| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
B, Phosphatidylinositol 3'-Kinase, and Mitogen-Activated Protein Kinase Signaling Allows Tumor Necrosis Factor 
-Evoked Bcl-2 Expression to Provoke Internal Ribosome Entry Site-Dependent Translation of Hypoxia-Inducible Factor 1
1 Department of Cell Biology, Faculty of Biology, University of Kaiserslautern, Kaiserslautern, Germany; 2 Division of Human Immunology and Hanson Institute, Institute of Medical and Veterinary Science, Adelaide, South Australia, Australia; and 3 Department of Medicine, The University of Adelaide, South Australia, Australia
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and HIF-1ß subunits coordinates pathophysiologic responses toward decreased oxygen availability. It is now appreciated that enhanced protein translation of HIF-1 under normoxia accounts for an alternative regulatory circuit to activate HIF-1 by hormones, growth factors, or cytokines such as tumor necrosis factor (TNF-). Here, we aimed at understanding molecular details of HIF-1 translation in response to TNF-. In tubular LLC-PK1 cells, activation of nuclear factor 
B (NFB) by TNF- resulted in HIF-1 protein synthesis as determined by [35S]methionine pulse experiments. Protein synthesis was attenuated by blocking NFB, phosphatidylinositol 3'-kinase (PI3k), and mitogen-activated protein kinase (MAPK). Use of a dicistronic reporter with the HIF-1 5'-untranslated region (5'UTR) between two coding regions indicated that TNF- promoted an internal ribosome entry site (IRES) rather than a cap-dependent translation. IRES-mediated translation required the functional integrity of the NFB, PI3k, and MAPK signaling pathways. Although no signal cross-talk was noticed between NFB, PI3k, and MAPK signaling, these pathways are needed to up-regulate the anti-apoptotic target protein Bcl-2 by TNF-. Expression of Bcl-2 provoked not only IRES-dependent translation but also HIF-1 protein synthesis. We conclude that Bcl-2 functions as an important determinant in facilitating HIF-1 protein expression by TNF- via an IRES-dependent translational mechanism. These observations suggest a link between Bcl-2 and HIF-1 expression, a situation with potential relevance to cancer biology. |
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and the aryl hydrocarbon nuclear translocator, known as HIF-1ß. Expression of HIF-1 is subjected to regulation by oxygen availability, with low expression or its absence under normoxia and protein appearance during hypoxia, whereas HIF-1ß is constitutively present. To achieve degradation of HIF-1 under normoxia, prolyl hydroxylases and the von Hippel-Lindau protein are of major importance (2
, 4
, 5)
. Prolyl hydroxylases target Pro-564 and/or Pro-402 of HIF-1 to hydroxylate them under ambient oxygen concentrations. Proline hydroxylation suffices for binding of von Hippel-Lindau protein to HIF-1 with concomitant degradation of HIF-1 by the ubiquitin/26S proteasome system. Hypoxia, transition metals, or iron chelators directly attenuate prolyl hydroxylase and Asn-803–hydroxylase activities, leading to HIF-1 stabilization and HIF-1 transactivation (6
, 7)
.
Besides hypoxia, it is now appreciated that expression of HIF-1 and activation of HIF-1 can be observed during normoxia, by growth factors, hormones, nitric oxide (NO), or cytokines. Reports on cytokines affecting HIF-1 stability regulation pointed to a role of interleukin-1ß and tumor necrosis factor (TNF-; refs. 8
and 9
). Mechanistically, formation of reactive oxygen species and activation of nuclear factor B (NFB) and/or phosphatidylinositol 3'-kinase (PI3k) are involved in conveying the TNF- signal that resulted in HIF-1 protein expression (9, 10, 11, 12)
.
Several factors such as interleukin-1ß, insulin, HER2/neu, Src, and epithelial growth factor are known to enhance expression of HIF-1, and it is suggested that multiple pathways, including among others a PI3k–Akt–FKBP-rapamycin-associated protein pathway and the mitogen-activated protein kinase (MAPK) pathway, play crucial roles (1)
. Inhibitors of the MAPK pathway either affect HIF-1 transactivation by attenuating phosphorylation of transcriptional cofactors (13
, 14)
or control HIF-1 protein level (15
, 16)
.
