INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THERE IS ACCUMULATING evidence implicating the renal medullary microvasculature in the control of arterial pressure. In particular, the level of renal medullary blood flow (MBF) appears to be an important determinant of sodium and water reabsorption (4, 5) and is perhaps also important for the release of the putative renal medullary depressor hormone (2). Furthermore, although there is evidence to the contrary (16, 17), some studies have shown that renal MBF can be poorly autoregulated, at least under volume-expanded conditions (4, 5). The renal medullary microcirculation may therefore be well placed to transduce changes in arterial pressure into homeostatic responses that restore normal arterial pressure.

One technique that has been a useful tool for studying the role of the renal medullary microcirculation in the long-term control of arterial pressure has been infusion of vasoactive substances into the renal medulla. Cowley and colleagues (4-6, 13, 14, 19, 23) employed this technique in rats, combined with laser-Doppler flowmetry for evaluation of regional kidney blood flow (22). They showed that chronic medullary interstitial infusion of vasoconstrictor agents such as N^G-nitro-L-arginine methyl ester (19) and the vasopressin V₁-agonist [Phe²,Ile³,Orn⁸]vasopressin (23), at doses that reduce medullary but not cortical blood flow (CBF), results in the development of sustained hypertension. Conversely, medullary interstitial infusion of captopril in spontaneously hypertensive rats, which increases medullary but not CBF, ameliorates their hypertension (13).

In longitudinal studies such as those described above, there are considerable advantages to employing larger species. In the case of conscious rabbits, we are able to obtain long-term and simultaneous data regarding hormonal status (10), cardiac output (months; 9), renal blood flow (weeks; 25), and renal sympathetic nerve activity (weeks; 18). Therefore, in the present study we investigated methods for infusion of vasoactive agents into the renal medulla of rabbits. We chose norepinephrine (NE) as our test agent, because it is rapidly broken down in vivo, minimizing the confounding effects of spillover into the systemic circulation. We first tested whether outer medullary interstitial infusion of NE via acutely implanted needle catheters in anesthetized rabbits affected renal hemodynamics and function in the infused and contralateral kidney. We paid particular attention to the effects of the infusion on cortical and medullary perfusion, determined by laser-Doppler flowmetry. This was correlated with examination of the distribution of radiolabel after medullary interstitial infusion of [³H]NE and the removal of the radiolabel from the infusion site via the renal venous and urinary excretory routes. Subsequently, similar studies were performed 7-14 days after implantation of catheters designed for chronic implantation, in which we compared the effects, distribution, and disposition of NE/radiolabel during infusion of NE into the outer and inner medulla. These studies suggest that medullary interstitial infusion of vasoactive agents provides a useful method for selectively manipulating renal MBF in rabbits. However, the positioning of the catheter tip is critical for the distribution of the infused agent and the effects observed.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

Experimental Preparation

1

1

Both kidneys were denervated, and both ureters were cannulated. A transit-time ultrasound flow probe (type 2SB; Transonic Systems, Ithaca, NY) was placed around the renal artery, and the tips of three single-fiber laser-Doppler flow probes (0.5 mm diameter; University of Linköping, Sweden) were placed 0.5 (cortical), 0.5 (cortical), and 10 mm (medullary), respectively, below the cortical surface (see Ref. 8 for detailed description). For protocols 2 and 3 (see below), a branch of the ileolumbar vein was isolated and cannulated (polyvinyl chloride tubing, 0.8 mm ID, 1.2 mm OD; Critchley Electrical, Auburn, NSW, Australia), the tip of the cannula being advanced so that it lay in the renal vein for collection of renal venous blood. To maintain catheter patency, heparinized (50 IU/ml heparin sodium, Monoparin, Fisons) Hartmann's solution was infused at a rate of 2 ml/h.

Implantation of Medullary Interstitial Catheters

1

1

Chronically positioned catheters. These were implanted 7-14 days before the experiment under halothane (1-4%, Fluothane; ICI, Victoria, Australia) anesthesia and sterile conditions (see Ref. 25). A left flank incision was made, and the kidney was gently exteriorized. The tip of a single polyethylene catheter (ID 0.28 mm, OD 0.61 mm; Critchley Electrical) was introduced into the ventral side of the kidney, slightly rostral to the midline aspect, at an angle directed toward the renal pelvis. The catheter was then advanced so that its tip lay either 8.5 (n = 8) or 10.5 mm (n = 9) below the surface of the kidney. Correct insertion resulted in minimal bleeding, which stopped almost immediately. A small piece of nylon mesh (1.5 cm diameter; Hilton Hosiery, Coolaroo, Victoria, Australia) attached to the catheter was anchored to the surface of the kidney with cyanoacrylate glue (Loctite; Caringbah, NSW, Australia). The catheter was tunneled subcutaneously so that its end lay between the shoulder blades, and an osmotic mini pump (Alzet 2ML2; 5 µl/h for 14 days, Alza, Palo Alto, CA) filled with 154 mM saline was attached to the end of the catheter to maintain catheter patency.

In preliminary studies we found that the depth of 8.5 mm (outer medullary catheters) corresponds approximately with the junction of the inner and outer stripes of the outer medulla, whereas the depth of 10.5 mm placed the catheter in the white, inner medulla, but not in the papilla. When possible, gross postmortem examination of the kidneys was performed in this and other studies (Ref. 8, unpublished observations), and the catheters were always found to be correctly placed. Furthermore, no evidence of gross disruption of kidney tissue or scar tissue due to implantation of the medullary interstitial catheters was found from examination of the frozen sections taken for autoradiographic analysis (see below).

Experimental Protocols

1

1

1

1

1

1

1

1

Protocol 2. Acutely positioned catheters: disposition and distribution of radiolabel during intramedullary infusion of [³H]NE (9 rabbits). [³H]NE (16-24 nCi · kg¹ · min¹ in 100 ng · kg¹ · min¹ NE; one-half the dose via each of the 2 catheters) was infused into the medullary interstitium for 20 min. Urine produced by both kidneys was collected during the 2 min before the infusion commenced and for each 2-min period throughout the [³H]NE infusion. Ear arterial and renal venous blood samples (0.5 ml each) were collected at the midpoint of each urine collection period.

At the completion of the 20-min [³H]NE infusion, both the infused and contralateral kidneys were quickly retrieved, decapsulated, and halved coronally. For half of each kidney, portions of the cortex, medulla, and papilla were dissected and weighed in separate preweighed 20-ml vials, to which 50 µl of tissue solubilizer (NCS II, 0.5 N solution, Amersham, Buckinghamshire, UK) was added per milligram of wet tissue weight. Radioactivity in each region of the kidney was subsequently determined by liquid scintillation counting (see below). For five of these rabbits, the remaining kidney halves were frozen in liquid nitrogen and stored at 70°C for subsequent autoradiographic analysis (see below).

Protocol 3. Chronically positioned outer and inner medullary catheters (17 rabbits). First, increasing doses of NE (0, 30, 100, and 300 ng · kg¹ · min¹) were infused into the inner (n = 9) or the outer (n = 8) medulla via the catheters implanted 7-14 days previously. Each dose of NE was infused for 20 min. Second, after a 20- to 40-min recovery period, during which all variables returned to their baseline levels, NE (300 ng · kg¹ · min¹) was infused intravenously for 20 min. Finally, after a further 20-40 min was allowed for recovery from intravenous NE, [³H]NE (16-24 nCi · kg¹ · min¹ in 100 ng · kg¹ · min¹ NE) was infused via the chronically implanted medullary interstitial catheter for 20 min. Arterial and renal venous blood samples and urine samples were collected during the infusion, and the kidneys were harvested and processed at the completion of the infusion, as described for protocol 2.

Measurements

Renal function (protocol 1). Clearance measurements of glomerular filtration rate (GFR, ml/min) and effective renal plasma flow (which was corrected for hematocrit to provide effective renal blood flow; ERBF, ml/min) and determinations of urine and sodium excretion were made as previously described (8). At the completion of experiments in which renal clearance measurements were made (protocol 1), the kidneys were removed and desiccated, and the dry weight was determined. Therefore, for this protocol, RBF_probe, ERBF, GFR, urine flow rate, and urinary sodium excretion are expressed per gram of kidney dry weight (mean = 1.86 ± 0.07 g).

