Quantification of 15N-Nitrate in Urine with Gas Chromatography

Anal. Chem. 2010, 82, 601–607

Quantification of 15N-Nitrate in Urine with Gas Chromatography Combustion Isotope Ratio Mass Spectrometry to Estimate Endogenous NO Production Els Houben,† Henrike M. Hamer,† Anja Luypaerts,† Vicky De Preter,† Pieter Evenepoel,‡ Paul Rutgeerts,† and Kristin Verbeke*,† Department of Gastrointestinal Research, University Hospitals Leuven and Leuven Food Science and Nutrition Research Centre (LFoRCe), Leuven, Belgium, and Department of Nephrology and Renal Transplantation, University Hospitals Leuven, Leuven, Belgium The use of stable isotope labeled substrates and subsequent analysis of urinary nitrate, forms a noninvasive test for evaluation of the in vivo NO metabolism. The present paper describes a new method for simultaneous quantification of 15N-nitrate and total nitrate with gas chromatography combustion isotope ratio mass spectrometry (GC-C-IRMS). Nitrate, isolated from urine with a nitrate selective resin, was reduced to nitrite using copperized cadmium. Subsequently, Sudan I was formed by diazotation. Sudan II was added as internal standard, and both molecules were analyzed with GCC-IRMS as tert-butyldimethylsilyl derivatives. The accuracy was determined during a recovery study of two different known nitrate concentrations and two 15Nenrichments. A recovery of 101.6% and 103.9% for total nitrate and 107.6% and 91.2% for 15N-nitrate was obtained, respectively. The validated method was applied on complete 72 h urine collections after intravenous administration of 15N-nitrate and 15N-arginine in humans. On average, 51.8% (47.0-71.0%) of administered 15N-nitrate was excreted, while 0.68% (0.44-1.17%) of 15N-arginine was metabolized to nitrate. In conclusion, this method can be used for accurate simultaneous determination of 15N-nitrate and total nitrate concentrations in urine and can be applied in clinical studies for noninvasive evaluation of NO metabolism in vivo. Endogenous nitric oxide (NO) is an important signaling molecule involved in many physiological processes, such as the regulation of the vascular tone, cardiac contractility, intestinal peristalsis, immune function, neurotransmission, and hormonal secretion. NO is formed by oxidation of the terminal guanidino * Corresponding author. Kristin Verbeke, Ph.D., Department of Gastrointestinal Research, University Hospital Leuven, Herestraat 49-3000 Leuven, Belgium. Phone: +32 16 34 43 97. Fax: +32 16 34 43 99. E-mail: Kristin.Verbeke@ uz.kuleuven.ac.be. † Department of Gastrointestinal Research, University Hospitals Leuven and Leuven Food Science and Nutrition Research Centre (LFoRCe). ‡ Department of Nephrology and Renal Transplantation, University Hospitals Leuven. 10.1021/ac9019208  2010 American Chemical Society Published on Web 12/14/2009

