Total nitric oxide production is low in patients with chronic renal disease

Abstract

Total nitric oxide production is low in patients with chronic renal disease.

Background

A deficiency of the endogenous vasodilator nitric oxide (NO) has been implicated as a potential cause of hypertension in chronic renal disease (CRD) patients. This study was conducted to determine whether 24-hour NO_X (NO₂ and NO₃) excretion (a qualitative index of total NO production) is reduced in patients with CRD.

Methods

Measurements were made in 13 CRD patients and 9 normotensive healthy controls after 48 hours on a controlled low-NO_X diet. Urine was collected over the second 24-hour period for analysis of 24-hour NO_X, and cGMP and blood drawn at the completion. Plasma levels of arginine (the substrate for endogenous renal NO synthesis), citrulline (substrate for renal arginine synthesis), and the endogenous NO synthesis inhibitor asymmetrical dimethylarginine (ADMA) and its inert isomer and symmetrical dimethylarginine (SDMA) were also determined.

Results

Systolic blood pressure was higher in CRD patients (12 of whom were already on antihypertensive therapy) than in controls (P < 0.05). Twenty-four–hour urinary NO_X excretion was low in CRD patients compared with controls despite similar dietary NO intake, suggesting that net endogenous NO production is decreased in renal disease. In contrast, the 24-hour urinary cGMP did not correlate with U_NOXV. Plasma citrulline was increased in CRD patients, possibly reflecting reduced conversion of citrulline to arginine. Plasma arginine was not different, and plasma ADMA levels were elevated in CRD versus controls, changes that would tend to lower NO synthase.

Conclusion

These results suggest that NO production is low in CRD patients and may contribute to hypertension and disease progression in CRD.

It has been suggested that the physiologically important vasodilator nitric oxide (NO) is deficient in chronic progressive renal disease (CRD) and in end-stage renal failure (ESRD) [1, 2]. This could result from substrate (arginine) deficiency [3] caused by a loss of functional renal mass, increased endogenous NO synthase (NOS) inhibitors that accumulate in renal failure [2], and/or other causes, such as increased oxidant stress [4]. In addition to being caused by CRD, low NO production may contribute to and/or exacerbate the progression of CRD by both hemodynamic and renal growth-promoting actions [5].

The plasma concentration and urinary excretion of NO₂ and NO₃ (NO_X; the stable oxidation products of NO) are now being widely used to give a measure of total NO production in vivo. In rats with subtotal nephrectomy, 24-hour NO_X excretion is reduced [6–8], suggesting reduced total NO production (assuming that NO_X intake is constant). Reduced urinary NO_X excretion has also been reported in children with CRD [9], in adults with IgA nephropathy and hypertension [10], and in adults with CRD caused by a variety of causes [11]. Although suggestive of decreased total NO production in CRD, one concern is that no clinical study to date has employed dietary control with a low NO_X intake. As we have discussed previously, controlled NO_X intake is essential for interpretation of NO_X excretion data [12]. Accordingly, we conducted the present study during low and controlled dietary NO_X intake to determine whether 24-hour NO_X excretion (a qualitative index of total NO production) is reduced in patients with CRD of various etiologies.

METHODS

Thirteen CRD patients were asked to participate in the study. Nine normotensive, healthy subjects were recruited from spouses of CRD patients and from West Virginia University Health Science Center faculty, students, and staff to serve as the control group. Controls were required to be normotensive with no known illnesses, and they could not be taking cardiovascular medications. Of the 13 CRD patients, the causes of renal disease were variable (Table 1), and 12 out of 13 were on one or more antihypertensive medications (Table 1); however, no patient was on nitrates.

Table 1.

