Olig2-Induced Neural Stem Cell Differentiation Involves Downregulation of Wnt Signaling and Induction of Dickkopf-1 Expression

Understanding stem cell-differentiation at the molecular level is important for clinical applications of stem cells and for finding new therapeutic approaches in the context of cancer stem cells. To investigate genome-wide changes involved in differentiation, we have used immortalized neural stem cell (NSC) line (HB1.F3) and Olig2-induced NSC differentiation model (F3.Olig2). Using microarray analysis, we revealed that Olig2-induced NSC differentiation involves downregulation of Wnt pathway, which was further confirmed by TOPflash/FOPflash reporter assay, RT-PCR analysis, immunoblots, and immunocytochemistry. Furthermore, we found that Olig2-induced differentiation induces the expression of Dickkopf-1(Dkk1), a potent antagonist of Wnt signaling. Dkk1 treatment blocked Wnt signaling in HB1.F3 in a dosage-dependent manner, and induced differentiation into astrocytes, oligodendrocytes, and neurons. Our results support cancer stem cell hypothesis which implies that signaling pathway for self-renewal and proliferation of stem cells is maintained till the late stage of differentiation. In our proposed model, Dkk1 may play an important role in downregulating self-renewal and proliferation pathway of stem cells at the late stage of differentiation, and its failure may lead to carcinogenesis.


Introduction
It has been widely accepted until recently that no new neurons are generated after neurogenesis is completed during the early embryonic development (i.e., there are no resident stem cells in the nervous system) [1]. More recent studies, however, led to the isolation of neural stem cells (NSCs) from the embryonic mammalian central nervous system (CNS) [2][3][4], followed by the isolation of NSCs from the adult mammalian CNS [5,6]. These discoveries revealed the regenerative power of the CNS, which may be used for therapeutic purposes [7].
Currently, there are four main strategies in NSCs and their progenitor cell-based therapy: transplantation of oligodendrocyte progenitor cells for treating myelin disorders; transplantation of neuronal progenitor cells to treat diseases of discrete loss of a single neuronal phenotype, such as Parkinson disease; implantation of mixed progenitor pools to treat diseases resulting from the loss of several phenotypes, such as spinal cord injury; mobilization of endogenous neural progenitor cells to treat neurodegenerative diseases [8]. Despite significant progress that has been made for clinical application of NSCs, key questions about global perspectives for the differentiation pathway remain to be answered including molecular determinants of neural and glial fates and distinctive stages of differentiation [9].
Understanding differentiation is important for at least two reasons. Firstly, differentiation is a process of acquiring specific functions of committed cells. Therefore, understanding each step of differentiation, and characterizing differentiation phenotypes are the basis of stem cell engineering. Future stem cell research is likely to focus on improving the ability to guide the differentiation of stem cells and to control their survival and proliferation for clinical application [10]. Secondly, understanding differentiation may provide an important clue for treating cancers. According to the newly emerging cancer stem cell hypothesis, tumors seem to arise from small populations of cancer stem cells that originate from the transformation of normal stem cells [11]. In this hypothesis, a tumor can be viewed as an aberrant organ initiated by a cancer stem cell that undergoes processes analogous to the self-renewal and differentiation of normal stem cells [12]. Although similar to normal stem cells in many ways, cancer stems cells are critically different in that their transit-amplifying progeny do not mature and die as do the progeny of normal stem cells (maturation arrest) [13]. Therefore, understanding differentiation may ultimately lead to the development of differentiation therapy, which is directed toward reversal of the maturation arrest, thus allowing the cancer cells to differentiate and die eventually [14].
To identify genes and pathways that could play a role in the differentiation of NSCs, we performed microarray analysis using immortalized neural stem cell line (HB1.F3) and its oligodendrocyte progeny (F3.Olig2) in which olig2 is over-expressed. It has been shown that olig2 overexpression can induce the in vitro differentiation of NSCs into mature oligodendrocytes [15]. HB1.F3 has the ability to self-renew and differentiate into cells of neuronal and glial lineages in both in vivo and in vitro [16,17]. F3.Olig2 cells express oligodendrocyte markers and represent a model of NSC differentiation (Fig. 1).

