Characterizing Pleiotropic Effects of Glucocorticoids in Mice Using Heavy Water Labeling and Mass Spectrometry

Glucocorticoids are widely prescribed to treat autoimmune and inflammatory diseases. While they are extremely potent, their utility in clinical practice is limited by a variety of adverse effects, including osteoporosis, skin thinning and increased susceptibility to bruising, muscle wasting, fat redistribution, cognitive impairments and mood disorders, and insulin resistance and diabetes. Development of compounds that retain the potent immunomodulating and anti-inflammatory properties of classic glucocorticoids while exhibiting reduced adverse actions is therefore a priority. Much attention has been focused on developing selective glucocorticoid receptor modulators (SGRMs) that would dissociate adverse effects from therapeutic immunomodulatory and anti-inflammatory effects. Here we present several approaches to the field of SGRM development: characterization of a novel putative SGRM (compound L5); characterization of the effects of a glucocorticoid receptor mutant mouse (the GRdim/dim mouse); and comparison of acute vs. chronic effects of glucocorticoids using a model of chronic glucocorticoid excess (the CRH-transgenic mouse). The phenotypic actions of glucocorticoids were measured in these mice by a heavy water labeling strategy combined with mass spectrometry that allows measurement of fluxes through multiple target metabolic pathways concurrently. Pathways monitored included collagen synthesis in bone and skin, protein synthesis in skeletal muscle, triglyceride dynamics in adipose tissue and liver, and proliferation of cells in spleen, pancreatic islets, and hippocampal neuronal precursors. To quantify accumulation or distribution of bone mass, lean tissue mass, and fat mass, we used DEXA and microCT. To quantify changes in glucose homeostasis, we used the deuterated glucose disposal test, a modified oral glucose tolerance test that also reveals glucose uptake and metabolism by peripheral tissues in addition to the standard measurements of glucose and insulin concentrations. We demonstrate for the first time that L5, a compound belonging to a novel class of potential SGRMs, exhibits clearly selective actions on disease-relevant pathways compared to prednisolone. Prednisolone reduced bone collagen synthesis, skin collagen synthesis, muscle protein synthesis, and splenic lymphocyte counts, proliferation, and cell death, whereas L5 had none of these actions. In contrast, L5 was a more rapid and potent inhibitor of hippocampal neurogenesis than was prednisolone; and L5 and prednisolone induced insulin resistance equally. We also demonstrate that, contrary to expectations, some metabolic effects of glucocorticoid treatment are exaggerated in adipose tissue of GRdim/dim mice compared to wildtype mice. The GRdim/dim mouse has a mutation in the dimerization domain of GR, and has been shown to have attenuated transactivation with intact repression. The predominant, current hypothesis is that GRdim/dim mice should experience fewer or less severe effects of glucocorticoids, because the GRdim/dim mutation is a loss of function mutation, from a molecular standpoint. Thus, it would be surprising, a priori, if GRdim/dim mice were to show any exaggerated responses to glucocorticoids. Our labeling studies demonstrate that adipose tissue triglyceride synthesis and de novo lipogenesis exhibit greater stimulation by glucocorticoid treatment in GRdim/dim mice than in wildtype controls. More concordant with the predominant hypothesis were our findings that glucocorticoid-dependent inhibition of bone and skin collagen synthesis and decreases in insulin sensitivity were diminished in GRdim/dim mice compared to wildtype mice. Wildtype and GRdim/dim mice were equally sensitive to glucocorticoid-dependent decreases in muscle protein synthesis. Thus the consequences of mutation in the dimerization domain of the glucocorticoid receptor vary on a pathway-by-pathway basis, and are not restricted to reduced responses to glucocorticoids.Finally, we demonstrated that increased futile cycling between triglycerides and free fatty acids occurs in both abdominal and subcutaneous fat depots in a model of chronic glucocorticoid exposure, CRH-transgenic (CRH-Tg+) mice, and in these mice triglyceride accumulation is favored over net lipolysis. In these studies, we characterized multiple pathways in mice exposed to both chronically and acutely-elevated levels of glucocorticoids (CRH-Tg+ mice and wildtype mice administered glucocorticoids, respectively) to explore how glucocorticoids act on these pathways over time. In bone and skin, fractional collagen synthesis rates were dramatically inhibited by acute glucocorticoid exposure, but in CRH-Tg+ mice, which had reduced bone mass and thinner skin compared to controls, fractional collagen synthesis rates were only modestly lower in bone and are similar in skin compared to controls. In muscle, fractional synthesis of total muscle protein was not dramatically inhibited by either acute or chronic exposure to glucocorticoids, although CRH-Tg+ mice had measurably less muscle mass relative to controls. In adipose tissue there were no remarkable changes in triglyceride dynamics in young mice given glucocorticoids acutely, but in CRH-Tg+ mice we observed more triglyceride synthesis in subcutaneous and abdominal fat depots, more de novo lipogenesis in the abdominal depot only, more triglyceride accumulation in both fat depots and no difference in glyceroneogenesis in either depot compared to controls. Fat accumulation had occurred much more slowly than measured triglyceride synthesis in both depots in CRH-Tg+ mice, thus while gradual accumulation, rather than reduction, of fat is favored, both triglyceride synthesis and lipolysis occurs simultaneously at a high rate. We conclude that futile cyling between triglycerides and free fatty acids occurs at a high rate, in response to chronic, but not acute, glucocorticoid exposure, and that chronic, endogenous glucocorticoid exposure has different metabolic consequences than acute, exogenous administration. We draw several conclusions relevant to novel glucocorticoid agents. First, ligands for the glucocorticoid receptor can exhibit phenotypic selectivity. Second, although L5 may not be a therapeutically attractive candidate, it is a SGRM, and further testing on other compounds of the same class is warranted. Third, dose response ranges vary depending on the ligand, the pathway measured, and the tissue in which it is measured, so a variety of pathways need to be measured in different tissues to accurately establish therapeutic index. Fourth, while the expectation is that SGRMs will activate only a subset of glucocorticoid receptor-dependent pathways, it is important to assess the tissue-specific effects of all potential SGRMs on global transcription and GC-dependent metabolic pathways in many different tissues to accurately establish therapeutic indices of compounds. Fifth, targeting the glucocorticoid receptor dimerization domain pharmacologically is not likely to result in simple diminution of glucocorticoid actions due to attenuated transactivation with intact repression, but may result in stimulation of certain processes. Sixth, there are several advantages of this heavy water labeling method combined with mass spectrometry, in particular that small amounts of drug, subjects, and tissues are required relative to the large amount of information gained. Finally, this method can be adapted to characterize ligands for the glucocorticoid receptor and other nuclear hormone receptors, can be applied to disease models and transgenic mice, and can be integrated into existing clinical studies