INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

ADENOSINE is a purine nucleoside composed of adenine and ribose joined by a glycosidic bond. It is a metabolic precursor for nucleic acid biosynthesis. All cells are possible sources of adenosine because all cells can use free energy derived from catabolism of ATP to perform their various functions. Many of the cells that "produce" adenosine also have receptors for adenosine. Adenosine receptors on the cell surface membrane are members of the G protein-coupled receptor family and are categorized as A₁, A_2a, A_2b, and A₃ (and possibly A₁ subtypes) by pharmacological interaction with a variety of agonists and antagonists (10, 39). Adenosine is a regulator of numerous bodily functions associated with the cardiovascular, renal, respiratory, neurological, immunologic, gastrointestinal, and metabolic systems (39).

Physiological effects of adenosine on the cardiovascular system include vasodilation throughout the body except in the kidney where vasoconstriction occurs, bradycardia, decreased arteriovenous nodal conduction velocity, reduction of atrial contractility, inhibition of cardiac pacemaker activity, inhibition of platelet aggregation, anti--adrenergic effects, inhibition of renin release, and renal sodium retention (36). Evidence for novel cardiovascular actions mediated by the different adenosine receptors is accumulating. The importance of adenosine as a negative feedback modulator of cell and organ energy demand and consumption is a basic tenet of adenosine biology (6, 8, 30).

It is now well known that angiogenesis is often a compensatory response to prolonged imbalances between the metabolic requirements of the tissues and the perfusion capabilities of the vasculature (2). Although angiogenesis is associated with tissue hypoxia and increased metabolic rate, the mechanisms that regulate angiogenesis under these conditions is less well understood. Adenosine is attractive as a negative feedback modulator of angiogenesis (1, 2, 11, 38) because adenosine levels in tissues are elevated during hypoxic conditions and when metabolic rate increases (7, 19, 21, 35), and adenosine can stimulate endothelial proliferation in vitro (12, 13, 28, 29) as well as growth of blood vessels in a variety of in vivo models (1-3).

Mounting evidence suggests that hypoxia-induced vascular endothelial growth factor (VEGF) plays an important regulatory role in angiogenesis in both physiological and pathological conditions. Recent evidence has shown that adenosine can enhance VEGF expression in a variety of cultured cells (15, 16, 23, 34, 37). Hashimoto et al. (23) reported that 5'-(N-ethylcarboxamido)adenosine, an adenosine analog, upregulated VEGF mRNA expression in cultured U-937 cells under normoxic conditions. Takagi et al. (37) found that 8-(3-chlorostyryl)caffeine (CSC), an adenosine A₂-receptor antagonist, reduced hypoxic stimulation of VEGF mRNA by 68% in cultured bovine retinal capillary endothelial cells. However, the mechanism by which adenosine induces blood vessel growth and the relation between adenosine and VEGF in the regulation of angiogenesis under normoxic and hypoxic conditions are still poorly understood.

The present study seeks to determine whether adenosine has a differential effect on VEGF expression under normoxic and hypoxic conditions and whether the effects of adenosine are mediated by adenosine A₁ or A₂ receptors in cultured myocardial vascular smooth muscle cells (MVSMCs) from dog coronary artery. The results suggest that adenosine, acting by way of A₂ receptors, plays a dominant role in establishing basal levels of VEGF during normoxic conditions and a relatively minor role during hypoxic conditions.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Chemicals. Adenosine was purchased from Fujisawa USA (Deerfield, IL). N⁶-cyclopentyladenosine (CPA), R(-)-N⁶-(2-phenylisopropyl)adenosine (R-PIA), N⁶-[2-(3,5-dimethoxyphenyl)-2-(2-methylphenyl)]ethyl adenosine (DPMA), 1,3-diethyl-8-phenylxanthine (DPX), and CSC were obtained from Research Biochemicals International (Natick, MA). [³²P]dCTP was obtained from DuPont New England Nuclear (Boston, MA). Multiprime DNA-labeling system was obtained from Amersham International (Amersham, UK). VEGF ELISA kits were provided by R & D systems (Minneapolis, MN).

