Functional role of mesenchymal stem cells in the treatment of chronic neurodegenerative diseases

1 INTRODUCTION

Stem cells have the ability to renew themselves continuously or differentiate into many cell types. In adult life, stem cells have been identified in several tissues where, by the continuous production of new cells, they guarantee physiological conditions and repair mechanisms after injuries or diseases. In many organs, such as the blood, skin, or gastrointestinal tract, stem cells extensively work throughout the organism's entire life, contributing to rapid replacement of dead or worn out cells. Other tissues, in which repair mechanisms are not so efficient, do not easily recover after injuries or extensive degenerative events.

This is the case of the central nervous system (CNS), where for a long time it was believed that only microglia, astrocytes, and oligodendrocytes proliferate in the adult organism, whereas, neurons were considered unable to divide. Indeed, this limitation has to be partially revised, since it has been demonstrated that in at least two areas of the brain, the dentate gyrus of the hippocampus and the subventricular zone adjacent to the lateral ventricles (SVZ), neural stem cells (NSCs) can continuously generate new functional neurons (Gage, 2002). However, they likely account for neurogenesis only within limited regions of the brain and are not able to counteract diffuse neuronal death as it occurs in neurodegenerative diseases (Lie, Song, Colamarino, Ming, & Gage, 2004). In these cases, the gradual and progressive loss of neural cells severely impairs the quality of life of patients, mostly because of cognitive and behavioral dysfunctions.

As pharmacological approaches have induced in many instances poor effects, stem cell-based brain transplantation therapies have been extensively investigated (Chan et al., 2014). Different lines of stem cells have been explored (Blundell & Shah, 2015). Embryonic stem cells have been widely studied for their pluripotency and very high proliferative potential. However, they raise ethical/religious issues and involve the risk of easily producing tumors. Induced pluripotent stem cells, reprogrammed from autologous somatic cells could also differentiate to neural cells, but their safe use has not yet been properly assessed. Considering their nature, NSCs would appear the most suitable tool (Kim, Lee, & Kim, 2013). In fact, they already give origin to the three main cell types of the mammalian CNS: neurons, astrocytes, and oligodendrocytes. Unfortunately, having to be extracted from the nervous tissue, autologous NSCs are hardly available in sufficient amounts without significant harm for the patient (Chan et al., 2014). In addition, data on NSC transplantation are controversial, since when injected in vivo, they may remain in the undifferentiated form. These cells are also susceptible to immune responses upon allogeneic transplantation (Rossignol et al., 2014).

The use of MSCs appears more suitable since they are available in satisfactory amounts, feature low immunogenicity, can differentiate into multiple cell lineages (Arutyunyan, Elchaninov, Makarov, & Fatkhudinov, 2016; Li & Ikehara, 2013; Lo Furno, Mannino, Cardile, Parenti, & Giuffrida, 2016), including neurons and glial cells when cultured under specific conditions (Lo Furno et al., 2013). Moreover, they can be used in clinical treatments without raising significant ethical or religious issues (Tanna & Sachan, 2014). Among many available tissues, the most studied MSCs are those from umbilical cord (UC-MSCs), bone marrow (BMSCs) and adipose tissue (ASCs). Some other sources have also been examined, such as dental pulp (Xiao & Tsutsui, 2013) or even menstrual blood (Zemel'ko et al., 2013).

In the present review, we focus on results obtained using MSCs as a potential therapeutic tool for the treatment of some disabling neurodegenerative disorders such as Alzheimer's disease (AD), Amyotrophic Lateral Sclerosis (ALS), Huntington's disease (HD), Parkinson's disease (PD). The large number of positive effects reported, both at histological and behavioral levels, have encouraged a progressively increasing number of therapeutic approaches in humans.

2 ALZHEIMER'S DISEASE

Alzheimer's Disease, the most common type of dementia, is characterized by degeneration and death of neurons throughout the brain, mainly in the basal forebrain, amygdala, hippocampus, and cerebral cortex. AD is considered a multifactorial and heterogeneous disease and is generally known as a late-onset disease, accompanied by cognitive dysfunction and associated with the accumulation of amyloid β (Aβ) plaques and neurofibrillary tangles. As the disease progresses, memory and cognitive functions progressively decline, patients become demented and prematurely die (Chan et al., 2014; Chang, Lee, & Suh, 2014; Kim et al., 2013). Although the pathogenesis of the disease is not clearly defined, genetic factors such as rare mutations in amyloid precursor protein (APP), apolipoprotein E, and presenilins (PS) seem to be involved in about 5% of cases.