Demographic and clinical characteristics of the study population

CRD patient #	Age	Sex	BSA	CCr	Disease	Medications
1	29	M	1.67	46	SK	None
2	51	M	2.22	23	Wegener's	BB, CCB, MIN
3	56	M	1.75	23	CIN	CAD, CCB, D
4	58	M	2.30	19	GN	BB, CCB, CAD
5	62	F	2.59	46	DM	BB, CEI, D
6	63	F	2.60	31	DM	CEI, D
7	65	M	2.10	32	FSGS	CAD, CCB, D
8	67	M	1.92	40	DM	CEI, D
9	74	F	1.81	28	DM	CCB, D
10	78	M	1.84	19	DM	D
11	59	M	2.24	46	IgA+SK	CEI
12	32	F	1.34	10	Obstr Nephr	CEI
13	67	M	2.19	44	IgA	CEI, D
All CRD (N = 13)	59±4	4F:9M	2.04±0.10	31±3a
CRD subgroup (N = 6)	57±7	2F:4M	1.98±0.18	33±6a	3DM,1SK,1IgA + SK, Obstr Nephr
Control (N = 9)	59±6	5F:4M	2.07±0.11	124±21	None	None

Twelve of 13 CRD patients were on one or more antihypertensive agents. Abbreviations are: BB, beta blocker; CAD, centrally acting adrenergic blocker; CCB, calcium channel blocker; CEI, angiotensin-converting enzyme inhibitor; D, diuretic; MIN, minoxidil; M, male; F, female; FSGS, focal and segmental glomerulosclerosis; SK, solitary kidney; CIN, chronic interstitial nephritis; GN, glomerulonephritis; DM, diabetic nephropathy; Obstr Nephr, obstructive nephropathy; C_Cr, creatinine clearance; BSA, body surface area. Patients on nitrates were excluded from the study.

^aP < 0.05 vs. control

All participants were provided with complete meals containing low NO_X (330 μmol/day, approximately 25% of normal daily dietary NO₃ intake) [13]. Supplemental calories, if required, were provided in the form of zero NO_X, zero sodium content foods. Subjects were given verbal and written instructions at the beginning of the diet period, and were instructed to eat only what food was provided during this period and to report any un-eaten food. Subjects were again questioned at the end of the 48-hour dietary period to assure compliance. Subjects collected all urine during the second 24-hour period of the low-NO_X diet. Blood pressure was measured (2 to 3 estimates at 5-minute intervals while subjects were quietly seated), and a blood sample was obtained at the end of the 48-hour dietary period, before 10 a.m., and after a fast of at least 12 hours, coincident with the end of the 24-hour urine collection. As discussed previously, subjects come into NO_X balance within 24 to 48 hours of a controlled NO_X diet [12].

Sample collection and preparation

Urine was collected for 24 hours into sterile, dry, plastic containers with 2% boric acid as an antibacterial agent. Urine volumes were measured and samples aliquoted and stored frozen for later analysis. Blood samples were taken into chilled tubes + ethylenediaminetetraacetic acid (EDTA) and were centrifuged at 4°C, and plasma was aliquoted, frozen, and stored at –80°C for later analysis, except for endogenous methylarginine analysis, in which heparin was used as the anticoagulant. The plasma to be used for cGMP analysis was stored with the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (IBMX; 30 μL of 10 mmol/L IBMX to 500 μL plasma) to prevent the breakdown of cGMP in vitro.

Analyses

The following analyses were conducted: NO_X concentration of plasma and urine was measured using the Griess assay after conversion of NO₃ to NO₂ with the NO₃ reductase enzyme, as we previously described [14]. Plasma and urine cGMP were measured by the Cayman Chemicals ELISA kit. Plasma arginine and citrulline were measured by reverse-phase high-pressure liquid chromatography (HPLC) with precolumn derivatization and fluorescence detection using a modification of the AccQ Tag system for amino acid analysis (Waters, Milford, MA, USA) as we described previously [25]. To measure the endogenous NOS inhibitors asymmetrical dimethylarginine (ADMA) and symmetrical dimethylarginine (SDMA), reverse-phase HPLC was used as described by Anderstam, Katzarski, and Bergstrom with minor modifications as we previously described [16, 25]. Recovery was 98 ± 4% (on plasma spiked with 2.5 μmol/L, N = 4). Replicate measurements (N = 5) of a 4 μmol/L sample gave a coefficient of variation of 5.7%. Interassay variability (N = 10 runs) for a plasma sample spiked with 2.5 μmol/L was 9.2%. Because we established the HPLC method for ADMA after most of these studies were undertaken, ADMA, SDMA, and n-monomethyl l-arginine (L-NMA) were measured in only 11 out of 13 of the CRD patients. Some of these measurements were made at varying time intervals (the average was 7 months, range 0 to 14 months after the 24-hour U_NOXV collection, CRD subgroup). However all measurements were made simultaneously in six of the CRD patients.

