FM19G11 Favors Spinal Cord Injury Regeneration and Stem Cell Self-Renewal by Mitochondrial Uncoupling and Glucose Metabolism Induction§


  • Author contributions: F.J.R.J., M.S, and V.M.M: Conceived and designed the experiments; V.M.M and M.S.: Financial support to perform the work; F.J.R.J., A.A.A, S.E., and V.M.M: Performed the experiments; F.J.R.J. and V.M.M., Wrote the paper and conceived and designed the experiments.

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

  • §

    First published online in STEM CELLSEXPRESS August 3, 2012.


Spinal cord injury is a major cause of paralysis with no currently effective therapies. Induction of self-renewal and proliferation of endogenous regenerative machinery with noninvasive and nontoxic therapies could constitute a real hope and an alternative to cell transplantation for spinal cord injury patients. We previously showed that FM19G11 promotes differentiation of adult spinal cord-derived ependymal stem cells under hypoxia. Interestingly, FM19G11 induces self-renewal of these ependymal stem cells grown under normoxia. The analysis of the mechanism of action revealed an early increment of mitochondrial uncoupling protein 1 and 2 with an early drop of ATP, followed by a subsequent compensatory recovery with activated mitochondrial metabolism and the induction of glucose uptake by upregulation of the glucose transporter GLUT-4. Here we show that phosphorylation of AKT and AMP-activated kinase (AMPK) is involved in FM19G11-dependent activation of GLUT-4, glucose influx, and consequently in stem cell self-renewal. Small interfering RNA of uncoupling protein 1/2, GLUT-4 and pharmacological inhibitors of AKT, mTOR and AMPK signaling blocked the FM19G11-dependent induction of the self-renewal-related markers Sox2, Oct4, and Notch1. Importantly, FM19G11-treated animals showed accelerated locomotor recovery. In vivo intrathecal sustained administration of FM19G11 in rats after spinal cord injury showed more neurofilament TUJ1-positive fibers crossing the injured area surrounded by an increase of neural precursor Vimentin-positive cells. Overall, FM19G11 exerts an important influence on the self-renewal of ependymal stem progenitor cells with a plausible neuroprotective role, providing functional benefits for spinal cord injury treatment. STEM Cells2012;30:2221–2233


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. In fact, transplantation of epSPCs immediately after severe spinal cord injury (SCI) significantly rescues the lost locomotor activity 1 week after lesion and transplantation [1]. Recent findings have demonstrated the relevance of the cellular bioenergetics, the metabolic state, and the mitochondrial contribution to the control of stem cell fate and somatic cell reprogramming [2–6]. Some attempts to characterize the temporal mitochondrial bioenergetics after SCI had the goal of identification of the therapeutic window for further pharmacological interventions and clearly indicate the relevance of mitochondrial dysfunction in this trauma [7]. Mitochondrial uncoupling is a term that refers to a functionally disconnected flow of protons from the production of ATP through the electron transport system. Mitochondrial uncoupling proteins (UCPs) belong to the superfamily of mitochondrial anion transporters that can dissociate oxidative phosphorylation causing uncoupled mitochondrial respiration from ATP synthesis and therefore play a critical role in energy balance. UCP-mediated proton leak may compete with ATP synthesis and in fact, ectopic expression of the most studied component of UCPs family, UCP1 in HepG2 cells significantly decreased the production of ATP [8]. Mitochondrial uncoupling stimulates a compensatory mechanism to restore ATP levels, because a lack of compensation may have severe consequences for the cells [9]. The resulting diminution of ATP leads to an increase in glucose uptake for anaerobic generation of ATP. The mitochondrial uncoupling agent 2′,4-dinitrophenol (DNP) compromises ATP production by disrupting the mitochondrial electron transport chain and in addition causes an increase in the level of glucose transporter 4 (GLUT-4), which increases the rate of glucose uptake [10]. Although a long-term and complete uncoupling of mitochondria and prolonged reductions in ATP would be detrimental for the cells, a transient or mild mitochondrial uncoupling produced by nontoxic agents could confer neuroprotection and has been proposed as a target for therapeutic interventions after neuronal injury [11, 12]. Thus, DNP and strategies that uncouple mitochondria may hold potential as a pharmacological therapy for acute SCI [13] or 6 weeks after injury [14]. Activation of two crucial molecules, the AKT protein kinases and the cellular sensor of energy status AMP-activated kinase (AMPK), regulate GLUT-4 traffic and glucose uptake [15]. The compensatory increase of glucose uptake involves AMPK activation, which elevates cell surface GLUT-4 by retarding its endocytosis and AKT activation that increases its exocytosis [10]. As reviewed elsewhere [16], there is an active crosstalk between both AMPK and AKT pathways and some stimuli cause the parallel activation including stress that can regulate energy supply [17]. Insulin and AMPK activators can act synergistically to induce AKT overactivation [18]. It is also known that, under different cellular energy conditions, the activated AMPK can also regulate the downstream target of AKT, mTOR [19], which is positioned as an upstream activator of the hypoxia-inducible transcription factors (HIF). FM19G11 was first identified as an inhibitor of HIFα protein expression and transcriptional activity under hypoxic conditions repressing a variety of key genes involved in stemness [20]. In contrast, FM19G11 augmented mTOR activation and HIF1α expression in cells grown under normoxia [21]. Differentiation experiments demonstrated that this HIFα inhibitor favors oligodendrocyte differentiation, possibly through the modulation of Sox2 and Oct4 expression and by allowing neural stem and/or precursor cells from adult spinal cord to differentiate under hypoxia. It was then suggested that the low toxicity profile of this drug may favor pharmacological approaches that enable to act on spinal cord regeneration. It has been recently shown that increased expression of glycolytic genes occurs prior to induction of pluripotency [4]. In contrast to differentiated cells, glucose uptake is less coupled to oxidative phosphorylation in human pluripotent stem cells (hPSCs) that predominantly use glycolysis to produce ATP [22]. In addition, the potentiation of the bioenergetic conversion from somatic oxidative mitochondria toward an ATP-generating glycolytic state increases the efficiency to reprogram somatic cells into stem cells [2]. Therefore, stemness can also be regulated by energy metabolism, and differences in energy status can distinguish human pluripotent stem cells from differentiated cells [23]. This work shows that the mitochondrial uncoupling protein UCP2 regulates energy metabolism and self-renewal potential of hPSCs. In summary, therapies based on the reduction of mitochondrial dysfunction and increment of stem cell population could constitute beneficial pharmacological treatments for acute SCI. Discovery of potential pharmacological tools that favor functional recovery after SCI still constitute a challenge but is crucial to minimize secondary damage. Here we show a new chemical compound, FM19G11 that is able to modify the mitochondrial uncoupling process, which induces glucose uptake by activation of AMPK as well as AKT signaling pathways. The consequence of these molecular events is an increase self-renewal of epSPC, as indicated by the increment of pluripotent markers, showing favoring characteristics of FM19G11 for its potential use in regenerative therapies such as spinal cord disorders.


Ependymal/Progenitor Cell Isolation, Culture, and Drug Treatments

epSPC were harvested from adult female Sprague–Dawley rats, isolated and cultured as previously described [1]. Neurospheres formed from epSPC were treated with FM19G11 (Sigma Chemical, MO, or its vehicle (DMSO) at the indicated concentrations (0–1 μM). Rapamycin was purchased from Calbiochem (San Diego, CA, and used at 100 pM. Wortmanin was used at 2.5 μM (Sigma Chemical), Compound C was used at 10 μM (Calbiochem), and DNP was used at 50 μM (Sigma Chemical).

Human Embryonic Stem Cell/Human Inducible Pluripotent Stem Cell Culture

Primary human embryonic stem cell (hESC, H9 line; WiCell, Madison, WI, or human inducible pluripotent stem cell (hIPSC) [24] colonies were cultured as previously described [25]. For drug stimulation, the colonies were seeded onto Matrigel (1:10; BD Bioscience, San Diego, CA,

Human Skeletal Muscle Cell, Adipocytes, and mC2C12 Culture

Human skeletal muscle cell (SKMC) and the human white preadipocytes (HWP) were purchased from (PromoCell, Heidelberg, Germany, Both primary human lines were kept and treated for the different assays in growth media (GM) up to passage 5–6. The components of hSKM-GM were: 10% FBS, 50 μg/ml Bovine Fetuin, 10 ng/ml epidermal growth factor (EGF), 1 ng/ml basic fibroblast growth factor (bFGF), 10 μg/ml insulin, and 0.4 μg/ml dexamethasone; HWP-GM consisted of 10% FBS, 2% endothelial cell growth supplement, 10 ng/ml EGF, and 1 μg/ml hydrocortisone. The mouse myoblast (mC2C12) cell line is an immortalized cell line that was cultured in high glucose Dulbecco's modified Eagle's medium supplemented with 10% FBS.

RNA Isolation and Semiquantitative Real Time Polymerase Chain Reaction

Total RNA was extracted by using the RNeasy Mini-kit (Qiagen, Germany, according to the manufacturer's instructions. One microgram of total RNA was reverse-transcribed in a total reaction volume of 50 μl at 42°C for 30 minutes using random hexamer primers. The semiquantitative polymerase chain reaction (PCR) amplification was performed as previously described [20]. The specific bands were analyzed by densitometry, and the ratio with GAPDH expression was represented. For quantitative analyses, we used LightCycler SYBR Green I technology, and the relative expression of mRNA transcripts was analyzed by the ABI PRISM 5700 Sequence Detection System (Applied Biosystems, Foster City, CA, As template, we used 40 ng of cDNA to analyze the expression of target and housekeeping genes (GAPDH) in separate tubes for each pair of primers. The upstream (FW) and downstream (RV) primers as well as the amplicon length are detailed in Supporting Information Table. The comparative threshold cycle (CT) method was used to calculate the relative expression as followed: ΔCT [ΔCT = CT (test gene) – CT (GAPDH)]. ΔCT for FM19G11-treated samples was then subtracted from the ΔCT for vehicle-treated samples, to generate ΔΔCT [ΔΔCT = ΔCT (FM19G11) − ΔCT (vehicle)]. The mean of these ΔΔCT measurements was used to calculate the fold change in gene expression (2−ΔΔCT). Representative results were presented as the mean ± SE.

DNA Micro-Array Hybridization and Analysis

epSPC were obtained from four different rats and were individually assayed and divided into two groups, DMSO or FM19G11 (500 nM) treated for 48 hours. Labeled cRNA (3 μg) was hybridized with Whole Rat Genome Oligo Microarray Kit (Agilent p/n G2519F-014879). The protocol as well as the raw and normalized data was deposited in the Array Express MIAME compliant database (; ArrayExpress accession: E-MEXP-2549). The gene profile was sorted by differential expression levels between the two experimental conditions and clustered into biological functional profiling using FatiGO application [26].

ATP Cell Content Quantification

The assayed cells, floating neurospheres of epSPC, hSKM, and HWP were plated onto opaque 96-well plates, 24 hours before the experiment started. We strictly followed the manufacturer's instructions for the Adenosine 5′-triphosphate ATP bioluminescent somatic cell assay kit (FLASC; SigmaAldrich, U.K.). The luminescence was measured by a Multilabel Counter VICTOR3 (Perkin Elmer, Waltham, MA,

MitoTracker Staining/Flow Cytometry Analysis

For fluorescent microscopy detection, epSPC were seeded into coverslips covered by Matrigel (BD Biosciences, San Jose, CA, for 24 hours and then treated with 500 nM FM19G11 or vehicle (DMSO) in a time-response manner (15 minutes-24 hours). Fifteen minutes before fixation with 2% paraformaldehyde (PFA), the cells were incubated with MitoTracker Red CMXRos (1 μM; Invitrogen Ltd.) at 37°C. For fluorescence-activated cell sorting (FACS) detection, 105 cells were treated with FM19G11 and MitoTracker at the indicated dose and times under neurosphere cell-forming culture conditions. Labeled mitochondria were identified on a FACScalibur flow cytometer (BD Biosciences) on the basis of Forward and Sideward Scatter parameters and red fluorescence.

