Earlier studies show that Hes1 expression is oscillatory in neural stem cells but sustained and high in the roof plate and the floor plate, and that such different dynamics of Hes1 expression (oscillatory versus sustained) regulate different proliferation and differentiation characteristics of these cells (active in neural stem cells but rather dormant in roof/floor plate cells). The mechanism of how different dynamics of Hes1 expression is controlled remains to be determined. Here, we found that the seed sequence of microRNA-9 (miR-9) is complementary to the 3′-UTR sequence of Hes1 mRNA. MiR-9 is highly expressed in the ventricular zone of the developing brain, which contains neural stem cells, but it is not expressed in the roof plate or the floor plate. Over-expression of miR-9 negatively regulates the Hes1 protein expression by interacting with the 3′-UTR of Hes1 mRNA, thereby inducing cell cycle exit and neuronal differentiation. Conversely, knockdown of miR-9 inhibits neuronal differentiation. Furthermore, knockdown of miR-9 inhibits the oscillatory expression of Hes1 mRNA in neural stem cells. These results indicate that miR-9 regulates the proliferation and differentiation of neural stem cells by controlling the dynamics of Hes1 expression in the developing brain.
The basic helix-loop-helix transcriptional repressor gene Hes1 regulates the maintenance of neural stem cells by repressing proneural gene expression (Kageyama et al. 2007). In the absence of Hes1, proneural genes such as Mash1 and Neurogenin2 are up-regulated and accelerate neuronal differentiation, resulting in neural tube defects and eye morphogenesis defects (Ishibashi et al. 1995; Tomita et al. 1996; Hatakeyama et al. 2004). It has been shown that Hes1 expression oscillates with a period of 2–3 h in many cell types including neural stem cells (Hirata et al. 2002; Shimojo et al. 2008; Kobayashi et al. 2009), and that this oscillation is important for cell proliferation and differentiation as sustained Hes1 expression inhibits both processes (Baek et al. 2006; Yoshiura et al. 2007). However, not all cells express Hes1 in an oscillatory manner, and cells in the roof plate and the floor plate express Hes1 in a sustained manner (Baek et al. 2006). These cells proliferate slowly or stay quiescent and usually do not differentiate into neurons, suggesting that the oscillatory versus sustained Hes1 expression dynamics are important for cell proliferation and differentiation (Baek et al. 2006). Hes1 oscillation is regulated by delayed negative feedback, and this oscillatory expression critically depends upon the instability of both Hes1 mRNA and Hes1 protein (Hirata et al. 2002). Although the instability of Hes1 protein is regulated by the ubiquitin-proteasome system (Hirata et al. 2002), the mechanism of how the instability of Hes1 mRNA is regulated remains to be determined.
MicroRNAs (miRNAs) are a class of small noncoding RNAs that induce degradation and/or repress translation of target mRNAs (He & Hannon 2004). Inactivation of Dicer1, which is essential for the production of almost all miRNAs, results in severe hypotrophy of the neocortex, suggesting that miRNAs play an important role in brain development (De Pietri Tonelli et al. 2008). There are over 200 miRNAs in the mouse, and among them, microRNA-9 (miR-9) was shown to be essential for neural development. There are three different precursors for miR-9: miR-9-1, miR-9-2 and miR-9-3, and ablation of miR-9-2 and miR-9-3 decreases Pax6 expression, increases Gsh2 and Foxg1 expression, and impairs brain development (Shibata et al. 2011). It has been shown that miR-9 also targets her5 and her9 in zebrafish (Leucht et al. 2008) and hairy1 in Xenopus (Bonev et al. 2011). It was also suggested that miR-9 targets mouse Hes1, a homolog of her and hairy genes (Leucht et al. 2008). These results suggest that miR-9 may regulate neural development via multiple transcription factors including Hes1.
Here, we examined the relationship between miR-9 and Hes1 expression and found that miR-9 regulates Hes1 oscillation by controlling the stability and expression efficiency of Hes1 mRNA.