The concept of intratumoral hypoxia as a driving force in cancer progression is crucial for various aspects such as angiogenesis, energy metabolism, invasion, and cell survival. Immunohistochemical analysis revealed that HIF-1 is overexpressed in many human cancers, and some correlation between increased HIF-1 expression and patient mortality and/or treatment failure exists (for references, see 1
). In addition, there is direct evidence that hypoxia in tumors selects for cells with decreased potential for apoptosis through the loss of p53 or overexpression of Bcl-2 (17)
. Bcl-2 is a prototype anti-apoptotic factor whose expression is regulated in response to hypoxia. However, contradictory studies either showed Bcl-2 up-regulation by HIF-1 (18)
or noticed decreased expression of Bcl-2 during hypoxia (19)
. Iervolino et al. (20)
observed enhanced HIF-1 accumulation under hypoxic conditions in combination with Bcl-2 overexpression, which positioned HIF-1 as a downstream target of Bcl-2, although mechanistic details remained obscure.
Here, we provide molecular explanations to understand TNF-–evoked HIF-1 protein expression. TNF- activates NFB, PI3k, and MAPK with concomitant downstream expression of Bcl-2. TNF- and Bcl-2 share the ability to provoke internal ribosome entry site (IRES)-dependent HIF-1 mRNA translation, and inhibition of NFB, PI3k, and MAPK abrogates the TNF- response. We conclude that TNF- uses divergent signaling amplifiers to regulate HIF-1 protein expression via Bcl-2–mediated IRES-dependent translation. These observations may provide groundwork to understand signaling between two proteins, often found highly expressed in tumors.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, sulfasalazine, LY294002, wortmannin, PD98059, and SB203580 were bought from Sigma (St. Louis, MO). Kinase inhibitors staurosporine, KT5720, calphostin C, KT5823, ML-7 were ordered from Biomol Research Laboratories (Plymouth Meeting, PA). A protein assay kit was from Bio-Rad (Hercules, CA). Protease inhibitor mixtures and T4 ligase came from Roche Diagnostics (Indianapolis, IN). Nitrocellulose membrane, ECL detection system, and horseradish peroxidase-labeled antimouse or antirabbit secondary antibodies were from by Amersham Biosciences (Uppsala, Sweden). HIF-1 and HIF-1ß antibodies were purchased from Becton Dickinson (Franklin Lakes, NJ). Akt, p44/42, p38, phospho-Akt (Ser-473), phospho-p44/42 MAPK (Thr-202/Tyr-204), phospho-p38 MAPK (Thr-180/Tyr-182) antibodies came from Cell Signaling Technology Inc. (Beverly, MA). Anti-Bcl-2 antibody was from R&D Systems (Minneapolis, MN). Reporter plasmid cap-Luc, luciferase reporter assay systems, and dual-luciferase reporter assay system were supplied by Promega (Madison, WI). Reporter plasmids pRF and pRhifF (21)
, reporter plasmid NFB-Luc (22)
, and expression plasmid pRc-Bcl-2 (23)
were described previously. The plasmids pCMV-IBM encoding a Ser-to-Ala mutant of IB were purchased from Becton Dickinson, whereas pSR-
p85 encoding the dominant-negative PI3k subunit p85 was kindly provided by Dr. W. Ogawa (Kobe University, School of Medicine, Kobe, Japan).
Cell Culture.
Proximal tubular LLC-PK1 cells were cultured in Dulbecco’s modified Eagle’s medium with 1 g/L D-glucose. Human embryonic kidney 293 cells were cultured in Dulbecco’s modified Eagle’s medium with 4.5 g/L D-glucose. Medium was supplemented with 10% fetal calf serum, 2 mmol/L glutamine, 100 units/mL penicillin, 100 µg/mL streptomycin, and 1 mmol/L sodium pyruvate (only for human embryonic kidney 293 cells). Cells were transferred twice a week, and medium was changed before experiments. Cells were kept in a humidified atmosphere of 5% CO2 in air at 37°C.
Plasmid Construction.
The pRrevhifF plasmid was constructed by excising the HIF-1 5'-untranslated region (UTR) between SpeI (5' end) and NcoI (3' end) sites from pRhifF plasmid and recloning the inverted HIF-1 5'UTR. Amplification of the inverted HIF-1 5'UTR fragment was performed by PCR using the forward 5' primer containing the NcoI site and backward 3' primer containing the SpeI site.