Downregulation of Wnt pathway in F3.Olig2
Microarray analysis revealed global gene expression changes between HB1.F3 and F3.Olig2; more than 60% of genes that are present in HB1.F3 are absent in F3.Olig2. Since the global gene expression changes violate basic assumptions of statistical analysis of microarray data that most genes are not differentially expressed [18], we have employed the knowledge-based Gene Set Enrichment Analysis (GSEA) (Materials and Methods), instead of using conventional statistical analysis such as t-test, to investigate expression changes in functional groups of genes. Since the Wnt pathway is known to be involved in neural stem cell-differentiation in contra-acting ways (i.e., maintain stemness versus inducing differentiation [19][20][21], the investigation of the microarray data was focused on Wnt pathway-related gene sets. Using this method, we identified significant enrichment of Wnt pathway genes, genes upregulated by Wnt [22], and Wnt pathway target genes in HB1.F3, an immortalized neural stem cell line ( Fig. 2A-C) (see Table S1, S2, S3 for detailed information). To obtain further evidence that Wnt pathway is active in HB1.F3 and suppressed in F3.Olig2, we transfected a transcription factor (TCF) reporter gene (TOPflash) containing five optimal TCF-binding sites or the mutant control plasmid (FOPflash) into HB1.F3 and F3.Olig2. pRL-TK was included to normalize data for transfection efficiency [23]. As shown in Fig. 2D, the reporter activity of Wnt pathway is more than three times higher in HB1.F3 than in F3.Olig2. The addition of dominant negative TCF plasmids (dnTCF) decreased the reporter activity in HB1.F3 close to the level of F3.Olig2, indicating that the Wnt activity in F3.Olig2 is at the background level.

Suppression of Wnt signaling and induction of differentiation by Dkk1
As shown in Fig. 3D, Dkk1, a potent antagonist of Wnt signaling, is expressed only in F3.Olig2. We also tested the effect of Dkk1 on HB1.F3. When HB1.F3 cells were treated with Dkk1, Wnt signaling in HB1.F3 was inhibited in a dosage-dependent manner (Fig. 5A). Dkk1 treatment decreased the expression of cmyc, a Wnt pathway target gene, in a dosage-dependent manner (Fig. 5B). Dkk1 treatment also induced the expression of oligodendrocyte markers such as Olig2 and CNPase in HB1.F3 (Fig. 5C). Furthermore, Dkk1 treatment induced differentiation of HB1.F3 into astrocytes, neurons, and oligodendrocytes (Fig. 6, Fig.  S1). As for differentiation efficiency, astrocytes were the highest, oligodendrocytes the second, and neurons the lowest.

Discussion
In the present study, we showed that Olig2-induced differentiation of NSCs leads to downregulation of Wnt pathway, which is known to regulate the balance between self-renewal and differentiation in CNS [33]. Although Wnt signaling can influence cell lineage decisions such as neural differentiation of NSCs [19], differentiation of embryonic stem cells into dorsal interneurons [34], and differentiation of NSCs into dopaminergic neurons [21], Wnt signaling predominates in stem cell proliferation and neural stem cell expansion [20], and inhibits differentiation [21,35].
Unlike these studies that focused on modulating Wnt signaling and analyzing its effects on stem cells, we found, in the present study, that the differentiation-inducing event (i.e., overexpression of Olig2) may precede the downregulation of Wnt pathway.
We found that most of genes, receptors, co-receptors, and target genes expressions were increased in HB1.F3, but that of Wnt inhibitor was increased in F3.Olig2. And b-catenin was observed in cytoplasm of F3.Olig2 and nucleus of HB1.F3, whereas p-bcatenin was expressed only in the nucleus of F3.Olig2. This means canonical Wnt pathway may be activated in HB1.F3, and decreased in F3.Olig2.
Furthermore, we showed that the expression of Dkk1, a potent antagonist of Wnt signaling, is induced in F3.Olig2. Dkk1 treatment blocks Wnt signaling in HB1.F3, and induces differentiation into astrocytes, oligodendrocytes, and neurons. These findings comply with previous findings that blocking Wnt pathway induces differentiation [21], but not lineage-specific. It is generally accepted that there is the balance between self-renewal and differentiation [33], which may be manifested in two different ways. When Olig2, a differentiation-inducing signal, was overexpressed, this led to lineage-specific differentiation of neural stem cells and downregulation of Wnt pathway (i.e., self-renewal pathway) as demonstrated by F3.Olig2. When HB1.F3 cells were treated with Dkk1, a Wnt inhibitor, this led to downregulation of Wnt pathway and lineage-non-specific differentiation.
According to previous findings, Dkk1 is a direct target of the bcatenin/TCF transcription complex that mediates Wnt signaling [36][37][38]. Although these studies indicate that Dkk1 forms a novel feedback loop in Wnt signaling, our results suggest that the expression of Dkk1 is induced by a different pathway in F3.Olig2 since Wnt signaling as well as Wnt genes and receptors are suppressed in F3.Olig2. Previous reports showed that the expression of Dkk1 can be induced, independent of Wnt signaling, by differentiation-promoting reagents such as 1a, 25-dihydroxyvitamin D3 [39] and retinoic acids [40]. Dkk1 can be also induced by p53 [41].
Evidences from our experiments provide a probable link between stem cell maturation arrest and carcinogenesis at the molecular level. According to cancer stem cell hypothesis, tumors arise from maturation arrest of stem cells [42], which implies that signaling pathway for self-renewal and proliferation of stem cells is maintained till the late stage of differentiation. In our proposed model (Fig. 6), Wnt signaling, which is important for self-renewal and proliferation of NSCs, is turned off at the late stage of differentiation by Dkk1, which is turned on not by Wnt pathway but by a differentiation-related pathway. The feasibility of this model is supported by experimental evidences that Dkk1 is epigenetically silenced in many tumors including gastrointestinal tumors [43,44], cervical cancers [45], leukemia [46], and medulloblastoma [47]. Also, in HeLa cells, Dkk1 is required for tumorigenicity [48]. Altogether, these evidences may indicate that Dkk1 play an important role in downregulating self-renewal and proliferation pathway of stem cells at the late stage of differentiation, and its failure may lead to carcinogenesis.