N-atoms of L-arginine by the enzyme NO synthase and is rapidly converted into nitrite (NO2-) and nitrate (NO3-). In plasma, nitrite is oxidized to nitrate, which is excreted in urine.1,2 As NO plays a key role in several pathological conditions, including renal failure, cardiovascular disease, diabetes, and gastrointestinal inflammation, there is increasing interest in the assessment of the activity of the L-arginine/NO pathway. Although urinary excretion of nitrate has often been correlated to the in vivo production of NO, urinary nitrate does not exclusively originate from NO conversion but also from dietary nitrate, bacterial nitrate synthesis, and organic nitrate drugs.2 Therefore, quantification of the urinary excretion of 15N-labeled NO3- following intravenous administration of guanidine-L-[15N2] arginine (H215N+d CNH(15NH2)(CH2)3sCH(NH2)COOH) is a more accurate measurement of whole body NO synthesis.1 Previously, different methods have been described for the determination of 15NO3-. A frequently described method is the nitration of aromatic compounds, such as benzene,3 trimethoxybenzene,4 toluene,5 or mesitylene,6 in the presence of sulfonic acid3,4 or trifluoroacetic anhydride5,6 and subsequent analysis by gas chromatography/mass spectrometry (GC/MS). The major drawback of these derivatization procedures is the use of hazardous reagents and the formation of isomers. Tsikas et al.7 have reported a GC/MS method for the simultaneous determination of nitrate and nitrite that involves a derivatization procedure with pentafluorobenzyl bromide (PFB-Br) that results in the formation of PFB-NO2 and PFB-ONO2. However, the formation of PFB-ONO2 was incomplete and was found to proceed considerably slower compared to PFB-NO2. Therefore, quantification of PFB-ONO2 was difficult using gas chromatography combustion isotope ratio (1) Luiking, Y. C.; Deutz, N. E. Curr. Opin. Clin. Nutr. Metab. Care 2003, 6, 103–108. (2) Baylis, C.; Vallance, P. Curr. Opin. Nephrol. Hypertens. 1998, 7, 59–62. (3) Green, L. C.; Wagner, D. A.; Glogowski, J.; Skipper, P. L.; Wishnok, J. S.; Tannenbaum, S. R. Anal. Biochem. 1982, 126, 131–138. (4) Gutzki, F. M.; Tsikas, D.; Alheid, U.; Frolich, J. C. Biol. Mass Spectrom. 1992, 21, 97–102. (5) Smythe, G. A.; Matanovic, G.; Yi, D.; Duncan, M. W. Nitric Oxide 1999, 3, 67–74. (6) Jackson, S. J.; Siervo, M.; Persson, E.; McKenna, L. M.; Bluck, L. J. Rapid Commun. Mass Spectrom. 2008, 22, 4158–4164. (7) Tsikas, D. Anal. Chem. 2000, 72, 4064–4072.

Analytical Chemistry, Vol. 82, No. 2, January 15, 2010

601

mass spectrometry (GC-C-IRMS). Furthermore, Forte et al.8 developed an ammonium diffusion method in which urinary nitrate was converted to ammonia using Devarda’s alloy. Subsequently, the produced ammonia was collected by entrapment on a filter paper disk and the 15N/14N ratio analyzed using a continuous flow gas isotope ratio mass spectrometer (C-IRMS). A separate assay using the Griess reaction was required for quantification of total nitrate.8 Another method for determination of 15NO3- in seawater was described by Preston et al.9 This method required previous reduction of nitrate to nitrite using cadmium columns, followed by a modified Griess reaction, derivatization with N-(tert-butyldimethylsilyl)-N-methyltrifluoroacetamide(MTBSTFA), and analysis using GC/MS. In contrast to seawater, urine samples have been shown to contain unknown compounds that interfere with the Griess reaction.10 In this study, the method described by Preston et al.9 was modified to develop an accurate and reproducible method for the simultaneous determination of total nitrate concentration and 15 N/14N ratio in urine samples using GC-C-IRMS. This method was applied in a pilot study using intravenously administrated 15 N-labeled nitrate and 15N-labeled L-arginine for the determination of NO synthesis in healthy volunteers and, as a proof of principle, one patient with renal insufficiency was included. EXPERIMENTAL SECTION Chemicals and Standards. Lewatit Sybron Ionac SR-7 (15-50 mesh), cadmium (purum p.a. for filling reductors), aniline sulfate (>99%), 2-naphthol (>99%), acetylchloride, methanol, and 4-hydroxyazobenzene (>90%) were obtained from Fluka (Steinheim, Germany). Sodium nitrite (>99.5%), Sudan I (97%), and Sudan II (>96%) were purchased from Sigma Aldrich (Steinheim, Germany); hydrochloric acid 37% (p.a.), formic acid, and sodium nitrate (p.a.) were purchased from Merck (Mu¨nchen, Germany); and citric acid (p.a.), trifluoroacetic acid anhydride, sodium sulfate, and cyclohexane (HPLC-grade) were purchased from Acros Organics (Geel, Belgium). Sodium chloride (reagent grade) was supplied by Biosolve BV (Valkenswaard, The Netherlands), ammonium chloride (98%) by MP Biomedicals (Eschwege, Germany), ethyl acetate (p.a.) by Chemlab NV (Zedelgem, Belgium), MTBSTFA by Pierce (Rockford), sodium hydroxide (p.a.) by Riedel-de Hae¨n AG (Seelze, Germany), copper sulfate (p.a.) by UCB (Leuven, Belgium), sodium [15N]nitrate (15N, 98%+) and L-[15N]arginine (15N, 98%+) by Cambridge Isotope Laboratories (Andover), and ammonium hydroxide 30% by Carlo Erba Reagenti (Milan, Italy). Isotopic reference nitrates with known enrichments (USGS34, 0.3651 atom %; IAEA-NO3, 0.3680 atom %; USGS34, 0.4320 atom %) were purchased from the International Atomic Energy Agency (Vienna, Austria). Standard solutions of nitrate and nitrite (5 mM) were prepared in MQ water (Millipore, Billerica, MA) and stored at 4 °C, whereas standard solutions of Sudan I (5 mM) and Sudan II (5 mM) were prepared in ethyl acetate and stored at -20 °C. (8) Forte, P.; Smith, L. M.; Milne, E.; Benjamin, N. Methods Enzymol. 1999, 301, 92–98. (9) Preston, T.; Bury, S.; McMeekin, B.; Slater, C. Rapid Commun. Mass Spectrom. 1998, 12, 423–428. (10) Tsikas, D.; Gutzki, F. M.; Rossa, S.; Bauer, H.; Neumann, C.; Dockendorff, K.; Sandmann, J.; Frolich, J. C. Anal. Biochem. 1997, 244, 208–220.

