br Results br Discussion Glucose homeostasis is regulated by

Results
Discussion Glucose homeostasis is regulated by a complex and intricate signaling network involving multiple organs. The BA nuclear receptor FXR is integrated into this regulatory network and participates in glucose handling and metabolism. Intestinal FXR favors glucose Fluconazole and induces FGF15/19 secretion which, through signaling via the hepatic β-Klotho/FGFR4 membrane receptor, inhibits GSK3β, hence increasing glycogenesis. Liver FXR inhibits ChREBP activity, hence decreasing glycolysis and pancreatic FXR potentiates glucose-induced insulin secretion. All these effects contribute to maintain glucose homeostasis in the post-prandial state and led to the prediction that FXR activation could favorably impact on glucose metabolism. Prolonged in vivo activation of FXR by natural or synthetic agonists led to unclear results. Contrasting with cholic acid treatment of C57BL/6J mice (five days), GW4064 treatment did not modulate gluconeogenic gene expression. GW4064 treatment for seven days in C57Bl6 mice increased the expression of gluconeogenic genes without detectable increases in plasma glucose Fasting plasma glucose was increased in high fat diet-fed C57BL/6J mice treated for three months with GW4064 but decreased after a six-week treatment. Thus, long term interference of the FXR signaling pathway by either whole body gene knockout or prolonged agonist treatment did not provide information on a potential role of FXR in the highly dynamic physiological fasting response, which we investigated in this study. In our study, FXR acts positively on the gluconeogenic pathways through two arms. The first positive arm is controlled through the novel glucagon/cAMP/PKA/FXR pathway, which potentiates gluconeogenic gene transcription. This synergy requires PKA-catalyzed phosphorylation of FXR on S325 and S357 and CREB, which co-localizes with FXR at the Fbp1 and Pck1 URRs. ChIP-PCR assays showed that PKA activation correlates with increased FXR DNA binding. Three-dimensional structures of the FXR ligand binding domain bound to natural or synthetic ligands show that S325 is located in helix 7 (H7) which constitutes part of the coactivator LXXLL binding groove and is poorly exposed to solvent in the agonist-bound, coactivator-FXR complex. S357 localizes on the β-loop connecting H7 and H8 and is more accessible to solvent than S325 in this configuration. In light of these structural data, it is still unclear how phosphorylation of S325 and S357 might increase/stabilize DNA binding. Nevertheless, our data add PKA to the growing list of metabolism-sensitive FXR modifiers that includes O-GlcNAc transferase, AMPK, protein kinase C alpha, sirtuin 1 (Sirt1) and p300.[47], [48] A pending question is how these post-translational modifications (PTMs) vary according to the metabolic status. Prolonged energy shortage could lead to selective activation of AMPK and FXR inhibition, whereas PKA would predominantly specify FXR activity in normal fasting conditions. These PTMs have been studied independently and Sirt1-mediated deacetylation and activation of FXR is likely to directly superimpose its regulatory effect on FXR transcriptional activity. As daily variation of protein subcellular localization, phosphorylation and activities[49], [50] are very likely to top on these metabolically-regulated FXR functional alterations, it becomes mandatory to decipher the PTM code of FXR during fasting/feeding periods to fully appreciate how this affects FXR-regulated biological output(s). These outputs might extend beyond metabolic control, as we recently showed that liver FXR may also regulate other specific gene sets and biological pathways as it interacts with other TRs. The prototypical FXR target gene Shp/Nr0b2 controls BA synthesis and lipogenesis and has been proposed to be a negative regulator of gluconeogenesis through interactions with the pro-gluconeogenic glucocorticoid receptor, HNF4α, Foxo-related transcription factors or C/EBPα. This repressive activity provides a direct link between the observed plasma glucose lowering effect of prolonged FXR agonism in mice and gluconeogenic gene transcription. Our data however show that FoxA2 could serve as a repressor of FXR transcriptional activity on a limited number of genes, including Shp/Nr0b2, in short-term fasting conditions which represents another example of signal integration at specific genes. FOXA2-mediated repression of FXR activity proceeds from a DNA binding-independent mechanism, and affects a limited number of genes with no common function, as studied by gene set enrichment analysis or gene ontology term enrichment analysis (data not shown). The repression of Shp gene transcription by FOXA2 is intriguing in light of the ability of SHP to prevent FOXA2 DNA binding in vitro. However, the physiological significance of these findings is unclear, as these investigations were carried out without including hormonal signals such as glucagon, which triggers FOXA2 acetylation and subsequent activation to control fatty acid oxidation and ketogenesis in a process involving Sirt1 and p300. It is important to note here that our mechanistic investigations were carried out at normal glucose concentrations to avoid any confounding effects related to either energy depletion, hence activating Sirt1 and/or AMPK, or to glucose overload, hence activating the hexosamine biosynthetic pathway. A highly integrative approach combining biochemical, proteomic, epigenomic and transcriptomic approaches is required to fully understand PTM-dependent FXR activity variations in the physiologically-varying fasting and fed conditions, to which we now add the glucagon/cAMP pathway as an important regulator of FXR. Whether this physiological mechanism is dysregulated in type 2 diabetes remains to be explored. A decreased activity of both hepatic FOXA2 and FXR through phosphorylation and acetylation, respectively, has been reported in rodent models of T2D,[47], [56] and we observed that exposure of MPHs to glucolipotoxic conditions abolished FXR’s contribution to glucose production (data not shown). However, no correlation between FXR expression and that of its cognate target genes could be established with the diabetic status of human patients (Table S2), suggesting that this novel regulatory pathway is more likely at play in physiological conditions.