are interpreted [37, 38].

In a weight drop model of TBI, a solid object falls freely in a guiding tube from a predetermined height onto a subject’s head, which is usually anesthetized during the procedure to induce TBI [39–42]. Enhanced reactive gliosis and astrogliosis have been shown with several variations of the weight drop model in both mouse and rat subjects [43–46], in addition to increased proinflammatory cytokine [e.g., interleukin (IL)-1, tumor necrosis factor-alpha (TNF-α)] levels and their gene expression [41, 47]. These neuroinflammatory responses are enhanced with repeated injuries [1, 44]. Another method to induce closed-head TBI is the utilization of pneumatically or electronically controlled piston as the impactor instead of a free-falling weight [48–52]. Similar to the findings in the weight drop models, drastically increased GFAP and Iba1 levels have been reported in this TBI model [53–55]. However, elevation of proinflammatory cytokines has not been consistently observed with this model [53, 54, 56].

An alternative to the weight drop- and piston-based animal models of TBI are the projectile impactor models. These models use accelerated projectiles as impactors and have the ability to model the head impacts from high-speed small mass objects [57–61]. Using this approach, it has been demonstrated that repeated impacts to the head causes extensive increases in neuroinflammatory markers and microgliosis, including increased M1-type microglia [61, 62].

The blast-induced TBI model is another example of a closed-head TBI model. In blast-induced TBI models, the animal is secured in a blast tube where a shock wave travels and reaches the animal’s head to create TBI [36]. An acute neuroinflammatory response occurs acutely in blast-induced TBI models and levels of various chemokines and cytokines [e.g., IL-1α, IL-6, IL-10, interferon-gamma (IFN-γ), monocyte chemoattractant protein-1] increase in a short time frame (2–4 h after blast TBI) [63, 64].

Open-head models

The controlled cortical impact (CCI) model is one of the most utilized animal models of TBI that was originally developed in ferrets and then applied predominantly to rodents [65, 66]. In CCI, anesthetized animals undergo a craniectomy or craniotomy procedure, depending on the experimental design, and the dura is exposed [65, 66]. During the injury, a piston impactor moving at a predetermined speed hits the exposed dura to create TBI. Modified versions of CCI in which the skull is kept intact and the piston impacts the skull directly should be considered a separate model [67]. CCI was originally thought to produce focal injury around the impact site, but later studies showed that it also produces more diffuse injury and neurodegeneration throughout the brain distant from the cortical impact site [68]. CCI has been shown to induce acute increases in chemokines (i.e., CXCL₁) and cytokines (i.e., IFN-γ, TNF-α) and also certain anti-inflammatory cytokines (IL-4, IL-13) along with activated microglia [69, 70]. Fluid percussion injury (FPI) is another commonly used open-head TBI animal model [71–74]. FPI and CCI are similar in terms of their application of craniectomy to open a window on the skull. Through this opening, however, instead of blunt force impact from a piston, rapid and transient hydraulic pressure is applied through the opening in the skull to create TBI [75]. FPI induces both focal and diffuse injury as well [25, 75]. FPI increases activated microglia and macrophages in the brain [76]. In a more recent study, FPI has been shown to initiate a neuroinflammatory response acutely and increase levels of various chemokines and cytokines in the brain [77].

Even though relatively uncommon among civilian TBI patients, penetrating injury to the brain from a ballistic high-speed object (i.e., a bullet) is a particular concern for the military. The penetrating ballistic-like brain injury (PBBI) model was developed to address this particular type of injury [78]. In PBBI, similar to CCI and FPI, a skull window is opened through craniectomy. After that, a probe with an inflatable tip is inserted into the brain. To create injury, the probe tip is inflated very rapidly. This mimics effects of a bullet wound in the brain [78]. PBBI has been shown to induce loss of brain tissue, extensive intracranial hemorrhages and edema [78–80]. Neuroinflammation after PBBI is widespread and occurs rapidly, and given the major disruption of the blood-brain barrier (BBB), the central events are rapidly mirrored in the periphery. Levels of proinflammatory cytokines (IL-1β, IL-6, IFN-γ, TNF-α) and markers of microglial reactivity increase acutely after PBBI [81, 82].

THE IMMUNE SYSTEM: A BRIEF SUMMARY

The immune system includes structures, cells, and molecules that together are charged with defending the self against pathogens, malignant cells, transplanted tissue and, most commonly, microorganisms, as well as with eliminating damaged tissue that consumes resources and represents a risk of systemic toxicity, and ultimately sepsis. The innate immune system is characterized by rapid and non-specific responses and limited capacity to cascade the response based on the memory of previous encounters with antigens. It plays an essential role in preventing or limiting the invasion by microorganisms, and in the removal of debris, wound healing, and tissue remodeling. The innate immune system includes anatomical barriers (e.g., skin, mucosa) and physiological gradient barriers (e.g., pH, temperature), antimicrobial peptides, cellular effectors such as innate lymphoid cells (e.g., natural killer cells), phagocytes (e.g., neutrophils, macrophages, microglia, and dendritic cells), and granulocytes (e.g., eosinophils, basophils, mast cells, and neutrophils) that mainly release immune molecular mediators from their granules [83, 84]. To sense danger signals or pathogens and to trigger immune responses, these martial cells use “pattern recognition receptors”, including nucleotide-binding oligomerization domain (NOD)-like receptors (NLRs), Toll-like receptors (TLRs), and scavenger receptors [83, 85]. Chemokines and cytokines are released to amplify and initiate a cascade of immune responses and recruit other immune cells. Distal signaling via the vascular system and neural input to the brain may affect behavior, ultimately centralizing behavior towards the bare essentials necessary to survive the threat (i.e., “sickness behavior”, involving an overall retreat from the environment, a decrease in energy and purposeful activity, an increase in sleepiness and propensity to sleep, and decrease in appetite and food intake).

Adaptive immunity involves the cascading activation and expansion of T and/or B lymphocytes targeting specific antigens, and a memory of previous antigen exposure allowing a targeted, intense, and rapid response to future antigen challenges [86, 87]. T lymphocytes (CD4⁺ and CD8⁺ T cells) have cytotoxic, helper, or regulatory functions, and are activated and reactivated by recognizing antigens displayed in major histocompatibility complexes. B lymphocytes produce immunoglobulins and have reciprocal interactions with T cells and cells of the innate immunity [88]. Adaptive immunity is also involved in autoimmune responses, as well as atopy.

SPECIFIC COMPONENTS OF THE IMMUNE RESPONSE TO TBI

The timeline of immune response to TBI, reflecting the elevation of various components of the immune system, is summarized in Fig. 1. Several methods, including pharmacological probing, targeting with antibodies, and utilizing genetically modified animals, have been used to infer the role of molecular signaling pathways and cellular immunity implicated in TBI (described in Tables 1A and 1B). Molecular pathways and cell types involved in the inflammatory response to TBI are well-described in other mechanism-oriented reviews [11, 30, 89, 90]. Below, we list some of the components of this immune response.

Fig. 1.

Timeline of the immune response to TBI. After impact, rapid tissue release of damage-associated molecular patterns (DAMPs) prompts resident cells to secrete chemokines and cytokines. These molecules attract neutrophils, which contribute to circumscribing the injury site and promoting the removal of damaged and dead cells and debris. Infiltrating monocytes and activated glia begin to predominate 3–5 days post-injury to defend against infection and perform reparative functions. T and B cells can also be recruited to sites of brain pathology at later time points in the response (3–7 days post-injury). The timeline is modeled to human TBI across modalities, and is likely to correspond to the great majority of animal models of TBI (Reproduced from McKee CA, Lukens JR (2016) Emerging roles for the immune system in traumatic brain injury. Front Immunol 7, 556).

Table 1.

Rodent models used to characterize the role of signaling pathways and immune cell types in TBI. Table 1A describes molecular signaling pathways and Table 1B describes the cellular components of the immune system in TBI, illustrating approaches based on antibody targeting, as well as genetically modified mice (Reproduced from McKee CA, Lukens JR (2016) Emerging roles for the immune system in traumatic brain injury. Front Immunol 7, 556)

Table 1A. Molecular signaling pathways
Inflammatory mediator	Animal line/model	Purpose	Major findings in animals exposed to TBI	Reference
IL-1	Anti-IL-1β antibody	Blockade of IL-1β signaling	Reductions in macrophages/microglia, neutrophils, and T cell numbers in the brain, improvement in learning tasks, and decreased tissue loss	[289, 290]
	IL-1 R antagonist	Neutralize IL-1	Higher expression of proinflammatory cytokines in macrophages	[291]
ASC	Anti-ASC	Limit inflammasome assembly	Reduced caspase-1 activation and IL-1β production, decreased lesion volume	[292]
	ASC knockout	Abrogate inflammasome assembly	No improvements in lesion volume, histopathology, cell death, or motor function	[293]
NLRP1	NLRP1 knockout	Prevent NLRP1 inflammasome assembly	No improvements in lesion volume, histopathology, cell death, or motor function	[293]
IL-6	IL-6 knockout	Ablation of IL-6 signaling	Fewer reactive astrocytes and macrophages, increased neuronal death	[294]
	IL-6 knockout	Ablation of IL-6 signaling	Poor behavioral performance, higher IL-1β levels in the cortex	[295]
	GFAP-IL-6 overexpression	Increase IL-6 expression in astrocytes	Greater recruitment of glia and immune cells to the lesion, decreased oxidative stress and neuronal death	[296]
	Anti-IL-6 antibody	Neutralize IL-6	Reduced some inflammatory and behavioral effects of post-injury hypoxia	[297]
TNF-α	TNF-α inhibitor post-TBI	Inhibit TNF-α signaling	Early administration improved cognitive performance, and decreased neuronal apoptosis and astrogliosis	[298]
	TNFR1 knockout	Disrupt TNF-α signaling through TNFR1	Improved neurological function and neuronal survival/lesion volume, decreased numbers of CD11b+cells in the brain	[299]
	TNFR2 knockout	Reduce TNFR2 signaling	Worsened neurological function and no protection from tissue loss	[299]
	TNFR2/Fas knockout	Abrogate TNF-α signaling through TNFR2	Impaired motor and cognitive performance	[142]
G-CSF	G-CSF injection post-TBI	Enhance G-CSF signalling	Improved cognitive performance and increased hippocampal neurogenesis, higher glial activation and production of BDNF and GDNF	[300]
GM-CSF	GM-CSF knockout	Disrupt GM-CSF signalling	Worsened cognitive deficits as well as cell and tissue loss, reduced astrogliosis	[301]
Type 1 IFN	IFNAR knockout or IFNAR blocking antibody	Block type 1 IFN signalling	Reduced lesion volume, more anti-inflammatory cytokine signaling, increased glial activation, these effects were hematopoietic cell-dependent	[302]
IL-10	IL-10 knockout, IL-10 injection	Modulate IL-10 signaling	Diminished protective effects of hyperbaric oxygen treatment, including lesion volume, edema, cognitive improvement, and decreased cytokine production in IL-10 knockout mice, while IL-10 injection improved these outcomes	[303]
TGF-β	TAK1 inhibition	Disrupt signaling downstream of TGF-β	Improved neuronal survival and motor function, decreased NF-κB signaling and inflammatory cytokine production	[304]
	TGIF shRNAv knockdown	Ablation of downstream TGF-β signalling	Decreased infarct volume and microglia numbers, improved motor function	[305]
APOE	APOEϵ4 overexpression	APOEϵ4 overexpression	Worsened brain pathology, BBB breakdown, and neurological impairments	[306, 307]
TREM2	TREM2 knockout	Abrogate TREM2 signaling	Altered macrophage distribution, hippocampal neuroprotection, and fewer cognitive deficits	[308]
Table 1B. Cellular mediators
Cell type	Animal line/model	Purpose	Major findings in animals exposed to TBI	Reference
Neutrophils	IgM RP-3	Neutrophil depletion	No significant decrease in BBB permeability	[151]
	Anti-Gr1 antibody	Neutrophil depletion	Decreased edema, apoptosis, and microglia/macrophage activation, no significant changes in BBB integrity	[153]
	CXCR2 knockout	Reduce neutrophil infiltration	Decreased cell death, no significant changes in BBB permeability or behavior	[148]
	Neutrophil elastase knockout	Reduce neutrophil effector functions	Decreased edema and apoptotic neurons, but no decrease in tissue volume loss or behavioral improvement	[154]
Macrophages and microglia	CD11b-TK	Deplete CD11b-expressing cells	Reductions in microglia numbers in the brain, no improvement in axonal injury, treatment toxic at high dosage	[309]
	CD11b-DTR	Deplete CD11b-expressing cells	No change in lesion size, treatment caused inflammatory response without injury	[310]
	CCX872 (CCR2 antagonist)	Reduce CCR2 signaling functions	Reduced macrophages in the brain, altered pro- and anti-inflammatory cytokine expression, less cognitive dysfunction	[162]
	CCR2 knockout	Limit CCR2-mediated recruitment of monocytes	Reduced numbers of infiltrating monocytes, improved learning and memory	[161]
	CCR2RFP/RFP	Disrupt recruitment of monocytes	Reduced monocyte recruitment, cavity volume, and axonal pathology	[160]
	CX3CR1 knockout	Abrogate CX3CR1 signaling functions in macrophages and microglia	Short-term neuroprotection and lower inflammatory response, long-term functional impairments and elevated myeloid cell activation	[311]
T cells	Rag1 knockout	Genetic ablation of B and T cells	No changes in neurological outcome, BBB integrity, pro- or anti-apoptotic mediators, hippocampal architecture, or astroglial activation	[312]
	FTY720	Sequester lymphocytes and reduce their migration to the brain	Decreased circulating lymphocytes, decreased neutrophils and macrophages/microglia in ipsilateral hemisphere	[313]