602


Safety Considerations. Cadmium and aniline sulfate are highly toxic. Ammonium chloride and MTBSTFA cause irritation to eyes and skin. Inhalation and contact with skin and eyes should be avoided. Copper sulfate, aniline sulfate, 2-naphthol, and Sudan I are (very) toxic to aquatic organisms. Therefore, release into the environment has to be avoided. Sudan I and aniline sulfate have limited evidence of carcinogenic effects. All work should be performed in a well-ventilated fume hood, and suitable protective clothing and gloves should be worn. Analytical Procedure. Isolation of Nitrate. To remove interfering components present in urine, nitrate was isolated using a nitrate selective anion exchange resin. Extract-Clean SPE, 4 mL reservoirs (Grace, Deerfield) were filled with 500 mg of Lewatit Sybron Ionac SR-7 granules and conditioned with 50 mL of 15% (w/v) NaCl. A total of 5 mL of a standard nitrate solution (range 500-3000 nmol nitrate) or 5 mL of urine was acidified to pH 3 with 0.1 M HCl and 1 M HCl, respectively. Samples or standards were passed through the conditioned resin at a flow rate of 0.5-1 mL/min. The nitrate was eluted with 15 mL of 15% NaCl, and the eluate was stored at 4 °C. Fractions of 3 mL (containing 100-600 nmol of nitrate) were used for further analysis. Reduction of Nitrate. Nitrate was reduced to nitrite using cadmium columns. Cadmium was activated with copper as described previously.11,12 Briefly, 42 g of Cd granules (0.3-1.6 mm) were soaked in 100 mL of 1 M HCl for 10 min. A volume of 100 mL of 80 mM CuSO4 was rinsed with MQ water and added to the cadmium granules for 10 min. Extract-Clean SPE columns (Grace) were filled with 7 g of Cu/Cd granules and were immediately used or kept in 0.11 M NH4Cl (pH 8.3) at 4 °C. A volume of 3 mL of the (diluted) isolated urinary nitrate fraction or nitrate standard solutions (range 100-600 nmol) were adjusted to 10 mL with MQ water, and 1.5 mL of 4.7 M NH4Cl (pH 8.5) was added. The aliquots were passed through the Cu/Cd column at a flow rate of 0.5-1 mL/min, washed with 10 mL 0.11 M NH4Cl (pH 8.3), and stored at 4 °C. Sudan I Synthesis. To the reduced nitrate aliquot, 200 µL of 28 mM aniline sulfate in 3 M HCl (pH ≈ 2) was added. After 10 min of incubation at 40 °C, 200 µL of 28 mM 2-naphthol in 3 M NaOH (pH ≈ 8) was added and incubated for another 10 min at 40 °C. Samples were acidified with 130 µL of 1 M citric acid to adjust the pH to 5-6. The formed 1-phenylazo-2-naphthol (Sudan I) was extracted using 3 mL of cyclohexane. The Sudan I concentration was quantified spectrophotometrically (Scientific Evolution 160 UV-vis, Thermo Fisher Scientific Waltham) at 487 nm. GC-C-IRMS Analysis of 15N-Sudan I as tert-Butyldimethylsilyl (TBDMS) Derivatives. Before derivatization with MTBSTFA, 40 µL of 5 mM Sudan II was added to 2 mL of the cyclohexane layer, as an internal standard. The cyclohexane layer, containing Sudan I and Sudan II, was dried and derivatized with 50 µL of MTBSTFA at 75 °C overnight. The Sudan I concentration and nitrogen isotope ratios were measured with an isotope ratio mass spectrometer (IRMS), (Deltaplus XP, Thermo Fisher Scientific) coupled online to a gas chromatograph (GC) (Trace GC, Thermo Fisher Scientific) via a combustion (C) interface (Finnigan GC-C/TC III, Thermo Fisher Scientific). A splitless injection of 1 µL by a CTC CombiPal (11) Morris, A. W.; Riley, J. P. Anal. Chim. Acta 1963, 29, 272–279. (12) Johnston, A. M.; Scrimgeour, C. M.; Henry, M. O.; Handley, L. L. Rapid Commun. Mass Spectrom. 1999, 13, 1531–1534.