Cell cultures. MVSMCs were isolated from adult male mongrel dogs with a method previously described (20). Briefly, the left anterior descending coronary artery was removed from the heart and cut into small pieces. The tissue pieces were digested with 0.1% collagenase (Sigma), 0.02% elastase (Sigma), and 0.02 trypsin inhibitor (Sigma) in DMEM (Sigma) for 2 h at 37°C. The cells were seeded into sterile culture flasks at ~5 × 10⁴ cells/cm² and incubated at 37°C in a humidified atmosphere of 5% CO₂-20% O₂-75% N₂. The culture medium was 50% DMEM plus 50% medium 199 (GIBCO) supplemented with 10% fetal bovine serum (Hyclone), 100 U/ml penicillin, 100 µg/ml streptomycin, and 0.25 µg/ml amphotericin B. The primary culture reached confluence in 5-7 days. The purity of the culture was confirmed by immunohistochemical staining with mouse anti--smooth muscle actin-fluorescein isothiocyanate conjugate (Sigma). The cell lines were used between passages 4 and 8 in the experiments. When monolayers of MVSMCs reached ~80% confluence, standard medium was replaced with media having 4% heat-inactivated fetal bovine serum in the absence and presence of adenosine and a variety of A₁- or A₂-receptor agonists and antagonists. Samples were taken at different periods of incubation in normoxic (20% O₂-5% CO₂-75% N₂) and hypoxic (1% O₂-5% CO₂-94% N₂) environments.

Exposure to hypoxia. The experiments were performed in a water-jacketed triple gas incubator (Queue System, Parkersburg, WV) that controls the O₂ and CO₂ to within ±0.1% of the set-point value. After 24 h of incubation in a normoxic environment, the medium was replenished with similar complete media and the ~80% confluent MVSMC monlayers were exposed to a hypoxic environment (1% O₂-5% CO₂-94% N₂) in the absence and presence of various adenosine agonists and antagonists. Cells exposed to hypoxia still excluded trypan blue dye (>95%) and showed no morphological changes by light microscopy, and the levels of lactate dehydrogenase (LD-L 20 assay kit, Sigma) were not increased in the media. MVSMCs incubated under normoxic conditions (20% O₂-5% CO₂-75% N₂) from the same batch and passage were used as controls.

Measurement of VEGF protein. The amount of VEGF was measured in the normoxic and hypoxic conditioned media of cultured MVSMCs with sandwich ELISA (R & D Systems). In these assays, the angiogenic growth factor in the test sample is sandwiched between a monoclonal antibody against human recombinant VEGF. A second polyclonal antibody against the growth factor is conjugated to horseradish peroxidase and then added to the mixture. Color develops by addition of hydrogen peroxide and chromogen tetramethylbenzidine, and the intensity is measured at 450 nm. VEGF protein levels were normalized to the total amount of cellular protein and expressed as picograms per milligrams of total cell protein. Cell protein content was determined in duplicate with BSA as the standard (Bio-Rad Protein Assay Kit, Bio-Rad Laboratories).

Northern blot analysis. Total RNA was prepared with a total RNA isolation kit (catalog no. 1910; Ambion, Austin, TX). Total RNA (20 µg) was electrophoresed through a 1% agarose gel containing 2.2 M formaldehyde in 1× MOPS buffer. The gel was stained with ethidium bromide to visualize rRNA, and the RNA was transferred to nitrocellulose paper (Schleicher & Schuell, Keene, NH) overnight by capillary action in 20× saline sodium citrate (SSC; 3 M NaCl, 0.3 M Na₃citrate · 2H₂O, pH 7.0). After transfer, the nitrocellulose paper was baked at 80°C under vacuum for 2 h and then prehybridized for 3 h at 42°C in a buffer containing 25 mM KPO₄ (pH 7.4), 6× SSC, 5× Denhardt's solution, 50% formamide, and 50 µg/ml salmon sperm DNA. Dextran (10%, final concentration) was then added to the prehybridization solution. cDNA probes were denatured by boiling and added to the hybridization solution. After hybridization overnight at 42°C, the blots were washed at 42°C in 0.2× SSC/0.1% sodium dodecyl sulfate until the background radioactivity was removed.

