Dysfunctional Renal Nitric Oxide Synthase as a Determinant of Salt-Sensitive Hypertension

Animals, BP, and Body Weight

Male salt-sensitive (DS; n = 7 to 10 per group) and salt-resistant (DR; n = 6 to 8 per group) Dahl rats (13 wk of age; Charles River WIGA, Sulzfeld, Germany) were randomly assigned to control diet (standard chow), high salt diet (NaCl 4%; Harlan Teklad, Madison, WI), or high salt diet in combination with LU135252, an orally active, ET_A receptor-selective endothelin antagonist. LU135252 was a gift from Knoll AG (Ludwigshafen, Germany) and administered by chow (60 mg/d + NaCl 4%) for 8 wk. Body weight and food intake were continuously monitored, and systolic BP was measured at the beginning, after 4 wk, and at the end of treatment as described (9,24). Study design and experimental protocols were approved by the institutional animal care committee.

Arterial and Renal Tissue Preparations

Rats were anesthetized (thiopental, 50 mg/kg body wt, intraperitoneally) and sacrificed. The left and right main renal artery and the thoracic aorta were isolated, removed, and placed into cold (4°C) Krebs Ringer bicarbonate solution (in mmol/L): 118.6 NaCl, 4.7 KCl, 2.5 CaCl₂, 1.2 KH₂PO₄, 1.2 MgSO₄, 25.1 NaHCO₃, 0.026 ethylenediaminetetra-acetic acid (EDTA) calcium disodium, and 11.1 glucose. Arteries were dissected in cold Krebs solution under a microscope (Wild-Heerbrugg, Switzerland), cleaned from perivascular tissue, and rinsed with a cannula to remove residual blood cells. Two rings of each renal artery (length: DS 2.37 ± 0.03 mm, n = 67 rings; DR 2.43 ± 0.03 mm, n = 51 rings) were used for organ chamber experiments. For endothelium-dependent contractions, aortic rings were used for comparison (9,25). Remaining renal artery tissue was snap-frozen in liquid nitrogen and stored at -80°C. The right kidney was cut in half, and one half was fixed in phosphate-buffered saline-buffered paraformaldehyde (4%) at 4°C, processed for histologic analysis, and embedded in paraffin. The other half and the left kidney were decapsulated, sliced horizontally, separated between outer medulla and inner cortex, immediately snap-frozen in liquid nitrogen, and kept at -80°C until assayed.

Renal Artery Endothelial Function and Structure

Renal artery rings were suspended to fine tungsten holders (diameter, 100 μm) in organ chambers containing 25 ml of Krebs-bicarbonate solution (37°C, pH 7.4, 95% O₂ and 5% CO₂) and equilibrated for 1 h. Resting tension was gradually increased, and rings were repeatedly exposed to 100 mmol/L KCl until the optimal tension for generation of force during isometric contraction was reached (DS group: 1.48 ± 0.01 g, n = 90 rings; DR group: 1.46 ± 0.01 g, n = 72 rings, NS). After equilibration for 30 min, rings were exposed to cumulative concentrations of ET-1 (10^-11 - 3 × 10^-7 mol/L) or norepinephrine (10^-10 3 × 10^-5 mol/L). In quiescent renal artery and aortic rings, basal NO release was assessed by contractions to NOS inhibitor N^G-nitro-L-arginine methyl ester (L-NAME; 10^-5 - 3 × 10^-4 mol/L). In rings pretreated with L-NAME (3 × 10^-4 mol/L, 30 min of preincubation), acetylcholine (10^-8 - 3 × 10^-4 mol/L) in the presence or absence of the thromboxane A₂/prostaglandin H₂ receptor antagonist SQ 30741 (10^-7 mol/L, 30 min of preincubation) or cyclooxygenase inhibitor indomethacin (10^-5 mol/L, 30 min of preincubation) was used to study endothelium-dependent contractions. Other rings were precontracted with norepinephrine (1 × 10^-8 - 2 × 10^-7 mol/L) until a stable plateau was reached (approximately 70% of contraction induced by 100 mmol/L KCl); precontraction was not different between groups (data not shown). Relaxations to acetylcholine (10^-10 - 3 × 10^-5 mol/L)—with or without indomethacin (10^-5 mol/L, 30 min of preincubation)— and to sodium nitroprusside (10^-10 - 3 × 10^-5 mol/L) were then performed. In some preparations, the endothelium was removed and its absence was confirmed by the lack of relaxation in response to acetylcholine (3 × 10^-6 mol/L). For the assessment of vascular hypertrophy, rings were blotted dry after the experiments and weighed, and arterial surface area (mm²) was calculated by microscopic planimetry as described previously (9). Arterial diameter of renal artery rings was calculated using the following equations: Given that the circumference (c) of the artery equals the length of the transversely cut, opened arterial strip, the formula (c = 2πr) was used (π = 3.1415, r = vessel radius). Renal artery diameter (2r = c/π) was calculated by use of formulas for diameter (d = 2r) and radius (r = c/2π) of a cylinder for each individual ring, and values were averaged. Mean diameter of renal arteries was 444 ± 14 μm (DS rats, n = 67 rings) and 428 ± 11 μm (DR rats, n = 51 rings), with no statistical difference between treatment groups (Table 1).

