| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Departments of Medicine and Pharmacology, University of California San Diego and Veterans Affairs San Diego Health Care System, San Diego, California; and Institute of Pharmacology and Toxicology, Medical Faculty, University of Tuebingen, Tuebingen, Germany
ABSTRACT I. INTRODUCTION II. ADENOSINE GENERATION IN THE KIDNEY A. Renal Tissue Content of Adenosine 1. Normoxia and ischemia 2. Maleic acid 3. Hypertonic saline 4. S-adenosylhomocysteine hydrolase binds adenosine B. Extracellular Adenosine Concentration 1. Microdialysis technique 2. Sources of extracellular adenosine A) NUCLEOSIDE TRANSPORTERS. B) ECTO-5'-NUCLEOTIDASE IN GLOMERULI AND LUMINAL MEMBRANES OF TUBULES. C) ECTO-5'-NUCLEOTIDASE IN PERITUBULAR SITES. D) INTERSTITIAL ATP AS A PRECURSOR OF ADENOSINE. E) INTERSTITIAL CAMP AS A PRECURSOR OF ADENOSINE. C. Renal Excretion of Adenosine D. Concluding Remarks III. ADENOSINE RECEPTORS IN THE KIDNEY A. Adenosine A1 Receptors B. Adenosine A2a and A2b Receptors C. Adenosine A3 Receptors D. Coupling of Adenosine Receptors E. Concluding Remarks IV. VASCULAR ACTIONS OF ADENOSINE IN THE KIDNEY A. Exogenous Adenosine 1. Effects of intrarenal administration of adenosine on renal blood flow and GFR 2. Adenosine effects on pre- and postglomerular arteries 3. Adenosine actions on human kidneys B. Endogenous Adenosine 1. Postocclusive ischemia 2. Hypertonic saline infusion 3. Maleic acid 4. Application of adenosine receptor antagonists C. Mechanisms of Adenosine-Mediated Vasoconstriction and Vasodilation D. Factors That Modulate the Vascular Response to Adenosine 1. Dietary NaCl status and renin-angiotensin system 2. Renal artery pressure and prostaglandins 3. Inhibitors of NO synthase 4. Renal nerves 5. Insulin-dependent diabetes mellitus E. Concluding Remarks V. ADENOSINE AND TUBULOGLOMERULAR FEEDBACK A. Tubuloglomerular Feedback B. Altering TGF Responses by Manipulating Adenosine Receptor Activation or Adenosine Formation C. Absence of TGF Response in Adenosine A1 Receptor-Deficient Mice and Consequences on the Single-Nephron Level D. Adenosine Is a Mediator of TGF E. Concluding Remarks VI. ADENOSINE AND RENIN RELEASE A. Effects of Exogenous Adenosine Agonists on Renin Release B. Role of Endogenous Adenosine in the Control of Renin Release C. Concluding Remarks VII. ROLE OF ADENOSINE IN FLUID AND ELECTROLYTE TRANSPORT IN THE KIDNEY A. Proximal Tubule B. Medullary Thick Ascending Limb C. Distal Convoluted Tubule and Cortical Collecting Duct D. Inner Medullary Collecting Duct E. Concluding Remarks VIII. ADENOSINE AND METABOLIC CONTROL OF ORGAN FUNCTION: DOES IT APPLY TO THE KIDNEY? IX. PATHOPHYSIOLOGICAL ASPECTS A. Radiocontrast Media-Induced Acute Renal Failure B. Ischemia-Reperfusion Injury X. SUMMARY AND PERSPECTIVES GRANTS ACKNOWLEDGMENTS REFERENCES
 |
ABSTRACT |
|---|
 Top NextReferences |
|---|
|
I. INTRODUCTION |
|---|
|
TopPreviousNextReferences |
|---|
The way the mammalian kidney developed to fulfill its role in body fluid and NaCl homeostasis includes a high glomerular filtration rate (GFR, 
180 l/day in humans). Subsequently, nearly all of the filtered fluid and NaCl is reabsorbed along the nephron such that only 1% of the glomerular filtered amounts are excreted in the urine. As a consequence, GFR and reabsorption have to be coordinated to prevent excessive renal losses of fluid and NaCl. Intrarenal mechanisms that contribute to this coordination from minute to minute include glomerulotubular balance and tubuloglomerular feedback (TGF). According to glomerulotubular balance, an increase in GFR and thus the filtered amounts of NaCl causes a near-proportional increase in NaCl reabsorption in all segments of the tubular and collecting duct system. In absolute amounts, this is particularly evident in the proximal tubule and the medullary and cortical thick ascending limb of Henle (TAL) where the bulk of the filtered NaCl is reabsorbed. The NaCl load at the end of the TAL is sensed, and the TGF establishes an inverse relationship between this tubular NaCl load and the GFR of the same nephron. This stabilizes and limits the NaCl load to the further distal segments, which have a limited capacity to alter NaCl reabsorption from minute to minute, but which, under systemic neurohumoral control, are the site of fine regulation of NaCl balance.
Because urinary fluid and NaCl excretion closely match intake and because variations in fluid and NaCl intake are minor compared with the total amounts filtered, it follows that GFR is the major determinant of renal fluid and NaCl reabsorption. Reabsorption of NaCl requires energy, and the GFR is thus the major determinant of renal O2 consumption. It follows that GFR is an important determinant of salt balance but also of metabolic aspects of kidney function. Therefore, intrarenal mechanisms that limit GFR when the ratio of renal O2 supply to demand is significantly reduced could be beneficial. Blood flow to the kidneys amounts to 20% of cardiac output, and the cortical blood flow, which determines GFR, is also rather high and thus the O2 supply of the kidney cortex is generous. The situation is quite different, however, in the medulla, where blood flow derives from the postglomerular circulation of the deep cortex. To be able to concentrate the urine, the kidney uses a mechanism that involves a low blood flow to the renal medulla and a counter-current system. As a consequence, the O2 supply to the renal medulla is low, although active NaCl reabsorption in the medullary TAL is essential for the counter-current system. This situation asks for an intrarenal metabolic control to prevent hypoxic injury in the medulla. Considering on the other hand that cortical blood flow is high but determines via GFR the tubular NaCl load and thus transport work in cortex and medulla, it follows that an intrarenal metabolic control of kidney function requires differential effects on the vasculature and transport systems of cortex and medulla.
