Regulation of Kidney Function and Metabolism: A Question of Supply and Demand

Abstract

Kidney blood flow and glomerular filtration rate (GFR) are maintained relatively constant by hormonal influences and by efficient autoregulation. However, the kidney remains at risk for ischemia and acute kidney injury. Increases in kidney blood flow cause parallel increments in GFR, thereby dictating tubular reabsorption and increased oxygen/metabolic demands. Coordination between kidney blood flow and GFR with tubular reabsorption is maintained by the tubuloglomerular feedback (TGF) system whereby delivery of NaCl to the macula densa varies inversely with nephron GFR. Metabolic products, ATP and adenosine, are the mediators of TGF via afferent arteriolar vasoconstriction, and nitric oxide; COX-2 products and angiotensin II are modulators of acute TGF responses and temporal adaptation of TGF. Oxygen requirements and metabolic efficiency of Na transport in the kidney are significant variables that are regulated by both mediators and modulators of TGF. These metabolic and hormonal substances efficiently regulate both kidney supply and demand.

Introduction

The kidney must be normally viewed as the champion among organs of the body when it comes to stability of organ blood flow and, in the case of the organ itself, stability of glomerular filtration rate. This regulation is normally attributed to a) highly efficient autoregulation of kidney blood flow in response to normal variations in systemic blood pressure (1), b) a multiplicity of complex interacting humoral vasoconstrictor and vasodilator influences (2,3) and, c) tubuloglomerular feedback system, an intrinsic system which interrelates tubular reabsorption and the rate of glomerular filtration via regulation of kidney vascular resistances (4–6). In spite of this plethora of regulatory devices, the kidney remains at risk of ischemia and persistent high likelihood of clinical acute kidney injury (AKI) or acute renal failure. This clinical observation raises further questions regarding the adequacy of regulation of oxygen and substrate delivery and the utilization of oxygen/substrate or the metabolic demands of the kidney (7,8).

These apparent conflicting observations can be resolved when certain facts are considered. Glomerular filtration rate (GFR) is highly plasma flow dependent, changing more or less in parallel in normal conditions (3,9). This linkage means that maintenance or increases in renal blood flow (RBF) not only satisfy the supply of oxygen and substrate but also dictates maintenance or increases in GFR, which is largely reabsorbed by kidney tubules, a process which requires ATP, oxygen and substrate. This makes the relationship highly complex, since restoration or increases in kidney blood flow not only augment metabolic supply but also increase metabolic demands required for increased tubular reabsorption (7). Normally, the arterio-venous oxygen difference in the kidney is small, implying a significant oxygen and substrate reserve for prevention of ischemia. However, the presence of a preglomerular countercurrent oxygen diffusion shunt produces a kidney environment which is relatively hypoxic (10); the pO₂ in the kidney cortex is normally low at 40–45 mmHg (11) and progressively even lower in the medulla and papilla (8,12). In certain experimental conditions, diabetes and hypertension, the pO₂ in the kidney cortex may be even lower (11,13). Therefore, any consideration of the risk of kidney ischemia must take into account both oxygen and substrate delivery and adequacy of kidney blood flow and the glomerular filtration rate, but also the metabolic demands for tubular reabsorption.

In the kidney there is a unique linkage between a) blood flow and GFR (3,9) and, b) tubular reabsorption provided in major part by the tubuloglomerular feedback (TGF) system and by glomerulo-tubular balance (5,6). The TGF relationship dictates a negative feedback system which predicts that increases in delivery and reabsorption at the macula densa (MD) dictate a reduction in the filtered load via reductions in kidney blood flow (Figure 1). The mediators of TGF, ATP and adenosine, a product of AMP, are metabolically linked (5,14–16), and the modulators of TGF activity, nitric oxide (NO) from nitric oxide synthase-1 (NOS-1) (17–19), angiotensin II (AII) (20) and, potentially, cyclooxygenase-2 (COX-2) products (21) also exert major metabolic impacts (Figure 2). Prior investigations have shown a linear relationship between kidney oxygen consumption (Q_O2) and tubular reabsorption, (22,23) and, therefore, GFR, with a positive y intercept which defines the metabolic demands of the non-filtering kidney. However, more recent observations have suggested that this relationship between kidney Na reabsorption and oxygen consumption is highly variable, affected, in major part, by hormonal influences which also regulate kidney blood flow and TGF (11,12,19,21,24–27). Mediators and modulators of TGF function are either critical metabolic byproducts or regulators of metabolism (5). Therefore, kidney ischemia is dictated by regulation of both metabolic supply and demand, and the regulatory factors may be similar hormonal/metabolic systems, which greatly simplifies these biologic processes.

Fig. 1.

Schematic operation of the tubuloglomerular feedback system (TGF). Increases in delivery and reabsorption of NaCl at the macula densa cell elicits vasoconstriction of the afferent arteriole resulting in a reduction of renal blood flow and the glomerular hydrostatic pressure as well as reductions in the glomerular ultrafiltration coefficient, K_f. The net effect of this TGF response is to reduce glomerular filtration rate. This reduction constitutes a negative feedback system which returns tubular fluid delivery and reabsorption of NaCl at the macula densa to control levels. The net effect of this TGF system is to stabilize late proximal flow rate, augment autoregulation of renal blood flow (RBF) and prevent major urinary losses of NaCl and water by overwhelming distal tubular reabsorptive capacities.

Fig. 2.

The signal transduction mechanism for transmission of the macula densa tubuloglomerular feedback signal to the afferent arteriole and glomerulus. Increased delivery of fluid to the macula densa promotes increased NaCl reabsorption. NaCl reabsorption causes the immediate release of ATP, which is degraded extracellularly to AMP and, specifically, further degraded by ecto-5′-nucleotidase (e-5′-NT) and to adenosine, a substance which is a vasoconstrictor at the afferent arteriole. Increased NaCl reabsorption in the macula densa cell also causes generation of nitric oxide and may release prostaglandins via COX-2. In addition to acting as primary vasodilators which counteract the effects of adenosine and ATP, NO also inhibits e-5′-NT thereby reducing the generation of adenosine. NaCl transport at the macula densa also modifies the generation of angiotensin II (AII) via renin. NO and COX-2 are also modulators of AII generation. ATP and adenosine mediators of the acute tubuloglomerular feedback response, and NO, prostaglandins from COX-2, and AII are further modulators of the tubuloglomerular feedback response and contribute to resetting or adaptation of tubuloglomerular feedback.

Methods

Studies were performed in Wistar and Wistar Frömter rats, the latter with surface glomeruli and raised in a colony at the San Diego VA Animal Research Facility. These studies were conducted in conformance with the local VA/UCSD IACUC protocol (A3659-01) dated November 5, 2005.

General micropuncture.

Under terminal Inactin anesthesia, the kidney is placed in a cup and covered with saline solution, and the configuration of distal tubules is identified by inserting a micropipette (1–3 μm tip) into urinary space of surface glomeruli to inject small boluses of stained artificial tubular fluid (ATF). A pipette is inserted into the last proximal or first distal accessible tubular loop to perform timed (2–3 min) collection of tubular fluid under free-flow conditions, using a short mineral oil block (28). Tubular fluid collections are analyzed for tubular flow rate and Na⁺, K⁺, and Cl⁻ concentrations as well as ³H-inulin concentration to determine single nephron glomerular filtration rate (SNGFR). Absolute and fractional reabsorption of fluid or Na⁺, K⁺, and Cl⁻ between the glomerulus and the late proximal or early distal tubule are calculated as previously described (28,29).

Analysis of proximal reabsorption.

When proximal reabsorption (Jprox) changes due to a change in SNGFR, this is normal glomerulotubular balance (GTB). When Jprox changes due to a change in the avidity of the tubule, this constitutes a “primary” change in tubular reabsorption. To test for primary changes in tubular reabsorption it is necessary to control for the influence of GTB (29,30). We accomplished this by using TGF as a tool for manipulating SNGFR. SNGFR will be manipulated by perfusing the nephron downstream from a late proximal wax block. SNGFR and Jprox will be determined by late proximal collections during perfusion at 8 and 40 nl/min and Jprox expressed as a function of SNGFR.

Micropuncture analytic methods.

Late proximal flow (V_LP) and early distal tubular flow rate (V_ED) are determined after transferring the tubular fluid collections to a constant bore capillary. We also examined changes in electrical conductivity of the early distal nephron fluid (T_ED) measured continuously as a legitimate surrogate for the ionic content of fluid flowing past the macula densa (MD). T_ED is determined from the electrical conductance across the tip of a micropipette (3–5 μM diameter filled with 1.2 M NaCl). Tubular fluid is allowed into the tip by gentle suction and pipette is wired to an electrical circuit such that the rat and pipette in series constitute a segment of a voltage and prior divider (31). Measurements of [Na⁺] and [K⁺] are expressed as a fraction of the conductance of artificial proximal tubular fluid and determined as described previously (29), using a micro flame photometer. [Cl⁻] is measured in samples with a micro adaptation of the electrometric titration method (Microtitre ET-1; World Precision Instruments, Sarasota, FL) (29). Concentrations of ³H-inulin in plasma, tubular collections, and urine are measured by liquid phase scintillation counting.

Glomerular hemodynamics.

In certain micropuncture studies, glomerular hemodynamics were evaluated by measuring glomerular capillary, urinary space, and efferent arteriolar hydrostatic pressure (32,33). In conjunction with efferent arteriolar protein samples (approximately three per period) and SNGFR, we can compute nephron plasma flow, the ultrafiltration coefficient (LpA) and effective filtration pressure.

Micropuncture analysis of tubuloglomerular feedback and glomerulotubular balance Conceptual framework for TGF.

Briefly, the TGF system operates as a negative feedback loop in which two parameters (SNGFR and V_LP) are linked by TGF and glomerulotubular balance (GTB). In the first type of experiment, SNGFR and V_LP are measured during orthograde perfusion of Henle's loop in wax-blocked nephrons. Measurements are made in each nephron during high and low rates of loop perfusion. These data establish a range for the TGF response. The second type of experiment involves measuring the fractional compensation for flow perturbations in free-flowing nephrons in order to reveal the behavior of TGF near the natural operating point. Data from these two sets of experiments are combined to calculate the slope of the TGF curve to generate complete TGF curves, and to predict the impact which a primary change in tubular reabsorption will have on SNGFR by altering the TGF stimulus.

TGF range and operating point SNGFR.

SNGFRmax refers to SNGFR during perfusion of Henle's loop at 8 nl/min. SNGFRmin refers to SNGFR during perfusion at 40 nl/min. SNGFRmax minus SNGFRmin equals the range. SNGFR at the TGF operating point is determined by collection from the early distal tubule and is referred to as SNGFRd. SNGFR at the inflection point of the sigmoid TGF curve is simply the average of SNGFRmax and SNGFRmin or SNGFRmid. As an index of TGF activation, SNGFRd is compared with SNGFRmid. SNGFRd-SNGFRmid will increase when conditions external to the juxtaglomerular apparatus (JGA) influence SNGFRd to increase or when conditions within the JGA influence TGF to reset downward.

Measurement of TGF efficiency in free-flowing nephrons.

Ambient proximal flow and the fractional compensation for perturbations in ambient flow are measured in free-flowing late proximal nephrons (6,34–36). Tubular flow is measured by a non-invasive optical technique videometric flow velocitometry (VMFV). A Hample nanoliter pump is employed to perturb flow. Tubular flow, V_M, is measured just upstream from the pump pipette. Perturbations, V_H, are applied by adding or subtracting fluid in the following sequence: V_H(nl/min) = 0, −4, +4, −8, +8, and 0. Each perturbation is for three minutes and data are collected over the last 60 seconds. V_M is expressed as a function of V_H by polynomial curve fitting or linear regression and fractional compensation at the operating point is calculated from C = (−dV_M/dV_H) for V_H = 0.

Videometric flow velocitometry (VMFV).

Small boluses of artificial tubular fluid containing rhodamine B dextran as a fluorescent marker are injected into the proximal tubule by a pneumatic microinjection pump. The dye is excited by a laser reflected onto the kidney surface. The image of the kidney surface is filtered to maximize resolution of emitted fluorescence and monitored with a video microscope. The intensity of the video image is monitored at two points downstream along the nephron. The time required for fluid to traverse the two points is calculated by cross-correlating the video densities. Geometry of the tubular segment is measured from digitized images. Tubular fluid flow rate, V_M, is calculated from the cross-correlations and the geometry of the tubule.

Enzyme assays for 5′-NT.

A sieving technique is used to separate the glomeruli from the tubules by gently pressing the cortex through a 106 μm stainless steel sieve. The passed through suspension is forced through an 18 gauge needle to decapsulate the glomeruli and then passed through a 75 μm sieve. The glomeruli are trapped on the top while the tubules pass through. 20 μl of each sample are added to a tube containing 20 μM Dipyridamole, 10 μM EHNA, and either buffer or 50 μM, aβ-methylene adenosine diphosphate (MADP). MADP is supposed to inhibit plasma membrane bound ecto- and endo-5′-NT but not AMP-specific cytosolic 5′-NT. This mixture is incubated for 10 minutes at 37°C in a shaking water bath. ¹⁴C AMP and 10 μM AMP or IMP were added and the incubation continued for 15 minutes. 5 μl of 17 N Formic acid is added to stop the reaction. The reaction tubes are spun down and a portion of each reaction supernate is separated by thin layer chromatography on cellulose sheets with fluorescent indicator in water: methanol (1:1). The reaction product (adenosine) and substrate (cold AMP) are located by nonradioactive markers. The spots are marked, cut out and quantitated by liquid scintillation counting. Plasma membrane bound ecto- and endo-5′-NT activity represents the MADP-inhibitable phosphatase activity (37).

Whole kidney clearance and O₂ consumption studies.

