J Appl Physiol 96: 765–773, 2004;
10.1152/japplphysiol.00836.2003.
HIGHLIGHTED TOPIC
Invited Review
Oxygen Sensing in Health and Disease
Functional genomics approach to hypoxia signaling
Karen A. Seta and David E. Millhorn
Department of Genome Science, Genome Research Institute, University of Cincinnati, Cincinnati, Ohio 45237
subtractive suppression hybridization; microarray; real-time polymerase chain
reaction; RNA interference
MAMMALIAN CELLS REQUIRE A constant supply of oxygen to
maintain energy balance. Hypoxia leads to reduced oxidative
phosphorylation and depletion of cellular ATP. Sustained hypoxia can result in cell death. It is therefore not surprising that
sophisticated adaptive mechanisms have evolved that enhance
cell survival during hypoxia. Although the nature of the adaptive mechanisms remains unknown, reduced metabolic activity
and conservation of ATP are often involved (33). However,
reduced activity is not a viable option for oxygen-sensing cells
(e.g., carotid body type I cells) whose role in maintaining
oxygen homeostasis often requires increased activity during
hypoxia. These cell types have evolved the ability to respond to
reduced oxygen tension by altering protein and gene expression levels to increase oxygen delivery to hypoxic tissues and
organs. To carry out these functions, oxygen-sensing cells must
alter the expression patterns of potentially hundreds of genes
and proteins. As a result, oxygen-sensing cells have evolved
the remarkable ability to tolerate hypoxia. During the past 10
or so years, there have been a growing number of reports on
hypoxia-induced transcription of specific genes that mediate
such cellular functions as erythropoiesis, pulmonary ventilation and blood flow, angiogenesis, and energy metabolism. In
Address for reprint requests and other correspondence: D. E. Millhorn, Dept.
of Genome Science, Genome Research Institute, Univ. of Cincinnati, 2180 E.
Galbraith Rd., Cincinnati, OH 45237 (E-mail: [email protected]).
http://www.jap.org
this review, we describe a unique experimental approach that
utilizes focused cDNA libraries coupled to microarray analyses
to identify hypoxia-responsive signal transduction pathways
and genes that confer the hypoxia-tolerant phenotype.
CELLULAR RESPONSE TO HYPOXIA: A BRIEF OVERVIEW
Exposure of pheochromocytoma (PC12) cells to reduced O2
leads to membrane depolarization and an increase in intracellular free Ca2⫹ (87). This membrane depolarization is mediated by the O2-sensitive potassium channel Kv1.2 (12). This
depolarization and Ca2⫹ increase is not seen in nonexcitable,
O2-sensing cell types such as HepG2 and Hep3B cells. Several
critical signaling pathways regulate gene expression during
hypoxia. These include the cAMP-protein kinase A (PKA) (3),
Ca2⫹-calmodulin (3), p42/44 mitogen-activated protein kinase
(MAPK) (15), stress-activated protein kinase (SAPK; p38
kinase) (16), and phosphatidylinositol 3-kinase/Akt (6) pathways.
The hypoxia-inducible factor (HIF) family of transcription
factors plays a critical role in mediating the hypoxic regulation
of several genes (63, 64). The functional unit of these transcription factors consists of ␣/␤-heterodimers. HIF-1␤ is identical to the aryl hydrocarbon nuclear translocator and is necessary for transactivation of target genes by the ␣-subunits.
The HIF ␣-subunits are basic helix-loop-helix PAS domain
proteins. Both prokaryotes and eukaryotes express PAS do-
8750-7587/04 $5.00 Copyright © 2004 the American Physiological Society
765
Downloaded from http://jap.physiology.org/ by 10.220.33.5 on June 18, 2017
Seta, Karen A., and David E. Millhorn. Functional genomics approach to
hypoxia signaling. J Appl Physiol 96: 765–773, 2004; 10.1152/japplphysiol.
00836.2003.—Mammalian cells require a constant supply of oxygen to maintain
energy balance, and sustained hypoxia can result in cell death. It is therefore not
surprising that sophisticated adaptive mechanisms have evolved that enhance cell
survival during hypoxia. During the past few years, there have been a growing
number of reports on hypoxia-induced transcription of specific genes. In this
review, we describe a unique experimental approach that utilizes focused cDNA
libraries coupled to microarray analyses to identify hypoxia-responsive signal
transduction pathways and genes that confer the hypoxia-tolerant phenotype. We
have used the subtractive suppression hybridization (SSH) method to create a
cDNA library enriched in hypoxia-regulated genes in oxygen-sensing pheochromocytoma cells and have used this library to create microarrays that allow us to
examine hundreds of genes at a time. This library contains over 300 genes and
expressed sequence tags upregulated by hypoxia, including tyrosine hydroxylase,
vascular endothelial growth factor, and junB. Hypoxic regulation of these and other
genes in the library has been confirmed by microarray, Northern blot, and real-time
PCR analyses. Coupling focused SSH libraries with microarray analyses allows one
to specifically study genes relevant to a phenotype of interest while reducing much
of the biological noise associated with these types of studies. When used in
conjunction with high-throughput, dye-based assays for cell survival and apoptosis,
this approach offers a rapid method for discovering validated therapeutic targets for
the treatment of cardiovascular disease, stroke, and tumors.
Invited Review
766
GENOMICS OF HYPOXIA SIGNALING
J Appl Physiol • VOL
81). AP-1 is induced by a variety of environmental stresses,
including hypoxia. Regulation of the various subunits may be
cell specific. For example, in HepG2 cells, c-Jun protein
accumulates in response to hypoxia (50). We have not seen
activation of the JNK pathway or c-Jun phosphorylation in
PC12 cells (16), although others have reported a modest induction
of c-Jun in response to hypoxia in this cell line (55). Other
transcription factors regulated by hypoxia include CREB (4),
GATA-1 (88) and GATA-2 (77), and forkhead protein (25, 78).
The transcription factors regulated by hypoxia work in
concert to fine tune the hypoxic response. Many hypoxiainducible genes contain promoter and enhancer elements for
multiple transcription factors in their regulatory regions. For
example, the tyrosine hydroxylase (TH) promoter contains
AP-1, AP-2, cAMP response element, and multiple HRE sites
(18, 38, 53), and all are involved in the hypoxic regulation of
this gene (38, 48, 49, 51, 53). Hypoxic induction of endothelin-1 in vascular endothelial cells requires its HRE as well as
AP-1, GATA-2, and nuclear factor-1 binding sites (84). IL-8
gene expression by hypoxia in human ovarian cancer cells
requires cooperation of AP-1 and NF-␬B binding sites, as
mutation of either abolishes induction of an IL-8 reporter
construct (73, 83). In addition, these transcription factors
regulate the expression and activity of each other. HIF-1␣ and
p53 interact in vivo (1). p53 promotes murine double minute2mediated ubiquitination and proteasomal degradation of the
HIF-1␣ subunit of HIF-1 (57), whereas HIF-1␣ promotes
stabilization of p53 (1). HIF-1␣ also interacts with Jab1 (Jun
activation domain-binding protein-1), a coactivator of AP-1,
which promotes HIF-1␣ stability at least in part by interfering
with the binding of HIF-1␣ to p53 (2). In addition, NF-␬B
binds to the p53 promoter and regulates p53 expression (82). It
is obvious that a thorough understanding of how cells respond
to hypoxia will require an understanding of how the many
signaling pathways and transcription factors involved in the
hypoxic response are coordinately regulated.
