University of Southern California Dissertations and Theses

Utilizing zebrafish and mouse models to uncover the underlying genetics of human craniofacial anomalies

(USC Thesis Other)

Utilizing zebrafish and mouse models to uncover the underlying genetics of human craniofacial anomalies

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Utilizing Zebrafish and Mouse Models to Uncover the
Underlying Genetics of Human Craniofacial Anomalies
by
Camilla Sue Teng
A dissertation submitted in conformity with the requirements
for the degree of Doctor of Philosophy
in Genetic, Molecular, and Cellular Biology
Faculty of the University of Southern California Graduate School
May 2019
© Copyright by Camilla Sue Teng 2019

ii

Acknowledgments
I would like to express my gratitude to my advisors, Drs. Robert Maxson and Gage Crump, for
their guidance, support, and generosity. Their breadth of knowledge and expertise in scientific
writing, scientific critique, and rational thinking continually impresses me. I am truly fortunate to
have had such devoted, superb, and caring mentors.
I am grateful for my committee members, Drs. Neil Segil and Amy Merrill, for their constructive
advice and encouragement year after year.
I would like to offer special thanks to Drs. Man-Chun Ting, D’Juan Farmer, and Hai-Yun Yen
for their contributions to experiments and insightful discussions on research directions.
I would also like to acknowledge Dr. Seth Ruffins, Director of the Microscopy Core Facility of
USC Stem Cell, for his assistance in microscopy training and image analysis; Megan Matsutani
and Jennifer DeKoeyer Crump for exceptional fish care; and members of the Maxson and Crump
labs for stimulating conversations, strength, and cheer.
Finally, I thank my parents for their continuous support, patience, confidence, and so much
more. I am also appreciative of my extended family, boyfriend, friends, and classmates for their
encouragement, understanding, and listening ears.

iii

Table of Contents
ACKNOWLEDGMENTS II
TABLE OF CONTENTS III
LIST OF TABLES VI
LIST OF FIGURES VII
CHAPTER 1 SHIFTING BOUNDARIES: EVOLUTION, DEVELOPMENT, AND DISEASE OF THE VERTEBRATE
SKULL 9
ABSTRACT 9
1.1 DEVELOPMENT 9
1.1.1 TISSUE-TISSUE BOUNDARIES: NEURAL CREST AND MESODERM MIXING IN THE HEAD SKELETON 9
1.1.2 BONE-BONE BOUNDARIES: SUTURE MESENCHYME 12
1.2 EVOLUTION 13
1.2.1 EVOLUTION OF THE SHIFTING CORONAL SUTURE 13
1.2.2 DEFINING THE CORONAL SUTURE 15
1.3 DISEASE: BROKEN BOUNDARIES 16
1.4 ZEBRAFISH MODEL FOR CALVARIAL DEVELOPMENT 17
1.5 REFERENCES 19
CHAPTER 2 ALTERED BONE GROWTH DYNAMICS PREFIGURE CRANIOSYNOSTOSIS IN A ZEBRAFISH
MODEL OF SAETHRE-CHOTZEN SYNDROME 23
ABSTRACT 23
2.1 INTRODUCTION 23
2.2 RESULTS 27
2.2.1 SPECIFIC LOSS OF THE CORONAL SUTURE IN TCF12; TWIST1B MUTANT ZEBRAFISH 27
2.2.2 LOSS OF TCF12 PARTIALLY SUPPRESSES THE EMBRYONIC DEFECTS AND LETHALITY OF TWIST1 DEFICIENCY 30
2.2.3 EXPRESSION OF TCF12 AND TWIST1B AT MULTIPLE SUTURES IN ZEBRAFISH 32
2.2.4 ALTERED CALVARIAL BONE GROWTH PREFIGURES CORONAL SUTURE LOSS IN TCF12; TWIST1B MUTANT FISH 33
2.2.5 ALTERED DYNAMICS OF BONE FRONT CELLS IN TCF12; TWIST1B MUTANT FISH 36

iv

2.2.6 ALTERED CALVARIAL BONE GROWTH PRECEDES CORONAL SUTURE LOSS IN TCF12
+/-
; TWIST1
+/-
MICE 39
2.2.7 SELECTIVE REDUCTION OF THE OSTEOPROGENITOR POOL AT THE MUTANT CORONAL SUTURE 43
2.2.8 TISSUE-SPECIFIC ROLES FOR TWIST1 IN CALVARIAL BONE GROWTH AND SUTURE PATENCY 46
2.3 DISCUSSION 48
2.4 MATERIALS AND METHODS 50
2.4.1 ANIMALS 50
2.4.2 GENERATION OF ZEBRAFISH MUTANT LINES 51
2.4.3 SKULL PREPARATIONS 53
2.4.4 MICRO-COMPUTED TOMOGRAPHY 53
2.4.5 PARAFFIN EMBEDDING, SECTIONING, AND TISSUE HISTOLOGY 53
2.4.6 IN SITU HYBRIDIZATION 54
2.4.7 ZEBRAFISH LARVAE SKELETAL STAINING 54
2.4.8 SEQUENTIAL LIVE STAINING AND IMAGING 55
2.4.9 BRDU TREATMENTS AND IMMUNOHISTOCHEMISTRY 55
2.4.10 QUANTITATION AND STATISTICAL ANALYSES 56
2.5 REFERENCES 57
CHAPTER 3 REQUIREMENT FOR JAGGED1-NOTCH2 SIGNALING IN PATTERNING THE BONES OF THE
MOUSE AND HUMAN MIDDLE EAR 61
ABSTRACT 61
3.1 INTRODUCTION 61
3.2 RESULTS 64
3.2.1 LOSS OF JAG1 OR NOTCH2 IN NCCS RESULTS IN CRANIOFACIAL AND MIDDLE EAR DEFECTS 64
3.2.2 JAG1 IS NECESSARY FOR EARLY PATTERNING OF THE STAPES CARTILAGE BUT NOT FORMATION OF THE STAPEDIAL
ARTERY 67
3.2.3 JAG1 AND TWIST1 INTERACT TO PATTERN THE INCUS AND RETROTYMPANIC PROCESS 69
3.2.4 JAG1 IS REQUIRED IN NCCS FOR NORMAL HEARING IN MICE 71
3.2.5 CONDUCTIVE HEARING LOSS AND ANOMALIES IN MIDDLE EAR BONES IN PATIENTS HETEROZYGOUS FOR JAG1 LOSS-
OF-FUNCTION MUTATIONS 74
3.3 DISCUSSION 81
3.4 MATERIALS AND METHODS 86
3.4.1 MOUSE MUTANTS AND GENOTYPING 86

v

3.4.2 MICRO-COMPUTED TOMOGRAPHY OF MICE 86
3.4.3 SKULL PREPARATION 87
3.4.4 STAPES AND STAPEDIAL ARTERY VISUALIZATION 87
3.4.5 HEARING TESTS IN MICE 87
3.4.6 HEARING TESTS AND COMPUTED TOMOGRAPHY SCANS ON HUMAN SUBJECTS 87
3.5 REFERENCES 88
CHAPTER 4 PERSPECTIVES ON THE FUTURE 94
ABSTRACT 94
4.1 DIMERIZATION PARTNERS AS KEY TO DIFFERENTIAL REGULATION 94
4.2 CORONAL SUTURE SENSITIVITY TO GENETIC CHANGES 98
4.3 REFERENCES 100

vi

List of Tables
Table supplement 2-1. Summary of phenotypes observed in combinatorial zebrafish mutants 30
Table supplement 2-2. TALEN targeting and mutant genotyping 52
Table 3-1. Middle ear defects detected in Jag1, Notch2, and Twist1 mutant mice 66
Table 3-2. Hearing loss and middle ear defects in human subjects 78
Table supplement 3-1. Hearing test results of Alagille syndrome patients 79

vii

List of Figures
Figure 2-1. Coronal suture loss in tcf12; twist1b mutant zebrafish 26
Figure supplement 2-1. Zebrafish TALEN mutants 28
Figure supplement 2-2. Patent sutures in tcf12 single mutants 28
Figure supplement 2-3. Zebrafish skull phenotypes in different mutant combinations 29
Figure 2-2. Loss of tcf12 partially rescues twist1a; twist1b jaw cartilage defects and viability 31
Figure 2-3. Expression of tcf12 and twist1b in multiple sutures of zebrafish 33
Figure 2-4. Altered bone growth dynamics precede craniosynostosis in mutant zebrafish 35
Figure supplement 2-4. Accelerated bone fronts in tcf12; twist1b mutants transgenic for
sp7:EGFP 37
Figure 2-5. Altered proliferation and osteoblast production at mutant zebrafish bone fronts 38
Figure supplement 2-5. Additional examples of BrdU and Sp7 staining in tcf12; twist1b mutant
zebrafish 39
Figure 2-6. Altered bone growth dynamics in Tcf12
+/-
; Twist1
+/-
mice 41
Figure supplement 2-6. Increased calvarial bone front thickness in mutants 42
Figure 2-7. Reduced osteoprogenitor pool at the mutant coronal suture 44
Figure supplement 2-7. Quantification of progenitor marker and Fgf pathway transcripts in tcf12;
twist1b mutants 45
Figure 2-8. Tissue-autonomous bone overgrowth in Twist1 conditional mutants 47
Figure 3-1. Craniofacial and middle ear defects in mice deficient for Jag1 or Notch2 63
Figure supplement 3-1. Complete penetrance of stapes defects in Jag1 CKO mice 67
Figure 3-2. Mispatterning of middle ear cartilages and formation of the stapedial artery in Jag1-
deficient mice 68

viii

Figure 3-3. Incus and retrotympanic defects in Jag1; Twist1 compound mutants 70
Figure 3-4. Hearing loss in mice lacking Jag1 in NCCs 72
Figure supplement 3-2. Normal hearing in five-week-old mice lacking one copy of Jag1 in NCCs
73
Figure supplement 3-3. Normal gross inner ear structures in Jag1 CKO mice 74
Figure 3-5. Hearing loss and middle ear defects in patients heterozygous for JAG1 mutations
and/or diagnosed with AGS 76
Figure supplement 3-4. Barx1 expression in conditional Jag1 mutants 83
Figure 4-1. Jaw cartilage defects in twist1a; twist1b; twist3 mutants 97
Figure 4-2. Craniosynostosis detected in Tcf12
+/-
; E2A
+/-
mouse 98
Figure 4-3. Model of differential progenitor pools 100

9
Chapter 1
Shifting Boundaries: Evolution, Development, and Disease of the
Vertebrate Skull
Abstract
Proper skull development is regulated by boundaries – ones not to be broken but shifted.
Boundaries of various types can be found in the vertebrate skull. These include boundaries
between germ layer tissues that together form bones and physical boundaries between calvarial
bones known as sutures. I will review where these boundaries exist and how they are established
during development, explore how these boundaries have shifted during evolution, and discuss the
consequences of disruption of these boundaries.
1.1 Development
Human skull development begins well before birth; thus children are born with soft bones
shaping their heads. However, the brain volume of a full term child at birth is roughly 40 percent
that of an adult, and so the skull must continue to grow postnatally to accommodate brain
expansion. A mechanism to allow for this growth is cranial sutures. Cranial sutures are fibrous
joints that connect skull bones, and this connection serves multiple purposes throughout
development. The sutures act as a stem cell niche to facilitate continued bone growth of opposing
bones and also as a barrier to simultaneously inhibit fusion of adjacent bones. In addition to these
capabilities, the sutures impart flexibility to the skull, which is of particular importance for
natural childbirth and later brain growth. During childbirth, the head of the child must be
malleable to the external forces for a smoother birth. During brain maturation post-birth, the
sutures allow for internal pressures to be released.
1.1.1 Tissue-tissue boundaries: neural crest and mesoderm mixing in the
head skeleton
The head skeleton is comprised of bones derived from two tissue origins. Whereas most
elements are of a single tissue origin, some structures harbor two tissues that have merged to
form seamless boundaries. There is long-standing interest as to how these structures are formed
and where the tissue boundaries lie. An example of a bone of dual origin is the interparietal,
which lies posterior to the paired parietal bones and, in some cases, fuses to the supraoccipital
bone. Mouse neural crest lineage tracing labels only the interparietal medial area (Jiang et al.,
10

2002), while mesoderm lineage tracing marks the lateral area (Yoshida et al., 2008). From
observing fetal or perinatal stages of over three hundred species, Koyabu et al. found that the
interparietal bone could be singular, as in the Japanese macaque; paired, as in the Amur
hedgehog; tripartite, as in the bottle-nosed dolphin; or quadripartite, as in the cow. Moreover,
this varying number of interparietal bones can even be detected within a single species, as in
hydraxes (Koyabu et al., 2012). These results argue that the interparietal bone is formed from the
fusion of four bones, two located medially and likely derived from neural crest, and two laterally
and derived from mesoderm (Koyabu et al., 2012). Interestingly, the order of fusion of these
bones is not conserved, as evidenced by the pattern of bony elements in assorted animals. The
Amur hedgehog and a number of bat species have paired interparietals, suggesting the two lateral
mesodermal bone elements have fused to their adjacent medial neural crest elements. The bottle-
nosed dolphin was found having two lateral and one enlarged median interparietals, suggesting
the two medial neural crest elements have fused. Koyabu et al. also note other examples in which
the lateral mesodermal elements fuse first to the supraoccipital bone. Further investigation into
such diversity may provide hints to processes utilized in evolution to create or eliminate specific
cranial sutures resulting in wide-ranging skull shapes.
The chick frontal bone is another example of a bone having both a neural crest-derived
component and mesoderm-derived component. Early fate-mapping studies investigating neural
crest contributions to the chick craniofacial skeleton relied on quail-chick chimeras that produced
disagreeing outcomes, all of which have since been revisited in detail (Gross and Hanken, 2008).
Briefly, grafting experiments by Le Lièvre in 1978 (Le Lievre, 1978) and Noden in 1982
(Noden, 1982) suggested only partial neural crest contribution in the frontal bone. On the other
hand, Couly et al. in 1993 (Couly et al., 1993) claimed that the entirety of the frontal bone is of
neural crest origin. Differences in the timing, amount of tissue, and position from which tissue
was taken for grafting, may provide an explanation for the discrepancy. In 2006, Evans and
Noden followed up with more precise labeling of both neural crest and mesoderm cells with
retroviral injections (Evans and Noden, 2006). These higher resolution results supported Noden’s
earlier finding that the chick frontal bone is of dual origin. Similar to the interparietal bone, the
chick frontal bone is thought to be derived from two ossification centers. A more recent fate-
mapping study finds that this frontal bone is formed from at least two distinct condensations
(Abzhanov et al., 2007). However, in the same study, bone staining across a series of stages does
not clearly show separate bones in early stages. So, perhaps neural crest and mesoderm cells
11

separately form adjacent condensations, before joining together to ossify and produce the chick
frontal bone.
While the interparietal bone is likely formed from fusion events of mature neural crest bone and
mesoderm bone, some other bones develop after neural crest and mesoderm bone progenitors
have already intermingled. One such example is the mammalian stapes, a middle ear ossicle.
This bone is a stirrup-shaped structure that serves to enhance hearing through air conduction.
The mouse stapes consists of neural crest-derived head and crus, and in contrast a footplate of
both neural crest and mesoderm origin (Thompson et al., 2012). The stapes is a bone that forms
through a cartilage precursor; and as early as embryonic day 12.5, Sox9
+
chondrocytes of the
footplate are a combination of Wnt1-Cre traced neural crest derivatives and non-neural crest
(presumably mesoderm) derivatives (Thompson et al., 2012). Thompson et al. also discovered
that the mesoderm-derived regions fail to form in the absence of neural crest derivatives,
suggesting the mesoderm is dependent on signaling cues from the neural crest. Another example
is the zebrafish frontal bone. Two groups independently generated sox10 reporters, traced
expression to adulthood, and found that sox10 labeled neural crest cells only contributed to the
anterior portion of zebrafish frontal bones (Kague et al., 2012; Mongera et al., 2013). The
unlabeled posterior portion and parietal bones are presumably mesoderm derived. Repeated live
imaging performed through early stages of zebrafish skull development with a RUNX2 reporter
for osteoblast revealed that the zebrafish frontal bone forms from a single ossification (Kague et
al., 2016), suggesting that neural crest and mesoderm cells are mixed early on, prior to bone
matrix formation.
With both neural crest and mesoderm producing bone, many have wondered if the tissue origin
confers functional distinctions. In the middle ear, endoderm and neural crest together form a
continuous epithelium to line the mouse middle ear cavity and auditory tube, yet only the
endoderm-derived portion is ciliated (Thompson and Tucker, 2013). The endodermal epithelium
thus has the distinct function of debris clearing, but not the neural crest-derived epithelium. So,
in the same vein, it is possible that neural crest-derived bone may have different properties than
mesoderm-derived bone. Some studies have focused on the prominent bones of the skullcap,
mainly the paired frontal bones and paired parietal bones. Although these bones are similar in
morphology and structure, the mammalian frontal bones are neural crest derived, and parietal
bones are mesoderm derived (Jiang et al., 2002; Yoshida et al., 2008). The function of these
bones is identical – each acts as armor to shield the brain from impact. A key question is thus
12

whether there is a biological difference if the armor is made of neural crest or mesoderm. In
attempts to resolve possible differences, Quarto et al. used a combination of in vitro and in vivo
wound healing assays on mouse models. They find that osteoblasts of the neural crest lineage in
mouse frontal bones have superior osteogenic potential, as compared to the parietal bone
osteoblasts of a mesoderm lineage, which suggested that embryonic origins may impart different
biological activity (Quarto et al., 2010). However, this distinction could be explained in other
ways such as variations in external or positional cues. Because the frontal bones of zebrafish and
chick are of neural crest and mesoderm, perhaps similar experiments performed on these animal
models may prove more insightful. Thus, differences in neural crest-derived calvarial bone and
mesoderm-derived calvarial bone remain unsettled.
1.1.2 Bone-bone boundaries: suture mesenchyme
At the borders of calvarial bones are cranial sutures that have dual functions of continuing bone
growth, while keeping apart the abutting bones. The suture mesenchyme contains progenitor
cells that maintain bone growth and contribute to bone healing, as several recent studies have
demonstrated. The mature mouse suture at one month of age specifically expresses Gli1, though
the expression pattern is much broader earlier in development (Zhao et al., 2015). When these
Gli1
+
cells are traced, starting at one month of age, osteocytes in the nearby bone are also labeled
at eight months of age. Zhao et al. also found that the Gli1
+
suture cells turn highly proliferative
in response to calvarial bone injury, and the injured site is later positive with osteocytes derived
from these Gli1
+
cells. A comparable set of experiments by Maruyama et al. identified restricted
Axin2 expression in the suture mesenchyme, and the Axin2
+
cells are also capable of skeletal
maintenance and repair (Maruyama et al., 2016). Similarly, Prrx1 also marks cells in the cranial
sutures that are responsible for bone healing (Takarada et al., 2015; Wilk et al., 2017), though
expression of Prrx1 seem much broader than those of Gli1 and Axin2. As these studies are
relatively recent, it is unclear the extent to which these three genes mark similar populations of
suture mesenchyme. Ongoing single-cell RNA sequencing experiments by independent groups
and the FaceBase consortium should provide a better understanding.
Also unclear is when these newly-identified postnatal suture markers are first expressed, and
whether the markers for the mature suture is the same as the developing suture. Gli1, as well as
Grem1 that has been identified as a general skeletal stem cell marker (Worthley et al., 2015),
have recently been shown to be expressed in the developing suture region of mice and zebrafish
13

(Teng et al., 2018), though long-term lineage tracing of Grem1
+
cells is still lacking.
Nevertheless, the same study confirms conservation of gli1, prrx1a, and grem1a in the
developing and mature zebrafish sutures. An earlier study of the mouse embryonic sagittal suture
found Msx1/2 highly concentrated in the region between parietal bones (Kim et al., 1998). Kim
et al. also noted several interestingly specific expression patterns. First, Bmp2 was detected
condensed at the osteogenic fronts of the parietal bones at embryonic day (E) 15 but dispersed
and weak by E18 (Kim et al., 1998). Second, Shh and Ptc expression first appeared at E18
osteogenic fronts and increased in postnatal stages; additionally, expression was high in metopic
and sagittal sutures but deficient in coronal sutures (Kim et al., 1998). This apparent differential
timing and patterning of expression may be a clue to the larger underlying network regulating
suture formation and maintenance. Future work following various markers through a long
duration of prenatal and postnatal development should be particularly insightful.
1.2 Evolution
In some of the earliest observational studies of animal anatomy, animal skeleton nomenclature
was established by directly transferring terminology from the human skeleton (Schultze, 2008).
While this may be adequate for some bones such as a radius and ulna of a chicken, the system
would not apply to skeletal elements that have no physical equivalents in humans. For example,
an adult human skull is composed of 22 bones, but a zebrafish skull consists of 74 bones
(Cubbage and Mabee, 1996). Thus, mammalian evolution involved an overall progressive
reduction or simplification in the number of skull bones, along with repurposing and creation of
new and highly specialized bones. The three mammalian middle ear bones, malleus, incus, and
stapes, are highly derived and function for the sole purpose of amplifying sound waves that
travel through air. These three bones are thought of being potentially homologous to the
endochondral portion of the proximal end of Meckel’s, the palatoquadrate, and the hyomandibula
of fish (Medeiros and Crump, 2012), and they are similarly regulated by Jagged-Notch signaling
across fish, mouse, and human (Teng et al., 2017). Consequently, it is important to note that the
current nomenclature was established according to physical morphology, and did not have
knowledge of homology as defined by embryonic tissue origins or genetic regulations.
1.2.1 Evolution of the shifting coronal suture
Over the course of evolution, the neural crest-mesoderm boundary and cranial sutures have
shifted in relation to each other. A prominent neural crest-mesoderm boundary runs transversely
14

across the skull of many animals. However, the position of this border is quite variable across
vertebrate species, and in many cases does not demarcate a clear border between bones.
Fascinatingly, the relationship between this neural crest-mesoderm boundary and coronal suture
is not even conserved amongst animals in the same clade. In mice, this tissue boundary has been
clearly demonstrated to lie between the frontal and parietal bones. A Wnt1-Cre allele marking all
neural crest derivatives labeled the frontal bones (Jiang et al., 2002), and a Mesp1-Cre allele
tracing all mesoderm derivatives marked the coronal sutures and parietal bones (Yoshida et al.,
2008). As previously mentioned, quail-chick grafting, and retroviral injection labeling
experiments in the chick, places the neural crest-mesoderm boundary within the frontal bones
(Evans and Noden, 2006; Noden, 1982). The mouse and chick are both amniotes derived from
the lobe-finned fish clade, but the boundary position in the chick is instead similar to the
zebrafish, a ray-finned fish. In zebrafish, neural crest contributes to only the anterior frontal
bone, as seen in genetic lineage tracings of the Sox10-marked neural crest population (Kague et
al., 2012; Mongera et al., 2013). The mesoderm population was not traced and presumed to be
the regions negative of lineage labeling.
Representing the class of amphibians, frogs and axolotls have been examined. Long-term fate
mapping in Mexican axolotl, Ambystoma mexicanum, was possible through transplantation of
neural crest or mesoderm from a GFP-positive donor to a wild-type host. To date, the axolotl,
along with mice and chick, are the only animal models in which both neural crest and mesoderm
populations were traced to resolve their respective contributions and the tissue boundary. Results
illustrate a neural crest-only frontal bone and mesoderm-only parietal bone, with no mentioning
of the coronal suture mesenchyme (Maddin et al., 2016). Hence, the neural crest-mesoderm
boundary in the axolotl is similar to that of mice. In the African clawed frog, Xenopus laevis,
neural crest cells from donors ubiquitously expressing GFP were grafted into wild-type hosts.
Assessment of bones after metamorphosis revealed considerable neural crest contribution
throughout the skull (Piekarski et al., 2014). Frogs have a single frontoparietal bone that is
thought to be a fusion of the frontal and parietal bones. Piekarski et al. claims that this bone is
solely neural crest derived, though supporting data is still lacking. Especially since it is known
from chick and zebrafish that bones can be of dual origin, whole skull analyses would be
advantageous over tissue section analyses. Nonetheless, if Piekarski et al.’s claim is true, then
this is yet another case of divergence in boundary position within a clade.
15