Determination of NOS Activity in Renal Tissue

NOS activity was determined as the formation of L-[2,3,4,5-³H]-citrulline from L-[2,3,4,5-³H]-arginine of tissue homogenates from renal cortex and medulla. After excision, kidneys were dissected into cortex and medulla, frozen in liquid nitrogen, and stored at -80°C until assay. For determination, tissue samples were homogenized using an Ultra-Turrax in 1.5 ml of buffer (pH 7.4) containing Tris (50 mmol/L), sucrose (320 mmol/L), EDTA (1 mmol/L), dithiothreitol (1 mmol/L), aprotinin (2 mg/L), and phenylmethylsulfonyl fluoride (100 mg/L). The homogenates were used for measuring protein concentration using the Bradford method and NOS activity. All measurements were performed in duplicate. Total (Ca²⁺-dependent plus Ca²⁺-independent) activity was measured by incubating 50 μl of homogenate during 30 min at 37°C with 50 μl of KH₂PO₄ buffer (50 mmol/L, pH 7.2) containing (final concentrations) dithiothreitol (1 mmol/L), NADPH (0.5 mmol/L), tetrahydrobiopterin (0.3 mmol/L), calmodulin (300 U/ml), CaCl₂ (1 mmol/L), L-citrulline (1 mmol/L), L-arginine (0.01 mmol/L), and L-[2,3,4,5-³H]-arginine (3.7 KBq). Ca²⁺-independent activity was measured by incubation in the presence of EDTA (10 mmol/L) instead of CaCl₂, and nonspecific activity was determined by incubation in the presence of L-NAME (100 mmol/L) instead of NADPH. Lower concentrations of EDTA and L-NAME were found to give incomplete inhibition of Ca²⁺-dependent and total NOS activity, respectively. Incubations were stopped by placing the incubation mixtures in ice, followed by addition of 1 ml of ice-cold Hepes buffer (20 mmol/L, pH 5.5). The mixtures were loaded onto 1-ml Dowex 50×8-200 (Na⁺ form) columns, followed by washing with 1 ml of stop buffer. Citrulline was eluted with 2 ml of stop buffer, collected into scintillation vials, and counted in a beta counter in the presence of 15 ml of scintillation fluid. By addition of L-[¹⁴C]-citrulline before or after the incubation step, the recovery of this procedure was found to be >95%. After correction for nonspecific counts, total and Ca²⁺-independent NOS activity was calculated from the percent conversion of [³H]-arginine into [³H]-citrulline in the presence of CaCl₂ or EDTA, respectively, and expressed as picomoles per milligram of protein per minute. Ca²⁺-dependent activity was calculated by subtraction of Ca²⁺-independent activity from total activity. In additional experiments, we compared NOS enzyme activity in membrane fractions from 20,000 × g supernatants with the activity measured in crude tissue homogenates.