Early immune “danger” signaling in TBI

In the damaged CNS, the initiation of immune signaling is started by recognition of damage-associated molecular patterns (DAMPS), while microorganism invasion in the CNS is signaled by pathogen-associated molecular patterns (PAMPS) [91]. These molecules interact with receptors, such as TLRs, NLRs, and scavenger receptors that serve as “danger” sensors and help to initiate the inflammatory cascade [85, 92–94].

The DAMPS’ recognition leads to the activation of nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kB)-inducing kinase [94] and elevation of proinflammatory cytokines, such as TNF-α and IL-6 [95]. A reduced expression of inflammatory markers and smaller brain lesions after impact were observed in TLR4-deficient mice relative to wild-type controls [96], and less edema was observed after TBI after administration of VGX-1027, a selective TLR4 blocker [97]. Brain injury in humans and experimental animals induces expression of High Mobility Group Box 1 protein (HMGB1) [98], a protein binding to TLR4 and initiating inflammatory cascades [99–101], and its inhibition diminishes both inflammation as well as BBB breakdown [102]. Other proteins triggering inflammation after TBI are the protein chaperones involved in tissue reorganization and wound healing [103, 104], and have immunomodulatory and neuroprotective effects [105–108]. Other inflammation triggering and perpetuating mechanisms include the activation of the purinergic receptors by adenosine tri-phosphate (ATP) released in the extracellular space secondary to tissue damage, leading to activation of inflammatory cascades [109, 110] and promotion of neutrophil recruitment [111, 112], as well as the formation of inflammasomes in the cytosol [113, 114]. Multiple cells in the CNS—such as neurons, astrocytes, microglia, macrophages, and astrocytes—can contain assembled inflammasomes [115, 116], and levels of inflammasome proteins in the cerebrospinal fluid (CSF) generally predict a poorer outcome in TBI [117].

Inflammatory gene expression

After TBI, a program of immune gene expression is induced that determines the progression of the subsequent inflammatory cascade [118]. Qualitatively, genes implicated in antigen presentation [CD74, CD86, major histocompatibility complex (MHC) II], phagocytosis [C3, C4, Fc gamma receptor (FcgR) and FCGR4], astrocyte activation [Aquaporin (AQP)4 and GFAP), chemotaxis [chemokine ligand (CCL)2, CCL4, chemokine (C-X-C motif) ligand (CXCL)1 and CXCL4], cytokine signaling [IL-1β, IFN-γ, IL-6, IL-12, IL-10 and transforming growth factor-beta (TGF-β)] are, surprisingly, similarly upregulated in both mild and severe TBI [118], and thus, are unlikely to be key mediators of severity of TBI outcome.

Peripheral innate immune activation after TBI

It is known that the body responds to TBI by engineering both central and peripheral inflammatory responses [119–122], and that a single brain injury can elicit complex changes in the systemic immune system. In fact, the immunologic complex known as the systemic inflammatory response syndrome characterized by alterations in circulating leukocyte numbers, complement proteins, coagulants, and inflammatory cytokines [120, 122], seems to result from post-TBI disruption of the BBB and CNS vasculature along with consequent leakage of central inflammatory mediators into the periphery [119–122].

The complement system

Various complement components have been found in the CSF of TBI patients [123, 124] and correlate with the severity of BBB dysfunction [124], and interrupting the complement cascade in various animal models of TBI has been consistently found to be neuroprotective [125–132].

Cytokines

Cytokines are cellular messengers that are secreted by multiple cells implicated in TBI that can convey proinflammatory or anti-inflammatory signals to other immune and nonimmune cells [133]. Often produced during sterile injury immune responses, IL-1β is elevated in the CSF of patients with severe TBI for at least 24 h, with levels that are linked with a less favorable prognosis [134, 135]. This signal is not just a marker of severity, but likely a component of a mediating adaptive mechanism, as administering an IL-1β antagonist in animal models reduces neurological symptoms, improves neurological functioning, and reduces the lesion volume [136].

Similarly, IL-18 is detected early post-TBI [137] and administering IL-18 antagonists in rodents one hour post-brain injury led to improved symptomatic recovery [137]. Thus, IL-1β and IL-18 have been associated with negative outcomes post-TBI. In contrast, other cytokines have been associated with either a good or a bad prognosis. For instance, a classical proinflammatory cytokine, TNF-α, shows heterogenous links with outcome in TBI. On one hand, it is elevated in TBI patients and, moreover, neuroprotective effects can be elicited by its blockade in rodent TBI models [138, 139]. However, while less severe neurological symptoms and functional deficits occur one week after TBI in mice that are genetically deficient in TNF-α, greater deficits and motor symptoms were seen at 2–4 weeks after injury with larger lesions at later time points [140]. It also appears that the experimental model of TBI plays a strong moderating effect on the functional outcome of TNF-α inhibition. In particular, concomitant Fas receptor and TNF-α inhibition appears to lead to worse functional outcomes in a mouse concussion model with no structural brain damage based on impact and head acceleration, while in models of cerebral contusion, the Fas receptor and TNF-α inhibition improved functional outcome and reduced tissue damage [141, 142]. This suggests that TNF-α has a dual role in contributing to both tissue repair and recovery, as well as to neuroinflammation and excitotoxicity.

