Differential Regulation of Proliferation and Differentiation in Neural Precursor Cells by the Jak Pathway§


  • Author contributions: Y.H.K. and E.J.B.: conception and design; Y.H.K. and J.-I.C.: provision of study material and collection of data; Y.H.K., H.G.W., Y.-S.J., S.H.L., C.-H.M, H.S.-K., and E.J.B.: data analysis and interpretation; Y.-S.J., S.H.L., C.-H.M, and E.J.B.: financial support; Y.H.K., H.G.W., and E.J.B.: manuscript writing; E.J.B.: final approval.

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

  • §

    First published online in STEM CELLS EXPRESS August 31, 2010.


Neuronal precursor cells (NPCs) are temporally regulated and have the ability to proliferate and differentiate into mature neurons, oligodendrocytes, and astrocytes in the presence of growth factors (GFs). In the present study, the role of the Jak pathway in brain development was investigated in NPCs derived from neurosphere cultures using Jak2 and Jak3 small interfering RNAs and specific inhibitors. Jak2 inhibition profoundly decreased NPC proliferation, preventing further differentiation into neurons and glial cells. However, Jak3 inhibition induced neuronal differentiation accompanied by neurite growth. This phenomenon was due to the Jak3 inhibition-mediated induction of neurogenin (Ngn)2 and NeuroD in NPCs. Jak3 inhibition induced NPCs to differentiate into scattered neurons and increased the expression of Tuj1, microtubule associated protein 2 (MAP2), Olig2, and neuroglial protein (NG)2, but decreased glial fibrillary acidic protein (GFAP) expression, with predominant neurogenesis/polydendrogenesis compared with astrogliogenesis. Therefore, Jak2 may be important for NPC proliferation and maintenance, whereas knocking-down of Jak3 signaling is essential for NPC differentiation into neurons and oligodendrocytes but does not lead to astrocyte differentiation. These results suggest that NPC proliferation and differentiation are differentially regulated by the Jak pathway. STEM CELLS 2010;28:1816–1828


Neural stem cells (NSCs) and neuronal precursor cells (NPCs) can proliferate and differentiate into mature neurons. These cells are essential for brain development and the normal physiological functions of the brain. During the aging process and neurodegenerative conditions, such as Alzheimer's disease, neurogenesis is impaired [1, 2]. The survival, proliferation, and differentiation of NSCs/NPCs are important for neuronal regeneration and functional recovery because NSCs/NPCs can proliferate and differentiate into functional neural cells. Therefore, a better understanding of the fine regulatory mechanism of neuronal differentiation from progenitors is required.

Environmental factors are important for NPC survival and maintenance during development. Growth factor (GFs), such as epidermal growth factor (EGF) and basic fibroblast growth factor (bFGF, also abbreviated FGF-2) are particularly critical for NSC/NPC maintenance and expansion [3]. EGF and bFGF promote NSC/NPC proliferation, maintaining the undifferentiated state [4]. These GFs are also necessary for further differentiation. The removal of EGF and FGF-2 from neurosphere cultures of a human NPC line for short periods of time increases neuronal differentiation and neurite extension [5]; however, mechanism is not yet well-characterized.

In the normal developmental process, NSCs initially proliferate, subsequently giving rise to neurons first and glial cells later. During the proliferation process, notch signaling is activated, which provokes the expression of repressor-type basic helix-loop-helix (bHLH) genes Hairy and Enhancer of split homolog (Hes)1 and Hes5. Hes1 and Hes5 repress activator-type genes, including mammalian achaete-scute complex homolog (MASH)1, mammalian atonal homolog (MATH)1, and neurogenin (Ngn), which are known to promote neuronal fate determination [6]. In differentiating neurons, notch is not activated, and the activator-type bHLH factors are expressed, inducing Hes6 expression. This, in turn, inhibits Hes1 function and reinforces the neurogenic process [7]. However, it is not clear how NPC proliferation and differentiation are finely regulated.

The molecular switch regulating the transition from neurogenesis to gliogenesis in NPCs is internally programmed [8]. The timing between neurogenesis and astrogliogenesis is controlled by several transcription factors. Ngn1, a bHLH transcription factor, drives neurogenesis by activating the expression of neuronal genes and inhibits astrocyte differentiation [9]. Moreover, the gp130-Jak-STAT pathway, which is activated by the neurotrophic cytokine cardiotrophin-1, is an essential mechanism regulating the neurogenic-to-gliogenic transition for timed astrocyte formation [10]. The members of the Jak/STAT family of proteins modulate gene expression during the different stages of brain maturation [11]. Among these factors, Jak2 is known to play a critical role during astrogliogenesis involving the survival pathway of erythropoietin and ciliary neurotrophic factor (CNTF) in the chick embryo [12–14]. However, the expression and role of Jak3 in brain development have not yet been studied. Jak3 is abundant in the hematopoietic system and plays an important role in hematopoiesis and inflammation [15], which implies a role during neurogenesis via inflammatory processes [16]. In addition, a specific Jak3 inhibitor has been shown to aid survival in a transgenic lateral amyotrophic sclerosis mouse model [17]. These results suggest that Jak3 plays a critical role in neuronal differentiation. Thus, we investigated the role of Jak3 in NPC proliferation and differentiation during neurogenesis and compared its involvement with that of Jak2.



