Rapid Determination of Orotic Acid in Urine by Liquid Chromatography–Electrospray Tandem Mass Spectrom

Clinical Chemistry 49, No. 3, 2003 Rapid Determination of Orotic Acid in Urine by Liquid Chromatography–Electrospray Tandem Mass Spectrometry, Mohamed S. Rashed,1* Minnie Jacob,1 Mohamed AlAmoudi,1 Zuhair Rahbeeni,2 Moeen A.D. Al-Sayed,2 Lujane Al-Ahaidib,1 Amal A.A. Saadallah,1 and Sharon Legaspi1 (Departments of 1Genetics and 2Medical Genetics, King Faisal Specialist Hospital and Research Centre, Riyadh 11211, Saudi Arabia; * address correspondence to this author at: Metabolic Screening Laboratory, King Faisal Specialist Hospital and Research Centre, PO Box 3354, Riyadh 11211, Saudi Arabia; fax 966-1-442-4546, e-mail rashed@kfshrc.edu.sa) Orotic acid (ORA) is an important biochemical marker for uridine monophosphate synthase deficiency, an autosomal recessive disease characterized by macrocytic hypochromic megaloblastic anemia, growth retardation, orotic aciduria, and crystalluria (1 ). In addition, patients with urea cycle diseases excrete increased amounts of urinary ORA (2 ). Thus, ORA aciduria is observed in patients with ornithine carbamoylasetransferase deficiency (OCTD), an X-linked disorder, and could reveal heterozygosity after a protein load, and in citrullinemia, argininosuccinic aciduria, and argininemia (2, 3 ). Currently there are two widely accepted approaches to the determination of urinary ORA, stable-isotope-dilution gas chromatography–mass spectrometry (GC/MS), and ion-exchange HPLC methods with ultraviolet detection, each of which has difficulties and limitations (4 –7 ). Ito et al. (8 ) recently described a liquid chromatography– electrospray tandem mass spectrometry (LC-ESI-MS/MS) method for a large number of urinary metabolites, including ORA. In the present work we describe a more specific, high-throughput, sensitive stable-isotope-dilution LCESI-MS/MS method for the determination of urinary ORA. ORA was purchased from Sigma, and [1,3-15N2]orotic acid used as internal standard (IS) was obtained from Cambridge Isotope Laboratories. Acetonitrile and glacial acetic used for HPLC were purchased from Fisher Scientific. Syringe-driven membrane filter units were of LCR type (0.45 ␮m pore size) obtained from Millipore. Glass autosampler vials (2-mL capacity) and 300-␮L vial inserts were purchased from Agilent. Urine samples were from pediatric patients investigated for metabolic diseases. Adult urine samples were from healthy laboratory personnel. All urine samples were stored at 20 °C until analysis. MS/MS analysis was carried out on a Micromass QuattroLC triple quadrupole mass spectrometer interfaced with a Z-spray ESI source and equipped with a Jasco HPLC pump Model PU-980 and a Jasco Model AS-950 autosampler. The ESI source was operated in the negative-ion mode with the source kept at 150 °C and a desolvation temperature of 350 °C. The capillary voltage was 3.0 kV, cone voltage was 18 V, extraction voltage was 4 V, and collision energy was 10 eV. Nitrogen was used as the nebulizing gas at a flow rate of 120 L/h and for 499 desolvation at a flow rate of 450 L/h. Argon was used as collision gas at a pressure of 0.2 Pa. The collision-induced dissociation spectra of the deprotonated ORA (m/z 155) and its IS (m/z 157) were acquired in the continuous flow injection analysis mode. Under collision-induced dissociation conditions, both compounds gave one major fragment corresponding to the neutral loss of carbon dioxide [M-H  CO2]. For LC-MS/MS analysis, a short Supelco ABZ-plus column [3.3 cm  2.1 mm (i.d.); 3-␮m particles] was used. The mobile phase consisted of acetonitrile–water– glacial acetic acid (30:70:0.15 by volume) in the isocratic mode at a flow rate of 250 ␮L/min. Chromatography was carried out at ambient temperature. The effluent from the injector was passed through an online filter Model A-315 (Upchurch Scientific) that was connected to the column via PEEK tubing. The column was interfaced directly to the ESI source, with no splitting. For selected-reaction monitoring (SRM) analysis, two transitions were monitored, [M-H] at m/z 155 to product ion at m/z 111 for ORA and m/z 157 to product ion at m/z 113 for the IS. The autosampler was programmed to inject a sample every 5 min. The isotopic purity of the IS was determined by