Slit/Robo signaling underlies the spatial patterning of spiral ganglion neurons to shape the peripheral auditory circuitry assembly

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

SLIT/ROBO SIGNALING UNDERLIES THE SPATIAL PATTERNING OF
SPIRAL GANGLION NEURONS TO SHAPE THE PERIPHERAL AUDITORY
CIRCUITRY ASSEMBLY

by
Shengzhi Wang

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PHYSIOLOGY AND BIOPHYSICS)

May 2013

Copyright 2013 Shengzhi Wang

i

Dedication

This work is dedicated to my parents, Mr. Chenghui Wang and Mrs. Shujian Li, for their
unconditioned love and support.

ii

Acknowledgements
I would like to express my deepest gratitude to people who helped me over the course of
my Ph.D training at University of Southern California.
First, I want to thank my mentors, Dr. Li I Zhang and Dr. Huizhong Whit Tao for the
invaluable guidance, generous support and exceptional patience they persistently offered
to me throughout my study and training in the lab. I feel deeply indebted to them for
their tremendous help both scientifically and personally. It is truly fortunate for me to
work with them during the past six and a half years and I really learned a lot from them.
This wonderful experience will be treasured forever.
I also want to thank my committee members, Dr. Jeannie Chen and Dr. Robert Chow, for
their thoughtful suggestions and tremendous help.
I want to thank all the current and former lab members for their generous help. Special
thanks go to Leena A. Ibrahim and Young J. Kim, who work together with me side by
side to solve the problems together and to share the joy together. They gave me so much
help that makes my life a lot easier.
I also want to thank Daniel A. Gibson, Haiwen C. Leung and Dr. Le Ma for their
generosity to provide technical and material support for my project.
Last but not least, I feel so blessed to have my wife, Jinghua Liu as well as other family
members always standing by me. It’s them that make my life full.

iii

Table of Contents
Dedication ..................................................................................................................................... i
Acknowledgements ................................................................................................................. ii
List of Figures ............................................................................................................................. v
List of Tables ............................................................................................................................ vii
Abstract .................................................................................................................................... viii
Chapter 1: Slit/Robo signaling underlies spatial patterning of cochlear
innervation ................................................................................................................................. 1
1.1 Introduction to the assembly of peripheral auditory circuitry ........................... 4
1.1.1 Cochlea hair cells: the “mechanical” device for sound perception ................. 8
1.1.2 Formation of spiral ganglion neurons from a developmental view ............. 14
1.1.3 Afferent innervation from spiral ganglion neurons to hair cells ................... 15
1.2 Transcriptome analysis to identify molecular candidates underlying the
spatial patterning of peripheral innervation ........................................................................ 21
1.2.1 Novel methods to purify cochlea hair cells and non-hair cells ....................... 23
1.2.2 Microarray study ...................................................................................................................... 26
1.2.3 Screening for guidance molecules enriched in non-hair cell in the cochlea
29
1.3 The role of Slit family molecules in shaping peripheral auditory circuitry
assembly ............................................................................................................................................. 30
1.3.1 Background of Slit and Robo family molecules ....................................................... 30
1.3.2 Expression of Slit molecules in cochlear tissue ....................................................... 32
1.3.3 Mis-positioning of spiral ganglion neurons in Slit2 mutant ............................. 34
1.3.4 SGNs are progressively relocated to the cochlear epithelium from their
normal position in the Slit2 mutant ............................................................................................... 37
1.4 The role of Robo family molecules in shaping peripheral auditory circuitry
assembly ............................................................................................................................................. 40

iv

1.4.1 Robo is expressed in SGNs as well as supporting cells in the organ of Corti
40
1.4.2 SGNs are mispositioned as individuals in the Robo1/Robo2 double mutant
at E18 41
1.4.3 Expansion of SGN boundary in the Robo1/Robo2 double mutant ................. 44
1.4.4 Developmental progression of SGN territory expansion in Robo mutants
48
1.5 Discussion and Perspectives ......................................................................................... 50
Slit-Robo interactions in spatial restriction of spiral ganglion neurons ................... 50
The relationship between SGN soma positioning and neurite path finding............ 52
The function of Slit1 and additional roles of Robo in the cochlea development .. 53
Intracellular mechanisms underlying the restriction of SGN by Slit/ Robo
signaling ........................................................................................................................................................ 54
1.6 Experimental Procedures and Materials .................................................................. 55
Mouse Strains ............................................................................................................................................. 55
Tissue dissection, FM1-43 staining and FACS sorting .......................................................... 55
RNA amplification and microarray data analysis ................................................................... 56
Immunohistochemistry and in situ hybridization ................................................................. 57
Chapter 2: Development of STARS technique for lineage tracing in the inner
ear ................................................................................................................................................ 58
2.1 Introduction to Single Cell Labeling in the nervous system .............................. 60
2.2 Proof of Principle and Development of STARS Technique ................................. 64
2.3 Generating STARS mouse ............................................................................................... 74
2.4 Perspectives and Future Direction ............................................................................. 80
2.5 Experimental Design and Methods ............................................................................. 83
Bibliography ............................................................................................................................ 87

v

List of Figures

Figure 1: The Anatomy of Inner Ear................................................................................... 7
Figure 2: Hair bundle is the mechanosensitive organelle of hair cells. .............................. 9
Figure 3: Development of Hair cells. ............................................................................... 13
Figure 4: Schematic illustration of the assembly of peripheral auditory circuitry. .......... 16
Figure 5: The mature innervation pattern between SGN and HC. .................................... 22
Figure 6: Novel methods to purify cochlear hair cells and nonsensory cells in the cochlea.
........................................................................................................................................... 25
Figure 7: Microarray study result. .................................................................................... 27
Figure 8: Screening for guidance cues enriched in the nonsensory cells.. ....................... 29
Figure 9: Slit molecules are expressed in cochlear tissue during SGN innervating hair
cells. ................................................................................................................................ 33
Figure 10: Spiral ganglion neurons are mislocated in Slit2 mutant cochlea. .................... 36
Figure 11: Developmental change of SGN mispositioning in Slit2 mutant cochlea. ....... 39
Figure 12: Robo are expressed in Cochlear tissue during SGN innervating HC.. ............ 41
Figure 13: Individual SGNs are mislocated in Cochlea epithelium in Robo mutant. . ..... 42
Figure 14: Overall arrangement of cochlea sensory epithelium is not altered in Robo
mutant. .............................................................................................................................. 43
Figure 15: Spiral ganglion territory is expanded towards HC progressively in Robo
mutant cochlea. . ............................................................................................................... 47
Figure 16: A proposed model for Slit/Robo signaling in restricting SGN positioning. .. 51
Figure 17: STARS strategy.. ............................................................................................. 65
Figure 18: Test of STARS hypothesis.. ........................................................................... 68
Figure 19: Relationship between the level of sparseness and the ratio of length between
recombination units.. ......................................................................................................... 69

vi

Figure 20: Examination of the identities of M-GFP positive cells. . ............................... 71
Figure 21: Extension of the sparseness level of STARS technique.. ............................... 73
Figure 22: A two-step process to target STARS transgene to ROSA26 locus ................. 76
Figure 23: PCR genotyping of first step targeting. . ........................................................ 77
Figure 24: PCR genotyping the second step STARS targeting.. ...................................... 78
Figure 25: Sparse activation of EYFP expression in STARS ES cell. . .......................... 79
Figure 26: Strategy to make STARS Knock-in Mouse based on Ai9 ES cell. ................. 81

vii

List of Tables

Table 1. Test of validity and sensitivity of our microarray study ..................................... 28

viii

Abstract

During development, proliferating neuroblasts delaminate from the otocyst to generate
spiral ganglion neurons which settled down in Rosenthal’s canal medial to the cochlea
sensory epithelium. Subsequently spiral ganglion neurons extend their peripheral axons
back to the cochlea by penetrating through the spiral limbus and Kolliker’s organ and
ultimately form the tonotopically organized innervations with cochlea hair cells. How
spiral ganglion neurons territory is defined along the medial-lateral axis is largely
unknown. In this study, we performed microarray analysis and identified Slit molecules
as potential candidates involved in this process. Analysis of Slit2 mutant mouse embryos
showed that a significant number of spiral ganglion neuron soma are not restrained in
Rosenthal’s canal but rather spread randomly over the cochlea tissue, some even go
beyond the hair cells. These mis-positioned spiral ganglion neurons extend their merits
largely along the longitudinal axis and randomly travel within the cochlea tissue without
innervating hair cells. Similar phenotypes are also observed in embryos with mutations
for Robo, the receptors for Slit molecules. Furthermore, the spiral ganglion territory is
dramatically expanded towards the sensory epithelium in Robo mutant in addition to
those individually displaced cells. In situ hybridization showed Robo1 and Robo2 are
expressed in spiral ganglion neurons and Slit 2 and Slit3 are expressed in the greater
epithelium ridge region (Kolliker’s organ) as well as spiral limbus region during the time
spiral ganglion neurons innervate cochlea hair cells. Developmental studies revealed that
these SGNs progressively moved to ectopic positions from their normal locations in the
Rosenthal’s canal during the time SGNs innervate HCs. We propose that this disruption
of spatial patterning of SGNs is attributed to the loss of the restriction force imposed by

ix

Slit/Robo signaling, which serves to refrain SGNs from invading the cochlear epithelium
and to ensure the formation of precise innervation patterns in the peripheral auditory
circuitry.
In the second part of the thesis, we want to dissect the cell lineage relationship between
cochlear hair cells and their neighboring supporting cells, which is largely elusive
primarily due to the lack of a straightforward genetic method to perform the lineage
tracing in the inner ear. To address this need, we developed a method termed STARS:
stochastic gene activation with genetically regulated sparseness. The stochastic
expression was achieved by two cross-linked, mutually-exclusive Cre-mediated
recombinations. The stochastic level was further controlled by regulating Cre/lox reaction
kinetics through varying the intrachromosomal distance between the lox sites mediating
one of the recombinations. We further explore the possibility to extend the sparseness
level by mutagenizing lox sites employed in the STARS transgene. In mammalian cell
lines stably transfected with a single copy of different STARS transgenes, the
activation/knockout of reporter genes was specifically controlled to occur in from 5% to
50% of the cell population and further down to 1% when combined with lox variants with
lower recombination efficiency. STARS can potentially provide a convenient way for
genetic labeling as well as gene expression/knockout in a population of cells with a
desired sparseness level and hence will provide a useful tool to the dissection of the cell
lineage in the inner ear.

1

Chapter 1: Slit/Robo signaling underlies spatial patterning of
cochlear innervation

Neural circuit assembly relies on the proper control of lamina formation during
development. This unique anatomical feature has been best studied in synaptic
connections in the target regions of the cortex, retina, and spinal cord, where molecular
mechanisms have begun to be identified (Sanes and Yamagata 2009; Huberman,
Clandinin et al. 2010; Sanes and Zipursky 2010; Matsuoka, Nguyen-Ba-Charvet et al.
2011; Matsuoka, Jiang et al. 2012). However, assembly of proper synaptic connections
also requires the patterning of neuronal cell bodies into distinct layers or regions.
Although cell migration regulated by extracellular cues such as Reelin provides an
important mechanism to establish different layers (Marin and Rubenstein 2003; Frotscher
2010), how the boundary of each layer is determined and maintained, and how cell body
patterning contribute to circuit development are not well understood.
The peripheral auditory system provides another model to study lamina patterning of both
cell body and synaptic connection during circuit assembly. Here, sound signals
converted by cochlear hair cells (HCs) are conveyed by spiral ganglion neurons (SGNs)
to cochlear nucleus neurons in the brain stem. SGNs are the bipolar afferent neurons in
the medial surface of the cochlear duct that run in parallel with HCs in the cochlear
epithelium. They form stereotyped connectivity with HCs in adult cochleae. During
development, SGNs migrate out from the otocyst and cluster in the Rosenthal’s canal to
form a spatially segregated lamina away from the cochlear epithelial layer (Ruben 1967).
SGNs subsequently extend their peripheral neurites into the cochlear sensory epithelium

2

to innervate HCs, forming a topographic innervation pattern (Appler and Goodrich 2011;
Nayagam, Muniak et al. 2011; Yang, Kersigo et al. 2011). Although previous studies
have characterized in details the differentiation and morphological development of HCs
and SGNs (Kelley 2006; Grimsley-Myers, Sipe et al. 2009; Kelly and Chen 2009; Lu,
Appler et al. 2011; Coate, Raft et al. 2012; Groves and Fekete 2012) , the molecular
mechanisms underlying the spatial patterning of SGNs during their innervation of HCs
are largely elusive.
Axon guidance molecules are important in determining neuronal innervation patterns in
both peripheral and central nervous systems (Tessier-Lavigne 2002; Yamamoto, Tamada
et al. 2002; Huberman, Feller et al. 2008; Salinas and Zou 2008; Cho, Prince et al. 2009;
Shen and Cowan 2010). Previous studies have shown specific expression of axon
guidance molecules in the developing inner ear (Webber and Raz 2006; Fekete and
Campero 2007; Appler and Goodrich 2011), but their potential roles in the development
of spatial organization of cochlear cells have not been examined. Of the guidance
molecules, Slits are of particular interest as they have been shown to play essential roles
in various processes such as axon guidance, branching, fasciculation as well as cell
migration through Roundabout (Robo) (Brose, Bland et al. 1999; Kidd, Bland et al. 1999;
Wang, Brose et al. 1999; Ma and Tessier-Lavigne 2007; Ypsilanti, Zagar et al. 2010).
Recently, it has also been implicated in controlling lamina-specific connections in the
optic tectum (Xiao, Staub et al. 2011). However, their function in the inner ear
development and patterning is not known.
Here we demonstrate a role of Slit in controlling laminar patterning in cochlear
development. First, based on a microarray-based screening, we found Slit2 specifically

3

expressed in the cochlear epithelium. We confirmed this non-sensory cell expression by
RNA in situ hybridization, and the result is consistent with previous studies in chick
(Holmes and Niswander 2001; Battisti and Fekete 2008) and rodents (Holmes, Negus et
al. 1998; Yuan, Zhou et al. 1999; Marillat, Cases et al. 2002; Webber and Raz 2006). We
also found complementary expression of Robo receptors in SGNs. In mouse mutants for
either Slit2 or Robo1 and Robo2, the spatial restriction of SGNs was disrupted with more
severe phenotypes in the Robo1/2 mutant. By tracing developmental changes, we found
that the displacement of SGNs occurs after their initial normal positioning at an earlier
developmental stage. Thus, Slit/Robo signalling serves to prevent SGNs from invading
the cochlear epithelium. Our results revealed a previously unrecognized role of Slit/Robo
signalling in maintaining the spatial segregation between different types of cells. Such
cell assembly organization is an important basis for the formation of precise innervation
patterns in the peripheral nervous system.

4

1.1 Introduction to the assembly of peripheral auditory circuitry

Our sensation of the environment critically depends on the precise neuronal circuitry
connectivity. During hearing, sound information, such as sound frequency, intensity and
timing, was first decoded by a very special type of receptor cells called hair cells located
in the inner ear. This information was subsequently relayed to the central nervous system
for further processing via the bipolar spiral ganglion neurons (SGN). Each spiral ganglion
neuron sends one peripheral process to innervate hair cell via the highly specialized type
of synapse called ribbon synapses and one central process to the cochlear nucleus in the
hindbrain. This information flow from hair cells to spiral ganglion neurons along the
peripheral auditory pathway is essentially indispensable for our sense of hearing. In fact,
most of the sensorineural deafness in human are accounted by the loss of either hair cells
or the spiral ganglion neurons. To date, the best treatment for sensory hearing loss is the
cochlea implant which depends on the effective electrical stimulus directly on the spiral
ganglion afferent fibers. Hence a thorough understanding of the molecular basis of the
connectivity between spiral ganglion neurons and hair cells will provide new insights for
the development of novel therapies for sensorineural deafness.
Mammals are able to hear a very wide range of sound, from 20Hz to 20kHz in frequency
and a difference of 120db (a range of one-million) in amplitude. They are able to respond
to the sound induced vibrations at the atomic scale and amplify the sound signal over 100
fold (Vollrath, Kwan et al. 2007; Gillespie and Muller 2009; Kazmierczak and Muller
2012). This incredible sensitivity and robustness pose a daunting challenge to our
auditory modality. Amazingly, our peripheral auditory circuitry has evolved into an
elegant system tailored to this demanding task. The cochlea epithelium extends

5

throughout the cochlea duct along the basal-apex axis. Each hair cell, depending on its
location along the basal-apex axis, responds to a particular narrow range of sound
frequencies. Consequently, hair cells located on the basal part of the cochlea with wider
open space will respond to high frequency and those located on apex part will respond to
low frequency so to form the tonotopicity along the longitudinal (base to apex) axis.
There are generally two types of hair cells: one row inner hair cells and three rows of
outer hair cells. Inner hair cells are the primary receptors for sound perception and outer
hair cells are believed to perform as an amplifier to increase the sound detection
sensitivity (Liberman, Gao et al. 2002; Dallos, Wu et al. 2008). For spiral ganglion
neurons, more than 90% percent of them belong to type I neurons, which extend
unbranched fiber to make contact with one inner hair cell only; the rest 10% belongs to
type II neuron, which extend the process bypassing the inner hair cell and turn towards
the base and make contacts with multiple outer hair cells along the path (Rubel and
Fritzsch 2002; Meyer, Frank et al. 2009; Appler and Goodrich 2011). Noticeably, one
single inner hair cell is connected with 10 to 20 type I spiral ganglion neurons depending
on the locations of the hair cell along the base-apex axis (Rubel and Fritzsch 2002;
Meyer, Frank et al. 2009), possibly to increase the probability to detect soft sound
(Meyer, Frank et al. 2009). Interestingly, spiral ganglion neurons which innervate hair
cells of the similar frequencies organize their afferent fibers into radial bundles during the
fiber outgrowth. Reminiscent to the organization of tonotopicity along the basal-apex axis
in the hair cells, spiral ganglion neurons also manifest such a pattern, largely due to the
fact that spiral ganglion neurons radially innervate the nearest hair cells despite a few
exceptions (Rubel and Fritzsch 2002). Such topographic organization of tonotopicity is

6

preserved throughout the auditory pathway to ensure the faithful transmission of sound
frequency information.
Extensive efforts have been put into the study of molecular and cellular basis of circuitry
assembly in sensory system like retina, olfactory, spinal cord, etc., which are readily
accessible and ease for genetic or biochemical studies (Dodd and Jessell 1988; Tanabe
and Jessell 1996; Charron and Tessier-Lavigne 2005; Chao, Ma et al. 2009; Lu, Wang et
al. 2009; Adam and Mizrahi 2010; Liang and Luo 2010; Zipursky and Sanes 2010; Imai
and Sakano 2011; Kolodkin and Tessier-Lavigne 2011). In contrast, peripheral auditory
circuitry has lagged behind probably due to the inaccessibility of the cochlea tissue,
which is deeply embedded in the temporal bone and the limited materials available. For
example, Hair cells are very limited in numbers, 5000 per ear in mouse and 8000 in
human. Nevertheless, groundbreaking work has started to shed light on the cellular and
molecular substrates shaping the construction of the peripheral auditory pathway (Morris,
Maklad et al. 2006; Koundakjian, Appler et al. 2007; Meyer, Frank et al. 2009; Appler
and Goodrich 2011; Lu, Appler et al. 2011; Yang, Kersigo et al. 2011; Coate, Raft et al.
2012). Here I will first briefly summarize the current knowledge about the genesis and
differentiation for the two core operating components of the peripheral auditory circuitry:
the cochlea hair cells and the spiral ganglion neurons. Then I will focus more on the
events during each step of the circuitry assembly both anatomically and mechanistically.