Dysfunction of the presynaptic cholinergic system is considered one of the causes of cognitive disorders. Thus, pharmacological therapy has been mainly based on molecules aimed at increasing the acetylcholine concentration, by inhibiting acetylcholinesterase. In some people, these treatments can temporarily improve symptoms or slow down their progression, but they are mostly palliative, without real protection against progressive cell degeneration (Amemori, Jendelova, Ruzicka, Urdzikova, & Sykova, 2015). Searching for more effective treatments, a cell-based therapy could represent a promising tool to ameliorate neuropathological deficits.

Most preclinical studies have been carried out in double transgenic animal models characterized by expression of mutated APP and PS1 (APP/PS1). Otherwise, the animal model was generated by intracerebral/intracerebroventricular injections of Aβ aggregates. Less frequently, some other animal models have been adopted. The main findings are summarized in Table 1.

3 AMYOTROPHIC LATERAL SCLEROSIS

Amyotrophic lateral sclerosis (ALS) is a severe, persistent neurodegenerative disease. It is characterized by the progressive loss of motor neurons in the cerebral cortex, brain stem, and spinal cord. Clinical symptoms consist of a gradual weakening and atrophy of muscles. Death usually occurs within 5 years from diagnosis and is mostly due to paralysis of respiratory muscles (Kim et al., 2013). Although the pathogenesis is still unknown, probable mechanisms involved in this multifactorial disease include glutamate-mediated excitotoxicity, oxidative damage, protein aggregation, mitochondrial dysfunction, altered axonal transport, and insufficient production of neurotrophic factors (Boillée, Vande Velde, & Cleveland, 2006). Most cases are sporadic (90%), whereas, the remaining cases are familial (FALS). In 15–20% of FALS cases, mutations of the gene encoding superoxide-dismutase 1 (SOD1) have been detected. This genetic basis has permitted the development of transgenic rodent models featuring clinical symptoms and histopathological features of both familial and sporadic human ALS (Bendotti & Carri, 2004). Among various mutant human SOD1 genes (G37R, G85R, G93A), G93A is the most commonly used. The main findings are summarized in Table 2.

4 HUNTINGTON'S DISEASE

Huntington's disease (HD) is an autosomal dominant, late-onset neurodegenerative disease. Although it may occur before 20 years of age, in more than 90% of cases it develops at an age of 35–40 years (Fink et al., 2015). Characterized by progressive neuronal cell death, HD symptoms consist of a cognitive decline, behavioral changes, involuntary choreic movements, and psychiatric disturbances. HD is associated with an expanded and unstable CAG trinucleotide repeat in the huntingtin gene (htt), which yields abnormal accumulation of the Htt protein and cellular toxicity. As a result, a significant loss of neurons occurs, particularly GABAergic medium-sized spiny neurons in the caudate nucleus and putamen. Various other brain regions may, however, be involved, especially in later stages of the disease. In HD patients, a substantial reduction (about 19%) of total brain volume can be found, associated with an enlargement of lateral ventricles. Currently, available pharmacological treatments can only relieve symptoms to improve the quality of life in HD patients. The disease constantly progresses and culminates in death within 15–20 years after the onset of motor symptoms.

Preclinical investigations to develop cell-based therapeutic approaches have been mainly carried out in two categories of animal models (Kerkis, Haddad, Valverde, & Glosman, 2015; Lescaudron, Naveilhan, & Neveu, 2012; Maucksch, Vazey, Gordon, & Connor, 2013). In the first type, the disease is simulated by inducing intrastriatal neuronal death by excitotoxic lesions caused by injections of kainic acid, quinolinic acid (QA), or 3-Nitroproprionic acid (3-NP). The second class includes genetically modified animal models overexpressing Htt protein (N1T1-82Q2, R6/2, R6/2-J2, N171-82Q, YAC128). The main findings are summarized in Table 3.