repeated injections of 0.16 mmol/L solution (n  10) under the SRM conditions mentioned above. Peak areas were measured for both transitions, m/z 1553111 for ORA and 1573113 for IS. The averaged data showed that the isotopic purity of the IS was 99.584%; it thus contributed 0.416% to the unlabelled ORA channel, whereas the unlabelled ORA contribution in the IS channel was 0.28%. Urine samples were membrane-filtered before use to remove any particulate material. For samples with creatinine concentrations ⱕ5 mmol/L, we used 25 ␮L, and for samples with creatinine concentrations 5 mmol/L, we diluted the samples 1:1 with water and placed 25 ␮L in microcentrifuge tubes with 20 ␮L of IS solution (1 ␮g) and 300 ␮L of mobile phase. The tubes were vortex-mixed, and 100 ␮L of each of the mixtures was transferred to inserts in glass autosampler vials, capped, and injected. The dilution scheme described was necessary because of signs of overloading as evidenced by peak tailing and/or peak splitting in urine samples with high creatinine concentrations. To prevent overloading of the short, narrow-bore column used, we injected only 3-␮L volumes into the mass spectrometer. Under these conditions, the retention time for ORA was reproducible (CV, 3.4%; mean, 3.02 min; n  50). Some peak tailing was still observed at the high concentration points of the calibration curve and in the abnormal urine samples. This did not affect the quantification of ORA because there were no interfering peaks and because the isotope-labeled IS eluted in the exact retention window, allowing accurate definition of the retention window for peak integration. Shown in Fig. 1A are two overlaid SRM chromatograms for ORA in a pooled control urine (4.65 ␮mol/L; 2 ng injected on column gave a signal-to-noise ratio of 40:1) and for the IS. Fig. 1B shows two overlaid SRM chromatograms for ORA in urine from a male OCTD patient (1570 500 Technical Briefs Fig. 1. LC-MS/MS negative-ion SRM chromatograms for the m/z 1553111 transition for ORA and the m/z 1573113 transition for IS added to urine. (A), pooled pediatric control urine; (B) urine from a male OCTD patient. ␮mol/L) and for the IS. The short ABZ-plus column provided excellent retention of ORA, whereas use of a relatively high acetonitrile concentration in the mobile phase was necessary to enhance ESI signal intensity. The addition of acetic acid to the mobile phase improved peak shapes but decreased the retention time. The decreased retention time occurred only after 50 injections, presumably after saturation of binding sites on the column. This behavior was reported previously for ORA on C18 reversed-phase columns (6 ). No further change was observed over the next 300 injections. To demonstrate the unique retention properties of the ABZ-plus stationary phase for ORA, we analyzed several clinically relevant compounds under the same conditions. Among these were hippuric acid, homovanillic acid, uracil, 5-fluorouracil, and tryptophan. All five compounds eluted before 1.5 min vs the 3.0-min retention for ORA (data not shown). We also observed that the behavior of the column, at least for ORA, was not truly reversed-phase because the retention time was shorter for the ORA and IS calibration solutions in acidified mobile phase containing 200 mL/L acetonitrile than in acidified mobile phase containing 300 mL/L acetonitrile, and was the reverse with 400 mL/L acidified acetonitrile. Calibrators were prepared in pooled control pediatric urine (n  20). Serially diluted solutions of ORA were used to achieve concentrations of 0, 1.282, 3.846, 9.615, 320.50, 961.54, 1282.10, and 2564.20 ␮mol/L. The samples were then treated as described above. A mobile phase blank was injected before and after each calibration curve. The carryover between injections was consistently
