ICANCER
RESEARCH
54, 5273—5279,
October15, 19941
Advances in Brief
Hypoxic Activation of Nuclear Factor-icB Is Mediated by a Ras and Raf Signaling
Pathway and Does Not Involve MAP Kinase (ERK1 or ERK2)1
Albert C. Koong, Eunice Y. Chen, Nahid F. Mivechi, Nicholas C. Denko, Peter Stambrook,
and Amato J. Giaccia2
Cancer Biology Research Laboratory, Department of Radiation Oncology, Stanford University School of Medicine, Stanfor4 California 94305-5468 (A. C. K., E. Y. C., N. F. M.,
A. J. G.J and Department ofAnatomy and Cell Biology, University of Cincinnati School ofMedicine, Cincinnati. Ohio 45221 [N. C. D., P. 1. 5.1
MEK (MAP@-kinasekinase) (8), which then phosphorylates and ac
Abstract
We have previously shown that hypoxia causes the activation of nuclear
factor-.cB
(NF-icB), and the phosphorylation
of Its inhibitory
subunit,
IicBa, on tyrosine residues.With the use of dominant negativemutants of
Ha-Ras
and Raf-1, we investigated
some of the early signaling
events
leading to the activation of NF-.cB by hypoxia. Both dominant negative
alleles of Ha-Ras and Real-i inhibited NF-.cB induction by hypoxia, sug
gaMingthat the hypoxia-inducedpathway of NF-.cBInduction Is depend
ent on Ras and Raf-1 klnase activity. Furthermore, although conditions of
low oxygen can also activate mitogen-activated protein kinases (ERK1 and
ERK2), these kinases do not appear to be involved in regulating NF-icB by
tivates MAP kinase via tyrosine and threonine phosphorylation (9).
Therefore, constitutive activation of Ha-ras or Raf-1 permits the cell
to escape the regulation of normal growth factor control. In addition
to growth factor regulation, MAP kinases have been shown to be
induced by cellular exposure to free radical generating agents such as
ionizing radiation, hydrogen peroxide, and UV light, suggesting that
MAP kinases also respond to nongrowth factor-mediated signals as
well (10, 11). The link between Ras and Raf-1 in the MAP kinase
activation pathway has been genetically demonstrated through the use
of oncogenic and dominant inhibitory mutants of Ras and Raf-1 (12).
lowoxygenconditions,as dominant negativemutants of mitogen-activated
We have previously reported that the transcription factor, NF-,cB,
protein kinase do not inhibit NF-.cB activation by hypoxia. Since Ras and
was activated during conditions of low oxygen or hypoxia (13). The
Raf-1 have been previouslyshownto work downstream from membrane
translocation of NF-tcB from the cytoplasm to the nucleus during
associated tyrosine kinases such as Src, we determined If the Src mem
exposure to low oxygen conditions was enhanced by the phosphoryl
brane-assoclated
kinase was also activated by low oxygen conditions. We
detected an increase in Src proto-oncogene
activity within 15-30 mm of
cellular exposure to hypoxia. We postulate that Src activation by hypoxia
may be one of the earliest events that precedes Ran activation in the
signaling cascade which ultimately leads to the phosphorylatlon and
dissocintion
of the inhibitory
subunit
of NF-@cB, I.cBa@
ation of its cytoplasmic inhibitory subunit, I.cBa, on tyrosine residues
and the subsequent degradation of I.cBa. The dissociation and deg
radation of I.cBa from NF-KB permit NF-scB to translocate from the
cytoplasm to the nucleus where it functions to enhance transcriptional
activation of a variety of genes. Using a combination of immunopre
cipitation and immunoblotting techniques, we found that the kinetics
of tyrosine phosphorylation of IKBa preceded its degradation. If cells
Introduction
The Ras family of proto-oncogenes encodes a set of guanine
nucleotide-binding proteins (GTP-binding proteins) that are responsi
ble for mediating a variety of signal transduction cascades linking
events at the cellular membrane from receptor tyrosine kinases or the
Src kinase to transcriptional control of early response genes in the
nucleus (See Refs. 1—3,for reviews). In addition to their role in
regulating different aspects of normal and transformed cell growth,
members of this superfamily are involved in cytoskeletal organization
and cellular trafficking between different compartments of the cell.
The stimulation of Ras GDP-GTP exchange and subsequent activation
of Ran is mediated by the recruitment of the Ras exchange factor SOS
to the plasma membrane by the GRB2 adapter protein. Since Ras
proteins have low intrinsic GTPase activity, their inactivation in vivo
is dependent on GTPase-activating proteins (1—3).
Oncogenic mutants of Ras proteins at positions 12, 13, or 61 are
resistant to GTPase-activating protein-mediated GTPase stimulation
causing Ras to be permanently locked in its GTP-bound or active state
(4, 5). Activated Ras triggers a kinase cascade involving the senne
threonine kinase Raf-1 (6, 7) which phosphorylates
were pretreated with tyrosine kinase inhibitors before hypoxia, the
degradation and tyrosine phosphorylation of I.cBa was inhibited,
suggesting that tyrosine phosphorylation of IscBa is important in the
targeting of I.cBa for degradation. Transfection of a dominant inhib
itory allele of Raf-1 in Jurkat T cells blocked the degradation of IicBa
by hypoxia, indicating that Raf-1 was part of a specific hypoxia
induced signaling pathway leading to the activation of NF-.cB (13).