7

Figure 1: The Anatomy of Inner Ear. Schematic drawing showing the structure of the
inner ear, whole mount bony cochlear, cross section view of cochlear duct, organ of Corti
and hair bundles alignment of the hair cells, sequentially. (Adapted from Neuroscience,
Dale Purves et al)

Inner pillar cell outer pillar cell
Deiters’ cell
cells of hensen
cells of boettcher
Modified from Neuroscience, DALE PURVES et al

8

1.1.1 Cochlea hair cells: the “mechanical” device for sound
perception

At the heart of sound perception is the conversion of mechanical force elicited by sound
waves into electrical signals, the process of which is called mechanotransduction. This
critical step takes place at cochlea hair cells sitting on the sensory epithelium called the
organ of Corti in the inner ear (Figure 1). Organ of Corti is composed of mechanosenstive
hair cells and heterogeneous nonsensory supporting cells (pillar cell, Deiter cell, Hensen
cell, etc.), which are mixed together to form an ordered alternative mosaics (Kelley 2006;
Kelley 2007; Kelly and Chen 2007; Kelly and Chen 2009). The organ of Corti itself sits
on the basilar membrane and is covered by the tectorial membrane on the top which is
attached to the hair bundle of the outer hair cells. During hearing, sound wave caused
fluid pressure travels down the cochlea duct and leads to the vibration of the basilar
membrane. This vibration causes deflection of hair bundle on the apical surface of the
hair cells and leads to the opening of the mechanically gated ion channels on the tip of
hair bundle and depolarization of hair cells (Nayak, Ratnayaka et al. 2007; Vollrath,
Kwan et al. 2007; Beurg, Fettiplace et al. 2009; Gillespie and Muller 2009; Schwander,
Kachar et al. 2010; Kazmierczak and Muller 2012).
1.1.1.1 Morphological and functional characteristics of hair
cells

Hair bundle is the ultimate mechanosensitive organelle of the hair cell. It consists of two
or more rows of actin-filled structure called stereocilia with increasing height (Figure 2).
There are numerous extracellular filament links running to connect neighboring

9

Figure 2: Hair bundle is the mechanosensitive organelle of hair cells. A. Visualization
of hair bundles of cochlear hair cells as well as the tip link connecting the tip of the
shorter stereocilia to the side of its higher neighbors based on scanning electron
microscopy. B. (a) Cross section view of hair bundle and the apical surface of hair cell.
Hair bundles are composed of several rows of actin filled structure called stereocilia and
one single kinocilium filled with microtubules. Stereocilia and the kinocilium are
connected to each other via extracellular links including kinociliary links, tip links, top
connectors and ankle links. (b) Schematic drawing of the molecular constitutes of the tip
link complex. CDH23 homodimers and PCDH15 homodimers form the upper and lower
part of the tip link, respectively. Shaded areas illustrate the electron-dense regions from
the transmission electron microscopy image close to the upper and lower anchorage sites
of tip links. (Adapted from (Kazmierczak and Muller 2012))

Hair cell Stereocilium Tip link
A
B
Kazmierczak and Muller, 2011

10

stereocilia: tip links connect the tip of the shorter stereocilium to the shoulder of its
higher neighbor and is thought to gate the putative mechanical gated ion channel directly
(Nayak, Ratnayaka et al. 2007; Vollrath, Kwan et al. 2007; Beurg, Fettiplace et al. 2009;
Gillespie and Muller 2009; Schwander, Kachar et al. 2010; Kazmierczak and Muller
2012), top connectors and ankle links are remodeled during the development and
important to maintain the structural integrity and coherence of the hair bundles (Vollrath,
Kwan et al. 2007; Gillespie and Muller 2009). The stereocilia manifest itself in a ‘U’ or
‘V’ shape pointing towards the lateral axis and forms the medial-lateral planar polarity
(Figure 2) (Jones and Chen 2007; Kelly and Chen 2007). This polarity is important
because it dictates the directional sensitivity to mechanical stimulation: deflection of the
hair bundle in the direction of the longest stereocilia increases the probability of channel
opening and depolarizes the hair cell; conversely, deflections in the opposite direction
will reduce channel open probability and lead to hyperpolarization. (Vollrath, Kwan et al.
2007; Gillespie and Muller 2009; Kazmierczak and Muller 2012).
The molecular identities of mechanotransduction machinery have begun to be revealed
largely due to the genetic approaches characterizing deafness genes. Among those
significant players are the tip link proteins: cadherin 23 and protocadherin15, homodimer
of which constitutes the upper and lower half of tip link filament respectively (Figure 2)
(Siemens, Lillo et al. 2004; Sollner, Rauch et al. 2004; Kazmierczak, Sakaguchi et al.
2007; Muller 2008; Sakaguchi, Tokita et al. 2009; Sotomayor, Weihofen et al. 2010);
myosin motor proteins, myosin3a, myosin7a and myosin15, are involved in regulating the
length of hair bundle, and myosin1c involved in adaptation and adaptor proteins
harmonin and SANS acting as the scaffold to bridge different protein players together

11

(Vollrath, Kwan et al. 2007; Petit and Richardson 2009; Schwander, Kachar et al. 2010;
Kazmierczak and Muller 2012).
The molecular constitute of the most wanted protein in the hearing field, the mechanical
transduction channel, has remained elusive (Vollrath, Kwan et al. 2007; Gillespie and
Muller 2009). However, we do know substantially about its biophysical property. For
example, the channel opens within micro-seconds scale in response to mechanical
stimulation, arguing it is directly gated by force rather than chemical signaling which
usually happens at millisecond scale (Corey and Hudspeth 1983; Crawford, Evans et al.
1989). In addition, hair cells display adaptation to the continuing mechanical stimulus on
both a fast and slow time scale, presumably to increase its dynamic range facilitating its
frequency selectivity and intensity amplification (Wu, Ricci et al. 1999; Fettiplace and
Ricci 2003; Vilfan and Duke 2003; Gillespie and Cyr 2004; Stauffer and Holt 2007;
Hudspeth 2008).
Taken together, hair cell along with the accessory structure in the organ of Corti is
designed as elegant machinery with unparalleled sensitivity, robustness and high dynamic
range to fulfill the task of sound detection happened every single second in our daily life.
1.1.1.2 Generation and differentiation of hair cells

The whole inner ear is originated from the structure called ‘otic placode’, which is the
thickening of the ectoderm on the outer surface of an embryo adjacent to hindbrain and is
readily observable after embryonic day 8 (E8) (Figure 3). The otic placode invaginates
from the surface to form the hollow structure called ‘otic vesicle ’ or otocyst around E10.
From this time point on, various signaling molecules including BMP, SHH, Wnts and

12

FGFs operate within the otocyst to set up subdomains which will give rise to different
structures of the future cochlea (McKay, Lewis et al. 1996; Liu, Chu et al. 2003; Wright
and Mansour 2003; Ozaki, Nakamura et al. 2004; Solomon, Kwak et al. 2004;
Riccomagno, Takada et al. 2005; Fritzsch, Pauley et al. 2006; Ohyama, Mohamed et al.
2006; Kelley 2007; Kelly and Chen 2007; Kelly and Chen 2009). Soon after cochlea duct
starts to form, the ventral part of the otocyst show clear boundaries marked by expression
of certain molecules with Sox2 marking the prosensory domain (Figure 3). Around E14,
cell cycle exit of the precursors starts from the apex and moves towards the base driven
by certain cyclin-dependent kinase inhibitors, including p27 (Ruben 1967; Chen and
Segil 1999; Lee, Liu et al. 2006; Kelly and Chen 2007; Kelly and Chen 2009). The
expression of Math1, a bHLH gene specifying the fate of hair cells both necessarily and
sufficiently (Bermingham, Hassan et al. 1999; Zheng and Gao 2000; Izumikawa, Minoda
et al. 2005; Jones, Montcouquiol et al. 2006; Gubbels, Woessner et al. 2008), marks the
onset of differentiation in hair cells. The wave of differentiation initiates from the
midbasal region of the cochlea and progresses towards both base and apex directions
(Figure 3). Besides this longitudinal gradient, the differentiation process also progresses
along the medial-lateral axis (from inner hair to outer hair cell). As the differentiation
proceeds, hair bundles start to emerge on the apical surface between E16 and E17 and
ultimately achieve the uniform orientation under the effect of planar cell polarity (PCP)
signaling pathway (Jones and Chen 2007; Kelly and Chen 2007). Terminal
differentiation and maturation during late embryonic and early postnatal stages will
ultimately make hair cells ready to do the job before the onset of hearing around 12 days

13

Figure 3: Development of Hair cells. A. The timeline of hair cell specification and
differentiation. The whole inner ear structure comes from the invagination of the otic
placode which forms otocyst around E9. Cell cycle exit for cochlear epithelium cells
finish around E14. Specific markers start to emerge in hair cells from E14. Hair cells
start to show mechanosensitivity around P0 and continue to mature during the early
postnatal period. Kinocillum disappear from hair cells around P6 and finally the onset of
hearing starts around P12. B. Key events underlying hair cell specification and
differentiation. (a). The prosensory domain (outlined with bracket) is initially marked by
the expression of Sox2 (yellow color) under the action of Sox2, Lmx1a and Notch as well
as Hedgehog (HH) signaling. The diagrams show the whole-mount cochlea (top) and the
cross-section of the cochlea respectively. (b) The cell cycle exit of the progenitors
progresses from the apex towards the base under the action of the cyclin-dependent
kinase inhibitors (e.g. p27, red color). (c) Hair cell differentiation (marked by Math1
expression, green color) starts from the base and moves towards the apex. As a second
gradient, hair cells are also specified along the medial-lateral axis of the cochlea from the
inner to outer hair cells. (d) Inductive and inhibitory signals shape the cellular patterning
of the organ of Corti, including Notch signaling dependent and independent mechanisms.
(e) After hair cells and their supporting cells are specified, the planar cellular polarity
(PCP) signaling patterns the organ of Corti and leads to the uniform orientation of the
‘V’-shaped hair bundles on the apical surfaces of hair cells. (Adapted from (Doetzlhofer,
White et al. 2004), (Kelly and Chen 2009))

Kelly and Chen, 2009
E8
A
B

14

after birth in mice. It is obvious that the generation and differentiation of hair cells is a
highly regulated and coordinated process involving multiple signaling pathways
operating together.
1.1.2 Formation of spiral ganglion neurons from a developmental
view

Similar to hair cells, spiral ganglion neurons also derive from the proneurosensory
domain in the anteroventral quadrant of the otocyst (Rubel and Fritzsch 2002; Appler and
Goodrich 2011). Within the proneurosensory domain, the expression of Sox2, a SRY-
related high-mobility-group box (Sox) transcription factor belonging to the family of
SoxB1 (Wegner and Stolt 2005; Lefebvre, Dumitriu et al. 2007), seems to drive the
proliferation of precursors for both neurons and sensory tissues within the inner ear and
mutations with Sox2 lead to the loss of Spiral ganglion in mouse (Puligilla, Dabdoub et
al. 2010). Subsequently, the expression of a proneural bHLH gene, Ngn1, promotes
proneural cell fate within the neurogenic domain. The expression another differentiation
bHLH gene, NeuroD, at later stage drives the differentiation of neuron precursor towards
vestibule-cochlear fate. Both Ngn1 and NeuroD are essential for the genesis of all the
neurons within the inner ear as their mutation lead to complete (Ngn1) or near complete
(NeuroD) loss of vestibulecochlear ganglion (Ma, Chen et al. 1998; Liu, Pereira et al.
2000; Ma, Anderson et al. 2000; Kim, Fritzsch et al. 2001; Jahan, Kersigo et al. 2010). At
the same time this bHLH transcriptional network operating to set up the neuron cell fate,
neuroblasts start to delaminate from the ventral otocyst around E10.5 to coalesce into the
cochlear - vestibular ganglion (CVG) to the medial part of otic epithelium. As
development progresses, CVG is segregated into the spiral ganglion medially which

15

coiled throughout the cochlea duct and vestibular ganglion laterally which innervate
sensory epithelium of vestibular part.
It is generally believed that spiral ganglions exit the cell cycle rapidly primarily around
E11.5 to E12.5 starting from the base and moving towards the apex (Ruben 1967; Matei,
Pauley et al. 2005; Koundakjian, Appler et al. 2007) . This is in sharp contrast to the cell
cycle exit in hair cells, which progresses from the apex to the base.

1.1.3 Afferent innervation from spiral ganglion neurons to hair
cells

The establishment of functional connectivity between cochlear hair cell with spiral
ganglion neurons is a multiple step, highly orchestrated and regulated process. Similar to
the neuronal circuitry assembly in other central as well as peripheral systems, this process
largely consists of axon guidance to the target cells, formation of initial synaptic
connections and synaptic pruning and remodeling to the final mature synaptic
connectivity patterns (Defourny, Lallemend et al. 2011). A cohort of key mechanisms
came into play to shape the precise innervation pattern between spiral ganglion neuron
and hair cells and more mechanisms are just beginning to be unraveled.
1.1.3.1 Afferent fiber outgrowth towards sensory epithelium

Soon after Spiral ganglion neurons are specified and settled down in the Rosenthal’s
canal close to the modiolus roughly around E13, they start to extend one central axon
towards cochlear nucleus as well as one peripheral axon towards the developing sensory

16

OC
SGN
GER
SL
LER
Mesenchyme
SGN
~E10
TOP VIEW
SIDE VIEW
E13
SGN
Modiolus
Otocyst
SGN
E14
SGN
GER
SL
LER
Mesenchyme
Prosen
-sory
Modiolus
Rosenthal’s canal
SGN
IHC
OHC
SL
GER
E16
/E18
Rosenthal’s canal
L M
V
D
SGN
Otocyst
OC
SGN
GER
SL
LER
IHC
Mesenchyme Rosenthal’ Canal
Figure 4: Schematic illustration of the assembly of peripheral
auditory circuitry. Early in development (around E10), spiral
ganglion neurons delaminate from developing otocyst and coalescent
to form a tight cluster in the Rosenthal’s canal close to the modiolus
around E13. They subsequently extend their peripheral neurites to
innervate hair cells located in the organ of corti at cochlear epithelium
while their soma are restricted within the Rosenthal’s canal , forming
a topograhically organized connectivity.

17

epithelium (Figure 4). At this time, the prosensory region in the cochlea is still
developing and no hair cells have been specified yet. The peripheral axons of SGN are
largely branched and actively explore the mesenchyme cells surrounding the cochlea
epithelium as well as the epithelial cells in the Kolliker’s organ to the medial part of the
cochlea prosensory domain (Farinas, Jones et al. 2001; Huang, Thorne et al. 2007; Appler
and Goodrich 2011; Defourny, Lallemend et al. 2011).
Starting from E14, hair cells begin to be differentiated from base to apex following the
wave of cell cycle exit from apex to base. In addition to this longitudinal direction of
differentiation, hair cells are also differentiated from the medial to the lateral axis so
inner hair cells are generated before outer hair cells (Kelley 2006; Lee, Liu et al. 2006;
Kelley 2007; Kelly and Chen 2009). As they are differentiated, hair cells migrate to the
apical surface of the sensory epithelium and start to express several signature genes like
Myosin VII a, Math1 and Calretinin, etc. At this time, the peripheral axons of SGN are
bundled into radial fibers and begin to rigorously invade sensory epithelium as
unbranched processes. Depending on the cell type, the axons of type I spiral ganglion
neurons stop at the level of inner hair cells and form a specialized type of synapse called
ribbon synapses with them. In contrast, the axons of type II spiral ganglion neurons cross
the tunnel of organ of Corti, turn towards the basal directions and make connections with
multiple outer hair cells along their path. This exclusive innervation pattern of different
types of spiral ganglion neurons to different hair cell types may not be true early in
development. It was reported that type I axons also make transient connections with
outer hair cells during the establishment of the innervations from spiral ganglion to hair

18

cells while type II axons make exclusive connections with outer hair cells (Echteler 1992;
Simmons 1994; Huang, Thorne et al. 2007).

1.1.3.2 Molecular basis of spiral ganglion neuron path
finding

Accumulating evidence suggests traditional axon guidance molecules are involved in the
patterning of spiral ganglion neuron innervation to hair cells (Webber and Raz 2006;
Fekete and Campero 2007; Huang, Thorne et al. 2007; Appler and Goodrich 2011; Yang,
Kersigo et al. 2011). Among these candidates, many of them have been shown to express
in the inner ear of chick and/or mammals. Examples are: Ntn1 and its receptor (Gillespie,
Marzella et al. 2005; Matilainen, Haugas et al. 2007; Abraira, del Rio et al. 2008; Lee and
Warchol 2008) , Semaphorins and their receptors (Miyazaki, Furuyama et al. 1999;
Murakami, Suto et al. 2001; Chilton and Guthrie 2003), Wnts and their receptors
(Sienknecht and Fekete 2008; Sienknecht and Fekete 2009), Ephrins and the Ephs
(Bianchi and Liu 1999; Pickles, Claxton et al. 2002; Coate, Raft et al. 2012) and finally,
Slit and Robo (Holmes, Negus et al. 1998; Yuan, Zhou et al. 1999; Holmes and
Niswander 2001; Marillat, Cases et al. 2002; Webber and Raz 2006; Battisti and Fekete
2008).
Despite extensive studies, few reports have demonstrated the functional role of any these
guidance cues in the peripheral auditory circuitry assembly in vivo except the recent
study reports that ephrin-B2/EphA4 interaction promotes axon fasciculation spiral
ganglion neurons during their innervation to hair cells (Coate, Raft et al. 2012). The roles

19

of these well-known guidance molecules await further analysis use genetic approaches,
e.g. analysis of gene knockout animals.
1.1.3.3 The role of other cell types of the cochlea in the
peripheral auditory innervation

It has been suggested that the initial setup of spiral ganglion neuron innervation with
cochlea hair cells is independent of hair cells based on the following evidence: first, the
peripheral neurites of SGN reach the cochlear epithelium before HCs are differentiated
(Farinas, Jones et al. 2001; Koundakjian, Appler et al. 2007; Appler and Goodrich 2011);
second, in mouse mutants lacking HCs or mutants with immature HCs, SGN are correctly
positioned and are able to innervate the sensory epithelium correctly (Xiang, Gao et al.
1998; Xiang, Maklad et al. 2003; Fritzsch, Matei et al. 2005). In fact, spiral ganglion
neuron axons make contact with numerous other nonsensory cells in the cochlea during
their trajectory to innervate hair cells. It is very likely those nonsensory cells play an
active role in the peripheral auditory circuitry assembly rather than simply being a
structural component in the cochlear. I will summarize a few examples below as
previously reported (Appler and Goodrich 2011).
Otic mesenchymal cells: mesenchyme cell fill in the space of the otic capsule and
separate the spiral ganglion with cochlea epithelium. They are actually the first cells
spiral ganglion encounter as they extend the axon towards cochlea epithelium and also
the place where their axons are bundled to form the radial fibers. Consistent with this
observation, Coate et al recently reported that otic mesenchyme establishes an
Eph/ephrin-mediated fasciculation signal that promotes inner radial bundle formation
(Coate, Raft et al. 2012). Kolliker’s organ, also called GER (greater epithelial ridge) is a

20

transient cell population which lies immediately adjacent to hair cells on the medial side
in the embryonic cochlea epithelium. Epithelial cells in Kolliker’s organ have been
shown to hold the potential to differentiate into hair cells upon the introduction of Math1
(Zheng and Gao 2000; Woods, Montcouquiol et al. 2004) or neurons upon the forced
expression of Ngn1 (Puligilla, Dabdoub et al. 2010). In addition, it has been shown that
cells in Kolliker’s organ secrete ATP to depolarize inner hair cells and leads the
transmitter release triggering action potentials in spiral ganglion neurons (Tritsch, Yi et
al. 2007). This sensory input independent event serves to provide the spontaneous activity
to the peripheral auditory circuitry prior to the onset of hearing and is replaced by sound
input as hearing starts.
Schwann cells: schwann cells are the only cell type in the inner ear originated outside the
otic placode. They come from the neural crest to the inner ear around E10 and form
myelination with spiral ganglion neurons. It has been shown that Schwann cells
contribute to axon guidance as well as neuronal survival through the analysis of ErbB2
mutant mouse (Morris, Maklad et al. 2006). In this mutant mouse, spiral ganglion
neurons migrated beyond their normal position and were positioned closer to the
modiolus. In addition, their neurites are disorganized and the neurons themselves are lost
presumably due to the reduction of neurotrophin expression.

21

1.2 Transcriptome analysis to identify molecular candidates
underlying the spatial patterning of peripheral innervation

One of the key questions in the assembly of peripheral auditory circuitry is what are the
mechanisms for defining the spatial boundary of spiral ganglion neurons (SGN) at the
time of their innervations of cochlear hair cells. As I described above in the introduction
part, in spite of extensive studies focused on the development of hair cells and spiral
ganglion neurons, the mechanisms that shape the connectivity between them are largely
elusive. Recent studies have begun to shed light on the molecular underpinnings
regulating the developmental patterning of the spiral ganglion. However, the detailed
mechanisms to set up the clear spatial separation between the spiral ganglion and its
target sensory epithelium tissue (Figure 5) have not been elucidated.
Previously it has been observed that in mouse mutants lacking HCs or mutants with
immature HCs, SGNs are correctly positioned and are able to innervate the sensory
epithelium correctly (Xiang, Gao et al. 1998; Xiang, Maklad et al. 2003; Fritzsch, Matei
et al. 2005). These results lead to our hypothesis that the molecular signals regulating the
spatial positioning of SGNs are provided by nonsensory cells within the cochlea. We
thus screened for genes that are enriched in nonsensory cells of the cochlea in comparison
to HCs by performing a microarray transcriptome analysis.

