Potential of mesenchymal stem cells alone, or in combination, to treat traumatic brain injury.

Abstract Traumatic brain injury (TBI) causes death and disability in the United States and around the world. The traumatic insult causes the mechanical injury of the brain and primary cellular death. While a comprehensive pathological mechanism of TBI is still lacking, the focus of the TBI research is concentrated on understanding the pathophysiology and developing suitable therapeutic approaches. Given the complexities in pathophysiology involving interconnected immunologic, inflammatory, and neurological cascades occurring after TBI, the therapies directed to a single mechanism fail in the clinical trials. This has led to the development of the paradigm of a combination therapeutic approach against TBI. While there are no drugs available for the treatment of TBI, stem cell therapy has shown promising results in preclinical studies. But, the success of the therapy depends on the survival of the stem cells, which are limited by several factors including route of administration, health of the administered cells, and inflammatory microenvironment of the injured brain. Reducing the inflammation prior to cell administration may provide a better outcome of cell therapy following TBI. This review is focused on different therapeutic approaches of TBI and the present status of the clinical trials.

to what is observed in the real-world population with 49% having mild TBI, 34% having moderate TBI, and 17% having a severe injury.
About half of the study population did not return to their previous work after 1 year, and ~28% never returned to work of any kind. 8 Also, long-term disability is seen occasionally even in those with mild TBI. 9 Thorough reviews of TBI epidemiology have recently been published. 10,11 Thus, while TBI is a significant public health problem, unfortunately there is no single therapy that has proved efficacious in its treatment. Similar to the situation with other brain injuries (such as the failure of neuroprotective glutamate receptor antagonists and antioxidant treatments in clinical trials for stroke 12,13 ) and neurodegenerative diseases, there have been myriad-positive preclinical studies in TBI models and all of these promising therapies have failed in clinical trials. Various reasons have been advanced for these failures, including, but not limited to, differences in brain anatomy and physiology between rodents and humans, inadequate animal models, failure to test the treatment in a clinically relevant way coupled with failure to remain faithful to the preclinical testing parameters in the clinical trials, underpowered studies, heterogeneity of TBI injury, and insensitive outcome measures in both preclinical and clinical studies. There is no dearth of discussions in the literature identifying these shortcomings in the therapeutic development and testing of potential new treatments for TBI. [14][15][16][17][18] What we are left with for treatments is a general approach that is akin to crisis management. According to the current Brain Trauma Foundation Guidelines, based on the best available medical evidence for the management of severe TBI, it is imperative to provide adequate nutrition, support breathing by tracheostomy, and perform a large decompressive craniectomy. 19 The underlying problems for developing an effective treatment for TBI are 2-fold. First, the injury can be unique to the patient, depending on the type of TBI and the region of the brain affected. Second, once that injury occurs, a complicated neurodegenerative cascade is triggered; resolving any one of these pathological processes is not enough to prevent or terminate the others. In this review, we will discuss the pathophysiology of TBI with emphasis on immune and inflammatory function. We will also discuss the evidence for the development of a mesenchymal stem cell (MSC)-based treatment that can suppress immune and inflammatory/degenerative cascades and provide neuroprotection.