Although one previous study by Li and Sedivy (14) suggests that
Raf-1 directly phosphorylates IscBa, we believe that their result is due
to the use of a Raf-1 mutant, which lacks a regulatory domain, thereby
permitting Raf-1 to non-specifically phosphorylate substrates it would
not phosphorylate if its regulatory domain was intact. Furthermore,
since Raf-1 is a serine-threonine kinase and not a tyrosine kinase, it
could not be the kinase that phosphorylates IicBa during hypoxia. In
contrast to phosphorylation of IscBa on tyrosine residues by hypoxia,
cells treatedwith phorbol esters or 1'NF exhibited no detectable tyrosine
phosphorylation,further suggesting that different stresses may modulate
the activation of NF-@cBby multiple pathways downstream of Raf-1.
In these studies, we investigated whether the activation of NF-KB
by hypoxia is dependent on Ran and Raf-1 activity. To examine this
relationship, we stably overexpressed dominant negative alleles Ha
and activates
rasNl7 and raf3Ol in NIH3T3 cells before exposing them to hypoxia
and phorbol esters. Since membrane-associated kinases seemed to be
Received 8/15/94; accepted 8/31/94.
The costs of publicationof this articlewere defrayedin partby the paymentof page
charges. This article must therefore be hereby marked advertisement in accordance with
18 U.S.C. Section 1734 solely to indicatethis fact.
1 This
work
supported
by NIH
Grants
CA03353
(A.
J. G.),
CA54093
(N.
F. M.),
P30
3 The
ES06096 (P. S.), and a NIH Predoctoral Training Grant (A. C. K., N. C. D.).
2 To
whom
requests
for
reprints
should
be
addressed,
at Department
of
important in signaling I.cBa for degradation, we investigated the
Radiation
Oncology, Division of Radiation Biology, Stanford University School of Medicine,
Stanford, CA 94305-5468.
abbreviations
used
are:
MAP,
mitogen-activated
protein;
TNF,
tumor
necrosis
factor; FCS, fetal calf serum; TPCK, tosyl phenylalanine chloromethyl ketone; TLME,
tosyl lysine methyl ester; PMA, phorbol 12-myristate 13-acetate; IPTG, isopropyl-1-thio
fJ-o-galactopyranoside; ECL, enhanced chemilluminescence; NF-icB, nuclear-factor-KB;
I@cBa,cytoplasmic inhibitory subunit of NF-scB.
5273
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1994 American Association for Cancer Research.
PAThWAY FOR NF-sB ACFIVATION BY HYPOXIA
activity of Src kinase under low oxygen conditions and compared it to
other stimulators of NF-KB activity. We have found that Src kinase
activation is the earliest event in the cellular response to hypoxia. We
also sought to determine whether hypoxia activates MAP kinases
(ERK1, ERK2), and again through the use of dominant negative
mutants, we sought to determine whether MAP kinase was involved
in the pathway
that leads to NF-KB activation
by low oxygen
condi
tions. Finally, we show the importance of IicBa degradation on the
activation of NF-KB by hypoxia through the use of chymotrypsin-like
serine protease inhibitor (TPCK) and a trypsin-like protease inhibitor
(TLME). In conclusion, it seems that hypoxia induces a series of
membrane-associated kinases, which by direct phosphorylation of
IKBct,causethe dissociationof IKBa from NF-scBwhich thensignals
it for degradation by a serine-like protease.
experimentswere performedat leastthreetimes.Littlefluctuationwas seen among
experiments
Immunoblot
Analysis.
Cells were lysed in a buffer containing
137 mM
NaCl, 20 mMTris-HC1 (jH 8.0), 2 mM EDTA, 1 mM phenylmethylsulfonyl
fluoride, 1 mM sodium orthovanadate, 100 nM okadaic acid, 0.4% Nonidet
P-40, 10% glycerol, 10 @g/mlleupeptin, 1 @g/mlaprotinin, and 1 mg/mI
pepstatin by sonication with 4.5-s pulses of 75 W while they were on ice. To
separate nuclear and cytoplasmic extracts, cells were lysed in the above
mentioned buffer and incubated for 1 h on ice. Nuclei were pelleted by
centrifugation at 14,000 X g for 20 mm at 4°C.Protein concentrations of each
sample were estimated by the bicinchinonic acid method (Rockford, IL).
Immunoblots were performed as described previously (13). Briefly, cell cx
tracts were mixed in an equal volume
with 2X sodium
dodecyl
sulfate
sample
buffer,denaturedby heatingto 97°Cfor 5 mm, and separatedby polyacryl
amide gel electrophoresis.
The gel was then transferred onto Hybond ECL
(Amersham, Arlington Heights, IL). The gels were stained with 0.15% Coo
massie blue to ensure a uniform transfer of proteins to the membrane. The
membranes were then probed with either l.cBa antibody (kindly provided by
Materials and Methods
CellCulture and Treatments. NIH3T3cellswereroutinelymaintainedin
WarnerGreene,Universityof California,San Francisco),p21 ras antibody
(SantaCruzBiotechnology,SantaCruz,CA), or ERK1antibody(K-23; Santa
Dulbecco's modified Eagle's medium supplemented with 10% FCS. NIH3T3
Cruz Biotechnology).
cells that were stably transfected
using ECL according to the protocol of the manufacturer (Amersham).
