it.

‘

x ,

3m........r..
2‘. .1.
2

.
Kit. I «.
hafiflmmw
.2.
JWh..lxit ..

"r!“

1.. .x.

.1: V8}.

i
.v; 1124'

.‘ogat‘flv‘u‘ . v

talk: :1?

. 3:"...1.

.3.“

:T

ii

 

l]
‘be‘

mp

 

LIBRARY
Michigan State
University

 

 

This is to certify that the
thesis entitled

THE ROLE OF HIF10t SIGNALING IN METAL-INDUCED
TOXICITY

presented by

Ajith Vengellur

has been accepted towards fuIfillment
of the requirements for the

Ph.D degree in GENETICS

 

 

 

 

MSU is an affirmative-action, equal-opportunity employer

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/CIRC/DateDue.indd-p.1

 

THE ROLE OF HIF1or SIGNALING IN METAL-INDUCED TOXICITY

By

Ajith Vengellur

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
GENETICS

2007

ABSTRACT

THE ROLE OF HIF1a SIGNALING IN METAL-INDUCED TOXICITY

By

Ajith Vengellur

Oxygen is critical for all aerobic organisms and they have well developed cell
signaling systems in place to maintain homeostasis under oxygen deprivation.
Hypoxia inducible factor 1a (HIF1a) is the major transcription factor that
mediates the cellular response to hypoxia in higher eukaryotes. HIF1a protein
undergoes proteolytic degradation under norrnoxia by the concerted activity of
HIF pronI hydroxylase (PHD) that modifies HIFla and the Von hippel-Iindaeu
(VHL) protein that recognizes the hydroxyproline and instigates HIF1a’s 26$
proteosome dependent proteolysis. PHDs require oxygen and 2-oxoglutarate as
substrates and iron and ascorbate as cofactors. DivaIent metals such as cobalt
and nickel and iron chelator desferoxamine act as hypoxia mimic by inhibiting
PHDs. Global gene expression profiling in Hep3B cells revealed that cobalt,
desferoxamine and hypoxia modulate the expression a battery of common
genes. We hypothesized that divalent metal ions may cause toxicity to cells by
their ability to aberrantly activate HIF1 signaling under normoxia. Using wild type
and HIF1or null mouse embryonic fibroblasts, we showed that cobalt and hypoxia
treatments induce similar HIF1or dependent gene expression patterns (is. cobalt

exposure includes ‘signature hypoxia genes’). For example, cobalt exposure

leads to elevated levels of genes with characterized hypoxia response elements,
including many glycolytic enzymes, glucose transporters and the pro-death
factors, NIX and BNip3. BNIP3 expression was only observed in the wild type
cells and its increase had functional significance, leading to caspase-
independent cytotoxicity. Moreover, shRNA that specifically reduced the levels
of BNIP3 within the cell were shown to be protective against cobalt-induced cell
damage. In contrast, cadmium treatment does not stabilize HIF1a protein and is
more toxic to HIF1a null cells compared to wild type cells, and cadmium
exposure induces apoptotic cell death characterized by chromatin condensation
and caspase-3 activity. Cadmium treatment also causes increased oxidative
stress, and cadmium induced toxicity could be prevented by co-treatment with N-
acetyl cysteine and melatonin. The increased susceptibility of HIF1a -/- cells to
cadmium-induced damage could be due, in part, to the fact that these null cells
had elevated levels of ROS in the untreated condition and a compromised ROS
scavenging system. Taken together, our results suggest that cobalt-induced cell
damage is promoted through HIF1a stabilization, and subsequent transcriptional
activation of BNip3 gene and cadmium-induced cytotoxicity is due to ROS
generation and the compromised nature of the ROS scavenging system in the
HIF 1a -/- cells. This study suggests that HIF1 signaling is an important factor in
the underlying pathology of injuries seen in workers exposed to metals, such as
cobalt and cadmium, and understanding this connection is critical to our ability to

cope with these injuries.

DEDICATION

To My Teachers

iv

ACKNOWLEDGEMENTS

From the beginning of my graduate school life I have been fortunate to have a
great many number of people who taught me very valuable lessons both
scientific and personal. First and foremost I want to acknowledge Dr. John
LaPres, my mentor who guided me to the completion of my PhD. He has taught
me to think logically and positively even when things are going tough. I will be
always grateful to him for the freedom he bestowed upon me while planning and
executing experiments. I would like to thank my guidance committee members
Dr. David Arnosti, Dr. Robert Roth and Dr. Tim Zacharewski for their valuable
comments and advice during the course of my work. It has certainly helped me
to tackle scientific problems in creative ways. I would also like to thank Dr. John
Wang and, Dr. Ron Patterson for scientific advice, Dr. Louis King of the MSU
flow cytometry facility and members of the Genomic Technology Support Facility
for technical support. Dr. Helmut Bertrand, Dr. Barb Sears and Jeannine Lee of
the Genetics program made my life easy in dealing with academic requirements.
I would like to thank Genetics program and MSU graduate school for financial

suppon.

I would also like to thank present and past members of LaPres lab; Barbara
Woods, Eric Jerks, KangAe Lee, Dorothy Tappenden, Yogesh Saini, Scott Lynn,
Jeremy Lynd, Steve Proper and Andrea MacFadden as well as Zacharewski lab
members; Darrell Boverhof, Ed Dere, Josh Kwekel, Lyle Burgoon, Cora Fong

and Jeremy Burt who made my stay in East Lansing very pleasant and

memorable. I am grateful to Fraker lab members; Joe Frentzel and Jim Mann as
well as Roth lab members; Dr. Jane Maddox, Dr. Umesh Hanumegowda, Dr.
Bryan Copple and Dr. Jesus Olivero-Verbal for assisting me in running various
assays. Finally I would like to thank my family and all my friends at MSU and

beyond.

vi

TABLE OF CONTENTS

LIST OF TABLES ........................................................................... x
LIST OF FIGURES ........................................................................ xii
INTRODUCTION ........................................................................... 1
l. Metals .............................................................................. 1
LA. History ..................................................................... 1
LB. Metal Ions in Biological Systems ................................... 2
LC. Transition Metals ........................................................ 3
ID. Cobalt ...................................................................... 8
I.D.1. Occurrence .......................................................... 8
I.D.2. Exposure ............................................................. 9
I.D.3. Biological Functions .............................................. 9
I.D.4. Toxicity ............................................................... 10
LE. Cadmium ................................................................. 12
I.E.1. Occurrence .......................................................... 12
I.E.2. Exposure ............................................................. 12
l.E.3. Biological Function and Toxicity ................................ 13
l.F.Molecular Effects of Transition Metal Exposure .................... 15
l.F.1. Effects on Metal Ion Transport and Metalloproteins ....... 15
l.F.2. Metal-induced ROS and Biomolecule Damage ............ 16

l.F.3. Metal-induced Changes in Gene Expression
and Signaling Pathways ......................................... 17
ll. Hypoxia ............................................................................ 20
HA. Hypoxic Signaling ...................................................... 21
ll.A.1. Physiological Changes to Hypoxia ............................ 22
"AZ. Tissue Responses to Hypoxia .................................. 23
ll.A.3. Cellular Responses to Hypoxia ................................ 25
MB. Hypoxia Inducible Factors (HlFs) ................................... 27
II.B.1. Regulation of Hypoxia Inducible Factors .................... 28

ll.B.1.a Transcriptional and Translational

Regulation of HIF ........................................... 30
Il.B.1.b. Post-translational Regulation of HIF .................... 30
ll.B.2.Regulation of Gene Expression by HIF1a .................. 35
Il.B.2.a. Survival Response ......................................... 36
ll.B.2.b. Cell Death Response ....................................... 37
".0. Role of Hypoxic Signaling in Human Diseases ............... 38
Ill. Reactive Oxygen Species .................................................... 41
"LA. Source of Reactive Oxygen Species (ROS) ..................... 42
lll.A.1 . Mitochondria ........................................................ 42
lll.A.2.Transition Metals .................................................. 45
lll.A.3. Hypoxia and Hypoxia Reperfusion ............................ 47
lll.A.4. Extra-mitochondrial Sources ................................... 47

vii

III.B. Protection against ROS ............................................... 48
IV. Hypothesis and Specific Aims ................................................ 49
References ............................................................................ 53

CHAPTER 1 — Gene Expression Profiling of Hypoxia Signaling in Human

Hepatocellular Carcinoma Cells ......................................................... 72
Abstract ....................................................................................... 73
Introduction .................................................................................. 74
Materials and Methods .................................................................... 75
Results ........................................................................................ 79
Discussion .................................................................................... 93
Acknowledgements ........................................................................ 1 01
References ................................................................................... 102

CHAPTER 2 - Gene Expression Profiling of the Hypoxia Signaling
Pathway in Hypoxia Inducible Factor 1a Null Mouse Embryonic

Fibroblasts .................................................................................... 105
Abstract ........................................................................................ 106
Introduction ................................................................................... 1 07
Materials and Methods .................................................................... 108
Results ......................................................................................... 116
Discussion .................................................................................... 135
Acknowledgements ......................................................................... 142
References ................................................................................... 143
Appendix ...................................................................................... 147
CHAPTER 3 — The Role of Hypoxia Inducible Factor 1a in Cobalt Chloride

Induced Cell Death in Mouse Embryonic Fibroblasts ............................. 154
Abstract ....................................................................................... 155
Introduction .................................................................................. 1 56
Materials and Methods .................................................................... 158
Results ........................................................................................ 163
Discussion .................................................................................... 176
Acknowledgements ........................................................................ 180
References ................................................................................... 181

CHAPTER 4 — The Role of HIF1a in Metal-Induced Toxicity in Mouse

Embryonic Fibroblasts ..................................................................... 185
Abstract ....................................................................................... 186
Introduction .................................................................................. 188
Materials and Methods .................................................................... 188
Results ........................................................................................ 194
Discussion .................................................................................... 21 1
Acknowledgements ........................................................................ 217
References ................................................................................... 218

viii

SUMMARY AND DISCUSSION ........................................................ 221

ix

LIST OF TABLES

INTRODUCTION

Ta_ble__1_: Metalloproteins in Organisms ................................................ 5
CHAPTER 1

Meg: Primers used in qRT-PCR ..................................................... 80
Mei: Top 20 Hypoxia Up and Down-regulated Probe Sets .................. 82

Table 4: The 62 Probe Sets Representing the Overlap
Between all Four Treatments ............................................................ 85

Table 5: Top 20 up and Down-regulated Genes

Specific to Hypoxia (1% 02) Exposure ................................................ 91
Table 6: Top 20 up and Down-regulated Genes

Specific to COCIz (100uM) Exposure ................................................... 92
Table 7: Top 20 up and Down-regulated Genes

Specific to Deferoxamine (100uM) Exposure ........................................ 94
Table 8: Top 20 up and Down-regulated Genes

Specific to NiCI2 (100uM) Exposure .................................................... 95
CHAPTER 2

Table 9: Primers for Microarray Production and RT-PCR ........................ 111
Table 10: Significantly influenced Genes in WT Cells

Treated with Hypoxia or Cobalt ......................................................... 122
CHAPTER 2 APPENDIX

Table 11: Genes Significantly Influenced by Cobalt
and Hypoxia in HIF1a 4- Cells .......................................................... 148

Table 12: Genes Significantly Influenced by Cobalt
and Hypoxia in Wild Type and HIF1a -/- Cells ...................................... 150

Table 13: Genes Significantly Influenced by Cobalt
in Wild Type and HIF1a 4- Cells ........................................................ 151

Table 14: Comparison of Genes in Table 2 with those
Presented in Jiang et. al Ref. 36 ....................................................... 152

CHAPTER 3
Table 15: qRT-PCR Primers ............................................................ 161
CHAPTER 4
Table 16: qRT-PCR Primers ............................................................ 191

Table 17: shRNA Sequences ........................................................... 191

xi

LIST OF FIGURES

INTRODUCTION

flgLe‘I: Fenton Reaction ............................................................... 7
flgLeZ: Graphical representation of HIF1oc Protein .............................. 29
53M: HIF1a and Hypoxia Signaling ............................................... 33

Figure 4: Sites of ROS Production in Mitochondria ................................ 44

CHAPTER 1

Figure 5: Venn Diagrams Representing Genes whose
Expression was Significantly Influenced by Treatment ........................... 83

Figure 6: qRT-PCR Verification of 12 Genes ......................................... 88

Figure 7: Analytical Comparison of qRT—PCR
and Microarray Data ........................................................................ 89

CHAPTER 2
Figure 8: Experimental Design of Loop ............................................... 112
Figure 9: HIF1a Western Blot ........................................................... 117

Figure 10: Hierarchical Cluster Diagram of
Entire Data Set ............................................................................. 119

Figure 11: Analysis of CoClz and Hypoxia Treatments
in Wild Type Cells ........................................................................... 121

Figure 12: Venn Diagrams of Various Combinations
of Cell Types and Treatments ............................................................ 126

Figure 13: Glycolytic Enzyme Cluster ................................................ 129
Figure 14: Comparison of Microarray and rRT-PCR Results .................. 131

Figure 15: Time Course Analysis of the PHDs by rRT-PCR .................... 134

xii

Figure 16: Analysis of the Basal Levels of GST-u,
SOD1, SODZ and Catalase by rRT-PCR ............................................. 136

Figure 17: Transient Transfection of Hep3B Cells ................................. 137
CHAPTER 3
Figure 18: Growth Curve ................................................................. 164

Figure 19: Dose Response and Time Course Analysis of
CoClz-induced Cytotoxicity ............................................................... 165

Figure 20: qRT—PCR Data for BNip3 and BNip3L ................................ 167

Figure 21: qRT-PCR Data for BNip3 Expression under
COCIz and Hypoxia Time Course in Wild Type Cells .............................. 169

Figure 22: BNIP3 and BNIP3L (NIX) Western Blot ............................... 171
Figure 23: Cell Morphology ............................................................. 173
Figure 24: Caspase-3 Assay ............................................................ 175
CHAPTER 4

Figure 25: Cytotoxicity of CoClz and CdCIz to MEFs ............................. 195
Figure 26: Characterization of Cadmium-induced Cell Death .................. 197
Figure 27: Hypoxia Signaling and Cytotoxicity to COCIz and CdClz ........... 198
Figure 28: Oxidative Stress in CoClz and CdCI2 mediated Cytotoxicity ....... 203
Figure 29: Expression and Activity of Superoxide Dismutases ................ 209

Figure 30: Cellular Glutathione Levels under CoCIz
and CdCl2 Treatments .................................................................. 212

(Images in this dissertation are presented in color)

xiii

INTRODUCTION

I. Metals

I.A. History

The earliest form of life is thought to have evolved on the face of earth
approximately four billion years ago from a pre-biotic soup, rich in various
chemical nutrients (1). Current theory suggests that RNA molecules were
produced in the early days by random chemical reactions. These RNA’s later
acquired the ability to perform catalysis and self replication and were possibly
enclosed in micelles leading to the origin of the early life form (2). It is also
thought that the catalytic activity of RNA was dependent on the presence of
divalent cations (3). As time passed, organisms evolved into more complex life
forms through multiple cycles of mutation and natural selection. Even as DNA
rose in prominence and the major role of catalysts were taken over by proteins,
the role of divalent cations in facilitating various chemical reactions was

conserved (4).

Meanwhile, the process of evolution and natural selection lead to the creation of
advanced life forms like vertebrates and subsequently mammals such as Homo
sapiens that require certain metals as essential nutrients. It was not until the late
Stone Age, approximately 8000 8.0. that the use of metal or rather metal rich
mineral rocks were first introduced as tools for hunting (5). Metals, such as gold
and silver, which occur more or less in the pure form, were the first to be used

along with rocks rich in copper. The Bronze Age, which occurred around 3500

B.C., began a new age with the production of tools for farming and hunting thus

ushering in the new era of metallurgy.

It took a further two millennia for the advent of the Iron Age when iron tools and
steel production began in earnest. By this time people began combining different
metals and minerals to produce alloys with distinct properties. The discovery of
new metals and alloys heralded an era where human civilization was pushed to
new heights by the introduction of advanced tools that took advantage of the
strength and malleability of these new materials (6). Many new metals, such as
cobalt, cadmium, uranium were discovered after the advent of the new field of
chemistry and advances in metallurgy. By this time the exploitation of natural
resources by mining and further refining had created enormous global industries,
fostering advances in many fields such as aerospace and atomic energy. This
also resulted in increased exposure of worker, as well as, the general population,

to various potentially toxic metals and minerals.

LB. Metal Ions in Biological systems

Metals ions are associated with almost one third of all enzymes and are essential
for their activity (7). A metal’s ability to rapidly transfer electrons facilitates
enzyme catalyzed reaction and without the metal a metalloenzyme’s activity is
drastically reduced. Since metals ions are generally positively charged and act
as electrophiles, they can operate as a buffer system facilitating

reduction/oxidation (redox) reactions (8, 9). In addition, metals are essential for

maintaining pH in cells via their role in ion transport, and play equally critical roles
in cellular energy production via their role in the electron transport chain (ETC)
(10). Most organisms employ coupled oxidoreductase systems for the
production of energy, a process known as oxidative phosphorylation. In higher
life forms, the ETC employs a variety of metalloproteins to transfer electrons
obtained through oxidation of carbons to oxygen, producing ATP and water. The
metalloproteins of the ETC utilize the energy produced as the electrons move
through the ETC to transport H+ ions across the inner-membrane of the
mitochondria. This process is coupled to energy production by ATP synthase
(also known as Complex V) that utilizes the membrane potential to drive the
production of ATP. Therefore, metals are essential to energy production and

thus life of a cell.

Nevertheless, not all metal ions are utilized by living things. Many of the heavy
metals (i.e. metals that are heavier than the rare earth metals), such as mercury
and arsenic are not employed by cells and are generally toxic to most life forms
(11). Even metals that play an essential role in supporting life the levels are
maintained within a strict limit. Any change in the homeostasis of these metals is
detrimental to the cells. They can have adverse reactions with cellular

components such as proteins, lipids and nucleic acids (9).

LC. Transition Metals

In chemistry, a metal (Greek: Metallon) is an element that readily forms ions
(cations). The metals are one of the three groups of elements, along with the
metalloids and nonmetals, each distinguished by their ionization and bonding
properties. The transition metals are distinguished by the number and
distribution of electrons in their outer two shells and include metals such as iron,
zinc, manganese, molybdenum, nickel, cobalt, and cadmium. The properties
inherent in the distribution of these electrons, accord these metals certain

features, such as high tensile strength and different oxidation states.

Metals, such as iron, calcium, and zinc, play critical roles in major biological
processes at molecular level (12). Among all these metals, iron has a central
metabolic role as it facilitates many critical cellular processes, such as oxygen
transport (e.g. hemoglobin and myoglobin) and energy production (e.g. ETC).
Evolution of complex life forms, such as vertebrates, was made possible by the
ability to deliver oxygen to distant tissues, an iron dependent process (13).
Moreover, the evolution of the electron transport chain, a system that requires
iron and copper, increased the efficiency of energy production. Zinc, another
essential nutrient, plays a major role as a prosthetic group for various enzymes
and associates with many transcriptional factors (Table 1). Calcium plays a
major role in various intercellular signaling processes and along with magnesium
and sodium, plays a central role in regulating cellular pH levels and neural

signaling (14). Because of the major role played by these metals, they are

 

Table 1: Metalloproteins in Organisms

 

 

 

 

 

 

Metal Class Protein

Iron - Heme Oxygen Transport Hemoglobin
Cytochromes Cytochrome C
Cytochrome Peroxidases Horseradish peroxidase
Cytochrome P-450 Cytochrome P-450
Oxygenases Heme Oxygenase

 

Oxido-Reductases

Cytochrome c oxidase

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Iron — Non-Heme Iron-Sulfur Clusters Rubredoxin
Mononuclear Iron Proteins Fe-SOD
Dinuclear Iron Proteins Ribonucleotide Reductase
Iron Storage Ferritin
Iron Transport Transferrin
Nickel Urease
Nickel-Iron Hydrogenases
PeLtide Deformylase
Selenium Peroxidase Glutathione peroxidase
Manganese Mn-Superoxide Dismutase
Arginase
Aminopeptidase P
Cobalt Glutamate Mutase
Methylmalonyl CoA Mutase
Methionine Synthase
Vanadium Vanadium Haloperoxidase
Molybdenum / Nitrrogenase
Tungsten Aldehyde Oxidoreductase
CO Dehydrogenase
COpper Type1 Copper Proteins Plastocyanin
Type-2 Copper Enzymes Cu-Zn Superoxide Dismutase
Type-3 Copper Enzymes Hemocyanin
Binuclear Copper Binuclear Copper A
Multicopper enzymes Ascorbate oxidase
Copper storage Copper Metallothionein
Zinc Oxidireductases
Transferases Protein prenyltransferase
Hydrolases: Ester bonds 5’-Nucleotidase
H drolases: Pgtide bonds Matrix metalloproteinase
Other Hydrolases GTP Cyclohydrolase 1
Lyases Carbonic Anhydrase
Zinc-Fingers RING Domain proteins
Zinc Stme Metallothioneins
Other Zinc proteins Insulin
Calcium EF-H and Caz“ bindinL Calmodulin
EGF-Domains
GLA Domains
CZ-Like Domains
Magnesium Chlorophyll hfiprotoporphyrin IX chelatase
ATPases MgQ+) transport ATPase,
Transporters Mamesium transporter protein 1

 

 

considered as essential nutrients, however; under various conditions, these

metals can also become toxic and aberrantly influence cellular function.

The reactive nature of these metal ions is essential for their biological activity and
is also responsible for cellular damage following exposure to excess metal levels.
These metals tend to react with acids and other functional groups, forming
various adducts, and tend to undergo oxidation reactions producing various
reactive species (15). Several metals, such as iron and cobalt, are capable of
undergoing Fenton-like reactions (Figure 1). Fenton-like reactions lead to the
production of reactive oxygen species (ROS) when the metal reacts with
hydrogen peroxide, producing extremely reactive hydroxyl radicals. These
hydroxyl radicals can begin a cascade of chemical reactions producing other

types of ROS within the cell, causing cellular damage (16).

Belonging to the same class of periodic table, transition metals have similar
chemical properties and valence states. Because of this, it is important for the
cells to regulate the levels of various metals so as to prevent them from inhibiting
each other’s chemical function. An increase in exposure to metals such as
cobalt and nickel which are essential for the function of manyenzymes and
proteins but are not considered essential nutrient, may lead to the disruption of
metal ion homeostasis (17). Many proteins that contain co-coordinately bonded
metal ions such as non-heme iron and zinc at their catalytic site are prone to

replacement by nickel and cobalt ions because of their similarity in charge and

Fenton Reaction

Fe2+ + H202 ——> Fe3+ + OH' + 'OH

Fe3+ + 02' —-——-D F92+ + 02

Haber-Weiss Reaction

H202 + 02- —> 02 + OH' + 'OH

Fenton-like Reaction of Metals

Mn+ + H202 ———> Min“) + OH' + 'OH

M(n+1) + 02. —_—’ MI'H' "I' 02

Figure 1. Fenton Reaction

Iron, Cobalt, Nickel, Copper and Manganese can be oxidized by H202
producing hydroxyl anion and hydroxyl radicals. The oxidized metal ions
gets reduced by superoxide anion. This can lead to a continuous cycle of
ROS production.

size (18). Moreover, metals such as cadmium and arsenic can bind to proteins,

forming adducts that can disrupt normal cellular processes (19).

LB. Cobalt

I.D.1.0ccurrence

Cobalt is a magnetic element with ferrous-like properties. It occurs as arsenides,
oxides and sufides in nature and its introduction into manufactured goods began
in Egypt and Persia in second century BC. as a dye for pottery. Later, Leonardo
Da Vinci used cobalt for its bright color in oil paints. The actual element was
isolated by the Swedish chemist, George Brandt in the early 18th century and
TO. Bergman identified it as an element in 1780 (20). The first cobalt-containing
permanent magnetic alloy, called alnico, was developed in 1930. Cobalt is
usually not seen in the free form and exists as ores with other metals such as
copper, silver and nickel. The main cobalt ores are cobaltite and glaucodot
((Co,Fe)AsS), erythrite ((C03(AsO4)2-8H20)), and skutterudite ((Co,Ni,Fe)Asa)
and the main producers are Democratic Republic of Congo, China, and Russia.
The radioactive form of the element, 60Co, was used to irradiate food stuffs and in
radiotherapy. Industrial application of cobalt, In high-speed drills and cutting
tools as a hard metal in combination with tungsten was started in 20th century.
Today it is primarily used in the production of superalloys, corrosion resistant
alloys, high speed steels, magnets, wear resistant coatings and various

metallurgical applications.

I.D.2. Exposure

Cobaltous minerals are primarily found mixed with silver and other minerals of
copper, iron, lead, nickel. Cobalt metal is usually produced during the refining of
copper or nickel (20). Metal refinery workers and those involved in the
production of certain alloys, such as tungsten carbide, are at the greatest risk for
cobalt-induced toxicity, however, the general public has become increasingly at

risk of cobalt exposure by various mechanisms.

Cobalt is present in air as a result of natural processes, such as erosion, volcanic
eruptions, and by various human actions, such as the burning of fossil fuels,
sewage sludge, and phosphate fertilizer (21). Being a rare trace element, cobalt
content of soils, especially the one derived from sandstone, limestone etc., are
inadequate to support basic biological demand. Diet is the main source of
exposure for the general population and the average daily intake ranges from 5-
45 pg. People with high dietary intake of sea foods, cocoa, bran and molasses
ingest more cobalt, as these foods have a higher natural content of this metal. In
addition, certain organ meats, especially liver, and fat tissues have high cobalt
content (22). Generally, the concentration of cobalt in the drinking water is low
(less than 5pg/L) (20). Cigarette smokers are at an increased risk for cobalt-
induced injury, primarily because tobacco smoke contains 0.3-2.3mg Co/kg dry

weight (23).

I.D.3. Biological Functions

Cobalt is essential for the production of vitamin B12 (also known as cobalamin),
which is required for the synthesis of methionine and the metabolism of purines
and folates (24). The deficiency of cobalt leads to pernicious anemia. Humans
cannot synthesize vitamin B12; “they” must obtain it through diet. Cobalt is also
associated with the regulation of a number of cofactors such as acetyl coenzyme
A and enzymes such as lipoprotein lipase (25). Cobalt was used in mid 20th
century for the treatment of anemia because of its ability to induce the production

of erythropoietin, the red blood cell stimulating factor (26, 27).

I.D.4. Toxicity

Most cobalt toxicity occurs in an occupational setting, as exposure to the general
public is usually rare. Even as cobalt was being used to treat anemic patients, its
excessive uptake was reported to cause thyroid failure and goiter (28). In
European and Canadian breweries cobalt was used to reduce the effect of soap
residues on foam formation in beer production. Brewery workers who drank
cases of beers showed symptoms of pericardial effusion, elevated hemoglobin
levels and congestive heart failure (29). Post-mortem histology showed
myocardial fiber degeneration, increased vacuolation and interstitial edema and
extensive damage to mitochondria. In Andean mining communities in South
America, polycythemia was seen in miners with excessive levels of serum cobalt

(30).

IO

Cobalt exposure leads to a variety of effects. Cobalt is a known allergen and
causes allergic dermatitis (31). Acute exposure to high levels of cobalt through
inhalation in humans results in respiratory problems such as decrease in
ventillatory function, congestion, edema and hemorrhage of the lung. Chronic
exposure to cobalt through inhalation leads to respiratory irritation, wheezing,
asthma, pneumonia and fibrosis (20). During the World War II, severe
pulmonary problems were reported among workers exposed to hard metal
particles. Exposure to cobalt dust among diamond polishing workers was
reported to cause pneumoconiosis, which was mostly the result of exposure to
both tungsten and cobalt. This disease is characterized by intense alveolitis and
in the end stage, pulmonary fibrosis. Pneumoconiosis rarely occurs because of
cobalt alone and usually is a combined effect with other metals such as tungsten

(25).

Other effects of cobalt exposure include cardiovascular damage such as cardiac
myopathy, as well as congestion of the liver, kidneys and conjunctiva, and
immunological effects. In humans, gastrointestinal effects such as vomiting,
diarrhea, liver injury and allergic dermatitis were also reported in case of oral
exposure to cobalt (32). There have been no reports of reproductive toxicity of
cobalt in either oral or inhalational exposure (33). Although cobalt has been
shown to be weekly mutagenic in bacterial systems, conclusive evidence of

cobalt genotoxicity in mammals has not been reported (34).

11

I.E. Cadmium

I.E.1. Occurrence

Pure cadmium is a silvery white, lustrous, soft and ductile metal, which is rare in
nature and occurs mostly as inorganic compounds, as organo-cadmium
compounds are extremely unstable (35). It is generally seen with other metal
ores, most often zinc sulfide, which is called greenockite or cadmium blende (36).
Cadmium occurs in small quantities in air, water and soil and is produced as a
by-product of the refinement of other metals, as cadmium concentration is not
high enough in minerals for a dedicated process. Although known for many
years, the industrial use of cadmium began in 19303 and has steadily risen with
approximately 20,000 tons produced annually by the end of twentieth century.
The general consumption of cadmium has increased as the use of Ni—Cd
batteries in electronics has risen. Although cadmium had been used in pigments,
PVC stabilizers and, electroplating, consumption for these purposes is

decreasing rapidly (37).

I.E.2. Exposure

Cadmium is released into the biosphere by both natural and anthropogenic
activities, including volcanoes and weathering of rock. Each plays a major role in
the global cadmium cycle but those processes do not result in elevated
environmental levels. The main sources of emission into the atmosphere are
volcanoes, sea spray, and forest fires (38). The major source of cadmium by

human activity results from the mining of cadmium-containing minerals such as

12

zinc ores, coal and lime. Significant environmental release of cadmium also
occurs in countries where extensive waste incineration is performed (38).

Land filling of discarded products and, contaminated waste are major sources of
cadmium in the soil. Recently, the use of treated sewage sludge as fertilizer for
crops has resulted in an increase in soil contamination (39). These activities are
leading to the accumulation of cadmium in the top soil and ground water. Apart
from atmospheric deposition, domestic waste water, non-ferrous metal smelting
and refining and manufacturing of chemicals and metals have contributed to the

localized environmental release of cadmium.

l.E.3. Biological Function and Toxicity

Cadmium is not known to have any endogenous functional role in the animal
system, however; in plants it has been shown that sub-toxic levels of cadmium
can block the viral invasion of nutrient transport system (40). Cadmium and its
compounds are relatively water soluble and mobile in soil, making it more
bioavailable and increasing its bioaccumulation in plants, microbes and marine

organisms.

In animals, cadmium concentrates in the internal organs such as kidney and
liver, and in the body cadmium levels usually increase with age, as only a small
proportion of the body burden of cadmium is excreted through urine and feces
(41). The acceptable daily intake of cadmium should not exceed 0.4—0.5

mg/week according to WHO guidelines (42). Chronic cadmium exposure leads

13

to a variety of toxic effects including kidney damage and lung emphysema, in a
variety of organisms including birds and mammals (43). Cadmium exposure can
also lead to bone loss due to its ability to impair the metabolism of calcium and

vitamin D.

In humans where cadmium exposure can occur through food, water and cigarette
smoke, the renal accumulation of cadmium results in kidney dysfunction. This
renal toxicity leads to impaired re-absorption of proteins, glucose and amino
acids (44). In Japan, a syndrome called itai-itai has been associated with chronic
cadmium exposure in humans. The affected individuals, mostly elderly women,
experience softening of bones (osteomalacia), lumbar pain, myalgia and
spontaneous fractures with skeletal deformation (45). Occupational cadmium
exposure can lead to a variety of lung changes, including chronic obstructive
airway disease and cancer (43). In addition, workers exposed to high cadmium
levels are at an increased risk of prostate cancer (46). Interestingly, zinc is
antagonistic to cadmium-induced toxicity and has been shown to reduce the toxic
effects of cadmium exposure, such as damage to ovaries and testes,

hypertension etc. (47).

Although there has been no report of teratogenic effects in the human population
exposed to cadmium, effects are reported in rats administered with 1.25mg/kg
cadmium (48). This may be due to the fact that cadmium inhibits the placental

transport of zinc. Cadmium has also been shown to produce sarcomas at the

14

sites of injection and interstitial tumors in the testes of rats (49). Even though
inhalation of cadmium chloride has resulted in primary lung cancer in rats (50)

gastrointestinal exposure did not produce any tumors (51).

IF. Molecular Effects of Transition Metal Exposure

The essential role that various metals play in sustaining life processes at the
cellular level means that maintenance of intracellular concentrations of these
metals is vital. Exposure to toxic doses of essential metals or other non-essential
metals can result in perturbations to homeostasis of these elements and lead to
cellular dysfunction. The changes at the molecular level upon metal exposure
can be grouped into three categories — inhibition of function or transport of
nutrient metal ions within the cell, generation of reactive oxygen species and

resulting damage to biomolecules, and changes in gene expression.

l.F.1. Effects on metal ion transport and metalloproteins

Divalent non-ferrous metal insult can disturb iron homeostasis by disrupting the
transport of iron, as well as inhibiting iron metalloproteins and enzymes. For
example, increased cadmium levels in the diet can decrease the uptake of iron
due to competitive binding of cadmium to ferritin, an iron storage protein (52).
Similarly cadmium can interfere with calcium metabolism, resulting in
hypercalciurea and osteoporosis (53). Metals such as lead and mercury, can
also interfere with neurotransmitter signaling by disturbing the calcium and

sodium transporters in the plasma membrane of neurons and muscle cells (54).

15

Many metalloenzymes, such as exopeptidases, arginase, keto acid
carboxylases, and phosphatases do not show absolute metal specificity and may
be activated by many different metals (12). Divalent transition metals, such as
cobalt, nickel and cadmium, activate mammalian alkaline phosphatase, a
metalloenzyme that has magnesium at its catalytic center (55). In contrast,
increased levels of cellular zinc can inhibit cytochrome c oxidase, an important
component of the electron transport chain that utilizes copper for enzymatic
function, impairing cellular energy production (56). Mercury and copper ions also

inhibit rat liver arginase (57).

l.F.2. Metal-induced ROS and biomolecule damage

Other major cellular endpoints of metal insult are the production of various
reactive radicals and depletion of the cellular antioxidant pool. Most redox active
transition metals produce ROS through Fenton-like reactions. These reactions
produce highly reactive hydrogen peroxide and hydroxyl radicals that in turn
react with the surrounding proteins, nucleic acids and lipids, resulting in their
oxidation (21). The oxidized products will be non-functional and are
subsequently degraded. High levels of ROS production cause massive cellular

damage, ultimately leading to cell death (see section III. p.41).

Metals such as arsenic, mercury and cadmium, have electron-sharing affinities
that result in the formation of covalent attachments, especially with sulfhydryl

groups of proteins (58). For example, arsenic can inhibit the pyruvate

16

dehydrogenase complex by covalently linking to the active sulfur residue of the
Iipoic acid covalently attached to E2 subunit, rendering the complex inactive (59).
Cadmium is known to bind covalently the sulfhydryl group of glutathione, thus
depleting the cellular antioxidant reserves (60). Mercury interacts with many
proteins, including albumins, forming S-Hg-S bonds (61). Lead also forms
mercaptides with protein SH groups and less stable complexes with other

amninoacid side chains (62).

l.F.3. Metal-induced changes in gene expression and signaling pathways

As described earlier, all living organisms utilize the diverse substitution and redox
properties of transition metals in a variety of regulatory functions. Metals and
metalloproteins also influence the expression of a wide range of genes (63).
These genes encode proteins meant for metal detoxification, uptake, storage,
transport and homeostasis, as well as various other metabolic pathways. Metal-
induced changes in gene expression may be triggered by processes involving a
specific metal ion acting as a sensor, changes in the cellular redox status or
oxidative stress, and aberrant activity and regulation of various metalloproteins
and their downstream effectors (63). Many times one or more of these factors

together cause changes in gene expression.

A change in intracellular or extracellular metal concentrations may activate a

metal ion sensor, causing downstream changes in expression of genes involved

in an adaptive response. For example, heme containing proteins play an

17

important role in sensing the changes in iron concentration, influencing the
expression of genes involved in iron and heme homeostasis (64). Similarly an
increase in the level of copper ions causes a copper ion binding regulatory
protein, ACE1, to bind to the 5’ end of the promoter of metallothionein gene and
increase its expression in the yeast, Candita glabrata (65). Metallothioneins in
turn act as a detoxifier by transporting the excess copper ions out of the cell. In

most cases such adaptive responses are not harmful for the cells.

By regulating the activity of various redox sensitive transcription factors, metals
can cause gene expression changes and activate various signaling pathways.
NF-KB is an important redox sensitive transcription factor that is affected by
metal exposure and resulting oxidative stress. Critical cysteine residues in the
DNA binding region of NF-kB protein that binds zinc are sensitive to ROS.
Alternatively, metals such as cadmium and copper interfere with the DNA binding
activity of NF-kB (66, 67). Nickel and cobalt ions, however, have a positive effect
on the DNA binding ability of NF-kB (68). Another important transcription factor,
AP1, is induced by certain metals. For example, arsenic, vanadium, chromium,
nickel, cadmium, lead, cobalt and iron are known to activate AP-1 in a variety of
cells. It is thought that AP-1 activation upon metal exposure is mediated by the
activity of several protein kinase signaling pathways that are activated by the
oxidative stress (69). p53 is another critical protein that is affected by metal-
induced oxidative stress. Inorganic arsenic, chromium (VI) and nickel have been

shown to induce p53 expression (69).

18

Metal transcription factor 1 (MTF1) is a protein whose activity is induced by zinc
ions and oxidative stress. MTF1 regulates the expression of detoxifying genes
such as metallothioneins. Exposure to zinc, cobalt and various other metals
result in the expression of metallothioneins, which are major proteins involved in
metal detoxification (70). Cadmium also increases metallothionein expression
but the mechanism is thought to be MTF1-independent. In case of all metals,
metallothioneins act as transporters and storage proteins to limit the free metal
concentration within the cell. For example, cadmium is bound by metallothionein
inside the body and transported to the kidney for elimination (71). Most metal
exposures also lead to changes in the expression of a large number of genes
through unknown mechanisms. Cadmium exposure causes the expression of

heat shock chaperone proteins, DNA repair enzymes, p53, c-fos, etc. (72-76).

Since the early 19505, ferrous-like metals such as cobalt, have been known to
induce the expression of erythropoietin and was used to treat anemia (77).
Later, cobalt and nickel were shown to induce a wide variety of genes involved in
angiogenesis, such as vascular endothelial factor (VEGF), and glycolysis, such
as hexokinase and pyruvate kinase (78, 79). The ability of these metals to alter
this wide expression pattern changes lies with their ability to influence the activity

of the transcription factor, hypoxia inducible factor 1a (HIF1a).

19

Cobalt and nickel are known to stabilize HIF1or, causing the expression of a large
number of genes in the hypoxia signaling pathway. HIF1a protein function is
tightly regulated by a family of non-heme iron-containing dioxygenases called,
prolyl hydroxylase domain containing enzymes (PHDs), which hydroxylate and
target HIF1a for proteosomal degradation under normoxia (see section ll.B.1.b
p.30). Cobalt and nickel interfere with the ability of PHDs to hydroxylate HlFs
and thus cause aberrant stabilization and activity of the transcription factor. This
ability of metals such as cobalt and nickel, to influence PHD activity has three
possible explanations. First, cobalt and nickel might replace the non-covalently
held iron atom from the PHDs rendering them catalytically inactive. Second,
metal-induced ROS production might affect the redox status of iron that is critical
for hydroxylase activity. Finally, it has also been reported that cobalt and nickel
can chelate ascorbic acid, an essential cofactor for PHD activity, in the cell (80).
It is probable that each of these plays a role in the regulation of PHDs and
ultimately influences HIF1a signaling. For these reason, the ferrous-like metals

are considered “hypoxia mimics”.

II. Hypoxia

When life started approximately 4 billion years ago, the world is thought to have
had a reducing atmosphere. The passage of time led to the advent of
photosynthetic organisms that utilized water and carbon dioxide to produce
energy. The principal by-product of this process was oxygen, which began to

accumulate in the atmosphere. The reducing atmosphere gradually changed into

20

an oxidizing one and in modern times, a majority of life forms use aerobic
pathways as a means to produce the energy necessary to sustain life (81). The
process of oxidative phosphorylation, the source of greater than 95% of the
adenosine triphosphate (ATP) produced in a cell, requires molecular oxygen as a
final electron acceptor. Without this final step, ATP generation is inhibited and a
cell’s survival potential is compromised. The ability to sense and cope with
changes in available oxygen, therefore, is critical for life (81). Most of the higher
organisms have evolved elaborate sensing mechanism to cope with decreases in

available oxygen, a condition known as hypoxia.

ILA. Hypoxic Signaling

Hypoxia compromises a cell’s ability to produce ATP through oxidative
phosphorylation and perform enzymatic reactions that require oxygen, such as
hydroxylations. All aerobic life forms, therefore, have evolved unique pathways
to respond to changes in oxygen partial pressure (82). The molecular
components sense the oxygen tension, transduce the signal to various effectors
for facilitating adaptive responses, and promote the physiological and molecular
processes associated with these components. The various processes in the
hypoxic signaling pathway can be loosely classified, into three groups:
physiological changes that occur at an organismal level, coordinated changes
that happen at the tissue level, and molecular changes that occur at cellular

level.

21

ll.A.1. Physiological Changes to Hypoxia

All multicellular organisms have elaborate systems in place for the delivery of
nutrients to metabolically active tissues and cells. In the presence of hypoxia,
these systems respond to adapt to maintain balance. In higher organisms,
hypoxia leads to a rapid modulation of pulmonary ventilation, perfusion, and
blood circulation to optimize the supply of oxygen to tissues. This process is
directed by carotid bodies (specialized chemoreceptor cells in the arterial
system), neuroepithelial bodies in the aiwvay, and vascular smooth muscle cells

(83).

In the presence of decreased oxygen tension, peripheral blood vessels dilate to
allow greater blood volume and thus maximum oxygen supply, to be delivered to
the peripheral tissues. Hypoxic-mediated vasodilatation in oxygen—deprived
tissues is a fast process, mediated by potassium ATP channels of vascular
smooth muscle cells. These channels open upon a decrease in ATP levels, a
condition that occurs under hypoxic conditions (84). It is also thought that Ca2+
channels are involved in this process. Hypoxic vasodilation is highly activated in
the blood vessels of tissues with a high metabolic load, such as the heart and
brain (85). Conversely, pulmonary vasculature constricts to limit the amount of
blood being delivered to poorly ventilated regions, matching ventilation to
perfusion (86). The hypoxia-induced pulmonary vasoconstriction is also a fast
response, mediated by the rapid constriction of vascular smooth muscle cells and

is initiated by the depolarization resulting from the inhibition of several K+

22

channels (87). The depolarization activates voltage gated Ca2+ channels leading

to increase in cytosolic Ca2+ and causes myocyte constriction (85).

On exposure to low p02, ainrvay neuroepithelial bodies and carotid bodies create
cardiorespiratory adjustments through chemosensory fibers (88). Carotid bodies
that reside in the carotid arteries are highly vascularized organs that contain
type1 glomus cells. These glomus cells contain neurotransmitters,
catecholamines and acetylcholine and are in synaptic contact with efferent
sensory fibers of the carotid sinus nerve. Membrane K+ channels in these cells
are inhibited by low p02, leading to membrane depolarization and Ca2+ influx
through voltage gated Ca2+ channels (82). The resulting increased calcium

concentration leads to neurotransmitter release and activation of sensory fibers.

