Recent advances in the treatment of traumatic brain injury with autologous and non-autologous multipotent stem and progenitor cells: preclinical models and clinical trials

Introduction

Traumatic brain injury (TBI) is one of the main causes of death and disability in the world [19,91]. The annual incidence rate of TBI in various countries are 55-64 cases per 100,000 people in China [96], 823 cases per 100,000 people in the United States, 811 cases per 100,000 people in New Zealand [62], 107 cases per 100,000 people in Australia [43], 262 cases per 100,000 people in Europe [74]. Moreover, the TBI incidence rates of paediatrics varies greatly from country to country. For example, TBI affects more than 486 children per 100,000 people in Australia and more than 280 children per 100,000 people in the United Kingdom ever year [36,67]. The World Health Organization (WHO) reports that TBI will be the main source of morbidity and mortality with being one of the top health issues burdening financial platforms in the coming years [41,62]. TBI is mainly caused by traffic accidents, falls, gunshots, blasts, warfare and sports-related events [85]. Patients suffering from TBI can manifest motor, sensory and perceptive disturbances along with severe depression, anxiety, and personality changes. In the long term, TBI may also give rise to neurodegenerative diseases, such as Parkinson’s disease and Alzheimer’s disease [81,93]. Damages of neuronal tissues associated with TBI fall into two categories: (i) Primary injury, which is directly caused by mechanical forces during the initial insult. The immediate impact of different mechanical insults to the brain can cause two types of primary injuries: focal and diffuse brain injuries. Studies have demonstrated that the co-existence of both types of injuries is common in patients who suffered from moderate to severe TBI however, diffuse axonal injury (DAI) accounts for approximately 70% of TBI cases [86]. (ii) Secondary injury, which refers to further tissue and cellular damages following the primary insult. The biochemical, cellular and physiological events that occur during the primary injury often progress into delayed and prolonged secondary damages which can last from hours to years. Mechanistically, a number of factors contribute to secondary injuries, which include excitotoxicity, mitochondrial dysfunction, oxidative stress, lipid peroxidation, neuroinflammation, axon degeneration and apoptotic cell death [76].
Based on severity, TBI is classified as mild, moderate and severe with mild being the most common one [33]. The Glasgow Coma Scale (GCS) (Table I) which is a clinical method for evaluating the consciousness level of patients after brain injury was used. The GCS for mild brain injury is 13-15, moderate brain injury is 9-12, and severe brain injury is 3-8 [33,34,94]. TBI initially causes focal or diffuse brain damage due to cutting, tearing and stretching [83,85]. If the violence is severe enough, it can cause damage of the blood-brain barrier (BBB), destruction of cerebral vessels, neurodegeneration, axonal injury, brain oedema and neuroinflammation [41,50,103]. Secondary events caused by TBI may occur hours to days to months post-trauma which leads to cellular, metabolic and molecular changes in the injured brain that further damages neurons [50,93]. Primary focal brain injury can be detected by computed tomography (CT) and magnetic resonance imaging (MRI), while diffuse brain injury requires diffusion tensor imaging (DTI) [46,56,68]. There is no effective therapy for the treatment of TBI. Currently, the clinically available treatment options include hyperbaric oxygen, brain stimulation and rehabilitation [15,19,100]. Due to the lack of promising curative options for TBI-induced primary injury, current research focuses on preventing secondary injury [100]. Moreover, the recent use of various stem cell types such as mesenchymal stem cells (MSCs), neural stem cells (NSCs) and their promising therapeutic effects have brought great hopes for curing both primary and secondary injures of TBI [16,48]. The two key factors affecting the effect of stem cell therapy are the optimal transplantation time and the transplantation route. Various time intervals, such as 1, 2, 7 and 14 days post-injury, as well as routes of administration such as intracranial, intrathecal, intravenous, intra-arterial have been tried, but the ideal time and route is yet to be agreed upon [20]. This article reviews the recent research progress in clinical trials and preclinical models of autologous and non-autologous pluripotent stem and progenitor cells for the treatment of brain trauma.