The VEGF cDNA probe is a 580-bp EcoR I-BamH I fragment of the murine VEGF cDNA cloned into pBluescript plasmid [kindly provided by Dr. Werner Risau (33) at Max-Planck Institute]. The cDNA probe was labeled with [-³²P]dCTP (New England Nuclear, Boston, MA) with a multiprime random-primed DNA labeling kit (Amersham, UK). Quantification of VEGF mRNA expression was performed on phosphor images of blots collected with a PhosphoImager (Molecular Dynamic; software: ImageQuant, version 3.3, Molecular Dynamics). To verify the relative amount of total RNA, filters were hybridized with a ³²P-labeled 28S rRNA antisense oligonucleotide probe (Ambion). The VEGF mRNA was normalized against 28S rRNA in each sample.

Statistical analysis. All determinations were performed in duplicate, and each experiment was repeated at least three times. Where indicated, data are presented as means ± SD. Differences were considered statistically significant when P < 0.01 by paired t-test. All statistical calculations were performed with StatView software (BrainPower, Calabasa, CA).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

VEGF protein expression in normoxia. VEGF protein levels in media were measured with sandwich ELISA (R & D Systems) after MVSMCs were cultured under normoxic conditions in the absence and presence of adenosine and adenosine A₁- and A₂-receptor agonists and antagonists. Adenosine administered into culture media caused a dose-related increase in VEGF protein expression (Fig. 1A). VEGF protein levels, expressed as nanograms per milligrams total cell protein, increased by 30, 57, 87, and 187%, respectively, in MVSMCs after 18-h exposure to adenosine at 1 µM, 10 µM, 50 µM, and 500 µM, compared with control (control VEGF concentration, 0.60 ± 0.02). However, exposure to 1,000 µM adenosine caused VEGF protein levels to decrease to 77% of control (0.60 ± 0.02 vs. 0.46 ± 0.04, P < 0.01).

View larger version (18K): [in this window] [in a new window]   Fig. 1.   A: dose-related effect of adenosine (Ado) on vascular endothelial growth factor (VEGF) protein expression in myocardial vascular smooth muscle cells (MVSMCs) cultured under normoxic conditions. Cells were cultured in absence (control as 0) and presence of various concentrations of Ado for 18 h under normoxic conditions (5% CO2-20% O2-75% N2). B: effects of Ado and Ado A1- and A2-receptor agonists on VEGF protein expression in MVSMCs cultured under normoxic conditions. When monolayers reached ~80% confluence, cells were cultured in absence (control) and presence of Ado (5 µM), A1-receptor agonists [R(-)-N6-(2-phenylisopropyl)adenosine (R-PIA), 100 nM and N6-cyclopentyladenosine (CPA), 100 nM], and A2-receptor agonist {N6-[2-(3,5-dimethoxyphenyl)-2-(2-methylphenyl)]ethyl adenosine (DPMA), 100 nM} for 18 h under normoxic conditions. C: effects of Ado and Ado A1- and A2-receptor antagonists on VEGF protein expression in MVSMCs cultured under normoxic conditions. Cells were cultured in absence (control) and presence of Ado (5 µM), A2-receptor antagonist [8-(3-chlorostyryl)caffeine (CSC), 5 µM], and A1-receptor antagonist [1,3-diethyl-8-phenylxanthine (DPX), 5 µM] for 18 h under normoxic conditions. VEGF protein levels in media were measured by ELISA and expressed as ng/mg total cell protein. Data are means ± SD from 2 independent series of experiments (n = 8). * P < 0.01 vs. control. ** P < 0.01 vs. Ado.

