Hedgehog Signaling in Anterior Development of the Mammalian Embryo
by
Chandra Lynn Clarke Davenport
Department of Cell Biology
Duke University
Date:_______________________
Approved:
___________________________
John Klingensmith, Supervisor
___________________________
Blanche Capel, Chair
___________________________
Terry Lechler
___________________________
Mary Hutson
___________________________
Fan Wang
Dissertation submitted in partial fulfillment of
the requirements for the degree of Doctor
of Philosophy in the Department of
Cell Biology in the Graduate School
of Duke University
2013
i
v
ABSTRACT
Hedgehog Signaling in Anterior Development of the Mammalian Embryo
by
Chandra Lynn Clarke Davenport
Department of Cell Biology
Duke University
Date:_______________________
Approved:
___________________________
John Klingensmith, Supervisor
___________________________
Blanche Capel, Chair
___________________________
Terry Lechler
___________________________
Mary Hutson
___________________________
Fan Wang
An abstract of a dissertation submitted in partial
fulfillment of the requirements for the degree
of Doctor of Philosophy in the Department of
Cell Biology in the Graduate School of
Duke University
2013
i
v
Copyright by
Chandra Lynn Clarke Davenport
2013
Abstract
Sonic hedgehog (Shh) is a critical secreted signaling molecule that regulates
many aspects of organogenesis. In the absence of Shh, many organs, including the
foregut, larynx, palate, cerebellum and heart do not form properly. However, the
cellular details of the roles of Shh, including the relevant domains of Shh expression and
reception, have not been elucidated for many of these processes.
The single embryonic foregut tube must divide into the trachea and esophagus,
which does not occur in the Shh-null mutant. In Chapter 5, I use Cre-Lox technology to
determine that the ventral foregut endoderm is the relevant source of Shh for this
process and the mesoderm must directly receive that Shh signal. Surprisingly, this
signaling event appears to occur two days before the foregut begins to divide, indicating
an early essential role for Shh in foregut division.
Shh is also expressed at later stages in the maturing trachea and esophagus. In
Chapter 6, I demonstrate that these domains serve to establish differentiated mesoderm.
In the trachea, Shh from the endoderm signals directly to the mesoderm to form the
tracheal cartilage rings. In the esophagus, the roles of Shh are more complex. Shh
regulates the size of the esophagus and controls patterning of the concentric rings of
esophageal mesoderm, however this process seems to be indirect, requiring autocrine
Shh signaling within the esophageal endoderm.
iv
The laryngeal apparatus is entirely absent in the Shh-null mouse. I n Chapter 3, I
dissect the domains of Shh expression and reception required for laryngeal development
and demonstrate that loss of endodermal Shh expression causes
laryngotracheoesophageal clefts and malformed laryngeal cartilages. As much of
laryngeal morphogenesis poorly understood, I also utilize dual mesodermal and neural
crest fate maps to determine the embryonic origins of various laryngeal tissues. Finally,
as Shh signaling often occurs in concert with Bone Morphogenic Protein (BMP)
signaling, I investigate the roles of BMP signaling in laryngeal development.
Much of Shh signaling occurs at the primary cilium, to which Smoothened, a
critical pathway member, must translocate upon Shh signal transduction. This process
requires a Smo-Kif3a-βarretin complex in mammalian cell culture. However, the roles of
βarrestins in mouse development, and their relationship to Shh signaling have not been
investigated in vivo. To do so, in Chapter 4, I analyze the phenotypes of the βarr1/βarr2
double knockout embryos and demonstrate that they have palatal, cerebellar,
cardiovascular and renal defects consistent with a specific impairment of mitogenic Shh
signaling.
Altogether, my work dissects the cellular details of Shh signaling during multiple
organ systems in the mouse embryo. I further analyze the consequences of absent or
misregulated Shh signaling across multiple developmental contexts and determine that
Shh plays critical and diverse roles in organogenesis.
v
Dedication
This dissertation is dedicated to my family. To my parents, who instilled in me a
love of learning and discovery from a young age. And, to my wonderful husband, Sam,
whose endless love and support made this work possible.
vi
Contents
Abstract .........................................................................................................................................iv
List of Tables ................................................................................................................................. xi
List of Figures ..............................................................................................................................xii
List of Abbreviations .................................................................................................................. xv
Acknowledgements ................................................................................................................. xvii
1. Introduction .............................................................................................................................. 1
1.1 The mouse as a model for organogenesis .................................................................... 2
1.2 Molecular regulation of Hedgehog signaling and the link to βarrestins ................ 4
1.3 Established roles for Shh signaling in organogenesis ................................................ 8
1.4 Cre-Lox and CreER-Lox approach .............................................................................. 12
2. Materials and Methods.......................................................................................................... 18
2.1 Mouse strains and genotyping .................................................................................... 18
2.2 Tamoxifen and BrdU administration.......................................................................... 22
2.3 Tissue preparation and histology................................................................................ 22
2.4 Immunohistochemistry................................................................................................. 23
2.5 Whole-mount E-cadherin staining .............................................................................. 24
2.6 Bone and cartilage preparations .................................................................................. 24
2.7 LacZ staining .................................................................................................................. 25
3. Embryonic origins and morphogenesis of the larynx ....................................................... 26
3.1 Summary......................................................................................................................... 26
vii
3.2 Acknowledgements........................................................................................................ 27
3.3 Introduction.................................................................................................................... 27
3.4 Results ............................................................................................................................. 36
3.4.1 Early laryngeal development.................................................................................. 36
3.4.2 Laryngeal compartmentalization ........................................................................... 39
3.4.3 Shh and laryngeal clefting ....................................................................................... 46
3.4.4 BMP and laryngeal clefting ..................................................................................... 49
3.4.5 Laryngeal cartilage development and deformities .............................................. 52
3.4.6 BMP in laryngeal cartilage development .............................................................. 60
3.4.7 Shh and laryngeal cartilage ..................................................................................... 63
3.4.8 Laryngeal musculature, epithelia, and connective tissue development........... 66
3.4.9 Recanalization of the larynx.................................................................................... 68
3.4.10 Maturation of the larynx....................................................................................... 71
3.5 Discussion and future directions................................................................................. 73
4. βarrestins and Shh in the developing embryo .................................................................... 76
4.1 Summary......................................................................................................................... 76
4.2 Acknowledgements........................................................................................................ 77
4.3 Introduction.................................................................................................................... 77
4.4 Results ............................................................................................................................. 81
4.4.1 βarrestins are mostly dispensable for murine development .............................. 81
4.4.2 βarr DKO pups exhibit respiratory distress despite having normal lungs ...... 85
4.4.3 Cleft secondary palate is responsible for most βarr DKO lethality................... 87
viii
4.4.4 Barr DKO cleft palate corresponds to decreased Shh signaling ........................ 92
4.4.5 βarr DKO pups also exhibit cardiovascular, skeletal, and renal defects .......... 95
4.4.6 Underfoliation of βarr DKO cerebella is due to reduced Shh-stimulated
proliferation......................................................................................................................... 99
4.4.7 All βarr DKO phenotypes may be due to reduced mitogenic Shh signaling 102
4.5 Discussion and future directions............................................................................... 102
5. Roles for Shh in foregut morphogenesis........................................................................... 107
5.1 Summary....................................................................................................................... 107
5.2 Introduction.................................................................................................................. 107
5.3 Results ........................................................................................................................... 111
5.3.1 Shh expression from the ventral FGE is not required during foregut division
............................................................................................................................................. 111
5.3.2 The dorsal-to-ventral shift in endodermal Shh expression is not required for
foregut division................................................................................................................. 113
5.3.3 Shh from the floorplate is not required for foregut division ............................ 116
5.3.4 Shh expression from the notochord is not required during foregut division 118
5.3.5 A decrease in Shh expression at E8.5 produces broncho-esophageal fistulas123
5.3.6 Hh reception within the FGE is not required for foregut division ................... 126
5.3.6 Hh reception within the FGE is not required for foregut division ................... 126
5.3.7 Hh reception is not required in the ventral foregut mesoderm after E9.5 ..... 128
5.4 Discussion and future directions................................................................................ 131
6. Roles for Shh in development of foregut mesoderm ...................................................... 133
6.1 Summary....................................................................................................................... 133
ix
6.2 Introduction.................................................................................................................. 133
6.3 Results ............................................................................................................................ 134
6.3.1 Shh is required for esophageal size and patterning .......................................... 134
6.3.2 Ablation of autocrine Shh signaling within the esophageal endoderm causes
esophageal dilation .......................................................................................................... 137
6.3.3 Overactivation of autocrine Hh signaling within the esophageal endoderm
produces submucosal vascular defects ......................................................................... 140
6.3.4 Shh expression from the ventral FGE is required for tracheal cartilage
development ..................................................................................................................... 143
6.3.6 Shh signals directly to the foregut mesoderm for tracheal cartilage
development ..................................................................................................................... 145
6.4 Discussion and future directions............................................................................... 147
7 Discussion .............................................................................................................................. 151
7.1 Roles for Shh in foregut development ...................................................................... 151
7.2 Shh in laryngeal compartmentalization ................................................................... 153
7.3 Origin of the thyroid cartilage ................................................................................... 154
7. 4 Noggin and homeotic transformations in laryngeal cartilages............................ 154
7.5 Models of tracheomalacia........................................................................................... 155
7.6 Secondary phenotypes of βarr DKO embryos ........................................................ 155
7.7 Links between Hh signaling and βarrestins in the embryo .................................. 158
References .................................................................................................................................. 162
Biography ................................................................................................................................... 178
x
List of Tables
Table 1: Cre recombination patterns ....................................................................................... 17
Table 2: Mouse strains ............................................................................................................... 19
Table 3: Genotyping primers .................................................................................................... 20
Table 4: Antibodies used for immunohistochemistry........................................................... 23
Table 5: Temporal ablations of Shh ....................................................................................... 124
xi
List of Figures
Figure 1: Regulation of Shh signaling at the primary cilium ................................................. 7
Figure 2: Shh-null phenotype ..................................................................................................... 9
Figure 3: Cre-Lox genetic technology ..................................................................................... 14
Figure 4: Anatomy of the larynx .............................................................................................. 30
Figure 5: Classes of laryngotracheoesophageal clefts ........................................................... 35
Figure 6: Neural crest and mesoderm fate maps ................................................................... 38
Figure 7: Laryngeal compartmentalization at E11.5 ............................................................. 41
Figure 8: Laryngeal compartmentalization ............................................................................ 44
Figure 9: Shh and Nog expression in the E12.5 larynx ........................................................ 45
Figure 10: Shh in laryngeal clefting ......................................................................................... 48
Figure 11: BMP in laryngeal clefting ....................................................................................... 51
Figure 12: Complementary Wnt1-Cre and Mef2C-Cre lineage traces ................................ 56
Figure 13: Development of laryngeal cartilages .................................................................... 59
Figure 14: Noggin and laryngeal cartilage ............................................................................. 62
Figure 15: Shh and laryngeal cartilage .................................................................................... 65
Figure 16: Recanalization of the infraglottis........................................................................... 70
Figure 17: Barr DKO verification ............................................................................................. 83
Figure 18: βarrestins are not required for embryonic viability ........................................... 84
Figure 19: Normal histology and differentiation of the Barr DKO lung ............................ 86
Figure 20: Postnatal Barr DKO lethality is due to cleft secondary palate .......................... 89
xii
Figure 21: Hypoplastic and dysmorphic palatal shelves in the Barr DKO ....................... 91
Figure 22: Shh signaling is decreased during development of the Barr DKO palate ....... 94
Figure 23: Cardiovascular and skeletal defects in Barr DKO pups..................................... 97
Figure 24: Renal defects in Barr DKO pups............................................................................ 98
Figure 25: Shh induced cerebellar neurogenesis is diminished in Barr DKO pups ....... 101
Figure 26: The Shh-null foregut tube does not divide into trachea and esophagus ....... 108
Figure 27: Domains of Shh expression in and around the developing foregut .............. 110
Figure 28: Ablation of Shh from the vFGE does not prevent foregut division ............... 112
Figure 29: Constitutive activation of Hh signaling does not impair foregut division ... 115
Figure 30: Sox17-iCre; Shhflox/null have and undivided foregut but also notochord
recombination ............................................................................................................................ 117
Figure 31: Ablation of Shh from the notochord and floor plate ........................................ 119
Figure 32: 50% of Shh-null type have an undivided foregut ............................................. 121
Figure 33: The single remaining spatiotemporal domain of Shh expression required for
foregut division is the early vFGE .......................................................................................... 122
Figure 34: ShhCreER/flox with tamoxifen at E8.5 produces broncho-esophageal fistulas .... 125
Figure 35: Autocrine endodermal Shh signaling is not required for foregut division ... 127
Figure 36: Shh reception in the ventral foregut mesoderm is not required for foregut
compartmentalization after E9.5 ............................................................................................. 130
Figure 37: Shh expression in the developing esophagus regulates esophageal size and
patterning ................................................................................................................................... 136
Figure 38: Ablation of autocrine Shh signaling within the foregut endoderm causes
dilation of the esophagus ......................................................................................................... 139
xiii
Figure 39: Over-activation of Hh signaling within the foregut endoderm causes a
reduction of the presumptive esophageal vasculature ........................................................ 142
Figure 40: Shh expression from the vFGE is required for tracheal cartilage development
..................................................................................................................................................... 144
Figure 41: Shh signals directly to ventral foregut mesoderm for tracheal cartilage
development .............................................................................................................................. 146
xiv
List of Abbreviations
βarr
β-arrestin
BEF
Broncho-Esophageal Fistula
BMP
Bone Morphogenic Protein
BMPr1a
BMP Receptor 1A
BrDU
Bromodeoxyuridine
CA
Constitutively Active
dFGE
Dorsal Foregut Endoderm
DAPI
4',6-Diamidino-2-Phenylindole
DKO
Double Knockout
E
Embryonic day
EGL
External Germinal Layer
EL
Epithelial Lamina
ER
Estrogen Receptor
FGE
Foregut Endoderm
GCP
Granule Cell Precursor
GRK
G Protein-Coupled Receptor Kinase
H&E
Hematoxylin and Eosin
Hh
Hedgehog
xv
LE
Laryngeal Endoderm
LTS
Laryngotracheal Sulcus
NCC
Neural Crest Cells
NFP
Notochord and Floor Plate
Nog
Noggin
PAS
Periodic Acid-Schiff
PBST
Phosphate-Buffered Saline with Tween
PPS
Primitive Pulmonary Sac
Ptch
Patched
R26R
Rosa26-Reporter
Shh
Sonic Hedgehog
Smo
Smoothened
vFGE
Ventral Foregut Endoderm
xvi
Acknowledgements
Many people at Duke made this dissertation possible. First and foremost, my
advisor, John Klingensmith, whose support and guidance was critical for this process.
My thesis committee, Blanche Capel, Fan Wang, Terry Lechler, and Mary Hutson,
provided a sounding board for new ideas and helped guide my work.
Former Klingensmith and Meyers lab members, Matt Goddeeris, Laura Custer,
and Andy Ravanelli guided my early graduate career and patiently introduced me to the
world of developmental biology. My current "lab family" of Sarah Fausett, Maiko
Matsui and Katie Kretovich have made working in the lab a fun and productive
environment. Their ideas, helpful criticism, and collaboration have greatly influenced
my thesis work.
The βarrestin work would not have been possible without Bob Lefkowitz and
Jeff Kovacs, who not only provided the mice, but also helped me navigate though the
foreign world of receptor biology.
Last, but certainly not least, none of this would have been possible without
excellent technical support from Kathy Carmody, who lovingly cared for and genotyped
my mice. A heartfelt thanks to all of you.
xvii
1. Introduction
Embryonic development is one of the most tightly orchestrated and complex
processes found in nature. An enormous variety of cell signaling events and molecular
processes must work in concert for the embryo to grow in size, specify tissues, and form
organs. An error in any one of these process can result in congenital malformations,
commonly termed “birth defects.” 3-5% of all births result in congenital malformations
(Robinson and Linden, 1993) and more than 120,000 babies are born with a birth defect
in the United States every year (www.marchofdimes.com). While the complete etiology
of most birth defects is unknown, many are due, at least in part, to underlying genetic
lesions. Mutation of even a single DNA base can at times change the function of a
protein or inactivate a gene entirely. Distressingly often, congenital malformations are
fatal. In fact, 20-30% of all infant deaths are due to genetic disorders (Berry, et al., 1987).
If we hope to impact the frequency, severity, and consequences of birth defects,
we must first understand normal embryonic development. We must not only identify
the genes involved in developmental processes, but also the molecular function of the
proteins and how various genes interact with each other. Furthermore, we need to
identify the tissue(s) in which critical genes are expressed, the timing of that expression,
and, in the case of secreted molecules, the relevant tissue target(s) of the signal.
My thesis work focuses on the genetics underlying normal and abnormal
development of structures in the anterior embryo, particularly the role of the secreted
1
signaling molecule, Sonic hedgehog (Shh). Using the mouse as a model system, I
identify novel, tissue-specific roles for Shh in morphogenesis of the anterior embryo and
analyze the consequences of misregulated or absent Shh signaling. By doing so, I
provide insight into the etiology of multiple birth defects.
1.1 The mouse as a model for organogenesis
In Victorian England, “fancy” mice were widely bred and traded, resulting in a
variety of phenotypic expression. The establishment of genetically pure, inbred mouse
strains by Clarence Cook Little in 1909 heralded the beginning of the house mouse, Mus
musculus, as a genetic model organism. Initially, mouse geneticists were limited to
studying the consequences of spontaneous and induced random mutations. With the
advent of gene targeting, researchers had the ability to make targeted knockout mouse
lines of their gene of interest. Much insight was made into the functions of various
genes during embryonic development by analyzing the consequences of these gene
ablations.
Unfortunately, knockout of many genes results in early lethality or severe
phenotypes which preclude investigation of the roles of these genes in later embryonic
development. Cre-Lox technology (discussed in Section 1.4) has proven to be a valuable
tool in dissecting the finer points of embryonic development. With these new genetic
tools, tissue-specific gene manipulations can be performed, ablating a single domain of
gene expression while leaving gene expression intact in the rest of the developing
2
embryo. Most recently, CreERs have allowed an even greater level of control, as genes
of interest can be manipulate genes not only in embryonic space, but also in
developmental time.
Organogenesis, or the formation of organs in the developing embryo, largely
occurs between the 3rd and 8th week of human gestation. During this period, the heart
forms and begins to beat, the common foregut tube divides into the trachea and
esophagus, and the lungs, liver, kidneys, and the rest of the internal organs develop. My
work has encompassed many of these organ systems. Division of the foregut tube is
described in Section 5.2. Formation and morphogenesis of the larynx is covered in
Section 3.2. Development of the palate, cerebellum, heart, lungs, and kidneys are all
covered in Chapter 4.
There is remarkable similarity between murine and human organ development.
In the mouse, the period of organogenesis corresponds to embryonic day (E) 8.5-14.5.
Nearly all of the same genes and developmental processes important for human organ
development are also used in the mouse. As it is not feasible to directly study
embryonic organ development in humans, the mouse serves as a useful and tractable
substitute model organism.
Organogenesis occurs as a multi-step process, although these phases can
temporally overlap. First, the organ must be specified from a multipotential progenitor
tissue. Typically, a combination of secreted molecules "instructs" this tissue to become a
3
particular organ. That organ must then be patterned, or subdivided, into the multiple
tissue components that comprise the mature organ. Next, morphogenesis of that organ
begins. During this period, the specified tissue changes its shape to resemble the mature
organ. Finally, the organ primordium must both grow in size and differentiate. As
embryos grow rapidly, the tissue must proliferate to keep pace. Differentiation is the
cellular process of becoming a mature tissue. It is during this final stage that individual
cells convert into their final form, such as muscle or cartilage.
1.2 Molecular regulation of Hedgehog signaling and the link to
βarrestins
Hedgehogs (Hhs) are a family of secreted molecules of which there are three
mammalian homologues: Sonic (Shh), Desert, and Indian. Shh is widely expressed in
the developing embryo, primarily in the endoderm. As illustrated in Figure 1, Hhs
signal through Gli transcription factors. In the absence of Hh signaling, the 12transmembrane receptor Patched inhibits 7-transmembrane receptor Smoothened (Smo),
and repressor Glis prevent transcription of Hh target genes. However when Hh ligand
is present, the inhibition of Smo by Patched is relieved, and Smo is phosphorylated by
GRK2 (Chen, et al., 2011). In mammalian cell culture, a β-arrestin 2 (βarr2)-Kif3a
complex translocates active Smo to the primary cilium (Kovacs, et al., 2008). Through a
poorly understood process, repressor Glis are themselves repressed and activator Glis
translocate to the nucleus to transcribe Hh target genes. Patched itself is a Hh target
gene, often serving as a reporter for active Hh signaling.
4
5
Figure 1 – Regulation of Shh signaling at the primary cilium
In the absence of an Hh ligand, Patched inhibits Smoothened and
transcription of Hh target genes does not occur. Binding of a
Sonic Hh to Ptch relieves its inhibition of Smo, which translocates
to the primary cilium via a Barr2-Kif3a complex. Repressor Glis
are then converted to activator Glis that translocate to the nucleus
to transcribe Hh target genes. (Details in text.)
6
Figure 1: Regulation of Shh signaling at the primary cilium
7
1.3 Established roles for Shh signaling in organogenesis
Shh is a critical embryonic signaling molecule, regulating many developmental
processes. Shh plays a role in the node to regulate establishment of the left-right axis
(Tsukui, et al., 1999). Mammals born without Shh exhibit holoprosencephaly, a failure
of midline structures to form. As seen in Figure 2, the Shh-null mouse embryo exhibits
small size, aborted limb development, loss of midline craniofacial structures, and
cyclopia. Internally, Shh-null embryos have an undivided foregut tube, which normally
compartmentalizes into the trachea and esophagus, as discussed in Section 5.2. These
embryos die at birth due to cardiovascular defects: the cardiac outflow tract is not
divided into aorta and pulmonary artery, preventing proper circulation of oxygenated
blood (Washington-Smoak et al., 2005).
In addition to formation of midline structures, Shh has critical functions during
the formation of many organs. Shh is typically well-known for its expression in the limb
bud, where it serves to pattern the digits (Harfe, et al., 2004). Shh is also expressed in the
oral epithelium, which promotes fusion of the palatal shelves to prevent cleft palate
(Cobourne, et al., 2009), discussed in Section 4.3.4. Proliferation and foliation of the
cerebellum, discussed in Section 4.3.6, is also driven by Shh expression from granule cell
precursors (Wechsler-Reya and Scott, 1999). Shh expression from the notochord is
required to pattern the neural tube (Ericson, et al., 1996).
8
Figure 2: Shh-null phenotype
In comparison to a wild-type littermate, the E12.5 Shh-null mouse
embryo is small in size and exhibits holoprosencephaly. Scale bar
indicates 5mm.
9
Additionally, Shh is robustly expressed throughout most of the endoderm
during various stages of embryonic development. As such, Shh expression (or lack
thereof) plays a significant role in the development of endodermal organs. Shh from the
pharyngeal endoderm controls development of the mandible (Melnick, et al., 2005). This
domain also regulates development and branching morphogenesis of the
submandibular gland (Jaskoll, et al., 2004) and proliferation and differentiation of
thymocytes (Outram, et al, 2000). Exclusion of Shh from pharyngeal pouches is critical
for initiation of thyroid development (Fagman, et al, 2004) and positioning of the
parathyroid (Moore-Scott and Manley, 2005). In Chapter 3, I investigate the later roles of
Shh from this tissue, which becomes the laryngeal endoderm.
Shh expression is critical for the division of the common foregut tube into the
trachea/lung and esophagus (Litingtung, et al., 1998). I explore the relevant Shh domain
for this process in Chapter 5. Later in development, Shh is required in the lung
endoderm for branching morphogenesis, to prevent pulmonary cysts (Pepicelli, et al.,
1999). As I demonstrate in Section 6.3.4, development of the tracheal cartilage rings is
also regulated by Shh expression from the endoderm. Section 6.3.1 reveals the function
of Shh in the esophagus, where it plays a key role in regulating both esophageal size and
development of the concentric rings of esophageal smooth muscle.
Rostral endodermal Shh expression also controls development of some nonendodermal organs, such as the heart. Shh expression from the pharyngeal endoderm is
10
important for division of the cardiac outflow tract into the aorta and pulmonary artery
(Washington Smoak, et al.. 2005). Additionally, Shh from the lung endoderm
(Hoffmann, et al., 2009) plays a crucial role in septation of the heart into four chambers
by regulating migration and differentiation of the dorsal mesocardium (Goddeeris, et al.,
2008). Details regarding the roles of endodermal Shh in cardiac development are found
in Section 4.4.
More caudally, endodermal Shh expression in the stomach regulates
proliferation and differentiation of gastric glands and prevents intestinal transformation
of the stomach lining (Ramalho-Santos, et al., 2000). Exclusion of Shh from a small
region of the endoderm is required for proper initiation of the pancreas (Hebrok, et al.,
1998) and the Shh-null mouse has an annular pancreas (Ramalho-Santos, et al., 2000).
Downregulation of Shh expression in the liver is important for cellular proliferation
(Hirose, et al., 2009). Shh is required for formation of the kidney, as the Shh-null mouse
has bilateral renal aplasia (Hu, et al., 2006). Later, Shh indirectly controls kidney
development by regulating proliferation and differentiation of smooth muscle of the
ureter (Yu, et al, 2002), discussed in Section 4.4.
In the small intestine, Shh regulates both proliferation and patterning of the
crypt-villus axis (Madison, et al., 2005) and in the colon, Shh controls gut rotation and
development of smooth muscle (Ramalho-Santos, et al., 2000). Shh is also important for
the formation of smooth muscle in the bladder (Shiroyanagi, et al., 2007). Shh is
11
expressed in the urogenital sinus epithelium where it regulates proliferation,
differentiation, and branching morphogenesis of the prostate (Freestone, et al., 2003).
Genital and rectal development are also controlled by Shh, as the Shh-null mouse
has persistent cloaca and lacks external genitalia (Haraguchi, et al., 2001). The relevant
domain of Shh expression for genital tubercle development is the urethral epithelium
(Perriton, et al., 2002). Inactivation of Shh later in development results in hypospadias
(Seifert, et al., 2009).
Thus, Shh expression and signaling is critical for proper formation of nearly
every organ in the mammalian body. Shh (or its absence) is involved in the specification
of certain organs, such as the pancreas and thyroid. Shh is responsible for patterning of
other organs and structures, like the digits and intestine. For many organs, Shh plays a
key role in morphogenesis, such as branching of the lung and submandibular gland
epithelium. Shh is also a potent mitogen, inducing proliferation of many tissues,
including the cerebellum, prostate, and liver. In other organs, including the foregut
(trachea and esophagus), Shh is essential but its cellular roles have not been elucidated.
1.4 Cre-Lox and CreER-Lox approach
The study of knockout mice has revealed a multitude of roles of Shh during
development. However, the gross severity of the Shh-null phenotype has precluded
investigation into many later roles of Shh during organogenesis and maturation. By
utilizing a Cre-Lox approach, I dissect the roles of Shh in anterior organogenesis.