A MODEL SYSTEM
Gaining a deeper knowledge of how oxygen-sensing cells
transduce reduced O2 tension into signals that alter gene
expression has proved difficult due to the low abundance of
these cells in intact tissues. To address this problem, clonal cell
lines that respond to hypoxia in a fashion similar to oxygensensing cells in vivo have been established as model systems.
One such cell line is the PC12 cell line. Oxygen-sensing
capabilities have been documented in both rat (13, 87) and
mouse (23) pheochromocytoma cell lines.
There is now a considerable body of evidence that PC12
cells are an attractive model for carotid body type 1 cells
(reviewed in Ref. 8). PC12 cells closely resemble type I cells
morphologically and phenotypically (28). Both carotid body
type I cells and PC12 cells depolarize rapidly (within seconds)
during hypoxia via inhibition of an O2-sensitive outward K⫹
current (13, 19, 46, 86). This depolarization is followed by an
increase in intracellular Ca2⫹ (45, 80). Both cell types synthesize and release dopamine in response to acute hypoxia (17, 18,
24, 26, 72). Carotid body type I cells and PC12 cells have
similar longer-term responses to hypoxia (within hours), including induction of TH (17, 18) c-fos and junB (53) mRNAs.
Thus this cell line is a useful system in which to study the
96 • FEBRUARY 2004 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.5 on June 18, 2017
main proteins. To date, three mammalian HIF ␣-subunits have
been identified: HIF-1␣, HIF-2␣ or EPAS, and HIF-3␣.
HIF-1␣ is expressed in a wide variety of cells and tissues,
whereas HIF-2␣ and HIF-3␣ have a more limited distribution.
HIF-2␣ is the predominant HIF found in PC12 cells.
The levels of all three HIF ␣-subunits are rapidly increased
by hypoxia as a result of protein stabilization due to the
inhibition of ubiquitin-mediated proteasomal degradation (29,
34, 76). Once activated by reduced O2, HIF-␣ dimerizes with
HIF-1␤ in the nucleus and binds to the hypoxia response
element (HRE), an enhancer element with the consensus sequence 5⬘-RCGTG-3⬘, located within the regulatory region of
many hypoxia-responsive genes (65, 67). The binding of the
HIF-␣/␤ complex during hypoxia regulates genes involved in
a wide variety of physiological process, including angiogenesis, proliferation, cell survival and death, erythropoiesis, energy metabolism, and oxygen chemoreception. HIF-3␣ lacks
one of the two transactivation domains found in the COOH
terminus of the other HIF-␣ proteins and might be a negative
regulator of hypoxia-inducible gene expression, possibly by
competing for limited quantities of HIF-1␤ (31).
The roles of the various HIF proteins in the hypoxic response are unclear. In mouse embryo fibroblasts, only HIF-1␣
undergoes oxygen-dependent protein degradation. HIF-2␣ is
detected under all oxygenation conditions and is localized to
the cytoplasm. Also, endogenous HIF-2␣ is transcriptionally
inactive in hypoxia, but overexpressed HIF-2␣ could translocate to the nucleus and stimulate expression of hypoxiainducible genes (54). Loss of HIF-1␣, but not HIF-2␣, protects
embryonic stem cells from hypoxia-induced apoptosis. However, loss of either HIF-1␣ or HIF-2␣ protects these cells from
hypoglycemia-induced apoptosis (9). HIF-2␣ has been reported to
play a critical role in hematopoiesis in bone marrow (62).
Although HIF proteins mediate much of the transcriptional
response to hypoxia, a variety of other factors are also involved. Hypoxia also increases levels of the p53 tumor suppressor protein (27). Agents that inhibit the accumulation of
p53 by DNA-damaging agents do not prevent p53 accumulation by hypoxia, indicating that different stresses induce p53 by
different mechanisms (27, 58). Although the majority of p53 is
nuclear and cytoplasmic, hypoxia causes p53 accumulation in
the mitochondria (60). Mitochondrial function appears to be
critical for hypoxia-induced p53 accumulation (11).
NF-␬B is member of the Rel-related proteins (p50, p52,
v-Rel, c-Rel, RelA/p65, and RelB; reviewed in Ref. 59). In the
cytoplasm, inactive NF-␬B dimers interact with I␬B. On stimulation with cytokines or stress agents, I␬B becomes phosphorylated and dissociates from NF-␬B, exposing a nuclear translocation signal on NF-␬B. The p50 and p52 proteins are
constitutively nuclear and lack the I␬B binding site and transactivation domain. Homodimers of p50 and p52 may act as
transcriptional repressors. Cytokine stimulation of NF-␬B
causes serine phosphorylation and subsequent degradation of
I␬B (22). In contrast, stimulation of NF-␬B by hypoxia is via
tyrosine phosphorylation of I␬B (43) and does not always
involve I␬B degradation (35). Tyrosine phosphorylation of I␬B
prevents its phosphorylation on serine (35, 75).
Activator protein-1 (AP-1) is a transcription factor that
consists of dimers of jun/jun (c-jun, junB, junD) or jun/fos
[s-fos, fosB, fra-1, fra-2, ATF-2, cAMP response element
binding protein (CREB)] subunits (reviewed in Refs. 47 and
Invited Review
GENOMICS OF HYPOXIA SIGNALING
767
molecular and cellular basis of O2 chemosensitivity and the
mechanisms by which O2-responsive genes are regulated by
hypoxia.
Using this model system, our laboratory has identified a
number of genes that are regulated by hypoxia (reviewed in
Refs. 14 and 70). The adenosine A2A receptor is upregulated by
hypoxia (39, 42). In contrast, the N-methyl-D-aspartate receptor
(41), PKA, and Ca2⫹/calmodulin-dependent protein kinase II
(41) are downregulated. The transcription factors CREB (4, 5)
and HIF-2␣ (also known as EPAS1) (15) are activated. Our
laboratory has also implicated several signal transduction pathways in hypoxia-induced gene regulation in PC12 cells. The
p42/p44 MAPK and p38␣/p38␥ SAPK pathways are activated
by hypoxia (16). HIF-2␣ trans-activation of a transfected
HRE-reporter construct depends on MAPK activation but is
independent of Ras and requires Ca2⫹/calmodulin (15). In
contrast, regulation of MAPK phosphatase-1 (MKP-1), which
dephosphorylates MAPKs and SAPKs, is Ca2⫹ independent
and is partially mediated by p38 SAPK but not by MAPK (69).