The rats are prepared for surgery as previously described (6,35), and animals anesthetized with Inactin (100 mg/kg i.p.). Rats are maintained with systemic infusion of 1.5 ml/hr of isotonic NaCl/NaHCO₃ containing appropriate amounts of ³H inulin 5 μCi/hr for clearance studies and 60–80 μCi/hr for micropuncture evaluation of single nephron filtration rate (SNGFR) (6,14). After anesthesia, rats are placed on a thermostatically controlled surgical table to maintain body temperature at 37°C. After surgery, the left renal artery is then separated from the renal vein through a midline incision plus a flank extension. The left kidney blood flow (ml/min) is monitored with a perivascular ultrasonic transit time flow probe (Transonics T206, Ithaca, NY, USA) connected to a computer for continuous recording (26,27). A 23 gauge needle is bent at 90° and inserted into the proximal left renal vein for sampling of venous blood (19). GFR (ml/min) is measured by clearance of ³H-inulin or calculated from RBF, hematocrit (HCT) and filtration fraction (FF) where FF is calculated according to the Fick principle from ³H-inulin content of arterial (R_A) and renal venous (R_V) plasma: FF = 1 − (R_V/R_A). Oxygen consumption is computed from A-V difference in O₂ multiplied by RBF. The cost of sodium transport (Q_O2/T_Na) is the ratio of total amount of renal Q_O2 over the total amount of sodium reabsorbed (T_Na). T_Na is equal to the total amount of Na filtered minus the amount of Na excreted. In certain studies we examined GFR and renal plasma flow (PAH clearance) in awake, chronically catheterized rats as previously described.

In vitro assessment of proximal tubules.

Isolation and purification of rat renal proximal tubules are performed according to previously published methods (26,27,38) with some modifications. The rat kidneys are first flushed with 40 ml cold oxygenated perfusion buffer containing (in mM, pH 7.1) NaCl: 112, NaHCO₃: 20, KCl: 5, CaCl₂: 1.6, Na₂HPO₄: 2, MgSO₄: 1.2, glucose: 5, HEPES: 10, mannitol: 10, glutamine: 1, sodium butyrate: 1, and sodium lactate: 1 and then perfused with 30 ml warm collagenase containing buffer (2 mg/ml in perfusion buffer, Collagenase Type II, Worthington) for in situ digestion within 10 min. The dissected and chopped cortex were submitted to another 5 min enzyme digestion. After straining through a sieve (pore size 125 μm), the tubules are washed and purified by 50% Percoll centrifugation. Oxygen consumption (Q_O2, pmole/min/μg protein) is measured polarographically in a 0.6 ml chamber (Instech Laboratories, Inc.) with a water jacket maintained at 37°C by using a Clark-type oxygen electrode and an YSI model 5300 oxymeter (Yellow Spring Instrument). The reaction chamber is filled with chamber buffer and a baseline recorded. Then 0.1 ml of the tubular suspension is added to the chamber for Q_O2 measurements. Once a stable recording (basal Q_O2) is achieved various reagents were added into the chamber via a port on the upright side of the chamber using an extra long tip to allow the deposition of the drugs near the bottom of the chamber with continuous tracing recording. Q_O2 is calculated assuming that the concentration of O₂ in the buffer under saturating conditions at 37°C is 544 nmol/ml as described in a recent publication (27).

Radioimmunoassay of Ang II.

Plasma is extracted using a Bondelut C18 column (Varian, Harbor City, CA) previously washed with methanol and a triethylamine and formic acid buffer. The Ang II is eluted with acetonitrile:triethylamine/formic acid (70/30), lyophilized on a Speed-Vac (Savant, Farmingdale, NY) overnight, and kept at −20°C until assayed. This plasma extraction yielded 98.5% recovery of Ang II. Details of RIA are provided in recent publications (20,39). In the case of nanoliter volumes (100–200 nl) of tubular fluid, no extraction is required and the sample is directly analyzed as described (20). Ang II concentrations were calculated using a computer-aided logit/log transformation of the standard curve. Cross-reactivity of this Ang II antibody with angiotensin I is 0.33% and with Ang III is 69%.

Results and Discussion

The kidney exhibits extraordinary efficiency of blood flow autoregulation and a multiplicity of overlapping neurohormonal factors that combine to regulate kidney blood flow and glomerular ultrafiltration (1–3,5,6), but in spite of these attributes, acute kidney injury or acute renal failure remains a major clinical problem. Such a finding may imply a breakdown in the balance of supply and demand of oxygen and substrates (7). Experimental findings would suggest that the tubuloglomerular feedback system normally provides exquisite coordination between a) kidney blood flow and load of glomerular filtrate and, b) tubular reabsorption (4–6,14). The TGF system adapts over time using complex hormonal modulators (17–21). The mediators and modulators of temporal adaptation of TGF are either substances critical to the metabolic processes or regulators of metabolic rate (7,24–26).

