Ecto-5‘-nucleotidase (cd73)-dependent and -independent generation of adenosine participates in the mediation of tubuloglomerular feedback in vivo

【摘要】  Tubuloglomerular feedback (TGF) describes a sequence of events linking salt concentrations in tubular fluid at the macula densa to the vascular tone of the afferent arteriole and thus to the glomerular filtration rate (GFR) of the same nephron. The signal transduction pathways of TGF remain incompletely understood, but both ATP release from macula densa cells and local formation of adenosine were suggested to be involved in the process. To test the role of extracellular formation of adenosine by ecto-5'-nucleotidase (cd73) in TGF, in regulation of GFR, and in tubular reabsorption, renal clearance and micropunture experiments were performed in cd73 wild-type (cd73 +/+ ) and knockout mice (cd73 -/- ). The cd73 -/- mice presented normal mean arterial blood pressure, but modestly lower whole kidney and single nephron GFR (SNGFR). Fractional reabsorption of Na + and K + up to the late proximal tubule, distal tubule, as well as urine were not significantly different between cd73 -/- and cd73 +/+ mice. Lack of cd73 resulted in a diminished TGF response, as indicated by smaller changes of stop-flow pressure in response to increasing loop of Henle perfusion from 0 to 25 nl/min, smaller differences in SNGFR determined from paired proximal and distal tubular collections, and by smaller fractional changes of distal SNGFR in response to adding 6 nl/min of artificial tubular fluid to free-flowing proximal tubules. The TGF response in cd73 +/+ mice and the residual TGF response in cd73 -/- mice were completely inhibited by adenosine A 1 -receptor blockade. The results suggest that extracellular formation of adenosine by ecto-5'-nucleotidase (cd73) is dispensable for normal fluid, Na +, or K + reabsorption along the nephron, but contributes to the regulation of GFR. Adenosine generated by both ecto-5'-nucleotidase (cd73)-dependent and -independent mechanisms participates in the mediation of TGF in vivo.

【关键词】  micropuncture stopflow pressure tubular reabsorption

THE TUBULOGLOMERULAR FEEDBACK (TGF) describes a sequence of events linking salt concentrations in the tubular fluid at the macula densa to the vascular tone of the afferent arteriole of the same nephron. When salt concentrations at the macula densa increase, single nephron glomerular filtration rate (SNGFR) is reduced by preglomerular vasoconstriction. By inversely linking the salt concentration at the macula densa with SNGFR, the TGF mechanism provides an efficient way to stabilize the salt and fluid load to the distal nephron, facilitating fine regulation of water and electrolyte homeostasis at the latter site.

The signal transduction pathways involved in the TGF response remain poorly understood, but both ATP release from macula densa cells ( 1 ) as well as local formation of adenosine were suggested to play important roles in the process ( 23, 24, 25, 30 ). With regard to the role of adenosine, the TGF response is completely absent in adenosine A 1 -receptor knockout mice ( 3, 32, 36 ) and significantly suppressed during local inhibition of adenosine synthesis by pharmacological blockade of 5'-nucleotidase ( 26, 33 ). Moreover, clamping local adenosine A 1 -receptor activation at constant levels inhibits TGF activity ( 33 ). These findings suggested that a normal TGF response requires adenosine to fluctuate with changes in salt concentration at the macula densa, indicating that adenosine acts as a mediator of TGF in vivo ( 33 ). By integrating roles of ATP release and adenosine A 1 -receptor activation in one concept, it seems possible that ATP is released from macula densa cells into the interstitium of the extraglomerular mesangium, where it is broken down by nucleoside triphosphate diphosphohydrolases to AMP ( 1, 14, 15 ). The latter is converted to adenosine by cd73 to activate adenosine A 1 receptors and cause local vasoconstriction ( 26, 35 ). The cd73 is a glycosylphosphatidylinositol-anchored ubiquitous enzyme that catalyzes the extracellular conversion of adenine nucleotides to adenosine ( 14, 15, 19, 39 ). Extracellular formation of adenosine is thought to contribute to renal physiology and pathophysiology ( 8, 14, 15 ). With regard to TGF, a recent study using the gene-targeting approach provided further evidence for the outlined concept by showing that TGF was impaired in ecto-5'-nucleotidase (cd73) knockout mice ( 5 ). These knockout mice were generated by a gene exchange strategy that leaves the neomycin-resistance cassette within the targeted locus and using the C57BL/6J mouse strain as an inbred host.