Adenosine (5 µM) and the adenosine A₂-receptor agonist DPMA (100 nM) increased VEGF protein levels by 51% (0.77 ± 0.09 vs. 1.16 ± 0.04, P < 0.01) and 132% (0.77 ± 0.09 vs. 1.79 ± 0.21, P < 0.01), respectively, when MVSMCs were cultured under normoxic conditions (Fig. 1B). However, the adenosine A₁-receptor agonists, CPA and R-PIA did not increase VEGF protein levels at A₁-receptor stimulatory concentrations (Fig. 1B). The A₂-receptor antagonist CSC (5 µM) completely blocked adenosine-induced increases in VEGF protein expression and decreased baseline VEGF protein levels by 44% under normoxic conditions (Fig. 1C: 0.77 ± 0.09 vs. 0.43 ± 0.05, P < 0.01). The A₁-receptor antagonist DPX (5 µM) had no effect on VEGF protein levels (Fig. 1C).

The changes in VEGF expression caused by adenosine and A₂-receptor agonist and antagonist that occurred after 18 h of treatment (Fig. 1) were also present after 2, 6, and 12 h of treatment (Fig. 2). In these experiments, small samples of media (200 µl) were taken from T-75 flasks (media volume, 15 ml) at various times after cells were exposed to fresh media in the absence (control) and presence of adenosine (5 µM) and A₂-receptor agonist DPMA (100 nM) or antagonist CSC (5 µM). VEGF protein levels in media increased progressively in control cultures throughout the 12-h experiment (Fig. 2). Adenosine (5 µM) and A₂-receptor agonist DPMA (100 nM) significantly increased VEGF protein levels by 27 and 60%, respectively, after only 2 h of treatment compared with the 2-h control group (P < 0.05), and the effects of adenosine and DPMA to increase VEGF levels in media persisted for the next 10 h (Fig. 2). The A₂-receptor antagonist CSC (5 µM) decreased VEGF protein levels by 16% (P < 0.05) after only 2 h of treatment, and the inhibitory effects of CSC continued throughout the experiment.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Time-related effects of Ado and A2-receptor agonist or antagonist on VEGF protein expression in MVSMCs cultured under normoxic conditions. Samples (200 µl) were taken from T-75 flasks at various times (2, 6, and 12 h) after cells were exposed to fresh media in absence (control as 0) and presence of Ado (5 µM), A2-receptor agonist (DPMA, 100 nM), or A2-receptor antagonist (CSC, 5 µM) under normoxic conditions. VEGF protein levels in media were measured by ELISA and expressed as ng/mg total cell protein. Data are means ± SD from 2 independent series of experiments (n = 8). * P < 0.01.

VEGF protein expression in hypoxia. Hypoxia (hypoxic control) caused a 4.3-fold increase in VEGF protein levels in media after 18 h of treatment (Fig. 3), compared with the normoxic control group (normoxic control, 0.77 ± 0.09 vs. hypoxic control, 3.28 ± 0.63). Neither adenosine (5 µM), A₂-receptor agonist DPMA (100 nM), nor A₁-receptor agonist DPX (5 µM) could increase VEGF protein expression beyond that of the hypoxic control group (Fig. 3). Interestingly, A₂-receptor antagonist CSC (5 µM) reduced VEGF protein levels by only 20% during 18 h of incubation under hypoxic conditions, compared with the hypoxic control group (Fig. 3; hypoxic control, 3.28 ± 0.63 vs. hypoxic CSC, 2.62 ± 0.50).

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Effects of Ado and Ado-receptor agonists and antagonists on VEGF protein expression in MVSMCs cultured under hypoxic conditions (5% CO2-1% O2-94% N2). Cells were cultured in absence [control (C)] and presence of Ado (5 µM), A2-receptor agonist (DPMA, 100 nM), A2-receptor antagonist (CSC, 5 µM), and A1-receptor antagonist (DPX, 5 µM) for 18 h under hypoxic (H) or normoxic (N) conditions. VEGF protein levels in media were measured by ELISA and expressed as ng/mg total cell protein. Data are means ± SD from 2 independent series of experiments (n = 8). * P < 0.01 vs. hypoxic control.