All pharmacological signaling pathway inhibitors AG490 and WHI-p154 were obtained from Calbiochem (San Diego, CA, http://www.emdbiosciences.com).

Cortical Primary Neurosphere Cultures

All animal experiments were performed following Ajou University institutional guidelines. For neurosphere cultures, brains were isolated from imprinting control region mouse (Korea Bio-link, South Korea, http://www.dhbiolink.net) E13 embryos. Briefly, cortices were freed of meninges prior to mechanical dissociation and gentle trituration several times in culture medium using a flame-polished Pasteur pipette. Cells were plated in a flask at a cell density of 2 × 105/ml. The NPCs were cultured in Dulbecco's modified Eagle's medium (DMEM)/F12 media (Gibco, Carlsbad, CA, USA, http://www.invitrogen.com) supplemented with 5 mM HEPES (Sigma, St. Louis, MO, USA), 20 ng/ml bFGF (Invitrogen, Carlsbad, CA, USA, http://www.invitrogen.com), 20 ng/ml EGF (Invitrogen), and N-2 supplements (insulin, transferrin, progesterone, putrescine, and selenite, 10 mM; Invitrogen). The media was replaced every 2 days. For secondary neurosphere formation assays, primary neurospheres were collected and dissociated. A fraction (400 μl) of the cell suspension (2 × 105 cells/well) was transferred to each well of a 24-well plate in culture medium containing FGF2 and EGF. Neurospheres were passaged every 6 days and dissociated with Accutase (Invitrogen). The subcultured neurospheres were used as neurospheres themselves or redissociated single cells for differentiation in plates coated overnight with 100 μg/ml poly-D-lysine (Sigma-Aldrich, St. Louis, MO, http://www.sigmaaldrich.com) and 4 μg/ml laminin (Sigma-Aldrich) in DMEM/F12 media (Gibco). Cells treated with Jak inhibitors were maintained at 37°C in 5% CO2 incubators.

Assessment of Neural Progenitor Cell Differentiation In Vitro

Under a light microscope, the migration of NPCs from the neurosphere was evaluated after 24 hours. The plated neurospheres had a diameter of approximately 100 μm, after growing for 5 days in GF media. The neurosphere diameter was calculated by Axio vision (Carl Zeiss, Peabody, MA, USA, http://www.zeiss.com). For neurite outgrowth assays, purified single NPCs were cultured for 2 days with Jak2 or Jak3 inhibitor. The length of neurites in MAP2-positive cells was also calculated by the Axio vision program. Differentiated cells were immunostained and counted in four to five randomly selected fields, with 200–500 cells counted in each field. All experiments were performed in triplicate.

Western Blot

NPCs were grown in coated 6-well plates and cultured until indicated, as described earlier. NPCs were lysed in lysis buffer (100 μl/well; 20 mM Tris HCl, 10 mM NaCl, 30 mM Na4O7P2 (pH 7.2), 30 mM NaF, 1 mM Na3VO4, 1% sodium deoxycholate, and 1% triton X-100) containing 1 mM EDTA, 10 μg/ml leupeptin, 0.5 mM phenylmethylsulfonyl fluoride, 1 μg/ml pepstatin, and 10 μg/ml aprotinin. The cell lysates were further homogenized by brief sonication and centrifuged at 14,000g for 20 minutes at 4°C. The resulting supernatant was collected. To isolate proteins from the brain tissues, the brain cortices were isolated and stored at −70°C overnight, and then homogenized in 1 ml ice-cold buffer containing 50 mM Tris HCl (pH 7.4), 1 mM sodium orthovanadate, 150 mM NaCl, 1% triton X-100, 5 mM EDTA, 0.5% sodium deoxycholate, 0.1% SDS, 1 mM MgCl2, and protease inhibitors. Homogenates were centrifuged at 15,000g for 30 minutes at 4°C. The supernatants were collected, and the protein concentration determined using bicinchoninic acid reagent. Equal amounts of protein were subjected to 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), and the separated proteins were electrophoretically transferred to polyvinylidenedifluoride (PVDF) membranes. The blot was blocked with 5% nonfat dried milk, incubated overnight with anti-Jak2 (Cell Signaling Technologies, Beverly, MA, http://www.cellsignal.com, 1:1,000), anti-pJak2 (Cell Signaling Technology, 1:1,000), anti-Jak3 (Santa Cruz Biotechnology, 1:1,000), anti-pJak3 (Santa Cruz Biotechnology, CA, http://www.scbt.com, 1:1,000), anti-nestin, anti-MAP2 (Chemicon, Billerica, MA, http://www.millipore.com, 1:1,000), anti-Tuj1 (Chemicon, 1:1,000), anti-glial fibrillary acidic protein (GFAP; Sigma-Aldrich, 1:500 dilution), or anti-neuroglial protein (NG)2 (Chemicon, 1:1,000). After washing with PBS containing Tween-20 (PBST), a 1:5,000 dilution of secondary horseradish peroxidase-conjugated antibody in PBST was applied for 1 hour. Chemiluminescent signals were quantified by scanning the PVDF membranes with the LAS-1000 system (Fuji) using Image Gauge software (version 3.12). To normalize the protein content in each line, blots were stripped and reprobed with anti-β-actin antibody (Sigma-Aldrich) in blocking solution.


