INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

NITRIC OXIDE (NO) contributes importantly to the regulation of arteriolar tone and blood flow in the intestine (8, 9, 24). Under normal resting conditions, there is continuous NO release from the arteriolar endothelium that is due in large part to the shear stress associated with blood flow (9). However, we have obtained considerable evidence that during periods of increased sympathetic nerve activity, the constriction of intestinal arterioles is limited by sustained endothelial NO release despite a precipitous fall in shear stress (25, 26, 33, 34). Moreover, this modulating influence of NO on sympathetic constriction can be prevented by minimizing the fall in microvascular and/or tissue PO₂ that normally accompanies the decrease in network blood flow (33, 34), suggesting that a reduction in local O₂ levels becomes the main stimulus for endothelial NO release under these low shear conditions. Direct measurement with NO-sensitive microelectrodes has verified that a reduction in blood O₂ delivery is a potent stimulus for arterial NO release in the intestine (9), and reduced O₂ levels lead to increased endothelial NO production in other vascular beds as well (1, 11, 28, 29).

We have recently determined that during sympathetic constriction in the intestine, the fall in microvascular and/or tissue PO₂ is linked to endothelial NO release through the local formation of adenosine, which then binds to endothelial A₁ receptors that are coupled to the NO synthesis pathway. Our findings also suggest that this adenosine is most likely formed within 50 µm of the arteriolar wall (34). In addition to the possible release of adenosine from periarteriolar parenchymal cells or from cells within the arteriolar wall itself, recent studies raise the possibility that adenosine could also be formed extracellularly via the breakdown of ATP that is released from circulating erythrocytes when luminal O₂ levels fall in either the arterioles or nearby venules (13, 15, 16). Consistent with this possibility, microvascular endothelial cells have a high activity of extracellular membrane ectonucleotidase, which readily degrades adenine nucleotides to adenosine (35). The purpose of the current study was to determine whether this latter mechanism contributes to local adenosine formation, and therefore sustained endothelial NO release, during sympathetic vasoconstriction in the intestine.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All surgical and experimental procedures were approved by the West Virginia University Animal Care and Use Committee. Male Sprague-Dawley rats aged 8-9 wk (Harlan Sprague Dawley; Indianapolis, IN) were anesthetized with thiopental sodium (100 mg/kg ip) and placed on a heating mat to maintain a 37°C rectal temperature. To ensure adequate gas exchange, rats were intubated and ventilated with a rodent ventilator (Harvard Apparatus; South Natick, MA). Arterial pressure was measured from the cannulated right carotid artery with a Gould P23 ID pressure transducer (Cleveland, OH). The small intestine was prepared for microscopic observation as previously described (25) and continuously superfused with a physiological electrolyte solution (in mM: 119 NaCl, 25 NaHCO₃, 6 KCl, and 3.6 CaCl₂) that was warmed to 37°C and equilibrated with either 5% O₂-5% CO₂-90% N₂ to mimic normal in vivo conditions (3) or 20% O₂-5% CO₂-75% N₂ to create an O₂-enriched environment. Isoproterenol (10 mg/l, Sigma; St. Louis, MO) and phenytoin (20 mg/l, Parke-Davis; Morris Plains, NJ) were added to the superfusate to suppress intestinal motility. At these concentrations, neither agent alters resting arteriolar tone in this vascular bed (10). With most of the preparation covered by polyvinyl film and superfusate flow directed beneath the film, solution PO₂ immediately above the tissue is maintained at 40-50 mmHg under normal conditions (6).

After surgery, the rat was transferred to the stage of an Olympus BHTU intravital microscope (Hyde Park, NY) fitted with a CCD video camera (Dage MTI; Michigan City, IN). Video images were displayed on a Panasonic high-resolution video monitor and stored on videotape for offline analysis. Arteriolar inner diameters were measured with a video caliper (Microcirculation Research Institute, Texas A&M University) during videotape replay.