Western Blotting Analysis

Cells were collected and proteins extracted by using 2% SDS-TrisCl lysis buffer. A total of 20 μg of protein was loaded per lane in 10% SDS-polyacrylamide gels and resolved by standard SDS-PAGE. Proteins were electrophoretically transferred onto PVDF membranes. Membranes were blocked with 5% nonfat dry milk in PBST for 60 minutes and incubated overnight with specific antibodies against UCP1, UCP2, Sox2, Oct4, Notch1, GLUT-4, (Abcam, Cambridge, U.K.,, vimentin (Millipore, Billerica, MA,, P-AKT, P-AMPK, P-CREB, AMPcyclic, P-STAT3 (Cell Signaling, U.K.,, at 1:500 dilution. β-Actin at dilution 1:5,000 (Sigma) was used as a loading control. Subsequently, membranes were incubated with anti-mouse, anti-rabbit, or anti-goat horseradish peroxidase-conjugated secondary antibodies (1:5,000) (Sigma). Blots were visualized with the ECL (Amersham, U.K., detection system.

Glucose Transport Assay

epSPC neurosphere-forming culture, hSKMC, or mC2C12 were treated for 24 hours with FM19G11 (500 nM) or vehicle (DMSO). In addition, mC2C12 cells were also pretreated with wortmanin (2.5 μM) or Compound C (10 μM) for 30 minutes, rapidly washed in KRH buffer at pH 7.4 (25 mM HEPES-NaOH, 120 mM NaCl, 5 mM KCl, 1.2 mM MgSO4, 1.3 mM CaCl2, 1.3 mM KH2PO4) and incubated with or without 100 nM insulin (100 nM) for 10 minutes as a positive control. Then, 2-deoxy-d-(2,6-H3) glucose (GE Healthcare, U.K., (1 μCi) was added for an additional 10 minutes. Glucose uptake was determined after three washes with cold phosphate buffered saline (PBS) from cells lysed in RIPA buffer (50 mM TrisCl, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 0.1% SDS, 0.5% DOC, 1% Igepal) and counted for radioactivity.

RNA Interference by Small Interfering RNA Duplex Transfection for GLUT-4 and UCP1 or UCP2

The pool solution of small interfering RNA (siRNA) sequences ON-TARGETplus (SMART pool L-098209-01-0020, Thermo Scientific Dharmacon, USA, targeting rat GLUT-4 were used at 500 nM for lipotransfection. The annealed siRNA duplexes for UCP1 (Cat. No. s128565) and UCP2 (Cat. No. s222032) were purchased from Applied Biosystems and used at 500 nM for electroporation by using the Rat Neural Stem Cell Nucleofector Kit (Cat. No VPG-1005. Lonza, USA) and the Nucleofector Device (Cat. No. AAD-1001. Lonza, USA). Trasfections were performed in line with the manufacturer's instructions using the Nucleofector program A-33.

Telomerase Activity

The TRAPEZE Telomerase Detection Kit (Chemicon International, Temecula, CA, was used, and the manufacturer's instructions were strictly followed to detect the telomerase activity in treated and nontreated epSPC with FM19G11 for 2 days [1].

Clonogenic Assay

For single cell suspension, neurospheres were treated for 10 minutes with accutase (Sigma). A total of 1.6 × 104 viable cells were plated onto 60 mm ultra-low attachment plates in growth medium plus bFGF and EGF (20 ng/ml each), in the presence of FM19G11 (500 nM) or DMSO (vehicle). After 4 days of plating, the number of neurospheres (containing more than three cells) were quantified from three independent experiments.

Spinal Cord Contusion and Intrathecal Drug Administration

SCI by traumatic contusion was performed as was previously described [1]. Severe contusion (250 kdyn, “Infinitive Horizon Impactor,” Precision Systems, Kentucky, IL, at thoracic segment T8 with intrathecal administration of DMSO (vehicle, n = 9) or of FM19G11 (n = 14). Partial laminectomy of L6 allows introducing the catheter (previously filled with 0.9% saline solution) through a hole done in dura mater, up to the injured segments (T8). The osmotic pump, Model 1007D (Alzet Corp. Germany,; previously filled and incubated overnight at 37°C in a saline solution) delivers 0.5 μl per hour of a 9 mM FM19G11 solution or DMSO. The rats were premedicated with subcutaneous morphine (2.5 mg/kg) and Baytril (enrofloxacine, 5 mg/kg, Bayer, Germany, Leverkusen, and anesthetized with 2% isofluorane in a continuous oxygen flow of 1 l/minute. The open field locomotor test of Basso, Beattie, and Bresnahan (BBB) was used during 4 weeks as functional locomotor analysis. The experimental protocol was approved by the Animal Care Committee of the Research Institute Principe Felipe (Valencia, Spain) in accordance with the National Guide to the Care and Use of Experimental Animals (Real Decreto 1201/2005).

Magnetic Resonance Experiments

In vivo 1H magnetic resonance studies were performed at the NMR facility (SeRMN) of the Autonomous University of Barcelona in a 7 Tesla horizontal magnet (BioSpec 70/30, Bruker BioSpin, Ettlingen, Germany, equipped with actively shielded gradients (B-GA20S) using a quadrature 72 mm inner diameter volume resonator. Imaging parameters for these images were: effective echo time (TEeff) = 36 millisecond, repetition time (TR) = 2 second, echo train length (ETL) = 8, field of view (FOV) = 7 × 4 cm2, matrix size (MTX) = 128 × 128, slice thickness (ST) = 1.5 mm, and number of averages = 4. Using these scout images, a high resolution T2-weigthed respiration gated image was acquired in the sagital plane through the center of the spinal cord with the following parameter: TEeff = 45 millisecond, TR = 2 second, ETL = 4, FOV = 6 × 4 cm2, MTX = 512 × 256, and ST = 1.5 mm, number of averages = 8.

Eosin/Hematoxylin Staining for Quantification of Cavity/Cyst Formation

After 2 months of surgery, the animals were transcardially perfused with a 0.9% saline solution followed by 4% PFA in PBS, and 2 days incubation time in 30% sucrose before inclusion in Tissue-Teck OCT (Sakura Finetek U.S.A, Sagittal cryosections of 10 μm thickness were used for immunoassays. Every fifth section was collected for Eosin and hematoxylin staining to determine the cavitation/cystic area. The area of cavities/cysts was quantified from sagittal sections using ImageJ software from a collage of consecutive pictures taken using phase contract microscopy at ×10 magnification of approximately 20 mm2 of medullar tissue (including the epicenter of the lesion).

