THE CONSTRUCTION AND CHARACTERIZATION OF HYPOXIA RESPONSAE REPORTER GENES FOR USE IN TRANSGENIC MICE Lorraine Tarnar Howard A thesis submitted in conformity with the requirements for the degree of Master of Science Graduate Department of Molecular and Medical Genetics University of Toronto O Copyright by Lorraine Tamar Howard 2001
127
Embed
CONSTRUCTION AND CHARACTERIZATION HYPOXIA …collectionscanada.gc.ca/obj/s4/f2/dsk3/ftp05/MQ63131.pdf · 3 by diffusion. As the embryo continues to grow, it reaches a point where
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
THE CONSTRUCTION AND CHARACTERIZATION OF HYPOXIA
RESPONSAE REPORTER GENES FOR USE IN TRANSGENIC MICE
Lorraine Tarnar Howard
A thesis submitted in conformity with the requirements for the degree of Master of Science
Graduate Department of Molecular and Medical Genetics University of Toronto
O Copyright by Lorraine Tamar Howard 2001
---- * - ..2 uisi@ns and Acq$sitiom et ~ o g n p h i Sewices se-s 6BTo(lraphtqu8s
The author has granted a non- exclusiw licence aUowing the National Library of Canada to reproduce, loan, distniute or sell copies of this thesis in microform, paper or electronic formats.
The author retains ownership of the copyright in ÜUs thesis. Neither the
L'auteur a accordé une Licence non exclusive pemetiant à la Biôli0thé:que nationale du Canada de reproduire, prêter, distribuer ou vendre des copies de cette thèse sous la forme de microfiche/nlm, de reproduction sur papier ou sur format Clectronique .
L'autew conserve la propriété du droit d'auteur qui protège cette thèse.
thesis nor substantial extracts fkom it Ni la thèse ni des extraits substantiels may be printed or otherwise de celle-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son penni ssion . autorisation.
Abstract
-=-A
THE CONSTRUCTION AND C H ~ C T E ~ Z A T I O N OF HYPOXIA RESPONSIVE
REPORTER GENES FOR USE IN TRANSGENIC MICE
Lorraine Tamar Howard
Master of Science 200 1
Molecuhr and Meàical Genetics
University of Toronto
Evidence in the literature suggests that conditions of low oxygen may have a role in the
development of the vasculature. To examine the areas of hypoxia in a deviloping mouse
embryo, traasgenes were developed coupling hypoxia responsive elements (HRE) to exogenous
promoter-reporter cassenes. Transient expression assays with these constructs show that HRE
sequences fiom different sources provide significantly different levels of induction in constant
backgrounds. Furthemore, evidence is presented to show that an HRE is not always sufficient
to confer hypoxia-sensitive activity to a transgene. Finally, preliminary evidence demonstrates
that two HRE transgenes show activity under physiologically relevant oxygen concentrations.
The use of HRE-based transgenes in the determination of embryonic hypoxia is discussed.
Acknowledgements
"You look-attrees ancf tabefttiemjust sa, (for trees are 'trees' and growing is 'to grow')"
"Yet trees are not 'trees', until so named and seen" "He sees no stars who does not see hem f h t of living silver.. ."
-- J.R.R. Tolkien, "Mythopoeia"
There are so many people that 1 would like to thank for helping me over the course of this
degree, but foremost among them is my supervisor Janet Rossant. Janet, thank you for your
support and guidance over the past three years - it bas been an incredible experience workiog in
your lab, and I consider myself extremely fortunate to have had the chance to leam from you.
Thank you also for giving me the chance to figure out what it is that 1 want to do: the most
strongly held ideas are not always the best ones. Many thanks also go to my cornmittee members
Alan Cochrane and Andras Nagy for their advice and guidance, especially when 1 hit the tough
parts. Th& you for helping me leam to become a better student.
Thanks also to al1 of the members of the Rossant lab for your fkiendship, advice, and
encouragement-you guys are great! Special thanks go to Masatsugu Ema, Laura Corson, Dan
Strumpf, and Perry (Tex?) Liao for helping me to leam several techniques, and providing
reagents, but especially for giving me moral support, advice, and encouragement!
My thesis could not be complete without mentioning the bunch at Knox College who have
made the last three years unforgettable. Among these are "Third East", including veteran
damsels Jen, Doma, Jess, and Kim, "the Crafiers" Susan, Claudia, and Rebecca, and al1 of the
people past and present who believed in the wonders of blue mice. Thanks to Daniela D'Anie110
for going above and beyond countless times. To my friend and fellow lab rat Albert Chang,
thanks for keeping me sane, especially when "sane" was somewhat relative. To my good nlend
Sheela Rupal thank you for being there, 1 kaow that you will go far along whatever path you
choose.
Finally, thank you to my family for al1 of your love, encouragement, and support. Whether it
came in the form of cwkies, a c raq phone call, deep thoughts, or a holiday visit, 1 appreciate it.
I love you-this one is for you.
iii
TABLE OF CONTENTS
................................................................................................................ II . Literature Suwey .4
2 . The Drosophila trachea as a mode1 of branching morphogenesis ................................... 13
3 . Molecular determinants of mammalian vascular patteming ............................................ 25
4 . The hypoxic response in mammalian vascular development .......................................... 39
................................................................... III. Introduction to the experimental approach ..O4
CHAPTER 2: THE CONSTRUCTION AND CHARACTERIZATION OF HYPOXIA
RESPONSIVE REPORTER GENES FOR USE IN TRANSGENIC MICE
II . Discussion ..................................... ~ . o o ~ ~ o o o * o o ~ o a m ~ ~ ~ o o o o * * * o * o * o o m ~ ~ ~ m o a ~ m ~ ~ o m ~ ~ ~ o ~ m o o ~ ~ o o o ~ ~ ~ m ~ ~ ~ ~ o m o o o o o m o o o m m o o o o o ~ o 9 3
1 I. Future Directions ............................................................................................................. 113
............................................................ 1 . An examination of HRE structure and fhction 113
............................... 2 . The optimization and implementation of an HRE based transgene 114
3 . Identification of areas of hypoxia during normal embryonic development ................... 117
III . Final Commenfg.. ......mm.................................................m...m.............................................. 120
N . References ....................................................................................................................... 121
CHAPTER 1
INTRODUCTION
1. General Introduction and Rationale: - .. . .. - - - - - -
The vasculature is one of the most important and complex organs in the mammalian body.
The first fuactional organ to form during embryonic development, the intricately branched
nehkrork of endothelial, and supporting periendothelial cells is essential for the transportation of
oxygen and nutrients to, and the removal of waste products nom the tissues. Serious dismptions
in the formation of the vascular network are lethal early in post-implantation development, while
the maintenance of vessel integrity and the control of vessel physiology and hemodynamics have
important consequences throughout embryonic and adult life.