5 PARKINSON'S DISEASE

Parkinson's disease (PD) is a common neurodegenerative disorder. It is characterized by a progressive and extensive degeneration of dopamine-producing neurons in the pars compacta of Substantia Nigra (SNpc) at mesencephalic level, as well as their terminals in the striatum. However, other brain regions, such as the cerebral cortex and lower brain stem, can also be affected. Because of the functional role of nigrostriatal projections, typical motor symptoms of PD include resting tremor, muscle rigidity, bradykinesia, postural instability, and impaired ability to control voluntary movements. Other symptoms include cognitive dysfunctions, sleep disturbances and hyposmia. PD is commonly considered a non-genetic disorder since viral infections, programmed cell death, or environmental factors may play a critical role in the etiology. However, in about 5% of cases, PD seems associated with specific gene mutations. In most of these cases, mutations of the α-synuclein gene is responsible for intracytoplasmic eosinophilic inclusions (Lewy bodies), and spontaneous cell death (Kalia, Kalia, McLean, Lozano, & Lang, 2013). The main findings are summarized in Table 4.

Effective treatments for PD are not available at the present. Since the disease is associated with the lack of nigrostriatal dopaminergic projections, dopamine replacement is commonly used to relieve PD symptoms. However, long-term treatment with L-DOPA induces severe side effects, including drug-induced motor complications, and above all, abnormal, uncontrollable movements, known as dyskinesia (Lescaudron et al., 2012). For these reasons, alternative approaches have been designed.

Transplantation of human fetal brain tissue into the striatum of patients with advanced PD has been able to improve their cognitive and motor behavior (Kim et al., 2013). However, these strategies imply some limitations because of the difficulty to obtain enough fetal tissue and for ethical and religious issues. These difficulties may be overcome by stem cell-based therapies, especially using MSCs. Experiments were mostly carried out on animal models, in which PD-like symptoms can be induced by intracerebral injections of 6-hydroxydopamine (6-OHDA; especially for rat models) or after intravenous/intraperitoneal administration of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP; more suitable for mouse and non-human primate models). Another animal model featuring PD symptoms can be generated by intraperitoneal/subcutaneous administration of rotenone, which replicates many aspects of the human disease.

6 GENETICALLY ENGINEERED MESENCHYMAL STEM CELLS

A large body of evidence demonstrates that MSCs transplanted into the brain promote functional recovery mainly by producing trophic factors that favour survival and regeneration of host neurons. Their typical trophic support can be enhanced using genetically engineered MSCs as delivery vehicles. For treatment of chronic neurodegenerative diseases, MSCs have been most commonly genetically modified to overexpress neurotrophic factors (mainly NGF, BDNF, GDNF) whose neuroprotective/neuroregenerative actions are widely acknowledged. (Deng et al., 2016; Huang, Tabata, & Gao, 2012; Wyse, Dunbar, & Rossignol, 2014). In addition, surface receptor expression can be similarly manipulated to enhance MSC migration toward injured sites. Different strategies have been investigated for engineering MSCs (Muroski, Levenson, & Strouse, 2014). Though efficient, electroporation techniques imply significant loss of viable cells; viral vectors are commonly used, despite considerable biosafety restrictions, such as risks of secondary infection, immunogenic responses, excessive secretion of neurotrophins; non-viral approach are also available, for example by using solid gold nanoparticles. The main findings are summarized in Tables 1-4, for related diseases.

Gene-modified BMSCs to overexpress NGF were transplanted into the hippocampus of an AD rat model, in which the β-amyloid protein was injected bilaterally into the hippocampus (Li et al., 2008). Transplanted cells survived and migrated into surrounding tissue and other sites (corpus striatum and other hippocampal structures) acquiring a tendency to express choline acetyltransferase-like immunopositivity. Compared to controls, a reduced neuronal loss was observed in the hippocampus of these animals, showing significant progress in learning and memory.