Immunocytochemistry and Immunohistology

Cells or cryosectioned tissues (10 μm) were fixed with 4% paraformaldehyde at room temperature for 10 minutes. After permeabilization with 0.5% Triton 100% solution containing 2% goat serum (blocking solution), the primary antibodies were incubated overnight at 4°C. Vimentin (α-mouse, clone V9 Cat. MAB3400; Millipore Corporation, Billerica, MA, USA, http//, TUJ1 (α-mouse; Cat. MO15052; Neuromics, Herford, Germany,, glial fibrillary acidic protein (GFAP, α-Rabbit; Cat. Z0334; DAKO, Germany,, GLUT-4 (α-mouse; Cat. mAb48547; Abcam) primary antibodies were diluted 1:200 in blocking solution. After being rinsed three times with PBS, the cells were incubated with Oregon Green-Alexa488 dye conjugated goat anti-mouse IgG or Alexa555 goat anti-rabbit IgG 1:400 (Invitrogen, CA, secondary antibodies for 1 hour at room temperature. All cells were counterstained by incubation with 4,6-diamidino-2-phenylindole dihydrochloride from Molecular Probes (Invitrogen) for 3 minutes at room temperature followed by washing steps. Signals were visualized by Confocal Microscopy (Leica, Germany,

Enzyme-Linked Immunosorbent Assay for UCPs

The assay for quantification of UCP1 and UCP2 in rat with SCI was performed by using the enzyme-linked immunosorbent assay (ELISA) kit, (Uscn, Life Science Inc., Wuhan 430056, P.R. China, for UCP1 (E95557Ra) or UCP2 (E92586Ra). Briefly, 1 cm of the rat spinal cord containing the injured tissue was homogenized in 250 μl of ice-cold PBS with a glass homogenizer on ice. The resulting suspension was sonicated three times with an ultrasonic cell disrupter Bioruptor UCD-200TM (Diagenode, Liège, Belgium, Homogenates were centrifuged for 5 minutes at 5,000g. The supernatants were collected and used for UCP protein quantification according to the manufacturer's instructions. The results of three independent experiments were represented as the concentration of UCPs (pg/ml).

Statistical Analysis

Statistical comparisons were assessed by the Student's t test. All p values were derived from a two-tailed statistical test using the SPSS 11.5 Software. A p-value <.05 was considered statistically significant.


FM19G11 Increases Cell Metabolism in Rat epSPC

In order to reveal the cell effect of FM19G11 under normoxia, we compared the gene expression profile of rat epSPC when treated with FM19G11 at 500 nM for 48 hours versus control (DMSO) (; ArrayExpress accession: E-MEXP-2549). The differentially expressed genes were organized according to gene ontology (GO) by using the corresponding gene-GO association and FatiGO-implemented analysis. As shown in Figure 1A, after biological function clustering, a clear influence of metabolic and cellular processes by FM19G11 was observed. Both metabolic and cellular processes appeared to be related to the activated ATP biosynthetic processes.

Figure 1.

FM19G11 activates cell metabolism. (A): FatiGO analysis of epSPC treated with 500 nM FM19G11 in comparison with vehicle alone for 48 hours. Three biological functional groups were over-represented in FM19G11-treated sample after hierarchical clustering. Scores over 0 imply positive regulation. (B): ATP cellular content quantification in epSPC (left graphic) and SKMC or HWP (right graphic) treated with FM19G11 (500 nM) or vehicle (DMSO) at different time points (15, 30, 60, 90, and 120 minutes; 24 and 48 hours). Relative values were normalized by the initial protein content and represented as percentage of control. Results were obtained from three independent experiments. Statistical comparisons were performed with Student's t test, *, p <.05. Abbreviations: epSPC, ependymal stem progenitor cell; HWP, human white preadipocyte; SKMC, skeletal muscle cell.

FM19G11 Causes an Early Induction of UCP1 and UCP2 Expression Coincidental with Reduced ATP Production and Mitochondrial Activity in Ependymal Stem Cells

In order to validate the previous microarray experiment (Fig. 1A), we quantified the ATP cellular content of epSPC, human skeletal muscle cells (hSKMC), and HWP in the presence of FM19G11 (500 nM) at short (0–120 minutes) or long (24 and 48 hours) incubation times (Fig. 1B). We observed that coincidental with the increment of the mitochondrial uncoupling proteins UCP1 and UCP2 (Fig. 2A), an initial rapid consumption of ATP occurs after treatment with FM19G11 (500 nM). In addition, we observed a reduction of mitochondrial activity at short times (15 minutes-4 hours), as visualized by MitoTracker labeled cells (Fig. 2C) or quantified by flow cytometry acquisition after 30 minutes of FM19G11 treatment (Fig. 2D). However, in accordance with the microarray analysis, we observed a significant (p <.05) accumulation of cellular ATP after 24 or 48 hours of treatment with FM19G11 (Fig. 1B) that correlates with an increment of mitochondrial activity in epSPC after 24 hours (1.51-fold change) (Fig. 2C). The evolution of the mitochondrial activity measured by flow cytometry is depicted in the graphic (Fig. 2C, right). Increased expression of mitochondrial genes, cytochrome B (complex I) (CYB) and NADH dehydrogenase, subunit 5 (complex I) (ND1) in rat epSPC and in the human embryonic stem cells (hESC) or in the hIPSC indicates an increase in mitochondrial biosynthesis after 24 hours of FM19G11 treatment (Fig. 3A). In addition, the murine mitochondrial ATPg and COX5 genes were also detected in C2C12 after 24 hours of FM19G11 treatment (Fig. 3A).

Figure 2.

