Abstract
We recently demonstrated that cAMP added to the perfusate increased the renal venous recovery of adenosine in the isolated rat kidney, an effect blocked by inhibition of ecto-phosphodiesterase and ecto-5′-nucleotidase. Although our previous study established the cAMP-adenosine pathway, i.e., the conversion of cAMP to adenosine, as a viable metabolic pathway within the kidney, that study did not determine whether conversion of arterial cAMP to adenosine recoverable in the venous effluent occurred in the tubulesversus nontubular sites. In the current study, we addressed this issue by determining the effects of blocking cAMP transport into the renal tubules with probenecid (0.1, 0.3 and 1 mM) on the increase in renal venous output of adenosine induced by adding cAMP (30 μM) to the perfusate of isolated rat kidneys. Addition of cAMP to the perfusate caused a marked increase in renal venous secretion of adenosine, an effect that was augmented, rather than inhibited, by probenecid. To test the hypothesis that the renal vasculature supports a cAMP-adenosine pathway, cultured rat preglomerular vascular smooth muscle cells were incubated with cAMP (30 μM) for 1 hr in the presence and absence of 3-isobutyl-1-methylxanthine (a phosphodiesterase inhibitor). Incubation with cAMP increased extracellular adenosine levels 41-fold, and this effect was abolished by 3-isobutyl-1-methylxanthine. In a third experimental series, addition of cAMP (0.3, 1, 3, 10 and 30 μM) to the perfusate of isolated rat kidneys and mesenteric vascular beds increased the renal venous, but not mesenteric venous, output of AMP, adenosine and inosine. We conclude that the renal vasculature supports a cAMP-adenosine pathway, that administering cAMP into the renal artery and measuring adenosine in the venous effluent of the perfused rat kidney most likely monitors primarily the renal vascular cAMP-adenosine pathway and that the quantitative importance of the cAMP-adenosine pathway is not equivalent in all vascular compartments.
Renal adenosine participates importantly in the regulation of renin release, renal hemodynamics, tubuloglomerular feedback, erythropoietin production and tubular transport (Jackson, 1997). It is important, therefore, to determine the mechanisms of adenosine formation in the kidney. One such mechanism appears to be the cAMP-adenosine pathway (Jackson, 1991). In this regard, hormonal activation of adenylyl cyclase is linked to egress of intracellular cAMP via an active transport mechanism (Barber and Butcher, 1983). The cAMP-adenosine pathway hypothesis posits that ecto-phosphodiesterase metabolizes cAMP, as it egresses from the cell, to AMP, which in turn is converted to adenosine by ecto-5′-nucleotidase. This highly compartmentalized formation of adenosine would then function in an autocrine/paracrine manner to modify and/or expand the initial hormonal stimulus.
The existence of a cAMP-adenosine pathway in the kidneys is supported by two lines of evidence. First, infusion of IBMX (a phosphodiesterase inhibitor) into the renal cortical interstitium via a microdialysis probe decreases renal cortical interstitial levels of adenosine and inosine (a metabolite of adenosine) (Mi et al., 1994). This study suggests that a pathway involving phosphodiesterase(s) accounts for a significant portion, perhaps 50%, of renal adenosine formation. Second, administration of exogenous cAMP to the isolated perfused rat kidney increases the renal secretion rates (as measured in the venous outflow) of both AMP and adenosine, with the increase in AMP being greater than the increase in adenosine (Mi and Jackson, 1995). Moreover, this increase in adenosine secretion induced by exogenous cAMP is inhibited by blockade of total phosphodiesterase, ecto-phosphodiesterase and ecto-5′-nucleotidase (Mi and Jackson, 1995).