Another line of ecto-5'-nucleotidase (cd73)-deficient mice was now independently generated using the conditional knockout strategy based on the cre-loxP system, to ensure the effective removal of the selected cassettes ( 17 ), and the NMRI (Naval Medical Research Institute) strain was chosen as an inbred host ( 7 ). It has been recently reported that this strain of cd73-deficient mice has disrupted degradation of AMP to adenosine as well as significantly altered coronary vascular tone, platelet activation, and leukocyte adhesion to the vascular endothelium ( 17 ). The present study was performed to assess TGF in this line of cd73-deficient mice, which were generated by a different gene-targeting approach and a different genetic background. Because the previous line of mutant mice still had 50% preserved TGF activity ( 5 ), the present study aimed to further assess whether the residual TGF activity in the absence of cd73 is also mediated by adenosine A 1 receptors or depends on ATP-activated purinergic P2X receptors. The latter possibility could arise due to local ATP accumulation when cd73 is lacking. Based on the findings that adenosine has diverse effects on tubular fluid and electrolyte transport ( 4, 14, 15, 28 ) and that acute local blockade of 5'-nucleotidase may affect tubular fluid reabsorption ( 33 ), we also analyzed glomerular filtration rate (GFR), SNGFR, and fluid and electrolyte reabsorption along the nephron in cd73 wild-type (cd73 +/+ ) and knockout (cd73 -/- ) mice.

METHODS

Animal care and generation of mutant mice. Mice were kept on a standard mouse chow (Altromin, Heidenau, Germany) and were taken care of in accordance with the European Convention on Animal Care. Animal experiment protocols were conducted according to the German law for the welfare of animals and were approved by local authorities. Mice deficient for ecto-5'-nucleotidase (cd73) were generated as previously described ( 17 ). Heterozygous mice were interbred to produce homozygous cd73 +/+ and littermate cd73 -/- mice. Both male and female mice were used for the current experiments.

Histochemical detection of cd73/ecto-5'-nucleotidase activity in kidney sections. Histochemistry was performed in frozen kidney sections using a lead phosphate method (pH 7.4), as described ( 2 ). To block alkaline phosphatase activity, kidney sections were preincubated for 15 min in a Tris-maleate sucrose buffer with 5 mM levamisole (Sigma-Aldrich Chemie, Steinheim, Germany) at room temperature. AMP was added to the substrate solution at a concentration of 1 mM. After the sections were washed with demineralized water, lead orthophosphate, which was precipitated as a result of cd73 activity, was visualized as brown deposits ( 2 ).

Surgical preparation for clearance and micropuncture experiments. Surgical preparation was performed as described before ( 9, 38 ). Briefly, mice were anesthetized using 100 mg/kg ip Inactin (Sigma-Aldrich Chemie) and 100 mg/kg im ketamine (CuraMED Pharma, Karlsruhe, Germany). Animals were then placed on a temperature-controlled operating table to keep rectal temperature at 37°C. Tracheostomy was performed, and the right jugular vein was cannulated for continuous infusion of 2.25% BSA dissolved in 0.85% NaCl at a rate of 350 µl·h -1 ·30 g body wt -1. [ 3 H]inulin was added in the infusion for evaluation of whole kidney GFR and SNGFR. Arterial blood pressure was recorded via a catheter inserted into the left femoral artery. The left kidney was exposed by flank incision, carefully freed of perirenal fat, and immobilized in a lucite cup. The kidney was covered with prewarmed paraffin oil. After completion of surgery and stabilization of the animals for 30 min, timed urine collections were performed via a catheter placed in the urinary bladder. Arterial blood samples were taken before and after completion of the urine collection period.

Free-flow collections and SNGFR measurements from distal and late proximal tubules. Proximal and distal tubular configuration of nephrons was identified by injecting small volumes of stained artificial tubular fluid (ATF) into a random proximal segment ( 10 ). After identification of nephron configuration, those nephrons, having both superficial proximal and distal loops, were subjected to paired free-flow tubular fluid collections. A collecting pipette (5- to 6-µm tip) was inserted into the first accessible distal tubular loop to perform a timed collection of tubular fluid (at least 3 min in duration) under free-flow conditions, utilizing a short mineral oil block. After the distal collection was finished and the short mineral oil block was washed out, another collecting pipette (10- to 11-µm tip) was inserted into the last proximal tubular loop of the same nephron to perform a paired free-flow collection. The collected tubular fluid was then transferred into constant-bore capillaries for determination of volume, [ 3 H]inulin, and Na + and K + concentrations (see below).