Neuroepithelial bodies are located in ainrvay bifurcations and comprise neuron
derived cells that synapse with branches of afferent and efferent neurons and
transmit signals to brain respiratory centers through the release of
neurotransmitters such as serotonin (89). The activation process is similar to
one that is seen in carotid bodies but the individual channel proteins may vary.
The actual oxygen sensor in both carotid bodies and neuroepithelial bodies are
thought to be a heme-containing protein, associated closely with oxygen

sensitive potassium channel.

ll.A.2. Tissue responses to Hypoxia

23

An important aspect of tissue hypoxia is the increase in angiogenesis. This is
mediated by the activity of various pro-angiogenic factors, most importantly
vascular endothelial growth factor (VEGF), which mediates the growth of blood
vessels to hypoxic regions (90). VEGF promotes cell division of endothelial cells
forming new capillaries which supply oxygenated blood to hypoxic tissues (91).
Another adaptive response to hypoxia is the production of erythropoietin (EPO)
which increases the oxygen carrying capacity of the blood by promoting red
blood cell formation (92). Erythropoietin is produced by the kidney and liver and
is released into the blood which promotes the formation of erythrocytes from their
progenitor cells (93). Together these two changes lead to an increase in oxygen

supply to hypoxic tissues.

Chronic hypoxia also leads to vascular remodeling by the release of vasoactive
substances. Endothelial cells release vasoconstrictors, growth factors, matrix
modifying proteins and adhesion molecules under chronic hypoxia (94). Platelet
derived growth factor (PDGF), platelet activating factor (PAF), endothelin,
leukotrienes, serotonin etc., are produced, leading to increased proliferation (95-
98). Hypoxia increases the production of cytokines such as interleukin-1 (IL-1)
and IL-6 mRNA levels (99). The increased production of cell adhesion molecules
such as intracellular adhesion molecule (lCAM)-1 and vascular adhesion
molecule (VCAM)-1 on endothelial cells along with lL-1 and lL-6 release provide
prothrombotic and proinflammatory interactions with circulating cells, facilitating

the recruitment and migration of progenitor cells into vessel wall (100, 101).

24

Matrix metalloproteinases, MMP-9 and MMP-3, enzymes involved in remodeling
of extracellular matrix facilitating cell invasion, are also induced by hypoxia

promoting vascular remodeling of hypoxic tissues (102).

ll.A.3. Cellular Responses to hypoxia

Oxygen is essential for the survival of cells due to its central role as an acceptor
of the electrons in the electron transport chain (ETC). The ETC is coupled to
ATP synthase, the enzyme responsible for generating ATP through oxidative
phosphorylation. ATP is critical for all energy-requiring cellular processes
including protein synthesis, DNA replication, metabolic pathways, and
maintenance of intracellular pH. Among these, intracellular pH maintenance by
ATP-dependent Na"lK+ ATPase consumes 20-80% of cell’s resting stage
metabolic output (83). When there is a steep drop in ATP levels, these ATPase
pumps fail, leading to membrane depolarization and Ca2+ influx through voltage
gated calcium channels. The rapid rise in intracellular Ca2+ and subsequent
activation of calcium dependent phospholipases and proteases leads to cellular
swelling and necrotic cell death (103). Even a transient lack of oxygen, therefore,

forces cells to undergo molecular adaptation to cope with a drop in energy levels.

At the cellular level, reductions in ATP levels lead to adaptive response to
hypoxia. This is achieved mainly by maximizing the energy production, as well
as decreasing the energy consuming processes. Protein, RNA, and DNA

synthesis are inhibited almost immediately and ion motive ATPases that perform

25

Na+lK+ transport and Ca2+ cycling are favored, as they are essential to maintain a
favorable intracellular environment (104). This phenomenon, known as oxygen
conformance, involves regulatory mechanisms, including translational arrest
(105). Cells show differences in their susceptibility to cell death depending on
their electric activity. Electrically active cells, such as neurons, have high ion
motive ATPase activity that utilizes 80% of cellular ATP, and are more

susceptible to hypoxic cell death than skeletal muscle cells (106).

Under normal oxygen tension oxidative phosphorylation is the major source of
cellular ATP. Under hypoxia, cells switch to aneorobic glycolysis, a less efficient
source of ATP for the oxygen-stewed cells. The increase in glycolysis is
achieved by increases in the production and activity of several glycolytic
enzymes. Phosphofructokinase, a major regulator of carbon flux through
glycolysis, is allosterically activated by ADP and AMP and is inhibited by ATP,
thus deciding the glycolytic rate (107). Phosphofructokinase-2 (PFK-2) is
another important allosteric enzyme involved in regulating the glycolytic pathway.
This enzyme is activated by phosphorylation by the AMP-activated protein kinase
(AMPK). PFK-2 stimulates the production of fructose-2, 6-bisphosphate, a
critical activator of glycolysis, thus increasing the ATP production capacity of the
hypoxic cell (108). The transcript level of PFK-2 is also increased under hypoxia
(109). AMPK also regulates other pathways that help the cell cope with the loss
in oxygen-dependent ATP production, such as glucose uptake controlled by

translocation of various glucose transporter (GLUT) proteins to the cell

26

membrane, the tricarboxylic acid cycle and respiratory chain (110) In addition,
AMPK inhibits anabolic pathways such as fatty acid, triglyceride and sterol
synthesis, as well as the expression of enzymes involved in fatty acid and

gluconeogenesis pathways (11 1-1 13).

It can be seen that many of these adaptive responses are dependent on
transcriptional activation of genes. Cells undergoing hypoxia show regulated
expression of several genes involved in various adaptive responses such as
increased glycolysis and angiogenesis (114). How cells sense the lack of
oxygen had been a point of contention for quite some time until the discovery of
hypoxia inducible factor 1 alpha (HIF1a), a transcription factor that is regulated
by oxygen tension (115). This transcription factor mediates the transcriptional
regulation of a library of genes involved in the cellular and physiological response
to tissue hypoxia (116). Although there may be other transcription factors
regulated by hypoxia, HIF1a remains the most important, regulating a large

majority of hypoxic gene expression.

".8. Hypoxia Inducible Factors
The hypoxia inducible factors (HlFs), which include HlF1a, HlF20, HIF3o, aryl
hydrocarbon receptor nuclear translocator - ARNT (also known as HIF1B) and

ARNT2, control the cellular response to hypoxia in mammals (117). All five HIF
proteins belong to the basic-Helix-Loop-Helix (bHLH) containing PER-ARNT-SIM

(PAS) domain super-family. PAS domains are also found in other transcription

27

factors involved in circadian regulation and xenobiotic metabolism (118-120).
The HlFas dimerize with ARNT or ARNT2 to form a hetero-dimeric transcription
factor. HIF1, a dimer of HIF1a and ARNT, is the most widely studied hypoxia
signaling factor. HlF1a and ARNT are essential for normal development, as the
HIF1a and ARNT null alleles are embryonic lethal due to vascularization defects

(121, 122).

The predominant role that HlF1oc plays in the cellular response to hypoxia makes
it the target of intense investigation. Besides its bHLH-PAS domains, the
transcription factor also contains a transcriptional activation domain (TAD)
towards the carboxy terminus. The TAD is divided into two separate functional
regions, the more N-terminal portion, N-TAD and the more C-terminal, C-TAD
(Figure 2). The bHLH acts as a contact surface for DNA and a primary
interaction surface between HIF1a and ARNT or ARNT2. The PAS domain of
HIF1a is a secondary dimerization surface thought to provide specificity to the
HIF1a:ARNT interaction (123-125). Finally, HIF1a contains an oxygen-
dependent degradation domain (ODD) that is responsible for regulating the

protein’s stability under varying oxygen tensions.

II.B.1. Regulation of Hypoxia Inducible Factors
HIF1a expression and function is translationally and post-translationally
regulated. ARNT is constitutively expressed protein and is not known to play a

major role in the regulation of HIF1 activity. Under normal oxygen tension, HIF1a

28

$52838.

2...”: “:1 9553:. .98". use 301% 83.386? 3.9g “:2 >3 3x063: mezzo 83.9663 Be

55 8:28. oc§mam< ecu 8:05 9: 9m 2 ucm n. 5.028032 EmEoQ cozm>=o< 3:25:85; $58.2
-o ecu .mc_:=9.z d<ho can 05.2 _ cozmuflmoc Emu—Loewe com>xo -ooo . CE<.E_w.Lon_ -w<n_ _x__o.._-aoo..
3.2.. 2me I I.=._n M 9a 2.58% 26:? 9: "E89... BEE 05 no 53558.52 32.320 .N 2:2”.

 

29

mRNA is constitutively expressed, translated and the protein is quickly targeted
for proteosome-mediated degradation by an ubiquitin-dependent process (126-

128).

lI.B.1.a. Transcriptional and Translational Regulation of HIF

In almost all cell types, hypoxia leads to HIF1a protein stabilization whereas
there is no change in the mRNA levels. A variety of other factors, such as
epidermal growth factor, fibroblast growth factor 2, insulin, insulin like growth
factor 1 and 2, and interleukin 1a induce the expression of HIF1oc, HIF1 DNA
binding and HIF1 target gene expression under normoxic'conditions (129-131).
This is mostly through phosphoinositide-3 kinase (PI3K) signaling pathway,
which increases the protein levels of HIF1a by increased translation. This
response varies in different cell types (132, 133). Inhibition of mammalian target
of rapamycin kinase (mTOR), a kinase that works downstream of PI3K and AKT,

reduces the expression of HIF1a (134).

II.B.1.b. Post-Translational Regulation of Hypoxia Inducible Factors

HIF1a-dependent signaling and transcriptional activity is primarily regulated via
protein stability and post-translational modifications. Originally, it was thought
that this process required a heme containing protein, since HIF1a stability can be
influenced by carbon monoxide, iron-like metals and iron chelators (135, 136).

Recently, the signaling mechanism controlling HIF1a stability has come to light.

30

It involves motifs within the ODD, ubiquitination and various post-translational

modifying enzymes.

The degradation of HIF1a requires the two ODD motifs found in the carboxy
terminus of protein (Figure 2) (127). These ODD motifs each contain a proline
residue that, when hydroxylated, creates a binding surface for the Von Hippel
Lindau (VHL) tumor suppressor protein. Upon binding HIF1a, VHL recruits the
necessary ubiquitination machinery required for tagging the transcription factor
for degradation. In fact, these 0008 can render non-homologous proteins labile
in the presence of oxygen (127). The hydroxylation of.the conserved prolines in

the ODD are performed by a family of non-heme iron-dependent hydroxylases.

The family of prolyl hydroxylase domain containing proteins (PHDs) that modify
HIF1a was first identified in C. elegans and later shown to have three
mammalian homologs (137, 138). The three hydroxylase family members
(PHD1-3) are iron and a-ketoglutarate (or-KG) dependent enzymes that utilize
oxygen as a co-substrate (139). PHDs hydroxylate prolines within an LXXLAP
motif (i.e. human HIF1or residues 402 and 564 and mouse HIF1a residues 402
and 577 (137, 138, 140). These enzymes also require ascorbate to maintain
proper function. During the reaction cycle, the a-KG is decarboxylated to yield
succinate and the oxygen is transferred to the proline. The involvement of TCA
cycle intermediates (i.e. a—KG and succinate) directly links HIF1a stability to the

metabolic state of the cell. This link led researchers to hypothesize that hypoxia

31

signaling is involved in cancer that develops in families that carry mutations in
genes such as succinate dehydrogenase and fumarate hydratase. VHL can only
interact with HIF1a and direct its degradation after HIF1a is hydroxylated. Under
hypoxic conditions, the PHDs are unable to hydroxylate the cytoplasmic HlFs
(141). The unmodified HIFs are stabilized and signaled by an unknown
mechanism to translocate to the nucleus, where they are free to form dimers with
ARNT or ARNT2 (128, 142). The HleARNT heterodimers bind DNA at
sequence specific sites termed hypoxia response elements (HREs) of target

genes (143) (Figure 3).

HIF1a is also functionally regulated by hydroxylation. Hypoxic conditions
promote the ability of the CTAD of HIF1a to interact with the transcriptional co-
activator, p300 (144, 145). Hydroxylation of a conserved asparagine (residue
803 in hH|F1or and 813 in mHlF1 a) in the CTAD inhibits the interaction between
CTAD and p300, preventing the recruitment of the basal transcription machinery
under norrnoxic conditions (145-147). HIF asparaginyl hydroxylase also known
as factor inhibiting HIF1 (FIH1). This enzyme requires oxygen and a—KG as
substrates and Fe2+ as a co—factor. Its reaction mechanism is similar to the
PHDs, and FIH can also be inhibited by iron chelators, Co” and hypoxia (140,
148-150). Tight control of the HIF transcriptional response under hypoxia is
achieved by the dual hydroxylation steps, coupled with difference in Km values for
02 of the PHDs and FIH. PHDs that hydroxylate prolines, causing proteosomal

degradation have a higher Km value for 02 than the asparagine hydroxylase, FIH

32

ES

mac—.0 u. 1:39 0:0 15.050 £95.25.

.03.... 320302.000 $19 .0 00000.0 0." 38.2.8 00:00200 «00.0000 .: 1:39 50. .0000 .0 :0 <.....-
00002.03 0030...:03: 0:0 009000.69 .uIU «000.400 0» 0:0 sin. 00 0:00:00 0:0 :0? 0m 0 00.008:
c:a0q :<00x.0 01: 50 0.00020 0.. .8: 0.30.088 0. 0.0.0:. 300.0 ..x0 0000.. 0:0 220.. .uIUm 0.0 :0.
00.20. .0098 .0 1.39 0.00._.N0..0:. 2.0.00. ”8:068:09 0.303020: 2...: >32... 0:0 5:30.072.
203020.63. 00.2020: 0. 3.0050 «0000:020 00:00. >352 0560: 0:0 .u0 00002.02 0:N<30 .u....
3992.000 0 00:00200 000040.030 400.90 600.00 .: 50 #0501330. 00..<0».0: 0030.: 999 0:00.
:0::0x.0 0~0<0:..:c 50 5.0.09.0: 00.2.00: 1.39 0:0 0009208.. 023300 320050.

 

 

\\

050.: J

 

 

 

 

  

I...

 

 

 

hzw IIPI E...

 

0E...—

If

 

0.80.2.0 :0...

J

0:.
I U

a

 

 

£2.00

2.80»... k

 

 

f
\\ 0....0042. J *W”

 

IX maggonnO

5U

 

30.5:330xé
+~ 0“.
\ 002.5

00.32005... .0205 0.080....—

 

 

 

f

1......)
r

I.“—

 

O

 

.0...
_

:

-fi 0....
2...: \

suompuoo uOIXOdAHu

VIXOWUON

34

(151). The difference in these two enzymes allows the HlF1a protein to be

stable at moderate hypoxia while transcriptionally active only at severe hypoxia.

Studies indicate that HlFs undergo other covalent modifications, such as
phosphorylation. P42, AKT and MAP kinases were found to phosphorylate
HlF1a and regulate its activity (152-154). Recently Mylonis et. al. showed that
serine-641 and serine-643 of human HIF1a were targets of phosphorylation by
MAP kinase, which promotes the nuclear accumulation of the protein (155). It
was also shown that treatment with MAPK inhibitors can inhibit both nuclear

accumulation and transcriptional activity of HlF1a.

In addition to hydroxylation and phosphorylation it has been reported that in
mammalian cells ARD1, an acetyl transferase, modifies a lysine (residue 532 in
hHIF1 a) in the ODD of HIF1 a. This modification is thought to enhance the
interaction between HIF1a and VHL and leads to augment HlF1a ubiquitination
(156). Recent reports, however, suggest that the effect of ARD1 on HlF1oc are
cell type-specific and its overall role in hypoxia-mediated gene transcription

requires further clarification (157, 158).

".32. Regulation of gene expression by HIF1a
Since the discovery of HIF1a as the transcriptional factor that binds to the
promoter of erythropoietin genes, a large number of genes have been found to

be directly regulated by HIF1oc (159). Under hypoxic conditions HIF1a protein

35

that escapes proteolytic cleavage enters the nucleus, binds ARNT and contact
DNA at HREs. This assembly helps in the recruitment of transcriptional
mediators such as CBP/p300, which in turn recruits the basal transcriptional
machinery (147). The library of genes influenced by HIF1 is varied and can be

categorized into survival response or cell death response groups.

Il.B.2.a. Survival Response

Under hypoxia, the stabilized HIF1a protein rapidly induces the expression of
genes that increase the oxygen availability to cells such as erythropoietin (EPO)
that promotes erythropoiesis. It also induce vascularization by expressing
vascular endothelial growth factor (VEGF) (160). In general these genes, along
with others such as adrenomedullin, (161), IGF2, a potent mitogen (162) and
carbonic anhydrase, a regulator of intracellular pH (163), promote cell survival
under hypoxic stress. In addition, HIF1 plays a central role in the switch from
aerobic to anaerobic metabolism during hypoxic stress by promoting the
expression of glycolytic enzymes, such as hexokinase, enolase1, glyceraldehyde
phosphate dehydrogenase (GAPDH), lactate dehydrogenaseoc(LDHa), and
phosphoglycerate kinasel (PGK1) (79). This switch to glycolytic centered
energy production is further facilitated by the increased expression of various
glucose transporters such as glut-1 (164). HIF1or also plays a major role in
growth and development by promoting cell division by upregulating cyclinGZ,

TGFa and IGF2, as well as wound healing through increased expression of

36

extracellular matrix proteins such as prolyl-4-hydroxylase a(P4Ha), matrix

metalloprotease 2 (MMP2), and plasminogen activator inhibitor1(PAl1) (165).

ll.B.2.b. Cell death response

Among more than 70 genes that are HRE:HIF1 regulated, a majority of them
have a role in promoting cell survival under hypoxia. HlF1a, however, also
regulates the expression of a battery of genes involved in the promotion of cell
death, such as BCL2/adenovirus E1B 19kDa interacting protein 3 (BNip3),
BNip3-like (BNip3L) and RTP801 (166, 167). BNIP3 and BNIP3L are BH3-
domain containing proteins that associate with the mitochondrial outer membrane
and can cause the membrane pore to open, leading to loss of membrane
potential and cell death in a necrotic manner (168). In primary cardiac myocytes
BNIP3 is also known to induce apoptotic cell death under severe hypoxia (169).
The necrotic cell death requires a decreased cellular pH when tested in vitro and
interestingly, chronic and severe hypoxia can produce tissue acidity (170).
Intracellular acidity is also linked to genetic mutations leading to an increased risk
of cellular transformation. The expression of BNIP3 may be a method to
eliminate potential pre-malignant cells by forcing them to undergo cell death

(171).

RTP801 was seen to be regulated in a cell-specific manner by HIF1a in an

ischemic animal model (167). In non-dividing neuron-like cells, RTP801

expression caused cell death, while in MCF7 cells, RT801 showed cellular

37

protection against glucose deprivation or hypoxia when it was expressed prior to
the incident. So by regulating a mix of adaptation and cell death promoting

genes, HIF1a may play an important role in the survival of the organism.

The role of HIF1oc in maintaining the balance between survival and cell death is a
critical step that is misregulated in various human diseases and toxic insults. In
various solid tumors, the rapid growth of tumor cells and resulting hypoxia favors
an adaptive response that might promote even faster tumor growth. Similarly in
solid tumors, necrosis of hypoxic regions leads to accumulation of inflammatory
immune cells leading to conditions favorable for genetic instability (172-174).
Various divalent metals such as cobalt and nickel exposure can also lead to

abnormal regulation of hypoxic balance between cell survival and cell death.

ll.C. Role of hypoxic signaling in human diseases

In higher organisms, hypoxia resulting from localized or global deficiency in
oxygen can lead to life-threatening complications. Organs that have a high
metabolic rate, such as the brain and heart, are extremely susceptible to
hypoxia-induced damage. For example, the pathological complications of stroke
(cerebral ischemia) and heart infarction (myocardial ischemia) arelargely due to
hypoxia-induced and reperfusion injuries. The influx of oxygen upon tissue
reoxygenation after hypoxia produces reactive oxygen species that interact with
proteins and lipids, causing cellular damage (175). In addition, hypoxia and

HIF1a signaling play a central role in cellular transformation and tumorigenesis.

38

The brain offers a special case, given its high energy use and dependence on
proper blood flow to supply the nutrients adequately to maintain this high energy
production. Stroke-induced damage to the brain in which, the brain stem,
hippocampus and cerebral cortex are most susceptible, can cause irreversible
brain injury, brain cell death, paralysis and death (176). Even reperfusion, which
is required to protect the brain tissues, causes cell death as it produces reactive
oxygen species leading to inflammatory cell infiltration (177). In case of a short,
not so severe exposure to hypoxia, a reversible phenomenon, called “penumbra”,
occurs and is characterized by suppression of protein synthesis and
spontaneous electrical activity (176). Brief ischemic injuries are usually
reversible and do not cause necrotic cell death and usually lead to a process

called stunning (178).

Obstructions in coronary arteries can lead to severe incidents of ischemia.
These prolonged ischemic instances can cause necrotic cell death spreading
from subendocardium to subepicardium (179). Immediately after the ischemic
insult, cells shift to glycolysis for energy production (See section II.A.3 p.25).
This shift to anaerobic metabolism increases the production of lactate within the
cell and leads to cellular acidosis and edema. Intracellular Ca2+ also increases
and finally leads to cell necrosis (180). Reperfusion is essential to save the
cardiac myocytes but leads to reperfusion injury; very similar to what happens to

brain cells. The reestablishment of blood flow delivers a burst of oxygen to the

39

affected tissue and the resulting rapid production of reactive oxygen species can
lead to further alterations in Ca2+ homeostasis. This disruption in Ca2+
homeostasis can cause the myocardium to loose its contractile function for a

short period (181).

Hypoxia and HIF1a are essential in tumorigenesis (121, 182). HIF1a is a
positive factor in solid tumor growth. In a tumor transplant model, HIF1a null
cells fail to produce tumor Many human cancers show aberrant HIF1a protein
and hypoxia-mediated transcription (183). The increases in HlF1a signaling
might be the result of the loss of tumor suppressor proteins, such as VHL, or the
hypoxic environment inside solid tumor tissues (183, 184). As tumors grow
larger than 1 mm3, hypoxic regions appear inside the tumor due to an inability of
the tumor to maintain adequate vascularization. HIF1a induces the expression of
glycolytic genes and growth factors, such as VEGF, to enable the tumor to
maintain ATP production and to re-establish blood flow to the hypoXic region of
the tumor, respectively (83). HIF1a also increases the expression of genes
involved in extracellular modification, causing increased metastatic potential
(185). Furthermore, in glioblastomas, HIF1a is overexpressed in viable cells
surrounding the areas of necrosis, suggesting that it might be promoting cell

survival (186).

Increased expression of HlF1a has also been shown to be associated with

aggressive forms of cancer. Increased intracellular levels of HIF1a protein are

40

associated with poor prognosis and resistance to therapy in the case of head and
neck tumor, ovarian cancer and esophageal cancer (187, 188). Ryan et al
showed that tumor growth and angiogenesis were highly reduced by the loss of
HlFla (121). HIF1a is also known to induce multi-drug resistance protein 1 and
P-glycoprotein under hypoxia and might play a role in tumor resistance to
chemotherapy (189). Radiation therapy is used to treat many tumors and
primarily works by promoting the production of ROS. The increased oxidative
stress upon radiation causes DNA strand breaks and ultimately the death of the
cancer cell. Under hypoxic conditions, ROS production is inhibited and studies
have shown that many primary tumors with high levels of HIF1a expression show

increased radiation resistance (183).

Ill. Reactive Oxygen Species

It is important for a cell’s survival to maintain a reducing environment since cells
utilize selective oxidation steps to derive energy for running various biological
processes. At the same time, the oxidation step necessitates the presence of
oxidizing agents such as molecular oxygen. Oxygen is a potent oxidant with two
unpaired electrons and can react with various cellular components such as
proteins, lipids, nucleic acids. It can also react with transition metals and can
produce reactive oxygen species (ROS) which include singlet oxygen (02"),
hydroxyl radicals (OH'), and superoxide anions (02"). Generally these
molecules are highly reactive and easily diffusible (190). Hydrogen peroxide

(H202) is another species which is highly reactive but has no unpaired electron

41

and can be produced from these reactive oxygen species. An important source
of ROS is the mitochondrion, which maintains the apparatus for oxidative
phosphorylation. This paradox of oxygen as a necessary evil has forced cells to
maintain mechanisms to protect themselves from the harmful effects of ROS
while utilizing oxygen for energy production. Some time during evolution
organisms began using these byproducts of aerobic life as messengers for
biological signaling, as well as weapons against hostile invaders (191).
Antioxidant enzymes such as superoxide dismutases, as well as antioxidant
molecules like ascorbic acid and glutathione are main components of the host

defense against cellular ROS damage (192).

"LA. Sources of Reactive Oxygen Species (ROS)

Reactive oxygen species are produced in every cell either through enzymatic or
random chemical reactions (193). The major sources of reactive oxygen species
are mitochondria and transition metal oxidation. There are also other
cytoplasmic enzymes that produce ROS while utilizing oxygen as a substrate for
other chemical reactions. The sources of ROS can be classified as
mitochondrial, transition metal reactions, hypoxia and hypoxia reperfusion, and

extra-mitochondrial sources.

lll.A.1. Mitochondria

Electrons derived from NADH and FADHz, synthesized primarily during the TCA

cycle, flow through the mitochondrial ETC, terminating with the transfer of 4

42

electrons to an oxygen atom, producing water. As these electrons travel through
the ETC, the energy is used to pump protons into the inter-membrane space of
the mitochondria, creating a membrane potential of approximately 200 mV. This
membrane potential is coupled to the production of ATP via ATP synthase (also
known as Complex V). The oxidation steps are not always complete, leading to
the production of superoxide anions (194). Complex 1 and 3 of the ETC, located
in the mitochondrial inner-membrane are the major sites of ROS production (195)
(Figure 4). Ubisemiquinones at these sites transfer the electrons to oxygen,
generating superoxide. Superoxide dismutases in the mitochondria or cytoplasm
inactivate superoxide producing H202, which may in turn be detoxified by
glutathione peroxidase. Any H202 that escapes detoxification might cause
downstream cell signaling or react with cellular components causing damage
(196). Hydroxyl radicals might also be formed from superoxide through Haber-
Weiss reaction or from H202 by Fenton reaction (Figure 1). Hydroxyl radicals are
very reactive and cause extensive damage to nucleotides (197). H202 on the
other hand, is known to be involved in cell signaling pathways acting as a long
range second messenger as it can easily diffuse through the membranes (198).

Complex II of the electron transport chain may also produce superoxide when the
enzyme is damaged by oxidative stress or aging (199). Recent data suggests
that sites other than electron transport chain may produce ROS within the
mitochondria (200). For example, alpha ketoglutarate dehydrogenase, a TCA
cycle enzyme has been shown to produce H202 in isolated brain mitochondria.

Reduced flavins in other flavoprotein enzymes may also produce superoxide

43

.89.... 2... 2:25:05 new woman ocm.nEmE..oE_ 2:. 8380 E 53:80

.3 683.8... 9.0 on. new 83m _m_.uco..oo._c. 8:5 2:. $380 _ onEOo am “.2283 2.6.35
.80... comes .0 concave. anES... .6 8:80.. 5 new _ meEoo .m 82.8.2.3 .0 5:050... 05

w. 282.9. 5.82m .6 8:3 25 .0 638.95 9.0 dracocoozE on. .o ocmBEoE 5:... o... 30.8
2.205 v6 EoEo>oE on. 2 9.2805 .0 .so: 0... $338 .>_-_ moonEoov Ego toamcmz 8.82m och

3.22.8.5 c. 5.832.. mom .o 8.5 .4 2:9”.

. .No

44

Sufiouzm 8225...

226 <8

 

(199). Monoaminooxidase, an enzyme bound to the outer mitochondrial

membrane that oxidizes biogenic amines may also produce H202 (201).

lll.A.2. Transition Metals

Many transition metals are known to produce reactive oxygen species through
Fenton like or Haber-Weiss reactions (15) (Figure 1). Redox active metals such
as iron, cobalt, and nickel, produce superoxide and hydroxyl radical by reacting
with molecular oxygen. Redox inactive metals such as lead, arsenic, mercury
and cadmium are thought to deplete a cell’s sufhydryl antioxidant reserves,
causing oxidative damage, as well as bonding with the sufhydryl group of other

proteins (19).

The redox state of the cell is generally thought to be highly coupled with the iron
redox state. Iron homeostasis is maintained by regulating the absorption,
transport and storage of iron through specific iron binding proteins (202). Free
iron can readily participate in one electron transport reactions, producing
hydroxyl radicals through Fenton-like reactions and these radicals can go on to
cause lipid peroxidation and oxidative damage of proteins and DNA (21). This
also causes the production of a variety of other ROS species by the
decomposing peroxides. The superoxide insult can further increase the free form

of iron by the release of iron from binding proteins such as ferritin.

45

Cobalt ions can form hydroxyl radicals by reacting with endogenous H202,
causing oxidative damage to cells (15). Cobalt metal particles are also able to
produce hydroxyl radicals by reacting with dissolved oxygen in the presence of
superoxide dismutase in EPR spin trapping experiments (203). Surprisingly, in
vitro systems have demonstrated that Co2+ can produce hydroxyl radicals using
Fenton reaction, in the presence of low molecular weight chelators such as
glutathione and anserine (203). Nickel produces much lower levels of ROS.
Both cobalt and nickel’s ability to produce ROS are thought to play a critical role
in their ability to act as hypoxia mimic (204).. Data suggests that NiCI2 and NiS2
can produce ROS. Nickel is also shown to deplete GSH, as well as have pro-
oxidant effect when present along with GSH (205). Nickel-induced oxidative

stress has been linked to DNA damage and cellular transformation.

Although cadmium does not produce significant levels of ROS directly, it is
thought to produce oxidative stress indirectly (19). Through a slow reduction,
cellular glutathione forms a complex with hexavalent chromium and yields Cr(V),
which can bind DNA or produce ROS by a Fenton-like reaction (206). Cr(V) may
be further reduced by cellular antioxidants such as ascorbic acid and glutathione
producing Cr(IV) and Cr(lll) that are capable of producing R08 (207). Cells
utilize the glutathione/glutathione peroxidase system as a scavenger of
intracellular oxygen radicals. Cadmium treatment increased the cellular GSH
levels at moderate doses, however; GSH levels were depleted at high doses of

the metal, presumably due to the oxidative load (208). Cadmium can also inhibit

46

 

the activity of SOD and increase the levels of lipid peroxidation in treated animals
(209). Low concentrations of cadmium treatment in human-hamster hybrid cells
caused the attenuation of the removal of H202, suggesting the inhibition of

peroxide-scavenging enzymes (210).

lll.A.3. Hypoxia and Hypoxia Reperfusion

Cells might experience a lack of oxygen for various reasons, most notably due to
a blockage in blood vessels (ischemia). As described above (section ll.C. p38),
this can have disastrous consequences in case of brain, heart and other organs.
ROS production under ischemia is a paradox. While ROS production is
suppressed in tissues such as neurons during ischemia, its production increases
in heart, lung and skeletal muscle (211). Damage to the mitochondria increases
the ROS production during ischemia because decreased flux through ETC
coupled with damage to the complexes leads to electron leakage and oxyradical
formation (212). Studies with rotenone and antimycin A showed that complex 3
is the primary site of ROS production during hypoxic stress (213). Although a
restoration of blood (oxygen) supply is essential for the survival of the organism,
the sudden arrival of oxygen also produces a burst of reactive oxygen species,
presumably due to inability of the scavenging system to keep up with the sudden
burst in oxygen supply. The restoration of blood flow can result in reperfusion

injury of organs (175).

lll.A.4. Extramitochondrial Sources

47

Various cellular reactions produce ROS as byproducts. Monooxygenases,
cytochrome P-450 (CYP) enzymes are multi-enzyme systems found in the
endoplasmic reticulum and include FAD/FMN containing NADPH-cytochrome
P450 reductase. Enzymes of this CYP superfamily are involved in
biotransformation reactions of xenobiotics and endogenous compounds, such as
steroids. The enzymes perform substrate oxidation reactions in an O2 and
NADPH-dependent manner (214). CYP enzymes produce superoxide and H202
during the oxidation reaction by the accidental release of electrons by the flavins
in the NADPH2P450 enzyme (215). Another important source of ROS is
peroxisomes, which play a major role as antioxidants, primarily scavenging
cytosolic H202. Aged peroxisomes lose catalase activity thus becoming a source
of ROS, as the balance between production and removal is shifted to the
production side (216). The peroxisomal flavin oxidases that are involved in 8-
oxidation reactions of fatty acids can produce H202. Xanthine oxidase,
responsible for catalyzing the 1 electron reduction of molecular 02 using
hypoxanthine, xanthine and NADH, is a major source of ROS (217). NADPH
oxidases are important source of superoxides in activated macrophages,
neutrophils, and endothelial cells (218). The ROS species are used to destroy
invading pathogens by the activated immune cells and for signaling vascular

remodeling by endothelial cells.

III.B. Protection against R08

48

Cells have a well developed system in place to deal with ROS that includes
antioxidant enzymes, such as glutathione peroxidase and superoxide dismutase,
as well as small antioxidant molecules derived from dietary intake of fruits and
vegetables , such as ascorbate and tocopherol (219, 220). This defense system
primarily includes six classes of molecules as described by Beckman and Ames
(221): These include (i) enzymatic scavengers of ROS such as superoxide
dismutases that convert superoxide into H202, catalases that convert H202 into
water and O2, and glutathione peroxidase that decompose H202 to water; (ii)
ascorbate, urate and glutathione that act as hydrophilic ROS scavengers; (iii)
tocopherols, flavonoids, carotenoids and ubiquinol that work in the hydrophobic
environment; (iv) enzymes (e.g. GSH reductase, dehydroascorbate reductase)
that play a critical role in the recycling of small antioxidant molecules and
maintenance of thiol groups of proteins (e.g thioredoxin reductase) under
oxidative stress; (v) enzymes that generate reducing equivalents such as
glucose-6-phosphate dehydrogenase and help to maintain the appropriate
cellular environment by replenishing NADPH levels and (vi) various transcription
factors and other proteins such as NF-KB, MTF-1, NRF-2 that promote an
adaptive response to an ROS insult. Various components of the antioxidant
pathways act in parallel as different antioxidants have similar roles. In addition,
they are highly co-operative. For example, SOD and glutathione peroxidase

work in tandem to convert superoxide into water and 02.

IV. Hypothesis and Specific Aims:

49

In daily life we are exposed to metals that are ubiquitous in our environment and,
once internalized, can influence HIF1 activity through direct PHD inhibition or
Indirectly via ROS generation. The aberrant HIF1 activity can disrupt the normal
balance established by hypoxia signaling between cell death and adaptation,
resulting in cellular damage. Moreover, metal-induced ROS generation might
have other cellular consequences outside of HIF1 modulation that also influence
cell viability. The published literature overviewed earlier and our own
experiments suggest that divalent metal-ions, such as cobalt and nickel, can
mimic hypoxia to the extent that there is a significant overlap between the mRNA
expression profiles of hypoxia and cobalt chloride-treated mouse embryonic
fibroblast cells (MEFs). Understanding how metals, including both hypoxia-
mimics (i.e. cobalt and nickel) and non-hypoxia mimics (i.e cadmium) influence
cell function and viability, and determining the role that the hypoxia signaling
cascade plays in this process has led me to the following hypothesis and specific

aims:

Hypothesis: cobalt and similar divalent metals cause toxicity in MEFs by

virtue of their ability to interfere with hypoxic signaling.

The fogr chagters in this dissertation focus on testing the above hypothesis with
the following specific aims.
A) Characterize the global gene expression profiles under hypoxic

conditions and upon exposure to ‘hypoxia mimics’.

50

1) Determine the global expression changes under hypoxia, cobalt
chloride, and desferoxamine treatments using HepBB cells

2) Characterize the extent of overlap in gene expression profiles between
hypoxia, cobalt chloride, and desferoxamine treatments.

8) Characterize the HIF1a-dependent gene expression profiles following
hypoxia and cobalt chloride treatments using wild type and HIF1a 4-
mouse embryonic fibroblasts.

1) Characterize hypoxia mediated gene expression changes in MEFs.

2) Determine the HlF1a —dependent gene expression changes upon
hypoxia and CoCI2 treatment

3) Identify shared genes between hypoxia and cobalt treated MEFs with
the goal of identifying possible mechanism of action for metal-induced
toxicity

C) Elucidate the role of HIF1a in mediating cobalt induced toxicity in MEFs.

1) Determine the time course and dose response to CoCI2 mediated
toxicity in wild type and HIF1a -/- cells.

2) Determine the changes in mRNA and protein levels of cytotoxic genes
in CoCI2 treated wild type and HlF1a -/- cells.

3) Characterize the morphological changes associated with CoCI2 induced
cell death in wild type and HIF1a -/- cells.

D) Determine the mechanism(s) underlying the difference in toxicity to
cobalt chloride and cadmium chloride in wild type and HIF1a -I- mouse

embryonic fibroblasts.

51

1) Characterize the cytotoxicity to CoCI2 and CdCI2 treatment in wild type
and HIF1a -/- MEFs.

2) Determine the changes in the cell morphology to CoCI2 and CdCI2
treatment in wild type and HIF1a -/- MEFs.

3) Characterize the role of BNip3 in CoCI2 and CdCI2 induced toxicity in
wild type and HlFla -/- MEFs.

4) Determine the levels of oxidative stress in CoCI2 and CdCI2 mediated
toxicity and HIF1a -/- MEFs.

5) Characterize the levels of ROS scavenger enzyme function under
CoCI2 and CdCI2 treatment in wild type and HIF1a -/- MEFs.

6) Determine the levels of cellular antioxidants in CoCI2 and CdCI2 treated

wild type and HIF10L -/- MEFs.

52

References

10.

11.

12.

13.

Shapiro, R. Small molecule interactions were central to the origin of life. Q
Rev Biol, 81: 105-125, 2006.

Bartel, D. P. and Unrau, P. J. Constructing an RNA world. Trends Cell
Biol, 9: M9-M13, 1999.

Muller, U. F. Re-creating an RNA world. Cell Mol Life Sci, 63: 1278-1293,
2006.

Cohen, R, Schelling, D. L., and Stark, M. J. Remarkable similarities
between yeast and mammalian protein phosphatases. FEBS Lett, 250:
601-606, 1989.

Parr, G. J. Man, Metals and Modern Magic, 3rd edition, Vol. 1, p. 238.
Cleveland, OH: American Society for Metals, 1967.

Aitchison, A. A History of Metals, 1 edition, Vol. 2, p. 343. New York:
lnterscience Publishers, Inc., 1960.

Glusker, J. P. Structural aspects of metal liganding to functional groups in
proteins. Adv Protein Chem, 42: 1-76, 1991.

Baker, J. O. Metal-buffered systems. Methods Enzymol, 158: 33-55, 1988.

Nelson, N. Metal ion transporters and homeostasis. Embo J, 18: 4361-
4371, 1999.

Nord, E. P., Brown, S. E., and Crandall, E. D. Characterization of Na+-H+
antiport in type II alveolar epithelial cells. Am J Physiol, 252: 0490-498,
1987.

Waalkes, M. P., Coogan, T. P., and Barter, R. A. Toxicological principles
of metal carcinogenesis with special emphasis on cadmium. Crit Rev
Toxicol, 22: 175-201 , 1992.

Lehninger, A. L. Role of metal ions in enzyme systems. Physiol Rev, 30:
393-429, 1950.

Clementi, M. E., Condo, S. G., Castagnola, M., and Giardina, B.
Hemoglobin function under extreme life conditions. Eur J Biochem, 223:
309-317, 1994.

53

14.

15.

16.

17.

18.

19.

20.
21.

22.

23.

24.

25.

26.

Strickholm, A. and Clark, H. R. Ionic permeability of K, Na, and CI in
crayfish nerve. Regulation by membrane fixed charges and pH. Biophys J,
19: 29-48, 1977.

Stohs, S. J. and Bagchi, D. Oxidative mechanisms in the toxicity of metal
ions. Free Radic Biol Med, 18: 321-336, 1995.

Hamai, D., Bondy, S. C., Becaria, A., and Campbell, A. The chemistry of
transition metals in relation to their potential role in neurodegenerative
processes. Curr Top Med Chem, 1: 541-551, 2001.

Ho, V. T. and Bunn, H. F. Effects of transition metals on the expression of
the erythropoietin gene: further evidence that the oxygen sensor is a heme
protein. Biochem. Biophys. Res. Comm, 223: 175-180, 1996.

Hartwig, A. Carcinogenicity of metal compounds: possible role of DNA
repair inhibition. Toxicol Lett, 102-103: 235-239, 1998.

Ercal, N., Gurer-Orhan, H., and Aykin-Burns, N. Toxic metals and
oxidative stress part I: mechanisms involved in metal-induced oxidative
damage. Curr Top Med Chem, 1: 529-539, 2001.

Barceloux, D. G. Cobalt. J Toxicol Clin Toxicol, 37: 201-206, 1999.

Valko, M., Morris, H., and Cronin, M. T. Metals, toxicity and oxidative
stress. Curr Med Chem, 12: 1161-1208, 2005.

Cobb, A. G. and Schmalzreid, T. P. The clinical significance of metal ion
release from cobalt-chromium metal-on-metal hip joint arthroplasty. Proc
Inst Mech Eng [H], 220: 385-398, 2006.

Wehner, A. P., Busch, R. H., Olson, R. J., and Craig, D. K. Chronic
inhalation of cobalt oxide and cigarette smoke by hamsters. Am Ind Hyg
Assoc J, 38: 338-346, 1977.

Stangl, G. l., Schwarz, F. J., Jahn, B., and Kirchgessner, M. Cobalt-
deficiency—induced hyperhomocysteinaemia and oxidative status of cattle.
Br J Nutr, 83: 3-6, 2000.

Lison, D. Human toxicity of cobalt-containing dust and experimental
studies on the mechanism of interstitial lung disease (hard metal disease).
Critical Reviews in Toxicology, 26: 585-616, 1996.

Brown, T. E. and Meineke, H. A. Presence of an active erythropoietic

factor (erythropoietin) in plasma of rats after prolonged cobalt therapy.
Proc Soc Exp Biol Med, 99: 435-437, 1958.

54

27.

28.

29.

30.

31.

32.

33.

35.

36.

37.

38.

39.

Goldwasser, E., Jacobson, L. 0., Fried, W., and Plzak, L. F. Studies on
erythropoiesis. V. The effect of cobalt on the production of erythropoietin.
Blood, 13: 55-60, 1958.

Kriss, J. P., Carnes, W. H., and Gross, R. T. Hypothyroidism and thyroid
hyperplasia in patients treated with cobalt. J Am Med Assoc, 157: 117-
121, 1955.

Centeno, J. A., Pestaner, J. P., Mullick, F. G., and Virmani, R. An
analytical comparison of cobalt cardiomyopathy and idiopathic dilated
cardiomyopathy. Biol Trace Elem Res, 55: 21-30, 1996.