The identification and characterization of genes involved in the development and maintenance
of the vasculature is a diverse and rapidly expanding field with numerous applications to medical
research, and the treatment of disease. Several hereditary conditions, including venous
malformations, and the Alagille syndrome have been shown to be caused by defects in genes
involved in the development of the vasculature; other diseases such as diabetic retinopathy,
ischemia, and arthritis, also have important vascular components. Especially interesthg is the
discovery that growing tumours and their metastases cospt host blood vessels to allow growth
beyond a defined size; clinical trials of anti-angiogenic cancer therapies are ongoing, with many
more to be tested in the near fiiture.
Clearly the development of a healthy, fuoctioning vasculature is of critical importance to the
survival of both embryo and adult; despite this fact, many questions remain unanswered. How
are vessels pattemed? What mechanisms ensure that al1 tissues have access to the blood? What
genes are involved in the differentiation, development, and maturation of the vascular network?
Examination of the development of other branched structures, such as the mammalian lung, and
Drosophila trachea show that both intriasic programs and extrinsic cues are required to ensure
correct delivery of oxygen to the tissues. 1s the mammalian vasculahire pattemed solely by
intrinsic developmental programs, or cm it respond to extrinsic cues encountered over the course
of embryogenesis? 1s then a role for hypoxia in normal vascular development?
The idea that hypoxia may have a role in the developing embryo is intuitively attractive. In
the early (pre-implantation) stages of development embryonic cells obtain oxygen and nutrients
3 by diffusion. As the embryo continues to grow, it reaches a point where diffision is insufficient
to supply the tissues, forcing it to develop a more efficient method of transport. The formation a - - - -
of the yolk sac, with its extensively branched vasculatwe, is one method by which a growing
embryo can obtain oxygen and nutrients; the large surface area of the yolk sac allows diffision
of oxygen into the vasculatwe, for transport into the embryo. As the embryo continues to grow,
this too becomes insufficient, and later stages of development connect the embryo to the
matemal circulatory system through the placenta. Given this context, it is reasonable to
hypothesize that hypoxic conditions arising nahually during development might have a role in
initiating the early stages of differentiation, or guiding some of the later branching events which
occur in the developing embryo. In this thesis is descnbed the construction and in vitro analysis
of a senes of transgenes designed to examine the extent and localization of hypoxic tissues in the
developing mouse embryo.
II. - - Literature Survey 1. Vascular Embryology
Throughout human history, man bas speculated on the role of the circulatory system. Once
thought as a seat for emotion and reason, later as the source of a body's "humours", work
throughout the twentieth cenhüy has shown the cardiovascular system to be a complex,
intricately branched network. Developed through a combination of intrinsic developmental
programs, extrinsic adaptation, and hemodynamic constraints, a functioning vasculature is
essential to ensure sufficient oxygen and metabolites to al1 of the tissues in the body.
Although many groups are currently working to identiQ the specific molecular mechanisms
involved in mammalian vascular development, the fundamental principles of vesse1 growth were
discovered using the chick. Light and electron micrograph analyses, coupled with the
production of quaii-chick chimeras, pioneered by F. Dieterlen-Lievre, and N. Le Douarin, have
been critical to understanding the basic processes, morphogenetic movements and lineages that
occur to produce the vasculature lo4. Current ideas hold that a primaty capillary plexus, formed
by vasculogenesis, is remodeled through sprouting angiogenesis to form a mature branched
network (Figure 1). Furthemore, observations made on chick vascular development, coupled
witb recent embryological work by the lab of Dieterlen-Lievre have shown a close association
between the endothelial and hematopoietic lineages in specific areas of the developing chick.
This observation is one of several pieces of evidence supporting the existence of the
hemangioblast, a bipotential precursor capable of forming bodi the endothelial and hematopoietic
Iineages.
Early work on vascular development was performed by Florence Sabin in the early 1900s
I .~rom ber detailed studies of the chick blastoderm she proposed the existence of angioblasts,
specialized cells easily distinguishable fkom the surroundhg mesencbyme, that could
differentiate into cells of both the endothelial and hematopoietic lineages. She also observed
"masses of cells that.. . develop hemoglobin and become eythroblasts," which she terrned blood
islands. Interestingly, Sabin concluded that vascular lumiiia were derived fkom cytolysis and the
production of an intracellular space, rather than h m an extracellular origin.
Figure 1 : A schematic o v e ~ e w of vascular developrnent
Hematopoietic
precursor
Hematopoietic
lineages
Hemangioblast
Endothelial precursor
Primary capillary plexus
Mature branched network
- . 6
Figure 1 :An ovewiew of the stages in vascular development. A multipotential precursor cell, the
a -- hemangioblast, - produces cells - of both - the hematopoietic and endothelial lineages. Endothelhl cells migrate and divide to form the primary capillary plexus. This primitive
vascular network is remodeled to form a branched network. The association between the
endothelium and smooth muscle cells is omitted for clarity. 39
Further detail on vascular development - - would wait until the advent of electron microscopy
and the work of Gonzalez- Crussi, Hiruma, and Hirakow 295. Study of thin sections of 6-16
somite stage embryos confirmed the formation of angioblastic clusters, with lumina existing as
regions of extracellular space, gradually enclosed by endothelial cells. Gonzalez-Cmssi also
described the formation of vascular plexi, onginally lacking a basement membrane, that are
remodeled as development progresses 2. Furthemore, arterial pressure caused by flowing blood
was detected in vessels formed solely fiom endothelial cells; arguments were made that later
changes in vessel structure could be caused in part by hemodynamic pressure.
Observations made of endothelial ce11 biology have shown that immature endothelial cells
have a rounded morphology, with a small surface-area to volume ratio 2,596. As the vesse1
matures, the endothelial cells elongate to fom thin-walled tubes. These new vessels are fragile,
and gradually becorne smunded by periendothelial support cells. Detailed study of vessel
morphology has s h o w that the endothelium fashions two types of tubes. Larger vessels have
lumina originally derived fiom extracellular space, bounded by junctions behveen neighbouring
endothelial cells, or an endothelial ce11 with itself. Such contact results in a "seamed" vessel. It
has been s h o w that "unseameà" capillaries lack these areas of cell-ce11 contact, and are thought
to have intracellular lumina. The formation of multicellular and unicellular endothelial branches
is not unlike the formation of the Dmsophila tracheal network described by Shilo and Krasnow
798 (described below).
A new tool for dissecting apart the mechanisms behind vascular growth was described by the
lab of Fmcois Dieterlen-Livre in 1987 9. The QH1 monoclonal antibody was found to bind to
quail, but not chicken endothelial and hematopoietic cells. At the time of this study, it had been
shown that clusters of endothelial and hematopoietic cells, known as blood islands, arose in the
extraembryonic tissues (area opaca), and that vessels were later observed in the developing
embryo (area pellucida) 10. Pardanaud et al. used QHI immunohistochemistry on quai1
blastodiscs, to confirm that: 1) ~ ~ l ' c e l l s could be fomd in blood islands, which interconnected
to fonn the extraembryonic vasculature. 2) The embryonic vasculature fomed separately fiom
that of the extraembryonic regions, eventually connecthg to form a complete network. These
8 results confirmed, and extended earlier work on quail-chick yok sac chimeras, that had shown
tbat the embryonic and extraernbryonic vasculahire formed independently of one another.