12
Specifically, I determine where and when the Shh signal is produced and received
during various developmental processes.
In Cre-Lox genetics, a transgenic mouse expressing bacterial Cre recombinase
under a tissue-specific promoter is bred to a second transgenic line with "floxed" alleles
of the gene of interest (Figure 3). Critical portions of these "floxed" genes are flanked by
LoxP sites, 34 base pair recognition sequences of Cre recombinase. When a cell normally
expresses that tissue-specific promoter, Cre recombinase is produced, enters the nucleus,
and binds to its LoxP recognition sequences. It then recombines out the critical region of
the gene of interest, deleting all DNA sequence between the two LoxP sites. Now, this
cell is a "knockout" and no longer expresses the gene of interest. As this is a permanent
genetic change, all descendents of this cell will also be knockouts. Nearby cells that
don't express Cre retain function of the gene of interest. For example, if a mouse which
expressed Cre specifically in the endoderm were crossed to a Shh-floxed mouse, mutant
progeny would lack all Shh expression from the endoderm.
13
Figure 3: Cre-Lox genetic technology
In Cre-Lox genetics, a mouse expressing Cre recombinase is mated
to a floxed mouse. When expressed, Cre recombines between the
LoxP sites, deleting the intervening DNA sequence. In nearby
cells not expressing Cre, gene function remains intact, generating
a tissue-specific knockout (details in text).
14
For additional fine-tune control of gene ablations, genes of interest can be
temporally manipulated with CreERs. CreER mice also express Cre under a tissuespecific promoter. However, that Cre is altered by conjugation to a mutant version of
the estrogen receptor (ER), which responds to tamoxifen, an estrogen receptor
antagonist. In the absence of tamoxifen, the ER sequesters Cre in the cytoplasm. Only
upon tamoxifen administration does the ER translocate to the nucleus, allowing Cre to
recombine the LoxP sites. This allows temporal ablation of genes by altering the timing
of tamoxifen dosage to the pregnant dam. For example, if a mouse carrying an
endodermal-specific CreER were crossed to a Shh-floxed mouse, and the pregnant dam
was injected with tamoxifen at E8.5, the resulting mutant progeny would lack Shh
expression in the endoderm after E9.5 (allowing approximately 24 hours for
recombination).
While Cre and CreER strains are commonly used to ablate genes, some genes can
be over-activated, or ectopically expressed, using Cre transgenics. These mice have a
"flox-stop-flox" sequence (a stop codon flanked by two LoxP sties) immediately
preceding the coding sequence for a mutant, activated version of a gene of interest. For
example, the flox-stop-flox CA-Smo mouse expresses a constitutively active version of
Smo that is immune to inhibition by Patched. If this mouse is crossed to an endodermspecific Cre mouse, mutant progeny will have ectopic Hh signaling specifically in the
endoderm, even when a Hh ligand is not present.
15
Additionally, Cre-Lox and CreER-Lox genetics can be used as a tool for longterm fate mapping of individual cell populations. In this technique the R26-reporter
(R26R) mouse is crossed to a Cre line to label all cells expressing Cre. The R26R mouse
contains a "flox-stop-flox" sequence preceding the gene for β-galactosidase, all inserted
under the control of a ubiquitous promoter. Thus, all cells that express Cre under a
tissue-specific promoter, and their cellular descendents, express β-gal in that specific
tissue. The presence of β-gal can be revealed by LacZ staining, in which X-gal (5-bromo4-chloro-indolyl-β-D-galactopyranoside ) is hydrolized to create an insoluble blue
compound. As an example, if a neural crest-specific Cre (Wnt1-Cre) were crossed to the
R26R reporter and tissues were LacZ stained, all neural crest cells and their derivatives
would be labeled blue. As this is a permanent genetic change, this technique allows one
to follow the fate of a particular group of cells throughout development.
In Chapters 3 and 5, I utilize many different Cre recombinase-containing mouse
strains. For reference, Table 1 lists the relevant recombination patterns and time of
recombination for the Cre strains used.
16
Table 1: Cre recombination patterns
Foxa2-NFP-Cre
Foxg1-Cre
Mef2C-Cre
Nkx2.5-Cre
Shh-GFP-Cre
Shh-CreER
Sox17-iCre
Wnt1-Cre
Notochord (E7.5) and floorplate
(E8.0). Ectopic recombination in
endoderm.
Foregut endoderm and mesoderm
from E8.0.
Ectopic recombination in
notochord.
Ventral foregut mesoderm from
E9.5.
Ventral foregut endoderm and
mesoderm from E9.5.
Notochord and floorplate (E7.5 and
E8.0). Ventral (E8.5) and dorsal
(E11.0) foregut endoderm.
Notochord and floorplate (E7.5 and
E8.0). Ventral (E8.5) and dorsal
(E11.0) foregut endoderm.
Foregut endoderm from E8.0.
Ectopic recombination in
notochord.
Neural crest cells (E8.0).
17
2. Materials and Methods
2.1 Mouse strains and genotyping
Table 2 lists the mouse strains and sources used. Mice were housed and bred
according to Duke IACUC regulations. Embryonic day (E) 0.5 designated as day of
vaginal plug. Pregnant dams were euthanized by cervical location and embryonic
tissues rapidly harvested on ice.
DNA was isolated from embryo tails and/or caudal tissues via standard ethanol
precipitation. Genotyping was performed using standard PCR protocols with the
primers listed in Table 3. In all Figures, "wild-type" indicates non-mutant littermates,
confirmed by PCR genotyping.
18
Table 2: Mouse strains
Strain
Foxa2-NFP-Cre
Foxg1-Cre
Mef2C-Cre
Nkx2.5-Cre
Shh-GFP-Cre
Shh-CreER
Sox17-iCre
Wnt1-Cre
R26R
Shh-flox
Shh-null
Smo-flox
Smo-null
CA-Smo
Ptch-LacZ
Nog-flox
Nog-null
CA-BMPr1a
Barr1-null
Barr2-null
Formal nomenclature
Tg(Foxa2*-cre)#Mku
Foxg1tm1(cre)Skm
Tg(Mef2c-cre)2Blk
Nkx2-5tm1(cre)Rjs
Shhtm1(EGFP/cre)Cjt
Shhtm2(cre/ERT2)Cjt
Sox17tm2.1(icre)Heli
Tg(Wnt1-cre)11Rth
Gt(Rosa)26Sortm1Sor
Shhtm2Amc
ShhTm1Chg
Smotm2Amc
Smotm3Mey
Gt(ROSA)26Sortm1(Smo/YFP)Amc
Ptch1TmxMS
Nogtm1.1Rmh
N/A
Gt(ROSA)26Sortm1(CAG-
Reference
Kumar, et al., 2007
Hebert and McConnell, 2000
Verzi, et al., 2005
Moses, et al., 2001
Harfe, et al., 2004
Harfe, et al., 2004
Engert, et al., 2009
Jax #003829
Jax #003474
Jax #004293
Chiang, et al., 2006
Jax #004288
Goddeeris, et al., 2007
Jax #005130
Jax #003081
Stafford, et al., 2011
Unpublished (S. Fausett)
Bmpr1a)Que
Rodriguez, et al., 2010
Arrb1tm1Jse
Arrb2tm1Rjl
Conner, et al., 1997
Bohn, et al., 1999
19
Table 3: Genotyping primers
Cre
Shh-GFP-Cre
Foxg1-Cre
Sox17-iCre
Wnt1-Cre
R26R
Shh-flox
Smo-flox
CA-Smo
Ptch-LacZ
Nog-LacZ
Nog-null
Nog-flox
CA-BMPr1a
5’-CGGTGAACGTGCAAAACAGG-3’
5’-CCCCAGAAATGCCAGATTACG-3’
5’-CTCGGCTACGTTGGGAATAA-3’
5’-GGGACAGCTCACAAGTCCTC-3’
5’-GGTGCGCTCCTGGACGTA-3’
5’-TCCTATAAGTTGAATGGTATTTTG-3’
5’-AGTATTGTTTTGCCAAGTTCTAAT-3’
5’-TCACGAAGCACTTGTTGAGG-3’
5’-GAACGGCAGTACGAGAAGC-3’
5’-GTGTATAAGCCCGAGATGG-3’
5’-CTCAACTGTTCAAGTGGCAG-3’
5’-GATCTATGGTGCCAAGGATGAC-3’
5’-GCGGTCTGGCAGTAAAAACTATC-3’
5’-GTGAAACAGGATTGCTGTCACTT-3’
5’-GGAGCGGGAGAAATGGATATG-3’
5’-AAAGTCGCTCTGAGTTGTTAT-3’
5’-GCGAAGAGTTTGTCCTCAACC-3’
5’-ATGCTGGCTCGCCTGGCTGTGGAA-3’
5’-GAAGAGATCAAGGCAAGCTCTGGC-3’
5’-CCACTGCGAGCCTTTGCGCTAC-3’
5’-CCCATCACCTCCGCGTGCCA-3’
5’-GGAGCGGGAGAAATGGATATG-3’
5’-TGCTCAGGTAGTGGTTGTCG-3’
5’-CGTGATCTGCAACTCCAGTC-3’
5’-AAGTTCATCTGCACCACCG-3’
5’-GTTTTCCCAGTCACGACGTT-3’
5’-CAGCAGCGGGTCTGTCAC-3’
5’-GCATGGAGCGCTGCCCCAAGC-3’
5’-GAGCAGCGAGCGCAGCAGCG-3’
5’-AAGGGCGATCGGTGCGGGCC-3’
5’-GAGTGCCGCTGAAACTTGACC-3’
5’-CGCTGGAGTAATCTCGGATGC-3’
5’-CCATTTTGATTTCCCCTCCCTT-3’
5’-ACAGAGAAACAAGACGCC-3’
5’-TTGTTAACCCTTTGCACGCC-3’
5’-GTTGGTGATTTGTACCCAGA-3’
5’-GGTATGTAACGCGGAACTCC-3’
5’-CACTTGCTCTCCCAAAGTCG-3’
20
Shh-null
Smo-null (LacZ)
Barr1-/-; Barr2+/-
5’-TAGTCTAACTCGCGACACTG-3’
5’-AAATAGGAGACAGCCTCGGAA-3’
5’-GTTCCACATACACTTCATTCTCAG-3’
5’-GACCATGTCTGCACACTTAGGTTCC-3’
5’-GAAGGCCAGGAGGAGAAGGCTCAC-3’
5’-ATCCTCTGCATGGTCAGGTC-3’
5’-CGTGGCCTGATTCATTCC-3’
5’-GCTAAAGCGCATGCTCCAGA-3’
5’-GATCAAAGCCCTCGATGATC-3’
5’-ACAGGGTCCACTTTGTCCA-3’
21
2.2 Tamoxifen and BrdU administration
Tamoxifen suspended in corn oil at 20 mg/ml was injected intraperitoneally at
0.5-1.0 mg per gram body weight as a single dose. For BrdU labeling, pregnant dams
were injected intraperitoneally with 50ug/g body weight BrdU. Labeling proceeded for
60 minutes before embryos were harvested.
2.3 Tissue preparation and histology
Embryonic tissues or whole embryos were dissected in PBST and fixed in 4%
paraformaldehyde from 2 hours to overnight. Paraffin samples were dehydrated
through an alcohol series, cleared in xylene, and embedded in paraffin for sectioning at
10um. Frozen samples were passed through a sucrose gradient to 30% sucrose then
equilibrated overnight in 1:1 30% sucrose:OCT at 20C. The following day, samples were
equilibrated in 1:3 30% sucrose:OCT for one hour and then embedded in 100% OCT.
Frozen samples were stored at -80C and cryosectioned at 8-12um.
For histological staining, slides were dewaxed in xylene and rehydrated through
a graded alcohol series. Hematoxylin and Eosin staining was performed using standard
procedures. Histological Alcian Blue staining was performed by incubating slides in 10
mg/ml Alcian Blue 8GX in 3% acetic acid for 20-40 minutes. For PAS-Schiff staining,
slides were oxidized in 0.5% periodic acid for 5 minutes and stained with Schiff reagent
for 15-20 minutes.
22
2.4 Immunohistochemistry
Antibody staining was performed using standard procedures. Antigen retrieval
was performed by boiling slides in 0.1M Sodium Citrate for 10 minutes and then slides
were cooled to room temperature. Slides were blocked in "modified Hogan block"
(MHB: 1%BSA, 5% fetal bovine serum, 1% normal donkey serum in PBST) and primary
antibodies were used at the following concentrations in MHB at 4C overnight.
Table 4: Antibodies used for immunohistochemistry
Antigen
Conc.
Supplier
Species
Β-galactosidase
1:5,000
Cappel
Rabbit
Cleaved
caspase 3
E-cadherin
1:1000
Accurate
Chem.
Cell
Signaling
Zymed
GFP
1:500
AbCam
Rabbit
α-Smooth
muscle actin
1:500
Sigma
Mouse
Sox9
1:1000
Chemicon
Rabbit
BrdU
1:500
1:500
Rat
Rabbit
Rat
Secondary
Donkey αrabbit AF594
Goat α-rat Cy5
Donkey αrabbit AF594
Goat α-rat Cy5
Donkey αrabbit AF594
Donkey αmouse AF488
Donkey αrabbit AF594
Supplier
Invitrogen
Jackson
Invitrogen
Jackson
Invitrogen
Invitrogen
Invitrogen
Slides were incubated with AlexaFluor-conjugated secondary antibodies (1:500 in MHB)
for 2 hours at room temperature and nuclei were counterstained with DAPI. Slides
were imaged on an a basic fluorescent microscope and images edited in ImageJ and
Picassa
23
2.5 Whole-mount E-cadherin staining
For visualization of the tracheal and esophageal lumens, foreguts were
microdissected from E12.5-E14.5 embryos and fixed in 4% paraformaldehyde for 15-30
minutes, rocking at 4C. Tissues were then permeabilized for 15-30 minutes in PBSTx
with 1.0% Triton, rocking at 4C and blocked in PBSTx + 5% FBS 2 hours to overnight.
Rat anti-E-cadherin was added to samples at 1:1000 and incubated rocking at 4C
overnight. Peroxidase-conjugated donkey anti-rat secondary antibody was used at
1:2000 for 2 hours rocking at 4C. After washing off excess antibody, samples were
developed in 1X DAB in Stable Peroxidase Buffer. After 3 minutes, staining was
stopped by washing samples with excess PBS. Foreguts were immediately imaged or
post-fixed in 4% PFA overnight for storage.
2.6 Bone and cartilage preparations
Late stage embryos were dissected, removing skin and abdominal organs. For
tracheal cartilage staining, the rib cage and clavicle were opened. Samples were fized in
95% EtOH at room temperature overnight, stained in Alcian Blue solution (100ml acetic
acid, 400ml 95% EtOH, 75mg Alcian Blue) for 2-3 days at room temperature. Samples
were next incubated in 2% KOH overnight, then stained with Alizarin red solution (5g
KOH, 37.5 mg Alizarin Red in 500ml dH20) for 1-2 days at room temperature. Embryos
were cleared in 1% KOH in 20% glycerol, changed daily, for two weeks and stored in 1:1
EtOH:glycerol.
24
2.7 LacZ staining
Embryos or isolated organs were fixed in LacZ fix (4% PFA, 5mM EGTA, 2mM
MgCl2 in 0.1M sodium phosphate buffer) for 30 minutes to 2 hours, depending on
sample size, and then washed in LacZ wash buffer (0.1M sodium phosphate with 2mM
MgCl2, 0.01% Deoxycholate, 0.02% NP-40). Large embryos were permeabilized in
PBSTx with 1% Triton for 30-60 minutes. Next, samples were incubated in LacZ Stain
(1X K-ferro, 1X K-ferri, 10ul/ml X-gal or Blu-o-gal in LacZ wash buffer) at 37C overnight,
protected from light. Stained samples were post-fixed in 4% PFA overnight and
photographed or processed for histology.
25
3. Embryonic origins and morphogenesis of the larynx
3.1 Summary
The larynx is a structurally complex organ in the neck that houses the vocal
cords, or true vocal folds; it is commonly known as the voice box. As the connection
between the mouth and the trachea, the larynx is essential for breathing, vocalization,
and protection against food aspiration. Congenital defects of the larynx are relatively
common in humans, occurring at approximately 1 in 2000 births (McIntosh et al., 1954).
Despite this great clinical significance, little is known about the etiology of laryngeal
defects. In this chapter, I review the normal anatomy of the larynx, congenital defects of
the larynx, and existing mouse models of laryngeal malformations. Multiple aspects of
normal embryonic laryngeal development remain unclear or controversial in the
literature. As such, I employed lineage-mapping tools in the mouse to resolve some of
these long-standing questions. I establish the position of the rostral trachea in the early
embryo and map the tissue origins of laryngeal cartilages. I also investigate the role of
two key developmental signaling pathways, activated by Hedgehog or Bone
Morphogenetic Protein ligands, in laryngeal morphogenesis. Tissue-specific genetic
manipulation of these pathways recapitulates human laryngeal birth defects, and
indicates key tissue sources and targets for signaling via these pathways. These studies
provide new mouse models for the molecular cell signaling events underlying normal
and defective formation of the larynx.
26
3.2 Acknowledgements
The work in this chapter was performed in collaboration with Sarah Fausett and
Maiko Matsui in my lab. In particular, Sarah contributed to Figures 8, 9, and 11 and
Maiko's images are found in Figure 14. All are used with their permission.
3.3 Introduction
The larynx is a complex cartilaginous box which houses the true vocal folds
(vocal cords), twin infoldings of mucous membrane that produce noise upon
vocalization. The vocal folds vibrate, as air is expelled from the lungs during phonation.
Low frequency vibrations produce lower tones while faster vibrations produce higher
tones. Mammals have great vocal versatility as the length, tension, and shape of the
vocal folds can all be modulated by the laryngeal musculature.
To understand the etiology of congenital defects of the larynx, one must first
have a basic understanding of normal laryngeal structure. The hyoid bone, while not
officially part of the larynx, supports the top of the thyroid cartilage, coupling it to the
tongue (Figure 4A). The thyroid cartilage is the largest component of the larynx,
forming the "Adam's apple." Caudally, the laryngeal “box” is continued with the cricoid
cartilage, which forms a complete ring around the larynx. Sitting atop the dorsal cricoid
cartilage are the smaller paired arytenoid, corniculate, and cuneiform cartilages. The
vocal folds are formed from the free borders of the conus elasticus membrane, attaching
dorsally to the arytenoid cartilages and ventrally to the thyroid cartilage. The outer
27
edges of the vocal folds are attached to laryngeal muscles while the inner edges are free
to vibrate. The glottis is defined as the true vocal folds and the space in between (also
referred to as the rima glottidis) (Figure 4B).
28
Figure 4: Anatomy of the larynx
A) Schematic model of the larynx and individual laryngeal
cartilages. Connective tissue and muscles are colored pink. The
level of the vocal cords (glottis) is indicated by a red dotted line.
B) Cross-section of the larynx at the level of the glottis, viewed
from above. The vocal folds (bright red) stretch horizontally
across the larynx. The lateral aspects of the thyroarytenoid muscle
(dark red) have been removed for visualization of the subglottic
structures.
29
Figure 4: Anatomy of the larynx
30
The interior of the larynx is regionally defined by relationship to the glottis, the space
between the vocal folds; it is divided into supraglottic, glottic, and subglottic sections.
Cranially, the supraglottis begins with the epiglottis, a cartilaginous flap that guards the
entrance to the larynx. During swallowing, the larynx is pulled upwards and the tongue
forces the epiglottis down to cover the glottis. As a result, ingested solids and fluids are
prevented from entering the trachea. The subglottis, comprised mostly of the cricoid
cartilage and its lumen, is the narrowest part of the respiratory system and thus
regulates intrapulmonary pressure.
Tetrapods have diverse laryngeal structures, with mammalian larynges being the
most complex. The most primitive larynx is that of the lungfish, comprised of a simple
muscular sphincter in the floor of the pharynx. Of animals with laryngeal cartilage,
salamanders have only arytenoids and frogs have a cricoid and arytenoids but lack vocal
folds. Most reptiles have an annular thyrocricoid cartilage and arytenoid cartilages, but
only geckos and chameleons have true vocal folds. Some species, including crocodilians,
have a flap of mucous membrane that serves as an epiglottic analogue, but only
mammals have a true cartilaginous epiglottis. Moreover, only mammals have distinct
thyroid and cricoid cartilages, creating a mobile crico-thyroid joint that permits
additional modulation of vocal fold length and tension. (Reviewed by Kirchner, 1993.)
Rather than vocal folds, birds have a sound-producing organ called a syrinx located at
the base of the trachea where it bifurcates into the bronchi. The syrinx works much like
31
the larynx: sounds are produced by forced air, which vibrates membranous walls to
produce noise.
Among mammals, the development and structure of the larynx is highly
conserved. At birth, the larynges of mice and humans are nearly identical, located at the
level of the second and third cervical vertebrae. The most significant difference arises
postnatally, as the human larynx descends to the level of C3-C6 by age six while the
murine larynx remains high in the neck. As a result of this remarkable similarity
between human and murine larynges, the mouse serves as a tractable model for
investigation into normal and abnormal laryngeal development. Indeed, many mutant
strains have reported laryngeal defects. Unfortunately, consideration of the larynx in
the vast majority of these reports is cursory in nature, limited to descriptions of
dysmorphic or absent laryngeal cartilages. As such, surprisingly few mechanistic details
of laryngeal development have been elucidated from mouse genetic models to date.
While previous studies have elegantly reconstructed 3D wax or computer-aided
models of developing human or murine larynges, several controversies regarding
laryngeal development remain. Most fundamentally, there is no consensus regarding
the embryonic origin of multiple laryngeal structures. There are also conflicting
conclusions in the literature regarding the boundary between the posterior larynx and
the rostral trachea. During development, a portion of the laryngeal lumen becomes
obliterated by an epithelial seam, termed the epithelial lamina, that later recanalizes.
32
However, whether the epithelial lamina represents the infraglottis (Kallius, Frazer), the
glottis (Walander), or the supraglottis (Zaw-Tun, Henick) is disputed. Confounding the
debate, it is unclear whether the trachea arises by outgrowth of the respiratory
diverticulum (lengthening caudally) or by an active septation process (lengthening
rostrally), affecting the position of the boundary between the trachea and infraglottis.
Some (Zaw-Tun and Burdi, 1985) believe the laryngeal ventricles to be the permanent
structure derived from the fifth pharyngeal pouch, while others disagree (Viejo et al,
2012). Finally, the origins of the laryngeal cartilages, whether from neural crest or
mesoderm, are debated. In this chapter, I apply mouse genetic lineage tracing
technology to address many of these questions.
Laryngeal and laryngotracheoesophageal clefts (LTE clefts) are abnormal
connections between the larynx (and possibly trachea) and the esophagus. LTE clefts
form when the process of laryngeal compartmentalization is disrupted or incomplete.
These clefts often allow ingested substances to enter the airway, or air to enter the
alimentary tract. LTE clefts are classified by the length of the aberrant communication;
for example, a Type I cleft does not continue caudal to the vocal folds, whereas a Type
IV cleft extends into the thoracic trachea (Figure 5). Submucous or occult clefts, in which
the soft tissue is intact despite a cartilaginous cleft, are often asymptomatic and thus
have only recently been recognized as a Type 0 (Tucker and Maddalozzo, 1987).
33
The complex morphology of the larynx and its origins from multiple precursor
lineages suggests multiple layers of interactions must occur between tissues to achieve
compartmentalization of the larynx from other structures and organs. Of particular
importance are the tissue interactions which allow the larynx to be fully partitioned
from the esophagus, as openings connecting these organs can be catastrophic. Such
interactions are often mediated by intercellular signaling molecules and pathways. For
example, the Ephrin-B2 knockout mouse may provide insight into the molecular control
of laryngeal and tracheoesophageal compartmentalization. Reverse Ephrin-B2 signaling
has been implicated in multiple septation events at the embryonic midline (Dravis and
Henkemeyer, 2011). A very recent report posits that Eph-ephrin interactions regulate
midline fusion of the posterior larynx and reveals that the eprhin-B2 knockout mouse
also has type III or IV laryngotracheoesophageal clefts (Neilan, et al., 2012).
Two additional intercellular signaling pathways, activated by hedgehog (Hh) or
bone morphogenetic protein (BMP) ligands, seem to be particularly good candidates for
regulating the tissue interactions that underlie laryngeal compartmentalization. Both
sonic hedgehog (Shh) and multiple BMPs are expressed in the ventral rostral foregut
endoderm in the timeframe of active laryngeal morphogenesis. My analysis of Hh and
BMP signaling mouse mutants provides new data, discussed below, demonstrating that
disruption of either pathway can result in laryngeal clefting and revealing the tissue
specificity of these signaling interactions.
34
Figure 5: Classes of laryngotracheoesophageal clefts
Dorsal views of the larynx and trachea with red indicating the
connection to the esophagus. A Type I laryngotracheoesophageal
(LTE) cleft does not extend into the cricoid cartilage. Type II clefts
include the cricoid region. Type III LTE clefts extend through the
cricoid into the cervical trachea. Type IV clefts extend well into
the thoracic trachea, and can reach the level of the bronchi. CC,
cricoid cartilage; TC, thyroid cartilage
35
3.4 Results
3.4.1 Early laryngeal development
In the mouse embryo, the respiratory primordium is first identifiable at Carnegie
stage 11 (17-21 somites, E9.5) as an epithelial thickening of the ventral foregut
endoderm. At stage 12 (22-27 somites, E10.0) the respiratory primordium is an
outpocketing of the foregut lumen and the lung buds are visible. This respiratory
diverticulum can be thought of as being subdivided into the rostral laryngotracheal
sulcus (LTS) and caudal primitive pulmonary sac (PPS) (Figure 6C). Controversy exists
regarding the geographic location of the laryngeal anlage within this respiratory
primordium, predominantly regarding the relative positions of the glottis and
infraglottis in the early embryo.
To resolve these issues, I employed genetic fate-mapping tools to trace the level
of the glottis back through development. The tissues surrounding the laryngeal
endoderm are predominantly derived from two sources: neural crest cells (NCCs) and
splanchnic mesoderm. Mice containing Cre recombinase expressed under neural crest
(Wnt1-Cre) or mesodermal (Mef2C-Cre) promoters were crossed to R26R reporter mice,
which express -galactosidase after Cre-mediated recombination. As this is a permanent
genetic modification, all cells which express Cre and the descendents of those cells will
be labeled blue after LacZ staining. The nature of the visualization chemistry is such
that neighboring cells, those which did not express Cre in their forebears, are not
36
labeled. Consequently, this technique permits a long-term, genetic lineage trace of
individual cell populations.
At birth, the ventral tissues above the glottis are labeled with Wnt1-Cre; R26R
(Figure 12EF) whereas the ventral tissues below the glottis are labeled with Mef2C-Cre;
R26R (Figure 12GH). Thus the morphological boundary between these the Wnt1-Cre
and Mef2C-Cre domains marks the level of the glottis.
