Nephrol Dial Transplant (1996) 11: 2449-2452
Nephrology
Dialysis
Transplantation
Original Article
Concentration of dimethyl-L-arginine in the plasma of patients
with end-stage renal failure
R. J. MacAllister1, M. H. Rambausek3, P. Vallance2, D. Williams2, K.-H. Hoffmann3 and E. Ritz3
'Department of Pharmacology and Clinical Pharmacology, St George's Hospital Medical School, London,
Centre for Clinical Pharmacology, Cruciform Project, University College London Medical School, London, and
3
Department of Internal Medicine, Ruperto Carola University, Heidelberg, Germany
2
Abstract Abstract. A^A^dimethyl-L-arginine (asymmetric dimethyl-L-arginine; ADMA) and N0]^0
dimethyl-L-arginine (symmetric dimethyl-L-arginine;
SDMA) are naturally occurring analogues of L-arginine, the substrate for nitric oxide (NO) synthesis.
ADMA is a potent inhibitor of NO synthesis, and
accumulates in the plasma of patients with renal failure.
However the precise concentration of ADMA and
SDMA in renal patients is still controversial. This study
was performed to measure plasma ADMA and SDMA
concentrations by two different HPLC techniques in
nine healthy controls and 10 uraemic subjects, and to
investigate the effects of haemodialysis. In controls, the
mean (±SEM) plasma concentrations of ADMA and
SDMA were 0.36 + 0.09 and 0.39 + 0.05 umol/1 respectively, yielding an ADMA/SDMA ratio of 1.2 + 0.17. In
uraemic patients, the plasma concentrations of ADMA
and SDMA were 0.9 + 0.08 umol/1 (P < 0.001 compared
to controls) and 3.4 + 0.3 umol/1 (P<0.001 compared
to controls) with an ADMA/SDMA ratio of
0.27 ±0.015 (P<0.00l). In the course of one4h haemodialysis session, ADMA concentrations decreased from
0.99 + 0.13 to 0.77 + 0.3 umol/1 and SDMA concentrations from 3.38 + 0.44 to 2.27 + 0.21 umol/1. The plasma
ADMA/creatinine ratio tended to increase from 1.26 +
0.20 x 10~3 to 2.01 ±0.41 x 10~3. It is concluded that
there is a modest (3-fold) but definite increase in plasma
ADMA concentration in uraemic patients compared to
controls. SDMA accumulates to a greater degree
(8-fold increase) and more closely parallels creatinine
concentration than ADMA. The change in the
ADMA/SDMA ratio is not accounted for by greater
renal or dialysis clearance of ADMA, and, even though
alternative explanations are not excluded, greater metabolism of ADMA than SDMA is the most likely
explanation. Although small in magnitude, the increase
in ADMA concentration might be biologically
significant.
Correspondence
and offprint requests
to: R. J.
MacAllister,
Department of Pharmacology and Clinical Pharmacology, St
Georges Hospital Medical School, London, UK.
Key words: uraemia; dialysis; NO synthase; ADMA;
SDMA
Introduction
A^A^dimethyl-L-arginine (asymmetric dimethyl-L-arginine; ADMA) and N0]^0 dimethyl-L-arginine (symmetric dimethyl-L-arginine; SDMA) accumulate in the
plasma of patients with renal failure, and of these,
ADMA is a potent inhibitor of nitric oxide (NO)
synthase (NOS) [ 1 ]. This guanidine compound inhibits
each of the three isoforms of NOS (endothelial, neuronal and macrophage). It accumulates in the plasma of
patients with renal failure and might contribute to
certain pathophysiological features of this condition.