A second important experimental tool had its roots several years earlier. In the late 1960's,
Nicole Le Douarin observed that the interphase nuclei of quai1 cells had a large aggregation of
heterochromatin in the nucleus, making them easily distinguishable nom chick cells 3. This
difference laid the foundation for a vast array of quail-chick grafting experiments; by replacing,
or inserting a piece of quai1 tissue into a chick embryo, and allowing it to develop, one could
determine to which tissues the descendants contributed. Orthotopic or heterotopic grafis could
be performed allowing cornparisons to be made between different graft tissues in a standard
environment 1 1 . Additionally, the heritable label provided by the quai1 nucleoli was
advantageous for fate mapping experiments, as it would not be diluted through ce11 divisions,
unlike the conventional fluorescent dyes 12.
By the late 1980's it was known that that endothelial cells could differentiate in situ to fom
vascular plexi, through a process termed vascuiogenesis. Additionally, Judah Folkman had
observed a second process of vascular modeling, in which new endothelial tubes sprouted fiom
existing vessels 13. This phenomenon, termed sprouting angiogenesis, had been extensively
snidied in tumours, but it was not as well understood in the embryonic context. Some models of
vascular development held that the bone marrow, brain, and kidney were vascularized by
angiogenesis, but it was not known if vasculogenesis occurred concurrently with angiogenesis in
the developing embryo 4. A paper published in 1989 by the Dieterlen-Lievre lab took advantage
of the Q H ~ ' antibody and the nucleolar propertîes of quai1 cells to address some of these
questions, and examine the methods by which tissues were vascularized 4.
With these points in mind, Dieterlen-Lievre's group performed a series of chicldquail grafts
to examine the "angiogenic potential" of different types of tissue. In the first experiments pieces
of chick limb bud, representative somatopleute (ectoderm/mesodemi) derived tissue, were
grafted onto a quai1 embryo 4 (Figure 2). Interestingly, when the embryos were sacrificed a few
days later, the graft tissue was vascularized by QH 1 + endothelial cells. Clearly the chick explant
had been vascularized almost exclusively by quai1 host endothelium. The reverse experiment, a
quail graft into a chick host confhed these observations, as the somatopleuric quai1 grafi was
Figure 2: Schematic diagram of chick embryo (transverse section)
-.Lw-.- -A.. - - -
1 Ec todem
Neural tube
Somatic lateral plate
mesodenn
Coelom
Splanchnic lateral plate mesoderm
Endodenn
Dorsal aorta - Notochord
10 Figure 2: Transverse section of a chick embryo, (Stage 11) showing the arrangement of
-A--- - splanchnic and somatopleuric - lateral - plate mesoderm relative to the endoderm, ectordem, and donal aorta. Figure is adapted fkom Plate 1 Id, Bellairs and Osmond
1998. 142
vascularized by the QH1' chick endothelium. Expenrnents performed with splanchnopleuric - - - - - - - =' - - -
(endoderm/mesodem) organ rudiments gave the opposite result. Embryos with sections of chick
splanchnopleure in a quai1 host gave a chick-denved vascular network; a quail gr& produced a
~ ~ l ' ( q u a i 1 denved) network in chick. A final experiment using a splenic graft (mesodennal
origin) showed similar results to that of the splanchnopleure grafts; the vasculature was able to
extend out fiom the graA and form chimeric vessels with the host endothelium. In al1 cases,
quail endothelial cells were visualized by QHI+ immunoreactivity, while non-endothelial grafted
tissue could be differentiated through the nucleolar stain. In the conclusion of their 1989 paper,
Pardanaud et al. claimed that, "rudiments composed of mesoderm and ectodenn are sites for
By the early 1990s, it had become clear that vasculogenesis and angiogenesis were processes
that occurred in parallel during development to produce the mature vasculature. Extending this
idea, the lab of Francois Dieterlen-Lievre demonstrated that the mechanism of vascularization
depended on the origins of the tissue 14915. Somatopleural tissues, such as the body wall and
limbs, were vascularized by external angioblasts, which migrated into the tissue to form a
vascular network. Intemal organs of splanchnopleuric origin (heart, h g , digestive organs)
would be vascularized by intrinsic endothelial precursors. In the mid 1990s, Pardanaud et al.
worked to M e r dissect the mechanisms of vascular development and their relationship to
hematopoiesis.
A year later, Pardanaud et al. examined the vasculogenic potential of segmental plate
mesoderm, taterat plate mesodem, and tait bnd grafts, concindmg that there were two separate
lineages of angioblastic cells which contributed to the embryo vasculature 16. The tust lineage,
a line of angioblastic cells derived fiom the sornites, and paraxial mesodem, were capable of
produciag endothelial cells, that colonized the body wall, the kidney, and die wall and roof of the
dorsal aorta. A second cell lineage, produced fkom the splanchnopleuric mesoderm was capable
of forming both endothelial and, in specific regions, hematopoietic cells; splanchnopleure-
derived angioblasts were capable of colonizing the visceral organs, and floor of the dorsal aorta,
in addition to the areas of the somatopleure. It bad long been known that the dorsal aorta was
one of the intraembryonic sites of hematopoiesis; micrographs and other studies of the vesse1
12 showed that the ventral (floor) of the vessel produced clusters of cells, which changed
morphology,-and took on some of thecharacteristics of hematopoietic cells 2. The finding that
splanchnopleuric, but not somatopleuric mesoderm could contribute to these lineages was an
important step towards understanding the origins of the vasculature, and the relationship between
the endothelial and hematopoietic lineages.
A fmal piece of work combined some of the accumulating data on molecular replators of
vascular development with the (now-classic) embryological techniques. In the most recent
experiments, the lab of Dieterlen-Lievre bave used the observation that only splanchnopleure-
derived angioblasts cm colonize the visceral organs and floor of the dorsal aorta, to assay the
developmental potential of treated somitic tissue '. As shown previously, angioblasts fkom
somitic tissue were only capable of colonizing somitopleunc tissues; such angioblasts were
unable to colonize the visceral organs, nor were they found to contribute to the endothelial or
hernatopoietic lineages on the floor of the dorsal aorta. By transiently culturing somitic tissue
with endodenn ptior to grafting, Pardanaud et al. found that they were able to change the
properties of the angioblasts, making them capable of colonizing the visceral organs, and the
floor of the dorsal aorta. Interestingly, a similar effect was observed when somitic tissue was
cultured in the presence of VEGF, TGFBl, or bFGF.
Conversely, when Pardanaud et al. transiently cultured splanchnoplewic tissue with ectodenn,
EGF, or TGFa they reduced the potential of angioblasts to migrate fiom the explant 17.
Splanchnopleuric grafts treated in this way were unable to colonize visceral organs, nor were
they capable of contributing to the hematopoietic clusters on the floor of the dorsal aorta.
