The effects of microRNAs on human neural stem cell differentiation in two- and three-dimensional cultures

Introduction Stem cells have the ability to self-renew or to differentiate into numerous cell types; however, our understanding of how to control and exploit this potential is currently limited. An emerging hypothesis is that microRNAs (miRNAs) play a central role in controlling stem cell-fate determination. Herein, we have characterized the effects of miRNAs in differentiated human neural stem cells (hNSCs) by using a cell line currently being tested in clinical trials for stroke disability (NCT01151124, Clinicaltrials.gov). Methods HNSCs were differentiated on 2- (2D) and 3-dimensional (3D) cultures for 1 and 3 weeks. Quantification of hNSC differentiation was measured with real-time PCR and axon outgrowth. The miRNA PCR arrays were implemented to investigate differential expression profiles in differentiated hNSCs. Evaluation of miRNA effects on hNSCs was performed by using transfection of miRNA mimics, real-time PCR, Western blot, and immunocytochemistry. Results The 3D substrate promoted enhanced hNSC differentiation coupled with a loss of cell proliferation. Differentiated hNSCs exhibited a similar miRNA profiling. However, in 3D samples, the degree and timing of regulation were significantly different in miRNA members of cluster mi-R17 and miR-96-182, and hsa-miR-302a. Overall, hNSC 3D cultures demonstrated differential regulation of miRNAs involved in hNSC stemness, cell proliferation, and differentiation. The miRNA mimic analysis of hsa-miR-146b-5p and hsa-miR-99a confirmed induction of lineage-committed progenitors. Downregulated miRNAs were more abundant; those most significantly downregulated were selected, and their putative target mRNAs analyzed with the aim of unraveling their functionality. In differentiated hNSCs, downregulated hsa-miR-96 correlated with SOX5 upregulation of gene and protein expression; similar results were obtained for hsa-miR-302a, hsa-miR-182, hsa-miR-7, hsa-miR-20a/b, and hsa-miR-17 and their target NR4A3. Moreover, SOX5 was identified as a direct target gene of hsa-miR-96, and NR43A, a direct target of hsa-miR-7 and hsa-mir-17 by luciferase reporter assays. Therefore, the regulatory role of these miRNAs may occur through targeting NR4A3 and SOX5, both reported as modulators of cell-cycle progression and axon length. Conclusions The results provide new insight into the identification of specific miRNAs implicated in hNSC differentiation. These strategies may be exploited to optimize in vitro hNSC differentiation potential for use in preclinical studies and future clinical applications.


Introduction
Stem cell research has the potential to support future medical advances, especially in the unmet need for chronic diseases [1]. Stem cells play central roles, both in development of the organism and repair of damaged tissue. CTX0E03 is a clonal conditionally immortalized human neural stem cell (hNSC) line [2]. This cell line has defined quality characteristics that are required for cell banking under good manufacturing practice (GMP), to ensure reliable and reproducible stocks of cells suitable for clinical application [3]. HNSCs can differentiate into neurons, glia, and oligodendrocytes and have been shown to ameliorate neurologic deficits in a rodent model of focal ischemia after transplantation into the brain [2,4,5].
Recently this hNSC line was tested in a human clinical trial for stroke disability in Scotland, the PISCES trial (NCT01151124, Clinicaltrials.gov). Although the functional properties of hNSCs have been studied extensively, the molecular mechanisms underlying neural stem differentiation are not fully understood.
MiRNAs have received emerging attention over the last years as significant regulatory molecules [6]. They constitute a subpopulation of small RNAs of average 22 nucleotides in length. Unlike messenger RNAs (mRNAs), miRNAs do not encode proteins, but rather bind 3′ untranslated region (3′ UTR) of mRNAs, regulating their stability and translation into proteins. Functional studies indicate that miRNAs participate in the regulation of a number of cellular processes, including development, proliferation, and differentiation [7]. The discovery of miRNAs has offered new potential for modulating stem cell lineage commitment and differentiation by posttranscriptional gene regulation [8,9]. Many studies have demonstrated that transient overexpression or inhibition of brain-specific miRNAs in stem cells significantly directed their differentiation toward neuronal cell lineages [10]. Several miRNAs have been implicated in regulating self-renewal of neural stem cells and neuronal fate specification [11]. Herein we compared miRNA profiling obtained from in vitro assays of a clinical grade hNSC line to investigate further miRNA functionality and effects on neuronal and glial differentiation potential.
In two-dimensional (2D) standard differentiation protocols, the complexity of the in vivo environment is not reflected, and consequently, the induction and regulation of hNSC differentiation is not optimal. In vivo, cells are surrounded by other cells as well as many extracellular ligands, including many types of collagens, laminin, and other matrix proteins (extracellular matrix, ECM) in a three-dimensional (3D) microenvironment. Previous studies showed that specific topologic architecture of the materials mimicking ECMs and their geometry influence cell phenotype and fate [12][13][14][15][16][17][18][19]. A number of commercially available 3D culture systems are available, including Alvetex (Amsbio, UK). Alvetex is a highly porous polystyrene scaffold engineered with a well-defined and uniform architecture into a 200-μm-thick membrane, which provides a 3D space into which cells can invade and differentiate [20]. The scaffold is formed by polymerization in a biphasic emulsion, consisting of an aqueous and a nonaqueous monomer/surfactant phase, termed a high-internalphase emulsion, or HIPE [21,22]. The resulting polymer (poly-HIPE) consists of a relatively homogeneous porous network of voids, linked by interconnecting pores. Moreover, such scaffolds have been proven amenable for 3D in vitro cell culture [20].
In the present study, 2D and 3D differentiation assays were used to monitor mRNAs of neuronal and glial markers and neurite outgrowth to identify miRNAs involved in the regulation of neuronal/glial differentiation processes.