Disposition of radiolabel during infusion of [³H]NE (protocols 2 and 3). The concentrations of [³H]NE and its metabolites in each of the samples were measured by liquid scintillation counting and expressed in terms of the total dose of NE infused (100 ng · kg¹ · min¹). The rate of disposition of radiolabel by the infused and contralateral kidneys was then calculated as the concentration (above) multiplied by the flow rate (ml · kg¹ · min¹). Urine flow was determined gravimetrically. Renal venous blood flow was taken as renal arterial blood flow determined by the transit-time ultrasound flow probe and was not corrected for urine flow.

Analysis of solubilized kidney tissue (protocols 2 and 3). Once dissolved (after at least 7 days of incubation in tissue solubilizer at room temperature), triplicate 20-µl samples of each solubilized tissue sample were added to a water-soluble scintillation fluid (ACS, Aqueous counting fluid; Amersham) and subjected to liquid scintillation counting.

Autoradiography (protocols 2 and 3). Coronal 50-µm sections of the frozen left kidney were cut on a cryostat at 19°C and mounted on glass slides (subbed with 10% gelatin; BDH Chemicals, Poole, UK). These sections were left to dry for 1-2 h at room temperature. Subsequent to drying, slides were apposed to tritium-sensitive film (Amersham Hyperfilm) in the presence of tritium microscales (Amersham) for 6-8 wk. Developed autoradiograms were quantified using an MCID M4 image analysis system (Imaging Research) as previously described (1). Each kidney section was divided into four regions, cortex, outer stripe of the outer medulla (outer stripe), medulla (excluding the outer stripe and papilla), and the papilla (defined as the portion of the inner medulla that protrudes into the renal pelvis), for separate quantification.

We chose to use both methods (analysis of solubilized kidney tissue and autoradiography) because both have inherent advantages and disadvantages and provide complementary data. For example, the results from counting solubilized tissue are potentially subject to variation due to the dissection, the degree of solubilization, and extrapolation errors from counting aliquots. On the other hand, whereas the autoradiograms provide greater sensitivity and anatomic resolution, the data they provide are representative only of the relatively thin (50 µm) sections taken.

Statistical Analyses

Protocol 1. The data were analyzed as the average level of each variable over each 20-min clearance period, because, at least for the renal clearance measurements, this was the maximum level of resolution. To test for an effect of medullary interstitial infusion of saline, an ANOVA was partitioned to contrast the levels of each variable during the 20-min saline infusion (the second experimental period of the first experimental run) with the 20-min periods before and after the saline infusion. To test for effects of medullary interstitial infusion of NE on systemic and renal hemodynamic variables, we used the interaction term of time and treatment (vehicle or NE). For renal clearance variables, the interaction term of time and kidney (infused and contralateral) was used to control the confounding effects of changes in arterial pressure on renal function. To test for differences between the responses of CBF and MBF during NE infusion, the main effect of kidney region (cortex and medulla) tested whether, across all three doses of NE, changes on MBF were greater than those in CBF.

Protocol 2. The levels of radioactivity in the various kidney regions were subjected to ANOVA, the factors comprising rabbit, side (infused or contralateral kidney), and kidney region.

Protocol 3. ANOVAs were partitioned to test the specific hypotheses that medullary interstitial infusion of NE dose relatedly influenced the levels of systemic and renal hemodynamic variables and that intravenous infusion of NE influenced these variables. To test for differences between outer and inner medullary interstitial [³H]NE infusion in terms of the levels of radioactivity entering the kidney via the renal artery and exiting the kidney via the urine and the renal vein, the interaction term between catheter position (outer or inner medulla) and time was used. Data concerning the levels of radiolabel in the infused and contralateral kidneys were analyzed as for protocol 2.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Protocol 1. Acutely Positioned Outer Medullary Interstitial Catheters: Effects of Medullary Interstitial Infusion of Saline and NE on Systemic and Renal Hemodynamics and Renal Excretory Function