Immunopreclpitatlon. Immunoprecipitation studies were carried out as
with a Ha-ras-inducible
gene under the
control of a SVlacO-inducible promoter were cultured in Dulbecco's modified
Eagle's medium supplemented with 10% F@Sand 0.4 mg/ml of G418 antibi
otic to maintain selective pressure (15). To address the question of what early
Target proteins of these antibodies were then visualized
described previously (13). Briefly, ERK1 or bcBa antibody was incubated at
ras-inducible NIH3T3 cell lines with dominant negative alleles of Ha-ras
4°C
with proteinA beadsfor 16 h with constantmixing.The mixtureof beads
and antibody was washed and incubated with cell lysates while turning at 4°C
for 16 h. The iinmunoprecipitateswere then mixed with 2X sodium dodecyl
(rasNl7)
sulfate sample buffer and electrophoresed
events are necessary for the activation of NF-.cB, we stably transfected
and Raf-I
(raf3OI)
that were kindly
provided
by Dr. Channing
Der
(University of North Carolina, Chapel Hill, NC). Overexpression of these
mutants was assessed for functionality by testing for their ability to prevent
Ha-Ras-induced colony forming ability in soft agar. Jurkat T cells were
cultured in RPMI supplemented with 10% FCS. For studies investigating the
inhibition of NF-KB binding by hypoxia, cells were pretreated with either 25
@LMTPCK
(Sigma,
St
Louis,
MO)
or
25
pM
TLME
(Sigma)
before
on a 15% polyacrylamide
gel. The
immunoprecipitatedproteins were then subjected to Western blotting by trans
ferring to Hybond ECL and probing the membranes with antiphosphotyrosine
antibody (UBI Biotechnology, Lake Placid, NY).
Src Activity Assays. Immunoprecipitation of 150 p.g of cell lysates with
antibody specific to Src (kindly provided by Dr. Christine Cartwright, Stanford
University, Stanford, CA) was performed as described above. Next, the im
exposure
to hypoxia. TPCK is a chymotrypsin protease inhibitor, and liME is a trypsin
protease inhibitor. For Src activity measurements, cells were irradiated with a
254-nm UV lamp at a fluence rate of 0.75 J/m2/s,and treated with 10 g.@M
PMA
or 10 ng/ml TNF for various times before harvesting.
Hypoxia TreatmenL Cells were treated in 60-mm glass tissue culture
dishes (Corning, Inc.) with notched sides to allow gas exchange between the
medium in the dish and the environment (16). The dishes were placed in
specially designed aluminum hypoxia chambers that were prewarmed to 37°C,
sealed, and subjected to successive rounds of evacuation, followed by flushing
with 95% N2-5% CO2 while being slowly agitated on a reciprocating shaker.
The chambers were then placed in a 37°Cincubator and kept on a reciprocating
shaker for the duration of the hypoxic treatment. After one cycle of evacuation
and flushing to 2% oxygen, oxygen concentration in the media was reduced to
0.2%afterthreecycles,andto 0.02%afterfive cycles.Oxygenconcentration
munoprecipitates were incubated in a kinase buffer (30 s.d)consisting of 100
mM piperazine-N,N'-bis (2-ethane sulfonic acid), 100 mM dithiothreitol, 1 M
MnCl2,100m@ienolase,and 10 @CiJ,.&l
‘y[32P]ATP.
After incubationfor 15
min at room temperature, the reaction was stopped with the addition of 2X
sample buffer, and proteins were separated by polyacrylamide
gel electro
phoresis. Src and its substrate enolase were visualized by autoradiography, and
quantitation of src activity was done by AMBIS scanning (16).
Gel Mobifity Shift Assays. Gel shift studies were performed as described
previously (13). Briefly, 5—[email protected]
of nuclear extracts were prepared and
incubated in the presence of a radiolabeled
oligonucleotide
containing
the
NF-xB binding site from the HIV-LTR. The binding reaction was carried out
at room temperature for 20 mm in a buffer containing 500 ng salmon sperm,
10 mM 4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid (pH 7.8), 5 mM
MgCl2, 60 mti KCI, 0.5 mM dithiothreitol, 1 mM spermidine, and 10% glycerol
was measured with the use of a Clark-type electrode (Controls Katharobic,
in a reaction volume
Edmonton, Alberta, Canada). In the combined drug and hypoxia treatments,
cells were incubated for 0.5 h at 37°C
prior to the initiation of hypoxia. In some
experiments, hypoxic treatments were done in an anaerobic glove box to
perform all manipulations under low oxygen conditions (Macrobe Systems).
demonstrate the specificity of the binding. After electrophoresis on a 6%
polyacrylamide gel, the gel was fixed in 7% acetic acid, dried onto Whatman
Transient Transfection Studies. Two @i.g
ofreporter plasmid [fos-luciferase
which contains the human c-fos promotor (17) kindly provided by Dr. David
Brenner, University ofNorth Carolina, Chapel Hill, NC] were mixed with 8 gg of
expression vector alone (ERK1, ERK2, kindly provided by Dr. Melanie Cobb,
University of Texas Southwestern Medical Center, Dallas@DC) or with 4 @gof
expression vector and 4 @gof dominant negative mutants (KRErkJ, K52RErk2,
kindly provided by Dr. David Brenner) and added to 3 X 10@NIH3T3 cells. Cells
were electroporatedin 800 pi of RPM! media supplementedwith DEAE-dextran
ata concentration
of 5 @gJml
at roomtemperature
at250V with a capacitance
of
960 @F,
and then plated into tissueculturedishes.Transfectedcells were exposed
to ll@FGfor 24—36
h afterelectroporationto induceHa-Rasactivity,washedtwice
in phosphate-bufferedsaline, lysed in luciferaselysis buffer [0.1 Msodium phos
phate (pH 7.8)0.5% Triton-X 100-1 m@idithiothreitol],and then assayed for
luciferaseactivity using a luminometer.To eliminate differencesin transfection
paper,
and visualized
of 20 @.il.Competitors
by autoradiography
were added in a 50-fold
(Hyperfilm,
excess
to
Amersham).