Olfactory ensheathing cells

Olfactory ensheathing cells (OECs) isolated in 1991 are unique glial cells that protect and support the regeneration of olfactory receptor neurons (ORNs) all the way from peripheral olfactory epithelium to the glomeruli of the olfactory bulb in the central nervous system [4,77]. OECs originally come from neural crest and can be extracted either from the olfactory bulb (Fig. 1) called olfactory bulb-olfactory ensheathing cells (OB-OECs) or from olfactory mucosa named OECs [4,31,77]. The most specific markers for the identification of OECs are P75, S100b, and glial fibrillary acidic protein (GFAP) [3]. The expression of p75 which is a marker for OECs progenitor cells is inhibited by axonal contact. For example, the OECs will not express p75 if they are placed in contact with axons during culture [98,101]. To explore the neuroprotective effects of OECs in TBI, Wang et al. extracted cells from newborn Sprague-Dawley (SD) rats aged 2 to 3 days, identified by p75, labelled with Hoechst 33342, and immediately transplanted into the surroundings of the lesion of a modified controlled cortical impact (CCI) rat model. In order to prevent the occurrence of transplant rejection, the rats were injected with cyclosporine for three days before the injury until the tissue was harvested. On the 14th day post-implantation, neurological function was significantly improved. Histological analysis showed that OECs survived and migrated, growth-associated protein 43 (GAP43) and synaptophysin were enhanced, but the number of apoptotic cells was significantly reduced. In addition, at the molecular level, the Bcl2-associated agonist of cell death (BAD) gene is down-regulated, which is considered to be a potential mechanism for changes in OECs in transplanted brains [98]. Fu et al. tried either OECs or bone marrow derived-mesenchymal stem cells (BM-MSCs) alone or OECs and BM-MSCs co-transplantation to treat TBI. OECs were derived from SD rats, identified with p75 antibody, and labelled with Hoechst 33342. BM-MSCs were extracted from transgenic green fluorescent protein (GFP) mice and identified with CD44 antibody. Cells were transplanted singly or jointly around the lesion immediately after injury. To control the rejection of transplantation, the immunosuppressant cyclosporine A was given every day from 3 days before transplantation to the end of the experiment. Fourteen days after implantation, the neural function was improved in all three groups, being more obvious in the co-transplanted group. Histological studies showed increased number of surviving neurons and reduced reactive astrocytosis in the host brain. Moreover, molecular studies have shown that the expression of ciliary neurotrophic factor (CNF) is low, and the Janus kinase/signal transducer and activator of the transcription (JAK/STAT) signalling pathway is activated [25].
Liu et al. were interested to know the therapeutic effects of combined NSCs and OECs in TBI. They isolated NSCs from mouse embryonic hippocampus and OECs from the olfactory bulb of neonatal rats. NSCs and OECs were labelled with Hoechst 33342 and chloromethyl-benzamidodialkyl carbocyanine (CM-Dil), respectively. Cells were engrafted into the penumbra of the rat TBI model alone or in combination. After 14 days of follow-up, the neurological function was greatly improved in the co-grafted group, with sparing more cortical neurons and minimum apoptotic cells. Furthermore, NSCs and OECs survived, migrated, and differentiated under inhibitory levels of interleukin 6 (IL-6) and BAD [57]. Transplantation of OECs for the treatment of TBI has attracted the attention of researchers in the field. Studies have shown that co-transplantation of MSCs and OECs or NSCs and OECs is more effective than transplantation of OECs alone in the treatment of rat TBI models. This might be due to the special biological features of OECs and their accompanying MSCs which have an apparent effect on the regeneration of axons. OECs, as a unique type of neuro-glial cells, are not only available in the epithelium of the nose and olfactory nerve but also in the olfactory bulb of the central nervous system. Although these studies have non-autologous preclinical properties, OECs are still a promising choice for autologous transplantation in clinical practice.