Adenosine is a well-studied candidate that participates in intraorgan control mechanisms including acute responses to increased work load (25, 26, 33, 234, 271, 299). The latter increases ATP hydrolysis and adenosine generation. Extracellular adenosine acts on specific G protein-coupled receptors and, in organs like brain, heart, or skeletal muscle, induces vasodilation to match the delivery of O2 and metabolic fuel to consumption. Additional potential defense mechanisms include a suppression of the release of stimulatory neurotransmitters and a reduction of cell activity (see Fig. 1). The idea of adenosine acting as a "retaliatory metabolite" or as a "homeostatic metabolite" has been reviewed (234, 299). Before this concept can be applied to the kidney and intrarenal control mechanisms, it is necessary to understand mechanisms of adenosine formation in the kidney, the effects of adenosine on kidney function, and the determinants of kidney energy consumption and supply.
|
|
II. ADENOSINE GENERATION IN THE KIDNEY |
|---|
|
TopPreviousNextReferences |
|---|
When rat kidneys were snap frozen, an adenosine tissue content of 5 nmol/g wet wt was found in normoxic kidneys and the tissue content increased severalfold within a few minutes of renal ischemia (by occlusion of the renal artery) (253). These findings were confirmed and extended to the kidneys of dogs and cats (215). The increase of adenosine tissue content during ischemia preceded the increase of inosine and hypoxanthine, whereas ATP levels were rapidly reduced (245, 253) (see Fig. 2). In vitro studies with suspensions of medullary TAL demonstrated a hypoxia-stimulated adenosine release into the medium. This increase was completely blocked by furosemide or ouabain, indicating that adenosine release is related to electrolyte transport (21). Thus the kidney makes no exception with respect to enhanced adenosine generation and cellular release following ATP breakdown during oxygen deficiency or enhanced organ work. Further evidence is provided by the renal response to maleic acid and hypertonic saline as outlined in the following section.
|
ATP depletion without inducing ischemia can be achieved in the kidney by the use of maleic acid. This stereoisomer of fumarate forms a stable complex with coenzyme A, resulting in a fall in ATP levels mainly in the proximal tubules (190, 256, 284). When in rat kidneys ATP levels were reduced by intravenous administration of maleate, adenosine tissue content was increased threefold (245, 247). In dogs, maleate increased adenosine release into renal venous blood and into urine severalfold while leaving arterial adenosine unchanged (10). The changes of kidney function after maleate administration are discussed in section IV.
An experimental maneuver to induce a rapid increase in renal transport work that leads to a fall in ATP and an increase of adenosine tissue content is the short time infusion of hypertonic saline into the thoracic aorta (251). This maneuver increases renal NaCl reabsorption because the amount of NaCl filtered significantly increases due to the rise in plasma NaCl concentrations and a concomitant increase in GFR by 20%. A rise in GFR in response to short time exposure to hypertonic saline (within 10 min) is in accordance with findings of Young and Rostorfer (378). In contrast, prolonged infusion of hypertonic saline directly into the renal artery leads to sustained vasoconstriction that can be blocked by theophylline (94). The results of this experimental series are shown in Figure 3. The experiments demonstrated a reciprocal relationship between the fall in ATP and the increase in adenosine tissue levels depending on the absolute amount of Na+reabsorbed by the kidney. Intravenous infusion of hypertonic saline also led to a fall in renal ATP tissue levels, which returned to control levels following furosemide administration (80, 250). The hemodynamic response of the kidney to intra-arterial infusion of hypertonic saline is discussed in section IV.
|
An unsolved issue is the free intracellular adenosine concentration in the kidney. Considering a renal tissue content of adenosine of 5 nmol/g wet wt (253) and an extracellular adenosine concentration of 100 nM under normoxic conditions (see sect. IIB), a calculated cytosolic adenosine concentration of 5 µM appears to be unrealistically high. Therefore, one has to assume that >90% of intracellular adenosine is protein bound, which makes it unavailable for deamination by the intracellular enzyme adenosine deaminase (ADA). One candidate protein for binding adenosine intracellularly is the cytosolic enzyme S-adenosylhomocysteine (SAH) hydrolase (163, 341) (see Fig. 2). This enzyme hydrolyzes reversibly SAH into adenosine and homocysteine. Two binding sites for adenosine of the SAH hydrolase could be identified, one with high affinity (kD1 = 9.2 nM) and one with low affinity (kD2 =1.4 µM) (166, 167). Both adenosine binding sites of SAH hydrolase are controlled by the ratio of NAD+/NADH (162, 163, 165, 166). Based on the intracellular concentration of SAH hydrolase in the kidney of 2.2 µM and its binding capacity, one can calculate that 20% of intracellular adenosine is bound to SAH hydrolase under normoxic conditions (124, 125, 341). Thus other proteins that can bind intracellular adenosine have to be identified. The formation of SAH in the isolated perfused guinea pig hearts was used by Deussen et al. (63) to calculate that free adenosine concentrations in the cytosol are 80 nM. Notably, these estimated free cytosolic adenosine concentrations are remarkably similar to the basal adenosine concentrations measured in the kidney interstitium (see below). This may not be unexpected given the presence of equilibrative nucleoside transporters particularly in the basolateral membranes of kidney tubules (see below) across which adenosine should equilibrate.