Examination of this boundary at earlier timepoints permits us to follow the
glottis backwards through development, thus identifying the origins of the Wnt1-Cre(+)
supraglottis and the Mef2C-Cre(+) infraglottis. When examined at E10.0, the tissues
surrounding the LTS and rostral PPS are Wnt1-Cre(+) (Figure 6A) whereas those
surrounding the caudal PPS are Mef2C-Cre(+) (Figure 6B). These data indicate that the
future infraglottis, trachea, and lung develop from the PPS, whereas the future
supraglottis forms from the LTS (Figure 6C).
37
Figure 6: Neural crest and mesoderm fate maps
(A) Sagittal view of E10.0 Wnt1-Cre; R26R with the LTS stained Wnt1Cre(+). (B) Sagittal view of E10.0 Mef2C-Cre; R26R demonstrating the
ventral PPS is Mef2C-Cre(+). (C) Schematic view of the E10.0 larynx
showing the relative domains of Wnt1-Cre (pink) and Mef2C-Cre (green).
PPS, primitive pulmonary sac; LTS, laryngotracheal sulcus. Scale bars
indicate 250 µm.
38
3.4.2 Laryngeal compartmentalization
At stage 13 (28-33 somites, E10.5) the bronchi emerge as the carina, the point at
which the bronchi bifurcate, separates from the top of the respiratory diverticulum. The
LTS becomes compressed laterally by growing mesodermal anlagen. The foregut
begins to compartmentalize into the trachea/larynx and esophagus via a poorly
understood process (reviewed by Fausett and Klingensmith, 2011). On the dorsal side of
the developing larynx/trachea, controversy relates to the existence (or lack thereof) of an
ascending tracheoesophageal septum, which potentially changes the boundary between
the infraglottis and rostral trachea. Unfortunately, the Mef2C-Cre lineage-trace we
employ here only labels the ventral side of the developing larynx/trachea and therefore
cannot address this issue (the necessary Cre driver to resolve this issue does not appear
to exist). Sanudo and Domenech-Mateu (1990) measured the length of the primary LTS
from stage 11-15 and found no change in length, which they interpreted as evidence
against a septation process. However the entire embryo grows rapidly in size from E9.5E12.0, so the fact that the distance between the top of the septum and the fourth
pharyngeal pouch is constant is consistent with a septation mechanism. However, as the
primary LTS only becomes the supraglottis, these measurements do not address
separation of the trachea and infraglottis from the developing esophagus.
During stage 14 (34-36 somites, E11.0), the arytenoid swellings form and are
pushed craniomedially into the pharyngeal cavity by the developing 4th pharyngeal arch
39
artery and the growing laryngeal mesoderm (Zaw-Tun and Burdi, 1985). These
swellings begin to compartmentalize, or separate, the developing larynx from the dorsal
esophagus (Figure 7A). At the beginning of stage 15 (37-44 somites, E11.5-E12.0) the
epithelial lamina forms ventrally where the laryngeal lumen is compressed by the lateral
walls and continues rostrally between the arytenoid swellings (Figure 7B). This
developing endodermal plate closes the ventral aspect of the lumen. Using the Mef2CCre and Wnt1-Cre lineage traces, one can identify the future supraglottis, glottis, and
infraglottis. As discussed above and in Figure 6, all supraglottic tissues are Wnt1-Cre(+)
whereas all subglottic tissues are Mef2C-Cre(+). Surrounded by Mef2C-Cre(+), Wnt1Cre(-) cells, the epithelial lamina will form the infraglottis.
40
Figure 7: Laryngeal compartmentalization at E11.5
A-D) H&E staining of E11.5 larynx in transverse sections demonstrates
the process of laryngeal compartmentalization. E-H) E11.5 Mef2C-Cre
fate map (Mef2C-Cre; R26R) shows the epithelial lamina is completely
surrounded by Mef2C(+) mesoderm. I-L) While surrounding structures
are Wnt1(+) (Wnt1-Cre; R26R), the tissue immediately around the
epithelial lamina is Wnt1(-), indicating a mesodermal origin. AS,
arytenoid swellings; EL, epithelial lamina; Es, esophagus; Ph, pharynx;
Tr, trachea. Scale bars indicate 250 µm.
41
By stage 16 (45-51 somites, E12.0-E12.5), the epithelial lamina have juxtaposed
and formed an epithelial seam, which blocks the entire primitive laryngopharynx except
the narrow dorsal pharyngotracheal duct. While other studies have described the
pharyngotracheal duct as patent throughout development, I was unable to see a lumen
rostrally after E11.5 (compare Figure 7C to 8C). The rostral ends of the arytenoid
swellings and associated epithelial lamina protrude into the pharyngeal lumen (Figure
8BC), thus separating the rostral larynx from the developing esophagus. A cluster of
apoptotic cells are present at the point at which this epithelial seam detaches from the
esophageal epithelium (Figure 8GH). This compartmentalization process creates a band
of mesodermal tissue, continuous with the tracheoesophageal septum, that separates the
larynx from the esophagus (Figure 8E).
Molecular regulation of laryngeal compartmentalization has not previously been
investigated. Shh and Nog, both critical secreted signaling molecules, are both
expressed in the epithelium of the larynx during the period of compartmentalization.
Shh is expressed in the dorsal and ventral midline of the rostral esophagus and strongly
expressed in the EL (Figure 9A). Intriguingly, Shh expression is significantly lower at the
dorsal end of the EL, the point where it remodels to separate from the esophagus. Nog
is expressed in a pattern complementary to that of Shh, with highest levels in the lateral
pouches of the rostral esophagus and low levels in the EL (Figure 9B). As both Shh and
42
Nog are present in the compartmentalizing larynx, we wanted to determine if either
play a role in laryngotracheoesophageal compartmentalization.
43
Figure 8: Laryngeal compartmentalization
A) Undivided pharyngeal region. B) Arytenoid swellings
separate the future larynx from the esophagus. C) Epithelial
lamina caudal to arytenoid swellings. D) Epithelial lamina
separates from the esophagus. E) Trachea and esophagus. (F-I)
Rostral to caudal sections through the compartmentalizing larynx
and esophagus demonstrating local pattern of apoptosis (cleaved
caspase 3, red). As, arytenoid swelling; EL, epithelial lamina; Es,
esophagus; Ph, pharynx; Tr, trachea. Scale bars indicate 250 µm.
44
Figure 9: Shh and Nog expression in the E12.5 larynx
(A) The Shh expression pattern at E12.5 is concentrated medially,
with strong expression in the EL. A slight downregulation of Shh
at the point of compartmentalization is observed. (ShhGFP-Cre
embryo stained with anti-GFP). (B) Noggin is expressed in a
complementary pattern to Shh, with highest expression in the
lateral pouches of the rostral esophagus. Low-level expression is
present in the EL. (Nog-LacZ embryo stained with anti-βgal).
Scale bars indicate 250 µm.
45
3.4.3 Shh and laryngeal clefting
As discussed above, the arytenoid swellings and epithelial lamina divide the
pharynx into laryngeal and esophageal aspects. The arytenoid swellings enlarge to fill
the pharyngeal lumen, the epithelial lamina forms to compartmentalize the larynx, and
apoptotic remodeling occurs along the dorsal epithelial lamina to separate the laryngeal
and esophageal epithelia. Laryngeal clefts may form when any step in this process is
disrupted. As reported by Segai et al., (2009), ablation of the “pharyngeal endoderm
enhancer” of Shh expression results in hypoplastic arytenoid swellings. We recreate
this phenotype with endoderm-specific ablation of Shh, using the Nkx2.5-Cre
recombination driver. In these mutants, no identifiable arytenoid swellings are present
(Figure 10AB). This creates a Type I laryngeal cleft, or a failure of the most-rostral
compartmentalization of the larynx from the esophagus. Thus, Shh expression from the
pharyngeal endoderm is required for normal laryngeal compartmentalization.
As Shh is a secreted ligand, we wanted to determine the domain of Shh reception
required for normal laryngeal compartmentalization. We first examined whether the
pharyngeal endoderm itself required Shh in an autocrine fashion by ablating the Shh
receptor Smo with ShhGFP-Cre (not shown). This removes the receptor from the cells
that express Shh. As this manipulation did not produce laryngeal clefts, we next
examined whether ventral mesoderm was the relevant target by ablating Smo with
Mef2C-Cre (not shown). We also used Nkx2.5-Cre, which would ablate Smo in both
46
ventral mesoderm and endoderm (not shown). Shh signaling is not required in either
ventral mesoderm or endoderm, as none of these ablations produced a laryngeal cleft
phenotype. Finally, we analyzed whether NCCs (or their derivatives) were the target of
Shh signaling during laryngeal compartmentalization. Ablation of Smo with Wnt1-Cre,
which deletes Smo only in nascent and migrating NCCs, produces laryngeal clefts
indistinguishable from Nkx2.5-Cre; Shhflox/- (Figure 10BCDE). Together, these tissuespecific genetic manipulations of Shh expression and signal reception indicate that Shh
ligand produced by the pharyngeal endoderm must be received by neural crest cells as a
critical signaling component of laryngeal compartmentalization.
47
Figure 10: Shh in laryngeal clefting
A) Rostral to caudal sections through a wild-type, compartmentalized
larynx and esophagus at E18.5. B) Ablation of endodermal Shh results in
a laryngeal cleft and the larynx separates from the esophagus in an
abnormally caudal position. (CDE) Ablation of Smo from NCCs with
Wnt1-Cre recapitulates the laryngeal cleft phenotype. Es, esophagus;
LTEC, laryngotracheoesophageal cleft; Lx, larynx; Tr, trachea. Scale bars
indicate 500 µm.
48
3.4.4 BMP and laryngeal clefting
Shh signaling often acts in concert with BMP signaling antagonism, as in neural
tube patterning (Liem, et al., 2000). Accordingly, we addressed whether genetic
reduction of a key BMP antagonist expressed in the spatiotemporal context of laryngeal
development might also cause disruptions. During patterning and morphogenesis of
the future neck region, Noggin expression in the laryngeal endoderm (Figure 9B)
coincides with Shh expression and is present in a partially complementary pattern. We
thus wanted to determine if this domain of Nog expression plays a role in laryngeal
compartmentalization.
As Nog is a secreted BMP antagonist, absence of Nog will result in excess BMP
signaling in cells that normally express Nog and the surrounding tissue. When we used
ShhGFP-Cre to ablate Nog in the laryngeal endoderm and notochord, the mutants
(ShhGFP-Cre; Nogflox/null embryos) have a laryngeal cleft (Figure 11B), connecting the
larynx to the esophagus. This cleft extends through the cricoid cartilage, splitting it
dorsally, and is thus a Type II LTE cleft. Therefore, BMP antagonism in the larynx is
required for proper compartmentalization of the larynx.
As this manipulation increases BMP signaling in both the laryngeal endoderm
and surrounding tissue, it was not clear what the relevant target of Nog was for
laryngeal compartmentalization. Nog could either prevent BMP signaling in the
laryngeal mesoderm or in the laryngeal endoderm per se. To distinguish between these
49
two possibilities, we also used a constitutively active allele of BMP receptor 1A to cellautonomously increase BMP signaling. When this genetic manipulation was performed
in the endoderm (ShhGFP-Cre; CABmpr1a embryos, not shown), similar LTE clefts were
obtained. Thus, for laryngeal compartmentalization, Nog expression from the laryngeal
endoderm is required to prevent excess BMP signaling in the endoderm itself.
We sought to determine a possible cellular basis for these defects. In these
ShhGFP-Cre; caBmpr1a mutants, the dorsal cluster of apoptotic cells in the remodeling
epithelial lamina is never observed (Figure 11D). On the other hand, a cluster of cleavedcaspase 3 positive cells is present during remodeling of the wild-type epithelial lamina
(Figure 11C). It is intriguing to speculate that this apoptosis is required to separate the
foregut into the larynx and esophagus. However, the Apaf1-null mouse, in which
apoptosis is inhibited (Nagasaka, et al., 2010), has no report of laryngeal cleft, indicating
that apoptosis is not absolutely required for foregut division. The apoptosis observed in
wild-type and absent with excess BMP signaling is likely a byproduct of the remodeling
process.
50
Figure 11: BMP in laryngeal clefting
(A) Wild-type larynx caudal to separation point from esophagus
at E15.5. (B) Activation of CA-Bmpr1a within the endoderm
causes a laryngeal cleft at the level of the cricoid cartilage. (C)
Apoptotic cells at site of epithelial remodeling in the wild-type
larynx at E12.5. (D) No cluster of apoptosis is observed when CABMPr1a is activated in the endoderm. Scale bars indicate 250 µm
in A-D and 100 µm in C' and D'.
51
3.4.5 Laryngeal cartilage development and deformities
The scaffolding of the larynx is provided by three large single cartilages
(epiglottic, thyroid, and cricoid) and three smaller paired cartilages (arytenoid,
corniculate, and cuneiform) (Figure 4A). The thyroid and cricoid are hyaline cartilages
that house the majority of the larynx. The thyroid cartilage forms the “Adam’s apple”
on the ventral side of the larynx. The cricoid cartilage, the only cartilage to form a
complete ring around the larynx, forms the dorsal wall of the larynx and connects to the
top of the trachea. The pairs of arytenoid (mostly hyaline), corniculate, and cuneiform
(elastic) cartilages are located in the interior of the larynx and influence the position and
tension of the vocal folds. The hyoid bone, while not technically a component of the
larynx, connects to the thyroid cartilage via the thyrohyoid membrane and serves to
anchor and couple the tongue to the larynx. Cartilaginous in the embryo, the hyoid
becomes bone via endochondrial ossification.
Short term, dye based fate mapping studies have established that the laryngeal
skeletal structures are formed from the pharyngeal arches. The pharyngeal arches are
transient embryonic structures located at the lateral sides of the neck that become
populated by a mix of mesoderm and neural crest cells (NCCs). The lesser horns and
the superior portion of the hyoid body are derived from the second arch, while the
greater horns and inferior portion of the hyoid body are derivatives of the third arch.
The epiglottic and thyroid cartilages are derived from the fourth pharyngeal arch and
52
the rest come from the sixth arch (Cochard, 2002). While the hyoid is widely believed to
be of neural crest in origin, there is no consensus in the literature regarding whether the
more caudal cartilages are derivatives of NCCs (Matsuoka et al., 2005), lateral plate
mesoderm (Kaufman and Bard, 1999), or a combination of both (Jeong et al., 2004). As
the only laryngeal cartilage to express GFAP (widely expressed in the CNS), the
epiglottic cartilage has been hypothesized to be a derivative of the neural crest (Kepes et
al., 1998, Katori et al., 2011).
To resolve this controversy, we employed a genetic lineage trace in the mouse,
Wnt1-Cre; R26R, in which all NCC derivatives are permanently marked, to establish the
first long-term fate map of neural crest contributions to the larynx. By E10.5, neural
crest cells (NCCs) are present along the lateral sides of the future larynx (Figure 12A).
At E12.5, this NCC-derived domain encircles all except the dorsal-most aspect of the
larynx. The arytenoid swellings, in the interior of the larynx, are notably non-neural
crest in origin (Figure 12B). These data indicate that the rostral cartilages are
predominantly NCC-derived and suggest that the caudal cartilages are mesodermal.
The epiglottis (Figure 12C) and hyoid bone (Figure 12D) are entirely Wnt1-Cre(+) and
Mef2C-Cre(-), indicating a NCC origin, whereas the arytenoid cartilages (Figure 12EF)
and cricoid cartilage (Figure 12GH) are Wnt1-Cre(-) and Mef2C-Cre(+), therefore derived
from mesoderm. Interestingly, while the horns and the majority of the body of the
53
thyroid cartilage are Wnt1-Cre(+), Mef2C-Cre(-), the ventralmost caudal portion is Wnt1Cre(-), Mef2C-Cre(+) (Figure 12GH), indicating mesodermal contributions.
54
Figure 12: Complementary Wnt1-Cre and Mef2C-Cre lineage traces
A) Transverse section through the larynx of E10.5 Wnt1-Cre; R26R
embryo. LacZ(+) NCCs have migrated into the laryngeal region.
B) By E12.5, NCCs almost entirely surround the developing
arytenoid swellings, which are mesodermal in origin. C-F) E17.5
transverse. The epiglottic cartilage (C), hyoid bone (D), and
majority of the thyroid cartilage (E) are derived from NCCs.
However, the arytenoid cartilages (E) and cricoid cartilage (F) are
not. A small portion of the ventrocaudal thyroid cartilage (* in F) is
also not of NCC origin (F). Using the Mef2C-Cre mesodermal
lineage trace, we confirm that the arytenoid cartilages (G) cricoid
cartilage (H) and the small portion of the thyroid cartilage (* in H)
are mesodermal. AC, arytenoid cartilage; AC, arytenoid cartilage;
AS, arytenoid swelling; CC, cricoid cartilage; EC, epiglottic
cartilage; Es, esophagus; HB, hyoid bone; Lx, larynx; TC, thyroid
cartilage. Scale bars indicate 250 µm.
55
Figure 12: Complementary Wnt1-Cre and Mef2C-Cre lineage traces
56
At Carnegie stage 15, the “mesodermal anlagen” have condensed into distinct
hyoid and cricothyroid parts and begin expressing pre-chondrocyte marker Sox9 (Figure
13A). During stages 17-18, chondrification centers appear in the hyoid (E13.0) and
cricoid (E14.0) primordia (Zaw-Tun and Burdi, 1985). Lateral and dorsal fusion of the
cricothyroid anlagen occurs. During stages 19-23, chondrification of the hyoid, thyroid,
and cricoid occurs by E15.0, followed by the epiglottis at E16.0 (Katori et al, 2011)
(Figure 13B). The cuneiform and corniculate cartilages fuse with the apices of the
arytenoids.
The subglottis is the region of the larynx immediately caudal to the vocal folds,
encompassing the cricoid cartilage region and extending to the tracheal opening. In the
infant, the subglottis is the narrowest part of the larynx. Congenital subglottic stenosis,
the third most common laryngeal defect, is a narrowing of the larynx below the level of
the glottis. Subglottic stenoses can be categorized into soft tissue (usually acquired) or
cartilaginous (usually congenital) in etiology. Soft tissue stenosis is caused by an
overgrowth of the laryngeal mucosa. The submucosa proliferates rapidly during cricoid
chondrification, transiently herniating into the dorsal subglottis. If the cricoid does not
proportionally increase in size, persistence of this excess tissue could cause stenosis.
Cartilaginous stenosis is caused by a small or misshapen cricoid cartilage. Most
commonly, the cricoid cartilage is elliptical (narrow transversely) but may instead be
flattened (narrow anteroposteriorly) or generally thickened (Holinger, 1999). The first
57
tracheal ring can sometimes telescope within a flattened cricoid, trapping it and
narrowing the airway (Kakodkar et al, 2012). Individuals with Fraser syndrome often
have hypoplasia or stenosis of the larynx (van Haelst et al., 2007). The 22q11 monosomy
syndromes have an increased incidence of congenital subglottic stenosis, laryngeal webs
and clefts.
Many fundamental cellular processes are essential for proper laryngeal cartilage
development. Sox9 is an essential and early transcription factor during chondrogenesis.
Haploinsufficiency of Sox9 causes thin hyoid and laryngeal cartilages (Bi et al., 2001)
whereas ablation of Sox9 from NCCs (with Wnt1-Cre) results in a loss of the hyoid and
thyroid cartilages (Mori-Akiyama et al., 2003). Similarly, loss of Fibulin-1, an ECM
protein which regulates cell mobility, causes gross abnormalities of NCC-derived
structures, including a reduction of the hyoid and thyroid cartilages (Cooley et al., 2008).
Stereotypic developmental signaling pathways, including Hh, Wnt, and BMP,
are also essential for laryngeal cartilage development. Ablation of R-spondin2, an
activator of Wnt/β-catenin signaling, results in dysmorphic laryngeal cartilages and
fusions between the arytenoids and cricoid (Yamada et al., 2009). Simultaneous
ablation of Wnt-1 and Wnt-3a causes malformations of the thyroid cartilage and absence
of the main body of the hyoid bone (Ikeya et al., 1997).
58
Figure 13: Development of laryngeal cartilages
(A) Transverse view of the developing laryngeal cartilages.
Developing pre-cartilage condensations of the future cricoid and
thyroid cartilages are visualized with anti-Sox9 staining (red). (B)
Alcian blue staining of mature laryngeal cartilages, ventral view.
CC, cricoid cartilage; TC, thyroid cartilage. Scale bars indicate 500
µm.
59
3.4.6 BMP in laryngeal cartilage development
Bmp-7-deficient embryos have an absent cricoid and malformed thyroid cartilage
and hyoid bone (Zouvelou et al., 2009). Crossveinless 2 (Cv2; Bmper) is a BMP-binding
protein that positively regulates the sizes of laryngeal cartilages (Ikeya et al., 2006).
Similarly, ablation of Follistatin-like 1 (Fstl 1), a TGF-β family antagonist, causes gross
disorganization of the laryngeal cartilages (Geng et al., 2011).
Many mouse mutants have small or hypoplastic laryngeal cartilages; however
the signaling events that regulate chondrification of the larynx have not been previously
elucidated. In addition to laryngeal clefts, we report that ablation of Noggin is the first
genetic model to have enlarged laryngeal cartilages. Specifically the thyroid cartilage
and hyoid bone are much thicker than those of wild-type littermates (Figure 14B).
Additionally, both lateral and midline fusions are present between the hyoid bone and
thyroid cartilage. Noggin is expressed in multiple tissues around the developing larynx,
including the notochord, floor plate, endoderm, and in neural crest cells, but it is unclear
which of these domains are responsible for limiting laryngeal cartilage size. Ablation of
Noggin from the notochord, floor plate, and endoderm with ShhGFP-Cre does not affect
the size of the laryngeal cartilages. However, ablation of Noggin from neural crest cells
with Wnt1-Cre (Figure 14C) cause the large laryngeal cartilages seen in the Noggin-null
mouse, with premature ossification centers. When autocrine BMP signaling is increased
even more, by activation of Bmpr1a within the neural crest, the over-ossification
60
phenotype is even more dramatic (Figure 14D). Collectively, these data indicate that
neural crest cells per se produce Noggin to protect themselves from surrounding BMPs.
61
Figure 14: Noggin and laryngeal cartilage
A) Wild-type laryngeal cartilage and hyoid bone with single
midline ossification center. B) Noggin-null hyoid bone is large
and exhibits premature ossification, lateral and midline fusions
with the thyroid cartilage, and a cleft in the cricoid cartilage. C)
Ablation of Nog from NCCs causes extensive overgrowth and
over ossification of the hyoid bone, as well as thyroid cartilage
deformities. D) Activation of Bmpr1a in NCCs exacerbates the
phenotypes, precluding identification of individual skeletal
elements. Scale bars indicate 1mm.
62
3.4.7 Shh and laryngeal cartilage
In addition to being important for proper laryngeal compartmentalization, Sonic
hedgehog (Shh) is also essential for laryngeal cartilage development. Hedgehog
signaling is likewise important for development of the laryngeal apparatus. Ablation of
Hh signaling from the neural crest results in absence of the hyoid and thyroid cartilages
(Jeong et al., 2004). A knockout of Kinesin-2 subunit Kif3a, essential for intraflagellar
transport and Hh signaling, phenocopies this mutant (Kolpakova-Hart et al., 2007). In
the Shh-null mouse, no laryngeal cartilage is present (Figure 15A) and no laryngeal
structures are identifiable. In fact, there is no communication between the oral cavity
and the foregut tube. However, the relevant domain(s) of Shh signal production and
transduction are unknown. Ablation of the “pharyngeal endoderm enhancer” of Shh
expression, MFCS4, causes deformation of the hyoid, loss of the epiglottis, and
hypoplasia of the arytenoid swellings (Sagai et al., 2009). Unlike the Shh-null, however,
the cricoid and thyroid cartilages are easily identifiable and morphologically normal.
We thus hypothesized that axial Shh expression (from the notochord and floor plate) is
responsible for the development of these caudal laryngeal cartilages. To address the
axial domain, we ablated Shh with Foxa2-NFP-Cre; however, we did not see a laryngeal
phenotype (not shown). This led us to reconsider the entire anterior endodermal
domain of Shh expression, more broad than the MFCS4 enhancer, with an endodermaldeletion of Shh with Nkx2.5-Cre. Surprisingly, we found that the endodermal domain
63
of Shh is indeed responsible for laryngeal morphogenesis. In the absence of Shh from
the nascent foregut endoderm, the hyoid bone, thyroid cartilage, and arytenoid
cartilages do not form, and the cricoid cartilage is severely malformed (Figure 15BC).
While we determined that the endodermal domain of Shh is essential for
laryngeal cartilage development, which cells receive that secreted signal is not known.
Three likely domains of Shh reception exist: neural crest, mesoderm, and the endoderm
itself. First, we examined the role of autocrine Shh signaling, back to the endoderm
itself by ablating Smoothened with ShhGFP-Cre (Figure 15D) and found no role for
autocrine Shh signaling in laryngeal development. Next, we confirmed the work of
Jeong et al. (2004) by ablation of Smoothened in the neural crest (with Wnt1-Cre). As
expected, loss of Hh reception in the neural crest resulted in loss of the hyoid and
thyroid cartilages (Figure 15E), the laryngeal derivatives of neural crest. Finally, we
ablated Smoothened in the caudal laryngeal mesoderm with Mef2C-Cre, and found that
the arytenoid and cricoid cartilages are lost, further supporting a mesodermal origin
(Figure 15F). Consistent with a small portion of the thyroid cartilage being mesodermal
in origin, ablation of Smo within the laryngeal mesoderm did produce a small defect in
the caudal, ventral thyroid cartilage. Thus, Shh expression from the foregut endoderm
signals directly to neural crest for the formation of the rostral laryngeal cartilages and
also signals directly to mesoderm for the caudal cartilages.
64
Figure 15: Shh and laryngeal cartilage
(A) Shh-null larynx and trachea exhibit nearly complete absence
of cartilage. (B) Alcian blue staining of wild-type laryngeal
cartilage. (C) Ablation of endodermal Shh results in a loss of the
thyroid cartilage and deformity of the cricoid cartilage. (D)
Ablation of Shh reception from the endoderm does not cause
laryngeal defects. (E) Ablation of Shh reception from splanchnic
mesoderm causes a defect of the cricoid cartilage and tracheal
rings, but the thyroid cartilage is mostly present, missing only a
small ventrocaudal portion (*). Scale bars indicate 1mm.
65
3.4.8 Laryngeal musculature, epithelia, and connective tissue
development
By stage 17-18, the primordia of intrinsic and extrinsic laryngeal muscles are
visible. The musculature and connective tissue of the larynx can be subdivided in to the
intrinsic elements that are confined to the larynx and the extrinsic elements that provide
connections to surrounding structures. The intrinsic muscles join together the cartilages
of the larynx and change the shape of the laryngeal cavity to adjust pitch and tone
during vocalization. All intrinsic muscles of the larynx come from the sixth branchial
arch mesoderm except the cricothyroid, a fourth arch derivative. The extrinsic muscles
support the larynx and influence the relationship between the hyoid bone and the larynx
during swallowing and phonation. While originally believed to be derived from
(paraxial) somitic mesoderm, it was recently established that the majority of the extrinsic
musculature of the larynx comes from Isl1(+) splanchnic mesoderm (Nathan et al., 2008).