Retro-orbital injections of genetically modified hUC-MSCs were carried out in a pre-symptomatic G93A mouse model of ALS (Rizvanov et al., 2008). UC-MSCs were transiently transfected by electroporation to express human VEGF and the mouse neural L1 cell adhesion molecule (L1CAM). As a result, UC-MSCs successfully grafted into nervous tissue of ALS mice and survived for over 3 months. According to the authors, the expression of L1CAM was responsible for increased homing ability and proliferation of transplanted cells at the site of neurodegeneration in the spinal cord parenchyma. However, rather than neural, they differentiate into endothelial cells forming new blood vessels, likely because of increased VEGF expression. This might, however, enhance neuroprotective/neuroregenerative effects because more neurotrophic factors would be delivered to the damaged regions from newly formed blood vessels.

Positive data were reported by injecting BMSCs, genetically modified using a retrovirus to overexpress NGF and/or BDNF, into the striatum of a YAC128 mice model of HD (Dey et al., 2010). Compared to controls, transplanted animals exhibited reduced clasping and improvements on the rotarod task. In addition, striatal neuronal nuclei (NeuN) positive cell counts were restored almost to those of wild-type mice.

By lentiviral transduction, hBMSCs, engineered to overexpress BDNF, were transplanted into the striatum of immune-suppressed HD transgenic mouse models (Pollock et al., 2016). Decreased striatal atrophy and reduced anxiety were detected in YAC128 mice. A significant increase in neurogenesis-like activity and the mean lifespan were observed in R6/2 mice.

Genetically, modified hBMSCs were implanted into the striatum of 6OHDA rats. Cells were first transfected with the intracellular domain of the Notch1 gene to generate SB623 cells; then, lentiviral transduction was performed to enhance their expression of GDNF (Glavaski-Joksimovic et al., 2010). After 1 week, histological analysis showed that numerous engrafted cells were surviving. At 2 weeks, their number was significantly reduced, and only cell debris was found at 4 weeks. However, although no surviving SB623 cell was present, numerous rejuvenated TH-positive fibers were observed after 5 weeks. The authors suggest that, even if GDNF was secreted only shortly after the grafting, it was able to exert longer functional recovery in this rat PD model. This conclusion is supported by statistically significant decreases in the amphetamine-induced rotation, evaluated at 4 weeks post implantation. They also hypothesize that GDNF secreted into the striatum could be uptaken by nigrostriatal fibers and retrogradely transported to neuronal cell bodies in the SN. Similar conclusions were drawn by Hoban, Howard, and Dowd (2015) in an inflammatory rat model of PD (the lipopolysaccharide model). After intrastriatal injection of BMSCs, genetically engineered to overexpress GDNF, dense areas of TH-immunoreactivity were revealed in proximity to the transplant site. It is concluded that, although most transplanted cells die within 1 week, GDNF secretion is able to provide a longer-lasting protection of nigrostriatal terminals, as evaluated 21 days after transplantation.

Promising results have also been described in an MPTP monkey model of PD, using autologous BMSCs, genetically modified to overexpress GDNF (Ren et al., 2013). In particular, 2 weeks before intravenous MPTP administration, unilateral injections of GDNF-MSC were made into the striatum and SN. In these monkeys, motor functions were better preserved in the contralateral limbs. Moreover, higher dopamine levels and greater dopamine uptake were found in the striatum of the grafted hemisphere, along with more numerous TH-positive fibers between the SN and striatum. It is suggested that this procedure represents a safe vehicle to deliver GDNF, able to preserve nigral and striatal functions, although the loss of SN dopaminergic neurons was not prevented in this work. The authors admit that, for a closer simulation of human therapeutic applications, further studies should be differently designed. More precisely, MPTP induction of the disease should precede GDNF-MSC transplantation. In fact, therapeutic approaches occur in patients already at a late PD stage.

7 CLINICAL TRIALS

Based on positive outcomes from animal models, some clinical trials have already been performed and many others are currently in progress (clinicaltrials.gov). Many of them have been designed to assess the safety of treatments. The main findings are summarized in Tables 1, 2, and 4 for related diseases.