In addition to adenosine binding, SAH hydrolase may serve as a target of intracellular adenosine actions (167). Because adenosine inhibits in vitro SAH hydrolase activity in nanomolar concentrations, this action would increase cytosolic SAH levels and thus diminish the methylation potential [the ratio of S-adenosylmethionine (SAM) to SAH], which regulates transmethylation reactions in the cell. Moreover, it was shown that renal tissue content of SAH increases severalfold after ischemia (161). In this respect, it is of interest to note that in addition to being expressed in the cytosol of nearly all kidney cells, a prominent staining of SAH hydrolase can be seen in the nuclei of podocytes (164) and that SAH hydrolase accumulates in nuclei of transcriptional activated cells (275). In summary, as in other organs, most of the adenosine content of the kidney is sequestered to intracellular adenosine binding proteins including SAH hydrolase, and much has to be learned about the functional consequences of the interactions between intracellular adenosine and SAH hydrolase.
B. Extracellular Adenosine Concentration
With the use of the microdialysis technique in rat kidneys, it was found that mean values of interstitial fluid adenosine concentrations ([ADO]ISF) are 55 nM in cortex and 212 nM in medulla (381). Infusion of ATP-MgCl2 resulted in a roughly twofold elevation of adenosine, inosine, hypoxanthine, and uric acid, indicating the capacity of the kidney to metabolize exogenous ATP (381) (see Fig. 2). Employing microdialysis tubes inserted into both kidney cortex and medulla in rats on a normal-NaCl diet, Siragy and Linden (308) found [ADO]ISF in the dialysate from the cortex of 63 nM and from the medulla of 157 nM. Notably, rats consuming a high-NaCl diet had renal cortical and medullary dialysate adenosine concentrations that were increased about sevenfold, whereas low-NaCl diet lowered [ADO]ISF by 64% in both kidney regions compared with normal diet (308). The mechanisms involved are not absolutely clear but may relate to the fact that rats can respond to a high NaCl diet with an increase in GFR (346). As a consequence, absolute and fractional renal NaCl excretion are increased to match increased intake but at the same time the associated increase in GFR enhances absolute renal NaCl reabsorption (primarily in proximal tubule and thick ascending limb)(346), and thus possibly adenosine formation. Further studies are required to clarify this issue. In another study using the microdialysis technique in rabbit kidneys, [ADO]ISF was found to be 293 nM in the cortex under basal conditions and increased threefold after induction of systemic hypoxia (238). In dogs, renal [ADO]ISF was 117 nM and did not change during reduction of renal perfusion pressure within the autoregulatory range when GFR remained constant (236). One study actually looked at the time course of [ADO]ISF in the effluent of the microdialysis tubes in conscious rats: [ADO]ISF revealed immediately after implantation very high levels that fell subsequently to concentrations between 100 and 200 nM within 2–6 days (237). Although the microdialysis technique may have some limitations, the data show clearly that adenosine is present in the renal interstitium at concentrations sufficient to activate G-coupled adenosine receptors (see sect. III), i.e., in the mid to high nanomolar range under normal-NaCl diet and normoxic conditions with concentrations being about two- to fourfold greater in medulla than in cortex.
2. Sources of extracellular adenosine
To what extent interstitial adenosine is derived from intracellular or from extracellular sources is incompletely understood. It is established that AMP is a major precursor for intracellular and extracellular adenosine formation (see Fig. 2). However, it would be important to know the free concentration of cytosolic AMP and adenosine; the potential efflux rates of ATP, AMP, and adenosine from the cytosol into the interstitium; and the contribution of ecto-5'-nucleotidases to the interstitial adenosine concentration. With the use of the technique of nuclear resonance spectroscopy for 31P phosphorus, it was found in cardiac tissues that free AMP amounts to <5% of total extracted AMP. The measured (96) and calculated (39, 355) concentrations of free AMP in the cytosol of cardiac tissue under normoxic conditions are in the range of 200 nM. These low concentrations are close to the assumed levels of free adenosine in the cytosol (39). Changes in the AMP-adenosine cycle via the activities of cytosolic 5'-nucleotidase and adenosine kinase (see Fig. 2) can lead to significant changes of intracellular adenosine formation (58, 96, 211). The cytosolic 5'-nucleotidase has recently been cloned (for review, see Ref. 138138), which should be helpful to further delineate this issue.
A) NUCLEOSIDE TRANSPORTERS. Cellular uptake and release of adenosine is mediated by nucleoside transporters. The nucleoside transporter proteins differentiated so far are divided into five distinct superfamilies that are functionally characterized and vary in substrate specificity. The concentrative nucleoside transporters CNT1-CNT3 [solute carrier (SLC) 28A1–28A3], which mainly localize to the apical membrane of renal epithelium, mediate the intracellular flux of nucleosides. The equilibrative nucleoside transporters ENT1–2 (SLC29A1-SLC29A2), on the other hand, primarily localize to basolateral membranes and mediate bidirectional facilitated diffusion of nucleosides and may thus contribute to cellular adenosine release when cytosolic concentrations increase. However, the knowledge base is far from clear to define satisfactorily how these transporters work, alone or in concert, and under varying intra- and extracellular conditions. Much has to be learned in this regard, and the interested reader is referred to recent reviews on the topic (17, 101, 331).
B) ECTO-5'-NUCLEOTIDASE IN GLOMERULI AND LUMINAL MEMBRANES OF TUBULES. Ecto-5'-nucleotidase is expressed in glomeruli including mesangial cells (45, 134, 189), where it may contribute to the generation of adenosine, which mediates the TGF mechanism (see sect. V). In the first loops of the proximal convoluted tubule, prominent staining of ecto-5'-nucleotidase is visible in the luminal brush-border membrane (56, 91, 112). In the cortical pars recta of the proximal tubule, ecto-5'-nucleotidase staining is low but increases slightly towards the medullary pars recta of the proximal tubule. Also in segments of the distal tubule (intercalated cells), a luminal staining was shown. The luminal localization of ecto-5'-nucleotidase is likely to be involved in the purine salvage pathway of the filtered or luminal released nucleotides that after dephosphorylation can affect luminal adenosine concentrations or be taken up by nucleoside carriers in the brush-border membrane (4, 306). It is, however, unlikely that these luminal ecto-5'-nucleotidase activities produce changes in adenosine concentrations at the basolateral sites of the tubular epithelium and in the interstitium (189, 283).