Taking together observations from the different animal models, several, non-mutually exclusive
arguments could be made. First, one of the patterns represents the ancestral condition, and
deviations arose multiple times during evolution. Conceivably, the tissue boundary positioned
within frontal bones seen in chick and zebrafish is closest to the common ancestor, suggesting
boundary shifts occurred after the split between lobe-finned and ray-finned fishes. Second, the
high frequency that this neural crest-mesoderm boundary has shifted as observed in the few
species examined suggests little importance in skull development. Rather, the shifting of the
neural crest-mesoderm boundary relative to the coronal suture may reflect a fluid nature of
cranial sutures that is advantageous for skull evolution. Given that adult humans have fewer skull
bones than fish, many cranial sutures must have been lost. Perhaps the number and position of
sutures are deterministic of skull shape variability. These changes in tissue boundary and suture
positions have fueled calls for nomenclature reformation. Some suggest the use of quotation
marks to distinguish fish bones that are similarly named as mammalian counterparts but have
derived from a different embryonic origin (Schultze, 2008). Others propose that the chick frontal
bone is a fusion of a neural crest frontal and mesoderm parietal, and thus the frontal bone should
be termed “frontoparietal” and parietal bone “postparietal” instead to better reflect homology
(Maddin et al., 2016).
1.2.2 Defining the coronal suture
The mouse coronal suture lies at the neural crest-mesoderm boundary between the frontal and
parietal bones, yet the zebrafish coronal suture lies wholly within the boundary of mesoderm-
derived bones. What then defines this suture, or at least its position? If the coronal suture must
coincide with the neural crest and mesoderm boundary, then such a suture does not exist in
certain animals such as the zebrafish and chick. In accordance with this logic, the currently
named zebrafish coronal suture that lies between the mesoderm portion of the frontal bone and
the mesoderm-derived parietal bone is more likely a post-parietal suture that is not seen in mice
and humans. On the other hand, perhaps a physical landmark should be a defining feature of a
coronal suture. In the mouse, short cartilage bars exist on the lateral sides of the head where the
coronal suture begins to form. In the zebrafish, a cartilage rod, the epiphyseal bar, running left to
right across the skull sits underneath where neural crest and mesoderm meet in the frontal bone.
The conserved positioning of this cartilage at the neural crest-mesoderm boundary in both mice
and zebrafish is further evidence arguing that the zebrafish coronal suture is not a true coronal
suture. Yet another landmark that may demarcate the coronal suture is the forebrain-midbrain
16

boundary. Fabbri et al. evaluated skull morphology in relation to the brain in multitudes of
Reptilia representatives, ranging from extinct dinosaurs to extant alligators and avian, and found
a significant correlation between the forebrain-midbrain boundary and the coronal suture
(referred to also as frontoparietal suture) (Fabbri et al., 2017). Although the zebrafish and chick
similarly have frontal bones of mixed embryonic origins, the zebrafish coronal suture in fact
more highly correlates with the midbrain-hindbrain boundary. Interestingly, the zebrafish
forebrain-midbrain boundary instead coincides with the neural crest-mesoderm boundary as well
as the epiphyseal bar. The collective observations point to the position where the coronal suture
is expected but is not present in the zebrafish. Unlike the brains of chick and zebrafish that have
a relatively linear morphology, human brains are more globular in shape. Perhaps because the
human is more derived, the forebrain constitutes much of the human brain volume. The human
forebrain is so enlarged that it nearly encapsulates the midbrain and hindbrain, and hence the
human coronal suture overlies the forebrain. Thus, the forebrain-midbrain boundary as a
reference point for the coronal suture may not be conserved in humans. As an extension of this
idea, the neural crest-mesoderm boundaries in humans are unknown, so the positioning of this
boundary at the coronal suture may also not be conserved in humans.
1.3 Disease: broken boundaries
While the benefits of cranial sutures are multi-fold, they do present as weak points of the skull as
well. Thus, once the human brain has fully developed, the calvarial bones eventually fuse to
eliminate the cranial sutures. This process of bone fusion must be well-timed, as premature loss
of this suture boundary, termed craniosynostosis, will constrain brain expansion. The prevalence
of human craniosynostosis is roughly one in 2,500 live births, and surgical intervention is often
required to release intracranial pressure and provide space for the growing brain. The crude
corrective surgery of sawing apart the fused bones is followed by a long period of helmet
wearing and sometimes a second fusion event, critically compromising the quality of life. Hence,
the genetics and molecular biology behind suture development needs to be well analyzed in order
to even begin imagining a cure.
To date, a number of genetic mutations have been identified to cause craniosynostosis, often as a
phenotype of developmental disorders. Correspondingly, the roles of these identified genes in
normal suture and calvarial bone development has been studied using genetic animal models.
Results from a number of studies coupling genetic mutants with neural crest or mesoderm
17

lineage tracing revealed an unusual mixing of these two tissue types and also improper migration
of osteogenic cell into the suture mesenchyme (Merrill et al., 2006; Ting et al., 2009; Yen et al.,
2010). Accordingly, craniosynostosis is a disease resulting from disruption of the tissue-tissue
boundary as well as the osteogenic and non-osteogenic boundary.
Mouse calvarial development initiates and occurs mostly in utero, which makes it difficult to
continuously track cellular events. Indeed, many studies have revolved around static images
taken at embryonic time points to stitch together a broader picture of the developmental process.
While this is sufficient for a general understanding of normal growth, a comprehensive
assessment of the disease state would benefit from the ability to continuously follow an affected
individual. This is particularly important in cases of incomplete phenotype penetrance to
establish any early changes to be causative of later defects. In attempts to address this concern
and overcome the uterus barrier, groups have experimented with short-term explant cultures of
embryonic mouse calvaria. With explants, precise questions, such as the role of dura mater in
regulating proper suture formation or effects of localized protein expression can be investigated
(Connerney et al., 2008; Kim et al., 1998; Rice et al., 2000). While the advantages of explants
have not yet been thoroughly exploited and continues to be powerful when paired with live
imaging, this tool provides only at best a restricted window into the lengthy process of calvarial
development, with the greatest limitation being the duration the tissue can survive and continue
normal development in culture conditions.
1.4 Zebrafish model for calvarial development
While comparative developmental biology has illustrated the evolutionary nuances of skull
development in the animal kingdom, a major goal in science is to conduct comparative and
functional studies to ultimately understand human biology and disease. To this end, as long as
differences and caveats of a model organism are accounted for in interpretive studies, then the
model is an effective one regardless of naming. The zebrafish model is thoroughly used to study
early embryonic patterning, organ development, and regeneration capacities. Zebrafish gained
popularity as an oviparous animal, which makes the animals convenient for manipulations and
observations. Aided by their removable chorions and transparent bodies, fish embryos can be
imaged for an extended period of time at any point of their development. With advancing
microscopy technology allowing faster and higher resolution imaging, repeated manipulations
and observations at juvenile and adult stages are also feasible (Kague et al., 2016; Teng et al.,
18

2018). Additional measures, such as intubation, to deliver constant anesthesia can further
prolong the imaging window (Rasmussen et al., 2018). As fertilization and development occur
entirely external of a womb, the entire calvarial developmental process is traceable and hence
visible. Although the zebrafish head is more elongated compared to a rounder human head, the
general anatomy of a zebrafish skull vault is rather similar to that of humans, as introduced
earlier. Furthermore, similar gene expression profiles have also been found at the zebrafish bone
fronts and suture region to that of mice (Topczewska et al., 2016). Hence, the fish is an attractive
model to study calvarial development and disease, complementing the current mouse and mouse
explant culture models.
In recent years, several groups have taken up the zebrafish model for investigating the genetics
and signaling pathways in calvarial development. One of the first cases is a hypomorphic allele,
named stocksteif, for the retinoic acid-inactivating enzyme cyp26b1. The mutant animals are
semiviable, with adults being extremely undersized and presenting with severe craniofacial
abnormalities including coronal synostosis and separation of the parietal bones at the sagittal
suture (Laue et al., 2011). As altered retinoic acid levels have a teratogenic effect, the etiology of
the craniosynostosis phenotype is difficult to separate from other defects. In this case, the
extreme facial hypoplasia could potentially be a primary defect driving the secondary suture
defect. Although the overall phenotype is comparable to a human subject (Laue et al., 2011), the
group turned to chemical treatments instead in a follow-up study to reveal amplified osteoclast
activity with increased retinoic acid (Jeradi and Hammerschmidt, 2016). The result provided an
explanation to the fragmented skull bones and holes at where the sagittal suture should be, but it
is unclear whether this model also presents with craniosynostosis. Another genetic model is the
devoid of blastema mutant that harbors an fgf20a mutation. These mutants have reduction of the
midface and upper jaw and, interestingly, branched sutures (Cooper et al., 2013). As FGF20
binds FGFR2, which when mutated results in an array of craniofacial related syndromes, Cooper
et al. believes that this fish model can be used to understand certain aspects of Apert and
Crouzon syndromes. Additionally, the authors argue that the fgf20a mutants can provide insights
into skull evolution in terms of feeding diversification, as Fgf20a has a role in regulating jaw
shape (Cooper et al., 2013). On the other end of the calvarial defect spectrum is an sp7 (also
known as osterix) mutant that has a relatively normal body skeleton but malformed craniofacial
skeleton, with defects including midface hypoplasia and wormian, or ectopic, skull bones (Kague
et al., 2016). By combining this zebrafish mutant with transgenic reporter lines, Kague et al.
19

learned that the wormian bones are a result of ectopic bone initiation sites and not fragmentation
of a larger bone. This is reflective of a rarely identified human patient harboring an OSX
mutation and presenting with midface hypoplasia, wormian bones, and osteogenesis imperfecta
(Lapunzina et al., 2010). Further solidifying the zebrafish model as relevant for human
craniosynostosis is a tcf12; twist1b mutant model for Saethre-Chotzen syndrome that I generated
for my thesis project (see Chapter 2). The synostosis phenotype in fish mutants was similar to
what was previously described in mouse mutants and human patients, offering definitive
evidence of conserved genetic regulation. This result addresses the worries stemming from the
shifted neural crest-mesoderm boundary relative to the coronal sutures in the evolutionary
context. From repeated live imaging of this model, I detected dynamically altered bone growth in
mutants prior to synostosis, a finding difficult to uncover from static images of mouse calvarial
development.
Lessons from these zebrafish calvarial studies have made a convincing case for the zebrafish as a
suitable paradigm in uncovering the genetic, cellular, and molecular basis of human
craniosynostosis. The mouse has been the long preferred disease model and has indeed been
useful in laying the foundation to our current understanding. The discussed fish studies, while
individually having specific foci, all repeatedly establish the great extent of commonalities
between fish and mammalian calvarial development. Though the zebrafish proves to be a faithful
model in these cases, there remain other conditions that the fish may not be able to recapitulate.
For example, if a particular human coronal synostosis case is a secondary result from a primary
neural crest defect, then the zebrafish may not be reflective of the disease state. Overall, the
zebrafish is an untapped resource that can continue to be utilized, with caution on a case-by-case
basis, to fill in the gaps of knowledge necessary for a comprehensive picture of human calvarial
development and disease.
1.5 References
Abzhanov, A., Rodda, S.J., McMahon, A.P., Tabin, C.J., 2007. Regulation of skeletogenic
differentiation in cranial dermal bone. Development 134, 3133-3144.
Connerney, J., Andreeva, V., Leshem, Y., Mercado, M.A., Dowell, K., Yang, X., Lindner, V.,
Friesel, R.E., Spicer, D.B., 2008. Twist1 homodimers enhance FGF responsiveness of the cranial
sutures and promote suture closure. Developmental biology 318, 323-334.
20

Cooper, W.J., Wirgau, R.M., Sweet, E.M., Albertson, R.C., 2013. Deficiency of zebrafish fgf20a
results in aberrant skull remodeling that mimics both human cranial disease and evolutionarily
important fish skull morphologies. Evolution & development 15, 426-441.
Couly, G.F., Coltey, P.M., Le Douarin, N.M., 1993. The triple origin of skull in higher
vertebrates: a study in quail-chick chimeras. Development 117, 409-429.
Cubbage, C.C., Mabee, P.M., 1996. Development of the cranium and paired fins in the zebrafish
Danio rerio (Ostariophysi, cyprinidae). Journal of Morphology 229, 121-160.
Evans, D.J., Noden, D.M., 2006. Spatial relations between avian craniofacial neural crest and
paraxial mesoderm cells. Developmental dynamics : an official publication of the American
Association of Anatomists 235, 1310-1325.
Fabbri, M., Mongiardino Koch, N., Pritchard, A.C., Hanson, M., Hoffman, E., Bever, G.S.,
Balanoff, A.M., Morris, Z.S., Field, D.J., Camacho, J., Rowe, T.B., Norell, M.A., Smith, R.M.,
Abzhanov, A., Bhullar, B.S., 2017. The skull roof tracks the brain during the evolution and
development of reptiles including birds. Nature ecology & evolution 1, 1543-1550.
Gross, J.B., Hanken, J., 2008. Review of fate-mapping studies of osteogenic cranial neural crest
in vertebrates. Developmental biology 317, 389-400.
Jeradi, S., Hammerschmidt, M., 2016. Retinoic acid-induced premature osteoblast-to-
preosteocyte transitioning has multiple effects on calvarial development. Development.
Jiang, X., Iseki, S., Maxson, R.E., Sucov, H.M., Morriss-Kay, G.M., 2002. Tissue origins and
interactions in the mammalian skull vault. Developmental biology 241, 106-116.
Kague, E., Gallagher, M., Burke, S., Parsons, M., Franz-Odendaal, T., Fisher, S., 2012.
Skeletogenic fate of zebrafish cranial and trunk neural crest. PloS one 7, e47394.
Kague, E., Roy, P., Asselin, G., Hu, G., Simonet, J., Stanley, A., Albertson, C., Fisher, S., 2016.
Osterix/Sp7 limits cranial bone initiation sites and is required for formation of sutures.
Developmental biology 413, 160-172.
Kim, H.J., Rice, D.P., Kettunen, P.J., Thesleff, I., 1998. FGF-, BMP- and Shh-mediated
signalling pathways in the regulation of cranial suture morphogenesis and calvarial bone
development. Development 125, 1241-1251.
Koyabu, D., Maier, W., Sanchez-Villagra, M.R., 2012. Paleontological and developmental
evidence resolve the homology and dual embryonic origin of a mammalian skull bone, the
interparietal. Proceedings of the National Academy of Sciences of the United States of America
109, 14075-14080.
Lapunzina, P., Aglan, M., Temtamy, S., Caparros-Martin, J.A., Valencia, M., Leton, R.,
Martinez-Glez, V., Elhossini, R., Amr, K., Vilaboa, N., Ruiz-Perez, V.L., 2010. Identification of
a frameshift mutation in Osterix in a patient with recessive osteogenesis imperfecta. American
journal of human genetics 87, 110-114.
Laue, K., Pogoda, H.M., Daniel, P.B., van Haeringen, A., Alanay, Y., von Ameln, S.,
Rachwalski, M., Morgan, T., Gray, M.J., Breuning, M.H., Sawyer, G.M., Sutherland-Smith, A.J.,
21

Nikkels, P.G., Kubisch, C., Bloch, W., Wollnik, B., Hammerschmidt, M., Robertson, S.P., 2011.
Craniosynostosis and multiple skeletal anomalies in humans and zebrafish result from a defect in
the localized degradation of retinoic acid. American journal of human genetics 89, 595-606.
Le Lievre, C.S., 1978. Participation of neural crest-derived cells in the genesis of the skull in
birds. Journal of embryology and experimental morphology 47, 17-37.
Maddin, H.C., Piekarski, N., Sefton, E.M., Hanken, J., 2016. Homology of the cranial vault in
birds: new insights based on embryonic fate-mapping and character analysis. Royal Society open
science 3, 160356.
Maruyama, T., Jeong, J., Sheu, T.J., Hsu, W., 2016. Stem cells of the suture mesenchyme in
craniofacial bone development, repair and regeneration. Nature communications 7, 10526.
Medeiros, D.M., Crump, J.G., 2012. New perspectives on pharyngeal dorsoventral patterning in
development and evolution of the vertebrate jaw. Developmental biology 371, 121-135.
Merrill, A.E., Bochukova, E.G., Brugger, S.M., Ishii, M., Pilz, D.T., Wall, S.A., Lyons, K.M.,
Wilkie, A.O., Maxson, R.E., Jr., 2006. Cell mixing at a neural crest-mesoderm boundary and
deficient ephrin-Eph signaling in the pathogenesis of craniosynostosis. Human molecular
genetics 15, 1319-1328.
Mongera, A., Singh, A.P., Levesque, M.P., Chen, Y.Y., Konstantinidis, P., Nusslein-Volhard, C.,
2013. Genetic lineage labeling in zebrafish uncovers novel neural crest contributions to the head,
including gill pillar cells. Development 140, 916-925.
Noden, D.M., 1982. Patterns and organization of craniofacial skeletogenic and myogenic
mesenchyme: a perspective. Progress in clinical and biological research 101, 167-203.
Piekarski, N., Gross, J.B., Hanken, J., 2014. Evolutionary innovation and conservation in the
embryonic derivation of the vertebrate skull. Nature communications 5, 5661.
Quarto, N., Wan, D.C., Kwan, M.D., Panetta, N.J., Li, S., Longaker, M.T., 2010. Origin matters:
differences in embryonic tissue origin and Wnt signaling determine the osteogenic potential and
healing capacity of frontal and parietal calvarial bones. Journal of bone and mineral research :
the official journal of the American Society for Bone and Mineral Research 25, 1680-1694.
Rasmussen, J.P., Vo, N.T., Sagasti, A., 2018. Fish Scales Dictate the Pattern of Adult Skin
Innervation and Vascularization. Developmental cell 46, 344-359 e344.
Rice, D.P., Aberg, T., Chan, Y., Tang, Z., Kettunen, P.J., Pakarinen, L., Maxson, R.E., Thesleff,
I., 2000. Integration of FGF and TWIST in calvarial bone and suture development. Development
127, 1845-1855.
Schultze, H.-P., 2008. Nomenclature and homologization. Mesozoic Fishes 4, 23-48.
Takarada, T., Nakazato, R., Tsuchikane, A., Fujikawa, K., Iezaki, T., Yoneda, Y., Hinoi, E.,
2015. Genetic analysis of Runx2 function during intramembranous ossification. Development.
22

Teng, C.S., Ting, M.C., Farmer, D.T., Brockop, M., Maxson, R.E., Crump, J.G., 2018. Altered
bone growth dynamics prefigure craniosynostosis in a zebrafish model of Saethre-Chotzen
syndrome. eLife 7.
Teng, C.S., Yen, H.Y., Barske, L., Smith, B., Llamas, J., Segil, N., Go, J., Sanchez-Lara, P.A.,
Maxson, R.E., Crump, J.G., 2017. Requirement for Jagged1-Notch2 signaling in patterning the
bones of the mouse and human middle ear. Scientific reports 7, 2497.
Thompson, H., Ohazama, A., Sharpe, P.T., Tucker, A.S., 2012. The origin of the stapes and
relationship to the otic capsule and oval window. Developmental dynamics : an official
publication of the American Association of Anatomists 241, 1396-1404.
Thompson, H., Tucker, A.S., 2013. Dual origin of the epithelium of the mammalian middle ear.
Science 339, 1453-1456.
Ting, M.C., Wu, N.L., Roybal, P.G., Sun, J., Liu, L., Yen, Y., Maxson, R.E., Jr., 2009. EphA4 as
an effector of Twist1 in the guidance of osteogenic precursor cells during calvarial bone growth
and in craniosynostosis. Development 136, 855-864.
Topczewska, J.M., Shoela, R.A., Tomaszewski, J.P., Mirmira, R.B., Gosain, A.K., 2016. The
Morphogenesis of Cranial Sutures in Zebrafish. PloS one 11.
Wilk, K., Yeh, S.A., Mortensen, L.J., Ghaffarigarakani, S., Lombardo, C.M., Bassir, S.H.,
Aldawood, Z.A., Lin, C.P., Intini, G., 2017. Postnatal Calvarial Skeletal Stem Cells Expressing
PRX1 Reside Exclusively in the Calvarial Sutures and Are Required for Bone Regeneration.
Stem cell reports 8, 933-946.
Worthley, D.L., Churchill, M., Compton, J.T., Tailor, Y., Rao, M., Si, Y., Levin, D., Schwartz,
M.G., Uygur, A., Hayakawa, Y., Gross, S., Renz, B.W., Setlik, W., Martinez, A.N., Chen, X.,
Nizami, S., Lee, H.G., Kang, H.P., Caldwell, J.M., Asfaha, S., Westphalen, C.B., Graham, T.,
Jin, G., Nagar, K., Wang, H., Kheirbek, M.A., Kolhe, A., Carpenter, J., Glaire, M., Nair, A.,
Renders, S., Manieri, N., Muthupalani, S., Fox, J.G., Reichert, M., Giraud, A.S., Schwabe, R.F.,
Pradere, J.P., Walton, K., Prakash, A., Gumucio, D., Rustgi, A.K., Stappenbeck, T.S., Friedman,
R.A., Gershon, M.D., Sims, P., Grikscheit, T., Lee, F.Y., Karsenty, G., Mukherjee, S., Wang,
T.C., 2015. Gremlin 1 identifies a skeletal stem cell with bone, cartilage, and reticular stromal
potential. Cell 160, 269-284.
Yen, H.Y., Ting, M.C., Maxson, R.E., 2010. Jagged1 functions downstream of Twist1 in the
specification of the coronal suture and the formation of a boundary between osteogenic and non-
osteogenic cells. Developmental biology 347, 258-270.
Yoshida, T., Vivatbutsiri, P., Morriss-Kay, G., Saga, Y., Iseki, S., 2008. Cell lineage in
mammalian craniofacial mesenchyme. Mechanisms of development 125, 797-808.
Zhao, H., Feng, J., Ho, T.V., Grimes, W., Urata, M., Chai, Y., 2015. The suture provides a niche
for mesenchymal stem cells of craniofacial bones. Nature cell biology 17, 386-396.
23

Chapter 2
Altered bone growth dynamics prefigure craniosynostosis in a
zebrafish model of Saethre-Chotzen syndrome
Abstract
Cranial sutures separate the skull bones and house stem cells for bone growth and repair. In
Saethre-Chotzen syndrome, mutations in TCF12 or TWIST1 ablate a specific suture, the coronal.
This suture forms at a neural-crest/mesoderm interface in mammals and a mesoderm/mesoderm
interface in zebrafish. Despite these differences, results show that the coronal suture of zebrafish
lacking TCF12 and TWIST1 homologs is similarly lost. Sequential bone staining reveals an
initial, directional acceleration of bone production in the mutant skull, with subsequent localized
stalling of bone growth prefiguring coronal suture loss. Mouse genetics further reveal
requirements for Twist1 and Tcf12 in both the frontal and parietal bones for suture patency, and
to maintain putative progenitors in the coronal region. These findings reveal conservation of
coronal suture formation despite evolutionary shifts in embryonic origins, and suggest that the
coronal suture might be especially susceptible to imbalances in progenitor maintenance and
osteoblast differentiation.