A bipolar platinum electrode was used to stimulate the sympathetic postganglionic efferents running along a mesenteric artery-vein pair upstream from the arteriole under study. The electrode and artery-vein pair were briefly raised above the superfusate, and the nerves were stimulated with square-wave pulses at supramaximal voltage (5-6 V) and a pulse duration of 10 ms. These stimulation parameters elicit frequency-dependent arteriolar constrictions that are abolished by the selective ₁-receptor antagonist prazosin (27), verifying that these responses are due to sympathetic nerve activation.

Experimental protocols. The overall aim of this study was to evaluate the hypothesis that during sympathetic arteriolar constriction in the intestine, the adenosine that links reduced local O₂ levels to endothelial NO release is formed by the sequential hydrolysis of extracellular ATP that has been released from erythrocytes. The framing of this hypothesis gives rise to a number of testable predictions. First, increased extracellular ATP levels in the immediate vicinity of the arteriolar wall must reduce arteriolar tone by the same mechanism through which endogenous adenosine acts during sympathetic constriction, i.e., the activation of A₁ adenosine receptors and stimulation of arteriolar NO release (34). Second, this vasoactive effect of ATP must depend on the activity of 5'-ectonucleotidase, the extracellular membrane-bound enzyme that converts AMP to adenosine (the final step of the postulated adenosine formation pathway). Third, inhibition of 5'-ectonucleotidase activity should abolish the NO-mediated modulation of arteriolar constriction during sympathetic nerve stimulation. Fourth, inhibition of 5'-ectonucleotidase activity should disrupt the relationship between reduced luminal O₂ levels and NO activity that we have previously documented during sympathetic nerve stimulation (33). These predictions were tested in the experiments described below.

5

3

3

4

4

5

4

4

5

4

3

Data and statistical analysis. For each arteriole, the level of resting tone (T) was calculated as follows: T = [(D_max  D_c)/D_max] × 100, where D_max is passive diameter under adenosine and D_c is the diameter measured during the control period.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In all protocols combined, 58 first-order arterioles were observed in 36 rats. The rats averaged 264 ± 4 g body wt at 56 ± 1 days of age, with a mean arterial pressure under anesthesia of 100 ± 1 mmHg. Arteriolar resting and passive diameters averaged 54 ± 1 and 90 ± 1 µm, respectively, with an average calculated arteriolar tone of 39 ± 1%.

Eight first-order arterioles were studied to evaluate the effectiveness of AOPCP as a 5'-ectonucleotidase inhibitor in the rat intestine. AOPCP did not have a significant effect on resting arteriolar diameter (55 ± 2 µm under control conditions vs. 57 ± 2 µm during AOPCP exposure), but it completely abolished arteriolar responses to directly applied AMP. AMP applied at ejection pressures of 10, 20, and 30 psi caused dilations of 6 ± 1, 13 ± 1, and 18 ± 1 µm under control conditions vs. 0 ± 0, 0 ± 0, and 1 ± 1 µm in the presence of AOPCP (Fig. 1).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Arteriolar responses to AMP applied via pressurized micropipette directly to the arteriolar wall under the normal superfusate before and then during exposure to 4.5 × 105 M ,-methyleneadenosine 5'-diphosphate (AOPCP). All data are paired. n = 10 vessels. *P < 0.05 vs. control.

To assess the ability of ATP to stimulate arteriolar NO synthesis in this vascular bed, we evaluated the effect of L-NMMA on arteriolar responses to directly applied ATP (n = 8 vessels). L-NMMA had no significant effect on resting arteriolar diameters (47 ± 3 µm under control conditions vs. 48 ± 2 µm during L-NMMA exposure), but it completely abolished arteriolar responses to ATP. ATP applied to the arteriolar wall at 10, 20, and 30 psi induced respective dilations of 7 ± 1, 12 ± 1, and 16 ± 2 µm under control conditions versus 0 ± 0, 1 ± 1, and 1 ± 1 µm in the presence of L-NMMA (Fig. 2A). Eight additional vessels were studied to determine the importance of adenosine A₁ receptors in the effect of ATP on arteriolar tone. DPCPX had no significant effect on resting arteriolar diameter (49 ± 3 µm under control conditions and during DPCPX exposure) but significantly reduced arteriolar responses to ATP. In these experiments, ATP application at 10, 20, and 30 psi induced respective dilations of 8 ± 1, 14 ± 1, and 16 ± 2 µm under control conditions versus 1 ± 1, 2 ± 1, and 4 ± 1 µm in the presence of DPCPX, a reduction of 75-88% (Fig. 2B).