The increase in adenosine A2A receptor levels requires Ca2⫹
via protein kinase C rather than through calmodulin (42).
Table 1. Frequency of hypoxia-induced genes isolated in the
SSH library
Gene Name
Ref.
No. of
Copies
No. of
Fragments
Frequency
JunB
TH
VEGF
Hexokinase II
Bnip3
A2 adenosine receptor
Phosphoglycerate kinase
Plasminogen activator inhibitor
Pyruvate kinase
53, 55
18
74
52
75a
39
66
24a
66
14
5
3
2
2
1
1
1
1
2
7
4
1
2
1
1
3
3
5.2%
1.9%
1.1%
0.7%
0.7%
0.4%
0.4%
0.4%
0.4%
TH, tyrosine hydroxylase; SSH, subtractive suppression hybridization.
J Appl Physiol • VOL
Fig. 2. Quality control standards (std) for subtractive suppression hybridization (SSH) microarrays. Plant gene mRNAs (SpotReport 1–10) were spiked
into total RNA samples before they were labeled for microarray analysis.
SpotReport 1–3 were added in equal amounts to the Cy3 and Cy5 reactions to
produce spots with a 1:1 intensity ratio (solid line). SpotReport 6 and 7 were
spiked in at 2-fold greater concentrations in the Cy5 and Cy3 reactions,
respectively (dashed line). SpotReport 5 and 8 were spiked into the Cy5 and
Cy3 reactions with a 5-fold difference (dotted line). SpotReport 4 and 9 were
spiked to show a 10-fold difference (dotted-dashed line). SpotReport 10 was
added only to the Cy3 reaction. Background-subtracted median pixel intensities are plotted. Each point on the scatter plot represents one spot on the array.
96 • FEBRUARY 2004 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.5 on June 18, 2017
Fig. 1. Overview of the signal transduction pathways activated by hypoxia. ATF-2, activating transcription factor-2; CREB, cAMP
response element binding protein; EPAS-1, endothelial PAS protein-1; FKHR, forkhead transcription factor; GRB2, growth-factorreceptor-bound protein 2; GSK3, glycogen synthase kinase-3; HIF, hypoxia-inducible factor; LRG, leucine-rich glycoprotein;
MEK, MAPK kinase; MEKK, MAPK kinase kinase; MKP-1, MAPK phosphatase-1; PDK1, 3-phosphoinositide-dependent protein
kinase 1; PI3-kinase, phosphatidylinositol 3-kinase; Ptdins(3,4)P2, phosphatidylinositol 3,4-bisphosphate; Ptdins(3,4,5)P3, phosphatidylinositol 3,4,5-trisphosphate; Pyk2, proline-rich tyrosine kinase 2; RLPK, RSK-like protein kinase; RSK, ribosomal S6
kinase; SAPK, stress-activated protein kinase; SRF, serum response factor.
Invited Review
768
GENOMICS OF HYPOXIA SIGNALING
Interestingly, adenosine attenuates the hypoxia-induced increase in intracellular Ca2⫹ through a mechanism that requires
PKA, and membrane excitability is thought to be maintained
by a compensatory decrease in PKA activity (39, 40). The
Akt/phosphatidylinositol 3-kinase pathway, which is thought
to be critical for cell survival, is also activated (6). The
transcription factor CREB is activated through a novel, as-yet
unidentified, signaling mechanism (4, 5). It is clear that these
pathways are not acting independently; rather, there is a great
deal of cross-talk between them (see Fig. 1).
FOCUSED CLONE SETS SIMPLIFY MICROARRAY ANALYSIS
Although traditional molecular techniques have yielded a
great deal of information about the biology of oxygen sensing,
the study of hypoxia-regulated genes individually by traditional methods does not provide a comprehensive picture of
hypoxia-regulated gene expression. To gain an understanding
of the many genes involved in the hypoxic response, a more
global approach was needed. The development of highthroughput genomic technologies such as microarrays has
dramatically changed the way we can evaluate regulation of
Downloaded from http://jap.physiology.org/ by 10.220.33.5 on June 18, 2017
Fig. 3. Comparison between SSH microarrays and database-derived microarrays. A: PC12 cells were exposed to
normoxia (21% O2) or hypoxia (1% O2) for 6 h. Total RNA
was isolated, and 10 ␮g were converted to 32P-labeled
cDNA in a standard reverse transcriptase reaction. The
radiolabeled cDNA was hybridized to rat cDNA filter arrays
(GeneFilter 2 array, Research Genetics, Carlsbad, CA) that
contained 5,299 rat sequences. The hybridized arrays were
washed and exposed to a phosphorimager screen and
scanned with a Storm XXX system (Molecular Dynamics).
B: for each spot on the array in A, the ratio of the signal
intensities of the hypoxia sample to the normoxia sample
was calculated. The graph depicts the number of genes that
produced a given ratio. C: PC12 cells were exposed to
normoxia (21% O2) or hypoxia (1% O2) for 6 h. Total RNA
was isolated and converted to Cy3- or Cy5-labeled cDNA
as indicated. Labeled samples were mixed and hybridized to
glass arrays onto which PCR products from the SSH library
had been spotted. The arrays were washed and scanned at
the appropriate wavelength for each dye (532 nm for Cy3
and 635 nm for Cy5) with an Axon GenePix 4000A scanner
(Axon Instruments, Union City, CA). The data are plotted
as the background-subtracted median pixel intensities for
each spot on the array. Note that the hypoxia-regulated
genes junB, tyrosine hydroxylase (TH), and VEGF center
around the line of unity in the normoxia vs. normoxia
hybridization, whereas they are strongly upregulated in the
hypoxia vs. normoxia hybridization. D: median Cy3-to-Cy5
ratio was calculated for each gene on the array. The number
of genes that produced a given ratio are depicted.
J Appl Physiol • VOL
96 • FEBRUARY 2004 •
www.jap.org
Invited Review
769
GENOMICS OF HYPOXIA SIGNALING
gene expression. Use of microarrays greatly accelerates the
rate of discovery of genes involved in the response to a
stimulus, while also providing information about global patterns of gene regulation.
Database-derived arrays (e.g., Affymetrix and many commercially available cDNA arrays) offer the advantage of containing typically thousands to tens of thousands of genes.
Despite their large size, however, there is no guarantee that the
genes involved in the process under study will be included on
these arrays, especially if one is working with an organism
other than yeast, human, mouse, or rat (this disadvantage will
disappear as the genomes of more organisms are completed).
