Development of suitable method for large-scale urinary glucocorticoid analysis by liquid chromatography-mass spectrometry.

Abstract

Levels of urinary glucocorticoids and their concentration ratios have been analyzed as potential markers for various pathological statuses. Large-scale studies may possibly accelerate the investigations; however, a suitable method needs to be established. Analytical conditions for measurement of urinary glucocorticoids with LCMS were examined. Electrospray ionization in the positive ion mode was applied for detection of cortisol (precursor>product ion: 363.3>121.0), cortisol-d4 (internal standard, IS, 367.4>121.1), and cortisone (361.2>163.2). To maximize ionization, acetic acid-ammonium acetate buffer (18mM) at pH 5.3 was employed as eluent A. A C18 column (100mm×2.1mm, 2.7μm) at 50°C was used for the 9.5min binary gradient separation starting with 60% eluent A with methanol being eluent B. Linear correlations were observed between the concentrations and the peak areas in the concentration range of 1-300ng/mL with correlation coefficients (r) of 0.998 and 0.997 for cortisol and cortisone, respectively, without IS adjustment, and 0.999 with IS adjustment for both cortisol and cortisone. Solid-phase extraction (SPE) using a 2mL centrifuge column was performed for the urine samples, with the original and final volumes being 100μL. The SPE of 12 urine specimens could be performed within 30min. The effect of the sample matrix on the quantification of endogenous compounds present in the urine extract was limited (coefficient of variation (CV) of IS-adjusted matrix factor: 4.4-8.1%; urine extracts of 8 individuals); however, substantial peak reduction of cortisol was observed at low concentrations. Exogenous contaminants originating from the SPE centrifuge column seemed to be a main cause for this phenomenon because the pure-water extract showed similar peak reduction. A recovery of ∼50% was obtained for both cortisol and cortisone. Adjustment with the IS improved the apparent recovery, with ∼100% being obtained for both cortisol and cortisone. The recovery rate decreased when the urine samples were concentrated in the SPE step; the reduction was greater for cortisol than for cortisone. The lower limit of quantification (LLOQ) was set at 2.5ng/mL when the injection volume was 10μL, based on the reproducibility of the standards which were measured (CV of 12 repetitions: 10.1% for 0.5ng/mL cortisol and 19.6% for 1ng/mL cortisone), the matrix effect (-55% at 2ng/mL concentrations of cortisol), and the recovery rate (∼50%). Furthermore an alternative approach for preparation of the cortisol standards was required for low concentration range (2.5-20ng/mL) because of the effect of the matrix. Degradation of original urine specimens at room temperature was minimal during the first 24h. The extracted urine samples degraded over time; however, their concentrations were corrected with the IS, allowing for analysis up to 5days after extraction. In conclusion, an analytical method for urinary glucocorticoids was established, which is fast, sensitive, and well suited for practical application to large-scale study.