Morphologic Analysis of Glomerular Injury

Renal injury was assessed as described previously (26). Briefly, paraffin-embedded sections of whole kidneys (5 to 7 μm) stained with periodic acid-Schiff reagent were viewed by light microscopy at a magnification of ×40 using a Zeiss microscope. One hundred glomeruli per slide were evaluated. Morphologic evaluation of glomerular injury was performed by two of the authors (I.V., H.J.G.) blinded to the groups, using semiquantitative scoring methods. Lesions were graded by glomerulosclerosis (grade 0 to 4, i.e., 1 to 25, 26 to 50, 51 to 76, and 76 to 100% sclerosis), mesangiosclerosis, mesangiolysis, and mesangioproliferation (grade 1 to 3), ischemia, and thrombosis. The glomerular injury score was calculated by summarizing the products of severity score (index) and the percentage of glomeruli or arteries displaying the same degree of severity. The total injury index was defined as total amount of morphologic changes.

Quantification of Tissue Endothelin Protein Content

Determination of tissue ET-1 and ET-3 protein content was performed in a blinded manner. Frozen renal artery tissue was pulverized and homogenized using a polytron for 60 s in ice-cold chloroform: methanol (2:1 dilution) containing 1 mmol/L N-ethylmaleimide and 0.1% trifluoroacetic acid. Frozen tissue from renal cortex and medulla (300 mg) was homogenized without pulverization. ET-1 and ET-3 protein was extracted from tissue homogenates as described (24). RIA measurements and reversed-phase HPLC was used for ET-1 and ET-3 protein content determination according to previously published protocols (24,27), and vascular and renal ET-1 and ET-3 tissue content was related to tissue weight (pg/g) (24).

Materials

Acetylcholine chloride, calmodulin, CaCl₂, dichlorodiphenyltri-chloroethane, Dowex resin AG 50W-X4, EDTA, ET-3, indomethacin (dissolved in 5 mmol/L sodium carbonate), L-arginine, L-NAME, NADPH, norepinephrine bitartrate salt, phenylmethylsulfonyl fluoride, potassium chloride, sucrose, sodium nitroprusside dihydrate, and Tris salt were purchased from Sigma Chemical Co. (St. Louis, MO). ET-1 was from Calbiochem/Novabiochem AG (Läufingen, Switzerland). Rabbit antibodies against synthetic ET-1 and ET-3 were from Peninsula Laboratories (San Carlos, CA). Radiolabeled ¹²⁵I-ET-1, ¹²⁵I-ET-3, and L-[2,3,4,5-³H]-arginine was purchased from Amersham (Amersham, Buckinghamshire, United Kingdom). SQ 30741 was a gift of Bristol-Myers Squibb (Princeton, NJ), and pentobarbital was from Abbott Laboratories (Chicago, IL).

Statistical Analyses

Data are given as mean ± SEM, and n equals the number of animals used. Relaxations are expressed as percent contraction to norepinephrine, and contractions are given as percent contraction to potassium chloride. For multiple comparisons, results were analyzed using ANOVA followed by Bonferroni correction. For comparison between two values, the unpaired t test or the nonparametric Mann-Whitney test were used when appropriate. The Pearson correlation coefficient was calculated by linear regression analysis. A P value <0.05 was considered significant.

BP and Body Weight

Systolic BP increased in DS (P < 0.05) but not in DR rats on high salt diet after 4 wk and increased further after 8 wk (Figure 1). Hypertensive DS rats had significantly lower body weight than salt-loaded DR rats (P < 0.05 versus control) (Table 1). ET_A receptor antagonist LU135252 treatment only in part reduced the BP increase (P < 0.05) (Figure 1) but largely prevented the inadequate weight gain (P < 0.05) (Table 1). No significant effect on BP or body weight was observed in DR rats.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

BP in Dahl salt-sensitive (DS) and salt-resistant (DR) rats. (A) In DS but not DR rats on a high salt diet, systolic BP increased after 4 and 8 wk compared with DS control rats (P < 0.05). Concomitant LU135252 in part reduced the increase in pressure in DS rats (P < 0.05), and had no effect in DR rats. Data are mean ± SEM. ^*P < 0.05 versus control; †P < 0.05 versus salt.