Dose-related inhibition of VEGF expression. Dose-related inhibitory effects of A₂-receptor antagonist (CSC) on VEGF protein expression were examined in MVSMCs incubated in hypoxic and normoxic environments (Fig. 4). Cells were cultured in the absence and presence of various concentrations of CSC (0.5 µM, 5.0 µM, and 50 µM) for 18 h. VEGF protein levels in media were measured by ELISA and expressed as percentage of control. CSC caused a greater percent decrease in VEGF protein expression during normoxic conditions compared with hypoxic conditions at each concentration (Fig. 4). At the highest dosage of CSC (50 µM), VEGF protein levels in media were decreased by ~60% under normoxic conditions and by ~25% under hypoxic conditions. This suggests that adenosine may be more important for establishing basal levels of VEGF during normoxic conditions, compared with hypoxic conditions.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Dose-related inhibitory effect of Ado A2-receptor antagonist (CSC) on VEGF protein expression in MVSMCs cultured under hypoxic and normoxic conditions. Cells were cultured in absence (control as 0) and presence of various concentrations of CSC for 18 h. VEGF protein levels in media were measured by ELISA and expressed as percent of control. Data are means ± SD from 2 independent series of experiments (n = 8). * P < 0.01.

VEGF mRNA expression in normoxic and hypoxic conditions. Northern blot analyses were used to determine the effect of adenosine and adenosine-receptor agonists and antagonists on VEGF mRNA expression in MVSMCs cultured under normoxic (Fig. 5) and hypoxic (Fig. 6) conditions. Adenosine (5 µM) and the A₂-receptor agonist DPMA (100 nM) increased VEGF mRNA by 44 and 90%, respectively, after 18-h treatment, compared with the normoxic control group (Fig. 5; control, 1.03 ± 0.16; adenosine, 1.48 ± 0.38; DPMA, 1.96 ± 0.49; P < 0.01 for both). The A₂-receptor antagonist CSC (5 µM) and CSC (5 µM) plus adenosine (5 µM) decreased VEGF mRNA by 41 and 17%, respectively, whereas A₁-receptor antagonist DPX (5 µM) had no effect (Fig. 5). Hypoxia alone caused a 4.1-fold increase in VEGF mRNA expression compared with the normoxic control group (Fig. 6: normoxic control, 1.03 ± 0.16 vs. hypoxic control, 4.27 ± 1.01). A₂-receptor antagonist CSC (5 µM) caused a 17% decrease in VEGF mRNA expression (hypoxic CSC, 3.54 ± 0.76) compared the hypoxic control group (P < 0.05). There were no significant differences in VEGF mRNA expression in MVSMCs treated with adenosine and adenosine plus CSC compared with the hypoxic control group (Fig. 6).

View larger version (33K): [in this window] [in a new window]   Fig. 5.   Effects of Ado and Ado-receptor agonists and antagonists on VEGF mRNA expression in MVSMCs cultured under normoxic conditions. Cells were cultured in absence (control) and presence of Ado (5 µM), A2-receptor agonist (DPMA, 100 nM), A2-receptor antagonist (CSC, 5 µM), CSC plus Ado, and A1-receptor antagonist (DPX, 5 µM) for 18 h under normoxic conditions. VEGF mRNA expression was determined by Northern blot analysis. Data are means ± SD from 2 independent series of experiments (n = 8). * P < 0.01.