The above-mentioned studies in the isolated perfused rat kidney strongly suggest that in the kidney cAMP is converted to adenosine extracellularly. However, those studies do not determine whether conversion of perfusate cAMP to adenosine recoverable in the venous effluent occurs mostly in the tubules versus nontubular sites such as the renal vasculature. Several studies demonstrate that cAMP is efficiently transported by the probenecid-inhibitable organic anion transport system in the proximal tubule (Coulson and Bowman, 1974; Coulson et al., 1974; Gogel et al., 1983;Ullrich et al., 1991). Consequently, administration of cAMP to the kidneys via the renal artery would deliver cAMP to the tubular lumen where it could be converted to adenosine. Back-diffusion of adenosine into the peritubular capillaries could then account for adenosine in the venous effluent.
The main goal of the present study was to determine whether adenosine recoverable in the renal venous outflow during cAMP administered into the renal artery is formed by metabolism of cAMP to adenosine predominantly in the tubules versus nontubular sites. The experimental strategy used to address this issue was to measure the amount of adenosine appearing in the renal venous effluent during intrarenal artery infusions of cAMP in the absence and presence of probenecid, a blocker of the organic anion transport system in the proximal tubule. Our findings support the conclusion that adenosine measured in the renal venous outflow during intrarenal cAMP infusions is not formed predominantly in the renal tubules. In additional studies in cultured preglomerular vascular smooth muscle cells, we established the existence of a renovascular cAMP-adenosine pathway which most likely contributes significantly to the renal venous outflow of adenosine during cAMP administration to the kidney. Finally, the relative importance of the cAMP-adenosine pathway in the kidneyversus the mesenteric vascular bed was determined by comparing the effects of intra-arterial administration of cAMP on venous adenosine levels in the perfused rat kidney versusthe perfused rat mesenteric vascular bed.
Methods
Animals.
Male Sprague-Dawley rats (Charles River; Wilmington, MA) weighing 341 to 584 g (mean, 445 g) were used in these studies. Animals were housed at the University of Pittsburgh Animal Facility and fed Prolab RMH 3000 (PMI Feeds, Inc., St. Louis, MO) containing 0.26% sodium and 0.82% potassium.
Probenecid studies in the perfused rat kidney.
Each rat was anesthetized with sodium pentobarbital (45 mg/kg i.p.), and a midline incision was made. The left kidney, left renal artery, abdominal aorta and left ureter were dissected free from surrounding tissue, the left ureter was cannulated (PE-10 tubing), and the abdominal aorta below the left kidney was cannulated (PE-50 tubing). The suprarenal aorta was ligated, and the left kidney was immediately flushed (2.5 ml/min) with oxygenated Tyrode’s solution containing 100 U/ml of heparin. Without interrupting perfusion, the left kidney was isolated and placed in a water-jacketed organ chamber that was kept at 37°C with a thermostatically controlled water circulator (Thermocirculator, Harvard Apparatus, South Natick, MA) that circulated warm water through the jacket of the chamber. Kidneys were perfused (nonrecirculating) at 5 ml/min by a Harvard model 1210 peristaltic pump with a Tyrode’s solution (composition in mM: NaCl, 137; KCl, 2.7; CaCl2, 1.8; MgCl2, 1.1; NaHCO3, 12; NaH2PO4, 0.42;d(+)-glucose, 5.6). The perfusate was gassed with 95% O2 and 5% CO2 and was pumped through a warming coil (37°C) that was fitted with a bubble trap. Perfusion pressure was monitored with a Statham pressure transducer (model P23ID, Statham Division, Gould Inc., Oxnard, CA) connected to an access port located above the kidney on the perfusion cannula and was displayed on a Grass model 79D polygraph (Grass Instruments, Quincy, MA).