In addition, only a handful of genes are differentially expressed
in response to a stimulus; therefore, in a database-derived clone
set, which is essentially randomly selected, the false-positive
rate is high. Other methodologies for genome-wide analysis
J Appl Physiol • VOL
Table 2. Validation of genes isolated in the SSH library
Gene Name
Microarraya
PCRb
JunB
TH
VEGF
Hexokinase II
Bnip3
A2 adenosine receptor
Phosphoglycerate kinase
Plasminogen activator inhibitor
Pyruvate kinase
⫹
⫹
⫹
⫹
⫹
⫹
1.67-fold
⫹
1.99-fold
⫹
⫹
⫹
⫹
⫹
⫹
⫹
⫹
NS
For microarray, ⫹ indicates induction of ⬎2-fold (average of minimum of
2 experiments). For PCR, ⫹ indicates statistically greater levels in hypoxia
samples than in normoxia samples (P ⬍ 0.001 by t-test; n ⫽ at least 4). NS ⫽
not significant.
96 • FEBRUARY 2004 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.5 on June 18, 2017
Fig. 4. Validation of gene regulation by real-time PCR. PC12 cells were
exposed to normoxia (dashed lines) or hypoxia (1% O2, solid lines) for 6 h.
Total RNA was isolated, and quantitative real-time PCR (SYBR Green I dye
intercalation method) was performed as described in the text. Growth curves
are plotted for TH (A) and GAPDH (B). The threshold cycles (Ct) were
calculated by the second-derivative method and represent the point at which
the growth curves enter log-linear phase. In this experiment, GAPDH is
unregulated, whereas TH is strongly induced. Insets: melt curves, which
indicate the formation of single products.
include differential display and serial analysis of gene expression (SAGE). Differential display allows direct side-by-side
comparisons of most of the mRNAs between or among two or
more samples by amplifying expressed sequences in a systematic and sequence-specific manner using multiple primer-pair
combinations (44). In SAGE, short sequence tags from all
expressed mRNAs are generated and ligated together into long
molecules that are cloned into vectors and sequenced (79). The
expression of a given gene is proportional to the number of
times its sequence is represented. These methods offer many
advantages, chief among them being relatively low cost and no
requirement for specialized equipment.
We have used the subtractive suppression hybridization
(SSH; PCR-Select kit, Clontech, Palo Alto, CA) method to
create a cDNA library enriched in hypoxia-regulated genes and
have used this library to create microarrays that allow us to
examine hundreds of genes at a time (7, 68). SSH libraries
provide a powerful biological approach for obtaining clones
that are differentially expressed between two populations. The
libraries consist of a subset of genes that are involved in
mediating the phenotype of interest. The protocol combines
high-subtraction efficiency with an equalized representation of
abundant and rare mRNAs (20, 21, 30). The method can be
used with any two populations whose gene expression profiles
differ (e.g., treated vs. untreated, different developmental
states, different tissue types). In addition, our library, which is
enriched for genes upregulated by 6 h of hypoxia (1% O2) in
PC12 cells, contains about 250 known genes and ⬃60 unidentified genes and expressed sequence tags. Among the identified
genes are many genes that have previously been associated
with hypoxic regulation, including TH (17), VEGF (74), and
junB (53, 55) (Table 1). The frequency of these genes in the
library corresponds well with their induction by hypoxia. In
addition to the prototypical hypoxia-regulated genes TH, junB,
and VEGF, the SSH library contains genes involved in glucose
metabolism, apoptosis, and neurotransmission. One adaptation
to hypoxia involves generating ATP via the glycolytic pathway
instead of oxidative phosphorylation. Several enzymes in the
glycolytic pathway, such as pyruvate kinase (52), phosphoglycerate kinase (52), and hexokinase II (66), have HIF-1
binding sites and are regulated by hypoxia. One of the apoptosis-related proteins identified by microarray analyses was
Bnip-3, a recently described proapoptotic member of the BH3only Bcl-2 protein family (see Ref. 85 for review). These few
Invited Review
770
GENOMICS OF HYPOXIA SIGNALING
examples touch only briefly on the many phenotypic alterations
that occur in response to hypoxia.
Microarrays have been produced by amplifying the clone
inserts from the SSH library and spotting the PCR products
onto glass slides (32, 61, 71). All of the widely used microarray
normalization methods (56) assume that the average expression
ratio for all the genes on the array is equal to one. This is not
the case for microarrays made from SSH libraries because the
gene content is biased. To avoid normalization and scanner
artifacts, we have included 10 exogenous plant genes (SpotReport 1–10, Stratagene, La Jolla, CA) scattered throughout
Fig. 6. Microarray analysis of the dependence of hypoxic gene regulation on extracellular Ca2⫹. PC12 cells were incubated for 6 h
under normoxic (21% O2; A) or hypoxic (1%
O2; B) conditions, in the presence (C) or
absence (D) of extracellular Ca2⫹ as indicated. Total RNA was isolated and converted
to Cy3- or Cy5-labeled cDNA as indicated.
Labeled samples were mixed and hybridized
to glass arrays onto which PCR products
from the SSH library had been spotted, as
described in Fig. 3. Data are plotted as the
background-subtracted median pixel intensities for each spot on the array.
J Appl Physiol • VOL
96 • FEBRUARY 2004 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.5 on June 18, 2017
Fig. 5. Schematic of modified loop design for microarray analysis of signal
transduction pathways. Thick solid arrows indicate hybridizations that would
be performed in a Kerr and Churchill loop design. The arrowhead represents
one dye (either Cy3 or Cy5), and the ball at the other end indicates the other
dye. Thin solid arrows represent corresponding dye-flip hybridizations, in
which the Cy3 and Cy5 dyes are reversed. Dashed arrows indicate additional
hybridizations recommended as part of the modified loop design.
the body of the array. SpotReport 1, 2, and 3 are also spotted
at the top of the array, and their mRNAs are added in equal
amounts (5 ng, 0.5 ng, and 0.05 ng per 100 ␮g of total RNA,
respectively) to each sample labeling reaction to provide a
known set of genes that should have a ratio of 1:1. After the
slide has been scanned, the SpotReport 1–3 spots at the top of
the array are quickly analyzed. The scanner settings are adjusted, and the slide is rescanned until the ratios for these spots
are as close as possible to 1:1 (we generally accept ratios of
0.95–1.05). The SpotReport 1–3 spots in the body of the array
are used to demonstrate that there is no positional bias on the
arrays. The other seven mRNAs are spiked into the two
labeling reactions at different known ratios. When the background-subtracted fluorescence values for each wavelength are
plotted, these spots fall along their predicted lines (Fig. 2). This
system provides a good control for all the technical aspects of
microarray analysis (labeling, hybridization, and scanning).
