Molecular Regulation of Gliogenesis in Xenopus Laevis During Primary

Molecular Regulation of Gliogenesis in Xenopus laevis during

Primary Neurogenesis

------

A Dissertation Presented to

the Faculty of the Department of Biology and Biochemistry

University of Houston

------

In Partial Fulfillment

of the Requirements for the Degree

Doctor of Philosophy

------

Christina Helen Ulrich

August 2019

Molecular Regulation of Gliogenesis in Xenopus laevis during

Primary Neurogenesis

______Christina H. Ulrich

APPROVED:

______Dr. Amy K. Sater, Chair

______Dr. Brigitte Dauwalder

______Dr. Arne Lekven

______Dr. Rachel K. Miller

______Dean, College of Natural Sciences and Mathematics

In memory of my mom, and for my dad.

iii

Acknowledgements

First, I would like to thank my mentor and friend, Dr. Amy K. Sater. She has provided countless hours of teaching and feedback and without her continual support and encouragement I could not have completed my studies. I would also like to thank my committee members, Dr. Brigitte Dauwalder, Dr. Rachel K. Miller, and Dr. Arne Lekven.

They have provided valuable insight and asked questions that opened new avenues of thinking in my project. I would also like to recognize Dr. Robert Schwartz, he provided support and encouragement during difficult periods of my graduate studies.

I am grateful to my family; my parents, Nancy and Sam Ulrich, and my siblings,

Samantha, Marta, and Ben, who patiently listened to me practice my science communication on them. Without their love and support, I would never have started or finished graduate school. I want to thank my incredibly supportive partner Jon, his love of science and unflagging enthusiasm offsets my more pessimistic view of the world. I want to acknowledge Ella, who never passed up an opportunity to keep me company in lab and is an unparalleled frog detector.

I would like to express my deepest thanks to my colleagues and the members of my lab, both past and present. Dr. Vrutant Shah, Dr. Ruth Ritter, Ray Torres, Melissa

Zamora, Sydnee Eldridge, and Mahmoud Al Homouz have all provided wisdom, conversation, and assistance whenever I needed. Finally, I would like to recognize and thank the undergraduates I have been lucky enough to mentor, Anna Subonj, Devanshi

Singh, Frida Mora, and Eric Murphy.

Molecular Regulation of Gliogenesis in Xenopus laevis during

Primary Neurogenesis

------

An Abstract of a Dissertation

Presented to

the Faculty of the Department of Biology and Biochemistry

University of Houston

------

In Partial Fulfillment

of the Requirements for the Degree

Doctor of Philosophy

------

Christina Helen Ulrich

August 2019

Abstract

Glia play essential roles in the vertebrate nervous system. In adults, glia can become mis-regulated, leading to the formation of tumors, including gliomas and glioblastoma (GBM). Understanding glial development is integral to elucidating the mechanisms responsible for the formation and progression of GBM. However, the signals required for initiation of glial development have yet to be completely characterized.

Our preliminary studies in Xenopus laevis indicate that glial specification is initiated between Nieuwkoop and Faber (NF) stages 16-18 in the anterior spinal cord.

Glial specification progresses posteriorly down the neural tube, occurring anterior of the midbrain-hindbrain boundary only at NF stage 24. Gene expression analysis on isolated explants consisting of either mid-gastrula neural ectoderm or animal cap ectoderm overexpressing the BMP inhibitor Noggin demonstrated that neural plates transcribed glial associated genes at levels akin to those of their whole embryo siblings. In contrast, the expression of glial genes in Noggin-injected animal caps (NogAC), such as olig3, sox10, glast, and glt-1, was significantly reduced. These findings suggest that (1) inhibition of BMP signaling alone is insufficient to induce gliogenesis; and (2) signals from the dorsal mesoderm during gastrulation are required for the initiation of gliogenesis at later stages. Pair-wise comparisons between the transcriptomes of mid-gastrula (NF

Stage 11) and mid-neurulation (NF Stage 18) neural plates and NogACs elucidated underlying differences accounting for their distinct developmental potentials. Functional annotation of the differentially expressed genes revealed that glial transcription factors such as pou3f2, sox9, sox10, olig3, and rfx4 and members of Wnt, FGF, and RA signaling

pathways were differentially expressed in neural plates. Pharmacological inhibition of these signaling pathways in neural plates suggested distinct functions in glial differentiation. Ectopic expression of TAZ, the master regulator of mesenchymal differentiation in GBM, in X. laevis embryos encouraged persistence of an undifferentiated neural progenitor population, and potentially inhibited glial specification.

Our results indicate that signals important for gliogenesis are active in the Xenopus embryos as early as mid-gastrulation and have direct implications in further

GBM research.

vii

Table of Contents

Chapter I: Introduction………………………………………………………….…………1 1.1 Overview of Glial Cells……………………………………………………………. 2 1.2 Glial Development and Timing of the Gliogenic Switch…………………………...6 1.3 Xenopus laevis as a Developmental Model for Gliogenesis……………………….11 1.4 Xenopus Neural Development and Specification………………………………….12 1.5 Extrinsic Signals in Gliogenesis…………… ……………………………………15 1.6 Glia in Disease……………………………………………………………………..17 1.7 Project Outline……………………………………………………………………..21 Chapter II: Materials and Methods………………………………………………………23 2.1 Xenopus laevis Care and Embryo Culture…………………………………………24 2.2 Microinjection of Xenopus laevis Embryos………………………………………..25 2.3 Explant Isolation and Culture……………………………………………………...27 2.4 Cloning…………………………………………………………………………….28 2.4a Fragment Cloning…………………………………………………………..28 2.4b Vector Construction………………………………………………………...32 2.4c Plasmid DNA Isolation……………………………………………………..33 2.4d Sequence Confirmation.…………………………………………………….34 2.4e TAZ Subcloning…………………………………………………………….35 2.5 RNA Isolation……………………………………………………………………...35 2.6 cDNA Synthesis……………………………………………………………………38 2.7 Quantitative RT-PCR………………………………………………………………40 2.8 RNA Synthesis……………………………………………………………………..44 2.8a Plasmid DNA Linearization………………………………………………...44 2.8b In situ Probe Synthesis……………………………………………………...45 2.8c Capped RNA Synthesis……………………………………………………..46 2.9 In situ Hybridizations……… …………………………………………………….47 2.10 Sectioning, Imaging, and Image analysis………………………………………...51

viii

2.10a Sectioning………………………………………………………..………..51 2.10b Imaging………………………………………………………………...….52 2.10c Image Analysis……………………………………………...……………..53 2.11 Poly(A) Selected Total RNA Library Preparation………………………………..53 2.12 Transcriptome Sequencing Analysis……………………………………………...55 2.13 Small Molecule Inhibition of Wnt, FGF, RA, and Notch Signaling in NP………61 Chapter III: Initiation of Gliogenesis in Xenopus laevis…………………………………62 3.1 Overview…………………………………………………………………………...63 3.2 Temporal Expression of Glial Associated Genes in X. laevis …………………….67 3.3 Spatial Expression of Glial Associated Genes in X. laevis……………….………..73 3.4 Summary…………………………………………………………………………...81 Chapter IV: Regulation of the Initiation of Gliogenesis by Extrinsic Signals…………...82 4.1 Overview…………………………………………………………………………...83 4.2 Competency of Xenopus Explants in Initiating Gliogenesis……………...………..88 4.3 Sufficiency of sox9 and nfix in Gliogenesis During Primary Neurogenesis...……100 4.4 Discovery Based Identification of Differences Between Noggin Animal Cap and Neural Plate Explants During Gastrulation and Neurulation………….……………...103 4.5 In vivo Verification of Identified Genes and Signaling Pathways in Gliogenesis..130 4.6 Summary………………………………………………………………………….141 Chapter V: Effects of Aberrant Expression of TAZ on neurogenesis and gliogenesis...145 5.1 Overview of Canonical and Aberrant Roles of TAZ in Development and Disease146 5.2 Summary of Preliminary Xenopus laevis Research …………………………..….149 5.3 Replication of 2-cell Stage Injections……...……...……………………………...150 5.4 Effects of TAZ Overexpression on Neurogenesis……..…………………………156 5.5 Impact of Overexpression of Human TAZ on Gliogenesis……….……………...162 5.6 Effects of Overexpression of Human TAZ on Neural Crest Differentiation and Mesenchymal Gene Expression……………..………………………………………..165 5.7 Summary………………………………………………………………………….167 Chapter VI: Discussion…………………………………………………………………169

6.1 Timing of the Gliogenic Switch in X. laevis……………………………………...170 6.2 Competence of Ectodermal Explants……………………………………………..172 6.3 Roles of Signaling in the Gliogenic Switch………………………………………174 6.4 Effects of Aberrant Expression of Taz on Glial Development………...…………177 6.5 Future Directions and Research Implications……………………………..……...178 References………………………………………………………………………………180

List of Figures

Figure 1.1 Early glial specification …………………………………………………..…...5 Figure 1.2 Glial specification transcription factor network ………………………………9 Figure 2.1 Fragment amplification scheme……………………………………………...29 Figure 2.2 cDNA synthesis scheme……………………………………………………...39 Figure 2.3 Quantitative reverse transcriptase reaction scheme…………………………..41 Figure 2.4 Transcriptome pipeline……………………………………………………….55 Figure 3.1 Timing of the gliogenic switch…………………………………………….....66 Figure 3.2 Graph of onset of glial-specific gene expression during Xenopus development………………………………………………………………….…………..68 Figure 3.3 Temporal expression of glial associated genes……………………...……….71 Figure 3.4 Spatial expression of glial associated genes during Gastrulation and neurulation……………………………………………………………………….………76 Figure 3.5 Spatial expression of glial associated genes during tailbud stages………..….78 Figure 3.6 Aldh1l expression in the notochord……………………………………….….79 Figure 3.7 Astroglial development along the anterior-posterior axis…………………....80 Figure 4.1 Experimental summary……………………………………………………….86 Figure 4.2 Astroglial specification in neural plates with or without mesoderm…………90 Figure 4.3 Expression of glast in neural plates and noggin animal caps…………..…….94 Figure 4.4 Effects of bmp4 inhibition on astroglial specification………………………..95 Figure 4.5 Effects of persistent BMP4 inhibition on astroglial specification……………98 Figure 4.6 Gliogenesis experimental system…………………………………………….99 Figure 4.7 Sufficiency of sox9 and nfix to initiate gliogenesis…………………………102 Figure 4.8 Gastrula transcriptome summary……………………………………………106 Figure 4.9 Neurula transcriptome summary……………………………………………108 Figure 4.10 Gastrula highly significantly expressed genes…………………………….112 Figure 4.11 Neurula highly significantly expressed genes…………………………...... 116 Figure 4.12 GO DAVID analysis of differentially expressed genes…………………...122

Figure 4.13 Gastrula principal component analysis of top 510 differentially expressed genes……………………………………………………………………………………126 Figure 4.14 Neurula principal component analysis of top 534 differentially expressed genes……………………………………………………………………………………127 Figure 4.15 Rescue of glial associated genes through overexpression of ascl2………..131 Figure 4.16 Notch inhibition in neural plates…………………………………………..134 Figure 4.17 Wnt inhibition in neural plates…………………………………………….135 Figure 4.18 FGF inhibition in neural plates………………………………………...…..138 Figure 4.19 Preliminary retinoic acid receptor inhibition in neural plates……………..140 Figure 5.1 Experimental injection design……………………………………………....148 Figure 5.2 Comparison of human TAZ sequences……………………………………..153 Figure 5.3 Early neural and gliogenic transcription factor networks…………………...154 Figure 5.4 2-cell TAZ variant injections………………………………………………..155 Figure 5.5 TAZ regulation of neural proliferation and neuronal differentiation………..158 Figure 5.6 Expression of sox2 in the neural plate in TAZ variant embryos…………….159 Figure 5.7 Expression of n-tubulin cells in the neural plate in TAZ variant embryos….161 Figure 5.8 TAZ regulation of glial differentiation………………………………………164 Figure 5.9 Effects of TAZ expression on mesenchymal and neural crest genes………..166 Figure 5.10 Expression of the neural crest marker twist in TAZ variant embryos……...166 Figure 6.1 Timing and contributing factors in Xenopus astroglial development……….176

xii

List of Tables

Table 2.1 Microinjection Concentrations and Experiments……………………………..26 Table 2.2 Fragment Amplification Reaction Mix………………………………………..29 Table 2.3 Cloned Genes of Interest………………………………………………………31 Table 2.4 cDNA Synthesis Mix I…………………………………………………….…..38 Table 2.5 cDNA Synthesis Mix II………………………………………………….……39 Table 2.6 Quantitative RT-PCR Mix……………………….……………………….…...40 Table 2.7 Quantitative RT-PCR Primers………………………………………………...42 Table 2.8 Plasmid Linearization Mix…………………………………………………….44 Table 2.9 Plasmids Used in RNA Synthesis…………………………………...………...47 Table 2.10 Summary of Poly(A) Selected Total RNA Sequencing Run Results………..56 Table 2.11 Software Used for Transcriptome Processing………………...……………..57 Table 4.1 Genes Involved in Neural Specification and Differentiation……………….…87 Table 4.2 Top 30 Differentially Expressed Genes in Gastrula Neural Plate…………...113 Table 4.3 Top 30 Differentially Expressed Genes in Gastrula NogAC…………….…..114 Table 4.4 Top 30 Differentially Expressed Genes in Neurula Neural Plate……………117 Table 4.5 Top 30 Differentially Expressed Genes in Neurula NogAC………………...118 Table 4.6 Expression of Nuclear Factor 1 Genes in Neurula Transcriptome…………..119 Table 4.7 Differntial Expression of Genes Associated with Neural Development and Anterior-Posterior Identity and Development………………………………………….123 Table 4.8 Top 50 Genes Accounting for Variation in Gastrula and Neurula Transcriptomes………………………………………………………………………….128

xiii

Chapter I

Introduction

1.1 Overview of Glial Cells

Glia are the supportive cells in the central and peripheral nervous systems. They are non-conductive cells that support neuronal function in adults in multiple ways; they sheathe the axons of neurons to buffer signaling, modulate synaptic signaling, supply nutrients and remove waste from neurons, identify and remove damaged tissue, and isolate neurons while providing structural support within the brain (Temburni & Jacob,

2001). During development, glia are integral to establishing neural networks through pruning of inactive or minimally active synapses (Barres, 2008; Bosworth & Allen, 2017;

Zuchero & Barres, 2015). In the adult central nervous system (CNS), there are five broad categories of glia; astrocytes, oligodendrocytes, ependymal cells, microglia, and radial glia (Falk & Götz, 2017; Hockfield & McKay, 1985; Kremer, Jung, Batelli, Rubin, &

Gaul, 2017; Molofsky & Deneen, 2015). Glial cells typically arise from neuroectoderm, however microglia are mesodermal in origin (Low, 2018). Glia of the peripheral nervous system (PNS) include Schwann cells, and Satellite cells, both of which sheathe axons, but

Schwann cells migrate to the edges of the PNS whereas Satellite cells remain in the ganglia (Bhatheja & Field, 2006; Weider & Wegner, 2017; Zuchero & Barres, 2015). Our studies focused on the glia in the central nervous system, which will be discussed further below.

Radial glia are the main progenitor cells during development for neurons and macroglia (astrocytes, oligodendrocytes, ependymal cells) and contribute to the neural stem cell population in adult mammalian brains (Todd E Anthony, Klein, Fishell, &

Heintz, 2004; DeCarolis et al., 2013; Falk & Götz, 2017; Zuchero & Barres, 2015).

During mouse development, neural progenitor cells transform into radial glial cells starting at embryonic day 9 (E9). This transformation is mediated through activated

Notch signaling in radial glial cells (B. A. Patten, Sardi, Koirala, Nakafuku, & Corfas,

2006) and leads to them extending long processes throughout all layers of the developing

CNS. During this time, radial glial cells begin to express markers such as nestin, fabp7, and glastage (T. E. Anthony, Mason, Gridley, Fishell, & Heintz, 2005; Howard et al.,

2008; Liour & Yu, 2003; Shibata et al., 1997). During development, the processes of radial glial cells provide scaffolding used by differentiating neurons and macroglia to guide their migrations to specific regions of the developing CNS (Howard et al., 2008;

Zuchero & Barres, 2015). Interestingly, some studies suggest that radial glial cells expressing fabp7 are more likely to differentiate into neuronal cell populations, whereas those marked by glast trend towards a gliogenic fate (Howard et al., 2008). Embryonic radial glial cells and proliferating neural progenitors give rise to all of the neuroectodermal lineages of the nervous system and are the cell types in which the gliogenic switch occurs (Deneen et al., 2006; Howard et al., 2008; Zuchero & Barres,

2015). The gliogenic switch is the process where a neural progenitor cell combines the inhibition of neurogenesis and becomes capable of initiating gliogenesis (Deneen et al.,

2006; Tchieu et al., 2019).

Macroglia are the astrocytes, oligodendrocytes, and ependymal cells in the central nervous system, and are grouped together based on their developmental origin. These three cell types are produced following the gliogenic switch, when proliferating neural progenitor cells gain the ability to initiate gliogenesis (Figure 1.1) (Molofsky & Deneen,

2015; Spassky et al., 2005; Zuchero & Barres, 2015). Initially, macroglia were thought 3

simply to provide structural support in the central nervous system; however their functions as currently understood are far more diverse (Kremer et al., 2017).

Oligodendrocytes are a highly migratory group of cells responsible for creating the myelin sheaths which insulate the axons of neurons (Goldman & Kuypers, 2015).

Ependymal cells are ciliated cells lining the ventricles of the brain and spinal cord and are responsible for maintaining brain homeostasis by transporting cerebrospinal fluid and maintaining epithelial integrity at the interior surface of the CNS (Jiménez, Domínguez-

Pinos, Guerra, Fernández-Llebrez, & Pérez-Fígares, 2014; Spassky et al., 2005). Finally, astrocytes are the macroglia with the most diverse set of functions. They regulate brain metabolism, promote and maintain synaptic function, contribute to the formation and maintenance of the blood-brain barrier, and respond to insults to the brain by re-entering the cell cycle and isolating the injured area from healthy tissue (Bushong et al., 2002;

John Lin et al., 2017; Molofsky & Deneen, 2015; Oberheim et al., 2009; Rothstein et al.,

1996; Tabata, 2015). This cell type was initially characterized by its stellate morphology by Michael von Lenhossek in 1891, which is the typical morphology of a fibrous, or white matter, astrocyte (Parpura & Verkhratsky, 2012). The other broad category of astrocyte is a protoplasmic, or gray matter, astrocyte, which have fewer processes than a typical fibrous astrocyte. Astrocytes are the most functionally diverse macroglial cell, which has led to regional specialization in the mammalian brain (Cragnolini,

Montenegro, Friedman, & Mascó, 2018). For instance, astrocytes found in the subventricular zone retain many of the characteristics seen in radial glia, whereas fibrous astrocytes mostly support synaptic function and regulate brain homeostasis (Kriegstein &

Alvarez-Buylla, 2009).

Microglia are the immune system of the central nervous system and initially colonize the developing nervous system in response to apoptotic neurons (Xu, Wang,

Wu, Jin, & Wen, 2016). While macroglia are derived from neural ectoderm, microglia arise from the yolk sac in mammals and the peripheral hematopoietic tissues in amphibians (Low, 2018; Xu et al., 2016). In an adult organism, their phagocytic capabilities render them essential to responding to brain trauma (Low, 2018). As microglia have a separate developmental origin than macroglia, they are outside the scope of this study and will not be discussed further.

Figure 1.1 Early glial specification

Figure 1.1 Lineage of Neural Ectoderm derived cell types. Adapted from Kreigstein et al., 2009.

1.2 Glial Development and Timing of the Gliogenic Switch Rudolf Virchow first introduced the concept of a glial cell in 1856, and originally, they were thought to be derived from mesoderm. However, work by Wilhelm His in the late 1890’s indicated that neurons and glial cells both originated from an ectodermal cell population (Jacobson, 1991). How glial specification and differentiation takes place has undergone a number of iterations, with the most recent theory outlined in Figure 1.1

(Kriegstein & Alvarez-Buylla, 2009).

During gastrulation, three different germ layers are established: ectoderm, mesoderm, and endoderm. The ectodermal germ layer then becomes specified for either epidermal ectoderm or neural ectoderm. This initial specification is caused by the inhibition of bmp4 signaling in ectoderm fated to be neural ectoderm, and active bmp4 signaling in epidermal ectoderm (Lamb et al., 1993; Smith & Harland, 1992). Neural ectoderm, which is first observed in mice at E9 and is also referred to as neural epithelium, is an actively replicating cell population marked by high expression levels of

Sox2 and activated Notch signaling (Grandbarbe et al., 2003; Kriegstein & Alvarez-

Buylla, 2009; Molofsky & Deneen, 2015). This neural epithelium generates the initial set of neurons, and as the neural tube expands in size, neural epithelial cells transform into a radial glial cell population. This proliferating neural progenitor population retains activated Notch signaling and high levels of Sox2 expression, and begins to express the radial glial markers nestin, fabp7, and glast (T. E. Anthony et al., 2005; DeCarolis et al.,

2013; B. A. Patten et al., 2006; Shibata et al., 1997). Early radial glial cells initially divide asymmetrically to make an intermediate neural progenitor population (as identified by low expression of Sox2 and increasing expression of neuronal-specific genes) and

retain a population of radial glia (Kriegstein & Alvarez-Buylla, 2009; Zuchero & Barres,

2015). As the number of neurons increases, the radial glial progenitor population acquires the ability to differentiate into a glial progenitor cell (Molofsky & Deneen, 2015;

Reiprich & Wegner, 2015; Zuchero & Barres, 2015). This acquisition of competence to generate glia is referred to as the gliogenic switch and was originally defined as both the inhibition of neuronal differentiation and the promotion of glial differentiation (Deneen et al., 2006). However, as neuronal differentiation continues throughout the remainder of neurogenesis, the first condition appears too strict. Our lab defines the gliogenic switch as an epigenetic remodeling event, as yet undefined, that allows for the transcription of glial associated genes.

The glial progenitor cell is first observed between E11.5 and E12.5 in mice and will then differentiate into astrocytes and oligodendrocytes (Chaboub et al., 2016; Laug,

Glasgow, & Deneen, 2018). Near the end of embryonic neurogenesis, the radial glial population either differentiates into ependymal cells or into sub-ventricular zone astrocytes (Kriegstein & Alvarez-Buylla, 2009). This model for glial development is broadly consistent across mammalian and avian species (nuances apply to regions of the developing brain); however this isn’t entirely conserved in developing zebrafish (Lyons

& Talbot, 2015). In zebrafish it appears that radial glial-like cells do differentiate into both neurons and oligodendrocytes, however a differentiated population of astrocytes appears to be absent. Radial glial cells are observed in adult zebrafish, and it has been suggested that radial glia specialize to fulfill the many functions that would otherwise be performed by astrocytes (Lyons & Talbot, 2015).

The molecular basis for the mammalian and avian gliogenic switch has been elucidated in the spinal cord and is described in Figure 1.2 (Deneen et al., 2006).

Immediately prior to the gliogenic switch, the notch-inducible transcription factors Hes1 and Hes5 direct an increase in expression of Sox9 in the proliferating neural progenitor cell population (Laug et al., 2018). Sox9 is traditionally known for its role in neural crest development(Cheung & Briscoe, 2003), research has shown that in the developing central nervous system Sox9 fulfills roles in glial development (Kang et al., 2012). The transcription factor Pou3f2 then drives the radial glial populations towards differentiation into glial precursors, which is followed by the Sox9 and pou3f2-dependent expression of the transcription factor nuclear factor 1 A (Nfia) (Deneen et al., 2006; Glasgow et al.,

2017; Kang et al., 2012a; Laug et al., 2018; Stolt et al., 2003). Sox9 and Nfia then initiate a glial- specific transcriptional cascade that results in the specification of glial progenitor cells (Namihira et al., 2009). It is this point that is most commonly defined as the gliogenic switch, and it is marked by the exponential increase in expression of the glutamate transporter, Glast (Shibata et al., 1997). Nfia has been shown to directly initiate transcription of Glast (Deneen et al., 2006). In this glial progenitor cell population, if Nfia is transcriptionally active, either as homodimers or as heterodimers with other binding partners such as pStat3, the glial progenitor cell population differentiates into astrocytes. If high levels of transcriptionally active Sox10, Sox8, or

Olig3 are present, the glial progenitor cell population differentiates into oligodendrocytes

(Glasgow et al., 2014; Stolt et al., 2003). Sox10 directly antagonizes activity of Nfia through directly binding Nfia and keeping it transcriptionally silent (Glasgow et al.,

2014). Sox9 and Nfia are considered necessary and sufficient for initiating gliogenesis and Nfia is considered necessary and sufficient for astrogenesis.

Figure 1.2 Glial specification transcription factor network

Figure 1.2 Simple transcription factor network associated with glial specification in mouse and chick (Glasgow et al., 2017; Kang et al., 2012; Laug et al., 2018; Namihira et al., 2009).

We have determined that the functional ortholog of Nfia in Xenopus laevis is nuclear factor 1 x (nfix). The most recently released genome (X. laevis assembly 9.4 released in Fall 2018) included a gene identified as nfia. However, there is no expression data available for nfia and the identified coding sequence is truncated; therefore, it is likely that the function of nfia is fulfilled by one of the other homologues. There are three other members of this family: nfib, nfic, and nfix. All are expressed in adult brain and eye tissue (Session et al., 2016); however expression of nfib was observed only following NF stage 40, which excludes it as a possible regulator of the gliogenic switch in primary neurogenesis (Session et al., 2016). Nfic and nfix are expressed in parallel, however nfix transcript is expressed at four-fold levels as compared to nfic during primary neurogenesis, leading us to believe that it is the key transcription factor governing glial development.

The timing of the mammalian gliogenic switch, which coincides with expression of glast (Shibata et al., 1997), and is directly preceded by the expression of Nfia (Deneen et al., 2006), is likely governed by epigenetic changes in availability of glial transcriptional targets (Takouda, Katada, & Nakashima, 2017; Weider & Wegner, 2017).

There is evidence that the promoter regions of glial-associated genes are directly silenced through DNA methylation, and that Nfia is thought to promote the demethylation of glial- associated genes such as Olig1 and Gfap (Sanosaka et al., 2017; Tchieu et al., 2019). The promoter regions of the glial transcription factors, Sox9 and Nfia, along with pioneer transcription factors such as Ascl1, undergo a distinct decrease in DNA methylation in their promotor regions preceding the gliogenic switch (Sanosaka et al., 2017; Takouda et al., 2017). 10

While the role of DNA demethylation in the gliogenic switch is best established, other mechanisms of chromatin remodeling may also play roles in gliogenesis. For example, the SoxE factor Sox10 has been shown to recruit chromatin modifying complexes such as HDAC1/2 or the BAF complex, and contributes to the timing of oligodendrocyte differentiation, if not directly to timing of the gliogenic switch (Reiprich

& Wegner, 2015; Weider & Wegner, 2017). One additional way Nfia might regulate the initiation of gliogenesis is to gradually lengthen G1 phase (Tchieu et al., 2019). The authors found that the proportion of time cells spent in each phase contributed to the timing of the gliogenic switch. While great strides have been made in defining the gliogenic switch, enough questions remain that further investigation is required.

1.3 Xenopus laevis as a Developmental Model for Gliogenesis

The African clawed frog Xenopus laevis has been a popular model organism since the early 1900s, when it was found that frogs ovulated when exposed to a pregnant woman’s urine (Nikos, 2012; Sive, Grainger, & Harland, 2010). The hormone responsible for this response is human chorionic gonadotropin, and Xenopus females can be induced to ovulate in a controlled fashion by injecting them with this hormone.

Developmental biologists utilized this characteristic to collect large numbers of externally developing embryos. In addition, the externally developing embryos are simple to inject with capped RNAs, proteins, oligonucleotides, fluorescent dextrans, or other substances, and the embryos are suitable for microsurgery and explant culture (Sive et al., 2010).

Genomic studies are also commonly performed in X. laevis, and conservation of genes and proteins between frogs and humans is roughly 80% (Hellsten et al., 2010). One

challenge of working with Xenopus laevis is that the genome is allotetraploid, with 36 chromosomes. However, a concerted effort by the X. laevis community increased the resources and support for genomic studies, which has rendered genomic studies relatively straightforward (Session et al., 2016). Overall, this set of characteristics makes Xenopus laevis an ideal model organism to study the gliogenic switch, as the current chick and mouse models are difficult to observe directly during the gliogenic switch and obtaining appropriately timed tissue for genomic studies is challenging.

1.4 Xenopus Neural Development and Specification

Xenopus laevis have two stages of neurogenesis: primary neurogenesis spanning from NF stages 14-35 and sary neurogenesis occurring between NF stages 46 until metamorphosis (Henningfeld, Locker, & Perron, 2007; Nieuwkoop & Faber, 1994;

Wullimann, Rink, Vernier, & Schlosser, 2005). Each wave of neurogenesis is characterized by the rapid increase in numbers of neurons and glia (Thuret, Auger, &

Papalopulu, 2015; Wullimann et al., 2005). Primary neurogenesis quickly establishes a simple nervous system from the neural plate in the first 48-50 h post fertilization

(Nieuwkoop & Faber, 1994). The embryonic brain is distinguished from the neural tube by perceptible segregation of the forebrain, midbrain, hindbrain, and spinal cord regions, which are morphologically distinct by NF stage 22-28 (Xenbase). Primary neurogenesis is followed by a period of neural expansion that is referred to as quiescence (Thuret et al.,

2015). Sary neurogenesis expands the embryonic nervous system over a period of several weeks. The initiation of sary neurogenesis is detected by an increase in expression of neurogenic genes and transcription factors, such as Delta1, Ngn1, and NeuroDs, but is

also marked by a distinct lengthening of the cell cycle in neural progenitors between tailbud phase and tadpole phase (Thuret et al., 2015; Wullimann et al., 2005).

Interestingly, despite the limited time period, primary neurogenesis also exhibits a significant increase in cell cycle length in neural progenitors, doubling in time from ~6 h at NF stage 14 to ~12 h at NF stage 35 (Thuret et al., 2015). In the spinal cord, 30% of all spinal cord neurons are born at NF stage 14, whereas only 2% of new spinal cord neurons are born at NF stage 35 (Thuret et al., 2015). We chose to focus our studies on primary neurogenesis, as we were interested in clearly defining factors contributing to the initial gliogenic switch.

One important aspect of primary neurogenesis is the establishment of anteroposterior (A-P) identity. The groundwork of A-P patterning is laid during gastrulation and is partially described by the activation-transformation model. In this model, morphogens released from the organizer during gastrulation activate anterior neural fate in cells through their inhibition of BMP signaling (P. D. Nieuwkoop &

Nigtevecht, 1954; Pieter Dirk Nieuwkoop, 1952). Inhibition of BMP signaling is considered sufficient to initiate neural fate specification (Mehler, Mabie, Zhang, &

Kessler, 1997; Rogers, Moody, & Casey, 2009). The neural-inducing morphogens released from the organizer are the BMP antagonists noggin and chordin, which inhibit bmp4 produced on the ventral side of the embryo (Lamb et al., 1993; Piccolo, Sasai, Lu,

& EM., 1996; Smith & Harland, 1992; Zimmerman, De Jesús-Escobar, & Harland,

1996). The presumptive neural ectoderm, or neural plate, is then posteriorized by further signaling from the organizer, with differing contributions from the Wnt, FGF, and RA signaling pathways (Green, Whitener, Mohanty, & Lekven, 2015; Lamb & Harland, 13

1995; Patapoutian & Reichardt, 2000; Schilling, 2008). In this model, a Wnt signaling gradient accounts for the majority of posteriorization, with regions of high Wnt activity corresponding to presumptive spinal cord and hindbrain, whereas in the forebrain Wnt signaling is actively inhibited (Green et al., 2015; Henrique, Abranches, Verrier, &

Storey, 2015). FGF signaling contributes to posteriorization of the neural plate in conjunction with Wnt signaling (Green et al., 2015). In animal caps, common explants used in Xenopus, joint expression of the neural inducer noggin and fgf2 results in animal caps with anterior to posterior identity and expression of the hindbrain associated gene, egr2 (krox-20) (Lamb & Harland, 1995). However, ectopic expression of FGF signaling in zebrafish forebrain does not induce egr2 expression (Woo & Fraser, 1997). Activity from both Wnt and FGF signaling is complemented by retinoic acid signaling, which contributes to the anterior-posterior identity of developing motor neurons (Mahony et al.,

2011). These signaling pathways work in part through directly initiating transcription of specific genes at transition points in the anteroposterior body plan (Durston, 2019). Wnt and retinoic acid signals are particularly important for setting up the midbrain-hindbrain boundary, and FGF signaling contributes to the transition between portions of the anterior spinal cord located at the neck and body (Durston, 2019; Holder & Hill, 1991; Rhinn,

Lun, Luz, Werner, & Brand, 2005). The activation-transformation model appears to explain the early anteroposterior patterning of the forebrain, midbrain, and hindbrain; however recent research suggests spinal cord cells arise in part from neuromesodermal progenitors during axial elongation (Henrique et al., 2015; Metzis et al., 2018; Polevoy et al., 2019).

Several lines of evidence suggest that A-P axial identity may result from early chromatin modifications, resulting in tissue that is primed to make spinal cord (Metzis et al., 2018). A recent study shows that ectopic expression of dkk1, an inhibitor of canonical

Wnt signaling, ablates expression of hindbrain markers in Xenopus; however expression of spinal cord associated genes was unaffected (Polevoy et al., 2019). The authors also observed that overexpression of bmp4 in animal caps induced spinal cord markers and appears to do this through FGF3 and FGF8. In addition, recent ATAC-seq experiments observed distinct chromatin modifications between spinal cord and other regions of the brain prior to neural induction (Metzis et al., 2018). Together, these support an element of axial identity in specification neural ectoderm. How, or if, neural induction and the establishment of the anterior-posterior axis contributes to glial specification is unknown.

1.5 Extrinsic Signals in Gliogenesis

During development, spatial and temporal specialization of paracrine and juxtacrine signaling contribute to the creation of the incredible diversity of cells types required to make a multicellular organism. How these extrinsic signals influence the specification and differentiation of neurons has been extensively studied; however, how paracrine signaling effects the gliogenic switch is poorly understood (Grandbarbe et al.,

2003; Perrimon, Pitsouli, & Shilo, 2012; Rankin et al., 2018; Rogers et al., 2009a).