Similarly, the membranes and ligaments of the larynx may also be subdivided into
extrinsic and intrinsic categories. The lateral cricothyroid membranes, referred to as the
conus elasticus, connect the thyroid, cricoid, and arytenoid cartilages.
During the "fetal stage," goblet cells and submucosal glands develop. The larynx
develops neurologic reflexes: myenteric plexuses and ganglion cells are differentiated by
13 weeks (Kakodar, et al., 2012). The free edge of the vocal folds, the anterior glottis, is
covered with stratified squamous epithelium, 5-25 cells thick (Hopp, 1955). The
66
posterior glottis is covered with pseudostratified, ciliated epithelium, with microridges
and microvilli. The true vocal folds also contain skeletal muscle, the vocalis.
The glossopharyngeal nerve (CN IX), derived from the third branchial arch,
sends fibers to the upper epiglottis. However, the majority of the larynx is innervated
by two branches of the vagus nerve (CN X). The superior laryngeal nerve, from the
fourth branchial arch, innervates the lower epiglottis, provides sensory innervations to
the glottis and laryngeal vestibule, and motor innervations of the cricothyroid muscle.
The recurrent laryngeal nerve, from the sixth arch, creates sensory innervations to the
subglottis and motor innervations to all other muscles of the larynx.
Multiple congenital laryngeal defects, including cri-du-chat and vocal cord
paralysis, may be due to defects in laryngeal innervations. In cri-du-chat syndrome, the
posterior portion of the glottis remains open during phonation, producing a
characteristic “mewing” cry. Cri-du-chat syndrome is associated with 5p15 deletion
(Church, et al., 1995). Vocal cord immobility is the second most common defect of the
larynx. While unilateral paralysis is usually the result of injury, bilateral paralysis is
typically congenital. As idiopathic vocal cord paralysis usually resolves with age, it is
hypothesized to be the result of an immature vagus nerve. Arnold-Chiari malformation
can cause hydrocephalus, which may cause bilateral vocal cord paralysis.
Disruption of fundamental cellular processes underlying laryngeal
morphogenesis leads to expected laryngeal defects in the mouse. Ablation of myogenin,
67
a MyoD family bHLH transcription factor that activates skeletal muscle differentiation,
results in a near-complete loss of laryngeal skeletal musculature (Tseng et al., 2000). The
authors hypothesize that the myogenin knockout mouse may be unable to adequately
close its glottis, which may be similar to cri-du-chat syndrome in human infants.
3.4.9 Recanalization of the larynx
At stage 15, a transverse pouch develops between the arytenoid swellings and
the hypobranchial eminence (future epiglottis), termed the laryngeal cecum. However,
because the arytenoid swellings block a portion of the laryngeal lumen, this pouch is
blind and does not connect to the trachea. Between stages 19-23, or E14.5-E16.0 in the
mouse, the epithelial lamina begins to recanalize in a dorsocranial to ventrocaudal
direction (Figure 16). This brings the pharyngoglottic duct in continuity with the
laryngeal cecum. It has been previously debated whether this recanalizing portion of
the larynx represents the infraglottis or supraglottis. As the site of recanalization is
surrounded by Mef2C-Cre labeled mesoderm, I demonstrate that the recanalized portion
of the larynx is, in fact, the infraglottis.
The lateral sides of the laryngeal cecum expand to form ventricular outgrowths.
These outgrowths are more pronounced in humans (Zaw-Tun and Burdi, 1985) than in
the mouse (Henick, 1993). It is debated whether the laryngeal ventricles represent the 5th
pharyngeal pouch (Zaw-Tun and Burdi, 1985) or develop from the midline of the
pharyngeal floor between the third and fourth pouches (Viejo et al., 2012). The lumen
68
caudal to the EL, Zaw-Tun’s infraglottis, enlarges and the pharyngoglottic duct assumes
a vertical position along the dorsal margin of the EL. During stage 17-18 (52-64 somites,
E13.0-E14.0) the laryngeal cecum grows caudally to the level of the glottis.
69
Figure 16: Recanalization of the infraglottis
A) Sagittal section through E18.5 pharynx, larynx, and trachea. BD) Sections of recanalizing larynx through the planes indicated in
(A). Scale bars indicate 500 µm.
70
3.4.10 Maturation of the larynx
During Carnegie’s “fetal stage” or E16.5-E19.5 in the mouse, recanalization of the
epithelial lamina is completed. By the 10th week in humans, only a perforated
membrane (“web”) remains at the level of Zaw-Tun’s presumptive glottis. During the
12th week, the EL disappears completely and the laryngeal lumen is continuous.
Defects at the level of the glottis interfere with breathing and phonation. Laryngeal
atresia, while rare, is the congenital failure of the laryngeal opening to develop and
requires an immediate tracheotomy at birth. The three classes of laryngeal atresia are
interpreted to be the result of developmental arrest during various stages of
recanalization of the epithelial lamina (Zaw-Tun, 1998). In Type I atresia, it appears that
no recanalization has occurred: a shallow cleft is present between the apices of the
arytenoids, but caudally no lumen is patent because the epithelial lamina is intact. Type
II atresia may result from arrest mid-recanalization, creating a partially-formed
vestibule. As the epithelial lamina at the level of the glottis is last to recanalize, arrest at
this stage may cause Type III laryngeal atresia, also referred to as glottic webs.
Congenital laryngeal webs are abnormal membranous tissue present across the lumen of
the larynx, most common at the level of the glottis (75%) but may also be supra- or infraglottic. Currently, no mouse models of laryngeal atresia or laryngeal webs have been
reported.
71
Also during the fetal stage, the bilateral outgrowths from the laryngeal cecum
dilate and become the laryngeal ventricles. The upper and lower “lips” of the
ventricular openings become the false and true vocal folds, respectively (Zaw-Tun and
Burdi, 1985). Vocal processes form from the arytenoid swellings . The lingual surface
of the epiglottis is covered by a stratified squamous epithelium. The majority of the
laryngeal surface, like most of the larynx, is covered by respiratory epithelium:
pseudostratified, with ciliated columnar cells and scattered mucous-secreting goblet
cells (Hopp, 1955).
Caudal to the epiglottis are the vestibular folds, or false vocal folds, which are
not used for vocalization but are important for resonance. The vestibular folds are
formed from the free edges of the quadrangular membranes and do not contain muscle.
The laryngeal ventricle (also known as the laryngeal sinus or Morgangni’s sinus) is a
spindle-shaped indentation between the vestibular folds and true vocal folds, bordered
laterally by the thyroarytenoid muscle. The anterior portion of the ventricle has an
opening to the laryngeal saccule (aka laryngeal pouch, Hilton’s pouch), a narrow
membranous diverticulum that extends rostrally to the level of the top of the thyroid
cartilage. The laryngeal saccule contains mucous-secreting glands which lubricate the
vocal folds.
Other supraglottic congenital defects of the larynx include, lymphangiomas,
saccular cysts, and laryngocoeles. Lymphangiomas, or abnormal development of
72
lymphatic tissue, can develop to block the upper airway. Saccular cysts and
laryngocoeles are abnormal dilations or herniations of the laryngeal saccule. A
laryngocoele is patent with the laryngeal ventricle and filled with air. On the other
hand, saccular cysts do not open into the ventricle and are often filled with mucous.
Saccular cysts are thought to form when patency between the laryngeal saccule and
ventricle is not maintained during embryogenesis.
3.5 Discussion and future directions
Much of this work revolves around the dual fate maps of Mef2C-cre
(mesodermal) and Wnt1-Cre (NCC). Mef2C-Cre labeled laryngeal tissues up to the
cricoid cartilage and Wnt1-Cre labeled those down to an including the thyroid cartilage.
This serendipitously marked the level of the glottis as the interface between these two
domains. These dual fate maps served three scientific purposes. First of all, they
resolved the controversy regarding embryonic origins of the supraglottis, glottis, and
infraglottis, demonstrating that the primitive pulmonary sac contributes only to the
infraglottis and the laryngotracheal sulcus becomes the supraglottis. Secondly, the dual
fate maps resolved the controversy regarding what part of the larynx is formed by the
epithelial lamina, and, thus, what part must recanalize to create an open connection
between the pharynx and the trachea. Positioned between the Mef2C-Cre-labeled
mesodermal arytenoid swellings, the epithelial lamina clearly becomes the infraglottis.
Lastly, comparison of these fate maps at late stages demonstrates that the hyoid and
73
most of the thyroid cartilage are derived from NCCs, whereas the arytenoids, cricoid,
and a small portion of the thyroid cartilage are derivatives of splanchnic mesoderm.
Shh plays multiple key roles in laryngeal development. Shh expression from the
vFGE is required for laryngeal compartmentalization, and I demonstrate that this
domain of Shh signals to NCCs to promote separation of the larynx from the esophagus.
Experiments demonstrated that, unlike many other contexts, the role of Noggin is to
cell-autonomously protect the cells that secrete it. Regarding laryngeal
compartmentalization, Nog produced by the endoderm protects the endoderm from
excess BMP signaling. Likewise, during development of NCC-derived laryngeal
cartilages, Nog secreted by NCCs is essential for preventing BMPs from signaling to
those NCCs.
With respect to development of the laryngeal cartilages, I determined that Shh
signals directly from the laryngeal endoderm to NCCs and foregut mesoderm. While
the mechanism of Shh's action was not determined, the simplest model is that Shh acts
as a survival or proliferation signal. When this signal is ablated, both NCC and
mesoderm-derived laryngeal cartilages do not form. We also determined that Noggin
negatively regulates the size of laryngeal cartilages and, with the ablation of Nog,
provide the first mouse model of thickened laryngeal cartilages. Thus, Noggin is likely
to serve as an anti-proliferative signal, controlling laryngeal cartilage size.
74
Altogether, this work resolves many debates regarding laryngeal morphogenesis
impossible to answer without modern genetic tools. Additionally, I demonstrate key
functions of Shh and BMP signaling in compartmentalization of the larynx from the
esophagus and in the development of the laryngeal cartilages. I determined the relevant
domains of Shh and Nog expression and reception in the developing larynx, which is
the first step towards a greater molecular understanding of laryngeal development.
75
4. βarrestins and Shh in the developing embryo
4.1 Summary
Arrestins, a family of multifunctional cytosolic adaptor proteins, are widely
studied for the ability to desensitize and internalize activated G-protein coupled
receptors (GPCRs). β-arrestins, the non-visual arrestins, are furthermore known to
mediate signaling through a wide variety of other receptor families, many of which
control key signaling events during embryogenesis. As both the βarrestin1 (Arrb1) and
βarrestin2 (Arrb2) knockout mouse strains are viable, we sought to determine the
essential roles of β-arrestins during development. Given the large number of signaling
pathways potentially misregulated in the absence of βarrestins, we were surprised to
find that βarr1/βarr2 double knockout (DKO) mice are born alive without major external
defects. However, βarrestin DKO pups are small in size, quickly become cyanotic, and
do not survive 48 hours postnatally, mostly due to complications from a severe cleft
secondary palate. Internally, βarrestin DKO pups also have cardiovascular, cerebellar,
renal, and skeletal anomalies. Most βarrestin DKO phenotypes are consistent with
specific impairment of mitogenic Shh signaling; in contrast, DKO pups do not have
anomalies in tissues for which Shh acts principally as a morphogen. Thus, many
signaling pathways in which βarrestins have been shown to act in vitro do not require
βarrestins during development. However, mitogenic Sonic hedgehog signaling appears
to require βarrestins to fulfill its developmental roles.
76
4.2 Acknowledgements
This work was performed in collaboration with Jeff Kovacs from the Lefkowitz
lab. In particular, the qPCR arrays (Figures 22 and 25G) were partially performed by
Jeff.
4.3 Introduction
Proper signaling through transmembrane receptors is critical for basic cellular
function during development, homeostasis, and disease. The largest family of receptors,
G-protein coupled receptors (GPCRs), respond to a wide variety of extracellular ligands,
whose binding induces a conformational change required for recognition and
phosphorylation by GPCR kinases (GRKs) (Premont and Gainetdinov, 2007). Arrestins
are a family of multifunctional signaling and adaptor proteins with the ability to
desensitize and internalize ligand-activated GPCRs (Lefkowitz et al., 2005). Arrestins
are recruited to phosphorylated GPCRs, where they physically inhibit signaling and/or
mediate clathrin-dependent endocytosis of the activated GPCR. Classically,
GPCR/GRK/arrestin interactions result in receptor desensitization, or termination of
signaling to G-proteins. However, arrestins can also potentiate a diverse array of
signaling events, including the generation of “signalosomes”, in which distinct signaling
events are enhanced by physically assembling critical components on arrestin scaffolds
(Lefkowitz et al, 2006).
77
Recently, many interactions between β-arrestins and non-GPCRs, especially
developmentally important receptors, have been reported in a wide variety of contexts,
from C. elegans to human cell culture. (Reviewed in Kovacs, et al., 2009.) ARR-1, the sole
C. elegans arrestin, and kurtz, the single non-visual arrestin in Drosophila both have
traditional arrestin functions, modulating sensory input in olfactory neurons. Arr-1
mutants, are not defective in chemotaxis (Fukuto, et al., 2004), but have deficiencies in
olfactory adaptation and recovery, likely the result of faulty receptor endocytosis
(Palmitessa et al., 2005). ARR-1 regulates C. elegans longevity through a interactions
with DAF-2 and DAF-18, the orthologues of IGF-1R and PTEN, respectively (Palmitessa
and Benovic, 2010). ARR-1 also participates in the unfolded protein response and a p38
MAPK signaling pathway required for innate immunity through regulation of GPCR
signaling in C. elegans neurons (Singh and Aballay, 2012).
Kurtz is an essential neural gene in Drosophila melanogaster (Roman, et al, 2000)
that helps maintain olfactory sensitivity (Ge, et al., 2006) and exploratory activity (Liu, et
al., 2007), functioning canonically as a desensitizer of GPCR signaling (Johnson, et al.,
2008). Kurtz also participates in Notch signaling, forming a complex with Deltex and
Notch and promoting the E3 ubiquitin ligase activity of Deltex on Notch, resulting in
Notch degradation (Mukherjee, et al., 2005). Additionally, Kurtz can inhibit both MAPK
ERK and NF-κB signaling (Tipping, et al, 2010). Recent reports have implicated Kurtz in
negative regulation of Hedgehog signaling, promoting internalization and degradation
78
of Smoothened in an activity-independent manner (Molnar, et al., 2011) parallel to
internalization by ubiquitination (Li, et al., 2012).
In Xenopus, xβarr2 is required for convergent extension and midline intercalation
of mesoderm during embryonic axis elongation. xβarr2 is essential for non-canonical
Wnt/planar cell polarity signaling by regulation of Dishevelled, but does not participate
in the G-protein mediated non-canonical Wnt/Ca2+ signaling pathway (Kim et al., 2007).
In fact, the balance between xBarr2 and casein kinase 1/2 may direct which noncanonical Wnt signaling pathway is activated (Bryja, et al., 2008). Simultaneously,
xβarr2 is involved in the negative regulation of canonical Wnt signaling, as axis
duplication by ectopic Wnt8 is suppressed in the xβarr2 morphant (Bryja, et al., 2007).
Experiments in mammalian cell culture have further implicated βarrestins in
developmental pathways. Dishevelled1/2 were identified as βarr1-interacting proteins
in the human heart, and were found to regulate β-catenin stabilization and LEFmediated transcription in HEK-293 cells (Chen, et al., 2001). In (non-canonical) Wnt5a
stimulation of the same cell type, Dvl2 recruits βarr2 to mediate endocytosis of Frizzled4
(Chen, et al., 2003a). β-arrestins have recently been implicated in multiple novel roles in
the TGFβ signaling pathway. After recruitment by TGFβRII, βarr2 mediates the
internalization of activated TGFβRIII to downregulate TGFβ signaling in HEK-293 cells
(Chen, et al., 2003b).
79
βarr2 is a positive regulator of Hh signaling during zebrafish embryogenesis, as
knockdown of βarr2 in the Zebrafish causes a reduction of Hh signaling (Wilbanks, et
al., 2004). In mammalian cell culture, βarr2 & GRK2 mediate clathrin-dependent
internalization of activated Smoothened to potentiate Shh signaling (Chen, et al, 2004).
Mammalian Smo translocates to the primary cilium in a ligand-dependent manner to
transduce hedgehog signaling (Corbit, et al., 2005). This ciliary localization of Smo is
dependent on a βarrestin/Kif3a/Smo complex (Kovacs, et al., 2008). Unlike most GPCRs,
all of these receptors with non-canonical β-arrestin interactions are essential for
embryonic development.
β-arrestin1 (Arrb1) and β-arrestin2 (Arrb2), the two mammalian non-visual
arrestins, are ubiquitously expressed in adult tissues. In the absence of β-arrestin1,
misregulated signaling through β-adrenergic receptors, angiotensin receptors, and many
other GPCRs has been reported by us and others (e.g. Conner et al., 1997; Kohout, et al.,
2001; Rajagopal, et al., 2006). Similarly, βarr2 knockout mice have an increased analgesic
response to morphine, due to misregulation of μ-opioid receptor desensitization (Bohn,
et al, 1999). Ablation of either βarr1 or βarr2 in the mouse has been associated with a
wide variety of adult physiological processes and disease states, diet-induced obesity
(Zhuang, et al., 2011), insulin resistance (Luan, et al., 2009), allergic asthma (Walker, et
al., 2003), lung tumor growth and angiogenesis (Raghuwanshi, et al., 2008), and
initiation and progression of myeloid leukemia (Fereshteh, et al., 2012). In contrast, lack
80
of either βarr1 or βarr2 does not prevent apparently normal mouse development
(Conner, et al, 1997; Bohn, et al, 1999), indicating either redundancy or non-requirement
during embryogenesis. As over 1600 GPCRs have been identified in the murine genome
(www.informatics.jax.org) and given the enormous size of the βarrestin interactome
(reviewed by Shukla, et al, 2011), unimportance seems unlikely.
Indeed, when βarr1 and βarr2 knockout strains were intercrossed, double
knockout (DKO) progeny were never observed. The GPCR/GRK/βarr paradigm is very
large, so many signaling events are potentially misregulated in the βarr DKO embryo.
Additionally, as few GPCRs are individually required for development, GPCR
misregulation causing βarr DKO lethality would likely be combinatorial. Furthermore,
in mammalian cell culture, knockdown of β-arrestins has been reported to cause
misregulation of a wide variety of global cellular processes, including DNA damage
response (Hara, et al., 2011), apoptosis (Ahn, et al., 2009), protein translation (DeWire, et
al., 2008), and actin filament severing (Zoudilova, et al., 2010).
4.4 Results
4.4.1 βarrestins are mostly dispensable for murine development
Given the number of reported βarrestin interactors across various contexts, there
were many possible phenotypes owing to complete absence of βarrestins. As such, we
sought to determine the earliest requirement for β-arrestins in mammalian development
by analyzing the β-arrestin DKO mouse embryo.
81
As in the adult mouse, both βarr1 and βarr2 are ubiquitously expressed in all
tissues examined at E12.5 and E15.5, with minor variances in relative expression (Figure
17A). To generate βarrestin1, βarrestin2 double knockout (βarr DKO) mice, we
intercrossed βarr1-/-; βarr2+/- mice generated from the original βarr1 (Conner, et al,
1997) and βarr2 (Bohn, et al, 1999) knockout lines, back-crossed 12 generations onto
C57Bl/6 (Figure 17B). As βarr DKO pups were never observed in the colony, we sought
to determine the timing of lethality. Due to the large number of signaling interactions
possibly disrupted, we were surprised that βarr DKO embryos were present at nearMendelian ratio at E18.5 (Figure 18A) and appeared viable and outwardly normal, other
than small in size (Figure 18B, Figure 17C). The wet weight of lung, kidney, and brain
were proportional to body weight (Figure 17DE and data not shown). Careful
observation revealed that DKO pups were born alive, indicating that βarrestins are not
required for embryonic viability.
As the ablation of both βarrs produced a significantly milder phenotype than
expected, we wanted to ensure that upregulation of SAG and Xarr (the two mammalian
visual arrestins, normally expressed in the eye) had not occurred in non-visual tissues.
Neither transcript was detected at significant levels in DKO or WT non-visual tissues at
E12.5 or E15.5 (not shown). Thus, gross compensation by ectopic visual arrestins had
not occurred in the absence of βarrestins.
82
Figure 17: Barr DKO verification
(A) βarr1 and βarr2 transcripts are expressed in all embryonic
tissues examined at E15.5, with minor variation in relative
expression level. (B) Mating strategy used to generate βarr DKO
embryos. (1) βarr1-/- and βarr2-/- strains were crossed to generate
double heterozygotes, which were backcrossed to βarr1-/- (2) to
generate “βarr triples”. (3) Non-consanguineous “βarr triples”
were intercrossed to generate the experimental litters. (C) βarr
DKO pups were significantly smaller than wild-type littermates at
birth (p<0.02) (D) βarr DKO lungs are smaller than wild-type
littermates’. (E) βarr DKO lungs are proportional to body size. (F)
Some βarr DKO pups exhibited abnormal subcutaneous bruising,
but no obvious internal bleeding. Scale bars indicate 2mm.
83
Figure 18: βarrestins are not required for embryonic viability
(A) Genotyping results indicate that βarr DKO embryos are
present at near Mendelian ratio and viable at E19.0. (B) βarr DKO
pups have reduced body size and weight at birth, but appear
outwardly normal. (C) While most βarr DKO pups survive two or
more hours after birth, all fostered DKO pups died within 24
hours. (D) The majority of both βarr DKO pups and littermates
established spontaneous breathing and inflated their lungs. Scale
bars indicate 5mm.
84
4.4.2 βarr DKO pups exhibit respiratory distress despite having
normal lungs
I wanted to determine the cause of the βarr DKO postnatal lethality. Due to a
high rate of maternal cannibalism, experimental litters were extracted by Cesarean
section and fostered to recently postpartum ICR dams.
Upon manual or maternal
stimulation most DKO pups gasped, moved, and squeaked, 75% established a regular
breathing pattern and normal oxygenation. Intriguingly, βarr DKO pups were twice as
likely to exhibit moderate to severe bruising after birth or C-section as littermates
(Figure 17G). However, βarr DKO pups became cyanotic after birth, and appeared to
die of respiratory distress.
No C-sectioned βarr DKO pups survived past 30 hours (Figure 18C), although a
very few DKO pups were observed in the colony at P2 and P3. Atelectasis, or failure to
inflate the lung, was observed in a similar percentage of C-sectioned DKO pups and
littermates (Figure 18D). Although βarr DKO lungs were normal in size and
appearance, defects in the lung epithelium can cause respiratory distress and/or
atelectasis. However, the βarr DKO lung appears histologically normal and functional
(Figure 19ABC), with ample surfactant (Figure 19DE), and no excess glycogen (Figure
19FG).
85
Figure 19: Normal histology and differentiation of the Barr DKO lung
(A) - Analysis of lung structure and histology demonstrates that inflated βarr
DKO lungs (A') were histologically identical to those of WT littermates. (B) - The
number, size, and distribution of terminal air sacs, assessed in histological
section by number of alveolae per unit area, is similar in βarr DKO and WT
littermate lungs. (C) - The average width of alveolar septa, not significantly
different in WT and βarr DKO lungs, is an inverse measure of lung maturity, as
the septal walls remodel and thin immediately before birth. (D) - The epithelium
of βarr DKO lungs appears histologically mature at birth, with ample surfactantC (labeled in green) is present in the medial and distal alveolar compartment
(identified by lack of underlying pulmonary smooth muscle layer, labeled with
anti-smooth muscle actin (red). (E) - The absolute proportion of mature type II
pneumocytes, identified by expression of SP-C (green in D) and normalized to
the total number of nuclei (blue in D), is normal in the βarr DKO lung. (F) Pulmonary maturity, as measured by persistence of embryonic glycogen in the
postnatal airways, can be quantified by PAS-Schiff staining (purple) in frozen
section. No excess glycogen (arrowheads) was apparent in the distal airways of
P0 WT or βarr DKO lungs after 1-2 hours of spontaneous breathing. (G) - The
overall percentage of PAS+ cells is normal in βarr DKO lungs 1-2 hours
postnatally. Scale bars indicate 500 µm in A and 100 µm in D and F.
86
4.4.3 Cleft secondary palate is responsible for most βarr DKO
lethality
Upon necropsy, large amounts of air, and very little milk, were found in the
stomach and intestine of DKO pups (Figure 20A), indicating a defect in the respiratory
and/or alimentary systems. Examination by serial histological section revealed no
tracheoesophageal fistulas or laryngeal clefts (not shown), which would allow passage
of inhaled air into the stomach. However, 83% of βarr DKO pups have an overt cleft
secondary palate (Figure 20B), allowing free communication between the oral and nasal
cavities, and preventing isolation of the respiratory and alimentary systems. In the βarr
DKO pups with cleft palate, every swallow included a substantial amount of air. While
all newborn mouse pups swallow air during the first few postnatal hours, βarr DKO
pups were unable to prevent massive amounts of air from entering the stomach.
Upon further examination of skeletal structure by Alcian blue and Alizarin red
staining, all βarr DKO pups have severe defects of the basisphenoid and palatine bones
and surrounding cartilages, which include and support the secondary palate (Figure
20CD) The palatine bone, dramatically hypoplastic in the DKO, normally forms the rear
of the hard palate. Even in DKO pups without overt cleft palate, the palatine hypoplasia
creates a condition referred to as occult or submucous cleft palate, in which the oral and
nasal cavities are only separated by mucous, not bony, tissue. The secondary palate
defects in the βarr DKO appear isolated, as careful measurement of other craniofacial
87
structures in whole-mount and skeletal preparations did not reveal any significant
differences (not shown).
88
Figure 20: Postnatal Barr DKO lethality is due to cleft secondary palate
(A) Shortly after birth, βarr DKO pups begin to accumulate air in
the stomach, eventually resulting in extreme distension of the
abdomen, due to large amounts of air in the stomach and
intestine. (B) The wild-type palate forms the roof of the mouth.
(C) A large cleft secondary palate (**) allows for free
communication between the oral and nasal cavities of the DKO
mouse (D) In a view of the base of the skull, cartilage is stained
with Alcian Blue and ossified bone with Alizarin Red. (E) These
structures, which comprise and support the palate, are markedly
hypoplastic in all DKO pups examined, even those without an
overt cleft secondary palate. (FG) This hypoplasia can be
recapitulated by knockdown of Shh directly in Shh-CreER/flox
embryos and tamoxifen administration at E10.5. 1°P, primary
palate; 2°P, secondary palate; CP, cleft palate; int, intestine; PS,
palatal shelf; st, stomach. Scale bars indicate 500 µm in A, D-G
and 5mm in B and C.