Medullary interstitial infusion of NE. Medullary interstitial infusion of NE (30, 100, and 300 ng · kg¹ · min¹) was accompanied by dose-related increases in MAP (by 3 ± 4, 10 ± 2, and 16 ± 4% of resting, respectively) and small but statistically significant reductions in HR by (1 ± 1, 3 ± 1, and 3 ± 1% of resting, respectively). In the infused (left) kidney, RBF_probe was dose relatedly reduced by (8 ± 1, 16 ± 3, and 30 ± 4% of resting, respectively), as was CBF (by 6 ± 4, 7 ± 1, and 19 ± 4% of resting, respectively) and MBF (by 17 ± 5, 37 ± 11, and 45 ± 9% of resting, respectively). The reductions in MBF were significantly greater than those of CBF (P < 0.001) (Fig. 1), consistent with the notion that tissue levels of NE are greater in the medulla and/or inner cortex than in the outer cortex during medullary interstitial infusion. In later protocols, this hypothesis was more directly tested by analysis of the intrarenal distribution of radiolabel during medullary interstitial infusion of [³H]NE.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Effects of medullary interstitial infusion of norepinephrine (0, 30, 100, and 300 ng · kg1 · min1) or its vehicle (154 mM NaCl; 20 µl · kg1 · min1) on systemic and renal hemodynamic variables. MAP, mean arterial pressure; HR, heart rate; RBFprobe, renal blood flow determined by transit-time ultrasound flow probe; CBF, renal cortical perfusion (laser-Doppler flux signal); MBF, renal medullary perfusion. Columns and error bars represent means ± SE of data from 8 rabbits (except for CBF and MBF, where n = 6 because laser-Doppler flowmetry was not employed in 2 rabbits), for average levels over each 20-min experimental period. P values represent outcomes of interaction terms (treatment × dose) from repeated-measures ANOVA (degrees of freedom 3, 30-42) testing for nonparallelism in responses to vehicle and norepinephrine.

During medullary interstitial infusion of saline for four consecutive 20-min periods, the patterns of the responses of ERBF, GFR, urine flow, and sodium excretion were indistinguishable in the infused (left) and contralateral (right) kidneys. In contrast, during medullary interstitial infusion of NE, ERBF, GFR, and urine flow were reduced in the infused kidney relative to the contralateral kidney. A similar pattern of responses was observed for urinary sodium excretion, although this was not statistically significant (P = 0.09) (Fig. 2). Neither medullary interstitial infusion of saline nor NE appeared to affect the fractional excretion of urine (urine flow/GFR) or sodium (sodium clearance/GFR) (P always  0.3; data not shown).

View larger version (29K): [in this window] [in a new window]   Fig. 2.   Effects of medullary interstitial infusion of norepinephrine (0, 30, 100, and 300 ng · kg1 · min1; right) or its vehicle (NaCl 154 mM; 20 µl · kg1 · min1; left) on renal clearance variables. ERBF, effective renal blood flow determined by [14C]para-aminohippurate clearance; GFR, glomerular filtration rate determined by [3H]inulin clearance. Columns and error bars represent means ± SE of data from 8 rabbits. P values represent outcomes of repeated-measures ANOVA testing for nonparallelism in responses of left (infused, solid bars) and right (contralateral, open bars) kidneys.

Protocol 2. Acutely Positioned Catheters: Disposition and Distribution of Radiolabel During Intramedullary Infusion of [³H]NE

1

1

1

1

1

1

1

1

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Disposition of radiolabel [expressed as [3H]norepinephrine equivalents (ng · kg1 · min1) see METHODS for description of calculations] during a 20-min infusion of [3H]norepinephrine (16-24 nCi · kg1 · min1 in 100 ng · kg1 · min1 norepinephrine) into renal outer medulla of left kidney via 2 acutely positioned catheters. Coordinates represent means ± SE of 9 sets of observations. , Radiolabel spilling over from infused kidney into renal vein; , radiolabel reentering infused kidney via renal artery; , radiolabel excreted by infused (left) kidney; , radiolabel excreted by contralateral (right) kidney.