| PATHOPHYS I OLOGY OF TB I
The pathophysiologic mechanisms of TBI are poorly understood as the anatomy of the brain is uniquely complex with multiple cell types (neurons, astrocytes, oligodendrocytes, and microglia) and multiple subtypes of these cells. While we now know that neural stem cells (NSC) exist within the adult brain and some degree of axonal, dendritic, and synaptic plasticity occurs, we have yet to fully exploit the brain's regenerative capacity to repair an injury. Add to this the complicated neuronal networks throughout the brain and neural repair is a daunting task. Ramon Y Cajal, widely regarded as one of the fathers of modern neuroscience, stated in his treatise on neural development and regeneration "…once development was ended, the founts of growth and regeneration of the axons and dendrites dried up irrevocably. In the adult centers the nerve paths are something fixed, ended and immutable. Everything must die, nothing may be regenerated. It is for the science of the future to change, if possible, this harsh decree." 20 When we consider additional factors such as the confined space within the skull that contributes to increased intracranial pressure as edema develops and the blood-brain barrier (BBB), which can make therapeutic access to the brain difficult, the task of repairing the brain after TBI may seem insurmountable.
The pathophysiology of TBI occurs in two main phases-the primary insult and the secondary sequelae. The primary insult is the initial blow to the head. It can be a penetrating wound or a closed-head injury. The nature of the injury can be focal, involving a very circumscribed area of the brain, or it can be more diffuse causing widespread axonal injury. Then, there is the special case of the coup-countercoup injury as the brain impacts the skull in two sites on opposite sides of the brain; this is observed with trauma that involves deceleration of the head. Depending on the nature of the trauma, neurons, astrocytes, and oligodendrocytes may be destroyed, bleeding may occur, axons may be severed, and a contusion may form. 21 The immediate pathological consequences of these injuries are similar to those observed with cerebral ischemia-excitotoxicity, changes in ion flux (Ca ++ , Na + , and K + ) across the cell membrane, loss of ATP, lactate production, induction of cortical spreading depression, cytokine production, and loss of barrier function at the BBB. [22][23][24] The severity of the initial blow and the immediate pathophysiologic changes that occur determines the severity of any subsequent secondary degenerative processes. In the acute period, neurons and axons continue to die, and there is damage to the vascular endothelium; this, in turn, allows blood components to leak into the brain parenchyma, including peripheral immune cells, which then contribute to the pro-inflammatory environment. Astrocytes swell and tissue edema occurs. If edema is not controlled, intracranial pressure increases, which can lead to compression of arteries and decreased cerebral blood flow; cerebral ischemia commonly occurs under these conditions leading to a vicious cycle of increasing edema, increasing intracranial pressure, and increased ischemia that can lead to death.

| INFL AMMATI ON IN TB I
The main architects of the local inflammation at the site of injury are the microglia. Once a TBI has occurred, it is the microglia that proliferate and migrate to the site of the damage. As these microglia work to remove the cellular debris at the lesion site, they produce cytokines and chemokines that activate pattern recognition receptors to bind damage-associated molecular patterns, and attract and polarize peripheral immune cells. The first peripheral immune cells into the damaged tissue are neutrophils 25 followed 24-48 hours later by monocytes or macrophage, and T cells all of which are releasing cytokines and chemokines. Once the peripheral immune cells have established a strong pro-inflammatory response in the brain in the acute to subacute stage of TBI, then tissue damage is likely to be exacerbated. Further, the T cells may become activated through antigen presentation by microglia and macrophage, mobilizing the adaptive immune system 26 and potentially leading to autoimmunity. 26,27 In addition to local neuroinflammatory processes within the brain after TBI, systemic immune and inflammatory processes are also impacted and have recently been reviewed. 28,29 Peripheral immune cells are mobilized from the bone marrow, 30 thymus, 31 and spleen 31 early after the TBI. While there is a brief recovery in thymocytes and classical monocytes in thymus and spleen within the first 2 weeks postinjury, over time these cell populations decline again and may be responsible for the post-TBI immune suppression that is observed in patients. 29,31 Also implicated in the peripheral immune response and long-term immune suppression is both the activation of the hypothalamus-pituitary-adrenal axis, through the release of glucocorticoids and the sympathetic nervous system, through release of catecholamines. 29,32 Ideally, inflammation will be a self-limiting process. There are endogenously formed products of arachidonic acid metabolism that actively inhibit pro-inflammatory responses. 33 These include the lipoxins, resolvins, protectins, and maresins, which decrease pro-inflammatory cytokine secretion, 34 alter migratory signals for peripheral immune cells, 35 and stimulate neuroprotective and tissue regeneration processes. 36 There has been very little research on these resolving mediators in TBI, but administering the lipoxin, LXA4, into the lateral ventricles 10 minutes after a TBI induced by weight drop, reduced pro-inflammation, BBB disruption, and lesion size. 34 Further, its receptor, ALX/FPR2, is upregulated in astrocytes. 37 Unfortunately, in many instances inflammation may not resolve and becomes chronic. 38 In the aged brain, there is increased recruitment of peripheral macrophage into the TBI brain. 39 In addition to cellular infiltration of pro-inflammatory immune cells after TBI, sustained complement C3 activation leads to chronic inflammation by activating microglia and astrocytes in the region around the initial lesion and contributes to further neuronal loss 30 days post-TBI. 40 Using a combination of magnetic resonance imaging, magnetic resonance spectroscopy, and positron emission tomography in a rat lateral fluid percussion model, inflammation was shown to be present still 6 months post-TBI. 41 Even 12 months after a controlled cortical impact in mouse, there is increased immunolabeling for IBA1 (microglia) and glial fibrillary acidic protein (GFAP). 42 What's more, this chronic inflammation is associated with continued behavioral deficits.

| Pharmaceuticals
The focus of current pharmacological interventions after TBI is to manage level of consciousness, neuropsychiatric, neurocognitive, and neurobehavioral symptoms that may arise. 43 With the occurrence of so many interconnected neuro-immune-inflammatory pathologic cascades engaged after TBI, it is not surprising that therapies targeting one specific degenerative pathway have failed to demonstrate efficacy in clinical trials. Drug interventions that have been studied can generally be categorized by their therapeutic target. One class of drugs are those that prevent calcium ion flux. An example of such an approach is the calcium channel blocker, Nimodipine, which showed promising effects in rodents but exhibited only a small effect on TBI patients. 44,45 Disruption of intracellular calcium signaling may also improve outcome after TBI. Cyclosporin is a T-cell immunosuppressant that acts by binding to cyclophilin; the cyclosporine-cyclophilin complex binds to calcineurin preventing dephosphorylation of NFAT, translocation to the nucleus, and increased transcription of interleukin (IL) 2. 46 In the absence of cyclosporine, calcineurin is regulated by calcium and calmodulin. Cyclosporin prevents calcium ion transport into the mitochondria in animal models but not in TBI patients. 47 There are also a number of studies that have targeted excitotoxicity, specifically glutamate release and overstimulation of the NMDA receptor. For example, Selfotel is a NMDA antagonist and the first glutamate antagonist to enter into Phase III clinical trial.
This trial was discontinued because of high mortality and a failure to improve Glasgow Outcome Score. 48 The results from clinical studies of other NMDA receptor antagonists also failed to demonstrate any efficacy of treatment. 49 While not a true antagonist, magnesium blocks the NMDA receptor calcium channel. Increasing available magnesium also had no effect on TBI outcome. 50 Another approach has been to target oxidative stress produced by oxygen radical formation and lipid peroxidation. The lipid peroxidation inhibitor, Tirilazad mesylate, which is an approved drug in Europe to treat aneurismal subarachnoid hemorrhage, showed promising neuro-and vaso-protective responses in animal models of moderate-to-severe TBI but failed to show improvement over placebo control in Phase III clinical trials involving human TBI patients. 51,52 Pegylated superoxide dismutase, a free radical scavenger, was found to be effective in preventing secondary injury in preclinical and Phase I clinical studies but failed to show reduction in mortality or improve neurologic outcome in Phase III trials. 53,54 Another antioxidant that may be promising is N-acetylcysteine. When administered to patients within 24 hours of mild TBI, symptoms were significantly better compared to a placebo-controlled group. 55 Another target that has been examined is the treatment of inflammation with corticosteroids, statins, 56 cannabinoids, and bradykinin B2 receptor antagonists. 50 In addition, the gonadal hormones, estrogen and progesterone, both showed promising results in preclinical studies but failed to show beneficial effects in clinical trials. [57][58][59] The more recent approach has been to search for potential therapies that target more than one pathway. One such strategy is to use a pharmacologic that interacts with multiple receptor types, which, thereby, produces more than one effect. For example, sigma receptor agonists selective for either sigma 1 (σ1) or σ2 receptors (or both) have both neuroprotective and anti-inflammatory effects in rodent models of stroke. 60 More recently, σ1-selective agonists have been shown to decrease neuroinflammation, 61 while σ2-selective agonists are neuroprotective after TBI. 62

| Combination drug treatment regimens for TBI
In an effort to increase treatment efficacy, multiple drug combinations have been administered together to target multiple neurodegenerative pathways. Based on reported success in the treatment of HIV/AIDS with HAART, 63 recently a combination drug therapy was designed for the treatment of TBI 64 combining vitamin D3, progesterone, omega 3 fatty acids, and glutamine administration for the first 72 hours for TBI patients with a poor prognosis; all patients improved beyond original expectations. However, these case studies included only three patients, there were no controls, and a larger study has not been performed to validate these observations. Another example of a combination of drug therapy is the progesterone and 1,25-dihydroxyvitamin D3 combination, which was effective in reducing neuroinflammation as compared to treatment with the drugs separately. 65,66

| CELL THER APY IN TB I
A number of different cell types have been examined as potential therapeutics for TBI. The first studies in the field focused on replacing neurons in order to rebuild the neural circuitry. The earliest cells examined were already postmitotic neurons, 67-69 and there was variable therapeutic success. Gradually, these studies were replaced by studies using stem cells. In a mouse model, NSC survived, differentiated, migrated to the lesion site, and improved motor and cognitive function after TBI. 70 Also, embryonic stem cells were shown to improve functional outcome after TBI in rodents, but tumors also formed. 71 Later studies involved predifferentiating stem cells into more lineage-restricted precursors in order to reduce the likelihood of tumorigenesis. 72,73 In an effort to reduce reliance on embryonic or fetal tissue, there have been a number of studies more recently focused on induced pluripotent stem cells derived from adult somatic cells. [74][75][76] In addition to these embryonic and NSC for which there are direct developmental pathways to produce neurons, astrocytes, and oligodendrocytes, another source of stem cells that has received a great deal of attention is MSCs. Originally, MSCs were isolated from bone marrow where they support hematopoiesis. However, it has become clear that MSCs reside in many tissues in the body, 77 which may explain why they appear to be efficacious for treating so many different injuries and diseases. Because of their pleiotropic characteristics, these cells have significant therapeutic potential for various diseases including TBI. Following administration, MSCs have shown to penetrate the BBB, migrate to the site of injury, and secrete several growth factors including brain-derived neurotrophic factor (BDNF), glial-derived neurotrophic factor (GDNF), vascular endothelial growth factor (VEGF), nerve growth factor (NGF), and regenerate BBB and neuronal and glial tissues. [78][79][80][81] MSCs also modulated inflammation by inhibiting interleukin six (IL-6) and IL1-β and enhancing IL-10. 82 The anti-inflammatory effect of MSC was reported in a study involving combined administration of MSC and NSC that led to increased recovery from stroke-induced cerebral damage in rats as compared to MSC or NSC alone. 83 Early studies demonstrated that intravenous, 84 intraarterial, 85 and intracerebral 86 administration of these cells improved motor and neurological outcome after TBI. Other methods of MSC delivery that have been used since include intranasal, [87][88][89] intrathecal, 90 and intracisternal. 91 The postulated mechanism of repair in these early studies was transdifferentiation of the cells into neural cells. A number of studies were able to demonstrate that some of the transplanted MSCs expressed neuronal and astrocytic markers in vivo, but few cells survived. 84,85,[92][93][94] The results of in vitro studies were more positive 95 ; however, the transformation from MSC to mature neuron occurred within hours, 96 was reversible when the initiating stimuli were removed, and occurred in the presence of protein synthesis inhibition. 97 It has been suggested that the "transdifferentiation" of MSCs into neurons requires MSCs to be in a toxic, stressful environment. 96,97 However, the ultimate proof that the cells transdifferentiate into neurons is still lacking; these "neurons" have never been shown to produce an action potential. 98 In the ensuing years, several other putative mechanisms of repair for these cells have been studied. Thus, it has been suggested that MSCs may induce brain repair through the production of trophic factors or stimulating release of trophic factors from endogenous cells. [99][100][101] These paracrine mechanisms of repair have recently been reviewed in some detail. [102][103][104] More recently, MSC-derived exosomes have been examined as the paracrine source of neuroprotection and anti-inflammation. These exosomes improved cognitive function and reduced inflammation as determined by decreased IL-1 and increased IL-10 in the brain after TBI. 105 Later studies are consistent with these results. 106,107 MSC-derived exosomes have also been tested in porcine model of TBI coupled with hemorrhagic shock; consistent with the rodent data, the pigs had fewer cognitive deficits as determined with a neurological severity score and recovered faster than nontreated pigs. 108 There is also evidence for MSC inducing immune suppression 109,110 and anti-inflammation 106

| Clinical translation of cell therapy for TBI
As described in the previous section, many different approaches have shown promise for treating TBI-induced pathologies and stimulating tissue regeneration in animal models. However, none of these have thus far translated into therapeutic benefit in human patients and the early clinical trials have not used pure MSC. Table 1 shows the current comprehensive status of clinical trials of stem cell therapy for TBI. The earliest clinical studies reported in the literature used bone marrow-derived mononuclear cells (BMMNCs), which is a heterogeneous mixture of immune cells and stem cells including MSCs. The first major cell therapy trials for TBI using BMMNCs were conducted on children by Cox and colleagues in 2011. 122 In the preclinical studies and Phase 1 clinical trial, treatment reduced BBB permeability and neuroinflammation after TBI. In a second pediatric study from the same research group, the TBI-induced increase in intracranial pressure was reduced in the cell treatment group. 123 There are also clinical trials of BMMNCs in the adults with severe TBI. In this population, treatment with BMMNCs resulted in structural preservation of the corpus callosum and corticospinal tract and these changes were correlated to neurocognitive outcomes; in addition, there was a reduction in the pro-inflammatory cytokine response to injury (NCT01575470). 124

| Potential of MSC and anti-inflammatory combination treatments for TBI
Because of inflammation and other ongoing neurodegenerative cascades, the brain environment post-TBI is a hostile environment for One approach has been to delay transplantation so that the pathophysiology can stabilize; for example, in a study comparing TA B L E 1 Current status of cell therapy clinical trials of traumatic brain injury (TBI)

NCT04063215
A clinical trial to determine the safety and efficacy of hope biosciences autologous mesenchymal stem cell therapy for the treatment of traumatic brain injury and hypoxic-ischemic encephalopathy MSCs with MSCs in a scaffold, 130 MSCs were administered two months post-TBI so that regenerative and repair physiologic processes would dominant in the parenchyma around the TBI lesion.
Another approach has been to administer a drug treatment along with the cells or prior to cell transplantation. For example, Mahmood and associates administered MSCs intravenously 24 hours after TBI. 131 We have taken a similar approach, but instead of modifying sympathetic activity in conjunction with targeting inflammation, we combined two inflammation-modulating treatments. Based on earlier studies in which we identified chemokine (C-C motif) ligand 20 (CCL20) as being significantly elevated after TBI, 135,136 we combined treatment with pioglitazone, a peroxisome proliferator-activated receptor gamma agonist that inhibits CCL20, with treatment with MSCs. 88 The pioglitazone was administered once a day for 5 days after TBI. On day 5, MSCs were administered intranasally. The combination of pioglitazone and MSCs was significantly better than either treatment alone on measures of anxiety, inflammation in the brain, and endogenous NSC proliferation. A similar approach was taken by Peruzzaro and associates when they engineered MSC's to overexpress IL-10 137 before transplantation in a TBI model.
Growth or trophic factor delivery in conjunction with cell administration is another approach that has been studied. These growth/ trophic factors may favorably condition the environment in the TBI brain, or they may protect the transplanted cells from cell death.
GDNF, epidermal growth factor, and VEGF have been shown to protect the brain from neuronal injury and increase regeneration of different cell types. [138][139][140] However, growth factor delivery can be problematic. Growth factors have very short half-lives, necessitating continual local delivery. 141 Further, systemic delivery is often associated with side effects. 142 As a result, it is imperative to deliver the growth factors into the brain near the injury. Liu et al 143  This approach has shown to improve cell injury and motor activity of injured rats and improved metabolic and nutritional outcomes following TBI. 145,146 However, even growth factor delivery directly into the central nervous system (CNS) can lead to adverse effects. In a recent review of NGF trials for CNS diseases, the authors point out significant adverse effects such as pain and weight loss led to the discontinuation of the studies. 142,147 The most common procedure for combining a growth factor treatment with the delivery of MSCs has been to transfect the cells with specific growth factors. For example, overexpressing FGF21 in MSC resulted in improved performance on the Morris water maze, increased hippocampal neurogenesis and dendritic outgrowth. 148 Other investigators have focused on the neurotrophins. Wu and associates overexpressed NT3 in MSCs and observed decreased glial activation, a smaller brain lesion, and decreased edema in the brain, 149 while a number of investigators have focused on increasing BDNF expression in the cells, essentially improving functional outcome. 150,151 More recently, transplanting MSCs with a bioactive scaffold or biomaterial that secretes growth factors has been used to enhance survival, migration, and differentiation of transplanted cells. 152 In one version of a scaffold, Bhang and associates suspended MSCs in a fibrin gel laced with FGF2. 153 The MSCs in the fibrin glue with FGF2 decreased lesion size and apoptotic cell death more than MSCs in glue alone. When a functional peptide derivative of BDNF was incorporated in a self-assembling hydrogel scaffold prior to seeding with MSCs and activated astrocytes, the resulting structure was able to reduce TBI lesion size. 154  prophylactically (such as baby aspirin or statins) or to treat common chronic diseases (beta blockers). For example, aspirin altered the immune and inflammatory profile of both endogenous monocytes harvested from stroke patients or healthy controls and MSCs in culture. 155 When rats were treated with aspirin and cord blood-derived MSCs singly or in combination, outcome tended to be somewhat worse with the combination treatment. 156 These considerations are especially important as the population ages, since the elderly are at risk for TBI and polypharmacy is an issue in this population.

