JCI - Modulation of bone morphogenetic protein signaling in vivo regulates systemic iron balance

Selective regulation of hepcidin by BMP/TGF-β superfamily members. TGF-β superfamily members were tested for their ability to regulate hepcidin using both a hepcidin promoter reporter assay (Figure 1A) and quantitative real-time RT-PCR (Figure 1B) in Hep3B hepatoma-derived cells. Relative concentrations of BMP/TGF-β superfamily ligands used are similar to those previously used by others to compare responses among superfamily ligands (23, 30, 31). BMP-2, BMP-4, BMP-5, BMP-6, BMP-7, and BMP-9 robustly increased hepcidin promoter luciferase activity 20- to 100-fold over baseline and increased hepcidin mRNA expression by 160- to 1,100-fold. In contrast, TGF-β1, -β2, and -β3 increased hepcidin expression by only 1.5- to 3-fold over baseline by both methods. BMP-3, BMP-11, GDF-5, GDF-6, and GDF-7 showed no or comparatively little hepcidin induction by both methods. Activin A increased hepcidin promoter relative luciferase activity by 10-fold but increased hepcidin mRNA expression to a lesser extent relative to BMP-2, BMP-4, BMP-5, BMP-6, BMP-7, and BMP-9 as analyzed by real-time RT-PCR. Biologic activity of all ligands was verified by luciferase assay using a BMP-responsive firefly luciferase reporter (30) and a TGF-β/activin–responsive firefly luciferase reporter (31). Results using both methods correlated well with each other, suggesting that the hepcidin promoter luciferase assay is a good surrogate for hepcidin mRNA expression by quantitative real-time RT-PCR. Thus, many TGF-β superfamily members can positively regulate hepcidin expression in vitro; however, BMP-2, BMP-4, BMP-5, BMP-6, BMP-7, and BMP-9 are much more potent regulators of hepcidin compared with other superfamily members, including all 3 TGF-β ligands.

BMP-2 administration in vivo increases hepcidin expression and decreases serum iron. We next investigated whether BMP-2 regulates hepcidin expression and iron metabolism in vivo. Purified BMP-2 at 1 mg/kg was injected retroorbitally into mice, followed by determination of serum iron levels and hepatic hepcidin mRNA expression 4 hours after injection. BMP-2 administration increased hepatic hepcidin mRNA expression 1.8-fold over mice injected with vehicle alone (Figure 2A; P = 0.02). BMP-2 administration also decreased serum iron levels from 170 μg/dl to 114 μg/dl (Figure 2B; P = 0.02). This is consistent with a role for BMP-2 as a positive regulator of hepcidin expression in vivo.

Soluble HJV.Fc selectively inhibits BMP signaling in vitro. Soluble receptors such as the soluble TNF receptor etanercept have previously been used to inhibit ligand activity in vitro and in vivo, presumably by binding to ligands and preventing their interaction with membrane-bound receptors (32). Interestingly, soluble hemojuvelin has been detected in human sera and has been shown to inhibit hepcidin expression in cultured cells, although the mechanism for this inhibition was not investigated (33). We therefore generated purified soluble HJV.Fc (Figure 3A), the murine homolog of which we have previously shown can bind to BMP-2 and BMP-4 ligands (22). We then investigated whether HJV.Fc inhibited basal hepcidin expression and BMP induction of hepcidin expression in vitro. In hepatoma-derived HepG2 cells, which have higher basal hepcidin expression, HJV.Fc inhibited basal hepcidin mRNA expression by 80% (Figure 3B; P = 0.03). These results are consistent with a prior report using soluble hemojuvelin that does not contain an Fc tail (33), suggesting that the Fc tail does not affect the function of soluble hemojuvelin. HJV.Fc also inhibited BMP-2 induction of hepcidin expression (Figure 3C; P = 0.009) and BMP-2–induced activation of the hepcidin promoter in a dose-dependent fashion (Figure 3D). HJV.Fc inhibition of BMP ligands was selective: HJV.Fc inhibited more than 90% of hepcidin promoter activation induced by BMP-2, BMP-4, BMP-5, and BMP-6 but did not inhibit BMP-9 even at lower ligand concentrations (Figure 3D). There was a trend toward low-level inhibition of BMP-7 (Figure 3D).