Plasma and urine sodium and potassium were measured by flame photometer. Creatinine concentrations were done by colorimetric assay of the Janovski complex (Sigma kit #555-A; Sigma Chemical Co., St. Louis, MO, USA).

All data were reported as mean ± SE, and statistics were by paired and unpaired t-test and one-way analysis of variance (ANOVA). Statistical significance was defined where P < 0.05.

RESULTS

Demographics and clinical characteristics of the study population are presented in Table 1. The groups were of comparable age and body surface area (BSA), and the average creatinine clearance in the CRD patients was approximately 25% of control. The causes of CRD were varied, with five individuals having diabetic nephropathy, three with various glomerulonephritides (one status postnephrectomy), one with biopsy-proven primary focal segmental glomerulosclerosis, one with a solitary kidney, one with chronic interstitial nephritis, one with obstructive nephropathy, and one with Wegener's granulomatosis. CRD patients were proteinuric, whereas controls had normal urine protein excretion (Table 2). Systolic blood pressure was higher in CRD patients versus controls at the end of the 48-hour study period (Table 2) despite the fact that 12 out of 13 CRD patients were taking one or more antihypertensive medications at the time of study (Table 1). Urinary excretion of NO_X over 24 hours was significantly lower in the CRD group compared with the controls, and since intake was similar, the endogenous NO production was much lower in CRD patients than controls. As shown in Table 2, the values of 24-hour U_NOXV, U_cGMPV, protein excretion, and creatinine clearance were similar in the entire CRD group and the subgroup of CRD patients (N = 6) in whom all measurements were made contemporaneously. The 24-hour urinary excretion of cGMP was not measured in every subject because of analytical problems, but in seven CRD patients and seven controls, the 24-hour U_cGMPV was not statistically different, although numerically the values were increased approximately 200% in CRD. Low nonsignificance may have contributed to the lack of statistical significance in this group, but given the fall in U_NOXV, the predicted parallel relationship with U_cGMPV was clearly absent. In fact, there was no correlation between U_cGMPV and U_NOXV in control (r² = 0.3895, P = 0.84) or CRD (r² = 0.0038, P = 0.74; Table 2).

Table 2.

Blood pressure, renal function and 24 hour U_NOXV and U_cGMPV excretions in chronic renal disease patients and controls

	BP mm HG		24 h CCr mL/min	UprotV mg/24 h	UNOXV μmol/24 h	UcGMPV nmol/24 h
	Systolic	Diastolic
CRD (N = 13)	142 ± 7a	82 ± 4	31 ± 3a	3591 ± 1055a	410 ± 56a	473 ± 153b
CRD subgroup (N = 6)	137 ± 9	79 ± 1	33 ± 6a	3292 ± 1335a	322 ± 75a	549 ± 163
Controls (N = 9)	120 ± 5	73 ± 3	124 ± 21	63 ± 11	914 ± 129	148 ± 51b
P value	<0.05	NS	<0.001	<0.001	<0.001	NS

^aDifferent vs. control

^bN = 7

Plasma creatinine was higher in CRD patients compared with controls, reflecting a loss of renal clearance in the CRD group (Table 3). Plasma NO_X was not significantly higher in the CRD group. The increased plasma cGMP was at the borderline of significance (P = 0.05), and plasma citrulline values were elevated versus controls. Plasma arginine concentrations were similar in controls and CRD patients (Table 3), whereas P_ADMA and P_SDMA were elevated in CRD versus controls. Since plasma levels of L-NMA were not detectable in most samples, they are not reported. Although the average plasma creatinine and ADMA values were elevated in the CRD group, as shown in Figure 1, there was no correlation between individual plasma values of creatinine and ADMA in the CRD patients.

Table 3.

Plasma chemistry of CRD patients and controls

	PCr mg/dL	PNOX μmol/L	PcGMP pmol/mL	Parginine	Pcitrulline	PADMA	PSDMA
				μmol/L
CRD N = 13	3.1 ± 0.3a	62 ± 15	15 ± 2	67 ± 8b	63 ± 7ab	1.26 ± 0.30ac	0.60 ± 0.11ac
CRD subgroup N = 6	3.0 ± 0.4a	45 ± 12	13 ± 2	—	—	1.45 ± 0.50	0.60 ± 0.13a
Controls N = 9	0.8 ± 0.1	35 ± 5	8 ± 2	68 ± 10	30 ± 3	0.40 ± 0.10	0.14 ± 0.04

^aDifferent vs. control

^bN = 9

^cN = 10

Fig. 1.

Regression analysis showing lack of a relationship between plasma asymmetrical dimethylarginine (ADMA) and plasma creatinine levels in 10 chronic renal disease (CRD) patients (r² = 0.03).

DISCUSSION

A majority of patients with CRD suffer from hypertension, irrespective of the etiology of their renal disease. While hypertensive nephrosclerosis accounts for approximately 20 to 25% of all diagnoses leading to ESRD, patients with ESRD from diabetic nephropathy (35% of all ESRD and nearly 50% of our CRD patients) also usually have hypertension [17–19]. Uncontrolled hypertension has been associated with stroke and cardiovascular morbidity and mortality, and contributes greatly to morbidity in CRD patients, many of whom are diabetic and thereby likely to have accelerated coronary artery and atherosclerotic vascular disease [20].

Blood pressures were significantly higher in our clinically euvolemic CRD patients versus controls, despite the fact that 12 out of 13 were on antihypertensive therapy, suggesting a persistent hypertensive stimulus other than fluid overload. NO deficiency has been implicated as a potential cause of high blood pressure in CRD, and animal data suggest that net renal NO deficiency occurs and is pathogenic. For example, chronic dietary arginine supplementation is renoprotective in a range of CRD models, including ablation of renal mass, ureteral obstruction, and cyclosporine-induced renal damage [1, 21]. Chronic experimental NOS inhibition in normal rats produces hypertension and kidney damage, the severity of which increases with increasing levels of NOS inhibition [5]. Chronic NOS inhibition [22] and endogenous NOS deficiency [23] accelerate the rate of progression of renal ablation-induced CRD. Direct evidence of low intrarenal NOS activity and NOS expression has been reported in the 5/6 ablation model of progressive CRD [7, 8].

The finding that 24-hour NO_X output is low in our CRD patients suggests that total systemic NO production is also low in humans with impaired renal function. This is the first observation in subjects on controlled low NO_X intake, an essential feature of the experimental design if one attempts to use 24-hour NO_X excretion as a measure of total NO production [12]. Thus, we have established that total NO production is uniformly low in patients with CRD because of a variety of causes. The assumption that is widely promulgated in the literature is that low total NO production (from U_NOXV) reflects low cardiovascular/renal NO production, the “hemodynamically active” pool of NO. Unfortunately, this is only likely to be true if “total NO” and “hemodynamically active NO” always change in parallel. The contribution of the low levels of endothelial and intrarenal NO is likely to be overwhelmed by the high level of production from sites such as cerebellum, which do not have any obvious hemodynamic impact. Thus, although our observation of reduced 24-hour U_NOXV may reflect a deficiency of “hemodynamically active” NO in CRD, this measurement alone is merely suggestive.