Material and methods
HNSC derivation, culture, and differentiation CTX0E03 is a fully manufactured conditionally immortalized hNSC line, originally derived from ethically sourced human fetal brain cortical tissue of 12 weeks' gestation and described in [2]. To set up differentiation assays, a singlecell suspension of hNSCs, obtained from passage 30 to 36 cell cultures, was achieved by trypsinization, and the number of cells determined by using a hemocytometer. Cells were seeded either on standard cell-treated plastic vessels (BD Biosciences) or on polystyrene scaffolds (Alvetex®, Amsbio) in serum-free medium. Cells were maintained at 37°C in a humidified, 5% CO 2 incubator for 1 and 3 weeks, and cultured in the same defined medium without the mitogens, EGF, bFGF, and 4-OHT (Sigma).

Measurement of axon-process outgrowth
Measurement of axon-process outgrowth was performed on differentiated and undifferentiated hNSCs fixed with 4% paraformaldehyde (PFA; Pioneer Research Chemicals) and incubated with β3-tubulin (TUBB3) primary antibody (Sigma). After rinsing, the cells were incubated with antimouse Alexa Fluor 568 (Invitrogen) secondary antibody. Cell nuclei were counterstained with Hoechst (Sigma, 1 μM). Measurement of axon-process length was carried out on a minimum of three representative fields by using Image-Pro Plus 7 software (Media Cybernetics).

Measurement of cell proliferation
Measurement of cell proliferation was performed on differentiated and undifferentiated hNSCs fixed with 4% PFA and stained with Ki67 primary antibody, 1:100, (Thermo Scientific). After rinsing, the cells were incubated at room temperature with anti-rabbit Alexa Fluor 568 (Invitrogen) secondary antibody. Cell nuclei were counterstained with Hoechst (Sigma, 1 μM). A minimum of three representative fields was analyzed for each condition.

Stem cells and developmental pathways-focused miRNA PCR-array analysis
The stem cells and developmental pathways-focused miRNA PCR array (Qiagen, Sabiosciences) was carried out according to manufacturer's instructions. For each array, a minimum of 250 ng total RNA was retrotranscribed by using miScript II RT Kit (Qiagen) according to the manufacturer's instruction. Each condition was run in triplicate. Array analysis was performed by using miScript miRNA PCR-array data analysis [28].

MiRNA mimic transfection
Before transfection, hNSCs were seeded into a 24-well plate. MiRNA-transfection optimization was performed by using allstars negative control siRNA AF 488 (Qiagen); by using the following condition, the miRNA transfection efficiency was found to be 100%. Each of three miRNA mimics (5 nM; Qiagen), hsa-miR-146b-5p, UG AGAACUGAAUUCCAUAGGCU, hsa-miR-23b, AUCA CAUUGCCAGGGAUUACC, and hsa-miR-99a, AACC CGUAGAUCCGAUCUUGUG was combined with HiPer-Fect (Qiagen) according to manufacturer's instructions. Control samples were transfected with green fluorescent protein (GFP) plasmid. Samples were collected after 1 week. MiRNA-mimic cell internalization after transfection was measured by qRT-PCR using miScript PCR Starter kit (Qiagen) according to manufacturer's instructions. A minimum of three biologic replicants was evaluated for each marker and condition.
MiRNA KEGG pathway analysis of selected miRNAs was performed by using DIANA LAB [30,35].