View larger version (32K): [in this window] [in a new window]   Fig. 6.   Effects of Ado and A2-receptor antagonist on VEGF mRNA expression in MVSMCs cultured under hypoxic conditions. Cells were cultured in absence (control) and presence of Ado (5 µM), A2 antagonist (CSC, 5 µM), and Ado + CSC for 18 h under hypoxic (H) and normoxic (N) conditions. VEGF mRNA expression was determined by Northern blot analysis. Data are means ± SD from 2 independent series of experiments (n = 8). * P < 0.05 vs. hypoxic control.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The results show that exposing MVSMCs to media containing adenosine or the A₂-receptor agonist DPMA caused a significant increase in the expression of VEGF protein and mRNA under normoxic conditions. Previous studies suggested this might be so. Fischer et al. (16) demonstrated that adenosine upregulated VEGF mRNA expression in a dose- and time-dependent manner in porcine brain-derived microvascular endothelial cells under normoxic conditions. Takagi et al. (37) reported that the adenosine A₂ agonist DPMA increased VEGF mRNA expression in cultured bovine retinal pericytes under normoxic conditions. However, the role of adenosine in the mediation of hypoxia induced upregulation of VEGF expression is less well understood. The results shown on Fig. 3 indicate that neither adenosine nor the adenosine A₂-receptor agonist DPMA could upregulate VEGF expression in MVSMCs under hypoxic conditions. The results also indicate that the A₂-receptor antagonist CSC (5 µM) reduced the expression of VEGF protein and mRNA by only ~20% under hypoxic conditions, whereas a ~40% reduction was observed in normoxic cultures. Therefore, these findings support the hypothesis that adenosine, acting by way of A₂ receptors, plays a major role in the regulation VEGF expression during normoxic conditions and has a relatively minor role during hypoxic conditions.

Adenosine is thought to have a physiological role as an angiogenic factor (1-3, 8, 38). This hypothesis can be attributed to the following findings: 1) adenosine can stimulate the proliferation of endothelial cells in vitro (12, 13, 28, 29) but not the proliferation of other types of cells (12). 2) Blood vessel growth in chick embryos is stimulated by dipyridamole, an adenosine reuptake inhibitor, and greatly attenuated by aminophylline, a nonselective adenosine-receptor antagonist (2, 3). 3) Adenosine production is increased in hypoxic tissues as well as in tissues with increased metabolic rate (7, 19, 35); in fact, the interstitial concentration of adenosine can increase from a normal value on the order of ~100 nM (32) to as high as 40 µM (21) during hypoxic conditions. 4) Exogenous adenosine can stimulate blood vessel growth in the intact animal, and both hypoxia-induced angiogenesis and endothelial proliferation can be partly attributed to adenosine accumulation in the hypoxic environment (for review, see Refs. 1 and 2).

The mechanism by which adenosine stimulates blood vessel growth is still poorly understood. However, the results of the present study as well as studies from other laboratories suggest that adenosine can enhance VEGF expression in a variety of cultured cells (15, 16, 23, 34). VEGF is a key mediator of angiogenesis in both physiological and pathological conditions (14, 17, 25) and can be upregulated by various cytokines, growth factors, steroid hormones, and mutated oncogenes in tumor cells (14, 31); however, the induction of VEGF expression by hypoxia is considered to be most significant from a physiological perspective. VEGF regulation by hypoxia occurs at both the transcriptional and posttranscriptional levels. Transcriptional regulation of VEGF is mediated by hypoxia-inducible factor 1, which accumulates under hypoxic conditions and activates VEGF transcription by binding to specific promoter sequences (18). In addition, hypoxia leads to stabilization of VEGF mRNA (24, 26), which acts to increase steady-state levels of VEGF mRNA. Recent evidence suggests that the protein kinase C pathway (9), the Ca²⁺ channel (27), and the cAMP-dependent pathway (9) can also regulate VEGF gene expression. Fischer et al. (16) reported that rolipram, an inhibitor of cAMP phosphodiesterase, completely blocked the upregulation of VEGF mRNA induced by adenosine in cultured microvascular endothelial cells. Pueyo et al. (34) found that a 2.2-fold stimulation of VEGF mRNA expression by adenosine was enhanced after cotreatment with cobalt chloride, a hypoxia-mimicking agent.

The results of the present study indicate that the adenosine A₂ antagonist CSC could not block more than ~20-25% of the VEGF expression induced by hypoxia. These results together with those from studies discussed previously support the hypothesis that adenosine has a relatively minor role in the upregulation of VEGF expression caused by exposure to a hypoxic environment. However, this contention is based entirely on in vitro experimentation in which tissue hypoxia was simulated by exposing cells to a 1% oxygen atmosphere for a relatively short period of time. The quantitative role of adenosine in the mediation of hypoxia-induced VEGF expression in vivo is likely to depend on the degree of hypoxia to which the tissues are exposed as well as the duration and frequency of the hypoxic exposure. Additional studies will be required to confirm or refute our present hypothesis that adenosine has a relatively minor role in the hypoxic induction of VEGF.