Western Blot Analysis.
Cells were treated, scraped off, lysed in 300 µL of buffer A [50 mmol/L Tris, 150 mmol/L NaCl, 5 mmol/L EDTA, 0.5% NP-40, and protease inhibitor mixture (pH 7.5)], and sonicated, followed by centrifugation (15,000 x g, 15 minutes). Eighty µg of protein were loaded onto and resolved on 7.5% or 10% SDS-polyacrylamide gels. Gels were washed with blotting buffer [25 mmol/L Tris, 192 mmol/L glycine, and 20% methanol (pH 8.3)] for 5 minutes, proteins were blotted onto nitrocellulose by a semidry transfer cell, and unspecific binding sites were blocked with 5% milk/Tween-Tris buffer saline (TTBS) [50 mmol/L Tris/HCl, 140 mmol/L NaCl, and 0.05% Tween-20 (pH 7.2)] for 1 hour. The primary antibody was added and incubated overnight at 4°C. Afterward, nitrocellulose membranes were washed three times for 5 minutes each with TTBS. For protein detection, blots were incubated with a horseradish peroxidase-labeled antimouse or antirabbit secondary antibody for 1 hour and washed three times for 5 minutes each with TTBS, followed by ECL detection.
Trichloroacetic Acid Protein Precipitation.
Trichloroacetic acid (TCA) at a final concentration of 20% was added to cell lysates, incubated for 15 minutes on ice, and centrifuged at 10,000 x g at 4°C for 15 minutes. The supernatant was carefully removed, and the pellet was washed three times, each with 300 µL of cold acetone. Finally, the air-dried pellet was resuspended to allow scintillation counting.
[35S]Methionine Labeling, Scintillation, and Autoradiography.
LLC-PK1 cells were treated with 500 ng/mL TNF- in the presence or absence of 300 µmol/L sulfasalazine for 8 hour. Cells were then starved for 1 hour in serum- and methionine-free medium. Thereafter medium was changed to methionine-free medium containing 10% fetal calf serum supplemented with 100 µCi/mL [35S]methionine for 1 hour. TNF- and, if appropriate, sulfasalazine were present throughout the experiment. Cells were then washed with PBS and lyzed in buffer B [50 mmol/L Tris, 150 mmol/L NaCl, 5 mmol/L EDTA, 0.5% NP-40, 5% glycerol, and protease inhibitor mixture (pH 7.5)]. For scintillation counting, lysates containing 1 mg of total protein were TCA-precipitated. Pellets were resuspended in 1 mL of buffer B and transferred into scintillation vials. After adding 10 mL of scintillation fluid, radioactivity was counted in a liquid scintillation counter (PerkinElmer Life Sciences, Boston, MA), and counts were corrected for decay, quenching, and counting efficiency. For autoradiography, HIF-1 was immunoprecipitated from lysates containing 1 mg of total protein using 1 µg of anti-HIF-1 antibody as described previously (12)
. After running 10% SDS-PAGE, gels were dried and exposed to X-ray film to detect 35S-labeled HIF-1.
Cell Transfection and Reporter Assay.
LLC-PK1 cells (2 x 105) were seeded in 6-cm dishes 1 day before transfection. At a rate of 60% confluence, cells were transfected with reporter plasmids, using the calcium phosphate precipitation method (12)
. In brief, plasmids in the presence of 125 mmol/L CaCl2 and HBS buffer [25 mmol/L HEPES, 140 mmol/L NaCl, and 0.75 mmol/L Na2HPO4 (pH 7.05)] were incubated for 15 minutes at room temperature and added dropwise to cells. Sixteen hours later, medium was changed, and incubations were continued for another 8-hour period. At this point the medium was replaced, and cells were treated as indicated for an additional 8- or 16-hour period. Cells were lysed, and firefly and/or Renilla luciferase activity was measured using commercial kits.
Densitometric Quantification.
Densitometric quantification (expressed as quantum levels) of, e.g., HIF-1 signals, was performed with the Aida Image Software (Raytest Isotopenmessgeraete GmbH, Straubenhardt, Germany). Quantum levels are expressed relative to controls (set as 100).