Endothelium-Dependent Relaxations to Acetylcholine

Despite hypertension, relaxations to acetylcholine were unaffected in DS rats (NS) (Figure 2); however, concomitant treatment with LU135252 slightly enhanced relaxations in DS rats (P < 0.05 versus salt). Surprisingly, relaxations were reduced in salt-treated DR rats after chronic treatment with LU135252 (P < 0.05) (Figure 2A, right panel), which could be prevented by cyclooxygenase inhibition with indomethacin in vitro (P < 0.05) (Table 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Endothelium-dependent relaxations to acetylcholine and endothelium-dependent contractions to nitric oxide (NO) synthase inhibitor N^G-nitro-L-arginine methyl ester (L-NAME) in renal arteries. (A) High sodium diet alone had no effect on relaxations to acetylcholine in DS or DR animals. However, in animals concomitantly treated with LU135252, relaxations were slightly but significantly increased (left panel), whereas they were reduced in DR rats (right panel) (P < 0.05). (B). High salt diet attenuated the contraction to L-NAME in DS rats, whereas salt treatment increased contractions in DR rats. Concomitant LU135252 treatment prevented the attenuation of basal NO release in DS rats and further enhanced contractions to L-NAME in DR rats. Data are mean ± SEM. ^*P < 0.05 versus control; ^†P < 0.05 versus salt.

Endothelium-Dependent Contractions to NOS Inhibitor L-NAME

NOS inhibitor L-NAME caused concentration-dependent contractions in quiescent renal artery and aortic rings. Treatment with high salt diet attenuated contractile responses to L-NAME in DS rats, while contractions increased in DR rats (Figure 2B) (P < 0.05 versus low salt). In DS rats, LU135252 treatment normalized the attenuated response (P < 0.05 versus salt). In comparison, maximal contractions to L-NAME (300 μmol/L) in the aorta of DS control rats were about one-third in magnitude compared with the renal artery (10 ± 2% of KCl) (P < 0.01 versus renal artery) and unaffected by high-salt diet or LU135252 (Table 2).

Endothelium-Dependent Contractions to Acetylcholine

The magnitude of responses was similar in renal arteries from DS and DR control rats (NS) (Figure 3). High salt diet markedly potentiated contractions in DS rats (approximately threefold, from 25 ± 3 to 78 ± 4%, P < 0.05 versus control) and, to a lesser extent, in DR rats (approximately 1.6-fold) (Figure 3). Contractions were correlated with vascular ET-1 protein content in DS rats only (r = 0.7, P < 0.01). LU135252 treatment completely normalized the increased contractions (Figure 3) (P < 0.05 versus high salt). Responses to acetylcholine were markedly inhibited by pretreatment with the thromboxane receptor antagonist SQ 30741 (10^-7 mol/L, P < 0.05) (Table 2) or indomethacin (10^-5 mol/L) in all groups (data not shown). In the aorta of DS rats, contractions to acetylcholine were only about 1/20 in magnitude compared with the renal artery (DS control rats 3 ± 2% and DS rats 4 ± 2% on a high salt diet, NS) and unaffected by concomitant LU135252 treatment (Table 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Endothelium-dependent, prostanoid-mediated contractions to acetylcholine in renal arteries. In renal arteries of control animals, contractions to acetylcholine (in the presence of L-NAME, 0.3 mmol/L) were similar between strains. High salt diet markedly enhanced contractions in DS and, to a lesser extent, in DR rats (P < 0.05); the effect of high salt diet was completely prevented by LU135252 in both strains. Data are mean ± SEM. ^*P < 0.05 versus control; ^†P < 0.05 versus salt.

