Liquid Chromatography tandem Mass Spectrometry for Routine Measurement of Steroid Hormones in Human Serum

Steroid hormones are physiologically important molecules that are regularly measured in human serum to aid in the diagnosis or follow-up of a variety of conditions. Routine steroid hormone testing is currently often performed on completely automated platforms within clinical laboratories. The reaction principles behind these methods are invariably antibody-based. Although these assays are easy-to-use, some fundamental weaknesses remain such as cross-reactivity, interference from auto- or anti-reagent antibodies, suboptimal performance for low concentrated analytes, and between-platform differences. Cross reactivity is particularly relevant for steroid hormones. Human serum is a pool with multiple circulating structurally resembling steroid hormones, metabolites, or structurally similar drugs. Also, some important steroid hormones are very low concentrated (pico-molar) in relevant patient groups (17β-estradiol in children, postmenopausal women and postmenopausal breast cancer patients taking aromatase inhibitors; 1α,25-dihydroxyvitamin D (vitamin D hormone), and testosterone in women and children). Last, the lipophilic steroid hormones circulate almost completely protein bound in human serum. A complete displacement from the respective binding proteins is necessary to measure total steroid hormone concentrations. This step is often not completely effective performed in automated immunoassays.

Liquid chromatography-tandem mass spectrometry (LC-MS/MS) methods are able to cope with these shortcomings, but routine implementation in clinical laboratories lags behind because of steep capital investment costs, the need for expert technicians, limited service options from vendors, and multiple manual steps. Furthermore, it remains a true challenge to develop LC-MS/MS methods for the low concentrated steroid hormones that do not rely on laborious sample pretreatment protocols, despite the inherent analytical sensitivity and specificity of LC-MS/MS methods.

First, we developed and validated LC-MS/MS assays for the above-mentioned challenging steroids using minimal sample pretreatment. We therefore combined published technical concepts (2-dimensional chromatography, adduct formation, buffer additives) and explored novel concepts (signal summing) to increase sensitivity. Difficulties and opportunities encountered by this simple approach were illustrated by the discovery of a new metabolite of vitamin D that is an isobaric, stereo-isomeric analogue of 1α,25-dihydroxyvitamin D, 1β,25-dihydroxyvitamin D, which has potential clinical relevance. The clinical applicability of these assays was demonstrated in relevant patient groups.

Second, we examined the feasibility and cost-effectiveness of LC-MS/MS assays for routine analyses, performed by non-expert technicians with acceptable performance and turn-around-time (TAT). We therefore developed and validated LC-MS/MS methods for the measurement of 25-hydroxyvitamine D in human serum (18,000 samples/year) and immunosuppressant levels in whole blood (35,000 samples/year, rather stringent turn-around-times in transplantation centers). We automated the protein precipitation sample pretreatment on a liquid handler (LH) equipped with needles for the sampling of whole blood as well as serum. We streamlined interface and data-transfer for non-expert use. Our developed approach allowed non-experts to run LC-MS/MS assays, also outside normal working hours, with acceptable TATs (<4h) and performance. Protein precipitation sample pretreatment was key to achieve a cost-effective transfer from immunoassay to LC-MS/MS and facilitated pretreatment automation. Robustness was still suboptimal for these systems (especially LH) and backup is needed. Verification is needed if manual pipetting is used as backup for LH automated pretreatment. Some differences can be expected even if identical pretreatment protocols are used manually and on LH