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Arteriolar responses to ATP applied via pressurized micropipette before and then during exposure to 104 M NG-monomethyl-L-arginine (L-NMMA; A) or 4 × 104 M 8-cyclopentyl-1,3-dipropylxanthine (DPCPX; B). Data within each group of experiments (n = 8 vessels) are paired. *P < 0.05 vs. control.

Eight arterioles were studied to determine whether the activity of 5'-ectonucleotidase is necessary for arteriolar responsiveness to ATP (Fig. 3). In these experiments, neither AOPCP nor L-NMMA had an effect on resting arteriolar diameters (50 ± 2 µm under control conditions vs. 51 ± 2 µm with AOPCP and 52 ± 1 with AOPCP + L-NMMA). ATP application at 10, 20, and 30 psi induced respective dilations of 10 ± 1, 14 ± 1, and 17 ± 1 µm. In the presence of AOPCP, these responses were reduced to 2 ± 1, 3 ± 1, and 5 ± 1 µm, respectively (reductions of 70-80%). With the subsequent addition of L-NMMA, these residual responses were virtually abolished (dilations of 0 ± 0, 0 ± 1, and 1 ± 1 µm, respectively).

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Arteriolar responses to ATP applied via pressurized micropipette under the normal superfusate before and then during exposure to 4.5 × 105 M AOPCP and then during exposure to 4.5 × 105 M AOPCP + 104 M L-NMMA. All data are paired. n = 8 vessels. *P < 0.05 vs. control; P < 0.05 vs. AOPCP.

Seventeen arterioles were studied to explore the relationship between 5'-ectonucleotidase activity and NO activity during periods of increased sympathetic nerve activity. Under the normal superfusate, resting arteriolar diameter averaged 62 ± 1 µm, and sympathetic nerve stimulation at 3 and 8 Hz induced frequency-dependent constrictions of 11 ± 1 and 19 ± 1 µm, respectively (Fig. 4). With the addition of L-NMMA, resting arteriolar diameters were unchanged (60 ± 1 µm), but responses to each level of sympathetic nerve stimulation were significantly enhanced (constrictions of 19 ± 1 and 29 ± 2 µm at 3 and 8 Hz, respectively). Under the high O₂ superfusate, resting arteriolar diameters were unchanged (61 ± 1 µm), but control responses to nerve stimulation were significantly greater than those under the normal superfusate (constrictions of 18 ± 1 µm at 3 Hz and 30 ± 2 µm at 8 Hz). L-NMMA did not alter resting arteriolar diameters (62 ± 1 µm) under the high O₂ superfusate. However, in contrast to the normal superfusate, L-NMMA did not enhance arteriolar responses to sympathetic nerve stimulation under the high O₂ superfusate (constrictions of 20 ± 1 µm at 3 Hz and 30 ± 2 µm at 8 Hz).

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Effect of L-NMMA on arteriolar responses to 3- and 8-Hz sympathetic nerve stimulation under normal superfusate (left) and high O2 superfusate (right). All data are paired. n = 8 vessels. *P < 0.05 vs. control response to 3-Hz stimulation under normal superfusate; P < 0.05 vs. control response to 8-Hz stimulation under normal superfusate.