Currently there are believed to be two separate sources of endothelial cells in the chick embryo,
which have differeat potentials for invasion and hematopoiesis. in one lineage, angioblasts
derived fiom somatopleural and axial mesodem produce endothelial cells that can invade the
body wall and somatopleural tissues. A second lineage, derived h m the splanchnopleure, and
dependent on a transient contact with endoderm, vascularizes the visceral organs, and is capable
of contributing to hematopoiesis in specific regions of the dorsal aorta.
While electron microscopie analysis of the developing chick vasculature has provided some
useful information on the ultrastnictwe of a vessel, it was the work of Francois Dieterlen-Lievre
13 and CO-workers that has laid the foundations towards understanding the mechanisms in vascular
forqation and patteming. Through the c l ~ s i c embryological techniques of grafting and lineage
mapping, Dieterlen-Lievre et al. have described some of the g e m layer interactions,
differentiation, and migration, that must occur to form the endothelial network. From these
origins have sprung many of the studies of the marnmalian vasculature. The terni vasculogenesis
has corne to describe the mechanism by which endothelial cells differentiate and interact to form
networks in the extraembryonic tissues, and some areas of the embryo 18. Angiogenesis, a terni
originally describing the sprouting of branches fiom pre-existing vessels, has been expanded to
include the stages of branching, remodeling, and occasionally the maturation of the vessel; in
essence the changes that occur in the vascular plexus to produce the mature branched vascular
tree.
2. The DrosophiIu trachea as a mode1 of branching morphogenesis
While much work has been done to understand the fiuidaxnental principles behind vascular
development, many problems remain to be solved. Although we have an understanding of the
origins of endothelial precursors, it is not yet known how the vasculature is pattemed, how
lumina are fomed, or how the vasculature is remodeled to ensure that oxygen is delivered to al1
of the tissues. To address these issues, the labs of Mark Krasnow, Ben-Zion Shilo 8919 and
others have turned to Drosophila. The stereotyped pnmary branches, and intricate arborization
of the Drosophila trachea have proven to be valuable models for dissecting the formation of
branched networks in nature.
The Drosophila trachea is an intricately branched network of tubes used to convey oxygen
from the spiracles to the tissues of the Drosophila larvae; like the mammalian vasculature, the
Drosophila trachea is characterized by a series of stereotyped branches, and variable tracheole
formation. Growth and patternhg of the Drosophila trachea begins with the differentiation of
ten ectodermal clusters on each side of the developing embryo 8.20. These clusters invaginate to
form sacs comprised of approximately 80 cells; subsequent ce11 migration and branch formation
occurs without proüferation or apoptosis. Primary branch formation occurs with the migration of
tracheal cells to form the six major branches of each metamete: the dorsal branch, lateral tnink,
and ganglionic branches migrate along the dorsiventral axis, while the dorsal trunk and visceral
14 branches migrate anteriorally. Secondary branches form with the expression of pantip markers
in the terminal cells; these temiinal cells differentiate to fonn unicellular branches that invade the - -+L- - - - - --- - -- - - - - - - - - -
surrounding tissue. Finally, a finely ramified network of tubules are fomed nom the secondary
branches to become the terminal branches of the trachea. The formation of these seamless
tracheoles has been shown to be responsive to conditions of low oxygen in the surrounding
tissues. (Figure 3)
Three major types of cells comprise the Drosophilu trachea. Terminal cells form the
extensive network of intracellular tubes required to ensure oxygen delivery to the tissues 2 1.
Stalk cells provide the channels through which oxygen can pass, while fusion cells are required
to comect the segmented tracheal network 22. At least two fusion cells, located in the dorsal
trunk and dorsal branches send out fine processes to contact the fusion ce11 of an adjacent
tracheal metamere. In this way tracheal segments can fonn an interconnected network
throughout the embryo 23.
While early models of tracheal development proposed that modifications made to an iterative
genetic program could account for the formation of the complete tracheal network 24, current
ideas hold that several different pathways are required to produce a correctly branched,
functional, trachea. Furthetmore, work perfomed in the labs of Affolter, Shilo, and Krasnow
have shown that three distinct patteming mechanisms are active in tracheal development.
Firstly, it is known that tracheal cells are assigned to a specific branch, and that fusion cells are
specified prior to the onset of migration 22,23325. This intriasic control of tracheal branching is
produced through a combination of Dpp, EGF, and Notch signaling, resulting in the specification
of tracheal ce11 fate, and an alteration of the cellular response to directional cues. A second
method of conîrol is provided by the Brealless (btl) and Branchless ( h l ) genes. Tracheal cells
expressing the Breuihless FGF receptor migrate toward dynamic sources of the Brarchless FGF
ligand provided by the surrounding ectoderm 24; ectopic expression of Branchless results in
migration of tracheal branches towards the expressing cells. The chemotactic response of btl
expressing tracheal cells towards bnl expressing tissues shows that the trachea can be pattemed
by extrinsic cues. A third method of tracheal patternhg is a specialized form of extrinsic
response important in the migration of terminal branches. Tracheoles will grow toward areas of
low tissue oxygenation, through a modification of the bnllbtl pathway 2 1.
--A----- - - - Figure . .- . 3: An overview of the Drosophila trachea
A.
Dorsal Branch
Dorsal T d (anterior) Dorsal Trunk
(posterior)
Visceral branch
Ganglionic branc h Lateral
Tnink
Figure-% A-)A schematicdiagram ofmegmeat of thstncheaI- network. Differentiating tracbai
precmors form a cluster of approxiamately 80 cells, temed the tnicheal placode. In the
early stages of tracheal development, cells migrate nom the placode to produce a series
of stereotyped primary branches. The major branches produced from one such placode
are shown and labelled in A). The Drosophila tracheal network fonns from the
19 intercomection of 20 placodes. B) The formation of secondary and tertiary tracheal
branches. Terminal cells of a primary brancb form a pair of unicellular secondary
branches, tbat can rami@ into a network of subcellular tertiary branches. (Adapted fkom
Krasnow 1997). 8
i. Intrinsic patterning of the Drosophila trachea
The idea that tracheal ce11 fate was detemined pcior to branch formation began with the
observation that mutations in the Dpp receptors punt and thick vein (th) resulted in defects in
specific branches of the developing trachea 25. Embryos lacking tkv orpunt were able to
produce normal anterior-posterior branches, such as the visceral branch and dorsal trunk, but
were unable to produce a dorsal branch, and had defects in the branches migrating ventrally.
Ectopic expression of Dpp prevented the antenor growth of the dorsal trunk, and increased the
number of cells available to migrate dorsally. Studies of tkv and punt expression showed that
both genes were expressed in the placode prior to tracheal ce11 migration; conversely, the Dpp
ligand was expressed as a pair of stripes in the ectoderm dorsal md ventral to the tracheal pits.
Expenments performed with constitutively active and dominant negative tkv receptor confimed
that Dpp activation was necessary for tracheal cell migration. Interestingly, the time of Dpp
activity was important: ectopic Dpp receptor activation had no effect on tracheal patteming
during ce11 migration. From these data, Vincent et al. proposed that Dpp acts to regionalize the
tracheal placode, not as a chemoattractant.