autosampler (CTC Analytics, Zwingen; Switzerland) was performed at an injection temperature of 250 °C. Chromatographic separation was achieved with an AT5-MS column (30 m × 0.25 mm i.d., 0.25 µm film thickness (Grace) and with a constant helium flow of 2.3 mL/min. The initial oven temperature of 80 °C was isothermal for 1.75 min, ramped to 210 at 35 °C/min, ramped to 310 at 20 °C/min, and held for 8 min. The separated GC-effluents were combusted to NOx, CO2, SO2, (SiO2)x, and H2O in an oxidation furnace (CuO/NiO/Pt) at 980 °C. Because of the formation of (SiO2)x, the oxidation column had to be renewed every 500 samples. Subsequently, the generated nitrogen oxides were reduced to N2 at 650 °C in a reduction furnace with copper. Water was evacuated using a water permeable membrane (Nafion, Thermo Fisher Scientific), and CO2 was eliminated by a liquid nitrogen trap. Nitrogen gas was lead into the ion source where total N2 was measured as m/z ) 28 (14N14N), 29 (14N15N), and 30 (15N15N) with reference to a calibrated laboratory standard (i.e., N2 gas 5.0, high purity (>99.999 vol %)). Data from the GC-C-IRMS were processed using Isodat NT (version 2.0). Isotope ratios are expressed in atom percent (at %) calculated as follows: 15

at % )

N

14

( N+

15

N)

× 100

Total nitrate concentrations were calculated from the total peak area of Sudan I divided by the total peak area of Sudan II (S1/ S2). Correction for Isotopic Dilution. The diazotation of nitrite with aniline sulfate introduces one additional nitrogen atom (not originating from nitrate) into the corresponding TBDMS-Sudan I resulting in a dilution of the 15N-enrichment. To correct for this dilution, the at % of standard nitrate solutions (100 µg N) with known enrichments varying between 0.363 93 and 0.571 60 at % was measured using a continuous flow elemental analyzer IRMS (C-IRMS; ANCA-2020, Europa Scientific, Crewe, U.K.) and correlated to the at % of the corresponding TBDMS-Sudan I derivatives measured by GC-C-IRMS using Deming regression. Concentrations of urinary 15NO3- were calculated by multiplying total nitrate concentrations with the corrected at %. Method Validation. Recoveries of the Consecutive Steps of the Sample Preparation. The recoveries of the nitrate isolation, nitrate reduction, and Sudan I formation were evaluated spectrophotometrically. For each step, all solutions were prepared and analyzed in triplicate and this was repeated for 4 consecutive days. The recovery of the nitrate isolation was determined by comparing Sudan I formation of nitrate standards (range 100-600 nmol) that were passed over the SR-7 column with nitrate standards. Reduction yields using the Cu/Cd columns were calculated after conversion of reduced nitrate solutions (range 100-600 nmol) and standard nitrite solutions to Sudan I. The yield of the Sudan I synthesis and extraction was determined by comparing standard nitrite solutions (range 100-600 nmol) with a calibration curve of commercially available Sudan I that was treated in the same manner. Finally, the extraction yield of Sudan I in cyclohexane was determined spectrophotometrically by