Assessment of TGF response by changes in stop-flow pressure to loop of Henle perfusion. As described previously, TGF response was assessed as the fall in early proximal tubule stop-flow pressure (SFP) in response to alteration of late proximal tubular perfusion rate ( 11 ). Briefly, a perfusion pipette filled with ATF (in mM: 130 NaCl, 10 NaHCO 3, 4 KCl, 2 CaCl 2, 7.5 urea, stained with 0.1% FD&C fast green, pH 7.4) and directly attached to a microperfusion pump was inserted into the last surface loop of the proximal tubule. Two immobile wax blocks were injected into the proximal tubule: one into the last accessible loop upstream from the perfusion pipette, and the other into the tubular site originally used for identification. Another micropipette (1- to 3-µm tip), filled with sodium chloride (1.5 M) and connected with a servo-nulling device (World Precision Instruments, New Haven, CT), was inserted upstream from the proximal wax block to monitor early proximal SFP during perfusion of Henle's loop at 0 and 25 nl/min.

Assessment of TGF response by changes in SNGFR to increasing late proximal flow rate. First, distal tubular segments on kidney surface were identified, and timed collections of tubular fluid were performed utilizing short mineral oil blocks to determine early distal tubular flow rate and Na + concentration, as well as basal distal SNGFR. Then a perfusion pipette containing ATF (see above) was inserted into the last surface loop of the proximal tubule to add 6 nl/min, thereby enhancing the salt load to loop of Henle and macula densa. Under these conditions, the first distal tubular segments were punctured a second time for recollection of tubular fluid and assessment of the above-mentioned parameters.

Effects of P2X-receptor and adenosine A 1 -receptor blockade on the TGF response. To clarify whether the remaining TGF response observed in cd73 -/- was mediated by adenosine and to assess a potential contribution of accumulating ATP in the absence of cd73, SNGFR responses were determined, as described in the previous paragraph, in response to adding ATF to the free-flowing late proximal tubule (6 nl/min), with or without the P2X-receptor blocker NF279 (Tocris, Avonmouth, UK) ( 13 ) or adenosine A 1 -receptor blocker 8-cyclopentyl-1,3-dipropylxanthine (DPCPX; Sigma-Aldrich Chemie) ( 31 ) in the perfusate. NF279 is a potent and highly selective antagonist of P2X receptors ( 6, 16, 18, 27 ) and is ineffective on adenosine A 1 receptors ( 18, 27 ). NF279 at 20 µM, the concentration used in the present studies, has been reported to completely block pressure-induced and ATP-induced afferent arteriolar vasoconstriction in rat isolated juxtamedullary nephrons ( 13 ). Results from Lohse and coworkers ( 20 ) showed that DPCPX has 700-fold selectivity for adenosine A 1 receptors over the A 2 receptor and does not significantly inhibit the soluble cAMP phosphodiesterase activity, even at millimolar concentrations. DPCPX in concentrations of 100 µM has been shown before to inhibit the TGF response when the perfusion rate of the loop of Henle is increased as well as adenosine-induced afferent arteriolar vasoconstriction ( 31, 37 ).

Analytical and statistical methods. Kidney weight was determined after decapsulation at the end of micropuncture experiments. Urinary flow rate was determined gravimetrically. Urine and plasma were analyzed for Na + and K + concentration by flame photometry (ELEX 6361, Eppendorf). Tubular Na + and K + concentration in the collected tubular fluid were determined as previously described ( 10, 11, 38 ), employing a microflame photometer. Concentration of [ 3 H]inulin in plasma, urine, and tubular collections was measured by liquid-phase scintillation counting. GFR and SNGFR were calculated according to standard formula. All data were tested for significance using paired or unpaired Student t -test as appropriate, and only results with P < 0.05 were considered statistically significant.

RESULTS

Renal histochemistry of ecto-5'-nucleotidase (cd73) activity. As shown in Fig. 1, AMPase activity was found in luminal membranes of proximal tubules and was especially abundant in the glomerular mesangium and in the cortical interstitium of cd73 +/+ mice. The presence of levamisole in the reaction mixture precluded any staining due to alkaline phosphatase activity. These data are in accordance with previous observations by Le Hir and Kaissling ( 19 ). Importantly, no activity of ecto-5'-nucleotidase was detectable in the kidney sections of cd73 -/- mice.

Fig. 1. Enzyme histochemical staining of kidney sections of cd73/ecto-5'-nucleotidase wild type (cd73 +/+; A ) and knockout mice (cd73 -/-; B ). Levamisole (5 mM) was applied to block alkaline phosphatase activity. AMP-hydrolyzing activity was found in glomeruli (asterisks), at luminal membranes of proximal tubules (PTs) (star), as well as in the interstitium (arrows) in cd73 +/+ mice. Distal tubules (DTs) were devoid of enzyme activity (circle). No staining was found in cd73 -/- mice, indicating that the AMP-hydrolyzing activity detected in cd73 +/+ mice reflected cd73 activity. Bar represents 100 µM ( A and B ).