Statistical Analysis.
Each experiment was performed at least three times, and representative data are shown. Data in bar graphs are given as the mean ± SD. Means were checked for statistical differences using the t test with error probabilities of P < 0.05.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Provokes Hypoxia-Inducible Factor 1 Synthesis Using Nuclear Factor B, Phosphatidylinositol 3'-Kinase, and Mitogen-Activated Protein Kinase Pathways. provoked accumulation of HIF-1 in LLC-PK1 cells (12)
. It was our intention to gain additional insights into molecular mechanisms of TNF- signaling and to understand how this cytokine contributes to HIF-1 expression. In a first experiment, we showed that TNF- led to the accumulation of HIF-1. This response was antagonized by the NFB inhibitor sulfasalazine. In contrast, expression of HIF-1ß was neither altered in response to TNF- nor affected by sulfasalazine (Fig. 1A)
. For these studies, we added recombinant human TNF- to pig LLC-PK1 cells, which may explain high concentrations of TNF- being used. In human A549 cells, HIF-1 accumulation was achieved with 10 ng/mL TNF-, which is a dose reported to work in MCF-7 cells as well (11)
. In the next set of experiments, we determined HIF-1 protein synthesis by [35S]methionine pulse experiments. Therefore, LLC-PK1 cells were stimulated with 500 ng/mL TNF- for 8 hours followed by [35S]methionine pulse labeling. Subsequently, we determined the synthesis of radioactive HIF-1 during a 1-hour time period under the impact of TNF- (Fig. 1B)
. Compared with unstimulated cells (with the relative quantum level of the corresponding bands set to 100), TNF- provoked protein synthesis (relative quantum level of the bands amounted to 1420 ± 153), whereas 300 µmol/L sulfasalazine reduced the rate of 35S-HIF-1 protein synthesis (relative quantum level of the bands amounted to 134 ± 18).
|
B luciferase reporter assay confirmed that sulfasalazine indeed inhibits NFB activity (Fig. 1C)
. TNF-–stimulated luciferase activity was suppressed by sulfasalazine to near control values. To prove specificity for TNF- in expressing HIF-1, we controlled for total protein synthesis by TCA-precipitating all protein, followed by scintillation counting. As seen in Fig. 1D
, neither TNF- nor sulfasalazine affected protein labeling significantly, thus implying that the impact of TNF- is not a general effect on protein turnover.
Puzzled by the observation not only that inhibition of NFB reduced TNF-–evoked HIF-1 expression but also that PI3k inhibitors such as wortmannin revealed some interference (9)
, we decided to determine phosphorylation events in general. Therefore, we screened several kinases inhibitors for modulating TNF-–elicited HIF-1 expression (Fig. 2A)
. Despite an inhibitory impact of the pan kinase inhibitor staurosporine, other interventions such as blocking protein kinase A, protein kinase C, protein kinase G, or myosine light chain kinase showed no interference with HIF-1 expression.
|
–stimulated HIF-1 expression (Fig. 2B)
. We also found that the MEK inhibitor PD58059 lowered HIF-1 expression and that the p38 MAPK inhibitor SB203580 decreased HIF-1 expression as well (Fig. 2C)
. Having established that TNF- provoked enhanced HIF-1 translation in LLC-PK1 cells and considering that signaling via NFB, PI3k, and MAPK was needed for HIF-1 expression, we asked whether corresponding inhibitors would attenuate translation of HIF-1 mRNA. Because mRNA translation may result from either cap- or IRES-dependent mechanisms, we discriminated these possibilities in additional studies.
Tumor Necrosis Factor Enhances Hypoxia-Inducible Factor 1 Translation by an Internal Ribosome Entry Site-Dependent Mechanism.