An interesting finding in the present study was that the adenosine A₂ antagonist CSC reduced the expression of VEGF protein and mRNA by >60% under normoxic conditions. It is unlikely that the suppression of VEGF resulted from a toxic and/or nonspecific effect of the molecule for the following reason: a 10-fold increase in the concentration of CSC (from 5 to 50 µM) did not cause a further lowering of VEGF protein levels in the hypoxic group, as shown in Fig. 4. Additional studies showed that basal levels of VEGF protein in the media of rat cardiomyoblasts cultured under normoxic conditions could be decreased from a control value of 1.84 ± 0.05 to a value of 0.74 ± 0.06 ng/mg total cell protein by subjecting the cells to 10 U/ml adenosine deaminase in the media for 18 h (n = 6, P < 0.01). Because adenosine deaminase converts adenosine to inosine, an inactive metabolite, it is clear that adenosine could account for at least 60% of the basal levels of VEGF protein in the media under normoxic conditions. These findings support the hypothesis that adenosine plays an autocrine role to stimulate VEGF expression.

On the basis of these findings, we can speculate that adenosine, acting by way of VEGF protein, is a maintenance factor for the vasculature. What is a vascular maintenance factor? The concept of a "vascular maintenance factor" is based on the knowledge that the size of a capillary network is dependent on the metabolic activity of the tissues that it serves. Factors that increase metabolic activity can stimulate angiogenesis in skeletal muscles; whereas decreases in metabolic activity caused by immobilizing a limb in a plaster cast, cutting the Achilles tendon, or unloading hindlimb muscles by tail suspension in rats can decrease capillarity in the affected skeletal muscles by as much as 50% within a few weeks (for review, see Ref. 2). Therefore, a normal level of muscular activity seems to be required to maintain the structural integrity of the capillary network. We are speculating that adenosine is a vascular maintenance factor not only because adenosine may have an autocrine role to maintain basal levels of VEGF in the muscles, but also because the levels of adenosine in a muscle are proportional to metabolic rate of the muscle (5), as discussed further below.

Inasmuch as the concentration of vascular maintenance factor in a muscle should be proportional to the capillarity of the muscle, we would predict that oxidative muscles with their relatively high capillarity should have higher levels of vascular maintenance factor under normoxic conditions compared with glycolytic muscles that have relatively low capillarity. The following findings are consistent with this prediction: 1) the concentration of 5'-nucleotidase, which converts AMP to adenosine, is higher in oxidative compared with glycolytic muscles in many different species of animals, suggesting that adenosine levels are higher in oxidative muscles (5). 2) Rabbit oxidative muscles (semitendinosis and soleus) have higher levels of VEGF protein compared with glycolytic muscles (plantaris and tibialis anterior) under basal, resting conditions (4). 3) When a glycolytic muscle is converted to an oxidative muscle by prolonged electrical stimulation of a motor nerve, an approximately twofold increase in capillarity is associated with an approximately twofold increase in basal levels of VEGF mRNA in the muscle tissues (22).

Therefore, the available data from intact animals show correlations between muscle capillarity and basal levels of adenosine and VEGF protein in the muscles under normoxic conditions. These correlations, along with our new finding that endogenous adenosine has an autocrine role to establish basal levels of VEGF protein under normoxic conditions, are all consistent with the supposition that adenosine, acting by way of VEGF, is a maintenance factor for the vasculature. It will be interesting to determine whether increasing endogenous levels of adenosine in the muscles of intact animals will increase basal levels of VEGF in the tissues and therefore increase the capillarity under normoxic conditions or whether adenosine antagonists will decrease the basal levels of VEGF and cause rarefaction of the vasculature under normoxic conditions.