Endothelium-Independent Responses

Potassium chloride-induced contractions were comparable between groups (NS) (Table 2). Contractions to ET-1 but not those to norepinephrine were attenuated in salt-loaded DS rats; however, this was prevented by concomitant LU135252 treatment (Figure 4, A and B). LU135252 normalized the response to ET-1 (P < 0.05 versus salt) (Figure 5, right panel). Endothelium-independent relaxation to sodium nitroprusside was unaffected by the treatments (NS) (Table 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Contractions to endothelin-1 (ET-1) and norepinephrine in isolated renal arteries. (A) In control animals, the response to ET-1 was greater in DS than DR rats (P < 0.05). High salt diet reduced contractions in DS but not DR rats. LU135252 treatment restored this attenuation in DS rats, whereas contractions were slightly enhanced in DR rats. (B) Contractions to norepinephrine were unaffected by either treatment in DS or DR rats. Data are mean ± SEM. ^*P < 0.05 versus control; ^†P < 0.05 versus salt.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Renal nitric oxide synthase (NOS) activity. In DR rats (right panel), high salt diet induced a threefold increase in total NOS activity in renal cortex and medulla. No increase in response to salt loading was observed in DS rats (left panel). However, concomitant treatment with LU135252 slightly increased renal NOS activity in kidneys of DS rats only (P < 0.05, left panel). Note that NOS activity was higher in the medulla than in the cortex in both strains. Data are mean ± SEM. ^*P < 0.05 versus control; ^†P < 0.05 versus salt; ^#P < 0.05 versus cortex.

Renal Artery Structure

Vascular hypertrophy of the renal artery (expressed as ratio of tissue weight in mg per mm² vascular surface area) was observed after salt loading in DS rats only. The ratio increased from 0.17 ± 0.01 to 0.23 ± 0.01 mg/mm², and was completely normalized by concomitant LU135252 treatment (0.18 ± 0.01 mg/mm², P < 0.001 versus salt). High salt diet with or without LU135252 had no effect on vascular structure in DR rats (NS) (Table 1).

Renal NOS Activity

Total NOS activity is given as the L-NAME-inhibitable portion of L-citrulline formation. L-NAME concentrations <100 mmol/L showed incomplete inhibition of total NOS activity. Total activity was comparable between untreated DS and DR rats and higher in the medulla than in the cortex (Figure 5). Total NOS activity was of the same order of magnitude as reported previously in Dahl rats, Wistar Kyoto rats, and spontaneously hypertensive rats (8,28). The majority of renal NOS activity was calcium-independent (Table 3). Salt loading markedly increased total NOS activity (about three-fold) in DR rats (P < 0.05). In contrast, DS rats failed to increase total NOS activity after salt loading (Figure 5). Interestingly, LU135252 treatment was associated with an increase of total NOS activity in both cortex (74 ± 10%) and medulla (58 ± 8%) in DS rats only (P < 0.05) (Figure 5). Using membrane fractions instead of crude homogenates resulted in no net difference with regard to enzyme activity and had no effect on the ratio of calcium-dependent and -independent conversion (data not shown).

Glomerular Injury

Compared with DS control animals, the number of structurally normal glomeruli (grade 0 renal injury) in hypertensive DS rats was reduced to 49 ± 2% (P < 0.05). Hypertension in DS rats was associated with pronounced glomerular damage with thickening of Bowman's capsule, adhesion formation, mesangial proliferation, and hypertrophy of the preglomerular artergial and arterioles (Figure 6, middle panels). Concomitant LU135252 treatment increased the number of healthy glomeruli from 49 ± 2 to 79 ± 4% (P < 0.05 versus high salt) (Figure 7) and markedly reduced structural glomerular changes (P < 0.05) (Figure 6, right panels and Figure 7). In salt-treated DR rats, renal morphology was essentially identical to untreated DR rats, which showed no signs of renal damage (authors' unpublished observation and reference 26).