Blood pressure and whole kidney function in anesthetized mice. Body weight, mean arterial blood pressure, as well as arterial hematocrit and Na + and K + concentrations in plasma were similar in cd73 +/+ and cd73 -/- ( Table 1 ). Urinary flow rate and fractional urinary excretion of fluid, Na +, and K + were not significantly different between the genotypes ( Table 1 and Fig. 2 ). However, two kidney GFR as well as kidney wet weight were modestly lower in cd73 -/- than cd73 +/+ mice ( Table 1 ).

Table 1. Systemic and whole kidney parameters in anesthetized cd73/ecto-5'-nucleotidase wild type (cd73 +/+ ) and knockout mice (cd73 -/- )

Fig. 2. Fractional delivery of fluid, Na +, and K + to late PT, early DT, and urine in cd73 +/+ (19 nephrons/5 mice) and cd73 -/- (21 nephrons/5 mice). * P < 0.05 vs. cd73 +/+.

Single nephron function. Consistent with results for two kidney GFR, SNGFR measured from both early distal (DT) and from late proximal tubular (PT) sites was significantly lower in cd73 -/- than cd73 +/+ mice ( Fig. 3 ). Lower SNGFR was associated with lower flow rates at both late proximal and early distal tubular sites in cd73 -/- vs. cd73 +/+ mice (proximal: 6.3 ± 0.4 vs. 9.4 ± 0.8 nl/min, and distal: 2.3 ± 0.3 vs. 4.4 ± 0.4. nl/min, P < 0.05 each). Fractional delivery of Na + and K + to the late proximal tubule, to the distal tubule, and to the final urine was similar in both genotypes ( Fig. 2 ). Fractional delivery of fluid to the last proximal loop on the surface and to the final urine was likewise similar in cd73 -/- and cd73 +/+ mice ( Fig. 2 ). Fractional delivery of fluid up to the early distal tubule was modestly reduced in cd73 -/- vs. cd73 +/+ mice (see DISCUSSION ).

Fig. 3. A : single nephron filtration rate (SNGFR) measured from late PT and early DT of the same nephrons in cd73 +/+ (16 nephrons/5 mice) and cd73 -/- (14 nephrons/4 mice). Distal SNGFR-to-proximal SNGFR ratios ( B ) and absolute differences between SNGFR from proximal and distal measurements ( C ) are shown in both genotypes. * P < 0.05 vs. cd73 +/+.

TGF response. In cd73 +/+ mice, SNGFR determined from early distal tubular collections, i.e., with ambient flow at the macula densa, was significantly lower than SNGFR determined from late proximal collections, i.e., with acute reduction in the flow passing the macula densa, consistent with TGF-mediated suppression of SNGFR under ambient conditions. In comparison, a significant difference in SNGFR between proximal and distal measurements was not detected in cd73 -/- mice ( Fig. 3 ). When the late proximal perfusion rate was increased from 0 to 25 nl/min, SFP decreased from 38.3 ± 0.3 to 32.1 ± 0.8 mmHg in cd73 +/+, whereas, in cd73 -/-, SFP fell from 37.5 ± 0.4 to 34.2 ± 1.4 mmHg. Both the absolute and fractional changes of SFP were significantly reduced in cd73 -/- vs. cd73 +/+ mice ( Fig. 4 ). Consistent with TGF activation, cd73 +/+ mice responded to the addition of 6 nl/min of ATF to the last surface loop of the proximal tubule with an increase in distal Na + delivery and a reduction in SNGFR, as determined by distal tubular collection ( Fig. 5 ). This response of SNGFR was significantly attenuated in cd73 -/- mice, consistent with an attenuated TGF response. As a consequence, the increase in distal Na + delivery was significantly greater in cd73 -/- mice.

Fig. 4. Change in stop-flow pressure (SFP) in response to increasing Henle's loop perfusion rate from 0 to 25 nl/min in cd73 +/+ (24 nephrons/7 mice; A ) and cd73 -/- mice (15 nephrons/6 mice; B ). Absolute ( C ) and relative ( D ) changes of SFP in cd73 +/+ and cd73 -/- mice are shown. * P < 0.05 vs. cd73 +/+.

Fig. 5. Fractional changes in SNGFR and early distal sodium delivery ([Na + ] delivery ) in response to adding 6 nl/min of artificial tubular fluid (ATF) ± P2X receptor blocker NF279 (20 µM) or adenosine A 1 -receptor blocker 8-cyclopentyl-1,3-dipropylxanthine (DPCPX; 100 µM) into the last surface loop of the PTs under free-flow conditions. N = 14-20 nephrons per group. * P < 0.05 vs. cd73 +/+. # P < 0.05 vs. respective ATF only.

Effects of adenosine A 1 -receptor or P2X-receptor blockade. The adenosine A 1 -receptor blocker DPCPX (100 µM) completely blunted the SNGFR response to adding ATF to the free-flowing late proximal tubule in cd73 +/+ mice and completely inhibited the residual SNGFR response in cd73 -/- mice. As a consequence, the increases in early distal Na + delivery were greater in both genotypes when DPCPX was added to the perfusate ( Fig. 5 ). In comparison, the P2X-receptor blocker NF279 (20 µM) did not alter SNGFR responses or early distal Na + loads in cd73 +/+ or cd73 -/- mice compared with control ATF in a statistically significant way ( Fig. 5 ). A tendency for somewhat smaller SNGFR responses to NF279 than control ATF, however, appeared apparent, especially in cd73 -/- mice (see DISCUSSION ).

DISCUSSION

In a recent report by Castrop et al. ( 5 ), the TGF response was found to be impaired in ecto-5'-nucleotidase (cd73)-deficient mice using the C57BL/6J mouse strain as an inbred host. The present cd73 knockout mice were inbred with the NMRI strain and were generated by a different gene knockout strategy ( 17 ), but revealed a similar phenotype with regard to TGF responses. The fall of SFP in response to increasing the loop of Henle perfusion rate from 0 to 25 nl/min was reduced by 48% in cd73 -/- vs. cd73 +/+ mice. This indicates that extracellular adenosine formation via ecto-5'-nucleotidase (cd73) is involved in TGF signaling, but does not account for the entire TGF response, at least in response to supraphysiological flow rates at the macula densa. To gain insights into the ambient influence of fluid and salt passing the macula densa on SNGFR, we performed free-flow tubular collections to compare distal to proximal tubular measurements of SNGFR, i.e., during ambient flow and salt concentrations at the macula densa vs. an acute reduction in these parameters. Evidence for a requirement of cd73 in the ambient macula densa-mediated control of SNGFR was provided, because, only in mice with intact cd73, proximal SNGFRs were significantly greater than distal SNGFRs.

The TGF mechanism operates by sensing changes in early distal salt concentrations in the tubular fluid at the macula densa and transmitting this signal to the afferent arteriole. Adenosine was first considered as a possible mediator of TGF by Osswald and colleagues, as adenosine antagonism inhibited, while adenosine accumulation potentiated, TGF ( 23, 24, 25 ). An increase of salt concentrations at the macula densa segment forces enhanced transport through the Na + -2Cl - -K + cotransporter, thus potentially increasing ATP consumption and adenosine formation in the macula densa cells. Alternatively or in addition, ATP itself may be released with increasing salt concentrations at the macula densa. Evidence for this was provided in the isolated juxtaglomerular apparatus by Bell and colleagues ( 1 ), who showed ATP release from macula densa cells when luminal NaCl concentration was increased. Considering the evidence that adenosine formation and activation of adenosine A 1 receptors are required for intact TGF ( 3, 32, 33, 36 ), it seems possible that the ATP released by the macula densa is converted, at least in part, to adenosine. In accordance, pharmacological blockade of 5'-nucleotidase inhibited TGF in the rat ( 33 ) and rabbit ( 26 ). The evidence that two knockout mice for cd73, which have different genetic backgrounds and were generated independently by different strategies, both present impaired TGF responses in vivo (Ref. 5 and present study) further supports this concept.

Unlike adenosine A 1 -receptor knockout mice, in which TGF was absolutely absent ( 3, 32, 36 ), cd73 knockout mice still have some TGF response. The TGF activity in cd73 +/+ mice, as well as the preserved TGF response in cd73 -/- mice, was completely inhibited by the selective adenosine A 1 -receptor blocker DPCPX, suggesting that adenosine may also mediate the preserved TGF activity in the absence of cd73. DPCPX at micromolar concentrations given in the late proximal tubular fluid has been shown before to inhibit the normal TGF response and adenosine receptor activation-induced afferent arteriolar vasoconstriction in rats ( 31, 37 ). Although the possibility that DPCPX at this concentration may mask the direct effects of this agent on other receptors in afferent arterioles could not be absolutely excluded, our observation implied that adenosine A 1 -receptor activation is one important component for a full TGF response in the presence or absence of cd73 in mice. Together, these data indicate that ecto-5'-nucleotidase (cd73)-dependent and -independent generation of adenosine participates in the mediation of TGF in vivo. Thus, in addition to extracellular formation, adenosine may be directly generated and released from macula densa cells. Intracellular generation of adenosine includes 1 ) the classic sequential dephosphorylation of intracellular ATP to adenosine, and 2 ) the hydrolysis of S -adenosyl- L -homocysteine to L -homocysteine and adenosine (for review, see Refs. 14, 15 ). These intracellular adenosine pathways may contribute to the preserved TGF activity found in the cd73 -/-.