By beginning with a phenotype instead of a database for
gene selection, this hybrid approach allows one to specifically
study genes of interest while reducing much of the biological
noise associated with microarray analysis. Comparison of array
results showed that the percentage of genes upregulated at least
twofold by hypoxia is an order of magnitude higher with our
focused array (49%) than with a commercial filter-based array
(4.6%) (Fig. 3). When the absolute number of genes under
consideration is reduced and one can bias the clone set in favor
of genes related to a phenotype, biological noise is minimized
and data analysis is greatly simplified. Reproducibility is also
high, with ⬎40% of genes showing twofold or greater upregulation in at least two of three replicate experiments (not
shown).
Invited Review
GENOMICS OF HYPOXIA SIGNALING
VALIDATION OF RESULTS
ANALYSIS OF SIGNAL TRANSDUCTION PATHWAYS
The use of microarray analysis as a screening tool to identify
interesting genes for further study is relatively straightforward.
Analysis of gene expression patterns related to hypoxia-responsive signal transduction pathways is considerably more
complicated because multiple treatments (e.g., a signal transduction pathway activator/inhibitor and hypoxia) are given to
each sample. In the simplest scenario, four different conditions
are possible (vehicle ⫹ normoxia, vehicle ⫹ hypoxia, drug ⫹
normoxia, and drug ⫹ hypoxia). In a single-gene analysis
experiment, expression levels of the gene of interest would be
measured in each sample and statistical comparison would be
made by two-way ANOVA. In a two-color fluorescent hybridization, multiple hybridizations of combinations of these four
conditions must be made. Kerr and Churchill (36, 37) have
outlined principles for experimental design and data analysis
for these types of experiments. Figure 5 depicts the experimental conditions and hybridizations that would be used to study
the effects of Ca2⫹ on hypoxia-induced gene expression and
the information that is gained from each set of hybridizations.
This general experimental design can be used to dissect signaling pathways and their downstream targets activated by
hypoxia. The feasibility of this approach is demonstrated in an
experiment in which PC12 cells were incubated in the presence
J Appl Physiol • VOL
or absence of extracellular Ca2⫹ under either normoxic or
hypoxic (1% O2) conditions. The hybridizations in the basic
loop design (solid arrows in Fig. 5) were performed. Removal
of extracellular Ca2⫹ had a small effect on the resting (normoxic) gene expression levels (Fig. 6A) and a somewhat larger
effect on hypoxic gene expression levels (Fig. 6B). In contrast,
the presence or absence of extracellular Ca2⫹ has a dramatic
effect on hypoxia-induced gene expression (Fig. 6, C and D).
Differential analyses such as these, using pharmacological and
oligo-based inhibition (e.g., RNA interference) of specific
signal transduction pathways, offer a powerful tool to examine
hypoxia-responsive signal transduction pathways and their
gene targets.
FUTURE DIRECTIONS: RELATING GENE EXPRESSION
PATTERNS TO PHENOTYPES
The ultimate goal of gene expression analysis is to elucidate
gene expression patterns associated with a particular phenotype. We have begun to try to relate gene expression patterns
with specific phenotypic outcomes in response to hypoxia,
namely, cell survival or death. The coupling of focused microarray analysis described above with high-throughput, dyebased assays for cell survival and apoptosis offers a rapid
approach for discovering validated therapeutic targets for the
treatment of cardiovascular disease, stroke, and tumors.
GRANTS
This work was supported by National Institutes of Health Grants HL-33831,
DK-58811, HL-66312, HL-59945, and HL-07571, the Parker B. Francis
Foundation (KAS), and a Grant from the U.S. Army.
REFERENCES
1. An WG, Kanekal M, Simon MC, Maltepe E, Blagosklonny MV, and
Neckers LM. Stabilization of wild-type p53 by hypoxia-inducible factor
1␣. Nature 392: 405–408, 1998.
2. Bae MK, Ahn MY, Jeong JW, Bae MH, Lee YM, Bae SK, Park JW,
Kim KR, and Kim KW. Jab1 interacts directly with HIF-1␣ and regulates
its stability. J Biol Chem 277: 9–12, 2002.
3. Beitner-Johnson D, Leibold J, and Millhorn DE. Hypoxia regulates the
cAMP- and Ca2⫹/calmodulin signaling systems in PC12 cells. Biochem
Biophys Res Commun 241: 61–66, 1998.
4. Beitner-Johnson D and Millhorn DE. Hypoxia induces phosphorylation
of the cyclic AMP response element-binding protein by a novel signaling
mechanism. J Biol Chem 273: 19834–19839, 1998.
5. Beitner-Johnson D, Rust RT, Hsieh T, and Millhorn DE. Regulation of
CREB by moderate hypoxia in PC12 cells. Adv Exp Med Biol 475:
143–152, 2000.
6. Beitner-Johnson D, Rust RT, Hsieh TC, and Millhorn DE. Hypoxia
activates Akt and induces phosphorylation of GSK-3 in PC12 cells. Cell
Signal 13: 23–27, 2001.
7. Beitner-Johnson D, Seta K, Yuan Y, Kim HW, Rust R, Conrad PW,
Kobayashi S, and Millhorn DE. Identification of hypoxia-responsive
genes in a dopaminergic cell line by subtractive cDNA libraries and
microarray analysis. Parkinsonism Related Disorders 7: 273–281, 2001.
8. Beitner-Johnson D, Shull GE, Dedman JR, and Millhorn DE. Regulation of gene expression by hypoxia: a molecular approach. Respir
Physiol 110: 87–97, 1997.
9. Brusselmans K, Bono F, Maxwell P, Dor Y, Dewerchin M, Collen D,
Herbert JM, and Carmeliet P. Hypoxia-inducible factor-2␣ (HIF-2␣) is
involved in the apoptotic response to hypoglycemia but not to hypoxia.
J Biol Chem 276: 39192–39196, 2001.
10. Bustin SA. Absolute quantification of mRNA using real-time reverse
transcription polymerase chain reaction assays. J Mol Endocrinol 25:
169–196, 2000.
11. Chandel NS, Vander Heiden MG, Thompson CB, and Schumacker
PT. Redox regulation of p53 during hypoxia. Oncogene 19: 3840–3848,
2000.
96 • FEBRUARY 2004 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.5 on June 18, 2017
Microarray analysis requires validation by a second methodology. Northern blots (traditional and virtual), SAGE, differential display, and quantitative end-point RT-PCR have all
been widely used to validate microarray results. We have used
Northern and Western blot analyses to confirm the hypoxic
regulation of MKP-1 (69). In addition, the relatively new
technique of quantitative real-time RT-PCR is a valuable tool
for large-scale analysis of microarray results (reviewed in Ref.
10). The SYBR Green I dye intercalation method is particularly well-suited for this purpose because it does not require
special primers and is therefore far less time consuming and
expensive than other real-time PCR methods. Primers are
designed to produce a short amplicon (⬃80–100 bp), and the
amount of product is measured in each cycle by measuring the
fluorescence emitted by dye bound to the double-stranded
DNA. A threshold cycle (Ct) is generated by measuring either
the cycle in which the growth curve crosses a predetermined
threshold or the cycle in which the growth curve enters the log
linear phase. The Ct is related to the relative abundance of a
particular gene in the original RNA sample. A lower Ct in one
sample relative to another indicates a greater abundance of the
gene in that sample. Calibration curves can be generated to
relate a change in Ct value to a change in gene expression.
