Anal. Chem. 1996, 68, 3520-3523
High-Performance Liquid Chromatographic Assay for the Quantitation of L-Arginine in Human Plasma Vidhya Gopalakrishnan, Pat J. Burton, and Terrence F. Blaschke*
Division of Clinical Pharmacology, Stanford University Medical Center, Stanford, California 94305-5113
L-Arginine is metabolized to nitric oxide by nitric oxide synthase, and abnormalities in nitric oxide production have been implicated in the pathogenesis of some diseases involving the vasculature. Thus, there has been interest in the effects of pharmacologic doses of L-arginine in patients with cardiovascular and renal diseases. To study the disposition of exogenous doses, an HPLC method was developed to analyze plasma samples for L-arginine. The assay involves precolumn derivatization of arginine with naphthalenedicarboxaldehyde and cyanide followed by HPLC with UV detection. Only a simple deproteinization of the plasma samples was required. The derivatized arginine was stable (less than 5% degradation in 20 h), facilitating batch sample processing and analysis in an autosampler. Calibration curves were generated in Ringer’s lactate solution instead of plasma to correct for endogenous plasma L-arginine. Recovery in plasma, compared to Ringer’s solution (n ) 4), was 103%. Mean intraday assay precision (n ) 6), expressed as coefficient of variation, was 3.4%. Interassay precision (n ) 6) was 7%. The assay was applied for the quantitation of Larginine in plasma samples from a normal subject who had been given a single oral (10 g) and a single intravenous dose (30 g) of exogenous L-arginine. L-Arginine
is a nonessential amino acid which has been recognized since the 1880s.1 L-Arginine is involved in several metabolic pathways including the synthesis of urea, creatine, and agmatine. L-Arginine is also the physiological precursor of nitric oxide,2 a mediator released by vascular endothelial cells3 that accounts for the biological activity of endothelium-derived relaxing factor.4,5 In animal4,6 and human7 tissues, NO is a potent vasodilator8 and an inhibitor of platelet aggregation9 and adhesion.10 NO is stereospecifically synthesized from L-arginine by the enzyme nitric oxide synthase, releasing l-citrulline as a byproduct.11 Abnormalities in nitric oxide production have been implicated in the pathogenesis of a number of diseases involving
the vasculature including hypertension,12,13 diabetes,14 and atheroma.15,16 It is hypothesized that provision of exogenous L-arginine might increase the synthesis of NO17 and consequently restore normal vascular function. Thus, there has been significant interest in the effects of pharmacologic doses of L-arginine in patients with cardiovascular and renal diseases.18-24 In order to study the disposition of exogenous doses, an HPLC method has been developed to analyze plasma samples for L-arginine. This article describes the development and validation of an HPLC assay for the quantitation of L-arginine in plasma. The assay involves precolumn derivatization of arginine with naphthalenedicarboxaldehyde (NDA) and cyanide followed by high-performance liquid chromatography (HPLC) with UV detection. No cleanup of the plasma samples other than a simple deproteinization was required. The derivatized arginine was quite stable (less than 5% degradation in 20 h), facilitating batch sample processing and analysis in an autosampler. Overall, the assay for quantitation of L-arginine in plasma samples, as described in this article, is fairly simple, rapid, inexpensive, and efficient for use in the analysis of a large number of samples. The assay was applied for the quantitation of L-arginine in plasma samples from a normal subject who had been given a single oral (10 g) and a single intravenous dose (30 g) of exogenous L-arginine. EXPERIMENTAL SECTION L-Arginine, l-norleucine, tetramethylammonium chloride, sodium tetraborate, sodium citrate, sodium cyanide, and Dowex H+ were from Sigma (St. Louis, MO). HPLC grade acetonitrile, methanol, sodium hydroxide, phosphoric acid, and perchloric acid were from J. T. Baker (Phillipsburg, NJ), NDA was from Molecular
* E-mail:
[email protected]. (1) Schulze, E.; Steiger, E. Z. Physiol. Chem. 1887, 11, 43-65. (2) Palmer, R.; Rees, D.; Ashton, D.; Moncada, S. Biochem. Biophys. Res. Commun. 1988, 153, 1251-1256. (3) Palmer, R.; Ashton, D.; Moncada, S. Nature 1988, 333, 664-666. (4) Palmer, R.; Ferrige, A.; Moncada, S. Nature 1987, 327, 524-526. (5) Furchgott, R.; Zawadzki, J. Nature 1980, 373-376. (6) Calver, A.; Collier, J.; Vallance, P. Clin. Sci. 1991, 81, 695-700. (7) Yang, Z.; von Segesser, L.; Bauer, E.; Stulz, P.; Turina, M.; Luscher, T. Circ. Res. 1991, 68, 52-60. (8) Moncada, S.; Radomski, M.; Palmer, R. Biochem. Pharmacol. 1988, 37, 2495-2501. (9) Radomski, M.; Palmer, R.; Moncada, S. Br. J. Pharmacol. 1990, 101, 325328. (10) Radomski, M.; Palmer, R.; Moncada, S. Lancet 1987, (2), 1057-1058.