To investigate whether an IRES element facilitates the TNF- response in LLC-PK1 cells, we transfected 0.5 µg of pRF, pRhifF, or pRrevhifF plasmids and stimulated cells for 16 hours with 500 ng/mL TNF-. pRF contains a short linker sequence between the Renilla and firefly luciferase cistrons, and pRrevhifF contains the inverted HIF-1 5'UTR, whereas pRhifF contains the HIF-1 5'UTR inserted between the cistrons (Fig. 3A)
. The inverted HIF-1 5'UTR-containing plasmid pRrevhifF showed similar firefly/Renilla ratios compared with the control plasmid pRF. However, the presence of the HIF-1 5'UTR increased expression of the downstream firefly luciferase relative to Renilla luciferase roughly 5-fold under control conditions and up to 20-fold with TNF- stimulation, suggesting that the HIF-1 5'UTR can indeed promote translation by allowing internal ribosome entry.
|
treatment when transfected into LLC-PK1 cells (Fig. 3B)
. The relative ratio of IRES-dependent reporter activity to cap-dependent reporter activity and the lack of cap-evoked reporter activity on its own implies that TNF- uses an IRES to increase HIF-1 translation and thus to enhance HIF-1 protein expression.
Internal Ribosome Entry Site-Dependent Hypoxia-Inducible Factor 1 Translation Requires Nuclear Factor B, Phosphatidylinositol 3'-Kinase, and Mitogen-Activated Protein Kinase Signaling.
Findings on the impact of NFB, PI3k, and MAPK inhibitors in attenuating TNF-–elicited HIF-1 expression and the observation on an IRES-facilitated translational control raised the question of whether the divergent signaling systems can be positioned upstream of IRES translation. To answer this question, we transfected LLC-PK1 cells with 0.5 µg of pRhifF plasmid followed by stimulation for 16 hours with 500 ng/mL TNF- in the absence or presence of inhibitors (Fig. 4A)
. The relative ratios of firefly to Renilla luciferase activities were measured and normalized to control ratios determined in the absence of the agonist TNF-.
|
B with sulfasalazine significantly attenuated the TNF-–stimulated ratio of firefly versus Renilla luciferase activity, thereby suggesting the involvement of NFB signaling in IRES-stimulated translation. Similar results were obtained by blocking PI3k with LY294002 and wortmannin, by employing the MEK inhibitor PD58059, or by shutting off p38 MAPK with SB203580. Results thus far suggest a role of NFB, PI3k, and MAPK in IRES-dependent HIF-1 translational regulation. To exclude potential side effects of drugs, we eliminated NFB and PI3k signaling by employing a molecular approach (Fig. 4B)
. Transfection of a dominant-negative noncleavable mutant of IB or a dominant-negative mutant of p85-PI3k attenuated the TNF-–stimulated ratio of firefly versus Renilla luciferase activity, thereby confirming effects seen with pharmaceutical interventions.
In the following studies, it was our intention to see whether any cross-talk between the major transducing pathways exists that may further help to understand IRES translational control (Fig. 5)
. In a first set of experiments, we checked the impact of PI3k and MAPK inhibitors on NFB. To this end, we transfected a NFB–luciferase reporter construct into LLC-PK1 cells followed by stimulation with TNF- for 8 hours. As seen in Fig. 5A
, neither LY294002 or wortmannin nor PD58059 or SB203580 inhibited activation of NFB. Variations in reporter activity did not reach the level of significance. Thus, blocking PI3k or MAPK pathways did not affect NFB signaling.
|
B pathway on PI3k or MAPK signaling (Fig. 5B)
. An increased phosphorylation of Akt, ERK1/2, and p38 on TNF- addition indicated activation of the PI3k and MAPK pathways. Blocking NFB with 500 µmol/L sulfasalazine did not affect protein phosphorylation of Akt, ERK1/2, or p38. However, inhibition was visible when using specific inhibitors such as LY294002, PD58059, or SB203580. Activity variations, i.e., the degree of protein phosphorylation of Akt, p44/42, or p38, was verified in relation to the expression of the nonphosphorylated proteins to show that the protein amounts did not change. We conclude that the NFB, PI3k, and MAPK kinase pathways are independently required for activation of IRES-dependent HIF-1 translation.
Tumor Necrosis Factor Provokes Nuclear Factor B–, Phosphatidylinositol 3'-Kinase–, and Mitogen-Activated Protein Kinase–Dependent Bcl-2 Expression, Which in Turn Stimulates Hypoxia-Inducible Factor 1 Translation.