Anti-inflammatory cytokines, such as IL-10 and TGF-β, counteract the effects of proinflammatory cytokines and reduce neuroinflammation, and overall, improve the functional and symptomatic outcomes and lesion dimension. For instance, the cytokine IL-10 appears to be neuroprotective if administered early and peripherally (but not intrathecally) in a fluid percussion concussion [143], suggesting that its anti-inflammatory effect has to act, at least initially, on the peripheral immune system for beneficial effects to occur. Similarly, TGF-β1 appears to be neuroprotective in a weight drop model in rats, where the cytokine is expressed in perilesional neurons and astrocytes, and consistently, knockout of the TGF-β1 results in increased neuronal death, larger lesion size, and worse neurological deficits [16].

Phagocytes/glial cells

Post-TBI, myeloid cells in the CNS are first to be activated followed by neutrophils, which are usually among the first peripheral immune cells to be mobilized towards the site of injury within just a few hours (h) [111, 144–146]. There is a significant surge in the number of peripheral neutrophils in the early hours after TBI that lasts until 48 h post-injury [147]. Perivascular spaces and meninges seem particularly vulnerable to alignment of neutrophils after cortical injury [37, 111]. Promoting complexity, there are data to define neutrophils as neuroprotective [111, 148], actively involved in healing after neurological injury [149, 150], neutral [151, 152], or alternatively, neurotoxic [121, 151–154]. Similarly, monocytes and macrophages are directed by temporal and biological contextual factors towards either tissue damage or repair [155, 156].

The perivascular spaces and brain parenchyma become hosts for monocytes within the first two days after injury. These monocytes then differentiate into macrophages and remain in the same location for weeks after the original injury [157]. The chemokine CCL2 signals the deployment of monocytes [158, 159] after TBI and elicits CCR2 + monocyte recruitment. Some studies suggest that interfering with this pathway decreases the lesion size and may lead towards improved neurological outcomes in animal models of TBI [158, 160–162]. Yet, other studies demonstrate the repair capacity of macrophages after CNS injury [163–165]. For microglial priming, the reader is directed to discussion later in the article at the section on “Age effects/microglial priming.”

Astrocytes

Astrocytes respond directly and indirectly to TBI via HMGB1/Receptor for advanced glycation end products (RAGE) signaling and activation of the NF-kB pathway [98], leading to secretion of chemokines/cytokines and enhanced astrocyte phagocytic activity. Yet, post-TBI reactive astrogliosis limits the spread of damage and inflammatory response to the unaffected CNS area, with astrocytes also providing metabolic support for neurons and their synapses [166, 167]. Astrocytes interact with other immune cells, such as microglia, in releasing various growth factors like insulin like growth factor 1 (IGF1) and nerve growth factor (NGF) that promote post-TBI healing [4].

T cells

Brain injury results in a cell-mediated immune activation [168, 169] with recruitment of T cells in the brain [170] independent of their antigenic specificity. Detailing the adaptive immunity activation in detail is beyond the intended scope of this review, However, it is important to mention that while overreactive adaptive immunity has been thought to represent a negative prognostic factor, the CD4⁺ T cells autoreactive to CNS antigens appear to be mainly neuroprotective, in particular for injured axons [5–7].

LIPIDS, IMMUNE MODULATION, AND TBI

Polyunsaturated fatty acids (PUFA)

Lipids are a diverse molecular class, important for signaling and immunomodulation. A number of studies reported increases in levels of free fatty acids (FFA), including polyunsaturated fatty acids (PUFA), in tissues, CSF, and serum in both human TBI and experimental models of TBI [171, 172]. Notably, two independent mass spectroscopy imaging studies of TBI reported PUFA-containing lipid signatures within the injury site in rat models, including multiple phosphatidylcholines [173], diacylglycerols, and sphingolipids [174].

PUFAs are important substrates for the cyclooxygenase (COX), lipoxygenase (LOX), and cytochrome P450 (CYP) pathways, ultimately yielding a range of immunomodulatory molecules, such as prostaglandins, thromboxanes, leukotrienes, and resolvins that mediate signaling through G-protein coupled receptors [175, 176]. Two classes of PUFAs, ω-3 (i.e., docosahexaenoic acid, DHA; eicosapentaenoic acid, EPA) and ω-6 (arachidonic acid) are generally implicated in opposing anti- and proinflammatory functions, respectively, that have been extensively studied [177–179].

Multiple regimens for supplementation with ω-3 s have been implemented to limit putative immune-mediated pathology in various rodent models of TBI [180–183]. Furthermore, in rodent models of mild to severe TBI, treatment with DHA or fish oil (containing DHA and other ω-3 s) showed significant improvements in cognitive function and mitigation of neuronal damage [181–184]. In humans, there are no adequate data to make a determination about efficacy and safety of ω-3 s treatment in TBI, and despite anecdotal evidence, pathophysiological targeting, as well as advocacy [185], well-designed human clinical trials are necessary. Additionally, a theoretical possibility of impaired initiation of clotting, and thus, increased risk of bleeding due to decreased eicosanoid synthesis exists [186], even if this is disputed based on analogical reasoning [185], and to our knowledge, is unsupported by large-scale safety data.

Statins

Cholesterol-lowering drugs, in particular, statins, have also been implicated in reduction of TBI pathogenesis, as well as in anti-inflammatory effects. The primary effect of statins is inhibition of 3-hydroxy-3-methyl-glutaryl-coenzyme A (HMG-CoA) reductase (a rate-limiting enzyme in the mevalonate pathway), but there are well-recognized pleiotropic effects of statins, including direct immunomodulation [187].

The positive significant immune effects of statins in animal models are reductions of IL-6, TNF-α, and intracellular adhesion molecule 1 (ICAM-1), attenuation of cerebral edema and glial cell activation, and reduction of the permeability of the BBB [188, 189]. The upstream immunomodulatory mechanism of statins is the inhibition of TLR4 and NF-κB [188]. It has been shown that statins have an anti-apoptotic effect via downregulating the caspase pathways, which can be overly activated after TBI [190]. In animal models of TBI, statins have resulted in reduction of parenchymal hemorrhage [191], improved cerebral blood flow [192], decreased cerebral edema [193], improved neuronal survival and regenerative angiogenesis and synaptogenesis [194], improved neurological outcome [195], reduced intravascular thrombosis, improved spatial learning [196], and reduced cognitive deficits [197], including in severe forms of TBI [198]. In a small double-blind randomized trial in patients with moderate-severe TBI, 72 h after TBI, rosuvastatin significantly decreased the level of TNF-α but not of IL-1β, IL-6, or IL-10, and led to a reduction in the disability score [199]. Preinjury atorvastatin use leads to a reduction in mortality and improved functional outcomes in elderly exposed to TBI [200].