Results and Discussion
Previous data from several groups have suggested the involvement
of cytoplasmic tyrosine kinases from the src family in the activation
of NE-KB(18, 19). To test whetherSrc kinase activationwas also an
early step in NF-aB activation by hypoxia, we examined the kinetics
of src activity in cells exposed to low oxygen conditions. Using
immune complex kinase assays, we determined Src activity in extracts
from equal numbers of untreated or hypoxia-treated cells by their
ability to phosphorylate enolase, a standard Src protein kinase sub
strate. A consistently reproducible 5—7-foldincrease in Src activity
was detected within 15—30mm of exposing cells to stringent (0.02%
oxygen) hypoxic conditions (Fig. 1, A and B). Since Src kinase
efficiencies,cellswerealsocotransfectedwith 1.0pg of pSV2CATplasmidto
activation may represent an early event in the NF-scB pathway by
normalizeagainst fluctuationsin plasmid uptake and expression.All transfection
different stresses, we compared Src kinase activity by hypoxia with
5274
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1994 American Association for Cancer Research.
PAThWAY
@
A
FOR NF-KB ACTIVATION BY HYPOXIA
, hypoxia
•
UVc ThFc@@TPA@
other stresses for similar periods of time that have been shown
previously to activate NF-KB (Fig. 1, A and B). From this and other
gels, we conclude that the activation of Src by hypoxia was similar in
magnitude to the activation of Src by UV light, greater in magnitude
to the activation of Src by TNF or phorbol esters, and possessed
slower kinetics. The upper bands in these gels represent the autophos
phorylation of src kinase. The activation of Src by hypoxia is the
earliest cellular response we have detected by this stress and precedes
the degradation of IKBa and the activation of NF-KB by hypoxia.
Ras and Raf-1 have previously been shown to be downstream
@
src-•.
- .@
effector kinases of Src in some cell lines and are activated by similar
agents that induce Src and NF-icB (18, 20—22).To investigate whether
—
hypoxia-induced IicBa degradation was mediated by Ran and Raf-1
enolase —@.@
kinase activity, we tested whether dominant negative alleles of Ha
Ras (Ha-rasNl7) and Raf-1 (raf-301) would block the activation of
NF-KB
by hypoxia
and phorbol
esters.
Using
a cell line in which
Ha-ras was conditionally inducible, we could directly test whether
increasing Ha-ras protein levels would increase IicBa degradation and
whether inhibiting Ha-ras activity would prevent IicBa degradation. In
these NIH3T3 cells, the activated Ha-ras gene is regulated by an
B
inducible SVlacO promoter that is kept inactive by the lac repressor.
When these cells are treated with IPTG, the lac repressor is sterically
hindered from binding to the lacO in the promoter, and Ha-ras is
transcriptionally induced. Unlike many inducible promoters, little
detectable Ha-ras activity is present in uninduced cells, which display
. Src-auto
@
no enhanced ability to grow in soft agar. In addition, we could also
Enolase
test the functional ability of dominant negative alleles of Ha-Ras or
6=
Raf-1 to inhibit transformation in these conditionally inducible
NIH3T3 cells when Ha-ras is activated. Fig. 14 is an immunoblot
showing the kinetics of Ha-ras protein induction as a function of time
exposed to 20 mt@iof the inducing agent IPTG. The kinetics of IicBa
degradation (Fig. 2B) and the increase in NF-KB binding (Fig. 2C)
0
U
I@
correlate with the kinetics of Ha-ras protein synthesis induced by
2
IPTG.
As further
evidence
that Ras is involved
in signaling
NF-.cB
activation, we compared NF-.cB binding activity in the untransfected
parental NIH3T3 cell line with a Ha-ras-transfected cell line that had
lost its ltic repressor, permitting constitutive Ha-ras expression. These
is
30
45
Time exposed to 0.02% oxygen (mm)
genetically matched “revertants―
overexpress Ha-ras and possess an
increased ability to form colonies in soft agar compared to their
wild-type parent cells. The conditional Ha-ras-inducible cell line that
lost its repressor activity displayed enhanced constitutive binding to the
KB element compared to parental wild-type
. [email protected].
cells. Taken together, both
inducible and constitutive Ha-ras expressions increase NF-icB activation.
I Enolase-TNF
U Enolase.-PMA
To investigate whether Ha-ras and Raf-1 activation are important in
the signaling for IscBa degradation by low oxygen conditions, we
investigated the kinetics of IKBa degradation by hypoxia and phorbol
esters in wild-type NIH3T3 cells stably transfected with dominant
0
4-I
U
negative alleles of these genes. RasNl7 (23) and raf3Ol (24) are
thought to function as dominant negative mutants by binding the same
C
substrates as their wild-type proteins, but due to a mutation that
inactivates their kinase activity, they are unable to phosphorylate their
0
U-
bound substrates. Thus, they are relatively selective in their inhibition
of kinase activity. However, although it is clear that Ha-rasNl7 acts
as a true dominant
5
10
negative
allele
of wild-type
Ha-ras,
it is still
possible that raJ3Ol is acting as a dominant negative allele by forming
Timeexposedto agent(mm)
a complex with Ha-Ras and, therefore, not directly inhibiting Raf-1
Fig. 1. Hypoxia stimulates Src activity. A, NIH3T3 cells that were exposed to hypoxia
for 15 and 30 mm, or 40 J/m2 UV, 10 ng/ml TNF, 10 ps@t12-O-tetradecanoylphorbol
13-acetate for 5 and 10 min before Src kinase activity assays. Arrows, the bands on the
autoradiogram corresponding to the phosphorylation of enolase and autophosphorylation
of Src. B, graphicrepresentationof Src activityin cells exposed to low oxygen, UV, 1'NF,
and TPA as described in A.
activity (7). Ha-ras-inducible
cells stably transformed
with Ha
rasNI7 and raf3Ol had 10—15%
of colony forming ability in soft agar
of Ha-ms-inducible cells that did not possess these dominant negative
alleles. When we tested the effects of overexpression of Ha-rasNi 7
and raf-301 dominant negative alleles on the inactivation and degra
dation of IKBa by hypoxia, we found that both dominant negative
alleles inhibited IKBa degradation by hypoxia (Fig. 3A). This data is
5275
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1994 American Association for Cancer Research.