Notch signaling plays the most obvious role in timing of gliogenesis, and in the early 2000’s was considered solely responsible for inducing the gliogenic switch

(Grandbarbe et al., 2003). When Notch signaling was disrupted, gliogenesis was lost

(Hatakeyama et al., 2004; Laug et al., 2018; Taylor, Yeager, & Morrison, 2007).

However, precocious Notch signaling does not accelerate the gliogenic switch, indicating that it cannot be solely responsible for the gliogenic switch (Deneen et al., 2006; Park &

Appel, 2003). The current body of research shows that Notch signals maintain the radial glial/stem cell population in the developing nervous system: its effectors Hes1 and Hes5 turn on transcription of radial glial-associated genes such as fabp7, and ablation of these effectors results in precocious differentiation of radial glial cells into neurons (De Rosa et al., 2012; Hatakeyama et al., 2004). Notch signaling contributes to gliogenesis through a two-step process: first, it actively inhibits neurogenesis, and, s, Notch signaling effectors activate expression of glial-specific genes (Grandbarbe et al., 2003; M. L. Lee, Martinez-

Lozada, Krizman, & Robinson, 2017; Park & Appel, 2003; Taylor et al., 2007). The first way Notch signaling effects gliogenesis is in the developing spinal cord, where Hes1 and

Hes5 can initiate transcription Sox9 and Nfia (Deneen et al., 2006; Laug et al., 2018;

Taylor et al., 2007). Notch signaling effectors have also been shown to work in partnership with Nfia to initiate transcription of glial fibrillary acidic protein (GFAP) in a subset of astrocytes (Namihira et al., 2009). Notch signaling in a glial progenitor population encourages them to differentiate into astrocytes, while preventing differentiation into oligodendrocytes (Glasgow et al., 2017; Namihira et al., 2009; Park &

Appel, 2003). In mature astrocytes, expression of the glutamate transporter glt-1 is induced by ependymal cells through a notch-dependent mechanism (M. L. Lee et al.,

2017). Notch signaling is necessary but insufficient for the gliogenic switch; in Xenopus the other components contributing to gliogenesis have yet to be elucidated.

Paracrine signaling has been shown to affect different aspects of gliogenesis.

While inhibition of Bmp signaling is sufficient to initiate neurogenesis, Bmp signaling 16

has established roles in late development and adulthood in directing differentiation of astroglial cells (Bonaguidi et al., 2005; Mabie et al., 1997; Mehler et al., 1997; Srikanth,

Kim, Das, & Kessler, 2014; Zimmerman et al., 1996). In mature sub-ventricular astrocytes, expression of the bmp4 inhibitor noggin inhibits glial differentiation (Lim et al., 2000; Niu et al., 2013). In the developing spinal cord, canonical Wnt signaling orchestrates the timing of oligodendrocyte differentiation and directly inhibits astrogenesis (Dai et al., 2014; Guo et al., 2015; H. K. Lee et al., 2015; Shimizu et al.,

2005; Sun et al., 2019). FGF signaling has been shown to regulate expression of the glial transcription factor Sox9 in the teleost hindbrain, and partners with notch signaling effectors in the mammalian telencephalon to promote the formation of radial glial cells

(Esain, Postlethwait, Charnay, & Ghislain, 2010; Yoon et al., 2004). Retinoic acid (RA) signaling has been reported to promote astrogenesis in E17 mouse embryos, however, at

E13 RA signaling inhibits astrogenesis (Faigle, Liu, Cundiff, Funa, & Xia, 2008). Also,

RA-deficient mouse embryos exhibited decreased numbers of glial precursors (

Kriegstein & Alvarez-Buylla, 2009; Paschaki et al., 2013). How, or if, any of these signaling pathways directly contribute to the initiation of the gliogenesis is unknown.

1.6 Glia in Disease

In the adult central nervous system, neurons are typically in a post-mitotic quiescent state whereas astroglial cells are more capable of re-entering the cell cycle

(Frade & Ovejero-Benito, 2015). Correspondingly, astroglial cells are more susceptible to aberrant function, and their roles in cancer are by the far the best studied. These roles vary greatly, ranging from the degradation of the myelin sheath built by oligodendrocytes

in multiple sclerosis to the roles of astrocyte dysregulation in neurodegenerative diseases and epilepsy (Balasubramaniyan et al., 2015; Chaboub & Deneen, 2013; Chamberlain,

Nanescu, Psachoulia, & Huang, 2016; Dulamea, 2017; John Lin et al., 2017; Shandra et al., 2019; Verkhratsky, 2019).

One of the most pressing examples of aberrant astroglial function is seen in gliomas, the macroglial-derived tumors of the central nervous system (Claes, Idema, &

Wesseling, 2007; Laug et al., 2018). Gliomal cells appear to resemble glial precursor populations, in that they are frequently associated with elevated levels of Sox9, pou3f2, and Nfia; these genes are expressed at correspondingly higher levels in GBMs than in other glioma types (Laug et al., 2018). Further indication that gliomas are akin to glial progenitor cell populations is the presence of activated Notch signaling and pSTAT3, though their effects on tumorgenicity have been shown to be context- dependent

(Giachino et al., 2016; Hai et al., 2018; Stockhausen, Kristoffersen, & Poulsen, 2010,

2012). Notch signaling promotes differentiation and apoptosis in some gliomas, while also encouraging cell proliferation in others. pSTAT3 in conjunction with Nfia can encourage differentiation, however, this combination can also drive tumorigenicity in conjunction with oncogenic factors such as PTEN (Laug et al., 2018).

Gliomas are subdivided by their cell type of origin: ependymomas are derived from ependymal cells, oligodendrogliomas from oligodendrocytes, and astrogliomas from astrocytes. There is a fourth type of glioma, the glioblastoma (GBM), which are partially derived from astrocytomas, and have the worst prognosis by far (Balasubramaniyan et al., 2015; Laug et al., 2018; S. C. Lee, 2018; Maher et al., 2006; Parsons et al., 2008).

GBMs account for 52% of all primary brain tumors and have a median survival rate with all treatment options of 15 months following diagnosis (Balasubramaniyan et al., 2015;

Bi & Beroukhim, 2014). GBMs have four main molecular subtypes; proneural, neural, classical, and mesenchymal. Proneural and neural subtypes are associated with the greatest survival rate (Cohen, Holmen, & Colman, 2013; Hai et al., 2018; Parsons et al.,

2008; Verhaak et al., 2010), whereas mesenchymal subtype GBMs have the poorest patient prognosis (Verhaak et al., 2010). The transcriptional co-activator Taz was found to be a master regulator of mesenchymal differentiation in glioma stem cells (Bhat et al.,

2011) and several transcription factors associated with astroglial differentiation are expressed at very low levels in mesenchymal GBM (Sox10 and Sox8) (Cancer Genome

Atlas Research Network et al., 2013). The relationship between TAZ and these genes is unknown (unpublished data from K. Bhat) and would be an interesting area of study.

Artificially induced differentiation has been suggested as a potential treatment for these tumor types (Balasubramaniyan et al., 2015). Further research is needed, however, to decipher the relationship between differentiation and proliferation genes before these treatments can be designed.

Astroglial cells also play an integral role in the response to traumatic brain injury

(TBI) (Burda, Bernstein, & Sofroniew, 2016; Karve, Taylor, & Crack, 2016). Astrocytes become reactive and signal to microglial cells, the brain’s immune system, following primary brain injury (damage caused by the initial trauma). Reactive astrocytes then play a key role in sary brain injury (damage caused by persistent cellular response to the initial injury) by reinforcing inflammation at and around the area of the initial injury (Ortega-

Pérez & Amaya-Rey, 2018). TBI consists of an initial insult to the central nervous 19

system, which can lead to axonal shearing, damage to blood vessels, disruption to the blood-brain barrier, and outright death of neurons and glia. The mechanical damage of the initial injury signals to the surrounding astrocytes to become reactive by re-entering the cell cycle and migrating to the location of the injured tissue (Karve et al., 2016). Once the astrocytes have reached the site of injury, they signal to microglial cells through chemokines and cytokines (such as signaling through Toll-like receptors, or the stimulation of IFN-γ or TGF-β signaling) and together they remove damaged tissue and form a glial scar surrounding the area (Silver & Miller, 2004). This inflammation response allows the brain to respond quickly to TBI, however, this response can persist inappropriately, eventually damaging healthy tissue and impairing neural function further

(Burda et al., 2016; Karve et al., 2016; Silver & Miller, 2004). Astrocytes promote both the pro-inflammatory response outline above, but also encourage a pro-neuroprotective response simultaneously (Karve et al., 2016). This neuro-protective response is poorly understood, and further research is needed to promote a neuroprotective response following TBI.

Perhaps the most interesting astrocyte-mediated response to TBI from a developmental perspective is the ability of local fibrous astrocytes to de-differentiate into a neural precursor state and re-differentiate into nascent neurons (Götz, Sirko, Beckers, &

Irmler, 2015; Kato, Losada-Perez, & Hidalgo, 2018; Hedong Li & Chen, 2016; J. P.

Magnusson et al., 2014; Jens P Magnusson & Frisén, 2016). It has been widely accepted that subventricular zone astrocytes fulfill many of the functions of embryonic radial glia; however, that all astrocytes might retain a neurogenic potential in traditionally static areas of the brain is highly exciting (Kriegstein & Alvarez-Buylla, 2009). One 20

mechanism for how astrocytes accomplish this centers around decreasing notch signaling, which allows for an increase in Ascl1 activity (Jens P Magnusson & Frisén, 2016).

Currently, the number of astrocytes reprogramming themselves into neurons is functionally insignificant; more complete understanding of the gliogenic switch could lead to controlled reprogramming of astrocytes in adults following injury.

These examples illustrate how imperative proper function of astroglial cells are in vertebrates.

1.7 Project Outline

In this project, we aim to answer several questions that will allow us to describe the initiation of gliogenesis in Xenopus laevis. The first question pertains to the timing of the gliogenic switch during primary neurogenesis. Xenopus laevis is a new model organism for studying glial development, and this will allow us to pinpoint the temporal frame of the gliogenic switch. Once we have identified the timing of the gliogenic switch, we can then exploit the strengths of X. laevis and identify the signaling conditions that contribute to the gliogenic switch. While Notch signaling plays an integral role in the gliogenic switch, how signals from the organizer contribute to gliogenesis is unknown.

While pursuing the answer to these questions, we will also be able to address the question of conservation of glial development among vertebrates; currently, a divide is observed between higher vertebrates (mouse and chick) and teleosts. My final question addresses how aberrant expression of the transcriptional co-activator TAZ affects neuronal and glial development. Expression of TAZ in GBMs have extremely deleterious effects on the

patient; however, how TAZ would effect neural development in a healthy embryo is unknown.

To answer these questions, we employed a combination of molecular and transcriptomic techniques. Using in situ hybridizations and quantitative RT-PCR, we identified the gliogenic switch as occurring sequentially starting as early as NF Stage 16-

18, beginning at the very posterior hindbrain and anterior spinal cord. The gliogenic switch does not occur in the presumptive forebrain and midbrain until NF stage 24. By observing the transcriptomes of two different explants, we identified a factor from the spinal cord that appears to be necessary to initiate gliogenesis in the very anterior portions of the nervous system. Using small molecules, we have excluded Wnt signaling as a candidate, and suggested FGF or RA signaling as potentially contributing to the differentiation of astrocytes. While pursuing this, we also found evidence of divergence between lower and higher vertebrates in the initiation of the gliogenic switch. Ectopic expression of the transcription factors sox9 and nfix failed to initiate gliogenesis, suggesting that they are not sufficient for gliogenesis in X. laevis. Interestingly, we found that aberrant TAZ expression appears to encourage an undifferentiated neural state and may actively discourage glial differentiation by significantly lowering levels of transcription factors nfix and sox10. In conclusion, we have defined the timing of gliogenesis in X. laevis, and have started to elucidate extrinsic factors that contribute to it.

Chapter II

Materials and Methods

2.1 Xenopus laevis Care and Embryo Culture

Adult Xenopus laevis animals were housed in an automatic recirculating housing system procured from Aquaneering, Inc, (Model XS360) grouped in 10-12 per tank, and maintained on a diurnal light cycle. Input water was filtered and adjusted to ideal pH and conductivity using bicarbonate and sea salt. Animals were fed three times per week using frog brittle (cat # SA05961(LM), NASCO, Inc).

In vitro fertilization was accomplished as described in Sive, Grainger, & Harland,

(2010). Testes were isolated from anesthetized (0.05% w/v tricaine (ethyl 3- aminobenzoate methanesulfonate salt)(cat # A5040,Sigma) and sacrificed (through cervical dislocation) male X. laevis and maintained in filtered 1X Marc’s Modified

Ringer Solution (MMR, 100 mM sodium chloride, 1 mM magnesium chloride, 20 mM potassium chloride, and 5 mM HEPES; adjusted to pH 7.5 using sodium hydroxide) supplemented with 10% w/v normal rabbit serum (cat #R4505, Sigma) and 1% w/v antibiotics/antimycotics (penicillin, streptomycin, and amphotericin (PSA, cat # A-9909,

Sigma). Testes were stored at 4 °C for two to three weeks and the media was changed every seven days. Ovulation was induced in adult female X. laevis through injection of

300 units of human chorionic gonadotropin (hCG) (Sigma CG-10, Sigma). Injected females produced oocytes sixteen h after hCG induction in an 18 °C incubator, for 8 to 10 h. Adult females were manually encouraged to lay their oocytes, and collected oocytes were fertilized with macerated, previously collected testes. Fertilized embryos were cultured in 1/3 X MMR (pH of 7.6 using sodium hydroxide) at either room temperature

(RT) or 18 °C until developing until their desired stages. Staging was determined by

Nieuwkoop and Faber staging (Pieter Dirk Nieuwkoop, 1952).

2.2 Microinjection of Xenopus laevis Embryos

The peptidoglycan jelly coat surrounding the Xenopus laevis embryos was removed by washing with 3.5% cysteine (cat # C7880, Sigma, pH 7.5-7.8) for up to five min and rinsed 3x with 1/3 MMR prior to microinjection. Xenopus embryos were bilaterally injected at either 2-cell or 8-cell dorsal blastomeres (cells fated to contribute neural ectoderm) in 4% filtered ficoll (cat # F9378, Sigma) with a tapered glass injection needle

(external diameter 1.0 mm and inner filament diameter 0.58 mm, needles were shaped by

PUL-1 needle puller (both items supplied by World Precision Instruments). Needles were filled either back filled or allowed to passively fill through capillary action. Embryos were either injected with 5 nanoliters (nl) per cell at the 2-cell stage or 3 nl per cell at the

8-cell stage. Injected embryos were allowed to develop to the desired stage. Correctly injected embryos were collected for further experimentation.

To induce anterior neural cell fate, we bilaterally injected equal amounts of capped noggin RNA, an extracellular BMP antagonist, into each cell at the two cell stage in X. laevis embryos. To test whether the sox9 and nfix could induce glial development we bilaterally injected the two cell stage embryos with each previously mentioned capped

RNA either individually or together, and with noggin capped RNA either individually or together. Injection scheme was thus; sox9, nfix, sox9+nfix, noggin + sox9, noggin + nfix, and noggin + sox9+nfix. Embryos were injected with potential pioneering transcription factor ascl2 in conjunction with noggin (both capped RNA) to investigate ascl2’s ability

to initiate gliogenesis. Injected embryos were allowed to develop until either stage 18

(total RNA sequencing experiments) or stage 28 (quantitative RT-PCR experiments).

Targeted injections of capped RNA synthesized from human TAZ and it’s variants were performed at X. laevis’ 8-cell stage. The two dorsal blastomeres fated to give rise to the anterior neural plate were bilaterally injected. Injection solutions are as follows: WT TAZ, silent TAZ, or active TAZ each in addition to 0.25% Alexa-Dextran

488 (cat # D22910, Thermo Fisher). Incorrectly injected embryos (embryos with fluorescence in places other than the neural plate) were removed from the sample pool.

Correctly injected embryos were collected at NF stage 18.

Table 2.1 Microinjection Concentrations and Experiments

2.3 Explant Isolation and Culture

Animal caps were created from embryos bilaterally injected at the 2-cell stage with either noggin capped RNA or some combination of capped RNAs (See previous section). Injected embryos were allowed to develop until mid-blastula stage (stage 8). At mid-blastula, vitelline envelops were manually removed using sharp and fine forceps.

The animal cap, the animal pole region of the blastula embryo directly above the blastocoel, was dissected off the injected embryos and transferred to a new dish with filtered 1X MMR. After all necessary animal caps were dissected 1% w/v of PSA was added to the dish. Animal caps were cultured in the previously mentioned media until reaching their desired stage at either RT or 18 °C.

Neural Plates ± Mesoderm were isolated from uninjected embryos developed until early (stage 10.5, neural plate) or mid-gastrulation (stage 11, neural plate + mesoderm).

At the desired stage vitelline envelopes were manually removed from the embryo using sharp and fine forceps. Using an eyebrow hair knife cuts were made bracketing the dorsal lip and extended upward on the animal pole until the middle (peak) of the animal pole. A s set of cuts was made across the middle of the animal pole, followed by a third set of cuts 15 cell diameters above the dorsal lip for the neural plate or directly above the dorsal lip for the neural plate + mesoderm. Isolated neural plate ± mesoderm were cultured in

1X MMR + 1% w/v PSA until the desired stage at either RT or 18 °C.

Explants were grown to NF stage 18 for Total RNA Sequencing or stage 28 for quantitative RT-PCR. Staging was determined by observation of whole embryo siblings cultured under the same conditions as the animal cap or neural plate ± mesoderm

explants. For Total RNA Sequencing 75-100 animal caps and 60-75 neural plates were cut and cultured per biological replicate, for quantitative RT-PCR 15-20 animal caps and

12-15 neural plates ± mesoderm were cut and cultured per biological replicate.

A detailed description of both animal cap isolation and neural plate ± mesoderm isolation are detailed in the microdissection chapter of Early development of Xenopus laevis : a laboratory manual (Sive et al., 2010).

2.4 Cloning

2.4a Fragment Cloning

Primers targeting the gene of interest were designed after identifying the correct sequence from Xenbase and NCBI and subsequently ordered from Sigma Aldrich. The size of the amplified fragment from the gene of interest’s sequence was determined by its purpose; primers designed to make in situ hybridization probes amplified fragments between 300-500 bps, and primers used to clone the entire gene amplified fragments specific to the gene of interestage Amplification of the target sequence occurred from temporally and spatially specific cDNA. Expression of the gene of interest in the specific cDNA was predicted during the initial sequence selection and primer design, starting on

Xenbase. To amplify the target gene the following reaction (Table 2.2) was set up:

Table 2.2 Fragment Amplification Reaction Mix

Substance Volume (Concentration) cDNA 250 ng/reaction 10x Reaction Buffer 10 µl 25 mM MgCl2 4 µl (1mM) 10 mM dNTPs 5 µl (0.5 mM) 10 mM Primer Mix (F&R) 5 µl (0.5 mM) DNA Polymerase (Taq/Q5 etc.) 1 µl Nuclease Free Water Remainder Total Volume 100 µl

Suitable polymerase for each reaction was determined based off the size of the fragment. For short fragments, ≥1000 bps fragment, the Taq polymerase was sufficient

(cat # 10966-018, Invitrogen). For fragments longer than 1000 bps the high-fidelity polymerase Q5 (cat # M0491S, NEB) was chosen. Amplification of target fragment was accomplished using a Mastercycler nexus GX2 (material # - 2231000289, Eppendorf).

The thermocycler program conditions were tailored to the fragment getting amplified, with an example thermocycler program listed below (Figure 2.1).

Figure 2.1 Fragment amplification scheme

Figure 2.1 Thermocycler reaction scheme to amplify fragment DNA.

Annealing temperature was determined based off the length and concentration of the primers used for amplification and confirmed using two online Tm calculators. The

Invitrogen Tm calculator (https://www.thermofisher.com/us/en/home/brands/thermo- scientific/molecular-biology/molecular-biology-learning-center/molecular-biology- resource-library/thermo-scientific-web-tools/tm-calculator.html) was used for Platinum

Taq polymerase and the NEB Tm calculator was used for Q5 polymerase

(https://tmcalculator.neb.com/#!/main). Elongation time was determined by the length of the fragment being amplified, with the target of every 1000 bps requires 1 min of elongation time.

Once the thermocycler reaction was complete, 10% (10 µl) of the reaction mix was run on a 1.2% agarose gel gel (0.7 g agarose in 60 ml 1X TAE buffer + 3 µl 1% ethidium bromide) for 30-60 min at 90 vs and visualized using a NucleoTech

NucleoVISION UV Cabinet coupled with ImageJ. Promega’s 100 bp or 1 kb DNA fragment ladders (cat #’s G2101 and G5711) were added onto the get to confirm the length of the amplified nucleotide fragment.

Once the fragment size had been confirmed, the remaining 90 µl of the reaction mix was cleaned using the Sater lab In-House DNA Purification Method. First, 110 µl of nuclease free water was added to the remaining reaction volume. Next, an equal amount

(200µl) of phenol:chloroform (Sigma) was added to the reaction mix and rapidly vortexed for 15 s. Immediately following vortexing the mix was spun at 14,000xg for 2 min at room temperature. The top aqueous layer was transferred to a sterile 1.7 ml

Eppendorf tube and two volumes of 100% ethanol and 1/10 volume of 3 M sodium

acetate and 1µl of glycoblue (cat # AM9515, Invitrogen) was added to the aqueous layer.

Sodium acetate is added to neutralize phosphate backbone sugars and make the fragment

DNA less soluble in water. This process, referred to as “salting out”, is conserved across different types of nucleotide cleaning (ie, RNA, plasmid DNA). The mix was incubated at -20 °C for at least an h and followed by DNA precipitation through a 30-min spin at

14,000xg at 4 °C. The DNA fragment pellets, marked by glycoblue, were washed at least twice in 80% ethanol and resuspended in 15 µl nuclease free water.

Alternatively, if multiple DNA bands were amplified, the remainder (90 µl) of the reaction mix was run on a 1.2 % agarose gel. After running the gel for 30-60 min at 90 vs, the band corresponding to the predicted fragment size was excised from the gel using a new razor blade. The fragment DNA was then extracted from the gel using the

Zymoclean Gel DNA Recovery kit (cat # D4007, Zymo Research) or an equivalent kit according to the manufacturer’s specifications.

Table 2.3 Cloned Genes of Interest

2.4b Vector Construction

Amplified DNA fragments described in the previous section were ligated into a

TOPO backbone (TOPO TA (cat # K450001, Invitrogen) for Taq amplified fragments, and TOPO Blunt (cat # K2800J10, Invitrogen) for Q5 amplified fragments) according to the manufacturer’s protocol. Optimal conditions for fragment insertion were determined using the ligation calculator created by the University of Duesseldorf

(http://www.insilico.uni-duesseldorf.de/Lig_Input.html) which took into account the length of the backbone, the concentration of the backbone, the length of the fragment, and molar ratio of backbone to fragment (1 to 3 was typically used.) The newly constructed vector (target DNA + TOPO backbone) was transformed into competent

DH5α E.coli for plasmid amplification. Transformation protocol was as follows: competent DH5α cells were thawed on ice for 10 min, or until completely free of ice, before use. 1-5 µl of the constructed vector was introduced for 50 µl of competent DH5α cells and incubated on ice for 30 min. Following incubation the competent cells + vector were heat shocked for 30 s at 42 °C to allow for plasmid uptake by the bacterial cells.

The transformed cells were then allowed to recover in nutrient rich SOC media (2% w/v tryptone, 0.5% w/v yeast extract, 8.56 mM NaCl, 2.5 mM KCl, 10 mM MgCl2, and 20 mM glucose) for one h at 37 °C with shaking. TOPO TA is resistant to ampicillin and its derivatives and TOPO Blunt is resistant to kanamycin and zeocin. The competent cells + vector were then plated on a pre-warm LB plate containing the backbone specific antibiotic (100 µg/ml ampicillin or 50 µg/ml kanamycin). The plates were then grown overnight at 37 °C. Colonies observed the next day were counted and individually grown

in 3 ml LB broth at 37 °C with shaking and selective antibiotic (final concentration: 100

µg/ml ampicillin OR 50 µg/ml kanamycin) for 16 h.

2.4c Plasmid DNA Isolation

The Sater lab has two different methods to isolate DNA; the first method results in a lower concentration but cleaner product through using a commercially supplied kit, such as the DNA Clean and Concentrator kit (cat # D4013, Zymo Research). The s method, the in-house plasmid isolation protocol, results in larger concentrations of DNA, but a less clean product. Plasmid isolation was performed according to the manufacturer’s protocol if the first method was chosen, and the in-house method is an adaptation of the alkaline lysis method initially described in Molecular Cloning. A

Laboratory Manual (T Maniatis, 1982). Before plasmid isolation occurred with either method, a glycerol stock was made with 550 µl liquid culture and 450 µl 60% filtered glycerol (Sigma). Glycerol concentration in bacterial stocks is optimized between 20-

30%, with the previous calculation glycerol concentration is at 27%. Bacterial glycerol stocks are stored at -80 °C.

Once a bacterial stock had been made, the remaining liquid culture was spun at

14,000x g to pellet the bacterial cells. The supernatant was removed and the pellet was suspended in 100 µl cold Solution I (50 mM glucose, 25 mM Tris-HCl pH 8.0, and 10 mM EDTA pH 8.0). The bacterial cells were lysed with the addition of Solution II (0.2

N NaOH and 1% sodium dodecyl sulfate or SDS) and gently inverted six times to avoid

DNA shearing. Following inversion, the mix was incubated at RT to allow for complete lysis to occur. Cold Solution III (5M potassium acetate and 1M glacial acetic acid) was

then added (150 µl) to the lysed mix and vortexed to initiate precipitation of cell debris.

After vortexing for 15 s the mix was spun at 14,000x g for 5 min to pellet the cell lysate.

The supernatant was transferred into a new 1.7 ml Eppendorf tube and 2 µl of 20 mg/ml

RNAse A (Sigma) was added. The mix was then incubated at 37 °C for 15 min, followed by the addition of an equal volume of phenol:chloroform. The mix was vortexed, spun at14,000x g for 2 min, and the top aqueous layer was transferred to a new 1.7 ml

Eppendorf tube. An equal volume of 100% ethanol and 1/10 of a volume of 3M sodium acetate was added to the aqueous layer and incubated at -20 °C for at least one h. Plasmid

DNA was precipitated through spinning the mixture for 30 min at 14,000xg in 4 °C. The plasmid DNA pellet was then rinsed with 80% ethanol three times and resuspended in 20-

40 µl of nuclease free water (depending on pellet size). The resuspended pellet was then nano-dropped for concentration and purity using an IMPLEN NanoPhotometer.

2.4d Sequence Confirmation

Individual sequencing was outsourced to Lone Star Labs (http://www.lslabs.com/) according to the company’s specifications. Sequencing was performed using either universal primers known to exist on the backbone of the plasmid being sequenced (i.e.

TOPO plasmids have M13 forward and M13 reverse bracketing the multiple cloning sequence) or specific primers for the gene of interestage Sequencing results and chromatograms were received the following day. Sequence chromatograms were observed using SnapGene Viewer Software (GSL Biotech; available at snapgene.com).

Sequence files were read either with SnapGene Viewer or Serial Cloner 2.6.1

(http://serialbasics.free.fr/Softwares.html). Sequences were confirmed by first removing any sequence formatting with The Lab Notebook’s DNA tools 34

(http://www.thelabnotebook.com/sequence.php) followed by comparing them to their reference sequence using T-COFFEE Multiple Sequence Alignment tool

(http://tcoffee.crg.cat/apps/tcoffee/do:mcoffee) (Di Tommaso et al., 2011).

2.4e TAZ Subcloning

Human TAZ WT (Accession # - NM_015472.5) and variants (4SA (TAZ-active) and 4SA + S51A(TAZ-silent)) in pBABE plasmid was obtained from K. Bhat’s lab. TAZ was removed from pBABE plasmid using the restriction enzymes BamHI and EcoRI (cat

# R0136S and R0101S respectively, NEB) and ligated into pCR™-Blunt II-TOPO®

(cloning kit cat # 450245, Thermo Fisher) that had been previously digested with the same restriction enzymes. The ligation was performed according to NEB’s T4 ligase (cat

# M0202S) protocol. TAZ Topo Blunt plasmid sequences were then confirmed through sequencing at Lonestar Labs (Houston,TX).

2.5 RNA Isolation

Whole embryos or animal caps and neural plates were collected at the appropriate stage in 1ml of TRIzol (cat # 15596026, ThermoFisher Scientific) in a new 1.7 ml

Eppendorf tube after all remaining growth media had been removed. Samples were stored at -80 °C until at least three biological replicates were obtained. One biological replicate is defined as such for quantitative RT-PCR experiments; n=5-7 whole embryos, n=15-20 animal caps, and n=12-15 neural plates ± mesoderm. One biological replicate for total RNA sequencing experiments are defined in Section

2.3. After isolation was complete, all RNA for every sample and experiment type was stored at -80 °C indefinitely.

Samples used in quantitative RT-PCR experiments were removed from -80

°C and thawed on ice. Once samples were completely thawed they were incubated at RT with mild vortexing to disrupt the nucleoprotein complexes for five min.

Following this, 200 µl chloroform (Sigma) per ml of TRIzol was added to the mixture. The TRIzol-chloroform mixture was vortexed for 15 s and centrifuged at

12,000x g for 15 min at 4 °C. The top aqueous layer containing the partially cleaned RNA was transferred to a new 1.7 ml Eppendorf tube and 1/10 volume of chloroform was added to the mixture, vortexed for 15 s, and centrifuged for 10 min at 12,000xg at 4 °C. The top aqueous layer was again transferred to a new 1.7 ml Eppendorf tube and an equal volume of 100% isopropanol and 1/10 volume of

3 M sodium acetate was added to the mix. The entire mixture was incubated for at least one h at -20 °C, and after the completion of the incubation was spun at

12,000 x g for 30 min at 4 °C to precipitate the RNA. Pelleted RNA was then washed three times in 80% ethanol diluted in DEPC (Sigma) and autoclaved water.

The pellet was resuspended after air drying in 10-25 µl of nuclease free water depending on the size of the pellet. Resuspended RNA was then nano-dropped on using an IMPLEN NanoPhotometer for concentration and quality.

Following resuspension in nuclease free water 2 µg of RNA was treated with TURBO DNase (cat # AM2238, ThermoFisher Scientific) for 30 min at 37 °C according to the manufacturer’s specifications. Following DNase treatment, 1 µl of 0.5 M EDTA was added to inactivate the TURBO DNase and the samples were centrifuged at 12,000 x g for 3 min to pellet extra salt. The supernatant was

transferred into a new, RNase free Eppendorf tube and nano-dropped using the machine listed above for concentration and quality.

Isolation of RNA for total RNA sequencing was accomplished one of two ways; the first method used the Direct-zol RNA Miniprep Kit (cat #11-330, Zymo

Research) to isolate RNA before cDNA library preparation, the s method initially isolated RNA through the in-house preparation mentioned above then cleaned the

RNA using RNA Clean & Concentrator Kit (cat # R1013, Zymo Research). The first method was faster and retained a higher percentage of RNA, the s method produced RNA of higher quality. Samples were nano-dropped for concentration and purity, then sent to Seq-N-Edit, the University of Houston’s sequencing core, for quality analysis. Each RNA sample was assigned a RNA Integrity Number

(RIN) showing level of sample degradation. RNA samples with RINs above 7 are acceptable to use in further sequencing projects. All samples were placed at -80 °C for long term storage.

2.6 cDNA Synthesis

Whole embryo cDNA libraries used for quantitative RT-PCR experiments were created using 500 ng of DNased RNA (see previous section for RNA isolation description.) RNA isolated from animal cap or neural plate ± mesoderm samples was directly used to create cDNA libraries and were created using 500 ng of RNA. All heating and cooling steps were accomplished using a Mastercycler nexus GX2 (material

# - 2231000289, Eppendorf.)The following reaction mix, Mix I, was set up for each sample in a 0.6 ml Eppendorf tube:

Table 2.4 cDNA Synthesis Mix I

Substance Volume (Concentration) RNA 500 ng 50 µM Random Hexamers 2.5 µl (6.3 µM) (cat # N8080127, Thermo Fisher) Oligo dT (12-18 bps) 1 µl (cat # 18418012, Thermo Fisher) 10 mM dNTPs 1 µl (0.5 M) (cat # 10297018, Thermo Fisher) Nuclease free water Remainder Total Volume 13 µl

Mix I was then incubated for five min at 65°C after which it was cooled on ice for at least two min. During this incubation period the reagents were thawed for Mix II, and upon completion the following reagents were added directly into the 0.6 ml Eppendorf tube containing Mix I:

Table 2.5 cDNA Synthesis Mix II

Substance Volume (Concentration) Mix I 13 µl 5X RT Buffer 4 µl (1X) 0.1 M DTT 1 µl (5mM) 200 units/µl Superscript III 1 µl (10 units/µl) (cat # 18080093, Thermo Fisher) RNase OUT 1 µl (cat # 10777-019, Thermo Fisher) Total Volume 20 µl

Once Mix I and Mix II had been combined the 20 µl reaction was run through the cDNA synthesis thermocycler program consisting of the following incubations: 5 min at

25 °C, 60 min at 50 °C, 15 min at 70 °C, then held at 4 °C. The mix was then nano- dropped for concentration and quality and either used immediately or stored at -20 °C indefinitely. The following reaction scheme (Figure 2.2) illustrates the thermocycler program used.

Figure 2.2 cDNA synthesis scheme

Figure 2.2 Thermocycler reaction scheme to create a cDNA fragment library from an RNA sample.

2.7 Quantitative RT-PCR

cDNA was diluted to a concentration of 200 ng/reaction and performed in either duplicate or triplicate for each gene of interest observed for each sample. Expression levels were identified through target primers specific to each gene of interest that were designed to amplify 70-150 bps of the gene of interest’s mRNA reference sequence.

Reference sequences were identified through Xenbase and NCBI, primers were designed between melting temperatures 57-63 °C with optimal Tm set to 60 °C, and primers were ordered from Sigma Aldrich. Table 2.7 lists the quantitative RT-PCR primer sequences, accession numbers, and amplicon lengths. The quantitative RT-PCR experiments were performed using Applied Biosystems StepOnePlus Real-Time PCR System coupled with

StepOne software version 2.3. Each sample and target gene well consisted of a 10 µl reaction that consisted of the following reagents:

Table 2.6 Quantitative RT-PCR Mix

Substance Volume (Concentration) 2X SYBR Select Master Mix 5 µl (1X) (cat # 4472908, Thermo Fisher) 10 mM Target Gene Primer (F & R) 1 µl (1 mM) 200 ng/µl cDNA 1 µl (20 ng/µl) Nuclease Free Water 3 µl Total Volume 10 µl

The quantitative RT-PCR reaction consisted of a 2-min 50 °C warm up, 40 cycles of 95 °C denaturation step (15 s) followed by 60 °C annealing and elongation step (60 s) and culminated in either a final dissociation or melt curve stage (Figure 2.3). Cycle threshold (CT), the cycle during which the fluorescence produced exceeds background

levels, was recorded for each target gene. Sample wells recording lower CT values indicate higher expression levels of the targeted gene. Sample wells recording CT values above 35 were discarded due to the manufacturer’s published limit of detection of the

SYBR used.

Figure 2.3 Quantitative reverse transcriptase reaction scheme

Figure 2.3 Thermocycler reaction scheme for quantitative RT-PCR.

Quantitative RT-PCR results were analyzed by first calculating the ΔCT for each sample by normalizing each target gene’s CT value to the calculated geometric mean CT for the two housekeeping genes, histone H4 (hish4) and ornithine decarboxylase (odc) for that sample. The geometric mean is the equal to square root of the first housekeeping gene multiplied by the s housekeeping gene (√∆퐶푇1 ∗ ∆퐶푇2). If three housekeeping genes were used the geometric mean would be 3√∆퐶푇1 ∗ ∆퐶푇2 ∗ ∆퐶푇3). Once the ΔCT had been calculated for both experimental and control samples, the ΔΔCT for each experimental sample was calculated by subtracting the ΔCTcontrol from the ΔCTexperiment.

Fold change between each experiment and the control is equivalent to 2^-ΔΔCT, which accounts for the log base 2 increase between each cycle. Error bars were calculated using the standard error for each condition, and significance is calculated using the unpaired

Student’s T-test, while assuming unequal variance of the two samples.

Table 2.7 Quantitative RT-PCR Primers

Table 2.7 continued

2.8 RNA Synthesis

2.8a Plasmid DNA Linearization

RNA was synthesized from genes of interest for capped RNA or in situ probes.

Genes of interest were cloned into appropriate plasmid backbones and were linearized using a restriction enzyme specific to the backbone. See Table 2.8 for description of plasmids containing genes of interest for RNA synthesis. RNA synthesis reactions used

10-15 µg of linearized plasmid DNA in each reaction. A typical reaction mix is listed below in Table 2.8.

Table 2.8 Plasmid Linearization Mix

Substance Volume (Concentration) Plasmid DNA 10-15 µg 10 X Restriction Enzyme Buffer 10 µl 10 mg/ml Bovine Serum Albumin 1 µl (100 µg/µl) 10 units/µl Restriction Enzyme 5 µl (0.5 units/µl) Nuclease Free Water Remainder Total Volume 100 µl

The reaction mix listed above was then incubated at 37 °C for 4 h to ensure complete linearization of plasmid DNA. Restriction enzymes were purchased from either

Promega or NEB.

Following linearization of plasmid DNA, a 1.2% agarose + ethidium bromide gel was run for one h at 90 vs and visualized using a NucleoTech NucleoVISION UV

Cabinet coupled with ImageJ. The gel was set up as follows: lane 1 – Promega’s 1 kb

DNA Ladder (cat # G5711), lane 2 – restriction enzyme digested plasmid DNA (linear),

lane 3 – undigested stock plasmid DNA (supercoiled). During visualization three characteristics were evaluated: position of the linear plasmid band in relation to the uncut plasmid band (supercoiled uncut band would travel further), the number of bands in Lane

2, and if the linearized plasmid length was the predicted size of the backbone + gene of interest insert. Numerous bands in Lane 2 means that either 1) the full 10-15 µg of plasmid DNA weren’t digested or 2) that the restriction enzyme cut more than one spot on the plasmid DNA. Comparison to the uncut plasmid DNA in lane 3 allows the researcher to determine which situation has occurred. Linearized plasmid DNA band was excised from the gel using a new razor blade and the DNA was extracted from the gel using Zymoclean Gel DNA Recovery kit (cat # D4007, Zymo Research) according to the manufacturer’s specifications. After linear DNA was recovered the sample was nano- dropped for concentration and purity using an IMPLEN NanoPhotometer. Purity of linear

DNA, as measured by 260/230 value, was ideally above 1.8. DNA with a 260/230 less than 1.8 has larger amounts of salt and protein contamination, both of which interfere with efficient RNA synthesis. Linear DNA with low 260/230 values would need to be cleaned with DNA Clean and Concentrator Kit (cat # D4013, Zymo Research) before

RNA synthesis can proceed.

2.8b In situ Probe Synthesis

Linearized and cleaned DNA digested with the restriction enzyme allowing for antisense in vitro transcription from section 2.8a was then used to create an in situ hybridization probe. Hybridization probes incorporate non-radioactive digoxigenin-11-

UTP into the synthesized complementary sequence, allowing for visual detection of the target mRNA. 45

Hybridization probes were synthesized using Invitrogen’s MAXIscript T7/T3

Transciption Kit (cat # AM1326, Thermo Fisher) according to the manufacturer’s specifications. Roche’s DIG RNA labeling mix (cat # 11277073910, Sigma-Aldrich) was substituted for the supplied NTPs. Following completion of the RNA Synthesis reaction, the newly synthesized RNA was nano-dropped for concentration and purity. The RNA probe was diluted in hybridization buffer (50% formamide, 5X SSC, 1mg/ml Torula

RNA, 100 mg/ml heparin, 1X Denhardt’s, 0.1% Tween 20, 0.1% CHAPS, and 10mM

EDTA) to either 10x or 1X and stored at -20 °C.

2.8c Capped RNA Synthesis

Capped RNA was synthesized from sense-linearized and cleaned plasmid DNA described in Section 2.8a. Table 2.8 lists the genes of interest and linearization restriction enzymes used to make capped RNA. Capped RNA was synthesized with Ambion’s mMESSAGE mMACHINE SP6 transcription kit (cat # AM1340, Thermo Fisher) according to the manufacturer’s protocol. Capped RNA was nano-dropped for concentration and purity and 1µl of undiluted capped RNA was run on a 6% TBE-urea gel (cat # EC6865BOX, Thermo Fisher) for one h at 90 v. Properly synthesized capped

RNA presented as one distinct band. Capped RNA was diluted to 2 µg/µl concentration, aliquoted, and stored at -80 °C.

Table 2.9 Plasmids used in RNA Synthesis

2.9 In situ Hybridizations

Samples were collected in glass vials and fixed in MEMFA (0.1M MOPS, pH 7.4,

2 mM EGTA, 1 mM MgSO4 and 4% formaldehyde) for one h at RT, or overnight at 4

°C. Samples collected ranged from albino uninjected whole embryos between stages 12-

28, explants collected at NF stage 28, and correctly injected embryos at stages 14-15.

Following fixation samples were washed three times in 1X diethyl pyrocarbonate phosphate buffered saline (DEPC PBS, 1.85 mM NaH2PO4.H2O, 8.4 mM Na2HPO4.H2O,

175 mM NaCl), dehydrated in two washes of 100% methanol, and stored at -20 °C.

In situ hybridizations were performed according to an adapted Cold Spring

Harbor press protocol (Sive et al., 2010). In situ hybridization took place over a period of five days and is described below. All steps were taken at room temperature with gentle rocking on a nutator unless otherwise stated. For low copy number genes of interest the first day (except for proteinase K permeabilization) took place at 4 °C according to the

Blum lab protocol (Tisler, Thumberger, Schneider, Schweickert, & Blum, 2017).

On the first day samples were removed from storage and allowed to return to room temperature before in situ hybridization commenced. Samples were rehydrated through a four-step serial rehydration from 100% methanol to 1X DEPC PBS+ 0.1%

Tween 20 (1X DEPC PBSw). Each rehydration step lasted 10 min. Samples were washed three times in 1X DEPC PBSw for 5 min each wash. Following rehydration, samples of different biological replicates were pooled by their stage or control/experiment condition and randomly distributed into different vials. Random distribution of samples allowed for multiple genes of interest to be targeted.

Following distribution samples were then permeabilized with 8µg/ml Proteinase

K (cat # P2308, Sigma) in 1X DEPC PBSw for 10-15 min. Samples were then washed three times in triethanolamine (TEA, 0.27 g/15 ml DEPC water, pH 7.5, cat # 90279,

Sigma) for five min each wash. The final 2 mls of TEA was retained and 5 µl of acetic anhydride (cat # 242845, Sigma) was added to the TEA and incubated for 5 min. Acetic anhydride was added once more for a s 5-min incubation. TEA and acetic anhydride reduce non-specific binding through acetylating exposed fatty acids. The samples were then washed three times 1X DEPC PBSw for 5 min each wash, re-fixed in MEMFA for

15 min, and again washed three times in 1X DEPC PBSw for 5 min each wash. Samples were washed in 20% hybridization buffer for 10 min and then pre-hybridized in 100% hybridization buffer for 3-h at 60 °C with gentle horizontal shaking. Upon completion of the 3-h incubation, pre-warmed 1X in situ hybridization probe (enough to cover all the samples, ~300 µls) was added to each vial and the samples were hybridized overnight at

60 °C with gentle horizontal shaking.

On the second day, the in situ hybridization probe was removed and stored at -20

°C for re-use. Embryos were washed once in pre-warmed 100% hybridization buffer at

60 °C for five min. Next, samples were washed three times for 20 min each wash in 2X

SSC + 0.1% Tween (Na3C6H5O7 60 mM, NaCl 0.2 M) pH 7 at 65 °C, followed by two

20-min incubations in 0.2X SSC pH 7. Samples were then moved to RT and washed twice for 5 min each wash in 1X maleic acid buffer (MAB, maleic acid 100 mM (cat #

M0375, Sigma), NaCl 150 mM, pH 7.5). Samples were then blocked in 2% blocking reagent (BMB, cat # 11096176001, Sigma) + 10% normal lamb serum in 1X MAB for one h at RT. Following blocking 1 µl of anti-dioxigenenin antibody (anti-DIG, cat #

11093274910, Sigma) was added to 1 ml of the blocking solution described above.

Samples were incubated in blocking solution + anti-DIG overnight at 4 °C.

Day three consisted of at least 5-h long washes of the samples in MAB. Samples were washed overnight in 1X MAB at 4 °C. Following washing, samples were incubated twice in alkaline phosphatase buffer (AP, 100 mM Tris pH 9.5, 100 mM NaOH, 50 mM

MgCl, and 2 mg/10 ml levamisole (cat # 196142, Sigma)) to remove phosphate groups from exposed protein, preventing non-specific color from developing. Samples then

underwent the color reaction, using one NBT/BCIP tablet (cat # 11697471001, Roche) dissolved in 10 ml of milli Q water. Sample vials were wrapped in aluminum foil to allow for the color reaction to proceed without light. Samples were allowed to react for

30 min to overnight, depending on the copy number of the transcript being probed for.

High copy number transcripts developed quickly, whereas low copy number transcripts, such as those for transcription factors, required longer color reactions. Once the samples had reached the desired level of color, the NBT/BCIP solution was removed and samples were washed three times in 1X MAB for at least five min per wash. They were then fixed in Bouin’s fixative (25% formaldehyde, 5% glacial acetic acid in milli q water) overnight at RT.

The fifth day of whole mount in situ hybridization involved bleaching the native pigment in whole embryos. After fixation in Bouin’s fixative, all samples were washed three times in 1X PBS, five min per wash. Samples were then either imaged directly or underwent bleaching. Bleaching solution (0.5X SSC, 5 % formamide, 0.33% hydrogen peroxide in milli Q water) was added to each sample and samples were placed under an ultraviolet let for several h. After sample pigmentation reached desired levels of transparency, samples were washed three times in 1X PBS and imaged. Once samples had been imaged they were dehydrated in 100% methanol for storage and clearing.

2.10 Sectioning, Imaging, and Image analysis

2.10a Sectioning

Sectioning was performed two different ways; paraffin embedding and sectioning using a Microm HM 340 E (cat # 905190, Thermo Fisher), or cryo-sectioning using a

Leica CM1950. Both types of sectioning used MX35 Premier Plus low-profile blades (cat

# 3052835, Thermo Fisher).

Dehydrated samples (in 100% ethanol) were cleared in Histochoice (cat # H2779-

1L, Sigma) before being embedded in paraffin. Sample clearing consisted of one wash of

50% ethanol and 50% Histochoice for 1-h at RT, followed by 3-h long washes in 100%

Histochoice at RT. Histochoice was disposed following EHS’s guidelines on hazardous chemicals. After clearing, samples were transitioned into paraffin with one wash in 50%

Histochoice and 50% paraffin at 65 °C. Following the transition wash, samples were rotated through three, at least one h long, 100% paraffin washes at 65 °C before positioning and embedding in a paraffin block. Once the paraffin block containing the samples had solidified, the paraffin blocks were sectioned. Section thickness was between 8-15 µm thick depending on the sample. Following sectioning, samples were floated at 35 °C on diH2O to remove wrinkles and allow for sections to better stick to the positively charged slide. Slides were stored at RT until samples were either stained or sectioned.

Fixed samples (using either MEMFA or Dents Fix (20% DMSO and 80% methanol, samples were fixed for less than 24 h to prevent masking or destroying antigens) were (re)hydrated and embedded in optimal cutting temperature (O.C.T., cat #

23-730-571, Thermo Fisher) compound blocks. Samples were flash frozen in liquid nitrogen and stored at -80 °C until used. Samples were cryo-sectioned to thicknesses between 10-30 µms, collected on positively charged slides, and stored at -20 °C until either stained or imaged.

2.10b Imaging

Whole mount samples and explants were imaged using Nikon SMZ800 stereoscope with the NIS-Elements BR version 5.02.00. Images were taken either in filtered 1X PBS in a 10-20% agar dish for hydrated samples, or in 2:1 Benzyl

Benzoate/Benzyl Alcohol (BB/BA) clearing solution in a glass well for dehydrated and cleared samples. Hydrated samples were illuminated from above and to both sides of the sample, with the light source positioning adjusted to minimize the samples shadow and light reflections on the sample. The background was cleared of all debris, and the exposure time and white balance were optimized in the Nikon software. Light sources for dehydrated and cleared samples were positioned laterally to the sample to distinguish sample from background.

Sectioned samples, either from cryo-sectioning or paraffin sectioning, were imaged at the Department of Biology and Biochemistry’s Imaging Core using the Leica

SP8 Upright Confocal DM6000 CFS Microscope combined with LAS AF software.

2.10c Image Analysis

Images were analyzed using FIJI (FIJI is just ImageJ for Windows). Each set of images was analyzed by two independent observers. Embryos stained against sox2 were measured at three sections of the neural plate (anterior open neural plate, posterior open neural plate, and spinal cord) and each measurement was compared to the diameter of the embryo. Ratios of each section/embryo diameter were calculated for all sample embryos and results were averaged for each variant (WT, active, silent.) Standard error for each set of samples was calculated, and significance between the experiments (WT, active) and the control (silent) was calculated using the unpaired Student’s T-Test, assuming unequal variance of each sample set. Results for each set of independent measurements for both target genes were compared and common regions of significance were verified.

2.11 Poly(A) Selected Total RNA Library Preparation

Transcriptome sequencing was performed on two explants, neural plates and noggin animal caps, and at two timepoints, mid-gastrulation (stage 11) and mid- neurulation (stage 18). Mid-gastrulation samples had four biological replicates each for neural plates and noggin animal caps, with approximately 50 pooled explants per biological replicate. Mid-neurulation samples had either five (noggin animal caps) or six

(neural plates) biological replicates. One noggin animal cap biological replicate for mid- neurulation sequencing consisted of 75-100 explants, and one biological replicate for neural plates consisted of 60-75 explants. One biological replicate was collected in 1 ml of Trizol and stored at -80 °C.

RNA used to make cDNA libraries was isolated according to the outlined procedure in Section 2.5. RNA with RIN’s below 7 were excluded for further use. RNA was used at a starting concentration of 1-5 µg. cDNA libraries were created using

NEBNext Poly(A) mRNA Magnetic Isolation Module (NEB # E7490) according to a modified manufacturer’s procedure. Modifications included a fragmentation time of 7 min and double-sided RNA selection. Size selection was performed using AMPure XP beads (cat # A63881, Beckman Coulter) and fragments were selected from lengths between 200-600 nucleotides, with optimal fragment length being 350-450 nucleotides.

Following fragmentation, cDNA libraries were created and barcoded using NEBNext®

Multiplex Oligos for Illumina® (Index Primers Set 1) (cat # E7335S, NEB) for each individual sample and sent to UH Seq-N-Edit for quality analysis on an Agilent

BioAnalyzer. Libraries containing fragments between 200-600 nucleotides in length were each diluted to a final concentration of 4 nM using the Ilumina Pooling calculator

(https://support.illumina.com/help/pooling-calculator/pooling-calculator.html) and pooled into one sample containing up to 12 individually barcoded libraries. Pooled libraries were sequenced at the UH Seq-N-Edit sequencing core using Ilumina’s NextSeq 500 at a sequencing depth of 76 paired ends (PE). A full transcriptome pipeline is outlined in

Figure 2.4.

Figure 2.4 Transcriptome pipeline

Figure 2.4 Bioinformatics pipeline followed for transcriptome analysis.

2.12 Transcriptome Sequencing Analysis

UH Seq-N-Edit shared sequencing results through Ilumina’s basespace platform

(https://basespace.illumina.com). Quality analysis of the sequencing run is provided; with sample metrics such as percent of reads that pass and average quality of the reads among those listed. Table 2.10 is a summary of the samples used for analysis in Chapter 5.

Table 2.10 Summary of Poly(A) Selected Total RNA Sequencing Run Results

Sequences were downloaded onto an external hard drive, then uploaded onto

University of Houston’s Maxwell computing cluster for analysis. Results for each sequenced sample was is received divided into paired fastq files (R1 and R2) from each lane of the four lanes of the sequencing flow cell. Before further analysis, the forward pair files (R1.1, R1.2, R1.3, and R1.4) for each sample were combined sequentially into one file (R1.1→4). This was followed by sequentially combining reverse pair files into one file (R2.1→4). Once files were combined further analysis could proceed. Software used for analysis of sequences is listed below in Table 2.11.

Table 2.11 Software Used for Transcriptome Processing

Paired sequence files were trimmed of remaining adaptor sequences and zero length pairs using the software Cutadapt (Martin, 2011). Cutadapt was run with the following options; -a and -A to filter out the adaptor sequences,

(AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC and

AGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGTAGATCTCGGTGGTCGCCGT

ATCAT) -m to remove sequences lower than the length listed, -o to determine where the output is stored, and -p to indicate that the sequences being input are paired. Once that had been completed, FastQC was used to analyze the sequences and create quality reports

(Andrews, 2010). The FastQC report covers basic statistics (the most relevant are total number of sequences analyzed, and lengths of sequences,) per base sequence quality, per tile sequence quality, per tile sequence scores, per base sequence content, per sequence

GC content, per base N content, sequence length distribution, sequence duplication levels, overrepresented sequences, adapter content, and kmer content. It also rates each of these categories and flags categories that fail FastQC Reports standards. A detailed manual for FastQC can be found at https://dnacore.missouri.edu/PDF/FastQC_Manual.pdf. FastQC reports are also useful for determining if any of the overrepresented sequences were associated with the barcode 57

sequence for that sample. If barcode sequence was detected, a s cutadapt program was run to remove paired sequences containing the barcode.

Once samples had passed the quality report, and been trimmed of unwanted sequences, they were aligned to the Xenopus laevis v9.2 genome assembly and counted.

We accomplished this through utilizing splice aware RSEM version 1.2.25 (See Table

2.11 above), coupled with Bowtie2 2.2.3 (B. Li & Dewey, 2011). We used RSEM with the following options; -p to determine the number of threads used (how quickly the software runs, higher thread count finishes faster, but uses more memory), --paired-end to indicate that the input files were paired sequences, --bowtie2 to tell RSEM which alignment tool to use, --bowtie2-path to indicate to the program where to find the

Bowtie2 software, --estimate-rspd to indicate that the read start position should be estimated, and --output-genome-bam to indicate that the output file should be in .bam format, and will be sorted and indexed by samtools (H. Li et al., 2009). The sample output text file contains the gene numbers, paired RNA sequences that correspond to that gene, the total length of the gene, the effective length of the gene, the expected count, and then two normalized counts, the TPM (transcripts per kilobase million) and FPKM

(fragments per kilobase million). The expected counts were downloaded and compiled into one text file. The gene number in the compiled text file was then opened with excel and translated to its corresponding Xenbase ID (gene name) according to the reference file located at https://sourceforge.net/projects/csbb- repository/files/Laevis_Map.txt/download (Chaturvedi, Chetal, Ponny, & Pabla, 2018).

All data visualization was performed in R Studio version 3.6.0 with the packages listed below. Differential expression was calculated for the genes identified above using

EdgeR (Robinson, McCarthy, & Smyth, 2010). EdgeR was chosen in place of the Tuxedo

Suite as CummeRbund is deprecated (Schurch et al., 2016). Samples were input into R

Studio and genes with low sample counts (less than 2 expected counts) were removed from the sample set (command = rowSums). Samples were then normalized using TMM normalization to account for composition biases between each sample

(calcNormFactors). TMM normalization is appropriate as less than half of the genes are expected to be differentially expressed in the noggin animal caps and neural plates.

Following normalization, we employed a binomial distribution to identify the statistically significant, differentially expressed genes (estimateDisp, glmQLFit, makeContrasts, decideTests) and write the top 500 differentially expressed genes into a new file (Chen,

Mccarthy, Ritchie, Robinson, & Smyth, 2008).

Alternatively, we also identified differentially expressed genes using the R package DESeq2 (Love, Huber, & Anders, 2014). Gene duplicates were removed using the dplyr and the gene table was transformed into a matrix. Each condition was set

(NogAC vs NP) and differential expression was then calculated. The most differentially expressed genes based off their log2fold change and p-adjusted value were sorted and written into a separate table. MA plots, heatmaps, volcano plots, and multidimensional scaling plots were created using the ggplot2 R package and the EnhancedVolcano package (Blighe, 2019; Wickham, 2016).

The expected counts for the top 500 differentially expressed genes were then transferred into a new text file and used for principal component analysis (PCA). PCA allows the user to identify the genes that are responsible for the greatest variation seen identified between variables. It also identifies sample groupings, giving us insight into how similar our sample replicates are (Starmer, 2017). Samples counts were normalized and transformed logarithmically, before the PCA was calculated. To illustrate the variation accounted for and the sample groupings, we used the R package factoextra to calculate scree plots and biplots (Kassambara & Mundt, 2017). PCA also allows us to identify the top genes that are responsible for the variation. We plotted these genes in a heatmap and wrote them into a new text file to use for GO Analysis. Principal component analysis was incorporated into the Sater lab pipeline through the efforts of Seth Uzman.

Gene ontology functional analysis was performed on the sample sets identified above. This allowed the functional pathways associated with the differentially expressed, or the genes responsible for variation to be identified (Huang, Sherman, & Lempicki,

2009c, 2009a). GO DAVID version 6.8 and GO PANTHER were used for our studies as they allowed for Xenbase ID’s (gene names) to be used for functional annotation (Mi,

Muruganujan, Ebert, Huang, & Thomas, 2019). We also adapted our genes into the corresponding Human ENSEMBL and ENTREZ gene ID’s to better functionally annotate our gene set.

2.13 Small Molecule Inhibition of Wnt, FGF, RA, and Notch Signaling in Neural Plate Neural plates (without mesoderm) were isolated according to the procedure outlined in Section 2.3. After isolation of the neural plate from wild-type embryos at

Stage 10.5 neural plates were allowed to partially heal in filtered 1X MMR for 30 min

(until the neural plates had slightly rolled up). Partially healed neural plates were then randomly divided into one of four conditions; filtered 1X MMR + dimethyl sulfoxide

(DMSO, Sigma), filtered 1X MMR + IWR-1 (Wnt signaling inhibition, cat # I0161,

Sigma), filtered 1X MMR + SU5402 (FGF signaling inhibition, cat # C7854-1s,

Cellagen), filtered 1X MMR + RO4929097 (Notch signaling inhibition, cat # C7649-2s,

Cellagen), or filtered 1X MMR + 4-diethylaminobenzaldehyde (DEAB) (Retinoic Acid

Receptor inhibition, cat # D86256, Sigma). Each small molecule was suspended in

DMSO at a stock concentration of 10 mM, aliquoted, and stored at -20 °C to avoid repeated freeze-thaw cycles. During experimentation each small molecule was diluted in filtered 1X MMR to a final working concentration of 100 µM (5 µM for DEAB).

Working concentration was determined from previously published protocols (Fletcher &

Harland, 2008; Myers & Appleby, 2014; Rankin et al., 2018). Neural plates were developed in their culture conditions until after the gliogenic switch had occurred, then collected for their RNA in 1 ml of TRIZOL. Media was changed every 24 h to ensure activity of each small molecule. Neural plates were collected between stages 22-28, and

12-15 neural plates were collected in each sample. Samples were stored at -80 °C.

Chapter III

Initiation of Gliogenesis in

Xenopus laevis

3.1 Overview In neural ectoderm development, the time at which an actively replicating neural progenitor population acquires the ability to differentiate into a glial progenitor population, is referred to as the gliogenic switch (Glasgow et al., 2017; Kang et al.,

2012). Studies in mouse and chick models have shown that the gliogenic switch occurs at

E11.5 in mice, and E5/HH stage 22 in chick (Figure 3.1) (Chaboub et al., 2016). For both model organisms, the gliogenic switch occurs following the closing of the neural tube

(Armit, Richardson, Hill, Yang, & Baldock, 2015; Chaboub & Deneen, 2013; Chaboub et al., 2016; Hamburger & Hamilton, 1992). As gliogenesis and the gliogenic switch have not been investigated in Xenopus laevis, we first had to identify the time at which gliogenesis is initiated.

Xenopus laevis has two phases during which neurogenesis occurs: primary neurogenesis, which occurs between NF stages 14-35, and sary neurogenesis, which begins at NF stage 46 and persists through metamorphosis (Henningfeld, Locker, and

Perron 2007; Wullimann et al. 2005). For the purposes of this chapter, I focus on primary neurogenesis, which encompasses most of neurulation (NF stage 13-21) and the tailbud phase (NF stage 22-44) (Pieter Dirk Nieuwkoop, 1952).

We turned to the mouse and chick models for developmental events and astroglial markers associated with the gliogenic switch. As previously mentioned, neural tube closure (E10.5 for mice and HH St 10 for chick) precedes the gliogenic switch (Figure

3.1) (Armit et al., 2015; Lawson, Anderson, & Schoenwolf, 2001). Neural tube fusion is

first seen during mid-neurulation at NF stage 18 in X. laevis (Davidson & Keller, 1999), giving us an initial stage to investigate.

Previous work in both amniote models also identified a set of transcription factors and other genes that are associated with gliogenesis and glial function. The first is the transcription factor Sox9, which has traditionally been associated with neural crest development (Cheung & Briscoe, 2003). Sox9 functions in mammalian gliogenesis to activate transcription of nuclear factor 1 A (Nfia), though Sox9 is also involved in directing glial differentiation into either astrocytes or oligodendrocytes (Glasgow et al.,

2017; Kang et al., 2012; Stolt et al., 2003). Nfia is the transcription factor described as controlling the onset of gliogenesis (Deneen et al., 2006). It accomplishes this through transcription of astroglial genes such as Glast, the earliest marker of a glial progenitor population (Deneen et al. 2006). Nfia also actively encourages suppression of neurogenesis through its maintenance of activated notch signaling (Kang et al., 2012).

Nfia expression in mice is first identified at E11, in conjunction with glast; Nfia has been shown to be necessary and sufficient to induce glast expression in the ventricular zone

(Deneen et al., 2006; Shibata et al., 1997). Continued expression of Sox9 and Nfia in the glial progenitor population drives its differentiation into astrocytes in mouse and chick, and pulsed expression of Nfia in human neural stem cells resulted in astrocyte differentiation (Glasgow et al., 2014; Kang et al., 2012; Tchieu et al., 2019; Tiwari et al.,

2018). We evaluated the expression patterns of the X. laevis homolog of each transcription factor, sox9 and nfix, to determine whether their expression is consistent with a role in the timing of the gliogenic switch during in primary neurogenesis. The neural progenitor marker, sox2, was included as a control. 64

Genes associated with astroglial cells are divided into those that are initially associated with a glial precursor population, and continue to be expressed in functional astrocytes, and genes that are observed in mature astrocytes and oligodendrocytes. Glial precursor-associated genes in mice are Glast, Fabp7, and Fgfr3; with Glast being the most specific to a glial lineage (Shibata et al. 1997; Owada, Yoshimoto, and Kondo

1996; Molofsky and Deneen 2015). Glast expression has been shown to coincide with the gliogenic switch (Deneen et al., 2006; Molofsky & Deneen, 2015). Glast and Fabp7 have both been shown to be expressed in radial glia, which fulfill the role of nervous system stem cells among many other functions (De Rosa et al., 2012; DeCarolis et al., 2013;

Owada et al., 1996; Shibata et al., 1997). As we were also interested in when specified astrocytes were first observed, we included the radial glia marker nestin as a control. We reasoned that expression of glast and fabp7 in excess of nestin expression, which is observed as early as E11 in mice (Hockfield & McKay, 1985), was likely to be indicative of a functional astrocyte cell population.

Numerous genes associated with mature astrocytes or oligodendrocytes have been identified in mice and chick. For astrocytes, genes such as GFAP, Glt1, S100β, Aldh1l1, and Aquaporin-4 (Aqp4) are used alone or in combination with each other in mouse and chick models (Molofsky et al. 2012; Molofsky and Deneen 2015). For oligodendrocytes, commonly associated genes are Apcdd1 and MBP (Lee et al., 2015; Yoshida, 1997). We chose to observe the expression of the Xenopus homologs for Glt1, Aldh1l1, and Aqp4 for mature astrocytes as there are no identified homologs for GFAP or S100β in Xenopus species. We also chose to observe the expression of Apcdd1 for oligodendrocytes, as

MBP has been shown to be heavily expressed prior to and during gastrulation in X. laevis 65

embryos, and this early expression would obscure the emergence of expression in oligodendrocytes (Session et al., 2016).

Figure 3.1 Timing of the gliogenic switch

Figure 3.1 Timing of initiation of gliogenesis (gliogenic switch) in amniotes (A,B) and Xenopus (C). A, B. Glial progenitor cells are first observed at E12.5 in mice and E6 in chick. The gliogenic switch occurs following the complete closure of the neural tube(~E10.5) for mice, and twelve to fifteen stages after the initiation of neural tube closure in chicks (Armit et al., 2015; Lawson et al., 2001). Glial specification refers to the initial timepoint where an astrocyte or oligodendrocyte progenitor population is observed. C. The gliogenic switch occurs as early as NF stage 16 in the developing X. laevis presumptive hindbrain/anterior spinal cord and appears to follow neural tube closure from this initiation point. Information for A and B was adapted from previously published research by Chaboub et al (Chaboub & Deneen, 2013).

3.2 Temporal Expression of Glial Associated Genes in X. laevis As we had postulated the correlation between closure of the neural tube and initiation of gliogenesis following a period of neurogenesis, we initially investigated the transcript and protein levels of neuronal (n-cadherin) and glial associated genes (glast, fabp7, and glt-1) from previously published transcriptome and proteome datasets (Figure

3.2) (Peshkin et al., 2019; Session et al., 2016). N-cadherin transcript and protein levels increased in expression immediately following initiation of neurulation (NF Stage 13).

Transcript levels decreased throughout the remainder of tailbud phase; however, protein levels didn’t experience the same decrease in expression. Glast transcript levels started to increase between NF stages 15-18, with fabp7 transcript levels most drastically increasing following NF stage 25. Protein levels for glast started accumulating between

NF stages 20 and 22, near the end of neurulation, whereas fabp7 levels saw their greatest increase past mid tailbud phase. Glast transcription is initiated slightly earlier than the known initiation of neural tube closure (NF stage 18) however, as the protein doesn’t accumulate until between NF stages 20 and 22, it could still be consistent. Both protein and mRNA transcripts were observed in mice at E11 (Shibata et al., 1997). The presence of glt-1 protein indicates that specified astrocytes are observed as early as the beginning of tailbud phase (NF stage 22-24).

Figure 3.2 Graph of onset of glial-specific gene expression during Xenopus development A. 400 n-cadherin glast fabp7 350 300 250 200 150

Expression Expression (TPM) 100 50 0 9 10 12 15 20 25 29-30 35-36 40 Nieuwkoop and Faber Stage

B. 0.9 n-cadherin glast fabp7 glt-1 0.8 0.7 0.6 0.5 0.4 0.3 0.2

Relative Protein Expression Protein Relative 0.1 0 9 12 17 20 22 24 29-30 42

Nieuwkoop and Faber Stage Figure 3.2 Expression of neuronal and glial associated genes during primary neurogenesis, based on published datasets. Neuronal associated (n-cadherin) transcript (A) and protein (B) levels increase in expression immediately following initiation of neurulation (NF stage 13). A. Expression of astroglial associated (glast, fabp7) transcript levels increases exponentially during mid-late neurulation. (NF stage 18-20) B. Protein accumulation of glial associated genes (glast, glt-1) begins to rise at late neurulation (NF stage 20-21), with levels of fabp7 drastically increasing at mid tailbud phase. (NF stage 29) Transcript (Session et al., 2016) and protein (Peshkin et al., 2019) data identified from previously published datasets available on xenbase.org.

After observing an increase in transcript and protein levels of the glial progenitor markers glast and fabp7 as early as mid-neurulation (potentially the onset of the gliogenic switch), we confirmed these results using quantitative RT-PCR in whole embryos to detect glial-associated genes described at the end of section 3.1 throughout neurulation and tailbud phase (Figure 3.3). We accomplished this by collecting embryos every two stages from NF stage 12 (end of gastrulation) until NF stage 28 (late tailbud phase.)

