Anal Bioanal Chem (2012) 404:133–140 DOI 10.1007/s00216-012-6099-z
ORIGINAL PAPER
Highly sensitive and selective measurement of underivatized methylmalonic acid in serum and plasma by liquid chromatography-tandem mass spectrometry Chao Yuan & Jessica Gabler & Joe M. El-Khoury & Regina Spatholt & Sihe Wang
Received: 12 March 2012 / Revised: 1 May 2012 / Accepted: 3 May 2012 / Published online: 23 May 2012 # Springer-Verlag 2012
Abstract Methylmalonic acid (MMA) is a functional biomarker of vitamin B12 deficiency. Measurement of plasma MMA is challenging due to its small molecular weight and hydrophilic nature. Several liquid chromatography-tandem mass spectrometry (LC-MS/MS) methods have been developed for measuring plasma MMA. However, these methods involve lengthy sample preparation, long chromatographic run time, inadequate sensitivity, or interference from succinic acid (SA). Here we report a novel LC-MS/MS method for quantitation of underivatized MMA in serum or heparinized plasma with high sensitivity and selectivity. Sample preparation involved only strong anion exchange solid phase extraction. The extract was purified by online turbulent flow and analyzed on an Organic Acids column. MS/ MS analysis was performed in negative electrospray mode, and the analytical time was 6 min. The use of ion ratio confirmation in combination with chromatographic resolution from SA greatly enhanced the selectivity. No interference was observed. This method was linear from 26.2 to
Electronic supplementary material The online version of this article (doi:10.1007/s00216-012-6099-z) contains supplementary material, which is available to authorized users. C. Yuan : J. Gabler : J. M. El-Khoury : R. Spatholt : S. Wang Department of Clinical Pathology, Cleveland Clinic, LL3-129, 9500 Euclid Ave, Cleveland, OH 44195, USA J. M. El-Khoury : S. Wang (*) Department of Chemistry, Cleveland State University, Cleveland, OH 44115, USA e-mail:
[email protected] Present Address: J. Gabler Unity Lab Services, ThermoFisher Scientific, West Palm Beach, FL 33407, USA
26,010.0 nM with an accuracy of 98–111 %. Total coefficient of variation was less than 4.6 % for three concentration levels tested. Comparison with a reference laboratory LCMS/MS method using leftover patient serum specimens (n0 48) showed a mean bias of −2.3 nM (−0.61 %) with a Deming regression slope of 1.016, intercept of −6.6 nM, standard error of estimate of 25.3 nM, and a correlation coefficient of 0.9945. In conclusion, this LC-MS/MS method offers highly sensitive and selective quantitation of MMA in serum and plasma with simple sample preparation. Keywords Plasma . Serum . Methylmalonic acid . Liquid chromatography . Mass spectrometry . LC-MS/MS
Introduction Vitamin B12 (cobalamin) deficiency is an important and often under-recognized problem, particularly in the elderly, that can lead to irreversible neurological damage, anemia, osteoporosis, and cerebrovascular and cardiovascular diseases [1]. Plasma methylmalonic acid (MMA; reference interval, 60–360 nM [2]) is widely accepted as a functional biomarker of vitamin B12 deficiency [3, 4]. A variety of methods have been reported for the measurement of plasma MMA, including capillary electrophoresis [5], high-performance liquid chromatography (HPLC) [6], gas chromatography–mass spectrometry (GC–MS) [7], and liquid chromatography-tandem mass spectrometry (LC-MS/ MS) [8]. The vitamin B12 experts from the National Health and Nutrition Examination Survey (NHANES) encourage the use of LC-MS/MS for the analysis of plasma MMA [9]. The measurement of plasma MMA is confounded by its hydrophilic and non-volatile nature, low molecular weight (118 Da), and isobaric interference from succinic acid (SA)
134
C. Yuan et al.
(Fig. 1) which can have circulating concentrations 100-fold higher than MMA [8, 10]. Several LC-MS/MS methods for the measurement of plasma MMA have been developed [8, 11–16]. However, many of these methods have limitations including tedious sample preparation involving derivatization, lengthy chromatography, or insufficient sensitivity and/ or specificity. Important characteristics of these methods are summarized in Table 1. In this study, our primary aim was to develop an LC-MS/ MS method to measure serum and plasma MMA using a simple sample preparation and a short chromatography with high sensitivity and selectivity.