In accordance with previous reports ( 5, 17 ), cd73 -/- mice have normal blood pressure and plasma Na + and K + concentrations. There was no difference between genotypes with regard to urinary excretion of fluid, Na +, and K +. It was observed previously that acute and local application of 5'-nucleotidase blocker, -methylene adenosine diphosphate can reduce proximal fluid reabsorption in anesthetized rats ( 33 ). The present study revealed that reabsorption of fluid, Na +, and K + in proximal tubules was not altered in cd73 -/- mice, possibly indicating efficient compensatory mechanism in the gene-disrupted animals. Likewise, fractional reabsorption of Na + and K + to the early distal tubule as well as to the urine was not affected in mice lacking cd73, arguing against an indispensable role of cd73 in the control of Na + and K + transport along the nephron. The observed modest increase in fractional fluid reabsorption up to the early distal tubule in cd73 -/- mice was possibly a consequence of a lower SNGFR in this group and the imperfect nature of glomerulo-tubular balance, i.e., changes in SNGFR cause modest inverse changes in fractional reabsorption ( 34 ).

Whole kidney GFR as well as SNGFR measured from both proximal and distal tubule were modestly but significantly lower in cd73 -/- than cd73 +/+ mice. Kidney weight was likewise modestly lower in the absence of cd73, such that GFR related to kidney weight was normal. We did not find disarranged renal structures or disturbed renal morphology in kidneys of cd73 -/- mice, and thus the reduced kidney weights could, at least in part, be the result of a primary reduction in GFR. The mechanisms contributing to lower GFR in cd73 -/- mice remain unclear. The findings indicate, however, that ecto-5'-nucleotidase (cd73)-dependent regulation of GFR is not restricted to its local role in TGF, because an attenuated TGF-dependent suppression of GFR itself (as observed in cd73 -/- mice) would not be expected to result in lower, but rather higher, GFR. The previously published study on cd73 -/- mice reported normal renal blood flow, whereas GFR or SNGFR were not determined ( 5 ). Notably, renal expression of cd73 protein includes glomerular mesengial cells, as well as smooth muscle cells of both afferent and efferent glomerular arterioles ( 19 ). Considering further that local levels of ATP may be elevated in cd73 -/- mice due to reduced conversion of adenine nucleotides to adenosine, it seems possible that accumulating ATP, a known renal vasoconstrictor ( 13, 14 ), could contribute to the observed lower GFR in the absence of cd73. In contrast to adenosine A 1 -receptor blockade, luminal application of the P2x-receptor blocker NF279 did not significantly alter SNGFR responses to increasing late proximal tubular flow rate in cd73 +/+ or cd73 -/- mice. However, minor differences, too small to be detected in a statistically significant way, cannot be excluded, and, especially in cd73 -/- mice, NF279 tended to lower the fall in SNGFR. This tendency may reflect a minor contribution of P2X receptors in TGF under these particular conditions or reflect TGF-independent increases in SNGFR in response to blocking P2X receptors on mesangial cells or at the afferent arteriole, which are stimulated by elevated local ATP concentrations. Further studies in cd73 -/- mice are required to assess renal ATP and adenosine metabolism, the renal expression of the respective purinergic receptors, and the regulation of afferent and efferent arteriolar resistances.

In conclusion, our results support the notion that extracellular formation of adenosine by ecto-5'-nucleotidase (cd73) is dispensable for normal fluid, Na +, or K + reabsorption along the nephron, but contributes to the regulation of GFR. Adenosine being generated by both ecto-5'-nucleotidase (cd73)-dependent and -independent mechanisms participates via activation of adenosine A 1 receptors in the mediation of TGF in vivo.

GRANTS

The presented work was supported by the Federal Ministry of Education and Research, the Interdisciplinary Clinical Research Center (IZKF) Tübingen, the Department of Veterans Affairs, and National Institutes of Health (Grants DK-56248, DK-28602, and GM-66232).

【参考文献】
  Bell PD, Lapointe JY, Sabirov R, Hayashi S, Peti-Peterdi J, Manabe K, Kovacs G, and Okada Y. Macula densa cell signaling involves ATP release through a maxi anion channel. Proc Natl Acad Sci USA 100: 4322-4327, 2003.