22

Figure 5: The mature innervation pattern between SGN and HC. SGN (green, Tuj1
immunostaining) soma are restricted in Rosenthal’s canal while their peripheral axons
penetrate through otic mesenchyme and GER to innervate HC (red, Myo6
immunostaining) in OC at E18 (A and B: top view, C and D: cross sectioning view).
SGN: spiral ganglion neuron, RF: radial fiber, OC: organ of Corti, SL: spiral limbus,
GER: greater epithelial ridge, LER: lesser epithelial ridge (or outer sulcus)

SGN
RF
IHC OHC
SGN
RF
IHC
OHC
SGN
IHC
OHC
SL
GER
A B
C D
OC
L
SGN
GER
SL
LER
IHC
Mesenchyme
Rosenthal’ Canal
M
V
D
Modiolus

23

1.2.1 Novel methods to purify cochlea hair cells and non-hair
cells

To make a comparison of gene expression profiles between non-sensory cells in the
cochlea and hair cells, we need to purify hair cells and nonsensory cells respectively to
get a certain amount of materials for microarray experiments. This posed a significant
challenge especially for hair cells due to the very limited number of cells ( 3000
cells/cochlea in mouse) and the inaccessibility of the cochlear tissue embedded deeply in
the inner ear. Previous microarray studies of the inner ear used the entire cochlea as the
materials (Chen and Corey 2002; Chen and Corey 2002; Hertzano, Montcouquiol et al.
2004; Gong, Karolyi et al. 2006; Sajan, Warchol et al. 2007) and have provided important
insights into the gene expression profiles in the inner ear, however, these approaches will
dilute the gene expression of a particular cell type (e.g. Hair cells, which accounts for less
than 5% of the total population in the cochlear tissue) and lose a lot of valuable
information. To address this problem, we took two complementary approaches to label
hair cells specifically in surgically dissected mouse cochlear epithelium at a proper
developmental stage, and collect separately substantial numbers of hair cells and their
sibling supporting cells with fluorescence activated cell sorting (FACS).
In the first approach, we utilized the styryl dye FM1-43, which can permeate the
mechanotransduction channel of HCs and has been used to study the acquisition of
mechanosensitivity in HCs (Gale, Marcotti et al. 2001; Geleoc and Holt 2003; Meyers,
MacDonald et al. 2003; Lelli, Asai et al. 2009). Consistent with previous reports, FM1-43
was able to label HCs specifically whereas other cells in the cochlea were not labelled
(Figure 6A, top panel). In the second approach, we screened several Cre-driver mouse

24

lines with the Cre expression controlled by the promoters of calcium binding proteins.
We found that in the Parvalbumin-Cre line, Cre activity in the cochlea was specifically
limited to HCs, although outer hair cells were labelled in a mosaic pattern, as shown by
the pattern of red fluorescence when crossed with a Cre-dependent tdTomato reporter
line, Ai14 (Madisen, Zwingman et al. 2010) (Figure 6A, bottom panel). SGNs could be
significantly labelled by both approaches (data not shown) and were trimmed away
during the tissue dissection step.
Combing the two approaches, we purified HCs and nonsensory cells of the cochlea by
using fluorescence-activated cell sorting (FACS) (Figure 6B and methods). After the
sorting, we observed over 90% of purity for HCs in the fluorescence-positive population
(Figure 6C) and no fluorescent cells in the negative population (Figure 6D). We further
performed RT-PCR for hair cell markers MyosinVIIa in both positive and negative
populations and found the amplification product only in the positive but not negative
populations (Figure 6E), suggesting there is no or negligible contaminations between the
two groups of cells. We then proceed with the two populations of cells to the microarray
experiment and performed data analysis using bioinformatics techniques to get the
normalized expression values for each individual gene probe on the microarray chip (see
methods).

25

Figure 6: Novel methods to purify cochlear hair cells and nonsensory cells in the
cochlea. A. Top panel, DIC and fluorescence images of a wild type mouse cochlea at
postnatal day 6 (P6) stained with the styryl dye FM1-43. Bottom panel, images of a
cochlea of a Parvalbumin-Cre; Ai14 (ROSA-LSL-tdtomato) mouse at P6. B: Gating
configuration of FACS sorting for separating HCs (fluorescence positive population as
Positive Population Negative Population
DIC PV-tdTomato Merge
DIC FM1-43 Merge
A B
C D
E
Fluorescence Intensity
+
-

26

illustrated by +) from other cochlear cells (negative population as illustrated by -) based
on fluorescence intensity. C. Representative images of cells from the positive population
after sorting. Most of them had fluorescence and displayed typical shape of hair cells. D.
Representative images of cells from the negative population after sorting and no
fluorescent cells were detected. E. RT-PCR for hair cell marker MyosinVIIa showed
signal only in positive population but not in negative population.

1.2.2 Microarray study

We compared the expression values of each individual probe between the two cell
populations (Figure 7). The established hair cell markers were found to be highly
enriched in the fluorescence-positive (i.e. HC) population (Gene 1 to 10, highlighted by
red circles), while markers for cochlear supporting cells were highly enriched in the
negative population (Gene 11 to 17, highlighted by green circles) (Figure 7 and Table 1).
This result confirmed the validity and sensitivity of our experimental method.

27

Figure 7: Microarray study result. A scatter plot of the expression value of each
individual probes from microarray experiments between the positive cell population and
negative cell population. Fold of change of 1x, 2x and 5x was illustrated by the yellow,
red and blue lines respectively. Selected markers for hair cell specific genes (1-10 in the
table 1) and supporting cells specific genes (11-17 in the table 1) were highlighted by the
red circles and green circles respectively

Intensity (Positive Cells)
Intensity (Negative Cells)
1x
5 fold
2
7
5
1
3
4
8
10
6
17
16
11
12
13
14
15
9
5 fold
2 fold
2 fold
0
5
10 15
0 5 10 15

28

Table 1. Test of validity and sensitivity of our microarray study. A selected list of
probes and their fold of change (positive population/ negative population) for hair cell
specific genes and supporting cell specific genes from the microarray experiment. P
value and related human deafness form associated with each gene are also listed. Gene
index corresponds to the number listed in the scatter plot in Figure 7

Hair Cell Specific Markers
Index Gene name Probe ID Deafness Form Fold of Change(+/-) P value
1 Atoh1 1449822_at 7.97 0.000936988
2 Cdh23 1432346_a_at
DFNB12, Usher syndrome 1D
29.29 3.90E-06
3 Myo3a 1431983_at
DFNB30
94.26 6.68E-07
4 Myo7a 1421385_a_at
DFBN2;DFNA11; Usher syndrome 1B
13.45 0.000794132
5 Myo15a 1421615_at
DFNB3
15.20 1.42E-05
6 Otof 1420419_a_at
DFNB9
47.28 1.43E-06
7 Pcdh15 1444317_at
DFNB23, Usher syndrome 1F
13.82 1.04E-05
8 Slc26a5 1421725_at
DFNB61
58.01 3.34E-06
9 Tmhs 1429266_at
DFNB67
140.48 7.31E-07
10 Tmc1 1421626_at
DFNA36
38.78 3.68E-05
Supporting Cell Specific
Markers
11 Coch 1423285_at
DFNA9
0.16 0.003925
12 Col11a2 1423578_at
DFNA13
0.25 0.0002835
13 Crym 1416776_at 0.15 0.0009579
14 Dfna5 1417903_at 0.19 0.004602797
15 Gjb6 1448397_at
DFNA3B
0.09 4.80E-06
16 Jag1 1434070_at 0.27 0.003679259
17 Pou3f4 1422164_at
DFNX2
0.08 2.96E-05

29

1.2.3 Screening for guidance molecules enriched in non-hair cell
in the cochlea

Based on the result we obtained the microarray experiment, we screened of a selected list
of known axon guidance molecules and morphogens to see which of them are highly
enriched in the non-sensory cell population. We found that Slit molecules, especially in
particular Slit2, stand out as the top candidates potentially involved in the regulation of
SGN innervations of cochlear HCs, as they are highly enriched in the nonsensory cell
population (Figure 8)

Figure 8: Screening for guidance cues enriched in the nonsensory cells. Plot of fold
of change between negative cell population and positive cell population for selected
known guidance molecules and morphogens from the microarray gene expression
analysis. N= 3. Bar = SD.
F o ld o f C h a n g e
(N e g a tiv e /P o s itiv e )
B m p 1
B m p 4
B m p 6
B m p 1 5
E fn a 2
E fn b 1
E fn b 2
E fn b 3
F g f1
F g f2
N tn 1
N tn 3
N tn 4
N tn 5
S e m a 3 a
S e m a 3 b
S e m a 3 g
S h h
S lit1
S lit2
S lit3
0
2
4
6
8
1 0

30

1.3 The role of Slit family molecules in shaping peripheral auditory
circuitry assembly

To understand the role of Slit family molecules in regulating the peripheral auditory
circuitry assembly, we first confirmed their expression in the cochlear tissue using in situ
hybridization. Then we took advantage of knockout mouse of Slit1, Slit2 and Slit3 to
examine the spatial innervation pattern from spiral ganglion to cochlear hair cells at
embryonic day 18 (E18). We further performed developmental studies of those mutants
to examine the innervation pattern change over the course of assembly of the peripheral
auditory circuitry.
1.3.1 Background of Slit and Robo family molecules

Slit are large secreted proteins that bind their transmembrane receptors Roundabout
(Robo) (Brose, Bland et al. 1999; Kidd, Bland et al. 1999). Slit/Robo forms one of the
most crucial axon guidance cues in the development of the nervous system. Both Slit and
Robo are identified from mutant screening in Drosophila for genes regulating
commissural axons midline crossing (Rothberg, Hartley et al. 1988; Rothberg, Jacobs et
al. 1990; Seeger, Tear et al. 1993; Kidd, Brose et al. 1998). In Slit/Robo mutant,
commissural axons make multiple recrossing of the midline instead of staying away in
contralateral side. Ever since, homologs of both Slit and Robo have been identified in
many other species as well including mammals (Chedotal 2007).
There are three Slit molecules expressed in the nervous system as well as other organs in
mammals (Marillat, Cases et al. 2002). Slit is generally composed of the following
structures sequentially from the N terminus to the C terminus: an N-terminal signal

31

peptide, four domains (D1-D4) with leucine-rich repeats (LRR), several EGF-like
sequences, a laminin-G domain and a C-terminal cysteine-rich domain (Ypsilanti, Zagar
et al. 2010). It has been shown that the full length Slits can be cleaved to produce a
longer N-terminal fragment responsible for the binding to Robos and a shorter C-
terminal fragment with unknown functions (Wang, Brose et al. 1999; Nguyen Ba-
Charvet, Brose et al. 2001; Ypsilanti, Zagar et al. 2010).
There are generally three Robo genes expressed in the nervous system for most
vertebrates (Chedotal 2007; Ypsilanti, Zagar et al. 2010). As a cell adhesion molecule of
the immunoglobulin (Ig) superfamily, Robo generally consists of five Ig motifs, three
fibronectin type III domains and four cytoplasmic domains (Ypsilanti, Zagar et al. 2010).
All Robo receptors can be alternatively spliced, producing different isoforms varying in
either on the N terminus (Robo 1 and Robo2) or C terminus (Robo3) (Kidd, Brose et al.
1998; Yue, Grossmann et al. 2006; Chedotal 2007; Ypsilanti, Zagar et al. 2010).
Slit usually binds to Robo as a dimer and this binding has been shown to be stabilized by
heparan sulphate proteoglycans (HSPGs), a proteoglycan core protein attached to HS
chains either in the extracellular space or attached to the cell membrane. HSPG binds the
D4 domain of Slit and potentates Slit activity (Seiradake, von Philipsborn et al. 2009).
Besides their typical roles in axon guidance, Slit/Robo signaling has been shown to play
essential roles in a variety of other developmental processes, including axon branching,
axon fasciculation, cell proliferation and cell migration in numerous systems (Plump,
Erskine et al. 2002; Grieshammer, Le et al. 2004; Nguyen-Ba-Charvet, Picard-Riera et al.
2004; Andrews, Barber et al. 2007; Guan, Xu et al. 2007; Ma and Tessier-Lavigne 2007;

32

Andrews, Barber et al. 2008; Legg, Herbert et al. 2008; Ypsilanti, Zagar et al. 2010;
Zheng, Geng et al. 2011; Borrell, Cardenas et al. 2012; Cariboni, Andrews et al. 2012;
Jaworski and Tessier-Lavigne 2012; Liu, Lu et al. 2012)
1.3.2 Expression of Slit molecules in cochlear tissue

To verify the expression of Slit molecules in the cochlea, we performed in situ
hybridization for all three Slit genes on in the whole mount cochlear tissue.
At both E14 and E16, Slit1 signal was located in spiral ganglion neurons (SGNs) (Figure
9, A, D and G). On the other hand, very strong Slit2 signals were located in the spiral
limbus (SL) region as well as part of the GER (greater epithelial ridge or Kolliker’s
organ) at both E14 and E16 (Figure 9, B, E and H). Slit3 expression was located in
similar regions as Slit2 although the signal was weaker (Figure 9, C, F and I). These data
are consistent with our microarray expression data for Slit molecules and largely agree
with previous descriptions of Slit expression patterns in the mouse cochlea (Webber and
Raz 2006) and suggest an important role of Slit molecules in shaping the connectivity of
SGNs with cochlear HCs.

33

Figure 9: Slit molecules are expressed in cochlear tissue during SGN innervating
hair cells. In situ hybridization of Slit1 (A, D and G), Slit2 (B, E and H) and Slit3 (C, F
and I) molecules for whole mount cochlea at E14 and E16. Slit1 is expressed in spiral
ganglion regions at both E14 and E16 (A, D and G). Both Slit2 and Slit3 are expressed in
part of GER region as well SL region (B, E, H and C, F, I, respectively. G, H and I are
enlarged view from D, E and F respectively. LER: lesser epithelial ridge (or outer
sulcus), OC: organ of Corti, GER: greater epithelial ridge (or Kolliker’s organ, SL: spiral
limbus, SGN: spiral ganglion neuron

GER
SGN
OC
LER
GER
SGN
OC
LER
Slit1 Slit2 Slit3
SL
SL
E14 E16
A B C
D E F
G H I

34

1.3.3 Mis-positioning of spiral ganglion neurons in Slit2 mutant

To elucidate the roles of different Slit molecules in the peripheral auditory circuit
assembly, we first examined the innervation pattern between SGNs and HCs at E18
(embryonic day18). In the wild type animal, SGN (Green, Tuj1 immunostaining) cell
bodies were well restricted in the Rosenthal’s canal close to the modiolus and their
peripheral neuritis were bundled into radial fibers (RFs) to innervate both inner hair cells
and outer hair cells (Figure 10 A, first row). No obvious defect was observed in Slit1-/-
(Figure 10A, second row) or Slit3-/- cochlea (Figure 10 A, fourth row). However, in
Slit2-/- mutant, a significant number of SGN cell bodies were not restrained within the
Rosenthal’s canal but randomly distributed within the cochlear epithelium. Some of them
were even dispersed more laterally beyond hair cells (Red arrows, Figure 10 A, third
row). The neurites of these mispositioned SGNs mostly travelled randomly in the
cochlear tissue without innervating hair cells (marked by white arrows in Figure 10).
Noticeably, the majority of SGNs were located correctly within the Rosenthal’s canal
(Figure 10 A, third row). Their peripheral neurites innervated hair cells correctly as
evidenced by the normal organization of type I and type II fibers (Figure 10 B, bottom
row) compared with wildtype (Figure 10 B, top row), although there were some
overshooting fibers presumably from mispositioned SGNs. 3D Projection of confocal
images showed that the somas of those misplaced neurons were positioned at a lower part
of the sensory epithelium, close to the basal membrane (data not shown). We quantified
the average number of mispositioned SGNs (mSGNs) in each cochlea (Figure 10 C).
Only Slit2-/- mutants showed mispositioning of SGNs in the cochlear epithelium.

35

Figure 10
A

h
E18
TUJ1 TUJ1/MYO6 TUJ1/MYO6
Slit1-/-
E18
TUJ1 TUJ1/MYO6 TUJ1/MYO6
Slit2 -/-
SGN
SGN
E18
SGN
TUJ1 TUJ1/MYO6 TUJ1/MYO6
HC
HC
SGN
SGN
SGN
HC
WT
HC
HC
HC
RF
RF
RF
E18
SGN
TUJ1 TUJ1/MYO6 TUJ1/MYO6
HC
SGN
Slit3 -/-
HC
RF

36

Figure 10 Continued

Figure 10: Spiral ganglion neurons are mislocated in Slit2 mutant cochlea. A:
Representative images of SGN (Green, Tuj1 immunostaining) innervation of HCs (red, MyoVI
immunostaining) in the whole mount cochlea at E18 of wildtype (top row), Slit1-/-(second row), Slit2-/-
(third row) and Slit3-/-(fourth row) embryos. A significant number of SGNs are mispositioned in the
cochlear epithelium in the Slit2-/- cochlea (pointed by arrows). Scale bar: 50µ m. B: Representative images
of SGN fibers (Green) and HCs (red) at E18. Scale bar: 25µ m. C: Average number of mSGN per cochlea
of wildtype and Slit mutant mice at E18. Bar=SD. Number of embryos examined N=6 for all genotypes.
D: Average fasciculation Index (quantified as the ratio of summed thickness of radial fiber bundles over the
total width along the white dotted line shown in A (last column) at E18. Bar=SD. Number of embryos
examined N=6 for all genotypes.

TUJ1/MYO6
WT
TUJ1/MYO6
MYO6
MYO6 TUJ1
TUJ1
Type I
Type II
Type I
Type II
Slit2 -/-
IHC
IHC
OHC
OHC
B
F a s ic u la tio n In d e x
W T s lit1 -/- S lit2 -/- S lit3 -/-
0 .0
0 .2
0 .4
0 .6
0 .8
C
m S G N n u m b e r
0
1 0
2 0
3 0
4 0
5 0
W T S lit1 -/- S lit2 -/- S lit3 -/-
D

37

Since Slit molecules have been shown to regulate axon fasciculation (Jaworski and
Tessier-Lavigne 2012), we also quantified the fasciculation of radial fibers of SGNs but
did not find significant differences between wild type and any of the Slit mutants (Figure
10A, the last column and Figure 10 D). These data suggest that Slit is not involved in the
regulation of fasciculation of SGN axons.
1.3.4 SGNs are progressively relocated to the cochlear epithelium
from their normal position in the Slit2 mutant

The mispositioning of SGNs in the Slit2 mutant cochlea at E18 could be explained by
two possible mechanisms: first, those SGNs might fail to delaminate from the otocyst
early in development; alternatively, all SGNs might migrate out successfully from the
otocyst but some segregate from the large SGN population located in the Rosenthal’
canal and move towards the cochlear epithelium during the time of SGN innervating
HCs. If the first case is true, we should expect a relatively constant number of
mispositioned SGNs across different developmental stages in the Slit2 mutant.
Conversely, if the second case is true, we would expect to see progressively more and
more SGNs accumulating in the cochlear epithelium.
To address these possibilities, we examined SGN patterning at different developmental
stages in the Slit2 mutant. In wild type embryos at E13, SGNs had migrated out from the
otocyst, settled down in Rosenthal’s canal and began to extend peripheral neurites
towards the sensory epithelium (Figure 11 A and B). No mispositioned SGN was
observed in the Slit2 mutant at this stage (Figure 11 C and D). However, one day later at
E14, a small number of SGNs were found scattered in the cochlear epithelium (Figure 11

38

G and H, pointed by red arrows) but not in their wild type littermates (Figure 11 E and F).
Some of them had already extended neurites of substantial lengths while some others just
started to grow neurites (Figure 11 H, red arrows). At E16, the peripheral neurites of
SGNs in wild type embryos formed connections with inner HCs and started to explore
outer HCs, with their cell bodies restricted within the Rosenthal’s canal (Figure 11 I-K).
In the Slit2 mutant, a number of SGNs were mislocated in the cochlear epithelium similar
as at E14, but the number was significantly larger (Figure 11 L-N, pointed by red
arrows). As shown in the summary, the average number of mispositioned SGNs per
cochlea in the Slit2-/- mutant steadily increased from E14 to E16 (Figure 11O). These
observations collectively suggest that in the Slit2 mutant, SGNs are dispersed from their
normal location to the cochlea epithelium in a progressive manner starting from as early
as E14. Based on the fact that Slit2 is strongly expressed in the SL and GER regions
separating the spiral ganglion from the cochlear epithelium, our data suggest that Slit2
molecules provide a restriction force to prevent SGN cell bodies from leaving the
Rosenthal’s canal while allowing their neurites to extend into the cochlear epithelium.

39

Figure 11: Developmental changes of SGN positioning in the Slit2 mutant cochlea. A-
D: Representative images of SGN (Green, Tuj1 immunostaining) positioning in the
whole mount cochlea from wildtype (A, B, E, F and I-K) and Slit2-/-(C, D, G, H and L-
M) mouse embryos at E13 (A-D), E14 (E-H) and E16(I-N), respectively. White dotted
lines mark the lateral boundary of the SGN cluster and the cochlear lateral wall (LW) in
A-H. M is the maximum intensity projection from a subset images from the Slit2-/-
cochlea to show the mispositioned SGN while others are the projections from the
complete set of image series. Red arrows mark the mispositioned SGN in H, M and N.
Scale bar: 50µ m. O: Average number of mSGNs per cochlea in Slit2 mutant and their
wildtype littermates at different developmental stages. Number of embryos examined N
=4, 5, 5, 6, 4, 4, 6, 6 for the genotypes listed from left to right respectively. Bar=SD, **:
p< 0.01, one-way ANOVA with Tukey's multiple comparisons test.
m S G N n u m b e r
0
2 0
4 0
6 0
8 0
E 1 3 E 1 4 E 1 6 E 1 8 E 1 3 E 1 4 E 1 6 E 1 8
W T S lit2 -/-
**
**
**
**
**
Slit2-/- WT
TUJ1/MYO6 TUJ1 TUJ1/MYO6
E16
E16
SGN
SGN
SGN SGN
SGN SGN
HC
HC
HC
HC
RF
RF
Slit2 -/- WT
E13
E13
TUJ1
TUJ1
TUJ1
TUJ1
SGN
SGN
LW
LW
SGN
E13
SGN
E13
Slit2 -/- WT
E14
E14
TUJ1
TUJ1
TUJ1
TUJ1
SGN
SGN
SGN
SGN
LW
LW
A
O
B
C D
E F
G H
I J
L M
K
N

40

1.4 The role of Robo family molecules in shaping peripheral auditory
circuitry assembly

1.4.1 Robo is expressed in SGNs as well as supporting cells in the
organ of Corti

We next asked the question of whether the effect of Slit2 on SGN soma position is
mediated through its receptor, Robo. First, we sought to examine whether Robo is
expressed in SGNs during the time they innervate hair cells. We performed in situ
hybridization for three Robo molecules in the whole mount cochlea, and found that both
Robo1 and Robo2 were strongly expressed in SGNs as well as the supporting cells
around the organ of Corti at E16 and E18 (Figure 12). We did not detect significant
Robo3 expression in the cochlea (data not shown). These expression data suggest that
SGNs might depend on Robo1 and Robo2 to interpret the signals of Slit2 secreted from
the spiral limbus/GER region. If that is the case, the mutants of Robo1 and 2 should
recapitulate the phenotype of the Slit2 mutant.