Intrarenal distribution of radiolabel. The concentration of radioactivity, determined from the solubilized kidney tissue, was significantly greater in the infused compared with the contralateral kidney (P value of infused vs. contralateral kidney = 0.03; Fig. 4A). Because these data were clearly not normally distributed, as evidenced by the proportionality of the means and their attendant SEs, they were log transformed and subjected again to ANOVA. Analysis of the log-transformed data showed significant heterogeneity of variance according to region in both the infused and contralateral kidney. Within the infused kidney, this can be attributed to the fact that the radiolabel was more concentrated in the medulla (495 ± 388 ng/g) and papilla (592 ± 347 ng/g) than in the cortex (75 ± 44 ng/g). The opposite appeared to be the case in the contralateral kidney in which radiolabel was more concentrated in the cortex (49 ± 29 ng/g) than the medulla (23 ± 12 ng/g) or papilla (25 ± 10 ng/g).

View larger version (21K): [in this window] [in a new window]   Fig. 4.   A: distribution of radiolabel in infused (left, solid bars) and contralateral (right, open bars) kidneys at end of 20-min outer medullary infusion of [3H]norepinephrine (16-24 nCi · kg1 · min1 in 100 ng · kg1 · min1 norepinephrine) via 2 acutely positioned catheters. Radiolabel in solubilized kidney tissue is expressed as [3H]norepinephrine equivalents (ng/g). Columns and error bars represent means ± SE for 9 rabbits. P values are outcomes of ANOVA testing for differences between levels of radioactivity (log10 transformed) in infused and contralateral kidneys (Pside) and for heterogeneity of variance among different kidney regions (Pregion). B: levels of radiolabel, expressed as disintegrations per minute per square mm (dpm/mm2), determined from autoradiographic analysis of 5 infused kidneys. Columns and error bars represent means ± SE. P value represents outcome of a partitioned ANOVA testing for a difference between levels of radioactivity in cortex compared with other three kidney regions. C: typical autoradiogram from an infused kidney. C, cortex; OS, outer stripe of outer medulla; M, medulla; P, papilla. See METHODS for definitions of these regions.

Consistent with the above observations, autoradiographic analysis of the coronal kidney sections demonstrated that the levels of radioactivity in the cortex of the infused kidney were 3.4 ± 0.6-, 5.6 ± 0.8-, and 8.0 ± 0.8-fold lower than those in the outer stripe, medulla, and papilla, respectively (Fig. 4, B and C).

Protocol 3. Chronically Positioned Outer and Inner Medullary Catheters

1

1

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Systemic and renal hemodynamic effects of outer (n = 8; left) and inner (n = 9; right) medullary interstitial infusion of norepinephrine delivered via chronically positioned catheters. Each dose (0, 30, 100, 300 ng · kg1 · min1) was infused over consecutive 20-min periods. Columns and error bars represent means ± SE of each variables over final 10-min period when all variables were stable. Solid bars, periods when norepinephrine was infused; open bars, periods during which vehicle (154 mM NaCl; 20 µl · kg1 · min1) was infused. P values represent outcome of a partitioned ANOVA testing for a dose-dependent effect of infusions on each hemodynamic variable. Recovery period was excluded for this analysis.

In contrast to outer medullary infusion, the dose-dependent pressor responses to inner medullary infusion of NE (30, 100, and 300 ng · kg¹ · min¹) were lesser in magnitude (increasing by 3 ± 3, 6 ± 3, and 8 ± 4% of its resting level, respectively), and no significant effects on CBF or MBF were observed (Fig. 5, right).

Effects of intravenous infusion of NE. Intravenous NE (300 ng · kg¹ · min¹) increased MAP (by 7 ± 1% of its resting level) and reduced RBF_probe (by 15 ± 5% of its resting level) and CBF (by 14 ± 7% of its resting level) but did not significantly alter MBF (4 ± 4% change) (Fig. 6).

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Effects of intravenous infusion of norepinephrine (300 ng · kg1 · min1) on systemic and renal hemodynamic variables (n = 17). Columns and error bars represent means ± SE of each variable during final 10 min of each 20-min infusion period. Open bars, periods before and after; solid bars, period during norepinephrine infusion. P values represent outcomes of partitioned ANOVA, testing whether levels of each variable during intravenous norepinephrine infusion differed from those preceeding and following it.

