Concise Review: Spinal Cord Injuries: How Could Adult Mesenchymal and Neural Crest Stem Cells Take Up the Challenge?


  • Virginie Neirinckx,

    1. Groupe Interdisciplinaire de Génoprotéomique appliquée (GIGA), Neurosciences Unit, Liège, Belgium
    Search for more papers by this author
  • Dorothée Cantinieaux,

    1. Groupe Interdisciplinaire de Génoprotéomique appliquée (GIGA), Neurosciences Unit, Liège, Belgium
    Search for more papers by this author
  • Cécile Coste,

    1. Groupe Interdisciplinaire de Génoprotéomique appliquée (GIGA), Neurosciences Unit, Liège, Belgium
    Search for more papers by this author
  • Bernard Rogister,

    1. Groupe Interdisciplinaire de Génoprotéomique appliquée (GIGA), Neurosciences Unit, Liège, Belgium
    2. GIGA, Development, Stem Cells and Regenerative Medicine Unit, University of Liège, Liège, Belgium
    3. Department of Neurology, Centre Hospitalier Universitaire de Liège, Liège, Belgium
    Search for more papers by this author
  • Rachelle Franzen,

    1. Groupe Interdisciplinaire de Génoprotéomique appliquée (GIGA), Neurosciences Unit, Liège, Belgium
    Search for more papers by this author
  • Sabine Wislet-Gendebien

    Corresponding author
    1. Groupe Interdisciplinaire de Génoprotéomique appliquée (GIGA), Neurosciences Unit, Liège, Belgium
    • Correspondence: Sabine Wislet-Gendebien, Ph.D., GIGA—Neuroscience, University of Liège, Tour de Pathologie 2, Avenue de l'Hôpital, 1, 4000 Liège, Belgium. Telephone: 32-4-366-39-88; Fax: 32-4-366-23-14; e-mail:

    Search for more papers by this author


Since several years, adult/perinatal mesenchymal and neural crest stem cells have been widely used to help experimental animal to recover from spinal cord injury. More interestingly, recent clinical trials confirmed the beneficial effect of those stem cells, which improve functional score of patients suffering from such lesions. However, a complete understanding of the mechanisms of stem cell-induced recovery is seriously lacking. Indeed, spinal cord injuries gathered a wide range of biochemical and physiopathological events (such as inflammation, oxidative stress, axonal damage, demyelination, etc.) and the genuine healing process after cell transplantation is not sufficiently defined. This review aims to sum up recent data about cell therapy in spinal cord lesions using mesenchymal or recently identified neural crest stem cells, by describing precisely which physiopathological parameter is affected and the exact processes underlying the observed changes. Overall, although significant advances are acknowledged, it seems that further deep mechanistic investigation is needed for the development of optimized and efficient cell-based therapy protocols. Stem Cells 2014;32:829–843

Mesenchymal and Neural Crest Stem Cells: Definition and Rationale for Spinal Cord Injury Treatment

Mesenchymal stem cells (MSCs) in the adult bone marrow stroma were first identified in the last 1970s by Friedenstein et al. [1, 2] as colony-forming unit fibroblast-like cells. Those mesoderm-derived cells were subsequently described to be self-renewable and highly multipotent, giving rise to different cells of mesodermal origin such as adipocytes, chondrocytes, and osteocytes [3, 4]. In the following years, MSCs were isolated from other adult tissues (e.g., fat tissue [5], muscle [6], synovium [7], peripheral blood, circulatory system [8], and so forth), where they contribute and regulate organ physiology and homeostasis. In addition, it was shown that MSCs were also present in perinatal tissues like umbilical cord blood (UCB) [9] or Wharton's jelly [10], amniotic fluid [11], and placenta [12] (Table 1).

Table 1. MSCs and NCSCs in adult and perinatal tissues
  1. Abbreviations: AT-MSCs: adipose tissue mesenchymal stem cells; BMSCs: bone marrow stromal cells; DPSCs: dental pulp stem cells; EPI-NCSCs: epidermal neural crest stem cells; HBCs: horizontal basal cells; LP-NCSC: lamina propria neural crest stem cells; OECs: olfactory ensheathing cells; SHED: stem cells from human exfoliated deciduous teeth; SKPs: skin-derived precursors; UCB-MSCs: umbilical cord blood mesenchymal stem cells.


Bone marrow stromal cells (BMSCs) were especially considered for cell therapy in neurological lesions regarding their capacity to give rise to neural-like cells [13-15]. However, in vivo neural differentiation is currently matter of debate and it seems that adult BMSCs would rather help lesion recovery through many other mechanisms than in-host differentiation [16, 17]. Indeed, besides the immunomodulatory effects and the secretion of several neurotrophic factors, those cells have an anti-inflammatory effect, making them attractive candidates as the most efficient treatment for spinal cord injury (SCI), so far, is a high-dose of methylprednisolone (anti-inflammatory drug) [18]. Moreover, a recent clinical trial showed that BMSCs transplantation inside the cerebrospinal fluid of spinal cord-injured patients modestly enhanced motor and sensitive functions. As this work provided first clues about BMSCs relevancy for SCI treatment, the underlying ways of recovery observed in these cases were not explained [19].

More recently, neural crest stem cells (NCSCs) were identified inside the adult bone marrow [20, 21]. Those cells initially derive from neural crest, which arises at the borders of the neural tube during embryonic development of the nervous system. NCSCs then migrate toward different organs, where they differentiate to give rise to peripheral neurons and glia, melanocytes, chondrocytes, smooth muscle cells, and so forth. Adult NCSCs were identified in several postnatal organs [22], intermingled with MSCs inside the bone marrow [15, 20, 23, 24], skin [25-27], gut [28], teeth [29, 30], heart [31], palatum [32], cornea [33], and olfactory tissue [13-15] (Table 1).