In the initial study [1], ADMA was measured using a
non-derivatized HPLC method. More recently,
Anderstamm et al. [2], using a derivatized method,
confirmed plasma ADMA concentration to be elevated
in dialysed patients and decreased by dialysis, although
the reported concentrations were lower than those
described previously [ 1 ]. Whilst Arese et al.[3] reported
that the majority of plasma samples of dialysed patients
inhibited each isoform of NOS in vitro, about 20% of
samples actually stimulated NOS. Noris et al. [4] also
suggested the presence in plasma of factors that
enhanced conversion of L-arginine to L-citrulline in
human umbilical vein endothelial cells, and there are
also preliminary reports of increased NO production
during dialysis, as assessed by NO measurements in
expiratory air [5]. Increased NO production was related
to intradialytic hypotension [6] and was reproduced in
vitro by blood-dialysis membrane interaction [7]. The
issue is further complicated by the heterogeneity of
guanidine compounds that accumulate in the plasma
of uraemic patients, certain of which (including methylguanidine) might also inhibit NOS [8].
Because of these discrepancies in the literature concerning the precise concentration of methylated
derivatives of L-arginine in renal failure, and because
of the considerable biological relevance of this issue,
© 1996 European Renal Association-European Dialysis and Transplant Association
R. J. MacAllister et at.
2450
we measured ADMA and SDMA in patients with endstage renal failure about to start dialysis, patients established on dialysis (before, during and after one dialysis
session) and healthy controls. Plasma samples were
analysed using two different methods, one identical to
that described previously [ 1 ] and a novel method (in
which methylated arginines were derivatized) developed for the measurement of the related methylarginine
(A^monomethyl-L-arginine; L-NMMA) that is in clinical trials for the management of septic shock.
Results
Concentration of ADMA and SDMA
The concentrations of ADMA, SDMA and L-arginine
in the plasma of healthy subjects (determined by the
non-derivatized HPLC method) were 0.36 + 0.09,
0.39±0.05 and 75.3 +13 umol/1, respectively (n = 9).
The mean ratios of ADMA to SDMA and ADMA to
L-arginine were 1.2 + 0.17 and 0.007 + 0.001, respectively (n = 9). In patients with end-stage renal failure, the concentrations of ADMA and SDMA were
higher (0.9 ±0.08 and 3.4 + 0.03 umol/1, respectively;
Subjects and methods
P< 0.001; «=10), but L-arginine concentration was
not significantly different (35.4 + 3.1 umol/1; /><0.8;
«=4). Subgroup analysis of uraemic plasma showed
Design of the study
that the ADMA concentrations were similar in anuric
We studied three groups of individuals: nine healthy subjects patients (1.0 + 0.23 umol/1; « = 3), in patients with
(three female, six male, mean age 24 + 4 years), four subjects residual urine output (0.98 + 0.18 umol/1; « = 3) and in
(two male, two female, mean age 64 + 4.9 years) with end- pre-dialysis patients (0.77 + 0.08 umol/1; ;? = 4; /">0.05
stage renal failure who were being prepared to start dialysis by ANOVA).
(S-creatinine > 500 umol/1), and six dialysis patients (four
The plasma ADMA/SDMA ratio was reduced
male, two female, median age 66.5 + 7 years). The latter
patients had been maintained on haemodialysis for a median (0.27 + 0.015; /><0.001; n=10) and the A D M A / L of 2.0 years (range 1-7) and were studied during one 4h arginine ratio was increased (0.02 + 0.001; / > <0.01;
dialysis session (measured blood flow 200 ml/min; polysul- n = 4). The clearance of ADMA, SDMA and creatinine
fone dialyser F8 (1.8 m2 surface); dialysate flow 0.5 1/min; in patients with renal failure was similar (5.7 + 2.8,
dialysate with 32 mmol/1 HCO3; Braun Secura HD machine). 5.0+ 2.6 and 5.1 + 1.3 ml/min, respectively; P>0.5; « =
Heparinized blood samples were drawn at times 0, 1, 2 and 4). Subgroup analysis of the ADMA/SDMA ratios
4h after beginning of dialysis, and 0.5, 1, 2, 3 and 6h
following dialysis. Twenty-four hour urine samples were in uraemia revealed no differences between anuric
collected in the four pre-dialysis patients and on the day patients (0.29+0.04; n = 3), patients with residual urine
prior to dialysis in three uraemic patients with residual urine output (0.27 + 0.04; n = 3) or pre-dialysis patients
(0.24 + 0.025; n = 4; P>0.5 by ANOVA).
output (three patients on dialysis were anuric).