A paper published by Ben Shilo's group a few months later reported that antagonism between
the EGF and Dpp pathways regionalized the tracheal placode 22. In basic EGF signaling, the
EGF receptor, der is b o n d by an active f o m of the spitz ligand. Inactive spitz is ubiquitously
expressed, and becomes functional only after cleavage by rhomboid (rho) and star; specificity of
the EGF signaling is conferred in part by the tightly regulated expression pattern of rho.
Interestingly, EGF signaling was implicated in üacheal patternhg when it was observed that rho
was expressed in tracheal pits; it was later found that mutations in rho resulted in tracheal
defects. Similarly, an examination of spi& group mutants showed that the dorsal trunk and
visceral branches were poorly developed, if not entirely absent. Finally, embryos deficient for
the EGF receptor (der&-'-) did not have correct dorsal tnink migration, and showed fusion
defects. By transgenically expressing the EGF receptor (der) in the trachea of embryos deficient
in der function, Wappner et al. were able to rescue dorsal trunk migration and fusion.
18 Several experiments demonstrated that Dpp and EGF act antagonistically to pattern the
@achea. Firstly ectopic local or trachea1 specific Dpp activation resulted in the loss of the dorsal
tnink, and a reduction in the visceral branch; furthemore, instead of remaining in the tracheal
placode, these cells contribute to the dorsal branch. In the converse experiment, activated spi&
did not alter branch identity in wild-type embryos. When the experiment was repeated in
embryos with reduced levels of Dpp signalhg (tkv +/O), Wappner et al. observed normal dorsal
trunk and visceral branch formation, with a corresponding reduction in the dorsal and visceral
branches. If the EGF and Dpp pathways were antagonistic to one another, then one would expect
the phenotype caused by a hypomorphic mutation in one pathway to be rescued by a reduction in
signaling by the other. To test this prediction, Wappner et al. examined the trachea of embryos
carrying hypomorphic alleles of punt and flb. Double mutant embryos had general
abnonnalities, but were able to produce a continuous dorsal t d . From these data Wappner et
al. concluded that the EGF pathway was essential to assign tracheal cells to the dorsal trunk and
visceral branch fate; the Dpp pathway detennined the cells that would form the dorsal and lateral
tnink branches. Thus the EGF and Dpp pathways act antagonistically to one another to confer
branch identity and pattern the cells of the tracheal placode.
A second exarnple of intriasic patteming came fiom the Sarnakovlis lab 23. In this paper, it
was reported that the decrease in dorsal bmch production in tkv " embryos was accompanied by
a decrease in expression of the fusion ce11 markers headcuse (hdc) and escargot (ex), and fusion
ce11 defects. To m e r examine this phenornenon, Steneberg et al. specifically disrupted Dpp
signaling in fusion cells, and found an increase in fusion defects. Conversely, ectopic activation
of Dpp in tracheal cells resulted in the production of extra fusion cells in a few of the dorsal, and
dorsaI trunk branches studied; fiom this work Steneberg et al. concluded that Dpp was capable of
inducing fusion ce11 fate in the trachea. Interestingly, Notch signaling became implicated in the
detemination of fusion ce11 fate an en'ancer trap screen for genes expressed in fusion cells
identified Delta (a Notch ligand). Using a temperature sensitive Notch mutant, it was observed
that trachea deficient in notch signaling produced additional fusion cells at the expense of stalk
cells. Conversely, ectopic expression of the notch receptor throughout the trachea resulted in
conectly branched metameres that were unable to fuse. Such trachea lacked the expression of
fusion ce11 markers. From these data, Steneberg et al. concluded that Notch-Delta signaling was
essential for the correct specificatioa of the nision cell. In this case, the Delta-expressing fusion
ce11 is thought to activate the Notch receptor on neighbouring cells, preventing them fiom
19 adopting the fusion ce11 fate. Samakovlis' group M e r hypothesize that activated Notch
prevents - - - stalk cells fiom responding to the - Dpp - signal. Although fùrther experiments must be
performed to test this model, it is clear that inûinsic responses to Dpp, Spitz, and Notch signaling
are essential to pattern the tracheal placode, and eaable fuiun branch cells to respond to the
correct developmental cues.
ii. Extrinsic patterning of the Drosophila trachea
While intrinsic mechanisms clearly play a role in establishing ceIl fates, they do not determine
the position of a branch relative to the suiroundhg tissue. Extrinsic patteming mechanisms, such
as the chernotactic responses mediated by Branchless/Breathless interactions are critical for the
correct development of the trachea. The Breathless gene was first identified by Shilo's group,
who noticed that embryos lacking the Breathless gene product were unable to produce a
branched trachea 26. Further studies of the btl locus showed that it encoded a homologue of
mammalian FGFR-1; homozygous mutations at this locus inhibited the correct migration, but not
the early differentiation of tracheal cells, in addition to causing defects in the glial cells of the
midline and the salivary glands 27,*8. Studies of downstream effectors of Breathless signaling
showed that the FGF receptor activates the RaslRaf pathway through DoflStumps 20. Further
work with embryos expressing mutant and cbimeric foms of btl demonstrated that 1 ) btl was
required for the onset of tracheal migration, but not for the determination of the tracheal placode;
2) btl was not required continuously for tracheal morphogenesis, but appeared to be required at
specific stages in the branching sequence 27. Finally, Reichman-Fned et al. concluded that btl
bad a "permissive" but wt an " i n s ~ t i v e " role in patterning the trachea.
While work to this point had demonstrated that btl was necessary for the formation of
primary, secondary, and terminal branches, it was not known if this requirement was for the
correct quantitative or spatial regulatioa of the receptor. Further studies on btl fùnction
demonstrated that constitutively active btl receptor expressed at high levels could not rescue
defects in btl deficient embryos 28. In fact, while tracheal-specific expression of wild-type btl
was sufiicient to rescue the tracheal defects in btl " embryos, expression of constitutively active
btl interfered with the rescue of such embryos. It was also obsewed that the trachea of wild-type
embryos were formed incorrectly on the addition of constitutively active btl receptor. Lee et al.
20 concluded fiom these and other experiments that the Breathless receptor tyrosine kinase is
requued for the cell migration fonning the primary, secondary, and terminal branches. In this - -- -- - - - .
paper, it is also speculated that Breathless activity may be required in a space-specific manner,
since hi@ levels of active Breathless are not sufficient to pattern the trachea.
Evidence for btl-mediated chernotaxis was described with the discovery and characterization
of the Branchless ligand, published by the lab of Mark Krasnow in 1996 24. Identified through
an enhancer-trap screen, embryos carrying strong bnl mutant alleles produced tracheal sacs, but
were unable to complete normal branching. Furthemiore, the bnl locus was haploinsufficient,
with heterozygotes also missing some branches. Sequence information, coupled with genetic
and biochemical tests suggested that bnl was an FGF homologue that could act through the btl
receptor to stimulate branching. The possibility of bnl mediated chemotaxis arose when it was
observed that bnl was expressed by ectodemal (and some mesodennal) cells surrounding the
tracheal sec, at the positions of fùture branch outgrowth. Further analysis of the Bnl expression
pattern has shown it to be highly dynamic: as tracheal branches grow toward bnl expressing
clusters, bnl expression is decreased in the original cells, and is activated in new areas.