89
As cleft secondary palate appears to be largely responsible for neonatal βarr
DKO lethality, we sought to determine which cellular processes required for
palatogenesis are disrupted in the βarr DKO embryo. While palatal shelves appeared
normal at E12.5 (3/3), examination at later stages revealed defects both in size and
morphogenesis of the βarr DKO palate. In normal palatal development, bilateral palatal
shelves elevate above the tongue at E13.5 (Figure 21A) to meet at the midline by E14.5
(Figure 21C) where they fuse into a single palatal structure to separate the oral and nasal
cavities. In the absence of βarrestins, however, palatal development is disrupted at
multiple stages. At E13.5, many βarr DKO palatal shelves had failed to coordinately
elevate above the tongue (Figure 21B). Furthermore, examination at E14.5 revealed that
DKO palatal shelves that did elevate were still markedly hypoplastic and do not meet at
the midline (Figure 21D). Both of these defects in elevation and medial growth lead to a
failure of palatal fusion, causing cleft palate.
90
Figure 21: Hypoplastic and dysmorphic palatal shelves in the Barr DKO
(A) At E13.5, littermate palatal shelves (PS) have elevated above
the tongue and grow towards the midline. (B) In some βarr DKO
embryos, the palatal shelves fail to coordinately elevate. (C) By
E14.5, wild-type palatal shelves meet at the midline to separate
the oral and nasal cavities. (D) Most βarr DKO palates are
markedly hypoplastic, resulting in a dramatic cleft palate (** in D)
Chondrogenesis, marked by Sox9 (red), does not appear to be
impaired. L-PS, left palatal shelf, MEE, midline epithelial edge;
NC, nasal cavity; NS, nasal septum; OC, oral cavity; R-PS, right
palatal shelf; T, tongue. Error bars indicate 250 µm.
91
4.4.4 Barr DKO cleft palate corresponds to decreased Shh signaling
Unfortunately, the combination of palatal phenotypes, incomplete penetrance of
cleft palate, and large number of signaling pathways implicated in both cleft palate and
βarrestin involvement made identification of the signaling interactions misregulated in
the βarr DKO palate difficult. We therefore used an unbiased approach, employing
signal transduction PathwayFinder qPCR arrays (SABiosciences) to analyze the
expression level transcriptional targets for 18 distinct signal transduction pathways. Of
84 transcriptional targets present on the qPCR array, 22 were significantly
downregulated (p<0.05) in the E13.5 DKO palatal shelves (Figure 22A). Of these,
Hedgehog signaling was the most prominent pathway, with multiple target genes
downregulated (Figure 22B). Thus, it appears that misregulation of Hh signaling in the
absence of βarrestins leads to cleft palate in the DKO embryo.
Because all βarr DKO pups have skeletal abnormalities in the palatal region, it is
possible that these defects produce cleft palate in the majority of βarr DKO embryos.
Thus, the decrease in Shh target gene expression may be an unrelated phenomenon on
top of a preexisting cleft palate. To determine whether the skeletal hypoplasia and cleft
palate are etiologically linked or isolated events, we attempted to recapitulate the
skeletal phenotype by specifically modulating Shh expression in wild-type mice. To do
this, we used a tamoxifen-inducible cre driver driven by the Shh promoter, Shh-CreER,
to reduce Shh expression by tamoxifen administration at E10.5. While no cleft palate
92
was observed, all pups with reduced Shh expression at E10.5 have skeletal hypoplasia
(Figure 20FG) identical to that in the βarr DKO, indicating that reduced Shh expression
can account for the entire spectrum of palatal phenotypes seen in the βarr DKO mice.
93
Figure 22: Shh signaling is decreased during development of the Barr DKO palate
Downregulation of Hedgehog pathway targets in E13.5 βarr DKO palate.
94
4.4.5 βarr DKO pups also exhibit cardiovascular, skeletal, and renal
defects
In addition to cleft palate, βarr DKO pups exhibit a number of other phenotypes,
either subtler in nature or at lower penetrance. While the normal aorta arches to the left
(Figure 23A), 41% of βarr DKO pups have a right-sided aortic arch (Figure 23B). Of
these, 75% have a retro-esophageal left subclavian artery (Figure 23C) and 30% have
persistent truncus arteriosus (not shown). However, circulation appears otherwise
normal in the majority of these mice and the presence of cardiac defects did not correlate
with postnatal survival.
Occasionally DKO pups had cardiac defects which would have contributed to
early postnatal lethality. In one βarr DKO pup, a double aortic arch created a vascular
ring which may have compressed the trachea and esophagus (not shown). However,
the vast majority of right-sided aortic arches were isolated, with no aberrant left
ligamentum arteriosus or association with tetralogy of Fallot. Small ventricular
septation defects were present in approximately 15% of DKO pups, but also would not
impact immediate postnatal survival.
While the majority of the βarr DKO skeleton is normal, all βarr DKO pups have
an isolated defect in lumbar vertebral development, including 85% with only an isolated
defect in the L4 vertebra (Figure 23E). The 5th ossification center of the sternum is
underdeveloped or absent in 7/9 DKO pups examined (Figure 23G).
95
6/9 βarr DKO pups had dysmorphic kidneys (Figure 24B), which often consisted
of little more tissue than one to three large cystic pockets (Figure 24D). Gross molecular
examination of laminin (for basement membranes) and E-cadherin (for epithelia)
outlines the normal glomerular structure of littermate kidneys (Figure 24E), and
demonstrates the near-complete lack of organization in the βarr DKO kidney (Figure
24F).
96
Figure 23: Cardiovascular and skeletal defects in Barr DKO pups
(A) In wild-type littermates, the cardiac outflow tract (OFT) is
septated into the pulmonary artery and aorta, which arches to the
left above the heart before continuing down into the abdomen. (B)
Approximately 50% of βarr DKO mice have an abnormal rightsided aortic arch, and 20% of these were associated with a failure
to septate the cardiac outflow tract, similar to the common human
birth defect, persistent truncus arteriousus. (C) Many (40%) of
βarr RAAs also had an aberrant retroesophageal left subclavian
(RELS) artery. (D) Wild-type littermate skeleton stained with
Alcian blue and Alizarin red. (E) Although reduced in size, the
βarr axial skeleton is generally indistinguishable from that of
littermates, however all βarr DKO pups have defects in
ossification of the 4th lumbar vertebra, compare D’ with E’, which
occasionally (~20%) extends to include L1-L4. (F) At birth, six
ossification centers are present in the wild-type sternum. (G) βarr
DKO pups also have a minor defect in the sternum, with a specific
loss of the 5th ossification centers, but no abnormalities in vertebral
or rib number. Ao, aorta; Es, esophagus; L4, 4th lumbar vertebra,
LAA, left aortic arch, LCC, left common carotid; LSC, left
subclavian; Pa, pulmonary artery; RAA, right-sided aortic arch;
RELS, retroesophageal left subclavian; Tr, trachea. Error bars
indicate 500 µm in A-C, D', and E' and indicate 2mm in D-G.
97
Figure 24: Renal defects in Barr DKO pups
(A) Wild-type littermate kidneys. (B) Many βarr DKO kidneys are
dysmorphic at birth. (C) Cross-section through wild-type kidney.
(D) βarr DKO kidneys typically contain one or two large fluidfilled sacs. (E) Normal glomerular structure is highlighted by
epithelially expressed E-cadherin (yellow) and laminin (green) in
the basement membrane. (F) βarr DKO kidneys lack almost all
recognizable organized structure. Error bars indicate 2mm in A
and B and indicate 200 µm in C-F.
98
4.4.6 Underfoliation of βarr DKO cerebella is due to reduced Shhstimulated proliferation.
A decrease or delay in cerebellar foliation was observed in βarr DKO pups.
While most cerebellar growth and foliation occurs postnatally, 3-5 distinct lobes were
observed in littermate cerebella at P0 (Figure 25A). In contrast, 4/6 βarr DKO cerebella
were markedly underfoliated (Figure 25B). As cerebellar foliation is driven by
proliferation of granule cell precursors (GCPs) in the external germinal layer (EGL) of
the cerebellum, we analyzed cerebellar cell proliferation at E18.5 by
immunohistochemistry against phospho-histoneH3 (pHH3) (Figure 25CD) or
bromodeoxyuridine (BrdU) one hour post labeling by intraperitoneal injection of the
pregnant dam. By both methods, a substantial (50%) decrease in proliferating cells was
observed in the EGL. As Shh is a known mitogen for GCPs, we examined markers of Hh
signaling and cell proliferation by qPCR. Hh transcriptional targets Ptch1 and Gli1 were
both downregulated in the DKO cerebellum, as were CyclinD1 and nMyc (Figure 25D).
Because the early postnatal lethality of βarr DKO pups precludes further analysis of the
cerebellar defect, we used an unbiased approach to determine whether concomitant
misregulation of another signaling pathway could cause the cerebellar proliferation
defect. Again, we made use of the signal transduction PathwayFinder qPCR arrays
(SABiosciences). Eleven transcripts were consistently downregulated in all samples
(Figure 25F). Supporting our hypothesis, downregulation of multiple Shh target genes
was the only unambiguous functional group of transcripts misregulated in the cerebellar
99
samples (Figure 25E). Interestingly, βarrestins may not only promote the mitogenic
response to Shh, but qPCR results hint at suppression of apoptosis in the βarr DKO
cerebellum.
100
Figure 25: Shh induced cerebellar neurogenesis is diminished in Barr DKO pups
(A) While most cerebellar foliation occurs postnatally, 3-5 distinct
lobes were observed in WT P0 cerebella. (B) On the other hand,
3/4 βarr DKO cerebella were markedly underfoliated. (C) Rapid
expansion of granule cell precursors (GCPs) in the external
germinal layer (EGL), known to drive cerebellar foliation, is
quantified by anti-pHH3 labeling (red) of proliferating cells. (D)
GCP proliferation in the βarr DKO cerebellum is dramatically
decreased. EGL, external germinal layer; GCP, granule cell
precursor; pHH3, phospho-histoneH3. Scale bars indicate 200
µm.
101
4.4.7 All βarr DKO phenotypes may be due to reduced mitogenic Shh
signaling
Given the association between loss of βarrestins and decreased mitogenic
hedgehog signaling in the cerebellum, we began to consider the consequences of a
globally suppressed response to Shh in the developing embryo. Nearly all of the βarr
DKO phenotypes observed at lower penetrance could be the consequence of decreased
Shh signaling. In particular, the cerebellar, palatal, renal, and cardiovascular defects are
all attributable to decreased mitogenic Shh signaling. In contrast, transcripts isolated
from the E12.5 limb bud, in which Shh acts as a patterning cue (Zhu, et al., 2008) did not
show a decrease in levels of Shh target genes (not shown).
4.5 Discussion and future directions
βarr1 and βarr2 are functionally redundant during embryonic development, as
both individual knockout lines are viable, while complete absence of βarrestins is
incompatible with postnatal life. βarr1 and βarr2 are individually critical for adult
homeostatic function and may be involved in many disease states, often through
misregulation of GPCRs. In addition to GCPRs, βarrs have many newly reported
functions including varied and novel interactions with many non-GPCRs essential for
embryonic development in the mouse.
An essential functional requirement for βarrestins in signaling through any of
these receptors would be reflected in the phenotype of the β arrestin1, β-arrestin2
102
double knock-out (βarr DKO) mouse. The consequence of complete ablation of
βarrestins in the mouse embryo was difficult to predict, given the number of known βarr
interactors in the genome. We expected to find many disparate defects resulting in
gestational lethality; the early phenotypes would reflect the earliest requirements for
βarrestins during embryogenesis. Here, we demonstrate that selective enhancement of
mitogenic Shh signaling is the first critical function of βarrestins during mouse
development.
βarr DKO embryos are present at E18.5 at near-Mendelian ratios, appearing
outwardly normal other than small size. Their presence indicates that βarrestins are
dispensable for most signaling events in the developing mouse embryo. βarr DKO
pups exhibited spontaneous movement and established normal breathing patterns,
initially indistinguishable from littermates. Only after initially “pinking up” did most
DKO pups begin to exhibit respiratory distress and cyanosis. Unlike previous reports
(Zhang, et al., 2010; Zhang, et al., 2011) we did not find atelectasis or immaturity of the
pulmonary epithelium to be a major contributor to postnatal βarr DKO lethality. All
inflated DKO lungs examined were histologically and molecularly indistinguishable
from littermates. However, a small percentage of DKO pups did not breathe at birth, or
died without establishing a regular breathing pattern, indicating multiple causes of
lethality from βarr ablation. The lungs of a single DKO animal that did not have cleft
palate or spontaneous breathing looked normal when examined by molecular markers
103
(data not shown). However, because of small sample size examined, it is possible that
βarrestins do contribute to the maturity of the lung epithelium in a subset of βarr DKO
embryos.
Approximately 80% of DKO embryos (and a small percentage of littermates)
have a severe cleft secondary palate, causing perinatal lethality due to respiratory
distress. Cleft palate is a common human birth defect (usually associated with cleft lip)
and often requires immediate surgical intervention at birth to prevent both respiratory
and alimentary difficulties in newborn babies. During normal palatal development,
cranial neural crest cells (CNCCs) migrate ventrally from the dorsal neural folds to
populate the first pharyngeal arch (PA1), where CNCCs join mesodermal arch
mesenchyme to proliferate rapidly and undergo morphogenesis into the ventral
viscerocranium, including the palatal shelves and mandible. The bilateral palatal
shelves continue to increase in size, then elevate above the level of the tongue, fuse at the
midline, and ossify to form the roof of the mouth, or palate. (Reviewed in Gritli-Linde,
2007.) Defects during any stage in this process result in a cleft, or hole, in the roof of the
mouth, allowing for free communication between the oral and nasal cavities. Early
defects in generation, proliferation, or migration of cNCCs into the first pharyngeal arch
(PA1) all result in PA1 hypoplasia and, consequently, cleft palate. A cNCC defect could
also account for the high incidence of cardiac outflow tract defects observed in βarr
DKO pups. However, βarr DKO mice do not display other phenotypes consistent with a
104
global defect in cNCC, such as PA1 hypoplasia and micrognathia. Indeed, Shh regulates
adhesion and migration of cNCCs independently of Ptch-Smo-Gli signaling (Testaz, et
al., 2001). Reduced proliferation, increased apoptosis, or dysmorphogenesis of palatal
shelves per se can also cause cleft palate. While no PA1 hypoplasia was seen at or before
E12.5, most DKO palates examined at E13.5-E15.5 were markedly hypoplastic.
Occasionally, DKO embryos had normal sized palatal shelves, but they failed to
coordinately elevate and extend laterally to the midline. Failure of epithelial fusion or
apoptosis at the midline can also cause cleft palate; however this is not likely to be
occurring in βarr DKO embryos, as no DKO palatal shelves observed at E15.5 had
progressed to meet at the midline, while all littermate palates had begun epithelial
fusion.
Just as disruption of many cellular processes can cause cleft palate, many
signaling pathways are involved. βarr signaling interactions may be very complicated,
especially given variability in early palatal phenotypes. We identified decreased Hh
signaling, in developing palate of βarr DKO embryos examined by targeted qPCR
against Ptch1 and Gli1, transcriptional targets of Hh signaling. These results are
consistent with previous studies demonstrating Shh is a central player in the complex
transcriptional network required for normal palatal development. Shh expression from
the palatal epithelium (regulated by homeobox transcription factor Msx1) is necessary
for Bmp2-mediated proliferation of palatal mesenchyme and maintenance of mitogenic
105
expression of cyclins D1 and D2 (Lan and Jiang, 2009). Furthermore, restoration of Shh
expression is key to rescuing the Msx1-null cleft palate by ectopic Bmp4 (Zhang, et al.,
2002). However, as a number of known βarr interactors are involved in palatogenesis, it
is possible that the observed decrease in Shh signaling is insufficient to cause cleft
palate.
A lower percentage of non-DKO littermates also displayed craniofacial
anomalies consistent with decreased Shh signaling: cleft palate (2/87), micrognathia
(3/87), and even complete cyclopia (1/87). These phenotypes are all generally consistent
with an inverse relationship between the absolute level of βarrestins and the magnitude
of the Shh signal transduced.
106
5. Roles for Shh in foregut morphogenesis
5.1 Summary
In the early embryo, the foregut is comprised of a single tube that
compartmentalizes into the trachea and esophagus via a poorly understood process.
Sonic hedgehog (Shh) is required for this process as the Shh-null foregut never divides
into two tubes (Figure 26). Using Cre-Lox recombination, I determine where and when
Shh is expressed for foregut division.
5.2 Introduction
In the midgestational embryo (E10.0-E11.5), the common foregut tube divides
into two tubes. The ventral tube further differentiates to form the trachea and the dorsal
tube forms the esophagus. In the Shh-null mouse, division of the foregut tube never
occurs. The Shh-null foregut tube remains entirely undivided, connecting to both the
lungs and stomach (Figure 26). This phenotype is incompatible with postnatal life as
the newborn cannot properly breathe and eat. Thus, some domain(s) of Shh expression
in the developing embryo are essential for foregut compartmentalization into trachea
and esophagus.
107
Figure 26: The Shh-null foregut tube does not divide into trachea and esophagus
Whole-mount E-cadherin staining outlines the foregut endoderm.
By E11.5 the wild-type foregut (left) has divided into ventral
tracheal and dorsal esophageal tubes. In the Shh-null mutant
(right) this does not occur and the foregut remains a single tube to
which the lungs and stomach both connect. Scale bars indicate
500 µm.
108
As diagrammed in Figure 27, Shh is expressed in the notochord from E7.5, in the
floor plate of the neural tube from E8.0, and in the ventral foregut endoderm (vFGE)
from E8.5. During the window of tracheoesophageal separation, Shh expression shifts
from the vFGE (future trachea) to the dorsal FGE (future esophagus). Using a Cre-Lox
approach, I aimed to genetically dissect these domains of Shh expression to determine
the relevant source of Shh in foregut compartmentalization. Furthermore, as Shh is a
secreted signaling molecule, I wanted to determine the tissue(s) that must receive the
Shh signal for correct division of the foregut tube
109
Figure 27: Domains of Shh expression in and around the developing foregut
Shh is first expressed in the notochord (red) around E7.5 and turns
on shortly thereafter in the floor plate of the neural tube (orange).
Shh is expressed in the ventral FGE (blue) at E8.5 and switches to
the dorsal FGE (green) during tracheo-esophageal separation
110
5.3 Results
5.3.1 Shh expression from the ventral FGE is not required during
foregut division
Immediately prior to foregut compartmentalization, Shh is expressed in the
vFGE. As this domain is closest spatiotemporally to the dividing foregut, I
hypothesized that Shh produced by the vFGE is required for foregut division. Nkx2.5Cre recombines in the vFGE domain by E9.5, labeling the ventral foregut endoderm and
mesoderm (Figure 28AB). Shh is not expressed in the foregut mesoderm, so the relevant
overlapping domain is Shh and Nkx2.5 expression in the vFGE. Thus, Nkx2.5-Cre;
Shhflox/null embryos lack endodermal Shh expression from E9.5
Surprisingly, these Nkx2.5-Cre; Shhflox/null mutants do not have defects in foregut
division. As seen at E18.5 in Figure 28DE, distinct tracheal and esophageal tubes are
present. These mutants do not resemble the Shh-null phenotype. Thus, I conclude that,
from E9.5 onwards, Shh expression from the vFGE is not required for foregut division.
111
Figure 28: Ablation of Shh from the vFGE does not prevent foregut division
(ABC) LacZ staining of the Nkx2.5-Cre domain in Nkx2.5-Cre;
R26R embryos. (A) The Nkx2.5-Cre domain is broad and
encompasses the entire rostrocaudal extent of the developing
foregut at E8.5, seen in sagittal view. (B) Nkx2.5-Cre is expressed
robustly in the vFGE and E10.5, the time of foregut separation.
(C) Long-term fate mapping reveals that the Nkx2.5-Cre domain
becomes the majority of the tracheal endoderm and mesoderm.
(D) Wild-type littermate has distinct trachea and esophagus when
examined in transverse section. (E) Ablation of Shh from the
vFGE with Nkx2.5-Cre does not impact foregut separation:
distinct tracheal and esophageal tubes are present although the
esophagus is small and mispatterned. Scale bars indicate 250 µm.
112
5.3.2 The dorsal-to-ventral shift in endodermal Shh expression is not
required for foregut division
Surprised by this result, we next investigated whether the ventral-to-dorsal shift
in Shh expression is pertinent to compartmentalization. This domain appeared relevant
as Shh expression shifts from the ventral to the dorsal side of the foregut as it
compartmentalizes. To disrupt this shift, I utilized a constitutively active form of Smo
which produces ongoing Hh signaling even in the absence of a Hh ligand, as the CASmo allele is immune to repression by Patched. I first created ectopic Hh signaling in
the future tracheal mesoderm by crossing Mef2C-Cre to CA-Smo. (The Mef2C-Cre
recombination domain can be seen in Figure 42). Constitutive activation of Hh signaling
in the ventral foregut mesoderm, which would normally cease around E11.0, did not
create a foregut phenotype (not shown). Similarly, activation of Smo throughout the
vFGE and ventral mesoderm with Nkx2.5-Cre (that is Nkx2.5-Cre; CA-Smo embryos)
did not produce a compartmentalization phenotype (Figure 29) Together these data
indicate that cessation of Shh expression in the future trachea is not required for foregut
division.
It was possible that turning on expression in the dFGE (future esophagus) was
instead the critical domain of Shh expression. To address this case, I next activated Smo
throughout the entire foregut (ventral and dorsal) endoderm and mesoderm with
Foxg1-Cre (Foxg1-Cre; CA-Smo embryos). This genetic manipulation too failed to
produce a foregut compartmentalization phenotype.
113
Finally, to address the possibility that axial domains (instead of endodermal) of
Shh expression were relevant to foregut division, I crossed the CA-Smo line to one
carrying Cre under the Shh promoter (i.e. Shh-GFP-Cre; CA-Smo). In these mutant
embryos, there is constitutive activation of Hh signaling in all Shh-producing cells,
including the notochord and floor plate. As this manipulation did not produce foregut
compartmentalization defects, I concluded that no change in autocrine HH signaling
impacts foregut division. Collectively, these data indicate that the ventral-to-dorsal shift
of Shh expression in the developing foregut is not required for foregut division. No
ectopic activation of Hh signaling was able to recapitulate the Shh-null foregut
compartmentalization defect.
114
Figure 29: Constitutive activation of Hh signaling does not impair foregut
division
(AB) Both wild-type and mutants in which the ventral foregut
endoderm and mesoderm constitutively signal through Smo
(Nkx2.5-Cre; CA-Smo) have normally divided foregut tubes. (C)
Summary of results from Smo-activation experiments. Scale bars
indicate 250 µm.
115
5.3.3 Shh from the floorplate is not required for foregut division
Further attempts to address early endodermal Shh were hampered by lack of an
early-endodermal Cre line that does not also express in the notochord. Sox17-iCre is
reported to recombine specifically in the endoderm from E8.0. However, we find that
50% of notochordal cells ectopically express Cre (Figure 30A). Importantly, Sox17i-Cre
does not recombine in the floor plate of the neural tube. Thus, this Cre line addresses
only the endodermal and notochordal domains of Shh expression.
When Sox17-iCre is used to ablate Shh (Sox17-iCre; Shhflox/null embryos), 100% of
mutants have an undivided foregut tube (Figure 30D). Thus, the domain of Shh
expression relevant to foregut division is either the endoderm or notochord, not the
floor plate of the neural tube.
116
Figure 30: Sox17-iCre; Shhflox/null have and undivided foregut but also notochord
recombination
(A) Sox17-iCre recombines robustly in the endoderm (Sox17-iCre;
R26R) but also in approximately 50% of notochord cells. (BC)
Wild-type littermate foregut tube has properly divided into
trachea and esophagus. (D) Ablation of Shh from endoderm and
50% of notochord with Sox17-iCre (Sox17-iCre; Shhflox/null) results
in an undivided foregut tube. Scale bars indicate 2.5mm in A and
250 µm in B-D.
117
5.3.4 Shh expression from the notochord is not required during
foregut division
The results from the previous experiments led me to consider whether the
relevant domain of Shh expression was actually axial tissues (the notochord and floor
plate of the neural tube) and not in the foregut per se. To investigate the role of Shh
expression from axial tissues, I used Foxa2-NFP-Cre, in which a fragment of the Foxa2
promoter is reported to drive recombination specifically in the notochord and floor plate
(NFP) (Figure 31AB), to ablate Shh in these domains. Unfortunately this Cre also drove
ectopic expression scattered throughout the endoderm, as seen in Figure 31C.
When I ablated axial Shh with Foxa2-NFP-Cre (Foxa2-NFP-Cre; Shhflox/null
embryos) two mutant phenotypes emerged (Figure 31 DEF). Approximately 40% of
mutants resembled a Shh-null, including complete holoprosencephaly. The other 60%
displayed a posterior truncation phenotype. Both classes of mutants had similar degrees
of ectopic endodermal recombination (compare Figure 31 D'' to E''). Thus, the more
severe Shh-null-like phenotype is not simply due to greater recombination in the
endoderm. More likely, this Cre driver has variability in the onset of expression and
may have turned on earlier in the Shh-null class of mutants.
118
Figure 31: Ablation of Shh from the notochord and floor plate
(AB) Normal recombination pattern of Foxa2-NFP-Cre in the
notochord and floor plate of the neural tube. (C) Ectopic
endodermal recombination found in the ventral trachea. (D) Shhnull-like class of Foxa2-NFP-Cre; Shhflox/null mutants exhibit small
size and holoprosencephaly. (E) Posterior truncation class of
Foxa2-NFP-Cre; Shhflox/null mutants. (F) Wild-type littermate. Scale
bars indicate 200 µm in A-C and indicate 5mm in D-F.
119
Internally, none of the posterior truncation mutants had defects in foregut
compartmentalization (not shown). Thus, as this Cre is robustly expressed in axial
tissues from E8.5, Shh expression from notochord and floor plate is dispensable for
correct formation of the trachea and esophagus. Intriguingly, half of the Shh-null class
of mutants had an undivided foregut, like the true Shh-null (Figure 32B). However, all
of these mutants also had ectopic endodermal recombination, precluding analysis of the
origin of the relevant Shh ligand. Despite this fact, these data do hint that the temporal
role of Shh is very early in foregut development. As discussed above, the Foxa2-NFPCre likely turns on earlier in the Shh-null class of mutants than in the posterior
truncation class, missing the critical window.
Collectively, these results eliminate multiple spatiotemporal domains of Shh,
allowing me to narrow in on the domain necessary for foregut compartmentalization :
the early vFGE, before E9.5 (Figure 33). The vFGE after E9.5 is not required, as Nkx2.5Cre; Shhflox/null have a separate trachea and esophagus. The notochord and floor plate are
also not the relevant domains, as Foxa2-NFP-Cre; Shhflox/null do not have consistent
foregut defects. (The low incidence of foregut defects observed is likely attributable to
ectopic endodermal recombination.) Foxg1-Cre and Sox17-iCre, which both express in
the early foregut endoderm, are the only manipulations that recapitulated the Shh-null
foregut defect. Thus, by process of elimination, the early (before E9.5) vFGE is the
source of Shh required for foregut compartmentalization.
