The results obtained here in int Gr and int Gr

The results obtained here in int-Gr+/+ and int-Gr−/− mice suggested that Gr is required to maintain basal level of Fgf15 expression in the ileum, whereas DEX treatment is able to reduce Fgf15 mRNA in a Gr-independent manner. In previous in vitro studies we have already shown that GC are able to affect the interaction of FXR with a consensus IR-1 response element [1]. An important contribution of the present study concerns the physiological implications of our findings. As, in addition to the insulin/glucagon tandem, the hormonal pair FGF19/FGF21 is involved in the control of energy metabolism, the effect of GC at non-hepatotoxic doses on Cyp7a1 expression suggests that when the mobilization of substrate resources from stores is needed, for instance during fasting, stimulation of de novo synthesis of additional BA is not required, because under these circumstances the enterohepatic circulation is slowed down and there are no available lipids for intestinal digestion/absorption. This effect of FGF21 is consistent with the fact that an increased secretion of FGF21 decreases lipogenesis by inhibition of SREBP-1c [28] and reverses hepatic steatosis [50]. Accordingly, the balanced effect of GC on FGF19 and FGF21 may play a role in disconnecting the hormonal signaling that controls energy supply from that regulating BA synthesis. In a recent study, Zhang et al. have induced AAV-mediated Fgf21 overexpression in mice. In that experimental model, Cyp7a1 was up-regulated and consequently BA pool size was enlarged [51]. The authors concluded that Fgf21-induced changes in BA homeostasis were mainly due to an antagonistic effect on Fgf15 function on liver βKlotho/FGFR4 receptor complex, which resulted in inhibition of Fgf15-mediated Cyp7a1 down-regulation. Our results agree with this hypothesis on the existence of a role of FGF21 in the control of BA homeostasis, but differ in the explanation of the mechanism of action. Model-associated effects can explain this apparent discrepancy. Marked elevation of serum levels of Fgf21 in mice due to Fgf21 overexpression could carry out, as suggested by the authors, a non-physiological whatever between Fgf15 and Fgf21 through their C-terminal domains for the binding site on the co-receptor βKlotho [51]. In general, in most experimental conditions assayed in the present study, reduction in Fgf15 expression was accompanied by Fgf21 up-regulation, which, according to the mechanism proposed by Zhang et al., would result in markedly enhanced Cyp7a1 expression. However, this up-regulation was not found in any of the experimental models used in the present study, except for the case of Fgf15−/− mice. In this strain of mice, GC treatment resulted in Fgf21 up-regulation, and this was not accompanied by down-regulation but a moderate up-regulation of Cyp7a1. The reason for this exception is unclear, which leads to conclude that although others and we have provided important information for the understanding of the mechanisms accounting for Fgf21 interaction with Cyp7a1 regulation, the story is still not clearly cut. Our results from in vitro experiments suggested that FGF21-dependent modulation of CYP7A1 expression occurs by inhibition of the transcriptional activity of CYP7A1 promoter through the autocrine signal mediated by FGF21 released by HepG2 cells and presumably interacting with βKlotho/FGFR1 at the plasma membrane. Because it has been reported that in male hamster, DHP induces Cyp7a1 up-regulation [26], we have used DHP to modify CYP7A1 expression in HepG2 cells. This study showed that DHP indeed affected the transcriptional activity of prCYP7A1 but, at least in our experimental setting, this interaction resulted in an inhibitory rather than stimulatory effect. Moreover, in cells in which prCYP7A1 was inhibited by co-transfection with FGF21, the addition of DHP further reduced Luc2 expression. In conclusion, under non-hepatotoxic conditions, GC can modify the intestine-liver cross-talk by interfering with ileal FXR-FGF19/Fgf15-mediated regulation of hepatic CYP7A1. This involves the enhanced expression of FGF21, which inhibits CYP7A1 promoter activity through an autocrine mechanism. Inversely, as it has been previously reported in animals treated with hepatotoxic doses of GC, in patients with endogenous hypersecretion of GC, such as in Cushing syndrome, or by chronic therapy with GC, these steroids can impair the intestine/liver regulatory crosstalk resulting in enhanced BA levels, which can contribute to aggravate liver damage and favor the alteration in lipid metabolism that is often present in these patients. During the preparation of the revised version of the present manuscript a paper on the role of FGF21 as a negative regulator of BA synthesis has been published [52]. Our results are fully in agreement with the concept of a direct effect of FGF21 and provide additional complementary information on the underlying mechanism as well as on the role in BA homeostasis of the interaction between FXR/FGF19/FGF21 axis and other hormonal signals such as these mediated by GC.