Mitochondrial effects induced by FM19G11 in stem cells. (A): Evaluation of UCP1 and UCP2 expression in epSPC after treatment with vehicle (DMSO) (time = 0) or FM19G11 (500 nM) for 15, 30, 60, 90, or 120 minutes. (B): Cells treated with DMSO (time = 0) or FM19G11 (15, 30, 60, 240 minutes or 24 hours) and stained with MitoTracker Red (1 μM) or DAPI. (C): Quantification of cells stained with MitoTracker by flow cytometric determination after treatment with DMSO or FM19G11 (500 nM) for 30 minutes. The mean fluorescence intensity for DMSO and FM19G11-treated cells is indicated in the red square. The evolution of the percentage of gated cells stained by MitoTracker after treatment with FM19G11 (500 nM) at different time points is depicted in the graphic. Abbreviations: DAPI, 4,6-diamidino-2-phenylindole dihydrochloride; DMSO, dimethyl sulfoxide; epSPC, ependymal stem progenitor cell; UCP, uncoupling protein.

Figure 3.

FM19G11 causes changes in mitochondrial activity and gene expression in stem cells through activation of AKT and AMPK pathways. (A): Transcriptional expression of CYB and ND1 mitochondrial genes was quantified by real-time quantitative polymerase chain reaction (PCR) analysis in rat epSPC as well as hESC and hiPSC after treatment with DMSO or FM19G11 (500 nM). Relative mRNA expression of mouse mitochondrial genes ATPg and COX5 was quantified in mC2C12 cells after treatment with FM19G11 (500 nM). Results were standardized by the housekeeping gene GAPDH. mRNA levels were calculated by the 2−ΔΔCT method. (B): Evolution of AKT and AMPK phosphorylation in rat epSPC and mC2C12 after treatment with DMSO (time 0) or FM19G11 (500 nM) for 15, 30, 60, 90, or 120 minutes was examined by Western blotting. AMPc and CREB phosphorylation was also detected in epSPC. Different concentrations of FM19G11 (0.1, 0.2, 0.5 or 1, 5 μM) were used to examine AKT and AMPK phosphorylation in hESC after 24 hours of treatment. (C): Effects of blocking AKT and AMPK signaling on expression of mitochondrial genes were determined by real-time quantitative PCR analysis in epSPC cells after 24 hours of exposure to DMSO, FM19G11 (500 nM), or a combination of FM19G11 with the AKT inhibitor (Wortmanin) or FM19G11 with the AMPK inhibitor (Compound C). Statistical comparisons were performed with Student's t-test (significant increase *, p <.05); significant inhibition §, p <.001). Abbreviations: CYB, cytochrome B (complex I); DMSO, dimethyl sulfoxide; hESC, human embryonic stem cell; hIPSC, human inducible pluripotent stem cell; epSPC, ependymal stem progenitor cell; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; ND1, NADH dehydrogenase, subunit 5 (complex I).

FM19G11 Induces AKT and AMPK Pathways with Influence on Mitochondrial Biosynthesis

We evaluated the activation of AKT and AMPK pathways during a time-course treatment with FM19G11 (500 nM). A rapid phosphorylation of AMPK and AKT was caused by FM19G11 treatment in epSPC as well as in mC2C12. Consequently, downstream targets such as cyclic-AMP and P-CREB were also induced in epSPC by FM19G11 (Fig. 3B). In hESC, their activation was detected 24 hours after treatment with 5 μM FM19G11 (Fig. 3B). We explored whether mitochondrial biosynthesis was affected by activation of these two pathways by studying CYB and ND1 gene expression in the presence of specific inhibitors. In epSPC, the induction of both gene expression after 24 hours of FM19G11 treatment (500 nM) was impaired by using rapamycin (100 nM, mTOR inhibitor) or Compound C (10 μM, AMPK inhibitor) (Fig. 3C).

FM19G11 Induces Expression of the Glucose Transporter GLUT-4 Through Activation of AKT and AMPK Pathways

We detected an induction of protein expression of GLUT-4 at the cytosolic membrane in cells treated with FM19G11 for 24 hours (Fig. 4A, upper panels). Further examination of the differentially expressed gene list from the microarray experiment, an over-representation of genes involved in GLUT translocation such as Flot1, Vamp47, or Snap27 [27] was discovered, alongside hexokinase II, an enzyme that plays a key role in glucose homeostasis [28]. In all cases, microarray data were validated by semiquantitative PCR, and the average of the densitometric analysis of four individual samples showed significant differences between both groups, vehicle- or FM19G11-treated cells (Supporting Information Fig. 1). PI3K/AKT- and AMPK-mediated signaling have been extensively studied as regulatory mechanism for the activity of glucose transporters [29]. Both pathways were involved in the FM19G11 induction of GLUT-4 expression at the cytosolic membrane as evidenced through the use of specific inhibitors, Wortmannin (PI3K/AKT inhibitor) or Compound C (AMPK inhibitor). Moreover, the use of RNAi to knockdown GLUT4 expression revealed a direct implication on FM19G11-dependent induction of ND1 and CYB (Fig. 4B).

Figure 4.

A gain in surface GLUT-4 levels after FM19G11 treatment depends on PI3K and AMPK activation. Glucose uptake induction by FM19G11 in stem cells and mC2C12. (A): Immunocytochemical detection of GLUT-4 on the cell surface of epSPC which were pretreated for 30 minutes with Wortmanin (PI3K inhibitor, 2.5 μM) or compound C (AMPK inhibitor, 10 μM) and then treated with vehicle (DMSO) or FM19G11 (500 nM) for 24 hours. Scale bar = 50 μM. (B): 500 nM each siRNA duplex, scramble (Scr, nonspecific probe) or GLUT-4-specific rat probe was used for transfection of epSPC. After 24 hours of transfection, epSPC were exposed to FM19G11 (500 nM) or DMSO for 24 hours and the expression of mitochondrial genes CYB and ND1 was quantified by real time quantitative polymerase chain reaction analysis. *, p <.05 when compared with DMSO Scr-siRNA; §, p <.05 when compared with FM19G11 Scr-siRNA. (C): Rat epSPC, human hSKM, and mC2C12 were stimulated with DMSO or FM19G11 (500 nM) for 10 minutes with 100 nM insulin and then incubated for additional 10 minutes with 1μCi of 2-deoxy-d-(2,6-3H) glucose in KRH transport buffer and quantified after cell lysis. Glucose uptake of FM19G11-treated cells (24 hours) was represented as a percentage of change in comparison to vehicle-treated culture at the same time points. Wortmanin antagonizes the effect of FM19G11 on glucose uptake in mC2C12. Statistical comparisons were performed with Student's t test, *, p <.05 when compared with DMSO alone; §, p <.05 when compared with FM19G11 alone. Abbreviations: CYB, cytochrome B (complex I); DMSO, dimethyl sulfoxide; epSPC, ependymal stem progenitor cell; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; GLUT4, glucose transporter 4; hESC, human embryonic stem cell; IPSC, inducible pluripotent stem cells; siRNA, short interfering RNA.