120
Figure 32: 50% of Shh-null type have an undivided foregut
Ventral and dorsal views of the foregut phenotype found in half
of the Shh-null class of Foxa2-NFP-Cre; Shhflox/null mutants. The
foregut tube has not divided into the trachea and esophagus,
remaining a single tube like that found in the true Shh-null. Scale
bars indicate 2.5 mm in A, A', B, and B', and indicate 1mm in A''
and B''.
121
Figure 33: The single remaining spatiotemporal domain of Shh expression required
for foregut division is the early vFGE
122
5.3.5 A decrease in Shh expression at E8.5 produces bronchoesophageal fistulas
In an attempt to ablate Shh expression from the vFGE prior to
compartmentalization, we used Shh-CreER; Shhflox/null embryos with successively earlier
tamoxifen administration. Tamoxifen before E8.5 did not recombine in the endoderm
and administration after E8.5 did not produce a compartmentalization phenotype (Table
3). However, tamoxifen at E8.5 (recombination between E9.0-E10.0) does produce a high
incidence of broncho-esophageal fistulas, an aberrant connection between the distal lung
and caudal esophagus (Figure 34). Unfortunately, tamoxifen administration earlier than
E8.5 caused gross lethality of mutant and wild-type embryos, precluding analysis of the
earlier Shh knockout phenotypes. Thus, even though these experiments did not produce
an undivided foregut tube, they hint at a critical window for the role of Shh in
compartmentalization just prior to E9.5. Broncho-esophageal fistulas are a mild foregut
compartmentalization defect that likely occur as Shh is ablated during the middle of a
critical window.
123
Table 5: Temporal ablations of Shh
Tamoxifen
administration
E6.5
E7.5
E8.0
E8.5
E9.5
E10.5
Recombination
window
E7.0 - E8.0
E8.0 - E9.0
E8.5 - E9.5
E9.0 - E10.0
E10.0 - E11.0
E11.0 - E12.0
Recombination
domain
NC
NC (NFP)
NFP
NFP & vFGE
NFP & vFGE (dFGE)
NFP & dFGE
Shh KO
phenotype
tamox. lethality
tamox. lethality
tamox. lethality
BEF (50%)
None
None
BEF, broncho-esophageal fistula; dFGE, dorsal foregut endoderm
NC, notochord; NFP, notochord and floor plate; vFGE, ventral
foregut endoderm
124
Figure 34: ShhCreER/flox with tamoxifen at E8.5 produces broncho-esophageal fistulas
(ABC) Broncho-esophageal fistula, or aberrant connection
between the lung and caudal esophagus, produced by ablation of
Shh after E8.5 tamoxifen injection. (B) Ventral view. (C) Dorsal
view. Scale bars indicate 250 µm.
125
5.3.6 Hh reception within the FGE is not required for foregut division
As Shh is an endodermally secreted ligand, to fully understand the role of Shh in
foregut compartmentalization I must further determine the spatiotemporal location of
Shh reception and signaling. Shh is often involved in epithelial-to-mesenchymal
signaling, so the foregut mesoderm is a likely target of Shh reception. On the other
hand, a widely used Shh-reporter allele, Ptch-LacZ, is weakly expressed in the foregut
endoderm prior to compartmentalization, indicating a low level of Hh signaling within
the FGE per se. Thus, it is also possible that Shh produced by the endoderm must signal
back to the endoderm in an autocrine fashion to permit foregut division.
To determine the role of Shh signaling in the endoderm, I ablated Smo
throughout the endoderm with two different Cre lines: ShhGFP-Cre (Figure 35) and
Sox17-iCre (not shown). Ablation of Smo with these Cre drivers specifically inhibits
autocrine Shh signaling within the endoderm, but does not impair endodermal Shh
secretion or Shh signaling to the mesoderm. As seen in Figure 35, ablation of autocrine
Shh signaling (ShhGFP-Cre; Smoflox/null embryos) does not prevent foregut division.
Throughout the foregut, distinct tracheal and esophageal tubes are present. Identical
results were produced with Sox17-iCre (Sox17-iCre; Smoflox/null embryos, not shown). As
neither manipulation produced an undivided foregut, I concluded that Shh signaling
within the FGE is not required for foregut compartmentalization.
126
Figure 35: Autocrine endodermal Shh signaling is not required for foregut division
(A) Throughout the rostrocaudal length of the wild-type foregut,
both trachea and esophagus are present. (B) When autocrine Shh
signaling is ablated (ShhGFP-Cre; Smoflox/null embryos), foregut
division also occurs normally. The entire foregut, from thymus
(A) to the division of the main bronchi (A'') contains distinct
tracheal and esophageal tubes. Br, bronchi; Es, esophagus; OFT,
cardiac outflow tract; Tr, trachea. Scale bars indicate 250 µm.
127
5.3.7 Hh reception is not required in the ventral foregut mesoderm
after E9.5
As Shh signaling within the endoderm is not important for foregut division, it
appeared likely that Shh signals directly to the surrounding tissue, the foregut
mesoderm, for foregut compartmentalization. To investigate this possibility, I aimed to
ablate Shh reception in this tissue. However, these experiments were hampered by lack
of an appropriate early mesodermal Cre driver.
I sought to address the role of Shh signaling in the mesoderm immediately prior
to and during foregut compartmentalization by ablation of Smo with Mef2C-Cre or
Nkx2.5-Cre. Both of these Cre lines recombine in the ventral mesoderm from E9.5, but
unfortunately are not expressed in the dorsal foregut. Thus, ablation of Smo in these
domains tests the role of Shh signaling specifically in the ventral foregut mesoderm,
immediately prior to and during foregut compartmentalization. Additionally, Nkx2.5Cre recombines in the vFGE, but Shh reception in that domain has already been
determined to be nonessential for foregut division (Section 5.3.6).
Neither manipulation (Mef2C-Cre; Smoflox/null embryos, not shown, or or Nkx2.5Cre; Smoflox/null, Figure 36) resulted in a compartmentalization defect. Separate tracheal
and esophageal tubes were present throughout the rostrocaudal length of the foregut.
Thus, the ventral foregut mesoderm after E9.5 is not likely to be the location of Shh
signaling required for foregut division.
128
By process of elimination, Shh signaling is thus required in either early ventral
foregut mesoderm (before E9.5) or late dorsal mesoderm for foregut
compartmentalization. As Shh expression appears to be an early requirement in foregut
separation (Section 5.3.5), I concluded that Shh reception is likely required in the ventral
foregut mesoderm prior to E9.5.
129
Figure 36: Shh reception in the ventral foregut mesoderm is not required for foregut
compartmentalization after E9.5
(A) Wild-type trachea and esophagus. (B) When Shh reception is
ablated in the ventral foregut endoderm and mesoderm (Nkx2.5Cre; Smoflox/null embryos), foregut division is not impaired and
separate tracheal and esophageal tubes are present throughout the
length of the foregut. Es, esophagus; Tr, trachea. Scale bars
indicate 250 µm.
130
5.4 Discussion and future directions
Altogether, the combination of Cre-Lox genetic manipulations in this Chapter
reveal that Shh expression from the early vFGE signals directly to the foregut
mesoderm. Axial Shh expression (from the notochord and floor plate) was eliminated as
the relevant domain, as was autocrine Shh signaling within the FGE. Only
manipulations that had strong endodermal recombination before E8.5, such as Foxg1Cre; Shhflox/null and Sox17-iCre; Shhflox/null produced foregut compartmentalization defects.
While it may not be surprising that endodermal Shh expression regulates rostral
endodermal compartmentalization, the timing of this event is intriguing. Nkx2.5-Cre;
Shhflox/null embryos, in which the vFGE domain of Shh expression is eliminated after E8.5,
do not have foregut compartmentalization defects. The requirement for Shh expression
and reception appears to occur before E8.5, a full day and a half before the foregut
begins to divide. These results are strengthened by the temporal ablations of Shh
(ShhCreER/flox embryos), in which mild compartmentalization defects (bronchoesophageal
fistulas, BEFs) are present only with tamoxifen administration at E8.5.
The creation of BEFs in these embryos is interesting, as these defects do not easily
reconcile with any proposed model for foregut separation. While it is debated whether
the foregut compartmentalizes by tracheal outgrowth or an ascending
tracheoesophageal septum, it is well established that the division process occurs in a
caudal-to-rostral process. Thus, one would expect an interruption midway through
131
foregut division to create only rostral compartmentalization defects, and BEFs are, by
definition, at the caudal end of the foregut. Therefore, BEFs are unlikely to be the result
of aborted foregut division. Instead, the unknown role of Shh in foregut
compartmentalization may, in fact, occur in a rostral-to-caudal manner, with
interruptions in this process creating caudal defects. It would be most interesting to
analyze foregut compartmentalization in ShhCreER/flox embryos and tamoxifen
administration before E8.5. I would hypothesize that earlier ablations of Shh would
create more severe foregut defects, such as tracheoesophageal fistulas or even an
undivided foregut tube, as seen in the Shh-null mouse. Unfortunately, tamoxifen
lethality precluded these experiments.
While I did not determine the molecular function of Shh signaling in foregut
compartmentalization, this work is a first essential step in understanding the role of Shh
in this process. By establishing where and when Shh signals for foregut division, future
researchers will know the correct spatiotemporal domains of Shh expression and
reception, and thus will begin analysis of the function of Shh in the foregut at earlier
developmental timepoints.
132
6. Roles for Shh in development of foregut mesoderm
6.1 Summary
In addition to division of the single foregut tube into two, both the trachea and
esophagus must be correctly patterned and increase in size. This not only includes the
foregut endoderm, but also the surrounding mesoderm, which differentiates into many
tissues, including smooth muscle and cartilage. During periods after foregut
compartmentalization, Shh is expressed in both the trachea and esophagus. I
hypothesized that these later domains of Shh expression regulate the growth and
development of the trachea and esophagus. In this Chapter, I reveal varied and complex
roles for Shh in development of tracheal cartilage and esophageal size and patterning.
6.2 Introduction
To properly form a trachea and esophagus, the endodermal foregut tube not only
must compartmentalize, but the surrounding foregut mesoderm must differentiate into
mature tissues. For the esophagus, the mesoderm becomes concentric layers of tissue: a
mostly acellular submucosal layer just beneath the epithelium, two rings of smooth
muscle, vasculature, and two enteric plexuses. The ventral tracheal mesoderm
differentiates into cartilaginous rings that encompass the ventral two-thirds of the
trachea and the dorsal tracheal mesoderm becomes smooth muscle, the trachealis.
Finally, all of these tissues must proliferate rapidly to grow in size with the developing
embryo.
133
As the Shh-null foregut tube lacks both smooth muscle and cartilage, I
hypothesized that, in addition to its role in foregut compartmentalization, Shh
expression in the developing trachea and esophagus regulates differentiation of foregut
mesoderm. Immediately after foregut compartmentalization (E10.5), Shh is expressed in
the esophagus and Shh is re-expressed at low levels in the trachea from E12.5. However
the roles of these expression domains in later esophageal and tracheal development have
not been elucidated.
To analyze the functions of these later domains of Shh expression, I used Cre-Lox
technology to spatially and temporally manipulate Shh expression and reception in both
endodermal and mesodermal tissues of the trachea and esophagus.
6.3 Results
6.3.1 Shh is required for esophageal size and patterning
In Chapter 5, I established that the esophageal domain of Shh expression, present
from E10.5, does not play a role in foregut compartmentalization. I hypothesized that
this domain was responsible for aspects of later esophageal development. To test this
possibility, I temporally ablated Shh expression from the esophagus after the foregut
had compartmentalized with ShhCreER. As discussed in Section 5.2, the foregut divides
into the trachea and esophagus from E10.0-E11.0. Thus, by administering tamoxifen at
E10.5 and allowing for a 24 hour recombination window, I prevented the esophagus
from expressing Shh from E11.5 onwards.
134
In wild-type littermates at E15.5, the esophagus is roughly the same size as the
trachea, is surrounded by a thick submucosal layer, and has two rings of smooth muscle
(Figure 37A). However, ablation of Shh expression from the developing esophagus
(ShhCreER/flox embryos, tamoxifen at E10.5) causes a small, mispatterned esophagus that
lacks two distinct smooth muscle rings and has a diminished submucosal layer. Thus,
Shh expression from the esophageal endoderm after E11.5 appears to play a role in
regulation of esophageal size and radial patterning.
135
Figure 37: Shh expression in the developing esophagus regulates esophageal size and
patterning
(A) Wild-type trachea and esophagus at E15.5 are similar in size
and two distinct layers of smooth muscle are visible (green, αsmooth muscle actin). (B) When Shh expression is specifically
knocked down in the esophageal endoderm (ShhCreER/flox embryos
and tamoxifen at E10.5), the mutant esophagus is small in size,
lacks distinct muscular rings, and has diminished submucosal
mesenchyme. (Yellow, Sox9 and E-Cadherin; blue, DAPI). Es,
esophagus; Tr, trachea. Scale bars indicate 250 µm.
136
6.3.2 Ablation of autocrine Shh signaling within the esophageal
endoderm causes esophageal dilation
Shh produced by the endoderm is essential for patterning of the esophageal
mesoderm and controls esophageal size. However, the tissue target(s) of the Shh signal
were unknown. I hypothesized that endodermal Shh is received by the mesoderm to
regulate patterning and esophageal size. However, it was also possible that the
endoderm per se requires Shh signaling in an autocrine fashion to produce a secondary
signal that controls esophageal size and patterning.
To determine whether esophageal mesoderm patterning defects were direct (Shh
signaling from endoderm to mesoderm) or indirect (autocrine Shh signaling within
endoderm), we ablated Shh signal reception within the esophageal endoderm (ShhGFPCre; Smoflox/null embryos). Surprisingly, ablation of autocrine Hh signaling produced
an opposite phenotype to ablation of esophageal Shh: a dilated, mispatterned
esophagus lacking submucosal mesenchyme and distinct smooth muscle rings (Figure
38B). These data indicate that autocrine Shh signal transduction within the endoderm
regulates esophageal dilation and mesodermal patterning. Thus, this function of Shh is
indirect; a secondary signal from the endoderm, regulated by autocrine Hh signaling,
controls these processes.
Ablation of the ligand (Nkx2.5-Cre; Shhflox/null embryos, Figure 37) does not
phenotypically resemble endodermal ablation of the receptor (ShhGFP-Cre; Smoflox/null
embryos, Figure 38), but both produce defects in mesodermal patterning. Thus, the
137
roles of Shh in esophageal patterning must be complex. It is likely that Shh from the
endoderm also signals directly to the mesodermal layers for patterning and regulation of
esophageal size. Unfortunately, no Cre driver exists specifically for esophageal
mesoderm, so I could not directly test this possibility.
138
Figure 38: Ablation of autocrine Shh signaling within the foregut endoderm causes
dilation of the esophagus
(A) In an E18.5 wild-type embryo, the esophagus is much smaller
than the trachea. (A') Distinct, concentric rings of mesoderm
surround the esophagus. (B) When Shh reception in the
endoderm is ablated (ShhGFP-Cre; Smoflox/null embryos), the
trachea and esophagus are approximately equal in size. (B')
Distinct layers of mesoderm are not visible surrounding the
mutant esophagus. Scale bars indicate 500 µm.
139
6.3.3 Overactivation of autocrine Hh signaling within the esophageal
endoderm produces submucosal vascular defects
As ablation of autocrine Shh signaling within the esophageal endoderm
produced a dilated esophagus, I hypothesized that overactivation of Hh signaling
within that tissue would produce an opposite phenotype. To do so, I utilized the CASmo line, a Cre-inducible mutant allele of Smo that is immune to inhibition by Patched,
resulting in ectopic Hh signal transduction even in the absence of the Shh ligand. To
overactivate Hh signaling specifically in the endoderm while leaving mesodermal Hh
signaling intact, I again used ShhGFP-Cre, this time to activate the CA-Smo allele.
Overactivation of autocrine Hh signaling (ShhGFP-Cre; CA-Smo embryos) did
not affect esophageal size (compare Figure 39A to B). Furthermore, development of the
two smooth muscle rings (green in Figure 39AB) was not impaired. Histologically, the
esophagus of these mutants looked indistinguishable from wild-type. However, when
examined immunohistochemically with anti-laminin, differences between wild-type and
mutant were apparent.
At E15.5, the wild-type esophagus (Figure 39B) has approximately 15 round
laminin-positive structures (red in Figure 39AB) in the submucosal layer, between the
inner muscular ring and the basement membrane of the endoderm (which also contains
laminin). These laminin-positive bodies are presumptive vascular structures that will
form the submucosal blood vessels of the esophagus. On the other hand, in ShhGFP-
140
Cre; CA-Smo mutants, far fewer of these structures were observed (Figure 39B). When
quantified (Figure 39C), these results were statistically significant at p<0.05.
While these results were not pursued further, they indicate that ShhGFP-Cre;
CA-Smo mutants do have subtle defects in development of submucosal mesoderm in
the esophagus. As Hh signaling was only overactivated in the endoderm, not the
mesoderm, these data indicate that excess autocrine endodermal Hh signaling produces
a secondary signal, likely secreted, that influences development of the submucosal
vasculature.
141
Figure 39: Over-activation of Hh signaling within the foregut endoderm causes a
reduction of the presumptive esophageal vasculature
(A) The wild-type esophagus has two distinct muscular rings (αsmooth muscle actin, green) and approximately 15 lamininpositive bodies (presumptive developing vasculature, red) in the
submucosal layer. (nuclei are stained with DAPI, blue). (B)
Ectopic over-activation of Hh signaling in the endoderm (ShhGFPCre; CA-Smo embryos) produces a dramatic reduction in lamininpositive submucousal bodies (C) Quantification of presumptive
vascular bodies. Scale bars indicate 250 µm.
142
6.3.4 Shh expression from the ventral FGE is required for tracheal
cartilage development
As Shh plays a critical role in establishment of esophageal mesoderm, I wanted
to determine whether Shh functions in later development of the other derivative of the
foregut tube, the trachea. While Shh expression turns off in the trachea during foregut
compartmentalization, it is re-expressed weakly from E12.5. During this time period,
the ventral tracheal mesoderm is transforming into cartilage. To analyze the role of Shh
in tracheal cartilage development, I performed Alcian blue staining on various Shh
mutant embryos.
Nkx2.5-Cre recombines in ventral foregut endoderm and mesoderm around E8.5.
This domain encompasses the majority of the future trachea, as seen by R26R lineage
trace (Nkx2.5-Cre; R26R embryos, Figure 40C). Shh is not expressed in the mesoderm,
so this Cre can be used to specifically ablate Shh expression from the vFGE.
In wild-type embryos, approximately 15 cartilage rings are present in the trachea
and more extend down the main bronchi (Figure 40B). However, in the absence of Shh
expression from the vFGE (Nkx2.5-Cre; Shhflox/null embryos, Figure 40C), the tracheal
cartilages are nearly absent and grossly mispatterned. Instead of evenly spaced rings,
the only cartilage present is a few midline blebs. Thus, Shh secreted by the vFGE is
required for tracheal cartilage development, likely controlling both patterning and
proliferation.
143
Figure 40: Shh expression from the vFGE is required for tracheal cartilage
development
(A) Nkx2.5-Cre labels the ventral tracheal endoderm and mesoderm
(Nkx2.5-Cre; R26R embryos). (B) Wild-type trachea has well organized
cartilage rings (C) When Shh is ablated from the vFGE with Nkx2.5-Cre
(Nkx2.5-Cre; Shhflox/null embryos), a severe tracheal cartilage phenotype
emerges. Very little tracheal cartilage is present and forms abnormal
ventral midline blebs. Es, esophagus; Tr, trachea. Scale bars indicate 250
µm in A and indicate 1 mm in B and C.
144
6.3.6 Shh signals directly to the foregut mesoderm for tracheal
cartilage development
While Shh expression from the vFGE is required for tracheal cartilage
development, the target tissue that must receive this signal was not known. As
prevention of autocrine Shh signaling within the esophagus severely impacted
development of esophageal mesoderm (Section 6.3.2) I wanted to determine whether
the Shh reception domain for tracheal cartilage was endodermal or mesodermal. The
tracheal mesoderm is derived from the ventral foregut mesoderm. Mef2C-Cre is
expressed in this tissue from E9.5, labeling the future tracheal mesoderm from which the
cartilage rings develop (Figure 41A).
Thus, to abolish reception of the Shh signal in this domain, I ablated Smo in
ventral foregut mesoderm with Mef2C-Cre (Mef2C-Cre; Smoflox/null embryos). As
anticipated, this manipulation severely impacted cartilage development. Overall,
significantly less cartilage is present in these mutants than in wild-type littermates
(Figure 41B). Furthermore, the cartilage that is present is severely mispatterned. Instead
of evenly spaced, separate ventral rings, the cartilage forms abnormal connections and
blebs. Thus, Shh signaling from the vFGE directly to the ventral mesoderm is required
for both expansion and patterning of the tracheal cartilage rings.
145
Figure 41: Shh signals directly to ventral foregut mesoderm for tracheal cartilage
development
(A) Mef2C-Cre is expressed in the ventral foregut mesoderm,
which becomes the trachea. Expression is absent in the endoderm.
(B) When Hh reception is ablated from the ventral tracheal
mesoderm (Mef2C-Cre; Smoflox/null embryos) the tracheal cartilage
rings are hypoplastic and mispatterned. Scale bars indicate 250
µm in A and 1 mm in B and C.
146
6.4 Discussion and future directions
In this Chapter, I investigated the roles of Shh expression and reception in later
development of the trachea and esophagus. Multiple functions were revealed,
indicating that Shh is critical not only for foregut compartmentalization, but also for the
growth and maturation of the trachea and esophagus. In both organs, Shh has key
functions in the development of mesodermal tissues.
In the esophagus, Shh is required for two separate functions: esophageal size and
patterning of the concentric rings of mesoderm. When Shh was ablated specifically from
the esophagus (ShhCreER/flox, tamoxifen at E10.5), the resulting esophagus was
dramatically reduced in size. Shh is a potent mitogen in many contexts, and likely
serves that function in the growing esophagus. Furthermore, the mesodermal rings
which surround the ShhCreER/flox esophagus were mispatterned or absent, as was the
mesoderm of the ShhGFP-Cre; Smoflox/null esophagus. Specifically, two distinct smooth
muscle rings were not visible and the submucosal layer was almost completely absent.
Together, these results indicate that Shh expression from the esophageal endoderm must
be received by the esophageal endoderm per se to correctly pattern the mesodermal
rings. Thus, this patterning event requires autocrine endodermal Shh signaling, which
must produce a secondary signal received by the mesoderm. By analogy to the role that
Shh plays in the small intestine (Sukegawa, et al., 2000), BMP4 is a likely candidate.
147
While endogenous levels of autocrine Shh signaling within the endoderm are
required for mesodermal patterning, the consequences of ectopically high Shh signaling
(ShhGFP-Cre; CA-Smo) were interesting. While one might expect an opposite
phenotype (expanded submucosa and smooth muscle), those layers of mesoderm
appeared unaffected. The only phenotype was subtle: a decrease in the presumptive
submucosal vasculature. Significantly fewer laminin-positive bodies were observed in
the submucosal layer at E15.5, revealing an unexpected consequence of constitutively
active Hh signaling within the endoderm. It will be of interest to examine these mutants
at birth to determine the functional consequence of this loss of vascular development.
Shh clearly plays a critical role in growth of the esophagus, as the size of the
ShhCreER/flox esophagus, in which Shh expression from the esophagus is ablated, was
dramatically reduced. On the other hand, removal of Shh reception from the endoderm
(ShhGFP-Cre; Smoflox/null created an opposite phenotype: a large, dilated esophagus.
Thus, the role of Shh in the regulation of esophageal size is complex and may be
multiphasic. The cellular processes underlying the dilated esophagus remain to be
determined, with at least two possibilities to consider. Most simply, ablation of
autocrine endodermal Shh signaling may somehow promote proliferation. However,
histologically, the esophagus appeared more dilated, not just increased in size: the
normal infoldings of esophageal endoderm were absent, resulting in a completely round
esophagus. This phenotype is reminiscent of Hirschsprung's disease, in which NCCs do
148
not migrate into and ennervate the colon. As Hh-mediated gliogenesis has been
implicated in this disorder (Ngan, et al., 2011) it is possible that ablation of autocrine Shh
signaling in the esophagus creates an analogous esophageal phenotype. It will be of
interest to determine whether NCCs properly migrate and differentiate into ganglion
cells in the myenteric and submucosal plexus of the esophagus.
With regards to the trachea, Shh does not appear to play a critical role in
regulating tracheal size. Ablation of Shh from the vFGE and trachea with Nkx2.5-Cre
(Nkx2.5-Cre; Shhflox/null embryos) does not impact the overall size of the trachea.
However, the cartilage rings which encompass the ventral side of the trachea are mostly
absent. Unlike in the esophagus, I revealed that this signaling is direct from the
endoderm to the mesoderm; there is no endodermal autocrine component in this
process.
Not only is less cartilage present in these mutants, the existing cartilage is
mispatterned. Instead of segmented rings, the cartilage is concentrated in blebs at the
midline. Thus, Shh appears to play a dual role in tracheal cartilage development:
regulation of chondrocyte proliferation or survival and segmented patterning. Indeed, a
role for Shh in patterning of the cartilage rings has previously been proposed (Park, et
al., 2010). In explant culture experiments, Shh patterns segmentation of Sox9 expression,
a marker of pre-chondrodytes. As this process can be modulated by exogenous addition
149
of BMP4 and Noggin (Park et al., 2010), it will be interesting to determine whether these
factors play an endogenous role in the embryo.
150
7 Discussion
7.1 Roles for Shh in foregut development
The timing of the requirement for Shh in foregut compartmentalization is
intriguing. If Shh does not play an active role during foregut separation, what is the
critical early function of Shh for foregut compartmentalization? One possibility is that
vFGE Shh is required to pattern the foregut tube into ventral (future trachea) and dorsal
(future esophagus) domains. These tissues can typically be labeled with Nkx2.1 and
Sox2, respectively. However, when the Shh-null foregut tube was analyzed in this
manner, the results were inconsistent. Both molecularly and histologically, Shh-null
foregut tubes contain regions that appear esophageal (Sox2-positive) and regions that
appear tracheal (Nkx2.1-positive), however, these regions are not simply divided into
dorsal and ventral domains. One would expect an undivided, but correctly patterned,
foregut tube to have a dorsal esophageal domain and a ventral tracheal domain.
Instead, entire portions of a Shh-null foregut tube may resemble an esophagus rostrally
and a trachea caudally, or vice-versa, and these patterns were not consistent between
embryos. Thus, while Shh must be important for foregut patterning, its role is not
simple.
Confounding this matter, Nkx2.1 does not only mark the future trachea, but also
labels thyroid tissue. In the Shh-null foregut, ectopic thyroid nodules are present
throughout the rostrocaudal length of the tube. (Fagman, et al, 2004). Thus, any domain
151
molecularly labeled with Nkx2.1 may actually be thyroid, not tracheal tissue. In fact, my
histological observations reveal that much of the Shh-null foregut has converted into
ectopic thyroid tissue. One interesting possibility is that this transformation is so
extensive that there is insufficient foregut remaining to create separate tracheal and
esophageal tubes, resulting in an undivided foregut. This model could be tested by
crossing the Shh-null mouse to an athyrotic line, such as the FGF10 knockout mouse
(Ohuchi, et al., 2000). If the foregut were unable to transform into thyroid tissue, it is
possible that foregut compartmentalization would be rescued.