2.1 Introduction
Craniofacial anomalies are among the most common congenital defects, encompassing cleft lip
and palate, facial malformations, and abnormalities in the flat bones forming the top of the skull.
Craniosynostosis involves the premature fusion of the skull bones at sutures, fibrous structures
that join the bones. During normal development, cranial sutures hold bones of the skull in place
while providing malleability required during childbirth. Cranial sutures also contain stem cells
that allow for continued skull bone growth as the brain and head structure expands (Zhao et al.,
2015). While human skull bones eventually fuse later in life, precocious bone fusion in
craniosynostosis correlates with region-specific defects in bone growth that often negatively
impact growth of the underlying brain. Genetic causes of craniosynostosis include mutations in
genes that participate in developmental signaling pathways, including FGFRs 1, 2, and 3
(Hajihosseini, 2008), TGFBRs 1 and 2 (Loeys et al., 2005), and the Notch ligand JAGGED1
(Kamath et al., 2002). A central unanswered question is the extent to which defects in these
24

pathways result in a failure to maintain postnatal stem cells once the sutures have formed, or
whether these pathways have roles in the specification and maintenance of bone progenitors
during earlier phases of bone growth preceding suture formation (Flaherty et al., 2016). Most
studies have focused on the maintenance of postnatal sutural stem cells, as these stem cells have
only been recently marked with a variety of reporters in mice, for example based on Gli1 (Zhao
et al., 2015), Axin2 (Maruyama et al., 2016), and Prrx1 (Wilk et al., 2017). Less is known,
however, about how potential defects in the embryonic progenitors that grow the skull bones
may prefigure suture loss, although a few reports describe altered bone formation during
embryonic stages in craniosynostosis mutants (Merrill et al., 2006). The ex utero development
and transparency of zebrafish provide a unique opportunity to track the progenitors that form the
skull bones, and to better understand how defects in the dynamics of bone growth relate to a later
ability to form and maintain sutures once the bones come together (Quarto and Longaker, 2005;
Laue et al., 2011; Kague et al., 2016; Topczewska et al., 2016).
A striking feature of many forms of craniosynostosis is that only particular sutures are affected.
For example, in Saethre-Chotzen syndrome, the second most common form of craniosynostosis,
the coronal suture is selectively lost. The majority of Saethre-Chotzen patients harbor
heterozygous loss-of-function mutations in TWIST1 or TCF12, which encode basic helix-loop-
helix transcription factors (el Ghouzzi et al., 1997; Howard et al., 1997; Sharma et al., 2013).
Similarly, mice lacking one copy of Twist1, or compound heterozygous for Twist1 and Tcf12,
display loss of just the coronal suture (Sharma et al., 2013). Analysis of mouse models has led to
an appreciation of proper cell migration (Yoshida et al., 2008; Ting et al., 2009; Roybal et al.,
2010; Deckelbaum et al., 2012), segregation of osteogenic and non-osteogenic cells at the sutural
boundary (Merrill et al., 2006; Ting et al., 2009; Yen et al., 2010), and maintenance of postnatal
stem cells (Zhao et al., 2015) in suture patency. However, it remains unknown the extent to
which defects in early progenitors versus postnatal stem cells account for suture loss. As fish are
amenable to repeated live imaging of calvarial bone growth and suture development outside the
mother, this model provides an opportunity to correlate early defects in osteoprogenitors and
bone growth with later suture loss in individual mutants, particularly in cases where the
synostosis phenotype is variably penetrant. However, a potential complication is that the coronal
suture of zebrafish (Kague et al., 2012; Mongera et al., 2013), as well as those of amphibians
(Piekarski et al., 2014) and birds (Matsuoka et al., 2005), forms at a mesoderm/mesoderm
boundary, contrasting with the mammalian coronal suture that forms at a unique interface
25

between the neural-crest-derived frontal bone and the mesoderm-derived parietal bone (Ishii,
2015) (Figure 2-1A). Mammalian sutures may therefore not be considered evolutionarily
homologous to the sutures of these other vertebrates, at least from a strict embryological
perspective. Should the neural-crest/mesoderm boundary be a factor in coronal sensitivity, non-
mammalian coronal sutures might not be susceptible to the same genetic perturbations that cause
coronal-specific synostosis in mammals.
I report in this study the generation of a zebrafish model of Saethre-Chotzen syndrome that
faithfully recapitulates the craniosynostosis phenotype seen in mice and humans with
heterozygous mutations in TCF12 and TWIST1. The similarity in the genetic interaction between
Twist1 and Tcf12 in mice, humans, and fish, despite differences in the cell lineages that give rise
to the bones, suggests that the underlying processes of coronal suture development and
craniosynostosis are deeply conserved. I demonstrate that in tcf12; twist1b mutant fish, the
frontal and parietal bones grow abnormally. In mutants, skull bones initiate normally, yet early
bone growth is accelerated across the skull. However, subsequent bone growth selectively stalls
at the future coronal suture that is destined to fuse. Moreover, sequential live imaging of
individual mutant fish shows that the degree of later bone stalling predicts which animals will
lose the coronal suture. A similar misregulation of bone growth in Tcf12
+/-
; Twist1
+/-
mutant
mice is observed, with tissue-specific removal of Twist1 resulting in selective overgrowth of the
frontal or parietal bones. Further, Twist1 function must be perturbed within both neural crest- and
mesoderm-derived bones, and not just the mesoderm-derived postnatal sutural mesenchyme, to
prevent suture formation. At least in mice, these altered bone growth dynamics may be due to
changes in osteoblast proliferation and eventual depletion of Gli1
+
and Gremlin1
+
putative bone
progenitors prior to the fusion of the bones. These findings demonstrate that Tcf12 and Twist1
have a conserved early function during skull bone growth to regulate the sustained production of
osteoblasts, possibly through maintenance of osteoprogenitors, with the coronal suture being the
most sensitive to defects in this process.

26

Figure 2-1. Coronal suture loss in tcf12; twist1b mutant zebrafish

(A) Diagrams of zebrafish, mouse, and human skulls, with neural crest contributions in
turquoise and mesoderm contributions in gold. The coronal suture is at a mesoderm-mesoderm
boundary in zebrafish and a neural-crest-mesoderm boundary in mouse and human. Instead of a
suture, an epiphyseal bar cartilage (eb) is present at the neural-crest-mesoderm boundary in
zebrafish. ms, metopic suture; ss, sagittal suture.
27

(B) Dissected skullcaps of adult fish stained with Alizarin Red S show loss of the coronal suture
(asterisks) in tcf12
-/-
; twist1b
-/-
double mutants but not single mutants. Scale bar, 1 mm.
(C) Micro-CT scans of adult fish heads show unilateral (left) and bilateral (right) coronal suture
loss in tcf12
-/-
; twist1b
-/-
mutants. Shading indicates bones derived from neural crest (turquoise)
and mesoderm (gold). Panels below are digital sections through the coronal sutures indicated by
the dotted lines above. Arrowhead indicates the wild-type suture.
(D) Hematoxylin and eosin stained sections show loss of the coronal suture mesenchyme
(arrowhead) in tcf12
-/-
; twist1b
-/-
mutants. Scale bar, 100 µm.

2.2 Results
2.2.1 Specific loss of the coronal suture in tcf12; twist1b mutant zebrafish
In order to investigate requirements for Tcf12 and Twist1 homologs in zebrafish suture
formation, I designed TALE nucleases to generate mutant alleles for tcf12 and both zebrafish
Twist1 homologs, twist1a and twist1b. The tcf12
el548
, twist1a
el570
, and twist1b
el571
alleles result in
premature truncations before the helix-loop-helix domains required for DNA-binding and
dimerization, thus likely abrogating all protein function (Figure supplement 2-1). Whereas
individual homozygous mutants displayed no gross defects as embryos or adults and had patent
sutures across the head (Figure 2-1B and Figure supplement 2-2), 38% of tcf12; twist1b double
mutant adults developed unilateral or bilateral coronal synostosis, as revealed by Alizarin Red
staining of bone (Figure 2-1B). I confirmed loss of the coronal suture by micro-computed
tomography scans and histology (Figure 2-1C,D). As in mice and humans, coronal suture loss
correlated with reduced anterior-posterior growth of the frontal and parietal bones, and in cases
where the suture was lost unilaterally we consistently observed reduced anterior-posterior growth
of that side of the skull (Figure 2-1C). In contrast, I did not detect a requirement for twist1a in
suture development; sutures formed normally in tcf12
-/-
; twist1a
-/-
mutants, and loss of twist1a
alleles did not increase the severity or penetrance of suture defects in tcf12
-/-
; twist1b
-/-
fish
(Figure supplement 2-3, Table supplement 2-1). Loss of tcf12 appears essential for synostosis, as
rare adult viable twist1a
-/-
; twist1b
-/-
fish had normal sutures. While I only detected fusions of the
coronal suture in tcf12
-/-
; twist1b
-/-
fish, I did observe other abnormalities in skull bones in
different mutant combinations, including ectopic sutures and small gaps between the parietal
28

bones (Figure supplement 2-3, Table supplement 2-1). My results demonstrate that, as in humans
and mice with reduced TCF12 and/or TWIST1 dosage, mutations in the homologous genes in
zebrafish primarily result in loss of the coronal suture, although other calvarial defects are also
rarely observed.
Figure supplement 2-1. Zebrafish TALEN mutants

The sites of nucleotide changes induced by TALEN cleavage are shown for each mutant allele.
Schematics show the predicted protein truncations caused by the frame-shift mutations, relative
to the DNA-binding basic helix-loop-helix domain (HLH, purple).

Figure supplement 2-2. Patent sutures in tcf12 single mutants

Skull bones of an adult tcf12
-/-
fish were stained with Alizarin Red. Scale bar, 1 mm. Sections of
the coronal suture (cs) and sagittal suture (ss) were stained with hematoxylin and eosin to show
presence of suture mesenchyme. Scale bar, 100 µm.

29

Figure supplement 2-3. Zebrafish skull phenotypes in different mutant combinations

Dorsal views of adult zebrafish skull bones stained with Alizarin Red. We observed coronal
suture loss (asterisks) in tcf12
-/-
; twist1b
-/-
mutants of any twist1a genotype (D-F). Animals
lacking both Twist1 homologs occasionally develop ectopic sutures (arrows) in the anterior
frontal bone region (B,C, enlarged in F). We also occasionally observed gaps (arrowhead, D) and
ectopic sutures (arrows, enlarged in G) in the posterior region of the parietal bone. Scale bars, 1
mm.

30

Table supplement 2-1. Summary of phenotypes observed in combinatorial zebrafish
mutants
Genotype N
Coronal
Synostosis
Coronal
Synostosis
Index
Sagittal
Gaps
Ectopic
Suture -
Metopic
Ectopic
Suture -
Sagittal
12
-/-
; 1a
+/+
; 1b
+/+
10 0% 0 10% (1/10) 0% 10% (1/10)
12
+/-
; 1a
-/-
; 1b
+/-
9 0% 0 0% 0% 22% (2/9)
12
+/+
; 1a
-/-
; 1b
-/-
1 0% 0 0% 100% (1/1) 0%
12
+/-
; 1a
-/-
; 1b
-/-
7 0% 0 0% 43% (3/7) 0%
12
-/-
; 1a
-/-
; 1b
-/-
3 33% (1/3) 0.667 0% 33% (1/3) 0%
12
-/-
; 1a
+/-
; 1b
-/-
22 41% (9/22) 1 9% (2/22) 0% 0%
12
-/-
; 1a
+/+
; 1b
-/-
13 38% (5/13) 1.077 15% (2/13) 0% 0%
Abbreviations: 12, tcf12; 1a, twist1a; 1b, twist1b.

2.2.2 Loss of tcf12 partially suppresses the embryonic defects and lethality
of Twist1 deficiency
Given the synergistic effect of tcf12 and twist1b loss on coronal suture formation, I examined
whether tcf12 also interacts genetically with Twist1 genes in earlier craniofacial development.
Similar to previous reports of zebrafish with antisense morpholino reduction of twist1a and
twist1b (Das and Crump, 2012), and mice with neural-crest-specific deletion of Twist1 (Bildsoe
et al., 2009), twist1a
-/-
; twist1b
-/-
zebrafish embryos displayed defects in the specification of
skeletogenic ectomesenchyme from the neural crest. In wild-type embryos, neural crest
expression of sox10 is down-regulated by 20 hours post-fertilization (hpf) as ectomesenchyme
neural crest cells populate the pharyngeal arches. In twist1a
-/-
; twist1b
-/-
embryos, we observed
persistent sox10 expression in arch ectomesenchyme and reductions in facial cartilage and bone
at 5 days post-fertilization (dpf) (Figure 2-2A,B). Dorsal facial cartilages (e.g. palatoquadrate
and hyosymplectic) were most affected (Figure 2-2C), potentially reflecting the greater
sensitivity of these elements to general neural crest defects (Cox et al., 2012) and/or post-
migratory roles of Twist1 genes in branchial arch development (Askary et al., 2017).
Interestingly, loss of tcf12 suppressed rather than enhanced the severity of facial skeletal defects
in twist1a
-/-
; twist1b
-/-
larvae, with partial suppression seen with loss of just one tcf12 allele
(Figure 2-2B,D). The suppression did not appear to be due to a rescue of ectomesenchyme
specification, as similarly persistent sox10 was evident in twist1a
-/-
; twist1b
-/-
embryos with or
without tcf12 loss (Figure 2-2A). I also observed that loss of at least one copy of tcf12 enhanced
31

adult viability of twist1a
-/-
; twist1b
-/-
mutants (Figure 2-2E). These findings indicate temporally
distinct genetic interactions between Twist1 and Tcf12, with Tcf12 acting antagonistically to
Twist1 during arch development and synergistically during later skull bone growth and suture
formation.
Figure 2-2. Loss of tcf12 partially rescues twist1a; twist1b jaw cartilage defects and viability

(A) In situ hybridizations at 20 hpf show abnormal persistence of sox10 expression in arch
ectomesenchyme (boxed region) in twist1a
-/-
; twist1b
-/-
and tcf12
-/-
; twist1a
-/-
; twist1b
-/-
mutants.
Arrowheads indicate persistent sox10 expression in arches.
(B) Unilateral dissections of the first and second arch skeletons stained with Alcian Blue
(cartilage) and Alizarin Red (bone) at 5 dpf. Compared to the reductions of the upper facial
skeleton in twist1a
-/-
; twist1b
-/-
mutants, tcf12
-/-
; twist1a
-/-
; twist1b
-/-
triple mutants display less
severe defects. Scale bars, 250 µm.
32

(C) Quantitation of wild-type and twist1a
-/-
; twist1b
-/-
jaw cartilage areas show specific
reductions in more dorsal cartilages, the palatoquadrate (Pq) and hyosympletic (Hs). M,
Meckel’s cartilage; Ch, ceratohyal.
(D) Qualitative scoring of facial skeletal defects from Grade 0 (unaffected) to Grade 4 (most
affected). Loss of one or two copies of tcf12 improved the facial skeletal morphology of twist1a
-
/-
; twist1b
-/-
mutants. Wild type (wt, n = 20), twist1a
-/-
; twist1b
-/-
(12
+/+
, 1a
-/-
, 1b
-/-
, n = 25),
tcf12
+/-
; twist1a
-/-
; twist1b
-/-
(12
+/-
, 1a
-/-
, 1b
-/-
, n = 32), tcf12
-/-
; twist1a
-/-
; twist1b
-/-
(12
-/-
, 1a
-/-
, 1b
-/-
,
n = 22). Using a Fisher’s Exact Test, p = 0.0032 for 12
+/+
, 1a
-/-
, 1b
-/-
versus 12
-/-
, 1a
-/-
, 1b
-/-
and p
= 0.001 for 12
+/+
, 1a
-/-
, 1b
-/-
versus 12
+/-
, 1a
-/-
, 1b
-/-
.
(E) Reduction of tcf12 dosage improves adult viability of twist1a
-/-
; twist1b
-/-
mutants. From an
incross of tcf12
+/-
; twist1a
+/-
; twist1b
+/-
fish, we obtained twist1a
-/-
; twist1b
-/-
mutants and
assessed their viability to 3 months of age. After genotyping for tcf12, we observed a 4:36:12
ratio of tcf12
+/+
: tcf12
+/-
: tcf12
-/-
, compared to the predicted 13:26:13 ratio, which was
significantly skewed as determined by a Chi-squared test (p = 0.0062).

2.2.3 Expression of tcf12 and twist1b at multiple sutures in zebrafish
Given the selective loss of the coronal suture in mutants, I examined whether tcf12 and twist1b
genes might be selectively expressed in this suture. However, at a stage when sutures have just
formed (14 mm standard length), I observed expression of tcf12 and twist1b within the
mesenchyme of not only the coronal but also the metopic and sagittal sutures (Figure 2-3).
Whereas twist1b expression was largely restricted to the suture mesenchyme, tcf12 expression
was observed more broadly in the suture mesenchyme and cells surrounding the skull bones.
Expression of tcf12 and twist1b at multiple sutures in fish is consistent with similar broad suture
expression of Twist1 in mice (Rice, D. P. et al., 2000) and argues against genetic sensitivity of
the coronal suture being due to selective expression of tcf12 and twist1b at this suture.

33

Figure 2-3. Expression of tcf12 and twist1b in multiple sutures of zebrafish

(A) Schematic of the zebrafish skull depicting positions of sections (dotted lines) used for
RNAscope in situ hybridizations. cs, coronal suture; ms, metopic suture; ss, sagittal suture; fb,
frontal bone; pb, parietal bone.
(B-D) In situ hybridizations on sections taken from zebrafish at 14 mm standard length. Red
puncta indicate positive expression. DapB (B) was included as a negative control, with suture
mesenchyme outlined in a dashed white line for reference. Expression of tcf12 (C) and twist1b
(D) was detected in the metopic, coronal and sagittal suture mesenchyme, with tcf12 also
expressed broadly outside the sutures. n = 3 for each experiment. Scale bar, 50 µm.

2.2.4 Altered calvarial bone growth prefigures coronal suture loss in tcf12;
twist1b mutant fish
I next investigated whether suture defects in tcf12; twist1b mutants might result from an earlier
misregulation of skull bone growth, to which the coronal suture might be particularly sensitive.
In wild-type zebrafish, the anlage of the frontal and parietal bones can first be seen at 6 mm
standard length, with the coronal suture forming at the interface of these bones in a lateral to
34

medial progression (Kague et al., 2016). Live mineralization stains revealed that initiation of the
frontal and parietal bones was unaffected in mutants, yet accelerated growth of these bones
became detectable in mutants by 8 mm, and more so by 9 mm (Figure 2-4A). Staining with the
mineralization dye Calcein Green revealed accelerated frontal bone fronts by 10.25±0.25 mm in
mutants, with no difference in the degree of increased bone between sides that developed
synostosis and those that did not (Figure 2-4B,C). Although the overall area of the mutant
parietal bone was comparable to that of wild types at 10.25±0.25 mm, both the mutant frontal
and parietal bones were aberrantly shaped. In particular, both the frontal and parietal bones
exhibited enhanced growth along an axis diagonal to the anterior-posterior and medial-lateral
axes, which themselves exhibit little to no changes in directional growth (Figure 2-4C). Such
enhanced diagonal growth would be expected to bring the frontal and parietal bones closer
together at the forming coronal suture, particularly in the medial region most commonly affected
in mutants (Figure 2-4B). Likewise, the lack of enhanced medial-lateral growth of the parietal
bones correlates with the sagittal suture being unaffected in mutants.
Subsequent sequential staining with Alizarin Red unveiled a marked decrease in bone growth in
mutants from 10.25±0.25 to 14±0.5 mm at the future coronal but not the metopic or sagittal
sutures (Figure 2-4D). We further took advantage of the variable penetrance of suture defects to
correlate the degree of bone stalling with later suture loss in individual tcf12; twist1b mutants.
Consistently, synostotic sides exhibited greater reductions in bone growth compared to mutant
sides that had patent sutures. Thus, the degree to which the growth of the mutant parietal and
frontal bones slows in the coronal zone predicts later loss of this suture. Reciprocally, the
absence of bone stalling at the future metopic and sagittal zones is consistent with these sutures
being unaffected in mutants.

35

Figure 2-4. Altered bone growth dynamics precede craniosynostosis in mutant zebrafish

(A) Dorsal views of the developing skull bones in the same wild-type and mutant individuals
across four developmental stages. Live fish were stained with Calcein Green at 6, 7, and 8 mm
and Alizarin Red at 9 mm. For the right sides, arrows show initiation of the frontal bone at 6 mm
and dashed lines show the frontal (left) and parietal (right) fronts at successive later stages.
(B) Individual wild-type and tcf12; twist1b mutant fish were stained with Calcein Green at
10.25±0.25 mm, recovered, and then stained again with Alizarin Red and imaged at 14±0.5 mm.
These same fish were then grown to 20 mm, at which stage they were fixed and stained again
36

with Alizarin Red to assess suture patency. White dotted lines indicate bone generated by 10.25
mm. Arrowheads indicate missing coronal suture. Scale bars, 1 mm.
(C) Quantification of calvarial bone growth. Bone produced by 10.25±0.25 mm was calculated
based on the area (µm) stained with Calcein Green (white outlines in B). At 10.25 mm,
compared to control frontal bones (n = 12), tcf12; twist1b mutant frontal bones that developed
synostosis later (n = 11) and those that did not (n = 11) showed similar increases in bone
formation. Bone shape was assessed by overlaying tracings of posterior frontal bones and
parietal bones for wild types and mutants. Specific growth directionality in the anterior-posterior,
medial-lateral, and diagonal axes were measured and quantified (green arrows).
(D) Bone growth from 10.25±0.25 mm to 14±0.5 mm was analyzed in respect to prospective
suture zones. Growth in the metopic (yellow) and sagittal (blue) zones did not differ significantly
in controls versus tcf12; twist1b mutants, which correlated with no defects in these sutures in
mutants. In contrast, growth in the coronal zone was reduced in tcf12; twist1b mutants, with a
more pronounced decrease in mutant sides that later developed synostosis. p values were
determined by Student’s t-tests; error bars represent standard error of the mean.

2.2.5 Altered dynamics of bone front cells in tcf12; twist1b mutant fish
I next examined the cellular mechanisms underlying altered bone growth in mutants. Analysis of
an sp7:EGFP transgenic line, which labels Sp7
+
osteoblasts, revealed accelerated frontal and
parietal bone fronts along the diagonal axis in tcf12; twist1b mutants versus sibling controls at 10
mm (Figure supplement 2-4), consistent with our analysis using mineralization dyes. I then used
BrdU incorporation in combination with anti-Sp7 antibody staining to assess the numbers of
proliferative cells at the growing fronts of the frontal and parietal bones at an earlier 9 mm stage.
Analysis of Sp7
+
osteoblasts revealed acceleration of frontal and parietal bone growth, with the
leading edges of the bones appearing uneven (Figure 2-5A; additional examples in Figure
supplement 2-5). After digitally extracting the bone fronts to avoid signals from the highly
proliferative skin, I quantified the numbers of proliferative cells at the fronts. Along both the
mutant frontal and parietal bone fronts, I observed an increase in the percentage of Sp7
+

osteoblasts undergoing proliferation, and a trend toward increased numbers of proliferative Sp7
-

cells just ahead of the bone fronts (Figure 2-5B,C). These results suggest that increased
37

proliferation of early osteoblasts, and potentially also osteoprogenitors, contribute to the initial
acceleration of bone growth across the mutant skull.
Figure supplement 2-4. Accelerated bone fronts in tcf12; twist1b mutants transgenic for
sp7:EGFP

Live imaging of sp7:EGFP was performed on wild-type sibling controls (n = 11) and tcf12;
twist1b mutants (n = 7) at 10 mm standard length. Dorsal views of the skull bones showed
accelerated and dysmorphic bone fronts in all mutants examined. To the right, bone fronts of
individual fish were traced and overlaid according to genotype. Scale bar, 250 µm.

38

Figure 2-5. Altered proliferation and osteoblast production at mutant zebrafish bone fronts

(A) Dissected skullcaps were stained for BrdU (magenta) and Sp7 protein (green) at 9 mm. Top
panels show maximum intensity projections of whole skull volumes, and middle panels are the
same volumes but processed to extract the bone fronts (note that much of the BrdU staining in
the center of the top images is in the skin). Bottom panels show enlarged regions of the
osteogenic fronts (dotted rectangles). White arrows show proliferative osteoblasts (BrdU
+
/Sp7
+
)
and magenta arrows show adjacent proliferative Sp7
-
cells. fb, frontal bone; pb, parietal bone.
Scale bars, 300 µm for whole skull view, 100 µm for enlarged view.
(B, C) Based on the extracted osteogenic fronts (middle panels in A), we quantified the
percentage of Sp7
+
osteoblasts that were BrdU
+
(B) and the number of adjacent BrdU
+
/Sp7
-
cells
per area (C). Wild-type controls, n = 20; tcf12; twist1b mutants, n = 28. p values were
determined by a Student’s t-test; error bars represent standard error of the mean.

39

Figure supplement 2-5. Additional examples of BrdU and Sp7 staining in tcf12; twist1b
mutant zebrafish

Dorsal views of 9 mm zebrafish skullcaps stained with Sp7 antibody (green) and BrdU
(magenta) show an array of altered bone shapes in mutants compared to wild type. fb, frontal
bone; pb, parietal bone.

2.2.6 Altered calvarial bone growth precedes coronal suture loss in Tcf12
+/-
;
Twist1
+/-
mice
Given the similarity of coronal suture defects in tcf12
-/-
; twist1b
-/-
fish and Tcf12
+/-
; Twist1
+/-

mice, I worked with Dr. Man-Chun Ting, a postdoctoral scholar in Dr. Robert Maxson’s
laboratory, to investigate whether earlier alterations in calvarial bone growth might also
prefigure suture loss in mice. At embryonic day (E) 13.5, when skull bone rudiments are first
apparent, alkaline phosphatase staining revealed accelerated frontal and parietal bones that were
40

in closer apposition in Tcf12
+/-
; Twist1
+/-
mice versus wild-type sibling controls (Figure 2-6A).
At birth, mutant frontal and parietal bones were more closely apposed than in controls and had
abnormal shapes (Figure 2-6B). Next, we assessed proliferation rates and osteoblast density at
the forming coronal suture (E14.5) and sagittal suture (E16.5) (Figure 2-6C, outlined regions of
bottom panels). In mutants, we observed thicker bone fronts, as marked by Sp7
+
cells (Figure
supplement 2-6), as well as a marked increase in the number of Sp7
+
osteoblasts at the bones
fronts predestined for the coronal and sagittal sutures (Figure 2-6D). We also observed an
increase in the number of proliferative Sp7
-
cells immediately adjacent to osteoblasts at the bone
fronts, with this increase more evident at the forming coronal versus sagittal suture (Figure 2-
6E). In contrast to fish, we noted a decrease in the number of proliferative osteoblasts at the
forming coronal suture, and no change at the forming sagittal suture (Figure 2-6F). These
findings largely support a conserved role for Tcf12 and Twist1 in negatively regulating the
number of proliferative cells at the growing bone fronts in fish and mice.