Searching for a common denominator to explain NFB–, PI3k-, and MAPK-dependent IRES translation in response to TNF-, we considered classical TNF--responsive genes and succeeded in identifying Bcl-2. Expression of Bcl-2 was low in resting LLC-PK1 cells but increased upon TNF--stimulation (Fig. 6A)
. Expression regulation was time dependent with an increase seen at 2 hours that persisted up to 24 hours, although a lower expression was noticed at 16 and 24 hours compared with 8 hours.
|
for 4 hours, we noticed attenuated expression of Bcl-2 with the addition of sulfasalazine, LY294002, PD58059, or SB203580 (Fig. 6B)
. These results imply that TNF- uses NFB–, PI3k-, and MAPK-signaling pathways to stimulate expression of the common downstream protein target Bcl-2.
To prove causation of Bcl-2 in affecting HIF-1 translation, we transfected human embryonic kidney 293 cells with 0.25 µg of the reporter plasmid pRhifF or pRrevhifF in combination with 0.25 µg of pRc-Bcl-2 versus control plasmid (Fig. 7A)
. The ratio of firefly to Renilla luciferase activity was increased by forced Bcl-2 expression as well as with 500 ng/mL TNF- as seen before, which indicates translational regulation via the HIF-1 5'UTR.
|
translation via the 5'UTR to a similar extent as TNF-. This response was not observed with using the plasmid containing the 5'UTR of HIF-1 in reverse orientation. Additional experiments in human embryonic kidney 293 cells showed that overexpression of Bcl-2 promoted HIF-1 expression as determined by Western blot analysis (Fig. 7B)
. We conclude that TNF- enhances the expression of Bcl-2, which in turn increases HIF-1 protein level and thus provokes formation of the transcriptionally active HIF-1 complex. These signaling circuits require active NFB, PI3k, and MAPK pathways. |
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
in response to cytokines or growth factors often results from enhanced mRNA translation. We corroborated these observations for TNF- signaling and went on to demonstrate the involvement of the NFB, PI3k, and MAPK pathways. Enhanced mRNA translation is facilitated by an IRES activity that is attenuated by blocking NFB-, PI3k- and MAPK-signaling. Activation of these pathways in response to TNF- resulted in Bcl-2 expression. Overexpression of Bcl-2 promoted IRES-dependent HIF-1 mRNA translation as well as HIF-1 protein synthesis. These results, schematically presented in Fig. 8
, provide an intriguing link between Bcl-2 expression as seen in many tumors and pathophysiologic consequences evoked via HIF-1 signaling.
|
B, PI3k, and MAPK pathways by TNF- in LLC-PK1 cells increases the rate of HIF-1 protein synthesis rather than affecting HIF-1 stability as known for hypoxia. This was proven by three independent experimental approaches involving cycloheximide addition to follow HIF-1 disappearance (12)
, radioactive pulse experiments, and reporter gene assays. The effect of TNF- signaling may be such that sufficient amounts of HIF-1 are produced to make von Hippel-Lindau protein limiting (3)
, resulting in the failure to degrade all of the HIF-1, thus explaining activation of HIF-1 under normoxia.
Previous studies have demonstrated that hypoxia affects HIF-1 protein stability via blocked ubiquitination to increase the HIF-1 protein amount, whereas TNF- leaves ubiquitination unaltered but increases mRNA translation. Translation of many eukaryotic mRNAs involves interaction of the mRNA 5'm7GpppN cap with the eIF4E subunit of the eIF4F translation initiation complex. Once the eIF4F complex is assembled on the mRNA, the small ribosomal subunit is recruited, and scanning for a favorable AUG codon starts. The availability of eIF4E largely depends on the phosphorylation state of its inhibitory binding partners 4EBP1 and 4EBP2. Phosphorylation of 4EBP1/2 relieves eIF4E to promote cap-dependent translation. An alternative mode of translation initiation recruits the translation initiation complex by a 5'UTR IRES. This may involve p70S6K, which regulates translation of a group of mRNAs possessing a 5'-terminal oligopyrimidine tract. The HIF-1 5'UTR contains these tracts, including a conserved sequence in the extreme 5' terminus (24)
. Using dicistronic reporter assays, we demonstrated that TNF- uses an IRES to enhance translation that required active NFB, PI3k, and MAPK pathways. IRES-dependent HIF-1 translation has been noticed for HER2/neu and angiotensin II as well as HIF-1 translation during normoxia and hypoxia (21
, 25
, 26)
. HER2/neu uses the PI3k–Akt–FKBP-rapamycin-associated protein pathways (25)
with the likelihood phosphorylation of p70S6K and additional signal transduction via phosphorylation of the S6 protein to control IRES-dependent translational mechanisms (27)
. In some analogy, angiotensin II increases HIF-1 translation by reactive oxygen species (ROS)-dependent activation of PI3k through the 5'UTR of HIF-1 (26)