My work has demonstrated that Shh plays varied and critical roles in the
morphogenesis and maturation of the foregut. Not only is Shh responsible for
partitioning the foregut endoderm into trachea and esophagus, but also for the
development of differentiated foregut mesoderm. Of interest, the cellular roles of Shh
appear to be mechanistically different between the trachea and esophagus. In
development of the tracheal cartilages, Shh signaling is paracrine, direct from the
endoderm to the mesoderm. However, with regards to the esophageal mesoderm,
patterning is regulated by autocrine Shh signaling within the esophageal endoderm.
This indicates the presence of a secondary secreted factor, possibly BMP4, which is
responsible for endoderm-to-mesoderm signaling.
152
7.2 Shh in laryngeal compartmentalization
My work demonstrated that Shh expression from the vFGE is required for
laryngeal compartmentalization. This source of Shh signals directly to NCCs to promote
separation of the larynx from the rostral esophagus, as the Wnt1-Cre; Smoflox/null embryo
has similar laryngeal clefts. This result is intriguing, as the mechanism by which NCCs
are required to compartmentalize the larynx is not clear. The simplest model is that
NCCs are required to form the cartilaginous scaffold of the larynx, namely the thyroid
cartilage. If the presence of the thyroid cartilage were required for laryngeal separation,
its absence would result in Type I LTE clefts. The thyroid cartilage is absent in both
endodermal Shh mutants (Nkx2.5-Cre; Shhflox/null) and NCC Shh reception mutants
(Wnt1-Cre; Smoflox/null), supporting this hypothesis. However, it is also possible that
NCCs serve not as a scaffold, but as a signaling center, and produce a secondary Shhdependent signal required for laryngeal compartmentalization.
To resolve this issue, one would have to separate the laryngeal cartilage and cleft
phenotypes. It may be possible to perform this with in vitro explant experiments, by
reintroducing Shh or inhibiting Smo to the compartmentalizing larynx after cartilage
formation has initiated. It may also be possible to perform these experiments
genetically, especially if the timing of Shh requirements for compartmentalization and
cartilage development are temporally distinct. In this case, utilization of a CreER or
153
doxycycline-inducible system may be able to impact compartmentalization without
affecting laryngeal cartilage development.
7.3 Origin of the thyroid cartilage
My demonstration that the thyroid cartilage, mostly derived from NCCs, also
has a small mesodermal contribution is intriguing. As more basal animals often have a
single thyrocricoid cartilage, it appears as if the cricoid and thyroid were originally a
single cartilage which split later in development. In fact, only mammals have a mobile
crico-thyroid junction, providing an increased range of vocalization. It would be
interesting to examine the single cricothyroid cartilage in non-mammalian models and
determine whether it is predominantly of NCC origin or whether it is half mesodermal.
7. 4 Noggin and homeotic transformations in laryngeal
cartilages
The aberrant fusions between the laryngeal cartilages observed in the various
Nog/BMP manipulations resemble homeotic transformations. Homeobox genes play a
role in establishing of the segmental body plan; as such, mutations in Hox genes often
cause fusions between laryngeal skeletal elements. For example, ablation of Hoxa3
causes a shortened, thickened larynx in which the greater horn of the hyoid is often
fused with the superior horn of the thyroid cartilage (Chisaka and Capecchi, 1991).
Simultaneous ablation of paralogs (e.g. Hoxa3/Hoxb3) exacerbates this laryngeal
phenotype to include fusions between the cricoid and thyroid cartilages (Condie and
154
Capecchi, 1994) (Manley and Capecchi, 1997). Mutations in Hoxa5 cause fusions
between the cricoid cartilage and first tracheal ring (Aubin et al., 1997) (Tabarias et al.,
2007). It will be of interest to determine whether Nog interacts with homeobox
patterning of the larynx.
7.5 Models of tracheomalacia
Tracheomalacia, or airway flaccidity, is caused by mispatterning or hypoplasia of
the tracheal cartilage rings that provide rigid support and maintain airway patency. In
Chapter 6, I create two new models of mild (Mef2C-Cre; Smoflox/null) and severe (Nkx2.5Cre; Shhflox/null) tracheomalacia. In both of these models, tracheal cartilage development
is severely impaired. In the severe model, tracheal cartilage is nearly absent, present in
only a few midline blebs. In the mild model, the tracheal cartilage is slightly hypoplastic
and not patterned correctly into ventral rings, possibly representing a more biological
model of tracheomalacia. In the future, these models may lead to a better understanding
of the causes and consequences of tracheomalacia.
7.6 Secondary phenotypes of βarr DKO embryos
Other than cleft palate, a variety of internal phenotypes were observed in βarr
DKO mice: none immediately postnatal lethal, but nearly all consistent with a decreased
proliferative response to Shh signaling. In particular, the underfoliated cerebellum of
βarr DKO mice provides a direct and tractable link to mitogenic hedgehog signaling:
Shh produced by Purkinje cells is the principal mitogen for granule cell precursors
155
(GCPs), whose proliferation drives cerebellar foliation (Corrales, et al., 2006). Both
targeted and unbiased qPCR analysis revealed downregulation of Shh-target genes,
including Hhip, Ptch, and Gli1, indicating Shh signaling is reduced in the DKO
cerebellum. However, the mechanism by which loss of βarrs reduces Shh-mediated
proliferation is unclear.
βarrestins may promote Shh signaling by facilitating Smo trafficking at the
primary cilium. Smo functions at primary cilium (Corbit, et al., 2005), where it is
regulated by Ptch1 (Rohatgi, et al., 2007) in mammalian cell culture. βarrestins have
been demonstrated to regulate the ciliary localization of Smo (Kovacs, et al., 2008). In
this model, βarrestins promote Shh signaling by facilitating the ciliary shuttling of Smo.
In the DKO embryo, the absence of βarrestins may be most acutely felt by rapidly
proliferating tissues, which require a fast mitogenic response.
However, we have no evidence that βarrs act directly on Smo in the mouse
embryo. It remains possible that βarrs promote proliferation by other means. It is
tempting to speculate that βarrs act directly on the Hh pathway via regulation of
Smoothened localization. However, downregulation of p53 inhibitors, including Mdm2,
in the βarr DKO cerebellum opens the possibility of additional regulatory roles for βarrs
in mitogenic Hh signaling. Shh stimulation of GCPs promotes inhibitory binding of
Mdm2 to p53. In mice with reduced levels of Mdm2, Shh signaling is downregulated,
GCP proliferation is decreased, and cerebellar foliation is diminished (Malek, et al.,
156
2011). Constitutive activation of Smo or overexpression of Gli inhibits accumulation of
p53 by stimulating binding to Mmd2 (Abe, et al., 2008). p53 has been demonstrated to
inhibit Gli1-mediated proliferation and in turn, Gli1 represses p53, establishing an
inhibitory loop (Stecca and Ruiz I Altaba, 2009). As βarrestins have been implicated in
both activation of Mdm2 and scaffolding of the Mdm2-p53 inhibitory interaction (Hara,
et al., 2011), it is possible that βarrestin-mediated attenuation of Shh signaling occurs via
modulation of p53 stability, in addition to a role in Smoothened trafficking (Figure 25E).
Interestingly, a recent study has found that Shh expression increases βarr1
localization to the nucleus of cultured GPCs and that knockdown of βarr1 increases GCP
proliferation via suppression of p27 (Parathath, et al., 2010). While we also see that
ablation of βarrs leads to downregulation of p27 in both the cerebellum and palate, we
do not see evidence of increased proliferation in any tissue examined. These differences
may reflect the difference between an acute knockdown and prolonged absence of
βarrestins, or may be due to complexities of βarr1/βarr2 redundancies during embryonic
development.
Anomalies of the aortic arch are present in approximately 50% of βarr DKO
embryos and a small percentage of littermates. These defects can be caused by a
number of genetic insults, primarily to the cardiac neural crest or second heart field.
While we did not examine gene expression levels in either of these tissues, both require
Shh signaling for proper formation of the cardiac outflow tract (Goddeeris, et al., 2007).
157
Ablation of Smoothened from the second heart field causes aortic arch defects, including
right-sided aortic arches (Lin, et al., 2006). Furthermore, Shh signaling has been shown
to have a mitogenic function in the second heart field (Dyer and Kirby, 2009). Thus, the
cardiac defects observed in βarr DKO mice may also be due to reduced proliferation in
the second heart field.
The large fluid-filled kidneys observed in DKO pups require further
investigation. Downstream of Shh and Bmp4, Tea3 promotes uteral smooth muscle
proliferation. When this mitogenic signaling pathway is disrupted, back-pressure from
hydroureter causes hydronephrosis (Yu, et al., 2002), histologically similar to the kidney
defects observed it the βarr DKO. Immediately adjacent to the murine kidney, L4
vertebral defects were observed in all DKO pups, sometimes extending from L1-L4. The
etiology of this isolated vertebral defect is unknown, but may be secondary to the renal
defects or related to Shh-mediated somitic development. This unusual defect of an
isolated split L4 also observed in Pax1-null (Wilm, et al., 1998).
7.7 Links between Hh signaling and βarrestins in the embryo
Interestingly, no defects were observed in other molecular functions of Shh
signaling. βarrestins are clearly not required for all aspects of Hedgehog (such as a Smonull, Zhang, et al., 2001), or Sonic Hh (Shh-null, Chiang, et al., 1996) signaling. The βarr
DKO does not have all classical phenotypes of a Shh-hypomorph, such as those
observed in IFT mutants. Surprisingly, only defects in the mitogenic activity of Shh are
158
observed: in the palate, cerebellum, cardiac outflow tract, and kidney. But not all
mitogenic contexts of Shh signaling appear to be affected, as proliferation in the
developing adrenal gland (Huang, et al., 2010) and gut mesenchyme (Mao, et al., 2010)
appear unaffected.
Most notably, no phenotypes in the DKO were consistent with misregulation of
other, non-mitogenic Shh activity. Shh is a classical morphogen for many tissues,
however no defects were observed consistent with misregulated patterning. For
instance, no heterotaxy, holoprosencephaly, micrognathia, or other midline defects were
observed (Chiang, et al., 1996). Patterning of limbs/digits (Harfe, et al., 2004), hair
follicles (Chiang, et al., 1999) bilateral kidney positioning (Tripathi, et al., 2010) were all
normal, as were induction of the pancreas (Hebrok, et al., 1998) and parathyroid (MooreScott and Manley, 2004). Locomotion and reaction to stimuli were initially normal in
βarr DKO pups, indicating no gross defect in neural/spinal neurogenesis, patterning, or
survival (Cayuso, et al., 2005) (Cai, et al., 2008). Altogether, the most critical function of
βarrestins appears to be promotion of mitogenic Shh signaling in rapidly proliferating
embryonic tissues.
Perhaps it is not surprising that βarrestins are involved in mitogenic Shh
signaling in the mouse embryo. Hh's role as a mitogen (not just morphogen) is
evolutionarily conserved: in the Drosophila eye, Hh regulates cellular proliferation, via
cyclins D and E, (Duman-Scheel, et al., 2002). βarr2 morphant Zebrafish also display
159
reduced cellular proliferation downstream of Shh, resulting in underdeveloped axial
musculature, and a loss of midline structures causing in partial cyclopia and severely
hypoplastic palatal structures (Wilbanks, et al., 2004).
The molecular function of βarrs in Smo signaling is still unknown. In Zebrafish,
βarr2 acts downstream of the Shh ligand, but upstream of immediate transcriptional
targets. (Wilbanks, et al., 2004). Indeed, it is likely that βarr2 promotes Hh signaling at
the level of Smo, because GRK2, which promotes Shh signaling in Zebrafish (Philipp, et
al., 2008), also enhances the interaction between activated Smo and βarr2 in mammalian
cell culture (Meloni, et al., 2006). Furthermore, this interaction may be required for
proliferation, as GRK2 interacts directly with Patched to promote cyclinB1-mediated
proliferation, in a kinase-independent manner (Jiang, et al., 2009).
It also remains possible that the βarr phenotypes are the consequence of gross
misregulation of fundamental cellular processes, such as suppression of apoptosis (via
BAD phosphorylation, Ahn, et al., 2009), maintenance of genomic instability (via β2AR,
Hara, et al., 2011), or global regulation of phosphorylation (Xiao, et al., 2010). However,
these models require the observed transcriptional downregulation of Shh targets to be a
coincidental, secondary consequence.
Altogether, my results support a model in which the critical role of βarrestins
during embryogenesis is in promotion of mitogenic Shh signaling. Whether βarrs act
directly on Smoothened in the embryo remains to be determined.
160
Interestingly, the βarr DKO does not display foregut defects. While I have no
evidence that foregut compartmentalization requires mitogenic Shh signaling, the
development of differentiated foregut mesoderm is very likely to have a proliferative
component. However, the tracheal cartilages and esophageal mesoderm are normal in
the βarr DKO embryo. As discussed above, βarrs may not be required in all mitogenic
contexts of Shh signaling.
One intriguing possibility is that βarrs are critical for only the most rapid Shhinduced proliferation. Perhaps ciliary translocation of Smo is only required in the fastest
proliferating cells, or perhaps Smo can translocate to the cilium in the absence of βarrs,
but at a slower pace. In these models, βarrs serve as a kinetic enhancer of Shh signaling.
161
References
Abe Y, Oda-Sato E, Tobiume K, Kawauchi K, Taya Y, Okamoto K, Oren M, Tanaka N.
Hedgehog signaling overrides p53-mediated tumor suppression by activating
Mdm2. Proc Natl Acad Sci U S A. 2008 Mar 25;105(12):4838-43.
Ahn S, Kim J, Hara MR, Ren XR, Lefkowitz RJ. {beta}-Arrestin-2 Mediates Antiapoptotic Signaling through Regulation of BAD Phosphorylation. J Biol Chem.
2009 Mar 27;284(13):8855-65.
Aubin J, Lemieux M, Tremblay M, Berard J, Jeannotte L. Early postnatal lethality in
Hoxa-5 mutant mice is attributable to respiratory tract defects. Dev Biol. 1997
Dec 15;192(2):432-45.
Berry RJ, Buehler JW, Strauss LT, et al. 1987. Birth weight-specific infant mortality due to
congenital abnormalities, 1960 and 1980. Public Health Report 102:171-81
Bertrand P, Navarro H, Caussade S, Holmgren N, Sanchez I. Airway anomalies in
children with Down syndrome: endoscopic findings. Pediatr Pulmonol. 2003
Aug;36(2):137-41.
Bose J, Grotewold L, Ruther U. Pallister-Hall syndrome phenotype in mice mutant for
Gli3. Hum Mol Genet. 2002 May 1;11(9):1129-35.
Bi W, Huang W, Whitworth DJ, Deng JM, Zhang Z, Behringer RR, de Crombrugghe B.
Haploinsufficiency of Sox9 results in defective cartilage primordia and
premature skeletal mineralization. Proc Natl Acad Sci U S A. 2001 Jun
5;98(12):6698-703.
Bohn LM, Lefkowitz RJ, Gainetdinov RR, Peppel K, Caron MG, Lin FT. Enhanced
morphine analgesia in mice lacking beta-arrestin 2. Science. 1999 Dec
24;286(5449):2495-8.
Bryja V, Gradl D, Schambony A, Arenas E, Schulte G. Beta-arrestin is a necessary
component of Wnt/beta-catenin signaling in vitro and in vivo. Proc Natl Acad
Sci U S A. 2007 Apr 17;104(16):6690-5.
Bryja V, Schambony A, Cajanek L, Dominguez I, Arenas E, Schulte G. Beta-arrestin and
casein kinase 1/2 define distinct branches of non-canonical WNT signalling
pathways. EMBO Rep. 2008 Dec;9(12):1244-50.
162
Cai C, Thorne J, Grabel L. Hedgehog serves as a mitogen and survival factor during
embryonic stem cell neurogenesis. Stem Cells. 2008 May;26(5):1097-108.
Cayuso J, Marti E. Morphogens in motion: growth control of the neural tube. J
Neurobiol. 2005 Sep 15;64(4):376-87.
Chen W, Hu LA, Semenov MV, Yanagawa S, Kikuchi A, Lefkowitz RJ, Miller WE. betaArrestin1 modulates lymphoid enhancer factor transcriptional activity through
interaction with phosphorylated dishevelled proteins. Proc Natl Acad Sci U S A.
2001 Dec 18;98(26):14889-94.
Chen W, Kirkbride KC, How T, Nelson CD, Mo J, Frederick JP, Wang XF, Lefkowitz RJ,
Blobe GC. Beta-arrestin 2 mediates endocytosis of type III TGF-beta receptor and
down-regulation of its signaling. Science. 2003 Sep 5;301(5638):1394-7.
Chen W, Ren XR, Nelson CD, Barak LS, Chen JK, Beachy PA, de Sauvage F, Lefkowitz
RJ. Activity-dependent internalization of smoothened mediated by beta-arrestin
2 and GRK2. Science. 2004 Dec 24;306(5705):2257-60.
Chen W, ten Berge D, Brown J, Ahn S, Hu LA, Miller WE, Caron MG, Barak LS, Nusse R,
Lefkowitz RJ. Dishevelled 2 recruits beta-arrestin 2 to mediate Wnt5Astimulated endocytosis of Frizzled 4. Science. 2003 Sep 5;301(5638):1391-4.
Chen Y, Sasai N, Ma G, Yue T, Jia J, Briscoe J, Jiang J. Sonic Hedgehog dependent
phosphorylation by CK1α and GRK2 is required for ciliary accumulation and
activation of smoothened. PLoS Biol. 2011 Jun;9(6):e1001083.
Chiang C, Litingtung Y, Lee E, Young KE, Corden JL, Westphal H, Beachy PA. Cyclopia
and defective axial patterning in mice lacking Sonic hedgehog gene function.
Nature. 1996 Oct 3;383(6599):407-13.
Chiang C, Swan RZ, Grachtchouk M, Bolinger M, Litingtung Y, Robertson EK, Cooper
MK, Gaffield W, Westphal H, Beachy PA, Dlugosz AA. Essential role for Sonic
hedgehog during hair follicle morphogenesis. Dev Biol. 1999 Jan 1;205(1):1-9.
Chisaka O, Capecchi MR. Regionally restricted developmental defects resulting from
targeted disruption of the mouse homeobox gene hox-1.5. Nature. 1991 Apr
11;350(6318):473-9.
Church DM, Bengtsson U, Nielsen KV, Wasmuth JJ, Niebuhr E. Molecular definition of
deletions of different segments of distal 5p that result in distinct phenotypic
features. Am J Hum Genet. 1995 May;56(5):1162-72.
163
Cobourne MT, Xavier GM, Depew M, Hagan L, Sealby J, Webster Z, Sharpe PT. Sonic
hedgehog signalling inhibits palatogenesis and arrests tooth development in a
mouse model of the nevoid basal cell carcinoma syndrome. Dev Biol. 2009 Jul
1;331(1):38-49.
Cochard Netter’s Atlas of Human Embryology (2002)
Condie BG, Capecchi MR. Mice with targeted disruptions in the paralogous genes hoxa3 and hoxd-3 reveal synergistic interactions. Nature. 1994 Jul 28;370(6487):304-7.
Erratum in: Nature 1994 Oct 6;371(6497):537.
Conner DA, Mathier MA, Mortensen RM, Christe M, Vatner SF, Seidman CE, Seidman
JG. beta-Arrestin1 knockout mice appear normal but demonstrate altered
cardiac responses to beta-adrenergic stimulation. Circ Res. 1997 Dec;81(6):10216.
Cooley MA, Kern CB, Fresco VM, Wessels A, Thompson RP, McQuinn TC, Twal WO,
Mjaatvedt CH, Drake CJ, Argraves WS. Fibulin-1 is required for morphogenesis
of neural crest-derived structures. Dev Biol. 2008 Jul 15;319(2):336-45.
Corbit KC, Aanstad P, Singla V, Norman AR, Stainier DY, Reiter JF. Vertebrate
Smoothened functions at the primary cilium. Nature. 2005 Oct
13;437(7061):1018-21.
Corrales JD, Blaess S, Mahoney EM, Joyner AL. The level of sonic hedgehog signaling
regulates the complexity of cerebellar foliation. Development. 2006
May;133(9):1811-21.
DeWire SM, Kim J, Whalen EJ, Ahn S, Chen M, Lefkowitz RJ. Beta-arrestin-mediated
signaling regulates protein synthesis. J Biol Chem. 2008 Apr 18;283(16):10611-20.
Dravis C, Henkemeyer M. Ephrin-B reverse signaling controls septation events at the
embryonic midline through separate tyrosine phosphorylation-independent
signaling avenues. Dev Biol. 2011 Jul 1;355(1):138-51.
Duman-Scheel M, Weng L, Xin S, Du W. Hedgehog regulates cell growth and
proliferation by inducing Cyclin D and Cyclin E. Nature. 2002 May
16;417(6886):299-304.
Dyer LA, Kirby ML. Sonic hedgehog maintains proliferation in secondary heart field
progenitors and is required for normal arterial pole formation. Dev Biol. 2009
Jun 15;330(2):305-17.
164
Engert S, Liao WP, Burtscher I, Lickert H. Sox17-2A-iCre: a knock-in mouse line
expressing Cre recombinase in endoderm and vascular endothelial cells.
Genesis. 2009 Sep;47(9):603-10.
Ericson J, Morton S, Kawakami A, Roelink H, Jessell TM. Two critical periods of Sonic
Hedgehog signaling required for the specification of motor neuron identity. Cell.
1996 Nov 15;87(4):661-73.
Fagman H, Grande M, Gritli-Linde A, Nilsson M. Genetic deletion of sonic hedgehog
causes hemiagenesis and ectopic development of the thyroid in mouse. Am J
Pathol. 2004 May;164(5):1865-72.
Fausett, S. R. and Klingensmith, J. (2012), Compartmentalization of the foregut tube:
developmental origins of the trachea and esophagus. WIREs Dev Biol, 1: 184–202.
Fereshteh M, Ito T, Kovacs JJ, Zhao C, Kwon HY, Tornini V, Konuma T, Chen M,
Lefkowitz RJ, Reya T. B-Arrestin2 mediates the initiation and progression of
myeloid leukemia. Proc Natl Acad Sci U S A. 2012 Jul 31;109(31):12532-7.
Frazer JE. The Development of the Larynx. J Anat Physiol. 1910 Jan;44(Pt 2):156-91. No
abstract available.
Freestone SH, Marker P, Grace OC, Tomlinson DC, Cunha GR, Harnden P, Thomson
AA. Sonic hedgehog regulates prostatic growth and epithelial differentiation.
Dev Biol. 2003 Dec 15;264(2):352-62.
Fukuto HS, Ferkey DM, Apicella AJ, Lans H, Sharmeen T, Chen W, Lefkowitz RJ, Jansen
G, Schafer WR, Hart AC. G protein-coupled receptor kinase function is essential
for chemosensation in C. elegans. Neuron. 2004 May 27;42(4):581-93.
Ge H, Krishnan P, Liu L, Krishnan B, Davis RL, Hardin PE, Roman G. A Drosophila
nonvisual arrestin is required for the maintenance of olfactory sensitivity. Chem
Senses. 2006 Jan;31(1):49-62.
Geng Y, Dong Y, Yu M, Zhang L, Yan X, Sun J, Qiao L, Geng H, Nakajima M, Furuichi T,
Ikegawa S, Gao X, Chen YG, Jiang D, Ning W. Follistatin-like 1 (Fstl1) is a bone
morphogenetic protein (BMP) 4 signaling antagonist in controlling mouse lung
development. Proc Natl Acad Sci U S A. 2011 Apr 26;108(17):7058-63.
Ghyselinck NB, Dupe V, Dierich A, Messaddeq N, Garnier JM, Rochette-Egly C,
Chambon P, Mark M. Role of the retinoic acid receptor beta (RARbeta) during
mouse development. Int J Dev Biol. 1997 Jun;41(3):425-47.
165
Goddeeris MM, Schwartz R, Klingensmith J, Meyers EN. Independent requirements for
Hedgehog signaling by both the anterior heart field and neural crest cells for
outflow tract development. Development. 2007 Apr;134(8):1593-604.
Goddeeris MM, Rho S, Petiet A, Davenport CL, Johnson GA, Meyers EN, Klingensmith
J. Intracardiac septation requires hedgehog-dependent cellular contributions
from outside the heart. Development. 2008 May;135(10):1887-95.
Gritli-Linde A. Molecular control of secondary palate development. Dev Biol. 2007 Jan
15;301(2):309-26.
Hebert JM, McConnell SK. Targeting of cre to the Foxg1 (BF-1) locus mediates loxP
recombination in the telencephalon and other developing head structures. Dev
Biol. 2000 Jun 15;222(2):296-306.
Hara MR, Kovacs JJ, Whalen EJ, Rajagopal S, Strachan RT, Grant W, Towers AJ,
Williams B, Lam CM, Xiao K, Shenoy SK, Gregory SG, Ahn S, Duckett DR,
Lefkowitz RJ. A stress response pathway regulates DNA damage through B2adrenoreceptors and B-arrestin-1. Nature. 2011 Aug 21;477(7364):349-53.
Haraguchi R, Mo R, Hui C, Motoyama J, Makino S, Shiroishi T, Gaffield W, Yamada G.
Unique functions of Sonic hedgehog signaling during external genitalia
development. Development. 2001 Nov;128(21):4241-50.
Harfe BD, Scherz PJ, Nissim S, Tian H, McMahon AP, Tabin CJ. Evidence for an
expansion-based temporal Shh gradient in specifying vertebrate digit identities.
Cell. 2004 Aug 20;118(4):517-28.
Hebrok M, Kim SK, Melton DA. Notochord repression of endodermal Sonic hedgehog
permits pancreas development. Genes Dev. 1998 Jun 1;12(11):1705-13.
Henick DH. Three-dimensional analysis of murine laryngeal development. Ann Otol
Rhinol Laryngol Suppl. 1993 Mar;159:3-24.
Hirose Y, Itoh T, Miyajima A. Hedgehog signal activation coordinates proliferation and
differentiation of fetal liver progenitor cells. Exp Cell Res. 2009 Sep
10;315(15):2648-57.
Hoffmann AD, Peterson MA, Friedland-Little JM, Anderson SA, Moskowitz IP. sonic
hedgehog is required in pulmonary endoderm for atrial septation. Development.
2009 May;136(10):1761-70.
166
Holinger LD. Histopathology of congenital subglottic stenosis. Ann Otol Rhinol
Laryngol. 1999 Feb;108(2):101-11.
Hopp ES. The development of the epithelium of the larynx. Laryngoscope. 1955
Jul;65(7):475-99.
Hu MC, Mo R, Bhella S, Wilson CW, Chuang PT, Hui CC, Rosenblum ND. GLI3dependent transcriptional repression of Gli1, Gli2 and kidney patterning genes
disrupts renal morphogenesis. Development. 2006 Feb;133(3):569-78.
