The Role of Microrna-9 in Vertebrate Neural Development Date: 27-09-2011

THE ROLE OF MICRORNA-9 IN

VERTEBRATE NEURAL DEVELOPMENT

A thesis submitted to the University of Manchester for the Wellcome Trust 4 Year PhD in the Faculty of Life Sciences

2011

Boyan Bonev Table of Contents

TITLE PAGE ...... 1

TABLE OF CONTENTS ...... 2

TABLE OF FIGURES ...... 3

ABSTRACT ...... 5

ACKNOWLEDGEMENTS ...... 7

ABBREVIATIONS ...... 8

RATIONALE FOR SUBMITTING IN ALTERNATIVE FORMAT ...... 9

AUTHOR CONTRIBUTIONS ...... 10

CHAPTER 1. INTRODUCTION ...... 11

1. DEVELOPMENT OF THE VERTEBRATE CENTRAL NERVOUS SYSTEM ...... 11

1.1. Extrinsic signals regulating neuronal differentiation ...... 13

1.1.1. Wnt signaling ...... 13

1.1.2. Fibroblast Growth Factor (FGF) signaling ...... 14

1.1.3. Notch signaling ...... 16

1.2. Intrinsic mechanisms of cell fate regulation during neural development ...... 17

1.2.1. Chromatin state determines the probability of transcription ...... 18

1.2.2. DNA methylation ...... 19

2. MORPHOLOGICAL CHANGES DURING TERMINAL NEURONAL DIFFERENTIATION ...... 20

2.1.1. Axon specification ...... 20

2.1.2. Axon elongation ...... 21

2.1.1. Axon branching ...... 22

2.1.2. Local control on axonal mRNA translation ...... 23

3. MICRORNAS IN NEURAL DEVELOPMENT ...... 24

3.1. Biogenesis, mechanism and function ...... 24

3.2. microRNA function in the development of the nervous system ...... 27

3.3. miRNA-9 – an important regulator of vertebrate neural development ...... 29

AIMS AND OBJECTIVES ...... 31

2 CHAPTER 2. MICRORNA-9 REVEALS REGIONAL DIVERSITY OF NEURAL PROGENITORS ALONG THE ANTERIOR-POSTERIOR AXIS ...... 33

CHAPTER 3. MICRORNA-9 MODULATES HES1 ULTRADIAN OSCILLATIONS BY FORMING A DOUBLE NEGATIVE FEEDBACK LOOP ...... 34

CHAPTER 4. MICRORNA-9 REGULATES AXON EXTENSION AND BRANCHING BY TARGETING MAP1B IN MOUSE CORTICAL NEURONS ...... 61

CHAPTER 5. DISCUSSION ...... 93

REFERENCES ...... 104

Total word count (excluding publications): 26,796

3 Table of Figures

FIGURE 1. CELL TYPES COMPRISING THE VERTEBRATE CENTRAL NERVOUS SYSTEM ...... 12

FIGURE 2. WNT SIGNALING IN NEURAL DEVELOPMENT...... 14

FIGURE 3. FGF8 SIGNALING IN NEURAL DEVELOPMENT ...... 15

FIGURE 4. NOTCH SIGNALING GENERATES HES1 MOLECULAR OSCILLATIONS ...... 16

FIGURE 5. SPECIFICATION AND ELONGATION OF THE AXON DURING NEURONAL MATURATION ... 21

FIGURE 6. AXON BRANCHING IS AN ESSENTIAL STEP DURING NEURONAL MATURATION...... 23

FIGURE 7. BIOGENESIS OF MICRORNAS...... 25

FIGURE 8. MICRORNAS MECHANISMS OF TARGET RECOGNITION AND SILENCING ...... 26

4 ABSTRACT The University of Manchester Boyan Bonev Wellcome Trust 4 Year PhD Thesis Title: The Role of microRNA-9 in vertebrate neural development Date: 27-09-2011

During neural development proliferating cells in the ventricular zone undergo repeated self-renewal to maintain the progenitor pool or, alternatively exit the cell cycle and differentiate into neurons. This process is regulated by the coordination of cell intrinsic signals and regional and temporal specific external cues which determine the type and the amount of neurons generated. After a neural progenitor has committed to differentiation the process of balancing internal and external signals continues during maturation to guide the initiation, elongation and branching of the axon. One major remaining question is how biological regulation can be integrated with the developmental context to produce a specific outcome. Here, I show that microRNAs and particularly, microRNA-9 (miR-9) plays a very important role in vertebrate neural development; a role, which is highly dependent on the context. In X. tropicalis, miR-9 reveals regional specific progenitor heterogeneity – it is required for cell cycle exit and differentiation throughout the neural tube, but forebrain progenitors additionally, and uniquely, require miR-9 for their survival. I have shown that the major miR-9 target in this context is the hairy and enhancer of split gene hairy1. When miR-9 is absent, hairy1 domain expands and selectively activates different signaling pathways in the forebrain and the hindbrain, culminating in the regional specific differences we observed. In the mouse, the homologue of hairy1 – Hes1 expression is oscillatory which is essential for progenitor maintenance. I have shown that Hes1 is also a subject of miR-9 regulation, which, in this context, is necessary for maintaining the oscillations. Furthermore, miR-9 is also regulated by Hes1, which leads to opposite-phase oscillations of miR-9 primary transcripts, but step-wise accumulation of the mature miR-9 form due to its stability. I propose that miR-9 levels act as output to measure the number of Hes1 cycles and at certain critical threshold levels this leads to dampening of the oscillations and allows progenitors to differentiate. MiR-9 is also expressed in differentiated neurons in the forebrain, where its function is completely unknown. We have shown that in this context it promotes axon branching and inhibits axon growth through the microtubule-associated protein 1B (MAP1B). We have also provided evidence that brain-derived neurotrophic factor (BDNF) can modulate this process by regulating miR-9 levels in the axon. Overall, these findings contribute to our understanding of neural development and provide novel explanation of how to cell fate decisions are integrated with context-specific signals to generate a specific outcome.

5 DECLARATION

No portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning.

The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the “Copyright”) and s/he has given The University of Manchester certain rights to use such Copyright, including for administrative purposes.

Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with licensing agreements which the University has from time to time. This page must form part of any such copies made.

The ownership of certain Copyright, patents, designs, trade marks and other intellectual property (the “Intellectual Property”) and any reproductions of copyright works in the thesis, for example graphs and tables (“Reproductions”), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions.

Further information on the conditions under which disclosure, publication and commercialization of this thesis, the Copyright and any Intellectual Property and/or Reproductions described in it may take place is available in the University IP Policy (see http://www.campus.manchester.ac.uk/medialibrary/ policies/intellectualproperty.pdf), in any relevant Thesis restriction declarations deposited in the University Library, The University Library’s regulations (see http://www.manchester.ac.uk/library/aboutus/regulations) and in The University’s policy on presentation of Theses.

6 ACKNOWLEDGEMENTS

I would like to thank my supervisor Nancy Papalopulu for her incredible help and support during all those years. It is extremely rare to work with somebody who cares so much not only about your scientific advancement but also about your personal career development. I cherish that and I am extremely grateful that I had a chance to learn so much during these years. Thanks for trusting me so much and believing in me, even when I did not.

I would also like to thank all the members of the lab for their help and support. For showing me the basics and helping me master the techniques. Also thanks for the great, relaxed atmosphere in the lab, which encouraged discussions and mutual help.

I would also like to thank my advisor Stephen Taylor, Matthew Ronshaugen for helping me set up microRNA – related assays and very helpful discussion, Dave Thornton for his support during the years and last, but not least, the Wellcome Trust for the generous funding which made this study possible.

7 Abbreviations

CNS central nervous system NSC neural stem cell FGF fibroblast growth factor NPC neural progenitor cell MHB mid-hindbrain boundary ZLI zona limitans intrathalamica NICD notch intracellular domain bHLH basic helix-loop-helix EGF epidermal growth factor PcG polycomb group complex ESC embryonic stem cell TRD transcriptional repression domain RISC RNA-induced complex UTR untranslated region miR/miRNA microRNA CR Cajal-Retzius

8 Rationale for Submitting in Alternative Format

During my PhD, I have examined the role of one microRNA – miR-9 in vertebrate neural development by focusing on 3 different model systems and uncovered a potential function for miR-9 in each of those:

- Using X. tropicalis as a model system, we have shown that miR-9 establishes progenitor diversity in space along the anterior-posterior axis by targeting the bHLH factor hairy1 – Chapter 2

- Using mouse and neural progenitor cell lines, we have uncovered a role for miR-9 to promote progenitor heterogeneity in time by modulating the oscillatory expression of Hes1 – Chapter 3

- In differentiated cortical neurons, we have shown that miR-9 regulates axon growth and branching by targeting and essential component of the cytoskeleton - MAP1B – Chapter 4

These findings describe different model systems and physiological processes, but are unified by the role of miR-9 in each of them. The first story was published in January, 2011 in Developmental Cell; the second part is planned to be submitted by the end of 2011 and the third story is currently under revision in Nature Neuroscience. Therefore, I consider the alternative format to be the most suitable to present my findings.

9 Author Contributions

Paper 1: BB designed and carried out the experiments, designed the figures and drafted the manuscript. AP participated in the neuroblastoma cell line experiments (Figure 5I). NP conceived the study, participated in its design and coordination and helped draft the manuscript.

Paper 2: BB designed and carried the experiments, designed the figures and drafted the manuscript. NP conceived the study, helped design the experiments and draft the manuscript.

Paper 3: FB and BB designed the experiments and drafted the manuscript, FB carried out function analysis in cortical neurons, BB performed in situ hybridizations, target identification and validation. PG and FG carried out the in vivo experiments. NP helped coordinate the study and participated in drafting the manuscript.

10 Chapter 1. Introduction

During development neural progenitor cells need to be able to interpret the external signals in order to decide whether to proliferate or differentiate. Mature neurons on the other hand need to be able to extend long projections (axons) to their respective targets in order to communicate with them. Both neuronal differentiation and neuronal maturation are regulated by the coordination of internal signals with external regional or temporal specific cues. Examining how these processes are integrated with developmental context to regulate cell fate and cell morphogenesis will improve our understanding of neural development. In addition, understanding the molecular signals that regulate progenitor heterogeneity is crucial for any cancer therapies based on neural stem cells, while understanding the mechanisms behind axon morphogenesis will improve our chances to develop a therapy for neurodegenerative diseases.

The following introduction provides an overview of the development of the central nervous system (CNS) focusing on both the external and intrinsic factors that drive progenitor heterogeneity during development. In addition I will also discuss the morphological changes that occur during the maturation of differentiated neurons and particularly how axonal specification, elongation and branching are regulated. I will then discuss the role of a class of non-coding RNAs called microRNAs during neural development, focusing on one brain specific in details – microRNA-9. The introduction ends with the specific aims and objectives for this investigation.

1. Development of the vertebrate central nervous system

The vertebrate brain is a complex, highly organized structure, which contains many different neuronal cell types and a variety of glial cells. It is the region of the CNS responsible for cognitive function, sensory perception and consciousness and as such has undergone a pronounced expansion and development during vertebrate evolution (Finlay and Darlington, 1995). Therefore, knowledge of how neural development is molecularly determined is fundamental to our understanding of the logic behind brain architecture.

11 Central nervous system develops from a small number of cells, which can proliferate, acquire multiple regional identities and produces many different types of cells. These progenitor cells have been identified as neural stem cells (NSCs) on the basis that they can self-renew in vitro and differentiate into multiple lineages (Temple et. al., 2001), which consist of many neurons of different types, as well as oligodendrocytes and astrocytes (Figure 1).

During the early stages of neural development, NSCs divide symmetrically, leading to a rapid increase of their numbers and expansion of the progenitor pool, but later they undergo asymmetrical Figure 1. The vertebrate central nervous system consists of divisions generating a three primary cell types – neurons, oligodendrocytes and astrocytes generated in distinct waves during development progenitor, which remains close to the ventricle, and a daughter cell that migrates outward (Anthony et al., 2004; Haubensak et al., 2004). Committed neuronal progenitors also become specified to produce a particular type of neurons – a step called subtype specification, which is particularly evident during the specification of the mouse cerebral cortex (Molyneaux et al., 2007).

Neurons are electrically excitable cells, which can process and transmit information via neurotransmitters, while oligodendrocytes and astrocytes are important for the structural integrity and proper functions of the brain. During development, the generation of these cell types occurs sequentially in distinct waves – neurogenesis first, followed by the generations of astrocytes and finally oligodendrocytes – Figure 1 (Hirabayashi and Gotoh, 2005; Miller and Gauthier, 2007).

It is apparent that such a highly complicated process must be carefully controlled at multiple levels to ensure the precise timing of lineage commitment

12 and differentiation. This regulation occurs in two ways – first, by the intrinsically established developmental program, and second - by positional cues and extrinsic signaling cascades so that the correct numbers of the different cell types are produced at the appropriate time points. And one of the fundamental unresolved challenges in developmental neurobiology is how general decisions whether neural progenitors should proliferate or differentiate are coupled with external context-specific signals to assure correct development and patterning.

1.1. Extrinsic signals regulating neuronal differentiation

Signaling pathways implicated in neuronal development provide positional information and ensure proper communication with the surrounding tissue. However, the outcome of a particular pathway could be both temporarily and spatially specific based on the context and the recipients of the signal. Some of the major pathways involved in neuronal differentiation that are of significant importance in this work include wnt signaling in the posterior brain regions (mid- and hindbrain), fibroblast growth signaling (FGF) in the thalamic region and notch signaling.

1.1.1. Wnt signaling

Wnt ligands are identified by their unusual post-translational modifications – palmitoylation at a conserved cysteine, which is essential for their proper function. It has been shown that wnt participates in many aspects of the development of the nervous system, with canonical wnt (mediated by β-catenin) promoting cell-cycle progression and progenitor proliferation in the early stages of cortical development in mouse (Chenn and Walsh, 2003). Interestingly, during later stages wnt signaling promotes neuronal differentiation through induction of proneural genes such as Ngn2 (Hirabayashi et al., 2004), suggesting that neural progenitor cells (NPCs) can respond to the same signals differently depending on their internal developmental program.

13 Figure 2. Wnt signaling controls the proliferation of neuronal progenitors by regulating G1 to S phase transition Wnt1 and Wnt3a are expressed in the most dorsal area of the hindbrain and spinal cord and form a mitogen DV gradient. Mechanism - wnt leads to stabilization of β-catenin and transcriptional activation of cell cycle promoting cyclins such as CyclinD1 and D2 (Megason and McMahon, 2002).

Binding of wnt ligands to the respective receptors (mostly members of the frizzled family) leads to stabilization of β-catenin and its nuclear translocation where it forms a complex with the transcription factor TCF, activating the transcription of various downstream target genes. In the context of neuronal development in the posterior brain (mid-, hindbrain and spinal cord), mitogenic activity of wnts expressed from the dorsal midline (Wnt1 and Wnt3a) often leads to promoting G1-S phase transition at least partially through induction of Cyclin D1 and D2 – Figure 2 (Megason and McMahon, 2002).

Thus wnt signaling from the roof plate act primarily as a morphogen gradient along the dorso-ventral axis to promote self-renewal throughout the spinal cord, while increasing distance from the source of the signal leads to mitotic arrest, suggesting that it has the potential to determine both patterning and cell fate (Megason and McMahon, 2002).

1.1.2. Fibroblast Growth Factor (FGF) signaling

Several growth and differentiation factors are selectively expressed in restricted domains in the telencephalon and the mid-hindbrain boundary (MHB), and have been proposed to pattern the developing neuroepithelium. Among them members of the FGF family are expressed in the rostral patterning centres in the telencephalon including Fgf8, Fgf15, Fgf17, Fgf18 (Bachler and Neubüser, 2001). Of

14 those Fgf8 signaling is of particular importance since it has been shown to pattern the developing thalamus and participate in the specification of the ventral telencephalon (Kataoka and Shimogori, 2008; Storm et al., 2006).

Fgf8 expression in the developing diencephalon overlaps with Sonic Hedgehog (Shh) and has been shown to be restricted primarily to the Zona Limitans Intrathalamica (ZLI) - a boundary region between the thalamic and the prethalamic progenitor domains (Larsen et al., 2001), which is determined by the expression of Otx1l, Otx2 and Irx1b (Scholpp et. al., 2007). Boundary regions usually represent areas formed by specialized progenitor cells, which have properties different that those in the neighbouring compartments - slow proliferation and delayed neurogenesis (Kiecker et. al., 2005; Baek et. al., 2006). ZLI and has been shown to regulate the development of the adjacent areas through secretion of shh and fgf8 ligands (Storm et al., 2006; Vieira et al., 2005) thus acting as an organizing centre.

In addition to patterning, Fgf8 has been shown to promote progenitor proliferation and suppress differentiation through regulation of Wnt8b, Bmp4 and Shh and partially through activation of the forkhead transcription factor FoxG1 (Storm et al., 2006) (which also coordinated wnt and shh signaling to pattern the telencephalon – (Danesin et. al., 2009). In addition, activation of the MAPK and mTOR pathway through Fgf8 could also contribute to its anti-differentiation role during forebrain development (Hentges et al., 2001; Shinya et al., 2001).

Figure 3. Fgf8 signaling in neural development (A) Fgf8 is expressed at the border region between p2 and p3 in the developing diencephalon (ZLI), where it overlaps with sonic hedgehog. (B) Fgf8 signaling regulates progenitor maintenance through FoxG1, MAPK and mTOR signaling pathways (Storm et al., 2006; Suzuki-Hirano and Shimogori, 2009)

15 1.1.3. Notch signaling

Notch signaling is one of the major ways of direct cell-to-cell communication between adjacent cells during development and it is considered to be one of the key regulatory mechanisms for both embryonic and adult neurogenesis. There, it has been shown to regulate glial formation and neuronal differentiation (Ables et al., 2011; Louvi and Artavanis-Tsakonas, 2006), as well as influence the formation of boundary regions such as the mid-hindbrain boundary (MHB) and the rhombomere boundaries in the hindbrain (Yoon et. al., 2005).

Notch signaling acts by a mechanism called lateral inhibition, which prevents two neighbouring cells from adopting neuronal fate simultaneously – Figure 4A. In the classical view of lateral inhibition, cells that express higher levels of the proneural genes Mash1 and Ngn2 induce the expression of Notch ligands (such as delta-like 1 - Dll and serrate), which are localized on the plasma membrane. They in turn bind to Notch receptors on the neighbouring cells leading to cleavage of the notch intracellular domain (NICD) and its translocation into the

A. B.

Figure 4. Notch signaling generates Hes1 molecular oscillations A. Mechanism of lateral inhibition – Dll1 activation leads to transcription of the notch effector genes Hes1/Hes5 in the neighbouring cell, preventing them from adopting the same cell fate (Kageyama et al., 2008). B. Hes1 oscillates with a period of 2-3h due to negative feedback, delay and high instability of the protein and mRNA.

16 nucleus (Selkoe and Kopan, 2003). There, NICD forms a complex with the transcription factor RBPj, activating the transcription of Notch effectors such as the basic helix-loop-helix (bHLH) factors Hes1/5 in adjacent cells (Ohtsuka et al., 1999). They, in turn, repress the expression of proneural genes, thereby preventing the premature depletion of the neuronal pool.

Initially, studies on notch signaling focused on the idea that small stochastic differences in Hes1 levels were amplified through lateral inhibition to give rise to its salt-and-pepper pattern of expression in the mouse telencephalon (Artavanis- Tsakonas et al., 1999). However, recent evidence suggests that Hes1 and other proneural genes are in fact expressed in an oscillatory manner with a period of 2- 3h (Shimojo et al., 2008). This type of gene expression is remarkably similar to the Hes1 and Hes7 expression during somitogenesis, with one major difference – while Hes1 oscillations are synchronized in the presomitic mesoderm (Masamizu et. al., 2006), they appear to be asynchronous (in single cells) in neural progenitors. The oscillatory pattern of Hes1 is essential to maintain neural progenitors in a proliferative state as persistent high levels of Hes1 result in lower proliferation rate, while low Hes1 levels lead to neural differentiation (Shimojo et al., 2008).

There are three essential prerequisites to drive Hes1 molecular oscillations: a negative feedback loop generated by Hes1 protein repressing its own transcription; delay due to biological processing (transcription, splicing, translation, etc. …) and, importantly, the high instability of both Hes1 protein and mRNA (Hirata et al., 2002) – Figure 4B.

1.2. Intrinsic mechanisms of cell fate regulation during neural development

In vitro culture systems, where embryonic stem cells are triggered to differentiate into neural progenitors and subsequently neurons, as well as clonal analysis of NPCs have revealed that cell-intrinsic programmes play a key role in the “responsiveness” of NPCs to external cues during development. This is particularly evident during the switch from neuron-generating mode to gliogenesis in single clones (Shen et al., 2006), which mimics remarkably well the in vivo timing. This implies the existence of an intrinsic ‘‘timer’’ that modulates the fate of NPCs over time. How does this intrinsic timer relate to the extrinsic instructive signals?

17 One example of such coordination is the accumulation of the epidermal growth factor (EGF) receptor during cortical development resulting in increased responsiveness to its ligands and enhanced commitment to a glial fate (Burrows et al., 1997). However in other examples (such as the differential effect of wnt signaling on cortical development) the exact mechanism responsive for the switch in response to the same signaling pathway remains elusive (Hirabayashi and Gotoh, 2005). The two major epigenetic mechanisms, considered to be essential to reinforce the timing and the outcome of neural development, include chromatin modifications, DNA methylation and small non-coding RNAs (discussed in chapter 3).

1.2.1. Chromatin state determines the probability of transcription

Recently, there is considerable evidence that chromatin modifications and in particular, histone H3 methylation at specific lysine residues within a promoter region, determines whether a particular gene would be transcribed or not (Martin and Zhang, 2005). These modifications can be divided in two broad groups: activations marks (H3K4me3, H3K36me3) – mediated primarily by members of the trithorax family of protein; and restrictive marks (H3K9me3, H3K27me3) – mediated by polycomb group (PcG) complexes (Martin and Zhang, 2005).

Importantly, there is increasing evidence that histone methylation plays an important role in regulating cell fate during NSCs differentiation. Initially, in embryonic stem cells (ESCs) neural genes such as Ngns, Pax6, Sox1, Sox3, Nkx2.2 and Mash1 have both repressive and permissive marks (a state referred to as “bivalent”) but lose the repressive mark during differentiation into the neural lineage (Mikkelsen et al., 2007). During that initial specification genes related to pluripotency (oct4, Nanog) acquire H3K27me3 marks and therefore become silenced (Hirabayashi and Gotoh, 2010). At the same time, a subset of genes functioning in terminally differentiated neurons become bivalent during ESC to NPC transition and gradually lose the repressive marks during neuronal differentiation, ultimately leading to expression of those genes only in differentiated neurons (Mohn et al., 2008). Chromatin modifications have been shown to play an important role during the neuronal-to-astrocytes transition during NPC differentiation, specifically by accumulating of restrictive histone

18 marks on the promoters of neurogenins by PcG proteins leading to a gradual decrease in their expression (Hirabayashi et al., 2009).

However, a major unresolved question remains how, during the course of neural fate commitment, chromatin modification complexes discriminate between their targets, given the fact that they do not seem to have sequence specificity (Schuettengruber et al., 2007). Several mechanisms including specific recruitment of PcG complexes by long non-coding RNAs to particular loci, DNA methylation patterns and chromatin structure have been proposed to explain this controversy (Schuettengruber et al., 2007).

1.2.2. DNA methylation

The methylation status of a given promoter is an important indicator for the activity of this gene, as heavily methylated DNA regions are transcriptionally silent (Goll and Bestor, 2005). The predominant form of DNA methylation in mammals is the symmetric modification of cytosine in the 5’ position in CpG nucleotides (Shi et al., 2008). Importantly, upon commitment of ESCs to the neuronal lineage several hundred promoters including pluripotency and germ-line specific genes become DNA methylated, which contributes at least partially to their silencing (Mohn et al., 2008). In addition, promoters marked by repressive histone modifications often become methylated, leading to long-term silencing (Mohn et al., 2008).

One of the most prominent genes involved in Xenopus neurogenesis is the methyl-cytosine binding protein MeCP2 (Stancheva et al., 2003). It binds to methylated CpG islands and silences nearby genes through an additional transcriptional repression domain (TRD). It has been shown to be important for maturation and maintenance of neurons at late stages but not critical for embryonic neurogenesis in the mammalian brains (Kishi et al., 2005).

The combination of cell intrinsic changes and external guidance cues during neurogenesis leads to some progenitors exiting the cell cycle and initiating neuronal differentiation, which is accompanied with migration (in some cases over long distance) to reach their predetermined position in the cortex and maturation of the newly postmitotic neurons.

19 2. Morphological changes during terminal neuronal differentiation

One of the most important morphological processes during neuronal maturation is the polarization of the neuron and the formation two types of projections, which usually consist of a single long axon and multiple dendrites. An axon is usually the longest cellular process, and transmits signals to cells that could be located meters away. Dendrites, on the other hand, are elaborate short, but extensive branches, and collect signals from a large number of neurons. The information is usually received by the dendrites, processed in the cell body and transmitted through the axon, which underlies the directional flow of information transfer in the central nervous system. Therefore it is important to understand the processes that govern both the specification of the axon and its development, including elongation and branching.

2.1.1. Axon specification

The process of axon-dendrite polarization is usually initiated during neuronal migration in vivo. While some neurons derive this type of polarity directly from the apical-basal polarity of their progenitor cells (retinal ganglion cells and bipolar neurons in the retina), most cortical progenitors undergo extensive morphological changes leading to a polarized outgrowth of their dendrites and axon (Barnes and Polleux, 2009). In vitro cultures of cortical neurons isolated from E17-E18 mouse forebrain are quite often used to examine signaling pathways which regulate axon specification and development, however a further validation using in vivo methods such as in utero electroporation or gene knockout studies is desirable to confirm potential regulatory molecules.

PIP3 signaling and more specifically, the lipid kinase phosphatidylinositol 3-kinase (PI3K) and its counterpart PTEN, has been shown to play a role in the specification of the axons by means of inhibition and overexpression studies (Jiang et al., 2005; Yoshimura et al., 2006). PIP3 enriched membrane regions have been shown to occur selectively within single neurite (Menager et al., 2004), which is then specified to become an axon in part due of local enhancement of AKT signaling through PIP3 recruitment (Shi et al., 2003). In addition, conserved polarity pathways such as LKB1 and SAD kinases (Par4/Par1 dyad) have been

20 identified using genetic perturbation studies to be important for establishing axonal identity (Barnes and Polleux, 2009).

2.1.2. Axon elongation

Once the axon becomes specified, it undergoes an initial process of elongation and growth, which is followed by extensive branching, once it reaches its target. This requires the regulation of the dynamic assembly of cytoskeletal polymers - filaments of actin and a/b-tubulin that are highly dynamic and undergo rapid cycles of polymerisation and depolymerisation (Dent et al., 1999) – Figure 5. Even though axon initiation and elongation are largely intrinsic properties of neurons in vitro, there are multiple extracellular cues regulating these processes.

Figure 5. Specification and elongation of the axon during neuronal maturation During neuronal polarization one of the minor neuritis extends rapidly to become the future axon, a process driven by actin and microtubule filaments. Growth cone of the future axon is larger and has more dynamic cytoskeleton compared to the growth cone of the dendrites, which allows for extended growth (Stiess and Bradke, 2011).

Multiple signaling pathways such as RAF/MEK/ERK, PIP3 and JNK have been shown to regulate axon elongation often converging on the cytoskeleton (Polleux and Snider, 2010). In addition, GSK-3 signaling and its effectors (such as APC) have critical role both in the initiation and elongation of the axon presumably by enhancing microtubule stability (Aoki and Taketo, 2007; Shi et al., 2004). While

21 axon assembly is achieved by microtubule stabilization and polymerization in the proximal part of the axon (Stiess and Bradke, 2011), distal parts close to the growth cone require dynamic regulation between actin and microtubule filaments in order to destabilize the cytoskeleton at the growing tip. One example of a molecule regulating this process is the microtubule-associated protein MAP1B. It is the first MAP that is expressed and is especially prominent in developing neurons, being highly concentrated at the distal tip of growing axons where it associates with tyrosinated microtubule (Black et. al, 1994). In addition, neurons deficient of MAP1B have reduced rate of axonal elongation, while the number of terminal branch points is increased (Gordon-Weeks and Fischer, 2000; Montenegro- Venegas et al., 2010). This suggests that MAP1B can provide a mechanism to coordinate axon elongation and branching, however the regulatory pathways upstream of it are largely unknown.

2.1.1. Axon branching

Axon branching can occur at multiple points during the development of the axon: arborization (extensive repetitive branching once the axon has reached its target region), bifurcation (also occurs at terminal points but generates two ends going in the opposite direction) and collateral formation (branch point is away from the axon terminals and often innervates targets different from the main axon shaft) – Figure 6 (Gibson and Ma, 2011). Branch points form by protrusion of actin filaments, which is subsequently invaded by axonal microtubules as the branch matures and continues to extend. Because this process is dependent on dynamic microtubule reorganization at the branch point, regulation of the polymerization and destabilization of the cytoskeleton is crucial for branching, in addition to its role in elongation (Gallo et. al., 2011).

Signaling cascades regulating axonal elongation can often be utilized to produce branches in a different context. Brain-derived neurothropic factor (BDNF) is a good example of a signaling which can regulate both growth and branching. The biological function of mature BDNF is mediated by the TrkB receptor tyrosine kinase, which activates the downstream MAPK, phospho- tidylinositol-3 kinase and PLC-γ pathways, which converge on regulating cytoskeleton dynamics (Segal, 2003). Importantly, it has been recently shown that transient activation of BDNF

22 signaling leads to neurite and axon elongation, while sustained BDNF results in increased branching (Hoshino et al., 2010; Ji et al., 2010).

Figure 6. Axon branching is an essential step during neuronal maturation. A. An interneuron derived from layer 2/3 from the rat cortex shows the complexity that axon branching can create (Schmidt and Rathjen, 2010) B. Mechanisms to create axon branches are conserved in the vertebrate nervous system (Gibson and Ma, 2011).

2.1.2. Local control on axonal mRNA translation

The extreme distances between the cell body and the distal parts of the axons makes it difficult to transport all necessary proteins from the soma. Local protein synthesis and regulation of mRNA synthesis at the distal parts of the axons are thought to contribute to this process. However there is accumulating evidence that only a small fraction of the axonal proteins are locally synthetized. Most of the locally translated mRNAs appear to have functions in enabling rapid morphological plasticity of the growth cone in response to a signal rather than to maintain the protein composition in distal parts of the axon (Hengst et. al., 2007).

During axonal development it has been shown that translation machinery is present in the distal parts of the axon and local mRNA translation occurs (Leung et al., 2006). However once axons have reached their final destination and synapses have formed they seem to lack functional translation machinery (Kleiman et al., 1994). The majority of locally translated mRNAs are usually visualized as small puncta present at the axon on the growth cone and have been implicated in regulating the plasticity of the cytoskeleton and its ability to respond to external cues (Wu et al., 2005). However, only a subpopulation of these mRNAs undergoes translation at any given point (Leung et al., 2006; Wu et al., 2005).

23 Since there is increasing evidence that a novel class of non-coding RNAs – microRNAs can be important for the local suppression of mRNA in the axon, in addition to their well-established role in the earlier steps of neural development, I will devote the next section to describing their biogenesis and proposed function.

3. MicroRNAs in neural development

3.1. Biogenesis, mechanism and function

Members of the miRNA family were initially discovered as small temporal RNAs that regulate developmental transitions in C. elegans (Pasquinelli and Ruvkun, 2002), but since then they have been found in both vertebrate and invertebrate. There are estimated to be at least 700 miRNAs (and there might be as many as 1000) in the human genome, comprising 1–4% of all expressed human genes, which makes miRNAs one of the largest classes of gene regulators (Bentwich et al., 2005; Berezikov et al., 2005).

MicroRNAs are short regulatory non-coding RNAs (19-25nt) encoded in the genomes of invertebrates, plants and vertebrates. Many microRNAs reside in the intronic sequences of pre-mRNAs, suggesting that at least some miRNAs are not transcribed from their own promoters (Aravin et al., 2003). Nevertheless, the majority of the miRNAs in human derive from independent transcription units. Clustering of miRNA-encoding genes, both functionally related and not, is common for Drosophila genome and they are often expressed as a multi-cistronic primary transcript (Robins et al., 2005). However, the majority of human miRNA genes are isolated and not clustered (Lim et al., 2003).

24 MicroRNAs are generally transcribed as long (>1kB) primary transcripts by RNA Polymerase II (Pol II) which appear to be capped and polyadenylated even in the absence of clear polyadenylation signals (Ohler et al., 2004). The primary miRNA transcripts are characterized by specific hairpin secondary structure, which is recognized by the RNase III enzyme, Drosha (Lee et al., 2003), and the double-stranded-Pasha (also known as DGCR8), RNA-binding protein and processed into ~70-nucleotide pre-miRNAs (Lee et al., 2002). Importantly, Drosha cuts the RNA duplex with a staggered cut typical for RNAse III endonucleases, producing 5’phospate and ∼2nt 3’ overhang (Basyuk et al., 2003). This can lead to destabilization of the pre-miRNA ends, which could be important for the

Figure 7. Biogenesis of microRNAs MicroRNAs are processed from primary transcripts by the endonucelase Drosha and its cofactor Pasha and exported to the cytoplasm through RAN- GTPase/Exportin5. Subsequent cleavage by another endonuclease – Dicer leads to the formation of the miRNA/miRNA* duplex. One strand is incorporated in the RISC-silencing complex and represses the translation of its mRNA targets or initiates RNA degradation through deadenylation and decapping (Stefani and Slack, 2008).