Neural stem cells

Neural stem cells are self-replicating, multipotent progenitors [49], and have the potential to differentiate into neurons, astrocytes and oligodendrocytes [70,108]. In the mouse model, it was found that the progenitor cells of NSCs were formed between 13.5 days and 15.5 days of the embryonic stage and remained inactive until birth [26]. There are two potential niches that host NSCs in the adult mammalian brain, the subventricular zone (SVZ) of the lateral ventricle and the subgranular zone (SGZ) of the hippocampus [66]. The neurogenesis of SVZ and SGZ is essential for maintaining the olfactory bulb and spatial memory [9,42,110]. In order to study the long-term therapeutic effects of the human foetal neural progenitor cell (hfNPC) on TBI, Skardelly et al. injected human fetal neural progenitor cells (hfNPCs) directly into the lesion or intravenously to a CCI rat model 24 hours after the injury. Twelve weeks post-transplantation, the neurosensory and mental functions were improved in both groups and the volume of the lesion measured by MRI was also decreased with a prominent reduction in systemically transplanted rats. In histological investigations, reactive astrogliosis was decreased, angiogenesis and neuronal survival was increased, but engrafted cells differentiation was observed in very rare cases, indicating that the improvements are mainly due to inhibition of reactive gliosis and induced angiogenesis in the host brain [87]. Similarly Haus et al. investigated the long-term effects of human neural stem cells (hNSCs) on cognition improvement in the TBI model. They deployed hNSCs into the lesion of the CCI model of immunodeficient athymic nude (ATN) rats nine days after the injury. Five months after the operation, spatial memory improved, but non-spatial memory and emotional deficits did not boost significantly. Moreover, hNSCs differentiated into mature neurons and glial cells, but unlike Skardelly’s findings, their observations showed that the size of the lesion did not decrease [35]. Furthermore, Ma et al. revealed the potential therapeutic effect of rat embryonic NSCs in TBI and its possible underlying mechanism. They extracted NSCs from the forebrain of a 14-day-old rat embryo and injected them directly into the injured brain of a CCI rat model 72 hours after the trauma. Eight weeks after transplantation, neuronal function was improved, NSCs survived, and their differentiation into neurons was observed. In addition, histological analysis demonstrated overexpression of synaptophysin (SYP) and regenerationassociated protein (GAP43) which is considered to be the underlying mechanism for NSCs mediated mental function improvement in the native brain [61].
Due to the variations of the clinical situation in TBI, different animal models have been developed. Even though, bigger animals are more similar in physiology to the human being, but the standardized methods of measuring outcomes and their reasonable price made rats the most used models in TBI [22,95]. The most commonly used models are CCI, weight drop injury (WDI), fluid percussion injury (FPI), electronic device brain injury (EDBI) and blast injury (BI) [11,64]. In general, the rat is cleaned and anesthetized, fixed on injury platform, injured with the required parameters and then followed up for the targeted experimental time [1].
Sun et al. tried multipotent neural stem/progenitor cells (NS/NPCs) derived from the SVZ of adult Fisher 344 (F344) rats for the treatment of TBI. They injected NS/NPCs around the lesion of the FPI rat model two days post-injury. After two to four weeks, motor and cognitive function was improved, and NS/NPCs survived, migrated from the site of deposition and differentiated into location-specific astrocytes and oligodendrocytes [90]. As gap junctions provide direct communication between cells by means of tiny particles and ions transportation therefore Yu et al. were interested to investigate their connecting role between the transplanted NSCs and host brain cells of the TBI model. They implanted NSCs around the lesions of the CCI model 72 hours after the Wister rats were injured. Gap junction-associated connexin 43 (Cx43) was significantly overexpressed at the transplant site and the injury edge 1-4 weeks after transplantation, indicating that the transplanted cells may be involved in the circuit of native brain neurons [106]. In addition, Lin et al. demonstrated that human neural progenitor cells (hNPCs) can survive, differentiate and integrate into the neural circuit of the native brain in the rat TBI model [55]. In order to investigate the effects of hNSCs on the modulation of host immune response and their contribution to neuroprotection after TBI, Gao et al. primed hNSCs and injected them intracerebrally to the C57BL/6 mice CCI model 24 hours after injury. Six days after engraftment, there was no remarkable decrease in the volume of the lesion but there was a significant reduction in injury-dependent accumulation of the amyloid precursor protein (APP). Moreover, there was a decrease in activation of microglia and production of pro-inflammatory interferon g (IFN-g) while production of anti-inflammatory interleukin 4 (IL-4) was increased. Thus, primed hNSCs neuroprotection in the TBI model is at least partially due to modulation of the host immune response [27].
To reveal the effects of transplanted NSCs on the level of expression of B-cell lymphoma-extra-­large (Bcl-xL) in the TBI model, Pang et al. introduced mouse embryonic NSCs into the traumatically injured brain of rats within 24 hours post-injury. Seven days after transplantation, motor function was improved, cellular apoptosis was reduced and Bcl-xL gene was overexpressed. Furthermore, herpes simplex virus (HSV) 1 carrying Bcl-xL recombinant alone was administered to the surroundings of the lesion, which resulted in the improvement of neuronal deficits as well as inhibition of apoptosis in the host brain. These findings suggest that Bcl-xL increased expression can be considered as an alternative for NCS engraftment in the treatment of TBI [73]. Xiong et al. attempted to justify the idea of brain-derived neurotropic factor (BDNF) mediated neuroplasticity being the underlying mechanism for transplanted NSCs therapy in TBI. They applied normal NSCs or BDNF knockdown NSCs to the traumatized brain regions of mice 24 hours after the injury. Two to four weeks post-operation, normal NSCs engrafted mice showed improvement of neurologic disabilities and upregulation of BDNF with synaptophysin, while rats that received BDNF knockdown NSCs not only demonstrated a lower level of synaptophysin but also exacerbated functions of the nervous system, indicating the critical therapeutic role of BDNF in the TBI model [102]. Likewise, Tao Chen et al. extracted NSCs from the rat embryo and introduced the BDNF gene to the cells. Encoded-NSCs or naïve-NSCs were injected directly into the damaged area of the rat’s brain. One to four weeks after implantation, synaptic proteins such as TrkB and postsynaptic density protein 95 (PSD-95) were increased, the MAPK/Erk1/2 signalling pathway was activated, and the antioxidant protein thioredoxin (Trx) was upregulated [13]. Those results indicate that engrafted NSCs ameliorate TBI-induced neuro­logic deficits. The underlying mechanisms mostly seem to be inhibition of neuroinflammation and apoptosis, with increased expression of neurotrophic factors and synaptic proteins. The idea that transplanted NSCs substitute the damaged neurons and decrease the lesion size remains contradicting and unconvincing. Even though the penetrating traumatic brain injury (PTBI) is less common than closed brain injury but its prognosis is far worse than for closed brain trauma. Therefore, Spurlock et al. tested the potential role of hNSCs in the improvement of PTBI. They implanted 4 × 105 hNSCs into the surroundings of the lesion of the Sprague Dawley (SD) rat TBI model one week after injury. The neurointellectual performance improved after 8 weeks of transplantation, with significant differentiation and migration, but the differentiation was not obvious after 16 weeks of transplantation, and the cell morphology was very similar to that of native brain cells [88]. Similarly, Hu et al. wanted to know the ideal location for the deposition of hNSCs when treating PTBI. They transplanted hNSCs into the core or around the lesion one week after the injury. After 12 weeks, neuronal differentiation and motor function improvement were observed in both groups but sparing of a larger volume of intact brain including cortex makes the peri-lesion administration a preferable choice [40]. Beretta et al. made an attempt to see the effects of hNSCs on posttraumatic hyperexcitable brain. Due to lack of a reliable model for testing posttraumatic hyperexcitability of the brain, they traumatized one cerebral hemisphere of the immunodeficient ATN rat with CCI model and contralateral hemisphere was electrically stimulated to produce spontaneous seizures. Human NSCs were transplanted to either sham or injured brains. Fourteen weeks after implantation, they observed a twofold increase in survival and differentiation of hNSCs in injured brains compared to shams. Moreover, surviving and differentiation was the worst in stimulated hemispheres in either groups, indicating that kindling is harmful for differentiation and survival of hNSCs. It was also observed that mental function was ameliorated by hNSCs while these improvements were vanished by kindling meaning that hNSCs engraftment is not useful for animal or human with posttraumatic seizures [6]. Those results suggest that the implantation of the cells to the surrounding of the lesion is more effective than intralesional administration. Moreover, traumatized brains with a previous history of seizures may not be cured with NSCs therapy.
Lee et al. revealed the potential curative role of human parthenogenetic neural stem cells (hpNSCs) in the treatment of TBI. They implanted hpNSCs into the brains of a CCI rat model 72 hours post-trauma. Three months after engraftment, neurological, motor and intellectual functions were improved. In histological studies, biomarkers related to myelination increased, while reactive gliosis, neuroinflammation and splenic inflammation decreased [51]. To elucidate the difference between the effects of embryonic neural stem cells (ENSC) or differentiated cells (DC) in the treatment of TBI, ENSCs/DCs were isolated from E14 mouse embryo and injected into the lesion of CCI rat models 7 days post-injury. Following up for one to four weeks, it was noticed that microglia were activated, astrogliosis was inhibited and neurogenesis due to inducing of neuronal progenitor cells was enhanced. In addition, there was no significant difference between the ENSCs group and the dendritic cell group in terms of improvement, indicating that the implantation of ENSCs/dendritic cells can improve TBI by reducing neuroinflammation and inducing neurogenesis in the host brain [71]. Furthermore, Abad et al. derived human neural stem/progenitor cells (hNS/PCs) from the brain of a human undergoing epilepsy surgery. The cells were cultured in Pura Matrix hydrogel (PM) and deposited into the lesion of a TBI rat model 30 minutes post-injury. In the evaluation 1 to 28 days after transplantation, motor function and spatial memory improved, lesion volume reduced, neuroinflammation and reactive gliosis were inhibited, which strongly supports the use of PM-cultured hNS/PCs for the treatment of TBI [44]. After decades of conceptual assumption of the fixed regenerative capacity of the adult brain, the fascinating discovery of adult brain NSCs and their continued ability of neurogenesis throughout the life brings a new era for deeper understanding of the brain. NSCs not only maintain the neural tissue, but also provide expanded plasticity to the key brain structures that are vital for learning and memory. The rapid and vast progress in the field has elaborated on the biological as well as clinical potential of the cells. Although transplanted embryonic or adult NSCs can survive, migrate, differentiate and improve neurologic and cognitive functions in preclinical TBI models, the realistic and ethical sources of NSCs for autologous clinical translation have not yet been achieved. The clinical promise of NSCs still requires thorough understanding of their properties, mechanism of regulation and eventually application through collaborative efforts of both basic and translational research.