Western blot analysis
HNSC cell monolayers were lysed with 1× SDS sample buffer and dithiothreitol (DTT) reducing agent (AMS Biotechnology), and loaded onto a PAGEgel (Invitrogen). After electrophoresis, the proteins were blotted on nitrocellulose membrane. Immunodetection was performed by using rabbit anti-SOX5 and anti-Nor (NR4A3) polyclonal antibodies (Santa Cruz Biotechnology Inc.) and detected by using a horseradish peroxidase-conjugated anti-rabbit antibody (Cell Signaling Technology). The nitrocellulose membrane was then processed by using chemiluminescence-detection reagents (Thermo Scientific). The blots were stripped and reprobed by using anti-α-tubulin (Sigma, 1:1,000) to act as an internal loading-level standard. Western blot images were captured by using BioRad FluorS Imaging.

Statistics
Data were analyzed by Student t test, and P < 0.05 was considered statistically significant. All error bars indicate ± standard deviation of the mean (SDM). In stem cells and developmental pathways focused miRNA PCR array analysis, significant changes were defined as ±1.5-fold up-and down-regulation to a statistically significant extent. A two-sample Student t test was used to analyze each comparison.

Results
Quantification of differentiation by axon outgrowth and real-time PCR molecular analysis, and quantification of cell proliferation Assessment of the ability of hNSCs to differentiate into neurons, glia, and oligodendrocytes by immunocyto- and glial (GALC, GFAP, and S100B) markers performed on hNSCs differentiated on 2D and 3D for 1W, and 3W, and expressed as fold change compared with undifferentiated control. Statistical analysis showed significant differences between 2D versus 3D samples at the same time point; ± SDMs. *P < 0.05; **P < 0.001, ***P < 0.005, Student t test.

Human cell differentiation and development miScript miRNA PCR array profiles
Stem cells and developmental pathways focused miRNA PCR array (Sabioscience; Qiagen) was implemented to investigate miRNAs differential expression profiles after hNSC 2D and 3D differentiation for 1W and 3W.

Mimic miRNA transfection and real-time RT-PCR quantification of neuronal and glial gene markers
Based on array analysis we selected three up-regulated miRNAs to substantiate their effects on neuronal/glial differentiation. Hsa-miR-99a, hsa-miR-146b-5p, and hsa-miR-23b mimics were transfected into hNSCs. Quantification of each miRNA mimic up-take and expression of GALC, GFAP, TUBB3, S100B, DCX, and MAP2 mRNAs were measured by qRT-PCR. All selected mimic miRNAs were appropriately transfected and quantified in hNSCs ( Figure 4A) and they significantly increased GFAP, DCX, and MAP2 markers when compared to control. In contrast significant increases in gene expression of the oligodendrocyte precursor marker GALC and the astrocytic glial marker S100B were solely induced by hsa-miR-99a, whereas the neuronal precursor marker TUBB3 was solely induced by hsa-miR-146b-5p ( Figure 4B).
resulting target sites was quite high, ranging from 1955/ 973 to 302/190 total target sites per gene identified. To overcome this issue we only investigated top ranked miRNA target genes that were previously referenced to brain development and function or were target sites of multiple investigated miRNAs. Validation of predicted targets was performed by qRT-PCR. SOX5, NR4A3 (NOR1), and FOXN3 were found to be significantly up-regulated in differentiated samples compared with control. In addition DUSP10 was only found to be up-regulated in 3D differentiated cultures at 3W. Among the putative target mRNAs, only EIF4G3 was not up-regulated and therefore did not correlate with the prediction ( Figure 5A). To support miRNA target prediction we also examined protein expression of SOX5 and NR4A3. We performed Western blot analysis for two of the predictive target mRNA, SOX5, and NR4A3 (NOR1). The differential expression of these proteins was verified in 2D and 3D differentiated hNSCs at both time points (1W and 3W) and compared with proliferative controls ( Figure 5B).

