Quantitative analysis of 11‐dehydrocorticosterone and corticosterone for preclinical studies by liquid chromatography/triple quadrupole mass spectrometry

Rationale The activity of the glucocorticoid activating enzyme 11β‐hydroxysteroid dehydrogenase type‐1 (11βHSD1) is altered in diseases such as obesity, inflammation and psychiatric disorders. In rodents 11βHSD1 converts inert 11‐dehydrocorticosterone (11‐DHC) into the active form, corticosterone (CORT). A sensitive, specific liquid chromatography/tandem mass spectrometry method was sought to simultaneously quantify total 11‐DHC and total and free CORT in murine plasma for simple assessment of 11βHSD1 activity in murine models. Methods Mass spectrometry parameters were optimised and a method for the chromatographic separation of CORT and 11‐DHC was developed. Murine plasma was prepared by 10:1 chloroform liquid–liquid extraction (LLE) for analysis. Limits of quantitation (LOQs), linearity and other method criteria were assessed, according to bioanalytical method validation guidelines. Results Reliable separation of 11‐DHC and CORT was achieved using an ACE Excel 2 C18‐AR (2.1 × 150 mm; 2 μm) fused core column at 25°C, with an acidified water/acetonitrile gradient over 10 min. Analytes were detected by multiple reaction monitoring after positive electrospray ionisation (m/z 345.1.1 ➔ 121.2, m/z 347.1 ➔ 121.1 for 11‐DHC and CORT, respectively). The LOQs were 0.25 and 0.20 ng/mL for 11‐DHC and CORT, respectively. Conclusions This LC/MS method is suitable for the reliable analysis of 11‐DHC and CORT following simple LLE of murine plasma, bringing preclinical analysis in line with recommendations for clinical endocrinology and biochemistry.

The activity of 11βHSD1 is altered in several disease states including obesity, inflammation and psychiatric disorders. 2 The importance of this enzyme has been revealed by murine models of global or tissue-specific disruption 3,4 or over-expression 5 of the enzyme. Demonstrating the activity of this enzyme is essential in the validation of these murine models and in the understanding of the role of this enzyme in health and disease. The preclinical field is hampered by the lack of a robust assay for the enzyme substrate, 11-dehydrocorticosterone . Antibodies to 11-DHC are not commercially available, and only rarely reported. 6 Immunoassays for CORT have variable cross-reactivity with other endogenous steroids, of which there are many. Similar problems in clinical biochemistry have been overcome by the use of tandem mass spectrometry (MS/ MS); non-selective immunoassays for sex steroids are now no longer acceptable for publication. 7 Furthermore, the advantage of LC/MS analysis is the ability to analyse more than one compound in a sample. In 2005 Ronquist et al 8 detailed a 16-min LC/MS method that measured levels of CORT and 11-DHC in murine liver and adipose; however, this was not applied to blood and used trifluoracetic acid as a modifier in the chromatographic method. A 6min on-line extraction LC/MS method for CORT and 11-DHC analysis has been described and applied to human and rat placenta, 9 but again it has not been applied to blood. Peti et al reported an 11min LC/MS method for CORT, 11-DHC, progesterone, aldosterone, cortisol and cortisone analysis using a high-resolution TripleTOF 5600 mass spectrometer, 10 where the limit of quantitation (LOQ) of CORT and 11-DHC was 3.9 ng/mL. Li et al developed a 6-min LC/ MS method for CORT analysis (LOQ of 1 ng/mL), but not 11-DHC, in mouse plasma. 11 A recent human clinical study by Taylor et al 12 describes a 19.7-min-long LC/MS method for a panel of 13 steroids in serum. The method includes CORT, with an LOQ of 0.25 ng/mL, but 11-DHC is not included in this panel. A 14-min steroid profiling human plasma method includes CORT with an LOQ of 0.5 ng/mL, but not 11-DHC. 13 To date there has not been a validated method focusing only on CORT and 11-DHC in murine plasma.
In this study we have used a murine model of inflammation known to increase 11βHSD1 activity 14 to report a validated, sensitive liquid chromatography/tandem mass spectrometry (LC/MS/ MS)-based method for quantifying plasma levels of the substrate as well as the product of 11βHSD1, 11-DHC and CORT. This method for CORT analysis was sensitive enough to analyse the "free" CORT levels in murine plasma, i.e. the circulating steroid unbound to corticosteroid binding globulin (CBG), which is thought to be the biologically active portion of CORT.

| Stock solutions and calibration standards
Stock solutions (1 mg/mL) of analytes (CORT, 11-DHC) and internal standard (d4-cortisol (d4F) and epi-CORT), in methanol, were stored at −20°C and were further diluted in methanol on the day of use.
Calibration standards were diluted in methanol from 0.01 to 500 ng/ mL on the day of preparation.

| Administration of LPS intra-nasally
Mice were administered LPS (1 mg/mL) intra-nasally at 09:00 h. A control group was left untreated. Mice were culled by decapitation 24 h later and blood collected for steroid analysis.