41

Figure 2-6. Altered bone growth dynamics in Tcf12
+/-
; Twist1
+/-
mice

(A) Lateral views of E13.5 mouse heads show Alkaline phosphatase staining of developing
frontal and parietal bones. Dotted lines indicate the bone fronts that will form the coronal suture.
Compared to wild types (n = 5), the fronts were accelerated in all Tcf12
+/-
; Twist1
+/-
mutants (n =
12). Scale bar, 1 mm.
(B) Dorsal views of skull bones stained with Alizarin Red at birth (P0). Compared to wild types
(n = 5), the fronts were closer together in all Tcf12
+/-
; Twist1
+/-
mutants (n = 11). Scale bar, 1
mm.
(C) Sections of E14.5 coronal sutures and E16.5 sagittal sutures stained for BrdU (magenta), Sp7
protein (green), and DAPI (blue, nuclei). Boxed regions are magnified in lower panels, with
42

yellow dotted lines indicating the regions of interest (ROI) used for quantification. fb, frontal
bone; pb, parietal bone. Scale bar, 100 µm.
(D-F) Quantification of Sp7
+
osteoblasts per ROI (D), BrdU
+
Sp7
-
bone front cells per ROI (E),
and the percentage of Sp7
+
osteoblasts that are BrdU
+
in the ROI (F). Cell counts were
performed at the developing coronal sutures (cs, 4 wild types, 3 mutants) and sagittal sutures (ss,
6 wild types, 6 mutants). p values were determined by Student’s t-tests; error bars represent
standard error of the mean.

Figure supplement 2-6. Increased calvarial bone front thickness in mutants

(A) Sections of mouse E14.5 coronal sutures and E16.5 sagittal sutures stained for Sp7 protein
(green) and DAPI (blue, nuclei) were used to assess bone thickness. Yellow lines indicate
thickness measurements used for quantification. fb, frontal bone; pb, parietal bone. Scale bar,
100 µm.
(B) Bone front thickness of the coronal (4 wild types, 3 mutants) and sagittal sutures (6 wild
types, 6 mutants) were quantified. p values were determined by Student’s t-tests; error bars
represent standard error of the mean.
43

2.2.7 Selective reduction of the osteoprogenitor pool at the mutant coronal
suture
I next investigated whether the lack of continued bone growth at the mutant coronal fronts might
reflect an exhaustion of osteoprogenitor cells. In mouse, sutural stem cells express Prrx1 (Wilk
et al., 2017) and Gli1 (Zhao et al., 2015). Grem1 also marks skeletal stem cells throughout the
animal (Worthley et al., 2015), yet a role in the skull and sutures has not been previously
examined. Using RNAscope in situ hybridization technology in zebrafish and with the help of
Dr. D’Juan Farmer, a postdoctoral scholar in Dr. Gage Crump’s laboratory, I find that prrx1a is
broadly expressed at the parietal and frontal bone fronts destined for the coronal and sagittal
sutures, as well as the periosteum, at 10 mm, and in the sutures and periosteum at adult stages
(Figure 2-7A,B). The expression of gli1 and grem1a appears more restricted to the growing
bone fronts and suture mesenchyme, although we also see more general periosteal expression at
earlier stages. In tcf12; twist1b mutants, we still observe cells expressing gli1, grem1a, and
prrx1a at the forming coronal and sagittal sutures, as well as within the periosetum (Figure 2-
7A,B), with quantitation revealing similar levels of each gene on a per cell basis (Figure
supplement 2-7A). At the coronal suture, we observed that the mutant frontal and parietal bones
were more closely apposed than in stage-matched wild-type siblings, which could possibly be a
consequence of depleted progenitors at the bone fronts. In adult mutants with fused coronal
sutures, we failed to detect gli1
+
, grem1a
+
, or prrx1a
+
cells in the coronal suture region, whereas
prrx1a was still expressed in the periosteum (Figure 2-7A). Normal expression of all three
markers was observed at the patent sagittal suture. We also examined Fgf signaling in zebrafish
mutants, as the expression of Fgfr2 has been reported to be altered in mouse Twist1 heterozygous
animals (Rice, D. P. C. et al., 2000; Connerney et al., 2006). As with gli1, grem1a, and prrx1a,
we still observed cells expressing fgfr2 and the Fgf target gene dusp6 at the mutant coronal
suture (Figure supplement 2-7B), arguing against the calvarial phenotypes being due to
wholesale loss of Fgfr2 signaling at the bone fronts.
Given the difficulty in quantitating the numbers of osteoprogenitors at the forming coronal suture
zone of mutant zebrafish, owing in part to the small sizes of these bone fronts, we along with Dr.
Ting also examined putative progenitors in mutant mice. As in zebrafish, we observed cells
expressing Gli1 and Grem1 protein at and around the fronts of the embryonic frontal and parietal
bones. In Tcf12
+/-
; Twist1
+/-
mice, we observed a marked reduction in the number of Gli1
+
and
44

Grem1
+
putative progenitors at and around the developing coronal but not the sagittal bone fronts
(Figure 2-7C,D). These findings highlight a conserved molecular signature of putative
osteoprogenitors and sutural stem cells of zebrafish and mice and suggest, at least in mice, a
selective exhaustion of osteoprogenitors at the developing coronal suture.
Figure 2-7. Reduced osteoprogenitor pool at the mutant coronal suture

(A, B) Sections of forming coronal and sagittal sutures of 10 mm fish and fully formed sutures of
adult fish were assessed for gli1, grem1a, and prrx1a mRNA expression (white) by RNAscope in
situ hybridization. Orange dotted lines indicate bones, and the boxed regions of the coronal
45

suture regions are magnified below. For adult sutures, yellow arrowheads show expression in the
suture mesenchyme, yellow arrows show expression of prrx1a in the periosteum, and grew
arrows show lack of expression of gli1 and grem1a in the periosteum. Scale bar at 10 mm stage,
20 µm; scale bar at adult stage, 50 µm.
(C, D) Sections of E14.5 coronal sutures and E16.5 forming sagittal sutures stained for
Gli1/Grem1 (magenta) and Sp7 protein (green). Gray arrowheads indicate progenitor region in
forming coronal sutures. Boxed regions of parietal bone fronts in the forming sagittal sutures are
magnified in lower panels. Nuclei are stained blue by DAPI in all images. Scale bars, 100 µm.

Figure supplement 2-7. Quantification of progenitor marker and Fgf pathway transcripts
in tcf12; twist1b mutants

(A) The number of gli1, grem1a, and prrx1a transcripts at the forming coronal and sagittal suture
regions were quantified per region area (as indicated in magenta in Figure 2-7A). Although the
progenitor regions expressing these genes were reduced (Figure 2-7B), no significant difference
in transcript levels were observed in equivalent areas of remaining progenitors between wild
types (n = 4) and tcf12; twist1b mutants (n = 4, with the exception of n = 2 mutants for prrx1a
46

transcripts). p values were determined by Student’s t-tests; error bars represent standard error of
the mean.
(B) RNAscope in situ hybridizations show fgfr2 and dusp6 transcripts (white dots) relative to all
nuclei (DAPI, blue) at the forming coronal sutures of 10 mm standard length fish. Orange dotted
lines indicate bones and magenta dotted lines the forming suture region in which fgfr2 and dusp6
transcripts were quantified. Although coronal progenitor regions expressing fgfr2 and dusp6
were reduced, these genes were expressed at similar levels per area of progenitor regions in wild
types (n = 5) and tcf12; twist1b mutants (n = 5). p values were determined by Student’s t-tests;
error bars represent standard error of the mean. Scale bar, 20 µm.

2.2.8 Tissue-specific roles for Twist1 in calvarial bone growth and suture
patency
To investigate whether Twist1 functions tissue-intrinsically for proper skull bone growth, the
mouse model is suitable with the unique germ-layer origins of the frontal and parietal bones. Dr.
Mia Brockop, a previous student of Dr. Robert Maxson, had removed one copy of Twist1 in each
tissue. Working with Dr. Ting again, we found that reduced dosage of Twist1 in the neural-crest-
derived precursors of the frontal bone, in Wnt1-Cre; Twist1
f/+
mice at postnatal day (P) 21,
resulted in the overgrowth of the frontal bone relative to the parietal bone, which we quantified
by measuring the ratio of the sagittal to metopic suture (Figure 2-8). Reciprocally, removing one
copy of Twist1 from the mesoderm-derived parietal bone, in Mesp1-Cre; Twist1
f/+
mice, resulted
in its overgrowth relative to the frontal bone. Reduced dosage of Twist1 in both the neural crest
and mesoderm, in Wnt1-Cre; Mesp1-Cre; Twist1
flox/+
mice, normalized the relative sizes of the
frontal and parietal bones and resulted in loss of the coronal suture, a phenotype not seen upon
deletion of Twist1 in neural crest or mesoderm alone. We conclude that Twist1 negatively
regulates bone growth in both the neural-crest- and mesoderm-derived portions of the skull, and
that Twist1 must be mutated in not only the mesoderm-derived parietal bone and suture
mesenchyme, but also the neural-crest-derived frontal bone, to impact coronal suture formation.

47

Figure 2-8. Tissue-autonomous bone overgrowth in Twist1 conditional mutants

(A) Dorsal views of Alizarin-stained skulls of three-week-old mice. In the accompanying
diagrams, turquoise indicates the neural-crest-derived frontal bones and gold the mesoderm-
derived parietal bones. The relative lengths of the metopic suture (ms) and sagittal suture (ss)
serve as a proxy for bone size. Compared to wild type (n = 0/10), Wnt1-Cre; Twist1
flox/+
(n =
0/8), and Mesp1-Cre; Twist1
flox/+
(n = 0/4); Wnt1-Cre; Mesp1-Cre; Twist1
flox/+
mice (n = 2/3)
displayed coronal synostosis (arrowhead, average craniosynostosis index of 2.33). Scale bar, 1
mm.
(B) Quantification of the relative length of the sagittal over the metopic suture. p values were
determined by a one-way ANOVA with post-hoc Tukey-Kramer HSD test; error bars represent
standard error of the mean.

48

2.3 Discussion
Discovery of a selective requirement for tcf12 and twist1b in coronal suture formation in
zebrafish has allowed us to gain a better understanding of the developmental basis of suture loss
in Saethre-Chotzen syndrome. The similarity of coronal suture defects from humans to mice to
zebrafish is striking, although each species displays unique dosage sensitivities to loss of
TWIST1 and TCF12. In humans, haploinsufficiency of TWIST1 or TCF12 can lead to suture loss.
In mice, haploinsufficiency of Twist1 also results in coronal suture loss, yet haploinsufficiency of
Tcf12 does not, despite enhancing the penetrance of suture defects in Twist1
+/-
mice (Sharma et
al., 2013). In zebrafish, only loss of one of two Twist1 homologs (twist1b) along with loss of
tcf12 results in suture loss. It remains unclear why humans are more sensitive to Twist1 and
Tcf12 dosage than mice and zebrafish, although a similar phenomenon has been observed with
other synostosis genes (e.g. JAG1) (Teng et al., 2017).
Repeated live imaging of individual mutant fish revealed a strong correlation between the extent
of altered bone growth and later suture loss. Whereas the initiation of the frontal and parietal
bones was largely unaffected in mutants, we observed increased proliferation and osteoblast
production at the mutant bone fronts in both fish and mice. There were some subtle differences
between fish and mice, with larger increases in proliferative osteoblasts in mutant fish than mice,
which might reflect species-specific or staging differences. Nonetheless, mutants in both species
displayed accelerated growth of both the frontal and parietal bones, which often resulted in
abnormal shapes likely due to growth variations along individual bone fronts. In particular, we
found that increased diagonal growth in mutants brought the parietal and frontal bones together
at the prospective medial regions of the coronal suture much earlier than in wild type animals,
which correlates with the medial region of the coronal suture being most commonly fused in
mutants. This altered directional growth might be one reason why the coronal suture is
preferentially affected in both zebrafish and mouse Twist1/Tcf12 mutants, despite the different
origins of the coronal suture in these species.
Another prominent finding was a lack of continued bone growth at the future coronal but not
other sutures in mutants, with the degree of bone stalling predicting synostosis in individuals. We
identified Gli1/gli1 and Grem1/grem1a as conserved markers of putative osteoprogenitors in
both the growing bone fronts and mature sutures of mice and zebrafish, although both genes
display broader expression at early stages in calvarial bone development, including in some
49

periosteal cells. Future lineage tracing experiments should help reveal the extent to which gli1 in
fish and Grem1/grem1a in both species mark similar populations of embryonic osteoprogenitors
and postnatal sutural stem cells. A previous study had observed reduced expression of Gli1 in the
sutures of adult Twist1
+/-
mice (Zhao et al., 2015). Here we extend this finding to embryonic
stages when the coronal suture is forming, and uncover the existence of Grem1+ cells in and
around the bone fronts of the nascent coronal suture that become depleted in mutant mice.
Whether similar osteoprogenitors become exhausted at the developing coronal suture of
zebrafish remains to be determined, as it was difficult to precisely quantify the numbers of these
progenitors by RNA expression alone. One possibility is that reduced Twist1 and Tcf12 function
alters the balance between long-term sutural stem cells (i.e. those marked by Gli1) and
proliferative osteoblasts and their immediate progenitors. In such a model, the early increase in
osteoblast production would come at the expense of long-term progenitors, thus leading to a later
failure of continued bone growth and a loss of the sutural stem cells that would normally separate
the skull bones. In addition to further verifying this model, an important next step will be to
determine why the fronts of the parietal and frontal bones abutting the future coronal suture are
most sensitive to progenitor exhaustion in mutants. We did not observe preferential expression of
tcf12 and twist1b at the coronal suture in fish, consistent with similarly broad sutural expression
of Twist1 in mice (Rice, D. P. et al., 2000). Instead, osteoprogenitors in the coronal zone could
be fewer in number at initial stages and/or more sensitive to loss of Tcf12 and Twist1, for
example due to compensation by related genes at other sutures.
Further support for suture defects arising from much earlier changes in bone growth come from
our conditional Twist1 deletion experiments in mouse. Whereas the frontal bone arises from
neural crest and the parietal bone from mesoderm, the mesenchyme within the postnatal coronal
suture is largely mesoderm-derived (Yen et al., 2010). However, suture loss was only observed
upon conditional deletion of one allele of Twist1 from both the embryonic mesoderm and neural
crest. This finding is inconsistent with Twist1 functioning solely in the mesoderm-derived
postnatal suture mesenchyme for suture patency, instead suggesting that misregulated growth of
both the frontal and parietal bones is required to later disrupt the coronal suture in Twist1
mutants.
Our study highlights a selective role for Tcf12 in the later growth of the skull bones and patency
of the coronal suture. In contrast to animals lacking Twist1 homologs, zebrafish lacking tcf12 do
not display embryonic lethality or defects in ectomesenchyme and facial cartilage formation.
50

Instead, loss of tcf12 partially suppresses the facial cartilage defects and lethality of twist1a;
twist1b mutants. Suppression of embryonic defects could be due to loss of competition of Tcf12
with other Twist binding partners, such as Hand2 (Firulli et al., 2005). In this scenario, maternal
Twist1a/b and/or other Twist family members (e.g. twist2 and twist3) could compensate for lack
of zygotic Twist1a/b. Loss of tcf12 would then allow remaining Twist proteins to more
effectively form homodimers or alternate heterodimers important for embryogenesis. Whereas a
previous report indicates that Tcf12 and Twist1 can form heterodimers (Connerney et al., 2006),
it is also possible that Tcf12 has Twist1-independent functions that antagonize Twist1 during
embryogenesis. Future efforts to directly visualize Tcf12-Twist1 complexes will help to resolve
how Tcf12 promotes Twist1 function during skull bone growth while counteracting it during
embryonic neural crest development.
Development of homologous structures often employs ancestrally conserved gene regulatory
networks. For example, a requirement for Pax6 genes in eye development in a wide range of
animals has been used to argue for deep homology of eye structures (Gehring and Ikeo, 1999).
Here, we show remarkably specific loss of a single anatomical structure, the coronal suture, in
zebrafish and mice lacking Tcf12 and Twist1, despite this suture occurring at a
mesoderm/mesoderm interface in zebrafish and a neural-crest/mesoderm interface in mice.
Hence, sensitivity of the coronal suture in Saethre-Chotzen syndrome is unlikely to be due to its
location at a unique neural-crest/mesoderm interface. In addition, there are several other sutures
occurring at a neural-crest/mesoderm interface in mice and fish that are not affected by loss of
Tcf12 and Twist1. There has been on-going debate, given these distinct tissue boundaries, as to
whether coronal sutures are truly homologous across vertebrates (Maddin et al., 2016). Our data
indicate that the conserved genetic dependence of the coronal suture in fish and mammals likely
reflects a similar sensitivity to early bone growth changes, perhaps owing to similar
developmental and anatomical constraints irrespective of the embryonic origins of the bones
flanking this suture.
2.4 Materials and Methods
2.4.1 Animals
The University of Southern California Institutional Animal Care and Use Committee approved
all animal experiments, and all methods were performed in accordance with the relevant
guidelines and regulations. Zebrafish (Danio rerio) embryos were raised in Embryo Medium
51

(Westerfield, 2007) at 28.5°C. Juvenile and adult fish were housed in groups of 10-15. Mutant
lines were maintained on a mixed Tubingen wild-type (Haffter et al., 1996), casper (White et al.,
2008), and sp7:EGFP background. Three targeted mutant lines (tcf12
el548
, twist1a
el571
, and
twist1b
el570
) were generated for this study (see below). Lines were propagated by genotyping fin
biopsies. As we found that twist1a loss did not effect the penetrance or expressivity of suture loss
in tcf12
-/-
; twist1b
-/-
fish (Table supplement 2-1), we pooled tcf12
-/-
; twist1b
-/-
fish with any
twist1a genotype for the experiments. Mice (Mus musculus) were housed in cages with no more
than five adults or three adults with one litter per cage. The Tcf12 (Wojciechowski et al., 2007),
Twist1 (Chen and Behringer, 1995), conditional Twist1-flox (Bildsoe et al., 2009), Wnt1-
cre (Danielian et al., 1998), and Mesp1-cre (Saga et al., 1999) alleles were genotyped as
described.
2.4.2 Generation of zebrafish mutant lines
Zebrafish mutant for tcf12, twist1a, or twist1b were generated with TALEN-based targeted
mutagenesis (Figure supplement 2-1). TALEN constructs were generated using the PCR-based
platform (Sanjana et al., 2012) and digested with StuI (New England Biolabs, Ipswich, MA).
RNAs were synthesized from linearized constructs using the mMessage mMachine T7 Ultra kit
(Ambion/Life Technologies, Carlsbad, CA, USA). TALEN RNAs were injected at 100 ng/μl into
1-cell-stage embryos, and we identified mosaic germline founders by sequencing their progeny.
The tcf12
el548
allele includes base pair changes and insertions that disrupt a common exon shared
in all protein-coding transcript variants. The twist1a
el571
and twist1b
el570
alleles are deletions that
interrupt the coding sequence close to the translation start site. All three alleles are frame-shift
mutations that result in premature stop codons upstream of the helix-loop-helix DNA-binding
domains. Detailed sequences of TALEN targets and genotyping primers are listed in Table
supplement 2-2.

52
Table supplement 2-2. TALEN targeting and mutant genotyping

Gene TALEN Targets Genotyping Primers
Annealing
Temperature
(ºC)
Restriction
Enzyme
Product sizes
(bp)
tcf12
L: TATGGGGGAATGCTGGGAGG
R: TGTAGTTTCCAGACTGTGGC
F: CCACATCTTCAAAGCTGGAAA
R: TTGATCTGATCCCGCAGAG
58 BsaBI
WT: 145+99
MT: 244
twist1a
L: TGGACAGTCTGGGAAACAGC
R: TGACGCGCTTCGGTTGTCGC
F: TGGACAGTCTGGGAAACAGC
R: GTGGGACTGTCGGAATCCT
58 --
WT: 109
MT: 102
twist1b
L: TCAGCAACAGCGACGGAGAG
R: TCTTTTCCTTGCGCACCTTT
F: GCGGACAGTCTCAGCAACAG
R: TTCTTGCTCGACCGTCTTTT
58 --
WT: 80
MT: 69

53

2.4.3 Skull preparations
Adult zebrafish were fixed overnight at 4°C with 4% paraformaldehyde, washed with 0.5% KOH
for at least half an hour, cleared with 3% H2O2 in 0.5% KOH for several hours until
pigmentation was removed, washed with 35% NaBO4 for at least half an hour, incubated with
1% trypsin in 35% NaBO4 for several hours until tissue was reasonably cleared, washed with
10% glycerol in 0.5% KOH for at least one hour, stained with 0.02% Alizarin Red S (Amresco
9436) pH 7.5 in 10% glycerol and 0.5% KOH overnight, washed with 50% glycerol/0.5% KOH
until residual stain was removed, and stored in 100% glycerol. Skullcaps were dissected and
imaged using a Leica S8 APO stereomicroscope.
The heads of newborn mice were skinned and cleared with 1% KOH for 1 to 2 days, stained with
2% Alizarin Red S in 1% KOH until mineralized bone was red, and stored in 100% glycerol.
Mouse skulls were imaged using a Leica MZ125 stereomicroscope.
2.4.4 Micro-computed tomography
Data was collected on a Nikon Metrology Xt S 225 ST with the following parameters: energy at
120kV; current at 26 uA; 3141 projections at 2 frames/sec and averaging 2 frames; no filter; and
at 6 μm resolution. Two-dimensional slices were rendered into three-dimensional reconstructions
using Arivis (Phoenix, AZ).
2.4.5 Paraffin embedding, sectioning, and tissue histology
Whole zebrafish were embedded into paraffin according to standard protocol. Briefly, fish heads
were fixed with 4% paraformaldehyde at 4°C or 10% neutral buffered formalin at room
temperature overnight, washed with PBS for half an hour each time, and then separated into
heads and trunks. Heads were decalcified with 20% EDTA pH 8.0 for 10 days at room
temperature, washed with DEPC-treated water, dehydrated through a series of sequentially
increasing ethanol to DEPC water ratios, treated through a series of sequentially increasing
Hemo-De (xylene substitute) to ethanol ratios, washed in 50% paraffin in Hemo-De at 65°C for
one hour, incubated at 65°C in 100% paraffin overnight, embedded in paraffin in molds, and
allowed to solidify at room temperature. Paraffin blocks were cut into 5 μm sections using a
54

Shandon Finesse Me+ microtome (cat no. 77500102) and collected on superfrost plus slides
(Thermo Fisher Scientific).
For tissue histology, hematoxylin and eosin staining was performed according to standard
protocol. Briefly, sections were deparaffinized in xylene and ethanol, rinsed in water, stained
with hematoxylin, rinsed in 4% glacial acetic acid, rinsed with water, washed with blueing
solution, rinsed with water, dried with ethanol, stained with eosin, washed with ethanol and
Hemo-De, and mounted with cytoseal.
For whole-mount alkaline phosphatase staining, E13.5 mouse heads were fixed in 4%
paraformaldehyde in PBS, bisected mid-sagitally after fixation, and stained with NBT and BCIP.
Sections were imaged on a Leica DM2500 compound microscope.
2.4.6 In situ hybridization
Colorimetric in situ hybridization using a sox10 digoxigenin-labeled riboprobe was performed on
20 hpf zebrafish embryos as described (Cox et al., 2012). Briefly, embryos were fixed,
dehydrated with an increasing methanol series, and stored in 100% methanol at -20°C until use.
Embryos were then rehydrated with a decreasing methanol series, treated with 1 μg/ml
Proteinase K for 6.6 min, post-fixed with 4% PFA, incubated with hybridization buffer (50%
formamide, 5X SSC, 100 μg/ml yeast RNA, 50 μg/ml heparin, 0.125% Tween-20, citric acid to
pH 6), hybridized with probe, washed with solution series with decreasing formamide and SSC,
incubated in blocking solution (5 mg/ml BSA, 5% sheep serum), treated with sheep digoxigenin-
AP at 1:10,000, stained with NBT and BCIP (Roche), and color was allowed to develop for two
and a half hours. Stained embryos were imaged on a Leica DM2500 compound microscope.
RNAscope in situ hybridization was performed with the RNAscope® 2.5 HD Assay – RED
(Advanced Cell Diagnostics, Newark, CA) or the RNAscope® Multiplex Fluorescent Kit v2
(Advanced Cell Diagnostics, Newark, CA) according to the manufacturer’s protocol for
formalin-fixed paraffin embedded sections. Probes include twist1b (C1), tcf12 (C2), prrx1a (C1),
grem1a (C1), gli1 (C3), fgfr2 (C1), and dusp6 (C3).
2.4.7 Zebrafish larvae skeletal staining
Alcian Blue (cartilage) and Alizarin Red S (bone) staining was performed on 5 dpf zebrafish
larvae as previously described (Walker and Kimmel, 2007). Briefly, larvae were fixed for one
hour in 2% paraformaldehyde, rinsed in 100 mM Tris pH 7.5 in 10 mM MgCl2, incubated
55

overnight in Alcian Blue solution (0.04% Alcian Blue, 80% ethanol, 100 mM Tris pH 7.5, 10
mM MgCl2), rehydrated through a sequentially decreasing series of ethanol to 100 mM Tris pH
7.5 and 10 mM MgCl2 ratio, bleached with 3% H2O2 in 0.5% KOH under a lamp, washed with
25% glycerol in 0.1% KOH, stained with Alizarin Red S (0.01% Alizarin Red, 25% glycerol,
100 mM Tris pH 7.5), and de-stained with 50% glycerol in 0.1% KOH. Dissected jaw cartilages
were mounted in 50% glycerol on a slide and imaged with a Leica DM2500 compound
microscope.
2.4.8 Sequential live staining and imaging
Fish were anesthetized, measured for body length, and recovered in Calcein Green stain (3
mg/30 ml, Molecular Probes C481) overnight in the dark. The following day, fish were washed
in fish system water at least twice for one hour each time before imaging with a Zeiss AxioZoom
microscope and returned to tanks on system. This process was repeated at a later time point with
Alizarin Red S (1 mg/30 ml).
2.4.9 BrdU treatments and immunohistochemistry
Anesthetized fish were measured for body length and incubated in 4.5 mg/ml BrdU solution
(B5002, Sigma Aldrich) for two hours in the dark. Fish were then transferred to fish system
water or embryo medium for 15 minutes before euthanasia. Heads were fixed in 4%
paraformaldehyde overnight before skullcaps were dissected and stored in PBS at 4ºC. For
mouse embryos at E14.5 and E16.5, BrdU was injected into the pregnant female (200 μg/g body
weight) 2 hours prior to dissection. Heads of embryos were embedded in OCT medium
(Histoprep, Fisher Scientific) before sectioning. Frozen sections were cut at 10 µm.
Immunohistochemistry was performed using rat anti-BrdU (MCA2060 GA, Bio-Rad), rabbit
anti-Osx/Sp7 (sc-22536-r, Santa Cruz), goat anti-Grem1 (PA5-47973, Thermo Fisher), or goat
anti-Gli1 (AF3455, R&D Systems) diluted in 1% BSA/PBS and incubated overnight at 4ºC.
Detection of primary antibodies was performed by incubating goat anti-rat FITC (sc-2011, Santa
Cruz), goat anti-rabbit Alexa Fluor 568 (A-11011, Thermo Fisher Scientific), or donkey anti-
goat Alexa Fluor 488 (ab150129, Abcam) for 1 hour at room temperature followed by DAPI
counterstaining.
56