In three subjects the ADMA concentrations determined by the non-derivatized and derivatized HPLC
HPLC methods
method were (respectively) 0.6 + 0.05 and 0.4 +
0.05 umol/1 (P<0.05), the SDMA concentrations were
Non-derivatized method. Samples were purified by sequential 1.4 + 0.2 and 1.3 + 0.3 umol/1 (/>>0.1) and the
passage through Bond Elut SCX and CBA columns (Jones ADMA/SDMA ratios were 0.42 + 0.06 and 0.35 + 0.08
Chromatography Ltd, Hengoed, Mid Glamorgan, Wales, (/>>0.1) (Fig. 1).
UK), and eluted with ammonia/methanol [1]. Separation of
amino acids was achieved on an ODS C18 HPLC column
(Phase Separation, Queensferry, Wales, UK) using a mobile
phase containing 0.025 mol/1 phosphoric acid (pH 5.0),
0.01 mol/1 hexane sulphonic acid (Romil Laboratories,
Loughborough, UK) and 1% (v/v) acetonitrile (Romil
Laboratories). Flow was maintained at 1 ml/min and amino
acids were detected by UV absorbance at 200 nm.
Derivatized method. Samples were purified by centrifugation
(Millipore Ultrafree MC 5 kDa filters) and derivatized using
o-phthalaldehyde [9]. Separation of amino acids was achieved
on an SGE C18 HPLC column by gradient elution with
acetonitrile, the second solvent containing ammonium acetate
(25 mmol/1), citric acid (10 mmol/1) and potassium chloride
(330 umol/1; pH 5.5). Derivatized amino acids were detected
by electrochemical oxidation [9].
ADMA and SDMA concentrations were determined by
reference to synthetic standards.
Statistics
Data are expressed as means+ SEM. Comparisons were
made using Wilcoxon's test for paired or unpaired samples
or by analysis of variance (ANOVA) as appropriate.
Change ofplasma ADMA and SDMA concentrations
during one 4 h haemodialysis session (Table 1 and Figs 2
and 3)
As shown in Table 1, a modest decrease of ADMA
concentration was seen at the end of the 4 h dialysis
session, but this was not statistically significant. There
was, however, a marked and significant increase in the
plasma ADMA/creatinine and SDMA/creatinine
ratios. There was a rapid rebound of plasma ADMA
and SDMA concentrations after dialysis.
Discussion
The results of this study demonstrate that in uraemia,
the plasma concentrations of ADMA and SDMA are
higher than in controls. This observation is consistent
with previous reports [1], although the levels found in
the present study are somewhat lower. Dialysis was
associated with a modest decrease in the total concentration of dimethylarginines. The concentrations of
2451
Dimethyl-L-arginine in end-stage renal failure
0.03 n
0.02 O
C
0.01 o
0 2
< C/5
-0.01 15
10
20
25
30
Time (min)
Fig. 1. Chromatogram of human plasma. Non-derivatized amino acids are detected by UV absorbance at 200 nm. The amounts of arginine,
ADMA and SDMA are derived from the area under the chromatogram peaks with reference to synthetic standards; in this chromatogram
of healthy human plasma, the areas under the ADMA and SDMA peaks are similar.
molar ratio of plasma
ADMA/creaxlCf3
molar ratio of plasma
SDMA/creaxlO"3
ADMA/ena ratio
200
-10
8DMA/«raa rttlo
2-
-5
100-
60
post
pre
dialysis
post
pre
dialysis
Fig. 2. Molar ADMA/creatinine and SDMA/creatinine ratios in six
haemodialysis patients before and after one 4 h haemodialysis.
ADMA and SDMA pre- and during dialysis are consistent with those reported by Arese et al. [3]. In view
of the controversy concerning the absolute values [2]
it is of note that, in this study, the values obtained
with the original HPLC method from partially purified
plasma [ 1 ] were largely confirmed in this study by an
independent derivatized method, developed to measure
methylarginines for drug regulatory purposes.