FM19G11 Induces Glucose Uptake in epSPC

We confirmed a significant glucose uptake in epSPC, hSKMC, and mC2C12 after 24 hours of treatment with FM19G11 (500 nM) (Fig. 4C). The increased glucose uptake caused by FM19G11 in mC2C12 was inhibited by Wortmanin and Compound C, indicating a relevant role of AKT and AMPK pathways in this process.

FM19G11 Favors Self-Renewal of the epSPC Population

Neurospheres were treated with vehicle (DMSO) or FM19G11, and the clonogenic assay confirmed a significant induction of proliferation in epSPC treated with FM19G11 (500 nM) in comparison with stem cells treated with DMSO (Fig. 5A). In addition, immunodetection of phospho-histone 3 (p-H3) (a mitotic marker) revealed a significant increment in epSPCs treated with FM19G11 for 24 hours (Fig. 5B). The induction of telomerase activity and TERT expression (telomerase reverse transcriptase) in epSPC-FM19G11-treated cultures (+) as compared with nontreated cells (−) associates this proliferative activity to the self-renewal of the epSPC population (Fig. 5C). Hence, the expression levels of Sox2, Oct4, Nanog, and Notch1, critical factors for pluripotency, self-renewal, and the regulation of cell differentiation [30], were induced by FM19G11 in a dose-responsive manner (Fig. 5D, left). STAT3, a transcription factor related to stem self-renewal in neural precursors [31] was rapidly phosphorylated by FM19G11 after 15 minutes of treatment (Fig. 5D, right). We examined whether the increment of the mentioned markers of stemness-related proteins depend on GLUT-4. For this purpose, we ablated the expression of GLUT-4 by siRNA showing a blockage on FM19G11-depedent induction of all tested pluripotency-related markers (Fig. 5E). Moreover, we also explored the putative contribution of the AKT and AMPK pathways in the increment of Sox2 and Oct4 caused by FM19G11. We treated the cells with DMSO or FM19G11 (500 nM) for 24 hours and a combination of FM19G11 plus the AKT inhibitor (Wortmanin, wt), the mTOR inhibitor (Rapamycin, rap), and the AMPK inhibitor (Compound C, cc) for further comparison. We observed that the mentioned inhibitors blocked the increment of Sox2 and Oct4 caused by FM19G11 (Fig. 5F). In addition, we also explored a putative influence of UCPs in the increment of Sox2 and Oct4 expression caused by FM19G11. Therefore, to elucidate whether UCPs exert a relevant role in the increment of pluripotent markers caused by FM19G11, we used siRNA to silence the expression of UCPs. Although, no conclusion after UCP1 silencing was obtained due to inefficient ablation, no induction of Sox2 was observed when UCP2 was ablated (Fig. 5G).

Figure 5.

FM19G11 induces epSPC self-renewal. (A): Evaluation of the clone formation capability of epSPC after treatment with FM19G11 (500 nM). (B): Increased expression of phospho-H3, a marker of dividing cells, in epSPC after treatment with FM19G11 (500 nM). (C): Analysis of telomerase activity and TERT mRNA expression in epSPC after treatment with FM19G11 (500 nM). (D): Protein levels of pluripotency factors Sox2, Oct4, and Notch1 determined by Western blotting after treatment with different concentrations of FM19G11 (62.5, 125, 250, or 500 nM). Evaluation of phosphorylation of the self-renewal marker STAT3 in epSPC after treatment with FM19G11 (500 nM) at different time points. (E): Effect of GLUT-4 ablation on the expression of FM19G11-induced genes Sox2, Oct4, and Notch-1. Two microgram of each siRNA duplex, scramble (Scr, nonspecific probe) or GLUT-4 specific rat probe was transfected 48 hours before cell stimulation with DMSO or FM19G11 (500 nM). (F): Effects of blocking AKT/mTOR and AMPK activation caused by FM19G11 (500 nM) by Wortmanin (2.5 μM)/Rapamycin (100 pmol) and Compound C (10 μM), respectively, on the expression of Sox2 and Oct4 pluripotent markers. (G): Effects of UCP2 ablation in the expression of the FM19G11-induced gene Sox2. Two microgram of each siRNA duplex, scramble (Scr, nonspecific probe) or UCP1/2 specific rat probe was transfected 48 hours before cell stimulation with DMSO or FM19G11 (500 nM). β-Actin served as loading control. Statistical comparisons were performed with Student's t test, *, p <.05. Abbreviations: DMSO, dimethyl sulfoxide; epSPC, ependymal stem progenitor cell; GLUT4, glucose transporter 4; siRNA, small interfering RNA; UCP, uncoupling protein.

FM19G11 Treatment Improves Locomotor Recovery After SCI

We show in Figure 6A a schematic representation of the intrathecal administration system for DMSO or FM19G11. A sustained delivery system was created from an osmotic pump within a catheter located in the subarachnoid space up to the lesion area, immediately after SCI. A solution of FM19G11 (9 mM) from the osmotic pump delivers a 1.2 mmol per day in 0.5 μl per hour for the first week after SCI. The analysis of locomotor activity in an open field by using the BBB rating scale showed a significant increase in locomotion recovery of hind limbs in animals treated with FM19G11 in comparison with vehicle 4 weeks after injury (Fig. 6B; Supporting Information Video).

Figure 6.