Cell lines
Stable clonal human NSC line, HB1.F3, was generated by retroviral transduction of primary fetal human neural stem cells (hNSCs) with an avian v-myc cell cycle regulatory gene as previously reported (Kim et al, 2008; Production and characterization of immortal human neural stem cell line with multipotent differentiation property [49]. F3.Olig2 was generated by overexpressing Olig2 in HB1.F3. Briefly, Olig2 cDNA was ligated into multiple cloning sites of the retroviral vector pLHCX (Clontech, Mountain View, CA). PA317 amphotropic packaging cell line was infected with the recombinant retroviral vector, and the supernatants from the packagaing cells were added to the HB1.F3 cells. Stably transfected colonies were selected by hygromycin resistance (Kim et al., in submission).

Microarray analysis
Five replicates of each cell line were used for isolating and purifying RNA by Qiagen RNeasy kit. The resulting total RNA samples were further assessed for integrity prior to chipping using BioRad Experion System. Microarray experiments were performed using Sentrix Human-6 Whole Genome Expression BeadChips (Illumina), analyzing over 46,000 known genes, gene candidates, and splice variants according to manufacturer's instructions.

Bioinformatics
Microarray data were normalized using median normalization method. Pathway analysis of the gene expression data was performed using the Gene Set Enrichment Analysis (GSEA), which is able to detect coordinate changes in the expression of groups of functionally related genes [50,51]. An enrichment score (ES) was calculated for each gene set in GSEA and the statistical significance of the ES was estimated by an empirical permutation test using 1,000 gene permutations to obtain the nominal p-value. GSEA software was downloaded from http://www.broad.mit. edu/gsea/ Luciferase reporter assay Cells were transiently transfected using the basic nucleofector kit for primary mammalian neural cells (Amaxa biosystems) according to the manufacturer's instructions. Briefly, after trypsinization, cell counting using hemocytometer, and centrifugation, ,5610 6 cells were re-suspended in 100 ml of Basic Nucleofector Solution (Amaxa biosystems), and transfected with 1 mg of reporter plasmids (pTOPflash or pFOPflash), and 1 mg of internal control pRL-TK; with or without 4 mg of dominant negative (DN)-TCF plasmids (program A33). After transfection, ,1610 4 cells were transferred to each well of a 24-well plate containing fresh, prewarmed DMEM and maintained for 48 hrs at 37uC and 5% CO 2 . Then, the cells were lysed in lysis buffer, and 20 ml of each lysate was monitored for luciferase activity using Dual-Luciferase Reporter Assay System (Promega). Light units were measured using Lmax II 384 (Molecular devices).
A control reporter, pRL-TK contains a herpes simplex virus thymidine kinase promoter driving a Renilla luciferase gene, and Renilla luciferase activity was used to normalize the results for transfection efficiency. The reporter activities were shown as the ratios of TOPflash to FOPflash luciferase activity from triplicate experiments.

RT-PCR
RNA was isolated from six biological replicates from each group using Qiagen RNeasy MiniKit (Qiagen), pooled, and subjected to first-strand cDNA synthesis using Reverse Transcription System (Promega, A3500) according to the manufacturer's instructions. Amplification was carried out at 94uC for 2 min, followed by 30 cycles at 94uC for 1 min, at appropriate annealing temperatures for each primer for 1 min, and at 72uC for 1 min 30 s. Sequences of primers used for RT-PCR are summarized in Supplementary  Table S1.

Immunofluorescent analysis
Cells were grown on Lab-Tek II chamber slide (Nalge Nunc Int., Naperville, IL), rinsed in PBS, fixed in 4% paraformaldehyde for 20 min, and rinsed again in PBS. The cells were incubated for overnight at 4uC with following antibodies: b-catenin antibody   Supporting Information Figure S1 Phase contrast(DIC) and merged images of the immunohistochemical staining of HB1.F3 or HB1.F3 with Dkk1 treatment for various markers of stem cell and differentiated cells. Merged images of DAPI and astrocyte markers (GFAP, S100), oligodendrocyte markers (CNPase, O4), neuron markers (NeuroD, NeuN) and neural stem cell markers(Nestin, CD133) were compared with DIC images in HB1.