MiR-9 is expressed in the mouse brain during development
To understand the relationship between miR-9 and Hes1 expression, we first carried out in situ hybridization by using a DIG-labeled LNA-miR-9 probe. Whole-mount in situ hybridization analysis showed that miR-9 is highly expressed mainly in the central nervous system at E11.5, including the telencephalon and the mesencephalon (Fig. S1A in Supporting Information). A probe for primary miR-9 (pri-miR-9) was also used for whole-mount in situ hybridization analysis, and pri-miR-9 was found to exhibit similar expression patterns to mature miR-9 (Fig. S1B in Supporting Information). At E12.5, pri-miR-9-1 and pri-miR9-2 were widely expressed in the ventricular zone of the nervous system such as the telencephalon, diencephalon, and spinal cord, where Hes1 is also expressed (Fig. 1). In these regions, Hes1 expression is oscillatory, resulting in variable levels from cell to cell (Shimojo et al. 2008). However, in the roof plate and the floor plate, where Hes1 is steadily expressed at high levels (Fig. 1G,H,K,L,O,P, arrowheads) (Baek et al. 2006), the expression of miR-9-1 and miR9-2 precursors was almost absent (Fig. 1E,F,I,J,M,N, arrows). These results indicate that miR-9 expression occurs in the regions where Hes1 expression is oscillatory but not in the regions where Hes1 expression is at high and sustained levels.
MiR-9 regulates Hes1 expression by interacting with 3′-UTR of Hes1 mRNA
Bioinformatics analyses using PICTAR and TargetScan showed that there is a miR-9-binding site, 5′-gccaaaga-3′, in the 3′-untranslated region (UTR) of mouse, rat, and human Hes1 and zebrafish her1 mRNAs (Fig. 2A), suggesting that miR-9 is involved in regulation of Hes1 expression (Bonev et al. 2011). To determine whether Hes1 expression is regulated by interaction of miR-9 with the 3′-UTR, we examined the luciferase reporter expression in the presence or absence of a miR-9 binding site of the 3′-UTR of Hes1 mRNA (Fig. 2B). HEK293T cells were transfected with control or miR-9 expression vector together with the luciferase reporters. As a control, we used a luciferase reporter containing SV40 late 3′-UTR, whose expression was not affected by over-expression of miR-9 (Fig. 2B,C). When the wild-type 3′-UTR of Hes1 mRNA was present, the reporter expression was down-regulated by miR-9 whereas this down-regulation did not occur when the miR-9-binding site in the 3′-UTR of Hes1 mRNA was mutated (Fig. 2B,C), suggesting that Hes1 expression is negatively regulated by miR-9. Similarly, elongation factor 1 (EF) promoter-driven miR-9 over-expression, which generated a mature form of miR-9 (Fig. S2A in Supporting Information), down-regulated the endogenous Hes1 expression in embryonic stem cells (Fig. S2B in Supporting Information). These results indicate that miR-9 negatively regulates Hes1 expression by interacting with the 3′-UTR of Hes1 mRNA.
Because Hes1 mRNA is known to have a short half-life (Hirata et al. 2002), we next examined whether miR-9 is involved in the instability of Hes1 mRNA. The Hes1 mRNA level in NIH3T3 cells was measured after transcription was blocked by Actinomycin-D. In the control, the half-life of Hes1 mRNA was approximately 38 min, whereas it was approximately 12 min when miR-9 was over-expressed (Fig. 2D). The half-life of the GFP mRNA, which did not contain any miR-9-binding sites, was not affected by over-expression of miR-9 (Fig. 2E). We also examined whether miR-9 is involved in the stability of Hes1 protein. The Hes1 protein level was measured after translation was blocked by cycloheximide, but the stability of Hes1 protein was not significantly affected by miR-9 (Fig. S3 in Supporting Information). These results suggest that the stability of Hes1 mRNA, but not Hes1 protein, is negatively regulated by miR-9.