The kidneys were first allowed to stabilize for 1 hr before beginning the protocol. The protocol consisted of four cAMP treatment periods, each 5 min in duration, in which 30 μM cAMP was added directly to the Tyrode’s solution. Each 5-min cAMP treatment period was separated by 15 min. One minute before adding the cAMP and from 4 to 5 min into the cAMP treatment, perfusate exiting the renal vein was collected on ice and frozen at −40°C for later analysis of purines. In six perfused kidneys, beginning 10 min before the second, third and fourth cAMP treatment periods, probenecid (100, 300 and 1000 μM, respectively) was added to the perfusate. Coulson and Bowman (1974) have demonstrated previously that 900 μM probenecid abolishes tubular transport of cAMP in the perfused rat kidney. We avoided higher concentrations of probenecid because of concern that higher concentrations would inhibit phosphodiesterase activity (Podevin et al., 1980). In four perfused kidneys, probenecid was not added to the perfusate to assess whether purine secretion was stable during the four cAMP treatment periods.
The cAMP-induced change in the renal venous secretion rate of AMP, adenosine or inosine was calculated by subtracting the secretion rate just before infusing cAMP from the secretion rate during the infusion of cAMP. These secretion rates were compared across the four levels of probenecid concentrations (0, 100, 300 and 1000 μM) with a repeated measures one-factor analysis of variance followed by a Fisher’s Least Significant Difference test if the analysis of variance indicated significant differences among the means. Statistical analysis was performed with the Number Cruncher Statistical System (Kaysville, UT), and all values in the text, figures and table refer to means ± S.E.M.
Studies comparing cAMP metabolism in the perfused rat kidneyversus the perfused rat mesenteric vascular bed.
Rat kidneys (n = 5) were isolated and perfused as described above. Rat mesenteric vascular beds (n = 5) were isolated by the technique of McGregor (1965) and as described previously (Campbell and Jackson, 1979; Jackson and Campbell, 1980). Mesenteric vascular beds were perfused in the same organ perfusion system and under the same conditions (composition of perfusate, perfusion rate, temperature, etc.) as described above for the perfused rat kidney. After a 1-hr equilibration period, a 1-min sample of venous effluent was collected. Next, increasing concentrations of cAMP (0.3, 1, 3, 10 and 30 μM) were added to the perfusate at 10-min intervals, and a 1-min sample of venous effluent was collected just before the end of each treatment level of cAMP. Venous effluent samples were collected on ice and frozen at −40°C for later analysis of purines. Renal and mesenteric venous secretion rates of AMP, adenosine and inosine were calculated, and the effects of cAMP on secretion rates of each purine in each tissue were determined with a repeated measures one-factor analysis of variance followed by a Fisher’s Least Significant Difference test if the analysis of variance indicated significant differences among the means.
Studies with cultured preglomerular vascular smooth muscle cells.
Preglomerular (primarily interlobular and afferent arteriolar) vascular smooth muscle cells were cultured from explants of preglomerular microvessels by the method of Dubey et al.(1992). Wistar-Kyoto rats (Taconic Farms, Germantown, NY) were anesthetized with pentobarbital, and the aorta was exposedvia a midline incision. A PE-190 catheter was placed in the abdominal aorta below the orifice of the left renal artery, and the mesenteric artery was ligated. The aorta above the orifice of the right renal artery was ligated, the renal veins were severed and the kidneys were flushed clear of blood with phosphate-buffered saline. A suspension (1% wt/vol; 5 ml) of iron oxide particles (Aldrich Chemical Co., Milwaukee, WI) was flushed into the renal arteries. The iron oxide particles were suspended in culture medium I (Dulbecco’s Modified Eagle’s Medium [GIBCO Laboratories, Grand Island, NY] containing penicillin, 100 U/ml; streptomycin, 100 μg/ml; amphotericin B, 100 U/ml; polymyxin B, 50 μg/ml; and HEPES, 25 mM). The kidneys were rapidly removed from the rat, put in culture medium I (4°C) and decapsulated. All steps from this point in the procedure until cell culture were conducted at 4°C, with the exception of the incubation with collagenase (see below). Cortical tissue was obtained and placed in 10 ml of culture medium I. After mincing the cortical tissue, the tissue was dispersed by passing it through a sterile wire screen (stainless steel, 30 mesh, Small Parts, Inc., Miami, FL). The cortical suspension was placed in sterile tubes and diluted to 30 ml with culture medium I. Next, a magnet was placed against the tube thus securing the iron-laden microvessels in the tube while decanting the remaining tissue. The microvessels were washed five times in 15 ml of culture medium I and placed in 25 ml of culture medium I containing 0.6 mg/ml of collagenase type IV (Sigma Chemical Co., St. Louis, MO). After incubating with shaking for approximately 30 min at 37°C, the microvessels were again separated from the remaining tissue with a magnet. The microvessels were then suspended in 10 ml of culture medium II (Dulbecco’s Modified Eagle’s Medium F-12 containing penicillin, 100 U/ml; streptomycin, 100 μg/ml; amphotericin B, 100 U/ml; polymyxin B, 50 μg/ml; HEPES, 25 mM; NaHCO3, 13 mM; and 20% fetal calf serum [Hyclone Laboratories Inc., Logan, Utah]).