Samples from each stage were pooled (five whole embryos/sample) and were considered one biological replicate. At least three biological replicates were utilized for each timepoint and target gene. We compared transcript levels of the identified target genes to the levels observed during gastrulation (NF stage 12) under the assumption that expression levels for genes associated with glial specification would be low during gastrulation.

Initially, we measured the expression of putative astroglial transcription factors sox9 and nfix (Figure 3.3 A). Expression levels of the neural progenitor transcription factor sox2 were also measured as a control. Sox2 levels remained constant throughout neurulation and tailbud phase as compared to those observed in gastrulation, which is consistent with published levels (Session et al., 2016). Increased levels of sox9 were first observed at NF stage 14 and remained consistently upregulated throughout neurulation and late tailbud stages. Interestingly, nfix transcripts showed increased levels of expression only by NF stage 20, and this increased level of expression only persisted through NF stage 24, returning to gastrulation levels at NF stage 28. This is inconsistent with the potential gliogenic switch timepoint suggested by glast expression levels from

Sessions et al (Figure 3.2). It is also inconsistent with the observation in mice that Nfia, 69

the homologue of nfix, initiates transcription, then continues to be transcribed in the nervous system throughout the remainder of the organism’s life (Chaboub et al., 2016).

Results on expression of all nuclear factor one transcription factors is discussed in

Chapter 4.

We then observed the expression of glial progenitor-associated genes (Figure 3.3

B). Nestin, the intermediate filament protein specific for neural progenitor cells and radial glia (Dahlstrand, Zimmerman, McKay, & Lendahl, 1992; Hockfield & McKay, 1985), was found to increase in expression at NF stage 16, with its expression levels remaining relatively constant until late tailbud phase. Glast, the earliest marker of a glial progenitor cell population (Deneen et al., 2006; Shibata et al., 1997), showed an initial increase in expression at NF stage 16; expression then increased exponentially throughout the remainder of neurulation and tailbud phase. This expression pattern was echoed by fabp7, an astroglial progenitor marker (Owada et al., 1996) starting at NF stage 20 and exponentially increasing throughout the remainder of tailbud phase. Transcript expression of fabp7 and glast confirm the previously published data from Sessions et al., and protein accumulation closely follows that of transcript expression for both genes (Figure 3.2 A,

B). Finally, we observed the expression of mature astrocyte-associated genes (Figure 3.3

C). Glt-1, the glutamate transporter commonly expressed in mature astrocytes, increases in expression at NF stage 24 and 28, in agreement with protein data published by Peshkin et al (Figure 3.2 B). Aldh1l1 and apcdd1, mature astrocyte and oligodendrocyte markers, respectively, showed no significant changes in expression throughout neurulation and tailbud phase, as compared to levels seen during gastrulation. Transcript levels of aqp4 transiently increased at NF stage 20, but quickly returned to baseline expression. 70

Figure 3.3 Temporal expression of glial associated genes A. Transcription factors 25 sox2 sox9 nfix 20

10 Fold Change Fold 5

0 14 16 18 20 24 28 Nieuwkoop and Faber Stages B. Radial glial and early glial progenitor associated genes 10000 nestin glast fabp7 1000

100 Fold Change Fold 10

1 14 16 18 20 24 28 Niewkoop and Faber Stages C. Differentiated glial associated genes 60 aldh1l1 glt-1 aqp4 apcdd1 50

20 Fold Change Fold 10

0 14 16 18 20 24 28 Niewkoop and Faber Stages

Figure 3.3 A temporal profile of expression of genes associated with glial development and differentiation in whole embryos during primary neurogenesis. A. Quantitative RT- PCR analysis of expression levels of transcription factors associated with neural progenitors (sox2) and glial specification and initiation of differentiation (sox9 and nfix) as compared to expression levels during gastrulation. Levels of sox2 remain relatively constant throughout primary neurogenesis, whereas sox9 experiences gradual increases in expression levels. Expression levels of nfix experience increases starting at NF stage 20 and 24 before dropping at NF stage 28. B. Expression levels of early glial progenitor genes. Levels of nestin, a radial glial marker, increases by NF stage 16 as compared to gastrula embryos. Expression levels of glast are consistent with those observed for nestin until NF stage 18, at which point they start to increase exponentially. Fabp7 lags behind glast, increasing dramatically after NF stage 24. C. Expression levels of mature astrocyte and oligodendrocyte markers. Aldh1l1 and apcdd1 expression levels remain constant through all stages of primary neurogenesis. Levels of glt-1 increased by NF stage 24, and levels of aqp4 increased at NF stage 20, but immediately returned to baseline levels at NF stage 24. Fold change was calculated by comparing expression levels to gastrulation stage levels. Error bars represent S.E.M. N=3 for sox9, nestin, and fabp7. N=7 for sox2, nfix, glast, glt-1, and apcdd1.

3.3 Spatial Expression of Glial Associated Genes in X. laevis Quantitative RT-PCR of astroglial transcription factors, early glial progenitors, and differentiated glial cell markers demonstrate that the gliogenic switch occurs as early as NF stages 16-18. They also cast doubt on a conserved role for nfix, the homolog of

Nfia in Xenopus, in being necessary and sufficient for expression of glast (See previous section). To further investigate the roles for previously identified genes in initiating gliogenesis, we evaluated their spatial patterns of expression through in situ hybridizations on albino embryos during gastrulation, neurulation, and the tailbud phase

(Figure 3.4 and Figure 3.5).

Expression of sox9 is absent during gastrulation (Figure 3.4 A) and is first observed during neurulation (Figure 3.4 B-D). However, this expression appears to be limited to regions associated with early neural crest (Pieper, Eagleson, Wosniok, &

Schlosser, 2011). It isn’t until the tailbud phase (Figure 3.5 A,B) that expression of sox9 appears in the developing neural tube, though expression of sox9 might be present internally at the very end of neurulation (Figure 3.4 D). Expression of nfix appears to be either absent or expressed at extremely low levels during gastrulation and neurulation,

(Figure 3.4 E-H) with expression rising to observable levels in the tailbud phase (Figure

3.5 C,D). Since neither sox9 nor nfix are expressed at detectable levels during mid-late neurula, as glial-specific gene expression is initiated, these observations contradict previously established roles for sox9 and nfix in mouse and chick models.

Nestin is lightly expressed as early as NF stage 17 (Figure 3.4 K) and remains present through the remainder of the stages pictured (Figure 3.4 L, Figure 3.5 E, F).

Initially, punctate expression of glast is observed in mid to late neurulation (Figure 3.4 O,

P) before expression expands broadly throughout the neural tube (Figure 3.5 G, H).

Interestingly, aldh1l1 is expressed during gastrulation (Figure 3.4 Q), and expressed in the epidermis during neurulation (Figure 3.4 R, S, T). Expression of aldh1l1 is difficult to discern in the neural plate and tube during this time and expression is drastically reduced during tailbud phase (Figure 3.5 I, J). When these embryos were cleared in BB/BA solution, we found that aldh1l1 was observed in the intercalated disk pattern characteristic of expression in the notochord (Figure 3.6 B, C). When we observed NF stage 14/15 embryos laterally, we saw strong expression of aldh1l1 in the notochord and scattered expression in the epidermal ectoderm, but little to no expression in the neural plate (Figure 3.6 A). As expression of aldh1l1 appeared to be decreasing as neurulation progresses (Figure 3.6 C), we observed the previously published dataset from

Sessions et al, which showed that by late tailbud phase, transcript levels were had decreased close to baseline levels for aldh1l1 (Figure 3.6 D). This indicates that while aldh1l1 is a marker for a specified astrocyte population in mice (Molofsky & Deneen,

2015), the same does not hold true for Xenopus laevis during primary neurogenesis.

Expression of glast appeared only after neural tube closure (Figure 3.4 O, P), which is consistent with findings published in mouse and chick (Molofsky & Deneen,

2015; Shibata et al., 1997). Glast expression consistently appeared in the region corresponding to the presumptive hindbrain, and anterior notochord (Figure 3.4 O). To address if there was an anterior – posterior identity for the gliogenic switch, we performed an in situ on NF stage 20 embryos to detect expression of engrailed2 alone

(Figure 3.7 A) and in combination with glast (Figure 3.7 B). Engrailed2 is uniquely expressed anterior to the midbrain-hindbrain boundary (MHB) (Brivanlou & Harland,

1989). We observed that glast expression was observed posterior to engrailed2 (Figure

3.7 B), indicating that the gliogenic switch occurs first in the hindbrain and anterior spinal cord. The gliogenic switch then moves posteriorly down the neural tube (Figure

3.4 P) before moving anteriorly into the midbrain and forebrain portions of the developing neural tube (Figure 3.5 G, H).

Figure 3.4 Spatial expression of glial associated genes during gastrulation and neurulation

Figure 3.4 In situ hybridizations of glial associated genes during gastrulation and neurulation. A-D. Sox9, the transcription factor which turns on Nfia in mice, (Kang et al., 2012) is expressed in regions associated with neural crest cell populations during neurulation. (B-D) Sox9 is absent during gastrulation, and isn’t observed in the neural plate (B,C) or neural tube.(D) E-H. Expression of nfix, the transcription factor considered necessary and sufficient for astroglial development, has negligible expression during neurulation, though may be lightly expressed during gastrulation.(A) I-P. Expression of nestin and glast are observed in the neural plate midway (NF Stage 16-18) through neurulation. Q-T. Aldh1l1 is observed in the presumptive neural plate during gastrulation. (Q) However, expression of adh1l1 is observed in the epidermis, along with the midline of the neurulation embryo. (R-T) Albino embryos were used to observe endogenous expression, and images chosen are representative of each sample. Embryos pictured in O and P were cleared in BB/BA clearing solution. Scale bar is 500 µm. Albino, but otherwise wildtype, embryos were collected from four separate fertilizations and mixed. Each gene of interest had between 7-10 embryos from the mixed origin stained for each stage shown. Images chosen were representative for staining patterns observed.

Figure 3.5 Spatial expression of glial associated genes during tailbud stages

Figure 3.5 In situ hybridizations of glial associated genes during tailbud stages. A, B. Expression of sox9 is observed along the neural tube and in the neural crestage C, D. Nfix is initially observed in the anterior neural tube (C) before expanding throughout the neural tube in late tailbud phase. (D) E, F. Low levels of nestin are observed in the throughout the neural tube. G, H. Glast is heavily expressed throughout the neural tube during tailbud phase. I, H. Aldh1l1 is absent during tailbud phase. Embryos pictured in A, C, and D were cleared in BB/BA solution. Scale bar is 500 µm. Number of albino embryos stained is described in Figure 3.4 legend.

Figure 3.6 Aldh1l expression in the notochord

Figure 3.6 In situ hybridization of aldh1l in cleared whole embryos at NF stage 14/15 (A), NF stage 20 (B), and NF stage 28 (C). A. Expression of aldh1l1 is observed in the notochord, located in between the dorsal mesoderm, and underlying the neural plate. No expression of aldh1l1 is observed in the neural plate. B,C. Aldh1l1 also exhibits the typical intercalated disks observed in notochord. By NF stage 28 (C) Expression of aldh1l1 is present but reduced throughout the notochord. Black arrow indicates expression of aldh1l1, red arrow indicates bubble caught in embryo. Black dots are debris on embryo. A = anterior, P = posterior. This is echoed by the observed expression levels in whole embryos (D). D. Expression of aldh1l1 transcript is highest during gastrulation and neurulation, with expression levels low after the tailbud phase. Data identified from Sessions et al. previously published dataset (Session et al., 2016). Scale bar is 500 µm. 7- 10 albino embryos were used at each stage from n=4 independent fertilizations.

Figure 3.7 Astroglial development along the anterior-posterior axis

A. B.

Posterior

Right Left

engrailed2 engrailed2 + glast

Figure 3.7 In situ hybridizations against A) engrailed2 or B) engrailed2 + glastage At NF stage 20, expression of glast is observed posterior to expression of engrailed, MHB gene. This indicates astroglial cells are observed first in the presumptive hindbrain and spinal cord. Scale bar is 500 µm. Wildtype albino embryos from two independent fertilizations were mixed and 5 embryos were stained for engrailed2, and 7 embryos were stained for engrailed2 + glast. Images chosen were representative of observed phenotype.

3.4 Summary The temporal and spatial expression patterns cast doubt on sox9 and nfix directing gliogenesis in X. laevis, indicating a divergence from their identified roles in mouse and chick. The initial expression of glast, correlated to the induction of the gliogenic switch, is seen between NF stages 16-18 for mRNA transcript expression, with protein accumulation by the end of neurulation. This potentially introduces doubt about whether neural tube closure occurs before the gliogenic switch; however, in situ hybridization showed punctate expression of glast in portions of the neural tube that were on the verge of fusing. With X. laevis’ rapid rate of development as compared to mouse embryos, functional glial cells could be needed earlier in development, keeping pace with the closure of the neural tube. Comparisons of expression patterns for glast and the MHB gene engrailed2, indicate that the gliogenic switch is initiated at different timepoints depending upon the position along the anteroposterior axis. The gliogenic switch is initiated in the anterior spinal cord by NF stage 16, and then moves posteriorly along the spinal cord; initiation of gliogenesis is not detected in anterior neural tissue (midbrain and forebrain) until NF stage 24. This is the first clear demonstration of anterior – posterior regionalization of the gliogenic switch in a model organism.

Chapter IV

Regulation of the Initiation of Gliogenesis by Extrinsic Signals

4.1 Overview Contributions to gliogenesis by various signaling pathways have been studied since the early 1990’s. The most extensively studied pathway in relation to gliogenesis is the Notch signaling pathway. Initial interest stemmed from the dual roles of Notch in maintaining a neural progenitor population and inhibiting neuronal differentiation; however it was found that Notch signaling actively promotes gliogenesis beyond its role in maintaining the population of proliferating neural progenitors (Grandbarbe et al.,

2003; Lowell, 2000; Rogers, Moody, & Casey, 2009; Taylor et al., 2007). Notch signaling encourages glial differentiation through several avenues, the first of which is the activation of delta, which then commits neural progenitor cells to a glial fate (Lowell,

2000). Sly, notch intracellular domain (NICD) activates transcription of hes1 and hes5, which each then drive glial differentiation in neural progenitors (Hojo et al., 2000).

Finally, Notch signaling also encourages the formation of a radial glial population by directly activating transcription of the astroglial marker fabp7 (T. E. Anthony et al.,

2005; B. A. Patten et al., 2006). Wnt and Bmp signaling also play significant roles in aspects of oligodendrocyte and astrocyte differentiation, though specific roles in the initiation of gliogenesis have not been clearly established for either pathway. Wnt signals are integral to orchestrating the timing of oligodendrocyte differentiation (Dai et al.,

2014; H. K. Lee et al., 2015; Shimizu et al., 2005) and Bmps have been shown to induce astrocyte differentiation from a mammalian glial precursor population (Gross et al., 1996;

Mabie et al., 1997).

While each of these pathways is known to contribute to gliogenesis, no one signaling pathway appears to be entirely responsible for initiating the gliogenic switch.

Notch signaling is necessary for the gliogenic switch to occur, however it is not sufficient for the activation of gliogenesis, as notch signaling is active both before and after the gliogenic switch (Hojo et al., 2000). The currently accepted model for activation of the gliogenic switch is that Notch signaling contributes to the timing of the gliogenic switch through Hes5, which contributes to Nfia transcription, and Nfia then initiates transcription of glial associated genes (Deneen et al., 2006). However, our findings described in

Chapter 3 are inconsistent with this model and suggest that mechanisms underlying the initiation of gliogenesis are not strongly conserved between the proposed mouse model and Xenopus glial development. To address this question, we utilized one of the main strengths of the Xenopus laevis system for developmental studies, our ability to isolate robust explants. These explants are specific to various tissue types and allow the researcher to exhibit very fine control over the microenvironmental conditions for the explant. This allows us to address questions that would be difficult to investigate in other model organisms.

For the experiments discussed in this chapter, we used two distinct explants, noggin injected animal caps, and isolated neural plates (Figure 4.1). Animal caps are a region of pluripotent cells isolated from the animal region of mid-blastula stage embryos, which will form the most anterior neural tissue in response to the inhibition of BMP

(Lamb et al., 1993). This is accomplished through overexpression of noggin, a factor secreted from the dorsal lip (Spemann organizer) during gastrulation. Our s explant is called a neural plate, a group of cells isolated from mid-gastrula embryos that corresponds to the early neural ectoderm. This explant offers a more balanced representation of the entire central nervous system reflecting the entire anteroposterior 84

pattern across the explant. The patterning of the neural plate is due to the combination of noggin and fgf2 signaling from the organizer, which synergistically act to establish A-P identity in the explant (Lamb & Harland, 1995). Neural plates can be isolated with or without their underlying mesoderm. Neural plates isolated away from mesoderm receive an initial burst of organizer signals, but then grow without further external signals, whereas neural plates isolated with their underlying mesoderm continue to receive signals that a neural tube will be exposed to in a whole embryo. The mesoderm isolated with the neural plate will differentiate into the notochord, the dorsal most mesoderm in a developing embryo, and an important signaling center. Mesodermal signaling to the overlying neural ectoderm is important for proper morphogenesis and patterning of the neural plate (Araya et al., 2014; Patten & Placzek, 2000).

Each explant will allow us to answer a specific question. Noggin injected animal caps (NogACs) will test whether inhibition of bmp4, which is considered sufficient for establishment of neural ectoderm, is sufficient to initiate glial differentiation. Neural plates without mesoderm will allow us to investigate the role of initial signals from the organizer in gliogenesis, and neural plate with mesoderm will let us examine how continued signaling from the notochord affects gliogenesis.

Figure 4.1 Experimental summary

Figure 4.1 A. Experimental timeline. B. White dashed line indicates the portion of NF stage 8 embryo isolated that is referred to as an animal cap. Left alone, an animal cap has high levels of bmp4 and is specified to form epidermal ectoderm. Animal caps isolated from embryos injected at the 2-cell stage with noggin, a bmp signaling antagonist, are specified to form neural ectoderm. C. White dashed lines indicate where excisions were made to isolate the presumptive neural plate at NF stage 10.5 and NF stage 11, respectively. Neural plates cut at NF stage 10.5 were isolated independently of the underlying mesoderm, whereas neural plates at NF stage 11 were excised in conjunction with their tightly bound underlying mesoderm. The underlying mesoderm will differentiate into the notochord and somites. All explants were cultured in 1X MMR until they reached the desired stage.

Table 4.1 Genes Involved in Neural Specification and Differentiation

Gene of Function in Neural Ectoderm Specification and Differentiation Interest

noggin BMP4 inhibitor, promotes neural development in all vertebrates.

sox2 Expressed in proliferating neural progenitors

neurod1 Transcription Factor in neuronal differentiation pathway

ngn1 Transcription Factor in neuronal differentiation pathway, inhibits astrocyte differentiation in mouse

nestin Intermediate filament protein expressed in radial glia

sox9 Transcription Factor with multiple roles in neural development; initiates mammalian glial specification.

nfix Xenopus orthologue of mammalian Nfia. Nfia is the transcription factor necessary and sufficient for astrocyte differentiation in mice.

olig3 Transcription Factor expressed in glial specification.

sox10 Transcription Factor required for oligodendrocyte differentiation pathway.

fabp7 Fatty acid binding protein, also known as blbp. Expressed in both astrocytes and radial glia.

apcdd1 Expressed in mammalian astrocyte progenitors.

glast Glutamate transporter. Expressed in early forming astrocytes and radial glia. Initial expression congruent with the gliogenic switch

glt-1 Glutamate transporter. Expressed in mature astrocytes.

mbp Myelin basic protein. Expressed in mature oligodendrocytes

4.2 Competency of Xenopus Explants in Initiating Gliogenesis We discovered that the gliogenic switch occurs between NF stages 16-24 depending on the anterior to posterior location in the neural tube. Gliogenesis is initiated in the anterior spinal cord, first extending posteriorly down the spinal cord before it is initiated in the midbrain and forebrain (Chapter 3). During gastrulation, cells forming the presumptive forebrain are quickly pushed away from Spemann’s organizer by the convergence and extension of the presumptive hindbrain and spinal cord (Keller, Shih,

Sater, & Moreno, 1992) limiting the exposure of highly concentrated signals coming from the organizer. We utilized isolated neural plate without the underlying mesoderm, neural plate with the underlying mesoderm, and animal caps overexpressing noggin to identify the roles signaling plays in gliogenesis (Figure 4.1) These explants were isolated at NF stage 10.5, NF stage 11, and NF stage 8 respectively, and allowed to grow until mid-tailbud stage (NF Stage 28) to allow for the complete activation of the gliogenic switch along the anteroposterior axis of intact sibling embryos.

Initially we isolated neural plates ± mesoderm and observed them throughout their development until NF stage 22. The explants of both types exhibit outwards signs of development akin to the neural plate in a whole embryo, with a fold appearing in the explant similar to the median hinge point in a whole embryo (Figure 4.2 A, Figure 4.3 A).

There were also layers within the developing neural plates, with light cream-colored early neural tissue covered by a layer of epidermal ectoderm (Figure 4.2A). Neural plates retain both the organization and the epidermal covering until they are collected at NF stage 28 (Data not shown).

Following this, we collected the neural plate ± mesoderm samples at mid-tailbud stage, isolated their RNA, and compared the expression of genes associated with neural specification and differentiation with that of their whole embryo siblings (See Table 4.1 for genes of interest and their roles in neural development). We found that both neural plates with or without mesoderm are expressing neural progenitor (sox2), neuronal

(neurod1, ngn1), and glial-associated genes (nestin, sox9, nfix, sox10, olig3, apcdd1, glast, glt-1, and fabp7) at levels that are at least consistent with those seen in their whole embryo siblings (Figure 4.2 B). Interestingly, while expression levels in neural plates with mesoderm never significantly changed from those observed in whole embryos, neural plates without mesoderm appeared to be weighted toward initiating oligodendrocyte differentiation. This was indicated by the increased levels of expression of oligodendrocyte transcription factors sox10, olig3, and apcdd1, a negative regulator of canonical Wnt signaling with previously established roles in stimulating oligodendrocyte differentiation (H. K. Lee et al., 2015). This was coupled with significantly decreased expression of nfix, the X. laevis orthologue of Nfia, which fits a model of a mutually antagonistic system balancing astrocyte and oligodendrocyte differentiation described in mice and chick (Glasgow et al., 2014). Overall, it appears the continued signaling from the Spemann organizer and then the underlying mesoderm potentially plays a role in balancing the levels of differentiation between astrocytes and oligodendrocytes.

Figure 4.2 Astroglial specification in neural plates with or without mesoderm

Figure 4.2 Astroglial specification in neural plates (± mesoderm) as compared to whole embryos. A. NF stage 22 neural plates exhibit a crease that resembles the neural fold in a whole embryo. The overlying epidermis was partially removed to observe the neural plate layers. B. As compared to their whole embryo siblings, neural plates in either the presence or absence of mesoderm are equally competent at undergoing neuronal and astroglial specification. Neural plates without mesoderm exhibited significantly increased levels of oligodendrocyte associated genes (sox10, olig3, apcdd1) and significantly decreased nfix. Samples were normalized to the geometric mean. Neural plates were collected at NF stage 28. Standard error was calculated for error bars. Unpaired student’s TTEST was performed to calculate significance. N=5, p ≤ 0.05. Scale bar is 250 µm long.

We then addressed whether the initiation of bmp4 signaling was sufficient to initiate gliogenesis. Inhibition of bmp4 signaling is sufficient to initiate neuronal differentiation (Mehler et al., 1997; Sater, Steinhardt, & Keller, 1993) and overexpression of noggin, the bmp antagonist secreted from the Spemann organizer (Smith & Harland,

1992), in isolated animal caps is capable of directing the animal caps into neural differentiation (Rogers, Moody, & Casey, 2009). By isolating noggin animal caps at the blastula, rather than the gastrula, stage we prevented the animal caps from receiving any endogenous mix of signals from the organizer, allowing us to examine the effect of bmp4 inhibition in isolation. We also isolated uninjected animal caps to confirm that high levels of endogenous bmp4 prevent neuronal differentiation.

Isolated animal caps were observed at NF stage 20 (Figure 4.3 B) and at the time of collection, NF stage 28 (Figure 4.4 A). Animal caps overexpressing noggin at NF stage 20 are covered by an epidermal layer but exhibit no signs of the physical cell rearrangements observed in neural plates at the corresponding timepoint (Figure 4.3 A, B).

Following further development, noggin animal caps appear to consist of only light cream- colored neural tissue and an early cement gland (brown spot) (Figure 4.4 A). This contrasts with the appearance of uninjected animal caps at NF stage 28 which mostly consisted of epidermal ectoderm.

Once uninjected and noggin animal caps were collected, we compared their levels of neuronal and glial associated gene expression against those observed in their whole embryo siblings (Figure 4.4 B). Uninjected animal caps showed significantly reduced levels of expression for most early neural, neuronal, and glial genes compared to whole

embryo siblings. These results are consistent with expectations, as uninjected animal caps with high levels of bmp4 should undergo epidermal differentiation only. Levels of the transcription factors sox9 and nfix remained consistent with expression levels observed in whole embryos, and expression of apcdd1 increased. Unchanged levels of sox9 and nfix as compared to whole embryos are explained by their known expression in epidermal ectoderm (Session et al., 2016) and could be fulfilling other roles in these tissues. As

Wnt signaling plays significant roles in epidermal ectoderm development (Patthey &

Gunhaga, 2014), it is possible that the increase in apcdd1 is associated with epidermal development.

When we observed the results for noggin animal caps, we were shocked to find that while noggin animal caps expressed genes characteristic of neural progenitors and initiation of neuronal differentiation, they had very low levels of expression of astroglial transcription factors and associated genes (sox10, olig3, glast, glt-1). This appeared to affect both astrocytes (glast, glt-1) and oligodendrocytes (sox10, olig3) although radial glial markers (nestin, fabp7) were unaffected (Figure 4.4 B). Though fabp7 is both a radial glial and astrocyte marker, at this timepoint it is predominantly expressed in radial glia. Expression of fabp7 transcript and protein did not increase exponentially until after

NF stage 28, thus we reasoned that this increase in expression is associated with astrocyte expression (Chapter 3, Figure 3.2, and 3.3 B). This appears to suggest that noggin animal caps may not be capable of, or have a greatly reduced ability, to initiate gliogenesis despite the elevated expression of sox9 and nfix.

To evaluate spatial organization of gliogenesis in these explants, we isolated both neural plates (glial specification-competent) and noggin animal caps, collected both explants at NF Stage 20, and performed an in situ hybridization against glast, the astroglial gene we have used to evaluate the gliogenic switch (Figure 4.3 C, D). We observed strong expression of glast in portions of the neural plate; however, there was little to no expression of glast in noggin animal caps. This confirms what we observed in the quantitative RT-PCRs from Figures 4.2 and 4.4.

Figure 4.3 Expression of glast in neural plates and noggin animal caps

Figure 4.3 NF stage 20 NP (A, C) and NogAC (B,D) explants. NP explants exhibit morphological patterning (A) whereas noggin animal caps do not (B). In situ hybridization against glast in NPs (C) vs NogACs (D). NPs have distinct expression of staining whereas noggin animal caps exhibit little to no expression of glastage Neural plates and NogACs were collected from 2 independent experiments, with an n=7 for NPs and n=13 for NogACs. 5/7 NPs showed high expression of glast, whereas only no NogACs had significant levels of glast. Scale bar is 250 µm long.

Figure 4.4 Effects of bmp4 inhibition on astroglial specification

Figure 4.4 Astroglial specification in noggin animal caps as compared to whole embryos. A. Morphological differences between NF stage 28 animal caps. Uninjected animal caps consist almost entirely of epidermal ectoderm. Noggin animal caps retain epidermis at NF stage 20 (Figure 4.2 B), however they have little to no epidermal ectoderm by NF stage 28. Noggin animal caps at this late stage exhibit mostly white presumptive neural tissue. B. Animal caps isolated from NF stage 8 uninjected embryos show significantly decreased levels of neural progenitor, (sox2) neuronal, (neurod1, ngn1) and astroglial (nestin, sox10, olig3, glast, glt-1, fabp7) associated genes as compared to their whole embryo siblings. There is no change in expression for sox9 and nfix; acpdd1 expression is significantly increased in uninjected animal caps. Expression levels of sox2, neurod1, and ngn1 are significantly increased in noggin animal caps. In contrast, astroglial genes associated with potential differentiation (sox10, olig3, glast, glt-1) were very significantly decreased in expression. Expression levels of early astroglial transcription factors sox9 and nfix, and radial glial associated genes (nestin, fabp7) were not unchanged in expression as compared to whole embryos. Samples were normalized to the geometric mean. Animal caps were collected at NF stage 28. Error bars represent S. E. M. Unpaired student’s TTEST was performed to calculate significance. Uninjected animal cap n= 4, noggin animal caps n=6. p ≤ 0.05. Scale bar is 250 µm long.

Bmp signaling is important for astrocyte differentiation, and in the adult subventricular zone, inhibition of Bmp signaling by noggin is considered inhibitory to glial differentiation (Bonaguidi et al., 2005; Hu et al., 2010; Lim et al., 2000; Nakashima,

Yanagisawa, Arakawa, & Taga, 1999; Niu et al., 2013; Srikanth et al., 2014). We addressed the possibility that continued inhibition of Bmp signaling in noggin animal caps was preventing gliogenesis. To do this, we injected capped noggin mRNA into whole embryos, then isolated their neural plates during gastrulation before allowing them to develop until tailbud stage. Following collection, we measured the expression of the neuronal and astroglial genes in the noggin neural plates as compared to whole embryos

(Figure 4.5). We found that expression of astroglial genes such as sox10, olig3, glast, glt-

1, fabp7, or mbp in noggin neural plates were slightly increased as compared to their whole embryo siblings, though not significantly. This indicates that noggin-expressing neural plates were equally competent as their whole embryo siblings to initiate gliogenesis.

Figure 4.5 Effects of persistent BMP4 inhibition on astroglial specification

Figure 4.5 Astroglial specification in noggin (bmp4 inhibited) neural plates as compared to whole embryos. Neural plates isolated from embryos injected with noggin at 2-cell stage did not exhibit the decrease in astroglial genes such as sox10, olig3, glast, and glt-1 indicative of bmp being required for their expression. Instead, noggin neural plates had slightly higher expression levels of these genes, consistent with observations from uninjected neural plates. Samples were normalized to the geometric mean. Neural plates were collected at NF stage 28. Error bars represent S.E.M. An unpaired student’s TTEST was performed to calculate significance. N=3. p ≤ 0.05.

We determined that neural plates both with and without mesoderm are equally capable of initiating gliogenesis. This contrasts with noggin animal caps where bmp4 is inhibited, and neuronal differentiation is initiated, but glial differentiation is not. This indicates that a pivotal change occurs between late blastula and early gastrula to allow for the eventual gliogenic switch to occur. The most obvious difference is the mix of signals that neural plate explants receive from their short exposure to activity from the organizer.

This difference in developmental potency also allowed us to establish an experimental system that permitted us to determine whether a given gene of interest is necessary or sufficient to initiate gliogenesis (Figure 4.6).

Figure 4.6 Gliogenesis experimental system

Figure 4.6 Explant that are capable of initiation gliogenesis (neural plates) and incapable of initiating gliogenesis (noggin animal cap.) This allows us to add or remove specific transcription factors or signaling pathway components to test their effect on gliogenesis. Scale bar is 250 µm long.

4.3 Sufficiency of sox9 and nfix in Gliogenesis During Primary Neurogenesis. In mouse and chick, sox9 and nfix are considered sufficient to initiate gliogenesis, and nfix has been shown to be necessary and sufficient for astrocyte differentiation

(Caiazzo et al., 2015; Deneen et al., 2006; Kang et al., 2012b; Stolt et al., 2003; Tiwari et al., 2018). However, their presence in noggin animal caps, which do not appear to initiate gliogenesis, are inconsistent with a putative role in gliogenesis during this stage of

Xenopus development (Chapter 4 Figure 4.4 B).

To assess a role for these transcription factors in gliogenesis, we overexpressed them in animal cap explants. We overexpressed sox9 and nfix either by themselves or together in tissue that was specified as epidermal ectoderm (uninjected animal caps), and tissue that was specified as neural ectoderm (noggin animal cap) to determine what conditions, if any, were required to initiate the gliogenic switch. Animal caps were isolated at NF stage 8 and collected at NF stage 28 and expression levels of genes of interest were compared to those of uninjected animal caps or noggin animal caps (Figure

4.7 A, B respectively).

Our results indicated that overexpression of these transcription factors in either uninjected animal caps or noggin animal caps was insufficient to initiate glial specification and differentiation. If sox9 and nfix were capable of activating gliogenesis, the levels of glial associated genes (glast, glt-1, mbp) would significantly increase as compared to those levels seen in either uninjected or noggin injected animal caps. In mouse and chick embryos, Glast is a direct transcriptional target of Sox9 and Nfia (Kang et al., 2012b) and the X. laevis glast promoter has several potential nfix binding sites.

100

While slight increases in glast expression are observed in both uninjected and noggin injected animal cap backgrounds, none of these increases are significant. There would potentially also be an increase in oligodendrocyte-associated transcription factors following initial specification of a glial progenitor cell population, which is absent in either background. These results continue to cast doubt on the sufficiency of sox9 and nfix for the gliogenic switch to occur during primary neurogenesis in X. laevis.

101

Figure 4.7 Sufficiency of sox9 and nfix to initiate gliogenesis

Figure 4.7 Induction of gliogenesis by sox9 and nfix on uninjected (A) and noggin (B) animal caps. Injections of sox9, nfix, or sox9+nfix capped RNA at the 2-cell stage was unable to increase levels of astroglial genes glast, glt-1, acpdd1, or mbp in either uninjected or noggin animal caps. Samples were normalized to the geometric mean and compared to expression levels in animal caps (± noggin) siblings. Animal caps were collected at NF stage 28. Error bars represent S. E. M. Unpaired student’s TTEST was performed to calculate significance. N=3 for all conditions tested. p ≤ 0.05.

102

4.4 Discovery Based Identification of Transcriptome Differences Between Noggin Animal Cap and Neural Plate Explants During Gastrulation and Neurulation. In mouse and chick embryos, Sox9 and Nfia fulfill pivotal roles in the initiation of gliogenesis. As our results suggest an evolutionary departure from this paradigm in X. laevis during primary neurogenesis, we performed a series of experiments comparing transcriptome profiles of gastrulation (NF stage 11) and neurulation (NF stage 18) noggin animal cap and neural plate explants. The gastrulation timepoint was chosen to identify the initial differences between neural plates (NP) (gliogenesis permissive) and noggin animal caps (NogACs) (gliogenesis deficient). We examined the neurulation timepoint to identify the differences between NPs and NogACs following initiation of the gliogenic switch.

We isolated four biological replicates each consisting of ~50 explants for NogAC and NP explants for the gastrulation comparison, and five (each replicate was set of 75-

100 pooled animal caps) NogAC and six (60-75 pooled neural plates per biological replicate) NP biological replicates for the neurulation comparison. Sequenced transcripts were aligned and counted using RSEM software (B. Li & Dewey, 2011) and differential expression of transcripts was identified using DESeq2 (Love et al., 2014). DESeq2 calculates the p-adjusted value by employing the Benjamini-Hochberg procedure to correct for false rate of discovery in large datasets. In our data, this removed genes that were significantly expressed due to overrepresentation in one biological replicate of an explant type. Following DESeq2 calculation of differential expression, data were shrunk using the Approximate Posterior Estimation for GLM (apeglm) to allow for visualization of significantly differentially expressed genes without the bias introduced by genes with

103

total counts of less than 2 (Zhu, Ibrahim, & Love, 2018). Transcript counts were also used for principal component analysis (PCA) with the procedure established by S.

Uzman.

Initial analysis of the gastrulation transcriptome found that 45,038 unique genes were identified (the .L and .S forms are counted individually). Of that, 36,344 genes were expressed at greater than zero levels, though 10,435 genes had combined mean count levels of less than 2 transcripts for each condition (NogAC or NP) (Figure 4.8 A).

NogAC and NP explants had roughly equal numbers of differentially expressed genes

(adjusted p-values less than or equal to 0.1) (Figure 4.8 A, B) and the total number of differentially expressed genes accounted for 24.53% of the total number of genes. We then compared differences between the biological replicates of each explant and found that the NogACs and NPs biological replicates were grouped tightly together with others of their explant (Figure 4.8 C,D). This allowed us to proceed with further evaluation without having to remove a dissimilar sample that might alter or confound the results.

Similarity between samples was calculated using distance heatmaps (Figure 4.8 C) and

PCA (Figure 4.9 D) on the entire gastrulation sample set.

We identified 49,088 unique genes (.L and .S forms continue to be counted separately) in the neurulation transcriptome. 38,377 were expressed in either the NogAC or NP samples, though 9,531 of those genes were expressed at mean counts of less than 2

(Figure 4.9 A). Interestingly, as opposed to the gastrulation transcriptome comparison,

NP explants had almost twice as many differentially expressed genes as NogAC samples

(4,289 vs 2,577, respectively) but the total overall number of differentially expressed

104

genes decreased (17.89% of the overall observed genes were differentially expressed)

(Figure 4.8 A and Figure 4.9 A). The relative log2fold distribution of differentially expressed genes (as shown by the MA-Plot in Figure 4.9 B) shifted towards the highly differentially expressed in NP samples (log2fold increases of greater than 2) whereas in the NogAC samples, the shift in genes towards high differential expression is less prominent. We confirmed that each biological replicate is highly comparable to the other replicates in that explant type (Figure 4.9 C, D).

When observing the summaries of both gastrulation and neurulation transcriptomes, it appears that as either the NogAC or NP samples progress in their development, they become more homogeneous; i.e. there is greater homogeneity among biological replicates at the neurula stage than the gastrula stage. This homogeneity is both among biological replicates within each type of explant (as evidenced by the closer clustering of NPs and NogACs in Figure 4.9 C, D as compared to Figure 4.8 C, D) and between NPs and NogACs as a whole. This conclusion is also supported by the finding that the percentage of differentially expressed genes during neurulation (17.89%. Figure

4.9A) was substantially lower than the corresponding percentage of differentially expressed genes in gastrulation. (24.53%, Figure 4.8 A) This was reassuring because as development progresses, the differences observed become more likely to account for the initiation of gliogenesis in NPs, and the corresponding lack of gliogenesis in NogACs.

105

Figure 4.8 Gastrula transcriptome summary

106

Figure 4.8 Differential expression comparison of NF stage 11 noggin animal caps to NF stage 11 neural plates. A. The gastrulation transcriptome had 45,038 unique genes identified and of that, 36,344 had genes with read counts above zero in at least one sample. Roughly equal numbers of genes were differentially expressed in noggin animal caps vs neural plates (padj <0.1). There were 249 genes classified as outliers (genes where one sample in a category was clearly increased compared to the range seen in other samples) and ~28% of the identified genes were classified as having a low, or mean of less than two, count. B. Shrunk MA plot describing the data from A. Shrinkage of the data using R package apeglm (Zhu et al., 2018) was applied to account for the large number of genes with counts below 2 identified in A. Genes that are highly differentially expressed (abs(log2fold value) >2) are visualized as red dots at the ± 2 line. Sample similarities and differences were visualized using distance heatmaps (C) and principal component analysis (D). The distinct groupings of noggin animal caps and neural plates indicates that all samples (4 noggin animal caps 4 neural plate samples) are highly similar to others within each grouping. Neural plate 1 and 3 were the most similar with neural plate 4 most dissimilar. Noggin animal cap 3 and 4 were most similar with noggin 2 the most dissimilar. The distance heatmap (C) and principal component analysis were plotted using rlog transformed counts for all identified genes. Counts were identified using RSEM software (B. Li & Dewey, 2011). Differential expression was calculated in R using DESeq2, (Love et al., 2014) and graphs were created in R with the gplots (Warnes et al., 2019) and ggplot2 packages (Wickham, 2016).

107

Figure 4.9 Neurula transcriptome summary

108

Figure 4.9 Differential expression comparison of NF stage 18 noggin animal caps to NF stage 18 neural plates. A. The neurulation transcriptome had 49,088 unique genes identified with 38,377 genes with non-zero read counts. Neural plates had 4,289 differentially expressed genes (padj <0.1), as compared to the 2,577 differentially expressed genes in noggin animal caps. Fifty-eight genes were classified as outliers and 9,531 genes fell into the low count category. B. Shrunk MA plot describing the data from A. Shrinkage of the data using R package apeglm (Zhu et al., 2018) was applied to account for the large number of low counts identified in A. Genes that are highly differentially expressed (abs(log2fold value) >2) are visualized as red dots at the ± 2 line. Similarities between each sample grouping (noggin animal cap n=5, neural plates n=6) were visualized using distance heatmaps (C) and principal component analysis (D). Neural plate samples 1, 2, 3, and 5 were the most similar; neural plate 4 and 6 were also highly similar to each other though more removed from the other four neural plate samples. Noggin animal cap samples 2, 3, and 4 were tightly grouped, however overall the noggin animal cap samples were highly similar among all five samples (D). The distance heatmap (C) and principal component analysis (D) were plotted using rlog transformed counts. Counts were identified using RSEM software. (B. Li & Dewey, 2011) Differential expression was calculated in R using DESeq2, (Love et al., 2014) and graphs were created in R with the gplots (Warnes et al., 2019) and ggplot2 packages. (Wickham, 2016)

109

Once we established that both the gastrulation and neurulation datasets were providing reproducible results, we evaluated the differentially expressed genes at each stage to identify how each condition diverged. We identified the highly differentially expressed genes at each stage and explant type by graphing the log2fold values against their -log10 p-value (Figure 4.10 and Figure 4.11). As an example, a p-value of 0.00001 would translate into a -log10 value of 5. Genes were categorized as highly differentially expressed (the red category) when their absolute log2fold values were greater than 2 and their p-values were lower than 0.00001. Genes that only had an absolute log2fold value of greater than 2 are marked in green, and genes that only had a p-value lower than

0.00001 were marked in blue.

We identified 350 highly differentially expressed genes in gastrula NPs and 163 highly differentially expressed genes in gastrula NogACs (Figure 4.10). The top 30 differentially expressed genes for each condition are listed in Table 4.2 (Gastrula NP) and

Table 4.3 (Gastrula NogAC). The top 30 genes were ranked based on their p-adjusted value, for further stringency.

The most significant differentially expressed genes in gastrulation NPs have functions ranging from early neural development (foxb1, pax3, irx4) to components of various signaling pathways (bmpr1b, cish) and seventeen of the top 30 are transcription factors. Excitingly, one of the most significantly expressed genes was the transcription factor ascl2. Ascl1, a close family member of ascl2, is a known pioneer transcription factor. In its hypo-phosphorylated state, Ascl1 is capable of opening otherwise closed chromatin to allow for further transcription (Ali et al., 2014). Ascl2 in Xenopus tropicalis

110

initiates its expression at NF stage 11 (Owens et al., 2016) and is widely expressed along the along the neural tube at NF stage 13 in X. laevis (Takada, Hattori, Kitayama, Ueno, &

Taira, 2005). The expression pattern and timing of ascl2 lends itself to contributing to the gliogenic switch.

The top significantly expressed genes in the NogAC gastrula set feature genes involved in cell function and cell interactions but does not include transcription factors or signaling genes (Table 4.3). This caused us to focus our attention on the transcription factors in NPs that might allow for gliogenesis to occur.

111

Figure 4.10 Gastrula highly significantly expressed genes

Figure 4.10 Highly differentially expressed genes in the gastrulation transcriptomes. A. Volcano plot identifying genes with log2fold changes of greater than 2 or less than -2 and p-value of less than 0.00001. This identified 350 highly significantly differentially expressed genes for neural plates (Table 4.2) and 163 highly significantly differentially expressed genes for noggin animal caps (Table 4.3). Volcano plots were created using the R package EnhancedVolcano (Blighe, 2019).

112

Table 4.2 Top 30 Differentially Expressed Genes in Gastrula Neural Plate - Genes marked by an asterisk are transcription factors Genes Log2Fold Change P -Value P-adjusted nr4a1.L -5.32349 5.80E-49 4.96E-45 bmpr1b.L -4.76721 2.19E-42 1.41E-38 nkx6-2.S* -5.1466 4.78E-40 1.75E-36 plxna2.S -4.86505 4.46E-33 1.04E-29 ascl2.S* -5.17807 1.25E-30 2.46E-27 nkx6-2.L* -7.07364 8.54E-29 1.57E-25 tox.L -5.52518 2.41E-27 3.87E-24 dbx1.S* -7.8385 4.02E-27 5.73E-24 prrt1.S* -4.87316 6.40E-27 8.64E-24 rfx4.S* -5.76235 4.55E-25 5.31E-22 pax3.L* -6.63331 1.75E-24 1.87E-21 scrt1.S -6.74258 8.27E-24 7.58E-21 sp5.S* -7.85423 1.75E-23 1.45E-20 irf8.S* -4.99907 2.42E-23 1.94E-20 cish.L -6.61079 8.24E-22 5.56E-19 cish.S -6.77082 1.32E-21 8.46E-19 foxb1.L* -4.82029 2.44E-21 1.43E-18 tlx3.L* -4.86027 1.26E-18 6.12E-16 myoc.L -5.49712 1.92E-18 8.82E-16 st6gal1.S -5.35813 2.51E-18 1.13E-15 pax3.S* -5.7973 3.71E-18 1.61E-15 c3.L -4.76169 2.99E-17 1.18E-14 angpt4.L -6.01952 1.17E-16 4.10E-14 foxb1.S* -5.41939 2.75E-16 9.28E-14 hoxb1.S* -7.22447 1.48E-14 3.39E-12 irx4.L* -4.7531 1.87E-14 4.22E-12 apold1.L -4.96327 2.18E-14 4.86E-12 pitx2.S* -5.76163 1.27E-13 2.50E-11 hnf1b.L* -9.24017 5.37E-13 8.88E-11

113

Table 4.3 Top 30 Differentially Expressed Genes in Gastrulation NogAC - Genes marked by an asterisk are transcription factors Genes Log2Fold Change P-Value P-Adjusted synj2.L 2.854818 3.16E-26 4.05E-23 nt5c3a.L 4.897874 8.01E-24 7.58E-21 parp3.L 3.461512 5.91E-15 1.52E-12 slc20a2.L 3.339914 1.69E-14 3.83E-12 c3ar1.L 3.210656 2.09E-13 3.83E-11 gng10.L 2.828667 5.41E-13 8.90E-11 c15orf39.S 2.925959 1.79E-12 2.66E-10 ulk1.S 2.914634 1.23E-11 1.57E-09 march8.S 3.767583 2.93E-11 3.36E-09 ift140.L 2.831903 2.96E-11 3.38E-09 gjc2.L 3.896742 4.55E-11 4.86E-09 aplnr.L 3.860815 1.61E-10 1.49E-08 eda2r.L 2.891267 6.79E-10 5.19E-08 stard4.L 2.820627 1.26E-09 8.83E-08 aicda.S 5.315878 3.99E-09 2.37E-07 filip1.L 3.212157 5.16E-09 2.94E-07 znf577.S 3.103425 5.45E-09 3.08E-07 ccnb1.L 2.975423 5.85E-09 3.26E-07 rbl2.L 3.209431 6.23E-09 3.43E-07 atp13a4.S 3.05723 6.59E-09 3.60E-07 kcnn1.S 6.799145 2.75E-08 1.17E-06 col17a1.L 3.785151 3.79E-08 1.55E-06 diras3.L 3.606851 6.02E-08 2.34E-06 adora2a.S 4.104488 7.80E-08 2.86E-06 chst8.L 2.899597 9.08E-08 3.24E-06 il15.S 2.80807 1.46E-07 4.81E-06 lancl3.S 3.191956 1.58E-07 5.17E-06 stx11.S 3.055584 1.89E-07 6.07E-06 mep1a.L 4.342716 2.48E-07 7.68E-06 tcn2.S 3.254989 3.11E-07 9.28E-06

114

In the neurulation transcriptome we identified 660 significantly differentially expressed genes in the NP sample set, and 41 significantly differentially expressed genes for NogAC samples (Figure 4.11). The top 30 differentially expressed genes for the NPs had representatives of early neural development, (pax3, irx3, pou3f2) differentiated neuronal cells, (n-tubulin, slc7a8) oligodendrocyte transcription factors, (sox8, sox10).

Excitingly, the s mostly highly differentially expressed gene in the NP transcriptome was rfx4.L, a transcription factor and orthologue of a gene that was described as being expressed specifically in astrocytes in mice (Cahoy et al., 2008). Interestingly, pou3f2, a transcription factor that in mouse spinal cord that directly activates transcription of Nfia, is the 9th most differentially expressed gene in the transcriptome (Table 4.4) (Laug et al.,

2018). Genes associated with Wnt signaling (wnt4, wnt2b) and anterior-posterior identity

(engrailed2, krox-20) were also present in the NP top 30 differentially expressed genes.

In contrast, the highly significantly differentially expressed genes in neurulation

NogAC had high numbers of transcription factors that were linked to anterior neural specification or eye development (foxg1, foxe1, vax1, vax, fezf1, lhx2, rax) (12 out of 30 genes) which is consistent with traditional characterization of NogACs (Table 4.5) (Lamb et al., 1993).

115

Figure 4.11 Neurula highly significantly expressed genes

Figure 4.11 Highly differentially expressed genes in the neurulation transcriptomes. Volcano plot identifying genes with log2fold changes of greater than 2 or less than -2 and p-value of less than 0.00001. We identified 660 highly significantly differentially expressed genes for neural plates (Table 4.4) and 41 highly significantly differentially expressed genes for noggin animal caps (Table 4.5). Volcano plots were created using the R package EnhancedVolcano (Blighe, 2019).

116

Table 4.4 Top 30 Differentially Expressed Genes in Neurula Neural Plate - Genes marked by an asterisk are transcription factors Genes Log2Fold Change P-Value P-Adjusted slc7a8.S -5.36943 2.20E-94 1.58E-90 rfx4.L* -7.77626 6.03E-77 3.47E-73 pax3.S* -6.88798 1.19E-72 4.88E-69 pax3.L* -6.45623 5.55E-71 2.00E-67 mmp28.L -5.92912 4.68E-70 1.50E-66 egr2.L* (krox-20) -6.34095 5.78E-65 1.66E-61 irx3.S* -6.47723 1.92E-57 4.26E-54 n-tubulin.S -5.32321 1.70E-55 3.26E-52 pou3f2.L* -5.06207 9.71E-49 1.40E-45 adamtsl4.S -5.83524 2.87E-46 3.76E-43 wnt4.S -5.45005 4.34E-46 5.43E-43 nrp1.L -5.13089 5.98E-46 6.89E-43 dmbx1.L* -5.06234 6.44E-46 7.13E-43 sp5.L* -5.75232 1.96E-40 1.66E-37 fstageL -5.34042 2.15E-40 1.77E-37 irx3.L* -6.13708 5.04E-40 4.03E-37 engrailed2.S* -8.40856 1.06E-39 8.22E-37 sox8.S* -6.56349 1.46E-37 1.00E-34 pnhd.L -5.49152 1.86E-34 1.01E-31 twist1.S -8.19962 3.18E-32 1.58E-29 ltbp1.S -5.32619 3.13E-32 1.58E-29 ebf2.S* -5.31358 2.42E-31 1.12E-28 nhlh1.S* -7.34245 1.86E-30 7.78E-28 foxb1.L* -6.57745 1.85E-29 7.20E-27 pdgfra.L -5.76816 2.02E-28 7.36E-26 hoxb9.S* -8.17127 4.32E-28 1.52E-25 sox10.S* -9.68291 1.19E-27 4.07E-25 foxb1.S* -7.45066 1.28E-27 4.29E-25 vgll3.L -6.68313 5.13E-27 1.62E-24 wnt2b.L -6.6628 1.02E-26 3.20E-24

117

Table 4.5 Top 30 Differentially Expressed Genes in Neurula NogAC - Genes marked by an asterisk are transcription factors Genes Log2Fold Change P-Value P-Adjusted foxg1.S* 4.422074 9.55E-37 6.25E-34 vax1.L* 4.418675 1.83E-22 3.85E-20 nog2.L 2.197963 3.08E-21 5.58E-19 six3.S* 2.659677 7.98E-19 1.13E-16 cyp27c1.S 2.195895 1.08E-18 1.50E-16 foxg1.L* 2.903985 3.80E-18 4.95E-16 foxe1.S* 3.22447 7.45E-18 9.27E-16 dscaml1.L 2.253135 2.63E-15 2.38E-13 lhx2.S* 2.195551 9.61E-15 7.90E-13 fezf1.S* 2.496844 1.25E-14 1.00E-12 fzd8.L 2.691408 1.41E-14 1.12E-12 nsmf.S 2.20944 7.43E-14 5.43E-12 cyp27c1.L 2.177108 1.44E-13 1.01E-11 vax2.L* 2.991195 2.38E-13 1.63E-11 slc2a1.L 2.014731 7.33E-13 4.70E-11 dusp5.S 2.018259 9.77E-13 6.15E-11 bhlhe22.S 2.051971 2.55E-12 1.50E-10 foxn4.S* 2.090063 7.24E-12 3.96E-10 ntn4.L 2.730113 2.64E-11 1.31E-09 six3.L* 2.748212 4.15E-11 1.99E-09 foxe1.L* 5.843482 4.68E-11 2.21E-09 cdc42ep1.L 2.200583 8.38E-11 3.82E-09 gjc2.L 2.88287 2.25E-10 9.60E-09 fzd8.S 2.168757 2.70E-10 1.14E-08 tbx3.S* 2.21313 4.54E-10 1.83E-08 rax.L* 2.104514 4.94E-10 1.98E-08 adra2a.L 2.497655 5.14E-10 2.06E-08 myl4.L 2.367483 2.78E-09 9.45E-08 tcn2.S 2.816563 3.14E-09 1.05E-07 vax1.S* 4.946095 4.15E-09 1.37E-07

118

We also investigated the differential expression for each nuclear factor one (NF1) transcription factor during neurulation (Table 4.6). This family of transcription factors had no significant differential expression for any of its members in either NogACs or

NPs. This agreed with our quantitative RT-PCR results, which showed that the level of nfix expression was not significantly changed from whole embryo siblings in either

NogACs or NPs.

Table 4.6 Expression of Nuclear Factor 1 Genes in Neurula Transcriptome

Genes Log2Fold P-Value Adj. P-Value Significance Change nfix.L 0.73487672 0.53944234 0.69162038 None nfix.S -1.0750759 0.60243347 NA None nfia.L -0.4043073 0.89739194 NA None nfia.S 1.96345156 0.01458144 0.05179399 None nfib.L 1.15045602 0.66504733 NA None nfib.S 0.62572639 0.85626464 NA None nfic.L -1.2182652 0.7302858 NA None nfia-like.1.L 1.08647411 0.68293472 NA None nfia-like.1.S -1.6364859 0.35024827 NA None

After identifying the top differentially expressed genes for each type of explant at each stage, we performed a gene ontology analysis for each list using DAVID 6.8 software and PANTHER v14 (Huang et al., 2009c; Huang, Sherman, & Lempicki,

2009b; Mi et al., 2019). The initial gene list was poorly identified (less than 80% of genes matched to a functional category) so we broadened the list to include all genes that had an absolute log2fold greater than 2 for each condition. We identified and simplified the functional categories for each condition (Figure 4.12) and compared the results for each condition.

119

We confirmed that NogACs are more developmentally restricted (Figure 4.12 A,

C), the genes that are highly differentially expressed in NogACs are associated with anterior neural development. In contrast, NPs exhibited both a broader range of gene functions associated with neural development, including functional categories such as midbrain development and general neural development. Genes differentially expressed in the neural plate also included large numbers of genes associated with signaling pathways and tissue patterning. One of the most interesting differences between the NogACs and the NPs was that NPs differentially expressed hox genes, the gene family associated with the anterior to posterior identity of the body (and the nervous system up until the midbrain hindbrain boundary). Engrailed2 and krox-20, the genes which determine the position of the midbrain hindbrain boundary, are almost exclusively expressed in NPs.

Table 4.7 lists of many of the differentially expressed genes associated with neural development and anteroposterior for all conditions. Also of interest was the differential expression of neuronal specific transcription factors neurod1/4, ngn1/2/3, and ebf2, and neuronal differentiation marker n-tubulin in neurulation NPs.

Overall, it appeared that both explants faithfully reproduce early neural development. NogACs become restricted to anterior neural development as they continue to grow, whereas NPs express genes that are found throughout entire developing nervous system. These findings are consistent with reports from previously published data (Lamb et al., 1993; Sater et al., 1993). NPs express genes that translate into diffusible signaling factors that establish anteroposterior identity, and initiate transcription of genes characteristic of a greater array of cell types. This introduces the possibility that (1) identity along the anteroposterior axis is important for the timing of the onset of astrocyte 120

development; and (2) in NPs the signals establishing posterior identity may initiate transcription of some combination of factors that signal to the anterior portion of the explant. These signals may then permit the initiation of gliogenesis in the forebrain and midbrain. We know that the synergistic action between noggin and fgf2 induce a balanced identity in isolated neural ectoderm (NPs) (Lamb & Harland, 1995); this balanced identity may contribute to initiation of gliogenesis in previously unidentified ways.

121

Figure 4.12 – GO DAVID analysis of differentially expressed genes

Figure 4.12 Gene ontology functional analysis for differentially expressed genes in gastrulation (A, B) and neurulation (C,D) transcriptomes. A, B, C, D. Simplified functional categories for gastrulation (A) and neurulation (B) noggin animal caps and gastrulation (B) and neurulation (D) neural plates. Throughout both stages, neural development in noggin animal caps remains restricted to anterior neural processes (forebrain and eye development) whereas neural plates are have differentially expressed genes that are normally expressed throughout the entire developing nervous system. See Table 4.8 for further delineation. GO analysis was performed using DAVID v6.8 software (Huang, et al., 2009; Huang, et al., 2009).

122

Table 4.7 Differntial Expression of Genes Associated with Neural Development and Anterior-Posterior Identity and Development

Gastrulation- NogAC Gastrulation- NP Neurulation- NogAC Neurulation- NP Gene Function Gene Function Gene Function Gene Function mdm1 Eye barhl1 Midbrain fezf1 Forebrain pou3f1/2 Brain development development pattern development, specification early glial dev ift140 Eye foxb1 Midbrain six3 Forebrain irx1/2/3 Brain development development pattern development specification/ eye development lamb2 Eye phox2a Midbrain lhx2 telencephalon sox8 Oligodendrocte development development regionalization development fezf1 Forebrain btg2 Anterior/ dlx2 Olfactory bulb sox9 Glial pattern posterior development specification specification patttern specification six6 Eye cyp26c1 Anterior/ hesx1 Regional sox10 Oligodendrocte development posterior specification of development patttern anterior head specification six3 Forebrain gli1/2/3 Anterior/ vax1 Ventral en2 Midbrain- pattern posterior forebrain hindbrain specification/ patttern development boundary eye specification development lhx2 eye emx2 Forebrain vax2 Ventral eye ngn1/2/3 Neuronal development development development differentiation hesx1 Regional pax6 Forebrain six6 Eye neurod1/4 Neuronal specification dorsal/ventral development differentiation of anterior pattern head formation foxe3 Eye fezf2 Forebrain and hoxa2/ Anterior - development olfactory bulb 3/5/7/10/ posterior formation and 11 identity patterning olig2 Thalmus rax Eye hoxb1/2/3/ Anterior - development development 4/5/7/9/ posterior 10/11/12 identity ptchd1 Thalmus nkx2-6 Hypothalamus hoxc3/4/5 Anterior - development development /6/8/9/ posterior 10/11/12 identity isl1 Spinal cord hoxd3/4/ Anterior - development 8/9 posterior identity hoxb1 Anterior - wnt1 Midbrain- posterior hindbrain identity boundary expression hoxc4/6/8 Anterior - foxb1 Midbrain posterior development identity hoxd4/10 Anterior - olig3/4 Oligodendrocte posterior development identity fabp7 Glial marker slc1a3 Glial marker

123

We performed principal component analysis on the top ~500 differentially expressed genes to have an unbiased comparison of the two explant types. The top 500 genes were identified based on their absolute log2fold value. Before PCA analysis occurred, we normalized the raw read counts for each of the top 500 differentially expressed genes and transformed the normalized counts by log10 to account for variation.

The PCA for the gastrula transcriptome is summarized below in Figure 4.13. The first two principal components accounted for just under 60% of the observed variation

(Figure 4.13 A) and the two conditions were distinctly grouped in Figure 4.13 B.

However, when we observed the top 10 genes accounting for variation in PCA1, the genes identified were relatively homogenous in expression across both NogAC and NP samples, with the variation seen in one sample (NP2) seeming to sway the results (Figure

4.13 C). Condition clustering between NPs and NogACs is also less distinct, leading me to the conclusion that most of the genes identified would fall into the outlier category, and are less reliable. However, while it is likely that the top10 genes identified in PCA1 are outliers, when the top 500 genes are considered, each condition is tightly clustered, leading that gene list to be more reliable.

The neurula PCA proved to be more informative, with each condition clustering reproducibly (Figure 4.14 B, C). In fact, the levels of gene expression of the five neurulation NogACs are so consistent across samples that the biplot graphs them on top of each other (Figure 4.14 B). The first principal component also accounts for slightly over 65% of the observed variation, and when the top 10 genes were observed, the clustered genes confirmed several of the genes previously identified in the highly

124

significantly differentially expressed dataset (n-tubulin, nhlh1) and identified a notch signaling pathway member (hes5_X2). We observed the top 50 genes accounting for variation in PCA1 and PCA2 for both gastrula and neurula stages (See Table 4.8 for the entire list for each condition). It wasn’t until PCA2 at both stages that genes differentially expressed in NogACs started to make a noticeable contribution to the variation between samples. Interestingly, the genes that were the most significantly differentially expressed

(Tables 4.2, 4.3, 4.4, 4.5) accounted for less variation in our PCA analysis. The DESeq2 normalizes genes based on a regularized log value, and the differences between normalization techniques could account for the observed differences. However, the genes that were reproducibly identified (mostly hox genes) are highly relevant to the differences between the two explants.

After considering all the transcriptome results, it seems likely that the difference in the competence to initiate gliogenesis for each explant is due to the presence of a factor, or set of factors, in NPs, rather than the presence of a factor in NogACs that prevents gliogenesis. Expression of nfix was not significantly increased in either explant type or at either stage. Our analysis (1) identified specific transcription factors that could be playing roles in gliogenesis; (ascl2, rfx4), (2) supported previous findings that notch signaling is important for gliogenesis through significantly increased expression of notch pathway genes in NPs, and (3) suggests that signaling pathways that establish anterior- posterior identity in the central nervous system (Wnt signaling, FGF signaling, RA signaling) could be playing as yet unexplored roles in gliogenesis.

125

Figure 4.13 Gastrula principal component analysis of top 510 differentially expressed genes

Figure 4.13 Principal component analysis on the top 510 differentially expressed genes from NF Stage 11 transcriptomes. Log2fold values were set to be greater than 3.7 (noggin animal cap expressed) or less than ≤ 3.7 (neural plate expressed.) A. Scree plot indicating what portion of variation each principal component analysis (PCA) explains. PCA1 accounts for 45.2% of the variation seen among the top differentially expressed genes. B. Biplot of the variation observed in PCA1 and PCA2. NP 2 was most dissimilar (as compared to NP 4 over the entire transcriptome set Figure 4.7) for the neural plate sample set. Nog2, in agreement with the PCA analysis on the entire transcriptome data set in Figure 4.7, was the most dissimilar. C. Expression of top 10 genes in PCA1 accounting for variation. All genes were differentially expressed in neural plates. Graphs were created using R package factoextra (Kassambara & Mundt, 2017).

126

Figure 4.14 – Neurula principal component analysis of top 534 differentially expressed genes

Figure 4.14 Principal component analysis on the top 534 differentially expressed genes from NF Stage 18 transcriptomes. Log2fold values were set to be greater than 3.5 (noggin animal cap expressed) or less than ≤ 3.5 (neural plate expressed.) A. Scree plot indicating what portion of variation each PCA explains. PCA1 accounts for 66.2% of the variation seen among the top differentially expressed genes. B. Biplot of the variation observed in PCA1 and PCA2. NP 3 and 5 were highly similar, with NP 1, 2, 4, and 6 also being highly similar. Interestingly, the all noggin animal cap samples were extremely similar. These results are similar to what was observed for the whole transcriptome PCA performed in Figure 4.8, though more variation was observed among noggin animal caps. C. Expression of top 10 genes in PCA1 accounting for variation. All ten genes were differentially expressed in neural plates. Graphs were created using R package factoextra (Kassambara & Mundt, 2017).

127

Table 4.8 Top 50 Genes Accounting for Variation in Gastrulation and Neurulation Transcriptomes

PCA1 - PCA2 - PCA1 - PCA2 - Gastrulation Gastrulation Neurulation Neurulation 1 myot.L nr4a1.L nhlh1.L egr2.L 2 sftpa1.L h2afx.S tubb2b.S hba2.L 3 megf6.L sipa1.S slc18a3.S hdac9.S 4 ptchd1.S nr1h5.S ltbp1.S plp1.L 5 dus4l.S tmem255b.S fam181b.L tal2.L 6 tex2.1.L st6gal1.S lpar5.L vax1.S 7 tmem132e.S cplx1.L hes5_X2.S slc34a2.L 8 celf2.S tie1.S nhlh1.S dhx35.L 9 hoxc11.L cux2.S lmo2.S pou3f2.L 10 mob3b.L chrm1.S mmp28.L nmi.L 11 trim55.L gpr174.S hapln3.L ND4 12 hgf.S frk.L pou4f1.2.L hhex.L 13 ttpa.L nkx6-2.S ebf2.L cd28.L 14 mmp3.L hrh1.S hoxc11.S sox17b.2.L 15 lrrc66.L dpysl4.S paqr9.L lrrc66.L 16 acta2.S sp5.S pou3f4.S ebf2.S 17 prkg1.S crb2.S neurod4.L plscr4.S 18 nppc.S nmur1.S hoxb6.S syt5.L 19 lingo1.L cldn18.L plk5.L hoxc11.S 20 tldc2.L slc32a1.S capn9.S slc18a3.S 21 neto2.L prkaa2.S tbx6.S zap70.L 22 adm.L sparc.S ednra.S adm.L 23 msr1.L tbl1x.S pcdh8.2.S capn9.S 24 unc79.L cish.S myf5.L snai2.L 25 hoxa3.S myo10.L onecut1.2.S prodh2.L 26 syt5.L hoxc4.S hoxa2.L slco2b1.L 27 wnt10a.L fam212b.L nkx6-1.L myot.L 28 pck1.S gpr124.L hoxd11.L lgals9b.L 29 chrnd.L agrp.L neurod4.S ptprc.L 30 hsd3b7.S lef1.S slc7a8.S prdm13.S 31 bai3-like.L c3.L hoxc10.L rdh9.L 32 foxb2.L fosl1.L hoxc5.L twist1.S 33 ass1.S kcnq3.S hoxd9.L map3k13.S 34 vwa7.L foxe3.L ebf2.S tmem132c.S 35 abcb1.L aplnr.S pcolce2.S hhip.L 36 trpm1.L hoxc3.L hoxb9.S ihh.L

128

Table 4.8 continued

PCA1 – PCA2 – PCA1 – PCA2 - Gastrulation Gastrulation Neurulation Neurulation 37 rpl3l.L pitx2.S map3k13.S ptprq.S 38 tpd52l1.L slc35d3.S lhx1.S pcdh8.2.L 39 cldn1.L flnc.S neurod1.S ablim2.L 40 nmi.S socs3.S pax3.L foxe1.L 41 otop3.L col17a1.L ebf3.L hoxb2.L 42 fam167a.S mpl.L frmpd1.L col9a2.S 43 grm2.S stil.S neurog1.L hspa12a.L 44 s1pr1.L hoxb1.S scrt1.S foxb1.S 45 cxcr3.S robo1.L ntn3.L tmc5.S 46 mab21l1.S junb.L hoxa3.L vax1.L 47 bcl3.S apold1.S hoxb7.L neurod4.S 48 h2afx.S sfrp1.L tfap2e.S pou4f1.2.L 49 sipa1.S tox.L hoxd10.L foxg1.S 50 nr1h5.S rhoc.S hoxa3.S jade2.S

129

4.5 In vivo Verification of Identified Genes and Signaling Pathways in Gliogenesis

To establish whether the identified transcription factors (ascl2 and rfx4) and members of signaling pathways responsible for anteroposterior development did fulfill roles in gliogenesis, we re-introduced factors to noggin animal caps or inhibited them in neural plates.