Figure 5 illustrates the effect of L-NMMA on arteriolar responses to sympathetic nerve stimulation under the normal and high O₂ superfusates in the presence of AOPCP. As in the other experiments, neither AOPCP nor L-NMMA altered resting arteriolar diameters in these experiments (60 ± 1 µm under control conditions vs. 60 ± 1 µm in the presence of AOPCP and 58 ± 1 µm in the presence of L-NMMA). Similar to the above findings, in the presence of AOPCP, arteriolar responses to sympathetic nerve stimulation were enhanced by L-NMMA under the normal superfusate but not under the high O₂ superfusate. Under the normal superfusate, 3- and 8-Hz stimulation induced arteriolar constrictions of 10 ± 1 and 17 ± 1 µm under control conditions versus 17 ± 1 and 30 ± 1 µm in the presence of L-NMMA. Under the high O₂ superfusate, 3- and 8-Hz stimulation induced arteriolar constrictions of 19 ± 1 and 30 ± 1 µm under control conditions versus 16 ± 1 and 30 ± 1 µm in the presence of L-NMMA.

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Effect of L-NMMA on arteriolar responses to 3- and 8-Hz sympathetic nerve stimulation during continuous exposure to 4.5 × 105 M AOPCP under normal superfusate (left) and high O2 superfusate (right). All data are paired. n = 9 vessels.

To explore the possibility that erythrocytes in nearby first-order venules may be a significant source of ATP and/or adenosine during increased sympathetic nerve activity, we determined whether proximity to a venule was a critical factor in the limitation of arteriolar constriction by reduced PO₂ or NO. Figure 6, which is derived from the data shown in Figs. 4 and 5, illustrates the lack of a relationship between arteriolar-venular separation distance and the effect of either L-NMMA or increased O₂ availability on arteriolar responses to sympathetic nerve stimulation. For either treatment, there was no significant correlation between the magnitude of the response enhancement and arteriole-venule separation distance (see Fig. 6 for line equations, correlation coefficients, and P values).

View larger version (21K): [in this window] [in a new window]   Fig. 6.   First-order linear regression plots of the increase in arteriolar sympathetic constriction vs. arteriole-venule separation distance at nerve stimulation frequencies of 3 Hz (A and C) or 8 Hz (B and D) during exposure to L-NMMA (A and B) or to the high O2 superfusate (C and D). A: y = 0.11x + 79.3, r = 0.03, P = 0.91. B: y = 0.87x + 42.9, r = 0.42, P = 0.09. C: y = 0.83x + 65.8, r = 0.23, P = 0.38. D: y = 0.83x + 50.7. All data are paired. n = 17 vessels.

Seven first-order arterioles were studied to evaluate the effect of AOPCP on vascular smooth muscle responsiveness to NO (Fig. 7). As in the other experiments, AOPCP did not alter resting arteriolar diameter (57 ± 2 µm under control conditions vs. 58 ± 2 µm in the presence of AOPCP). Under control conditions, SNP induced dilations of 11 ± 2, 23 ± 2, and 35 ± 3 µm at ejection currents of 5, 20, and 40 nA, respectively. Exposure to AOPCP did not significantly alter these responses (dilations of 12 ± 2, 22 ± 3, and 32 ± 3 µm at ejection currents of 5, 20, and 40 nA, respectively).

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Arteriolar responses to iontophoretically applied sodium nitroprusside (SNP) under normal superfusate before and during exposure to AOPCP. All data are paired. n = 7 vessels.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our recent investigations of this vascular bed have led us to conclude that during periods of increased sympathetic outflow to the arterioles, a flow-dependent fall in local PO₂ leads to local formation of adenosine, which in turn maintains arteriolar NO synthesis in the face of reduced shear stress via activation of endothelial A₁ receptors (34). In an attempt to determine the site of adenosine formation under these conditions, we sought in this study to test the hypothesis that ATP, released from erythrocytes in response to reduced O₂ levels, is sequentially hydrolyzed to adenosine via extracellular nucleotidase activity.