Although absolute quantification is rarely necessary, it is also
possible by generating standard curves consisting of samples
containing known amounts of mRNA for the gene of interest.
Specificity is determined by analysis of the melting point of the
PCR product. It is also desirable to verify the sequence of the
PCR product. Figure 4 shows typical real-time PCR growth
curves for a gene that is upregulated by hypoxia in PC12 cells
(TH) and a gene that is not regulated (GAPDH). Preliminary
analysis of 55 genes isolated in our SSH library indicates a
75% agreement between real-time PCR and microarray results
(Table 2).
771
Invited Review
772
GENOMICS OF HYPOXIA SIGNALING
J Appl Physiol • VOL
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
ney: suppression of HIF-mediated gene expression by HIF-3␣. Biochem
Biophys Res Commun 287: 808–813, 2001.
Hegde P, Qi R, Abernathy K, Gay C, Dharap S, Gaspard R, Hughes
JE, Snesrud E, Lee N, and Quackenbush J. A concise guide to cDNA
microarray analysis. Biotechniques 29: 548–562, 2000.
Hochachka PW, Buck LT, Doll CJ, and Land SC. Unifying theory of
hypoxia tolerance: molecular/metabolic defense and rescue mechanisms
for surviving oxygen lack. Proc Natl Acad Sci USA 93: 9493–9498, 1996.
Huang LE, Gu J, Schau M, and Bunn HF. Regulation of hypoxiainducible factor 1 is mediated by an O2-dependent degradation domain via
the ubiquitin-proteasome pathway. Proc Natl Acad Sci USA 95: 7987–
7992, 1998.
Imbert V, Rupec RA, Livolsi A, Pahl HL, Traenckner EB, MuellerDieckmann C, Farahifar D, Rossi B, Auberger P, Baeuerle PA, and
Peyron JF. Tyrosine phosphorylation of I␬B-␣ activates NF-␬B without
proteolytic degradation of I␬B-␣. Cell 86: 787–798, 1996.
Kerr MK and Churchill GA. Experimental design for gene expression
microarrays. Biostatistics 2: 183–201, 2001.
Kerr MK and Churchill GA. Statistical design and the analysis of gene
expression microarray data. Genet Res 77: 123–128, 2001.
Kilbourne EJ, Nankova BB, Lewis EJ, McMahon A, Osaka H, Sabban
DB, and Sabban EL. Regulated expression of the tyrosine hydroxylase
gene by membrane depolarization. Identification of the responsive element
and possible second messengers. J Biol Chem 267: 7563–7569, 1992.
Kobayashi S, Beitner-Johnson D, Conforti L, and Millhorn DE.
Chronic hypoxia reduces adenosine A2A receptor-mediated inhibition of
calcium current in rat PC12 cells via downregulation of protein kinase A.
J Physiol 512.2: 351–363, 1998.
Kobayashi S, Conforti L, Pun RYK, and Millhorn DE. Adenosine
modulates hypoxia-induced responses in rat PC12 cells via the A2A
receptor. J Physiol 508.1: 95–107, 1998.
Kobayashi S and Millhorn DE. Regulation of N-methyl-D-aspartate
receptor expression and N-methyl-D-aspartate-induced cellular response
during chronic hypoxia in differentiated rat PC12 cells. Neuroscience 101:
1153–1162, 2000.
Kobayashi S and Millhorn DE. Stimulation of expression for the adenosine A2A receptor gene by hypoxia in PC12 cells. A potential role in cell
protection. J Biol Chem 274: 20358–20365, 1999.
Koong A, Chen EY, and Giaccia AJ. Hypoxia causes the activation of
nuclear factor ␬B through the phosphorylation of I␬B␣ on tyrosine
residues. Cancer Res 54: 1425–1430, 1994.
Liang P and Pardee AB. Differential display of eukaryotic messenger
RNA by means of the polymerase chain reaction. Science 257: 967–971,
1992.
Lopez-Barneo J. Oxygen-sensing ion channels and the regulation of
cellular functions. Trends Neurosci 19: 435–440, 1996.
Lopez-Barneo J, Lopez-Lopez JR, Urens J, and Gonzalez C. Chemotransduction in the carotid body: K⫹ current modulated by PO2 in the type
I chemoreceptor cells. Science 241: 580–582, 1988.
Michiels C, Minet E, Michel G, Mottet D, Piret JP, and Raes M. HIF-1
and AP-1 cooperate to increase gene expression in hypoxia: role of MAP
kinases. IUBMB Life 52: 49–53, 2001.
Millhorn DE, Norris M, and Raymond R. Stimulation of tyrosine
hydroxylase gene transcription during hypoxia: Regulation by cis-trans
interactions. In: Stress: Molecular Genetic and Neurobiological Advances,
edited by McCarty R, Aguilera G, Sabban E, and Kvetnansky R. New
York: Gordon and Breach Science, 1996, p. 629–645.
Millhorn DE, Raymond R, Conforti L, Zhu W, Beitner-Johnson D,
Filisko T, Genter MB, Kobayashi S, and Peng M. Regulation of gene
expression for tyrosine hydroxylase in oxygen sensitive cells by hypoxia.
Kidney Int 51: 527–535, 1997.
Minet E, Michel G, Mottet D, Piret JP, Barbieux A, Raes M, and
Michiels C. c-JUN gene induction and AP-1 activity is regulated by a
JNK-dependent pathway in hypoxic HepG2 cells. Exp Cell Res 265:
114–124, 2001.
Mishra RR, Adhikary G, Simonson MS, Cherniack NS, and Prabhakar NR. Role of c-fos in hypoxia-induced AP-1 cis-element activity
and tyrosine hydroxylase gene expression. Brain Res Mol Brain Res 59:
74–83, 1998.
Niitsu Y, Hori O, Yamaguchi A, Bando Y, Ozawa K, Tamatani M,
Ogawa S, and Tohyama M. Exposure of cultured primary rat astrocytes
to hypoxia results in intracellular glucose depletion and induction of
glycolytic enzymes. Brain Res Mol Brain Res 74: 26–34, 1999.
96 • FEBRUARY 2004 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.5 on June 18, 2017
12. Conforti L, Bodi I, Nisbet JW, and Millhorn DE. O2-sensitive K⫹
channels: role of the Kv1.2-subunit in mediating the hypoxic response.
J Physiol 524: 783–793, 2000.
13. Conforti L and Millhorn DE. Selective inhibition of a slow-inactivating
voltage-dependent K⫹ channel in rat PC12 cells by hypoxia. J Physiol
502: 293–305, 1997.