For AD patients, various types of MSCs and different administration routes have been explored worldwide (Duncan & Valenzuela, 2017). In a Korean phase 1 clinical trial, hUC-MSCs were stereotactically injected into the hippocampus and precuneus of nine patients with mild-to-moderate AD (Kim et al., 2015). Although no attenuations of the pathology are described, no serious adverse events are reported. Therefore, the authors conclude that stereotactic administration of hUCB-MSCs into the brain of patients is feasible, safe, and well tolerated.

Numerous clinical trials are being made in ALS patients, especially using BMSCs. Most of them (phase I/II) have been designed to demonstrate the safety of transplantation procedures, but some progress has already been observed (Czarzasta et al., 2017; Mazzini, Vescovi, Cantello, Gelati, & Vercelli, 2016). The functional status of patients was generally assessed by the Amyotrophic Lateral Sclerosis Functional Rating Scale (ALS-FRS), the most widely used scale to evaluate disease progression. This scale analyzes 12 parameters: speech, salivation, swallowing, writing, eating and utensils use, dressing and hygiene, turning in bed, walking, stairs climbing, breathing, orthopnea, and respiratory insufficiency.

In a phase II clinical study, early improvements of symptoms were described after injection of autologous BMSCs directly into the brainstem and cervical spinal cord of 13 patients in Turkey (Deda et al., 2009). After a follow-up period of 1 year, although three patients died of lung infection and myocardial infarct, nine of them were in better conditions than the preoperative period, even though with some degree of decline. One patient remained stable. According to the authors, these effects would be due both to implanted BMSCs and activation of endogenous stem cells.

A phase I/II clinical trial was carried out in Israel by Karussis et al. (2010). In patients with ALS and multiple sclerosis, intrathecal and intravenous delivery of autologous BMSCs was designed to maximize their migration to the nervous system through the cerebrospinal fluid (CSF) and blood circulation. Magnetic resonance imaging showed a possible propagation of BMSCs from the lumbar site of inoculation to the occipital horns, meninges, spinal roots, and spinal cord parenchyma. In addition, beneficial immunomodulatory influences were also obtained as early as 4 hr after BMSC transplantation. The ALS-FRS score showed a slight deterioration in the 2 month interval before BMSC administration, whereas a clinical stabilization or even amelioration was observed in some patients during a follow-up of 6–25 months, without any significant immediate or late adverse effects.

In a phase I clinical trial carried out in Italy on 10 ALS patients (Mazzini et al., 2010), autologous BMSCs were injected into the spinal cord at a high thoracic level, after in vitro expansion and suspension in autologous CSF. As no serious adverse effects were noted in treated patients, the authors assume that this procedure is safe enough to be further used and developed for future approaches.

The safety of intraspinal (posterior spinal cord funiculus) infusion of autologous bone marrow mononuclear cells was proven on ALS patients in Spain (Blanquer et al., 2012). Four patients died for reasons unrelated to the procedure, whereas no serious transplant-related adverse events were observed in the others. Although no significant clinical improvement is reported, no acceleration was noticed in the rate of decline of forced vital capacity, ALS-FRS, and other specific scales. Notably, autopsy samples showed a higher number of motoneurons in the anterior horns at the level of the transplanted spinal segments.

Two repeated intrathecal autologous BMSC injections were performed in a Korean phase I clinical study on seven ALS patients (Oh, Kim et al., 2015). The authors believe that this procedure is safe and feasible, as no serious adverse events occurred during the 12-month follow-up period. They also claim that this procedure would maximize BMSC migration and strengthen therapeutic outcomes in brain and spinal cord parenchyma.

Autologous BMSCs were transplanted in ALS patients in Belarus by a combined administration method (Rushkevich et al., 2015). After 6–7 days from intravenous delivery of undifferentiated BMSCs, neural-induced BMSCs were injected via lumbar puncture. The authors assume that this cell therapy is safe since no serious side effects or complications occurred after a 12-month follow-up period. They also report a slowed disease progression and an improvement of the functional status, as evaluated by the ALSFRS score.

In a phase I/IIa clinical trial, autologous BMSCs were isolated, expanded, and injected via lumbar puncture into the CSF of 23 ALS patients in the Czech Republic (Syková et al., 2017). Results indicate that this intrathecal application of BMSCs was safe since no serious adverse reactions were observed after an 18-month follow-up period. The authors also report a slowing down in the progression of the disease since a reduction in the ALSFRS score decline was revealed at 3 months and, in some cases, for 6 months after BMSC administration. Forced vital capacity remained stable or above 70% for 9 months.