2.4.10 Quantitation and statistical analyses
Severity of coronal synostosis was quantified using a coronal synostosis scoring index that was
adapted from Oram and Gridley’s craniosynostosis index (Oram and Gridley, 2005). Left and
right sutures were given scores of 0 (no fusion), 1 (< 50% fused), 2 (≥ 50% fused), or 3
(completely fused). A composite score for each animal was calculated by adding the sum of both
suture scores. The index value was determined by the sum of composite scores in each genotype
group divided by the number of animals in the group.
The degree of jaw cartilage defect was determined by the number of cartilage elements affected,
with more affected equating to a higher grade of severity. Flat-mounted cartilages were measured
for area using Fiji.
Directional bone growth was quantified using Fiji. For bone produced by 10.25 mm and
subsequent growth by 14 mm, bone regions were drawn freehand and measured for area. For
assessment of BrdU
+
and Sp7
+
cells, tiled z-stacks were captured of stained zebrafish skulls caps
using a Zeiss LSM 800. The bone fronts were then digitally extracted using a 30-pt. brush in
Amira software, and cells were manually counted in a 3D view on Imaris software. In mouse
coronal suture sections, the region of interest was determined by a defined length across the bone
fronts and suture. In mouse sagittal suture sections, defined lengths included several cell
diameters medial from the last Sp7
+
cell at the bone fronts and accounted for bone curvature.
Five sections per animal were quantified and averaged. For bone thickness quantifications, a
perpendicular measurement across the broadest point of Sp7+ cells was measured for each
section. All measurements and counting were completed in Fiji. For the quantification of
RNAscope experiments, regions of interests were defined by the edge of the calvarial bones
(imaged by DIC microscopy) in Fiji. The coronal progenitor region was defined as the space
between the neighboring frontal and parietal bones, and the sagittal progenitor region was
defined as an approximately three-cell layer length region ahead of the parietal bone. RNAscope
signal was quantified across a Z-stack projection using the 3D Object Counter tool, and
measurements above a pre-defined unit of one transcript were adjusted to account for closely
packed transcripts. The final transcript count was normalized to the area of the region of interest.
Three sections per animal were averaged for each probe. Unpaired t-tests were performed for all
statistical analyses, and all samples were scored blindly.
57

2.5 References
Askary, A., Xu, P., Barske, L., Bay, M., Bump, P., Balczerski, B., Bonaguidi, M.A., Crump,
J.G., 2017. Genome-wide analysis of facial skeletal regionalization in zebrafish. Development.
Bildsoe, H., Loebel, D.A., Jones, V.J., Chen, Y.T., Behringer, R.R., Tam, P.P., 2009.
Requirement for Twist1 in frontonasal and skull vault development in the mouse embryo.
Developmental biology 331, 176-188.
Chen, Z.F., Behringer, R.R., 1995. twist is required in head mesenchyme for cranial neural tube
morphogenesis. Genes & development 9, 686-699.
Connerney, J., Andreeva, V., Leshem, Y., Muentener, C., Mercado, M.A., Spicer, D.B., 2006.
Twist1 dimer selection regulates cranial suture patterning and fusion. Developmental dynamics :
an official publication of the American Association of Anatomists 235, 1345-1357.
Cox, S.G., Kim, H., Garnett, A.T., Medeiros, D.M., An, W., Crump, J.G., 2012. An essential role
of variant histone H3.3 for ectomesenchyme potential of the cranial neural crest. PLoS genetics
8, e1002938.
Danielian, P.S., Muccino, D., Rowitch, D.H., Michael, S.K., McMahon, A.P., 1998.
Modification of gene activity in mouse embryos in utero by a tamoxifen-inducible form of Cre
recombinase. Current biology : CB 8, 1323-1326.
Das, A., Crump, J.G., 2012. Bmps and id2a act upstream of Twist1 to restrict ectomesenchyme
potential of the cranial neural crest. PLoS genetics 8, e1002710.
Deckelbaum, R.A., Holmes, G., Zhao, Z., Tong, C., Basilico, C., Loomis, C.A., 2012. Regulation
of cranial morphogenesis and cell fate at the neural crest-mesoderm boundary by engrailed 1.
Development 139, 1346-1358.
el Ghouzzi, V., Le Merrer, M., Perrin-Schmitt, F., Lajeunie, E., Benit, P., Renier, D., Bourgeois,
P., Bolcato-Bellemin, A.L., Munnich, A., Bonaventure, J., 1997. Mutations of the TWIST gene
in the Saethre-Chotzen syndrome. Nature genetics 15, 42-46.
Firulli, B.A., Krawchuk, D., Centonze, V.E., Vargesson, N., Virshup, D.M., Conway, S.J.,
Cserjesi, P., Laufer, E., Firulli, A.B., 2005. Altered Twist1 and Hand2 dimerization is associated
with Saethre-Chotzen syndrome and limb abnormalities. Nature genetics 37, 373-381.
Flaherty, K., Singh, N., Richtsmeier, J.T., 2016. Understanding craniosynostosis as a growth
disorder. Wiley Interdiscip Rev Dev Biol 5, 429-459.
Gehring, W.J., Ikeo, K., 1999. Pax 6: mastering eye morphogenesis and eye evolution. Trends in
genetics : TIG 15, 371-377.
Haffter, P., Granato, M., Brand, M., Mullins, M.C., Hammerschmidt, M., Kane, D.A., Odenthal,
J., van Eeden, F.J., Jiang, Y.J., Heisenberg, C.P., Kelsh, R.N., Furutani-Seiki, M., Vogelsang, E.,
Beuchle, D., Schach, U., Fabian, C., Nusslein-Volhard, C., 1996. The identification of genes
with unique and essential functions in the development of the zebrafish, Danio rerio.
Development 123, 1-36.
Hajihosseini, M.K., 2008. Fibroblast growth factor signaling in cranial suture development and
pathogenesis. Frontiers of oral biology 12, 160-177.
Howard, T.D., Paznekas, W.A., Green, E.D., Chiang, L.C., Ma, N., Ortiz de Luna, R.I., Garcia
Delgado, C., Gonzalez-Ramos, M., Kline, A.D., Jabs, E.W., 1997. Mutations in TWIST, a basic
helix-loop-helix transcription factor, in Saethre-Chotzen syndrome. Nature genetics 15, 36-41.
58

Ishii, M., Sun, J., Ting, M.-C., Maxson, R.E., 2015. The Development of the Calvarial Bones and
Sutures and the Pathophysiology of Craniosynostosis. Current Topics in Developmental Biology.
Kague, E., Gallagher, M., Burke, S., Parsons, M., Franz-Odendaal, T., Fisher, S., 2012.
Skeletogenic fate of zebrafish cranial and trunk neural crest. PloS one 7, e47394.
Kague, E., Roy, P., Asselin, G., Hu, G., Simonet, J., Stanley, A., Albertson, C., Fisher, S., 2016.
Osterix/Sp7 limits cranial bone initiation sites and is required for formation of sutures.
Developmental biology 413, 160-172.
Kamath, B.M., Stolle, C., Bason, L., Colliton, R.P., Piccoli, D.A., Spinner, N.B., Krantz, I.D.,
2002. Craniosynostosis in Alagille syndrome. American journal of medical genetics 112, 176-
180.
Laue, K., Pogoda, H.M., Daniel, P.B., van Haeringen, A., Alanay, Y., von Ameln, S.,
Rachwalski, M., Morgan, T., Gray, M.J., Breuning, M.H., Sawyer, G.M., Sutherland-Smith, A.J.,
Nikkels, P.G., Kubisch, C., Bloch, W., Wollnik, B., Hammerschmidt, M., Robertson, S.P., 2011.
Craniosynostosis and multiple skeletal anomalies in humans and zebrafish result from a defect in
the localized degradation of retinoic acid. American journal of human genetics 89, 595-606.
Loeys, B.L., Chen, J., Neptune, E.R., Judge, D.P., Podowski, M., Holm, T., Meyers, J., Leitch,
C.C., Katsanis, N., Sharifi, N., Xu, F.L., Myers, L.A., Spevak, P.J., Cameron, D.E., De Backer,
J., Hellemans, J., Chen, Y., Davis, E.C., Webb, C.L., Kress, W., Coucke, P., Rifkin, D.B., De
Paepe, A.M., Dietz, H.C., 2005. A syndrome of altered cardiovascular, craniofacial,
neurocognitive and skeletal development caused by mutations in TGFBR1 or TGFBR2. Nature
genetics 37, 275-281.
Maddin, H.C., Piekarski, N., Sefton, E.M., Hanken, J., 2016. Homology of the cranial vault in
birds: new insights based on embryonic fate-mapping and character analysis. Royal Society open
science 3, 160356.
Maruyama, T., Jeong, J., Sheu, T.J., Hsu, W., 2016. Stem cells of the suture mesenchyme in
craniofacial bone development, repair and regeneration. Nature communications 7, 10526.
Matsuoka, T., Ahlberg, P.E., Kessaris, N., Iannarelli, P., Dennehy, U., Richardson, W.D.,
McMahon, A.P., Koentges, G., 2005. Neural crest origins of the neck and shoulder. Nature 436,
347-355.
Merrill, A.E., Bochukova, E.G., Brugger, S.M., Ishii, M., Pilz, D.T., Wall, S.A., Lyons, K.M.,
Wilkie, A.O., Maxson, R.E., Jr., 2006. Cell mixing at a neural crest-mesoderm boundary and
deficient ephrin-Eph signaling in the pathogenesis of craniosynostosis. Human molecular
genetics 15, 1319-1328.
Mongera, A., Singh, A.P., Levesque, M.P., Chen, Y.Y., Konstantinidis, P., Nusslein-Volhard, C.,
2013. Genetic lineage labeling in zebrafish uncovers novel neural crest contributions to the head,
including gill pillar cells. Development 140, 916-925.
Oram, K.F., Gridley, T., 2005. Mutations in snail family genes enhance craniosynostosis of
Twist1 haplo-insufficient mice: implications for Saethre-Chotzen Syndrome. Genetics 170, 971-
974.
Piekarski, N., Gross, J.B., Hanken, J., 2014. Evolutionary innovation and conservation in the
embryonic derivation of the vertebrate skull. Nature communications 5, 5661.
Quarto, N., Longaker, M.T., 2005. The zebrafish (Danio rerio): a model system for cranial suture
patterning. Cells, tissues, organs 181, 109-118.
59

Rice, D.P., Aberg, T., Chan, Y., Tang, Z., Kettunen, P.J., Pakarinen, L., Maxson, R.E., Thesleff,
I., 2000a. Integration of FGF and TWIST in calvarial bone and suture development.
Development 127, 1845-1855.
Rice, D.P.C., Aberg, T., Chan, Y.S., Tang, Z.Q., Kettunen, P.J., Pakarinen, L., Maxson, R.E.,
Thesleff, I., 2000b. Integration of FGF and TWIST in calvarial bone and suture development.
Development 127, 1845-1855.
Roybal, P.G., Wu, N.L., Sun, J., Ting, M.C., Schafer, C.A., Maxson, R.E., 2010. Inactivation of
Msx1 and Msx2 in neural crest reveals an unexpected role in suppressing heterotopic bone
formation in the head. Developmental biology 343, 28-39.
Saga, Y., Miyagawa-Tomita, S., Takagi, A., Kitajima, S., Miyazaki, J., Inoue, T., 1999. MesP1 is
expressed in the heart precursor cells and required for the formation of a single heart tube.
Development 126, 3437-3447.
Sanjana, N.E., Cong, L., Zhou, Y., Cunniff, M.M., Feng, G., Zhang, F., 2012. A transcription
activator-like effector toolbox for genome engineering. Nat Protoc 7, 171-192.
Sharma, V.P., Fenwick, A.L., Brockop, M.S., McGowan, S.J., Goos, J.A., Hoogeboom, A.J.,
Brady, A.F., Jeelani, N.O., Lynch, S.A., Mulliken, J.B., Murray, D.J., Phipps, J.M., Sweeney, E.,
Tomkins, S.E., Wilson, L.C., Bennett, S., Cornall, R.J., Broxholme, J., Kanapin, A., Whole-
Genome Sequences, C., Johnson, D., Wall, S.A., van der Spek, P.J., Mathijssen, I.M., Maxson,
R.E., Twigg, S.R., Wilkie, A.O., 2013. Mutations in TCF12, encoding a basic helix-loop-helix
partner of TWIST1, are a frequent cause of coronal craniosynostosis. Nature genetics 45, 304-
307.
Teng, C.S., Yen, H.Y., Barske, L., Smith, B., Llamas, J., Segil, N., Go, J., Sanchez-Lara, P.A.,
Maxson, R.E., Crump, J.G., 2017. Requirement for Jagged1-Notch2 signaling in patterning the
bones of the mouse and human middle ear. Scientific reports 7, 2497.
Ting, M.C., Wu, N.L., Roybal, P.G., Sun, J., Liu, L., Yen, Y., Maxson, R.E., Jr., 2009. EphA4 as
an effector of Twist1 in the guidance of osteogenic precursor cells during calvarial bone growth
and in craniosynostosis. Development 136, 855-864.
Topczewska, J.M., Shoela, R.A., Tomaszewski, J.P., Mirmira, R.B., Gosain, A.K., 2016. The
Morphogenesis of Cranial Sutures in Zebrafish. PLoS One 11, e0165775.
Walker, M.B., Kimmel, C.B., 2007. A two-color acid-free cartilage and bone stain for zebrafish
larvae. Biotechnic & histochemistry : official publication of the Biological Stain Commission 82,
23-28.
Westerfield, M., 2007. THE ZEBRAFISH BOOK: A guide for the laboratory use of zebrafish
(Danio rerio), 5th Edition ed. University of Oregon Press, Eugene, OR.
White, R.M., Sessa, A., Burke, C., Bowman, T., LeBlanc, J., Ceol, C., Bourque, C., Dovey, M.,
Goessling, W., Burns, C.E., Zon, L.I., 2008. Transparent adult zebrafish as a tool for in vivo
transplantation analysis. Cell stem cell 2, 183-189.
Wilk, K., Yeh, S.A., Mortensen, L.J., Ghaffarigarakani, S., Lombardo, C.M., Bassir, S.H.,
Aldawood, Z.A., Lin, C.P., Intini, G., 2017. Postnatal Calvarial Skeletal Stem Cells Expressing
PRX1 Reside Exclusively in the Calvarial Sutures and Are Required for Bone Regeneration.
Stem cell reports 8, 933-946.
Wojciechowski, J., Lai, A., Kondo, M., Zhuang, Y., 2007. E2A and HEB are required to block
thymocyte proliferation prior to pre-TCR expression. Journal of immunology 178, 5717-5726.
60

Worthley, D.L., Churchill, M., Compton, J.T., Tailor, Y., Rao, M., Si, Y., Levin, D., Schwartz,
M.G., Uygur, A., Hayakawa, Y., Gross, S., Renz, B.W., Setlik, W., Martinez, A.N., Chen, X.,
Nizami, S., Lee, H.G., Kang, H.P., Caldwell, J.M., Asfaha, S., Westphalen, C.B., Graham, T.,
Jin, G., Nagar, K., Wang, H., Kheirbek, M.A., Kolhe, A., Carpenter, J., Glaire, M., Nair, A.,
Renders, S., Manieri, N., Muthupalani, S., Fox, J.G., Reichert, M., Giraud, A.S., Schwabe, R.F.,
Pradere, J.P., Walton, K., Prakash, A., Gumucio, D., Rustgi, A.K., Stappenbeck, T.S., Friedman,
R.A., Gershon, M.D., Sims, P., Grikscheit, T., Lee, F.Y., Karsenty, G., Mukherjee, S., Wang,
T.C., 2015. Gremlin 1 identifies a skeletal stem cell with bone, cartilage, and reticular stromal
potential. Cell 160, 269-284.
Yen, H.Y., Ting, M.C., Maxson, R.E., 2010. Jagged1 functions downstream of Twist1 in the
specification of the coronal suture and the formation of a boundary between osteogenic and non-
osteogenic cells. Developmental biology 347, 258-270.
Yoshida, T., Vivatbutsiri, P., Morriss-Kay, G., Saga, Y., Iseki, S., 2008. Cell lineage in
mammalian craniofacial mesenchyme. Mechanisms of development 125, 797-808.
Zhao, H., Feng, J., Ho, T.V., Grimes, W., Urata, M., Chai, Y., 2015. The suture provides a niche
for mesenchymal stem cells of craniofacial bones. Nature cell biology 17, 386-396.

61

Chapter 3
Requirement for Jagged1-Notch2 signaling in patterning the
bones of the mouse and human middle ear
Abstract
Whereas Jagged1-Notch2 signaling is known to pattern the sensorineural components of the
inner ear, its role in middle ear development has been less clear. My laboratory previously
reported a role for Jagged-Notch signaling in shaping skeletal elements derived from the first two
pharyngeal arches of zebrafish. Here I show a conserved requirement for Jagged1-Notch2
signaling in patterning the stapes and incus middle ear bones derived from the equivalent
pharyngeal arches of mammals. Mice lacking Jagged1 or Notch2 in neural crest-derived cells
(NCCs) of the pharyngeal arches display a malformed stapes. Heterozygous Jagged1 knockout
mice, a model for Alagille Syndrome (AGS), also display stapes and incus defects. I find that
Jagged1-Notch2 signaling functions early to pattern the stapes cartilage template, with stapes
malformations correlating with hearing loss across all frequencies. I observe similar stapes
defects and hearing loss in one patient with heterozygous JAGGED1 loss, and a diversity of
conductive and sensorineural hearing loss in nearly half of AGS patients, many of which carry
JAGGED1 mutations. My findings reveal deep conservation of Jagged1-Notch2 signaling in
patterning the pharyngeal arches from fish to mouse to man, despite the very different functions
of their skeletal derivatives in jaw support and sound transduction.

3.1 Introduction
Despite their critical importance in sound transduction, we still know relatively little about the
developmental patterning of the diminutive ossicles of the mammalian middle ear. The malleus
and incus bones are derived from NCCs that populate the first (i.e. mandibular) pharyngeal arch,
and the stapes bone is derived from the second (i.e. hyoid) arch (O'Gorman, 2005). In a
fascinating evolutionary transition, these tiny middle ear bones are thought to have arisen by
modification of the ancestral fish jaw-support skeleton, with the malleus, incus, and stapes being
homologous to portions of the Meckel’s, palatoquadrate, and hyomandibular elements of fish,
respectively (Medeiros and Crump, 2012) (Figure 3-1A). In a genetic screen in zebrafish, the lab
previously identified a loss-of-function mutation in the JAG1 homolog, jag1b, which resulted in
62

specific malformations of the palatoquadrate and hyomandibular cartilages (Zuniga et al., 2010).
Subsequently, we found that Jag1b works through the Notch2 and Notch3 receptors to regulate
bone and cartilage differentiation in the dorsal portions of the mandibular and hyoid arches,
regions from which the incus and stapes bones arise in mammals (Barske et al., 2016). I
therefore asked in this study whether loss of Jagged-Notch signaling might similarly disrupt
development of the stapes and incus bones.
In humans, heterozygous loss-of-function mutations in JAGGED1 (JAG1) have been found in
94% of AGS patients (Li et al., 1997; Oda et al., 1997; Warthen et al., 2006), with a small
proportion of AGS patients harboring heterozygous loss-of-function mutations in NOTCH2
(McDaniell et al., 2006), which encodes a receptor for JAG1. Clinical diagnosis of AGS is based
on reduced numbers of intrahepatic bile ducts in the liver, accompanied by cholestasis, a
characteristic facial appearance, and defects in the heart, eyes, and skeleton (Alagille et al.,
1975). Although not part of the clinical diagnosis for AGS, there are some reports of hearing loss
in patients with AGS and/or mutations in JAG1. Work in mouse and chick has uncovered roles
for Jagged-Notch signaling in the patterning of the prosensory domain, which gives rise to the
hair and support cells of the cochlea (Hartman et al., 2010; Kiernan et al., 2006; Neves et al.,
2011). These findings would appear to suggest that hearing loss in AGS is primarily due to
sensorineural defects. However, there are also a few isolated reports of conductive hearing loss
in AGS, which indicates potential structural defects of the middle ear. In one large AGS kindred,
mild conductive hearing loss was noted (LaBrecque et al., 1982). Another kindred, with
missense mutations in JAG1 yet only the cardiac defects typically associated with AGS,
displayed mixed hearing loss (i.e. combined conductive and sensorineural components) (Le
Caignec et al., 2002). In a separate post-mortem analysis of the temporal bone, two AGS
individuals were described as having a “bulky” stapes, with one also displaying a “bulky” incus,
yet the precise morphological changes of these middle ear bones were unclear (Okuno et al.,
1990). These studies raise the possibility that defects in not only the neural components of the
inner ear but also the structural components of the middle ear might contribute to hearing loss in
AGS patients.

63

Figure 3-1. Craniofacial and middle ear defects in mice deficient for Jag1 or Notch2

(A) Diagrams of the heads of zebrafish, mouse, and human (left to right) show homology
between the fish jaw skeleton and the mammalian middle ear ossicles. The fish hyomandibula is
homologous to the mammalian stapes (green); the fish palatoquadrate is homologous in part to
the mammalian incus (Soo et al.); and the proximal portion of the fish Meckel’s cartilage is
homologous to the mammalian malleus (brown).
(B) Micro-CT scans of mouse skulls at three weeks of age. Compared to control Notch2
f/f
mice,
Wnt1-Cre; Jag1
f/f
and Wnt1-Cre; Notch2
f/f
mice exhibit a persistent foramen (arrowheads in
dorsal view) and midfacial hyperplasia resulting in an abnormally shaped skull and malocclusion
(lateral view).
64

(C, D) Dissected middle ear ossicles of three-week-old mice stained with Alizarin Red S.
Compared to Jag1
f/+
controls, Wnt1-Cre; Jag1
f/f
and Wnt1-Cre; Notch2
f/f
mice display a fully
penetrant columellar stapes phenotype. Wnt1-Cre; Notch2
f/f
mice also rarely display an ectopic
process from the anterior medial edge of the incus (arrow). Compared to wild-type siblings,
some Jag1
+/-
mice display a columellar stapes and a small ectopic process from the posterior
medial edge of the incus body (arrow and inset).

