Post-transcriptional regulation of vascular endothelial growth factor by hypoxia.

The major control point for the hypoxic induction of the vascular endothelial growth factor (VEGF) gene is the regulation of the steady-state level of the mRNA. We previously demonstrated a discrepancy between the transcription rate and the steady-state mRNA level induced by hypoxia. This led us to examine the post-transcriptional regulation of VEGF expression. Actinomycin D experiments revealed that hypoxia increased VEGF mRNA half-life from 43 ± 6 min to 106 ± 9 min. Using an in vitro mRNA degradation assay, the half-life of VEGF mRNA 3′-untranslated region (UTR) transcripts were also found to be increased when incubated with hypoxic versus normoxic extracts. Both cis-regulatory elements involved in VEGF mRNA degradation under normoxic conditions and in increased stabilization under hypoxic conditions were mapped using this degradation assay. A hypoxia-induced protein(s) was found that bound to the sequences in the VEGF 3′-UTR which mediated increased stability in the degradation assay. Furthermore, genistein, a tyrosine kinase inhibitor, blocked the hypoxia-induced stabilization of VEGF 3′-UTR transcripts and inhibited hypoxia-induced protein binding to the VEGF 3′-UTR. These findings demonstrate a significant post-transcriptional component to the regulation of VEGF.

The major control point for the hypoxic induction of the vascular endothelial growth factor (VEGF) gene is the regulation of the steady-state level of the mRNA. We previously demonstrated a discrepancy between the transcription rate and the steady-state mRNA level induced by hypoxia. This led us to examine the post-transcriptional regulation of VEGF expression. Actinomycin D experiments revealed that hypoxia increased VEGF mRNA half-life from 43 ؎ 6 min to 106 ؎ 9 min. Using an in vitro mRNA degradation assay, the half-life of VEGF mRNA 3-untranslated region (UTR) transcripts were also found to be increased when incubated with hypoxic versus normoxic extracts. Both cis-regulatory elements involved in VEGF mRNA degradation under normoxic conditions and in increased stabilization under hypoxic conditions were mapped using this degradation assay. A hypoxia-induced protein(s) was found that bound to the sequences in the VEGF 3-UTR which mediated increased stability in the degradation assay. Furthermore, genistein, a tyrosine kinase inhibitor, blocked the hypoxia-induced stabilization of VEGF 3-UTR transcripts and inhibited hypoxia-induced protein binding to the VEGF 3-UTR. These findings demonstrate a significant post-transcriptional component to the regulation of VEGF.
Hypoxia has been shown to be an important stimulus for the new blood vessel formation seen in coronary artery disease (1), tumor angiogenesis (2), and diabetic neovascularization (3). VEGF, 1 also known as vascular permeability factor, is a potent angiogenic and endothelial cell-specific mitogen (4 -6), which is regulated by hypoxia in vitro (2,7,8) and in vivo (2, 3, 9 -11). The major control point for the hypoxic induction of the VEGF gene is the regulation of the steady-state level of mRNA (2,8) which is determined by the relative rates of mRNA synthesis and decay.
We have previously demonstrated that hypoxia induces VEGF steady-state mRNA 25.0 Ϯ 11.4 and 12.0 Ϯ 0.6 fold in rat primary cardiac myocytes (8) and rat pheochromocytoma PC12 cells (12), respectively. However, nuclear runoff transcription assays demonstrated that the transcription rate for VEGF was increased only 3.1 Ϯ 0.6-fold by hypoxia in the PC12 cells (12). Rat genomic sequences encoding VEGF were cloned and a 28-bp element in the 5Ј promoter was identified that mediates a significant portion of this hypoxia-inducible transcription in transient expression assays. This element was shown to have sequence and protein binding similarities to the hypoxia-inducible factor 1 binding site within the erythropoietin (Epo) 3Ј enhancer (12). These studies demonstrated that, while increased transcription rate can account for a portion of the increase in the steady-state level of VEGF mRNA in the PC12 cells, it does not account for all of the increase and suggested that a post-transcriptional mechanism plays a significant role in the hypoxic induction of VEGF mRNA, as well.