NPCs were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) for 10 minutes and washed with PBS containing 0.1% triton X-100. Antibodies against nestin, Ki67 (BD Biosciences, San Jose, CA, http://www.bdbiosciences.com), Tuj1, MAP2, NeuN (Chemicon), NG2, and GFAP were diluted 1:500 in PBS. The cells were incubated with a primary antibody overnight at 4°C, washed with PBS containing 0.1% triton X-100, and incubated for 1.5 hours at room temperature with either a 1:1,000 dilution of goat anti-mouse IgG (Alexa Fluor 594; Invitrogen or Molecular probes, Eugene, OR, http://probes.invitrogen.com) or 1:1,000 dilution of goat anti-rabbit IgG (Alexa Fluor 488; Invitrogen or Molecular probes). Finally, the cells were washed with PBS, counterstained with 300 nM Hoechst for 5 minutes at room temperature and washed with PBS. Images were acquired with the Axio Vision photomicroscope and confocal microscopy (LSM, Carl Zeiss). For brain immunostaining, formalin-fixed paraffin-embedded sections of embryo brain were dissected with a microtome (RM2245, Leica). The fixed cells were incubated for 30 minutes with a solution containing 5% bovine serum albumin to block nonspecific antibody binding.

Reverse Transcription Polymerase Chain Reaction

Total RNA was extracted from cultured NPCs and embryonic brain tissue at different developmental stages using Easy Blue (Intron, South Korea, www.intron.co.kr). First-strand cDNA was synthesized from 500 ng of total RNA in a 20 μl reaction with random (dT) primers. The same amount of first-strand cDNA from each sample was used in polymerase chain reaction (PCR) amplification to detect the spatial and temporal expression of RNAs using specific primers for Jak3, nestin, Tuj1, Otx2, Olig2, NCAM, GFAP, Hes1, Hes6, Ngn2, and NeuroD (see Supporting Information Table 1).

Assessment of Small Interfering RNA Tansfection

Small interfering RNAs (siRNAs) targeting Jak2 and Jak3 (NCBI Reference Sequence: NM_010589) were obtained from Bioneer, South Korea, www.bioneer.co.kr: Jak3-s, 5′-GGUUAUACUCAUGGCACCU-3′, Jak3-as, 5′-AGGUGCCAUGAGUAUAACC-3′; Jak2-s, 5′-GUGGUAUUACGCCUGUGUA-3′, Jak2-as, 5′-UACACAGGCGUAAUACCAC-3′. Jak2 and Jak3 siRNA were combined to a final concentration of 3 or 10 nM. Neurosphere dissociation and transfection were performed according to the protocols described earlier [18]. Cells (2 × 105 cells/well) were transfected with negative control (10 nM; Santa Cruz, sc-37007), Jak2 or Jak3 siRNA using a Microporator MP-100 (1,100 V, 40 ms with a single pulse; Digital Bio Technology) and seeded in a 24-well plate. Two days later, the cells were induced to differentiate on a 24-well plate coated with poly-D-lysine/laminin and fixed to analyze the expression patterns of specific proteins. After eletroporation of mock-fluorescein isothiocyanate (10 nM; Bioneer) into 50-μm neurospheres, fluorescence was detected at 4 hours with the Axio Vision photomicroscope.

Statistical Analysis

The data were expressed as the mean ± SE from 3 to 5 independent experiments. An analysis of variance (ANOVA) followed by Dunnett's multiple comparison test was used for statistical comparisons (Sigma Stat, Jandel Scientific). A p value <.05 was considered significant.


Effect of Jak3 Inhibitor on Neurosphere NPCs

In the present study, two types of neurosphere NPC cultures were used: subcultured neurospheres and single redissociated NPCs. NPCs obtained from E13 embryo brain cortices by floating neurospheres robustly proliferated in GF-rich (EGF/bFGF) media. These growing neurospheres were subcultured for experiments with or without redissociation into single cells. The NPCs were then cultured on poly-D-lysine and laminin-coated tissue culture plates with GF-rich media. In the neurosphere cultures without redissociation, NPCs initially differentiated into neurons as observed by Tuj1 and Map2 staining. The differentiated cells migrated from the neurosphere core, but the redissociated NPCs proliferated with cluster formation and concomitantly underwent neuronal differentiation, generating neuronal sprouts. After this point, nestin-positive NPCs survived for more than 10 days when given a continuous supply of fresh GF. After 4 days, the NPCs differentiated into NG2-/O4-positive and GFAP-positive glial cells (Fig. 1A).

Figure 1.