Mesenchymal stem cells

Mesenchymal stem cells are multipotent in nature, exist in adults, and have the ability to self-replicate and differentiate into various cell types [69]. As TBI is a very complex disease and there is no effective treatment available in the clinical practice, therefore, stem cell therapy has been the focus of research in the recent years. In spite of great progress in animal models, the clinical translation of stem cells for TBI is limited. However, there are some issues regarding the clinical trials, such as smaller sample size, majorly no controls, short follow-up time and being single-centred, but yet, it remains a great hope for the future of TBI patients. Especially, autologous stem cell therapy, which has no risk of immune reaction and rejection gives an enormous optimism to the future of the field. Mesenchymal stem cells can differentiate into neural cells including neurons, but their differentiation without any stimulation is not very effective. In order to improve the survival and differentiation of MSCs into nervous cells, they are usually reprogrammed before being transplanted [59,63]. Four different strategies are used to reprogram MSCs towards neuronal cells lineage namely, small molecules, epigenetic modifications, psychotropic drugs and enriched media [7]. Moreover, retinoic acid which is a vitamin derivative is also known for its ability to differentiate MScs and progenitor cells into neurogenic cells [28]. In the following sections, we will review them according to their source of origin, such as autologous MSCs and non-autologous MSCs with their subtypes.

Autologous mesenchymal stem cells

Autologous MSCs studies are further sub-sectioned into clinical trials and preclinical models.