Disposition of radiolabel during outer and inner medullary infusion of [³H]NE. The profile of renal disposition of the radiolabel during outer medullary infusion of [³H]NE was similar to that found using acutely positioned catheters (protocol 2). That is, much of the infused radiolabel spilled over into the renal vein (39.1 ± 25.2 ng · kg¹ · min¹ during the final 6 min of the infusion), but only small amounts were excreted by the infused kidney (3.7 ± 2.0 ng · kg¹ · min¹ during the final 6 min of the infusion). Not surprisingly therefore, progressive increases were observed in the amount of radiolabel reentering the infused kidney via the renal artery (9.6 ± 0.4 ng · kg¹ · min¹ during the final 6 min of the infusion) and the amount excreted by the contralateral kidney (1.1 ± 0.1 ng · kg¹ · min¹ during the final 6 min of the infusion) (Fig. 7, left).

View larger version (28K): [in this window] [in a new window]   Fig. 7.   Disposition of radiolabel (expressed as [3H]norepinephrine equivalents; ng · kg1 · min1) during a 20-min infusion of [3H]norepinephrine (16-24 nCi · kg1 · min1 in 100 ng · kg1 · min1 norepinephrine) into outer (left) and inner (right) medulla of left kidney via chronically positioned catheters. Coordinates represent means ± SE of 8 (outer medullary infusion) or 9 (inner medullary infusion) sets of observations. P values represent outcomes of repeated-measures ANOVA, testing whether location of catheter tip (outer vs. inner medulla) influenced disposition of radiolabel.

In contrast, during inner medullary infusion of [³H]NE, only small amounts of the radiolabel spilled over into the renal vein of the infused kidney (4.8 ± 2.7 ng · kg¹ · min¹ during the final 6 min of the infusion), whereas most of the infused radiolabel was excreted by the infused kidney (63.6 ± 12.8 ng · kg¹ · min¹ during the final 6 min of the infusion). Consistent with this, only small quantities of radiolabel reentered the infused kidney via the renal artery (1.6 ± 1.1 ng · kg¹ · min¹ during the final 6 min of the infusion) or were excreted by the right kidney (0.1 ± 0.1 ng · kg¹ · min¹ during the final 6 min of the infusion) (Fig. 7, right).

Intrarenal distribution of radiolabel. There was considerable variation associated with the levels of radiolabel in the solubilized kidney tissue, particularly during outer medullary infusion of radiolabel. It was clear, however, that much greater levels of the radiolabel remained in the kidney during outer medullary, compared with inner medullary, infusion of [³H]NE (Fig. 8A).

View larger version (34K): [in this window] [in a new window]   Fig. 8.   A: distribution of radiolabel in infused (left, solid bars) and contralateral (right, open bars) kidneys at end of 20-min infusions of [3H]norepinephrine (16-24 nCi · kg1 · min1 in 100 ng · kg1 · min1 norepinephrine) into outer (left) and inner (right) medulla via chronically positioned catheters. Radiolabel in solubilized kidney tissue, expressed as [3H]norepinephrine equivalents (ng/g). Columns and error bars represent means ± SE for 7 (outer medulla) or 9 (inner medulla) rabbits. P values are as for Fig. 4. Note different scales for different routes. B: levels of radiolabel, expressed as dpm/mm2, determined from autoradiographic analysis of infused kidneys. Columns, error bars, and P values are as in Fig. 3B. C: typical autoradiograms for infused kidneys.