Braun N, Zhu Y, Krieglstein J, Culmsee C, and Zimmermann H. Upregulation of the enzyme chain hydrolyzing extracellular ATP after transient forebrain ischemia in the rat. J Neurosci 18: 4891-4900, 1998.

Brown R, Ollerstam A, Johansson B, Skott O, Gebre-Medhin S, Fredholm B, and Persson AE. Abolished tubuloglomerular feedback and increased plasma renin in adenosine A 1 receptor-deficient mice. Am J Physiol Regul Integr Comp Physiol 281: R1362-R1367, 2001.

Bucheimer RE and Linden J. Purinergic regulation of epithelial transport. J Physiol 555: 311-321, 2004.

Castrop H, Huang Y, Hashimoto S, Mizel D, Hansen P, Theilig F, Bachmann S, Deng C, Briggs J, and Schnermann J. Impairment of tubuloglomerular feedback regulation of GFR in ecto-5'-nucleotidase/CD73-deficient mice. J Clin Invest 114: 634-642, 2004.

Damer S, Niebel B, Czeche S, Nickel P, Ardanuy U, Schmalzing G, Rettinger J, Mutschler E, and Lambrecht G. NF279: a novel and selective antagonist of P2X receptor-mediated responses. Eur J Pharmacol 350: R5-R6, 1998.

Dyrberg T, Baekkeskov S, and Lernmark A. Specific pancreatic beta-cell surface antigens recognized by a xenogenic antiserum. J Cell Biol 94: 472-477, 1982.

Hansen PB and Schnermann J. Vasoconstrictor and vasodilator effects of adenosine in the kidney. Am J Physiol Renal Physiol 285: F590-F599, 2003.

Huang DY, Wulff P, Volkl H, Loffing J, Richter K, Kuhl D, Lang F, and Vallon V. Impaired regulation of renal K + elimination in the sgk1-knockout mouse. J Am Soc Nephrol 15: 885-891, 2004.

Huang DY, Osswald H, and Vallon V. Eukaliuric diuresis and natriuresis in response to the KATP channel blocker U37883A: micropuncture studies on the tubular site of action. Br J Pharmacol 127: 1811-1818, 1999.

Huang DY, Osswald H, and Vallon V. Intratubular application of sodium azide inhibits loop of Henle reabsorption and tubuloglomerular feedback response in anesthetized rats. Naunyn Schmiedebergs Arch Pharmacol 358: 367-373, 1998.

Inscho EW, Cook AK, Imig JD, Vial C, and Evans RJ. Renal autoregulation in P2X1 knockout mice. Acta Physiol Scand 181: 445-453, 2004.

Inscho EW, Cook AK, Imig JD, Vial C, and Evans RJ. Physiological role for P2X1 receptors in renal microvascular autoregulatory behavior. J Clin Invest 112: 1895-1905, 2003.

Jackson EK and Dubey RK. The extracellular cyclic AMP-adenosine pathway in renal physiology. Annu Rev Physiol 66: 571-599, 2004.

Jackson EK and Dubey RK. Role of the extracellular cAMP-adenosine pathway in renal physiology. Am J Physiol Renal Physiol 281: F597-F612, 2001.

Klapperstuck M, Buttner C, Nickel P, Schmalzing G, Lambrecht G, and Markwardt F. Antagonism by the suramin analogue NF279 on human P2X 1 and P2X 7 receptors. Eur J Pharmacol 387: 245-252, 2000.

Koszalka P, Ozuyaman B, Huo Y, Zernecke A, Flogel U, Braun N, Buchheiser A, Decking UK, Smith ML, Sevigny J, Gear A, Weber AA, Molojavyi A, Ding Z, Weber C, Ley K, Zimmermann H, Godecke A, and Schrader J. Targeted disruption of cd73/ecto-5'-nucleotidase alters thromboregulation and augments vascular inflammatory response. Circ Res 95: 814-821, 2004.

Lambrecht G, Damer S, Niebel B, Czeche S, Nikel P, Rettinger J, Schmalzing G, and Mutschler E. Novel ligands for P2 receptor subtypes in innervated tissues. Prog Brain Res 120: 107-117, 1999.

Le Hir M and Kaissling B. Distribution and regulation of renal ecto-5'-nucleotidase: implications for physiological functions of adenosine. Am J Physiol Renal Fluid Electrolyte Physiol 264: F377-F387, 1993.

Lohse MJ, Klotz KN, Lindenborn-Fotinos J, Reddington M, Schwabe U, and Olsson RA. 8-Cyclopentyl-1,3-dipropylxanthine (DPCPX)-a selective high affinity antagonist radioligand for A1 adenosine receptors. Naunyn Schmiedebergs Arch Pharmacol 336: 204-210, 1987.