Recently, Ellsworth (15) reported that a reduction in PO₂ from 85 to 35 mmHg increases ATP release from rat erythrocytes by more than 75%. We have found that 3- to 8-Hz sympathetic nerve stimulation reduces blood flow in the intestine by 33-66% (23), which causes first-order arteriolar wall PO₂ to fall from 67 ± 3 mmHg to as low as 41 ± 6 mmHg (33). Because arteriolar wall PO₂ is almost identical to luminal PO₂ (14), it seems possible that under our experimental conditions, there could have been an increase in erythrocyte ATP release similar to that reported by Ellsworth (15). Erythrocytes in nearby venules could also be source of ATP under these conditions. We have not measured first-order venular PO₂ in the rat intestine during sympathetic nerve stimulation, but measurements made by others in this preparation under similar experimental conditions (7, 36) suggest that the resting PO₂ of these venules is probably around 50 mmHg, or roughly 25% below that of the paired first-order arterioles. Therefore, during sympathetic nerve stimulation, we would expect venular lumen PO₂ to fall as low as if not lower than the minimum PO₂ we measured in the nearby arterioles. Chemical communication can occur between paired arterioles and venules (17, 21, 37), and adenosine in particular can readily diffuse from venules to arterioles in vasoactive amounts, even when the vessels are separated by >150 µm (21). In the current experiments, the adenosine responsible for stimulating endothelial NO synthesis is most likely formed within 50 µm of the arteriolar wall (34). Because most of the arterioles we studied were within 50 µm of a paired venule (mean separation distance = 29 ± 5 µm), it is possible that at least some of the adenosine responsible for stimulating arteriolar NO synthesis during sympathetic constriction could have either diffused from the venular blood or have been formed from ATP at the venular wall. However, we found no correlation between arteriole-venule separation distance and the effect of either L-NMMA or the high O₂ superfusate on arteriolar sympathetic constriction (Fig. 6). Therefore, it appears unlikely that a significant fraction of the vasoactive pool of adenosine (or its hypothesized precursor, ATP) originated within the venules.

Consistent with the possibility that ATP from arteriolar erythrocytes serves as an extracellular source of adenosine in this preparation, we found that ATP applied to the arteriolar wall causes a NO-dependent dilation (Fig. 2A) that is largely due to A₁ receptor activation (Fig. 2B) and that largely depends on 5'-ectonucleotidase activity (Fig. 3). Taken together, these findings suggest that, in the intestine, extracellular ATP in the immediate vicinity of the arteriolar wall can stimulate arteriolar NO release through the ultimate conversion of AMP to adenosine. The smooth muscle of intestinal and mesenteric arterioles contains P_2x purinergic receptors that should elicit vasoconstriction when stimulated by ATP (19, 31). However, the amounts of ATP that we applied to the arteriolar wall were relatively low (the highest amount causing only about 35% of maximal dilation) and did not cause even a transient constriction. Because most of that ATP was quickly hydrolyzed to adenosine (Fig. 3), the amount of unhydrolyzed ATP was presumably insufficient to activate the P_2x receptors. Consistent with this interpretation, we found in preliminary experiments that application of ATP in higher quantities did cause arteriolar constriction (unpublished results).

If endogenous ATP were sequentially hydrolyzed to adenosine as local O₂ levels fell during sympathetic constriction, then inhibition of 5'-ectonucleotidase should have abolished the limiting influence of NO on arteriolar constriction because, as mentioned above, adenosine serves to link the fall in O₂ to NO production (34). A comparison of Figs. 4 and 5 reveals that AOPCP had no effect on the ability of L-NMMA to enhance arteriolar responses to sympathetic nerve stimulation or on the abrogation of the effect of L-NMMA under the high O₂ superfusate. Therefore, these results suggest that extracellular ATP, regardless of its origin, probably does not serve as a significant source of the adenosine that sustains arteriolar NO release under the current experimental conditions.