Inside the bone marrow stroma, both MSCs and NCSCs accumulate at the bone epiphysis, and they are consequently often studied together and referred as BMSCs. This fact makes BMSCs even more attractive for SCI treatment as beside the inflammatory modulation effects provided by MSCs, NCSCs may be the perfect candidate for cell replacement therapy. Indeed, due to their neural origin, neural crest cells are closely related to neural tube stem cells, which makes them in close ontological relationship with the spinal cord [37]. Furthermore, NCSCs express the NCSC molecular signature genes that were initially used to create induced pluripotent stem cells (iPSC) [38]. This last fact makes adult NCSCs more attractive than iPSC or embryonic stem cells (ESCs) according to the guidelines for the clinical translation of stem cells, published by the International Society for Stem Cell Research: “…maximum effort should be made to minimize the risks for all possible adverse events associated with stem cell-based therapy.” Therefore, adult bone marrow NCSCs are more suitable as they are not genetically modified like iPSC and are less immunogenic (similarly to BMSCs) than ESCs.

According to the last update reported by Lee et al. [39], the global incidence of traumatic SCIs was estimated in 2007 at 23 cases per million worldwide. Reported SCI cases mainly concern young adults, among them 80% of men, who are for the most part victim from motor vehicle accidents and falls [40]. The cervical spine and lumbar spine are the most commonly affected regions, and regarding the level of injury, those patients could suffer from para- or tetraplegia. Besides locomotor impairments, neuropathic pain, reflexive, and sensitive troubles also occur, accompanied by highly disabling social and financial issues. In the last decade, although numerous reports have shown significant improvements in medical management and clinical recuperation after SCI, there is still no effective treatment that completely allows functional recovery.

The development of such effective treatments should first be based on the full understanding of SCI physiopathological events, which are gathered in three major phases. Briefly, the first events occuring in the acute phase following traumatic SCI and spinal shock encompass neuronal necrosis, and axonal disruption, blood supply default, edema and accumulation of electrolytes, calcium but also potassium and excitotoxic neurotransmitters. The intermediate phase starts a few minutes after the lesion, and lasts for several weeks. It is characterized by further ischemia, oxidative stress taking place by free-radical production, and lipid peroxidation as well as by the recruitment of neutrophils and lymphocytes which secrete cytokines and promote the development of an inflammatory environment. The chronic phase arises after few months, covering continuous alteration of ionic balance, apoptosis of oligodendrocytes and demyelination, formation of cavities, and astroglial scar formation (reviewed by Oyinbo, in 2011 [41] and by Ronaghi et al., in 2010 [42]) (Fig. 1).

Figure 1.

Physiopathological events occurring after spinal cord injury, contributing to the harmful environment that hampers axonal regrowth and recovery.

Altogether, those unfavorable events hamper axonal regrowth and functional recovery. Therefore, in this review, we will detail updated evidences (coming from either in vitro or in vivo preclinical experiments performed in the last 7 years) about the different abilities of adult MSCs and NCSCs to manage neural tissue-associated oxidative stress, inflammation or apoptosis, neurodegeneration, and demyelination, which could be highly relevant in spinal cord injury cases and provide new clues about the precise mechanisms that could trigger recovery. All the data gathered in this review are classified in an order which follows the temporal progression of SCIs as suggested by the most relevant studies in that field. Still we keep in mind that transplanted MSCs/NCSCs can act on different events concomitantly, soon after the injury or later.

Various Properties of MSCs and NCSCs and Spinal Cord Injury Treatment

Acute Phase Events: Axonal degeneration, Ion Balance Disruption, and Excitotoxicity

Neurotrophic Support by MSCs/NCSCs and Protection Against Glutamate Toxicity

Traumatic damages to the spinal cord induce axonal interruptions that are more or less complete according to the degree of injury. Regrowth of disrupted nervous fibers is highly hampered first by the hostile environment and later on by the astroglial scar, which significantly inhibits nerve regeneration and prevents axons to go through. Synergistic strategies could be considered in order to enhance axonal regrowth, combining neurotrophic factor supply to degenerated axons; and afterward, degradation of extracellular matrix/glial scar and reconstitution of their myelin sheaths (see below).

Different papers described that MSCs/NCSCs could serve as trophic mediators in diverse situations of tissue lesions [43, 44], which make their use relevant in the context of spinal cord damages. Indeed, several studies reported that MSCs/NCSCs were able to secrete neurotrophic factors that exhibit substantial effects on neuron survival and neurite outgrowth, in both in vitro and in vivo paradigms (Table 2). For instance, BMSCs were able to support and direct axonal growth of adult dorsal roots ganglia (DRG) neurons in vitro, by secreting extracellular matrix that orientate neurite extension, but especially by secreting molecules (as BMSCs-conditioned medium [BMSCs-CM] treatment generated the longest neurites) [45]. Besides, it was shown that BMSCs and BMSCs-CM (even when microvesicle-depleted) were able to protect cortical neurons against N-methyl-d-aspartate excitotoxicity [46]. Unfortunately, no details about the precise molecules present inside the CM were provided.

Table 2. Acute phase events: Neurotrophic properties of MSCs/NCSCs, axonal regrowth and protection against excitotoxicity
 Cell typeCell stateModelModalitiesResultsRef
In vitroh BMSCNaiveBMSC layer or CM treatment on DRG neurons inline image Neurite extension/length[45]
 h BMSCNaiveCo-culture with SH-SY5Y cells and DRG explants inline image Cell viability inline image Neurite extension/length = Role of BDNF and NGF secretion by BMSC[47]
 m BMSCNaiveCo-culture or CM treatment on cortical neurons for 24 hours: – NMDA 50 µM + Glycine 10 µpM Co-culture or CM treatment on retinal ganglion cells for 12 hours: inline image Cell viability inline image Mitochondrial dehydrogenase activity inline imageNMDAR subunits expression inline imageNMDAR-mediated calcium influx[46]
 Cell typeCell stateHostModelParametersResultsRef
  1. Abbreviations: 5-HT: 5-hydroxytryptamine (serotonin); AT-MSC: adipose tissue mesenchymal stem cells; BBB: Beattie Basso Bresnahan; BDNF: brain-derived neurotrophic factor; BMSC: bone marrow stromal cells; CM: conditioned medium; CSF: cerebrospinal fluid; CST: corticospinal tract; DC: dorsal column; DRG: dorsal root ganglion; GFAP: glial fibrillary acidic protein; h: human; IGF: insulin-like growth factor; IL: interleukin; IV: intravenous; LIF, leukemia inhibitory factor; MΦ: macrophages; NGF: nerve growth factor; NF: neurofilament; NMDA: N-methyl-d-aspartate; NMDAR: NMDA receptor; PDGF: Platelet-derived growth factor; PKC: Protein kinase C; r: Rat; RECA-1: Rat endothelial cell antigen 1; RST: Rubrospinal tract; SC: Spinal cord; T: Thoracic; vWF: von Willebrand factor.

In vivoh AT-MSC h BMSCNaiveRatSC section (CST, RST, OC) T9To depth of central canal2 × 105 cells Intraspinal − 1 × 105 cells (2 mm rostral) − 1 × 105 cells (2 mm caudal) Immediately after SCIAT-MSC > BMSC inline imagevWF+ blood vessels inline imageNF200+ fibers inline image 5-HT+ fibers inline image Cavity size inline imageED1+ MΦ inline image BDNF inline image BBB score[48]
 h BMSCBDNF transduction +PKH26RatSC section (DC) T91mm depth1.2  × 105 cells Intraspinal − 2  ×  3.104 cells (0.5 mm rostral) − 2  ×  3.104 cells (0.5 mm caudal) Immediately after SCIhBMSCBDNF > hMSC inline imagePKCγ+ CST fibers inline image 5-HT fibers inline image BBB score, Rotarod inline image PKCγ+ CST fibers = Role of BDNF secretion byBMSC[49]
 r BMSCBrdURatSC Compression T9-T1015 minute 15 µl volume balloon1 × 105 cells IV 7 days post-op inline image BBB score. Grid navigation inline image NGF expression inline image RECA1+blood vessels[50]
 r BMSCNaiveRatSC compression T81 minute 26g clip2 × 106 cells CSF infusion inline image NGF, LIF, IGF-1, IL-lβ inline imagePDGF-A[51]

Crigler et al. identified those factors and showed that BMSCs secreted significant levels of brain-derived neurotrophic factor (BDNF) and nerve growth factor (NGF), thereby partially favoring neurite extension when cocultured with neuroblastoma cells or DRG neurons [47]. The involvement of BDNF and NGF was then confirmed in animal models of SCI. Indeed, grafts of adipose tissue-derived MSCs (AT-MSCs) into the injured spinal cord of mice improved functional recovery, in correlation with increased BDNF expression and enhanced regrowth of serotonergic fibers. They also noticed that improvements were greater compared to BMSCs transplantation conditions [48]. Neuronal rescue coupled with pathological and behavioral progresses was also observed after graft of BDNF-hypersecreting-BMSCs in spinal cord-injured rats [49], suggesting a trophic role for grafted cells. NGF also seemed to be involved in SCI motor recovery and tissue sparing [50]. Finally, Hawryluk et al. showed that the expression levels of NGF, leukemia inhibitory factor, insulin-like growth factor-1, and transforming growth factor-β1 were significantly reduced in the rat spinal cord, 3 weeks after being compressed, and that intrathecal transplantation of BMSCs induced an upregulation of those trophic factors [51].

Intermediate Phase Events: Inflammation, Oxidative Stress, and Apoptotic Death of Host Cells

Immunomodulating Properties of MSCs/NCSCs and Regulation of Inflammation

The immunomodulatory properties of MSCs/NCSCs have been associated with both molecule secretions and cell-cell contact. First of all, it was shown that MSCs/NCSCs were able to suppress T-cell proliferation [52] and monocyte maturation into dendritic cells [53]. Moreover, MSCs/NCSCs impaired the functionality of dendritic cells, their antigen-presenting properties, and cytokine secretion [54], and also hampered the proper function of natural killer cells and their interleukin (IL)-2 secretion [55]. Many other studies also reported immunomodulating functions of MSCs/NCSCs which were of significant importance in graft-versus-host disease cases [56], and reviewed in details by Uccelli et al. [57], Prockop and Oh [58], and Singer and Caplan [59].

In the past few years, several papers also described more precisely the immunomodulative properties of MSCs/NCSCs in the particular case of SCI (Table 3). Globally, all those experimental studies reported a potent anti-inflammatory action of MSCs/NCSCs when transplanted in spinal cord-injured animals.

Table 3. Intermediate phase events: Immunomodulation, regulation of inflammation, antioxidative, and antiapoptotic abilities of MSCs/NCSCs
 Cell typeModel and modalitiesResultsRef
In vitror MAPCsNaiveCoculture (or CM treatment) with NR8383 MΦ and/or DRG dystrophic neurons (aggrecan gradient) inline imageDRG axonal dieback[60]
    inline imagearginase1+ MΦ (M2) 
    inline imageINOS+ MΦ (M1) 
    inline imageDRG neurons neurite outgrowth 
 h TGSCsNaiveCoculture with SH-SY5Y cells: inline imageCell viability[61]
   − H2O2 (300 µM) inline imageGPX, GST, CAT, SOD 
   − 6-OHDA (80 µM) inline imageAnnexin V+ cells 
   − Aβ (1–42) (20 µM) inline imageBax/Bcl2+ ratio 
     inline imageCasp3 expression 
 h AT-MSCsNaiveCM treatment of dermal fibroblasts: inline imageGPX, SOD[62]
   − tbOOH (up to 400 µM) for 6 hours  
 h AT-MSCsNaiveCoculture with NSC inline imageNSC viability[63]
   − H2O2 (500 µM) inline imageAnnexin V+ NSC 
   − Serum deprivation inline imageBax/Bcl2+ ratio 
   − Both  
   _ under normoxia (21% O2) or hypoxia (1% O2)  
 r/g BMSCsNaiveCM treatment on hippocampal/cortical neural cells: inline imageApoptotic nuclei (max effect with 40%CM)[64]
   − Staurosporine (300 nM) inline imageMAPK/Erk1,2 
   − Aβ (10 µM)= Confirmed with inhibitor 
     inline imageAKT 
     inline imagePI3K 
    = Confirmed with inhibitor 
 Cell typeCell stateHostLesionParametersResultsRef
  1. Abbreviations: 6-OHDA: 6-hydroxydopamine; Aβ: amyloid peptide β; AT-MSC: adipose tissue mesenchymal stem cells; Bad: Bcl-2-associated death promoter; Bak: Bcl-2 homologous antagonist/killer; Bax: B-cell lymphoma 2-associated X protein; BBB: Beattie Basso Bresnahan; Bcl-: B-cell lymphoma; Bid;-: BH3 interacting-domain death agonist; Bik: Bcl-2 interacting killer; BMSC: bone marrow stromal cells; C: cervical; Casp: caspase; CAT: catalase; CM: conditioned medium; CSF: cerebrospinal fluid; DC: dorsal column; DRG: dorsal root ganglion; eGFP: enhanced green fluorescent protein; ERK: extracellular signal-regulated kinase; FLIP: FLICE-inhibitory protein; g: Gerbil; GFAP: glial fibrillary acidic protein; GPX: glutathione peroxidase; GST: glutathione-s-transferase; h: human; Iba1: ionized calcium-binding adapter molecule 1; IGF: insulin-like growth factor; IL: interleukin; iNOS: inducible nitric oxide synthase; IV: intravenous; LIF: leukemia inhibitory factor; MΦ: macrophages; MAPC: multipotent adult progenitor cells; MAPK: mitogen-activated protein kinase; MDA: malondialdehyde; NGF: nerve growth factor; NSC: neural stem cells; p53: protein 53; PARP-1: poly(ADP-ribose) polymerase 1; PDGF: platelet-derived growth factor; r: rat; SC: spinal cord; SSEP: somato-sensory evoked potentials; SOD: superoxide dismutase; T: thoracic; tbOOH: tert-butyl hydroperoxide; TGSC: tooth germ stem cells; TNF: tumor necrosis factor; TNFR: TNF receptor; TUNEL: terminal deoxynucleotidyl transferase dUTP nick end labeling; UCB-MSC: umbilical cord blood mesenchymal stem cells; WM: white matter; XIAP: X-linked inhibitor of apoptosis protein.