Experiments in which bnl was ectopically expressed in wild type and bnl deficient embryos
showed that tracheal branches grew toward the source of bnl in most of the segments of the
embryo. While much more work needs to be perfomed to understand the genetic control of bnl,
it is clear that bnl-btl mediated chemotaxis is essential for the correct patteming of the
Drosophila trachea (Figure 4).
Wolf and Schuh have recently published an interesting twist on the idea of chemotactic
responses in the Drosophila trachea 2 . in theu paper, they observe that a single hunchback (hb)
expressing ceIl could be observed at the posterior-lateral margin of each tracheal cluster. In situ
analyses of this ceIl over several stages of branching morphogenesis suggest that it could connect
the posterior and anterior branches of the dorsal mink, behaving as a "bridge" to ensure proper
comection of the two branches. This hypothesis was supported by the finding that ectopic
expression of hunchback in a cell near the tracheal placode resulted in the misdirection of the
dorsal trunk branch, and the incomct fusion of the anterior and posterior branches. Further
support for an alternate chemotactic mechanism comes fiom studies of trachea formation in bnl,
Figure 4: Extrinsic patternhg of the Drosophila trachea
Figure 4: The tnicheal placode is patterned by extnnsic mechanisms. Branchless (black) is
- - - - -= expressed in cells surrounding the tracheal placode (white). As the tracheal branches
begin to grow toward areas of Branchless expression, Branchless is down regulated in the
original cells (broken lines), and upregulated in a new population of cells (black). 8
Figure adapted fiom Krasnow 1997. 8
23 btl, and hb embryos. Firstly, areas of dorsal tmnk formation have been observed in embryos
lacking h l OF btl funetim, dorsal tnink- bim defetts-have been noted in irb deficient embrycm.
Secondly bnllbtl expression is unaffected in hb deficient embryos, just as hb expression is
normal in bnl or btl deficient embryos. Finally, tunel staining of hb deficient embryos show that
the bridge cells form bnefly, and rapidly undergo apoptosis. From these data, Wolf and Schuh
propose that a hunchbuck expresshg bridge ce11 guides the migration and fusion of the dorsal
txunk, irrespective of the areas of bnl or btl expression; the absence of functional hb results in
bridge cell apoptosis and dorsal mi& fusion defects. While further work must be done to
support or refute the bridge ce11 hypothesis, it does provide new and intriguing possibilities for
the extriasic regulation of tracheal patteming.
iii. Hypoxia and extrinsic patteming
How is an organ modified to fulfill the specific requirements of the organism? Int~itively we
recognize that muscle structure and function can be altered by exercise, and that erythrocyte
content in blood is altered by chronic exposure to reduced oxygen levels *1,3*. How do organs
respond to chemical and physical cues fiom the surrounding environment? A seminal paper
published in 1999 demonstrated that conditions of low oxygen act to pattern the terminal
branches of the trachea; such a mechanism may have important implications for the patteming of
other branched organs, including the mammaüan vasculature 2 1.
To examine the effect of oxygen concentration on the production of terminal branches
exnbryos were grown tmâer 5%, 2t%, and 60.h oxygen and scored for the number of tenninal
branches 21. Through this work, Jarecki et al. demonstrated an inverse correlation between the
concentration of oxygen and the number of terminal branches; 68% more branches were formed
in embryos grown under 5% oxygen compared with those grown under 21% oxygen.
Additionally, branches produced in embryos grown under low oxygen tended to be long, highly
branched, and tortuous when compared witb branches produced under normoxia. By creating
tracheal clones deficient in terminal branching (blistered 3 or lumen formation (synuptobrevin - '-1, Jarecki et al. obsewed that branches fiom neighbouriog segments grew into areas that were
insufficiently oxygeaated by the mutant trachea. To begin to elucidate a mechanism, bnl
expression was examined during the stages of terminal branching. In situ analysis of bnl
24 expression during terminal branching showed that it was nomally expressed in a few cells in al1
tracheated tissues; ubiquitous expression of bnl resulted in tangled masses of tracheal branches. - -- - - -
These observations were consistent with the idea that bnl might be acting as a chemoattractant
for trachea terminal branches. Further expenments demonstrated that bnl protein is upregulated
in larva grown under 5% oxygen compared with siblings grown at 21%. More convhcingly,
areas of poor tracheation, presumed to be hypoxic, expressed increased levels of bnl. Although
more work must be done to understand the mechanisms by which bnl expression is up-regulated
by hypoxia, Jarecki et al. have demonstrated that conditions of low oxygen are instrumental in
patterning the terminal branches of the Drosophifu trachea. Such a finding may be important to
understanding the formation of the mammalian vasculahire.
Over the past decade, the Drosophila trachea has proven to be a valuable mode1 of the
formation of branched networks in vivo. Like the tracheal network, the mammalian vasculature
is comprised of large, multicellular primary, unicellular secondary, and subcellular tertiary
branches. Both the Drosophila trachea and the mammalian vasculature have structures that are
stereotyped between organisms, in addition to more variable components. From a physiologic
perspective, both systems use a semi-iterative tree structure to ensure an extensive area of
coverage, for the delivery of oxygen, and recovery of carbon dioxide. Furthemore, the
formation of a functional lumen of defined size is required for each system. Finally, both
systems are able to adjust to the tissue requirements of the organism after the organ has begun to
fuaction. While it is not yet known if the genetic mechanisms controlling tracheal development
will have direct parallels with those controlling mammalian vascular development, it is clear that
the general mechanisms of intrinsic, extrinsic, and possibly hypoxic regulation are important in
vascular dwelopmcnt and pattcrning. Thm rcprtstntative signaling pathways: VEGF, hg-Tic,
and Ephtin, will be discussed as examples of the current understanding of mammalian vascular
patteminp.
3. Mdecular determinants of mammalian vascular patternhg - =L - A -
i. The VEGF pathway
Originally identified as a vascular permeability factors, the VEGF family of growth factors
are a group of homodimeric glycoproteins whose carefully regulated activity is essential for the
formation and modeling of the vasculahire 31332. Four VEGF genes (A-D) have been identified
in humans, with a fiAh (VEGF-E) produced by members of the poxviridae 31933. Five isofonns
of the VEGF-A gene are produced by altemate splicing in humans; the mouse VEGF gene has
been shown to produce three isofonns of 120, 164, and 1 88 amino acids (a.a.) 34335. Of these
the 120, and 164 a.a. proteins are the most abundant, witb the 164 a.a. isoform acting as the
strongest mitogen. Additionally, it bas been found that the VEGF isoforms differ in the ability to
interact with heparin sulfate proteoglycans; the larger VEGF isoforms can bind heparin, and
associate with the extracellular matrix, while the smallest isoform has been shown to diffuse
freely (discussed in 31). It has been hypothesized that the different isoforms could act in
combination to mediate endothelial ce11 mitogenesis, differentiation, and proliferation.