Jak3 Inhibition in NPC culture. (A): Subcultured neurospheres or redissociated NPCs from neurospheres were cultured in a coated plate with growth factor-rich media. NPCs proliferated in a cluster, differentiating first into neurons and later into glial cells. (B): Subcultured neurospheres without redissociation were plated onto coated plates and WHI-p154, a Jak3 inhibitor, added at various concentrations (0, 1, 3, and 10 μM) for 24 hours. The NPCs began to differentiate on the coated plate, which was increased by WHI-p154 in a dose-dependent manner. The differentiated neurons developed long neurites and migrated further from the core. (C): After 18 hours, neurospheres were stained for Tuj1 (green) and pJak3 (red). Jak3 inhibition decreased pJak3 expression and increased neurite growth on Tuj1-positive cells in a dose-dependent manner. The NPCs with downregulated pJak3 expression at the outer portion of the neurosphere showed more differentiation. Scale bar = 100 μm. Abbreviation: NPC, neuronal precursor cell.

The Jak3 inhibitor WHI-p154 induced NPC differentiation and migration in a dose-dependent manner (Fig. 1B). The NPCs with processes moved further from the core of neurospheres. The NPC processes were nestin-positive and Tuj1-positive. NPC spheres were immunochemically stained with Tuj1 and pJak3 18 hours after seeding (Fig. 1C). The inhibition of Jak3 increased the number of Tuj1-positive neurite extensions and concomitantly decreased pJak3 expression. NPCs treated with a higher concentration of WHI-p154 showed more differentiation with more and longer extensions. The NPCs in the outer portion of the neurospheres that were downregulated by pJak3, differentiated earlier than NPCs in the core, relative to increased pJak3. Therefore, NPCs with more Jak3 inhibition underwent more differentiation.

Effect of Jak3 Inhibition on Redissociated NPCs

The effect of Jak3 inhibition was investigated by transfecting redissociated NPCs with Jak3 siRNA. The transfected NPCs were cultured on poly-D-lysine and laminin-coated plates. Jak3 knockdown induced earlier differentiation compared with the control NPCs (Fig. 2A). Treatment with pJak3 inhibitor WHI-p154 at various concentrations (0.1, 1, and 3 μM) also induced more differentiation. After 2 days, NPCs grew in clusters, which might be due to proliferation and differentiation under GF-rich conditions. When the NPCs were treated with a higher concentration of WHI-p154, pJak3 was expressed at a lower level with more concomitant differentiation (Fig. 2B). WHI-p154 significantly inhibited the expression of pJak3 protein as shown by the western blot analysis (Fig. 2C).

Figure 2.

Effect of pJak3 inhibition on redissociated NPCs. (A): Single NPCs were purified and transfected with siRNA against Jak3 by electroporation, which induced differentiation and neurite extension on poly-D-lysine and laminin-coated plates. After 18 hours, reverse transcription polymerase chain reaction analysis revealed Jak3 expression in transfected NPCs. Scale bar = 25 μm. (B): On coated plates, WHI-p154 reduced the expression of pJak3, and the NPCs with downregulated pJak3 showed more differentiation. (C): In the western blot, pJak3 expression decreased with time, and WHI-p154 reduced the expression. (D): NPCs were cultured for 1, 4, and 6 days and double-stained for nestin and MAP2 and GFAP and neuroglial protein 2. The proportion of immunochemically stained cells relative to nuclei was counted in culture days. Scale bar = 50 μm. Abbreviations: GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GFAP, glial fibrillary acidic protein; MAP2, microtubule-associated protein 2; NPC, neuronal precursor cell.

Next, we investigated the time-dependent regulation of differentiation by Jak3 inhibition. After 1, 4, and 6 days in GF media, NPCs were double-stained with MAP2/nestin and NG2/GFAP. The NPCs initially grew in a cluster. Nestin-positive cells appeared in the cluster first, followed by MAP2-positive cells. Astrocytes grew abundantly after 4 days with simultaneous increase in NG2-positive cells, which were known to develop into oligodendrocytes [19]. Treatment with WHI-p154 significantly increased the proportion of MAP2-positive cells from day 1, and increased the proportion of NG2-positive cells later. However, the proportion of nestin-positive and GFAP-positive astrocytes decreased (Fig. 2D).

The first day after seeding, nestin-positive cells were observed in the cluster and Tuj1-positive cells with neurites began to appear. On day 3, the NPC cluster was composed of round nestin-positive cells, and MAP2-positive cells appeared at the outer boundary of the cluster, implying that the differentiated cells migrated further from the core. Jak3 inhibition changed the morphology of the nestin-positive cells into long spindle-shaped cells on day 1, which occurred later in the control group (Fig. 3). On day 4, nestin-positive cells had a spindle-shape with processes, which were present in the cluster. However, WHI-p154 induced NPC sprouting in a dose-dependent manner with a scattered appearance. On day 8, all nestin-positive cells in the control cultures showed a spindle shape. Jak3 inhibition increased the number of NG2-positive cells, as shown by mature oligodendrocytes with more processes. GFAP was highly expressed, but neurons looked feeble under EGF/bFGF-rich conditions. Jak3 inhibition differentiated NPCs into MAP2-positive neurons with bulging cytoplasm and longer neurites and significantly decreased the number of GFAP-positive cells in a dose-dependent manner (Fig. 3).

Figure 3.