In vivor MAPCsNaiveRatDC crush2 × 10s forceps, 1 mm depth200,000 cells inline imageAxonal dieback[65]
    C8Intraspinal (dorsal column) inline imageED1+ 
      –200,000 cells (0.5 mm caudal and 0.5 mm lateral) inline imageAxonal growth 
      Immediately after SCI  
 h UCB-MSCsNaiveRatSC contusion10g1 × 106 cells inline imageBBB score[66]
    T925 mm heightIV inline imageSSEP latency 
      Immediately after SCI inline imageIL1β, IL-6, 
        inline imageIL-10 
        inline imageAmeboid Iba1+ 
 r BMSCseGFP transductionRatSC contusion12.5 mm (mild) or 25 mm height (severe)300,000 cells inline imageSensitivity to mechanical stimuli[60]
    T11-T12Intraspinal inline imageGFAP+ astrocytes 
     –150,000 cells (epicenter) inline imageED1+ 
      –75,000 cells (1mm caudal) inline imageCavity size 
      –75,000 cells (1mm rostral) inline imageWM tissue sparing 
      7 days post-op inline imageBBB score (severe lesion) 
 h BMSCsPKH26RatSC contusion200 kDy1 × 106 cells inline imagearginase1+ and CD206+ MΦ (M2)[67]
    T9-T10 Intraspinal inline imageiNOS+ and CD16/32+ MΦ (M1) 
      –1 × 106 cells (epicenter) inline imageWM tissue sparing 
      3 days post-op  
 r BMSCsNaiveRatSC compression1 minute 26 g clip2 × 106 cells inline imageNGF, LIF, IGF-1, IL-1β[51]
    T8 CSF infusion inline imagePDGF-A 
 h AT-MSCsNaiveRatSC Compression10 minutes clip300,000 cells inline imageNSC survival[63]
    T9 Intraspinal  
      –300,000 cells (epicenter)  
      7 days post-op  
 rb BMSCsNeuroglobin transductionRabbitSC Compression2 minutes5 × 106 cells inline imageNeuroglobin over time[68]
   T9-T102 atm (balloon)Intraspinal inline imageBBB score 
      –5 × 106 cells (epicenter) inline imageMDA levels 
      24 hour post-op inline imageApoptotic index 
 r BMSCsNaiveRatSC contusion10 g, 12.5 mm height2.5 × 105 cells inline imageTUNEL+ cells[69]
    T10 Intraspinal inline imageCasp3+ cells 
      –2.5 × 105 cells (Epicenter)= In neurons but especially oligodendrocytes 
      7 days post-op inline imageCasp8+, Casp10+ cells 
        inline imageFasL 

Indeed, Seo et al. demonstrated that rats with contused spinal cord directly followed by UCB-MSCs intravenous (IV) injection, presented significant improvements in motor/sensory score. These improvements were correlated with a reduction of inflammatory events, as both a decrease in IL-1β and IL-6 expression and an increase in IL-10 expression were observed. Moreover, the number of activated macrophages was reduced in those conditions [66]. Those results were consistent with previous observations from Abrams et al., showing that BMSCs transplantation decreased astrocytic reactivity and microglial activation inside the lesioned spinal cord, associated with a reduced injury-induced response to mechanical stimuli [60].

Still, those experiments did not provide any clue on the real interaction of MSCs/NCSCs with host cells underlying these anti-inflammatory effects. Nakajima et al. showed that MSC/NCSCs transplantation into the epicenter of a spinal cord contusion was associated with a switch in host macrophage phenotype, from classically activated (M1) to alternatively activated (M2) [70] (which are known to have anti-inflammatory properties), and even promoted regrowth of sensory axons in a SCI model. This was also associated with an elevation of IL-4 and IL-13 levels of expression and reduction in IL-6 and tumor necrosis factor (TNF)-α levels, at the lesion site. All those events were correlated with a significant recovery of locomotor function in contused rats, as assessed by the Basso Beattie Bresnahan scoring which appraises motor function, especially based on walk evaluation [67]. Busch et al. previously reported this M1 to M2 switch, while studying the beneficial effects of multipotent adult progenitor cells or MAPCs (a very immature BMSCs subtype) [71] on DRG dystrophic neurons, in vitro. Besides, this study also described a valuable effect of MAPCs transplantation on macrophage activation after dorsal column crush lesion, associated with enhanced axonal growth [65]. At the opposite, Hawryluk et al. demonstrated that intrathecal injection of BMSCs induced an increase in IL-1β in the epicenter of a spinal cord compression [51] which rather reveals an induction of inflammation.

Antioxidative Actions of MSCs/NCSCs

Free radical formation, lipid peroxidation, and further oxidative damage events are generated during acute SCI and are tightly linked with mitochondrial dysfunctions, disruption in ion balance, glutamate-release, and associated excitotoxicity. Different therapeutical strategies based on antioxidant compounds are therefore currently tested in experimental animal models and in clinical trial procedures. Unfortunately, those drugs often lack in selectivity and trigger side effects (reviewed by Hall [72]). Very few detailed data about antioxidant potency of stem cells are recorded in SCI preclinical models. Still, several in vitro studies provided clues on the mechanisms involved in the protection against oxidative damage and reactive oxygen species (ROS) (Table 3).

First of all, BMSCs were reported to be able to efficiently manage in vitro-induced oxidative stress. Indeed, it was shown that cell viability was not altered by ROS [73], and that resistance to ROS was linked to glutathione availability. Therefore, it would be conceivable to use BMSCs to reduce oxidative stress-induced damages, in vivo. However, other studies about UCB-MSCs showed that these cells suffered from senescence and genomic alterations once undergoing oxidative stress [74].

In vitro tests showed that stem cells could protect different types of cells against oxidative stress. For instance, tooth-germ stem cells are able to protect SH-SY5Y cells (human neuroblastoma-derived neuronal cell line) from amyloid-beta (Aβ) peptide or 6-hydroxydopamine-induced cell death in culture, by increasing the activity of antioxidant proteins such as catalase, glutathione-peroxidase (GP), or superoxide dismutase (SOD) [61]. Similarly, AT-MSCs-CM helped dermal fibroblasts to resist to free radicals in culture by inducing an enhancement of their SOD and GP activity with a decrease of apoptotic cells [62]. In the same line of evidence, Oh et al. demonstrated that the coculture of neural stem cells (NSCs) with AT-MSCs protected NSCs against hydrogen peroxide and serum deprivation-insult, in both normoxia and hypoxia conditions (as showed by increased survival and decreased apoptosis). Likewise, they showed that the cotransplantation of NSCs and AT-MSCs enhanced the survival of grafted NSCs inside the compressed spinal cord of rats [63]. Finally, Lin et al.'s study showed that neuroglobin-expressing BMSCs grafting improved functionality after SCI in rabbits, and this was associated with a decrease in lipid peroxidation (as assessed by reduced malondialdehyde levels) and decreased apoptotic index [68].

Antiapoptotic Properties of MSCs/NCSCs

Another aspect that has been quite well-described in injured spinal cord concerns apoptotic events. Indeed, after SCI, post-traumatic tissue necrosis occurs but is followed by apoptotic events from myelinating oligodendrocytes, microglial cells, and neurons [75]. Apoptosis pathways subsequent to SCI imply either Fas-dependent and/or TNF-α signalization, followed by activation of caspase cascade or JNK pathway [76, 77].

As described above, different studies showed that MSCs/NCSCs treatment is able to lessen cell apoptosis, either in a SCI model or under unfavorable culture conditions. Dasari et al. characterized more accurately the diverse antiapoptotic actions of MSCs/NCSCs in the context of SCI (Table 3). They demonstrated the decrease of caspase-3 expression by oligodendrocytes and neurons, lowering the number of terminal deoxynucleotidyl dUTP nick end labeling (TUNEL)-positive cells (TUNEL—is a method to detect cells undergoing apoptosis) when an impacted spinal cord was transplanted with BMSCs. The expression of Fas was downregulated in neural/glial cells inside the lesion site, triggering the decrease of caspase-8, caspase-10, and caspase-3 activation [69]. Besides, BMSCs treatment upregulated the expression of FLICE-inhibitory protein and X-linked inhibitor of apoptosis protein (two inhibitors of apoptosis) and nuclear cleavage of poly(ADP-ribose) polymerase 1 was also reduced. They further completed those data by evidencing the activity of UCB-MSCs on TNF-α and NF-κB-mediated apoptosis pathway [78] and the implication of the upregulation of the phosphoinositide-3-kinase (PI3K)/Akt pathway in antiapoptotic activity of UCB-MSCs once grafted in SCI rats [79]. PI3K/Akt pathway was also activated in embryonic neurons by MSCs in vitro and was associated with a neuroprotective effect [64].

Chronic Phase Events: Demyelination of Axons and Glial Scar Formation

Promyelinating Abilities of MSCs/NCSCs

Different in vitro evidences established that MSCs are able to promote maturation of glial precursors into myelinating cells (Table 4). Indeed, BMSCs-CM primes oligodendroglial cell fate decision of proliferating neural precursors cells in vitro (assessed by A2B5, NG2, O4, galactocerebroside [GalC], and 2′,3′-cyclic-nucleotide 3′-phosphodiesterase staining) by modulating Id2/Olig2 expression [80]. This was previously observed in cocultures of BMSCs with NSCs [81] and in cotransplantation experiments with NSCs on hippocampal slices [82]. Interestingly, the same group compared BMSCs-CM to ciliary neurotrophic factor (CNTF) treatment, and showed that if CNTF only influences differentiation/maturation of neural precursor cells into oligodendrocytes, BMSCs-CM also instructs cell fate decisions and commitment [83].

Table 4. Chronic phase events: Promyelinating properties and matrix-degrading actions of MSCs/NCSCs for promoting axonal regrowth
 Cell typeCell stateModel and modalitiesResultsRef
In vitror BMSCNaiveCM treatment on NPC (neurospheres) inline imageCNPase, A2B5, NG2, O4, RIP[80]
 r BMSCNaiveCoculture with NSC or CM treatment on NSC inline imageMBP[81]
 r BMSCNaiveCotransplantation with NSC into hippocampal slices (350 µm) (5,000 total cells in CA and DG) inline imageGFAP inline imageId2[82]
   Or CM treatment on NSC before transplantation into slices. inline imageOlig2 
 r BMSCNaiveCM treatment on NPCMBP[83]
     inline imageGFAP 
    = Even w/o CNTF in the CM 
 h LP-SCNaiveCoculture (or CM treatment) with:LP-MSC > BMSC[84]
 h BMSC − OPC inline imageCaspr 
   − OEC inline imageO4 
   − OEC and Spinal cord axons inline imageNPC and OEC proliferation and process extension 
 Cell typeCell stateHostModel ParametersResultsRef
  1. Abbreviations: 5-HT: 5-hydroxytryptamine (serotonin); APC: adenomatous popyposis coli; BBB: Beattie Basso Bresnahan; BMS: Basso mouse scale (BBB equivalent for mice); BMSC: bone marrow stromal cells; CA: Cornu ammonis; Caspr: contactin-associated protein; CM: conditioned medium; CNPase: 2′,3′-cyclic-nucleotide 3′-phosphodiesterase; DC: dorsal column; DG: dentate gyrus; DRG: dorsal root ganglion; EtBr: ethidium bromide; GAP43: growth-associated protein 43; GFAP: glial fibrillary acidic protein; h: human; Id2: inhibitor of DNA-binding protein 2; LP-MSCs: lamina propria mesenchymal stem cells; MBP: myelin basic protein; MMP: matrix metalloproteinase; NF: neurofilament; NPC: neural precursor cells; NSC: neural stem cells; NT-3: neurotrophin 3; OEC: olfactory ensheathing cells; Olig2: oligodendrocyte transcription factor 2; OPC: oligodendrocyte precursor cells; P0: protein 0; r: rat; RIP: receptor interaction protein; SC: spinal cord; SCEP: spinal cord evoked potentials; SKP: skin-derived precursors; T: thoracic; TH: tyrosine hydroxylase; TIMP: tissue inhibitor of metalloproteinases; UCB-MSC: umbilical cord blood mesenchymal stem cells; ≠: predifferentiated.

In vivor BMSCNT-3 transductionRatEtBr-demyelination 1 × 105 cellsinline imageMBP around NF+ axons[85]
    DC Intraspinalinline imageNG2, APC 
    T10 − Lesion siteinline imageBeam walking score 
      3 days post-opinline imageLatency of SCEP 
       = MSC also take part to myelination themselves (MBP, association with axons) 
 h UCB-MSCNaïveRatSC contusion25 mm height3 × 105 cellsinline imageNG2[86]
    T9 Intraspinalinline imageCNPase 
      − 3 × 105 cells (Epicenter)inline imageMBP 
      7 days post-opinline imageGFAP 
       inline imageNestin 
       inline imageBBB score 
 h SKP≠ Schwann cellsRatSC contusion15 mm height8 × 105 cellsh SKPSchwann > h SKP[87]
    T9280–290 kDyIntraspinalinline imageSpared tissue 
      − 8 × 105 cells (Epicenter)inline image5-HT, TH axons 
      7 days post-opinline imageCaspr 
       inline imageP0 sheaths 
       inline imageBBB score 
       inline imageBladder run walking 
       = MSC also take part to myelination themselves (PO, Caspr, association with axons) 
 h BMSC≠ Schwann cellsRatSC contusion10g, 25 mm height2 × 106 cells in Matrigelh BMSCSchwann > h BMSC[88]
    T9 Intraspinalinline imageGAP43, 5-HT, TH fibers 
      − 2 × 106 cells (Epicenter)inline imageMyelin sheats 
      7 days post-opinline imageCavity size 
       inline imageBBB score 
 m BMSCCoculture withMouseSC compression8.3 g clip3 × 104 cellsm BMSC+Schwann cells > m BMSC[89]
  Schwann cells (1:1 ratio, for 1–10 days) T4 Intraspinalinline imageBMS score 
     − 3 × 104 cells (Epicenter)inline imageWM sparing 
     7 days post-opinline image5-HT axons 
       inline imageSchwann cell infiltration 
 h UCB-MSCNaiveRatSC contusion10g, 12.5 mm height2.5 × 105 cellsinline imageMMP2[90]
    T10 Intraspinalinline imageTIMP2 
      − 2.5 × 105 cells (Epicenter)inline imageGFAP 
      8 days post-op  

Lamina propria stem cells (LP-SCs) were identified to present typical BMSCs features, but LP-SCs or LP-SCs-CM are more efficient (compared to BMSCs of BMSCs-CM) to promote in vitro myelination of dissociated spinal cord neurons [84]. Additionally, LP-SCs and LP-SCs-CM enhanced proliferation and process extension of oligodendrocyte progenitor cells and olfactory ensheathing cells.

Remyelination abilities of MSCs were further described in in vivo experimental models (Table 4). Neurotrophin-3 (NT3)-overexpressing BMSCs were injected in a rat model of ethidium bromide-induced spinal cord demyelination [85]. The cell implantation was coupled to strong myelin-basic protein expression in or around the demyelinated area. NT3-BMSCs seemed to participate to remyelination, but more specifically promoted the enrollment of endogenous myelinating oligodendrocytes, hence improving significantly locomotor function. Park et al. also described enhanced oligodendrogliogenesis inside the contused spinal cord of rats, once they were treated with UCB-MSCs [86].

Another strategy consists into differentiating MSCs/NCSCs into myelinating cells just before transplanting them inside the lesioned spinal cord. For instance, neuroprotection and motor improvements were much more drastic using skin-derived precursor (SKPs) or BMSCs previously differentiated into Schwann cells than non-predifferentiated cells [87, 88]. Indeed, results revealed that both cell types reduced the size of the contusion cavity, myelinated host axons, and recruited endogenous Schwann cells. SKPs-derived Schwann cells also provided a bridge across the lesion site, increased the extent of spared tissue, promoted myelination of spared axons, reduced gliosis, and provided an environment that highly favors the axonal growth. Finally, SKPs/BMSCs-derived Schwann cells provided enhanced locomotor recovery relative to native cells. In the same way, cocultivating BMSCs with Schwann cells improves their therapeutic effects when grafted in spinal cord-injured mice [89].

Degradation of Extracellular Matrix by MSCs/NCSCs

Matrix metalloproteinases (MMPs) are endopeptidases that degrade extracellular matrix proteins, proteinases, and membrane receptors, and play a dual role in cases of SCI. After SCI, MMPs contribute to secondary pathogenesis by promoting inflammation, inducing blood-brain-barrier disruption, degrading myelin, and so forth. Conversely, it has been shown that they are able to regulate axonal guidance and regrowth by remodeling the extracellular matrix [91, 92].

UCB-MSCs upregulate the expression of MMP-2 inside a lesion, 3 weeks following rat spinal cord contusion [90] (Table 4). The activity of MMP2 was increased in the treated spinal cords and this was coupled to a reduction in the glial scarring, confirming the already-reported beneficial effect of MMP-2 on astrocyte reactivity, and chondroitin sulfate proteoglycans accumulation after SCI [93].

Proangiogenic Abilities of MSCs/NCSCs

Managing blood supply and revascularization is mandatory for the repair of hypoxic and inflamed environment of spinal cord lesioned tissue. The secretion of proangiogenic factors such as vascular endothelial growth factor (VEGF) by MSCs/NCSCs could help in this vascular remodeling (Table 5). New blood vessels and enhanced blood supply could provide valuable help in lesion recovery whatever the acute, intermediate, or chronic timing.

Table 5. Proangiogenic abilities of MSCs/NCSCs
 Cell typeCell stateHostModel ParametersResultsRef
  1. Abbreviations: 5-HT: 5-hydroxytryptamine (serotonin); AT-MSC: adipose tissue mesenchymal stem cells; BBB: Beattie Basso Bresnahan; BDNF: brain-derived neurotrophic factor; BMSC: bone marrow stromal cells; CST: corticospinal tract; DC: dorsal column; FGF2: fibroblast growth factor 2; h: human; HIF: hypoxia-inducible factor; IL: interleukin; IV: intravenous; MΦ: macrophages; MBP: maltose-binding protein; NGF: nerve growth factor; NF: neurofilament; r: rat; RECA-1: rat endothelial cell antigen 1; RST: rubrospinal tract; SC: spinal cord; SMA: smooth muscle actin; T: thoracic; TNF: tumor necrosis factors; VEGF: vascular endothelial growth factor; vWF: von Willebrand factor.

In vivor BMSCBrdURatSCCompression T9-T1015 minute 15 µl volume balloon1×105 cells IV 7 days post-op inline image BBB score, Grid navigation inline image NGF expression inline image RECA-1+ blood vessels[50]
 h AT-MSCh BMSCNaiveRatSC section (CST, RST, DC) T9To depth ofcentral canal2 ×105 Intraspinal –1×105 cells (2 mm rostral) –1×105 cells (2 mm caudal) Immediately after SCIAT-MSC > BMSC inline image vWF+ blood vessels inline image NF200+ fibers inline image 5-HT+ fibers inline imageCavity size inline imageEDI+ MΦ inline image BDNF inline image BBB score[[48]]
 h AT-MSC3D cluster on MBP-FGF2 surfaceRatSC compression T910 minute, dip10.3×105 cells Intraspinal –2.5 ×105 cells (Epicenter) Immediately after SCIhAT-MSC3D > hAT -MSC inline image vWF inline imageCD31 inline imageSMA[94]
 r BMSC3D cluster in gelatin-spongeRatSC section T10 1.5 mm segment removal 1.5 mm gelatin scaffold containing MSC Immediately after segment removal. inline imagevWF close relation between BMSC and blood vessels inline imageVEGF, HIF-lα inline imageCavity sixe inline image CD68+ MΦ inline imageIL-lβ, TNFα[95]

The study of Quertainmont et al. revealed a higher number of blood vessels stained for rat endothelial cell antigen 1 in the epicenter of spinal cord lesions after BMSCs graft, even if VEGF was not overexpressed in MSC-treated spinal cord extracts [50]. The same observation was reported after AT-MSCs transplantation inside acutely injured spinal cord, as assessed by the higher number of von Willebrand Factor (vWF)-positive blood vessels (which was superior compared to BMSC transplantation conditions) [48].

Oh et al. showed that AT-MSCs enhanced vasculogenesis after being transplanted in the lesion-epicenter of spinal cord-injured rats [94]. Moreover, the vWF, CD31, or smooth muscle actin-positive areas were especially higher when the AT-MSCs were transplanted as a three-dimensional (3D)-cell mass (inserted in maltose-binding protein-fibroblast-growth factor 2 surface).

The same concept of 3D-cluster was used by Zeng et al., who removed a segment of the spinal cord of rats and replaced it by a gelatin sponge scaffold containing BMSC. They observed that regenerated blood vessels crossed the junction between host spinal cord tissue and scaffold. vWF-positive area was also higher in those conditions compared with a scaffold without cells. They also showed that MSC-surrounding blood vessels expressed HIF-1α and VEGF [95].


Traumatic SCIs represent a critical issue in clinical situation nowadays, and injured patients undergo severe physiological, psychological, and social failures. Even if slight recovery can be achieved with intensive training programs and treatments, this recuperation is slow and limited, and depends on the degree of injury.

Cell-based therapies could consequently provide additional expectations in the case of spinal cord lesions. Recent remarkable papers reviewed the interest of stem cell therapy in SCIs and other neurological diseases, from experimental models to clinical application (iPCs and ESCs [42, 96, 97]; MSCs [98-100]; comparison of diverse types of stem cells [101-103]).

However, after considering all those papers that were somehow informative about several aspects of stem cell therapies, we wanted to dig deeper about the precise mechanisms of stem cell-associated positive effects, and more particularly stromal stem cells. Indeed, as initially mentioned, we are convinced that MSCs and NCSCs, both isolated from adult bone marrow could constitute a great tool for stem cell therapy and we will discuss their advantages thereafter.

In the past few years, several prospective clinical trials confirmed that AT-MSC or BMSC transplantation in spinal cord-injured patients did not induce tumors or any other side effects, but promoted preliminary motor and somatosensory improvements, which still have to be confirmed with further investigations in randomized trials, at a larger scale [19, 104, 105]. Nonetheless, it seems that those clinical analyses were somewhat prematurely executed. Indeed, the exact(s) way(s) by which those stem cells exert their beneficial effects are not fully understood and still need to be characterized.