SDMA accumulated to a greater extent (8-fold) in
uraemia than ADMA (3-fold), and this is reflected by
the significantly lower plasma ADMA to SDMA ratio
in renal patients compared to controls. The change in
ADMA/SDMA ratio is not accounted for by differ-
0
1 2
3 4
end of dialysis
8
9
10
ft
12
(h)
Fig. 3. Time course of the mean plasma ADMA and SDMA concentrations as well as molar ADMA/creatinine and SDMA/creatinine
ratios during and after one 4 h haemodialysis session.
ences in the renal clearance of the two enantiomers,
since this ratio was similar in patients with renal failure
not on dialysis, and the clearance of ADMA and
SDMA by dialysis appeared to be similar. Although
our data do not exclude the possibility that more
SDMA is synthesized, the most likely explanation is
selective metabolism of ADMA but not SDMA. Indeed
the enzyme dimethylarginine dimethylaminohydrolase
(DDAH) that metabolizes ADMA but not SDMA is
found in a variety of human tissues, including kidney, pancreas and human blood vessels [10,11].
Consequently, plasma concentrations of ADMA are
R. J. MacAllister et al.
2452
Table 1. Plasma ADMA and SDMA concentrations in six haemodialysis patients before and after one 4 h haemodialysis session
ADMA(umol/l)
Molar ADMA/creatinine
ratio (x 10~3)
SDMA
Molar SDMA/creatinine
ratio (x 10~3)
Pre-dialysis
Post-dialysis Difference*
0.99 ±0.13
1.25±0.20
0.77 ±0.03
2.01 ±0.41
NS**
0.05
3.78 + 0.44
4.24±0.56
2.23 ±0.25
5.71+0.73
NS
0.05
Means±SEM.
*Pre-dialysis vs post-dialysis (Wilcoxon's test for paired samples).
"Difference controls vs dialysis patients (Wilcoxon's test for random
samples), /><0.05.
presumably not a passive reflection of decreased glomerular filtration rate, since the kidney (and other tissues) might play an active role in ADMA metabolism.
The wide range of ADMA/SDMA ratios measured
(0.15-0.5) is consistent with variable DDAH activity
in patients with renal failure. The existence of an
enzyme that metabolizes only methylarginines that are
potent NO synthase (NOS) inhibitors highlights the
potential biological importance of ADMA. A rapid
rebound increase in dimethylarginine concentration
was observed after dialysis, consistent with either
higher production rates after dialysis or, more likely,
re-equilibration of the plasma with other
compartments.
Whether or not the plasma of uraemic subjects alters
nitric oxide production is still controversial [4,5], but,
at least in principle, accumulation of ADMA would
have the potential to inhibit NO synthesis. Although
the recorded concentrations are low (1 umol range),
very low (sub-pressor) concentrations of NOS inhibitors impair renal function and cause salt-sensitive
hypertension [12] and promote atherogenesis [13].
Low micromolar concentrations of L-NMMA (that is
equipotent with ADMA) increase vascular tone and
blood pressure of patients with septic shock [14]. In
addition, plasma ADMA concentrations might not be
representative of local concentrations, similar to the
situation with circulating concentrations of endothelin.
ADMA and SDMA are synthesized by human endothelial cells [11] and extracellular ADMA is concentrated within cells (approximately 5-fold) [11]. Another
possibility to consider is that other guanidinocompounds that inhibit NOS might also accumulate
in renal failure [8]. In addition the net effect of NOS
inhibitors depends on the availability of the substrate
for NO synthesis, L-arginine, and it is of note that the
increase in the concentration of ADMA is mirrored
by a similar increase in the plasma ADMA/L-arginine
ratio. Indeed, L-arginine causes stereospecific vasodilatation in the kidney [15] and ameliorates renal damage
in animal models [16], observations that are consistent
with competition for NOS between L-arginine and
inhibitors such as ADMA.
The following conclusions can be drawn from the
above observations.