(11) Palmer, R.; Moncada, S. Biochem. Biophys. Res. Commun. 1989, 158, 348352. (12) Winquist, R.; Bunting, P.; Baskin, E.; Wallace, A. J. Hypertens. 1984, 2, 541-545. (13) Lockette, W.; Otsuka, Y.; Carretero, O. Hypertension 1986, 8, II-61-II-66. (14) Kamata, K.; Miyata, N.; Kasuya, Y. J. Pharmacol. 1989, 97, 614-618. (15) Ludmer, P.; Selwyn, A.; Shook, T.; Wayne, R.; Mudge, G.; Alexander, R.; Ganz, P. N. Engl. J. Med. 1986, 315, 1046-4051. (16) Chester, A.; O’Neil, G.; Moncada, S.; Tadjkarimi, S.; Yacoub, M. Lancet 1990, 336, 897-900. (17) Girerd, X.; Hirsch, A.; Cooke, J.; Dzau, V.; Creager, M. Circ. Res. 1990, 67, 1301-1308. (18) Baudouin, S.; Bath, P.; Martin, J.; Du Bois, R.; Evans, T. Br. J. Clin. Pharmacol. 1993, 36, 45-49. (19) Cohen, R.; Zitnay, K.; Haudenschild, C.; Cunningham, L. Circ. Res. 1988, 63, 903-910. (20) Drexler, H.; Zeiher, A. Hypertension 1991, 18, II-90-II-99. (21) Creager, M.; Gallagher, S.; Girerd, X.; Coleman, S.; Dzau, V.; Cooke, J. J. Clin. Invest. 1992, 90, 1248-1253. (22) Drexler, H.; Fischell, T.; Pinto, F.; et al. Circulation 1994, 89, 1615-1623. (23) Zeiher, A.; Drexler, H.; Wollschlager, H.; Just, H. Circulation 1991, 83, 391-401. (24) Cooke, J.; Andon, N.; Girerd, X.; Hirsch, A.; Creager, M. Circulation 1991, 83, 1057-1062.
3520 Analytical Chemistry, Vol. 68, No. 19, October 1, 1996
S0003-2700(96)00613-0 CCC: $12.00
© 1996 American Chemical Society
Probes (Eugene, OR), and lactated Ringer’s solution was from Abbott Laboratory (Chicago, IL). [3H]-L-arginine was from Moravak Biochemicals (Brea, CA). All water was deionized and purified with a Milli-Q water system (Millipore, Bedford, MA). Recovered plasma, for the control reactions, was obtained from the Stanford Blood Center. Borate buffer (0.1 M, pH 9.2) was prepared by dissolving the required amount of sodium tetraborate in water. Sodium citrate buffer (20 mM) containing tetramethylammonium chloride (5 mM) was made up in water and the pH adjusted to 2.8 with phosphoric acid. NDA (0.1 M) was prepared in methanol, and sodium cyanide (0.1 M) was made up in borate buffer. NDA and cyanide were freshly made on a daily basis. A 10 mM solution of l-norleucine was made up in water. Preparation of Calibration Standards. L-Arginine (free base, 0.5 g) was dissolved in 100 mL of water to give a stock solution of 5000 µg/mL. This was then serially diluted with Ringer’s solution to give standard solutions of 0, 10, 25, 50, and 100 µg/mL. All dilutions were made in volumetric flasks. Where plasma calibration curves were required, similar serial dilutions were made with recovered plasma obtained from the blood bank. Sample Treatment. To 125 µL of plasma (containing arginine) in a 1.7 mL Eppendorf tube was added 10 µL of 10 mM l-norleucine followed by 125 µL of perchloric acid (PCA, 10%). The plasma was vortexed and centifuged (Eppendorf microfuge Model 5414 at 15600g) for 10 min. Supernatant (230 µL) was pipetted and transferred to another Eppendorf tube. This was neutralized with NaOH (63 µL, 2.0 M). Samples containing more than 100 µg/mL arginine were diluted 10-fold with Ringer’s solution and assayed by an identical procedure. Ringer’s solution was similarly treated, except that the PCA and the NaOH were replaced with water. Derivatization Reaction. To the above Ringer’s solution or neutralized plasma (293 µL) was added 187 µL of 0.1 M borate buffer (pH 9.2) followed by 10 µL of sodium cyanide (0.1 M) and 10 µL of NDA (0.1 M), respectively. The reaction was incubated at room temperature for 30 min. The occurrence of the reaction is visually obvious by the appearance of a fluorescent green color. The absence of the fluorescent green color or the appearance of any other color is indicative of an incorrect or incomplete reaction. In this event, the borate, NDA, and cyanide need to be freshly remade. A 40 µL aliquot of the reaction was added to 60 µL of HPLC eluent A containing 30% acetonitrile in an amber HPLC vial (with an insert), and the entire volume (100 µL) was injected on the HPLC. Data Analysis. The ratio of the peak