. We conclude that TNF- shares with HER2/neu and angiotensin II the ability to express HIF-1 by facilitating IRES-dependent translation making use of the PI3k pathway. Because the PI3k pathway is also implicated in cap-dependent translation of HIF-1, e.g., in response to insulin or insulin-like growth factor (16
, 28)
, one can speculate that a distinct phosphorylation pattern of 4EBP1/2, p70S6K, or the S6 protein reflects these differences and that cosignals determine the route and efficacy of HIF-1 protein synthesis. Among the cosignals generated by TNF-, we found activation of NFB and MAPK. Activation of NFB is an established response of TNF-receptor-1 activation helping to understand why TNF- does not usually trigger apoptosis in such cells. As previously noticed, NFB is essential in allowing HIF-1 expression and, as shown here, is demanded for IRES translation. A role of MAPK in regulating HIF-1 is established as well, but contrasting observations may point to cell-specific effects (28
, 29)
. There are reports that the MAPK pathway can stimulate HIF-1 expression or up-regulate HIF-1 activity by promoting the formation of the HIF-p300 complex to enhance transactivation. We extend existing information on multiple pathways in regulating HIF-1 expression by showing their participation in IRES translation. Moreover, our data suggest that signaling via NFB, PI3k, and MAPK independently of each other contribute to HIF-1 expression. Interestingly, transmission by these pathways culminated in Bcl-2 expression. There is ample evidence that NFB-induced Bcl-2 expression (30)
and particularly TNF- activate NFB to up-regulate Bcl-2 (31)
. Moreover, PI3k as a survival signaling pathway provokes Bcl-2 expression (32)
, activation of the MEK/ERK pathway is known to induce Bcl-2 (33)
, and activation of p38 MAPK results in Bcl-2 expression (34)
and/or Bcl-2 mRNA stabilization (35)
. In human embryonic kidney 293 cells, we succeeded in demonstrating a direct impact of Bcl-2 on IRES-dependent HIF-1 translation and HIF-1 protein expression, and similar results were obtained in LLC-PK1 and HepG2 cells. This not only identifies Bcl-2 as a common denominator of NFB, PI3k, and MAPK signaling in eliciting IRES-dependent translation but also merges actions of an anti-apoptotic protein with those of HIF-1. In patients with early-stage esophageal cancer, the combined overexpression of HIF-1 and Bcl-2 was associated with a failure to respond to photodynamic therapy (19)
. Bcl-2 overexpression will provide a significant survival advantage toward apoptosis, a phenomenon seen in subpopulations of hypoxic tumor cells that are refractory to radiotherapy and some forms of chemotherapy (36)
. Protecting tumor cells from cell demise may require additional adaptive responses to ensure transcription of genes that are involved in crucial aspects of cancer biology. The products of the genes that HIF-1 regulates will fulfill these criteria and it is attractive to directly link Bcl-2 signaling to increased HIF-1 activity. It was previously noticed that Bcl-2 overexpression enhanced hypoxia-evoked protein amounts of HIF-1 and increased hypoxia-induced VEGF expression and thus contributes to angiogenesis (20)
. Furthermore, it has been demonstrated that Bcl-2 overexpression in hypoxia increases Sp1 expression and activity through ERK signaling, which results in enhanced urokinase plasminogen activator receptor expression (37)
. Although these data suggest the possibility that Bcl-2 affects gene activation and our study expands this notion by demonstrating enhanced IRES-dependent translation under the control of Bcl-2, molecular details remain obscure. One may speculate on the involvement of ROS, considering some reports indicating ROS involvement in (de)stabilizing HIF-1. Bearing in mind that Bcl-2 has been found to modulate ROS (38)
, quenching of ROS may contribute to HIF-1 stabilization. On the other side, previous studies also noticed a role of ROS in Bcl-2 mRNA and protein expression (39)
. As a more general concept, it is attractive to hypothesize that stressed cells may up-regulate Bcl-2 to fight unfavorable environmental living conditions and/or to escape apoptosis. Cellular stress often attenuates cap-dependent translation in favor of mRNA translation via IRES elements. Expression of HIF-1 via Bcl-2 may in turn use the HIF-1 transcriptional machinery to gain a growth advantage by, e.g., improving energy metabolism via induction of HIF-1 target genes. Unfortunately, this system can be misused in tumors that show excessive Bcl-2 expression. It will be important to further identify the intimate relation between overexpression of Bcl-2 and HIF-1 in tumors and to question to what extent Bcl-2 contributes to HIF-1 expression and signaling. The observation that the cytokine TNF- stabilizes HIF-1 via expression regulation of Bcl-2 may add to our understanding of how inflammatory mediators, e.g., TNF- contribute to cancer development (40)
.
Our experiments have shown that TNF- activates distinct signaling pathways such as NFB, PI3k, and MAPK that lead to Bcl-2 expression, which in turn provokes IRES-dependent HIF-1 mRNA translation and HIF-1 protein synthesis. The fact that Bcl-2 and HIF-1 are heavily expressed in a number of tumors may argue for an important medical relevance of our observations.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Bernhard Brüne, University of Kaiserslautern, Faculty of Biology, Department of Cell Biology, Erwin Schroedinger-Strasse, 67663 Kaiserslautern, Germany. Phone: 49-631-205-2406; Fax: 49-631-205-2492; E-mail: bruene{at}rhrk.uni-kl.de
Received 4/23/04. Revised 9/ 9/04. Accepted 10/ 6/04.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Frede, C. Stockmann, S. Winning, P. Freitag, and J. Fandrey Hypoxia-Inducible Factor (HIF) 1{alpha} Accumulation and HIF Target Gene Expression Are Impaired after Induction of Endotoxin Tolerance J. Immunol., May 15, 2009; 182(10): 6470 - 6476. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Tacchini, E. Gammella, C. De Ponti, S. Recalcati, and G. Cairo Role of HIF-1 and NF-{kappa}B Transcription Factors in the Modulation of Transferrin Receptor by Inflammatory and Anti-inflammatory Signals J. Biol. Chem., July 25, 2008; 283(30): 20674 - 20686. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. J. Carver, B. Gaston, K. deRonde, and L. A. Palmer Akt-Mediated Activation of HIF-1 in Pulmonary Vascular Endothelial Cells by S-Nitrosoglutathione Am. J. Respir. Cell Mol. Biol., September 1, 2007; 37(3): 255 - 263. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Zhou, A. E. Damdimopoulos, G. Spyrou, and B. Brune Thioredoxin 1 and Thioredoxin 2 Have Opposed Regulatory Functions on Hypoxia-inducible Factor-1{alpha} J. Biol. Chem., March 9, 2007; 282(10): 7482 - 7490. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Aggarwal, H. Ichikawa, Y. Takada, S. K. Sandur, S. Shishodia, and B. B. Aggarwal Curcumin (Diferuloylmethane) Down-Regulates Expression of Cell Proliferation and Antiapoptotic and Metastatic Gene Products through Suppression of I{kappa}B{alpha} Kinase and Akt Activation Mol. Pharmacol., January 1, 2006; 69(1): 195 - 206. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Kozak A second look at cellular mRNA sequences said to function as internal ribosome entry sites Nucleic Acids Res., November 28, 2005; 33(20): 6593 - 6602. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Frede, P. Freitag, T. Otto, C. Heilmaier, and J. Fandrey The Proinflammatory Cytokine Interleukin 1{beta} and Hypoxia Cooperatively Induce the Expression of Adrenomedullin in Ovarian Carcinoma Cells through Hypoxia Inducible Factor 1 Activation Cancer Res., June 1, 2005; 65(11): 4690 - 4697. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Ichikawa, Y. Takada, A. Murakami, and B. B. Aggarwal Identification of a Novel Blocker of I{kappa}B{alpha} Kinase That Enhances Cellular Apoptosis and Inhibits Cellular Invasion through Suppression of NF-{kappa}B-Regulated Gene Products J. Immunol., June 1, 2005; 174(11): 7383 - 7392. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