1. Dimethylarginines accumulate in renal disease and
such increase is cross-validated by comparison of
two different analytical methods.
2. As an NOS inhibitor, endogenous ADMA might
contribute to the pathophysiology of renal dysfunction and complications of renal failure.
3. The more marked accumulation of SDMA compared to ADMA points to a potential role for
DDAH in maintaining low concentrations of
ADMA in renal failure.
The latter aspect warrants further investigation.
References
1. Vallance P, Leone A, Calver A et al. Accumulation of an
endogenous inhibitor of nitric oxide synthesis in chronic renal
failure. Lancet 1992; 339: 572-575
2. Anderstam B, Katzarski K, Bergstrom J. Methylarginines in
uraemia. JASN 1994; 5: 572a
3. Arese M, Strasly M, Ruva C, Costamagna C, Ghigo D,
MacAllister R, Verzetti G, Tetta C, Bosia A, Bussolini F.
Regulation of nitrix oxide synthesis in uraemia. Nephrol Dial
Transplant 1995; 10: 1386-1397
4. Noris M, Benigni A, Boccardo P et al. Enhanced nitric oxide
synthesis in uraemia: implication for platelet dysfunction and
dialysis hypotension. Kidney Int 1993; 44: 445-450
5. Vimay P, Madore F, Prudhomme L, Austin JS, Blaise G,
Prudhomme M. Nitric oxide in expiratory air in dialysed
patients. JASN 1995; 6: 590a
6. Yokokawa K, Mankus R, Saklayen MG, Kohno M, Yasunan K,
Minami M, Kano H, Horio T, Takeda T, Mandel AK. Increased
nitric oxide production in patients with hypotension during
haemodialysis. Ann Int Med 1995; 123: 35-37
7. Anore A, Bonaudo R, Ghigo D, Arese M, Costamagna C,
Cirina P, Gianoglio B, Perugini L, Coppo R. Enhanced production of nitric oxide by blood-dialysis membrane interaction.
J Am Soc Nephrol 1995; 6: 1278-1283
8. MacAllister RJ, Whitley GStJ, Vallance P. Effect of guanidino
and uraemic toxins on nitric oxide pathways. Kidney Int 1994;
45: 737-742
9. Canevari L, Vieira R, Aldegunde M, Dagani F. High performance liquid chromatographic separation with electrochemical
detection of amino acids focusing on neurochemical application.
AnalBiochem 1992; 205: 137-142
10. Kimoto M, Whitley GStJ, Tsuji H, Ogawa T. Detection of
A^A^-dimethylarginine dimethylaminohydrolase in human tissues using a monoclonal antibody. J Biochem 1995; 117: 237-238
11. MacAllister RJ, Fickling SA, Whitley SJ, Vallance P.
Metabolism of methylarginines by human vasculature; implications for the regulation of nitric oxide synthesis. Br J Pharmacol
1994; 112:43-48
12. Salazar FJ, Pinilla JM, Lopez F, Romero JC, Quesada T. Renal
effects of prolonged synthesis inhibition of endothelium-derived
nitric oxide. Hypertension 1992; 20: 113-117
13. Cayatte AJ, Palacino JJ, Horten K, Cohen RA. Chronic inhibition of nitric oxide production accelerates neointima formation
and impairs endothelial function in hypercholesterolaemic
rabbits. Arterioscler Thromb 1994; 14: 753-759
14. Petros A, Lamb G, Leone A, Moncada S, Bennett D, Vallance P.
Effects of a nitric oxide synthase inhibitor in humans with septic
shock. Cardiovasc Res 1994; 28: 34-39
15. Kumagai K., Suzuki H, Ichikawa M, Jimbo M, Murakami M,
Ryuzaki M, Saruta T. Nitric oxide increases renal blood flow
by interacting with the sympathetic nervous system. Hypertension
1994; 24: 220-226
16. Reyes AA, Karl IE, KJahr S. Role of arginine in health and
renal disease. Am J Physiol 1994; 267: F331-F346
Received for publication: 29.5.96
Accepted in revised form: 27.7.96