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Morphology of glomerular injury in kidneys from DS rats. (Left) Paraffin sections from a DS control rat kidney showing a normal glomerulus and preglomerular artery (top panel) and normal intrarenal artery (bottom panel). (Middle) Glomerulosclerosis (top panel) and marked intrarenal artery hypertrophy (below) in a salt-loaded, hypertensive DS rat. (Right) Glomerular injury (top panel) and vascular hypertrophy (bottom panel) were largely prevented by ET_A antagonist LU135252. Magnification, × 100 (oil immersion), paraffin, periodic acid-Schiff (PAS) stain.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

Morphologic assessment of glomerular injury in DS rats. Salt-induced hypertension markedly reduced the number of healthy glomeruli (grade 0 renal injury) and was associated with renal injury (grades 1 to 4, black bars), which was inhibited by chronic ET_A receptor blockade. No morphologic changes were observed in DS control rats or DR rats (not shown). Data are mean ± SEM. ^*P < 0.05 versus control; ^†P < 0.05 versus salt.

Endothelin Protein Content in Renal Artery and Kidney

In hypertensive DS rats, renal artery ET-1 protein content was increased from 165 ± 33 to 484 ± 57 pg/g tissue (P < 0.05). This increase was prevented by chronic LU135252 treatment (124 ± 11 pg/g tissue, P < 0.05 versus salt). No changes occurred in DR rats (Table 1). LU135252 also prevented the increase in ET-1 protein content in renal cortex and medulla of DS rats only (Figure 8) (P < 0.05). Renal ET-3 protein content, which was predominantly expressed in the medulla of both DS and DR rats, was unaffected by high salt diet (NS) (Figure 8B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 8.

ET-1 and ET-3 protein content in renal cortex and medulla. (A) In control DR animals, ET-1 protein content in the medulla was higher than in the cortex (right panel, P < 0.05). No difference was observed in DS rats (left panel). High salt diet increased ET-1 content in DS rats only, and this increase was prevented by chronic ET_A receptor blockade. (B) In all groups of rats, ET-3 protein content in the medulla was higher than in the cortex (P < 0.05) and not significantly affected by treatments. Data are mean ± SEM. ^*P < 0.05 versus control; ^†P < 0.05 versus salt; ^#P < 0.05 versus cortex.

Salt-Induced Hypertension: A Failure to Increase Renal NOS Activity

The key finding of our study was the observation that DS rats failed to increase renal NOS activity in response to salt loading. In contrast, and consistent with increased renal NOS2 protein expression in salt-loaded Sprague Dawley rats (17), renal NOS activity in DR rats increased markedly after high salt diet treatment. Neither in Sprague Dawley rats (17) nor in DR rats in the present study did salt loading affect BP. The observed failure in DS rats to adequately increase NOS activity could be linked to a genetically determined defect of NOS or its regulation. In this context, it is of interest that the gene of calcium-independent Nos2 cosegregates with BP in DS rats (12,13). A functional defect of NOS is also suggested by studies demonstrating that treatment with the NOS substrate L-arginine abolishes hypertension and inhibits renal injury in this model (14,15,29).

An unexpected and novel finding in contrast to previous studies (8,28) was that in all groups more than half of renal NOS activity was calcium-independent. It is important to note that in our experiments L-NAME concentrations <100 mmol/L did not completely block NOS activity in renal tissue. The concentration used was 100-fold higher compared to previously published protocols (8,28), and may in part explain the difference between calcium-dependent and -independent NOS activity compared with these previous studies. Unlike other calcium-independent NOS enzymes, expression of Nos2 is constitutive in Dahl and Wistar-Kyoto rat kidneys (30,31), and can be stimulated by salt loading (17). Also, sodium-mediated effects in humans may involve the L-arginine/NO pathway, as L-arginine-induced increases in sodium excretion and inhibition of fractional sodium reabsorption are dependent on salt loading (32). NOS2 regulates BP under normal conditions (18), and chronic inhibition of NOS2 is associated with reduced urinary nitrate excretion and proteinuria (33), indeed suggesting a role for NOS2-derived NO in renal function. Thus, compensatory upregulation of Nos2 expression and/or NOS activity in the kidney (as observed in DR rats in the present study) may represent a novel mechanism regulating the increased volume load and BP. Two distinct forms of Nos2 mRNA—macrophage Nos2 and vascular smooth muscle cell Nos2—have been identified in the rat kidney (30). In addition, evidence from a very recent study indicates that Nos2 mRNA is present in both rat cortex and medulla and that the enzyme is functionally active (34). Also, Nos2 is present in vascular smooth muscle of intrarenal vessels (30), and reduced nitrate production by Nos2 has been demonstrated in vascular smooth muscle cell of DS rats (13).