Consistent with the findings described above, a retrospective analysis found a significant reduction of hospital mortality in statin-using patients admitted for TBI, as compared to nonusers [201]. In contrast, a recent observational study did not confirm any benefit of statins in TBI [202]. Meanwhile, another study in Asia [203] did not find any benefit of a statin pretreated status in patients with moderate-severe TBI—a difference that could be attributed to: 1) using multiple statins rather than purely atorvastatin; 2) a predominantly Asian population; 3) potential pharmacokinetic differences; or, finally, 4) withholding the statin medications during the inpatient hospitalizations. In a randomized clinical trial of atorvastatin 20 mg as compared to placebo in patients after moderate-severe TBI (32 participants in each group), no statin–placebo difference in the volume of the contusion and its expansion was found at any time point, and yet, significantly superior scores on the Glasgow Outcome Scale and the Disability Rating Scale were observed in the statin group at the 3-month follow-up period, with no differences at baseline [204]. An important area of research is to evaluate the differential effects of statins that more easily penetrate the BBB (lipophilic ones) vs. those that act mainly at the periphery on neuronal cholesterol levels, neuroprotection, neuroproliferation, neuroinflammation, and oxidative stress. This will guide the selection of specific statins for specific molecular pathways implicated in TBI.

OTHER IMMUNE-TARGETING INTERVENTIONS IN ANIMAL MODELS OF TBI

Metabotropic glutamate receptor 5 agonists

Metabotropic glutamate receptor 5 (MGluR5) agonists, such as (RS)-2-chloro-5-hydroxyphenylglycine (CHPG), attenuate microglial activation by actions at MGlurR5 on microglia and neurons and inhibit caspase-dependent apoptosis. CHPG attenuates microglial activation through NADPH oxidase [205] with a potent neuroprotective effect observed early in treatment [206]. Moreover, a delayed administration reduced neurodegeneration and long-term inflammation. Specifically, if administered after 1 month, it lowered subsequent activation of microglia expressing NADPH oxidase subunits, inhibited hippocampal neuronal loss, and lesion progression measured by T2-weighted MRI (at one, two, and three months), as well as white matter loss measured by diffusion tensor imaging at four months post-TBI [207].

Minocycline

Minocycline, a common antibiotic, has been explored for its ability to attenuate inflammation after brain injury and improve prognosis. Minocycline is a lipophilic, tetracycline-derivative antibiotic that is most commonly associated with treatment of acne and various infections. Its anti-inflammatory and anti-edematous properties at higher concentrations were investigated in the treatment of TBI, given its favorable safety profile and ability to cross the BBB [2, 12, 208]. The drug decreases secretion of cytokines and chemokines that promote inflammation. Minocycline results in decreased downstream nitric oxide production and decreases the activation of macrophages and microglia, and decreases expression of multiple cytokines (TNF-α, IL-1β, IL-6, and MMP-9) [2, 90, 209]. Additionally, minocycline promotes neuroprotective responses after brain injury in mice [210]. Several experimental studies have shown promise for use of minocycline as an intervention after TBI. These results indicated that minocycline decreased cerebral edema lasting up to 24 h post-injury by decreasing immediate release of inflammatory markers like IL-1β and MMP-9, but with no difference in oxidative stress as measured by glutathione levels [208]. On a macroscopic level, minocycline has been shown to decrease gray and white matter changes after TBI in mice [209]. Lesions after TBI are tempered, as evidenced by decreased focal brain lesion volume and decreased olfactory bulb atrophy in treated mice [209, 211]. Minocycline was able to attenuate memory deficits, an effect that was sustained until at least 13 weeks post-injury [212]. Other studies have indicated mice with minocycline treatment after TBI perform better on the tasks of spatial learning and memory [12].

Considering the difficulty in extrapolating the results from animal studies into practice, there are a number of guidelines for future immune targeting interventions to be developed. Studies should be conducted across models and species to provide evidence of molecular or cellular target engagement, and should add histological (whenever possible) or imaging outcomes to the behavioral ones. Factors such as brain penetration as well as pharmacokinetic and pharmacodynamic measurements with optimal dosage should be identified [213] and the therapeutic window for any intervention intended to become a treatment should not be less than 6 h [213].

THE TIME COURSE OF TBI-INDUCED INFLAMMATION

As previously discussed, mechanical forces leading to TBI lead to direct, primary damage that results in axonal injury, contusion, or hemorrhage [16]. The secondary damage occurs hours to days after the initial insult [16] as a result of widespread neuroinflammation [11], oxidative stress [214], and excitotoxicity [9]. Within the inflammatory milieu, TBI is not an isolated event. Subsequent, co-occurring, and especially previous immune challenge and inflammation-mediated medical conditions interact with TBI with synergistic, antagonistic, or additive effects.

Both central and peripheral nervous systems are involved in the complex and dynamic inflammatory response induced by TBI, with genetics, location and severity of the trauma, secondary injury cascades, sex, and age as important moderators [11]. It has been proposed that modulation of the immune response could be central in identifying clinically relevant and effective therapeutic interventions for TBI, as all brain injuries trigger inflammatory responses [12].

A common immune phenomenon in the long-term recovery after TBI is the tardive enhanced inflammatory response that occurs post-injury. Experimental TBI results in changes that are usually coupled with prolonged elevation in glial/macrophage reactivity as a component of neuroinflammation [215–218], which is consistent with studies in humans that identified increased binding of PK-[11C](R)PK11195 (expressed by activated microglia) ligand on imaging between 11 months and 17 years post-injury [219], and higher mRNA expression of microglial markers CD68 and OX-6 at 1-year post-injury [220]. Many age-related neurodegenerative disorders, including Alzheimer’s disease, also have cell-mediated neuroinflammation as a prominent feature [221].

Age effects/microglial priming

Whereas microglia display an amplified proinflammatory response with aging that is referred to as “microglial priming” (first described in a model of prion disease) [222], macrophages display an age-related reduction in the proinflammatory response to an immune challenge [223]. Characteristics of a primed profile of microglia, in addition to morphological changes, include: 1) reduced activation threshold to express and release proinflammatory molecules; 2) exaggerated inflammatory response to immune challenge; and 3) increased basal expression of inflammatory markers and mediators [224]. Additionally, in older rodents, in contrast to younger rodents, although white matter demyelination injury induces a reduced macrophage/microglia response [225], a gray matter injury provokes an enhanced macrophage/microglia response [70, 226] (Fig. 2).

Fig. 2.

Traumatic brain injury-induced macrophage response varies in reaction to immune stressors that occur before, with, and after the injury. A solid black line and gray shading depicts normal, age-related health burden. A) Traumatic brain injury (TBI) occurring in the absence of Aβ (dotted black line) or tau results in acute macrophage-related neuroinflammation that subsides over time. TBI in the presence of Aβ (solid red line) results in an acute blunted macrophage response that increases at chronic post-injury time points. TBI in the presence of pathological tau (solid blue line) results in an enhanced macrophage response to TBI that remains elevated at chronic post-injury time points. B) Over time, macrophage-related neuroinflammation increases with normal health burden. Single TBI (dotted black line) results in acute macrophage-related neuroinflammation that subsides over time. Pre-injury peripheral immune challenge at sub-threshold levels (red line) attenuates the post-injury macrophage-related inflammatory response to TBI. Post-injury peripheral immune challenge (solid blue line) causes a hyper-active macrophage response correlating with behavioral dysfunction. Repetitive post-injury immune challenge (dotted blue line), similar to what is observed in repetitive TBI, increases macrophage-related neuroinflammation and correlates with the advanced neuropathology. The figure represents a model of what occurs in human TBI, translating specific findings in rodent TBI studies. (Reproduced from Kokiko-Cochran ON, Godbout JP (2018) The inflammatory continuum of traumatic brain injury and Alzheimer’s disease. Front Immunol 9, 672.)

PERSISTENT PROINFLAMMATORY PROFILES AFTER TBI

Reactive microglia

Brain trauma induces a persistent proinflammatory profile with increased inflammatory cytokines TNF-α and IL-6 (detectable in the CSF for up to 12 months and directly associated with measures of disinhibition and functional impairment [227, 228]) and priming in microglia [229–231] (Fig. 3). Increased CR3/43⁺ (MHC-II+) and CD68⁺ reactive microglia have been reported several months after brain injury in an immunohistochemical analysis of autopsied brains [232]. Especially, white matter in these autopsied brains had an increased immune load of microglial reactivity in the presence of traumatic axonal injury [232]. An association of CR3/43⁺ cells with increased myelin basic protein long-term after brain trauma (2–8 years) has also been reported in other studies [220].

Fig. 3.

Double hit model illustrates how during the first hit microglia become activated, and then potentially primed. A second hit may include any source of inflammation, such as intense stress, autoimmune disease reactivation, allergen exposure in atopic individuals, and most commonly, infections (acute, chronic, chronic reactivated) and/or repeat TBI. The primed microglia convey a risk for developing behavioral, affective, and cognitive disorders, and exposure to subsequent inflammation driven “hits” exponentially increases that probability, as well as symptomatic intensity and functional deterioration. The double hit model is expected to fully represent all TBI models in rodents, as well as its translation to clinical TBI, with the difference that in human TBI, a multitude of individual immune stressors interact with TBI in contrast to experimental animals living in conditions characterized by limited variation between individuals in exposure to various immune modifiers. This is a top-of-the-list possible explanation of why immune targeting interventions for TBI that are beneficial in rodents do not translate in humans, with few exceptions (e.g., aerobic exercise).

Multiple Immune Challenges and Multi-hit Models

Pre-injury peripheral mild immune challenges may improve recovery following TBI:

Experimental models of TBI have a reported improved behavioral recovery and reduced lesion volume after pretreatment with low doses of LPS [233]. Additionally, attenuation of IL-1β and TNF-α overexpression and neuronal loss was found in the hippocampus when LPS treatment was administered before TBI [234]. Even a higher dose of LPS administered over shorter duration (1.0 mg/kg, single intraperitoneal dose for 4 days) prior to a lateral cryogenic brain injury has been demonstrated to decrease lesion volume and neuronal death [235]. In sum, a pre-treatment inflammatory stimulus may help protect, in part, the CNS from a secondary, generally more potent, stimulus, such as TBI, likely via pre-conditioning leading to the induction of neuroprotective responses (see Fig. 2, upper graph).

Immune challenges after TBI

A primed microglial phenotype, defined by increased expression of MHC-II and CD68 and altered morphology is consistently induced by experimental models of TBI [231, 236]. Primed microglia show some increased reactivity at baseline, but a strongly increased reactivity after an immune trigger [30].

In rats that sustained TBI via impact acceleration, an exaggerated inflammatory response associated with cognitive impairments and depressive-like behavior 3 months post-injury has been reported to be induced by administering LPS, as early as 5 days after the primary injury [237]. Findings from these studies substantiate the fact that post-TBI immune challenges negatively impact symptoms and functional outcomes and lead to increased deficits, as well as highlight the long period of increased vulnerability to subsequent immune challenges. However, a decreased production of inflammatory cytokines and reduced neuronal injury were reported if LPS treatment was given 1 day after repetitive mild TBI (3 TBIs, 5 days apart) [43]. In contrast, if the administration of LPS was delayed to 5 days after repetitive TBI, worsening of neuronal damage characterized by aggregation and phosphorylation of tau, impaired behavioral recovery, and elevation of inflammatory cytokines was observed [43]. Together, these findings emphasize the substantial temporality of the immune response to TBI and its interactions with other immune triggers.

Common clinically related immune challenges, such as allergies [238, 239], autoimmune disorders [240, 241] and/or infections [242, 243], produce a proinflammatory response and lower the threshold for priming of microglia [244, 245]. Thus, further clinical studies should investigate the role of these immune-mediated conditions and their treatment in altering the course of TBI. Clinically, there is potential benefit to be uncovered from assertively diagnosing and aggressively treating these conditions when co-occurring with TBI.

Psychological stress, a common occurrence in reaction to TBI and events that contributed to it and surrounding it, has been consistently demonstrated to upregulate immune responses acutely and chronically, with secondary loops related to hypothalamic-pituitary-adrenal (HPA) axis dysregulation (with both suppression and hyperactivity) contributing to dynamic sequential changes in immune function [246] and progression of symptomatic and functional deficits. HPA axis suppression and the diminished modulatory role of HPA axis on immune alterations post-TBI, and vulnerability to stress, sleep deprivation, and circadian dysregulation, represent important domains that may provide translational value for minimizing early symptoms and long-term consequences after TBI [246].

Immune challenges in aged- and TBI-exposed mice lead to an exaggerated immune response

Microglial priming, characterized by altered morphology and higher expression of complement receptor-3 (CD11b) and MHC-II, has been reported in aged rodents [247, 248]. As compared to young adult mice being challenged with LPS, aged animals have an increased microglial expression of IL-1β upon induction of TBI, leading to a depressive-like phenotype [249] and prolongation of sickness behavior [250], as well as greater spatial memory impairments in parallel to the higher expression of IL-6, TNF-α, and IL-1β in the hippocampus [251, 252]. As compared to TBI-exposed mice that were given saline, TBI-exposed mice that received a peripheral LPS challenge 30 days after brain injury were reported to have an exaggerated microglial response, which was depicted by higher depressive-like behaviors and increased TNF-α, MHC-II, and IL-1β expression [231]. Moreover, memory recall deficits in TBI mice are exaggerated by administration of LPS 30 days after brain injury [236].

CHRONIC NEURODEGENERATIVE CONSEQUENCES OF TBI

History of TBI is predictively associated with future onset and severity of all forms of neurodegenerative diseases, including chronic traumatic encephalopathy (CTE), Alzheimer’s, Parkinson’s, Huntington’s, and vascular dementias [253], as well as psychiatric sequelae such as depression [254] believed to be, at least in part, consequences of microglial priming and perpetuation of neuroinflammation. However, complete replacement of primed microglia with nonreactive microglia failed to reduce astrocyte-mediated inflammation, and more importantly, markers of neuronal death (reviewed by Tapp et al. [246]). A single TBI produces long-term plastic changes in multiple aspects of the immune function (i.e., acute phase response, TLR signaling, NF-κB signaling, complement system) that take part in reciprocally activating and perpetuating immune molecular networks [255]. In repetitive TBI, the first TBI is a priming event, which creates the foundation for an elevated inflammatory response that occurs upon subsequent exposure to more brain injuries. It is important to note that TBI has persistent and profound effects on systemic immunity that show similarities to changes linked with aging [256].

Chronic traumatic encephalopathy

Development of CTE, a neurodegenerative disease characterized by abnormal tau accumulation in the sulci of the cortex, is an example of chronic cognitive impairment associated with repetitive TBI via participation in contact sports [257]. Even though some cases of CTE may have presence of amyloid-β (Aβ) plaques, their location and distribution are different from those of Alzheimer’s disease subjects [258–260]. A common feature of CTE is glial reactivity, including microglial reactivity as well as astrocytic accumulation of abnormal tau [257, 261]. In a study utilizing PET imaging, retired National Football League players were reported to have increased translocator protein binding in the entorhinal cortex, supramarginal gyrus, hippocampus, and the parahippocampal cortex, as compared to sex- and age-matched control subjects who did not have a history of repetitive head trauma [262]. In parallel to these findings, other studies on American football players showed increased CD68⁺ microglia in their brains, which could have partially mediated the deposition of abnormal tau [263].

A recent review describes the animal models of repetitive mild TBI and includes the discussion on available evidence related to the development of tau and Aβ pathology after the brain injury [264]. The number of injuries in repetitive TBI varies from 2–10 over a period of several days or weeks, but its severity is usually referred to as “mild.” Recently, a highly repetitive mouse model of TBI has been described that includes 30 injuries [265]. Generally, on one injury day, administration of 1 or 2 brain injuries is done. At chronic post-injury time points, consistent presence of increased behavioral impairment, amyloid-β protein precursor, and phosphorylated tau have been reported in non-transgenic mice exposed to repetitive mild TBI [265].

Alzheimer’s disease

Dementia risk later in life is elevated by history of TBI [266, 267]. Chronic inflammation induced by TBI has been suggested to be a major contributor to the development of neurodegenerative disease [220, 268]. Indeed, many studies have documented an association between TBI and vulnerability to dementia, with underlying mechanisms involving damage to white matter tracts and neuronal networks, impaired cognitive resources, and deposition of a Aβ and tau [266, 269], with neurotoxic forms of tau and amyloid leading to neuroinflammation and perpetuating a cascading vicious circle [270]. The interruption of this vicious circle by targeting persistent posttraumatic immune responses is a logical aim of future clinical trials in TBI [271]. The recent confirmation of long-term dysregulation of immunoregulatory genes provides optimism through targetable molecular networks that can be modified during a much longer therapeutic window, e.g., months after brain injury [255]. Future interventional studies in animals should consider strong interactive factors that can robustly modify the immunological consequences of TBI (e.g., dementia continuum), and the application in humans should account for immunological triggers and precipitants (by design, inclusion, and adjustment) that originate in the environment, lifestyle (e.g., exercise, sleep patterns), and the temporal dynamic of immune-mediated medical conditions [272].

MICROBIOME AND PROBIOTIC INTERVENTIONS

One factor contributing to increases in chronic inflammatory disorders in Western countries is thought to be failing immunoregulation, attributable to reduced exposure to the microbial environment within which the mammalian immune system co-evolved [273]. Our understanding of the “gut-brain axis” [274] is steeply increasing. The three proposed mechanisms bidirectionally linking the gut with the brain are: 1) changes in sympathetic sensorimotor function or vagal tone, 2) alterations in neuroendocrine signaling, and 3) alterations in immune signaling and gut, as well as BBB permeability [275]. The metabolites of the microbiome’s colonizing bacteria can be both anti-inflammatory and neuroprotective [274] and the gut microbiome has a prominent role in maintaining a delicate homeostasis of the immune system [276], and integrity of barrier functions. Rodents with germ-free gastrointestinal systems have decreased tight junction integrity (and increased BBB permeability) likely contributing to increased microglial activation [274].

Changes in gut microbiome are associated with increased neuroinflammatory and autoimmune reactions in multiple sclerosis patients [277] and have been implicated in other autoimmune-mediated diseases mediated by T cells including rheumatoid arthritis, inflammatory bowel disease, and type 1 diabetes [274]. Krezalek and colleagues coined the term “pathobiome” to describe changes in the microbiome after TBI [276]. There is a proposed feedback loop in which TBI and subsequent induction of the “pathobiome” result in a neuroinflammatory cascade, thus, further augmenting the effects of brain injury by cell death, BBB disruption, edema, and hemorrhage [276]. Changes in the gut microbiome occur as soon as 2 h after a TBI in rodent studies [276], with subsequent increases in permeability in the intestinal wall making TBI recipients susceptible to translocation of bacterial pathogens, resulting in exponentially increased local, systemic, and CNS inflammation [276].

Certain studies on probiotic administration have reported increases in anti-inflammatory cytokines such as IL-10 and decreases in inflammatory cytokine production in the intestine, and decreases in stressrelated hormone production induced by brain injury [276]. A 2017 review by Brenner identifies three studies indicating a benefit for the treatment of TBI [278]. Two randomized clinical trials in TBI patients in the ICU showed patients who received probiotics had decreased ICU length of stay, decreased infections, infection with fewer number of pathogens, treatment with fewer types of antibiotics, and decreased duration (in days) on artificial ventilation. As initial studies are promising in the use of probiotics in TBI patients immediately following their injury during hospital stay [279], further research on long-term effects and benefits across various severities and types of TBI is much needed.

In sum, in TBI, in contrast to many other neuropsychiatric domains, there is a disquieting disconnect between bench and bedside, with successful preclinical studies failing to translate [255]. It is commonly stated that the lack of efficacy demonstrated in human clinical trials of very promising interventions developed in animal models of TBI, including molecules directed at immune targets, is due to the incomplete representation of human TBI by the specific animal models used, specifically the much higher pathogenetic heterogeneity of clinical TBI. Yet, we see two factors that are equally important. First, human TBI does not occur in isolation—concurrent stressors, levels of immunomodulatory hormones and vitamins, variability in fitness and common endocrine dysregulation and metabolic syndrome, common infections, allergies and exposure to allergens, past TBIs, including barely noticed concussive or subconcussive events, exposure to heat and cold, and photoperiodic and circadian factors, are all known to modify immune response, and some are known to modify neuroprotection and neurogenesis. Second, many immune factors are highly context-dependent, with a good number of them playing either a protective or degenerative role, dependent on the time sequence, dose, and interaction with other immune stressors or immune-mediated clinical conditions, either as a trait or state. Reciprocal cascading interactions exist—while there is long-term immune activation in many TBI patients, TBI has an immediate peripheral immunosuppressive effect that is similar to advanced aging [256], and that can lead to vulnerability to acute infections and reactivation of chronic infections that lead to immediate increased mortality [280], and long-term negative outcomes in TBI [281, 282]. Prior brain trauma, prior extracranial trauma, polytrauma, and their interactions with infections [283], in particular nosocomial infections, may result in a different course of brain injury and immune contribution to brain injury than if it would happen in isolation. Sleep abnormalities are often present in TBI patients [284] and in animal models of TBI [285], and are known to affect immune function, functional changes of HPA axis, and stress vulnerability [246]. These have not yet been studied in interventional animal studies of TBI that could mimic more closely the human experience of TBI. These processes are currently oversimplified in animal models that generate treatment candidates but are not formally tested before being transposed to humans. Ideally, these factors would be analytically studied in animals, determining the specific association of immune-modifying factors in interaction, considering simultaneity, precedence, sequence, repetitiveness, and dose-response [272]. While life in experimental animals is forcefully homogenized, with identical genetic background, similar food, and sleep-wake and rest-activity opportunities, it is not the case in humans. Sources of heterogeneity include genetic variation, exposure to activities that modify the response of the immune system to TBI, such as acute, latent, and chronic infections, atopy with allergen exposure, autoimmune conditions, exercise, degree of general as well as context-specific psychological stress (e.g., realization of personal loss of physical, social and economical abilities via TBI and potential loss through demise or severe sequelae in others who have been co-exposed to the same trauma as the TBI patient). Additionally, regular or recent intake of common medications for common conditions that have immunomodulatory properties (such as statins, metformin and psychotropic medications) can markedly increase immune response heterogeneity in clinical samples. Additive or multiplicative interactions between these immune-modifying factors make the specific manipulation of candidate therapeutic molecules, identified in animal research and utilized in humans, to be overly simplified, and call for concentric interventions addressing individual combinations of interacting sources of immune dysregulation in animal models. Their translation to human studies requires large multicenter collaborations allowing the adequate number of participants to select specific interacting immune-modifying factors to be addressed by specific immune targeting interventions.

TAKE HOME MESSAGE

Attenuation of secondary injury has been a major goal of neuroprotective treatment, and despite strong experimental data, many trials for human TBI have failed to produce results [213]. In our evaluation, the complexity and interindividual differences of interactive factors that up or downregulate immune function prior to TBI in humans relative to the homogenous condition of experimental animals is the main cause of poor animal to human translation of significant interventional results with beneficial outcomes in animal models. The failure of the human trials to show consistency have also been attributed to: 1) considering secondary mechanisms in isolation rather than in interaction; 2) inadequate attention to animal model species, strains, age, gender, and timing; 3) the major difference of anesthetics being used in animal models; 4) the relative impermeability of the BBB; 5) the definition of meaningful translational behavioral outcomes; 6) epigenetic variability in humans resulting in pharmacogenetic heterogeneity; 7) not considering predicting biomarkers that would justify physiologically the need of the intervention in a particular participant; 8) inadequate sample size [213]; and 9) less diffuse axonal injury because of much smaller brains in rodents. Animal studies generally do not make pharmacokinetic and pharmacodynamic measurements, and do not try to optimize dosage to tissue concentration.

While we are noticing the disturbing lack of translation between targeting the immune system in animal models and clinical TBI, there is, however, one element of congruence—aerobic exercise that has shown significant benefits (cognitive performance) in patients with mild and moderate TBI [286], as it adequately translates the rodent study on pre-conditioning with exercise (daily running for four weeks) leading to reduced increases in IL-1β, TNF-α, and markers of BBB permeability and lesser decreases in IL-10 post-brain injury, as well as less motor function impairment after fluid percussion injury in rats [287]. Similar results have been found in a mouse study suggesting that there is a therapeutic window, specifically late after TBI (5 weeks after brain trauma) for reduction of longer-term inflammation and cognitive impairment [288]. This late administration of exercise is important given that there are concerns regarding aerobic exercise early after TBI, considering that, as shown previously in this article, immune activation appears to be beneficial in the early phase of TBI for cleaning of the impacted tissue, as well as for protection of infections, in addition to increasing brain oxygen demand and generation of additional free radicals with exercise. Furthermore, the later administration of an immunomodulating intervention is consistent with literature suggesting that targetable molecular networks can be modified during a much longer therapeutic window than initially thought, e.g., months after brain injury [255], although a more exact translation of the therapeutic window in humans is highly desirable. Finding the temporal sweet-spot for switching from immunotherapies geared toward clearance of debris, reparative and neuroprotective effects of inflammation, and those geared toward decreasing chronic unremitting inflammation [11], should represent a central topic for future research.