The present study indicates that adenosine has differential effects on the regulation of VEGF expression in MVSMCs cultured under normoxic compared with hypoxic conditions. This conclusion is based on the following findings: 1) exogenous adenosine or the adenosine A₂ agonist DPMA did not increase VEGF expression in hypoxic cultures of MVSMCs; however, both factors upregulated VEGF protein and mRNA expression under normoxic conditions. 2) The adenosine A₂ antagonist CSC (5 µM) reduced the expression VEGF protein as well as mRNA by ~40% in normoxic cultures but only by ~20% under hypoxic conditions. 3) A dose-response experiment (Fig. 4) showed that the maximal effects of the adenosine A₂ antagonist CSC could block only 25% of the VEGF protein expression induced by hypoxia but could decrease basal levels of VEGF protein by >60% under normoxic conditions. Why does exogenous adenosine or adenosine A₂ agonist induce VEGF expression in normoxic cultures but not in hypoxic cultures? A likely possibility is that hypoxia greatly increased the adenosine levels in the cultures so that maximum adenosine effects were achieved in the absence of exogenous adenosine or adenosine A₂ agonist. This differential response to adenosine in regulating VEGF expression under normoxic and hypoxic conditions could have a significant clinical impact in terms of developing angiogenic therapy with adenosine and adenosine agonists or developing antiangiogenic therapy with adenosine antagonists.

Adenosine caused a dose-related increase in VEGF protein expression with a maximal approximately threefold increase in VEGF protein levels at an adenosine concentration of 500 µM, as shown in Fig. 1A. However, 1,000 µM adenosine decreased VEGF protein levels by 23%, compared with the control group. These results are consistent with a study by Ethier et al. (12), which showed that adenosine caused a dose-related increase in the proliferation of human umbilical vein endothelial cells over a wide range of adenosine concentrations but caused a reduction in cell number at an adenosine concentration of 1,000 µM. We can speculate that the inhibitory effects of adenosine on VEGF expression and endothelial cell growth are caused by toxic effects of adenosine at high concentrations.

In comparing the effectiveness of the various adenosine A₁- or A₂-receptor agonists and antagonists, our data demonstrate that the effects of adenosine on VEGF expression are mediated by adenosine A₂ receptors but not A₁ receptors in cultured MVSMCs. These findings are consistent with previous reports from Hashimoto et al. (23) and Takagi et al. (37). Hashimoto et al. (23) reported that exogenous adenosine upregulated VEGF mRNA via adenosine A₂ receptors in cultured U-937 cells under normoxic conditions. Takagi and associates (37) found that adenosine stimulated VEGF gene expression through the stimulation of adenosine A_2a receptors in cultured retinal vascular cells. However, Fischer and associates (16) reported that the effect of adenosine on VEGF mRNA expression was mediated via A₁ rather than A₂ receptors in cultured pig cerebral capillary endothelial cells. This disparity may reflect species differences or cell or organ specific differences in the distribution of A₁ and A₂ receptors. For example, A₁ receptors are found in greatest numbers in the brain, spinal cord, testis, and adipose tissue, and they are found in lesser numbers in many other tissues, including kidney, spleen, and heart (39).

In conclusion, we have demonstrated that exposing MVSMCs to media containing adenosine or the adenosine A₂ agonist DPMA caused a significant increase in the expression of VEGF protein and mRNA under normoxic conditions but not under hypoxic conditions. The adenosine A₂ antagonist CSC greatly decreased the expression of VEGF protein and mRNA under normoxic conditions but had a relatively small effect in hypoxic cultures. These findings suggest that 1) adenosine, acting by way of A₂ receptors, plays a major role in regulating VEGF expression during normoxic conditions and minor role during hypoxic conditions; 2) adenosine has an autocrine role to stimulate VEGF secretion and can account for the majority of basal VEGF secretion by MVSMCs under normoxic conditions; and 3) most of the hypoxic induction of VEGF expression is independent of adenosine accumulation in the microenvironment of the hypoxic cells.