25 subsequent processing and selection of guiding strand. The precursor miRNAs are then exported into the cytoplasm by the RAN GTP-dependent transporter Exportin 5 (Lund et al., 2004) and undergo an additional processing step which involves the RNAse III enzyme, Dicer. Dicer is thought to recognize preferentially the 5’phosphate and 3’ overhang of the base of the stem loop (Hutvagner et al., 2001). Subsequently, it cuts both strands leading to the excision of the terminal base pairs and pre-miRNA loop and the liberation of a dsRNA of approximately 22 nucleotides in length (miRNA:miRNA* duplex). The biogenesis and processing of a microRNA is schematically presented in Figure 7.

For most of the known miRNAs one strand is then preferentially incorporated in the functional RNA-induced silencing complex (RISC) and guides the binding to its target mRNA, while the other strand (called miRNA*) is present at much lower copy numbers and is usually targeted for degradation (Lau et al., 2001). However, some of the miRNA* retain functional importance and are also incorporated into the RISC silencing complex, albeit at lower frequency (RFAM database).

The mature miRNA strand is preferentially retained in the functional miRISC complex and guides the binding to its target mRNA. Various hypotheses to explain the mechanism of miRNA function have been proposed, including translational inhibition at the level of initiation and elongation, rapid degradation of the nascent peptide, mRNA sequestration into P bodies and mRNA Figure 8. miRNA target recognition and silencing mechanism (Liu et al., 2007) degradation – Fig. 7b (Pillai et al., 2007). However recent evidence suggest that the majority of the mammalian microRNAs regulate their target expression primarily at the RNA level through

26 deadenylation and decapping of the message, contrary to what was thought before (Giraldez et al., 2006; Guo et al., 2010; Lim et al., 2005).

Experimental and bioinformatics approaches have shown that the most important determinant of target RNA recognition is a perfect or near-perfect match between the proximal (5') region of the miRNA and the mRNA, also known as the "seed" region (Lewis et al., 2003). This model has recently been refined (Lagos-Quintana et al., 2002) to account for the presence of secondary structures and other features of the 3' untranslated region (UTR) sequence surrounding the target site, and for the ability of complementarity at the 3' end of the cognate miRNA to compensate for imperfect seed matching (Grimson et al., 2007), (Long et al., 2007). Recent findings have challenged the classic assumption that miRNAs can target only the 3’UTR of a given mRNA and have shown that miRNA to regions in the coding sequence are also effective in translation repression (Tay et al., 2008a), however there is very little evidence supporting this and the majority of the targets that have been identified so far are located in the 3’UTR.

3.2. microRNA function in the development of the nervous system

Because of their function as post-transcriptional regulators of gene expression and the ability to target multiple genes, microRNAs are attractive candidates for regulation of key developmental processes, particularly in the nervous system. More than 70% of the known mammalian miRNAs are expressed in the brain (Cao et al., 2006) and the levels of some of them change dramatically during neurogenesis and brain development (Miska et al., 2004). MiR-124 and miR-128 are preferentially expressed in neurons; miR-23, miR-26 and miR-29 are predominant in astrocytes while miR-9 and miR-125 are more evenly distributed (Smirnova et al., 2005).

Initial studies on the function of microRNAs in neurogenesis focused on genetic analysis of animals lacking crucial components of the miRNA-processing machinery. Giraldez et al. demonstrated that knockout of Dicer leads to severe defects in brain morphogenesis and neuronal differentiation in zebrafish, including establishing the mid-hindbrain boundary and axon specification in the brain, confirming the importance of microRNAs in neural development (Giraldez et al., 2005). More recently, conditional Dicer knockout in mouse forebrain was shown to

27 lead to a reduction in the size of the neonatal cortex, as well as marked reduction in radial thickness (De Pietri Tonelli et al., 2008). In this study microRNAs were shown to be required for proper neuronal differentiation and survival of post- mitotic neurons, as implicated by increased neuronal (starting at stage E12.5) and progenitor (after E13.5) apoptosis.

A major focus of the research so far has been the most abundant miRNA in brain – miR-124. It is expressed specifically and abundantly in the mouse CNS and in P19 pluripotent cells as they differentiate into neurons (Sempere et al., 2004). It is thought to help acquire neuronal identity by directly repressing a large number of target genes including the polypyrimidine tract-binding protein PTBP1, an important splicing regulator (Makeyev et al., 2007) and the small C-terminal domain phosphatase-1 SCP1, which is an important activator of REST (Visvanathan et al., 2007). Interestingly, REST is thought to repress miR-124 forming an inhibitory loop, but is also found on the promoter sequence of other brain-specific miRNAs such as miR-9 and miR-132 (Conaco et al., 2006). Expression of a dominant negative form of REST lead to an increase in the levels of both miR-9 and miR-124, while introducing exogenous REST inhibits the expression of these miRNAs. However, as stated above, miR-9 and miR-124 have different expression patterns, which implies that there are additional levels of regulation involved.

One of the important insights about the role of microRNAs in the control of neural progenitor maintenance and/or neural differentiation comes from a study on NSCs (Krichevsky et al., 2006). They demonstrated that miR-124a and miR-9, acting together, are able to affect neural differentiation by shifting the fate towards more neurons at the expense of astrocytes. Furthermore, selective inhibition of miR-9 by 2’-orthomethyl oligonucleotides not only negatively affected neuronal differentiation, but also led to significant increase of phosphorylated STAT3 levels, which could be important for the switch to glial generating division.

During later stages of neurogenesis - in neuronal maturation, several dendritic miRNAs have been identified to regulate synaptic function. MiR-134 is an example of a microRNA enriched in the dendrites of hippocampal neurons, which regulates the size of the dendritic spines through its target Limk1 (Schratt et al.,

28 2006). In addition, the component of the RISC complex in drosophila - Armitage has been shown to participate in the regulation of α-calcium/calmodulin- dependent protein kinase II (α-CaMKII) at the synapse, although the precise microRNAs involved have not yet been identified. In mammalian neurons, Armitage homologue – MOV10 was also found to play an important role in α- CaMKII and Lypla1 regulation at the synapse, which was mediated by miR-138 (Banerjee et al., 2009). The importance of microRNA-mediated silencing for axonal pathfinding and growth was confirmed shown by the presence of the RISC complex at the 3’UTR of RhoA (an important factor for growth cone collapse and local protein synthesis) (Hengst et al., 2006). However, the actual microRNAs involved in their precise function remains largely unexplored (Natera-Naranjo et al., 2010).

3.3. miRNA-9 – an important regulator of vertebrate neural development

MiR-9 is a brain-specific microRNA, which belongs to a highly conserved family found in worms, flies and vertebrates (Lagos-Quintana et al., 2002), and has been show to be expressed in the eye, the spinal cord and the CNS, particularly in the proliferating cells of the brain (Kloosterman et al., 2006). While several functions of miR-9, mediated by a variety of targets, have been proposed, often with controversial conclusions, its precise role in neural development and whether that is context-dependent remains unknown.

Initial studies of miR-9 focused on its role in Drosophila neurogenesis, where Cohen et al. found that it maintains cells in proliferative state by repressing the proneural gene senseless (Cohen et al., 2006). This observation, however, was disputed by recent finding about the role of miR-9 in vertebrate neurogenesis, which concluded that miR-9 promotes differentiation (Krichevsky et al., 2006; Leucht et. al., 2008; Zhao et. al., 2009; Shibata et. al., 2011), although the proposed target was different.

Leucht et al. described a role for miR-9 in the organization of the mid- hindbrain boundary in zebrafish (Leucht et al., 2008). They showed that miR-9 targets several components of the FGF pathway and thus defines the boundary of the mid-hindbrain organizer. Importantly, while most of the effects of miR-9 on the MHB seemed to be rescued by a Her5 target protector (a morpholino, which selectively disrupts the interaction between the 3’UTR of a given mRNA and the

29 miRNA), the general effect of miR-9 on neurogenesis could not be attributed to her-5/9 inhibition alone. This supports the idea that miR-9 might be regulating an additional set of genes in different areas of the brain.

Shibata et al. described a novel function for miR-9 in the mouse forebrain (Shibata et al., 2008). They propose that miR-9 selectively targets the transcription factor FoxG1 thereby controlling the proper generation of the Cajal-Retzius cells (CR) (one of the first newly born neurons in the mouse cortex). Furthermore, miR- 9 electroporation at E11.5 leads to ectopic induction of CR cells while miR-9 inhibition by antisense nucleotides causes the reduction of the CR cells molecular marker Reelin, as well as the factors NeuroD1 and p73. However, the authors did not perform rescue or target protector experiments to unambiguously show that the effects of miR-9 overexpression on neurogenesis are due to FoxG1, thus one cannot exclude the possibility that FoxG1 downregulation and the defects in CR neurons are secondary effects due a general problem with neurogenesis.

Recently, the same authors characterized the phenotype of miR-9-2/miR-9- 3 knockout mouse (Shibata et al., 2011). Lack of miR-9 led to a reduction of the cortical layer, malformation of each cortical projection, and defects in the tangential migration of interneurons into the pallium. Furthermore, the authors suggested that early during cortical development (E11-E16.5) miR-9 regulates directly FoxG1 while the increased expression of the RNA-binding protein Elav2 after E16.5 counteracts this repression. In addition to FoxG1, Meis2 and Gsh2 were suggested as putative miR-9 targets. However, similar to the authors’ previous approach the lack of rescue or target protector experiments raises doubt whether the effects observed were direct or as a consequence of a general perturbation of neurogenesis.

Using adult neural stem cells, miR-9 was found to regulate the expression of the nuclear transcription factor NR2E1 (TLX), forming a negative feedback loop (Zhao et al., 2009). In addition to the inhibitory effect on stem cell proliferation, miR-9 overexpression leads to precocious migration, which the authors attributed to the reduced levels of TLX. However, overexpression is notoriously prone to artifacts and side effects and should be accompanied by loss-of-function studies, which the authors did not carry out.

30 Recently, loss of miR-9 has been shown to suppress proliferation in neural progenitors derived from human ES cells, albeit by a small degree. In this system, loss of miR-9 promoted migration of neural progenitors (Delaloy et al., 2010). While the authors convincingly showed that these defects are mediated by the direct target of miR-9 – stathmin (regulates microtubule stability), it is unclear how these findings obtained from an in vitro study are directly applicable to embryonic development.

From these studies the emerging theme is that in most systems, miR-9 is necessary for neuronal differentiation but the effect on proliferation is highly variable. Differences in the results obtained may be partly due to different model systems or experimental methodology, however, these discrepancies also raise the possibility that the function of miR-9 in neurogenesis and proliferation is highly context dependent. In addition even though there is evidence that miR-9 is expressed in mature neurons in the telencephalon (Deo et al., 2006; Leucht et. al., 2008), its function there is completely unknown.

Aims and objectives

Determining what the major function of miR-9 is during neural development and whether it is context dependent on regional or temporal differences is the major objective of this thesis. In order to address this problem, I’ll try to answer two major questions:

1) What is the function of miR-9 in neural progenitors?

Most studies looking at miR-9 function use in vitro generated neural stem cells or have concentrated only in one specific area of the CNS (MHB, telencephalon). Using X. tropicalis as a model organism will allow me to examine miR-9 function throughout the neural tube, focusing on potential differences. Specific aims include:

• Examine the expression of both primary miR-9 transcripts and the mature form of miR-9 during development • Investigate the consequences of miR-9 knockdown for neural development

31 • Identify the molecular mechanism of miR-9 function – major targets and downstream signaling cascades

Using such detailed analysis, I will then compare the expression and the major mechanism of function of miR-9 between Xenopus and mouse and will try to answer the question if and how miR-9 function has changed during evolution.

2) What is the function of miR-9 in differentiated neurons?

MiR-9 is present in differentiated neurons, yet its role is completely unknown. Using mouse cortical neurons in vitro, I will determine the function of miR-9 during neuronal maturation focusing on three major questions:

• How is mature miR-9 expressed in differentiated neurons? • What are the consequences of its overexpression and loss-of-function and can we attribute them to one or more major targets? • What is the molecular basis of miR-9 function in neurons and can it be regulated by extracellular signaling?

Addressing these questions will contribute to our understanding of neural development and will provide a potential mechanism to integrate cell fate decisions with context-specific signals.

Chapter 2. microRNA-9 reveals regional diversity of neural progenitors along the anterior-posterior axis

Publication Number One:

Bonev, B., Pisco, A. and Papalopulu, N. (2011) 'MicroRNA-9 Reveals Regional Diversity of Neural Progenitors along the Anterior-Posterior Axis', Dev Cell 20(1): 19-32.

33 Developmental Cell Article

MicroRNA-9 Reveals Regional Diversity of Neural Progenitors along the Anterior-Posterior Axis

Boyan Bonev,1 Angela Pisco,1 and Nancy Papalopulu1,* 1Faculty of Life Sciences, Michael Smith Building, University of Manchester, Oxford Road, Manchester M13 9PT, UK *Correspondence: [email protected] DOI 10.1016/j.devcel.2010.11.018

SUMMARY neurons in a region-specific manner (Falk et al., 2008; Jessell, 2000; Lee and Pfaff, 2003; Marklund et al., 2010). Neural progenitors self-renew and generate neurons MicroRNAs are a class of small noncoding RNAs, which have throughout the central nervous system. Here, we been shown to play key roles in many developmental processes uncover an unexpected regional specificity in the including stem cell proliferation and differentiation (Gangaraju properties of neural progenitor cells, revealed by and Lin, 2009; Kosik, 2006; Stefani and Slack, 2008). They the function of a microRNA—miR-9. miR-9 is ex- are particularly attractive for their potential to coordinate the pressed in neural progenitors, and its knockdown response of many target genes, thereby acting as point of information integration. Knockout of the essential component of results in an inhibition of neurogenesis along the microRNA-processing Dicer has shown that microRNAs are anterior-posterior axis. However, the underlying indispensable for proper neural development in zebrafish (Giral- mechanism differs—in the hindbrain, progenitors dez et al., 2005) and mouse (De Pietri Tonelli et al., 2008), fail to exit the cell cycle, whereas in the forebrain although the key miRs and their precise molecular targets have they undergo apoptosis, counteracting the prolifera- not been fully examined. tive effect. Among several targets, we functionally miR-9 is a highly conserved microRNA, which is expressed identify hairy1 as a primary target of miR-9, regulated primarily in the CNS (Kapsimali et al., 2007; Wienholds et al., at the mRNA level. hairy1 mediates the effects of 2005). In vertebrates the function of miR-9 has been studied in miR-9 on proliferation, through Fgf8 signaling in the fish and mice with loss and gain-of-function approaches. In the forebrain and Wnt signaling in the hindbrain, but fish, miR-9 has been shown to be necessary to define the mid- affects apoptosis only in the forebrain, via the p53 hindbrain boundary (MHB), a non-neurogenic boundary zone with organizer properties (Leucht et al., 2008). However, with pathway. Our findings show a positional difference respect to the role of miR-9 in neuronal differentiation and prolif- in the responsiveness of progenitors to miR-9 deple- eration, the results obtained by the loss-of-function experiments tion, revealing an underlying divergence of their in different systems have not always been consistent. In the properties. anterior hindbrain, where miR-9 is expressed, a decrease in neuronal differentiation was reported, which, however, was not accompanied by an increase in progenitor proliferation (Leucht INTRODUCTION et al., 2008). This is similar to the result obtained in the embryonic mammalian forebrain, where miR-9 knockdown caused a reduc- During neurogenesis, proliferating neural cells (neural progenitor tion of early-born Cajal-Retzius neurons but did not have an or neural stem cells), located in the ventricular zone (VZ), effect on progenitors (Shibata et al., 2008). In another study, undergo self-renewal to replenish the progenitor population or, miR-9 knockdown caused a reduction in neural progenitors alternatively, engage in asymmetric divisions associated with derived from mouse ES cells, accompanied by a slight increase the generation of neurons (Go¨ tz and Huttner, 2005). The process in GFAP+ astrocytes, although the effects on proliferation were of neurogenesis is tightly coupled with the process of regional not directly tested (Krichevsky et al., 2006). However, the oppo- specification, which dictates the identity of neurons born in site result was obtained in neural stem cells derived from adult different areas of the central nervous system (CNS) (Gaspard mammalian forebrain, where miR-9 knockdown caused a small and Vanderhaeghen, 2010). Neural stem cells themselves have increase in proliferating cells (1.37-fold) but did not change different positional identity and can give rise to tumors with differentiation (Zhao et al., 2009). Finally, in neural progenitors different signatures depending on their origin (Lee da et al., derived from human ES cells, loss of miR-9 has been shown to 2010; Palm and Schwamborn, 2010). suppress proliferation, albeit by a small degree. In this system, However, how regional specificity is integrated with the funda- loss of miR-9 promoted migration of neural progenitors (Delaloy mental cellular decisions that drive neurogenesis is not well et al., 2010). From these studies the emerging theme is that in understood. Both intrinsic and external factors are thought to most systems, miR-9 is necessary for neuronal differentiation, contribute to the correct execution and the transition from the but the effect on proliferation is highly variable. Differences in transcriptional programs of neural stem cells to differentiated the results obtained may be partly due to different model

Developmental Cell 20, 19–32, January 18, 2011 ª2011 Elsevier Inc. 19 Developmental Cell MicroRNA-9 Function in Xenopus Neurogenesis

systems or experimental methodology; however, these discrep- miR-9b, consistent with previous results (Walker and Harland, ancies also raise the possibility that the function of miR-9 in neu- 2008). rogenesis and proliferation is highly context dependent. During neural development progenitors divide in the VZ, and Here, we have undertaken a systematic analysis of miR-9 daughters that exit the cell cycle, migrate laterally to the marginal expression and function along the anterior-posterior (AP) axis zone where they differentiate (Figure 1C). Sections showed that during X. tropicalis development and uncovered an unexpected miR-9 transcripts have widespread expression in the forebrain regional specificity. In the forebrain, miR-9 is expressed in but were restricted to the VZ in the more posterior areas (Fig- both neural progenitors and developing neurons, whereas in ure 1D; Figure S1C). These spatial differences became even the more posterior regions of the brain (mid- and hindbrain), it more apparent later during development (stage 36, Figure S1D). is restricted to neural progenitors only. Using loss-of-function To determine whether miR-9 was also present in post-mitotic experiments, we demonstrate that even though miR-9 is re- neurons in the forebrain or expressed only in progenitors along quired for neuronal differentiation, regardless of the position the AP axis, we used fluorescent in situ hybridization (FISH) for along the AP axis, it regulates neural progenitors in a region- miR-9a-1 combined with immunostaining for Sox3 (marker for specific manner—it limits progenitor proliferation and promotes neural progenitors) at stages 30 and 36. We found that in the neuronal fate throughout the neural tube; in addition, in the fore- forebrain, miR-9 was transcribed in both Sox3-positive and brain it is important for progenitor survival. We have identified Sox3-negative cells, whereas it appeared to be restricted several genes that contain miR-9 binding sites in their 30UTR to the Sox3-positive domain in the hindbrain (Figure 1E; Fig- and respond to miR-9 in vitro and in vivo. However, functional ure S1E). This suggests that miR-9 expression differs along the analysis showed that hairy1 is the single key target that mediates AP axis within a single species and raises the question whether the effects of miR-9 in the forebrain and the hindbrain. hairy1 is it has the same function in different populations of neural a member of the Hes family of genes, and we show that, unlike progenitors. other Hes genes, it is primarily expressed in neurogenic rather than boundary areas of the CNS (Baek et al., 2006). Finally, we miR-9 Is Required for Neuronal Differentiation provide a molecular explanation for the regional-specific effects: In order to gain insight about miR-9’s role during neural develop- miR-9 regulates proliferation by feeding into the network con- ment, we decided to examine its loss-of-function phenotype. trolling cyclinD1/p27 expression in both areas, through Wnt We used an anti-miR-9 specific morpholino (miR-9 MO), which signaling in the hindbrain and Fgf8 signaling in the forebrain, interferes with both the processing of miR-9 precursors and but affects apoptosis via the mdm2/p53 pathway specifically in inhibits the activity of the mature miRNA (Kapsimali et al., the forebrain. These findings suggest that the positional embry- 2007; Martello et al., 2007) (see Figure S2A for schematic). Injec- onic origin of neural progenitors is an important parameter that tion of miR-9 MO led to an almost complete knockdown of dictates their response to the same microRNA and that in the mature miR-9 at early tadpole stage compared to wild-type case of miR-9 the specificity of response is generated down- (WT) embryos, whereas miR-9 levels were increased in embryos stream of a key target, hairy1. They show a regional diversity in injected with miR-9-2 precursor (Figure 2A), as shown using the properties of neural progenitors and highlight the importance semiquantitative RT-PCR. Knockdown was also confirmed of taking into account the positional origin of stem cells in using in situ hybridization and real-time PCR for the mature designing rational strategies to manipulate their proliferative form of miR-9 (Figures S2B and S2C). potential. Next, we injected miR-9 MO in one cell of a two-cell stage embryo and compared the injected to the control side at stage RESULTS 30 at the forebrain and hindbrain level (Figure 2B). Depletion of miR-9 negatively affected neuronal differentiation, as indicated miR-9 Expression Differs along the AP Axis by the decreased expression of N-tubulin (n = 14/25) and Neu- First, we examined miR-9 expression during the development of roD1 (n = 18/24) (Figure 2C, arrows). The number of Myt1-posi- X. tropicalis using in situ hybridization (miR-9 LNA probe). miR-9 tive cells (a transcription factor expressed in post-mitotic expression was evident in the prospective forebrain region in the neurons; Bellefroid et al., 1996) was also reduced in the miR-9 anterior neural plate at stage 18/19. At stage 23/24 mature miR-9 MO-injected side (Figure 2D), but not when control MO was was also detected in the developing eye and retina but later on its used (Figure S2D). Quantification of the results showed that expression in the neural tube expanded to the more posterior miR-9 depletion caused a reduction of the number of Myt1-posi- parts of the brain, including the mid- and hindbrain at stage tive cells to about 51% of the control in the forebrain (n = 7 30–36 (Figure 1A). There are four predicted miR-9 encoding embryos; p < 0.001), and 53% of the control in the hindbrain loci in the genome of X. tropicalis, which give rise to nearly iden- (n = 9 embryos; p < 0.001) (Figure 2E). These results indicate tical mature miR-9 after processing (see Figure S1A available on- that miR-9 is required for neuronal differentiation, regardless of line). Expression of the individual transcripts was similar to the the position along the AP axis. expression of mature miR-9 (Figure 1B; Figure S1B); however, miR-9a-1 was expressed at higher levels than the others. Tran- miR-9 Knockdown Promotes the Proliferation of Neural scripts were present in the forebrain, the eye, and in the mid- Progenitors in the Hindbrain and hindbrain, but no expression was detected in the MHB (Fig- We hypothesized that miR-9 depletion could interfere with the ure 1B; Figure S1B, marked with asterisk), in agreement with onset of the neurogenic program by preventing cell-cycle exit, reports in the zebrafish (Leucht et al., 2008), and no expression resulting in an increase in the number of progenitors. To test was evident in the spinal cord. We could not detect a signal for this we measured the area occupied by Sox3-positive neuronal

20 Developmental Cell 20, 19–32, January 18, 2011 ª2011 Elsevier Inc. Developmental Cell MicroRNA-9 Function in Xenopus Neurogenesis

A Stage 19 Stage 24 Stage 28 Stage 30 Stage 36 miR-9

B miR-9a-1 miR-9a-2 miR-9-3 C

* **

Stage 30 lateral view

D Forebrain Hindbrain dorsal view E miR-9a-1 miR-9a-1 Sox3 Sox3 Forebrain DAPI Midbrain miR-9a-1 miR-9a-1 Sox3 Sox3 DAPI Hindbrain Forebrain Hindbrain

Figure 1. miR-9 Expression Differs along the AP Axis (A) Whole-mount in situ hybridization for miR-9 expression in X. tropicalis using LNA probe. (B) Expression of miR-9 primary transcripts at stage 30—dorsal view. Dashed line indicates the plane of sectioning in (D). MHB is indicated with an asterisk. Scale bar, 200 mm. (C) Schematic representation of the different regions in the neural tube (red, forebrain; green, midbrain; blue, hindbrain) and transverse sections from the forebrain and hindbrain (red, progenitors; purple, intermediate zone; blue, neurons). (D) In situ hybridization for miR-9 precursors in transverse sections from stage 30 embryos. CNS tissue is outlined with a dashed line. Scale bar, 20 mm. (E) FISH for miR-9a-1 (in red) combined with immunohistochemistry (IHC) for Sox3 (marker for neural progenitors) in stage 30 embryo. CNS tissue is outlined with a dashed line. DNA is stained with DAPI (4,6-diamidino-2-phenylindole). Scale bar, 20 mm. progenitors per section in miR-9 MO-injected embryos. As ex- and pH3-positive cells was due to a change in cell proliferation, pected, in the hindbrain there was an increase of the progenitor we performed double labeling for pH3 and Sox3 and found domain by 28% compared to the control (n = 9; p < 0.001) that the labeling index (pH3+/Sox3+ cells in the hindbrain) is (Figures 3A and 3B). However, in the forebrain the Sox3-positive increased upon miR-9 knockdown (Figure 3E, p < 0.01). The area was not increased, and if anything it was slightly decreased increased rate of proliferation of the hindbrain progenitors was by 14% compared to the control (n = 7; p = 0.008). In the hind- also conﬁrmed using BrdU labeling of the proliferating progeni- brain some Sox3-positive cells were found further away from tors (Figures 3F and 3G, p < 0.001). the ventricle (data not shown), thus found in positions where These observations suggest that miR-9 function in the hind- differentiated cells would normally reside. brain is important for limiting progenitor proliferation and To ﬁnd out if there was a corresponding increase in the number promoting the onset of the neurogenic program and raises of cells undergoing mitosis, we examined the number of phos- interesting questions about how (and why) that differs in the pho-histone H3 (pH3)-positive cells in both areas. miR-9 knock- forebrain. down led to an almost 2-fold increase in the number of pH3-positive cells in the hindbrain, but there was no apparent change in miR-9 Depletion Causes Apoptosis in Forebrain the forebrain (Figures 3C and 4D, p < 0.001). Injection of control Progenitors MO had no effect on either Sox3 or pH3 expression (Figures S2E One possibility for the decrease of differentiated neurons in and S2F). To examine whether the increase in the Sox3-positive the forebrain is increased apoptosis. Indeed, TUNEL analysis

Developmental Cell 20, 19–32, January 18, 2011 ª2011 Elsevier Inc. 21 Developmental Cell MicroRNA-9 Function in Xenopus Neurogenesis

A C Forebrain Hindbrain

Control miR-9-2 miR-9 MO 100 miRNA-9

100

snRNA U2 N-tubulin co inj co inj B miR-9 MO NeuroD1 Stage 2 Stage 30 oc inj oc jni

D Forebrain Hindbrain E Control miR-9 Myt1 Myt1 120 *** ***

100

40 cells per section, % + 20

oc jni oc jni Myt1 0 Forebrain Hindbrain Myt1 Myt1 FITC FITC DAPI DAPI

oc jni oc jni

Figure 2. miR-9 Is Required for Neuronal Differentiation (A) Semiquantitative PCR analysis of mature miR-9 levels in stage 30 WT embryos, injected with miR-9-2 precursor or miR-9 MO at one cell stage. The snRNAU2 is used as a loading control. (B) Experimental outline. miR-9 MO was injected in one cell of the two-cell stage embryo, and the injected side was compared to the control at stage 30. (C) In situ hybridization (whole-mount and transverse sections from the forebrain and hindbrain) with markers for differentiated (N-tubulin) and differentiating neurons (NeuroD1). Note the reduced expression of both markers (arrows) in the miR-9 MO-injected side. (D) Immunohistochemistry on sections for the transcription factor Myt1 indicates impaired neuronal differentiation upon miR-9 knockdown. The FITC tag on miR-9 MO was used to identify the injected side; DAPI was used to stain the DNA. (E) The percentage of Myt1-positive cells in miR-9 MO-injected side relative to the control side in the forebrain (n = 6 embryos, p < 0.001) and hindbrain(n=9 embryos, p < 0.001). Error bars represent SEM. In all images, scale bars represent 20 mm and CNS tissue is outlined with a dashed line. showed that miR-9 MO caused an increase in apoptosis in the Alternatively, miR-9 depletion could reduce the survival of the forebrain, which was specific for that area, and it was not forebrain progenitors, which would be consistent with the loca- observed in the hindbrain (Figures 4A and 4B, p < 0.001). tion of the majority of the apoptosing cells (see above). In order Apoptotic cells were present throughout the forebrain but were to distinguish between these possibilities, we blocked cell death most frequent in the VZ (Figure 4A, arrows). No increase in by injecting a pan-caspase inhibitor together with miR-9 MO or apoptotic cells was apparent when control MO was used control MO. Cell death was efficiently prevented, as evident by (Figure S2G). the reduction of the number of apoptotic cells compared to in- An important question is whether the cells undergoing apo- jecting miR-9 MO alone (Figure S2H). Coinjection of caspase ptosis represent neuronal progenitors or differentiating neurons. inhibitor together with miR-9 MO led to an expansion of the Because miR-9 knockdown caused only a modest reduction of Sox3-positive area in the forebrain, compared to miR-9 MO the progenitor domain but a significant decrease in the number alone (Figures 4C and 4D), whereas the number of differentiating of neurons (see Figure 2), one may hypothesize that it is the fore- neurons was still reduced (Figures 4E and 4F). Effectively, pre- brain neurons that undergo apoptosis in the absence on miR-9. venting apoptosis made the miR-9 loss-of-function phenotype

22 Developmental Cell 20, 19–32, January 18, 2011 ª2011 Elsevier Inc. Developmental Cell MicroRNA-9 Function in Xenopus Neurogenesis

A B Forebrain Hindbrain Control Sox3 *** Sox3 150 miR-9 MO FITC FITC DAPI DAPI **

) domain, % ) domain, 100 +

co inj oc jni (Sox3 Progenitor 0 Forebrain Hindbrain C Forebrain Hindbrain D pH3 pH3 Control 250 *** FITC FITC DAPI DAPI miR-9 MO 200

150

100 cells per section, % + 50 pH3 oc jni oc jni 0 Forebrain Hindbrain E F G BrdU *** 15 60 ** FITC

Sox3 , % + , % + 10 40 / Sox3 + / Sox3 +

5 ,Sox3

+ 20 pH3 pH3 labeling index BrdU labeling index BrdU BrdU 0 co inj 0 Control miR-9 MO Control miR-9 MO

Figure 3. miR-9 Regulates Progenitor Proliferation in a Region-Specific Manner (A) Immunohistochemistry on sections for Sox3 shows expansion of the progenitor domain in the hindbrain. (B) Area occupied by Sox3-positive cells (progenitor domain) per section in miR-9 MO-injected side expressed relative to the control side in the forebrain (n = 7, p = 0.008) and hindbrain (n = 9, p < 0.001). (C and D) Transverse sections from the forebrain or hindbrain of miR-9 MO-injected embryos analyzed for the mitotic marker pH3 show a hindbrain-specific increase in the number of mitotic cells (n = 11, p < 0.001), but no change in the forebrain (n = 9). (E) pH3-labeling index (pH3+ cells over Sox3+ cells) in the hindbrain (n = 6, p = 0.004). (F and G) Rate of proliferation of the hindbrain progenitors is increased, as determined by BrdU incorporation for 30 min. BrdU-labeling index is calculated as the percentage of BrdU+ and Sox3+ cells over the total population of Sox3+ cells (n = 7, p < 0.001). In all panels, scale bars represent 20 mm, FITC staining shows the MO-injected side, DNA was counterstained with DAPI, CNS tissue is outlined with a dashed line, and error bars represent SEM. in the forebrain more similar to the one observed in the hindbrain. along the AP axis but has different functions in different axial Taken together, this suggests that miR-9 is necessary for the levels (Figure S3A). transition of progenitors to neurons across the AP axis, and in As a starting point we used bioinformatic analysis using the addition it is required for the survival of progenitors in the overlap of the targets predicted by the algorithms PicTar (Krek forebrain. et al., 2005) and TargetScan (Lewis et al., 2003) to identify more than 500 potential miR-9 targets based on target site hairy1 Is an Endogenous Target of miR-9 In Vivo conservation in mammals (data not shown). This data set was To understand how the differences in miR-9 loss-of-function further refined using GO analysis (Figure S3B) conservation of phenotype along the AP axis arise at molecular level, we set to the seed in Xenopus (data not shown), luciferase reporter assay determine the potential miR-9 targets in X. tropicalis in relation in HeLa cells (Figures S3D and S3E), and whole-mount in situ to the phenotype we observed. One possibility was that miR-9 hybridization expression screen (Figure S3F). We decided to might regulate two or more regionally restricted targets, which focus on the members of the hes (hairy and enhancer of split) in turn mediate functional specificity in different areas of the family, which have been shown to play crucial roles in maintain- CNS. Alternatively, miR-9 specificity of function might be gener- ing neural progenitors (Baek et al., 2006; Ohtsuka et al., 2001) ated downstream of one primary target, which is expressed Among them, Hes1 was present in all three GO categories, its

Developmental Cell 20, 19–32, January 18, 2011 ª2011 Elsevier Inc. 23 Developmental Cell MicroRNA-9 Function in Xenopus Neurogenesis

A B Forebrain Hindbrain 300 *** TUNEL TUNEL Control FITC FITC miR-9 MO DAPI DAPI 200

100 cells per section, % +

oc jni oc jni

TUNEL 0 Forebrain Hindbrain C miR-9 MO + Casp. Inh. E miR-9 MO + Casp. Inh. Sox3 Sox3 Myt1 Myt1

oc jni oc jni oc jni oc jni

Sox3 Sox3 Myt1 Myt1 FITC FITC FITC FITC DAPI DAPI DAPI DAPI

oc jni oc jni oc jni oc jni

D F 150 *** 125 ** 125 100

100 75 75 50 50 cells per section, % section, per cells cells per section, % section, per cells + + 25 25 Myt1 Sox3 0 0 miR-9 MO miR-9 MO miR-9 MO miR-9 MO + Caspase Inhibitor + Caspase Inhibitor

Figure 4. miR-9 Depletion Negatively Affects the Survival of Forebrain Neural Progenitors (A) TUNEL staining shows increased apoptosis upon miR-9 depletion in the forebrain (arrowheads), but not in the hindbrain. (B) Percentage of the TUNEL+ cells in the injected compared to the control side in the forebrain (n = 6, p < 0.001) and in the hindbrain (n = 6). Error bars represent SEM. (C and D) Sox3-positive domain is expanded in miR-9 MO-injected side when apoptosis is prevented (n = 7, p < 0.001). (E and F) The reduced number of differentiating neurons (Myt1+) upon miR-9 depletion is not rescued by caspase inhibitor block of apoptosis (n = 6, p = 0.003). In all images FITC staining shows the MO-injected side; DNA was counterstained with DAPI. Neural tube is outlined with a dashed line. Scale bars, 20 mm. Error bars represent SEM.