14. Conrad PW, Conforti L, Kobayashi S, Beitner-Johnson D, Rust RT,
Yuan Y, Kim HW, Kim RH, Seta K, and Millhorn DE. The molecular
basis of O2-sensing and hypoxia tolerance in pheochromocytoma cells.
Comp Biochem Physiol B 128: 187–204, 2001.
15. Conrad PW, Freeman TL, Beitner-Johnson D, and Millhorn DE.
EPAS1 trans-activation during hypoxia requires p42/p44 MAPK. J Biol
Chem 274: 33709–33713, 1999.
16. Conrad PW, Rust RT, Han J, Millhorn DE, and Beitner-Johnson D.
Selective activation of p38␣ and p38␥ by hypoxia. Role in regulation of
cyclin D1 by hypoxia in PC12 cells. J Biol Chem 274: 23570–23576,
1999.
17. Czyzyk-Krzeska MF, Bayliss DA, Lawson EE, and Millhorn DE.
Regulation of tyrosine hydroxylase gene expression in the rat carotid body
by hypoxia. J Neurochem 58: 1538–1546, 1992.
18. Czyzyk-Krzeska MF, Furnari BA, Lawson EE, and Millhorn DE.
Hypoxia increases rate of transcription and stability of tyrosine hydroxylase mRNA in pheochromocytoma (PC12) cells. J Biol Chem 269: 760–
764, 1994.
19. Delpiano MA and Hescheler J. Evidence for a PO2 sensitive K⫹ channel
in the type-I cell of the rabbit carotid body. FEBS Lett 249: 195–198,
1989.
20. Diatchenko L, Chenchik A, and Siebert P. Suppression subtractive
hybridization: a method for generating subtracted cDNA libraries starting
from poly(A⫹) or total RNA. In: RT-PCR Methods for Gene Cloning and
Analysis, edited by Siebert P and Larrick J. Boston, MA: BioTechniques
Books, 1998, p. 213–239.
21. Diatchenko L, Lau YFC, Campbell AP, Chenchik A, Moqadam F,
Huang B, Lukyanov S, Lukyanov K, Gurskaya N, Sverdlov ED, and
Siebert PD. Suppression subtractive hybridization: a method for generating differentially regulated or tissue-specific cDNA probes and libraries.
Proc Natl Acad Sci USA 93: 6025–6030, 1996.
22. DiDonato J, Mercurio F, Rosette C, Wu-Li J, Suyang H, Ghosh S, and
Karin M. Mapping of the inducible I␬B phosphorylation sites that signal
its ubiquitination and degradation. Mol Cell Biol 16: 1295–1304, 1996.
23. Evinger MJ, Cikos S, Nwafor-Anene V, Powers JF, and Tischler AS.
Hypoxia activates multiple transcriptional pathways in mouse pheochromocytoma cells. Ann NY Acad Sci 971: 61–65, 2002.
24. Fishman M, Greene WL, and Platika D. Oxygen chemoreception by
carotid body cells in culture. Proc Natl Acad Sci USA 82: 1448–1450,
1985.
24a.Fitzpatrick TE and Graham CH. Stimulation of plasminogen activator
inhibitor-1 expression in immortalized human trophoblast cells cultured
under low levels of oxygen. Exp Cell Res 245: 155–162, 1998.
25. Ghafar MA, Anastasiadis AG, Chen MW, Burchardt M, Olsson LE,
Xie H, Benson MC, and Buttyan R. Acute hypoxia increases the
aggressive characteristics and survival properties of prostate cancer cells.
Prostate 54: 58–67, 2003.
26. Gonzales CY, Kwok Y, Gibb J, and Fidone S. Effects of hypoxia on
tyrosine hydroxylase activity in rat carotid body. J Neurochem 33: 713–
719, 1979.
27. Graeber TG, Peterson JF, Tsai M, Monica K, Fornace AJ Jr, and
Giaccia AJ. Hypoxia induces accumulation of p53 protein, but activation
of a G1-phase checkpoint by low-oxygen conditions is independent of p53
status. Mol Cell Biol 14: 6264–6277, 1994.
28. Green LA and Tischler AS. PC12 pheochromocytoma cultures in neurobiological research. Adv Cell Neurobiol 3: 373–414, 1982.
29. Gu YZ, Moran SM, Hogenesch JB, Wartman L, and Bradfield CA.
Molecular characterization and chromosomal localization of a third alphaclass hypoxia inducible factor subunit, HIF3␣. Gene Expr 7: 205–213,
1998.
30. Gurskaya NG, Diatchenko L, Chenchik A, Siebert PD, Khaspekov
GL, Lukyanov KA, Vagner LL, Ermolaeva OD, Lukyanov SA, and
Sverdlov ED. Equalizing cDNA subtraction based on selective suppression of polymerase chain reaction: cloning of Jurkay cell transcripts
induced by phytohemaglutinin and phorbol 12-myristate 13-acitate. Anal
Biochem 240: 90–97, 1996.
31. Hara S, Hamada J, Kobayashi C, Kondo Y, and Imura N. Expression
and characterization of hypoxia-inducible factor (HIF)-3␣ in human kid-
Invited Review
GENOMICS OF HYPOXIA SIGNALING
J Appl Physiol • VOL
72. Shaw K, Montague W, and Yoshizaki K. Biochemical studies on the
release of dopamine from the rat carotid body. In Vitro 1013: 42–46, 1989.
73. Shi Q, Le X, Abbruzzese JL, Wang B, Mujaida N, Matsushima K,
Huang S, Xiong Q, and Xie K. Cooperation between transcription factor
AP-1 and NF-␬B in the induction of interleukin-8 in human pancreatic
adenocarcinoma cells by hypoxia. J Interferon Cytokine Res 19: 1363–
1371, 1999.
74. Shweiki D, Itin A, Soffer D, and Keshet E. Vascular endothelial growth
factor induced by hypoxia may mediate hypoxia-initiated angiogenesis.
Nature 359: 843–845, 1992.
75. Singh S, Darnay BG, and Aggarwal BB. Site-specific tyrosine phosphorylation of I␬B␣ negatively regulates its inducible phosphorylation
and degradation. J Biol Chem 271: 31049–31054, 1996.
75a.Sowter HM, Ratcliffe PJ, Watson P, Greenberg AH, and Harris AL.
HIF-1-dependent regulation of hypoxic induction of the cell death factors
BNIP3 and NIX in human tumors. Cancer Res 61: 6669–6673, 2001.
76. Srinivas V, Zhang LP, Zhu XH, and Caro J. Characterization of an
oxygen/redox-dependent degradation domain of hypoxia-inducible factor
␣ (HIF-␣) proteins. Biochem Biophys Res Commun 260: 557–561, 1999.
