INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

ADENOSINE MEDIATES NUMEROUS physiological and cardioprotective effects (20). However, the local adenosine concentration that is necessary to mediate these effects at the cell surface receptors is unknown, mainly because of rapid adenosine metabolism (19) and a high degree of compartmentalization (23). With respect to the vascular effects, it is traditionally assumed that adenosine is produced in cardiomyocytes or endothelial cells (23) and acts on vascular smooth muscle A₂ receptors (21). This simplified model regards the smooth muscle cell as an effector cell and does not consider the influence of smooth muscle metabolism on the adenosine concentration near the respective surface receptors. However, several lines of evidence indicate that smooth muscle cells have an active adenosine metabolism: 1) Inhibition of adenosine deaminase, adenosine kinase, and adenosine membrane transport, respectively, enhance the adenosine concentration in the supernatant of cultured smooth muscle cells (8). 2) Smooth muscle cells release ATP (4) and hydrolyze adenine nucleotides to adenosine via an extracellular nucleotidase cascade (22). 3) Smooth muscle cells may change their adenosine release in response to changes in the acid-base status (2). However, the flux rates of adenosine metabolism of vascular smooth muscle cells (VSMC) are unknown, as is the associated cellular surface concentration of adenosine.

The regulation of the adenosine surface concentration of VSMC is not only important for purine receptor-mediated regulation of vessel diameter. Extracellular ATP and adenosine exert antagonistic effects on the proliferation of VSMC, inasmuch as ATP stimulates (10), whereas adenosine inhibits (8, 9), cell growth. It is therefore of crucial importance to understand quantitatively the native production of extracellular adenine nucleotide-derived adenosine and its significance for the resulting cell surface adenosine concentration. Thus the aims of the present study were threefold: 1) to quantify adenosine flux terms in VSMC, 2) to determine the transmembrane concentration gradient of adenosine, and 3) to assess the adenosine cell surface concentration that results from smooth muscle cell adenosine metabolism only. For comparison we studied these objectives also in porcine aortic endothelial cells (PAEC), for which a rather complex data set is available (6, 23).

	MATERIAL AND METHODS

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

Cell isolation. VSMC were isolated by enzymatic treatment of porcine coronary arteries as follows. For each isolation procedure, the epicardial arteries (1-4 mm diameter) of two hearts were dissected from the surrounding tissue and cut into 1- to 2-mm-thick rings. Rings were incubated for 45 min in 10 ml of an enzyme solution containing 160 mg of collagenase (Serva, Heidelberg, Germany), 800 mg of BSA, 6.2 µg of dithiothreitol, and 80 mg of soybean trypsin inhibitor dissolved in 40 ml of HEPES buffer [containing (in mM) 130 NaCl, 0.1 CaCl₂, 2 MgCl₂, 6 KCl, 10 taurine, 10 glucose, and 10 HEPES]. Cells isolated by this procedure were separated from undigested tissue by filtration through a 100-µm nylon mesh and concentrated by centrifugation (10 min, 250 g). For an optimal cell yield, enzymatic treatment of vessel rings was repeated twice for 45 min each. Cell pellets were resuspended and cultivated in medium 199 (GIBCO, Eggenstein, Germany) supplemented with 20% newborn calf serum (PAA Laboratories, Cölbe, Germany), 5% amphotericin B (Fungizone, 250 µg/ml), 5% penicillin-streptomycin (5,000 IU/ml), and 0.5% gentamicin (50 µg/ml, all from GIBCO). After 24 h, the concentration of antibiotics was reduced to 2% amphotericin B and 2% penicillin-streptomycin; no more gentamicin was supplied.

Cell identification. VSMC were identified by their extended growth pattern in phase-contrast light microscopy and staining with FITC-coupled anti--smooth muscle actin antibody (24) (Sigma Chemical, Munich, Germany). PAEC were identified by their typical appearance in phase-contrast light microscopy and uptake of fluorescence-labeled acetylated low-density lipoprotein (27) (Paesel, Frankfurt, Germany). Purity of cell cultures was quantified by a double-staining technique using the respective cell-specific immunofluorescent marker and the nucleus-specific, cell-unspecific fluorescent dye bisbenzimide. Purity was expressed as numbers of cells labeled with the specific fluorescence relative to the number of nuclei.

Experimental setup. Confluent VSMC and PAEC (usually after 10-14 days) were trypsinized and transferred to Biosilon microcarrier beads (Nunc, Wiesbaden-Bieberich, Germany) with a seeding density of 10⁶-10⁷ cells/g beads. Confluence of VSMC and PAEC on beads was routinely assessed by phase-contrast light microscopy. Microcarrier beads (2 ml) covered with a confluent cell layer were transferred into a watertight chromatography column (10 mm ID) and bathed in 37°C water. The column was perfused (2 ml/min) with a HEPES-buffered Krebs-Ringer solution equilibrated with room air in a reservoir at 37°C (6). Small filter plates (pore size <150 µm) at both ends of the column prevented loss of microcarrier beads during perfusion. Experiments were started 30 min after column loading (6). For determination of adenosine release, individual samples were collected over 3 min, kept on ice during the remaining experiment, and frozen at 20°C until analysis.

Total adenosine production. Total adenosine production rates for VSMC and PAEC were assessed during concerted block (7) of adenosine deaminase (EHNA) and adenosine kinase (ITU), respectively. On the basis of the assumption that membrane transport is not a rate-limiting factor, the steady-state adenosine release rate under this condition represents the total adenosine production rate.

Intracellular adenosine production. In the presence of EHNA plus ITU, adenosine membrane transport was inhibited with NBTI. If adenosine is mainly produced intracellularly, adenosine release is expected to decrease during infusion of NBTI (7). Additional infusion of AOPCP, which blocks ecto-5'-nucleotidase, was used to provide independent evidence for a contribution of this enzyme to total adenosine production.

Transmembrane adenosine concentration gradient. Adenosine membrane transport was blocked under control conditions. A decrease in adenosine release in response to infusion of NBTI alone would indicate a concentration gradient directed from the inside to the outside of the cell. An increase in adenosine release during this intervention, however, would indicate a concentration gradient directed from the outside to the inside of the cell. During NBTI treatment, infusion of AOPCP or levamisole was used to provide independent evidence for extracellular adenosine production via ecto-5'-nucleotidase or alkaline phosphatase, respectively.

Adenosine surface concentration. In the cell-column model, the inflow adenosine concentration is typically zero, while the outflow adenosine concentration ranges from 10 to 30 nM. Previous model simulations (14) have indicated that, for isolated perfused tissue systems, the underlying axial adenosine concentration profile increases monotonically along the system path length and that the slope of this axial concentration gradient is a function of the transit time. With long transit times, the axial concentration profile will level off toward the outflow. If such a steady-state concentration would be reached in our experimental model, the outflow concentration would represent a valid estimate of the cell surface concentration. The presence of such an axial steady state can be tested by infusing adenosine into the cell-column inflow while measuring the adenosine concentration of the column effluent perfusate. If enhancement of the adenosine inflow concentration above the control adenosine outflow concentration would not change the outflow adenosine concentration, this would be experimental evidence for the existence of an axial steady-state concentration. Therefore, the inflow adenosine concentration was increased stepwise from 0 to 60 nM, and, starting 5 min after each concentration change, three effluent samples were collected for 3 min each. In additional experiments, whether the adenosine surface concentration is changed under conditions of impaired intracellular adenosine metabolism was tested. Hence, during simultaneous infusion of TUB and EHNA, the adenosine surface concentration was determined by infusion of 15-200 nM exogenous adenosine. To prove the inflow adenosine concentration experimentally, in both experimental sets, samples were also taken directly from the column inflow.

Cytosolic volume. The cytosolic volume was determined using a modification of the method by Kletzien et al. (13). Briefly, confluent cell layers of VSMC and PAEC (density 6 × 10⁴ cells/cm²) were thoroughly washed with PBS and incubated with 0.5 ml of HEPES buffer [containing (in mM) 140 NaCl, 3.8 KCl, 1 CaCl₂, 1.2 MgCl₂, 5 glucose, 0.5 methylglucose, and 20 HEPES, pH 7.4] containing 3 µl of D-[3-O-methyl-¹⁴C]glucose (250-370 mCi, Dupont de Nemours, Mechelen, Belgium) per well (area 9.6 cm²). Inasmuch as D-[3-O-methyl-¹⁴C]glucose enters the cells via facilitated diffusion but is not metabolized, an equilibrium between the extracellular and the cytosolic concentration is reached within 10 min. Cells were then washed three times with PBS containing the glucose transport blocker phloretin (0.1 mM). After 30 min of incubation with 2 ml of SDS (1%) at 37°C, detached cells were resuspended by pipetting. Aliquots (1 ml each) were used to determine radioactivity and protein content (17).

Adenosine measurement. For desalting and concentrating, effluent perfusate samples (6 ml) were passed over C₁₈ Sep-Pak chromatography precolumns (Waters, Eschborn, Germany). Purines were eluted with 66% (vol/vol) methanol-water solution, evaporated to dryness, and redissolved in 200 µl of distilled water (6). Adenosine was quantified by HPLC using a linear ammonium acetate (25 mM, pH 5)-methanol (66%) gradient over 12.5 min. Absorbance of column eluate was continuously measured at 254-nm wavelength. Adenosine was identified by comparison of retention times and quantified by software-guided peak area integration using known external standards.

Statistics. Values are means ± SD. Normal distribution of data was tested after Shapiro-Wilks in the modification of Lillefors. Similarity of variances was tested after Levene. Differences of mean adenosine release rates during the additive infusion protocol of inhibitors and during a stepwise increase of the adenosine inflow concentration were assessed by Student's t-test for paired data, testing each experimental condition vs. the next preceding condition. The significance of differences between VSMC and PAEC was determined using one-way ANOVA. P < 0.05 was taken to indicate a significant difference. Statistical tests were performed using SPSS software for Windows (version 7.5).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

Cell cultures. In phase-contrast microscopy, VSMC exhibited an elongated shape, and, when reaching confluence, they revealed a "hill- and valley-like" growth pattern. After VSMC were stained with FITC-coupled anti--smooth muscle actin, the filamentous actin skeleton became apparent in fluorescent microscopy. PAEC revealed the contact-inhibited "cobblestone-like" growth pattern and, after ingestion of fluorescent-labeled acetylated low-density lipoprotein, a patchy cytosolic pattern in fluorescent microscopy. On the basis of double-staining techniques, only cell cultures that exhibited a purity of >95% were studied.

Relationship between basal adenosine release, cytosolic volume, and column volume. A quantitative comparison of adenosine release from VSMC and PAEC requires common reference points. Therefore, adenosine release rates were determined in relation to column volume and cell cytosolic volume (Table 1). Mean column volume was systematically larger (36%) in PAEC than in VSMC. However, when related to column volume, the mean protein content and mean cytosolic volume did not differ between VSMC and PAEC, respectively. Furthermore, the cell volume-protein relation was the same. Averaged basal adenosine release rates related to column volume amounted to 16.4 ± 8.2 and 17.2 ± 8.0 pmol · min¹ · ml column volume¹ in VSMC (n = 26) and PAEC (n = 26), respectively, and thus were similar. Relative to cytosolic volume, mean basal adenosine release rates were 1.2 pmol · min¹ · µl¹ in VSMC and PAEC (Table 1).

Adenosine production rates. The effects of inhibition of adenosine deaminase (EHNA) and adenosine kinase (ITU) on adenosine release are shown in Fig. 1 for VSMC (n = 8) and PAEC (n = 5). Infusion of EHNA alone did not significantly change adenosine release of VSMC or PAEC (data not shown). However, combined infusion of ITU and EHNA significantly enhanced adenosine release rates from 1.6 ± 0.8 pmol · min¹ · µl¹ under control conditions to 12.1 ± 2.7 pmol · min¹ · µl¹ in VSMC and from 1.2 ± 0.5 to 7.9 ± 1.4 pmol · min¹ · µl¹ in PAEC. Under this condition, the adenosine release rate represents the minimal total adenosine production rate.

View larger version (25K): [in this window] [in a new window]   Fig. 1.   Adenosine release of vascular smooth muscle (VSMC, top) and porcine aortic endothelial cells (PAEC, bottom) during inhibition of adenosine deaminase [erythro-9-(2-hydroxy-3-nonyl)-adenine (EHNA), 5 µM] plus adenosine kinase [iodotubericidine (ITU), 10 µM], additional inhibition of adenosine membrane transport [n-nitrobenzylthioinosine (NBTI), 1 µM], and inhibition of ecto-5'-nucleotidase [,-methylene-adenosine 5'-diphosphate (AOPCP), 50 µM]. Compounds were infused cumulatively. Values are means ± SD. Ctr, control conditions. *P < 0.05; **P < 0.01 vs. the preceding condition. §§P < 0.01, VSMC vs. PAEC.

1

1

1

1

1

1

1

1

Extracellular adenosine production and transmembrane adenosine concentration gradient. The effects of adenosine membrane transport inhibition under control conditions are shown in Fig. 2. Infusion of NBTI enhanced adenosine release in VSMC (n = 7) from 1.4 ± 0.4 to 2.8 ± 0.8 pmol · min¹ · µl¹ and in PAEC (n = 7) from 1.0 ± 0.2 to 2.4 ± 0.5 pmol · min¹ · µl¹. This result indicates a transmembrane concentration gradient directed from the outside to the inside of the cell under control conditions. Additional infusion of AOPCP in the presence of NBTI reduced adenosine release from 2.8 ± 0.8 to 1.3 ± 0.4 pmol · min¹ · µl¹ in VSMC and from 2.4 ± 0.5 to 0.6 ± 0.2 pmol · min¹ · µl¹ in PAEC. This provides evidence for extracellular adenosine production via ecto-5'-nucleotidase.

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Adenosine release of VSMC (top) and PAEC (bottom) during inhibition of adenosine membrane transport (NBTI) and additional inhibition of ecto-5'-nucleotidase (AOPCP). See Fig. 1 legend for more information. **P < 0.01 vs. the preceding condition. §§P < 0.01, VSMC vs. PAEC.

Cell surface adenosine concentration. At an inflow adenosine concentration of zero, the control adenosine outflow concentration amounted to 12 ± 4 nM in VSMC and 17 ± 7 nM in PAEC (Table 2). These concentrations were not significantly different from each other. During a stepwise increase of the inflow adenosine concentration, the outflow concentration tended to increase in VSMC and PAEC, respectively. However, while the small increase was determined to be significant in VSMC, no significant effect was noted in PAEC. Most importantly, however, for the range of enhanced inflow concentrations studied, the outflow concentration in VSMC and PAEC remained below the column inflow concentration. The respective differences of inflow minus outflow concentration are plotted in Fig. 3 as a function of the adenosine inflow concentration. As can be deduced from these curves, the concentration gradient along the column axis was rising at low inflow concentrations and falling at higher inflow concentrations. For VSMC and PAEC, the curves intersected the abscissa at ~18 nM. For this inflow concentration, no axial adenosine concentration gradient existed. That is, extracellular and cytosolic adenosine concentrations were equal, and the net adenosine flow across the cell membrane was zero.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Inflow minus outflow concentration difference of adenosine vs. inflow adenosine concentration. , PAEC; , VSMC. Dashed lines connect measured values; dotted and solid lines represent linear regression curves. Insets: correlation coefficients and regression slopes. Each data point represents a mean value of 6 experiments.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Inflow minus outflow concentration difference of adenosine vs. inflow adenosine concentration during partial inhibition of adenosine metabolism (adenosine kinase, adenosine deaminase). See Fig. 3 legend for more information. Each data point represents a mean value of 3 experiments.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

The present study shows that VSMC from the pig coronary artery have a surprisingly active adenosine metabolism. They have a 63% higher total production rate (Fig. 1) and consistently release adenosine at a higher rate under similar experimental conditions than macrovascular endothelial cells from the pig aorta (PAEC). The greater total production rate of adenosine by VSMC is most likely due to higher intracellular production. However, it is noteworthy that also extracellular adenosine production of VSMC, most likely via ecto-5'-nucleotidase, exceeds that of PAEC (Fig. 2). Both cell species exhibit membrane net inward transport of adenosine under control conditions, indicating that this feature recently reported for the total heart (7) has its equivalent on the cellular level and may reflect a rather general feature of the adenosine transmembrane concentration gradient under physiological conditions. However, under conditions of reduced salvage of adenosine, the adenosine cell surface concentration is increased and the transmembrane concentration gradient may be reversed (Table 3, Fig. 4).

The capability of VSMC to metabolize adenosine via different pathways was previously described (8, 9). However, the flux rates of adenosine production and removal have not been analyzed. Therefore, a previously established set of inhibitor experiments (6, 7) was used to differentiate the individual metabolic pathways as well as membrane transport. The present experiments indicate that the total adenosine production rate of VSMC is in a range similar to that found for PAEC (this study; 6). While adenosine kinase seems of major importance for rapid metabolism of adenosine under control physiological conditions, adenosine deaminase seems of less importance. However, in a previous study (9), a considerable importance of the deaminase pathway was suggested on the basis of experiments on VSMC incubated in FCS and treated with 10-200 µM EHNA. Inasmuch as cardiac tissues were previously shown to have no residual adenosine deaminase activity at 5 µM EHNA and the plasma adenosine deaminase isoform ADA₂ is much less susceptible to inhibition by EHNA (26), it seems likely that the high adenosine deaminase activity of VSMC suggested previously (9) was mostly contributed by the presence of a plasma fraction. Inasmuch as cells were superfused with a buffer in the present study, ADA₂ activity does not interfere with the results obtained.

There exists no agreement on the physiological surface concentration of adenosine on endothelial or smooth muscle cells. While isolated (14) and in vivo (25) heart studies using collection of transudate to assess the interstitial adenosine concentration suggest values in the micromolar range (0.1-1.0 µM), mathematical model analysis (7) that integrates comprehensive measurements of venous adenosine release and the free cytosolic adenosine concentration (SAH technique) suggests considerably lower endothelial cell surface concentrations (7.3 nM luminal, 34 nM abluminal). In the present study we made a first attempt to determine the surface adenosine concentration of isolated VSMC and PAEC directly. The approach uses the feature that the concentration of an uncharged substrate in the perfusate region will achieve an equilibrium with the cell surface concentration if the transit time through the perfusion column is sufficiently long and the substrate is not metabolized in the perfusate. This equilibrium will be achieved more easily if the concentration at the system inflow is raised to reduce the inflow-outflow concentration difference. Following this approach, we found that increasing the inflow adenosine concentration resulted in much smaller or even missing increases of the outflow concentration (Table 2). Thus, while the adenosine concentration was rising along the column length at an inflow concentration of zero, it was falling from inflow to outflow at enhanced inflow adenosine concentrations (Fig. 3). The reversal of this axial concentration gradient occurred at inflow concentrations of ~18 nM in VSMC and PAEC. At this concentration, there was obviously no net release or uptake of adenosine by the cells. Thus the perfusate concentration must not only have equaled the surface concentration but also the average cytosolic adenosine concentration. Inasmuch as the column outflow adenosine concentration under this condition of diffusion equilibrium was still in the range of that determined under control conditions (zero inflow concentration), it can be concluded that the outflow concentration at zero inflow reflects the normal cell surface adenosine concentration reasonably well. It seems noteworthy that the estimates for PAEC are close to those calculated for the venous (10.1 nM) and the endothelial cell region (8.0 nM) of the isolated perfused guinea pig heart by mathematical model analysis based on experimental results (7). Further experiments using coronary endothelial cells from the guinea pig heart indicate that the remaining nearly twofold concentration difference between the present experimental estimate and the previous model estimate can be attributed to species differences (pig vs. guinea pig) (18).