All experimental data described in this review summarized the diverse properties of adult mesenchymal or NCSCs that could be relevant in spinal cord injury context. Still, is has to be highlighted that (Fig. 2): (a) BMSCs could modulate immune response and inflammatory reaction after SCI by adjusting cytokine secretion and macrophage activation, resulting in a potent anti-inflammatory effect for the most part. This inflammatory modulation has been observed also for AT-MSCs and UCB-MSCs suggesting that these effects should be attributed to the presence of MSCs, rather than NCSCs, as no NCSC has ever been identified in AT or UCB so far. (b) BMSCs also protect cells against oxidative damage by increasing the expression of antioxidant proteins such as glutathione-associated enzymes, catalase, and SOD. All those neuroprotective activities are obviously associated with a decrease in apoptotic events, which appear to be mediated through different pathways (Caspases, PI3K, Akt, etc.), and well regulated by stem cell graft. (c) Besides inducing this newly favorable lesion environment, MSCs are also able to promote axonal sparing and sprouting of new neurites, by means of neurotrophic support and efficient remyelination of fibers. The formation of new myelin sheaths can be both due to recruited endogenous glial cells and to transplanted BMSCq themselves. Interestingly, this last characteristic has only been demonstrated for BMSC or SKPs. As both populations contain NCSCs, it is tempting to suggest that this last property could be attributed to those cells.

Figure 2.

Adult mesenchymal stem cells/neural crest stem cells properties and the different ways they can contribute to functional recovery after spinal cord injury. Abbreviations: Bak: Bcl-2 homologous antagonist/killer; Bax: B-cell lymphoma 2-associated X protein; Bcl: B-cell lymphoma; Bid: BH3 interacting-domain death agonist; BDNF: brain-derived neurotrophic factor; Casp: Caspase; CAT: Catalase; FLIP: FLICE-inhibitory protein; GPX: Gluthatione peroxidase; GST: Gluthathione-s-transferase; HIF: Hypoxia-inducible factor; IGF-1: insulin-like growth factor-1; IL: Interleukin; MΦ: Macrophages; MMP2: Matrix metalloproteinase 2; NGF, nerve growth factor; NMDAR: N-methyl-D-aspartate receptor; PDGF: platelet-derived growth factor; SOD: Superoxide dismutase; TNF: Tumor necrosis factor; TNFR: TNF receptor; VEGF: vascular endothelial growth factor; XIAP: X-linked inhibitor of apoptosis protein.

To the light of those observations, it appears that MSCs and NCSCs could be used concomitantly: MSCs could be injected in acute phase to modulate inflammation, whereas NCSCs could then be injected in chronic phase to improve neuronal recovery and axonal sprouting. Adult MSCs/NCSCs present enormous interest regarding SCI therapy and could efficiently regulate a multitude of phenomena occurring at the time of a lesion. Moreover, it is interesting to highlight that both cell types could be isolated from the same patient, limiting the risk of immune reaction in case of heterologous graft.

Nonetheless, we think that this wide-ranging beneficial effect of MSCs/NCSCs lacks in systematic and mechanistic explanations. Although the different molecules that are secreted by MSCs were largely studied and identified in vitro (e.g., in CM analyses) [106], details are missing about the genuine mechanisms by which the physiopathological and clinical recovery of experimental SCI animals are achieved.

A common criticism concerning BMSCs is the lack of reproducibility. Therefore, several points have to be taken into consideration: (a) donor age: it appears that for human above 30 years old, MSCs have a drastically decreased secreting factors and stemness characteristics [107, 108]. It is therefore suggested to use MSCs from young donors, excluding autologous graft for oldest patients. (b) MSC and NCSC isolation protocols: since NCSCs are systematically mistaken for MSCs, new specific panels of MSC or NCSC markers should be defined. Indeed, all CD markers classically defined by the International Society for Cellular Therapy to characterize clinical grade MSCs [109] are not specific enough to discriminate NCSCs. Similarly, technical questions remain regarding the design of MSCs/NCSCs transplantation and its procedural parameters. (c) Technical questions remain regarding the design of MSCs/NCSCs transplantation and its procedural parameters. (a) The finest way to implant cells still needs to be defined. It was recently published that intracisternal transplantation of BMSCs led to the best motor recovery in rats, followed by intralesional and IV administration [110], while another study described the same functional improvement after intralesional or IV cell injection [111]. (b) The impact of immunosuppressive treatment has to be evaluated, in order to allow or prevent allografts. (c) The right timing of injection should be identified, as well as the importance of SCI kinetics on the transplantation time. As suggested above, injection of MSCs in acute phase, then NCSCs in chronic phase might be the right protocol. (d) Finally, the maturation state of transplanted stem cells should be characterized. As several studies use stem cells under their initial state; other papers have used predifferentiated cells in SCI models. Both neuron-differentiated and glia-differentiated cells have shown valuable benefits in experimental animals. Questions also remain concerning cell culture conditions before grafting: confluence, number of passages, and so forth. All those technical issues should be deeply studied in order to answer these questions and develop better strategies for cell therapy protocols.

In conclusion, as recently suggested by Vawda and Fehlings [112], the explosion in the number of studies examining the secretomes of stromal cells as well as the number of submitted patent reflects a recognition by the scientific communities of the vastly untapped potential of MSCs/NCSCs; however, evidence on mechanistic organization and technical procedures should be provided in order to ensure efficient and accurate cell-therapy protocols in SCI patients.


This work was supported by a grant from the Fonds National de la Recherche Scientifique (FNRS) of Belgium, by the Belgian League against Multiple Sclerosis associated with the Leon Frédericq Foundation, and by the Fonds Spéciaux à la Recherche of the University of Liège.

Authors Contributions

V.N.: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing, and final approval of manuscript; D.C., C.C., and R.F.: data analysis and interpretation and final approval of manuscript; B.R.: conception and design, final approval of manuscript, financial support, and administrative support; S.W.: conception and design, manuscript writing, final approval of manuscript, financial support, and administrative support.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.