comparing the absorbance of Sudan I in cyclohexane after extraction of Sudan I with standards of Sudan I in cyclohexane. Limit of Blank, Detection, and Quantification. The limit of blank (LoB), limit of detection (LoD), and limit of quantification (LoQ) were calculated as described by Armbruster and Pry.13 The LoB is the highest apparent analyte concentration expected to be found when replicates of a blank sample containing no analyte are tested; this is defined as meanblank + 1.645 (SDblank). The LoD of total nitrate was the lowest total nitrate concentration that could be reliably distinguished from the blank and is defined as LoB + 1.645 (SDlow concentration sample). The LoQ of total nitrate was defined as the lowest total nitrate concentration that results in a coefficient of variation e20% (CV ) (SD)/(mean) × 100, %). The LoQ of 15N-nitrate is influenced by the LoQ of total nitrate as well as by the accuracy of at % measurements. At low nitrate concentrations, variability in at % measurements increases. To evaluate the variability in 15N/14N measurements at different concentrations, the at % of a nitrate standard with a known enrichment (at % ) 0.374 138) was determined along the entire range of the calibration curve (100-600 µM). The limit of quantification of 15N-nitrate was defined as the lowest total nitrate concentration for which the CV of the at % e 0.5% and the CV of total NO3- e 20%. Calibration Curve. Linearity was studied over a range of concentrations (0-600 µM) of nitrate obtained from a nitrate standard solution in water. Three aliquots of each calibration point were prepared and measured for 2 days as described above. A standard curve was fitted through the mean of each calibration point by the linear regression equation y ) ax + b, where y represents the ratio of the total peak area of Sudan I versus total peak area of Sudan II (S1/S2) and x the nitrate concentration (micromolar). Accuracy, Recovery, and Imprecision. The accuracy of the enrichment measurements was determined by analysis of three isotopic reference nitrates with known enrichments (0.3651 at %, 0.3680 at %, 0.4320 at %). The accuracy, expressed as percentage error, was calculated as follows: calculated at % - known at % × 100 known at % The accuracy of total nitrate and 15N-nitrate concentration was estimated by a recovery study after the addition of two known concentrations of (14N + 15N)-nitrate to the urine matrix (spike 1 and spike 2). Spike 1 contained 250 µM 14NO3- + 0.55 µM 15 NO3-, and spike 2 contained 500 µM 14NO3- + 2.13 µM 15 NO3-. On 20 consecutive days, two aliquots of these samples were derivatized and measured in duplicate as described above. The recovery of added nitrate was calculated as follows: final concentration - initial concentration × 100 added concentration Imprecision was determined by calculation of the interday variability of three concentrations (low, normal, and above-normal), which were prepared and measured in duplicate during 20 days. The interday variability was expressed as CV (%). (13) Armbruster, D. A.; Pry, T. Clin. Biochem. Rev. 2008, 29, S49–S52.


603

Figure 1. Overview of the method for the simultaneous determination of the total nitrate concentration and 15N/14N ratio in urine samples using gas chromatography combustion isotope ratio mass spectrometry (GC-C-IRMS). Below a chromatogram with TBDMS-Sudan I and TBDMSSudan II determined with GC-C-IRMS is presented.

Experimental Design for the Assessment of Human Whole Body NO Production. Six healthy volunteers (3 women, 3 men, median age, 29 (27-40) years) with a normal BMI (21.9 (20.8-23.6)) and one male patient with renal insufficiency (29 years; BMI, 22.3) participated in the pilot study. The ethics committee of the university of Leuven approved the study, and all subjects gave informed consent. Each volunteer underwent two tests separated from each other by at least 1 week. The first test consisted of an intravenous injection of Na[15N]-nitrate (0.25 mg/5 mL of 0.9% NaCl) followed by a complete 72 h urine collection in 5 fractions: 0-6, 6-12, 12-24, 24-48, and 48-72 h for the determination of the total nitrate concentration and the ratio of 15N/14N. The second test consisted of an intravenous injection of L-[15N2]-arginine (100 mg/5 mL of 0.9% NaCl) followed by a complete 72 h urine collection in the same 5 fractions. The day before each administration of the tracer and during the urine collection, volunteers were instructed to consume a diet low in nitrate. A basal urine sample was collected before each intravenous injection. Urine was collected in receptacles to which 750 mg of neomycin was added to prevent bacterial growth. After measurement of the volume, urine samples were filtered through a GF/F filter (Whatman, Royston, U.K.). Urine samples were stored at -80 °C until further analysis. Creatinine content in urine was measured by standard laboratory techniques. 604


Results for 15N in the pilot study were expressed as the percentage of the administered dose of 15N recovered in the 0-6, 6-12, 12-24, 24-48, and 48-72 h urine collections and are presented as the median (range).14 The percentage of administered dose of 15N recovered was calculated as follows:

% dose 15NO3- ) 100 ×

µmol excess 15NO3µmol 15NO3- adm

where µmol excess 15NO3- )

at %t - at %bas nNO-3 100

and at %t is the 15N-nitrate enrichment of a specified urine sample, at %bas is the 15N-nitrate enrichment of a basal urine sample, and nNO-3 is the amount of nitrate (micromoles) in a specified urine sample. RESULTS AND DISCUSSION To determine total nitrate concentration and 15NO 3- enrichment simultaneously in urine, a suitable method was developed (14) Evenepoel, P.; Claus, D.; Geypens, B.; Hiele, M.; Geboes, K.; Rutgeerts, P.; Ghoos, Y. Am. J. Physiol. 1999, 277, G935–G943.

Figure 2. The at % of TBDMS-Sudan I of each point of the calibration curve (100-600 µM) determined by GC-C-IRMS.

Figure 3. Atom % measured by GC-C-IRMS versus C-IRMS to correct for the isotopic dilution. The at % measured by C-IRMS reflects the nitrate enrichment, and the at % measured by GC-C-IRMS reflects the Sudan I enrichment. The measured Sudan I enrichment can be converted to nitrate enrichment using the following regression line obtained by Deming regression: at % (TBDMS-Sudan I) ) 0.4362 at % (nitrate) + 0.207 24 (R 2 ) 0.998).

using GC-C-IRMS (Figure 1). Our method was based on a modified Griess reaction that has previously been applied to the determination of 15NO3- enrichment in seawater.9 Although the Griess reaction has often been used to quantify nitrate in urine,15-20 it has been shown that urine contains varying amounts of unknown compounds that interfere with the Griess reaction.10,21 Indeed, when the modified Griess reaction was applied to urine (15) Forte, P.; Dykhuizen, R. S.; Milne, E.; McKenzie, A.; Smith, C. C.; Benjamin, N. Gut 1999, 45, 355–361. (16) Forte, P.; Ogborn, M. R.; Lilley-Chan, T. Pediatr. Res. 2006, 59, 736–741. (17) Lewicki, J.; Wiechetek, M.; Souffrant, W. B.; Karlik, W.; Garwacki, S. Can. J. Physiol. Pharmacol. 1998, 76, 850–857. (18) Schmidt, R. J.; Baylis, C. Kidney Int. 2000, 58, 1261–1266. (19) Smith, S. D.; Wheeler, M. A.; Lorber, M. I.; Weiss, R. M. Kidney Int. 2000, 58, 829–837. (20) Wagner, D. A.; Schultz, D. S.; Deen, W. M.; Young, V. R.; Tannenbaum, S. R. Cancer Res. 1983, 43, 1921–1925. (21) Tsikas, D. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2007, 851, 51–70.

samples from different subjects, spiked with 500 µM NO2-, recoveries varied from 19-73% suggesting large interindividual differences in interfering compounds. Therefore, an additional cleaning step with a nitrate selective resin was introduced to isolate urinary nitrate. However, the selectivity of both this isolation step and the Griess reaction precluded the use of an internal standard in the sample preparation. Others have used 15NO3- as an internal standard when applying the Griess reaction.7,22 This was impossible in our method since 15NO3- is the analyte of interest. An internal standard (Sudan II) was added prior to derivatization with MTBSTFA and analysis with the GC-C-IRMS to correct for variability in the analysis. A typical chromatogram of TBDMSSudan I and TBDMS-Sudan II is presented in Figure 1. (22) Tsikas, D. Clin. Chem. 2004, 50, 1259–1261.