Xenopus homolog hairy1 showed a prominent effect in the Xenopus hairy1 (xHairy1) and mouse Hes1 (mHes1) using lucif- reporter assays, and was also expressed in the CNS, which is erase-based reporter assay. Both xhairy1 30UTR (xHairy1-WT) why we decided to examine it further. and mHes1 30UTR were signiﬁcantly repressed by synthetic The X. tropicalis hairy1 is most closely related to the mamma- miR-9 precursors, whereas this effect was absent when a mutant lian Hes1 based on sequence conservation (72%) (Jouve et al., reporter lacking the seed-complementary sequence (xHairy1_ 2000; data not shown). miR-9 binding site is highly conserved Mut) was used. In order to validate the speciﬁcity of the repres- in the vertebrate homologs of Hes1, with 100% sequence sion, we used a target-protector approach to block miR-9 homology in the seed-complementary region (Figure 5A). In binding site (Choi et al., 2007). A hairy1 target protector order to test whether miR-9 regulates hairy1 in vitro, we tested morpholino (hairy1 TP) was designed to overlap with the

24 Developmental Cell 20, 19–32, January 18, 2011 ª2011 Elsevier Inc. Developmental Cell MicroRNA-9 Function in Xenopus Neurogenesis

A C

D B

E G

F H I

Figure 5. miR-9 Regulates the Expression of hairy1 In Vivo (A) Sequence alignment of the predicted miR-9 binding site in HES1 homologs in human, mouse, Xenopus, and zebrafish. Positions that have a single, fully conserved residue are marked with an asterisk. Seed-complementary region is boxed in red. (B) HeLa cells were transfected with WT Xenopus hairy1 (xhairy1_WT), mouse Hes1 (mHes1), or mutant hairy1 (xHairy1_Mut) reporter together with either scrambled (Control) or miR-9 precursors (miR-9). Luciferase expression was normalized and expressed relative to the control levels. Error bars represent SD. (C) Design of target protector morpholino (Hairy1 TP) directed against hairy1 miR-9 binding site. Seed region is boxed in red. (D) Hairy1 TP alleviates the repression of hairy1 luciferase reporter when cotransfected with miR-9 precursors but has no effect on the repression of other miR-9 targets. Error bars represent SD. (E) In situ hybridization for miR-9 (miR-9a-1 transcript) and hairy1 in stage 30 embryos. Shown are whole mounts and transverse sections through the respective brain areas. (F) Double-fluorescent in situ for hairy1 (red) and miR-9a-1 (green) shows mutually exclusive pattern of expression along the AP axis. (G) miR-9 MO and hairy1 TP lead to expansion of the hairy1-positive domain (red arrowheads) along the AP axis, as shown by in situ hybridization. (H) Quantification of the change in hairy1 mRNA expression using qRT-PCR. (I) Hes1 mRNA levels in N1E neuroblastoma cells are downregulated when miR-9 is overexpressed and increased when it is knocked down using LNA inhibitors. In all graphs, data are presented as mean values, and error bars represent SEM. seed-complementary sequence on hairy1 and extend further in reporter assays confirmed that hairy1 TP is able to partially alle- the 30 direction to confer specificity (Figure 5C). Next, we exam- viate the repression of miR-9 on the hairy1 luciferase reporter ined the efficiency and specificity of hairy1 TP. Luciferase when introduced in vitro together with miR-9 mimics, but not

Developmental Cell 20, 19–32, January 18, 2011 ª2011 Elsevier Inc. 25 Developmental Cell MicroRNA-9 Function in Xenopus Neurogenesis

of a reporter carrying the 30UTR of other miR-9 targets such as These findings suggest that that the mechanism of miR-9 regu- hairy2, TLX, and Onecut1 (Plaisance et al., 2006)(Figure 5D). lation is evolutionarily conserved and that miR-9 acts by desta- These results show that miR-9 is able to repress hairy1 in vitro. bilizing the mRNA rather than repressing protein translation. This is consistent with recent reports that contrary to what was hairy1 and miR-9 Expression Is Mutually Exclusive previously thought, decreasing mRNA levels is the main mode To gain insight into the miR-9-hairy1 interaction, we compared of repression for mammalian microRNAs (Guo et al., 2010). their expression in vivo. Hairy1 has been cloned from Xenopus before (Palmeirim et al., 1997), but here we described its expres- Hairy1 TP Functionally Mimics miR-9 MO Phenotype sion in the nervous system in detail. During early brain develop- In order to determine the contribution of hairy1 repression to ment (stages 21–26), hairy1 is expressed in a broad region in the miR-9 function, we examined the effect of hairy1 TP on neuronal forebrain (data not shown) but later becomes restricted to the differentiation, progenitor proliferation, and apoptosis. Injection roof plate and an intermediate patch of progenitors, which repre- of hairy1 TP resulted in decrease in the expression of N-tubulin sents the zona limitans intrathalamica (ZLI)—a boundary region (Figure 6A, n = 11/17), whereas TP designed against another between the thalamus and the prethalamus (Figure 5Eb). In this potential miR-9 target—NR2E1/TLX had no effect on N-tubulin region hairy1 expression overlaps with the known marker of expression (data not shown). Furthermore, the number of the ZLI Shh (Ishibashi and McMahon, 2002) and is immediately Myt1-positive cells was also negatively affected in both the fore- adjacent to the expression of Irx3, which marks the thalamic brain and the hindbrain (Figures 6B and 6C). As with the miR-9 region in chick and mouse (Kiecker and Lumsden, 2004)(Fig- MO, neuronal reduction was accompanied by an increase in ure S4A). Conversely, in the more posterior areas, hairy1 tran- apoptotic cells in the forebrain (Figures 6D and 6E) and an scripts are present ventrally in the midbrain but are absent increase in proliferating cells in the hindbrain (Figures 6F and from the mid- and hindbrain boundary, contrary to the expres- 6G). Furthermore, electroporation of hairy1 construct lacking sion of Hes1 in the mouse and the hairy-related genes her5/9 the 30UTR together with LacZ DNA as a tracer led to a reduction in zebrafish. In the hindbrain hairy1 expression is restricted to in N-tubulin expression (Figure 6H, n = 16/18 embryos), confirm- distinct domains—in a ventral region adjacent to the floor plate ing the ability of hairy1 to repress the neurogenic program in both and in an intermediate region of progenitors (Figure 5Ec). areas. Electroporation of LacZ alone had no effect on N-tubulin Mammalian Hes1 is also expressed at high levels in the ZLI expression (data not shown). and in an intermediate zone of progenitors in the hindbrain These results show that alleviation of miR-9 repression on (Baek et al., 2006), but in addition it is also expressed throughout hairy1 mimics miR-9 MO phenotype and suggest that posttran- the VZ in the telencephalon and in the boundary regions such as scriptional regulation of hairy1 is one of the essential aspects of MHB, the roof plate, and the floor plate. The zebrafish her5 is miR-9 function during neural development. They also point out also expressed in boundary regions such as the MHB (Geling that the specificity of miR-9 function is generated downstream et al., 2003). Thus, Xenopus tropicalis hairy1 shows similarities of hairy1. and differences with hes1; both are expressed in the ZLI but unlike hes1, hairy1 is not expressed in the roof plate or the floor Changes in Cyclin D1 and p27Xic1 Expression plate or the MHB, with the exception of the roof plate in the fore- Contribute to Increased Progenitor Proliferation brain. Instead, Hairy1 is expressed in a subset of dorsoventrally To understand the mechanism by which miR-9 affects prolifera- restricted progenitors within the neurogenic compartments. This tion and apoptosis, we looked at molecular pathways that may expression pattern appears complementary to that of mir-9 in be regulated by miR-9 through hairy1. Injection of either miR-9 whole mounts and sections (Figure 5E). Double FISH for MO or hairy1 TP led to an expansion of the expression domain miR-9a-1 and hairy1 confirmed that their expression is mutually of cyclin D1 (miR-9 MO: n = 18/28 embryos; n = 15/24 hairy1 exclusive along the AP axis with the exception of a few double- TP), which promotes G1-S phase progression and to the down- stained cells in the ventral hindbrain (Figure 5F). regulation of p27Xic1 expression, a cyclin-dependent kinase inhibitor (miR-9 MO, 12/19; hairy1 TP, 10/19). This was observed miR-9 Regulates hairy1 In Vivo both in the hindbrain and the forebrain (Figures 7A and 7B, In order to determine whether miR-9 regulates hairy1 in vivo, we arrowheads), consistent with an effect on proliferation on both examined hairy1 expression in morphant embryos using areas. In mammals, p27 has been shown to be a direct target in situ hybridization. Both miR-9 MO (n = 22/36) and hairy1 TP of Hes1 (Murata et al., 2005); therefore, the interaction of hairy1 (n = 20/35) led to an expansion of the hairy1-positive domain with p27Xic1 is likely to be direct. By contrast the upregulation of along the AP axis: in the forebrain the expression in the roof plate cyclin D1 by hairy1 is likely to be indirect (diagram in Figure 7E). and in the ZLI region was expanded, whereas posteriorly the Cyclin D1 is a direct downstream target of Wnt1 (Megason and hairy1-positive domain expanded both laterally and dorsally McMahon, 2002), which was also increased in the injected (Figure 5G). side (miR-9 MO, eight of 14; hairy1 TP, ten of 18) (Figure 7B). The expansion of the expression domain of Hairy1 suggests Hairy1 may affect Wnt1 expression through Zic1, a transcription that miR-9 acts at the mRNA level. Indeed, miR-9 MO and hairy1 factor known to promote the proliferation of neural progenitors TP led to an increase in hairy1 mRNA levels, as shown by real- (Aruga et al., 2002; Elsen et al., 2008). Zic1 positively regulates time PCR (Figure 5H). In addition, miR-9 overexpression in wnt signaling both in Xenopus and zebrafish (Elsen et al., 2008; a neuroblastoma cell line (N1E-115) decreased the RNA level Merzdorf and Sive, 2006), and in addition another member of of the murine homolog, Hes1, and conversely, inhibition of the hes family in Xenopus, hairy2, has been previously shown endogenous miR-9 with miR-9 LNA increased it (Figure 5I). to regulate Zic1 (Nichane et al., 2008). Therefore, it is very likely

26 Developmental Cell 20, 19–32, January 18, 2011 ª2011 Elsevier Inc. Developmental Cell MicroRNA-9 Function in Xenopus Neurogenesis

A Forebrain Hindbrain E 600 Control *** 500 miR-9 MO

400 co inj co inj 300

200 cells per section, % B Forebrain Hindbrain + 100 Myt1 Myt1

FITC FITC TUNEL 0 DAPI DAPI Forebrain Hindbrain F Forebrain Hindbrain pH3 pH3 FITC FITC DAPI DAPI

oc jni oc jni

C Control miR-9 120 *** *** oc jni oc jni

100 G 80 Control 250 *** 60 miR-9 MO 200 40 cells per section, % + 20 150

Myt1 0 Forebrain Hindbrain 100 cells per section, % D Forebrain Hindbrain + 50 pH3 TUNEL TUNEL 0 FITC FITC Forebrain Hindbrain DAPI DAPI H Forebrain Hindbrain

oc jni oc jni op

Figure 6. Hairy1 Target Protector Mimics miR-9 MO Phenotype (A) In situ hybridization (whole-mount and transverse sections from the forebrain and hindbrain) for N-tubulin in hairy1 TP-injected embryos. (B) The number of differentiating neurons (Myt1+ cells) is decreased upon injection of hairy1 TP. (C) Quantification of the Myt1+ cells in the forebrain (n = 7, p < 0.001) and the hindbrain (n = 7, p < 0.001). Myt1+ cells in the injected side were expressed as a percentage of the control side. (D) Hairy1 TP leads to forebrain-specific induction of apoptosis as indicated by TUNEL staining. (E) Quantification of the TUNEL-positive nuclei in the forebrain (n = 7, p < 0.001) and in the hindbrain (n = 5). (F) Immunostaining for pH3 in embryos injected in one side with hairy1 TP. (G) Relative number of pH3+ cells in the hairy1 TP-injected compared to the control side in the forebrain (n = 11) and the hindbrain (n = 9, p < 0.001). (H) In situ hybridization for N-tubulin (purple) in embryos electroporated in one side with hairy1 D30UTR and lacZ DNA as a tracer. Light-blue staining indicates the electroporated area. op, olfactory placodes. Scale bars, 20 mm. In all panels, FITC was used to identify the MO-injected side; DNA was counterstained with DAPI; CNS tissue is outlined with a dashed line; and error bars represent SEM.

that Zic1 lies between hairy1 and wnt1. Indeed, injection of either it can be involved in the regulation of other signaling pathways miR-9 MO (n = 12/20) or hairy1 TP (n = 10/23) led to a lateral (such as BMP signaling) in addition to wnt. expansion of the Zic1 domain in the hindbrain (Figure 7B), sug- In the forebrain, wnt1 is not expressed, and Zic1 expression is gesting that Zic1 may mediate the hairy1 regulation on wnt1 not affected by miR-9 knockdown (data not shown); therefore, pathway. However, we cannot rule out the possibility that the effect of hairy1 on proliferation may be mediated via an inter- miR-9 might affect wnt signaling independently of Zic1 or that mediate regulator other than wnt. Fgf signaling is known to

Developmental Cell 20, 19–32, January 18, 2011 ª2011 Elsevier Inc. 27 Developmental Cell MicroRNA-9 Function in Xenopus Neurogenesis

A miR-9 MO Hairy1 TP B miR-9 MO Hairy1 TP Figure 7. Mechanism of miR-9 Function (A) Forebrain sections of miR-9 MO or hairy1 TP-injected embryos analyzed for CyclinD1, p27Xic1, and Fgf8 expression by in situ hybridization. (B) Hindbrain sections of miR-9 MO or hairy1 CyclinD1

CyclinD1 co inj co inj TP-injected embryos analyzed for CyclinD1,

oc jni oc jni p27Xic1, Wnt1, and Zic1 expression. (C) Representative western blot for endogenous p53 protein levels in forebrain or hindbrain tissue isolated from X. tropicalis embryos injected with control MO, miR-9 MO, or hairy1 TP. Numbers p27Xic1 co inj co inj represent the mean from three experiments, p27Xic1 60 embryos each. oc jni oc jni (D) Real-time PCR analysis for Mdm2 expression, normalized for the ribosomal protein RPL8 (n = 3 experiments, 20 embryos each). Error bars repre- Wnt1 sent SEM. oc jni co inj (E) Model for miR-9 function in cell survival and Fgf8 progenitor proliferation.

oc jni oc jni

C Zic1 rol MO oc jni co inj ont Control MOmiR-9 MO Hairy1 TP C miR-9 MO Hairy1 TP well with the region-specific induction of 1.00 2.06 2.00 1.00 1.02 0.94 E p53 apoptosis we observed in the embryo miR-9 Forebrain Hindbrain using TUNEL (Figures 4A and 4B) and suggests that activation of p53 pathway α-Tubulin Hairy1 may be responsible for this phenotype. Forebrain Hindbrain +X Zic1 We examined whether the upregulation of p53 is mediated through its regulator D Control MO Mdm2 Fgf8 miR-9 MO p27Xic1 Wnt1 Mdm2 (Haupt et al., 1997) using quantita- 1.25 Hairy1 TP tive RT-PCR. We found that miR-9 MO or p53 1.00 CyclinD1 hairy1 TP led to an approximately 30% ** ** 0.75 decrease in Mdm2 mRNA expression in Apoptosis Proliferation the forebrain but did not significantly 0.50 affect Mdm2 levels in the hindbrain (Fig- 0.25 ure 7D, n = 3 experiments, 20 embryos Relative Mdm2 RNA levels RNA Mdm2 Relative 0.00 each). Hes1 has been previously shown Forebrain Hindbrain to activate the p53 pathway through Mdm2; however, this probably requires specific cofactors because Hes1 cannot promote proliferation in the developing forebrain (Storm et al., bind Mdm2 promoter per se (Huang et al., 2004). Nevertheless, 2006), and Fgf8 is expressed in the ZLI area (Kataoka and Shi- the lack of mdm2 repression by hairy1 in the hindbrain provides mogori, 2008). Using double FISH, we found that hairy1 overlaps a molecular explanation for the lack of an effect on apoptosis in with Fgf8 (Figure S4C) and furthermore, injection of miR-9 MO this region when miR-9 is depleted. (n = 8/14) and hairy1 TP (n = 10/18) led to an expansion of the Fgf8-positive domain (Figure 7A). Therefore, even though hairy1 is expressed in a restricted domain, it regulates neurogenesis DISCUSSION throughout the neural tube via non-cell-autonomous signaling pathways. In this study we have examined the role of miR-9 during Xenopus neurogenesis, focusing on the regional differences in its expres- p53 Contributes to miR-9 MO-Induced Apoptosis sion and function. Conflicting results about miR-9 expression in the Forebrain and function in different model systems have been obtained, To understand the molecular pathway behind the differential even though its sequence is 100% conserved (Delaloy et al., effects on miR-9 on apoptosis in the forebrain versus the hind- 2010; Leucht et al., 2008; Zhao et al., 2009). In regard to expres- brain, first, we examined p53 expression in these two areas sion, previous studies were either based mainly on the location of because p53 has been shown to mediate Notch-induced the expressing cells in relation to the VZ or did not examine apoptosis in the forebrain (Yang et al., 2004). Injection of either different AP levels (Delaloy et al., 2010; Deo et al., 2006; Leucht miR-9 MO or hairy1 TP led to approximately 2-fold increase in et al., 2008; Shibata et al., 2008). Here, by comparison to other p53 protein levels in the forebrain, but not in the hindbrain (Fig- markers, we have shown that miR-9 expression differs along ure 7C, n = 3 experiments, 60 embryos each). This correlated the AP axis, even in a single species—it is expressed in both

28 Developmental Cell 20, 19–32, January 18, 2011 ª2011 Elsevier Inc. Developmental Cell MicroRNA-9 Function in Xenopus Neurogenesis

neurons and progenitors in the forebrain but becomes restricted our results suggest that miR-9 falls into the growing category to progenitors in the more posterior brain regions, namely the of miRNAs that have just one or few important targets, although midbrain and hindbrain. many more can be bioinformatically predicted (reviewed in In addition to regional differences in expression, our work has Flynt and Lai, 2008). Such miRNAs tend to be involved in ‘‘devel- uncovered a regional difference in the function of miR-9 in opmental genetic switching’’ rather than ‘‘fine tuning,’’ a hypoth- progenitor cells. Using a loss-of-function approach, we have esis that is consistent with the proposed role of miR-9 in found that in the absence of miR-9, neurogenesis fails along the neurogenesis. AP axis. At the same time, in miR-9 MO embryos the number of What is the significance of hairy1 as a miR-9 target? Hairy1 is progenitors increases in the hindbrain, but paradoxically, it a member of the hes (hairy and enhancer of split), helix-loop- slightly decreases in the forebrain. However, an underlying helix (bHLH) type transcriptional repressors. Several Hes genes, increase in forebrain progenitors is uncovered when apoptosis such as Hes1, Hes3 in the mouse and the her5 in zebrafish, are is blocked. We propose that miR-9 is necessary for cell-cycle expressed at high levels in boundary of the nervous system exit throughout its AP domain of expression, but neuronal (Baek et al., 2006). Such boundary regions, exemplified by the progenitors in the forebrain additionally and uniquely require ZLI, the MHB, the floor, and roof plate, are characterized by miR-9 for their survival. Therefore, in the forebrain, in the absence secretion of morphogens, slow proliferation of progenitors, and of miR-9, extra proliferation is counterbalanced by increased lack of neurogenesis. Hes1 is also expressed at variable levels apoptosis, resulting in no net increase in the number of forebrain in adjacent neural compartments where neurogenesis actively progenitors, and even a slight decrease. Such context-depen- takes place. These expression data and the results of functional dent activity of miR-9 based on the regional identity of progenitor analysis gave rise to a model whereby the high persistent cells may explain previously conflicting results with respect to levels of Hes1 observed in boundaries suppress neurogenesis, miR-9 function (Delaloy et al., 2010; Zhao et al., 2009). whereas in compartments the variable levels permit neurogene- To our knowledge, an effect on apoptosis in neural develop- sis when protein levels are low (Baek et al., 2006). The variable ment by miR-9 knockdown has not been reported before. Hes1 levels are in fact oscillatory (Shimojo et al., 2008), and However, Dicer ablation in the mouse forebrain led to increased such oscillations are thought to be driven both by mRNA and cell death in committed neuronal progenitors (De Pietri Tonelli Hes1 protein instability, although factors that mediate the et al., 2008), and miR-24a is required to prevent apoptosis in mRNA instability are not known (Davis et al., 2001; Hirata et al., the retina (Walker and Harland, 2009). Our studies have not re- 2002; Shimojo et al., 2008). It is tempting to speculate that vealed a function of miR-9 in forebrain neurons because their miR-9 is involved in hairy1/Hes1 oscillations by regulating formation is mostly prevented by the loss of function. Alternative mRNA stability because, indeed, the role of mRNA stability in strategies will be needed to address this question. the Hes1 oscillator has been previously theoretically predicted Having shown distinct effects on proliferation and apoptosis, (Hirata et al., 2002; Xie et al., 2007). This would be consistent an important question is whether these are mediated by one with the expression of both miR-9 and Hes1 in proliferating primary miR-9 target or the coordinate regulation of several progenitors in the VZ of the mammalian telencephalon (Baek targets. Theoretically, microRNAs are capable of regulating et al., 2006; Delaloy et al., 2010) and our observations that many target genes, and miR-9 is no exception to this. Indeed, there is conserved miR-9 binding site in the 30UTR of Hes1, our bioinformatic analysis followed by luciferase assay verifica- and that both hairy1 and hes1 are regulated by miR-9 at mRNA tion identified several genes as potential miR-9 targets. Similarly, level. Because Hairy1 is a primary target of miR-9, regional spec- previous reports in other species have identified several miR-9 ificity is generated downstream of hairy1, culminating in the targets. In the fish, several components of the FGF pathway differential effect on apoptosis in the forebrain versus the hind- and Her5 have been proposed as targets involved in the forma- brain. In turn, this specificity may be mediated by the presence, tion of the MHB and Her9 in the control of neurogenesis (Leucht availability, or activity of cofactors, some of which may be tissue et al., 2008). In the mouse, proposed targets include FoxG1 in the specific. Indeed, several cofactors for the Hes family of genes developing mouse telencephalon (Shibata et al., 2008), NR2E1/ have been identified, such as Id and Groucho (Bai et al., 2007; TLX in adult neural stem cells (Zhao et al., 2009), and stathmin McLarren et al., 2001). in human embryonic stem cell-derived neural progenitors (Dela- To summarize, we propose that in normal development, miR-9 loy et al., 2010). However, with the exception of Her5, target- promotes neurogenesis by lowering the levels of hairy1 such that protector experiments (Leucht et al., 2008), where the endoge- cells can exit the proliferative compartment. In the absence of nous putative target is specifically protected from miR-9 binding, miR-9, hairy1 levels remain high, and progenitor cells cannot have not been performed; therefore, it is very difficult to evaluate complete the differentiation program. A regional specificity of the contribution of these targets to the miR-9 loss-of-function action is evident in that forebrain progenitors that fail to exit phenotype. the cell cycle undergo apoptosis. Therefore, in the forebrain In our work, Hairy1 target-protector experiments recapitulated the proliferative effect of miR-9 depletion can only be seen the miR-9 MO phenotype in vivo, including the regional-specific when apoptosis is also blocked. These findings complement effects in apoptosis. These results suggest that a single target, the miR-9/ her5 regulation in the zebrafish MHB (Leucht et al., hairy1, mediates the effects of miR-9 on neurogenesis, prolifera- 2008) and show that miR-9 regulation of hairy genes is more tion, and apoptosis. In this scenario the regional specificity of widespread, occurring well outside boundary regions. function is regulated downstream of Hairy1, rather than directly Our results have far-reaching implications for any cancer ther- downstream of miR-9. Although we cannot exclude the possi- apies and stem cell expansion that rely on manipulating miR-9 bility that other targets mediate other aspects of miR-9 activity, levels. In terms of stem cell expansion, their positional identity

Developmental Cell 20, 19–32, January 18, 2011 ª2011 Elsevier Inc. 29 Developmental Cell MicroRNA-9 Function in Xenopus Neurogenesis

may determine whether they will undergo proliferation or apo- were counted across two consecutive sections in at least six embryos and ptosis in response to blocking miR-9. On the other hand, an averaged. Values were expressed relative to the number of apoptotic cells in inhibitor of miR-9 may have therapeutic potential in forebrain- the control side. Apoptosis was blocked using a pan-caspase inhibitor (Z-VAD (OMe)-FMK; derived tumors, inducing apoptosis of progenitors, but may Calbiochem), which was injected at two-cell stage at a final concentration of have an undesirable effect in tumors of hindbrain origin, enhanc- 2 ng/ml. ing their proliferation. The regional-specific effect of miR-9 on neural progenitors underscores the importance of taking into RNA Isolation, RT-PCR, and Quantitative Real-Time PCR Analysis account the positional identity of cells when testing miR-9 func- Total RNA was extracted from either whole embryos or forebrain/hindbrain tion in normal development and disease. tissue using TRIzol (Invitrogen) and retrotranscribed using RT-AMV (Invitrogen) according to the manufacturer’s instructions. Mature miR-9a levels were assessed using modified semiquantitative RT-PCR as previously described EXPERIMENTAL PROCEDURES (Martello et al., 2007). Quantitative real-time PCR was performed in an ABI Ste- pOnePlus Sequence Detection System (Applied Biosystems) using TaqMan DNA Constructs and Electroporation Fast Real-Time PCR Master Mix and probes purchased from Applied Biosys- For the generation of luciferase reporter constructs, 30UTR of predicted miR-9 tems. The expression of X. tropicalis genes was normalized for Rpl8, whereas targets (or 1 kb region containing the seed-complementary sequence if the Hes1 expression was normalized to Gapdh in mouse. miR-9 expression was 30UTR was not annotated) was PCR amplified from X. tropicalis genomic examined using TaqMan microRNA assay (ABI). DNA and cloned downstream of Renilla luciferase coding sequence in the psiCHECK-2 vector (Promega). miR-9-2 WT and miR-9-2 Mut were amplified from genomic DNA as described previously (Shibata et al., 2008) and cloned in Cell Culture and Luciferase Reporter Assay the pCS2+ vector. pCS2-Hairy1 construct lacking the 30UTR was electropo- HeLa cells were maintained in DMEM supplemented with 10% serum and anti- rated together with LacZ DNA as a tracer into the brain of stage 26 embryos biotics. N1E-115 neuroblastoma cell line was obtained from ECACC and main- using SD9 stimulator (Grass Technologies) as previously described (Falk tained in DMEM supplemented with 10% serum and GlutaMAX (Invitrogen). et al., 2007). For Hes1 expression analysis 24 hr after transfection with miR-9 precursors (30 nM) or miR-9 LNA inhibitor (50 nM), cells were synchronized by serum star- vation as previously described (Hirata et al., 2002). For luciferase reporter Morpholino Design and Injection assays, cells were seeded at a density of 104 cells/well in a 96-well plate The anti-miR-9 morpholino (50-CTCATACAGCTAGATAACCAAAGAT-30), the and transfected after 24 hr with 25 ng of the reporter and either 30 nM of hairy1 target protector morpholino (50-AAGAGCATTCCATGTCTTTGGCA scrambled or miR-9 precursors (Ambion). Luciferase expression was analyzed TC-30), and the standard Negative Control Morpholino (50-CCTCTTACCT after 48 hr using Dual Luciferase Assay system (Promega). Renilla luciferase CAGTTACAATTTATA-30) were purchased from Gene Tools LLC and used at activity was normalized by the coexpressed Firefly Luciferase and expressed the following amounts: control MO (one side, 10 ng; whole embryo, 20 ng); as a percentage of the control. All assays were repeated at least three times miR-9 MO (one side, 7.5 ng; whole embryo, 15 ng); and hairy1 TP (one side, and performed in triplicate each time. 10 ng; whole embryo, 20 ng). All morpholinos were conjugated to FITC, and the injected side was identified using primary mouse anti-FITC (1:250; Roche) and anti-mouse Alexa 488 (1:500; Molecular Probes) antibodies. Statistical Analysis For Myt1, pH3, or TUNEL analysis, positive cells were counted across two consecutive sections in the corresponding brain area and the numbers aver- In Situ Hybridization aged per embryo. Sox3 expression was quantified by drawing a border around Whole-mount in situ hybridizations were performed as previously described the area containing Sox3-positive cells and measuring the area using ImageJ. (Bourguignon et al., 1998). Mature miR-9 was detected using miR-9 DIG- Values were expressed relative to the control side. N numbers represent labeled LNA probe (TCATACAGCTAGATAACCAAAGA; Exiqon) and the number of embryos from at least three experiments unless otherwise indi- following modifications to the standard in situ protocol: additional fixation cated. Statistical analysis of the data (two-tailed unpaired Student’s t test, using 1-ethyl-3-(3-dimethyl-aminopropyl) carbodiimide for 1 hr (adapted calculation of SEM) was done using SigmaStat 3.0 (Aspire Software). Statis- from Pena et al., 2009) and hybridization temperature 52 C. Fluorescent in tical significance is indicated as follows: *p < 0.1, ** p < 0.01, *** p < 0.001. situ hybridization (FISH) was performed as previously described (Vize et al., 2009) with the following modification - signal was detected using tyramide signal amplification (Perkin Elmer). Detailed protocols are available upon SUPPLEMENTAL INFORMATION request. Neural tube boundary was drawn based on high-magnification DAPI or bright-field images. Supplemental Information includes four figures and can be found with this article online at doi:10.1016/j.devcel.2010.11.018. Cryosectioning, Antibody Staining, and Immunoblotting For immunohistochemistry, embryos fixed in MEMFA (0.1 M MOPS [pH 7.4], ACKNOWLEDGMENTS

2 mM EGTA, 1 mM MgSO4, 3.7% formaldehyde) were sectioned on a Leica CM3050 S cryostat after embedding in 25% fish gelatin/15% sucrose and We thank Drs. M. Ronshaugen and F. Dajas-Bailador for discussions. We stained as described previously (Chalmers et al., 2003; Regad et al., 2007). would like to thank the Wellcome Trust for funding. N.P. is a WT senior The following primary antibodies were used: anti-Sox3 (1:2000; gift from research fellow, B.B. is a WT 4 year PhD student, and A.P. is a DTC in Systems Klymkovsky laboratory); anti-Myt1 (1:1000; Sabherwal et al., 2009); anti-pH3 Biology PhD student. (1:500; Upstate); and anti-p53 (1:1000; Abcam). Appropriate secondary antibodies were obtained from Molecular Probes. Received: June 4, 2010 For western blot, primary mouse anti-p53 (1:100; Abcam), mouse anti-a- Revised: October 11, 2010 tubulin (1:5000; Sigma), and secondary anti-mouse HRP (1:2000; DakoCyto- Accepted: November 19, 2010 mation) were used. Experiment was repeated three times (with 60 embryos Published: January 18, 2011 each), and results were quantified using Intelligent Quantifier software (Bio Image Systems). REFERENCES

TUNEL Staining and Apoptosis Inhibitor Aruga, J., Tohmonda, T., Homma, S., and Mikoshiba, K. (2002). Zic1 promotes TUNEL staining was performed using TMR red In Situ Cell Death Detection kit the expansion of dorsal neural progenitors in spinal cord by inhibiting neuronal according to the manufacturer’s instructions (Roche). TUNEL-positive cells differentiation. Dev. Biol. 244, 329–341.