To test whether ascl2 plays a role in initiating gliogenesis, we cloned the WT gene and overexpressed it in noggin animal caps (Figure 4.15). Both ascl2.L and ascl2.S were significantly differentially expressed in neural plates during gastrulation.

We found that in comparison to sibling noggin-injected animal caps, noggin + ascl2 animal caps were unable to rescue expression of glial associated genes sox10, olig3, and glast to levels consistent with those observed in whole embryos. In fact, the overexpression of ascl2 was associated with significant decreases in expression of the radial glial marker nestin, and nfix.

While initially disappointing, our cloned ascl2 had all 27 serine residues intact out of a 161 amino acid protein (Accession # NP_001079106.1). Given the prevalence of serines in ascl2, it seems possible that the exogenous ascl2 was hyperphosphorylated and was potentially unable to act as a pioneer transcription factor. Further experiments with an ascl2 mutated to reduce its ability to be phosphorylated would need to be pursued before this avenue was excluded.

130

Figure 4.15 Rescue of glial associated genes through overexpression of ascl2

Figure 4.15 Overexpression of ascl2 in noggin injected animal caps. No significant increase in expression was observed for astroglial associated genes that are otherwise low in noggin animal caps as compared to whole embryos (sox10, olig3, glast). Samples were normalized to the geometric mean. Noggin + ascl2 animal caps were compared to noggin animal caps. Animal caps were collected at NF stage 28. Error bars reflect S. E. M. Unpaired student’s TTEST was performed to calculate significance. N=3, p ≤ 0.05.

131

To confirm the role notch signaling played in gliogenesis, we first identified genes that were differentially expressed in each of the transcriptomes (Figure 4.16 A).

Interestingly, no genes associated with Notch signaling were identified in NogACs at either stage. Conversely, Notch signaling genes were enriched in NPs at both stages. We then inhibited notch signaling in neural plates using the small molecule RO4929097.

RO4929097 is a γ-secretase inhibitor which prevents NICD cleavage and prevents transcriptional activity (Myers & Appleby, 2014). We isolated the neural plates without the underlying mesoderm and grew them until NF stage 28 in a 1X MMR with 100 µM

RO4929097 diluted in DMSO or DMSO alone. Positive controls were included in the form of genes known to be directly affected by each signaling pathway inhibited.

Inhibition of notch signaling led to significantly reduced expression of olig3 and glast (Figure 4.16 B). We observed a similar expression decrease in noggin animal caps.

Expression of sox9 decreased, though not significantly. Interestingly, fabp7, which is unchanged in expression in noggin animal caps, significantly decreased. This is consistent with previously published research that suggests NICD directly activates transcription of fabp7 in radial glial cells (T. E. Anthony et al., 2005). Oddly, expression of acpdd1 significantly decreased and expression of glt-1 was unchanged. The lack of change in glt-1 expression suggests that the significant decrease in fabp7 and glast, coupled with the decrease of nestin expression, may be correlated with a reduction in the number of radial glial cells along with decreases in astrocytes and oligodendrocytes. This experiment did confirm the importance of notch signaling in aspects of glial specification and differentiation.

132

We noted expression of numerous genes involved in Wnt signaling in noggin animal cap and neural plate transcriptomes during gastrulation and neurulation (Figure

4.17 A). This signaling pathway was the most prolifically expressed in both types of explants from both stages, although far greater numbers of Wnt signaling pathway members were observed in neural plates. We inhibited Wnt signaling in isolated neural plates using the small molecule IWR at 100 µM concentration. IWR inhibits canonical

Wnt signaling by stabilizing axin proteins in the β-catenin destruction complex (Myers &

Appleby, 2014), thus blocking the accumulation of β-catenin. Samples were collected at

NF stage 28 for transcript expression analysis.

In comparison to untreated control neural plates, Wnt inhibited neural plates exhibit no change in expression of astroglial genes such as sox10 or olig3, and an increase (though insignificant) expression of glast, glt-1, and fabp7 (Figure 4.17 B). This indicates that Wnt-inhibited neural plates are equally competent as untreated neural plates in initiating glial development. The only gene that decreased significantly in expression was mbp, a gene associated with differentiated oligodendrocytes. However, these results are consistent with findings from human and mouse cell lines implicating Wnt signaling in the timing of oligodendrocyte differentiation (Guo et al., 2015; H. K. Lee et al., 2015;

Shimizu et al., 2005).

133

Figure 4.16 – Notch inhibition in neural plates

A. Notch Signaling Pathway Significantly Differentially Expressed Genes NogAC Neural Plate NogAC Neurulation Neural Plate Gastrulation Gastrulation Neurulation angpt4.L jag1(L/S) hhex.S hes2.L hes7.L hes7.2.L mdk.S hes5.2(L/S) snai2.L hes5_X2(L/S) hes5.1(L/S) hes9-1(L/S) hes8.L

Figure 4.16 A. List of genes related to notch signaling that were differentially expressed transcriptomes of each explant type. Genes were identified through DAVID and PANTHER gene ontology analysis. B. Inhibition of Notch signaling by 100 µM RO4929097. Significant reduction in expression of astroglial-associated genes olig3, acpdd1, glast, and fabp7 was observed. No change was observed in early neural progenitor (sox2) or neuronal associated gene (neurod1, ngn1). Positive controls hes4 and meis1 decreased in expression, with primary NICD target hes4 significantly downregulated. Samples were normalized to the geometric mean and neural plates treated with RO4929097 were compared to untreated neural plates. Neural plates were collected at NF St 28. Error bars represent the S.E.M. Unpaired student’s TTEST was performed to calculate significance. N=4, p ≤ 0.05.

134

Figure 4.17 Wnt inhibition in neural plates A. WNT Signaling Pathway Significantly Differentially Expressed Genes NogAC Neural Plate NogAC Neural Plate Gastrulation Gastrulation Neurulation Neurulation ppp3r1.L wnt8a(L/S) fzd8(L/S) wnt1.L nkd1.S myh6(L/S) fzd5.L pcdh8.L csnk1g2.L nr4a2(L/S) pcdh8.2(L/S) snai2.L pcdh8l.S lrp4.S act3(L/S) axin2.(L/S) en(L/S) cpz.S wnt8a(L/S) hhex.S prkch.L kremen2 (L/S) wnt3.L pitx2(S/L) wnt3a.L tle1.S wnt10b.L isl11 (L/S) wnt10a.L egr1(L/S) wnt11.L prickle1.L wnt11b.L ror2.S pcdh20(L/S) wnt1.L axin2.S itpr3.S cdh6(L/S) act3(L/S) dvl1.L bmpr1b(L/S) mmp3.L mycn(L/S) wnt2b.L cer1(L/S) wnt4.S cdh20.S cdh11(L/S) pcdh7.S actc1(L/S) lef1.S fzd10(L/S) cdh2.L fzd7(L/S) actc1(L/S) fzd1.L xarp.L rspo2.L acpdd1.L pax3(L/S)

135

Figure 4.17 A. List of genes related to Wnt signaling that were differentially expressed in transcriptomes from each type of explant. Genes were identified through DAVID and PANTHER gene ontology analysis. B. Inhibition of Wnt signaling through use of the small molecule, IWR, at a 100 µM concentration. Significant reduction of expression was observed only in oligodendrocyte associated gene mbp. This is consistent with previously reported roles for wnt signaling in oligodendrocyte differentiation(Guo et al., 2015; H. K. Lee et al., 2015; Shimizu et al., 2005). Positive controls hoxd1 and twist exhibited no change in expression. Samples were normalized to the geometric mean and treated neural plates were compared to untreated neural plates. Neural plates were collected at NF Stage 28. Error bars represent S. E. M. Unpaired student’s TTEST was performed to calculate significance. N=5, p* ≤ 0.05.

136

Two remaining signaling pathways play a role in anteroposterior signaling in the developing neural tube: FGF and retinoic acid. As both pathways establish anteroposterior identity through concentration gradients which regulate HOX gene expression (Schilling, 2008), both are likely actively increased in NPs as compared to

NogACs (Table 4.8). Further examination of differentially expressed genes in the transcriptomes for each explant type and stage confirmed that both pathways had more signaling pathway members expressed in NPs at both gastrulation and neurulation stages, and that genes associated with either FGF or RA signaling were absent in neurulation

NogACs (Figure 4.18 A and Figure 4.19 A). RA signaling had fewer genes differentially expressed in NPs than the number of genes differentially expressed in NPs for FGF signaling. FGF genes that were differentially expressed are also the actual signaling proteins (Figure 4.18 A). This suggests that FGF signaling may play a more direct role in gliogenesis.

We used the FGF receptor inhibitor SU5402 at a 100 µM concentration to block

FGF signaling in neural plates (Fletcher & Harland, 2008). SU5402 inhibits FGF signaling by blocking tyrosine kinase activity, which in turn prevents phosphorylation, and subsequent activation, of FGFR3, VEGFR, and PDGFR. Neural plates were collected at NF stage 28 for quantitative RT-PCR.

Neural plates treated with SU5402 exhibited significantly decreased expression levels of neurod1, ngn1, sox9, glast, glt-1, and fabp7 as compared to untreated neural plates (Figure 4.18 B). This suggests that neurogenesis and astrogenesis were affected by

FGF inhibition, whereas oligodendrocyte differentiation was unaffected.

137

Figure 4.18 FGF inhibition in neural plates

A. FGF Signaling Pathway Significantly Differentially Expressed Genes NogAC Neural Plate NogAC Neural Plate Gastrulation Gastrulation Neurulation Neurulation pik3ca.L fgf8.L fgf14.S ppp2r2b.L spry1(L/S) prkch.L fgf16.L fgf3.S fgf6.L fgf7.L fgf20.L fgfbp2.L

Figure 4.18 A. List of genes related to FGF signaling that were differentially expressed transcriptomes for each type of explant. Genes were identified through DAVID and PANTHER gene ontology analysis. B. Inhibition of FGF signaling by SU5402, at a concentration of 100 µM. Highly significant reduction in expression of neuronal transcription factors neurod1 and astroglial marker fabp7 was observed. A significant reduction in expression of neuronal transcription factor ngn1 astroglial associated genes glast and glt-1. The positive controls hes4 and meis1 were unchanged in expression. Samples were normalized to the geometric mean and neural plates inhibited by SU5402 were compared to uninhibited neural plates. Neural plates were collected at NF St 28. Error bars represent S. E. M. Unpaired student’s TTEST was performed to calculate significance. N=4, p* ≤ 0.05, p** ≤ 0.001.

138

We inhibited retinoic acid signaling using 4-diethylaminobenzaldehyde (DEAB) at a 5 µM concentration. Neural plates inhibited by DEAB were collected at NF stage 25

(unlike all other experiments discussed in Chapter 4) as the small molecule proved toxic at efficacious (≥5 µM) quantities. DEAB is an aldehyde dehydrogenase inhibitor, which prevents retinol from being converted to retinoic acid (Rankin et al., 2018). DEAB results described below are preliminary until further biological replicates are collected.

Neural plates treated with DEAB exhibited significantly decreased expression levels of neurod1, ngn1, sox9, glt-1 and fabp7 as compared to untreated neural plates and

DEAB inhibited neural plates had obviously, though not significantly decreased levels of glast (Figure 4.19 B).

These results were interesting as they implied a role for RA signaling in encouraging both initiation of neuronal differentiation and initiation of astrocyte differentiation. However, as there is no change in oligodendrocyte associated genes for

RA-inhibited neural plates, it appears unable to affect oligodendrocyte differentiation.

These findings are in strong agreement with those observed for FGF signaling.

Overall, it appears that neither FGF nor RA signaling directly contributes to the gliogenic switch in neural plates. Therefore, their presence or absence in noggin animal caps would not affect the lack of gliogenesis observed.

139

Figure 4.19 Preliminary retinoic acid receptor inhibition in neural plates A. Retinoic Acid Signaling Significantly Differentially Expressed Genes NogAC Neural Plate NogAC Neural Plate Gastrulation Gastrulation Neurulation Neurulation pik3ca.L rarb.S cd38.L ppp2r2b.L rarg.S igfbp2.S

Figure 4.19 A. List of genes related to retinoic acid signaling that were differentially expressed in each transcriptome condition. Genes were identified through DAVID and PANTHER gene ontology analysis. B. Preliminary results of inhibition of retinoic acid receptor through use of the small molecule, 4-diethylaminobenzaldehyde (DEAB), at a 5µM concentration. Significant reduction of neuronal transcription factors neurod1 and ngn1, along with early glial transcription factor sox9 was observed in RAR inhibited neural plates. We also observed a significant reduction in expression of early glial associated gene fabp7 and astrocyte associated gene glt-1. Expression of glast was reduced in neural plates inhibited by DEAB. No other significant change in expression was observed, including in the signaling controls cdx1 and ets1. The signaling controls were chosen due to their purported direct upregulation by retinoic acid signaling.(Balmer & Blomhoff, 2002) Samples were normalized to the geometric mean, and neural plates inhibited by DEAB were compared to uninhibited neural plates. Neural plates were collected at NF St 25. Error bars represent S. E. M. Unpaired student’s TTEST was performed to calculate significance. N=2, p ≤ 0.05.

140

4.6 Summary

We observed that neural plates with or without mesoderm were both capable of initiating expression of gliogenic genes at levels comparable to those observed in whole embryos. Additionally, neural plates without mesoderm expressed higher levels of oligodendrocyte-associated genes (Figure 4.2 B). Inhibition of bmp4, which is sufficient to initiate expression of neurogenic genes, is insufficient to initiate gliogenesis and the continued inhibition of bmp4 is not responsible for the lack of gliogenesis (Figure 4.4 B,

Figure 4.5).

These results are generally consistent with the observation made in Chapter 3 showing that the gliogenic switch is initiated in different regions along the anteroposterior axis, starting in the anterior spinal cord/posterior hindbrain. Neural plate explants incorporate the dorsal ectoderm that corresponds to the entire anteroposterior axis of developing neural tube; from the hindbrain and spinal cord precursor cells located

10-15 cell diameters above the dorsal lip, to the cells corresponding to the presumptive forebrain. Noggin animal caps isolate only the most animal pole grouping of cells in blastula embryos. Neural plates are exposed to the initial burst of posteriorizing signals from the organizer, whereas noggin animal caps do not receive these signals. This leads to the NogAC developing strictly into the most anterior neural tissue (Lamb & Harland,

1995; Lamb et al., 1993). However, anterior neural tissue is the last place to initiate the gliogenic switch. This in intriguing, as we have observed that NogACs do not initiate gliogenesis, whereas NPs do. It appears likely that signals from either the posterior portion of the early spinal cord or the organizer prepare the anterior portion of the neural plate to acquire the competence to initiate gliogenesis.

141

This hypothesis is supported by the enrichment of genes associated with Wnt,

FGF and RA signaling in the neural plates (Figure 4.17 A, 4.18 A, Figure 4.19 A). When

Wnt signaling was inhibited in neural plates, the only significant gene expression reduction was observed in the oligodendrocyte associated gene, mbp (Figure 4.17 B).

This indicates that the influence of Wnt signaling on glial development appears to be limited to oligodendrocyte differentiation. This finding is consistent with current research

(Guo et al., 2015; H. K. Lee et al., 2015; Shimizu et al., 2005). Inhibition of FGF signaling inhibition led to the reduction in neuronal (neurod1, ngn1) and astrocyte associated genes (sox9, fabp7, glast, glt-1) as compared to untreated neural plates. This indicates that FGF signaling has effects on neuronal differentiation and astrocyte differentiation but doesn’t appear to affect oligodendrocyte differentiation. This agrees with recently published work linking FGF signaling to increased expression of glt-1 and enhanced astrocyte proliferation (Savchenko et al., 2019). Preliminary results on the effect of RA signaling inhibition on neural plates saw reduction in neuronal (neurod1, ngn1) and astrocyte associated genes (sox9, fabp7, glt-1). These findings were highly similar to those observed in FGF inhibited neural plates. RA may encourage neuronal and astrocyte differentiation in Xenopus laevis. Taken together, the data support roles for

FGF and RA signaling in astrocyte differentiation, and a function for Wnt signaling in oligodendrocyte differentiation.

One area of concern with these experiments is failure of the positive controls to decrease in expression following Wnt, FGF, or RA signaling inhibition (Figure 4.17 B,

4.18 B, 4.19 B). This is likely due to the genes chosen not being directly impacted by the

142

signaling pathway chosen. New positive controls have been chosen and will be tested to confirm true inhibition of these pathways.

Inhibition of notch signaling in neural plates led to the reduction in expression of olig3, apcdd1, glast, and fabp7 (Figure 4.16 B) as compared to untreated neural plates.

These astroglial genes are associated with both astrocyte (glast, fabp7) and oligodendrocyte (olig3, apcdd1) lineages. This indicates that inhibition of Notch signaling affects the differentiation of both glial lineages in Xenopus primary neurogenesis. There was no decrease in expression of either sox9 or nfix. This is interesting, as in mice, Hes1 and Hes5 are reported to encourage the transcription of these genes’ orthologs (Deneen et al., 2006; Taylor et al., 2007). These experiments continue to support an evolutionary divergence in gliogenesis between anmiotes and Xenopus.

Inhibition of notch signaling in neural plates affects both glial lineages (reduced expression in sox9, fabp7, glast, glt-1), however this effect appears slightly different than the lack of gliogenesis observed in NogACs (decreased expression in sox10, olig3, glast, glt-1). These differences can potentially be attributed to a combination of severely reduced Notch, FGF, or RA signaling. It is also possible that the absence of other transcription factors in the NogACs is responsible for the phenomenon observed. Four potential candidates are the transcription factors pou3f2, rfx4, or pstat3. Pou3f2 and rfx4 are highly differentially expressed in neural plates and all three genes have previously identified roles in gliogenesis in mouse and chick models (Ali et al., 2014; Bonaguidi et al., 2005; Cahoy et al., 2008; Laug et al., 2018). This makes them likely to play roles in gliogenesis in Xenopus.

143

Finally, our results support an evolutionary divergence between gliogenesis during Xenopus laevis primary neurogenesis, and gliogenesis in mouse and chick models

(Figure 4.7). Overexpression of sox9 and nfix in noggin animal caps was unable to rescue the initiation of gliogenesis in this explant, and thus are not sufficient to initiate the gliogenic switch. We also found that no nuclear factor transcription family member was differentially expressed in neural plates, where they could contribute to gliogenesis

(Table 4.6). However, Xenopus laevis has two phases of neurogenesis; sary neurogenesis, during which the tadpole brain becomes more similar to the brains of amniotes, may undergo gliogenesis in a fashion more akin to mouse and chick models.

144

Chapter V

Functional Analysis of TAZ on

neurogenesis and gliogenesis

145

5.1 Overview of Canonical and Aberrant Roles of TAZ in Development and Disease

The transcriptional coactivator with PDZ-binding motif (TAZ) is known for its role in maintaining cell proliferation and delaying differentiation (Dobrokhotov,

Samsonov, Sokabe, & Hirata, 2018; Noguchi, Saito, & Nagase, 2018; Pfleger, 2017;

Wang, Liu, Heallen, & Martin, 2018; Zhang et al., 2009). TAZ and Yes-associated protein (YAP), its transcriptional coactivator, accomplish this through interactions with various binding partners to affect transcription. The best characterized binding partners are transcription factors in the TEA domain/transcription enhancer family (TEAD/TEF) which promote or suppress growth in the presence or absence of YAP/TAZ, respectively

(Pfleger, 2017; Zhang et al., 2009). The HIPPO pathway is often responsible for downregulating YAP/TAZ cell proliferation (Noguchi et al., 2018).

TAZ, also known as WW domain containing transcription regulator 1 (WWTR1), has been shown to be aberrantly expressed in glioblastomas (GBM) (Balasubramaniyan et al., 2015; Bhat et al., 2011). GBMs account for half of all primary tumors in the central nervous system, and patients with GBM have a median survival rate of 12 months without treatment, and 15-18 months with treatment, after their initial diagnosis (Bi &

Beroukhim, 2014; Laug et al., 2018). Microarray and genomic analyses of patient GBM samples have identified four distinct molecular subtypes: proneural, neural, classical, and mesenchymal (Lei et al., 2011; Phillips et al., 2006; Verhaak et al., 2010). GBM subtypes have clinical relevance: the proneural subtype is associated with a 50% survival rate of 4 years following initial diagnosis, whereas mesenchymal subtype tumors have a

50% survival rate of ~ 1 year following diagnosis (Noushmehr et al., 2010; Parsons et al.,

146

2008). A mesenchymal molecular subtype is also associated with recurrent GBM

(Verhaak et al., 2010).

TAZ expression is associated with the mesenchymal GBM subtype. In association with TEAD, TAZ has been shown to initiate the transcription of mesenchymal genes such as CD44, and TAZ has been described as a master regulator of mesenchymal differentiation in GBM (Balasubramaniyan et al., 2015; Bhat et al., 2011; Zhang et al.,

2009). In contrast, the transcriptional profiles of “TAZ-low” GBM were enriched with neuronal and glial transcription factors such as LHX2, LHX4, NEUROD2, HOXB3,

SOX10, and SOX8 among others (Cancer Genome Atlas Research Network et al., 2013).

These analyses prompted our interest in the effect of TAZ on neurogenesis and gliogenesis in healthy neural tissue. In collaboration with Krishna Bhat at MD Anderson

Cancer Center, we sought to determine whether aberrant expression of TAZ promotes persistence of early neural cell proliferation, potentially weakening neurogenesis and gliogenesis in Xenopus laevis. To do this, we performed a series of overexpression studies, using wild-type and variant forms of TAZ, via microinjection of synthetic capped mRNA. By varying the site of injection, TAZ could be overexpressed either throughout the whole embryo or targeted to the prospective anterior neural tissue (Figure 5.1).

147

Figure 5.1 Experimental injection design

Figure 5.1 Xenopus embryos were injected with capped TAZ RNA in either both cells at the 2-cell stage, or two of 8 cells at the 8-cell stage. Two-cell injections, which had been used in preliminary research by Dr. Pierre McCrea’s lab, affect development of the entire embryo. In contrast, injection of the two dorsal animal blastomeres at the 8-cell stage leads to overexpression restricted to the anterior neural plate, allowing a more direct observation of the effect of TAZ overexpression on neural development. In our studies, targeted 8-cell injections were used to study neurogenesis, gliogenesis, and the effects of TAZ overexpression on mesenchymal genes, whereas 2-cell TAZ injections were used to observe the effects of TAZ overexpression on mesenchymal genes.

148

5.2 Summary of Preliminary Xenopus laevis Research

Preliminary studies by Dr. Pierre McCrea and colleagues suggested that aberrant TAZ expression in Xenopus laevis affected neural differentiation. These studies included a series of TAZ injections at the two-cell stage (Figure 5.1) and observed the effects on anterior and eye development at mid tailbud stage (NF stage 28). The defects ranged from small or missing eyes to complete loss of anterior structures in the embryo.

TAZ overexpression led to a reduction in the expression of genes associated with neuronal differentiation, (xash3, n-tubulin) eye and glial development, (otx2, nefm) and neural crest differentiation (twist, slug) as compared to uninjected siblings. In contrast, some neuronal differentiation (xif3, neurod4) and mesoderm associated genes (xhox3, myod, xnot) showed no change in expression. In situ hybridizations against neurod1, twist, and krox20 in embryos injected with TAZ into one of the two cells at 2-cell stage led to a marked decrease or absence of expression of these genes in the injected half.

Embryos that had been injected with an inactive TAZ showed no obvious reduction of expression for these genes between the injected and uninjected halves. Taken together, these results suggest TAZ might play an antagonistic role towards neuronal, neural crest, and glial differentiation.

149

5.3 Replication of 2-cell Stage Injections

In previously published literature, TAZ variants are referred to as TAZ4SA for the variant associated with greatly increased transcriptional activity, and TAZ4SA-S51A for the variant that is transcriptionally silent (Bhat et al., 2011; Zhang et al., 2009). For the purposes of this chapter, they will be referred to as “TAZ-active” in place of TAZ4SA, and

‘TAZ-silent” instead of TAZ4SA-S51A.

Our lab received plasmids containing the human TAZ variants (TAZ-WT, TAZ- active, and TAZ-silent) from Krishna Bhat’s lab and sequenced them to confirm the presence of the expected mutations (Figure 5.3). Previous research had identified four serine residues that were important for regulating TAZ activity (Lei et al., 2011; Zhang et al., 2009). Mutation of these four serines to alanines (66, 89, 118, and 311) generated a constitutively active TAZ (“TAZ active”). If a final serine residue (51) was mutated to alanine, the resulting protein would bind target sequences, but was transcriptionally inactive (“TAZ silent”). Human and Xenopus TAZ proteins are 77% similar overall, their

DNA binding sequences are identical, and the five identified serine residues that contribute to TAZ activity are conserved. Given this high degree of conservation, we feel confident that human TAZ will have a similar functional effect in Xenopus as those seen in human tissue.

When we evaluated our newly sequenced clones and aligned our predicted protein sequences against the TAZ RefSeq sequence (Accession # NP_056287.1) we identified unexpected mutations in each variant. TAZ-WT had two unexpected differences, F32C and E123D. The first mutation is the most worrying, as it introduces a potential cysteine

150

disulfide bond, though cysteine is relatively scarce in the TAZ protein. The s mutation retains its negative charge and is less likely to interfere with protein sary and tertiary structures. TAZ-active had three of the four expected serine-to-alanine mutations (S66A,

S89A, S118A), along with an unexpected mutation, E309G, that may have been introduced when attempting to mutate serine 311. The change from glutamic acid to glycine might strongly affect the protein folding and function of this variant. Outside of the four mutations outline above, TAZ-active most closely resembled the TAZ reference sequence. Finally, TAZ-silent had the five expected serine to alanine mutations, (51, 66,

89, 118, and 311) but also had two unexpected mutations. The first, the E123D mutation outlined above and shared by TAZ WT, is not expected to affect the overall function of the protein. The s mutation, K148Q, replaces a positively charged residue (lysine) with a neutrally charged amino acid (glutamine). This could potentially modulate protein function, as this mutation is in the DNA binding site and might be preventing correct binding of TAZ targets. We believe that these clones were used in Dr. Bhat’s current research. Dr. Bhat’s group has found that when TAZ is expressed in glioma stem cells that mesenchymal differentiation is induced (Bhat et al., 2011). We are interested in exploring how in vivo aberrant expression of TAZ effects neural development.

To confirm the earlier findings on TAZ overexpression in Xenopus, we injected embryos at the two-cell stage with capped RNA made for each TAZ variant from the clones described above (See Figure 5.1 for description of microinjections). Embryos were allowed to develop until NF stage 18 and stage 25 and collected for quantitative RT-PCR.

TAZ-active and TAZ-WT injected embryos were compared to siblings injected with TAZ- silent (Figure 5.4). Genes associated with early neural development (geminin, sox2), 151

neuronal differentiation (neurod1), glial differentiation (sox9, nfix, sox10), and mesenchymal associated genes (tgbf1, cd44, ccn1, ccn2) showed no significant change in expression at either stage (NF stage 18 or 25) or for either TAZ variant injected. For neurod1, this is inconsistent with the results of whole mount in situ hybridizations in Dr.

McCrea’s studies (See section 5.2). Interestingly, at NF stage 25, expression of twist was significantly downregulated in TAZ-active embryos (Figure 5.4B). There was no change in expression for twist at NF stage 18, or at either stage for TAZ-WT embryos. This partially confirms the previously recorded observation that showed lowered or abolished expression of twist in TAZ-active embryos as compared to uninjected embryos or embryos expressing TAZ-silent.

152

Figure 5.2 Comparison of human TAZ sequences

Figure 5.2 Comparison of human TAZ sequences. K. Bhat introduced mutations to human TAZ (mRNA Accession # NM_015472.5, protein Accession # NP_056287.1) to create inactive (silent) and highly active (active) versions of TAZ. We sequenced these variants along with the WT variant and confirmed the following mutations as compared to the RefSeq protein. TAZ-WT had two identified differences, F32C and E123D. TAZ- silent had eight mutations, M36I, S51A, S66A, S89A, S118A, E123D, K148Q, S311A. Serine to Alanine mutations 66, 89, 118, and 311 increase activation of TAZ; however, the serine to alanine mutation at amino acid 51 causes a loss of activity of the TAZ protein. TAZ-active had four observed mutations, S66A, S89A, S118A, E309G. Outlined in blue is the FLAG tag added to each TAZ variant. In green is the DNA binding motif. Expected mutations are marked in red, with new mutations marked in purple.

153

Figure 5.3 Early neural and gliogenic transcription factor networks.

A. B.

Figure 5.3 Early neural and gliogenic transcription factor networks. A. Early neural transcription factor network. Transcription factors shown play key roles in stabilizing neural fate while encouraging neural progenitor proliferation. Geminin has been shown to control neuronal differentiation through brahma-related gene 1 (brg1), one of two catalytic components of the SWI/SNF chromatin remodeling complex. Brg1 interacts with both neurogenenin1 (ngn1) and neuroD proteins, which both lead to neuronal differentiation (Seo et al., 2005). Figure adapted from Rogers et al (2009). B. Regulation of mammalian gliogenesis. Increased Sox9 and Pou3f2 expression in neural progenitor cells directly initiates transcription of Nfia, which forms a complex with Sox9 that initiates expression of genes mediating glial specification. Continued expression of Nfia with other transcription factors initiates astrogenesis. Increased expression of transcription factors Sox10 and Olig3 initiate oligodendrogenesis. Figure adapted from Namihira et al (2009), Kang et al (2012) and Glasgow et al (2014).

154

Figure 5.4 2-cell TAZ variant injections

Figure 5.4 Effects of TAZ overexpression on gene expression in embryos injected at the 2-cell stage. Genes assayed include neural (geminin, sox2, neurod1, sox9, nfix, sox10), neural crest (twist), and mesenchymal (tgfb1, cd44, ccn1, ccn2) associated genes. Quantitative RT-PCR on NF stage 18 (A) and stage 25 (B) whole embryos. A. The genes identified to verify and expand Dr. McCrea’s set showed no significant changes in either TAZ-WT or TAZ-active embryos during neurulation. B. At NF Stage 25 (post-neural tube closure), levels of twist, the neural crest associated gene, was significantly downregulated for TAZ-active embryos. No other genes for either TAZ variant were significantly changed. Samples were compared to their silent TAZ overexpressing siblings. Error bars represent S.E.M. Unpaired student’s T-TEST was performed to calculate significance. N=3 *p≤0.05.

155

5.4 Effects of TAZ Overexpression on Neurogenesis Our preliminary experiments in 2-cell whole embryo injections provided no evidence that overexpression of active TAZ inhibits expression of neuronal transcription factors and associated genes, which is inconsistent with the earlier findings from Dr.

McCrea’s lab (Sections 5.2 and 5.3). To evaluate the specific effects of overexpression of

TAZ in the neural ectoderm, we injected mRNA encoding each of the TAZ-variants into the two dorsal animal blastomeres of the 8-cell embryo, which contribute to the anterior portion of the neural plate (Figure 5.1). The injected embryos were cultured until mid- neurulation, and either used for in situ hybridizations (NF stage 15-16,) or collected in

TRIzol for quantitative RT-PCR (NF stage 18). Each set of experiments was compared to

TAZ-silent injected embryos.

Genes of interest were chosen both from the preliminary results outlined in section 5.2 and from the early neural transcription factor network (Figure 5.3).

Expression of TAZ-WT and TAZ-active affected transcription factors associated with the early neural gene regulatory network shown in Fig. 5.4, such as geminin and sox2.

Geminin was significantly increased in embryos injected with either TAZ-WT or TAZ- active and sox2 was significantly increased in TAZ-WT. Targeted expression of TAZ had no effect on the other early neural transcription factors tested (foxd41l1, zic2, sox11, otx2), suggesting only a limited connection between aberrant TAZ expression and increased neural progenitor proliferation. Expression of the neurogenic transcription factors neurod1, ngn1, and the differentiated neuronal marker n-tubulin did not exhibit any changes when either TAZ-WT or TAZ-active was expressed (Figure 5.4).

156

To determine whether aberrant expression of TAZ leads to alterations in the spatial patterns of sox2 expression, an in situ hybridization was performed against sox2 on NF stage 15 embryos that had been injected the TAZ variants in dorsal animal blastomeres as described previously (Figure 5.5). Sox2 showed significant spatial increases in the mid neural plate (MNP) of both TAZ-WT or TAZ-active injected embryos, as compared to embryos injected with TAZ-silent. TAZ-WT injected embryos were also significantly spatially increased in the spinal cord (SC) as compared to TAZ-silent.

Neither TAZ-WT or TAZ-active injections significantly increased the width of the anterior neural plate (ANP) as compared to TAZ-silent embryos. Interestingly, the width of neural plates in embryos injected with TAZ-active and stained against sox2 fluctuated between embryos; some TAZ-active neural plates were similar to uninjected or silent TAZ injected embryos (Figure 5.6 D ii), whereas other TAZ-active neural plates were greatly expanded

(Figure 5.6 D i). This heterogeneity could explain why a statistically insignificant increase in sox2 transcript expression was observed in TAZ-active embryos shown in

Figure 5.5.

157

Figure 5.5 TAZ regulation of genes associated with neural proliferation and neuronal differentiation.

Figure 5.5 Effects of TAZ overexpression in targeted injections on genes associated with neural proliferation and neuronal differentiation. Quantitative RT-PCR experiment on NF stage 18 whole embryos specifically overexpressing TAZ-WT, TAZ-active, or TAZ-silent in the presumptive neural ectoderm. No significant change was observed for most early neural transcription factors (foxd41l1, zic2, sox11, otx2), neurogenic transcription factors (neurod1, ngnr1), or markers of neuronal differentiation (n-tubulin) in embryos overexpressing TAZ-WT or TAZ-active. Expression of two early neural transcription factors, geminin and sox2, were significantly increased in TAZ-WT expressing embryos, and expression of geminin was also significantly elevated in embryos overexpressing TAZ-active. Expression of neurod1 experienced an average fold change of 28.97, with a standard error of 18.47 in TAZ-active embryos. Graph was simplified to allow for easier viewing. Samples were compared to their TAZ-silent overexpressing siblings. Standard error was calculated for error bars. Unpaired student’s TTEST was performed to calculate significance. N=4 *p≤0.05

158

Figure 5.6 Expression of sox2 in the neural plate in TAZ variant embryos.