Under the present experimental conditions, VSMC and PAEC take up into the cell about one-half of the extracellularly produced adenosine. This is indicated by the effects of NBTI (Fig. 2), which inhibits adenosine-facilitated membrane diffusion (5). The motor of adenosine uptake is largely provided by rapid intracellular phosphorylation of adenosine to AMP via adenosine kinase (23). To maintain the physiological transmembrane concentration gradient, a continuous extracellular production of adenosine must be provided. The source of this extracellular adenosine production is most likely extracellular 5'-AMP dephosphorylated via ecto-5'-nucleotidase (Fig. 2). While native adenine nucleotide release was previously shown for PAEC (6), it has not been determined in VSMC. The present experiments, however, suggest that the adenine nucleotide release of VSMC should at least amount to ~3 pmol · min¹ · µl¹ to account for the adenosine release in the presence of NBTI (Fig. 2). As to the mechanism of cellular adenine nucleotide release, it remains to be clarified whether cAMP release (15) can quantitatively explain extracellular adenosine production and/or the possible importance of the multidrug resistance (MDR1) gene product or the cystic fibrosis transmembrane conductance regulator for adenine nucleotide release of smooth muscle and endothelial cells (1, 10).

Block of adenosine membrane transport doubled the extracellular adenosine concentration (Fig. 2). This effect may be pertinent to understanding the mechanism by which clinically used adenosine membrane transport blockers, e.g., dipyridamole, cause coronary vasodilation. Adenosine produced on the surface of VSMC is obviously prevented by transport blockers from reaching its physiological site of metabolism. The enhanced surface concentration may then stimulate smooth muscle surface receptors and evoke relaxation in an autocrine mode.

Study limitations. To assess the adenosine production rates of VSMC and PAEC, this study used competitive inhibitors of adenosine metabolism and membrane transport. Therefore, to interpret the obtained results, the specificity and potency of the inhibitors are of importance. As outlined recently (7), EHNA, ITU, and NBTI cause a highly specific and nearly complete inhibition of the respective enzymes and transporter. With an in vitro inhibition constant in the nanomolar range (29), the efficacy of 50 µM AOPCP as used in the present experiments to inhibit ecto-5'-nucleotidase may be assumed to be complete. However, previous blood plasma and isolated heart studies (3, 19) have shown that this concentration yields only a 60-80% inhibition of ecto-5'-nucleotidase. Thus the persisting adenosine release of VSMC and PAEC during combined block of membrane transport and ecto-5'-nucleotidase is in line with results from previous studies and may reflect a residual enzyme activity. Even though the other blockers were highly effective under control conditions, it still needs to be considered that the efficacy of the blockers will decrease if the concentration of the natural substrate (adenosine) increases. Although the intracellular adenosine concentration was not directly assessed in this study, it is likely that the intracellular concentration increased, particularly under the combined treatment with EHNA and ITU. Such a rise of the substrate concentration would increase the residual fluxes through the adenosine deaminase and adenosine kinase pathway and thus lead to an underestimation of the true adenosine production rate (7).

Future directions. We have shown that VSMC represent a region of active adenosine metabolism and exhibit a remarkable similarity to PAEC with respect to 1) the pattern of production and degradation of adenosine, 2) the transmembrane adenosine concentration gradient, and 3) an adenosine surface concentration of ~18 nM. While the endothelium has previously been suggested to contribute significantly to total cardiac adenosine production (23), the possible role of smooth muscle cells has not been quantified. While it seems appropriate to include only endothelial and parenchymal cell regions in the analysis of tissue adenosine metabolism for the capillary region (14, 25), such concepts need a further extension toward inclusion of the smooth muscle region for noncapillary vessel segments. The integration of vascular smooth muscle adenosine metabolism in the control of the interstitial concentration will be of particular importance for future studies that aim to quantify the relationship between the adenosine concentration and coronary flow (16) or smooth muscle cell migration and proliferation (10, 8).