MiR-9 regulates proliferation and differentiation of neural stem cells
The above results indicate that over-expression of miR-9 down-regulates Hes1 expression by reducing the expression efficiency and the stability of Hes1 mRNA. Because Hes1 plays an important role in proliferation and differentiation of neural stem cells, miR-9 may also regulate these processes. Control or miR-9 expression vector was electroporated into the dorsal telencephalon at E13.5, and the behavior of the transfected cells were examined at E15.5. Control cells were mostly present in the ventricular zone or the subventricular zone, and only some of them migrated into the cortical plate at E15.5 (Fig. 3A,B, control). By contrast, when miR-9 was over-expressed, more cells migrated out of the ventricular zone, and more cells reached the cortical plate (Fig. 3A,B). Furthermore, over-expression of miR-9 increased the population of transfected cells that became negative for Ki67, a marker of proliferating cells (Fig. 3A). In addition, although many of control cells expressed phosphohistone H3 (PH3), another marker of proliferating cells, miR-9-over-expressing cells that expressed PH3 were much fewer, suggesting that over-expression of miR-9 induced cell cycle exit (Fig. 3C). These results indicate that miR-9 induces cell migration and cell cycle exit. However, coelectroporation of the Hes1 expression vector that lacked the miR-9-binding site antagonized these miR-9 effects (Fig. 3A,B). These results suggest that miR-9 regulates neural development via repression of Hes1.
To examine the function of miR-9 further, control or miR-9 expression vector was transfected into neural stem cells at E13.5 by electroporation, and the Hes1 expression was measured at E15.5. Over-expression of miR-9 by electroporation at E13.5 decreased the Hes1 expression level in neural stem cells in the ventricular zone at E15.5 (Fig. 4A,B). Knockdown of miR-9 at E13.5 decreased the generation of Tuj1-positive neurons (Fig. 4C,D) but increased the number of nestin-positive neural stem cells at E15.5 (Fig. 4E), an effect similar to up-regulation of Hes1 expression. These data suggest that miR-9 promotes neuronal differentiation by repressing Hes1 expression.
MiR-9 regulates the dynamics of Hes1 expression
We showed that miR-9 regulates the efficiency of Hes1 protein expression and the stability of Hes1 mRNA, and that miR-9 is expressed by neural stem cells, which express Hes1 in an oscillatory manner (Shimojo et al. 2008). Because the expression efficiency and the stability of gene products are important for oscillatory expression (Lewis 2003; Hirata et al. 2004), the above results suggest that miR-9 may regulate the oscillatory expression of Hes1. Hes1 oscillations have been simulated by a delayed differential equation model (Jensen et al. 2003; Lewis 2003; Monk 2003; Hirata et al. 2004). According to this model, robust Hes1 oscillations can be simulated with the Hes1 mRNA half-life of 25 min and the protein synthesis rate from 4.5/min (Fig. 5A). To simulate the miR-9 knockdown effect, we increased the Hes1 mRNA half-life from 25 to 40 min and the protein synthesis rate of 4.5 to 9/min. Under this condition, Hes1 oscillations would be dampened rapidly (Fig. 5A). To evaluate this mathematical prediction, we examined the effect of miR-9 knockdown on Hes1 oscillation by using the time-lapse imaging system (Masamizu et al. 2006; Shimojo et al. 2008).
The Hes1 promoter-driven destabilized luciferase reporter (Fig. 5B) was transfected with the control or miR-9 knockdown hairpin inhibitor into neural stem cells in the E13.5 cortex by electroporation, and 24 h later these cells were dissociated and cultured. Bioluminescence imaging showed that Hes1 expression oscillated in an unstable manner in many neural stem cells when the control hairpin inhibitor was co-tranfected (Fig. 5C, Control, and Movies S1 and S2 in Supporting Information), as previously reported (Shimojo et al. 2008). By contrast, Hes1 expression became nonoscillatory or stabilized in many neural stem cells when the miR-9 knockdown hairpin inhibitor was co-tranfected (Fig. 5C, miR-9 KD, and Movies S3 and S4 in Supporting Information). These results indicate that miR-9 plays an important role in the oscillatory expression of Hes1.
MiR-9 regulates proliferation and differentiation of neural stem cells by controlling Hes1
In this study, we showed that miR-9 regulates the expression efficiency and the stability of Hes1 mRNA by interacting with its 3′-UTR, and that over-expression of miR-9 decreases the Hes1 level and promotes cell cycle exit and neuronal differentiation whereas knockdown of miR-9 inhibits neuronal differentiation. Furthermore, knockdown of miR-9 inhibits oscillatory expression of Hes1 in neural stem cells. These results indicate that miR-9 regulates the proliferation and differentiation of neural stem cells by controlling the Hes1 expression dynamics. During preparation of this work, Bonev et al. (2012) reported basically the same results to ours. In addition, they showed that Hes1 periodically represses pri-miR-9 expression, resulting in out-of-phase oscillation between Hes1 and pri-miR-9 expression. However, because the mature form of miR-9 is stable, pri-miR-9 oscillation leads to gradual accumulation of miR-9, which may eventually antagonize Hes1 expression, leading to neuronal differentiation (Bonev et al. 2012). Therefore, it is possible that miR-9 is involved in the timing of neuronal differentiation.
