INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ADENOSINE IS AN IMPORTANT modulator of renal vascular resistance. In the normal kidney, adenosine elicits both vasoconstriction and vasodilation (reviewed in Refs. 16 and 28). Physiological or pathophysiological elevations of adenosine are generally associated with renal afferent arteriolar vasoconstriction and a reduction in glomerular filtration rate (GFR). This vasoconstrictor response is mediated by activation of the adenosine A₁ receptor subtype (24) and is purported to contribute to a metabolic regulation of GFR (18, 25). Abnormal adenosine-induced vasoconstriction has also been demonstrated to contribute to renal pathophysiology in models of acute renal insufficiency (1, 7). Less is known of the role of adenosine-induced renal vasodilation. Experimentally, vasodilation is usually observed only at very high adenosine concentrations and is believed to be mediated by the activation of a low-affinity adenosine A₂ receptor subtype coupled to adenylate cyclase (16, 28). However, studies in conscious animal models and anecdotal reports in humans subjects suggest important additional renal vasodilatory actions of adenosine (1, 26).

In the present study, we examined the renal microvascular effects of adenosine in the in vitro perfused hydronephrotic rat kidney (15). Using this model, we could demonstrate both of the renal vascular responses previously described (adenosine A₁ receptor-mediated vasoconstriction and low-affinity adenosine A₂ receptor-induced vasodilation). However, in addition to these two actions of adenosine, we observed a third type of adenosine response. Submicromolar concentrations (10-300 nM) of adenosine were capable of reversing or inhibiting pressure-induced afferent arteriolar vasoconstriction. This novel vasodilatory response occurred at concentrations well below that required to elicit adenosine A₁-induced vasoconstriction, and was blocked by both 4-(2-[7-amino-2-(2-furyl)[1,2,4]triazolo[2,3-a][1,3,5]triazin-5-yl-amino]ethyl)phenol (ZM-241385) and glibenclamide. Thus our findings indicate that the renal microvasculature is capable of expressing high-affinity adenosine A_2a receptors and that the afferent arteriolar vasodilatory response mediated by the activation of this receptor type is linked to the activation of ATP-sensitive K channels.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Isolated perfused hydronephrotic kidney. The isolated hydronephrotic rat kidney was used to examine the effects of adenosine on the renal afferent arteriole. Unilateral hydronephrosis was produced by ligating the left ureter under methoxyflurane-induced anesthesia. Hydronephrotic kidneys were harvested after 6 wk. Rats were anesthetized with methoxyflurane, the renal artery of the hydronephrotic kidney was cannulated, and the kidney was excised for in vitro perfusion. During the initial cannulation and throughout the isolation procedure, kidneys were continuously perfused with medium to avoid a disruption of nutritive flow or exposure to hypoxia or ischemia.

The perfusing apparatus used in the present study employed a single-pass presentation of medium to the kidney (13, 15, 23). The medium was pumped on demand through a heat exchanger to a pressurized reservoir, supplying the renal artery. Perfusion pressure was monitored within the renal artery and altered by adjusting the pressure within the reservoir. Kidneys were perfused with Dulbecco's modified Eagle's medium (GIBCO; Life Technologies, Gaithersburg, MD) containing 30 mM bicarbonate, 5 mM glucose, and 5 mM HEPES (GIBCO) and passed through 0.2-µm filters before use. The perfusate was equilibrated with 95% air-5% CO₂ (PO₂ = 150 Torr). Temperature and pH were maintained at 37°C and 7.40, respectively.