Future studies will determine whether the increase in calcium-independent NOS activity observed in DR rats is derived from NOS2. In this context, it is interesting to note that the paradigm of calcium dependency of NOS isoenzymes has recently been modified by studies showing that Nos2 activity—although completely calmodulin-independent—can be partly inhibited by removal of calcium (35). On the other hand, calcium-independent actions of Nos3 have recently been reported (36,37). Although we did not investigate gene expression of NOS isoenzymes, our study unequivocally demonstrates that salt sensitivity appears to involve a failure to increase renal NOS activity in response to salt loading, and that a substantial portion of the activity is calcium-independent. These findings are in agreement with the previously mentioned molecular data and a recent preliminary report (38).

Renal Artery Endothelial Dysfunction: Role of NO and Vasoconstrictor Prostanoids

Impaired renal vasodilatory capacity after salt loading has been described in Dahl hypertension (6). Here, we provide evidence that “selective” endothelial dysfunction may be one of the mechanisms interfering with renal artery vasodilation. Unexpectedly, and in contrast to aorta (9,25) and mesenteric artery (39), endothelium-dependent relaxations in the renal artery were unaffected by hypertension. However, endothelial dysfunction due to abnormal increased basal NO release and prostanoid-mediated contractions was present, which could result in inappropriate vascoconstriction. To our knowledge, this is the first study reporting the release of endothelium-derived vasoconstrictor prostanoids as part of endothelial dysfunction in Dahl hypertension. These mechanisms are likely to contribute to impaired renal vasodilatory capacity (6). Furthermore, our data suggest that endothelium-dependent relaxation per se should not be used as the sole indicator to indicate “preserved” endothelial function.

Interestingly, our study also shows that endothelin plays an important role in these functional changes. Although chronic ET_A receptor blockade only partly reduced hypertension in DS rats, the treatment completely prevented the attenuation of basal NO release and the enhanced vasoconstrictor prostanoid release. Renal NOS activity was also slightly increased, suggesting that ET_A receptor blockade can interfere with the L-arginine/NO pathway in vivo, in line with recent in vitro observations in mesangial (40) and vascular (41) cells in culture and normotensive animals in vivo (20). Both ET-1 (through ET_A receptor-mediated mechanisms) (42,43) and NO deficiency (44) have been implicated in vasoconstrictor prostanoid formation in vitro. As chronic treatment with LU135252 completely normalized prostanoid-mediated vaso-constriction and increased tissue ET-1 levels, increased vascular ET-1 concentrations may well contribute to endothelial vasoconstrictor prostanoid release in vivo. As observed previously in other vascular beds (9,39), contractions to exogenous ET-1 but not those to norepinephrine were reduced in the renal artery, and this was likely caused by agonist-induced receptor downregulation due to increased vascular ET-1 content. Again, treatment with LU135252 prevented both the increase of ET-1 protein and the attenuation of contractile responses.