41

.

1.4.2 SGNs are mispositioned as individuals in the Robo1/Robo2
double mutant at E18

To address whether Robo1 and Robo2 indeed mediate the Slit2 function, we examined
the SGN-HC innervation pattern in Robo mutant embryos at E18. A small number of
SGNs were found in the cochlear epithelium in the Robo1-/-; Robo2-/- double mutant
(Figure 13 D-H). This defect is similar to that found in the Slit2 mutant, but the number
of identifiable mispositioned cells was much smaller (Figure 13 I). In contrast, no
E16
Robo1 Robo2
P0
A B
c
GER
SPG
OC
LER
SL
C
D E F
Figure 12: Robo are expressed in Cochlear tissue during SGN innervating HC.
In situ hybridization of Robo 1(A-C) and Robo2 (D-F) in whole mount cochlea at E16
and P0. Both of them are expressed in spiral ganglion as well as the supporting cells in
the organ of Corti. LER: lesser epithelial ridge (or outer sulcus), OC: organ of Corti,
GER: greater epithelial ridge (or Kolliker’s organ, SL: spiral limbus, SPG: spiral
ganglion neuron.

42

0
2
4
6
8
m S G N N u m b e r
R o b o 1 + /-;
R o b o 2 + /-
R o b o 1 -/-;
R o b o 2 -/-
Robo1 Robo2
c
GER SGN
OC
LER SL
TUJ1/MYO6
Robo1-/-; Robo2 -/-
E18
E18
TUJ1/MYO6
MYO6
MYO6 TUJ1
TUJ1
Type I
Type II
Type I
Type II
Robo1+/-; Robo2 +/-
IHC
OHC
IHC
OHC
TUJ1
TUJ1
I
A B C D
E F G H
J
K
L
M
0
2
4
6
8
m S G N N u m b e r
R o b o 1 + /-;
R o b o 2 + /-
R o b o 1 -/-;
R o b o 2 -/-
Figure 13: Individual SGN are mislocated in Cochlea epithelium in
Robo mutant. A-H: Representative projection images from confocal Z
stack of SGN (Green, Tuj1 immunostaining) innervation of HCs (red,
MyoVI immunostaining) in the whole mount cochlea from Robo1+/-;
Robo2+/- (A-D) and Robo1-/-; Robo2-/- (E-H) embryos. D and H are
independent examples from A-C and E-G, respectively. H is the
maximum intensity projection from a subset images from the Robo1-/-;
Robo2-/- to show the mispositioned SGN while A-G are maximum
intensity projections from the complete image series. Yellow arrows
mark the individually mispositioned SGN in G and H. Scale bar: 25µ m
in A-C and E-G; 50 µ m in D and H. I: Average number of mispositioned
SGN (mSGN) per cochlea at E18 in Robo1+/-; Robo2+/- and Robo1-/-;
Robo2-/- embroys. Number of embryos examined N = 4 for both
genotypes.
I

43

mispositioned SGN was observed in the Robo1+/-; Robo2+/- littermates (Figure 13 A-
D). Also similar as in the Slit2 mutant, the neurites of the mispositioned SGNs were
misguided and did not innervate HCs (Figure 13 E-H). The similarity of the defects
found in Slit2 and Robo1/2 mutants suggests that Slit2 indeed interacts with Robo1 and
Robo2 to ensure the correct positioning of SGN cell bodies.
Since Robos are also expressed in the supporting cells around the organ of Corti, we
examined the structural arrangement of the cochlear epithelium using phalloidin staining.
However, we did not observe any significant alterations of the overall cochlear structure
in Robo1/2 double mutants compared with their heterozygous littermates (Figure 14).

OC
LER
GER
SL
OC
LER
GER
SL
Robo+/-; Robo2+/- Robo-/-; Robo2-/-
Phalloidin Alexa488 Phalloidin Alexa488
Figure 14: Overall arrangement of cochlea sensory epithelium is not altered in
Robo mutant. Representative images of whole mount cochlea sensory epithelium
from Robo1-/-; Robo2-/- embryos(Left panel) compared with their Robo1+/-;
Robo2+/- littermates(right panel) at E18 as illustrated by Phalloidin staining (Green).
LER: lesser epithelial ridge (or outer sulcus), OC: organ of Corti, GER: greater
epithelial ridge (or Kolliker’s organ), SL: spiral limbus,

44

1.4.3 Expansion of SGN boundary in the Robo1/Robo2 double
mutant

In the Slit2 mutant, most SGNs are in their normal position despite a small number
scattered in the cochlear epithelium. In sharp contrast, in the Robo1-/-; Robo2-/- double
mutant at E18, besides individual mispositioned SGNs, the large population of SGNs
were not restricted within the Rosenthal’s canal any more but shifted towards HCs as an
entirety (Fig. 15C, 15D,15G,15H). This was not observed in the Robo1+/-; Robo2+/-
littermates (Fig. 15A, 15B, 15E, 15F). We measured the radial distance between the
lateral boundary of the SGN lamina and inner HCs (SGN-IHC Distance, shown as red
lines with double arrow head in Fig. 15B, 15D,15F,15H) along the basal-apical axis of
the entire cochlea (Fig. 15Y). In the middle region of the cochlea, the SGN-IHC distance
was significantly reduced in Robo1-/-; Robo2-/- embryos compared with Robo1+/-;
Robo2+/- or wild type littermates (Fig. 15Z), while the total radial distance from the
medial boundary of the SGN lamina to HCs remained largely unchanged (data not
shown). As a result, the SGN width (blue line with double arrowhead in Fig. 15B, 15D)
was significantly increased in Robo1-/-; Robo2-/- embryos in comparison with Robo1+/-;
Robo2+/- or wild type littermates (Fig. 15A’). There were numerous crossing fibers
running in the longitudinal direction without innervating HCs in the Robo1-/-; Robo2-/-
embryos but not in their Robo1+/-; Robo2+/- littermates (Fig. 15C, 15D).
In an independent experiment to verify Robo mutant phenotypes, we set to
genetically label SGNs on the Robo mutant background by crossing mouse lines carrying
the following alleles: Neurogenin1-CreER (Koundakjian, Appler et al. 2007), Ai14
(Madisen, Zwingman et al. 2010) and Robo1+/-; Robo2+/-. In embryos carrying all four

45

Figure 15

Base
Apex
IHC
Robo1 -/- ; Robo2 -/- Robo1 +/-; Robo2 +/-
e
f
TUJ1/MYO6
TUJ1/PV
E16
Base
Apex
IHC
E18
IHC
IHC
TUJ1/MYO6
TUJ1/MYO6 TUJ1/MYO6
E18
E18 E18
Robo1
+/-
; Robo2
+/-
Robo1-/-; Robo2-/-
E18 E18
E18 E18
tdTomato tdTomato
tdTomato tdTomato
Robo1 +/-; Robo2 +/- Robo1 -/-; Robo2 -/-
SGN SGN
LW
LW
SGN
SGN
TUJ1 TUJ1
TUJ1 TUJ1
E14 E14
E14 E14
A B C
D
E F G H
I J
L K
M N
P O

46

Figure 15 Continued

S G N W id th (  m )
0
1 0 0
2 0 0
3 0 0
W T
E 1 6
R o b o 1 + /-;
R o b o 2 + /-
E 1 6
R o b o 1 -/-;
R o b o 2 -/-
E 1 6
R o b o 1 + /-;
R o b o 2 + /-
E 1 8
R o b o 1 -/-;
R o b o 2 -/-
E 1 8
W T
E 1 8
W T
E 1 4
R o b o 1 + /-;
R o b o 2 + /-
E 1 4
R o b o 1 -/-;
R o b o 2 -/-
E 1 4
**
**
**
**
S G N D e n s ity
0
1 0
2 0
3 0
W T R o b o 1 + /-;
R o b o 2 + /-
R o b o 1 -/-;
R o b o 2 -/-
*
*
SGN-IHC
Distance(µ m)
Robo KO E16
Robo HET E18
Robo HET E16
Robo KO E18
S G N -IH C /L W
D is ta n c e (  m )
0
5 0
1 0 0
1 5 0
2 0 0
2 5 0
W T
E 1 8
R o b o 1 + /-
R o b o 2 + /-
E 1 8
R o b o 1 -/-
R o b o 2 -/-
E 1 8
W T
E 1 6
R o b o 1 + /-
R o b o 2 + /-
E 1 6
R o b o 1 -/-
R o b o 2 -/-
E 1 6
W T
E 1 4
R o b o 1 + /-
R o b o 2 + /-
E 1 4
R o b o 1 -/-
R o b o 2 -/-
E 1 4
**
**
**
**
SGN-IHC/LW
Distance(µ m)
Base
Apex
IHC
SGN
Robo1 -/- ; Robo2 -/- Robo1 +/-; Robo2 +/-
Base
Apex
IHC
SGN
IHC
SGN
IHC
SGN
TUJ1/MYO6
TUJ1/MYO6
TUJ1/MYO6
TUJ1/MYO6
E16 E16
E16 E16
SGN
SGN
SGN
SGN
IHC
IHC
IHC
IHC
Q R S T
U V W X
Y
Z
A’ B’

47

Figure 15: Spiral ganglion territory is progressively expanded towards HCs in the
Robo mutant. Representative images of SGN (Green, Tuj1 immunostaining)
innervation of HCs (red, MyoVI immunostaining) in the whole mount cochlea from
Robo1+/-; Robo2+/- and Robo1-/-; Robo2-/- mice at E18 (A-H), E14 (M-P) and E16
(Q-X), respectively. The red lines with double arrow heads illustrate the distance
between the lateral SGN boundary and inner hair cells (SGN-IHC distance) in B, D, F,
H, R, T, V and X or the distance between the lateral SGN boundary and the lateral wall
(SGN-LW distance) in M and N, respectively. The blue line with double arrow heads
illustrate the SGN width in B, D, F, H, R, T, V and X. I-L: Representative images of
sparsely labeled SGN (white signal, tdTomato expression) distribution in the cochlea at
E18 from Robo1+/-; Robo2+/- (I and K) and Robo1-/-; Robo2-/- (J and L) mice which
also carry Neurogenin1-CreERT2 and Ai14 alleles. Y: Plot of SGN-IHC distance along
the base-apex axis for Robo1-/-; Robo2-/- mutants at E16 (red line) and E18 (green line)
as well as for their Robo1+/-; Robo2+/- littermates (blue line for E16 and purple line for
E18, respectively). Z: Average SGN-LW distance (for E14) or SGN-IHC distance (for
E16 and E18) in the middle part of the cochlea for wildtype and Robo1/2 mutant
embryos. Number of embryos examined N=5 for all genotypes at E14, N=6 for all
genotypes at E16 and N=4 for all genotypes at and 18. Bar=SD. **: p <0.01 one-way
ANOVA with Tukey's multiple comparisons test. A': Average SGN width in the middle
part of the cochlea. Number of embryos examined N=5 for all genotypes at E14, N=6 for
all genotypes at E16 and N=4 for all genotypes at E18. Bar=SD. **: p <0.01 one-way
ANOVA with Tukey's multiple comparisons test. B': Average SGN density as
quantified by the number of SGNs per 10000 square micrometers for in wildtype and
Robo mutant embryos at E18. Number of embryos examined N=4 for all genotypes.
Bar=SD. *: p <0.05 one-way ANOVA with Tukey's multiple comparison test. Scale
bar: 50µ m in A-P; 100 µ m in Q-X

48

alleles, SGNs were labelled sparsely by tdTomato expression due to the leaky activity of
Neurogenin1-CreER (Fig. 15 I-L). While the sparsely labelled SGNs were tightly packed
in the central cochlea in Robo1+/-; Robo2+/- embryos (Fig. 15 I, 15K), they were much
more dispersed towards HCs in the Robo1-/-; Robo2-/- mutant (Fig. 15J, 15L), consistent
with an extension of the lateral SGN boundary towards HCs.

1.4.4 Developmental progression of SGN territory expansion in
Robo mutants

We also performed developmental studies in Robo mutants to determine when the
expansion occurs in Robo mutants. At E14, SGNs were normally located in the central
cochlea in both Robo1+/-; Robo2+/- and Robo1-/-; Robo2-/- embryos (Fig. 15M-P).
There was no significant difference in SGN-LW distance (the radial distance from the
lateral SGN boundary to the lateral wall blue line with double arrowhead in Fig. 15M,
15N) or SGN width (red line with double arrowhead in Fig. 15M, 15N) between wild
type, Robo1+/-; Robo2+/- and Robo1-/-; Robo2-/- embryos (Fig. 15Z, 15A’, respectively
). These results suggest that the initial migration of SGNs from the otocyst and their
settlement in the Rosenthal’ canal is not altered in the Robo mutant. At E16, however,
the SGN-IHC distance was significantly shorter in Robo1-/-; Robo2-/- mutants in
comparison with Robo1+/-; Robo2+/- littermates (Fig. 15R-X, 15Y,15Z). At the same
time, the SGN width became broader in the Robo mutants in comparison with their
heterozygous littermates (Fig. 15R-X, 15A’). Consistent with the notion of an expansion
of SGN territory, we found that at E18 the SGN density was significantly reduced in

49

Robo1-/-; Robo2-/- embryos in comparison with their Robo1+/-; Robo2+/- or wild type
littermates (Fig. 15B’). Taken together, these data indicate that, in Robo mutants, the
entire SGN territory is expanded progressively towards HCs after the initial migration
and settlement of SGNs is accomplished, similar to the developmental progression of the
defect found in the Slit2 mutant.

50

1.5 Discussion and Perspectives

Slit-Robo interactions in spatial restriction of spiral ganglion neurons

The assembly of cochlear circuits follows discrete steps during embryonic development.
The specific pattern of cochlear innervations provides a nice model to examine molecular
mechanisms underlying the local organization of different cell populations and their
interactions. Using this model, our study provides the first molecular cues that are
critical in patterning the synaptic connection at this stage. Our genetic analysis has
revealed a novel functional role of Slit-Robo signalling: to restrict spatial positioning of
periphery neurons, i.e. to stabilize the position of SGNs in the Rosenthal’s canal and
prevent them from migrating out abnormally during the assembly of peripheral auditory
circuits. Previously it was reported that in the ErbB2 mutant mice SGNs migrate beyond
their normal position and settle down abnormally in the modiolus (Morris, Maklad et al.
2006). Here we found that after the initial migration is accomplished, SGNs are
progressively dispersed towards HCs as an entirety in Robo mutants and to a lesser
degree are dispersed as individual cells in Slit2 mutants (depicted in Fig. 16). It is likely
that Slit/Robo and ErbB2 signaling prevents SGNs from escaping into the cochlear
epithelium and modiolus respectively, thus ensuring their precise spatial patterning.
Indicated by its expression pattern, Slit2 is likely secreted from the spiral limbus and
GER regions, setting up the lateral boundary of the SGN lamina. Since Slit is not the
only ligand for Robo (Ypsilanti, Zagar et al. 2010), the massive expansion of SGN
territory in Robo mutants but not in Slit2 mutants (Fig. 16) suggests that other Robo
ligands might act synergistically with Slit2 to impose the restriction force as to maintain

51

OC
SGN
GER
SL
LER
Mesenchyme
OC
SGN
GER
SL
LER
Mesenchyme
~E10
E16/E18
WT
Slit2 Mutant
Robo Mutant
E13.5
E14
L M
V
D
SGN
IHC
OHC
SL
GER
SGN
IHC
OHC
SL
GER
SGN
IHC
OHC
GER
SGN
Otocyst OC
SGN
GER
SL
LER
Mesenchyme
OC
SGN
GER
SL
LER
Mesenchyme
OC
SGN
GER
SL
LER
IHC
Mesenchyme Rosenthal’ Canal
OC
GER
SL
LER
SGN
Mesenchyme
Figure 16: A proposed model for Slit/Robo signaling in restricting SGN
positioning. Early in development (around E10), SGNs delaminate from the otocyst,
migrate towards the modiolus and finally settled down in the Rosenthal’ canal at
around E13.5. Subsequently, SGNs extend their peripheral axons towards the OC at
E14 and begin to form synaptic connections with HCs at E16. The SGN somas are
restrained within the Rosenthal’ canal during the formation of SGN-HC innervations
in the wildtype cochlea. Slit2 (and possibly Slit3) secreted from the SL and GER
region act on Robo receptors expressed in SGNs and provides the restriction force to
refrain SGNs from invading the cochlear epithelium. In the Slit2 mutant cochlea, a
number of SGNs disperse from the Rosenthal’s canal and move into the cochlear
epithelium progressively. Their neurites travel largely along the longitudinal direction
and fail to innervate HCs. In the Robo mutant, a small number of SGNs are scattered
in the cochlear epithelium similar as in the Slit2 mutant. In addition, the entire SGN
territory expands progressively towards HCs starting from E14, resulting in a shorter
SGN-HC distance as well as a broader SGN width. The more severe defect in the
Robo mutant compared with the Slit2 mutant suggests additional factors acting
synergistically with Slit2 on Robo receptors to restrict SGN soma within the
Rosenthal’s canal, which is essential for the formation of precise patterning of SGN-
HC connections

52

an intact SGN lamina. Thus, by restricting the SGN position, Slit/Robo signalling
maintains a clear spatial segregation between SGNs and their target tissue, the cochlear
sensory epithelium. This separation ensures the compartmentalization of distinct cell
populations and prevents them from intermingling into a scrambled organization in the
cochlea. Such molecular function is reminiscent of the restriction of cell intermingling
between hindbrain segments by the interaction of Eph receptors and their Ephrin ligands
(Mellitzer, Xu et al. 1999; Xu, Mellitzer et al. 2000).

The relationship between SGN soma positioning and neurite path
finding

Axon guidance molecules are important in determining neuronal innervation patterns in
both peripheral and central nervous systems (Tessier-Lavigne 2002; Yamamoto, Tamada
et al. 2002; Huberman, Feller et al. 2008; Salinas and Zou 2008; Cho, Prince et al. 2009;
Shen and Cowan 2010). Previous studies have shown specific expression of axon
guidance molecules in the developing inner ear (Webber and Raz 2006; Fekete and
Campero 2007; Appler and Goodrich 2011), but their potential roles in the development
of spatial organization of cochlear cells have not been examined. Of the guidance
molecules, Slits are of particular interest as they have been shown to play essential roles
in various processes such as axon guidance, branching, fasciculation as well as cell
migration through Roundabout (Robo) (Brose, Bland et al. 1999; Kidd, Bland et al. 1999;
Wang, Brose et al. 1999; Ma and Tessier-Lavigne 2007; Ypsilanti, Zagar et al. 2010).
Recently, it has also been implicated in controlling lamina-specific connections in the
optic tectum (Xiao, Staub et al. 2011). Their function in the inner ear development and

53

patterning is not known. To our knowledge, this is the first demonstration that Slit/Robo
signalling plays an essential role in shaping the spatial patterning of distinct cell
populations in the nervous system. The correct assembly and positioning of SGN cell
bodies appears critical for their afferent axons to navigate to HCs without errors. As
shown in both Slit2 and Robo1 and 2 mutants (Fig. 10,11 and 15), the neurites of the
individually mispositioned neurons travel randomly within the cochlear epithelium and
never innervate HCs. On the other hand, the rest of SGNs whose cell bodies stay in the
right location form radial fibers that normally and correctly innervate HCs. One possible
explanation for this difference is that the guidance cues for neurite pathfinding are not
present in the ectopic location where SGNs are misplaced, resulting in the misrouting of
their fibers. Our results also indicate that Slit/Robo signalling in the developing cochlea
is not primarily involved in guiding SGN axons to innervate HCs, as no aberrant axon
was observed for the entire population of properly positioned SGNs in either mutant.
Taking together, the spatially precise positioning of SGN cell bodies by Slit/Robo
interactions may provide an essential control for the precise formation of functional
cochlear innervations.
The function of Slit1 and additional roles of Robo in the cochlea
development

Since Slit1 is expressed in SGNs, it is possible for Slit1 to regulate SGN development in
a cell-autonomous manner and possibly the axon innervation of hair cells via
autocrine/paracrine mechanisms. However, we did not observe any obvious defects in
Slit1 mutants compared with wild type animals, possibly due to the redundancy of Slit1.
In addition to SGNs, Robo is also strongly expressed by the supporting cells around the

54

organ of Corti. It might define the boundary between sensory and nonsensory regions
within the cochlea through homophilic interactions as previously suggested (Battisti and
Fekete 2008). However, we did not observe any obvious defect of cell arrangement
patterns within the organ of Corti. The role of Robos in the cochlea development needs to
be further studied. All the mutants analyzed in this study are lethal at perinatal stages. In
order to assess the physiological impact of Slit/Robo signalling on the auditory function,
conditional knockout animals are needed to maintain the survival of animals to much
later stages.