@
@
t
.@
PAThWAY
FOR NF.sB ACTIVATION BY HYPOXIA
C
comp
m
ILQ.
@
@
C
HoursoflPlG Z
(I)
C
£
(‘I
£
l.-+
+
£
•@
ii ;@
.%°@
A
Fig. 2. Induction of Ras and NF-scBbinding in
@
@
Hours of IPTG
conditionally inducible NIH3T3 cells. A, kinetics
of Ha-ras induction as detected by immunoblot
.@
_ Nf-KB—ø@
. . S4—N
Ui
c.,@
analysis. B, kinetics of I,cBa protein degradation
and resynthesisafterHa-rasinduction.C, kinetics
of NF-icB binding as a function of inducible or
constitutive Ha-ras activity. comp, competitor.
ras-ø
Hours of IPTG
B
@
I
It)@
04
• •
IKBO—*'
—
in agreement with our previous studies showing that transient trans
fections of Jurkat T cells with raf3Ol also prevented the degradation
of IicBa during hypoxia (13). We also examined whether these same
dominant negative alleles could inhibit the activation of NF-scB by
phorbol esters, since no direct evidence has been presented to show
that the phorbol ester pathway involved in NP-KB activation is me
diated through Ras or Raf-1 kinases. IscBa degradation induced by
during
phorbol esters is also blocked in cells expressing Ha-rasNl7 and raf3Ol,
the phosphorylation
compared to wild-type NIH3T3 cells alone (Fig. 3B), suggesting that
these kinases are also important in NF-@cBactivation by this activator.
In many cell types, Ras and Raf-1 are part of a pathway that leads
to the activation of MAP kinase (ERK1, ERK2) (12, 25, 26). Since
Ras and Raf-1 are involved in the pathway signaling NF-icB activa
tion, next we examined whether the hypoxia-induced signaling path
way that leads to NF-KB activation involves MAP kinase or diverges
before the activation of MAP kinase. Transfection of dominant neg
ative alleles of ERK1 and ERK2 (27, 28) did not block the degrada
tion of IKBa by hypoxia, suggesting that neither ERK1 nor ERK2 are
involved in NF-KB activation (Fig. 4A). These results cannot defi
nitely rule out the possibility of effectors downstream of MAP kinase
acting in a circular or divergent pathway to activate NP-KB, but at
least they rule out a direct role for MAP kinase in NF-KB activation.
The dominant negative mutants of ERK1 and ERK2 used in these
A
B
@HoursofHypoxla@
0
kBa-*
.
C@
assays were functionally able to block wild-type ERK1- and ERK2-
induced transactivation of the c-fos promoter, a known MAP kinase
regulated gene (Fig. 4B).
Although hypoxia-induced activation of NF-icB did not appear to
involve either ERK1 or ERK2, we wanted to be certain that the
signaling pathway that activates MAP kinase was still functional
kinase
hypoxia.
To address
is activated
this question,
we assayed
for changes
in
status of ERK-1 during hypoxia. Since MAP
by phosphorylation
of its tyrosine
and threonine
residues (29), the active or phosphorylated form of MAP kinase
results in a protein that migrates more slowly than the unphosphory
lated form of the protein (Fig. 5A). To compare the kinetics of tyrosine
phosphorylation of ERK-1 and IKBa by hypoxia, we immunoprecipi
tated proteins from hypoxia-treated cells with antibodies specffic for
ERK1 or I,cBa, respectively. The immunoprecipitates were electro
phoresed and transferred to Hybond membranes, which were then
probed with an antiphosphotyrosine antibody. The results of these
studies suggest that MAP kinase (Fig. SB) and ItcBa (Fig. SC) are
both phosphorylated on tyrosine residues by hypoxia, and that this
tyrosine phosphorylation is inhibited by a specffic tyrosine kinase
inhibitor, herbimycin A. Thus, the pathway leading to MAP kinase
activation is activated by hypoxia, parallels the phosphorylation of
IKBa, but it is not involved in the activation of NF-icB.
Previous studies have shown that ItcBa degradation is important for
NF-@Bactivation (30). To examine whether cellular treatment with serine
protease inhibitors could prevent NF-scB activation by hypoxia, we pro
treated cells with serine protease inhibitors which have been shown
e
_
313.29
previously to inhibit IiBa degradation and NF-icB binding by TNF and
PMA. Hypoxia-induced binding to the NF-icB element is greafly reduced
raf@1
unaffected
in the presenceof a chymotrypsinpmteaseinhibitor, TPCK, and is
kBa'-@'
by a trypsin
protease
inhibitor,
TLME
(Fig. 6). From this
result and others, it seems that a serine protease is responsible for the
degradationof I.Ba by many stresses including low oxygen conditions.
@
kBa-*
—-
N17
These experiments suggest that the activation of NF-.cB by hypoxia is
Fig. 3. Effectof inhibitingRas andRaf-1activityon I.cBa degradationby low oxygen
conditions. A, kinetics of I.cBa degradation as a function of time exposed to hypoxia in
control NIH3T3.29 cells or in 3T3 cells stably transfected with dominant negative alleles
raf3Ol or rasNI7. B, effect of raf3OI and rasNl7 dominantnegativealleleson PMA
induced degradation of IKBa.
dependent on both the phosphorylationand degradation of I.cBa.