30 Developmental Cell 20, 19–32, January 18, 2011 ª2011 Elsevier Inc. Developmental Cell MicroRNA-9 Function in Xenopus Neurogenesis

Baek, J.H., Hatakeyama, J., Sakamoto, S., Ohtsuka, T., and Kageyama, R. Haupt, Y., Maya, R., Kazaz, A., and Oren, M. (1997). Mdm2 promotes the rapid (2006). Persistent and high levels of Hes1 expression regulate boundary degradation of p53. Nature 387, 296–299. formation in the developing central nervous system. Development 133, Hirata, H., Yoshiura, S., Ohtsuka, T., Bessho, Y., Harada, T., Yoshikawa, K., 2467–2476. and Kageyama, R. (2002). Oscillatory expression of the bHLH factor Hes1 Bai, G., Sheng, N., Xie, Z., Bian, W., Yokota, Y., Benezra, R., Kageyama, R., regulated by a negative feedback loop. Science 298, 840–843. Guillemot, F., and Jing, N. (2007). Id sustains Hes1 expression to inhibit preco- Huang, Q., Raya, A., DeJesus, P., Chao, S.H., Quon, K.C., Caldwell, J.S., cious neurogenesis by releasing negative autoregulation of Hes1. Dev. Cell 13, Chanda, S.K., Izpisua-Belmonte, J.C., and Schultz, P.G. (2004). 283–297. Identification of p53 regulators by genome-wide functional analysis. Proc. Bellefroid, E.J., Bourguignon, C., Hollemann, T., Ma, Q., Anderson, D.J., Natl. Acad. Sci. USA 101, 3456–3461. Kintner, C., and Pieler, T. (1996). X-MyT1, a Xenopus C2HC-type zinc Ishibashi, M., and McMahon, A.P. (2002). A sonic hedgehog-dependent finger protein with a regulatory function in neuronal differentiation. Cell 87, signaling relay regulates growth of diencephalic and mesencephalic primordia 1191–1202. in the early mouse embryo. Development 129, 4807–4819. Bourguignon, C., Li, J., and Papalopulu, N. (1998). XBF-1, a winged helix tran- Jessell, T.M. (2000). Neuronal specification in the spinal cord: inductive signals scription factor with dual activity, has a role in positioning neurogenesis in and transcriptional codes. Nat. Rev. Genet. 1, 20–29. Xenopus competent ectoderm. Development 125, 4889–4900. Jouve, C., Palmeirim, I., Henrique, D., Beckers, J., Gossler, A., Ish-Horowicz, Chalmers, A.D., Strauss, B., and Papalopulu, N. (2003). Oriented cell divisions D., and Pourquie´ , O. (2000). Notch signalling is required for cyclic expression asymmetrically segregate aPKC and generate cell fate diversity in the early of the hairy-like gene HES1 in the presomitic mesoderm. Development 127, Xenopus embryo. Development 130, 2657–2668. 1421–1429. Choi, W.Y., Giraldez, A.J., and Schier, A.F. (2007). Target protectors reveal Kapsimali, M., Kloosterman, W.P., de Bruijn, E., Rosa, F., Plasterk, R.H.A., and dampening and balancing of Nodal agonist and antagonist by miR-430. Wilson, S.W. (2007). MicroRNAs show a wide diversity of expression profiles in Science 318, 271–274. the developing and mature central nervous system. Genome Biol. 8, R173. Davis, R.L., Turner, D.L., Evans, L.M., and Kirschner, M.W. (2001). Molecular Kataoka, A., and Shimogori, T. (2008). Fgf8 controls regional identity in the targets of vertebrate segmentation: two mechanisms control segmental developing thalamus. Development 135, 2873–2881. expression of Xenopus hairy2 during somite formation. Dev. Cell 1, 553–565. Kiecker, C., and Lumsden, A. (2004). Hedgehog signaling from the ZLI regu- De Pietri Tonelli, D., Pulvers, J.N., Haffner, C., Murchison, E.P., Hannon, G.J., lates diencephalic regional identity. Nat. Neurosci. 7, 1242–1249. and Huttner, W.B. (2008). miRNAs are essential for survival and differentiation of newborn neurons but not for expansion of neural progenitors during early Kosik, K.S. (2006). The neuronal microRNA system. Nat. Rev. Neurosci. 7, neurogenesis in the mouse embryonic neocortex. Development 135, 3911– 911–920. 3921. Krek, A., Grun, D., Poy, M.N., Wolf, R., Rosenberg, L., Epstein, E.J., Delaloy, C., Liu, L., Lee, J.A., Su, H., Shen, F., Yang, G.Y., Young, W.L., Ivey, MacMenamin, P., da Piedade, I., Gunsalus, K.C., Stoffel, M., et al. (2005). K.N., and Gao, F.B. (2010). MicroRNA-9 coordinates proliferation and migra- Combinatorial microRNA target predictions. Nat. Genet. 37, 495–500. tion of human embryonic stem cell-derived neural progenitors. Cell Stem Krichevsky, A.M., Sonntag, K.-C., Isacson, O., and Kosik, K.S. (2006). Specific Cell 6, 323–335. microRNAs modulate embryonic stem cell-derived neurogenesis. Stem Cells Deo, M., Yu, J.Y., Chung, K.H., Tippens, M., and Turner, D.L. (2006). Detection 24, 857–864. of mammalian microRNA expression by in situ hybridization with RNA oligonu- Lee da, Y., Yeh, T.H., Emnett, R.J., White, C.R., and Gutmann, D.H. (2010). cleotides. Dev. Dyn. 235, 2538–2548. Neurofibromatosis-1 regulates neuroglial progenitor proliferation and glial Elsen, G.E., Choi, L.Y., Millen, K.J., Grinblat, Y., and Prince, V.E. (2008). Zic1 differentiation in a brain region-specific manner. Genes Dev. 24, 2317–2329. and Zic4 regulate zebrafish roof plate specification and hindbrain ventricle Lee, S.K., and Pfaff, S.L. (2003). Synchronization of neurogenesis and motor morphogenesis. Dev. Biol. 314, 376–392. neuron specification by direct coupling of bHLH and homeodomain transcrip- Falk, J., Drinjakovic, J., Leung, K.M., Dwivedy, A., Regan, A.G., Piper, M., and tion factors. Neuron 38, 731–745. Holt, C.E. (2007). Electroporation of cDNA/Morpholinos to targeted areas of Leucht, C., Stigloher, C., Wizenmann, A., Klafke, R., Folchert, A., and Bally- embryonic CNS in Xenopus. BMC Dev. Biol. 7, 107. Cuif, L. (2008). MicroRNA-9 directs late organizer activity of the midbrain-hind- Falk, S., Wurdak, H., Ittner, L.M., Ille, F., Sumara, G., Schmid, M.-T., brain boundary. Nat. Neurosci. 11, 641–648. Draganova, K., Lang, K.S., Paratore, C., Leveen, P., et al. (2008). Brain area- Lewis, B.P., Shih, I.H., Jones-Rhoades, M.W., Bartel, D.P., and Burge, C.B. specific effect of TGF-beta signaling on Wnt-dependent neural stem cell (2003). Prediction of mammalian microRNA targets. Cell 115, 787–798. expansion. Cell Stem Cell 2, 472–483. Marklund, U., Hansson, E.M., Sundstrom, E., de Angelis, M.H., Przemeck, Flynt, A.S., and Lai, E.C. (2008). Biological principles of microRNA-mediated G.K., Lendahl, U., Muhr, J., and Ericson, J. (2010). Domain-specific control regulation: shared themes amid diversity. Nat. Rev. Genet. 9, 831–842. of neurogenesis achieved through patterned regulation of Notch ligand Gangaraju, V.K., and Lin, H. (2009). MicroRNAs: key regulators of stem cells. expression. Development 137, 437–445. Nat. Rev. Mol. Cell Biol. 10, 116–125. Martello, G., Zacchigna, L., Inui, M., Montagner, M., Adorno, M., Mamidi, A., Gaspard, N., and Vanderhaeghen, P. (2010). Mechanisms of neural specifica- Morsut, L., Soligo, S., Tran, U., Dupont, S., et al. (2007). MicroRNA control tion from embryonic stem cells. Curr. Opin. Neurobiol. 20, 37–43. of Nodal signalling. Nature 449, 183–188. Geling, A., Itoh, M., Tallafuss, A., Chapouton, P., Tannha¨ user, B., Kuwada, McLarren, K.W., Theriault, F.M., and Stifani, S. (2001). Association with the J.Y., Chitnis, A.B., and Bally-Cuif, L. (2003). bHLH transcription factor Her5 nuclear matrix and interaction with Groucho and RUNX proteins regulate the links patterning to regional inhibition of neurogenesis at the midbrain-hindbrain transcription repression activity of the basic helix loop helix factor Hes1. boundary. Development 130, 1591–1604. J. Biol. Chem. 276, 1578–1584. Giraldez, A.J., Cinalli, R.M., Glasner, M.E., Enright, A.J., Thomson, J.M., Megason, S.G., and McMahon, A.P. (2002). A mitogen gradient of dorsal Baskerville, S., Hammond, S.M., Bartel, D.P., and Schier, A.F. (2005). midline Wnts organizes growth in the CNS. Development 129, 2087–2098. MicroRNAs regulate brain morphogenesis in zebrafish. Science 308, 833–838. Merzdorf, C.S., and Sive, H.L. (2006). The zic1 gene is an activator of Wnt Go¨ tz, M., and Huttner, W.B. (2005). The cell biology of neurogenesis. Nat. Rev. signaling. Int. J. Dev. Biol. 50, 611–617. Mol. Cell Biol. 6, 777–788. Murata, K., Hattori, M., Hirai, N., Shinozuka, Y., Hirata, H., Kageyama, R., Guo, H., Ingolia, N.T., Weissman, J.S., and Bartel, D.P. (2010). Mammalian Sakai, T., and Minato, N. (2005). Hes1 directly controls cell proliferation microRNAs predominantly act to decrease target mRNA levels. Nature 466, through the transcriptional repression of p27Kip1. Mol. Cell. Biol. 25, 4262– 835–840. 4271.

Developmental Cell 20, 19–32, January 18, 2011 ª2011 Elsevier Inc. 31 Developmental Cell MicroRNA-9 Function in Xenopus Neurogenesis

Nichane, M., de Croze´ , N., Ren, X., Souopgui, J., Monsoro-Burq, A.H., and Shimojo, H., Ohtsuka, T., and Kageyama, R. (2008). Oscillations in notch Bellefroid, E.J. (2008). Hairy2-Id3 interactions play an essential role in signaling regulate maintenance of neural progenitors. Neuron 58, 52–64. Xenopus neural crest progenitor specification. Dev. Biol. 322, 355–367. Stefani, G., and Slack, F.J. (2008). Small non-coding RNAs in animal develop- Ohtsuka, T., Sakamoto, M., Guillemot, F., and Kageyama, R. (2001). Roles of ment. Nat. Rev. Mol. Cell Biol. 9, 219–230. the basic helix-loop-helix genes Hes1 and Hes5 in expansion of neural stem Storm, E.E., Garel, S., Borello, U., Hebert, J.M., Martinez, S., McConnell, S.K., cells of the developing brain. J. Biol. Chem. 276, 30467–30474. Martin, G.R., and Rubenstein, J.L. (2006). Dose-dependent functions of Fgf8 in Palm, T., and Schwamborn, J.C. (2010). Brain tumor stem cells. Biol. Chem. regulating telencephalic patterning centers. Development 133, 1831–1844. 391, 607–617. Vize, P.D., McCoy, K.E., and Zhou, X. (2009). Multichannel wholemount Palmeirim, I., Henrique, D., Ish-Horowicz, D., and Pourquie, O. (1997). Avian fluorescent and fluorescent/chromogenic in situ hybridization in Xenopus hairy gene expression identifies a molecular clock linked to vertebrate embryos. Nat Protoc 4, 975–983. segmentation and somitogenesis. Cell 91, 639–648. Walker, J.C., and Harland, R.M. (2008). Expression of microRNAs Pena, J., Sohn-Lee, C., Rouhanifard, S., Ludwig, J., Hafner, M., Mihailovic, A., during embryonic development of Xenopus tropicalis. Gene Expr. Patterns Lim, C., Holoch, D., Berninger, P., Zavolan, M., et al. (2009). miRNA in situ 8, 452–456. hybridization in formaldehyde and EDC-fixed tissues. Nat. Methods 6, 139– 141. Walker, J.C., and Harland, R.M. (2009). microRNA-24a is required to repress apoptosis in the developing neural retina. Genes Dev. 23, 1046–1051. Plaisance, V., Abderrahmani, A., Perret-Menoud, V., Jacquemin, P., Lemaigre, F., and Regazzi, R. (2006). MicroRNA-9 controls the expression of Granuphilin/ Wienholds, E., Kloosterman, W.P., Miska, E., Alvarez-Saavedra, E., Berezikov, Slp4 and the secretory response of insulin-producing cells. J. Biol. Chem. 281, E., de Bruijn, E., Horvitz, H.R., Kauppinen, S., and Plasterk, R.H.A. (2005). 26932–26942. MicroRNA expression in zebrafish embryonic development. Science 309, 310–311. Regad, T., Roth, M., Bredenkamp, N., Illing, N., and Papalopulu, N. (2007). The neural progenitor-specifying activity of FoxG1 is antagonistically regulated by Xie, Z.R., Yang, H.T., Liu, W.C., and Hwang, M.J. (2007). The role of microRNA CKI and FGF. Nat. Cell Biol. 9, 531–540. in the delayed negative feedback regulation of gene expression. Biochem. Biophys. Res. Commun. 358, 722–726. Sabherwal, N., Tsutsui, A., Hodge, S., Wei, J., Chalmers, A.D., and Papalopulu, N. (2009). The apicobasal polarity kinase aPKC functions as Yang, X., Klein, R., Tian, X., Cheng, H.-T., Kopan, R., and Shen, J. (2004). a nuclear determinant and regulates cell proliferation and fate during Notch activation induces apoptosis in neural progenitor cells through a p53- Xenopus primary neurogenesis. Development 136, 2767–2777. dependent pathway. Dev. Biol. 269, 81–94. Shibata, M., Kurokawa, D., Nakao, H., Ohmura, T., and Aizawa, S. (2008). Zhao, C., Sun, G., Li, S., and Shi, Y. (2009). A feedback regulatory loop MicroRNA-9 modulates Cajal-Retzius cell differentiation by suppressing involving microRNA-9 and nuclear receptor TLX in neural stem cell fate deter- Foxg1 expression in mouse medial pallium. J. Neurosci. 28, 10415–10421. mination. Nat. Struct. Mol. Biol. 16, 365–371.

32 Developmental Cell 20, 19–32, January 18, 2011 ª2011 Elsevier Inc. Developmental Cell, Volume 20

Supplemental Information

MicroRNA-9 Reveals Regional Diversity of Neural Progenitors along the Anterior-Posterior Axis Boyan Bonev, Angela Pisco, and Nancy Papalopulu Figure S1. miR‐9 expression in Xenopus Tropicalis ‐ relates to Figure1

Figure S2. Injection of control morpholino has no effect on neural development – relates to Figure 2, 3 and 4

Figure S3. Target screen identified several miR‐9 targets – relates to Figure 5

Figure S4. Hairy1 is expressed in the ZLI boundary region and overlaps with Fgf8 domain – relates to Figure 5 and 7

A. miR-9 miR-9* xtr-miR-9a-1 GGGGUUGGUUGUUAUCUUUGGUUAUCUAGCUGUAUGAGUGUUGUCAAUCCUUCAUAAAGC UAGAUAACCGAAAGUAAAAAUAACCCCA xtr-miR-9a-2 GGAAGUGGUUGUUAUCUUUGGUUAUCUAGCUGUAUGAGUGUAUUGGU--UUUCAUAAAGC UAGAUAACCGAAAGUAAAAACUCCUUC- xtr-miR-9-3 ------GUUUCUAUCUUUGGUUAUCUAGCUGUAUGAGUGUAAAUAAGCCGUCAUAAAGC UAGAUAACCGAAAGUAGGAAUCA----- xtr-miR-9b ------GUUUCUGUCUUUGGUUACCUAGCUGUAUGAGUAUAACUAA--UGUCAUAAAGC UAGACAACCGAACGUAUAAACCA----- *** * ********** ************** * ********* **** ******* *** ** B. miR-9 LNA C. ForebrainMidbrain Hindbrain

D.Forebrain Hindbrain E. Forebrain Hindbrain miR-9a-1 miR-9a-1 Sox3 Sox3

Figure S1. miR-9 expression in Xenopus tropicalis (A) Sequence alignment of miR-9 predicted precursors. Positions that have an identical single residue are marked with an asterisk. Mature miR-9 sequence is shown in red. (B) Whole mount in situ hybridization for mature miR-9 (LNA-based probe). Scale bar = 200 μm (C) Transverse sections of embryos stained for mature miR-9. Scale bar = 20 μm (D) In situ hybridization for miR-9a-1 primary transcript in stage 36 X.tropicalis embryos in whole mount and transverse sections. Scale bar = 20 μm (E) Fluorescent in situ hybridization (FISH) for miR-9a-1 (in red) combined with immunohistochemistry for Sox3 (marker for neural progenitors) in st. 36 embryo. Scale bar = 20 μm A. Drosha Dicer D. Control MO x x Myt1 Myt1 FITC DAPI x antisense MO miR-9

B. miR-9 LNA C. Forebrain 1.4 co inj co inj 1.2 Myt1 Myt1 1.0 FITC DAPI Control 0.8

0.6

0.4

0.2 miR-9 relative levels Hindbrain 0.0 miR-9 MO Control MO miR-9 MO co inj co inj E. Forebrain Hindbrain G. Forebrain Hindbrain Sox3 Sox3 TUNEL TUNEL FITC FITC FITC FITC DAPI DAPI DAPI DAPI

co inj co inj co inj co inj

F. H. miR-9 MO + Casp. Inh. pH3 pH3 TUNEL TUNEL FITC FITC FITC FITC DAPI DAPI DAPI DAPI

co inj coinj inj co inj co inj

Figure S2. Injection of control morpholino has no effect on neural development (A) Representative schematic of the mechanism of action of miR-9 MO. (B) In situ hybridization for mature miR-9 in control and miR-9 MO injected embryos, Scale bars = 200μm (C) Real-time PCR for mature miR-9 in control MO and miR-9 MO injected embryos, n=3 (10 embryos each), error bars represent SEM Injection of control MO has no effect on the number of Myt1 positive cells (D), Sox3 positive area (E), phosphohistone H3 positive cells (F) or apoptosis (G). (H) Injection of caspase inhibitor together with miR-9 MO prevents the apoptosis in the forebrain as indicated by TUNEL staining. In all images FITC was used to identify the injected side and DAPI to counterstain the DNA. Neural tube is outlined with a dashed line. Scale bars = 20μm Stage 30X.tropicalisembryos. (F) Wholemountinsituhybridization forhairy1,hairy2,NR2E1/TLX, Tis21/Btg2, Sirt1or TGF expressed relativetothecontrollevels.Errorbarsrepresent s.d. (miR-9-2 WT)ormutantMut)miR-9-2.Luciferase expressionwasnormalizedand (E) A subsetofthetargetsin(D)werefurthervalidatedusingcontrolvector(Control), wild-type sion wasnormalizedandexpressedrelativetothecontrol levels.Errorbarsrepresents.d. target togetherwithscrambledprecursors(Control) ormiR-9precursors(miR-9).Luciferaseexpres- (D) HeLacellsweretransfectedwithluciferasereporter carryingthe3’UTRofpredictedmiR-9 promoter regionsdrivingtheexpressionofRenilaand Fireflyluciferaserespectively. of thepsi-CHECK2luciferasereportervectorused in thisstudy. SV40andHSV-TK represent cell cycleprogressionandapoptosisamongthesignificantcategories.(C)Schematicrepresentation (B) GO Analysis usingmiR-9predictedtargetsinmammalsidentifiesneuronaldifferentiation, G1/S along the A-P axisor Targets A andBareexpressedonlyintheforebrainorhindbrainrespectively. (A) Possiblemechanismoftheregional-specificmiR-9MOphenotype: Target A (yellow)isexpressed Figure S3.Target screenusedtoidentifyseveralputativemiR-9targets C. .B. A. .F. E. HSV-TK

Relative Luciferase Expression,100 %120 SV40 20 40 60 80 0 o3URHiy ar2TxTs1Sr1TFR BL Onecut1 MBNL1 TGFBR2 Sirt1 Tis21 Tlx Hairy2 Hairy1 No 3'UTR Renila-Luc Firefly-Luc Stage 30tadpole Control 3’UTR oftargetgene miR-9-2 WT D. miR-9-2 Mut Target Target Target Relative Luciferase Expression, % 100 120 20 60 80 40 No 3'UTR 0 A B A

F oxG1

Hairy1

Rest1

Development and F unction unction and F Cell Death cell death 3.06E-03 3.06E-03 death cell Death Cell Category Fu nction P-value Mo lecules lecules Mo P-value nction Fu Category nPTB ycle C Cell Nervous System Hairy1 Tis21

ID4

S o x2 Differentiation Differentiation G1/ S phase S G1/ of neurons of neurons

Sirt1 transition Control TGFBR2

NR2E1/Tlx 2.13E-02 2.13E-02 3.72E-05 3.72E-05 Hairy2 Hen1 Sirt1

HDAC5 miR-9

Fzd2 NR2E1, NTRK3,REST, RET, SHC1 GAB2, HES1,ISL1,KCNMA1, BTG2, CNTFR,EFNA1,EN1,GAB2, HES1, MAP2K3,NCOA3,SOX2 ATF3, CCNE1,CCNE2ELAV1, REST, RET, RUNX1,SHC1 LMX1A, NR2E1,NTRK3,PSEN1, ID4, ISL1,KCNMA1,LIFR, EN1, EN2,FGF5,GAB2,HES1, BTG2, CNTFR,CNTN4,EFNA1

CNTFR

PSEN1

MBNL1 TGFBR2 TLX Hairy2

β B T

R2 in G2/Tis21

Onecut1 A. a. Shh B. Irx3 Hairy1 Shh Hairy1 Hairy1

c. Fgf8 C. Fgf8 Hairy1 Hairy1

b. Irx3 Hairy1

Figure S4. Hairy1 is expressed in the ZLI boundary region and overlaps with Fgf8 domain (A) Hairy1 co-localizes with the marker of the ZLI Shh in the forebrain region as shown by double fluorescent in situ hybridization. The boxed region is enlarged in (a). (B) Hairy1 is expressed immediately adjacent to Irx3 as shown by double fluorescent in situ hybridization. The boxed region is enlarged in (b). (C) Hairy1 expression overlaps with Fgf8 in the developing dorsal forebrain. The boxed region is enlarged in (c). In all images DAPI is used to stain the DNA (blue). Scale bars = 20 μm

Chapter 3. microRNA-9 modulates Hes1 ultradian oscillations by forming a double negative feedback loop

Manuscript in preparation

34 microRNA-9 modulates Hes1 ultradian oscillations by forming a negative feedback loop

Boyan Bonev and Nancy Papalopulu*

Faculty of Life Sciences, Michael Smith Building, University of Manchester, Oxford Road, Manchester, M13 9PT, UK

* author for correspondence E-mail: [email protected] Tel:+44-161-3068907 Fax:+44=161-2755082

Keywords: microRNA, miR-9, Hes-1, neural progenitor, oscillation, mRNA stability, negative feedback, neurogenesis

35 Abstract

The expression of Hes1, a Notch signaling effector, oscillates in neural progenitors with short periodicity of 2-3 hours. Hes1 oscillations maintain progenitors in a proliferative state, while constant repression of Hes1 leads to neuronal differentiation. Hes1 oscillations are driven by auto-repression, coupled with high protein and RNA instability. It is unknown how Hes1 mRNA stability is controlled and how cells exit oscillations in order to differentiate. Here, we show that a microRNA - miR-9, controls the stability of Hes1 mRNA and that miR-9 overexpression dampens Hes1 oscillations. In addition, we show that Hes1 represses the transcription of miR-9, resulting in out of phase oscillations of primary mir-9 transcripts. However, mature miR-9 is very stable and accumulates over time, providing a mechanism for cells to escape Hes1 oscillation cycles and differentiate. Our findings identify miR-9 as a novel component of ultradian oscillations and a new mechanism to promote neural differentiation by dampening Hes1 oscillations.

36 Introduction

The correct development of the CNS relies on a balance of differentiation and progenitor maintenance. Neural progenitors undergo a period of proliferation to expand the progenitor pool, followed by controlled cell cycle exit to generate neurons and glial cells, at specific time points of development and in a defined order (Alvarez-Buylla et al., 2001; Temple, 2001; Miller and Gauthier, 2007). One of the key challenges is to understand how neuronal differentiation is promoted, while sufficient numbers of progenitors are maintained to generate neurons over a prolonged period of time.

One of the major signaling pathways involved in regulating progenitor maintenance is the Notch signaling pathway, which operates by a process known as lateral inhibition (Gaiano and Fishell, 2002). In neural development, the classical view of lateral inhibition is as follows: proneural genes such as Mash1 and Ngn2 upregulate the expression of Notch receptor ligands (Delta-like 1 – Dll1 or Serrate) in one cell. Delta then binds to notch receptor on the neighboring cell, inducing the cleavage of the intracellular notch domain (NICD) and its translocation into the nucleus (Selkoe and Kopan, 2003). There, NICD forms a complex with the transcription factor RBPj and activates expression of Notch effectors, notably the basic helix-loop-helix (bHLH) transcriptional repressors of the Hairy-Enhancer of split (Hes) family, such as Hes1 and Hes5 (Ohtsuka et al., 1999). These genes are master repressors of neuronal differentiation, as shown by the outcome of knockdown and overexpression studies in several vertebrate species (Ishibashi et al., 1994; Ohtsuka et al., 1999). They suppress autonomously the expression of the proneural factors Mash1 and Ngn2 and thus inhibit differentiation.

With this feedback mechanism, a group of cells that start off with approximately the same level of Notch signaling would be able to amplify small stochastic differences (Artavanis-Tsakonas et al., 1999), until some cells express sufficiently high levels of Dll1/proneural factors to differentiate, while the neighbouring cells remain as progenitors. This model predicts that Hes genes, and other components of the Notch signaling pathway, would be expressed in a salt- and-pepper manner and at variable levels, reflecting the selection of cells by lateral

37 inhibition and the gradual progression towards differentiation. Indeed, in areas undergoing neuronal differentiation, such as the ventricular zone (VZ) of the mouse telencephalon, Hes1 is expressed at varying levels, in a salt-and-pepper manner.

However, this classic view of gradual amplification and linear cell fate progression has been recently challenged by the observation that Hes1, and other genes in the Notch signaling pathway, display short-period (ultradian) oscillatory expression in a number of cell types, including neural progenitors, in vitro and in vivo (Shimojo et al., 2008). This observation has led to a radically revised view of lateral inhibition where, instead of a gradual amplification of small differences, Hes1 expression appears in a salt and pepper pattern because cells are in different points of an oscillatory curve. This pioneering work, led to the proposal that Hes1 oscillations are necessary for efficient proliferation of progenitor cells and they serve to keep a population of cells in a progenitor state. Thus, 3 states of Hes1 expression can be recognized in the developing CNS; high and persistent expression of Hes1 in regions that contain slow cycling, non-neurogenic, progenitors (e.g. “boundary” regions such as the MHB and the zona limitans intrathalamica – ZLI) (Baek et al., 2006), low or no expression in cells undergoing neuronal differentiation and variable, oscillatory, expression in neural progenitors.

Given the importance of Hes1 oscillations for progenitor maintenance a major remaining question is how such oscillations are controlled. The core requirement appears to be a negative feedback loop, whereby Hes1 protein represses its own transcription. Both Hes1 mRNA and protein are also extremely unstable, with a half-life in the order of 20mins; the rapid degradation of Hes1 protein and Hes1 mRNA results in release from inhibition and initiation of the next cycle of expression (Hirata et al., 2002). However, while the mechanisms of protein degradation have been largely elucidated (Hirata et al., 2002), how mRNA stability is regulated is not well understood. A 25nt element, which is important for the mRNA degradation, has been identified in the 3‘UTR of the frog homologue hairy2a (Davis et al., 2001), although it remains unclear if that is the major regulatory motif and how it affects oscillations. Furthermore, the mechanism whereby cells

38 terminate oscillations and permanently downregulate Hes1 to enter the neuronal differentiation pathway, are completely unknown.

MicroRNAs are a class of small non-coding RNAs, which regulate gene expression at the post-transcriptional level (Bartel, 2009). Recently, microRNAs have been shown to regulate their target expression primarily at the RNA level through deadenylation and decapping of the message (Lim et al., 2005; Giraldez et al., 2006; Guo et al., 2010), suggesting that they are prime candidates for controlling mRNA stability. MiR-9, a highly conserved microRNA, is expressed predominantly in the CNS of the developing embryo (Wienholds et al., 2005) and is of particular importance in the development of the CNS, in numerous organisms (Leucht et al., 2008; Zhao et al., 2009; Delaloy et al., 2010; Shibata et al., 2011). Recently, we have shown that xenopus miR-9 binds to the a highly conserved site in the 3’UTR of the Hes1 homologue, hairy1 and that it regulates its expression at the RNA level (Bonev et al., 2011).

Here, we investigate whether miR-9 has a role in controlling the stability of hes1 mRNA and wheter it is a missing player in the hes1 oscillator. We show that miR-9 regulates Hes1 at the RNA level and that interfering with the binding of miR- 9 in the the Hes1 3’UTR, stabilizes Hes1 mRNA, suggesting that miR-9 normally contributes to its instability. Both overexpression and inhibition of miR-9 alter Hes1 osillations; overexpression of miR-9 dampens serum induced Hes1 oscillations, while inhibiting miR-9 increases their amplitude. We show that miR-9 and Hes1 are both expressed in a proliferative area of the mouse telencephalon but when viewed with single cell resolution, the levels of miR-9 and Hes1 are inversely related. In addition, we show that Hes1 regulates miR-9 levels by inhibiting the transcription of its primary transcripts, leading to coupled out-of-phase oscillations. Surprisingly, the levels of mature miR-9 do not oscillate, which is due to high stability of the mature miR-9 transcript. We show that periodic trascription of pri-miR-9, coupled with stability of the mature form, leads to a stepwise accumulation of miR-9 over time. Since overexpression of miR-9 dampens Hes1 oscillations, we propose that accumulation of mir-9 would eventually lead the cells to exit the oscillatory phase and initiate differentiation. Our findings identify an essential, but previously unknown, component of the Hes1 molecular oscillator

39 and furthermore, provide a plausible mechanism for the elusive problem of how oscillations are terminated.

Results miR-9 regulates mouse Hes1 at the RNA level

We have previously shown that miR-9 regulates the xenopus homologue of Hes1 – hairy1 and that the miR-9 binding site is highly conserved in its vertebrate homologues with 100% sequence homology in the seed-complementary region (Bonev et al., 2011). In order to determine whether the mouse Hes1 is also regulated by miR-9 in vitro, we designed a luciferase reporter fused to either the wild-type (WT) or mutated (seed-complementary region is deleted) Hes1 3’UTR. While the expression of the WT luciferase reporter in HeLa cells was significantly repressed by artificial miR-9 precursor mimics, the expression of the mutated luciferase reporter was not affected, confirming the direct repression by miR-9 (Figure 1A). Importantly when the luciferase reporter constructs were introduced into a neural progenitor cell line expressing endogenous miR-9 (c17.2) the expression levels of the Hes1 3’UTR Mut reporter were significantly higher than those of the WT reporter (Figure 1B), suggesting that endogenous miR-9 is able to repress the expression of the WT, but not of the Mut reporter.

In order to further evaluate the regulation of Hes1 by miR-9, we decided to use a locked-nucleic acid (LNA) antisense probe to knock down miR-9 in the c17.2 cell line. Additionally, to validate the specificity of the repression we used a target protector LNA-modified oligo to specifically disrupt miR-9/Hes1 3’UTR binding – Figure 1C (based on Choi et al., 2007). We confirmed that miR-9 LNA inhibitor reduced the levels of endogenous miR-9 to about 16% of the control, while Hes1 TP had no effect on miR-9 levels as predicted (Figure 1D).

To confirm the interaction between miR-9 and Hes1, we manipulated miR-9 levels in c17.2 cells and examined endogenous Hes1 protein using western blot. Overexpression of miR-9 significantly reduced the amount of Hes1 protein compared to scrambled control, while inhibiting miR-9 using LNA inhibitor led to

40 an increase in the levels of Hes1 protein (1.32 fold increase compared to the control). Transfecting Hes1 TP also upregulated Hes1 levels, but had a more modest effect (1.12 fold increase compared to the control).

Next, we asked whether endogenous Hes1 is regulated by miR-9. Transfection of either miR-9 LNA or Hes1 TP in c.17.2 cells led to an increase in the endogenous Hes1 RNA levels measured by qRT-PCR (Figure 1F). In order to determine if this increase is due to RNA stabilization, we examined whether miR-9 promotes Hes1 degradation. We used actinomycin D to block transcription in a serum-synchronized c17.2 cells as previously described (Hirata et al., 2002). Then, we examined the degradation rate of Hes1 mRNA over 3hrs using RT-PCR and found out that while in cells transfected with the control LNA inhibitor, Hes1 mRNA had a half-life of 25±2.3 min (similar to reports in other cell lines), transfection of either miR-9 LNA and Hes1 TP led to a significant increase in the stability of the Hes1 mRNA to 32.3±2.25min and 35±5min respectively (Figure 1G and 1H). In addition, overexpression of miR-9 using precursor mimics led to a reduction in Hes1 mRNA half-life to 20±3.2min (Figure 1H), while scrambled precursors showed no effect on the stability (data not shown).