Intracellular mechanisms underlying the restriction of SGN by Slit/
Robo signaling

Several GTPase-activating proteins (GAPs), e.g. Slit-RoboGAPs (srGAPs) and
CrossGAP/Vilse, have been identified to act downstream of Robo receptors to regulate
the activity of small GTPases of the Rho family, such as RhoA, Rac1 and Cdc42 and
lead to cytoskeleton rearrangement in the presence of Slit binding (Wong, Ren et al.
2001; Lundstrom, Gallio et al. 2004; Hu, Li et al. 2005; Ypsilanti, Zagar et al. 2010). It
remained to be further tested whether any of these molecules are involved in the role of
Slit/Robo interaction revealed in this study.

55

1.6 Experimental Procedures and Materials

Mouse Strains

Mice were handled according to the protocols approved by the Institutional Animal Care
and Use Committees at University of Southern California following the National
Institutes of Health regulations. All animals were generated in a mixed CD-
1/129Sv/C57BL/6 background. To generate Slit mutants, we crossed Slit1 −/−; Slit2 +/−,
Slit2 +/−, Slit3 +/− animals and obtained the desired single or double mutant. Robo
mutants were obtained by crossing double heterozygous Robo1;Robo2 animals, which
have the two mutant alleles (1.8 megabases apart) already linked to the same
chromosome 16(Ma and Tessier-Lavigne 2007). Genotyping of Slit1, Slit2, Slit3, Robo1,
and Robo2 was done by PCR as described previously (Plump, Erskine et al. 2002; Yuan,
Rao et al. 2003; Grieshammer, Le et al. 2004; Sabatier, Plump et al. 2004). Genotyping
of Neurogenin1-cre/ERT2 and Ai14 (ROSA-LSL-tdTomato) were done by PCR
following previously described (Koundakjian, Appler et al. 2007; Madisen, Zwingman et
al. 2010).

Tissue dissection, FM1-43 staining and FACS sorting

Cochlea epithelium was dissected from mouse pups between postnatal day 5(P5) to
postnatal day 7 (P7) as previously described (Doetzlhofer, White et al. 2006; Lelli, Asai
et al. 2009). For FM1-43 dye staining, whole mount cochlea from wildtype mouse were
bathed with 5 μM FM1-43 (Invitrogen) for 30 s followed by an extensive wash with PBS.

56

Spiral ganglion was trimmed away from both FM1-43 stained wildtype cochlea and
parvalbumin-Cre; Ai14(ROSA-LSL-tdtomato)cochlea. Tissues from both sources are
then treated with activated papain (Worthington, 20 units/ml plus one millimeter L-
cysteine) for 20 minutes followed by 2 minutes crude trypsin (0.5 mg/ml, Sigma)
treatment. DMEM (invitrogen) plus 10%FBS were added and tissue were triturated to
achieve fully dissociation. FACS sorting was performed at Flow Cytometry Core Facility
at USC. Cell suspensions were fed into BDAriaII sorter and purified using 488nm laser
excitation and 100-μm CytoNozzle. Distinct cell populations were collected into DMEM
plus 10%FBS and pelleted down through centrifuging.

RNA amplification and microarray data analysis

Cells from different experiments were pooled. RNA was extracted using the PicoPure
RNA isolation kit (Arcturus) from distinct purified cell populations. Three independent
pools of RNA from each population are amplified using WT-Ovation Pico amplification
kit (Nugen) and labeled for microarray experiment carried out at CHLA genomic core.
Samples were profiled on GeneChip Mouse Genome 430 2.0 Array (Affymetrix).
Microarray data analysis was conducted within R environment. The raw expression data
was normalized with MAS5.0 algorithm and filtered before logarithm transformation
(Bolstad, Irizarry et al. 2003). Pair wise comparison for each probe set between the two
groups was performed by Empirical Bayes method, and differentially expressed probe
sets was identified as having an absolute signal log ratio ≥1.0, and a FDR value ≤

57

5%(Smyth 2004). Gene annotation and ontology information was obtained from the
National Center for Biotechnology Information, NetAffx, the Gene Ontology
Consortium, and the Kyoto Encyclopedia of Genes and Genomes. Significant
enrichment of specific gene sets was assessed with gene set analysis method(Efron and
Tibshirani 2007). Gene set information was obtained from Gene Set Enrichment
Analysis website(Subramanian, Tamayo et al. 2005).
Immunohistochemistry and in situ hybridization

Whole mount cochlea were dissected from embryos at particular stage using timely
pregnant females and fixed with 4% PFA overnight. The cochleae was first
permeabilized with 0.5% Triton X-100 for 10 minutes, blocked with 10% serum plus
3%BSA in PBS for 2 hours at room temperature and then incubated with the first
antibody overnight at 4° C. After washout, Alexa conjugated secondary antibodies
(Invitrogen, 1:800) were added for 2 hours at room temperature. Confocal z stack images
were obtained using Fluoview1000 (Olympus), projected using NIH-ImageJ, and then
further processed using Inkscape. Antibodies used in this study and their dilution:
Alexa488 conjugated Tuj1 mouse antibody (Covance, 1:300 dilution), MyoVI (rabbit
polyclonal antibody, Millipore, 1:600 dilution), Alexa546 conjugated goat Anti–Rabbit
IgG antibodies (Invitrogen, 1:800 dilution).
In situ hybridization was performed on whole mount cochlea using digoxigenin (DIG)-
labeled cRNA probes following previously described (Zine, Aubert et al. 2001). Probes
were generated using the templates previously published: Slit1, Slit2, Slit3, Robo1,
Robo2, and Rig1 (Brose, Bland et al. 1999).

58

Chapter 2: Development of STARS technique for lineage
tracing in the inner ear

In the cochlear sensory epithelium, hair cells are surrounded by various types of
supporting cells forming a checkerboard organization. Despite extensive studies on the
cell fate specification of hair cells and supporting cells in the cochlear sensory
epithelium, their detailed lineage relationship remains largely elusive(Fekete,
Muthukumar et al. 1998; Kelley 2006). This might be in large part due to the fact that the
inner ear is deeply embedded in the temporal bone which makes the traditional tracing
methods such as retrovirus injection difficult to perform. The quest to probe this question
calls for a straightforward genetic approach that can label a small subset of progenitors at
an early developmental stage. If the genetic labeling can reach a low enough sparseness
level, cells come from the same progenitor will form a cluster which will be spatially
separated from the cells originated from the other progenitor. In this way, the cell lineage
relationship can be dissected.
To address this question, in this chapter, we developed a method termed STARS:
stochastic gene activation with genetically regulated sparseness. The stochastic
expression was achieved by two cross-linked, mutually-exclusive Cre-mediated
recombinations. The stochastic level was further controlled by regulating Cre/lox reaction
kinetics through varying the intrachromosomal distance between the lox sites mediating
one of the recombinations. We further explore the possibility to extend the sparseness
level by mutagenizing lox sites employed in the STARS transgene. In mammalian cell
lines stably transfected with a single copy of different STARS transgenes, the
activation/knockout of reporter genes was specifically controlled to occur in from 5% to

59

50% of the cell population and further down to 1% when combined with lox variants with
lower recombination efficiency. STARS can potentially provide a convenient way for
genetic labeling as well as gene expression/knockout in a population of cells with a
desired sparseness level.

60

2.1 Introduction to Single Cell Labeling in the nervous system

Cre/lox recombination system provides a powerful tool for conditional gene
expression control. Cre recombinase derived from bacteriophage P1 is a member of λ
integrase superfamily and specifically recognizes a pair of lox site which has two 14 base
pair inverted repeat separated by an 8 basepair asymmetric spacer, the later dictating the
direction of the lox site (Sauer 1993). Depending the location and orientation of the lox
site, Cre can efficiently catalyze excision, inversion, translocation and integration event
and has been widely applied in genetic analysis (Lewandoski 2001; Schnutgen,
Doerflinger et al. 2003; Luo 2007; Luo, Callaway et al. 2008). To achieve conditional
gene expression control, people flanked an essential part of the gene with a lox pair to
generate the conditional allele without perturbing the gene function without
recombination. Once the mice are crossed to a tissue or cell type specific Cre mouse line,
recombination will excise the floxed sequence and lead to the allele inactivation with
spatial and temporal accuracy dictated by the promoter drives Cre expression
(Lewandoski 2001). Alternatively, one can put a floxed “STOP” element (usually a
multi-polyA sequence) between the promoter and the coding sequence of a gene so that
the gene expression is inactive unless it is crossed to a specific Cre line to kick out the
STOP signal (Dragatsis, Levine et al. 2000).
Current approaches have their limitations to achieve genetic “Golgi” staining as
well as genetic manipulation in a small population of cells in nervous tissue in an
efficient way. To image and dissect the complex neural circuitry in vivo, an initial step
would be to label individual neurons and then figure out how they connect to other
neurons. Several genetic methods have been developed for achieving gene

61

expression/knockout in a small population of cells: 1) screening transgenic lines with
variegated gene expression (Feng, Mellor et al. 2000; Xiao, Roeser et al. 2005; Young,
Qiu et al. 2008); 2) Cre/CreER-lox mediated conditional intrachromosomal
recombination (Hayashi and McMahon 2002; Badea, Wang et al. 2003; Kuhlman and
Huang 2008); and 3) Cre-lox mediated interchromosomal recombination during mitosis
(Liu, Jenkins et al. 2002; Zong, Espinosa et al. 2005; Muzumdar, Luo et al. 2007; Wang,
Warren et al. 2007) (e.g. Mosaic analysis with double markers in mice, MADM). The
application of variegation effect depends on the screening of transgenic lines with the
desired expression pattern, which is not predictable and labor intensive. The method of
CreER will require careful titration of the dose of tamoxifen for every specific tissue
(Hayashi and McMahon 2002; Badea, Wang et al. 2003). As for MADM, the
requirement for mitotic recombination, although will benefit the tracing of cell lineage,
may not be efficient for marking differentiated cells. It also depends on the establishment
of MADM marker on the specific arm of each chromosome. In all these methods, a
straightforward genetic control of the percentage of cells with expression/knockout of the
gene of interest cannot be readily achieved. Recently, a Brainbow transgenic strategy has
been developed (Livet, Weissman et al. 2007), in which Cre/lox recombination was
utilized to create a stochastic choice of expression among three different fluorescence
proteins. Cells could be labeled with as many as tens of different colors according to the
combination of differential doses of fluorescence proteins expressed in each cell. This
method enables imaging a large number of individual cells in the same circuit but is
limited by the technical challenges to resolve so many colors. The uniformity of different
fluorescent proteins throughout the entire cell volume could be a concern, especially in

62

nervous tissues with dauntingly complex process, which might result in different hues at
different locations from the same cell. It is also difficult to couple genetic manipulations
with labeling using this method.
STARS can provide a straightforward way to achieve genetic “Golgi” staining as
well as genetic manipulations in the same cells labeled. Using STARS strategy to
express a maker fluorescent protein (e.g. membrane bound EGFP) with low levels of
sparseness, genetic “Golgi” staining will be achieved when the transgenic mice are
crossed to tissue or cell type specific Cre lines. In addition, the sparseness level can be
further controlled based on the examination in HEK293 cells or ES cells and might not be
influenced by the locus in the genome STARS targeted. The latter is because it actually
depends on the competition between two recombination cassettes which is harbored in
the same STARS trangene and targeted to the same genomic locus no matter where it is,
which has nothing to do with Cre’s efficiency for that particular locus. Furthermore, if an
exogenous gene (e.g. dominant negative form of a gene of interest) coupled with IRES
element is expressed in the same way, genetic mosaics could be generated allowing
visualizing the mutant cells along with their wild type siblings. This will permit the study
of molecular and cellular mechanisms underlying neuron circuit development, tracing
cell lineage and analyzing the cell autonomous function of a gene, etc.
STARS technique can achieve a pure genetic control of the probability of gene
activation/knockout in a population of cells. Techniques are available to tune gene
expression levels within a single cell (Deans, Cantor et al. 2007). An interesting analogue
would be to tune the probability of gene expression in a population of cells, which cannot
be readily achieved so far in a straightforward manner. By combining the two strategies

63

described in specific aim 1 and 2, STARS can provide a direct control of gene expression
in the population context. Specifically, STARS will address two challenges in genetics:
first, to achieve differential gene expression in cells which otherwise has identical
“transcriptome” profile; second, to achieve specific gene expression/knockout in a
desired percentage of cells, i.e. the level of sparseness. Based on this, it could be
foreseen that STARS will provide a more efficient way to generate genetic mosaics to
model cancer development, sporadic genetic diseases and other interesting biological
questions.

64

2.2 Proof of Principle and Development of STARS Technique

Development of STARS hypothesis. To develop a method that can stochastically
activate gene expression with desired sparseness, we first considered the reaction kinetics
of Cre/lox-mediated intrachromosomal recombination. A previous in vivo study (Zheng,
Sage et al. 2000) demonstrated that Cre can effectively induce intrachromosomal
recombination between two loxP sites separated as far as several mega basepairs on
mouse chromosome 11 and that this efficiency is reduced when the substrate length
increases. We hypothesize that as the length of DNA sequence between two lox sites
increases, it may take a longer time for the two lox sites to be brought together to form
the synapsed structure (Guo, Gopaul et al. 1997), which is required for recombination to
occur. We designed our strategy for stochastic gene activation/knockout with regulated
sparseness (STARS) as shown in Figure 17A: two independent recombination units (A
and B, which contain different lox variants, lox2272 and loxP, respectively.) are cross
linked with the first one containing gene X and the other containing gene Y. Because
Cre-mediated recombination can only occur between two identical lox sites (Sauer 1993;
Lee and Saito 1998; Feng, Mellor et al. 2000; Schnutgen, Doerflinger et al. 2003), and
excision by one recombination event removes a lox site necessary for the other to occur,
the choice between X and Y expression is made stochastic and mutually exclusive
(Figure 17 B). Under our hypothesis, if the lengths of the two recombination units (A
and B) are equal, their reaction kinetics, reflected by the rate constant k
1
and k
2

respectively, should be comparable and the chances of X and Y expression should be
similar. On the other hand, if the sequence in unit A is longer than that in unit B, the

65

Length Reaction Rate Outcome
A = B k
1
= k
2
A > B k
1
< k
2
A >> B k
1
<< k
2
X Y
Y
Y
X
X
DNA Gene X
lox2272 lox2272 loxP loxP
A
B
Gene Y
Cre
DNA Gene X Gene Y
loxP loxP lox2272
lox2272 loxP
Gene X
lox2272 loxP
Gene Y
Y activation& X knockout
lox2272
k
1
k
2
X activation& Y knockout
A
B
C
Figure 17: STARS strategy. A, The structure of STARS
transgene: gene X flanked by lox2272 pair (Unit A) and gene Y
flanked by loxP pair (Unit B) are crossly linked and subjected to
Cre action. B, Two Cre molecules bind to each lox site and two
identical lox sites are brought together for the recombination to
occur. Cre stochastically excises recombination unit A or B (the
reaction kinetics reflected by the rate constant k1 and k2,
respectively) and leads to mutually exclusive expression of gene
Y and gene X. C, We hypothesize that the length of
recombination unit may affect the reaction kinetics and leads to
differential outcomes after Cre action.

66

reaction time needed for excising A will be longer than that for B, hence the
recombination will be more likely to occur in B, resulting in a higher probability of
expressing gene X than Y (Figure 17 C). Thus, by varying the length of one unit while
keeping the other constant, we can regulate the probability of the desired recombination,
and control the sparseness level of gene activation.
Regulation of the reaction kinetics of Cre/lox recombination through varying the
length of DNA substrate. To test the effect of substrate DNA length on the reaction
kinetics of Cre/lox recombination, a set of constructs were made, which contains
membrane bound-GFP (M-GFP as gene X) flanked by lox2272 sites, mCherry (gene Y)
flanked by loxP sites, and a spacer in unit A of various lengths. Here, M-GFP was
defined as the gene of interest and mCherry was defined as the reporter gene. We
predicted that an increase in the spacer length would result in a higher probability of Cre-
mediated recombination in unit B, leading to the removal of mCherry cassette. A third
marker, fluorescent protein mKusabira orange(mKO) flanked by loxN pair (loxN-mKO-
pA-loxN) was incorporated upstream of the STARS cassette to facilitate selection of cell
lines that have the construct integrated in genome, and to prevent cells from expressing
M-GFP in the absence of Cre (Figure 18A). Single copy of each of these constructs was
inserted into the same locus of HEK293 cell genome with a targeted method (see
experimental design). Selected stable cell lines for each construct were transfected with
the plasmid expressing Cre recombinase and imaged five days post transfection. In the
absence of Cre, all cells were mKO-positive, and no expression of M-GFP or mCherry
was observed (data not shown). Five days after Cre transfection, cells expressing M-GFP
and mCherry were found to be separate populations, consistent with the stochastic and

67

STARS
Transgene
1-1
Spacer(Kb)
1
4
12
8
0
1-2
1-5
1-9
1-13
mCherry
M-GFP
merge
P
CAGGS
loxP lox2272 lox2272 loxP
M-GFP
pA pA
mKO
loxN loxN
Spacer mCherry
A
B
pA
0 1 4 8 12
0.0
0.5
1.0
mCherry+ M-GFP +
STARS Spacer Length (Kb)
Population Percentage
** ** *
A
B
C

68

mutually exclusive choice of expression and supporting the notion that the selected cell
lines had only single copy STARS transgene in the genome (Figure 18 B). In addition,
these cells all lacked mKO fluorescence (data not shown), suggesting that 5 days is
sufficient for the turnover of mKO protein.
To examine the effect of DNA length on the recombination reaction, we derived the ratio
between the number of cells expressing M-GFP and mCherry. For the cell line derived
from transgene without spacer (STARS1-1), M-GFP and mCherry were expressed in
roughly equal number of cells after Cre transfection (53% ± 3% vs. 47% ± 3%,
respectively, mean ± SD ) (Figure 18 B & Figure 18 C), consistent with the observation
with Brainbow constructs
5
. Interestingly, as the length of the spacer in unit A increased
from 0 kb to 12kb (STARS 1-1 to STAR1-13), the number of mCherry-positive cells
dropped monotonically from 47% down to 5% of the total cell population, while M-GFP
positive cells increased to 95% (Figure 18B & Figure 18 C). Here, each cell line with a
specific STARS transgene was tested five times independently. Our data indicated that,
by varying the length of the fragment flanked by lox2272 sites up to 13 kb, mCherry
Figure 18: Test of STARS hypothesis. A, Schematic drawing of the STARS construct.
CMV β-actin enhancer (CAGGS) promoter drives the expression of STARS transgene.
The GFP recombination cassette is flanked by lox2272 pair, with a spacer of various
lengths incorporated. The crosslinked mCherry cassette is flanked by loxP pair. MKO (an
orange FP) flanked by loxN pair is used to mark the stable cell lines and prevent the
expression of M-GFP without Cre action. B, Example images of HEK293 cells stably
expressing various STARS transgenes after a confirmed thorough action of Cre. The
names of STARS constructs (left) and the corresponding spacer length (kb, right) are
indicated. C, Average percentage of cells expressing mCherry (red) or M-GFP (green)
under each STARS transgene expression. Data from five independent experiments for
each cell line were averaged. Bars represent standard deviation.

69

activation (or M-GFP knockout in another word) could be controlled to occur in from
around 50% to 5% of cells (Figure 18 C). Because all STARS transgenes were integrated
into the same locus of HEK293 cell genome (see experimental design), the differential
sparseness is unlikely attributed to the variation in Cre efficiency. These results support
our hypothesis that the longer DNA fragment flanked by lox pair, the longer time it takes
for Cre-mediated recombination to occur.

1 5 9 13
0
10
20
30
40
50
Figure 19: Relationship between the level of sparseness and the ratio of length
between recombination units. In our STARS application, the gene Y in unit B was
used as a reporter with a fixed length of about 1 kb. The regulation is achieved
through unit A. Here, X-axis represents the ratio of DNA length between unit A and
B. Y-axis represents the percentage of cells expressing gene Y (mCherry). Data
point represents each individual experiment. The red curve represents the best fit
(r
2
= 0.97) with a second order exponential function: y = y
0
+ a
1
* e^(-x/t
1
) + a
2
* e^(-
x/t
2
), where y
0
is 0, a
1
is 556.79, t
1
is 0.32, a
2
is 23.81, t
2
is 10.00.

70

The relationship between the percentage of cells expressing mCherry and the ratio
between the lengths of two recombination units can be best described with a second order
exponential decay curve (see Figure 19). This curve may provide guidance for construct
design to achieve desired sparseness level. The dependence of reaction kinetics on the
length of flanked DNA fragment also provides a potential explanation for the apparent
missing of cross recombination between copies of Brainbow transgenes in their
transgenic mice.