VEGF ligands have been shown to bind to three major receptors, VEGFR-l/flt-1, VEGFR-
2 / f k 1, and VEGFR-3/flt4, in addition to at least one "accessory receptor", Neuropilin- 1 3 132.
The VEGFRI -3 receptors are characterized by the presence of seven immunoglobulin-like
domains used for binding to the VEGF ligand, and an inhacellular kinase domain (Figure 5).
VEGF binding induces homodimerization of the VEGFR-I and VEGFR-2 receptors 32, followed
by autophosphorylation and activation of the downstream signaling cascade. A single report
claims to show VEGF-mediated heterodimerization of soluble VEGFR-1IFlt-1, and the
extracellular domain of VEGFR-2/Flk-1, however, this finding is unsubstantiated, and its
relevance in vivo is unclear 36. Of the two most-well characterized teceptors, VEGFR-I and 2,
it bas been shown that VEGFR-1 has a ten-fold higher affiity for VEGF than VEGFR-2 32.
Interestingly, WGFR-1 undergoes little detectable phosphorylation when bound to VEGF, whiie
VEGF binding to WGFR-2 results in autophosphorylation at four major sites, followed by the
activation of the Raf/MapK pathway 31.32. In terms of the other receptors, Neuropilin-1 is
thought to act as a CO-receptor with VEGFR-2, while VEGFR-3 signaling is not well understood.
26 Figure 5: Schematic oveMew of the interactions between the VEGF receptors and their ligands
---- - - - Fig-2-1: A) Sepuence of the wild-tyge and mutated forms of the mouse VEGF hypoxia responsive element, with HIF- 1 consensus binding site in bold. B) Schematic diagram of
VEGF HRE Hsp Lac2 transgenes. Three copies of wild-type or mutated mouse VEGF
HRE were cloned into the Hind III (H) site 20 bp h m the start of the Hsp Lac2 cassette.
64 Table 1 : Oligonucleotide sequences and PCR pnmers used in the preparation of HRE containing
p & i Ï Ï a d e 1 ol'mber 1 Sequence
mVEGFWTHRE-F elements 1 1 1 ACGTTACACAGTGCATACGTGGGTTTCCACAGGTCGTCTCAC
To test whether lack of induction was due to the HRE or the particular promoter used, a
transgene was consûucted in which four copies of the EPO HRE and the SV40 promoter were
cloned before a Lac2 reporter. In al1 replicate experiments, cells transfected with the EPO HRE
SV40 LacZ, but not the SV40 Lac2 construct measurably induced LacZ activity under hypoxic
conditions (Figure 2-4).
71 Figure 2-4: The EPO HRE can activate the SV40 but not the Hsp promoter in hypoxic HeLa
-- - - cells
Activity of Lac2 transgenes in HeLa cells exposed to 1% or 20% oxygen
hEPO n=4 Hsp LacZ pBOS Construct
O Average 20% Average 1%
Activity of Lac2 transgenes in HeLa cells exposed to 1% or 20% oxygen
Average Average
Construct
Figure 2-4: A) A representative experiment in which EPO HRE Hsp LacZ or pBOS transgenes
were cotransfected with a constitutive Luciferase expression plasmid into HeLa cells.
Transfected cells were exposed to 1% or 20% oxygen, harvested, and assayed for p- galactosidase and Luciferase activity as described in the Materials and Methods. Bars
represent the absorbance for each construct averaged over two replicates after
standardizing for transfection and harvesting efficiency, using the Luciferase data. EPO
Hsp Lac2 activity was assayed in seven separate experiments; no evidence was found for
an oxygen-concentration dependent effect on EPO Hsp LacZ activity. Data shown are
fiom one of six replicate experiments. B) An experiment showing the activity of EPO
HRE SV40 LacZ, and SV40 LacZ transgenes in HeLa cells exposed to 1% or 20%
oxygen. As in A), p-galactosidase readings for 1% and 20% samples were standardized
for transfection and harvesting using the Luciferase data. Bars correspond to the average
absorbance of three replicates for each sample. (** indicates that activity under 1% and
20% oxygen is significantly different by Student's t-test, p<0.005.)
73 4. Minimal HREs from dlfterent KIF-1 target genes have diffennt responses in HeLa ceUs
exposed to low oxygen
Since the discovery of HIF-1, more than twenty-eight genes have been shown to be
upreguleted in hypoxic cells 10. Minimal hypoxia responsive elements have been identified in
some of these genes including the phosphoglycerate kinase (PGK-1) gene, the vascular
endothelial growth factor (VEGF), and the erythrocyte stimulating hormone erythropoietin
(EPO). As shown in Figure 2-5, many groups have demonstrated the sufficiency of these
elements to impart hypoxia inducibility on a transgene, although differences in experimental
conditions, reporter context, and ceIl lines make it impossible to compare them directly. Are
there differences between the activities of minimal HRE elements derived fiom different
sources?
To address this question constructs containing three copies of wild-type (wt) EPO, wt PGK,
wt VEGF, and mutated (mut) EPO HRE were cloned in consistent orientation and distance to an
SV40 Luciferase transgene. One of these constmcts, or a constitutively expressed SV40
Luciferase driven by the SV40 Enhancer @SVLuc+), was transiently transfected into HeLa cells
with a constitutively expressed bgalactosidase @BOS). As shown in Figure 2-6 the consûuct
containing three copies of the wild-type, but not the mutated EPO HRE was capable of mediating
4-fold induction of Luciferase activity. The construct containing three copies of the mouse
VEGF element gave variable activity, ranging fkom no induction in several cases, to 22-fold
induction on one occasion. (The average value is shown in Figure 2-6.) Neither the SV40
Luciferase, nor the CMV Luciferase demonstrated hypoxic induction. The addition of four or
three copies of EPO HRE, three copies of PGK HRE, three copies of VEGF HRE, or the SV40
enhancer produced significantly different inductions compared with that observed with SV40
Luci ferase alone @ varies from <0.0 1 (VEGF HRE), to <0.0005 (PGK HRE). The addition of
three copies of mutated EPO HRE produced no sipificant difference in induction when
compared with the SV40 Luciferase transgene. Sûikingly, the construct containing the wild-type
PGK element gave consistently higher induction than either the EPO H E or VEGF HRE
containing conshucts. (Figure 2-6: 52-fold compared with Cfold, and 9-fo ld) (p<0.0005).
Figure 2-5: An alignment of HRE elements used in the literatwe
2 Copies in TK GH 1 Copy in TK CGT 3 Copies in Hsp68 Lac2 1 Copy in TK Luci ferase 3 Copies in TK GH 1 Copy in SV40 Luci ferase 1 Copy of 372 bp fragment of IGFBPl intron 1 in Hsp7O Luci ferase
CeU Llne Induction Reference Hep 3B 50 Fold 8
HeLa I 8-9 Fold
R1 ES None This thesis
PC12 1 2.8 Fold 1 13 1 Hep G2 18 Fold 1 1 Hep 3B 34 Fold 14
---- - - - F i p - - 2.5: A survey of the literature showed - @at - - HF& sequences have been identified in many different target genes; nine representative elements are shown with details as to the
transgene and ce11 line in which they were tested. HIF-1 consensus binding sites are
highlighted, aad the EPO, VEGF, and PGK elements that were used in this study are
marked with an asterix.