Marxer-Meier A, Hegyi I, Loffing J, and Kaissling B. Postnatal maturation of renal cortical peritubular fibroblasts in the rat. Anat Embryol (Berl) 197: 143-153, 1998.

Nishiyama A and Navar LG. ATP mediates tubuloglomerular feedback. Am J Physiol Regul Integr Comp Physiol 283: R273-R275, 2002.

Osswald H, Muhlbauer B, and Schenk F. Adenosine mediates tubuloglomerular feedback response: an element of metabolic control of kidney function. Kidney Int 32: S128-S131, 1991.

Osswald H, Nabakowski G, and Hermes H. Adenosine as a possible mediator of metabolic control of glomerular filtration rate. Int J Biochem 12: 263-267, 1980.

Osswald H, Muhlbauer B, and Vallon V. Adenosine and tubuloglomerular feedback. Blood Purif 15: 243-252, 1997.

Ren YL, Garvin JL, Liu R, and Carretero OA. Role of macula densa adenosine triphosphate (ATP) in tubuloglomerular feedback. Kidney Int 66: 1479-1485, 2004.

Rettinger J, Schmalzing G, Damer S, Muller G, Nickel P, and Lambrecht G. The suramin analogue NF279 is a novel and potent antagonist selective for the P2X 1 receptor. Neuropharmacology 39: 2044-2053, 2000.

Rieg T, Steigele H, Schnermann J, Richter K, Osswald H, and Vallon V. Requirement of intact adenosine A 1 receptors for the diuretic and natriuretic action of the methylxanthines theophylline and caffeine. J Pharmacol Exp Ther 313: 403-409, 2005.

Savic V, Blanchard A, Vlahovic P, Stefanovic V, Ardaillou N, and Ardaillou R. Cyclic adenosine monophosphate-stimulating agents induce ecto-5'-nucleotidase activity and inhibit DNA synthesis in rat cultured mesangial cells. Arch Biochem Biophys 290: 202-206, 1991.

Schnermann J. Adenosine mediates tubuloglomerular feedback. Am J Physiol Regul Integr Comp Physiol 283: R276-R277, 2002.

Schnermann J, Weihprecht H, and Briggs JP. Inhibition of tubuloglomerular feedback during adenosine A1 receptor blockade. Am J Physiol Renal Fluid Electrolyte Physiol 258: F553-F561, 1990.

Sun D, Samuelson LC, Yang T, Huang Y, Paliege A, Saunders T, Briggs J, and Schermann J. Mediation of tubuloglomerular feedback by adenosine: evidence from mice lacking adenosine 1 receptors. Proc Natl Acad Sci USA 98: 9983-9988, 2001.

Thomson S, Bao D, Deng A, and Vallon V. Adenosine formed by 5'-nucleotidase mediates tubuloglomerular feedback. J Clin Invest 106: 289-298, 2000.

Thomson SC, Deng A, Bao D, Satriano J, Blantz RC, and Vallon V. Ornithine decarboxylase, kidney size, and the tubular hypothesis of glomerular hyperfiltration in experimental diabetes. J Clin Invest 107: 217-224, 2001.

Vallon V. Tubuloglomerular feedback and the control of glomerular filtration rate. News Physiol Sci 18: 169-174, 2003.

Vallon V, Richter K, Huang DY, Rieg T, and Schnermann J. Functional consequences at the single-nephron level of the lack of adenosine A1 receptors and tubuloglomerular feedback in mice. Pflügers Arch 448: 214-21, 2004.

Weihprecht H, Lorenz JN, Briggs JP, and Schnermann J. Synergistic effects of angiotensin and adenosine in the renal microvasculature. Am J Physiol Renal Fluid Electrolyte Physiol 266: F227-F239, 1994.

Wulff P, Vallon V, Huang DY, Volkl H, Yu F, Richter K, Jansen M, Schlunz M, Klingel K, Loffing J, Kauselmann G, Bosl MR, Lang F, and Kuhl D. Impaired renal Na(+) retention in the sgk1-knockout mouse. J Clin Invest 110: 1263-1268, 2002.

Zimmermann H. 5'-Nucleotidase: molecular structure and functional aspects. Biochem J 285: 345-365, 1992.

作者单位：1 Institute of Pharmacology and Toxicology, University of Tübingen, Tübingen, Germany; 2 Departments of Medicine and Pharmacology, University of California San Diego and Veterans Affairs Medical Center San Diego, San Diego, California; 3 Institute of Zoology, University of Frankfurt, Frank