| Blood collection
Trunk blood from the mice was collected into EDTA-coated tubes, centrifuged (1000 g, 10 min, 4°C) and the plasma transferred to labelled vials and stored at −80°C prior to steroid analysis.

| Preparation of plasma for analysis of free CORT
Plasma (150 μL) was incubated (37°C; 30 min) before being applied to an Ultracel-30 membrane in an Amicon ultra-centrifugal filter unit (Millipore, Livingstone, UK) and subjected to centrifugation (14,000 g, 37°C; 30 min). The ultrafiltrate (150 μL) was subjected to 10:1 chloroform steroid extraction as described in section 2.5.1, and the extract was assessed for CORT levels which represents the free component.

| Reproducibility
The accuracy and precision were determined by assessing calibration standards, following extraction, at the LLOQ and at low, medium and

| Method application
The amounts of total and free CORT and total 11-DHC were quantified in murine plasma. The concentrations of endogenous steroids were compared before and after LPS treatment in two separate mouse experiments; intra-peritoneal injection and intranasal injection in C57BL6 mice.

| Chromatographic development
Due to the mass difference in molecular mass between 11-DHC and CORT of only 2 Da, the natural 2 H and 13 C isotopologues of 11-DHC register signals within the corticosterone MRM transition.
Thus, chromatographic separation of 11-DHC and CORT was essential. Initially, a Sunfire C18 column (100 × 2.1 mm; 3.5 μm; Waters) was trialled for separation of the steroids, but selectivity was improved using an ACE Excel C18-AR column (150 × 2.1 mm; 2 μm; ACT, Aberdeen, UK), due to the longer column length, smaller particle size and a modified stationary phase. This not only afforded improved sensitivity but also robust, reliable chromatographic separation. Under these conditions, the isobaric internal standard, epi-CORT, was resolved temporally (Figure 1), which was also necessary.

| Assay specificity for LLE
No interfering peaks were seen in chromatographic traces of LLE-extracted plasma for the internal standards d4F or epi-CORT.
The quantifier/qualifier ratio of 11-DHC and CORT in six different

| Sensitivity (LOD and LOQ) and linearity
The LOD and LOQ were calculated by extrapolation to be 0.1 ng/mL and 0.25 ng/mL for 11-DHC and 0.10 ng and 0.20 ng/mL for CORT.
Based on regression parameters, d4F was consistently found to be the best internal standard for 11-DHC while epi-CORT was the best internal standard for CORT. This was determined by assessing the widest dynamic range and the best line fit for the calibration curves.
The calibration curves were linear over the range 0.1-500 ng/mL for 11-DHC and CORT. The mean regression coefficients of the standard curves (n = 6) were r 2 = 0.995 ± 0.003 for 11-DHC and r 2 = 0.997 ± 0.002 for CORT, with weighting of 1/x applied for optimal fitting of the lowest amounts. This is an improvement on previously reported analysis of CORT in murine plasma by tandem mass spectrometry (LOQ of 1 ng/mL 11 ). Of importance, their method did not detect or report the amount of 11-DHC in the murine plasma.

| Accuracy, precision, reproducibility
The precision and accuracy were acceptable for 11-DHC and CORT at low, medium and high levels (2.5, 10 and 150 ng/mL) ( Table 2). When applying the LOQs to a volume of 150 μL the levels of 11-DHC and CORT fell comfortably above the LOQ (0.25 ng/mL for 11-DHC and 0.20 ng/mL for CORT) and the upper limit of the assay (500 ng for 11-DHC and CORT).

| Sample stability
The stability of the extracts was acceptable upon short-term storage,

| Method application
The to 15% of the total CORT, following LPS treatment ( Figure 2).

TABLE 2
Recovery and indices of intra-and inter-day precision and accuracy for quantitation of mouse plasma enriched at LOQ (0.25 and 0.20 ng/mL), low (2.5), medium (10) and high (150) levels of 11-DHC and CORT

| CONCLUSIONS
The use of immunoassays in steroid biochemistry is gradually being replaced by chromatographic/mass spectrometric methods in the clinical research arena but this presents additional challenges in the preclinical field due to limited sample volumes from small animals.
Nonetheless, advancing technology is now bringing these within reach allowing improvements in the specificity of biochemical data.
The LC/MS/MS method presented here allows reliable analysis of active and inactive glucocorticoids in plasma from individual animals.
In contrast to previous methods of plasma steroid extraction, some of which used more costly solid-phase extraction (SPE), the present method uses a relatively simple and cost-effective liquid-liquid extraction (LLE) with highly efficient recoveries of the steroid analytes (~90%). It is likely that other steroids will be present within the plasma extract, offering possibilities of broader spectrum data for individual animals, depending upon the research question. Under current conditions, this approach consumed 150 μL of plasma, but with technological advances already available beyond the instrumental specifications described here it will probably lead to gains in this field of preclinical research.