Effect of WHI-p154, a Jak3 inhibitor, on neuronal precursor cells (NPCs) over time in culture. Single NPCs were cultured on a coated plate. After 1 day, NPCs grew in a cluster. Nestin-positive NPCs were present in the cluster and Tuj1-positive neurons appeared scattered at the boundary of the cluster. The presence of WHI-p154 induced the process of nestin-positive cells and MAP2-positive cells. After 4 days, the cluster appeared in the control group and Jak3 inhibitor significantly increased the number of MAP2-positive cells. After 8 days, the expression of NG2 and GFAP was prominent in the control group. The number of NG2-positive cells increased, and their prominent processes differentiated in the presence of Jak3 inhibition. Neurogenesis was apparent, but astrogliogensis was diminished by Jak3 inhibition. Scale bar = 50 μm. Abbreviations: GFAP, glial fibrillary acidic protein; MAP2, microtubule associated protein 2; NG2, neuroglial protein 2.

Jak2 and Jak3 Comparison

Jak2 is known to play a critical role in astrogliogenesis; therefore, we compared the effects of Jak2 and Jak3 inhibition during neurogenesis. We performed reverse transcription PCR studies on subcultured NPCs treated with AG490 or WHI-p154 for 18 hours. AG490 is a potent Jak2 inhibitor and significantly reduced nestin, Otx2, Tuj1, GFAP, Olig2, and NCAM expression in a dose-dependent manner. In contrast, Jak3 inhibition significantly increased Tuj1, Olig2, and NCAM mRNA levels and slightly decreased nestin and GFAP mRNA levels (Fig. 4A). Protein expression was analyzed by western blot after treatment with Jak2 or Jak3 inhibitor for 2 days. Both AG490 and WHI-p154 significantly reduced nestin and GFAP expression. However, WHI-p154 significantly increased the expression of Tuj1, MAP2, and NG2, which were not altered by AG490 over the same period of time (Fig. 4B). Over the course of the 2 days, some NPCs differentiated with neurite formation. WHI-p154 induced differentiation and the NPCs had longer MAP2-positive neurites (Fig. 4C). In the control group and AG490-treated group, the main population of cells had shorter neurites (less than 10 μm), whereas the longest neurites (>50 μm) were observed in the WHI-p154 treated group. Jak3 inhibition significantly increased the number of MAP2-positve cells and the neurites length.

Figure 4.

Comparison of Jak3 and Jak2 inhibition. Neuronal precursor cells (NPCs) were treated with the indicated doses of Jak3 inhibitor (WHI-p154) and Jak2 inhibitor (AG490) for 18 hours in growth factor (GF)-enriched media. (A): Reverse transcription polymerase chain reaction analysis showing that, after 18 hours, neuronal differentiation genes Tuj1, Olig2, and NCAM were upregulated by Jak3 inhibition but not Jak2 inhibition. (B): Single NPC was treated with the indicated doses of WHI-p154 and AG490 for 2 days in GF-enriched media. Jak3 inhibition significantly increased the expression of neuronal genes Tuj1 and MAP2 and preoligodendrocyte gene NG2, but decreased the expression of the astrocyte gene GFAP. (C): After 2 days, NPCs grew in clusters due to the proliferation. AG490 generated the smaller neurospheres and WHI-p154 induced the differentiation into cells with neurites. The length of the neurites was measured and numbers of MAP2-positive cells counted. (D): NPCs treated with Jak2 or Jak3 inhibitor were immunochemically stained after 2 days. NPCs growing in clusters were nestin-positive and they were decreased in the presence of Jak2 inhibitor, but differentiated in the presence of Jak3 inhibitor. The number of MAP2-positive neurons with prominent neurites increased in the presence of Jak3 inhibitor, but the expression of GFAP was scarce. (E): After 6 days, NPCs were stained, counted, and statistically analyzed. The number of Ki67-positive and nestin-positive NPCs profoundly decreased in the presence of Jak2 inhibition. Jak3 inhibition increased the numbers of NG2- and MAP2-positive cells, but it decreased the number of GFAP-positive cells. Scale bar = 50 μm. Abbreviations: CTL, control; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GFAP, glial fibrillary acidic protein; MAP2, microtubule-associated protein 2; NCAM, N-cadherin adhesion molecule; NG2, neuroglial protein 2.

In GF-rich media, NPCs grew in clusters due to proliferation. NPCs treated with AG490 generated smaller neurospheres and exhibited less differentiation than untreated controls. In contrast, WHI-p154 induced all NPCs to differentiate into cells with process. With Jak3 inhibition, nestin-positive cells had processes, the number of MAP2-positive neurons with longer neurites increased, and GFAP-positive cells concomitantly decreased, though GFAP expression at this time was very weak, even in the control group (Fig. 4D).

After 6 days, nestin-positive cells exhibited a spindle shape, and MAP2-positive cells and GFAP-positive astrocytes were well-differentiated. Jak2 inhibition significantly reduced the number of cells positive for nestin, MAP2, NG2, or GFAP in a dose-dependent manner. However, Jak3 inhibition increased the number of MAP2-positive and NG2-positive cells but decreased the number of nestin-positive and GFAP-positive cells. In particular, the NG2-positive cells treated with Jak3 inhibition showed the morphology of mature oligodendrocytes with abundant processes, and these cells were positive for O4, a mature oligodendrocyte marker (Fig. 4E).