Tubuloglomerular feedback system: The mediators and the modulators

Increases in flow rate to the macula densa (MD) and NaCl reabsorption at the MD elicits reduction in the filtered load to that nephron unit (4–6), mediated primarily by constriction of the afferent arteriole (40,41). This TGF effect tends to stabilize late proximal tubular flow rate, prevents overload of distal tubular reabsorptive capacities and contributes to autoregulation (Figure 1). Such a linkage between tubular reabsorption and filtered load had been demonstrated repeatedly both in vitro and in vivo, the latter via alterations in the MD flow at the single nephron level and with whole kidney studies in which proximal tubular reabsorption is inhibited with application of agents such as carbonic anhydrase inhibitors (17,19,21,27,42,43). The mediators of this acute TGF response have been investigated extensively, both in vitro, in the perfused MD, glomerulus and afferent arteriolar preparation, and in vivo using micropuncture and pharmacologic blockade. Concurrent with increases in MD NaCl delivery and reabsorption is the release of ATP producing afferent arteriolar vasoconstriction via purinergic receptors from both in vitro and in vivo studies (44). However, ATP is rapidly metabolized extracellularly to AMP by nucleotide triphosphate diphosphohydrolases to AMP. This compound is then converted to adenosine by abundant kidney ecto-5′-nucleotidase enzyme in glomeruli, mesangial, and tubular cells (37,45) (Figure 2). Adenosine can then produce afferent arteriolar constriction via the adenosine A-1 receptor. Clearly, ATP release is required and may contribute to acute TGF responses, but evidence has accumulated that adenosine may maintain TGF elicited vasoconstrictor responses over a longer time period (Figure 2). In vivo micropuncture experiments utilizing adenosine “clamp” constraints, in which adenosine levels cannot change, completely eliminate TGF responses (14). In support of the viewpoint that adenosine is a critical mediator are the findings that mice lacking ecto-5′-nucleotidase exhibit greatly diminished acute TGF responses (44,45) and mice minus the adenosine A1 receptor also demonstrate nearly absent TGF responses (15,46). What is critical to this discourse on ATP versus adenosine as primary mediators is the fact that the candidate substances are major products of normal metabolism, suggesting that TGF function is linked to the kidney metabolic requirements for NaCl reabsorption.

TGF is a negative feedback system which exhibits a rapid response to maintain homeostatic efficiency for the process and preserves stability of tubular flow rates. Acute TGF responses are accompanied by generation and release of NO, primarily derived from NOS-1, located primarily within the MD but also other renal cells (47–49). NO, a vasodilator, has been shown to ameliorate the intensity of the acute TGF response. However, the TGF system cannot be truly relevant to the control of kidney function if it remains static in spite of changes in physiologic conditions. In order to maintain homeostatic efficiency, the TGF system must be capable of “resetting” or temporally adapting to new physiologic conditions. The nature of this TGF temporal adaptation has been examined in vivo after inhibition of proximal reabsorption over periods from one to 24 hours (17,19,21,34,35) and in the in vitro perfused MD/glomerulus/afferent arteriolar (AA) preparation (36). In normal conditions, acute TGF responses persist for up to 30-minutes but then nephron blood flow and AA diameter return to normal control values, permitting a readjustment of the relationship between MD NaCl delivery and filtered load or afferent arteriolar constriction, and consequent restoration of homeostatic efficiency of TGF and the capacity to respond to further perturbations. In practical clinical terms, this would imply that tubular injury should not lead to permanent kidney vasoconstriction or pre-renal conditions, but that kidney hemodynamic responses should be transient and temporally adapt (2,17). In vitro, TGF induces afferent constriction followed within 30 minutes by a full restoration of AA diameter (36). Studies in vivo have addressed the factors or mediators contributing to temporal adaptation of TGF. We have shown earlier that application of benzolamide, a carbonic anhydrase inhibitor which reduces proximal tubular reabsorption by ∼50% and activates TGF for a period of 18–24 hours, produces a rebound increase in kidney GFR (17) after withdrawal of this agent, generating temporal adaptation of TGF or resetting of the relationship of MD NaCl delivery and GFR. Application of agents which inhibit kidney NOS-1 prevented this hyperfiltration response and, in fact, caused a sustained reduction in GFR which is characteristic of the acute TGF response to benzolamide. More acute studies demonstrated temporal adaptation of kidney blood flow after benzolamide within 30–60 minutes, and this return of RBF to control values was prevented by concurrent application of SMTC, a selective NOS-1 inhibitor (19). In a pattern identical to NOS-1 inhibition, concurrent administration of COX-2 inhibitors also completely prevented temporal adaptation (21). It is of interest to speculate that this physiologic effect of COX-2 inhibition to prevent normal temporal adaptation of TGF may contribute to the NaCl retention and the hypertension frequently observed with administration of COX-2 inhibitors.

The most physiologically relevant example of normal temporal adaptation is the transition between low and high NaCl intakes. This adaptation requires an adjustment of the relationship between MD NaCl delivery and kidney GFR. In spite of this adjustment and contrary to earlier speculations and publications, we found in nephron micropuncture studies that TGF maintained homeostatic efficiency and normal acute TGF responses while rats were maintained on high NaCl intakes (20). Reports have suggested that kidney adenosine content is actually increased on high NaCl intake (50,51). We have developed assays for ecto-5′-nucleotidase (ecto-5′-NT) and found that glomerular enzyme activity increases in parallel with NaCl intake in rats. We also observed that NO donors significantly inhibit glomerular enzyme activity, which might provide a further mechanism whereby NO, derived from NOS-1 could cause temporal adaptation of TGF by decreasing adenosine formation (36).