Autologous BMSCs, in vitro induced to secrete neurotrophic factors, were transplanted in patients with ALS in Israel (Petrou et al., 2016). In this phase I/II clinical study, intramuscular (six patients at early stage of ALS) or intrathecal (six patients at more advanced stages) injections were made. In phase IIa, 14 patients with early stage ALS received a combined intramuscular and intrathecal transplantation. Overall, the forced vital capacity and the ALSFR score were improved, compared to the expected decline, especially in patients receiving intrathecal or both injections. Since only mild and transient treatment-related adverse effects were noticed over the 6 months of follow-up, the procedure was considered safe and well tolerated, to be further used in clinical trials for patient benefit.

Clinical trials on PD patients have also been already started. In an Indian clinical trial, autologous BMSCs were transplanted stereotactically into the SVZ of seven patients with advanced PD (Venkataramana et al., 2010). This brain region was chosen since it physiologically contains the largest number of NSCs involved in neurogenesis. In a follow-up period of 10–36 months, amelioration of clinical conditions was observed by using the Unified Parkinson's Disease Rating Scale (UPDRS), Hoehn and Yahr (H&Y), and Schwab and England (S&E) scales. Subjective improvements were also reported for symptoms such as facial expression, gait, and freezing episodes. Moreover, the dosage of L-DOPA was significantly reduced in two patients. Being an open-label study, the authors do not entirely exclude a placebo effect, although the persistence of clinical improvements over 20–26 months makes it unlikely. Even if the exact mechanisms are not clearly understood, this trial has confirmed the safety of the procedure, as none of the patients showed clinical worsening until the end of the observation period. In addition, cranial magnetic resonance imaging did not show any abnormal parenchymal changes.

In an Italian pilot phase I study, infusion of BMSCs into the cerebral arteries was tested in five patients affected by progressive supranuclear palsy (PSP), a rare, severe, and no-option form of Parkinsonism (Canesi et al., 2016). Most patients were alive after 1 year from the administration, except one who died for reasons not associate with the treatment (accidental fall). Clinical conditions of all patients were assessed using specific rating scales (UPDRS, H&Y, PSP rating scale) before cell administration and after 1, 3, 6, and 12 months. Rating scale scores indicate that motor functions remained stable for at least 6 months during the 1-year follow-up. These findings are considered encouraging since in these patients motor function deterioration would have been normally faster. The next phase II study will definitively provide information on the efficacy of this approach, easily applicable to other neurodegenerative disorders.

Altogether, although no serious adverse effects have been reported, the safety of treatments has to be carefully evaluated when using cell-based therapies in humans. In fact, it is not entirely known whether surviving cells may give rise to abnormal nervous tissue, maintain their stemness features, or even evolve into different cell lines and induce unexpected events. Another potential risk could be related to their angiogenetic activity and their tendency to home to the hypoxic regions near tumors (Joyce et al., 2010). As these features could became extremely dangerous, a careful screening must be done in selecting patients for stem cell implantations. Cell-free strategies, using exosomes or conditioned media, are perhaps closer to widespread clinical applications as ideal carriers to delivery therapeutic agents to nervous tissue that is otherwise difficult to reach (Lee et al., 2016). Notably, the potential benefit-to-risk ratio is another major factor to be considered. Particularly for those patients in advanced stages of disabling diseases, when no effective treatments are available.

8 CONCLUSIONS

Data from the studies above described show that the use of MSCs may induce appreciable amelioration in neurodegenerative diseases, where only poor results can be achieved by pharmacological approaches or other therapeutic strategies. Beneficial MSC-induced effects primarily rely on their particular tendency to home to injured areas, in particular to hypoxic, apoptotic, or inflamed areas (Figure 1). Notably, their ability to pass the BBB has permitted different administration routes (Kerkis et al., 2015). In fact, MSC engrafting at the injured nervous tissue has been reported not only upon intracerebral injection, but also after intracerebroventricular, intracarotid, intravenous, or even intranasal administration.