To confirm our observations, we compared the roles of Jak2 and Jak3 in NPC neurospheres using siRNAs-mediated knockdown. NPCs transfected with a higher concentration of Jak3 siRNA expressed lower levels of Jak3 mRNA. Most NPCs were transfected, though some had less siRNA and green fluorescence (Fig. 5A). The Jak3 knockdown cells differentiated earlier than mock-transfected NPCs. Transfected NPC neurospheres were grown in GF-rich media and then submitted to GF withdrawal. The NPCs differentiated first into neurons and then into glial cells. Jak3 inhibition resulted in increased neuronal differentiation. NPCs with greater Jak3 inhibition showed more neuronal differentiation than astroglial differentiation. The length of MAP2-positive neurites was longer in the presence of Jak3 inhibition (Fig. 5B).

Figure 5.

Effect of Jak2 and Jak3 siRNA transfection. (A): A photo of neurospheres mock-transfected with FITC. Neurospheres transfected with 3 nM and 10 nM siRNA were analyzed by reverse transcription polymerase chain reaction after 18 hours. (B): On coated plates, neuronal precursor cells (NPCs) from neurospheres transfected with Jak3 siRNA had more MAP2 (red) staining with long neurites and less GFAP (green) staining 3 days after transfection. Nuclei were counterstained with Hoechst (blue). The length of the neurites further increased with a higher concentration of siRNA. (C): NPC spheres were transfected with Jak3 or Jak2 siRNAs. Immunostaining showed that Jak2 inhibition reduced Ki67, nestin, Tuj1, NG2, and GFAP expression. However, Jak3 inhibition increased the differentiation of Tuj1- and NG2-positive cells. Nuclei were counterstained with Hoechst (blue). Scale bar = 50 μm. *, p < .05 versus control. Data are representative of more than three independent experiments. Abbreviations: FITC, fluorescein isothiocyanate; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GFAP, glial fibrillary acidic protein; NG2, neuroglial protein 2.

We also examined whether Jak2 and Jak3 regulate the proliferation properties of NPCs. NPCs transfected with Jak2-targeted siRNA reduced the number of Ki67-positive cells, whereas Jak3 inhibition resulted in NPC scattering and reduced their ability to proliferate less than Jak2 inhibition (Fig. 5C). The effects of Jak2 and Jak3-targeted siRNAs on the expression of nestin, Tuj1, NG2, and GFAP were the same as the effects seen with AG490 and WHI-p154 treatment. This finding suggests that Jak2 inhibition reduces the proliferation of NPCs and inhibits the further differentiation of NPCs, whereas Jak3 inhibition initiates neuronal differentiation, promotes oligodendrogliogenesis, and blocks sequential astrogliogenesis.

Interplay with Intrinsic Regulators of Neurogenesis

To observe the effect of Jak3 inhibition on the initiation of differentiation, we investigated the effects of Jak3 inhibition on intrinsic regulators of neurogenesis. Hes1 is an important repressor-type bHLH gene in neurogenesis that inhibits neurogenic bHLH factors such as MASH, MATH, and Ngn [20]. Hes6, a Hes1 inhibitor, is induced during neurogenesis [7]. Hes1 is also an important transcription factor for stemness in NSCs and astrocyte differentiation [21]. Jak3 inhibition decreased Hes1 and increased Hes6 expression levels (Fig. 6A). We also examined the expression level of neurogenic markers Ngn2 and NeuroD after Jak3 and Jak2 inhibition. Jak3 inhibition induced the expression of these markers, whereas Jak2 inhibition decreased their expression (Fig. 6B). These data strongly suggest that Jak3, but not Jak2, plays a critical role in the inhibition of neuronal differentiation.

Figure 6.

Jak3 inhibition triggers cascades of neuronal-differentiation genes. (A): Jak3 inhibitor downregulated the expression of Hes1, a neurogenesis repressor, and increased the expression of Hes6, a Hes1 inhibitor. (B): Neurogenic-specific factors, such as Ngn2 and NeuroD, were induced by Jak3 inhibition. These effects were not observed with Jak2 inhibition. *, p < .05 versus each control. Abbreviations: CTL, control; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; Ngn2, neurogenin 2.

In Vivo Expression of Jak3 in the Brain During Development

To determine pJak3 expression during neurogenesis in the brain, the brains of E12 embryos were prepared and oriented on the coronal plane. We found pJak3 to be coexpressed with the neural progenitor marker nestin (Fig. 7A), which was expressed in an area near the lateral ventricle and the edge of the cortical plate and lateral ganglionic eminence. The expression of pJak3 occurred in similar area, as well as the cytosol, as observed by high magnification. Tuj1, an immature neuronal marker, was also expressed in the E12 brain, though little colocalization of nestin and pJak3 was observed.

Figure 7.