AII is also a modulator of TGF function whereby increased AII has been shown to increase TGF responses. AII in physiologic concentrations also increases proximal tubular reabsorption (33). Multiple studies have shown that the traditional RAAS system is suppressed by high NaCl intakes. Studies in rats on high NaCl intakes from our laboratory also demonstrated lower plasma and kidney values for AII by RIA (20). However, when we evaluated AII levels in proximal tubular fluid samples we noted a 50% increase in AII in rats on high NaCl intake (8 mEq/day) in comparison with rats on normal NaCl intake (∼2 mEq/day). This increase in AII in the intrarenal microenvironment exerts unexpected physiologic impacts in that administration of AT-1 receptor blockers produced major reductions in proximal tubular reabsorption and significant inhibition of acute TGF responses (20). The net effect was significantly elevated late proximal tubular flow rates in high NaCl intake rats receiving blockers of AII activity. These unexpected results regarding changes in AII within the proximal tubule may help explain why patients receiving inhibitors of AII as therapy, but are noncompliant with dietary NaCl restriction, still appear to benefit from this therapy in terms of both natriuresis and blood pressure control. AII also functions as an important modulator of TGF by both influencing the magnitude of TGF during temporal adaptation and the stability of proximal reabsorption and tubular flow rates.

The metabolic costs of kidney tubular reabsorption

Oxygen consumption by the kidney has been noted to relate directly and predominantly to Na reabsorption (22,23). Recent studies have supplied evidence that the slope of this relationship is variable among physiologic and pathophysiologic conditions and affected by hormonal factors (11–13). Given the finding of temporal adaptation of TGF (17), one might postulate that during this period there are metabolic adaptations and changes in tubular reabsorption that permit or justify “resetting” of the relationship between GFR and MD NaCl delivery/reabsorptive rate. NOS-1 is a major modulator of TGF and is critical to temporal adaptation of TGF (17,19). NO exerts several chemical interactions which influence oxidative metabolism. NO suppresses the citric acid cycle enzyme, aconitase, and inhibits mitochondrial pyruvate uptake, but the major “braking” effect of NO on oxidative metabolism is accomplished via inhibition of cytochrome c oxidase (52–56). Although effects are more complex than we will describe in detail, NO can inhibit mitochondrial respiration in vitro by up to 85%. Therefore it is possible that NO generated during temporal adaptation of TGF subsumes functions which reduce epithelial metabolism or increase metabolic efficiency of tubular reabsorption during this period of TGF adaptation. There is evidence that epithelial NOS-1 and possibly NO generated within mitochondria constitutively act as a “metabolic brake” on tubular oxygen consumption (25,26).

Earlier studies have observed a reduction in the metabolic efficiency of kidney Na reabsorption as well as cortical pO₂ in the spontaneously hypertensive rat (SHR) (11,57). The authors attributed the increased Q_O2/T_Na in SHR to a deficiency of NO generation. Laycock also demonstrated in the dog that application of non-selective NOS inhibitors produced major increases in kidney Q_O2 (24). Koivisto et al., also demonstrated an effect of NO on oxygen consumption in isolated kidney proximal tubules (25). We have recently demonstrated a major increase in kidney Q_O2 and Q_O2/T_Na after NOS inhibition, and that the NO which influences Q_O2 is entirely derived from NOS-1, an isoform which may be located within mitochondria (Figure 3) (26). The effects of NOS inhibition were not influenced by blockade of AII activity via the AT-1 receptor. There were also major increases in oxygen consumption in freshly harvested proximal tubules after SMTC, an NOS-1 inhibitor, studied while in motion in a specially designed metabolic chamber (Figure 3). These collective results raise the possibility that NO derived from NOS-1 may function as a “metabolic brake” in kidney epithelial cells. The increase in NOS-1 activity after TGF activation and temporal adaptation or resetting of TGF may further amplify these influences acting to suppress oxygen consumption and enhance metabolic efficiency of Na reabsorption, thereby influencing both metabolic “supply” and “demand”.

Fig. 3.

Effects of NO and NOS-1 blockers on oxygen consumption in freshly harvested isolated proximal tubules. Panel A demonstrates the profile of declining O2 % in a metabolic chamber containing proximal tubules. When the NOS-1 inhibitor, SMTC, is added, oxygen consumption increases, designated by the increase in slope of decline. When an NO donor, NONOate is applied, oxygen consumption decreases dramatically. In Panel B absolute oxygen consumption is depicted in control tubules and after SMTC, the NOS-1 blocker. The increase in oxygen consumption after NOS-1 blockade is totally reversed by application of the NO donor, suggesting NO specificity to the phenomenon. These data suggest a major role for intracellular NOS activity in regulation of oxygen consumption, probably at the mitochondrial level.