Huang CC, Miyagawa S, Matsumaru D, Parker KL, Yao HH. Progenitor cell expansion
and organ size of mouse adrenal is regulated by sonic hedgehog. Endocrinology.
2010 Mar;151(3):1119-28.
Ikeya M, Kawada M, Kiyonari H, Sasai N, Nakao K, Furuta Y, Sasai Y. Essential proBmp roles of crossveinless 2 in mouse organogenesis. Development. 2006
Nov;133(22):4463-73.
Ikeya M, Lee SM, Johnson JE, McMahon AP, Takada S. Wnt signalling required for
expansion of neural crest and CNS progenitors. Nature. 1997 Oct
30;389(6654):966-70.
Jackson Labs Informatics, www.informatics.jax.org. Accessed Feb 7th, 2013
Jaskoll T, Leo T, Witcher D, Ormestad M, Astorga J, Bringas P Jr, Carlsson P, Melnick M.
Sonic hedgehog signaling plays an essential role during embryonic salivary
gland epithelial branching morphogenesis. Dev Dyn. 2004 Apr;229(4):722-32.
Jeong J, Mao J, Tenzen T, Kottmann AH, McMahon AP. Hedgehog signaling in the
neural crest cells regulates the patterning and growth of facial primordia. Genes
Dev. 2004 Apr 15;18(8):937-51.
Jiang X, Yang P, Ma L. Kinase activity-independent regulation of cyclin pathway by
GRK2 is essential for zebrafish early development. Proc Natl Acad Sci U S A.
2009 Jun 23;106(25):10183-8.
Johnson EC, Tift FW, McCauley A, Liu L, Roman G. Functional characterization of
kurtz, a Drosophila non-visual arrestin, reveals conservation of GPCR
desensitization mechanisms. Insect Biochem Mol Biol. 2008 Nov;38(11):1016-22.
Kakodkar KA, Schroeder JW Jr, Holinger LD. Laryngeal development and anatomy.
Adv Otorhinolaryngol. 2012;73:1-11.
167
Katori Y, Takeuchi H, Rodriguez-Vazquez JF, Kitano H, Murakami G, Kawase T. Fetal
development of the human epiglottis revisited: appearance of GFAP-positive
mesenchymal cells and fibrous connections with other laryngeal and lingual
structures. Ann Anat. 2011 Mar;193(2):149-55.
Kaufman and Bard Anatomical Basis of Mouse Development (1999)
Kelly RG, Jerome-Majewska LA, Papaioannou VE. The del22q11.2 candidate gene Tbx1
regulates branchiomeric myogenesis. Hum Mol Genet. 2004 Nov 15;13(22):282940.
Kepes JJ, Perentes E. Glial fibrillary acidic protein in chondrocytes of elastic cartilage in
the human epiglottis: an immunohistochemical study with polyvalent and
monoclonal antibodies. Anat Rec. 1988 Mar;220(3):296-9.
Kim GH, Han JK. Essential role for beta-arrestin 2 in the regulation of Xenopus
convergent extension movements. EMBO J. 2007 May 16;26(10):2513-26.
Kirchner JA. Joseph H. Ogura Memorial Lecture. The vertebrate larynx: adaptations and
aberrations. Laryngoscope. 1993 Oct;103(10):1197-201.
Kohout TA, Lin FS, Perry SJ, Conner DA, Lefkowitz RJ. beta-Arrestin 1 and 2
differentially regulate heptahelical receptor signaling and trafficking. Proc Natl
Acad Sci U S A. 2001 Feb 13;98(4):1601-6.
Kolpakova-Hart E, Jinnin M, Hou B, Fukai N, Olsen BR. Kinesin-2 controls
development and patterning of the vertebrate skeleton by Hedgehog- and Gli3dependent mechanisms. Dev Biol. 2007 Sep 15;309(2):273-84.
Kovacs JJ, Hara MR, Davenport CL, Kim J, Lefkowitz RJ. Arrestin development:
emerging roles for beta-arrestins in developmental signaling pathways. Dev
Cell. 2009 Oct;17(4):443-58.
Kovacs JJ, Whalen EJ, Liu R, Xiao K, Kim J, Chen M, Wang J, Chen W, Lefkowitz RJ.
Beta-arrestin-mediated localization of smoothened to the primary cilium.
Science. 2008 Jun 27;320(5884):1777-81.
Kumar A, Yamaguchi T, Sharma P, Kuehn MR. Transgenic mouse lines expressing Cre
recombinase specifically in posterior notochord and notochord. Genesis. 2007
Dec;45(12):729-36.
168
Lan Y, Jiang R. Sonic hedgehog signaling regulates reciprocal epithelial-mesenchymal
interactions controlling palatal outgrowth. Development. 2009 Apr;136(8):138796.
Lefkowitz RJ, Rajagopal K, Whalen EJ. New roles for beta-arrestins in cell signaling: not
just for seven-transmembrane receptors. Mol Cell. 2006 Dec 8;24(5):643-52.
Review.
Lefkowitz RJ, Shenoy SK. Transduction of receptor signals by beta-arrestins. Science.
2005 Apr 22;308(5721):512-7.
Li S, Chen Y, Shi Q, Yue T, Wang B, Jiang J. Hedgehog-regulated ubiquitination controls
smoothened trafficking and cell surface expression in Drosophila. PLoS Biol.
2012 Jan;10(1):e1001239.
Liem KF Jr, Jessell TM, Briscoe J. Regulation of the neural patterning activity of sonic
hedgehog by secreted BMP inhibitors expressed by notochord and somites.
Development. 2000 Nov;127(22):4855-66.
Lin L, Bu L, Cai CL, Zhang X, Evans S. Isl1 is upstream of sonic hedgehog in a pathway
required for cardiac morphogenesis. Dev Biol. 2006 Jul 15;295(2):756-63.
Litingtung Y, Lei L, Westphal H, Chiang C. Sonic hedgehog is essential to foregut
development. Nat Genet. 1998 Sep;20(1):58-61.
Liu L, Davis RL, Roman G. Exploratory activity in Drosophila requires the kurtz
nonvisual arrestin. Genetics. 2007 Mar;175(3):1197-212.
Luan B, Zhao J, Wu H, Duan B, Shu G, Wang X, Li D, Jia W, Kang J, Pei G. Deficiency of
a beta-arrestin-2 signal complex contributes to insulin resistance. Nature. 2009
Feb 26;457(7233):1146-9.
Madison BB, Braunstein K, Kuizon E, Portman K, Qiao XT, Gumucio DL. Epithelial
hedgehog signals pattern the intestinal crypt-villus axis. Development. 2005
Jan;132(2):279-89.
Malek R, Matta J, Taylor N, Perry ME, Mendrysa SM. The p53 inhibitor MDM2
facilitates Sonic Hedgehog-mediated tumorigenesis and influences cerebellar
foliation. PLoS One. 2011 Mar 18;6(3):e17884.
169
Manley NR, Capecchi MR. Hox group 3 paralogous genes act synergistically in the
formation of somitic and neural crest-derived structures. Dev Biol. 1997 Dec
15;192(2):274-88.
Mao J, Kim BM, Rajurkar M, Shivdasani RA, McMahon AP. Hedgehog signaling
controls mesenchymal growth in the developing mammalian digestive tract.
Development. 2010 May;137(10):1721-9.
March of Dimes, www.marchofdimes.com. Accessed Feb 7th, 2013.
Matsuoka T, Ahlberg PE, Kessaris N, Iannarelli P, Dennehy U, Richardson WD,
McMahon AP, Koentges G. Neural crest origins of the neck and shoulder.
Nature. 2005 Jul 21;436(7049):347-55.
McElhinney DB, Jacobs I, McDonald-McGinn DM, Zackai EH, Goldmuntz E.
Chromosomal and cardiovascular anomalies associated with congenital
laryngeal web. Int J Pediatr Otorhinolaryngol. 2002 Oct 21;66(1):23-27.
McINTOSH R, MERRITT KK, RICHARDS MR, SAMUELS MH, BELLOWS MT. The
incidence of congenital malformations: a study of 5,964 pregnancies. Pediatrics.
1954 Nov;14(5):505-22.
Melnick M, Witcher D, Bringas P Jr, Carlsson P, Jaskoll T. Meckel's cartilage
differentiation is dependent on hedgehog signaling. Cells Tissues Organs.
2005;179(4):146-57.
Meloni AR, Fralish GB, Kelly P, Salahpour A, Chen JK, Wechsler-Reya RJ, Lefkowitz RJ,
Caron MG. Smoothened signal transduction is promoted by G protein-coupled
receptor kinase 2. Mol Cell Biol. 2006 Oct;26(20):7550-60.
Mendelsohn C, Lohnes D, Decimo D, Lufkin T, LeMeur M, Chambon P, Mark M.
Function of the retinoic acid receptors (RARs) during development (II). Multiple
abnormalities at various stages of organogenesis in RAR double mutants.
Development. 1994 Oct;120(10):2749-71.
Molnar C, Ruiz-Gamez A, Martin M, Rojo-Berciano S, Mayor F, de Celis JF. Role of the
Drosophila non-visual B-arrestin kurtz in hedgehog signalling. PLoS Genet. 2011
Mar;7(3):e1001335.
Moore-Scott BA, Manley NR. Differential expression of Sonic hedgehog along the
anterior-posterior axis regulates patterning of pharyngeal pouch endoderm and
pharyngeal endoderm-derived organs. Dev Biol. 2005 Feb 15;278(2):323-35.
170
Mori-Akiyama Y, Akiyama H, Rowitch DH, de Crombrugghe B. Sox9 is required for
determination of the chondrogenic cell lineage in the cranial neural crest. Proc
Natl Acad Sci U S A. 2003 Aug 5;100(16):9360-5.
Moses KA, DeMayo F, Braun RM, Reecy JL, Schwartz RJ. Embryonic expression of an
Nkx2-5/Cre gene using ROSA26 reporter mice. Genesis. 2001 Dec;31(4):176-80.
Mukherjee A, Veraksa A, Bauer A, Rosse C, Camonis J, Artavanis-Tsakonas S.
Regulation of Notch signalling by non-visual beta-arrestin. Nat Cell Biol. 2005
Dec;7(12):1191-201.
Nagasaka A, Kawane K, Yoshida H, Nagata S. Apaf-1-independent programmed cell
death in mouse development. Cell Death Differ. 2010 Jun;17(6):931-41.
Nathan E, Monovich A, Tirosh-Finkel L, Harrelson Z, Rousso T, Rinon A, Harel I, Evans
SM, Tzahor E. The contribution of Islet1-expressing splanchnic mesoderm cells
to distinct branchiomeric muscles reveals significant heterogeneity in head
muscle development. Development. 2008 Feb;135(4):647-57.
Neilan RE, Shao D, Dravis C, Henkemeyer M, Lee KH. Characterization of the larynx in
ephrin-B2 knockout mice: a novel animal model for laryngeal clefts. Arch
Otolaryngol Head Neck Surg. 2012 Oct;138(10):969-72.
Ngan ES, Garcia-Barceló MM, Yip BH, Poon HC, Lau ST, Kwok CK, Sat E, Sham MH,
Wong KK, Wainwright BJ, Cherny SS, Hui CC, Sham PC, Lui VC, Tam PK.
Hedgehog/Notch-induced premature gliogenesis represents a new disease
mechanism for Hirschsprung disease in mice and humans. J Clin Invest. 2011
Sep;121(9):3467-78.
Ohuchi H, Hori Y, Yamasaki M, Harada H, Sekine K, Kato S, Itoh N. FGF10 acts as a
major ligand for FGF receptor 2 IIIb in mouse multi-organ development.
Biochem Biophys Res Commun. 2000 Nov 2;277(3):643-9.
Ondrey F, Griffith A, Van Waes C, Rudy S, Peters K, McCullagh L, Biesecker LG.
Asymptomatic laryngeal malformations are common in patients with PallisterHall syndrome. Am J Med Genet. 2000 Sep 4;94(1):64-7.
Outram SV, Varas A, Pepicelli CV, Crompton T. Hedgehog signaling regulates
differentiation from double-negative to double-positive thymocyte. Immunity.
2000 Aug;13(2):187-97. Erratum in: Immunity 2000 Oct;13(4):585.
171
Palmitessa A, Benovic JL. Arrestin and the multi-PDZ domain-containing protein MPZ1 interact with phosphatase and tensin homolog (PTEN) and regulate
Caenorhabditis elegans longevity. J Biol Chem. 2010 May 14;285(20):15187-200.
Palmitessa A, Hess HA, Bany IA, Kim YM, Koelle MR, Benovic JL. Caenorhabditus
elegans arrestin regulates neural G protein signaling and olfactory adaptation
and recovery. J Biol Chem. 2005 Jul 1;280(26):24649-62.
Parathath SR, Mainwaring LA, Fernandez-L A, Guldal CG, Nahle Z, Kenney AM. BArrestin-1 links mitogenic sonic hedgehog signaling to the cell cycle exit
machinery in neural precursors. Cell Cycle. 2010 Oct 1;9(19):4013-24.
Park J, Zhang JJ, Moro A, Kushida M, Wegner M, Kim PC. Regulation of Sox9 by Sonic
Hedgehog (Shh) is essential for patterning and formation of tracheal cartilage.
Dev Dyn. 2010 Feb;239(2):514-26.
Parrish M, Ott T, Lance-Jones C, Schuetz G, Schwaeger-Nickolenko A, Monaghan AP.
Loss of the Sall3 gene leads to palate deficiency, abnormalities in cranial nerves,
and perinatal lethality. Mol Cell Biol. 2004 Aug;24(16):7102-12.
Pepicelli CV, Lewis PM, McMahon AP. Sonic hedgehog regulates branching
morphogenesis in the mammalian lung. Curr Biol. 1998 Sep 24;8(19):1083-6.
Review.
Perriton CL, Powles N, Chiang C, Maconochie MK, Cohn MJ. Sonic hedgehog signaling
from the urethral epithelium controls external genital development. Dev Biol.
2002 Jul 1;247(1):26-46.
Pezzettigotta SM, Leboulanger N, Roger G, Denoyelle F, Garabedian EN. Laryngeal
cleft. Otolaryngol Clin North Am. 2008 Oct;41(5):913-33, ix.
Philipp M, Fralish GB, Meloni AR, Chen W, MacInnes AW, Barak LS, Caron MG.
Smoothened signaling in vertebrates is facilitated by a G protein-coupled
receptor kinase. Mol Biol Cell. 2008 Dec;19(12):5478-89.
Premont RT, Gainetdinov RR. Physiological roles of G protein-coupled receptor kinases
and arrestins. Annu Rev Physiol. 2007;69:511-34.
Raghuwanshi SK, Nasser MW, Chen X, Strieter RM, Richardson RM. Depletion of betaarrestin-2 promotes tumor growth and angiogenesis in a murine model of lung
cancer. J Immunol. 2008 Apr 15;180(8):5699-706.
172
Rajagopal K, Whalen EJ, Violin JD, Stiber JA, Rosenberg PB, Premont RT, Coffman TM,
Rockman HA, Lefkowitz RJ. Beta-arrestin2-mediated inotropic effects of the
angiotensin II type 1A receptor in isolated cardiac myocytes. Proc Natl Acad Sci
U S A. 2006 Oct 31;103(44):16284-9.
Ramalho-Santos M, Melton DA, McMahon AP. Hedgehog signals regulate multiple
aspects of gastrointestinal development. Development. 2000 Jun;127(12):2763-72.
Robinson A. and Linden MG. 1993. Clinical Genetic Handbook, Boston, Blackwell
Scientific Publications
Rodriguez P, Da Silva S, Oxburgh L, Wang F, Hogan BL, Que J. BMP signaling in the
development of the mouse esophagus and forestomach. Development. 2010
Dec;137(24):4171-6.
Rohatgi R, Milenkovic L, Scott MP. Patched1 regulates hedgehog signaling at the
primary cilium. Science. 2007 Jul 20;317(5836):372-6.
Roman G, He J, Davis RL. kurtz, a novel nonvisual arrestin, is an essential neural gene
in Drosophila. Genetics. 2000 Jul;155(3):1281-95.
Sanudo JR, Domenech-Mateu JM. The laryngeal primordium and epithelial lamina. A
new interpretation. J Anat. 1990 Aug;171:207-22.
Sagai T, Amano T, Tamura M, Mizushina Y, Sumiyama K, Shiroishi T. A cluster of three
long-range enhancers directs regional Shh expression in the epithelial linings.
Development. 2009 May;136(10):1665-74.
Seifert AW, Bouldin CM, Choi KS, Harfe BD, Cohn MJ. Multiphasic and tissue-specific
roles of sonic hedgehog in cloacal septation and external genitalia development.
Development. 2009 Dec;136(23):3949-57.
Shiroyanagi Y, Liu B, Cao M, Agras K, Li J, Hsieh MH, Willingham EJ, Baskin LS.
Urothelial sonic hedgehog signaling plays an important role in bladder smooth
muscle formation. Differentiation. 2007 Dec;75(10):968-77.
Shukla AK, Xiao K, Lefkowitz RJ. Emerging paradigms of B-arrestin-dependent seven
transmembrane receptor signaling. Trends Biochem Sci. 2011 Sep;36(9):457-69.
Singh V, Aballay A. Endoplasmic reticulum stress pathway required for immune
homeostasis is neurally controlled by arrestin-1. J Biol Chem. 2012 Sep
28;287(40):33191-7.
173
Stafford DA, Brunet LJ, Khokha MK, Economides AN, Harland RM. Cooperative
activity of noggin and gremlin 1 in axial skeleton development. Development.
2011 Mar;138(5):1005-14.
Stecca B, Ruiz i Altaba A. A GLI1-p53 inhibitory loop controls neural stem cell and
tumour cell numbers. EMBO J. 2009 Mar 18;28(6):663-76.
Subbarayan V, Kastner P, Mark M, Dierich A, Gorry P, Chambon P. Limited specificity
and large overlap of the functions of the mouse RAR gamma 1 and RAR gamma
2 isoforms. Mech Dev. 1997 Aug;66(1-2):131-42.
Sukegawa A, Narita T, Kameda T, Saitoh K, Nohno T, Iba H, Yasugi S, Fukuda K. The
concentric structure of the developing gut is regulated by Sonic hedgehog
derived from endodermal epithelium. Development. 2000 May;127(9):1971-80.
Tabarias S, Lemieux M, Aubin J, Jeannotte L. Comparative analysis of Hoxa5 allelic
series. Genesis. 2007 Apr;45(4):218-28.
Testaz S, Jarov A, Williams KP, Ling LE, Koteliansky VE, Fournier-Thibault C, Duband
JL. Sonic hedgehog restricts adhesion and migration of neural crest cells
independently of the Patched- Smoothened-Gli signaling pathway. Proc Natl
Acad Sci U S A. 2001 Oct 23;98(22):12521-6.
Thompson DM. Abnormal sensorimotor integrative function of the larynx in congenital
laryngomalacia: a new theory of etiology. Laryngoscope. 2007 Jun;117(6 Pt 2
Suppl 114):1-33.
Tipping M, Kim Y, Kyriakakis P, Tong M, Shvartsman SY, Veraksa A. B-arrestin Kurtz
inhibits MAPK and Toll signalling in Drosophila development. EMBO J. 2010
Oct 6;29(19):3222-35.
Tripathi P, Guo Q, Wang Y, Coussens M, Liapis H, Jain S, Kuehn MR, Capecchi MR,
Chen F. Midline signaling regulates kidney positioning but not nephrogenesis
through Shh. Dev Biol. 2010 Apr 15;340(2):518-27.
Tseng BS, Cavin ST, Booth FW, Olson EN, Marin MC, McDonnell TJ, Butler IJ.
Pulmonary hypoplasia in the myogenin null mouse embryo. Am J Respir Cell
Mol Biol. 2000 Mar;22(3):304-15.
Tsukui T, Capdevila J, Tamura K, Ruiz-Lozano P, Rodriguez-Esteban C, Yonei-Tamura
S, Magallan J, Chandraratna RA, Chien K, Blumberg B, Evans RM, Belmonte JC.
Multiple left-right asymmetry defects in Shh(-/-) mutant mice unveil a
174
convergence of the shh and retinoic acid pathways in the control of Lefty-1. Proc
Natl Acad Sci U S A. 1999 Sep 28;96(20):11376-81.
Tucker GF Jr, Maddalozzo J. Occult" posterior laryngeal cleft." Laryngoscope. 1987
Jun;97(6):701-4.
van Haelst MM, Scambler PJ; Fraser Syndrome Collaboration Group, Hennekam RC.
Fraser syndrome: a clinical study of 59 cases and evaluation of diagnostic criteria.
Am J Med Genet A. 2007 Dec 15;143A(24):3194-203.
Vermot J, Niederreither K, Garnier JM, Chambon P, Dolle P. Decreased embryonic
retinoic acid synthesis results in a DiGeorge syndrome phenotype in newborn
mice. Proc Natl Acad Sci U S A. 2003 Feb 18;100(4):1763-8.
Verzi MP, McCulley DJ, De Val S, Dodou E, Black BL. The right ventricle, outflow tract,
and ventricular septum comprise a restricted expression domain within the
secondary/anterior heart field. Dev Biol. 2005 Nov 1;287(1):134-45.
Viejo F, Pascual-Font A, Rivas L, Hernandez-Morato I, Vazquez T, RodriguezNiedenfuhr M, McHanwell S, Sanudo JR. Morphogenesis of the human
laryngeal ventricles. Head Neck. 2013 Mar;35(3):361-9.
WALANDER A. The mechanism of origin of congenital malformations of the larynx.
Acta Otolaryngol. 1955 Sep-Oct;45(5):426-32.
Walker JK, Fong AM, Lawson BL, Savov JD, Patel DD, Schwartz DA, Lefkowitz RJ.
Beta-arrestin-2 regulates the development of allergic asthma. J Clin Invest. 2003
Aug;112(4):566-74.
Washington Smoak I, Byrd NA, Abu-Issa R, Goddeeris MM, Anderson R, Morris J,
Yamamura K, Klingensmith J, Meyers EN. Sonic hedgehog is required for
cardiac outflow tract and neural crest cell development. Dev Biol. 2005 Jul
15;283(2):357-72.
Wechsler-Reya RJ, Scott MP. Control of neuronal precursor proliferation in the
cerebellum by Sonic Hedgehog. Neuron. 1999 Jan;22(1):103-14.
Wilbanks AM, Fralish GB, Kirby ML, Barak LS, Li YX, Caron MG. Beta-arrestin 2
regulates zebrafish development through the hedgehog signaling pathway.
Science. 2004 Dec 24;306(5705):2264-7.
175
Wilm B, Dahl E, Peters H, Balling R, Imai K. Targeted disruption of Pax1 defines its null
phenotype and proves haploinsufficiency. Proc Natl Acad Sci U S A. 1998 Jul
21;95(15):8692-7.
Xiao K, Sun J, Kim J, Rajagopal S, Zhai B, Villen J, Haas W, Kovacs JJ, Shukla AK, Hara
MR, Hernandez M, Lachmann A, Zhao S, Lin Y, Cheng Y, Mizuno K, Ma'ayan A,
Gygi SP, Lefkowitz RJ. Global phosphorylation analysis of beta-arrestinmediated signaling downstream of a seven transmembrane receptor (7TMR).
Proc Natl Acad Sci U S A. 2010 Aug 24;107(34):15299-304.
Yamada W, Nagao K, Horikoshi K, Fujikura A, Ikeda E, Inagaki Y, Kakitani M,
Tomizuka K, Miyazaki H, Suda T, Takubo K. Craniofacial malformation in Rspondin2 knockout mice. Biochem Biophys Res Commun. 2009 Apr
10;381(3):453-8.
Yu J, Carroll TJ, McMahon AP. Sonic hedgehog regulates proliferation and
differentiation of mesenchymal cells in the mouse metanephric kidney.
Development. 2002 Nov;129(22):5301-12.
Zaw-Tun HA, Burdi AR. Reexamination of the origin and early development of the
human larynx. Acta Anat (Basel). 1985;122(3):163-84.
Zaw-Tun HI. Development of congenital laryngeal atresias and clefts. Ann Otol Rhinol
Laryngol. 1988 Jul-Aug;97(4 Pt 1):353-8.
Zhang M, Liu X, Zhang Y, Zhao J. Loss of betaarrestin1 and betaarrestin2 contributes to
pulmonary hypoplasia and neonatal lethality in mice. Dev Biol. 2010 Mar
15;339(2):407-17.
Zhang M, Teng H, Shi J, Zhang Y. Disruption of B-arrestins blocks glucocorticoid
receptor and severely retards lung and liver development in mice. Mech Dev.
2011 Sep-Dec;128(7-10):368-75.
Zhang XM, Ramalho-Santos M, McMahon AP. Smoothened mutants reveal redundant
roles for Shh and Ihh signaling including regulation of L/R symmetry by the
mouse node. Cell. 2001 Jul 27;106(2):781-92.
Zhang Z, Song Y, Zhao X, Zhang X, Fermin C, Chen Y. Rescue of cleft palate in Msx1deficient mice by transgenic Bmp4 reveals a network of BMP and Shh signaling
in the regulation of mammalian palatogenesis. Development. 2002
Sep;129(17):4135-46.
176
Zhu J, Nakamura E, Nguyen MT, Bao X, Akiyama H, Mackem S. Uncoupling Sonic
hedgehog control of pattern and expansion of the developing limb bud. Dev
Cell. 2008 Apr;14(4):624-32.
Zhuang LN, Hu WX, Zhang ML, Xin SM, Jia WP, Zhao J, Pei G. Beta-arrestin-1 protein
represses diet-induced obesity. J Biol Chem. 2011 Aug 12;286(32):28396-402.
Zoudilova M, Min J, Richards HL, Carter D, Huang T, DeFea KA. beta-Arrestins
scaffold cofilin with chronophin to direct localized actin filament severing and
membrane protrusions downstream of protease-activated receptor-2. J Biol
Chem. 2010 May 7;285(19):14318-29.
Zouvelou V, Luder HU, Mitsiadis TA, Graf D. Deletion of BMP7 affects the
development of bones, teeth, and other ectodermal appendages of the orofacial
complex. J Exp Zool B Mol Dev Evol. 2009 Jun 15;312B(4):361-74.
177
Biography
Chandra Lynn Clarke was born on December 14th, 1983 in suburban Detroit,
Michigan. She attended Cranbrook Kingswood in Bloomfield Hills from kindergarten
through high school, where she met her future husband, Sam Davenport. In 2006, she
was awarded an S.B. in Biology from the Massachusetts Institute of Technology.
Chandra and Sam moved down to North Carolina in 2006 so that Chandra could study
at Duke, and were married the following year.
Chandra joined John Klingensmith's lab in 2007 and was awarded a Graduate
Research Fellowship from the National Science Foundation in 2008 to support three
years of her thesis work. She has co-authored papers with Matt Goddeeris (Intracardiac
septation requires hedgehog-dependent cellular contributions from outside the heart,
Development 2008) and Jeff Kovacs (Arrestin development: emerging roles for betaarrestins in developmental signaling pathways, Developmental Cell, 2009). Chandra now
lives in Durham, North Carolina with her husband, Sam, and dog, Rothko. Her hobbies
include gardening, scuba diving, and traveling the world.
178