C) ECTO-5'-NUCLEOTIDASE IN PERITUBULAR SITES. The peritubular staining of ecto-5'-nucleotidase in the kidney has first been attributed to the cells of blood capillaries (112). Le Hir and Kaissling (189), however, demonstrated that the ecto-5'-nucleotidase-positive perivascular cells were in fact fibroblasts. Endothelial cells of the capillaries were negative. The fibroblasts in the interstitium of the kidney make contact with tubular cells and peritubular capillaries and form a sheath around afferent and efferent arterioles of the glomerulus. Under normoxic conditions the ecto-5'-nucleotidase-positive cells are exclusively located in the cortex and cannot be demonstrated in the medulla (56, 91). The intensity of fibroblast staining in the cortex changes in parallel to the increased production of erythropoietin. Under normal conditions, the staining is located predominantly in the deep cortex. The staining increases, however, throughout the whole cortex under challenges of hypobaric oxygen breathing or anemia (189). The exact role of ecto-5'-nucleotidase on the fibroblasts remains to be determined.
D) INTERSTITIAL ATP AS A PRECURSOR OF ADENOSINE. Recently, renal cortical interstitial ATP concentrations assessed by the microdialysis technique in anesthetized dogs were found to be 6.5 nM at a renal artery pressure of 131 mmHg. Stepwise reduction of renal perfusion pressure to 105 and 80 mmHg lowered ATP concentrations to 4.5 and 2.8 nM, respectively. Interstitial adenosine concentrations were 117 nM and remained unaltered by these changes of renal perfusion pressure, which did not affect renal blood flow or GFR (236). A similarly low interstitial ATP/adenosine ratio was reported in the isolated perfused rat heart (203). The 20- to 40-fold higher concentrations of adenosine than those of ATP do not readily support the assumption that this ATP is the major precursor of adenosine in the bulk phase of the interstitial fluid. ATP can serve as a major precursor of adenosine, however, if higher ATP concentrations exist in the unstirred layer at the surface of the plasma membranes, which are equipped with ectoenzymes to metabolize adenine nucleotides (see Fig. 2) and/or the generation of adenosine by these pathways is faster than the downstream adenosine metabolism or cell uptake. Extracellular generation of adenosine from ATP may contribute to the signaling mechanisms of the tubuloglomerular feedback (see sect. V).
E) INTERSTITIAL CAMP AS A PRECURSOR OF ADENOSINE.
According to the scheme of Figure 2, extracellular adenosine can be generated from cAMP, which is released by cells of the tubular or vascular system. Jackson and co-workers examined this possibility in several experiments. The cAMP added to the perfusate of isolated perfused kidneys can be converted to AMP by ecto-phosphodiesterases and subsequently to adenosine by ecto-5'-nucleotidases resulting in a release of AMP, adenosine, and inosine into the venous effluent (213). Addition of the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (IBMX) to the perfusate resulted in an almost complete block of AMP, adenosine, and inosine release. However, when the inhibitor of ecto-5'-nucleotidases, 
,
-methylene-adenosine-5'-diphosphate (AMPCP), was infused together with cAMP, the release of AMP was unchanged but adenosine and inosine release were nearly completely inhibited (213). Treating the isolated perfused kidneys with isoproterenol, the endogenously released cAMP is also extracellularly converted to AMP and adenosine yielding a threefold increase of adenosine, inosine, and hypoxanthine (214). The -blocker propranolol, IBMX, and AMPCP blocked the isoproterenol-induced increase of purines (214). Also in the isolated guinea pig gallbladder, a model of a transporting epithelium, substantial cAMP release into the extracellular space was found after stimulation with prostaglandins (265). Thus the formation of adenosine from extracellular cAMP suggests that adenosine by activation of adenosine A1 receptors, which can be coupled to an inhibitory Gi protein (see sect. IIID), can function as a local feedback inhibitor of adenylyl cyclase. This pathway may play a role for effects of adenosine on proximal tubular reabsorption (see sect. VII) as well as adenosine-mediated inhibition of both, renin release (see sect. VI) and vasopressin-stimulated transport in the inner medullary collecting duct (see sect. VII).
C. Renal Excretion of Adenosine
Is the renal adenosine excretion of physiological or pathophysiological significance? Only few data in the literature address this question. Thompson et al. (329) analyzed in anesthetized dogs the renal arterial-venous difference of adenosine and adenosine excretion kinetics following single injections of radiolabeled adenosine into the renal artery (329). Under basal conditions the concentrations of endogenous adenosine in plasma of renal vein and artery were 52–60 nM and not statistically different. Urinary adenosine concentration was 312 nM, and the excretion rate was 0.67 nmol/min. With the use of the single injection method, it was found that 12% of the injected adenosine was recovered in urine and 11% in the renal vein. This low venous recovery was due to cellular uptake by nucleoside transporters as evidenced by a threefold increase of adenosine recovery in the renal vein by the nucleoside transport inhibitor dipyridamole (329). Heyne et al. (127) studied renal adenosine excretion in 12 healthy volunteers under basal conditions, after water loading, and following low and high Na+ intake. It was found that adenosine excretion (3.2 nmol/min) was independent of urinary flow rate (2–19 ml/min), indicating that passive tubular back-diffusion does not significantly contribute to net adenosine excretion. Moreover, low- and high-Na+ diet did not change adenosine excretion per milliliter of GFR. These remarkably constant values under normal conditions could provide a basis for the evaluation of renal adenosine excretion as a marker of renal injury in various clinical settings. In fact, enhanced renal adenosine excretion rates were found during renal ischemia (215), maleic acid (10), radiocontrast media application (154), and methotrexate (12).
The available data have shown that adenosine is present in the normoxic kidney. The tissue content, 5 nmol/g wet wt, represents mainly the cytosolic fraction of renal adenosine, whereas only 2–5% of this amount is present in the extracellular compartments, such as tubular fluid/urine and interstitium. Since a calculated cytosolic adenosine concentration of 5 µM appears to be unrealistic, most of the intracellular adenosine must be bound to intracellular proteins including SAH hydrolase. Sources of extracellular adenosine in the kidney include cellular adenosine release as well as extracellular adenosine formation from ATP, AMP, and cAMP being released from the cells. After ATP depletion by ischemia, maleic acid, or hypertonic saline, extracellular adenosine concentrations increase indicating that in the kidney as in other organs adenosine generation is controlled by the phosphorylation potential of the cell. The functional implications are discussed in section VIII.
|
III. ADENOSINE RECEPTORS IN THE KIDNEY |
|---|
|
TopPreviousNextReferences |
|---|
With the use of the selective ligands cyclohexyladenosine (CHA) or N6-p-hydroxy-phenyl-isopropyl-adenosine (PIA), adenosine A1 receptors were first identified by autoradiography in sections of human and guinea pig kidney (257, 358). Moreover, binding sites on glomerular structures for adenosine A1 receptor ligands were reported in both species (55, 337). Specific binding was more recently reported of 3H-labeled 1,3-dipropyl-8-cyclopentylxanthine (DPCPX), an adenosine A1 receptor antagonist, to plasma membranes of immortalized cells derived from normal adult human proximal tubule (328). Among the adenosine receptors, the A1 subtype was the first to be cloned (208). The respective gene, in humans, was allocated to chromosome 1q32.1 (336). Several studies using molecular techniques showed the presence of the adenosine A1 receptor in the rodent kidney. As depicted in Table 1, adenosine A1 receptors are present in afferent arterioles, glomeruli including mesangial cells, juxtaglomerular cells, vasa recta, as well as in various segments of the tubular and collecting duct system including proximal tubule, thin limbs of Henle, TAL, and collecting ducts. In spite of the numerous renal effects of adenosine A1 receptor activation in humans, data on the localization of this receptor subtype in the human kidney on the molecular level have not been reported so far.
|
The adenosine A2 receptor family consists of two subtypes, the A2a and the A2b receptor, which possess a high and a low agonist affinity, respectively. The adenosine A2 receptor was first cloned from a canine thyroid cDNA library (196). The human adenosine A2a receptor gene was localized to chromosome 22q11.2 (181, 205). Similar to the adenosine A1 receptor, renal A2a receptors, so far, have been identified only in rodents (see Table 1). Both the mRNA and the protein of the adenosine A2a receptor were demonstrated in whole kidney preparations. Furthermore, mRNA for the adenosine A2a receptor was detected in the papilla of the rat kidney (357) and in glomeruli of rat and mouse kidney (356). Finally, adenosine A2a receptor mRNA was found in the outer medullary descending vasa recta (176).
Adenosine A2b receptors were first cloned from brain regions of human (272) and rat (282). The responding human gene was localized to chromosome 17p12 (151). There is only sparse information on the presence of adenosine A2b adenosine receptors in the kidney (see Table 1). Adenosine A2b receptor mRNA or protein was detected in whole kidney preparations. In addition, mRNA was detected in the cortical TAL and in the distal convoluted tubule (356) as well as in the outer medullary descending vasa recta (176), and more recently, adenosine A2b receptor mRNA was reported in baby hamster kidney cells (218) and at the protein and mRNA level in rat preglomerular vessels (150).
The fourth distinct adenosine receptor is the A3 subtype. It was first cloned from the rat striatum (379). The recombinant striatal adenosine A3 receptor differs completely compared with the other adenosine subtypes in agonist or antagonist binding. The human adenosine A3 receptor gene was localized to chromosome 1p (221). Adenosine A3 receptors have been detected, both on the mRNA and on the protein level, in whole kidney preparations of various species (see Table 1). In contrast, no distinct intrarenal localization has been reported so far. From radioligand binding studies, the presence of the adenosine A3 receptor in brush-border membranes isolated from pig kidney had been suggested (32). Interestingly, adenosine A3 receptor mRNA in the kidneys of young rats increased with age from the newborn state to early adolescence by more than one order of magnitude (225).
D. Coupling of Adenosine Receptors
The adenosine receptors belong to the superfamily of G protein coupling receptors. According to a consensus definition, adenosine A1 receptors induce, via pertussis toxin-sensitive Gi and Go proteins, adenylyl cyclase inhibition and phospholipase C (PLC) stimulation, whereas adenosine A2a and A2b receptor stimulation leads, via cholera toxin-sensitive stimulatory G proteins, to adenylyl cyclase activation (for detailed information, see reviews in Refs. 79, 85, 86, 88, 229). The adenosine A3 receptor appears to couple in a similar fashion as A1 receptors, via inhibitory Gq/11 protein, to adenylyl cyclase and PLC (for review, see Ref. 8686). Thus adenosine receptor subtypes appear to couple to more than one G protein and/or effector system. As outlined in Table 2 for adenosine A1 and A2 receptors, the coupling of adenosine receptors to the various effector systems in the kidney is, in general, in agreement with the above-mentioned concepts derived from extrarenal cell types.
|
Each of the family of adenosine receptors has been demonstrated in virtually all organs. In the last years, the knowledge on expression and signal transduction of adenosine receptors in the kidney has grown but is still sparse. In particular, differentiated localization of the adenosine receptors in this organ is incompletely defined. In general, the existing data, however, indicate that the mechanisms of adenosine receptor coupling as obtained from other tissues also apply to the kidney.
|
IV. VASCULAR ACTIONS OF ADENOSINE IN THE KIDNEY |
|---|
|
TopPreviousNextReferences |
|---|
1. Effects of intrarenal administration of adenosine on renal blood flow and GFR
Two reports in 1964 described adenosine-induced renal vasoconstriction in the anesthetized dog (333) and in the blood-perfused dog kidney (115), respectively. These findings were confirmed by other investigators (116, 243, 325). Importantly, also conscious dogs respond to intrarenal adenosine injection with vasoconstriction (27). In anesthetized dogs, rats, and cats, intra-arterial single injections of adenosine elicited a renal vasoconstriction, which is rapid in onset and short in duration (243, 252, 319). Continuous intra-arterial infusion of adenosine leads to an initial fall in renal blood flow that lasts, however, for only 1–2 min and then whole renal blood flow returns to or slightly above preinfusion levels. After cessation of adenosine infusion, a short-lasting increase of renal blood flow ("overshoot") can be observed (8, 243, 246, 255, 325) (see Fig. 4). The factors that can modulate the renal response to adenosine are discussed in section IVD.
|
A different approach of intrarenal adenosine administration was chosen by Pawlowska et al. (261) who infused adenosine into the renal interstitial space of rats via implanted capsules. The data clearly demonstrated that adenosine and two stable analogs, 2-chloro-adenosine and NECA, decreased GFR by 50–80% while leaving total renal blood flow unchanged (261). To analyze the intrarenal blood flow distribution during interstitial adenosine infusion, Laser-Doppler flow probes were used to assess medullary and cortical blood flow in rat kidneys. Cortical blood flow was reduced by adenosine infusion for 1 min while medullary blood flow increased above preinfusion levels after an initial short-lasting decrease. The adenosine A1 receptor agonist CPA reduced flow in both regions, whereas the specific adenosine A2 receptor agonist CGS-21680C increased medullary blood flow to 1.8-fold above baseline (2). When adenosine was infused directly into the medullary interstitium, outer and inner medullary blood flow were increased together with urine and Na+ excretion, a response that was sensitive to the adenosine A2 receptor antagonist 3,7-dimethyl-1-propargylxanthine (DMPX) (381).
Studies in the isolated in vitro perfused outer medullary descending vasa recta (OMDVR) revealed that abluminal adenosine application causes vasoconstriction at low concentrations (10 pM to 0.1 µM), whereas high concentrations (1–10 µM) reversed the vasoconstriction (307). These vessels of the OMDVR are equipped with adenosine A1 and A2a/A2b receptors (176) consistent with the notion that low adenosine concentrations caused vasoconstriction via the adenosine A1 receptors, whereas higher concentrations can vasodilate via adenosine A2 receptors.
2. Adenosine effects on pre- and postglomerular arteries
Direct videometric assessment of the vascular reactivity of pre- and postglomerular arteries can be achieved using the "split-hydronephrotic" kidney technique. In this experimental model adenosine induced a preglomerular constriction, when applied topically from the interstitial side, which was most prominent in the distal part of the afferent arterioles (64). With the use of adenosine agonists and antagonists with different receptor subtype specificity, the data supported the concept that activation of adenosine A1 receptors leads to constriction mainly of afferent arterioles near the glomerulus, whereas adenosine A2 receptor activation leads to dilation mainly of the postglomerular arteries (65, 90, 133). Another study in the isolated perfused hydronephrotic rat kidney preparation provided evidence for both adenosine-induced constriction and dilation of afferent arterioles, and these effects were dependent on activation of adenosine A1 and A2a/A2b receptors, respectively (327).
Experiments in isolated perfused rabbit afferent arterioles indicated an essentially exclusive presence of adenosine A1 receptors in the glomerular entrance segment of the afferent arteriole, whereas in more proximal regions adenosine A2 receptors appear to be expressed in low density (359). Notably, studies in rat juxtamedullary afferent arterioles using the in vitro blood-perfused juxtamedullary nephron technique revealed that the metabolically stable adenosine analog 2-chloroadenosine reduced the vessel diameter at a concentration of 1 µM, whereas at 100 µM afferent vasodilation was observed (140, 141). Similarly, studies in the blood-perfused rat juxtamedullary nephron preparation also indicated the presence of adenosine A1 and A2a receptors on afferent and efferent arterioles of juxtamedullary nephrons, such that adenosine A2a receptor-mediated vasodilation partially buffers adenosine-induced vasoconstriction in both pre- and postglomerular segments of the renal microvasculature (235). In comparison, isolated-perfused afferent arterioles from superficial cortex of rabbit kidneys constricted in a dose-dependent manner when exposed to adenosine added to either lumen or bath. Adenosine A1 receptor antagonists blocked this adenosine effect. Higher micromolar concentrations of adenosine added to either lumen or bath only induced vasodilation in the presence of an adenosine A1 receptor antagonist. This effect was blocked by an adenosine A2 receptor antagonist (376). These data indicate that adenosine A1 receptor-mediated afferent arteriolar constriction may dominate in superficial nephrons, whereas juxtamedullary or deep cortical nephrons can respond to high adenosine concentrations with adenosine A2 receptor-mediated vasodilation. This would be consistent with persistent superficial cortical vasoconstriction and deep cortical vasodilation in response to adenosine infusion (see above).