Materials and methods Chemicals, standards, calibrators, and quality controls Type I water was from a Millipore Synergy System (Billerica, MA, USA). Burdick and Jackson High Purity Solvents (methanol, acetonitrile, isopropanol, and acetone) were purchased from VWR (West Chester, PA, USA). Formic acid (98 % pure) was from EMD Chemicals (Gibbstown, NJ, USA). Ammonium acetate and sodium bicarbonate of ACS grade were from ThermoFisher Scientific (Waltham, MA, USA). MMA (99 %), SA (≥99 %), and bovine albumin (≥96 %) were from Sigma-Aldrich (St. Louis, MO, USA). Saline solution (0.9 % sodium chloride) was from Baxter Healthcare (Deerfield, IL, USA). Deuterium-labeled MMA (d3-MMA, 98 % D) was acquired from Medical Isotopes (Pelham, NH, USA). Bond-Elute strong anion exchange (SAX) SPE cartridges (100 mg bed/1 mL) were from Agilent (Milford, MA, USA). MMA, d3-MMA, and SA stock solutions were prepared in 20 mM sodium bicarbonate water solution. MMA calibrators were prepared in 5 % bovine albumin saline solution at concentrations of 50, 100, 200, 400, and 800 nM. A working internal standard (IS) solution was prepared in 20 mM sodium bicarbonate water solution with a d3MMA concentration of 10 μM. Stock solutions and calibrators were stored at −20 °C until use. A low-level quality control (QC) was a pool of normal patient serum, and a high-level QC was made from spiking the low QC with a concentrated MMA stock solution. QC samples were aliquoted into 1.5-mL fractions and stored at −70 °C until use. O
HO HO
OH
OH O
MMA Fig. 1 Molecular structures of MMA and SA
SA
Five hundred microliters of sample (calibrator, QC, serum, or plasma) was mixed with 500 μL of water and 25 μL of working IS solution. Bond Elute SAX SPE cartridges were conditioned with methanol (2×), 10 M formic acid (2×), methanol (1×), and water (3×). After sample loading, the cartridges were washed with water (2×) and eluted with 125 μL of 18 M formic acid. Ten microliters of the eluate was injected for analysis. LC-MS/MS method A transcend TLX-4 multichannel HPLC system was procured from ThermoFisher Scientific, and the LC channel used for this test consisted of a quaternary loading pump and a binary elution pump. The controlling software was Aria (version 1.6.2). A Cyclone-MAX TurboFlow column (50×0.5 mm, ThermoFisher Scientific) was used for online extraction, and a mixing column (Agilent, Santa Clara, CA, USA) was placed between injector and the TurboFlow column. An Allure® Organic Acids column (150×3.2 mm, 5 μm particle size) from Restek (Bellefonte, PA, USA) was used for chromatographic separation. A diagram of online extraction and LC pluming is shown in Fig. 2. Compositions of mobile phase buffers and chromatography details are listed in Table 2. Total analytical cycle time was 6 min. MS/MS monitoring started at 2 min and ended at 3 min. MS/MS detection was performed on a TSQ Quantum Access triple quadrupole mass spectrometer (ThermoFisher Scientific). Multiple reaction monitoring (MRM) was performed in negative ESI mode. Ion source parameters were composed of a spray voltage of 500 volts, a vaporizer temperature of 350 °C, a sheath gas pressure of 50 units, an auxiliary gas of 5 units, and a capillary temperature of 250 °C. Collision gas pressure was set at 1.5 mTorr. The Q1 and Q3 resolutions were set at 0.7 FWHM (full width at half-maximum). One MRM transition (120.1→76.1) was monitored for d3-MMA, whereas two MRM transitions, a quantifier (117.1→73.1) and a qualifier (117.1→55.2), were monitored for MMA. MMA concentration was calculated using the LCquan software (version 2.6, ThermoFisher Scientific). Range of qualifier-to-quantifier ion ratio used for MMA peak identification was set between 2.5 % and 7.5 % based on tolerance windows defined by the Clinical and Laboratory Standards Institute (CLSI) guidelines C50-A. Method validation