3.2 Results
3.2.1 Loss of Jag1 or Notch2 in NCCs results in craniofacial and middle ear
defects
Homozygous deletion of Jag1 or Notch2 results in early embryonic lethality in mice (Hamada et
al., 1999; Xue et al., 1999). The lab had previously reported that conditional deletion of Jag1 in
NCCs (using Wnt1-Cre) resulted in a persistent foramen in the frontal bone (Yen et al., 2010),
and an independent group reported midfacial hypoplasia in these Wnt1-Cre; Jag1
f/f
conditional
knockout (Jag1-CKO) mice, reminiscent of the facial characteristics of AGS (Hill et al., 2014;
Humphreys et al., 2012). Micro-computed tomography (uCT) scans confirmed these previously
reported phenotypes in Jag1-CKO mice, and revealed a similar persistent foramen and midfacial
hypoplasia in Wnt1-Cre; Notch2
f/f
(Notch2-CKO) mice (Figure 3-1B). I therefore analyzed the
effects of removing both copies of Jag1 or Notch2 in NCCs on middle ear bone development. At
postnatal day 21 (P21), Alizarin Red staining of bone in Jag1
f/f
controls shows that the stapes
consists of two cruces with a prominent foramen. In 100% of Jag1-CKO (n = 8/8) and Notch2-
CKO (n = 8/8) mice, the stapes was narrower than in controls and had a reduced or absent
foramen (Figure 3-1C and Table 3-1). Whereas the stapes of Jag1-CKO mice were consistently
reduced in size, there was some variability in the amount of ectopic mineralization in the reduced
foramen (Figure supplement 3-1). This single crus phenotype of the stapes has previously been
referred to as “columellar” or “monopode” (Hoshino, 1980; Jahrsdoerfer et al., 1989; Kurosaki et
al., 1995; Scheer, 1967). In contrast, the malleus and incus bones of Jag1-CKO and Notch2-
CKO mice were less affected, with just 12.5% (n = 1/8) of Notch2-CKO mice and no Jag1-CKO
mice having an ectopic process extending from the anterior medial edge of the incus body.
As heterozygous loss-of-function mutations in human JAG1 result in AGS, a previous student,
Dr. Hai-Yun Yen, also examined the effects of removing just one copy of Jag1 throughout the
65

whole mouse (Figure 3-1D and Table 3-1). In Jag1 heterozygotes, we observed a similar
columellar stapes in 25% of animals (n = 5/20), and an ectopic process extending from the
posterior medial edge of the incus body in 75% of animals (n = 15/20). As with Jag1-CKO mice,
the malleus was unaffected in Jag1 heterozygotes. These findings indicate that development of
the stapes and incus is especially sensitive to reduced dosage of Jag1-Notch2 signaling in mice.

66
Table 3-1. Middle ear defects detected in Jag1, Notch2, and Twist1 mutant mice

Summary of incus and stapes defects in mice deficient for Jag1, Notch2, and Twist1.

Genotype N
Incus Stapes
N
Retrotympanic
process

ectopic posterior
medial process
ectopic anterior
medial process
small or
no lumen

broken
Wild type 18 3 (16.7%) 0 (0%) 0 (0%) 23 1 (4%)
Jag1
f/+
9 0 (0%) 0 (0%) 0 (0%) N.D. -
Wnt1-Cre; Jag1
f/+
9 0 (0%) 0 (0%) 0 (0%) N.D. -
Wnt1-Cre; Jag1
f/f
8 1 (12.5%) 0 (0%) 8 (100%)
d
N.D. -
Notch2
f/+
5 0 (0%) 0 (0%) 0 (0%) N.D. -
Wnt1-Cre; Notch2
f/+
7 0 (0%) 0 (0%) 0 (0%) N.D. -
Wnt1-Cre; Notch2
f/f
8 0 (0%) 1 (12.5%) 8 (100%)
e
N.D. -
Jag1
+/-
20 15 (75%)
a
0 (0%) 5 (25%)
f
15 0 (0%)
Twist1
+/-
16 0 (0%) 6 (38%)
b
0 (0%) 18 2 (11%)
Jag1
+/-
; Twist1
+/-
13 0 (0%) 9 (69%)
c
1 (8%) 22 15 (68%)
g

For statistical analysis, we performed a multi-group comparison using a Fisher Exact Test or Chi-Square Test, followed by post-hoc pair-
wise comparisons. N.D. = not determined.
a
= p < 0.005 vs wild type, Twist1
+/-
, and Jag1
+/-
; Twist1
+/-

b
= p < 0.01 vs wild type and Jag1
+/-

c
= p < 0.025 vs Twist1
+/-

d
= p < 0.001 vs Jag1
f/+
and Wnt1-Cre; Jag1
f/+

e
= p < 0.001 vs Notch2
f/+
and Wnt1-Cre; Notch2
f/+

f
= p < 0.05 vs wild type
g
= p < 0.001 vs wild type, Jag1
+/-
and Twist1
+/-

67
Figure supplement 3-1. Complete penetrance of stapes defects in Jag1 CKO mice

Dissected stapes bones were stained with Alizarin Red S. Wnt1-Cre; Jag1
f/f
mice display a fully
penetrant columellar stapes phenotype, although there is some variability in the extent of ectopic
ossification within the reduced foramen.

3.2.2 Jag1 is necessary for early patterning of the stapes cartilage but not
formation of the stapedial artery
The lack of a foramen in the mutant stapes could be due to earlier mispatterning of the cartilage
template, ectopic mineralization, or loss of the stapedial artery that runs through the normal
foramen. We therefore examined the middle ear cartilages of newborn mice, before they are
ossified, using Alizarin Red and Alcian Blue to stain for bone and cartilage (Figure 3-2A,B and
Table 3-1). In wild-type, Jag1
f/f
, and Wnt1-Cre; Jag1
f/+
controls, the stapes cartilage had a
prominent foramen. In contrast, Jag1
+/-
and Wnt1-Cre; Jag1
f/f
mice had a smaller stapes cartilage
with a reduced or absent foramen, consistent with the later columellar phenotype of the ossified
stapes bone. We next examined the stapedial artery, which normally passes through the stapes
and serves as a bridge connecting the external and internal carotid arteries. To visualize this
artery in conjunction with the stapes cartilage, I bred the conditional Jag1 mutants onto a
Rosa26-Tomato reporter background and, with the help of Juan Llamas in the laboratory of Dr.
Neil Segil, injected India ink into the artery after dissecting out the intact middle and inner ear
from newborn mice. As in controls, we still observed a prominent stapedial artery in Jag1-CKO
68

mice (n = 3/3), which curved around the misshapen stapes cartilage (Figure 3-2C). These
findings indicate that the stapes defects in Jag1-deficient mice are likely due to an early
mispatterning of the cartilage template rather than loss of the stapedial artery or ectopic
mineralization.
Figure 3-2. Mispatterning of middle ear cartilages and formation of the stapedial artery in
Jag1-deficient mice

(A, B) Newborn mice were stained with Alcian Blue for cartilage and Alizarin Red S for bone.
Close-up views show the developing middle ear, which is diagrammed below for wild-type and
Jag1 heterozygous mice (malleus, brown; incus, red; stapes, green). Dissected middle ear
cartilages are shown for conditional mutants. Arrows point to the stapes cartilage, which is
reduced in size in both heterozygous and conditional Jag1 mutant mice.
(C) The stapes of Wnt1-Cre; Jag1
f/+
; Rosa26-Tomato and Wnt1-Cre; Jag1
f/f
; Rosa26-Tomato
mice fluoresce red and the stapedial arteries appear black from India ink injection. The artery is
still present in Jag1-CKO mice, where it deviates around the misshapen stapes cartilage.
Arrowheads point to the stapes.

69

3.2.3 Jag1 and Twist1 interact to pattern the incus and retrotympanic
process
Because we had previously uncovered a genetic interaction between Jag1 and Twist1 in coronal
suture development (Yen et al., 2010), we next examined whether the variable penetrance of
middle ear defects in Jag1 heterozygotes might be due in part to genetic interactions with Twist1.
In humans, heterozygous loss of TWIST1 results in Saethre-Chotzen syndrome (Twigg and
Wilkie, 2015), a variable feature of which is conductive or mixed hearing loss (Lamonica et al.,
2010; Rosen et al., 2011). In Twist1
+/-
mice, we observed that the stapes and malleus bones were
normal, with loss of one Twist1 allele failing to enhance the stapes defects of Jag1
+/-
mice (Table
3-1). In contrast, 38% (n = 6/16) of Twist1
+/-
mice developed a prominent ectopic process from
the anterior medial edge of the incus body, and in Jag1
+/-
; Twist1
+/-
mice the penetrance of this
phenotype increased to 69% (n = 9/13) (Table 3-1 and Figure 3-3A). The retrotympanic process,
a posterior extension of the squamosal bone that lies just above the incus, was also reduced in
size and fragmented in Twist1
+/-
but not Jag1
+/-
single mutants. The penetrance of this phenotype
increased from 11% (n = 2/18) to 68% (n = 13/22) in compound heterozygotes (Table 3-1 and
Figure 3-3B). These findings reveal a selective interaction between Jag1 and Twist1 in patterning
the mandibular arch from which the incus and retrotympanic process derive, and not the hyoid
arch from which the stapes derives. As both Twist1 and Jagged-Notch signaling inhibit skeletal
differentiation in the head (Bialek et al., 2004; Merrill et al., 2006; Yen et al., 2010; Zanotti and
Canalis, 2016), inappropriate skeletal differentiation might underlie both the suture and middle
ear bone phenotypes in animals deficient in these factors. Further analysis will be required to
determine whether Twist1 and Jagged-Notch signaling function in a linear pathway or in parallel
for mandibular arch patterning.

70

Figure 3-3. Incus and retrotympanic defects in Jag1; Twist1 compound mutants

(A) Dissected incus bones of three-week-old mice stained with Alizarin Red S. Wild types and
this Jag1
+/-
example display a normal incus. In contrast, Twist1
+/-
and Jag1
+/-
; Twist1
+/-
mice
have an extra process (black arrows) extending from the anterior medial edge of the incus body.
Accompanying diagrams illustrate the ectopic processes with orange lines.
(B) Views of the temporal bone in three-week-old mice stained with Alizarin Red S. The dashed
box in the illustration shows the approximate region being imaged. Compared to wild-type and
Jag1
+/-
mice, Twist1
+/-
and Jag1
+/-
; Twist1
+/-
mice display reduction and fragmentation of the
retrotympanic process (shown in orange in the adjacent diagrams).

71

3.2.4 Jag1 is required in NCCs for normal hearing in mice
The stapes and incus bones are essential for normal hearing (Baba et al., 2004; Teunissen and
Cremers, 1993). To test whether the ossicular defects of Jag1-CKO mice result in hearing loss, I
worked with Juan Llamas to measure auditory brainstem response (ABR) at P18. In a click
stimulus test, in which a large range of frequencies is presented simultaneously, we observed an
approximately 30 decibel shift in the sound pressure level (dB SPL) in Jag1-CKO mice
compared to Wnt1-Cre; Jag1
f/+
or Jag1
f/+
controls (Figure 3-4A). Hearing level thresholds were
then measured at the specific frequencies of 4, 8, 12, 16, 24, and 32 kilohertz (kHz). We found a
roughly 22 db SPL threshold shift in Jag1-CKO mice across all frequencies, significantly higher
than Wnt1-Cre; Jag1
f/+
, Jag1
f/+
, and Jag1
f/f
mice (Figure 3-4B). We also found a small but
statistically significant high-frequency threshold shift in mice lacking just one copy of Jag1 in
NCCs, compared to Jag1
f/+
controls; however this shift appeared to be largely attributable to one
animal and was no longer apparent when mice were retested at five weeks of age (Figure
supplement 3-2). As we deleted Jag1 solely in NCCs, these results are consistent with hearing
loss being due to structural defects of the stapes and/or other components of the middle ear. It
remains unclear whether the variability in hearing acuity of individual Jag1-CKO animals could
be explained by the small differences in stapes morphology observed (Figure supplement 3-1).
As neural crest-derived cells also make a small contribution to the inner ear, defects in these
structures might also contribute to the degree of hearing loss. However, the semicircular canals
of the inner ear, which are dysmorphic in Alagille Syndrome (Le Caignec et al., 2002) and
conventional Jag1 heterozygous mice (Kiernan et al., 2006; Tsai et al., 2001; Vrijens et al.,
2006), were unaffected in Jag1-CKO mice (Figure supplement 3-3), arguing against canal
defects being the cause of hearing loss.

72

Figure 3-4. Hearing loss in mice lacking Jag1 in NCCs

(A) A click stimulus test was performed in P18 mice, a stage at which wild-type mice show
normal hearing. Compared to Jag1
f/+
(n = 7) and Wnt1-Cre; Jag1
f/+
(n = 7) mice, the threshold at
which Wnt1-Cre; Jag1
f/f
mice (n = 2) could hear was significantly higher. Each point represents
one individually tested ear. **p < 0.01; differences were measured by one-way ANOVA with
post-hoc Tukey-Kramer HSD test. Error bars represent standard error of the mean.
(B) Auditory brainstem responses were recorded at a range of frequencies in P18 mice.
Compared to Jag1
f/+
(n = 8), Jag1
f/f
(n = 3), and Wnt1-Cre; Jag1
f/+
(n = 8) mice, Wnt1-Cre;
Jag1
f/f
mice (n = 4) showed significantly higher thresholds across all frequencies as determined
by a one-way ANOVA and subsequent post-hoc Tukey-Kramer HSD test. Hearing thresholds of
Wnt1-Cre; Jag1
f/+
mice were similar to Jag1
f/+
and Jag1
f/f
control mice at lower frequencies but
significantly different from Jag1
f/+
controls at 24 and 32 kHz. Circles represent averages, and
lines represent individually tested ears. *p < 0.05, **p < 0.01; error bars represent standard error
of the mean.

73

Figure supplement 3-2. Normal hearing in five-week-old mice lacking one copy of Jag1 in
NCCs

At five weeks of age, Wnt1-Cre; Jag1
f/+
(n = 7) and Jag1
f/+
(n = 7) mice did not have
significantly different hearing thresholds. Lack of difference at 4 kHz (p = 0.24), 8 kHz (p =
0.35), 12 kHz (p = 0.10), 16 kHz (p = 0.56), 24 kHz (p = 0.53), and 32 kHz (p = 0.08) was
determined by two-tailed student’s t-tests. Circles represent averages, and lines represent
individually tested ears. Error bars represent the standard error of the mean. See also Figure 4.

74

Figure supplement 3-3. Normal gross inner ear structures in Jag1 CKO mice

At P0, inner ears of Jag1
f/+
(n = 6) and Wnt1-Cre; Jag1
f/f
(n = 6) mice were not noticeably
different. In particular, the anterior (assc), posterior (pssc), and lateral (lssc) semicircular canals
are intact. Canals were filled by India ink injection of dissected ears.

3.2.5 Conductive hearing loss and anomalies in middle ear bones in
patients heterozygous for JAG1 loss-of-function mutations
Given the stapes defects in heterozygous Jag1 mutant mice, we investigated whether
heterozygous loss of JAG1 might also affect middle ear development and hearing in humans.
Conductive and mixed hearing loss have been described in a few kindreds with mutations or
deletions in JAG1 (LaBrecque et al., 1982; Le Caignec et al., 2002; Okuno et al., 1990), yet the
prevalence of such hearing loss in AGS was unclear (Krantz et al., 1997). My mentor, Dr. Gage
Crump, and collaborators, Dr. Pedro Sanchez-Lara and Dr. Bea Smith, therefore attended the
Alagille Alliance meetings in 2011 and 2014 and conducted hearing tests on participants with
AGS clinical diagnosis and/or known heterozygous mutations in JAG1. Of the 44 subjects tested,
75

16 have known mutations in JAG1 and the remaining 28 have not yet been determined (Table
supplement 3-1). The most common finding was conductive hearing loss (27% of left ears, 30%
of right ears), followed by mixed hearing loss (14% of left ears, 9% of right ears) and then
sensorineural hearing loss (4% of left ears, 11% of right ears) (Figure 3-5A and Table
supplement 3-1). We found conductive hearing loss to be primarily mild or moderate, compared
to sensorineural and mixed hearing loss, which could be in the severe to profound range (Figure
3-5A).
For five participants – one with conductive, two with mixed, one with sensorineural, and one
with no hearing loss – and a non-AGS control, we obtained computed tomography scans of the
temporal bone to visualize middle ear structures (summarized in Table 3-2). CT scans have been
shown to be fairly accurate in diagnosing stapes defects (Lagleyre et al., 2009). All five subjects
showed abnormalities in the posterior semicircular canals. With the help of Dr. John Go, a
radiologist specializing in head and neck imaging, we found that three showed defects in the
superior semicircular canals, consistent with previous reports of hypoplasia or absence of
semicircular canals in AGS (Koch et al., 2006; Le Caignec et al., 2002). In a 49-year-old male
with severe mixed hearing loss in the left ear and high-frequency hearing loss in the right ear, we
observed bilateral malformations of the stapes such that it appeared columellar (i.e. lacking
distinct anterior and posterior cruces) (Figure 3-5B,C), nearly identical to what we observed in
Jag1-deficient mice. This subject has a mutation in exon 19 (c.2345-2A>G) of JAG1 that alters
the splicing consensus sequence and leads to heterozygous JAG1 loss-of-function. He also had
normal compliance of the tympanic membrane in both ears. These findings are consistent with
partial defects in the ossicular chain resulting in high-frequency hearing loss (Anderson and Barr,
1971). We also observed an ectopic process extending from the left incus towards the posterior
wall of the tympanic cavity in a 9-year-old subject with sensorineural hearing loss in the right
ear, and irregular orientation of the petrous part of the temporal bone, which was sloping
upwards in a lateral to medial manner, in an 8-year-old subject with mixed hearing loss in both
ears (Table 3-2). In a 7-year-old subject with mild low-frequency conductive hearing loss in the
left ear, middle ear bones were normal yet the aperture of the cochlear nerve was reduced. A 16-
year-old subject with normal hearing, who had no defects in the middle ear bones, also displayed
abnormal calcification of the left oval window (Figure 3-5D). This lesion is suggestive of
otosclerosis, in which calcified bone fixes the stapes to the oval window and in some cases limits
the ability to transmit sound. In summary, while conductive hearing loss was associated with
76

stapes defects in one subject, our findings show a diversity of structural defects of the middle ear
in AGS patients that are variably associated with hearing loss.
Figure 3-5. Hearing loss and middle ear defects in patients heterozygous for JAG1
mutations and/or diagnosed with AGS

(A) The ratio of types of hearing loss for the right and left ears were calculated based on the 44
subjects tested. The degrees of hearing loss encompassed in each type of hearing loss were also
analyzed. Table S1 lists the degrees of hearing loss as a range indicating the loss at its best
frequency and worst frequency. The categorical analyses illustrated by the pie charts have taken
into account only the loss at the worst frequency.
77

(B) CT scans of the temporal bone in the axial plane from a control 69-year-old male without
AGS and subject 5 who is heterozygous for a JAG1 loss-of-function mutation. Magnified areas
of the dashed box regions and accompanying diagrams are shown below. Compared to the right
stapes (orange) from the control subject, the right and left stapes of subject 5 appear as a single
rod (i.e. columellar). Adjacent sections showed relatively normal articulation of the stapes with
the incus. For better comparison, the orientation of the left stapes is flipped horizontally in the
magnified view.
(C) Audiogram of subject 5 (see Table 2) indicates mild to profound mixed hearing loss in the
left ear and normal to mild sensorineural hearing loss with a potential high-frequency conductive
component in the right ear.
(D) Compared to a control 69-year-old male without AGS, CT scans of the temporal bone in the
coronal plane show inappropriate ossification of the oval window in the left ear of subject 1. The
control right ear is flipped horizontally in the magnified area and accompanying diagram.

78
Table 3-2. Hearing loss and middle ear defects in human subjects

Hearing tests and CT data of selected subjects from the 2011 and 2014 Alagille Alliance meetings.

Audiologic Data CT Data
Subject Age Mutation Ear
Type of
Hearing
Loss
Degree of
Hearing
Loss
Tympanogram
Affected
Structures
1

16 JAG1 R - - normal PSCC, CoA

L - - normal OW, PSCC, CoA
2
a
9 JAG1 R sens mild normal SSCC, PSCC

L - - normal incus
3 8 JAG1 R mixed mild
negative ME pressure/
normal compliance
OW, RW, SSCC,
PSCC

L mixed mild-severe
negative ME pressure/
normal compliance
OW, RW, SSCC,
PSCC
4 7 N.D. R - - normal PSCC, CoA

L cond mild normal PSCC
5
b
49 JAG1 R sens normal-mild normal
stapes, SSCC,
PSCC, CoA

L mixed
mild-
profound
normal
stapes, SSCC,
PSCC, CoA

cond = conductive hearing loss; CoA = cochlear aperature; CT = computed tomography; L = left; ME = middle ear; N.D. = not
determined OW = oval window; PSCC = posterior semicircular canal; R = right; RW = round window; sens = sensorineural hearing loss;
SSCC = superior semicircular canal; - = within normal limits.
a
Previous test had showed conductive hearing loss in right ear.
b
This subject has mutations in JAG1 but had not previously been diagnosed with AGS.

79
Table supplement 3-1. Hearing test results of Alagille syndrome patients

Results from hearing tests at the 2011 and 2014 Alagille Syndrome Alliance meetings. Unless
otherwise noted, all subjects had clinical diagnosis of AGS.

Subject
Age
Tested
Mutation
s
Ear Tympanogram
Type of
Hearing
Loss
Degree of
Hearing Loss
1 16 JAG1 R normal - -

L normal - -
2
a
9 JAG1 R
slightly reduced
eardrum mobility with
normal pressure
sens mild

L normal - -
3 8 JAG1 R
negative ME pressure/
normal compliance
mixed mild

L
negative ME pressure/
normal compliance
mixed mild-severe
4 7 N.D. R normal - -

L normal cond mild
5
b
49 JAG1 R normal sens normal-mild

L normal mixed mild-profound
6 5 N.D. R no ear drum mobility cond moderate

L no ear drum mobility cond moderate
7 5 N.D. R
negative ME pressure/
normal compliance
cond mod-normal

L
negative ME pressure/
normal compliance
cond mod-normal
8 6 N.D. R
negative ME pressure/
normal compliance
cond mild-normal

L normal - -
9 33 N.D. R normal sens mild-profound

L normal sens mild-profound
10 7 N.D. R normal - -

L normal cond normal-mild
11 3 N.D. R
negative ME pressure/
reduced compliance
cond unknown

L
negative ME pressure/
normal compliance
cond unknown
12 11 N.D. R normal - -

L normal - -
13 13 N.D. R normal mixed normal-mod

L normal mixed normal-mod
14 15 N.D. R normal - -

L normal - -
15 11 N.D. R normal - -

L normal - -
16 15 N.D. R normal - -

L normal - -
80

17 36
JAG1
mosaic
R normal - -

L normal - -
18 16 N.D. R no ear drum mobility mixed normal-mod

L
slightly reduced
eardrum mobility
mixed normal-mild
19 16 N.D. R normal mixed normal-mild

L normal cond normal-mild
20 26 N.D. R normal sens normal-mild

L normal sens normal-mild
21 3 N.D. R no ear drum mobility cond normal-mild

L no ear drum mobility cond normal-mild
22 15 N.D. R normal - -

L normal cond mild-mod
23 11 N.D. R normal - -

L normal - -
24 41 N.D. R normal - -

L normal - -
25 9 JAG1 R normal - -

L normal - -
26 32 JAG1 R normal - -

L normal - -
27 12 JAG1 R normal cond normal-mild

L normal - -
28 3 JAG1 R
negative ME pressure/
normal compliance
cond normal-mild

L
negative ME pressure/
normal compliance
cond mild
29 5 N.D. R CNT - -

L CNT - -
30 3 JAG1 R
negative ME pressure/
normal compliance
- -

L
negative ME pressure/
normal compliance
- -
31 10
JAG1
(de novo)
R normal cond normal-mild

L normal - -
32 7 JAG1 R CNT - -

L CNT - -
33 19 N.D. R reduced compliance - -

L reduced compliance - -
34 10 JAG1 R normal cond normal-mild

L normal - -
35 12 N.D. R normal - -

L normal - -
36 10 N.D. R normal - -

L normal - -
37 4 N.D. R
negative ME pressure/
normal compliance
cond mild
81

L CNT cond mild
38 24 N.D. R no ear drum mobility cond normal-mild

L no ear drum mobility mixed normal-mild
39 13
JAG1
(deletion)
R no ear drum mobility sens mild-mod

L CNT mixed mod-severe
40 16
JAG1
(de novo)
R hypercompliance - -

L normal - -
41 5 JAG1 R reduced compliance cond mild

L normal - -
42 24 N.D. R CNT - -

L CNT cond normal-mild
43
c
12 N.D. R no ear drum mobility cond normal-mild

L no ear drum mobility cond normal-mild
44
d
23 N.D. R CNT - -

L CNT - -

cond = conductive hearing loss; CNT = could not test; L = left; mod = moderate; ME = middle
ear; N.D. = not determined; R = right; sens = sensorineural hearing loss; - = within normal limits.
a
Previous test results showed conductive hearing loss in right ear.
b
This subject has mutations in JAG1 but was not previously clinically diagnosed with AGS.
c
Previous test results showed hearing at 250Hz was affected in both left and right ears.
d
Previous test results showed conductive hearing loss in both left and right ears at 250-500Hz.