Three-milliliter portions of the final suspension were placed in 75-cm2 culture flasks and were incubated in 5% CO2/95% air at 37°C and 98% humidity. Approximately 48 hr later, an additional 2 ml of culture medium II was added to each flask. After 6 days the medium was replaced with fresh culture medium II, and the medium was changed approximately every 48 hr until the cells became confluent (approximately 2 weeks).
Although the vascular smooth muscle cells did not appear to be contaminated with fibroblasts, the method of Aviv et al.(1983) was used to remove any occult fibroblast contamination. Two milliliters of calcium- and magnesium-free Hanks’ balanced salt solution (GIBCO) containing 0.25% trypsin and 1 mM ethylenediaminetetraacetic acid was placed in each flask. Thirty seconds later, most of the medium in each flask was aspirated, leaving a film of trypsin-containing medium behind. Next, 10 ml of culture medium II was added, and the cells were lifted from the flask bottom by gentle swirling. A magnet was applied to the flask bottom, and the detached cells were aspirated and divided between two other culture flasks (5 ml each), leaving the iron-laden microvessels in the original flask. Twenty minutes later, the culture medium from the second flask was aspirated and placed in a third flask, and this procedure was repeated two more times. Because fibroblasts attach rapidly to culture surfaces, whereas vascular smooth muscle cells do not, this method allowed effective separation of vascular smooth muscle cells from fibroblasts. In approximately 2 weeks, the passaged preglomerular vascular smooth muscle cells (in culture medium II) were confluent.
On the day of the study, 24 wells containing confluent cells in second passage were washed twice with 1 ml of phosphate-buffered saline and incubated for 60 min with 0.5 ml of phosphate-buffered saline containing either no additions (n = 6) or cAMP (30 μM), IBMX (300 μM) or cAMP + IBMX (n = 6 for each treatment). After the 60-min incubation period, the medium was aspirated and frozen at −40°C for latter analysis of purines.
The concentration of adenosine in the medium was normalized to cell number, which was determined by use of a Coulter counter. These values were compared among the four experimental groups with a one-factor analysis of variance followed by a Fisher’s Least Significant Difference test.
Sample analysis.
Purines in samples of perfusate were analyzed with an Isco (Lincoln, NE) high-pressure liquid chromatographic system (pump model 2350, gradient programmer model 2360, 4.6 × 250 mm C18 column with 5-μm particle size, ChemResearch Data Management System) with UV detection as described previously (Mi and Jackson, 1995). Adenosine in the culture media was analyzed with the same chromatographic system but with fluorescence detection as described previously (Jackson et al., 1996).
Results
As shown in table 1, in time-control experiments, i.e., no probenecid in the perfusate, cAMP (30 μM) added to the perfusate increased the renal secretion of AMP, adenosine and inosine. The cAMP-induced increases in purine secretion were reproducible across all four experimental periods. Regardless of concentration, addition of probenecid to the perfusate did not reduce the cAMP-induced changes in secretion rates of AMP, adenosine or inosine (fig. 1). The highest dose of probenecid (1 mM) significantly increased the secretion rate of adenosine and tended to increase the secretion rate of inosine (fig.1).
Addition of cAMP (30 μM) to cultured preglomerular vascular smooth muscle cells increased extracellular levels of adenosine approximately 41-fold (fig. 2), and this effect was abolished by the phosphodiesterase inhibitor, IBMX.
Increasing concentrations of cAMP added to the perfusate of isolated kidneys caused increasing renal secretion rates of AMP, adenosine and inosine (fig. 3). In contrast, mesenteric vascular beds, perfused under conditions identical with those used for the kidneys, did not convert cAMP to AMP, adenosine or inosine (fig.3).
Discussion
In a previous study (Mi and Jackson, 1995) we reported that administration of exogenous cAMP to the perfused rat kidney increases the renal venous secretion rate of adenosine. The intrarenal conversion of cAMP to adenosine was blocked by inhibition of phosphodiesterase with IBMX, by inhibition of ecto-phosphodiesterase with 1,3-dipropyl-8-p-sulfophenylxanthine and by inhibition of ecto-5′-nucleotidase with α,β-methyleneadenosine-5′-diphosphate. These studies indicated that the kidney supports a cAMP-adenosine pathway that involves the conversion of cAMP to AMP by ecto-phosphodiesterase and the conversion of AMP to adenosine by ecto-5′-nucleotidase.
Our previous study, however, did not allow inferences regarding whether the adenosine measured in the venous effluent during intrarenal administration of cAMP is derived from cAMP converted to adenosine predominantly in the renal tubules versus nontubular sites. Although ecto-5′-nucleotidase is expressed on the surface of mesangial cells (Stefanovic et al., 1989), endothelial cells (Gordonet al., 1986) and vascular smooth muscle cells (Gordonet al., 1989), it is most abundant on proximal tubule brush border and on peritubular fibroblasts in the cortical labyrinth (Le Hir and Kaissling, 1993). Moreover, renal arterial cAMP is secreted efficiently into the proximal tubule by the organic anion transport mechanism (Coulson and Bowman, 1974; Coulson et al., 1974;Gogel et al., 1983; Ullrich et al., 1991). Indeed, Coulson and Bowman (1974) have demonstrated that cAMP extraction by the isolated perfused rat kidney is mediated mainly by secretion of cAMP into renal cells via a probenecid-sensitive secretory mechanism. Therefore, it is possible that cAMP when infused into the renal artery is transported into the tubules and converted to AMP and adenosine by tubular elements, and this newly formed tubular adenosine then diffuses back into the vascular compartment to appear in the renal venous effluent. The main purpose of the present study was to determine whether the tubular cAMP-adenosine pathway accounts for the increase in renal venous purine levels during intrarenal artery administration of cAMP.
Probenecid is the prototypical inhibitor of the proximal tubular organic anion transport system (Cunningham et al., 1981;Moller and Sheikh, 1982), and effectively blocks the transport of renal arterial cAMP into the proximal tubule (Coulson and Bowman, 1974;Coulson et al., 1974; Gogel et al., 1983; Ullrichet al., 1991). If the adenosine appearing in the renal venous effluent during intrarenal administration of cAMP is formed predominantly in the tubules, then inhibition of cAMP transport into proximal tubules with probenecid would diminish the renal venous recovery of adenosine during intrarenal cAMP infusions because less of the administered cAMP would be available to the tubules for conversion to adenosine. Conversely, if the adenosine appearing in the renal venous effluent during intrarenal administration of cAMP is formed predominantly in the nontubular compartment, then probenecid would increase renal venous recovery of adenosine during intrarenal cAMP infusions. This is because probenecid would reduce the secretion of cAMP into the tubular compartment and thereby would increase the amount of cAMP available for conversion to adenosine in the nontubular compartment. As demonstrated in the present study, high concentrations of probenecid significantly increase the appearance of adenosine in the renal venous effluent, which suggests that adenosine measured in the renal venous outflow during intrarenal cAMP infusions is formed predominantly in a nontubular (presumably vascular) compartment.
It is possible that cAMP administered into the renal artery diffuses into the interstitial space, is converted to adenosine by cells residing in the interstitial compartment and then diffuses back into the vascular lumen to be measured in the venous outflow. Although this possibility certainly cannot be ruled out, given the low permeability of cell membranes to cAMP and the high rate of uptake of adenosine by nearly all cell types, it seems unlikely that this pathway contributes importantly to renal venous adenosine during cAMP administration. Thus, cAMP would have to negotiate several diffusion barriers and adenosine would have to escape multiple uptake sinks for this pathway to account for the renal venous adenosine measured during intrarenal administration of cAMP. Moreover, in the present study we also demonstrate that cultured preglomerular vascular smooth muscle cells convert exogenous cAMP to adenosine by a pathway that is abolished by inhibition of phosphodiesterase with IBMX. By directly establishing the existence of the cAMP-adenosine pathway in vascular smooth muscle cells obtained from renal microvessels, this observation strengthens the inference that measurement of renal venous adenosine during intrarenal administration of cAMP assesses primarily the renal vascular cAMP-adenosine pathway.
The present study demonstrates the existence of a cAMP-adenosine pathway in the renal vasculature, and a cAMP-adenosine pathway also exists in cultured aortic vascular smooth muscle cells (Dubey et al., 1996). However, to what extent one can generalize these findings to all vascular smooth muscle cells/vascular beds is unknown. In the present study, mesenteric vascular beds that were perfused under conditions identical with those used for the kidney experiments did not convert cAMP to adenosine. Therefore, either the vascular cAMP-adenosine pathway is restricted in distribution in vivoor the pathway requires some unknown but necessary condition(s) that is (are) absent from the perfused mesenteric vascular bed.
We would like to emphasize that the present study in no way rules out the existence of a tubular cAMP-adenosine pathway. Rather, the present study suggests that the vascular cAMP-adenosine pathway is the one primarily monitored when adenosine is measured in the renal venous outflow from perfused rat kidneys during intrarenal cAMP administration. It is possible, indeed likely, that the renal tubules also convert cAMP to adenosine, and this hypothesis needs to be addressed by additional experiments. Indeed, we have found that intrarenal infusions of cAMP into the intact rat kidney in vivo greatly increases urinary adenosine levels (work in progress). Thus, a cAMP-adenosine pathway may exist in several renal compartments including the tubules.
In summary, this study establishes the existence of a renal vascular cAMP-adenosine pathway and indicates that measurement of renal venous adenosine during intrarenal administration of cAMP most likely monitors the renovascular cAMP-adenosine pathway. Moreover, this study demonstrates that under the conditions of in vitroperfusion, the renovascular cAMP-adenosine pathway is much more efficient than the mesenteric vascular cAMP-adenosine pathway. The present study does not, however, exclude the possibility of a tubular cAMP-adenosine pathway.
Footnotes
-
Send reprint requests to: Edwin K. Jackson, Ph.D., Center for Clinical Pharmacology, University of Pittsburgh Medical Center, 623 Scaife Hall, 200 Lothrop Street, Pittsburgh, PA 15213-2582.
-
↵1 This work was supported by National Institutes of Health grants HL40319, HL35909 and HL55314.
- Abbreviations:
- cAMP
- adenosine 3′,5′-monophosphate
- AMP
- adenosine 5′-monophosphate
- IBMX
- 3-isobutyl-1-methylxanthine
- PE
- polyethylene
- HEPES
- N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid
- Received March 14, 1997.
- Accepted June 30, 1997.
- The American Society for Pharmacology and Experimental Therapeutics