Autoradiographic analysis of the coronal kidney sections demonstrated that with outer medullary infusion, the levels of radioactivity in the cortex of the infused kidney were 10.3 ± 2.3-, 14.0 ± 3.3-, and 8.8 ± 2.8-fold lower than those in the outer stripe, medulla, and papilla, respectively. During inner medullary infusion, the levels of radioactivity in the cortex of the infused kidney were 6.6 ± 1.6-, 8.7 ± 2.2-, and 8.2 ± 2.4-fold lower than those in the outer stripe, medullary, and papillary regions, respectively (Fig. 8, B and C).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The primary objective of this study was to design and validate techniques for delivery of pharmacological agents into the renal medullary interstitium of rabbits. Similar techniques have been developed in rats (13, 14, 19, 23), and these have helped provide considerable information regarding the role of the renal medulla and in particular the renal medullary microcirculation in the control of blood pressure (4-6). Our results suggest that in the rabbit, outer medullary interstitial infusion of pharmacological agents, using either acutely or chronically implanted catheters, may be a useful method for targeting pharmacological agents to the renal medulla and in particular for altering MBF. Thus delivery of NE into the outer medullary interstitium via acutely positioned or chronically implanted catheters dose relatedly reduced MBF more than it did CBF or RBF_probe. In contrast, intravenous infusion of NE reduced RBF_probe and CBF but not MBF. When [³H]NE was infused into the outer medulla, the concentration of the radiolabel was found to be much greater in the renal medulla of the infused kidney than in its cortex or in the contralateral kidney. We therefore conclude that outer medullary interstitial infusion of NE achieves high concentrations of this agent at vascular sites important in the control of MBF, in the medulla, and/or in the inner cortex.

One of the strengths of the present study was the estimation, during medullary interstitial infusion of [³H]NE, of the amount of radiolabel spilled over from the kidney into the renal vein, the amount of this radiolabel that reentered the kidney via the renal artery, and the amounts of radiolabel excreted by the infused and contralateral kidneys. Although much of the radiolabel in these biological fluids probably reflects metabolites of [³H]NE, we argue that most small, uncharged molecules should be handled similarly by the kidney during medullary interstitial infusion. In other words, the pattern of renal disposition of radiolabel during [³H]NE infusion probably reflects that expected for stable, uncharged small molecules. In the case of NE and other molecules that are rapidly metabolized in vivo the proportion of the radiolabel that represents intact [³H]NE must become less in proportion to the distance from the infusion site. In this study therefore, we should not expect complete agreement between the localization of the radiolabel and the effects of the NE infusions. Nevertheless, with this caveat in mind, we suggest that this analysis provides two important observations that illustrate the limitations of the medullary interstitial infusion technique.

First, although we were able to achieve much greater concentrations of radiolabel within the medulla compared with the cortex during outer medullary infusion of [³H]NE, we also found that much of the radiolabel (~40% using the chronically implanted catheter) spilled over into the renal vein and so recirculated systemically. This spillover is unlikely to greatly confound the interpretation of studies in which relatively unstable compounds (e.g., NE and angiotensin II) or compounds that are metabolized in the pulmonary circulation (e.g., bradykinin and endothelin-1) are infused. Nevertheless, the dose-related pressor effect that we observed clearly indicates that significant quantities of intact NE do spill over during medullary interstitial infusion. Furthermore, in studies where more stable compounds are infused, spillover into the systemic circulation will almost certainly confound the interpretation of the results. This appears to have been the case in a recent study in which we compared the effects of outer medullary, renal arterial, and intravenous infusion of the vasopressin V₁ agonist [Phe²,Ile³,Orn⁸]vasopressin (8). The pressor and bradycardic effects of this agent were indistinguishable by all three routes, as was the reduction in MBF. Systemic spillover during outer medullary interstitial infusion therefore appears not to be a unique property of NE but also occurs with other small molecules.

Second, we observed a major effect of the site of the medullary interstitial infusion on the way in which the kidney handled the infused [³H]NE. Thus, in contrast to the outer medullary infusion, where much of the infused radiolabel spilled over into the systemic circulation, during inner medullary infusion ~60% of the infused radiolabel was excreted by the infused kidney. The reason for this difference between the outer medullary and inner medullary infusions remains to be determined unequivocally but may relate in part to the presence of mechanisms for tubular secretion of NE (11) and to the relatively lower blood flow in the inner medulla compared with the outer medulla (20). It is unlikely to reflect leakage of [³H]NE due to damage to the papillary tissue from implantation of the catheter, because no such damage was observed in the frozen sections submitted for autoradiography. This effect of the infusion site on the disposition of the infused radiolabel probably explains why the levels of radioactivity were considerably less in the excised kidneys that had received an inner medullary infusion of NE compared with those that had received the outer medullary infusion. This, in turn, may help explain why MBF was not reduced during inner medullary infusion of NE and why the pressor effect of NE was considerably less with inner medullary infusion than with outer medullary infusion of NE.