3.3 Discussion
Our findings indicate a conserved role for Jag1-Notch2 signaling in NCCs for the patterning of
the stapes and incus bones of the mammalian middle ear. The fully penetrant stapes defects seen
upon NCC-specific deletion of Jag1 or Notch2 are consistent with findings in zebrafish that
Jag1b functions in NCCs to pattern the homologous hyomandibular cartilage (Zuniga et al.,
2010). Whereas hyomandibular defects in zebrafish lacking jag1b, or notch2 and notch3,
correlate with ectopic expression of the cartilage condensation marker barx1 in the hyoid arch
(Barske et al., 2016), we observed no differences in Barx1 expression in the hyoid arches of
Jag1-CKO mice at E10.5 and E11.5 (Figure supplement 3-4). However, it remains possible that
subtle differences in Barx1 expression escaped our detection, especially given the small number
of arch NCCs that contribute to the diminutive stapes bone. On the other hand, it seems less
likely that stapes defects are a secondary consequence of stapedial artery defects. Jag1 has been
shown to be required in endothelial cells, which are not of neural crest origin, for vascular
development (Benedito et al., 2009). Further, we show that NCC-specific loss of the Notch2
82

receptor, which is expected to act cell-autonomously in endothelial cells and not NCCs for artery
development, causes a similar stapes defect to NCC-specific Jag1 loss. Whereas the stapedial
artery is later associated with pericytes, which are of neural crest origin, pericytes are recruited
only after artery formation and thus are unlikely to affect the initial routing of the stapedial artery
(Benjamin et al., 1998). Further, the foramen of the stapes homolog (hyomandibula) in zebrafish
jag1b mutants is similarly lost, despite this foramen being associated with the facial nerve
instead of an artery (Zuniga et al., 2010). Thus, rerouting of the stapedial artery appears to be a
secondary consequence of the reduced stapes anlagen and not vice versa. Nonetheless, additional
studies are clearly needed to elucidate the developmental basis of stapes defects in Jag1-deficient
mice.

83

Figure supplement 3-4. Barx1 expression in conditional Jag1 mutants

At E10.5, Wnt1-Cre; Jag1
f/f
(n = 4) and Jag1
f/f
controls (n = 4) mice did not have noticeable
differences in Barx1 expression. At E11.5, Barx1 expression was also similar between Wnt1-
Cre; Jag1
f/f
(n = 5) and Jag1
f/f
controls (n = 4). Ventral views of the pharyngeal arches are shown
at the top of each panel, with lateral views of the left and right sides shown below. Anterior is to
the top in each image. As a reference in the top left, a lateral view of an E10.5 wild-type embryo
stained for Barx1 is shown, with the pharyngeal arch region indicated by the dashed box. Whole-
mount in situ hybridization was carried out as previously described
1
using the published Barx1
probe
2
.

Wnt1-Cre; RBP-J
f/f
mice with global loss of Notch signaling in NCCs display similar defects in
the frontal bone as we observe upon NCC-specific deletion of Jag1 or Notch2 (Mead and
84

Yutzey, 2012). Although middle ear bones were not examined in Wnt1-Cre; RBP-J
f/f
mice, the
similarity of stapes defects in Jag1-CKO and Notch2-CKO mice suggest that JAG1 and
NOTCH2 are the likely major ligand and receptor for both calvarial and middle ear bone
development. Of note, global but not NCC-specific heterozygous loss of Jag1 results in partially
penetrant stapes defects, as well as incus defects, suggesting additional requirements for Jag1 in
non-NCC tissues for ossicle patterning. One candidate tissue is the first endodermal pouch,
which displays strong Jag1 expression in both mice (Mitsiadis et al., 1997) and zebrafish
(Zuniga et al., 2010) and develops in close association with the stapes and incus anlagen. Jag2
and Notch3 may also partially compensate for Jag1 and Notch2 in other aspects of NCC skeletal
differentiation. Jag2 is required for development of the NCC-derived palate in mice (Jiang et al.,
1998), and notch3 reduction enhances craniofacial defects in notch2 mutant zebrafish (Barske et
al., 2016). However, deletion of Notch1 in neural-crest-derived cells does not cause midfacial
hypoplasia characteristic of Alagille Syndrome
20
, and a previous study did not detect expression
of either Notch1 gene in the developing zebrafish face (Zuniga et al., 2010). Interestingly, the
liver and heart defects of AGS have not been observed in the analogous Jag1 heterozygous mice
(McCright et al., 2002; Xue et al., 1999), yet we did observe similar stapes defects in both mice
and humans lacking one copy of JAG1. These results indicate that organs have independent
dosage sensitivities to JAG1 across species.
While we found a correlation between stapes defects and hearing loss in one individual
heterozygous for a JAG1 mutation, the presence of conductive or mixed hearing loss but
apparently normal middle ear bones in other AGS patients suggests that middle ear bone defects
alone cannot account for conductive hearing loss in this syndrome. NCC-specific Jag1 mutant
mice have midfacial hypoplasia, and the development and function of the Eustachian tube is
closely associated with the craniofacial skeleton (Algudkar et al., 2013; Kemaloglu et al., 2000).
Thus, other factors such as Eustachian tube dysfunction may contribute to hearing loss in
subjects with apparently normal ossicles. Similarly in NCC-specific Jag1 mutant mice, we
cannot rule out the possibility that defects in other NCC-derived ear structures besides the stapes
contribute to hearing loss, especially as the ABR test we performed does not distinguish between
conductive and sensorineural components. For example, NCC derivatives contribute to the early
cochleovestibular ganglion and utricle, as well as a few rare cells in the semicircular canals and
stria vascularis (Freyer et al., 2011). While the cochleovestibular ganglion and stria vascularis
are involved in auditory functions, the utricle and semicircular canals are responsible for balance.
85

Indeed, studies on three mouse mutants – Slalom (Tsai et al., 2001), Headturner (Kiernan et al.,
2001), and Ozzy (Vrijens et al., 2006) – have described specific point mutations in Jag1 resulting
in malformations of the semicircular canals, and we and others observe frequent losses and
anomalies of the superior and posterior semicircular canals in AGS patients (Koch et al., 2006;
Le Caignec et al., 2002; Okuno et al., 1990). Further, the role of Jagged-Notch signaling in the
development of the vestibular system also appears to be conserved in zebrafish, as jag1b mutants
were independently isolated based on semicircular canal defects (Ma and Zhang, 2015; Obholzer
et al., 2012). However, we found that NCC-specific loss of Jag1 in mice causes hearing loss
without affecting canal morphology, consistent with defects in the stapes and/or other middle and
inner ear structures being responsible for hearing defects in these mice. Nonetheless, superior
semicircular canal dehiscence can lead to autophony and conductive hearing loss, and dehiscence
of the canal can cause disruption of the normal endolymph flow, resulting in a lower bone
conduction and higher air conduction threshold (Bi et al., 2017). Although there are currently no
reports linking Jag1 and AGS to canal dehiscence, we cannot rule out a possible canal
dehiscence component in AGS-associated hearing loss. In addition, Jag1 has a well-known
requirement in patterning the ectodermal sensory placode from which the hair and support cells
of the inner ear derive (Hartman et al., 2010; Kiernan et al., 2006; Neves et al., 2011), consistent
with the finding of sensorineural hearing loss in several AGS patients. These observations
indicate a complex etiology of hearing loss in AGS, likely affecting multiple structures in both
the middle and inner ear, particularly for those patients presenting with mixed hearing loss.
However, for those patients with isolated middle ear bone defects, surgical correction might be
an option to improve hearing (Lippy et al., 2003).
To date, mutations in only a few genes, such as NOGGIN (Gong et al., 1999) and ANK (Kornak
et al., 2010), have been linked to congenital conductive hearing loss in humans. A contribution
of JAG1 mutations to congenital conductive hearing loss thus expands our knowledge of human
middle ear development. However, as AGS is relatively rare, we were only able to examine the
middle ears of a small number of patients. Thus, the extent to which stapes and/or incus defects
occur in this syndrome and contribute to hearing loss remains to be determined. While the
similarity in stapes defects between Jag1 mutant mice and an individual with heterozygous loss
of JAG1 strongly supports the view that human stapes malformations are due to JAG1
deficiency, we cannot rule out that second site mutations act synergistically with JAG1 mutations
to cause middle ear bone defects. For example, conditional deletion studies in mice have
86

revealed requirements for Tbx1 in middle and inner ear structures (Arnold et al., 2006), and
DiGeorge Syndrome, which is associated with heterozygous deletion of TBX1, encompasses an
array of defects that overlap with Alagille Syndrome, including tetralogy of fallot and hearing
loss. On the other hand, the patient with stapes defects did not display the multi-organ defects
typically associated with AGS, despite having a child with the same JAG1 mutation and typical
AGS features. Together with a previous study showing mixed hearing loss in a kindred with a
JAG1 missense mutation but only a subset of AGS features (Le Caignec et al., 2002), and the
presence of JAG1 mutations presenting with only one or two AGS features (Guegan et al., 2012),
our results suggest that reduced JAG1 function can cause hearing loss largely independently of
other AGS features. It will therefore be informative to examine the extent to which family
members of AGS patients who exhibit hearing loss despite the lack of an AGS diagnosis, as well
as unrelated patients with defects in the stapes and/or incus bones, carry JAG1 or NOTCH2
mutations.

3.4 Materials and Methods
3.4.1 Mouse mutants and genotyping
Animal experiments were approved by the University of Southern California IACUC committee,
and all methods were performed in accordance with the relevant guidelines and regulations.
Genotyping was performed as previously published for Jag1 (Xue et al., 1999), Twist1 (Chen
and Behringer, 1995), Jag1-flox (Brooker et al., 2006), Notch2-flox (McCright et al., 2006),
Wnt1-Cre (Danielian et al., 1998), and Rosa26-Tomato (Boddupally et al., 2016) mouse lines.
3.4.2 Micro-computed tomography of mice
Imaging was performed using a MicroCT 50 (Scanco Medical AG, Switzerland), scanning at
high resolution [2040×2040 in-plane image matrix; 0.18 degree rotational step (DRS)] and a
field of view of 20.4 mm. Scans were conducted at an energy setting of 70 kVp, current intensity
of 200 µA, and an integration time of 500 ms/projection. Two-dimensional slices taken at 10-
micron increments were rendered into three-dimensional reconstructions using Exposure Render
(Kroes et al., 2012).
87

3.4.3 Skull preparation
The heads of three-week-old mice were skinned and cleared with 1% KOH for 1 to 2 days,
stained with 2% Alizarin Red S in 1% KOH until mineralized bone is red, and stored in 100%
glycerol. Middle ear ossicles were then dissected out for analysis. The heads of P0 newborn mice
were skinned and double stained with Alcian Blue and Alizarin Red S for cartilage and bone as
described (Morriss-Kay, 1999) with minor modifications. Washes in tap water pre and post bone
staining were omitted; Alizarin Red S at 0.1% was used at 1:50 dilution in potassium hydroxide
(Kitazawa et al.); decolorizing treatment after bone staining was performed with 1% KOH
overnight; and the dehydration process was completed through sequentially increasing
proportions of 100% glycerol:1% KOH (1:3, 1:1, 3:1, 1:0). Samples were imaged using a Leica
S8 APO stereo microscope.
3.4.4 Stapes and stapedial artery visualization
Wnt1-Cre; Jag1
f/+
; Rosa26-Tomato
ki/+
and Jag1
f/f
; Rosa26-Tomato
ki/ki
mice were bred to
generate Wnt1-Cre; Jag1
f/f
; Rosa26-Tomato mutants. The heads of P0 newborn mice were
skinned, fixed overnight in 4% paraformaldehyde, and stored in phosphate buffered saline
(PBS). Whole ear structures were carefully dissected out, and the stapedial artery was injected
with India ink using glass capillary needles. The malleus, incus, and other tissues were removed
after injections and before imaging to better reveal the stapes. Image z-stacks were taken using a
Zeiss Axiozoom and processed using extended depth of field by Zeiss LSM software.
3.4.5 Hearing tests in mice
ABR was performed through inserted earphones, using closed-field acoustics. The sound
pressure level (SPL) of the stimuli ranged between 20 and 105 decibels (dB). In determining
ABR thresholds, 300 responses with artifacts less than 30 microvolts were averaged.
Presentation of stimuli and averaging of responses were both controlled by BioSig
software. When ABR threshold was above maximum output range, it was classified as 105 dB.
Hearing level thresholds were measured at specific frequencies of 4, 8, 12, 16, 24, and 32
kilohertz (kHz).
3.4.6 Hearing tests and computed tomography scans on human subjects
Studies on human patients were approved by the Institutional Review Board at the University of
Southern California Keck School of Medicine, and all methods were performed in accordance
88

with the relevant guidelines and regulations. Written informed consent to obtain samples for
genetics research was obtained from each subject and/or subject’s parent or guardian. Hearing
tests were performed at the Alagille Alliance meetings of 2011 and 2014. Otoscopy was
performed to check for occluding wax, fluid, or infection, as well as wellness of the canal and
middle ear system. The testing used was tympanometry and behavioral audiometry. For
tympanometry, a small probe tip is inserted into the ear canal to create a seal, and then air
pressure is directed into the external auditory ear canal. The mobility of the eardrum is recorded
to confirm the health of the middle ear system. For behavioral audiometry, thresholds of hearing
are tested from 250 to 8000 Hz. Test procedures are based on the age of the test subject. If a
subject is able to test conventionally by raising their hand only when a test frequency is heard,
the standard 10 dB up and 5 dB down testing procedure is performed. Frequencies are tested in
conventional order - 1000, 2000, 4000, 8000, 250, 500 Hz. If the difference between octaves is
greater or equal to 20 dB, then inner octave threshold is measured. In cases where a subject is too
young, unreliable or distracted to complete this testing procedure, conditioned play audiometry is
completed. In this form of testing, the subject is trained to perform an action in response to
sound, such as dropping a block in a bucket. Frequencies are not presented in any standard order
to keep the subject interested. If the patient becomes unreliable, then thresholds are not
recorded. In standard air and bone conduction tests, a hearing level of -10 to 25 dBHL (decibels
hearing level) is classified as normal, 25 to 40 dBHL as mild, 40 to 70 dBHL as moderate, 70 to
90 dBHL as severe, and 90 dBHL and above as profound. Conductive hearing loss is indicated
by decreased air but normal bone conduction, sensorineural hearing loss by decreased air and
bone conduction thresholds, and mixed hearing loss by decreased air and bone conduction
thresholds with the bone conduction threshold being 10 dB higher than for air. Temporal bone
CT scans on subjects 1 and 5 were acquired at the University of Southern California Keck
School of Medicine using standard clinical procedures. Temporal bone CT scans for subjects 2,
3, and 4 were kindly provided by the patients’ physicians.

3.5 References
Alagille, D., Odievre, M., Gautier, M., Dommergues, J.P., 1975. Hepatic ductular hypoplasia
associated with characteristic facies, vertebral malformations, retarded physical, mental, and
sexual development, and cardiac murmur. The Journal of pediatrics 86, 63-71.
89

Algudkar, A., Lim, B., Fan, K., Bentley, R., 2013. Sustained bilateral middle ear effusions post
orthognathic surgery successfully treated with grommet insertion. JRSM short reports 4,
2042533313489822.
Anderson, H., Barr, B., 1971. Conductive high-tone hearing loss. Arch Otolaryngol 93, 599-605.
Arnold, J.S., Braunstein, E.M., Ohyama, T., Groves, A.K., Adams, J.C., Brown, M.C., Morrow,
B.E., 2006. Tissue-specific roles of Tbx1 in the development of the outer, middle and inner ear,
defective in 22q11DS patients. Human molecular genetics 15, 1629-1639.
Baba, S., Ikezono, T., Pawankar, R., Yagi, T., 2004. Congenital malformations of the middle ear
with an intact external ear: a review of 38 cases. ORL; journal for oto-rhino-laryngology and its
related specialties 66, 74-79.
Barske, L., Askary, A., Zuniga, E., Balczerski, B., Bump, P., Nichols, J.T., Crump, J.G., 2016.
Competition between Jagged-Notch and Endothelin1 Signaling Selectively Restricts Cartilage
Formation in the Zebrafish Upper Face. PLoS genetics 12, e1005967.
Benedito, R., Roca, C., Sorensen, I., Adams, S., Gossler, A., Fruttiger, M., Adams, R.H., 2009.
The notch ligands Dll4 and Jagged1 have opposing effects on angiogenesis. Cell 137, 1124-
1135.
Benjamin, L.E., Hemo, I., Keshet, E., 1998. A plasticity window for blood vessel remodelling is
defined by pericyte coverage of the preformed endothelial network and is regulated by PDGF-B
and VEGF. Development 125, 1591-1598.
Bi, W.L., Brewster, R., Poe, D., Vernick, D., Lee, D.J., Eduardo Corrales, C., Dunn, I.F., 2017.
Superior semicircular canal dehiscence syndrome. Journal of neurosurgery, 1-9.
Bialek, P., Kern, B., Yang, X., Schrock, M., Sosic, D., Hong, N., Wu, H., Yu, K., Ornitz, D.M.,
Olson, E.N., Justice, M.J., Karsenty, G., 2004. A twist code determines the onset of osteoblast
differentiation. Developmental cell 6, 423-435.
Boddupally, K., Wang, G., Chen, Y., Kobielak, A., 2016. Lgr5 Marks Neural Crest Derived
Multipotent Oral Stromal Stem Cells. Stem cells 34, 720-731.
Brooker, R., Hozumi, K., Lewis, J., 2006. Notch ligands with contrasting functions: Jagged1 and
Delta1 in the mouse inner ear. Development 133, 1277-1286.
Chen, Z.F., Behringer, R.R., 1995. twist is required in head mesenchyme for cranial neural tube
morphogenesis. Genes & development 9, 686-699.
Danielian, P.S., Muccino, D., Rowitch, D.H., Michael, S.K., McMahon, A.P., 1998.
Modification of gene activity in mouse embryos in utero by a tamoxifen-inducible form of Cre
recombinase. Current biology : CB 8, 1323-1326.
Freyer, L., Aggarwal, V., Morrow, B.E., 2011. Dual embryonic origin of the mammalian otic
vesicle forming the inner ear. Development 138, 5403-5414.
Gong, Y., Krakow, D., Marcelino, J., Wilkin, D., Chitayat, D., Babul-Hirji, R., Hudgins, L.,
Cremers, C.W., Cremers, F.P., Brunner, H.G., Reinker, K., Rimoin, D.L., Cohn, D.H., Goodman,
90

F.R., Reardon, W., Patton, M., Francomano, C.A., Warman, M.L., 1999. Heterozygous
mutations in the gene encoding noggin affect human joint morphogenesis. Nature genetics 21,
302-304.
Guegan, K., Stals, K., Day, M., Turnpenny, P., Ellard, S., 2012. JAG1 mutations are found in
approximately one third of patients presenting with only one or two clinical features of Alagille
syndrome. Clinical genetics 82, 33-40.
Hamada, Y., Kadokawa, Y., Okabe, M., Ikawa, M., Coleman, J.R., Tsujimoto, Y., 1999.
Mutation in ankyrin repeats of the mouse Notch2 gene induces early embryonic lethality.
Development 126, 3415-3424.
Hartman, B.H., Reh, T.A., Bermingham-McDonogh, O., 2010. Notch signaling specifies
prosensory domains via lateral induction in the developing mammalian inner ear. Proc Natl Acad
Sci U S A 107, 15792-15797.
Hill, C.R., Yuasa, M., Schoenecker, J., Goudy, S.L., 2014. Jagged1 is essential for osteoblast
development during maxillary ossification. Bone 62, 10-21.
Hogan, B. Manipulating the mouse embryo : a laboratory manual, (Cold Spring Harbor
Laboratory Press, Plainview, N.Y., 1994).
Hoshino, T., 1980. Stapes anomaly and cochlear sensory cell changes. A scanning electron
microscope study. Archives of oto-rhino-laryngology 228, 93-99.
Humphreys, R., Zheng, W., Prince, L.S., Qu, X., Brown, C., Loomes, K., Huppert, S.S.,
Baldwin, S., Goudy, S., 2012. Cranial neural crest ablation of Jagged1 recapitulates the
craniofacial phenotype of Alagille syndrome patients. Human molecular genetics 21, 1374-1383.
Jahrsdoerfer, R.A., Aguilar, E.A., Yeakley, J.W., Cole, R.R., 1989. Treacher Collins syndrome:
an otologic challenge. The Annals of otology, rhinology, and laryngology 98, 807-812.
Jiang, R., Lan, Y., Chapman, H.D., Shawber, C., Norton, C.R., Serreze, D.V., Weinmaster, G.,
Gridley, T., 1998. Defects in limb, craniofacial, and thymic development in Jagged2 mutant
mice. Genes Dev 12, 1046-1057.
Kemaloglu, Y.K., Kobayashi, T., Nakajima, T., 2000. Associations between the eustachian tube
and craniofacial skeleton. International journal of pediatric otorhinolaryngology 53, 195-205.
Kiernan, A.E., Ahituv, N., Fuchs, H., Balling, R., Avraham, K.B., Steel, K.P., Hrabe de Angelis,
M., 2001. The Notch ligand Jagged1 is required for inner ear sensory development. Proceedings
of the National Academy of Sciences of the United States of America 98, 3873-3878.
Kiernan, A.E., Xu, J., Gridley, T., 2006. The Notch ligand JAG1 is required for sensory
progenitor development in the mammalian inner ear. PLoS genetics 2, e4.
Kitazawa, T., Fujisawa, K., Narboux-Neme, N., Arima, Y., Kawamura, Y., Inoue, T., Wada, Y.,
Kohro, T., Aburatani, H., Kodama, T., Kim, K.S., Sato, T., Uchijima, Y., Maeda, K., Miyagawa-
Tomita, S., Minoux, M., Rijli, F.M., Levi, G., Kurihara, Y., Kurihara, H., 2015. Distinct effects
of Hoxa2 overexpression in cranial neural crest populations reveal that the mammalian
hyomandibular-ceratohyal boundary maps within the styloid process. Developmental biology.
91

Koch, B., Goold, A., Egelhoff, J., Benton, C., 2006. Partial absence of the posterior semicircular
canal in Alagille syndrome: CT findings. Pediatric radiology 36, 977-979.
Kornak, U., Brancati, F., Le Merrer, M., Lichtenbelt, K., Hohne, W., Tinschert, S., Garaci, F.G.,
Dallapiccola, B., Nurnberg, P., 2010. Three novel mutations in the ANK membrane protein
cause craniometaphyseal dysplasia with variable conductive hearing loss. American journal of
medical genetics. Part A 152A, 870-874.
Krantz, I.D., Piccoli, D.A., Spinner, N.B., 1997. Alagille syndrome. Journal of medical genetics
34, 152-157.
Kroes, T., Post, F.H., Botha, C.P., 2012. Exposure render: an interactive photo-realistic volume
rendering framework. PloS one 7, e38586.
Kurosaki, Y., Kuramoto, K., Matsumoto, K., Itai, Y., Hara, A., Kusakari, J., 1995. Congenital
ossification of the stapedius tendon: diagnosis with CT. Radiology 195, 711-714.
LaBrecque, D.R., Mitros, F.A., Nathan, R.J., Romanchuk, K.G., Judisch, G.F., El-Khoury, G.H.,
1982. Four generations of arteriohepatic dysplasia. Hepatology 2, 467-474.
Lagleyre, S., Sorrentino, T., Calmels, M.N., Shin, Y.J., Escude, B., Deguine, O., Fraysse, B.,
2009. Reliability of high-resolution CT scan in diagnosis of otosclerosis. Otology & neurotology
: official publication of the American Otological Society, American Neurotology Society [and]
European Academy of Otology and Neurotology 30, 1152-1159.
Lamonica, D.A., Maximino, L.P., Feniman, M.R., Silva, G.K., Zanchetta, S., Abramides, D.V.,
Passos-Bueno, M.R., Rocha, K., Richieri-Costa, A., 2010. Saethre-Chotzen syndrome,
Pro136His TWIST mutation, hearing loss, and external and middle ear structural anomalies:
report on a Brazilian family. The Cleft palate-craniofacial journal : official publication of the
American Cleft Palate-Craniofacial Association 47, 548-552.
Le Caignec, C., Lefevre, M., Schott, J.J., Chaventre, A., Gayet, M., Calais, C., Moisan, J.P.,
2002. Familial deafness, congenital heart defects, and posterior embryotoxon caused by cysteine
substitution in the first epidermal-growth-factor-like domain of jagged 1. American journal of
human genetics 71, 180-186.
Li, L., Krantz, I.D., Deng, Y., Genin, A., Banta, A.B., Collins, C.C., Qi, M., Trask, B.J., Kuo,
W.L., Cochran, J., Costa, T., Pierpont, M.E., Rand, E.B., Piccoli, D.A., Hood, L., Spinner, N.B.,
1997. Alagille syndrome is caused by mutations in human Jagged1, which encodes a ligand for
Notch1. Nat Genet 16, 243-251.
Lippy, W.H., Battista, R.A., Berenholz, L., Schuring, A.G., Burkey, J.M., 2003. Twenty-year
review of revision stapedectomy. Otol Neurotol 24, 560-566.
Ma, W.R., Zhang, J., 2015. Jag1b is essential for patterning inner ear sensory cristae by
regulating anterior morphogenetic tissue separation and preventing posterior cell death.
Development 142, 763-773.
McCright, B., Lozier, J., Gridley, T., 2002. A mouse model of Alagille syndrome: Notch2 as a
genetic modifier of Jag1 haploinsufficiency. Development 129, 1075-1082.
92