FM19G11 promotes locomotor recovery after spinal cord injury. (A): Illustration of intrathecal administration of FM19G11 by osmotic pump in the area of the spinal cord lesion. (B): Locomotor performance of injured rats was evaluated weekly for 1 month by BBB scoring. Constant levels of DMSO or FM19G11 (1.2 mM) were maintained by intrathecal administration during the first week. (C): Quantification of UCP1 and UCP2 by ELISA in injured spinal cord of rats treated for 1 month with DMSO or FM19G11 by intrathecal injection. (D): Evaluation by Western blotting of pluripotent markers Sox2, Oct4, and Notch1 in injured spinal cord of rats treated for 1 week with DMSO or FM19G11 by intrathecal injection. Statistical comparisons were performed with Student's t test, *, p <.05. Abbreviations: BBB, Basso, Beattie, and Bresnahan; DMSO, dimethyl sulfoxide; UCP, uncoupling protein.

Differential Expression of UCP Proteins After Long-Term Intrathecal Administration of FM19G11

Since UCP might have a crucial role in the effects of FM1G11 observed at initial stages in epSPC (Figs. 2A, 5G), we also quantified by ELISA protein expression of UCPs in vivo after long-term FM19G11 intrathecal administration around the damaged medullar tissue. We observed that, UCP1 expression in FM19G11-treated rats was at similar levels in comparison with DMSO-treated animals. However, a significant increment was detected for UCP2 in those animals treated with FM19G11 (Fig. 6C). Moreover, in parallel to UCP2 increased expression, Western blotting analysis showed increased expression of the pluripotent markers Oct4, Notch1, and Sox2 in FM19G11-treated animals (Fig. 6D).

Higher Tissue Recovery of the Injured Spinal Cord After FM19G11 Treatment

In vivo nuclear magnetic resonance (NMR) was used to evaluate the state of the injured spinal cord 1 week after injury with constant FM19G11 supply (Fig. 7A). A smaller injured area was observed in the NMR images 1 week after injury in FM19G11-treated animals. However, 4 weeks after injury, when the animals were sacrificed and the lesioned spinal cords were analyzed, no significant differences were found either on the quantification for scar area (GFAP negative staining) or in the cyst/cavities formation (Fig. 7B). Despite the fact that no significant improvement regarding the scar area and the cycts/cavities was observed, immunostaining analysis of TUJ1, a protein that contributes to microtubule stability in neuronal cell bodies and axons with a relevant role in axonal transport, shows an increase in positive crossing axons through the injured area in those FM19G11-treated rats (Fig. 7C). Moreover, in those animals treated with FM19G11, we detected higher immunostaining of Vimentin, a neuronal precursor marker that stains the ependymal cell population (Fig. 7D).

Figure 7.

In vivo effects of FM19G11 in the regeneration of injured spinal cord tissue. (A): Photomicrographs obtained by in vivo 1H magnetic resonance imaging studies of the region of spinal cord lesion in animals treated for 1 month with DMSO or FM19G11 by intrathecal injection. (B): The area of scar and cysts/cavities in damaged spinal cord tissue of rats treated for 1 week with DMSO or FM19G11 by intrathecal injection was quantified by Meta-Morph Offline software and represented as the mean ± SD. The percentage of scar as well as cysts/cavities is the measure of their area in reference to the total analyzed area from the epicenter (1.5-mm rostral to the lesion). (C): Immunodetection of the astrocyte lineage marker GFAP and the neural lineage marker β-tubulin in the damaged spinal cord of animals treated for 1 week with DMSO or FM19G11 by intrathecal administration. (D): Immunodetection of the astrocyte lineage marker GFAP and the neuronal precursor marker Vimentin in the damaged tissue of animals treated for 1 week with DMSO or FM19G11 by intrathecal administration. Abbreviations: DAPI, 4,6-diamidino-2-phenylindole dihydrochloride; DMSO, dimethyl sulfoxide; GFAP, glial fibrillary acidic protein.