Kidneys were allowed to equilibrate for at least 1 h before initiating study protocols. To assess the effects of DPCPX and diltiazem on the contractile responses to adenosine, initial responses to 1.0 µM adenosine were assessed; the kidneys were then washed 10-20 min with control media or media containing either blocker, and the adenosine responses were reassessed. The effects of 4-[(3-butoxy-4-methoxyphenyl) methyl]-2-imidaxolidinone (Ro-20-1724) were assessed using a similar experimental design, in which initial responses were obtained, and then responses were reassessed after 10-min pretreatment with 0.5 µM Ro-20-1724. In studies examining the effects of adenosine and adenosine analogues on the response of the afferent arteriole to stepped increases in pressure, we first obtained two to three reproducible control pressure responses. Renal arterial pressure was increased from 60 to 180 mmHg in 20-mmHg steps and held at each pressure for 1 min. Pressure was then returned to 80 mmHg, kidneys were pretreated for 10 min with the adenosine analogue at each concentration, and reactivity was reassessed. We then sought to assess the impact of the K_ATP channel blocker glibenclamide (10 µM) on the responses to each adenosine analogue. In these studies, glibenclamide was added at the onset of the study and was present throughout the experiment. To cover the full concentration range for adenosine, 25 kidneys were used. Of these, 11 were exposed to only 4 concentrations, and 14 were exposed to each concentration. These data were combined. The single-pass adaptation of the in vitro perfused hydronephrotic kidney model is very stable, and reproducible constrictor responses to angiotensin II and pressure ramps are obtained for at least 6-8 h (e.g., see Refs. 15, 23). Moreover, K_ATP appears to be quiescent in the basal state of this preparation, as we have previously found glibenclamide to have no effect on the basal reactivity to either angiotensin II or elevated pressure (15, 22, 23).

Materials. Adenosine, glibenclamide, DCPCX, 2-phenylaminoadenosine (CV-1808), 5'-N-ethylcarboxamidoadenosine (NECA), and Ro-20-1724 were purchased from Research Biochemicals International (Natick, MA). ZM-241385 was purchased from Tocris Cookson (Ballwin, MO).

Analysis of data. Video images were transmitted to computer with a video acquisition board (model IVG-128; Datacube, Peabody, MA) for online analysis. Custom software was used to measure vessel diameter from the digitized video image as described in detail in previous publications (13). Vessel segments (20-30 µm in length) were scanned automatically at 1- to 2-s intervals. The mean-diameter measurements obtained during the plateau of the response were then averaged. Typically, one diameter determination was derived from the mean of 30-40 individual measurements, each in turn representing the mean diameter over the length of arteriolar segment.

Throughout the text, data are expressed as the mean followed by the standard error of the mean, as an index of dispersion. The number of replicates refers to the number of afferent arterioles examined. Differences between means were evaluated by analysis of variance followed by Student's t-test (paired or unpaired). Probabilities (P) less than 0.05 were considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Figures 1 and 2 demonstrate the classic renal adenosine A₁ and A₂ responses observed in the in vitro perfused hydronephrotic rat kidney preparation. As depicted in Fig. 1, the administration of 1.0 µM adenosine elicited a transient afferent arteriolar vasoconstriction that spontaneously abates after 5-10 min. This vasoconstriction was blocked by 0.1 µM 8-cyclopentyl-1,3-dipropylxanthine (DPCPX) (Fig. 1, A and C), an adenosine A₁ antagonist. The vasoconstriction was also blocked by diltiazem (Fig. 1D) and thus exhibits characteristics quite similar to adenosine A₁-mediated responses reported using other renal microvascular models (24).

View larger version (31K): [in this window] [in a new window]   Fig. 1.   Adenosine-induced afferent arteriolar vasoconstriction. A: tracing of transient afferent arteriolar vasoconstrictor response to 1.0 µM adenosine (ADO) and blockade of this response by the A1 adenosine antagonist 8-cyclopentyl-1,3-dipropylxanthine (DPCPX, 0.1 µM). B: mean time control data illustrating sequential responses to adenosine (n = 5). C: blockade of response by DPCPX (n = 5). D: blockade of the response by 10 µM diltiazem (n = 5). ADO-1 and ADO-2 refer to the first and second applications of adenosine (compare with A). CTL-1 and CTL-2 refer to pre-adenosine (control) diameters.

Figure 2 illustrates the vasodilation seen in response to stimulation of low-affinity adenosine A₂ receptors. In these experiments, afferent arteriolar tone was increased by elevating medium potassium concentration to 30 mM, and adenosine-induced vasodilation was observed over a range of 3-30 µM. In these studies, KCl decreased afferent arteriolar diameter from 12.3 ± 0.5 to 3.2 ± 0.2 µm (n = 5). The addition of 1 µM adenosine did not cause a further decrease in afferent arteriolar diameter in the presence of 30 mM KCl (3.4 ± 0.4 µm), indicating either that adenosine has no further effect on depolarized vessels or that the response to KCl was maximal. At 3, 10, and 30 µM adenosine, diameters increased to 4.5 ± 0.6, 6.4 ± 0.6 (P < 0.01), and 7.9 ± 0.9 µm (P < 0.01), respectively. It is important to emphasize that we have previously shown that K_ATP-dependent vasodilation is abolished by the presence of 30 mM KCl in this preparation (15, 22). The vasodilation elicited by 1-30 µM adenosine was potentiated by 0.5 µM Ro-20-1724, a selective phosphodiesterase inhibitor (11). In these experiments, KCl reduced diameters from 12.4 ± 0.8 to 3.8 ± 0.3 µm (n = 5). The addition of 0.3 µM Ro-20-1724 did not significantly alter diameter (3.9 ± 0.3 µm) alone, but potentiated the vasodilatory actions of adenosine. Thus, 1, 3, 10, and 30 µM adenosine increased diameters to 5.7 ± 0.5 (P < 0.005), 7.7 ± 0.7 (P < 0.005), 8.6 ± 0.8 (P < 0.001), and 10.1 ± 0.5 µm (P < 0.001), respectively. These data are plotted as the percent inhibition of KCl-induced vasoconstriction in Fig. 2. Ro-20-1724 potentiated the response at each concentration of adenosine. These findings are consistent with the purported role of cAMP in renal response to stimulation of low-affinity A_2b receptors (28).

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Adenosine-induced vasodilation of arterioles preconstricted with 30 mM KCl (to eliminate any contribution of altered K channel activity). Under control conditions (open circles), vasodilation was initiated only at adenosine concentrations above 1 µM. This vasodilation was potentiated by 0.5 µM Ro-20-1724 (solid circles). *P < 0.05 vs. control; n = 5.

Figure 3 illustrates the effects of a much lower concentration of adenosine (100 nM) on the vasoconstrictor response of the afferent arteriole to elevated renal perfusion pressure. The tracing in Fig. 3A illustrates the protocol. Pressure-dependent afferent arteriolar tone was established by elevating perfusion pressure from 80 to 180 mmHg. In this setting, the administration of 100 nM adenosine elicited a reversible vasodilation. This response was abolished by pretreatment with the adenosine A_2a receptor antagonist ZM-241385 (100 nM). The mean data obtained from five such experiments are summarized in Fig. 3B. Additional studies were conducted in which sequential responses to 100 nM adenosine were assessed as controls. Afferent arteriolar diameters for 80, 180, and 180 mmHg plus adenosine were 14.7 ± 1.5, 6.3 ± 1.0, and 13.4 ± 1.6 µm, respectively, for the first response and 14.2 ± 1.3, 6.6 ± 1.0, and 13.1 ± 1.5 µm for the second response (n = 3, P > 0.80).

View larger version (24K): [in this window] [in a new window]   Fig. 3.   A: tracings of the high-affinity vasodilatory response to adenosine observed during pressure-induced (myogenic) afferent arteriolar vasoconstriction. Adenosine (ADO) at 100 nM dilated vessels preconstricted by elevated pressure. This response was prevented by pretreatment with the adenosine A2a-selective antagonist 4-(2-[7-amino-2-(2-furyl)[1,2,4]triazolo[2,3-a][1,3,5]triazin-5-yl-amino]ethyl)phenol (ZM-241385). B: mean data (n = 5).

This high-affinity vasodilatory response was also potentiated by 0.5 µM Ro-20-1724. Thus elevating perfusion pressure from 80 to 180 mmHg decreased afferent arteriolar diameter from 14.0 ± 0.7 to 6.2 ± 0.6 µm (n = 5). At concentrations of 1, 3, 10, and 30 nM, adenosine increased diameters to 6.3 ± 0.7, 6.7 ± 0.7, 7.8 ± 0.7 (P < 0.01), and 8.5 ± 0.9 µm (P < 0.01), respectively. In the presence of Ro-20-1724, elevating perfusion pressure reduced afferent arteriolar diameters from 13.0 ± 0.6 to 6.5 ± 0.5 µm (n = 5) and 1, 3, 10, and 30 nM adenosine increased diameters to 7.7 ± 0.6 (P < 0.01), 9.2 ± 0.7 (P < 0.01), 10.2 ± 0.6 (P < 0.001), and 11.2 ± 0.7 µm (P < 0.001), respectively. These data are plotted as the percent inhibition in Fig. 4.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Potentiation of high-affinity adenosine-induced vasodilation (1-30 nM adenosine) by the phosphodiesterase inhibitor 4-[(3-butoxy-4-methoxyphenyl)methyl]-2-imidaxolidinone (Ro-20-1724, 0.5 µM; solid circles). *P < 0.05 vs. control (open circles); n = 5.

The effects of adenosine pretreatment on the response of the afferent arteriole to stepped increases in renal perfusion pressure are illustrated in Figs. 5-7. In these studies, the renal arterial pressure is reduced to 60 mmHg and then increased in 20-mmHg steps to a final pressure of 180 mmHg. The pressure at each step is held for 1 min, and constrictor responses are assessed. Stable control responses were obtained and these data are represented by the solid circles in Figs. 5 and 6. The kidneys were then pretreated with adenosine and responses were reassessed. Responses in the presence of increasing concentrations of adenosine are depicted by the open symbols in Figs. 5 and 6. Note that over a concentration range of 1-100 nM, adenosine exerted a concentration-dependent inhibition of myogenic reactivity (Fig. 5, left). For reasons that are not presently known, 100 nM adenosine appeared to be more effective in reversing preestablished pressure-induced vasoconstriction (Fig. 3) than in preventing the response. The myogenic response was further attenuated when adenosine concentrations were increased to 10 µM and 30 µM.

View larger version (29K): [in this window] [in a new window]   Fig. 5.   Effects of adenosine on myogenic reactivity of renal afferent arterioles. Adenosine was added cumulatively to concentrations depicted to right of each curve. To facilitate comparisons, the control response is depicted by solid circles in both the right and left. Note that adenosine inhibited myogenic reactivity at 1-100 nM. At higher concentrations, myogenic reactivity partially recovered (1.0 µM). Increasing adenosine levels further (3-30 µM) resulted in further inhibition.

The effects of K_ATP channel-blockade on this action of adenosine are illustrated in Fig. 6. The same protocol was repeated using a separate group of kidneys in the presence of 10 µM glibenclamide. In the presence of glibenclamide, 1-100 nM adenosine had no effect on the myogenic response (Fig. 6, left). However, the inhibitory effects observed at the higher adenosine concentrations (10-30 µM) were preserved. This effect of glibenclamide on the actions of adenosine are summarized in Fig. 7, in which concentration-response curves to adenosine were constructed in the absence and presence of glibenclamide. To facilitate this comparison, only data obtained at the highest pressure (180 mmHg) are plotted. In the absence of glibenclamide (solid circles in Fig. 7), adenosine elicited a biphasic response, characterized by vasodilatory components at 10⁸ to 10⁷ M and at 10⁵ to 10⁴ M adenosine. Note that in the kidneys pretreated with glibenclamide, only the low-affinity component (10⁵ to 10⁴ M adenosine) was seen.

View larger version (25K): [in this window] [in a new window]   Fig. 6.   Effects of adenosine on myogenic reactivity in presence of 10 µM glibenclamide. Note that the inhibitory effects of 1-100 nM adenosine normally seen in the absence of glibenclamide (Fig. 5) were abolished by this treatment. In contrast, the vasodilatory effects of adenosine at concentrations of 10-30 µM were preserved; , control.

View larger version (18K): [in this window] [in a new window]   Fig. 7.   Summary of effects of glibenclamide on adenosine-induced inhibition of myogenic vasoconstriction. Percent inhibition of vasoconstrictor response to elevating perfusion pressure to 180 mmHg is plotted. During control (), a biphasic vasodilatory response to adenosine was observed, whereas only the low-affinity response was observed in the presence of glibenclamide (). Data were taken from Figs. 5 and 6. Number of replicates is indicated in parentheses as "(control, glibenclamide)." *P < 0.05 vs. control.

Figures 8 and 9 summarize studies examining the effects of glibenclamide on the inhibition of myogenic reactivity elicited by the adenosine agonists CV-1808 and NECA. In each case, cumulative dose-response data were obtained, and separate kidneys were used for the glibenclamide studies. As shown in Fig. 8, CV-1808 significantly inhibited myogenic vasoconstriction with an IC₅₀ of 100 nM (n = 5). Over this concentration range, CV-1808 has been shown to activate adenosine A_2a receptors (9). These actions of CV-1808 were completely prevented by 10 µM glibenclamide (n = 6). Figure 9 illustrates that NECA inhibited myogenic responses over concentrations ranging from 3 to 30 nM (n = 5). The dose response to NECA was right-shifted following pretreatment with 10 µM glibenclamide (n = 6). As seen with adenosine, the actions of NECA appeared to be mediated by both glibenclamide-sensitive and glibenclamide-insensitive mechanisms. Figure 10 compares the potencies of adenosine, CV-1808, and NECA against myogenic afferent arteriolar vasoconstriction. The order of potency for these vasodilatory actions was NECA > CV-1808 = adenosine.

View larger version (20K): [in this window] [in a new window]   Fig. 8.   Inhibition of myogenic vasoconstriction by 2-phenylaminoadenosine (CV-1808). Glibenclamide completely prevented inhibitory effects of this adenosine agonist; n = 6 and n = 5 for control () and glibenclamide (), respectively. *P < 0.05 vs. control.

View larger version (15K): [in this window] [in a new window]   Fig. 9.   Inhibition of myogenic vasoconstriction by 5'-N-ethylcarboxamidoadenosine (NECA). Like adenosine, NECA exhibited glibenclamide-sensitive and glibenclamide-insensitive vasodilatory actions; n = 6 for both control () and glibenclamide (). *P < 0.05 vs. control.

View larger version (22K): [in this window] [in a new window]   Fig. 10.   Comparison of the potencies of NECA (solid circles), CV-1808 (open circles), and adenosine (gray-shaded circles) in inhibiting myogenic vasoconstriction in renal afferent arteriole. Data are taken from Figs. 6, 8, and 9.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this report, we partially characterize an unusual renal vasodilatory response to adenosine that we have observed in the in vitro perfused hydronephrotic rat kidney. In this preparation, adenosine exerts a vasodilatory action on afferent arteriolar myogenic vasoconstriction at submicromolar concentrations. This response was mimicked by NECA and CV-1808, indicating that it is mediated by an adenosine A₂ receptor subtype. The fact that this response was elicited at submicromolar levels of adenosine, the relatively high sensitivity to CV-1808 (EC₅₀ of approximately 100 nM, Fig. 8), and our observation that this response was blocked by ZM-241385 all suggest an involvement of the high-affinity adenosine A_2a receptor subtype (9).

This novel adenosine response was potentiated by Ro-20-1724 and was prevented by glibenclamide, implicating a role of cAMP and ATP-sensitive K channels in the receptor-effector coupling mechanism. Both A_2a and A_2b receptors are known to be coupled through G_s proteins to the activation of adenylate cyclase (9). Furthermore, cAMP-dependent protein kinase has been implicated in the activation of K_ATP in vascular smooth muscle (12, 21), and adenosine responses have been shown to be coupled to K channel activation in other vascular beds (7, 8, 10, 12).

The glibenclamide-sensitive afferent arteriolar vasodilation we observed in our model was blocked by 100 nM ZM-241385. At this concentration, ZM-241385 is highly selective against the adenosine A_2a receptor (19, 20). In the normal kidney, adenosine A₂ receptors are generally considered to be low-affinity A_2b type and are activated only at adenosine concentrations above that eliciting vasoconstriction (2, 28). We did observe a comparable low-affinity adenosine-induced vasodilation in our model, which we attribute to activation of A_2b receptors. The vasodilatory actions of adenosine observed during KCl-induced vasoconstriction appear to be mediated exclusively by the low-affinity A_2b receptor. The lack of effect of the high-affinity A_2a receptor in this setting is consistent with a vasodilatory mechanism requiring K channel activation. We have previously demonstrated that K_ATP activation by hypoxia or pinacidil does not induce afferent arteriolar vasodilation in the presence of 30 mM KCl (15, 22). Low-affinity A_2b receptors also likely account for inhibition of myogenic responses by high adenosine concentrations in the presence of glibenclamide. Adenosine concentrations of 10-30 µM were required for both responses (compare Figs. 2 and 7).

Our findings thus indicate that both high-affinity A_2a and low-affinity A_2b adenosine receptors are expressed in the renal microvasculature of our model. It is not clear whether both receptor subtypes are present on the same cell type (e.g., the afferent arteriolar myocyte). Both responses were potentiated by Ro-20-1724, suggesting each is coupled to cAMP. However only the A_2a-mediated response requires K_ATP. It is difficult to imagine a signaling pathway in which cAMP mediates either K_ATP-dependent or K_ATP-independent vasodilation within a single cell, depending on which receptor (A_2a or A_2b) is activated. Conceivably, this phenomenon could reflect a compartmentalization of cAMP-dependent responses within the cell. Alternatively, differing cell types could be involved in each response. For example, adenosine is reported to have endothelial-dependent vasodilatory actions, and adenosine-induced endothelial activation could conceivably induce hyperpolarization in the underlying arteriolar smooth muscle cells (4, 17, 27).

We interpret our findings as indicating that the renal microcirculation has the potential of expressing the high-affinity adenosine A_2a receptor and that activation of this receptor modulates myogenic afferent arteriolar vasoconstriction through K_ATP. As discussed above, this type of adenosine response has not been reported in the normal kidney. Indeed, adenosine infusion or elevation of endogenous adenosine by uptake blockade normally elicits renal vasoconstriction via activation of A₁ adenosine receptors (2, 28). We consider it very likely that our ability to demonstrate the A_2a-mediated response is related to the hydronephrotic state. The nonfiltering hydronephrotic kidney would be anticipated to have lower metabolic activity and reduced renal interstitial levels of adenosine, which could induce an upregulation of the adenosine A_2a receptor.

Alternatively, it is possible that the conditions of our study (e.g., low basal adenosine levels, elevated pressure-induced arteriolar tone, low levels of endogenous vasoconstrictors) may have facilitated the observation of this response. Microdialysis studies report cortical renal interstitial adenosine concentrations to be approximately 200 nM in the anesthetized rat (3), and the glibenclamide-sensitive adenosine response we observed may not be readily demonstrated in the presence of high basal levels of adenosine. Adenosine levels may be lower in conscious animal preparations, but conscious animals generally have very low basal renal vascular tone, and renal vasodilation is difficult to demonstrate in the absence of significant renal vasoconstriction (e.g., see Ref. 14). Certainly, the ability of the normal kidney to generate myogenic tone in the anesthetized preparation, despite presumably high basal levels of adenosine, would argue against the presence of this receptor or, at least, its chronic activation in the normal kidney.

The presence of circulating or intrarenal vasoconstrictors may also impact on the ability to observe K_ATP-dependent responses in the normal kidney. We have recently reported that the K_ATP component of the afferent arteriolar actions of calcitonin gene-related peptide (CGRP), readily observed during pressure-induced vasoconstriction, is eliminated during angiotensin II-induced vasoconstriction, suggesting angiotensin II inhibits K_ATP activation (23). Studies in isolated smooth muscle cells indicate that vasoconstrictor agonists such as angiotensin II are capable of inhibiting K_ATP by stimulating protein kinase C (5). Accordingly, specific experimental conditions, including low basal intrarenal adenosine levels, elevated pressure-induced renal vascular tone and low levels of agonist-induced vasoconstriction, may be required to observe this unusual adenosine response in the normal kidney. Of interest, an anecdotal report by Smits et al. (26) demonstrated that submicromolar levels of adenosine administered into the renal artery of conscious hypertensive patients caused a twofold increase in renal blood flow. Similarly, a study by Arakawa et al. (1) indirectly suggested the presence of a high-affinity adenosine vasodilatory mechanism in the normal, conscious dog. It is interesting to speculate that the high-affinity A_2a receptor may play some role in the regulation of normal renal microvascular reactivity or contribute to abnormal myogenic reactivity in pathophysiological states. Further investigations will be required to address these issues.

In summary, in the present study we describe three distinct actions of adenosine on the renal afferent arteriole of the in vitro perfused hydronephrotic rat kidney. In addition to adenosine A₁-induced vasoconstriction, we find two vasodilatory responses. One response corresponds to that previously ascribed to activation of low-affinity adenosine A_2b receptors and is elicited at adenosine concentrations of 10 µM and above. The second response was elicited at 10-100 nM adenosine and was blocked by glibenclamide, suggesting a mechanism coupled to activation of K_ATP. We suggest that this high-affinity response is mediated by renal microvascular adenosine A_2a receptors.