= Emerging. More than 5 years before clinical availability. (16.60%, 2024)
= Expected to be clinically available in 1 to 4 years. (37.02%, 2024)
= Clinically available now. (46.38%, 2024)
MSACL 2024 : Farris

MSACL 2024 Abstract

Self-Classified Topic Area(s): Troubleshooting > Tox / TDM / Endocrine > none

Poster Presentation
Poster #1b
Attended on Wednesday at 18:00

Ordeals of Developing a Method to Measure Low-Level Concentrations of Serum Testosterone and my Troubleshooting Journey

Leslie Farris (1), Richard Giles (2), Jessica Colón-Franco (3), Katie Troike (4), Drew Payto (5)
Department of Laboratory Medicine, Cleveland Clinic Foundation, Cleveland, OH, USA

Leslie Farris, B.S. (Presenter)
Cleveland Clinic


1. Problem

We needed to improve sensitivity to measure low concentrations (down to ~1 ng/dL) of total testosterone in serum by LC-MS/MS. Several issues affecting sensitivity and accuracy (to a lesser extent) were encountered throughout method development, as listed below (a-d). Through systematic examination of each potential variable, multiple modifications were implemented during method development, particularly to improve sensitivity.

a. Contamination: Initial injections of neat standards prepared in 50% acetonitrile were linear from 0.5 - 1000 ng/dL, but the low end of the curve was intermittently visible due to high background. Additionally, the baseline of neat injections was increasing over time. We suspected that either testosterone or another analyte with a similar ionization pattern was accumulating on the column or system.
b. Calibration preparation: The calibrators prepared in the candidate matrix generated biased results (>30%) relative to an outside reference lab. The sample comparison showed no bias (≤-5%) with a set of commercial calibrators.
c. Mobile phase modifier: Testosterone areas were higher with more labor-intensive sample preparations such as SPE and LLE, compared to protein precipitation methods. To glean the benefits of LLE without the partitioning time and supernatant transfer we opted for a SLE sample preparation. The cleanliness of the extracts improved the signal to noise ratio of our LLOQ (1 ng/dL), but the absolute area of the quantification peak was still lower than desired (<10,000).
d. Signal loss after sample extraction in plates: The SLE extraction procedure was optimized in cartridges but was switched to plates to improve throughput. A significant loss (3-fold) in signal was observed when the SLE extraction was switched from cartridges to plates, and the collection from glass vials to plastic plates.

2. Method Information (Optimized)
• Sample Extraction
o Sample volume 200uL
o Supported liquid extraction (SLE)
o Elution solvent: 9:1 Hexane:Ethyl acetate
o Concentrate and reconstitute in 200 uL of 60% methanol
• Instrument Information
o Binary Pumping system
 Multiplex over 2-channels
o MS Triple Quadrupole
• Mobile Phase A: 0.2 mM Ammonium Fluoride in LCMS grade water
• Mobile Phase B: 0.2 mM Ammonium Fluoride in LCMS grade (7:3)(Methanol:Acetonitrile)
• 7-minute gradient, 0.5 mL/min flow rate
o Multiplexing run time ~ 5 minutes
• Column: C18 100x3 mm, 2.6 um with guard cartridge (40°C)
• Injection volume 50 uL
• Quantitative MRM acquisition
• Calibration range 1.0-2000 ng/dL in ultra-low testosterone stripped serum

3. Troubleshooting Steps
a. LC Cleaning/flushing and MS Bakeout
b. Evaluation of mobile phase reagents (i.e., water) and equilibration step
c. Assessment of candidate matrix for calibrators and preparation protocol
d. Comparing different sample preparation and mobile phase additives
e. Evaluation of plastic collection plate vs glass vials

4. Outcomes
a. Contamination:
The LC system was thoroughly cleaned/flushed and the MS was bakeout (Step 3a.) to remove any residual contaminants. To investigate further, we injected blank using a double gradient after a prolonged equilibration period at initial conditions (90% aqueous) in mobile phase A prepared using CLRW. The resulting chromatography showed a noticeable increase in the baseline signal at the analyte retention time. We repeated this step using mobile phase A prepared using bottled DI or LCMS grade water (Step 3b). LCMS grade water effectively reduced the background and risk of contamination but not the DI water.

b. Calibration preparation:
We investigated if the candidate matrix was contributing to the bias. Not surprisingly, several commercial serums and prepared alternatives using BSA and PBS screened were found to contain testosterone. We identified a double stripped ultra-low testosterone serum as a suitable candidate matrix for preparing calibrators (Step 3c.). The bias was corrected only after implementing the following steps to prepare in-house calibrators in double stripped ultra-low testosterone: sonicate the CRM in its ampule, dilute CRM in the same carrier solvent it was shipped in, sonicate and rock the high-level calibrator in matrix before preparing additional calibrators levels.

c. Mobile phase modifier:
Replacing the formic acid mobile phase modifier for a low concentration of ammonium fluoride (Step 3d.) resulted in a 3-fold increase in signal.

d. Signal loss after sample extraction in plates:
Through a series of investigations alternating the sample pre-treatment vessel (plastic vs glass), the SLE format (cartridge vs plate) and collection vessel (plastic vs glass), the root cause of the signal loss was attributed to the plastic collection plate. This was rectified by collecting the samples in glass-coated plates (Step 3e.).

Financial Disclosure

Board Memberno
IP Royaltyno

Planning to mention or discuss specific products or technology of the company(ies) listed above: