Activated Spinal Cord Ependymal Stem Cells Rescue Neurological Function§


  • Author contributions: V.M.-M., F.J.R.-J., and M.S.: conceived and designed the experiments; V.M.-M., F.J.R.-J., M.G.-R., S.L., S.E., M.T.C., M.R., and M.L.: performed the experiments; V.M.-M., F.J.R.-J., R.P.-C., J.M.S.-P., and M.S.: analyzed the data; V.M.-M., F.J.R.-J., R.P.-C., J.M.S.-P., and M.S.: wrote the article.

  • Disclosure of potential conflicts of interest is found at the end of this article.

  • §

    First published online in STEM CELLSExpress January 28, 2009.


Spinal cord injury (SCI) is a major cause of paralysis. Currently, there are no effective therapies to reverse this disabling condition. The presence of ependymal stem/progenitor cells (epSPCs) in the adult spinal cord suggests that endogenous stem cell-associated mechanisms might be exploited to repair spinal cord lesions. epSPC cells that proliferate after SCI are recruited by the injured zone, and can be modulated by innate and adaptive immune responses. Here we demonstrate that when epSPCs are cultured from rats with a SCI (ependymal stem/progenitor cells injury [epSPCi]), these cells proliferate 10 times faster in vitro than epSPC derived from control animals and display enhanced self renewal. Genetic profile analysis revealed an important influence of inflammation on signaling pathways in epSPCi after injury, including the upregulation of Jak/Stat and mitogen activated protein kinase pathways. Although neurospheres derived from either epSPCs or epSPCi differentiated efficiently to oligodendrocites and functional spinal motoneurons, a better yield of differentiated cells was consistently obtained from epSPCi cultures. Acute transplantation of undifferentiated epSPCi or the resulting oligodendrocyte precursor cells into a rat model of severe spinal cord contusion produced a significant recovery of motor activity 1 week after injury. These transplanted cells migrated long distances from the rostral and caudal regions of the transplant to the neurofilament-labeled axons in and around the lesion zone. Our findings demonstrate that modulation of endogenous epSPCs represents a viable cell-based strategy for restoring neuronal dysfunction in patients with spinal cord damage. STEM CELLS2009;27:733–743


Ependymal cells, lining the central canal, rapidly proliferate, migrate, and differentiate to regenerate the cord in lower vertebrates on spinal cord injury (SCI) [1–3]. In mammals, including humans, proliferation of ependymal cells and their progeny is a frequent event during embryonic and early postnatal periods of development. However, the turnover of spinal cord ependymal derived stem/progenitor cells (epSPCs) declines significantly in the postnatal period [4, 5]. Nevertheless, extensive epSPC proliferation has been observed in response to several types of trauma, including compression [6–8], contusion injury [9], dorsal funiculus incision [10], transection [11], or crushed Filum Terminale [12]. After contusion, nearly two million new cells are produced at the injury site within a month, peaking at 3-7 days after injury [reviewed in13]. Endogenous epSPCs respond to intrathecally delivered epidermal growth factor (EGF) and fibroblast growth factor two (FGF-2) by increasing cell proliferation, resulting in fewer cysts and more white matter at the lesion epicenter after SCI [14]. Accumulating evidence suggests that the inflammatory response induced after spinal cord trauma strongly contributes to the endogenous epSPC activation by inducing cell proliferation and differentiation [15]. Implantation of activated macrophages, induction of autoimmune T-cells, or transplantation of dendritic cells results in functional recovery of the spinal cord [15–17]. This beneficial effect has been attributed to the dialogue between T-cells and endogenous microglia, including the secretion of cytokines which protect neurons and induce the differentiation of stem/progenitor cells to neurons and (OPC) oligodendrocytes [18]. Moreover, the activation of the JAK/STAT and mitogen-activated protein kinase (MAPK) signaling pathways in vivo induces proliferation of spinal cord-derived stem/progenitor cells [19]. In contusive injuries, significant compressive force is applied to the spinal cord, producing an initial trauma associated with oligodendrocyte, neuronal, and precursor cell death. Ischemia-associated damage contributes to the progressive degeneration of neurons and axons which, in turn, leads to the formation of astrocytic scars [20]. Transplantation of undifferentiated stem cells has revealed that either these cells fail to differentiate in vivo or that differentiation yields a predominant astrocytic lineage which contributes to the glial scar [21]. Therefore, the main goal of therapeutic strategies aimed at treating SCI is to alter the fate of stem cells by restricting differentiation to the generation of oligodendrocytes or motoneurons to specifically replace the loss of functional units [22, 23]. Previous studies have demonstrated that transplantation of adult spinal cord-derived neurospheres from healthy donors slightly improved motor recovery but with aberrant axonal sprouting associated with allodynia [24]. Parr et al. have recently shown a regenerative potential but only when epSPCs were transplanted 1 week after spinal cord compression [25]; otherwise, the survival rate of the transplanted cells was very poor [26]. Here, we demonstrate that acute transplantation of epSPCs reverses the paralysis associated with spinal cord contusion in rats. Recovery of function was more pronounced in animals transplanted with ependymal stem/progenitor cells injury (epSPCi) than with epSPCs. Our studies revealed that cultures of epSPCi display a higher rate of proliferation, more robust self-renewal, and better yields of oligodendrocytes and motoneurons when subjected to directed differentiation. Genetic profile analysis of epSPCi revealed an important influence of inflammation in epSPC-mediated signaling after injury. Although the present findings clearly demonstrate the potential of epSPCs to promote recovery from spinal cord damage, further investigation is needed to elucidate how to amplify and modulate the endogenous regenerative machinery to efficiently rescue neurological dysfunction.


Spinal Cord Contusion and Posttraumatic Cares

Female Sprague Dawley rats (∼200 g) were divided into two groups: (a) Control: SCI by severe contusion (n = 14); (b) epSPCi: SCI by severe contusion followed by transplantation of epSPCi (n = 13); (c) OPCi: SCI by severe contusion followed by transplantation of oligodendrocyte precursor cells (OPC) (n = 6); and (d) Sham: laminectomy with no lesion (n = 6). The maintenance and use of all animals was in accordance with the National Guide for the Care and Use of Experimental Animals (Real Decreto 1201/2005 of the Ministerio de Presidencia and the Animal Care Committee of the Research Institute Príncipe Felipe.

Surgery Procedure

Rats were premedicated with morphine (2.5 mg/kg) (Braun Medical Inc., Bethlehem, PA, and Baytril (enrofloxacine, 5 mg/kg, Bayer, Leverkusen, Germany, and subsequently, were anesthetized with isofluorane (Braun Medical Inc.). A laminectomy at the T8-T9 level was performed to induce a severe traumatic lesion by applying 250 kilodyne using the “Infinitive Horizon Impactor” (IHI; Precision Systems, Kentucky, IL, Records of the impact magnitudes and tissue displacement indicated no significant differences in impact parameters between groups that might account for behavioral differences. The epSPCi and the OPCi for transplantation were obtained from homozygote transgenic rats (SD-eGFP+/+ (SD-Tg(GFP)2BalRrrc) rats), with constitutive expression of eGFP [27] 1 week after spinal cord contusion. A total of 106 cells were transplanted by stereotaxis at a distance of 2-mm rostral and caudal to the lesion in four positions, from anterior to posterior.

Postsurgery Care of Injured Rats

Bladders were drained manually three times each day until autonomous functioning was restored. All groups were subjected to rehabilitation procedures consisting initially of passive mobilization through a full range of movement to maintain joint flexibility and reflexes in the hind limbs for 15 minutes every day. Subsequently, animals were subjected to active rehabilitation by forcing limb movement on a treadmill at constant speed (8-15 cm/seconds), plus 5 minutes of exercise in a swimming pool.

Behavioral Tests

Basso, Beattie, and Bresnahan Scale.

Before surgery, the animals were trained for 2 weeks and baseline measurements were recorded. At 7 days after surgery, the testing sessions were performed in weekly intervals for up to 60 days after surgery. Open-field locomotion was evaluated by using the 21-point Basso, Beattie, and Bresnahan (BBB) locomotion scale [28]. In each testing session, the animals were individually videotaped for 4 minutes which were scored blind by two unbiased observers.

Narrow Beam Crossing Test.

Animals were forced to walk on a 1-m long horizontal metal beam elevated 30 cm from the ground. Analysis was performed by identifying the animals that were able to walk through the beam.

Foot Print Analysis.

This analysis was done using CatWalk (Noldus, Asheville, NC, video-based system for automated gait analysis. Paw contact was quantified by counting high-intensity pixels as the mean of at least three rounds per analysis [29].

Morphological Analysis

Two months after surgery, the animals were transcardially perfused with a 0.9% saline solution followed by 4% paraformaldehyde (PFA) in phosphate buffered saline (PBS), and 2 days incubation time on 30% sucrose before Tissue-Teck OCT blocks (Miles, Elkhart, IN, preparation. Sagittal and coronal 20 μm criosections were performed for immunoassays. Every seventh section was collected for phosphotungstic acid hematoxylin (PAH) staining to determine the cavitation/cystic area. The area of cyst/cavities (pixels) was quantified from coronal sections using the Metamorph Offline software (Molecular Devices Corp., Union City, CA,

Spinal Cord Tissue Isolation and Spinal Cord Stem/Progenitor Cell Culture and Differentiation

epSPCs were harvested from female Sprague Dawley adult transgenic SD-eGFP+/+ rats (∼200 g) for cell-transplantation experiments, and SD-eGFP−/− were used for in vitro assays. epSPCs were obtained from a noninjured rat, and epSPCi were isolated 1 week after severe contusion of spinal cord (250 kilodyne at T8-T9). Once the overlying meninges and blood vessels were removed, the dissected tissue was cut into 1-mm3 pieces and homogenized without enzymatic treatment. epSPCs were cultured as neurospheres with complete medium (CM) as previously described [30].

Differentiation of Oligodendrocyte Precursor Cells

Differentiation was performed as previously described by Keirstead et al. [22]. Briefly, epSPCs were cultured with glial restriction media (GRM) which consists of DMEM:F12, B27 supplement (Invitrogen, Carlsbad, CA,, 25 μg/ml insulin, 6.3 ng/ml progesterone, 10 μg/ml putrescin, 50 ng/ml sodium selenite, 50 μg/ml holotransferin, 40 ng/ml tri-iodo-thyroidin, and supplemented with 4 ng/ml basic fibroblast growth factor (bFGF) and 10 ng/ml EGF (Sigma, St. Louis, MO, for 1 day. Subsequently, cells were incubated with 20 ng/ml EGF and 10 μM of all-trans retinoic acid (ATRA) for 1 week. ATRA was then removed and the cells were exposed to GRM supplemented with 20 ng/ml EGF for 25 days. At day 28, the spheres were plated in Petri dishes coated with 1:30 Matrigel for 1 week and cultured on GRM supplemented with 20 ng/ml EGF. For terminal differentiation, at day 35, OPCs were seeded on poly-L-lysine and human laminin (Sigma) coated slides. At day 0, 3, 7, 28, and 35, the cells were harvested for total RNA or immunocytochemical staining.

Differentiation of Motoneurons

epSPCs were grown in MIM medium: DMEM/F12, N2 supplement, heparin 2 μg/ml, and penicillin/streptomycin supplemented with bFGF (20 ng/ml) and ATRA (0.1 μM) for 1 week. Cells were then maintained in ornithine/laminin chamber slides with MM medium (Neurobasal medium, N2 supplement, camp 1 μM, ATRA 0.1 μM, and penicillin/streptomycin) supplemented with 200 ng/ml of recombinant human (rh) SHH and cultured for an additional week. The final differentiation medium included 10 ng/ml of rhBDNF, rhGDNF, rhIGF-I, and 50 ng/ml of rhSHH in which they were maintained for up to 2 weeks.


All cell recordings were obtained with an EPC-10 amplifier (HEKA, Lambrecht, Germany, The holding potential (Vh) was always set at −70 mV. The bath solution consisted of 140 mM NaCl; 4 mM KCl; 2 mM CaCl2; 10 mM HEPES; 5 mM Glucose; 20 mM Mannitol, pH 7.4. The electrode solution contained 144 mM KCl; 2 mM MgCl2; 10 mM HEPES; 5 mM EGTA, pH 7.2. The pipette resistance was 1.5-2 MΩ. Series resistance was compensated for (>70%) and leak subtraction performed. Data were low pass-filtered at 3.3 kHz and sampled at 10 kHz. Applied pulse protocols and analysis were carried out by Pulse software (HEKA). Action potentials were evoked by applying a variable pulse of current (20-100 pA) to neurons in current-clamp configuration. In voltage clamp configuration, squared pulse depolarization voltage steps were applied from −30 mV or to +60 mV from the holding potential of −70 mV. Tetrodotoxin (TTX; Tocris Laboratories, Bristol, U.K., was used to block TTX-sensitive voltage-dependent Na+ channels and tetraethylammonium (TEA) blocked outward K+ currents. Recordings were obtained from three different experiments with epSPCs and four experiments with epSPCi. Data are represented as the mean ± SD.

DNA Microarray Analysis

epSPCs and epSPCi were obtained from four different rats. The quality of all RNA was confirmed by RNA 6000 Nano Bioanalyzer (Agilent Technologies, Palo Alto, CA, assay. A total RNA of 500 ng were used to produce Cyanine 3-CTP-labeled cRNA using the Low RNA Input Linear Amplification Kit, PLUS, One-Color (Agilent p/n 5188-5339) according to the manufacturer's instructions. Following “One-Color Microarray-Based Gene Expression Analysis” protocol Version 5.5 (Agilent p/n 5188-5977), 3 μg of labeled cRNA was hybridized with Whole Rat Genome Oligo Microarray Kit (Agilent p/n G2519F-014879) containing 41,000+ rat genes and transcripts. Arrays were scanned in an Agilent Microarray Scanner (Agilent G2565BA) according to the manufacturer's protocol and data extracted using Agilent Feature Extraction Software 9.5.3 following the Agilent protocol GE1-v5_95_Feb07 and the QC Metric Set GE1_QCM_Feb07.

The gene profile was sorted by differential expression levels between the two experimental conditions (epSPCi vs. epSPCs) and clustered into biological functional profiling using FatiGO application [31]. FatiGO converts the two lists of genes into two lists of GO terms using the corresponding gene-GO association table. Subsequently, a Fisher's exact test for 2 × 2 contingency tables is used to check for significant overrepresentation of GO terms in one of the sets with respect to the other one. Multiple test correction to account for the multiple hypothesis tested (one for each GO term) was applied as previously described [31].

Proliferation/Cell Viability Assay

CellTiter-Glo luminiscent cell viability kit (Promega, Madison, WI, was used to assess proliferation. Three × 105 viable cells were plated in quadruplicate on low-attachment 96-well plates, and cells were allowed to growth for 3 days. For the assay, cells were transferred into opaque 96-well plates containing the same volume of medium (100 μl per well) of substrate mix (containing Beetle Luciferin). The luminescence was measured by a Multilabel Counter VICTOR [3] (Perkin Elmer, Waltham, MA,

Fluorescence-Activated Cell Sorting, Proliferation, and Apoptosis Analysis

The DNA content was assessed by staining ethanol-fixed cells with propidium iodide and monitoring by FACSCaliber (Becton Dickinson, Franklin Lakes, NJ, The percentages of cells in the G0/apoptosis or S-phase were determined by using ModFIT software (Verity Software House, Topsham, ME,

Immunocytochemistry and Immunohystochemistry

Cells or tissue sections were fixed with 4% PFA at room temperature for 10 minutes, permeabilized with 0.1% Triton X-100, and subsequently blocked with 1% fetal bovine serum in PBS. The primary antibodies were incubated overnight (1:200) at 4°C: mouse oligodendrocyte cell marker (2′,3′-cyclic nucleotide 3′-phosphodiesterase)-non-compact myelin protein (RIP), A2B5, MAP2 or rabbit NG2, glial fibrillary acidic protein (GFAP), HB9 (Millipore, Billerica, MA,, Sox2 or goat Oct4 antibody (Abcam, Cambridge, U.K., After washing, secondary antibodies (Texas Red conjugated goat anti-rabbit IgG, 1:400; Jackson ImmunoResearch Laboratories, Newmarket, U.K.,; Oregon Green 488 goat anti-mouse IgG, or Alexa Fluor 633 mouse anti-goat 1:400; Invitrogen) were incubated for 1 hour at room temperature. All cells were counterstained by incubation with 4,6-diamidino-2-phenylindole dihydrochloride (DAPI; Invitrogen). Signals were visualized by Confocal Microscopy (Leica, Wetzlar, Germany,

RNA Isolation and Quantitative RT-PCR

Total RNA was extracted by using Rneasy Mini-kit (Qiagen, Barcelona, Spain, according to manufacturer's instructions. One microgram of total RNA was reverse transcribed (RT) in a total reaction volume of 50 μl through incubation at 42°C for 30 minutes using random hexamer primers. mRNAs were amplified and quantified by using SYBR Green (Applied Biosystems, Foster City, CA, Each reaction was carried out in duplicate from six different animals for each group. The sequences of specific primers for quantitative PCR, designed as SG, are described in the supporting information Table 1. As template, 40 ng of cDNA from target and housekeeping gene (GAPDH) were prepared in separate tubes for each primer master mix reaction. The comparative threshold cycle (CT) method was used to calculate the relative expression [32].

Western Blot Analysis

Cells were collected and proteins extracted by using 2% SDS-TrisCl lysis buffer. Twenty micrograms of protein was loaded onto a 10% SDS-polyacrylamide gel and resolved by standard SDS-PAGE. Proteins were electrophoretically transferred onto PVDF membranes. Membranes were blocked with 5% nonfat dry milk in PBS for 60 minutes and incubated overnight with specific antibodies against Sox2, Oct4 (Abcam, U.K.), Phospho, or total STAT3 (New England Biolabs, Ipswich, MA, at 1:500 dilution. β-actin at dilution 1:5,000 (Sigma) was used as loading control. Subsequently, membranes were incubated with rabbit anti-mouse, rabbit anti-goat, or anti-rabbit horseradish peroxidase-conjugated secondary antibody (1:5,000; Sigma). Blots were visualized by the ECL (Amersham, Barcelona, Spain).

Statistical Analysis

Statistical comparisons were assessed by Student's t test. All p values are derived from a two-tailed statistical test using the SPSS 11.5 Software. A p-value of <.05 was considered statistically significant. The in vitro experiments were always represented as a mean of at least three different assays ± SD.


Characterization of Cell Proliferation and Self-Renewal Properties of epSPCs After SCI

In agreement with previous reports [33], a significant increase of stem/progenitor-forming neurospheres was observed in cultures of epSPC isolated from rats 1 week after spinal cord contusion (epSPCi) as compared with cultures obtained from noninjured control rats (epSPCs). This increase was detected after 10 days in culture but was much more apparent at day 30 (Fig. 1A). In addition, proliferation in vitro was more than 10 times higher in cultures of epSPCi as compared with epSPCs (Fig. 1B). The lower proliferative activity of epSPCs was accompanied by an increase in the percentage of apoptotic cells (18.8 ± 5 in epSPCs vs. 9.6 ± 4 in epSPCi, 10 days after cell culture, and 38 ± 6.7 in epSPCs vs. 10.4 ± 3.5 in epSPCi, 20 days after cell culture). The epSPCi also display enhanced self-renewal properties (Fig. 1C, 1D). Telomerase activity, indicative of the amplification of the telomeric repeat sequences, is associated with the stemness and self-renewal capability of cells [34]. Interestingly, telomerase activity was augmented in epSPCi as compared with epSPCs (Fig. 1C). Hence, we also analyzed the expression levels of both Sox2 and Oct4, factors which are critical for pluripotency, self-renewal, and the regulation of the differentiation [35]. As shown in Figure 1D, protein (upper panel) and mRNA levels (lower panel) of Sox2 and Oct4 were significantly upregulated in epSPCi.

Figure 1.

epSPCi displays a higher rate of proliferation and improved self-renewal. (A): Phase-contrast micrographs of epSPCs and epSPCi at 10 days after cell culture (Scale bar 100 μm). Quantification of neurospheres demonstrates a significant (p < .05) increase when epSPCs were isolated 1 week after spinal cord contusion (epSPCi) in comparison with noninjured spinal cord (epSPCs). (B): After the first 10 days of culture, the proliferative activity of epSPCi was significantly (p < .05) increased as assessed by measuring intracellular ATP. (C): At day 10 of culture, telomerase activity was determined in both epSPCs and epSPCi. Longer telomere repeats were obtained in epSPCi (lane 3) as compared with epSPCs (lane 2) after polymerase chain reaction (PCR) amplification. An extract of HeLa cells was used as a positive control of telomere amplification (lane 1) and epSPCi-heated cell extract at 80°C for 10 minutes was included as a negative control (lane 4). (D): Ten days after neurosphere culture, higher expression levels of Sox2 and Oct4 were found in epSPCi than in epSPCs from noninjured cords at both protein (upper part) and mRNA (lower part) levels. (E): Differential expression of cell fate markers. Cell markers of ependymal cells (Dlx2, PDGF-Rp, vimentin), neuroblasts (NCAM-1), astrocytes (GFAP), or oligodendrocyte precursor cells (Olig2, Nkx2.2, NG2) were assayed by PCR for both cell cultures epSPCi and epSPCs. All experiments were performed from cell cultures derived from three different animals. Statistical comparisons were performed with Student's t test (∗compared with epSPCs, p < .05). Abbreviations: epSPCi, ependymal stem/progenitor cells injury; epSPCs, ependymal stem/progenitor cells; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; GFAP, glial fibrillary acidic protein; HeLa, Henrietta Lacks cervical cancer cell line; NCAM, neural cell adhesion molecule; PDGF-Rp, platelet-derived growth factor; RLU, relative luminescence units.

Enrichment of Ependymal-Derived Progenitor Cells After SCI

Recently, the use of differential antigens to distinguish the endogenous ependymal cells within a mixed population of proliferating cells has been reported [36]. Thus, the platelet-derived growth factor (PDGF) receptor is expressed in ependymal cells and in neural stem cells, and neuroblasts, vimentin, and Dlx2 are unique markers for primitive ependymal cells although NCAM is a marker for neuroblast [36]. Given the heterogeneous nature of the cell culture preparations, all markers were detected in both epSPCs and epSPCi after 10 days of in vitro culture. Interestingly, the expression of primitive ependymal marker Dlx2 was higher in epSPCs, epSPCi displayed enhanced expression of the parenquimal progenitor cell markers NCAM-1, PDGF-Rp, Olig2, Nkx2.2, and NG2 (Fig. 1E). After injury, the population of oligodendrocyte progenitors has been reported to increase as characterized by high levels of Olig2, Nkx2.2, NG2, and PDGF receptor expression [37]. Additionally, we observed a significant enrichment of NCAM-expressing neuroblasts in cultures of epSPCi (Fig. 1E). On the other hand, expression of the glial marker GFAP was more prominent in undifferentiated epSPC cultures (Fig. 1E).

Differential Transcriptional Profiles of Stem/Progenitor Cells After SCI

During SCI, a variety of molecular signals occur simultaneously that contribute to neurogenesis, inflammation, and cell death. To identify the complex signaling cascades involved in activation of epSPCs after injury, the gene expression profile of epSPCi was compared with epSPCs by the Agilent DNA microarray technology. We detected 1,321 genes in which expression was significantly (p < .05) different in epSPCi from samples of four different animals. The two different lists of genes were organized according to Gene Ontology (GO) or Kyto Encyclopedia of genes and genomes (KEGG) terms by using the corresponding gene-GO or gene-KEGG association tables to obtain FatiGO or KEGG implement analysis. In concordance with the data of Figure 1 and with the previous reports, the underrepresented terms in epSPCi revealed a gene profile associated with active cell turn-over and proliferation (data not shown). Moreover, the KEGG terms showed upregulation of the Jak/Stat and VEGF/MAPK signaling pathways in epSPCi (Table 1). These results are consistent with our previous observations in which inflammatory mediators, present in the injured tissue, may influence the proliferative activity of epSPCs in vitro via the mentioned pathways. Moreover, when we stimulate these pathways by exogenous bFGF, the proliferative difference between epSPCi and epSPCs was significantly diminished under use in vitro conditions (supporting information Fig. 1).

Table 1. Jak/Stat and VEGF/MAPK signaling pathways are unregulated in ependymal stem/progenitor cells injury
inline image

epSPCs Differentiate Efficiently to Oligodendrocytes

According to Keirstead et al. [22], the oligodendrocyte differentiation protocol was extended up to 42 days (Fig. 2A). On days 0, 3, 7, 21, 28, and 35 in vitro, samples were harvested for immunoassay (Fig. 2B) or semiquantitative reverse transcriptase polymerase chain reaction (Fig. 2C). The expression of the astrocytic marker GFAP was diminished in the presence of ATRA (from day 3) with further recovery after ATRA removal. From day 7, GFAP expression was downregulated more rapidly in epSPCi in comparison with epSPCs (Fig. 2B, 2C). Expression of Sox10, Olig1, Olig2, and Nkx2.2 homeodomain transcription factors as well as the platelet-derived growth factor receptor (PDGFR) are linked to oligodendrocyte specification during spinal cord development [38, 39]. The expression of all of these markers was consistently higher during the differentiation of epSPCi as compared with epSPCs (Fig. 2C). Nkx2.2 is expressed at low levels during early phases but the second phase of higher expression is reported to be a hallmark of oligodendrocyte differentiation in vivo [38]. We have observed a similar phenomenon in vitro; the expression of Nkx2.2 increased notably at day 28 of differentiation in cultures of epSPCi (Fig. 2C). Typical oligodendrocyte markers such as A2B5 or RIP were widely expressed in OPC obtained from the differentiation of epSPCi and epSPCs (Fig. 2B). However, the NG2 proteoglycan, which is induced after spinal cord insults and exhibits remyelinating properties [40], was more abundantly expressed in epSPCi (Fig. 2B).

Figure 2.

Directed differentiation of epSPCs to oligodendrocytes. (A): Schematic representation of differentiation protocol. (B): Immunocytochemical analysis of the astrocytic (GFAP) and oligodendrocyte markers for early and later differentiation stages (A2B5 and NG2 or RIP, respectively) of cells plated on Matrigel at day 0, 3, 7, and 28 of the differentiation protocol. (C): reverse transcriptase polymerase chain reaction analysis of mRNA expression of transcriptional factors which regulate oligodendrocyte differentiation process performed at 0, 3, 7, 21, or 28 days before culturing on Matrigel. These results are representative of three experiments. Abbreviations: ATRA, all-trans retinoic acid; bFGF, basic fibroblast growth factor; CM, complete medium; DAPI, 4,6-diamidino-2-phenylindole dihydrochloride; EGF, epidermal growth factor; epSPCi, ependymal stem/progenitor cells injury; epSPCs, ependymal stem/progenitor cells; GFAP, glial fibrillary acidic protein; GRM, glial restriction media; PDGFR, platelet-derived growth factor receptor; PLL, poly-L-lysine; RIP, oligodendrocyte cell marker (2′,3′-cyclic nucleotide 3′-phosphodiesterase)-non-compact myelin protein.

In addition to the differential expression of oligodendrocyte markers, we also noted that morphological changes associated with differentiation were delayed in epSPCs. After day 3, epSPCi-differentiated cells displayed a predominantly bipolar morphology which is characteristic of early OPC in contrast to the fibroblast-like shape of epSPCs. Nevertheless, in both epSPCi and epSPCs cultures, mature-oligodendrocyte-like phenotype was observed after 28 days of differentiation on either Matrigel or laminin (Fig. 2B). The levels of myelin binding protein (MPB) mRNA were equivalent between epSPCs and epSPCi-derived OPC at the end of the differentiation protocol (cells grow on laminin from day 35 to 42, data not shown).

Differentiation of epSPCs to Functional Motoneurons

Spinal motoneurons were generated by culturing epSPCi or epSPCs with ATRA to neutralize other signals and to establish a caudal position, and subsequently by addition of SHH to promote a ventral position [41] (Fig. 3A). By day 25 of the differentiation protocol, a substantial proportion of both epSPCs or epSPCi displayed a motor neuron-specific transcriptional pattern with acquisition of immunohistochemical features such as HB9 (transcription factor induced during motoneuron differentiation) and MAP2 (a neuronal marker) protein expression (Fig. 3B). The excitability of the resulting motoneuron-like cells from epSPCi (epSPCi-Mot) or epSPCs (epSPCi-Mot) at day 25 of the differentiation protocol was studied by the current-clamp mode of the patch-clamp technique. Representative examples obtained from epSPCi-Mot are shown in Figure 3C. In the voltage-clamp configuration, a sodium current was observed by giving a squared voltage-pulse at −10 mV from a Vh = −70 mV in 21% (n = 28) of epSPCi-Mot cells, with a mean current amplitude of 930 ± 470 pA (Fig. 3C, upper trace) and only 7% (n = 29) of epSPC-Mot cells with a mean current of 170 ± 139 pA (data not shown). Outward currents activated by a pulse at +60 mV were blocked ∼70% by applying the voltage-dependent K+ channel blocker TEA at 10 mM (Fig. 3C, lower trace). This K+ current was present in 22% (n = 28) and 32% (n = 28) of epSPCs and epSPCi, respectively. The K+ current amplitude was not significantly different between both groups. A single-spike action potential was always observed. Cells that failed to respond to the depolarizing pulses most likely represent an immature phenotype.

Figure 3.

Directed differentiation of epSPCi to motoneuron. (A): Scheme of differentiation protocol. (B): Immunocytochemical analysis of MAP2 as a marker of mature neurons and HB9 for spinal motoneurons. The histogram shows the quantification of HB9 and MAP2 positive cells at day 25 of the differentiation protocol from three independent experiments. (C): Electrophysiological characterization of epSPCi-motoneuron differentiated cells at day 25 of the differentiation protocol. Outward currents evoked at +60 mV Vh = −70 mV were from ∼70% blocked by applying the voltage-dependent K+ channel blocker TTA (10 mM). Abbreviations: AMPc, cyclic adenosine monophosphate; ATRA, all-trans retinoic acid; BDNF, brain derived growth factor; bFGF, basic fibroblast growth factor; CM, complete medium; EGF, epidermal growth factor; epSPC, ependymal stem/progenitor cells; epSPCi, ependymal stem/progenitor cells injury; FMM, final motoneuron medium; GDNF, glial derived growth factor; IGF-I, insulin like-growth factor one; MM, motoneuron medium; SHH, sonic Hedgehog homolog; TEA, tetraethylammonium.

Thus, we have developed efficient protocols to obtain functional OPC (Fig. 2) and motoneurons (Fig. 3) by in vitro-directed differentiation of epSPCs and epSPCi. We consistently obtained a better yield of differentiated cells from epSPCi cultures, isolated from injured animals, as compared with homologous epSPCs from healthy donors (OPC: 87.3% ± 16% vs. 21.1% ± 9.8%; motoneurons: 45.6% ± 8.2% vs. 12% ± 5.6%).

epSPCi Accelerates Functional Locomotor Recovery in Rats with Spinal Cord Contusions