3. Adenosine actions on human kidneys
Edlund and Sollevi (72) infused adenosine intravenously in eight healthy awake volunteers in doses of 60–80 µg · kg–1 · min–1. GFR was lowered by 28% while blood pressure and renal blood flow were unchanged. As a result of systemic vasodilation, heart rate and plasma concentrations of epinephrine and norepinephrine were increased while total peripheral resistance was decreased. Plasma renin activity was unchanged despite the activation of the efferent sympathetic tone (72) (see sect. VI). Similar results were reported by Balakrishnan et al. (16). In a subsequent study, adenosine was directly infused into the renal artery of volunteers at a dose of 2–10 µg · kg–1 · min–1, i.e., nearly the same dose applied in the dog experiments depicted in Figure 4. The infusion of adenosine reduced GFR significantly in the volunteers and tended to increase renal blood flow determined by p-aminohippurate (PAH) clearance (71). For comparison, 40 µg · kg–1 · min–1 adenosine infused into the renal vein induced no systemic effects on the cardiovascular system or on plasma catecholamines, indicating that intra-arterial adenosine infusion is unlikely to change renal function by systemic responses. Moreover, single injections of adenosine into the renal artery at doses between 0.01 and 1,000 µg adenosine (dissolved in 1 ml saline) induced a rapid fall in renal blood flow, as measured by an intravascular catheter connected to Doppler flowmetry, which, similar to the experiments in dog (243), returned to preinjection levels within 5–30 s (209). Thus the renal hemodynamic response to injections or infusions of adenosine is very similar in dogs and humans.
The interruption of arterial blood supply to an organ and the observation of postocclusive flow patterns are used as methods to study the relationship between energy metabolism and blood flow regulation (26). In organs like heart, brain, and skeletal muscle, a postocclusive hyperemia was observed (26). More than a century ago it was found that the kidney makes an exception compared with other organs as the reperfusion following a release of an artery clamp was virtually absent (31). Direct measurements of renal blood flow in rats demonstrated that in the kidney the postocclusive blood flow pattern is characterized by a short-lasting vasoconstriction (253). Because adenosine tissue content increases within 30 s of ischemia threefold and because theophylline, in doses which can antagonize exogenously applied adenosine, also blocked the postocclusive vasoconstriction, we concluded that endogenous adenosine was responsible for this unique vascular response (253). A typical tracing from this type of experiments is shown in Figure 5.
|
As illustrated in part in Figure 3, measurements of ATP, ADP, and AMP in the rat kidney with and without hypertonic saline infusion into the thoracic aorta revealed that the increased renal work load by hypertonic saline led to a fall of renal ATP and the energy charge (171, 250) associated with a threefold increase in renal adenosine levels (251). The renal vasculature responded to prolonged hypertonic saline infusion into the renal artery with sustained vasoconstriction, and the renin secretion was reduced to virtually zero (93). In subsequent studies it was shown that theophylline could block both vasoconstriction and suppression of renin release (94, 318). Using a nonxanthine adenosine receptor antagonist, CGS 15934 A, which is devoid of inhibitory action on phosphodiesterases (368), Callis et al. (43) could antagonize in anesthetized dogs the hypertonic saline-induced renal vasoconstriction. Several studies confirmed the observation of sustained renal vasoconstriction due to hypertonic saline and supported the assumption that adenosine is mediating this response (62, 92, 202). Micropuncture experiments revealed that acute hypernatriemia decreases SNGFR measured by distal collection (with intact flow at the macula densa), whereas SNGFR from proximal collections (with interruption of flow to the macula densa) was increased, indicating that the TGF mechanism was responsible for the fall in GFR during hypernatriemia (291). Taken together, hypertonic saline infusion leads to an increase of the tubular Na+ load to which the kidney responds with 1) a fall in ATP, 2) an increase in adenosine tissue levels, 3) sustained renal vasoconstriction due to TGF activation, and 4) reduced rate of renin secretion. The latter two effects are sensitive to adenosine receptor antagonism. The role of adenosine in renin release is outlined in section VI.
As discussed in section IIA2, maleic acid lowers ATP and increases adenosine levels in the kidney (245, 247). This maleate action is associated with transport inhibition mainly in the proximal tubule (e.g., glucose, amino acids, bicarbonate, and phosphate). As a consequence, maleate reduced SNGFR in the rat kidney by 42% due to a rise in proximal tubular hydrostatic pressure and activation of TGF (190). A role for TGF was indicated by the finding that the ratio of proximal to distal measurements of SNGFR, i.e., without and with intact flow at the macula densa, was significantly increased by maleate. However, maleate also reduced SNGFR from proximal tubular collections, i.e., without intact flow at the macula densa, pointing to a TGF-independent component of SNGFR reduction in response to maleate. The adenosine receptor antagonist theophylline can attenuate the fall in GFR in response to maleate by 50% (247). Similar results were obtained in dogs (10). These data indicate that endogenous adenosine mediates, at least in part, the reduction of GFR in response to maleate, possibly by activating TGF as well as by adenosine-induced vasoconstriction outside the TGF pathway. The role of adenosine in TGF is outlined in section V.
4. Application of adenosine receptor antagonists
Application of receptor antagonists can provide information on the ambient activity of the system inhibited. In healthy, male subjects oral application of the adenosine A1 receptor antagonist FK-453 significantly increased GFR by 20% (determined by clearance of 51Cr-labeled EDTA after 2 and 3 h). FK-453 tended to increase effective renal plasma flow (determined by clearance of 125I-hippuran) without significant changes in mean arterial blood pressure (15). These data suggest that endogenous adenosine through activation of adenosine A1 receptors elicits a tonic suppression of GFR. This suppression may reflect at least in part the tonic influence of TGF on GFR as adenosine through activation of adenosine A1 receptors mediates TGF-induced afferent arteriolar constriction (see sect. V). In comparison, no effect on GFR could be observed in response to the adenosine A3 receptor antagonists MRS-1191 and MRS-1220 in the rat (227).