Autologous bone marrow-derived MSCs in clinical trials

In order to evaluate the safety and feasibility of autologous BM-MSCs in the treatment of TBI, a number of studies have been conducted (Table II). Zhang et al. conducted a prospective, non-randomized, open label clinical trial on seven patients aged six to fifty-five years old. The Barthel index (BI) (Table III) which is an ordinal scale used to assess the performance in activities of daily living (ADL) was used and all the patients have BI score less than 40. They administered 107-109 BM-MSCs directly into the lesion during the operation, and injected 108-1010 cells intravenously for the second time 4 to 12 days later. After 6 months of follow-up, the neurologic performance of all patients was significantly improved, except for one patient who had 2 seizures in the first 2 months after surgery and received conservative treatment. There was no death or serious adverse reaction caused by transplantation [109]. Cox et al. have a total of four clinical trials, two of which have been completed, the other two are registered at ClinicalTrials.gov, and the final results are still on the way. The completed studies are prospective, non-random, open label, single-centre clinical trials on children and adults. In their first trial, ten children aged five to fourteen years with severe TBI were given intravenous 6 × 106 autologous bone marrow-derived mononuclear cells (BMMNCs) per kg body weight forty 8 hours after TBI. After 6 months of follow-up, the motor and sensory functions of all patients improved, and 3 patients recovered completely. No seizures, refractory intracranial pressure, altered cerebral perfusion or new ischemic events were noticed. MRI did not show any significant morphological changes at 1 and 6 months after surgery [17]. Moreover, in the second trial of Cox, 15 patients aged 18-55 with GCS scores of 5-8, age- and severitymatched controls, were treated with intravenous autologous BMMNCs. The patients were divided into three groups, and engrafted with 6, 9 and 12 million cells within 48 hours post injury. MRI and DTI were used to evaluate critical structures preservation in the brain. Six months post intervention, the patient’s neurological and perceptive functions improved, the white matter anisotropy score (FA) was retained, and inflammatory biomarkers such as interleukin 1b (IL-1b) and IFN-g, tumour necrosis factor a (TNF-a) and interleukin 10 (IL-10) were significantly reduced. There were no severe adverse events, except for mild pulmonary toxicity in high-dose recipients, which was not clinically significant [18]. Since Cox’s first trial did not have a control group, Liao et al. tried a second retrospective, cohort, phase I clinical study. They included the patients treated in Cox research as the treatment group, and collected 19 children of matching age and severity as the control group. They were treated with standard treatment regimens without autologous bone marrow stromal cell transplantation. To evaluate and compare the clinical outcome between the two groups, paediatric intensity level of therapy (PILOT) and paediatric logistic organ dysfunction (PELOD) scores and length of neurointensive care unit (NICU) stay were considered. In patients treated with BMMNCs, PILOT and PELOD scores and NICU hospitalization were significantly reduced [54]. In order to investigate the efficacy and safety of bone marrow MSCs in the treatment of subacute TBI, Tian et al. conducted a prospective, non-randomized and open-label clinical trial. The subjects of the study were 97 patients aged 18 to 50 years with a persistent vegetative state or motor disturbance post-injury. The patients were injected intrathecally with 4 × 106 autologous BM-MSC within two months after trauma and followed up for fourteen days. 27 out of 73 patients showed improvement in motor function, and 11 out of 24 patients showed improvement in the conscious level. There were no serious transplantation-related complications or adverse effects. Younger and earlier treated patients benefited more from the treatment [92]. In order to evaluate the long-term efficacy and safety of autologous BMMNCs in the treatment of chronic TBI, Sharma et al. conducted two prospective, openlabel, non-randomized clinical trials. In the first trial, the number of patients was fourteen, they were aged 12-65 years, and had a history of TBI for about five years. A dose of 1 × 106 BMMNCs per kg body weight were intrathecally injected via lumbar puncture. The follow-up was done for 6 months and the functional independence measure (FIM) scoring system was used to analyse the improvement of symptoms. The electroencephalograph (EEG) taken before transplantation was abnormal for ten patients of which seven patients had a history of seizures. Positron emission tomography (PET) brain scan showed improved metabolism in the brain. The neurological activities were enhanced with the rates of 7 of 13, 6 of 11, 7 of 13, 8 of 11, 9 of 13, 6 of 12 and 3 of 8 in upper limbs, lower limbs, trunk, balance, voluntary control, speech and cognition, respectively. Except for one case of epileptic seizures within 3 months after transplantation, there were no complications or side effects [79].
Furthermore, in Sharma’s second trial, 50 patients aged between 7 and 64 were studied for 22 months after implantation. Fifteen patients received a second dose of cells after 6 months of intervention. The criteria for categorizing symptomatic improvement of less than 30%, 30-60%, and more than 60% into mild, moderate, and significant were adopted. Sixteen out of fifty patients had a history of seizures. 50% of the patients had significant improvement, 16% of the patients had moderate improvement, and 26% of the patients had mild improvement. Except for 2 cases of single seizures, which were cured by anti-epileptic drugs, there were no obvious adverse reactions. The patients who received a second dose of cells did not differ from single time recipients in their clinical outcome. The candidates who were younger and had acute mild injury showed better response [80]. In order to reveal the safety and feasibility of autologous BM-MSCs in non-acute severe TBI, Wang et al. attempted a prospective, non-randomised, open-label clinical trial on ten patients who had brain injury in about past sixty days and were aged 21-55 years. The patients were followed up for six months. Seven patients received 4 × 107 cells intravenously and 3 patients received 2 × 107 cells intrathecally. The motor and sensory deficits of 7 patients improved to varying degrees, and none of the patients deteriorated. There were no reports of deaths or serious adverse events. On the 3rd and 7th day after the intervention, serum nerve growth factor (NGF) and BDNF levels increased significantly [98]. Despite the good clinical results and minor adverse events, the shortcomings of these trials, such as small sample size, short-term follow-up, being single-centre, mostly uncontrolled, and non-randomized, need to be addressed in future studies. Moreover, the ideal time of administration, the optimal dose and the potential therapeutic mechanism of the cells remain to be studied.

Autologous bone marrow-derived MSCs in preclinical models

To study the neuroprotective effects and the underlying mechanism of autologous BMMNCs in TBI, Bedi et al. intravenously injected 2 × 106 cells per kg body weight to the CCI rat model 72 hours post-injury. After 24 hours of implantation, the permeability of the blood-brain barrier was significantly decreased, and the apoptosis of activated microglia was noticeably increased. Four weeks later, neurocognitive function and spatial memory were remarkably improved, which is proposed to be due to the preservation of the blood-brain barrier and the inhibition of active gliosis [5]. Liu et al. were interested to know whether intrathecal injection of autologous BM-MSCs could reach the site of brain injury, and whether intrathecal or intravenous injection is more effective for the treatment of TBI. BM-MSCs were isolated from the rabbit, labelled with GFP and 2 × 106 cells per kg body weight were implanted either intravenously or intrathecally 24 hours post-injury. Four weeks after engraftment, the motor function was improved in all treated rabbits with greater enhancement in the intrathecal group, and the cells were observed in the lesion area of the brain with fluorescence microscopy [58]. Jiang et al. studied the therapeutic effects of autologous BM-MSCs transplantation on the CCI rat model along with their role in ameliorating pathological changes. They injected a million cells into the site of injury during the surgery. The mice were scarified on day 1, 7, 14, 21, 28 and neurologic and pathologic studies were done respectively. Motor and intellectual functions improved fourteen days post engraftment. Histological studies showed that new blood vessels were formed, cerebral oedema was improved and reactive gliosis was inhibited. Moreover, the level of brainderived neurotrophic factor was increased seven days post-implantation [45]. Despite multiple studies, the autologous nature of these studies makes them special among the preclinical models which can give a good insight into the underlying mechanism for autologous stem cell therapy. Furthermore, these data suggest that intrathecal transplantation is more effective than intravenous administration.

Non-autologous mesenchymal stem cells

Non-autologous MSCs research studies are categorized into the following four sub-types (Fig. 2).