area of the derivatized arginine to that of the internal standard (peak area ratio, PAR) was plotted as a function of concentration of L-arginine to obtain slopes. For estimating arginine amounts in unknown samples, the normalized PAR (PARnorm) for the calibration standards was calculated by dividing the PAR from the standards by the respective arginine concentration [PAR/(µg/mL)]. The mean and coefficient of variation of these normalized PARs were determined. The PAR of the sample divided by the mean PARnorm of the calibration standards gave the concentration of arginine in the unknown samples in micrograms per milliliter. The intraday assay precision was determined on six sets of plasma samples at four different arginine concentrations, one with no added arginine and the others at 10, 50, and 250 µg/mL added
arginine. The result was expressed as the coefficient of variation of the PARs. For interday assay precision, the coefficient of variation of the normalized PARs for six calibration curves in Ringer’s solution and in blood bank plasma on six different days was evaluated. Stability of the derivatized arginine was estimated by repeated injection of the same reaction (in Ringer’s solution and in plasma) over a 20 h time period. The percent difference in PAR between the first and last injection was used as a measure of the stability of the product. Assay recovery was assessed by simultaneously performing the assay with the calibration standards and control plasma spiked with arginine such that the added concentrations were equivalent to that in the calibration standards. The percent ratio of the normalized PARs in plasma corrected for the endogenous arginine concentration to that in Ringer’s solution gave the percent recovery in plasma. (The intercept obtained on plotting absolute arginine peak areas as a function of arginine concentration was used as the measure of endogenous arginine concentration in plasma.) HPLC. HPLC was performed on a HP 1050 series system (Hewlett Packard, Mountain View, CA) equipped with an autosampler. The instrument was controlled by Chemstation software (Hewlett Packard) in a Windows environment. Fluorescent detection was through a Jasco FP-920 detector (Jasco Corp., Tokyo, Japan), and the resulting chromatogram was plotted on a SP4290 integrator (Spectraphysics, San Jose, CA). The runs were performed on a Waters Nova-Pak C-18 reversedphase column, 4.6 × 250 mm, 4 µm particle size packing, and a guard column packed with the same material. HPLC elution was done with eluent A, containing 20 mM sodium citrate and 5 mM tetramethylammonium chloride adjusted to pH 2.8 with phosphoric acid and eluent B, which was 60% acetonitrile. The following gradient was used: 0-3 min 100% A; 8 min 30% B; 24 min 60% B; 30 min 80% B; 35 min 90% B; 38 min 100% B; 44 min 100% B; 46-52 min 100% A; flow rate 1.0 mL/ min. Sample Collection from a Human Subject. The plasma samples were collected before and up to 24 h after separate administration of oral (10 g) and intravenous (30 g) L-arginine to a normal subject. The plasma samples were collected in heparinized tubes and stored at -70 °C. RESULTS AND DISCUSSION The most commonly used methods for the analysis of amino acids have been (a) gas chromatography (GC),25 (b) ion-exchange chromatography of the amino acids followed by postcolumn derivatization with ninhydrin (amino acid analyzer),26 and (c) HPLC with pre- or postcolumn derivatization and detection by fluorescence,27 electrochemical,28 or chemiluminescence29 detectors. Specifically, the analysis of L-arginine in plasma or serum samples has been achieved by (i) a fluorometric method,30 (ii) HPLC with fluorescent detection,31 (iii) GC with (a) flame (25) Frank, H.; Rettenmeier, A.; Weicker, H.; Nicholson, G,; Bayer, E. Clin. Chim. Acta 1980, 105, 201-211. (26) Spackman, D.; Stein, W.; Moore, S. Anal. Chem. 1958, 30, 1190-1205. (27) de Montigny, P.; Stobaugh, J.; Givens, R.; Carlson, R.; Srinivasachar, K.; Sternson, L.; Higuchi, T. Anal. Chem. 1987, 59, 1096-1101. (28) Lunte, S.; Mohabbat, T.; Wong, O.; Kuwana, T. Anal. Biochem. 1989, 178, 202-207. (29) Kwakman, P.; Koelewijn, H.; Kool, I.; Brinkman, U.; DeJong, G. J. Chromatogr. 1990, 511, 155-166. (30) Miura, T.; Kashiwamura, M.; Kimura, M. Anal. Biochem. 1984, 139, 432437.