McCright, B., Lozier, J., Gridley, T., 2006. Generation of new Notch2 mutant alleles. Genesis
44, 29-33.
McDaniell, R., Warthen, D.M., Sanchez-Lara, P.A., Pai, A., Krantz, I.D., Piccoli, D.A., Spinner,
N.B., 2006. NOTCH2 mutations cause Alagille syndrome, a heterogeneous disorder of the notch
signaling pathway. Am J Hum Genet 79, 169-173.
Mead, T.J., Yutzey, K.E., 2012. Notch pathway regulation of neural crest cell development in
vivo. Developmental dynamics : an official publication of the American Association of
Anatomists 241, 376-389.
Medeiros, D.M., Crump, J.G., 2012. New perspectives on pharyngeal dorsoventral patterning in
development and evolution of the vertebrate jaw. Developmental biology 371, 121-135.
Merrill, A.E., Bochukova, E.G., Brugger, S.M., Ishii, M., Pilz, D.T., Wall, S.A., Lyons, K.M.,
Wilkie, A.O., Maxson, R.E., Jr., 2006. Cell mixing at a neural crest-mesoderm boundary and
deficient ephrin-Eph signaling in the pathogenesis of craniosynostosis. Human molecular
genetics 15, 1319-1328.
Mitsiadis, T.A. & Drouin, J. Deletion of the Pitx1 genomic locus affects mandibular tooth
morphogenesis and expression of the Barx1 and Tbx1 genes. Developmental biology 313, 887-
896 (2008).
Mitsiadis, T.A., Henrique, D., Thesleff, I., Lendahl, U., 1997. Mouse Serrate-1 (Jagged-1):
expression in the developing tooth is regulated by epithelial-mesenchymal interactions and
fibroblast growth factor-4. Development 124, 1473-1483.
Morriss-Kay, G.M., 1999. Treatment of mice with retinoids in vivo and in vitro. Skeletal
staining. Methods in molecular biology 97, 33-39.
Neves, J., Parada, C., Chamizo, M., Giraldez, F., 2011. Jagged 1 regulates the restriction of Sox2
expression in the developing chicken inner ear: a mechanism for sensory organ specification.
Development 138, 735-744.
O'Gorman, S., 2005. Second branchial arch lineages of the middle ear of wild-type and Hoxa2
mutant mice. Dev Dyn 234, 124-131.
Obholzer, N., Swinburne, I.A., Schwab, E., Nechiporuk, A.V., Nicolson, T., Megason, S.G.,
2012. Rapid positional cloning of zebrafish mutations by linkage and homozygosity mapping
using whole-genome sequencing. Development 139, 4280-4290.
Oda, T., Elkahloun, A.G., Pike, B.L., Okajima, K., Krantz, I.D., Genin, A., Piccoli, D.A.,
Meltzer, P.S., Spinner, N.B., Collins, F.S., Chandrasekharappa, S.C., 1997. Mutations in the
human Jagged1 gene are responsible for Alagille syndrome. Nat Genet 16, 235-242.
Okuno, T., Takahashi, H., Shibahara, Y., Hashida, Y., Sando, I., 1990. Temporal bone
histopathologic findings in Alagille's syndrome. Archives of otolaryngology--head & neck
surgery 116, 217-220.
93

Rosen, H., Andrews, B.T., Meara, J.G., Stoler, J.M., Mulliken, J.B., Rogers, G.F., 2011.
Audiologic findings in Saethre-Chotzen syndrome. Plastic and reconstructive surgery 127, 2014-
2020.
Scheer, A.A., 1967. Correction of congenital middle ear deformities. Archives of otolaryngology
85, 269-277.
Soo, K., O'Rourke, M.P., Khoo, P.-L., Steiner, K.A., Wong, N., Behringer, R.R., Tam, P.P.L.,
2002. Twist Function Is Required for the Morphogenesis of the Cephalic Neural Tube and the
Differentiation of the Cranial Neural Crest Cells in the Mouse Embryo. Developmental biology
247, 251-270.
Teunissen, E.B., Cremers, W.R., 1993. Classification of congenital middle ear anomalies. Report
on 144 ears. The Annals of otology, rhinology, and laryngology 102, 606-612.
Tsai, H., Hardisty, R.E., Rhodes, C., Kiernan, A.E., Roby, P., Tymowska-Lalanne, Z., Mburu, P.,
Rastan, S., Hunter, A.J., Brown, S.D., Steel, K.P., 2001. The mouse slalom mutant demonstrates
a role for Jagged1 in neuroepithelial patterning in the organ of Corti. Human molecular genetics
10, 507-512.
Twigg, S.R., Wilkie, A.O., 2015. New insights into craniofacial malformations. Hum Mol Genet
24, R50-59.
Vrijens, K., Thys, S., De Jeu, M.T., Postnov, A.A., Pfister, M., Cox, L., Zwijsen, A., Van Hoof,
V., Mueller, M., De Clerck, N.M., De Zeeuw, C.I., Van Camp, G., Van Laer, L., 2006. Ozzy, a
Jag1 vestibular mouse mutant, displays characteristics of Alagille syndrome. Neurobiology of
disease 24, 28-40.
Warthen, D.M., Moore, E.C., Kamath, B.M., Morrissette, J.J., Sanchez-Lara, P.A., Piccoli, D.A.,
Krantz, I.D., Spinner, N.B., 2006. Jagged1 (JAG1) mutations in Alagille syndrome: increasing
the mutation detection rate. Human mutation 27, 436-443.
Xue, Y., Gao, X., Lindsell, C.E., Norton, C.R., Chang, B., Hicks, C., Gendron-Maguire, M.,
Rand, E.B., Weinmaster, G., Gridley, T., 1999. Embryonic lethality and vascular defects in mice
lacking the Notch ligand Jagged1. Human molecular genetics 8, 723-730.
Yen, H.Y., Ting, M.C., Maxson, R.E., 2010. Jagged1 functions downstream of Twist1 in the
specification of the coronal suture and the formation of a boundary between osteogenic and non-
osteogenic cells. Developmental biology 347, 258-270.
Yoshida, T., Phylactou, L.A., Uney, J.B., Ishikawa, I., Eto, K., Iseki, S., 2005. Twist is required
for establishment of the mouse coronal suture. Journal of anatomy 206, 437-444.
Zanotti, S., Canalis, E., 2016. Notch Signaling and the Skeleton. Endocrine reviews 37, 223-253.
Zuniga, E., Stellabotte, F., Crump, J.G., 2010. Jagged-Notch signaling ensures dorsal skeletal
identity in the vertebrate face. Development 137, 1843-1852.

94

Chapter 4
Perspectives on the Future
Abstract
Results of an experiment often answer a particular question and raise many more questions. This
is particularly true regarding the calvarial suture study described in Chapter 2. The study on
middle ear bone patterning regulated by Jagged-Notch signaling was more descriptive by nature,
and proposed future work would be to increase clinical observations of middle ear phenotypes in
Alagille syndrome patients. Thus, this chapter is an extended discussion in response to
unanswered research questions regarding calvarial sutures, with particular focus on Twist1
dimerization partners and coronal suture sensitivity.
4.1 Dimerization partners as key to differential regulation
Given the multiple roles Twist1 has in craniofacial development, multiple previous studies
attempted to understand the protein function. Twist1 encodes a basic helix-loop-helix (bHLH)
transcription factor that binds to sequences known as E-boxes. As class II bHLH proteins such as
Twist1 are known to act on downstream target genes as homodimers or heterodimers with class I
E-box proteins, an understanding of dimerization complexes may reveal functional capacity. The
relatively ubiquitously expressed class I bHLH proteins include Tcf12, also known as HEB, and
E2A, also known as Tcf3. E2A encodes E12 and E47 proteins, and are collectively referred to as
E2A. Firulli et al. took a biochemical approach and focused on particularly conserved threonine
and serine sites in the helix I region of Twist1 (Firulli et al., 2005). They found that affinities for
homodimerization and heterodimerization could be controlled via phosphorylation of these sites.
The same year, Funato et al. described a TWIST1 mutation identified from a patient presenting
with right unicoronal synostosis. This mutation was mapped to the nuclear localization signal of
TWIST1, but an in vitro reporter assay performed on a human osteoblastic cell line suggested that
the mutated protein still retained its ability to inhibit E2A-induced transcriptional activation of
p21 (Funato et al., 2005), which is thought to be the pathway for cell cycle arrest and then
differentiation of osteoblast precursors (Funato et al., 2001). Intriguingly, the mutant protein was
found localized to the cytoplasm and nucleus, yet specifically localized to the nucleus only with
E12 expression. This suggests the importance of interaction with E12 in this scenario for proper
nuclear translocation. Connerney et al. confirmed Twist1 and E12 could function as a
95

heterodimer through an electrophoretic mobility shift assay with force-fused heterodimers
(Connerney et al., 2006). They further found that osteogenesis and suture related genes, namely
periostin, thrombospondin-1, and Fgfr2, are differentially regulated by Twist1 homo- and
heterodimers. Thus, they proposed that the ratio of Twist1-Twist1 homodimers to Twist1-E2A
heterodimers might be different at the osteogenic fronts of calvarial bones versus the suture
mesenchyme (Connerney et al., 2006). Another study by Bialek et al. used mouse models to
show a genetic interaction between Twist1 and Runx2 and suggested that Twist1 functions to
inhibit Runx2-dependent osteogenic differentiation (Bialek et al., 2004). This result supports a
role for Twist1 in the suture maintaining the identity of a progenitor pool. Bialek et al. further
narrowed it down to a 21 amino acid domain in the Twist1 C-terminus, which they termed the
“Twist box,” that physically interacts with Runx2. This result revealed an added layer of
complexity in Twist1 protein function, apart from the more focused bHLH domains.
Collectively, these studies highlighted the possible versatility of a single protein in binding
specific factors to carry out specific regulations. As bHLH transcription factors are classified into
six types, the possibilities of interactions for distinct regulatory actions are vast and remain
unexplored.
In my own work, I have begun to investigate different dimerization scenarios for Twist1;
however, much more work in the future is necessary in utilizing these newly made tools. As
discussed in Chapter 2, I found Tcf12 to be a suture-specific partner for Twist1. Briefly, Tcf12
mouse and zebrafish mutants do not have facial phenotypes, and tcf12
-/-
on the twist1a
-/-
; twist1b
-
/-
background suppresses the jaw cartilage phenotype in fish. This had led me to hypothesize that
Twist1 either homodimerizes with other Twist members or heterodimerizes with other Tcf
members. To begin addressing these possibilities, I first looked into other Twist family members.
Zebrafish have four genes, including twist2 and twist3 in addition to the twist1a/b already
discussed (Germanguz et al., 2007). Of these four genes, twist1a, twist1b, and twist3 are
expressed in the branchial arches during early zebrafish development. So, twist3, which does not
have a mammalian homolog, may have redundant roles with twist1a/b in jaw development and
act as a “homodimer” partner. Thus, I used CRISPR/Cas9 technology to create a twist3 mutant
allele (Figure 4-1A). Bone and cartilage staining of twist1a; twist1b; twist3 triple mutants at five
days post fertilization revealed jaw cartilage defects more consistent of the severe category, as
compared to twist1a/b double nulls (Figure 4-1B). As the phenotype of the triple mutant is still
not as severe as the mouse neural crest conditional Twist1 null, there remains a possibility that
96

the zebrafish twist2 is playing a compensatory role. Future work assessing twist1a/1b/2/3
quadruple mutants will provide a complete story of Twist function in fish jaw cartilage
development. On the other hand, Twist1 may be forming heterodimers with other Tcf family
members. This motivated a look into the zebrafish E2A or Tcf3 homologs, tcf3a and tcf3b.
Nonsense mutant alleles for both genes have been identified, but phenotypes have not yet been
thoroughly assessed. Unpublished, preliminary work from Dr. Mia Brockop in Dr. Robert
Maxson’s laboratory suggested a genetic interaction between E2A and Twist1 in mouse coronal
suture development. This together with the previously discussed in vitro studies confirming
Twist1-E2A dimers proves E2A to be a promising candidate. Given that Tcf12 and E2A have
similar roles in T-cell development (Barndt et al., 2000), they may also have redundant roles in
craniofacial development. Indeed, I have obtained a Tcf12
+/-
; E2A
+/-
mouse that exhibits partial
coronal synostosis (Figure 4-2). The double heterozygous mutant was weak and runty at the time
of collection, which was approximately postnatal day 8-10. No animals of this genotype were
recovered at weaning age. Altogether, these models suggest strong genetic interactions and
support probable dimerizations for differential regulations.

97

Figure 4-1. Jaw cartilage defects in twist1a; twist1b; twist3 mutants

(A) The sites of nucleotide changes induced by CRISPR/Cas9 are shown for each mutant allele.
Schematics show the predicted protein truncations caused by the frame-shift mutations, relative
to the DNA-binding basic helix-loop-helix domain (HLH, purple).
(B) Qualitative scoring of facial skeletal defects from Grade 0 (unaffected) to Grade 4 (most
affected). Loss of one or two copies of twist3 enhanced the facial skeletal phenotype of twist1a
-/-
;
twist1b
-/-
mutants. Wild type (n = 20), twist1a
-/-
; twist1b
-/-
(tw1a
-/-
; tw1b
-/-
; tw3
+/+
, n = 17),
twist1a
-/-
; twist1b
-/-
; twist3
+/-
(tw1a
-/-
; tw1b
-/-
; tw3
+/-
, n = 11), twist1a
-/-
; twist1b
-/-
; twist3
-/-
(tw1a
-/-
;
tw1b
-/-
; tw3
-/-
, n = 4).

98

Figure 4-2. Craniosynostosis detected in Tcf12
+/-
; E2A
+/-
mouse

Bone staining of litter mates at roughly 1-1.5 weeks old reveal partial loss of the coronal suture
in Tcf12
+/-
; E2A
+/-
mutant (n =1). Lower panels are enlarged images of corresponding boxed
regions. Arrow indicates synostosis.

To examine the physical interaction between Twist1 and Tcf12 proteins, new models with
detectable fusion proteins are necessary. As current antibodies targeting Twist1 and Tcf12 have
not been shown to be sufficiently specific, tagged alleles would prove valuable for future
molecular studies. Not only would tagged alleles allow for dimer detection, but also the alleles
combined with next generation sequencing could definitively reveal gene sets that are co-
regulated and individually regulated by Twist1 and Tcf12. Furthermore, adding different tags for
the same protein would allow for interrogating both heterodimers and homodimers. As a
previous study by (Connerney et al., 2006) suggested, the ratios of homodimers to heterodimers
may be different at the osteogenic front versus the suture. Tagged alleles would allow for studies
to confirm this model.
4.2 Coronal suture sensitivity to genetic changes
To date, a number of genetic mutations have been identified to cause craniosynostosis, often as a
phenotype of developmental disorders. Coronal synostosis is a principle characteristic of
Saethre-Chotzen syndrome; and mutations in TWIST1 have been linked to roughly 70 percent of
the cases (el Ghouzzi et al., 1997; Howard et al., 1997). More recently, human geneticists have
99

found mutations in TCF12 in Saethre-Chotzen syndrome patients without TWIST1 mutations
(Sharma et al., 2013), partially explaining the remaining 30 percent of Saethre-Chotzen
syndrome cases. Mutations in JAGGED1 and NOTCH2 contribute to approximately 80 percent
of Alagille syndrome cases, which often present with liver malfunction, heart defects, and facial
abnormalitites, and a subgroup of these patients has coronal craniosynostosis (Kamath et al.,
2002). Fgf signaling also has roles in proper development. Various mutations in FGFR2 or
FGFR3 result in Apert syndrome, Crouzon syndrome, Beare-Stevenson cutis gyrata syndrome,
and Muenke syndrome – all of which is characterized by coronal synostosis (Teven et al., 2014).
Distinct facial features additionally accompany each of these syndromes. Furthermore, numerous
genetic mutations are responsible for non-syndromic craniosynostosis cases, which are equally,
if not more, prevalent (Morriss-Kay and Wilkie, 2005). As seen from these examples of
syndromic craniosynostosis, the coronal suture is affected at a higher frequency compared to
other sutures. What then makes the coronal suture so sensitive to genetic perturbations?

My research has ruled out the neural crest-mesoderm boundary as well as differential twist1b and
tcf12 expression to be reasons for coronal suture sensitivity; however, one still plausible
explanation is a natural differential distribution of osteoprogenitors at the various sutures. In the
mutant condition where progenitors are presumably prematurely differentiating into osteoblasts,
the progenitor pool at the coronal suture may be very quickly depleted, resulting in the
neighboring frontal and parietal bone matrix fusing without cells to maintain the separation.
Conversely, if the progenitor pool is larger at the metopic and sagittal sutures to begin with, then
premature progenitor depletion at these sutures may not have as extreme of an effect. The
remaining progenitors would still be able to self-renew and allow continuous bone growth while
keeping the paired frontal bones or paired parietal bones apart (Figure 4-3). As I have now
shown that progenitor markers that were identified in postnatal mouse sutures are indeed
conserved in the zebrafish developing and mature suture, future work in utilizing these markers
to trace and quantify the progenitor population over time may offer evidence to support or refute
this model. Moreover, tracing these markers may reveal any differences in expression pattern or
timing of pattern in the different sutures. As some genes, such as Shh and Ptc, have been
previously reported to be restricted to the metopic and sagittal sutures, it is likely that there are
inherent differences amongst the cranial sutures.

100

Figure 4-3. Model of differential progenitor pools

Cartoon models of developing wild-type and mutant zebrafish skulls. Gray represents bone, and
green progenitor cells.

4.3 References
Barndt, R.J., Dai, M., Zhuang, Y., 2000. Functions of E2A-HEB heterodimers in T-cell
development revealed by a dominant negative mutation of HEB. Molecular and cellular biology
20, 6677-6685.
Bialek, P., Kern, B., Yang, X., Schrock, M., Sosic, D., Hong, N., Wu, H., Yu, K., Ornitz, D.M.,
Olson, E.N., et al. (2004). A twist code determines the onset of osteoblast differentiation.
Developmental cell 6, 423-435.
Connerney, J., Andreeva, V., Leshem, Y., Muentener, C., Mercado, M.A., and Spicer, D.B.
(2006). Twist1 dimer selection regulates cranial suture patterning and fusion. Developmental
dynamics : an official publication of the American Association of Anatomists 235, 1345-1357.
el Ghouzzi, V., Le Merrer, M., Perrin-Schmitt, F., Lajeunie, E., Benit, P., Renier, D., Bourgeois,
P., Bolcato-Bellemin, A.L., Munnich, A., Bonaventure, J., 1997. Mutations of the TWIST gene
in the Saethre-Chotzen syndrome. Nature genetics 15, 42-46.
Firulli, B.A., Krawchuk, D., Centonze, V.E., Vargesson, N., Virshup, D.M., Conway, S.J.,
Cserjesi, P., Laufer, E., and Firulli, A.B. (2005). Altered Twist1 and Hand2 dimerization is
associated with Saethre-Chotzen syndrome and limb abnormalities. Nature genetics 37, 373-381.
Funato, N., Ohtani, K., Ohyama, K., Kuroda, T., and Nakamura, M. (2001). Common regulation
of growth arrest and differentiation of osteoblasts by helix-loop-helix factors. Molecular and
cellular biology 21, 7416-7428.
101

Funato, N., Twigg, S.R., Higashihori, N., Ohyama, K., Wall, S.A., Wilkie, A.O., and Nakamura,
M. (2005). Functional analysis of natural mutations in two TWIST protein motifs. Human
mutation 25, 550-556.
Germanguz, I., Lev, D., Waisman, T., Kim, C.H., Gitelman, I., 2007. Four twist genes in
zebrafish, four expression patterns. Developmental dynamics : an official publication of the
American Association of Anatomists 236, 2615-2626.
Howard, T.D., Paznekas, W.A., Green, E.D., Chiang, L.C., Ma, N., Ortiz de Luna, R.I., Garcia
Delgado, C., Gonzalez-Ramos, M., Kline, A.D., Jabs, E.W., 1997. Mutations in TWIST, a basic
helix-loop-helix transcription factor, in Saethre-Chotzen syndrome. Nature genetics 15, 36-41.
Kamath, B.M., Stolle, C., Bason, L., Colliton, R.P., Piccoli, D.A., Spinner, N.B., Krantz, I.D.,
2002. Craniosynostosis in Alagille syndrome. American journal of medical genetics 112, 176-
180.
Morriss-Kay, G.M., Wilkie, A.O., 2005. Growth of the normal skull vault and its alteration in
craniosynostosis: insights from human genetics and experimental studies. Journal of anatomy
207, 637-653.
Sharma, V.P., Fenwick, A.L., Brockop, M.S., McGowan, S.J., Goos, J.A., Hoogeboom, A.J.,
Brady, A.F., Jeelani, N.O., Lynch, S.A., Mulliken, J.B., Murray, D.J., Phipps, J.M., Sweeney, E.,
Tomkins, S.E., Wilson, L.C., Bennett, S., Cornall, R.J., Broxholme, J., Kanapin, A., Whole-
Genome Sequences, C., Johnson, D., Wall, S.A., van der Spek, P.J., Mathijssen, I.M., Maxson,
R.E., Twigg, S.R., Wilkie, A.O., 2013. Mutations in TCF12, encoding a basic helix-loop-helix
partner of TWIST1, are a frequent cause of coronal craniosynostosis. Nature genetics.
Teven, C.M., Farina, E.M., Rivas, J., Reid, R.R., 2014. Fibroblast growth factor (FGF) signaling
in development and skeletal diseases. Genes & diseases 1, 199-213.

Abstract (if available)

Abstract Development of the skull is a lengthy process, which, if disrupted, results in craniofacial anomalies that are often associated with birth defects. The work presented in this dissertation discusses two particular elements of the skull—cranial sutures and middle ear ossicles. Cranial sutures are physical boundaries that separate the skull bones and also house stem cells for bone growth. The coronal suture is one that appears to have shifted multiple times in relation to germ layer boundaries during evolution. I review these boundaries in the context of development, evolution, and disease. The coronal suture is commonly affected in a human birth defect called craniosynostosis, and this is a defining characteristic in Saethre-Chotzen syndrome. I provide original evidence that human, mouse and zebrafish endure coronal suture loss when the genes TCF12 and TWIST1 are mutated, despite evolutionary differences in skull development of these three species. Using the zebrafish model, I find differential skull bone growth in mutants. As part of my dissertation, I also studied the middle ear ossicles, a chain of three small bones that amplifies and transmits sound waves to the inner ear. While these are unique mammalian structures, they are homologous to jaw cartilages of zebrafish. Previous work from my laboratory confirmed a role for Jagged-Notch signaling in shaping skeletal elements derived from the first two pharyngeal arches of zebrafish. My work shows a conserved requirement for Jagged1-Notch2 signaling in patterning the stapes and incus middle ear bones derived from the equivalent pharyngeal arches of mammals. Overall, my findings reveal deep genetic conservation in cranial suture development and pharyngeal arch patterning from fish to mouse to man.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Jagged1 functions downstream of Twist1 in the specification of the coronal suture and the formation of a boundary between osteogenic and non-osteogenic cells

PDF

Jagged-notch signaling: patterning the vertebrate face

PDF

TWIST1 functions in both mesoderm and neural crest derived cranial tissues to establish and maintain coronal suture patency

PDF

Elucidating the role of neural crest specific Stat3 signaling in maintaining coronal suture patency during embryonic development

PDF

Development and evolution of skeletal joints: lessons learned from studying zebrafish

Asset Metadata

Creator Teng, Camilla Sue (author)

Core Title Utilizing zebrafish and mouse models to uncover the underlying genetics of human craniofacial anomalies

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Genetic, Molecular and Cellular Biology

Publication Date 04/27/2021

Defense Date 11/28/2018

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

Tag Alagille Syndrome,coronal suture,craniofacial,craniosynostosis,Development,genetics,human,Jagged1,middle ear ossicles,mouse,Notch2,OAI-PMH Harvest,Saethre-Chotzen syndrome,skull bones,stapes,Tcf12,TWIST1,zebrafish

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Segil, Neil (committee chair), Crump, Gage (committee member), Maxson, Robert (committee member), Merrill-Brugger, Amy (committee member)

Creator Email camilla.teng@gmail.com,cteng@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-153898

Unique identifier UC11660978

Identifier etd-TengCamill-7332.pdf (filename),usctheses-c89-153898 (legacy record id)

Legacy Identifier etd-TengCamill-7332.pdf

Dmrecord 153898

Document Type Dissertation

Format application/pdf (imt)

Rights Teng, Camilla Sue

Type texts

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

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

Repository Name University of Southern California Digital Library

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