| FUTURE PER S PEC TIVE S
In another example, when type I diabetic rats underwent a middle cerebral artery occlusion followed by MSC transplantation 24 hours later, the cells induced hemorrhagic transformation of the stroke and increased BBB leakage. 157 The addition of niacin to the MSC therapy prevented these adverse effects. 158 It is, therefore, essential that MSCs alone or in combination with other drugs must be tested for efficacy against a background of commonly prescribed drugs or medical conditions.
Another consideration with dosing is whether or not a single injection of cells is enough to maintain long-term improvements in the functional outcome. As mentioned previously, TBI is accompanied by chronic inflammation. 38,[40][41][42] As MSCs (1 × 10 6 cells/kg) 28 days apart and followed for 1 year. 160 These patients had no severe adverse events and function as measured with the ALS Functional Rating Scale-Revised stabilized.
Another issue that needs to be addressed in this field is the reliance on rodents in the early testing of potential new therapeutics.
While using mice and rats is cost-effective and allows manipulation of the genome, there are significant differences in the structure of the rodent brain and the human brain, not only in size, but also in, cerebrovascular volume, oxygen and glucose requirements, lissencephalic vs gyrencephalic architecture, and the amount of white matter present. 161 It is imperative that larger animal models are used for the testing of new therapies including cell therapies so that we can distinguish good candidate therapies that are likely to succeed in clinical trials from those that will not. There has been some work in the development and characterization of porcine and ovine models of TBI, but there are currently no studies of MSC therapies and only a handful of MSC-derive exosome studies 108 in these larger animal models TBI models. 161,162

| SUMMARY
The preclinical data on MSCs both alone and in conjunction with other treatment strategies are promising. These cells have entered at least Phase I (safety) clinical trials for multiple nervous system diseases and injuries, most notably cerebral ischemia, 163-166 multiple sclerosis, [167][168][169][170][171][172][173][174] Alzheimer's disease, 175 and TBI. 133,150,176 While we have a better understanding of the pathologic cascades triggered after TBI and the mechanisms by which MSCs repair the brain and improve functional outcomes, we are still years away from realizing an effective regenerative medicine therapy for TBI that is widely available to patients.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.