Reply to Armani et al. Can cortisol stimulate adipogenesis without the glucocorticoid receptor?

The letter by Armani et al.¹ brings up a number of interesting questions that underline the importance of selecting and defining experimental conditions and experimental models for studies of adipogenesis. The nuclear receptors may have to battle out who is the ‘leader’ in specific physiological and pathophysiological situations, but we think it is more likely that they play complementary roles in achieving optimal adipose tissue growth and remodeling, as well as metabolic and endocrine function.

The Caprio group cites several lines of evidence that activation of the mineralcorticoid receptor (MR) enhances adipogenesis, and is required for the adipogenic effect of corticosterone/cortisol in murine and human adipocytes. In contrast, we observed no effect of MR knockdown in human subcutaneous adipocytes, while glucocorticoid receptor (GR) knockdown clearly blocked the proadipogenic effects of cortisol.² Stable GR knockdown in 3T3-L1 preadipocytes (decreased GR protein by over 90%) also prevented differentiation (Figure 1), so a species difference is unlikely. In these experiments, dexamethasone was added during the first 2 days when activation of GR is thought to prime cells for adipogenesis.^3,4 However, in the protocol employed by the Caprio group,⁵ aldosterone or dexamethasone was added during after the induction period, and aldosterone stimulated and dexamethasone inhibited adipogenesis/maturation. Thus, it seems likely that both the timing of GR action and the magnitude of the knockdown are critical for the outcome of these experiments.

Figure 1.

Knockdown of the GR inhibits 3T3-L1 adipogenesis. Stable GR knockdown of 3T3-L1 preadipocyte lines were created with lentivirus delivery of short hairpin RNA (shRNA) targeted to the murine GR (GR shRNA) or a control, non-targeting shRNA (ctrl). After confirming GR knockdown, 3T3-L1 preadipocytes were differentiated following the standard protocol, induction in differentiation cocktail for 2 days (167 nM insulin, 1 μM dexamethasone, 0.5 mM IBMX in 10% fetal bovine serum (FBS) supplemented Dulbecco's modified Eagle's medium (DMEM)) and maintenance in DMEM+10% FBS. Cells were harvested and expression levels of GR and FABP4 (a maker of adipogenesis) were measured in cell lysates prepared before (d0) and after differentiation (d7). Data are from two independent GR shRNA sequences that showed more than 90% reduction in GR at the mRNA (a) and protein levels (b), and repeated three times with similar results.

The experiment shown in the letter by Armani et al.¹ analyzed human visceral (omental) cultures, without prior subculturing, and apparently with cortisol present throughout. Only a minority of visceral stromal cells differentiated into adipocytes in controls, and MR knockdown decreased adipogenesis. The effect of GR knockdown was not assessed, and the reproducibility of this finding in cultures from multiple donors is not clear. Whether the subculturing procedure we employed is the critical factor determining the importance of GR vs MR in adipogenesis also remains to be determined.

A key question is if there are depot differences in the relative importance of MR and GR in adipogenesis. Although MR is expressed at much lower levels (several hundred fold), its expression is higher in omental than subcutaneous adipose tissue (our unpublished observation). Thus, the contribution of glucocorticoid activation of MR pathway might be increased in visceral depot. Another important factor is the local interconversion of cortisone and cortisol. We and others find that mRNA levels and the activity of 11-beta-hydroxysteroid dehydrogenase-2 (HSD2), which inactivates cortisol, is higher in visceral than subcutaneous human adipose tissue.^6,7 By lowering intracellular cortisol, HSD2 may increase the importance of the higher affinity MR in visceral adipocytes. Analyses of the dose-, time-, depot-dependent effects of cortisol via MR and GR on adipogenesis and experiments that address the relevance of each experimental system to in vivo physiology are important avenues for future work.