Infusion of the adenosine A1 receptor antagonist DPCPX into the renal medulla of rats did not change medullary blood flow, while the selective adenosine A2 receptor antagonist DMPX decreased medullary flow (381). In accordance, renal interstitial infusion of the specific adenosine A2 receptor agonist CGS-21680C in rats increased medullary blood flow to 1.8-fold above baseline (2). These data suggest that endogenous adenosine at physiological concentrations dilates medullary vessels via adenosine A2 receptors.
C. Mechanisms of Adenosine-Mediated Vasoconstriction and Vasodilation
To further elucidate the signaling mechanisms involved in adenosine-mediated vasoconstriction, Hansen et al. (111) performed studies in perfused afferent arterioles from mouse kidney. They observed that adenosine, when added to the bath, caused constriction in the concentration range of 10–9 to 10–6 M. Adenosine-induced vasoconstriction was stable for up to 30 min and was most pronounced in the most distal part of the afferent arterioles. Adenosine did not cause vasoconstriction in arterioles from mice lacking adenosine A1 receptors confirming a role of this receptor in adenosine-mediated vasoconstriction in the kidney. Further studies indicated that the constriction response to adenosine in afferent arterioles is mediated by adenosine A1 receptors coupled to a pertussis toxin-sensitive Gi protein and subsequent activation of PLC, presumably through 
-subunits released from Gi (111). Similarly, studies in the isolated perfused rat kidney indicated a pertussis toxin-sensitive step between the occupation of adenosine A1 receptors on renal vascular smooth muscle cells and vasoconstriction induced by increased Ca2+ influx through potential-operated Ca2+ channels (286). With regard to adenosine-induced renal vasodilation, a study using an isolated-perfused hydronephrotic rat kidney preparation showed that adenosine-induced dilation of afferent arterioles was mediated by adenosine A2a receptor-dependent activation of KATP channels (327). A recent study in mice using intravenous infusion of adenosine indicated that the resulting renal vasodilation is due to adenosine A2a receptor-mediated activation of endothelial NO synthase (109). Whether the same pathway contributes to medullary vasodilation in response to endogenous adenosine, which probably derives from the abluminal site, remains to be determined.
D. Factors That Modulate the Vascular Response to Adenosine
1. Dietary NaCl status and renin-angiotensin system
One prominent factor that modulates the renal response to low doses of adenosine is the NaCl diet through activation of the renin-angiotensin system. In rats and dogs, the kidney is rendered insensitive to the vasoconstrictive action of adenosine at conditions of a high NaCl diet and volume expansion when plasma renin activity is low (11, 249, 252, 254). Correspondingly, high renin states of the animal are associated with an increased potency of adenosine to induce vasoconstriction and to lower GFR (249, 252, 254). The vasoconstrictive action of exogenous and endogenous adenosine can be antagonized by angiotensin II receptor antagonists (64, 207, 249, 320). Likewise, inhibitors of angiotensin I converting enzyme can block adenosine-induced renal vasoconstriction (65). The assumed interaction of angiotensin II and adenosine in preglomerular vessels was confirmed in dogs (106, 107). More detailed studies employing micropuncture experiments in the rat and in isolated perfused afferent arterioles from rabbits revealed a mutual dependency and cooperation of adenosine and angiotensin II in producing afferent arteriolar constriction (360). Further evidence for this conclusion was provided by the observation that angiotensin II AT1 receptor knockout mice show a markedly reduced vasoconstrictor response to the specific adenosine A1 receptor agonist CHA (338). Vice versa, deficiency of adenosine A1 receptors diminishes the effectiveness of angiotensin II to constrict afferent arterioles and to reduce GFR (108). The cellular mechanisms involved in this mutual dependency and cooperation remain to be determined.
2. Renal artery pressure and prostaglandins
A reduction of renal perfusion pressure to 65–70 mmHg by clamping the aorta above the origin of the renal arteries nearly abolishes adenosine-induced vasoconstriction in dogs and rats (103, 210). In rats, administration of indomethacin to inhibit prostaglandin synthesis can restore the adenosine-induced vasoconstriction under these conditions (103) and was shown to potentiate the vasoconstrictive action of endogenous and exogenous adenosine (268). Similarly, a 100-fold increase in sensitivity to vasoconstriction induced by single injections of adenosine into the renal artery of cat kidneys was achieved by pretreatment of the animals with meclofenamate (319). These data indicate that prostaglandins counteract adenosine-induced renal vasoconstriction. Adenosine-induced vasodilation, mainly in the deep cortex, however, is not sensitive to inhibitors of the cyclooxygenase (2, 316).
Another system that counteracts the vasoconstrictive action of many autacoids of the kidney is the production of NO by different NO synthases (NOS) in the kidney. After administration of unselective NOS inhibitors like N
-nitro-L-arginine, the sensitivity of the kidney to the vasoconstrictive action of adenosine is enhanced at least 10-fold (18, 269). This is consistent with the finding that adenosine infusion-induced vasodilation via adenosine A2a receptors is mediated by activation of endothelial NOS (109).
In virtually all organs, adenosine inhibits the transmitter release from nerve endings by activation of presynaptic adenosine A1 receptors. This inhibition may involve regulation of the Ca2+ influx via voltage-dependent calcium channels and hyperpolarization of the presynaptic nerve terminal (for review, see Ref. 70). In the kidney, adenosine can inhibit stimulation-evoked norepinephrine overflow, but it was also found to potentiate the postsynaptic norepinephrine-induced vasoconstriction (120, 121). Thus the numerous observations on the interactions between adenosine and the firing rate of nerve endings most likely apply also to the kidney.
5. Insulin-dependent diabetes mellitus
Renal function in early insulin-dependent diabetes mellitus (IDDM) is characterized by glomerular hyperfiltration. The hypothesis was brought forward that a deficiency of intrarenal vasoconstrictive mechanisms contributes to the observed hyperfiltration (347). Therefore, dipyridamole, an adenosine reuptake inhibitor, was administered to rats with streptozotocin-induced IDDM for 4 wk. Thereafter, m