0.5%. The peak-area ratio of ORA to IS for the first point in the curve was subtracted from all subsequent points to eliminate the contribution of endogenous ORA. The peak-area ratios were then plotted vs ORA concentration to con- struct calibration curves (y  mx  b). The calculation involved corrections for contribution of IS to the ORA channel and vice versa. We obtained excellent linear calibration curves for the determination of ORA in urine in the range of 1.282–2564 ␮mol/L. Because of the large calibration range, we constructed two calibration plots, one covering the reference interval (up to 20 ␮mol/L) and the other covering exceedingly abnormal ORA values. The calibration results based on the averages of six separate experiments for each of the ranges were as follows: mean (SD) slope, intercept, and correlation coefficient (r2) for the reference interval curve were 0.0036 (0.0002), 0.0006 (0.0004) ␮mol/L, and 0.9996 (range, 0.9991– 0.9998), respectively. The mean (SD) slope, intercept, and correlation coefficient (r2) for the abnormal range curve were 0.0042 (0.0001), 0.0243 (0.0130) ␮mol/L, and 0.9998 (range, 0.9996 – 0.9999), respectively. The intra- and interday accuracy and precision of the method were determined for the control and three known concentrations (9.615, 96.15, 961.5 ␮mol/L) from six experiments run on 6 consecutive days, with interim storage at 4 °C. The average recovery at all concentrations was 101.12% (range, 99.95–101.71%) with intraday CVs of 0.38 –2.5%), and interday CVs of 0.6 –2.4% (n  6 for each concentration). The ORA concentrations in urine from randomly selected pediatric patients with negative blood MS/MS and urine GC/MS results, adult female and male volunteers, and from male patients with OCTD, citrullinemia, or argininosuccinic aciduria are shown in Table 1. We also tested urine from nine patients with propionic acidemia. ORA was detectable in all nine samples from patients with propionic acidemia, with one sample borderline high at 25.54 ␮mol/mmol creatinine (Table 1). Despite of the wide calibration range used, it was necessary to further dilute the urine for 3 of 18 abnormal patient samples because the ORA concentration exceeded the highest point in the curve (2564 ␮mol/L). Our results were in excellent agreement with the stable-isotope GC/MS method reported by McCann et al. (4 ) and with the column-switching ion-exchange HPLC method reported by Ohba et al. (7 ). Table 1. Urine concentration of ORA in controls and patients. ORA, ␮mol/mmol creatinine a Controls 0–1 years 1–12 years Adultsb Patients OCTD Citrullinemia Argininosuccinic aciduria Propionic acidemia a b Number of individuals. 10 males and 10 females. n Mean ⴞ SD 40 18 20 1.51 1.15 0.78 3 5 10 9 1.36 1.52 0.33 1519.7 190.8 1071.8 243.5 855.8 993.2 4.75 7.89 Range 0.05–5.88 0.05–5.98 0.31–1.60 1316–1694 657–1250 22–3271 0.783–25.54 501 Clinical Chemistry 49, No. 3, 2003 In conclusion, we have developed and validated a simple, rapid, and specific method for the determination of ORA in urine. The method can be used for routine screening of hyperammonemia or urea cycle patients and may also be used for heterozygosity testing for OCTD. References 1. Webster DR, Becroft DMO, Suttle DP. Hereditary orotic aciduria and other disorders of pyrimidine metabolism. In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds. The metabolic bases of inherited disease, 7th ed. New York: McGraw-Hill, 1995:1799 – 837. 2. Potter M, Hammond JW, Sim KG, Green AK, Wilcken B. Ornithine carbamoyltransferase deficiency: improved sensitivity of testing for protein tolerance in the diagnosis of heterozygotes. J Inherit Metab Dis 2001;24:5–14. 3. Brusilow SW, Horwich AL. Urea cycle enzymes. In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds. The metabolic bases of inherited disease, 7th ed. New York: McGraw-Hill, 1995:1187–238. 4. McCann MT, Thompson MM, Gueron IC, Tuchman M. Quantification of orotic acid in dried filter-paper urine samples by stable isotope dilution. Clin Chem 1995;41:739 – 43. 5. Brusilow SW, Huaser E. Simple method of measurement of orotic acid and orotidine in urine. J Chromatogr 1989;493:388 –91. 6. Seiler N, Grauffel C, Therrien G, Sarhan S, Knoedgen B. Determination of orotic in urine. J Chromatogr 1994;653:87–91. 7. Ohba S, Kidouchi K, Katoh T, Kibe T, Kobayashi M, Wada Y. Automated determination of orotic acid, uracil and pseudouridine in urine by highperformance liquid chromatography with column switching. J Chromatogr 1991;568:325–32. 