Overestimation of M-GFP expressing cells was negligible in our quantification.
Because Cre expression is not long-lasting in transfected cells, a possibility exists that M-
GFP positive cells resulted from simple excision of the mKO unit without recombination
in the mCherry unit. This can lead to an overestimation of the number of M-GFP
expressing cells that directly result from the stochastic choice between competing M-GFP
and mCherry units. To estimate the number of M-GFP cells with mKO excision only
(i.e. “false” M-GFP cells), we sorted out M-GFP-positive cells after Cre action by using
FACS and seeded them in 96-well plates at single-cell per well (see experimental
design). After cells had proliferated and expanded to colonies with thousands of cells,
we further transfected individual colonies with Cre. As “false” M-GFP cells have an
intact STARS cassette, Cre expression will lead to the expression of mCherry in at least
some cells of the colony. This is due to a bona fide action of Cre on the STARS cassette.
As shown in Figure 20, from all colonies collected, only one colony exhibited mCherry
expressing cells. This result indicated that Cre action was in fact highly efficient and
quite complete, and that overestimation of M-GFP expressing cells was negligible.

71

Figure 20: Examination of the identities of M-GFP positive cells. After
Cre action on HEK293-STARS cell lines, M-GFP positive cells were sorted
out and seeded at single cell/well in 96-well plates for proliferation. When the
colonies contain thousands of cells (takes about 2 weeks), they were
transfected with the plasmid expressing Cre. Clones that have cells expressing
mCherry in 5 days post transfection were counted as positive and summarized
in the table.

Cre FACS
Cre
+
-
or
trypsin
2 weeks later
147 67 32 161 98
Total number of
clones examined
0 0 0 0 1
Number of clones with mCherry
postive cells
1-13 1-9 1-5 1-2 1-1
HEK293-STARS
8 12 4 1 0
Spacer (Kb)
147 67 32 161 98
Total number of
clones examined
0 0 0 0 1
Number of clones with mCherry
postive cells
1-13 1-9 1-5 1-2 1-1
HEK293-STARS
8 12 4 1 0
Spacer (Kb)
A
B

72

Expand the sparseness level by mutagenizing lox sequence
So far we can reach the sparseness level down to 5% among the total population by
utilizing lox 2272 paired with 12 Kb spacer in the STARS construct (Figure 18&Figure
19). It is difficult to further extend the sparseness level by increasing the spacer length
beyond 12 Kb due to the technical limitations (e.g. constructs become unstable, etc.) . To
address this challenge, we explored the possibility to employ different lox sites variants
in the STARS construct instead of lox 2272. Previously it has been reported that
mutagenized Lox sites show significant variance in terms of recombination efficiency
upon Cre ‘s action as examined in vitro (Lee and Saito 1998). We hypothesized that if we
replace lox2272 in the STARS transgene with those lox variants with much lower
efficiency, we could further decrease the sparseness level STARS can reach when
coupled with the spacer effect. Indeed, when we replaced lox 2272 in STARS construct
with lox 4271 or lox 3172 , the sparseness level can be extended to 3% coupled with 8Kb
spacer and 0.5% -1% coupled with 6Kb spacer, respectively (Figure 21).

73

P
CAGGS
loxP Lox X Lox X loxP
pA
loxN loxN
Spacer mCherry
A
B
pA STOP
lox X
Sequence
Recombination Efficiency (normalized to loxP)
Spacer length in lox X unit (kb)
0 1 4 6 8 12
lox 2272
ATAACTTCGTATA AaGTATcC TATACGAAGTTAT
47.10% 19.51% 13.10% 8.08% 5.51%
lox 3172
ATAACTTCGTATA ATaTATcC TATACGAAGTTAT
17.31% 10% 0.5-1% <0.001%
lox 4271
ATAACTTCGTATA ATGaATaC TATACGAAGTTAT
18.63% 3.26%
M-GFP
Figure 21: Extension of the sparseness level of STARS technique. Top: the construct
for examining the sparseness level for different lox variants (loxX). The CMV β-actin
enhancer (CAGGS) promoter is used to drive the expression of the STARS transgene.
loxN-STOP-loxN element prevents the expression without Cre action. The spacer
regulates the length of unit A. The table summarizes our findings on different
combinations of lox variants and the length of spacers. The numbers indicate the
sparseness level through measuring the percentage of cells expressing mCherry.

74

2.3 Generating STARS mouse

To implement STARS technique we have developed (Wang, Liu et al. 2009) to study cell
lineage relationship in the inner ear and potentially other systems, we aim to make a
reporter mouse line using STARS system to achieve sparse labeling of cell type of
interest to around 1% sparseness level when crossed with a specific Cre mouse line.
Based on the result summarized above , we decide to use Lox 3172 coupled with 6Kb
spacer to make the final STARS targeting construct to integrate into ROSA26 locus at
mouse chromosome 6(Figure 22). The Rosa26 locus was previously identified in a
gene-trap screen using mouse ES cells and found to support ubiquitous gene expression
for exogenous transgene element integrated at this locus at both embryonic and adult
stages(Friedrich and Soriano 1991; Zambrowicz, Imamoto et al. 1997). Despite general
agreement that ROSA26 locus support high efficiency gene targeting ( generally around
25% positive rate) for regular transgene knock-in (Hohenstein, Slight et al. 2008), we
have substantial difficulty to target our STARS transgene into ROSA26 locus using
traditional homologous recombination approaches as described previously (Srinivas,
Watanabe et al. 2001). This might be primarily due to the extremely long length of
STARS transgene we are trying to integrate into the genome.
To address this challenge, we took advantage of the FLP recombinase assisted gene
targeting approach to target our STARS transgene into mouse ES cells. The Flp-In or
similar recombinase based gene targeting approach have been demonstrated to facilitate
transgene targeting especially for those with large size to a particular genomic locus
harboring the corresponding recombination site with extremely high efficiency

75

5’ ARM 3’ ARM
ROSA genomic locus
PGK
promoter
GFP pA
Frt
PGK promoter
Neo pA
Targeting vector
1
st
step
Pick GFP positive, G418 resistant colonies,
PCR genotyping using primer 1+2
Primer 1
Primer 2
5’ ARM 3’ ARM
ROSA genomic locus
PGK
promoter
GFP pA
Frt
PGK promoter
Neo pA
Primer 1
Primer 2
ROSA genomic locus modified
PGK DTA pA
Flp recombination, pick Hygromycine resistant,
mcherry positive, GFP negative colonies, PCR
genotyping using primer 3+4 and 3+5
Targeting vector
2
nd
step
Primer 3 Primer 4 Primer 5
ROSA genomic locus
with STARS
STEP ONE
STEP TWO
STARS
construct
Intermediate
construct
Figure 22

76

Figure 22: A two-step process to target STARS transgene to ROSA26 locus:
In the first step, we electroporated feeder free low passage E14Tg2a.4 stem cell(P14)
purchased from BayGenomics with ROSA targeting vector harboring Neo coding and
GFP coding sequence as well as a single Frt site immediately upstream of GFP coding
under its PGK promoter. Colonies surviving G418 selection and with GFP color were
picked and genotyped with PCR primer using the primer pair 1 and 2, whereas primer 1
is on the chromosome outside ROSA 3' ARM and primer 2 is on the Neo coding
sequence. In the second step, we inserted STARS transgene into a plasmid harboring
Hygromycin coding sequence without promoter but has a single Frt site immediately in
front of its ORF with matching frame. Then we co-electrporated the ES cell clone
obtained from the first step with our STARS plasmid with pOG44 plasmid encoding a
Flp recombinase with the 1to 9 ratio and selected the colonies using Hygromycin. Flp/Frt
mediated recombination will knock the whole STARS plasmid into the particular Frt site
on the ROSA site we have in step one to displace GFP. So in the correctly targeted event,
GFP is no longer expressed due to the loss of promoter and hygromycin is expressed
instead. Surviving colonies with mCherry color but without GFP color were genotyped
with primer pair 3+4 and 3+5 to confirm the correct targeting.

77

(Wallace, Marques-Kranc et al. 2007; Wirth, Gama-Norton et al. 2007). In fact, this
approach is actually what we have been using to generate HEK293 cells harboring single
copy STARS transgene in their genome when we test our STARS hypothesis.
We took two steps to implement FLP-In strategy to insert STARS transgene into ROSA26
locus (Figure 22). In the first step, we use traditional homologous recombination
technique to target an intermediate construct harboring harboring Neomycin coding and
GFP coding sequence as well as a single Frt site to ROSA26 locus of a E14Tg2a.4 stem
cell (from BayGenomics). We selected the colonies using G418 and picked forming
colonies with GFP color. Then we genotyped those colonies with PCR primer pair 1 plus
2 (Figure 22) to confirm the correct targeting. We identified two colonies, G4 and H2
matching the criteria (Figure 23) and using one of them, H2 (with normal karyotype) to
proceed with the second step targeting.

Figure 23: PCR genotyping of first step targeting. Colonies with GFP color and
surviving the G418 selection were genotyped using primer1 on the chromosome outside
ROSA 3' ARM and primer 2 on the Neo sequence. 5.3Kb DNA band were amplified
from the correctly targeted clones(clone G4 and H2).
ES cell line(ROSA-PGKfrtGFP-PGKneo) 3’ ROSA external PCR 5351bp
G4 H2, H3 H4 H5 H6 H7 H8 F1 to F8
1
st
step targeting genotyping using primer 1+2
5.3Kb

78

In the second step, we inserted STARS transgene into a plasmid harboring Hygromycin
coding sequence without promoter but has single Frt site immediately in front of its ORF
with matching frame. Then we co-electrporated ES cell clone H2 with our STARS
plasmid as well as the pOG44 plasmid encoding a Flp recombinase and selected the
colonies using Hygromycin. Flp/Frt mediated recombination will knock the whole
STARS plasmid into the particular Frt site on the ROSA site we have in step one to
displace GFP coding sequence. So in the correctly targeted event, GFP is no longer
expressed due to the loss of promoter and hygromycin is expressed instead. We picked
emerging colonies with mCherry color but without GFP color and then genotyped with

Figure 24: PCR genotyping the second step STARS targeting. Colonies surviving the
Hygromycin selection with mCherry color but not GFP color were PCR genotyped using
primer 3 on PGK promoter plus primer 4 or 5 on the hygromycin sequence. 500 bp and
980 bp DNA bands were amplified from the correctly targeted clones (clone #5 and #9)
using primer 3+4 and 3+5, respectively.
2
nd
step targeting PCR genotyping of ES FlpIn cell
line #5 and #9
500bp
980bp
#9 #5 #9 #5
primer 3 + 4 Primer 3+5
Primer 3 Primer 4 Primer 5

79

primer pair 3+4 and 3+5. Both #5 and #9 colonies were identified positive(Figure 24) and
with normal karyotype(data not shown). Upon Cre transfection, both clones showed
activation of membrane bound EYFP expression sparsely, with the prediction of
sparseness level around1% among the total population (Figure 25), consistent with the
initial test result from STARS-HEK293 cell lines (Figure 21).

Figure 25: Sparse activation of EYFP expression in STARS ES cell. ES cell correctly
targeted with STARS at ROSA26 locus(clone #5 and #9) were transfected with Cre
expressing plasmid(bottom row) or mock transfected (top row). Four days after the
transfection, a sparse number of ES cells switched the color from mCherry to membrane
bound EYFP only in the presence of Cre (bottom row) but not in the mock transfected
case (top row).

+Cre -Cre
mCherry M-EYFP Merge

80

2.4 Perspectives and Future Direction

After generating ES cells with STARS transgene correctly targeted at ROSA26 locus
(Figure 25), we next injected them into C57BL/6 female mouse blastocysts to produce
chimeric mice. So far, out of the 4 chimeras we obtained, we were not able to harvest the
germline transmitted clones. We suspect this might be due to the loss of germline
competence during our two step manipulations in the process of targeting STARS
transgene.
As an alternative measure, we are very grateful to obtain Ai9 mouse ES cell line which
harbors a pair of FRT sites at ROSA26 locus at mouse chromosome 6 from Allen
Institute with high germline competence (Madisen, Zwingman et al. 2010). This Ai9 ES
cell will serve as the substrate where we will target our STARS transgene into ROSA26
site using one step Flp-in approach . As illustrated in Figure 26, as the recombination
happens, the STARS transgene cassette will replace the Lox-STOP-Lox-tdTomato
cassettes in the original Ai9 ES cell and endows the hygromycin resistance as well the
mCherry fluorescence to the recombinant clones.
After we confirmed the positive clone with correct STARS transgene targeting with PCR
or southern blot, we will inject two positive clones into a C57BL/6 blastocyst to produce
chimeric mice hopefully to obtain higher germline transmission rate. This will be
confirmed by PCR designed to discriminate knock-in alleles for all lines. The resulting
knocking mice will be bred to different Cre lines for further analysis. For example, as the
STARS line is breeded with Foxg1-Cre line which has the activity in the progenitor of
epithelium cells from the cochlear sensory epithelium (Hebert and McConnell 2000;

81

Grimsley-Myers, Sipe et al. 2009; Kopecky, Santi et al. 2011; Yamamoto, Chang et al.
2011), we will be able to see a small number of cells clustered together in the cochlear
epithelium, indicating they are from the same cell lineage. This will enable us to dissect
the lineage relationship between different cell types within the inner ear.

Figure 26: Strategy to make STARS Knock-in Mouse based on Ai9 ES cell. Flp-In
strategy was employed to replace original reporter cassette(lox-Stop-lox-tdTomato)
flanked by a pair of FRT sites at ROSA26 locus in Ai9 ES cells with our STARS
transgene. After contransfect the Ai9 ES cell with STARS transgene and a plasmid
expressing FLP recombinase, the proper recombined clone will be selected based on the
following criteria: first, the original Ai9 ES cell clones will lose their resistance to
Neomycin as the promoter for Neo is lost in the recombined clone. Instead, hygromycin
gene will be placed under the CAG promoter and endow the hygromycin resistance(hyg-
r) to the properly recombined clone. Second, the mCherry will be expressed
constitutively in the STARS transgene, thus will mark the positive clone with mCherry
fluorescence. Positive clones matching passing these two screening will be picked and
further screened with PCR using the primers flanking the FLP recombination junction
site. (Adapted from Madisen, 2010)

Frt Frt
STARS
Frt Frt
CAG Pr
Flp-modified Rosa26 locus
Neo pA
Hyg-r & mCherry +
Hyg
CAG Pr
lox 3172
mCherry
6K pA
lox 3172
lox P lox P Lyn-EYFP
WPRE
CAG Pr
mCherry
lox P lox P Lyn-EYFP
Hyg
lox 3172
6K pA
lox 3172
WPRE

82

In conclusion, STARS technique can allow three types of applications. First, by
combining tissue or cell-type specific Cre expression, a genetic “Golgi”-like staining of a
small number of cells of a particular type can be achieved. Second, overexpression of
exogenous genes in a small population of cells (e.g. by replacing mCherry with any gene
of interest) can be achieved. Finally, it can provide a potential way for sparse knockout of
endogenous genes with simultaneous labeling of the cells undergoing the knockout event,
e.g. in the STARS constructs, mCherry expression can be also viewed as GFP knockout.
With STARS, the level of sparseness might not be influenced by the locus of gene to be
knocked out in the genome. A major advantage of STARS is to be able to regulate the
level of sparseness through controlling the stochastic level of gene activation. Based on
the results from the present study, the STARS strategy can be potentially applied to
create genetic mosaics for studying the population dosage of genes underlying sporadic
genetic diseases, to induce loss of heterozygosity for modeling cancer development and
screening for tumor suppressor genes, to trace neuronal circuit and cell fate, as well as to
determine cell-autonomous function of a gene, etc. Taken together, as a purely genetic
method, STARS can be conveniently and broadly applied, with tissue- or cell-type
specificity provided by specific Cre lines.

83

2.5 Experimental Design and Methods

Generation of STARS constructs and CAGGS-Cre plasmid:
All constructs were made using standard molecular cloning methods.

Briefly, CMV β-
actin enhancer (CAGGS) promoter (Niwa, Yamamura et al. 1991; Okada, Lansford et al.
1999) was digested from pCAGGS-ES plasmid (gift of Dr. Le Ma) using restriction
enzyme SpeI plus EcoRI and filled to blunt end at EcoRI site with T4 DNA
polymerase(Promega). This fragment was subcloned into pBlue-script SK+/- digested
with SpeI plus SmaI. To obtain lox (sequence underlied in primers listed below)-flanked
XFP, loxN pair was introduced in the primers amplifying mKusabira orange (Karasawa,
Araki et al. 2004)

(5’ primer: ctgcagataacttcgtatagtataccttatacgaagttataccatggtgagtgtgatta
a, 3’primer: ctgcagataacttcgtataaggtatactatacgaagttatattaaaaaac ctcccacacct)and lox2272-
loxP pair was introduced in the primers amplifying both membrane bound (M)-GFP(or
lynGFP)(Koster and Fraser 2001) (5’primer:agatatcctgcagataacttcgtataggatactttatacgaa
gttataccatgggatgtattaaatca, 3’primer: ccatgggaagcttataacttcgtataatgtatgctatacgaagttatgaatt
cggcgcgccggccggccgaattcgtcgaggccgcgaattaaaaaacc)and mCherry (Shaner, Campbell et
al. 2004)( 5’ primer: ccatgggaagcttataacttcgtataggatactttatacgaagttatcacgtgccaccatggtga
gcaagggcgagga, 3’primer:ggaattcctcgagataacttcgtataatgtatgctatacgaagttatagatctgtcgaggcc
gcgaattaaaaaacc). Polymerase chain reaction(PCR) was then performed to generate PstI-
lox2272-M-GFP-SV40polyA-EcoRI-AscI-FseI-EcoRI-loxP-HindIII, PstI-loxN-mKO-
SV40polyA- loxN-PstI and HindIII-lox2272-mCherry-SV40polyA-loxP-XhoI fragments,
which were sequentially inserted downstream of CAGGS promoter in pBlue-script SK+/-
by PstI plus HindIII, PstI and HindIII plus XhoI respectively to make STARS 1-1. We
then generated multiples of tetra-poly(A) fragment (Jackson, Willis et al. 2001) (4kb,

84

8kb, 12kb)flanked by AscI and FseI in pCR-script plasmid and inserted these fragments
or a single poly(A)(1kb) fragment flanked by EcoRI into the spacer of STARS1-1 by
either AscI plus FseI or EcoRI respectively to generate other STARS constructs. To
make stable HEK293 cell lines, the entire STARS construct cassettes were subcloned
using XhoI plus NotI sites into pcDNA5/FRT expression vector (Invitrogen) previously
modified to remove its CMV promoter. CAGGS-Cre construct was made by subcloning
Cre cDNA plus HSV-TKpolyA from pACN plasmid (Bunting, Bernstein et al. 1999) to
pCAGGS-ES vector using EcoRI plus NotI sites.

Generation of stable HEK293 cell lines with single-copy STARS integration.
HEK293 cell lines stably transfected with single copy of various STARS constructs were
made using Flp-In
TM
system (Invitrogen). Briefly, HEK293 Flp-In cells were cultured in
DMEM,

containing 10% FCS, 1% penicillin/streptomycin and 100ug/ml zeocin in a
temperature- and humidity-controlled incubator (95% air,

5% CO
2
as the gas phase) at
37° C. Subconfluent cells were cotransfected with pOG44 (Invitrogen): pcDNA/ FRT/
STARS at 9:1 ratio using electroporation methods. 48hrs after transfection, cells were
split into 10cm tissue culture dishes and selected with 100ug/ml hygromycin until
colonies are visible. Individual colonies were further screened for mKO expression and
picked to expand until subconfluence for transfection. Due to the fact that Flp-In system
utilized Flp recombinase mediated transgene integration into single Frt site previously
inserted in the genome of Flp-In
TM
HEK293 cells, the clones generated should be isogenic
and have single copy integration of STARS transgenes. This is further supported by the

85

fact that in all tested cell lines M-GFP-positive and mCherry-positive cells are separate
populations after Cre action.
Cell Transfection, FACS sorting and data analysis.
HEK293 Flp-In cells with single copy STARS integration were cultured under
conditions described above until subconfluence, and then transfected with CAGGS-Cre
plasmid using lipofectamine 2000 (Invitrogen) according to the product manual. Five
days after transfection, cells were imaged with a custom-built epi-fluorescence
microscope equipped with a 10X objective (Olympus). For each line with desired
STARS transgene, around 800 cells were counted for each transfection and data from five
independent transfections were averaged to derive the ratio of the number of M-GFP
positive versus mCherry positive cells. In several other experiments, we included 1µ M
cytosine arabinoside in the medium during transfection to slow down cell proliferation as
to limit potential loss of Cre plasmid during cell division. No significant difference was
observed between results with and without cytosine arabinoside treatment. In addition,
we infected cells with lentivirus expressing CreER under CMV promoter with a MOI
(multiplicity of infection) of 8 and cultured cells in the presence of 1.25µ M 4-hydroxy-
tamoxinfen. Again, no significant difference was observed compared to the data from
Cre plasmid transfection. The following filter sets (Chroma Technology) were used to
separate signals between GFP/mKO/mCerry channels without detectable cross-talk: for
GFP, excitation 465/37, emission 512/35, and 490DCLP; for mKO: 540/10, 570/20, and
550DCLP; for mCherry, 580/10, 630/50, and 600DCLP. Images were analyzed and
edited using Image J software.

