Analysis of Testosterone Using the Xevo TQ-XS for Clinical Research

This technology brief demonstrates the high analytical sensitivity and quantitative performance of the Xevo TQ-XS for the analysis of serum testosterone for clinical research using 100 μL serum.

Benefits

• The Xevo TQ-XS provides high analytical sensitivity for the analysis of testosterone from a small sample volume.

Testosterone is synthesized from androstenedione in the adrenal cortex, and is a key biomarker in many clinical research studies. Measurement of testosterone at low concentrations in serum by more traditional immunoassay methods are sometimes problematic due to cross-reactivity of structurally-related steroids (e.g. dihydrotesterone), causing high imprecision and inaccuracy.

LC-MS/MS is an advanced analytical technique that provides high selectivity – differentiating between structurally-similar steroid hormones through selective sample preparation, chromatographic separation, and specific mass detection using Multiple Reaction Monitoring (MRM).  

The Xevo TQ-XS (Figure 1) is a high-performance benchtop tandem quadrupole mass spectrometer expanding the scope of ultimate analytical-sensitivity analysis. It features StepWave XS ion transfer optics delivering enhanced analytical sensitivity, robustness, and reliability. In addition, the new Xtended Dynamic Range (XDR) Detector allows six orders of linear dynamic range. This ensures that methods are easily transferable onto this instrument, and that wide concentration ranges can be measured in a single analysis.

The Xevo TQ-XS provides high analytical sensitivity for the analysis of testosterone from a small sample volume.

Sample preparation and LC-MS/MS analysis

Serum samples (100 µL) were precipitated with methanol and diluted with water prior to centrifugation. Sample supernatant was loaded directly onto the Oasis PRiME HLB µElution Plate. Consecutive washes with 0.1% ammonia in 35% methanol and 0.1% formic acid in 35% methanol were performed. Samples were eluted with 85:15 (v/v) acetonitrile–methanol and diluted with water.  

Samples were subsequently injected on an ACQUITY UPLC I-Class System and Xevo TQ-XS Mass Spectrometer, utilizing a water/methanol/ammonium acetate/formic acid gradient, and an ACQUITY UPLC HSS T3 Column with a VanGuard T3 pre-column. 

MRM transitions monitored for testosterone were m/z 289.2 > 97.0 (quantifier) and 109.0 (qualifier). The internal standard testosterone-¹³C₃ had a MRM transition of m/z 292.2 > 100.0.

To convert conventional mass units (ng/mL) to SI units (nmol/L), multiply by 3.470.

Results

Using a 5 pg/mL solution of testosterone, triplicate injections were performed to demonstrate system analytical sensitivity for the analyte (Figure 2). The Xevo TQ-XS provides reproducible peak areas at this concentration (peak area <3% RSD, n=3).

The low calibration serum standard used for the LC-MS/MS method was assessed on the system over three separate days. This evaluation demonstrates reproducibility of a sample extracted and injected on the Xevo TQ-XS, on three separate occasions (Figure 3).

An accuracy evaluation using European Reference Material (ERM) was performed, demonstrating the excellent quantitative performance of the LC-MS/MS method on the Xevo TQ-XS. ERM DA345a and DA346a were extracted and analyzed, demonstrating method bias within ±3.3% (Figure 4).

Calibration lines (0.005–20 ng/mL) performed on three separate occasions were >0.997 (Figure 5). Total precision of the LC-MS/MS method was assessed over three separate days, with three replicates on each day at 0.02, 0.1, 1, and 14 ng/mL (n=18). Total precision and repeatability was ≤4.4% (Figure 6).

The Xevo TQ-XS has demonstrated excellent quantitative performance for the analysis of testosterone for clinical research, which includes:

• High analytical sensitivity testosterone in stripped serum at 5 pg/mL (0.017 nmol/L)
• Excellent analytical precision (≤4.4%) for testosterone in stripped serum, ranging from 0.02–14 ng/mL
• Excellent agreement for ERM testosterone in serum, demonstrating method bias within ±3.3%