Mass spectrometry imaging of steroids
Cobice, Diego Federico
MetadataShow full item record
Glucocorticoids are steroid hormones involved in the stress response, with a well-established role in promoting cardiovascular risk factors including obesity and diabetes. The focus of glucocorticoid research has shifted from understanding control of blood levels, to understanding the factors that control tissue steroid concentrations available for receptor activation; it is disruption of these tissue-specific factors that has emerged as underpinning pathophysiological mechanisms in cardiovascular risk, and revealed potential therapeutic targets. However, the field is hampered by the inability at present to measure concentrations of steroid within individual tissues and indeed within component cell types. This research project explores the potential for steroid measurements using mass spectrometry-based tissue imaging techniques combining matrix assisted laser desorption ionization with on-tissue derivatisation with Girard T and Fourier Transform Ion Cyclotron Resonance Mass Spectrometry (OTCD-MALDIFTICRMS). A mass spectrometry imaging (MSI) platform was developed and validated to quantify inert substrate and active product (11-dehydrocorticosterone (11DHC), corticosterone (CORT) respectively) of the glucocorticoid-amplifying enzyme 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1) in rodent tissues. A novel approach to derivatising keto-steroids in tissue sections using Girard T reagent was developed and validated. Signals were boosted (10⁴ fold) by formation of GirT hydrazones compared to non-derivatised neutral steroids. Active and inert glucocorticoids were detected in a variety of tissues, including adrenal gland and brain; in the latter, highest abundance was found in the cortex and hippocampus. The MSI platform was also applied to human biopsies and murine tissues for the analysis of other ketosterols such as androgens and oxysterols. Proof-of-principle validation that the MSI platform could be used to quantify differences in enzyme activity was carried out by following in vivo manipulation of 11β-HSD1. Regional steroid distribution of both substrate and product were imaged at 150-200μm resolution in mouse brain sections, and the identification confirmed by collision induced dissociation/liquid extraction surface analysis (CID-LESA). To validate the technique, the CORT/11DHC ratios (active/inert) were determined in 11β- HSD1 deficient mice and found to be reduced (KO vs WT; cortex (49 %*); hippocampus (46 %*); amygdala (57 %)). Following pharmacological inhibition by administration of UE2316, drug levels peaked at 1 h in tissue and at this time point, a reduction in CORT/11DHC ratios were also determined, although to a lesser degree than in KO mice, cortex (22%), hippocampus (25 %) and amygdala (33 %). The changes in ratios appeared driven by accumulation of DHC, the enzyme substrate. In brains of mice with 11β-HSD1 deficiency or inhibition, decreases in sub-regional CORT/11DHC ratio were quantified, as well as accumulation of an alternative 11β- HSD1 substrate, 7-ketocholesterol. MSI data correlated well with the standard liquid chromatography tandem mass spectrometry (LC-MS/MS) in whole brain homogenates. Subsequently, the MSI platform was also applied to measure the dynamic turnover of glucocorticoids by 11β-HSD1 in metabolic tissues using stable isotope tracers (Cortisol-D4 (9,11,12,12-D4) (D4F). D4F was detected in plasma, liver and brain after 6 h infusion and after 48 h in adipose. D3F generation was detected at 6 h in plasma and liver; at 24 h in brain specifically in cortex, hippocampus and amygdala; and at 48 h in adipose. The spatial distribution of d3F generation in brain by MSI closely matched enzyme localisation. In liver, an 11β-HSD1-riched tissue, substantial generation of d3F was detected, with a difference in d4F/d3F ratios compared with plasma (ᴧTTRᴧ 0.18± 0.03 (6 h), 0.27± 0.05 (24 h) and 0.38±0.04 (48 h)). A smaller difference in TTR was also detected between plasma and brain (ᴧTTR 0.09 ± 0.03 (24 h), 0.13±0.04 (48 h)), with no detectable regeneration in adipose. After genetic disruption of 11β-HSD1, d3F generation was not detected in plasma or any tissues, suggesting that 11β-HSD1 is the only enzyme carrying out this reaction. After pharmacological inhibition, a similar pattern was seen. The circulating concentration of drug peaked at 2 h and declined towards 4 h, with same pattern in liver and brain. The ᴧTTR ratios 2HPD between plasma and liver (0.27±0.08vs. 0.45± 0.04) and brain (0.11±0.2 vs. 0.19± 0.04) were smaller following drug administration than vehicle, indicating less d3F generation. Extent of enzyme inhibition in liver responded quickly to the declining drug, with ᴧTTR returning to normal by 4 h (0.38± 0.06). ᴧTTR had not normalised 4HPD in brain (0.12±0.02, suggesting buffering of this pool. In adipose, UE2316 was not detected and nor were rates of d3F altered by the drug. Two possible phase I CYP450 metabolites were identified in the brain differing in spatial distribution. In conclusion, MSI with on-tissue derivatisation is a powerful new tool to study the regional variation in abundance of steroids within tissues. We have demonstrated that keto-steroids can be studied by MALDI-MSI by using the chemical derivatisation method developed here and exemplified its utility for measuring pharmacodynamic effects of small molecule inhibitors of 11β-HSD1. This approach offers the prospect of many novel insights into tissue-specific steroid and sterol biology.