Analytical Chemistry, Vol. 68, No. 19, October 1, 1996
3521
ionization detection32 and (b) mass spectrometric detection,33 and (iv) an enzymatic assay.34 GC has not been routinely used due to problems with precision and reproducibility,25 while use of an amino acid analyzer is expensive and has lower sensitivity of detection.35 The enzymatic assay involved several reaction steps36 and was not simple enough for routine analysis. We were interested in exploring the use of LC with UV or fluorescent detection as the method of choice for the quantitation of derivatized L-arginine in plasma samples. The most commonly used derivatizing agent for the detection of trace amounts of amino acids has been the primary aminespecific o-phthalaldehyde (OPA).37,38 Optimal use of this reagent for batch analysis of samples is achieved by automation of the derivatization reaction. The NDA-CN reagent, developed as an improvement over the OPA, forms relatively more stable adducts with a high quantum efficiency for fluorescent detection.27,39 In view of its greater stability, we decided to explore the use of the NDA-CN reagent for the derivatization and analysis of arginine in our study. The NDA-CN reaction, like the OPA reaction, is nonspecific in that it reacts with any primary amine to form a fluorescent N-substituted 1-cyanobenz[f]isoindole (CBI) derivative.27 Since interference from other endogeneous amino acids and amines in plasma would be an issue, we were also interested in a derivatization strategy that would be more specific for arginine. It has been reported that NDA reacts specifically with arginine in the presence of β-cyclodextrin to form a fluorescent complex with λex ) 462 nm, λem ) 520 nm.30 We explored the use of the NDA-β-cyclodextrin derivatization reactions as well as the NDA-CN reactions for the quantitation of arginine in plasma, and the results of our investigations were (1) the NDAβ-cyclodextrin derivatization reaction produced more peaks on HPLC than expected if the reaction was highly specific for arginine. Additionally, there was a peak that eluted very close to the derivatized arginine peak in plasma that proved difficult to resolve. The NDA-CN derivatization reaction did yield several peaks as expected, but the derivatized arginine and norleucine (used as internal standard) peaks could be well resolved from the other peaks by the use of an acidic mobile phase. (2) The sensitivity of the reaction (UV detection at 260 nm and fluorescent detection) in the latter case was at least 1 order of magnitude higher than that of the former. On the basis of these results, it was decided to use the NDA-CN derivatization for the analysis of arginine in plasma. Assay Method. The method described here for the quantitation of L-arginine in plasma involves (1) deproteinization of the arginine-containing plasma samples by PCA (10%) and precipitation and neutralization of the resultant protein-free plasma with sodium hydroxide, (2) precolumn derivatization of the plasma samples by cyclization with NDA and CN reagents in borate buffer (pH 9.2), and (3) reversed-phase HPLC of the derivatized plasma with UV detection at 260 nm. (31) Mitchell, J.; Hecker, M.; Anggard, E.; Vane, J. Eur. J. Pharmacol. 1990, 182, 573-576. (32) Williams, A. Biochem. Pharmacol. 1993, 46, 2097-2099. (33) Castillo, L.; Ajami, A.; Branch, S.; Chapman, T.; Yu, Y.-M.; Burke, J.; Young, V. Metabolism 1994, 43, 114-122. (34) van Haeften, T.; Konings, C. Clin. Chem. 1989, 35, 1024-1026. (35) Williams, A. J. Chromatogr. 1986, 373, 175-190. (36) Konings, C. Clin. Chim. Acta 1988, 176, 185-194. (37) Benson, J.; Hare, P. Proc. Natl. Acad. Sci. U.S.A. 1975, 72, 619-622. (38) Roth, M. Anal. Chem. 1971, 43, 880-882. (39) Carlson, R.; Srinivasachar, K.; Givens, R.; Matuszewski, B. J. Org. Chem. 1986, 51, 3978-3983.