Discussion
Highly orchestrated programs of gene expression act to shape the developing nervous system. More recently, it has become clear that gene expression can also be modulated by several classes of small RNAs. In this study, we profiled miRNA differential expression patterns in hNSCs differentiated in 2D and 3D culture systems. HNSC differentiation was assessed by measuring processes/axons outgrowth and gene expression of well-established neural and glial markers. MiRNA profile of 2D versus 3D was very similar in terms of miRNA types; but the degree and timing of miRNA differential regulation was significantly different for miRNAs involved in both maintenance of stemness (hsa-miR-302a), and cell proliferation (the clusters miR-17-92 and miR-96-183). The upregulated miR-NAs: hsa-miR-146b-5p, hsa-miR-23b, and hsa-miR-99a were selected, and their mimics were transfected into hNSCs to validate their correlation with differentiation of progenitors into neurons or glia. GALC and S100B gene expression was significantly upregulated after transfection of hsa-miR-99a mimics, suggesting a role in promoting differentiation of hNSC into glial restricted progenitors and oligodendrocyte progenitor cells. Similarly, hsa-miR-146b-5p transfection promoted upregulation of TUBB3 gene expression. Taken together, hsa-miR-99a and hsa-miR-146b-5p appear to be involved in early-stage commitment of glial and neuronal precursors, respectively. The most significantly downregulated miRNAs were selected and analyzed to assess their putative target mRNAs with the intent of determining their functionality and identifying KEGG pathway maps for biological interpretation. By using online-available software analysis tools (DIANA Lab, PicTar, and TargetScans), we identified and verified a set of target mRNAs correlated with selected downregulated miRNAs. A single miRNA can recognize hundreds of targets. However, several miRNAs can target one gene. NR4A3, a target of the hsa-miR-7, hsa-miR302a, and miR-17 family, is a member of the NR4A family of nuclear receptors, which, depending on their level of expression, are involved in the  Figure 5 Validation of miRNA target-predicted genes. (A) qRT-PCR analysis performed on 1W and 3W differentiated hNSCs cultured on 2D and 3D substrates and expressed as fold change compared with proliferative control. Statistical analysis showed significant difference compared with control; ±SDMs, *P < 0.05, **P < 0.001, ***P < 0.005, Student t test. (B) SOX5 and NR4A3 protein quantification performed by Western blot on 1W and 3W differentiated hNSCs cultured on both 2D or 3D substrates and proliferative control. (C) Dual luciferase report assay. Measurement of the relative luciferase activity of SOX5 and NR4A3 3′-UTR constructs transfected with hsa-miR-96, and hsa-miR-7 and 17, respectively. Data are expressed as mean values ± SDMs and are shown as percentage of control (cells transfected with either SOX5 or NR4A3 3′-UTR constructs and control microRNA). Each bar represents values from three independent experiments, measured in triplicate. The relative activity of firefly luciferase expression was normalized to renilla luciferase activity. Data were analyzed with Student t test, ***P < 0.005. (D) MiRNA KEGG pathway analysis results obtained by using DIANA Lab. differentiation, survival, apoptosis, and regulation of hippocampal axon guidance [50]. Similarly, SOX5, a target of hsa-miR-96, is reported to control cell-cycle progression in neural progenitors [51], axon length [52], migration, postmigratory differentiation, and projections of neurons [53]. NR4A3 and SOX5 were selected as possible candidate regulators of cell-cycle progression and axonal regulation.
QRT-PCR and Western blot analysis confirmed upregulation of both NR4A3 and SOX5. Furthermore, SOX5 and NR4A3 were identified as direct target genes of hsa-miR-96, and hsa-miR-7 and 17, respectively, by luciferase reporter assays. These findings, together with miRNA KEGG pathway analysis, support a possible regulatory role of these identified miRNAs as cellular regulators of axon guidance and outgrowth regulation. By mimicking a tissue-like environment, 3D scaffolds may enhance environmental guidance cues that modulate miRNA expression to promote axon outgrowth.

Conclusions
Overall, 3D hNSC culture was associated with greater arrest of cell proliferation, increased incidence of differentiation, and differential degree and timing of regulation in miRNA expression compared with 2D systems. These observations suggest that 3D surface topography influences hNSC molecular behavior by modulating miRNAs associated with cell proliferation and stemness maintenance, thereby promoting cell differentiation.
Additionally, 3D architecture may provide contact guidance that could regulate miRNAs involved in developmental processes, including growing axons and neuritogenesis.