Knockdown of miR-9 stabilized Hes1 expression or dampened Hes1 oscillation in many neural stem cells but not all. Because the knockdown method did not completely block miR-9 production, remaining miR-9 could maintain Hes1 oscillation in some neural stem cells. Alternatively, other miRNAs could be responsible for the expression efficiency and the stability of Hes1 mRNA. It has been shown that other miRNAs, miR-199b-5p in medulloblastoma (Garzia et al. 2009), and miR-124 in P19 cells (Wang et al. 2010) negatively regulate Hes1 expression, and these miRNAs could be also involved in oscillatory expression of Hes1.
Interestingly, miR-9 is not expressed in the roof plate or the floor plate, where Hes1 is expressed steadily at high levels (Baek et al. 2006). Sustained expression of Hes1 inhibits proliferation and differentiation of neural stem cells, suggesting that oscillatory expression of Hes1 is required for efficient proliferation and differentiation of these cells (Baek et al. 2006). The precise mechanism of how the oscillatory versus sustained Hes1 expression is regulated is unknown, but it is likely that the absence of miR-9 expression in the roof plate and floor plate leads to stabilization of Hes1 mRNA, resulting in steady expression of Hes1. It remains to be determined whether ectopic expression of miR-9 is sufficient to induce the oscillatory expression of Hes1 in the roof plate and the floor plate.
MiR-9 targets many genes in addition to Hes1
It has been shown that miR-9 targets many genes in addition to Hes1. For example, miR-9 induces Cajal-Retzius cell differentiation by repressing Foxg1 expression (Shibata et al. 2008), although it also promotes neuronal differentiation by repressing TLX (Zhao et al. 2009). The final outcome of miR-9-induced repression of Foxg1 and TLX is similar to that of miR-9-induced repression of Hes1. Therefore, miR-9 regulates neural stem cell differentiation by coordinately controlling many target genes. In addition, miR-9 regulates the expression of stathmin and REST in the nervous system (Packer et al. 2008; Delaloy et al. 2010). It is likely that miR-9 may be involved in many different steps of neural functions by regulating different target genes.
Dicer is not necessarily required for oscillatory expression
Hes1 expression oscillates with a period of approximately 2–3 h widely in many cell types (Hirata et al. 2002), but miR-9 expression seems to be rather specific to the nervous system, suggesting that other miRNAs such as miR-199b-5p and miR-124 (Garzia et al. 2009; Wang et al. 2010) might regulate the instability of Hes1 mRNA in nonneural cells. However, miRNAs are not necessarily required for oscillatory expression. For example, oscillatory expression of Hes7 in the presomitic mesoderm regulates somite segmentation, which occurs every 2 h in mouse embryos (Bessho et al. 2001). In Dicer knockout mice, however, the somite segmentation seems to proceed normally, although the size of somites is smaller than the wild type (Zhang et al. 2011). Therefore, although Dicer is important for proliferation and survival of somitic cells, the expression of the segmentation clock genes such as Hes7 is likely to oscillate normally without miRNAs. Further analyses will be required to determine the mechanism of regulation of the stability of Hes7 mRNA as well as of Hes1 mRNA in nonneural systems such as the segmentation clock.
Whole-mount and section in situ hybridization
Whole-mount in situ hybridization with DIG-labeled LNA probe specific for detecting the mature form of miR-9 (Exiqon) was carried out with E11.5 embryos according to manufacturer's instruction. Section in situ hybridization was carried out with 16-μm cryosections, as described previously (Shimojo et al. 2008; Tan et al. 2012).