3522 Analytical Chemistry, Vol. 68, No. 19, October 1, 1996
Figure 1. Chromatogram of NDA-CN-derivatized L-arginine and l-norleucine (IS) in Ringer’s lactate solution. Inset: chromatogram of the same reaction with fluorescent detection. λex ) 420 nm, λem ) 490 nm. See text for experimental details and chromatographic conditions.
Figure 2. Chromatogram of NDA-CN-derivatized L-arginine and l-norleucine (IS) in plasma. See text for experimental details and chromatographic conditions.
l-Norleucine was used as an internal standard and similarly derivatized. Calibration curves were generated in Ringer’s lactate solution instead of plasma to correct for the endogenous L-arginine present in plasma. Deproteinization of Plasma. Deproteinization was achieved by the addition of an equal volume of 10% PCA to plasma followed by the separation of the precipitated proteins (pellet) from the supernatant. That there was no significant loss of arginine in the precipitate was ensured by the use of [3H]-L-arginine in plasma and quantitative recovery of the radioactivity counts. The deproteinized supernatant was neutralized with sodium hydroxide (2 M) prior to derivatization. Derivatization Reaction and Chromatography. An HPLC chromatogram of the reaction of L-arginine with NDA and CN in boratebuffered Ringer’s solution is shown in Figure 1. As seen from Figure 1, the retention time of the derivatized arginine is 21.8 min and that of the internal standard is 40.6 min. The peaks at 16.6 and 31.3 min are from the NDA reagent. The integrity of the arginine and norleucine CBI products was ascertained by concurrent fluorescent detection of the derivatization reaction (in Ringer’s solution) at λex ) 420 nm, λem ) 490 nm (Figure 1 inset). Only the NDA-CN-derivatized amino compounds fluoresce at 490 nm when excited at 420 nm.27 A reaction control with no added arginine (in Ringer’s solution) gave no peaks under fluorescent detection at the same excitation and emission wavelengths. Figure 2 shows the chromatogram of derivatized plasma. The derivatized arginine (retention time 21.4 min) and internal standard (retention time 40.8 min) were chromatographically resolved from the other endogeneous, derivatized amino acids in plasma by the use of an acidic mobile phase (20 mM sodium
citrate containing tetramethylammonium chloride, pH 2.8)40 and gradient elution (Experimental Section). The use of a more common phosphate buffer mobile phase, isocratically31 or with gradient elution,41 failed to resolve the derivatized arginine from an overlapping impurity peak. In an attempt to decrease the number of the interfering peaks from plasma, a cation-exchange (Dowex H+) cleanup of plasma was attempted. However, there was no significant change in the chromatographic profile of the derivatized plasma following cation-exchange cleanup. Hence plasma was used without any cleanup. Since the sensitivity of the reaction was quite high, very small volumes of the reaction can be injected on the HPLC column, thus eliminating the danger of any buildup of contaminants from plasma or reagents on the HPLC column. The ratio of the derivatized arginine peak area to the internal standard peak area (peak area ratio or PAR) was used as a measure of the arginine concentration. For estimating arginine amounts in unknown samples, the normalized PAR (PARnorm) for the calibration standards was calculated by dividing the PAR from the standards by the respective arginine concentration [PAR/(µg/mL)]. The mean and coefficient of variation of these normalized PARs was determined. The PAR of the sample divided by the mean PARnorm of the calibration standards gave the concentration of arginine in the unknown samples in micrograms per milliliter.