In vivo expression of Jak3 in the developing brain. (A): In the coronal plane of an E12 brain cortex, nestin, pJak3, and Tuj1 protein were expressed. (B): In E11-E15, P6, and 21-day-old mouse brain cortices, Jak3 mRNA was expressed. The mouse spleen was used as a positive control. (C): Tissue homogenates derived from cerebral cortices of E11-E15, P6, P21, and adult (3-months old) were subjected to SDS-PAGE using antibodies against phospho-Jak3, Jak3, phospho-Jak2, and Jak2. Abbreviations: CP, cortical plate; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; LGE, lateral ganglion eminence; LV, lateral ventricle.

To show the developmental expression patterns of Jak3 and pJak3, cortices were collected from the immature brains of E11, E12, E13, E15, postnatal day 6 (P6) and P21 mice, as well as the mature brain and spinal cord of a 3-month-old mouse. Jak3 was present in the brain and spinal cord, and it was also found in the spleen and thymus, where it is known to be abundant. Interestingly, Jak3 mRNA expression fluctuated throughout development. From E11 to E15 and P6, Jak3 mRNA expression increased, but it decreased until adulthood. The expression of Jak3 mRNA in the mature brain and spinal cord was much less than the expression found in the immature brain and spinal cord (Fig. 7B).

Phosphorylated Jak3 protein was also expressed in the brain at E11 and increased until P6, followed by a decrease after P6 and downregulated in the adult brain (Fig. 7C), similar to the mRNA observations. Most cortical cells in the E13 brain were NSCs or NPCs. The developmental expression of pJak3 in brain cortices coincided with cerebral cortex development. To expand the brain cortex, NPCs need to proliferate and differentiate into functional neurons and other cells, such as astrocytes and oligodendrocytes. During this period, Jak3 and Jak2 levels and phosphorylated forms were increased. Therefore, Jak3 and Jak2 were activated during neurogenesis/astrogliogenesis, because they are phosphorylated during activation. However, the expression of pJak3 decreased after P6, whereas pJak2 increased until the adulthood.


During cortical development in the mammalian brain, neural stem cells (NSCs) located within the periventricular generative zones give rise to migrating chains of neurons and radial glial cells, followed by oligodendrocytes and astrocytes [8]. To organize precisely timed and progressive maturation, the intrinsic program of NSCs and appropriate extrinsic environmental cues are essential [22, 23], but the molecular and cellular regulatory mechanisms have not yet been defined.

During early neurogenesis, NPC proliferation and subsequent differentiation and migration are essential processes for cerebral cortex development. These processes require a continuous and adequate supply of nutrients and GFs. Specifically, nestin-positive NSC/NPCs must maintain their proliferative capacity to expand the volume of the telencephalon [24–26]. NSC self-renewal is only maintained in the presence of an appropriate supply of GFs, such as EGF and FGF, which maintains an undifferentiated state with proliferative capacity and the ability to differentiate into neurons and all types of glial cells. EGF and FGF are known to be critical factors for NSC survival and proliferation. Targeted deletion of these EGF and FGF receptors causes forebrain cortical dysgenesis and embryonic lethality [3, 27]. In the present study, NSC/NPCs survived only in the presence of a continuous supply of EGF/bFGF.

Nestin-positive NPCs can proliferate, as demonstrated by Ki67 staining. Compared with Jak3 inhibition, Jak2 inhibition was more effective at reducing GF-induced proliferation. One possible explanation is that Jak2 signaling may be more important for GF-mediated proliferation. In ST14A cells, Jak2, but not Jak1 or Tyk2, is susceptible to activation through the stimulation of an exogenously expressed cytokine receptor, ultimately leading to cell proliferation [28]. Jak2-knockout mice are not viable, whereas Jak3 knockout mice are viable but severely immunocompromised. Therefore, in the present study, Jak2 may be more important for NSC survival and proliferation, and Jak3 inhibition induced precocious neuronal differentiation of the NPCs.

NPCs have the potential for self-renewal and to differentiate into neurons and these cells are finally internally programmed to be astrocytes. During this process, intrinsic cues and environmental factors play important roles. Among extrinsic factors, cell-cell interactions mediated by Notch signaling, bFGF, and CNTF are well known [29]. CNTF in particular regulates the neuronal and glial differentiation of retinal stem cells by concentration-dependent recruitment of the mitogen-activated protein kinase (MAPK) and Jak/STAT pathway [30]. Jak2/STAT signaling is known to be critical for astrogliogenesis during the neurogenesis-to-astrogliogenesis transition. However, in our study, the proliferation and differentiation of NPCs were differentially regulated by Jak2 and Jak3. During the determination of lineage into neurons/polydendrocytes versus GFAP-positive astrocytes, Jak3 signaling is decisive.

After the early stages of NSC expansion and self-renewal, NSCs differentiate and mature into neurons, astrocytes, and oligodendrocytes in a lineage-restricted manner. Proneural bHLH proteins, such as Ngn2 and NeuroD, are key regulators of neurogenesis. Jak3 inhibition reduced the expression of Hes1, a mammalian bHLH transcriptional repressor that inhibits neuronal differentiation [20]. In contrast to Hes1, Hes6 promotes neuronal differentiation by antagonizing Hes1 function [7]. In the present study, Jak signaling was associated with the regulation of Hes1 and Hes6, which directly determined the fate of NSCs. Therefore, Jak3 inhibition promotes the exit from self-renewal in NSC/NPCs and finally induces precocious differentiation into neurons by directly regulating Hes1 and Hes6.

In the present study, Jak3 inhibition first accelerated neurogenesis and later blocked astrogliogenesis. This effect can also be explained by the effect of Jak3 inhibition on the Hes genes. The inactivation of Hes1 leads to the upregulation of proneural genes, acceleration of neurogenesis, and premature depletion of NSC [25]. Conversely, the overexpression of Hes1 leads to the inhibition of neurogenesis and the maintenance of NSCs [31, 32]. At later stages of development, Hes1 promotes gliogenesis [20]. During neurogenesis, the “neurons first-glia second” principle seems to be valid; however, the mechanism regulating this process is not fully understood. Inducers of neuronal differentiation, such as retinoic acid, block premature astroglial differentiation if present at defined stages [33].

Neurogenic transcription factors, including MASH, MATH, and Ngn, inhibit astrogliogenesis by inhibiting the differentiation of primed glial progenitors [25]. Hes1 inhibits neuronal differentiation and promotes astrogliogenesis [31]. In the present study, Jak3 inhibition directly regulated Hes1 and Hes6 expression levels and induced neurogenic factors such as Ngn and NeuroD. Therefore, Jak signaling is critical for the maintenance of NSCs and determining the neuronal cell lineage fate.

The effect of Jak3 inhibition on neurite outgrowth was very prominent. FGF and adhesion molecules are essential environmental factors for the protrusion of neurites [34]. Adhesion molecules, such as L1, cadherins, and NCAM, are known to be associated with neurite growth [35]. These adhesion molecules are linked with the FGF receptor, which regulates the intracellular signaling associated with the rearrangement of cytoskeletal proteins [36, 37]. Jak3 inhibition remarkably induced neurite outgrowth in cells grown in any media condition, with or without FGF, this was also true when cells were grown on plates with or without a poly-D-lysine or laminin coating. This observation is in contrast to Jak2 inhibition, which had little effect on neurite protrusion. Jak3 inhibition significantly increased the expression of adhesion molecules, such as NCAM, with prominent neurite growth, and it induced adhesion molecules to bind the FGF receptor. Receptor activation may be associated with survival and later with neurite growth. In other stem cell studies, cell surface molecules, which include cell adhesion molecules, integrins, selectins, and immunoglobulins, have been shown to regulate cell commitment within different tissue microenvironments [34, 38].

NG2-positive cells, also known as polydendrocytes, have been equated with oligodendrocyte precursor cells [39]. NG2-positive cells differentiate into oligodendrocytes during development and their persistence is thought to mark the multipotent cells that can give rise to neurons and astrocytes but are not colocalized with GFAP-positive cells. Astrocytes, which originate from polydendrocytes, are protoplasmic astrocytes in gray matter and are not fibrous astrocytes in white matter [40]. Direct evidence exists that endogenous NG2 polydendrocytes give rise to oligodendrocytes in vivo [40, 41], but it is debatable as to whether polydendrocytes generate neurons. In the adult central nervous system, the core protein of NG2 directly binds to platelet-derived growth factor as well as bFGF [42, 43]. NG2-positive cells undergo proliferation and morphological changes in response to demyelination or inflammation [44]. Both EGF and bFGF promote the proliferation of NSCs. bFGF has the potential to induce differentiation of the NSCs into oligodendroglial lineage, but the number of differentiating neurons is decreased [45]. In the present study, neurons were feeble in the presence of EGF and bFGF, but neurons induced by Jak3 inhibition were well differentiated. Though the inhibition of Jak3 induced differentiation into neurons and NG2-positive cells, it blocked differentiation into astrocytes. The NG2-positive cells induced by Jak3 inhibition did not overlap with Tuj1 or MAP2-positive cells and GFAP astrocytes, which exhibited morphology distinct from the other cells. The effect of Jak3 inhibition suggests the presence of polydendrocytes, which give rise to neurons and oligodendrocytes but not GFAP-positive astrocytes. However, in the present study, whether the NG2-positive cells are the multipotent cells that give rise to neurons and oligodendrocytes is uncertain. Certainly, the NG2-positive cells looked like oligodendrocyte precursor cells, and Jak3 inhibition significantly increased the number of NG2 positive cells, which had the morphology of mature oligodendrocytes with abundant processes. In the present study, Jak3 inhibition increased the portion of Tuj1-positive and NG2-positive cells, which are strongly implicated as potential repair targets. Jak3 inhibition triggered the process of polydendrocytes development instead of being blocked into fibrous astrocytes, which might explain the beneficial effects of Jak3 inhibitors in the amyotrophic sclerosis model [17].


Jak2 is responsible for the proliferation and further differentiation of NPCs. In contrast, the off-on Jak3 signaling finely regulates the initiation of neuronal differentiation and astrogliogenesis. NPCs can survive, proliferate, and differentiate over time in an appropriate local environment via Jak2 and Jak3 signaling.


This study was supported by the Korea Science and Engineering Foundation through the Chronic Inflammatory Disease Research Center (R13-2003-019).


The authors indicate no potential conflicts of interest.