O
O
Sample preparation
Method validation was performed using an in-house developed protocol, which was based on guidelines from CLSI and US Food and Drug Administration (FDA) [17]. Absolute ion suppression was assessed by a post-column infusion experiment [18]. Briefly, a constant flow (5 μL/
Measurement of MMA in serum and plasma by LC-MS/MS
135
Table 1 Summary of reported LC-MS methods for MMA measurement in serum, plasma, and/or urine Reference
Sample preparation
LC run time (min)
Ions monitored (1) MMA (2) IS
LLOQ (nM)
HLOQ (nM)
Lakso et al. [8]
Plasma (0.2 mL) →protein precipitation (PPT) with acidified acetonitrile Plasma (0.1 mL) →ultrafiltration →acidification
10
Quantifier ions (1) 117.2m/z (2) 120.2m/z
90a
200,000
5
Quantifier ions (1) 117→73m/z
No data
2,000b
Plasma, serum or [urine] (1 mL or [0.1 mL+0.9 mL water]) →liquid–liquid extraction (LLE) with acidified methyl tert-butyl ether (MTBE)→derivatization with acidified n-butanol
1
100
150,000b
Blom et al. [11]
Kushnir et al. [12]
(2) 120→76m/z Qualifier ions (1) 231→119m/z (2) 234→122m/z Quantifier ions (1) 231→175m/z (2) 234→178m/z
Schmedes and Brandslund [13]
Magera et al. [14]
Fasching and Singh [15]
Pedersen et al. [16]
a
Plasma or serum (0.75 mL) →automated solid phase extraction (SPE) →derivatization with acidified n-butanol Plasma or [urine] (0.6 mL or [0.1 mL]) →SPE→derivatization with acidified n-butanol Plasma (0.2 mL) →ultrafiltration →acidification Serum (0.025 mL) →LLE with acidified MTBE→derivatization with acidified n-butanol
5.5
Quantifier ions (1) 231→119m/z (2) 234→122m/z
48
200,000b
3
Quantifier ions (1) 231→119m/z (2) 234→122m/z
30c
16,700b
3
Quantifier ions (1) 117→73m/z (2) 120→76m/z
25
3,000
4
Quantifier ions (1) 231→119m/z (2) 234→122m/z
16c
50,000b
Estimated LLOQ based on S/N of 10:1
b
Based on upper limit of linearity as reported by the authors
c
Estimated by converting S/N of a low level sample reported by authors to corresponding concentration with S/N010
min) of d3-MMA (10 μM) was infused into the post-column flow path using a T junction, while processed patient samples (three females and three males) without IS were injected. MRM of d3-MMA was monitored for the entire LC gradient. The obtained chromatograms were qualitatively inspected for obvious ion suppression or enhancement compared to that of 18 M formic acid injection. In order to evaluate whether the 5 % albumin in saline solution could be used as the calibrator matrix and as the diluent for this assay, a mixing study was performed to quantitatively determine the relative matrix effect. Briefly, 5 % albumin in saline was spiked with 800 nM of MMA, and aliquots of this spiked solution were mixed 1:1 with individual patient serum (n06). Spiked 5 % albumin in saline, patient serum, and 1:1 mixed samples were then extracted and analyzed. The measured analyte/IS peak area ratio of each 1:1 mixed sample was compared with the theoretical value which was the mean of the spiked 5 %
albumin in saline and the corresponding patient sample. A difference