Immunohistochemistry and antibodies
To carry out immunohistochemistry, embryos of different ages were fixed in 4% PFA for 2 h, cryo-protected in 20% sucrose overnight and cryo-sectioned with a thickness of 16 μm. Sections were blocked with 5% normal goat serum and incubated with primary antibodies at 4 °C overnight. Antibodies used in this research are as follows: Hes1 (Imayoshi et al. 2008), Ki67 (1 : 100; BD Pharmingen), Nestin (1 : 500; BD Pharmingen), PH3 (1 : 500; Sigma H6409), and TujI (1 : 500; Babco).
Plasmids and luciferase assay
To overexpress miR-9, ~300 bp sequences flanking mature miR-9 were cloned into pBos vector downstream of EF promoter. Primer sets for amplifying miR-9 are listed in Table 1. To overexpress Hes1, rat Hes1 cDNA excluding the 3′-UTR was inserted into EcoRI site of pCLIG as described previously (Ohtsuka et al. 2001). To carry out miR-9 knockdown assay, we used control (Thermo Scientific Dharmacon IH-001005-01-05) and miR-9 (Thermo Scientific Dharmacon IH-310400-08) hairpin inhibitors. Full-length Hes1 3′-UTR was cloned downstream of the luciferase coding sequence into the XbaI site of pGL3-Basic Vector (Clontech). Mutation of a miR-9 binding site was introduced into the luciferase construct using a site-directed mutagenesis method with the primers listed in Table 1. Luciferase assay was carried out with 293T cells in 24-well tissue culture plates. Briefly, cells were plated 1 day before, and transfected the following day with pEF-control or pEF-miR-9 expression vector, together with pGL3-Basic, pGL3-Basic with wild type Hes1 3′-UTR or point-mutated 3′-UTR vectors. pTk-Renilla was co-transfected as a transfection control. Transfection was carried out with Lipofectamine LTX (Invitrogen) according to manufacturer's protocol. All experiments were carried out in triplicates, and the experiments were repeated at least 3 times.
Table 1. Primer list
Amplification of Pri-miR-9
Hes1 3′UTR mut
Mutation of Hes1 3′UTR
mRNA half-life assay
mRNA half-life assay was carried out using NIH3T3 cells. Cells were transfected with pEF-control or pEF-miR-9 expression vector 1 day before. Then, cells were treated with Actinomycin-D and collected at T = 0, 20, 40, 60, 90, 120, 150 and 180 min. RNA was extracted with Trizol according to manufacturer's protocol (Invitrogen), followed by superscript reverse transcription and quantitative real-time PCR (qRT-PCR) to detect endogenous Hes1 mRNA expression. Hes1 mRNA levels were normalized with an internal control, GAPDH. qRT-PCR was carried out as previously described (Tan et al. 2012). Primer sets for qRT-PCR are listed in Table 1.
In utero electroporation
In utero electroporation was carried out as described previously (Shimojo et al. 2008). Briefly, 1–2 μL of 2 μg/μL DNA was injected into the lateral ventricle of E13.5 embryos and 5 pulses of 50 V electricity were applied to each embryo. Embryos were killed and fixed at E15.5.
Time-lapse imaging was carried out as described previously (Shimojo et al. 2008). Briefly, pHes1-UbLuc was electroporated into E13.5 embryos with control or miR-9 hairpin inhibitor. Twenty-four hours later, neural stem cells were harvested and plated onto PLL-coated glass-bottom plates. Before time-lapse imaging, luciferin was added into the culture medium and equilibrated for at least 30 min. Images were analyzed with ImageJ software.
The mathematical modeling was previously described (Hirata et al. 2004) with some changes as follows.
These equations include the Hes1 protein p and mRNA m. We set the wild-type protein and synthesis rates a =4.5/min and k = 33/min, critical repression threshold pcrit = 40, protein and mRNA synthesis delays Tp= 8 min and Tm= 29 min and protein and mRNA half-lives τp = 20 min and τm= 25 min, respectively. Note the experimentally observed higher mRNA half-life τm = 25 min and Hill coefficient 2.5 than in the previous work (Hirata et al. 2004). In the knockdown simulation, we set higher protein synthesis rate a =9/min and mRNA half-life τm = 40 min.
We thank Itaru Imayoshi and Hiromi Shimojo for technical help. This work was supported by the Grants-in-aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan. S.L.T. was supported by MEXT scholarship of the Ministry of Education, Culture, Sports, Science and Technology of Japan.