Other measures that may also provide insight were examined in the present study. For example, plasma NO_X values are unchanged in CRD, which most likely reflects a combination of reduced total NO production and reduced renal clearance. In ESRD patients on dialysis, we find that plasma NO_X values are high (because of a complete loss of renal clearance) despite the low total NO production [24, 25]. Unfortunately, therefore, plasma NO_X levels, even when obtained under conditions of dietary NO_X control, do not give “stand-alone” information about NO production, particularly when renal function is impaired. The plasma cGMP levels (a major second messenger of NO) were borderline high in CRD versus controls, and there was no correlation between 24-hour U_cGMPV and U_NOXV, although U_cGMPV showed an anomolous, nonsignificant tendency to rise in CRD. There are several reasons for cGMP not being a reliable index of NO production: NO signals through other messengers. cGMP is also a second messenger to atrial natriuretic peptide. Tissue, rather than plasma, levels of cGMP correlate with NO deficiency [26]. Therefore, the cGMP level in body fluids is not readily interpretable in the context of “hemodynamically active NO” production.

We anticipate reduced arginine synthesis in CRD since the kidney (a major site of endogenous arginine production) is failing [1]. This is supported by the finding that citrulline levels (the precursor of arginine [3]) are markedly elevated in CRD, which may reflect reduced citrulline to arginine conversion as well as reduced renal clearance. However, these indirect indicators are not compelling evidence of a selective reduction in vascular/renal NO production. Plasma levels of the endogenous NOS inhibitor ADMA are reported to rise in CRD in humans and in rats [27, 28], and ESRD patients [26, 27], because it has been suggested to reduce renal clearance [2], although there is controversy over whether the increase in P_ADMA in ESRD is sufficient to influence NOS [16, 29]. An increase in circulating ADMA (if sufficiently great) will inhibit systemic endothelial NOS and renal NOS, leading to reduced “hemodynamically active” NO and, hence, hypertension and acceleration of the progression of renal disease. In the present study, plasma ADMA levels were significantly elevated in the CRD group, which may contribute to reduced endothelial NO synthesis. Our findings complement a recent study by Wever et al in which in vivo NOS activity was measured by determining the conversion of 15N-arginine to 15N-citrulline in a group of patients with CRD [30]. In this study, total body NO production (estimated from 15N-citrulline accumulation) was significantly reduced in CRD.

In addition to the increased plasma ADMA, plasma levels of SDMA were also increased in CRD. Although SDMA does not inhibit NOS activity, it does compete with the cationic amino acid transporter in the endothelial cell membrane [31], and this could worsen the situation by accentuating any intracellular arginine deficiency.

Of note, P_ADMA did not correlate well with creatinine clearance or plasma creatinine within the CRD group (Fig. 1), suggesting that factors in addition to reduced renal clearance contribute to the increased P_ADMA in CRD patients. Interestingly, a recent study reported an association between atherosclerotic disease with elevated P_ADMA in patients with normal renal function [32], and in vivo NOS activity tends to be low in atherosclerosis-prone individuals with hypercholesterolemia [30]. Hemodialysis-dependent ESRD patients with known atherosclerotic disease have higher P_ADMA values compared with ESRD patients on hemodialysis without known atherosclerosis [33]. Six of the CRD patients in the present study had either coronary artery disease (by catheterization or known acute myocardial infarction), cerebrovascular disease, or known peripheral vascular occlusive disease.

Another possible mechanism to account for increases in P_ADMA (in addition to reduced renal clearance) is a reduced activity of dimethylarginine dimethylaminohydrolase (DDAH), the enzyme responsible for degradation of ADMA. Indeed, decreased DDAH activity has been reported in hypercholesterolemic rabbits where ADMA is high despite normal renal function [34].

Taken together, these results suggest that a low total NO production occurs in CRD and may contribute to hypertension, atherosclerosis, and progression of renal disease. While most of the measured indices in our study are indicative of reduced “hemodynamically active” NO production, these measures remain indirect. Further studies of the predictive value of both P_ADMA and low 24-hour NO_X excretion in renal disease are warranted.