[Arg] (µg/mL) )
sample PAR standard PARnorm
The stability of the derivatized L-arginine as determined by repeated injection of the same sample over a 20 h time period and comparison of the PARs at 20 h to that at 0.5 h was 98%. This facilitates batch processing of samples and the use of an autosampler for sample analysis. Assay Validation. Plasma contains endogenous arginine, and to correct for this, calibration curves were generated in Ringer’s lactate solution. The curves were linear in the 0-100 µg/mL range, and the coefficient of variation of the means of the normalized PARs of the calibration standards (0-100 µg/mL, n ) 6) was 6.7%. The intraday assay precision (n ) 6) for quantitation of the endogenous and three exogenously added arginine concentrations in control plasma (incremental concentrations of 10, 50, and 250 µg/mL), expressed as a coefficient of variation, was 4.6, 2.9, 2.2, and 4.1%, respectively. Effects due to intersubject variability in the plasma or mode of blood collection (EDTA or heparinized tubes) was investigated by analysis of plasma samples from different subjects (n ) 6) that were collected in heparin or EDTA tubes. The HPLC chromatograms of these samples indicated a similar number of peaks and retention times with the only difference being in peak heights. Thus, any intersubject variability in the plasma composition or the effect of any extraneous chemicals introduced into the plasma from the collection tube did not interfere with the arginine or internal standard peaks in the HPLC chromatograms. The use of Ringer’s solution for the generation of calibration curves was justified on the basis of a recovery study. The percent ratio of the mean normalized PARs in plasma to that in Ringer’s solution (n ) 4) was 103%. This quantitative recovery ratio indicates the absence of any matrix effect and justified the generation of calibration curves in Ringer’s solution. Interassay precision, expressed as a (40) Gilbert, R.; Gonzales, G.; Hawel, L., III; Byus, C. Anal. Biochem. 1991, 199, 86-92. (41) Halawa, I.; Baig, S. Biomed. Chromatogr. 1994, 5, 216-220.
Figure 3. Semilog plot of human plasma L-arginine concentration versus time from a normal volunteer who was administered an oral dose (10 g) and IV infusion (30 g over 30 min) of L-arginine on separate occasions. The arginine concentration in the plasma samples were obtained by following the described assay.
coefficient of variation of the normalized PARs for six calibration curves in blood bank plasma and in Ringer’s solution (0-100 µg/ mL, n ) 6) on six different days, was 7.2 and 6.7%, respectively. Assay Application. The utility of the assay was demonstrated by monitoring the plasma concentration of L-arginine in a normal volunteer subject after an oral dose (10 g) and an IV infusion (30 g over 30 min) administered on separate occasions. Plasma that had been stored at -70 °C, from blood samples drawn from the volunteer at various times in heparinized tubes, was subjected to the described assay procedure. The experimental data are given in Figure 3, which shows the semilog plot of arginine concentration as a function of time after the IV and oral dose. CONCLUSION This article describes an HPLC assay for the quantitation of L-arginine in plasma. The assay involves precolumn derivatization of arginine with NDA-CN followed by HPLC with UV detection. The sensitivity provided by UV detection was sufficient for our application and hence the use of UV detection in the described method. However, the derivatized arginine is also fluorescent, and sensitivities in the low-picomole to femtomole ranges can be achieved with fluorescent detection.27 The method can also be extended to other analogs of L-arginine such as asymmetrical dimethylarginine, which is an endogenous inhibitor of NO synthesis and is known to accumulate in patients with chronic renal failure.42 ACKNOWLEDGMENT Funding for this project was provided by NIH (Grant AG 05627) and (Grant GM 07065). We thank Oranee Tangphao, M.D., for providing the patient samples for analysis. Received for review June 20, 1996. Accepted August 5, 1996.X AC960613N (42) Vallance, P.; Leone, A.; Calver, A.; Collier, J.; Moncada, S. Lancet 1992, 339, 572-575. X Abstract published in Advance ACS Abstracts, September 1, 1996.
Analytical Chemistry, Vol. 68, No. 19, October 1, 1996
3523