Previous approaches based on spinal cord-derived stem/progenitor cells from either fetal or adult healthy donors reported only modest or no significant functional improvement during the acute phase of transplantation [26, 42]. Here, we demonstrate that transplants of epPSC obtained from injured tissue promote early functional locomotor recovery in rats with SCI (Fig. 4A, 4B). The BBB locomotor rating scale and CatWalk foot print analysis were utilized weekly to quantify the functional recovery [28]. The Control group (traumatic lesion by severe contusion) did not score beyond the 10th level (occasional weight-supported plantar steps: no forelimb-hindlimb coordination) on BBB ranking 1 month after lesion. In contrast, the transplanted groups (epSPCi or OPCi: acute transplantation of epSPCs or OPC from injured spinal cord of homozygote transgenic eGFP rats) showed consistent plantar stepping, capacity to support weight, and coordination when evaluated 1 month after transplantation (BBB scale: epSPCi: 17.76 ± 2.4; OPCi: 21 ± 0.2; Fig. 4A; see supporting information Videos 1 and 2 of a representative animal from the control and epSPCi groups, respectively). We also performed Narrow Beam crossing test to quantify the ability of the animals to maintain balance. It is well-known that walking on beams increases the difficulty with respect to the open field tests [43]. In this case, a precise paw placement is needed which is dependent on the integrity of the corticospinal tract. None of the control animals were able to cross the beam when tested after the surgical induction of SCI. However, 2 months after transplantation, six of the 13 epSPCi transplanted rats walked the length of the horizontal runway. Foot print analysis using the CatWalk device was recorded when rats crossed a glass plate, illuminated from the side (see a representative example in Fig. 4B, upper panel). Paw contact was quantified by counting high-intensity pixels [29]. Visualization and accurate timing of all paw placements during walking locomotion allowed the calculation of the parameters shown in Figure 4B (lower panel). Two months after surgery, the regulatory index (RI, a fractional measure of inter-paw coordination [29]) based on the step sequence normal patterns indicated a recovery of hind-limb coordination within the transplanted group (epSPCi) in comparison with nontransplanted rats (control). The stride length (SL) represents the distance between successive placements of the same paw (using the maximal contact data). A reduction of hind-SL observed in control animals (in comparison with sham animals) was significantly recovered in the epSPCi group (Fig. 4B, lower panel). Both base of support-left or -right hind analysis displayed a significant recovery in the animals transplanted with epSPCi. No abdomen contacts were recorded in the epSPCi group, and only two out of 13 rats registered positive tail contact recordings (Fig. 4B, lower panel). At the time of sacrifice, which was 2 months after SCI, no gross macroscopic differences were noted between the spinal cords of the control and transplanted groups (data not shown). However, the neuroanatomical stain PAH (Fig. 4C) revealed a reduction in the area of cysts/cavities in and around 1.5 mm of the lesion in the epSPCi transplanted group (n = 6). The epSPCi, which constitutively express eGFP and were acutely transplanted 2 mm rostral and caudal to the lesion, were detected in sagittal sections where they invaded the lesion zone which is delineated by the GFAP staining (Fig. 4D, upper panels). Therefore, these observations suggest that the transplanted cells migrated long distances from the rostral and caudal regions of the transplant, accompanying the neurofilament (NF)-labeled axons in and around the lesion zone (Fig. 4D, upper panels). We also detected MBP-debris surrounded by macrophages in the injured zone as positively labeled with ED1 (Fig. 4D, lower panels). The transplanted epSPCi were visualized crossing the injured zone, indicating that the lesion did not repel their migration. Although most of the eGFP-expressing cells colocalized with NG2 staining, they were rarely positive for markers of mature cells such as APC and RIP (for mature oligodendrocytes) or MAP2 (for mature neurons) (data not shown).

Figure 4.

Functional locomotor recovery by acute transplantation of epSPCi or OPCi. (A): Locomotor performance was evaluated weekly by BBB scoring and a significant improvement was observed in rats transplanted with epSPCI or OPCi when compared with nontransplanted ones (control). (B): CatWalk analysis, upper panel: step pattern analysis showed a defective printing for the right and left hind paws in a representative control animal when compared with transplanted one (epSPCI). Lower panel: all parameters tested in epSPCi transplanted animals showed a recovery comparable with the levels of the sham animals. Statistical comparisons were performed with Student's t test (∗compared with sham group; $, compared with control group, p < .05). (C): The area of the cavities/cysts was quantified 2 months after surgery in control and transplanted animals by MetaMorph Offline software. Representative pictures after phosphotungstic acid hematoxylin staining of control and epSPCi group were shown from epicenter and 1.5-mm rostral to the lesion. Area of the cavities/cysts were quantified in pixels, and the mean ± SD of the area in equivalent sections in all animals per group is represented in the table in absolute units (∗compared with control group, p < .05; n = 6). (D): eGFP-transplanted cells migrate to the lesion zone. The eGFP cells migrate at least 2 mm from rostral and caudal to the lesion zone accompanying the NF-labeled axons even through the GFAP-positive scar. The lower panels show immunodetection of MBP-debris (red) surrounded by the macrophages, positively labeled with ED1 (blue). The higher magnification picture shows a detail of the lesion zone. Scale bars 50 μm. Abbreviations: BBB, Basso, Beattie and Bresnahan; BSO-LH, base of support-left hind; BSO-RH, base of support-right hind; epSPCi, ependymal stem/progenitor cell injury; GFAP, glial fibrillary acidic protein; GFP, green flourescent protein; LFP, left front paw; LHP, left hind paw; MBP, myelin binding protein; NF, neurofilament; NP, normal pattern; OPCi, oligodendrocyte precursor cell injury; RFP, right front paw; RHP, right hind paw; RI, regulatory index; SL-H, stride length-hind.


Although various lines of evidence suggest the physiological relevance of the neurogenesis programs in adult mammals, no data exist to answer the following critical question: is adult neurogenesis capable of regenerating neurological dysfunction following injury? Given the serious social and health problems presented by diseases and accidents that destroy neuronal function, there is an ever-increasing interest in determining whether adult stem cells might be utilized as a basis of regenerative therapies. With the present study, we demonstrate that after SCI, the epSPCs undergo phenotypic and genotypic changes, most probably mediated by the cytokines secreted from infiltrating immune cells. These alterations include enhanced self-renewal properties, better response to differentiation signals, and improved regenerative capacity as demonstrated by more efficient recovery of locomotor function after SCI. Recently, Meletis et al. [37] have characterized the ependymal population within the spinal cord of adult healthy mice or after SCI by employing genetic fate mapping. This study has shown that nearly all properties of neural stem cells reside within ependymal cells. In response to the spinal cord lesion, the proliferation of ependymal cells increased significantly and these cells migrated to the injured zone. Moreover, the migrating cells lost their ependymal phenotype, based on changes in the expression of cell markers; most of the ependymal derived cells expressed Sox9 and vimentin, although a significant population within the injured site were Olig2 positive [37]. In the present study, we demonstrate that after SCI the number of epSPCs which forms neurospheres increases dramatically. Based on the expression of certain markers, the neurosphrere-forming cultures obtained from injured rats were enriched in neuroblast and parenchymal progenitor cells (high-expression levels of NCAM-1, PDGF-Rp, Olig2, and Nkx2.2), thereby minimizing the proportion of primitive Dlx2-expressing ependymal cells. The glial marker GFAP was more abundantly expressed in cultures of epSPCs as compared with epSPCi cultures. Thus, the presence of GFAP-positive astrocytes and microglia could conceivably contribute to the limited self-renewal of the epSPC culture [44].