Endothelin: Mediator of Renal Injury and Endothelial Dysfunction

In this study, we have demonstrated that both renal and vascular ET-1 protein levels, but not renal ET-3 protein, markedly increase in DS rats after salt loading. This was associated with renal artery hypertrophy and glomerulosclerosis in hypertensive DS rats, confirming previous reports (26,45). Although genetic analyses have linked the ET-3 gene locus (23) but not the ET-1 gene locus (46) to hypertension in DS rats, and despite altered renal ET-3 expression in other forms of renal injury (47), we found no evidence that ET-3 expression is modulated in hypertensive DS rats. Thus, ET-1 but not ET-3 appeared to be the likely mediator of salt-mediated injury. Consistent with this hypothesis, the structural and particularly the functional changes associated with hypertension were largely normalized by concomitant treatment with an orally active ET_A antagonist, which had only moderate effects on the salt-induced BP increase. It is noteworthy that no activation of the ET system was observed in DR rats in which NOS activity (in contrast to DS rats) almost tripled. We speculate that the failure to adequately increase NOS activity may facilitate activation of the ET system in hypertensive DS rats, which is regulated by NO (16,48). This “intrinsic” NO deficiency in DS rats, in turn, may promote the detrimental effect of ET-1 on renal structure and function. Indeed, chronic NO deficiency induced pharmacologically by L-NAME treatment is associated with hypertension (2) and with pronounced changes in renal morphology and function, which can be attenuated by ET receptor antagonism (11,49. One of the limitations of our study is that we did not use another antihypertensive agent to examine the effect of BP lowering. However, it is unlikely that normalization of the observed changes was due to the antihypertensive effects of LU135252, which lowered BP only in part. In line with this notion, a previous study demonstrated that antihypertensive therapy with the diuretic indapamide actually normalized BP, but did not provide effective renal protection (8). Unfortunately, these investigators did not examine vascular hypertrophy. Schiffrin and colleagues have suggested that the effects of ET-1 on vascular structure may be, at least in part, pressure-independent (50), which is in keeping with the present study and our observations in atherosclerotic mice (20).

We have reported previously that chronic ET_A receptor blockade normalizes elevated tissue levels of ET-1 protein in hypertension and atherosclerosis (9,20,24). We have confirmed these findings in the present study, in which ET-1 tissue levels in LU135252-treated animals were comparable to those seen in untreated DS control rats. The underlying mechanism is presently unknown. It is possible, however, that increased endothelial NO activity, which regulates ET-1 production (51) and was observed after LU135252 treatment, was involved. The effects of LU135252 may also involve ET_B receptor-mediated mechanisms such as NO formation, because ET_B blockade worsens renovascular function and injury in deoxycorticosterone acetate-salt hypertension (52). Other possible mechanisms include ET_A receptor-coupled autocrine regulation of ET-1 synthesis at the transcriptional (53) or protein (54,55) level.

Clinical Implications

Salt sensitivity in patients with essential hypertension is associated with increased cardiovascular risk (1) and is particularly common among African-Americans. These patients are often resistant to conventional antihypertensive therapy (56) and commonly develop hypertensive renal injury. Two recent studies in this population demonstrated both endothelial dysfunction (57) and increased renal vasoconstriction in response to salt loading (58), consistent with our observations in experimental salt-sensitive hypertension. The ET system appears to be activated in these patients because plasma levels of ET-1 are extraordinarily high (59). Therefore, it is reasonable to speculate that selective ET_A receptor antagonism will provide additional therapeutic benefit for the treatment of salt-sensitive forms of essential hypertension.

In conclusion, we have demonstrated: (1) a failure to increase NOS activity; (2) enhanced renal artery vasoconstriction due to selective endothelial dysfunction; and (3) a role for ET-1 as mediator of functional and structural changes that act as novel mechanisms in the pathogenesis of experimental salt-sensitive hypertension. Clinical studies will show whether ET_A receptor blockade has the therapeutic potential to inhibit endorgan damage in patients with salt-sensitive hypertension.

Note Added in Proof: While this manuscript was under revision, a study was published reporting that chronic inhibition of Nos2 by use of several isoform-specific enzyme inhibitors resulted in salt-induced hypertension without affecting endothelial function in resistance arteries in formerly salt-resistant DR rats (60).