Overall, these findings show that miR-9 regulates Hes1 mRNA expression directly by promoting its degradation. This suggests that in normal conditions miR-9 is required for destabilizing Hes1 mRNA and thus shows similar mechanisms of repression compared to hairy1 in X. tropicalis. miR-9 modulates serum-induced Hes1 oscillations

Hes1 oscillations are driven by negative feedback, delay and high instability of both the RNA and the protein (Hirata et al., 2002) – schematized in Figure 2A. First we examined whether Hes1 displays oscillatory behavior in the c17.2 cell line. We used ubiquitinated luciferase reporter driven by Hes1 promoter and containing either the wild-type Hes1 3’UTR (Hes1Pr-ubqluc-3’UTR WT) or a construct which carries mutated 3’UTR (the region corresponding to miR-9 seed was deleted) – Hes1Pr-ubqluc-3’UTR Mut (Shimojo et al., 2008) to image Hes1 oscillations in single cells. Oscillations were quite asynchronous and varied from cell to cell with some cells downregulating Hes1 expression over prolonged period of time (presumably due to them entering G1 – Shimojo et. al., 2008)

41 Next, we examined the number of cycles that were displayed over a period of 20h imaging and we found that cells transfected with Hes1Pr-ubqluc-3’UTR Mut construct display different distribution and has significantly reduced average number of cycles (p<0.01) (Figure 2C and 2D). The average period of the cells displaying 2 or more cycles of oscillations appeared to be unchanged, while the amplitude was increased in the Hes1-ubqluc-3’UTR Mut expressing cells (Figure 2E).

In order to examine how manipulation of miR-9 levels would affect Hes1 oscillations in a synchronized population of cells, we decided to measure Hes1 RNA levels over time upon serum stimulation as described previously (Yoshiura et. al., 2007). 24h after transfection with either scrambled control or miR-9 precursors c17.2 cells were shifted to low serum (0.2%) media for additional 24h. In order to stimulate Hes1 expression, serum was then added to the media to a final concentration of 10% at t=0 and Hes1 mRNA levels were followed over a period of 6 hours by qRT-PCR. While we were able to observe only 2 cycles of oscillations with this method, they appeared to be were remarkably similar to those observed in fibroblasts (Yoshiura et al., 2007). Transfection of miR-9 precursor mimics caused a small decrease in the amplitude of the first cycle and a significantly dampened second peak, which was apparent at t=90min post stimulation compared to t=180 for the scrambled precursors control (Figure 2F). This suggests that overexpression of miR-9 causes dampening of Hes1 oscillations and decreases their periodicity.

Next, we examined how inhibiting miR-9 or interfering with miR-9/Hes1 binding by using Hes1 TP would affect oscillations. Both Hes1 TP and miR-9 led to an increased amplitude of the first peak of Hes1 oscillations, while miR-9 LNA also caused a slight shift of the first peak in addition – from 30min to 60 min after serum addition (Figure 2F). Hes1 TP (and to a lesser extend miR-9 LNA inhibitor) also caused an increase in the basal levels of Hes1 after serum stimulation but also to higher amplitude in the second oscillation peak (Figure 2G – t=150 compare with control).

These results show that miR-9 regulation is important for establishing the oscillatory pattern of Hes1 and that overexpression leads to dampening of the

42 oscillations and reduced period, while inhibiting miR-9 causes increase in the amplitude of the oscillations with no significant change in the periodicity.

miR-9 and Hes1 are expressed in the same region of the mouse telencephalon

Since we have showed that miR-9 regulates Hes1 at the RNA level, we decided to examine how the two are expressed in relation to each other in vivo. In the mouse E12 dorsal telencephalon miR-9 expression was restricted to the ventricular zone and didn’t overlap with the marker for differentiated neurons acetylated tubulin (Figure 3A).

In order to compare the expression of miR-9 and Hes1 mRNA in the mouse VZ we performed a double fluorescent in situ hybridization (FISH). Hes1 was expressed in a salt-and-pepper manner throughout the neocortex, while mature miR-9 was abundant in the hippocampus and in the dorsal are of the neocortex (Figure 3B).

Next, we asked how miR-9 and Hes1 were expressed in the c17.2 mouse neural progenitor cell lines we used in our previous experiments. When grown in conditions promoting proliferation (with 10% serum supplement) these cells expressed both miR-9 and Hes1 (Figure 3C upper panels). Serum withdrawal (conditions promoting neuronal differentiation) led to an accumulation of miR-9 levels (confirming previous reports – (Laneve et al., 2010)), which was accompanied by a decrease in Hes1 levels, consistent with its role of inhibitor of neurogenesis (Figure 3D). The results obtained using qRT-pCR were confirmed using a double fluorescent in situ hybridization (Figure 3C lower panels).

Overall these results show that both miR-9 and Hes1 are expressed in the VZ zone of the developing mouse telencephalon and neural progenitor cells, but their expression is inversely related – when miR-9 is high, Hes1 is low and vice versa.

43 Hes1 inhibits the transcription of miR-9 primary transcripts

We have showed that miR-9 represses Hes1, which could explain the high miR-9/low Hes1 condition, but we wondered what is the mechanism responsible for reducing miR-9 levels when Hes1 is high. Since Hes1 is a known transcriptional repressor, we hypothesized that it can bind to miR-9 promoters and directly repress its transcription. We performed a bioinformatics search for Hes1 binding motifs in the promoter regions of the three pri-miR-9s. We found 2 putative N-box (5'-CACNAG-3') motifs and 4 E-boxes (5'-CANNTG-3') in the 2kB region upstream of the pre-miR-9-1; 1 N-box and 6 E-box motifs in the miR-9-2 promoter region and 4x E-box motifs in the putative miR-9-3 promoter (Figure 4A). To address whether Hes1 can repress miR-9 transcription we overexpressed Hes1 in the c17.2 cell line and examined the expression of miR-9 primary transcripts using qRT-PCR. Importantly, we found that all three primary miR-9 transcripts were significantly reduced to about 50% of the control levels (Figure 4B).

MiR-9-2 has been shown to be the most abundant primary transcript (Shibata et al., 2011) in the mouse brain, which we confirmed in the c.17.2 cell line using qRT-PCR (Figure 4C). Next, we designed a luciferase reporter vector under the control of the miR-9-2 promoter (identified by (Laneve et al., 2010)). Overexpression of Hes1 decreased the luciferase signal in a dose-dependent manner, confirming the specificity of Hes1 regulation (Figure 4D).

Next, we hypothesized that if Hes1 inhibits the expression of miR-9 precursors then miR-9 precursors will oscillate out-of-phase with Hes1. Indeed, the levels of miR-9-1 and miR-9-2 precursors after serum stimulation showed a burst of expression peaking around 120min after serum stimulation, which occurred when Hes1 levels were low (Figure 4E). Interestingly, the expression of mIR-9-3 was not affected by serum stimulation (data not shown) suggesting that release of Hes1 inhibition is not sufficient to induce the transcription of this particular transcript.

We wondered how transcriptional repression of pri-miR-9 by Hes1 translates into regulation at the mature miR-9 levels. To address this we overexpressed Hes1 in c17.2 cells and examined the expression of mature miR-9. Unexpectedly, even though the same treatment causes significant reduction at the

44 primary transcripts, mature miR-9 levels appeared to be unchanged (Figure 4F). Since most microRNAs have been shown to have a relatively high stability, we wondered whether the reason for this discrepancy was the low turnover rate of the existing miR-9 over the course of the experiment (48h). To test this, we examined the degradation of mature miR-9 over time after blocking transcription using actinomycin D. We observed that mature miR-9 was indeed very stable over 3 hours, compared to Hes1 mRNA (Figure 4G).

However, if mature miR-9 is stable and does not get degraded over a long period, we hypothesized that the bursts of pri-miR-9 transcription upon serum stimulation would lead to a gradual accumulation of mature miR-9 over time. Indeed when we followed mature miR-9 levels for 8 hours after serum stimulation we observed an increase which started at around t = 200-250min (Figure 4H) followed by a plateau region and another, smaller, increase. This also suggests that there is a 90-120min delay between the initial burst of primary miR-9 transcription (which occurred at t = 120min) and the start of its accumulation at the mature levels. Next we tested if this effect happens over longer time period by measuring mature miR-9 levels over the course of 3 days in cells grown in proliferative conditions. We confirmed that miR-9 levels increase even if there is no manipulation of the experimental system (Figure 4I), in agreement with the hypothesis that mature miR-9 autonomously accumulates over time in neural progenitors.

These results combined with the dampening of Hes1 oscillations which occurs when miR-9 is overexpressed lead us to propose the following model. In neural progenitors miR-9 primary transcripts oscillate out-of-phase with Hes1, which leads to gradual accumulation of mature miR-9 with increasing number of cycles. At a certain threshold levels of miR-9 Hes1 oscillations are dampened leading to increased probability of adopting neuronal fate (Figure 5).

Discussion

Oscillatory gene expression is a widespread and important phenomenon in living systems, from circadian clocks to oscillations of regulatory factors in the

45 immune system. Within the context of developing embryos, oscillations have been recognized relatively recently, because it is technically very challenging to image ultradian oscillations and moreover, to do so in vivo. One of the best-known cases of developmental oscillations is observed during somitogenesis, when blocks of somitic cells are periodically segregated during axis formation. More recently, oscillations in regulatory gene expression have been shown to take place in neural progenitors of the mouse cortex. Oscillations have been observed in components of the Notch signaling pathway (Hes1, Ngn2 and Dll1) and these findings have transformed our view of lateral inhibition from a linear amplification of stochastic differences to that of a dynamic, cyclical and mutual, inhibition of differentiation.

Oscillations in Hes1 have been shown to be important for efficient proliferation and for maintenance of a group of cells as progenitor. While both Hes1 protein and mRNA are unstable, mechanistic studies have focused on the regulation of protein stability. The regulation of mRNA stability had been overlooked because of a lack of an understanding of the molecular components that control it. However, mathematical modeling predicted that mRNA stability is important as Hes1 oscillates only within certain values of mRNA stability and proposed the existence of microRNA-mediated control of hes genes. Here, we have shown that miR-9 is involved in regulating the Hes1 oscillator.

Since Hes1 oscillations are asynchronous in neural progenitor cells, the population appears highly heterogeneous with regards to gene expression, at any given time. A role of microRNAs in promoting such heterogeneity, which may be important for cell fate choices, contrasts with the more widely held view that microRNAs serve to buffer transcriptional “noise”, promoting homogeneity of response across a population of cells. Thus, our findings expand the known roles of microRNAs from promoting robustness and controlling cell fate switch, to include heterogeneity.

Here, we have also shown that miR-9 and hes1 are coupled in a double negative loop where miR-9 regulates negatively the stability of Hes1 mRNA and Hes1 protein downregulates the transcription of miR-9 primary transcripts. There are two important outcomes of this interaction; first, pri-MiR-9s also oscillates

46 because it is periodically shut down but it does so out of phase with Hes1. This fits very well with the expression pattern of Hes1 and pri-miR-9 observed in the embryo and cell lines; both Hes1 and miR-9 are expressed in a salt and pepper manner but their levels are inversely related. Second, the mutually repressive interaction of miR-9 and Hes1 predicts that when Hes1 levels are constitutively high (as in boundary regions), the levels of miR-9 would be low, while when the levels of Hes1 would constitutively low (as in neurons) the levels of miR-9 would be high. This is indeed the case, as miR-9 is not expressed in boundary regions of the vertebrate CNS (Leucht et al., 2008; Bonev et al., 2011), where Hes1 is high; but it is expressed in post-mitotic neurons, where Hes1 is low (Bonev et al., 2011; unpublished).

While maintaining neuroepithelial cells in an oscillatory phase of Hes1 expression is important for progenitor maintenance, exiting oscillations with low Hes1 is equally important, as it allows neuronal differentiation to occur. However, the mechanism where this might be happening was completely unknown. We have provided evidence due to that while pri-miR-9 oscillates, mature miR-9 levels accumulate over time due to its high stability. We have also shown that experimentally raising miR-9 levels has a quantitative effect on reducing Hes1 mRNA stability, eventually dampening oscillations of Hes1. Taken these findings together, we propose a model whereby the accumulation of miR-9 over time allows the cells to escape the progenitor state by dampening Hes1 oscillations. Thus, the double negative feedback loop of pri-miR-9 and Hes1, coupled with high stability of the mature miR-9 RNA, provides the first mechanistically plausible explanation for the timed exit of cells from the progenitor compartment.

Acknowledgements

We are grateful to Prof. Mike White and Dr. David Spiller for their invaluable help with live imaging. The work was funded by the Wellcome Trust. NP is a WT Senior

47 Research Fellow and BB is a WT 4 yr PhD Student. The authors declare no conflict of interest.

Material and methods

In situ hybridization

Locked-nucleic-acid modified (LNA), digoxigenin labeled probes were obtained from Exiqon. Fluorescent in situ hybridization (FISH) on 12μm cryosections from E11.5 mouse forebrain were performed as previously described (Pena et al., 2009), with the following modifications – no proteinase K treatment, probe concentration was 25nM and primary anti-DIG-POD antibody (1:1000, Roche) incubation was performed overnight at 4°C. For double FISH Dig-labeled miR-9 LNA probe and FITC-labeled Hes1 probe were incubated simultaneously and miR-9 expression was detected first using TSA-Cy3 method (Perkin Elmer) while Hes1 expression was detected using anti-FITC mouse primary antibody (1:400, Roche) and anti- mouse Alexa-488 secondary antibody (1:500, Molecular Probes). To combine FISH with an immunostaining the respective primary antibody was incubated together with anti-Dig-POD and the secondary was incubated for 1 hour after TSA-Cy3 was completed. Detailed protocol for the FISH used to detect miR-9 in c17.2 is available upon request.

Cryosectioning and immunostaining

Stage E12 mouse embryos fixed in 4% PFA were sectioned on a LEICA CM3050 S cryostat after embedding in 25% fish gelatin/15% sucrose and stained as described previously (Chalmers et al., 2003; Regad et al., 2007). The following primary antibodies were used: acetylated tubulin (1:1000, Sigma), anti-pH3 (1:500, Upstate). Appropriate secondary antibodies were obtained from Molecular Probes and used at 1:500 dilution

48 Cell culture and Luciferase Reporter Assay c17.2 mouse neural progenitor cell line was a generous gift from Prof Mike White (University of Manchester). They were maintained in proliferative state in DMEM growth media supplemented with 10% serum and antibiotics. Neuronal differentiation was induced by serum withdrawal. For luciferase reporter assays cells were plated in a 24 well plate and transfected after 24hours with 100ng of the luciferase reporter and either 30nM of scrambled/miR-9 precursors (Ambion) or 50nM of control LNA/miR-9 LNA/Hes1 TP oligonulceotides. Luciferase expression was analyzed after 48 hours using Dual Luciferase Assay system (Promega). Renila luciferase activity was normalized by the co-expressed firefly luciferase and expressed as a percentage of the control. All assays were repeated at least 3 times and performed in triplicates each time.

RNA isolation, RT-PCR, and quantitative Real-Time PCR analysis

Cells were plated in a 24 well plate 24h prior to transfection. They were transfected at ≈80% density and were shifted to 0.2% Serum media after 24h to synchronize the population as previously described (Hirata et al., 2002). For analysis of Hes1 oscillations serum was added to the cells to a final concentration of 10% at t=0 and RNA was extracted every 30min using TRIZOL (Invitrogen). 1μg of total RNA was retrotranscribed using RT-AMV (Invitrogen) according to the manufacturer’s instructions. miR-9/U6 cDNA was reverse transcribed from 50ng total RNA using TaqMan microRNA reverse transcription kit (ABI) and the levels were accessed using Taqman microRNA assay (ABI) according to the manufacturer’s instructions. Quantitative real-time PCR was performed in an ABI StepOnePlus Sequence Detection System (Applied Biosystems) using Taqman Fast RealTime PCR master mix and probes purchased from Applied Biosystems.

Measurement of RNA half-life

C17.2 cells were transfected with control LNA, miR-9 LNA and Hes1 TP at ~80% confluency 24h after plating. After another 24h cells were washed with PBS and transferred to media containing 0.2% Serum to synchronize the population. Hes1 expression was stimulated by adding serum to the media to a final concentration

49 of 10% for 60min and 10μM actinomycin D was added at t=0 to block transcription. RNA was extracted at t=0, 15, 30, 45, 60, 90, 120 and 180 min using TRIZOL (Introgen). Hes1 levels were analyzed by qRT-PCR and normalized for GAPDH expression, while miR-9 levels were normalized for non-coding RNA U6. The relative expression levels were plotted on a log scale and the degradation rate k was determined using first-order exponential fit. Half-life of the Hes1 RNA was then calculated as t1/2=ln2/kdecay. Values were expressed as the mean of three independent experiments ± st dev.

Bioluminescence imaging of Hes1 expression in the c17.2 cell line

Stable c17.2 cell line expressing ubiquitinated luciferase under the control of Hes1 promoter, fused to Hes1 3’UTR, were cutured in DMEM medium containing 10% serum and 1mM luciferin. For measurement of bioluminescence, the dish was placed on the stage of inverted microscope and was maintained at 37C in 5% CO2. Bioluminescence was collected using the 4xobjective and was transmitted directly to a cooled charge-coupled device (CCD) camera as described elsewhere (Masamizu et al., 2006). The signal-to-noise ratio was increased by 4x4 binning and 30 min exposure.

References:

Alvarez-Buylla, A., Garcia-Verdugo, J. M. and Tramontin, A. D. (2001) 'A unified hypothesis on the lineage of neural stem cells', Nat Rev Neurosci 2(4): 287-93.

Artavanis-Tsakonas, S., Rand, M. and Lake, R. (1999) 'Notch signaling: cell fate control and signal integration in development', Science 284(5415): 770-6.

Baek, J. H., Hatakeyama, J., Sakamoto, S., Ohtsuka, T. and Kageyama, R. (2006) 'Persistent and high levels of Hes1 expression regulate boundary formation in the developing central nervous system', Development 133(13): 2467-76.

Bartel, D. P. (2009) MicroRNAs: target recognition and regulatory functions Cell, vol. 136.

Bonev, B., Pisco, A. and Papalopulu, N. (2011) 'MicroRNA-9 Reveals Regional Diversity of Neural Progenitors along the Anterior-Posterior Axis', Dev Cell 20(1): 19-32.

50 Chalmers, A. D., Strauss, B. and Papalopulu, N. (2003) 'Oriented cell divisions asymmetrically segregate aPKC and generate cell fate diversity in the early Xenopus embryo', Development 130(12): 2657-68.

Choi, W.-Y., Giraldez, A. J. and Schier, A. F. (2007) 'Target protectors reveal dampening and balancing of Nodal agonist and antagonist by miR-430', Science 318(5848): 271-4.

Davis, R. L., Turner, D. L., Evans, L. M. and Kirschner, M. W. (2001) 'Molecular targets of vertebrate segmentation: two mechanisms control segmental expression of Xenopus hairy2 during somite formation', Dev Cell 1(4): 553-65.

Delaloy, C., Liu, L., Lee, J.-A., Su, H., Shen, F., Yang, G.-Y., Young, W. L., Ivey, K. N. and Gao, F.-B. (2010) MicroRNA-9 coordinates proliferation and migration of human embryonic stem cell-derived neural progenitors Cell Stem Cell, vol. 6.

Gaiano, N. and Fishell, G. (2002) 'The role of notch in promoting glial and neural stem cell fates', Annu Rev Neurosci 25: 471-90.

Giraldez, A. J., Mishima, Y., Rihel, J., Grocock, R. J., Van Dongen, S., Inoue, K., Enright, A. J. and Schier, A. F. (2006) 'Zebrafish MiR-430 promotes deadenylation and clearance of maternal mRNAs', Science 312(5770): 75-9.

Guo, H., Ingolia, N. T., Weissman, J. S. and Bartel, D. P. (2010) 'Mammalian microRNAs predominantly act to decrease target mRNA levels', Nature 466(7308): 835-40.

Hirata, H., Yoshiura, S., Ohtsuka, T., Bessho, Y., Harada, T., Yoshikawa, K. and Kageyama, R. (2002) 'Oscillatory expression of the bHLH factor Hes1 regulated by a negative feedback loop', Science 298(5594): 840-3.

Ishibashi, M., Moriyoshi, K., Sasai, Y., Shiota, K., Nakanishi, S. and Kageyama, R. (1994) 'Persistent expression of helix-loop-helix factor HES-1 prevents mammalian neural differentiation in the central nervous system', EMBO J 13(8): 1799-805.

Laneve, P., Gioia, U., Andriotto, A., Moretti, F., Bozzoni, I. and Caffarelli, E. (2010) 'A minicircuitry involving REST and CREB controls miR-9-2 expression during human neuronal differentiation', Nucleic acids research.

Leucht, C., Stigloher, C., Wizenmann, A., Klafke, R., Folchert, A. and Bally-Cuif, L. (2008) 'MicroRNA-9 directs late organizer activity of the midbrain-hindbrain boundary', Nat Neurosci 11(6): 641-8.

Lim, L. P., Lau, N. C., Garrett-Engele, P., Grimson, A., Schelter, J. M., Castle, J., Bartel, D. P., Linsley, P. S. and Johnson, J. M. (2005) 'Microarray analysis shows that some microRNAs downregulate large numbers of target mRNAs', Nature 433(7027): 769-73.

Miller, F. D. and Gauthier, A. S. (2007) 'Timing is everything: making neurons versus glia in the developing cortex', Neuron 54(3): 357-69.

51 Ohtsuka, T., Ishibashi, M., Gradwohl, G., Nakanishi, S., Guillemot, F. and Kageyama, R. (1999) 'Hes1 and Hes5 as notch effectors in mammalian neuronal differentiation', EMBO J 18(8): 2196-207.

Pena, J., Sohn-Lee, C., Rouhanifard, S., Ludwig, J., Hafner, M., Mihailovic, A., Lim, C., Holoch, D., Berninger, P., Zavolan, M. et al. (2009) 'miRNA in situ hybridization in formaldehyde and EDC-fixed tissues', Nat Methods.

Regad, T., Roth, M., Bredenkamp, N., Illing, N. and Papalopulu, N. (2007) 'The neural progenitor-specifying activity of FoxG1 is antagonistically regulated by CKI and FGF', Nat Cell Biol 9(5): 531-40.

Selkoe, D. and Kopan, R. (2003) 'Notch and Presenilin: regulated intramembrane proteolysis links development and degeneration', Annu Rev Neurosci 26: 565-97.

Shibata, M., Nakao, H., Kiyonari, H., Abe, T. and Aizawa, S. (2011) 'MicroRNA-9 regulates neurogenesis in mouse telencephalon by targeting multiple transcription factors', J Neurosci 31(9): 3407-22.

Shimojo, H., Ohtsuka, T. and Kageyama, R. (2008) 'Oscillations in notch signaling regulate maintenance of neural progenitors', Neuron 58(1): 52-64.

Temple, S. (2001) 'The development of neural stem cells', Nature 414(6859): 112- 7.

Wienholds, E., Kloosterman, W. P., Miska, E., Alvarez-Saavedra, E., Berezikov, E., de Bruijn, E., Horvitz, H. R., Kauppinen, S. and Plasterk, R. H. (2005) 'MicroRNA expression in zebrafish embryonic development', Science 309(5732): 310-1.

Yoshiura, S., Yoshiura, S., Ohtsuka, T., Ohtsuka, T., Takenaka, Y., Takenaka, Y., Nagahara, H., Nagahara, H., Yoshikawa, K., Yoshikawa, K. et al. (2007) 'Ultradian oscillations of Stat, Smad, and Hes1 expression in response to serum', Proc Natl Acad Sci USA 104(27): 11292.

Zhao, C., Sun, G., Li, S. and Shi, Y. (2009) 'A feedback regulatory loop involving microRNA-9 and nuclear receptor TLX in neural stem cell fate determination', Nat Struct Mol Biol 16(4): 365-71.

Figure Legends:

Figure 1. miR-9 regulates mouse Hes1 expression at the RNA level (A) HeLa cells were transfected with a luciferase wild-type (WT) or miR-9 binding site deleted (Mut) Hes1 3’UTR reporter together with either scrambled (Control) or miR-9 precursors (miR-9). Luciferase expression was normalized and expressed

52 relative to the control levels. (B) Neural progenitor cells (c17.2) were transfected with the Hes1 3’UTR WT/Mut constructs and expression was measured after 48h. Luciferase activity was normalized and expressed relative to the WT reporter levels. (C) Design of Hes1 target protector – Hes1 TP (LNA- based) directed against hes1 miR-9 binding site. Seed region is boxed in red. (D) c17.2 cells were transfected with either control LNA 50nM (Control), miR-9 LNA or Hes1 TP and mature miR-9 levels were measured using qRT-PCR after 48h. (E) c17.2 cells were transfected with the corresponding constructs and synchronized for 24h by shifting to 0.2% serum media. Hes1 protein levels were examined 75min after serum addition to 10% using Western Blot. Gapdh was used as a loading control (F). c17.2 cells were transfected with either control LNA, miR-9 LNA or Hes1 TP and Hes1 mRNA levels were measured by qRT-PCR after 48h. (G) Hes1 mRNA degradation rate was measured after addition of actinomycin D at t=0. Relative Hes1 expression was plotted on a log scale and non-linear regression using exponential fit was used to determine the rate of degradation constant - k according to the equation y=Ae-k . (H) Half-life (in min) of Hes1 mRNA in c17.2 cells transfected with miR-9 precursor mimics (Pre-miR-9), miR-9 LNA inhibitor or

Hes1 TP. Half-life was determined by the formula t1/2=ln(2)/k

Figure 2. miR-9 modulates Hes1 oscillations

(A) Schematic model of Hes1 oscillator – Hes1 gene is transcribed into mRNA which is exported into the cytoplasm after processing. There it is translated into Hes1 protein, which imported back into the nucleus and inhibits its own transcription forming a negative feedback loop. In addition, both Hes1 mRNA and Hes1 protein are highly unstable and are degraded with a half-life of around 23min. (B). Bioluminescence imaging of single c17.2 cells expressing pHes1Pr-ubqluc- 3UTR and 3’UTR Mut (miR-9 binding site deleted) imaging vector.

(C) Distribution of the number of cells displaying different number of Hes1 oscillations over 20h period (WT n=34 cells; Mut n=31 cells) (D) Average number of oscillation cycles over 20h period (E) Average period and amplitude of Hes1Pr- ubqluc-3’UTR WT/Mut. Only cells displaying 2 or more cycles were considered.

53 Hes1 expression was induced by addition of fetal calf serum (FCS) to a 10% final concentration in serum-synchronized c17.2 cells, transfected with either scrambled/miR-9 precursors (F) or control LNA, miR-9 LNA, Hes1 TP oligos (G) RNA was extracted every 30min for 6 hours and Hes1 relative RNA fold change (normalized for GAPDH levels) was plotted vs time.

Figure 3. miR-9 and Hes1 expression are inversely related in vivo

(A) Fluorescent in situ hybridization (FISH) for mature miR-9 was combined with immunofluorescence for acetylated tubulin on E12 mouse telencephalon cryosections. (B) miR-9 (red) and Hes1 (green) are expressed in the VZ of the developing neocortex of a E11 mouse telencephalon. (C) Double FISH for mature miR-9 (red) and Hes1 mRNA (green) and in neural progenitors. mature miR-9 levels (D) and Hes1 protein levels (E) in cells grown at high (10%) or low (0.2%) serum conditions

Figure 4. Hes1 regulates miR-9 transcription, forming a double negative feedback loop

(A) Bioinformatic prediction for the presence of putative Hes1 binding elements in the promoters of the three primary miR-9 transcripts (B) qRT-PCR for relative levels of the primary miR-9 transcripts in c17.2 cells transfected with control pCS2 vector and pCS2Hes1 vector. (C) Relative expression of the three primary miR-9 transcripts in wild-type c17.2 cells grown in proliferative conditions. (D) Firefly luciferase reporter under the control of miR-9-2 promoter was transfected in c17.2 cells together with increasing concentrations of pCS2-Hes1. Luciferase activity was normalized and expressed relative to the control levels. (E) qRT-PCR measuring the relative fold change of Hes1 mRNA, pri-miR-9-1 and pri-miR-9-2 upon serum stimulation in synchronized c17.2 cells.

(F) qRT-PCR for relative expression of mature miR-9 in c17.2 cells transfected with control pCS2 vector and pCS2-Hes1 vector. (G) Stability of the mature miR-9 and Hes1 mRNA was followed using qRT-PCR for 3 hours after blocking transcription with Actinomycin D. (H) Mature miR-9 fold change upon serum stimulation in

54 synchronized c17.2 cells was examined using qRT-PCR (I) qRT-PCR for mature miR-9 levels in c17.2 cells grown in proliferating conditions (10% serum) 24, 48 and 72h after plating.

Figure 5. Model for the role of miR-9 regulation of Hes1 oscillations in neural development

During the proliferative mode of division of neural progenitors Hes1 oscillates with a 2-3h periodicity, which causes oscillations of pri-miR-9 in the opposite phase (time1). These bursts of miR-9 transcription are processes to the mature form of miR-9, which, due to its high stability, gradually accumulates with each Hes1 cycle. At a certain point (time 2), miR-9 concentration reaches a critical threshold, where it dampens Hes1 oscillations and enhances neural differentiation. During differentiation, Hes1 levels remain low which leads to further increase in miR-9 levels (time 3).

Figure 1.

A. B. C. 3’ ACACTAC-GGTTTCTACA -5’ Hes1 TP 5’ ATTTCTTTTTTTA--TGTGATG-CCAAAGATGTTTGAAAAT -3’ Hes1 3’UTR Control miR-9 3’AGTATGTCGAT CTATTGGTTTCT-5’ miR-9 120 *** 2.5

100 D. 2.0 1.5 ** 80 1.5 60 1.0 1.0 40

20 0.5 0.5 Relative Luciferase Activity, % Activity, Luciferase Relative

0 Activity Luciferase Relative Hes1 3'UTR_WT Hes1 3'UTR_Mut 0.0 Hes1_Luc WT Hes1_Luc Mut miR-9 levels Relative 0.0 F. Control miR-9 LNA Hes1 TP

E. 2.0

1 0.41 0.30 1 1.32 1.12 Hes1 1.5

1.0 Gapdh Control miR-9 miR-9 Control miR-9 Hes1 0.5 25nM 50nM LNA LNA TP 0.0 Relative Hes1 mRNA expression Relative Control LNA miR-9 LNA Hes1 TP

G. y = 0.7838e-0.029x H. 40 R = 0.90382 10 y = 2.3578e-0.022x R = 0.90805 30 y = 2.8969e-0.023x 1 R = 0.82857

20 0.1

Control LNA 0.01 10 miR-9 LNA

Hes1 TP half-life, min Hes1 mRNA 0.001 Relative Hes1 mRNA expression, log Relative Hes1 mRNA 0 0 50 100 150 200 Control Pre-miR-9 miR-9 LNA Hes1 TP Time after ActD treatment, min

56 Figure 2.

n Hes1Pr-ubqluc-3’UTR WT

A. o B. Hes1Pr-ubqluc-3’UTR WT i 2.5 s s e r

p 2.0 x e

Hes1 mRNA e s

a 1.5 r e f i c

u 1.0 l

d e z i

l 0.5

Hes1 protein a m r

o 0.0 N 0 5 10 15 20 25 Time, hours V Hes1Pr-ubqluc-3’UTR Mut

V n o i Hes1Pr-ubqluc-3’UTR Mut

s 6 s e r p x e

e 4 s a r e f i c u l Hes1 gene 2 d e z i l a m r

o 0

N 0 5 10 15 20 25 C. D. E. Time, hours Parameters of Hes1 WT oscillations Hes1Pr-ubqluc-3’UTR WT p=0.008 10 in c17.2 cells 40 Hes1Pr-ubqluc-3’UTR Mut 4 ** 8

30 3 6

2 20 4

10 1 2 Number of cycles 0 0

Number of cells, % of total Number of cells, 0 0 1 2 3 4 5 6 7 WT Mut WT Mut WT Mut Number of cycles Period, h Amplitude, FC F. G.

3 25 Scr 2.5 miR-9 20 Cntr LNA

2 Hes1 TP 15 1.5

relative expression 10 A 1

0.5 Hes1 mRNA relative expresion 5 Hes1 mRN 0 0 50 100 150 200 250 300 350 0 Time after serum stimulation, min 0 50 100 150 200 250 300 350 Time after Serum Induction, min

57 Figure 3.

A. B. miR-9 miR-9 Scr Scr Acet. Acet. Tubulin Tubulin DAPI DAPI

miR-9 miR-9

miR-9 miR-9 pH3 pH3 DAPI DAPI

Hes1 Hes1

C. miR-9 Hes1 DAPI merged 10% Serum (progenitors)

D. 3 E.

2 10% Serum 0.2% Serum

1 Hes1

0 Gapdh Mature miR-9 levels, FC High Serum Low Serum

58 Figure 4.

B. 1.5 A. + pCS2 (blank) + pCS2Hes1 miR-9-1 N-box 1.0 miR-9-2 E-box 0.5 miR-9-3

Relative RNA fold change Relative RNA 0.0 C. Pri-miR-9-1 Pri-miR-9-2 Pri-miR-9-3 15 D. 1.25

10 1.00

0.75

5 0.50

0.25 Relative RNA expression Relative RNA 0 0.00 Pri-miR-9-1 Pri-miR-9-2 Pri-miR-9-3 Relative miR-9-2 reporter activity E. Hes1 concentration 5 Pri-miR-9-1 4.5 F. 4 Pri-miR-9-2 1.5 3.5 Hes1 3 2.5 1.0 2 1.5 1 0.5 0.5

Relative RNA fold change Relative RNA 0 0 50 100 150 200 0.0 Relative mature miR-9 expression Time after serum stimulation, min + pCS2 (blank) + pCS2Hes1 G. H. 10 2 1.8

1 1.6 1.4 mature miR-9 1.2 0.1 Hes1 1 0.8 0.6 0.01 0.4 0.2

Relative RNA expression, log Relative RNA 0.001 0 Relative mature miR-9 levels 0 20 40 60 80 100 120 140 160 180 200 0 50 100 150 200 250 300 350 400 450 500 Time after ActD addition, min Time after serum stimulation, min I. s

l 5 e v e l 4 9 - R i m

3 e r u t

a 2 m

e v i

t 1 a l e

R 0 24h 48h 72h Time after plating (10% serum), h

59 Figure 5.

progenitors initiate proliferative state differentiation Neuronal maturation

pri-miR-9

Hes1 critical miR-9 treshold

mature miR-9

Chapter 4. microRNA-9 regulates axon extension and branching by targeting MAP1B in mouse cortical neurons

Nature Neuroscience – manuscript under revision

61 microRNA-9 regulates axon extension and branching by targeting MAP1B in mouse cortical neurons

Dajas-Bailador F.1*, Bonev B.1, Garcez P.2, Stanley P.1, Guillemot F.2, Papalopulu N.1

1 Faculty of Life Sciences

University of Manchester

Michael Smith Building

Manchester, M13 9PT, UK.