77. Tabata M, Tarumoto T, Ohmine K, Furukawa Y, Hatake K, Ozawa
K, Hasegawa Y, Mukai H, Yamamoto M, and Imagawa S. Stimulation
of GATA-2 as a mechanism of hydrogen peroxide suppression in hypoxiainduced erythropoietin gene expression. J Cell Physiol 186: 260–267,
2001.
78. Tang TTL and Lasky LA. The forkhead transcription factor FOXO4
induces the down-regulation of hypoxia-inducible factor 1 ␣ by a von
Hippel-Lindau protein-independent mechanism. J Biol Chem 278: 30125–
30135, 2003.
79. Velculescu VE, Zhang L, Vogelstein B, and Kinzler KW. Serial
analysis of gene expression. Science 270: 484–487, 1995.
80. Weir EK and Archer SL. The mechanism of acute hypoxic pulmonary
vasoconsctiction, the tale of two channels. FASEB J 9: 183–189, 1995.
81. Wisdom R. AP-1: one switch for many signals. Exp Cell Res 253:
180–185, 1999.
82. Wu H and Lozano G. NF␬B activation of p53. A potential mechanism for
suppressing cell growth in response to stress. J Biol Chem 269: 20067–
20074, 1994.
83. Xu L, Xie K, Mukaida N, Matsushima K, and Fidler IJ. Hypoxiainduced elevation in interleukin-8 expression by human ovarian carcinoma
cells. Cancer Res 59: 5822–5829, 1999.
84. Yamashita K, Discher DJ, Hu J, Bishopric NH, and Webster KA.
Molecular regulation of the endothelin-1 gene by hypoxia. Contributions
of hypoxia-inducible factor-1, activator protein-1, GATA-2, AND p300/
CBP. J Biol Chem 276: 12645–12653, 2001.
85. Zhang HM, Cheung P, Yanagawa B, McManus BM, and Yang DC.
BNips: a group of pro-apoptotic proteins in the Bcl-2 family. Apoptosis 8:
229–236, 2003.
86. Zhu H and Bunn F. Oxygen sensing and signaling: impact on the
regulation of physiologically important genes. Respir Physiol 115: 239–
247, 1999.
87. Zhu WH, Conforti L, Czyzyk-Krzeska MF, and Millhorn DE. Membrane depolarization in PC12 cells during hypoxia is regulated by an
O2-sensitive K⫹ current. Am J Physiol Cell Physiol 271: C658–C665,
1996.
88. Zon LI, Youssoufian H, Mather C, Lodish HF, and Orkin SH.
Activation of the erythropoietin receptor promoter by transcription factor
GATA-1. Proc Natl Acad Sci USA 88: 10638–10641, 1991.
96 • FEBRUARY 2004 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.5 on June 18, 2017
53. Norris ML and Millhorn DE. Hypoxia-induced protein binding of
O2-responsive sequences on the tyrosine hydroxylase gene. J Biol Chem
270: 23774–23779, 1995.
54. Park SK, Dadak AM, Haase VH, Fontana L, Giaccia AJ, and Johnson
RS. Hypoxia-induced gene expression occurs solely through the action of
hypoxia-inducible factor 1␣ (HIF-1␣): role of cytoplasmic trapping of
HIF-2␣. Mol Cell Biol 23: 4959–4971, 2003.
55. Prabhakar NR, Shenoy BC, Simonson MS, and Cherniack NS. Cell
selective induction and transcriptional activation of immediate early genes
by hypoxia. Brain Res 697: 266–270, 1995.
56. Quackenbush J. Computational analysis of microarray data. Nature Rev
Genetics 2: 418–427, 2001.
57. Ravi R, Mookerjee B, Bhujwalla ZM, Sutter CH, Artemov D, Zeng Q,
Dillehay LE, Madan A, Semenza GL, and Bedi A. Regulation of tumor
angiogenesis by p53-induced degradation of hypoxia-inducible factor 1␣.
Genes Dev 14: 34–44, 2000.
58. Renton A, Llanos S, and Lu X. Hypoxia induces p53 through a pathway
distinct from most DNA-damaging and stress-inducing agents. Carcinogenesis 24: 1177–1182, 2003.
59. Royds JA, Dower SK, Qwarnstrom EE, and Lewis CE. Response of
tumour cells to hypoxia: role of p53 and NF␬B. Mol Pathol 51: 55–61,
1998.
60. Sansome C, Zaika A, Marchenko ND, and Moll UM. Hypoxia death
stimulus induces translocation of p53 protein to mitochondria. Detection
by immunofluorescence on whole cells. FEBS Lett 488: 110–115, 2001.
61. Schena M, Shalon D, Davis RW, and Brown PO. Quantitative monitoring of gene expression patterns with a complementary DNA microarray. Science 270: 467–470, 1995.
62. Scortegagna M, Morris MA, Oktay Y, Bennett M, and Garcia JA. The
HIF family member EPAS1/HIF-2␣ is required for normal hematopoiesis
in mice. Blood 102: 1634–1640, 2003.
63. Semenza GL. HIF-1 and human disease: one highly involved factor.
Genes Dev 14: 1983–1981, 2000.
64. Semenza GL. HIF-1, O2, and the 3 PHDs: how animal cells signal
hypoxia to the nucleus. Cell 107: 1–3, 2001.
65. Semenza GL, Jiang BH, Leung SW, Passantino R, Concordet JP,
Maire P, and Giallongo A. Hypoxia response elements in the aldolase A,
enolase 1, and lactate dehydrogenase A gene promoters contain essential
binding sites for hypoxia-inducible factor 1. J Biol Chem 271: 32529–
32537, 1996.
66. Semenza GL, Roth PH, Fang HM, and Wang GL. Transcriptional
regulation of genes encoding glycolytic enzymes by hypoxia-inducible
factor 1. J Biol Chem 269: 23757–23763, 1994.
67. Semenza GL and Wang GL. A nuclear factor induced by hypoxia via de novo
protein synthesis binds to the human erythropoietin gene enhancer at a site
required for transcriptional activation. Mol Cell Biol 12: 5447–5454, 1992.
68. Seta KA, Ferguson TK, and Millhorn DE. Discovery of oxygen responsive genes in pheochromocytoma cells. In: Methods in Enzymology. In
press.
69. Seta KA, Kim R, Kim HW, Millhorn DE, and Beitner-Johnson D.
Hypoxia-induced regulation of MAPK phosphatase-1 as identified by
subtractive suppression hybridization and cDNA microarray analysis.
J Biol Chem 276: 44405–44412, 2001.
70. Seta KA, Spicer Z, Yuan Y, Lu G, and Millhorn DE. Responding to
hypoxia: lessons from a model cell line. Science’s STKE; http://www.
stke.org/cgi/content/full/sigtrans;2002/146/rell, 2002.
71. Shalon D, Smith SJ, and Brown PO. A DNA microarray system for
analyzing complex DNA samples using two-color fluorescent probe hybridization. Genome Res 6: 639–645, 1996.
773