76 Figure 2-6: HREs from different sources confer different induction on SV40 transgenes in
---,---A - - .. hypoxic HeLa cells
SV40 Luciferase activity of HeLa cells under 1% and 20% oxygen
1 2 3 4 5 6 7 8
Cons truc t
1 ~ ~ 4 0 Luciferase ~EPO n=4 SV40 Luciferase
Ipsv Luc+ ~ ~ E P O n=3 SV40 Luciferase
1 SV40 Luciferase
2 EPO n=4 SV40 Luciferase
3 CMV Luciferase
4 pSVLuc+
5 hEPû n=3 SV40 Luciferase
6 hEPO mut SV40 Luci feme
7 mWGF n=3 SV40 Luci ferase
8 mPGK n=3 SV40 Luciferase
Average Fold Std. Dev. n Induction 1.41 0.54 1 1
77 Figure 2-6: HeLa cells were transfected with the constitutive fhgalactosidase expressing plasmid
pBOS, and an SV40 Luciferase - - - transgene. - -- Transfected cells were incubated under 1% or
20% oxygen, harvested, and assayed for fl-galactosidase activity. The Fold Induction
was calculated by taking the ratio of Luciferase activity to ~galactosidase activity for the
sample grown at 1% oxygen, and dividing it by the same ratio for the sample grown at
20% oxygen. Data are presented as the average fold induction for n replicate samples;
88 Figure 2-10: TK Lac2 transgenes were electroporated into RI ES cells with the constitutive
w - Luciferase expression plasmid pSVLuc+. Transfected cells were split into two
populations and incubated under 1% or 20% oxygen as described in the Materials and
Methoûs. Data are presented as a pair of bars corresponding to the average Absorbance
for constmct, after standardizing for transfection and harvesting efficiency. Shown is
representative data fiom one of three replicate experiments. No significant differences in
TIC LacZ activity were seen in cells exposed to 1 % or 20% oxygen.
89 7. The PGK HRE SV40 Luciferase haasgene shows the gnatest change in acthity between
- 10% and 1% oxygen -- -
As discussed previously, the intention behind this work was to identiQ the areas of reduced
oxygen concentration that occur in a developing embryo through the expression of an oxygen
sensitive transgene. In the work to this point, HRE transgene expression has been investigated
by cornparhg transgene expression in cells exposed to 20% oxygen (normal incubator
conditions), to that observed in cells exposed to 1% oxygen (low oxygen conditions). In the
literature these conditions are often referred to as nomoxic and hypoxic respectively, and are the
standard conditions under which much of the work on hypoxia responsive genes has been
performed 1,13,14. Recently, it bas been hypothesized that the definition of 20% oxygen as
"normoxic" is misleading, and that the average tissue oxygen concentration is around 6%. This
estimate seems to be supported by the historical findings of clioicians 17. Given this
information, and the requirement that the transgene have low basal levels of activity under
nonnal conditions of oxygen, how suitable is an HRE-based transgene for the examination of
tissue oxygenation?
As shown in Figure 2-6, of the HRE SV40 Luciferase constnicts tested, the PGK HRE SV40
Luciferase transgene consistently gave the bighest increase in Luciferase activity in hypoxic
cells, averaging 52 f 16 fold induction (n=8). To examine the effect of oxygen concentration on
transgene activity HeLa cells were traasfected witb the PGK SV40 Luciferase, the EPO (n=3)
SV40 Luciferase, or the basic SV40 Luciferase transgene; transfected cells were split into two
plates and incubated at 20% oxygen, or under reduced levels of oxygen (l5%, IO%, 5%, or 1%)
for 40 hours. After incubation cells were harvested and assayed for Luciferase or p- galactosidase activity as described previously. Standardized induction ratios were calculated for
each sample set (see Matenals and Methods for calculations); Figure 2-1 1 shows the average
induction as a fiuiction of oxygen concentration. Neither the EPO nor the PGK transgene
showed a significant increase in activity between 20% and 10% oxygen (p < 0.10). Beween
10% and 5%, cells transfected with PGK HRE SV40 Luciferase showed a signifiant increase in
Luciferase activity ftom 2.6 to 11 times the level at 20% (p < 0.005). The EPO(n=3) SV40
Luciferase shows little change in activity between 10% and 5%. Between 5% and 1%, the PGK
SV40 Luciferase transgene shows a M e r increase in activity, fiom 1 1 fold increase at 5% to 24
-- - 90 fold increase at 1%. (Both compared with the level of Luciferase at 20%, p < 0.005.) At 1%
---A - - - - oxygen concentration the EPO construct showed - a modest increase to 6 times the Luciferase
activity at 20%, but there were not suficient data to conclude that the observed increase was
statistically signi ficant (p < 0.25).
91 Figure 2- 1 1 : The HRE SV40 Luciferase ûansgenes show dose dependent responses to changes
- ------ in oxygen levels.
Induction of SV40 Luciferase transgenes under increasing oxygen concentration
Oxygen concentration
-t- SV40 Luciferase
4 E P O (n=3) SV40 Luci ferase PGK (n=3) SV40 Luci ferase
Statistical analyses were perfonned using Student's t test, with reference to the table of t values
located at http:l/www.statsofi.com/textbook/sttable.h~l.
HRE &galactosidase results are presented as graphs of representative experiments showing
the $-ptactosidase activity meamcl m 20% and 1% ceti extracts, standdzed far the cumber
of cells. Unlike the constitutively expressed B-galactosidase used with the HRE Luciferase
experiments, p-galactosidase activity produced by the HRE-Lac2 constructs is almost
undetectable in most 20% extracts. Conversely, some samples, such as the HeLa cells
transfected with PGK HRE Lac2 produce nearly undetectable levels of pgalactosidase activity
under 20% oxygen, and measurable activity from cells exposed to 1% oxygen. Because of the
low level of activity observed under 20%, the calculation of a ratio between 1% and 20% & galactosidase activity is meaningless, mathematically equating to -/O. It is not known if the low
level of B-galactosidase activity detected in 20% samples represents "no expression" of the
107 transgene under 20% oxygen, or if the spectrophotometric assay is simply not sufficiently
-A--- - - sensitive to detect a response.
in the experiment to look at the effect of oxygen concentration on transgene activity (Figure
2-1 l), HeLa cells were tramfected as described above with EPO (n=3) SV40 Luciferase, PGK
Luciferase, or SV40 Luciferase, and pBOS. Transfected cells were split equally to two plates,
and incubated under 20% and a reduced level of oxygen (15% Oz, 5% CO2, 80% N2), (10% 0 2 ,