8. Ito T, van Kuilenburg ABP, Bootsma AH, Haasnoot AJ, van Cruchten A, Wada Y, et al. Rapid screening of high-risk patients for disorders of purine and pyrimidine metabolism using HPLC-electrospray tandem mass spectrometry of liquid urine or urine-soaked filter paper strips. Clin Chem 2000;46:445– 52. Accuracy of the Rapid Fetal Fibronectin TLi System in Predicting Preterm Delivery, Veronica Luzzi, Kelly Hankins, and Ann M. Gronowski* (Department of Pathology and Immunology, Division of Laboratory Medicine, Washington University School of Medicine, 660 South Euclid Ave., Box 8118, St Louis, MO 63110; * author for correspondence: fax 314-362-1461, e-mail gronowski@ pathology.wustl.edu) Numerous studies have demonstrated that fetal fibronectin (fFN) is an excellent marker of preterm delivery with a negative predictive value (NPV) 99% for predicting delivery within 7 or 14 days in symptomatic women (1–3 ). Interestingly, these studies have been performed by sending the sample to Adeza Biomedical for testing using an ELISA microtiter plate or by a rapid colloidal gold test marketed only outside the US. In the US, however, the only method available for rapid fFN analysis is the TLiTM system (Adeza Biomedical). Our objective was to assess the utility of the TLi rapid fFN system to predict delivery in symptomatic patients within 7, 14, or 21 days. We used 501 cervicovaginal samples consecutively received in the Barnes-Jewish Hospital Laboratory for physician-ordered fFN analysis during an 18-month period for the study (February 2001–August 2002). From this cohort, charts for 243 patients were available for review. This constitutes the group of patients who delivered at Barnes-Jewish Hospital. Study inclusion criteria included the following: signs and symptoms of preterm labor, intact membranes, cervical dilation
3 cm, and fFN collection at 24 –35 weeks. These are all criteria for fFN specimen collection included in the manufacturer’s package insert. Patients who had delivered by cesarean section or other forms of induced delivery within 21 days of fFN were excluded. If fFN measurement was performed more than once, for simplicity we arbitrarily chose the measurement made closest to delivery to be included in the study. A total of 15 patients had more than one fFN measurement performed. fFN measurements were performed according to the manufacturer’s instructions. Briefly, the patient sample was extracted from the Dacron swab into an extraction buffer (provided by the manufacturer), incubated at 37 °C in a water bath for 10 min, and filtered through a plunger filter (provided by the manufacturer). A 200-␮L portion of the filtered sample was applied to the Rapid fFN cassette, and at 20 min the TLi analyzer quantified the intensity of the lines. The instrument then provided a positive or negative result. Institutional Review Board approval was obtained for this study. Of 133 patients who met the inclusion criteria, 38 were positive for fFN and 95 were negative. The mean (SD) maternal ages at delivery for the positive and negative groups were 25.5 (5.8) and 24.1 (6.1) years, respectively (age ranges, 16.2– 40.8 and 12.9 – 41.4 years, respectively). The mean gestational ages at the time of collection for the positive and the negative groups were 29.6 weeks (SD, 2.4; range, 24.1–33.3 weeks) and 30.0 weeks (SD, 2.8; range, 25.3–34.4 weeks), respectively. The mean gestational ages at the time of delivery for the positive and the negative groups were 35.3 weeks (SD, 3.7; range, 24.7–39.8 weeks) and 37.4 weeks (SD, 2.1; range, 30.9 – 40 weeks), respectively. The mean times from collection to delivery for the positive and negative groups were 39.9 days (SD, 22.5; range, 0 –91 days) and 51.4 days (SD, 22.7; range, 2–107 days), respectively. Unpaired t-tests showed that only the gestational age at the time of delivery and the time from collection to delivery were significantly different between the fFN-positive and -negative groups (P  0.003 and 0.017, respectively). The NPV, positive predictive value (PPV), specificity, and sensitivity for predicting delivery within 7, 14, and 21 days of testing are shown in Table 1. We also calculated the likelihood ratio for a positive result [sensitivity/(1  specificity)]; the likelihood ratio was 2.12 for birth within Table 1. NPV, PPV, specificity, and sensitivity of TLi Rapid fFN for predicting delivery in symptomatic patients. Days to delivery Positive fFN, n Negative fFN, n Total, n NPV, % PPV, % Specificity, % Sensitivity, % <7 >7 <14 >14 <21 >21 4 3 7 34 92 126 96.8 10.5 73 57 6 6 12 32 89 121 93.7 15.8 73.6 50 9 6 15 29 89 118 93.7 23.7 75.4 60