Alterations in selective kidney transport processes, particularly in the proximal tubule, can also change the metabolic efficiency of Na reabsorption. The metabolic cost of Na transport in the in vivo proximal tubule is less than the more distal nephron because of the large component of passive NaCl reabsorption set up by proton secretion and NaHCO3 reabsorption in the early S-1 segment (58,59). Several of the mediators and modulators of TGF also regulate components of kidney tubular reabsorption, particularly NHE₃ and NHE₂ by adenosine, AII and NO, respectively.

We have recently published surprising results on the in vivo and in vitro effects of benzolamide, a carbonic anhydrase inhibitor, which exerts its primary effects in the proximal tubule (43). This agent reduces proximal tubular reabsorption by at least 50%, yet we observed a 50% increase in kidney oxygen consumption (Q_O2) and Q_O2/T_Na increased by approximately 80% (27). We postulated that the resulting decrease in luminal pH which accompanies carbonic anhydrase inhibition somehow promotes compensating active NaCl reabsorption while passive NaCl reabsorption is essentially eliminated. This in vivo phenomenon was duplicated in freshly harvested proximal tubules in vitro since NaCl transport persists in this preparation. Of interest to this discussion, application of adenosine A-1 receptor antagonists and NHE₃ blockers of proton secretion completely normalized Q_O2 and Q_O2/T_Na. These results demonstrate again that selective alterations in transport mechanisms can markedly influence overall oxygen costs of kidney function, dominantly within the proximal tubule, by either eliminating passive or promoting active transport (27,60,61). Adenosine, AII, and NO have major influences on TGF function and adaptation and all exert major potential impacts on kidney oxygen consumption and metabolic efficiency of kidney function.

Kidney oxygen consumption in pathophysiologic states

There are several pathophysiologic conditions in which increased Q_O2/T_Na has been observed, including models of hypertension, experimental diabetes and models of chronic kidney disease, created by reduction in renal mass (the 5/6^th nephrectomy model) (11,13,57,62). Factoring Q_O2 by total Na reabsorption may not be a reasonable assessment of metabolic oxygen utilization by the kidney in pathophysiologic states, since one should also consider other uses of oxygen by a) chemical oxidation reactions, b) gluconeogenesis, c) inefficiencies of Na reabsorption and, d) reabsorption of other filtered molecules such as albumin and other “middle molecules” as a result of increased glomerular leak of these substances. Nevertheless such phenomena will increase in demand for oxygen in pathophysiologic conditions.

In experimental diabetes, Q_O2 and Q_O2/T_Na are increased in hyperglycemic rats (13,63). We and others have also observed increased kidney NO generation which derives primarily from NOS-1 (64). NOS-1 is derived from cells other than the macula densa, the TGF sensing element or afferent limb. Blockade of NOS-1 in the diabetic kidney produces major reduction in renal blood flow (RBF) and GFR in contrast with little effects in control, non-diabetic rats. In addition, we have observed a significantly greater increase in kidney Q_O2 after SMTC, a NOS-1 inhibitor, in diabetic rather than non-diabetic rats (65). Such findings suggest that Q_O2 and Q_O2/T_Na are increased in diabetic kidneys, for reasons yet to be explained, and that NOS-1 activity is increased possibly as a compensatory “braking” mechanism in an attempt to maintain metabolic efficiency in the diabetic kidney.

In models of chronic kidney disease, pathogenetic roles for both reactive oxygen species (ROS) and kidney hypoxia have been postulated as critical to progression of disease (62). We have examined kidney oxygen consumption in a model of chronic kidney disease, the 5/6 nephrectomy model. In this model at one week we observed a remarkable increase in both absolute oxygen consumption (Q_O2) and Q_O2/T_Na (66). The kidney exhibited significant proteinuria at this early stage. We have examined the effects of chronic application of Losartan, an angiotensin II AT-1 receptor blocker. Losartan nearly doubled GFR and RBF, implying a major increase in AII activity in this early phase of this model of CKD. In spite this change in RBF or “supply”, oxygen consumption remained constant and Losartan essentially normalized Q_O2/T_Na. In normal rats, we have found no effect of AII, AT-1 receptor blockade, on kidney oxygen consumption (26). In the 5/6 nephrectomy model, Losartan therapy also eliminated proteinuria. Proximal tubular reabsorption of albumin and other proteins requires expenditure of ATP for proteasomal degradation of reabsorbed proteins. These observations raise questions as to whether some of the appreciable benefits of inhibition of the renin-angiotensin system in CKD, particularly with significant proteinuria, could be mediated by altering the metabolic efficiency and increased oxygen consumption in chronically diseased kidneys (66).