Post-transcriptional mechanisms of regulation have previously been suggested for Epo (13)(14)(15) and demonstrated for tyrosine hydroxylase (16), two other hypoxia-inducible genes. In the present study we examine the post-transcriptional regulation of VEGF mRNA expression under both normoxic and hypoxic conditions. We have employed several complementary techniques including actinomycin D chase experiments, in vitro mRNA degradation studies, and RNA electromobility shift assays.

MATERIALS AND METHODS
Cell Lines and Culture Conditions-PC12 rat pheochromocytoma cells were the generous gift of Dr. Eva J. Neer (Brigham and Women's Hospital, Boston, MA). H9c2 rat heart myocytes were obtained from the American Type Culture Collection (Rockville, MD) The cells were routinely grown in Dulbecco's modified Eagle's medium (DMEM) (Sigma) with 10% fetal bovine serum and used for all experiments at ϳ70% confluence. Cells were cultured under either normoxic conditions (5% CO 2 , 21% O 2 , 74% N 2 ) in a humidified Napco incubator at 37°C or hypoxic conditions (5% CO 2 , 1% O 2 , 94% N 2 ) in an Espec triple gas incubator (Tabai-Espec Corp., Osaka, Japan). Genistein (Sigma) was prepared as a 100 mM stock in Me 2 SO and added to cells 30 min prior to placement in the hypoxia chamber at a final concentration of 500 M.
Cloning and Sequencing of Rat VEGF cDNA-2 ϫ 10 6 bacteriophage clones from a gt11 oligo(dT)-primed PC12 cDNA library (Clontech, La Jolla, CA) were screened with two contiguous genomic fragments from the 3Ј-UTR of the VEGF gene (12), an 875-bp BamHI-EcoRI fragment (nucleotide 756-1642, GenBank TM accession no. U22372) and a 256-bp EcoRI-EcoRI fragment (nucleotide 1642-1855, GenBank TM accession no. U22372). Distinct VEGF cDNA clones hybridizing to both probes were isolated. The cDNA insert from each clone was isolated on a KpnI-SacI fragment (which contained both gt11 and VEGF sequences) and cloned into the Bluescript vector (Stratagene). Sequencing of the cDNA inserts was performed by the dideoxy chain-termination method using Sequenase (Stratagene) initially with oligonucleotide primers (5Ј-CCATCTGCTGCACGCGGAAGAAGGC-3Ј and 5Ј-CCT-TACGCGAAATACGGGCAGACATG-3Ј) corresponding to the gt11 sequence adjacent to the insert and subsequently, in a progressive fashion, with oligonucleotides complementary to the sequences obtained from the respective clones. Both strands of all clones were sequenced.
In Vitro Cell-free RNA Degradation Assay-Cells were grown under either normoxic or hypoxic conditions and S100 cytoplasmic extracts were prepared according to the method of Wang et al. (17). Briefly, cells were washed twice with ice-cold phosphate-buffered saline and then scraped into 10 ml of phosphate-buffered saline. The cells were then pelleted and resuspended in 2 volumes of homogenization buffer (10 mM Tris-HCl, pH 7.4, 0.5 mM dithiothreitol, 10 mM KCl, and 1.5 mM MgCl 2 ) and lysed with 20 strokes in a Dounce homogenizer (pestle B). 0.1 volume of extraction buffer (1.5 M KCl, 15 mM MgCl 2 , 100 mM Tris-HCl, pH 7.4, 5 mM dithiothreitol) was added, and the homogenate was centrifuged at 14,000 ϫ g for 2 min to pellet nuclei. The supernatant from this step was harvested and centrifuged at 100,000 ϫ g for 1 h at 4°C. Cytoplasmic extracts were immediately frozen on dry ice and were stored at Ϫ70°C. Protein concentrations were determined by Bradford protein assay (Bio-Rad) and were routinely 3-5 mg/ml. The entire procedure was performed at 4°C.
[ 32 P]CTP-labeled, capped, and polyadenylated transcripts were synthesized in vitro (18). The EcoRI site of pSP64 poly(A) (Promega) was transformed into an AseI restriction enzyme site by filling in EcoRIdigested pSP64 poly(A) with the Klenow fragment and then blunt-end ligating the vector. Restriction fragments containing the 3Ј-UTR of VEGF derived from clone 11.4 (12) were cloned into the multiple cloning site of this modified pSP64 vector. A series of deletions were made from the 3Ј end of these sequences using unique restriction sites in the VEGF 3Ј-UTR. Digestion of these plasmids with AseI-generated DNA templates containing a poly(dT) sequence that was transcribed into a 30base long poly(A) tail at the 3Ј end. Capped transcripts were synthesized from these templates with SP6 RNA polymerase using the MEGAscript in vitro transcription kit (Ambion, Austin, TX) according to the manufacturer's protocol with a 4:1 ratio of m 7 G5ЈpppG (cap analog) to GTP. Labeled RNA transcripts were produced by inclusion of [␣-32 P]CTP (3000 Ci/mmol) in the reaction. 80,000 cpm were used for each degradation assay.
Degradation assays were performed by incubating the transcript (10 6 cpm) with 130 g of cytoplasmic extract in a total volume of 39 l in a master mix at room temperature. At each time point the reaction was stopped by transferring 3 l from this master mix to a tube containing 15 l of H 2 O and 2 l of 5 M NH 4 OAc with 100 mM EDTA. The sample was extracted once with phenol:chloroform:isoamyl alcohol (25:24:1), and the supernatant was precipitated with 20 l of isopropanol. The pellets were washed once with 80% ethanol and air-dried. Samples were then electrophoresed on a formaldehyde-agarose gel and transferred to GeneScreen (DuPont NEN). Quantitation of the remaining primary (undegraded) transcript at the different time points was performed with a Molecular Dynamics PhosphorImager. All time points were performed in triplicate.
Measurement of VEGF mRNA Half-life in PC12 Cells-PC12 cells were grown under normoxic or hypoxic conditions in DMEM with 1% fetal bovine serum for 24 h prior to the addition of actinomycin D in T 75-cm 2 flasks (Corning). For hypoxic cell cultures, cells were grown in flasks with a solid rubber stopper containing two 18-gauge needles allowing for gas inflow and outflow. All flasks were connected in a parallel using small bore, triple-leg extension tubing (Braun Medical Inc., Bethlehem, PA) to a defined gas mixture containing 1% O 2 , 5% CO 2 , and balance N 2 . For hypoxic cultures actinomycin D (Sigma) was added to each flask through the 18-gauge outflow needle. This elaborate configuration allowed cells in each flask to be grown under hypoxic conditions for a defined period of time without any intervals of reoxygenation that would occur if all the flasks were kept in a common incubator or hypoxia apparatus. For the determination of VEGF mRNA half-life under hypoxic conditions the cells were grown under hypoxic conditions for 9 h prior to the addition of actinomycin D. For the determination of VEGF mRNA half-life under normoxic conditions, the cells were grown at 21% O 2 prior to the addition of the actinomycin D.
5 mg of actinomycin D were initially dissolved in 1 ml of Me 2 SO and subsequently diluted with DMEM to a concentration of 50 g/ml (10ϫ) stock solution. Flasks were harvested for RNA at 0, 5, 10, 15, 30, 60, 120, 240, and 480 min after the addition of actinomycin D. Total RNA was prepared from the flasks using RNA STAT-60 (Tel-Test "B," Inc., Friendswood, TX) and isolated according to the manufacturer's protocol.
VEGF and 18 S rRNA were detected by RNase protection analysis of 10 g of RNA isolated at the various time points. RNase protection assays were performed as described previously (8) to specifically detect the VEGF 165 isoform and 18 S rRNA. After electrophoresis on 6% polyacrylamide, 7 M urea gels, the protected fragments were quantitated using a PhosphorImager (Molecular Dynamics). The quantity of VEGF mRNA was normalized to the amount of 18 S rRNA (19) by calculating a VEGF/18 S ratio for each sample. All time points were performed in triplicate. The entire experiment was repeated three separate times. The half-life of VEGF mRNA was calculated by drawing the best fit linear curve on a log-linear plot of the VEGF/18 S ratio versus time. The time at half-maximal VEGF/18S ratio was taken to be the half-life.
Radiolabeled RNA transcripts (100,000 cpm/reaction) with or without nonradioactive competitor were incubated with 20 g of S100 cytoplasmic extract for 15 min at room temperature. 25 units of ribonuclease T1 were then added followed 10 min later by heparin to a final concentration of 5 mg/ml. Electrophoresis of RNA-protein complexes was carried out on 7% native polyacrylamide gel (acrylamide/methylene bisacrylamide ratio, 30:1) with 0.5 ϫ TBE (30 mM Tris, 30 mM boric acid, 0.06 mM EDTA, pH 7.3 at 20°C) at 4°C. The gels were preelectrophoresed for 1 h at 35 A followed by electrophoresis of RNA-protein complexes for 1.5 h at 30 A. Gels were dried and exposed to x-ray film (Kodak). The RNA-protein bands were quantitated by PhosphorImager analysis (Molecular Dynamics).
Statistical Analysis-Where indicated, data are presented as the mean Ϯ the standard error of the mean (23). Student's unpaired t test was employed to assess differences between two groups. Regression lines for determination of mRNA half lives were drawn using the Cricket Graph program (Cricket Software, Malverna, PA).

Determination of VEGF mRNA Transcription Termination
Site-32 cDNA clones for rat VEGF were isolated that hybridized to two contiguous fragments from the VEGF 3Ј-UTR as described under "Materials and Methods." Sequencing of six independent inserts verified a single transcription termination site determined by identification of a poly(A) tail preceded by sequence that was identical to the genomic DNA sequence for the VEGF gene previously reported (12). In all cases the transcription termination site was mapped to nucleotides 1875-1877 (GenBank TM accession no. U22372). The transcription termination site maps approximately 20 bp 3Ј to a consensus polyadenylation site (24) and is consistent with the site we previously described by Northern blot analysis (12). This polyadenylation site would generate a mRNA of 3.7 kb which is the size of the most abundant species seen by Northern blot.
Stabilization of VEGF mRNA by Hypoxia following Treatment with Actinomycin D-Quantitation of the half-life of the VEGF 165 isoform was determined as described under "Materials and Methods." The half-life was determined to be 43 Ϯ 6 min (n ϭ 3) under normoxic conditions and 106 Ϯ 9 min (n ϭ 3) under hypoxic conditions (Fig. 1) for a mean increase of 2.5 Ϯ 0.4. Similar results were obtained when the VEGF RNA was normalized to ␤-actin or to 18 S rRNA.
Mapping Instability Elements in the VEGF mRNA 3Ј-UTR in a Cell-free System under Normoxic Conditions-[ 32 P]CTP-labeled, capped, and polyadenylated transcripts containing the entire VEGF 3Ј-UTR or deletions from the 3Ј end of the UTR ( Fig. 2A) were synthesized and degradation assessed in vitro (Fig. 2, A-C) as described under "Materials and Methods." The half-life of each transcript was determined, and results were normalized to the half-life of the full-length transcript containing the entire 3Ј-UTR ( Fig. 2A). Deletion of specific sequences in the 3Ј-UTR resulted in a marked stabilization of the transcripts in this in vitro assay. Specifically, deletion of the NsiI-XbaI fragment and the StuI-NsiI fragment each resulted in an approximately 2-fold increase in transcript stability (Fig. 2, B  and C). Deletion of sequences 5Ј to the StuI site or 3Ј to the XbaI site had no significant effect on transcript stability. This would suggest the presence of two independent instability sequences in the VEGF 3Ј-UTR. Each of these instability regions co-localizes with one of the two nonameric consensus instability sequences (UUAUUUA(U/A)(U/A)) (25, 26) ( Fig. 2A).
Mapping Elements Which Mediate the Hypoxic Stabilization of the VEGF 3Ј mRNA UTR in Vitro-S100 cytoplasmic extracts from hypoxic cells were prepared in an identical fashion to those from normoxic cells. In vitro RNA degradation assays (Fig. 3, A-C) were performed as described under "Materials and Methods" and demonstrated that VEGF 3Ј-UTR transcripts had a significantly longer half-life in vitro when incubated with hypoxic versus normoxic extracts (ratio 1.5 Ϯ 0.1 n ϭ 12) (Fig.  3A). Progressive 3Ј deletion analysis of the VEGF 3Ј-UTR demonstrated that this preferential stabilization by hypoxia was lost upon deletion of the NsiI-XbaI fragment. Similar results were obtained with both PC12 and H9c2 cells.
RNA EMSA-RNA transcripts of different regions of the VEGF 3Ј-UTR incubated with S100 extract allowed for the identification by EMSA of both constitutive and hypoxia-induced VEGF mRNA binding proteins. The constitutive protein complex was found to map between the NsiI and StuI restriction enzyme sites (Fig. 4A). This complex could be completely inhibited by competition with excess unlabeled transcript from this region, but was not competed out with 500-fold excess ␤-actin or IRE transcripts (Fig. 4B). Proteinase K treatment of the extracts completely inhibited formation of the complex. Interestingly, 100-fold excess of a 162-base RNA transcript of the tyrosine hydroxylase 3Ј-UTR, corresponding to the region previously demonstrated to bind a hypoxia-inducible protein (16), completely inhibited formation of this complex. A region of the VEGF StuI-NsiI fragment was identified which is highly homologous to a region within a 28-base fragment specifically protected by the tyrosine hydroxylase RNA-binding protein (16) and to a region within the Epo 3Ј-UTR demonstrated to be the site for an RNA-binding protein (15) (Fig. 4C). Oligonucleotides were constructed from this region as described under "Materials and Methods" to define the binding site by competition studies. RNA derived from template WT 1 or WT 2 was capable of specifically competing with the protein complex binding to StuI-NsiI RNA transcripts, whereas template M, which contains a 3-nucleotide substitution in this region of homology, did not efficiently compete with the complex.
A hypoxia-inducible protein complex was mapped by EMSA between the NsiI site and the transcription termination site (Fig. 4A) using a template generated by a strategy described under "Materials and Methods." The RNA-protein complex was induced 2.2 Ϯ 0.2-fold (n ϭ 12) by EMSA using hypoxic versus normoxic S100 extracts (Fig. 4D). This complex could be inhibited by excess unlabeled transcript from this region, but was not displaced by a 500-fold excess of IRE transcript (Fig. 4D) or Epo transcripts. Proteinase K treatment of the extracts completely inhibited formation of the complex. 3Ј truncated forms of this template were generated using restriction endonucleases XbaI, EcoRI, HinfI, and MseI (Fig. 4A). RNA transcripts from these truncated templates allowed the binding site for this hypoxia-inducible species to be further defined within a MseI-XbaI fragment (nucleotide 1412-1754, GenBank TM accession no. U22372) (Fig. 4D).
Genistein Blocks Hypoxic Stabilization of VEGF 3Ј-UTR Transcripts in Vitro-PC12 or H9c2 cells were incubated with 500 M genistein, a tyrosine kinase inhibitor, for 30 min prior to their placement in the hypoxia chamber. Cells were exposed to hypoxia or normoxia for 4 h. Hypoxic and normoxic S100 extracts were then prepared in parallel from cells exposed to genistein and cells that were not exposed to genistein. As demonstrated in Fig. 5, A and B, genistein inhibited the pref- erential stabilization of hypoxic versus normoxic extracts in the in vitro RNA degradation assay. The change in the ratio of the half-life of the VEGF 3Ј-UTR transcript with hypoxic versus normoxic extracts in the presence of genistein (1.4 in the absence of genistein versus 0.5 in its presence) was equivalent to that seen when deletional analysis was performed on the VEGF 3Ј-UTR (Fig. 3A). No significant change was seen in the hypoxic induction of a previously described (12) VEGF promoter-driven luciferase construct in transient transfection assays in cells treated with genistein (data not shown). In addition, S100 extracts prepared from genistein-treated cells failed to demonstrate the hypoxia-induced increase in the electromobility shift of the NsiI transcription termination RNA fragment described above (Fig. 5C). DISCUSSION Hypoxia exerts its control on VEGF gene expression by increasing VEGF steady-state mRNA levels. Previous work has strongly suggested that an increase in transcription rate of the VEGF gene cannot account for all of the observed increase in the steady-state VEGF mRNA levels induced by hypoxia (12). These studies provide further evidence for a post-transcriptional mechanism contributing to VEGF mRNA induction by hypoxia. Several cis-acting elements that may mediate the turnover of VEGF mRNA under normoxic and hypoxic conditions are identified.
The half-life of the VEGF mRNA, determined using actinomycin D, is increased 2.5 Ϯ 0.4-fold by hypoxia. Steady-state kinetics (27) would predict that the increase in steady-state mRNA with hypoxia would be the product of the increase in the transcription rate and the increase in the mRNA half-life. These data therefore provide an adequate explanation for the discrepancy between the increase in the steady-state mRNA (12.0 Ϯ 0.6) and the increase in the transcription rate (3.1 Ϯ 0.6) in PC12 cells.
We have demonstrated using an in vitro RNA degradation assay that there are two distinct cis-acting instability elements in the VEGF 3Ј-UTR. The VEGF 3Ј-UTR contains two consensus nonameric sequences 5Ј-UUAUUUA(U/A)(U/A)-3Ј (25,26) that have been demonstrated to mediate the rapid turnover of multiple cytokine mRNAs. These nonameric consensus sequences fall within the fragments shown to significantly affect transcript stability in the in vitro degradation assays. We cannot rule out however that other sequences contained within these fragments are responsible for or contribute to the RNA instability.
The increased stability of VEGF 3Ј-UTR transcripts in vitro in the presence of hypoxic versus normoxic extracts has allowed us to map an element that when deleted from the UTR abrogates this increased stability with hypoxic extracts. From these studies one may hypothesize that a trans-acting factor that mediates stability binds to this region or, alternatively, the region is necessary for formation of a RNA secondary structure that mediates the change in RNA stability. RNA EMSA using the NsiI-StuI fragment as template was performed as described under "Materials and Methods." Unlabeled RNA transcripts for competition studies were generated from the following templates: NsiI-StuI (VEGF); IRE (22); tyrosine hydroxylase (TH) 162-bp fragment (16); oligonucleotides WT 1 , WT 2 , and M as described under "Materials and Methods." Proteinase K (PK) indicates extracts were first treated with proteinase K before adding the probe. The arrow points to the constitutive RNA-protein complex. The bracket encompasses free and degraded probe. C, sequence homology. Region of homology between the rat VEGF 3Ј-UTR, rat tyrosine hydroxylase 3Ј-UTR, and human Epo 3Ј-UTR. This region of the NsiI-StuI fragment of the VEGF 3Ј-UTR (nucleotide 1066 -1075) was demonstrated to bind a protein(s) in S100 cytoplasmic extracts. The tyrosine hydroxylase sequence is within a 28-bp sequence of the tyrosine hydroxylase 3Ј-UTR (nucleotide 1552-1579) (32) that is protected by a hypoxia-inducible protein (16). The Epo sequence (nucleotide 2831-2841) (33) is within a 120 bp sequence of the Epo 3Ј UTR shown to bind a Epo mRNA binding protein that is up-regulated by hypoxia in brain and spleen (15). Nucleotide sequence for rat VEGF, rat tyrosine hydroxylase, and human Epo refer to GenBank TM accession nos. U22372, M10244, and M11319, respectively. D, EMSA of the hypoxia-inducible complex. RNA EMSA using the NsiI-transcription termination fragment generated by PCR or agarose gel-purified restriction endonuclease digested subfragments of this fragment as templates for the generation of probe (labeled RNA transcript). [ 32 P]CTP labeled RNA transcripts were NsiI-transcription termination (N), NsiI-XbaI (X), NsiI-EcoRI (R), NsiI-HinfI (H), or NsiI-MseI (M). Unlabeled competitors were the NsiI-transcription termination transcript (N) or IRE transcripts (22) present in 100 ϫ molar excess to the labeled probe. Similar results were obtained with four different preparations of S100 extract. The arrows point to the hypoxia-inducible complex. The bracket encompasses the free and degraded probe.

FIG. 5. Genistein inhibits hypoxic stabilization of VEGF 3-UTR transcripts in vitro.
Cells were pretreated with 500 uM genistein, an inhibitor of the hypoxic induction of VEGF mRNA (32), for 30 min prior to beginning the hypoxic exposure. After 4 h extracts were prepared from normoxic and hypoxic cells. Degradation of full-length VEGF 3Ј-UTR transcript (construct A, Fig. 2A) was assessed as described under "Materials and Methods." A, representative autoradiograph of decay kinetics of the VEGF 3Ј-UTR with genistein treated or control cells using hypoxic and normoxic extracts. The arrow points to undegraded transcript. One of the RNA pellets in the triplicate 10 min 1% O 2 time point for genistein treated cells and one of the RNA pellets in the triplicate 10 min 21% O 2 time point for control cells were lost in processing and the data from these samples are not included. B, regression analysis of A demonstrating an inhibition in the stabilization of VEGF 3Ј-UTR transcripts in extracts prepared from genistein-treated cells. The experiment was performed three times with different preparations of extract. In the representative experiment shown, the ratio of the half-lives in hypoxia (Ⅺ) to normoxia (f) of the VEGF 3Ј-UTR transcript is decreased from 1.4 in control cells to 0.5 in genistein-treated cells. C, genistein inhibited formation of the hypoxia-inducible RNA-protein complex on EMSA. The NsiI-transcription termination template was used for EMSA analysis as described under "Materials and Methods." The competitor was an RNA transcript derived from the NsiI-transcription termination template (VEGF) or the IRE element (22). 1 G or 21 G indicates that the S100 extract was made under hypoxic or normoxic conditions, respectively, from cells treated with genistein. The arrow points to the hypoxia-inducible species.
In EMSA studies, the region of the VEGF mRNA 3Ј-UTR to which the hypoxia-induced protein complex bound correlated with the RNA degradation assays and points toward an important role for this complex in mediating the increased stabilization of VEGF mRNA by hypoxia. The hypoxia-inducible complex is occasionally seen to migrate as a doublet using the entire NsiI transcription termination site riboprobe (Fig. 4D). This has not been observed with truncated forms of the template (NsiI-EcoRI), although hypoxia-inducible binding is still seen. In addition, further truncation of this fragment (NsiI-HinfI) still results in binding of a protein by EMSA, but the complex is no longer hypoxia-inducible.
Genistein, a tyrosine kinase inhibitor, was recently shown to inhibit the hypoxic induction of VEGF mRNA (28) through its action on Src. A signal transduction cascade leading to the hypoxic induction of VEGF through Raf was demonstrated using the dominant inhibitory Raf-1 mutant Raf(301) (28,29). In other systems this cascade has been shown to proceed through mitogen-activated protein kinase and ultimately to modulate transcription of specific genes and phosphorylation of specific gene products (30,31). We have shown here that genistein interferes with the post-transcriptional induction of VEGF by hypoxia. 500 uM genistein had no effect on the hypoxic induction of 5Ј promoter reporter constructs but inhibited the preferential stabilization of VEGF 3Ј-UTR transcripts in vitro by hypoxic versus normoxic extracts and inhibited formation of a hypoxia-inducible protein-RNA complex by EMSA. One possibility is that the signal cascade through src/raf and possibly mitogen-activated protein kinase has as its terminal event a protein that binds to the VEGF mRNA 3Ј-UTR and mediates its hypoxic stabilization.
An understanding of the molecular basis of the regulation of VEGF by hypoxia forms the essential groundwork for the rational design of pharmacological agents to modulate VEGF expression and thereby augment or inhibit neovascularization. The in vitro degradation assays and EMSA described here should allow for the rapid and economic assessment of multiple agents that may affect VEGF mRNA stability.