We recently first identified FM19G11 as a new chemical entity that inhibits HIFα protein expression and transcriptional activity under hypoxic conditions repressing a variety of key genes involved in stemness [20]. Our reprogramming experiments showed that this HIFα inhibitor favors oligodendrocyte differentiation under hypoxic conditions, possibly through the modulation of Sox2 and Oct4 expression. More recently, we found a new mode of action of FM19G11. Under standard oxygen tension, in the absence of HIF1α, FM19G11 acts as inductor of mTOR causing an induction of the DNA damage response machinery, cell cycle arrest and cell death in human colon cancer cells [21]. Here, in a primary cell culture of neural precursor cells (epSPC) under normoxia, FM19G11 activates global cellular metabolism, with an increase in glucose uptake that causes improved stem cell renewal mediated by activation of the PI3K/mTOR/AKT and AMPK signaling pathways (also induced in hSKM and hWAdip, Supporting Information Fig. 2). The massive transcriptome analysis performed after micro-array hybridization of FM19G11-treated versus nontreated epSPC showed that there was a global increase in cell metabolism mechanisms. The possible influence of FM19G11 in stem cell metabolism was evaluated and we confirmed an early decline of ATP with a further increase of its biosynthesis possibly by compensatory glycolysis to restore ATP levels [9]. This led to an improved energy status which caused an increase in the proliferation of stem cells. Since mitochondrial UCPs regulate energy expenditure by reducing the synthesis of ATP [8], we evaluated whether FM19G11 may influence their expression and we observed an early upregulation of UCP1 and UCP2, which coincided with the drop of ATP at short times. We observed that after FM19G11 treatment, the profile of ATP supply runs in parallel with the mitochondrial activity observed by MitoTracker as well as the expression of mitochondrial genes in epSPC. There are two canonical ATP/nutrient-sensitive growth-signaling pathways, the kinases AKT/mTOR [32] and the AMP-activated protein kinase (AMPK) [15, 16]. After the initial diminution of mitochondrial activity, we detected a further induction of expression of mitochondrial genes and activity caused by FM19G11 that was blocked by inhibition of mTOR and AMPK pathways with Rapamycin and Compound C, respectively. The results suggest that the modification of the mitochondrial activity caused by the drug, at least partially, depends on the AMPK and AKT/mTOR activation. Energy deprivation through mitochondrial uncoupling triggers an adaptive cellular response that causes an increase of GLUT-4 in the surface of the cell membrane through a reduction in its endocytosis mediated by AMPK or increased exocytosis by AKT [10] causing an increase glucose influx to provide glucose for ATP production. However, further research is needed to disclose the existence of a preferential signaling pathway or the potential interaction between each other after FM19G11 treatment. Some stimuli increase AMPK phosphorylation but its activity is only essential for hypoxia and DNP to elevate cell surface GLUT-4 and glucose uptake [33, 34]. DNP is effective in preventing mitochondrial dysfunction with a plausible protection derived from synaptic (neuronal) and nonsynaptic (Glial and neurons) extracts after acute SCI contusion [13]. In addition, DNP causes an increase in white matter 6 weeks after SCI, even though the benefits caused by DNP in mitochondrial dysfunction start early (24 hours after SCI) [14]. The induction of GLUT-4 was impaired in epSPC and mC2C12 when we blocked both pathways by Rapamycin and Compound C. These results give a preponderant role for AKT and AMPK activation in the effects caused by FM19G11 on GLUT-4 expression as well as on glucose uptake. In addition, GLUT-4 seems to be a key element in mitochondrial adaptation to FM19G11 treatment, since the ablation of this glucose transporter by siRNA impairs the increase in the number of transcripts for mitochondrial genes caused by FM19G11. The results suggest that the increase of mitochondrial activity caused by FM19G11 after 24 hours depends on GLUT-4 and most likely on subsequent glucose uptake and ATP production. Ablation of GLUT-4 impairs the increase in expression of pluripotent markers Sox2 and Oct4 in ependymal stem cells, suggesting a relevant role of this glucose transporter in the maintenance of the undifferentiated state and cell proliferation of epSPCi caused by FM19G11 under normoxia. One of the most challenging objectives in cell therapy is restoring neurological function after SCI. It is possible to restore locomotor activity when epSPC are ectopically transplanted and available in a large quantity [1]. Alternative strategies to cell transplantation come from a wide variety of studies focused on identifying noninvasive therapies to support the increase of undifferentiated stem cell populations which in turn activate self-renewal [35]. The mitochondrial uncoupling protein UCP2 plays a relevant role in the energy-related metabolism of human pluripotent stem cells (hESC and hIPSC) facilitating glycolysis and impairing cell differentiation [23]. These authors showed that the repression of UCP2 is necessary for a proper transition from glycolysis to mitochondrial glucose oxidation that is required for the appropriate differentiation of these stem cells. Therefore, UCP2 is important for maintaining “stemness,” cell proliferation, and viability by impairing glucose oxidation and facilitating glycolysis. Increased expression of UCP2 correlates with neuronal survival, and neurological recovery was enhanced and might be especially useful if produced in the initial phase of neuronal disorders [12]. Results obtained by ELISA indicate a different regulation of protein expression of UCP1 and UCP2 after long-term intrathecal administration of FM19G11. It has been reported that UCP2 transports protons and in consequence exerts uncoupling activity in vitro but apparently not in vivo [36]. Although further studies need to be implemented these results suggest a different role or contribution of UCP1 and UCP2 to the injured tissue depending on the stage of evolution of the SCI. Previous data indicate that activation of AMPK and AKT pathways might contribute to restore neurological malfunction. The AMPK pathway links neuronal function with energy supply [37]. Some AMPK activators are anti-inflammatory and potential therapeutic agents in neurological disorders, particularly when used in combination with drugs, which protect oligodendrocyte cell loss, such as sPLA2 inhibitor [38]. The presence of GFAP-positive astrocytes and microglia could conceivably contribute to the limited self-renewal of the epSPC culture [39]. The effect of AMPK activation might be beneficial for SCI recovery because there would be a lower astroglial contribution to the astrocytic scar [40]. The discovery of drugs that potentiate self-renewal of stem cells and favor their predisposition to oligodendrocyte differentiation to the detriment of astrocyte differentiation might contribute to design pharmacological strategies for spinal cord regeneration. The AKT pathway may also play important roles in motoneuronal survival and nerve regeneration in vivo [41, 42]. It has recently been shown that activation of AKT/mTOR by the compound Bisperoxovanadium exerts a neuroprotective effect after SCI [43]. Thus, FM19G11 might positively contribute to spinal cord regeneration by activation of the AKT/mTOR and AMPK pathways, by favoring the directed differentiation to oligodendrocytes or a neuronal lineage that would replace the loss of functional units and would delay the demyelination process [20], by activation of neuroprotective mechanism of mitochondrial uncoupling [44], and by enhancing self-renewal of stem cells. Thus, we explored whether FM19G11 treatment might contribute to restore neuronal dysfunction as well as functional locomotor recovery in rats with spinal cord damage. The absence of differences in the glial scar formation or the number of cysts/cavities after FM19G11 treatment suggests that additional mechanisms apart from tissue preservation should be involved. Importantly, increased immunostaining of TUJ1 in FM19G11-treated animals indicates a higher axonal growth within the epicenter lesioned area. The increase of the cell marker for neural precursors Vimentin found in FM19G11-treated animals could improve neuronal activity and axonal growth [45]. It has also been suggested that Vimentin can contribute to the formation of scaffolds to facilitate the mature intermediate filament networks. Taken together with our findings, the increase of TUJ1 and Vimentin might contribute to explain the recovery in locomotion observed in animals treated with FM19G11. Thus, FM19G11 is an attractive new compound that may be potentially used to modulate stem cell populations by its amplification or differentiation depending on oxygen rates. Both characteristics endow FM19G11 with enough potential to warrant further studies under strict preclinical regulations.


FM19G11 induces functional regeneration after SCI. FM19G11 confers improved self-renewal capacity of ependymal progenitor stem cells by an early induction of a glycolytic-related response associated to a PI3K/AKT/mTOR signaling induction. Overall, FM19G11 constitute an attractive new compound that may be potentially used to modulate stem cells by its amplification or differentiation, depending on oxygen rates, for cell transplantation or endogenous activation on cell replacement therapeutical applications.


This work was supported by FISS PI10/01683, Instituto de Salud Carlos III (Cofinanciación FEDER), and The Spanish Consolider Ion Channel Initiative [CSD 2008-00005] MICINN grants. We thank Eric Lopez for his excellent technical support and Richard Griffeth for his English-language editing. Ana Alastrué Agudo has not participated in the manipulation or experimentation with human embryonic stem cells.


The authors indicate no potential conflicts of interest.