Jefferson, J. A., Escudero, E., Hurtado, M. E., Pando, J., Tapia, R.,
Swenson, E. R., Prchal, J., Schreiner, G. F., Schoene, R. B., Hurtado, A.,
and Johnson, R. J. Excessive erythrocytosis, chronic mountain sickness,
and serum cobalt levels. Lancet, 359: 407-408, 2002.

Fischer, T. and Rystedt, l. Cobalt allergy in hard metal workers. Contact
Dermatitis, 9: 115-121, 1983.

Lauwerys, R. and Lison, D. Health risks associated with cobalt exposure--
an overview. Sci Total Environ, 150: 1-6, 1994.

Paternain, J. L., Domingo, J. L., and Corbella, J. Developmental toxicity of
cobalt in the rat. J Toxicol Environ Health, 24: 193-200, 1988.

Bucher, J. R., Hailey, J. R., Roycroft, J. R., Haseman, J. K., Sills, R. C.,
Grumbein, S. L., Mellick, P. W., and Chou, B. J. Inhalation toxicity and
carcinogenicity studies of cobalt sulfate. Toxicol Sci, 49: 56-67., 1999.

Nordberg, G. F ., Goyer, R. A., and Clarkson, T. W. Impact of effects of
acid precipitation on toxicity of metals. Environ Health Perspect, 63: 169-
180, 1985.

Sonke, J. E., Hoogewerff, J. A., van der Laan, S. R., and Vangronsveld, J.
A chemical and mineralogical reconstruction of Zn-smelter emissions in
the Kempen region (Belgium), based on organic pool sediment cores. Sci
Total Environ, 292: 101-119, 2002.

Hiscock, S. A. Trends in the uses of cadmium (1970-1979). Ecotoxicol
Environ Saf, 7: 25-32, 1983.

Nriagu, J. 0. Global inventory of natural and anthropogenic emissions of
trace metals to the atmosphere. Nature, 279: 409-411, 1979.

Davis, R. D. Cadmium - a complex environmental problem. Part II.
Cadmium in sludges used as fertilizer. Experientia, 40: 117-126, 1984.

55

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

Carr, J. P. and Murphy, A. M. Cadmium blocks viral invasion in plants. Nat
Cell Biol, 4: E167-168, 2002.

Friberg, L. [Medical effects of toxic metals in the environment].
Lakartidningen, 80: 1707-1713, 1983.

Karpliuk, l. A., Volkova, N. A., Popov, V. l., and Stepanova, E. N.
[Substantiation of daily permissible levels of cadmium intake with food].
Vopr Pitan 70-73, 1987.

Stanescu, D., Veriter, C., Frans, A., Goncette, L., Roels, H., Lauwerys, R.,
and Brasseur, L. Effects on lung of chronic occupational exposure to
cadmium. Scand J Respir Dis, 58: 289-303, 1977.

Lee, J. S. and White, K. L. A review of the health effects of cadmium. Am
J Ind Med, 1: 307-317, 1980.

lshizaki, A. and F unkushima, M. [Studies on "ltai-itai" disease (Review)].
Nippon Eiseigaku Zasshi, 23: 271-285, 1968.

Piscator, M. Role of cadmium in carcinogenesis with special reference to
cancer of the prostate. Environ Health Perspect, 40: 107-120, 1981.

Liu, X. Y., Jin, T. Y., Nordberg, G. F., Rannar, S., Sjostrom, M., and Zhou,
Y. A multivariate study of protective effects of Zn and Cu against
nephrotoxicity induced by cadmium metallothionein in rats. Toxicol Appl
Pharmacol, 114: 239-245, 1992.

Chernoff, N. Teratogenic effects of cadmium in rats. TeratologY. 8: 29-32,
1973.

Heath, J. C. and Daniel, M. R. The Production of Malignant Tumours by
Cadmium in the Rat. Br J Cancer, 18: 124-129, 1964.

Oberdorster, G. Airborne cadmium and carcinogenesis of the respiratory
tract. Scand J Work Environ Health, 12: 523-537, 1986.

Waalkes, M. P. Cadmium carcinogenesis in review. J Inorg Biochem, 79:
241-244., 2000.

Huebers, H. A., Huebers, E., Csiba, E., Rummel, W., and Finch, C. A. The
cadmium effect on iron absorption. Am J Clin Nutr, 45: 1007-1012, 1987.

Watanabe, M., Shiroishi, K., Nishino, H., Shinmura, T., Murase, H., Shoji,
T., Naruse, Y., and Kagamimori, S. An experimental study on the long-
term effect of cadmium in mice fed cadmium-polluted rice with special
reference to the effect of repeated reproductive cycles. Environ Res, 40:
25-46, 1986.

56

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

Binah, 0., Meiri, U., and Rahamimoff, H. The effects of HGC|2 and
mersalyl on mechanisms regulating intracellular calcium and transmitter
release. Eur J Pharmacol, 51: 453-457, 1978.

Chen, Q. X., Zheng, W. 2., Lin, J. Y., Shi, Y., Xie, W. Z., and Zhou, H. M.
Effect of metal ions on the activity of green crab (Scylla serrata) alkaline
phosphatase. Int J Biochem Cell Biol, 32: 879-885, 2000.

Kuznetsova, S. S., Azarkina, N. V., Vygodina, T. V., Siletsky, S. A., and
Konstantinov, A. A. Zinc ions as cytochrome C oxidase inhibitors: two
sites of action. Biochemistry (Mosc), 70: 128-136, 2005.

Torrnanen, C. D. Inhibition of rat liver and kidney arginase by cadmium
ion. J Enzyme lnhib Med Chem, 21: 119-123, 2006.

Bondy, S. C., Guo-Ross, S. X., and Truong, A. T. Promotion of transition
metal-induced reactive oxygen species formation by beta-amyloid. Brain
Res, 799: 91-96, 1998.

Hu, Y., Su, L., and Snow, E. T. Arsenic toxicity is enzyme specific and its
affects on ligation are not caused by the direct inhibition of DNA repair
enzymes. Mutat Res, 408: 203-218, 1998.

lm, J. Y., Paik, S. G., and Han, P. L. Cadmium-induced astroglial death
proceeds via glutathione depletion. J Neurosci Res, 83: 301-308, 2006.

Hughes, W. L. and Dintzis, H. M. Crystallization of the Mercury Dimers of
Human and Bovine Mercaptalbumin. J Biol Chem, 239: 845-849, 1964.

Vallee, B. L. and Ulmer, D. D. Biochemical effects of mercury, cadmium,
and lead. Annu Rev Biochem, 41: 91-128, 1972.

O'Halloran, T. V. Transition metals in control of gene expression. Science,
261: 715-725, 1993.

Pimstone, N. R., Engel, P., Tenhunen, R., Seitz, P. T., Marver, H. S., and
Schmid, R. Inducible heme oxygenase in the kidney: a model for the
homeostatic control of hemoglobin catabolism. J Clin Invest, 50: 2042-
2050, 1971. '

Zhou, P. and Thiele, D. J. Rapid transcriptional autoregulation of a yeast
metalloregulatory transcription factor is essential for high-level copper
detoxification. Genes Dev, 7: 1824-1835, 1993.

Hainaut, P. and Milner, J. A structural role for metal ions in the "wild-type"

conformation of the tumor suppressor protein p53. Cancer Res, 53: 1739-
1742, 1993.

57

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

Hainaut, P., Rolley, N., Davies, M., and Milner, J. Modulation by copper of
p53 conformation and sequence-specific DNA binding: role for Cu(ll)/Cu(l)
redox mechanism. Oncogene, 10: 27-32, 1995.

Goebeler, M., Roth, J., Brocker, E. B., Sorg, C., and Schulze-Osthoff, K.
Activation of nuclear factor-kappa B and gene expression in human
endothelial cells by the common haptens nickel and cobalt. J Immunol,
155: 2459-2467, 1995.

Leonard, S. S., Harris, G. K., and Shi, X. Metal-induced oxidative stress
and signal transduction. Free Radic Biol Med, 37: 1921-1942, 2004.

Durnam, D. M. and Palmiter, R. D. Transcriptional regulation of the mouse
metallothionein-l gene by heavy metals. J Biol Chem, 256: 5712-5716,
1981.

Foulkes, E. C. Renal tubular transport of cadmium-metallothionein. Toxicol
Appl Pharmacol, 45: 505-512, 1978.

Courgeon, A. M., Maisonhaute, C., and Best-Belpomme, M. Heat shock
proteins are induced by cadmium in Drosophila cells. Exp Cell Res, 153:
515-521, 1984.

Goering, P. L., Kish, C. L., and Fisher, B. R. Stress protein synthesis
induced by cadmium-cysteine in rat kidney. Toxicology, 85: 25-39, 1993.

Liu, J., Corton, C., Dix, D. J., Liu, Y., Waalkes, M. P., and Klaassen, C. D.
Genetic background but not metallothionein phenotype dictates sensitivity
to cadmium-induced testicular injury in mice. Toxicol Appl Pharmacol, 176:
1-9, 2001.

Zheng, H., Liu, J., Choo, K. H., Michalska, A. E., and Klaassen, C. D.
Metallothionein-l and -II knock-out mice are sensitive to cadmium-induced
liver mRNA expression of c-jun and p53. Toxicol Appl Pharmacol, 136:
229-235, 1996.

Jin, P. and Ringertz, N. R. Cadmium induces transcription of proto-
oncogenes c-jun and c-myc in rat L6 myoblasts. J Biol Chem, 265: 14061-
14064, 1990.

Goldwasser, E., Jacobson, L. 0., Fried, W., and Plzak, L. Mechanism of
the erythropoietic effect of cobalt. Science, 125: 1085-1086, 1957.

Ladoux, A. and Frelin, C. Cobalt stimulates the expression of vascular

endothelial growth factor mRNA in rat cardiac cells. Biochem Biophys Res
Commun, 204: 794-798, 1994.

58

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

Semenza, G. L., Roth, P. H., Fang, H. M., and Wang, G. L. Transcriptional
regulation of genes encoding glycolytic enzymes by hypoxia-inducible
factor 1. J. Biol. Chem., 269: 23757-23763, 1994.

Salnikow, K., Donald, S. P., Bruick, R. K., Zhitkovich, A., Phang, J. M.,
and Kasprzak, K. S. Depletion of intracellular ascorbate by the
carcinogenic metals nickel and cobalt results in the induction of hypoxic
stress. J Biol Chem, 2004.

Haddad, G. G. a. L., G. Tissue Oxygen Deprivation : From Molecular to
Integrated Function, 1 edition, Vol. 1, p. 1085. New York: Marcel Dekker,
1996.

Lopez-Barneo, J., Pardal, R., and Ortega-Saenz, P. Cellular mechanism
of oxygen sensing. Annu Rev Physiol, 63: 259-287, 2001.

Michiels, C. Physiological and pathological responses to hypoxia. Am J
Pathol, 164: 1875-1882, 2004.

Post, J. M., Hume, J. R., Archer, S. L., and Weir, E. K. Direct role for
potassium channel inhibition in hypoxic pulmonary vasoconstriction. Am J
Physiol, 262: C882-890, 1992.

Dart, C. and Standen, N. B. Activation of ATP-dependent K+ channels by
hypoxia in smooth muscle cells isolated from the pig coronary artery. J
Physiol, 483 (Pt 1): 29-39, 1995.

Rubin, L. J. Pathology and pathophysiology of primary pulmonary
hypertension. Am J Cardiol, 75: 51A-54A, 1995.

Archer, S. L., Weir, E. K., Reeve, H. L., and Michelakis, E. Molecular
identification of 02 sensors and 02-sensitive potassium channels in the
pulmonary circulation. Adv Exp Med Biol, 475: 219-240, 2000.

Peers, C. and Kemp, P. J. Acute oxygen sensing: diverse but convergent
mechanisms in ainNay and arterial chemoreceptors. Respir Res, 2: 145-
149, 2001.

Cutz, E., Speirs, V., Yeger, H., Newman, 0, Wang, D., and Perrin, D. G.
Cell biology of pulmonary neuroepithelial bodies--validation of an in vitro
model. I. Effects of hypoxia and Ca2+ ionophore on serotonin content and
exocytosis of dense core vesicles. Anat Rec, 236: 41-52, 1993.

Shweiki, D., ltin, A., Soffer, D., and Keshet, E. Vascular endothelial growth
factor induced by hypoxia may mediate hypoxia-initiated angiogenesis.
Nature, 359: 843-845, 1992.

59

91.

92.

93.

94.

95.

96.

97.

98.

99.

100.

101.

102.

Leung, D. W., Cachianes, G., Kuang, W. J., Goeddel, D. V., and Ferrara,
N. Vascular endothelial growth factor is a secreted angiogenic mitogen.
Science, 246: 1306-1309, 1989.

Prentice, T. C. and Mirand, E. A. Effect of hypoxia on plasma
erythropoietin in the rabbit. Proc Soc Exp Biol Med, 106: 501-502, 1961.

DeGowin, R. L. and Johnson, S. Effect of endogenous erythropoiein on
replicating hemopoietic stem cells. Proc Soc Exp Biol Med, 126: 442-446,
1967.

Stenmark, K. R., Fagan, K. A., and Frid, M. G. Hypoxia-induced
pulmonary vascular remodeling: cellular and molecular mechanisms. Circ
Res, 99: 675-691, 2006.

Preston, l. R., Hill, N. S., Warburton, R. R., and Fanburg, B. L. Role of 12-
Iipoxygenase in hypoxia-induced rat pulmonary artery smooth muscle cell
proliferation. Am J Physiol Lung Cell Mol Physiol, 290: L367-374, 2006.

Kourembanas, S. and Bernfield, M. Hypoxia and endothelial-smooth
muscle cell interactions in the lung. Am J Respir Cell Mol Biol, 11: 373-
374, 1994.

Chen, Y. F. and Oparil, S. Endothelin and pulmonary hypertension. J
Cardiovasc Pharmacol, 35: 349-53, 2000.

Keith, I. M. The role of endogenous lung neuropeptides in regulation of the
pulmonary circulation. Physiol Res, 49: 519-537, 2000.

Ten, V. S. and Pinsky, D. J. Endothelial response to hypoxia: physiologic
adaptation and pathologic dysfunction. Curr Opin Crit Care, 8: 242-250,
2002.

Rafii, S. and Lyden, D. Therapeutic stem and progenitor cell
transplantation for organ vascularization and regeneration. Nat Med, 9:
702-712, 2003.

Yoon, C. H., Hur, J., Oh, I. Y., Park, K. W., Kim, T. Y., Shin, J. H., Kim, J.
H., Lee, C. S., Chung, J. K., Park, Y. B., and Kim, H. S. Intercellular
adhesion molecule-1 is upregulated in ischemic muscle, which mediates
trafficking of endothelial progenitor cells. Arterioscler Thromb Vasc Biol,
26: 1066-1072, 2006.

Romanic, A. M., Burns-Kurtis, C. L., Gout, B., Berrebi-Bertrand, l., and
Ohlstein, E. H. Matrix metalloproteinase expression in cardiac myocytes
following myocardial infarction in the rabbit. Life Sci, 68: 799-814., 2001.

60

103.

104.

105.

106.

107.

108.

109.

110.

111.

112.

113.

114.

115.

Hochachka, P. W. Defense strategies against hypoxia and hypothermia.
Science, 231: 234-241, 1986.

Tinton, S., Tran-Nguyen, Q. N., and Buc-Calderon, P. Role of protein-
phosphorylation events in the anoxia signal-transduction pathway leading
to the inhibition of total protein synthesis in isolated hepatocytes. Eur J
Biochem, 249: 121-126, 1997.

Buttgereit, F. and Brand, M. D. A hierarchy of ATP-consuming processes
in mammalian cells. Biochem J, 312 (Pt 1): 163-167, 1995.

Boutilier, R. G. and St-Pierre, J. Surviving hypoxia without really dying.
Comp Biochem Physiol A Mol Integr Physiol, 126: 481-490, 2000.

Hue, L. and Rider, M. H. Role of fructose 2,6-bisphosphate in the control
of glycolysis in mammalian tissues. Biochem J, 245: 313-324, 1987.

Marsin, A. S., Bertrand, L., Rider, M. H., Deprez, J., Beauloye, 0, Vincent,
M. F., Van den Berghe, G., Carling, D., and Hue, L. Phosphorylation and
activation of heart PF K-2 by AMPK has a role in the stimulation of
glycolysis during ischaemia. Curr Biol, 10: 1247-1255, 2000.

Minchenko, A., Leshchinsky, l., Opentanova, l., Sang, N., Srinivas, V.,
Arrnstead, V., and Caro, J. Hypoxia-inducible factor-1-mediated
expression of the 6-phosphofructo-2- kinase/fructose-2,6-bisphosphatase-
3 (PFKFB3) gene. Its possible role in the Warburg effect. J Biol Chem,
277: 6183-6187., 2002.

Kemp, B. E., Mitchelhill, K. l., Stapleton, D., Michell, B. J., Chen, Z. P.,
and Witters, L. A. Dealing with energy demand: the AMP-activated protein
kinase. Trends Biochem Sci, 24: 22-25, 1999.

Hardie, D. G. and Hawley, S. A. AMP-activated protein kinase: the energy
charge hypothesis revisited. Bioessays, 23: 1112-1119, 2001.

Ramaiah, A., Hathaway, J. A., and Atkinson, D. E. Adenylate as a
Metabolic Regulator. Effect on Yeast Phosphofructokinase Kinetics. J Biol
Chem, 239: 3619-3622, 1 964.

Hardie, D. G. Metabolic control: a new solution to an old problem. Curr
Biol, 10: R757-759, 2000.

Bartley, W., Clegg, E. J., and Butler, V. M. Enzyme changes in mice under
hypoxic conditions. Biochem J, 109: 48P, 1968.

Wang, G. L. and Semenza, G. L. General involvement of hypoxia-

inducible factor 1 in transcriptional response to hypoxia. Proc. Natl. Acad.
Sci. USA, 90: 4304-4308, 1993.

61

116.

117.

118.

119.

120.

121.

122.

123.

124.

125.

126.

Wang, G. L. and Semenza, G. L. Characterization of hypoxia-inducible
factor 1 and regulation of DNA binding activity by hypoxia. J. Biol. Chem.,
268: 21513-21518, 1993.

Hogenesch, J. 8, Chan, W. C., Jackiw, V. H., Brown, R. C., Gu, Y.-Z.,
Pray-Grant, M., Perdew, G. H., and Bradfield, C. A. Characterization of a
subset of the basic-helix-loop-helix-PAS superfamily that interact with
components of the dioxin signaling pathway. J. Biol. Chem., 272: 8581-
8593, 1997.

Pongratz, l., Antonsson, C., Whitelaw, M. L., and Poellinger, L. Role of the
PAS domain in regulation of dimerization and DNA binding specificity of
the dioxin receptor. Molecular & Cellular Biology, 18: 4079-4088, 1998.

He, 0., Cheng, P., Yang, Y., Wang, L., Gardner, K. H., and Liu, Y. White
collar-1, a DNA binding transcription factor and a light sensor. Science,
297: 840-843, 2002.

Crews, S. T. and Fan, C. M. Remembrance of things PAS: regulation of
development by bHLH-PAS proteins. Curr Opin Genet Dev, 9: 580-587,
1999.

Ryan, H. E., Lo, J., and Johnson, R. S. HIF-1 alpha is required for solid
tumor formation and embryonic vascularization. EMBO Journal, 17: 3005-
3015, 1998.

Kozak, K. R., Abbott, B., and Hankinson, O. ARNT-deficient mice and
placental differentiation. Developmental Biology, 191: 297-305, 1997.

Jiang, B. H., Zheng, J. 2., Leung, S. W., Roe, R., and Semenza, G. L.
Transactivation and inhibitory domains of hypoxia-inducible factor 1a|pha.

Modulation of transcriptional activity by oxygen tension. Journal of
Biological Chemistry, 272: 19253-19260, 1997.

Jiang, B. H., Rue, E., Wang, G. L., Roe, R., and Semenza, G. L.
Dimerization, DNA binding, and transactivation properties of hypoxia-
inducible factor 1. Journal of Biological Chemistry, 271 : 17771-17778,
1996.

Ruas, J. L., Poellinger, L., and Pereira, T. Functional analysis of hypoxia-
inducible factor-1 alpha-mediated transactivation. Identification of amino
acid residues critical for transcriptional activation and/or interaction with
CREB-binding protein. J Biol Chem, 277: 38723-38730, 2002.

Salceda, S. and Caro, J. Hypoxia-inducible factor 1a|pha (HIF-1a|pha)
protein is rapidly degraded by the ubiquitin-proteasome system under
norrnoxic conditions. Its stabilization by hypoxia depends on redox-
induced changes. J Biol Chem, 272: 22642-22647, 1997.

62

127.

128.

129.

130.

131.

132.

133.

134.

135.

136.

Huang, L. E., Gu, J., Schau, M., and Bunn, H. F. Regulation of hypoxia-
inducible factor 1a|pha is mediated by an OZ-dependent degradation
domain via the ubiquitin-proteasome pathway. Proc. Natl. Acad. Sci. USA,
95: 7987-7992, 1998.

Kallio, P. J., Wilson, W. J., S., O. B., Makino, Y., and Poellinger, L.
Regulation of the hypoxia-inducible transcription factor 1a|pha by the
ubiquitin-proteasome pathway. J. Biol. Chem., 274: 6519-6525, 1999.

Stiehl, D. P., Jelkmann, W., Wenger, R. H., and Hellwig-Burgel, T.
Normoxic induction of the hypoxia-inducible factor 1a|pha by insulin and
interleukin-1beta involves the phosphatidylinositol 3-kinase pathway.
FEBS Lett, 512: 157-162, 2002.

Richard, D. E., Berra, E., and Pouyssegur, J. Nonhypoxic pathway
mediates the induction of hypoxia-inducible factor 1a|pha in vascular
smooth muscle cells. J Biol Chem, 275: 26765-26771, 2000.

Thornton, R. D., Lane, P., Borghaei, R. C., Pease, E. A., Caro, J., and
Mochan, E. Interleukin 1 induces hypoxia-inducible factor 1 in human
gingival and synovial fibroblasts. Biochem J, 350 Pt 1: 307-312, 2000.

Arsham, A. M., Plas, D. R., Thompson, C. B., and Simon, M. C.
Phosphatidylinositol 3-kinase/Akt signaling is neither required for hypoxic
stabilization of HIF-1 alpha nor sufficient for HIF-1-dependent target gene
transcription. J Biol Chem, 277: 15162-15170, 2002.

Gomez-Manzano, C., Fueyo, J., Jiang, H., Glass, T. L., Lee, H. Y., Hu, M.,
Liu, J. L., Jasti, S. L., Liu, T. J., Conrad, C. A., and Yung, W. K.
Mechanisms underlying PTEN regulation of vascular endothelial growth
factor and angiogenesis. Ann Neurol, 53: 109-117, 2003.

Laughner, E., Taghavi, P., Chiles, K., Mahon, P. C., and Semenza, G. L.
HER2 (neu) signaling increases the rate of hypoxia-inducible factor 1a|pha
(HIF-1 alpha) synthesis: novel mechanism for HIF-1-mediated vascular
endothelial growth factor expression. Mol Cell Biol, 21 : 3995-4004, 2001.

Liu, Y., Christou, H., Morita, T., Laughner, E., Semenza, G. L., and
Kourembanas, S. Carbon monoxide and nitric oxide suppress the hypoxic
induction of vascular endothelial growth factor gene via the 5' enhancer.
Journal of Biological Chemistry, 273: 15257-15262, 1998.

Wang, G. L. and Semenza, G. L. Desferrioxamine induces erythropoietin
gene expression and hypoxia-inducible factor 1 DNA-binding activity:
implications for models of hypoxia signal transduction. Blood, 82: 3610-
3615, 1993.

63

137.

138.

139.

140.

141.

142.

143.

144.

145.

146.

Epstein, A. C., Gleadle, J. M., McNeill, L. A., Hewitson, K. S., O'Rourke,
J., Mole, D. R., Mukherji, M., Metzen, E., Wilson, M. l., Dhanda, A., Tian,
Y. M., Masson, N., Hamilton, D. L., Jaakkola, P., Barstead, R., Hodgkin,
J., Maxwell, P. H., Pugh, C. W., Schofield, C. J., and Ratcliffe, P. J. C.
elegans EGL-9 and mammalian homologs define a family of dioxygenases
that regulate HIF by prolyl hydroxylation. Cell, 107: 43-54., 2001.

Bruick, R. K. and McKnight, S. L. A conserved family of prolyl-4—
hydroxylases that modify HIF. Science, 294: 1337-1340., 2001 .

Ryle, M. J. and Hausinger, R. H. Non-heme iron oxygenases. Curr. Opin.
Chem. Biol, 6: 193-201, 2002.

Lando, D., Peet, D. J., Gorman, J. J., Whelan, D. A., Whitelaw, M. L., and
Bruick, R. K. FIH-1 is an asparaginyl hydroxylase enzyme that regulates
the transcriptional activity of hypoxia-inducible factor. Genes Dev, 16:
1466-1471, 2002.

Pereira, T., Zheng, X., Ruas, J. L., Tanimoto, K., and Poellinger, L.
Identification of residues critical for regulation of protein stability and the
transactivation function of the hypoxia-inducible factor-1 alpha by the von
Hippel-Lindau tumor suppressor gene product. J Biol Chem, 278: 6816-
6823, 2003.

Kallio, P. J., Okamoto, K., S., O. B., Carrero, P., Makino, Y., Tanaka, H.,
and Poellinger, L. Signal transduction in hypoxic cells: inducible nuclear
translocation and recruitment of the CBP/p300 coactivator by the hypoxia-
inducible factor-1a|pha. EMBO Journal, 17: 6573-6586, 1998.

Takahata, S., Sogawa, K., Kobayashi, A., Ema, M., Mimura, J., Ozaki, N.,
and Fujii-Kuriyama, Y. Transcriptionally active heterodimer formation of an
Arnt-like PAS protein, Arnt3, with HIF-1a, HLF, and clock. Bioch. Biophys.
Res. Comm, 248: 789-794, 1998.

Freedman, S. J., Sun, Z. Y., Poy, F., Kung, A. L., Livingston, D. M.,
Wagner, G., and Eck, M. J. Structural basis for recruitment of CBP/p300
by hypoxia-inducible factor-1 alpha. Proc Natl Acad Sci U S A, 99: 5367-
5372, 2002.

Dames, S. A., Martinez-Yamout, M., De Guzman, R. N., Dyson, H. J., and
Wright, P. E. Structural basis for Hif-1 alpha /CBP recognition in the
cellular hypoxic response. Proc Natl Acad Sci U S A, 99: 5271-5276,
2002.

Sang, N., Fang, J., Srinivas, V., Leshchinsky, |., and Caro, J. Carboxyl-
terminal transactivation activity of hypoxia-inducible factor 1 alpha is
governed by a von Hippel-Lindau protein-independent, hydroxylation-
regulated association with p300/CBP. Mol Cell Biol, 22: 2984-2992, 2002.

64

147.

148.

149.

150.

151.

152.

153.

154.

155.

156.

Arany, 2., Huang, L. E., Eckner, R., Bhattacharya, S., Jiang, C., Goldberg,
M. A., Bunn, H. F., and Livingston, D. M. An essential role for p300/CBP in
the cellular response to hypoxia. Proceedings of the National Academy of
Sciences of the United States of America, 93: 12969-12973, 1996.

Hewitson, K. S., McNeill, L. A., Riordan, M. V., Tian, Y. M., Bullock, A. N.,
Welford, R. W., Elkins, J. M., Oldham, N. J., Bhattacharya, S., Gleadle, J.
M., Ratcliffe, P. J., Pugh, C. W., and Schofield, C. J. Hypoxia-inducible
Factor (HIF) Asparagine Hydroxylase ls Identical to Factor Inhibiting HIF
(FIH) and Is Related to the Cupin Structural Family. J Biol Chem, 277:
26351-26355, 2002.

McNeill, L. A., Hewitson, K. S., Gleadle, J. M., Horsfall, L. E., Oldham, N.
J., Maxwell, P. H., Pugh, C. W., Ratcliffe, P. J., and Schofield, C. J. The
use of dioxygen by HIF prolyl hydroxylase (PHD1). Bioorg Med Chem Lett,
12: 1547-1550., 2002.

Lee, C., Kim, S. J., Jeong, D. G., Lee, S. M., and Ryu, S. E. Structure of
human FIH-1 reveals a unique active site pocket and interaction sites for
HIF-1 and von Hippel-Lindau. J Biol Chem, 278: 7558-7563, 2003.

Koivunen, P., Hirsila, M., Gunzler, V., Kivirikko, K. l., and MyItharju, J.
Catalytic properties of the asparaginyl hydroxylase (FIH) in the oxygen
sensing pathway are distinct from those of its prolyl 4-hydroxylases. J Biol
Chem, 279: 9899-9904, 2004.

Richard, D. E., Berra, E., Gothie, E., Roux, D., and Pouyssegur, J.
p42/p44 mitogen-activated protein kinases phosphorylate hypoxia-
inducible factor 1a|pha (HIF-1a|pha) and enhance the transcriptional
activity of HIF-1. J Biol Chem, 274: 32631-32637, 1999.

Sodhi, A., Montaner, S., Miyazaki, H., and Gutkind, J. S. MAPK and Akt
act cooperatively but independently on hypoxia inducible factor-1 alpha in
rasV12 upregulation of VEGF. Biochem Biophys Res Commun, 287: 292-
300, 2001.

Gradin, K., Takasaki, C., Fujii-Kuriyama, Y., and Sogawa, K. The
transcriptional activation function of the H IF-Iike factor requires
phosphorylation at a conserved threonine. J Biol Chem, 277: 23508-
23514., 2002.

Mylonis, |., Chachami, G., Samiotaki, M., Panayotou, G., Paraskeva, E.,
Kalousi, A., Georgatsou, E., Bonanou, S., and Simos, G. Identification of
MAPK phosphorylation sites and their role in the localization and activity of
hypoxia-inducible factor-1 alpha. J Biol Chem, 281: 33095-33106, 2006.

Jeong, J. W., Bae, M. K., Ahn, M. Y., Kim, S. H., Sohn, T. K., Bae, M. H.,
Yoo, M. A., Song, E. J., Lee, K. J., and Kim, K. W. Regulation and

65

157.

158.

159.

160.

161.

162.

163.

164.

165.

166.

destabilization of HIF-1a|pha by ARD1-mediated acetylation. Cell, 111:
709-720., 2002.

Fisher, T. 8., Etages, S. D., Hayes, L., Crimin, K., and Li, B. Analysis of
ARD1 Function in Hypoxia Response Using Retroviral RNA Interference.
J. Biol. Chem., 280: 17749-17757, 2005.

Kim, S. H., Park, J. A., Kim, J. H., Lee, J. W., Seo, J. H., Jung, B. K.,
Chun, K. H., Jeong, J. W., Bae, M. K., and Kim, K. W. Characterization of
ARD1 variants in mammalian cells. Biochem Biophys Res Commun, 340:
422-427, 2006.

Wang, G. L. and Semenza, G. L. Molecular basis of hypoxia-induced
erythropoietin expression. Current Opinion in Hematology, 3: 156-162,
1996.

Forsythe, J. A., Jiang, B. H., Iyer, N. V., Agani, F., Leung, S. W., Koos, R.
D., and Semenza, G. L. Activation of vascular endothelial growth factor
gene transcription by hypoxia-inducible factor 1. Mol. Cell. Biol., 16: 4604-
4613, 1996.

Nguyen, S. V. and Claycomb, W. C. Hypoxia regulates the expression of
the adrenomedullin and HIF-1 genes in cultured HL-1 cardiomyocytes.
Biochem Biophys Res Co, 265: 382-386., 1999.

Feldser, D., Agani, F., Iyer, N. V., Pak, B., Ferreira, G., and Semenza, G.
L. Reciprocal positive regulation of hypoxia-inducible factor 1a|pha and
insulin-like growth factor 2. Cancer Res, 59: 3915-3918, 1999.

Wykoff, C. C., Beasley, N. J., Watson, P. H., Turner, K. J., Pastorek, J.,
Sibtain, A., Wilson, G. D., Turley, H., Talks, K. L., Maxwell, P. H., Pugh, C.
W., Ratcliffe, P. J., and Harris, A. L. Hypoxia-inducible expression of
tumor-associated carbonic anhydrases. Cancer Res, 60: 7075-7083,
2000.

Ebert, B. L., Firth, J. D., and Ratcliffe, P. J. Hypoxia and Mitochondrial
Inhibitors Regulate Expression of Glucose Transporter-1 via Distinct Cis-
acting Sequences. J. Biol. Chem., 270: 29083-29089, 1995.

Semenza, G. L. Targeting HIF-1 for cancer therapy. Nat Rev Cancer, 3:
721-732, 2003.

Sowter, H. M., Ratcliffe, P. J., Watson, P., Greenberg, A. H., and Harris,
A. L. HIF1-dependent regulation of hypoxic induction of the cell death
factors BNIP3 and NIX in human tumors. Cancer Res, 61: 6669-6673.,
2001.

66

167.

168.

169.

170.

171.

172.

173.

174.

175.

176.

177.

178.

Shoshani, T., Faerman, A., Mett, l., Zelin, E., Tenne, T., Gorodin, S.,
Moshel, Y., Elbaz, S., Budanov, A., Chajut, A., Kalinski, H., Kamer, l.,
Rozen, A., Mor, 0., Keshet, E., Leshkowitz, D., Einat, P., Skaliter, R., and
Feinstein, E. Identification of a novel hypoxia-inducible factor 1-responsive
gene, RTP801, involved in apoptosis. Mol Cell Biol, 22: 2283-2293, 2002.

Vande Velde, C., Cizeau, J., Dubik, D., Alimonti, J., Brown, T., Israels, S.,
Hakem, R., and Greenberg, A. H. BNIP3 and genetic control of necrosis-

like cell death through the mitochondrial permeability transition pore. Mol

Cell Biol, 20: 5454-5468, 2000.

Kubasiak, L. A., Hernandez, O. M., Bishopric, N. H., and Webster, K. A.
Hypoxia and acidosis activate cardiac myocyte death through the Bcl-2
family protein BNIP3. Proc Natl Acad Sci U S A, 99: 12825-12830, 2002.

Kubasiak, L. A., Hernandez, O. M., Bishopric, N. H., and Webster, K. A.
Hypoxia and acidosis activate cardiac myocyte death through the Bcl-2
family protein BNIP3. Proc Natl Acad Sci USA, 99: 12825-12830, 2002.

Cifone, M. A., Myhr, B., Eiche, A., and Bolcsfoldi, G. Effect of pH shifts on
the mutant frequency at the thymidine kinase locus in mouse lymphoma
L5178Y TK+/- cells. Mutat Res, 189: 39-46, 1987.

Yuan, J., Narayanan, L., Rockwell, S., and Glazer, P. M. Diminished DNA
repair and elevated mutagenesis in mammalian cells exposed to hypoxia
and low pH. Cancer Res, 60: 4372-4376, 2000.

Murdoch, C. and Lewis, C. E. Macrophage migration and gene expression
in response to tumor hypoxia. Int J Cancer, 117: 701-708, 2005.

Weinmann, M., Belka, C., and Plasswilm, L. Tumour hypoxia: impact on
biology, prognosis and treatment of solid malignant tumours. Onkologie,
27: 83-90, 2004.

Li, C. and Jackson, R. M. Reactive species mechanisms of cellular
hypoxia-reoxygenation injury. Am J Physiol Cell Physiol, 282: 0227-241,
2002.

Erecinska, M. and Silver, l. A. Tissue oxygen tension and brain sensitivity
to hypoxia. Respir Physiol, 128: 263-276, 2001.

Biagas, K. Hypoxic-ischemic brain injury: advancements in the
understanding of mechanisms and potential avenues for therapy. Curr
Opin Pediatr, 11: 223-228, 1999.

Bhatia, G., Sosin, M., Leahy, J. F., Connolly, D. L., Davis, R. C., and Lip,
G. Y. Hibernating myocardium in heart failure. Expert Rev Cardiovasc
Ther, 3: 111-122, 2005.

67

179.

180.

181.

182.

183.

184.

185.

186.

187.

188.

189.

Buja, L. M. Myocardial ischemia and reperfusion injury. Cardiovasc Pathol,
14: 170-175, 2005.

Kloner, R. A. and Jennings, R. B. Consequences of brief ischemia:
stunning, preconditioning, and their clinical implications: part 2.
Circulation, 104: 3158-3167, 2001.

Braunwald, E. and Kloner, R. A. The stunned myocardium: prolonged,
postischemic ventricular dysfunction. Circulation, 66: 1146-1149, 1982.

Ryan, H. E., Poloni, M., McNulty, W., Elson, D., Gassmann, M., Arbeit, J.
M., and Johnson, R. S. Hypoxia-inducible factor-1a|pha is a positive factor
in solid tumor growth. Cancer Res, 60: 4010-4015, 2000.

Aebersold, D. M., Burri, P., Beer, K. T., Laissue, J., Djonov, V., Greiner, R.
H., and Semenza, G. L. Expression of hypoxia-inducible factor-1a|pha: a
novel predictive and prognostic parameter in the radiotherapy of
oropharyngeal cancer. Cancer Res, 61 : 291 1-2916, 2001.

Wiesener, M. S. and Eckardt, K. U. Erythropoietin, tumours and the von
Hippel-Lindau gene: towards identification of mechanisms and dysfunction
of oxygen sensing. Nephrol Dial Transplant, 17: 356-359, 2002.

Dachs, G. U. and Tozer, G. M. Hypoxia modulated gene expression:
angiogenesis, metastasis and therapeutic exploitation. Eur J Cancer, 36:
1649-1660, 2000.

Sondergaard, K. L., Hilton, D. A., Penney, M., Ollerenshaw, M., and
Demaine, A. G. Expression of hypoxia-inducible factor 1a|pha in tumours
of patients with glioblastoma. Neuropathol Appl Neurobiol, 28: 210-217,
2002.

Talks, K. L., Turley, H., Gatter, K. C., Maxwell, P. H., Pugh, C. W.,
Ratcliffe, P. J., and Harris, A. L. The expression and distribution of the
hypoxia-inducible factors HIF-1a|pha and HIF-2alpha in normal human
tissues, cancers, and tumor-associated macrophages. Am J Pathol, 157:
411-421, 2000.

Zhong, H., De Marzo, A. M., Laughner, E., Lim, M., Hilton, D. A., Zagzag,
D., Buechler, P., lsaacs, W. B., Semenza, G. L., and Simons, J. W.
Overexpression of hypoxia-inducible factor 1a|pha in common human
cancers and their metastases. Cancer Res, 59: 5830-5835, 1999.

Wartenberg, M., Ling, F. C., Muschen, M., Klein, F., Acker, H., Gassmann,
M., Petrat, K., Putz, V., Hescheler, J., and Sauer, H. Regulation of the
multidrug resistance transporter P-glycoprotein in multicellular tumor
spheroids by hypoxia-inducible factor (HIF-1) and reactive oxygen
species. Faseb J, 17: 503-505, 2003.

68

190.

191.

192.

193.

194.

195.

196.

197.

198.

199.

200.

201.

202.

Finkel, T. Oxygen radicals and signaling. Curr Opin Cell Biol, 10: 248-253,
1998.

Palmer, H. J. and Paulson, K. E. Reactive oxygen species and
antioxidants in signal transduction and gene expression. Nutr Rev, 55:
353-361, 1997.

Sies, H. Strategies of antioxidant defense. Eur J Biochem, 215: 213-219,
1993.

Byrne, J. A., Grieve, D. J., Cave, A. C., and Shah, A. M. Oxidative stress
and heart failure. Arch Mal Coeur Vaiss, 96: 214-221, 2003.

Brand, M. D., Affourtit, C., Esteves, T. C., Green, K., Lambert, A. J., Miwa,
S., Pakay, J. L., and Parker, N. Mitochondrial superoxide: production,
biological effects, and activation of uncoupling proteins. Free Radic Biol
Med, 37: 755-767, 2004.

Turrens, J. F. Mitochondrial formation of reactive oxygen species. J
Physiol, 552: 335-344, 2003.

Madesh, M. and Hajnoczky, G. VDAC-dependent perrneabilization of the
outer mitochondrial membrane by superoxide induces rapid and massive
cytochrome c release. J Cell Biol, 155: 1003-1015, 2001.

Mello F ilho, A. C. and Meneghini, R. In vivo formation of single-strand
breaks in DNA by hydrogen peroxide is mediated by the Haber-Weiss
reaction. Biochim Biophys Acta, 781: 56-63, 1984.

lmlay, J. A., Chin, S. M., and Linn, S. Toxic DNA damage by hydrogen
peroxide through the Fenton reaction in vivo and in vitro. Science, 240:
640-642, 1988.

Zhang, L., Yu, L., and Yu, C. A. Generation of superoxide anion by
succinate-cytochrome c reductase from bovine heart mitochondria. J Biol
Chem, 273: 33972-33976, 1998.

Starkov, A. A., Fiskum, G., Chinopoulos, C., Lorenzo, B. J., Browne, S. E.,
Patel, M. S., and Beal, M. F. Mitochondrial alpha-ketoglutarate
dehydrogenase complex generates reactive oxygen species. J Neurosci,
24: 7779-7788, 2004.

Cadenas, E. and Davies, K. J. Mitochondrial free radical generation,
oxidative stress, and aging. Free Radic Biol Med, 29: 222-230, 2000.

Touati, D. Iron and oxidative stress in bacteria. Arch Biochem Biophys,
373: 1-6, 2000.

69

203.

204.

205.

206.

207.

208.

209.

210.

211.

212.

213.

214.

Leonard, S., Gannett, P. M., Rojanasakul, Y., Schwegler-Berry, D.,
Castranova, V., Vallyathan, V., and Shi, X. Cobalt-mediated generation of
reactive oxygen species and its possible mechanism. J Inorg Biochem, 70:
239-244, 1998.

Salnikow, K., Blagosklonny, M. V., Ryan, H., Johnson, R., and Costa, M.
Carcinogenic nickel induces genes involved with hypoxic stress. Cancer
Research, 60:38-41, 2000.

Shi, X., Dalal, N. S., and Kasprzak, K. S. Generation of free radicals from
model lipid hydroperoxides and H202 by Co(ll) in the presence of
cysteinyl and histidyl chelators. Chem Res Toxicol, 6: 277-283, 1993.

O'Brien, T., Xu, J., and Patierno, S. R. Effects of glutathione on chromium-
induced DNA crosslinking and DNA polymerase arrest. Mol Cell Biochem,
222: 173-182, 2001 .

Kasprzak, K. S. Oxidative DNA and protein damage in metal-induced
toxicity and carcinogenesis. Free Radic Biol Med, 32: 958-967, 2002.

Chin, T. A. and Templeton, D. M. Protective elevations of glutathione and
metallothionein in cadmium-exposed mesangial cells. Toxicology, 77: 145-
156, 1993.

Hussain, T., Shukla, G. S., and Chandra, S. V. Effects of cadmium on
superoxide dismutase and lipid peroxidation in liver and kidney of growing
rats: in vivo and in vitro studies. Pharrnacol Toxicol, 60: 355-358, 1987.

Filipic, M. and Hei, T. K. Mutagenicity of cadmium in mammalian cells:
implication of oxidative DNA damage. Mutat Res, 546: 81-91, 2004.

Lesnefsky, E. J. and Hoppel, C. L. lschemia-reperfusion injury in the aged
heart: role of mitochondria. Arch Biochem Biophys, 420: 287-297, 2003.

Lesnefsky, E. J., Gallo, D. S., Ye, J., Whittingham, T. S., and Lust, W. D.
Aging increases ischemia-reperfusion injury in the isolated, buffer-
perfused heart. J Lab Clin Med, 124: 843-851, 1994.

Vanden Hoek, T. L., Becker, L. B., Shao, 2., Li, C., and Schumacker, P. T.
Reactive oxygen species released from mitochondria during brief hypoxia

induce preconditioning in cardiomyocytes. Journal of Biological Chemistry,
273: 18092-18098, 1998.

Nelson, D. R., Zeldin, D. C., Hoffman, S. M., Maltais, L. J., Wain, H. M.,
and Nebert, D. W. Comparison of cytochrome P450 (CYP) genes from the
mouse and human genomes, including nomenclature recommendations
for genes, pseudogenes and alternative-splice variants.
Pharmacogenetics, 14: 1-18, 2004.

70

215.

216.

217.

218.

219.

220.

221.

Jezek, P. and Hlavata, L. Mitochondria in homeostasis of reactive oxygen
species in cell, tissues, and organism. Int J Biochem Cell Biol, 37: 2478-
2503, 2005.

Schrader, M. and Fahimi, H. D. Mammalian peroxisomes and reactive
oxygen species. Histochem Cell Biol, 122: 383-393, 2004.

Cappola, T. P., Kass, D. A., Nelson, G. S., Berger, R. D., Rosas, G. O.,
Kobeissi, Z. A., Marban, E., and Hare, J. M. Allopurinol improves
myocardial efficiency in patients with idiopathic dilated cardiomyopathy.
Circulation, 104: 2407-2411, 2001.

Babior, B. M. The NADPH oxidase of endothelial cells. lUBMB Life, 50:
267-269, 2000.

Yu, B. P. Cellular defenses against damage from reactive oxygen species.
Physiol Rev, 74: 139-162, 1994.

Ames, B. N., Gold, L. S., and Willett, W. C. The causes and prevention of
cancer. Proc Natl Acad Sci U S A, 92: 5258-5265, 1995.

Beckman, K. B. a. A., B.N. Oxidants, Antioxidants and Aging. In: J. G.
Scandalios (ed.), Oxidative Stress and Molecular Biology of Antioxidant
Dfenses, Vol. 1, pp. 46. New York: Cod Spring Harbor Laboratory Press,
1997.

71

CHAPTER 1

Vengellur, Ajith, Phillips, Jennifer M., Hogenesch, John B., Lapres, John J.
(2005). Gene Expression Profiling of Hypoxia Signaling in Human Hepatocellular
Carcinoma Cells. thsiologicalfienomics. 22: 308-318.

72

Abstract

Cellular, local, and organismal responses to low oxygen availability occur during
processes such as anaerobic metabolism, wound healing, and pathological
conditions such as stroke and cancer. These responses include increases in
glycolytic activity, vascularization, breathing and red blood cell production, and
they are mediated in part by the hypoxia inducible factors (HlFs), which receive
information on 02 levels from a group of iron and oxygen-dependent
hydroxylases. Hypoxia mimics, such as cobalt chloride, nickel chloride, and
deferoxamine, act to simulate hypoxia by altering the iron status of these
hydroxylases. To determine if these mimics are appropriate substitutes for lower
oxygen tension evoked naturally, we compared transcriptional responses of a
Hep3B cell line using high-density oligonucleotide arrays. A battery of core
genes was identified that was shared by all four treatments (hypoxia, cobalt,
nickel and deferoxamine) including glycolytic enzymes, cell cycle regulators and
apoptotic genes. lmportantly, cobalt, nickel and deferoxamine influenced
transcription of distinct sets of genes that were not affected by cellular hypoxia.
These global responses to hypoxia indicate a balancing act between adaptation
and programmed cell death, and suggest caution in the use of hypoxia mimics as

a substitute for low 02 tension that occurs in vivo.

73

Introduction

Cells, tissues, and organisms are said to be hypoxic when they receive less than
normal levels of oxygen. Given the central role of oxygen in the production of
adenosine triphosphate (ATP) through oxidative phosphorylation, it is critical for
cells and tissue to respond rapidly to hypoxia. The importance of hypoxia
signaling is further highlighted by its essential role in mammalian development,
and several pathological conditions such as cardiovascular disease and cancer
(1). This response is regulated by a family of transcription factors called the
hypoxia inducible factors (HlFs). HIFs are members of the bHLH-PAS (basic-
helix-loop-helix-EER, ARNT, _S_|M) family of transcription factors (2). For a
detailed description of the pathway and its regulation see the “introduction”
chapter. Under norrnoxic conditions, HlF1or is ubiquitously transcribed,
translated, and subsequently degraded. Under hypoxic conditions, however,
HlF1a protein becomes stabilized (3-5). This oxygen-dependent degradation of
the HlF1a protein is primarily controlled by a family of non-heme oxygenases
called prolyl hydroxylase domain containing proteins (PHDs, also known as HlF
prolyl hydroxylase) (6, 7). The iron-dependent activity of the PHD enzymes may
help explain the ability of iron chelators and divalent metals to promote a
hypoxic-like response. In fact, cobalt chloride and nickel chloride, transition
metals capable of competing with iron at binding sites; and deferoxamine (DFO),
an iron chelator, are widely used hypoxia mimics. The appropriateness of these
hypoxia mimics has not been thoroughly tested. These current studies were

performed to characterize the battery of hypoxia regulated genes and to test the

74

hypothesis that cobalt chloride, nickel chloride, and DFO were capable of
modulating a similar battery of genes and act as true hypoxia mimics. To
address these goals, the hepatocellular carcinoma cell line, Hep3B, was exposed
to hypoxia, cobalt chloride, nickel chloride or DFO, total RNA extracted, and
transcriptional responses were evaluated using high density oligonucleotide
arrays. Comparisons among the four treatments suggest that there is substantial
overlap; however, there is also a large set of genes that are specific for the

individual treatments.

Materials and Methods

Cell culture: Hep3B cells were maintained in MEM (lnvitrogen) supplemented
with 10% fetal bovine serum (Hyclone, Logan, UT), 20 mM L-glutamine, 1 mM
MEM non-essential amino acids, 100 mM Hepes (pH=7.4), 1,000 units/ml
penicillin G and 1,000 pg/ml streptomycin sulfate (lnvitrogen). They were
approximately 70% confluent at the time of treatment. Cells were maintained at
37°C, 5% CO; and 21% 02 prior to treatment. Hypoxia treatment (1% 02) was
performed in oxygen-regulated incubator (Precision-NAPCO 7000, Winchester,
VA) at 37°C and 5% 002. Cobalt chloride (100 pM), nickel chloride (100 pM)
and deferoxamine (100 pM) treatments were performed at 37°C, 5% C02 and

21% 02.

75

RNA extraction: RNA was extracted by homogenization (Polytron, Kinematica,
Lucerne, Switzerland) in TRlzol reagent (GlBCO/BRL) (added to cell pellet) at
maximum speed for 90-120 sec. The homogenate was allowed to incubate for
five minutes at room temperature, a 1/5 volume of chloroform was added, the
tube was vortexed, and finally subjected to centrifugation at 12,000 x g for 15
minutes. The aqueous phase was isolated, and a half volume of isopropanol
was added to precipitate the RNA. Following this initial isolation, a secondary
purification was performed with the Qiagen RNAeasy Total RNA isolation kit
according to manufacturer’s specifications. The purified total RNA was finally
eluted in 10 p.L of DEPC treated H20, and quantity and integrity were
characterized using a Beckman DU640 UV spectrophotometer and Agilent

Bioanalyzer 2100.

RNA labeling: Briefly, 5 pg total RNA from separate biological replicates was
used to make first strand cDNA using the Superscript Choice system (Gibco
BRL) and a T7 promotor/oligodT primer (Gibco). Second strand cDNA was also
made with the Superscript Choice system. The resulting cDNA was subjected to
phenolzchloroform purification and ammonium acetate precipitation, and used as
a template to make biotinylated amplified antisense cRNA using T7 RNA
polymerase (Enzo kit, Affymetrix). Twenty micrograms cRNA was fragmented to
a range of 20 to 100 bases in length using fragmentation buffer (200 mM Tris-

acetate, pH 8.1, 500 mM KOAc, 150 mM MgOAc) and heating for 35 minutes at

76

94 °C. The quality of cRNA and size distribution of fragmented cRNA was

examined by both agarose and polyacrylamide gel electrophoresis.

Hybridization: 20 pg of cRNA was hybridized to a U95A version 1 gene chip
(Affymetrix) in with 1 X MES hybridization buffer using standard protocols
outlined in the Gene Chip® Expression Analysis Technical Manual (Affymetrix).
Hybridization was conducted in a GeneChip Hybridization Oven for 16 hours at
45 °C. Following hybridization, the arrays were washed on a Genechip Fluidics
Station 400 according to manufacturer’s instructions (Affymetrix). The arrays
were finally scanned using a Hewlett Packard 2500A Gene Array Scanner and
the raw images were analyzed using the MAS4 Software Suite (Affymetrix).
Finally, quality of cRNA was assessed by examining 3'l5' ratios for GAPDH

oligonucleotides present on the arrays.

Data Analysis: CEL files were condensed using the GCRMA algorithm in R
(www.r-proiect.org) (8-10). Pair-wise comparisons between control and
treatment conditions were conducted using a Students T-Test, two sample, equal
variance on logZ transformed data. Fold change calculations were performed in
Excel on data that was median-scaled to a global intensity target value of 100.
For each treatment vs. control condition, genes that changed were assigned

based on a t-test value of <0.05 and a fold change value of > 2.

77

Relative Real-Time PCR analysis: Changes in gene expression observed by
microarray analyses were verified by real-time PCR, performed on an Applied
Biosystems Prism 7000 Sequence detection System (Foster City, CA) as
described (11). Briefly, cDNA was synthesized from total RNA (1 pg per sample
per treatment, n=6) in a reverse transcriptase reaction in 20 pl of 1X First Strand
Synthesis buffer (lnvitrogen, Carlsbad, CA) containing 1 pg oligo (5’-T21VN-3’),
0.2 mM dNTPs, 10 mM DTT, and 200 IU of Superscript II reverse transcriptase
(lnvitrogen). The reaction mixture was incubated at 42 °C for 60 min, and
stopped by incubation at 75 °C for 15 min. Amplification of cDNA (1/20) was
performed using SYBR Green PCR buffer (1x AmpliTaqTM Gold PCR Buffer,
0.025 U/pl AmpliTaqTM Gold (Perkin-Elmer, Wellesley, MA), 0.2 mM dNTPs, 1
ng/pl 6-carboxy-X-rhodamine, 1:40,000 diluted SYBR® Green Dye and 3%
DMSO) and 0.1 pM primers. The thermal cycling parameters were 95°C for 10
min followed by 40 cycles of 95°C for 15 seconds and 60°C for 60 seconds. The
mRNA expression for each gene was determined by comparing it with a standard
curve of known quantities of the specific target. This measurement was controlled
for RNA quality, quantity, and RT efficiency by normalizing it to the expression
level of the hypoxanthine guanine phosphoribosyl transferase (HPRT) gene.
HPRT was used as a control gene because it was shown to be unaffected by any
treatment used. Each primer set produced a single product as determined by
melt-curve analysis and amplicons were of correct size as analyzed by agarose
gel electrophoresis. Statistical significance was determined using normalized

fold changes and student t-test.

78

Primers were designed using the web based application Primer3 (http://www-
genome.wi.mit.edu/cgi-bin/primer/primer3_www.cgi) biasing towards the 3’ end
of the transcript to maximize the likelihood of giving a gene specific product. The
settings used in Primer3 were 125 bp amplicon, 20mers, 60°C melting
temperatures and all other as defaults. Primer sequences were analyzed by
BLAST. Gene names, accession numbers, and forward and reverse primer

sequences are listed in Table 2.

Results

Gene expression measurements were generated following 24 hour exposure to
each of the five treatments, normoxia, hypoxia, cobalt, nickel and deferoxamine,
in Hep3b cells. To examine the global relatedness of each of these stimuli, we
performed hierarchal clustering on the entire data set. These results indicated
that there is considerable similarity among the four stimuli (data not shown). To
compare directly the effects of these treatments on the complete data set, an
analysis of variance (ANOVA) was performed. The results showed that greater
than 38% (3,404/12,626 probe sets) were significantly (p < 0.05) influenced by
treatment. These results suggest that hypoxia and hypoxia mimics have
profound effects on cellular homeostasis and that there is considerable similarity

among the four hypoxia treatments.

79

E Homoo “3m t
at w you w 8 m
out omom 2mm 66
u 2 mmmmooofimo m

022022808 9:
0 you mom “mum «o o
$08.03 .8886
gamma
“a «m 0mm twaoom
£823
[£33

  

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mozootouomommoouo mmvooo .22 2395
o «.8 £2883 8. 888 .22 100%?
88 .om «ommmom nomvoo .22 2.01.2
Em 08 z m «Scum 9.38 .22 <O<.2
m we ooommomom mm 85.8 .22 Epic.
«8880832228 $2.00 .22 ..<H.O...
02 «880% m8 3 023.2 m...
«802090 8 Stem @550 .22 aims: .252. 5me
to «am 2 new 0: 3800 .22 5.0a
. .88. .288 9° mmomoo .22 28.5
omowmoom 8mm... n...E:...
moommm zmoomomouoo. oKvmom< H.mum.>
«mom mommoa meow 3300 .22

 

 

 

 

 

swamp

 

 

  

c2833 .

 

  

 

monfima c. 89 205E .N asap

80

A direct comparison of each treatment was also performed to identify subsets of
genes that were shared, and those that were unique, among treatment groups.
Genes that showed significant changes in expression (p < 0.05 and greater than
2 fold change) were compared. At these significance levels, hypoxia (1% Oz)
influenced 451 different probe sets. These sets of responsive genes included
well known hypoxia-regulated genes such as glyceraldehydes-3-phosphate
dehydrogenase (GAPDH), vascular endothelial growth factor (VEGF), insulin
growth factor II (lGFll), carbonic anhydrase, and plasminogen activator inhibitor l
(PAl I) (12), (13) (Table 3). This list of hypoxia responsive genes also included
several genes that had not previously been demonstrated to be modulated by
low oxygen tension. For example, pyruvate dehydrogenase kinase 1 (PDK1) and
the creatine transporter, SLC6A8 were upregulated 4.7 and 5.7 fold respectively.
In addition, the NAD*-dependent 15 hydroxyprostaglandin dehydrogenase and
prohibitin genes were significantly down-regulated following hypoxia exposure.
These results confirm the appropriateness of our data analysis and suggest that

a more complete set of hypoxia responsive genes were identified.

The other three treatments, cobalt, DFO and nickel, also had profound effects on
the cells. Cobalt exposure (100 pM, 24 hours) led to the significant change (p<
0.05 and greater than 2 fold) in transcription of 303 probe sets (Figure 5). Of
these 303 probes sets, 85 (28%) overlapped with the hypoxia responsive subset.
Nickel exposure altered the expression of a similar number of genes as cobalt

(325 probe sets), however, a higher percentage overlapped with the hypoxia

81

Table 3. Top 20 hypoxia up and down-regulated probe sets

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

FOLD CHANGE
Probe Set HYP COB DFO NICK Descriptions
39352_at -13.1 -1.9 2.5 —6.7 thyroid-stimulating hormone alpha subunit
1025 g at -10.2 -5.9 -14.7 -8.2 HSCYP450 Human gene for cytochrome P(1)-450
32570_at -6.9 -5.8 -7.3 -5.4 15 hydroxyprostgggndin dehydrogenase (PGDH)
742_at -5.0 -9.6 -4.4 -11.4 Human mRNA for HGF activator like protein
40588_r_ at -4.7 1.0 -4.8 -2.9 p18 protein
41363_at -4.7 -1.7 -6.4 -2.6 survival of motor neuron protein interactiLg protein 1
39382_at -4.5 -1.5 -3.4 -2.2 mRNA for KIAA051Lgb=ABO11089
36592_at -4.5 -1.5 -8.3 -2.6 prohibitin
38680_at -4.4 -1.4 -1.6 -2.0 Alu repeats, region 5 to the SNRP E gene
39355_at -4.3 -1.4 -6.5 -2.6 2-5A binding protein
31880!at -4.1 -1.4 -6.6 -3.5 N9 Rep-ng=083767
36636_at -4.0 1.0 1.1 -1.9 ornithine aminotransferase
41772_at -4.0 -2.2 -4.7 -2.9 monoamine oxidase A
38323_at -4.0 1.1 -7.1 -2.8 BAC clone RG113D17jb=A0005162
33791_at -3.9 1.9 -2.6 -2.3 leukemia associated gene 1
38372_at -3.9 -1.1 -1.7 -2.6 U66042:Human clone 191B7
40788_at -3.9 1.0 -1.3 -2.6 adenylate kinase 2A (AK2A)
1969_s3t -3.8 -1.0 -1.8 -2.9 CDK activating kinase
41246_at -3.7 -1.4 1.8 -3.9 Homo sapiens $=Al743134
37297_at -3.7 -1.4 -3.7 -2.4 Homo sapiens DKFZp586A191
37758_s_at 4.6 -1.1 4.8 4.5 Homo sapiens gb=W28479
39827_at 4.7 4.4 4.7 6.8 Homo sapiens gb=AA522530
36386_at 4.7 10.8 14.9 3.9 pyruvate dehydrogenase kinase isoenzyme 1
40790_at 4.9 4.0 8.0 4.2 DECt
36101_s_at 5.0 3.6 7.2 4.1 vascular endothelial growth factor
38545_at 5.2 2.5 12.1 4.0 testicular inhibin beta-B-subunit
40926__at 5.8 12.1 25.5 6.7 creatine transporter (SLC6A8)
38125_at 6.3 10.6 22.8 7.9 plasminogen activator inhibitorl
HUMGAPDH 6.5 -1.0 2.9 2.7 glyceraldehyde-3-phosphate dehydrogenase
40888 _lj_ at 6.6 -1.3 3.6 3.3 Homo sapiens lb=W28170
34777_at 6.8 33.1 39.4 3.9 adrenomedullin precursor
32588_s_at 7.0 -2.5 1.3 8.5 ERF-2
40309_at 8.1 39.4 39.9 5.7 Carbonic anhydrases (MaTu MN)
31918_at 8.4 1.2 9.0 5.5 homeodomain protein (Prox 1)
34610_at 8.6 -1.0 1.6 3.1 Homo sapiens gb=W25845
1591!s_.1t 11.9 2.2 8.6 8.5 insulin-like growth factor II
36782_s_at 13.9 2.3 10.3 13.2 insulin-like growth factor II
672_at 14.2 7.4 13.5 3.3 plasminogen activator inhibitor-1
36933_at 79.9 127.7 241.0 61.5 RTP (N-myc downstream mated gene 1)
2079_s_at 255.6 27.6 140.0 202.7 insulin-like growth factor (lGF-ll)

 

82

 

Figure 5 Venn diagrams representing genes whose expression was
significantly influenced by treatment.

Genes whose expression was influenced by the four hypoxia treatments are
displayed as Venn diagrams. Total number of genes for each treatment group is
listed at the bottom next to the treatment label. Each section that shares the
same color and number represents the same subset of genes. (a) a complete list
of these genes can be found in Table 4. (b) a partial list of these genes can be
found in table 5. (c) a partial list of these genes can be found in table 6. (d) a
partial list of these genes can be found in table 7, (e) a partial list of these genes
can be found in table 8.

 

83

subset (65.2%, 212/325 probe sets) (Figure 5). DFO treatment led to altered
expression of a much larger number of probe sets under these conditions (100
pM, 24 hours). DFO exposure changed the expression of 1068 different probe
sets and 23.3% (249/1068 probe sets) were shared with the hypoxia subset
(Figure 5). These data suggest that there is a core set of genes that are
responsive to hypoxia and hypoxia mimics but a larger population may be unique

to each treatment.

To understand more thoroughly this core set of genes, a comprehensive analysis
of those genes that were shared among all four treatment groups was performed.
This analysis led to the identification of a core set of 62 probe sets, representing
55 different genes (Table 4). This list contains a large group of previously
characterized hypoxia responsive genes, including IGF ll, n-myc downstream
regulated (also known as NDR1, RTP, AND CAP43), PAI 1, VEGF and several
glycolytic enzyme genes. The list also contained several genes that had not
been characterized as hypoxia responsive. These include pyruvate

dehydrogenase kinase 1, PDGH, MAOA, inhibin and SLC6A8.

To verify the expression of some of the novel hypoxia regulated genes, as well
as some of the conventional hypoxia target, qRT-PCR was performed. The
genes verified included VEGF and carbonic anhydrase as controls for the

hypoxia treatments. These genes are known hypoxia genes and qRT-PCR and

84

  

3:95.35 .58 ..n 5933

05

flow

  

no on... .v 225...

85

  

$2953.: .58 =m Coogan-

2:

3cm

  

um 9:. .v 033.

86

microarray data were in good agreement. In addition, four other genes (PDK1,
SLC6A8, inhibin and MCT3) that were upregulated in all four treatments and had
not been characterized as hypoxia regulated were also verified. Each of the four
treatments led to the upregulation of the PDK1 and SLC6A8 genes in the
microarray and qRT-PCR analysis (Figure 6A). In addition, the qRT-PCR
expression pattern of inhibin correlated with the microarray data in three of the
four treatments, while MCT3 only verified in the DFO treatment group (Figure

6A).

There were also several down-regulated genes that were verified by qRT-PCR
(Figure 6B). Cyp1A1 was down-regulated in all four treatments in both the
microarray and q RT-PCR experiments and this is in good agreement with
previously published reports (14, 15). lL8 was upregulated in three of the four
treatments on the microarray and this pattern was verified in the qRT—PCR
results. Finally, four genes (HGFAL, IP30, MAOA and 15-PGDH) were down-
regulated in all four treatments on the microarray and this was confirmed on the

qRT_PCR, with one exception. Nickel induced a slight upregulation in the 15-

PGDH gene in the qRT-PCR (Figure 6B).

A direct comparison of the microarray and qRT-PCR data was performed by
compiling all of the values for all twelve genes under all four conditions and
plotting them on a Log2 chart (Figure 7). The graph was further analyzed by

linear regression and displayed a slope of 0.94 and an R2 value of 0.58. These

87

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

s
A' VEGF 5° Carb.Annh(MaTu)
5 4o
4 so
A ‘ 7 /-—-l
3 / 4 20
.E 2 ¢ (1 /
' 1o
=3 . 517, 2r .
.e Hyp cog DFO NICK HYP COB DFO NICK HYP COB DFO NICK
N
0 so 20
a 26 15 lnhlbln 5 MCT3
C
‘5 2° 10
5 15
U 10
E .5

 

HYP coe DFO NICK HYP coe oro NICK HYP COB DFO NICK

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Treatment

Be 3 2

. , u, my
.03 j .2 2L g.
g .9 “ g'
' .12 4’
g 45 .3
I; HYP COB DFO NICK HYP COB DFO NICK HYP COB DFO NICK
a a
5 0 new 1 it, ’ MAOA ’ 15mm:
3 / / 0 I v I 0 . . r
04412—ng ..?.élr fire
U .5 -/ fi 1 l ’2‘ / /
3 1 z / g / I;
u_ ,9 1 4 .4. / 2

-12 .o

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.5
HYP COB DFO NICK HYP COB DFO NICK flYP COB DFO NICK

Treatment

 

Figure 6. qRT-PCR verification of 12 genes.

The expression pattern for 6 upregulated genes (A) and 6 down-
regulated genes (B) was verified by qRT-PCR using sybr green as a
marker. The sybr green data (white bars) is directly compared to the
microarray data (hatched bars). Gene names are listed at the top of
each graph. HYP = 1% Oz, COB = 100 pM CoCI2, DFO = 100 pm
deferoxamine and NICK = 100 pM NiCI2

 

 

 

 

88

(:601) missmdxa Had-13b

 

 

 

 

 

 

 

 

 

 

 

 

 

6 . . ’
O //
4 e.‘ ’4
/ O
. £0 -’ g 0
2 e ,'
I'/
o . V, . .
... O/ . g
'2 ,o 8 e' " + y=0.9403x+0.3458
0 .‘o R’IO.5856
_4 4
-4 -2 O 2 4 6

GeneChip Expression (LogZ)

 

 

Figure 7. Analytical comparison of qRT-PCR and microarray
data.

Fold change values from qRT-PCR and microarray results were
L092 transformed and plotted on a linear gra h. Linear regression
was performed and equation of the line and R values are listed.

 

89

 

results validate the microarray data analysis and suggest that there is very good

correlation between the two assays.

The results suggest that there is considerable overlap between the various
treatments; however, there is also a large subset of genes that were altered in a
treatment specific manner. The expression of these genes passed significance
and fold change cut-offs in only one treatment. For example, there were 121
probe sets that showed change in expression (p < 0.05 and greater than 2 fold)
following hypoxia treatment but were not significantly changed in the other three
treatment groups (Figure 5, Table 5). This subset included probe sets for the
guanine nucleotide binding protein, c-jun, MAP kinase phosphatase 1, ubiquitin-
conjugating enzyme E2 and asparagine synthetase (Table 5). The role these

genes play in the cellular response to hypoxia is currently being investigated.

There were also 92, 676 and 49 probes sets that were specific to cobalt, DFO
and nickel, respectively (Figure 5). These subsets include growth factors

(eg. connective tissue growth factor, amphiregulin, and transforming growth
factor-beta), structural proteins (e.g. desmocollin, spectrin and adducin), and
genes for several important enzymes (e.g. glutathione-S-transferase, hepatic
dihydrodiol dehydrogenase, and s—adenosylmethionine synthetase). In addition,
each treatment altered the expression of various critical transcription factors and
cell cycle regulators. For example, cobalt exposure led to the decreased

expression of cyclin D1 (Table 6). Nickel altered the expression of the genes for

90

Table 5. Top 20 up and down regulated genes specific to hypoxia
(1% 02) exposure

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

FOLD CHANGE

Probe Set HYP COB DFO NICK Descriptions
38680_at -4.43 -1.37 -1.63 -2.05 Alu repeats in the small nuclear ribonucleoprotein E
36636_at -4.00 1.03 1.09 -1.91 ornithine aminotransferase
36671_at -3.43 -1.34 -1.52 -1.78 asparagine synthetase
40629_at -3.05 1.15 -1.61 -1.29 GPI-H mRNA
40074_at -3.05 -1.05 -1.31 -1.99 NAD—dependent methylene THF-DH cyclohydrolase
36604 !at -3.04 1.02 -1.66 -1.92 ubiquitin-conjugating enzyme E2
32777_at -3.00 1.14 -1.80 -1.90 CHDS protein
39983_at -2.90 -1.88 -1.92 -2.76 FKBP-associated protein FAP48
38029_at -2.80 -1.29 -1.07 -1.77 glycoprotein 4F2
41536_at -2.75 1.27 -1.21 -1.77 AL022726 clone 625H18
34366_g_at -2.75 1.10 -1.33 -1.76 cyclophilin-33B (CYP-33)
36857_at -2.70 -1.30 —1.69 -1.90 DNA repair exonuclease (REC1)
38983_at -2.69 -1.11 -1.85 -1.73 AI223047cDNA
33434_at -2.62 1.46 -1.87 -1.52 Bet1p homoLog (hbet1)
40250_at -2.60 -1.46 -1.97 -2.51 Rev interacting protein Rip-1
402_s_at -2.60 1.56 -1.08 -1.54 lCAM-3 mRNA
262_at -2.58 -1.10 -1.42 -1.57 S-adenosylmethionine decarboxylase
40568_at -2.56 -1.11 -1.50 -1.92 vacuolar H+-ATPase
34356_at -2.54 1.23 -1.57 -1.80 olymerase ll complex component SRB7
38725_s_at -2.50 -1.81 -1.62 -1.86 N36295 cDNA
37657_at 2.01 -1.03 1.86 1.50 PALM gene
38586_at 2.03 -1.67 1.10 1.14 fatty acid binding protein (FABP)
41657_at 2.03 1.33 1.23 1.54 serine threonine kinase 11
40186_at 2.04 1.54 1.15 1.83 MAP kinase phosphatase 4
1443_at 2.04 1.50 1.42 2.78 HSMAPKP4 MAP kinase phosphatase 4
41366_at 2.12 -1.04 -1.40 1.96 mRNA for KIAA1002
32583_at 2.13 1.51 1.81 1.70 c-jun proto oncogene (JUN)
38757_at 2.14 -1.14 1.42 1.87 PDGF associated protein
835_at 2.14 -1.89 1.71 1.87 PDGF associated protein
41061_at 2.17 -1.14 -1.15 1.61 huntingtin interactingprotein 1
34303_at 2.21 1.23 1.75 1.86 AL049949 mRNA
40960_at 2.28 -1.02 1.75 1.79 beta-1,4-galactosyltransferase
HSACO7 2.31 -1.00 1.01 1.09 beta-actin
37880_at 2.33 1.25 1.33 1.67 L-alanine-glyoxylate aminotransferase
39385_at 2.41 -1.15 1.48 1.79 aminopeptidase N/CD13
38295_at 2.47 -1.00 1.53 1.87 PBX2 mRNA
humJalugt 2.55 -1.18 1.69 1.68 Alu-Sq subfamily
37884_f_at 2.66 4.75 2.50 1.81 Al375033 cDNA
41872_at 2.69 1.32 -1.37 1.72 nonsyndromic hearing impairment protein (DFNA5)
34610_at 8.57 -1.01 1.65 3.13 guanine nucleotide bindingprotein

 

 

 

 

 

 

91

 

Table 6. Top 20 up and down regulated genes specific to CoCI2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(100 pM) exposure

FOLD CHANGE
”WEE—Sit HYP CD'BT‘UFU'TNICK Uesfilfit‘lfi'fis
41320_s_at 1.56 -3.76 -1.57 -1.07 transcrkLtional repressor (GCF2)
2017_s_at 1.90 -3.75 1.13 1.29 Human cyclin D (cyclin D1)
37637_at -1.20 -3.73 -1.52 -1.34 RGP3 mRNA
867_s_at 1.31 -3.58 1.36 1.26 thrombospondin-1
36638_at -1.27 -3.47 1.11 -1.74 connective tissue growth factor
1143_s_at -1.26 -3.41 -1.89 -1.68 Fibroblast Growth Factor Receptor K-Sam
38418_at -1.07 -3.10 -1.53 -1.23 PRADl mRNA for cyclin
33436_at -1.42 -2.95 -1.12 -1.26 SOX9 mRNA
34517_at 1.31 -2.90 -1.20 1.18 HMG-CoA—synthase
2020_at -1.31 -2.86 -1.86 -1.20 bcl-1 mRNA
33389_at -1.24 -2.75 -1.76 -1.26 cytochrome P450 (CYP51)
38686_at -1.42 -2.64 -1.26 -1.27 vacuolar proton ATPase
31521_f_at -1.64 -2.60 -2.00 -1.51 H4/egene for H4 histone
1787_at -1.08 -2.60 -1.30 -1.71 Cdk-inhibitor p57KlP2
34213_at -1.17 -2.57 1.20 -1.58 mRNA for KIAA0869 protein
404_at -1.17 -2.54 -1.92 -1.32 interleukin 4 receptor
1970_sLat -1.09 -2.48 -1.85 -1.34 FGFR2 mRNA
866_at -1.24 -2.43 -1.70 -1.23 thrombospondin-1
33369_at 1.05 -2.42 -1.95 1.44 Al535653 cDNA
40841_at -1.04 -2.37 1.12 -1.33 TACC1 (TACC1) mRNA
35284_f_at -1.63 2.83 -1.72 -1.78 W28620 cDNA
37002_at 1.33 2.91 1.12 1.32 biliverdin-leeta reductasel
38825 it 1.06 2.95 1.02 -1.06 fibrinogen alpha chain gene
1962_at 1.66 2.95 1.04 1.74 liver arginase
40838_at 1.56 3.11 1.46 1.31 mRNA for KIAA0530 protein
35214_at -1.57 3.32 -1.21 1.73 UDP-glucose dehydrogenase (UGDH)
38876!at -1.29 3.70 -1.06 1.34 AL080091 cDNA
31850_at -1.41 3.89 1.26 1.27 gamma-glutamylcysteine synthetase
35724_at -1.23 4.11 -1.46 1.50 Pirin, isolate 1
32664_at -1.16 4.62 2.07 1.56 RNase 4
35686_s_at -1.42 4.82 2.03 -1.34 MTCP1 gene
39365_i_at 1.19 4.95 1.70 1.12 protein phosphatase 1 (PPP1R5)
35194_at 1.12 5.05 1.00 1.05 glutathione peroxidase-like protein
748_s_at 1.20 5.79 3.44 1.35 Mxi1 protein
34342_s_at 1.16 6.93 1.01 1.07 osteopontin mRNA
39120_at -1.32 9.41 -1.32 -1.49 AA224832 cDNA
37399_at 1.11 9.57 1.00 1.03 mRNA for KIAA0119 gene
38379_at 1.16 11.24 1.02 1.07 NMB mRNA
32805_at -1.30 64.02 -1.95 -1.60 hepatic dihydrodiol dehydrogenase
32392_s_at 1.13 80.74 1.01 1.19 bilirubin UDP-glucuronosyltransferase isozyme 2

 

 

 

 

 

 

 

92

 

the transcription factors E2A and SL1 (Table 8). DFO had the largest unique set
of genes and it contained the transcription factors, c-fos, activating transcription
factor 3 and the farnesol receptor HRR-1 (Table 7). These results suggest that
each of these treatments has different functional consequences on cellular

homeostasis which distinguishes them from hypoxia exposure.

Discussion

Several dozen genes responded to both cellular hypoxia and its mimics. These
genes included known hypoxia regulated genes, such as lGFll, VEGF and
carbonic anhydrase, but also many previously uncharacterized genes that were
altered by all four treatments. Several genes were involved in the processes of
adaptation and cell death, indicating a delicate balance between these processes
under hypoxic stress. Adaptive genes included the previously known glycolytic
enzymes and carbonic anhydrase, but also the creatine transporter, SLC6A8,
through which creatine may function to regulate the ATP supply within the cell
(16). Another gene induced by hypoxia and its mimics, PDK1, is responsible for
inactivating the pyruvate dehydrogenase (PDH) complex and thus regulating the
amount of glucose that is ultimately converted to acetyl-CoA. The upregulation
of PDK1 and subsequent inactivation of the PDH complex may act to divert

pyruvate to other metabolic pathways to cope with the hypoxic environment.

Conversely, apoptotic or cell death-promoting genes and cell proliferation

inhibitors were also found to be regulated by all three hypoxic stimuli. These

93

Table 7. Top 20 up and down regulated genes specific to Deferoxamine

 

 

 

 

 

 

 

(100 pM) exposure

FOLD CHANGE
Probe Set HYP COB DFO NICK Descriptions
38519_at -1.40 -1.92 -27.06 -1.18 farnesol receptor HRR-1
527_at -1.89 -1.64 -10.50 -1.33 centromere protein-A (CENP-A)
1647_at -1.21 -1.81 -7.84 -1.22 RasGAP-related protein (IQGAP2)
40238_at -1.56 -1.33 -7.31 -2.03 Al674208 cDNA
1599_at -1.91 -1.62 -6.45 -1.76 protein tyrosine phosphatase (CIP2)

 

34851_at -1.56 -1.12 -6.30 -1.29 serine/threonine kinase (BTAK)
37173_at -1.96 -2.12 -6.22 -1.73 CENP-E mRNA

37649_at -1.98 -1.22 -6.03 -1.61 hydroxymethylbilane synthase gene
38099_r_at -1.30 -1.67 -5.99 -1.33 acyl-CoA synthetase 4 (ACS4)
38717_at 1.61 -1.27 -5.91 1.77 AL050159 cDNA

40508_at -2.00 1.35 -5.77 -1.53 glutathione S-transferase A4-4 (GSTA4)
33701_at -1.63 -1.83 -5.68 -1.51 phenylalanine hydroxylase @AH)
41211_at -1.29 -1.19 -5.64 -1.03 mRNA for KIAA0765 protein

34852 éat -1.43 -1.10 -5.60 -1.26 serine/threonine kinase (BTAK)
1945_at -1.75 -1.48 -5.56 -1.63 cyclin B

41352_at 1.19 1.26 -5.51 1.58 beta-galactoside alpha-2,6-sialyltransferase
37302_at -1.76 -1.86 -5.48 -1.75 mitosin mRNA

38824_at -1.86 1.49 -5.41 -1.39 Tat-interactiry protein T|P30

903_at -1.12 -1.93 -5.31 -1.54 phosphatase 2A BS6-alpha (PP2A)
37235 g at -1.78 -1.81 -5.24 -1.40 alpha-2-thiol proteinase inhibitor
39279_at 1.07 -1.03 6.44 1.02 transforming_grth factor-beta
35174_i_at 1.30 1.13 6.61 1.45 elongation factor 1 alpha-2

38790_at 1.03 1.94 6.69 1.29 p53/HEH epoxide hydrolase (EPHX)
38457 at 1.49 1.02 6.70 1.09 troponin I fast-twitch isoform mRNA

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_a 1.56 -1.28 . 1.07 Ea ufiation factor III
1890_at 1. 2 1.58 7.18 1.15 TG -beta superfamily protein
1733_at 1.16 1.01 7.66 1.08 transforming_grth factor-beta

 

39302_at 1.04 2.91 8.02 -1.10 desmocollins type 2a and 2b

35617_at 1.63 1.28 8.21 1.73 BMK1 alpha kinase

37111 g_at 1.35 1.48 8.51 1.35 fructose-6-phosphate,2-kinase

32899_s_at 1.32 1.13 9.05 1.11 orphan hormone nuclear receptor RORalpha1

 

 

 

 

 

 

1113_at -1.03 -1.39 10.03 -1.52 bone morphogenetic protein 2A
1915_s_at 1.28 1.05 10.25 1.06 cellular oncogene c-fos
287_Lat 1.02 -1.05 11.22 -1.09 activating transcription factor 3

 

40367_at 1.09 -1.23 11.28 -1.10 morphogenetic protein 2A

41038_at 1.47 3.58 11.46 -1.49 neutrophil oxidase factor (p67-phox)
39248! at -1.26 1.53 11.71 1.22 N74607cDNA

34661_at -1.03 -1.30 17.11 1.68 mRNA for KIAA0350

33127_at 1.20 1.71 21.51 1.09 lysyl oxidase-related protein (WSQ-14)
1916; s! at 1.37 1.12 29.73 1.12 cellular oncogene c-fos

 

 

 

 

 

 

 

 

 

 

 

 

 

94

Table 8. Top 20 up and down regulated genes specific to

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NiCI2 (100 pM) exposure
FOLD CHANGE
Probe Set HYP COB DFO NICK Descriptions
3489; at -2.16 -1.75 1.75 -2.92 amphiregulin (AR)
37701_at -1.97 -1.61 -1.34 -2.88 helix-loop—helix basic phosphoprotein (G088)
35803_at -2.91 -1.62 1.25 -2.61 RhoE=26 kda GTPase homolog
41662_at -4.46 1.10 1.00 -2.49 AL050272 cDNA
39351_at -1.85 -1.98 -1.29 -2.46 transmembrane protein (CD59)
36191_at -2.92 -1.32 -1.99 -2.45 mitochondrial transcription factor1
36079_at -2.63 -1.96 -1.49 -2.33 Pig3 (PIG3)
38678_at -1.62 -1.12 -1.78 -2.28 AA733050 cDNA
32088_at -1.78 1.04 -1.40 -2.28 basic-leucine zipper nuclear factor (JEM-1)
35842_at -1.75 -1.19 -1.10 -2.20 AL049265 cDNA
40269_at -2.08 -1.35 -2.00 -2.13 hPrp18 mRNA
35006_at -1.47 -1.51 -1.61 -2.11 transcription factor SL1
32102_at -1.56 -1.21 -1.04 -2.11ABO18273mRNA
39506_at -1.37 1.05 -1.04 -2.09 AA933984 cDNA
39224_at -1.24 -1.24 -1.33 -2.08 ABO11152 mRNA
1536_at -1.19 1.09 -1.34 -2.07 Cdc6-related protein (HsCDC6)
1119_at —2.07 -1.24 -1.43 -2.06 Human replication protein A
38381_at -1.49 -1.55 -1.63 -2.06 syntaxin 3
39533_at -1.81 1.14 1.14 -2.05 D87432 mRNA
339_at -1.96 -1.15 1.49 -2.05 caveolin-2
31431_at 1.71 1.19 1.24 2.06 lgG Fc receptor
1047_s_at 1.41 1.05 1.09 2.09 hepatocyte growth factor-like protein gene
41725_at 1.78 1.17 1.81 2.09 casein kinase I gamma 2
36784_at 2.40 -1.03 1.96 2.11 rowth hormone and chorionic somatomammotropin
38406_f_at 1.96 1.12 1.19 2.12 Al207842 cDNA
39358_at 1.61 1.28 1.54 2.12 silencing mediator of retinoid and thyroid hormone
35154_at 1.63 1.57 1.55 2.13 W68046 cDNA
33630_s_at 1.34 1.01 -1.09 2.13 beta Ill spectrin (SPTBN2)
1218_at 1.76 1.24 1.42 2.13 v-erbA related ear-2
32146_s_at 2.01 1.58 2.24 2.18 alpha adducin
1374 g_at 1.62 1.82 1.29 2.22 transcription factor (E2A)
39372_at 1.88 -1.50 -1.14 2.28 W26480 cDNA
38789! at 1.48 1.80 1.13 2.29 transketolase
40850_at 1.96 1.67 1.68 2.29 FK-506 bindig protein homologue (FKBP38)
446_at 1.85 -1.09 1.93 2.37 casein kinase I gamma 2
32800_at 1.88 1.04 1.29 2.53 retinoid X receptor alpha mRNA
1373_at 1.84 1.30 1.70 2.57 transcription factor (E2A)
39560_at 1.55 1.64 1.00 2.73 H10776 cDNA
3257L at 1.10 -1.08 -1.92 2.73 S-adenosylmethionine synthetase
35547_at 5.15 1.19 3.27 6.25 monocarboxylate transporter 2 (hMCT2)

 

95

 

include the known hypoxic targets BCL2/adenovirus E1B 19kDa-interacting
protein 3a (NIX), that can promote apoptosis or necrosis, depending on cell type
and circumstances and DEC1 (deleted in esophageal cancers 1) an anti-
proliferation and putative tumor suppressor gene (17-20). This list also included
the RTP801 (also known as DDIT4 and N-myc downstream regulated gene 1).
RTP801 (up regulated 80 fold by hypoxia) was previously described as a
hypoxia-inducible gene that plays a complex role in cellular viability (21). Under
certain conditions, it can act in a protective manner, however, under most
standard conditions, overexpression of RTP801 leads to cell death. Collectively,
these results indicate that hypoxia and its mimics induce a complex interplay
between adaptive and pro-apoptotic responses that ultimately dictate cell

survival.

Several genes that were identified and verified suggest a complex cellular
response to hypoxia. There is an initial upregulation of the glycolytic enzymes as
a way of coping with an inability to produce ATP through oxidative
phosphorylation. This upregulation may also be important to the maintenance of
critical cellular metabolites (22). This adaptive response is supported by other
cellular processes, such as creatine transport (SLC6A8) that may allow the cell to
survive in the hypoxic environment. The decrease in Cyp1A1 may be an attempt
to control oxygen usage in side reactions or partial limits due to aryl hydrocarbon
nuclear translocator (ARNT) availability. HGFAL (also known as GILT) is a

serine protease involved in cellular adhesion and down-regulated in all four

96

treatments. A putative tumor suppressor gene, inhibin B was also down-
regulated. lnhibin B is capable of inhibiting follicle-stimulating hormone when
bound to inhibin (1. Finally, IP30 (also known as gamma-interferon-inducible
protein (GILT)) is a lysosomal thiol reductase involved in disulfide bond reduction
at low pH and is down-regulated under all four conditions (23). The role these
proteins play in the adaptive or cell death response to hypoxia is currently

unknown.

Current genomic technology has begun to unravel the various signaling networks
that are influenced, both directly and indirectly, by hypoxia. Recent publications
have used HlF1a -/- cells to identify hypoxia-inducible genes using hypoxia and
hypoxia mimics such as cobalt and nickel chloride (11, 24). A direct comparison
between these genomic screens and those presented in this manuscript is
difficult due to limited published information and the differences in the gene
expression platforms. However, several points can be made: hypoxia and
hypoxia mimics (i.e. nickel, cobalt and deferoxamine) alter the expression of
glycolytic enzymes, apoptotic genes and several hydroxylases. The
hydroxylases include EGLN1 (also known as PHD2, HPH2), an enzyme
responsible for regulating HlF1a stability. In addition, the transcription factor Jun
B is represented by two different probe sets on the U95 chip (32786_AT and
2049_S_AT), and both were significantly influenced by hypoxia and DFO
treatment but were unaffected by exposure to either metal (data not shown). In

Salnikow et al., Jun B was influenced in a HIF1a-independent mechanism (24).

97

In contrast to Salnikow et al., we could detect no significant change for ATM or
focal adhesion kinase, perhaps because these genes were not detectably
expressed in Hep3B cells, not induced by the mimics employed in this

manuscript, or due to platform issues.

A direct comparison to the cDNA microarray results from our previously
published report and the current genomic screen is difficult due to platform,
species and cell type differences; however, several interesting points can be
made. First, when a direct comparison of Table 4 of this manuscript and table 3
(active genes from wild type fibroblasts under hypoxia and cobalt) of the previous
report was performed there were 4 genes that were common to both tables.
These include lactate dehydrogenase, prolyl 4 hydroxylase, aldolase and NIX.
The limited overlap is primarily due to platform and cell type differences.
Second, when a comparison of the 24 hour treatment times between the two
experiments is made, approximately 18% of the clones regulated by hypoxia in
the WT mouse embryonic fibroblasts are also influenced in the HepBB cells (51
of 287 clones). Third, most of these clones were known hypoxia target genes,
including the glycolytic enzymes and the apoptotic gene E1B 19K/Bcl-2-binding
protein Nip3 (BNIP3) (11, 25). Fourth, the list also included some novel hypoxia
targets, such as 17-beta-hydroxysteroid dehydrogenase. The 11-beta
hydroxysteroid dehydrogenase gene has previously been shown to be a target of

hypoxia mediated down-regulation and the addition of this family member may

98

suggest a global down regulation of steroid synthesis under hypoxic conditions

(26).

The genomic screen most similar to the one reported here was performed in
HepG2 cells and utilized the same platform (27). Of the 62 probe sets listed in
Table 4, 22 could be found in the Sonna et al. report. (27). Interestingly, only one
of these was a down-regulated gene (IP30, 925_at). When the analysis is
extended to the complete list of genes influenced by hypoxia at the 24 hour time
point (data not shown) and the Sonna et al. there was approximately 10%
overlap (47/452). These shared genes fell within the 95 confidence interval
reported in the Sonna et al. tables. Again, this overlap was weighted towards
upregulated genes (36 of 47 probe sets) even though a higher proportion was
down-regulated (314 of 452). There were some known hypoxia- regulated genes
found in the 24 hour chart not included in the Sonna et al. For example, the
hydroxylase genes, PLOD2 and P4H, were not mentioned by Sonna et al. but
were significantly altered in all four hypoxia treatments in this report (28, 29).
There were also several known hypoxia genes not included in Table 4 that were
found in the Sonna et al., including GAPDH and the glucose transporters.
Though each of these was influenced in one or more treatments, they did not
pass the significance and fold change cut-off in all four treatments. These results
suggest that there is a core set of hypoxia inducible genes in all cell types and
that each also harbors the ability to mount specific responses to low oxygen

availability.

99

In a recent publication, serial analysis of gene expression (SAGE) was used to
identify genes that showed changes in gene expression by the loss of the Von
Hippel Lindau (VHL protein) (30). VHL plays an essential role in regulating
hypoxia signaling by recognizing the hydroxylated form of the HIF and recruiting
the ubiquitination machinery necessary for its degradation. Therefore, we
hypothesized significant overlap between our results and those derived from their
SAGE experiments (30). Indeed, there was extensive overlap, as 30 of the
probe sets in table 4 were found in at least one supplemental data table from the
VHL paper (30). These genes include plasminogen activator inhibitor, BNIP3
and IP30. As expected, these results suggest that there is considerable overlap
between the loss of VHL and changes in hypoxia signaling. There were also
considerable differences, possibly stemming from the different techniques, cell

types and/or methodology, or underlying biology.

Adaptive responses to hypoxia involve a complex network of signaling pathways
and are necessary for cell and organismal survival. Current study of hypoxia
relies heavily on mimics such as cobalt chloride and DFO, which presumably
function by inhibiting the iron-dependent hydroxylases that regulate HlF1or
stability. However, the possibility that cobalt and DFO have other activities that
may complicate analysis of these experiments has not been fully addressed. Our
results indicate that the overlap between these mimics and hypoxia is limited,

and suggests caution in their use. In addition, we show that many genes are

100

induced by all three stimuli including glycolytic enzymes and other survival genes
(i.e. SLC6A8), hydroxylases and apoptotic genes. A detailed mechanistic
understanding of how cells respond to low oxygen, cobalt, and desferoxamine,
including their transcriptional programmes, will ultimately be required for a more
complete understanding of hypoxia- mediated signaling and how it relates to

critical processes of development and cancer.

Acknowledgements

We acknowledge Jennifer Villasenor and Mimmi Hayakawa for technical
assistance, John Walker for help with data processing and analysis, Dr.
Christopher Bradfield for his support during the initial phases of these

experiments.

101

References

10.

Bunn, H. F. and Poyton, R. O. Oxygen sensing and molecular adaptation
to hypoxia. Phys. Rev., 76: 839-885, 1996.

Gu, Y.-Z., Hogenesch, J., and Bradfield, C. The PAS superfamily:
Sensors of environmental and developmental signals. In: Annu. Rev.
Pharrnacol. Toxicol,, Vol. 40, pp. 519-561: Academic Press, 2000.

Iyer, N. V., Kotch, L. E., Agani, F., Leung, S. W., Laughner, E., Wenger, R.
H., Gassmann, M., Gearhart, J. D., Lawler, A. M., Yu, A. Y., and
Semenza, G. L. Cellular and developmental control of 02 homeostasis by
hypoxia-inducible factor 1 alpha. Genes Dev, 12: 149-162, 1998.

Blanchard, K. L., Acquaviva, A. M., Galson, D. L., and Bunn, H. F. Hypoxic
induction of the human erythropoietin gene: cooperation between the
promoter and enhancer, each of which contains steroid receptor response
elements. Mol Cell Biol, 12: 5373-5385, 1992.

Semenza, G. L., Nejfelt, M. K., Chi, S. M., and Antonarakis, S. E. Hypoxia-
inducible nuclear factors bind to an enhancer element located 3' to the
human erythropoietin gene. Proc. Natl. Acad. Sci. USA, 88: 5680-5684,
1991.

Bruick, R. K. and McKnight, S. L. A conserved family of prolyl-4-
hydroxylases that modify HIF. Science, 294: 1337-1340., 2001 .

Epstein, A. C., Gleadle, J. M., McNeiIl, L. A., Hewitson, K. S., O'Rourke,
J., Mole, D. R., Mukherji, M., Metzen, E., Wilson, M. l., Dhanda, A., Tian,
Y. M., Masson, N., Hamilton, D. L., Jaakkola, P., Barstead, R., Hodgkin,
J., Maxwell, P. H., Pugh, C. W., Schofield, C. J., and Ratcliffe, P. J. C.
elegans EGL-9 and mammalian homologs define a family of dioxygenases
that regulate HIF by prolyl hydroxylation. Cell, 107: 43-54., 2001.

Wu, Z., lrizarry, R. A., Gentlemen, R., Murillo, F. M., and Spencer, F. A
Model Based Background Adjustment for Oligonucleotide Expression
Arrays.
hflp:/Mw.bioinfomafi;a.unito.it/downloads/complexitv/gcpaper.pdf 1-29,
2004.

lhaka, R. and Gentlemen, R. R: A Language for Data Analysis and
Graphics. J. Comp. Graph. Stats, 5: 299-314, 1996.

Gentleman, R. C., Carey, V. J., Bates, D. M., Bolstad, B., Dettling, M.,

Dudoit, 8., Ellis, 8., Gautier, L., Ge, Y., Gentry, J., Hornik, K., Hothorn, T.,
Huber, W., Iacus, S., lrizarry, R., Leisch, F., Li, C., Maechler, M., Rossini,

102

11.

12.

13.

14.

15.

16.

‘17.

18.

19.

20.

A. J., Sawitzki, G., Smith, C., Smyth, G., Tierney, L., Yang, J. Y., and
Zhang, J. Bioconductor: open software development for computational
biology and bioinformatics. Genome Biol, 5: R80, 2004.

Vengellur, A., Woods, B. G., Ryan, H. E., Johnson, R. S., and LaPres, J.
J. Gene expression profiling of the hypoxia signaling pathway in hypoxia
inducible factor 1 null mouse embryonic fibroblasts. Gene Expression, 11 .'
181-197, 2003.

Semenza, G. L., Roth, P. H., Fang, H. M., and Wang, G. L. Transcriptional
regulation of genes encoding glycolytic enzymes by hypoxia-inducible
factor 1. J. Biol. Chem., 269: 23757-23763, 1994.

Levy, A. P., Levy, N. S., Wegner, S., and Goldberg, M. A. Transcriptional
regulation of the rat vascular endothelial growth factor gene by hypoxia. J
Biol Chem, 270: 13333-13340, 1995.

Fradette, C., Bleau, A. M., Pichette, V., Chauret, N., and Du Souich, P.
Hypoxia-induced down-regulation of CYP1A1/1 A2 and up-regulation of
CYP3A6 involves serum mediators. Brit. J. Pharm., 137: 881-891, 2002.

Fradette, C. and Du Souich, P. Effect of hypoxia on cytochrome P450
activity and expression. Curr. Drug Metab., 5: 257-271, 2004.

Salomons, G. S., van Dooren, S. J., Verhoeven, N. M., Cecil, K. M., Ball,
W. S., Degrauw, T. J., and Jakobs, C. X-linked creatine-transporter gene
(SLC6A8) defect: a new creatine-deficiency syndrome. Amer J Hum
Genet, 68: 1497-1500, 2001.

Sowter, H. M., Ferguson, M., Pym, C., Watson, P., Fox, S. 3., Han, C.,
and Harris, A. L. Expression of the cell death genes BNip3 and NlX in
ductal carcinoma in situ of the breast; correlation of BNip3 levels with
necrosis and grade. J Pathol, 201: 573-580, 2003.

Wykoff, C. C., Pugh, C. W., Maxwell, P. H., Harris, A. L., and Ratcliffe, P.
J. Identification of novel hypoxia dependent and independent target genes
of the von Hippel-Lindau (VHL) tumour suppressor by mRNA differential
expression profiling. Oncogene, 19: 6297-6305., 2000.

Nishiwaki, T., Daigo, Y., Kawasoe, T., and Nakamura, Y. Isolation and
mutational analysis of a novel human cDNA, DEC1 (deleted in esophageal
cancer 1), derived from the tumor suppressor locus in 9q32. Gene
Chromosome Canc, 27: 169-176, 2000.

Guo, K., Searfoss, G., Krolikowski, D., Pagnoni, M., Franks, 0., Clark, K.,
Yu, K. T., Jaye, M., and lvashchenko, Y. Hypoxia induces the expression
of the pro-apoptotic gene BNIP3. Cell Death Differ., 8: 367-376, 2001.

103

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

Shoshani, T., Faerman, A., Mett, |., Zelin, E., Tenne, T., Gorodin, S.,
Moshel, Y., Elbaz, S., Budanov, A., Chajut, A., Kalinski, H., Kamer, l.,
Rozen, A., Mor, O., Keshet, E., Leshkowitz, D., Einat, P., Skaliter, R., and
Feinstein, E. Identification of a novel hypoxia-inducible factor 1-responsive
gene, RTP801, involved in apoptosis. Mol. Cell. Biol., 22: 2283-2293,
2002.

Griffiths, J. R., McSheehy, P. M., Robinson, S. P., Troy, H., Chung, Y. L.,
Leek, R. D., Williams, K. J., Stratford, I. J., Harris, A. L., and Stubbs, M.
Metabolic changes detected by in vivo magnetic resonance studies of
HEPA-1 wild-type tumors and tumors deficient in hypoxia-inducible factor-
1beta (HIF-1 beta): evidence of an anabolic role for the HIF-1 pathway.
Cancer Research, 62: 688-695, 2002.

Luster, A. D., Weinshank, R. L., Feinman, R., and Ravetch, J. V.
Molecular and biochemical characterization of a novel gamma-interferon-
inducible protein. J. Biol. Chem., 263: 12036-12043, 1988.

Salnikow, K., Davidson, T., Kluz, T., Chen, H., Zhou, D., and Costa, M.
GeneChip analysis of signaling pathways effected by nickel. J Environ
Monitor, 5: 206-209, 2003.

Vengellur, A. and LaPres, J. J. The Role of Hypoxia Inducible Factor 1a in
Cobalt Chloride Induced Cell Death in Mouse Embryonic Fibroblasts. Tox.
Sci., 82: 638-646, 2004.

Heiniger, C. D., Kostadinova, R. M., Rochat, M. K., Serra, A., Ferrari, P.,
Dick, 3., Frey, B. M., and Frey, F. J. Hypoxia causes down-regulation of
11 beta-hydroxysteroid dehydrogenase type 2 by induction of Egr—1.
FASEB Journal, 17: 917-919, 2003.

Sonna, L. A., Cullivan, M. L., Sheldon, H. K., Pratt, R. E., and Lilly, C. M.
Effect of hypoxia on gene expression by human hepatocytes (HepG2).
Physiol Genomics, 12: 195-207, 2003.

Hofbauer, K. H., Gess, B., Lohaus, C., Meyer, H. E., Katschinski, D., and
Kurtz, A. Oxygen tension regulates the expression of a group of
procollagen hydroxylases. Eur J Biochem, 270: 4515-4522, 2003.

Takahashi, Y., Takahashi, S., Shiga, Y., Yoshimi, T., and Miura, T.
Hypoxic induction of prolyl 4-hydroxylase alpha (I) in cultured cells. J Biol
Chem, 275: 14139-14146, 2000.

Jiang, Y., Zhang, W., Kondo, K., cho, J. M., St Martin, T. B., Dufault, M.
R., Madden, S. L., Kaelin, W. G., Jr., and Nacht, M. Gene Expression
Profiling in a Renal Cell Carcinoma Cell Line: Dissecting VHL and
Hypoxia-Dependent Pathways. Mol Cancer Res, 1: 453-462., 2003.

104

CHAPTER 2

Vengellur, Ajith., Woods, Barbara G., Ryan Heather E., Johnson, Randall .S.,
LaPres, John J. (2003) Gene Expression Profiling of the Hypoxia Signaling

Pathway in Hypoxia-inducible Factor 1 or Null Mouse Embryonic Fibroblasts.
Gene Expression 11: 181-197.

105

Abstract

Hypoxia is defined as a deficiency of oxygen reaching the tissues of the body
and it plays a critical role in development and pathological conditions, such as
cancer. Once tumors outgrow their blood supply, their central portion becomes
hypoxic and the tumor stimulates angiogenesis through the activation of the
hypoxia inducible factors (HlFs). HlFs are transcription factors that are regulated
in an oxygen-dependent manner by a group of prolyl hydroxylases (known as
PHDs or HPHs). Our understanding of hypoxia signaling is limited by our
incomplete knowledge of HIF target genes. cDNA microarrays and a cell line
lacking a principal HIF protein, HIF 1a, were used to identify hypoxia regulated
genes. The microarrays identified a group of 286 clones that were significantly
influenced by hypoxia, and 54 of these were coordinately regulated by cobalt
chloride. The expression profile of HIF10L -l- cells also identified a group of
down-regulated genes encoding enzymes involved in protecting cells from
oxidative stress, offering an explanation for the increased sensitivity of HIF1a -l-
cells to agents that promote this type of response. The microarray studies
confirmed the hypoxia induced expression of the HIF regulating prolyl
hydroxylase, PHDZ. An analysis of the members of the PHD family revealed that
they are differentially regulated by cobalt chloride and hypoxia. These results
suggest that HIF1a is the predominant or HIF isoform in fibroblasts and that it
regulates a wide battery of genes critical for normal cellular function and survival

under various stresses.

106

Introduction

Hypoxia is defined as a state in which oxygen tension drops below the normal
limits of 30 - 60mm Hg found in tissue beds (1). Hypoxia plays a central role in
tumorigenesis but also has profound implications in normal cellular processes
and development (2). Organisms have developed a programmed response to
oxygen deprivation because of the critical role oxygen plays in energy production
and survival (3). Central to an organism’s response to hypoxia is the
transcriptional upregulation of genes capable of stimulating glycolysis,
angiogenesis and erythropoiesis (4-6) (see chapter “introduction”).
Understanding hypoxia signaling, its transcriptional control and a complete
assessment of target genes is critical to our ability to influence this signaling
cascade and thereby treat or prevent illness that results in decreases in available
oxygen. The HIF1azARNT dimer is the most widely studied hypoxia signaling
factor. The HIF1or null animal is embryonic lethal due to vascularization defects,

severely limiting its use for, in vivo studies (7). Recently, a conditional null animal

was created for HlF1or using the Cre-LoxP system (8). Subsequently, an
immortal mouse embryonic fibroblast (MEFs) cell line was established from this
mouse model (8). The conditional animal and cell line make it possible to
understand the role of HlF1or in vivo at various developmental time points,

tissues, and under various conditions.

We used cDNA microarrays and the HIF10t null MEF cell line to study global

gene expression pattern under hypoxia. Our results demonstrate that HlF1or is

107

the primary mediator of hypoxia signaling in MEFs. The microarray experiments
have detailed several groups of HlF1or dependent genes. Among these HlF1a
dependent genes are the glycolytic enzymes and several pro-apoptotic factors.
Most have been previously identified as hypoxia regulated, however some of
these had not been previously characterized as HIF driven genes. Among the
novel HlF1or dependent genes identified in the microarray experiments and
confirmed by relative real time PCR (rRT-PCR) was PHD2. In addition, PHD3
also showed HIF1or dependent regulation. Overexpression of the PHDs was also
shown to lead to a decrease in hypoxia induced gene transcription. The
microarray study also gave some insight into the role of HIF1a in cellular
homeostasis. The direct comparison of untreated wild type and HlF1a -/- cells
identified a group of genes influenced by the loss of the HlF1pt protein. These
included genes that encode for proteins involved in protection against oxidative
stress. These results suggest that the HIF1or protein regulates a wide variety of

genes and may be partially regulated by feedback inhibition.

Materials and methods

Materials: Tissue culture media, supplements and fetal bovine serum were
obtained from lnvitrogen, Inc, Carlsbad, CA. The synthesis of oligonucleotides
was performed by the macromolecular facility, Michigan State University, East
Lansing, MI. SYBR Green Dye and 6-carboxy-X-rhodamine were purchased
from Molecular Probes, Eugene, OR. AmpliTaqTM Gold PCR buffer and

AmpliTaqTM Gold DNA polymerase were purchased from Perkin-Elmer,

108

Wellesley, MA. All other chemicals were reagent grade and obtained from

Sigma-Aldrich Chemicals, St. Louis, MO.

Cell culture, nuclear extracts and Western blotting: Mouse embryonic fibroblast
cell lines were maintained in DMEM media (10% heat inactivated FBS, penicillin-
streptomycin (10 U/ml), non—essential amino acid (10 pglml), L-glutamine (2mM))
and grown at 37°C in 5% CO2. Hypoxia treatments were performed in an
oxygen-regulated incubator (Precision, Winchester, VA) or simulated by treating

the cells with 100 pM of CoCI2 (dissolved in H2O).

Wild type and HIF 1a -/- cells were grown under control (20% 02), or hypoxic (1%
02) conditions or in the presence of 100 pM CoCI2 for 6 hours. Nuclear extracts
were prepared as described previously with minor modifications (9). Briefly, cells
were washed with cold PBS (4°C) and removed from surface by scraping in cell
lysis buffer (10 mM Tris-HCI, pH 8.0, 1 mM EDTA, 150 mM NaCI, 1 pg/mL each
of Aprotinin, Leupeptin, Pepstatin A and 1 mM PMSF) and homogenized. The
nuclei were removed by centrifugation (4000 rpm, 15 min) and the supernatant
was saved. The nuclei were lysed upon addition of nuclei lysis buffer (20 mM
HEPES, pH 7.9, 400 mM NaCl, 1 mM EDTA, and 1 mM DTT,,ng/mL each of
Aprotinin, Leupeptin, Pepstatin A and 1 mM PMSF). Insoluble matter was
removed by centrifugation (10000 x g, 15 min). The protein concentration of the
supernatant was determined by standard colorimetric assay via manufacturer’s

instructions (Bio-RadTM, Hercules, CA). Equal amounts of protein for each

109

sample were separated by SDS-PAGE, and Western blotting was performed with
a HIF1or specific antibody (Novus Biologicals, Littleton, CO). A horseradish
peroxidase conjugated secondary antibody was used to facilitate detection. The
blot was stripped and reprobed with a B-actin specific antibody (a generous gift
from Dr. John Wang, Michigan State University). The blots were exposed using

a chemiluminescent assay kit (Amersham Biosciences, Piscataway, NJ).

cDNA microarray production: The National Institute of Aging 15K clone set was
used for the production of the microarray (10, 11). The NIA 15K set was
supplemented with an additional96 clones to bring the total number of clones to
15,360. Each clone was PCR amplified with modified M13 forward and reverse
primers (Table 9). Each primer was modified on the 5’ end with an amino linker
to facilitate linkage to the aldehyde-modified glass substrate. Amplicons were
precipitated and the size and quality verified by agarose gel electrophoresis, then
resuspended in 3X SSC. The full 15K set was printed on TelechemTM
Superaldehyde slides (Telechem International Inc., Sunnyvale, CA) using an
OmniGridTM robot (GenMachines®, San Carlos, CA) with Chipmaker 2

microarray pins (Telechem International Inc).

Microarray experimental design: The microarray experiments were performed
using a “Loop” design (12) (Figure 8). This design was chosen to increase the
statistical strength of the model, to allow for effective comparison among

treatments and cell types and to facilitate the long-term goals of the project.

110

   
 

moms;— can

Lo.— wpoEtn. .m wink

111

/Control\

 

 

\Null

Hypoxia ' Control

- Null

Cobalt: Cobalt
Null
\Hypoxia/

Figure 8. Experimental Design of Loop

Graphical representation of the 2V loop experimental design. Each arrow
represents a single microarray slide. The arrowhead represents the sample
labeled with Cy5 and the arrow tail represents sample labeled with Cy3.
Since the loop was performed twice, each treatment group sample was
labeled in 8 separate reactions, 4 with Cy3 and 4 with Cy5

112

Total RNA from two independent 10 cm tissue culture plates (approximately 1
x107 cells) within each treatment were labeled. Each RNA sample was labeled
four times, twice with Cy3 and twice with Cy5. Two complete loops were
performed; therefore, four independent cultures were labeled twice, once with
each dye, equaling a total of eight individual labeling reactions per treatment
group. It is important to note that the loop design does not utilize the calculation
of expression intensity ratios but relies on the averaged expression within each
treatment over all of the labeling reactions (in this case 8). Expression patterns
are then statistically analyzed to determine those genes for which expression is
significantly changed (below). Ratios can be calculated for each treatment by

direct comparison of the averaged expression between the treatments of interest.

RNA extraction and cDNA microarra y probe labeling and hybridization: RNA
extraction was performed using Trizol reagent (lnvitrogen) via the manufacturer’s
instructions and quantitated spectrophotometrically. 30 pg of total RNA from wild
type or HlF1a -/- cells grown for 24 hours under control (20% 02) or hypoxic (1%
02) conditions or in the presence of 100 pM CoCI2 was used to generate probes
as follows: Total RNA was incubated with 6 pg anchored oligo-dT primers (5’-
T21VN-3’). The reverse transcription was performed using SuperscriptTM ll RNase
H- Reverse Transcriptase (400 U, lnvitrogen) in the presence of aminoallyl dUTP
(aa-dUTP) (Amersham Biosciences, Piscataway, NJ), dTTP, dATP, dCTP and
dGTP (final dNTPs = 0.5 mM). The ratio of aa-dUTP to dTTP was 2:1. The

unincorporated nucleotides were removed using the QlAquickTM PCR purification

113

kit (Qiagen lnc., Valencia, CA) according to the manufacturer’s instructions, after
substituting the wash buffer with phosphate wash buffer (5mM KPO4, pH = 8.0 in
80% ethanol), and eluted with 4 mM KPO4 (pH = 8.5). The probes were dried
under vacuum centrifugation and resuspended in 0.1 M Na2C03 (pH = 9.0) and
conjugated to monoreactive Cy3 and Cy5 dyes (Amersham Biosciences).
Unincorporated dyes were removed using the QlAquickTM nucleotide removal kit
(QIAGEN |nc.). Microarrays were prehybridized in pre-hyb buffer (5X SSC, 0.1 %
w/v SDS, 1 % w/v BSA) for 2 hours at 42°C. Cy3 and Cy5 labeled probes were
resuspended in hybridization buffer (50% formamide, 5X SSC, 0.1% SDS, 20 pg
Cot1 DNA, and 20 pg Poly (A)-DNA) and denatured at 95°C for 3 minutes. The
probes were hybridized for at least 20 hours to the microarrays at 42°C under
Lifter slipsTM (Erie Scientific Company, Portsmouth, NH). Slides were washed as
follows: 1) brief wash with 1x SSC and 0.2% SDS at 42°C to remove the cover
slips; 2) 4 min in 1x SSC and 0.2% SDS at 42°C; 3) 4 min in 0.1x SSC and 0.2%
SDS at room temperature; and 4) twice for 2.5 min each in 0.1x SSC at room

temperature. Slides were dried and scanned immediately.

Microarray data analysis: Slides were scanned with an AffymetrixTM 428 scanner
(AffymetrixTM, Santa Clara, CA) at 532 nm (Cy3) and 635 nm (Cy5). The images
were analyzed using GenePixTM Version 3.0 software (AxonTM Instruments Inc.,
Union City, CA) and output files were analyzed using GP3
(http://www. bch.msu. edu/~zachareflmicroan'ay/gp3. html) (1 3). GP3

automatically flags spots with intensities below background levels, corrects for

114

background signal, and normalizes the two channels using a z-score method.
The GP3 output values were normalized within the loop using the Centralizer
program (14) (http://cartan.gmd.de/~zien/centralization/). Genes differentially
expressed between treatments were identified by t-test (two tailed, unequal
variance, p <= 0.05 cut-off). All genes called present (signal > [background + 2
standard deviations of background]) were hierarchically clustered using

GeneSpringTM, v 4.2.1 (Silicon Genetics, Redwood City, CA).

Relative Real-Time PCR analysis: Changes in gene expression observed by
microarray analyses were verified by real-time PCR, performed on an Applied
Biosystems Prism 7000 Sequence detection System (Foster City, CA) (for
detailed protocol see chapter 1 methods section). Gene names, accession

numbers, forward and reverse primer sequences are listed in Table 9.

Transient Transfection of Hep3B cells: HepSB cells were maintained in DMEM +
5% FBS (growth media). Prior to transfection, cells were washed twice with
serum free medium (DMEM). Cells were transfected using lipofectamine
(lnvitrogen, Carlsbad, CA) via manufacturer’s instructions using 5 ul of
lipofectamine (LF) with an incubation time of 6 hours prior to addition of growth
media. Cells were transfected with an HRE-driven Iuciferase reporter, a [3-
galactosidase (BGAL) expression vector for control of transfection efficiency and
the various PHD expression vectors (A generous gift of Dr. Peter Ratcliff,

University of Oxford, United Kingdom). All transfections were normalized for

115

DNA content with an empty expression vector. DNA was incubated with
rehydrated LF for 15 minutes prior to transfection. Cells were then treated with
DNAzLF mix for 6 hours. Following this, incubation medium was removed and
replaced with growth media and placed into control (20% O2) or hypoxia (1% O2)
incubator. Cells were incubated for 18 hours and assayed for Iuciferase and (3-
Gal activity via standard assays (Promega, Madison, WI). Luciferase values
were normalized to BGaI activity in each well and untreated empty expression

vector within experiments.

Results

Verification of the MEF cell line phenotype: A Western blot was performed on
the cell lines used in our microarray experiments to verify their phenotype. Wild
type and HlF1a -/- cells were exposed to CoCI2 (100 pM) or hypoxia (1% O2) for
6 hrs. Nuclear extracts were prepared from treated and untreated cells and
analyzed using a HlF1or specific antibody (Novus Biologicals, Littleton, CO). As
expected, HIF1a was only detectable in the wild type cells. A very small amount
was observed in untreated cells but greatly increased following exposure to
cobalt or hypoxia (Figure 9). There was also a migratory shift in these treated
fractions suggesting a protein modification. The blot was reprObed with an

antibody specific to B-actin to verify loading (Figure 9).

Hierarchical Cluster analysis of microarray data: cDNA microarrays have given

researchers an opportunity to understand signaling networks as they relate to

116

HIF1a +I+ HlF10r -I-

 

C ‘ Co” Hyp C Co2+ Hyp

   

HIF1a —» 7

Actin —’

Figure 9. HIF1a Western Blot

Vlfrld type (HIF1a +I+) and Null (HlF1a -l-) cells were exposed to CoCl2
(Cozt, 100 pM) or Hypoxia (Hyp, 1% 02) for 6 hours or left untreated (C).
Nuclear extracts were prepared and proteins were separated by SDS-
PAGE, transferred to nitrocellulose membrane and probed with a HIF1or
specific antibody (upper panel). Blot was reprobed with B-actin to verify
equal loading (lower panel).

117

global changes in gene expression. Here microarrays were used to characterize
hypoxia signaling and the role of HlF1a in this process. Two cell lines were
used, the wild type and the HlF1or -I- MEFs. Each cell line was left untreated
(Control, 20% 02) or treated with hypoxia (1% O2) or cobalt chloride (100 pM) for
24 hours. Cobalt has been extensively used as a hypoxia mimic and has been
shown to inhibit the activity of the PHDs (15). The microarray experiments were
performed in a “loop” design as described above (Figure 8) and statistically

analyzed (16).

Hierarchical cluster analysis of microarray data along treatments and
genes creates relationships between groups based on expression patterns
observed in the gene set as a whole. Here we used cluster analysis to identify
combinations of treatments and genotype that were similar at the gene
expression level and to identify groups of genes responding in a similar manner
among the various treatment and genotype combinations. As expected, results
from CoCI2 and hypoxia treatments in the wild type cells clustered together
(Figure 10). This similarity in gene expression between these two treatments did
not extend to the null MEFs. These results suggest that CoCI2 and hypoxia are
driving expression of a similar gene cluster and that HlF1a is essential for this
regulation. Interestingly, the HIFtor -/- cells under hypoxic conditions behaved
most like the null control cells (Figure 10). The loss of hypoxia signaling and
comparable expression pattern between hypoxia and control suggests that

HlF1or is central to hypoxia mediated signaling in these cells.

118

  

‘,','IIF""=
.‘ I Ill ”I” I “1“—

1 H? I: I7 |.||,IIIIIIFfimEI-ml

      
    

 

WT Hypoxia
WT Cobalt
Null Hypoxia
Null Control
WT Control
Null Cobalt

Figure 10. Hierarchical Cluster diagram of entire data set

The microarray data was filtered to remove genes for which expression
was determined to be insignificant (signal < {background + 2 st. dev.
Background}). Expression results from the filtered data were clustered by
treatment and gene. The relationships are represented as a dendrogram
in both directions. Red represents a high level of expression and green
represents low expression.

119

Statistical comparison of hypoxia and cobalt treatment in WT cells: The major
focus of these microarray studies was to identify a more complete battery of
hypoxia regulated genes. This was accomplished first by statistical assessment
of expression patterns within the data. Stringent criteria were set to minimize the
possibility of false positives. First, the results from wild type cells were screened
for all genes whose expression was significantly altered by treatment (p < 0.05
compared to untreated control). The results show that expression of 286 clones
was significantly (p s 0.05) altered by hypoxia and 252 were altered by cobalt
(Figure 11). Next, we determined the overlap within these two sets and found
that there were 54 clones representing 49 different genes that were in both of
these groups. These 54 clones were then analyzed for function and grouped
accordingly (Table 10). Finally, we determined how many of these genes were
also altered in the HlF‘IpL-I- cells under the same conditions. None of these 49
genes were altered in both hypoxia and cobalt in the null cells. However, seven
of the 49 genes in table 10 were significantly altered in the HlF1a -/- cells in
either cobalt or hypoxia treatment. The genes for metallothionein (clone A-2031)
and transglutaminase (clone H3137C06) were influenced in the HIF1a -/- cells by
cobalt. The HIF1or independent increase in metallothionein is most likely due to
its regulation by the metal transcription factor (17). Five of the 49 genes were
shown to be significantly influenced by hypoxia in the null cells. These genes
were pyruvate kinase 3 (clones H3030D10 and H3030D11), cathepsin L (clone
H3028F03), galectin 3 (clone H3016D10), tissue specific transplantation antigen

P35B (clone H3147H11) and one gene with unknown function (clone H3116A06).

120

Cobalt

(198)

 

Figure 11. Analysis of CoCI2 and hypoxia treatment in wild type

cells

The Venn diagram represents total number of genes unique to each
treatment group (in parentheses) and the number of genes shared
between the two treatments.

121

Table 10. Significantly influenced genes in WT cells treated
with or cobalt

 

122

Table 10. Significantly influenced genes in WT cells treated
with or cobalt

       

123

It should be noted that the pyruvate kinase 3 gene was significantly
downregulated in the null cells while being significantly upregulated in the WT

cells under hypoxia.

Only 12 of the 49 genes in Figure 11 and Table 10 had been previously
associated with hypoxia signaling. The 54 clones were separated into 7
categories by ontology. The largest group that could be assigned a function
consists of glycolytic enzymes (11 clones, 7 genes, Table 10). These enzymes
are well known targets of the hypoxia signaling cascade. There were several
genes in other groups that also had been identified as hypoxia targets. Most
notably are proline-4-hydroxylase (P4H, clene H3003D12) and BCL2/adenovirus
E18 19 kDa-interacting protein 3-like (BNIP3L or NIX, clone H3016D08). There
were also several groups that had not been previously identified as hypoxia
targets. These include a large group of matrix remodeling genes and cellular
antigens. The largest group included clones for which function has not yet been

assigned (Table 10).

Comparison of hypoxia and cobalt treatment in HlF1a -/- cells: As noted above,
only a handful of genes identified in Figure 11 and Table 10 were significantly
changed in null cells. None of these genes (metallothionein, pyruvate kinase,
transglutaminase, cathepsin L and galectin 3) were influenced in both treatments
in the null cells. To further clarify the HlF1or independently regulated genes from

the microarray studies, a comparison between cobalt and hypoxia treated HlF1a

124

-/- cells was performed. There were 376 clones significantly altered in these cells
upon exposure to hypoxia (1%, 24 hours) and 349 clones significantly altered
upon exposure to cobalt (100 pM, 24 hours) (data not shown). There were 65
clones that were shared between these two lists (Figure 12A). Of these 65
clones, none were found in the list for the wild type cells (Table 10, Appendix
data table 11). These 65 clones represented several different enzymes including
subunits of ATPase and biliverdin reductase B (BLRVB). BLRVB is downstream
of heme oxygenase 1 in the metabolism of heme. Heme oxygenase, a known
hypoxia regulated gene, converts heme into carbon monoxide, biliverdin and free
iron. BLRVB converts biliverdin into bilirubin (18-20). This suggests that the
pathway for heme degradation is not coordinately regulated under hypoxic
conditions. This group also contained a receptor for tumor necrosis factor and
raises the possibility that hypoxia and hypoxic mimics influence the expression of
the receptor of a factor known to play a role in hypoxia signaling (21, 22). These
65 clones presumably represent a subset of genes that can be influenced by
hypoxia and cobalt in a HlF1or independent manner and may involve HlF2a or

HlF3a since both are expressed in these cells at a much lower level then that of

HlF1or (data not shown).

Comparison of hypoxia treatment in wild type and HlF1a -/- cells: The
comparison described in Figure 11 and 12A were done using those genes that
were significantly changed by hypoxia and cobalt. To identify genes regulated in

a

125

 

HIF1or -/-

 

Hypoxia

    

Cobalt

Figure 12. Venn diagrams of various combinations of cell types and
treatments

Total number of genes unique to each treatment group (in parentheses) and
the number of genes shared between the two treatments (central shaded area).
(A) The Venn diagram representing genes significantly influenced by hypoxia
and/or cobalt in HlF1or -/- cells. (B) Comparison of WT and HIF1or -/- cells
following hypoxia treatment (1 % 02, 24 hrs). (C) Comparison of WT and HlF1a
-/- cells following cobalt treatments (100 pM, 24 hrs).

126

HlF1or independent manner by hypoxia alone, a comparison between hypoxia
treatments in wild type and HIF1or -/- cells was performed. There were 26 genes
that were significantly influenced by hypoxia in both cell types (Figure 128,
Appendix data table 12). These genes included those described above (Figure
11, Table 10), as well as, procollagen. Hypoxia-induced collagen synthesis has
been previously described and may be critical to several disease states (23, 24).
Interestingly, procollagen is upregulated in the wild type cells and down-regulated
in the HlF1a -/- cells upon exposure to hypoxia, similar to the pyruvate kinase 3

gene mentioned above.

Comparison of cobalt treatment in wild type and HIF1 a -/- cells: A comparison
between wild type and HIF1a -/- cells was also performed for genes significantly
influenced by cobalt. Though cobalt is used as a hypoxia mimic, there is no
overlap between the 14 genes identified in this comparison and those found for
the hypoxia comparison (Figure 128 and 120, Appendix data table 13). The 14
genes in this list include metallothionein and transglutaminase 2 (T92). T92 was

previously identified as a VHL/hypoxia target gene (25).

Comparison of WT ctri and HIF1a -/- ctri cells: Finally, a direct comparison
between untreated WT and HlF1a -/- cells was performed to determine the role
of HIF1a in cellular homeostasis. There were 234 genes influenced by the loss
of HlF1or_ (data not shown). Of these, H19 was significantly upregulated by the

loss of the HIF protein. The three clones, H3144BO7, H3144BO6, H3140G12,

127

displayed a WT control/HIF1a -/- control ratio of 0.067, 0.211 and 0.131
respectively (average 9 fold higher in HIF1or -/- cells). H19 is an imprinted gene
involved in development and not highly expressed following birth (26). The
reason for its aberrant expression following loss of the HIF1a protein is unknown.
Another group of genes also stood out in this comparison of control cells. The
HlF1g-l- cells displayed a decreased expression level of genes involved in the
cellular response to oxidative stress. These genes include glutathione S
transferase-p (GST—p, clones = H3134F05, H3119G08, H3159G05), superoxide
dismutase 1 (SOD1, clones = H313OB11, H313OB11), SOD2 (A-2062) and
catalase (A-2049). This decreased level of these protective genes may explain
why HIF1or -/- cells are more sensitive to oxidative stress (27). In addition, this
decrease adds to the growing evidence of the complex role hypoxia signaling

plays in normal cellular homeostasis and tumorigenesis.

The glycolytic cluster: One hypothesis of expression profiling is that genes that
cluster together are regulated in a similar manner. We wanted to identify novel
genes that may also play a role in the cellular response to hypoxia, and our
hypothesis was that they would behave similarly, at the expression level, to the
glycolytic enzymes. The glycolytic enzymes are critical to a cell’s response to
low oxygen tension and are direct targets of HIF regulated transcription. We
isolated a cluster of glycolytic genes from our microarray data and looked for
functional information for several of the non-glycolytic genes (Figure 13). First,

three genes involved in apoptosis were found within this Cluster. Two are known

128

r—w‘lfillrh

 

Ilfirilfififimflll

 

rH

Wi

Id Type

HIF1a -I-

 

0‘

trl

Co2+ Hyp

Ctrl

Co”

I.
l

 

I
‘<

Figure 13. Glycolytic enzyme Cluster
The cluster of glycolytic enzyme genes (Fig. 4) was further analyzed to
identify novel hypoxia regulated genes. Four subsets of genes were
identified; the glycolytic enzyme"s genes (blue), apoptotic genes (green),
hydroxylases (red) and others (black). Red represents a high level of
expression and green represents low expression.

1_-
N

9

”a

NO HIT (H3123A12)

TPI (H3149010)

a ENOLASE (H3027E07)
HMGP 1 (H3027D07)

LDI-I1 (H3023H12)

PGK (Hsozsooa)

a ENOLASE (Hsozaeos)
ENOLASEt (H3022l-l02)
ENOLASE (H3027Eos)
NAP1-L (H3030602)

igM HEAVY CHAIN (H3085G01)
ALDOLASE A (H301reos)
GAPDH (H3019E08)

GAPDH (H3025F11)

GAPDH (H3022E07)

NIX (H3103BOT)

ALDOLASE (H3150A01)
TRITHORAX-llke (H3013E11)
BNIP3 (H3136c07)

EST AI447so4 (H3059F11)
EGL9 HOMOLOGUE 1 (H3145803)
GAPDH (H3012A11)
ALDOLASE A (H3031 D03)
PS RECEPTOR (H3142H08)
BASIGIN (H3020G04)
GAPDH (H3101C10)

P4-H (H3003012)

(BNIP3 and NIX) and one is a novel (PS receptor) hypoxia—regulated gene (28,
29). The role that these genes play in the downstream events following acute

and Chronic hypoxia is currently being investigated.

Second, there were two hydroxylases identified within this glycolytic cluster
(Figure 13). The first, P4-H, is a pro-collagen prolyl 4-hydroxylase and a known
hypoxia regulated gene (30). It functions in a similar manner to that of the PHDs,
however, it is unable to influence HIF stability in C. elegans (15). The second is
EGL-9 homologue 1, also known as PHD2. As mentioned earlier, the PHDs
negatively influence HIF signaling by controlling HIF stability. This result
confirms previously published data and supports the hypothesis that HlF1a may
be partially controlled by feedback inhibition (15, 31). It should be pointed out
that PHDZ was not listed in Table 10 (genes altered by cobalt and hypoxia in WT
cells) because the increase seen following cobalt exposure for this Clone did not

pass statistical cut off (Figure 14).