Non-autologous bone marrow-derived MSCs

Bone marrow-derived MSCs have been proven to be effective in the treatment of TBI, and researchers have proposed several potential mechanisms (Table IV). Jiang et al. studied the therapeutic effects of human bone marrow stromal cells (hBMSCs) on TBI. They labelled the cells with superparamagnetic iron oxide (SPIO) and intravenously injected 3 million cells to the CCI rat model 5 days after the injury. Six weeks post-engraftment, mental function was improved. MRI demonstrated reduced systemic brain atrophy and increased cerebral blood flow (CBF) [53]. Osanai et al. tried intra-arterial implantation of bone marrow stromal cells (BMSCs) in the cortical freezing injury (CFI) rat model. Cells were extracted from rat’s bone marrow, labelled with fluorescent dye and injected to the ipsilateral internal carotid artery 7 days post-trauma. After four weeks of follow-up, motor function was significantly improved. The implanted cells differentiated into neurons, astrocytes and endothelial cells. Optical imaging demonstrated that these cells were engrafted into the brain in their first pass before the systemic circulation [72]. Furthermore, in order to observe whether the severity of TBI determines the therapeutic efficiency of BMSCs, Bonilla et al. injected 5 million cells into moderate and severe chronic rat TBI models via intravenous route. Two months post-implantation, the neurologic performance of moderately injured animals improved significantly [8].
In addition, Anbari et al. injected 3 million bone marrow-derived mesenchymal stem cells (BMMSCs) intravenously into a rat model 20 hours after trauma. After fourteen days of transplantation, the neuronal function was significantly improved. The transplanted cells differentiated into neurons and astrocytes, and were observed in the traumatized region of the brain [2]. Shen et al. wanted to observe the effects and underlying mechanism of engrafted BMSCs on TBI. They isolated cells from GFP-transgenic mice and immediately injected them into the penumbra of the CCI rat model. At two weeks’ follow-up, the brain activity and the survival rate of the rats were significantly improved. Histologic analysis manifested that apoptosis was significantly reduced and axonal regeneration increased. Molecular studies revealed increased levels of growth-associated protein 43 (GAP-43), synaptophysin and glial cell line-derived neurotrophic factor (GDNF). In addition, the recipient brain showed a down-regulation of BAD, Bcl-2-associated X-protein (BAX) genes [82]. Feng et al. tried to use BMSCs to treat TBI and injected 3 million cells intravenously into mentally traumatized rats within 30 minutes after the trauma. After fourteen days, neuromotor function was significantly improved. Furthermore, the implanted cells migrated to the injured brain area and differentiated into neurons and astrocytes. The expression level of vascular endothelial growth factor (VEGF), BDNF and synaptophysin were up-regulated [24].
Deng et al. used stromal cell-derived factor 1 (SDF-1) to induce the expression of C-X-C chemokine receptor 4 (CXCR4) in BMMSCs. In the rat model of TBI, normal cells or modified cells were injected into the lesion. Four weeks post-implantation, the spatial memory and neurologic function of the mice treated with the stimulated cells were significantly improved. Furthermore, the levels of expression of NGF and BDNF were also remarkably increased in induced cells implanted rats [21]. Li et al. revealed the effects of hBMSCs on the cerebral hemodynamics of TBI. They injected 3 million cells intravenously into a CCI rat model 6 hours and 1 week after injury. Neurointellectual function improved three months post-transplantation. Magnetic resonance imaging demonstrated early preservation of the blood-brain barrier and enhanced cerebral blood flow [52]. Shi et al. over-expressed superoxide dismutase 2 (SOD2) in BMMSCs, and implanted normal or modified BMMSCs into a CCI rat model through the intravenous route. Twenty one days post-intervention, the neuronal and intellectual functions improvement was more remarkable in SOD2-adjusted cells. Furthermore, the preservation of the blood-brain barrier and the reduction of neuroinflammation were more significant in the rat brain treated with modified cells [84]. Hu et al. studied the effects of transplanted BMMSCs on neovascularization in TBI. They injected 10,000 BMMSCs into the lateral ventricle of the WDI rat model 12 hours post-surgery. Twenty one days post-intervention, neuromotor function was significantly improved. Moreover, endothelial progenitor cells, angiogenetic, and neuronal markers were increased in peripheral blood [39].
Since calpain activation plays a key role in TBI-induced neuroinflammation, its inhibitory effect has been proven to be neuroprotective. Therefore, Hu et al. searched the effects of combined therapy of BMMSCs and calpain inhibitor MDL28170 in a WDI rat model. MDL28170 was injected to the lesion thirty minutes post-injury followed by 1 × 105 GFP-labelled BMMSCs engraftment 24 hours after trauma. Histopathological investigations showed improved microenvironment for survival of transplanted BMMSCs via inhibition of neuroinflammation. After a 4-week follow-up, there was a greater improvement of neurologic function in cells and MDL28170 co-grafted group compared to mice treated with cells alone [38]. Yuan et al. investigated the effects of hypoxic preconditioning (HP) on BMSCs survival and their therapeutic role in TBI. They injected hypoxic preconditioned- bone marrow stromal cells (HP-BMSCs) or normal BMSCs into the CCI rat model 24 hours post-injury. Sensorimotor and cognitive function was significantly improved in HP-BMSCs treated mice 35 days after intervention. Moreover, hypoxia-inducible factor-1a (HIF-1a), phosphorylated mechanistic target of rapamycin (p-mTOR) and VEGF were overexpressed in the HP-BMSCs-treated brain [107]. Guo et al. investigated the effects of HIF-1a/SDF-1/CXCR4 axis on the neuroprotective role of transplanted BMSCs in TBI. BMSCs were treated with HIF-1a and SDF-1 prior to transplantation, and then injected into the lateral ventricle of the EDBI rat model. 14 days post-implantation, the mice were sacrificed and their brains were analysed. Histologic and molecular investigations revealed that the HIF-1a/SDF-1/CXCR4 axis promoted the migration of transplanted cells, thereby reducing cell apoptosis and enhancing the neuroprotective effect of the host brain [32]. Although bone marrow stem cells have been tried in clinical trials, but to know their underling therapeutic mechanisms such preclinical studies are still vital.

