Abstract
Hepcidin induction during inflammation is partly due to direct transcriptional regulation by the IL6/STAT3 pathway. However, SMAD1/5/8 signaling is also believed to have a role in hepcidin regulation during inflammation, as inhibitors of BMP type I receptors or the BMP ligand antagonist ALK3-Fc block hepcidin induction, increase iron availability, and ameliorate anemia in different animal models of inflammation.
We previously observed that LPS stimulates liver Smad1/5/8 signaling even in Bmp6-deficient mice and our data suggested that, rather than Bmp6, activin B could be the activating ligand of this pathway during inflammation. There was indeed a dramatic induction of Inhbb mRNA, encoding activin B, in the liver of mice challenged with LPS, slightly preceding an increase in Smad1/5/8 phosphorylation and hepcidin (Hamp) mRNA. In liver cells in vitro, activin B stimulated not only canonical Smad2/3 but also non-canonical Smad1/5/8 signaling and hepcidin expression. Finally, pretreatment with a BMP type I receptor inhibitor showed that the effect of activin B on hepcidin expression in liver cells was entirely attributable to its effect on non-canonical Smad1/5/8 signaling. However, although these data demonstrate that activin B potently crossactivates non-canonical Smad1/5/8 signaling to induce hepcidin expression in hepatocytes in vitro, they do not definitively prove the role of activin B in hepcidin induction in vivo. Therefore, the goal of the present study was to challenge Inhbb-/- mice (deficient in activin B) with LPS or infect them with E. Coli and examine whether, as expected from the in vitro data, the lack of activin B prevents stimulation of both canonical Smad2/3 and non-canonical Smad1/5/8 signaling and induction of hepcidin in these mice.
We first showed that activin B is actually the ligand that in vivo induces hepatic Smad2/3 and Smad1/5/8 phosphorylation in response to inflammatory stimuli such as LPS and bacterial infections. Indeed, these signaling pathways are no longer activated in Inhbb-/- mice (Fig. 1A). Interestingly however, we found that the lack of activin B and, as a consequence, the lack of activation of Smad1/5/8 signaling does not impair the induction of hepatic hepcidin expression by these inflammatory stimuli (Fig. 1B), illustrating the limitations of in vitro studies in simulating what is actually going on inside a liver. In conclusion, although activin B is directly responsible for liver activation of Smad1/5/8 signaling in vivo, this signaling pathway is not governing upregulation of hepcidin production in animals submitted to inflammatory stimuli.
We also noticed that the level of Smad1/5/8 phosphorylation in the liver of mice challenged with LPS is not correlated with the expression of hepcidin. Indeed, although LPS-treated Bmp6-/- and wild-type mice have similar activation of Smad1/5/8 (Fig. 2A), the amount of circulating hepcidin in Bmp6-/- mice is about three times lower than in wild-type mice (Fig. 2B). This could indicate that induction of Smad1/5/8 signaling by inflammatory stimuli takes place in non-parenchymal cells rather than in hepatocytes and has no impact on hepcidin expression. Further investigations are necessary to determine in which liver cells activin B activates the canonical Smad2/3 and non-canonical Smad1/5/8 signaling observed in this study, and what are the exact target genes induced by this signaling.
Disclosures
No relevant conflicts of interest to declare.
In vivo and in vitro oncogenic effects of HIF2A mutations in pheochromocytomas and paragangliomas