Spinal cord contusion causes acute central hemorrhagic necrosis accompanied by profound activation of glia and leukocyte infiltration. During the first 24 hours, neutrophils accumulate within the lesion, and from day 3 to 7, there is a peak of lymphocyte and monocyte infiltration [45]. Although a beneficial role for neuroinflammation is supported by an extensive range of scientific publications, it remains controversial. Both the microglia/macrophages and lymphocytes secrete neuroprotective cytokines and growth factors that promote neuron protection and induce the differentiation of precursor cells to neurons and oligodendrocytes [18]. Moreover, the proliferation of endogenous epSPCs has been reported to be mediated by IL-6 and EGF via the JAK/STAT and MAPK signaling pathways [19]. In our studies, this biological scenario was identified and confirmed by transcriptome analysis of epSPCs versus epSPCi. The importance of the immune response in activating signaling cascades was reflected in the gene expression profile obtained from epSPCi. The activation of the mentioned signaling pathways by bFGF in epSPC induced a comparative proliferative profile in epSPCi group. Although all, IL-6, EGF, or bFGF-Jak/Stat stimulus efficiently induced the phosphorilation of Stat3, the first two did not show any effect on cell proliferation (data not shown and supporting information Fig. 1) under in vitro conditions. However, which stimulus appears to be crucial for biological activation in vivo deserves further studies because this will allow us to reproduce the entire phenomenon under in vitro conditions.

The ability to modulate the fate of epSPCs by specifically inducing neural progenitors may offer an effective alternative for cell replacement therapies to rescue the loss of functional units after SCI. With the present study, we demonstrate by in vitro assay that the multipotent capacity of epSPCs is comparable before or after SCI. However, a significantly larger population of cells completed the differentiation process when they were isolated after SCI, probably due to the limited life-span of the epSPCs from healthy donors because they were not able to expand for more than three passages (data not shown). Although postnatal neurogenesis in spinal cord was previously described [46], very little data have been reported about the specific differentiation process to obtain functional spinal motoneurons from adult epSPCs. Using our differentiation protocol, almost 90% of differentiated culture of epSPC stained positive for the motoneuron specific marker HB9. However, only 32% of these cells displayed potassium currents as characterized by electrophysiology, indicative of those with functional neuronal capacity. The electrophysiological study revealed the presence of an immature phenotype within the differentiated cell culture, suggesting that the incubation with the growth factors should be extended.

Transplantation of adult spinal cord-derived neurospheres from healthy donors was first shown to slightly improve motor recovery with aberrant axonal sprouting associated with allodynia [24]. Parr et al. have reported a poor survival rate when spinal cord-derived neurospheres were acutely transplanted into a compressed spinal cord, with no functional improvement [26]. Recently, the same group has confirmed a regenerative potential when epSPCs were transplanted 1 week after spinal cord compression [25]. The authors suggested that the inability of transplanted cells to recover function may likely be due to the severity of the injury. However, because the cell transplantation coincided with the peak of inflammatory response, a synergistic effect between transplanted and infiltrated cells could represent a beneficial mechanism as previously described [47]. We have been able to efficiently rescue the neuronal function in rats with SCI by exploiting the endogenous regenerative machinery as influenced by the microenvironment after SCI. In our model, the survival rate of transplanted epSPCi was high (data not shown). We have demonstrated a fast and consistent recovery of locomotor activity by epSPCi or OPCi in acute transplantation. The early functional recovery, which was apparent from the first week, might contribute strongly to a better functional outcome in the transplanted animals in comparison with the control group. Although no significant differences were detected between the two groups which received transplants, the OPCi rats showed better BBB scores and smaller individual differences compared with epSPCi group. The transplanted epSPCi can generate a paracrine environment for the release of growth factors and positively modulate the local immune response, thereby promoting neuronal protection and survival with axonal outgrowth [48]. Differentiation of the transplanted epSPCi occurred rarely, suggesting a regenerative mechanism that does not depend on a specific cell fate. The transplanted cells migrated to the injured zone and were found in direct proximity to neurofilament-immunoreactive axons in and around the lesion. These observations reinforce the idea that the epSPCi do not constitute a major impediment to axonal growth and may even suggest that they provide an axonal support. The observed reduction of cysts/cavities formation is also indicative of a potential protective mechanism provided by transplanted epSPCi.

In summary, our present experimental data strongly support the notion that the rat spinal cord contains functional regenerative machinery, perhaps residing in the epSPCi, with the potential to recover neurological function following injury. Additionally, our study describes the development of novel experimental procedures to efficiently produce oligoneurons and motoneurons from epSPCs that will expand our knowledge of the biology of this cell type in the design of novel regeneration therapies including cell transplantation and mobilization of endogenous stem cells.


We gratefully acknowledge Dr. Deborah Burks for her critical reading of the manuscript. We also acknowledge Eva Escorihuela for her technical support and the Confocal Microscopy service of the Centro de Investigacion Príncipe Felipe. This research was supported by the Spanish Program of Regenerative Medicine Conselleria de Sanidad de la Generalidad Valenciana-Instituto de Salud Carlos III, Fondo de Investigaciones Sanitarias-Instituto de Salud Carlos III, Project RD06/0010/1006 and the Ministerio de Educación, Ciencia y Tecnología, grants SAF2007-63714 and SAF2007-63193, respectively.


The authors have declared that no competing interests exist.