However, it should be pointed out that, although less invasive, systemic administrations have some limitations. After intravenous injections, many cells are trapped in the lungs and fail to reach target tissues. Intra-arterial delivery implies more invasive procedures and may lead to microthrombi formation, thus, compromising tissue microcirculation. The wide dispersion of cells in the blood stream would require a strong homing ability, which may be affected by the loss of expression of homing molecules during in vitro expansion. In this regard, different strategies have been developed. For example, by expanding MSCs under hypoxic conditions, by adding various chemicals or cytokines to the culture medium or by genetic engineering to overexpress homing molecules. These issues have been extensively addressed in a recent review by De Becker and Van Riet (2016).

Different mechanisms can be taken into account to explain MSC-mediated improvements. From in vitro studies, it has been reported that they can differentiate into neural elements, expressing typical neuronal and glial markers (Lo Furno et al., 2013). On this basis, the optimistic aim of MSC transplantation into the brain is to replace lost neurons and restore damaged nervous networks. Data obtained in vivo weakly corroborate this hypothesis. In most studies, only a small fraction of transplanted cells has been verified as stably engrafted, and only in some cases have they shown neuronal or glial differentiation. In addition, the number of grafted cells rapidly decreases, and often none of them can be detected after a few weeks from transplantation. On this basis, one must conclude that most beneficial actions are exerted rather by mechanisms that are different from MSC neural differentiation.

Most authors agree that the neuroprotective/neurotrophic effects observed upon MSC transplantation are mainly promoted by their paracrine activity. Indeed, the production of an impressive amount of neurotrophic growth factors, chemokines and cytokines has been demonstrated (Drago et al., 2013; Lee et al., 2016; Paul and Anisimov, 2013). By these paracrine actions, the local immune system can be suppressed, and apoptosis and free radicals levels reduced (Joyce et al., 2010). MSC paracrine activities seem improved by pro-inflammatory and hypoxic stimuli, by the exposure to apoptotic cell environment, and by activated microglia. Moreover, recruitment, differentiation and paracrine function of resident progenitor cells, neurons or glial cells can be enhanced. Additional positive mechanisms include angiogenesis stimulation, enhanced glutamate uptake, or production of extracellular matrix molecules. Stimulation of angiogenesis would help the diffusion of soluble factors along newly formed vessels in the injured tissue. By enhancing glutamate uptake, its neuro excitotoxicity would be decreased. The extracellular matrix molecules may stimulate growth and axonal extension. As modifications of cytokines and neurotrophic factors are also associated with various mental illness, MSC based therapies have also been suggested for psychiatric disorders (Colpo et al., 2015).

Some particular features of UC-MSCs, BMSCs, or ASCs deserve to be highlighted (Momin, Mohyeldin, Zaidi, Vela, & Quiñones-Hinojosa, 2010). UC-MSCs can be derived with no discomfort for the newborn or the mother, from the already detached umbilical cord. However, their use has some limitations. They can only be used for allogeneic transplantation and the cell yield might be limited. BMSCs were the first to be identified and have been the most studied and, especially for in vivo studies and clinical trials, are currently widely tested. However, this source of MSCs may not be the easiest one, since BMSC harvesting is characterized by limited cell yield and painful invasive procedures. In many respects, the use of ASCs seems more suitable for translational medicine. First, their availability is virtually unlimited from subcutaneous deposits of adipose tissue, which is easily harvested with minimally invasive procedures (Yeh et al., 2015). Second, in comparison with BMSCs, ASCs have often shown higher proliferation and differentiation ability for mesodermal and neural lineages (Calabrese et al., 2015; Zhang, Liu, Yao, Yang, & Xu, 2012). Not only can ASCs be quickly obtained for autologous use but, thanks to their low immunogenicity, they are also suitable for allogeneic transplantations.

It is reasonable to expect that more studies and clinical trials will be carried out in the near future using MSCs, either as naïve stem cells, or after neural induction or genetically engineered to enhance their innate paracrine and chemotactic potential, at least until more efficacious cell-replacement therapies are available to restore disease-disrupted brain circuitry.

CONFLICT OF INTEREST

The authors declare no conflict of interest.