For the correct interpretation of our findings, it was critical that we verify the specificity of our antagonists and also demonstrate that vascular smooth muscle responsiveness to NO was not altered by any of our treatments. Our finding that AOPCP completely abolished arteriolar responses to AMP (Fig. 1) confirms the efficacy of AOPCP as a 5'-ectonucleotidase inhibitor in the intestinal microcirculation and is consistent with studies conducted in other vascular beds (22, 23). In addition, we have previously demonstrated that 4 × 10⁴ M DPCPX completely abolishes arteriolar responses to A₁ receptor agonists in this preparation (34). The preservation of normal arteriolar responses to SNP in the presence of AOPCP (Fig. 7) argues against the possibility that AOPCP alters the vascular smooth muscle responsiveness to NO under our experimental conditions. Finally, we have previously demonstrated that neither the high O₂ superfusate nor DPCPX has any effect on vascular smooth muscle responsiveness to NO (33, 34).

Another important observation in this study is the lack of an effect of L-NMMA on steady-state arteriolar tone, which contrasts with findings in skeletal muscle (1, 24) but is consistent with our previous findings in this vascular bed (26, 33, 34). From recent direct measurements verifying that NO is continuously released from these vessels in the resting state (9), local NO synthase inhibition would be expected to reduce arteriolar diameter. In fact, we routinely did see a transient reduction in arteriolar diameter within the first few minutes of L-NMMA exposure, but then these vessels regained their initial resting diameter. This inability of L-NMMA to produce a sustained reduction in arteriolar diameter may reflect the fact that resting arteriolar tone results from the integration of numerous simultaneous vasoactive signals by the vascular smooth muscle. Local vascular regulatory mechanisms may be so highly developed in the intestine that the removal of any single influence (such as NO in the presence of L-NMMA) is accompanied by a compensatory change in the activity of any of these systems, thereby reestablishing the initial level of resting tone.

We also found that resting arteriolar diameters were not altered by exposure to the high O₂ superfusate, which is consistent with our earlier findings in this preparation (33, 34). This may be due to the fact that the high O₂ superfusate, which minimizes the fall in both arteriolar wall and local tissue PO₂ during sympathetic constriction, does not have any significant effect on these PO₂ values under resting conditions (33, 34). This is because O₂ levels in the arteriolar wall and immediately adjacent parenchymal tissue are overwhelmingly influenced by luminal blood O₂ delivery during normal flow conditions, with extraluminal O₂ delivery from the superfusate only becoming important when luminal blood flow falls during sympathetic constriction (4). Because resting wall and periarteriolar PO₂ values were not changed under the high O₂ superfusate, it is understandable that no arteriolar constriction occurred under these conditions.

Because molecular O₂ is a cosubstrate for all NO synthase isoforms (18, 38), the reduction in arteriolar wall PO₂ that accompanies sympathetic constriction (33) might have been expected to limit arteriolar NO production despite increased local adenosine levels. However, our previous measurements indicate that during sympathetic stimulation at the frequencies used here, arteriolar wall PO₂ never approaches the level reflecting the Michaelis constant (K_m) of O₂ for either isolated endothelial NO synthase (eNOS) (6 mmHg) (32) or eNOS in intact endothelial cells (38 mmHg) (38). Therefore, the reduction in arteriolar wall PO₂ during sympathetic constriction was probably not sufficient to limit endothelial NO synthesis in the current study. Our observations suggest that as arteriolar wall PO₂ falls, NO synthesis may in fact change in a biphasic manner. A fall in PO₂ from resting levels could initially stimulate NO synthesis through adenosine formation, whereas a continued fall to lower levels would begin to suppress NO synthesis as O₂ becomes limited as a NO synthase cosubstrate.

Although the current results do not suggest a contribution of extracellular ATP hydrolysis to adenosine formation during sympathetic nerve stimulation in the intestine, they do provide important, albeit indirect, information about the nature of adenosine release under our experimental conditions. From our previous investigations, we concluded that the adenosine involved in stimulating NO release during sympathetic stimulation must come from a site within 50 µm of the vessel wall (34). As reported here, ATP released from arteriolar and/or venular erythrocytes does not appear to contribute importantly to the formation of adenosine under the current experimental conditions, nor is there any evidence of the venular blood or venular wall cells serving as direct sources of adenosine. This leaves the arteriolar wall and/or the periarteriolar parenchymal cells as the most likely sources of adenosine under these conditions (11, 12).