| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
4 is a dominant-negative regulator of HIF-1 and is down-regulated in renal cell carcinoma
* Department of Laboratory Medicine and Pathobiology, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada;
Department of Pathology, University Health Network, Princess Margaret Hospital, Toronto, Ontario, Canada;
Department of Public Health and Molecular Toxicology, School of Pharmaceutical Sciences, Kitasato University, Minato-ku, Tokyo, Japan; and
Departments of Urology and Surgical Oncology, University of Toronto and Princess Margaret Hospital, University Health Network, Toronto, Ontario, Canada
1Correspondence: Department of Laboratory Medicine and Pathobiology, Faculty of Medicine, University of Toronto, 1 King’s College Circle, Toronto, Ontario, M5S 1A8, Canada. E-mail: michael.ohh{at}utoronto.ca
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
, signifying a critical role of the HIF pathway in cancer development. The recently identified third member of the human HIF- family, HIF-3, produces multiple splice variants that contain extra DNA binding elements and protein-protein interaction motifs not found in HIF-1 or HIF-2. Here we report the molecular cloning of the alternatively spliced human HIF-3 variant HIF-34 and show that it attenuates the ability of HIF-1 to bind hypoxia-responsive elements located within the enhancer/promoter of HIF target genes. The overexpression of HIF-34 suppresses the transcriptional activity of HIF-1 and siRNA-mediated knockdown of the endogenous HIF-34 increases transcription by hypoxia-inducible genes. HIF-34 itself is oxygen-regulated, suggesting a novel feedback mechanism of controlling HIF-1 activity. Furthermore, the expression of HIF-34 is dramatically down-regulated in the majority of primary renal carcinomas. These results demonstrate an important dominant-negative regulation of HIF-1-mediated gene transcription by HIF-34 in vivo and underscore its potential significance in renal epithelial oncogenesis.— Maynard, M. A., Evans, A. J., Hosomi, T., Hara, S., Jewett, M. A. S., Ohh, M. Human HIF-34 is a dominant-negative regulator of HIF-1 and is down-regulated in renal cell carcinoma.
Key Words: HIF-34 • VHL • RCC • hypoxia
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
2
3)
. The disruption of cellular processes regulated by oxygen tension has been associated not only with cancer but also chronic lung disease, stroke, and heart disease (1
2
3)
. Hypoxia-inducible factor (HIF) is the transcription factor that up-regulates the expression of > 60 genes in response to reduced oxygen availability (1
2
3)
. HIF is a heterodimeric complex composed of and ß (also known as aryl hydrocarbon receptor nuclear translocator or ARNT) subunits (1
2
3)
. There are three HIF- genes in humans: HIF-1, HIF-2, and HIF-3. The HIF- subunits and ARNT are members of the basic helix-loop-helix (bHLH) -Per/ARNT/Sim (PAS) family of DNA binding proteins (1
2
3)
. As such, each contains a bHLH domain in the N terminus preceding the PAS domain (1
2
3)
. The basic domain is essential for binding DNA and the HLH and PAS domains are required for heterodimerization and DNA binding (4)
. HIF-1/2 contain two transactivation domains (NAD and CAD, located in the N terminus and C terminus, respectively), while ARNT contains just one transactivation domain (TAD) in the C terminus (1
2
3)
. The transactivation property of HIF is in concert with the binding of transcriptional coactivators such as p300/CBP to CAD of HIF- (1
2
3)
. In addition, HIF- contains a unique oxygen-dependent degradation (ODD) domain that determines HIF- protein stability (1
2
3)
.
Unlike HIF-, ARNT is constitutively expressed and stable (1
2
3)
. Under hypoxia, HIF- is stabilized, recruits p300/CBP, and binds ARNT (1
2
3)
. This heterodimeric HIF complex binds to a 5'-RCGTG-3' hypoxia-responsive element (HRE) in enhancers/promoters to trigger the transcription of numerous hypoxia-inducible genes that promote, in adaptation to hypoxia, angiogenesis, erythropoiesis, and glycolysis, as well as genes involved in iron metabolism and cell survival (1
2
3)
.
The HIF- subunits are hydroxylated at conserved proline residues within the ODD domain by prolyl hydroxylase domain-containing proteins, PHD1-3, which belong to the 2-oxoglutarate-dependent oxygenase superfamily (1
2
3)
. The enzymatic activity of PHD is tightly dependent on the availability of oxygen (1
2
3)
. We and others have shown that von Hippel-Lindau (VHL) tumor suppressor protein is the substrate recognition component of the E3 ubiquitin ligase complex that selectively polyubiquitinates prolyl-hydroxylated HIF- subunits (1
2
3)
. Therefore, the ubiquitin-mediated, oxygen-dependent destruction of HIF- is mediated by the VHL-E3 ubiquitin ligase (1
2
3)
. Furthermore, functional inactivation of VHL is the cause of VHL disease and is also encountered in the vast majority of sporadic clear cell renal cell carcinoma (CC-RCC), which is the predominant form of kidney cancer (1
2
3)
. Individuals with VHL disease develop hypervascular tumors in multiple organs including the brain, spine, retina, adrenal gland, and kidney (1
2
3)
. Certain mutations in VHL also cause congenital polycythemia (5
6
7
8)
. Notably, all CC-RCC tumor-causing VHL mutants tested to date have shown a failure in either assembling into an E3 ubiquitin ligase complex or binding to HIF- (1
2
3)
. Concordantly, tumor cells including CC-RCC devoid of functional VHL have enhanced expression of HIF target genes, such as VEGF, GLUT-1, and EPO, irrespective of oxygen tension (1
2
3)
. The overexpression of these and other hypoxia-inducible genes likely explains the angiogenic phenotype of VHL-associated tumors, but also supports the notion that constitutive stabilization of HIF- may be causally linked to tumorigenesis (1
2
3)
.
In addition to inactivating mutations on VHL, other cancer-causing mutations often increase the expression of HIF-1. For example, loss of p53 in various tumors or increased expression of HER2 receptor tyrosine kinase in breast cancer enhances HIF-1 expression and intensifies HIF-1-dependent transcription, often correlating with tumor aggressiveness (9
, 10)
. Loss of PTEN, which has been observed in the brain tumor glioblastoma multiforme, activates the Akt/protein kinase B signaling cascade, resulting in increased HIF-1 levels (11)
. Mutations in succinate dehydrogenase (SDH) result in the cytosolic accumulation of succinate, which inhibits prolyl hydroxylases PHDs, leading to the stabilization and activation of HIF-1 (12)
. Mutations in TSC2 tumor suppressor gene increase the level of HIF-1 via the mammalian target of rapamycin (mTOR) -dependent and -independent mechanisms that may involve chromatin remodeling (13)
. These provide mechanistic explanation for the highly vascular tumors including CC-RCC that develop in the absence of VHL mutations. These examples support the notion that there are other important regulators of HIF activity, which when inactivated promote oncogenic transformation.
In 1998, Gu et al. identified HIF-3, the third member of the HIF- family in mammals. HIF-3 has an NAD and ODD, but no CAD in its C terminus (14
, 15)
, and therefore HIF-3 may function as a weak transcription factor in comparison to HIF-1/2 (15)
. However, similar to HIF-1/2, HIF-3 protein expression is regulated in an oxygen-dependent manner and we have demonstrated that human HIF-3 is targeted for proteasomal degradation by the VHL-E3 complex in the presence of oxygen (14
, 16)
. Both the mouse and human HIF-3 locus give rise to multiple splice variants (16
, 17)
. The mouse HIF-3 splice variant, termed inhibitory PAS domain protein (IPAS), lacks the ODD and transactivation domains of full-length HIF-3 (17)
. IPAS functions as a dominant-negative HIF subunit, attenuating the endogenous expression of VEGF in the mouse cornea under hypoxia (17
, 18)
. The Drosophila HIF complex is composed of the subunit Sima and the ß subunit Tango, and therefore Drosophila HIF is referred to as Sima-Tango (19)
. Sima-Tango can affect transcription of HRE-regulated genes in either a positive (for example, lactate dehydrogenase) or negative (for example, globin) manner under hypoxia (20)
. Recently, the Drosophila Sima locus was shown to give rise to an alternative splice isoform (svSima) under hypoxia (20)
. Similar to IPAS, svSima lacks ODD and transactivation motifs (20)
. svSima may likewise function as a dominant-negative inhibitor of HIF-mediated transcriptional activation and repression during the hypoxic response in Drosophila by sequestering Tango from full-length Sima (20)
. Clearly, the splicing of HIF- mRNA to dominant-negative isoforms represents an important regulatory pathway for HIF-mediated gene expression.
Here, we report the cloning of an alternatively spliced human HIF-3 isoform HIF-34 and show that HIF-34, which lacks ODD, NAD, and CAD, has a dominant-negative function of inactivating HIF-1-mediated transcription. In addition, HIF-34 expression itself is oxygen regulated, indicating a novel feedback mechanism of controlling HIF-1 activity. Finally, but perhaps most important, HIF-34 is dramatically down-regulated in the majority of patient CC-RCC samples tested to date, suggesting potential significance in renal tumorigenesis previously unrecognized.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
4 cDNA4fw 5'-GACTGGCGAGCCATGGCGCTGGGGCTGC-AGCGCGCAAGGT-3' and hHIF34rv 5'-GATCAGGGCCACCAAGGGGCTCAAAT-3' primer set was designed based on the predicted HIF-34 cDNA (accession no. BC026308), and PCR was performed using human cerebellum cDNA library (BioChain Institute, San Leandro, CA, USA). The amplified fragment was cloned into pGEM-T Easy vector (Promega, Madison, WI, USA) and directly sequenced using an automated sequencer.
Plasmids
pcDNA3-HA-HIF-34 plasmid was generated using the primers 5'-GACGACGGATCCGCGCTGGGGCTGCAGCGCGCA-3' with an EcoRI restriction site and 5'-GACGACGAATTCAGGGCCACCAAGGGGGCAAATG-3' with a BamHI restriction site. The PCR product was digested with EcoRI and BamHI, and ligated into pcDNA3-HA (Invitrogen, Carlsbad, CA, USA) containing the HA-tag sequence between HindIII and BamHI restriction sites. pcDNA3.1-T7-HIF-34 plasmid was generated using a forward primer 5'-GACGACGGATCCATGGCAAGCATGACTGGTGGACAGCAAATGGGTGCGCTGGGGCTG-CAGCGC-3' with a BamHI restriction site and a reverse primer 5'-GACGACGGATTCTCAGGGCCACCAAGGGGGCAA-3' with an EcoRI restriction site. The PCR product was digested with EcoRI and BamHI, and ligated into pcDNA3.1(+) (Invitrogen). pcDNA3-HA-HIF-1, pcDNA3-ARNT, and pcDNA3.1-(HRE)5-luc expression plasmids were described previously (15
, 16
, 21)
.
Cells
Osteosarcoma U2OS cells, CC-RCC 786-O cells, HepG2 hepatocellular carcinoma cells, and HEK293A adenovirus-transformed embryonic kidney cells were obtained from American Type Culture Collection (Rockville, MD, USA). Cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) (except HepG2 cells, which were maintained in -MEM) supplemented with 10% heat-inactivated fetal bovine serum (Sigma, St. Louis, MO, USA) and maintained at 37°C in a humidified 5% CO2 atmosphere.
Antibodies
Monoclonal anti-HA (12CA5) antibody was obtained from Roche Molecular Biochemicals (Mannheim, Germany). Monoclonal anti--tubulin antibody was from Sigma Aldrich. Monoclonal anti-T7 antibody was obtained from Novagen (Madison, WI, USA). Monoclonal anti-HIF-1 and polyclonal and monoclonal anti-ARNT were obtained from Novus Biologicals (Littleton, CO, USA). Polyclonal anti-GLUT-1 antibody was obtained from Alpha Diagnostic International (San Antonio, TX, USA).
In vivo binding assay
HEK293A cells were transiently transfected at 50% confluency on a 100 mm plate using 15 µL of Fugene 6 (Roche Molecular Biochemicals) and 2.5 µg of various expression plasmids. pcDNA3.1 plasmid (Invitrogen) was used to maintain a constant final amount of transfected DNA. Four hours before cell lysis, MG-132 (Boston Biochem, Cambridge, MA, USA) was added to a final concentration of 10 µM. Cells were lysed in RIPA buffer (200 mM Tris-HCl pH 7.4, 130 mM NaCl, 10% glycerol, 0.5% sodium deoxycholate, 0.1% SDS, and 1% Triton-X 100) with Complete Mini protease inhibitors (Roche, Mannheim, Germany) and immunoprecipitated with antibodies against the epitope tag or the protein itself with protein A-Sepharose beads (Amersham Biosciences, Uppsala Sweden). Beads were washed 5 times with NETN buffer (20 mM Tris (pH 8.0), 120 mM NaCl, 1 mM EDTA, 0,5% Nonidet P-40), eluted by boiling in SDS-containing sample buffer, and separated by SDS-PAGE. Western blot was performed as described (22)
.
Electrophoretic mobility shift assay
Probe preparation: The HRE (5'-GGATCTGTGAGACGTGCGGCTTCCGTTT-3') from the phosphoglycerate kinase-1 enhancer was radiolabeled: 600 ng of the sense or antisense oligos, 10x PNK buffer (New England Biolabs, Beverly, MA, USA), 1 mM spermidine, 50 µCi of 
-[32P]-ATP, and 10 units of polynucleotide kinase (NEB) were mixed to a final reaction volume of 25 µL. After a 60 min incubation at 37°C, the oligos were combined and purified using a Bio-Spin 30 column (Bio-Rad, Hercules, CA, USA). The probe was heated for 5 min at 95°C and incubated overnight at 37°C for annealing. Reaction mix: HA-HIF-, ARNT, and HA-HIF-34 were in vitro translated using the TNT T7 quick-coupled transcription/translation system (Promega). In vitro translated protein (2 µL) was combined with 2x buffer (20% glycerol, 25 mM Tris pH 7.5, 50 mM KCl, 5 mM KCl, 1 mg/ml BSA, and 0.1% Triton x100), 300 ng of salmon sperm DNA, and 2 µL of probe (50,000 cpm/µL) to a total volume of 20 µL. When HA-HIF-34 was added to HA-HIF-1/ARNT, 2, 4 and 8 µL of HA-HIF-34 were included in the reaction mixture before the addition of the probe. For competition of the HRE oligo, 15 min before adding the probe, either a 250x molar excess of wild-type competitor oligo or scrambled oligo (5'-CGATTGATACCCTAGATTCAGGGAGC-3') was incubated with the reaction mixture. Supershift assay was performed using 5.2 µg of anti-HIF-1 antibody added to the reaction mixture. Complexes were separated on a nondenaturing polyacrylamide gel and visualized by autoradiography.
Dual-luciferase assay
U2OS osteosarcoma cells grown on 6-well plates were transfected with a total of 2.5 µg of expression plasmids using Escort (Sigma). (HRE)5-luc (firefly luciferase, 0.9 µg) was used to measure HRE-mediated transcription and 0.1 µg of the renilla luciferase plasmid, pRL-SV40 (Promega), was used as a transfection control. pcDNA3.1 plasmid (Invitrogen) was used to maintain a constant final amount of transfected DNA. Where indicated, a combination of 0.1 µg of ARNT and 0.8 µg of HA-HIF-1 were used. A ratio of 3:4 of HA-HIF-34:HA-HIF-1 was used where HA-HIF-1 was transfected without ARNT. Increasing concentrations of HA-HIF-34 were mixed into HA-HIF1/ARNT transfection reactions at a ratio of 2:8, 3:8, and 6:8 (HA-HIF-34:HA-HIF-1). Transfections were performed in triplicate and luciferase activity was measured using the Dual-Luciferase Reporter Assay system (Promega) and microplate luminometer (EG and G Berthold, Bad Wildbad, Germany). Firefly luciferase relative light units (RLUs) were normalized against Renilla luciferase RLUs.
Endogenous GLUT-1 expression assay
U2OS cells were transiently transfected at 50% confluency on a 100 mm plate using 15 µL of Escort (Sigma) and 5 µg of HA-HIF-34. Cells were harvested after 48 h and prepared for immunoblotting as described previously (22)
.
Hypoxia treatment of cells
HepG2 and HEK293A cells were grown to 80% confluency on 100 mm plates at 21% O2. HA-HIF-34 transfected U2OS cells were incubated for 48 h after transfection before hypoxia treatment. Each cell line was incubated at <1% O2 for 4 h in a ThermoForma (Marietta, OH, USA) hypoxia chamber (5% CO2, 10% H2, 85% N2).
siRNA-mediated knockdown of HIF-34
Online siDESIGN software (Dharmacon, Austin, TX, USA) was used to design siRNA duplex targeting HIF-34’s unique 3' tail (target sequence: AAGGAACUGUCUCCUUCCU). An irrelevant siRNA duplex was used as a negative control (target sequence: CCAUUCCGAUCCUGAUCCG). HEK293A cells were transfected at 50% confluency on 100 mm tissue culture plates with control and HIF-34 siRNA at a final concentration of 200 nM. Briefly, 20 µL of Oligofectamine (Invitrogen) was incubated with 100 µL of Opti-MEM I (Gibco/Invitrogen) for 8 min. The oligofectamine mixture was added to the siRNA diluted in 500 µL of Opti-MEM I and incubated for 20 min before adding to 5 mL of Opti-MEM I on the tissue culture plates. After 4 h, 3 mL of 30% heat-inactivated FBS (Sigma) DMEM was added to the plates. RNA was extracted 48 h after transfection using the RNeasy kit (Qiagen, Mississauga, ON) treated with RNA-free DNase kit (Ambion, TX, USA) and first-strand cDNA synthesis was performed.
CC-RCC tumor samples
Fresh frozen tumor and normal kidney samples, obtained from 9 nephrectomy specimens, were collected with informed patient consent and stored in liquid nitrogen in the University Health Network Tissue Bank maintained in the Department of Pathology, University of Toronto. The tissue samples used in this study were part of a University Health Network Research Ethics Board-approved protocol concerning gene expression in RCC. Representative frozen sections were obtained from all tumor samples before inclusion in the tissue bank to confirm the presence of tumor. All samples in this study consisted of > 90% tumor cells relative to stroma.
RNA extraction
RNA was extracted from nephrectomy wedges of 9 primary CC-RCC tumors and normal kidney tissue using the standard Trizol (Invitrogen) protocol with a polytron homogenizer. RNA quality and concentration were verified using an Agilent Bioanalyzer (Agilent BioTechnologies, Palo Alto, CA, USA). It should be noted that RNA extracted from normal kidney tissue is, in general, of lower quality than that of tumors.
Quantitative real-time PCR
First-strand cDNA synthesis: 1 µL of oligo(dT)23 primer (Sigma) was incubated with 5–10 µg of RNA and dH2O (total reaction volume was 20 µL) for 10 min at 70°C in a thermal cycler (MJ Research, Boston, MA, USA). The mixture was cooled to 4°C at which time 4 µL of 5x 1st strand reaction buffer, 2 µL of 0.1 M DTT, 1 µL of 10 mM dNTPs, and 1 µL Superscript II reverse transcriptase (Invitrogen) were added. cDNA synthesis was performed for 1.5 h at 42°C, followed by 15 min at 70°C in the thermal cycler. Human genomic DNA standards (human genomic DNA was from Roche, Mannheim, Germany) or cDNA equivalent to 20 ng of total RNA were added to the qPCR reaction in a final volume of 10 µL containing: 1x PCR buffer (without MgCl2), 3 mM MgCl2, 0.25 units of Platinum Taq DNA polymerase, 0.2 mM dNTPs, 0.3 µL SYBR Green I, 0.2 µL ROX reference dye, and 0.5 µM primers (Invitrogen). Amplification conditions were performed as follows: 95°C (3 min), 40 cycles of 95°C (10 s), 65°C (15 s), 72°C (20 s), 60°C (15 s), 95°C (15 s). qPCR was performed using the ABI Prism 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA, USA). Gene-specific oligonucleotide primers designed using Primer Express (Applied Biosystems) were as follows: HIF-34 primer set 1 (5'-ACTGCCTCCCCACCTCAAC-3' and 5'-CGGCAAGGAAGGAGACAGTTC-3'), HIF-34 primer set 2 (5'-TCCCCACCTCAACACAAGCT-3' and 5'-GGGCAAGGAAGGAGACAGTTC-3'), HIF-1 primer set 1 (5'-CAATACCCTATGTAGTTGTGGAAGTTTATG-3' and 5'-ACCAACAGGGTAGGCAGAACATT-3'), HIF-1 primer set 2 (5'-AATGGAATGGAGCAAAAGACAATT-3' and 5'-ATTGATTGCCCCAGCAGTCTAC-3'), ARNT primer set (5'-CCCCACCCAAGGAGCAA-3' and 5'-AGAAAAGCCTGAGCGGGTAGT-3'), GLUT-1 primer set 1 (5'-CACCACCTCACTCCTGT-TACTTACCT-3' and 5'-CAAGCATTTCAAAACCATGTTTCTA-3'), GLUT-1 primer set 2 (5'-CTCCCAGCAGCCCTAAGGAT-3' and 5'-ATCTGTCAGGTTTGGAAGTCTCATC-3'), VEGF primer set (5'-CTCTCTCCCTCATCGGTGACA-3' and 5'-GGAGGGCAGAGCTGAGTGTTAG-3'), and ß-actin primer set (5'-GGATCGGCGGCTCCAT-3' and 5'-CATACTCCTGCTTGCTGATCCA-3'). SYBR I fluoresces during each cycle of the qPCR by an amount proportional to the quantity of amplified cDNA (the amplicon) present at that time (23)
. The point at which the fluorescent signal is statistically significant above background is defined as the cycle threshold (Ct) (24)
. Expression levels of the various transcripts were determined by taking the average Ct value for each cDNA sample performed in triplicate and measured against a standard plot of Ct values from amplification of serially diluted human genomic DNA standards. Since the Ct value is inversely proportional to the log of the initial copy number, the copy number of an experimental mRNA can be obtained from linear regression of the standard curve. A measure of the fold difference in copy number was determined for each mRNA. Values were normalized to expression of ß-actin mRNA and expressed relative to scrambled siRNA samples (taken as 1) and represent the average value of three independent experiments performed in triplicate ± SE.
Immunohistochemical staining
The same patient tumor samples used for quantitative real-time PCR were processed according to standard surgical pathology protocols for immunohistochemical staining. Briefly, the specimens were fixed in 10% neutral buffered formalin for 24–36 h. Representative sections of tumor with adjacent nontumor renal parenchyma, 3–4 mm in thickness, were embedded in paraffin and 5 µm sections were cut and placed on coated slides for light microscopy.
Tumor morphology and classification were assessed using standard hematoxylin and eosin (H&E) staining. The tumors were classified as conventional clear cell type according to criteria described in the WHO classification of renal tumors (25)
. Immunohistochemical staining for VHL was performed manually using a standard avidin-biotin-peroxidase complex method. Briefly, unlabeled anti-VHL (mouse anti-human IgG VHL; ref 26
) was used at a 1:2000 dilution and an overnight incubation after microwave pretreatment for antigen retrieval. Sections were then incubated with a biotinylated secondary antibody (horse anti-mouse IgG, 1:200 dilution) and the avidin-peroxidase complex. The color reaction was visualized using Nova Red as the chromagen. The tissue was then lightly counterstained with hematoxylin.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
4 from human cerebellum, we performed reverse transcriptase-PCR analysis on human cerebellum cDNA library using HIF-3 isoform-specific primers. HIF-34 isolated from the human cerebellum is identical to the sequence of the human genomic clone (accession no. AC007193). However, this HIF-34 cDNA has three single nucleotide and one amino acid substitution (159Arg
Pro) compared with the GenBank sequence BC026308 reported in our previous work (16)
. The correct HIF-34 mRNA sequence has been submitted to GenBank (accession no. AB118749). HIF-34 cDNA contains an open reading frame (ORF) of 1089 bp that encodes a protein of 363 amino acids (Fig. 1
). The ORF of HIF-34 cDNA starts at exon 1a and ends at intron 7, which is not spliced. HIF-34 contains the basic helix-loop-helix (bHLH) and two Per/ARNT/Sim (PAS) domains (PASa and PASb) (Fig. 1)
. IPAS (accession no. AF416641), a splice variant of mouse HIF-3, is similar in its domain structure to HIF-34 and is also expressed in the cerebellum (18)
. Therefore, we hypothesize that HIF-34 represents the human ortholog of IPAS.
|
The splicing events of HIF-34 and IPAS transcripts, however, are significantly different. In IPAS, there is alternative splicing of exons 3 (truncating the 5' end of the exon), 4a (an additional exon not included in the full-length mHIF-3), and 6 (truncating the 3' end of the exon followed by splicing to exon 16, which is the last exon of full-length mHIF-3) (17)
. In contrast, splicing of HIF-34 exons 2-7 is identical to those of the longest isoform HIF-31, and exon 7 of HIF-34 is fully incorporated into the transcript with intron 7, resulting in a novel carboxyl terminus to HIF-34. The result of these splicing events is that HIF-34 retains the full PASb domain, while IPAS retains prematurely truncated PASb domain.
HIF-34 interacts with HIF-1 and ARNT in vivo
IPAS, the mouse ortholog of HIF-34, negatively regulates mHIF-mediated transcription by binding to mHIF-1, but not to mARNT (18)
. To address whether HIF-34 functions as an analogous dominant-negative HIF in humans, we first asked whether HIF-34 interacts with HIF-1 and/or ARNT. HA-HIF-1 alone or in combination with T7-HIF-34 were transiently transfected into HEK293A human embryonic kidney cells. Immunoprecipitation with an anti-HIF-1 antibody immunoprecipitated both HA-tagged HIF-1 (Fig. 2
a, arrow) and endogenous HIF-1 (bracket) and coprecipitated T7-HIF-34 (Fig. 2a
, lanes 6 and 8). Notably, endogenous HIF-1 bound T7-HIF-34 (Fig. 2a
, lane 6). Whether ARNT is required for HIF-34 binding to HIF-1 is not known at present. The multiple banding pattern associated with HIF-1 is likely due to post-translational modifications (27
, 28)
. The expression of T7-HIF-34 generated additional slower migrating proteins (Fig. 2a
, lane 2, asterisks). However, the nature of these proteins is currently unknown.
|
Unlike IPAS, which contains a partial PASb domain (17)
, HIF-34 contains the full bHLH/PASa and PASb dimerization domains. Therefore, it is formally possible that HIF-34 binding to HIF-1 does not exclude an interaction with ARNT. HEK293A cells were transiently transfected with ARNT alone or in combination with T7-HIF-34. Immunoprecipitation of T7-HIF-34 coprecipitated ARNT (Fig. 2b
, lane 8). Reciprocal immunoprecipitation using an anti-ARNT antibody coprecipitated T7-HIF-34 with endogenous ARNT (Fig. 2c
). These experiments were performed under experimental normoxia, in the absence of proteasome inhibitors, where the expression of HIF-1 remains negligible. Therefore, the association between ARNT and HIF-34 is likely independent of HIF-1. These results taken together demonstrate that HIF-34 interacts with either HIF-1 or ARNT, and hence is dissimilar to IPAS.
HIF-34 inhibits binding of HIF-1 to hypoxia-responsive elements
To address whether the ability of HIF-34 to interact with HIF-1/ARNT inhibits the binding of HIF-1 complex to hypoxia-responsive elements (HRE), we performed electrophoretic mobility shift assays (EMSA) with radiolabeled oligonucleotide derived from the HRE of phosphoglycerate kinase-1. In vitro translated HA-HIF-1, HA-HIF-34, or ARNT alone were not capable of binding to HRE (Fig. 3
, lanes 1–3). As expected, HA-HIF-1 in combination with ARNT bound to HRE indicated by bands 2 and 3 (Fig. 3
, lane 5). HA-HIF-34 in combination with ARNT did not bind HRE (Fig. 3
, lane 4). The absence of the PAS-associated C-terminal (PAC) domain in HIF-34, as well as its novel carboxyl terminus generated by the inclusion of intron 7, may account for the lack of DNA binding activity after dimerization with ARNT, as the PAC domain has been implicated in contributing to the DNA binding activity of HIF-1(4). The shifts are specific for the HRE as they were competed with excess unlabeled wild-type HRE but not with scrambled HRE (Fig. 3
, lanes 6 and 7, respectively). In addition, HRE-specific bands were supershifted with a monoclonal anti-HIF-1 antibody (Fig. 3
, lane 11, band 1) or with a monoclonal anti-ARNT antibody (data not shown). This indicates that the band shifts contain at a minimum HA-HIF-1 and ARNT. The binding of HA-HIF-1/ARNT complex to the HRE was inhibited by the addition of HA-HIF-34 in a dosage-dependent manner (Fig. 3
, lanes 8–10). The additional band (represented by an asterisk) was not supershifted by either anti-HIF-1 (Fig. 3
, lane 11) or anti-ARNT (data not shown) antibody and the intensity correlated with the amount of total rabbit reticulocyte lysate in the reaction mixture. Therefore, it is likely a protein(s), yet unidentified, in the rabbit reticulocyte lystate that binds specifically or nonspecifically to HRE.
|
HIF-34 inhibits HRE-driven transcription
We next asked whether the ability of HIF-34 to block HIF-1 binding to HRE recapitulates in the inhibition of HIF-mediated transcription in vivo. A dual-luciferase assay was performed by transiently transfecting U2OS human osteosarcoma cells with firefly luciferase reporter construct driven by five HREs from the phosphoglycerate kinase-1 enhancer ((HRE)5-luc). Transfection of ARNT or HA-HIF-34 did not activate transcription of the reporter (Fig. 4
, lanes 3 and 4). Cotransfection of HA-HIF-34 and ARNT likewise had negligible transactivation activity (Fig. 4
, lane 5). This is consistent with EMSA data (see Fig. 3
) that showed undetectable binding of HIF-34 to HRE in the presence of ARNT and, as expected, for HIF-34 does not contain a transactivation domain. HA-HIF-1 alone modestly induced the reporter activity (Fig. 4
, lane 6). This is likely through dimerization with the endogenous ARNT. This minimal activity was reduced by cotransfecting HA-HIF-34 (Fig. 4
, lane 7). As expected, cotransfection of HA-HIF-1 and ARNT dramatically activated the reporter activity (Fig. 4
, lane 8). In accord with the EMSA data (Fig. 3)
, the addition of HA-HIF-34 significantly decreased HIF-1-mediated reporter activity in a dose-dependent manner (Fig. 4
, lanes 9–11).
|
HIF-34 inhibits endogenous expression of hypoxia-responsive gene
To investigate the effect of HIF-34 on the expression of a bona fide hypoxia-inducible gene in vivo, HA-HIF-34 was transiently transfected into VHL–/– 786-O CC-RCC cells and VHL+/+ U2OS osteosarcoma cells. In the absence of functional VHL, 786-O cells constitutively overexpress hypoxia-inducible genes including glucose transporter-1 (GLUT-1) (Fig. 5
a, lane 1) (22)
. Ectopic expression of HA-HIF-34 dramatically attenuated the expression of endogenous GLUT-1 (Fig. 5a
). U2OS cells express wild-type VHL and are responsive to the changes in oxygen tension. As expected, under experimental normoxia (21% O2), GLUT-1 levels were low (Fig. 5b
, lane 1) but are up-regulated significantly during hypoxia (Fig. 5b
, lane 3). Ectopic expression of HA-HIF-34 markedly decreased the expression of GLUT-1 in U2OS cells subjected to less than 1% O2 (Fig. 5b
, lane 4). This demonstrates a dominant-negative role of HIF-34 on hypoxia-inducible HIF-mediated gene expression independent of VHL. Under normoxia, the inhibitory effect of HA-HIF-34 is difficult to assess owing to the already low level of hypoxia-inducible gene expression (Fig. 5b
, lane 2). Similar results were observed at the transcriptional level, as determined by quantitative real-time PCR (data not shown).
|
Hypoxia induces down-regulation of HIF-34 expression
HIF-responsive genes are activated during hypoxia in an effort to counter the damaging effects of reduced oxygen availability. It follows then that the activity of a dominant-negative regulator, such as HIF-34, of HIF-mediated transcription would be suppressed during hypoxia. To test this hypothesis, we assayed for the expression of endogenous HIF-34 using quantitative real-time PCR in HepG2 hepatoma cells and HEK293A cells under hypoxia (Fig. 6
a). HIF-34 expression is a relatively low copy number mRNA, but both cell types showed a significant reduction in the expression of endogenous HIF-34 (
75% and 50%, respectively) after 4 h of hypoxia treatment relative to no treatment. As a control, the effectiveness of the hypoxic treatment was confirmed by the induction of GLUT-1 expression (Fig. 6a
). This result suggests a novel feedback mechanism of hypoxia-induced down-regulation of HIF-34, which leads to the unbridled activation of HIF-1 in hypoxia. Conversely and equally important, upon restoring normoxia, the expression of HIF-34 would increase and thereby secure an inhibition of HIF-1 activity (continued below).
|
siRNA-mediated knockdown of endogenous HIF-34 increases the expression of HIF target genes
There is low basal hypoxia-inducible gene expression under normoxia (see Fig. 6b
; data not shown). Therefore, suppression of HIF-34 expression under normal oxygen tension should lead to an accumulation of hypoxia-inducible genes, such as GLUT-1 or VEGF. HIF-34-specific small interfering RNA (siRNA) targeting the unique carboxyl terminus of HIF-34 was transfected into HEK293A cells and RNA extracted for quantitative real-time PCR. The fold change in the level of mRNA upon HIF-34 siRNA treatment was calculated relative to the scrambled control siRNA-treated samples. HIF-34 siRNA treatment achieved 70% reduction in the level of endogenous HIF-34 (Fig. 6b
). Endogenous GLUT-1 and VEGF expression correspondingly increased approximately twofold (Fig. 6b
). The mRNA expression of HIF-1 or ARNT was not increased by HIF-34 siRNA treatment. This result suggests that the increased expression of GLUT-1 and VEGF expression in normoxia was not due to increased HIF-1, but rather to the reduction in HIF-34.
Endogenous expression of HIF-34 is down-regulated in primary CC-RCC
Numerous hypoxia-inducible genes are invariably overexpressed in CC-RCC, which likely explains its hypervascular nature. CC-RCC- and other cancer-causing mutations on several tumor suppressor genes including VHL, TSC2, and SDH, which impinge on the oxygen-sensing pathway leading to the accumulation of HIF-1 and HIF-2, have been characterized (9
10
11
12
13)
. HIF-34 is a dominant-negative regulator of HIF-1. Therefore, a prediction is that HIF-34 is suppressed in CC-RCC that rely on amplified HIF-1/2 activity for tumor development/progression. To address this hypothesis, RNA from primary CC-RCC tumors and normal kidney tissue were used for cDNA synthesis and quantitative real-time PCR to determine the expression level of HIF-34. For the experiments, 9 CC-RCC tumor samples and normal kidney tissue from nephrectomy specimens were obtained with informed patient consent as part of an ethics board approved protocol concerning gene expression in RCC. Seven of 9 tumors had significantly decreased expression of HIF-34 relative to normal kidney (Fig. 7
a). In addition, VHL–/– CC-RCC cell lines 786-O and RCC4 likewise had negligible to undetectable levels of HIF-34 (data not shown).
|
Inactivating mutations or loss of VHL are the most common oncogenic events (80%) in the development of CC-RCC (1
2
3)
. Hematoxylin and eosin staining of the CC-RCC samples confirmed a distinct clear cell-type morphology that is characterized by nests of tumor cells with abundant, optically clear cytoplasm and delicate cell membranes surrounded by a network of small, thin-walled blood vessels; it meets the criteria according to the World Health Organization classification of renal tumors (25)
(Fig. 7b
). Each section of tumor studied contained nontumor renal parenchyma including proximal convoluted tubules within the renal cortex. Cytoplasmic and membranous anti-VHL staining shown by proximal convoluted tubules was used as an internal positive control in each case (29)
. The majority of tumors (7/9) had no detectable VHL immunoreactivity (Fig. 7c
), while some tumors (2/9) showed positive staining that was accentuated on the delicate tumor cell membranes (Fig. 7d
). VHL status was further confirmed by direct sequencing (data not shown). However, no obvious correlation between VHL status and the expression of HIF-34 was observed. In other words, both VHL-positive and VHL-negative CC-RCC samples had reduced levels of HIF-34. It is unknown whether the down-regulation of HIF-34 is a result of normal, albeit newly defined, physiological response to tumor hypoxia or the result of certain or yet uncharacterized tumorigenic mutations that potentiate neoplastic transformation of renal proximal tubule cells, the likely origin of CC-RCC.
HIF-1 and/or HIF-2 have been found to be elevated in a majority of human cancers, including CC-RCC, and increased expression frequently correlates with poor prognosis and disease progression (1
2
3)
. The increased level of HIF-1/2 irrespective of oxygen tension correlates with inappropriate overexpression of HIF target genes that contribute to tumorigenesis (1
2
3)
. HIF-34 attenuates HIF activity and functions in vivo to modulate HIF-mediated gene expression in normal kidney epithelium, probably by forming an abortive complex that fails to bind to the HREs. Our finding that HIF-34 is frequently down-regulated in primary CC-RCC tumors suggests that a decreased level of HIF-34 is conducive to renal tumor progression. Forced expression of stable HIF-2 in CC-RCC cells reconstituted with wild-type VHL has reversed the tumor suppressor function of VHL in mouse xenograft assays (30)
. As well, knockdown of HIF-2 in CC-RCC cells devoid of functional VHL was shown to inhibit the growth of these cells as tumors in mouse models(31). These findings support the notion that HIF-2 is the causal HIF contributing to renal tumorigenesis. As HIF-34 is capable of binding ARNT independent of HIF-1, it has the potential to inhibit HIF-2 in addition to HIF-1. Thus, the question of whether the down-regulation of HIF-34 is required for HIF-2-mediated renal tumorigenesis is being addressed.
HIF-34 is likely a human ortholog of IPAS based on their primary sequence similarity and analogous function as dominant-negative regulator of HIF-1. Both are HIF-3 splice variants and lack the ODD and transactivation domains while retaining the majority of the DNA binding and dimerization domains (17)
. There are, however, some significant differences between HIF-34 and IPAS. HIF-34 and IPAS are generated via dissimilar splicing events, and whereas IPAS has been reported to bind only mHIF-1(18), HIF-34 binds to HIF-1 and ARNT. The expression of IPAS has been shown to be up-regulated in the hypoxic corneal epithelium of the eye during sleep to prevent inappropriate angiogenic gene expression mediated by HIF-1(18). The expression of HIF-34 is instead down-regulated during hypoxia in the tested hepatoma and embryonic kidney cells. Whether this difference reflects a tissue-specific response to oxygen or is a result of oncogenic mutations remains to be determined. It is, however, clear that the expression of HIF-34 is responsive to changes in oxygen tension. Whereas no association between IPAS and tumor development has been reported, HIF-34 is down-regulated in the majority (albeit a small sample size) of CC-RCC, suggesting its potential importance in renal oncogenesis. Whether HIF-34 is dysregulated in other forms of cancer remains to be determined.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
4 nucleotide sequence reported in this paper has been submitted to the GenBank database under accession no. AB118749. Received for publication March 17, 2005. Accepted for publication May 5, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
locus. J. Biol. Chem. 277,32405-32408This article has been cited by other articles:
|
|