In this paper, we have shown that the degradation of I.cBa and
activation of NF-KB are mediated by a pathway that involves Ran and
Raf kinases and most probably, Src kinase. Furthermore, we also
5276
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1994 American Association for Cancer Research.
PAThWAY
A
FOR NF-xB ACI'IVATION BY HYPOXIA
the cytoplasm and to permit NF-KB nuclear translocation. Inhibition
of this phosphorylation inhibits I.cBa degradation by hypoxia. It
contrast, TNF and PMA treatment also signal for the dissociation of
Hrsof Hypoxia
C
@
IKBa—@°
,:
C@1
(f)
IicBa, but we could not detect any accumulation of phospho-tyrosine
residues on IKBa by either of these stresses. The link between
phosphorylation and degradation may be so rapid for these latter
stresses that the phosphorylated form oflicBa cannot be detected even
if it occurs, or that each stress induces multiple kinases that signal for
—
the dissociation of IKBa. This latter hypothesis is very attractive as it
could explain many conflicting reports about the role of phosphoryl
ation in IKBa degradation.
In vitro, numerous kinases, mainly of the serine and threonine
variety, have been shown to phosphorylate IKBa and cause its disso
ciation from the p65 subunit (14, 36). These results indicate that
B
I
5
(I)
tn
4)
phosphorylation of I,cBa, whether on serine/threonine residues or
tyrosine residues, is sufficient for dissociation. In vivo, the relation
ship of IKBa phosphorylation and degradation is still puzzling, as
inhibition of IscBa degradation by several stresses through the use of
unique chymotrypsin-like serine protease inhibitors will prevent its
degradation, presumably even though the molecule has been phos
phorylated (30). We speculate that phosphorylation is the signal for
‘D
U-
A
@
Hours
ofHypoxia
.i
AL
pCMV5
@
,-
ERK1
ERK2+
dom neg
ERKI+
dom neg
I
ERK-1 (
Fos-LuciferaseReporter
c,4
C.)
I
I
q
I
—
I
—
Fig. 4. Dominantnegative inhibitorymutantsof ERK1 and ERK2 do not block the
degradation ofbcBa by hypoxia. A, kinetics ofl.cBa degradation by hypoxia in NIH3T3
cellstransfectedandexpressingdominantnegativeallelesof ERK1(KRErKJ)andERK2
@
@
(K52RE-K2). B, demonstrationof functionalityof ERK1 and ERK2 dominant alleles on
fos-promoteractivation.The ERK1andERK2inhibitorymutantsweredemonstratedto be
functionally active by their ability to block the transactivation of afos-luciferase reporter
gene. Transfectionof wild-type ERK1 and ERK2 increases luciferase activity 6- and
Q
B
@‘.
4,
‘C.
‘C@
*
5'
Q
S
c@,
3-fold,respectively.Inthepresenceof thedominantnegativesof ERK1andERK2,c-fos
driven luciferaseactivity was reducedto controllevels.
show that MAP kinase is not involved in the activation of NF-KB by
hypoxia, which supports recent biochemical data by Diaz-Meco et a!.
ERK-1-+
@,
@..
.-/@_.
(31) that demonstrate that neither MAP nor MEK kinases could
phosphorylate IKBa in vitro. However, the activation of MAP kinase
by hypoxia could explain the increased transcriptional activity of the
c-fos proto-oncogene under low oxygen conditions (32).
The early events in the pathway for the inactivation of IKBa by
hypoxia possess slower kinetics than UV radiation TNF, and PMA
(13, 33). Although the latter three stresses signal for the inactivation
ofIKBa
@
presumably
through
a free radical
mechanism
that is inhibited
by radical scavengers (33), our preliminary evidence suggests that
hypoxic inactivation of I.cBa or increased NF-icB binding is not
C
inhibited by these scavengers.4 It is also noteworthy that hypoxia, UV
@
radiation, TNF, and PMA signal for the inactivation of IKBa through
Ranand Raf-1 kinases, and that Raf-1 kinase may activate divergent
kBct -0
.@
downstream kinases such as MEK (34) and I.cBa kinase depending on
the stress. Therefore, it seems that many pathways involved in the
activation of NF-KB are regulated by Raf-1 kinase. Differential reg
ulation of Raf-1 could be achieved through phosphorylation of spe
cific residues on the protein as exemplified by PKCa and the com
Fig. 5. Activationof ERK1 by hypoxia.A, hypoxia causes a shift in the mobility of
ERK1proteinas shownby immunoblotanalysis.B, detectionoftyrosine phosphorylation
on ERK1. InununoprecipitatingERK1 from hypoxia-treatedcells and then immunoblot
bination of Ran and Ick kinases (35).
Our previous data suggest that tyrosine phosphorylation of IKBGby
hypoxia
is necessary
to dissociate
it from the p65 subunit
of NF-KB in
ting the immunoprecipitates with an antiphosphotyrosine antibody shows an increase in
the phosphorylationof ERK1on tyrosineresidues.This phosphorylationis blockedin the
presence of the tyrosine kinase inhibitor, herbimycin A. C, I.cBa is phosphorylated
on
tyrosine residues by low oxygen conditions. Using the same combination of immunopre
cipitation and immunoblotting, we could detect similar kinetics of tyrosine phosphoryl
ation on IicBa as ERK1.
4 E. Y. Chen and A. I. Giaccia, unpublished results.
5277
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1994 American Association for Cancer Research.
PAThWAY
@
,_.
w
.j
I-.
E
0
+
0.
+
0.
0
a_
@
4)
FOR NF-sB ACtiVATION
2. Khosravi-Far,R., and Der, C. J. The ras signal transductionpathway. Cancer
Metastasis Rev., 13: 67—89,1994.