605

Table 1. Interday Variability of Three Concentrations (Low, Normal, and Above Normal) of Nitrate in Urine with Different 15N Enrichments

NO3-

mean (µM) SD (µM) CV (%)

NO3-

mean (µM) SD (µM) CV (%)

14

15

unspiked

spike 1

spike 2

214.52 21.60 10.1

469.03 38.03 8.1

736.64 79.98 10.9

0.51 0.05 8.9

2.32 0.25 10.6

This method does apply several toxic reagents as mentioned in the safety considerations. In particular, cadmium granules should be handled with caution. However, since the cadmium granules are contained in SPE columns which can be reused several times, continuous exposure is avoided and the contamination risk is limited. The use of these columns allows easy and complete removal of cadmium before further processing of the samples. In addition, unlike benzene which has been applied in previous methods,3 cadmium is not volatile. Method Validation. Recoveries of the Consecutive Steps of the Sample Preparation. The nitrate selective columns yielded a recovery of 88.7% ± 4.9% (n ) 12). The nitrate isolation was not affected by the urine matrix as shown by a recovery of 94.8% ± 3.4% (n ) 3) and 92.3% ± 0.8% (n ) 3) of added nitrate to MQ water and urine, respectively. Reduction of nitrate to nitrite was almost quantitative (96.7 ± 0.6% (n ) 12)). The conversion yield of nitrite to Sudan I in combination with the extraction of Sudan I in cyclohexane was 81.8 ± 3.5% (n ) 12). The extraction yield of Sudan I to cyclohexane was 101.5 ± 1.6% (n ) 9). The overall conversion yield of nitrate to Sudan I amounted to 72.0 ± 4.0%. Limit of Blank, Detection, and Quantification. The LoB and LoD of total nitrate concentration were 26 and 60 µM, respectively. The lowest point in the calibration curve was 100 µM. When only total nitrate is to be determined, an even lower concentration can be used as the lowest point of the calibration curve, since the CV for 100 µM was well below 20%. However, for accurate determination of the 15N enrichment, the contribution of the blank to the at % is significant at concentrations below 100 µM. The at % of TBDMS-Sudan I was constant along the entire concentration range of the calibration curve (100-600 µM) with a mean at % of 0.368 05 ± 0.000 68 (CV ) 0.2%) as shown in Figure 2. The CV of the at % of each nitrate concentration was lower than 0.3%. Therefore, the LoQ of 15N-nitrate was equal to the LoQ of total nitrate: 100 µM total nitrate. These results indicate that determination of the 15N/14N ratio using GC-C-IRMS is very accurate, especially when the total nitrate concentration of the sample is between 100 and 600 µM. Calibration Curve. Linearity was studied over a range of concentrations between 0 and 600 µM of nitrate. Because of the limited solubility of Sudan I in cyclohexane, the calibration curve was not extended to higher concentrations.23 The slope of the calibration curve was 0.002 051 ± 0.000 001 and the intercept was 0.058 ± 0.001. The correlation coefficient was 0.993. For all concentrations, the interday variability was e10%. (23) Abraham, M. H.; Amin, M.; Zissimos, M. Phys. Chem. Chem. Phys. 2002, 4, 5748–42.

606


Figure 4. Cumulative % of the intravenously administered dose of 15 N-nitrate (A) and 15N-arginine (B) recovered in urine versus time in healthy subjects (H1-H6) and one patient with renal failure (P1). (C) Cumulative % of the intravenously administered dose of 15N-arginine recovered in urine corrected for the 15N-nitrate excretion versus time in healthy subjects (H1-H6) and one patient with renal failure (P1).