A balance between glomerular filtration and tubular reabsorption must be maintained by the kidney. Tubuloglomerular feedback systems exert acute and temporally adapted control over the relationship of kidney blood flow and filtered NaCl and tubular reabsorption. The mediator and modulators of TGF have been well-defined (5,14,17). The kidney must also maintain a proper balance between metabolic supply and demand in order to prevent kidney hypoxia and ischemia. The metabolic costs of kidney reabsorption are also highly regulated such that the metabolic demands of epithelial cells become an important variable that contributes to the threshold for ischemia. We have observed that hormonal and metabolic products that regulate tubuloglomerular feedback systems play significant roles in the regulation of oxygen consumption in the kidney (26,27,65,66). (Figure 4). In fact, the critical operative substances tend to individually affect metabolic supply and demand in opposite directions, as noted with a) ATP/adenosine, b) NO from NOS-1, and, c) angiotensin II, and, d) possibly COX-2 products (Figure 4). Such regulatory efficiency makes sense from the standpoint of evolutionary biology in that regulatory substances should control both metabolic supply and demand.

Fig. 4.

The hormonal control of kidney oxygen supply and demand: Prevention of ischemia. Both oxygen and substrate supply and oxygen and substrate utilization or demand are controlled by a variety of hormonal and metabolic factors. There is a general pattern that substances which decrease supply tend to increase demand at the level of tubular epithelium, either by increasing reabsorption or increasing oxygen required for sodium transport. Such a relationship is logical from the standpoint of evolutionary biology in that metabolic and hormonal regulators will modify both supply and demand. The balance of these forces determines the threshold of kidney ischemia.

DISCUSSION

Mitch: Houston: Thank you. I was just wondering if you factored the increase in oxygen consumption with diabetes or even with the remnant kidney, which is growing rapidly in the proximal tubule. What is the contribution of protein synthesis and/or breakdown in those organs in those parts of the kidney?

Blantz: La Jolla: We factored it both by sodium transport, and as you have perceptively picked up, there are other functions that may be reabsorptive functions, such as protein absorption and things like that. Whether we factor it by sodium reabsorption or whether we factor it by protein content or even DNA, it is still elevated and there is quite a bit of variation from animal to animal, but the net average is high. As I remember, one of your ex-faculty, Harold Franch, did some elegant studies about protein synthesis in early diabetes. As I recall, protein synthesis actually was only amplified for the first few days after creating diabetes and the rest of the increase in protein was due to decreased degradation. In other words, there were lysosomal disposal problems of protein in diabetes that I think accounted for the extra protein. Right now, I'd say we don't know, and it is, I'm sure, of interest to the diabetologists as well as why the kidney as a peculiar organ seems to be inefficient in oxygen utilization. We've also noted, as Sharon Anderson at University of Oregon, the dominant NOS isoform in the diabetic kidney is actually brain NOS or NOS 1, which is not the case in normal kidney. Interestingly, the diabetic kidney does not allow for resetting or TGF adaptation. So, it already has a peculiar behavior when one changes salt intake as we've observed in other publications.

Luke: Cincinnati: I think you're suggesting a role, in that there have been a lot of papers recently saying that even a 0.3 or 0.4 increase in serum creatinine in the ICU after an MI or after an invasive procedure confers greater, or is associated with greater morbidity and mortality subsequently. Is the kidney just acting as a sensor for reactive oxygen species in that setting? I mean, you've just eluded to that.

Blantz: Well your statement is certainly true and we've got some studies that I didn't have a chance to mention that even cytokinemia or having more cytokines around, early sepsis before your kidney actually fails, the oxygen consumption by the kidney almost doubles, and this is not related to NO. We are trying to figure why in the world that occurs, but certainly, several of the acute kidney injury centers that are developing and sponsored by NIH are now interested in basically why this phenomenon of having a prior hit seems to set you up for progressive renal disease. I think that what we haven't taken serious consideration, is there a metabolic component that causes transformation or rather epithelial mesenchymal transformation of cells that derives from some kind of insult that led to a metabolic defect? All I am trying to say is that nephrologists, for 40 or 50 years, have been concerned about supply of blood to the kidney. I think now we ought to start thinking about what are the regulators that dictate the oxygen demands of the kidney, and are they factors that we can predict in particular problem patients as a consequence of the ICU setting.

Willerson: Houston: As a cardiologist, Roland, where is endothelin's role in the vasoregulatory cascade you've shown us; and Robin, I wondered where it was on your slides too?

Blantz: Well this wasn't a question of lack of loyalty. We just don't have a grant on endothelin. But endothelin is obviously one of the more potent vasoconstrictors, but it becomes a really complicated issue to study, and we haven't studied it yet because, as you well know, endothelin has a couple of different receptors, one of which liberates nitric oxide; and the question is: is all nitric oxide created equal? No, it isn't! It depends on which nitric oxide you are generating, and we also have some interest in what role the NMDA receptor has in the proximal tubule in regulating NO. It's hard to imagine that NO was primarily a metabolic regulator, but I think that nitric oxide and endothelin may have preceded the cardiovascular system, and that means preceded cardiologists, I presume. So, I don't know whether the biologic functions for having nitric oxide and endothelin might have derived from metabolic origins. Certainly, these substances, adenosine and NO, occurred before kidneys and cardiovascular systems.

Luke: Perhaps the cardiologists and nephrologists can continue this debate outside. The President's address is not allowed to be questioned, Dr. Willerson.