Relative Real time PCR: We verified our microarray data by rRT-PCR using
SYBR® Green as a fluorescent marker. The glycolytic enzyme glyceraldehyde-
3-phosphate dehydrogenase (GAPDH) is a known hypoxia regulated gene and
acted as a positive control for our SYBR® Green protocol. GAPDH was also
present on the microarray and shown to be regulated in a HlF1a—dependent
manner (Figure 11, 13 and Figure 14A). In addition, the loss of GAPDH

regulation by hypoxia in the HlF1a -/- suggests that HlF1a is the primary

130

    

A. Microarray a. rRT-PCR
3.0 -—
GAPDH - 4
:2 2‘ ‘5 GAPDH 1*
n 3-
D
g 2. 2
" 1 ft
m . m 2
‘” o
5 1. ,>
(I a
To 0 g 1
K ' m

 

 

 

Ctrl Co" Hyp

.L
0

Relative Expression P
N A O O

 

 

 

Ctrl Co" Hyp
PHDZ

N .N
0:

"<3

 

8
Relative Expression :I'I

Relative Expression !'“
0|

.°
01

 

   

Ctrl Co2+ Hyp cm Co2t Hyp
Treatment Treatment

Figure 14. Comparison of microarray and rRT-PCR results

GAPDH (A, B), P4H (C, D) and PHD2 (E, F) were analyzed for expression
levels in wild type (hatched bars) and HlF1a -/- cells (solid bars) by
microarray (A, C, E) and rRT-PCR (B, D, F) protocols. Each value was
normalized to the control level in the corresponding cell line. Ctrl =
untreated, Co2+ = 100 pM CoCl2 and Hyp = 1% 02 for 24 hours. Averaged
expression within each treatment group was used to generate ratios (see
experimental design). * Represents p < 0.05 when compared to
corresponding control

131

mediator of hypoxia signaling in MEFs. These results were confirmed in the rRT-
PCR. Six separate biological samples were analyzed for each separate
treatment in the two cell lines. There was a significant increase in GAPDH
message in the cobalt and hypoxia treated wild type cells when compared to the
control. This increase did not extend to the HIF10I -/- cell line (Figure. 148). A
similar HIF1a-dependent expression pattern was seen on the microarray for P4-
H (Figure 13 and 140). This result was also verified on the rRT—PCR (Figure.
14D). The results for the cobalt treatment were not as pronounced as that
following hypoxia exposure, though both were statistically significant. Finally, the
expression pattern of PHD2 was also confirmed by rRT-PCR (Figure. 14E and
14F). There was only a modest increase in expression when the microarray was
analyzed. The expression pattern among the treatment groups, however, was
identical to that of other known hypoxia regulated genes in the Cluster (i.e.
increased in cobalt and hypoxia in wild type cells, no change in HIF1or -l- cells).
The rRT-PCR results confirmed that PHD2 transcription increased in response to
cobalt (> 3 fold) and hypoxia (> 6 fold) when compared to the control treatment
only in the wild type cells. These results suggest that HlF1or may indirectly
control its own stability by affecting the level of the prolyl hydroxylase that marks
it for degradation. This data also show the level of data compression that is often
seen in microarray data due to the dynamic range of the microarray scanner,
normalization and other mathematical manipulation necessary for comparisons
between multiple experiments and labelings (32). The difference in fold change

in the microarray data is consistently underestimated. The relative compression

132

of the data for various genes ranged from 2 to 10 fold lower on the microarray

compared to rRT—PCR.

Given the implications of hypoxia-regulated PHD expression we also analyzed
the expression pattern of all three PHDs for their time and HIF1a dependent
expression by rRT-PCR. Wild type and HIF1a -/- cells were left untreated or
exposed to hypoxia (1% 02) or cobalt (100 pM) for 0.25, 2, 8, 24, 48 and 72
hours. PHD1 exhibited small but significant changes in expression at various
times, however these changes did not follow a pattern and did not appear
HlF1or dependent (Figure 15). PHD2 was significantly upregulated in a
HlF1or dependent manner in the wild type cells as early as 15 minutes after
exposure in both the hypoxia and cobalt treated cells. PHD3 showed a large
increase (> 11-fold) in the wild type cells following hypoxia but not cobalt
exposure (Figure 15). The time course of this induction was also notably
different from that of PHD2. The mechanism that controls the expression
differences between the various PHDs has not been determined. These results
indicate that the PHD genes are regulated by hypoxia in different ways. In
addition, the identification of PHD2 and PHD3 as hypoxia-regulated genes
suggests that the hypoxia signaling cascade may be partially Controlled by

feedback inhibition.

Finally, the microarray data suggests that loss of HIF1or leads to a down

regulation of the basal levels of oxidative stress genes. The basal level of

133

Wild Type

 

8

PHD1

 

n
as
I
l

.5
es

 

 

.3
N
|
l

 

Normalized Exp. (X10)
0

 

 

 

 

(x100)

 

 

.100

 

N
0}

 

3

Normalized Exp
M
q:

 

O

 

 

 

.3
O

7

j

 

Normalized Exp. (x100)
N
c;

 

O
I
’1

Time Points (hrs)

7

 

N '////////.

 

 

HIF1oI. -I-

 

PHD1

 

 

 

 

 

 

125

 

 

100

754

 

 

 

 

 

 

 

 

 

 

125 2 a

 

24 48 72

Time Points (hrs)

Figure 15. Time course analysis of the PHDs by rRT-PCR

The expression levels of the various PHDs were analyzed by rRT-PCR in wild
type (left) and HlF1oI -I- (right) cells. Each bar represents an average value of
6 samples normalized to the expression level of the control gene and graphed
as normalized expression. Black bars = untreated, White bars = 100 pM
CoCl2 and Hatched bars = 1% Oz. * Represents p < 0.05 when compared to

corresponding control.

134

 

expression of GST-p, SOD1, SOD2 and catalase were verified by rRT—PCR. As
seen in Figure 16, the basal levels of GST-p, SOD1 and 8002 were significantly
repressed following loss of the HlF1or proteins. Catalase was not influenced in
this comparison. The decreased level of expression of these protective enzymes
may explain why HlF1or -l- cells are more susceptible to oxidative stress induced

damage (27).

PHD effects on HIF mediated transcription: The possibility that upregulation of
the PHDs might lead to a feedback inhibition of HIF mediated signal was tested
in transient transfection experiments. Hep3B cells were transfected with an HRE
driven luciferase reporter and the various PHD expression plasmids. The cells
were left at control (20% O2, ctrl) or hypoxia (1% O2) for 18 hours and analyzed
for luciferase expression. PHD2 was capable of completely inhibiting HRE
mediated Iuciferase expression (Figure 17). PHD3 was also capable of inhibiting
this transactivation but to only 50% of controls (Figure 17). Finally, hypoxia
mediated transcription displayed a 40% reduction in the presence of PHD1;
however, this decrease was not significant. This supports the notion that these
enzymes play a role in regulating the initial HIF signal and are also involved in

attenuating the signal by feedback inhibition (15, 31, 33).

Discussion

Hypoxia mediated signaling is critical to normal development and is strongly

involved in several pathological conditions, including cancer and heart disease.

135

 

 

12 0.7

A. SOD1
10 ~ 0.6

$002

 

 

 

0.5-

 

2L1”

 

 

 

///'////2-

  
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

..._ ,,. 22:—8 I

E 3: c. a}... :3. as.-.
23121;? . I ”N
..._.\\ °-2-—\ ..
2::§ . 2&1

Figure 16. Analysis of basal levels of GST-p, SOD1, SODZ and catalase
by rRT-PCR

The basal level of expression of genes involved in the cellular response to
oxidative stress was analyzed by rRT-PCR. SOD1 (A), SOD2 (B), Catalase
(C) and GST-p (D) were analyzed in untreated control cells to determine
their relative levels of expression in wild type (hatched bars) vs. HlF1a -/-
cells (solid bars). ** Represents p < 0.05, * Represents p < 0.1.

136

 

0|
0

 

 

 

 

00 uh
O O
3!-

 

 

N

O

\
‘

 

.3
O

 

/

 

Relative Luciferase (arb. units)

 

 

cm P4H PHD1 PHD2 PHD3

Figure 17. Transient transfection of Hep3b cells

Hep3b cells were transiently transfected with a HRE driven reporter
and one of the PHDs or an empty expression vector as a control.
Transfected cells were left untreated (Dark Bars) or exposed to
hypoxia (1% 02 for 18 hours, Hatched Bars) Luciferase levels were
determined and normalized to co-transfected B—Gal activity and then

to untreated control.

137

Characterizing the upstream and downstream events following onset of hypoxia
is critical to our ability to treat and understand these conditions. Recent
publications have increased our understanding of the upstream signals involved
in HIF stabilization and the role that hydroxylation plays in mediating the HIF:VHL
interaction (15, 34, 35). However, our understanding of the downstream events
is not as complete. For example, we do not know the events and proteins
involved in translocating HIFs into the nucleus. In addition, we do not know what,
if any, protein is involved in controlling the ability of HlFs to interact with ARNT or
ARNT2. Finally, we do not know the complete battery of hypoxia-regulated
genes. The microarray is an excellent tool to address the latter knowledge gap.
To this end we have used the NIA 15K gene set to identify novel hypoxia

regulated genes.

Our microarray experiments confirmed the expression of several known hypoxia
regulated genes, including glycolytic enzymes, pro-apoptotic genes and a prolyl
hydroxylase (Figure 11 and 13 and Table 10). These act as positive controls to
confirm that our treatments, microarray experiments and data processing are
valid. In addition, it gave added confidence to the results that we observed for
the novel hypoxia-regulated genes identified. To further validate the analysis a
comparison of the 49 gene identified in Figure 11 and Table 10 was made with a
recently published report that used serial analysis of gene expression (SAGE) to
analyze the effect of VHL and hypoxia on normal and tumor derived renal cells

(36). Given that the species and platforms are different, the extent of overlap will

138

be underestimated. However, of the 49 genes identified as hypoxia responsive
in this study, 12 were found to be hypoxia responsive in the SAGE study
(Appendix data table 14 and figure 3 of (36)). This number is most likely higher
due to the nature of SAGE and the genes represented on the microarray. For
example, GLUT1 and GLUT2 were shown to be hypoxia responsive via SAGE
while GLUT4 was found in this study (Appendix data table 14). In addition,

GLUT1 and GLUT2 are not represented in the NIA 15K.

The microarray confirmed that the hypoxic mimic, CoC|2_ increases a battery of
genes similar to that of hypoxia. This overlap between expression patterns was
only partial, suggesting differences in secondary signals or downstream signaling
molecules. The exposure period was 24 hours and each treatment may initiate
other pathways and distinct secondary responses that will be observed at this
time point. It should also be pointed out that the results observed in these
studies with respect to the PHDs are in agreement with previously published
reports; however, the time course and differences seen between cobalt and
hypoxia treatments had not been previously described (15, 37, 38). Finally, the
microarray experiments suggest that HIF1or is the primary cytoplasmic controller
of hypoxia signaling in MEFs. The role of HlF2a and HlF3a has yet to be

determined in these cells.

The identification of PHD2 and PHD3 as hypoxia-regulated genes implies that

HIF1a is subject to feedback inhibition (31). This feedback is not surprising when

139

considering the implications of overstimulation of the hypoxia-signaling cascade.
For example, the VHL protein is thought to be a tumor suppressor in part
because of its ability to mediate the degradation of HIF (39). Once the VHL gene
is mutated, this ability is lost and HIF signaling goes unabated leading to the
overproduction of a number of factors such as VEGF. This unchecked
production of growth factors and other proteins may explain the
hypervascularization of VHL -/- neoplasms (39). The control of the hypoxia-
signaling cascade, both upstream and downstream, is therefore critical to normal
cells and tissue homeostasis. We speculate that the feedback loop would limit
the expression of the pro-apoptotic genes, thus giving the cell an alternative to
programmed cell death. For example, the oxygen concentration of the cell at
equilibrium may be “set” to limit damage due to small decreases in oxygen
tension. Once these small fluctuations are encountered, the HIF protein is
stabilized and the cell begins to transcribe the glycolytic, pro-apoptotic and PHD
genes. As the level of the PHDs increases, the equilibrium between the available
oxygen, PHDs and HIF stabilization is shifted in favor of HIF hydroxylation. The
increase in available PHDs would compensate for the small decrease in available
oxygen. However, at larger decreases in oxygen tension, the cell would not be
able to compensate by overproduction of the PHDs and the pro-apoptotic gene
transcription would continue until the cell began its programmed cell death. One
other feature of this feedback loop is the level of the individual PHDs within any
given tissue and cell type. PHD1 was not upregulated by cobalt or hypoxia so

presumably it would not be able to participate in this type of regulation.

140

Therefore, tissues that only express this isoform of the hydroxylase would not
exhibit feedback inhibition. This idea of tissue specific regulation and

hydroxylase differences is supported by recent publications (37, 38, 40).

The role of HlF1or in maintaining proper cellular function during development and
normal function is often neglected. The comparison of untreated WT and HlF1a
-/- cells clearly showed that HIF1a has an important role in normal homeostasis.
A battery of protective genes, such as SOD1 and SOD2, calmodulin (clones
H3019D01 and H3021E08) and genes encoding metal binding proteins
(metallothionein, Clone 1-2031 and CRIP1 clone H3108G04) were shown to be
affected negatively by the loss of the HIF1a protein (see Figure 16, data not
shown). In addition, several genes were positively influenced by the loss of
HIF1a. These results suggest that a basal level of HlF1a must be present to
maintain transcriptional regulation for these genes. Alternatively, the loss of
HIF1a may upset the signaling balance within the cell and these changes are the

result.

In summary, these observations suggest a role for the PHDs in the feedback
inhibition of the hypoxia-signaling cascade. The results also support the idea
that cobalt and hypoxia exposure lead to the increase in the transcription of a
number of gene families, including hydroxylases, glycolytic enzymes and pro-
apoptotic genes. The interplay among these families and the regulation of the

downstream events may help to explain the role of hypoxia in mediating the

141

damage following several pathological conditions, such as stroke and

cardiovascular disease.

Acknowledgements

The authors wish to thank the Tim Zacharewski laboratory for their careful
reading of this manuscript and the extensive bio-informatics support. We would
especially like to acknowledge Kirsten Fertuck, Lyle Burgoon and Darrell

Boverhof for their help in these endeavors.

142

References

10.

11.

Vaupel, P., Schlenger, K., Knoop, C., and Hockel, M. Oxygenation of
human tumors: evaluation of tissue oxygen distribution in breast cancers
by computerized 02 tension measurements. Cancer Res, 51 : 3316-3322.,
1991 .

Semenza, G. L. Regulation of mammalian 02 homeostasis by hypoxia-
inducible factor 1. Ann. Rev. Cell Dev. Biol., 15: 551-578, 1999.

Bunn, H. F. and Poyton, R. O. Oxygen sensing and molecular adaptation
to hypoxia. Phys. Rev., 76: 839-885, 1996.

Li, H., Ko, H. P., and Whitlock, J. P. Induction of phosphoglycerate kinase
1 gene expression by hypoxia. Roles of Arnt and HlF1alpha. J. Biol.
Chem., 271: 21262-21267, 1996.

Maltepe, E. and Simon, M. C. Oxygen, genes, and development: an
analysis of the role of hypoxic gene regulation during murine vascular
development. J. Mol. Med., 76: 391-401, 1998.

Semenza, G. L., Roth, P. H., Fang, H. M., and Wang, G. L. Transcriptional
regulation of genes encoding glycolytic enzymes by hypoxia-inducible
factor 1. J. Biol. Chem., 269: 23757-23763, 1994.

Ryan, H. E., Lo, J., and Johnson, R. S. HIF-1 alpha is required for solid
tumor formation and embryonic vascularization. EMBO Journal, 17: 3005-
3015, 1998.

Ryan, H. E., Poloni, M., McNulty, W., Elson, D., Gassmann, M., Arbeit, J.
M., and Johnson, R. S. Hypoxia-inducible factor-1a|pha is a positive factor
in solid tumor growth. Cancer Res, 60: 4010-4015, 2000.

Jiang, B. H., Semenza, G. L., Bauer, C., and Marti, H. H. Hypoxia-
inducible factor 1 levels vary exponentially over a physiologically relevant
range of 02 tension. American Journal of Physiology, 271: C1172-1180,
1996.

Kargul, G. J., Dudekula, D. B., Qian, Y., Lim, M. K., Jaradat, S. A.,
Tanaka, T. 8., Carter, M. G., and Ko, M. S. Verification and initial
annotation of the NIA mouse 15K cDNA clone set. Nat Genet, 28: 17-18.,
2001.

Tanaka, T. S., Jaradat, S. A., Lim, M. K., Kargul, G. J., Wang, X.,

Grahovac, M. J., Pantano, S., Sano, Y., Piao, Y., Nagaraja, R., Doi, H.,
Wood, W. H., 3rd, Becker, K. G., and Ko, M. S. Genome-wide expression

143

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

profiling of mid-gestation placenta and embryo using a 15,000 mouse
developmental cDNA microarray. Proc Natl Acad Sci U S A, 97: 9127-
91 32., 2000. ‘

Kerr, M. K. and Churchill, G. A. Statistical design and the analysis of gene
expression microarray data. Genet Res, 77: 123-128., 2001.

Fielden, M. R., Halgren, R. G., Dere, E., and Zacharewski, T. R. GP3:
GenePix post-processing program for automated analysis of raw
microarray data. Bioinformatics, 18: 771-773., 2002.

Zien, A., Aigner, T., Zimmer, R., and Lengauer, T. Centralization: a new
method for the normalization of gene expression data. Bioinformatics, 17:
S323-331., 2001.

Epstein, A. C., Gleadle, J. M., McNeill, L. A., Hewitson, K. S., O'Rourke,
J., Mole, D. R., Mukherji, M., Metzen, E., Wilson, M. I., Dhanda, A., Tian,
Y. M., Masson, N., Hamilton, D. L., Jaakkola, P., Barstead, R., Hodgkin,
J., Maxwell, P. H., Pugh, C. W., Schofield, C. J., and Ratcliffe, P. J. C.
elegans EGL—9 and mammalian homologs define a family of dioxygenases
that regulate HIF by prolyl hydroxylation. Cell, 107: 43-54., 2001.

Kerr, M. K., Martin, M., and Churchill, G. A. Analysis of variance for gene
expression microarray data. J Comput Biol, 7: 819-837, 2000.

Murphy, B. J., Andrews, G. K., Bittel, D., Discher, D. J., McCue, J., Green,
C. J., Yanovsky, M., Giaccia, A., Sutherland, R. M., Laderoute, K. R., and
Webster, K. A. Activation of metallothionein gene expression by hypoxia
involves metal response elements and metal transcription factor-1. Cancer
Res, 59: 1315-1322, 1999.

Lee, P. J., Jiang, B. H., Chin, B. Y., Iyer, N. V., Alam, J., Semenza, G. L.,
and Choi, A. M. Hypoxia-inducible factor-1 mediates transcriptional

activation of the heme oxygenase-1 gene in response to hypoxia. J Biol
Chem, 272: 5375-5381, 1997.

Dulak, J., Jozkowicz, A., Foresti, R., Kasza, A., Frick, M., Huk, l., Green,
C. J., Pachinger, O., Weidinger, F., and Motterlini, R. Heme oxygenase
activity modulates vascular endothelial growth factor synthesis in vascular
smooth muscle cells. Antioxid Redox Signal, 4: 229-240., 2002.

Yet, S. F., Melo, L. G., Layne, M. D., and Perrella, M. A. Heme oxygenase
1 in regulation of inflammation and oxidative damage. Methods Enzymol,
353: 163-176., 2002.

Jung, Y., lsaacs, J. S., Lee, S., Trepel, J., Liu, Z. G., and Neckers, L.
Hypoxia-inducible factor induction by tumour necrosis factor in norrnoxic

144

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

cells requires receptor-interacting protein-dependent nuclear factor kappa
B activation. Biochem J, 370: 1011-1017., 2003.

Lahat, N., Rahat, M. A., Ballan, M., Weiss-Cerem, L., Engelmayer, M., and
Bitterman, H. Hypoxia reduces CD80 expression on monocytes but
enhances their LPS-stimulated TNF-alpha secretion. J Leukoc Biol, 74:
197-205., 2003.

Tajima, R., Kawaguchi, N., Horino, Y., Takahashi, Y., Toriyama, K., lnou,
K., Torii, S., and Kitagawa, Y. Hypoxic enhancement of type IV collagen
secretion accelerates adipose conversion of 3T3-L1 fibroblasts. Biochim
Biophys Acta, 1540: 179-187., 2001.

Falanga, V., Zhou, L., and Yufit, T. Low oxygen tension stimulates
collagen synthesis and COL1A1 transcription through the action of TGF-
beta1. J Cell Physiol, 191: 42-50., 2002.

Wykoff, C. C., Pugh, C. W., Maxwell, P. H., Harris, A. L., and Ratcliffe, P.
J. Identification of novel hypoxia dependent and independent target genes
of the von Hippel-Lindau (VHL) tumour suppressor by mRNA differential
expression profiling. Oncogene, 19: 6297-6305., 2000.

Poirier, F., Chan, C. T., Timmons, P. M., Robertson, E. J., Evans, M. J.,
and Rigby, P. W. The murine H19 gene is activated during embryonic
stem cell differentiation in vitro and at the time of implantation in the
developing embryo. Development, 1 13: 1105-1114., 1991.

Williams, K. J., Telfer, B. A., Airley, R. E., Peters, H. P., Sheridan, M. R.,
van der Kogel, A. J., Harris, A. L., and Stratford, I. J. A protective role for
HIF-1 in response to redox manipulation and glucose deprivation:
implications for tumorigenesis. Oncogene, 21:282-290., 2002.

Bruick, R. K. Expression of the gene encoding the proapoptotic Nip3
protein is induced by hypoxia. Proc Natl Acad Sci U S A, 97: 9082-9087.,
2000.

Sowter, H. M., Ratcliffe, P. J., Watson, P., Greenberg, A. H., and Harris,
A. L. HIF-1-dependent regulation of hypoxic induction of the cell death
factors BNIP3 and NIX in human tumors. Cancer Res, 61: 6669-6673.,
2001.

Takahashi, Y., Takahashi, S., Shiga, Y., Yoshimi, T., and Miura, T.
Hypoxic induction of prolyl 4-hydroxylase alpha (l) in cultured cells. J Biol
Chem, 275: 14139-14146, 2000.

Berra, E., Richard, D. E., Gothie, E., and Pouyssegur, J. HlF-1-dependent
transcriptional activity is required for oxygen-mediated HIF-1a|pha
degradation. F EBS Lett, 491: 85-90., 2001.

145

32.

33.

34.

35.

36.

37.

38.

39.

40.

Kothapalli, R., Yoder, S. J., Mane, S., and Loughran, T. P., Jr. Microarray
results: how accurate are they? BMC Bioinformatics, 3:22., 2002.

Bruick, R. K. and McKnight, S. L. A conserved family of prolyl-4-
hydroxylases that modify HIF. Science, 294: 1337-1340., 2001 .

Jaakkola, P., Mole, D. R., Tian, Y. M., Wilson, M. |., Gielbert, J., Gaskell,
S. J., Kriegsheim, A., Hebestreit, H. F., Mukherji, M., Schofleld, C. J.,
Maxwell, P. H., Pugh, C. W., and Ratcliffe, P. J. Targeting of HIF-alpha to
the von Hippel-Lindau ubiquitylation complex by OZ-regulated prolyl
hydroxylation. Science, 292: 468-472., 2001.

Hon, W. C., Wilson, M. |., Harlos, K., Claridge, T. D., Schofield, C. J.,
Pugh, C. W., Maxwell, P. H., Ratcliffe, P. J., Stuart, D. l., and Jones, E. Y.
Structural basis for the recognition of hydroxyproline in HIF-1 alpha by
pVHL. Nature, 417: 975-978., 2002.

Jiang, Y., Zhang, W., Kondo, K., cho, J. M., St Martin, T. B., Dufault, M.
R., Madden, S. L., Kaelin, W. G., Jr., and Nacht, M. Gene Expression
Profiling in a Renal Cell Carcinoma Cell Line: Dissecting VHL and
Hypoxia-Dependent Pathways. Mol Cancer Res, 1: 453-462., 2003.

Cioffi, C. L., Qin Liu, X., Kosinski, P. A., Garay, M., and Bowen, B. R.
Differential regulation of HIF-1a|pha prolyl-4-hydroxylase genes by
hypoxia in human cardiovascular cells. Biochem Biophys Res Co, 303:
947-953., 2003.

Lieb, M. E., Menzies, K., Moschella, M. C., Ni, R., and Taubman, M. B.
Mammalian EGLN genes have distinct patterns of mRNA expression and
regulation. Biochem Cell Biol, 80: 421-426, 2002.

Kaelin, W. G., lliopoulos, O., Lonergan, K. M., and Ohh, M. Functions of
the von Hippel-Lindau tumour suppressor protein. Journal of Internal
Medicine, 243: 535-539, 1998.

Metzen, E., Berchner-Pfannschmidt, U., Stengel, P., Marxsen, J. H.,
Stolze, l., Klinger, M., Huang, W. 0., Wotzlaw, C., Hellwig-Burgel, T.,
Jelkmann, W., Acker, H., and Fandrey, J. Intracellular localisation of
human HIF-1 alpha hydroxylases: implications for oxygen sensing. J Cell
Sci, 116: 1319-1326., 2003.

146

CHAPTER 2 APPENDIX

This section contains the supplementary data tables described in the results and
discussion sections of chapter 2.

147

Appendix Table 11:
Genes significantly influenced by cobalt and hypoxia

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

in HIF‘Ia -l- cells
ID Gene Name Abbr
Nuclear (GO = 0005634)
H3014H11 DNA segment, Chr 15, ERATO Doi 417, expressed
H3017F05 lamin A Lmna
H3100G09 requiem Req
H3110H03 Btg3 associated nuclear protein Banp
H3118B02 prothymosin alpha Ptma
H3135G01 myeloid ecotropic viral integration site-related gene 1 Mrfl
Membrane Proteins (GO = 0016021)
H3029C03 podocalyxin-like Podxl
H3035F09 Ras and a-factor-converting enzyme 1 homolog Rce1
H3090D12 mannose—P-dolichol utilization defect 1 Mpdu1
H3112F03 degenerative spermatocyte homolpg (Drosophila) Degs
H3140C05 tumor necrosis factor receptor superfamily Tnfrsf12a
H3151H12 RIKEN cDNA 1110025J15 gene
Enzymes (Various GO)
H3013802 ATPase, H+ transporting, V1 subunit B, isoform 2 Atp6v1 b2
H3021E11 ATPase, H+transportinLV1 subunit A, isoform 1 Atp6v1a1
H3028F05 ornithine decarboxylase, structural OdC
H3042F09 dual specificity phosphatase 19 Dusp19
H3048G11 biliverdin reductase B (flavin reductase (NADPH)) Blvrb
H3095F09 brain-specific ang. lnh. 1-associated protein 2 Baiap2
H3157A06 PET112-like (yeast) Pet112l
H3157D12 microsomal glutathione S-transferase 3 Mgst3
Misc Functions (Various GO)
H3002F06 eukaryotic translation initiation factor 3 Eif3
H3003H10 actinin, alpha 1 Actn1
H3042H12 expressed sequence Al840044
H3078009 oxysterol binding protein-like 2 Osbpl2
H3108H07 cortactin Cttn
H3138G0 vacuolar protein sorting 26 yeast) Vp526
H3151 F07 procollaggn, type VI, alpha 1 Col6a1
Function Unknown
H3008808 RIKEN cDNA 633041 1 E07 gene
H3014E12 signal recogiition particle 72 Srp72
H3038G11 translocase of outer mitochondrial membrane Tomm20
H3043F08 protein phosphatase 1, regulatory subunit 2 Ppp1r2
H3057G09 RIKEN cDNA 6030440G05 gene
H3060C11 spermine oxidase Smox
H3071DOZ kelch domain containigng 2 thdc2
H3094E02 RIKEN cDNA A230106A15 gene
H3098D04 RIKEN cDNA 1300004C11 gene
H3103E10 Rho GTPase activating protein 18 Arhgap18
H3106H11 RIKEN cDNA 2400001 E08 gene
H3110E06 RIKEN cDNA 9130011E154ene
H3139F07 RIKEN cDNA 2410066K11 gene
H3144B07 H19 fetal liver mRNA H19
H3146A07 RIKEN CDNA C130083N04 gene

 

 

 

148

 

Appendix Table 11: (Continued)
Genes significantly influenced by cobalt and hypoxia
in HlF1oI -I- cells
H315GC1 RIKEN cDNA 1700052N19
Unknown
H301
H3014F
H3014G1
H3017CO1
H301

H3049F11
H3051 E02
H305681 1
H3065B07

H3094EO1

H3098801
H3101AO7
H3101A11
H3107B03
H3111BO7
H311
11
15000
154E06
159A07

 

149

Appendix Table 12:

Genes significantly influenced by hypoxia in Wild type and HlF1cr -l- cells

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ID Name Abbr.
H3003H10 Rattus norvegicus non-muscle alpha-actin Actn1
H3016D10 lectin, galactose binding, soluble 3 L als3
H3024804 Mus musculus chaperonin subunit 3 (gamma) Cct3
H3028F03 Mus musculus cell-line LX82 cathepsin L Ctsl
H3030D10 Mus musculus pyruvate kinase, muscle Pkm2
H3030D11 Mus musculus pyruvate kinase, muscle Pkm2
H3045G06 translocated promoter LegioMTPR) NM_003292.1

H3048G11 Similar to biliverdin reductase B Blvrb
H3059H07 clone RP23-20N14 on chromosome 11

H3067C04 Homo sapiens small nuclear ribonucleoprotein Snrpb2
H3088G10 Similar to Absent in melanoma 1 protein Mina
H3094E01 RP23-3209 on chromosome 2

H3103C03 Similar to R13F6.10.p [Caenorhabditis elegans]

H3103G04 ras association (RaIGDS/AF-6) domain family Rassf1
H3116A06 Mus musculus, Clone IMAGE:4024675

H3118801 Mouse mRNA for prothymosin alpha Ptma
H3128C03 RIKEN cDNA 5430419M09 gene L5430419M09Rik)

H3129G12 protease, serine, 15 Prss15
H3137C05 RAB14, member RAS oncogene family Rab14
H3145H03 mitggen activated protein kinase 3 Mapk3
H3146D12 clone RP23-135F6 on chromosome 11

H3147H11 tissue specific transplantation antigen P358 Tsta3
H3148E03 capping protein (actin filament) muscle Z-line, beta Capzb
H3151F07 procollagen, type VI, alpha 1 (Col6a1) Col6a1
H3153803 Mus musculus chromosome 8 clone RP24-324M11

H3157D03 clone RP23-384K10 on Chromosome 2

 

 

 

150

 

Appendix Table 13:
Genes significantly influenced by cobalt treatment in
Wild type and HlF1a -I- cells

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ID Gene Name Abbr.
A-2031 Metallothionein-l (image 1052401) Mt1
H3020E11 M.musculus mRNA fomclin F Ccnf
H3024G12 clone RP23-407M20 on chromosome 11

H3028F05 Mouse kidney ornithine decarboxylase mRN Odc
H3029811 Mouse chromatin nonhistone high mobility nga1
H3049F11 RIKEN cDNA 9930116P15 ge_ne (9930116P15Rik)
H3067E01 eukaryotic translation emetion factor 1 alpha 1 Eef1a1
H3098801 Unknown

H3131A07 Mus musculus peroxisomal/mitochondrial d Ech1
H3132E03 Homo sapiens chromosome 17, clone hRPK.1

H3137C06 Mouse transglutaminase (TGase) mRNA, com Tgm2
H3140C11 Homo sapiens ATPase, H+ transporting, Iy Atp6v0b
H3144807 Mus musculus H19 and muscle-specific th H19
H3151 H12 NMDA receptor glutamate-binding subunit Lag

 

151

 

Appendix Table 14:
Comparison of genes in Table 2 with those presented

in J et al. ref.36
Identical
1A metallothionein L26 L1
H3147H11 L15 L1
elF3 L29 L1
L13 1

enolase L21 1
Chronic
L13 L1
L37 L1
L22 L1
L39 L1
L41 L1
L3 L1
L12 L1

fibronectin Glut1 Glut4
lactate elF2 elF3
GAPDH elF1a elF3
P4H MMP7 P23
elF3 L9 L1

L21 L1

L37 L1

L13 L1

L28 (L17)

L41 (L17)
L22 1

lactate elF1a elF3
fibronectin elF4 elF3
268 elF2 elF3

MMP28 MMP23
MMP25
L13 L1
L28 L1
L20 L1
L15 L1
L18 L1
L7 L1

enolase L7 L1
L30 1
H31116A06 elF2 elF3
elF3 L5 1
L3 1

 

152

Appendix Table 14: (Continued)
Comparison of genes in Table 2 with those presented

in J et al. ref.36

3F Chronic
L22 L1
L3 L1
L12 L1
L13 L1
L37 L1
L39 L1
L21 L1
L41 1

enolase

L Glut1 Glut4

kinase Glut2 Glut4
H3147H11 elF1 elF3
GPI elF2 elF3
elF4 elF3

BNIP3
BNIP3L
GAPDH
aldolase
P4H
ADK2

T

enolase
MUM2
lutaminase 2
268

enolase
mitochondrial carrier
GAPDH

153

27a L1
L12 L1
L29 L1
L30 L1
L9 L1
L10 L1
L30 L1
L27 1

Chronic
L3 L1

L12 L1
L13 L1
L37 L1
L39 1
L21 L1
L41 1

kinase 3 ADK2

 

CHAPTER 3

Vengellur, Ajith., LaPres, John J. 2004. The Role of Hypoxia Inducible Factor
1a|pha in Cobalt Chloride Induced Cell Death in Mouse Embryonic Fibroblasts.
Toxicological Sciences. 822638-46.

154

Abstract

Cobalt has been widely used in the treatment of anemia and as a hypoxia mimic
in cell culture and it is known to activate hypoxic signaling by stabilizing the
hypoxia inducible transcription factor 1a (HIF1a). However, cobalt exposure can
lead to tissue and cellular toxicity. These studies were conducted to determine
the role of HlF1or in mediating cobalt-induced toxicity. Mouse embryonic
fibroblasts (MEFs) that were null for the HlF1or protein were used to show that
HlF1or protein plays a major role in mediating cobalt-induced cytotoxicity.
Previous work from our laboratory and others has shown that two 8H3 domain
containing cell death genes, BNip3 and NIX, are targets of hypoxia signaling.
These experiments document that BNip3 and NlX expression is HIF1or-
dependent, and cobalt induces their expression in a time and dose dependent
manner. In addition, their expression is correlated with an increase in BNIP3 and
NIX protein. Characteristically, the elevated level of BNIP3 was correlated with
an increased presence of Chromatin condensation, one marker for cell injury.
Interestingly, this increased chromosomal condensation was not coupled to
caspase-3 activation as usually seen in a typical apoptotic response. These
results show that HIF1qis playing a major role in mediating cobalt-induced
toxicity in mouse embryonic fibroblasts and may offer a possible mechanism for
the underlying pathology of injuries seen in workers exposed to environmental

contaminants that can influence the hypoxia signaling system, such as cobalt.

155

Introduction

Metals constitute a large percentage of the earth's crust and their biochemical
and geochemical cycles have been drastically altered by human activities.
Metals are stable and persistent environmental contaminants and tend to collect
in soils and sediments. Many of these metals, such as cobalt, nickel and
manganese are used in a wide variety of industrial applications, including the
production of alloys, paints and batteries. Exposure to these metals often occurs
in workers involved in these industrial processes. Environmental exposure has
now become a cause for concern. Metal smelting and other industrial processes
close to agricultural land and the use of bio-solids in the production of fertilizers
has increased the possibility that these metals are entering our food supply in
greater quantities (1). In addition, the recent approval of the gasoline additive,
methylcyclopentadienyl manganese tricarbonyl (MMT), increases the risk of
environmental contamination. Accordingly, understanding the underlying

molecular mechanism(s) of metal-induced toxicity is of increasing importance.

Certain metals are also essential for human health. For example, cobalt plays a
critical role in the synthesis of vitamin 812. In contrast, excessive exposure to
cobalt is associated with several conditions, including asthma, pneumonia, and
hematological abnormalities (2). In addition, nickel, cobalt, cadmium and other
metals are known or suspected carcinogens (3). Despite numerous reports of
metal-induced toxicity, the underlying mechanism remains unclear. Studies in

various systems have shown that exposure to certain metals, such as cobalt,

156

promotes a response similar to hypoxia. Hypoxia is defined as a state when
oxygen tension drops below normal limits and it plays a central role in
development and several pathological conditions including stroke, cardiovascular
disease and tumorigenesis (4). Due to oxygen’s critical role in energy
production, organisms have developed a programmed response to hypoxia that
increases glucose utilization and stimulates erythropoiesis and angiogenesis to
compensate for the decrease in available oxygen (5-8). The hypoxia inducible
factors (HIFs) are a family of transcription factors that mediate the response to
hypoxia by regulating the expression of genes capable of regulating glycolysis,
angiogenesis and erythropoiesis, such as erythropoietin (EPO), vascular

endothelial growth factor (VEGF), pyruvate kinase and many others (9-13).

Prolonged hypoxia can also induce genes involved in cell death (14). Cobalt and
nickel can activate hypoxia-mediated signaling pathways aberrantly under

normoxia by stabilizing the cytosolic hypoxia inducible factor 19 (HIF1a) (15).

For example, cobalt is thought to stabilize HlF1a by inhibiting the prolyl
hydroxylase domain- containing enzymes (PHDs), a family of enzymes that play
a key role in oxygen dependent degradation of the transcription factor (16). The
characterization of the PHD family of enzymes offers a direct link between metal
exposure and HIF mediated signaling. This direct link and the overlap between
gene expression patterns between hypoxia and cobalt exposure have led us to
hypothesize that HlFs may be necessary to the toxic effects of cobalt (17). This

hypothesis was tested using HlF1a -/- cells and several markers of toxicity. Our

157

results demonstrate that HlF1a plays an important role in metal-induced toxicity.
Among the HlF1or dependent genes whose expression was altered by cobalt
treatment are the pro-apoptotic factors, BNIP3 and NIX. Overall, these results
suggest that cobalt-induced toxicity is dependent upon the HlF1a protein and its

ability to induce the expression of cell death promoting genes.

Methods

Materials — Tissue culture media and supplements were obtained from
lnvitrogen, Inc. Cosmic calf serum was obtained from Hyclone. Oligonucleotide
synthesis was performed at the Macromolecular Facility, Michigan State
University, East Lansing, MI. SYBR Green T” real time PCR reagents were
purchased from Applied BiosystemsTM, CA. All other chemicals were reagent

grade and obtained from SigmaTM Chemical Company, MO.

Cell Culture and Toxicity assay — Mouse embryonic fibroblast cell lines were
maintained in modified DMEM media (10% heat inactivated FBS, penicillin-
streptomycin (10 Ulml), non-essential amino acid (10 pg/ml), L-glutamine
(2mM)). The cells were treated with 0, 1, 10, and 100 pM of CoCI2 for 48 hours.
Cells were trypsinized, counted with a hemacytometer and 10,000 cells were
plated in 6 cm cell culture dishes. Cells were counted with a hemacytometer on

1, 2, 3, 4, and 5 days after plating.

158

MTT Assay — Cells were grown in 96 well plates and treated with O, 50, 100,
150 and 200 pM of CoCI2 for 72 hours. The MTT assay was performed by
replacing the cobalt containing media with 100pl of media containing 5mg/ml of
MTT (3-(4,5-dimethylthiazolyl-2)-2, 5-diphenyltetrazolium bromide). The plates
were then incubated for 4 hours (37°C). Medium was then removed by
aspiration and 200pl of solvent (1:1 DMSO:ethanol) was added and the formazan
crystals solubilized with continuous agitation. OD measurements were taken at
550 and 630nm and the difference in OD relative to untreated controls was taken
as a measure of cell viability (18). For the CoCI2 time course experiment cells
were treated with 150 pM CoCI2 and MTT assay was performed on four
consecutive days and values are represented as a percent of time matched

controls within cell types.

Protein Extraction and Westem Blotting — Wild type and HlF1or -/- cells were
grown under control (20% O2), or hypoxic (1% 02) conditions (NAPCO 7000
incubator, NAPCO, VA) or in the presence of 100pM or 150pM CoCI2 and protein
extracts were prepared as described previously (19). Briefly, cells were washed
with cold PBS (4°C) and removed from surface by scraping on cold PBS and
collected by centrifugation. Soluble proteins were extracted with cell lysis buffer
(25 mM HEPES, pH=7.6, 2 mM EDTA, 10 % glycerol, 1 pglmL of aprotinin,
Ieupeptin, pepstatin A and 1 mM PMSF) and three rounds of sonication (5 secs.,
4°C). Insoluble material was removed by centrifugation (16,0009, 1hour).

Protein concentrations were determined using Bio-RadTM Bradford assay kit and

159

BSA standards (20). An equal amount of protein was separated by SDS-PAGE,
and Western blotting was performed with BNIP3 and NIX specific antibody
(Sigma, MO and Exalpha Biologicals, MA respectively) using ECL
chemiluminescent detection kit (Amersham Pharmacia). A B-actin specific
antibody (a generous gift of Dr. John Wang, MSU) was used to verify equal

loading.

RNA Extraction and Reverse Transcription -— RNA extraction was performed
using TriZol reagent (lnvitrogen) via manufacturer's instruction. Briefly, cells
were treated for the specific duration and washed in 1X PBS (4°C). Cells were
removed by scraping in the presence of 1 mL of TriZolTM reagent. Phase
separation was accomplished by addition of chloroform and centrifugation
(16,0009, 15min). RNA was precipitated using isopropanol and was quantitated
spectrophotometrically. 1 pg total RNA was used in subsequent reverse
transcription reaction using SuperscriptTM lI RNase H- Reverse Transcriptase

(lnvitrogen) via manufacturer’s instructions.

Real-Time Quantitative PCR Analysis —— The measurement of BNip3 and NIX
mRNA levels were performed using real-time PCR technology and SYBR® Green
as a detector on an Applied Biosystems Prism 7000 Sequence detection System
(Foster City, CA) (Table. 15). For detailed protocol see Chapter 1 methods

section.

160

22...... 5453 .2 gene

 

 

 

 

 

22v

<ooFFooeo<oeooeoeo 3528050544250 838 5.2 3.8.2..
<HooH<oo<oo<<ooo<o<< Ho<ooto<<o<oeoeoooo 888 .22 8.2m
<o<ooooe<o<ooo<5<t o<<oooo<oe<o<fiooo<< 885 .22 En...
6232... p.536". :o_3eoe< 280

 

 

 

 

 

161

Cell Staining and Caspase Assay - Cells were left untreated or exposed to CoCI2
(150 pM, 48 hrs or staurosporine (1pM, 3 hrs) and stained with Hoechst 33342
dye (1 pg/ml, 15 min.). Cells were viewed and photographed with fluorescent
microscopy. The percentage of apoptotic nuclei were determined by counting
the number of cells displaying a dense nuclear staining. At least 200 cells were
counted in at least 3 separate fields for each treatment and values are displayed

as a percent of total nuclei.

Caspase assays were performed using EnZChek caspase-3 assay kit #2
(Molecular Probes) via the manufacturer’s instructions. Briefly, cells were left
untreated or exposed to CoCI2 (150 pM, 48 hrs),or Staurosporine (1 pM, 3 hrs).
Cell extracts were obtained by scraping the cells in lysis buffer and cleared by
centrifugation. The caspase-3 activity in the supernatant was analyzed
spectrophotometrically (caspase activity in the cell lysate leads to the cleavage of
the non-florescent substrate into a florescent product). The specificity of the
caspase 3 activity was determined by the addition of a caspase-3 inhibitor (AC-

DEVD-CHO Inhibitor).