3. Boguski, M. S., and McCormick, F. Proteins regulating Ras and its relatives. Nature
(Lond.), 366: 643—654,1993.
4. Hall, A. The cellular function of small GTP-binding proteins. Science (Washington
U
>.
>.
>.
0.
+
0
z
=
z
I
BY HYPOXIA
DC), 249: 635—640, 1990.
5. Trahey, M., and McCormick, F. A cytoplasmic protein stimulates normal N-ras p21
>-
GTPase, but does not affect oncogenic mutants. Science (Washington DC), 238:
542—545,
1987.
6. Moodie, S. A., Willumsen, B. M., Weber, M. J., and Wolfman, A. Complexes of
ras-GTP with raf-1 and mitogen-activated protein kinase. Science (Washington DC),
260: 1658—1661,
1993.
7. Stokoe, D., Macdonald, S. G., Cadwallader, K., Symons, M., and Hancock, J. F.
Activation of Raf as a result of recruitment to the plasma membrane. Science
(Washington DC), 264: 1463-1467, 1994.
8. Crews, C. M., and Erikson, R. L Purification of a murine protein-tyrosine/threonine
kinase that phosphorylates and activates the Erk-1 gene product: relationship
to the fission yeast byrl gene product. Proc. NaIl. Aced. Sci. USA, 89: 8205—8209,
1992.
9. Anderson,N. G., Maller,J. L., Tonics,N. K., and Sturgill,T. W. Requirementon
integration of signals from two distinct phosphorylation
MAP kinase. Nature (Lond.), 343: 651—653, 1990.
pathways for activation of
10. Hibi, M., Lin, A., Smeal, T., Minden, A., and Karin, M. Identificationof an
oncoprotein-and uv-responsiveproteinkinase that binds and potentiatesthe c-jun
Nf-cB-*@
•
activation domain. Genes Dcv., 7: 2135—2148,1993.
11. Stevenson,M. A., Pollack,S. S., Coleman,C. N., andCalderwood,S. K X-irradiation,
phorbol esters, H202 stimulate mitogen-activatedprotein kinase activity in NIH3T3
cells through the formation of reactive oxygen intermediates. Cancer Res@,54: 12-15,
1994.
12. Howe, L R., Leevers, S. J., Gomez, N., Nakienly, S., Cohen, P., and Marshall, C J.
Activationofthe MAP kinasepathwayby the proteinkinasen4@.
Cell, 71: 335-342, 1992.
13. Koong, A. C., Chen, E. Y., and Giaccia, A. J. Hypoxia causes the activation of nuclear
factor-.cBthrough the phosphorylation oflacBa on tyrosine residues. Cancer Rca., 54:
1425—1430,1994.
14. Li, S., and Sedivy, J. M. Raf-1 protein kinase activates the NF-icB transcription factor
by dissociating the cytoplasmic NF-scB-IKB complex. Proc. NatI. Aced. Sci. USA, 90:
9247—9251,
1993.
15. Denko, N. C., Giaccia, A. J., Stringer, J. R., and Stambrook, P. J. The human Ha-ms
oncogene induces genomic instability in murine fibroblasts within one cell-cycle.
Proc. Natl. Acad. Sci. USA, 91: 5124-5128, 1994.
16. Giaccia,A. J., Auger, E. A., Koong, A., Terrace,D., Minchinton,A., Hahn,G. M.,
and Brown, J. M. Activation of the heat shock transcriptionfactor by hypoxia in
normal and tumor cell lines in vivo and in vitro. tnt. J. Radiat. Oncol. Biol. Phys., 23:
891—897,1992.
17. Rippe, R. A., Brenner, D. A., and Leffert, H. L DNA-mediated gene transfer into
adultrathepatocytesin primaryculture.Mol. Cell. Biol., 10: 689—695,1990.
18. Devary, Y., Rosette, C., DiDonato, J. A., and Karin, M. NF-.cB activation by
ultraviolet light not dependent on a nuclear signal. Science (Washington DC), 261:
Fr..
Probe
@
1442-1445, 1993.
19. Eicher,D. M., Tan,T.-H., Rice, N. R., O'Shea,J.J., andKennedy,I. C. S. Expression
of v-src in T cells correlateswith nuclearexpressionof Nf-.cB. J. Immunol.,152:
2710—2719,1994.
20. Kremer,N., D'Arcangelo,G., Thomas,S., DeMarco,M., Brugge,J., and Halegoua,
S. Signaltransductionby nervegrowthfactorandfibroblastgrowthfactorrequiresa
Fig. 6. The effect of proteaseinhibitorson NF-.cBactivation.DNA bindingactivityof
NF-KB by hypoxia is inhibited by a chymotrypsin protease inhibitor, TPCK A trypsin
proteaseinhibitor,ThME, had no effecton NF-,cBDNA bindingactivity.This work supports
studiesperformedby others,whichsuggestthata chymotrypsin-likeproteaseis responsiblefor
thedegradationof bcBa, andthus,the activationof NF-KB.comp,competitor.
sequence of Src and Ras actions. J. Cell Biol., 115: 809-819, 1991.
21. DeClue, J. E., Zhang, K., Redford, P., Vass, W. C., and Lowy, D. R. Suppression of
src transformation by overexpression of full length GTPase-activating protein (GAP)
or of the GAP C terminus.Mol. Cell. Biol., 11: 2819-2825, 1991.
22. Mulcahly, L. S., Smith, M. R., and Stacey, D. W. Requirement for ma protooncogene
function during serum-stimulated growth of NIH3T3 cells. Nature (Lond.), 313:
241—243,
1985.