Correction for Isotopic Dilution. To correct for the isotopic dilution, nitrate standard solutions in water with different at % in a range of 0.363 93-0.571 60% were analyzed as nitrate by C-IRMS and after conversion to Sudan I by GC-C-IRMS. The following equation was obtained using Deming regression: at % (TBDMSSudan I) ) 0.4362 at % (NO3-) + 0.207 24 (correlation coefficient ) 0.998) as shown in Figure 3. Accuracy, Recovery, and Imprecision. The accuracy of the enrichment was determined by analyzing three different known certified nitrate enrichments. The mean accuracy of the different solutions with enrichments of 0.365 65, 0.368 02, and 0.431 95 at %, was 2.6, 2.3, and 2.8%, respectively. The calculated recovery of the spiked urine samples (spike 1, 250 µM 14NO3- + 0.55 µM 15NO3-; spike 2, 500 µM 14NO3- + 2.13 µM 15NO3-) amounted to 101.6% and 103.9% for total nitrate and 91.2% and 107.6% for15N-nitrate, respectively.

The interday variability of three different urine samples containing different total nitrate concentrations and 15N-enrichments (unspiked, spike 1, and spike 2) determined over a period of 20 days is summarized in Table 1. The CV of total nitrate and 15N-nitrate in all samples was below 11%. The variability in the described method is due to the multiple steps of the procedure and the lack of a suitable internal standard for the first steps of the method. To improve the variability, we considered the application of the standard addition method.24 In this method, a sample is measured as such and again after addition of known amounts of the analyte. In this way, a calibration curve is obtained within the sample, and the intercept with the Y-axis provides the sample concentration. This method is mainly useful when matrix problems are encountered since the sample and the sample + additions will experience the same matrix interactions. However, in the current method, the variability is most likely not due to the matrix, since the same variability was observed in water and urine. Results Pilot Study. After validation, the method was applied to the samples obtained from the pilot study in which urinary 15Nnitrate excretion was determined in six healthy volunteers and one patient with renal insufficiency after intravenous administration of sodium 15N-nitrate and guanidine-L-[15N2] arginine. Healthy volunteers had a median creatinine clearance of 123 mL/min (range, 103-153 mL/min), while the patient with renal insufficiency had a creatinine clearance of 43 mL/min. Previous studies have indicated that about 60% of the administered 15NO3 is recovered in urine.20 Therefore, several studies use a correction factor of 1.67 () 100/60) when estimating the arginine-NO conversion.25 We intend to apply the method in patients with impaired kidney function. In this population, a (24) Ito, S.; Tsukada, K. J. Chromatogr., A 2002, 943, 39–46. (25) Castillo, L.; Beaumier, L.; Ajami, A. M.; Young, V. R. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 11460–11465.

general correction factor might not be appropriate. Therefore we determined the urinary 15N-nitrate recovery after administration of 15N-nitrate in a separate test. In healthy volunteers, 51.8% (range 47.0-71.0%) of the intravenously administered 15N-nitrate was excreted in urine during 72 h (Figure 4). These values correspond to the reported value of 60%, yet considerable interindividual variation was observed. In the patient with renal insufficiency only 33.2% of the intravenously administered 15N-nitrate was recovered during the same period. After administration of 15N-arginine, 0.35% (range 0.23-0.66%) was recovered as 15N-nitrate in the urine of healthy volunteers and 0.29% in the patient with renal insufficiency. After correction for the amount of nitrate recovered in urine, it can be calculated that 0.68% (range 0.44-1.17%) and 0.89% of the 15N-arginine was metabolized to nitrate in healthy volunteers and in the patient with renal insufficiency, respectively. Although no clinical conclusions can be drawn from this small number of included subjects, these results are in line with previous studies in healthy individuals, which show that about 0.5-1.2% of arginine is converted to NO.25-27 In conclusion, the described method can be used to accurately determine 15N-nitrate and total nitrate concentrations in urine. This method can be applied in clinical studies to evaluate the in vivo activity of the L-arginine/nitric oxide (NO) pathway noninvasively in different pathological conditions.

Received for review August 25, 2009. Accepted November 30, 2009. AC9019208 (26) Avogaro, A.; Toffolo, G.; Kiwanuka, E.; de Kreutzenberg, S. V.; Tessari, P.; Cobelli, C. Diabetes 2003, 52, 795–802. (27) Forte, P.; Copland, M.; Smith, L. M.; Milne, E.; Sutherland, J.; Benjamin, N. Lancet 1997, 349, 837–842.


607

Quantification of 15N-Nitrate in Urine with Gas Chromatography

Recommend Documents