Some of the present observations are at odds with similar studies performed in rats. For example, in rats, infusion of vasoconstrictor agents into the inner medulla reduces MBF, whereas similar infusions of vasodilator agents can increase MBF (4-6, 13, 14, 19). In contrast, in the present study we found that outer medullary, but not inner medullary, infusion of NE reduced MBF. There are a number of possible explanations for this apparent species difference, the most obvious being the difference in dimensions of the renal medulla in these two species. We hypothesize therefore, that in the rat (but not the rabbit, with a ~10-fold greater kidney weight), agents infused into the inner medulla may easily diffuse the relatively short distance to the outer medulla and inner cortex to influence vascular elements controlling MBF. It is also possible that differences in medullary countercurrent mechanisms between the two species might alter the renal handling of substances infused into the renal medulla. Indeed, Cowley and colleagues (5) argued previously that substances infused into the inner medulla are accumulated because of the efficient countercurrent exchanger in the vasa recta circulation. The results of the present study indicate that this is not the case in rabbits receiving medullary interstitial infusions of [³H]NE, in which much of the infused radiolabel either spilled over into the systemic circulation (outer medullary infusion) or was excreted in the urine of the infused kidney (inner medullary infusion). These apparent differences between rats and rabbits in the ability of the inner medulla to "trap" substances infused into the interstitium could reflect the differences in medullary structure between the two species (12). For example, the rabbit renal medulla has a "simple" structure, with relatively small vascular bundles containing only ascending and descending vasa recta. In the more "complex" rat renal medulla, larger vascular bundles are found that also contain descending thin limbs of short loops of Henle.

On the other hand, our results are in agreement with previous studies by Cowley and colleagues (5, 14) in that during medullary interstitial infusion of a radiolabeled small molecule ([¹⁴C]clentiazem in their case and [³H]NE in the present study) the radiolabel within the infused kidney was mostly concentrated in the medulla and papilla, with very little radiolabel in the cortex of the infused kidney or in the contralateral kidney. Consistent with these observations, medullary interstitial infusion of pharmacological agents in rats (23) and rabbits (present study) can have effects quite distinct from those of intravenous infusion of these agents. In the case of NE, we found that in contrast to intravenous NE, which reduced CBF and RBF_probe without significantly affecting MBF, outer medullary infusion of NE dose relatedly reduced MBF but had considerably smaller effects on CBF and RBF_probe.

Although we cannot be certain of the precise vascular elements that mediate the reduction in MBF during outer medullary infusion of NE, we can at least suggest some likely candidates. Within the kidney, NE can mediate vasoconstriction directly by acting on vascular ₁- and ₂-adrenoceptors (7, 27) or indirectly by -adrenoceptor-mediated stimulation of renin release (24). The fact that intravenous infusion of NE reduced CBF but not MBF suggests that an indirect stimulus, via renin release, is unlikely. NE directly constricts outer cortical afferent and efferent arterioles in situ (3) and outer medullary descending vasa recta in vitro (28). It is also possible that vascular sites in the inner medulla play some role in mediating the reduced MBF, because contractile elements have recently been identified in inner medullary descending vasa recta in rats (21). Our finding that inner medullary infusion of NE did not affect MBF does not exclude this possibility, because infusion of [³H]NE by this route resulted in relatively low levels of accumulated radiolabel in the inner medulla.

There was, however, clear evidence of cortical effects of the infused NE, because CBF and RBF_probe were reduced, and, at least in the case of the acutely positioned catheters, GFR was also reduced. Indeed, the reductions in urine flow and sodium excretion during outer medullary infusion of NE could be completely accounted for by the reduced GFR, indicating no net change in tubular sodium and water reabsorption. This latter observation seems at odds with the notion that reduced MBF should enhance tubular salt and water reabsorption (6), but may be explained by the confounding impact of the pressor effect of the medullary interstitial NE infusion. Clearly, further studies in which renal perfusion pressure is controlled are required to delineate the direct effects of medullary interstitial NE infusion on renal excretory function.

Perspectives