159

Figure 5.6 Changes in spatial expression of sox2 following targeted injection of TAZ. A. Diagram of NF stage 15/16 Xenopus laevis embryo. Neural plate shown in blue. Three areas of interest on the neural plate; anterior neural plate (ANP), mid neural plate (MNP), and spinal cord (SC) were measured for each embryo then compared to the diameter of the whole embryo. B-C. In situ hybridizations against sox2 in the presumptive neural plate for embryos overexpressing TAZ. B. TAZ-silent embryos exhibited normal spatial expression of sox2 neural plate. C. TAZ-WT embryos consistently showed increased sox2 spatial expression along in the ANP, MNP, and SC. D. TAZ-active embryos exhibited considerable variation in patterns of sox2 spatial expression. These ranged from patterns akin to that of embryos injected with TAZ-silent or uninjected embryos (ii) to very expanded expression of sox2 (i). Significant increases in spatial expression were observed only in the MNP. Embryos were imaged from a frontal (anterior) view. E. Quantification of changes in sox2 spatial expression in TAZ overexpressing embryos from B-D. ANPs marked by sox2 increased in diameter as compared to TAZ-silent embryos for both TAZ- WT and TAZ-active though not significantly. In contrast, the PNPs for embryos injected with either TAZ-WT or TAZ-active were significantly wider than controls. SC diameter was significantly increased in embryos injected with TAZ-WT, but not in embryos injected with TAZ-active. Quantification was performed independently three times; error bars show S.E.M. and unpaired Students TTEST was performed to calculate significance. TAZ-silent n = 15, TAZ-WT n = 8, TAZ-active n = 10. Scale bar is 500 µm. *p≤0.05.

160

Figure 5.7 Expression of n-tubulin cells in the neural plate in TAZ variant embryos.

Figure 5.7 N-tubulin marked cells following targeted expression of TAZ. A-C. In situ hybridizations against n-tubulin in NF stage 15/16 TAZ injected embryos. Nuclei expressing n-tubulin in the neural plate were observed in embryos overexpressing TAZ- silent (A), TAZ-WT (B), or TAZ-active (C). Insignificant differences in numbers of n- tubulin nuclei were observed between TAZ-silent embryos and either TAZ-WT or TAZ- active. Embryos were imaged from the dorsal aspect. D. Quantification of n-tubulin- positive nuclei from A-C. No significant difference was observed between TAZ-silent embryos and TAZ-WT or TAZ-active embryos. Embryos injected with TAZ-active displayed a greater variation in the of numbers of n-tubulin nuclei as compared to those overexpressing either TAZ-WT or TAZ-silent. Quantification was performed independently twice; unpaired Students TTEST was performed to calculate significance. TAZ-silent n = 6, TAZ-WT n = 7, TAZ-active n = 8. Scale bar is 500 µm.

161

5.5 Impact of Overexpression of Human TAZ on Gliogenesis

Preliminary studies in Dr. McCrea’s lab suggested that gliogenesis was affected by TAZ expression (Section 5.2). Expression of neurofilament medium (nefm), an intermediate filament protein expressed in glial cells, was significantly reduced. When we asked whether expression of glia-associated transcription factors was altered in embryos expressing TAZ variants homogeneously (Section 5.4), we detected no obvious differences (Figure 5.4). However, injections that targeted expression to the anterior neural plate of TAZ variants increased expression of early neural transcription factors

(Figure 5.5). Since targeted expression pinpointed small, though potentially important, changes in these transcription factors, we asked whether glial genes were also affected in this manner by carrying out a series of quantitative RT-PCR experiments. The experiments were set up identically to those described in Section 5.4. Genes of interest were selected from among those previously shown to regulate mammalian glial development (Figure 5.3B).

Of the nine glial-associated genes we measured, only two showed significant changes in expression, and only in embryos overexpressing TAZ-WT (Figure 5.8).

However, the genes that were responsive to TAZ overexpression were the transcription factors sox10 and nfix. Their orthologs in mice are important for oligodendrocyte and astrocyte specification, respectively, and their downregulation would potentially reflect a glial progenitor population remaining undifferentiated for longer periods of time. Genes expressed in glial progenitors (and eventually in differentiated astrocyte) such as nestin, glast, and fabp7 showed no change in expression in either TAZ-WT or TAZ-active

162

embryos, neither confirming nor denying the hypothesis outlined in the previous sentence. Genes associated with glial differentiation (glt-1 for astrocytes and mbp for oligodendrocytes) were also insignificantly changed in either TAZ WT or active embryos, though neither of these genes is significantly expressed at NF stage 18.

These results suggest that TAZ activity may alter or delay glial differentiation; however, a rigorous test of this hypothesis would require more experimentation.

163

Figure 5.8 TAZ regulation of glial differentiation.

Figure 5.8 Effects of TAZ overexpression in targeted injections on glial specification and differentiation. Quantitative RT-PCR experiment on NF stage 18 whole embryos specifically overexpressing TAZ-WT, TAZ-active, or TAZ-silent in the presumptive anterior nervous system. The transcription factors nfix and sox10 exhibited significantly decreased expression following injection of TAZ-WT RNA. However, this wasn’t followed by a subsequent decrease in expression of genes associated with glial differentiation (nestin, glast, fabp7, glt-1, mbp). Expression of TAZ-active had no discernible effect on transcription factors associated with gliogenesis (sox9, nfix, sox10) or glial differentiation (nestin, glast, fabp7, glt-1, mbp) as compared to TAZ-silent embryos. Samples were compared to their sibling embryos overexpressing TAZ-silent. Bars represent S.E.M. Unpaired student’s TTEST was performed to calculate significance. N=4 *p≤0.05

164

5.6 Effects of Overexpression of Human TAZ on Neural Crest Differentiation and Mesenchymal Gene Expression. The preliminary results from Dr. McCrea’s lab suggested that TAZ overexpression altered neural crest differentiation when TAZ variants were injected at the two-cell stage (Section 5.2). To determine whether this reflected TAZ activity in the neural ectoderm or in other tissues such as the dorsal mesoderm, we performed targeted injections of TAZ variants as described previously and evaluated expression of the neural crest transcription factor twist during mid-neurulation. Of the two dorsal blastomeres injected at 8-cell stage, one makes a major contribution to the early neural crest (Moody,

1987).

Expression of twist for both TAZ-WT and TAZ-active embryos was unchanged from that of controls both spatially and in intensity. This was surprising for TAZ-active, as preliminary two cell injection results suggested that expression of twist was significantly reduced (Figure 5.4). This could indicate that either TAZ has no effect on twist expression, or that the previously shown alteration in twist expression arises via the effect of TAZ on mesodermal gene expression instead of cell autonomous effects within the neural ectoderm.

8-cell injections also showed no difference in expression levels of mesenchyme- associated genes at NF stage 18, consistent with findings from embryos injected at the 2- cell stage in 2-cell injections (Figure 5.4).

165

Figure 5.9 Effects of TAZ expression on mesenchymal and neural crest genes.

Figure 5.9 Effects of TAZ overexpression in targeted injections on mesenchymal gene expression and an early neural crest marker (twist). Quantitative RT-PCR experiment on NF stage 18 whole embryos specifically overexpressing TAZ-WT, TAZ-active, or TAZ- silent in the presumptive anterior nervous system. Targeted expression of TAZ variants had no effect on expression of mesenchymal (tgfb1, cd44, ccn1, ccn2) associated genes or mes-ectoderm genes (twist). Samples were compared to their silent TAZ overexpressing siblings. Bars represent S.E.M. Unpaired student’s TTEST was performed to calculate significance. N=4 *p≤0.05.

5.10 Expression of the neural crest marker twist in TAZ variant embryos.

Figure 5.10 Expression of twist following targeted expression of TAZ. A-C. In situ hybridizations against twist in NF stage 15/16 TAZ injected embryos. Expression in the presumptive neural crest appears unchanged between TAZ variants. Embryos were imaged from the dorsal view. Scale bar is 500 µm. 166

5.7 Summary Overall, our results suggest that targeted expression of TAZ-WT in the anterior neural plate could be promoting an undifferentiated, actively proliferating neural progenitor population through increased expression of geminin and sox2 (Figure 5.5, 5.6).

Aberrant expression of TAZ-WT is also linked to decreased levels of two transcription factors important for glial differentiation, nfix and sox10 (Figure 5.7). Though the data supports the hypothesis that overexpression of TAZ causes the neural plate to remain in a proliferative state, while potentially delaying glial differentiation, it appears that neuronal differentiation is unaffected (Figure 5.5, Figure 5.8). This is surprising for several reasons; preliminary research had suggested that TAZ might suppress neuronal differentiation (Section 5.2) in whole embryo TAZ-active injections, and previously published research had indicated that geminin could delay neuronal differentiation (Seo et al., 2005). Geminin is purported to fill this role by sequestering neuronal transcription factors neurod1 and ngn1 (Seo et al., 2005). Since downregulation of geminin is an important component in timing of neurogenesis, effects on neurogenesis might become apparent if embryos were allowed to develop further.

Unexpectedly, aberrant expression of TAZ variants did not induce changes in mesenchymal marker expression in either whole embryo or targeted injections (Figure

5.4, 5.9, 5.10). It is possible that in healthy neural tissue, the embryo compensates for the increased expression of TAZ, preventing aberrant mesenchymal gene expression. The unexpected mutations in the all three TAZ clones could also have affected the co-factor binding properties of the TAZ variants (Figure 5.2). If these mutations specifically interrupted the ability of TAZ to interact with TEAD, then it would not be able to initiate

167

mesenchymal differentiation. TAZ-WT, with mutations in potentially less sensitive regions of the protein, was shown to be most efficient at disrupting normal neural development. We were also surprised to find a sharp divergence between our results and those previously reported in Dr. McCrea’s lab. These differences could be due to their comparison of injected embryos to uninjected controls or to the fact that their injections affected the entire embryo. Of these, the first would account for more of the variation seen, as our two-cell injection reproduction produces the most dissimilar results. The differences seen could also be due to the acquisition of the unexpected mutations outlined in Figure 5.2, as it is unknown when these mutations arose, they could account for most of the differences seen.

We have found that aberrant expression of TAZ in the developing anterior nervous system leads to an increase in expression of early neural transcription factors and reduces expression of glial associated transcription factors. Interestingly, TAZ expression had no effect on expression of mesenchymal genes, identifying a divergence between the function of TAZ in healthy neural tissue as compared to glioma stem cells.

168

Chapter VI

Discussion

169

6.1 Timing of the Gliogenic Switch in X. laevis

The study of gliogenesis in Xenopus laevis is in its infancy; our first step was to establish the timing of the gliogenic switch. Primary neurogenesis occurs from NF stages

14-35, so I assumed that gliogenesis would be initiated within this 48-h developmental window.

The gliogenic switch occurs in Xenopus laevis between NF stages 16-18. We know this because glast, the gene that is expressed concurrently with the gliogenic switch in mice and chick (Deneen et al., 2006), initiates its expression at this time. This is slightly surprising, as primary neurogenesis is initiated two stages prior to this, at NF stage 14. In mouse and chick models, 48 and 24 h, respectively, separate the initiation of neurogenesis from the gliogenic switch, with another 24 h for both models before glial specification is initiated (Chaboub et al., 2016). In contrast, only about 2-4 h separate the initiation of neurogenesis and the gliogenic switch in X. laevis. Glial specification, as marked by the exponential increase of glast transcription and the accumulation of glial associated proteins, also occurs more rapidly in Xenopus. Glial specification initiates between NF stages 20-24, only three to six h following the initial gliogenic switch. This is most likely necessitated by the external development characteristic of amphibians, which requires them to have a functional nervous system considerably earlier than their mouse or chick counterparts.

We also consistently observed a characteristic spatial pattern of gliogenesis. In X. laevis, different areas of the developing central nervous system undergo the gliogenic switch at different timepoints. The initial expression of glast at NF stage 16 is observed at

170

the very anterior portion of the spinal cord. By NF stage 20-22 the gliogenic switch, as revealed by expression of glast, has traveled posteriorly down the spinal cord. It is only following NF stage 24 that expression of glast is seen anterior to the midbrain-hindbrain boundary, indicating that the gliogenic switch has occurred in the midbrain and forebrain.

Current studies addressing the gliogenic switch are based on experiments in the mammalian or avian spinal cord or in cell lines; they do not consider the possibility that the gliogenic switch is under spatial regulation along the anteroposterior axis. However, the timeframe for these studies (days between timepoints as compared to the shorter intervals used for our studies in X. laevis) are large enough for differences in spatial positioning to be missed. The gliogenic switch appears to follow neural tube closure in mouse and chick models, whereas in X. laevis the earliest gliogenic switch occurs during the closing neural tube stage. These differences led us to two, potentially related, possibilities: first, that anterior to posterior signaling may contribute either directly or indirectly to the timing of the gliogenic switch, and s, that the presumptive spinal cord, which has been shown to be primed to initiate neurogenesis (Metzis et al., 2018; Polevoy et al., 2019) is also primed to initiate gliogenesis.

Our initial observations of the spatial and temporal expression of sox9 and nfix introduced doubt about their roles in the gliogenic switch in X. laevis. The temporal expression of sox9 is consistent with a role in gliogenesis; however, its spatial expression appeared to be limited to areas corresponding to the neural crest. Since high levels of transcript are identified by in situ hybridization, low levels of sox9 might be getting expressed in the neural tube. Neither the temporal nor the spatial expression of nfix, the

171

Xenopus ortholog Nfia, is consistent with a role in the initiation of the gliogenic switch in the spinal cord. Both sox9 and nfix could have contributed to initiating the gliogenic switch in the fore and midbrain.

6.2 Competence of Ectodermal Explants

One of the strengths of Xenopus as a model organism is in the explants that can be produced from their embryos. To determine whether anteroposterior positioning influences the gliogenic switch, we employed two different types of explants: animal caps overexpressing noggin (“noggin animal caps”) and mid-gastrula neural plates. The first explant, animal caps ectopically expressing the bmp4 inhibitor noggin, adopts an extreme anterior neural fate (Lamb et al., 1993; Zimmerman et al., 1996). They do this as they have undergone the activation, but not transformation portion of Nieuwkoop’s model. In contrast, neural plates isolated at the midgastrula stage can develop into all regions of the developing nervous system, along with neural crestage In addition, noggin animal cap explants also let us ask whether the inactivation of bmp, which is considered sufficient for neural specification, is sufficient for gliogenesis.

We were surprised to find that, even though both explants were cultured until the stage at which gliogenesis would have started in all regions of the brain, only the neural plates were capable of initiating gliogenesis. In situ hybridizations against glast in neural plates showed robust expression; in noggin animal caps, however, expression was largely undetectable. Noggin animal caps were very poor at initiating expression of astroglial- associated genes sox10, olig3, glast, and glt-1, whereas these genes were expressed at levels comparable to those of whole embryo controls in explanted neural plates. To

172

confirm that continued inhibition of bmp4 by noggin was not inhibiting gliogenesis, we isolated neural plates from embryos overexpressing noggin and found that they were still capable of initiating gliogenesis. This finding demonstrates that while inhibition of bmp is sufficient to initiate neurogenesis, it is insufficient to initiate gliogenesis. This also reinforced the possibility that anteroposterior positioning has a role in the gliogenic switch. The anterior-patterned noggin animal caps never initiated gliogenesis, whereas the neural plates, which reflect the full range of anteroposterior pattern were able to do so. This suggests that anteroposterior patterning may play an important role in the initiation of gliogenesis

The s interesting observation these studies identified is that while sox9 and nfix are sufficient to initiate the gliogenic switch in mouse and chick spinal cord, they are not sufficient to initiate gliogenesis in noggin animal caps. We initially observed that expression levels of sox9 and nfix were unchanged in noggin animal caps as compared to whole embryos. When we ectopically overexpressed sox9 and nfix, either alone or together in noggin animal caps, no combination of transcription factors was able to rescue expression of astroglial genes. This is significantly different from the published work on gliogenesis in mouse and chick embryos and is likely an indication of an evolutionary divergence between primary neurogenesis in frogs, and neurogenesis in higher vertebrates. In mammals, different regions of the brain have specialized astrocytes

(Tabata, 2015). This specialization could lead to different methods of glial specification and differentiation, that are not required by species with less specialized brains. It does suggest that nfix, unlike Nfia in mice and cell lines (Glasgow et al., 2017; Tchieu et al.,

2019; Tiwari et al., 2018), is unable to cause DNA methylation or chromatin remodeling 173

required for the transcription of astroglial-associated genes. Taken together, these findings do not support sox9 and nfix being sufficient to initiate Xenopus gliogenesis.

6.3 Roles of Signaling in the Gliogenic Switch

Our initial explant studies allowed us to establish an experimental system to test which factors are necessary and sufficient for the gliogenic switch to occur during X. laevis primary neurogenesis. One of the primary differences between noggin animal caps and neural plates is that noggin animal caps experience only the activation portion of neural development, whereas neural plates undergo both activation and transformation, through their initial burst of signals from the organizer before they are isolated. Given the competence of neural plates in making astroglia, it appeared likely that patterning signals played a role in priming the neural plate to initiate gliogenesis.

We carried out transcriptome profiling to identify genes that were differentially expressed in the neural plates and noggin animal caps during both gastrulation and in mid-neurulation and investigated their effects further. When we performed the GO analysis of the differentially expressed genes, we were struck by the number of genes related to anteroposterior patterning and were able to identify genes associated with the

Wnt signaling, FGF signaling, and RA signaling pathways. This confirms previously published works that indicate that noggin animal caps initiate anterior neurogenesis. We also identified differentially expressed transcription factors that have previously been shown to function in astroglial development in neural plates such as pou3f2, Sox9, sox10, sox8, olig3, and rfx4 (Cahoy et al., 2008; Laug et al., 2018; Stolt et al., 2003). Curiously,

174

we did not observe differential expression of nfix or any nuclear factor one family members in either explant.

Previous studies have shown that astroglial genes are methylated, and we and others anticipate that the gliogenic switch requires a reversal of DNA methylation at these loci. Therefore, the first gene we investigated was the potential pioneer transcription factor, ascl2, which was differentially expressed during gastrulation in the neural plates. Ectopic expression of ascl2 was unable to rescue initiation of gliogenesis in noggin animal caps, which led us to conclude that the phosphorylated ascl2 is unlikely to play a role in the gliogenic switch.

We investigated the roles Wnt, FGF, and RA signaling and confirmed the role of

Notch signaling on the gliogenic switch using small molecules to inhibit their function in neural plates. Similar to observations made in higher vertebrates, inhibition of notch signaling ablated the initiation of gliogenesis in Xenopus laevis. Our data also suggested that inhibition of Wnt signaling leads to delays in oligodendrocyte differentiation, which is also consistent with its role in gliogenesis in higher vertebrates. The most unexpected findings were that inhibition of FGF and RA signaling significantly reduced expression of astrocyte associated genes such as sox9, fabp7, glast, and glt-1. Reduction in these genes could decrease astrocyte differentiation, though how they may play this role is unclear. While no paracrine signaling pathway appeared to contribute directly to the gliogenic switch in X. laevis, our studies identified previously undetected roles for FGF and potentially retinoic acid signaling in astrocyte specification.

175

Figure 6.1 Timing and contributing factors in Xenopus astroglial development

Figure 6.1 Summary of the major findings affecting astroglial development in X. aevis during primary neurogenesis. A. As early as NF stage 16, glast expression (in red) is observed at the anterior portion of the presumptive spinal cord. This expression increases as the gliogenic switch occurs over a broader range, moving posteriorly down the spinal cord. It is only after NF stage 24 that the gliogenic switch is initiated anterior to the midbrain-hindbrain boundary. B. Alternate transcription factor network for gliogenesis in X. laevis. In the developing neural tube, neural progenitor populations marked by activated notch signaling experience increases in pou3f2 and sox9. Expression of these transcription factors either initiate an unknown transcription factor(s), or initiate transcription of early astroglial genes directly. It appears that transcription factors facilitating differentiation of oligodendrocytes (sox10, olig3) or astrocytes (rfx4) are conserved across tetrapods. We also found that FGF and RA signaling may play roles in astrocyte specification and differentiation and confirmed the role that Wnt signaling plays in oligodendrocyte differentiation.

176

6.4 Effects of Aberrant Expression of TAZ on Glial Development In glioblastoma, TAZ is a master regulator pushing glioma stem cell lines toward a mesenchymal, highly aggressive phenotype (Bhat et al., 2011). Glioblastomal tumors with this signature lose expression of both neuronal and astroglial associated markers such as Neurod2, Sox8, Sox10, and Lhx2. We sought to determine whether ectopic expression of human TAZ could replicate this in the developing X. laevis neural tube.

We found that overexpression of TAZ significantly increased expression of transcription factors sox2 and geminin, both of which are associated with a proliferative, non-differentiated phenotype. TAZ may also be indirectly delaying differentiation through geminin, which is known to antagonizes differentiation (Seo et al., 2005a). However, no other early neural transcription factor or any neuronal differentiation factors were affected, indicating that in healthy tissue aberrant TAZ expression has only a limited effect on neural development.

We also observed that ectopic expression of TAZ resulted in significantly lowered expression of sox10 and nfix. Lowered expression of sox10 is also observed in mesenchymal glioblastomas and would lead to decreased oligodendrocyte differentiation in healthy tissue. Significantly lowered nfix expression might indicate that TAZ is also able to decrease expression of Nfia in higher vertebrates. Decreased expression of Nfia would then lead to reduced astrocyte differentiation. Our findings are consistent with the role that TAZ plays in glioblastoma and would allow the tumor to remain in an actively replicating and growing state.

177

6.5 Future Directions and Research Implications

Our studies reflect an initial foray into gliogenesis in X. laevis and the results we found suggest further avenues of investigation. Potentially the most pressing experiment is determining which transcription factor, or combination of transcription factors, are required to initiate the gliogenic switch. As our data do not support a role for nfix, we need to determine whether pou3f2 or sox9 are required to initiate gliogenesis, or whether other known astroglial transcription factors such as rfx4 fulfill this role. We will accomplish this by ectopically expressing these genes in noggin animal caps or knocking them down in neural plates.

We will also pursue further research into the epigenetic regulation of the gliogenic switch. As we know from previous studies, the presumptive spinal cord is primed to initiate neurogenesis before it receives signals from the organizer during gastrulation

(Polevoy et al., 2019). It appears to be primed through an epigenetic rearrangement that then allows BMP signaling and FGF signaling to induce gliogenesis post-gastrulation

(Metzis et al., 2018; Polevoy et al., 2019). We know from several recent studies that part of the gliogenic switch is a change in the DNA methylation patterns that allows transcription of astroglial genes (Sanosaka et al., 2017; Takouda et al., 2017). We will isolate and sequence methylated DNA to identify the differences in DNA methylation patterns in noggin animal caps as compared to neural plates. This will allow us to identify genes that must be accessible for the gliogenic switch to occur.

The final thing I would like to pursue is whether spinal cord tissue is primed to initiate gliogenesis. Polevoy et al. (2019) induce noggin animal caps to express

178

spinal cord markers by co-injecting them with meis3 (Polevoy et al., 2019). It would be interesting to see if this artificially induced spinal cord intrinsically underwent the epigenetic rearrangements necessary to initiate gliogenesis. If they were able to make astroglial cells, it would also help identify which transcription factors were important for the gliogenic switch in X. laevis.

Glial development is currently grossly understudied in the amphibian research community. Our results provide a foundation for on which future research into the glial associated pathologies. We have initiated this by using X. laevis embryos as an in vivo model for the molecular basis of glioblastoma progression, we are also interested in applying this knowledge to traumatic brain injury studies. My dissertation expands the knowledge of gliogenesis in Xenopus laevis and establishes a reference for further exploration.

179

References Ali, F. R., Cheng, K., Kirwan, P., Metcalfe, S., Livesey, F. J., Barker, R. A., & Philpott, A. (2014). The phosphorylation status of Ascl1 is a key determinant of neuronal differentiation and maturation in vivo and in vitro. Development, 141(11), 2216– 2224. Andrews, S. (2010). FastQC: A quality control tool for high throughput sequence data. Anthony, T. E., Klein, C., Fishell, G., & Heintz, N. (2004). Radial glia serve as neuronal progenitors in all regions of the central nervous system. Neuron, 41(6), 881–890. Anthony, T. E., Mason, H. A., Gridley, T., Fishell, G., & Heintz, N. (2005). Brain lipid- binding protein is a direct target of Notch signaling in radial glial cells. Genes & Development, 19(9), 1028–1033. Araya, C., Tawk, M., Girdler, G. C., Costa, M., Carmona-Fontaine, C., & Clarke, J. D. (2014). Mesoderm is required for coordinated cell movements within zebrafish neural plate in vivo. Neural Development, 9(1), 9-9. Armit, C., Richardson, L., Hill, B., Yang, Y., & Baldock, R. A. (2015). eMouseAtlas informatics: embryo atlas and gene expression database. Mammalian Genome, 26(9– 10), 431–440. Balasubramaniyan, V., Vaillant, B., Wang, S., Gumin, J., Butalid. M.E., Sai, K., Mukheef, F., Kim, S.H., Boddeke, H.W., Lang, F., Aldape, K., Sulman, E.P., Bhat, K.P., & Colman, H. (2015). Aberrant mesenchymal differentiation of glioma stem- like cells: implications for therapeutic targeting. Oncotarget, 6(31), 31007–31017. Balmer, J. E., & Blomhoff, R. (2002). Gene expression regulation by retinoic acid. J Lipid Res, 43(11), 1773–1808. Barres, B. A. (2008). The mystery and magic of glia: a perspective on their roles in health and disease. Neuron, 60(3), 430–440. Bhat, K. P. L., Salazar, K. L., Balasubramaniyan, V., Wani, K., Heathcock, L., Hollingsworth, F., James J.D., Gumin, J., Diefes, K.L., Kim, S.H., Turski, A., Azodi, Y., Yang, Y., Doucette, T., Colman, H., Sulman, E.P., Lang, F.F., Rao, G., Copray, S., Vaillant, B.D., & Aldape, K.D. (2011). The transcriptional coactivator TAZ regulates mesenchymal differentiation in malignant glioma. Genes Dev, 25(24), 2594–2609. Bhatheja, K., & Field, J. (2006). Schwann cells: Origins and role in axonal maintenance and regeneration. Int J Biochem Cell Biol, 38(12), 1995–1999. Bi, W. L., & Beroukhim, R. (2014). Beating the odds: extreme long-term survival with glioblastoma. Neuro-Oncology, 16(9), 1159–1160.

180

Blighe, K. (2019). EnhancedVolcano: Publication-ready volcano plots with enhanced colouring and labeling. Retrieved from https://github.com/kevinblighe/EnhancedVolcano. Bonaguidi, M. A., McGuire, T., Hu, M., Kan, L., Samanta, J., & Kessler, J. A. (2005). LIF and BMP signaling generate separate and discrete types of GFAP-expressing cells. Development, 132(24), 5503–5514. Bosworth, A. P., & Allen, N. J. (2017). The diverse actions of astrocytes during synaptic development. Current Opinion in Neurobiology, 47, 38–43. Brivanlou, A. H., & Harland, R. M. (1989). Expression of an engrailed-related protein is induced in the anterior neural ectoderm of early Xenopus embryos. Development, 106(3), 611-617. Burda, J. E., Bernstein, A. M., & Sofroniew, M. V. (2016). Astrocyte roles in traumatic brain injury. Exp Neurol, 275(Pt 3), 305–315. Bushong, E. A., Martone, M. E., Jones, Y. Z., Ellisman, M. H., Zamanian, J. L., Christopherson, K. S., Xing, Y., Lubischer, J. L., Krieg, P. A., Krupenko, S. A., Thompson, W. J., & Barres, B. A. (2002). Protoplasmic astrocytes in CA1 stratum radiatum occupy separate anatomical domains. J Neurosci, 22(1), 183–192.

Cahoy, J. D., Emery, B., Kaushal, A., Foo, L. C., Zamanian, J. L., Christopherson, K. S., Xing, Y., Lubischer, J. L., Krieg, P. A., Krupenko, S. A., Thompson, W. J., & Barres, B. A. (2008). A transcriptome database for astrocytes, neurons, and oligodendrocytes: a new resource for understanding brain development and function. J Neurosci, 28(1), 264-278.

Caiazzo, M., Giannelli, S., Valente, P., Lignani, G., Carissimo, A., Sessa, A., Colasante, G., Bartolomeo, R., Massimino, L., Ferroni, S., Settembre, C., Benfenati, F., & Broccoli, V. (2015). Direct conversion of fibroblasts into functional astrocytes by defined transcription factors. Stem Cell Reports, 4(1), 25-36.

Cancer Genome Atlas Research Network, J. N., Weinstein, J. N., Collisson, E. A., Mills, G. B., Shaw, K. R. M., Ozenberger, B. A., Ellrott, K., Shmulevich, I., Sander, C., & Stuart, J. M. (2013). The cancer genome atlas pan-cancer analysis project. Nature Genetics, 45(10), 1113–1120.

Chaboub, L. S., & Deneen, B. (2013). Astrocyte form and function in the developing central nervous system. Seminars in Pediatric Neurology, 20(4), 230–235. Chaboub, L. S., Manalo, J. M., Lee, H. K., Glasgow, S. M., Chen, F., Kawasaki, Y., Akiyama, T., Kuo, C.T., Creighton, C. J., Mohila, C. A., & Deneen, B. (2016). Temporal profiling of astrocyte precursors reveals parallel roles for Asef during development and after injury. J Neurosci, 36(47), 11904–11917.

181

Chamberlain, K. A., Nanescu, S. E., Psachoulia, K., & Huang, J. K. (2016). Oligodendrocyte regeneration: Its significance in myelin replacement and neuroprotection in multiple sclerosis. Neuropharmacology, 110(Pt B), 633–643. Chaturvedi, P., Chetal, K., Ponny, S. R., & Pabla, S. (2018). CSBB-V2.1. Retrieved from https://sourceforge.net/projects/csbb-v2-1/ Chen, Y., Mccarthy, D., Ritchie, M., Robinson, M., & Smyth, G. (2008). edgeR: differential expression analysis of digital gene expression data. Cheung, M., & Briscoe, J. (2003). Neural crest development is regulated by the transcription factor Sox9. Development, 130(23), 5681–5693. Claes, A., Idema, A. J., & Wesseling, P. (2007). Diffuse glioma growth: a guerilla war. Acta Neuropathologica, 114(5), 443-458. Cohen, A. L., Holmen, S. L., & Colman, H. (2013). IDH1 and IDH2 mutations in gliomas. Current Neurology and Neuroscience Reports, 13(5), 345-345. Cragnolini, A. B., Montenegro, G., Friedman, W. J., & Mascó, D. H. (2018). Brain- region specific responses of astrocytes to an in vitro injury and neurotrophins. Molecular and Cellular Neuroscience, 88, 240–248. Dahlstrand, J., Zimmerman, L. B., McKay, R. D., & Lendahl, U. (1992). Characterization of the human nestin gene reveals a close evolutionary relationship to neurofilaments. J Cell Sci, 103(Pt 2), 589–597. Dai, Z.-M., Sun, S., Wang, C., Huang, H., Hu, X., Zhang, Z., Lu, Q. R., & Qiu, M. (2014). Stage-specific regulation of oligodendrocyte development by Wnt/β-catenin signaling. J Neurosci, 34(25), 8467–8473. Davidson, L. A., & Keller, R. E. (1999). Neural tube closure in Xenopus laevis involves medial migration, directed protrusive activity, cell intercalation and convergent extension. Development, 126(20), 4547-4556. De Rosa, A., Pellegatta, S., Rossi, M., Tunici, P., Magnoni, L., Speranza, M. C., Malusa, F., Miragliotta, V., Mori, E., Finocchiaro, G., & Bakker, A. (2012). A radial glia gene marker, fatty acid binding protein 7 (FABP7), is involved in proliferation and invasion of glioblastoma cells. PLoS ONE, 7(12), e52113.

DeCarolis, N. A., Mechanic, M., Petrik, D., Carlton, A., Ables, J. L., Malhotra, S., Bachoo, R., Götz, M., Lagace, D. C., & Eisch, A. J. (2013). In vivo contribution of nestin- and GLAST-lineage cells to adult hippocampal neurogenesis. Hippocampus, 23(8), 708–719.

182

Deneen, B., Ho, R., Lukaszewicz, A., Hochstim, C. J., Gronostajski, R. M., & Anderson, D. J. (2006). The transcription factor NFIA controls the onset of gliogenesis in the developing spinal cord. Neuron, 52(6), 953–968. Di Tommaso, P., Moretti, S., Xenarios, I., Orobitg, M., Montanyola, A., Chang, J.-M., Taly, J.-F., & Notredame, C. (2011). T-Coffee: a web server for the multiple sequence alignment of protein and RNA sequences using structural information and homology extension. Nucleic Acids Research, 39(Web Server issue), W13-7. Dobrokhotov, O., Samsonov, M., Sokabe, M., & Hirata, H. (2018). Mechanoregulation and pathology of YAP/TAZ via Hippo and non-Hippo mechanisms. Clinical and Translational Medicine, 7(1), 23-23. Dulamea, A. O. (2017). Role of oligodendrocyte dysfunction in demyelination, remyelination and neurodegeneration in multiple sclerosis. Adv Exp Med Biol, 958, 91–127. Durston, A. J. (2019). What are the roles of retinoids, other morphogens, and Hox genes in setting up the vertebrate body axis?. Genesis, e23296. Esain, V., Postlethwait, J. H., Charnay, P., & Ghislain, J. (2010). FGF-receptor signalling controls neural cell diversity in the zebrafish hindbrain by regulating olig2 and sox9. Development, 137(1), 33–42. Faigle, R., Liu, L., Cundiff, P., Funa, K., & Xia, Z. (2008). Opposing effects of retinoid signaling on astrogliogenesis in embryonic day 13 and 17 cortical progenitor cells. J Neurochem, 106(4), 1681–1698. Falk, S., & Götz, M. (2017). Glial control of neurogenesis. Curr Opin Neurobiol, 47, 188–195. Fletcher, R. B., & Harland, R. M. (2008). The role of FGF signaling in the establishment and maintenance of mesodermal gene expression in Xenopus. Dev Dyn, 237(5), 1243–1254. Frade, J. M., & Ovejero-Benito, M. C. (2015). Neuronal cell cycle: the neuron itself and its circumstances. Cell Cycle, 14(5), 712–720. Giachino, C., Boulay, J., Ivanek, R., Alvarado, A., Tostado, C., & Lugert, S. (2016). Notch signaling exhibits a tumor suppressor function in gliomas. Cancer Discovery, 6(2), 925-930. Glasgow, S. M., Carlson, J., Zhu, W., Chaboub, L. S., Kang, P., Lee, H. K., Clovis, Y., Lozzi, B., McEvilly, R. J., Rosenfeld, M. G., Creighton, C., Lee, S.-K., Mohila, C., Deneen, B. (2017). Glia-specific enhancers and chromatin structure regulate NFIA expression and glioma tumorigenesis. Nat Neurosci, 20(11), 1520-1528.