2 Division of Molecular Neurobiology

National Institute for Medical Research

Mill Hill

London, NW7 1AA, UK.

* corresponding authors

62 Summary

The capacity of neurons to develop a long axon and multiple dendrites defines their function and underlies the flow of information in the nervous system. Local protein translation has emerged as an important modulator of axonal extension and guidance, and although microRNAs are known regulators of mRNA translation, their role in axonal development has been largely unexplored. The highly conserved microRNA-9 (miR-9) is expressed in both neuronal precursors and some post-mitotic neurons, and here, we show its presence in the axons of primary cortical neurons. Via gain and loss of function experiments, we demonstrate that miR-9 controls axonal extension and branching by regulating the levels of Map1b protein, a key functional player in microtubule stability and axon development. Following microfluidic separation of the axon and the soma, we demonstrate that miR-9 represses Map1b translation in the axon and it is a functional target for the localized BDNF-dependent control of axon extension and branching. We propose that miR-9 links regulatory signalling processes with local protein synthesis, to control Map1b protein levels and axon development in cortical neurons.

63 Results and Discussion

The extension of an axon and its branching are controlled by signalling pathways that ultimately converge on the regulation of cytoskeletal components (Barnes and Polleux, 2009; Gibson and Ma, 2011). In turn, the overall structure of the cytoskeleton depends on post-translational modifications of regulatory or structural proteins, as well as local protein synthesis and degradation in response to external cues (Yoon et al., 2009). Several mRNAs have been identified in the axon, where their transport, stability and translation efficiency affects local protein levels (Yoon et al., 2009). Extracellular stimuli can specifically regulate localized levels of individual neuronal mRNAs, thus modulating localized protein synthesis (Willis et al., 2007). Indeed, in recent years, the localized axonal translation of cytoskeletal proteins has emerged as a fundamental process able to regulate axonal growth and pathfinding (Yoon et al., 2009). Therefore, there is increased interest in identifying the mechanisms by which the local translation of axonal proteins is regulated.

MicroRNAs (miRNAs) are short, non-coding RNAs that have emerged as key post-transcriptional regulators of gene expression, with the capacity to selectively modulate protein levels by controlling mRNA stability and translation efficiency (Krol et al., 2010). In neurons, selected miRNAs, such as miR-134, miR-138 and miR-125a, are enriched at distal sites in dendrites, where they can regulate synaptic function by inhibiting local translation of specific mRNAs (Krol et al., 2010; Swanger and Bassell, 2011). It follows that microRNAs would be attractive candidates for the regulation of local translation in the axonal compartment (Natera-Naranjo et al., 2010). Indeed, it has been shown that lin-4 miRNA can act as a developmental timing switch promoting axon initiation after cell cycle withdrawal in C. elegans (Olsson-Carter and Slack, 2010). More recently, miR-9- 2/3 double mutants in mice showed, among other defects in progenitor populations and telencephalic structures, the misrouting of thalamocortical and corticofugal axons (Shibata et al., 2011). In spite of these recent studies, the role of miRNAs in axon development and, in particular in the localized control of axonal protein synthesis remains unknown.

64 The highly conserved microRNA-9 (miR-9) is expressed in both neuronal precursors and some post-mitotic neurons (Gao, 2010), and although its function in neuronal progenitors has been extensively studied, its potential role in post- mitotic differentiated neurons remains largely unexplored (Delaloy et al., 2010; Bonev et al., 2011).

In the brain of late mouse embryos (E17) we observed miR-9 expression by Fluorescent In Situ Hybridization (FISH) in the progenitor area of the third ventricle (Suppl. Fig 1a), but also in the cingulate, somatosensory and motor cortex, as well as the hippocampus and visual cortex of more posterior brain regions (Fig 1a). Lower levels of miR-9 could also be seen in the intermediate zone, which at this stage contains both migrating neurons and developing axons (Suppl. Fig 1b). These results indicate that miR-9 is present in both progenitors and differentiated neurons.

To confirm the expression of miR-9 in differentiated neurons we looked at individual cortical neurons in culture from E17 mouse embryos using FISH. We found strong miR-9 presence in the soma and dendrites of neurons (Fig 1b), as reported previously in hippocampal tissue (Siegel et al., 2009). Moreover, after co- labelling with axonal markers (acetylated tubulin, Fig 1c, and Jip1, not shown) we also established the presence of miR-9 in the axons of cortical neurons (Fig 1b-c). The specificity of the miR-9 FISH signal was confirmed when compared to a scrambled control probe and to miR-1, another miRNA previously found in mouse neurons (Bastian et al., 2011). Unlike miR-9, both the scrambled and miR-1 probes showed no signal in the axons (Suppl. Fig 1d-e). Expression of miR-9 in these cortical cultures was also established by RT-PCR, confirming its high expression when compared to the non-neuronal miR-130 (Suppl. Fig 1c).

Cortical neurons in culture undergo a series of well-established morphological changes leading to polarization, where the elongation of the axon is the most significant step. To address the potential role of miR-9 in axon development, we transfected cortical neurons with a miR-9 precursor, after 1 day in culture. Neurons still showed a normal polarized morpholology when analyzed 5 days later, but axonal length was significantly decreased (Fig 1d). By contrast,

65 transfecting neurons with a miR-9 LNA inhibitor (LNA miR-9i) produced the opposite effect (Fig 1e), promoting axonal growth and indicating that the endogenous miR-9 regulates axonal extension in cortical neurons. The use of a miR-9* inhibitor demonstrated that the complementary strand of the miR-9 duplex (miR-9*) has no contribution to the observed effect in axon development (Fig 1e).

In addition to axonal growth, neuron connectivity depends on the elaboration of branching processes (Gibson and Ma, 2011). In cultured cortical neurons, axons branch upon extension and at 5 days in vitro, the neuronal population shows a normal distribution when analyzed for the number of axonal branch points (Suppl. Fig 2a). Transfection with a miR-9 precursor augmented axonal branching, as evidenced by the significant decrease in the number of neurons with no branches and the increase in the proportion with more than 5 branch points (Fig 1f). Inhibition of miR-9 had the opposite effect, increasing the percentage of neurons with no branches and decreasing the percentage of neurons with more than 5 branches (Fig 1g). Taken together, these experiments suggest that axonal behaviour is sensitive to changes in miR-9 levels; overexpression of miR-9 decreases axon length and promotes branching while loss of function of endogenous miR-9 augments axonal growth and reduces branching (for representative images see Suppl. Fig 2b).

miRNAs are capable of targeting multiple mRNAs, and we have previously identified more than 500 potential miR-9 targets based on sequence analysis (Bonev et al., 2011). In order to identify prospective candidates, we considered those involved in localized dynamics and stability of microtubules, as this is a fundamental aspect of axon development. One of these, the microtubule associated protein 1b (Map1b) appeared as a primary option, given its role in stabilizing axonal microtubules and regulating the cross talk between microtubule and actin dynamics. Map1b is known to regulate neurite growth/branching since neurons of Map1b -/- mice show decreased axonal growth and increased branching (Gordon- Weeks and Fischer, 2000; Montenegro-Venegas et al., 2010). We found that a luciferase reporter fused to the 3’UTR of Map1b (Luc-Map1b3’UTR) was repressed by miR-9 and this was abolished when the miR-9 seed-complementary sequence in the Map1b 3’UTR was mutated (Suppl. Fig 3a). The WT and mutated Map1b 3’UTR

66 were also tagged to destabilized-GFP (dGFP) and transfected into cortical neurons. The levels of expression for the mut-Map1b 3’UTR, as analysed by dGFP fluorescence, were more than double of those obtained with wild type construct, indicative of a lack of repression by endogenous miR-9 (Suppl Fig 3b). Interestingly, this difference in dGFP levels was found in the portion of the axon that is undergoing dynamic regulation, defined here as the distal axon (see materials and methods) but not in the cell body of cortical neurons. This observation provided an indication that miR-9 repression of Map1b protein expression could be specifically localized to the distal axon. The protein levels of Map1b in cortical neurons were indeed decreased following miR-9 overexpression (Suppl Fig 3c-d), while inhibition of miR-9 produced an increase (Suppl Fig 3c). The Map1b mRNA levels, however, were not affected (Suppl. Fig 3e), demonstrating that miR-9 regulates Map1b protein levels by interfering with translation, rather than promoting mRNA degradation.

To determine whether Map1b can rescue the miR-9 overexpression phenotype, we transfected cortical neurons with a Map1b construct lacking the UTRs (Map1b-Δ3’UTR), and thus not targeted by miR-9. Map1b-Δ3’UTR completely reverted the decrease in axon length observed after miR-9 overexpression (Fig 2a). Moreover, it also prevented the increase in axonal branching caused by exogenous miR-9 (Fig 2b-c). To further confirm the role of Map1b as a target of miR-9 in axonal development we specifically blocked the miR- 9 binding site in the Map1b mRNA using a target protector (Map1b-TP). This is an oligo that selectively protects Map1b from being targeted by miR-9, without affecting endogenous levels of miR-9 or its interaction with other mRNA targets (Suppl. Fig 4a and c). Transfection of cortical neurons with the Map1b-TP produced a significant increase in axonal length, together with a decrease in branching (Fig 2d-f), confirming that miR-9 controls axonal extension and branching by regulating the levels of Map1b protein. Both miR-9 and Map1b mRNA are present in the axon, where they can be detected as bright spots (particles) with at least partial co-localization (Fig 2g). Although we can also detect miR-9 particles in the dendritic processes of cortical neurons (Fig 1b), no effect was observed in their length after miR-9 overexpression (Suppl. methods). This most likely reflects

67 the preferential localization of Map1b to the developing axon (Gordon-Weeks and Fischer, 2000; Montenegro-Venegas et al., 2010).

To further understand the role of miR-9 in the regulation of axon growth, we decided to block miR-9 function specifically in the axons of cortical neurons. For this, neurons were grown in microfluidic chambers in order to isolate cell bodies and axons (Suppl. Fig 5a). To minimise the axonal toxicity that could be associated with lipid-based transfection reagents we employed a PNA miR-9 inhibitor (PNA miR-9i) linked to a cell penetrating peptide. In advance, we confirmed that the PNA miR-9i performed in the same way as the LNA miR-9i used previously (Suppl. Fig 5 c-d). When added to the axonal side of the microfluidic device, we observed a significant increase in axon length (Fig 2h-i). This increase was not observed when the PNA miR-9i was added to the somal side of the device (Suppl. Fig 5b), demonstrating that endogenous miR-9 can control axonal development by acting locally at the level of the axon.

In order to address if miR-9 regulates axon development in the mouse cortex in vivo, we electroporated the miR-9 inhibitor or Map1b-TP together with eGFP into the cerebral cortex of day 14.5 mouse embryos, and analyzed axon extension four days later. The inhibition of endogenous miR-9 in vivo at day 14.5 of development severely disrupted neuronal migration (data not shown), in agreement with its previously described role in neuronal differentiation and migration (Delaloy et al., 2010; Bonev et al., 2011; Shibata et al., 2011). This prevented the use of the inhibitor in the analysis of axonal extension on differentiated neurons. However, the specific disruption of the mir-9/Map1B interaction by electroporation of Map1B-TP produced a significant increase in axonal length (Fig 3a-d), assessed 4 days after electroporation. The Map1b-TP shows a more pronounced affect on axonal length in vivo compared to in vitro. This probably reflects the fact that electroporated neurons in utero are primed to growth rather than branch at the particular time of the experimental analysis (E 18.5). In order to analyze the effect on branching, mice brain cortices were analyzed at P10, when electroporated neurons develop cortical contralateral branching (see methods and Fig 3e). We reconstructed the distal axons of GFP positive neurons in the contralateral cortex (Fig 3f), and found that neurons

68 electroporated with the Map1b-TP had a significant reduction in the number of branches (Fig 3f-g). These results confirm that miR-9 targets Map1b in differentiated neurons, thus affecting axon development in vivo.

Taking into consideration this novel role of miR-9 in axon development we asked whether miR-9 could be a target for regulatory signalling pathways known to regulate axonal function. The brain-derived neurotrophic factor (BDNF), a member of the neurotrophin family of polypeptide growth factors, functions as a chemoattractant, but also as a branch-promoting factor (Cohen-Cory et al., 2010; Hoshino et al., 2010). This is thought to indicate a dose and temporal dependence where initial short-term (acute) stimulation promotes growth, while long term or gradual stimulation promotes branching (Hoshino et al., 2010; Panagiotaki et. al., 2010; Ji et al., 2010). To examine the potential regulation of miR-9 function by BDNF, we tested two concentrations of BDNF at short and long-term time points and measured miR-9 levels at the distal axon compartment using FISH.

Short (2h) stimulation with BDNF produced a significant decrease in miR-9 particles in the distal axon (Fig 3h), and this was accompanied by an increase in the levels of WT-dGFP-mMap1b-3’UTR, but not the mut-dGFP-mMap1b-3’UTR reporter (Fig 3i). In effect, short BDNF stimulation, by decreasing miR-9, brought the WT-3’UTR construct on a similar expression level to dGFP construct with the mutated miR-9 site. Given the role of miR-9 as discussed above, the functional prediction of this experiment would be that short stimulation with BDNF would promote axonal growth through an increase in Map1b protein. Indeed, we observed an increase in axonal length following short BDNF stimulation but this was completely prevented when miR-9 levels where augmented by overexpression of miR-9 (Fig 3j). Significantly, the block of BDNF-dependent axonal growth that follows the increase in miR-9 level, was completely reverted by co-transfection of the Map1b-TP (Fig 3j). The action of BDNF was local to the axon since the loss in miR-9 levels and increase in axon length were observed only when BDNF was applied to the axonal side, but not the cell body side of the microfluidic chambers (Suppl. Fig 6a-d). Unlike the effect seen in the axon, short BDNF stimulation produced no decrease in miR-9 levels in the cell body of cortical neurons (not shown).

69 In order to show that the changes in miR-9 levels are specific and not due to a general down-regulation in the mRNA processing machinery in the distal axon, we analyzed the localization and expression of GW-182, a member of the RISC complex. We show here that GW-182 is detected in bright particles along the axon and it is particularly enriched in the distal axon of cortical neurons (Suppl Fig 7a). Stimulation with BDNF did not significantly change expression levels of GW-182 in the distal axon (Suppl. Fig 7b).

We then asked whether miR-9 also lies downstream of BDNF in conditions of prolonged stimulation with high BDNF (48hr). Such stimulation increased the percentage of neurons with more than 5 branch points (Fig 3m), as shown previously (Panagiotaki et al., 2010), confirming a bi-phasic effect of BDNF (initial short term stimulation promotes growth, while long term stimulation stimulates branching). The prolonged stimulation with a high concentration of BDNF (100 ng/ml, 48 h) produced a significant increase in miR-9 density in the distal axon (Fig 3k), and lead to a decrease in the levels of WT-dGFP-mMap1b-3’UTR, but not the mut-dGFP-mMap1b-3’UTR (Fig 3l). No change in miR-9 levels was found in the cell body of cortical neurons. As predicted by our functional model, the LNA miR-9i prevented the branching effect seen after prolonged stimulation with high BDNF (48 h, Fig 3m). Thus, miR-9 acts downstream of BDNF in a bi-phasic manner; in the initial phase of BDNF stimulation, a reduction in miR-9 relieves the translational repression of Map1b leading to increased levels of Map1b protein, microtubule stabilization and axonal growth. On the other hand, prolonged BDNF action increases miR-9 levels and promotes axonal branching by decreasing Map1b protein. It would be interesting to establish whether the changes in miR-9 density are due to local production, stability and/or re-localisation.

Together, the results presented indicate that the presence of miR-9 in the distal parts of axons contributes locally to dynamic axonal growth, through the regulation of an important component of the cytoskeleton. The level of miR-9 responds locally to BDNF, constituting a functional target for the signalling processes controlling axonal extension and branching.

References:

Barnes, A. P. and Polleux, F. (2009) 'Establishment of axon-dendrite polarity in developing neurons', Annu Rev Neurosci 32: 347-81.

Bastian, I., Tam Tam, S., Zhou, X. F., Kazenwadel, J., Van der Hoek, M., Michael, M. Z., Gibbins, I. and Haberberger, R. V. (2011) 'Differential expression of microRNA-1 in dorsal root ganglion neurons', Histochem Cell Biol 135(1): 37-45.

Bonev, B., Pisco, A. and Papalopulu, N. (2011) 'MicroRNA-9 Reveals Regional Diversity of Neural Progenitors along the Anterior-Posterior Axis', Dev Cell 20(1): 19-32.

Choi, W.-Y., Giraldez, A. J. and Schier, A. F. (2007) 'Target protectors reveal dampening and balancing of Nodal agonist and antagonist by miR-430', Science 318(5848): 271-4.

Cohen-Cory, S., Kidane, A. H., Shirkey, N. J. and Marshak, S. (2010) Brain- derived neurotrophic factor and the development of structural neuronal connectivity Devel Neurobio, vol. 70

Gao, F.-B. (2010) Context-dependent functions of specific microRNAs in neuronal development Neural Development, vol. 5.

Gibson, D. A. and Ma, L. (2011) 'Developmental regulation of axon branching in the vertebrate nervous system', Development 138(2): 183-95.

Gordon-Weeks, P. R. and Fischer, I. (2000) 'MAP1B expression and microtubule stability in growing and regenerating axons', Microsc Res Tech 48(2): 63-74.

Hengst, U. and Jaffrey, S. R. (2007) 'Function and translational regulation of mRNA in developing axons', Semin Cell Dev Biol 18(2): 209-15.

Hoshino, N., Vatterott, P., Egwiekhor, A. and Rochlin, M. W. (2010) 'Brain- derived neurotrophic factor attracts geniculate ganglion neurites during embryonic targeting', Dev Neurosci 32(3): 184-96.

Ji, Y., Lu, Y., Yang, F., Shen, W., Tang, T. T.-T., Feng, L., Duan, S. and Lu, B. (2010) 'Acute and gradual increases in BDNF concentration elicit distinct signaling and functions in neurons', Nat Neurosci 13(3): 302-9.

Krol, J., Loedige, I. and Filipowicz, W. (2010) The widespread regulation of microRNA biogenesis, function and decay Nature Reviews Genetics, vol. 11.

Montenegro-Venegas, C., Tortosa, E., Rosso, S., Peretti, D., Bollati, F., Bisbal, M., Jausoro, I., Avila, J., Caceres, A. and Gonzalez-Billault, C. (2010) 'MAP1B regulates axonal development by modulating Rho-GTPase Rac1 activity', Mol Biol Cell 21(20): 3518-28.

Natera-Naranjo, O., Aschrafi, A., Gioio, A. E. and Kaplan, B. B. (2010) Identification and quantitative analyses of microRNAs located in the distal axons of sympathetic neurons RNA, vol. 16.

Olsson-Carter, K. and Slack, F. J. (2010) 'A developmental timing switch promotes axon outgrowth independent of known guidance receptors', PLoS Genet 6(8).

Pacary, E., Heng, J., Azzarelli, R., Riou, P., Castro, D., Lebel-Potter, M., Parras, C., Bell, D. M., Ridley, A. J., Parsons, M. et al. (2011) 'Proneural Transcription Factors Regulate Different Steps of Cortical Neuron Migration through Rnd- Mediated Inhibition of RhoA Signaling', Neuron 69(6): 1069-84.

Panagiotaki, N., Dajas-Bailador, F., Amaya, E., Papalopulu, N. and Dorey, K. (2010) Characterisation of a new regulator of BDNF signalling, Sprouty3, involved in axonal morphogenesis in vivo Development, vol. 137.

Park, J. W., Vahidi, B., Taylor, A. M., Rhee, S. W. and Jeon, N. L. (2006) Microfluidic culture platform for neuroscience research Nat Protoc, vol. 1.

Polleux, F. and Snider, W. (2010) 'Initiating and growing an axon', Cold Spring Harb Perspect Biol 2(4): a001925.

Shibata, M., Nakao, H., Kiyonari, H., Abe, T. and Aizawa, S. (2011) 'MicroRNA-9 regulates neurogenesis in mouse telencephalon by targeting multiple transcription factors', J Neurosci 31(9): 3407-22.

Siegel, G., Obernosterer, G., Fiore, R., Oehmen, M., Bicker, S., Christensen, M., Khudayberdiev, S., Leuschner, P. F., Busch, C. J. L., Kane, C. et al. (2009) A functional screen implicates microRNA-138-dependent regulation of the depalmitoylation enzyme APT1 in dendritic spine morphogenesis Nat Cell Biol, vol. 11.

Swanger, S. A. and Bassell, G. J. (2011) 'Making and breaking synapses through local mRNA regulation', Curr Opin Genet Dev.

Willis, D. E., van Niekerk, E. A., Sasaki, Y., Mesngon, M., Merianda, T. T., Williams, G. G., Kendall, M., Smith, D. S., Bassell, G. J. and Twiss, J. L. (2007)

72 Extracellular stimuli specifically regulate localized levels of individual neuronal mRNAs J Cell Biol, vol. 178.

Xu, X.-L., Li, Y., Wang, F. and Gao, F.-B. (2008) The steady-state level of the nervous-system-specific microRNA-124a is regulated by dFMR1 in Drosophila J Neurosci, vol. 28.

Yoon, B. C., Zivraj, K. H. and Holt, C. E. (2009) Local translation and mRNA trafficking in axon pathfinding Results Probl Cell Differ, vol. 48.

Zhao, C., Sun, G., Li, S. and Shi, Y. (2009) 'A feedback regulatory loop involving microRNA-9 and nuclear receptor TLX in neural stem cell fate determination', Nat Struct Mol Biol 16(4): 365.

Figure Legends

Figure 1.

(a) LNA FISH of mature miR-9 on mouse brain sections at E17, showing expression in the cingulate (cc), somatosensory (ssc), visual (vc) cortex and hipoccampus (hc). (b) LNA FISH of miR-9 expression in the cell body, dendrites (ds) and axons (ax) of primary cortical neurons. (c) Co-labelling of miR-9 (green) and acetylated-tubulin protein (red), as a marker of axonal projections. (d-e) miR-9 regulation of axonal length, and (f-g) miR-9 regulation of axonal branching in cortical neurons after transfection of precursor miR-9 (d, f) or LNA miR-9i (e, g). Scrambled precursor miR-9 (scr), LNA control and miR-9*i serve as controls. All data groups are expressed as the mean ± SEM; (*) indicates statistical significance with respect to controls (p < 0.05; Mann-Whitney Rank Sum test for d and e and Student’s t test for f and g).

Figure 2

(a-c) Map1b-Δ3’UTR overexpression rescues the decrease in axonal growth and increase in axon branching observed after transfection of the miR-9 precursor. (d- f) Blocking the targeting of Map1b by miR-9 with a target protector (Map1b-TP) increases axonal length and decreases branching. (g) Detection of miR-9 and Map1b mRNA in the axons of cortical neurons with FISH. (h-i) Representative images and quantification of microfluidic chamber devices showing the effect of local axonal inhibition of miR-9, using a PNA miR-9i, on axonal length when it is specifically applied on the axonal side. PNA-control serves as a negative control. All data groups are expressed as the mean ± SEM; (*) indicates statistical significance with respect to controls, unless indicated by brackets (p < 0.05; Mann-Whitney Rank Sum test for a and d and Student’s t test for b-c and e-f).

74 Figure 3

(a-c) Sections of electroporated brains of E14.5 mouse embryos, fixed at E18.5. (b) Axonal processes from neurons electroporated with eGFP (control) or eGFP plus Map1b-TP can be seen extending medially towards the midline, with Map1b-TP significantly increasing axonal length. Axons were measured by taking the edge of the GFP labelled region as the starting point (dotted line). The arrow indicates the end of the axonal tract in each condition. (c) High magnification of the axonal tract for control and Map1b-TP, with the arrow indicating the longest axons for each condition. (d) Quantification of axonal length in vivo. Data is expressed as the mean ± SEM; (*) indicates statistical significance with respect to control, (p < 0.05; Student’s t test). (e) Schematic diagram of the axonal extension and contralateral cortical branching of electroporated neurons at P10. (f) Representative traces of axon terminals in the contralateral cortical layers of transfected brains. Axons (black) and primary branches (red) are shown. (g) Quantification of contralateral axonal branching. Data shows number of primary branches per axon; mean ± SEM, (*) indicates statistical significance with respect to control, (p < 0.05; Student’s t test). (h-i) Effect of short stimulation with BDNF (2 h) on miR-9 particle density and levels of dGFP-mMap1b-3’UTR, either wild type or miR-9 binding site mutant (miR-9 mut), in the distal axon of cortical neurons. (j) The BDNF-dependent increase in axonal length is prevented by miR-9 precursor transfection and rescued by the Map1b-TP. (k-l) Effect of prolonged stimulation with BDNF (48 h), on miR-9 particle density and dGFP-mMap1b-3’UTR levels in the distal axon of cortical neurons. (m) Inhibition of miR-9 prevents the increase in axonal branching after long-term stimulation (48 h) with BDNF. All data groups for in vitro experiments (h-m) are expressed as the mean ± SEM; (*) indicates statistical significance with respect to controls (p < 0.05; Mann-Whitney Rank Sum test for j and Student’s t test for h-i and k-m).

75 Figure 1.

a b c ds vc ds ssc hc ds

500 μm 10 μm d e 140 140 * 120 120

100 100

80 * * 80 60 60

40 40

20 20 axon length (% of control) of (% length axon axon length (% of control) of (% length axon

0 0 control scrambled 25 nM 50 nM control LNA LNA LNA control miR-9i miR-9*i miR-9

f miR 9g LNA miR-9i 0 branch points >5 branch points 0 branch points >5 branch points 14 30 25 14 12 * * 25 20 12 10 20 10 8 15 * 8 15 6 * 10 6 10 4 4 5 2 5 2 0 0 0 number cell % of total 0 % of total cell number cell % of total control scr miR-9 control scr miR-9 control LNA LNA control LNA LNA control miR-9i control miR-9i

Figure 1 76 Figure 2.

a b c * 0 branch points >5 branch points

120 * * 50 * 30 * 100 25 40

80 20 30 60 * 15 20 40 10

20 10 5 % of total cell number cell % of total % of total cell number cell % of total axon length (% of control) of (% length axon 0 0 0 control miR-9 Map1b- controlmiR-9 miR-9 control miR-9 miR-9 Δ3’UTR Map1b-Δ3’UTR Map1b-Δ3’UTR miR-9 d ef 0 branch points >5 branch points

120 * * 30 * 16 14 100 25 12 80 20 * 10 60 15 8 * 6 40 10 4 20 5 2 % of total cell number cell % of total % of total cell number cell % of total axon length (% of control) axon 0 0 0 control Map1b-TP Map1b-TP controlMap1b-TP Map1b-TP control Map1b-TP Map1b-TP 25 nM 50 nM 25 nM 50 nM 25 nM 50 nM g h miR-9 (green)/Map1b (red) FISH somal axonal i

160 140 * 120

PNA-control 100

Map1b 60

20 axon length (% of control) of (% length axon miR-9

PNA-miR-9i 0 control PNA PNA control miR-9i 0.1 μM

Figure 2

77 Figure 3.

a b Control Map1b-TP d 200 * 150 Control

500 μm 100

50 axon length (% of control) of (% length axon Map1b-TP 0 100 μm 100 μm 500 μm control Map1b-TP c Control e g f Control 3 I

II-III

IV 50 μm 2 Map1b-TP V Map1b-TP I 1 II-III branches/axon *

50 μm IV

V 0 control Map1b-TP

i h DISTAL AXON DISTAL AXON j

120 240 * 140 * 100 200 * * 120

80 160 100

* 80 60 120

60 40 80

dGFP (% of control) of (% dGFP 40

20 40 miR-9 densitymiR-9 control) of (%

axon length (% of control) of (% length axon 20

0 0 0 c c control c 10 100 BDNF BDNF Map1b-TP 2h 2h miR-9 1nM BDNF (ng/ml) wild type miR-9 mut 2h BDNF 10ng/ml 2h dGFP-Map1b 3’ UTR k l m DISTAL AXON DISTAL AXON 140 * 240 >5 branch points

120 200 30 * * 100 25 160

80 20 120 60 15 * 80 10

40 control) of (% dGFP % of total cell number cell % of total miR-9 densitymiR-9 control) of (% 40 20 5

0 0 0 c 10 100 c BDNF c BDNF control LNA miR-9i BDNF (ng/ml) 48h 48h wild type miR-9 mut BDNF 100ng/ml 48h 48h dGFP-Map1b 3’ UTR

Figure 3 78 a b c

200 +m cp 150 iz ession r 100 e RNA exp

v 50 100 +m

cp Relati

0 U6 miR-9 miR-130

d Tubulin (red)+miR-1 (green) Tubulin (protein) miR-1 RNA

e Tubulin (red)+scr (green) Tubulin (protein) scr RNA

Supplementary Figure 1. FISH of miR-9 on mouse brain sections at E17, showing expression in the (a) progenitor layer of the third ventricle and (b) iz: intermediate zone + white matter; cp: cortical plate (c) RT-PCR analysis of miR-9 and miR-130 levels in cortical neurons in culture. Data are shown as percent of U6 RNA expression. (d-e) FISH of miR-1 and scrambled (scr) probes together with tubulin protein immu- nolabelling. Neither miR-1 nor scr probes show localisation in neuronal projections.

79 a b control Distal axon Cell body 120 miR-9 240 * 100 200

80 160

60 * 120

40 80

20 40 dGFP (% of control) of (% dGFP

Relative Luc. Activity (%) Luc. Relative 0 0 WT mut WT mut WT mut Luc-Map1b 3’UTR dGFP-Map1b 3’ UTR c OD (%) 100 94 64 98 124

Map1b

GAPDH control scrambled LNA miR-9i miR-9 25 nM miR-9 50 nM d GFP + Map1b GFP Map1b (protein)

120

100

scrambled 80

10 μm 10 μm 10 μm 60 *

Map1b control) of (% 20 miR-9

0 10 μm 10 μm 10 μm scr miR-9 50 nM

e 1.6

1.4

1.2

1.0

0.8

0.6

0.4

0.2

Relative Map1b RNA levels Relative 0.0 control scr miR-9 miR-9 control LNA 50 nM 100 nM LNA miR-9i 50 nM

Supplementary Figure 3. Map1b is a target for miR-9 in cortical neurons. (a) Expression of luciferase reporter (Luc-Map1b 3’UTR), wild type (WT) or with mutated (mut) miR-9 binding site, together with control or miR-9 precursor (b) Localized levels of expression of the same WT and mutated constructs but tagged to dGFP and overexpressed in cortical neurons. (c) Western blot of endogenous Map1b or GAPDH proteins from cortical neurons after transfection with miR-9 precursor or miR-9 LNA inhibitor. (d) Immuno!uorescent detection and quanti"cation of endogenous Map1b protein (red) in GFP positive cells, co-transfected with either scrambled or miR-9 precursor in cortical neurons. Arrowheads indicate the axon of a transfected neuron, note downregulation of Map1b signal when miR-9 precursor is transfected. (e) RT-PCR of Map1b RNA levels in cortical neurons after transfection with miR-9 precursor and LNA miR-9 inhibitor. Data is shown as fold change with respect to controls. For all graphs data are expressed as mean ± SEM. Statistical signi"cance compared to respective controls was determined with Student’s t test (* p < 0.05)

80 a b

1.5 10000 453* 2 ol) ol) r r ont ont c

c 1000

119* .9 1.0 100 els (FC of els (FC of v v 11*.95 10 0.5 1 e miR-9 le e miR-9 le

v 1 v * Relati Relati 0.0 0.1 Ctrl LNA LNA miR-9i Map1b-TP scr 1 nM 10 nM 30 nM 50 nM 50 nM miR-9

c control 3 ol)

r LNA miR-9i (50 nM) * ont c Map1b-TP (50 nM)

2 *

ession (FC of * r * * erase Exp

f 1 uci L e v Relati 0 Luc-Map1b 3’UTR Luc-Hes1 3’UTR Luc-TLX 3’UTR Luc-Onecut1 3’UTR

Supplementary Figure 4. Relative miR-9 levels after transfection with the (a) LNA miR-9i and Map1b- TP or with (b) miR-9 precursor in c17.2 cells. Data was normalized to snRNA U6 and expressed as fold change with respect to the control. Inhibition of miR-9 decreases mature miR-9 levels, as has been previously observed by Dellaloy et al., (2010), while the Map1b-TP has no effect. (c) Relative luciferase expression leveles in differentiated c17.2 cells co-transfected with either LNA miR-9i or Map1b-TP and respective luciferase reporters (Map1b, Hes1, TLX, Onecut1). Inhibition of miR-9 increases the expression levels of all targets, while the Map1b-TP only has an effect on the Map1b 3’UTR reporter. Data was expressed as fold change compared to the control. For all graphs data are shown as mean ± SEM. Statistical significance compared to respective controls was determined with Student’s t test (* p < 0.05).

81 a b soma side axonal side

120

100 ol) r ont

c 80

150 uM on length (% of

x 20 a

0 control PNA miR-9i 0.1 +M (cell body side) c d 140 1.4 y t 120 * * ol) 1.2 * * tivi r c A

ont 100 1.0 c ol) erase r 80 f 0.8 uci ont L c

60 e

v 0.6 (FC of 40 0.4 on length (% of x a 20 0.2 ap1b Relati M

0 0.0 control 0.05 +M 0.1 +M 0.5 +M control LNA control PNA miR-9i miR-9i PNA miR-9i 50 nM 0.1 +M

Supplementary Figure 5. (a) Schematic representation and representative image of a microfluidic chamber device, with its fluidically isolated axonal and somal compartments (red: axonal acetylated tubulin, green: somal otx-1). (b) Measurement of axonal length on the axon side after incubation with the PNA miR-9i only in the somatic side of the device. No effect is observed. (c) Measurement of axonal length in normal cortical cultures after incubation with the PNA miR-9i. Data for axonal length is shown as mean ± SEM. Statistical significance was determined using Mann-Whitney Rank Sum test for comparison with respective controls. (*p<0.05). (d) Increase in relative Luc-Map1b 3’UTR activity in c17.2 cells following incubation with LNA miR-9i or PNA-miR9-i. Data are expressed as fold change to controls (mean ± SEM). Statistical significance was determined using Student’s t test for comparison with respective controls. (*p<0.05).

82 BDNF: Axon side BDNF: Cell body side a b c d

DISTAL AXON 120 120 DISTAL AXON 160 120 * 140 100 100

100 ol) ol) ol) r ol) r r r 120 ont ont ont ont c 80 c 80 c 80 c 100 * 60 60

60 80 y (% of y (% of t t

60 40 40 40 on length (% of

on length (% of 40 x x a a 20 20 20 miR-9 densi miR-9 densi 20

0 0 0 0 control BDNF 2h control BDNF 2h control BDNF 2h control BDNF 2h 10 ng/ml 10 ng/ml 10 ng/ml 10 ng/ml

Supplementary Figure 6. (a-d) Quantification of miR-9 density and axonal length as measured on the axonal side of the microfluidic device, with BDNF applied only to the (a-b) axon or (c-d) cell body side of the chambers. Data is shown as mean ± SEM. For axonal length experiments statistical significance was determined using Mann-Whitney Rank Sum test for comparison with respective controls. (*p<0.05). For miR-9 density comparisons statistical significance was determined using Student’s t test for comparison with respective controls. (*p<0.05).

83 a b cell body distal axon 140

120

100

uorescence (% of control) of (% uorescence ! GW-182 20 10 μm

Ac. Tubulin (red)/GW182 (green)/DAPI (blue) (green)/DAPI (red)/GW182 Tubulin Ac. 0 control BDNF 2h 10 ng/ml c miR9*

DISTAL AXON 120

100

miR-9* densitymiR-9* control) of (% 20

0 control BDNF 2h 10 ng/ml

Supplementary Figure 7. (a) Protein localization of the RISC complex component GW-182 to the cell body and distal axon in cortical neurons. (b-c) Quantification of GW-182 and complementary miR-9* strand fluorescence in the distal axon after stimulation with BDNF. No changes were observed, indicating that the effect of BDNF is not due to a non-specific modification of miRNA levels or miRNA processing machinery. Data is shown as mean ± SEM

84 Supplementary Information

microRNA-9 regulates axon extension and branching by targeting MAP1B in mouse cortical neurons

Dajas-Bailador F., Bonev B., Garcez P., Stanley P., Guillemot F., Papalopulu N.

Materials and Methods

Animals

Mice were housed, bred and treated according to the guidelines approved by the UK Home Office under the Animal (Scientific Procedures) Act 1986.

Cell culture and Luciferase Reporter Assay

Primary cortical neurons were prepared as described previously (Dajas- Bailador et. al. 2008). Briefly, mouse cortices were dissected from C57/BL6 embryonic day 17 mice, carefully triturated and the cell suspension plated onto poly-L-ornithine coated culture plates. Neurons were cultured in Neurobasal medium (Invitrogen, supplemented with 5mM glutamax and 2 % B-27 supplement) for a maximum of 8 days. Neuronal transfections were performed 24 h after dissection using Lipofectamine 2000 (Invitrogen) for DNA, plus or minus RNA transfections. HeLa and C17 cells were maintained in DMEM supplemented with 10% serum and 100units/ml penicillin/streptomycin, and transfected using Lipofectamine 2000.

For luciferase reporter assays, HeLa cells were seeded at a density of 104 cells/well in a 96-well plate and transfected after 24 hr with 25 ng of the reporter and either 30 nM of scrambled or mir-9 precursors (Ambion). Luciferase expression was analyzed after 48h using Dual Luciferase Assay system (Promega).

85 Renilla luciferase activity was normalized by the co-expressed Firefly luciferase and expressed as a percentage of the control. All assays were repeated at least three times and performed in triplicate each time.

Compartmentalized culture in microfluidic chambers

Primary cortical neurons (E17) were cultured in microfluidic devices with 150-um long microgrooves (Xona Microfluidics; Park et. al., 2006). The chambers allow for fluidic isolation of the axonal from the cell body (somal) compartment. Microfluidic devices were assembled on poly-L-ornithine coated coverslips, and neurons plated on the cell body side were cultured for 5 days. Culture media was replenished every 48 h to ensure fluidic isolation at all times. After 5 days in culture, only axons could reach across to the other side of the chambers as the average dendrite length is < 50 um. Fluidic isolation was checked by the introduction of a fluorescent probe to the axonal side of the device, with no fluorescence detected in the cell body side (not shown). The cell permeable PNA miR-9i was applied 48 hours after neuron seeding, while microfluidic devices were removed and coverslips fixed 4 days later. The PNA miR-9i or BDNF were applied to either the cell body or axonal side of the chambers and fluidic isolation was maintained accordingly. Immunofluorescence and FISH analysis were carried out as described elsewhere. The length of the axons was established from the site where they enter the axonal side of chambers (see dotted line in Fig 2h) up to the growth cone. The length of the primary axonal projection (labelled with acetylated tubulin) for each neuron was measured using Image J and data are expressed as percent of respective controls (an average of 40 axons measured from each chamber; 6 chambers in total for each condition from 3 independent experiments). Density of miR-9 particles was established in the distal portion of the axon, as described elsewhere.

DNA Constructs and oligos

Map1b cDNA was kindly donated by Prof. Gordon-Weeks (King’s College, London). For the generation of luciferase reporter constructs, the 3’ UTR of

86 predicted miR-9 targets was PCR amplified from mouse brain cDNA. In the case of Map1b, a region of the Map1b 3' UTR (1094 nucleotides in length, including the predicted miR-9 binding site) was amplified using P1 (ctcgagTGTCATGGTGATGCAAGTCA) and P2 (gcggccgcAAAGAAGCCATTCCGGTCTT) and cloned downstream of Renilla luciferase coding sequence in the psi-CHECK-2 vector (Promega). For the destabilized-GFP (dGFP) constructs, the 3’ UTR of Map1b was cloned downstream of pCS2-d2eGFP (Clontech). To generate the mutant versions of these constructs we deleted 4 nucleotides in the seed complementary region (CCAAAG) of the Map1b 3’UTR.

The LNA control (scramble seq: GTGTAACACGTCTATACGCCCA; Cat N: 199002), LNA miR-9 inhibitor (seq: TCATACAGCTAGATAACCAAAG; Cat N: 410014), and custom designed Map1b-TP (seq: GCTTTGGTATTTGCTTCA) were from Exiqon. The scrambled miRNA precursor (cat N: AM 17110) and miR-9 precursor (cat N: AM17100/ID: PM10022) were from Life Technologies. The PNA control (seq: CTCCCTTCAATC; Cat N: PN-1001) and PNA miR-9 inhibitor (seq: GCTAGATAACCAAAG; Cat N: PI-1022) were from Cambridge Research Biochemicals.

In situ hybridization

Primary mouse cortical neurons (day 5-6 in vitro) cultured in poly-L- ornithine coated cover slips were fixed in 4% paraformaldehyde. Following fixation, the cells were washed in PBS and acetylated with 0.5% acetic anhydride in tri-ethanol-amine (TEA) buffer for 10 minutes.

After washing twice in PBS (5 minutes), neurons were permeabilised with 0.2% triton X100 in PBS for 5 minutes to allow more efficient penetration of the riboprobe during the hybridisation step. A one-hour incubation with EDC (Sigma, cat #39391) was used for additional fixation (Lu et. al. 2009). The cells were then washed in TBS twice for 10 minutes (the imidazole buffer also removes residual phosphates from the cells therefore all subsequent washes were performed using TBS) prior to the addition of the riboprobes in hybridisation buffer (DEPC H20 with 25 % deionised formamide, 4X SSC, 0.05 M EDTA, yeast tRNA 0.5mg/ml, 10 % w/v dextran sulphate, 1X denhardts solution, 10 U/ml of RNAse inhibitor). The

87 probes were diluted in hybridisation buffer to give a concentration of 25nM for the commercially available DIG labelled miR-9 probes (Exiqon) or 100ng/50ul for the MAP1B Fluorescein labelled riboprobes (Roche kit). The cover slips were laid over a 50ul drop of probe solution on parafilm and incubated over night at 55oC temperature in a sealed, humidified chamber.

Following hybridisation, the cover slips were removed from the chamber and washed in 2X SSC at room temperature for 10 minutes and then washed again in 2X SSC at 50oC below the hybridisation temperature for 10 minutes to remove excess probe. An increased stringency wash was then performed with 1X SSC at the same temperature. Cells were then incubated in a 3% solution of hydrogen peroxide in TBS for one hour, to quench endogenous peroxidase, followed by three 5-minute washes in TBS. The cover slips were then blocked in a 0.5% milk/TBS solution for one hour followed by over night incubation with the primary antibodies at the recommended dilutions in 0.5% milk/TBS. For the miR-9 probe, P.O.D. anti DIG antibody (Roche, cat# 11207733910, 1:1000) was used and for the Map1b mRNA, a P.O.D. anti FITC (Roche, 11426346910). The Map1b protein was detected with a primary goat anti-Map1b from Santa Cruz (sc-8970).

The cover slips were washed in TBS three times for 5 minutes and the signal detected using the tyramide signal amplification kit (Perkin Elmer). When Map1B protein was also detected, the tyramide reaction was followed by 3 x PBS wash and a blocking step with 0.5% milk/TBS. Secondary antibody incubation was carried out for 2 hours at room temperature. Excess secondary antibody was washed off with TBS for 5 minutes (three times). Following the relevant detection protocols, the cover slips were mounted on glass microscope slides with Vecta Shield mount including DAPI and analysed under a fluorescent microscope. All solutions and buffers in this protocol were prepared in RNAse free conditions. All washes were performed at room temperature unless otherwise stated.

RNA isolation, RT-PCR, and quantitative Real-Time PCR analysis

Total RNA was extracted from either whole mouse brain or cortical primary cultures using TRizol (Invitrogen) and retrotranscribed using RT-AMV (Invitrogen) according to the manufacturer’s instructions. Mature miR-9 levels

88 were assessed using taqman microRNA assay (Applied Biosystems) and the expression was normalized for the snRNA U6. Map1b mRNA expression was also measured using taqman probes and normalized for GAPDH. Real-time PCR was performed in an ABI StepOne Plus Sequence Detection System (Applied Biosystems) using TaqMan Fast Real-Time PCR Master Mix.

Immunoblotting

Samples were resolved by SDS-PAGE (10% gels) and transferred to Immobilon-P membranes (Millipore Inc.), which were immunoblotted with the following antibodies: anti-Map1b (Santa Cruz, sc-8970), anti-GAPDH (Applied Byosistems, cat# AM4300). Immunocomplexes were detected using HRP- conjugated secondary antibodies followed by enhanced chemiluminescence (Pierce). Results were quantified using Intelligent Quantifier software (Bio Image Systems).

Immunofluorescence

Cortical neurons were fixed using 4 % paraformaldheyde and transfected neurons detected by expression of green fluorescent protein. Map1b protein was detected using anti-Map1b (1:100, Santa Cruz, sc-8970). Acetylated Tubulin and GW-182 were obtained from Sigma and Santa Cruz Biotechnology respectively. Secondary antibodies were Alexa 488- or 594-labelled goat anti-mouse or anti- rabbit (Invitrogen). Images were obtained using a Nikon Eclipse 80i microscope attached to a DS-Qi1MC camera (Nikon).

In utero Electroporation and Tissue Processing

In utero electroporation was performed as described previously (Pacary et. al., 2011) pClG2-GFP plasmid (final concentration 1μg/μl) was co-injected with LNA-control, LNA miR-9 inhibitor or Map1b target protector (all at 50 nM), mixed with 0.05% Fast Green (Sigma) through the uterine wall into the telencephalic vesicle using pulled borosilicate needles (Harvard Apparatus) and a Femtojet

89 microinjector (Eppendorf). Five electrical pulses were applied at 30V (50 msec duration) across the uterine wall at 1 sec intervals using 5 mm platinum electrodes (Tweezertrode 45-0489, BTX, Harvard Apparatus) connected to an electroporator (ECM830, BTX). The uterine horns were then replaced in the abdominal cavity and the abdomen wall and skin were sutured using surgical needle and thread. The pregnant mouse was warmed on heating pad until it woke up. The whole procedure was complete within 30 min. Four days following the surgery, pregnant mice were sacrificed by neck dislocation and embryos were processed for tissue analyses. Embryonic brains were fixed in 4% PFA overnight and then transferred into liquid 3% low melting agarose (Sigma) and incubated on ice for 10 minutes. Embedded brains were cut coronally (150 μm) with a vibratome (Leica).

Sections were mounted with Antifade reagent (Invitrogen). The intermediate zone was identified based on cell density and visualized with DAPI nuclear staining (Invitrogen). All images were acquired with a laser scanning confocal microscope (LSM 710, Zeiss).

Image Data Analysis and Statistics

For the measurement of axons in culture, neurons were fixed at 5 d.i.v., and images from random fields for each condition were taken from at least four independent neuronal preparations. GFP positive neurons were assessed and axons were defined as a neurite that was longer than 80 mm and at least three times the length of other processes. The length of the primary axonal projection for each neuron was measured using Image J (http://rsb.info.nih.gov/ij/); and data are expressed as percent of respective controls (~ 300 axons measured for each condition from 4-6 independent experiments). Axon lengths were measured from the cell body to the distal extent of the central region of the growth cone. For determination of axon versus branch, the axon was defined as the process that remained parallel to the axon segment proximal to the branch point. Branches were defined as processes extending at orthogonal angles to the axon (Dent et. al., 2005). Total length of axon branches was not included in the measurement. The average length of axons in control groups was 197 mm ± 4 (mean ± SEM). Dendrite length was also measured in controls, with an average length of 46 um ± 1. Unlike

90 the effect observed on axons, the length of dendrites was not modified by miR-9 (45.2 um ± 0.9). For axon branching analysis, GFP positive neurons were assessed for the number of branches projecting from the axon. Data was expressed as the percentage of total neurons with 0 to >5 branch points (~ 200 neurons counted for each condition from 3-4 independent experiments). The proportion of neurons with 0 to >5 branch points had a normal distribution. For analysis of miR-9 and Map1b effects on axon branching we show data from the 0 and >5 groups. For analysis of miR-9 density in the axons, the final 20 mm of the axon were defined as the distal axonal field and miR-9 particles were counted. For evaluating the expression levels of dGFP and GW-182, fluorescence intensity in the distal axon and cell body was analysed using ImageJ. For all experiments, data are expressed as percent of respective controls (mean ± SEM).

For in utero electroporated brains, axonal length was also analysed using ImageJ. The length of individual axons was measured in brain sections, taking the edge of the GFP labelled region as the starting point (see dotted line in Fig 3b; Zhao et al., 2010). Three sections were analysed per brain, in six brains for each condition. 204 axons were measured in control brains and 192 axons in Map1b-TP electroporated brains. In all samples, the axons analyzed emanate from neurons located in the same area, i.e. in lateral position along the medio-lateral axis and in medial position along the rostro caudal axis (sensory cortex). The average length of axons measured was 198.04 mm in the control vs 286.5 mm in the Map1b-TP transfected neurons. Data are presented as mean axonal length ± SEM (n = 6) and are normalized to 100% for values obtained in control conditions.

For in vivo axonal branching experiments, brains were analysed 15 days post electroporation (i.e. at P10) and axons were labelled as described above. Image stacks were processed using Neurolucida (MBF Bioscience) and axonal arborization was measured by counting the number of primary branches (> 100 µm) extending from the longest neurite (Jeannetau et. al., 2010). Twenty axons from 3 different brains were reconstructed for each condition.

91 All data groups were expressed as the mean ± SEM and the probability distribution of the data set was analysed before further statistical analysis. For axonal length in cultured neurons, statistical significance was determined using Mann-Whitney Rank Sum test for comparison between individual groups. For axonal branching, in vivo axonal length experiments and miR-9, dGFP and GW-182 expression levels, statistical significance was determined using Student’s t-test. Statistical analysis was done using Sigma Stat 3.0 (Aspire Software) and significance compared to the respective control is indicated as follows: * p < 0.05. Individual comparisons between experimental groups are indicated with brackets.

Supplementary References

Dajas-Bailador, F., Jones, E. V. & Whitmarsh, A. J. The JIP1 scaffold protein regulates axonal development in cortical neurons. Curr Biol 18, 221-226, (2008).

Park, J. W., Vahidi, B., Taylor, A. M., Rhee, S. W. & Jeon, N. L. Microfluidic culture platform for neuroscience research. Nat Protoc 1, 2128-2136, (2006).

Lu, J. & Tsourkas, A. Imaging individual microRNAs in single mammalian cells in situ. Nucleic Acids Res 37, (2009).

Pacary, E. et al. Proneural Transcription Factors Regulate Different Steps of Cortical Neuron Migration through Rnd-Mediated Inhibition of RhoA Signaling. Neuron 69, 1069-1084, (2011).

Dent, E. W., Barnes, A. M., Tang, F. & Kalil, K. Netrin-1 and semaphorin 3A promote or inhibit cortical axon branching, respectively, by reorganization of the cytoskeleton. J Neurosci 24, 3002-3012, (2004).

Jeanneteau, F., Deinhardt, K., Miyoshi, G., Bennett, A. M. & Chao, M. V. The MAP kinase phosphatase MKP-1 regulates BDNF-induced axon branching. Nature Neuroscience, (2010).

92 Chapter 5. Discussion

In this study, I have characterized the function of one neural specific microRNA – miR-9 during development. I have identified a novel role for miR-9 in neural progenitors – to promote heterogeneity. I have shown that its function is highly context dependent, which can help explain controversial findings in the literature. In addition, a novel role for miR-9 in regulation of axonal elongation and branching was uncovered, which is the first described role for this miRNA in differentiated neurons and confirms the importance of microRNA mediated local translation in the axon.

A summary of the key findings is provided below, followed by a model of the current understanding of miR-9 function in different contexts – neural progenitors (Xenopus and mouse) and differentiated neurons.

Summary of findings

During neurogenesis, proliferating neural cells (neural progenitor or neural stem cells) undergo self-renewal to replenish the progenitor population or, alternatively, engage in asymmetric divisions associated with the generation of neurons. This is coordinated with position and time specific signals in order to determine the type and the amount of neurons generated. However, how biological regulation can be integrated with population heterogeneity to produce a specific outcome is not well understood. In the course of this study, we have uncovered a single microRNA – miR-9 can have different functions depending on the developmental context – in X. tropicalis it promotes regional diversity of the neural progenitors along the AP axis; in mouse miR-9 changes the differentiation potential of NSCs dynamically in time through regulation of Hes1 oscillations. In addition, in another context - in mature neurons, miR-9 controls the elongation and branching of the axon through regulation of the cytoskeleton.

93 miR-9 promotes regional diversity during Xenopus neural development

MicroRNA-9 is conserved at the sequence level in all vertebrate species which suggests that it has an essential function during development. Despite this, there is considerable conflicting evidence about its expression pattern and precise function in different systems (Delaloy et al., 2010; Leucht et al., 2008; Shibata et al., 2011; Zhao et al., 2009). These results raise the possibility that miR-9 function is highly context dependent, not only between species, but potentially within the same organism.

To examine this hypothesis in details we decided to use Xenopus tropicalis as a model system. Using a combination of fluorescent in situ hybridization and immunohistocheistry for neural progenitor specific markers we have shown that miR-9 expression differs along the AP axis even within a single species – it is expressed in both progenitors and differentiated neuron in the forebrain, while it becomes restricted to just the ventricular zone in the more posteriorly located regions of the brain (mid- and hindbrain). Consistent with previous reports, miR-9 appears to be absent from boundary regions (such as the MHB or the ZLI) and is present in regions undergoing active neurogenesis.

Using a loss-of-function approach to examine the function of miR-9 in the forbrain and the hindbrain, we uncovered an important difference in its role in these two regions. Knockdown of miR-9 interferes with neural differentiation in both areas, leading to a reduction in the number of differentiating neurons by ~50%. Importantly, while this is accompanied by an increase in the proliferation of the progenitors located in the hindbrain, the number of forebrain progenitors is not only not increased, but appears to be slightly downregulated in the absence of miR-9. We have shown that the underlying cause is a forebrain specific induction of apoptosis when miR-9 is absent. Most apoptotic cells are located close to the ventricle, suggesting that they represent either progenitors or immature neurons that have just initiated differentiation. Blocking apoptosis supports this observation as we observed an increase in the Sox3 positive progenitors. Therefore we propose that miR-9 is necessary for differentiation throughout its AP domain of expression, but neuronal progenitors initiating differentiation in the forebrain additionally and uniquely, require miR-9 for their survival. Interestingly,

94 knockout of Dicer in the mouse forebrain also led to increase in apoptosis in the committed neural progenitors, although the underlying microRNAs were not examined (De Pietri Tonelli et al., 2008). In addition, a recent study examined the role of Dicer in the NSCs derived from the mouse forebrain and found that while NSCs lacking Dicer were able to proliferate, they underwent apoptosis when exposed to differentiation conditions (Kawase-Koga et al., 2010). This, combined with our own observations leads us to propose that forebrain neural progenitors lacking miR-9 tend to die in the very early stages of differentiation, and when apoptosis is prevented they revert back to being progenitors rather then fully committing to neurons.

This finding highlights the importance of the origin of neural stem cells used for in vitro studies, because their positional identity will determine if they undergo apoptosis or proliferation in response to manipulating miR-9 levels. Furthermore, it reveals the importance of the regional signatures of neural tumours, since miR-9 inhibitor may be beneficial for forebrain-derived tumours due to the induction of apoptosis, but might be detrimental for tumours with hindbrain origin where it induces proliferation.

Hairy1 is the major miR-9 target and mediates the regional differences in its function by activating specific downstream signals

Since microRNAs are known to regulate multiple targets, an important question is what is the molecular basis of miR-9 loss-of-function phenotype – which are the major direct targets and downstream pathways? To address this we used prediction algorithms, GO analysis and luciferase reporter assays to identify several potential direct miR-9 targets in Xenopus. During the course of this study some of these targets were validated by others (NR2E1/TLX – (Zhao et al., 2009); Sirt1 – (Schonrock et al., 2011)), while the Xenopus homologues of some putative targets of miR-9 in mouse such as FoxG1 (Shibata et al., 2008) and Rest (Packer et al., 2008) appear to not be a subject of miR-9 regulation, even though the seed sequence is conserved. We have also identified several novel miR-9 targets, which had been implicated in regulating neural development such as the marker for basal progenitors Tis21 (Attardo et al., 2008) and the transcription factor expressed in

95 the retina Mbnl1 (Huang et al., 2008), which might mediate a different phenotype in another context/developmental stage. This suggest that even though bioinformatics predictions are useful start to search for microRNA targets, one should always verify these predictions using reporter assays and examine whether they are expressed at the right place and time to be a subject of microRNA- mediated regulation.

We decided to focus on the members of the hes (hairy and enhancer of split) family, which have been shown to play crucial roles in maintaining neural progenitors (Baek et al., 2006; Ohtsuka et al., 2001) Among them, hairy/hes1 was present in all three GO categories, showed a prominent effect in the reporter assays and was also expressed in the CNS, which is why we decided to examine it further. Using luciferase reporter assays, in situ hybridization and RT-PCR we confirmed that hairy1 is an in vivo target of miR-9 in X. tropicalis regulated at the RNA level. Furthermore, a target protector approach designed to specifically interfere with the binding of miR-9 to specific mRNA while leaving all other targets intact allowed us to confirm that miR-9/hairy1 repression is direct. Importantly, using hairy1 TP, we were able to recapitulate the miR-9 loss-of-function phenotype in vivo, including the regional specific effects in apoptosis, suggesting that hairy1 is the major target which mediates miR-9 function during Xenopus neural development. In contrast, we did not detect any changes in differentiation, proliferation or apoptosis in the neural tube using target protectors against other targets, which we identified in the screen - such as hairy2 (no obvious phenotype) and TLX (defects in eye morphogenesis).

These results confirm of the importance of target protector approaches (or analogous) in identifying the major microRNA-target interaction responsible for a particular phenotype. In addition, our findings suggest that miR-9 falls into the growing category of miRNAs that have just one or few important targets in a particular context, although many more can be bioinformatically predicted (Flynt and Lai, 2008). Such miRNAs tend to be involved in “developmental genetic switching” rather than “fine tuning”, a hypothesis that is consistent with the proposed role of miR-9 in neurogenesis.

96 Given that preventing miR-9 regulation on hairy1 mimicked the phenotype of miR-9 inhibition, we concluded that the regional specific differences that we observed in miR-9 morphant embryos reveal diversity in the response of neural progenitors to elevated hairy1 levels. We then identified some of the key transcription factors that are regulated by hairy1 throughout the neural tube to promote cell cycle reentry to be p27Xic1 and Cyclin D1. However, hairy1 appears to regulate different signaling pathways - fgf8 in the forebrain and wnt1 in the hindbrain, which, ultimately also contribute to increased proliferation. Importantly, increased hairy1 levels activate p53 only in the forebrain, which we propose is due to direct binding of hairy1 to mdm2 promoter only in this region. Since several cofactors for the Hes family of genes have been identified, such as Id and Groucho (Bai et al., 2007; McLarren et al., 2001), this specificity may be mediated by the presence, availability or activity of region specific co-factors or alternatively by phosphorylation which changes Hes1 binding affinity (reviewed in (Fischer and Gessler, 2007)).

What is the importance of hairy1 as miR-9 target? Hairy1 is usually expressed in very restricted domains (stripes) – the ZLI region in the forebrain and specific progenitor domain in the hindbrain VZ, which we hypothesize represent internal boundary regions. In the forebrain hairy1 enhances fgf8 signaling and regulates p53 activity, while in the hindbrain it appears that it limits the expression of wnt1 ventrally. MiR-9 is expressed in a mutually exclusive pattern with hairy1 and we propose that it helps establish the sharp boundaries of the hairy1 expressing domains by degrading hairy1 mRNA – thus limiting proliferation and promoting neuronal differentiation. However, a remaining question is whether the regulation of hes genes by miR-9 is conserved during evolution and if it fulfills the same function during neural development.

Mouse Hes1 expression is dynamic and oscillates in time

In the mouse hairy1 homologue – Hes1 is also expressed in boundary regions such as the ZLI and MHB and in the hindbrain it appears to be present in two stripe domains, similar to the Xenopus hairy1 (Baek et al., 2006). However, in addition, Hes1 is also present in a salt-and-pepper manner in the VZ of the developing

97 telencephalon – a pattern, which is evolutionary new as it is not evident for the Xenopus and zebrafish homologues. It has been recently shown that this expression is in fact oscillatory and that such mode of expression is essential for maintaining neural progenitors in a proliferative state, as sustained high expression leads to slow proliferation, while sustained low initiates neuronal differentiation (Shimojo et al., 2008). This findings, together with the observation that Hes1 induced oscillation drive other components of the Notch signaling pathway (Hes1, ngn2 and Dll1), have transformed our view of lateral inhibition from a linear amplification of stochastic differences to that of a dynamic, cyclical and mutual, inhibition of differentiation.

Even though Hes1 oscillations are well documented, little is known about the exact molecular components that determine whether Hes1 should oscillate (in the VZ of the telencephalon) or not (in the ZLI and MHB). In addition, while the components of the basic oscillatory loop have been identified, it is unclear how it is regulated and how neural progenitors escape from it in order to initiate differentiation. Thus it is important to understand the mechanistic details driving oscillations and to determine how they are regulated in vivo.

Some of the key requirements to generate oscillatory pattern are negative feedback, delay and instability of both hes1 protein and hes1 mRNA. While the feedback loop (Hirata et al., 2002), the delay (Takashima et al., 2011) and protein stability (Hirata et al., 2004) have been extensively studied, not much is known about the molecular mechanisms that determine the regulation of hes1 mRNA stability. However, mathematical modeling predicted that mRNA stability is important as hes1 oscillates only within certain values of mRNA stability and proposed the existence of microRNA mediated control of hes1 mRNA (Xie et al., 2007).

miR-9 regulates Hes1 mRNA stability and modulates its oscillatory expression

Using luciferase reporters we have confirmed that Hes1 is a target of miR-9, similar to its Xenopus homologue hairy1. In addition, we showed the mechanism of miR-9 regulation on Hes1 is also at the RNA level – by promoting RNA degradation

98 and therefore destabilizing Hes1 message. Importantly, this interaction is direct as results obtaining using an LNA – based inhibitor of miR-9 and Hes1 target protector oligo were indistinguishable.

We examined whether miR-9 contributes to the oscillating pattern of Hes1 in neural progenitors and showed that overexpression of miR-9 leads to dampening of the oscillations, while inhibiting miR-9 increases their amplitude. While the results of miR-9 overexpression match very well theoretical predictions, mathematical modeling also predicts dampening of the oscillations when miR-9 is stabilized. It is possible that such dampening occurred after 2 cycles, which prevents us from observing it using biochemical methods due to asynchronization of the population. Alternatively, the rate of protein translation (which could be affected by microRNA regulation) can have an effect on oscillation kinetics, in addition to RNA stability. Further experiments including single cell imaging in the presence/absence of miR-9 as well as mathematical modeling would be required to distinguish between these possibilities.

We propose that miR-9 mediated destabilization of Hes1 mRNA is necessary for Hes1 oscillations and thus, it appears to be another, previously unknown layer of regulation. Since hes1 oscillations are asynchronous in neural progenitor cells, the population appears highly heterogeneous with regards to gene expression, at any given time. While traditionally microRNAs are associated with buffering transcriptional noise and thus ensuring the homogeneity of the response, we have uncovered a novel role for microRNAs in promoting heterogeneity and thus allowing some cells to adopt different cell fate over time.

Hes1 represses miR-9 transcription forming a negative feedback loop

The observation that the majority of the cells in the mouse VZ had inversely related expression levels of miR-9 and Hes1 prompted us to ask whether Hes1 can also be regulating miR-9, forming a negative feedback loop. Using a bioinformatics approach we found several Hes1 binding motifs in the promoters of the three primary miR-9 transcripts and confirmed that overexpression of Hes1 inhibits the expression of all of them. In addition, upon serum stimulation pri-miR-9 levels were found to oscillate out of phase with Hes1, presumably due to release from

99 inhibition when Hes1 levels are low, leading to bursts of miR-9 transcription. The mutually repressive interaction of miR-9 and Hes1 predicts that when Hes1 levels are constitutively high (as in boundary regions), the levels of miR-9 would be low, while when the levels of Hes1 would constitutively low (as in neurons) the levels of miR-9 would be high. This is indeed the case, as miR-9 is not expressed in boundary regions of the vertebrate CNS (as shown by us and others – (Leucht et al., 2008)), but its expression increases during neuronal differentiation and is present at high levels in mature forebrain neurons (Laneve et al., 2010).

While the mechanisms, by which Hes1 oscillations can be initiated and maintained have been examined (Hirata et al., 2002; Nakayama et al., 2008), a major unresolved question is how neural progenitors exit the oscillatory cycle in order to differentiate and is it an autonomous process? We have provided evidence that while primary miR-9 transcripts oscillate out-of-phase with Hes1, due to its high stability the mature miR-9 levels do not exhibit such oscillatory behavior and instead accumulate gradually with each consecutive Hes1 cycle. This, coupled with the dampening of Hes1 oscillations when miR-9 is overexpressed, lead us to propose that accumulation of miR-9 over time allows the cells to escape the progenitor state by dampening Hes1 oscillations. Thus the amount of mature miR- 9 serves as an output of the oscillation clock. This process, while being primarily cell autonomous, can still be a subject of regulation, such as by modifying the binding of Hes1 on miR-9 promoters by co-factors or degrading miR-9 levels.

An important aspect of brain evolution is the extended proliferative state of neural progenitors in mammals leading to the expansion of the cortex. Since Hes1 oscillation are integral part of the progenitor maintenance in mouse it is tempting to speculate that interaction of miR-9 with Hes1 in the VZ of the forebrain (which is not present in Xenopus) has contributed to this process. Thus while in xenopus the role of miR-9 is to promote positional progenitor diversity by ensuring sharp boundaries of the hes1 positive domain, in mammals miR-9/Hes1 interaction establishes a different type of heterogeneity – temporal oscillations in gene expression in single cells making the population heterogeneous. At the same time the lack of miR-9 in regions such as the MHB and the ZLI doesn’t allow for such

100 oscillatory expression and instead Hes1 is expressed at persistently high levels, leading to slow proliferation and adoption of boundary features.

However, if regulation of hes genes were the sole function of miR-9, its transcription would not continue in differentiated neurons where hes1 is absent. Therefore we decided to examine whether miR-9 function might be different in this context.

miR-9 regulates axonal development in mature neurons through MAP1B

We have shown that miR-9 expression is also present in differentiated cortical neurons as well as in the VZ at stage E17.5 (similar to its expression in xenopus), where its function is completely unknown. Hes genes are not expressed in differentiated neurons and most of the identified targets for miR-9 are also absent in this area.

Using fluorescent in situ hybridization we have shown that mature miR-9 is enriched in the developing axon in cortical neuron in vitro. Moreover, manipulation of miR-9 by overexpression results in reduced length and increased branching, while inhibition leads to shorter axons and increased number of branch points. Since cytoskeleton components are crucial for the axonal polarization and growth we examined potential miR-9 targets, which might be involved in the regulation of actin or microtubule assembly. Using approach analogous to the one utilized in Xenopus (combination of bioinformatic predictions, GO analysis and luciferase reporter assays), we identified microtubule associated protein 1B (MAP1B) as a putative target of miR-9 in neurons. Contrary to most of the identified miR-9 targets (including hes genes in this study) miR-9 represses MAP1B at the protein level, without affecting RNA degradation. This could be important for dynamic and fast regulation of translation in distal regions such as the tip of the axon, where it would be more beneficial to temporarily silence a particular mRNA rather than degrade it.

What is the functional importance of MAP1B in mediating the axonal function of miR-9? Using a target protector approach (a LNA-modified oligo which interferes with miR-9/MAP1B 3’UTR binding) we rescued miR-9 overexpression

101 phenotype, while MAP1B TP alone mimicked both the increase in axon length and reduction in branch points induced by inhibition of miR-9. In addition, in utero electroporation of MAP1B TP in the cortex of an E14.5 embryo led to a significant increase in axon length, confirming the importance of post-transcriptional MAP1B regulation in vivo. The use of target protector allowed us to bypass the migration defects observed when miR-9 was knocked down and confirm the results obtained from the in vitro approach.

An important question is whether the observed effects on axon growth and branching are due to local translational repression by miR-9 or are secondary of the interaction between MAP1B and miR-9 in the cell body. Using microfluidic chambers we showed that locally inhibiting miR-9 only at the axonal side has the same effect on growth and branching (namely increasing length and decreasing branch points) compared to transfection in the cell body. This provides experimental evidence for the first time for the importance of local translational control for the development of the axon by microRNAs.

The dual role of BDNF signaling during axon development is mediated by miR-9/MAP1B

Axon development is ultimately regulated by external cues and therefore we wondered what the potential mechanisms for regulating miR-9/MAP1B interaction are. The role of brain-derived neurothropic factor (BDNF) is a signaling pathway in axon guidance and development is crucial, however the functional mechanism is not well understood. It has been shown previously that BDNF has a dual role in axon development and here we confirm this while providing mechanistic information – a short stimulation or low dose of BDNF leads to an increase in axon length, which is accompanied by a reduction in the number of miR-9 puncta at the distal part of the axon, while longer (24h) stimulation or higher BDNF concentration results in accumulation of miR-9 particles and a corresponding increase in the number of branch points. Furthermore, we have shown that this is mediated through MAP1B and thus it represents a new, previously unknown, aspect of BDNF dependent regulation on axon development.

102 Overall, these findings improve our understanding of how cytoskeleton modification (such as mediated by MAP1B proteins) are regulated in the distal parts of the growing axons by microRNA silencing, in order to guide the growth of the axon during its extension phase (when the concentration of extrinsic cues such as BDNF is low and directional) and promote its branching once it has reached the target (where the concentration of BDNF is high).

Significance of miR-9 mediated regulation during neural development

MicroRNAs are a relatively novel class of regulators of gene expression, which have been traditionally associated with cell fate choice and establishing gene “robustness”. In this study we have shown that function of one neuronal- specific microRNA – miR-9 is highly context dependent. In the developing Xenopus embryo it promotes positional diversity of the neural progenitors through repressing hairy1 and highlights the importance of taking into account the positional origin of stem cells in designing rational strategies to manipulate their proliferative potential. In the mammalian CNS, I propose that miR-9/Hes1 interaction has gained a new evolutionary significance – by driving Hes1 oscillations, it reveals neural progenitor diversity in time, rather than in space, Finally, the expression of miR-9 in the axons of mature neurons is required for proper axon development and in this context it regulates the local assembly of the cytoskeleton through MAP1B. Thus while microRNAs are predicted to have many different targets, we propose that they have primarily one major target in a different context – either in separate regions, or during different times in development.

103 References:

Ables, J.L., Breunig, J.J., Eisch, A.J., Rakic, P., 2011. Not(ch) just development: Notch signalling in the adult brain. Nat Rev Neurosci 12, 269-283. Anthony, T., Klein, C., Fishell, G., Heintz, N., 2004. Radial glia serve as neuronal progenitors in all regions of the central nervous system. Neuron 41, 881-890. Aoki, K., Taketo, M.M., 2007. Adenomatous polyposis coli (APC): a multi- functional tumor suppressor gene. J Cell Sci 120, 3327-3335. Aravin, A.A., Lagos-Quintana, M., Yalcin, A., Zavolan, M., Marks, D., Snyder, B., Gaasterland, T., Meyer, J., Tuschl, T., 2003. The small RNA profile during Drosophila melanogaster development. Dev Cell 5, 337-350. Artavanis-Tsakonas, S., Rand, M., Lake, R., 1999. Notch signaling: cell fate control and signal integration in development. Science 284, 770-776. Attardo, A., Calegari, F., Haubensak, W., Wilsch-Bräuninger, M., Huttner, W.B., 2008. Live imaging at the onset of cortical neurogenesis reveals differential appearance of the neuronal phenotype in apical versus basal progenitor progeny. PLoS ONE 3, e2388. Bachler, M., Neubüser, A., 2001. Expression of members of the Fgf family and their receptors during midfacial development. Mech Dev 100, 313-316. Baek, J.H., Hatakeyama, J., Sakamoto, S., Ohtsuka, T., Kageyama, R., 2006. Persistent and high levels of Hes1 expression regulate boundary formation in the developing central nervous system. Development 133, 2467-2476. Bai, G., Sheng, N., Xie, Z., Bian, W., Yokota, Y., Benezra, R., Kageyama, R., Guillemot, F., Jing, N., 2007. Id sustains Hes1 expression to inhibit precocious neurogenesis by releasing negative autoregulation of Hes1. Dev Cell 13, 283-297. Banerjee, S., Neveu, P., Kosik, K.S., 2009. A Coordinated Local Translational Control Point at the Synapse Involving Relief from Silencing and MOV10 Degradation. Neuron 64, 871-884. Barnes, A.P., Polleux, F., 2009. Establishment of axon-dendrite polarity in developing neurons. Annu Rev Neurosci 32, 347-381. Bartel, D.P., Chen, C.Z., 2004. Micromanagers of gene expression: the potentially widespread influence of metazoan microRNAs. Nat Rev Genet 5, 396-400. Basyuk, E., Suavet, F., Doglio, A., Bordonne, R., Bertrand, E., 2003. Human let-7 stem-loop precursors harbor features of RNase III cleavage products. Nucleic Acids Res 31, 6593-6597. Bentwich, I., Avniel, A., Karov, Y., Aharonov, R., Gilad, S., Barad, O., Barzilai, A., Einat, P., Einav, U., Meiri, E., Sharon, E., Spector, Y., Bentwich, Z., 2005. Identification of hundreds of conserved and nonconserved human microRNAs. Nat Genet 37, 766-770.

104 Berezikov, E., Guryev, V., van de Belt, J., Wienholds, E., Plasterk, R.H., Cuppen, E., 2005. Phylogenetic shadowing and computational identification of human microRNA genes. Cell 120, 21-24. Black, M.M., Slaughter, T., Fischer, I., 1994. Microtubule-associated protein 1b (MAP1b) is concentrated in the distal region of growing axons. J Neurosci 14, 857- 870. Burrows, R.C., Wancio, D., Levitt, P., Lillien, L., 1997. Response diversity and the timing of progenitor cell maturation are regulated by developmental changes in EGFR expression in the cortex. Neuron 19, 251-267. Cao, X., Yeo, G., Muotri, A.R., Kuwabara, T., Gage, F.H., 2006. Noncoding RNAs in the mammalian central nervous system. Annu Rev Neurosci 29, 77-103. Chenn, A., Walsh, C., 2003. Increased neuronal production, enlarged forebrains and cytoarchitectural distortions in beta-catenin overexpressing transgenic mice. Cereb Cortex 13, 599-606. Cohen, S.M., Brennecke, J., Stark, A., 2006. Denoising feedback loops by thresholding--a new role for microRNAs. Genes Dev 20, 2769-2772. Conaco, C., Otto, S., Han, J.-J., Mandel, G., 2006. Reciprocal actions of REST and a microRNA promote neuronal identity. Proc Natl Acad Sci USA 103, 2422-2427. Danesin, C., Peres, J.N., Johansson, M., Snowden, V., Cording, A., Papalopulu, N., Houart, C., 2009. Integration of telencephalic Wnt and hedgehog signaling center activities by Foxg1. Dev Cell 16, 576-587. De Pietri Tonelli, D., Pulvers, J.N., Haffner, C., Murchison, E.P., Hannon, G.J., Huttner, W.B., 2008. miRNAs are essential for survival and differentiation of newborn neurons but not for expansion of neural progenitors during early neurogenesis in the mouse embryonic neocortex. Development 135, 3911-3921. Delaloy, C., Liu, L., Lee, J.A., Su, H., Shen, F., Yang, G.Y., Young, W.L., Ivey, K.N., Gao, F.B., 2010. MicroRNA-9 coordinates proliferation and migration of human embryonic stem cell-derived neural progenitors. Cell stem cell 6, 323-335. Dent, E.W., Callaway, J.L., Szebenyi, G., Baas, P.W., Kalil, K., 1999. Reorganization and movement of microtubules in axonal growth cones and developing interstitial branches. J Neurosci 19, 8894-8908. Finlay, B.L., Darlington, R.B., 1995. Linked regularities in the development and evolution of mammalian brains. Science 268, 1578-1584. Fischer, A., Gessler, M., 2007. Delta-Notch--and then? Protein interactions and proposed modes of repression by Hes and Hey bHLH factors. Nucleic Acids Res 35, 4583-4596. Flynt, A.S., Lai, E.C., 2008. Biological principles of microRNA-mediated regulation: shared themes amid diversity. Nat Rev Genet 9, 831-842. Gallo, G., 2011. The cytoskeletal and signaling mechanisms of axon collateral branching. Dev Neurobiol 71, 201-220. Gibson, D.A., Ma, L., 2011. Developmental regulation of axon branching in the vertebrate nervous system. Development 138, 183-195.

105 Giraldez, A.J., Cinalli, R.M., Glasner, M.E., Enright, A.J., Thomson, J.M., Baskerville, S., Hammond, S.M., Bartel, D.P., Schier, A.F., 2005. MicroRNAs regulate brain morphogenesis in zebrafish. Science 308, 833-838. Giraldez, A.J., Mishima, Y., Rihel, J., Grocock, R.J., Van Dongen, S., Inoue, K., Enright, A.J., Schier, A.F., 2006. Zebrafish MiR-430 promotes deadenylation and clearance of maternal mRNAs. Science 312, 75-79. Goll, M.G., Bestor, T.H., 2005. Eukaryotic cytosine methyltransferases. Annu Rev Biochem 74, 481-514. Gordon-Weeks, P.R., Fischer, I., 2000. MAP1B expression and microtubule stability in growing and regenerating axons. Microsc Res Tech 48, 63-74. Götz, M., Huttner, W.B., 2005. The cell biology of neurogenesis. Nat Rev Mol Cell Biol 6, 777-788. Grimson, A., Farh, K.K., Johnston, W.K., Garrett-Engele, P., Lim, L.P., Bartel, D.P., 2007. MicroRNA targeting specificity in mammals: determinants beyond seed pairing. Mol Cell 27, 91-105. Guillemot, F., 2007. Cell fate specification in the mammalian telencephalon. Prog Neurobiol 83, 37-52. Guo, H., Ingolia, N.T., Weissman, J.S., Bartel, D.P., 2010. Mammalian microRNAs predominantly act to decrease target mRNA levels. Nature 466, 835-840. Haubensak, W., Attardo, A., Denk, W., Huttner, W., 2004. Neurons arise in the basal neuroepithelium of the early mammalian telencephalon: a major site of neurogenesis. Proc Natl Acad Sci U S A 101, 3196-3201. Hengst, U., Cox, L.J., Macosko, E.Z., Jaffrey, S.R., 2006. Functional and selective RNA interference in developing axons and growth cones. J Neurosci 26, 5727-5732. Hengst, U., Jaffrey, S.R., 2007. Function and translational regulation of mRNA in developing axons. Semin Cell Dev Biol 18, 209-215. Hentges, K.E., Sirry, B., Gingeras, A.C., Sarbassov, D., Sonenberg, N., Sabatini, D., Peterson, A.S., 2001. FRAP/mTOR is required for proliferation and patterning during embryonic development in the mouse. Proc Natl Acad Sci USA 98, 13796- 13801. Hirabayashi, Y., Gotoh, Y., 2005. Stage-dependent fate determination of neural precursor cells in mouse forebrain. Neurosci Res 51, 331-336. Hirabayashi, Y., Gotoh, Y., 2010. Epigenetic control of neural precursor cell fate during development. Nat Rev Neurosci 11, 377-388. Hirabayashi, Y., Itoh, Y., Tabata, H., Nakajima, K., Akiyama, T., Masuyama, N., Gotoh, Y., 2004. The Wnt/beta-catenin pathway directs neuronal differentiation of cortical neural precursor cells. Development 131, 2791-2801. Hirabayashi, Y., Suzki, N., Tsuboi, M., Endo, T.A., Toyoda, T., Shinga, J., Koseki, H., Vidal, M., Gotoh, Y., 2009. Polycomb limits the neurogenic competence of neural precursor cells to promote astrogenic fate transition. Neuron 63, 600-613.

106 Hirata, H., Bessho, Y., Kokubu, H., Masamizu, Y., Yamada, S., Lewis, J., Kageyama, R., 2004. Instability of Hes7 protein is crucial for the somite segmentation clock. Nat Genet 36, 750-754. Hirata, H., Yoshiura, S., Ohtsuka, T., Bessho, Y., Harada, T., Yoshikawa, K., Kageyama, R., 2002. Oscillatory expression of the bHLH factor Hes1 regulated by a negative feedback loop. Science 298, 840-843. Hoshino, N., Vatterott, P., Egwiekhor, A., Rochlin, M.W., 2010. Brain-derived neurotrophic factor attracts geniculate ganglion neurites during embryonic targeting. Dev Neurosci 32, 184-196. Huang, H., Wahlin, K.J., McNally, M., Irving, N.D., Adler, R., 2008. Developmental regulation of muscleblind-like (MBNL) gene expression in the chicken embryo retina. Dev Dyn 237, 286-296. Hutvagner, G., McLachlan, J., Pasquinelli, A.E., Balint, E., Tuschl, T., Zamore, P.D., 2001. A cellular function for the RNA-interference enzyme Dicer in the maturation of the let-7 small temporal RNA. Science 293, 834-838. Ji, Y., Lu, Y., Yang, F., Shen, W., Tang, T.T.-T., Feng, L., Duan, S., Lu, B., 2010. Acute and gradual increases in BDNF concentration elicit distinct signaling and functions in neurons. Nat Neurosci 13, 302-309. Jiang, H., Guo, W., Liang, X., Rao, Y., 2005. Both the establishment and the maintenance of neuronal polarity require active mechanisms: critical roles of GSK- 3beta and its upstream regulators. Cell 120, 123-135. Kageyama, R., Ohtsuka, T., Shimojo, H., Imayoshi, I., 2008. Dynamic Notch signaling in neural progenitor cells and a revised view of lateral inhibition. Nat Neurosci 11, 1247-1251. Kataoka, A., Shimogori, T., 2008. Fgf8 controls regional identity in the developing thalamus. Development 135, 2873-2881. Kawase-Koga, Y., Low, R., Otaegi, G., Pollock, A., Deng, H., Eisenhaber, F., Maurer-Stroh, S., Sun, T., 2010. RNAase-III enzyme Dicer maintains signaling pathways for differentiation and survival in mouse cortical neural stem cells. J Cell Sci 123, 586-594. Kiecker, C., Lumsden, A., 2005. Compartments and their boundaries in vertebrate brain development. Nat Rev Neurosci 6, 553-564. Kishi, Y., Takahashi, J., Koyanagi, M., Morizane, A., Okamoto, Y., Horiguchi, S., Tashiro, K., Honjo, T., Fujii, S., Hashimoto, N., 2005. Estrogen promotes differentiation and survival of dopaminergic neurons derived from human neural stem cells. J Neurosci Res 79, 279-286. Kleiman, R., Banker, G., Steward, O., 1994. Development of subcellular mRNA compartmentation in hippocampal neurons in culture. J Neurosci 14, 1130-1140. Kloosterman, W.P., Wienholds, E., de Bruijn, E., Kauppinen, S., Plasterk, R.H., 2006. In situ detection of miRNAs in animal embryos using LNA-modified oligonucleotide probes. Nat Methods 3, 27-29.

107 Krichevsky, A.M., Sonntag, K.-C., Isacson, O., Kosik, K.S., 2006. Specific microRNAs modulate embryonic stem cell-derived neurogenesis. Stem Cells 24, 857-864. Lagos-Quintana, M., Rauhut, R., Yalcin, A., Meyer, J., Lendeckel, W., Tuschl, T., 2002. Identification of tissue-specific microRNAs from mouse. Curr Biol 12, 735- 739. Laneve, P., Gioia, U., Andriotto, A., Moretti, F., Bozzoni, I., Caffarelli, E., 2010. A minicircuitry involving REST and CREB controls miR-9-2 expression during human neuronal differentiation. Nucleic acids research. Larsen, C.W., Zeltser, L.M., Lumsden, A., 2001. Boundary formation and compartition in the avian diencephalon. J Neurosci 21, 4699-4711. Lau, N.C., Lim, L.P., Weinstein, E.G., Bartel, D.P., 2001. An abundant class of tiny RNAs with probable regulatory roles in Caenorhabditis elegans. Science 294, 858- 862. Lee, Y., Ahn, C., Han, J., Choi, H., Kim, J., Yim, J., Lee, J., Provost, P., Radmark, O., Kim, S., Kim, V.N., 2003. The nuclear RNase III Drosha initiates microRNA processing. Nature 425, 415-419. Lee, Y., Jeon, K., Lee, J.T., Kim, S., Kim, V.N., 2002. MicroRNA maturation: stepwise processing and subcellular localization. EMBO J 21, 4663-4670. Leucht, C., Stigloher, C., Wizenmann, A., Klafke, R., Folchert, A., Bally-Cuif, L., 2008. MicroRNA-9 directs late organizer activity of the midbrain-hindbrain boundary. Nat Neurosci 11, 641-648. Leung, K.M., van Horck, F.P., Lin, A.C., Allison, R., Standart, N., Holt, C.E., 2006. Asymmetrical beta-actin mRNA translation in growth cones mediates attractive turning to netrin-1. Nat Neurosci 9, 1247-1256. Lewis, B.P., Shih, I.H., Jones-Rhoades, M.W., Bartel, D.P., Burge, C.B., 2003. Prediction of mammalian microRNA targets. Cell 115, 787-798. Lim, L.P., Glasner, M.E., Yekta, S., Burge, C.B., Bartel, D.P., 2003. Vertebrate microRNA genes. Science 299, 1540. Lim, L.P., Lau, N.C., Garrett-Engele, P., Grimson, A., Schelter, J.M., Castle, J., Bartel, D.P., Linsley, P.S., Johnson, J.M., 2005. Microarray analysis shows that some microRNAs downregulate large numbers of target mRNAs. Nature 433, 769- 773. Liu, W., Mao, S.Y., Zhu, W.Y., 2007. Impact of tiny miRNAs on cancers. World J Gastroenterol 13, 497-502. Long, D., Lee, R., Williams, P., Chan, C.Y., Ambros, V., Ding, Y., 2007. Potent effect of target structure on microRNA function. Nat Struct Mol Biol 14, 287-294. Louvi, A., Artavanis-Tsakonas, S., 2006. Notch signalling in vertebrate neural development. Nat Rev Neurosci 7, 93-102. Lund, E., Guttinger, S., Calado, A., Dahlberg, J.E., Kutay, U., 2004. Nuclear export of microRNA precursors. Science 303, 95-98.

108 Makeyev, E.V., Zhang, J., Carrasco, M.A., Maniatis, T., 2007. The MicroRNA miR- 124 promotes neuronal differentiation by triggering brain-specific alternative pre- mRNA splicing. Mol Cell 27, 435-448. Malatesta, P., Hack, M.A., Hartfuss, E., Kettenmann, H., Klinkert, W., Kirchhoff, F., Götz, M., 2003. Neuronal or glial progeny: regional differences in radial glia fate. Neuron 37, 751-764. Mansfield, J.H., Harfe, B.D., Nissen, R., Obenauer, J., Srineel, J., Chaudhuri, A., Farzan-Kashani, R., Zuker, M., Pasquinelli, A.E., Ruvkun, G., Sharp, P.A., Tabin, C.J., McManus, M.T., 2004. MicroRNA-responsive 'sensor' transgenes uncover Hox-like and other developmentally regulated patterns of vertebrate microRNA expression. Nat Genet 36, 1079-1083. Martin, C., Zhang, Y., 2005. The diverse functions of histone lysine methylation. Nat Rev Mol Cell Biol 6, 838-849. McLarren, K.W., Theriault, F.M., Stifani, S., 2001. Association with the nuclear matrix and interaction with Groucho and RUNX proteins regulate the transcription repression activity of the basic helix loop helix factor Hes1. J Biol Chem 276, 1578- 1584. Megason, S.G., McMahon, A.P., 2002. A mitogen gradient of dorsal midline Wnts organizes growth in the CNS. Development 129, 2087-2098. Menager, C., Arimura, N., Fukata, Y., Kaibuchi, K., 2004. PIP3 is involved in neuronal polarization and axon formation. J Neurochem 89, 109-118. Mikkelsen, T.S., Ku, M., Jaffe, D.B., Issac, B., Lieberman, E., Giannoukos, G., Alvarez, P., Brockman, W., Kim, T.-K., Koche, R.P., Lee, W., Mendenhall, E., O'Donovan, A., Presser, A., Russ, C., Xie, X., Meissner, A., Wernig, M., Jaenisch, R., Nusbaum, C., Lander, E.S., Bernstein, B.E., 2007. Genome-wide maps of chromatin state in pluripotent and lineage-committed cells. Nature 448, 553-560. Miller, F.D., Gauthier, A.S., 2007. Timing is everything: making neurons versus glia in the developing cortex. Neuron 54, 357-369. Miska, E.A., Alvarez-Saavedra, E., Townsend, M., Yoshii, A., Sestan, N., Rakic, P., Constantine-Paton, M., Horvitz, H.R., 2004. Microarray analysis of microRNA expression in the developing mammalian brain. Genome Biol 5, R68. Mohn, F., Weber, M., Rebhan, M., Roloff, T.C., Richter, J., Stadler, M.B., Bibel, M., Schübeler, D., 2008. Lineage-specific polycomb targets and de novo DNA methylation define restriction and potential of neuronal progenitors. Mol Cell 30, 755-766. Molyneaux, B.J., Arlotta, P., Menezes, J.R.L., Macklis, J.D., 2007. Neuronal subtype specification in the cerebral cortex. Nat Rev Neurosci 8, 427-437. Nakayama, K., Satoh, T., Igari, A., Kageyama, R., Nishida, E., 2008. FGF induces oscillations of Hes1 expression and Ras/ERK activation. Curr Biol 18, R332-334. Natera-Naranjo, O., Aschrafi, A., Gioio, A.E., Kaplan, B.B., 2010. Identification and quantitative analyses of microRNAs located in the distal axons of sympathetic neurons. RNA 16, 1516-1529.

109 Ohler, U., Yekta, S., Lim, L.P., Bartel, D.P., Burge, C.B., 2004. Patterns of flanking sequence conservation and a characteristic upstream motif for microRNA gene identification. RNA 10, 1309-1322. Ohtsuka, T., Ishibashi, M., Gradwohl, G., Nakanishi, S., Guillemot, F., Kageyama, R., 1999. Hes1 and Hes5 as notch effectors in mammalian neuronal differentiation. EMBO J 18, 2196-2207. Ohtsuka, T., Sakamoto, M., Guillemot, F., Kageyama, R., 2001. Roles of the basic helix-loop-helix genes Hes1 and Hes5 in expansion of neural stem cells of the developing brain. J Biol Chem 276, 30467-30474. Packer, A.N., Xing, Y., Harper, S.Q., Jones, L., Davidson, B.L., 2008. The bifunctional microRNA miR-9/miR-9* regulates REST and CoREST and is downregulated in Huntington's disease. J Neurosci 28, 14341-14346. Pasquinelli, A.E., Ruvkun, G., 2002. Control of developmental timing by micrornas and their targets. Annu Rev Cell Dev Biol 18, 495-513. Pillai, R.S., Bhattacharyya, S.N., Filipowicz, W., 2007. Repression of protein synthesis by miRNAs: how many mechanisms? Trends Cell Biol 17, 118-126. Polleux, F., Snider, W., 2010. Initiating and growing an axon. Cold Spring Harb Perspect Biol 2, a001925. Robins, H., Li, Y., Padgett, R.W., 2005. Incorporating structure to predict microRNA targets. Proc Natl Acad Sci U S A 102, 4006-4009. Schmidt, H., Rathjen, F.G., 2010. Signalling mechanisms regulating axonal branching in vivo. Bioessays 32, 977-985. Scholpp, S., Foucher, I., Staudt, N., Peukert, D., Lumsden, A., Houart, C., 2007. Otx1l, Otx2 and Irx1b establish and position the ZLI in the diencephalon. Development 134, 3167-3176. Schonrock, N., Humphreys, D.T., Preiss, T., Gotz, J., 2011. Target Gene Repression Mediated by miRNAs miR-181c and miR-9 Both of Which Are Down- regulated by Amyloid-beta. J Mol Neurosci. Schratt, G.M., Tuebing, F., Nigh, E.A., Kane, C.G., Sabatini, M.E., Kiebler, M., Greenberg, M.E., 2006. A brain-specific microRNA regulates dendritic spine development. Nature 439, 283-289. Schuettengruber, B., Chourrout, D., Vervoort, M., Leblanc, B., Cavalli, G., 2007. Genome regulation by polycomb and trithorax proteins. Cell 128, 735-745. Segal, R.A., 2003. Selectivity in neurotrophin signaling: theme and variations. Annu Rev Neurosci 26, 299-330. Selkoe, D., Kopan, R., 2003. Notch and Presenilin: regulated intramembrane proteolysis links development and degeneration. Annu Rev Neurosci 26, 565-597. Sempere, L.F., Freemantle, S., Pitha-Rowe, I., Moss, E., Dmitrovsky, E., Ambros, V., 2004. Expression profiling of mammalian microRNAs uncovers a subset of brain-expressed microRNAs with possible roles in murine and human neuronal differentiation. Genome Biol 5, R13.

110 Shen, Q., Wang, Y., Dimos, J.T., Fasano, C.A., Phoenix, T.N., Lemischka, I.R., Ivanova, N.B., Stifani, S., Morrisey, E.E., Temple, S., 2006. The timing of cortical neurogenesis is encoded within lineages of individual progenitor cells. Nat Neurosci 9, 743-751. Shi, S.H., Cheng, T., Jan, L.Y., Jan, Y.N., 2004. APC and GSK-3beta are involved in mPar3 targeting to the nascent axon and establishment of neuronal polarity. Curr Biol 14, 2025-2032. Shi, S.H., Jan, L.Y., Jan, Y.N., 2003. Hippocampal neuronal polarity specified by spatially localized mPar3/mPar6 and PI 3-kinase activity. Cell 112, 63-75. Shi, Y., Sun, G., Zhao, C., Stewart, R., 2008. Neural stem cell self-renewal. Crit Rev Oncol Hematol 65, 43-53. Shibata, M., Kurokawa, D., Nakao, H., Ohmura, T., Aizawa, S., 2008. MicroRNA- 9 modulates Cajal-Retzius cell differentiation by suppressing Foxg1 expression in mouse medial pallium. J Neurosci 28, 10415-10421. Shibata, M., Nakao, H., Kiyonari, H., Abe, T., Aizawa, S., 2011. MicroRNA-9 regulates neurogenesis in mouse telencephalon by targeting multiple transcription factors. J Neurosci 31, 3407-3422. Shimojo, H., Ohtsuka, T., Kageyama, R., 2008. Oscillations in notch signaling regulate maintenance of neural progenitors. Neuron 58, 52-64. Shinya, M., Koshida, S., Sawada, A., Kuroiwa, A., Takeda, H., 2001. Fgf signalling through MAPK cascade is required for development of the subpallial telencephalon in zebrafish embryos. Development 128, 4153-4164. Smirnova, L., Grafe, A., Seiler, A., Schumacher, S., Nitsch, R., Wulczyn, F.G., 2005. Regulation of miRNA expression during neural cell specification. Eur J Neurosci 21, 1469-1477. Stancheva, I., Collins, A.L., Van den Veyver, I.B., Zoghbi, H., Meehan, R.R., 2003. A mutant form of MeCP2 protein associated with human Rett syndrome cannot be displaced from methylated DNA by notch in Xenopus embryos. Mol Cell 12, 425- 435. Stefani, G., Slack, F.J., 2008. Small non-coding RNAs in animal development. Nat Rev Mol Cell Biol 9, 219-230. Stiess, M., Bradke, F., 2011. Neuronal polarization: the cytoskeleton leads the way. Dev Neurobiol 71, 430-444. Storm, E.E., Garel, S., Borello, U., Hebert, J.M., Martinez, S., McConnell, S.K., Martin, G.R., Rubenstein, J.L., 2006. Dose-dependent functions of Fgf8 in regulating telencephalic patterning centers. Development 133, 1831-1844. Suzuki-Hirano, A., Shimogori, T., 2009. The role of Fgf8 in telencephalic and diencephalic patterning. Semin Cell Dev Biol 20, 719-725. Takashima, Y., Ohtsuka, T., Gonzalez, A., Miyachi, H., Kageyama, R., 2011. Intronic delay is essential for oscillatory expression in the segmentation clock. Proc Natl Acad Sci U S A 108, 3300-3305.

111 Tay, Y., Zhang, J., Thomson, A., Lim, B., Rigoutsos, I., 2008a. MicroRNAs to Nanog, Oct4 and Sox2 coding regions modulate embryonic stem cell differentiation. Nature. Tay, Y.M., Tam, W.L., Ang, Y.S., Gaughwin, P.M., Yang, H., Wang, W., Liu, R., George, J., Ng, H.H., Perera, R.J., Lufkin, T., Rigoutsos, I., Thomson, A.M., Lim, B., 2008b. MicroRNA-134 modulates the differentiation of mouse embryonic stem cells, where it causes post-transcriptional attenuation of Nanog and LRH1. Stem Cells 26, 17-29. Temple, S., 2001. The development of neural stem cells. Nature 414, 112-117. Vieira, C., Garda, A.-L., Shimamura, K., Martinez, S., 2005. Thalamic development induced by Shh in the chick embryo. Dev Biol 284, 351-363. Visvanathan, J., Lee, S., Lee, B., Lee, J.W., Lee, S.K., 2007. The microRNA miR-124 antagonizes the anti-neural REST/SCP1 pathway during embryonic CNS development. Genes Dev 21, 744-749. Wu, K.Y., Hengst, U., Cox, L.J., Macosko, E.Z., Jeromin, A., Urquhart, E.R., Jaffrey, S.R., 2005. Local translation of RhoA regulates growth cone collapse. Nature 436, 1020-1024. Xie, Z.-R., Yang, H.-T., Liu, W.-C., Hwang, M.-J., 2007. The role of microRNA in the delayed negative feedback regulation of gene expression. Biochem Biophys Res Commun 358, 722-726. Yoon, K., Gaiano, N., 2005. Notch signaling in the mammalian central nervous system: insights from mouse mutants. Nat Neurosci 8, 709-715. Yoshimura, T., Arimura, N., Kawano, Y., Kawabata, S., Wang, S., Kaibuchi, K., 2006. Ras regulates neuronal polarity via the PI3-kinase/Akt/GSK-3beta/CRMP-2 pathway. Biochem Biophys Res Commun 340, 62-68. Zhao, C., Sun, G., Li, S., Shi, Y., 2009. A feedback regulatory loop involving microRNA-9 and nuclear receptor TLX in neural stem cell fate determination. Nat Struct Mol Biol 16, 365-371.

112