We have previously demonstrated that BMP-2 and BMP-4 are endogenously expressed in HepG2 cells (22), and we hypothesized that inhibition of these endogenous BMP ligands was the mechanism by which HJV.Fc decreased basal hepcidin expression in HepG2 cells. We therefore used RT-PCR to investigate whether other BMP ligands are endogenously expressed in HepG2. We then tested whether siRNA inhibition of these endogenously expressed BMP ligands inhibited basal hepcidin expression in a manner similar to HJV.Fc. BMP-2, BMP-4, and BMP-6 were endogenously expressed in HepG2 cells, with BMP-4 being the most abundant (Figure 4A). BMP-2, BMP-4, and BMP-6 siRNA each selectively and significantly reduced endogenous ligand expression in HepG2 cells by 65%, 90%, and 55%, respectively, as measured by real-time RT-PCR (Figure 4B). BMP-2, BMP-4, and BMP-6 siRNA each significantly inhibited basal hepcidin expression in HepG2 cells by approximately 10% (P = 0.012), 35% (P = 0.0027), and 15% (P = 0.0026), respectively, as measured by real-time RT-PCR (Figure 4C). As a negative control, neither a control siRNA nor a BMP-7–specific siRNA inhibited basal hepcidin expression. The relative ability of each ligand to inhibit basal hepcidin correlated with the relative mRNA abundance of the ligand and the strength of siRNA inhibition of ligand expression. These data suggest that endogenous BMP-2, BMP-4, and BMP-6 ligands all contribute to basal hepcidin expression in HepG2 cells. These data are consistent with our hypothesis that the mechanism by which HJV.Fc inhibits basal hepcidin expression in these cells is by inhibiting endogenous BMP signaling, presumably by binding and sequestering endogenously produced BMP ligands and preventing their interaction with BMP type I and type II receptors.

Soluble HJV.Fc inhibits hepatic BMP signaling, inhibits hepcidin expression, increases ferroportin expression, mobilizes reticuloendothelial cell iron stores, and increases serum iron in vivo. To test whether HJV.Fc administration could regulate hepcidin expression and iron metabolism in vivo, mice were injected with 25 mg/kg purified HJV.Fc or an equal volume of normal saline by i.p. injection 3 times weekly for 3 weeks. Western blot analysis of liver lysates from these mice showed decreased phosphorylated Smad1, Smad5, and Smad8 expression relative to total Smad1 expression in HJV.Fc-treated mice compared with control mice (Figure 5A; P = 0.0497), demonstrating that HJV.Fc decreases hepatic BMP signaling in vivo. Quantitative real-time RT-PCR analysis revealed a 10-fold decrease in hepatic hepcidin mRNA expression in HJV.Fc-treated mice compared with control mice (Figure 5B; P = 0.003). Consistent with the predicted effects of depressed hepcidin levels to increase ferroportin cell surface expression, increase intestinal iron absorption, and increase release of iron from reticuloendothelial stores, HJV.Fc treatment increased ferroportin expression in the spleen compared with control mice, as measured by western blot (Figure 5C). HJV.Fc treatment also increased serum iron levels from 177 ± 26 μg/dl to 309 ± 2 μg/dl (Figure 5D; P = 0.01) and increased serum transferrin saturation from 70% to 100% (Figure 5E; P = 0.004). Furthermore, HJV.Fc treatment increased hepatic tissue iron content by approximately 2-fold (Figure 5F; P = 0.03) and reduced splenic tissue iron content by almost 60% (Figure 5G; P = 0.009).

Soluble HJV.Fc inhibits IL-6 induction of hepcidin expression. Inflammatory cytokines induce hepcidin expression, and this hepcidin excess is thought to play a role in the anemia of chronic disease (1–6). We therefore investigated whether HJV.Fc could inhibit hepcidin induction by the inflammatory cytokine IL-6. IL-6 increased hepcidin expression 3.3-fold in HepG2 cells as measured by real-time RT-PCR (Figure 6; P = 0.003). Hepcidin induction by IL-6 was significantly abrogated when cells were incubated with HJV.Fc in combination with IL-6 (Figure 6; P = 0.0006 compared with cells treated with IL-6 alone).