Non-autologous umbilical cord/umbilical cord blood-derived MSCs

Dong et al. studied the distribution of human umbilical cord mesenchymal stem cells (hUCMSCs) in a rat model of TBI. They isolated hUCMSCs from fresh human umbilical cords, labelled with GFP and injected into the lateral ventricles of WDI rats. The animals were followed up for 14 days, and the cells were tracked using immunohistochemistry (IHC) and small animal imaging system. On the 10th day after transplantation, the cells were distributed in large blood vessels of the TBI model. Furthermore, VEGF was increased and neovascularization was noticed in the brains of engrafted mice [23]. Gincberg et al. conducted two studies on human umbilical cord blood-derived CD45+ pan-hematopoietic cells (HUCB-CD45+) and the TBI injury. In their first study, they investigated the therapeutic effects of HUCB-CD45+ cells in a rat TBI model. HUCB-CD45+ cells were isolated from vaginal or caesarean delivery HUCB and injected intravenously to the WDI rat model 24 hours post-injury. After 6 weeks of transplantation, the motor and rational functions of the two groups of mice improved [29]. In their second study, they wanted to know the relationship between NGF and its receptors with engrafted HUCB-CD45+ cells in TBI. The interrelation of HUCB-CD45+ cells, NGF and its receptors and TBI was investigated both in vivo and in vitro. In vitro the cells were stimulated with TBI brain extracts while in vivo the cells were studied in the brains of the TBI rat model. In either case, the HUCB-CD45+ cells were found to be expressing NGF and its receptors TrkA, p75NTR and a9b1 in response to TBI [30].
Qi et al. investigated therapeutic effects and the underlying mechanism of umbilical cord mesenchymal stem cells (UCMSCs) on TBI. They isolated cells from the umbilical cord of newborn rats, and transplanted three million cells into the lesion of the WDI rat model immediately post-injury. After 4 weeks of transplantation, motor and sensory functions were improved. In addition, neurotrophic factors such as GDNF and BDNF were up-regulated, while inflammatory factors such as IL-1b and TNF-a were down-regulated [75]. Xu et al. studied the importance of silencing histone deacetylase 1 (HDAC1) in hUCMSCs before their transplantation into the TBI model. The WDI rat model intravenously received three doses of 1 million normal hUCMSCs or HDAC1-silenced hUCMSCs (MSCs-siHDAC1) for three consecutive days 24 hours after injury. The mice were followed up for twenty eight days and the evaluation was done on day 1, 3, 7, 14, 21 and 28. Compared with the hUCMSCs treatment group, the intellectual and motor functions of the mice treated with MSCs-siHDAC1 were significantly improved. Furthermore, the survival rate, proliferation rate and migration rate of transplanted cells in the MSCs-siHDAC1 treatment group were significantly higher than those in the control group. The PI3K/AKT pathway was also activated in the MSCs-siHDAC1 treated group while inhibition of the pathway by LY294002 attenuated the neuroprotective effects of MSCs-siHDAC1 in the host brain [104].
Wang et al. explored the therapeutic effect of hUCMSCs in acute TBI. They isolated cells from fresh umbilical cord of human, labelled them with GFP and injected them into the lateral ventricle of the WDI rat model. Histological studies on day 10, 14, and 20 after transplantation showed increased expression levels of recipient brain microvessel density (MVD), BDNF and GFAP [97]. Srivastava et al. evaluated the role of human umbilical cord blood (HUCB) in TBI. They injected 25 million cells per kg body weight to the CCI rat model via the intravenous route either 24 or 72 hours post-injury. Far red fluorescent dye imaging showed that the integrity of the blood-brain barrier was improved 96 hours after injury. Furthermore, cytokine inhibition assay demonstrated reduced levels of TNF-a and IFN-g [89]. Hu et al. investigated if TBI can cause cardiac dysfunction and if human umbilical cord blood cells (HUCBCs) therapy can improve TBI-related neurologic and cardiac dysfunctions. Echocardiography showed decreased left ventricular ejection fraction (LVEF) and fractional shortening (LVFS) in the CCI rat model. The mice were intravenously injected with one million HUCBCs three days post-injury. After 30 days’ follow-up, neurocognitive function and cardiac dysfunction caused by brain trauma were significantly improved. Further studies revealed that myocardial cell apoptosis and myocardial fibrosis were reduced, and transforming growth factor b (TGF-b) and NADPH oxidase-2 (NOX2) were reduced [37].
Chen et al. studied the therapeutic effect of hUCMSCs in TBI and its underlying mechanism. One million cells were intravenously injected to the WDI rat model 3 hours post-injury. The neuronal deficit in the cell therapy group was significantly improved 28 days after transplantation. MRI showed that the volume of ischemic brain tissue in the treatment group was smaller than that in the untreated group. Furthermore, the expression levels of oxidative stress protein NOX2, apoptotic proteins BAX, Caspase-3 and brain oedema protein aquaporin-4 were downregulated while the expression levels of angiogenesis biomarkers VEGF, SDF-1a, CXCR4 and antifibrotic proteins phospho-Smad1/5 and bone morphogenetic protein-2 were upregulated in the treatment group [12]. Caplan et al. explored the therapeutic effects of human umbilical cord blood-derived T-cells (hUCB-TCs) in TBI. Ten million cells per kg body weight were injected through the tail vein of the CCI rat model either 6 or 24 hours after the injury. The blood-brain barrier integrity was not preserved 96 hours post intervention. Alteration of chronic microglial activation and modulation of immune response were observed thirty days post-engraftment in the brains of treated rats [10]. These data suggest that umbilical cord-derived MSCs are effective in the treatment of brain injury in animal models but their clinical translation for autologous transplantation does not seem to be realistic.