23. Feig, L., and Cooper, G. Inhibition of NIH3T3 cell proliferation by a mutant ras
protein with preferential affinity for GDP. Mol. Cell. Biol., 8: 3235—3243,1988.
IscBa degradation in vivo, and that if degradation of IscBa is inhibited
by a senne protease inhibitor, then IKBa can be dephosphorylated and
reassociated to the p65 subunit of NF-icB. Future studies are directed
at testing this hypothesis.
Finally, the activation of Src, Ras, Raf-1, and MAP kinase by
hypoxia suggests a physiologicalrole for hypoxia in transformationand
genomic instability.
causes chromosome
The unscheduled induction of Ha-ras activity rapidly
instabifity within one cell cycle (15). Cellular expo
sure to hypoxia has also been reported to induce genomic instability as
manifested
by gene amplification
(37, 38). Thus, hypoxia could represent
a physiological mechanism that will specifically increase genomic insta
bility in a tumor by deregulating
the expression
24. Kolch, W., Heidecker, G., Uoyd, P., and Rapp, U. Raf.1 protein kinase is required for
growth of induced NIH3T3 cells. Nature (Lond.), 349: 426—428, 1991.
25. Williams, N. G., Paradis, H., Agarwal, S., Charest, D. L, Pelech, S. L, and Roberts,
T. M. Raf-1 and p2l@'°'cooperate in the activationof mitogen-activatedprotein
kinase. Proc. NaIl. Acad. Sci. USA, 90: 5772—5776, 1993.
26. Williams, N. G., and Roberts, T. Signal transduction pathways involving the Raf
protooncogene. Cancer and Metastasis Rev. 13: 105—1
16, 1994.
27. Robbins, D. J., Then, E., Owaki, H., Vanderbilt, C. A., Ebert, D., Geppert, T. D., and
Cobb, M. H. Regulation and properties of extracellular signal-regulated protein
kinases 1 and 2 in vitro. J. Biol. Chem., 268: 5097—5106, 1993.
28. Westwick, J. K., Cox, A. D., Ocr, C. J., Cobb, M. H., Hibi, M., Karin, M., and
Brenner, D. A. Oncogenic Ras activates c-Jun via a separate pathway from the
activation of extracellular signal-regulated kinases. Proc. NatI. Acad. Sci. USA, 91:
6030—6034,1994.
29. Anderson,N. G., Maller,J. L., Tonks, N. K., and Sturgill,T. W. Requirementfor
integration of signals from two distinct phosphorylation pathways for activation of
MAP kinase. Nature (Lond.), 343: 651—653, 1990.
30. Henkel, T., Machleidt, T., Akalay, I., Kronke, M., Ben-Neriah, Y., and Baeuerle,
of these kinases.
P. A. Rapidproteolysisof 1KB-ais necessaryfor activationof Nf-kB. Nature(Land.),
365: 182—185,1993.
References
31. Diaz-Meco, M. T., Dominguez, I., Sans, L, Dent, P., Lozano, I., Municio, M. M.,
I. Bourne, H. R., Sanders, D. A., and McCormick, F. The GTPase superfamily:
a conservedswitch for diversecell functions.Nature(Lond.), 348: 125—132,
1990.
Berra, E., Hay, R. T., Strugill, T. W., and Moscat, J. @PKCinduces phosphorylation
and inactivationof I,cBcrin vitro. EMBO, 13: 2842—2848,
1994.
5278
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1994 American Association for Cancer Research.
PATHWAY FOR NF-icB ACflVATION BY HYPOXIA
32. Ausserer, W. A., Bourrat-Floeck, B., Green, C. J., Laderoute, K. R., and Sutherland,
R. M. Regulationof c-junexpressionduringhypoxicand lowglucosestress.Mol.
Cell. Biol., 14: 5032—5042,1994.
33. Devary, Y., Gottlieb, R. A., Smeal, T., and Karmn,M. The mammalianultraviolet
response is triggered by activation ofsrc tyrosine kinases. Cell, 71: 1081—1091,
1992.
zeller, G., Marme,D., and Rapp,U. R. Proteinkinase Ca activatesRAF-I by direct
phosphorylation. Nature (Lond.), 364: 249—252,1993.
36. Ghosh, S., and Baltimore,D. Activationin vitroof NF-.cBby phosphorylationof its
inhibitor 1KB.Nature (Land.), 344: 678—682,1990.
37. Youn& S. D., and Hill, R. P. Hypoxia induces overreplicationand enhances metastatic
potential of murine tumor cells. Proc. Natl. Mad. Sd. USA, 85: 9533-9537, 1988.
34. Huang, W., Alessandro, A., Crews, C. M., and Erikson, R. L Ref-I forms a stable
complex with Meki andactivatesMeki by serinephosphorylation.Proc.Nail. Aced.
Sci. USA, 90: 10947—10951, 1993.
35. Kolch, W., Heldecker, G., Kochs, G., Hummel, R., Vahidi, H., Mischak, H., Finken
38. Rice, G. C., Hoy, C., and Schimke, R. T. Transient hypoxia enhances the frequency
of dihydrofolatereductasegene amplificationin Chinese hamsterovary cells. Proc.
Natl. Aced. Sci. USA, 83: 5978—5982,1986.
5279
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1994 American Association for Cancer Research.
Hypoxic Activation of Nuclear Factor-κB Is Mediated by a Ras
and Raf Signaling Pathway and Does Not Involve MAP Kinase
(ERK1 or ERK2)
Albert C. Koong, Eunice Y. Chen, Nahid F. Mivechi, et al.
Cancer Res 1994;54:5273-5279.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/54/20/5273
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1994 American Association for Cancer Research.