183

Glasgow, S. M., Zhu, W., Stolt, C. C., Huang, T.-W., Chen, F., LoTurco, J. J., Neul, J. L., Wegner, M., Mohila, C., Deneen, B. (2014). Mutual antagonism between Sox10 and NFIA regulates diversification of glial lineages and glioma subtypes. Nat Neurosci, 17(10), 1322–1329.

Goldman, S. A., & Kuypers, N. J. (2015). How to make an oligodendrocyte. Development, 142(23), 3983–3995. Götz, M., Sirko, S., Beckers, J., & Irmler, M. (2015). Reactive astrocytes as neural stem or progenitor cells: In vivo lineage, In vitro potential, and Genome-wide expression analysis. Glia, 63(8), 1452–1468. Grandbarbe, L., Bouissac, J., Rand, M., Hrabé de Angelis, M., Artavanis-Tsakonas, S., & Mohier, E. (2003). Delta-Notch signaling controls the generation of neurons/glia from neural stem cells in a stepwise process. Development, 130(7), 1391–1402. Green, D., Whitener, A. E., Mohanty, S., & Lekven, A. C. (2015). Vertebrate nervous system posteriorization: Grading the function of Wnt signaling. Dev Dynam, 244(3), 507–512. Gross, R. E., Mehler, M. F., Mabie, P. C., Zang, Z., Santschi, L., & Kessler, J. A. (1996). Bone morphogenetic proteins promote astroglial lineage commitment by mammalian subventricular zone progenitor cells. Neuron, 17(4), 595–606. Guo, F., Lang, J., Sohn, J., Hammond, E., Chang, M., & Pleasure, D. (2015). Canonical Wnt signaling in the oligodendroglial lineage-puzzles remain. Glia, 63(10), 1671– 1693. Hai, L., Zhang, C., Li, T., Zhou, X., Liu, B., Li, S., Zhu, M., Lin, Y., Yu, S., Zhang, K., Ren, B., Ming, H., Huang, Y., Chen, L., Zhao, P., Zhou, H., Jiang, T., & Yang, X. (2018). Notch1 is a prognostic factor that is distinctly activated in the classical and proneural subtype of glioblastoma and that promotes glioma cell survival via the NF- κB(p65) pathway. Cell Death & Disease, 9(2), 158-158.

Hamburger, V., & Hamilton, H. L. (1992). A series of normal stages in the development of the chick embryo. Dev Dynam, 195(4), 231–272. Hatakeyama, J., Bessho, Y., Katoh, K., Ookawara, S., Fujioka, M., Guillemot, F., & Kageyama, R. (2004). Hes genes regulate size, shape and histogenesis of the nervous system by control of the timing of neural stem cell differentiation. Development, 131(22), 5539–5550. Hellsten, U., Harland, R. M., Gilchrist, M. J., Hendrix, D., Jurka, J., Kapitonov, V., Ovcharenko, I., Putnam, N. H., Shu, S., Taher, L., Blitz, I. L., Blumberg, B., Dichmann, D. S., Dubchak, I., Amaya, E., Detter, J. C., Fletcher, R., Gerhard, D. S., Goodstein, D., Graves, T., Grigoriev, I. V., Grimwood, J., Kawashima, T., Lindquist,

184

E., Lucas, S. M., Mead, P. E., Mitros, T., Ogino, H., Ohta, Y., Poliakov, A. V., Pollet, N., Robert, J., Salamov, A., Sater, A. K., Schmutz, J., Terry, A., Vize, P. D., Warren, W. C., Wells, D., Wills, A., Wilson, R. K., Zimmerman, L. B., Zorn, A. M., Grainger, R., Grammer, T., Khokha, M. K., Richardson, P. M., & Rokhsar, D. S. (2010). The genome of the western clawed frog Xenopus tropicalis. Science, 328(5978), 633–636. Henningfeld, K. A., Locker, M., & Perron, M. (2007). Xenopus primary neurogenesis and retinogenesis. Functional development and embryology, 1(1), 26-34. Henrique, D., Abranches, E., Verrier, L., & Storey, K. G. (2015). Neuromesodermal progenitors and the making of the spinal cord. Development, 142(17), 2864–2875. Hockfield, S., & McKay, R. D. (1985). Identification of major cell classes in the developing mammalian nervous system. J Neurosci, 5(12), 3310–3328. Hojo, M., Ohtsuka, T., Hashimoto, N., Gradwohl, G., Guillemot, F., & Kageyama, R. (2000). Glial cell fate specification modulated by the bHLH gene Hes5 in mouse retina. Development, 127(12), 2515–2522. Holder, N., & Hill, J. (1991). Retinoic acid modifies development of the midbrain- hindbrain border and affects cranial ganglion formation in zebrafish embryos. Development, 113(4), 1159-1170. Howard, B. M., Zhicheng Mo, Z., Filipovic, R., Moore, A. R., Antic, S. D., & Zecevic, N. (2008). Radial glia cells in the developing human brain. Neuroscientist, 14(5), 459–473. Hu, J.-G., Lü, H.-Z., Wang, Y.-X., Bao, M.-S., Zhao, B.-M., & Zhou, J.-S. (2010). BMP signaling mediates astrocyte differentiation of oligodendrocyte progenitor cells. Tohoku J Exp Med, 222(3), 195–200. Huang, D. W., Sherman, B. T., & Lempicki, R. A. (2009). Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res, 37(1), 1–13. Huang, D. W., Sherman, B. T., & Lempicki, R. A. (2009). Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Prot, 4(1), 44–57. Jacobson, M. (1991). Developmental neurobiology. Plenum Press, 3, 1-40. Jiménez, A. J., Domínguez-Pinos, M.-D., Guerra, M. M., Fernández-Llebrez, P., & Pérez-Fígares, J.-M. (2014). Structure and function of the ependymal barrier and diseases associated with ependyma disruption. Tissue Barriers, 2, e28426. John Lin, C.-C., Yu, K., Hatcher, A., Huang, T.-W., Lee, H. K., Carlson, J., Weston, M. C., Chen, F., Zhang, Y., Zhu, W., Mohila, C. A., Ahmed, N., Patel, A. J., Arenkiel,

185

B. R., Noebels, J. L., Creighton, C. J., & Deneen, B. (2017). Identification of diverse astrocyte populations and their malignant analogs. Nat Neurosci, 20(3), 396–405. Kang, P., Lee, H. K., Glasgow, S. M., Finley, M., Donti, T., Gaber, Z. B., Graham, B. H., Foster, A. E., Novitch, B. G., Gronostajski, R. M., & Deneen, B. (2012). Sox9 and NFIA coordinate a transcriptional regulatory cascade during the initiation of gliogenesis. Neuron, 74(1), 79–94. Karve, I. P., Taylor, J. M., & Crack, P. J. (2016). The contribution of astrocytes and microglia to traumatic brain injury. British Journal of Pharmacology, 173(4), 692– 702. Kassambara, A., & Mundt, F. (2017). factoextra: extract and visualize the results of multivariate data analyses. Retrieved from https://cran.r- project.org/package=factoextra Kato, K., Losada-Perez, M., & Hidalgo, A. (2018). Gene network underlying the glial regenerative response to central nervous system injury. Dev Dynam, 247(1), 85–93. Keller, R., Shih, J., Sater, A. K., & Moreno, C. (1992). Planar induction of convergence and extension of the neural plate by the organizer of Xenopus. Dev Dynam, 193(3), 218–234. Kremer, M. C., Jung, C., Batelli, S., Rubin, G. M., & Gaul, U. (2017). The glia of the adult Drosophila nervous system. Glia, 65(4), 606–638. Kriegstein, A., & Alvarez-Buylla, A. (2009). The glial nature of embryonic and adult neural stem cells. Annual Review of Neuroscience, 32, 149–184. Lamb, T. M., & Harland, R. M. (1995). Fibroblast growth factor is a direct neural inducer, which combined with noggin generates anterior-posterior neural pattern. Development, 121(11), 3627–3636. Lamb, T. M., Knecht, A. K., Smith, W. C., Stachel, S. E., Economides, A. N., Stahl, N., Yancopolous, G. D., & Harland, R. M. (1993). Neural induction by the secreted polypeptide noggin. Science, 262(5134), 713–718. Laug, D., Glasgow, S. M., & Deneen, B. (2018). A glial blueprint for gliomagenesis. Nat Rev Neurosci, 19(7), 393-403. Lawson, A., Anderson, H., & Schoenwolf, G. C. (2001). Cellular mechanisms of neural fold formation and morphogenesis in the chick embryo. The Anatomical Record, 262(2), 153–168. Lee, H. K., Laug, D., Zhu, W., Patel, J. M., Ung, K., Arenkiel, B. R., Fancy, S. P. J., Mohila, C., & Deneen, B. (2015). Apcdd1 stimulates oligodendrocyte differentiation after white matter injury. Glia, 63(10), 1840–1849.

186

Lee, M. L., Martinez-Lozada, Z., Krizman, E. N., & Robinson, M. B. (2017). Brain endothelial cells induce astrocytic expression of the glutamate transporter GLT-1 by a Notch-dependent mechanism. J Neurochem, 143(5), 489–506. Lee, S. C. (2018). Diffuse gliomas for nonneuropathologists: the new integrated molecular diagnostics. Archives of Pathology & Laboratory Medicine, 142(7), 804– 814. Lei, L., Sonabend, A. M., Guarnieri, P., Soderquist, C., Ludwig, T., Rosenfeld, S., Bruce, J. N., & Canoll, P. (2011). Glioblastoma models reveal the connection between adult glial progenitors and the proneural phenotype. PloS One, 6(5), e20041. Li, B., & Dewey, C. N. (2011). RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics, 12, 323-323. Li, H., & Chen, G. (2016). In vivo reprogramming for CNS repair: Regenerating neurons from endogenous glial cells. Neuron, 91(4), 728–738. Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N., Marth, G., Abecasis, G., & Durbin, R. (2009). The Sequence Alignment/Map format and SAMtools. Bioinformatics, 25(16), 2078–2079.

Lim, D. A., Tramontin, A. D., Trevejo, J. M., Herrera, D. G., García-Verdugo, J. M., & Alvarez-Buylla, A. (2000). Noggin antagonizes BMP signaling to create a niche for adult neurogenesis. Neuron, 28(3), 713–726. Liour, S. S., & Yu, R. K. (2003). Differentiation of radial glia-like cells from embryonic stem cells. Glia, 42(2), 109–117. Love, M. I., Huber, W., & Anders, S. (2014). Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biology, 15(12), 550. Low, D. (2018). Recent advances in the understanding of microglial development and homeostasis. Cellular Immunology, 330, 68–78. Lowell, S. (2000). Notch signalling: You make me feel so glial. Current Biology, 10(16), R595–R597. Lyons, D. A., & Talbot, W. S. (2015). Glial cell development and function in zebrafish. Cold Spring Harbor Perspectives in Biology, 7(2), 1–21. Mabie, P. C., Mehler, M. F., Marmur, R., Papavasiliou, A., Song, Q., & Kessler, J. A. (1997). Bone morphogenetic proteins induce astroglial differentiation of oligodendroglial-astroglial progenitor cells. J Neurosci, 17(11), 4112–4120. Magnusson, J. P., & Frisén, J. (2016). Stars from the darkest night: unlocking the neurogenic potential of astrocytes in different brain regions. Development, 143(7), 1075–1086.

187

Magnusson, J. P., Goritz, C., Tatarishvili, J., Dias, D. O., Smith, E. M. K., Lindvall, O., Kokaia, Z., & Frisen, J. (2014). A latent neurogenic program in astrocytes regulated by Notch signaling in the mouse. Science, 346(6206), 237–241. Maher, E. A., Brennan, C., Wen, P. Y., Durso, L., Ligon, K. L., Richardson, A., Khatry, D., Feng, B., Sinha, R., Louis, D. N., Quackenbush, J., Black, P., Chin, L., & DePinho, R. A. (2006). Marked genomic differences characterize primary and secondary glioblastoma subtypes and identify two distinct molecular and clinical secondary glioblastoma entities. Cancer Research, 66(23), 11502–11513.

Mahony, S., Mazzoni, E. O., McCuine, S., Young, R. A., Wichterle, H., & Gifford, D. K. (2011). Ligand-dependent dynamics of retinoic acid receptor binding during early neurogenesis. Genome Biology, 12(1), R2-R2. Martin, M. (2011). Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet.Journal, 17(1), 10-10. Mehler, M. F., Mabie, P. C., Zhang, D., & Kessler, J. A. (1997). Bone morphogenetic proteins in the nervous system. Trends in Neurosciences, 20(7), 309–317. Metzis, V., Steinhauser, S., Lovell-Badge, R., Luscombe, N. M., Briscoe Correspondence, J., Pakanavicius, E., Gouti, M., Stamataki, D., Ivanovitch, K., Watson, T., Rayon, T., Neda, S., Gharavy, M., & Briscoe, J. (2018). Nervous system regionalization entails axial allocation before neural differentiation. Cell, 175(4), 1105-1118. Mi, H., Muruganujan, A., Ebert, D., Huang, X., & Thomas, P. D. (2019). PANTHER version 14: more genomes, a new PANTHER GO-slim and improvements in enrichment analysis tools. Nucleic Acids Research, 47(D1), D419–D426. Molofsky, A. V., & Deneen, B. (2015). Astrocyte development: a guide for the perplexed. Glia, 63(8), 1320–1329. Molofsky, A. V, Krencik, R., Krenick, R., Ullian, E. M., Ullian, E., Tsai, H., Deneen, B., Richardson, W. D., Barres, B. A., & Rowitch, D. H. (2012). Astrocytes and disease: a neurodevelopmental perspective. Genes & Development, 26(9), 891–907. Moody, S. A. (1987). Fates of the blastomeres of the 16-cell stage Xenopus embryo. Dev Bio, 119(2), 560–578. Myers, C. T., & Appleby, S. C. (2014). Use of small molecule inhibitors of the Wnt and Notch signaling pathways during Xenopus development. Methods, 66(3), 380–389. Nakashima, K., Yanagisawa, M., Arakawa, H., & Taga, T. (1999). Astrocyte differentiation mediated by LIF in cooperation with BMP2. FEBS Letters, 457(1), 43–46.

188

Namihira, M., Kohyama, J., Semi, K., Sanosaka, T., Deneen, B., Taga, T., & Nakashima, K. (2009). Committed neuronal precursors confer astrocytic potential on residual neural precursor cells. Dev Cell, 16(2), 245–255. Nieuwkoop, P. D. (1952). Activation and organization of the central nervous system in amphibians. part i. induction and activation. J Experimental Zoology, 120(1), 1–31. Nieuwkoop, P. D., & Nigtevecht, G. V. (1954). Neural activation and transformation in explants of competent ectoderm under the influence of fragments of anterior notochord in urodeles. Development, 2(3), 175-193. Nikos, P. (2012). Xenopus laevis as a model system. Materials and Methods, 2. Niu, W., Zang, T., Zou, Y., Fang, S., Smith, D. K., Bachoo, R., & Zhang, C.-L. (2013). In vivo reprogramming of astrocytes to neuroblasts in the adult brain. Nat Cell Bio, 15(10), 1164-1175. Noguchi, S., Saito, A., & Nagase, T. (2018). YAP/TAZ Signaling as a molecular link between fibrosis and cancer. Intl J Molecular Sciences, 19(11), e3674. Noushmehr, H., Weisenberger, D. J., Diefes, K., Phillips, H. S., Pujara, K., Berman, B. P., Pan, F., Pelloski, C. E., Sulman, E. P., Bhat, K. P., Verhaak, R. G.W., Hoadley, K. A., Hayes, D. N., Perou, C. M., Schmidt, H. K., Ding, L., Wilson, R. K., Van Den Berg, D., Shen, H., Bengtsson, H., Neuvial, P., Cope, L. M., Buckley, J., Herman, J. G., Baylin, S. B., Laird, P. W., Aldape, K., & Cancer Genome Atlas Research Network. (2010). Identification of a CpG island methylator phenotype that defines a distinct subgroup of glioma. Cancer Cell, 17(5), 510–522.

Oberheim, N. A., Takano, T., Han, X., He, W., Lin, J. H. C., Wang, F., Xu, Q., Wyatt, J. D., Pilcher, W., Ojemann, J. G., Ransom, B. R., Goldman, S. A., & Nedergaard, M. (2009). Uniquely hominid features of adult human astrocytes. J Neurosci, 29(10), 3276–3287. Ortega-Pérez, S., & Amaya-Rey, M. C. (2018). Secondary brain injury. J Neurosci Nursing, 50(4), 220–224. Owada, Y., Yoshimoto, T., & Kondo, H. (1996). Spatio-temporally differential expression of genes for three members of fatty acid binding proteins in developing and mature rat brains. J Chem Neuroanatomy, 12(2), 113–122. Park, H.-C., & Appel, B. (2003). Delta-Notch signaling regulates oligodendrocyte specification. Development, 130(16), 3747–3755. Parpura, V., & Verkhratsky, A. (2012). Astrocytes revisited: concise historic outlook on glutamate homeostasis and signaling. Croatian Medical Journal, 53(6), 518–528.

189

Parsons, D. W., Jones, S., Zhang, X., Lin, J. C.-H., Leary, R. J., Angenendt, P., Mankoo, P., Carter, H., Siu, I-M., Gallia, G. L., Olivi, A., McLendon, R., Rasheed, B. A., Keir, S., Nikolskaya, T., Nikolsky, Y., Busam, D. A., Tekleab, H., Diaz, L. A., Hartigan, J., Smith, D. R., Strausberg, R. L., Marie, S. K. N., Shinjo, S. M. O., Yan, H., Riggins, G. J., Bigner, D. D., Karchin, R., Papadopoulos, N., Parmigiani, G., Vogelstein, B., Velculescu, V. E., & Kinzler, K. W. (2008). An integrated genomic analysis of human glioblastoma multiforme. Science, 321(5897), 1807–1812.

Paschaki, M., Schneider, C., Rhinn, M., Thibault-Carpentier, C., Dembélé, D., Niederreither, K., & Dollé, P. (2013). Transcriptomic analysis of murine embryos lacking endogenous retinoic acid signaling. PloS One, 8(4), e62274. Patapoutian, A., & Reichardt, L. F. (2000). Roles of Wnt proteins in neural development and maintenance. Current Opinion in Neurobiology, 10(3), 392–399. Patten, B. A., Sardi, S. P., Koirala, S., Nakafuku, M., & Corfas, G. (2006). Notch1 signaling regulates radial glia differentiation through multiple transcriptional mechanisms. J. Neurosci, 23(14), 6132–6140. Patten, I., & Placzek, M. (2000). The role of Sonic hedgehog in neural tube patterning. Cellular and Molecular Life Sciences, 57(12), 1695–1708. Patthey, C., & Gunhaga, L. (2014). Signaling pathways regulating ectodermal cell fate choices. Experimental Cell Research, 321(1), 11–16. Perrimon, N., Pitsouli, C., & Shilo, B.-Z. (2012). Signaling mechanisms controlling cell fate and embryonic patterning. Cold Spring Harbor Perspectives in Biology, 4(8), a005975. Peshkin, L., Lukyanov, A., Kalocsay, M., Gage, R. M., Wang, D., Pells, T. J., Karimi, K., Vize, P. D., Wühr, M., & Kirschner, M. W. (2019). The protein repertoire in early vertebrate embryogenesis. BioRxiv, e571174. Pfleger, C. M. (2017). The Hippo Pathway. Curr Top Dev Biol, 123, 181–228. Phillips, H. S., Kharbanda, S., Chen, R., Forrest, W. F., Soriano, R. H., Wu, T. D., Misra, A., Nigro, J. M., Colman, H., Soroceanu, L., Williams, P. M., Modrusan, Z., Feuerstein, B. G., & Aldape, K. (2006). Molecular subclasses of high-grade glioma predict prognosis, delineate a pattern of disease progression, and resemble stages in neurogenesis. Cancer Cell, 9(3), 157–173. Piccolo, S., Sasai, Y., Lu, B., & EM., D. R. (1996). Dorsoventral patterning in Xenopus: inhibition of ventral signals by direct binding of chordin to BMP-4. Cell, 86(4), 589– 598. Pieper, M., Eagleson, G. W., Wosniok, W., & Schlosser, G. (2011). Origin and segregation of cranial placodes in Xenopus laevis. Dev Biol, 360(2), 257–275.

190

Polevoy, H., Gutkovich, Y. E., Michaelov, A., Volovik, Y., Elkouby, Y. M., & Frank, D. (2019). New roles for Wnt and BMP signaling in neural anteroposterior patterning. EMBO Reports, 20(6), e45842. Rankin, S. A., McCracken, K. W., Luedeke, D. M., Han, L., Wells, J. M., Shannon, J. M., & Zorn, A. M. (2018). Timing is everything: reiterative Wnt, BMP and RA signaling regulate developmental competence during endoderm organogenesis. Dev Bio, 434(1), 121–132. Reiprich, S., & Wegner, M. (2015). From CNS stem cells to neurons and glia: Sox for everyone. Cell and Tissue Research, 359(1), 111–124. Rhinn, M., Lun, K., Luz, M., Werner, M., & Brand, M. (2005). Positioning of the midbrain-hindbrain boundary organizer through global posteriorization of the neuroectoderm mediated by Wnt8 signaling. Development, 132(6), 1261–1272. Robinson, M. D., McCarthy, D. J., & Smyth, G. K. (2010). edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics, 26(1), 139–140. Rogers, C. D., Moody, S. A., & Casey, E. S. (2009). Neural induction and factors that stabilize a neural fate. Birth Defects Research. Part C, Embryo Today: Reviews, 87(3), 249–262. Rothstein, J. D., Dykes-Hoberg, M., Pardo, C. A., Bristol, L. A., Jin, L., Kuncl, R. W., Kanai, Y., Hediger, M. A., Wang, Y., Schielke, J. P., & Welty, D. F. (1996). Knockout of glutamate transporters reveals a major role for astroglial transport in excitotoxicity and clearance of glutamate. Neuron, 16(3), 675–686. Sanosaka, T., Imamura, T., Hamazaki, N., Chai, M. C., Igarashi, K., Ideta-Otsuka, M., Miura, F., Ito, T., Fujii, N., Ikeo, K., & Nakashima, K. (2017). DNA methylome analysis identifies transcription factor-based epigenomic signatures of multilineage competence in neural stem/progenitor cells. Cell Reports, 20(12), 2992–3003. Sater, A. K., Steinhardt, R. A., & Keller, R. (1993). Induction of neuronal differentiation by planar signals in Xenopus embryos. Devel Dynam, 197(4), 268–280. Savchenko, E., Teku, G. N., Boza-Serrano, A., Russ, K., Berns, M., Deierborg, T., Lamas, N. J., Wichterle, H., Rothstein, J., Henderson, C. E., Vihinen, M., & Roybon, L. (2019). FGF family members differentially regulate maturation and proliferation of stem cell-derived astrocytes. Scientific Reports, 9(1), 9610-9610. Schilling, T. F. (2008). Anterior-posterior patterning and segmentation of the vertebrate head. Integrative and Comparative Biology, 48(5), 658–667. Schurch, N. J., Schofield, P., Gierliński, M., Cole, C., Sherstnev, A., Singh, V., Wrobel, N., Gharbi, K., Simpson, G. G., Owen-Hughes, T., Blaxter, M., & Barton, G. J.

191

(2016). How many biological replicates are needed in an RNA-seq experiment and which differential expression tool should you use?. RNA, 22(6), 839–851. Seo, S., Herr, A., Lim, J.-W., Richardson, G. A., Richardson, H., & Kroll, K. L. (2005). Geminin regulates neuronal differentiation by antagonizing Brg1 activity. Genes & Dev, 19(14), 1723–1734. Session, A. M., Uno, Y., Kwon, T., Chapman, J. A., Toyoda, A., Takahashi, S., Fukui, A., Hikosaka, A., Suzuki, A., Kondo, M., van Heeringen, S. J., Quigley, I., Heinz, S., Ogino, H., Ochi, H., Hellsten, U., Lyons, J. B., Simakov, O., Putnam, N., Stites, J., Kuroki, Y., Tanaka, T., Michiue, T., Watanabe, M., Bogdanovic, O., Lister, R., Georgiou, G., Paranjpe, S. S., van Kruijsbergen, I., Shu, S., Carlson, J., Kinoshita, T., Ohta, Y., Mawaribuchi, S., Jenkins, J., Grimwood, J., Schmutz, J., Mitros, T., Mozaffari, S. V., Suzuki, Y., Haramoto, Y., Yamamoto, T. S., Takagi, C., Heald, R., Miller, K., Haudenschild, C., Kitzman, J., Nakayama, T., Izutsu, Y., Robert, J., Fortriede, J., Burns, K., Lotay, V., Karimi, K., Yasuoka, Y., Dichmann, D. S., Flajnik, M. F., Houston, D. W., Shendure, J., DuPasquier, L., Vize, P. D., Zorn, A. M., Ito, M., Marcotte, E. M., Wallingford, J. B., Ito, Y., Asashima, M., Ueno, N., Matsuda, Y., Veenstra, G. J. C., Fujiyama, A., Harland, R. M., Taira, M., & Rokhsar, D. S. (2016). Genome evolution in the allotetraploid frog Xenopus laevis. Nature, 538(7625), 336–343. Shandra, O., Winemiller, A. R., Heithoff, B. P., Munoz-Ballester, C., George, K. K., Benko, M. J., Zuidhoek, I. A., Besser, M. N., Curley, D. E., Edwards, G. F., Mey, A., Harrington, A. N., Kitchen, J. P., & Robel, S. (2019). Repetitive diffuse mild traumatic brain injury causes an atypical astrocyte response and spontaneous recurrent seizures. J Neurosci, 39(10), 1944–1963. Shibata, T., Yamada, K., Watanabe, M., Ikenaka, K., Wada, K., Tanaka, K., & Inoue, Y. (1997). Glutamate transporter GLAST is expressed in the radial glia-astrocyte lineage of developing mouse spinal cord. J Neurosci, 17(23), 9212–9219. Shimizu, T., Kagawa, T., Wada, T., Muroyama, Y., Takada, S., & Ikenaka, K. (2005). Wnt signaling controls the timing of oligodendrocyte development in the spinal cord. Dev Bio, 282(2), 397–410. Silver, J., & Miller, J. H. (2004). Regeneration beyond the glial scar. Nature Reviews Neuroscience, 5(2), 146–156. Sive, H. L., Grainger, R., & Harland, R. M. (2010). Early development of Xenopus laevis: a laboratory manual. Cold Spring Harbor Laboratory Press. Smith, W. C., & Harland, R. M. (1992). Expression cloning of noggin, a new dorsalizing factor localized to the Spemann organizer in Xenopus embryos. Cell, 70(5), 829–840.

192

Spassky, N., Merkle, F. T., Flames, N., Tramontin, A. D., García-Verdugo, J. M., & Alvarez-Buylla, A. (2005). Adult ependymal cells are postmitotic and are derived from radial glial cells during embryogenesis. J Neurosci, 25(1), 10–18. Srikanth, M., Kim, J., Das, S., & Kessler, J. A. (2014). BMP signaling induces astrocytic differentiation of clinically derived oligodendroglioma propagating cells. Mol Cancer Research, 12(2), 283–294. Starmer, J. (2017). StatQuest: PCA in R. Stockhausen, M.-T., Kristoffersen, K., & Poulsen, H. S. (2010). The functional role of Notch signaling in human gliomas. Neuro-Oncology, 12(2), 199–211. Stockhausen, M.-T., Kristoffersen, K., & Poulsen, H. S. (2012). Notch signaling and brain tumors. Adv Exp Med Biol, 727, 289–304. Stolt, C. C., Lommes, P., Sock, E., Chaboissier, M.-C., Schedl, A., & Wegner, M. (2003). The Sox9 transcription factor determines glial fate choice in the developing spinal cord. Genes & Development, 17(13), 1677–1689. Sun, S., Zhu, X., Huang, H., Guo, W., Tang, T., Xie, B., Xu, X., Zhang, Z., Shen, Y., Dai, Z.‐M., & Qiu, M. (2019). WNT signaling represses astrogliogenesis via Ngn2‐ dependent direct suppression of astrocyte gene expression. Glia, 67(7), 1333–1343. T Maniatis, E. F. F. and J. S. (1982). Molecular Cloning. A Laboratory Manual. Cold Spring Harbor Laboratory. Tabata, H. (2015). Diverse subtypes of astrocytes and their development during corticogenesis. Frontiers in Neuroscience, 9, 114-114. Takouda, J., Katada, S., & Nakashima, K. (2017). Emerging mechanisms underlying astrogenesis in the developing mammalian brain. Proceedings of the Japan Academy, Series B, 93(6), 386–398. Taylor, M. K., Yeager, K., & Morrison, S. J. (2007). Physiological Notch signaling promotes gliogenesis in the developing peripheral and central nervous systems. Development, 134(13), 2435–2447. Tchieu, J., Calder, E. L., Guttikonda, S. R., Gutzwiller, E. M., Aromolaran, K. A., Steinbeck, J. A., Goldstein, P. A., & Studer, L. (2019). NFIA is a gliogenic switch enabling rapid derivation of functional human astrocytes from pluripotent stem cells. Nature Biotechnology, 37(3), 267-275. Temburni, M. K., & Jacob, M. H. (2001). New functions for glia in the brain. Proceedings of the National Academy of Sciences of the United States of America, 98(7), 3631.

193

Thuret, R., Auger, H., & Papalopulu, N. (2015). Analysis of neural progenitors from embryogenesis to juvenile adult in Xenopus laevis reveals biphasic neurogenesis and continuous lengthening of the cell cycle. Biology Open, 4(12), 1772–1781. Tisler, M., Thumberger, T., Schneider, I., Schweickert, A., & Blum, M. (2017). Leftward flow determines laterality in conjoined twins. Current Biology, 27(4), 543–548. Tiwari, N., Pataskar, A., Pé, S., Ló Pez-Mascaraque, L., Tiwari, V. K., & Correspondence, B. B. (2018). Stage-specific transcription factors drive astrogliogenesis by remodeling gene regulatory landscapes. Cell Stem Cell, 23, 557– 571. Verhaak, R. G. W., Hoadley, K. A., Purdom, E., Wang, V., Qi, Y., Wilkerson, M. D., Miller, C. R., Ding, L., Golub, T., Mesirov, J. P., Alexe, G., Lawrence, M., O'Kelly, M., Tamayo, P., Weir, B. A., Gabriel, S., Winckler, W., Gupta, S., Jakkula, L., Feiler, H. S., Hodgson, J. G., James, C. D., Sarkaria, J. N., Brennan, C., Kahn, A., Spellman, P. T., Wilson, R. K., Speed, T. P., Gray, J. W., Meyerson, M., Getz, G., Perou, C. M., Hayes, D. N., Cancer Genome Atlas Research Network, & T. C. G. A. R. (2010). Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell, 17(1), 98–110.

Verkhratsky, A. (2019). Astroglial calcium signaling in aging and alzheimer’s disease. Cold Spring Harb Perspect Biol. 11(7), a035188. Wang, J., Liu, S., Heallen, T., & Martin, J. F. (2018). The Hippo pathway in the heart: pivotal roles in development, disease, and regeneration. Nature Reviews Cardiology, 15(11), 672–684. Warnes, G. R., Bolker, B., Bonebakker, L., Gentleman, R., Huber, W., Liaw, A., Lumley, T., Maechler, M., Magnusson, A., Moeller, S., Schwartz, M., & Venables, B. (2019). gplots: Various R programming tools for plotting data. R package version 3.0.1.1.

Weider, M., & Wegner, M. (2017). SoxE factors: Transcriptional regulators of neural differentiation and nervous system development. Sem Cell & Dev Bio, 63, 35–42. Wickham, H. (2016). ggplot2: Elegant graphics for data analysis. Springer-Verlag. Woo, K., & Fraser, S. E. (1997). Specification of the zebrafish nervous system by nonaxial signals. Science, 277(5323), 254–257. Wullimann, M. F., Rink, E., Vernier, P., & Schlosser, G. (2005). Secondary neurogenesis in the brain of the African clawed frog, Xenopus laevis, as revealed by PCNA, Delta- 1, Neurogenin-related-1, and NeuroD expression. J Comparative Neurology, 489(3), 387–402.

194

Xu, J., Wang, T., Wu, Y., Jin, W., & Wen, Z. (2016). Microglia colonization of developing zebrafish midbrain is promoted by apoptotic neuron and lysophosphatidylcholine. Dev Cell, 38(2), 214–222. Yoon, K., Nery, S., Rutlin, M. L., Radtke, F., Fishell, G., & Gaiano, N. (2004). Fibroblast growth factor receptor signaling promotes radial glial identity and interacts with Notch1 signaling in telencephalic progenitors. J Neurosci, 24(43), 9497–9506. Yoshida, M. (1997). Oligodendrocyte maturation in Xenopus laevis. J Neurosci Res, 50(2), 169–176. Zhang, H., Liu, C.-Y., Zha, Z.-Y., Zhao, B., Yao, J., Zhao, S., Xiong, Y., Lei, Q.-Y., & Guan, K.-L. (2009). TEAD transcription factors mediate the function of TAZ in cell growth and epithelial-mesenchymal transition. J Biol Chem, 284(20), 13355–13362. Zhu, A., Ibrahim, J. G., & Love, M. I. (2018). Heavy-tailed prior distributions for sequence count data: removing the noise and preserving large differences. Bioinformatics, 35(12):2084-2092. Zimmerman, L. B., De Jesús-Escobar, J. M., & Harland, R. M. (1996). The Spemann organizer signal noggin binds and inactivates bone morphogenetic protein 4. Cell, 86(4), 599–606. Zuchero, J. B., & Barres, B. A. (2015). Glia in mammalian development and disease. Development, 142(22), 3805–3809.

195