86

Cells were trypsinized after imaging, suspended in PBS and subjected to FACS sorting
for M-GFP positive cells. The latter were seeded in 96-well plates at single cell/well for
expansion and subsequent transfection. The FACS was performed using a BD FACSAria
cell sorter.
ES cell culture and genotyping
ES cells were cultured according to the standard protocol described previously
(Valenzuela, Murphy et al. 2003; Friedel, Plump et al. 2005). In brief, ES cell were
cultured under feeder free condition using standard ES cell medium (GMEM plus 10%
Fetal Bovine Serum, 100µ M Non-Essential Amino Acids Solution, 2 mM L-Glutamine,
1mM Sodium Pyruvate Solution, 10
3
units/ml ESGRO® (LIF)) on gelation coated plate.
About 1x10
7
cell were harvested and electroporated(800 V, 10 µ F, exponential decay
mode) with 80-100 µ g DNA either linearized(first step targeting) or un-linearized
(second step for Flp-In approach) and selected with G418 ( 200µ g/ml) or hygromycin (
100µ g/ml) for a week.
DNA lysis of ES cells colonies were performed in a 96 well plate using mix sarcosyl-
DNA lysis buffer with Proteinase K(1 mg/ml) and extracted with ice cold salt-ethanol
mixture. Long range PCR genotyping was performed using Accuprime HD-Taq
system(Invitrogen).

87

Bibliography

Abraira, V. E., T. del Rio, et al. (2008). "Cross-repressive interactions between Lrig3 and
netrin 1 shape the architecture of the inner ear." Development 135(24): 4091-
4099.
Adam, Y. and A. Mizrahi (2010). "Circuit formation and maintenance--perspectives from
the mammalian olfactory bulb." Curr Opin Neurobiol 20(1): 134-140.
Andrews, W., M. Barber, et al. (2008). "The role of Slit-Robo signaling in the generation,
migration and morphological differentiation of cortical interneurons." Dev Biol
313(2): 648-658.
Andrews, W. D., M. Barber, et al. (2007). "Slit-Robo interactions during cortical
development." J Anat 211(2): 188-198.
Appler, J. M. and L. V. Goodrich (2011). "Connecting the ear to the brain: Molecular
mechanisms of auditory circuit assembly." Prog Neurobiol 93(4): 488-508.
Badea, T. C., Y. Wang, et al. (2003). "A noninvasive genetic/pharmacologic strategy for
visualizing cell morphology and clonal relationships in the mouse." J Neurosci
23(12657690): 2314-2322.
Battisti, A. C. and D. M. Fekete (2008). "Slits and Robos in the developing chicken inner
ear." Dev Dyn 237(2): 476-484.
Bermingham, N. A., B. A. Hassan, et al. (1999). "Math1: an essential gene for the
generation of inner ear hair cells." Science 284(5421): 1837-1841.
Beurg, M., R. Fettiplace, et al. (2009). "Localization of inner hair cell mechanotransducer
channels using high-speed calcium imaging." Nat Neurosci 12(5): 553-558.
Bianchi, L. M. and H. Liu (1999). "Comparison of ephrin-A ligand and EphA receptor
distribution in the developing inner ear." Anatomical Record 254(1): 127-134.
Bolstad, B. M., R. A. Irizarry, et al. (2003). "A comparison of normalization methods for
high density oligonucleotide array data based on variance and bias."
Bioinformatics 19(2): 185-193.
Borrell, V., A. Cardenas, et al. (2012). "Slit/Robo signaling modulates the proliferation of
central nervous system progenitors." Neuron 76(2): 338-352.
Brose, K., K. S. Bland, et al. (1999). "Slit proteins bind Robo receptors and have an
evolutionarily conserved role in repulsive axon guidance." Cell 96(6): 795-806.

88

Bunting, M., K. E. Bernstein, et al. (1999). "Targeting genes for self-excision in the germ
line." Genes Dev 13(10385621): 1524-1528.
Cariboni, A., W. D. Andrews, et al. (2012). "Slit2 and Robo3 modulate the migration of
GnRH-secreting neurons." Development 139(18): 3326-3331.
Chao, D. L., L. Ma, et al. (2009). "Transient cell-cell interactions in neural circuit
formation." Nat Rev Neurosci 10(4): 262-271.
Charron, F. and M. Tessier-Lavigne (2005). "Novel brain wiring functions for classical
morphogens: a role as graded positional cues in axon guidance." Development
132(10): 2251-2262.
Chedotal, A. (2007). "Slits and their receptors." Adv Exp Med Biol 621: 65-80.
Chen, P. and N. Segil (1999). "p27(Kip1) links cell proliferation to morphogenesis in the
developing organ of Corti." Development 126(8): 1581-1590.
Chen, Z. Y. and D. P. Corey (2002). "An inner ear gene expression database." J Assoc
Res Otolaryngol 3(2): 140-148.
Chen, Z. Y. and D. P. Corey (2002). "Understanding inner ear development with gene
expression profiling." J Neurobiol 53(2): 276-285.
Chilton, J. K. and S. Guthrie (2003). "Cranial expression of class 3 secreted semaphorins
and their neuropilin receptors." Developmental Dynamics 228(4): 726-733.
Cho, J. H., J. E. Prince, et al. (2009). "Axon guidance events in the wiring of the
mammalian olfactory system." Mol Neurobiol 39(1): 1-9.
Coate, T. M., S. Raft, et al. (2012). "Otic mesenchyme cells regulate spiral ganglion axon
fasciculation through a Pou3f4/EphA4 signaling pathway." Neuron 73(1): 49-63.
Corey, D. P. and A. J. Hudspeth (1983). "Kinetics of the receptor current in bullfrog
saccular hair cells." J Neurosci 3(5): 962-976.
Crawford, A. C., M. G. Evans, et al. (1989). "Activation and adaptation of transducer
currents in turtle hair cells." J Physiol 419: 405-434.
Dallos, P., X. Wu, et al. (2008). "Prestin-based outer hair cell motility is necessary for
mammalian cochlear amplification." Neuron 58(3): 333-339.
Deans, T. L., C. R. Cantor, et al. (2007). "A tunable genetic switch based on RNAi and
repressor proteins for regulating gene expression in mammalian cells." Cell
130(17662949): 363-372.
Defourny, J., F. Lallemend, et al. (2011). "Structure and development of cochlear afferent
innervation in mammals." Am J Physiol Cell Physiol 301(4): 13.
Dodd, J. and T. M. Jessell (1988). "Axon guidance and the patterning of neuronal
projections in vertebrates." Science 242(4879): 692-699.

89

Doetzlhofer, A., P. White, et al. (2006). "Prospective identification and purification of
hair cell and supporting cell progenitors from the embryonic cochlea." Brain Res
1091(1): 282-288.
Doetzlhofer, A., P. M. White, et al. (2004). "In vitro growth and differentiation of
mammalian sensory hair cell progenitors: a requirement for EGF and periotic
mesenchyme." Dev Biol 272(2): 432-447.
Dragatsis, I., M. S. Levine, et al. (2000). "Inactivation of Hdh in the brain and testis
results in progressive neurodegeneration and sterility in mice." Nat Genet
26(11062468): 300-306.
Echteler, S. M. (1992). "Developmental segregation in the afferent projections to
mammalian auditory hair cells." Proc Natl Acad Sci U S A 89(14): 6324-6327.
Efron, B. and R. Tibshirani (2007). "On testing the significance of sets of genes." Annals
of Applied Statistics 1(1): 107-129.
Farinas, I., K. R. Jones, et al. (2001). "Spatial shaping of cochlear innervation by
temporally regulated neurotrophin expression." J Neurosci 21(16): 6170-6180.
Fekete, D. M. and A. M. Campero (2007). "Axon guidance in the inner ear." International
Journal of Developmental Biology 51(6-7): 549-556.
Fekete, D. M., S. Muthukumar, et al. (1998). "Hair cells and supporting cells share a
common progenitor in the avian inner ear." J Neurosci 18(19): 7811-7821.
Feng, G., R. H. Mellor, et al. (2000). "Imaging neuronal subsets in transgenic mice
expressing multiple spectral variants of GFP." Neuron 28(11086982): 41-51.
Fettiplace, R. and A. J. Ricci (2003). "Adaptation in auditory hair cells." Curr Opin
Neurobiol 13(4): 446-451.
Friedel, R. H., A. Plump, et al. (2005). "Gene targeting using a promoterless gene trap
vector ("targeted trapping") is an efficient method to mutate a large fraction of
genes." Proc Natl Acad Sci U S A 102(37): 13188-13193.
Friedrich, G. and P. Soriano (1991). "Promoter traps in embryonic stem cells: a genetic
screen to identify and mutate developmental genes in mice." Genes Dev 5(9):
1513-1523.
Fritzsch, B., V. A. Matei, et al. (2005). "Atoh1 null mice show directed afferent fiber
growth to undifferentiated ear sensory epithelia followed by incomplete fiber
retention." Dev Dyn 233(2): 570-583.
Fritzsch, B., S. Pauley, et al. (2006). "Cells, molecules and morphogenesis: the making of
the vertebrate ear." Brain Res 26(1): 151-171.
Frotscher, M. (2010). "Role for Reelin in stabilizing cortical architecture." Trends
Neurosci 33(9): 407-414.

90

Gale, J. E., W. Marcotti, et al. (2001). "FM1-43 dye behaves as a permeant blocker of the
hair-cell mechanotransducer channel." J Neurosci 21(18): 7013-7025.
Geleoc, G. S. and J. R. Holt (2003). "Developmental acquisition of sensory transduction
in hair cells of the mouse inner ear." Nat Neurosci 6(10): 1019-1020.
Gillespie, L. N., P. L. Marzella, et al. (2005). "Netrin-1 as a guidance molecule in the
postnatal rat cochlea." Hear Res 199(1-2): 117-123.
Gillespie, P. G. and J. L. Cyr (2004). "Myosin-1c, the hair cell's adaptation motor." Annu
Rev Physiol 66: 521-545.
Gillespie, P. G. and U. Muller (2009). "Mechanotransduction by hair cells: models,
molecules, and mechanisms." Cell 139(1): 33-44.
Gong, T. W., I. J. Karolyi, et al. (2006). "Age-related changes in cochlear gene
expression in normal and shaker 2 mice." J Assoc Res Otolaryngol 7(3): 317-328.
Grieshammer, U., M. Le, et al. (2004). "SLIT2-mediated ROBO2 signaling restricts
kidney induction to a single site." Dev Cell 6(5): 709-717.
Grimsley-Myers, C. M., C. W. Sipe, et al. (2009). "The small GTPase Rac1 regulates
auditory hair cell morphogenesis." J Neurosci 29(50): 15859-15869.
Groves, A. K. and D. M. Fekete (2012). "Shaping sound in space: the regulation of inner
ear patterning." Development 139(2): 245-257.
Guan, C. B., H. T. Xu, et al. (2007). "Long-range Ca2+ signaling from growth cone to
soma mediates reversal of neuronal migration induced by slit-2." Cell 129(2):
385-395.
Gubbels, S. P., D. W. Woessner, et al. (2008). "Functional auditory hair cells produced in
the mammalian cochlea by in utero gene transfer." Nature 455(7212): 537-541.
Guo, F., D. N. Gopaul, et al. (1997). "Structure of Cre recombinase complexed with
DNA in a site-specific recombination synapse." Nature 389(9288963): 40-46.
Hayashi, S. and A. P. McMahon (2002). "Efficient recombination in diverse tissues by a
tamoxifen-inducible form of Cre: a tool for temporally regulated gene
activation/inactivation in the mouse." Dev Biol 244(11944939): 305-318.
Hebert, J. M. and S. K. McConnell (2000). "Targeting of cre to the Foxg1 (BF-1) locus
mediates loxP recombination in the telencephalon and other developing head
structures." Dev Biol 222(2): 296-306.
Hertzano, R., M. Montcouquiol, et al. (2004). "Transcription profiling of inner ears from
Pou4f3(ddl/ddl) identifies Gfi1 as a target of the Pou4f3 deafness gene." Hum
Mol Genet 13(18): 2143-2153.
Hohenstein, P., J. Slight, et al. (2008). "High-efficiency Rosa26 knock-in vector
construction for Cre-regulated overexpression and RNAi." Pathogenetics 1(1): 3.

91

Holmes, G. and L. Niswander (2001). "Expression of slit-2 and slit-3 during chick
development." Dev Dyn 222(2): 301-307.
Holmes, G. P., K. Negus, et al. (1998). "Distinct but overlapping expression patterns of
two vertebrate slit homologs implies functional roles in CNS development and
organogenesis." Mech Dev 79(1-2): 57-72.
Hu, H., M. Li, et al. (2005). "Cross GTPase-activating protein (CrossGAP)/Vilse links
the Roundabout receptor to Rac to regulate midline repulsion." Proc Natl Acad
Sci U S A 102(12): 4613-4618.
Huang, L. C., P. R. Thorne, et al. (2007). "Spatiotemporal definition of neurite
outgrowth, refinement and retraction in the developing mouse cochlea."
Development 134(16): 2925-2933.
Huberman, A. D., T. R. Clandinin, et al. (2010). "Molecular and cellular mechanisms of
lamina-specific axon targeting." Cold Spring Harb Perspect Biol 2(3).
Huberman, A. D., M. B. Feller, et al. (2008). "Mechanisms underlying development of
visual maps and receptive fields." Annu Rev Neurosci 31: 479-509.
Hudspeth, A. J. (2008). "Making an effort to listen: mechanical amplification in the ear."
Neuron 59(4): 530-545.
Imai, T. and H. Sakano (2011). "Axon-axon interactions in neuronal circuit assembly:
lessons from olfactory map formation." Eur J Neurosci 34(10): 1647-1654.
Izumikawa, M., R. Minoda, et al. (2005). "Auditory hair cell replacement and hearing
improvement by Atoh1 gene therapy in deaf mammals." Nat Med 11(3): 271-276.
Jackson, E. L., N. Willis, et al. (2001). "Analysis of lung tumor initiation and progression
using conditional expression of oncogenic K-ras." Genes Dev 15(11751630):
3243-3248.
Jahan, I., J. Kersigo, et al. (2010). "Neurod1 regulates survival and formation of
connections in mouse ear and brain." Cell Tissue Res 341(1): 95-110.
Jaworski, A. and M. Tessier-Lavigne (2012). "Autocrine/juxtaparacrine regulation of
axon fasciculation by Slit-Robo signaling." Nat Neurosci 15(3): 367-369.
Jones, C. and P. Chen (2007). "Planar cell polarity signaling in vertebrates." Bioessays
29(2): 120-132.
Jones, J. M., M. Montcouquiol, et al. (2006). "Inhibitors of differentiation and DNA
binding (Ids) regulate Math1 and hair cell formation during the development of
the organ of Corti." J Neurosci 26(2): 550-558.
Karasawa, S., T. Araki, et al. (2004). "Cyan-emitting and orange-emitting fluorescent
proteins as a donor/acceptor pair for fluorescence resonance energy transfer."
Biochem J 381(15065984): 307-312.

92

Kazmierczak, P. and U. Muller (2012). "Sensing sound: molecules that orchestrate
mechanotransduction by hair cells." Trends Neurosci 35(4): 220-229.
Kazmierczak, P., H. Sakaguchi, et al. (2007). "Cadherin 23 and protocadherin 15 interact
to form tip-link filaments in sensory hair cells." Nature 449(7158): 87-91.
Kelley, M. W. (2006). "Regulation of cell fate in the sensory epithelia of the inner ear."
Nat Rev Neurosci 7(11): 837-849.
Kelley, M. W. (2007). "Cellular commitment and differentiation in the organ of Corti."
Int J Dev Biol 51(6-7): 571-583.
Kelly, M. and P. Chen (2007). "Shaping the mammalian auditory sensory organ by the
planar cell polarity pathway." Int J Dev Biol 51(6-7): 535-547.
Kelly, M. C. and P. Chen (2009). "Development of form and function in the mammalian
cochlea." Curr Opin Neurobiol 19(4): 395-401.
Kidd, T., K. S. Bland, et al. (1999). "Slit is the midline repellent for the robo receptor in
Drosophila." Cell 96(6): 785-794.
Kidd, T., K. Brose, et al. (1998). "Roundabout controls axon crossing of the CNS midline
and defines a novel subfamily of evolutionarily conserved guidance receptors."
Cell 92(2): 205-215.
Kim, W. Y., B. Fritzsch, et al. (2001). "NeuroD-null mice are deaf due to a severe loss of
the inner ear sensory neurons during development." Development 128(3): 417-
426.
Kolodkin, A. L. and M. Tessier-Lavigne (2011). "Mechanisms and molecules of neuronal
wiring: a primer." Cold Spring Harb Perspect Biol 3(6).
Kopecky, B., P. Santi, et al. (2011). "Conditional deletion of N-Myc disrupts
neurosensory and non-sensory development of the ear." Dev Dyn 240(6): 1373-
1390.
Koster, R. W. and S. E. Fraser (2001). "Tracing transgene expression in living zebrafish
embryos." Dev Biol 233(11336499): 329-346.
Koundakjian, E. J., J. L. Appler, et al. (2007). "Auditory neurons make stereotyped
wiring decisions before maturation of their targets." J Neurosci 27(51): 14078-
14088.
Kuhlman, S. J. and Z. J. Huang (2008). "High-resolution labeling and functional
manipulation of specific neuron types in mouse brain by Cre-activated viral gene
expression." PLoS One 3(18414675).
Lee, G. and I. Saito (1998). "Role of nucleotide sequences of loxP spacer region in Cre-
mediated recombination." Gene 216(9714735): 55-65.

93

Lee, K. H. and M. E. Warchol (2008). "Promotion of neurite outgrowth and axon
guidance in spiral ganglion cells by netrin-1." Arch Otolaryngol Head Neck Surg
134(2): 146-151.
Lee, Y. S., F. Liu, et al. (2006). "A morphogenetic wave of p27Kip1 transcription directs
cell cycle exit during organ of Corti development." Development 133(15): 2817-
2826.
Lefebvre, V., B. Dumitriu, et al. (2007). "Control of cell fate and differentiation by Sry-
related high-mobility-group box (Sox) transcription factors." Int J Biochem Cell
Biol 39(12): 2195-2214.
Legg, J. A., J. M. Herbert, et al. (2008). "Slits and Roundabouts in cancer, tumour
angiogenesis and endothelial cell migration." Angiogenesis 11(1): 13-21.
Lelli, A., Y. Asai, et al. (2009). "Tonotopic gradient in the developmental acquisition of
sensory transduction in outer hair cells of the mouse cochlea." J Neurophysiol
101(6): 2961-2973.
Lewandoski, M. (2001). "Conditional control of gene expression in the mouse." Nat Rev
Genet 2(11584291): 743-755.
Liang, L. and L. Luo (2010). "The olfactory circuit of the fruit fly Drosophila
melanogaster." Sci China Life Sci 53(4): 472-484.
Liberman, M. C., J. Gao, et al. (2002). "Prestin is required for electromotility of the outer
hair cell and for the cochlear amplifier." Nature 419(6904): 300-304.
Liu, D., H. Chu, et al. (2003). "Fgf3 and Fgf8 dependent and independent transcription
factors are required for otic placode specification." Development 130(10): 2213-
2224.
Liu, M., F. A. Pereira, et al. (2000). "Essential role of BETA2/NeuroD1 in development
of the vestibular and auditory systems." Genes Dev 14(22): 2839-2854.
Liu, P., N. A. Jenkins, et al. (2002). "Efficient Cre-loxP-induced mitotic recombination in
mouse embryonic stem cells." Nat Genet 30(11740496): 66-72.
Liu, X., Y. Lu, et al. (2012). "Slit2 regulates the dispersal of oligodendrocyte precursor
cells via Fyn/RhoA signaling." J Biol Chem 287(21): 17503-17516.
Livet, J., T. A. Weissman, et al. (2007). "Transgenic strategies for combinatorial
expression of fluorescent proteins in the nervous system." Nature 450(17972876):
56-62.
Lu, B., K. H. Wang, et al. (2009). "Molecular mechanisms underlying neural circuit
formation." Curr Opin Neurobiol 19(2): 162-167.

94

Lu, C. C., J. M. Appler, et al. (2011). "Developmental profiling of spiral ganglion
neurons reveals insights into auditory circuit assembly." J Neurosci 31(30):
10903-10918.
Lundstrom, A., M. Gallio, et al. (2004). "Vilse, a conserved Rac/Cdc42 GAP mediating
Robo repulsion in tracheal cells and axons." Genes Dev 18(17): 2161-2171.
Luo, L. (2007). "Fly MARCM and mouse MADM: genetic methods of labeling and
manipulating single neurons." Brain Res Rev 55(17408568): 220-227.
Luo, L., E. M. Callaway, et al. (2008). "Genetic dissection of neural circuits." Neuron
57(18341986): 634-660.
Ma, L. and M. Tessier-Lavigne (2007). "Dual branch-promoting and branch-repelling
actions of Slit/Robo signaling on peripheral and central branches of developing
sensory axons." J Neurosci 27(25): 6843-6851.
Ma, Q., D. J. Anderson, et al. (2000). "Neurogenin 1 null mutant ears develop fewer,
morphologically normal hair cells in smaller sensory epithelia devoid of
innervation." J Assoc Res Otolaryngol 1(2): 129-143.
Ma, Q., Z. Chen, et al. (1998). "neurogenin1 is essential for the determination of neuronal
precursors for proximal cranial sensory ganglia." Neuron 20(3): 469-482.
Madisen, L., T. A. Zwingman, et al. (2010). "A robust and high-throughput Cre reporting
and characterization system for the whole mouse brain." Nat Neurosci 13(1): 133-
140.
Marillat, V., O. Cases, et al. (2002). "Spatiotemporal expression patterns of slit and robo
genes in the rat brain." J Comp Neurol 442(2): 130-155.
Marin, O. and J. L. Rubenstein (2003). "Cell migration in the forebrain." Annu Rev
Neurosci 26: 441-483.
Matei, V., S. Pauley, et al. (2005). "Smaller inner ear sensory epithelia in Neurog 1 null
mice are related to earlier hair cell cycle exit." Dev Dyn 234(3): 633-650.
Matilainen, T., M. Haugas, et al. (2007). "Analysis of Netrin 1 receptors during inner ear
development." International Journal of Developmental Biology 51(5): 409-413.
Matsuoka, R. L., Z. Jiang, et al. (2012). "Guidance-cue control of horizontal cell
morphology, lamination, and synapse formation in the mammalian outer retina." J
Neurosci 32(20): 6859-6868.
Matsuoka, R. L., K. T. Nguyen-Ba-Charvet, et al. (2011). "Transmembrane semaphorin
signalling controls laminar stratification in the mammalian retina." Nature
470(7333): 259-263.
McKay, I. J., J. Lewis, et al. (1996). "The role of FGF-3 in early inner ear development:
an analysis in normal and kreisler mutant mice." Dev Biol 174(2): 370-378.

95

Mellitzer, G., Q. Xu, et al. (1999). "Eph receptors and ephrins restrict cell intermingling
and communication." Nature 400(6739): 77-81.
Meyer, A. C., T. Frank, et al. (2009). "Tuning of synapse number, structure and function
in the cochlea." Nat Neurosci 12(4): 444-453.
Meyers, J. R., R. B. MacDonald, et al. (2003). "Lighting up the senses: FM1-43 loading
of sensory cells through nonselective ion channels." J Neurosci 23(10): 4054-
4065.
Miyazaki, N., T. Furuyama, et al. (1999). "Expression of mouse semaphorin H mRNA in
the inner ear of mouse fetuses." Neuroscience Letters 261(1-2): 127-129.
Morris, J. K., A. Maklad, et al. (2006). "A disorganized innervation of the inner ear
persists in the absence of ErbB2." Brain Res 1091(1): 186-199.
Muller, U. (2008). "Cadherins and mechanotransduction by hair cells." Curr Opin Cell
Biol 20(5): 557-566.
Murakami, Y., F. Suto, et al. (2001). "Differential expression of plexin-A subfamily
members in the mouse nervous system." Developmental Dynamics 220(3): 246-
258.
Muzumdar, M. D., L. Luo, et al. (2007). "Modeling sporadic loss of heterozygosity in
mice by using mosaic analysis with double markers (MADM)." Proc Natl Acad
Sci U S A 104(17360552): 4495-4500.
Nayagam, B. A., M. A. Muniak, et al. (2011). "The spiral ganglion: connecting the
peripheral and central auditory systems." Hear Res 278(1-2): 2-20.
Nayak, G. D., H. S. Ratnayaka, et al. (2007). "Development of the hair bundle and
mechanotransduction." Int J Dev Biol 51(6-7): 597-608.
Nguyen-Ba-Charvet, K. T., N. Picard-Riera, et al. (2004). "Multiple roles for slits in the
control of cell migration in the rostral migratory stream." J Neurosci 24(6): 1497-
1506.
Nguyen Ba-Charvet, K. T., K. Brose, et al. (2001). "Diversity and specificity of actions
of Slit2 proteolytic fragments in axon guidance." J Neurosci 21(12): 4281-4289.
Niwa, H., K. Yamamura, et al. (1991). "Efficient selection for high-expression
transfectants with a novel eukaryotic vector." Gene 108(1660837): 193-199.
Ohyama, T., O. A. Mohamed, et al. (2006). "Wnt signals mediate a fate decision between
otic placode and epidermis." Development 133(5): 865-875.
Okada, A., R. Lansford, et al. (1999). "Imaging cells in the developing nervous system
with retrovirus expressing modified green fluorescent protein." Exp Neurol
156(10328944): 394-406.

96

Ozaki, H., K. Nakamura, et al. (2004). "Six1 controls patterning of the mouse otic
vesicle." Development 131(3): 551-562.
Petit, C. and G. P. Richardson (2009). "Linking genes underlying deafness to hair-bundle
development and function." Nat Neurosci 12(6): 703-710.
Pickles, J. O., C. Claxton, et al. (2002). "Complementary and layered expression of Ephs
and ephrins in developing mouse inner ear." Journal of Comparative Neurology
449(3): 207-216.
Plump, A. S., L. Erskine, et al. (2002). "Slit1 and Slit2 cooperate to prevent premature
midline crossing of retinal axons in the mouse visual system." Neuron 33(2): 219-
232.
Puligilla, C., A. Dabdoub, et al. (2010). "Sox2 Induces Neuronal Formation in the
Developing Mammalian Cochlea." Journal of Neuroscience 30(2): 714-722.
Puligilla, C., A. Dabdoub, et al. (2010). "Sox2 induces neuronal formation in the
developing mammalian cochlea." J Neurosci 30(2): 714-722.
Riccomagno, M. M., S. Takada, et al. (2005). "Wnt-dependent regulation of inner ear
morphogenesis is balanced by the opposing and supporting roles of Shh." Genes
Dev 19(13): 1612-1623.
Rothberg, J. M., D. A. Hartley, et al. (1988). "slit: an EGF-homologous locus of D.
melanogaster involved in the development of the embryonic central nervous
system." Cell 55(6): 1047-1059.
Rothberg, J. M., J. R. Jacobs, et al. (1990). "slit: an extracellular protein necessary for
development of midline glia and commissural axon pathways contains both EGF
and LRR domains." Genes Dev 4(12A): 2169-2187.
Rubel, E. W. and B. Fritzsch (2002). "Auditory system development: primary auditory
neurons and their targets." Annu Rev Neurosci 25: 51-101.
Ruben, R. J. (1967). "Development of the inner ear of the mouse: a radioautographic
study of terminal mitoses." Acta Otolaryngol 220: 1-44.
Sabatier, C., A. S. Plump, et al. (2004). "The divergent Robo family protein rig-1/Robo3
is a negative regulator of slit responsiveness required for midline crossing by
commissural axons." Cell 117(2): 157-169.
Sajan, S. A., M. E. Warchol, et al. (2007). "Toward a systems biology of mouse inner ear
organogenesis: gene expression pathways, patterns and network analysis."
Genetics 177(1): 631-653.
Sakaguchi, H., J. Tokita, et al. (2009). "Tip links in hair cells: molecular composition and
role in hearing loss." Curr Opin Otolaryngol Head Neck Surg 17(5): 388-393.

97

Salinas, P. C. and Y. Zou (2008). "Wnt signaling in neural circuit assembly." Annu Rev
Neurosci 31: 339-358.
Sanes, J. R. and M. Yamagata (2009). "Many paths to synaptic specificity." Annu Rev
Cell Dev Biol 25: 161-195.
Sanes, J. R. and S. L. Zipursky (2010). "Design principles of insect and vertebrate visual
systems." Neuron 66(1): 15-36.
Sauer, B. (1993). "Manipulation of transgenes by site-specific recombination: use of Cre
recombinase." Methods Enzymol 225(8231893): 890-900.
Schnutgen, F., N. Doerflinger, et al. (2003). "A directional strategy for monitoring Cre-
mediated recombination at the cellular level in the mouse." Nat Biotechnol
21(12665802): 562-565.
Schwander, M., B. Kachar, et al. (2010). "Review series: The cell biology of hearing." J
Cell Biol 190(1): 9-20.
Seeger, M., G. Tear, et al. (1993). "Mutations affecting growth cone guidance in
Drosophila: genes necessary for guidance toward or away from the midline."
Neuron 10(3): 409-426.
Seiradake, E., A. C. von Philipsborn, et al. (2009). "Structure and functional relevance of
the Slit2 homodimerization domain." EMBO Rep 10(7): 736-741.
Shaner, N. C., R. E. Campbell, et al. (2004). "Improved monomeric red, orange and
yellow fluorescent proteins derived from Discosoma sp. red fluorescent protein."
Nat Biotechnol 22(15558047): 1567-1572.
Shen, K. and C. W. Cowan (2010). "Guidance molecules in synapse formation and
plasticity." Cold Spring Harb Perspect Biol 2(4): 10.
Siemens, J., C. Lillo, et al. (2004). "Cadherin 23 is a component of the tip link in hair-cell
stereocilia." Nature 428(6986): 950-955.
Sienknecht, U. J. and D. M. Fekete (2008). "Comprehensive Wnt-related gene expression
during cochlear duct development in chicken." Journal of Comparative Neurology
510(4): 378-395.
Sienknecht, U. J. and D. M. Fekete (2009). "Mapping of Wnt, Frizzled, and Wnt Inhibitor
Gene Expression Domains in the Avian Otic Primordium." Journal of
Comparative Neurology 517(6): 751-764.
Simmons, D. D. (1994). "A transient afferent innervation of outer hair cells in the
postnatal cochlea." Neuroreport 5(11): 1309-1312.
Smyth, G. K. (2004). "Linear models and empirical bayes methods for assessing
differential expression in microarray experiments." Stat Appl Genet Mol Biol 3:
12.

98

Sollner, C., G. J. Rauch, et al. (2004). "Mutations in cadherin 23 affect tip links in
zebrafish sensory hair cells." Nature 428(6986): 955-959.
Solomon, K. S., S. J. Kwak, et al. (2004). "Genetic interactions underlying otic placode
induction and formation." Dev Dyn 230(3): 419-433.
Sotomayor, M., W. A. Weihofen, et al. (2010). "Structural determinants of cadherin-23
function in hearing and deafness." Neuron 66(1): 85-100.
Srinivas, S., T. Watanabe, et al. (2001). "Cre reporter strains produced by targeted
insertion of EYFP and ECFP into the ROSA26 locus." BMC Dev Biol 1(4): 27.
Stauffer, E. A. and J. R. Holt (2007). "Sensory transduction and adaptation in inner and
outer hair cells of the mouse auditory system." J Neurophysiol 98(6): 3360-3369.
Subramanian, A., P. Tamayo, et al. (2005). "Gene set enrichment analysis: a knowledge-
based approach for interpreting genome-wide expression profiles." Proc Natl
Acad Sci U S A 102(43): 15545-15550.
Tanabe, Y. and T. M. Jessell (1996). "Diversity and pattern in the developing spinal
cord." Science 274(5290): 1115-1123.
Tessier-Lavigne, M. (2002). "Wiring the brain: the logic and molecular mechanisms of
axon guidance and regeneration." Harvey Lect 98: 103-143.
Tritsch, N. X., E. Y. Yi, et al. (2007). "The origin of spontaneous activity in the
developing auditory system." Nature 450(7166): 50-+.
Valenzuela, D. M., A. J. Murphy, et al. (2003). "High-throughput engineering of the
mouse genome coupled with high-resolution expression analysis." Nat Biotechnol
21(6): 652-659.
Vilfan, A. and T. Duke (2003). "Two adaptation processes in auditory hair cells together
can provide an active amplifier." Biophys J 85(1): 191-203.
Vollrath, M. A., K. Y. Kwan, et al. (2007). "The micromachinery of
mechanotransduction in hair cells." Annu Rev Neurosci 30: 339-365.
Wallace, H. A., F. Marques-Kranc, et al. (2007). "Manipulating the mouse genome to
engineer precise functional syntenic replacements with human sequence." Cell
128(1): 197-209.
Wang, K. H., K. Brose, et al. (1999). "Biochemical purification of a mammalian slit
protein as a positive regulator of sensory axon elongation and branching." Cell
96(6): 771-784.
Wang, S. Z., B. H. Liu, et al. (2009). "A genetic strategy for stochastic gene activation
with regulated sparseness (STARS)." PLoS One 4(1): e4200.
Wang, W., M. Warren, et al. (2007). "Induced mitotic recombination of p53 in vivo."
Proc Natl Acad Sci U S A 104(17360553): 4501-4505.

99

Webber, A. and Y. Raz (2006). "Axon guidance cues in auditory development." Anat Rec
A Discov Mol Cell Evol Biol 288(4): 390-396.
Wegner, M. and C. C. Stolt (2005). "From stem cells to neurons and glia: a Soxist's view
of neural development." Trends Neurosci 28(11): 583-588.
Wirth, D., L. Gama-Norton, et al. (2007). "Road to precision: recombinase-based
targeting technologies for genome engineering." Curr Opin Biotechnol 18(5):
411-419.
Wong, K., X. R. Ren, et al. (2001). "Signal transduction in neuronal migration: roles of
GTPase activating proteins and the small GTPase Cdc42 in the Slit-Robo
pathway." Cell 107(2): 209-221.
Woods, C., M. Montcouquiol, et al. (2004). "Math1 regulates development of the sensory
epithelium in the mammalian cochlea." Nature Neuroscience 7(12): 1310-1318.
Wright, T. J. and S. L. Mansour (2003). "Fgf3 and Fgf10 are required for mouse otic
placode induction." Development 130(15): 3379-3390.
Wu, Y. C., A. J. Ricci, et al. (1999). "Two components of transducer adaptation in
auditory hair cells." J Neurophysiol 82(5): 2171-2181.
Xiang, M., W. Q. Gao, et al. (1998). "Requirement for Brn-3c in maturation and survival,
but not in fate determination of inner ear hair cells." Development 125(20): 3935-
3946.
Xiang, M., A. Maklad, et al. (2003). "Brn3c null mutant mice show long-term,
incomplete retention of some afferent inner ear innervation." BMC Neurosci 4: 2.
Xiao, T., T. Roeser, et al. (2005). "A GFP-based genetic screen reveals mutations that
disrupt the architecture of the zebrafish retinotectal projection." Development
132(15930106): 2955-2967.
Xiao, T., W. Staub, et al. (2011). "Assembly of lamina-specific neuronal connections by
slit bound to type IV collagen." Cell 146(1): 164-176.
Xu, Q., G. Mellitzer, et al. (2000). "Roles of Eph receptors and ephrins in segmental
patterning." Philos Trans R Soc Lond B Biol Sci 355(1399): 993-1002.
Yamamoto, N., W. Chang, et al. (2011). "Rbpj regulates development of prosensory cells
in the mammalian inner ear." Dev Biol 353(2): 367-379.
Yamamoto, N., A. Tamada, et al. (2002). "Wiring of the brain by a range of guidance
cues." Prog Neurobiol 68(6): 393-407.
Yang, T., J. Kersigo, et al. (2011). "The molecular basis of making spiral ganglion
neurons and connecting them to hair cells of the organ of Corti." Hear Res 278(1-
2): 21-33.

100

Young, P., L. Qiu, et al. (2008). "Single-neuron labeling with inducible Cre-mediated
knockout in transgenic mice." Nat Neurosci 11(18454144): 721-728.
Ypsilanti, A. R., Y. Zagar, et al. (2010). "Moving away from the midline: new
developments for Slit and Robo." Development 137(12): 1939-1952.
Yuan, W., Y. Rao, et al. (2003). "A genetic model for a central (septum transversum)
congenital diaphragmatic hernia in mice lacking Slit3." Proc Natl Acad Sci U S A
100(9): 5217-5222.
Yuan, W., L. Zhou, et al. (1999). "The mouse SLIT family: secreted ligands for ROBO
expressed in patterns that suggest a role in morphogenesis and axon guidance."
Dev Biol 212(2): 290-306.
Yue, Y., B. Grossmann, et al. (2006). "Isolation and differential expression of two
isoforms of the ROBO2/Robo2 axon guidance receptor gene in humans and
mice." Genomics 88(6): 772-778.
Zambrowicz, B. P., A. Imamoto, et al. (1997). "Disruption of overlapping transcripts in
the ROSA beta geo 26 gene trap strain leads to widespread expression of beta-
galactosidase in mouse embryos and hematopoietic cells." Proc Natl Acad Sci U S
A 94(8): 3789-3794.
Zheng, B., M. Sage, et al. (2000). "Engineering mouse chromosomes with Cre-loxP:
range, efficiency, and somatic applications." Mol Cell Biol 20(10611243): 648-
655.
Zheng, J. L. and W. Q. Gao (2000). "Overexpression of Math1 induces robust production
of extra hair cells in postnatal rat inner ears." Nat Neurosci 3(6): 580-586.
Zheng, J. L. and W. Q. Gao (2000). "Overexpression of Math1 induces robust production
of extra hair cells in postnatal rat inner ears." Nature Neuroscience 3(6): 580-586.
Zheng, W., A. Q. Geng, et al. (2011). "Robo4 Regulates the Radial Migration of
Newborn Neurons in Developing Neocortex." Cereb Cortex 28: 28.
Zine, A., A. Aubert, et al. (2001). "Hes1 and Hes5 activities are required for the normal
development of the hair cells in the mammalian inner ear." J Neurosci 21(13):
4712-4720.
Zipursky, S. L. and J. R. Sanes (2010). "Chemoaffinity revisited: dscams, protocadherins,
and neural circuit assembly." Cell 143(3): 343-353.
Zong, H., J. S. Espinosa, et al. (2005). "Mosaic analysis with double markers in mice."
Cell 121(15882628): 479-492.

Abstract (if available)

Abstract During development, proliferating neuroblasts delaminate from the otocyst to generate spiral ganglion neurons which settled down in Rosenthal’s canal medial to the cochlea sensory epithelium. Subsequently spiral ganglion neurons extend their peripheral axons back to the cochlea by penetrating through the spiral limbus and Kolliker’s organ and ultimately form the tonotopically organized innervations with cochlea hair cells. How spiral ganglion neurons territory is defined along the medial-lateral axis is largely unknown. In this study, we performed microarray analysis and identified Slit molecules as potential candidates involved in this process. Analysis of Slit2 mutant mouse embryos showed that a significant number of spiral ganglion neuron soma are not restrained in Rosenthal’s canal but rather spread randomly over the cochlea tissue, some even go beyond the hair cells. These mis-positioned spiral ganglion neurons extend their merits largely along the longitudinal axis and randomly travel within the cochlea tissue without innervating hair cells. Similar phenotypes are also observed in embryos with mutations for Robo, the receptors for Slit molecules. Furthermore, the spiral ganglion territory is dramatically expanded towards the sensory epithelium in Robo mutant in addition to those individually displaced cells. In situ hybridization showed Robo1 and Robo2 are expressed in spiral ganglion neurons and Slit 2 and Slit3 are expressed in the greater epithelium ridge region (Kolliker’s organ) as well as spiral limbus region during the time spiral ganglion neurons innervate cochlea hair cells. Developmental studies revealed that these SGNs progressively moved to ectopic positions from their normal locations in the Rosenthal’s canal during the time SGNs innervate HCs. We propose that this disruption of spatial patterning of SGNs is attributed to the loss of the restriction force imposed by Slit/Robo signaling, which serves to refrain SGNs from invading the cochlear epithelium and to ensure the formation of precise innervation patterns in the peripheral auditory circuitry. ❧ In the second part of the thesis, we want to dissect the cell lineage relationship between cochlear hair cells and their neighboring supporting cells, which is largely elusive primarily due to the lack of a straightforward genetic method to perform the lineage tracing in the inner ear. To address this need, we developed a method termed STARS: stochastic gene activation with genetically regulated sparseness. The stochastic expression was achieved by two cross-linked, mutually-exclusive Cre-mediated recombinations. The stochastic level was further controlled by regulating Cre/lox reaction kinetics through varying the intrachromosomal distance between the lox sites mediating one of the recombinations. We further explore the possibility to extend the sparseness level by mutagenizing lox sites employed in the STARS transgene. In mammalian cell lines stably transfected with a single copy of different STARS transgenes, the activation/knockout of reporter genes was specifically controlled to occur in from 5% to 50% of the cell population and further down to 1% when combined with lox variants with lower recombination efficiency. STARS can potentially provide a convenient way for genetic labeling as well as gene expression/knockout in a population of cells with a desired sparseness level and hence will provide a useful tool to the dissection of the cell lineage in the inner ear.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Axon guidance cues in development of the mammalian auditory circuit

PDF

Physiology of the inner ear: the role of the biophysical properties of spiral ganglion neurons in encoding sound intensity information at the auditory nerve

Asset Metadata

Creator Wang, Shengzhi (author)

Core Title Slit/Robo signaling underlies the spatial patterning of spiral ganglion neurons to shape the peripheral auditory circuitry assembly

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Physiology and Biophysics

Publication Date 01/16/2013

Defense Date 12/10/2012

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag cochlea,hair cell,innervation,OAI-PMH Harvest,Robo,single neuron labeling,Slit,spatial patterning,spiral ganglion,Stars

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Zhang, Li I. (committee chair), Chen, Jeannie (committee member), Chow, Robert (committee member)

Creator Email shengzhi.wang@gmail.com,shengzhw@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-126064

Unique identifier UC11292137

Identifier usctheses-c3-126064 (legacy record id)

Legacy Identifier etd-WangShengz-1395.pdf

Dmrecord 126064

Document Type Dissertation

Rights Wang, Shengzhi

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA