INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

REDUCED OXYGEN TENSION has been suggested as a primary stimulus for VEGF expression and angiogenesis after exercise (25). This hypothesis is supported by numerous studies showing an augmented increase in VEGF levels after exercise in acute hypoxia (2, 4, 6). Hypoxia-inducible factor (HIF)-1 has also been shown to be an important transcriptional regulator of VEGF expression (15). However, whereas we have previously reported that VEGF mRNA levels were increased in the skeletal muscle of rats exposed to moderate hypoxia (rats breathing 12% O₂), HIF-1 protein levels were not altered (34). These findings imply that a HIF-1-independent mechanism for VEGF expression may also play a role in the hypoxia-induced VEGF response in skeletal muscle.

Hu proteins, first identified as specific tumor antigens in paraneoplastic neuropathy or Hu syndrome (12), are the homologues of Drosophila protein embryonic lethal abnormal visual (ELAV). ELAV proteins belong to a family of RNA binding proteins necessary for neuronal differentiation (17). So far, four Hu proteins have been identified. Three of these proteins, HuB (Nel-N1), HuC, and HuD, are expressed in terminally differentiated neurons and neuroendocrine tumors (1, 10, 21), whereas only HuR (HuA) is expressed ubiquitously in all tissues (28). Hu proteins bind with a high affinity and specificity to adenylate/uridylate-rich elements (AREs) located in the 3'-untranslated region (3'-UTR) of several mRNAs. The binding interaction of HuR to AREs either stabilizes mRNA transcripts, enhances translation, or a combination of these posttranscriptional regulatory steps (11, 13, 14, 30). Many biological conditions, including hypoxia (26), heat shock (16), and exposure to short-wave ultraviolet light (18), specifically increase the binding of HuR to specific AREs located in the 3'-UTR of mRNA species. Interestingly, heat shock increases HuR functional activity, leading to an attenuation of mRNA degradation without altering the total amount of HuR present in HeLa cells (16). After identification of a shuttling sequence (HNS) located in the hinge region between the second and third RNA recognition motifs of HuR, it has been proposed that the translocation of HuR between the nucleus and cytoplasm is a critical step that allows HuR to bind to its RNA consensus sequence and inhibit mRNA degradation (5, 22).

Posttranscriptional factor binding to AREs is a common regulatory mechanism for several immediate-early and short-lived genes, including c-myc, c-fos, Id, N-myc, GM-CSF, and GAP 43 (7, 10, 23, 24, 27, 29, 33). VEGF mRNA also contains several classic AREs located in its 3'-UTR. These specific consensus sequences have the potential to bind Hu proteins and regulate VEGF expression at the posttranscriptional level (25). Studies of cells cultured in vitro have identified a major site for HuR binding and posttranscriptional regulation between 1,631 and 1,678 bp of the VEGF 3'-UTR (26).

In the present study, we show that HuR binding activity to two distinct AREs located within the 3' regulatory region of VEGF mRNA is induced by acute ischemia in rat gastrocnemius muscle. This HuR binding activity is followed by an increase in VEGF expression.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal experiments. Six groups, each containing six young adult female Wistar rats (8-12 wk), were used in this experiment. After rats were anesthetized with pentobarbital sodium (40-60 mg/kg ip), the gastrocnemius muscle was exposed, and the hind legs were tightly clamped with plastic cable ties to completely abolish blood flow. The gastrocnemius muscle was then removed at one of the following times after the onset of ischemia: 0, 5, 10, 20, 30, and 60 min. The muscle was then immediately frozen in liquid nitrogen. Six rats were used at each of the above time points, and rats at each time point constituted each of the six groups.

Synthesis of RNA probes. To synthesize RNA probes for the EMSA, a series of truncated RNA probes, which span regions of the rat VEGF mRNA 3'-UTR sequence (25), were synthesized from DNA oligonucleotides and cloned into pBluscript II KS (Stratagene, CA) KpnI and PstI sites (Fig. 1). A short 66-bp region of pBluescript II KS (transcription of T3 to PstI site) served as the control probe. Plasmid constructs were digested with PstI, transcribed with T3 polymerase (Stratagene) in the presence of [³²P]UTP, and purified with a RNase-free spin column (Roche). For competition assays, a nonradioactive probe was synthesized from plasmid 59-71 using nonradioactive UTP.

View larger version (10K): [in this window] [in a new window]   Fig. 1.   Truncated and mutated RNA probes. The VEGF mRNA 3'-untranslated region (UTR) from base pairs 1,631-1,687 is shown. The numbers indicate the exact base pairs used as RNA probes in EMSAs. The bold letters indicate the replacement of UU with CG base pairs in site-directed probe mutations.

Electromobility shift assay. EMSA was modified from a protocol described previously (35). Frozen samples of the gastrocnemius muscle were homogenized in a buffer containing 100 mM NaCl, 50 mM Tris (pH 7.4), 0.5 mM Triton X-100, 1 mM DTT, 50 mM NaF, 0.5 mM NaVO₃, and an EDTA-free protease inhibitor cocktail tablet (Roche; Indianapolis, IN). Homogenates were centrifuged at 14,000 rpm for 10 min at 4°C, and the supernatants were collected. Protein concentrations were measured with Bio-Rad protein assay kits (Hercules, CA). HuR-mRNA binding assays included the 200-µg muscle protein sample, ³²P-labeled mRNA probe (10⁶ counts/min per reaction), and 50 µl of reaction buffer [50 mM NaCl, 50 mM KCl, 1 mM MgCl₂, 50 mM Tris · HCl (pH 7.6), 0.5% NP-40, 50 mM NaF, 5 mM DTT, and 250 µg/ml tRNA]. The reaction mixtures were incubated for 15 min at 25°C, followed by the addition of RNase T1 (2 µl, 25 U/ml), and incubated for an additional 30 min at 25°C. One microliter of 2% bromophenol blue was added to each reaction mixture, and the samples were immediately loaded onto a 6% native acrylamide gel in 1× Tris-borate-EDTA buffer and electrophoresed at 250 V for 3 h. Gels were dried and exposed to BioMax film (Kodak; Rochester, NY) for autoradiography.

Binding assay for VEGF mRNA 3'-UTR. Because there was only minimal binding of HuR to the VEGF mRNA in control protein samples, protein isolated from muscles subjected to 20 min of ischemia was used to delineate the specific HuR binding regions in the VEGF 3'-UTR. The whole binding region (probe 31-78) was analyzed and compared with a series of truncated probes (probes 59-78, 59-71, 59-65, 66-71, 31-71, 31-65, and 31-58) to determine the presence of the muscle-specific factors. In mutation assays, the core binding region [59-71 (5'-UUUUUUAAUUUUA-3')] was used as a control and compared with the binding ability of different mutated probes [60Mt (5'-UCGCGUAAUUUUA-3'), 68Mt (5'-UUUUUUAAUCGUA-3'), and 60,68Mt (5'-UCGCGUAAUCGUA-3')].

Nonradioactive probe competition and HuR antibody shift assay. The core probe 59-71 and the same protein samples used in the deletion assays were also used in assays to determine the specificity of binding. In the nonradioactive probe competition assay, either 10- or 100-fold of nonradioactive probe was added along with the radioactive probe in the reaction mixture. In the supershift assay, 1 µl HuR (0.2 µg) antibody 3A2 (Santa Cruz Biotechnology) was included in the reaction mixture. The reaction mixtures were subsequently analyzed for HuR binding activity by EMSA as described above.

Northern blot analysis of VEGF mRNA. Northern blot analyses were performed as described previously (4). Briefly, the muscle was homogenized in guanidinium buffer, and total RNA was isolated with the phenol-chloroform method (9). Total RNA (30 µg) was size fractionated by electrophoresis on a 6.6% formaldehyde-1% agarose gel and transferred to a Zeta probe membrane (Bio-Rad). The membranes were subsequently hybridized with rat VEGF cDNA probes (5) labeled with [³²P]dCTP using a random primer labeling kit (Stratagene). Blots were exposed to BioMax film (Kodak) and quantitated by densitometry. The same membranes were stripped and rehybridized with rat 18S cDNA probe labeled with [³²P]dCTP to control for uniformity of loading in each lane of the gel.

Western blot analysis for HuR and VEGF. Western blot analysis was modified from the method of Gallouzi et al. (16). The same protein samples analyzed in EMSA were used for the Western blot assays. Skeletal muscle protein (50 µg/lane) was denatured in loading buffer [0.25 M Tris (pH 6.8), 20% glycerol, 4% SDS, and 0.05% bromophenol blue] by boiling for 5 min. Samples were electrophoresed on 12% SDS-PAGE and transferred to Immobilon P membranes (Millipore; Bedford, MA). Membranes were incubated with blocking buffer [5% dry milk, 0.02% (vol/vol) Tween 20, and 0.01% (vol/vol) anti-foam A (Sigma; St. Louis, MO) in PBS] for 2 h to block nonspecific protein binding. Anti-HuR 3A2 (Santa Cruz Biotechnology) (18) or anti-VEGF (Santa Cruz Biotechnology) antibodies were incubated with the blots, respectively, as primary antibodies. After incubation overnight at 4°C, membranes were washed with PBS-0.02% Tween 20 and incubated with horseradish peroxidase-conjugated rabbit anti-mouse IgG (Amersham; Cleveland, OH) secondary antibodies. The VEGF- and HuR-specific signals were detected by chemiluminescence using an ECL kit (Amersham) and exposed to BioMax film (Kodak).

Data processing and statistics. The bands from the Northern and Western blots and EMSA were quantitated by densitometric analysis from digital scanned images using HP DeskScan (Hewlett-Packard; Wilmington, DE) and Gel-Pro Analyzer software (Media Cybernetics). Statistical significance was determined by ANOVA, and P < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Identification of VEGF mRNA 3'-UTR protein binding activity in rat skeletal muscle. In Fig. 2, no binding activity was detected using a nonspecific control probe. A clear band, reflecting slower mobility binding activity, was detected using probe 31-78 and protein extracts isolated from 20-min ischemic gastrocnemius muslce. Probe 31-78 alone (without protein sample) showed no shifted bands. Binding activity indicates that factors present in the gastrocnemius muscle bind to the core regulatory area (1,631-1,678 bp) of the VEGF mRNA 3'-UTR (17). Binding activity equivalent to that observed with probe 31-78 was also observed with probe 59-71 (Fig. 2), and this was competitively inhibited with a 10-fold excess of nonradioactive probe 59-71 (Fig. 3). Effective competition of the band observed in samples taken after 20 min of ischemia and the two bands observed in after 1 h of ischemia were also competitively inhibited by a 100-fold excess of nonradioactive probe (Fig. 3). These observations suggest that binding activities are specific for the ARE-containing sequence. Furthermore, HuR-specific antibody 3A2 added to the reaction mixture containing protein extracts isolated from muscle exposed to 20 min of ischemia resulted in a supershifted band that was observed in addition to the constitutive binding activity. This supershifted band confirms the presence of HuR in the constitutive binding activity.

View larger version (83K): [in this window] [in a new window]   Fig. 2.   Delineation of skeletal muscle protein binding activity to the VEGF mRNA 3'-UTR. Gel mobility shift assays of cellular extracts isolated from gastrocnemius exposed in vivo to 20 min of ischemia are shown. Each lane represents 200 µg of protein incubated with a 32P-labeled RNA probe that spans a region of the VEGF 3'-UTR. A slower mobility complex (arrow) is formed compared with the full-length labeled 3'-UTR probe alone (free probe). A nonspecific sequence transcribed from the Bluescript II T3 to PstI region was used as a negative control probe (NS). Probes 59-78, 59-71, 31-71, and 31-65 demonstrate binding activity similar to probe 31-78. Probe 59-65 demonstrated weak binding, whereas probes 66-71 and 31-59 showed no binding activity.

View larger version (47K): [in this window] [in a new window]   Fig. 3.   Specificity of Hu protein R (HuR)-VEGF mRNA binding activity. The specificity of VEGF mRNA 3'-UTR gastrocnemius protein binding activity in rats exposed to 20 min or 1 h of ischemia is shown. , Free-labeled probe that was not incubated with muscle extracts. In some lanes, 100-fold (100×) or 10-fold (10×) excess nonradioactive probe 59-71 was included in the reaction mixture and compared with a standard reaction containing labeled probe 59-71 alone. Nonradioactive probes competitively inhibited both constitutive and inducible binding activity. A supershift band is present in 20-min ischemia-exposed samples incubated with an HuR specific antibody (3A2).

Delineation of specific HuR binding sites in the core regulatory region of the VEGF mRNA 3'-UTR. In our analysis, by using a series of truncated probes that span regions of the VEGF 3'-UTR (Figs. 1 and 2), the maximum binding activity was detected from 1,659-1,671 bp (regions 1,631-1,658 and 1,672-1,678 both deleted). Binding activity did not appear to be enhanced with longer RNA probes 59-78, 31-71, or 31-78. Furthermore, binding activity was also detected with probe 59-65, which contains the UUUUUUA consensus sequence (Fig. 2). No binding activity was detected when very short RNA probe 66-71, containing the AUUUUA consensus sequence, was used for mobility shift assays. To further identify the precise nucleotides involved in these protein-RNA interactions, site-specific RNA mutant probes were used in a gel mobility shift assay (Fig. 4). Mutation of one ARE sequence downstream of probe 59-71 (5'-UUUUUUAAUUUUA-3') to mutated probe 68Mt (5'-UUUUUUAAUCGUA-3') or mutated probe 60Mt (5'-UCGCGUAAUUUUA-3') did not alter protein-RNA binding activity. However, mutated probe 60,68Mt (5'-UCGCGUAAUCGUA-3'), which altered both ARE consensus sequences, eliminated all protein-RNA binding activity.

View larger version (28K): [in this window] [in a new window]   Fig. 4.   Site-directed mutation of protein binding sites between 1,658 and 1,670 bp of the VEGF mRNA 3'-UTR. Gel mobility shift assay using probes with site-directed adenylate/uridylate-rich element (ARE) mutations is shown. Lane a, consensus probe 59-71; lane b, probe ARE 60Mt; lane c, probe ARE 68Mt; lane d, Probe ARE 60,68Mt. Mutation refers to changes in base pairs 1,660-1,663 (UUUU to GCGC), 1,668-1,669 (UU to GC), or in both regions. All probes were incubated with gastrocnemius muscle extracts isolated from rats exposed to 20 min of ischemia. Mutations of base pairs 1,660-1,663 (60MT) or 1,668-1,669 (68Mt) alone did not affect protein-RNA complex formation. Mutation of both AREs above (60,68Mt) completely blocked protein-RNA binding.

Increased HuR-VEGF mRNA binding activity in skeletal muscle exposed to acute ischemia. The constitutive HuR band progressively increased with the duration of ischemia (Fig. 5A). Increases above skeletal muscle protein values from nonischemic, control rats were as follows: 1.9 ± 0.17-fold at 5 min, 3.7 ± 0.25-fold at 10 min, 4.2 ± 0.31-fold at 15 min, 8.5 ± 0.76-fold at 20 min, 13.0 ± 1.86-fold at 30 min, and 12.3 ± 2.11-fold at 60 min (Fig. 5B). Upon complete hindlimb ischemia for 30 min or longer, an additional band with a slower electrophoretic mobility showing binding activity was observed in the gel mobility shift assay using the probe 59-71 (Fig. 5A). This ischemia-induced binding activity was competed by 100-fold excess nonradioactive probe 59-71 (Fig. 3).

View larger version (29K): [in this window] [in a new window]   Fig. 5.   HuR-VEGF mRNA binding in response to skeletal muscle ischemia. A: representative HuR-VEGF mRNA gel mobility shift assay with gastrocnemius muscle exposed to an acute ischemic bout. HuR binding activity was detected by EMSA using probe 59-71 and gastrocnemius muscle extracts from rats exposed to ischemia from 0 to 60 min. P, 32P-labeled probe alone. An increase in both a constitutive binding activity and a second induced binding activity is present in the 30- and 60-min ischemic samples. B: densitometric analysis of constitutive bands. *Significantly different from control (P < 0.05); n = 6.

Increased HuR-VEGF mRNA binding is temporally correlated with ischemia-induced VEGF expression. VEGF mRNA and protein levels in the ischemic gastrocnemius muscle were evaluated by Northern and Western blots, respectively. Northern blots (Fig. 6A) demonstrated that VEGF mRNA levels were increased 20 min after ischemia (2.4 ± 0.19-fold) and were further increased over the time of ischemia (2.9 ± 0.31-fold at 30 min and 3.6 ± 0.38-fold at 60 min; Fig. 6B). Western blots revealed that HuR protein levels did not change throughout the ischemic exposure (Fig. 7, A and B), whereas VEGF protein levels were significantly increased (Fig. 7, A and B): 2.1 ± 0.2-fold at 30 min and 3.2 ± 0.40-fold at 60 min after the onset of ischemia.

View larger version (45K): [in this window] [in a new window]   Fig. 6.   VEGF mRNA levels in skeletal muscle exposed to ischemic conditions. A: representative Northern blot illustrating skeletal muscle VEGF mRNA levels in response to ischemia. The 18S rRNA band is shown as a lane loading control. B: densitometric analysis of VEGF Northern blots. *Significantly different from control (P < 0.05); n = 6.

View larger version (35K): [in this window] [in a new window]   Fig. 7.   HuR and VEGF levels in skeletal muscle in response to ischemia. A: representative Western blot demonstrating an increase in VEGF, but not HuR, levels in response to skeletal muscle ischemia. B: Western blot densitometric analysis. *Significantly different from the control group (P < 0.05); n = 6.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

HuR-VEGF 3'-UTR binding in rat gastrocnemius muscle in vivo. HuR is a 36-kDa protein with three RNA binding domains. One is suggested to bind with the poly(A) tail, whereas the other two have been implicated in ARE recognition (28, 29). Like its homologue ELAV in Drosophila, HuR has been reported in vitro to interact with many mRNA 3'-UTR regions, including those found in VEGF mRNA (7, 10, 16, 23, 24, 27, 28, 33). In the current experiment, we established that HuR is present in rat skeletal muscle (Figs. 6 and 7) and that its binding activity to the VEGF 3'-UTR could be evaluated by HuR-mRNA EMSA (Fig. 2). HuR binding activity is highly specific, and this was confirmed by competition experiments with an excess of nonradioactive probe containing the ARE sequences as well as the ability of HuR-specific antibody to supershift binding activity (Fig. 3). Our RNA EMSA results also indicate that a probe containing 13 bp of nucleotides, UUUUUUAAUUUUA (1,659-1,671 bp in the VEGF 3'-UTR), is the minimal sequence required to detect HuR binding activity in protein extracts from the intact hindlimb gastrocnemius muscle.

HuR binding sites in VEGF mRNA 3'-UTR. AREs containing several adenosine and uridine ribonucleotides in the mRNA 3'-UTR have been identified in several mRNA species. This is often true for many mRNA species with a short half-life (8). In vitro RNA binding experiments have also shown that these ARE-containing 3'-UTR are regulated by HuR binding, which increases the stability of these early expressed transcripts (7, 10, 16, 23, 24, 27, 28, 33). The standard ARE specific for Hu proteins is AUUUUA (8), and this exact sequence is present in the VEGF mRNA 3'-UTR between 1,666 to 1,671 bp. This site has been suggested as an active binding site for HuR based on in vitro experiments in 3T3 cells (26). In the present work, we show that this AUUUUA site is also important for HuR binding in skeletal muscle in vivo. Mutations of the AUUUUA consensus sequence to AUGCUA abolish HuR binding to this region (Fig. 4). Interestingly, our results also indicate that a minimal sequence of AUUUUA by itself (probe 66-71) is not sufficient to retain binding activity of muscle-specific transcription factors (Fig. 2). These results suggest that base pairs immediately adjacent to the AUUUUA sequence are necessary to retain binding. Surprisingly, HuR efficiently binds to a second sequence between base pairs 1,659 and 1,665 containing UUUUUUA (Fig. 2). Mutations to this sequence also abolish HuR binding activity (Fig. 4). These results indicate that the core regulatory region of VEGF in the mRNA 3'-UTR is from 1,659 to 1,671 bp. Within this region, HuR can bind to two ARE sites, and binding of HuR to each of these AREs appears to occur with the same affinity. The functional relationship between these two binding sites will require further investigation.

Ischemia increases HuR-VEGF mRNA 3'-UTR binding in rat gastrocnemius muscle. Previous research has shown that the binding activity of HuR-VEGF mRNA 3'-UTR is oxygen sensitive. Upon exposure to 1% oxygen, the VEGF mRNA half-life increases dramatically in 293T cells, which overexpress HuR (26). In the present study, we found that the constitutive binding of HuR to the sequence containing base pairs 1,659-1,671 increased as early as 5 min after ischemia and gradually increased further over the next 30 min (Fig. 5, A and B). In addition to the constitutive HuR binding activity, an inducible binding activity to the 1,659-1,671 bp sequence was detected at 30 and 60 min after the onset of ischemia (Fig. 5A). This ischemia-inducible band was specific and could be competed by an excess of nonradioactive probe 59-71. However, this inducible binding activity could not be shifted by incubation with HuR antibody 3A2 in EMSA (Fig. 3). The possible reasons for absence of a supershifted HuR band might be that 1) a chemical modification of HuR under hypoxia blocks the binding of antibody 3A2, 2) more protein domains are bound during extended hypoxia and mask the HuR antibody binding site, or 3) another Hu-like protein may also bind to core AREs within the VEGF mRNA 3'-UTR but is not recognized by the particular HuR-specific antibody used in our experiments.

Skeletal muscle VEGF and HuR responses to ischemia. VEGF is a hypoxia-inducible growth factor (31). Its gene expression can be regulated by HIF-1 at the transcriptional level and by HuR at the posttranscriptional level (25). In our previous study (34), we found that both HIF-1 and VEGF levels in the rat gastrocnemius muscle increased similarly and substantially during exercise, whereas only VEGF but not HIF-1 increased in the gastrocnemius muscle exposed to moderate hypoxia. This implies that there is a sensitive-regulatory pathway for increasing VEGF gene expression in response to hypoxia that does not involve the expression of HIF-1 or its transacting binding to the VEGF promoter. In the present study, we found that the HuR binding activity specific for the VEGF mRNA 3'-UTR was elevated as early as 5 min after the onset of ischemia (Fig. 5), before the increase in VEGF mRNA and protein. Increased VEGF expression was detected 20 and 30 min after ischemic insult, respectively (Figs. 6 and 7). These results suggest an alternative HIF-independent and immediate posttranscriptional mechanism for increasing VEGF expression in response to acute ischemia. The mechanism by which HuR regulates mRNA levels has been proposed to involve HuR binding to hnRNA in the nucleus and subsequent translocation to the cytoplasm. In the cytoplasm, it is protected from degradation by nuclear RNase enzymes and is thus available for translation. HuR itself returns to nucleus by a HuR-specific import shuttling system. The translocation of HuR to the cytoplasm will enhance its protective function by attenuating mRNA degradation rates (5, 22). A similar mechanism has been observed in HeLa cells. Heat shock dramatically increases HuR binding to poly (A) RNA located in the cytoplasm of HeLa cells, whereas the total amount of HuR is unchanged (16). In the present experiments, we found that HuR protein levels were not altered throughout the entire 60-min ischemic period (Fig. 7). This suggests that in response to acute ischemia and hypoxia, skeletal muscle HuR expression is not the main mechanism for signaling VEGF upregulation. Alternatively, the binding activity of HuR to VEGF mRNA is increased in skeletal muscle after an acute ischemic bout (Fig. 5). These results suggest that, besides the translocation mechanism identified in reported in vitro experiments with HeLa cells, the binding activity of HuR to the mRNA 3'-UTR is also changed in skeletal muscle during hypoxia, possibly from ligand regulation of HuR. The observation that it requires only a few minutes for an increase in constitutive HuR binding activity would further suggest that a hypoxia-induced chemical modification, such as phosphorylation, is likely to be involved in this mechanism of gene regulation. The increased binding of HuR to the VEGF mRNA 3'-UTR will ultimately decrease its degradation and increase the level of VEGF mRNA in the cytoplasm available to the translational machinery.

Possible role of HuR in VEGF regulation at different levels of hypoxia. Many physiological and pathological conditions cause skeletal muscle hypoxia, including exercise, ischemia, and anemia. Even though it is hard to measure skeletal muscle tissue PO₂ directly, different methods have been used to reflect this parameter, for example, measurement of venous PO₂ (20), multiple channel surface electrode estimation of PO₂ (19, 32), and muscle microvascular measurements of PO₂ (3). Sjoberg et al. (32) reported that, after blood flow is totally blocked in the rat hindlimb, the muscle surface oxygen levels decrease to ~50% of normal levels in 10 min and approach zero by 20 min. In our experiments, we found that detectable HuR-VEGF 3'-UTR binding increased as early as 5 min after total blood flow was blocked (Fig. 5). This result demonstrates that HuR-VEGF regulation is very sensitive to hypoxia and can be activated at PO₂ levels above the 50% of normal values expected if blood flow was restricted for an additional 5 min as in the Sjoberg experiment (32). In addition to ischemic conditions, anemia, exercise, and high altitude have also been reported to decrease PO₂ in skeletal muscle. For instance, if the rabbit hemoglobin concentration is decreased to 8.6 g/dl compared with the normal value of 10.8 g/dl, extracellular tissue PO₂ in skeletal muscle falls to 47% of control values (19). Furthermore, rat gastrocnemius muscle microvascular PO₂ levels decrease to 30% of normal values after 90 s of electrically stimulated muscle contraction (3). Finally, popliteal venous PO₂ decreases by 57% when dogs are ventilated with 10% O₂-90% N₂ (20). In all these conditions, the potential for HuR-VEGF regulation to be a key regulatory step stimulating VEGF expression is high. More carefully controlled experiments under these various physiological conditions are needed to understand the role of HuR proteins in regulating VEGF expression and subsequent new capillary formation in skeletal muscle.