Statistics - Statistical analysis was performed between treated and untreated or

vehicle controls using t-test (two tailed, unequal variance, p <= 0.05 cut-off).

162

Results

Growth Curve Analysis of wild type and HlF1a -/- cells under CoCI2 treatment:
An equal number of untreated and cobalt exposed (100 pM, 48 hrs) Wild type
and HIF 1a -/- cells were plated onto 6 cm cell culture dishes. The cells in each
plate were counted on each of the subsequent 5 days. Wild types cells showed
little or no growth following CoCI2 (100pM) treatment (Figure 18). In contrast,
HlF1or -/- cells exposed to CoCI2 were initially slow to grow compared to
untreated controls, but had a much higher survival compared to wild type treated
cells. Wild type cells showed a lower rate of cell proliferation and the total cell
viability was lower than HlF1or -/- cells under cobalt treatment even after 5 days.
These results suggest that H|F_1or plays a role in mediating cobalt-induced

growth inhibition.

Cell Toxicity Assay of CoCI2 treated cells using MTT Assay: The growth curves
described in figure 18 suggest that HIF1oL plays an important role in cobalt-
induced growth arrest; however, it does not prove that HlF1a is involved in
cobalt-induced toxicity. To determine the role of HIF1or in the cytotoxicity
induced by CoCI2, wild type and HIF19 -/- cells were treated with varying
concentrations of cobalt (0, 50, 100, 150 and 200pM CoCI2) when cells had
reached approximately 40% confluence. After incubation of 72 hours, cell
viability was evaluated using the MTT assay. The cytotoxic effects of cobalt were
significantly attenuated in the HIF10I -/- cells when compared to the WT cells

(Figure 19A). To determine the time course of cell injury, MTT assay was

163

2.8 00:00: :88 ._.>> 0: 00.888 0.03 2.8 00:00: :88 .2 0E... :02; mod v a

2:80.002 .. 00>. ..8 55.3 003.9 P >00 0: 002.820: 0.02, m0:.0> .280 E03880 :0 002000.
0.03 2:38 :8 80 0:02.02: 5 00:20 E0 0 0: 00.0.0 0.02, 2.8 08.2. 80.500: 8.26:0“. 0:.

we 5.. .5... 2: .38. ~68 e 8820 .o .58 88:5 :0. so; 28 4. e$5.. use Eé 8.: 25>

02:0 5305 .3 059".

.233 8.3.x. 8.2 as:
m w n N _.

IO
N

 

 

O
In

 

 

In
N

 

 

 

. 2:.

 

A

m M 89.32 .0-
i

50.....z
* aeoés .
50.5 I
IILHIIIII

 

 

loom x) iunoo Iiao

 

 

 

 

IDOID
FFF

 

164

“00.0.58 0.03 003.? ...>> 8:0 .\. 0E... 0:: :0Es mod v a 2:009:02 ..

20::8 00:20E 0:... 0: 002.050: 0.02. m0:.0> t..2 .250: om-vm .020 >800 ts. 9.2.. 00.680
8; 2.50., =8 new N.08 2.53 2 88:0 20; 28 .28 0.8.2 -.- eE... use .28 2...;

....>>. 0%: .0..>> ”m .e\eoor o: .00 0.03 00>. =8 55.? 003.05 6:80 5080 .E..2 0:00:80 0 9.0:
8.88 me; 2:50., :00 .232. N: .2 .20. 8.82:: :2 a .21 8m new 2: .2: .8. N68 8
80.208088 030:? 0.. 000090 0.03 2.8 .80: 0.00.8 -\- a. H=1 0:0 .08 02:3 ....>>. 09.. 25> H<
3.0.0.3030 peosufikaoo .0 0.02.2.0 00.500 0:... 6:0 00:300.. 0000 .2. 0.59".

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N

.0..£0E_.::0E.00._._.~A5_1V _000
I a 8. 8. 2: 8 e a
w 2 m
3 [ON 3
m~—- ne—IeN -a m. m.
a lo: a
1:: M .4 18 M
4 I.
[am Ml ._. low M
on mu. co. .0.

4 I :32.

:szl m .250 ouwm
e... 2.1:...

m'
4i

165

performed on the wild type and HIF1 a -/- cells following exposure to 150 uM
CoCI2. As shown in figure 193, CoCI2 exposure resulted in progressive increase
in toxicity to wild type cells compared to HlF1a —l- cells. These results support
the growth curve analysis presented in Figure 18 and suggest that HlF1a plays

an important role in cobalt-induced cytotoxicity.

Quantitative Real-time PCR Analysis of the BCL2 family genes, BNip3 and NIX
genes: Our lab and others have described the regulation by hypoxia of two
members of the BCL2 family, BNip3 and NIX (17, 21-23). Global expression
analysis using cDNA microarrays of cobalt-induced genes in WT and HIF1oL -/-
cells has shown that this regulation is HIF1a dependent (17). The expression of
these genes was evaluated by quantitative real-time PCR (qRT-PCR) using
SYBR® Green as a detector. RNA was extracted from untreated and hypoxia- or
cobalt-exposed WT and HIF1a -/- cells. BNip3 was upregulated 8 fold in a
HlF1oc-dependent manner following cobalt (100 pM) or hypoxia (1% 02)
exposure (Figure 20A). In addition, NIX also showed a HlF1a-dependent
expression pattern and was induced more than 12 fold after cobalt exposure
(Figure 203). These results suggest that cobalt upregulates BNip3 and NIX in a
HlF1a-dependent manner and are consistent with cobalt-induced activation of

HlF1a-regulated cell-death signaling pathways.

Time course analysis of BNip3 mRNA levels under 00C]; or hypoxia treatment:

The time course of BNip3 upregulation by cobalt and hypoxia was also analyzed

166

._o=:00 050500280 0. 0000800 :0Es mod v a 9:000:31 . .930:

vm :0”. NO as; u 9f 0:0 N.000 .21 cor n +N00 00005:: n £0 .0:__ :00 050500980 05 :_ .05. 60:00
05 2 005.088: 002, 0:_0> :00w .:0=00m 00050.: :_ 00000000 00 3.00 0.00.9 0:00 4. 0F=I 0:0 00:00
92; .5: we: 0:; s 000.50 3 80:05 90; 29,0. 8009qu a: 220 .5220 20 90 320
.3920 0:0 320 .8 3% 5:5... .8 2:9".

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4.30.... I
:5 D
EA... 000 .00 a»... .000 .00
3
rl III e ..| IIJ .N m
.I III 0 .I. .l v 0
III, III] m 3
II III «F - 0 m
I.- l - a
an .0 s
. e I: i.
52 .— nezm .. m
8 2

 

 

 

m'
42

167

by qRT—PCR. WT cells were left untreated or exposed to CoCI2 (100 pM) or
hypoxia (1% 02). Total RNA was extracted at various time points (0.25 to 72
hours) after treatment and analyzed for BNip3 and HPRT expression. BNip3
expression levels reached maximum expression at 8 hours following exposure to
CoCI2 (Figure 21A). This expression level remained stable for the duration of the
time course. The BNip3 expression pattern following exposure to hypoxia in the
WT cells gradually increased and reached a maximum after 48 hours of
treatment (Figure 21A). BNip3 was not regulated by cobalt or hypoxia at any

time point tested in the HIF1a -/- cells (data not shown).

Dose response analysis of BNip3 mRNA levels under CoCI2: The regulation of
BNip3 mRNA expression in cells was determined by qRT-PCR following
exposure to 0, 50, 100, 150 and 200 pM CoCI2 for 24 hours. As shown in figure
21B, wild type cells showed marked upregulation of BNip3 mRNA at all doses
tested with a maximum effect at 1500M CoCI2. The HlF1a -l- cells displayed no
significant changes in Bnip3 transcript levels at any dose (Figure 213). The Ct
value and actual transcript levels of BNip3 in WT cells ranged from18.1 to 22.6
and 9.2E05 to 4.5E04 respectively (data not shown). The expression level of the
control gene HPRT was also analyzed and shown not to significantly change

following exposure to any concentration of CoCI2 in either cell type (Figure 21 B).

168

.0050 0505000000 00 00:0:E00 :0:>> 00.0 v : 0:009:01 .. .00:_0> .00:00

02000000: 0: 00 00N_.0::o: 0.05. 05.0000 0.0 0.05. 5000.90 ...m.n=.. 0.00000 02000000: 00 00.0.0000: 0:0
000:00E :. 0050.90 00 000050.00 0:0 0_0>0. :o_000::x0 mgzm .050: NN :00 «.000 .2100N 0:0 00.. .00: .00
5.; 5 090005 :2 .93 8220.. 5:: .98 .003 0220. 0:8 ... 2:: 0:0 .02 .00 5:: .92 92;
00.20 ....>>. 003 0:2, :_ «.00-...10 >0 00~>.0:0 0:02. 0.05. :0.0005x0 .En... 0:0 00.20 Mm .050: 3.0 00 .05.
.00:00 0000005 0:0 00 00~..0:.:0: 003 0:.0> :00m. 05009.0 N.000 .2109 0532.8 050: N5 0:0 0v ..N .0 .N
.0«.o .0 000050 was <20 .23 0003 .«0 $0 0.08%.. .5 .23 00222 .5... 8: .«oo 2 058%.. 052,20.
:0 .05: 0::3 :00. 0000005 00. 0:00 090 0:3 :_ «.00-...10 >0 00.5.05 0:02, 0.05. 5.000508 00.20 ”.4.

0:00 003 0:3 0. 00.500 0E0 008:»: 0:0 «.000 .000: 00.000508 00.2m :0: 300 moan—.10 ._.N 0.50.".

  

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.50.. «.000 .050... E05000 :0t«0 0E. ...
00N 00.. 00.. 00 0 Nu m/NI0
VMNI c

m... ,,..E W..- MI I. m...
04 y n W
. WI ii. .
I .l .l * a I. m m
_.x_._ 3
Ill-III. Eaz==zm 0d 0.“
0 I 320.52 I w . _..I m
.. 5me I 0 m. 080 «Fm.
0n..zm .. 3.20:5 D Eu .. 0n..zm .000 0.2m...

 

 

 

 

 

mi
<i

169

Western Blot Analysis of BNIP3 and NIX levels: Western blot analysis was
performed to determine the levels of BNIP3 and NIX proteins in CoCI2 treated
cells. Wild type and HlF1a -/- cells were treated with cobalt (100 and 150 pM) or
hypoxia (1% 02) for 48 hours and total protein was extracted and separated by
SDS-PAGE. There was a dramatic increase in BNIP3 levels following cobalt and
hypoxia exposure and this increase was dependent upon the presence of HlF1a
(Figure 22, upper panel). The multiple bands represented in the BNIP3 Western
blot are due to progressive proteolysis and has been previously reported to be
indicative of cellular stress (24). NIX protein levels also showed a marginal
increase in CoCI2 and hypoxia treated wild type cells (Figure 22, middle panel).
Both BNIP3 and NIX levels were unchanged in CoCI2 treated HlF,1or -/- cells. B-

actin was used to verify equal protein loading (Figure 22, lower panel).

Cell Morphology: The results suggest that cobalt may be promoting cell death by
inducing HlF1a-mediated upregulation of pro-apoptotic genes. Apoptosis is
correlated with morphological and biochemical changes within the cell, such as
nuclear condensation, DNA fragmentation, and caspase activation. However,
BNIP3 overexpression has been shown to induce a necrosis-like, caspase-
independent cell death (25). To determine if cobalt treated cells were undergoing
classic apoptosis or something similar to BNIP3 induced necrotic cell death,
chromosomal staining and caspase-3 assay were performed in the WT and

HIF10L -/— cells.

170

...0:00 .0300 0:500. .500 at? 00 :000... 5.3 000900. 0:0 0000.00 0.03 002m. $000.50

00.0000 00:00 0.00.Ev x.z 0:0 ..0:00 .0003 00.20 0 5.3 0000.0 0:0 0:0.0E0E 00o.:..0000.: 00
00:00:00 .m.0<n.-wn.w >0 0000:0000 0:0 0000000 002. c.0090 .0000. .050: 00 .0. ..«0 $5 0.x00>: .0
«.000 s... om. .«_0o0 .21 0.: o. 8898 .o .50. 8.09:5 :0. 90; 0.8 -.. 20.: 0:0 Es. 80 25>
8.0 5382. .52. 400.20 .23 2.20 .«« 230.".

 

 

 

 

 

« « o
O 0\o —. 9: car two 0 x; Om —. 2: two
0»: .«oo 08 E... 08 .«00

r I

1‘

4. a 0...... 05

r k
{

171

Cells were left untreated or exposed to cobalt (150 pM), or staurosporine
(0.01uM) for 48 hours and stained with Hoechst33342 nuclear dye.
Staurosporine is a known inducer of apoptosis and was used as a positive
control (26). Cobalt induced a moderate level of chromatin condensation in the

WT cells, while little, if any, was observed in the HlF1a -/- cells. In contrast,
staurosporine treated WT and HlF1or -/- cells showed marked chromatin
condensation (Figure 23A). The percentage of condensed nuclei in each
treatment was also determined. Cobalt treated wild type cells induced a 3 fold
increase in the percentage of condensed nuclei (Figure 233). These
experiments, taken together with the previous gene expression and Western blot
results, suggest that HlF1a-mediated gene activation may be involved in

promoting cobalt-induced cell death.

Caspase-3 Assay: As mentioned previously, caspase-3 activation is another
hallmark of apoptosis and is considered one of the final steps in the process (27).
To determine whether CoCI2 treatment causes activation of caspase-3 in a
HlF1a-dependent manner, WT and HlF1a -/- cells were exposed to cobalt (150
pM) or staurosporine (1 pM) and extracts were prepared (Figure 24A, B).
Caspase-3 activity in the extracts was assessed by fluorimetric analysis in the
presence and absence of a caspase-3 inhibitor (Ac-DEVD-CHO, Molecular
Probes). WT as well as HlF1or -/— cells showed no significant increase in
caspase-3 activity following CoCI2 treatment. This was not due to a general lack

of activity in the assay since the positive control, staurosporine, showed a

172

..0..000 00.000000000 0. 00.00800 0003 mod v 0 0.0009000

.. 0.30... 00 .015. 0000000590 5.15.0 .0 $00. N.000 .21 00. 0. 000098 .0 :00. 00.0000: 00. 0.00
0.00.0. 2.00 4. 0.0.... 000 0.00 0.....5 ES. 00>. 0...... 0. .2030 000000000 .0 000.000.00 ”m 00000000000
0000.200 0.05 .00.0.0000 >00.0...0.00. 0...... 0:00 0.0000 02600 0..0>> 000000.20. 00000.00 00.0: 0>0
00.050 <20 «000000000... 00.00 0020000 002. >00.0...0.0E :00 .0500 00 .0. cm. 0000000580 .20 5.0
.o .000. N.000 .21 8. o. 8890 .o .50. 8.09.5 00. em; 2.8 ..- eE... 9.0 Es. 8.0 2.2. .<

30.00035. :00 .3 0.30.".

173

 

 

 

 

 

 

 

 

 

 

 

 

. .w
00.0.00...
um +~oo .50 c 0 +~OU
o
a u
u D.
a
u
or a
D.
0. W
+3.... I an a. 2.0
E )
.U mu m
.m 4. awn—.1 ._.>> .<

174

0.00 0.00.0. .0..0.00. 0000000 0 .0 0.00 0002. I 00.20. .0..000 .00..0 .0 0000090

00. 0. 00000000 002. >000< $000.0 0.30.0.2. N... .... >0000 0-0000000 x000N0m 00.0: 00.00000.

00.... 3.2.00 00000000 00.000 0 .5... P ..w. 0000000580 .0 .0500 0.. .NO 00. .n.>.... 0080.... 00.000

0.. .2100. .000. 0000 o. .8890 .o ....0. 8.09.5 00. 0.0; 2.8 ..- 3.... .0. 20 E5. 2.... 2.2. 2.
>0000 0.0000000 .00 0.50.“.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5
.000..00.... .00E.00.... n
.0 .00.. 08 E0 .0 0.0.. 000 ...0
...: . m l! m
..n... 8. ..as
I . 8. ... . ...
I 8“ W I 08 W
... ...
I 80 m I 80 M
.a. 0.
. 08 n 4 08 n
02...... u . u
. - 0.3.0.0 I. ....>-< ...
0...... n. 80 m.” 80 M

 

 

 

 

 

 

marked activation of caspase-3 in the WT and HlF1a -/- cells. This activity was
abolished in the presence of the caspase-3 inhibitor confirming the specificity of
the assay (Figure 24). These results suggest that cobalt induced cytotoxicity
involves HlF1a—dependent upregulation of BNIP3 or NIX leading to a caspase-

independent necrosis-like cell death.

Discussion

Hypoxia is a characteristic feature of a number of pathophysiological conditions
such as cancer, stroke, cardiac ischemia etc. (28). Cobalt chloride has been
widely used as a hypoxia mimic in both in vivo and in vitro studies (29).
However, the role of aberrant hypoxic signaling in cobalt-induced toxicity has not
been addressed. Previous work has shown that on a global gene expression
level, both cobalt and hypoxia regulate a similar group of genes (17). The
observed similarity in gene expression appears to be dependent upon a
functional HlF1a protein (17, 30). The experiments described here characterized
the mechanism of action of cobalt chloride-induced cell death and determine the

role of HlF1oc in this process.

Cell viability and proliferation studies show that wild type cells are more
susceptible to cobalt-induced toxicity when compared to HlF1a -/- cells. Given
that HlF1q is a transcription factor, it seemed likely that this toxicity is dependent
upon gene activation. Previous genomic screens and other reports have

identified BNip3 and NlX as target genes of hypoxia signaling (14, 17, 23).

176

Expression of these pro-apoptotic factors was shown to be HlF1o; dependent and
to occur at a dose and in a time frame similar to that of cobalt-induced cell
damage (Figure 19 and 21). These results offer a direct link between the cobalt
exposure, hypoxia signaling and activation of genes involved in cell injury.
BNIP3 and NIX are BH3 domain-containing proteins belonging to the pro-
apoptotic family of genes and their increased expression is correlated with
increased cell death (31). Here we show that these mitochondrial proteins are
not mediating cell death through the classical caspase-3 activation pathway.
There have been earlier reports that increases in BNIP3 lead to caspase
activation in primary cardiac myocytes undergoing hypoxia (22). However, in
other cell types, BNIP3 induces a cell death similar to necrosis, which doesn’t
involve caspase activation (21). In addition, it has been shown that BNIP3
causes a necrosis-like cell death in cells through the mitochondrial permeability
transition pore which involves the loss of mitochondrial potential but without
caspase activation and cytochrome C release (25). This is not surprising, as
there are alternate pathways of apoptosis that do not require caspase activation
such as the AIF pathway (32). One important point is that caspase-dependent
apoptosis is an energy requiring process. At least during hypoxia, ATP levels in
the cell are low due to the inhibition of oxidative phosphorylation. Therefore, it is
possible that the cells initiate an apoptotic process but resort to a necrotic
pathway due to reduced ATP levels. At present it is not clear if ATP levels are
reduced under cobalt treatment however, the morphological study of CoCI2

treated cells show moderate chromatin condensation in the absence of caspase

177

3 activation. Taken together, it seems likely that the HlF1a-dependent increase
in BNIP3 and NIX leads to caspase-independent, necrotic-like cell death similar
to what has been demonstrated in 293T, MCF7, other MEFs and various tumors

(23, 25, 33).

Cobalt and nickel are known to activate hypoxic signaling, and nickel-induced
transformation of fibroblasts requires a functional hypoxia signaling pathway (34,
35). The mechanism of action of CoCI2 mediated stabilization of HlF1a under
normoxia is not completely elucidated. lt isthought to inhibit the iron containing
HIF prolyl hydroxylase enzyme, which plays a critical role in mediating the normal
hypoxic signaling by modifying HlF1a protein and targeting it for degradation.
The chemical characteristics of cobalt also allow it to compete for iron at reactive
sites of various enzymes, rendering these enzymes inactive. The first published
reports of the PHD family of enzymes characterized this inhibition and helped
explain the ability of cobalt to act as a hypoxic mimic (16). Recent reports also
suggest that cobalt exposure may not displace iron at the catalytic site within the
hydroxylase but may sequester the available ascorbate in the cell. Since
ascorbate is necessary for the transition of iron between oxidation states, this

would effectively inhibit PHD activity (36).

The hypoxic signaling pathway is known to activate cell survival genes involved

in glycolysis, angiogenesis and erythropoiesis (8, 37, 38). In addition, in some

cell types, cobalt and hypoxia exposure has been shown to inhibit apoptotic

178

pathways (39). Under chronic hypoxia, however, this pathway also activates
genes involved in cell cycle arrest and death including pro-apoptotic genes (14).
These reports and our current results suggest a complex and at times,
contradictory picture for cobalt induced damage. The protective effects of cobalt
were shown using a very different experimental paradigm and this may explain
the differences in results. Piret et al exposed HepGZ cells to tert-butyl
'hydroperoxide (t-BHP) under serum-free conditions to characterize cobalt’s
inhibitory effects (39). In contrast, we utilized MEFs in the presence of serum
and absence of outside apoptotic stimuli. Treatment time may have also been a
factor since the HepGZ cell’s cobalt exposure was limited to 8 hours in the serum
containing controls. These differences highlight the complexity in metal-induced
toxicity and suggest multiple pathways may be involved. For example, the
observation that cobalt toxicity was only partially attenuated in the HlF1a -/- cells
suggests that cobalt-induced cell death involves HlFla dependent and
independent mechanisms (Figure 18). HlF1a independent mechanisms may be
require other functioning HIFs in the MEFs or the possible disruption of one or
more of the essential enzymes that require iron as a cofactor, which may
ultimately affect cell viability. Also, the effect of oxidative stress due to the
production of reactive oxygen species on various cell processes and integrity of
cellular components such as proteins, DNA and lipid bi-layer cannot be
underestimated. Consistent with this hypothesis, cobalt chloride treatment is
known to induce stress responsive proteins such as metallothionein in wild type

and HIF1a -/- cells (17, 40).

179

In summary, WT and HIFIa -/- cells were treated with cobalt chloride and toxicity
was studied using cell count and MTT assays. Wild type cells showed a marked
decrease in viability as well as cell proliferation compared to HIF1a -/- cells. Wild
type cells showed an increase in the expression of the cell death gene, BNip3
mRNA and protein upon CoCI2 treatment. The cell death and expression of
BNip3 mRNA overlapped in both the time course and dose response studies.
This study shows that cobalt chloride exposure in mouse embryonic fibroblast
leads to a necrosis-like cell death which is dependent on the presence of
functional HIF1a protein. This indicates that the pathology of cobalt-induced
toxicity might be due to the activation of aberrant hypoxic signaling leading to the
increase in the cell death promoting genes such as BNIP3 and NIX levels and

subsequent necrosis.

Acknowledgements

We are grateful to Dr. Randy Johnson and Heather Ryan for providing the wild
type HlF1a -/- mouse embryonic fibroblast cell line. We thank Barbara Woods,
KangAe Lee and Darrell Boverhof for critical reading of the manuscript and Jane
Maddox, Brad Upham and Joseph Frentzel for assistance with cell staining and

caspase assays.

180

References

10.

11.

12.

Adamu, C. A., Bell, P. F., Mulchi, C., and Chaney, R. Residual metal
concentrations in soils and leaf accumulations in tobacco a decade
following Farmland application of municipal sludge. Environ. Poll., 56: 113-
126, 1989.

Lauwerys, R. and Lison, D. Health risks associated with cobalt exposur --
an overview. Sci Total Environ, 150: 1-6, 1994.

Hayes, R. B. The carcinogenicity of metals in humans. Cancer Causes &
Control, 8: 371-385, 1997.

Giaccia, A., Siim, B. G., and Johnson, R. S. HIF-1 as a target for drug
development. Nat Rev Drug Discov, 2: 803-811, 2003.

Bunn, H. F. and Poyton, R. O. Oxygen sensing and molecular adaptation
to hypoxia. Phys. Rev., 76: 839-885, 1996.

Li, H., Ko, H. P., and Whitlock, J. P. Induction of phosphoglycerate kinase
1 gene expression by hypoxia. Roles of Arnt and HlF1alpha. J. Biol.
Chem., 271: 21262-21267, 1996.

Maltepe, E. and Simon, M. C. Oxygen, genes, and development: an
analysis of the role of hypoxic gene regulation during murine vascular
development. J. Mol. Med., 76: 391-401, 1998.

Semenza, G. L., Roth, P. H., Fang, H. M., and Wang, G. L. Transcriptional
regulation of genes encoding glycolytic enzymes by hypoxia-inducible
factor 1. J. Biol. Chem., 269: 23757-23763, 1994.

Wang, G. L. and Semenza, G. L. Purification and characterization of
hypoxia-inducible factor 1. J. Biol. Chem, 270: 1230-1237, 1995.

Sandner, P., Wolf, K., Bergmaier, U., Gess, B., and Kurtz, A. Induction of
VEGF and VEGF receptor gene expression by hypoxia: divergent
regulation in vivo and in vitro. Kid. Int., 51: 448-453, 1997. '

Krieg, M., Marti, H. H., and Plate, K. H. Coexpression of erythropoietin and
vascular endothelial growth factor in nervous system tumors associated
with von Hippel-Lindau tumor suppressor gene loss of function. Blood, 92:
3388-3393, 1998.

Forsythe, J. A., Jiang, B. H., Iyer, N. V., Agani, F., Leung, S. W., Koos, R.
D., and Semenza, G. L. Activation of vascular endothelial growth factor

181

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

gene transcription by hypoxia-inducible factor 1. Mol. Cell. Biol., 16: 4604-
4613, 1996.

Gleadle, J. M. and Ratcliffe, P. J. Induction of hypoxia-inducible factor-1,
erythropoietin, vascular endothelial growth factor, and glucose transporter-
1 by hypoxia: evidence against a regulatory role for Src kinase. Blood, 89:
503-509, 1997.

Bruick, R. K. Expression of the gene encoding the proapoptotic Nip3
protein is induced by hypoxia. Proc Natl Acad Sci U S A, 97: 9082-9087.,
2000.

Ho, V. T. and Bunn, H. F. Effects of transition metals on the expression of
the erythropoietin gene: further evidence that the oxygen sensor is a heme
protein. Biochem. Biophys. Res. Comm., 223: 175-180, 1996.

Epstein, A. C., Gleadle, J. M., McNeill, L. A., Hewitson, K. S., O'Rourke,
J., Mole, D. R., Mukherji, M., Metzen, E., Wilson, M. l., Dhanda, A., Tian,
Y. M., Masson, N., Hamilton, D. L., Jaakkola, P., Barstead, R., Hodgkin,
J., Maxwell, P. H., Pugh, C. W., Schofield, C. J., and Ratcliffe, P. J. C.
elegans EGL-9 and mammalian homologs define a family of dioxygenases
that regulate HIF by prolyl hydroxylation. Cell, 107: 43-54., 2001.

Vengellur, A., Woods, B. G., Ryan, H. E., Johnson, R. S., and LaPres, J.
J. Gene expression profiling of the hypoxia signaling pathway in hypoxia
inducible factor 1 null mouse embryonic fibroblasts. Gene Expression, 11 :
181-197, 2003.

van de Loosdrecht, A. A., Nennie, E., Ossenkoppele, G. J., Beelen, R. H.
J., and Langenhuijsen, M. Cell-mediated cytotoxicity against U-937 cells
by human monocytes and macrophages in a modified colorimetric MTT
assay - a methodological study. J lmmun Meth, 141: 15-22, 1991.

LaPres, J. J., Glover, E., Dunham, E. E., Bunger, M. K., and Bradfield, C.
A. ARA9 Modifies Agonist Signaling through an Increase in Cytosolic Aryl
Hydrocarbon Receptor. J Biol Chem, 275: 6153-6159, 2000.

Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. Protein
measurement with the Folin phenol reagent. J Biol Chem, 193: 265-275,
1951.

Kubasiak, L. A., Hernandez, O. M., Bishopric, N. H., and Webster, K. A.
Hypoxia and acidosis activate cardiac myocyte death through the Bcl-2
family protein BNIP3. Proc Natl Acad Sci USA, 99: 12825-12830, 2002.

Regula, K. M., Ens, K., and Kirshenbaum, L. A. Inducible expression of
BNIP3 provokes mitochondrial defects and hypoxia-mediated cell death of
ventricular myocytes.[see comment]. Circ. Res., 91: 226-231, 2002.

182

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

Sowter, H. M., Ratcliffe, P. J., Watson, P., Greenberg, A. H., and Harris,
A. L. HlF-1-dependent regulation of hypoxic induction of the cell death
factors BNIP3 and NIX in human tumors. Cancer Res, 61: 6669-6673.,
2001.

Chen, 6., Ray, R., Dubik, D., Shi, L., Cizeau, J., Bleackley, R. C., Saxena,
S., Gietz, R. D., and Greenberg, A. H. The E1B 19K/BcI-2-binding protein
Nip3 is a dimeric mitochondrial protein that activates apoptosis. J Exp
Med, 186: 1975-1983, 1997.

Vande Velde, C., Cizeau, J., Dubik, D., Alimonti, J., Brown, T., Israels, S.,
Hakem, R., and Greenberg, A. H. BNIP3 and genetic control of necrosis-

like cell death through the mitochondrial permeability transition pore. Mol

Cell Biol, 20: 5454-5468, 2000.

Tamaoki, T., Nomoto, H., Takahashi, l., Kato, Y., Morimoto, M., and
Tomita, F. Staurosporine, a potent inhibitor of
phospholipid/Ca++dependent protein kinase. Biochem Biophys Res
Commun, 135: 397-402, 1986.

Nicholson, D. W., Ali, A., Thornberry, N. A., Vaillancourt, J. P., Ding, C. K.,
Gallant, M., Gareau, Y., Griffin, P. R., Labelle, M., Lazebnik, Y. A., and et
al. Identification and inhibition of the lCE/CED-3 protease necessary for
mammalian apoptosis. Nature, 376: 37-43, 1995.

Semenza, G. L., Agani, F., Feldser, D., Iyer, N., Kotch, L., Laughner, E.,
and Yu, A. Hypoxia, HIF-1, and the pathophysiology of common human
diseases. Adv Exp Med Biol, 475: 123-130, 2000.

Wang, G. L. and Semenza, G. L. Desferrioxamine induces erythropoietin
gene expression and hypoxia-inducible factor 1 DNA-binding activity:
implications for models of hypoxia signal transduction. Blood, 82: 3610-
3615, 1993.

Salnikow, K., Su, W., Blagosklonny, M. V., and Costa, M. Carcinogenic
metals induce hypoxia-inducible factor-stimulated transcription by reactive
oxygen species-independent mechanism. Cancer Res, 60: 3375-3378,
2000.

Chen, G., Cizeau, J., Vande Velde, C., Park, J. H., Bozek, G., Bolton, J.,
Shi, L., Dubik, D., and Greenberg, A. Nix and Nip3 form a subfamily of
pro-apoptotic mitochondrial proteins. J. Biol. Chem., 274: 7-10, 1999.

Lorenzo, H. K., Susin, S. A., Penninger, J., and Kroemer, G. Apoptosis
inducing factor (AIF): a phylogenetically old, caspase-independent effector
of cell death. Cell Death Differ, 6: 516-524, 1999.

183

33.

34.

35.

36.

37.

38.

39.

40.

Sowter, H. M., Ferguson, M., Pym, C., Watson, P., Fox, S. 8., Han, C.,
and Harris, A. L. Expression of the cell death genes BNip3 and NIX in
ductal carcinoma in situ of the breast; correlation of BNip3 levels with
necrosis and grade. J Pathol, 201: 573-580, 2003.

Salnikow, K., Wang, S., and Costa, M. Induction of activating transcription
factor 1 by nickel and its role as a negative regulator of thrombospondin I
gene expression. Cancer Res, 57: 5060-5066, 1997.

Salnikow, K., Davidson, T., Zhang, 0., Chen, L. C., Su, W., and Costa, M.
The involvement of hypoxia-inducible transcription factor-1-dependent
pathway in nickel carcinogenesis. Cancer Res, 63: 3524-3530, 2003.

Salnikow, K., Donald, S. P., Bruick, R. K., Zhitkovich, A., Phang, J. M.,
and Kasprzak, K. S. Depletion of intracellular ascorbate by the
carcinogenic metals nickel and cobalt results in the induction of hypoxic
stress. J Biol Chem, 2004.

Levy, A. P., Levy, N. S., Wegner, S., and Goldberg, M. A. Transcriptional
regulation of the rat vascular endothelial growth factor gene by hypoxia. J
Biol Chem, 270: 13333-13340, 1995.

Wang, G. L., Jiang, B. H., Rue, E. A., and Semenza, G. L. Hypoxia-
inducible factor 1 is a basic-helix-Ioop-helix-PAS heterodimer regulated by
cellular Oz tension. Proc. Natl. Acad. Sci. USA, 92: 5510-5514, 1995.

Piret, J. P., Lecocq, C., Toffoli, S., Ninane, N., Raes, M., and Michiels, C.

Hypoxia and CoCI2 protect HepG2 cells against serum deprivation- and t—
BHP-induced apoptosis: a possible anti-apoptotic role for HIF-1. Exp Cell

Res, 295: 340-349, 2004.

Murphy, B. J., Andrews, G. K., Bittel, D., Discher, D. J., McCue, J., Green,
C. J., Yanovsky, M., Giaccia, A., Sutherland, R. M., Laderoute, K. R., and
Webster, K. A. Activation of metallothionein gene expression by hypoxia

involves metal response elements and metal transcription factor-1. Cancer
Res, 59: 1315-1322, 1999.

184

CHAPTER 4

THE ROLE OF HIF1a IN METAL-INDUCED TOXICITY IN MOUSE

EMBRYONIC FIBROBLASTS

185

Abstract

Hypoxia inducible factor 10 is a major transcription factor that mediates cellular
response to hypoxia and is regulated by iron-dependent HIF prolyl hydroxylases
(PHDs) which modify and target HIF1oc protein for degradation under normoxia.
Divalent metal ions such as cobalt and nickel and iron chelators such as
desferoxamine can stabilize HlF1or. under normoxia by inhibiting the PHDs.
Previously, it was shown that cobalt’s ability to stabilize HlF1a is directly related
to its toxicity. To determine the role of HlF1oc in the toxicity of other divalent
metals, wild type and HlF1a null MEF cells were assessed for cadmium-induced
damage. Cobalt exposure leads to stabilization of HlF1a and activation of gene
transcription of pro-cell death BNip3 gene, leading to increased cell death in the
wild type cells. In contrast, cadmium treatment does not affect the HIF1
transcriptional signaling but leads to caspase activation and apoptotic cell death,
which is greater in the HlF1a null cells. Cadmium exposure also leads to greater
oxidative stress compared to cobalt exposure, which is again significantly higher
in the HlF1a null cells. Surprisingly, HlF1a null cells have a greater basal level of
reactive oxygen species compared to wild type cells and have reduced mRNA,
protein and activity of Cu/Zn superoxide dismutase enzyme (SOD1). Moreover,
HlF1a null cells had higher total glutathione levels following cobalt and cadmium
exposure compared to wild type cells, pointing to decreased capacity to cope
with oxidative stress. Overall, our results suggest that cobalt-induced cell
damage is promoted through HlF1a stabilization and subsequent transcriptional

activation of cell death genes. Cadmium-induced cytotoxicity is due to ROS

186

generation and the compromised nature of the ROS scavenging system in the

HlF1a -/- cells.

187

Introduction

Metals ions are critical for many processes necessary for cellular homeostasis,
including the function of almost one third of all enzymes (1). The reactive nature
of these metal ions also makes it necessary for cells to maintain a tight regulation
on their concentration and localization and any deviation from this normalcy can
adversely affect life processes by reacting with other cellular components such
as proteins, lipids and nucleic acids (2). Most living organisms require an aerobic
environment for survival and have a well developed signaling system to adapt to
any fluctuations in oxygen availability. Hypoxia inducible factor 10 (HIF1a) is a
major transcription factor that regulates the cellular and physiological response to

decreases in oxygen concentrations (i.e. hypoxia) (see “Introduction” chapter).

Insults that mimic hypoxia such as cobalt or nickel exposure and iron chelators
such as deferoxamine can activate HIF signaling by stabilizing HlF1a (3). In the
present study the effects of exposure to two divalent metals, cobalt and cadmium
were compared in two strains of mouse embryonic fibroblasts (MEFs) to
establish the role of Hl1a in mediating these effects. We show that HlF1a plays
a protective role against cadmium-induced toxicity while promoting cobalt-
induced toxicity. The results also indicate that basal expression of HIF1a under
normoxia plays a crucial role in maintaining the levels of cellular antioxidants

such as SOD1 and SODZ and glutathione levels.

Methods

188

Materials — Tissue culture media and supplements were obtained from
lnvitrogen, lnc. Cosmic calf serum was obtained from Hyclone. Oligonucleotide
synthesis was performed at the Macromolecular Facility, Michigan State
University, East Lansing, MI. SYBR Green T” real time PCR reagents were
purchased from Applied BiosystemsTM, CA. All other chemicals were reagent

grade and obtained from SigmaTM Chemical Company, MO.

Cell Culture and Toxicity assay — Mouse embryonic fibroblast cell lines were
maintained in modified DMEM media (10% heat inactivated FBS, penicillin-
streptomycin (10 Ulml), non-essential amino acid (10 pglml), L-glutamine (2mM)
and Hepes Buffer, pH 7.0 (10mM). Cells were grown in 96 well plates and
treated with 0, 100, 150, and 200 (M of CoCI2 for 72 hours or 0, 1, 2, 5 and 100M
CdCI2 for 72 hours. MTT assays were performed (see chapter 3 methods
section). For figure 27C wild type and HIF1oc -/- cells were treated with CoCI2,
CdCI2 and menadione alone or with N-acetyl cysteine, Tempol or melatonin for

72 hours and MTT assay was performed.

Protein Extraction and Western Blotting — Wild type and HlF1a -/- cells were
grown in normal medium or in the presence of 1500M CoCI2 or 1, 5 or 10pM
CdCI2 or 1500M CoCI2 and SIM CdCI2 and protein extracts were prepared (see
chapter 2 methods). For BNIP3, SOD1, and $002 westerns, total cell lysate
was used. Briefly, cells were washed with cold PBS (4°C) and removed from the

surface by scraping in cold PBS and collected by centrifugation. Soluble proteins

189

were extracted with cell lysis buffer (25 mM HEPES, pH-7.6, 2 mM EDTA, 10 %
glycerol, 1 pglmL of aprotinin, Ieupeptin, pepstatin A and 1 mM PMSF) and
sonication (5 secs., 4°C). Insoluble material was removed by centrifugation
(180009, 20 min.) and protein concentrations were determined using Bio-RadTM
Bradford assay kit and BSA standards (4). An equal amount of protein was

separated by SDS-PAGE, and western blotting was performed with anti-HIF1a

(Novus Biologicals, CO) and BNip3 and B-actin (Sigma) and SOD1 and $002
specific antibodies (Santacruz Biotech) using ECL chemiluminescent detection

kit (Amersham Pharmacia).

RNA Extraction and Reverse Transcription and Real- Time Quantitative PCR
Analysis — RNA extraction was performed using TriZol reagent (lnvitrogen) via
manufacturer’s instruction. The measurement of BNip3, SOD1and SODZ mRNA
levels was performed using real-time PCR, on an Applied Biosystems Prism
7000 Sequence detection System (Foster City, CA). For detailed protocol see
chapter 1 methods section. Primers were designed using the web based
application Primer3 (http:/lwww-genome.wi.mit.edu/cgi-
bin/primer/primer3_www.cgi) biasing towards the 3’ end of the transcript giving a
gene-specific product (Table 16). The relative abundance cf mRNA was

calculated from the Ct values normalized to HPRT.

Creation of stable BNip3 shRNA cells —

190

Table 16. qRT-PCR primers

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Gene Accession Forward Reverse

HPRT NM 013556 AAGCCTAAGATGAGCGC TTACTAGGCAGATGGCC
AAG ACA

BNip3 NM 009760 GGCGTCTGACAAC‘I'I’CC AACACCCAAGGACCATG
ACT CTA

SOD1 NM_011434.1 GAGACCTGGGCAATGTG TTG'ITI'CTCATGGACCAC
ACT CA

SOD2 NM_013671.3 AACTCAGGTCGCTCTTC GCTTGATAGCCTCCAGC
AGC AAC

Table 17. shRNA Sequences

Gene Accession Sequence

shCTRL TGCGTCTTGTTCATCTCCT

shBNip3 #1 NM 009760 GGCAGCCTGCGCCGCTCAGC

shBNip3 #2 NM 009760 TGCCA‘ITGCTGAAGTGCAGC

 

 

 

 

191

 

 

Target sequences were obtained using Ambion’s siRNA target finder
(http:llwww.ambion.com/techlib/misc/siRNA_finder.html) and were analyzed by
BLAST to determine specificity (Table 17). Oligos were designed and PCR was
performed to create a shRNA cassette carrying U6 promoter sequence, sense-
Ioop-antisense sequence and terminator sequence. The shRNA cassettes were
cloned into a lentivial vector and packaged using 293T cells via transient
transfection with helper plasmids to generate infectious virus (obtained form
University of California, San Diego). GFP construct was used as a control. Wild
type cells were plated at the density of 2 x 104 cells per 6 cm issue culture dish
and were infected with the virus overnight. Cells were split and individual
colonies were isolated and Puromycin-selected to create the stable shRNA
expressing cells. Scrambled target sequence with no similarity to any known

gene was also cloned and infected to use as shControl.

Cell Staining and Caspase Assay - Cells were left untreated or treated with 5 uM
CdCI2 for 24 hours and stained with Hoechst33342 dye (1 uglml, 15 min.). Cells
were viewed and photographed with fluorescent microscopy. Caspase assays
were performed using EnZChek caspase-3 assay kit (Molecular Probes) using
the manufacturer’s instructions. Briefly, cells were left untreated or exposed to
CdCI2 (5 pM, 24 hours,,or Staurosporine (1 pM, 4 hrs). Cell extracts were
obtained using the lysis buffer provided and the caspase-3 activity in the

supernatant was analyzed spectrophotometrically.

192

Determination of ROS levels — Cells were left untreated or treated with 2009M
CoCI2 or 5uM CdCI2 for 24 hours or 1mM H202 for 1 hour. Cell were trypsinized,
pelleted, and treated with 59M CM-HzDCFDA (Molecular Probes, OR) for 30
minutes in serum free media. Cells were washed with PBS and re-suspended in
PBS supplemented with 10% cosmic calf serum and analyzed on a BD
FACSDiva® flow cytometer. Fluorescence intensity (490nm excitation and
520nm emission) of 1 x 104 cells for each samples were measured and

experiments were performed in duplicate.

Determination of superoxide dismutase activity — Total cellular superoxide
dismutase activity was measured using superoxide dismutase assay (Cayman
Chemicals, Ml) as per manufacturer’s description. Briefly, treated cells were
harvested, pelleted, and lysed by sonication in buffer containing 20mM Hepes
(pH = 7.2), 1mM EGTA, 210mM mannitol and 70mM sucrose. Supernatant was
collected after centrifuging at 15009 for 5 minutes, and an aliquot was used for
performing the superoxide dismutase assay and enzyme activity was quantified
using a standard curve. Mn-SOD levels in the sample were determined using 3

mM KCN to inhibit the activity of Cu/Zn-SOD activity.