Non-autologous amnion-derived MSCs

Chen et al. studied the effects of human amnionderived multipotent progenitor cells (hAMPCs) on TBI-induced axonal degeneration in the penetrating brain injury (PBI) rat model. They extracted cells from human chorion, labelled with PKH26 and injected them into lateral ventricle at a dose of 2 × 106 immediately after the injury. Two weeks post-transplantation, the axonal degeneration of the corpus callosum and ipsilateral thalamus was significantly attenuated. In addition, cells were observed in the SVZ of both hemispheres, but cells were observed along the corpus callosum of the injured side, indicating that the cells first settled in the SVZ and then migrate towards the lesion [14]. Yan et al. explored the effects of human amnion-derived mesenchymal stem cells (AMSCs) and AMSC-derived neural stem-like cells (AMNSCs) in TBI. AMSCs were extracted from amniotic membrane of the human, co-cultured with differentiation medium in vitro, and segregated into AMNSCs. The CCI rat model was injected in penumbra either with AMSCs or AMNSCs four days post-injury. After 28 days’ follow-up, the neuronal performance and spatial memory of the mice treated with AMNSCs was significantly improved. Furthermore, the neurotrophic factors BDNF, GDNF, ciliary neurotrophic factor (CNTF), NGF, and NT-3 were all up-regulated, but the survival and differentiation of the grafted cells were poor [105]. Although amniotic mesenchymal stem cell therapy can improve animal brain trauma models, due to the lack of clinical trials, poor survival and differentiating quality of grafted cells and their unclear therapeutic mechanism are issues which need to be resolved in future research.

Non-autologous adipose tissue-derived MSCs

Martinez et al. investigated the therapeutic effects of adipose tissue-derived mesenchymal stem cells (ADMSCs) in TBI. They isolated the cells from mouse peritoneal fat, labelled with GFP, and injected 200 cells around the wound in a CCI rat model 24 hours after injury. The rats were followed up for twenty one days with regular evaluation at day 2, 7, 14 and 21. TBI caused motor dysfunction was improved while sensory deficit was not ameliorated. Moreover, enhanced neurogenesis was observed in the dentate gyrus of hippocampus of the host brain [65]. Kappy et al. searched the effects of human ADMSCs in the TBI rat model. ADMSCs were extracted from human fats, and injected into the tail vein of rat with a dose of 1 × 106 cells three hours post injury. Three days post-engraftment, TBI-induced neurologic impairment was not improved while the expression levels of TNF-a and b-amyloid precursor protein (b-APP) were decreased [47]. Ma et al. investigated whether HP can enhance therapeutic effects of ADMSCs in TBI. They isolated ADMSCs from transgenic-GFP mice, cultured under hypoxic conditions of 2.5% O2 for 18 hours. The CCI rat model was treated with ADMSCs or HP-ADMSCs, respectively, by applying 2 million cells to the damaged cerebral cortex 1 hour after the injury. Fourteen days post-engraftment, the mice treated with HP-ADMSCs experienced the greatest mental function recovery. The survival and differentiation of HP-ADMSCs in the host brain were more obvious than those of other groups. Real-time polymerase chain reaction (RT-PCR) showed that anti-inflammatory IL-10 and anti-apoptotic genes were up-regulated in treated mice [60]. Ruppert et al. tried to use human ADMSCs to treat acute and subacute brain injury in a CCI rat model. Three million cells per kg body weight were injected intravenously to the animals either three or fourteen days post-injury. Thirty-two days after intervention, the neurological impairments caused by TBI in both groups were significantly improved [78].
MSCs in the treatment of TBI has achieved enormous attention over the last decade. It is because MSCs are convenient to harvest, mainly autologous so they do not cause immune response, and they can differentiate into neuronal cells, thereby help to repair brain tissue after trauma exposure. Their healing phenomena of TBI is mainly due to affecting the cells of the host tissue. They basically decrease the inflammation in the host tissue, stimulate the production of neural growth factors, enhance axonal growth and improve neuronal plasticity. MSCs have the tendency to reside near the injury site and also have the ability to migrate across the injured tissue. However, these advantages are yet to be expanded to their full effect. More research is still required to better understand the mode of action of MSCS and their trophic effects on traumatized host tissue.

Conclusion and future directions

At present, there is no effective treatment for TBI in clinical practice. Currently available therapeutic options are hyperbaric oxygen, brain stimulation, and rehabilitation. Recent developments in the field of stem cell therapy have left a promising future for eliminating this fatal health problem. Although preclinical models have made great progress, clinical translation is still limited. There are few completed clinical trials only for BMMSCs with the drawbacks of smaller sample size, most of them without controls, short follow-up time and being single-centred. In addition, the issues concerning the exact mechanism of therapeutic effects, the limited available sources of autologous stem cells, and the determination of the optimal dose, time and route of administration for clinical use remain to be resolved.

Funding

This work was supported by the Science and Technology Bureau of Quanzhou (grant number 2020CT003).

Disclosure

The authors report no conflict of interest.

References