Determination of Glutathione levels - Total glutathione levels were determined
using spectrophotometric assay (Cayman chemicals, Ann Arbor, MI) as per
manufacturer’s instructions. Briefly, treated cells were harvested by scraping in

ice cold PBS and pelleted by centrifugation (10009, 5 minutes). Cells were

193

sonicated in 1X MES buffer and insoluble material was removed by centrifugation
(50009, 5 minutes). Supernatant was deproteinated by the addition of an equal
volume of 10% metaphosphoric acid and neutralized by adding 4M
triethanolamine. Glutathione concentration was determined by performing a
coupled kinetic assay. In this assay, glutathione is used by glutathione reductase
to convert a colorless tetrazolium salt into formazan which was measured

spectophotometrically. The results were normalized to total protein levels.

Statistics - Statistical analysis was performed between treated and untreated

samples using t-test (two tailed, unequal variance, p <= 0.05).

Results

Cadmium and Cobalt Toxicity to MEFs - Wild type and HIF1a -/- mouse
embryonic fibroblast (MEFs) cells were plated in 96 well plates and treated with
50, 100, 150 or 200uM CoCI2 (Figure 25A) and 1, 2, 5 or 109M CdCI2 (Figure
258) for 72 hours. Cell viability, determined by MTT assay, was markedly
reduced in both wild type and HlF1a -/- cell lines under cobalt and cadmium
treatment in a dose dependent manner. However, cobalt-induced cytotoxicity
was markedly higher in the wild type cells compared to HlF1a -/- cells as we
have reported earlier (5). In contrast, cadmium chloride treatment caused
significantly greater cytotoxicity in the HIF1a -/- cells compared to the wild type
cells. The results indicate that HIF1a might be playing different roles in cell

survival following cobalt and cadmium exposure.

194

Cell Viability

Cell Viability

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.5 ' U WT 7
1.2 - '- HlF1a 4. I
0.9 -
*
0.6 4
*
0.3 - FL- r""—_,
0.0 ‘— I I l
0 100 150 200
Cobalt chloride (pM)
1.2 - El
WT
1'0 d II HlF1a 4. J
0.8 . * ,,
0.6 -
*
0.4 -
0.2 -
0.0 -—
0 1 2 5 10
Cadmium Chloride (pM)

Figure 25. Cytotoxicity of CoCl2 and CdCl2 to MEFs

A. Wild type (WT, white bars) and HlF1a -I- (black bars) cells were left
untreated (O) or exposed to 100,150 or 2000M CoCI2 (100, 150, 200) for
72 hrs. Cell viability was assessed using a standard MTT assay.

Control values within cell type were set to 1. *, p<0.05. B. Wild type
(WT, white bars) and HlF1a -l- (black bars) cells were left untreated (0)
or exposed to 1, 2, 5, 10 9M CdCl2 (1, 2, 5, 10) for 72 hrs. Cell viability
was assayed using a standard MTT assay. Control values within cell
type were set to 1. . *, p<0.05.

195

Previously, we reported that CoCI2—induced cytotoxcity correlated with an
increase in nuclear condensation; however, this condensation occurred in the
absence of caspase-3 activation (5). To expand these findings, WT and HIF1pi -
/- MEFS were treated with CdCI2, and nuclear morphology and caspase-3
activation was assessed. Wild type and HlF1a -/- cells showed significant levels
of nuclear condensation characteristic of apoptosis following metal exposure,
however, the null cells were more sensitive (Figure 26A). As seen in Figure 263
HIF1a -/- cells displayed a greater caspase-3 activation following cadmium
exposure than the WT cells. This bias was not evident in the positive control,

staurosporine treated samples (Figure. 26B).

HlF 1a protein levels under cobalt and cadmium treatments — Given that HIF1 is
primarily regulated at the level of protein stability, the ability of both metals to
stabilize HlF1a was tested. Wild type cells were treated with various
concentrations of metals, and HlF1q protein levels were determined by western
blot analysis (Figure 27A). HlF1a -/- cell extract treated with 1500M CoCI2 was
used as a negative control. HlF1a protein levels were detected only in the wild
type cells under CoCI2 treatment. There were some non-specific bands present
in all the lanes. In contrast to published reports in 293 cells exposed to hypoxia,
cadmium co-treatment did not significantly reduce the HIF1a protein levels in
cobalt treated MEFs (Figure 27A) (6). These results suggest that the ability of
WT to be refractory to cadmium-induced cytotoxicity is not mediated by metal-

induced HlF1oi stabilization.

196

HIF1a -I-

 

WT
contml -
CdCI2 5PM -

   

9°
0
4|

 

 

 

 

 

 

I] wr Caspase-3 Activity
I HIT-'10 .1-
6.0 - *
8a
5 *
5 4 o -
g *
.2
2.0 - * I
0.0 -
CTRL CdCl2 (5 pM) Staurosporine

(1 HM)

Figure 26. Characterization of Cadmium-induced Cell Death.

A. Wild type (WT) and HlF1a -/- cells were left untreated (Control) or
exposed to 5 pM CdCI2 for 24 hours and nuclear morphology was
observed after staining with Hoechst 33342 dye using fluorescence
microscopy. B. Wild type (WT, white bars) and HlF1a -/- (black bars)
cells were left untreated (CTRL) or exposed to CdCl2 (CdCI2. 59M, 24
hours), or staurosporine (1 pM, 4 hours). Caspase-3 activity was
measured using EnZChek caspase-3 assay kit #2 (Molecular Probes). *,
p<0.05, n=4

197

Figure 27. Hypoxia signaling and cytotoxicity to CoCl2 and
CdCI2

A: Wild type (WT) cells were left untreated (CTRL) or exposed to
150 uM CoCI2, 1, 5 or 10 pM CdCI2 or 150 uM CoCl2 and 10 9M
CdCI2 for 24 hours. HlF1a -/- cells extract treated with 1509M CoCl2
was used as a negative control. Nuclear protein was extracted and
separated by SDS-PAGE, transferred to nitrocellulose membrane
and probed with a HlF1a (upper panel) or B-actin (lower panel)
specific antibody. B, C. BNip3 mRNA expression levels were
analyzed by qRT-PCR (SYBR Green) in wild type (WT, white bars)
and HlF1a -/- cells (black bars) as described in methods section.
Cells were left untreated (0), 1509M CoCl2 (B, 150) or 59M CdCI2
(C, 5) for 24 hours. Each value was normalized to the control level
in the corresponding cell line. *, p<0.05 compared to control with in
the cell type, n=4 D. BNip3 protein levels were determined in wild
type (1), HlF1a -/- (2), shControl (3), shBNip3 #1 (4), and shBNip3
#2 (5) after treatment with 1500M CoCI2 or 59M CdCl2 for 24 hours
using a BNip3 specific antibody and B-actin was used as a loading
control (Lower Panel). E. Wild type (WT), HlF1a -/-, shControl
(sthrl), shBNip3 #1 and shBNip3 #2 cells were exposed to 1509M
CoCl2 (Coz’z white bars) or 50M CdCI2 (Cd’f, black bars) for 72
hrs. Cell viability was assayed using a standard MTT assay.
Control values within cell type were set to 1. *, significant compared
to wild type treatments, p<0.05, n=4.

198

Fold Change

5.0 -
4.0 «-
3.0 -
2.0 «
1.0 d

HIF1a

lidlcthn

 

0.0 cl

HIF1a -I-
150pM
00"

 

 

1500M
Co”

 

101
cu»

5pm 10p." Com
Cd1*__ ca» ca:-

  

 

 

 

 

 

...

900': (IN)

 

1.5-

1.0-

Fold Change

 

0.0 J

199

 

I””"|

I HlFla .r

] it

 

 

 

099': (W)

Figure 27. Hypoxia signaling and cytotoxicity to CoCI2 and
CdCl2 (continued)

 

 

 

 

 

 

 

(3202" 3150 M Cd—I‘ILL)_+ 5 M
D. 2 3 4 5
BNip3 - .
B-Actin
E.
* D CO”
06 ' I w
*
0.5- *
£- o.4 -
‘3
S 0.3 '
3 0.2 -
0.1- *
OJ I I I
WT HlF1a -r- sthrl shBNip3 shBNip3
#1
Wild Type

200

BNip3 mRNA and protein eXpression following cobalt and cadmium treatments -
Previously our group and others have shown that the mRNA and protein levels of
BNip3, a BH3 domain containing, cell death promoting factor, are increased
under cobalt chloride treatment in a HlF1oc dependent manner (5, 7). In contrast,
CdCI2 treatment caused a reduction in the BNip3 mRNA levels in wild type and
HlF1a -/- cells which was not statistically significant (Figure 278 and 270). In
addition, cadmium was unable to induce BNIP3 protein levels in either cell type,
while cobalt treated cells led to an increase in BNIP3 protein in WT cells (Figure
27D). These results confirm our earlier observations and suggest that cadmium-

induced cell damage is not mediated by BNIP3 (5, 7, 8).

To further characterize the role of BNIP3 in cobalt-induced toxicity, we created
several cell strains with reduced BNIP3 expression levels using RNAi. WT cells
were infected with a lentiviral vector expressing one of two independent shRNA
constructs that specifically targeted BNip3 or a scrambled shRNA sequence as a
negative control. Following selection of stable cell strains, BNIP3 protein levels
were assessed by western blot analysis. The two independent BNip3 shRNAs
showed greater than 80% reduction in the BNIP3 protein levels (Figure 270).
The cell lines were then treated with 1509M CoCI2 for 72 hours and an MTT
assay was performed to determine the role of BNip3 in cobalt-induced
cytotoxicity. Both BNip3 shRNA expressing cell strains had reduced cobalt-
induced cell death compared to the scrambled control. In fact, the results were

comparable to HlF1a -/- treated cells (Figure 27E). Moreover, knock-down of

201

BNIP3 levels in wild type cells did not influence the cell viability upon cadmium
treatment significantly (Figure 27E). These results suggest that HlF1a-induced
BNIP3 protein expression plays a role in cobalt chloride induced toxicity but not

in cadmium-induced toxicity in MEFs.

Oxidative stress under cobalt and cadmium treatment — Since HlFla activation
and genes such as Bnip3 do not appear to play a role in cadmium induced
toxicity, oxidative stress was characterized in the two cell lines. Cadmium and
cobalt are known to produce reactive oxygen species (ROS) in various cell types
(9, 10). To begin to characterize the role of oxidative stress in metal-induced cell
death in the MEFs, the ROS levels were measured in WT and HlF1a -/- cells
following cobalt, cadmium, and hydrogen peroxide exposure. Cells were left
untreated or treated with 150pM CoCI2, 5uM CdCI2 or 1mM H202 for 24 hours,
and ROS were measured using flow cytometry and CH-HzDCFDA dye.
Cadmium treatment caused a significant increase in ROS only in the HlF1a -/-
cells while cobalt was unable to alter the ROS profile in either cell type (Figure
28A). The lack of response in the WT cells from either metal was not due to any
cell-specific mechanisms, as the positive control, H202 treated WT displayed a
marked increase in ROS generation. Interestingly, the HlF1a -/- cells have a
higher basal level of ROS compared to wild type cells (Figure 288). These
results suggest a possible mechanism for the specificity of cadmium-induced
toxicity seen in the HIF1a -/- cells and indicate that HlFta -/- cells are

experiencing an increased ROS load under basal conditions.

202

Figure 28. Oxidative stress in CoCI2 and CdCI2 mediated cytotoxicity.
A. Wild type (WT) and HIF1a -/- cells were left untreated (CTRL), or
exposed to 1mM H202 (H202, 2hours), ZOOpM CoCl2 (002*, 24 hours), or
5p.M CdCl2 (Cd2*, 24 hours). Reactive oxygen species generated in the cell
were measured using the ROS sensitive dye CH-HZDCF DA and flow
cytometry. B. Comparison of the levels of ROS generated in wild type (WT)
and HlF1oc -/- cells untreated. C. Wild type and HlF1a -/- cells were left
treated with 150pM CoCI2, (002+), 150pM CoCI2 , 5pM CdCl2 (Cd2+) or'
20pM menadione (MEN) alone or with 10mM N-acetyl cysteine (NAC),
1mM Tempol (TMP) or 1mM melatonin (MLT) for 72 hours and cell viability
was measured using MTT assay. Control values within cell type were set
to 1.*, significant compared to single treatment, p<0.05, n=4

203

 

CTRL

H202(1mM)

Coz“ (ZOOuM)

Cd2+(5p.M)

204

Numb-
-

Numb-r
4

Mun kw
..

HIF10L -I-

 

Figure 28. Oxidative stress in CoCI2 and CdCl2 mediated
cytotoxicity (continued)

U C O I l

   

HlF1a -I-

H-
a I n in I m-

205

Figure 28.
cytotoxicity (continued)

Oxidative stress in CoCI2 and CdCl2 mediated

 

 

 

 

 

 

 

C- N-Acetyl Cysteine (NAC) D V"
I HIF1a 4-
0.8 * *
1.2 *
*
5‘ 0.6 > o_9
= .= *
a 5
g 0-4 g 0.6
8 0.2 '1' III IL 3 0.3
0 . . v . . . v 0
MAC Coz' Com can cum MEN MEN+ .00“ Cd“Cd"+ MEN+
NAC "Ac NAC "Ac °°' NAC NAC MEN NAC
Melatonin (MLT)
0.4 0.8
g, 0.3 * E 0.0
35:? E
s 0.2 5 0.4
a =
o 0.1 m n '1] 8 0.2 *
o. . . . . .fi . n. o
, Com cam MEN+ Com cue cam MEN+
M” °°z MLT Cd" MLT ME" MLT ”LT °°” MLT MLT ME" MLT
Tempol(TMP)
0.4 0.0
E 0'3 Q 0.6
a =
g 0.2 g 0.4
8 0.1 n 3 0.2
a ll] ' Fl . fi . r‘I _ n. 0
Com , odm MEN+ TMP Co" Com cue cum MEN MEN+
TMP cw TMP Cd: TMP ME" TMP TMP TMP TMP

206

Protection against toxicity by antioxidants - Cadmium is known to create
oxidative stress by sequestering the cellular glutathione pool through interactions
with their sulfhydryl groups and effectively preventing their ability of to scavenge
ROS. Cobalt is a well known Fenton metal that generates hydroxyl radicals by
interacting with H202 and superoxide (11, 12). N-acetyl cysteine (NAC), a known
antioxidant, acts as a precursor for glutathione and can protect cells against
oxidative injury caused by metals and other agents (13). Wild type and HlF1a -/-
cells were treated with CoCI2, or CdCI2 in the presence and absence of NAC for
72 hours, and MTT assays were performed to assess cell viability. The results
show that NAC prevented metal-induced cell damage in both cell types
suggesting that ROS-mediated cobalt and cadmium-induced cytotoxicity. NAC
also protected against menadione, a redox cycling agent known to generate
oxidative stress in cells (Figure 280) (14). . Among the other antioxidants tested,
only melatonin at 1mM concentration showed any protection against cadmium
exposure in wild type and HIF10. -/- cells, and this effect was specific to
cadmium-induced toxicity. Tempol, an SOD mimetic did not protect against
either cobalt or cadmium exposure in either of the cells. Moreover, at 1mM
concentration and even at 2500M concentration tempol reduced the cell viability
in the control cells (Figure 28C, data not shown). These results suggest that
oxidative stress plays a greater role in cadmium-induced toxicity compared to

cobalt-induced toxicity.

207

Superoxide dismutase levels in CoCI2 and CdCI2 treated cells - Mammalian cells
generate ROS as a result of several different metabolic processes and have
evolved a variety of mechanisms for the efficient removal of these harmful
species (15). For example, superoxide is generated as a by—product of ETC
reactions due to incomplete electron transfer and is rapidly removed from the
cellular space by superoxide dismutases (SODs), which convert the superoxide
into oxygen and hydrogen peroxides (16). To determine if cobalt or cadmium
altered SOD expression, the mRNA levels of SOD1 and SOD2 were analyzed by
qRT-PCR. In agreement with our previously published report, SOD1 and SOD2
mRNA levels were significantly lower in the untreated HlF1a -/- cells compared to
WT cells (Figure 29A and B). CoCI2 treatment did not change the expression of
SOD1 and SOD2 in wild type or HlF1oc —/- cells. Although not statistically
significant, CdCI2 treatment caused a marginal reduction in SOD1 and SOD2
mRNA levels in HlF1a -/- cells. To determine if these changes in mRNA
expression correlated with changes in SOD protein levels, SOD1 and SOD2
were analyzed by western blot analysis. In agreement with the mRNA data,
SOD1 protein levels were lower in the HlF1a -/- control cells compared to wild
type cells (Figure 29A). Metal exposure did not alter the levels of SOD1 in the
wild type whereas cobalt exposure lead to increase in SOD1 protein levels in
HlF1a -/- cells (Figure 290). In contrast, SODZ protein levels were higher in the
HlF1a -/- cells compared to WT cells and CdCI2 treatment caused a reduction in
SOD2 protein levels. Finally, total SOD and SOD2 activity were determined and

SOD1 activity was deduced by subtracting the latter from the former value. The

208

Figure 29. Expression and activity of superoxide dismutases

Transcript levels of SOD1 (A) and SOD2 (B) were determined using qRT-
PCR in wild type (WT, white bars) and HlF1a -/- cells (black bars) as
described in methods section. Each value was normalized to the control level
in the corresponding cell line. Cells were left untreated (CTRL) or treated
with 150pM CoCl2 (Co’*) or 5pM CdCl2 (Cd’f) for 24 hours. C. Protein
levels of both SOD1 (Top Upper panel) and $002 (Bottom Upper panel)
were determined by Western blotting in wild type (WT) and HlF1a -/- cells. [3-
actin levels (Top and Bottom lower panels) were used a loading control.
Cells were treated as in A and B. (D) SOD1 enzyme activity in wild type (WT,
white bars) and HIF1a -/- (black bars) cells were determined as described
in the methods. Cells were treated as in A and B. *, significant compared to
untreated cells within the cell type, p<0.05 n=4. #, significant compared to
wild type cells, p<0.05, n=4.

209

A.

Relative Expression

SOD1

9.0 1

6.0 4

 

 

 

 

 

 

 

 

 

 

     

CTRL Cc," Cd”

WT HIF1a -I-
CTRL Co” Cd” CTRL Co” Cd”

8001 l— ”‘
s-Actin :' - ~ .
_ P-j

WT HIF1a -I-
CTRL Co” Cd” CTRL (3oat Cd”

3'

5°02 Fifi-E." -‘

B-Actin

 

 

 

 

 

 

I“

I HIF1a. -l-

 

 

Relative Activity

 

 

 

 

 

 

CTRL Co2+ Cd"

210

results show that SOD1 activity is consistently lower in all HlF1a -/- cell
treatments compared to WI' cells (Figure 29D). These results suggest that
HlF1a -/- cells have a reduced capacity to cope with superoxide stress compared
to their WT counterparts and that cadmium exposure alters the SOD expression

profile and activity in these null cells.

Glutathione concentration in cells treated with CoCI2 and CdCI2 - Glutathione
plays a critical role in various detoxification reactions within the cell by acting as a
reducing agent (17, 18). Cadmium is known to deplete the cellular glutathione
pool by covalently binding to its sulfhydryl groups and compromises the cell’s
ability to protect against oxidative stress (12). To determine if metal exposure
can alter the glutathione pool within the WT or HlF1a -/- cells, total cellular
glutathione was determined by a spectrophotometric assay (19). In contrast to
the SOD activity, HlF1oc -/- cells showed higher total glutathione levels under
control conditions compared to WT cells (Figure 30). Both cell types showed an
increase in total glutathione following metal exposure and the absolute levels

remained higher in the HlF1a -/- cells under these conditions. These results

suggest that HlF1a modulates a cell’s ability to regulate cellular glutathione

Discussion
Divalent metal ions play a critical role in various cellular processes in all living
things (1). Metals such as iron, manganese, magnesium, zinc, copper and cobalt

are essential elements because of their critical role in the activity of various

211

D m
I HlF‘la 4-

 

 

 

 

 

 

 

 

0.8 - *
a:
L” 0.6 - *
2
3
0 0.4 "
.2
E
g 0.2 ~
0 - . .
CTRL' Co”- Cd2+ -

Figure 30. Cellular Glutathione levels under CoCl2 and CdCl2
treatments

A. Total cellular glutathione levels in wild type (WT, white bars) and
HlF1a -/- (black bars) cells were determined as described in the
methods. Cells were left untreated (CTRL), with 1500M CoCl2 (Coz*) or
50M CdCl2 (Cd2“) for 24 hours. *, significant compared to respective
controls, p<0.05, n=4. #, significant compared to wild type cells, p<0.05,

=4.

212

enzymes and metalloproteins (20). However, overabundance of metals can lead
to imbalances in cellular processes due to their ability to form intra-molecular
interactions and generate reactive intermediates (21). These cellular imbalances
can lead to downstream effects and compromise a cell’s ability to cope with

oxidative stress. Hypoxia is a form of oxidative stress and the cellular response

to decreases in oxygen availability are predominantly regulated by HlF1a.

In this study we investigated the effect of two divalent metals, cobalt and
cadmium, on their ability to interfere with the HlF1oc pathway. The results
demonstrate that HlF1a -/- cells are more sensitive to cadmium-induced toxicity.
This is in contrast to the cobalt-induced toxicity, which is more toxic to wild type
MEFs. Presumably, this difference is due, in part, to the two metal’s ability to
activate of HlF1a pathway aberrantly (5). Cobalt is capable of stabilizing HlF1a
and promotes HIF1-mediated signaling; however, cadmium treatment does not
cause HlF1a stabilization and was unable to drive the expression of hypoxia-
regulated genes. It was reported that cadmium treatment can cause a reduction
in HlF1a protein stabilized under hypoxia in 293T cells (6). This is in contrast to
what was observed in MEFs, as cadmium was unable to inhibit CoClz-induced
HlF1a stabilization. This difference is most likely due to cell specific factors and

the differences in “hypoxia” treatment.

The aberrant activation of HIF10L signaling by cobalt leads to increased

expression of various adaptive and cell death promoting factors in the wild type

213

cells which are absent under cadmium exposure. For example, the pro-cell
death factor BNip3 was induced only in wild type cells following cobalt exposure
and this results in a caspase-independent necrotic cell death (5, 22). Under
cadmium exposure, wild type and HlF1a -/- cells undergo apoptosis
characterized by caspase activation and chromatin condensation but
independent of BNip3 expression. Interestingly, cadmium-induced changes in
caspase-3 activity and nuclear condensation were more pronounced in the
HlF1a -/- cells. The experiments using shBNip3 clones showed that BNip3
activation is a major factor in the increased cell death under cobalt treatment in
wild type cells. In contrast, BNip3 does not play any role in mediating cadmium
induced cell death (Figure 27E). However, cadmium has been shown to
modulate the expression of pro-apoptotic bax and cell cycle regulator p53
proteins in a variety of cell lines including alveolar type 2 cells promoting
apoptotic cell death. The role these proteins play in MEFs is currently under

investigation (23).

Another important characteristic of cadmium exposure was increased ROS levels
in the HlF1a -/- cells compared to wild type cells. Redox active metals such as
iron, cobalt and nickel are thought to produce ROS through Fenton-like reactions,
while redox inactive metals such as mercury and cadmium are thought to deplete
cellular sufhydryl antioxidant reserves and interfere with cellular antioxidant
enzymes such as glutathione transferases and glutathione peroxidases, leading

to oxidative stress (12, 24). Although cobalt is a more direct ROS producer

214

compared to cadmium, under these experiments there was much higher ROS
levels following cadmium treatment compared to wild type cells. This is
consistent with earlier reports that suggests the need for a much higher dose of
CoCI2 to generate any significant ROS generation (25). The most intriguing
observation was that HlF1a -l- cells had a much higher level of ROS in the
untreated cells compared to the wild type cells (Figure 288). Similar results were
observed by another group in MEF cells where they found increased production
of H202 in HlF1a -l— cells compared to wild type cells (26). Their data suggests
that HlF1a -/- cells have lower levels of pyruvate dehydrogenase kinase1
(PDK1), a key inhibitory enzyme for the pyruvate dehydrogenase complex. This
complex regulates the ability of pyruvate to enter the mitochondria, and the
authors argue that increased carbon flux through the tricarboxylic acid cycle
(TCA cycle) and subsequent electron flow through ETC in HlF1a -/- cells leads to
increased oxidative stress. Our data also suggest that the most effective method
to counter cadmium toxicity is to increase the glutathione reserves by treating
with N-acetyl cysteine (NAC), a precursor for glutathione. This is consistent with
many other studies (11, 13). NAC also marginally protected against CoClz
mediated toxicity in both cell types. NAC is a well known cobalt chelator and the
protection it offers against cobalt may primarily be due to the effective decrease

in functional cobalt within the cell (27).

Oxidative stress can be a result of reduced elimination of reactive species. We

have earlier reported that HIF1a -/- cells have a lower level of SOD1 and SODZ

215

mRNA compared to wild type cells (28)( Figure 29A, B). Overall levels of SOD1
protein and activity were also low in the HlF1a -/- cells compared to their wild
type counterparts. This might be a major factor in the increased oxidative stress
in basal HlF1a -/— cells and cadmium treatment might further increase the
oxidative stress. The basal level of SOD1 is transcriptionally regulated by the
binding of C/EBPa, SP1, EGRI and WT1 proteins (29). YY1, a protein that binds
to the upstream negative response element of SOD1 promoter, negatively
regulates its expression (30). It is not known if any of these proteins are
influenced by HIF1a protein. The protein levels of the major mitochondrial SOD,
SOD2 (Mn-SOD), were higher in the HlF1a -/- cells, but activity levels were not

significantly different compared to wild type cells.

Cellular glutathione pool shows a marked increase in wild type and HlF1a -/—
cells under CoCI2 and CdCI2 treatments. The higher glutathione levels in the
HlF1a -/- cells compared to wild type cells in all treatments studied may suggest
an increased demand for glutathione for a cell under oxidative stress or an
inability to utilize the available pool (Figure 30). We have shown earlier that
mRNA levels of two major mitochondrial xenobiotic enzymes, GSTp1 and
GSTa1, are significantly lower in the HIFIa -l- cells compared to wild type cells
(28, 31)(data not shown). A reduction in the levels and activity of glutathione
utilizing enzymes may result in an increased cellular glutathione pool. Also it is
not known how much of the total glutathione pool is in the reduced form that is

required for conjugation reactions.

216

We have performed a comparative analysis of the effect of two divalent metal
ions cobalt and cadmium on HlF1a signaling pathway. Taken together, our
results suggest that cobalt-induced cell damage is promoted through HlF1a
stabilization and subsequent transcriptional regulation where, cadmium-induced
cytotoxicity is due to ROS generation and the compromised nature of the ROS
scavenging system in the HIF 1a -/- cells. We show that HlF1a plays a protective
role against cadmium exposure putatively by its ability to maintain the activity of
cellular antioxidant capacity whereas aberrant and constitutive activation of
hypoxic signaling under cobalt exposure leads to cytotoxic effects by the activity
of pro-cell death factors. This study points to the significant role played by HlF1a

in the presence of oxygen and its role in maintaining cellular homeostasis.

Acknowledgements

We would like to thank Dr. Louis King of the MSU flow cytometry facility for help
with the ROS detection procedures, Dr. KangAe Lee for help with the lentiviral
infection procedures and Dr. Randall Johnson, UCSD for providing us with the

wild type and HlF1a -/- MEF cell lines.

217

References

10.

Glusker, J. P. Structural aspects of metal liganding to functional groups in
proteins. Adv Protein Chem, 42: 1-76, 1991.

Nelson, N. Metal ion transporters and homeostasis. Embo J, 18: 4361-
4371, 1999.

Maxwell, P. H., Wiesener, M. 8., Chang, G. W., Clifford, S. C., Vaux, E.
C., Cockman, M. E., Wykoff, C. C., Pugh, C. W., Maher, E. R., and
Ratcliffe, P. J. The tumour suppressor protein VHL targets hypoxia-
inducible factors for oxygen-dependent proteolysis [see comments].
Nature, 399: 271 -275, 1999.

Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. Protein
measurement with the Folin phenol reagent. J Biol Chem, 193: 265-275,
1951.

Vengellur, A. and LaPres, J. J. The Role of Hypoxia Inducible Factor
1{alpha} in Cobalt Chloride Induced Cell Death in Mouse Embryonic
Fibroblasts. Toxicol. Sci., 82: 638-646, 2004.

Chun, Y. S., Choi, E., Kim, G. T., Choi, H., Kim, C. H., Lee, M. J., Kim, M.
S., and Park, J. W. Cadmium blocks hypoxia-inducible factor (HlF)-1-
mediated response to hypoxia by stimulating the proteasome-dependent
degradation of HIF-1a|pha. European Journal of Biochemistry, 267: 4198-
4204, 2000.

Sowter, H. M., Ratcliffe, P. J., Watson, P., Greenberg, A. H., and Harris,
A. L. HlF-1-dependent regulation of hypoxic induction of the cell death
factors BNIP3 and NIX in human tumors. Cancer Res, 61: 6669-6673.,
2001.

Kubasiak, L. A., Hernandez, O. M., Bishopric, N. H., and Webster, K. A.
Hypoxia and acidosis activate cardiac myocyte death through the Bcl-2
family protein BNIP3. Proc Natl Acad Sci USA, 99: 12825-12830, 2002.

Waisberg, M., Joseph, P., Hale, B., and Beyersmann, D. Molecular and
cellular mechanisms of cadmium carcinogenesis. Toxicology, 192: 95-117,
2003.

Leonard, S., Gannett, P. M., Rojanasakul, Y., Schwegler-Berry, D.,
Castranova, V., Vallyathan, V., and Shi, X. Cobalt-mediated generation of
reactive oxygen species and its possible mechanism. J Inorg Biochem, 70:
239-244, 1998.

218

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

Im, J. Y., Paik, S. G., and Han, P. L. Cadmium-induced astroglial death
proceeds via glutathione depletion. J Neurosci Res, 83: 301-308, 2006.

Valko, M., Morris, H., and Cronin, M. T. Metals, toxicity and oxidative
stress. Curr Med Chem, 12: 1161-1208, 2005.

Shaikh, Z. A., Vu, T. T., and Zaman, K. Oxidative stress as a mechanism
of chronic cadmium-induced hepatotoxicity and renal toxicity and
protection by antioxidants. Toxicol Appl Pharmacol, 154: 256-263., 1999.

Kukielka, E. and Cederbaum, A. l. NADPH- and NADH-dependent oxygen
radical generation by rat liver nuclei in the presence of redox cycling
agents and iron. Arch Biochem Biophys, 283: 326-333, 1990.

Sies, H. Strategies of antioxidant defense. Eur J Biochem, 215: 21 3-21 9,
1993.

Jezek, P. and Hlavata, L. Mitochondria in homeostasis of reactive oxygen
species in cell, tissues, and organism. Int J Biochem Cell Biol, 37: 2478-
2503, 2005.

Pastore, A., Federici, G., Bertini, E., and Piemonte, F. Analysis of
glutathione: implication in redox and detoxification. Clin Chim Acta, 333:
19-39, 2003.

Rahman, Q., Abidi, P., Afaq, F., Schiffmann, D., Mossman, B. T., Kamp,
D. W., and Athar, M. Glutathione redox system in oxidative lung injury. Crit
Rev Toxicol, 29: 543-568, 1999.

Baker, M. A., Cerniglia, G. J., and Zaman, A. Microtiter plate assay for the
measurement of glutathione and glutathione disulfide in large numbers of
biological samples. Anal Biochem, 190: 360-365, 1990.

Lehninger, A. L. Role of metal ions in enzyme systems. Physiol Rev, 30:
393-429, 1950.

Stohs, S. J. and Bagchi, D. Oxidative mechanisms in the toxicity of metal
ions. Free Radic Biol Med, 18: 321-336, 1995.

Vande Velde, C., Cizeau, J., Dubik, D., Alimonti, J., Brown, T., Israels, S.,
Hakem, R., and Greenberg, A. H. BNIP3 and genetic control of necrosis-

like cell death through the mitochondrial permeability transition pore. Mol
Cell Biol, 20: 5454-5468, 2000.

Lag, M., Westly, S., Lerstad, T., Bjornsrud, C., Refsnes, M., and
Schwarze, P. E. Cadmium-induced apoptosis of primary epithelial lung
cells: involvement of Bax and p53, but not of oxidative stress. Cell Biol
Toxicol, 18: 29-42, 2002.

219

24.

25.

26.

27.

28.

29.

30.

31.

Ercal, N., Gurer-Orhan, H., and Aykin-Burns, N. Toxic metals and
oxidative stress part I: mechanisms involved in metal-induced oxidative
damage. Curr Top Med Chem, 1: 529-539, 2001.

Kotake-Nara, E., Takizawa, S., Quan, J., Wang, H., and Saida, K. Cobalt
chloride induces neurite outgrowth in rat pheochromocytoma PC-12 cells
through regulation of endothelin-2/vasoactive intestinal contractor. J
Neurosci Res, 81: 563-571, 2005.

Kim, J. W., Tchernyshyov, l., Semenza, G. L., and Dang, C. V. HIF-1-
mediated expression of pyruvate dehydrogenase kinase: a metabolic
switch required for cellular adaptation to hypoxia. Cell Metab, 3: 177-185,
2006.

Llobet, J. M., Domingo, J. L., and Corbella, J. Comparison of antidotal
efficacy of chelating agents upon acute toxicity of Co(|l) in mice. Research
Communications in Chemical Pathology & Pharmacology, 50: 305—308,
1985.

Vengellur, A., Woods, B. G., Ryan, H. E., Johnson, R. S., and LaPres, J.
J. Gene expression profiling of the hypoxia signaling pathway in hypoxia

inducible factor 1 null mouse embryonic fibroblasts. Gene Expression, 11 :
181-197, 2003.

Chang, M. S., Yoo, H. Y., and Rho, H. M. Transcriptional regulation and
environmental induction of gene encoding copper- and zinc-containing
superoxide dismutase. Methods Enzymol, 349: 293-305, 2002.

Cho, J. 8., Chang, M. S., and Rho, H. M. Transcriptional activation of the
human Cu/Zn superoxide dismutase gene by 2,3,7,8-tetrachlorodibenzo-

p-dioxin through the xenobiotic-responsive element. Mol Genet Genomics,
266: 133—141, 2001.

Raza, H., Robin, M. A., Fang, J. K., and Avadhani, N. G. Multiple isoforms

of mitochondrial glutathione S-transferases and their differential induction
under oxidative stress. Biochem J, 366: 45-55, 2002.

220

SUMMARY AND DISCUSSION

The present studies were focused on understanding the relationship between
environmental metals, such as cobalt and cadmium, and biological pathways that
are essential in sustaining life. The prevalence of metals in our environment is a
growing concern with the increase of “disposable” electronics, the EPA approved
use of bio-metal-containing solids and industrial waste in the production of
fertilizers and the industrial production necessary to expand the global economy.
The wholesale release of these metals into the environment can lead to their
large scale accumulation in soil and subsequently in the food chain. Therefore,
understanding the toxicity of these metals is a growing environmental issue.
Characterizing this toxicity is dependent upon comprehension of the basic
biology involved in their signaling and this type of knowledge is critical if we are
to understand the adverse effects metals have on human health. An attempt was
made to understand the molecular mechanism of the metal induced toxicity using

mouse fibroblasts as a model system.

Although certain metals are essential nutrients, excessive exposure to a wide
range of metals can lead to a variety of toxic endpoints. Metals such as cobalt,
nickel, and cadmium can interfere with several important pathways, one of which
is the hypoxia signaling cascade. The hypoxia signaling pathway is activated
under reduced levels of oxygen and helps the organism to maintain energy

levels. Improper regulation of this pathway is involved in diseases such as

221

cardiac ischemia, stroke, and cancer. Cobalt and nickel exposure can mimic
hypoxia by stabilizing HlF1a protein. Global gene expression profiling using
oligonucleotide array as described in chapter 2 showed that cobalt, hypoxia, and
desferoxamine treatments resulted in similar patterns of gene expression in
Hep3B cells. This suggested that a common signaling system was being
induced under these conditions and we hypothesized that certain divalent metals

might induce toxicity in cells by unnaturally activating HlF1a signaling.

We were fortunate that engineered mouse embryonic fibroblasts (MEFs) were
available to perform these experiments. Two MEF cell strains were used, a wild
type (WT) cell that was genotypically normal and a HlF1a -/- cell that was
identical to the WT cells except they lacked a functional HIF1a locus. These
cells enabled us to determine the role of HlF1a in mediating metal-induced
toxicity. As predicted, cobalt and hypoxia induced similar global changes in gene
expression in the wild type cells and this overlap was not observed in the HlF1a
null cells. These results suggest that cobalt and hypoxia are working in a similar
manner as observed in the HepBB cells and indicated that HlF1a is the major
HlFa isoform in MEFs. Cobalt exposure induced the HlF1a-dependent
expression of genes involved in adaptive response, including glucose transport,
glycolysis, angiogenesis, and erythropoiesis. There were also changes in non-
adaptive genes, such as BNip3, a BH3 domain containing pro-cell death factor
involved in apoptosis and necrosis in various cell types. BNip3 protein levels

showed HlF1a-, dose— and time-dependent increase upon cobalt exposure. Cell

222

morphology under cobalt exposure also displayed chromatin condensation in the
absence of caspase-3 activation suggesting this increased expression of Bnip3
had functional consequence. Finally, cell lines that expressed shRNA targeting
Bnip3 were partially protected against cobalt-induced damage and behaved like
HlF1a -/- following metal insult. This was a clear indication that metals like cobalt
that are capable of aberrantly stabilizing HlF1oc and activating the hypoxic
signaling cascade, can result in cellular damage through direct transcriptional

regulation of cell death promoting factors.

The in vivo effect of cobalt exposure during cobalt treatment of anemia is
characterized by polycythemia, an over—production of read blood cells. In vivo,
the production of red blood cells is primarily regulated by erythropoietin (EPO), a
growth factor produced by the kidney and liver and known HIF1 target gene. In
the Hep3B cell culture model, cobalt was capable of increasing EPO expression
and it is thought cobalt exposure, in vivo, would lead to a similar aberrant EPO
expression, leading to the observed polycythemia. Gene expression under
cobalt exposure in fibroblasts also suggests a potential role of various genes in
mediating other in vivo effects. For example, chronic cobalt exposure through
inhalation results in many pathological conditions, such as asthma, alveolitis and
pulmonary fibrosis. BNip3, a cobalt induced gene is known to be associated with
tissue necrosis and might play a role in tissue necrosis under cobalt exposure
eventually leading to fibrosis. The various genes expressed under cobalt
exposure might also be playing a role in other symptoms such as allergic

dermatitis and cardiomyopathy.

223

The fact that loss of HIF1a protects against cobalt-induced toxicity establishes a
role for the hypoxia signaling cascade in the cytotoxicty of hypoxic—mimicking
metals, but what part, if any, does it play in the toxicity of divalent metals that do
not stabilize HlF1a, such as cadmium? The results were very different; HlF1a
null cells were more susceptible to cadmium-induced damage than their wild type
counterparts. HlF1a null cells exposed to cadmium exhibited chromatin
condensation and caspase-3 activation characteristic of apoptotic cell death but
lacked the expression of BNip3. Manipulation of Bnip3 levels by shRNA also did
not affect the toxicity of cadmium. HlF1a protein is thought to be degraded
under normoxia without mediating any transcriptional activity. But it is clear that
HlF1a protein directly or indirectly affects cellular processes that lead to
protection against cadmium in wild type cells. These results suggest that the
HIF1a protein has cellular functions under normoxia, either non-transcriptional or

transcriptional.

Metals, by their nature can generate reactive intermediates causing oxidative
stress in cells. Cobalt is a Fenton metal that produces hydroxyl radicals by
reacting with hydrogen peroxide and superoxide. In contrast, cadmium depletes
cellular glutathione levels interfering with the activity of various cellular
antioxidant enzyme systems and generating oxidative stress indirectly. Cellular
ROS measurements indicated that ROS levels are greater under cadmium
treatment compared to cobalt exposure. Again HIF10L null cells had much higher

ROS levels under cadmium exposure compared to wild type cells suggesting a

224

possible explanation for increased cytotoxicity in those cells. Interestingly, there
was a higher basal ROS levels in the HlF1a null cells compared to wild type
cells. There are two possible reasons for such an increase in ROS levels:
increased ROS generation or decreased ROS scavenging activity. Recent data
and results presented in this thesis suggest that HIF1a regulates PDK1, the
kinase responsible for regulating pyruvate’s fate within the cell. It is possible that
the loss of HlF1oc in the null cells results in increased pyruvate flux through TCA
cycle due to disregulation of PDK1, even under normoxia. This would result in
more carbon moving into the mitochondria and the TCA cycle and lead to an
increase in the activity of the electron transport chains. The increased ETC flux
would tax the normal ROS maintenance machinery resulting in the change in
basal levels in ROS observed in the HIF1a -/- cell. In addition, this increased
oxidative stress would tax the cells in such a way that a metal capable of
promoting further oxidative damage through direct competition of these same
systems (eg. glutathione), such as cadmium, would be more cytotoxic under

these conditions.

Our data suggest another possible explanation for the increased ROS levels in
the HlF1a null cells. We observed differences in ROS scavenging enzymes,
SOD1 and SOD2, and the antioxidant glutathione between HIF 1a -/- and wild
type cells. For example, SOD1, the major cytosolic dismutase, levels were low in
cadmium treated HlF1a null cells potentially compromising their ability to cope

with the greater oxidative stress. These observations point to a potential role of

225

HIF1a in the maintenance of cellular redox status. This maintenance happens in
the absence of cellular stress and again suggests that HlF1a has a role in
cellular physiology outside of responding to hypoxic conditions. In fact, the
results suggest that this “norrnoxic” role for HlF1a in cellular homeostasis might
be equally important to its role in hypoxia adaptation. The maintenance and
possible surveillance role of HlF1a for the redox state of the cell might actually
establish its basal activity. For example, if ROS play a direct role in HlF1a
stability as has been proposed by various research groups, then the ability of
HlF1a to directly influence the basal expression level of genes, such as SOD1,
which in turn controls the ROS levels within the cell would create a feedback

circle to regulate oxidative stress.

Divalent metals, especially cobalt and cadmium are an important environmental
contaminants and their presence will only increase over the coming years. Our
dependence on disposable electronics and the ever increasing industrialization of
our world will ensure that metals will become prominent environmental
contaminants. Understanding the cellular and physiological response to these
metals will help our understanding of the toxicological concerns that they pose.
The research presented in this thesis also has implications outside of the field of
toxicology and the environment. As mentioned above, hypoxia signaling is
critical for several pathological conditions, as well as normal development and
immune function. In addition, the cascade has now become a viable target for

cancer chemotherapy due to its role tumorigenesis. We hope that the present

226

study will create a foundation for the analysis of this critical signaling pathway
and will allow us to relate downstream effects to HIF1a and other proteins
involved in the hypoxia signaling network. This research will establish a basis for
a systems biology approach to the study of hypoxia and its role in normal and

stress mediated biology.

227

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

lllljlllljljljjlllljjjjjjijllllljlllljjll